Vascular-Surgery-Books_1aim.net

Complications

in Vascular Surgery

Second Edition, Revised and Expanded

edited by

Jonathan B. Towne
Larry H. Hollier

To Jesse Thompson, M.D. Dr. Thompson was our teacher, mentor, and role model. He

played a significant role during our vascular fellowships and has been a friend and

mentor throughout our careers.

i

2

I

■c

iii c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

Preface to the Second Edition

Under the best of circumstances, confident vascular surgeons will encounter a myriad of
complications in the management of their patients. These problems reflect the complexity
of the surgical care that must be provided to individuals who are usually beyond their sixth
decade and are afflicted with a variety of associated diseases involving major organ
systems, which places them at high risk. In 1980, 1985, and in 1991, we published a volume
entitled Complications in Vascular Surgery. These initial three volumes were produced in
collaboration with Victor M. Bernhard, M.D. Larry H. Hollier, M.D., has helped with the
editorship of the present edition. The most significant changes in vascular surgery since
1991 have to do with the emergence of endovascular treatments and their increasing use in
peripheral vascular surgery. A section on these techniques has been added to this book so
as to make it more timely and helpful to the practicing vascular surgeon. The contributing
authors are experts in their fields and were kind enough to share their experiences. Most
physicians are reluctant to discuss complications; therefore these chapters represent
significant contributions.

We thank Roberta Sutton for her help with editorial review and oversight of
manuscript preparation.

Jonathan B. Towne
Larry H. Hollier

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

Preface to the First Edition

Under the best of circumstances, competent vascular surgeons will encounter a myriad of
complications in the management of their patients. These problems reflect the complexity of
surgical care that must be provided to individuals who are usually beyond their sixth decade
and who are afflicted with a variety of associated diseases involving major organ systems,
which places them at high risk. In 1980, and again in 1985, we asked a group of colleagues to
participate in symposia specifically directed toward the management of vascular surgery
complications. It was our perception that frank, objective, and timely appraisals of these
problems would lead to more opportune and accurate diagnosis and effective therapy and
would point the way to methods of prevention. Since the last symposium, many new
procedures have been added to the armamentarium of the vascular surgeon. These
procedures have in some instances replaced certain forms of therapy or have altered our
perspective in the design of the most appropriate management course. It is well established
that these interventions, as well as the more classic methods of management, may result in a
variety of adverse outcomes due to the nature of the procedure itself or to the manner in
which it is performed. Continued scrutiny of old and new techniques provides the basis for
scientific management of these problems and offers a wider opportunity to develop and "g

explore alternate forms of therapy. &

Six years have passed since the last symposium, and we believed that the time was ripe a

for reassessment of the current status of vascular surgery complications in light of our c

broader understanding of the disease processes encountered and recent advances in <j

technology. As in the past, we have invited experienced senior vascular surgeons whose >9

published experiences identify them as experts in the field. It is our hope that this volume, 4j

which embodies the discussions at our third symposium, will serve as a useful reference for Q

the clinical practitioner. |

S
©

vii f

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

viii PREFACE TO THE FIRST EDITION

We wish to thank Ann Hopkins, who supervised the details of organization and took
major responsibility for overseeing manuscript preparation and editorial review. We
would also like to thank Lynne Mascarella of the Office of Continuing Medical Education
at the University of Arizona College of Medicine for her help in organizing this sym-
posium.

Victor M. Bernhard
Jonathan B. Towne

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

Contents

Preface to the Second Edition v

Preface to the First Edition vii

Contributors xiii

1 . Pitfalls of Noninvasive Vascular Testing 1
Dennis F. Bandyk

2. Cardiopulmonary Complications Related to Vascular Surgery 15
W. Darrin Clouse and David C. Brewster

3. Renal Failure and Fluid Shifts Following Vascular Surgery 49
Gregory S. Cherr and Kimherly J. Hansen

4. Intimal Hyperplasia: The Mechanisms and Treatment of the Response

to Arterial Injury 67

Michael J. Englesbe and Alexander W. Clowes

5. The Healing Characteristics, Durability, and Long-Term Complications

of Vascular Prostheses 83 e

Glenn C. Hunter and David A. Bull e

|

6. Anastomatic Aneurysyms 139 -c

Alexander D. Shepard and Gary M. Jacobson <j

s
1 . Hypercoagulable States and Unexplained Vascular Graft Thrombosis 155 jf

Jonathan B. Towne §

8. Complications and Failures of Anticoagulant and Antithrombotic Therapy 179 |

John R. Hoch @

ix f

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

CONTENTS x

9. Gastrointestinal and Visceral Ischemic Complications of Aortic

Reconstruction 211

Daniel J. Redely and Hector M. Dourron

10. Spinal Cord Ischemia 221
Alfio Carroccio, Nicholas J. Morrissey, and Larry H. Hollier

11. Impotence Following Aortic Surgery 235
Richard Kempczinski

12. Complications Following Reconstructions of the Pararenal Aorta and

Its Branches 249

Kenneth J. Cherry, Jr.

13. Complications of Modern Renal Revascularization 261
Jeffry D. Cardneau and Louis M. Messina

14. The Diagnosis and Management of Aortic Bifurcation Graft Limb

Occlusions 279

Mark T. Eginton and Robert A. Cambria

15. Problems Related to Extra- Anatomic Bypass — Including Axillofemoral,
Femorofemoral, Obturator, and Thoracofemoral Bypasses 293
Kyle Mueller and William H. Pearce

16. Vascular Graft Infections: Epidemiology, Microbiology, Pathogenesis,

and Prevention 305

John M. Draus, Jr., and Thomas M. Bergamini

17. Aortic Graft Infections 317
G. Patrick Clagett

18. Detection and Management of Failing Autogenous Grafts 337
Jonathan B. Towne

19. An Approach to Treatment of Infrainguinal Graft Occlusions
Lloyd M. Taylor, Jr., Gregory J. Landry, and Gregory L. Moneta

20. Wound Complications Following Vascular Reconstructive Surgery
David R. Lorelli and Jonathan B. Towne

21. Complications in the Management of the Diabetic Foot
Gary R. Seabrook

22. Complications of Lower Extremity Amputation
Kenneth E. Mclntyre, Jr.

23. Complications of Vascular Access

Mark B. Adams, Christopher P. Johnson, and Allan M. Roza

24. Complications of Thoracic Outlet Surgery
David Rigberg and Julie A. Freishchlag

355
365
379
401
409
429

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

u

XI

CONTENTS

25. Stroke as a Complication of Noncerebrovascular Surgery 439
Sukru Dilege, Matthew I. Foley, and Gregory L. Moneta

26. Complications of Repair of the Supra-Aortic Trunks and the Vertebral

Arteries 457

Jeffery B. Dattilo and Richard P. Cambria

27. Prevention of Transient Ischemic Attacks and Acute Strokes After
Carotid Endarterectomy: A Critique of Techniques for Cerebrovascular
Protection During Carotid Endarterectomy 467
William H. Baker and Maureen Sheehan

28. Nonstroke Complications of Carotid Endarterectomy 475
Caron Rockman and Thomas S. Riles

Complications of Endovascular Intervention

29. Radiation Exposure and Contrast Toxicity
Evan C. Lipsitz, Frank J. Veith, and Takao Ohki

30. Complications in Peripheral Thrombolysis
Kenneth Ouriel

483
501

Venous Disease

31. Complications of Sclerotherapy

John J. Bergan and Mitchel P. Goldman

32. Complications of Subfascial Endoscopic Perforator Vein Surgery
and Minimally Invasive Vein Harvests

Peter Gloviczki and Manju Kalra

33. Complications of Venous Endovascular Lysis and Stenting
(Iliac, Subclavian)

Peter Neglen

34. Complications of Endovascular Intervention for AV Access Grafts
Abigail Folk

35. Complications of Vena Cava Filters

Enrico Ascher, Anil Hingorani, and William R. Yorkovich

36. Complications of Percutaneous Treatment of Arteriovenous
Malformations

Robert J. Rosen and Thomas Maldonado

Arterial Interventions

511

525

541
557
569

581

i

2

I

■c

37. Endovascular Complications of Angioplasty and Stenting
Gary M. Ansel

38. Complications of Carotid Stenting

Ramtin Agah, Patricia Gum, and Jay S. Yadav

597
615

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

CONTENTS xii

Endovascular Stent-Graft Complication in Aneurysm Repair

39. Endovascular Access Complications 625
Mark A. Farber and Robert Mendes

40. Device Failure 633
Tikva S. Jacobs, Larry H. Hollier, and Michael L. Marin

41. Endoleak 659
Hugh G. Beebe

42. Complications Following Endovascular Thoracic Aortic Aneurysm Repair 683
Alfio Carroccio, Sharif H. Ellozy, and Larry H. Hollier

43. Complications of Angiogenesis Therapy 691
Joshua Bernheim, Sashi Kilaru, and K. Craig Kent

Index 701

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i
2

I

■c

I

u

Contributors

Mark B. Adams, M.D. Professor of Surgery and Chair, Department of Surgery, Medical
College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

Ramtin Agah, M.D. Division of Cardiology and Program in Human Molecular Biology
and Genetics, University of Utah, Salt Lake City, Utah, U.S.A.

Gary M. Ansel, M.D., F.A.C.C. Director, Peripheral Vascular Intervention, Department
of Cardiology, Riverside Methodist Hospital, Columbus, Ohio, U.S.A.

Enrico Ascher, M.D. Professor of Surgery and Director, Division of Vascular Surgery,
Maimonides Medical Center, Brooklyn, New York, U.S.A.

William H. Baker, M.D. Professor Emeritus, Division of Vascular Surgery, Department of
Surgery, Loyola University Medical Center, Maywood, Illinois, U.S.A.

Dennis F. Bandyk, M.D., R.V.T. Professor of Surgery, Division of Vascular and Endo-
vascular Surgery, University of South Florida College of Medicine, Tampa, Florida,
U.S.A.

Hugh G. Beebe, M.D. Director Emeritus, Jobst Vascular Center, Toledo, Ohio, and
Adjunct Professor of Surgery, Dartmouth-Hitchcock Medical Center, Hanover, New
Hampshire, U.S.A.

Thomas M. Bergamini, M.D. Associate Clinical Professor of Surgery, Department of <j

Surgery, University of Louisville School of Medicine, and Surgical Care Associates, >9

Louisville, Kentucky, U.S.A. 4j

D

John J. Bergan, M.D., F.A.C.S., F.R.C.S.(Hon.), Eng. Professor, Department of Surgery, |

University of California, San Diego, School of Medicine, San Diego, California, and Pro- @

xiii f

I

i

2

I

■c

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

xiv CONTRIBUTORS

fessor of Surgery, Uniformed Services University of the Health Sciences, Bethesda, Mary-
land, U.S.A.

Joshua Bernheim, M.D. Vascular Surgery Fellow, Division of Vascular Surgery, New
York Presbyterian Hospital-The University Hospitals of Columbia and Cornell, New
York, New York, U.S.A.

David C. Brewster, M.D. Clinical Professor of Surgery, Massachusetts General Hospital
and Harvard Medical School, Boston, Massachusetts, U.S.A.

David A. Bull, M.D., F.A.C.S. Associate Professor of Surgery, Division of Cardiothoracic
Surgery, University of Utah Health Sciences Center, Salt Lake City, Utah, U.S.A.

Richard P. Cambria, M.D. Chief, Division of Vascular and Endovascular Surgery, and
Visiting Surgeon, Massachusetts General Hospital, and Professor of Surgery, Harvard
Medical School, Boston, Massachusetts, U.S.A.

Robert A. Cambria, M.D. Associate Professor of Surgery, Division of Vascular Surgery,
Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

Jeffry D. Cardneau, M.D. Fellow, Vascular Surgery and Interventional Radiology,
Division of Vascular Surgery, University of California, San Francisco, San Francisco,
California, U.S.A.

Alfio Carroccio, M.D. Assistant Professor, Division of Vascular Surgery, Mount Sinai
School of Medicine, New York, New York, U.S.A.

Gregory S. Cherr, M.D. Assistant Professor, Department of Surgery, State University of
New York at Buffalo, Buffalo, New York, U.S.A.

Kenneth J. Cherry, M.D. Professor of Surgery, Division of Vascular Surgery, Department
of Surgery, Mayo Clinic, Rochester, Minnesota, U.S.A.

G. Patrick Clagett, M.D. Professor and Chairman, Division of Vascular Surgery, and Jan
and Bob Pickens Distinguished Professorship in Medical Science, in memory of Jerry
Knight Rymer and Annette Brannon Rymer and Mr. and Mrs. W. L. Pickens, University of
Texas Southwestern Medical Center, Dallas, Texas, U.S.A.

W. Darrin Clouse, M.D. * Fellow in Vascular Surgery, Massachusetts General Hospital "§

and Harvard Medical School, Boston, Massachusetts, and Assistant Professor of Surgery, &

F. Edward Hebert School of Medicine, Uniformed Services University of the Health a

Sciences, Bethesda, Maryland, U.S.A. -c

Alexander W. Clowes, M.D. Professor and Chief of Vascular Surgery, Department of >9

Surgery, University of Washington, Seattle, Washington, U.S.A. jjj

^Current affiliation: Division of Vascular Surgery, Wilford Hall Medical Center, Lackland AFB,
Lexas, U.S.A.

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

u

CONTRIBUTORS xv

Jeffery B. Dattilo, M.D. Assistant Professor of Surgery, Division of Vascular Surgery,
Vanderbilt University Medical Center, Nashville, Tennessee, U.S.A.

Sukru Dilege, M.D. Assistant Professor of Surgery, Department of General Surgery,
Istanbul Medical Faculty, Istanbul, Turkey

Hector M. Dourron, M.D. * Fellow, Division of Vascular Surgery, Henry Ford Hospital,
Detroit, Michigan, U.S.A.

John M. Draus, Jr. General Surgery Resident, Department of Surgery, University of
Louisville School of Medicine, Louisville, Kentucky, U.S.A.

Mark T. Eginton, M.D. Division of Vascular Surgery, Medical College of Wisconsin,
Milwaukee, Wisconsin, U.S.A.

Sharif H. Ellozy, M.D. Division of Vascular Surgery, Mount Sinai School of Medicine,
New York, New York, U.S.A.

Michael J. Englesbe, M.D. Department of Surgery, University of Washington, Seattle,
Washington, U.S.A.

Abigail Falk, M.D. Assistant Professor, Department of Radiology, Mount Sinai Medical
Center, New York, New York, U.S.A.

Mark A. Farber, M.D. Assistant Professor, Division of Vascular Surgery, Department of
Surgery, University of North Carolina, Chapel Hill, North Carolina, U.S.A.

Matthew I. Foley, M.D. Vascular Surgery Resident, Division of Vascular Surgery, De-
partment of General Surgery, Oregon Health & Science University, Portland, Oregon,
U.S.A.

Julie A. Freishchlag, M.D. William Stewart Halsted Professor and Director, Department
of Surgery, Johns Hopkins School of Medicine, Baltimore, Maryland, U.S.A.

Peter Gloviczki, M.D. Professor of Surgery, Mayo Medical School, Chair, Division of
Vascular Surgery, and Director, Gonda Vascular Center, Department of Surgery, Mayo
Clinic, Rochester, Minnesota, U.S.A.

1
Mitchel P. Goldman, M.D. Associate Professor of Dermatology, University of California, 8

San Diego, School of Medicine, San Diego, California, U.S.A. ■§,

Patricia Gum, M.D. Department of Cardiovascular Medicine, The Cleveland Clinic
Foundation, Cleveland, Ohio, U.S.A.

"Current affiliation: Vascular Surgeon, Vascular Surgical Associates, P.C., Austell, Georgia, U.S.A.

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

xvi CONTRIBUTORS

Kimberly J. Hansen, M.D. Professor of Surgery and Head, Section of Vascular Surgery,
Division of Surgical Sciences, Wake Forest University School of Medicine, Winston-Salem,
North Carolina, U.S.A.

Anil Hingorani, M.D. Associate Professor of Surgery and Attending, Division of Vascular
Surgery, Maimonides Medical Center, Brooklyn, New York, U.S.A.

John R. Hoch, M.D. Associate Professor, Section of Vascular Surgery, Department of
Surgery, University of Wisconsin Medical School, Madison, Wisconsin, U.S.A.

Larry H. Hollier, M.D., F.A.C.S., F.A.C.C, F.R.C.S. (Eng) Professor of Surgery and
Dean, Louisiana State University Health Sciences Center School of Medicine, New Orleans,
Louisiana, U.S.A.

Glenn C. Hunter, M.D., F.R.C.S.(E), F.R.C.S.C. Professor, Department of Surgery,
University of Texas Medical Branch, Galveston, Texas, U.S.A.

Tikva S. Jacobs, M.D. General Surgery Resident, Division of Vascular Surgery, Mount
Sinai School of Medicine, New York, New York, U.S.A.

Gary M. Jacobson, M.D. * Fellow in Vascular Surgery, Henry Ford Hospital, Detroit,
Michigan, U.S.A.

Christopher P. Johnson, M.D. Professor of Surgery, Department of Transplant Surgery,
Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

Manju Kalra, M.D. Gonda Vascular Center, Department of Surgery, Mayo Clinic,
Rochester, Minnesota, U.S.A.

Richard Kempczinski, M.D. Professor Emeritus, Department of Surgery, University of
Cincinnati, Cincinnati, Ohio, U.S.A.

K. Craig Kent, M.D. Chief, Division of Vascular Surgery, New York Presbyterian Hos-
pital-The University Hospitals of Columbia and Cornell, New York, New York, U.S.A.

Sashi Kilaru, M.D. Vascular Surgery Fellow, Division of Vascular Surgery, New York
Presbyterian Hospital-The University Hospitals of Columbia and Cornell, New York, New
York, U.S.A.

Gregory J. Landry, M.D. Assistant Professor of Surgery, Division of Vascular Surgery,
Oregon Health & Science University, Portland, Oregon, U.S.A.

Evan C. Lipsitz, M.D., F.A.C.S. Assistant Professor of Surgery, Division of Vascular
Surgery, Department of Surgery, Montefiore Medical Center and Albert Einstein College of
Medicine, Bronx, New York, U.S.A.

*Current affiliation: Vascular Surgical Associates, P.C., Marietta, Georgia, U.S.A.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

CONTRIBUTORS xvii

David R. Lorelli, M.D., R.V.T. Attending Vascular Surgeon, Department of Vascular
Surgery, St. John Hospital and Medical Center, Detroit, Michigan, U.S.A.

Thomas Maldonado, M.D. Fellow, Peripheral Vascular Surgery, Department of Surgery,
New York University Medical Center, New York, New York, U.S.A.

Michael L. Marin, M.D. Professor and Chief, Division of Vascular Surgery, and Henry
Kaufman Professor of Vascular Surgery, Mount Sinai School of Medicine, New York, New
York, U.S.A.

Kenneth E. Mclntyre, Jr., M.D. Professor of Surgery and Chief, Division of Vascular
Surgery, University of Nevada School of Medicine, Las Vegas, Nevada, U.S.A.

Robert Mendes, M.D. Division of Vascular Surgery, Department of Surgery, University of
North Carolina, Chapel Hill, North Carolina, U.S.A.

Louis M. Messina, M.D. Professor of Surgery, Division of Vascular Surgery, Edwin J.
Wylie Endowed Chair in Surgery, and Vice Chair, Department of Surgery, University of
California, San Francisco, San Francisco, California, U.S.A.

Gregory L. Moneta, M.D. Professor and Chief, Division of Vascular Surgery, Department
of General Surgery, Oregon Health & Science University, Portland, Oregon, U.S.A.

Nicholas J. Morrissey, M.D. Assistant Professor, Division of Vascular Surgery, Mount
Sinai School of Medicine, New York, New York, U.S.A.

Kyle Mueller, M.D. Surgical Resident, Division of Vascular Surgery, The Feinberg School
of Medicine, Northwestern University, Chicago, Illinois, U.S.A.

Peter Neglen, M.D., Ph.D. River Oaks Hospital, Jackson, Mississippi, U.S.A.

Takao Ohki, M.D., Ph.D. Chief of End ovascular Programs, Division of Vascular Surgery,
Department of Surgery, Montefiore Medical Center, and Associate Professor of Surgery,
Albert Einstein College of Medicine, Bronx, New York, U.S.A.

Kenneth Ouriel, M.D. Chairman, Department of Vascular Surgery, The Cleveland Clinic
Foundation, Cleveland, Ohio, U.S.A.

1
William H. Pearce, M.D. Violet R. and Charles A. Baldwin Professor of Vascular Surgery, &

Division of Vascular Surgery, The Feinberg School of Medicine, Northwestern University, a

Chicago, Illinois, U.S.A. t

Daniel J. Reddy, M.D., F.A.C.S. D. Emerick and Eve Szilagyi Chair in Vascular Surgery, «

Department of Surgery, Henry Ford Hospital, Detroit, Michigan, U.S.A. -|

D

David Rigberg, M.D. Assistant Professor of Vascular Surgery, Department of Surgery, §

UCLA Medical Center, Los Angeles, California, U.S.A. |

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

xviii CONTRIBUTORS

Thomas S. Riles, M.D. The George David Stewart Professor and Chair, Department of
Surgery, New York University School of Medicine, New York, New York, U.S.A.

Caron Rockman, M.D., F.A.C.S. Assistant Professor of Surgery, Department of Surgery,
New York University School of Medicine, New York, New York, U.S.A.

Robert J. Rosen, M.D. Director, Interventional Radiology and Endovascular Surgery,
Department of Radiology, New York University Medical Center, New York, New York,
U.S.A.

Allan M. Roza, M.D. Professor, Division of Transplant Surgery, Department of Surgery,
Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

Gary R. Seabrook, M.D. Professor of Surgery, Division of Vascular Surgery, Medical
College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

Maureen Sheehan, M.D. Chief Surgical Resident, Division of Vascular Surgery, Depart-
ment of Surgery, Loyola University Medical Center, Maywood, Illinois, U.S.A.

Alexander D. Shepard, M.D. Senior Staff Surgeon, Department of Surgery, and Medical
Director, Vascular Lab, Henry Ford Hospital, Detroit, Michigan, U.S.A.

Lloyd M. Taylor, Jr., M.D. Professor of Surgery, Division of Vascular Surgery, Oregon
Health & Science University, Portland, Oregon, U.S.A.

Jonathan B. Towne, M.D. Professor of Surgery and Chairman, Division of Vascular
Surgery, Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

Frank J. Veith, M.D. Professor and Vice Chairman, The William J. von Liebig Chair in
Vascular Surgery, Department of Surgery, Montefiore Medical Center and Albert Einstein
College of Medicine, Bronx, New York, U.S.A.

Jay S. Yadav, M.D., F.A.C.C., F.S. C.A.I. Director, Vascular Intervention, Department of
Cardiovascular Medicine, The Cleveland Clinic Foundation, Cleveland, Ohio, U.S.A.

William R. Yorkovich, R.P.A. Physician Assistant, Division of Vascular Surgery, Maimo-
nides Medical Center, Brooklyn, New York, U.S.A.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

1

Pitfalls of Noninvasive VascularTesting

Dennis F. Bandyk

University of South Florida College of Medicine, Tampa, Florida, U.S.A.

Diagnostic testing of patients with vascular disease requires a thorough understanding of
the instrumentation, arterial and venous anatomy, and hemodynamics of blood circulation.
Although physical examination and vascular imaging studies — such as contrast arteriog-
raphy, magnetic resonance angiography, and contrast-enhanced computed tomography —
are indispensable in the management of peripheral vascular disease, noninvasive vascular
testing retains a prominent role in patient evaluation both prior to and following inter-
vention. Methods that use Doppler ultrasound — in particular duplex ultrasonography —
and plethysmography form the cornerstone of noninvasive vascular testing. Testing is
used to detect and grade the severity of cerebrovascular, peripheral arterial, and peripheral
venous disease and thereby assists in disease management. The accuracy of duplex ultra-
sonography coupled with indirect physiological testing methods is superior to clinical
evaluation alone and in many patients provides sufficient anatomical information of disease
extent and severity to proceed with surgical or endovascular intervention without con-
firmatory imaging studies. Newer enhancements of duplex ultrasonography — such as power
Doppler angiography, sonographic composite imaging, and the use of contrast agents —
have further improved anatomic resolution, which allows better characterization of the
extent and morphology of vascular disease.

Noninvasive vascular testing methods measure biophysical properties of the circulation
(e.g., pressure, pulse contour, blood-flow velocity, turbulence-disturbed flow); these mea- -a

surements can be used for disease localization and classification of severity. The diagnostic %

accuracy of each technique depends on the precision and reproducibility of the measure- J

ments. The measurements of blood pressure and velocity recorded by vascular laboratory !§

instrumentation should not always be assumed to be accurate, since testing can be affected ^

by a number of factors, including biological variability, instrumentation employed for test- &

ing, the skill and bias of the examiner, the pathological process being studied, and conditions §

under which the measurements are recorded. Interpretation of noninvasive testing studies q

requires an appreciation of the limitations, pitfalls, and artifacts associated with the various |

diagnostic techniques. Errors in any of these areas can result in an incorrect diagnosis and «

£
1 f

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

2 BANDYK

subsequent clinical decision making. In this chapter, the common pitfalls of noninvasive
vascular testing related to instrumentation, testing protocols, and diagnostic interpretation
are reviewed and techniques to minimize their occurrence are outlined.

I. CLASSIFICATION OF THE PITFALLS OF NONINVASIVE TESTING

Diagnostic errors associated with noninvasive vascular testing can result from procedural,
interpretative, or statistical pitfalls. Procedural pitfalls can be due to the instrumentation,
to deviations from the testing protocol, or to biological variability of the measurement.
Interpretative pitfalls can decrease diagnostic accuracy when the measurement or physician's
interpretation does not agree with a recognized "gold standard," such as arteriography.
Interpretative errors can occur despite the recording of a precise and reproducible measure-
ment. For example, the use of velocity criteria to grade internal carotid artery stenosis not
previously validated by comparison with angiographic results can cause consistent over-
classification of disease severity.

The accuracy of diagnostic testing is commonly expressed in terms of descriptive sta-
tistical parameters such as sensitivity, specificity, and positive and negative predictive values
(PPV and NPV). These parameters are useful to compare the diagnostic accuracy between
different testing methods or threshold criteria, but they are subject to statistical pitfalls
introduced by the reliability of the gold standard used for correlation as well as disease
prevalence and clinical status (symptomatic, asymptomatic) of the study population. For
example, the sensitivity (ability to detect the presence of a disease state) and specificity
(ability to recognize the absence of a disease state) of a specific test are not affected the
prevalence of the particular disease state in the study population used to calculate diagnostic
accuracy. But predictive values (PPV, NPV), which are better measures of clinical usefulness
in the management of an individual patient, are highly dependent on disease prevalence. It
is recommended that testing with the highest sensitivity be used to screen patients to rule out
a particular disease state (1). For an individual patient, a test with a high ( >90%) NPV can
be used to exclude the disease state. Similarly, testing with high specificity and PPV should
be used to confirm the presence of disease or to proceed with intervention — i.e., performing
carotid endarterectomy based on duplex ultrasonography. It is impossible to eliminate all
sources of measurement error associated with noninvasive vascular testing because of inher-
ent variability in instrumentation, anatomy, and the examination technique. Despite this
caveat, noninvasive vascular diagnostics, in particular duplex ultrasonography, demon-
strate sufficient accuracy to permit medical, surgical, or endovascular intervention with a
high degree of clinical certainty.

The noninvasive vascular laboratory typically employs a number of different instru-
ments for testing in the areas of cerebrovascular, peripheral arterial, peripheral venous,
and abdominal visceral disease. The most widely used instrumentation relies on ultra- ■§

sound to interrogate vessels for patency and flow. Duplex ultrasonography is necessary g

instrumentation in an accredited vascular laboratory, and the clinical applications of this a

technique are numerous in the evaluation cerebrovascular, peripheral venous, peripheral c

arterial, and visceral vascular disease (Table 1). Plethysmographic instruments, such as the <

pulse volume recorder or photoplethysmograph, coupled with the pressure transducer >9

(aneroid manometer), provide indirect hemodynamic measurements of peripheral arte- J

rial and venous flow and pressure. Compared with duplex ultrasound, these indirect tech- «

niques are vulnerable to their own unique diagnostic pitfalls. |

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PITFALLS OF NONINVASIVE VASCULAR TESTING 3

Table 1 Clinical Applications of Duplex Ultrasonography

Cerebrovascular Testing

Extracranial carotid and vertebral artery ultrasonography

Duplex detection of plaque morphology

Transcranial Doppler/duplex ultrasonography
Peripheral arterial testing

Lower/upper limb arterial duplex mapping

Infrainguinal graft surveillance

Evaluation of stent-graft aortic aneurysm exclusion

Dialysis access graft/fistula surveillance
Peripheral venous testing

Diagnosis of lower/upper limb deep venous thrombosis

Duplex evaluation of venous reflux — chronic venous insufficiency

Preoperative saphenous/arm vein mapping
Visceral vascular testing

Renal duplex ultrasonography

Mesenteric duplex ultrasonography

Evaluation of renal/liver organ transplants

II. PROCEDURAL PITFALLS
A. Instrumentation

Instrumentation in good operating condition and calibration is required for noninvasive
vascular testing. Measurements of limb blood pressure (ankle-brachial systolic pressure
index, or ABI), treadmill walking time, pulsatility index, and duplex-acquired velocity spec-
tral parameters (peak systolic velocity, end-diastolic velocity) demonstrate reproducibility
comparable to that of other common clinical (pulse rate, hemoglobin, serum creatinine)
measurements. At a 95% confidence level, a significant change between two measurements
has been found to be greater than 14% for ABI, greater than 120 for treadmill walking
time, greater than 0.4 for pulsatility index, and greater than 20% for duplex-acquired ve-
locity spectra and volume flow values (2).

Use of pressure cuffs of improper width relative to limb girth or noncalibrated manom-
eters can result in erroneous measurement of segmental limb systolic blood pressure. The
interpretation of the high-thigh pressure measurement to evaluate aortoiliac and common
femoral artery inflow occlusive disease is highly dependent on the size of the limb to relative
to the width of the cuff used. Theoretically, cuff width should be 20% greater than the
diameter of the limb (3). When a narrow (10- to 12-cm) thigh cuff is used, normal high-thigh
pressure is at least 20-30 mmHg higher than brachial pressure because of the artifact
produced by the relatively narrow cuff (e.g., normal high-thigh systolic pressure index is "8

>1.2; while the normal value of ankle-brachial systolic pressure index is >0.95) (1). §

Careful attention to instrument calibration is particularly important when plethysmo- a

graphic techniques are used for pulse volume recordings (PVR air plethysmography) or to •§

measure lower limb venous outflow (air or impedance plethysmography). Standardization ^

of cuff inflation pressure (approximately 65 mmHg) is mandatory to obtain reliable, repro- ^

ducible air plethysmography waveforms for the detection of arterial occlusive disease. For- jj

tunately, modern computer-based PVR instruments include an internal calibration system 2

that virtually eliminates operator errors in cuff application and technique. Improper cuff or I

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

4 BANDYK

photocell application can also produce artifacts and contribute to erroneous measurements
of digital pressure and venous recovery time.

When ultrasound is used to image blood vessels or interrogate blood flow patterns
within them, a variety of pitfalls can result in erroneous data or a study that cannot be
interpreted. These problems can be minimized if the operator has a thorough understanding
of ultrasound physics and the design features of the ultrasound system. A number of factors
relating to the scan-head design, Doppler system, frequency analyzer, and display devices
can affect imaging resolution and velocity spectral data (Table 2). Inappropriate selection
of the transducer's ultrasound frequency is a common pitfall of duplex ultrasonography.
High (7- to 15-MHz) B-mode imaging frequencies allow superior lateral and depth
resolution but are strongly attenuated by tissue, thereby limiting imaging to only superficial
(1- to 5-cm) vessels. Selection of an optimum transducer frequency should be based on the
depth of the vessel examined and the composition of overlying tissue. Table 3 lists the
transducer ultrasound frequency to obtain the strongest Doppler signal from vessels imaged
through different tissues; these frequencies are based on equations that account for ultra-
sound scatter, tissue attenuation, and vessel depth (4). For example, duplex testing of the
carotid artery in an obese patient that lies under fat and muscle at a depth of 7 cm may
require a transducer frequency of 3 MHz to record a strong Doppler signal.

Variation in beam pattern and focusing can result in lateral image and refractive dis-
tortion. Because ultrasound is a wave, the shape of the beam varies at distance from the
transducer. By increasing the transducer bandwidth, unwanted variations in the beam pat-
tern are smoothed out. Adjusting the focus (focal point) to just below the area of interest is
important to minimize lateral image distortion. In general, steering the ultrasound beam
perpendicular to vessel walls provides superior depth resolution. Refractive distortion re-
sults when the ultrasound travels through and crosses a boundary from one tissue to another
with a different ultrasound propagation speed. This can result in errors in dimensions and
the number of objects in the lateral direction of any ultrasound image. Duplication of the

Table 2 Factors Affecting Ultrasonic Imaging and Doppler
Data of Duplex Ultrasonography

Scan-head design

Transmitting ultrasound frequency
Ultrasound beam pattern and focusing
Beam steering instrumentation

Doppler ultrasound system

Time of image acquisition — frame rate

Sample volume size

Pulse repetition frequency 1

Frequency analyzer 8

Method of velocity spectrum analysis — fast Fourier a

transform c

Characteristic frequency and spectral width <

Display devices &

Real-time imaging and velocity spectral display jj

Compensation for aliasing g

Doppler sample volume/angle superimposed on B-mode image "§

Color power angiography g

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PITFALLS OF NONINVASIVE VASCULAR TESTING

Table 3 Ultrasound Frequency for the Strongest Doppler Signal
Relative to Vessel Depth and Type of Overlying Tissue

Ultrasound

Frequency (MHz)

Blood

Fat

Muscle

Bone

Depth (cm)

(0.1 8) a

(0.63)

(1.2)

(20)

0.5

97

28

5.0

0.9

2.0

24

7

7.0

0.2

5.0

10

3

0.5

0.1

8.0

6

2

0.3

-

15.0

3

1

0.5

-

a All values for attenuation expressed as dB/cm/MHz.
Source: Modified from Ref. (26).

subclavian artery is an example of a refractive distortion caused by reflection of ultrasound
from the pleura. Duplications appear in both B-mode and color images, and spectral
waveforms can be obtained from both images. The only defense against misdiagnosis is a
knowledge of anatomy and anatomical anomalies and the concept of refractive distortion.

Ultrasound beam steering can also introduce error measurement of blood-flow velocity.
An experimental study using a velocity-calibrated string phantom demonstrated significant
overestimation of recorded velocity when a multielement scan head was used in steered
versus unsteered modes (4). Differences in the range of 20-50% were recorded when end
elements of the scan head were used to record the Doppler signal. Errors of this magnitude
are worrisome, since measurements of peak systolic and end-diastolic velocity are used
clinically to recommend intervention for internal carotid artery or vein bypass graft stenosis.
The reasons for velocity overestimation are complex and related to characteristics of the
ultrasound system, including transducer beam width, aperture size, transmitting frequency,
and angle of Doppler beam insonation. Manufacturers should be encouraged to include
velocity calibration in routine instrument maintenance using test phantoms. In clinical
situations where peak velocity measurements approach thresholds levels, a linear-array scan
head should be used in the steered configuration with the Doppler cursor positioned at the
end of the transducer array, so that pulse Doppler signal recording will be at the lowest angle
of incidence to the axis of flow. Another strategy is to repeat the study using a phased-array
transducer and compare the values to those obtained with the linear-array transducer.

Duplex ultrasound systems continue to undergo rapid evolution in terms of image
resolution, color Doppler imaging, cost, and size. Some instruments provide velocity mea-
surements from peripheral arteries without any assumption about the Doppler examination ■§
angle (5). When color-flow imaging was introduced, it was hoped that diagnostic accuracy |
would improve. A comparison between color Doppler velocity and spectral waveform ve- a
locity demonstrates that values can differ due to angle adjustments or differences in signal c
processing. The definition of velocity is obscure, since duplex instrumentation records <
blood cell movement from a volume or voxel and the velocity components displayed in the >9
spectral display represent amplitude of multiple velocity vectors. A reproducible value can J
be obtained only when a consistent examination angle is used. This is important clinically «
for the diagnosis of disease progression. Serial duplex examinations should be performed |
with the same instrument and Doppler signals recorded at the same Doppler angle from an @

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

6 BANDYK

identical image as that acquired in the prior study. Because of the confusion produced by
real-time color imaging and color map aliasing, some instruments provide a feature called
"color power angiography," which shows pixels of blood motion in color without showing
direction. The resultant image permits the selection of pulsed Doppler recording sites at and
downstream from the location of maximum luminal stenosis. This minimizes the likelihood
that regions of slow blood flow or vessel segments insonated at high (80 -to 90-degree)
Doppler beam angles will be coded as showing no flow. When regions of no flow are
identified, it is essential to scan from at least two lines of sight to improve the angle at which
scan lines intersect with blood flow and thereby increase the likelihood of color-coding blood
flow if present. As a rule, if the tissue (i.e., blood) velocity is less than 1 cm/s, the velocity is
considered to be zero and color is not shown. If a wall filter is activated, velocities below
about 10 cm/s are not shown.

In "real-time" color Doppler ultrasound, the image is not formed instantly but pro-
cessed from the Doppler data from left to right over 30-50 ms. This produces a time dis-
tortion in all color-flow images. The speed of acquisition is approximately 150 cm/s which
is comparable to speeds in the vascular system: wall motion <1 cm/s, average arterial blood
velocity = 30 cm/s, pulse propagation speed = 1000 cm/s. Careful inspection of a single
color-flow image demonstrates that the speed of Doppler acquisition is slower than the
pulse propagation speed, resulting in a "systolic velocity bolus" approximately 2 cm long.
Thus, each color image frame depicts velocity data at a specific time during the pulse cycle
that is often different from that of the velocity spectra displayed below. The correct repre-
sentation of blood-flow patterns at an arterial stenosis cannot be depicted in a single color-
flow image.

B. Examination Technique

Obtaining high-quality noninvasive vascular data suitable for interpretation requires skill,
experience, and knowledge of vascular anatomy and hemodynamics. Lack of expertise
on the part of the examiner and adherence to the testing protocol are two important causes
of measurement error and variability. Peripheral arterial and venous Doppler studies
require the examiner to prepare the patient properly, manipulate the probe with care, and
develop a standard testing protocol with examination of both extremities for comparison.
Although indirect hemodynamic tests, such as segmental pressure measurement with
Doppler waveform analysis and pulse volume recording, are less subject to examiner error,
a number of pitfalls must be avoided (Table 4). When hemodynamic data are recorded
from the arterial or venous circulation, the patient should be rested before examination
and positioned appropriately for the testing method and indication for testing. To avoid
false-positive findings with indirect hemodynamic testing, it is recommended that when
abnormal values are obtained, the examination be repeated several times. ■§

Duplex ultrasonography requires considerable learned expertise in ultrasound phys- |

ics, vascular anatomy, and velocity spectral recording. To achieve a high level of correla- a

tion with contrast arteriography, the gold standard, a variety of pitfalls must be avoided c

(Table 5). The two most prominent problems are (a) assignment of the Doppler examina- <

tion angle and (b) aliasing. An understanding of these two concepts is essential, particularly >9

during the examination of tortuous or kinked arterial segments. In color Doppler instru- J

ments, color-coding of flow is based on the direction and velocity of blood flow. Since vas- «

cular anatomy is complex, with curves, branching, and dilatations, attention to the angle of |

insonation with the vessel is critical for assessing patency, directionality, and flow-pattern @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PITFALLS OF NONINVASIVE VASCULAR TESTING

Table 4 Testing Errors and Limitations of Segmental
Pressure Measurements

Causes of erroneously high systolic pressure measurement

Medial sclerosis — calcified arteries

Hypertension

Edema

Narrow cuff width relative to limb girth
Causes of erroneously low systolic pressure meassurement

Measurement after exercise or smoking

Low ambient temperature

Congestive heart failure

Inadequate resting period

Rapid ( >5 mmHg/s) cuff deflation

Unrecognized subclavian artery stenosis or occlusion

characteristics. The vascular group at the University of Washington has recommended that
an examination angle of 60 degrees be used for all studies and remain constant for serial
studies of the same artery. This is possible with current duplex ultrasound systems, which
allow electronic steering of the color image, pulsed Doppler beam, or both. The percentage
change in velocity relative to a velocity measured at 60 degrees increases dramatically at
insonation angles greater than 70 degrees (Fig. 1). In the range of 30-60 degrees, an angle
assignment error of ± 5 degrees will results in a 10-20% error in calculated velocity. By
contrast, at an angle of 70 degrees or greater, the error in calculated velocity will be ap-
proximately 25%; but it will exceed 100% at an angle of 80 degrees. Thus, velocity spectra
recorded at high ( >65 degrees) insonation angles cannot be used to calculate peak systolic or
end-diastolic velocity waveform values and thereby to classify stenosis severity or measure
volume flow. As the angle of insonation decreases toward zero, the Doppler frequency shift
will increase. When the detected frequency shift exceeds one-half the pulse repetition
frequency (PRF), known as the Nyquist limit of the instrument, aliasing occurs. Aliasing
in color Doppler results in improper color assignment of both direction and amplitude
data, with the resulting flow pattern erroneously interpreted as abnormal or turbulent
flow. Aliasing is easily recognized during pulsed Doppler spectral analysis; the artifact can
be decreased or eliminated by increasing the angle of insonation or PRF of the instrument
or by changing the scan plane of the transducer to decrease the depth of imaging. Accurate

Table 5 Potential Pitfalls of Color Duplex Ultrasonography

Vessel misidentification g

Improper Doppler angle assignment 8

Aliasing a

Incorrect color gain settings c

Calcified vessel wall producing acoustic shadowing <

Vessel tortuousity M

Tandem occlusive lesions jj

Incorrect sample volume placement g

Improper scanning technique "§

Overinterpretation of color-flow image g

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

BANDYK

jjfr 120

o

o

■q} 100 +

>

■g sol

1

o

60 ■■

O 40

a> 20

c

o

a>
H

<D
CL

•20 J
■40

30

10 50 60

Angle of Insonation

70

80

Figure 1 Percentage change in calculated velocity relative to duplex-derived velocity calculated at
a 60-degree angle of insonation.

color duplex imaging compels the examiner to be vigilant at keeping a 60-degree or less
angle between vessels walls (the axis of blood flow) and the transducer scan lines. Reports
that have analyzed the variability of duplex-acquired velocity measurements have found
variance to be within clinically acceptable levels and to be caused primarily by problems
with the examination technique (failure to record the Doppler signal downstream from a
stenosis, vessel tortousity precluding reproducible angle assignment) rather than inaccu-
rate measurement of waveform parameters or changes in patient hemodynamics (6,7).

III. INTERPRETATIVE AND STATISTICAL PITFALLS

Noninvasive vascular diagnostic testing is subject to interpretative errors independent of
problems related to the instrumentation or examination technique. Indirect physiological
testing and duplex ultrasonography can yield false-positive results because of anatomical
variation, multilevel or bilateral disease, or concomitant medical conditions (congestive
heart failure, hypertension, hypotension, vasospasm) that alter arterial flow. Elevation of
segmental pressure measurements caused by heavily calcified vessels that cannot be com-
pressed with cuff inflation is a common interpretative pitfall. Artery wall calcification, in
the media or an atherosclerotic plaque, also severely attenuates ultrasound transmission
and leads to blind areas (acoustic shadowing) in the B-mode image, rendering studies
uninterpretable or impairing the ability to accurately identify the site of maximum ste-
nosis. Acoustic shadowing is common in complex atherosclerotic plaques and is the pri-
mary reason why direct measurement of arterial luminal reduction is inaccurate for >50%
diameter-reduction stenosis (8,9). The incorporation of sensitive Doppler flow-sensing
capability, including color power angiography, in modern ultrasound systems decreases
the likelihood of interpretating a high-grade stenosis with plaque calcification as a total
occlusion.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PITFALLS OF NONINVASIVE VASCULAR TESTING 9

An abnormal hemodynamic state that increases basal flow is another interpretative
error associated with overestimation of stenosis severity. In patients with severe carotid
stenosis or internal carotid occlusion, compensatory collateral flow can be present in the
contralateral internal carotid artery (ICA) if the circle of Willis is intact and blood flow is
provided to both cerebral hemispheres. This condition will result in overestimation of ICA
stenosis, since diagnostic criteria depend on the measured levels of peak systolic and end-
diastolic velocity (10,11).

This effect is most evident in the misclassification of the 40-60% stenosis category to a
more severe and thus clinically important disease category. Similarly, the severity of periph-
eral arterial or infrainguinal graft stenosis may be overclassified if an arteriovenous fistula
or postrevascularization hyperemic state is present downstream.

The interpretation of noninvasive vascular testing is complicated by the absence of
uniform diagnostic criteria. Various criteria have been published for the grading of carotid
and peripheral arterial stenosis. In general, it is recommended that an individual vascular
laboratory initially utilize "published" criteria appropriate for their instrumentation and
examination protocol, and then validate these criteria by comparison with angiography.
Medical Directors of noninvasive vascular laboratories are responsible for ensuring "qual-
ity testing," which includes implementation of an independent, blinded correlation of test
interpretation with a gold standard. Such a review process should expose not only pitfalls
of testing related to inadequate diagnostic criteria but also problems related to the precise
of the instrumentation and technologist performance. Suggested standards for the report-
ing noninvasive vascular testing have been developed by the national vascular surgery
societies.

Angiography and venography have been considered the reference standard for com-
parison of noninvasive vascular testing. But these gold standards are also associated with
significant inter-and intraobserver interpretation variability. Investigators have shown
that the agreement between duplex ultrasound and angiography for both carotid and
peripheral arterial disease classification has a variability similar to the agreement of two
radiologists reading the same angiograms (sensitivity 87%, specificity 94%) (12,13). The
reliability of the gold standard remains an unresolved problem in defining the accuracy
of noninvasive testing, especially duplex ultrasonography. For selected clinical applica-
tions, such as the diagnosis of acute deep venous thrombosis (DVT), the accuracy of duplex
ultrasound is considered to be superior to that of contrast imaging studies. The inter-and
intraobserver variability of duplex studies can be minimized by adherence to a rigid testing
protocol and the use of diagnostic criteria based on nonsubjective data acquisition (6). As in
the case of other imaging studies, duplex classification of disease severity into minimal or
moderate categories accounts for the majority of variability in interpretation.

IV. PERIPHERAL ARTERIAL TESTING |

Testing for peripheral artery disease is used to detect, identify the extent, and grade the as

severity of obstructive and aneurysmal disease. Measurement of systolic pressure at mul- c

tiple levels in combination with pulse (Doppler pulse volume recording) waveform analysis <

is an essential component of upper and lower limb arterial testing. Pressure measurements >9

establish the presence and severity of arterial occlusive disease and, in conjunction with J

exercise testing, the diagnosis of intermittent claudication can thus be accurately deter- «

mined. Arterial wall calcification, obesity, and limb edema are important factors limiting |

the diagnostic accuracy of peripheral arterial testing by contributing to cuff artifact or @

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

10 BANDYK

impairing ultrasound imaging of the arteries. Multilevel disease — which is commonplace
in patients with symptomatic lower limb ischemia — also lessens the diagnostic accuracy
of segmental pressure measurements for disease localization (aortoiliac, femoropopliteal,
popliteal-tibial), and duplex mapping of disease in specific arteries (iliac, common femoral,
superficial femoral, popliteal, tibial). The presence of adjacent stenosis or occlusion re-
duces the diagnostic accuracy of duplex scanning in the aortoiliac segment from >90 to
63% and in the superficial femoral artery from 93 to 83% (14).

An important clinical limitation of indirect physiological testing is that it provides no
information as to the morphology (stenosis versus occlusion) of the occlusive process. This
characteristic limits its usefulness for preintervention planning (bypass versus angioplasty)
and postoperative surveillance following intervention to detect developing stenosis. The
diagnostic accuracy of segmental pressure measurements to detect any arterial segment
as abnormal is high (87%), with a positive predictive value of 96%, but these values are
diminished in the presence of diabetes (15).

Color duplex ultrasonography can be used to image the entire lower limb arterial tree
from the pararenal aorta to the pedal arteries; discriminating between occluded, stenotic,
and nonstenotic arterial segments (16). Mapping of the arterial tree permits patient coun-
seling regarding both diagnosis and the nature of intervention required to correct vascular
problem [percutaneous transluminal angioplasty (PTA), direct surgery], or whether disease
severity and extent dictates the need for a femoropopliteal or an infrageniculate bypass. In
general, when duplex scanning identifies single-level long-segment occlusion or focal high-
grade stenosis not amenable to PTA, arterial surgery can be performed without angiog-
raphy. The accuracy in predicting intervention by endovascular or "open" surgical bypass
exceeds 90% and, more importantly, PTA or bypass patency is not reduced compared to
patients evaluated by contrast angiography (17-20). In cases of severe multilevel disease,
particularly with involvement of the infrageniculate arteries, duplex scanning is plagued by
several pitfalls that impair obtaining an accurate, detailed anatomical and hemodynamic
definition of the arterial trees, including artifacts produced by calcific disease, poor imaging
of the tibioperoneal trunk, and examiners with insufficient experience to perform arterial
mapping studies.

V. CEREBROVASCULAR TESTING

Evaluation of cerebrovascular — i.e., extracranial carotid/vertebral and subclavian ar-
teries — disease differs from peripheral arterial testing in that detection of lesions with
embolic potential, in addition to pressure-and flow-reducing lesions, must be identified.
Duplex ultrasonography is required to image, detect, and grade obstructive or aneurysmal
lesions. The diagnostic accuracy of cerebrovascular testing is high (approximately 90%),

with positive predictive values in excess of 90-95% for the detection of >50% and >70% ■§

diameter reduction ICA stenosis. Negative predictive values exceed 95% for all categories of |

stenosis and the diagnosis of ICA occlusion. The clinical confidence and experience with a

carotid duplex scanning is such that carotid endarterectomy based on ultrasound studies c

alone is common (60-90% of procedures) in both community and academic vascular sur- <

gery practices (20,21). For symptomatic patients, lesions with velocity spectra of a >50% &

stenosis are clinically important. In asymptomatic patients, only a high-grade ICA stenosis J

increases the risk of stroke. Appropriate values of peak systolic and end-diastolic velocity «

should be used. It is recommended that criteria that correlate with a positive predictive |

value of 95% or greater be used to recommend carotid endarterectomy in asymptomatic @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PITFALLS OF NONINVASIVE VASCULAR TESTING

11

patients (22). Acoustic shadowing, slow flow in the distal ICA, vessel tortuosity, and com-
pensatory collateral flow are the major diagnostic pitfalls. In general, carotid duplex testing
tends to overclassify the severity of stenosis when correlated with blinded angiographic
grading of ICA stenosis.

VI. PERIPHERAL VENOUS TESTING

Duplex ultrasonography is the recommended test for the diagnosis of acute lower and upper
limb DVT. Diagnostic accuracy exceeds 95% with specificity approaching 100% when
an examination of technical adequacy is achieved (23,24). The addition of color Doppler
has enabled technologists to perform a more complete and rapid assessment of above-and
below-knee veins. Although imaging of vein segments transversely and then applying pres-
sure to coapt anterior and posterior walls is the best evidence of normal, thrombus-free vein,
pulsed Doppler waveform analysis is also important to confirm flow phasicity with respi-
ration and maneuvers to augment flow. Imaging permits diagnosis of thrombus in dupli-
cated venous segments as well as nonoccluding thrombus. Duplex ultrasound has replaced
both contrast venography and physiological testing for screening symptomatic and asymp-
tomatic patients for DVT. Air and photoplethysmographic techniques are used primarily
for the hemodynamic evaluation of chronic venous insufficiency to grade the severity of
venous obstruction (impaired venous outflow) or venous reflux (short venous refilling time,
ambulatory venous hypertension).

As in other areas of vascular testing, venous duplex testing is dependent on examiner
experience and the technical quality of the examination. Erroneous or equivocal studies
can be minimized by meticulous examination of the infrapopliteal vein, recognition of the
segmental incompressibility of the superficial femoral vein within the adductor canal as a
normal finding, and better estimation of the age of the thrombotic process (25). An intra-
luminal filling defect, a dilated vein, impaired flow, and homogenous echogencity of the
thrombus are important features of acute DVT. Chronic DVT demonstrates differences
in compressibility and thrombus imaging (Table 6). Obesity, limb edema, anatomically
deep vascular structures, small vein caliber, thrombus outside the scan field, and bowel gas
impairing imaging of the vena cava and iliac veins are other factors that limit the diagnostic
accuracy of venous duplex studies. When an equivocal study is encountered, repeat scan-
ning in 1-2 days is recommended. This approach has been demonstrated to be safe. The
development of fatal pulmonary embolus is rare, but proximal DVT developed in 2-5% of
patients based on serial examinations. In some patients, especially if central (vena cava,

Table 6 Differences in Duplex Ultrasound Criteria Between Acute and Chronic Deep Venous
Thrombosis (DVT)

Characteristic

Acute DVT

Chronic DVT

Compressibility Compress to medium probe pressure

Echogenicity Homogenous

Surface features Smooth

Attachment Free-floating

Vein caliber Dilated

Venous flow Low, continuous

None

Heterogenous

Irregular

Firm

Normal or contracted

Phasic with collaterals imaged

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

12 BANDYK

iliac, subclavian, brachiocephalic) venous thrombosis is suspected, venography should be
performed.

VII. GOALS OF VASCULAR TESTING

The application of noninvasive testing methods, especially duplex ultrasound, permits
detection of the clinical spectrum of vascular disorders involving the arterial and venous
systems. Although the duplex instrumentation in the various clinical applications is
similar, testing protocols and diagnostic criteria should be tailored to provide measure-
ments that characterize the vascular condition in terms of location, extent, severity, and
morphology. An important pitfall of vascular testing is the failure to consider the clinical
indication for the study. Interpretation of finings should be relevant to the individual
patient. For example, an abnormal carotid duplex study in an asymptomatic patient
should focus primarily on whether or not a high-grade stenosis has been identified. Plaque
surface features and composition suggesting embolic potential are more relevant in an
examination of a symptomatic patient. If intervention is based solely on duplex findings,
threshold criteria with a positive predictive value in excess of 90% should be used.

Decisions regarding extent of testing for arterial and venous disorders are complex.
The type of information required can vary depending on the need to establish only a
diagnosis versus determining a treatment plan or assessing the result of intervention or
diagnosing disease progression. Regardless of the indication for testing, the precision and
reproducibility of measurements are important. Avoidance of the known pitfalls of non-
invasive vascular instrumentation, testing protocols, and interpretation results in im-
proved diagnostic accuracy and patient care.

REFERENCES

1. Sumner DS. Evaluation of noninvasive testing procedures. Data analysis and interpretation.
In: Berstein EF, ed. Noninvasive Diagnostic Fechniques in Vascular Disease. 3d ed. St Louis:
Mosby, 1985:861-889.

2. Johnston KW, Hosang MY, Andrews DF. Reproducibility of noninvasive vascular laboratory
measurement of the peripheral circulation. J Vase Surg 1987; 6:147-151.

3. Kirkendall WM, Burton AC, Epstein FH, et al. Recommendations for human blood pressure
determination by sphygmomanometers: report of a subcommittee of the postgraduate edu-
cation committee, American Heart Association. Circulation 1978; 36:980-988.

4. Daigle RJ, Stavros AF, Lee RM. Overestimation of velocity and frequency values by multi-
element linear array Doppler. J Vase Fechnol 1990; 14:206-213.

5. Beach KW, Dunmire B, Overbeck JR, et al. Vector Doppler systems for arterial studies. J Vase
Invest 1996; 2:155-165.

6. Kohler F, Langlois Y, Roederer GO, et al. Sources of variability in carotid duplex examination — "§
A prospective study. Ultrasound Med Biol 1985; 1 1:571-576. |

7. Rizzo RJ, Sandager G, Astleford P, et al. Mesenteric flow velocity variations as a function of J
angle of insonation. J Vase Surg 1990; 11:694-699. f>

8. Comerota AJ, Cranley JJ, Katz ML, et al. Real-time B-mode carotid imaging: a three year §
multicenter experience. J Vase Surg 1984; 1:84-95. |

9. Erickson SJ, Mewissen MW, Foley WD, et al. Stenosis of the internal carotid artery: assess- £
ment using color Doppler imaging compared with angiography. Am J Radiol 1989; 152:1299- 11
1305. |

10. Spadone DP, Barkmeier LD, Hodgson KJ, et al. Contralateral internal carotid artery stenosis J?

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PITFALLS OF NONINVASIVE VASCULAR TESTING 13

or occlusion: pitfall of correct ipsilateral classification — A study performed with color flow
imaging. J Vase Surg 1990; 11:642-649.

11. Fujitani RM, Mill JL, Wang LM, et al. The effect of unilateral internal carotid arterial
occlusion upon contralateral duplex study: Criteria for accurate interpretation. J Vase Surg
1992; 16:459-468.

12. Chikos PM, Fisher LD, Hirsh JA, et al. Observer variability in evaluating extracranial carotid
artery stenosis. Stroke 1983; 14:885-889.

13. Jager KA, Phillips DJ, Martin RRL, et al. Noninvasive mapping of lower limb arterial lesions.
Ultrasound Med Biol 1985; 11:515-520.

14. Allard I, Cloutier G, Durand LG, et al. Limitations of ultrasonic duplex scanning for diag-
nosing lower limb arterial stenosis in the presence of adjacent segment disease. J Vase Surg
1994; 19:650-657.

15. AbuRahma AF, Khan S, Robinson PA. Selective use of segmental Doppler pressures and color
duplex imaging in the localization of arterial occlusive disease of the lower extremities. Surgery
1995; 118:496-503.

16. Wain RA, Berdejo GL, Delvalle WN, et al. Can duplex scan arterial mapping replace contrast
arteriography as the test of choice before infrainguinal revascularization. J Vase Surg 1999;
29:100.

17. Ligush J Jr, Reavis SW, Preisser JS, Hansen KJ. Duplex ultrasound scanning defines operative
strategies for patients with limb-threatening ischemia. J Vase Surg 1998; 28:482^190.

18. Ascher E, Mazzariol F, Hingorani A, et al. The use of duplex ultrasound arterial mapping as an
alternative for primary and secondary infrapopliteal bypasses. Am J Surg 1999; 178:162.

19. Mazzariol F, Ascher E, Hingorani A, Gunduz Y, Yorkovich W, Salles-Cunha S. Lower-
extremity revascularisation without preoperative contrast arteriography in 185 cases: lessons
learned with duplex ultrasound arterial mapping. Eur J Vase Endovasc Surg 2000; 19:509-515.

20. Proia RR, Walsh D, Nelson PR, et al. Early results of infragenicular reconstruction based
solely on duplex arteriography. J Vase Surg 2001; 33:1165.

21. Dawson DL, Zierler RE, Strandness DE Jr, et al. The role of duplex scanning and arteriography
before carotid endarterectomy: a prospective study. J Vase Surg 1993; 18:673-683.

22. Samson RH, Bandyk DF, Showalter, Yunis J. Carotid endarterectomy based on duplex
ultrasonography: a safe approach associated with long-term stroke prevention. Vase Surg 2000;
34:125-136.

23. Lensing AWA, Prandoni P, Brandjes D. Detection of DVT by real-time B-mode ultraso-
nography. N Engl J Med 1989; 320:342-345.

24. Rose SC, Zwiebel WJ, Nelson BD, et al. Sympotmatic lower extremity deep venous throm-
bosis: accuracy, limitations, and role of color duplex flow imaging in diagnosis. Radiology
1990; 175:639-644.

25. Wright DJ, Shepard AD, McPharlin M, Ernst CB. Pitfalls in lower extremity venous duplex
scanning. J Vase Surg 1990; 11:675-679.

26. Beach KW. Physics and instrumentation for ultrasonic duplex scanning. In: Strandness DE Jr,
ed. Duplex Scanning in Vascular Disorders. NewYork: Raven Press, 1990:196-227.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

Cardiopulmonary Complications
Related to Vascular Surgery

W. Darrin Clouse*

Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts,

and F. Edward Hehert School of Medicine, Uniformed Services University of the Health

Sciences, Bethesda, Maryland, U.S.A.

David C. Brewster

Massachusetts General Hospital and Harvard Medical School,
Boston, Massachusetts, U.S.A.

I. INTRODUCTION

In addition to having systemic atherosclerosis, patients undergoing vascular surgery are
also commonly advanced in age, afflicted with chronic obstructive pulmonary disease
(COPD), showing effects of tobacco abuse, and suffering from other medical comorbid-
ities. Accordingly, cardiopulmonary complications are the most frequent perioperative
problem facing this patient population. As with any complication, evaluating and inter-
vening in an attempt to avoid potential problems is preferred to attending them after they
occur. Hence, a considerable amount of investigation has been performed in an effort to
better define, profile, and preemptively address those at high risk for cardiopulmonary
difficulties after vascular surgery. Obviously, this strategy is an attempt to minimize the
short- and long-term morbidity and mortality from cardiopulmonary disease experienced |

by this population. Yet it remains difficult to clearly define specifics regarding the risks and a

benefits of preoperative evaluation; the ability to approach this analysis sensibly is one of J

the most important aspects in the management of the vascular surgical patient. a

Current affiliation: Wilford Hall Medical Center, Lackland Air Force Base, Texas, U.S.A.

15

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

16 CLOUSE and BREWSTER

II. CARDIAC COMPLICATIONS

A. Coronary Artery Disease in Peripheral Vascular Patients

Some amount of coronary artery disease (CAD) is present in nearly all vascular surgery
patients due to the systemic nature of atherosclerosis. In a sentinel report from the
Cleveland Clinic, Hertzer and colleagues performed coronary angiography in 1000 con-
secutive patients being evaluated for peripheral vascular surgery (1). Coronary artery
disease, to varying degrees, was present in 92%. Moreover, 25% had severe, correctable
CAD and almost 20% suffered from three-vessel coronary atherosclerosis. The severity of
CAD appeared to be independent of vascular disease distribution, as CAD was severe in
36% of those with abdominal aortic aneurysm (AAA), 28% of those with lower extremity
ischemia, and 32% of patients with cerebrovascular disease. Further, contemporary series
reveal clinical evidence of CAD in 30-50% of patients undergoing various peripheral
vascular procedures (2-15).

Not surprisingly, cardiac morbidity [usually defined as myocardial infarction, unstable
angina, congestive heart failure (CHF), dysrhythmia, and cardiac death] is among the
leading causes of perioperative morbidity and mortality in those undergoing peripheral
vascular operation. Vascular patients are apparently at highest risk for postoperative car-
diac events. Lee and colleagues, in evaluating 4315 patients undergoing noncardiac surgery,
found that AAA repair and other peripheral vascular operations had a 3.6 and 3.9 relative
risk, respectively, for cardiac events compared to other types of procedures (16). Acute
myocardial infarction (AMI) was responsible for the majority of these events and is the
leading postoperative cardiac event in most series. Although the pathophysiology of peri-
operative AMI is complex, numerous factors are likely involved. Both surgery and
anesthesia initiate processes that place added stress on the myocardium. Catechol release,
increased myocardial sensitivity to catechols, fluid shifts, blood pressure fluctuations, alter-
ations in oxygen delivery, transient hypercoagulability, and tachycardia all may compro-
mise the coronary circulation's ability to supply oxygen postoperatively.

Recent studies retrospectively delineating the incidence of postoperative AMI in vas-
cular patients reveal roughly a 1-7% chance of infarction in those having operations for
aneurysm, cerebrovascular disease, lower extremity ischemia, or other major peripheral
vascular disease (Table 1). But when studied prospectively with aggressive diagnostic physi-
ological parameters, AMI may occur in as many as 15% of patients (17). Several factors
could explain this disparity. Postoperative myocardial ischemia is commonly silent, can
occur any time during the postprocedural course, and is not uncommon in those without
preoperative evidence of CAD. Illustrating this last point, in Hertzer's angiographic eval-
uation of 1000 vascular patients, 15% of those without clinical evidence of CAD had severe
disease and 22% had advanced disease (1). Therefore, coronary disease is prevalent in this
population even without clinical indicators. Silent ischemia is a harbinger of myocardial ^

infarction and cardiac morbidity, and unexplained persistent tachycardia may be the only |

suggestion of ischemia postoperatively. Using perioperative ambulatory electrocardio- |

graphic (ECG) monitoring, Pasternak and coworkers reported that silent myocardial 5

ischemia occurred in some 60% of 200 vascular surgery patients, and both the duration ^

and number of ischemic episodes were significant predictors of postoperative AMI, which g

occurred in 4.5% (18). Even more disturbing, Krupski et al. documented postoperative jf

ischemic ECG changes in 57% of infrainguinal reconstructions and 31% of aortic opera- g

tions, with a staggering 98% being silent (15). Others have reported similarly large pro- "|

portions of vascular patients experiencing silent postoperative ischemia (19-21). AMI has 2

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 17

Table 1 Postoperative Myocardial Infarction (MI) Rate in Contemporary Vascular Surgery
Reports

Number of

Report"

Year

patients/operations

MI (%)

Aortic surgery

Cambria et al. (TAA) (106)

2002

337

3.9

Berry et al. (AAA) (128)

2001

856

1.3

Romero et al. (EAAA) (107)

2001

173

4.6

Hovsepian et al. (EAAA) (108)

2001

144

2.8

Becker et al. (EAAA) (109)

2001

305

0.6

Ponovost et al. (AAR) (110)

2001

2,987

3.3

Axelrodet al. (AAA) (111)

2001

1,001

3.1

Martin et al. (SR/TAAIII/IV) (112)

2000

165

5.0

Pearce et al. (AAA) (113)

1999

13,415

1.8

Carotid surgery

Sternbach et al. (114)

2002

550

1.1

Azia et al. (6)

2001

123

1.6

AbuRhama et al. (115)

2001

389

0.3

James et al. (116)

2001

324

0.9

Hamdan et al. (117)

1999

1,001

0.5

Pearce et al. (113)

1999

45,744

0.8

Caoetal. (118)

1998

1,353

0.4

Infrainguinal surgery

Conte et al. (all indications) (119)

2001

1,642

3.0

Kalra et al. (pedal bypass) (120)

2001

280

6.4

Faries et al. (all indications) (121)

2000

740

1.5

Nicoloff et al. (limb salvage) (122)

1998

112

9.0

Matsuura et al. (elective) (123)

1997

205

3.4

Major vascular surgery

(procedures combined)

Boersma (44)

2001

1,351

3.3

Landesberg et al. (23)

2001

185

6.5

Sprung et al. (3)

2000

6,948

1.5

Van Damme et al. (9)

1997

142

2.1

Landsberg et al. (35)

1997

405

4.7

Mamode et al. (22)

1996

204

6.9

a AAA = abdominal aortic aneurysm; AAR = abdominal aortic reconstruction (i.e., both aneurysm
and occlusive disease); EAAA = endovascular abdominal aortic aneurysm repair; SR = suprarenal
aortic aneurysm; TAA = thoracoabdominal aneurysm.

3

been estimated to be silent in nearly one-third of patients experiencing infarction after |j)

undergoing vascular surgery (13,22). A recent prospective evaluation by Landesberg et al. §

studied 185 vascular patients for ischemia and infarction with continuous postoperative J

ECG monitoring for 48-72 h and daily cardiac enzyme panels (23). Those with recent MI g"

or unstable angina were excluded. Myocardial ischemia developed in 38 (20.5%) patients, g

with AMI occurring in 12 (6.5%). Ischemia was silent in over 80%, and in 42% of those "S

with AMI. S

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

18 CLOUSE and BREWSTER

Postoperative silent myocardial ischemia in noncardiac surgery has also recently been
evaluated by Swear and colleagues and Higham et al. (20,21). Results were similar, with
40% of vascular patients having silent ischemia postoperatively; those going through
vascular surgery were at highest risk for its development. It is not surprising that myo-
cardial ischemia is silent in these patients, as it may be confounded by incisional pain and
drug effects. The proper recognition of and intervention in postoperative cardiac events in
terms of timing is another problem. Krupski and others have noted that a significant
proportion of cardiac events occur 3 or more days after vascular surgery, when patients
are usually not monitored or being routinely tested for laboratory evidence of myocardial
ischemia (3,15,22,24,25). These several factors make it difficult to consistently identify or
prepare for cardiac morbidity after vascular surgery.

Cardiac death is the most common cause of perioperative mortality in vascular surgery.
As mentioned, AMI accounts for the bulk of these events. It has been projected that up to
40% of postvascular surgery Mis result in death (26). LTtalien reported on 547 consecutive
patients referred for coronary evaluation prior to vascular surgery and found that 33% of
postoperative AMI resulted in mortality (10). During the early 1980s, several reports from
Hertzer's group at the Cleveland Clinic reported that half of their postoperative deaths
resulted from AMI in the vascular population (27-29). Sprung et al. recently reviewed
almost 7000 consecutive cases from the Cleveland Clinic vascular registry and found post-
operative AMI occurred in an enviable 1.5%, with 21% dying as a result (3). Also, cardiac
death is the leading reason for late demise in those with peripheral vascular disease.
Regardless of required procedure, authors have repeatedly implicated cardiac death as
the leading source of late mortality after vascular surgery (10,14). Therefore it is clear that
cardiac intervention could not only potentially lessen morbidity surrounding these oper-
ations but also perhaps improve long-term survival.

B. Clinical Risk Assessment

Because vascular patients are at an increased perioperative cardiac risk, any preoperatively
identified indicators of enhanced vulnerability to events may help in selecting patients for
further cardiac evaluation. Several studies have identified physical findings or historical
factors associated with increased cardiac risk in those undergoing surgery and have
combined these markers into indices designed to effectively quantitate risk. While some
large series have evaluated clinical predictors of cardiac events in all noncardiac surgical
patients, others have focused on those undergoing vascular surgery (Table 2). Based on these
studies, several clinical criteria have been identified as independently important in the
development of postoperative cardiac complications. To varying degrees these have in-
cluded age >70 years, angina, prior myocardial infarction, significant arrhythmias, severe
valvular disease, CHF, ECG abnormalities, prior stroke, diabetes mellitus, renal insuffi-
ciency and high risk (i.e., abdominal, thoracic, vascular, and emergent) surgery. 1
Recognizing a need for refinement of the factors predicting cardiac morbidity, the |
American College of Cardiology and the American Heart Association (ACC/AHA) Task j§
Force on Practice Guidelines has categorized these variables into major, intermediate, and ||
minor clinical predictors of perioperative cardiac risk (Table 3) (30,31). The surgical pro- ^
cedures were stratified into those having high, intermediate, and low cardiac risk. Aortic and "jl
peripheral vascular surgical procedures are considered high-risk procedures, while carotid s
endarterectomy is deemed intermediate in risk. Further, the task force described a historical 2
estimation of functional capacity based upon metabolic equivalents (METs), where 1 MET J

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY

19

Table 2 Clinical Risk Predictors for Adverse Postoperative Cardiac Events"

Report

Year

Independent predictors

Noncardiac surgery
Goldman et al. (99) 1977

Detsky et al. (101)

1986

Lee et al. (16)

1999

Vascular surgery
Eagle et al. (52)

Sprung et al. (3)

Boersma et al. (44)

1987

2000

2001

Age >70, MI in previous 6 months

S 3 gallop or JVD, aortic stenosis

Rhythm other than sinus or PAC, >5 PVC/min,

Poor general medical status (Pa0 2 <60 or

PCO 2 >50, K<3.0 or HC0 3 <20, BUN>50 or Cr

>3.0, elevated SGOT)
Intraperitoneal or intrathoracic or aortic or

emergency operation
Age>70, any prior MI, angina pectoris
Pulmonary edema, critical aortic stenosis
Rhythm other than sinus, >5 PVC/min
Poor general medical status (Pa0 2 <60 or

PCO 2 >50, K<3.0 or HC0 3 <20, BUN>50 or Cr

>3.0, elevated SGOT)
Emergency surgery
Ischemic heart disease
History of CHF
History of CVD

Insulin-dependent diabetes mellitus
Preoperative serum creatinine >2.0
High-risk surgery

Age >70

Q waves on ECG

History of angina pectoris

History of ventricular ectopy requiring therapy

Diabetes mellitus requiring therapy

Valvular heart disease (aortic or mitral)

Coronary artery disease

History of CHF

Operative transfusion

Prior CABG (protective)

History of CVD

History of CHF

Prior MI

Current or prior angina pectoris

Age >70

Beta-blocker therapy (protective)

11 BUN = blood urea nitrogen; CABG = coronary artery bypass grafting; CHF = congestive heart
failure; CVD = cerebrovascular disease; HC0 3 = serum bicarbonate; JVD = jugular venous dis-
tention; K = serum potassium; MI = myocardial infarction; PAC = premature atrial contraction;
PVC = premature ventricular contraction; Pa0 2 = arterial oxygen tension; PaC0 2 = arterial carbon
dioxide tension; SGOT = serum glutamic oxaloacetate transaminase.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

20 CLOUSE and BREWSTER

Table 3 Clinical Risk Factors for Adverse Postoperative Cardiac Event

Major

Unstable coronary syndromes

Acute or recent myocardial infarction with clinical symptoms or
noninvasive study evidence of ischemic risk

Unstable or severe angina pectoris
Decompensated heart failure
Significant arrhythmias

High-grade atrioventricular block

Symptomatic ventricular arrhythmias with underlying heart disease

Supraventricular arrhythmias with uncontrolled ventricular rate
Intermediate

Mild angina pectoris

Previous myocardial infarction by history or Q waves
Compensated or previous heart failure
Diabetes mellitus
Renal insufficiency
Minor

Advanced age
Abnormal ECG

Left ventricular hypertrophy

Left bundle branch block

ST-T abnormalities
Rhythm other than sinus
Low functional capacity
Stroke history
Uncontrolled hypertension

Sources: Refs. 30 and 31.

represents the oxygen consumption of a 40-year-old man in a resting state. Using these
scales, an algorithmic approach to preoperative cardiac risk assessment was then presented
(30,31). This algorithm delineated four specific clinical scenarios:

1. Patients who need emergency surgery, have undergone coronary revasculariza-
tion within 5 years without recurrent signs and symptoms of CAD, or have had
a recent favorable coronary evaluation without clinical changes should go directly
for operation.

2. Those with major clinical predictors should have elective surgery canceled and ^
coronary artery evaluation and intervention as necessary. E

3. Those with intermediate clinical predictors and poor functional capacity (<4 J
METs) should have noninvasive coronary evaluation, while those with moderate -g>
or excellent functional capacity ( >4 METs) should undergo noninvasive study §
when the proposed procedure is high risk in nature. When the procedure is inter- M
mediate or low in risk, the patient should proceed to surgery. M

4. In those with minor or no clinical CAD predictors and poor functional capacity «
( <4 METs) who are to undergo a high-risk operation, noninvasive coronary eval- §
uation is recommended. Those with moderate or excellent functional capacity ( >4 ©

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 21

METs) or those with poor capacity who are to undergo intermediate- or low-risk
procedures should proceed to operation.

If evidence of coronary ischemia is found on noninvasive testing, cardiology referral
should take place, with further coronary artery evaluation and revascularization when in-
dicated or, in those found to have less extensive disease, risk-factor modification prior to
operation.

Within this schema, vascular surgery patients are frequently within the category of
those requiring noninvasive evaluation. A recent retrospective evaluation of the ACC/
AHA guidelines in 133 consecutive patients undergoing aortic surgery was performed by
Samain and coworkers (5). All had routine cardiac evaluation prior to surgery. Following
the algorithm, 60 patients (45%) required noninvasive testing and the algorithm classified
all those with postoperative events as being at high risk. The authors concluded that the
ACC/AHA guidelines were valid in patients undergoing major vascular surgery. These
recommendations appear to have helped to streamline thinking about cardiac clinical risk
assessment prior to vascular surgery. However, while the use of clinical risk factors alone
does not provide complete postoperative cardiac prognosis, it allows surgeons to identify
those at increased risk who need further evaluation or procedure modification and those
who may proceed to surgery knowing that their risk, although not absent, is minimal.

C. Adjunctive Screening Methods for Cardiac Risk Stratification

Once the need for further preoperative cardiac evaluation beyond clinical assessment is
deemed necessary, another decision must be made on how best to accomplish this task.
While coronary angiography is appropriate in those patients with unstable angina, recent
myocardial infarction, decompensating CHF, and new arrhythmias, it is invasive, with
complications of its own, and provides only anatomical information. Thus, no functional
information is obtained and its predictive value is limited. Its principal use in preoperative
risk stratification is after noninvasive studies have determined the patient to be at high
risk for cardiac events and shown that catheterization may provide therapeutic benefit.
Thus, the use of routine coronary angiography to stratify vascular patients is historical. A
variety of noninvasive techniques have been developed that provide functional and/or
metabolic definition of the myocardium.

Exercise-stress ECG testing was the earliest method used to investigate CAD severity. It
is valuable in that this test provides data on functional workload capacity in addition
to identifying major CAD presence. Significant induced ischemia at low-level exercise
is clearly predictive of postoperative cardiac events, with negative predictive values of
over 90% (30,32). However, because of age, deconditioning, severe claudication, ischemic
ulcerations, prior amputation, stroke, severe lung disease, arthritis, or other infirmities,
peripheral vascular surgery patients are frequently incapable of adequate exercise to achieve ■§

the necessary 75-85% of the maximal predicted heart rate required for proper testing. |

Submaximal testing may decrease sensitivity in this population with a high prevalence of a

CAD and silent ischemia; because this is also the surgical population at highest risk of c

cardiac events, ECG stress testing plays a limited role in noninvasive evaluation. Ambula- <

tory ECG (Holter) monitoring with evidence of myocardial ischemia and left ventricular >9

hypertrophy has also been shown by some to correlate with postoperative cardiac events J

(18,19,33-35). One prospective study, however, has shown that evidence of myocardial «

ischemia on preoperative Holter monitoring did not predict postoperative AMI (36). Again, |

however, this method is limited for several reasons. First, there is no protocol to standard- @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

22 CLOUSE and BREWSTER

ize results using this method. Second, baseline ECG characteristics obviate the ability to
critically interpret ST segments in a significant proportion of vascular patients. And last,
while positive results may identify patients at increased perioperative cardiac risk, prog-
nostic stratification parameters are lacking, and how to manage those with positive findings
remains unclear.

Radionuclide ventriculography [gated blood pool scan, multiple gated acquisition
(MUGA) scan] is a noninvasive technique that provides for assessment of left ventricular
function and wall motion. Most studies have shown an increase in postoperative cardiac
events with a left ventricular ejection fraction (LVEF) less than normal (55%), and the risk
appears to be higher with LVEF less than 35% (17,37-39). Franco and colleagues, how-
ever, suggested that resting radionuclide ventriculography did not contribute insight into
postoperative cardiac events (40). Exercise testing may be added to enhance functional
changes in LVEF and wall motion indicative of CAD that is limiting myocardial
perfusion, and positive findings with this method correlate well with postoperative cardiac
risk (41). However, the development of echocardiography, which assesses both LVEF and
wall motion without the need for radioactive material — as well as improvements in
myocardial scintigraphy including newer tracer agents that allow both evaluation of
ventricular function as well as myocardial perfusion — has diminished enthusiasm to utilize
this technique for cardiac risk stratification (42).

Two techniques to noninvasively evaluate CAD in those with inadequate functional
capacity to undergo proper exercise testing have been devised. The newest method of non-
invasive cardiac testing is dobutamine stress echocardiography (DSE), which allows stress
evaluation of the myocardium without exercise. Dobutamine is a betai receptor-selective
agonist that increases heart rate and contractility and thus myocardial oxygen consumption.
When dobutamine is given in conjunction with echocardiography before and after admin-
istration, differences in left ventricular function and wall motion are detected. These findings
suggest unmasking of flow-limiting CAD with resultant myocardial dysfunction. Positive
ischemic endpoints are noted as new or worsening wall motion abnormalities (NWMA) and
the myocardium defined in segments. Although DSE has not been as extensively studied as
myocardial scintigraphy, its utility as an adjunct to cardiac risk assessment is being defined.
Indeed, the existence of wall motion abnormalities at rest correlates with postoperative
events in vascular patients (43,44). The positive predictive value of stress-induced ischemia
using DSE appears to be 10-40% and the negative predictive value without NWMA 95-
100% (9,30,43-45). Two studies have used DSE to report stratification schemes for those
requiring noninvasive testing. Poldermans and colleagues studied 300 vascular surgery
patients by clinical risk assessment using Detsky and Eagle criteria and DSE (43). All
patients who suffered a postoperative cardiac event had a positive DSE. Thus, the negative
predictive value was 100% and the positive predictive value 38%. Further, the authors
found that the lower the heart rate at the time of NWMA onset, the greater the risk of a ■§

postoperative event. Recently, Boersma et al. evaluated 1351 consecutive major vascular g

surgery patients (44). DSE was performed in 1079 patients and a positive DSE was hide- »

pendently associated with adverse cardiac events. Also, those with NWMA in five or more c

segments had a four- to sixfold increase in events compared to those with one to four <

segments affected. By incorporating the Lee clinical risk index, the authors found that in >9

those with fewer than three risk factors, the adverse cardiac event rate was 2%, and DSE %

added little to the patient's risk stratification. In those with more than three factors, DSE «

was predictive of adverse cardiac events, as the event rate was 4.6% in those with no |

NWMA, 13% in those with one to four affected, segments and 35% when five or more g

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 23

segments became ischemic. The protective effect of beta blockade in vascular patients was
impressive in this series as well and is discussed below. The major criticism of DSE as a
method of noninvasive evaluation is a lack of standardized protocols and criteria to describe
positive tests and the subjective nature of NWMA. Also, problems with dobutamine admin-
istration — such as arrhythmias, myocardial infarction, angina, hypotension and drug
reactions — may occur. However, at present, this appears to be a useful technique in the
risk stratification of vascular patients.

The second and most studied method of noninvasive cardiac evaluation in vascular
patients unable to exercise is dipyridamole-thallium scintigraphy. This method was first
applied to risk assessment prior to vascular surgery by Boucher in 1985, and we have used
it extensively at the Massachusetts General Hospital (46). The nuclear imaging technology
has continued to improve and newer radiotracers based on technitium 99m ( 99m Tc), with
shorter half-lives, allow both assessment of ventricular function as well as myocardial
viability imaging (42). Dipyridamole administered intravenously inhibits the natural
transport processes of adenosine, a potent coronary vasodilator, causing increased levels.
This results in coronary vasodilatation with increased flow in those coronary arteries
without significant disease, while arteries with significant atherosclerotic disease fail to
dilate. Flow is preferential to myocardial segments supplied by normal arteries, with
hypoperfusion in the distribution of severely disease coronary arteries. Adenosine itself is
now also used to precipitate pharmacologically induced coronary vasodilatation, and the
use of dobutamine-induced stress for scintigraphy has also been reported (9,42). Thallium
201 ( 201 Th) and the newer 99m Tc- labeled tracers enter myocardial cells in proportion to
blood flow; thus they enter viable areas. The myocardium normally has homogeneous
radiotracer uptake unless there are regions of hypoperfusion.

The current myocardial scintigraphy technique used for cardiac risk assessment prior
to vascular surgery at the Massachusetts General Hospital is described in Figure 1.
Intravenous adenosine is used for pharmacological stress and 99m Tc sestamibi (mibi) is the
currently used radiotracer. Imaging occurs with the use of stress single-photon-emission
computed tomography (SPECT). Initial resting scans are obtained 30-45 min after radio-
tracer injection (10 mCi m Tc mibi). Then the stress sequence begins with low-level exer-
cise and adenosine infusion (140 ug/kg/min). The radiotracer (30 mCi m Tc mibi) is again
injected 2 min after the adenosine is started. Once the stress sequence is complete, repeat
images are obtained 45 min after tracer injection, using ECG gating to provide LVEF
data. Normally perfused myocardium quickly absorbs the radiotracer and is visualized on
both resting and stress images. Viable ischemic myocardium, beyond a fixed coronary
artery stenosis that cannot dilate, takes up the radiotracer at a reduced rate and is seen
only in the resting baseline images (Fig. 2). Hence, any areas of hypoperfusion appear as
decreased radiotracer uptake or "cold spots" on the stress scans. If these initially under-
perfused areas show more homogeneous uptake of the radiotracer on the resting baseline ■a
images, this is referred to as "fill in" or redistribution (Fig. 3) (46). A scan is considered |
positive if regions of redistribution are confirmed. These areas represent regions of as
ischemic but viable myocardium at risk for perioperative ischemia or infarction. Regions c
of persistent decreased uptake represent nonviable myocardium from previous infarction <
(i.e., "scarred") and are considered not to be at risk for further events. Such a scan is not >9
considered positive (47). The sensitivity of scintigraphy may be enhanced by low-level J
exercise; we use low stress treadmill walking for this purpose. If patients experience any «
adverse reactions to dipyridamole or adenosine (i.e., flushing, wheezing, dizziness, angina, |
arrhythmias, and hypotension), the infusion is stopped. Aminophylline may need to be @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

24

CLOUSE and BREWSTER

Resting Radiotracer Injection: (lOmCi Tc-99m sestamibi)

I

30-45 Minutes

Resting/Baseline Scintigraphy

I

Treadmill at 0% grade

(0.5 to 1,7 miles per hour)

I

Ad e nosi ne I V ( 1 40ng/kg/m i n )

1

2 Minutes after Adenosine: Begin Radiotracer Injection (30mCi Tc-99m sestamibi)
Stop Exercise/ Adenosine after S Minutes

1

30-45 Minutes

Stress Scintigraphy (ECG Gating)

Figure 1 Our current myocardial scintigraphy protocol for cardiac risk stratification screening
prior to vascular surgery.

STRESS SCINTIGRAPHY IMAGES

RESTING/BASELINE SCINTIGRAPHY
IMAGES

I
•5

Figure 2 Myocardial scintigraphy images using the radiotracer technetium-99m sestamibi showing
apical (A) inferior (B) and lateral (C) "redistribution" consistent with viable ischemic myocardium
at risk.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 25

STRESS

Stress

Scintigraphy

IVDi|>yridamole Radiotracer

orAdenosine Injection

o **

/ w

Baseline or
Resting
Scintigraphy i

O % Q 3

Normal Fixed Defect Partial Complete

— Flll-ln

(i.e. Redistribution)
(Scar) (Viable Myocardium in Jeopardy)

Figure 3 Schematic representation of possible myocardial scintigraphy results. Normal stress and
resting myocardial scintigraphy showing homogenous uptake of radiotracer during both stress and
rest. Stress scintigraphy images obtained after pharmacological coronary vasodilation revealing
poor radiotracer uptake in a myocardial distribution may persist at rest as a fixed defect suggesting
scar, or resting images may partially or completely "fill in" the defect, suggesting viable myocardium
at risk.

used to compete for adenosine receptors. Patients taking methylxanthines must obviously
discontinue these prior to undergoing this test.

Stress scintigraphy has been widely evaluated in vascular patients in an attempt to iden-
tify those at highest risk of cardiac complications. L'ltalien found that in 547 consecutive
patients referred for dipyridamole-thallium scanning prior to aortic, infrainguinal, or ca-
rotid surgery, redistribution on scanning was the most significant predictor of postoperative
MI. Those with a positive test were over 3 1/2 times more likely to experience infarct (10).
The significance of redistribution on scintigraphy has also been suggested predictive of
postoperative cardiac events by others (10,1 1,48-51). Eagle et al. reported that the predictive
value of dipyridamole-thallium redistribution could be significantly increased by combining ■§

scan results with clinical criteria (52). This landmark study categorized patients into three g

risk groups based on the presence of these clinical factors. Low-risk patients were those with »

no clinical markers. One or two factors placed the patient in the intermediate-risk group, c

and three or more markers defined the high-risk group (Fig. 4). The risk of untoward post- <

operative cardiac ischemic events was small in the low-risk group (3% events, no deaths); >9

scintigraphy is not warranted in these patients. The event rate in the high-risk group was so -|

high (50%) that scan results would not improve the risk prediction or alter preoperative Q

workup or perioperative management. Consideration of a more extensive coronary artery |

evaluation (i.e., catheterization) and intervention as well as alternative surgical approaches @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

26 CLOUSE and BREWSTER

All Patients
200

I 1 1

64 116 20

No 1 or 2 Clinical Variables 3 or More

Clinical | | Clinical

Variable No Thallium Thallium Variables

l Distribution Distribution |

2 64 13.1%) 2/62(3 2%) 16'54 (29.6%) 10/20 (50°'ol Event

iCIOto.08) (CI to 08) (CI .16 to .44) (CI .29 to .71) Risk

' O wave on electrocardiogram, age > 70 yrs hisiory ol angina.

history ol ventricular eciopic activity requiring [realment and
diabeses me I lit us requiring treatment

Figure 4 Selective cardiac screening efficacy using clinical predictors and myocardial scintigraphy
to identify patients at increased risk for adverse cardiac events after vascular surgery. Event refers to
postoperative cardiac ischemic events, including unstable angina, ischemic pulmonary edema,
myocardial infarction, or cardiac death. CI = 95% confidence interval. (From Ref. 52.)

or cancellation would be more productive. However, in the moderate-risk patients, stress-
thallium scintigraphy was extremely helpful, as those with normal perfusion had a post-
operative event rate (3%) equating that of the low-risk cohort, whereas redistribution
indicated a 30% risk and suggested that a more detailed CAD assessment was prudent.
Using this strategy, the test's sensitivity and specificity for identifying postoperative ad-
verse cardiac events were 83% and 66%, respectively, with a positive predictive value of
30% and a negative predictive value of 96%. The authors concluded that the method's
ability to stratify patients was enhanced by combining clinical risk factors with selective
myocardial scintigraphy and that it could effectively limit the number of patients requiring
extensive coronary evaluation (52). Also, Levinson and colleagues have reported that quan-
tification of redistribution by ischemic segments, number of positive views, and number of
coronary territories affected further identifies those at highest risk for postoperative cardiac
events after vascular surgery (Table 4) (53). Others have also found quantitative interpre-

Table 4 Maximizing Myocardial Scintigraphy Interpretation

Variables Probability of cardiac event

Ischemic segments
<3 12%Q? = 0.03)

>4 38% |

Number of views with ischemia a

<1 0%(p = 0.005) §

>2 36% §

4

Number of coronary territories with ischemia jf

<3 13%(p = 0.007) |

>4 43% I

Source: Ref. 53. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 27

tation of redistribution to improve the positive predictive value of myocardial scintigraphy
(54,55).

However, there is no consensus regarding the use of myocardial scintigraphy for the
purpose of preoperative cardiac risk stratification and predicting cardiac morbidity and
mortality in vascular surgery patients. Several authors have suggested that routine stress
cardiac scintigraphy before vascular surgery is not helpful in predicting events (56-59).
Baron, et al. evaluated the usefulness of dipyridamole-thallium and LVEF in 513
consecutive patients who were to undergo elective abdominal aortic surgery (60). Those
with clear indications for coronary catheterization as well as those with incomplete clinical
or imaging data were excluded. In all, 457 patients were analyzed and the authors found
no correlation between preoperative redistribution and postoperative events. Most
recently, in a blinded, prospective study, de Virgilio and coworkers concluded that no
association existed between redistribution and postoperative cardiac events in 80 patients
with at least one Eagle risk factor (2). They reported a sensitivity and specificity of 44%
and 65%, respectively, with positive and negative predictive values of 14% and 90% in
their selective cohort.

At present, there is no clearly superior method of noninvasive evaluation for cardiac
stratification prior to vascular surgery. While exercise stress ECG is a powerful tool to
evaluate CAD, patients afflicted with peripheral vascular disease are, for the most part,
unable to undergo meaningful exercise stress testing for health related reasons. Also, a
significant proportion have baseline ECG abnormalities. Pharmacological stress testing
with dipyridamole, adenosine, or dobutamine is, therefore, useful to instigate myocardial
stress in these patients. Stress imaging by either DSE or myocardial scintigraphy appears
to help further define risks in those requiring noninvasive cardiac evaluation; these are the
most often used techniques in this population. Current reviews reveal a negative predictive
value of these noninvasive methods of roughly 98% (42). Further, there is evidence that
both myocardial scintigraphy and DSE provide long-term coronary artery prognostic
information in vascular patients (10,59,61-63). Even though their use in preoperative
stratification is still hotly debated, their strongest use may be in identifying those at
highest risk for long-term CAD events and in reducing late cardiac morbidity and mor-
tality.

D. Risk Stratification Postures for Cardiac Assessment Prior to
Vascular Surgery

Albeit clinical variables important to the development of postoperative cardiac compli-
cations are widely recognized and critical evaluation to find the elusive perfect noninvasive
technique is ongoing, what remains critical to vascular surgeons when their patients are
found to be at increased cardiac risk is what can be done to improve their patients' course ■§

and long-term health. A lowering of postoperative cardiac morbidity and mortality by g

aggressive preoperative evaluation for CAD and revascularization within this population »

remains, at present, unproven, and several general postures toward cardiac risk stratifi- c

cation have evolved. <

u

First, an aggressive stance may be taken that all patients who are to have a major >3

vascular surgical procedure should be critically evaluated and treated for the presence of ^

significant CAD in the hopes of both reducing perioperative morbidity and improving «

long-term survival. The cornerstone of this approach is the hypothesis that coronary |

revascularization improves postoperative outcomes after vascular surgery. The literature @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

28 CLOUSE and BREWSTER

over the last three decades has delineated a potential benefit, both perioperatively and in
the long term, in those who undergo coronary artery bypass grafting (CABG) prior to
noncardiac surgery. This was evident from the Coronary Artery Surgery Study (CASS),
when Eagle et al. reported on nearly 3400 patients undergoing various noncardiac
procedures. Previous CABG was independently protective in those undergoing high-risk
surgery and the protective effect was present for up to 6 years after coronary revascula-
rization (64). Hertzer and colleagues studied over 1000 vascular patients and found those
identified preoperatively with significant surgically reconstructable CAD who underwent
CABG had fewer postoperative cardiac complications after vascular surgery than those
who did not. Further, the long-term (5-year) survival in those who had coronary revas-
cularization was significantly enhanced (75%) compared to those without CABG (29%)
(1,65,66). This result has been echoed by several authors (67,68). Sprung and associates
reviewed the recent Cleveland Clinic experience with postoperative myocardial infarction
in nearly 7000 vascular surgical patients (3). While they did not find CABG protective for
the occurrence of AMI, they did find it to be protective against cardiac death. Those with
prior coronary grafting experienced a 79% reduction (27.9 vs. 7.7%; /; = 0.01) in
postoperative death due to cardiac causes.

Fewer reports have studied the effects of preoperative percutaneous coronary revascu-
larization on perioperative outcomes. Several small studies have implicated a protective
effect of percutaneous transluminal coronary angioplasty (PTCA) prior to noncardiac
surgery (69,70). One study by Elmore and colleagues from the Mayo Clinic evaluated 84
patients who underwent CABG and 14 with PTCA before AAA repair (71). They found that
both groups had similar postoperative cardiac event rates and 3-year survival. The PTCA
cohort was able to undergo AAA repair a mean of 10 days after coronary revascularization,
while those undergoing CABG had to wait 68 days. However, the PTCA group had sig-
nificantly more late cardiac events (57 vs. 27%; p = 0.002). Gottlieb and coworkers analyzed
194 patients who had PTCA within 18 months before undergoing vascular surgery (72).
Three-quarters of these patients underwent PTCA within 50 days of surgery and the
postoperative MI rate and occurrence of CHF were both an excellent 0.5%, suggesting a
beneficial effect to preoperative percutaneous revascularization. Recently, the Bypass vs.
Angioplasty Revascularization Investigation (BARI) reported its results in patients under-
going noncardiac surgery believed to require preoperative coronary revascularization (73).
Patients were randomly assigned to undergo CABG or PTCA prior to surgery. The
perioperative cardiac morbidity and mortality were no different between groups; however,
only 11% of the procedures were vascular and only one-third were high risk in nature.
Although small in number and biased by nonvascular operations, data suggest that PTCA
may provide similar perioperative cardiac protection compared to CABG in vascular
surgery patients with CAD appropriate for percutaneous intervention.

Not all reports conclude a salutary effect of coronary revascularization on cardiac ■§

events after vascular surgery (49). Mason and colleagues developed models evaluating the s

effect of preoperative coronary angiography and revascularization on perioperative vas- »

cular surgery results (74). Only when the vascular operative mortality was substantial c

(greater than 5%) and the risks of coronary revascularization low did a mortality benefit <

occur. This benefit was small (less than 1 %), however, and was only present when those with >9

unreconstructable CAD had their vascular surgery canceled. In all models performed, J

patients with preoperative coronary angiography and revascularization had more perioper- «

ative morbidity. Moreover, this strategy doubled the cost of patient care, but long-term |

consequences were not addressed. Massie and associates compared 70 vascular patients who @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 29

underwent coronary angiography due to redistribution on myocardial scintigraphy with 70
matched controls (75). Coronary revascularization was required in 25 of the 70 angiography
patients. Although fewer postoperative cardiac events occurred in the angiography/revas-
cularization cohort, complications secondary to the invasive coronary evaluation led to no
difference in perioperative outcome between groups. Also, the invasively studied/revascu-
larized patients had significantly fewer long-term cardiac events, but long-term survival was
similar. The authors concluded that in vascular surgery patients with abnormal myocardial
scintigraphy, coronary angiography and revascularization did not improve overall out-
comes. The above studies suggest that the additive morbidity and mortality of invasive
coronary evaluation and revasculariztion prior to vascular surgery may offset any benefit.
Thus, currently, there is no consensus regarding the efficacy of coronary revascularization
prior to vascular surgery in preventing postoperative cardiac events. To assist in answering
questions surrounding this controversy, the Coronary Artery Revascularization Prophy-
laxis for Elective Vascular Surgery Study (CARP) is ongoing and should shed light on how
best to approach coronary revascularization perioperatively in vascular patients with
significant CAD (76).

The lack of clear evidence supporting a protective effect of preoperative coronary re-
vascularization on postoperative adverse cardiac events has generated another outlook on
risk stratification. This posture toward preoperative cardiac risk assessment is a minimalist
approach. Taylor et al. prospectively evaluated 491 vascular surgery patients through 534
procedures during a 1-year period (13). The authors were unconvinced that preoperative
cardiac risk screening prior to vascular surgery was efficacious and wanted to study the
outcome with limited use of preoperative testing. Only when clinical evidence suggested
severe CAD did evaluation beyond history, physical examination and resting ECG take
place. In 31 (5.8%) patients, further testing was performed. The overall AMI rate was an
enviable 3.9% and no differences in event rates were found between patients with or
without clinical evidence of CAD or those undergoing detailed cardiac evaluation. Taylor
and colleagues concluded that preoperative cardiac evaluation in vascular patients should
be limited to those with severe symptomatic disease. This minimalist approach has recently
been advocated in several communications by de Virgilio and coworkers. Their work has
described no association between dipyridamole myocardial scintigraphy results and post-
vascular surgery events even in the intermediate-risk category (2,7). They conclude that, as
in selective stratification, those with no clinical risk factors should proceed to surgery and
those with major clinical risk factors continue with extensive coronary evaluation. How-
ever, in the intermediate-risk category, they propose proceeding to surgery under beta
blockade (77,78). Others have found no overall event reduction with preoperative cardiac
evaluation (58,75). Further, these groups argue that the complications of noninvasive
testing proceeding to invasive coronary evaluation and therapy are additive, and even if
they actually decrease the operative cardiac morbidity and mortality after vascular surgery, ■§

the complications due to the other prevascular procedures at least offset the gain (79). Also, |

they suggest that more money is spent per person undergoing extensive coronary evaluation a

for no apparent overall benefit (74). c

As the benefits of aggressive evaluation for CAD before vascular operation in all pa- <

tients remains questioned, and due to concern regarding the high CAD incidence in these >9

individuals, another position regarding preoperative stratification has developed. This J

method is referred to as selective screening for coronary disease. The basis for this approach «

lies in the use of clinical factors and noninvasive testing to identify those patients at increased |

cardiac risk and to intervene in them in some fashion to reduce risk. Thereby the cost, pre- @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

30 CLOUSE and BREWSTER

dictive value, and benefits of preoperative coronary studies may be enhanced. As mentioned
above, dipyridamole-thallium scintigraphy was used by Eagle and colleagues after clinically
scoring 200 patients who were to undergo vascular surgery (Fig. 4) (52). Those with no clin-
ical variables had a low (3%) cardiac event rate and those with three or more clinical
variables had a high risk for events (50%). The intermediate-risk group, however, was fur-
ther stratified by the use of myocardial scintigraphy. If redistribution was present on scan-
ning, 30% of patients suffered a postoperative adverse cardiac event, while those with no
redistribution had just as low a risk as those with no clinical risk factors (3%). This study
initiated the idea of selective screening, as it proved that higher-risk patients could be iden-
tified and therapies initiated in an attempt to reduce risk and that routine study of patients at
clinically low risk is not justified. This selective scheme using clinical and scintigraphy data
was subsequently validated by LTtalien and colleagues in 1081 vascular surgery patients
(1 1). The ability to select patients at higher cardiac risk was also noted by van Damme et al.
(9). They extensively studied 1 56 consecutive patients scheduled for major vascular surgery.
After clinical evaluation and noninvasive testing, 142 remained as 1 1 proceeded to coronary
angiography and 3 more had incomplete data. In those without clinical evidence of CAD,
the postoperative cardiac event rate was 0-3% regardless of dobutamine stress test results.
In those with clinically evident CAD and positive dobutamine stress myocardial scintig-
raphy, the event rate was 25%; in those with positive dobutamine stress echocardiography, it
was 18%. However, if scintigraphy or DSE was negative, the rates were reduced to 1.8 and
4%, respectively. The ability to reduce risk therapeutically within a noninvasively selected
stratification category was recently reported by Boersma and coworkers (44). Using DSE,
the authors found that in those with Lee cardiac risk indices of 3 or greater and NWM A in
one to four segments, those taking beta blockers had a 92% risk reduction compared to
those without beta blockade. Patients with five or more segments having NWMA were un-
affected by beta blockade, as postoperative event rates were 33 and 36% in those not taking
and taking beta blockers, respectively.

Thus, noninvasive evaluation in conjunction with clinical risk stratification may be
used to identify patients who are at highest risk for postoperative cardiac events and lead to
more aggressive therapeutic strategies for CAD in this group. Further, an intermediate-risk
group may be identified, within which proceeding to vascular surgery with risk-reducing
methods, such as beta blockade, is reasonable. Selective stratification is our preferred pos-
ture toward cardiac risk evaluation, given the experience of the Massachusetts General
Hospital with the use of stress myocardial scintigraphy. Our current selective process
incorporating clinical scoring with functional capacity along with noninvasive evaluation
(i.e., myocardial scintigraphy) is represented in Table 5. We believe that it optimizes the
predictive value and cost-effectiveness of noninvasive testing, identifies patients in whom
vascular surgical plans should be altered or canceled when feasible, adequately selects
patients who need further coronary evaluation, and may justifiably identify patients who ■a

can proceed to vascular surgery with risk-reducing strategies. |

The multitude of data in the literature supporting several approaches to preoperative as

cardiac screening and stratification have made this a difficult arena in which to clearly prove c

the best methodology to practice. However, as mentioned, the Prophylactic Coronary <

Artery Revascularization for Elective Vascular Surgery (CARP) trial is a prospective, >9

randomized trial of coronary revascularization (either CABG or PTCA based upon patient J

anatomy) versus no revascularization in intermediate-risk patients who have significant «

clinical risk and/or positive noninvasive testing (76). This trial should help answer questions |

regarding the benefits and proper extent of cardiac evaluation in intermediate-risk patients, @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY

31

Table 5 Selective Approach to Cardiac Stratification

Operative risk

Clinical risk predictors Functional capacity

Approach

(Poor <4 MET)

Emergent case

Coronary revascularization <5 years without symptoms

Recent favorable coronary evaluation

Operation

High risk (i.e., AAA)

Major

Any

Coronary evaluation
(cardiology)

Intermediate

Any

Noninvasive testing b

Minor/none

Good

Operation

Poor

Noninvasive testing b

Intermediate risk

Major

Any

Coronary evaluation

(e.g., CEA)

(cardiology)

Intermediate

Good

Operation

Poor

Noninvasive testing

Minor/none

Any

Operation

Low risk

Major

Any

Coronary evaluation

(e.g., amputation)

(cardiology)

Intermediate

Good

Operation

Poor

Noninvasive testing

Minor/none

Any

Operation

il MET = metabolic equivalent = oxygen consumption of 40 year-old male at rest.
b Noninvasive testing = our preference is myocardial scintigraphy.

the ability of coronary revascularization to protect against postoperative events, and the
comparability of CABG and PTCA prior to vascular surgery.

E. Risk-Reduction Techniques

1. Beta Blockade and Medical Therapies

Although the use and impact of coronary revascularization on cardiac outcome in
vascular surgery patients remains unclear, several other modes of risk reduction are
employed. Use of beta blockade to reduce the risk of postoperative cardiac events has been
intensively studied over the last several years. Mangano et al. studied 200 noncardiac
surgery patients who had clinical evidence of CAD and/or at least two risk factors for
CAD (80,81). The majority of enrolled patients were undergoing major vascular surgery.
The investigators performed a randomized, double-blinded trial of atenolol versus placebo
before induction of anesthesia and continuing through hospitalization; cardiac morbidity
and mortality were evaluated over the ensuing 2 years. Interestingly, patients administered
atenolol perioperatively had significantly lower long-term mortality and overall cardiac
morbidity than those taking placebo (80). During the postoperative period, no differences
in myocardial infarction or cardiac death were evident; however, the atenolol group had
significantly less myocardial ischemia detected on continuous ECG monitoring (81). Sub-
sequently, Poldermans and colleagues evaluated intermediate-risk vascular surgery pa-
tients (at least one clinical risk variable) who were not previously on beta blockers (82). All
patients underwent DSE, and those with extensive NWMA were excluded. The remaining
112 patients were randomized to the addition of beta blockade or standard perioperative

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

32 CLOUSE and BREWSTER

care. Bisoprolol was started 1 week before vascular surgery and titrated to keep the heart
rate at 60; it was then continued for 30 days afterward. Cardiac death and nonfatal AMI
were significantly reduced (91% risk reduction) in the postoperative period in the patients
taking beta blockers. Bisoprolol was continued in the treatment-arm patients postoper-
atively, and these authors have recently reported the long-term effects of beta blockade in
this cohort (83). They found that the protective effect of beta blockade was persistent, as
cardiac death and nonfatal myocardial infarction rates remained significantly less in the
treated group over the next 2 years.

Recently, this same group reported on their overall experience in screening 1351
patients for the above study, and they evaluated the utility of clinical scoring and DSE for
screening and the effects of beta blockade in select groups (44). The authors found the use of
clinical risk scoring (Lee's revised risk index) useful for identification, DSE valuable for
patient stratification, and the use of perioperative beta blockade to be protective for specific
patients. Regarding those either taking beta blockers before evaluation or placed on them
as part of the study protocol, event rates were reduced as clinical risk scoring increased.
Specifically, in those with no clinical risk factors not taking beta blockers, the cardiac event
rate was 1.2%, while no patient taking beta blockers suffered a postoperative event.
Patients with one to three risk factors not taking beta blockers suffered a 9% postoperative
cardiac event rate, versus 3% in those on the medication. Patients not on beta blockers with
more than three clinical risk variables and a normal DSE had a 5.8% adverse cardiac event
rate; this was reduced to 2% for those on beta blockers. In those with one to four NWMA,
the risk of event was reduced 92% by beta blocker usage (33 vs. 2.8%). This reduction led to
a rate approaching that of those patients with no clinical risk factors or NWMA. In the
group of vascular patients with five or more NWMA, beta blockers made no difference in
adverse cardiac outcome, as both groups had unacceptably high event rates (33 vs. 36%).
Although the proper dosage and timing of administration are not clearly defined, these
newer data suggest that vascular surgery patients with known or clinical evidence of CAD
and those with risk factors for CAD should be on perioperative beta blockade for the
purpose of reducing risk of cardiac complications.

Several other medications — such as nitroglycerin, calcium channel blockers, and alpha 2
adrenergic agonists — have been used perioperatively in an attempt to reduce cardiac risk in
vascular and noncardiac surgery patients (24,84,85). Currently, these studies have not been
conclusive enough to make specific recommendations regarding their use in this capacity.
Also of special mention is analgesia selection postoperatively. While considerable benefits to
the use of epidural analgesia in patients at risk for cardiac events after surgery have been
proposed, these benefits remain largely unproven. A stimulating metaanalysis combining
studies wherein randomized epidural use was continued for 24 h or more after surgery has
recently been performed (86). All but one small study consisted of vascular surgery patients,
and a combined total of nearly 1200 patients were included. Although in-hospital deaths ■a

were similar among epidural versus no epidural patients, the postoperative myocardial in- |

farction rate (6.3% overall) was significantly reduced by the use of epidural analgesia (minus as

3.8%; p — 0.05). Further, this reduction was more pronounced (minus 5.3%) with the use of c

thoracic epidural analgesia. Thus, it appears that the use of postoperative epidural analgesia <

may reduce the risk of postoperative AMI in vascular surgery patients. >5

I
2. Monitoring q

Initial reports on pulmonary artery catheter (PAC) use to aid in the management of |

vascular surgery patients to prevent adverse cardiac events were encouraging (87). However, @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 33

over the last several decades a modest body of literature has developed evaluating the
morbidity and mortality effects of such PAC use (88-90). In general, these studies have
found no reduction in perioperative cardiac events. A randomized, prospective trial by
Valentine and coworkers found no difference in overall or cardiac complications based on
PAC use in aortic surgery patients (91). Those with recent AMI, severe valvular disease, or
unstable coronary syndromes were excluded. Patients with PAC placement were preoper-
atively "optimized" by Starling curve performance. Intriguingly, those with PAC use had
significantly more cardiopulmonary intraoperative complications and received more intra-
venous fluid. The authors recommended continuing use of central venous access during
aortic surgery, however, for contingency purposes and to expedite PAC placement in those
with situations where it is deemed necessary. These findings were reinforced by Ziegler et al
(4), who performed a randomized, prospective trial of preoperative hemodynamic "opti-
mization" using the PAC to achieve central venous oxygen saturations of greater than
65%. The control group had PAC placement without therapy to achieve "optimization."
No differences in cardiovascular complications were found between groups, suggesting
that the use of PAC did not reduce complications. A recent strict metaanalysis of the pro-
spective, randomized studies of PAC use in moderate-risk vascular surgery patients by
Barone et al indicated no difference in postoperative mortality or morbidity based on man-
agement with a PAC (92). We concur that in the patient with moderate cardiac risk under-
going major vascular surgery the use of a PAC generally does not improve postoperative
cardiac outcome and is not indicated. While not clearly delineated, these catheters may still
have a place in high-risk patients and in select patients with cardiac complications after
vascular surgery.

The use of transesophageal echocardiography (TEE) intraoperatively to direct fluid
management and pharmacological modification of cardiac function during vascular
surgery has been described (93). Given the excellent quality of today's echocardiography
and Doppler measurement capabilities, this seemingly could improve or add to patient
management and perhaps reduce events in this patient population. Further, TEE is con-
sidered to detect myocardial dysfunction not observed by standard monitoring techniques
(94). Studies have suggested some degree of therapeutic usefulness in vascular and non-
cardiac surgery (95-98). However, prospective data supporting and defining the use of
TEE are lacking (31). Currently, it seems prudent to consider its use in high-risk patients,
such as those with valvular disease, congestive heart failure, or recent AMI requiring
vascular surgery.

F. Valvular Disease

As congestive heart failure (CHF) and left-sided valvular lesions have been shown to be

independent predictors of adverse cardiac events after surgery in several analyses, the ■a

impact valvular lesions may have on a vascular patient's course are significant. Goldman |

and associates reported a 20% incidence of postoperative CHF in patients with valvular a

disease (99,100). Stenotic lesions are particularly concerning, as they have been implicated c

in increasing postoperative mortality to the range of 10-20% as well as perioperative CHF <

and shock (31,99-102). And while a few small studies have suggested that recent advances >9

in anesthetic technique may have attenuated some of the risk in asymptomatic patients, J

severe and symptomatic aortic and mitral stenosis should be evaluated and repaired, either «

by valvulotomy or replacement, prior to major surgery (31,103,104). Patients with stenotic |

valvular lesions should be well hydrated, and both tachycardia and afterload reduction @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

34 CLOUSE and BREWSTER

should be avoided. Aortic or mitral regurgitation is, on the whole, better tolerated than
stenotic lesions. When left ventricular function is normal, noncardiac surgery may be
entertained with vigilant medical control and monitoring. However, symptomatic regur-
gitation should also undergo critical preoperative evaluation, as patients with reduced
ventricular function or multiple lesions may well be served by addressing these before
surgery. Proper fluid balance, afterload reduction, and postoperative diuresis are critical in
these patients. In patients with valvular disease, perioperative use of a PAC should be
strongly considered to aid in the above endeavors. Bacterial endocarditis prophylaxis
should be adhered to according to the AHA guidelines whenever patients with valvular
disease are being operated on (105).

III. PULMONARY COMPLICATIONS

A. Importance in Peripheral Vascular Patients

Pulmonary complications are a significant cause of morbidity after vascular surgery. While
not as extensively studied as cardiac complications, postoperative respiratory issues are
common and cause significant prolongation of intensive care unit (ICU) and hospital stay
(124,125). The impact pulmonary issues may have is not surprising. Advancing age, tobacco
abuse associated with chronic obstructive pulmonary disease (COPD), and other comorbid
states frequently accompany peripheral vascular disease. Vascular operations may also be
extensive, involving the thoracic and abdominal cavities and leading to inhibition of the
patient's ability to breathe fully. The incidence of pulmonary complications after vascular
surgery is dependent on how they are defined and the vascular surgery performed. Aortic
operations carry a more significant risk than those for peripheral occlusive disease or carotid
artery disease (126). Reported rates vary from as high as 40% for thoracoabdominal
aneurysm repair to the 1-2% range for carotid endarterectomy (126-128). Typical respi-
ratory complications that may occur include prolonged mechanical ventilation, need for
reintubation, respiratory failure, pneumonia, symptomatic pleural effusion, atelectasis/
collapse, and the acute respiratory distress syndrome (ARDS). Although the specifics of
these entities are beyond the scope of this chapter, great advances have been made in the
diagnosis and treatment of postoperative pneumonia and ARDS. These issues are critically
reviewed elsewhere (129-131). Our focus is on the identification of those at risk for these
complications and on a review of risk assessment and risk-reducing strategies.

B. Clinical Risk Assessment

Albeit perhaps not as extensive as the evaluation of clinical risk factors for cardiac
complications, studies have evaluated preoperative clinical variables attempting to identify

those at highest risk for pulmonary complications after vascular and noncardiac surgery. ■§

Important factors generally implicated are increasing age, smoking, COPD, chronic cough |

with changes in sputum character and amount, general functional status [both American a

Society of Anesthesiologists (ASA) and Goldman cardiac risk criteria], obesity, type and c

length of surgery, anesthesia used, and spirometric findings. These variables can be orga- <

nized into patient- and procedure-specific items. The data behind the assertion of these risk >9

factors have recently been reviewed (132). Many of these evaluations have been performed in J

noncardiac abdominal surgery patients (124,133-135). °

Four studies have been performed in patients undergoing aortic aneurysm surgery. |

Money and coworkers studied factors predictive of postoperative respiratory failure in 100 @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 35

consecutive thoracoabdominal aneurysm (TAA) resections (136). Increasing age, aneurysm
extent, amount of intraoperative blood transfusion, postoperative renal insufficiency, and
pneumonia were found to be independent correlates of respiratory failure. A similar report
by Svensson et al., in 1400 TAA patients, found that 112 (8%) had postoperative res-
piratory failure, with COPD, smoking, complications in other organ systems (cardiac and
renal) and spirometric evaluation, specifically the mid-expiratory flow (FEF25), being pre-
dictive (137). More recently, advanced age, smoking, aortic cross-clamp time, operative
transfusion, and diaphragmatic division were found to be independent predictors of pro-
longed ventilation after 387 TAA or thoracic aneurysm repairs by Engle and colleagues
(138). Age, obesity, and spirometric findings were also found to be important risk factors
for pulmonary complications by Calligaro and associates in elective abdominal aortic re-
construction (139).

Recently, the National Veterans Administration Surgical Quality Improvement Pro-
gram reported a clinical risk index predicting respiratory failure after noncardiac surgery
(140). Nearly 100,000 male patients were studied after model development, and post-
operative respiratory failure developed in 2746 (3.4%). Only clinical variables were con-
sidered. No objective preoperative testing data were included. Independently predictive
factors included type of surgery, serum albumin, blood urea nitrogen levels, independent
versus dependent functional status, COPD, and age. These critical factors were score-
weighted and a risk index developed (Table 6). The index is based upon point totals and its
categorization is presented in Table 7 and correlates with increasing risk of postoperative
respiratory failure. Of note, patients undergoing AAA repair and peripheral vascular
procedures were 14 times and 4 times more likely, respectively, to develop respiratory
failure. In fact, AAA repair was the operative procedure at highest risk. The authors sug-
gested this risk index be used to identify patients at increased risk for postoperative pul-

Table 6 Preoperative Clinical Predictors of Postoperative Respiratory Failure with Respiratory
Failure Index"

Variable

Odds ratio (95% CI)

Point value

Type of surgery

Abdominal aortic aneurysm

14.3 (12.0-16.9)

27

Thoracic

8.14 (7.17-9.25)

21

Peripheral vascular, upper abdominal, or neurosurgery

4.21 (3.80-4.67)

14

Neck

3.10 (2.40-4.01)

11

Emergency surgery

3.12 (2.83-3.43)

11

Albumin (<30 g/dL)

2.53 (2.28-2.80)

9

Blood urea nitrogen (>30 mg/dL)

2.29 (2.04-2.56)

8

Partially or fully dependent living status

1.92(1.74-2.11)

7

History of chronic obstructive pulmonary disease

1.81 (1.66-1.98)

6

Age (years)

>70

1.91 (1.71-2.13)

6

60-69

1.51 (1.36-1.69)

4

a Respiratory failure defined as mechanical ventilation for more than 48 h or reintubation after

extubation.

Source: Ref. 140.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

36

CLOUSE and BREWSTER

Table 7 Respiratory Failure Index"

Predicted probability 13

Point total

of respiratory failure

<10

0.5%

11-19

2.2%

20-27

5.0%

28-40

11.6%

>40

30.5%

11 Respiratory failure defined as mechan-
ical ventilation for more than 48 h or re-
intubation after extubation.
b Prediction based on model development
with cohort confirmation.
Source: Ref 140.

monary insufficiency who may benefit from further objective pulmonary testing prior to
operation.

C. Preoperative Evaluation

A detailed history and physical examination are invaluable in identifying patients with
significant pulmonary disease. Physical examination should detect hypoventilation in weak
or debilitated patients, and wheezing, rhonchi, prolonged expiration, and poor air move-
ment in those with COPD. Lung examination abnormalities clearly increase the risk of
adverse pulmonary events after operation (124). Although the degree of dyspnea on exertion
may be an indicator of significant pulmonary insufficiency, many vascular patients lead
sedentary lives, and the historical information given is of limited value. Obviously, as
indicated above, obtaining a smoking history is of critical value. The nature and quantity of
sputum production may provide insight into chronicity of disease and underlying infection.
Recent studies have found sputum production and chronic cough prior to noncardiac
surgery to be independent correlates with pulmonary complications (133,135). Preoperative
(or recent) anteroposterior and lateral chest x-ray is indicated in those over 40 years of age
with a history of or examination evidence of cardiopulmonary disease, smoking history or
active symptoms. Ostensibly, based on these criteria, chest x-ray should be routine for nearly
all vascular surgery patients.

Once a patient has been deemed at risk for pulmonary complications after vascular
surgery by clinical and physical stratification, further pulmonary function testing (PFT)
and arterial blood gas (ABG) evaluation is indicated. Pulmonary function testing provides
both volume (Fig. 5) and flow information implicating type and extent of lung disease
present, while room-air ABG analysis indicates the degree of current compensation. Most
evidence suggests association between PFT abnormalities — specifically flow variables such
as FEV! and FEF 2 s — and elevated Pa CQ 2 levels and the development of postoperative
pulmonary complications (Table 8) (126,132,137,141). This is not universal in the litera-
ture, however, and although there are no agreed upon criteria whereby pulmonary func-
tion is in and of itself prohibitive for operation, they may aid in identifying those with such
severe disease that it is prudent to alter the operation or anesthesia, cancel elective surgery,
or aggressively use risk-reduction strategies perioperatively (132).

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I
•5

I

COMPLICATIONS OF VASCULAR SURGERY

37

Figure 5 Spirometric lung volumes obtained during pulmonary function testing (PFT).
TLC = total lung capacity; TV = tidal volume; FRV = functional residual volume; IC = inspiratory
capacity; ERV = expiratory reserve volume; RV = residual volume; VC = vital capacity.

D. Risk Reduction

Vascular surgery procedures on the thoracic and abdominal aorta require thoracotomy or
laparotomy incisions. These incisions alter chest wall, diaphragmatic, and lung function.
Postoperative incisional pain causes splinting and decreases both tidal volume and func-
tional residual capacity (142). This pain also prevents the patients from coughing effec-
tively. The cumulative effects of these mechanical problems is closure of small airways
(microatelectasis) due to decreased volumes and bronchial occlusion and regional atelec-

Table 8 Preoperative Pulmonary Function Assessment

Normal

High Risk

Vital capacity

Forced expiratory volume, 1 s (FEV!)

Maximal midexpiratory flow (FEF 25 % 7;

Maximal voluntary ventilation

Pa 0i , room air

Paco,i room air

30-50 mL/kg; > 80% predicted < 30-50%

> 80% predicted < 40-50%

% ) 150-200 L/min; > 80% predicted < 35-50%

150-500 L/min; > 80% predicted < 35-50%

85 + 5 mmHg < 50-55 mmHg

40 + 4 mmHg > 45-55 mmHg

I
•5

Souce: Ref. 159.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

38 CLOUSE and BREWSTER

tasis (macroatelectasis) from poor pulmonary toilet (132,143,144). While enthusiasm for
retroperitoneal approaches for abdominal aortic surgery as opposed to transperitoneal
operations has been purported to reduce pulmonary complications this remains unproven;
our own prospective, randomized study revealed no benefit (145,146). Those identified by
clinical and objective data to be at highest risk for postoperative pulmonary complications
should be strongly considered for endovascular or extra-anatomic revascularization to re-
duce this potential.

Methods to expand the lungs and prevent this micro- and macrocollapse are proven
and easily accomplished. Two maneuvers with equal efficacy are deep-breathing exercises
and incentive spirometry (147-149). The greatest benefit by these modalities is achieved by
preoperative education and training prior to surgery (147,150). When patients are unable
to utilize these methods adequately after vascular surgery, intermittent positive-pressure
breathing (IPPB) or continuous positive airway pressure (CPAP) are beneficial, as they are
not dependent upon patient effort. However, complications occur more frequently, and
these measures are more expensive than patient-driven modalities such as incentive spi-
rometry (144,147,151). Chest physiotherapy and nasotracheal suctioning may be indicated
for those with substantial atelectasis or preexisting lung disease causing problems with the
clearance of secretions (144).

Beside preoperative education and training regarding lung-expansion techniques, other
preoperative strategies to reduce risk include smoking cessation at least 8 weeks before
surgery, optimization of respiratory flow dynamics in those with obstructive disease, and
treatment of any present respiratory infection well in advance of surgery. The impact that
cigarette smoking has on postoperative pulmonary complications was recently reiterated
by Bluman and associates (135). Patients undergoing noncardiac surgery who were smok-
ing at the time of operation were over five times more likely to have lung-related compli-
cations compared to patients who had never smoked. Intriguing was the fact that among
smokers, those who reduced or quit smoking within 1 month of surgery were at signifi-
cantly higher risk of pulmonary complications than those who did not. Warner et al. studied
patients undergoing CABG and found that those patients who quit smoking within 2
months of surgery had more pulmonary complications than those who quit at least 2 months
before surgery (152). While unexplained at present, these data suggest that, when possible,
smoking cessation should be strongly encouraged to occur at least 8 weeks before a planned
operation. Patients with obstructive lung disease are at increased risk for lung-related com-
plications postoperatively. When patients remain symptomatic prior to vascular surgery,
pulmonary medical consultation is wise. Ipratropium, beta agonists, and methylxanthines
should be added to the therapy regimen serially so as to maximize flow and the clearance of
secretions (132). When patients remain symptomatic despite aggressive, escalating therapy,
corticosteroids for 14 days may be initiated with minimal sequelae (153). Although there
is no place for antibiotics for preoperative "preparation," treatment of respiratory tract ■§

infection should be completed with clear resolution prior to entertaining operation. |

The conduct and effects of anesthesia have a profound influence on ventilatory media- as

nics and pulmonary function postoperatively. General anesthesia alters the mechanical c

properties of the chest wall and lung. The required positive-pressure circuit leads to atel- <

ectasis in the dependent areas, with decreased compliance. Inhalation anesthetics cause >9

respiratory muscle dysfunction and incoordination, thus contributing to this phenomenon J

(143). Pulmonary problems have been identified in those under neuromuscular blockade «

with pancuronium as opposed to the shorter-acting agents (154). This effect is related to the |

longer duration of action of pancuronium, thus leading to longer durations of neuro- @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 39

muscular blockade and hypoventilation. To this end, Mitchell and coworkers found the
duration of anesthesia to be an independent predictor of pulmonary complications after
elective surgery (133). The use of neuroaxial blockade (epidural/spinal) techniques, either as
"stand alone" anesthetic methods or in conjunction with general anesthesia, has proven
helpful in reducing postoperative pulmonary complications (132,134,143,155-158). These
strategies allow for excellent postoperative analgesia as well and reducing the medications
depressing respiratory function, such as systemic opioids, required both during operation
and postoperatively. This currently allows extubation in the operating room in over 90% of
our patients undergoing major aortic operations.

IV. CONCLUSION

The typical vascular surgery patient presents with systemic atherosclerosis, which includes
CAD to varying degrees as well as pulmonary disease such as COPD due to smoking. This
patient population is at highest risk for postoperative cardiac and pulmonary morbidity
and mortality. A dedicated approach to the preoperative identification, evaluation, and
stratification of patients with regard to cardiopulmonary risk helps to maximize benefit
when these patients undergo vascular procedures. Further, long-term outcomes beyond the
procedure at hand are also affected. Preoperative preparation, intraoperative techniques,
alternative surgical approaches, and postoperative care may all play a role in addressing
cardiopulmonary problems in this population, and the ability to approach these topics
sensibly is one of the most important in vascular surgery.

REFERENCES

1. Hertzer NR, Beven EG, Young JR, O'Hara PJ, Ruschhaupt WF III, Graor RA, DeWolfe
VG, Maljoved LC. Coronary artery disease in peripheral vascular patients: A classification of
1000 coronary angiograms and results of surgical management. Ann Surg 1984; 199:223-233.

2. de Virgilio C, Foosie K, Ephraim L, Elbassir M, Donayre C, Baker D, Narahara K, Mishkin
F, Lewis RJ, Chang C, White R, Mody FV. Dipyridamole-thallium/sestamibi before vascular
surgery: A prospective blinded study in moderate-risk patients. J Vase Surg 2000; 32:77-89.

3. Sprung J, Abdelmalak B, Gottlieb A, Mayhew C, Hammel J, Levy PJ, O'Hara PJ, Hertzer
NR. Analysis of risk factors for myocardial infarction and cardiac mortality and after major
vascular surgery. Anesthesiology 2000; 93:129-140.

4. Ziegler DW, Wright JG, Choban PS, Flancbaum L. A prospective randomized trial of
preoperative "optimization" of cardiac function in patients undergoing elective peripheral
vascular surgery. Surgery 1997; 122:584-592.

5. Samain E, Farah E, Leseche G, Marty J. Guidelines for perioperative cardiac evaluation from

the American College of Cardiology /American Heart Association task force are effective for ■o

stratifying cardiac risk before aortic surgery. J Vase Surg 2000; 31:91-99. §

6. Aziz I, Lewis JR, Baker JD, de Virgilio C. Cardiac morbidity and mortality following carotid S
endarterectomy: Fhe importance of diabetes and multiple eagle risk factors. Ann Vase Surg. j|
2001; 15:243-246. |

7. de Virgilio C, Wall DB, Ephraim L, Foosie K, Donayre C, White R, Elbassir M. An abnormal 6
dipyridamole thallium/sestamibi fails to predict long-term cardiac events in vascular surgery 'In-
patients. Ann Vase Surg 2001; 15:267-271. J

8. de Virgilio C, Foosie K, Lewis RJ, Stabile BE, Baker JD, White R, Donayre CE, Ephraim L. 2
Cardiac morbidity and operative mortality following lower-extremity amputation: Fhe I
significance of multiple Eagle criteria. Ann Vase Surg 1999; 13:204-208. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

40 CLOUSE and BREWSTER

9. van Damme H, Pierard L, Gillain D, Benoit T, Rigo P, Limet R. Cardiac risk assessment
before vascular surgery: A prospective study comparing clinical evaluation, dobutamine stress
echocardiography, and dobutamine Tc-99m sestamibi tomoscintigraphy. Cardiovasc Surg
1997; 5:54-64.

10. L'ltalien GJ, Cambria RP, Cutler BS, Leppo JA, Paul SD, Brewster DC, Hendel RC, Abbott
WM, Eagle KA. Comparative early and late morbidity among patients requiring different
vascular surgery procedures. J Vase Surg 1995; 21:935-944.

1 1 . L'ltalien GJ, Paul SD, Hendel RC, Leppo J A, Cohen MC, Fleisher LA, Brown KA, Zarich
SW, Cambria RP, Cutler BS, Eagle KA. Development and validation of a Bayesian model for
perioperative cardiac risk assessment in a cohort of 1 ,08 1 vascular surgical candidates. J Am
Coll Cardiol 1996; 27:779-786.

12. de Virgilio C, Bui H, Donayre C, Ephraim L, Lewis RJ, Elbassir M, Stabile BE, White RE.
Endovascular vs. open abdominal aortic aneurysm repair. Arch Surg 1999; 134:947-951.

13. Taylor LM, Yeager RA, Moneta GL, McConnell DB, Porter JM. The incidence of peri-
operative myocardial infarction in general vascular surgery. J Vase Surg 1991; 15:52-61.

14. Krupski WC, Layug EL, Reilly LM, Rapp JH, Mangano DT. Comparison of cardiac mor-
bidity rates between aortic and infrainguinal operations: Two-year follow-up. J Vase Surg
1993; 18:609-617.

15. Krupski WC, Layug EL, Reilly LM, Rapp JH, Mangano DT. Comparison of cardiac mor-
bidity rates between aortic and infrainguinal operations. J Vase Surg 1992; 15:354-365.

16. Lee TH, Marcantonio ER, Mangione CM, Thomas EJ, Polanczyk CA, Cook EF, Surgarbaker
DJ, Donaldson MC, Poss R, Ho KKL, Ludwig LE, Pedan A, Goldman L. Derivation and
prospective validation of a simple index for prediction of cardiac risk of major noncardiac
surgery. Circulation 1999; 100:1043-1049.

17. Yeager RA. Basic data related to cardiac testing and cardiac risk associated with vascular
surgery. Ann Vase Surg 1990; 4:193-197.

18. Pasternak PF, Grossi EA, Baumann G, Riles TS, Lamparello PJ, Giangola G, Primis LK,
Mintzer R, Imparato AM. The value of silent myocardial ischemia monitoring in the pre-
diction of perioperative myocardial infarction in patients undergoing peripheral vascular sur-
gery. J Vase Surg 1989; 10:617-625.

19. Mangano DT, Hollenberg M, Fegert G, Meyer LM, London MJ, Tubau JF, Krupski WC.
Perioperative myocardial ischemia in patients undergoing noncardiac surgery: I. Incidence
and severity during the 4 day perioperative period. J Am Coll Cardiol 1991; 17:843-850.

20. Sear JW, Foex P, Howell SJ. Effect of chronic intercurrent medication with p-adrenoceptor
blockade on postoperative silent myocardial ishemia. Br J Anaesth 2000; 84:311-315.

21. Higham H, Sear JW, Neill F, Sear YH, Foex P. Peri-operative silent myocardial ischaemia and
long-term adverse outcomes in non-cardiac surgical patients. Anaesthesia 2001; 56:630-637.

22. Mamode N, Scott RN, McLaughlin SC, McLelland A, Pollock JG. Perioperative myocardial
infarction in peripheral vascular surgery. Br Med J 1996; 312:1396-1397.

23. Landesberg G, Mosseri M, Zahger D, Wolf Y, Perouansky M, Anner H, Drenger B, Hasin Y,
Berlatzky Y, Weissman C. Myocardial infarction after vascular surgery: The role of pro-
longed, stress-induced, ST-depression-type ischemia. J Am Coll Cardiol 2001; 37:1839-1845.

24. Oliver MF, Goldman L, Julian DG, Holme I. The effect of mivazerol on perioperative cardiac j>
complications during non-cardiac surgery in patients with coronary artery disease: The Euro- <S
pean Mivazerol Trial (EMIT). Anesthesiology 1999; 91:951-961. |

25. Mangano DT, Wong MG, London MJ, Tubau JF, Rapp JA. Perioperative myocardial °|
ischemia in patients undergoing noncardiac surgery: II. Incidence and severity during the 1st **
week after surgery. J Am Coll Cardiol 1991; 17:851-857. |.

26. Mangano DT. Perioperative cardiac morbidity. Anesthesiology 1990; 72:153-184. J

27. Hertzer NR, Avellone JC, Farrell CJ, Plecha FR, Rhodes RS, Sharp WV, Wright GF. The q
risk of vascular surgery in a metropolitan community with observations on surgeon experience |
and hospital size. J Vase Surg 1984; 1:13-21. 2

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 41

28. Hertzer NR. Fatal myocardial infarction following lower extremity revascularization. Ann
Surg 1981; 193:492^198.

29. Hertzer NR. Fatal myocardial infarction following abdominal aortic aneurysm resection. Ann
Surg 1980; 192:667-673.

30. Eagle KA, Brundage BH, Chaitman BR, Ewy GA, Fleisher LA, Hertzer NR, Leppo JA, Ryan
T, Schlant RC, Spencer WH III, Spittell JA Jr, Twiss RD. Guidelines for perioperative
cardiovascular evaluation for noncardiac surgery: Report of the American College of Car-
diology/American Heart Association Task Force on Practice Guidelines (Committee on Peri-
operative Cardiovascular Evaluation for Noncardiac Surgery). Circulation 1996; 93:1278-
1317.

31. Eagle KA, Berger PB, Calkins H, Chaitman BR, Ewy GA, Fleischmann KE, Fleisher LA,
Froehlich JB, Gusberg RJ, Leppo JA, Ryan T, Schlant RC, Winters WL Jr, ACC/AHA
Guidelines update for perioperative cardiovascular evaluation for noncardiac surgery-executive
summary: A report of the American College of Cardiology/ American Heart Association Task
Force on Practice Guidelines (Committee to update the 1996 guidelines n perioperative
cardiovascular evaluation for noncardiac surgery). Circulation 2002; 105:2157-1267.

32. Cutler BS, Wheeler HB, Paraskos JA, Cardullo PA. Applicability and interpretation of
electrocardiographic stress testing in patients with peripheral vascular disease. Am J Surg
1981; 141:501-506.

33. Raby KE, Goldman L, Creager MA, Cook EF, Weisberg MC, Whittemore AD, Selwyn AP.
Correlation between preoperative ischemia and major cardiac events after peripheral vascular
surgery. N Engl J Med 1989; 321:1296-1300.

34. Mangano DT, Hollenberg M, Fegert G, Meyer ML, London MJ, Tubau JF, Krupski WC.
Perioperative myocardial ischemia in patients undergoing noncardiac surgery: I. Incidence
and severity during the 4 day perioperative period. The Study of the Perioperative Ischemia
(SPI) Research Group. J Am Coll Cardiol 1991; 17:851-857.

35. Landesberg G, Einav S, Christopherson R, Beattie C, Berlatzky Y, Rosenfeld B, Eidelman
LA, Norris E, Anner H, Mosseri M, Cotev S, Luria MH. Perioperative ischemia and cardiac
complications in major vascular surgery: Importance of the preoperative twelve-lead electro-
cardiogram. J Vase Surg 1997; 26:570-578.

36. Kirwin JD, Ascer E, Gennana M, Mohan C, Jonas S, Yorkovich W, Matano R. Silent myo-
cardial ischemia is not predictive of myocardial function in peripheral vascular surgery patients.
Ann Vase Surg 1993; 7:27-32.

37. Pasternack PF, Imparato AM, Bear G, Riles TS, Baumann FG, Benjamin D, Sanger J, Kramer
E, Wood RP. The value of radionuclide angiography as a predictor of perioperative myocardial
infarction in patients undergoing abdominal aortic aneurysm resection. J Vase Surg 1984; 1:
320-325.

38. Kazmers A, Cerquieira MD, Zierler RE. The role of preoperative radionuclide left ventricular
ejection fraction for risk assessment in carotid surgery. Arch Surg 1988; 123:416^419.

39. Kazmers A, Cerquieira MD, Zierler RE. The role of preoperative radionuclide ejection fraction
in direct abdominal aortic aneurysm repair. J Vase Surg 1988; 8:128-136.

40. Franco CD, Goldsmith J, Veith FJ, Ascer E, Wengerter KR, Calligaro KD, Gupta SK. Resting

gated pool ejection fraction: A poor predictor of perioperative myocardial infarction in pa- j>

tients undergoing vascular surgery for infrainguinal bypass grafting. J Vase Surg 1989; 10:656- <S

661- I

41. Beller GA, Gibson RS. Sensitivity, specificity, and prognostic significance of noninvasive test- 2
ing for occult or known coronary artery disease. Prog Cardiovasc Dis 1987; 29:241-270. ^

42. Lee TH, Boucher CA. Noninvasive tests in patients with stable coronary artery disease. N >9
Engl J Med 2001; 344:1840-1845. J

43. Poldermans D, Arnese M, Fioretti PM, Salustri A, Boersma E, Thomson IR, Roelandt JRTC, q
van Urk H. Improved cardiac risk stratification in major vascular surgery with dobutamine- |
atropine stress echocardiography. J Am Coll Cardiol 1995; 26:648-653. S

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

42 CLOUSE and BREWSTER

44. Boersma E, Poldermans D, Bax JJ, Steyerberg EW, Thomson IR, Banga JD, Van de ven LLM,
van Urk H, Roelandt JRTC. Predictors of cardiac events after major vascular surgery: Role of
clinical characteristics, dobutamine echocardiography, and [^-blocker therapy. JAMA 2001;
285:1865-1873.

45. Shaw LL, Eagle KA, Gersh BJ, Miller DD. Meta-analysis of intravenous dipyridamole-
thallium-201 imaging (1985-1994) and Dobutamine echocardiography (1991-1994) for risk
stratification before vascular surgery. J Am Coll Cardiol 1996; 27:787-798.

46. Boucher CA, Brewster DC, Darling RC, Okada RD, Strauss HW, Pohost GM. Determi-
nation of cardiac risk by dipyridamole-thallium imaging before peripheral vascular surgery.
N Engl J Med 1985; 312:389-394.

47. Eagle KA, Strauss HW, Boucher CA. Dipyridamole myocardial perfusion imaging for coro-
nary heart disease. Am J Cardiac Imaging 1988; 2:230-292.

48. Vanzetto G, Machecourt J, Blendea D, Fagret D, Borrel E, LucMagne J, Gattaz F, Guidicelli
H. Additive value of thallium single-photon emission computed tomography myocardial
imaging for prediction of perioperative events in clinically selected high cardiac risk patients
having abdominal aortic surgery. Am J Cardiol 1996; 77:143-148.

49. Cutler BS, Leppo JA. Dipyridamole thallium 201 scintigraphy to detect coronary artery dis-
ease before abdominal aortic surgery. J Vase Surg 1987; 5:91-100.

50. Eagle KA, Singer DE, Bewster DC, Darling RC, Mulley AG, Boucher CA. Dipyridamole-
thallium scanning in patients undergoing vascular surgery: Optimizing preoperative evalua-
tion of cardiac risk. JAMA 1987; 257:2185-2189.

51. Brewster DC, Okada RD, Strauss HW, Abbott WM, Darling RC, Boucher CA. Selection of
patients for preoperative coronary angiography: Use of dipyridamole-stress thallium myo-
cardial imaging. J Vase Surg 1985; 2:504-509.

52. Eagle KA, Coley CM, Newell JB, Brewster DC, Darling RC, Strauss HW, Guiney TE, Boucher
CA. Combining clinical and thallium data optimizes preoperative assessment of cardiac risk
before major vascular surgery. Ann Intern Med 1989; 1 10:859-866.

53. Levinson JR, Boucher CA, Coley CM, Guiney TE, Strauss HW, Eagle KA. Usefulness of
semiquantitative analysis of dipyridamole-thallium-201 redistribution for improving risk strati-
fication before vascular surgery. Am J Cardiol 1990; 66:406^110.

54. Hendel RC, Whitfield SS, Villegas BJ, Cutler BS, Leppo JA. Prediction of late cardiac events
by dipyridamole thallium imaging in patients undergoing elective vascular surgery. Am J
Cardiol 1992; 70:1243-1249.

55. Lette J, Waters D, Cerino M, Picard M, Champagne P, Lapointe J. Preoperative coronary
artery disease risk stratification based on dipyridamole imaging and a simple three-step, three-
segment model for patients undergoing noncardiac vascular surgery or major general surgery.
Am J Cardiol 1992; 69:1553-1558.

56. Mangano DT, London MJ, Tubau JF, Browner WS, Hollenberg M, Krupski W, Layug EL,
Massie B. Diypridamole thallium-201 scinitgraphy as a preoperative screening test: A reexam-
ination of its predictive potential. Circulation 1991; 84:493-502.

57. Schueppert MT, Kresowik TF, Corry DC, Jacobovicz C, Mohan CR, Slaymaker E, Hoballah
JJ, Sharp WJ, Grover-McKay M, Corson JD. Selection of patients for cardiac evaluation

before peripheral vascular operations. J Vase Surg 1996; 23:802-809. g

58. Seeger J, Rosenthal G, Self S, Flynn TC, Limacher MC, Harward TR. Does routine stress- <S
thallium cardiac screening reduce postoperative cardiac complications? Ann Surg 1994; 6: js
654-663. Jf

59. Stratmann H, Younis L, Wittry M, Amato M. Miller DD. Diypridamole technetium-99m <
sestamibi myocardial tomography in patients evaluated for elective vascular surgery: Prog- >9
nostic value for perioperative and late cardiac events. Am Heart J 1995; 131:923-929. ^

60. Baron JF, Mundler O, Bertrand M, Vicaut E, Barre E, Godet G, Samama M, Coriat P, q
Kieffer E, Viars P. Dipyridamole-thallium scintigraphy and gated radionuclide angiography |
to assess cardiac risk before abdominal aortic surgery. N Engl J Med 1994; 330:663-669. 2

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 43

61. Poldermans D, Mariarosaria A, Fioretti PM, Boersma E, Thomson IR, Rambaldi R, van Urk
H. Sustained prognostic value of dobutamine stress echocardiography for late cardiac events
after major noncardiac vascular surgery. Circulation 1997; 95:53-58.

62. Cohen MC, Curran PJ, L'ltalien, Mittleman MA, Zarich SW. Long-term prognostic value of
preoperative dipyridamole thallium imaging and clinical indexes in patients with diabetes
mellitus undergoing peripheral vascular surgery. Am J Cardiol 1999; 83:1038-1042.

63. Fleisher LA, Eagle KA, Shaffer T, Anderson GF. Perioperative and long-term mortality rates
after major vascular surgery: The relationship to preoperative testing in the medicare popu-
lation. Anesth Analg 1999; 89:849-855.

64. Eagle KA, Rihal CS, Mickel MC, Holmes DR, Foster ED, Gersh BJ. Cardiac risk of non-
cardiac surgery: Influence of coronary disease and type of surgery in 3368 operations. Circu-
lation 1997; 96:1882-1997.

65. Hertzer NR. The natural history of peripheral vascular disease: Implications for its manage-
ment. Circulation 1983; 83(S1):1-12.

66. Hertzer NR, Young J, Beven E, O'Hara PJ, Graor RA, Ruschhaupt WF, Maljovec LC. Late
results of coronary bypass in patients with peripheral vascular disease: II. Five-year survival
according to sex, hypertension, and diabetes. Cleve Clin J Med 1987; 54:15-23.

67. Rihal CS, Eagle KA, Mickel MC, Foster ED, Sopko G, Gersh BJ. Surgical therapy for coro-
nary artery disease among patients with combined coronary artery and peripheral vascular
disease. Circulation 1995; 91:46-53.

68. Toal K, Jacocks M, Elkins R. Preoperative coronary artery bypass grafting in patients
undergoing abdominal aortic reconstruction. Am J Surg 1984; 148:825-829.

69. Huber KC, Evans MA, Bresnahan JF, Gibbons RJ, Holmes DR Jr. Outcome of noncardiac
operations in patients with severe coronary artery disease successfully treated preoperatively
with coronary angioplasty. Mayo Clin Proc 1992; 67:15-21.

70. Posner KL, van Norman G, Chan V. Adverse cardiac outcomes after noncardiac surgery in
patients with prior percutaneous transluminal coronary angioplasty. Anesth Analg 1999;
89:553-560.

71. Elmore JR, Hallett JW Jr, Gibbons RJ, Naessens JM, Bower TC, Cherry KJ, Gloviczki P,
Pairoleiro PC. Myocardial revascularization before abdominal aortic aneurysmorrhaphy:
Effect of coronary angioplasty. Mayo Clin Proc 1993; 68:637-641.

72. Gottlieb A, Banoub M, Sprung J, Levy PJ, Beven M, Mascha EJ. Perioperative cardio-
vascular morbidity in patients with coronary artery disease undergoing vascular surgery after
percutanous transluminal coronary angioplasty. J Cardiothorac Vase Anesth 1998; 12:501-
506.

73. Hassan SA, Hlatky MA, Boothroyd DB, Winston C, Mark DB, Brooks MM, Eagle KA.
Outcomes of noncardiac surgery after coronary bypass surgery or coronary angioplasty in
the bypass angioplasty revascularization investigation (BARI). Am J Med 2001; 110:260-
266.

74. Mason JJ, Owens DK, Harris RA, Cooke JP, Hlatky MA. The role of coronary angiography
and coronary revascularization before noncardiac vascular surgery. JAMA 1995; 273:1919-
1925.

75. Massie MT, Rohrer MJ, Leppo JA, Cuter BS. Is coronary angiography necessary for vascular j>
surgery patients who have positive results of dipyridamole thallium scans? J Vase Surg 1997; <S
25:975-983. I

76. McFalls EO, Ward HB, Krupski WC, Goldman S, Littooy F, Eagle K, Nyman JA, Moritz Jf
T. McNabb S, Henderson WG. Prophylactic coronary artery revascularization for elective **
vascular surgery: Study design. Control Clin Trials 1999; 20:297-308. >9

77. Bui H, de Virgilio C. Preoperative evaluation and interventions before aortic surgery: Are they J
justified? Perspect Vase Surg 2000; 12:25-54. q

78. Romero L, de Virgilio C. Preoperative cardiac risk assessment: An updated approach. Arch |
Surg 2001; 136:1370-1376. S

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

44 CLOUSE and BREWSTER

79. Krupski WC, Nehler MR, Whitehill TA, Lawson RC, Strecker PK, Hiatt WR. Negative
impact of cardiac evaluation before vascular surgery. Vase Med 2000; 5:3-9.

80. Mangano DT, Layug EL, Wallace A, Tateo I. Effect of atenolol on mortality and cardio-
vascular morbidity after noncardiac surgery. N Engl J Med 1996; 335:1713-1720.

81. Wallace A, Layug B, Tateo I, Li J, Hollenberg M, Browner W, Miller D, Mangano DT.
Prophylactic atenolol reduces postoperative myocardial ischemia. Anesthesiology 1998; 88:
7-17.

82. Poldermans D, Boersma E, Bax JJ, Thomson IR, van de Ven LLM, Blankensteijn JD, Baars
HF, Yo T, Trocino G, Vigna C, Roelandt JRTC, van Urk H. The effect of bisoprolol on
perioperative mortality and myocardial infarction in high-risk patients undergoing vascular
surgery. N Engl J Med 1999; 341:1789-1794.

83. Poldermans D, Boersma E, Baxx JJ, Thomson IR, Paelinck B, van de Ven LLM, Scheffer
MG, Trocino G, Vigna C, Baars HF, van Urk H, Roelandt JRTC. Bisoprolol reduces cardiac
death and myocardial infarction in high-risk patients as long as 2 years after successful major
vascular surgery. Eur Heart J 2001; 22:1353-1358.

84. Dodds TM, Stone JG, Coromilas J, Weinberger M, Levy DG Prophylactic nitroglycerine
infusion during noncardiac surgery does not reduce perioperative ischemia. Anesth Analg
1993; 76:705-713.

85. Godet G, Coriat P, Baron JF, Bertrand M, Diquet B, Sebag C, Viars P. Prevention of
intraoperative myocardial ischemia during noncardiac surgery with intravenous diltiazem: A
randomized trial versus placebo. Anesthesiology 1987; 66:241-245.

86. Beattie WS, Badner NH, Choi P. Epidural analgesia reduces postoperative myocardial
infarction: A meta-analysis. Anesth Analg 2001; 93:853-858.

87. Whittemore AD, Clowes AW, Hechtman HB, Mannick JA. Aortic aneurysm repair: Reduced
operative mortality associated with maintenance of optimal cardiac performance. Ann Surg
1980; 192:414-421.

88. Berlauk JF, Abrams JH, Gilmour IJ, O'Connor SR, Knighton DR, Cerra FB. Preoperative
optimization of cardiovascular hemodynamics improves outcome in peripheral vascular
surgery. Ann Surg 1991; 214:290-299.

89. Bender JS, Smith-Meek MA, Jones CE. Routine pulmonary artery catheterization does not
reduce morbidity and mortality of elective vascular surgery. Ann Surg 1997; 226:229-237.

90. Isaacson IJ, Lowdon JD, Berry AJ, Smith RB III, Knos GB, Weitz FI, Ryan K. The value of
pulmonary artery and central venous monitoring in patients undergoing abdominal aortic
reconstructive surgery: A comparative study of two selected, randomized groups. J Vase Surg
1990; 12:754-760.

91. Valentine RJ, Duke ML, Inman MH, Grayburn PA, Hagino RT, Kakish HB, Clagett GP.
Effectiveness of pulmonary artery catheters in aortic surgery: A randomized trial. J Vase Surg
1998; 27:203-212.

92. Barone JE, Tucker JB, Rassias D, Corvo PR. Routine perioperative pulmonary artery
catheterization has no effect on rate of complications in vascular surgery: A meta-analysis.
Am Surg 2001; 67:674-679.

93. Gewertz BL, Kremser PC, Zarins CK, Smith JS, Ellis JE, Feinstein SB, Roizen MF.
Transesophageal echocardiographic monitoring of myocardial ischemia during vascular j>
surgery. J Vase Surg 1987; 5:607-613. |

94. Smith JS, Cahalan MK, Benefiel DJ, Byrd BF, Lurz FW, Shapiro WA, Roizen MF, Bouchard |
A, Schiller NB. Intraoperative detection of myocardial ischemia in high-risk patients: Elec- °|
trocardiography versus two-dimensional transesophageal echocardiography. Circulation 1985; **
72:1015-1021. |

95. Eisenberg MJ, London MJ, Leung JM, Browner WS, Hollenberg M, Tubau JF, Tateo IM, J
Schiller NB, Mangano DT. Monitoring for myocardial ischemia during noncardiac surgery. A q
technology assessment of transesophageal echocardiography and 12-lead electrocardiography. |

The Study of Perioperative Ischemia Research Group. JAMA 1992; 268:210-216. 2

©

!

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 45

96. Iafrati MD, Gordon G, Staples MH, Mackey WC, Belkin M, Diehl J, Schwartz S, Payne D,
OTJonnell TF. Transesophageal echocardiography for hemodynamic management of thoraco-
abdominal aneurysm repair. Am J Surg 1993; 166:179-185.

97. Suriani RJ, Neustein S, Shore-Lesserson L, Konstadt S. Intraoperative transesophageal
echocardiography during noncardiac surgery. J Cardiothorac Vase Anesth 1998; 12:274-280.

98. Schmidlin D, Bettex D, Bernard E, Germann R, Tornic M, Jenni R, Schmid ER. Transo-
esophageal echocardiography in cardiac and vascular surgery: Implications and observer var-
iablility. Br J Anesth 2001; 86:497-505.

99. Goldman L, Caldera DL, Nussbaum SR, Southwick FS, Drogstad D, Murray B, Burke DS,
O'Malley TA, Goroll AH, Caplan CH, Nolan J, Carabello B, Slater EE. Multifactorial index
of cardiac risk in noncardiac surgical procedures. N Engl J Med 1977; 297:845-850.

100. Goldman L. Cardiac risks and complications of noncardiac surgery. Ann Intern Med 1983;
98:504-513.

101. Detsky AS, Abrams HB, Forbath N, Scott JG, Hilliard JR. Cardiac assessment for patients
undergoing noncardiac surgery: A multifactorial risk index. Arch Intern Med 1986; 146:2131-
2134.

102. Detsky AS, Abrams HB, McLaughlin JR, Drucker DJ, Sasson Z, Johnston N, Scott JG,
Forbath N, Hilliard JR. Predicting cardiac complications in patients undergoing noncardiac
surgery. J Gen Intern Med 1986; 1:211-219.

103. Raymer K, Yang H. Patients with aortic stenosis: Cardiac complications in non-cardiac
surgery. Can J Anesth 1998; 45:855-859.

104. Torsher LC, Shub C, Rettke SR, Brown DL. Risk of patients with severe aortic stenosis
undergoing noncardiac surgery. Am J Cardiol 1998; 81:448-452.

105. Dajani AS, Taubert KA, Wilson W, Bolger AF, Bayer A, Ferried P, Gewitz MH, Shulman
ST, Nouri S, Newburger JW, Hutto C, Pallasch TJ, Gage TW, Levison ME, Peter G, Zuccaro
G Jr. Prevention of bacterial endocarditis: Recommendations by the American Heart Asso-
ciation. JAMA 1997; 277:1794-1801.

106. Cambria RP, Clouse WD, Davison JK, Dunn PF, Corey M, Dorer DJ. Thoracoabdominal
aneurysm repair: Results with 337 operations performed over a 15-year interval. Ann Surg
2002; 216:471^479.

107. Romero L, de Virgilio C, Donayre C, Stabile BE, Lewis RJ, Narahara K, Lippmann M,
White R, Chang C. Trends in cardiac morbidity and mortality after endoluminal abdominal
aortic aneurysm repair. Arch Surg 2001; 136:996-1000.

108. Hovsepian DM, Hein AN, Pilgram TK, Cohen DT, Kim HS, Sanchez LA, Rubin BG, Picus

D, Sicard GA. Endovascular abdominal aortic aneurysm repair in 144 patients: Correlation of
aneurysm size, proximal aortic neck length, and procedure-related complications. J Vase
Intervent Radiol 2001; 12:1373-1382.

109. Becker GJ, Kovacs M, Mathison MN, Katzen BT, Benenati JF, Zemel G, Powell A, Almeida
JI, Alvarez J Jr. Coello AA, Ingegno MD, Kanter SR, Katzman HE, Puente OA, Reiss IM,
Rua I, Gordon R, Baquero J. Risk stratification and outcomes of transluminal endografting
for abdominal aortic aneurysm: 7-year experience and long-term follow-up. J Vase Intervent
Radiol 2001; 12:1033-1046.

1 10. Pronovost P, Garrett E, Dorman T, Jenckes M, Webb TH III, Breslow M, Rosenfeld B, Bass 1

E. Variations in complication rates and opportunities for improvement in quality of care for <S
patients having abdominal aortic surgery. Langenbeck's Arch Surg 2001; 386:249-256. s

111. Axelrod DA, Henke PK, Wakefield TW, Stanley JC, Jacobs LA, Graham LM, Greenfield LJ, Jf
Upchurch GR. Impact of chronic obstructive pulmonary disease on elective and emergency **
abdominal aortic aneurysm repair. J Vase Surg 2001; 33:72-76. >9

1 12. Martin GH, O'Hara PJ, Hertzer NR, Mascha EJ, Krajewski LP, Beven EG, Clair DG, Ouriel J
K. Surgical repair of aneurysms involving the suprarenal, visceral, and lower thoracic aortic S
segments: Early results and late outcome. J Vase Surg 2000; 31:851-862. |

113. Pearce WH, Parker MA, Feinglass J, Ujiki M, Manheim LM. The importance of surgeon S

©

!

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

46 CLOUSE and BREWSTER

volume and training in outcomes for vascular surgical procedures. J Vase Surg 1999; 29:768-
778.

114. Sternbach Y, Illig KA, Zhang R, Shortell CK, Rhodes JM, Davies MG, Lyden SP, Green
RM. Hemodynamic benefits of regional anesthesia for carotid endarterectomy. J Vase Surg
2002; 35:333-339.

115. AbuRhama AF, Jennings TG, Wulu JT, Tarakji L, Robinson PA. Redo carotid endarterec-
tomy versus primary carotid endarterectomy. Stroke 2001; 32:2787-2792.

116. James DC, Hughes JD, Mills JL, Westerband A. The influence of gender on complications of
carotid endarterectomy. Am J Surg 2001; 182:654-657.

117. Hamdan AD, Pomposelli FB, Gibbons GW, Campbell DR, LoGerfo FW. Perioperative
strokes after 1001 consecutive carotid endarterectomy procedures without an electroencepha-
logram. Arch Surg 1999; 134:412-415.

118. Cao P, Giordano G, De Rango P, Zannetti S, Chiesa R, Coppi G, Palombo D, Spartera C,
Stancanelli V, Vecchiati E. A randomized study on eversion versus standard carotid en-
darterectomy: Study design and preliminary results: The EVEREST Trial. J Vase Surg 1998;
27:595-605.

1 19. Conte MS, Belkin M, Upchurch GR, Mannick JA, Whittemore AD, Donaldson MC. Impact
of increasing comorbidity on infrainguinal reconstruction: A 20-year perspective. Ann Surg
2001; 233:445-452.

120. Kalra M, Gloviczki P, Bower TC, Panneton JM, Harmsen WS, Jenkins GD, Stanson AW,
Toomey BJ, Canton LG Limb salvage after successful pedal bypass grafting is associated
with improved long-term survival. J Vase Surg 2001; 33:6-16.

121. Fanes PL, LoGerfo FW, Arora S, Hook S, Pulling MC, Akbari CM, Campbell DR,
Pomposelli FB. A comparative study of alternative conduits for lower extremity revascula-
rization: All-autogenous conduit versus prosthetic grafts. J Vase Surg 2000; 32: 1080-1090.

122. Nicoloff AD, Taylor LM, McLafferty RB, Moneta GL, Porter JM. Patient recovery after
infrainguinal bypass grafting for limb salvage. J Vase Surg 1998; 27:256-266.

123. Matsuura JH, Sobel M, Wong J, Dattilo JB, Poletti LF, Makhoul RG, Posner MP, Lee HM.
The limits of generalized cardiac screening tests for predicting cardiac complications after
infrainguinal arterial reconstruction. Ann Vase Surg 1997; 11:620-625.

124. Lawrence VA, Dhanda R, Hilsenbeck SG, Page CP. Risk of pulmonary complications after
elective abdominal surgery. Chest 1996; 110:744-750.

125. Kazmers A, Jacobs L, Perkins A. The impact of complications after vascular surgery in
Veterans Affairs medical centers. J Surg Res 1997; 67:62-66.

126. Kispert JF, Kazmers A, Roitman L. Preoperative spirometry predicts perioperative
pulmonary complications after major vascular surgery. Am Surg 1992; 8:491^495.

127. Rectenwald JE, Huber TS, Martin TD, Ozaki K, Devidas M, Wellborn MB, Seeger JM.
Functional outcome after thoracoabdominal aortic aneurysm repair. J Vase Surg 2002;
35:640-647.

128. Berry AJ, Smith RB III, Weintraub WS, Chaikof EL, Dodson TF, Lumsden AB, Salam AA,
Weiss V, Konigsberg S. Age versus comborbidities as risk factors for complications after
elective abdominal aortic reconstructive surgery. J Vase Surg 2001; 33:345-352.

129. Rowe S, Cheadle WG Complications of nosocomial pneumonia in the surgical patient. Am J g
Surg 2000; 179(suppl 2A):63S-68S. |

130. Croce MA. Diagnosis of acute respiratory distress syndrome and differentiation from venti- js
lator-associated pneumonia. Am J Surg 2000; 179(suppl 2A):26S-30S. Jf

131. Meade MO, Herridge MS. An evidence-based approach to acute respiratory distress syn- ^
drome. Respir Care 2001; 46:1368-1376. |

132. Smetana GW. Preoperative pulmonary evaluation. N Engl J Med 1999; 340:937-944. J

133. Mitchell CK, Smoger SH, Pfeifer MP, Vogel RL, Pandit MK, Donnelly PJ, Garrison RN, q
Rothschild MA. Multivariate analysis of factors associated with postoperative pulmonary |

complications following general elective surgery. Arch Surg 1998; 133:194-198. 2

©

!

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR SURGERY 47

134. Brooks-Brunn J A. Predictors of postoperative pulmonary complications following abdominal
surgery. Chest 1997; 111:564-571.

135. Bluman LG, Mosca L, Newman N, Simon DG. Preoperative smoking habits and post-
operative pulmonary complications. Chest 1998; 113:883-889.

136. Money SR, Rice K, Crockett D, Becker M, Abdoh A, Wisselink W, Kazmier F, Hollier LH.
Risk of respiratory failure after repair of thoracoabdominal aortic aneurysm. Am J Surg 1994;
168:152-155.

137. Svensson LG, Hess KR, Coselli JS, Safi HJ, Crawford ES. A prospective study of respiratory
failure after high-risk surgery on the thoracoabdominal aorta. J Vase Surg 1991; 14:271-282.

138. Engle J, Safi HJ, Miller CC III, Campbell MP, Harlin SA, Letsou GV, Lloyd KS, Root DB.
The impact of diaphragm management on prolonged ventilator support after thoraco-
abdominal aortic repair. J Vase Surg 1999; 29:150-156.

139. Calligaro KD, Azurin DJ, Dougherty MJ, Dandora R, Bajgier SM, Simper S, Savarese RP,
Raviola A, DeLaurentis DA. Pulmonary risk factors of elective abdominal aortic surgery.
J Vase Surg 1993; 18:914-921.

140. Arozullah AM, Daley J, Henderson WG, Khuri SF. Multifactorial risk index for predicting
postoperative respiratory failure in men after major noncardiac surgery. Ann Surg 2000; 232:
242-253.

141. Kroenke K, Lawrence VA, Theroux JF, Tuley MR, Hilsenbeck S. Postoperative compli-
cations after thoracic and major abdominal surgery in patients with and without obstructive
lung disease. Chest 1993; 104:1445-1451.

142. Meyers JR, Lembeck L, O'Kane H, Baue AE. Changes in functional residual capacity of the
lung after operation. Arch Surg 1975; 110:576-583.

143. Warner DO. Preventing postoperative pulmonary complications: The role of the anesthesi-
ologist. Anesthesiology 2000; 92:1467-1472.

144. Brooks-Brunn JA. Postoperative atelectasis and pneumonia. Heart Lung 1995; 24:94-115.

145. Cambria RP, Brewster DC, Abbott WM, Freehan M, Megerman J, LaMuraglia G, Wilson R,
Wilson D, Teplick R, Davison JK. Transperitoneal versus retroperitoneal approach for aortic
reconstruction: A randomized, prospective study. J Vase Surg 1990; 11:314-325.

146. Sicard GA. Surgical techniques for repair of abdominal aortic aneurysms. In: Gewertz BL
Schwartz LB, eds. Surgery of the Aorta and Its Branches. Philadelphia: Saunders 2000: 124-175.

147. Celli BR, Rodriguez KS, Snider GL. A controlled trial of intermittent positive pressure
breathing, incentive spirometry, and deep breathing exercises in preventing pulmonary
complications after abdominal surgery. Am Rev Respir Dis 1984; 130:12-15.

148. Chumillas S, Ponce JL, Delgado F, Viciano V, Mateu M. Prevention of postoperative
pulmonary complications through respiratory rehabilitation: A controlled clinical study. Arch
Phys Med Rehabil 1998; 79:5-9.

149. Thomas JA, Mcintosh JM. Are incentive spirometry, intermittent positive pressure breathing
and deep breathing exercises effective in the prevention of postoperative pulmonary compli-
cations after upper abdominal surgery? A systematic overview and meta-analysis. Phys Ther
1994; 74:3-16.

150. Castillo R, Haas A. Chest physical therapy: Comparative efficacy of preoperative and

postoperative in the elderly. Arch Phys Med Rehabil 1985; 66:376-379. 1

151. Stock MC, Downs JB, Gauer PK, Alster JM, Imrey PB. Prevention of postoperative <S
pulmonary complications with CPAP, incentive spirometry, and conservative therapy. Chest js
1985;87:151-157. Jf

152. Warner MA, Offord KP, Warner ME, Lennon RL, Conover MA, Jansson-Schumacher U. <
Role of preoperative cessation of smoking and other factors in postoperative pulmonary «
complications: A blinded, prospective study of coronary artery bypass patients. Mayo Clin J
Proc 1989; 64:609-616. a

153. Kabalin CS, Yarnold PR, Grammer LC. Low complication rate of corticosteroid-treated |
asthmatics undergoing surgical procedures. Arch Intern Med 1995; 155:1379-1384. S

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

48 CLOUSE and BREWSTER

154. Berg H, Roed J, Viby-Mogensen J, Mortensen CR, Engbaek J, Skovgaard LT, Krintel JJ.
Residual neuromuscular block is a risk factor for postoperative pulmonary complications: A
prospective, randomised, and blinded study of postoperative pulmonary complications after
atracurium, vecuronium, and pancuronium. Acta Anaesthesiol Scand 1997; 41:1095-1103.

155. Her C, Kizelshteyn G, Walker V, Hayes D, Lees DE. Combined epidural and general anes-
thesia for abdominal aortic surgery. J Cardiothorac Anesth 1990; 4:552-557.

156. Kehlet H, Holte K. Effect of postoperative analgesia on surgical outcome. Br J Anaesth 2001;
87:62-72.

157. Rodgers A, Walker N, Schug S, Kehlet H, van Zundert A, Sage D, Futter M, Saville G,
Clark T, MacMahon S. Reduction of postoperative mortality and morbidity with epidural or
spinal anaesthesia: Results from overview of randomised trials. Br Med J 2000; 321:1-12.

158. Ballantyne JC, Carr DB, de Ferranti S, Suarez T, Lau J, Chalmers TC, Angelillo A, Mosteller
F. The comparative effects of postoperative analgesic therapies on pulmonary outcome:
Cumulative meta-analysis of randomized, controlled trials. Anesth Analg 1998; 86:598-612.

159. Wakefield TW, Stanley JC. Cardiopulmonary assessment for major vascular reconstructive
procedures. In: Haimovici H, ed. Vascular Surgery. 4th ed. Cambridge, MA: Blackwell Science
1996:209-221.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

Renal Failure and Fluid Shifts Following
Vascular Surgery

Gregory S. Cherr

State University of New York at Buffalo, Buffalo, New York, U.S.A.

Kimberley J. Hansen

Wake Forest University School of Medicine, Winston- Salem, North Carolina, U.S.A.

The morbidity and mortality of acute renal failure (ARF) after vascular surgery was
recognized in the earliest series describing aortic reconstruction (1-3). With better under-
standing of perioperative fluid shifts, the incidence of renal dysfunction has subsequently
fallen, but ARF remains a frequent complication after vascular surgery (4-15). Because of
the limited ability to alter the course of established ARF and the high mortality associated
with ARF after vascular surgery, it is imperative that the vascular surgeon take measures
to prevent this complication.

Articles describing ARF use varying terminology and lack concise reporting standards.
Consequently, review of the topic is potentially confusing and direct comparisons between
studies are difficult (5,16). For the sake of clarity, in this review the term acute renal failure is
used to describe an abrupt rise in serum creatinine and/or blood urea nitrogen (BUN) with or
without oliguria (urine output <400 mL/day). Because patients with ARF do not necessarily
require renal replacement therapy, the distinction between ARF and renal replacement
therapy is noted. Following a brief description of normal renal physiology, the causes of
ARF are reviewed. From this reference point, the diagnosis, management, and potential -g

preventive strategies of ARF associated with vascular surgery are discussed. Because ARF &

occurs more commonly after aortic reconstruction and has been well described in this patient a

population, particular emphasis is placed on the management of these patients. c

%

I. THE INCIDENCE OF ACUTE RENAL FAILURE ASSOCIATED "J-

WITH VASCULAR SURGERY J

Because ARF is an unusual complication of infrainguinal or cerebrovascular reconstruc- |

tion, it has not been well described in the literature. Acute renal failure in these popu- @

49 I

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

50

CHERR and HANSEN

Table 1 The Incidence of Acute Renal Failure and Renal Replacement Therapy After Aortic
Surgery

Mortality

Mortality

Procedure

ARF (%)

with ARF (%)

RRT (%)

with RRT (%)

Infrarenal AAA (6)

CrCl>45:
CrCK45: 7.0

0.6

45.5

Infrarenal AAA (7)

13.9

1.8

Suprarenal AAA (7)

38.5

2.6

TAA (9)

16.0

11.9

41.0

Ruptured TAA (9)

22.2

TAA (10)

15.9

13.0

20.5 (comb)

TAA (11)

17.3

38.0

8.0

56.0

TAA (12)

15.0

2.5

49.0 (comb)

Ruptured AAA (13)

27.7

65.8

6.3

85.7

Ruptured AAA (14)

29.0

Ruptured AAA (15)

11.4

18.1

65.0 (comb)

Abbreviations: ARF, acute renal failure; RRT, renal replacement therapy; AAA, abdominal aortic
aneurysm; CrCl, creatinine clearance (mL/min); TAA, thoracoabdominal aneurysm; comb, com-
bined mortality for ARF and RRT.

lations is likely related to preexisting renal dysfunction and/or contrast nephrotoxicity.
However, ARF following aortic surgery is a relatively common complication associated
with high mortality (Table 1). Although it is reported to have an incidence ranging from 0-
13.9% after elective infrarenal aortic repair (4-8), the occurrence of ARF increases with
the addition of inciting factors such as urgent or emergency operation, proximal aortic
repair, preoperative renal dysfunction, adverse intraoperative and postoperative events,
and medical comorbidity (such as diabetes or coronary artery or liver disease) (7,9,13,17).
The incidence of ARF following repair of ruptured abdominal aortic aneurysm (AAA)
ranges from 20 to 29% (13,14,18) and has not changed appreciably in the past 25 years
(19-22). Moreover, the mortality of ARF requiring renal replacement therapy after repair
of intact or ruptured AAA is 58-86% and likely represents the mortality risk associated
with ARF as part of multisystem organ dysfunction (13,23-25).

II. NORMAL RENAL FUNCTION

While a complete discussion of normal renal physiology is beyond the scope of this
chapter, a basic understanding of intrarenal and excretory renal function is necessary to
understand abnormal renal function complicating the evaluation and management of
vascular disorders.

The kidney serves as the dominant site for maintenance of normal intravascular
volume and composition. Under normovolemic, unstressed conditions, the kidneys receive
approximately 25% of the cardiac output. Based on a cardiac output of 5 L/min, the
kidneys will receive approximately 900 L/day of plasma flow. Given the fact that the
glomeruli filter 20% of the renal plasma flow and that the normal 24-h urinary output for a
70-kg man is less than 1.8 L, the kidneys' tubular system must reabsorb more than 99% of
the 180 L/day of filtered plasma to maintain homeostasis. Moreover, the initial composi-
tion of the ultrafiltrate is the electrolyte and solute concentration of plasma. Therefore,
electrolytes and other solutes such as glucose must also be almost totally reabsorbed (26).

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i
2

I

■c

I

RENAL FAILURE AND FLUID SHIFTS 51

Reabsorption of electrolytes from the tubular fluid occurs both by active transport and
by passive back-diffusion. The sodium ion is reabsorbed in the early proximal tubule by its
cotransportation with organic solutes, bicarbonate, and divalent cations through an active
transport mechanism. Similarly, sodium is actively transported in the late proximal tubule in
combination with chloride transport. Since water freely follows this movement of solutes
and ions, the tubular fluid is iso-osmotic to plasma as it enters the loop of Henle (26).

Depending on their location, the tubular cells of the loop of Henle vary in their
permeability. This variable permeability establishes a hypotonic tubular fluid and medul-
lary osmotic gradient. Whereas the descending loop of Henle is permeable to water but
relatively impermeable to sodium and chloride, the ascending loop of Henle is impermeable
to water but actively transports the chloride ion, with sodium passively following. The
resulting countercurrent mechanism produces a medullary osmotic gradient that regulates
urine osmolarity from 50 to 1200 mOsm. Distal tubular reabsorption of sodium is also
active. In the distal tubule and the proximal collecting ducts, sodium is actively and almost
completely reabsorbed under the control of aldosterone. Of the approximately 25,000 mEq
of sodium filtered daily, only 50 to 200 mEq is ultimately excreted (less than 1%) (26).

Filtered potassium is almost totally reabsorbed in the proximal tubule and the loop of
Henle. Influenced by the electrochemical gradient and the intracellular concentration of
potassium, however, potassium is also passively secreted by the distal tubules and early
collecting ducts into the tubular lumen. Essentially all of the potassium in the urine is
transported there through this process (26,27).

A. Neuroendocrine Modulators of Renal Function

Intravascular volume is regulated primarily by a series of stretch or baroreceptors located
in the arterial tree and the atria. Since these receptors not only sense pressure or volume
changes (atrial receptors) but also monitor the rates of change during the cardiac cycle,
they govern the effective circulating volume. Factors that decrease cardiac performance
will alter the intravascular volume perceived by these receptors and thereby also alter the
renal function to retain water and increase the effective circulating volume. Similarly,
when the concentration of circulating plasma proteins is reduced, there is a net diffusion of
intravascular water into the extravascular space secondary to the decreased intravascular
oncotic pressure. This net decrease in circulating volume is sensed by these same receptors,
and neuroendocrine regulators of urinary output inhibit excretion of water to correct the
volume deficiency.

When the baroreceptors perceive a reduction in circulating volume, their afferent
signals are reduced, which decreases their tonic inhibition over the neuroendocrine system.
This leads to increased secretion of vasopressin, beta endorphins, growth hormone, and
adrenocorticotropic hormone through the central nervous system (CNS) and to an
increase in release of epinephrine from the adrenal medulla. Within the kidneys, at the
level of the nephron, baroreceptors within the macula densa cells of the juxtaglomerular

apparatus perceive a decrease in intravascular pressure or plasma ion concentration and c

stimulate juxtaglomerular cells to release renin. Renin, in turn, stimulates the production <j

of angiotensin I from angiotensinogen, which ultimately forms angiotensin II. Angiotensin >9

II functions to raise blood pressure by direct vasoconstriction and, through its stimulation 4j

of aldosterone, indirectly functions to increase circulating plasma volume. 2

The primary hormonal regulators of fluid and electrolyte balance are aldosterone, |

Cortisol, vasopressin, and angiotensin. However, the interactions between insulin, epi- @

nephrine, plasma glucose concentration, acid-base balance of the plasma, and other %

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

52 CHERR and HANSEN

factors play a vital role in modulating the release of these hormones and directly affect the
renal tubular management of water and the respective filtered solutes (26,27). Here,
discussion of these interactions and impact on renal function and fluid shifts is limited to
the effects of major vascular surgery.

III. FLUID SHIFTS ASSOCIATED WITH VASCULAR SURGERY

Because of the fluid shifts associated with vascular surgery, inappropriate fluid and elec-
trolyte administration after vascular reconstruction will place the patient at risk for ARF.
Causes of fluid shifts include tissue trauma from operative dissection, hemodynamic
response to arterial clamping, and operative blood loss. Due to the importance and pre-
dictability of these fluid shifts after vascular surgery, this physiology is briefly reviewed.

Movement of water and solutes from the intravascular to the extravascular, extra-
cellular space normally takes place at the precapillary level due to increased hydrostatic
pressure. The reentry of fluid into the intravascular space in the distal capillaries results
from the oncotic pressure gradient of intravascular proteins (predominantly albumin).
Operative dissection and disruption of lymphatic channels combined with inflammatory
mediators causing alterations in tissue perfusion result in increased capillary membrane
permeability to albumin (28). Reduced plasma albumin concentration leads to decreased
water reabsorption into the intravascular space. The resulting decreased intravascular vol-
ume causes activation of neuroendocrine mechanisms that decrease renal excretion of so-
dium and free water (29). In addition, ischemia-reperfusion and/or shock secondary to
blood loss causes changes in the cellular transmembrane potential with movement of sodium
and water into the intracellular space from the extracellular space (30). The normal response
to decreased intravascular volume is mobilization of extracellular fluid. Because of its
increased oncotic pressure after vascular surgery, the extracellular fluid is less available for
expansion of the intravascular space. Finally, anesthesia may cause decreased renal blood
flow through reductions in effective blood volume and reduced mean arterial pressure (16).
The net result of these mechanisms is the potential for renal hypoperfusion and ARF.

Because of the presence of fluid shifts and intravascular hypovolemia after vascular
surgery, determination of intravascular volume and serum electrolytes plays a vital role in
preventing dysfunction of the kidneys. Isotonic crystalloid is used for fluid resuscitation
during and after vascular surgery with blood transfusions reserved for significant reductions
of hemoglobin. However, routine albumin resuscitation in the postoperative period does
not appear beneficial after aortic reconstruction (31). Finally, time to mobilization of the
sequestered third-space fluid is variable, usually ranging from 2 to 5 days, depending on the
magnitude of perioperative stress, cardiac performance, and intravascular oncotic pressure.
If not managed with appropriate reduction in intravenous fluid administration (and occa-
sionally the addition of diuretic therapy), reabsorption of extracellular fluid can lead to
intravascular volume overload and acute congestive heart failure.

IV. CATEGORIES OF RENAL DYSFUNCTION

Potential causes of renal dysfunction are summarized in Table 2.

A. Prerenal Dysfunction

Prerenal causes are the most frequent source of ARF in the postoperative period. Renal
failure from a prerenal cause is usually the result of a contracted intravascular volume

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i
2

I

■c

I

u

RENAL FAILURE AND FLUID SHIFTS

53

Table 2 Potential Causes of Acute Renal Dysfunction in the Patient Undergoing Vascular
Reconstruction

Prerenal

Parenchymal

Postrenal

Low cardiac output/cardiogenic shock
Increased vascular space
Septic shock
Hypovolemia

Blood loss

Dehydration

Third-space sequestration

Nephrotoxic drugs
Acute tubular necrosis
Radiological contrast
Myoglobinuria
Other causes

Catheter kinking
Catheter clot
Bladder clot
Ureteral obstruction
Renal pelvic obstruction

resulting from inadequate replacement of intraoperative or postoperative fluid losses. Less
commonly, it is caused by a reduction in cardiac performance triggering neurohormonal
mechanisms that lead to increased reabsorption of sodium and water. In their pure forms,
these two causes of reduced renal function are clinically distinguishable. While hypovolemia
is associated with flat neck veins, dry mucous membranes, and reduced pulmonary artery
wedge and end-diastolic pressures, renal dysfunction from poor cardiac performance is
manifest by distended neck veins, clinical fluid overload, and elevated pulmonary artery
wedge pressure. Patients with hypovolemia are resuscitated with isotonic crystalloid (and
blood or blood products as needed), while those with ARF of cardiogenic origin require
improvement in myocardial performance (with afterload-reducing agents and/or inotropic
agents) and reduction of left ventricular preload (with diuretic and/or nitrate therapy).

Since the patient undergoing vascular surgery frequently has associated coronary
artery disease and impaired left ventricular systolic and diastolic function (3,33), dis-
tinction between hypovolemic and cardiogenic ARF can be difficult. Preexisting heart
disease may raise the baseline total body volume with associated higher central filling
pressures. In patients with diastolic dysfunction, normal or low-normal filling pressures
may actually reflect relative hypovolemia. In this clinical situation, the authors maintain a
constant infusion of both afterload-reducing and inotropic agents and cautiously admin-
ister small boluses of isotonic crystalloid while monitoring cardiac output, pulmonary
artery wedge pressure, and right ventricular end-diastolic volume. If urinary response is
negligible once filling pressures begin to rise, diuretic therapy may be required. The
frequent presence of left ventricular dysfunction in patients undergoing vascular recon-
struction requires that measures of cardiac function and filling pressures be established
before starting diuretic or inotropic therapy (32,33).

B. Postrenal Dysfunction

Postrenal dysfunction (obstruction of flow from the kidney) is an uncommon cause of
renal dysfunction after vascular surgery. The obstruction is usually at the level of the
urethra or urinary catheter and rarely at the ureters. Hematuria or traumatic catheter
insertion may lead to clot formation, catheter obstruction, and obstructive uropathy. For
this reason, an abrupt decline in urine output should prompt maneuvers — such as catheter
irrigation or replacement — to exclude mechanical causes. Urethral strictures, clots, or
other abnormalities during catheter insertion should be noted.

Ureteral or renal pelvic obstruction should be considered after other causes of
postrenal dysfunction have been excluded. Causes include kidney stones, iatrogenic injury,

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

54 CHERR and HANSEN

extrinsic compression of the ureter by a graft limb, or fibrotic reaction to surgery.
Preliminary diagnosis is suggested by renal ultrasound or renogram and confirmed by
retrograde urography. Delayed diagnosis and treatment reduces the chance of recovery of
renal function, so that prompt recognition of the pathology is paramount (34). Therapy
may require the placement of ureteral stents or percutaneous nephrostomy (16,35).

Acute urinary retention leading to obstructive uropathy may complicate the removal
of a bladder catheter. Patients at risk include those with prostatic hypertrophy or epidural
catheters for pain control. To avoid urinary retention, we generally allow 6-12 h time to
elapse after epidural analgesia is discontinued prior to removal of urinary catheters.

C. Renal Parenchymal Dysfunction

Parenchymal causes of ARF are diverse and pose the greatest risk for permanent damage
to the kidney. Potential sites of parenchymal injury include the renal tubule, vessels, or
glomerulus (16). Parenchymal dysfunction caused by injury of the renal tubules mediated
by ischemia or toxins is most relevant to the patient undergoing vascular surgery.
Common causes of renal ischemia after vascular surgery include suprarenal aortic cross-
clamping and systemic inflammatory response associated with multisystem organ dys-
function, hypovolemia, shock, or atheroembolism. Prerenal and parenchymal dysfunction
are linked when severe reductions in renal blood flow causes ischemic injury to the tubular
cells (16). Causes of toxic injury associated with vascular reconstruction include amino-
glycoside therapy, myoglobinuria, and radiological contrast. Other causes of parenchymal
dysfunction (such as acute interstitial nephritis and glomerulonephritis) are uncommon
after vascular surgery and are not further discussed in this chapter (16).

1. Acute Ischemic Injury

The pathophysiology of acute ischemic injury is likely twofold. First, as a consequence of
the magnitude and duration of ischemia, tubular cell swelling occurs following reperfu-
sion. This may cause tubular obstruction and reduction of glomerular filtration. Second,
ischemia may cause tubular cell necrosis, apoptosis, or loss of basement membrane
attachment (from interstitial edema during reperfusion), with sloughing of cells into the
tubule. The medullary thick ascending loop of Henle and the pars recta of the proximal
tubule are the segments of the tubular epithelium most sensitive to ischemia. Following
loss of the tubular cell, a back leak of glomerular filtrate into the renal parenchyma then
develops (36-38).

Aortic repair requiring a suprarenal cross clamp poses a significant risk for renal ische-
mia. The risk is greater for repair of thoracoabdominal aneurysm (TAA), where longer
periods of renal ischemia can be anticipated. Rates of ARF as high as 17% are reported in
larger series for elective repair of TAA (10-12). Predictors of renal dysfunction included
preoperative creatinine >1.5 mg/dL and cross-clamp time >100 min. These results were •§

similar to those from series in which partial left heart bypass and distal aortic perfusion were g

utilized (39). Recovery of renal function after suprarenal cross-clamping relates to preexist- a

ing renal dysfunction, patient age, and the duration of renal ischemia (cross-clamp time) c

(7,9,11,14,40). 5

Vascular surgery complicated by sepsis, myocardial dysfunction, or reperfusion injury >9

may cause renal ischemia leading to ARF. When it is part of the systemic inflammatory 4j

response syndrome, ARF may be mediated by increases in proinflammatory mediators Q

such as endotoxin, tumor necrosis factor, interleukin (IL)-l, IL-6, IL-8, prostaglandins, or |

leukotrienes (41). Recovery of renal function is dependent upon maintenance of adequate @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

RENAL FAILURE AND FLUID SHIFTS 55

renal perfusion and requires prompt elimination of the septic focus and improvement in
left ventricular performance. Hypotension from blood loss, myocardial dysfunction, or
sepsis can also diminish renal blood flow and incite acute ischemic injury (6,12,42).

Atheroembolism is an underrecognized cause of ARF. Embolization of atheromatous
debris from diseased segments of the pararenal aorta may complicate suprarenal cross-
clamping or dissection of the aorta. Likewise, catheter manipulation during aortography
may cause atheroembolization. These patients may also develop concomitant peripheral
ischemia from atheroembolization to the lower extremities. The presence of serum eosino-
philia and urine eosinophils may aid in the diagnosis (43). Acute renal failure from athero-
embolism is usually progressive and may be fatal.

2. Toxic Injury

Aminoglycosides appear to exert their renal toxicity at the tubular cell and cause impair-
ment of renal function in up to 20% of patients (44,45). Risk factors which can contribute
to nephrotoxicity with aminoglycoside use include high dosage or prolonged course of an
aminoglycoside, preexisting renal insufficiency, advanced age, extracellular volume deple-
tion with concomitant renal ischemia, or use of other nephrotoxins (44,45). Once-daily
dosing regimens for aminoglycosides are less costly and more convenient but do not
appear to reduce the risk of nephrotoxicity (46). A prudent approach would be to avoid
aminoglycoside therapy when possible.

Myoglobinuria is an important cause of ARF in patients undergoing vascular
reconstruction after a period of prolonged muscle ischemia. Myoglobin, released from
ischemic muscle, is freely filtered by the glomerulus and is thought to cause ARF through
direct tubular cell injury caused by myoglobin degradation products, precipitation within
the tubule, or abnormal renal blood flow (47-49). Myoglobinuria is suggested when the
urine is dipstick-positive for blood but no red cells are present on microscopic analysis.
Once diagnosed, myoglobinuria-induced injury to the kidney may be ameliorated by
alkalinizing the urine with sodium bicarbonate and maximizing urine output with
crystalloid infusion and mannitol administration (50).

Contrast agents used during angiography are a common cause of ARF. These agents
cause both direct tubular cell injury (51) and transient hypoperfusion of the kidneys (52).
The ionization and high osmolarity of contrast agents may contribute to their nephrotox-
icity. Conventional (ionic) contrast agents contain iodine, which absorbs x-ray photons
and thus allows visualization of the vasculature. Nonionic contrast agents provide
comparable absorption of x-ray photons yet are significantly less charged than traditional
agents. In comparing ionic and nonionic contrast agents, the rate of severe nephrotoxicity
is found to be similar for patients with normal renal function. Consequently, the lower
cost of ionic agents makes them preferable in this patient population (53). However, in
patients with preexisting renal dysfunction, nonionic contrast agents may reduce the risk ■g

of nephrotoxicity (54). In most patients with contrast-induced ARF, serum creatinine &

levels peak in 4-5 days and return to their baseline level within 2 weeks (55-57). a

Risk factors for contrast-induced nephrotoxicity as determined by metanalysis of c

prospective clinical trials include preexisting renal insufficiency, diabetes, heart failure, and <j

dose of contrast used (58). In patients with normal renal function, the incidence of contrast >9

nephropathy is <10%, but it increases with the addition of the previously mentioned risk 4j

factors (53,54,59). The risk of nephrotoxicity may increase exponentially for patients with 2

a serum creatinine >1.2 mg/dL (57). For diabetic patients with normal renal function [as |

indicated by a normal estimated glomerular filtration rate (EGFR)], the risk of contrast- @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

56 CHERR and HANSEN

induced nephrotoxicity is similar to that for nondiabetics (54,59,60). However, the
combination of diabetes and renal insufficiency increases the risk of contrast nephrotox-
icity to at least twice that expected for renal insufficiency alone (54,60,61). Among patients
commonly submitted to vascular surgery, diabetics with renal insufficiency appear to be at
greatest risk for ARF after angiography (62).

V. STRATEGIES TO PROTECT RENAL FUNCTION

Strategies to protect renal function are summarized in Table 3.

A. Preoperative Measures

Measures to protect renal function are begun in the preoperative period. The first goal is to
identify patients at increased risk for perioperative ARF. Excretory renal function should
be determined in every patient undergoing vascular surgery. The best overall estimate of
renal function is glomerular filtration rate. Direct measurement of glomerular filtration
rate, although preferred, is rarely performed in our clinical practice. Instead, EGFR is
determined using creatinine clearance calculated from measured serum creatinine (63).
Using EGFR, patients with preexisting renal dysfunction are identified and presumed to be
at increased risk for perioperative ARF. Elevated preoperative creatinine and associated
decreased EGFR are the most consistent predictors of ARF for patients submitted to aortic
reconstruction (4,7,9-15,40). Other risk factors for postoperative ARF include increased
age, poor cardiac function, diabetes, and renovascular disease (5,7,9).

Before vascular repair, the greatest risk of renal dysfunction occurs during angiog-
raphy. Strategies to minimize the effects of contrast-induced renal injury are controversial
and lack controlled studies to confirm their efficacy. Two easily modified risk factors for
contrast-induced nephrotoxicity are hypovolemia and volume of contrast agent used.
Intravenous crystalloid is started before angiography and urine output is used as a
measure of the adequacy of hydration. In patients with renal insufficiency undergoing
coronary angiography, hydration with 0.45% saline plus mannitol or furosemide has
demonstrated a greater risk of nephrotoxicity than 0.45% saline alone (61). Similarly,
renal dose dopamine (0.5-2.5 (ug/kg/min) does not appear to offer further advantages over

Table 3 Summary of Measures Taken to Prevent Acute Renal Dysfunction
During Vascular Reconstruction

Preoperative Intraoperative

Identify pts at risk (J, EGFR) Minimize renal ischemia time

Angiography Aortic cross-clamping
Preangiogram hydration Heparin

Minimize contrast use Mannitol

Acetylcysteine? Maintenance of euvolemia

Hold ACE inhibitors, loop diuretics, Bladder catheter

AT II receptor inhibitors Pulmonary artery catheter

Preoperative hydration Avoid atheroembolism

Abbreviations: pts, patients; EGFR, estimated glomerular filtration rate; ACE,
angiotensin-converting enzyme; AT II, angiotensin II.

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i
2

I

■c

I

u

RENAL FAILURE AND FLUID SHIFTS 57

crystalloid therapy (64,65). In our practice, patients at increased risk for nephrotoxicity
are routinely admitted for intravenous hydration prior to angiography (1.5 mL/kg/h for 12
h). Immediately before angiography, the patient usually receives a bolus of intravenous
fluid (3 to 5 mL/kg). Finally, intravenous hydration is continued for 4-6 h after
completion of the study.

Because no definitive limit on the "safe" amount of contrast agent exists, minimal
contrast doses should be used. Even small doses (30 mL) of contrast may induce dialysis-
dependent ARF in patients with extreme renal insufficiency (EGFR <15 mL/min). Con-
versely, 300 mL of contrast material has been safely administered to patients at low risk
for nephrotoxicity (66). The authors' practice is to limit the quantity of nonionized con-
trast agent to less than 50 mL in patients with renal insufficiency (EGFR <30 mL/min).
If additional contrast is required, further imaging is postponed to allow for recovery of
renal function.

Adjuncts or alternatives to conventional angiography are appropriate in many
instances. Digital subtraction techniques may be useful in limiting the quantity of contrast
material required. Carbon dioxide gas can be used for angiography with minimal renal
risk (67,68). Because it offers limited detail, CO2 angiography is often used to identify the
site of disease, which is then better defined with conventional contrast agents. Other
alternatives to conventional angiography that reduce or eliminate the risk of nephrotox-
icity include the use of gadolinium as a contrast agent for angiography (69), magnetic
resonance angiography (70), and abdominal ultrasound with visceral/renal artery duplex
sonography.

By scavenging reactive oxygen species, acetylcysteine may protect against contrast-
induced nephrotoxicity. Tepel and colleagues studied patients with chronic renal dysfunc-
tion who required nonionic contrast for computed tomography. They documented a sig-
nificant reduction in serum creatinine with the use of oral acetylcysteine and hydration
compared to placebo and hydration (71). Although further study is needed to better define
the role of acetylcysteine during arteriography, we administer two oral doses of acetyl-
cysteine (600 mg) before and after these studies in patients at high risk for contrast
nephrotoxicity.

Finally, high-dose loop diuretics, angiotensin-converting enzyme (ACE) inhibitors,
and angiotensin II receptor antagonists are held for at least 72 h prior to aortic recon-
struction. Selective beta blockers and calcium channel blockers (i.e., nifedipine) are sub-
stituted when necessary. With this approach, we hope to optimize renal perfusion through
avoidance of hypovolemia and renal vasoconstriction (72).

B. Operative Measures

Perioperative measures to protect renal function during vascular surgery are widely prac-
ticed by vascular surgeons. These include providing adequate circulating blood volume
through preoperative intravenous hydration, adequately replacing lost blood volume,

43
avoiding repetitive or prolonged renal ischemia, and optimizing myocardial performance. c

Additional modalities may include the use of dopamine or fenoldopam and mannitol to <j

enhance renal perfusion (73-76). It is our belief that these preventive measures may lessen >9

the severity and duration of ARF after vascular surgery (73,74). 4j

Particular attention to the patient's volume status is imperative. Patients are admitted 2

at least 12 h before operation for intravenous hydration. Without hydration, patients are |

frequently hypovolemic from the combined effects of bowel preparation and an overnight @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

58 CHERR and HANSEN

fast. Bladder catheters are placed to monitor urine output during and immediately after
the operation. For patients requiring aortic reconstruction, we routinely use pulmonary
artery catheters to help guide fluid resuscitation and monitor myocardial performance.
For patients with severe systolic and/or diastolic dysfunction, transesophageal echocar-
diography is used intraoperatively to guide fluid and vasoactive agent management,
especially during cross clamping of the aorta.

When possible, warm renal ischemia time should be limited by using efficient,
meticulous surgical technique. The normally perfused kidney can tolerate 45 min of warm
ischemia. For the chronically ischemic kidney, the duration of safe warm ischemic time
may be longer but ultimately depends on the amount and effectiveness of collateral blood
flow. Regional renal hypothermia may be a helpful adjunct for the protection of renal
function during periods of ischemia. Decreases in the core temperature of the kidney sig-
nificantly reduce metabolic needs and may reduce the incidence of postischemic ARF. In
animal models of renal ischemia, a modest decrease in core renal temperature (35°C) had a
protective effect on both renal tubular morphology and postprocedure serum creatinine
levels (77). Although we frequently use cold perfusion preservation techniques during
ex vivo renal artery repair, we do not routinely perfuse the in situ kidney with cold elec-
trolyte solution during aortic reconstruction. Rather, we use topical ice slush when warm
renal ischemia is likely to exceed 45 min during aortic repair.

A number of adjunctive measures may be employed at the time of aortic cross-clamp-
ing. We routinely administer heparin (1 mg/kg) and confirm systemic anticoagulation by
measurement of activated clotting time. We also use mannitol in small bolus doses (12.5-
25 g IV) during pararenal dissection and before and after periods of renal ischemia up to a
total dose of 1 g/kg of the patient is body weight. Potential protective actions of mannitol
include osmotic diuresis, decreased renovascular resistance that enhances cortical and
medullary blood flow, free-radical scavenging, and increased glomerular filtration during
renal hypoperfusion (50,78). When compared to saline administration before aortic cross-
clamping during infrarenal AAA repair, mannitol causes a reduction in subclinical glo-
merular and renal tubular damage (73).

Dopamine is frequently administered during vascular surgery. "Renal dose" dopa-
mine (0.5-3 (jig/kg/min) in healthy adults causes increased renal perfusion, glomerular
filtration rate, and urine output (79). However, the effect of low-dose dopamine in patients
undergoing vascular surgery is less well understood. During aortic reconstruction, dopa-
mine administration may cause increased urine output through its inotropic effects (74,80).
However, the clinical benefit of prophylactic dopamine administration in patients under-
going vascular surgery is unproved (74,81,82). Because dopamine may cause tachyarrhyth-
mias, myocardial ischemia, pulmonary shunting, or mesenteric vasoconstriction (81-83),
its routine use should be approached with caution.

Fenoldopam is a dopaminergic type 1 receptor agonist that may reduce the risk of ■g

ARF. Two dopamine receptors are found in the kidney: DAI and DA2. Activation of the &

DAI receptor causes increased glomerular filtration likely mediated by increased blood a

flow to the inner cortex and medulla of the kidney. Activation of the DA2 receptor causes c

a reduction in renal blood flow and glomerular filtration rate (84). Fenoldopam is a potent <j

antihypertensive agent used to treat patients with severe hypertension (85). Fenoldopam >9

significantly increases renal blood flow in healthy adults (76) and maintains kidney per- 4j

fusion in animal models of radiocontrast-induced nephrotoxicity (86) and aortic cross- 2

clamping (87). However, fenoldopam also causes significant reductions in cerebral blood |

flow in healthy adults (88). Although preliminary evidence indicates that fenoldopam may @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

RENAL FAILURE AND FLUID SHIFTS 59

prevent contrast-induced nephropathy (89) and renal dysfunction after aortic surgery (75),
additional investigation is needed to better elucidate the role of fenoldopam during
vascular surgery.

Specific intraoperative techniques may help to prevent microembolization of ath-
eromatous debris during juxtarenal aortic dissection and clamping. One should avoid
repetitive aortic cross-clamping, as this increases the risk of atheroembolization to the
renal arteries. Because the embolic potential of the debris cannot be judged definitively
until after the aorta is opened, one should assume its presence. When either suprarenal or
juxtarenal aortic control is required, we temporarily occlude renal artery flow immediately
before the application of the aortic clamp. Although we can provide only anecdotal
support for this maneuver, we believe that it has been an important adjunct in minimizing
the incidence of postoperative ARF due to atheroembolization among our patients.

Atheroembolism is the likely cause of ARF in patients without prolonged renal ische-
mia, excessive blood loss, hypotension, or other recognized nephrotoxic insult (90). The
quantity of emboli produced during manipulation of the aorta depends on the stability
and amount of atheromatous debris and the operative techniques employed. The clinical
impact of renal emboli depends on the amount of renal parenchyma affected and the
presence of concomitant renal insults. In patients without other risk factors for ARF or
preexisting renal dysfunction, large amounts of atheromatous emboli can occur without
immediate impact on renal function (91). In contrast, in patients with minimal renal
reserve, the added insult of minor emboli can lead to ARF. Because the prognosis for
recovery of renal function after atheroembolism is poor, prevention of this complication
is important.

Distal aortic perfusion may be used to maintain renal perfusion during repair of TAA.
This technique is most attractive during the repair of an isolated TAA (92) or when
complex disease precludes prompt completion of the proximal thoracic aortic anastamosis
(93). Distal aortic perfusion may be modified with "octopus" catheters to directly perfuse
the renal arteries during distal reconstruction (12). Since the routine use of distal aortic
perfusion has been associated with an increased incidence of ARF despite renal and spinal
cord protection in extensive (type II) TAA, we have preferred other strategies to provide
renal protection (12,94).

VI. DIAGNOSIS AND TREATMENT OF RENAL DYSFUNCTION

The diagnosis of ARF is based on an acute rise in serum creatinine and BUN with or
without a concomitant decrease in urine output. However, creatinine and BUN are rela-
tively insensitive markers of excretory renal function. The glomerular filtration rate may
fall by 50% before a rise in serum creatinine is noted due to the kidney's compensatory

increase in creatinine excretion (95,96). Conversely, creatinine and BUN may rise without ■g

an associated decrease in glomerular filtration. Causes of isolated increased creatinine &

without associated renal dysfunction include increased release from muscle and decreased a

secretion from the proximal tubules (e.g., trimethoprim or cephalosporin therapy). Causes c

of rising BUN without worsening renal function include increased protein intake, infusion <j

of amino acids, catabolism, gastrointestinal bleeding, and steroid therapy (97,98). Despite >9

these limitations, serial determinations of serum creatinine and BUN, calculation of 41

EGFR, and measurement of urine output are the mainstays of renal function monitoring. 2

An organized plan of diagnosis and treatment is important when faced with the patient |

with ARF after vascular surgery. The evaluation includes a thorough physical examination. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

60

CHERR and HANSEN

Table 4 Urinary and Blood Parameters that May Aid in the Evaluation of the Patient with Acute
Renal Dysfunction

Prerenal

Renal parenchymal

Postrenal

Characteristic

dysfunction

dysfunction

dysfunction

Urine specific gravity

> 1.020

< 1.020

< 1.020

Urine osmolarity (mOsm/L)

>400

<400

<400

Urine/plasma (U/P) osmolarity

>1.5

~1

~1

Urine Na (mEq/L)

<20

>30

<30 a

Fractional excretion of Na

<1%

>1%

<l% a

BUN:Cr

>20

<10

10-20"

U/PCr

>40

<20

<20

a First 24 h only.

Evidence of intravascular volume depletion, hemodynamic instability, sepsis or congestive
heart failure directs the differential diagnosis toward possible prerenal, renal, and postrenal
causes for renal dysfunction. Prerenal causes are the most frequent source of ARF in the
early postoperative period. The patient's intravascular volume status and cardiac perform-
ance should be evaluated. Patients with signs of volume depletion (flat neck veins, dry
mucous membranes, and reduced filling pressures) require fluid resuscitation with isotonic
crystalloid. Potassium-containing solutions and blood products are avoided until adequate
renal function is confirmed. Patients with signs of inadequate cardiac performance
(distended neck veins, S3 gallop, pulmonary edema, acute electrocardiographic changes,
dysrhythmias, decreased cardiac output, and elevated filling pressures) require judicious
inotropic support while indices of cardiac performance are measured. If correction of filling
pressures or myocardial performance fails to improve urinary output, samples of urine and
blood are obtained. Serum electrolytes, blood counts, and urine studies allow evaluation of
other possible sources of oliguria, such as myoglobinuria. Urine studies include urinalysis,
urine sodium, urea and creatinine concentrations, urine osmolality, and estimation of
fractional excretion of sodium. Interpretations of these blood and urinary parameters are
provided in Table 4.

VII. ESTABLISHED RENAL DYSFUNCTION AFTER
VASCULAR SURGERY

Acute renal failure following vascular surgery requires multidisciplinary therapy beyond the
scope of this review. Briefly, the initial goals are correction of extracellular volume deficits
and optimization of cardiac performance. With restoration of intravascular volume and
cardiac output, urinary output may be augmented with the addition of low-dose dopamine
(0.5-3 (|ig/kg/min) (99,100). The conversion of oliguric renal failure to a nonoliguric state
may delay the need for renal replacement therapy and simplify fluid management. It may
also be associated with fewer complications and improved survival (9,101,102), although
prospective data to support this notion are lacking.

Patients with radiocontrast-induced ARF require supportive care. Renal perfusion
should be optimized with fluid administration for patients with hypovolemia and diuresis
of those with fluid overload. If necessary, myocardial performance can be enhanced with

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

RENAL FAILURE AND FLUID SHIFTS 61

inotropic and/or vasodilatory agents. Further renal insults should be avoided with delay of
nonurgent operations or contrast studies. Finally, nephrotoxic drugs should be avoided.
Most patients' renal function will return to baseline within 2 weeks of the inciting angio-
gram (55-57).

As previously noted, the onset of renal failure requiring chronic replacement therapy
in the postoperative period has a grave prognosis. We prefer the use of continuous
venovenous hemofiltration/dialysis until the patient's hemodynamic status is stabilized.
Continuous hemodialysis reduces hemodynamic instability, allows better control of fluid
and metabolic status, and may affect outcome by removing deleterious cytokines
(103,104). Nutritional support of the patient with ARF is important, as protein-calorie
malnutrition is common in this population (105). The dosage of renally excreted med-
ications should be adjusted appropriately. Finally, it is important to maintain a frank and
realistic dialogue with the family of the patient requiring renal replacement therapy for
ARF after vascular repair.

REFERENCES

1. Powers SR, Boba A, Stein A. The mechanism and prevention of distal tubular necrosis
following aneurysmectomy. Surgery 1957; 42:156-162.

2. MacVaugh H, Roberts B. Results of resection of abdominal aortic aneurysm. Surg Gynecol
Obstet 1961; 113:17-23.

3. Payne JH, Wood DL, Goethel J A. Oliguria and renal failure in abdominal aortic surgery. Am
Surg 1963; 29:713-718.

4. Johnston KW. Multicenter prospective study of nonruptured abdominal aortic aneurysms:
Part II, variables predicting morbidity and mortality. J Vase Surg 1989; 9:437^147.

5. Novis BK, Roizen MF, Aronson S, Thisted RA. Association of preoperative risk factors with
postoperative acute renal failure. Anesth Analg 1994; 78:143-149.

6. Powell RJ, Roddy SP, Meier GH, Gusberg RJ, Conte MS, Sumpio BE. Effect of renal in-
sufficiency on outcome following infrarenal surgery. Am J Surg 1997; 174:126-130.

7. Breckwoldt WL, Mackey WC, Belkin M, O'Donnell TF. The effect of suprarenal cross-
clamping on abdominal aortic aneurysm repair. Arch Surg 1992; 127:520-524.

8. Berry AJ, Smith RB, Weintraub WS, Chaikof EL, Dodson TF, Lumsden AB. Age versus
comorbidities as risk factors for complications after elective aortic reconstructive surgery.
J Vase Surg 2001; 33:345-352.

9. Schepens MA, Defauw JJ, Hamerlijnck RP, Vermeulen FE. Risk assessment of acute renal
failure after thoracoabdominal aortic aneurysm surgery. Ann Surg 1994; 219:400-407.

10. Svensson LG, Crawford ES, Hess KR, Coselli JS, Safe HJ. Experience with 1509 patients
undergoing thoracoabdominal and aortic operations. J Vase Surg 1993; 17:357-368.

1 1 . Godet G, Fleron M-H, Vicaut E, Zubicki A. Risk factors for acute postoperative renal failure
in thoracic or thoracoabdominal aortic surgery: A prospective study. Anesth Analg 1997; 85:

1227-1232. '" '" |

12. San HJ, Harlin SA, Miller CC, Iliopoulos DC, Joshi A, Mohasci TG, Zippel R, Letsou GV, |
Tabor M. Predictive factors for acute renal failure in thoracic and thoracoabdominal aortic a
aneurysm surgery. J Vase Surg 1996; 24:338-344. |"

13. Panneton JM, Lassonde J, Laurendeau F. Ruptured abdominal aortic aneurysm: Impact of §
comorbidity and postoperative complications on outcome. Ann Vase Surg 1995; 9:535-541. =

14. Bauer EP, Redaelli C, von Segesser LK, Turina MI. Ruptured abdominal aortic aneurysms: jf
Predictors for early complications and death. Surgery 1993; 114:31-35. g

15. Hajarizadeh H, Rohrer MJ, Hermann JB, Cutler BS. Acute peritoneal dialysis following "g
ruptured abdominal aortic aneurysms. Am J Surg 1995; 170:223-226. g

©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

62 CHERR and HANSEN

16. Thadhani R, Pascual M, Bonventre JV. Acute renal failure. N Engl J Med 1996; 334:1448-
1460.

17. Nypaver TJ, Shepard AD, Reddy DJ, Elliot JP, Smith RF, Ernst CB. Repair of pararenal
abdominal aortic aneurysms. An analysis of operative management. Arch Surg 1993; 128:808-
813.

18. Gordon AC, Pryn S, Collin J. Outcome of patients who required renal support after surgery
for ruptured abdominal aortic aneurysm. Br J Surg 1994; 81:836-838.

19. Hicks GL, Eastland MW, DeWeese J A, Rob CG Survival improvement following aortic
aneurysm resection. Ann Surg 1975; 181:863-869.

20. McCombs PR, Roberts B. Acute renal failure following resection of abdominal aortic
aneurysm. Surg Gynecol Obstet 1979; 148:175-178.

21. Gornick CC, Kjellstrand CM. Acute renal failure complicating aortic surgery. Nephron 1983;
35:145-147.

22. Fielding JL, Black J, Ashton F, Slaney G. Ruptured aortic aneurysms: postoperative com-
plications and their aetiology. Br J Surg 1984; 71:487-491.

23. Braams R, Vossen V, Lisman BAM, Eikelboom BC. Outcome in patients requiring renal
replacement therapy after surgery for ruptured and nonruptured aneurysm of the abdominal
aorta. Eur J Vase Endovasc Surg 1999; 18:323-327.

24. Chen JC, Hildebrand HD, Salvian AJ, Taylor DC, Strandberg S, Myckatyn TM, Hsiang YN.
Predictors of death in nonruptured and ruptured abdominal aortic aneurysms. J Vase Surg
1996; 24:614-620.

25. Barratt J, Parajasingam R, Sayers RD, Feehally J. Outcome of acute renal failure following
surgical repair of ruptured abdominal aortic aneurysms. Eur J Vase Endovasc Surg 2000; 20:
163-168.

26. Robaczewski DL, Dean RH. Basic Science of Renovascular Hypertension. In: Sidawy AN,
Sumpio BE, Depalma RG, eds. The Basic Science of Vascular Disease. Armonk, NY: Futura,
1997 pp. 691-721.

27. Valtin H. Renal function: Mechanisms preserving fluid and solute balance in health. In: Valtin
H, ed. Renal Dysfunction: Mechanisms Involved in Fluid and Solute Imbalance, 3rd ed.
Boston, Little Brown, 1995: pp 110-129.

28. Weissman C. Ensuring perioperative fluid homeostasis in critically ill patients. J Crit Illness
1994; 9:1077-1084.

29. Schrier RW. A unifying hypothesis of body fluid volume regulation. J R Coll Phys Lond 1992;
26:295-306.

30. Smeets HJ, Kievit J, Dulfer FT, Hermans J, Moolenaar AJ. Analysis of post-operative hypo-
albuminemia: A clinical study. Int Surg 1994; 79:152-157.

31. Nielsen OM, Engell HC. Effects of maintaining normal plasma colloid osmotic pressure on
renal function and excretion of sodium and water after major surgery: A randomized study.
Dan Med Bull 1985; 31:182-185.

32. Hertzer NR, Beven EG, Young JR, O'Hara PJ, Rushchaupt WF, Graor RA, Dewolfe VG,
Maljovec LC. Coronary artery disease in peripheral vascular patients: A classification of 1000
coronary angiograms and results of surgical management. Ann Surg 1984; 199:223-233.

33. Blombery PA, Ferguson IA, Rosengarten DS, Stuchbery KE, Miles CR, Black AJ, Pitt A, 1
Anderson ST, Harper RW, Federman J. The role of coronary artery disease in complications g
of abdominal aortic aneurysm surgery. Surgery 1987; 101:150-155. js

34. Shapiro SR, Bennett AH. Recovery of renal function after prolonged unilateral ureteral ob- 2
struction. J Urol 1976; 115:136-140. ^

35. Naude G, Bongard F. Renal failure in the vascular patient. Semin Vase Surg 1996; 9:266-274. >9

36. Mason J, Joeris B, Welsch J, Kxiz W. Vascular congestion in ischemic renal failure: The role of jjj
cell swelling. Miner Electrolyte Metab 1989; 15:114-124. q

37. Lerman L, Textor SC. Pathophysiology of ischemic nephropathy. Urol Clin North Am 2001; |
28:793-803. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

RENAL FAILURE AND FLUID SHIFTS 63

38. Molitoris BA. New insights into the cell biology of ischemic acute renal failure. J Am Soc
Nephrol 1991; 1:1263-1270.

39. Kashyap VS, Cambria RP, Davison JF, L'ltalien GJ. Renal failure after thoracoabdominal
aortic surgery. J Vase Surg 1997; 26:949-955.

40. Crawford ES, Crawford JL, San HJ, Coselli JS, Hess KR, Brooks B, Norton HJ, Glaeser DH.
Thoracoabdominal aortic aneurysms: preoperative and intraoperative factors determining
immediate and long-term results of operations in 605 patients. J Vase Surg 1986; 3:389-404.

41. Breen D, Bihari D. Acute renal failure as a part of multiple organ failure: The slippery slope of
critical illness. Kidney Int 1998; 53:S25-S33.

42. Schepens MA, Defauw JJ, Hamerlijnck CRP, De Geest R, Vermeulen FE. Surgical treatment
of thoracoabdominal aortic aneurysms by simple cross-clamping: Risk factors and late results.
J Thorac Cardiovasc Surg 1994; 107:134-142.

43. Scolari F, Bracchi M, Valzorio B, Movilli E, Costantino E, Savoldi S, Zorat S, Bonardelli S,
Tardanico R, Maiorca R. Cholesterol atheromatous embolism: an increasingly recognized
cause of acute renal failure. Nephr Dial Trans 1996; 11:1607-1612.

44. Moore RD, Smith CR, Lipsky JJ, Mellits ED, Lietman PS. Risk factors for nephrotoxicity in
patients treated with aminoglycosides. Ann Intern Med 1984; 100:352-357.

45. Boucher BA, Coffey BC, Kuhl DA, Tolley EA, Fabian TC. Algorithm for assessing renal
dysfunction risk in critically ill trauma patients receiving aminoglycosides. Am J Surg 1990;
160:473-480.

46. Hatala R, Dinh T, Cook DJ. Once-daily aminoglycoside dosing in immunocompetent adults:
A meta-analysis. Ann Intern Med 1996; 124:717-725.

47. Zager RA. Heme protein-ischemic interactions at the vascular, intraluminal, and renal tubular
cell levels: Implications for therapy of myoglobin-induced renal injury. Renal Failure 1992;
14:341-344.

48. Zager RA. Rhabdomyolysis and myohemoglobinuric acute renal failure. Kidney Int 1996; 49:
314-326.

49. Vetterlein F, Hoffman F, Pedina J, Neckel M, Schmidt G. Disturbances in renal
microcirculation induced by myoglobin and hemorrhagic hypotension in anesthetized rats.
Am J Physiol 1995; 268:F839-F846.

50. Better OS, Stein JH. Early management of shock and prophylaxis of acute renal failure in
traumatic rhabdomyolysis. N Engl J Med 1990; 322:825-829.

51. Donadio C, Tramonti G, Lucceshi A, Giordani R, Lucchetti A, Bianchi C. Tubular toxicity is
the main renal effect of contrast media. Renal Failure 1996; 18:647-656.

52. Larson TS, Hudson K, Mertz JI, Romero JC, Knox FG Renal vasoconstrictive responses to
contrast medium. J Lab Clin Med 1983; 101:385-391.

53. Schwab SJ, Hlatky MA, Pieper KS, Davidson CJ, Morris KG, Skelton TN, Bashore TM.
Contrast nephrotoxicity: a randomized controlled trial of a nonionic and an ionic
radiographic contrast agent. N Engl J Med 1989; 320:149-153.

54. Rudnick MR, Goldfarb S, Wexler L, Ludbrook PA, Murphy MJ, Halpern EF, Hill JA,
Winniford M, Cohen MB, VanFossen DB. Nephrotoxicity of ionic and nonionic contrast
media in 1 196 patients: A randomized trial. The Iohexol Cooperative Study. Kidney Int 1995;

47:254-261. 1

55. Berns AS. Nephrotoxicity of contrast media. Kidney Int 1989; 36:730-740. |

56. Porter GA. Contrast-associated nephropathy. Am J Cardiol 1989; 64:22E-26E. js

57. Wish JB, Moritz CE. Preventing radiocontrast-induced acute renal failure. J Crit Illness 1990; £
5:16-25. ^

58. Barrett BJ. Contrast nephrotoxicity. J Am Soc Nephrol 1994; 5:125-137. 4

59. Davidson CJ, Hlatky M, Morris KG, Pieper K, Skelton TN, Schwab SJ, Bashore TM. Car- J
diovascular and renal toxicity of a non-ionic radiographic contrast agent after cardiac cath- q
eterization: A prospective trial. Ann Intern Med 1989; 110:119-124. |

60. Parfrey PS, Griffiths SM, Barrett BJ, Paul MD, Genge M, Withes J, Farid N, McManamon 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

64 CHERR and HANSEN

PJ. Contrast material-induced renal failure in patients with diabetes mellitus, renal insuf-
ficiency, or both. A prospective controlled study. N Engl J Med 1989; 320:143-149.

61. Solomon R, Werner C, Mann D, D'Elia J, Silva P. Effects of saline, mannitol and furosemide
to prevent acute decreases in renal function induced by radiocontrast agents. N Engl J Med
1994; 331:1416-1420.

62. Manske CL, Sprafka JM, Strony FT, Wang Y. Contrast nephrography in azotemic diabetic
patients undergoing coronary angiography. Am J Med 1990; 89:615-620.

63. Rolin HA, Hall PM, Wei R. Inaccuracy of estimated creatinine clearance for predictors of
iothalamate glomerular filtration rate. Am J Kidney Dis 1984; 4:48-54.

64. Hans B, Hans SS, Mittal VK, Khan TA, Patel N, Dahn MS. Renal function to dopamine during
and after arteriography in patients with chronic renal failure. Radiology 1990; 176:651-654.

65. Hall KA, Wong RW, Hunter GC, Camazine BM, Rappaport WA, Smyth SH, Bull DA,
Mclntyre KE, Bernhard VM, Misiorowski RL. Contrast-induced nephrotoxicity: The effects
of vasodilator therapy. J Surg Res 1992; 53:317-320.

66. Cigarroa RG, Lang RA, Williams RH, Hillis LD. Dosing of contrast material to prevent
contrast nephropathy in patients with renal disease. Am J Med 1989; 86:649-652.

67. Weaver FA, Pentecost MJ, Yellin AE. Carbon dioxide digital subtraction arteriography: A
pilot study. Ann Vase Surg 1990; 4:437-441.

68. Seeger JM, Self S, Harward TR, Flynn TC, Hawkins IF. Carbon dioxide gas as an arterial
contrast agent. Ann Surg 1993; 217:688-697.

69. Hammer FD, Goffette PP, Malaise J, Mathurin P. Gadolinium dimeglumine: An alternative
contrast agent for digital subtraction angiography. Eur Radiol 1999; 9:128-136.

70. Goyen M, Ruehm SG, Debatin JF. MR-angiography: The role of contrast agents. Eur J
Radiol 2000; 34:247-256.

71. Tepel M, van der Giet M, Schwarzfeld C, Laufer U, Liermann D, Zidek W. Prevention of
radiographic-contrast-agent-induced reductions in renal function by acetylcysteine. N Engl J
Med 2000; 343:180-184.

72. Bonventure JV. Mechanisms of ischemic acute renal failure. Kidney Int 1993; 43:1160-1178.

73. Nicholson ML, Baker DM, Hopkinson BR, Wenham PW. Randomized control trial of the
effect of mannitol on renal reperfusion injury during aortic aneurysm surgery. Br J Surg 1996;
83:1230-1233.

74. De Lasson L, Hansen HE, Juhl B, Paaske WP, Pedersen EB. A randomised, clinical study of
the effect of low-dose dopamine on central and renal haemodynamics in infrarenal aortic
surgery. Eur J Vase Endovasc Surg 1995; 10:82-90.

75. Gilbert TB, Hasnain JU, Flinn WR, Lilly MP, Benjamin ME. Fenoldopam infusion
associated with preserving renal function after aortic cross-clamping for aneurysm repair. J
Cardiovasc Pharmacol Ther 2001; 6:31-36.

76. Mathur VS, Swan SK, Lambrecht LJ, Anjum S, Fellmann J, McGuire D, Epstein M, Luther
RR. The effects of fenoldopam, a selective dopamine receptor agonist, on systemic and renal
hemodynamics in normotensive subjects. Crit Care Med 1999; 27:1832-1837.

77. Pelkey TJ, Frank RS, Stanley JJ, Frank TS, Zelenock GB, D'Alecy LG. Minimal physiologic
temperature variations during renal ischemia alter functional and morphologic outcome. J

Vase Surg 1992; 15:619-625. 1

78. Abbott WM, Austen WG The reversal of renal cortical ischemia during aortic occlusion by <S
mannitol. J Surg Res 1974; 16:482-489. ,1

79. Denton MD, Chertow M, Brady HR. "Renal-dose" dopamine for the treatment of acute renal 2
failure: Scientific rationale, experimental studies and clinical trials. Kidney Int 1996; 50:4-14. ^

80. Girbes AR, Lieverse AG, Smit AJ, van Veldhuisen KJ, Zwaveling JH, Miejer S, Rietsma WD. -9
Lack of specific renal haemodynamic effects of different doses of dopamine after infrarenal jjj
aortic surgery. Br J Anaesth 1996; 77:753-757. q

8 1 . Baldwin L, Henderson A, Hickman P. Effect of postoperative low-dose dopamine on renal |
function after elective major vascular surgery. Ann Intern Med 1994; 120:744-747. 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

RENAL FAILURE AND FLUID SHIFTS 65

82. Kellum JA. The use of diuretics and dopamine in acute renal failure: A systemic review of the
evidence. Crit Care 1997; 1:53-59.

83. Thompson BT, Cockrill BA. Renal-dose dopamine: a siren song? Lancet 1994; 344:7-8.

84. Kebabian JW, Calne DB. Multiple receptors for dopamine. Nature Lond 1979; 277:93-96.

85. Murphy MB, Murray C, Shorten GD. Fenoldopam: a selective peripheral dopamine-receptor
agonist for the treatment of severe hypertension. N Engl J Med 2001; 345:1548-1557.

86. Bakris GL, Lass NA, Glock D. Renal hemodynamics in radiocontrast medium-induced renal
dysfunction: A role for dopamine-1 receptors. Kidney Int 1999; 56:206-210.

87. Halpenny M, Markos F, Snow HM, Duggan PF, Gaffney E, O'Connell DP, Shorten GD. The
effects of fenoldopam on renal blood blow and tubular function during aortic cross-clamping
in anaesthetized dogs. Eur J Anaesth 2000; 17:491-498.

88. Prielipp RC, Wall MH, Groban L, Tobin JR, Fahey FH, Harkness BA, Stump DA, James
RL, Cannon MA, Bennett J, Butterworth J. Reduced regional and global cerebral blood flow
during fenoldopam-induced hypotension in volunteers. Anesth Analg 2001; 93:45-52.

89. Madyoon H, Croushore L, Weaver D, Mathur V. Use of fenoldopam to prevent radiocontrast
nephropathy in high-risk patients. Catheter Cardiovasc Intervent 2001; 53:341-345.

90. Iliopoulos JI, Zdon MJ, Crawford BG, Pierce GE, Thomas JH, Hermreck AS. Renal micro-
embolization syndrome. A cause for renal dysfunction after abdominal aortic reconstruction.
Am J Surg 1983; 146:779-783.

91. Smith MC, Ghose MK, Henry AR. The clinical spectrum of renal cholesterol embolization.
Am J Med 1981; 71:174-180.

92. Von Oppell U, Dunne T, DeGroot K, Zilla P. Spinal cord protection in the absence of collateral
circulation: meta-analysis of mortality and paraplegia. J Card Surg 1994; 9:685-691.

93. Coselli JS. Thoracoabdominal aortic aneurysms: Experience with 372 patients. J Card Surg
1994; 9:638-647.

94. Coselli JS, LeMaire SA, Miller CC, Schmittling ZC, Koksoy C, Pagan J, Curling PE.
Mortality and paraplegia after thoracoabdominal aortic aneurysm repair: A risk factor
analysis. Ann Thorac Surg 2000; 69:409^14.

95. Doolan PD, Alpen EL, Theil GB. A clinical appraisal of the plasma concentration and endog-
enous clearance of creatinine. Am J Med 1962; 32:65-69.

96. Bennett WM, Porter GA. Endogenous creatinine clearance as a clinical measure of glomerular
filtration rate. BMJ 1971; 4:84-86.

97. Berglund F, Killander J, Pompeius R. Effect of trimethoprim-sulfamethoxazole on the renal
excretion of creatinine in man. J Urol 1975; 114:802-808.

98. Anderson RA. Prevention and management of acute renal failure. Hosp Pract 1993; 28:61-72.

99. Szerlip HM. Renal-dose dopamine: Fact and fiction. Ann Intern Med 1991; 115:153-154.

100. Flancbaum L, Choban PS, Dasta JF. Quantitative effects of low-dose dopamine on urine
output in oliguric surgical intensive care unit patients. Crit Care Med 1994; 22:61-68.

101. Corwin HL, Bonventre JV. Factors influencing survival in acute renal failure. Semin Dial
1989; 2:220-231.

102. Lieberthal W, Levinsky NG Treatment of acute tubular necrosis. Semin Nephrol 1990; 10:571-
583.

103. Mehta RL. Therapeutic alternatives to renal replacement for critically ill patients in acute j>
renal failure. Semin Nephrol 1994; 14:64-82. |

104. Druml W. Metabolic aspects of continuous renal replacement therapies. Kidney Int 1999; s
72:S56-S61. Jf

105. Ikizler TA. Himmelfarb. Nutrition in acute renal failure patients. Adv Renal Repl Therap ^
1997; 4:54-63. £

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

Intimal Hyperplasia: The Mechanisms

and Treatment of the Response to

Arterial Injury

Michael J. Englesbe and Alexander W. Clowes

University of Washington, Seattle, Washington, U.S.A.

All forms of arterial reconstruction involve vessel wall injury and repair of the injury by
cells from adjacent normal tissues repair and possibly by circulating precursor cells.
Intimal hyperplasia is the hallmark of this healing process, often resulting in significant
luminal narrowing that predisposes the repair to failure. Since intimal hyperplasia affects
15-30% of all arterial interventions, strategies to control this injury response would have
significant clinical impact (1).

This chapter reviews the biology of vessel wall healing following vascular reconstruc-
tion and the molecular mechanisms involved in smooth muscle cell (SMC) mitogenesis,
migration, and extracellular matrix formation, three critical processes in intimal hyper-
plasia formation. Recent advances in the understanding of arterial injury response have
fostered new approaches to inhibit intimal hyperplasia. (Table 1) These novel clinical
approaches are discussed.

I. AUTOLOGOUS VEIN GRAFT HEALING

Vein grafts are the preferred conduits for most coronary and lower limb arterial bypasses.
Nonetheless, these arterial reconstructions fail because of intimal thickening, luminal
narrowing, and thrombosis. Graft failure is associated with 37-44% of aortocoronary

venous bypass grafts in the 3-10 years after surgery (2,3). Without careful surveillance, -c

patency is as low as 50% 2 years following infrainguinal bypass using saphenous vein as <j

the conduit (4). -9

The causes of graft failure are categorized into three groups. In the immediate 4j

postoperative period, graft occlusion occurs because of technical issues, such as twisting 2

or compression of the graft, poor distal arterial outflow, graft size mismatch, or flawed |

surgical technique. The treatment of early graft failure involves prompt revision of the ©

67 I

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

68

ENGLESBE and CLOWES

Table 1 Four General Categories of Therapeutic Approaches for Inhibiting Stenosis or Restenosis
from Intimal Hyperplasia

Mechanical
therapies

Cell-cycle
inhibitors

Growth factor
inhibitors

Antithrombotics
or anticoagulants

Balloon dilatation of
artery or stent

Stent placement in
artery or within
existing stent

Atherectomy

Brachytheraphy
Rapamycin (sirolimus)

(also inhibits migration)
E2F transcription factor

decoy oligonucleotides
Taxol (paclitaxel) (also

inhibits migration)
Gene therapy-

overexpression of RB,

p21, gax, dominant

negative mutants of ras,

NOS

Antibodies and inhibitors

of PDGFR a and p
Antibody to EGFR
Antibody to FGFR
Antibody to TGF-p
Recombinant TGF-

beta RII
Antibody to IGFR

Aspirin

Ticlopidine

Heparins

graft. From 3 months to 2 years, the graft wall thickens, or becomes "arterialized," along
its entire length. In addition, focal areas of intimal thickening develop, occurring at
anastomoses, valves, or vascular clamp sites. The resultant luminal stenoses reduce flow
and predispose the graft to thrombosis (5-7). There is currently no good treatment for
intimal hyperplastic lesions besides close graft surveillance and prompt surgical revision.
The third category of graft failure occurs after 2 years because of atherosclerosis in either
the graft or in the proximal or distal artery. Treatment entails aggressive risk-factor
intervention, including use of lipid-lowering agents.

An understanding of the biology of vein graft healing is fundamental to developing
new therapeutic approaches aimed at the prevention and regression of intimal hyperplasia.
Venous endothelial and smooth muscle cells (SMC) are injured at the time of graft harvest.
Techniques such as the method of vein harvesting, application of clamps, valvulotomy,
ischemia, and distention of the graft variably affect the degree of injury (8,9). In addition,
placement of a vein into the arterial system elicits an injury response in part due to
hemodynamic stress on the vein graft wall as well as to the surgery itself (10). The medial
hyperplastic response to increased wall stress is a reversible phenomenon. When flow
within the graft is returned to venous levels, the response is reversed (11).

Because the time course of intimal thickening in human venous grafts is difficult to
study, we must rely on animal models for a detailed description, even though these models
more often provide insight into vein wall adaptation rather than pathological intimal
thickening and luminal stenosis. In a rabbit model, when the external jugular vein is
transposed into the carotid artery, platelets, microthrombi, and leukocytes adhere to areas
of endothelial denudation (12). By 2 weeks, the areas of endothelial denudation have re-
generated. Over the first 4 weeks, there is substantial intimal SMC proliferation and mi-
gration and intimal thickening begins to develop. As proliferation and migration stop,
there is accumulation of extracellular matrix. In this rabbit model, graft wall thickness,
circumference, and cross-sectional area reach a maximum at 12 weeks (13). Following this
acute injury response, the graft continues to respond to environmental stimuli. The venous
graft is a dynamic entity that adapts to changes in hemodynamic forces. The chronic re-

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i
2

I

■c

I

u

INTIMAL HYPERPLASIA 69

sponse to arterialization likely mimics the arterial hypertrophic response to hypertension.
Also, shear stress likely affects venous graft diameter and wall thickness.

II. PROSTHETIC GRAFT HEALING

Prosthetic grafts are preferred for large vessel replacement and when autologous conduit is
inadequate. They function well when placed in vascular beds of high flow, such as in the
aortobifemoral position, but are prone to failure when used in beds of relatively low flow,
such as femoral-infrapopliteal bypasses (14). There are many differences between pros-
thetic vascular conduits and venous grafts. Prosthetic grafts are more prone to sponta-
neous thrombosis and infection. In animals, depending on the porosity of the vascular
graft, variable amounts of endothelium line the luminal aspect of the graft. Most clinically
relevant PTFE vascular grafts have low porosity (30-um pores or smaller). In such grafts,
the endothelial and SMC coverage of the luminal surface of the graft is limited to the first
few centimeters of each anastomosis. It is not known why these SMCs and endothelial
cells populate only the graft adjacent to the anastomosis. The central portion of the graft is
covered by a pseudointima containing fibrin, platelets, and leukocytes but no endothelial
cells (15). A detailed time course of graft healing in humans is not known because the
tissue is inaccessible. Nonetheless, intimal thickness is likely maximal within the peri-
anastomotic native artery and graft at approximately 12 weeks.

In contrast, polytetrafluoroethylene (PTFE) grafts with 60-um pores demonstrate a
markedly different process of graft healing. When these grafts are used in aortoiliac
bypasses in baboons, a fully developed, endothelialized intima develops by 2 weeks along
the entire length of the graft and continues to increase in size until 8-12 weeks (16). The
SMCs and endothelial cells presumably migrate in from perigraft granulation tissue.

III. ARTERIAL HEALING FOLLOWING BALLOON ANGIOPLASTY
AND STENTING

Percutaneous intravascular interventions have significant promise for patients with either
coronary artery disease or peripheral vascular disease. Important limitations of balloon
angioplasty are restenosis from intimal hyperplasia and pathological remodeling.

Inward or pathological remodeling is the major factor in restenosis after atherectomy
and angioplasty (17,18). Animal studies suggest that adventitial cicatrization is important
in inward remodeling (19,20). Time-course studies using intravascular ultrasound (IVUS)
indicate that inward remodeling occurs 1-6 months following a procedure (17). Late
regression of stenosis (6 months to 5 years) after angioplasty demonstrates that the ability
to outwardly remodel is restored, likely from the reestablishment of an intact and
functional endothelium (21). Intraluminal stent placement prevents pathological inward "g

or outward remodeling, and luminal narrowing depends solely upon the amount of in- &

stent intimal hyperplasia. a

The most extensively studied model for arterial healing and intimal formation after c

angioplasty utilizes the rat carotid balloon injury model (22) (Fig. 1). The balloon strips <j

away the endothelial cells, stretches the underlying media, and disrupts elastin layers. >9

Endothelial denudation alone is enough to significantly induce an intimal hyperplastic 41

response, but the response is augmented by balloon stretch of the vessel wall (22). Imme- 2

diately following injury, the exposed media is coated with platelets, which then degranulate. |

SMCs begin proliferating approximately 24 h after the injury. These smooth muscle cells @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

70 ENGLESBE and CLOWES

2 weeks after injury 4 weeks after injury

•;-..■ v-V

Intima IEL Intima ^ *' '.., .<■:'■!

Figure 1 Balloon-injured baboon saphenous artery. (H&E staining, x 100.) Demonstrates time
course of intimal growth from smooth muscle cell migration and proliferation. (IEL = internal elas-
tic lumina.)

increase from a basal level of proliferation of 0.06% per day to 10-30% per day (23). After 4
days, SMCs migrate into the intima, continue to proliferate, and secrete substantial amounts
of extracellular matrix. A steady state is reached at 12 weeks, with an intima composed of
20% cells and 80% extracellular matrix (24). Regrowth of the endothelium likely down
regulates intimal SMC proliferation and migration. If the zone of injury is larger than 3 cm,
endothelial repopulation of the artery never reaches the central segment of injured artery,
and smooth muscle cells at the luminal surface continue to turnover (25).

Intraluminal stent placement prevents arterial recoil and pathological remodeling, re-
ducing early (less than 6 weeks) restenosis. However, stent placement is associated with an
even more vigorous intimal hyperplastic response (26), and the overall rate of late (greater
than 6 weeks) stent restenosis remains approximately 15-30% in stented coronary arteries
(27). With stent placement, cytokines and growth factors induce multiple signaling path-
ways associated with SMC migration and proliferation (28,29). Coronary stent penetration
into the lipid core induces increased arterial inflammation, associated with increased
neointimal growth (A. Farb, R. Virmani, personal communication). Thus, the arterial wall
reacts not only to the underlying atherosclerotic process with its inflammatory components
and the acute wall injury but also to the chronic inflammation induced by the foreign body.

IV. MOLECULAR MECHANISMS OF INTIMAL HYPERPLASIA

The healing of vein grafts, prosthetic grafts, balloon-dilated and stented vessels, or
endarterectomized arteries is similar. There are three primary observations that apply
to all forms of arterial reconstruction: (a) medial SMC proliferation is in proportion to
the severity of injury, (b) all forms of injury cause endothelial denudation and accumu-
lations of platelets in the subendothelium, and (c) serum derived from degranulated
platelets and clotted blood induces SMC proliferation and migration. The best-studied
model of intimal hyperplasia is the rat carotid artery response to balloon injury. An in-

43
depth understanding of this relatively simple system of arterial injury is a critical step in -c

developing pharmacological approaches to inhibiting the deleterious effect of intimal <j

hyperplasia in patients. >3

In the rat, balloon injury leads to complete destruction of the endothelium as well as 41

death to medial SMCs (25). The SMC response to injury and the molecules involved in the 2

response have been categorized into four waves (25,30-32). Within the first 24 h, the first |

wave of injury response is initiated, which entails basic fibroblast growth factor (bFGF)- @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

INTIMAL HYPERPLASIA 71

and epidermal growth factor receptor (EGFR)- dependent medial SMC proliferation (33)
(A. Chan, unpublished results). Though this first wave of SMC proliferation has no
obvious relation to later SMC proliferation and migration, there is some evidence that
inhibition of this initial proliferation will lead to a diminution of the eventual size of the
final intimal hyperplastic lesion (34,35). The second wave involves the migration of SMCs
through the internal elastic lamina into the intima, first noted at 4 days after balloon injury
(25). The most important molecule in this response is thought to be platelet-derived
growth factor (PDGF), specifically PDGF-BB. Rats treated with blocking antibodies to
PDGF develop significantly smaller intimal lesions. These antibodies presumably inhibit
migration of SMCs, because there is no reduction in mitogenesis in the media or intima
(36). Similarly, infusion with PDGF-BB stimulates SMC migration and intimal thickening
but has little effect on SMC proliferation (37). Whether PDGF is a significant mitogen in
more complex lesion systems, such as in a human atherosclerotic vessel after angioplasty,
is unclear. Transfection of PDGF-B into swine arteries causes a marked increase in
replication (38). Such observations support the contention that PDGF-B may be a
mitogen in in vivo systems. Matrix metalloproteinases (MMPs) are also thought to have
a critical role in the migration associated with the second wave. MMPs are expressed in the
intima and are required for cell movement. MMP-3 and MMP-9 are expressed following
injury (39). The local overexpression of the tissue inhibitor of MMP-1 suppresses intimal
thickening by inhibiting SMC migration (40).

The third wave entails the replication of intimal SMCs, which may continue for weeks
to months (25). During this period, the intima overexpresses many growth factors and
receptors, including PDGF- A, Ang II receptor (AT-1), and transforming growth factor
(TGF-[J) (41,42). In addition, insulin-like growth factor I (IGF-1) is overexpressed in the
media, where there is also significant proliferation (43). The culprit molecule, blockade of
which would inhibit proliferation, has not been elucidated for the third wave in rats, but
there is some evidence in baboons that it is PDGF receptor (PDGFR-ct and PDGFR-p)
(M. Englesbe, A. Clowes, unpublished observations).

In a model of balloon arterial injury in baboons, there is a significant population of
inflammatory cells within the early intima. Inflammatory mediators may significantly
modulate the response to SMC to growth factors. For example, IL-1 p significantly inhibits
PDGF-BB-induced baboon SMC migration in an in vitro system while augmenting
PDGF-BB-induced mitogenesis (G. Daum and M. Englesbe, unpublished results). In a
mouse model, tumor necrosis factor alpha (TNF-a) and IL-1 modulate intimal hyper-
plasia induced by low shear stress (44).

The fourth wave is characterized by an increased intimal sensitivity to mitogens.
Specifically, the intima proliferates with infusion of endothelin, TGF-p, bFGF, and Ang II
(32). Proliferation and intimal size can be inhibited by blockade of the angiotensin II
receptor (45,46) or endothelin receptors (47,48). ■g

SMC migration is modulated by factors from platelets, specifically PDGF. PDGF is &

liberated at the time of platelet degranulation as well as synthesized and secreted by a

vascular wall cells (49). Intimal thickening in rats rendered thrombocytopenic is largely c

blocked even though the initial wave of SMC mitogenesis persists (50). Furthermore, <j

blockade of PDGFR-p can inhibit intimal development in both the rat and the baboon >9

(36,51). There are five known active forms of PDGF: PDGF AA, AB, BB, CC, and DD I

(52,53). Dimerization is required for high-affinity binding to the a and p receptor subunits. Q

PDGF-BB is thought to play a critical role in arterial injury response, but the role of the |

other isoforms is largely unknown (23). @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

72

ENGLESBE and CLOWES

Many novel approaches to inhibiting or reversing neointimal hyperplasia involve the
concept of negative growth control. SMCs in normal human arteries have a very low rate
of proliferation (0.01% per day), which translates into turnover once every 30 years. Only
upon injury do SMCs begin to proliferate. The inhibitory factors that prevent SMC
growth can be isolated from the endothelium, SMCs, inflammatory cells, and the matrix.

Certainly the endothelium is critical in modulating SMC quiescence. In the rat, upon
regeneration of the endothelium, SMCs cease proliferating (22). Nitric oxide produced by
endothelial cell nitric oxide synthase (eNOS) may be an important molecule in regulating
SMC growth. With removal of the endothelium and its associated NOS, the SMCs of the
intima proliferate. In contrast, in endothelialiazed baboon grafts, a change from high to
normal shear stress downregulates eNOS and mitogenesis occurs within the intima. Other
possible endothelial-derived inhibitors of SMC proliferation include COX-2 (54) and
prostaglandins I 2 and E 2 . T cells within the intima secrete interferon gamma, which induce
MHC II antigens and suppress SMC replication (55).

The matrix secreted in the intima helps maintain the quiescent state of vessel wall
SMCs. Within the matrix, heparan sulfates such as perlecan inhibit SMC migration and
proliferation in vitro (56-58). Perlecan has been proposed to bind to heparin-binding
mitogens, such as fibroblast growth factor-2 (FGF-2), and to prevent them from
stimulating smooth muscle cells (59). Matrix-mediated SMC inhibition may also occur
via heparin-mediated blockade of mitogen-associated protein kinases as well as cell-cycle
genes (e.g., c-myb) and proteases necessary for matrix degradation (60).

Approaches to the treatment of intimal hyperplasia target SMC migration and
proliferation. An alternative approach is treatment directed at lesions that already exist.
This approach would be applicable to the large number of patients currently suffering
from complications related to intimal hyperplasia. The baboon vascular graft model has

Normal Flow

High Flow

Lumen

Intima

V..,

«•'.

■mA.

S:M&?r y ' w - ; .: '•■■,

Graft

r V'' >, I

tiS ..*! ' " . it-,. >J* If v

s, ■ i ,: ■ - •: ..«.■ v*
, , '. '■ 'ft. '

i

2

I

■c

T . !*

Figure 2 Intimal regression induced by increased flow, demonstrating the dynamic nature of
intimal hyperplasia (H&E staining, x 40.) After 12 weeks of normal flow, the graft on the left
developed a significant intima. After 8 weeks of normal flow and then 4 weeks of high flow, the
intima lining the graft on the right atrophied significantly.

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

INTIMAL HYPERPLASIA 73

demonstrated the dynamic nature of intimal hyperplastic lesions (13). For example, with
the placement of an arteriovenous fistula distal to a vascular graft lined with an intima,
shear stress increases through the graft and induces intimal atrophy (61) (Fig. 2). The
atrophy is associated with a reduction in matrix content and cell number, with an
associated increase in intimal SMC apoptosis. Thus, matrix degradation and SMC
apoptosis are also potential targets for pharmacological manipulation.

Apoptosis is an important event in vasculogenesis and vascular injury response. In the
rat carotid artery injury model, even though SMCs continue to proliferate at the surface of
vessels lacking an endothelium, no net increase in SMC number occurs in these areas (25).
In addition, an increase in SMC apoptosis can be detected in these areas. Finally, apop-
tosis can not only be induced by increases in shear stress but it also may be responsible for
the termination of intimal thickening after injury (62-64). In all, control of apoptosis may
be a potential strategy for inducing intimal atrophy.

V. THERAPEUTIC APPROACHES TO INTIMAL HYPERPLASIA

A. Mechanical Approaches

The most widely used treatment for intimal hyperplasia is catheter-based direct mechan-
ical manipulation of hyperplastic lesions, generally by balloon dilation. It is estimated that
this approach is used for 150,000 coronary lesions annually (65). Balloon angioplasty of a
restenotic lesion within a stent is a safe short- and long-term approach for coronary and
peripheral lesions with inadequate initial expansion of the stent or focal neointimal
growths (66). In these procedures, of the total luminal enlargement, 56% is attributable
to further stent expansion, while 44% is from extrusion-axial redistribution of the intima
(67). In contrast, balloon treatment of diffuse intimal lesions in well-expanded stents has
poor intermediate and long-term outcomes (68). In such lesions, novel approaches have
been used such as atherectomy, in which blades excise the in-stent lesion. Unfortunately,
all of these mechanical approaches cause additional vascular wall injury in a patient
predisposed to intimal hyperplasia. Thus, even though mechanical approaches may
initially be effective, they do not address the injury response and stenosing intimal lesions
readily recur (69,70).

B. Cell Cycle-Directed Therapies

As discussed earlier, in response to vascular injury, SMCs enter the cell cycle and proliferate.
Cell-cycle inhibitors are therefore potential agents to prevent intimal hyperplasia. Gener-
ally, these agents must be locally delivered in order to prevent systemic toxicity.

The best-studied antiproliferative approach is vascular brachytherapy. Clinically

relevant approaches entail placement of "hot" intravascular catheters or stents. These ■g

devices emit either gamma or beta radiation. Ionizing radiation induces nonspecific breaks &

in chromosomal DNA and is a potent antiproliferative if the cells are actively dividing at a

the time of exposure. Initial trials have shown effective prevention of angiographic c

restenosis and reduced need for repeat revascularization (71-73). Unfortunately, vascular <j

brachytherapy has multiple drawbacks. The first is that the procedure can be technically >9

laborious and both the patient and the operators are at a significant risk of exposure to 4j

ionizing radiation. In addition, there may be an increased risk of late stent thrombosis, Q

since the radiation prevents endothelial regeneration (74). Finally, the most significant |

limitation of vascular brachytherapy is the phenomenon of "candy wrapper" restenosis @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

74 ENGLESBE and CLOWES

(75). In areas adjacent to a radioactive stent, intimal hyperplasia may actually be induced,
resulting in a patent stent with critical stenoses proximally and distally to the stent edge.
This phenomenon is due to lower doses of radiation delivered to the stent ends in the
setting of vessel wall balloon injury, leading to intimal hyperplasia and pathological
remodeling. Thus, after several years of research, it seems unlikely that radioactive stents
will soon become clinically useful tools with broad application.

Initially developed as a macrolide antibiotic and later used as an immunosuppressant,
Sirolimus (rapamycin) is a potent inhibitor of both SMC proliferation and migration.
This agent binds FK506-binding protein and inhibits TOR (target of Rapamycin), a
kinase that regulates cell cycle progression. Recent results from a phase I study using
rapamycin-coated stents in 30 patients have been encouraging. No significant toxicity was
noted and only 10% of the patients had > 15% stenosis from intimal hyperplasia 4 months
following stent placement (76). Rapamycin may prove to effectively reduce intimal hyper-
plasia, though more study of this novel approach is needed.

Another approach to antiproliferative therapy was used in the Project of Ex-vivo Vein
graft Engineering via Transfection (PREVENT) study (77). This ongoing trial examines
the delivery of E2F transcription factor decoy oligonucleotides as a means to inhibit vein
graft stenosis after peripheral bypass surgery. Phase I trials are encouraging with respect to
cellular access and were associated with minimal complications and a measurable decrease
in vein graft stenosis one year after peripheral vascular bypass.

Paclitaxel(Taxol) and related drugs are used as cancer chemotherapeutic agents. They
have diverse mechanisms of action on SMCs, including microtubule stabilization, arrest of
cell mitosis, and retardation of cell migration (78,79). Sustained local delivery of paclitaxel
by coating of intravascular stents prevented early in-stent restenosis in animal models
(80,81), but this inhibition was not maintained over the long term in one study (82).
Paclitaxel-coated stents have demonstrated encouraging results in one small human trial
(83), but there is a report of acute thrombosis of a paclitaxel-coated stent seven months
after implantation (84). A randomized, controlled human trial with taxol coated coronary
stents is in progress.

There are many approaches to introduce exogenous genetic material into the vascular
system. Current approaches involve primarily adenoviral vectors and cell-based gene
transfer. The complexity of vector optimization and techniques of transfection are beyond
the scope of this chapter. Nonetheless, gene therapy-based models have provided critical
tools in the study of vessel injury response and offer therapeutic promise. Many gene therapy
approaches to intimal hyperplasia focus on either inhibiting cell-cycle entry (cytostatic
approach) or causing death to cells that have entered the cell cycle (cytotoxic approach) (85).

Several genes have been targeted to induce SMC cytostasis at the site of injury. One
approach is to overexpress a constitutively active mutant form of the retinoblastoma
protein (RB), which complexes with the E2F family of transcription factors to block entry ■g

into the cell cycle (86). SMC proliferation and intimal development in a rat following &

balloon injury has been inhibited by overexpression of the endogenous cyclin kinase a

inhibitor protein p21 (87). This protein inhibits cyclin-dependent kinases and thus inhibits c

S-phase entry. Other genes that have shown cytostatic properties in in vivo models include <j

the homeodomain gene gax (88) and dominant-negative mutants of ras (89). >9

The cytotoxic approach generally involves the expression of enzymes capable of con- %

verting nucleoside analogues into toxic metabolites that interrupt DNA replication. The 2

result is death of transduced cells entering S phase in the presence of the nucleoside an- |

alogue (86). Intimal hyperplasia has been inhibited in animal models using a thymidine @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

INTIMAL HYPERPLASIA

75

kinase isozyme derived from the herpes simplex viral genome, which phosphorylates the
inactive drug gancyclovir and interrupts DNA synthesis (90,91). Induction of apoptosis is
another cytotoxic approach. As SMC apoptosis in the setting of arterial injury is better
understood, novel gene therapy approaches involving modulation of apoptosis will likely
develop.

There is growing evidence for a role of stem cells in arterial injury response. Statin
drugs may benefit patients with coronary artery disease by increasing the pool of
circulating bone marrow-derived endothelial progenitor cells (92). Vascular trauma
induces the release of circulating endothelial precursor cells (93). Such precursor cells
may be involved in endothelial regeneration at the site of vascular wall injury. Stem cell
technology offers many new approaches for preventing restenosis, but the field is too new
to foresee clinical applicability in the near future.

C. Growth Factor Inhibition and Matrix Modulation

Many growth factors modulate the vascular injury response (Table 2). These growth
factors are responsible for SMC migration and proliferation and act as survival signals
preventing initiation of the apoptotic cascade. Blocking antibodies, antisense strategies, or
gene therapy approaches can modulate these growth factors.

The most extensively studied growth factor system is PDGF. In multiple arterial
injury and vascular graft models, blockade of the PDGFR-fJ has been shown to inhibit
intimal hyperplasia (36,51,94,95). Blockade of the PDGFR-fJ is thought to function by
inhibiting the migration of SMCs into the intima. A human coronary stent trial using a
chimeric PDGFR-fJ-blocking antibody has failed to demonstrate significant angiographic
improvements in stent patency (P.W. Serruys, personal communication). Atherosclerotic
arteries likely express a larger quantity of PDGFR-a than the healthy vessels used in
animal models. In vitro, blockade of both the PDGFR-a and -[J simultaneously inhibits
migration and mitogenesis more than blockade of the PDGFR-fJ alone. In vivo studies to
block both receptors are ongoing in baboons.

The complexity of the injury response is evidenced in the number of potential
therapeutically targeted growth factors. Additional potential targets include IGF-I and
-II, thrombin, epidermal growth factor (EGF), heparin binding-EGF (HB-EGF), b-FGF,

Table 2 Summary of Growth Factors Relevant in Intimal Hyperplasia and Associated Inhibitors
that Have Demonstrated Efficacy at Inhibiting Intimal Hyperplasia in Vivo

Growth factor

Growth factor inhibitor

PDGF AA, AB, BB, CC, DD

EGF, HB-EGF

Endothelin

bFGF

MMP-3 and MMP-9

TGF-pl, TGF-P2, TGF-JJ3

Antibody to PDGFR a or [J (36,37,51)

PDGF chain aptamers (107)

PDGF-receptor tyrosine kinase inhibitors (108)

Antibody to EGFR a

Endothelin (A/B) receptor antagonists (47,48)

Antagonists to FGFR (109,110)

Tissue inhibitor of MMP-1 (111,112)

Antibody to TGF-fJ (113)

Recombinant TGF-beta RII (114)

i

2

I

■c

1 Chan A. Clowes A, unpublished observation.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

76 ENGLESBE and CLOWES

NOS, fibroblast growth factor I (FGF-I), and transforming growth factor (31 (TGF-p). In
addition, over-expression of NOS and local release of NO can inhibit SMC proliferation.
Modulation of the intimal matrix is another approach to inhibit intimal growth. Smooth
muscle cells transduced with replication-defective retrovirus encoding tissue inhibitor of
matrix metalloproteinase-1 (TIMP-1), when seeded onto the surface of the artery, suppress
intimal thickening (40). MMP-2 and MMP-9 are expressed in smooth muscle cells fol-
lowing injury. Blockade of these MMPs either with an active site inhibitor drug or local
overexpression of the natural inhibitor (TIMP-1), suppresses proteolytic activity and mi-
gration of smooth muscle cells.

D. Antiplatelet and Anticoagulation Agents

Adhesion of platelets and thrombus formation at the site of vascular wall trauma is involved
in the response to injury. Theoretically, prevention of early platelet adhesion and thrombus
formation and the resultant exposure of the activated SMCs to the multitude of growth
factors within platelet granules and thrombus may have long-term effects on patency.
Multicenter randomized trials have shown that an aspirin with or without dipyridamole
increased the long-term patency of aortocoronary saphenous vein grafts (96,97). Similarly,
ticlopidine improves the long-term patency of saphenous vein bypass grafts in the legs (98).
Unfortunately, vessels likely need exposure to these drugs immediately following the injury,
which may be associated with increased bleeding complications from surgical sites.

In addition, heparin alone inhibits intimal formation in animals following arterial
injury (99,100). The potential mechanisms include regulation of extracellular matrix
composition (101), inhibition of SMC migration and proliferation, inhibition of nuclear
transcription factors (102), and primary anti-inflammatory effects. Oral delivery of
heparin has recently shown promising results in animals in the prevention of in-stent
restenosis (103). Human trials have not demonstrated a similar inhibition of intimal
growth following balloon angioplasty using systemic delivery of either unfractionated
(104) or low-molecular-weight heparin (105). The poor results from these human trials
may be due to the inappropriate or inadequate dosing of heparin to the study patients. A
recent human trial with local delivery of low-molecular- weight heparin has demonstrated a
significant inhibition of intimal hyperplasia in coronary stents (106).

VI. CONCLUSIONS

Animal and clinical studies over the past two decades have documented that intimal hyper-
plasia is a response to injury and a complication of all forms of arterial reconstruction.
Although pathological remodeling (inward remodeling) is a more important mechanism
for luminal narrowing in vessels treated with balloon angioplasty, intimal hyperplasia "g

followed by stenosis or restenosis is the principal cause of failure in vein and synthetic &

grafts as well as stented arteries. Until recently, symptomatic restenosis of a stented artery a

or stenosis of a bypass graft could not be treated and usually required further surgical c

reconstruction. <j

y
Encouraging results have been obtained with brachytherapy and mechanical inter- «

ventions. Novel therapies should be compared to these treatments. Preliminary trials using 41

paclitaxel- and rapamycin-coated stents have generated very encouraging results. 2

The intimal lesion is composed primarily of SMCs. Unfortunately, most of the |

currently investigated therapeutic approaches are not SMC specific and also inhibit @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

INTIMAL HYPERPLASIA 77

endothelial regeneration and function. It is possible that failure of endothelial regeneration
will have significant implications such as late thrombosis. Adjuvant therapy currently
under development and targeted specifically at SMCs (e.g., PDGF or EGF) would be
expected to circumvent this problem.

REFERENCES

1. Williams DO, Holubkov R, Yeh W, et al. Percutaneous coronary interventions in the current
era compared with 1985-1986: The National Heart, Lung, and Blood Institute Registries.
Circulation 2000; 102:2945-2951.

2. Grondin CM, Campeau L, Lesperance J, Enjalbert M, Bourassa MG. Comparison of late
changes in internal mammary artery and saphenous vein grafts in two consecutive series of
patients 10 years after operation. Circulation 1984; 70(supp 1):208-212.

3. Bourassa MG, Enjalbert M, Campeau L. Progression of atherosclerosis in coronary arteries
and bypass grafts: Ten years later. Am J Cardiol 1984; 52:102C-107C.

4. Dalman RL, Taylor LM Jr. Basic data related to infrainguinal revascularization procedures.
Ann Vase Surg 1990; 4:309-312.

5. Thatte HS, Khuri SF. The coronary artery bypass conduit: I. Intraoperative endothelial injury
and its implication on graft patency. Ann Thorac Surg 2001; 72:S2245-S2252.

6. Gibson KD, Gillen DL, Caps MT, Kohler TR, Sherrard DJ, Stehman-Breen CO. Vascular
access survival and incidence of revisions: A comparison of prosthetic grafts, simple auto-
genous fistulas, and venous transposition fistulas from the United States Renal Data System
Dialysis Morbidity and Mortality Study. J Vase Surg 2001; 34:694-700.

7. Dilley JR, McGeachie JK, Prendergast FJ. A review of histologic changes in vein to artery
grafts, with particular reference to intimal hyperplasia. Arch Surg 1988; 123:691-699.

8. DePalma RG, Chidi CC, Sternfeld WC. Pathogenesis and prevention of trauma provoked
atheromas. Surgery 1977; 74:931-944.

9. Bonchek LI. Prevention of endothelial damage during preparation of saphenous veins for
bypass grafting. J Thorac Cardiovasc Surg 1980; 79:911-915.

10. Cox JL, Chiasson DA, Gotlieb AL Stranger in a strange land: The pathogenesis of saphenous
vein graft stenosis with emphasis on structural and functional differences between veins and
arteries. Prog Cardiovasc Dis 1991; 144:45-68.

11. Davies MG, Klyachkin ML, Dalen H, Svendsen E, Hagen PO. Regression of intimal hyper-
plasia with restoration of endothelium-dependant relaxing factor-mediated relaxation in ex-
perimental vein grafts. Surgery 1993; 114:258-271.

12. Zwolak RM, Adams MC, Clowes AW. Kinetics of vein graft hyperplasia: Association of tan-
gential stress. J Vase Surg 1987; 5:126-136.

13. Kohler T, Kirkman TR, Clowes AW. The effect of external rigid support on vein graft ad-
aptation to arterial circulation. J Vase Surg 1991; 9:277-285.

14. Veith FJ, Gupta SK, Ascer E. Six year prospective randomized comparison of autologous

saphenous vein and expanded polytetrafluoroethylene grafts in infrainguinal arterial re- •o

constructions. J Vase Surg 1986; 3:104-111. |

15. Clowes AW. Intimal hyperplasia and graft failure. Cardiovasc Pathol 1993; 2:179S-1786S. \

16. Clowes AW, Kirkman TR, Reidy MA. Mechanism of arterial graft healing: Rapid transmural H,

'Eh

capillary ingrowth provides a source of intimal endothelium and smooth muscle cell in porous a

PTFE prostheses. Am J Pathol 1986; 123:220-230. g

17. Kimura T, Kaburagi S, Tamura T, et al. Remodeling of human coronary arteries undergoing n-
angioplasty or atherectomy. Circulation 1997; 96:475-483. S

18. Post Mj, Borst C, Kuntz RE. The relative importance of arterial remodeling compared with !3
intimal hyperplasia in lumenal renarrowing after balloon angioplasty: A study in the normal I
rabbit and the hypercholesterolemic Yucatan micropig. Circulation 1994; 89:2816-2821. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

78 ENGLESBE and CLOWES

19. Mondy JS, Williams JK, Adams MR, Dean RH, Geary RL. Structural determinants of
luminal narrowing after angioplasty in atherosclerotic non-human primates. J Vase Surg
1997; 26:875-883.

20. Zalewski A, Shi Y. Vascular myofibroblasts: Lessons from coronary repair and remodeling.
Arterioscler Thromb Vase Biol 1997; 17:417-422.

21. Ormiston JA, Stewart FM, Roche AH, et al. Late regression of the dilated site after coronary
angioplasty: A five year quantitative angiographic study. Circulation 1997; 96:468-474.

22. Fingerle J, Au YP, Clowes AW, Reidy MA. Intimal lesion formation in rat carotid arteries
after endothelial denudation in absence of medial injury. Arteriosclerosis 1990; 10:1082-
1087.

23. Allaire E, Clowes AW. Endothelial cell injury in cardiovascular surgery: The intimal hyper-
plastic response. Ann Thorac Surg 1997; 63:582-591.

24. Nikkari ST, Jarvelainen HT, Wight TN, Ferguson M, Clowes AW. Smooth muscle cell ex-
pression of extracellular matrix genes after arterial injury. Am J Pathol 1994; 144:1348-1356.

25. Clowes AW, Clowes MM, Reidy MA. Kinetics of cellular proliferation after arterial injury.
Endothelial and smooth muscle cell growth in chronically denuded vessels. Lab Invest 1986;
54:295-303.

26. Kuntz RE, Baim DS. Prevention of coronary restenosis: Evolving evidence base for radiation
therapy. Circulation 2000; 101:2130-2133.

27. Mehran R, Dangas G, Mintz GS, et al. Patterns of in stent restenosis: Angiographic clas-
sification and implications for long-term clinical outcome. Circulation 1999; 100:1872-1878.

28. Farb A, Sangiorgi G, Carter AJ, Walley VM, Edwards WD, Schwartz RS, Virmani R.
Pathology of acute and chronic coronary stenting in humans. Circulation 1999; 99:44-52.

29. Grewe P, Deneke T, Machraoui A. Acute and chronic tissue response to coronary stent
implantation: Pathologic findings in human specimen. J Am Coll Cardiol 2000; 35:157-163.

30. Clowes AW, Schwartz SM. Significance of quiescent smooth muscle cell migration in the
injured rat carotid artery. Circ Res 1985; 56:139-145.

3 1 . Hanke H, Strohschneider T, Oberhoff M, Betz E, Karsch KR. Time course of smooth muscle
cell proliferation in the intima and media of arteries following experimental angioplasty. Circ
Res 1990; 67:651-659.

32. Schwartz SM, dsBlois D, O'Brien ERM. The intima: Soil for atherosclerosis and restenosis.
Circ Res 1995; 77:445^165.

33. Olson NE, Chao S, Linder V, Reidy MA. Intimal smooth muscle proliferation after balloon
catheter injury: The role of basic fibroblast growth factor. Am J Pathol 1992; 140:1017-1023.

34. Ebbecke M, Unterberg C, Buchwald A, Stohr S, Wiegand V. Antiproliferative effects of c-myc
antisense oligonucleotide on human arterial smooth muscle cells. Basic Res Cardiol 1992;
87:585-591.

35. Simons M, Edelman ER, Rosenberg RD. Antisense proliferating cell nuclear antigen oligo-
nucleotides inhibit intimal hyperplasia in a rat carotid artery injury model. J Clin Invest 1994;
93:2351-2356.

36. Ferns GAA, Raines EW, Sprugel KH, Motani AS, Reidy MA, Ross R. Inhibition of neo-
intimal smooth muscle accumulation after angioplasty by antibody to PDGF. Science 1991;
253:1129-1132. 1

37. Jawien A, Bowen-Pope DF, Lindner V, Schwartz SM, Clowes AW. Platelet-derived growth <S
factor promotes smooth muscle migration and intimal thickening in a rat model of balloon js
angioplasty. J Clin Invest 1992; 89:507-511. 2

38. Nabel EG, Yang Z, Liptay S, San H, Gordon D, Haudenschild CC, Nabel GJ. Recombinant ^
platelet-derived growth factor B gene expression in porcine arteries induce intimal hyperplasia «
in vivo. J Clin Invest 1993; 91:1822-1829. J

39. Bendeck MP, Zempo N, Clowes AW, Galardy RE, Reidy MA. Smooth muscle cell migration q
and matrix metalloproteinase expression after arterial injury in the rat. Circ Res 1994; 75:539- |
545. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

INTIMAL HYPERPLASIA 79

40. Forough R, Koyama N, Hasenstab D, Lea H, Clowes M, Nikkari ST, Clowes AW. Over-
expression of tissue inhibitor of matrix metalloproteinase-1 inhibits vascular smooth muscle
cell functions in vitro and in vivo. Circ Res 1996; 79:812-820.

41. Majesky MW, Linder V, Twardzik DR, Schwartz SM, Reidy MA. Production of trans-
forming growth factor (J-l during repair of arterial injury. J Clin Invest 1993; 92:2952.

42. Viswanathan M, Stromberg C, Seltzer A, Saavedra JM. Balloon angioplasty enhances the
expression of angiotensin II ATI receptors in neointima of rat aorta. J Clin Invest 1992; 90:
1707-1712.

43. Cercek B, Fishbein MC, Forrester JS, Helfant RH, Fagin JA. Induction of insulin like growth
factor I messenger RNA in rat aorta after balloon denudation. Circ Res 1990; 66: 1755-1760.

44. Rectenwald JE, Moldawer LL, Huber TS, Seeger JM, Ozaki CK. Direct evidence for cyto-
kine involvement in neointimal hyperplasia. Circulation 2000; 102:1697-1702.

45. Yuda A, Takai S, Jin D, Sawada Y, Nishimoto M, Matsuyama N, Asada K, Kondo K, Sasaki
S, Miyazaki M. Angiotensin II receptor antagonist, L-158,809, prevents intimal hyperplasia in
dog grafted veins. Life Sci 2000; 68:41-48.

46. Virone-Oddos A, Desangle V, Provost D, Cazes M, Caussade F, Cloarec A. In vitro and in
vivo effects of UP 269-6, a new potent orally active nonpeptide angiotensin II receptor anta-
gonist, on vascular smooth muscle cell proliferation. Br J Pharmacol 1997; 120:488-494.

47. Azuma H, Sato J, Masuda H, Goto M, Tamaoki S, Sugimoto A, Hamasaki H, Yamashita H.
ATZ1993, an orally active and novel nonpeptide antagonist for endothelin receptors and
inhibition of intimal hyperplasia after balloon denudation of the rabbit carotid artery. Jpn J
Pharmacol 1999; 81:21-28.

48. Marano G, Palazzesi S, Bernucci P, Grigioni M, Formigari R, Ballerini L. ET(A)/ET(B)
receptor antagonist bosentan inhibits neointimal development in collared carotid arteries of
rabbits. Life Sci 1998; 63:PL259-PL266.

49. Battegay EJ, Raines EW, Colbert T, Ross R. TNF-alpha stimulation of fibroblast prolif-
eration. Dependence on platelet-derived growth factor (PDGF) secretion and alteration of
PDGF receptor expression. J Immunol 1995; 154:6040-6047.

50. Fingerle J, Johnson R, Clowes AW, Majesky MW, Reidy MA. Role of platelets in smooth
muscle cell proliferation and migration after vascular injury in rat carotid artery. Proc Natl
Acad Sci USA 1989; 86:8412.

51. Davies MG, Owens EL, Mason DP, Lea H, Tran K, Vergel S, Hawkins SA, Hart CE, Clowes
AW. Effect of platelet-derived growth factor receptor-alpha and -beta blockade on flow-
induced neointimal formation in endothelialized baboon vascular grafts. Circ Res 2000; 86(7):
779-786.

52. Bergsten E, Uutela M, Li X, Pietras K, Ostman A, Heldin CH, Alitalo K, Eriksson U. PDGF-
D is a specific, pro tease-activated ligand for the PDGF beta-receptor. Nat Cell Biol 2001;
3:512-516.

53. Uutela M, Lauren J, Bergsten E, Li X, Horelli-Kuitunen N, Eriksson U, Alitalo K. Chro-
mosomal location, exon structure, and vascular expression patterns of the human PDGFC
and PDGFC genes. Circulation 2001; 103:2242-2247.

54. Topper JN, Cai J, Falb D, Gimbrone MA. Identification of vascular endothelial genes

differentially responsive to fluid mechanical stimuli: cyclooxygenase-2, manganese superoxide j>

dismutase, and endothelial nitric oxide synthase are selectively up-regulated by steady laminar <S

shear stress. Proc Natl Acad Sci USA 1996; 93:10417-10422. 1

55. Hansson GK, Holm J. Interferon-gamma inhibits arterial stenosis after injury. Circulation 2
1991; 84:1266-1272. " «!

56. Bingley JA, Campbell JH, Hayward IP, Campbell GR. Inhibition of neointimal formation by «
natural heparin sulfate proteoglycans of the arterial wall. Ann NY Acad Sci 1 997; 8 1 1 :238-242. jjj

57. Han RO, Ettenson DS, Koo EW, Edelman ER. Heparin/heparin sulfate chelation inhibits q
control of vascular repair by tissue engineered endothelial cells. Am J Physiol 1997; 273: |
H2586-H2595. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

80 ENGLESBE and CLOWES

58. Nugent MA, Nugent HM, Iozzo RV, Sanchack K, Edelman ER. Perlecan is required to
inhibit thrombosis after deep vascular injury and contributes to endothelial cell-mediated
inhibition of intimal hyperplasia. Proc Natl Acad Sci USA 2000; 97:6722-6727.

59. Forsten KE, Courant NA, Nugent MA. Endothelial proteoglycans inhibit bFGF binding and
mitogenesis. J Cell Physiol 1997; 172:209-220.

60. Au YPT, Kenagy RD, Clowes MM, Clowes AW. Mechanisms of inhibition of heparin of vas-
cular smooth muscle cell proliferation and migration. Haemostasis 1993; 23(suppl 1): 177—182.

61. Mattson EJR, Kohler TR, Vergel S, Clowes AW. Increased blood flow induces regression of
intimal hyperplasia. Arterioscler Thromb Vase Biol 1997; 17:2245-2249.

62. Schwartz SM, Bennett MR. Death by no other name. Am J Pathol 1995; 147:229-234.

63. Schwartz SM. Cell Death and the caspase cascade. Circulation 1998; 97:227.

64. Bennett MR, Evan GI, Schwartz SM. Apoptosis of rat vascular smooth muscle cells is
regulated by p-53 dependant and independent pathways. Circ Res 1995; 77:266-273.

65. Di Mario C, Marisco F, Adamian M, Karvouni E, Albiero R, Colombo A. New recipes for
in-stent restenosis: Cut, grate, roast, or sandwich the intima? Heart 2000; 84:471-475.

66. Wolf A, Schwarz F, Hoffmann M, et al. Repetitive balloon angioplasty for recurrent in-stent
restenosis: results of first and second PTCA in focal and diffuse stenosis [abstr]. J Am Coll
Cardiol 1999; 33:26A.

67. Mehran R, Mintz G, Popma J. Mechanisms and results of balloon angioplasty for the in-stent
restenosis. Am J Cardiol 1996; 78:618-622.

68. Eltchaninoff H, Koning R, Tron C, et al. Balloon angioplasty for the treatment of coronary
in-stent restenosis: Immediate results and 6 months angiographic recurrent restenosis rate. J
Am Coll Cardiol 1998; 32:980-984.

69. Harrington RA, Califf RM, Holmes DR Jr, Pieper KS, Lincoff AM, Berdan LG, Thompson
TD, Topol EJ. Is all unstable angina the same? Insights from the Coronary Angioplasty
Versus Excisional Atherectomy Trial (CAVEAT-I). The CAVEAT-Investigators. Am Heart J
1999; 137:199-203.

70. Cohen EA, Sykora K, Kimball BP, Bonan R, Ricci DR, Webb JG, Laramee L, Barbeau G,
Traboulsi M, Corbett BN, Schwartz L, Adelman AG. Clinical outcomes of patients more than
one year following randomization in the Canadian Coronary Atherectomy Trial (CCAT).
Can J Cardiol 1997; 13:825-830.

71. Teirstein PS, Massullo V, Jani S, et al. Catheter-based radiotherapy to inhibit restenosis after
coronary stenting. N Engl J Med 1997; 336:1697-1703.

72. Waksman R, White RL, Chan RC, et al. Intra-coronary gamma irradiation therapy after
angioplasty inhibits recurrence in patients with in-stent restenosis. Circulation 2000; 101:
2165-2171.

73. Leon MB, Teirstein PS, Moses J, et al. Localized intra-coronary gamma-radiation therapy to
inhibit the recurrence of restenosis after stenting. N Engl J Med 2001; 344:250-256.

74. Seabra-Gomes R. Radioactive stents to reduce restenosis: Time for an epitaph? Eur Heart J
2001; 22:621-623.

75. Albiero R, Adamian M, Kobayashi N, et al. Short and intermediate term results of 32 P
radioactive p-emitting stent implantation in patients with coronary artery disease. The Milan

dose response study. Circulation 2000; 101:18-26. j>

76. Sousa JE, Costa MA, Abizaid A, et al. Lack of neointimal proliferation after implantation of <S
sirolimus-coated stents in human coronary arteries. Circulation 2001; 103:192-195. js

77. Mann MJ, Whittemore AD, Donaldson MC, et al. Ex-vivo gene therapy of human vascu- 2
lar bypass grafts with EF2 decoy: The PREVENT single center, randomized, controlled trial. **
Lancet 1999; 354:1493-1498. ■§

78. Rowinsky EK, Donehower RC. Paclitaxel (Taxol). N Engl J Med 1995; 332:1004-1014. J

79. Axel DI, Kunert W, Goggelmann C, et al. Paclitaxel inhibits arterial smooth muscle cell q
proliferation and migration in vitro and in vitro using local drug delivery. Circulation 1997; |
96:636-645. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

INTIMAL HYPERPLASIA 81

80. Drachman DE, Edelman ER, Seifert P, et al. Neointimal thickening after stent delivery of
paclitaxel: Change in composition and arrest of growth over six months. J Am Coll Cardiol
2000; 36:2325-2332.

81. Heldman AW, Cheng L, Jenkins M. Paclitaxel stent coating inhibits neointimal hyperplasia at
4 weeks in a porcine model of coronary restenosis. Circulation 2001; 103:2289-2295.

82. Farb A, Heller PF, Shroff S, et al. Pathological analysis of local delivery of paclitaxel via a
polymer-coated stent. Circulation 2001; 104:473-479.

83. de la Fuente LM, Miano J, Mrad J, Penaloza E, Yeung AC, Eury R, Froix M, Fitzgerald
PJ, Stertzer SH. Initial results of the Quanam drug eluting stent (QuaDS-QP-2) Registry
(BARDDS) in human subjects. Catheter Cardiovasc Intervent 2001; 53:480-488.

84. Liistro F, Colombo A. Late acute thrombosis after paclitaxel eluting stent implantation.
Heart 2001; 86:262-264.

85. De Young MB, Dichek DA. Gene therapy for restenosis: Are we ready? Circ Res 1998; 82:
306-313.

86. Baek S, March KL. Gene therapy for restenosis: Getting nearer the heart of the matter. Circ
Res 1998; 82:295-305.

87. Chang MW, Barr E, Lu MM, Barton K, Leiden JM. Adenovirus mediated overexpression of
the cyclin/cyclin dependant kinase inhibitor, p21, inhibits vascular smooth muscle cell pro-
liferation and neointimal formation in the rat carotid artery model of balloon injury. J Clin
Invest 1995; 96:2260-2268.

88. Smith RC, Branellec D, Gorski DH, Guo K, Perlman H, Dedieu JF, Pastore C, Mahfoudi A,
Denefle P, Isner JM, Walsh K. p21CIPl -mediated inhibition of cell proliferation by over-
expression of the gax homeodomain gene. Genes Dev 1997; 11:1674-1689.

89. Ueno H, Yamamoto H, Ito S, Li JJ, Takeshita A. Adenovirus-mediated transfer of a dom-
inant-negative H-ras suppresses neointimal formation in balloon injured arteries in vivo.
Arterioscler Thromb Vase Biol 1997; 17:904-989.

90. Ohno T, Gordon D, San H, Pompili VJ, Imperiale MJ, Nabel GJ, Nabel EG. Gene ther-
apy for vascular smooth muscle cell proliferation after arterial injury. Science 1994; 265:781-
784.

91. Simari RD, San H, Rekhter M, Ohno T, Gordon D, Nabel GJ, Nabel EG. Regulation of
cellular proliferation and intimal formation following balloon injury in atherosclerotic rabbit
arteries. J Clin Invest 1996; 98:225-235.

92. Vasa M, Fichtlscherer D, Adler K, Aicher A, Martin H, Zeiher AM, Dimmeler S. Increase in
circulating endothelial progenitor cells by statin therapy in patients with stable coronary
artery disease. Circulation 2001; 103:2885-2890.

93. Gill M, Dias S, Hattori K, Rivera ML, Hicklin D, Witte L, Girardi L, Yurt R, Himel H, Rain
S. Vascular trauma induces rapid but transient mobilization of VEGFR2 + AC133 + endo-
thelial precursor cells. Circ Res 2001; 88:167-174.

94. Hart CE, Kraiss LW, Vergel S, Gilbertson D, Kenagy R, Kirkman T, Crandall DL, Tickle S,
Finney H, Tarranton G, Clowes AW. PDGFp. receptor blockade inhibits intimal hyperplasia
in the baboon. Circulation 1999; 99:564-569.

95. Sirois MG, Simons M, Edelman ER. Antisense oligonucleotide inhibition of PDGF-(i re-

ceptor subunit expression directs suppression of intimal thickening. Circulation 1997; 95:669- j>
676.

96. Chesebro JH, Clements IP, Fuster V, et al. A platelet-inhibitor drug trial in coronary artery js
bypass operations: benefit of perioperative dipyridamole and aspirin therapy on early post- 2
operative vein graft patency. N Engl J Med 1982; 307:73-78. ^

97. Lorenz RL, Schacky CV, Weber M, et al. Improved aorto-coronary bypass patency by low >9
dose aspirin ( 1 OOmg daily): Effects on platelet aggregation and thromboxane formation. Lancet jjj
1984;1:1261-1264. q

98. Becquemin JP. Effect of ticlopidine on the long-term patency of saphenous-vein bypass grafts |
in the legs. N Engl J Med 1997; 337:1726-1731. S

I
u

s

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

82 ENGLESBE and CLOWES

99. Clowes AW, Karnovsky MJ. Suppression by heparin of smooth muscle cell proliferation in
injured arteries. Nature 1977; 265:625-626.

100. Clowes AW, Clowes MM. Kinetics of cellular proliferation after arterial injury: II. Inhibition
of smooth muscle cell growth by heparin. Lab Invest 1985; 52:611-616.

101. Snow AD, Bolender RP, Wright TN, et al. Heparin modulates the composition of the
extracellular matrix domain surrounding arterial smooth muscle cells. Am J Physiol 1990; 137:
313-330.

102. Pukac LA, Castellot JJ, Wright TJ, et al. Heparin inhibits c-fos and c-myc mRNA expression
in vascular smooth muscle cells. Cell Regulation 1990; 1 :435^443.

103. Welt FGP, Woods TC, Edelman ER. Oral heparin prevents neointimal hyperplasia after ar-
terial injury. Circulation 2001; 104:3121-3124.

104. Ellis SG, Roubin GS, Wilentz J, et al. Effect of 18 to 24 hour heparin administration for
prevention of restenosis after uncomplicated coronary angioplasty. Am Heart J 1989; 117:
777-782.

105. Faxon DP, Spiro TE, Minor S, et al. Low molecular weight heparin in prevention of re-
stenosis after angioplasty. Results of the Enoxaparin Restenosis (ERA) Trial. Circulation
1994; 90:908-914.

106. Kiesz RS, Buszman P, Martin JL, Deutsch E, Rozek MM, Gaszewska E, Rewicki M,
Seweryniak P, Kosmider M, Tendera M. Local delivery of enoxaparin to decrease restenosis
after stenting: Results of initial multicenter trial: Polish-American Local Lovenox NIR As-
sessment study (The POLONIA study). Circulation 2001; 103:26-31.

107. Floege J, Ostendorf T, Janssen U, Burg M, Radeke HH, Vargeese C, Gill SC, Green LS,
Janjic N. Novel approach to specific growth factor inhibition in vivo: Antagonism of platelet-
derived growth factor in glomerulonephritis by aptamers. Am J Pathol 1999; 154: 169-179.

108. Bilder G, Wentz T, Leadley R, Amin D, Byan L, O'Conner B, Needle S, Galczenski H,
Bostwick J, Kasiewski C, Myers M, Spada A, Merkel L, Ly C, Persons P, Page K, Perrone M,
Dunwiddie C. Restenosis following angioplasty in the swine coronary artery is inhibited by an
orally active PDGF-receptor tyrosine kinase inhibitor, RPR101511A. Circulation 1999; 99:
3292-3299.

109. Neschis DG, Safford SD, Hanna AK, Fox JC, Golden MA. Antisense basic fibroblast growth
factor gene transfer reduces early intimal thickening in a rabbit femoral artery balloon injury
model. J Vase Surg 1998; 27:126-134.

110. Chen C, Mattar SG, Hughes JD, Pierce GF, Cook JE, Ku DN, Hanson SR, Lumsden AB.
Recombinant mitotoxin basic fibroblast growth factor-saporin reduces venous anastomotic
intimal hyperplasia in the arteriovenous graft. Circulation 1996; 94:1989-1995.

111. Forough R, Koyama N, Hasenstab D, Lea H, Clowes M, Nikkari ST, Clowes AW. Over-
expression of tissue inhibitor of matrix metalloproteinase-1 inhibits vascular smooth muscle
cell functions in vitro and in vivo. Circ Res 1996; 79:812-820.

1 12. Kanamasa K, Otani N, Ishida N, Inoue Y, Ikeda A, Morii H, Naito N, Hayashi T, Ishikawa
K, Miyazawa M. Suppression of cell proliferation by tissue plasminogen activator during the
early phase after balloon injury minimizes intimal hyperplasia in hypercholesterolemic rab-
bits. J Cardiovasc Pharmacol 2001; 37:155-162.

113. Wolf YG, Rasmussen LM, Ruoslahti E. Antibodies against transforming growth factor-beta j>
1 suppress intimal hyperplasia in a rat model. J Clin Invest 1994; 93:1172-1178. <S

1 14. Smith JD, Bryant SR, Couper LL, Vary CP, Gotwals PJ, Koteliansky VE, Lindner V. Soluble J
transforming growth factor-beta type II receptor inhibits negative remodeling, fibroblast 2
transdifferentiation, and intimal lesion formation but not endothelial growth. Circ Res 1999; ■"■
84:1212-1222. £

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

5

The Healing Characteristics, Durability,

and Long-Term Complications

of Vascular Prostheses

Glenn C. Hunter

University of Texas Medical Branch, Galveston, Texas, U.S.A.

David A. Bull

University of Utah Health Sciences Center, Salt Lake City, Utah, U.S.A.

Arterial and venous autografts remain the materials of choice to replace diseased or
damaged blood vessels. However, because of their limited supply, there is an increasing
need to develop arterial substitutes that are durable, are readily incorporated by host
tissues, possess a non- or hypothrombogenic flow surface, have compliance characteristics
that closely approximate the native vessel, are resistant to infection, and are easily sutured
(1,2). Although none of the currently available prostheses manifests all of the desired
characteristics of the ideal arterial replacement, large-diameter Dacron grafts used to
replace the abdominal aorta have proved adequate, with 5-year cumulative patency rates
of 85-90% (3-5). Late patency rates of 74 and 70% at 10 and 15 years, respectively, have
been reported by Nevelsteen et al. (5). Unfortunately, the longevity of small-diameter
prosthetic grafts (6 mm in internal diameter or less) is limited by the development of
anastomotic intimal hyperplasia and consequent thrombosis of the grafts. When small-
diameter prosthetic grafts such as those fabricated from polytetrafluoroethylene or ■§
Dacron are placed above the knee, cumulative patency rates range from 37.9 to 71%; g
below the knee, they range from 30 to 57% (6-9). a
The increasing longevity of the population and the more frequent execution of bypass c
procedures to the more distal vessels of the extremity have resulted in increases in both the <
number and complexity of the bypass procedures now performed. It is estimated that >9
approximately one-third of patients will require additional surgery related to their bypass J
graft within 2 years of the initial procedure (10). Furthermore, complications resulting from «
errors in technique or deterioration of the graft fabric are not uncommon, and the risk of |
reoperation is increased substantially by the progression of ischemic heart disease and other @

83

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

84 HUNTER and BULL

associated atherosclerotic risk factors in these patients. Reoperative mortality rates of up to
5% and major limb loss rates of 20% have been reported (11). As a consequence, patients
in whom complications related to their grafts develop may be unable to withstand the
secondary operations needed to replace them. These changes in the demographics of the
patients requiring repeated arterial reconstructive procedures and the limited supply of
autogenous graft material have increased the need for durable arterial substitutes that will
continue to function satisfactorily throughout the patient's remaining life.

In this chapter we discuss the fabrication, healing characteristics, clinical indications,
and complications related to grafts fabricated from polyethylene terephthalate (Dacron),
polytetrafluoroefhylene (PTFE), polyurethane, and the commonly used biological grafts —
human umbilical vein grafts, cryopreserved allografts, and bovine xenografts.

I. DACRON GRAFTS

Dacron is the most common prosthetic material used to replace diseased segments of the
aorta and its major branches. Although there are numerous variations in construction, all
Dacron grafts can be grouped into several major categories: woven versus knitted and
smooth versus veloured surfaces. Woven grafts have functioned satisfactorily when used
for repair of thoracic and abdominal aortic aneurysms, aortic replacement in patients with
known bleeding diatheses, or in the occasional patient who bleeds through the interstices
of a preclotted knitted graft immediately after restoration of flow. Knitted grafts are used
to bypass occlusive lesions of the aorta, and the iliac, common femoral, and superficial
femoral arteries. They have also been used successfully for axillofemoral or femorofemoral
bypass grafting.

Long-term patency rates of 85-90% for aortoiliac or aortofemoral bypass grafts and
5-year primary patency rates of 78-88% for axillofemoral, 71% for above-knee, and 57%
for below-knee femoropopliteal externally supported Dacron grafts have been reported
(3-5, 7, 12-14). The recent study by Robb et al. has demonstrated greater long-term
patency of aortofemoral grafts if both the superficial femoral and profunda femoris
arteries are patent (15).

A. Fabrication

A knowledge of the fabrication of Dacron grafts is essential for an understanding of the
factors that influence healing and/or contribute to the late failure of these grafts.

1 . Woven

Woven grafts constructed by interlacing two sets of yarn at right angles to each other
are the strongest prosthetic fabric grafts. About 45% of grafts implanted each year fall

into this category (1,2). Nonveloured woven grafts are dimensionally stable, relatively g

impervious to blood, noncompliant, and of high tensile strength. However, they are |

difficult to suture and tend to fray at their cut edges. The velour component, which can be s

added to woven grafts by incorporating additional nontextured yarns that are interlaced a

less frequently than the textured ground yarns, permits a reduction in tightness of the u

weave without altering permeability. This results in a softer graft that is easier to suture. y

2. Knitted a
Knitted grafts are of two basic types and are constructed with a set of yarns that are §

interlooped rather than interlaced. Weft-knitted grafts are formed from one set of yarns ©

interlocked in a circular fashion; warp-knitted grafts are fabricated from several sets of li

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 85

yarns interlooped in a zigzag pattern. Warp-knitted grafts, although more compliant than
woven grafts, are less compliant than weft-knitted grafts. Unlike woven grafts they do not
run, unravel, or fray at their cut edges.

Even though the walls of knitted grafts are thicker than woven (600 vs. 200 urn), they
are more porous because of their construction and as a result readily permit cellular
ingrowth through their interstices (2). The interlooping yarns used in the construction of
knitted grafts permit greater expansion of the yarn circumferentially than longitudinally;
consequently they have reduced dimensional stability and a tendency to dilate (1).

Unlike woven grafts, the greater porosity of knitted grafts requires that they be made
impervious to blood before implantation. This can be achieved either (a) by preclotting the
graft with blood (2) or by coating the luminal surface with proteinaceous sealants, such as
bovine collagen (16,17), gelatin (18,19), or albumin (18-20), or (b) by autoclaving the graft
in albumin or blood [as advocated by Bethea and Reemsta (21) and modified by Cooley et
al. (22)]. When first introduced, coated grafts were stiffer than uncoated prostheses,
making suturing difficult. In addition, the sealants were not uniformly applied, resulting in
unpredictable porosity, cracking, peeling, and embolization of the coated material (19). A
critical requirement for all coated prostheses is that the rate of absorption of the sealant
occur within a precise time frame so that it provides hemostasis without impairing healing.
The use of sealant-coated grafts has proved to be particularly advantageous when
replacing portions of the thoracic aorta or when repairing ruptured aortic aneurysms, in
which intraoperative interstitial and anastomotic bleeding are important causes of
morbidity and mortality.

Knitted (Gelseal) and woven Dacron (Gelweave) and, more recently, PTFE (seal
PTFE) grafts have all been impregnated with gelatin. The Hemashield graft is a warp-
knitted, double-velour graft impregnated with type 1 collagen. Although these grafts can
readily be implanted without bleeding in the majority of patients, significant needle-hole
bleeding has been observed in some patients having these grafts placed while on
cardiopulmonary bypass (23).

B. Healing Characteristics

The body attempts to incorporate the implanted Dacron graft by two distinct cellular
processes: (a) anastomotic pannus ingrowth, which extends approximately 1-2 cm from the
divided artery and is composed of smooth muscle and endothelial cells derived from
components of the arterial wall, and (b) perigraft fibrous tissue, which encircles the entire
external surface of the graft, resulting in encapsulation of the graft. Anastomotic pannus
ingrowth, however, does not possess any significant intrinsic tensile strength; hence the
anastomotic bond between graft and artery is entirely dependent on the suture material for
its integrity. Cells from the outer fibrous capsule readily invade the interstices of the graft ■a

but seldom penetrate beyond the midportion of the wall of the prosthesis (24,25). The |

addition of a veloured component to these grafts to enhance cellular ingrowth may induce as

excessive proliferation of perivascular fibrous tissue, which decreases the compliance of the c

graft or contributes to stenosis of an adjacent structure. <

The flow surface of a well-healed patent graft is characterized by a layer of compacted, >9

relatively acellular hypothrombogenic pseudointima, which is uniformly present except in J

areas adjacent to anastomoses or at bifurcations. If initially a prosthesis of too large a «

diameter has been used or if the diameter of the prosthesis has increased substantially over |

time, the thickness of this layer tends to increase, so that the lumen of the graft @

approximates in size that of the outflow vessel. The addition of an internal velour lining %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

86 HUNTER and BULL

may further increase the thickness of this layer and contribute to the ultimate failure of the
graft. Poor healing characterized by incomplete covering of the luminal surface has been
observed in explants from diabetic patients and those in poor general health (10).

Guidoin et al. (25) noted that the differences in the healing characteristics among the
various types of Dacron prostheses could be related solely to the varying thicknesses of the
internal fibrous pseudointimal layer and that of the external capsule. The thicknesses of
the inner and external linings of the grafts were minimal with woven grafts, moderate in
knitted grafts, and greatest for grafts with veloured surfaces (25). We have observed more
perigraft tissue ingrowth and greater thrombus adherence in coated grafts than in
uncoated prostheses. Reoperation for limb occlusion is technically difficult because of
the dense external capsule surrounding the graft. Removal of thrombus with thrombec-
tomy catheters or endarterectomy strippers is difficult and often incomplete. Catheter-
directed lytic therapy for graft limb occlusion has also been less successful in these patients
than in those with uncoated Dacron grafts. The use of albumin-coated grafts has been
associated with inconsistencies in graft porosity, cellular infiltration, and dissection or
embolism of the albumin layer and has largely been abandoned (19).

In a histological evaluation of collagen-coated graft explants in humans, Anderson (26)
has observed that the collagen component remains intact for up to 8 days after implantation.
Grafts removed beyond 1 year were characterized by a cellular pseudointimal layer with
numerous capillary loops and periadventitial infiltration of the graft. Experimental and
clinical studies indicate that the gelatin sealant on gelatin-coated grafts is resorbed within 1—
2 weeks (27). During the first week after implantation, fibrous thrombi and polymorph
neutrophils are present on the flow surface. By 2 weeks, most of the gelatin has been re-
absorbed, accompanied by a mild inflammatory response and return of the prostacyclin/
thromboxane ratio to 1, indicating that the healing process is well advanced (27,28).

Although direct comparisons between collagen-coated and gelatin-impregnated grafts
have shown no differences in patency, experimental studies suggest that gelatin-sealed
prostheses may be more resistant to infection with Staphylococcus epidermidis than un-
coated knitted velour polyester (29). In a prospective comparison between collagen-
coated, gelatin-impregnated, and stretch PTFE aortofemoral grafts, Prager et al. found
no statistical differences in patency (30). When the two Dacron grafts were compared
collectively with PTFE, they had a higher infection rate (3 vs. 0%) than the PTFE grafts.
The potential for both collagen-coated and gelatin-impregnated grafts to stimulate an
immune response after implantation appears low (31).

C. Complications

To date there have been no systematic follow-up studies of all Dacron grafts in otherwise
asymptomatic patients; most of the reports in the literature are confined to anecdotal case

studies of complications. The noninfectious complications associated with the implantation §

of these prostheses include (1) dilatation, (2) para-anastomotic aneurysms, and pseudo- |

aneurysms (3) anastomotic stenosis, (4) ureteric obstruction, and (5) neoplasia. Some 90% 3

of such failures occur within the first 3-5 years after implantation. a

1 . Dilatation s

Dilatation, defined as a permanent increase in the diameter of a graft caused by ^

pulsating stresses, has been reported with both arterial homografts and prosthetic grafts °

used as arterial conduits (1,32-42). Dilatation may involve the entire length of the graft or |

be confined to isolated portions, resulting in diffuse or focal aneurysmal change. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES

87

Significant dilatation of implanted grafts has been documented with all currently available
prosthetic materials, including knitted and woven Dacron, nylon, Teflon, Orion, and
PTFE (37-42). However, dilatation is more commonly observed with knitted Dacron
because of its inherent structural properties and its more frequent use as an arterial
substitute (Fig. 1). Woven grafts have a high initial modulus and therefore are resistant to
extension compared with knitted grafts, which have very little resistance to extension
because of their loop structure. Consequently, their interlocking loops straighten very
easily in the direction of greatest stress, resulting in an increase in diameter. The incidence
of graft dilatation reported in the literature, ranging from 1-3%, probably underestimates
the true incidence, since only symptomatic patients are likely to undergo imaging (43,44).
In a 30-year review of 390 cases of graft failure, Pourdeyhimi and Wagner (1) found
dilatation in 147 (38%), structural defects in 76 (20%), suture line defects in 56 (14%), and
graft infection or bleeding in 30 (8.9%). These complications are often interrelated because

I
•5

Figure 1 Abdominal aortogram showing dilation of an aortoiliac knitted graft with a right
common iliac aneurysm.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

88 HUNTER and BULL

graft dilatation may predispose the patient to pseudoaneurysm formation, perforation, or
rupture.

Using Doppler ultrasound, Nunn and colleagues (45) studied 95 Dacron aortic grafts
implanted for periods ranging from 2 weeks to 138 months in symptomatic patients. They
observed dilatation ranging from 0-84% (mean, 17.6%) in 85 of the 95 grafts. Dilatation
appeared more pronounced in hypertensive patients (21 vs. 15%). Lundqvist et al. (46), in
a study comparing 36 patients with symptomatic graft dilatation and 65 asymptomatic
patients, detected graft dilatation of between 25 and 50% in 42 of 101 patients (42%)
evaluated. In 12 patients, the diameter of the dilated graft exceeded the preimplantation
diameter by more than 50%. Four of these patients had false aneurysms. Furthermore, the
incidence of dilatation was greater in symptomatic patients, who also had a higher
incidence of false aneurysm formation compared with asymptomatic patients (14 vs.
3%). Berman et al. (47) evaluated 178 aortic grafts implanted for 0.2-233 months (mean
43.3 ± 3.2 months) with computed tomography (CT). A total of 143 Dacron prostheses
(74 woven, 69 knitted) and 35 PTFE prostheses were evaluated. The mean percentage
dilatation was 49.2 ± 4 for knitted prostheses, 28.5 ± 3.0 for woven prostheses, and 20.6 ±
1.9 for PTFE. The authors observed a significant correlation between graft dilatation of
>50% and knitted grafts.

Etiology. Permanent deformation of the fabric structure and fatigue of the yarn
appear to be important factors in the pathogenesis of graft dilatation. Dilatation of 10-
22% occurs when knitted grafts are bench tested at static pressures of 120-200 mmHg for
1 min (1). This corresponds to the initial dilatation seen when these grafts are exposed to
arterial pressure. Whether this increase in diameter reflects restoration of the initial
diameter (reduced by chemicals and heat used during the crimping process) or simply
represents straightening of the interlacing loop structure is unclear. Although bench
testing is a fairly reliable predictor of early dilatation, it does not accurately predict late
dilatation related to fiber elongation or yarn fatigue.

A number of factors are believed to contribute to deformation of fabric structure and
yarn fatigue of Dacron grafts. They include hypertension (45), mechanical or chemical
degradation of Dacron fibers (48-51), undetected flaws from the manufacturing process,
performance deformation or creep of the fabric due to loss of stitch density (knitted) or
fabric count (woven) (44,52), and damage inflicted by the injudicious application of
instruments intraoperatively. The contribution of any one of these factors to failure of a
given graft often is difficult to ascertain in the individual patient. Microscopic fracture of
the fibers with fibrillation (50) is typical of mechanical failure due to cyclical bending or
tension or torsion stresses on the graft.

In a comprehensive review of Dacron graft explants removed at autopsy or reopera-
tion, the most consistent finding observed by Guidoin (10) was a time-dependent loss of
stitch density found in all prostheses examined. The rate of loss of stitch density depends 13

on the type of fabric construction. Weft-knitted fabrics therefore would be less stable than |

the warp-knitted ones. As a result of this loss of stitch density, the fabric is susceptible to j§

irreversible dilatation when exposed to arterial blood pressure. Clagett et al. (52) observed £

a similar loss in stitch density in five patients with dilated grafts. Of considerable interest is ^

the size of the grafts implanted in the patients Clagett et al. (52) studied. The diameters of "jl

the grafts used were 19 x 9.5 mm (four patients) and 25 x 12.5 mm (one patient), s

considerably larger than would be chosen today. One can only speculate as to the 2

contribution of initial graft size to eventual failure. J

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES

89

Chemical biodegradation may further exacerbate the dilatation initially induced by a
loss of stitch density. King et al. (50), in a study of 19 prosthetic Dacron explants in
residence for intervals ranging from a few hours to 14 years using infrared spectroscopy,
observed a decrease in molecular weight and an increase in the content of carboxyl groups
proportional to the duration of implantation. It has been postulated that immunologically
active macrophages and monocytes that infiltrate the graft as part of the healing process
may also contribute to chemical deterioration of these grafts by recruitment of biologically
active cells that ingest the individual Dacron fibers (52) (Fig. 2).

Variations in the fabrication of grafts also have been implicated in cases of late fiber
deterioration and graft dilatation. In a study of 493 grafts, Berger and Sauvage (44) found
that graft dilatation between 17 and 43% was present in 15 cases that had been implanted
for 3-15.3 years. The authors postulated that reductions in fiber diameter and/or acute

I
•5

Figure 2 Photomicrograph of the outer surface of a Dacron graft covered with a layer of firmly
adherent connective tissue. Foreign-body giant cells and macrophages surrounding and ingesting the
Dacron fibers are present. (Original magnification x 200.)

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

90

HUNTER and BULL

breaks of individual filaments resulting from overheating during the crimping process,
frictional wear, or biodegradation were responsible for these changes.

Finally, damage to the graft can occur during implantation from careless handling
with instruments. Isolated areas of damaged fibers suggestive of vascular clamp injury
have been observed in graft explants (10). Thus careful handling during implantation is
essential to avoid damage that may increase later susceptibility to deterioration (44).

Diagnosis. Dilatation of Dacron grafts appears to be a biphasic phenomenon. Early
dilatation of approximately 10-22%, commencing immediately after the graft is exposed to
arterial blood pressure, plateaus within the first year (Fig. 3) (1). Late dilatation, caused by
yarn slippage or breakage, usually is seen within 2-3 years of implantation. Because of the
biphasic nature of graft dilatation, evaluation of abnormal dilatation appears unwarranted
within the first year after implantation of the graft, since this early dilatation is largely
attributed to restoration of the initial diameter of the graft. Late graft dilatation, however,
is progressive. Once dilatation is detected, such patients should be carefully monitored for
life.

Ultrasound. B-mode ultrasound is widely available, relatively inexpensive, and
particularly useful for assessing early and late complications of vascular grafts. Gooding
and coworkers (53), in a study of 87 patients with aortofemoral grafts, were able to suc-
cessfully image the anastomotic sites in 84% of the grafts. The authors discovered 23 un-
suspected anastomotic aneurysms (2 aortic and 21 femoral), and perigraft accumulation of
fluid or blood was observed in 16 patients. In a study of 127 patients with Dacron grafts,
Clifford et al. (54), using Doppler imaging and real-time ultrasound, found graft dila-
tation of 15-70% in five grafts. More recently we have studied patients with dilated grafts
using color-flow Doppler imaging and have observed nonlaminar turbulent flow patterns

LEJ
</}

+1

C
(S
d>

80

60

.1 40

Q

c 20

Q.

□ PTFE
H Knitted
E3 Woven

n=7

<1 Month <1 2 Months >12 Months
Implant Time

Figure 3 Histogram showing time course of aortic graft dilatation. Knitted Dacron compared with
PTFE and woven Dacron. * = p < 0.01 by Anova.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 91

within the grafts reminiscent of those seen in patients with aortic aneurysms. Although this
may explain the progressive increase in the pseudointimal lining in these grafts, to date
relatively few patients have been studied and no definitive conclusions can be drawn.

Ideally, color-flow Doppler imaging of all asymptomatic aortofemoral grafts should
be performed at 1 year and then at 5 years. If no significant alteration in the diameter of
the prosthesis is identified at 5 years after implantation, it seems unlikely that significant
dilatation will occur and the frequency of follow-up can be reduced. However, once
dilatation has been detected, continuing surveillance is indicated to detect anastomotic
aneurysms or an increase in luminal pseudointima, either of which may threaten the long-
term patency of the graft.

Computed Tomography. Contrast-enhanced computed tomography (CT) is helpful in
accurately assessing the size of anastomotic aneurysms and in excluding other complica-
tions such as perigraft collections of fluid or air. Unlike ultrasound, CT can be used to
evaluate grafts implanted in the abdomen or chest (Fig. 4) (26). Qvarfordt et al. (55), Brown
et al. (56), Berman et al. (57), Nunn (45,58), and Kalman (59) have described the useful-
ness of CT scans for detecting early and late complications of prosthetic grafts. CT scanning
of aortic grafts has demonstrated dilatation up to 367% of the aortic portion of knitted
prostheses. Dilatation can occur in the body and each of the limbs of an aortic bifurcated
graft but may not occur uniformly throughout the graft (Fig. 5). It can also occur in Dacron
grafts implanted in other locations such as femoropopliteal bypasses (Fig. 6). Berman et al.
(47) reported a complication rate of 13.5%. Kalman et al. found aneurysmal dilatation of
the abdominal or thoracic aorta in 13.8% of patients and dilatation of the iliac arteries in
patients with tube grafts in 15.4% (59). However, cost and attendant radiation exposure
make CT less desirable for routine follow-up and CT should be limited to the evaluation of
symptomatic patients or for those instances in which ultrasound examination fails to ade-
quately delineate the graft.

Complications. Progressive increases in diameter of prosthetic grafts have been
implicated in anastomotic aneurysm formation, thrombosis, and rupture of the affected
grafts (46,51,56). Embolization from a dilated graft containing laminated luminal throm-
bus also must be considered a potential risk; however, the frequency of this event is a
matter of speculation since its occurrence has not been well documented in the literature.

2. Para-Anastomotic Aneurysms and Pseudoaneurysms

Aortic Para-Anastomotic Aneurysms. Para-anastomotic aneurysms occur in 0.5-15%
of patients undergoing aortic operations for occlusive or aneurysmal disease and tend to
increase in frequency over time (60-65). The incidence of this complication is probably
underestimated, as systematic long-term follow-up of aortic prostheses is not currently

undertaken. ■§

These aneurysms can occur at either the aortic, iliac, or femoral anastomosis of bi- g

furcated grafts. They can be either true aneurysms or pseudoaneurysms and occur with »

varying frequency depending on the indication for aortic replacement. True aneurysms of c

the parental aorta occur almost entirely in patients undergoing aortic replacement for aneu- <

rysmal disease; whereas, progressive dilation of the iliac arteries is usually seen in patients >9

undergoing aortic tube graft replacement. Unrecognized or uncorrected preexisting whereas ^

progressive dilatation of aortic dilatation, ongoing matrix degradation as part of the an- «

eurysmal process, and the use of a prosthesis of too large a diameter are among the factors |

contributing to the development of true para-anastomotic aneurysms (Figs. 7 and 8). @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

92

HUNTER and BULL

Figure 4 Abdominal CT scans demonstrating (A) a dilated graft (arrow) and the surrounding
thrombus from an aortic anastomotic aneurysm and (B) dilatation of the aorta at the site of an end-
to-side anastomosis.

In contrast, para-anastomotic pseudoaneurysms are more common following aorto-
femoral bypass grafting for occlusive disease and occurs at either the proximal or distal
anastomosis with either end-to-end or end-to-side anastomotic configurations. Endar-
terectomy of the aortic cuff, suture failure, graft type, defects in the aortic wall, and a
mismatch in compliance all contribute to ultimate deterioration of the proximal aortic
anastomosis.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES

93

Figure 5 Abdominal arteriogram showing (A) focal dilatations in the limbs of an aortobifemoral
graft and (B) operative picture of the focal areas of dilatation seen in A.

The majority of patients with aortic para-anastomotic aneurysms are asymptomatic
and are detected incidentally on abdominal ultrasound or CT scanning. A herald bleed or
frank rupture in a patient with an aortic prosthesis should alert the surgeon to the
possibility of this complication.

The diagnosis is usually confirmed by CT scanning. Careful preoperative evaluation of
these patients is essential because of the higher morbidity and mortality rates associated
with these reoperative procedures. The size of the aneurysm and the general condition of
the patient are among the factors that need to be considered.

The retroperitoneal approach, if not previously used, offers an excellent method of
access for repair of these complex lesions. Extension of the graft more proximally, with or
without reimplantation of the visceral vessels, is usually required for true aneurysms.

Although endovascular techniques may be applicable for repair of pseudoaneurysms,
the small but definite risk of an infectious etiology remains when para-anastomotic
aneurysms are repaired surgically. Tissue excised at surgery should be sent for culture
and histological examination for the presence of microorganisms.

The operative mortality rates for repair of para-anastomotic aneurysms are high,
ranging from 20 to 24% for elective procedures and up to 73% for ruptured aneruysms
(62,65).

Whether para-anastomotic aneurysms can be prevented is a matter for debate.
However, a few technical precautions at the initial procedure seem prudent. Endarter-
ectomy of the proximal aortic cuff should be avoided and, if necessary, the aorta should be
reinforced with Teflon pledgets or placement of a prosthetic cuff overlapping the suture
line and the endarterectomized aorta.

Femoral Anastomotic Aneurysms and Pseudoaneurysms. Anastomotic pseudoaneu-
rysms involving the femoral anastomosis account for more than 80% of cases. A number

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

94

HUNTER and BULL

Figure 6 Femoral arteriogram showing diffuse enlargement and focal dilatation (arrow) of a
Dacron femoropopliteal graft.

of etiological factors — including weakness of the arterial wall (31%), hypertension (27%),
mechanical factors (12%), graft deterioration (12%), impaired wound healing (8%),
endarterectomy (7%), and suture failure (3%) — have been implicated in a comprehensive
study by Szilagyi et al. (66). Occasionally, there is true dilatation of the common femoral
artery at a femoral anastomosis. Here we focus on the contribution of graft dilatation to
the development of anastomotic pseudoaneurysms (Fig. 9).

How may dilatation of the prosthesis contribute to the development of anastomotic
pseudoaneurysms? Since knitted grafts dilate initially by 10-22%, selection of too large a
graft may be an important contributing factor. The diameter of the limbs of the graft
should approximate that of the outflow vessel. A graft-to-vessel ratio of 1.2—1.4:1, as
suggested by the experimental studies of Kinley et al. (67), may be too generous in view of
the inherent propensity of knitted grafts to dilate.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES

95

Figure 7 Abdominal aortogram demonstrating a para-anastomotic aneurysm in a patient who had
previously undergone repair of an abdominal aortic aneurysm. The abdominal CT scan in the upper
panel shows the aneurysm. The PTFE graft used to repair the infrarenal aneurysm is noted in the
lower panel (arrow).

The portion of the graft close to the suture line may fail from yarn slippage or
breakage. Yarn slippage is common with woven grafts, especially if they are cut on a bias,
as is commonly done in performing end-to-side anastomosis. Careful placement of su-
tures and heat sealing of the cut edges may minimize slippage of the yarn. Yarn breakage
was associated more commonly with weft-knitted grafts, which are no longer being
manufactured.

Dilatation and anastomotic aneurysms often occur in association (52,68). Kim et al.
(68) observed significant graft dilatation ranging from 50 to 150% in the patients they
studied. Anastomotic aneurysms were present in all instances. Perhaps suture line stress,
minimal when the graft-to-artery ratio is 1.4:1 or less, increases progressively as the graft
enlarges. Also, as the graft increases in diameter, it has a tendency to shorten in length,
thereby increasing suture line tension even further. Courbier and Aboukhater (69) recently
have hypothesized that scarring, caused by the graft exiting beneath the inguinal ligament,
may be an additional contributing factor. They recommend prophylactic division of the
inguinal ligament. We do not concur with this theory. Although the majority of grafts
adhere to the inguinal ligament, most do not exhibit false aneurysm formation. Further-
more, routine division of the inguinal ligament substantially increases the incidence of
subsequent inguinal or femoral hernias.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

96

HUNTER and BULL

Figure 8 Transaxillary aortogram of a dilated Dacron graft with an aortic anastomotic aneurysm.

In a series of 42 anastomotic aneurysms, Carson et al. (70) found a significant increase
in graft-to-artery ratio in 22 of 42 patients (52%), with a mean increase in diameter of
22%. Dilatation and an increase in graft-to-artery ratio (1.3—1.6:1 vs. 1—1.3:1) in patients
were significant etiological factors in the recurrence of anastomotic pseudoaneurysms.

Although pseudoaneurysms are subject to the same spectrum of complications as true
aneurysms, occlusion of a limb of the graft, especially at the femoral anastomosis, occurs
more often than rupture.

Thrombosis. As the diameter of the graft increases, the fibrous, relatively acellular
pseudointimal layer lining the luminal surface of the graft likewise increases progressively,
resulting in the simultaneous diminution of the diameter of the lumen to approximate that
of the outflow vessel. Dislodgment of this material by trauma or during arteriography or
the imposition of distal obstruction by progressive atherosclerosis or anastomotic intimal
hyperplasia results in occlusion of the limb (Fig. 10). The management of limb occlusion

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I
•5

I

HEALING OF PROSTHESES

97

Figure 9 Abdominal CT scan showing large bilateral femoral anastomotic pseudoaneurysms. The
outline of the dilated graft limb is indicated by the arrow.

secondary to an anastomotic pseudoaneurysm includes thrombectomy and/or the replace-
ment of a segment of the graft.

Rupture. The final strength of a textile prosthesis is determined by a number of
factors including the inherent properties of the basic polymer, the structure and number of
filaments as well as the yarn and fabric structure. Grafts manufactured prior to 1981,
made with T-62 yarns with trilobar filaments, were significantly weaker than the current
fabric grafts made with nontexturized T-56 yarns with cylindrical filaments (71). The
guideline, composed of carbon particles, is added to the T-56 yarns during multispinning
to both knitted and woven grafts, allowing proper alignment of the graft during
implantation (72). The chemical reactions required to insert the guideline results in
weakening of the prosthesis and produces a potential site for rupture. Construction of
earlier knitted double-velour grafts utilized alternating T-56 yarns twisted in both S and Z
configurations. The remeshing line is formed with two simultaneous knitted bands joined
together to form the tubular structure of the graft. Both the guidelines and the remeshing
lines are potential sites of weakness in Dacron grafts and frequently the sites of rupture. A
recent study of 20 human explants by Chakfe et al. (72) demonstrated that ruptures
occurred most often in areas of weakness within the prosthesis guideline (n = 6), remeshing
line («= 11), or both (w = 3). Scanning electron microscopy showed major fractures of the
tubular filaments, with complete disappearance of the velour in many instances. Ruptures
of the prosthesis caused by focal defects within the graft or occasionally disruption of the
entire prosthesis are rare but devastating complications. Focal rupture caused by fracture
and fragmentation of Dacron fibers or longitudinal tears in grafts manufactured with
inadequate tensile strength may result in single or multiple areas of false aneurysm for-

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

98

HUNTER and BULL

Figure 10 Operative arteriogram demonstrating a dilated Dacron graft limb with irregular pseu-
dointimal tissue lining the graft (arrows). The nondilated PTFE graft used to repair the anastomotic
aneurysm is well demonstrated. A stent to decompress the associated ureteral obstruction is present.

mation. These pseudoaneurysms may rupture into a hollow viscus, the retroperitoneum,
or freely into the peritoneal cavity (51,73,74). Rupture of an anastomotic pseudoaneu-
rysm, although uncommon in the absence of infection, is a notable cause of morbidity and
mortality in patients with prosthetic grafts.

Aortic or iliac anastomotic pseudoaneurysms may rupture into adjacent bowel, most
frequently the duodenum. However, other portions of the bowel may be involved also,
depending on what portion of the bowel happens to be adjacent to the anastomosis. More
rarely, these lesions rupture into the peritoneal cavity or the retroperitoneum.

Rarely do femoral anastomotic aneurysms rupture and usually only in patients with
very large, long neglected aneurysms.

Management of Graft Dilatation. Asymptomatic dilatation of prosthetic grafts with-
out anastomotic pseudoaneurysm formation seldom is detected clinically unless a com-
plication supervenes. Most often the patient presents with a painless groin swelling (the

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I
•5

I

HEALING OF PROSTHESES 99

anastomotic aneurysm) and dilatation of the graft is detected during the ensuing workup.
Acute or chronic occlusion of one or both of the limbs of an aortobifemoral graft caused
by thrombosis of an anastomotic pseudoaneurysm is a less common mode of presentation.

Patients with significant dilatation of their grafts require careful evaluation before
operative intervention is undertaken. Significant risk factors increasing operative morbid-
ity and mortality usually are present and reoperation should be tailored to the individual
patient. In addition, careful selection of the prosthesis to be used for the repair is essential
if additional complications are to be minimized.

The treatment of graft dilatation may entail (a) local repair of anastomotic aneurysms,
(b) thrombectomy and profundaplasty in patients with limb occlusion, or (c) graft
replacement if the entire graft is dilated more than 50% of its original diameter or if an
aortic anastomotic or para-anastomotic aneurysm is present.

A few salient points must be made about the operative techniques for repair of these
aneurysms:

1. When an anastomotic aneurysm or pseudoaneurysm associated with graft dila-
tation is being repaired, a prosthesis that approximates the diameter of the outflow
tract, not the dilated proximal graft, must be selected. We prefer to use PTFE in the
repair of noninfected anastomotic femoral aneurysms, since this material as cur-
rently manufactured does not dilate. Use of another Dacron graft approximating
the diameter of the dilated primary graft remains a common practice but may
further aggravate the problem, since the new segment of graft likewise is prone to
dilatation (Fig. 11).

2. When dilated aortic grafts are being replaced, small segments of the old graft
adjacent to the proximal aortic anastomosis may be left in place. This allows
replacement of the graft without extensive mobilization of aorta, which is often
encased in dense fibrous tissue.

3. The risk of injury to distal vessels may be minimized by using balloon catheter
occlusion following isolation of the inflow vessel.

It must be remembered that graft dilatation is a problem inherent with knitted Dacron
grafts. Although the tendency to yarn slippage can be minimized by increasing the number
of yarns used during manufacture, this is limited by the need to preserve the ability to
suture these grafts during implantation.

3. Anastomotic Stenosis

Stenosis caused by neointimal hyperplasia or progression of atherosclerosis occurs with
Dacron aortofemoral grafts but with lesser frequency than with PTFE grafts. Progression of
atherosclerosis in the profunda femoris or superficial femoral arteries and neointimal

hyperplasia are the most frequent causes of graft anastomotic stenosis (75) (Figs. 12 and ■§

13). Stenosis occurring at the proximal aortic anastomosis is usually a result of failure to g

remove residual atherosclerotic plaque, thrombus, or progression of disease. »

These patients usually present with symptoms of acute or chronic limb ischemia c

caused by occlusion of one or both limbs of the graft. Arteriography delineates the cause <

of the obstruction and guides direct repair. >5

Successful management of stenosis or occlusion at the distal anastomosis can be %

achieved in approximately 90% of patients (75). Surgical repair usually includes throm- «

bectomy of the graft plus endarterectomy of the profunda femoris artery with patch |

angioplasty (using autogenous vein, endarterectomized segments of the occluded super- @

ficial femoral artery, or prosthetic material) (54). Occasionally proximal stenosis may %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

100

HUNTER and BULL

Figure 11 Transaxillary aortogram demonstrating occlusion of the left limb of a dilated aorto-
bifemoral graft. The Dacron graft used to repair the right femoral anastomotic aneurysm has
subsequently also dilated (arrow).

require replacement of the graft after endarterectomy of the severely narrowed aortic
segment. In poor-risk patients with progressively increasing proximal aortic stenosis,
placement of a covered stent graft, or axillofemoral bypass grafting may be indicated.

4. Ureteric Obstruction

Ureteric obstruction is a well-recognized complication of aortic reconstructive proce-
dures (Fig. 14). The etiology of hydronephrosis includes ureteric ischemia, kinks, operative
trauma, anastomotic aneurysms, graft infection, graft limb thrombosis, and an incidental

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES

101

Figure 12 Abdominal aortogram showing a high-grade stenosis at the anastomosis of the Dacron
graft at its junction with the profunda femoris artery.

ureteric tumor. In approximately 1 % of patients, dense fibrosis, presumably associated with
incorporation of the graft, also encases the ureter and results in hydronephrosis (76,77).

Ureteric obstruction may be an incidental finding on workup for some other condition
or may present as obstructive uropathy or progressive deterioration in renal function.
Evaluation should include measures of renal function and delineation of the site and cause
of obstruction by intravenous pyelography, ultrasound, retrograde pyelography, or CT
scanning. The position of the ureter relative to the graft is best ascertained by the use of
contrast-enhanced CT scanning.

Ureteric obstruction caused by encasement of the ureter by perigraft fibrosis can be
avoided by careful positioning of the limbs of the graft posterior to the ureter when
performing an aortofemoral bypass. The management of ureteric obstruction depends on

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

102

HUNTER and BULL

Figure 13 Operative specimen from a patient with anastomotic stenosis due to neointimal
thickening.

the physical condition of the patient and the site and cause of the obstruction. Interven-
tional and surgical options include placement of indwelling ureteral stents, division, re-
routing and reanastomosis of prosthetic graft limbs if anterior to the ureter, or excision
and reanastomosis of the ureter for severe segmental fibrotic stenosis. Infection of the un-
derlying graft remains an ongoing concern with catheter intervention or surgical treatment
of ureteric obstruction in these patients.

5. Neoplasia

There are now a few case reports of angiosarcoma developing in patients with Dacron
grafts (78-81). The possible contribution of Dacron grafts to the development of angio-
sarcoma is unclear. Although some experimental evidence links plastic materials with
neoplastic change, the evidence in humans is less convincing (81). First, the incidence is
extremely low when one compares the number of tumors reported with the number of
prosthetic grafts implanted. Second, in the case reported by O'Connell and coworkers (79),
the tumor was almost certainly incidental to the Dacron graft in view of the short time
interval that elapsed between placement of the graft and diagnosis of the angiosarcoma.
However, Fehrenbacher et al. (80) reported an angiosarcoma thought to be related to the
use of a Dacron graft implanted 12 years previously. Although the time interval between
implantation of the graft and the development of the neoplasm is consistent with a

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I
•5

I

HEALING OF PROSTHESES

103

Figure 14 Right retrograde pyelogram shows hydroureteronephrosis with obstruction by the graft
limb at the pelvic brim.

possible cause-and-effect relationship, this remains speculative. Nonetheless, although a
direct cause-and-effect relationship between the use of Dacron grafts and the development
of angiosarcoma cannot be established, continued vigilance seems appropriate.

The diagnosis of angiosarcoma is seldom made antemortem. Progressive stenosis
resulting in obstructive or embolic symptoms is the usual clinical presentation. Contrast-
enhanced abdominal CT scanning or biplanar arteriography usually will demonstrate an
intraluminal filling defect. Excision/grafting and endarterectomy have been advocated for
treatment of these lesions. The prognosis is poor; subsequent patient survival usually is
only a few months in duration.

I
•5

II. EXPANDED PTFE GRAFTS

Expanded polytetrafluoroethylene (PTFE) grafts are widely used as arterial substitutes for
aortoiliofemoral, axillofemoral, and femorofemoral bypasses. They are also the preferred
conduit for carotid subclavian and visceral artery bypasses, dialysis access procedures, and

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

104 HUNTER and BULL

for infrainguinal bypass grafting procedures to the popliteal and tibial vessels when
autogenous vein is unavailable. PTFE grafts do not dilate significantly after implantation
and thrombus is readily removed with a thrombectomy catheter or lysed by thrombolytic
agents. Patency rates at 4-5 years for aortobifemoral grafts range from 91 to 95% (30,82).
When used for dialysis access, PTFE grafts have 1-year patency rates of 17-79% (83,84).
Primary patency rates for above- and below-knee femoropopliteal bypass grafts range
from 39.7 to 61% and 27 to 39%, respectively, between 5 and 8 years (5,7-9). Landry (85)
recently has reported a 5-year patency rate of 71% for ringed PTFE grafts placed as
axillofemoral bypasses.

A. Fabrication

Manufactured by mechanical extrusion of the chemically inert carbon fluorine-PTFE, the
resulting grafts consist of solid nodes of PTFE interconnected by longitudinally oriented
fibrils. The solid-node fibril structure of PTFE comprises only 15 to 20% of the volume of
the graft material; the remaining void is filled with air (Fig. 15). Thus PTFE is a porous
graft that readily accommodates tissue ingrowth (86).

Continuous deformation, known as "creep," occurs when polymeric materials such as
PTFE are subjected to arterial pressure. This was an important factor in aneurysmal
dilatation of nonreinforced grafts. Creep occurs in two phases: an initial deformation
caused by laxity of the structure is followed by gradual continuous deformation, tending
to progress with time (86). Presently, the commercially available PTFE prostheses are
made creep-resistant by increasing wall thickness (0.64 vs. 0.5 mm), by increasing the
density of the node fibril structure, or by the application of an external reinforcing sheath
also of PTFE.

These prostheses demonstrate very little propensity to dilate, hold sutures well, and do
not require preclotting; because of their smooth luminal surfaces, thrombus is easily
removed.

B. Healing Characteristics

Although the healing characteristics of PTFE grafts have been studied extensively in
experimental animals, few studies of explanted grafts in humans are available, as only
occluded grafts are usually analyzed. Experimentally, PTFE grafts are readily incorpo-
rated by dense fibrous tissue. The luminal flow surface is covered first by a layer of protein
and cellular elements derived from the blood (86,87).

Three distinct processes can be observed at both the proximal and distal anastomoses
of small-diameter grafts implanted in experimental animals. In animals with patent grafts,
pannus ingrowth originates from the host vessel and extends for approximately 1 cm; neo- ■§

intimal hyperplasia also is often seen. In occluded grafts granulation tissue and thrombus |

are found at the anastomosis. Clowes and associates (88,89) have demonstrated endo- a

thelial cell ingrowth extending approximately 1-1.25 cm from both proximal and distal c

anastomoses of 4-mm PTFE grafts implanted into baboons at 1 and 3 months postoper- <

atively. Although the tensile strength at an anastomotic suture line between Dacron and >9

the host vessel is entirely dependent on the suture material for its integrity, there is some J

evidence that healing may occur at anastomoses constructed with PTFE. Quinones-Bald- °

rich et al. (90), using polyglycolic acid suture to construct artery-graft and artery-artery |

anastomoses in experimental animals, have demonstrated an increase in anastomotic ten- @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES

105

(A)

(B)

Figure 1 5 Scanning electron photomicrograph demonstrating (A) node fibril structure of a stretch
PTFE graft and (B) the luminal surface of the same graft.

sile strength with PTFE grafts and double-veloured woven Dacron grafts compared with
knitted grafts. These data suggest greater anastomotic healing with the former grafts.

Mohring et al. (91) studied explanted PTFE grafts used for dialysis access. They
observed distinct differences in the healing between the two commercially available grafts.
In nonreinforced grafts, connective tissue ingrowth extending into the interstices of the
graft was more pronounced, and these grafts had a noticeably increased cellular internal
lining in contrast to the relatively cell-free neointimal lining of reinforced grafts.

PTFE grafts explanted from humans are characterized by an external fibrous capsule
of varying thickness, which is present in 39% of explants at 1 month. Thickness of the
capsule progressively increases with the length of implantation. Encapsulation, however, is

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

106

HUNTER and BULL

absent in 80% of infected grafts. The interstices of grafts retrieved within the first days
after implantation are infiltrated with red cells and fibrin and with proteinaceous material
at later time periods. Cellular invasion arising from the external capsule does not occur
to any significant degree. Instead, the luminal surface of PTFE grafts is usually covered
with an incomplete thin layer of acellular pseudointimal (Figs. 16 and 17). In a scanning
electron microscopic study of 298 human graft explants, Guidoin observed bacterial colo-
nization, leukocyte infiltration, and lipid deposition of flow surfaces. Bacterial coloniza-
tion was present in 56.3% of all grafts examined and leukocytes in 12.7%. Bacteria were
only seen in 58.4% of grafts excised for infection. Lipid deposition was usually present
in 31.4% of grafts, but cholesterol was detected in only 34.7% of these grafts. In an anal-
ysis of 79 explants using Fourier transform infrared spectroscopy, Guidon et al. found no
evidence of chemical degradation of PTFE grafts implanted for periods up to 6.5 years.
Collagen was detected on the luminal surface of 24 grafts, mainly in the region of the
anastomosis. The authors detected anastomotic intimal hyperplasia in only 4 cases (92,93).

The healing response of PTFE grafts used for arteriovenous grafts includes neo-
intimal thickening at either the arterial or venous anastomosis, granulation tissue ingrowth
along needle puncture tracts, cellular infiltration through the microstructure of the graft,
luminal pseudointima formation, and a foreign-body reaction at the graft-host tissue
interface (94).

Thrombosis is the major cause of failure of dialysis grafts and is most commonly the
result of venous outflow stenosis. Neointimal thickening at the arterial anastomosis,
hypercoagulable states, thrombus formation and needle puncture site, pseudoaneurysms,
and infection all contribute to the propensity to thrombosis of these grafts.

1\l

I

•5

Figure 16 Photomicrograph of the (A) luminal and (B) outer surface of a human PTFE explant.
The luminal surface is covered with a pseudointimal layer with occasional cells visible in the graft
interstices. Foreign-body giant cells (arrow) are present in the tissue along the external wrap (arrow).

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES

107

Figure 17 Photomicrograph of an occluded PTFE graft demonstrating thrombus (horizontal
arrow) containing neutrophils (oblique arrow) on the luminal surface. No bacteria were observed.

C. Enhancement of Small-Diameter PTFE Graft Patency

The major cause of failure of small-diameter prosthetic grafts is the development of
neointimal thickening, most often at the distal anastomosis, resulting in thrombosis of the
graft. There are ongoing attempts to improve the long-term patency of small-diameter
(<6 mm) prosthetic grafts by modifying either their flow surfaces or configuration of the
distal anastomosis. Methods of modifying the luminal flow surface of these grafts include
(a) carbon coating, (b) heparin bonding, and (c) endothelial cell seeding. A venous cuff or
patch and hooded modification of the graft are among the techniques used to modify the
configuration of the distal anastomosis.

1 . Reducing Graft Surface Thrombogenicity

Carbon Coating. While carbon impregnation of 25-30% of the wall of PTFE grafts
(Carboflow, Impra) has many theoretical advantages in preventing surface thrombus
formation, there appears to be no statistically significant benefit over conventional PTFE
with regard to their long-term performance (95). Bacourt reported primary and secondary
rates of 45 and 53% for carbon-impregnated PTFE compared to 35 and 36% for standard
PTFE at 2 years (96).

Heparin Bonding. The development of a thrombus-resistant flow surface to improve
the patency of small diameter prosthetic grafts appears desirable. Because of PTFE's
electronegative surface properties, it has been difficult to bond heparin to it. Heparin can

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

108 HUNTER and BULL

be bonded onto PTFE using covalent bonding with glutaraldehyde or thermal cross-
linking or immobilization onto a Carmeda bioactive surface (97-99). The use of glutar-
aldehyde has been associated with increased cell toxicity. The heparin content of thermally
cross-linked grafts was approximately 0.427 mg/cnr in the report by Iwai et al. (98). The
walls of thermally bonded heparin grafts are hard, but they become soft when the un-
bonded gelatin and heparin are removed. Experimentally, heparin-impregnated grafts
implanted into the carotid arteries of dogs were associated with increased anastomotic
bleeding and a predisposition to perigraft seromas.

Histological examination of heparin-coated grafts at 1 h showed inspissation of
interfibril spaces with gelatin and fibrin. By 7 days, cellular infiltration could be observed
in coated and uncoated grafts. Only a mild inflammatory response was observed on the outer
surface of both heparin-coated and control grafts. One of the theoretical concerns is that the
high concentration of heparin may inhibit tissue ingrowth. The heparin immobilized in the
fabric of vascular prostheses is released slowly and can still be detected at 5 days (98).

In a recent study comparing patency rates of heparin-bonded Dacron (HBD) with
PTFE used to bypass below- and above-knee occlusive disease, Devine et al. reported
patency rates of HBD grafts at 1, 2, and 3 years of 70, 63, and 55% compared with 56, 46
and 42% for PTFE. Whether heparin bonding will improve the long-term patency of small
diameter prosthetic grafts remains to be determined (99).

Endothelialization of PTFE Grafts. Following the initial report by Herring et al. ( 1 00),
there have been numerous attempts to endothelialize the luminal surface of PTFE grafts
used for dialysis access or femoropopliteal bypass grafting procedures. The endothelial
cells can be obtained from a variety of sources, including human umbilical vein endothe-
lial cells, saphenous, jugular or arm veins, microvascular endothelial cells, omentum, and
autologous bone. The cells can be harvested by either mechanical or chemical digestion or
liposuction. Cells are either seeded or sodded (ultraheavy seeding) onto the graft surface.
Cell retention remains a problem with some of these methods despite the use of enhancing
agents such as fibroblast growth factor or fibronectin (101-110).

Endothelialization via transmural ingrowth occurs normally in grafts implanted in
canine thoracic aortas by 16 weeks. In humans, endothelialization is limited to the zone of
pannus ingrowth within 1-2 cm of the anastomosis. There is usually no transinterstitial
full-wall ingrowth. Recently, Wu et al. (Ill) have elegantly demonstrated endothelial and
smooth muscle cells remote from the anastomosis in an explant from a patient with a
perigraft seroma, suggesting endothelial cell fallout from the blood.

In humans, mechanically or enzymatically separated endothelial cells (ECs) and
microvascular ECs have been used to seed PTFE and Dacron grafts used for dialysis
access and infrainguinal bypass grafting in patients with no viable autogenous conduits.
Swendenborg (101) evaluated the patency of saphenous vein-derived ECs allowed to grow
to confluence on PTFE grafts in sterile culture media and implanted in patients under- 13

going hemodialysis. In two of the functioning grafts, irregularities developed at needle |

puncture sites. In a series comparing the patency rates of 10 sodded grafts with 8 control j§

grafts, Berman et al. (112) found no statistical differences in patency between sodded and ||

standard wall PTFE grafts (63 vs. 44%). The characteristic finding in sodded grafts was ^

the marked cellular ingrowth of tissue resulting in a significant reduction of luminal "jl

diameter (Fig. 18). s

In a phase II study, Zilla et al. reported a 72.9% patency rate for femoropopliteal 2

bypass grafts. This was somewhat lower than the 84.7% 3-year and 73.8% 5-year patency J

rates reported in phase I of their trial. The inclusion of a larger number of "redo" patients ©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES

109

Figure 18 Femoral angiogram of a sodded PTFE graft 4 years after implantation shows ir-
regularity throughout the graft with a focal area of high-grade stenosis. A. The velocities at the site
of stenosis were 457/cm/s. B. The patient refused operative intervention. The graft remained patent
for 3 more years after this angiogram.

and the unavailability of fibronectin for improving cellular adhesions were among the
reasons advanced for the discrepancy in patency rates between the two studies (105).

While technically feasible, endothelialization of prosthetic grafts has not gained
widespread application because of the cumbersome nature of the techniques used.
Furthermore, variable cell adhesion and sterility remain concerns in the cell culture-
dependent procedure.

2. Modification of Anastomotic Configuration

Vein Cuffs and Patches. Neointimal thickening at either the proximal or, more com-
monly, the distal anastomosis of PTFE grafts is the major cause of late failure of these
grafts. A number of modifications of the distal anastomosis of such grafts have been used
to alter the flow patterns at this location. There is considerable evidence suggesting that
low wall shear stress, resulting in prolonged particle residence, may be one of the factors
predisposing to anastomotic neointimal thickening. In a study using PTFE aortic grafts in
baboons, Kraiss et al. found that an increase in wall shear stress was associated with a

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

110

HUNTER and BULL

significant reduction in anastomotic neointimal thickening and smooth muscle cell
proliferation (113). Mattsson et al. have shown that high wall shear may actually induce
regression of neointimal thickening in the same model (114).

The Linton and Taylor patches, vein interposition cuffs (such as the Miller cuff and
Tyrell cuff), and precuffed or hooded grafts are all currently employed to enhance the
patency of below-knee PTFE grafts. Taylor reported patency rates of 74 and 58% at 12
and 36 months (Fig. 19) (115-120). Results of the U.K. prospectively randomized trial
comparing infrainguinal PTFE bypass grafts with and without vein interposition cuffs
showed patency rates for vein cuff and no vein cuff above-knee bypasses of 80 and 84%,
respectively, at 1 year and 72 and 70%, respectively, at 2 years (1 17). In contrast, bypasses
to the below-knee popliteal artery showed a significant difference between the cuffed
(52%) and noncuffed (29%) grafts at 2 years. Batson et al. showed cumulative patency
rates of 65% at 24, 36, and 48 months using the Linton patch (115). In a comparison
between distal vein cuff (DVC) and distal arteriovenous fistulas (DAVF), Kreinenberg
found 3-year primary patency for DAVF and DVC was 48 and 38%, and secondary
patency was 48 and 47%, respectively (121). Neville et al., in an evaluation of 80 PTFE
distal bypasses with vein patch, reported 70% primary patency at 3 years and 62.9% at 4
years (120). Current data on the value of patches and cuffs in improving PTFE graft

Figure 19 Color duplex scan of a Miller cuff at the distal posterior tibial artery anastomosis of a
PTFE bypass graft.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 111

patency are encouraging. Whether the addition of a distal arteriovenous fistula will
provide additional improvement in long-term patency awaits further study.

Hooded or Precuffed Grafts. The results of the hemodynamic studies of Harris et al.
have led to the introduction of precuffed (hooded) PTFE grafts (122). Reporting for the
North American Prospective Trial Investigations, Panneton found no significant differ-
ences in 30- and 10-month patency rates between hooded and PTFE grafts with vein
modification (123).

D. PTFE Dialysis Grafts

PTFE is the most commonly used synthetic graft for hemodialysis. Primary patency rates of
17 and 79% at 12 months (stretch grafts) have been reported (83,84,127-134). Secondary
patency rates range from 20 to 80%. There are a number of causes of failure of these grafts,
including hypercoagulable states, thrombus or stenosis at the arterial anastomosis, and
venous outflow stenosis. In addition, needle puncture site trauma (two 15-gauge needles
three times a week) results in tearing of the prosthesis, with pseudoaneurysm formation. The
needle tracks seal with thrombus and heal by fibrosis and neovascular ingrowth (Figs. 20, 21
and 22). Approximately 220,000 patients undergo dialysis in the United States annually, and
this is anticipated to increase by 8-10% per year. Gibson et al., in an analysis of the U.S.
Research Data System Dialysis Morbidity and Mortality Wave 2, reported that autogenous
fistulas had a higher primary patency rate of 39.8%, versus 24.6% for PTFE at 2 years, and
an equivalent secondary patency rate of 64.3 vs. 59.5% (135). In an attempt to improve the
longevity of dialysis fistulas, the use of carbon-coated hooded grafts and Taylor and notched
vein patches has been employed (128,132,133). Early reports of comparisons between
hooded and conventional PTFE grafts show improved primary and secondary patency
rates (66.3 and 93.2% for hooded grafts vs. 40.5 and 49.7% for standard wall prostheses).
Pipinos et al., in a comparison of the conventional anastomosis Taylor patch and the
"notched" vein technique, reported 6-month patency rates of 47, 25, and 41%, which were
not statistically different (133).

E. Complications

Early in their use, focal and diffuse dilatation of PTFE grafts was a serious limitation.
Although some suggested that hypertension was an important contributing factor, in fact
not all patients with dilated grafts were hypertensive. Instead, inherent structural weakness
of the prosthesis related to inadequate wall thickness and creep was the most likely
causative factor (41,42,86). Modern PTFE prostheses demonstrate very little tendency to
dilate (less than 10%).

Late complications of PTFE grafts include anastomotic stenosis or occlusion caused
by progression of atherosclerosis or neointimal hyperplasia, thrombosis, and, rarely, pseu- ■§

doaneurysm formation. g

I
1 . Anastomotic Neointimal Hyperplasia .g>

The most frequent cause of failure of PTFE grafts is stenosis or occlusion, most <

frequently involving the distal anastomosis of femoropopliteal grafts and the venous >9

anastomosis of dialysis access grafts (88). A perplexing problem has been the observation J

that occlusion may occur suddenly without antecedent symptoms. Q

Anastomotic neointimal hyperplasia and progression of atherosclerosis are the most |

frequent pathological findings in occluded grafts (Fig. 23). Careful monitoring at regular @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

112

HUNTER and BULL

Figure 20 A. Normal intraoperative fistulogram of a recently placed PTFE graft showing some air
bubbles. B. Angiogram of a PTFE loop graft demonstrating pseudoaneurysm formation (arrow).

intervals with Doppler-derived pressure indices, flow velocity measurements, and color
imaging are essential to detect preocclusive lesions, which are more amenable to correction
before thrombosis occurs.

Once thrombosis has occurred, resolution of this occlusive process is essential. This can
be achieved either by lysis of the luminal thrombosis with Retavase/Alteplase or by surgical
or percutaneous thrombectomy. There have been no prospective randomized trials compar-
ing the efficacy of these newer agents with surgical or percutaneous thrombectomy. Lytic
therapy, which successfully delineates the responsible occlusive lesion in approximately 70%
of patients with occlusions of less than 14 days 7 duration, is probably the management of
choice in the absence of limb-threatening ischemia (136). There have not been any
comparative studies evaluating the efficacy of the newer thrombolytic agents with throm-
bectomy. The recent practice of combining these agents with a glycoprotein Ilb/IIIa
inhibitor requires further evaluation (137). Thrombectomy is preferable if the limb is at
risk and there has been no significant improvement after 24^-8 h of lyric therapy.

A number of interventions are available for managing anastomotic stenoses; here the
procedure must, however, be tailored to the prevailing circumstances in each patient.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES

113

Figure 21 Photomicrograph demonstrating granulation tissue infiltrating a needle puncture site of
a PTFE graft (arrow).

Balloon angioplasty and patch angioplasty are the appealing options. However, either
procedure is associated with significant reocclusion rates at 3 months. Consequently, our
preference is to extend the anastomosis to an uninvolved segment of the vein (for stenosis
at the venous end of grafts used for dialysis access) or an uninvolved segment of artery (in
patients with occlusion of a femoropopliteal PTFE bypass graft). Construction of an
autogenous vein bypass using the saphenous vein or a vein from the arm may be indicated
in patients who require a bypass to the tibial vessels.

2. Anastomotic Pseudoaneurysm

Anastomotic aneurysms associated with the use of PTFE grafts are quite uncommon in
the absence of infection. Chiesa et al. reported an incidence of femoral anastomotic
aneurysms of 1.3% in their series of 2112 patients. An additional patient had an aortic
pseudoaneurysm (126). However, we have recently observed several dialysis patients with
pseudoaneurysm formation at arteria anastomoses and along the course of PTFE dialysis
access grafts resulting from either inadequate compression of needle puncture sites or
disruption of node fibril structure.

Retained pieces of PTFE in patients with failed dialysis access grafts or following
above- or below-knee amputation are important reservoirs for bacteria. Nonhealing
wounds and infected anastomotic aneurysms, which present with local and systemic signs
and symptoms of infection, are not infrequent consequences of this practice. If possible, all

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

114

HUNTER and BULL

Figure 22 Fistulogram of a PTFE dialysis graft demonstrating the pseudointimal lining the inner
surface of the graft (arrow). A stent infiltrated with neointimal ingrowth is present at the venous
anastomosis.

residual prosthetic material should be removed from patients with infective lesions in the
involved extremity (138).

The principles of evaluation and management of pseudoaneurysm related to the use of
PTFE grafts are similar to those for anastomotic aneurysms occurring with the use of
fabric or biological grafts.

F. Comparison of Dacron vs. PTFE Bifurcation and
Femoropopliteal Grafts

The experience with early PTFE aortic grafts was somewhat limited due to their stiffness,
lack of conformability, and bleeding from suture lines. The longitudinally extensible
configuration (stretch grafts) has significantly improved handling properties and reduced
suture line bleeding.

Prospective randomized trials have not revealed any significant differences in long-
term patency between Dacron and PTFE used for aortic reconstruction. Other criteria,

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES

115

Figure 23 Arteriogram of an above-knee PTFE graft demonstrating (A) neointimal thickening at
the distal anastomosis and (B) occlusion of the distal anastomosis. Thrombus is present in the graft.

surgeon preference, graft dilatation, infectability, ease of transection, pseudoaneurysm
formation, and propensity to anastomotic intimal hyperplasia are among the factors to be
considered in choosing between these two types of prostheses (30,82,124,125).

In a comparison of knitted double-velour Dacron and PTFE grafts, Friedman et al.
found that the cumulative patency for PTFE was 95 and 86% for Dacron between 66 and
72 months (125).

In a review of 228 bifurcated PTFE grafts used to reconstruct aortoiliac occlusive
disease, Chiesa et al. reported 4 (1.7%) early graft thromboses, 8 (3.3%) late graft limb
thromboses, 1 aortic graft infection, 1 femoral graft infection, and 1 aortic and 3 femoral
pseudoaneurysms (126).

When using PTFE or aortic reconstruction, the graft edges should be cut with a knife
blade so as not to distort the fabric. We prefer to use polypropylene instead of Gore-Tex
suture to perform the anastomosis. The graft limbs must be tunneled under slight tension
and filled with blood at arterial pressure to minimize elongation and subsequent kinking of
the graft. We have encountered very little suture line bleeding except during completion of
the second limb anastomosis.

If it has been necessary to do an endarterectomy of the proximal aorta, the proximal
anastomosis can be reinforced with a cuff from the body of the graft from which the
external wrap has been removed and the graft gently stretched to the desired diameter.

There are few reports of the use of Dacron for femoropopliteal bypass. Massry et al.
in a study of 200 grafts reported patency rates of externally supported above-knee Dacron
grafts of 76%, 71%, and 50% at 3, 5, and 10 years, respectively. Below-knee grafts had
patency rates of 65% and 57% at 3 and 5 years (7).

In a multicenter prospective randomized comparison of Dacron Microvel Hemashield
with Gore-Tex, Green et al. (6) reported 5 years primary patency rates of 45% and 43%

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

116 HUNTER and BULL

and secondary rates of 68% vs. 68%. From the available data, it appears that Dacron and
PTFE have similar patency rates when used for femoropopliteal bypass.

G. Perigraft Seroma

The accumulation of fluid around knitted Dacron double-veloured or PTFE grafts, first
described by Kaupp et al. (139), is a rare complication of vascular bypass procedures. The
incidence of perigraft seroma ranges from 0.2 to 1% of major vascular reconstructions
(140). The perigraft reaction is characterized by painless, fluctuant swelling surrounding a
portion of the prosthesis or the prosthesis in its entirety, often with erythema of the
overlying skin. Grafts placed in extra-anatomic locations (i.e., axillofemoral, femorofe-
moral, axilloaxillary dialysis access, or subclavian pulmonary bypass grafts) seem partic-
ularly prone to this complication and account for 60-75% of cases. Grafts in anatomical
positions (aortofemoral, femoropopliteal) account for the remainder (139-149). Persistent
perigraft seromas may also occur in association with saphenous vein femoropopliteal
grafts, especially in patients with involvement of pelvic lymph nodes due to malignancy or
radiation. Ahn et al. (144) documented perigraft seromas in 4.2% of patients with extra-
anatomic bypasses, 1.2% with aortofemoral bypasses, and 0.3% with femoropopliteal
bypass procedures.

The interval from graft insertion to the clinical presentation of the seroma typically
ranges from 1-45 months, with a mean of approximately 25 months (140). Systemic signs of
infection are invariably absent and the straw-colored fluid (unless contaminated by repeated
attempts at aspiration) is usually bacteriologically sterile. The fluid that resides within a
fibrous capsule investing a portion of or the entire graft is biochemically a transudate of
serum. In a survey of the members of the North American chapter of the International
Society for Cardiovascular Surgery, Blumenberg and coworkers (141) noted that knitted
Dacron (54%) and PTFE (34%) grafts were the prosthetic grafts most frequently involved.

Histologically, the tissues surrounding Dacron grafts demonstrate a gradation of
changes ranging from an acute inflammatory cell infiltrate at 1 week to mature granulation
tissue at 4-6 weeks. Foreign-body giant cells attached to the Dacron fibers are frequently
observed. By contrast, the reaction to PTFE grafts is characterized by fibrin deposition with
scant giant cell reaction confined almost entirely to the outer surface of the graft (141).

The precise cause of perigraft fluid accumulation is unknown. We have observed two
large periaortic fluid collections after use of the stretch PTFE graft, which we believe is
related to flushing the graft with heparinized saline to assess the integrity of the proximal
anastomosis. CT scan revealed no evidence of para-anastomotic aneurysms (Fig. 24).
Needle aspiration of the perigraft collection after an arteriogram to eliminate a para-
anastomotic pseudoaneurysm revealed myxomatous tissue and no fluid. Flushing the graft
with heparinized saline may be one explanation for the occurrence of perigraft seromas after ■§

Blalock Taussig shunts (146). This practice has been discontinued and heparinized blood is g

currently used to flush the graft to determine if there are major leaks at the proximal »

anastomosis. However, the fundamental abnormality appears to be failure of incorporation c

of the graft by the host tissues. Fluid transport through the interstices of the graft, fluid <

exudation from surrounding tissues, an allergic or immune response, mechanical irritation >9

of the host tissue by repeated motion of the prosthesis, impaired fibrin formation in the graft J

interstices caused by the use of heparin, and the presence of a fibroblast inhibitory factor in «

the serum are among the many etiological factors advanced (139-149). Sladen and |

colleagues (150) have recently demonstrated a human fibroblast inhibitor (molecular @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES

117

Figure 24 CT scan of a PTFE aortoiliac graft demonstrating perigraft seroma surrounding the
limbs of a PTFE bifurcated graft.

weight, 2000 Da) in the serum of patients who develop seromas. Heparin by impairing fibrin
formation in the interstices of the graft, may disturb sealing of the graft, resulting in fluid
excitation.

Observation or repeated aspiration to control fluid accumulation is successful in
approximately two-thirds of patients with perigraft fluid collections; however, secondary
infection or graft thrombosis — reported to occur in 5-8% of patients — remains a serious
concern (141). Therefore, except in poor-risk patients, removal of the graft and replacement
with a prosthesis of different material is recommended. Removal of the prosthesis or plasma-
pheresis is accompanied by a reduction in the fibroblast inhibitory properties of the serum,
suggesting that the graft material may play a role in its induction of the inhibitor. Graft
replacement will result in cure of more than 90% of patients. The technique of Lowery et al .
(151), in which a communication lined with omentum is fashioned between the graft capsule
and the peritoneal cavity, seems a reasonable alternative when the fluid is sterile. Whether
modulation of the inhibitor by microfibrillar collagen, ginseng, or high-dose vitamin C will
be effective remains to be evaluated (148).

I
•5

H. Suture Line Failure

The tensile strength of a prosthetic graft-artery anastomosis is dependent entirely on the
suture material for its structural integrity. The desirable physical characteristics of the

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

118

HUNTER and BULL

ideal suture material for vascular anastomosis include long-term durability, high tensile
strength, a favorable stress-strain relationship, minimal biological reactivity and infection,
a low coefficient of friction, and knot security. Healing at the graft-anastomosis interface is
limited to pannus ingrowth, which consists of smooth muscle and endothelial cells
originating from the arterial wall adjacent to the anastomosis. Although this may extend
for approximately 1-2 cm onto the prosthetic graft, it provides little if any, tensile
strength. Because silk, nylon, and polyethylene sutures lose tensile strength with time,
they should not be used with prosthetic grafts.

Silk sutures were the first used in vascular anastomosis. Moore and Hall (152) reported
25 anastomotic aneurysms associated with the use of silk sutures that were found to be either
fractured or absent in their cases. This finding was not surprising, since Cutler and Dunphy
(153) had demonstrated deterioration of silk within 2 years of implantation. Anastomotic
aneurysms caused by dissolution of silk sutures still may occasionally be encountered in
patients who underwent bypass procedures in the early 1960s (Fig. 25).

Braided Dacron and monofilament sutures, such as polypropylene, which do not
deteriorate over time, are presently the sutures most often used in constructing vascular
anastomosis. However, they will fracture if carelessly handled with instruments (154).
Dissolution or fracture of a monofilament suture results in variable disruption of the
anastomosis and consequent false aneurysm formation. The latter may rupture or erode

I
•5

Figure 25 Transfemoral aortogram demonstrating (A) a dilated graft (lower arrow) and an
anastomotic aneurysm (upper arrow) in a patient whose anastomosis was constructed with silk
sutures in 1966. The femoral anastomotic aneurysm from the same patient is demonstrated in B.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 119

into adjacent bowel, causing graft enteric fistula. Although a polypropylene suture may
fracture, this has been extremely rare in our experience (155-158). Careful selection of
sutures of appropriate composition and with sufficient tensile strength is an essential first
step in ensuring the integrity of an anastomosis.

Polypropylene, the suture material most commonly used to perform vascular pros-
thetic anastomosis, demonstrates no decrease in tensile strength when subjected to load or
exposed to body tissue fluids for prolonged periods of time (154). However, excessive
manipulation of the suture with surgical instruments or inadvertent knotting may weaken
the suture mechanically and lead to breakage. Indeed, few complications are ascribable to
the failure of modern sutures.

PTFE suture is very inert and resistant to degradation by either hydrolysis or tissue
enzymes. The inflammatory response, once implanted, is characterized by a mild gran-
ulation tissue response composed of a few foreign-body giant cells and a thin capsule,
which is usually stabilized at about 30 days (159,160).

I. Carotid Patching

There is increasing evidence that routine carotid patching reduces the incidence of
perioperative stroke and late restenosis (161). However, the choice of patch material
remains the subject of debate. Saphenous and jugular vein, Dacron, PTFE, and bovine
pericardium are among the materials used (162-172).

Proponents of autogenous patches point to their ready availability, ease of suturing
and conformability, and a potential reduction in perioperative thrombosis and infection.
The major disadvantage of vein patches is the irregularity in their mechanical integrity,
which appears to vary with the site of harvest. Another concern is the infrequent but
serious complications of aneurysmal dilatation and rupture (163).

Prosthetic and biological patches, while more expensive than autogenous vein, are
readily available, resistant to patch dilatation and rupture, and do not have the morbidity
associated with vein harvest.

In a prospective randomized study comparing saphenous vein with Dacron in 195
patients undergoing 207 carotid endarterectomies (CEAs), O'Hara et al. found no
significant differences in stroke mortality or restenosis rates (162). In a series of 274
patients undergoing CEAs, Hayes et al. found that patients with Dacron patches had a
greater number of embolic events than those with a saphenous vein patch. However,
there was no difference in the number of patients with >50 emboli between the two
groups (163).

In a study of 200 CEAs comparing PTFE with collagen-impregnated Dacron
(Hemashield) as the patch material, AbuRahma et al. reported a higher incidence of
stroke (0 vs. 7%), carotid thrombosis (0 vs. 5%), and restenosis (2 vs. 5%) in patients ■§

treated with the Hemashield patch. The authors concluded that collagen-impregnated g

Dacron may be more thrombogenic than PTFE. Further prospective studies are necessary »

to verify the results of this study (164-167). c

Comparing primary closure (PC) with PTFE and vein patch closure (VPC), Abu- <

Rahma et al. reported an ipsilateral stroke rate of 5% with PC compared to 1 % with PTFE >9

and 0% with VPC. The incidence of restenosis and occlusion was 34% for PC versus 2% %

for PTFE and 9% for VPC. The incidence of restenosis requiring reoperation was 11% for «

PC, 1% for PTFE, and 2% of VPC (164,165). Archie, in a study comparing vein and 1

Dacron patch angioplasty, reported similar low perioperative thrombosis and stroke @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

120

HUNTER and BULL

rates. Vessels closed with Dacron were more likely to develop a stenoses >50% than those
closed with vein (168).

There have been few opportunities to evaluate specimens of healed Dacron or PTFE
patches in individuals with patent repairs. Shi and colleagues reported complete healing of
a carotid Dacron patch in a patient who died of congestive heart failure (169). The
specimen was completely incorporated by full-wall tissue ingrowth. The flow surface of the
patch, in place for 25 months, was lined by a smooth neointima consisting of smooth
muscle and endothelial cells.

Bovine pericardium (BioGuard, Biovascular Inc.) is growing in popularity as a patch
material (169,170). In a series of 1 12 patients treated with a pericardial patch, Grimsley et
al. reported no strokes and a 2% incidence of restenosis (>70-99%) (169). The authors
conclude that bovine pericardial patches are associated with a low incidence of midterm
complications. Reporting on their long-term experience with bovine pericardiual patches,
Biasi et al. (170) found no statistically significant differences in stroke rate or restenosis of
> 60% between vessels closed primarily and those closed with a pericardial patch.

It seems clear that carotid patching is a significant factor in reducing both early
thrombotic occlusion and late restenosis in patients undergoing carotid endarterectomy.
Each material has some disadvantages: groin wounds (4%), rupture and saphenous nerve
injuries with saphenous vein, needle hole bleeding with PTFE, the questionable durability
of bovine pericardium over the long term, and the possible increased incidence of
thromboembolism and restenosis with coated Dacron (Fig. 26).

■ ■■>'; '!#"•'

V.

1 1

I

•5

Figure 26 Photomicrograph of a coated Dacron patch removed from a patient with carotid
restenosis. Neointimal thickening is present on the luminal surface. The carbon pigment of the
guideline is indicated by the arrow. (Original magnification x 100.) Macrophages and multi-
nucleated giant cells can be seen ingesting the pigment and Dacron fibrils (arrow). (Original
magnification x 200.)

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

HEALING OF PROSTHESES 121

The prosthetic material for "redo" operations is also unclear. In a recent review of 82
carotid reoperations, Rockman et al. reported a higher restenosis rate (2.6 vs. 2.3%) in
patients reconstructed with vein than those performed with prosthetic material and
recommended prosthetic patch material be used in patients with restenosis (171).

J. Healing of Endovascular PTFE Grafts Used to Treat
Occlusive Disease

The use of transluminally placed self-expanding stented grafts has recently been advocated
to treat occlusive disease of the aortoiliac and femoropopliteal segments. A study
evaluating the safety and efficacy of PTFE-lined nitinol endoprostheses (Hemo Bahn) in
141 limbs in 127 patients reported cumulative patency rates of 98 ± 3% and 91 ± 4% for
patients treated for iliac lesions and 90 ± 3% and 79 ± 5% for femoral artery lesions at 6
and 12 months, respectively (173). These grafts are composed of thin-walled PTFE radially
reinforced with nitinol (wall thickness 0.1 mm, pore size 30 urn).

Marin (174), in a series of 26 stented grafts in 21 patients with limb-threatening
ischemia, observed that after 3 weeks, organizing thrombus was present on both the intra-
luminal surface of the artery and extraluminal surface of the implants. By 6 weeks, the
outer surface of the stent graft was firmly adherent to the wall of the native artery. The
neointimal lining the lumen of the graft consisted of a thin layer of fibrous tissue with an
overlying monolayer of endothelial cells within 2 cm of the graft-to-artery anastomosis. At
3 months, the neointima was present within 1-3 cm of the anastomosis and measured
between 40-150 p,m in thickness. By 7 months, the graft was well incorporated, with an
external capsule. The depth of insertion of the graft-i.e., media or periadvential plane —
determined the extent of healing. Grafts inserted into the media had less mononuclear and
foreign-body giant cell reaction than those placed within the periadventitial plane.

Plaque tissue in the iliac or femoral arteries underlying these stents was histologically
composed almost entirely of acellular fibrous tissue. In one graft segment, the authors
observed extrinsic smooth muscle cell proliferation of sufficient thickness to indent the stent.
Van Sambeek et al. noted only minimal luminal changes using intravascular ultrasound to
interrogate grafts implanted in the superficial femoral arteries of 12 patients (175).

K. Healing of PTFE Grafts in the Venous System

1. Venal Caval Replacement

Tumors of the liver, kidney, and adrenal gland and soft tissue sarcomas of the retro-
peritoneum may all invade the inferior vena cava. Primary or metastatic mediastinal
malignancies are the most common cause of obstruction of the superior vena cava (SVC).
Mediastinal fibrosis and thrombosis due to central lines and indwelling catheters are the

most frequent nonmalignant causes of SVC obstruction. Clinically, upper or lower ■§

extremity edema and venous engorgement should suggest the presence of vena caval g

obstruction. CT used to evaluate these patients has shown that vena caval obstruction may as

occur in the absence of significant symptoms (176,177). In an attempt to perform a c

curative resection in patients with tumors invading the inferior vena cava, excision of the <

vena cava and replacement with ringed PTFE have recently been undertaken. Sarkar et al., >9

in a report of 10 patients, showed that 7 of the 10 grafts were patent or functionally patent ^

at a mean interval of 9 months (176). Although the follow-up period is short and the long- «

term survival of these patients is presently unknown, vena caval replacement with PTFE |

does remain patent in the short term. Alimi et al., in a review of SVC reconstruction for @

nonmalignant disease, reported 30-day primary and secondary patency rates of 79 and %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

122 HUNTER and BULL

95%, respectively. Significant relief of symptoms was seen in 79% of patients followed
long term (177).

2. Portocaval Shunt

Sarfeh et al., in an attempt to reduce portal pressure without abolishing hepatopetal
flow, have popularized the use of partial shunts using 8- and 10-mm internal diameter
PTFE grafts (178-180). Initially, unsupported grafts were used, but recently almost all
grafts used have been ringed, externally supported PTFE grafts. In a series of 46 grafts
used in this position, Rypins reported an early thrombosis rate of 15% with PTFE
compared to 30-40% reported for Dacron grafts. Thrombosis of these shunts is usually
treated nonoperatively with catheter-based interventions (178).

3. Crossover Grafts

PTFE grafts have also been used to bypass obstructed venous segments in the lower
extremities in order to relieve symptoms of venous hypertension. The patency or externally
supported PTFE grafts can be enhanced by performing an adjunctive arteriovenous
fistula. Long-term anticoagulation with warfarin is usually necessary (57,181).

There is not a great deal of information regarding the healing characteristics of
prosthetic grafts implanted in the venous system. The desirable qualities of a graft
implanted in the venous system include reduced wall thickness, a smooth inner flow
surface, and adequate pore size. Patent explants of 60- and 90-um ePTFE show complete
neointimal lining involving the entire extent of the graft. Pannus ingrowth (1-2 cm) is
evident at both anastomosis. Endothelial cells originating from perigraft microvessels with
widely patent endothelial channels are readily evident (182,183).

In an attempt to improve the patency of venous conduits, investigators have wrapped
ePTFE grafts with peritoneum and infused heparin locally. The lined grafts had a similar
patency rate at 1 week (86%) as regular PTFE grafts, but patency had decreased to 57% in
lined grafts at 8 weeks. Histological examination of the implants removed at 6 weeks
showed stenosis of all PTFE grafts due to thrombosis, whereas lined grafts narrowed due
to proliferation of granulation and inflammatory tissue between the mesothelial lining and
the graft cover (184).

III. POLYURETHANE

The limitations of long-term vascular access using either an autogenous arteriovenous
fistula or PTFE interposition graft is that they must mature before they can be accessed.
This period of maturation is typically 6 weeks to 6 months for arteriovenous fistula or 10—
14 days for a PTFE bridge graft. Because of this delay, a temporary central venous

catheter with its incumbent problems is usually required. It is estimated that approx- ■§

imately 2.3 catheters per patient year are needed for the average patient requiring dialysis. |

In addition to the increased cost of these two-staged procedures, the complications a

associated with catheter placement, infection, thrombosis, and venous stenosis result in c

increased morbidity and mortality (185). <

Polyurethane grafts, such as Vascugraft/Vectra, offer the potential for early cannu- >9

lation, thus decreasing cost and morbidity. The Vascugraft comprises randomly oriented J

microfibers of varying thickness (0.1-5.0 |a.m), compared to 0.1-1 urn for PTFE, creating «

pores that communicate throughout the graft wall. Large globules up to 50 urn in diameter |

are also present. The Vascugraft is thinner than reinforced PTFE (0.46 ± 0.03 vs. 0.86 ± @

0.08 urn) and has somewhat less tensile and bursting strength than PTFE. Because of their %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 123

elasticity, polyurethane grafts show more elongation than PTFE (220 vs. 60%) at their
breaking point (186-188).

The Vectra graft has a trilayer design consisting of (a) an inner microporous layer
coated with a surface-modifying agent to minimize platelet adhesion; (b) a middle layer of
Thoralon to provide strength, flexibility, and sealing properties; and (c) an outer micro-
porous layer to permit tissue ingrowth. In a randomized study of 142 patients comparing
the Vectra graft with PTFE, Glickman et al. reported primary patency rates of 55% for
Vectra versus 47% for PTFE at 6 months and 44% versus 36% at 12 months, respectively.

Secondary patency rates were 87 and 90% Vectra versus PTFE at 6 months and 78
versus 80% for PTFE at 6 and 12 months, respectively. Some 20-33% of polyurethane
grafts were cannulated within 3 days and 53.9% by 8 days. No PTFE grafts were
cannulated at 9 days. The authors observed no differences in graft survival between the
two grafts (185).

Polyurethane grafts have also been used in the femoropopliteal region, with a primary
patency rate of 66% and secondary patency of 80% at 1 year(189). Histological examination
of occluded explants revealed that the grafts were surrounded by a capsule of varying
thickness. The luminal surface of the grafts were covered with a thin layer of thrombotic
material with pannus formation at the anastomosis. There was no evidence of endothelial-
ization of the luminal surface (190) or of any significant chemical changes within the grafts
examined.

The major advantage of polyurethane vascular grafts used for dialysis is early
cannulation. The disadvantages include the need of a vein of adequate diameter (>3
mm) and the absence of curves or muscle bulk to negotiate. These grafts are technically

Figure 27 Fistulogram of a Vectra graft showing kinks adjacent to the anastomosis and within the
graft.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

124 HUNTER and BULL

more difficult to implant; kinking and elongation are commonly encountered. An
interposition PTFE segment may be necessary if any turns or bends are encountered in
placing these grafts (Fig. 27).

IV. BIOLOGICAL GRAFTS

A. Umbilical Vein Grafts

Glutaraldehyde-stabilized human umbilical vein (HUV) grafts are presently used as
alternatives to PTFE grafts for femoropopliteal or tibial bypasses and for dialysis access.

1 . Preparation

These grafts are prepared from umbilical cords harvested in the delivery room, which
are manually cleaned and stripped. The collagen of cords with acceptable diameter and
quality is then cross-linked with glutaraldehyde; soluble proteins and excess Wharton's gel
are extracted with ethanol (191). The external surface is covered with a supporting Dacron
mesh. Glutaraldehyde starch has been shown to be superior to dialydehyde starch as a
cross-linking agent. This results in a more stable graft that is less prone to biodegradation.
Five-year primary patency rates range from 53 to 83% for above-knee bypasses. Bypasses
to the below-knee popliteal artery and crural vessels have reported secondary patency
rates of 71 and 56% (with distal arteriorenous fistulas), respectively (192,193).

2. Healing Characteristics

The healing characteristics of these grafts in humans have been described by Guidoin
et al. (194) and more recently by Batt et al. (195) in a study of 39 explants. Macroscop-
ically, they demonstrated irregular thickening of the walls of the grafts, with folds of
varying depths involving the luminal surface in virtually all explants. The thrombotic
material was present on the luminal surface, especially at anastomosis. The external sur-
faces of the grafts were covered by fibrous capsules of variable thickness, increasing pro-
portionately with the duration of implantation (194,195).

Scanning electron microscopy demonstrated evidence of lipid deposition in 18 of 39
prostheses; in 13% of the grafts, bacteria were present even though clinical infection was
not diagnosed. Cellular infiltrates, most severe when clinical infection was present, were
readily demonstrable. Also seen was delamination of the luminal surface (195).

3. Complications

Dilatation. The incidence of aneurysmal dilatation, a not infrequent complication of
HUV grafts, increases progressively beyond 5 years. Dilatation of HUV grafts is biphasic:
early enlargement of grafts from 4 to 6 mm at the time of implantation to approximately 9
mm in diameter is the norm. Using b-mode ultrasound, Dardik et al. (193) have dem-
onstrated a 21% incidence of dilatation and 36% incidence of focal aneurysms in patients e
studied beyond 5 years. Cranley et al. (196) found discrete aneurysms (diameter greater a
than 20 mm) in 11 of 25 patients (44%) surviving with patent grafts for 5 years. %

Two morphological variants of aneurysms have been described by Dardik et al. (192) =j

in association with HUV grafts: uniform diffuse dilatation of both the graft and its Dacron j

mesh or erosion of the graft with rupture of the mesh, resulting in multiple false g

aneurysms. «

The progressive increase in diameter of these grafts beyond 5 years is not entirely "g

unexpected. Reversal of glutaraldehyde-induced cross-linking and immunological mech- g

anisms have been proposed to explain aneurysm formation in these grafts (193). The 8

1
§

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 125

frequent observation of bacteria within these grafts in the absence of overt infection is
intriguing and deserves further study, as it may play a role in aneurysm formation. In their
most recent series, Dardik et al. report no aneurysms in 283 bypass grafting procedures
performed over a 10-year period.

The guidelines for excision and grafting of aneurysmal HUV grafts are not clear.
Dardik et al. (192) state that approximately 6% of these aneurysms in grafts implanted
more than 5 years previously are of sufficient size to require surgical excision and repair.
However, it appears prudent, from the studies of Cranley et al. (196), to consider resection
if the dilated segment exceeds 20-30 mm in size. Excision and grafting of segmental
aneurysmal change or replacement of the entire graft if diffusely dilated may be necessary.

Anastomotic Aneurysms. Anastomotic aneurysms have been reported infrequently
(1.4%) with the use of HUV grafts and appear to develop more often at anastomotic sites
between HUV and dacron grafts (9%) than between HUV host arteries (0.6%) (193).
Serial monitoring with duplex imaging, with examinations increasing in frequency beyond
the 5-year mark, is essential if significant complications are to be avoided.

Anastomotic Stenosis. Late thrombosis of HUV grafts caused by anastomotic
stenosis (in 1.8% of the patients reported by Dardik et al.) is usually associated with
neointimal hyperplasia or progression of atherosclerosis at the distal anastomosis (197).
Dissection of these grafts can often be extremely difficult because of the dense fibrous
reaction that frequently envelops them. If the distal anastomosis can be exposed, a
longitudinal arteriotomy is made and extended onto the native vessel beyond the occlusive
lesion. Adherent thrombus is removed carefully and the arteriotomy and graftotomy
closed with a vein patch. Thrombus within the grafts should be flushed out with hepari-
nized saline solution rather than extracted with a thrombectomy catheter, since these often
fracture the luminal surface, thus predisposing the patient to secondary thrombosis. Using
these techniques to restore patency to popliteal and crural vessels, Dardik et al. (192) have
reported early secondary patency rates of 44-56%. More recently, the authors have used
lytic therapy to restore patency to occluded grafts with successful lysis in 85% and limb
salvage in 74% of cases (197).

Although bacteria can be demonstrated in significant numbers of HUV graft explants,
the incidence of clinical infection nonetheless remains low. In a series of 907 patients,
Dardik et al. (192) reported an overall infection rate of 4.3%, which had declined to 3.2%
in their most recent update. The incidence was 0.6% in patients undergoing primary
revascularization and 3.7% after reoperation. The clinical presentation, evaluation, and
management of infection involving HUV grafts are similar to those seen with other
prosthetic and biological grafts and are fully discussed elsewhere (192,197).

B. Cryopreserved Allografts

Cryopreserved saphenous and femoral vein allografts are currently used for femoropo-

pliteal bypass grafting, dialysis access, and replacement of thrombosed iliac veins or the

inferior or superior vena cava and in patients with peripheral prosthetic graft infections .g>

with no other suitable autogenous conduits. Arterial allografts are used to replace the §

aortic root and thoracic and abdominal aorta in patients with bacterial endocarditis, my- M

cotic aneurysms, or prosthetic graft infection. M

1 . Preparation and Healing Characteristics ■§

Veins and arteries retrieved from tissue organ donors are carefully rinsed and stored in g

liquid nitrogen at -120(to -196(C in 15-20% dimethyl sulfoxide (DMSO). Electron 8

1
§

•c

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

126 HUNTER and BULL

microscopy of the vein segments has confirmed their structural integrity after preservation.
The endothelial lining of the veins retains its ability to grow in cell culture as well as to
produce prostacyclin.

2. Dialysis Grafts

Cryopreserved femoral veins are used for dialysis most often in the presence of infec-
tion. In a series of 48 cryopreserved dialysis grafts, Matsuura reported 1-year primary and
secondary patency rates of 49 and 75%, respectively, which compared favorably with the
65% primary and 78% secondary patency rates they achieved with brachial artery-to-
axillary vein prosthetic grafts (198).

In a study of 22 allograft explants, Johnson et al. (199) demonstrated an intact
endothelial lining with evidence of cellular damage and severe rejection as manifested by a
lymphocytic infiltrate with CD3, CD8, and CR3 cytotoxic granules in 29% of the explants.
The cells lining the lumen and those in the vessel wall were repopulated with cells from
the recipient, resulting in either a novel cellular constituency or a mosaic of host and
donor cells.

It has been suggested that cryopreserved veins should not be used for dialysis in
patients awaiting renal transplantation because a significant number will develop allo-
sensitization. The data, however, appear to be contradictory as to whether use of allografts
results in sensitization of the recipient (200).

3. Venous Replacement

There have only been anecdotal reports of the use of cryopreserved femoral veins to
bypass occluded segments of the major veins of the upper and lower extremities. Cryopre-
served venous valves are presently being evaluated to determine their efficacy in restoring
venous competence. In a phase I feasibility study in 10 patients, Dalsing et al. reported 6-
month valvular patency rates of 67 and 78% and freedom from valvular incompetence at
56% in patients undergoing femoral and popliteal vein valve transplants (201).

4. Arterial Replacement

Cryopreserved saphenous vein is presently being evaluated by many surgeons as an
alternative conduit in patients with inadequate or absent autogenous veins for infraingui-
nal bypass grafting procedures (202-204).

Patency rates for femoropopliteal or crural bypass procedures range from 36 to 66%.
In a recent review of 76 patients undergoing 80 bypass procedures followed for a mean of
17.8 ± 20.9 months, Harris et al. reported primary patency rates of 36.8% at 1 year and
23.6% at 3 years (204).

Patients undergoing femoropopliteal bypass grafting with cryopreserved vein should
be anticoagulated with warfarin sodium and antiplatelet agents (203). We have also placed
some patients on low-dose prednisone in an attempt to mitigate the immune response. The 1

presently available data suggest that the use of cryopreserved vein in infrainguinal bypass 1

grafting should be limited to those patients with limb salvage or infection with no other js

autogenous conduit. ||

5. Cryopreserved Arterial Allografts s
Arterial allografts are used to replace the aortic root, thoracic aorta, or abdominal J

aorta in patients with bacterial endocarditis, mycotic aneurysms, or infected prosthetic °

grafts (205,206). In a comparison between prosthetic graft replacement and cryopreserved |

allografts in patients with aortic infection, Vogt et al. reported lower perioperative (6 vs. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 127

18%) mortality and complications rates (24 vs. 63%) in patients with allografts compared
to those treated with prosthetic grafts (205). Explanted grafts were acellular, with
denudation of the endothelial lining and a low-grade B-cell lymphocyte infiltrate. The
elastic tissue was fragmented but the collagen layer was intact.

Leseche, in a report on 28 patients undergoing treatment for abdominal aortic
infection, reported a 17.8% mortality rate and a 67% 3-year survival in patients treated
with cryopreserved aortic allografts. Although few complications of the use of cryo-
preserved arterial conduits are discussed in the literature, anastomotic disruption due
to persistent infection and aneurysmal dilatation (3 of 23 Leseche) remain an ongoing
concern (206).

C. Bovine Xenografts

When first introduced, bovine xenografts were used for dialysis access and for femo-
ropopliteal bypass grafting. However, poor long-term patency rates of 50% at 5 years, a
3-6% incidence of aneurysmal dilatation, and a 6-7% incidence of infection when used for
infrainguinal bypass grafting have limited their use to dialysis access fistulas (207). There
are several advantages to their use for that purpose, including availability, ease of
implantation, and immediate use following insertion.

When used for dialysis access procedures, bovine xenografts have a 1-year patency of
26-91% and a 2-year patency between 42 and 83% (208-210) compared with patency rates
of 62-93% and 57-85% at 1 and 2 years, respectively, for PTFE grafts (209). Hurt et al.
(211), in a study of 140 grafts comparing PTFE grafts and bovine xenografts, were unable
to demonstrate any significant difference between the two. Similarly, Reese et al. found no
difference in 1-year patency rates (59 vs. 66%) or aneurysmal dilatation (2.2 vs. 2.6%) in a
study comparing bovine heterografts and PTFE (212).

The late complications of bovine xenografts used for dialysis access include throm-
bosis caused by hyperplasia at the venous anastomosis, pseudoaneurysm formation at
needle puncture sites, infection that may result in dissolution of segments of the graft, and
dilatation of the entire graft. In a comprehensive retrospective analysis of 385 bovine
heterografts, Brems and associates (208) reported 160 episodes of thrombosis, 18 cases of
dilatation, 18 cases of puncture site pseudoaneurysms, and 8 wound infections.

Thrombosis of bovine heterografts usually necessitates thrombectomy and revision of
the venous outflow tract.

Infection occurring within the first month of implantation almost invariably involves
the anastomosis, requiring removal of the prosthesis. Since late infection usually occurs at
needle puncture sites, replacement of segments of the graft may be undertaken. Dilatation
of the entire prosthesis or diffuse pseudoaneurysm may require complete removal of the
graft. Focal pseudoaneurysms in the absence of infection can be managed by local excision
and placement of an interposition graft to bypass the involved segment.

V. CONCLUSION

The introduction of prosthetic and biological grafts into clinical practice has substantially
improved the survival of patients with peripheral vascular disease. However, vascular
surgeons generally have not been critical in their evaluation of prostheses used to bypass
or replace diseased vessels. The more frequent implantation of grafts into younger patients
and the extension of surgical techniques to distal, small-vessel bypasses have heightened

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

128 HUNTER and BULL

the need for durable prostheses and an improved understanding of interactions at the
blood-surface interface. It is evident that large-diameter grafts (greater than 8 mm internal
diameter), although they generally function well, are nevertheless subject to complications
that may lead to loss of life or limb. The long-term patency of small-diameter grafts (less
than 6 mm) remains considerably lower than that of their larger-diameter counterparts.
Anastomotic intimal hyperplasia and thrombosis of these grafts have not been signifi-
cantly reduced by the use of antithrombotic and antiplatelet agents. Modification of
surface thrombogenicity and modulation of anastomotic wall shear stress by endothelial
cell seeding, carbon coating, heparin bonding, or patches and cuffs offers some improve-
ment in patency. However, the long-term efficacy of such measures has not yet been fully
established.

A number of changes are presently under way that will influence the long-term results
of prosthetic grafts. As greater emphasis is placed on endovascular graft placement, a
number of pitfalls loom on the horizon. First, the decision by DuPont to withdraw its
polymers from use in implantable devices has removed a stable, reliable source of polyester
fabric from the market. The reliability of alternate replacement biomaterials requires
careful monitoring. The use of ultrathin prostheses for endovascular grafts ignores the
lessons learned from the use of ultrathin implantable Dacron prostheses. In addition, the
practice of suturing or welding metals onto Dacron prostheses will result in focal areas of
weakness in the material, with a propensity to rupture. Concerted efforts by graft
manufacturers to produce durable, nonthrombogenic grafts — along with careful patient
selection, operative technique, modification of risk factors, improvement in our under-
standing of the interaction between the graft and the host tissues, and continued
surveillance — are essential if we are to improve the patient's long-term survival and graft
function.

REFERENCES

1. Pourdeyhimi B, Wagner D. On the correlation between the failure of vascular grafts and their
structural and material properties: A critical analysis. J Biomed Mater Res 1986; 20(3):375-409.

2. Sauvage LR. Biologic behavior of grafts in arterial system. In: Haimovici H, ed. Vascular
Surgery 3rd ed. East Norwalk: Appleton and Lange, 1989:136-160.

3. Malone JM, Moore WS, Goldstone J. Phe natural history of bilateral aortofemoral bypass
grafts for ischemia of the lower extremities. Arch Surg 1975; 1 10(1 1): 1300 1 306.

4. Szilagyi DE, Elliott JP Jr, Smith RF, Reddy DJ, McPharlin M. A thirty-year survey of
the reconstructive surgical treatment of aortoiliac occlusive disease. J Vase Surg 1986; 3(3):
421-436.

5. Nevelsteen A, Wouters L, Suy R. Aortofemoral dacron reconstruction for aorto-iliac oc-
clusive disease: a 25-year survey. Eur J Vase Surg 1991; 5(2):179-186.

6. Green RM, Abbott WM, Matsumoto P, Wheeler JR, Miller N, Veith FJ, et al. Prosthetic |
above-knee femoropopliteal bypass grafting: Five-year results of a randomized trial. J Vase «
Surg 2000; 31(3):417-425. J

7. el-Massry S, Saad E, Sauvage LR, Zammit M, Smith JC, Davis CC, et al. Femoropopliteal c
bypass with externally supported knitted Dacron grafts: A follow-up of 200 grafts for one to 5
twelve years. J Vase Surg 1994; 19(3):487^494. I

8. Johnson WC, Lee KK. A comparative evaluation of polytetrafluoroethylene, umbilical vein, §
and saphenous vein bypass grafts for femoral-popliteal above-knee revascularization: A g
prospective randomized Department of Veterans Affairs cooperative study. J Vase Surg 2000; "§
32(2):268-277. J

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 129

9. Quinones-Baldrich WJ, Prego AA, Ucelay-Gomez R, Freischlag JA, Ahn SS. Baker JD, et al.
Long-term results of infrainguinal revascularization with polytetrafluoroethylene: a ten-year
experience. J Vase Surg 1992; 16(2):209-217.

10. Guidoin R. A biological and structural evaluation of retrieved Dacron arterial prostheses US
Department of Commerce /National Bureau of Standards Implant Retrieval: Material and Bio-
logical Analysis. 1981.

11. Harris PL. Aorto-iliac-femoral re-operative surgery. Supplementary surgery at secondary
operations. Acta Chir Scand Suppl 1987; 538:51-55.

12. el-Massry S, Saad E, Sauvage LR, Zammit M, Davis CC, Smith JC, et al. Axillofemoral
bypass with externally supported, knitted Dacron grafts: A follow-up through twelve years. J
Vase Surg 1993; 1 7(1): 107 1 1 5.

13. Mii S, Mori A, Sakata H, Kawazoe N. Fifteen-year experience in axillofemoral bypass with
externally supported knitted Dacron prosthesis in a Japanese hospital. J Am Coll Surg 1998;
186(5):581-588.

14. Rutherford RB, Patt A, Pearce WH. Extra-anatomic bypass: A closer view. J Vase Surg 1987;
6(5):437-446.

15. Madiba TE, Mars M, Robbs JV. Aortobifemoral bypass in the presence of superficial femoral
artery occlusion: Does the profunda femoris artery provide adequate runoff? J R Coll Surg
Edinb 1998; 43(5):3 10-313.

16. Jonas RA, Schoen FJ, Levy RJ, Castaneda AR. Biological sealants and knitted Dacron: po-
rosity and histological comparisons of vascular graft materials with and without collagen and
fibrin glue pretreatments. Ann Thorac Surg 1986; 41(6):657— 663.

17. Quinones-Baldrich WJ, Moore WS, Ziomek S, Chvapil M. Development of a "leakproof,"
knitted Dacron vascular prosthesis. J Vase Surg 1986; 3(6):895-903.

18. Jonas RA, Ziemer G, Schoen FJ, Britton L, Castaneda AR. A new sealant for knitted Dacron
prostheses: Minimally cross-linked gelatin. J Vase Surg 1988; 7(3):414— 419.

19. Kadoba K, Schoen FJ, Jonas RA. Experimental comparison of albumin-sealed and gelatin-
sealed knitted Dacron conduits. Porosity control, handling, sealant resorption, and healing. J
Thorac Cardiovasc Surg 1992; 103(6): 1059-1067.

20. McGee GS, Shuman TA, Atkinson JB, Weaver FA, Edwards WH. Experimental evaluation
of a new albumin-impregnated knitted dacron prosthesis. Am Surg 1987; 53(12):695-701.

21. Bethea MC, Reemtsma K. Graft hemostasis: An alternative to preclotting. Ann Thorac Surg
1979; 27(4):374.

22. Cooley DA, Ramagnoli A, Nilam JD, Bossort MI. A method of predisposing woven Dacron
grafts to prevent interstitial hemorrhage. Cardiovasc Dis (Bull Texas Heart Inst) 1981; 8:48.

23. Adachi H, Mizuhara A, Yamaguchi A, Murata S, Kamio H, Ino T, et al. Clinical experience
of a new gelatin impregnated woven Dacron graft. Japan J Artif Organs 1996; 25(1):214-219.

24. Wesolow A. The healing of arterial prostheses — the state of the art. Thorac Cardiovasc Surg
1982; 30(4):196-208.

25. Guidoin R, Gosselin C, Martin L, Marois M, Laroche F, King M, et al. Polyester prostheses
as substitutes in the thoracic aorta of dogs. I Evaluation of commercial prostheses. J Biomed
Mater Res 1983; 17(6):1049-1077.

26. Anerson JM. Microvel with hemashield vascular grafts. A preliminary report of the healing j>
response in humans. Angiol Arch Bd 1985; 9:73-77. <S

27. Drury JK, Ashton TR, Cunningham JD, Maini R, Pollock JG. Experimental and clinical js
experience with a gelatin impregnated Dacron prosthesis. Ann Vase Surg 1987; l(5):542-547. °|

28. Ukpabi P, Marois Y, King M, Deng X, Martin L, Laroche G, et al. The Gelweave polyester ^
arterial prosthesis. Can J Surg 1995; 38(4):322-331. |

29. Farooq M, Freischlag J, Kelly H, Seabrook G, Cambria R, Towne J. Gelatin-sealed polyester ^
resists Staphylococcus epidermidis biofilm infection. J Surg Res 1999; 87(1): 57-61. Q

30. Prager M, Polterauer P, Bohmig HJ, Wagner O, Fugl A, Kretschmer G, et al. Collagen versus |
gelatin-coated Dacron versus stretch polytetrafluoroethylene in abdominal aortic bifurcation 2

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

130 HUNTER and BULL

graft surgery: Results of a seven-year prospective, randomized multicenter trial. Surgery 2001;
130(3):408^tl4.

31. The Canadian Multicenter Hemashield Study Grouplmmunologic response to collagen-
impregnated vascular grafts: A randomized prospective study. The Canadian Multicenter
Hemashield Study Group. J Vase Surg 1990; 12(6):741-746.

32. Knox WG, Miller RE. Long-term appraisal of aortic and arterial homografts implanted in
years 1954-1957. Ann Surg 1970; 172(6):1076-1078.

33. Humphries AW, Hawk WA, DeWolfe VG, Le Fevre FA. Clinicopathologic observations on
the fate of arterial freeze-dried homografts. Surgery 1959; 45:59-71.

34. Cooke PA, Nobis PA, Stoney RJ. Dacron aortic graft failure. Arch Surg 1974; 108(1):101-
103.

35. Perry MO. Early failure of Dacron prosthetic grafts. J Cardiovasc Surg (Torino) 1975; 16(3):
318-321.

36. Blumenberg RM, Gelfand ML. Failure of knitted Dacron as an arterial prosthesis. Surgery
1977; 81(5):493-496.

37. Nucho RC, Gryboski WA. Aneurysms of a double velour aortic graft. Arch Surg 1984;
119(10):1182-1184.

38. Creech OJ, Deterling RA, Edwards S, Julian OC, Linton RR, Schumacker H. Vascular
prostheses: Report of the Committee for the Study of Vascular Prostheses of the Society for
Vascular Surgery. Surgery 1957; 41:62-80.

39. Eastcott HHG Rupture of Orion aortic graft after six years. Lancet 1962; 2:75-76.

40. Hayward RH, White RR. Aneurysm in a woven Teflon graft. Angiology 1971; 22(4): 188-190.

41. Campbell CD, Brooks DH, Webster MW, Bondi RP, Lloyd JC, Hynes MF, et al. Aneurysm
formation in expanded polytetrafluoroethylene prostheses. Surgery 1976; 79(5):491-493.

42. Roberts AK, Johnson N. Aneurysm formation in an expanded microporous polytetrafluor-
ethylene graft. Arch Surg 1978; 1 13(2):21 1-213.

43. Trippestad A. Dilatation and rupture of Dacron arterial grafts. Acta Chir Scand Suppl 1985;
529:77-79.

44. Berger K, Sauvage LR. Late fiber deterioration in Dacron arterial grafts. Ann Surg 1981;
193(4):477^191.

45. Nunn DB, Freeman MH, Hudgins PC. Postoperative alterations in size of Dacron aortic
grafts: An ultrasonic evaluation. Ann Surg 1979; 189(6):741-745.

46. Lundqvist B, Almgren B, Bowald S, Lorelius LE, Eriksson I. Deterioration and dilatation of
Dacron prosthetic grafts. Acta Chir Scand Suppl 1985; 529:81-85.

47. Berman SS, Hunter GC, Smyth SH, Erdoes LS, Mclntyre KE, Bernhard VM. Application of
computed tomography for surveillance of aortic grafts. Surgery 1995; 118(1):8-15.

48. Yashar JJ, Richman MH, Dyckman J, Witoszka M, Burnard RJ, Weyman AK, et al. Failure
of Dacron prostheses caused by structural defect. Surgery 1978; 84(5):659-663.

49. Ratto GB, Truini M, Sacco A, Canepa G, Badini A, Motta G Multiple aneurysmal dil-
atations in a knitted Dacron velour graft. J Cardiovasc Surg (Torino) 1985; 26(6):589-591.

50. King MW, Guidoin R, Blais P, Gayton A, Cunasekera KR. Degradation of polyester arterial
prostheses: A physical or chemical mechanism? In: Fraker AC, Griffen CD, eds. Corrosion

and Degradation of Implant Materials. Second Symposium, ASTM STP 859. Philadelphia: g

American Society for Testing and Materials, 1985:294-307. <S

51. Rais O, Lundstrom B, Angquist KA, Hallmans G Bilateral aneurysm of dacron graft fol- js
lowing aorto-femoral graft operation. A case report. Acta Chir Scand 1976; 142(6):479-482. °|

52. Clagett GP, Salander JM, Eddleman WL, Cabellon S Jr, Youkey JR, Olson DW, et al. <
Dilation of knitted Dacron aortic prostheses and anastomotic false aneurysms: Etiologic >9
considerations. Surgery 1983; 93(1 Pt 1):9— 16. J

53. Gooding GA, Effeney DJ, Goldstone J. The aortofemoral graft: Detection and identification q
of healing complications by ultrasonography. Surgery 1981; 89(1):94-101. |

54. Clifford PC, Skidmore R, Woodcock JP, Bird DR, Lusby RJ, Baird RN. Arterial grafts 2
imaged using Doppler and real-time ultrasound. Vase Diagn Ther 1981; 2:43-58. «

I
§

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 131

55. Qvarfordt PG, Reilly LM, Mark AS, Goldstone J, Wall SD, Ehrenfeld WK, et al. Com-
puterized tomographic assessment of graft incorporation after aortic reconstruction. Am J
Surg 1985; 150(2):227-231.

56. Brown OW, Stanson AW, Pairolero PC, Hollier LH. Computerized tomography following
abdominal aortic surgery. Surgery 1982; 91(6):716-722.

57. Bergan JJ, Yao JS, Flinn WR, McCarthy WJ. Surgical treatment of venous obstruction and
insufficiency. J Vase Surg 1986; 3( 1): 174-1 8 1 .

58. Nunn DB, Carter MM. Donohue MT, Pourdeyhimi B. Dilative characteristics of Microvel
and Vasculour-II aortic bifurcation grafts. J Biomed Mater Res 1996; 30(1):41 — 46.

59. Kalman PG, Rappaport DC, Merchant N, Clarke K, Johnston KW. The value of late
computed tomographic scanning in identification of vascular abnormalities after abdominal
aortic aneurysm repair. J Vase Surg 1999; 29(3):442-450.

60. Biancari F, Ylonen K, Anttila V, Juvonen J, Romsi P, Satta J, et al. Durability of open repair
of infrarenal abdominal aortic aneurysm: A 15-year follow-up study. J Vase Surg 2002;
35(l):87-93.

61. Edwards JM, Teefey SA, Zierler RE, Kohler TR. Intraabdominal paraanastomotic aneu-
rysms after aortic bypass grafting. J Vase Surg 1992; 15(2):344-353.

62. Mii S, Mori A, Sakata H, Kawazoe N. Para-anastomotic aneurysms: Incidence, risk factors,
treatment and prognosis. J Cardiovasc Surg (Torino) 1998; 39(3):259-266.

63. Locati P, Socrate AM, Costantini E. Paraanastomotic aneurysms of the abdominal aorta: A
15-year experience review. Cardiovasc Surg 2000; 8(4):274-279.

64. Curl GR, Faggioli GL, Stella A, D'Addato M, Ricotta JJ. Aneurysmal change at or above the
proximal anastomosis after infrarenal aortic grafting. J Vase Surg 1992; 16(6):855-860.

65. Allen RC, Schneider J, Longenecker L, Smith RB III, Lumsden AB. Paraanastomotic aneu-
rysms of the abdominal aorta. J Vase Surg 1993; 18(3):424^432.

66. Szilagyi DE, Smith RF, Elliott JP, Hageman JH, DaH'Olmo CA. Anastomotic aneurysms
after vascular reconstruction: Problems of incidence, etiology, and treatment. Surgery 1975;
78(6):800-816.

67. Kinley CE, Paasche PE, MacDonald AS, Marble AE. Stress at vascular anastomosis in
relation to host artery: Synthetic graft diameter. Surgery 1974; 75(l):28-30.

68. Kim GE, Imparato AM, Nathan I, Riles TS. Dilation of synthetic grafts and junctional
aneurysms. Arch Surg 1979; 1 14(1 1): 1296 1303.

69. Courbier R, Aboukhater R. Progress in the treatment of anastomotic aneurysms. World J
Surg 1988; 12(6)742-749.

70. Carson SN, Hunter GC, Palmaz J, Guernsey JM. Recurrence of femoral anastomotic aneu-
rysms. Am J Surg 1983; 146(6)774-778.

71. Nunn DB. Structural failure of first-generation, polyester, double-velour, knitted prostheses.
J Vase Surg 2001; 33(5):1 131-1 132.

72. Chakfe N, Riepe G, Dieval F, Le J, Magnen F, Wang L, Urban E, et al. Longitudinal
ruptures of polyester knitted vascular prostheses. J Vase Surg 2001; 33(5): 10 1 5—102 1 .

73. Biedermann H, Flora G Fatigue problems in dacron vascular grafts. Int J Artif Organs 1982;
5(3):205-206.

74. Watanabe T, Kusaba A, Kuma H, Kina M, Okadome K, Inokuchi K. Failure of Dacron arterial g
prostheses caused by structural defects. J Cardiovasc Surg (Torino) 1983; 24(2):95-100. <S

75. Goldstone J. Management of late failures of aorto-femoral reconstructions. Acta Chir Scand js
Suppl 1990; 555:149-153. Jf

76. Johnston KW. Nonvascular complications of vascular surgery. Presented at the Seventeenth "•
Annual Symposium on Current Critical Problems and New Horizons in Vascular Surgery, New >S
York, 1990. J

77. Wright DJ, Ernst CB, Evans JR, Smith RF, Reddy DJ, Shepard AD, et al. Ureteral com- q
plications and aortoiliac reconstruction. J Vase Surg 1990; 1 1(1):29— 37. |

78. Burns WA, Kanhouwa S, Tillman L, Saini N, Herrmann JB. Fibrosarcoma occurring at the S
site of a plastic vascular graft. Cancer 1972; 29(l):66-72. «

I
§

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

132 HUNTER and BULL

79. O'Connell TX, Fee HJ, Golding A. Sarcoma associated with Dacron prosthetic material: Case
report and review of the literature. J Thorac Cardiovasc Surg 1976; 72(1):94— 96.

80. Fehrenbacher JW, Bowers W, Strate R, Pittman J. Angiosarcoma of the aorta associated with
a Dacron graft. Ann Thorac Surg 1981; 32(3):297-301.

81. Brand KC. Foreign body tumorogenesis, timing and location of preneoplastic events. J Natl
Cancer Inst 1971; 47:829.

82. Cintora I, Pearce DE, Cannon JA. A clinical survey of aortobifemoral bypass using two
inherently different graft types. Ann Surg 1988; 208(5):625-630.

83. Bacchini G, Del Vecchio L, Andrulli S, Pontoriero G, Locatelli F. Survival of prosthetic grafts
of different materials after impairment of a native arteriovenous fistula in hemodialysis
patients. Asaio J 2001; 47(l):30-33.

84. Derenoncourt FJ. PTFE for A-V access: six years of experience with 310 reinforced and
stretch grafts. In: Henry ML, Ferguson RM, eds. Vascular Access for Hemodialysis — IV.
Chicago: Precept Press, 1995:286-291.

85. Landry GJ, Moneta GL, Taylor LM Jr, Porter JM. Axillobifemoral bypass. Ann Vase Surg
2000; 14(3):296-305.

86. Boyce B. Physical characteristics of expanded polytetrafluoroethylene grafts. In: Stanley JC,
ed. Biologic and Synthetic Vascular Prostheses. New York: Grune and Stratton, 1982:553-
561.

87. Graham LM, Bergan JJ. Expanded polytetrafluoroethylene vascular grafts: Clinical and ex-
perimental observation. In: Stanley JC, ed. Biologic and Synthetic Vascular Prostheses. New
York: Grune and Stratton, 1982:536-586.

88. Clowes AW, Gown AM, Hanson SR, Reidy MA. Mechanisms of arterial graft failure. 1. Role
of cellular proliferation in early healing of PTFE prostheses. Am J Pathol 1985; 118(1):43-
54.

89. Clowes AW, Kirkman TR, Clowes MM. Mechanisms of arterial graft failure. II Chronic
endothelial and smooth muscle cell proliferation in healing polytetrafluoroethylene pros-
theses. J Vase Surg 1986; 3(6):877-884.

90. Quinones-Baldrich WJ, Ziomek S, Henderson T, Moore WS. Primary anastomotic bonding in
polytetrafluoroethylene grafts? J Vase Surg 1987; 5(2):31 1-318.

91. Mohring K, Osbach HW, Bersch W, Ikinger U, Schiller PL Clinical implications of patho-
morphological findings in vascular prostheses. In: Robinson BHB, Hawkins JB, eds. Dialysis,
Transplantation Nephrology. London: Pitman, 1978:582-583.

92. Guidoin R, Chakfe N, Maurel S, How T, Batt M, Marois M, et al. Expanded polytetra-
fluoroethylene arterial prostheses in humans: Histopathological study of 298 surgically excised
grafts. Biomaterials 1993; 14(9):678-693.

93. Guidoin R, Maurel S, Chakfe N, How T, Zhang Z, Therrien M, et al. Expanded polytetra-
fluoroethylene arterial prostheses in humans: Chemical analysis of 79 explanted specimens.
Biomaterials 1993; 14(9):-704694.

94. Anderson JM, Bennert KW, Johnson JM. The pathology and healing responses of expanded
polytetrafluoroethylene vascular access grafts. In: Wilson SE, ed. Vascular Access Surgery.
2nd ed. Chicago: Year Book, 1988:213-231.

95. Groegler FM, Kapfer X, Meichelbock W. Crural prosthetic revascularization: Randomized, j>
prospective, multicentric comparison of standard and carbon impregnated ePTFE grafts. 27th <S
Global Vascular Endovascular Issues Techniques Horizons Symposium, New York, 2000:12.1- js
12.3. Jf

96. Bacourt F. Prospective randomized study of carbon-impregnated polytetrafluoroethylene **
grafts for below-knee popliteal and distal bypass: Results at 2 years. The Association >9
Universitaire de Recherche en Chirurgie. Ann Vase Surg 1997; ll(6):596-603. J

97. Mohamed MS, Mukherjee M, Kakkar VV. Thrombogenicity of heparin and non-heparin- q

bound arterial prostheses: An in vitro evaluation. J R Coll Surg Edinb 1998; 43(3): 155— |

157. S

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 133

98. Iwai Y. Development of a thermal cross-linking heparinization method and its application to
small caliber vascular prostheses. Asaio J 1996; 42(5):M693-M697.

99. Devine C, Hons B, McCollum C. Heparin-bonded Dacron or polytetrafluoroethylene for
femoropopliteal bypass grafting: A multicenter trial. J Vase Surg 2001; 33(3):533-539.

100. Herring M, Baughman S, Glover J. Endothelium develops on seeded human arterial pros-
thesis: A brief clinical note. J Vase Surg 1985; 2(5):727-730.

101. Swedenborg J, Bengtsson L, Clyne N, Dryjski M, Gillis C, Rosfors S, et al. In vitro endo-
thelialisation of arteriovenous loop grafts for haemodialysis. Eur J Vase Endovasc Surg 1997;
13(3)272-277.

102. Magometschnigg H, Kadletz M, Vodrazka M, Dock W, Grimm M, Grabenwoger M, et al.
Prospective clinical study with in vitro endothelial cell lining of expanded polytetrafluoro-
ethylene grafts in crural repeat reconstruction. J Vase Surg 1992; 1 5(3):527— 535.

103. Deutsch M, Meinhart J, Vesely M, Fischlein T, Groscurth P, von Oppell U, et al. In vitro
endothelialization of expanded polytetrafluoroethylene grafts: A clinical case report after 41
months of implantation. J Vase Surg 1997; 25(4):757-763.

104. Yu H, Wang Y, Eton D, Rowe VL, Terramani TT, Cramer DV, et al. Dual cell seeding and
the use of zymogen tissue plasminogen activator to improve cell retention on polytetrafluoro-
ethylene grafts. J Vase Surg 2001; 34(2):337-343.

105. Zilla P, Deutsch M, Meinhart J, Fischlein T, Hofmann G. Long-term effects of clinical in vitro
endothelialization on grafts. J Vase Surg 1997; 25(6): 1 110-1 112.

106. Sipehia R, Martucci G, Lipscombe J. Transplantation of human endothelial cell monolayer
on artificial vascular prosthesis: The effect of growth-support surface chemistry, cell seeding
density, ECM protein coating, and growth factors. Artif Cells Blood Substit Immobil
Biotechnol 1996; 24(l):51-63.

107. Bellon JM, Garcia-Honduvilla N, Escudero C, Gimeno MJ, Contreras L, de Haro J, et al.
Mesothelial versus endothelial cell seeding: Evaluation of cell adherence to a fibroblastic
matrix using 111-In oxine. Eur J Vase Endovasc Surg 1997; 13(2): 142—148.

108. Shi Q, Wu MH, Hayashida N, Wechezak AR, Clowes AW, Sauvage LR. Proof of fallout
endothelialization of impervious Dacron grafts in the aorta and inferior vena cava of the dog.
J Vase Surg 1994; 20(4):546-557.

109. Scott SM, Barth MG, Gaddy LR, Ahl ET Jr, The role of circulating cells in the healing of
vascular prostheses. J Vase Surg 1994; 19(4):585-593.

1 10. Birchall IE, Field PL, Ketharanathan V. Adherence of human saphenous vein endothelial cell
monolayers to tissue-engineered biomatrix vascular conduits. J Biomed Mater Res 2001;
56(3):437^t43.

111. Wu MH, Shi Q, Wechezak AR, Clowes AW. Gordon IL, Sauvage LR. Definitive proof of
endothelialization of a Dacron arterial prosthesis in a human being. J Vase Surg 1995;
21(5):862-867.

112. Berman SS, Jarrell BE, Raymond MA, Kleinert L, Williams SK. Early experience with
ePTFE dialysis grafts sodded with liposuction-derived microvascular endothelial cells. Henry
ML, Ferguson RM, eds. Vascular Access for Hemodialysis 1995; Vol IV:. Chicago: Precept
Press, 1995:292-302.

113. Kraiss LW, Kirkman TR, Kohler TR, Zierler B, Clowes AW. Shear stress regulates smooth g
muscle proliferation and neointimal thickening in porous polytetrafluoroethylene grafts. |
Arterioscler Thromb 1991; 1 1(6): 1844-1852. g

1 14. Mattsson EJ, Kohler TR, Vergel SM, Clowes AW. Increased blood flow induces regression of °|
intimal hyperplasia. Arterioscler Thromb Vase Biol 1997; 17(10)2245-2249. ^

115. Batson RC, Sottiurai VS, Craighead CC. Linton patch angioplasty. An adjunct to distal «
bypass with polytetrafluoroethylene grafts. Ann Surg 1984; 199(6):684-693. J

116. Taylor RS, Loh A, McFarland RJ, Cox M, Chester JF. Improved technique for poly- q
tetrafluoroethylene bypass grafting: Long-term results using anastomotic vein patches. Br J |
Surg 1992; 79(4): 348-354. S

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

134 HUNTER and BULL

117. Stonebridge PA, Prescott RJ, Ruckley CV. Randomized trial comparing infrainguinal poly-
tetrafluoroethylene bypass grafting with and without vein interposition cuff at the distal
anastomosis. The Joint Vascular Research Group. J Vase Surg 1997; 26(4):543-550.

1 18. How TV, Rowe CS, Gilling-Smith GL, Harris PL. Interposition vein cuff anastomosis alters
wall shear stress distribution in the recipient artery. J Vase Surg 2000; 31(5): 1008—1017.

119. Kissin M, Kansal N, Pappas PJ, DeFouw DO, Duran WN, Hobsonll RW II. Vein inter-
position cuffs decrease the intimal hyperplastic response of polytetrafluoroethylene bypass
grafts. J Vase Surg 2000; 31(1 Pt l):69-83.

120. Neville RF, Tempesta B, Sidway AN. Tibial bypass for limb salvage using polytetrafluoro-
ethylene and a distal vein patch. J Vase Surg 2001; 33(2):266-272.

121. Kreienberg PB, Darling RC III, Chang BB, Paty PS, Lloyd WE, Shah DM. Adjunctive
techniques to improve patency of distal prosthetic bypass grafts: Polytetrafluoroethylene with
remote arteriovenous fistulae versus vein cuffs. J Vase Surg 2000; 31(4):696-701.

122. Harris PL, How TV. Haemodynamics of cuffed arterial anastomosis. Crit Ischaemia
9(l)1999;:20-26.

123. Panneton JM. Early results of the Distaflo randomized trial for critical limb ischemia 27th
Global Vascular Endovascular Issues Techniques Horizons Symposium, New York. 2000:14.1-
14.2.

124. Lord RSA, Nash PA, Raj BT, Stary DL, Graham AR, Hill DA, et al. Prospective randomized
trial of polytetrafluoroethylene and Dacron aortic prosthesis. I Perioperative results. Ann
Vase Surg 2(3)1988;:248-254.

125. Friedman SG, Lazzaro RS, Spier LN, Moccio C, Tortolani AL A prospective randomized
comparison of Dacron and polytetrafluoroethylene aortic bifurcation grafts. Surgery 1995;
117(1):7 10.

126. Chiesa R, Melissano G, Castellano R, Frigerio S. Extensible expanded polytetrafluoro-
ethylene vascular grafts for aortoiliac and aortofemoral reconstruction. Cardiovasc Surg 2000;
8(7):538-544.

127. Schuman ES, Standage BA, Ragsdale JW, Gross GF. Reinforced versus nonreinforced
polytetrafluoroethylene grafts for hemodialysis access. Am J Surg 1997; 173(5):407^10.

128. Soroma AJ, Hughes CB, McCarthy JT, Jenson BM, Prieto M, Panneton JM, Sterioff S,
Stegall MD, Nyberg SL. Prospective, randomized evaluation of a cuffed expanded polytetra-
fluoroethylene graft for hemodialysis vascular access. Surgery 2002; 132:135-140.

129. Hakaim AG, Scott TE. Durability of early prosthetic dialysis graft cannulation: Results of a
prospective, nonrandomized clinical trial. J Vase Surg 1997; 25(6): 1002-1006.

130. Palder SB, Kirkman RL, Whittemore AD, Hakim RM, Lazarus JM, Tilney NL. Vascular
access for hemodialysis. Patency rates and results of revision. Ann Surg 1985; 202(2):235-239.

131. Johnson JM, Anderson JM. Reasonable expectations for PTFE grafts in hemodialysis access.
Dialysis Transplant 1983; 12(4):238-240.

132. Escobar FS, Schwartz SA, Abouljoud M, Douzdjian V, Escobar MD, Besarab A, Elliott JP.
Comparison of a new "hooded" graft with a conventional ePTFE graft: A preliminary study.
In: Henry ML, ed. Vascular Access for Hemodialysis- VI. Chicago: WL Gore & Assoc and
Precept Press, 1999:205-212.

133. Pipinos II, Escobar FSI, Anagnostopoulos PV, Elliott JP, Schwartz S. Early experience with P
Taylor and notched vein patching in the construction of arteriovenous hemodialysis access <S
procedures: Results of a prospective study. Henry ML ed. Vascular Access for Hemodialysis js
1998; Vol VI:. Chicago: Precept Press, 1998:213-222. Jf

134. Enzler MA, Rajmon T, Lachat M, Largiader F. Long-term function of vascular access for ^
hemodialysis. Clin Transplant 1996; 10(6 Pt 1):51 1-515. |

135. Gibson KD, Gillen DL, Caps MT, Kohler TR, Sherrard DJ, Stehman-Breen CO. Vascular J
access survival and incidence of revisions: A comparison of prosthetic grafts, simple autoge- q
nous fistulas, and venous transposition fistulas from the United States Renal Data System |
Dialysis Morbidity and Mortality Study. J Vase Surg 2001; 34(4):694-700. 2

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 135

136. Graor RA, Risius B, Denny KM, Young JR, Beven EG, Hertzer NR, et al. Local throm-
bolysis in the treatment of thrombosed arteries, bypass grafts, and arteriovenous fistulas.
J Vase Surg 1985; 2(3):406-414.

137. Benenati J, Shlansky-Goldberg R, Meglin A, Seidl E. Thrombolytic and antiplatelet therapy
in peripheral vascular disease with use of reteplase and/or abciximab. The SCVIR Consul-
tants' Conference; May 22, 2000; Orlando, FL Society for Cardiovascular and Interventional
Radiology. J Vase Interv Radiol 2001; 12(7):795-805.

138. Rubin JR, Yao JS, Thompson RG, Bergan JJ. Management of infection of major amputation
stumps after failed femorodistal grafts. Surgery 1985; 98(4):810-815.

139. Kaupp HA, Matulewicz TJ, Lattimer GL, Kremen JE, Celani VJ. Graft infection or graft
reaction? Arch Surg 1979; 1 14(12): 1419-1422.

140. Paes E, Vollmar JF, Mohr W, Hamann H, Brecht-Krauss D. Perigraft reaction: Incom-
patibility of synthetic vascular grafts? New aspects on clinical manifestation, pathogenesis,
and therapy. World J Surg 1988; 12(6)750-755.

141. Blumenberg RM, Gelfand ML, Dale WA. Perigraft seromas complicating arterial grafts.
Surgery 1985; 97(2): 194-204.

142. Buche M, Schoevaerdts JC, Jaumin P, Ponlot R, Chalant CH. Perigraft seroma following
axillofemoral bypass: Report of three cases. Ann Vase Surg 1986; 1(3):374— 377.

143. Bhuta I, Dorrough R. Noninfectious fluid collection around velour Dacron graft: Possible
allergic reaction. South Med J 1981; 74(7):870-872.

144. Ahn SS, Machleder HI, Gupta R, Moore WS. Perigraft seroma: Clinical, histologic, and
serologic correlates. Am J Surg 1987; 154(2): 173-178.

145. Bolton W, Cannon JA. Seroma formation associated with PTFE vascular grafts used as
arteriovenous fistulae. Dial Transplant 1981; 10(1):60 (62-63,66).

146. Berger RMF, Bol-Raap G, Hop WJ, Bogers AJ, Hess J. Heparin as a risk factor for perigraft
seroma complicating the modified Blalock-Taussig shunt. J Thorac Cardiovasc Surg 1998;
116(2):286-293.

147. LeBlanc JG, Vince DJ, Taylor GP. Perigraft seroma: Long-term complications. J Thorac
Cardiovasc Surg 1986; 92(3 Pt l):451-454.

148. Ahn SS, Williams DE, Thye DA, Cheng KQ, Lee DA. The isolation of a fibroblast growth
inhibitor associated with perigraft seroma. J Vase Surg 1994; 20(2):202-208.

149. Sladen JG, Mandl MA, Grossman L, Denegri JF. Fibroblast inhibition: A new and treatable
cause of prosthetic graft failure. Am J Surg 1985; 149(5):587-590.

150. Schneiderman J, Knoller S, Adar R, Savion N. Biochemical analysis of a human humoral
fibroblast inhibitory factor associated with impaired vascular prosthetic graft incorporation. J
Vase Surg 1991; 14(1): 103 1 10.

151. Lowery RC Jr, Wicker HS, Sanders K, Peniston RL. Management of a recalcitrant peri-
prosthetic fluid collection. J Vase Surg 1987; 6(1)77-80.

152. Moore WS, Hall AD. Late suture failure in the pathogenesis of anastomotic false aneurysms.
Ann Surg 1970; 172(6):1064-1068.

153. Cutler EC, Dunphy JE. The use of silk in infected wounds. N Engl J Med 1941; 224:101-107.

154. Dobrin PB. Surgical manipulation and the tensile strength of polypropylene sutures. Arch

Surg 1989; 124(6):665-668. 1

155. Calhoun TR, Kitten CM. Polypropylene suture— is it safe? J Vase Surg 1986; 4(1):98-100. |

156. Myhre OA. Breakage of prolene suture. Ann Thorac Surg 1983; 36(1):121. |

157. Szarnicki RJ. Polypropylene suture fracture. Ann Thorac Surg 1983; 35(3):333. £

158. Aldrete V. Polypropylene suture fracture. Ann Thorac Surg 1984; 37(3):264. ^

159. Setzen G, Williams EF. Tissue response to suture materials implanted subcutaneously in a >9
rabbit model. Plast Reconstr Surg 1997; 100(7): 1788-1795. J

160. Gayle RG, Wheeler JR, Gregory RT, Snyder SO Jr. Evaluation of the expanded poly- q
tetrafluoroethylene (EPTFE) suture in peripheral vascular surgery using EPTFE prosthetic |
vascular grafts. J Cardiovasc Surg (Torino) 1988; 29(5):556-559. S

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

136 HUNTER and BULL

161. Moore WS, Kempczinski RF, Nelson JJ, Toole JF. Recurrent carotid stenosis: Results of the
asymptomatic carotid atherosclerosis study. Stroke 1998; 29(10):2018-2025.

162. O'Hara PJ, Hertzer NR, Mascha EJ, Krajewski LP, Clair DG, Ouriel K. A prospective,
randomized study of saphenous vein patching versus synthetic patching during carotid
endarterectomy. J Vase Surg 2002; 35(2):324-332.

163. Hayes PD, Allroggen H, Steel S, Thompson MM, London NJ, Bell PR, et al. Randomized
trial of vein versus Dacron patching during carotid endarterectomy: Influence of patch type
on postoperative embolization. J Vase Surg 2001; 33(5):994-1000.

164. AbuRahma AF, Khan JH, Robinson PA, Saiedy S, Short YS, Boland JP, et al. Prospective
randomized trial of carotid endarterectomy with primary closure and patch angioplasty with
saphenous vein, jugular vein, and polytetrafluoroethylene: Perioperative (30-day) results. J
Vase Surg 1996; 24(6):998-1007.

165. AbuRahma AF, Robinson PA, Saiedy S, Kahn JH, Boland JP. Prospective randomized trial
of carotid endarterectomy with primary closure and patch angioplasty with saphenous vein,
jugular vein, and polytetrafluoroethylene: Long-term follow-up. J Vase Surg 1998; 27(2):222-
234.

166. AbuRahma AF, Hannay RS, Khan JH, Robinson PA, Hudson JK, Davis EA. Prospective
randomized study of carotid endarterectomy with polytetrafluoroethylene versus collagen-
impregnated Dacron (Hemashield) patching: perioperative (30-day) results. J Vase Surg 2002;
35(1):125-130.

167. AbuRahma AF, Robinson PA, Hannay RS, Hudson J, Cutlip L. Prospective controlled study
of carotid endarterectomy with Hemashield patch: Is it thrombogenic?. Vase Surg 2001;
35(3): 167-1 74.

168. Archie JP. Carotid endarterectomy outcome with vein or Dacron graft patch angioplasty and
internal carotid artery shortening. J Vase Surg 1999; 29(4):654-664.

169. Shi Q, Wu MH, Sauvage LR. Clinical and experimental demonstration of complete healing of
porous Dacron patch grafts used for closure of the arteriotomy after carotid endarterectomy.
Ann Vase Surg 1999; 13(3):313-3I7.

170. Grimsley BR, Wells JK, Pearl GJ, Garrett WV, Shutze WP, Talkington CM, et al. Bovine
pericardial patch angioplasty in carotid endarterectomy. Am Surg 2001; 67(9):890-895.

171. Biasi GM, Sternjakob S, Mingazzini PM, Ferrari SA. Nine-year experience of bovine peri-
cardium patch angioplasty during carotid endarterectomy. J Vase Surg 2002; 36(2):271-
277.

172. Rockman CB, Riles TS, Landis R, Lamparello PJ, Giangola G, Adelman MA, et al. Redo
carotid surgery: An analysis of materials and configurations used in carotid reoperations and
their influence on perioperative stroke and subsequent recurrent stenosis. J Vase Surg 1999;
29(1):72-81.

173. Lammer J, Dake MD, Bleyn J, Katzen BT, Cejna M, Piquet P, et al. Peripheral arterial
obstruction: Prospective study of treatment with a transluminally placed self-expanding stent-
graft. International Trial Study Group. Radiology 2000; 217(1 ):95-104.

174. Marin ML, Veith FJ, Cynamon J, Sanchez LA, Bakal CW, Suggs WD, et al. Human trans-
luminally placed endovascular stented grafts: preliminary histopathologic analysis of healing

grafts in aortoiliac and femoral artery occlusive disease. J Vase Surg 1995; 21(4):595-604. g

175. van Sambeek MR, Hagenaars T, Gussenhoven EJ, Leertouwer TC, van der Lugt A, Hoedt <S
MT, et al. Vascular response in the femoropopliteal segment after implantation of an ePTFE js
balloon-expandable endovascular graft: An intravascular ultrasound study. J Endovasc Ther °|
2000; 7(3):204-212. ^

176. Sarkar R, Eilber FR, Gelabert HA, Quinones-Baldrich WJ. Prosthetic replacement of the «
inferior vena cava for malignancy. J Vase Surg 1998; 28(l):75-83. jjj

177. Alimi YS, Gloviczki P, Vrtiska TJ, Pairolero PC, Canton LG, Bower TC, et al. Recon- q

struction of the superior vena cava: Benefits of postoperative surveillance and secondary |

endovascular interventions. J Vase Surg 1998; 27(2):287-301. 2

©

!

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HEALING OF PROSTHESES 137

178. Rypins EB, Conroy RM, Sarfeh IJ. Advantages and disadvantages of polytetrafluoroethylene
(PTFE) grafts for portacaval shunting. Vase Surg 1988; 22(2):88-92.

179. Sarfeh IJ, Rypins EB, Mason GR. A systematic appraisal of portacaval H-graft diameters.
Clinical and hemodynamic perspectives. Ann Surg 1986; 204(4):356-363.

180. Sarfeh IJ, Rypins EB. Partial versus total portacaval shunt in alcoholic cirrhosis. Results of a
prospective, randomized clinical trial. Ann Surg 1994; 219(4):353— 361.

181. Sanders RJ, Rosales C, Pearce WH. Creation and closure of temporary arteriovenous fistulas
for venous reconstruction or thrombectomy: Description of technique. J Vase Surg 1987; 6(5):
504-505.

182. Kogel H, Vollmar JF, Cyba-Altunbay S, Mohr W, Frosch D, Amselgruber W. New obser-
vations on the healing process in prosthetic substitution of large veins by microporous
grafts — animal experiments. Thorac Cardiovasc Surg 1989; 37(2): 119-124.

183. Heydorn WH, Geasling JW, Moores WY, Lollini LO, Gomez AC. Changes in the manu-
facture of expanded microporous polytetrafluoroethylene: Effects on patency and histologi-
cal behavior when used to replace the superior vena cava. Ann Thorac Surg 1979; 27(2):
173-177.

184. Theuer CJ, Bergamini TM, Theuer HH, Burns CD, Proctor ML, Garrison RN. Vena cava
replacement with a peritoneum-lined vascular graft. Asaio J 1996; 42(4):266-270.

185. Glickman MH, Stokes GK, Ross JR, Schuman ED, Sternbergh WC III, Lindberg JS, et al.
Multicenter evaluation of a polytetrafluoroethylene vascular access graft as compared with the
expanded polytetrafluoroethylene vascular access graft in hemodialysis applications. J Vase
Surg 2001; 34(3):465^473.

186. King MW, Zhang Z, Ukpabi P, Murphy D, Guidoin R. Quantitative analysis of the surface
morphology and textile structure of the polyurethane Vascugraft arterial prosthesis using
image and statistical analyses. Biomaterials 1994; 15(8):621-627.

187. Huang B, Marois Y, Roy R, Julien M, Guidoin R. Cellular reaction to the Vascugraft
polyesterurethane vascular prosthesis: in vivo studies in rats. Biomaterials 1992; 13(4):209-
216.

188. Zhang Z, King MW, Guidoin R, Therrien M, Pezolet M, Adnot A, et al. Morphological,
physical and chemical evaluation of the Vascugraft arterial prosthesis: Comparison of a novel
polyurethane device with other microporous structures. Biomaterials 1994; 15(7):483-501.

189. Bull PG, Denck H, Guidoin R, Gruber H. Preliminary clinical experience with polyurethane
vascular prostheses in femoro-popliteal reconstruction. Eur J Vase Surg 1992; 6(2):217-224.

190. Zhang Z, Marois Y, Guidoin RG, Bull P, Marois M, How T, et al. Vascugraft polyurethane
arterial prosthesis as femoro-popliteal and femoro-peroneal bypasses in humans: Patho-
logical, structural and chemical analyses of four excised grafts. Biomaterials 1997; 18(2):113-
124.

191. Stanley JC, Lindenauer SM, Graham LM, Zelenock GB, Wakefield TW, Cronenwett JL.
Biologic and synthetic vascular grafts. In: Moore WS, ed. Vascular Surgery: A Comprehensive
Review. Philadelphia: Saunders, 1990:275-294.

192. Dardik H, Miller N, Dardik A, Ibrahim I, Sussman B, Berry SM, et al. A decade of experience
with the glutaraldehyde-tanned human umbilical cord vein graft for revascularization of the

lower limb. J Vase Surg 1988; 7(2):336-346. 1

193. Dardik H, Ibrahim IM, Sussman B, Kahn M, Sanchez M, Klausner S, et al. Biodegradation <S
and aneurysm formation in umbilical vein grafts. Observations and a realistic strategy. Ann js
Surg 1984; 199(l):61-68. Jf

194. Guidoin R, Gagnon Y, Roy PE, Marois M, Johnston KW, Batt M. Pathologic features of <
surgically excised human umbilical vein grafts. J Vase Surg 1986; 3(1): 146-1 54. >9

195. Batt M, Gagliardi JM, Avril G, Guzman R, Guidoin R, Hassen-Khodja R, et al. Human J
umbilical vein grafts as infrainguinal bypasses: Long-term clinical follow-up and pathological q
investigation of explanted grafts. Clin Invest Med 1990; 13(4): 155-1 64. 1

196. Cranely JJ, Karkow WS, Hafner CD, Flanagan LD. Aneurysmal dilatation in umbilical vein S

©

!

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

138 HUNTER and BULL

grafts. In: Yao JST, Bergan JJ, eds. Reoperative Arterial Surgery. New York: Grune and
Straton, 1986:343-358.

197. Dardik H, Wengerter K, Qin F, Pangilinan A, Silvestri F, Wolodiger F, et al. Comparative
decades of experience with glutaraldehyde-tanned human umbilical cord vein graft for lower
limb revascularization: An analysis of 1275 cases. J Vase Surg 2002; 35(1):64— 71.

198. Matsuura JH, Johansen KH, Rosenthal D, Clark MD, Clarke KA, Kirby LB. Cryopreserved
femoral vein grafts for difficult hemodialysis access. Ann Vase Surg 2000; 14(1):50— 55.

199. Johnson TR, Tomaszewski JE, Carpenter JP. Cellular repopulation of human vein allograft
bypass grafts. J Vase Surg 2000; 31(5):994-1002.

200. Benedetto B, Lipkowitz G, Madden R, Kurbanov A, Hull D, Miller M, et al. Use of
cryopreserved cadaveric vein allograft for hemodialysis access precludes kidney trans-
plantation because of allosensitization. J Vase Surg 2001; 34(1):139-142.

201. Dalsing MC, Raju S, Wakefield TW, Taheri S. A multicenter, phase I evaluation of cryo-
preserved venous valve allografts for the treatment of chronic deep venous insufficiency. J
Vase Surg 1999; 30(5):854-864.

202. Harris RW, Schneider PA, Andros G, Oblath RW, Salles-Cunha S, Dulawa L. Allograft vein
bypass: Is it an acceptable alternative for infrapopliteal revascularization? J Vase Surg 1993;
18(4):553-560.

203. Buckley CJ, Abernathy S, Lee SD, Arko FR, Patterson DE, Manning LG. Suggested treat-
ment protocol for improving patency of femoral-infrapopliteal cryopreserved saphenous vein
allografts. J Vase Surg 2000; 32(4):73 1-738.

204. Harris L, O'Brien-Irr M, Ricotta JJ. Long-term assessment of cryopreserved vein bypass
grafting success. J Vase Surg 2001; 33(3):528-532.

205. Vogt PR, Brunner-La Rocca HP, Carrel T, von Segesser LK, Ruef C, Debatin J, et al.
Cryopreserved arterial allografts in the treatment of major vascular infection: A comparison
with conventional surgical techniques. J Thorac Cardiovasc Surg 1998; 116(6):965-972.

206. Leseche G, Castier Y, Petit MD, Bertrand P, Kitzis M, Mussot S, et al. Long-term results of
cryopreserved arterial allograft reconstruction in infected prosthetic grafts and mycotic
aneurysms of the abdominal aorta. J Vase Surg 2001; 34(4):6 16-622.

207. Rosenberg N. Dialdehyde starch tanned bovine heterografts. In: Sawyer PN, Kaplitt MJ, eds.
Vascular Grafts. New York: Appleton-Century-Crofts, 1978:261-270.

208. Brems J, Castaneda M, Garvin PL A five-year experience with the bovine heterograft for
vascular access. Arch Surg 1986; 121(8):941-944.

209. Andersen RC, Ney AL, Madden MC, LaCombe MJ. Biologic conduits for vascular access:
Saphenous veins, umbilical veins, bovine carotid arteries. In: Sommer BG, Henry ML, eds.
Vascular Access for Hemodialysis. Chicago: Pluribus Press, 1989:65-83.

210. Sabanayagam P, Schwartz AB, Soricelli RR, Lyons P, Chinitz J. A comparative study of 402
bovine heterografts and 225 reinforced expanded PTFE grafts as AVF in the ESRD patient.
Trans Am Soc Artif Intern Organs 1980; 26:88-92.

211. Hurt AV, Batello-Cruz M, Skipper BJ, Teaf SR, Sterling WA Jr, Bovine carotid artery
heterografts versus polytetrafluoroethylene grafts. A prospective, randomized study. Am J
Surg 1983; 146(6):844-847.

212. Reese JC, Esterl R, Lindsey L, Aridge D, Solomon H, Fairchild RB, et al. A prospective j>
randomized comparison of bovine heterografts versus Impra grafts for chronic hemodialysis. <S
In: Henry ML, Ferguson RM, eds. Vascular Access for Hemodialysis — III. Chicago: Precept js
Press, 1993:157-163. 1

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

6

Anastomotic Aneurysms

Alexander D. Shepard and Gary M. Jacobson*

Henry Ford Hospital, Detroit, Michigan, U.S.A.

An anastomotic aneurysm (AA) is an aneurysm that occurs at an anastomotic interface

between a graft and an artery. Although such aneurysms may rarely involve autogenous

grafts, they are almost exclusively the result of prosthetic grafting procedures. Historically

such aneurysms are believed to be false aneurysms resulting from an anastomotic defect.

Extravasation of blood through such a defect leads to a reactive inflammatory response by

surrounding tissues and formation of a fibrous capsule. As a result of sustained arterial

pressure, the anastomotic defect and associated soft tissue capsule gradually expand

resulting in the formation of a false aneurysm. Recently, it has been recognized that an

increasing number of AAs may in fact be true aneurysms occurring at the junction of a

prosthetic graft and an artery. In this situation the aneurysm results from degeneration of

the native arterial wall; this aneurysmal degeneration may be primary (due to inherent

structural abnormalities in the arterial wall) or secondary to changes induced by the

prosthetic grafting procedure. True aneurysms involving the artery adjacent to an

anastomosis have also been termed juxta-anastomotic and lumped together with the more

traditional false AAs using the term para-anastomotic. In practice it is frequently difficult

to distinguish whether such aneurysms are true or false in nature and the diagnosis and

treatment are usually very similar. In the present discussion, the term anastomotic

aneurysm (AA) is used to refer to any aneurysm (true or false) occurring at an

anastomosis. Regardless of etiology most AAs require repair because of the risk of -a

complications (e.g., rupture, thrombosis) and the resulting potential for loss of life or limb. |

1
I. ETIOLOGY AND PATHOGENESIS |

Anastomotic aneurysms may occur as a result of numerous mechanisms or circumstances &

(Table 1). Typically, there is a failure of technique, graft/suture material, or host arterial wall jj

'Current affiliation: Vascular Surgical Associates, P.C., Marietta, Georgia, U.S.A.

139 |

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

140 SHEPARD and JACOBSON

Table 1 Proposed Etiological Factors

Patient factors

Native artery disease

Infection

Smoking

Diabetes

Hypertension

Healing complications (e.g., seroma, hematoma)
Material factors

Graft defects

Suture degradation or breakage

Prosthetic graft-arterial wall compliance mismatch
Technical factors

Inadequate suture bites

Excessive tension

Joint motion

"Redo" procedure

Endarterectomy

integrity. The most common cause appears to be degeneration of the native artery wall,
leading to weakness, fragmentation, and eventual dehiscence of the intact suture line (1). In
Szilagyi's 1975 landmark paper on AAs, "deficiency of the arterial wall" was the primary
causative factor in more than 30% of patients (1). In essence, the anastomotic sutures simply
pull through the degenerated wall — an outcome corroborated by the frequent operative
findings of an intact suture line attached to the end of the graft that has completely separated
from the native artery. A number of factors can contribute to this degeneration, including
aneurysmal disease in the grafted artery, endarterctomy, and hypertension (1-3).

Other patient factors implicated in the formation of AAs include gender, original
operative indication, and perioperative complications. In most series, males outnumber
females by 9 or 10 to 1, although female gender appears to be a risk factor for recurrent
AA (2,4). Patients undergoing aortofemoral bypass for occlusive disease appear to have a
higher incidence of AA than patients with aneurysmal disease (2,5,6). Perioperative
complications, both local and systemic, are associated with increased rates of AA
(1,4,6). Wound complications (e.g., infection, seroma, or lymph leaks) can lead to perigraft
fluid collections which prevent full graft incorporation (7). Connective tissue disorders
(e.g., Marfan's syndrome) have also been implicated in the genesis of AA following
thoracoabdominal aneurysm (TAAA) repair (8).

Infection is an important factor in the development of AAs. By attacking the graft-
native artery interface, bacterial proteolytic enzymes weaken and degrade the arterial wall. ■a
The possibility of a graft infection should always be considered whenever a suture line §
aneurysm is encountered. A proximal aortic A A should always alert the clinician to the a
possibility of an aortoenteric fistula. Seabrook and colleagues have identified occult graft c
infection as a frequent cause of groin AA (9). They noted a high rate of culture-positive <
pseudoaneurysms in the absence of clinical signs of infection. Coagulase-negative gram- &
positive cocci such as Staphylococcus species were the most common isolates. These J
bacteria have the ability to form biofilms and produce proteolytic enzymes and other Q
destructive substances that degrade host tissues. AAs resulting from infection tend to |
occur earlier after grafting than do noninfected AAs (3-4 vs. 5-6 years) (1,10). @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ANASTOMOTIC ANEURYSMS 141

Material defects, both in graft and suture, are important contributing factors in the
formation of AAs. Historically, degradable suture material (silk, braided polyester) is the
best-recognized cause of anastomotic disruption (11). In the era of silk suture use (prior to
the late 1960s), suture breakdown over time was common, occurring in up to one-quarter
of anastomoses. With the advent of modern polypropylene monofilament suture, primary
failures are uncommon. However, improper handling of modern monofilament sutures
can lead to filament fracture and suture failure. Early textile grafts, which were prone to
fragmentation and disruption, are no longer in use, although patients who received them
remain at risk for AA development.

Prosthetic graft dilatation predisposes to AA formation. According to LaPlace's law,
wall tension increases with increasing vessel diameter. An end-to-side anastomosis
effectively doubles the diameter of an artery at the anastomotic site. Dilation of the graft
leads to further increases in the effective vessel diameter and may tip the balance of forces
in favor of aneurysm formation. Dilatation of Dacron grafts, particularly the knitted
designs, is well recognized (12,13). In contrast, expanded polytetraflouroethylene (ePTFE)
grafts do not tend to dilate as much (12). Compliance mismatches may subject the
anastomosis to forces that predispose to aneurysm formation (14,15). With a relatively
stiff, noncompliant prosthesis, there may be preferential dilatation of the artery, resulting
in potentially disruptive stresses on the anastomosis. Prosthetic graft materials never
completely heal with the adjacent blood vessel. As such, prosthetic graft-to-native artery
anastomoses are always more prone to disruption. Autogenous tissues are more resistant
to AA complications but are still susceptible if conditions are otherwise unfavorable (e.g.,
infected field, anastomosis under tension, etc.).

Technical problems are another well-recognized cause of AA. Proper anastomotic tech-
nique requires that suture bites pass through a full thickness of healthy arterial wall as well as
an adequate margin of graft material. Grafts should be cut to adequate length to avoid
excessive tension on the graft-vessel interface. Anastomoses should be tension-free
throughout the patient's range of motion. Specific caution must be exercised in high-risk
areas (e.g., the proximal anastomosis of an axillofemoral bypass). The mechanical stresses
associated with repetitive joint flexion and extension undoubtedly play a role in the genesis of
some AAs (e.g., high incidence of femoral AA). Unfortunately these stresses can never be
eliminated completely, emphasizing the importance of controlling other risk factors. End-to-
side anastomoses are more often associated with AAs than end-to-end anastomoses (2,6,16).
As outlined above, an end-to-side anastomosis effectively double the vessel diameter, leading
to increased wall tension by LaPlace's law. Sewing to aneurysmal or even mildly dilated
arteries should be avoided if practically possible. The risk of juxta-anastomotic aneurysms of
the proximal aorta, including visceral patch aneurysms following TAAA repair, may
otherwise increase (6,8,10,17,18). The highest incidence of AA in one large study was in
patients undergoing bypasses for peripheral aneurysms (14). "Redo" operations also have a ■§

higher AA rate (9). Technical errors, along with infection, are associated with early AA |

formation (1,10). a

Regardless of etiology, the presence of an AA at one site is a risk factor for the c

development of AAs at other sites (2,10,19,20). Schellack and associates noted that 70% of <

groin AAs after aortofemoral grafting are bilateral, while 17% are associated with an >9

aortic AA (20). Despite the presence of an obvious etiology in many patients, not all J

patients have identifiable risk factors. The unpredictable occurrence of AAs in such «

individuals is one reason to maintain lifelong follow-up of all patients undergoing vascular |

reconstruction, especially when prosthetic materials are utilized. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

142 SHEPARD and JACOBSON

II. INCIDENCE

The overall incidence of AA ranges from less than 1% to more than 10% of anastomoses
at risk (1,2). The type of operation, duration of follow-up, and method of aneurysm
diagnosis all contribute to the wide-ranging incidence. In the largest series ever reported,
Szilagyi reviewed 4214 arterial reconstructions and found an AA in 3.9% of patients
(1.7% of anastomoses at risk). Although AAs can occur anywhere and after any type of
bypass grafting, they most commonly occur in the femoral artery as a complication of
prosthetic aortic grafting. Iliac and aortic anastomoses are much less frequently involved.
In reviewing the Henry Ford Hospital experience with aortic reconstruction for occlusive
disease, Szilagyi and associates documented AAs in 5.8% of femoral anastomoses, 2.4%
of iliac anastomoses, and only 0.2% of aortic anastomoses (7). Most of these aneurysms
were detected by clinical examination. The incidence of aneurysms remained relatively
constant over the 30-year span of the study. Others have reported very similar incidences
(14). Substantially higher rates of AA, however, have been reported by van den Akker and
colleagues in a very similar group of patients undergoing aortic reconstruction — 13.6%
femoral, 6.3% iliac, and 4.8% aortic (2). By 20 years postoperatively, nearly 50% of
patients in this series had developed an AA at one or more anastomoses. The higher
incidence of AAs in this report is at least partially due to the incorporation of screening
ultrasonography into the follow-up protocol during the last decade of the study.

Others have also reported an increasing incidence of aortic AA (6,16). Utilizing
routine ultrasound, Edwards and associates found aortic AAs in 10% of their aortic
reconstruction patients at a mean interval of 12 years following initial operation (6). The
reasons for this apparent increase are unclear but may well relate to increasing patient
longevity in addition to more aggressive screening.

Visceral patch AA following thoracoabdominal aortic aneurysm (TAAA) repair
utilizing the Crawford inclusion technique is a type of intra-abdominal AA that has only
recently been recognized. A 2001 report from Johns Hopkins documented the presence of
such aneurysms in 7.5% of patients (8 of 107) undergoing TAAA repair after a mean
follow-up of 6.5 years (8).

Although certain types of AA are being recognized with increasing frequency, the
incidence of femoral AA may actually be decreasing. A recent review of our registry
revealed a declining number of femoral AA repairs over the last decade despite a steady
increase in overall open procedures. Reasons for this decline are unclear but may include
improved technique and materials as well as a better understanding of the risk factors
associated with AA formation (see Sec. VI, below).

III. CLINICAL PRESENTATION AND DIAGNOSIS

The clinical presentation of AA depends on the location, size, and status of the aneurysm. |

Symptoms, when they occur, are usually similar to those of degenerative aneurysms at the a

same location. Femoral AAs most commonly present as an asymptomatic, pulsatile groin c

mass; patients note an enlarging mass associated with the scar from their previous surgery. <

Small femoral AAs may be first detected by the clinician during routine follow-up. Rarely >9

do they grow large enough to cause symptoms from nerve compression or venous obstruc- J

tion. Occasionally femoral AAs present urgently as rapidly enlarging, painful masses or «

with symptoms of limb ischemia. Ischemic complications result from thromboembolism of |

associated mural thrombus. Intrabdominal AAs usually remain clinically silent until they @

grow to a very large size or are detected on cross-sectional imaging studies. Occasionally %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ANASTOMOTIC ANEURYSMS 143

they present urgently, most frequently with rupture and hemorrhage (2). Because
associated scar tissue tends to tether adjacent structures to the prosthetic graft-arterial
interface, intrabdominal AAs have a greater tendency to present with fistula formation or
compressive symptoms than do degenerative aneurysms. Aortoenteric fistula from erosion
of a proximal aortic suture line AA into the overlying, adherent duodenum is the most
frequently recognized manifestation of this phenomenon. Hydronephrosis from ureteral
entrapment or iliac vein obstruction can result from iliac AAs.

The mean interval from the time of aortic reconstruction to the diagnosis of an AA
varies with location. Femoral AAs usually present within 5-6 years, while intrabdominal
AAs present 8-10 years after the initial reconstruction (2,6,17,20,21). Earlier presentations
are associated with technical error, infection, and recurrent AAs (1,4,10). The presence of
risk factors for AA (as discussed above) should heighten one's index of suspicion that an
AA may be present. In particular, patients who have developed an AA at one site are at
increased risk for developing another AA (2,10,19,20) (Fig. 1). Patients with an AA are
overwhelmingly male, with an average age of 62 years (2).

In most patients the diagnosis of a femoral AA is made by physical examination. In
questionable cases or obese patients, duplex ultrasound can be helpful in confirming the
diagnosis. Even in routine cases, duplex scanning is useful in sizing the aneurysm and
documenting the presence of mural thrombus — information that can be helpful in deciding
when to intervene. In contrast, intrabdominal AAs are frequently not detectable by
physical examination and more commonly present with symptoms (e.g., abdominal or
back pain) and/or complications (e.g., rupture, fistula formation) (10,18,21). Ultrasound
and cross-sectional imaging studies remain the most commonly employed diagnostic tests.
Angiography lacks the sensitivity of other imaging modalities for the diagnosis of AA but
is extremely helpful for preoperative planning.

A. Surveillance

The optimal approach to AA is early detection and elective operative repair to avoid the

morbidity /mortality associated with urgent presentations. An aggressive screening program

is the best way to make an early diagnosis. Our approach combines yearly physical

examination with selective ultrasonography or computed tomography (CT) scanning. Since

most femoral AAs are detectable on physical examination routine screening with ultrasound

is not necessary. In contrast, intra-abdominal AAs are rarely apparent on physical

examination unless they are large; routine screening with ultrasonography or CT scanning

is necessary (6,10,18). We currently favor CT scanning. Because most intrabdominal AAs

occur as a late (8-10 years) complication of aortic grafting, it seems reasonable to obtain a

first surveillance study at 5 years. Patients with risk factors for developing an early AA (e.g.,

previous AA repair, complicated perioperative course) should be screened earlier (2-3 years

postoperatively) (10). Subsequent studies can be obtained on an annual or biannual basis |

depending on circumstances. Observation with surveillance may be reasonable in patients g

with very small intrabdominal AAs or those who are poor operative candidates. ^

IV. MANAGEMENT

Treatment is indicated to prevent continuing expansion and rupture or limb-threatening
thromboembolic complications. The natural history of AAs differs from that of degener-
ative aneurysms in that AAs seem to grow more rapidly and unpredictably. The threshold
for intervention may therefore be lower than for a degenerative aneurysm of similar size.

■c

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

144

SHEPARD and JACOBSON

Figure 1 Arteriogram with right aortoiliac and left aortofemoral anastomotic aneurysms and right
hypogastric artery true aneurysm.

Many authors feel that the presence of an AA in itself is an indication for repair because of
the excess morbidity and mortality associated with conservative follow-up and emergency
surgery (5). However, as detailed earlier, not all AAs are false aneurysms and therefore
may not be at the same risk for rapid expansion, complications, etc. A selective approach
for small AAs may be reasonable assuming that patients are carefully followed at regular
intervals (20).

Because complications are uncommon with a femoral AA less than 2 cm in diameter,
most authorities do not consider repair of groin AAs until they reach at least 2 cm in
diameter (1). Even at this size, if they are stable and without significant intraluminal
thrombus, such aneurysms can probably be safely followed with serial duplex scanning.
Symptomatic aneurysms or those larger than 2.0-2.5 cm or containing significant mural
thrombus should be repaired when diagnosed. Because most intra-abdominal AAs are large
when detected, they should also be considered for repair at the time of diagnosis. The
management of small intra-abdominal AAs is controversial. Szilagyi has recommended that
aortoiliac AAs less than 50% the diameter of the host artery could be safely followed (1).
Rapid expansion of small aortic AAs (<4 cm), however, has been reported (21). Patients
with an intra-abdominal AA in whom repair is not performed at the time of diagnosis should

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ANASTOMOTIC ANEURYSMS 145

be followed very closely with imaging studies every 6 months. Occasionally patients with
severe comorbidities and limited life expectancy are managed nonoperatively.

Careful preoperative assessment is essential prior to operative intervention. Many
patients are a decade older than they were at the time of their original procedure, and
associated scarring can greatly increase the technical complexity of operation. High-
quality angiography is essential to define the patient's anatomy and allows selection of the
optimal surgical approach and method of repair. Cross-sectional imaging studies should
be obtained for all intra-abdominal AAs. Besides providing important information on the
size and location of the aneurysm, such studies can identify associated pathology, which
may alter the conduct of the repair (e.g., perigraft fluid, ureteral entrapment with
hydronephrosis). Spiral CT scanning with three-dimensional reconstruction is particularly
helpful and has recently eliminated the need for aortography on a high percentage of our
abdominal aortic aneurysm patients. Prior to surgery, an attempt should be made to
review old operative notes if available. In addition, the operating surgeon should always
make a risk assessment as to the chances of graft infection being the cause of the AA. The
operative approach for AAs associated with graft infections is much different than that for
sterile AAs.

A. Operative Repair

The basic principles of operative management for AAs are the same regardless of their
location. The "redo" nature of these procedures makes them technically challenging to
perform, even for experienced vascular surgeons. Sharp scalpel dissection is preferred
because of the associated, often dense scar tissue. Dissection must be precise to minimize
damage to adjacent structures and arterial branches. In severely scarred operative fields,
care must be taken to avoid the "easier" dissection plane sometimes found between the
adventitia and media; such "exarterectomy" makes subsequent reconstruction extremely
difficult. In contrast, dissection of most prosthetic grafts is greatly simplified by entry
through an enveloping pseudocapsule that can be stripped from the graft. Proximal
control is usually best obtained at a site somewhat removed from the origin of the
aneurysm. This approach avoids the most significant scarring, reduces the risk of
inadvertent entry into the aneurysm, and provides an adequate cuff of proximal graft or
artery for subsequent repair. With severe scarring, distal control is sometimes easier to
obtain from within the opened aneurysm using balloon occlusion catheters (or a Foley
catheter for the aorta).

During exposure, evidence of graft infection should always be carefully sought (e.g.,
nonincorporation of the graft, perigraft fluid) and appropriate samples (both fluid and
aneursym wall) sent for culture and Gram's stain. We do not routinely sonicate our
samples prior to culture unless we are dealing with a recurrent aneurysm (9). Once the ■a

aneurysm is open, an attempt should be made to determine the cause of anastomotic |

disruption if possible. Simple suture repair of a small anastomotic defect can occasionally a

be considered but is usually not successful in the long term. In the vast majority of cases, c

the most reliable method of repair is placement of a new interposition graft between the <

old graft and the host artery. >5

Segments of healthy, uninvolved artery and graft should be cleared of scar tissue in J

preparation for anastomosis. In some situations it is simply easier to excise the entire °

aneurysm and associated scar tissue (e.g., groin AA), while in others such debridement is |

too risky (e.g., a proximal aortic aneurysm with adherence to bowel). Whenever there is @

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

146 SHEPARD and JACOBSON

significant concern about the presence of infection, however, more rather than less
debridement would seem to be appropriate. The choice of graft material is probably
unimportant, although there is some theoretical and clinical support for using ePTFE,
which offers improved resistance to infection and less propensity to dilate compared to
Dacron (12,22,23). Anastomoses should be tension-free, taking generous, evenly spaced
bites into healthy arterial wall. If the arterial lumen is significantly compromised by
plaque, endarterectomy should not be avoided because of fears of a recurrent AA; leaving
significant plaque behind risks future graft thrombosis. End-to-end anastomoses are
preferred over end-to-side for the theoretical reasons discussed above in Sec. I; however,
no study has ever shown the superiority of one anastomotic configuration over another in
the prevention of recurrent aneurysm formation.

B. Femoral Anastomotic Aneurysms

Most femoral aneurysms can be safely approached through the old groin incision.
Occasionally, with larger or acutely expanded aneurysms, it is safer to obtain proximal
control through a suprainguinal, extraperitoneal approach. With badly scarred groins,
balloon occlusion catheters are helpful not only for distal control but also for controlling
bleeding from a patent proximal common femoral artery. Repair is usually easier after
excision of the pseudoaneurysm cavity and associated scar, taking care to avoid injury to
the femoral vein and nerve. Reconstruction is accomplished with an interposition graft
from the transected graft limb proximally to the outflow vessel(s) distally. The profunda
femoris is frequently the only outflow vessel and a local endarterectomy or profundoplasty
may be needed. Concomitant popliteal/tibial bypass is rarely necessary in our experience
(1). Rarely, a patent proximal common femoral artery is supplying critical pelvic perfusion
through retrograde flow. In this situation it may be necessary to maintain femoral artery
continuity by inserting an interposition graft between the proximal common femoral
artery and the distal outflow arteries and suturing a new graft extension to this prosthetic
"femoral artery." With large or acutely expanded groin aneurysms, obtaining adequate
coverage with healthy skin and soft tissue can be problematic; sartorius myoplasty may be
helpful in this situation.

C. Aortoiliac Anastomotic Aneurysms

Proximal aortic suture line aneurysms can be extremely challenging to repair because of
their frequent encroachment on the renal arteries and their occasional involvement of the
pararenal/visceral aorta. Such aneurysms present problems in terms of optimal exposure,
site of proximal aortic clamping, and avoidance of vital organ ischemic complications. A
complete discussion of these issues is beyond the scope of this chapter, but a brief
description of our approach may be worthwhile. Although some of these aneurysms can §

be safely accessed through a standard transabdominal inframesocolic approach, more |

extensive proximal abdominal aortic exposure is frequently needed (Fig. 2). We have s

found an extended left flank retropertioneal approach invaluable in dealing with these 2

aneurysms (24). This exposure not only provides excellent access to the entire abdominal u

aorta (from diaphragm to bifurcation) but also allows an approach to the aneurysm g

through undissected tissue planes, thus minimizing risk to adjacent structures (e.g., |j

duodenum, left renal vein), that may be involved in scar tissue. Aortic cross-clamping s

at a suprarenal or more proximal level is frequently necessary in the repair of these M

Q

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ANASTOMOTIC ANEURYSMS

147

Figure 2 Proximal aortic anastomotic aneurysm.

aneurysms, and appropriate precautions should be taken to minimize the risks of vital
organ ischemia and increased cardiac stress (10,17,18). Unless the pararenal aorta is
involved in the aneurysmal process, there is usually a short segment of infrarenal aorta
available for anastomosis to an interposition graft. If a suprarenal or more proximal aortic
clamp has been used, it can be moved on to this graft following completion of this
anastomosis to minimize renal (and visceral) ischemia. The distal anastomosis is per-
formed to the old graft after freeing it from its pseudocapsule and resecting any obvious
redundancy. No attempt should be made to excise the aneurysm cavity or associated scar
because of risk of injury to adjacent structures.

Iliac AAs can be approached either transperitoneally or through an extraperitoneal
lower abdominal incision, depending upon size, location, and potential involvement of

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

148

SHEPARD and JACOBSON

(a)

(b)

Figure 3 Endovascular treatment, (a). Iliac anastomotic aneurysm and hypogastric artery true
aneurysm, (b). Covered stent (bracket) and embolized coils in hypogastric aneurysm (arrowhead),
(c). Completion artiogram with both aneurysms excluded.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ANASTOMOTIC ANEURYSMS 149

(C)

Figure 3 Continued

adjacent structures. Because of their proximity to these aneurysms, the ureters are always
at risk during operative repair. Routine placement of a ureteral stent on the involved side
has proved invaluable in minimizing the risk of injury. Repair follows the guidelines
outlined previously. In the rare situation where hypogastric perfusion must be maintained
and the aneurysm involves the iliac bifurcation, the interposition graft can be sewn end-to-
end to the hypogastric artery distally and the external iliac artery reimplanted into the side
of the graft. Again, no attempt should be made to resect or debride the pseudoaneurysmal
cavity unless infection is strongly suspected.

D. Endovascular Repair

The repair of some intra-abdominal AAs has been greatly simplified by the recent
introduction of endoluminal stent graft techniques. The main anatomical restriction for
endovascular repair is the requirement for an adequate proximal "landing zone." While
this requirement is not a concern for most iliac AAs, it is necessary for many proximal
aortic suture line aneurysms that encroach on the renal arteries. Transrenally based stent
grafts may eliminate this constraint for many patients in the future. Currently, in patients
with suitable anatomy, this safer and less invasive approach holds great promise to avoid
the morbidity and mortality associated with open repairs (25,26). An example of a patient
who underwent successful endoluminal repair of an iliac AA is shown in Figure 3. !§>

E. Visceral Patch AA Following TAAA I

Perhaps the most technically challenging AAs to repair are those involving the visceral J

patch following TAAA surgery (Fig. 4). In addition to the scarring associated with any Q

AA, these repairs are also complicated by the extensive exposure and obligatory periods of |

prolonged vital organ ischemia required. There are several reconstruction options, but all ©

t

a

s

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

150

SHEPARD and JACOBSON

(a)

(b)

Figure 4 (a) CT scan of a visceral patch aneurysm following type I thoracoabdominal aneurysm
repair. Aneurysm wall with mural thrombus (arrowhead) and superior mesenteric artery (arrow), (b)
Aortogram of same aneurysm showing right renal artery originating from most distal portion of
aneurysm and bypass to left renal artery originating several centimeters lower.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ANASTOMOTIC ANEURYSMS 151

require insertion of a new interposition aortic graft to replace the prosthetic segment
involved with the aneurysm. The visceral/renal arteries can be reattached by fashioning a
smaller inclusion patch or reimplanting each vessel separately, taking deep bites through
their origins to exclude as much aortic wall as possible. The left renal artery can be
separately bypassed with a short prosthetic sidearm (8). We have utilized this approach
successfully in the one patient with this problem that we have encountered. Alternatively,
separate prosthetic bypasses off the new interposition graft can be used to reconstruct all
the visceral/renal branches (27). A novel technique employing sequential visceral bypass
from the descending thoracic aorta/prosthesis has also been employed with success (28).

V. OUTCOME

The results of femoral AA repair are generally good; the mortality rate is low (3-4%) even
when the procedure is performed in acutely symptomatic patients (1,20). Graft limb
thrombosis is the most frequent early and late complication, with resulting amputation in
less than half the affected patients (1,20). Mulder reported an early limb loss rate of 2.6%
(5). Graft infection occurs in up to 5% of patients (20). Recurrent femoral AA is another
well-recognized complication, affecting 6-19% of patients (4,5,20). In reviewing our
experience, Ernst documented the following risk factors for recurrent groin AAs: female
gender, groin wound complications at the time of AA repair, and development of an AA
within 4.5 years of the original aortofemoral grafting procedure (4). Repair of recurrent
AAs is also associated with good results (4).

The extensive scarring and frequent need for suprarenal aortic control and renal
artery reconstruction combine to increase the risk associated with the repair of proximal
aortic AAs. Mortality rates for elective repair average 10-15%, while those for emergent
repair exceed 50% (5,10,17,18,21). The morbidity associated with these repairs is also
substantial; bleeding, renal dysfunction, and pulmonary insufficiency are the most
commonly encountered problems (10,17,18,21). The outcome of iliac artery AA repair
is not well documented in the literature, perhaps because this entity is usually lumped with
aortic AAs. Risks are certainly less than with aortic AAs, though ureteral scarring can
present a problem. Due to the elevated risks associated with emergency open repair,
intrabdominal AAs should be repaired on an elective basis at the earliest practical
opportunity. Early reports of endovascular treatment of intrabdominal AAs have
demonstrated favorable results. The long-term durability and efficacy of this approach,
however, has not yet been proven (25,26,29,30).

Repair of visceral patch AA following TAAA surgery is a morbid undertaking (8).
Three of four patients undergoing elective repair in the Johns Hopkins series survived, but
only after prolonged intensive care unit and hospital stays (8).

Reducing the occurrence of AAs requires the application of good surgical technique
coupled with a thorough understanding of the associated risk factors. General principles
include the construction of a tension-free anastomosis using a monofilament suture and
taking generous, evenly spaced bites of healthy arterial wall. Endarterectomy should not
be avoided if leaving residual plaque will compromise the outcome of the reconstruction.
There is some theoretical evidence that ePTFE grafts are associated with a lower incidence

s

VI. PREVENTION §

'C

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

152 SHEPARD and JACOBSON

of AA, but the data are not strong enough to recommend one vascular prosthesis over
another. There are also theoretical as well as some clinical data to suggest that end-to-end
anastomoses are associated with less AAs than end-to-side anastomoses. Unfortunately, in
many situations, construction of an end-to-end anastomosis is not possible without risking
the interruption of critical arterial perfusion to a specific vascular territory. Placing distal
graft limb anastomoses above the inguinal ligament whenever feasible may reduce the
mechanical stresses associated with repetitive hip flexion and extension. The importance of
avoiding perioperative complications, both local and systemic, is obvious.

Measures to reduce the occurrence of intra-abdominal AAs include placing the
proximal aortic anastomosis close to the renals to minimize the risk of aneurysmal
degeneration in the retained infrarenal segment. Sewing to aneurysmal or even dilated
aorta should be avoided whenever possible. This dictum is particularly important in
utilizing Crawford's inclusion technique for reattaching visceral/renal branches to a
thoracoabdominal aortic graft. Patches of aorta containing the origins of the renal and
visceral vessels should be made as small as possible so as to exclude potentially degenerated
aortic wall from the anastomosis. Suture line bites should be taken as close to the ostia of the
reimplanted vessels as possible; occasionally it may be necessary to reimplant each vessel
separately. To reduce the size of the visceral/renal patch, we routinely reconstruct the left
renal artery with a short bypass graft from the aortic prosthesis.

Adoption of these techniques should help reduce the incidence of AAs, though it is
unlikely that this common vascular surgical complication will ever be completely
eliminated. For the aneurysms that do occur, prompt diagnosis and skilled repair remain
the keys for successful management.

REFERENCES

1. Szilagyi DE, Smith FR, Elliot JP, Hageman JH, Dall'Olmo CA. Anastomotic aneurysms after
vascular reconstruction: Problems of incidence, etiology and treatment. Surgery 1975; 178:800-
816.

2. van den Akker PJ, Brand R, van Schilfgaarde R, van Bockel JH, Terpstra JL. False aneurysms
after prosthetic reconstructions for aortoiliac obstructive disease. Ann Surg 1989; 210:658-666.

3. Mii S, Sakata H, Kawazoe N. Para-anastomotic aneurysms: incidence, risk factors, treatment
and prognosis. J Cardiovasc Surg 1998; 39:259-266.

4. Ernst CB, Elliott JP Jr, Ryan CJ, Abu-Hamad G, Tilley BC, Murphy RK, Smith RF, Reddy DJ,
Szilagyi DE. Recurrent femoral anastomotic aneurysms: A 30-year experience. Ann Surg 1988;
208(4):401-409.

5. Mulder EJ, van Bockel JH, Maas J, van den Akker PJ, Hermans J. Morbidity and mortality of
reconstructive surgery of noninfected false aneurysms detected long after aortic prosthetic
reconstruction. Arch Surg 1998; 133:45-49. -a

6. Edwards JM, Teefey SA, Zierler RE, Kohler TR. Intraabdominal paraanastomotic aneurysms |
after aortic bypass grafting. J Vase Surg 1992; 15:344-353. G

7. Szilagyi DE, Elliott JP, Smith RF, Reddy DJ, McParlin M. A thirty-year survey of the re- Jj,
constuctive surgical treatment of aortoiliac occlusive disease. J Vase Surg 1986; 3:421-436. s

8. Dardik A, Perler BA, Roseborough GS, Williams GM. Aneurysmal expansion of the visceral 6
patch after thoracoabdominal aortic replacement: An argument for limiting patch size? J Vase Tj
Surg 2001; 34:405-410. |

9. Seabrook GR, Schmitt DD, Bandyk DF, Edmiston CE, Krepel CJ, Towne JB. Anastomotic °
femoral pseudoaneurysm: An investigation of occult infection as an etiologic factor. J Vase I
Surg 1990; 11:629-634. g

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ANASTOMOTIC ANEURYSMS 153

10. Allen RC, Schneider J, Longenecker L, Smith RB, Lumsden AB. Paraanastomotic aneurysms
of the abdominal aorta. J Vase Surg 1993; 18:424-432.

1 1 . Moore WS, Hall AD. Late suture failure in the pathogenesis of anastomotic pseudoaneurysms.
Ann Surg 1970; 172:1064-1068.

12. Berman SS, Hunter GC, Smyth SH, et al. Application of computed tomography for surveil-
lance of aortic grafts. Surgery 1995; 118:8-15.

13. Nunn DB, Carter MM, Donohue MT. et al. Postoperative dilation of knitted Dacron aortic
bifurcation graft. J Vase Surg 1990; 12:291-297.

14. Mehigan D, Fitzpatrick B, Browne HI, Bouchier-Hayes DJ. Is compliance mismatch the major
cause of anastomotic aneurysms? J Cardiovasc Surg 1985; 26:147-150.

15. Gaylis H. Pathogenesis of anastomotic aneurysms. Surgery 1981; 90:509-515.

16. Mikati A, Marache P, Watel A, Warembourg H Jr, Roux JP, Noblet D, Soots G. End-to-
side aortoprosthetic anastomoses: Long-term computed tomography assessment. Ann Vase
Surg 1990; 4:584-591.

17. Hagino RT, Taylor SM, Fujitani RM, Mills JL. Proximal anastomotic failure following infra-
renal aortic reconstruction: Late development of true aneurysms, pseudoaneurysms, and oc-
clusive disease. Ann Vase Surg 1993; 7:8-13.

18. Curl GR, Faggioli GL, Stella A, D'Addato M, Ricotta JJ. Aneurysmal change at or above the
proximal anastomosis after infrarenal aortic grafting. J Vase Surg 1992; 16:855-860.

19. Gautier C, Borie H, Lagneau P. Aortic false aneurysms after prosthetic reconstruction of the
infrarenal aorta. Ann Vase Surg 1992; 6:413-417.

20. Schellack J, Salam A, Abouzeid MA, Smith RB, Steward MT, Perdue GD. Femoral anas-
tomotic aneurysms: A continuing challenge. J Vase Surg 1987; 6:308-317.

21. Treiman GJ, Weaver FA, Cossman DV, Cossman DV, Foran RF, Cohen JL, Levin PM,
Treiman RL. Anastomotic false aneurysms of the abdominal aorta and the iliac arteries. J Vase
Surg 1998; 8:268-273.

22. Schmitt DD, Bandyk DF, Pequet AJ, Towne JB. Bacterial adherence to vascular prostheses. J
Vase Surg 1986; 3:732-740.

23. Carson SN, Hunter GC, Palmaz J, Guernsey JM. Recurrence of femoral anastomotic
aneurysms. Am J Surg 1983; 146:774-778.

24. Shepard AD, Tollefson DJF, Reddy DJ, Evans JR, Elliot JP Jr, Smith RF, Ernst CB. Left
flank retroperitoneal exposure: A technical aid to complex aortic reconstruction. J Vase Surg
1991; 14:283-291.

25. Yuan JG, Marin ML, Veith FJ, Ohki T, Sanchez LA, Suggs WD, Cynamon J, Lyon RT.
Endovascular grafts for noninfected aortoiliac anastomotic aneurysms. J Vase Surg 1997;
26:210-221.

26. Morrissey NJ, Yano OJ, Soundararajan K, Eisen L, McArthur C, Teodorescu V, Kerstein M,
Hollier L, Marin M. Endovascular repair of para-anastomotic aneurysms of the aorta and iliac
arteries: Preferred treatment of a complex problem. J Vase Surg 2001; 33:503-512.

27. Carrel TP, Signer C. Separate revascularization of the visceral arteries in thoracoabdominal
aneurysm repair. Ann Thorac Surg 1999; 68:573-575.

28. Ballard JL. Thoracoabdominal aortic aneurysm repair with sequential visceral perfusion: A

technical note. Ann Vase Surg 1999; 13:216-221. 1

29. White RA, Donayre CE, Walot I, Wilson E, Jackson G, Kopchock G. Endoluminal graft ex- <S
elusion of a proximal para-anastomotic pseudoaneurysm following aortobifemoral bypass. js
J Endovasc Surg 1997; 4(1): 88-94. Jf

30. Criado E, Marston WA, Ligush J, Mauro MA, Keagy BA. Endovascular repair of peripheral ^
aneurysms, pseudoaneurysms, and arteriovenous fistulas. Ann Vase Surg 1997; 11:256-263. >S

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

Hypercoagulable States and Unexplained
Vascular Graft Thrombosis

Jonathan B. Towne

Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

Hypercoagulable states as a cause of unexplained vascular thrombosis present a difficult
clinical problem. Most graft failures in the perioperative period are presumed to occur
because of technical errors in the construction of the anastomosis, problems with the
conduit, or poor patient selection. The diagnosis of an abnormal hypercoagulable state is
often made only after all of these other factors have been excluded. Although failure of
heparin to prevent clotting in the operative field or immediate thrombosis of a vascular
repair suggests abnormal coagulation, the diagnosis can be confirmed only by the blood
coagulation laboratory. The clotting disorder must be detected early in the course of the
disease to obtain a favorable outcome. Abnormal thrombosis falls into five general cate-
gories: (a) heparin-induced platelet aggregation, (b) abnormalities in the antithrombin
system, (c) abnormalities of the fibrinolytic system, (d) thrombosis caused by lupus-like
anticoagulant, and (e) a miscellaneous category consisting primarily of abnormal platelet
aggregation and protein C and protein S deficiency.

I. HEPARIN-INDUCED THROMBOSIS

Paradoxical thrombotic complications of heparin sodium anticoagulant therapy are un-
common but potentially limb-threatening and occasionally fatal. Several investigators 1
have identified a chemically induced, immune thrombocytopenia as the cause of heparin- 1
induced intravascular thrombosis, which usually occurs after 4-10 days of continued s
exposure to the drug (1-5). The immune factor that triggers the thrombocytopenia has been
identified as an IgG antibody, which produced agglutination of normal platelets when
either porcine gut or beef lung heparin is added. The IgG protein is stimulated by the
heparin/platelet factor 4 complex and activates the platelet via the platelet F c receptor (6).
The thrombi that occur with heparin-induced thrombosis have an unusual grayish white
appearance in contradistinction to the red color of most thrombi. The white color is

155

c

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

156 TOWNE

secondary to the creation of fibrin-platelet aggregates, which can be clearly identified on
electron microscopy (7).

Rhodes et al. (8) found a heparin-dependent IgG antibody in the serum of several
patients by means of the complement lysis inhibition test. They also demonstrated a
residual heparin-platelet aggregating effect 12 days to 2 months after patient recovery from
the initial exposure to heparin. In these patients, a 24-h infusion of heparin caused a mean
reduction of platelet count of 197,000/mm 3 . Since heparin preparations are not pure
substances, it is also possible that a high-molecular-weight contaminant not eliminated by
the extraction procedure may cause the antiplatelet defect.

Up to 30% of patients may manifest a decrease in their platelet count after starting
heparin therapy, but the incidence of significant thrombocytopenia and resulting throm-
botic or hemorrhagic complications is approximately 5% (9). Two types of heparin-induced
thrombocytopenia are described. Type I, or the acute form, occurs relatively early and
results in a benign course with improvement in the platelet count during continued heparin
therapy. Type II, or the delayed form, occurs 5-14 days after the institution of heparin
therapy in a patient not previously exposed to heparin and after 3-9 days in patients with a
history of previous heparin therapy. Type II heparin-induced thrombocytopenia is reported
to have a 23-60% thrombotic or hemorrhagic complication rate and a 12-18% mortality
rate. Early recognition and treatment results in a significant improvement in the associated
morbidity and mortality (10,11). In type I heparin-induced thrombocytopenia, the mech-
anism of action is thought to be a non-immune-mediated direct effect of heparin on platelets
that causes aggregation. Type II heparin-induced thrombocytopenia is due to an immune
mediated (IgG and IgM) platelet aggregation.

A. Clinical Presentation

Heparin-induced intravascular thrombosis can occur following a wide variety of indications
for heparin administration, including thrombophlebitis with and without pulmonary
embolus, perioperative heparin prophylaxis in patients at risk for thrombophlebitis, car-
diac surgery, and vascular reconstruction. Platelet aggregation induced by heparin can
result from both porcine gut and bovine lung heparin and can affect either the arterial or
venous circulation. Both subcutaneous and intravenous heparin administration can pro-
duce this phenomenon (12). Even heparin-coated catheters can cause heparin-induced
thrombocytopenia. Laster and Silver (12) reported the development of heparin-induced
thrombocytopenia in 10 patients whom heparin-coated pulmonary artery catheters were
inserted. Despite discontinuation of all other sources of heparin, the thrombocytopenia
persisted. Although all of the patient were also given heparin, it is theoretically possible that
heparin-coated catheters alone could have caused abnormal platelet aggregation.

The clinical features of this syndrome are often dramatic. In any patient who has had
thrombotic complications while receiving heparin therapy, heparin-induced aggregation of 13

platelets should be considered. This is especially important in patients with arterial occlu- |

sions who do not have any other evidence of atherosclerotic vascular disease. At operation, j§

the finding of a white clot at thrombectomy should alert the surgeon to the possibility of ||

heparin-induced thrombosis. In contrast to several reports in literature, increased heparin ^

sensitivity rather than increased heparin resistance was noted in several of our patients (7). "jl

The cause of this is uncertain, but it is presently believed that it is unrelated to the heparin- s

induced aggregative immunoglobulin. 2

It was initially felt that arterial thrombosis was more prevalent than venous thrombosis J

with this complication. However, some prospective studies by Warkinton et al. demon- ©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HYPERCOAGULABLE STATES AND THROMBOSIS 157

strated that there is actually a prevalence of venous to arterial emboli at a 4-to-l ratio (6).
Many of the venous thromboses are not detected unless studies such as duplex scanning
doesn't search these out. The most common arterial location of thrombosis is the
extremities, primarily the lower extremities, followed by the cerebral circulation, and finally
manifesting as myocardial infarctions. The exact cause of this distribution of prevalence of
thrombosis is uncertain, but it is certainly the authors' view that the thrombosis occurs more
commonly on diseased vessels and that the incidence of diseased vessels in the extremities,
particularly the lower extremities, is much higher than in many other vessels. This is followed
by carotid bifurcation disease and coronary artery disease.

Skin lesions have been noted in patients with heparin-induced thrombosis. These are
often seen at the site of the subcutaneous injection. They can present as a painful ery-
thematous plaques, which can progress to skin necrosis. These can be unaccompanied by
thrombocytopenia. With the prevalence of subcutaneous injection, the incidence of these
findings has increased (13).

B. Diagnosis

Definitive diagnosis of heparin-induced intravascular thrombosis is obtained by perform-
ing platelet aggregation tests. Two patterns of response have been noted. The more com-
mon pattern is for the patient's platelet-poor plasma to aggregate donor platelets on the
addition of heparin, indicating the presence of a relative nonspecific platelet-aggregating
factor in the patient's plasma. The less common pattern is for the patient's plasma to be
active against only the patient's platelets and have no effect on donor platelets. Other more
sensitive tests include C 14 serotonin release testing and enzyme-linked immunosorbent
essay (ELISA) testing for the antibody to the heparin PF4 complex.

Other clotting factors are usually normal: fibrinogen level is normal, the level of fibrin
split products may be mildly elevated but not in the range seen with intravascular coagu-
lation, and prothrombin time is normal or slightly prolonged. All patient have a marked
reduction in platelet count of less than 100,000/mm 3 or a 50% decrease from admission
level. In our series (7), the platelet count averaged 37,500/mm 3 with a range of 6000 to
73,000/mm 3 .

Patients with arterial thrombosis often present with unique angiographic findings.
These lesions consist of broad-based, isolated, lobulated excrescences that produce a varia-
ble amount of narrowing of the arterial lumen. Usually these findings have an abrupt
appearance, with prominent luminal contour deformities in arterial segments that are
otherwise normal. This distribution of disease is unusual and distinct from findings com-
monly seen with atherosclerosis. These changes occur in both the suprarenal and infrarenal
portions of the abdominal aorta and represent adherent mural thrombi composed of
aggregates of platelets and fibrin incorporating varying amounts of leukocytes and ery-
throcytes. Platelet aggregation tests should also be performed on any patient in whom 1
recurrent pulmonary embolism developed while he or she was receiving adequate heparin |
therapy. j§

The diagnosis of heparin-induced thrombosis is primarily a clinical one. All the current ||

laboratory tests have a relatively high percentage of false-negative rates and the more ^

difficult-to-perform serotonin release and ELISA testing are not available in all hospitals. "jl

In a patient in whom the clinical syndrome of low platelet count and abnormal thrombosis s

is noted and in whom the tests are negative should be treated with a presumptive diagnosis 2

of heparin-induced thrombosis and the tests repeated. Although false-negative tests are J

reported, false-positive tests are quite unusual. ©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

158 TOWNE

C. Treatment

Currently there are three approaches to treating patients with heparin-induced thrombosis
(6). The first is the use of danaparoid (a heparinoid), a mixture of glycosaminioglycans
(heparin-sulfate) and dermatan sulfate. This is quite effective, but it has a 10-40% cross
reactivity with patients having heparin-induced platelet aggregation. Therefore, prior to
instituting this therapy, the patient must have a platelet test against danaparoid to make sure
that it does not cause platelet aggregation. The second course of therapy is the use of
lepirudin, a recombinant form of the medicinal leech salivary protein hirudin, which is a
direct thrombin inhibitor that can be quite effective. Patients are monitored by obtaining an
activated thromboplastin time (A-PTT), which is kept at the level of 1.5-3 times normal.
Following an adequate therapeutic response, the patient can be converted to warfarin for
long-term anticoagulation. The third choice is Argatroban, a synthetic direct thrombin
inhibitor, derived from L-arginine. Like lepirudin, this is monitored by following activated
PTT levels.

When heparin-induced thrombocytopenia is diagnosed, heparin treatment should be
reversed immediately with protamine sulfate, and dextran 40 should be administered for
its antiaggregating and rheologic effects. Warfarin therapy also should be initiated and
continued for several months. In patients with arterial occlusive manifestations of heparin-
induced thrombosis, long-term warfarin therapy is recommended because of the possi-
bility of coexisting latent venous occlusive disease.

The response of the platelet count to discontinuation of heparin therapy is usually
prompt, often resulting in thrombocytosis, with a platelet count of 500,000 to 6000,000/
mm 3 being achieved in several days.

Coagulation tests distinguish heparin-induced platelet aggregation from other clotting
disorders. The fibrinogen level and prothrombin time are usually normal. The level of fibrin
split products and prothrombin time are normal or slightly elevated. The sole patient in our
series with a noticeable elevated level of fibrin split products was the initial patient, in
whom the diagnosis was not made antemortem. Heparin therapy was not stopped, and
before her death (caused by an intracerebral hemorrhage), she had massive venous
thrombosis involving both upper and lower extremities, which resulted in an elevated
level of fibrin split products. Early identification of heparin-induced thrombosis is
necessary to minimize the catastrophic complications of major limb amputation and death.

This experience suggests that it is imperative, in all patients receiving heparin therapy,
to have serial platelet counts done from the fourth day of heparin therapy onward. It is
our policy to perform platelet counts every other day starting on the fourth day of heparin
therapy. If thrombocytopenia develops, platelet aggregation studies should be performed
immediately. With early recognition of complications, the mortality and morbidity of
major amputation can be prevented. Morbidity and mortality rates reported in the
literature vary from 22 to 61% and 12 to 33%, respectively (14,15).

II. STRATEGIES FOR PATIENTS WITH HEPARIN-INDUCED
PLATELET AGGREGATION

Patients who require subsequent heparin therapy for other vascular or cardiac surgery
procedures require special management. In patients in whom heparin-induced platelet
aggregation develops, the platelet aggregation tests usually revert to normal from 6 weeks
to 3 months. Vascular or cardiac surgery procedures are preferably delayed until these tests

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I
•5

I

HYPERCOAGULABLE STATES AND THROMBOSIS 159

revert to normal. We test the patient at 6 weeks and then every 2 weeks thereafter to deter-
mine when the platelet aggregation tests are negative. When they are negative, the patient is
then admitted to the hospital for surgery. Cardiac catheterization or angiography is done as
required without the use of heparin flush solutions, since even small amounts of heparin in
the flush solutions can stimulate the development of heparin-induced antiplatelet antibodies.
The vascular or cardiac surgery procedure is then performed with the usual administration
of heparin. At the conclusion of the procedure, the heparin is reversed with protamine and
care is taken during the postoperative period to ensure the patient does not receive heparin
inadvertently through the flushing of either central venous catheters or arterial lines. By
using this procedure, we have not had any difficulty with reexposure to heparin.

However, a different strategy is necessary for those patients who require an additional
vascular or cardiac surgery procedure and who cannot wait until the results from heparin-
induced platelet aggregation tests are negative. In patients requiring procedures that can
be done without the use of heparin, such as resection of abdominal aortic aneurysm,
heparin is not used. However, in patients who require complex lower extremity revascula-
rization or cardiopulmonary bypass, some sort of anticoagulation is necessary. There are
basically two approaches. That favored by Laster et al. (16) involves administering aspirin
and dipyridamole (Persantine) preoperatively and then using heparin for the operative
procedure, as is customary. In addition to aspirin and dipyridamole, we prefer also to use
low-molecular-weight dextran, which, in addition to its rheological properties, coats the
platelets and interferes with platelet adhesion. In some patients, however, as noted by
Kappa et al. (17), the administration of aspirin has no effect on heparin-induced platelet
aggregation. Makhoul et al. (18) noted that although aspirin abolished platelet aggrega-
tion in 9 of 16 patients with heparin-induced platelet aggregation, it only decreased platelet
aggregation in the remaining 7, suggesting that aspirin is not able to reverse abnormal
platelet aggregation in all patients. Based on these reports, our procedure is to administer
aspirin and dipyridamole for several days before the operative procedure. On the day of
operation, the platelet aggregation tests are performed with the addition of heparin. If the
heparin causes abnormal platelet aggregation, iloprost can then used to prevent heparin-
induced platelet aggregation during the procedure. The use of iloprost can be complicated,
particularly since it is a very potent vasodilator and rather large doses of adrenergic agents
are often required to support blood pressure. Also, it has been approved for use by the
U.S. Food and Drug Administration (FDA) and may be used off label.

Sobel et al. (19) reported an alternate technique in which patients received warfarin
anticoagulant combined with dextran as a means of preventing intraoperative thrombosis
during reconstruction. This is a reasonable alternative for peripheral vascular reconstruc-
tions but is not possible for cardiopulmonary bypass. In the future, different substances
may be available to allow for adequate anticoagulation. Makhoul et al. (18) noted in vitro
that heparinoids did not cause platelet aggregation. These new anticoagulant agents are ■§

being developed in Europe and may, in the future, be available in the United States. g

Latham et al. have described the use of recombinant hirudin for treatment of a patient »

who had heparin-induced platelet aggregation and who required cardiopulmonary bypass c

(20). They were successfully able to anticoagulate the patient and place him on bypass <

without any untored results. >5

Cole and Bormanis (21) have reported the use of ancrod, which is made from the ^

venom of the Malaysian pit viper (Agkistrudon rbodastoma), as an anticoagulant in patients °

who have heparin-induced platelet aggregation. Ancrod acts enzymatically on the fibri- |

nogen molecule to form a product that cannot be clotted by physiological thrombin. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

160 TOWNE

III. ANTITHROMBIN DEFICIENCY

Antithrombin III (AT III) is an alpha globulin manufactured in the liver and perhaps by
vascular endothelium, with a molecular weight of approximately 60,000 Da and a half-life
of 2.8 days (22). It is a serine proteinase inhibitor that binds in equimolar ratios to several
enzymes participating in the intrinsic pathway of blood coagulation, including thrombin
factors, Ixa, Xa, and Xia (23,24). Heparin significantly accelerates the rate at which AT III
neutralizes these enzymes, limiting sequential clotting reactions and preventing fibrin
formation. In 1965, Egeberg (19) described a family with an inborn defect of AT III.
Subsequent research has confirmed the genetic transmission of this deficiency (25-28). The
frequency of this defect is approximately 1 in 2000 to 1 in 5000 in the general population
(29,30). There are probably at least two types of AT III deficiency. In the classic form,
both the protein level of AT III (as determined by measurement of its protein concen-
tration) and its activity level are reduced in the patient's plasma (31). However, there are
other patients in whom the concentration of AT III is normal or even slightly elevated as
measured by protein level, but the biological function as measured by activity level tests is
abnormal (32). This suggest that these patients are manufacturing a defective antithrom-
bin molecule. Acquired AT III deficiency can occur in patients with severe liver disease,
nephrotic syndrome, hypoalbuminemia, malnutrition, and disseminated intravascular
coagulation and in some patients taking oral contraceptives. AT III deficiency may be
an indicator of significant protein catabolism. Flinn et al. (33) noted low AT III activity.
There was low serum albumin ( <3 mg/dL) in 48% of these patients, which was associated
with an increased incidence of early graft failure.

A. Clinical Presentation

Although AT III deficiency is inherited, it is rare for episodes of thrombosis to be clinically
manifest before the second decade of life. Despite continuously depressed levels of AT III
in these patients, thrombotic episodes are often related to predisposing factors such as
surgery, childbirth, and infection; they rarely occur spontaneously. This deficiency can
cause venous thrombosis, pulmonary embolism, dialysis fistula failure, arterial graft
occlusion, and spontaneous arterial occlusion. In our initial report, we identified 7 patients
(5 men, 2 women; age range, 21-65 years) with antithrombin deficiencies as a cause of
thrombosis (34). Three presented with early thrombosis of femorodistal grafts (Fig. 1). In
two of these patients the grafts became occluded shortly after surgery in the brief interval
between completion of the anastomoses and performance of an operative angiogram.
Despite repeated thrombectomies, graft patency could not be obtained for longer than 5
min. Two patients presented with spontaneous arterial thrombosis. One had acute
ischemia of the right lower extremity with angiographic demonstration of multiple areas
of thrombosis of the distal superficial femoral, popliteal, and tibial systems. The other ■§

presented with ischemia of one arm and both legs secondary to extensive thromboses of |

the brachial artery and its branches of one arm and the femoral, popliteal, and tibial a

systems of both legs. Extensive arteriography of the entire aorta and cardiac evaluation c

failed to reveal a proximal origin of embolic material. Another patient had spontaneous <

thrombosis of the tibial outflow vessels to the foot while undergoing an extended >9

profundoplasty. What distinguishes these patients from those with thrombotic occlusive J

disease secondary to atherosclerotic disease is the unique history, the distribution of «

occluded vessels, unusual angiographic findings, and absence of any proximal source of |

embolic material. Often, clot formation in the operative field despite heparin adminis- @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HYPERCOAGULABLE STATES AND THROMBOSIS

161

Figure 1 A. Patient with thrombosed femoroperoneal graft the night following surgery. Following
graft thrombectomy, patient had evidence of residual thrombi, as noted by arrows. B. Rethrombosis
of graft. More widespread residual thrombi are evident. Antithrombin III deficiency was diagnosed
and treated, resulting in long-term (>24 months) patency.

tration is the first clue that the patient may have an AT III deficiency. The presence of
multiple thrombi on operative angiograms is suggestive of a clotting abnormality.

B. Diagnosis

Results of routing coagulation tests are normal in patients with AT III deficiency. Generally,
reductions in AT III are measured by both immunological (tests that measure the total
amount of the protein) and functional (tests that measure activity of the AT III molecules)
assays. The best screening test is the antithrombin heparin cofactor assay (35,36).

Initially, patients with repeated episodes of venous thrombosis were identified because
of the reduction of AT III to levels 50-60% of normal values. Subsequent research has
identified patients with arterial thrombosis secondary to low AT II levels. Lynch et al. (37)
demonstrated a correlation between low preoperative plasma functional AT III levels and
the occurrence of postoperative thrombotic complications following cardiac and vascular
surgery. Thrombotic complications included arterial thrombosis, graft thrombosis, deep
venous thrombosis (DVT), cerebral vascular thrombosis, spinal infarction, and embolic
cortical blindness.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

162 TOWNE

A decrease in AT III levels causes the increase thrombotic tendency in patients taking
oral contraceptives. Sagar et al.(38) demonstrated that AT III activity was significantly
lower in patients taking oral contraceptives than in control patients. AT III activity fell in
both the contraceptive and control groups, but the decline was greater in patients taking
oral contraceptives. The only patients in whom DVT developed postoperatively as
determined by the (Iodine 125) fibrinogen test were those taking oral contraceptives. Of
the 31 patients taking oral contraceptives, 5 had an AT III activity level below 50%; in 3 of
these patients, DVT developed.

More recent study has demonstrated that the administration of heparin tends to lower
the AT III level. Conrad et al. (39) demonstrated that it is the presence of heparin and not
the rate of administration that determines the decrease in AT III; he noted that both
subcutaneous and intravenous heparin causes AT III levels to drop the same amount. Since
heparin is dependent on AT III for its antithrombotic action, the AT Ill-lowering effect
of heparin in patients with an already low AT III concentration probably indicates that
patients are at risk of thrombosis for two reasons. First, heparin is relatively ineffective in
patients with low levels of AT III. Second, because heparin binds with AT III, the already
low level is decreased even further, possibly to dangerous levels. This is the theoretical basis
for the paradoxical thrombotic episodes occasionally seen in patients following cessation
of heparin. Since heparin administration decreases the level of AT III, sometimes to sig-
nificantly dangerous levels, cessation of heparin is followed by a period when the patient
is hypercoagulable because the lower AT III level is not counteracted by the heparin. This
is why warfarin administration should be overlapped with heparin cessation when treating
thrombotic problems. Patients with a congenital AT III deficiency should receive chronic
long-term warfarin therapy because of the risk of recurrent thrombolic episodes. In addi-
tion to its anticoagulant effect, warfarin increases the level of AT III by an as yet unde-
termined mechanism. More recently, purified AT III can be obtained by recombinant
techniques, which will in the future change the treatment of this deficiency. AT III con-
centrate was used as factor-specific replacement by Tengborn and Bergquist (40) in patients
with AT III deficiency.

IV. DEFECTS IN THE FIBRINOLYTIC SYSTEM

The fibrinolytic system has become better understood in recent years and has been found
to be the source of coagulation abnormalities. The components of the fibrinolytic system
include plasminogen; plasminogen activators, including human tissue-type plasminogen
activator (t-PA) and urokinase, and inhibitors directed against plasminogen activators,
plasmin inhibitor (the most important of which is a-antiplasmin) and cellular plasmin
inhibitors, which have been identified in platelets and endothelial cells and are very poorly
characterized at the present time (41,42). The degradation of fibrin is normally carried out ■§

by the proteolytic enzyme plasmin, which is formed from the proenzyme plasminogen by g

the activation action of plasminogen activators such as t-PA and urokinase. The process as

is regulated at many levels, resulting in localized plasmin formation at the fibrin surface c

t-PA is the most important activator and is produced and released from vascular <

endothelium (43). >9

Plasminogen is a normal plasma protein consisting of a single polypeptide chain, with ^

a molecular weight of 90,000-94,000 Da (44). Thin-layer gel electro-phoresis coupled with «

immunofixation can demonstrate up to 10 different forms of plasminogen, with each |

variant having a glutamic acid as its terminal amino acid. Plasminogen is converted to @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HYPERCOAGULABLE STATES AND THROMBOSIS 163

plasmin by activators, many of which are released from endothelial cells. Plasmin, a serine
proteinase, is an important member of the fibrinolytic system that acts by cleaving
fibrinogen and fibrin (45—47).

The biosynthesis of plasminogen and fibrinolytic inhibitors is probably under genetic
control. In 1978, Aoki et al. (48) reported a patient with recurrent thrombosis who had a
hereditary molecular defect of plasminogen. This was followed by similar reports by
Kazama et al. (49) in 1981 and Soria et al (50) in 1983. These authors demonstrated that
abnormal plasminogen does not have the functional ability of normal plasminogen,
resulting in a discrepancy between the biological activity and the amount of plasminogen
detected in the serum by radioimmunoassay. These patients had normal concentrations of
plasminogen antigen, with approximately one-half of the activity of normal plasminogen.
Using electrofocusing techniques coupled with immunofixation and zymograms, they were
able to identify ten additional bands, each of which was located on the basic side in close
proximity to the corresponding normal band.

Determination of amino acid sequence has demonstrated defects in the arginine 516-
valine bond and the substitution of alanine 600 by threonine (51). The major function of
the fibrinolytic system in vivo is the limitation of fibrin deposition. A reduction of
fibrinolytic activity may provoke a thrombotic tendency by allowing the growth and
development of thrombi after the initiating thrombotic event. Most patients with
abnormal plasminogen are characterized by a normal antigen concentration and decreased
functional activity. Liu et al. (52) reported a plasminogen characterized by both low
functional activity and low antigen concentration and called it plasminogen San Antonio.

Ikemoto et al. (53) reported that the genetic characteristics of this disorder follow an
autosomal codominant inheritance pattern, with both alleles being completely expressed.
The results of the study of two families in our series concur with these findings. The clinical
history of recurrent phlebitis in one or our patients and his sister supports the genetic
aspect of this disease.

The immunoelectrophoresis technique is quicker, simpler, and less costly than iso-
electric focusing and is more applicable to the screening of large groups of patients (54).
The significance of abnormal plasminogen is uncertain. It is present in 10% of the normal
population and its presence does not ensure that a patient will experience thrombotic
complications. More likely, the presence of abnormal plasminogen results in the relative
defect of the fibrinolytic system, which places the patient at increased risk should he or she
be in a thrombosis-prone situation. Our data suggest, however, that once a thrombotic
episode occurs, it is likely to recur, emphasizing the need to identify and treat these
patients with long-term warfarin therapy.

A. Clinical Presentation ■%

We have noted thrombosis occurring on both the arterial and venous sides of the |

circulation in patients with abnormal plasminogen. In our initial report of 8 patients, a

the age of onset of the first thrombotic episode ranged from 21 to 57 years (54); 3 had c

venous thrombosis, 2 had spontaneous arterial thrombosis, 2 had occlusion of an arterial <

reconstruction in the early postoperative period, and 1 had separate episodes of both >9

arterial and venous occlusions. %

Thrombosis involving the venous system occurred in 4 patients; 2 had complete «

obstruction of the iliofemoral venous segment and inferior vena cava, 1 had primarily |

popliteal vein thrombosis, and the remaining patient had axillary and subclavian vein @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

164

TOWNE

thrombosis. Of these patients, 2 had concomitant pulmonary emboli, which occurred in 1
patient 4 months after the thrombotic event. Arterial thrombosis occurred in 5 patients; 2
presented with spontaneous thrombosis of the iliofemoral segment. Following thrombec-
tomy with Fogarty catheters, there was no evidence of inflow obstruction, and a complete
evaluation for the proximal source of emboli was negative (Fig. 2). Postoperative
occlusions of arterial reconstructions occurred in 2 patients. In the first patient, the site
of Dacron patch angioplasty of the vertebral artery orifice became occluded the night
following surgery. At reexploration, no stricture of the repair was found, and the

I
•5

Figure 2 A. Venogram demonstrating subclavian and axillary vein thrombosis in a patient who
had separate episodes of arterial and venous thrombosis. B. Angiogram of patient 9 months later,
demonstrating occlusion of the left common iliac artery. C. Angiogram obtained 37 months later,
showing reocclusion of the common iliac artery. (From Ref. 54.)

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

HYPERCOAGULABLE STATES AND THROMBOSIS

165

thrombosis was limited to the area of the patch angioplasty without distal propagation. In
the second patient, thrombosis occurred in a saphenous vein femoral-posterior tibial graft
the night following surgery. At reexploration, the thrombus formation was limited to the
graft and did not extend to the runoff vessels. Complete angiograms following both initial
surgery and reoperation demonstrated technically satisfactory anastomoses. Postopera-
tively, the graft remained patent with chronic progressive thrombosis of the runoff vessel
in both legs. This patient was not diabetic and initially presented with a 6-month history of
rest pain in both feet that progressed to digital gangrene. The ankle brachial index (ABI)
of the right lower extremity was 0.95, with a toe pressure of 44 mmHg. The ABI of the left
leg was 0.65, with a toe pressure of 10 mmHg. Arteriography demonstrated a normal
aortoiliac system and a 50% stenosis of the left superficial femoral artery, but occlusion of
the pedal and metatarsal arteries was present.

Of the 8 patients in the study of Towne et al. (55), 6 had recurrent thrombosis, 2 had
three recurrences, and 4 had two recurrences. The interval between thrombotic episodes
ranged from 4 to 36 months. Significantly, 5 patients who had recurrent thrombosis were
treated with warfarin following the first episode. Recurrent episodes of thrombosis
occurred 2 weeks to several months following cessation of warfarin therapy, and recurrent
thrombosis did not develop in any patient while he or she was receiving anticoagulation
therapy. We subsequently identified 4 patients who had abnormal plasminogen, as
detected by an abnormal arc on immunoelectrophoresis in whom severe thrombosis in
the upper extremities developed (Fig. 3) (55). The lack of atherosclerosis in the upper
extremities as well as the absence of any proximal embolic source further points out the
sometimes catastrophic consequences that patients with abnormal plasminogen may
experience. In our experience of over 30 patients in whom we have detected abnormal

I
•5

Figure 3 A. Nonoccluding thrombi of subclavian artery occurring several days after carotid
angiography from a femoral approach. B. Evidence of distal embolization of the distal radial artery
and palmar arch in same patient. (From Ref. 55.)

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

166 TOWNE

plasminogen over the last 5 years, recurrent thrombosis has developed in only 1 patient
while receiving warfarin therapy.

B. Methods of Testing

A complete coagulation profile on each patient should be performed, including tests of
platelet aggregation, prothrombin time, partial thromboplastin time, fibrinogen level, and
platelet count. Functional assays of AT III and plasminogen a2 antiplasmin should be
performed. Measurements of antigenic activity levels of AT III, plasminogen, a 2 anti-
trypsin, and « 2 macroglobulins likewise should be obtained. With immunoelectrophoresis,
abnormal plasminogen presents as an abnormal band that is separate on the electro-
phoretic pattern located nearer the anode and distinct from the normal band. We have
also noted one patient in whom plasminogen was demonstrated as separate from the main
band but was not confined to a distinct band. Indeed this may represent still another
species. Several investigators are involved in ongoing studies to further characterize the
molecular defect in these plasminogens and to assess the functional impairment. This
research requires rather sophisticated techniques to determine amino acid sequencing and
to test the functional ability of the various components of the plasminogen molecule.

1 . t-Pa and Anti-t-Pa

With the development of a method to measure t-PA, investigators have discovered that
levels of t-PA can vary in relation to the occurrence of thrombotic disease (56). Also, the
presence of an anti-t-PA that counteracts the effects of t-PA has been detected (57,58).
Several studies have identified patients who are thrombosis-prone because of increased
levels of anti-t-PA (59-61). Both the mechanisms and the effect of alterations of these
mechanisms are poorly understood at the present time. Wiman (56) first developed the test
to measure t-PA. In a study of patients with DVT, he found that 40% of his patients had a
reduced fibrinolytic potential, which was found to be caused by a reduced capacity to release
t-PA or an increased plasma level of an anti-t-PA or a combination of these (60). He also
noted a significant correlation between plasma anti-t-PA and the levels of serum trigly-
cerides in patients below the age of 45 with myocardial infarction. Obviously these are
primarily data, but they emphasize the need for ongoing investigation to determine more
precisely the role of the fibrinolytic system in the pathogenesis of thrombotic disorders.

V. PROTEIN C DEFICIENCY

Protein C is a vitamin K-dependent proenzyme that is involved in the control of clotting
and fibrinolysis. Protein C itself is activated by thrombin, but slowly. This activation is
increased up to 20,000-fold when thrombin forms a complex with an endothelial cell

membrane called thrombomodulin (Fig. 4). Activated protein C combined with phospho- ■§

lipids, calcium, and protein S inactivates the cofactors of the two rate-limiting steps of g

coagulation, factors Va and Vila (62-64). a

Protein S is likewise a vitamin K-dependent factor. It acts as a cofactor for the c

anticoagulant activity of activated protein C by promoting its binding to lipid and platelet <

surfaces, thus localizing protein C activity (65,66). Protein C, in conjunction with protein >9

S, also acts as a profibrinolytic agent by increasing plasmin activity through the J

inactivation of the major inhibitor of t-PA (62,63). °

Heterozygous protein C deficiency is inherited in an autosomal dominant fashion. In |

hereditary protein C deficiency, the homozygous state is associated with a very high risk of @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HYPERCOAGULABLE STATES AND THROMBOSIS 167

Protein C

Thrombin Thrombomodulin

Endothelium

Inactivated x P™*" s Activated ^ Fibrinolysis

factors Va and Villa protein C

Protein C
inhibitor

Inactivated
protein C

Figure 4 Protein C is produced in the liver and circulates in blood in inactive form. When
thrombin becomes bound to endothelial cofactor (thrombomodulin), the complex that is formed
rapidly activates protein C. Activated protein C is a potent plasma anticoagulant. It inactivates
cofactors of two rate-limiting steps of coagulation (factors Va and Villa) and enhances fibrinolysis.
These processes require the presence of protein S. Protein C is inactivated by protein C inhibitor in a
one-to-one fashion. (From Ref. 71.)

thrombosis (67,68). It usually presents as massive venous thrombosis in the neonatal
period and is often fatal. The in utero survival of affected infants may reflect the protection
afforded by maternal transfer of protein C or the reduced synthesis of other procoagulants
by the fetal liver, which thus compensates for the deficiency of protein C.

In the heterozygous form of the deficiency, a protein C level of 50% is sufficient to
predispose individuals to venous thrombosis (69). The incidence of thrombophlebitis in
patients who are heterozygous for this deficiency is uncertain. Some kindreds have been
identified in which there is a very high incidence of venous thrombosis (up to 80%) by the
age of 40, and there are others in which the occurrence of thrombosis is sporadic (69,70).
Acquired protein C deficiency can be observed in patients in the acute phase of
thrombosis, in patients with disseminating intervascular coagulation, in patients with
liver disease, and in postoperative patients.

Protein C deficiency generally manifests itself with venous thrombosis, either as DVT
of the lower extremity, often accompanied by pulmonary embolism, or as mesenteric
venous thrombosis. We recently reported 5 patients (age range, 28-41 years) with protein
C deficiency: 4 had DVT of the lower extremity as the initial thrombotic event and 1 had ■a

mesenteric venous thrombosis with small bowel necrosis (71). Two patients had recurrent |

lower extremity thrombosis, which was bilateral in one. One patient experienced only one a

clinical episode of DVT, but venous stasis ulceration developed, suggesting multiple c

episodes of subclinical phlebitis. One patient had a pulmonary embolus. Green et al. (72) <

evaluated 8 consecutive patients with splanchnic venous thrombosis and demonstrated >9

decreases in the levels of AT III and protein C in all. They were unable to document J

whether the low levels of protein C and AT III were a result or cause of the thrombosis. Q

Two patients had had a history of venous thrombotic problems, and evaluation of six |

patients following a period of 1-6 months revealed a persistent low level of protein C, @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

168 TOWNE

which certainly suggest a congenital etiology. The only case of arterial thrombosis
secondary to protein C reduction was reported by Coller et al. (73); they treated a patient
who experienced the onset of the first of three episodes of thrombophlebitis at age 23. At
age 27, the patient had a pulmonary embolus; at 28, he had a myocardial infarction and
superficial femoral artery thrombosis; and at age 31, simultaneous radial and ulnar artery
occlusions developed, followed shortly by the development of left ventricular thrombus,
which caused embolus to the leg.

A. Methods of Testing

Standard testing includes a radiolabeled Laurell electroimmunoassay to determine human
protein C antigen in plasma samples. Normal values should be between 70 and 130% of
normal activity. In our patients with venous thrombosis, the level of protein C ranged
from 34 to 67% (71). As with the evaluation of all patients with unusual or unexplained
thrombosis, measurement of AT III and protein S and routine coagulation studies should
be performed simultaneously. We evaluated the family of our patient with mesenteric
venous occlusion for protein C deficiency and found the results to indicate an autosomal
dominant type of transmission. No thrombotic episodes have been reported by other
family members with low protein C levels. Since thrombosis does not develop in all family
members with low protein C levels, asymptomatic patients with low levels of protein C
should be monitored closely and should not receive prophylactic anticoagulant therapy.
However, they should receive prophylactic anticoagulants preoperatively if major surgery
or prolonged immobilization is required. In those in whom thrombotic events develop, the
onset typically occurs between 15 and 30 years of age. This delay in onset of the first
thrombotic episode and the fact that protein C rarely causes arterial thrombosis — as
contrasted with our experiences with either abnormal plasminogen or AT III abnormal-
ities — are not well understood. It may be that protein C deficiency requires the slower-
moving blood and increased endothelial surface area found in the venous system to
manifest itself. However, when protein C deficiency is homozygous, thrombosis is wide-
spread, resulting in death in infancy unless treated.

Because of the risk of recurrent thrombotic events, with the possible sequelae of
pulmonary emboli and venous stasis disease, long-term warfarin therapy is recommended.
No loading dose should be administered, as this could precipitate warfarin-associated skin
necrosis (65,66). Such necrosis occurs 2-5 days following the initiation of warfarin therapy
and presents as an erythematous patch on the skin that progresses rapidly to a hemorrhagic
area, which can become gangrenous. There is a propensity for involvement of the breasts,
abdomen, buttocks, and thighs. The proposed mechanism is one of a transient hyper-
coagulable state that is created by bolus loading doses of warfarin given to initiate
anticoagulation. Because of it short half-life, protein C levels fall faster than levels of factor
X and prothrombin; and thus the inhibitory effect of protein C on the coagulation cascade is 13

further diminished. If these levels fall below a critical point, the procoagulant effects of the 8

coagulation cascade proceed unabated and thrombosis ensues. Administration of oral -|

warfarin 5 mg daily should be started to gradually attain a prothrombin time of 1.5-2 times •§

the control value. Heparin therapy and warfarin therapy should overlap by 4-5 days. ^

VI. PROTEIN S DEFICIENCY I

Protein S is a vitamin K-dependent protein that functions as a cofactor of anticoagulant §

activity of activated protein C. The liver is the major location of synthesis, although more ©

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HYPERCOAGULABLE STATES AND THROMBOSIS 169

recently the endothelial cells and megakaryocyte were identified as other sites of synthesis.
Protein S functions by expediting the binding of activated protein C to the lipid and
platelet surfaces. To date, only patients with heterozygous protein S deficiency have been
reported (68). Symptomatic patients often have protein S levels 50% of the normal value,
and — like protein C deficiency — protein S deficiency primarily causes venous thrombosis
(74-76). Protein S deficiency has been estimated by some to be the cause of approximately
10% of cases of spontaneous venous thrombosis. Coller et al. (73) also reported the only
known case history of a protein S-deficient patient having arterial occlusive problems.
This patient had had recurrent episodes of thrombophlebitis over several years, resulting
in venous stasis disease. At age 21, he experienced thrombotic problems in the legs, which
resulted in a below-knee amputation. As with patients with protein C problems, patients
with protein S deficiency have clotting abnormalities that tend to be recurrent; therefore it
is essential that they remain on long-term warfarin therapy. The association of deficiencies
in protein C and its cofactor protein S with hypercoagulable states has only recently been
appreciated. Data now suggest that the incidence of protein C and protein S deficiencies is
more common than either AT III or plasminogen abnormalities. In a recent report
evaluating 139 individuals who had at least one major venous thrombotic event, 7% were
deficient in protein C, 5% were deficient in protein S, 2% were deficient in plasminogen,
and 3% were deficient in AT III (77). A majority (79%), however, had no coagulopathy
detectable with current testing methods.

VII. ANTIPHOSPHOLIPID ANTIBODIES

Lupus-like anticoagulants are IgG or IgM antibodies that are directed against phospho-
lipids participating in coagulation disorders. They are present in 16-33% of patients with
lupus erythematosus, but they are also associated with a variety of other disorders and are
even found in normal individuals (78-80). These antibodies belong to a family of
antiphospholipid antibodies that were initially detected because of their effect in vitro
on the prolongation of plasma coagulation times. Most commonly, there is a prolongation
of activated partial thromboplastin time and, in some patients, also a prolongation of
prothrombin time. There have been only rare reports of bleeding tendencies related to the
demonstration of a lupus-like anticoagulant; however, in the last decade, there have been
an increasing number of reports of the presence of lupus-like anticoagulants associated
with abnormal thrombosis in both the arterial and venous systems, spontaneous abortion
secondary to placental thrombosis, cerebrovascular accidents, and thrombocytopenia.
Lupus-like anticoagulants also cause false-positive tests for syphilis. On occasion, lupus-
like anticoagulants can develop after administration of phenothiazines, procainamide, or
penicillin; following viral infections in children; and in patients with AIDS suffering from
Pneumocystis carinii pneumonia. E

A. Clinical Presentation 3

Recurrent thromboses have been reported in about one-third of patients with lupus-like
antiocoagulant (81). The most common manifestation is venous thrombosis, usually
involving the lower extremities. Pulmonary hypertension caused by recurrent pulmonary
emboli or intrapulmonary thrombosis may develop. Repeated strokes have been reported
in 15-55% of these patients. Obstetric complications (e.g, spontaneous abortions, intra-
uterine growth retardation, fetal death) occurring in the second and third trimesters have

c

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

170 TOWNE

been reported in 25-35% of women with lupus-like anticoagulant (78-80). Ahn et al. (82)
in a study of patients with lupus-like anticoagulant who were undergoing surgery, noted
that 9 of 18 vascular surgery procedures were complicated by thrombosis. Seven of these
patients suffered multiple postoperative thrombotic complications, resulting in amputa-
tion in three. The mechanism of action of lupus-like anticoagulants is not known. Several
theories have been suggested, including an inhibitory effect on prostacyclin (PGI2), which
is a potent in vivo inhibitor of platelet aggregation. IgG fractions with lupus-like
anticoagulant activity have been shown experimentally to block the production of
prostacyclin in rat aortic endothelial cells (83). Other investigators suggest that lupus-like
anticoagulant inhibits protein C activation, which is important in preventing thrombosis.
Tsakiris et al. (84) believe that the inhibition of the catalytic activity of thrombomodulin
might be explained by the direct attachment of lupus-like anticoagulant to thrombomo-
dulin or to adjacent phospholipids of the cell membrane, preventing thrombin and/or
protein C from binding to thrombomodulin.

B. Diagnosis

Often the only indication that a patient has lupus-like anticoagulant is an abnormally
prolonged activated partial thromboplastin time. On occasion, such a patient can also
have a prolonged prothrombin time. An abnormal rabbit brain neutralization procedure
and an ELISA for presence of anticardiolipin antibodies can more precisely identify lupus-
like anticoagulant (82).

C. Treatment

Because the precise mechanism by which lupus-like anticoagulant causes intravascular
thrombosis is not known, treatment has varied, including the administration of anti-
platelet medications (aspirin and dipyridamole), anticoagulation with warfarin and
heparin, and the administration of steroids. The basis for treatment with antiplatelet
medication is that some researchers believe that the lupus-like anticoagulant causes a
decrease in the availability of arachidonic acid, which is necessary for the synthesis of
prostacyclin inhibitor or platelet aggregation in vessel walls. In obstetric patients, it has
been reported that steroid and aspirin administration is effective in preventing sponta-
neous abortion. Prednisone has been shown to suppress production and/or activity of
lupus-like anticoagulant as measured by lessened prolongation of the activated partial
thromboplastin time. Until more information is available, we prefer to initiate antiplatelet
therapy with aspirin and dipyridamole before surgical procedures. We administer dextran
routinely in all vascular reconstructions, and patients are given heparin perioperatively.
Postoperatively, heparin therapy was converted to warfarin therapy.

VIM. ACTIVATED PROTEIN C RESISTANCE 1

I
A recurring theme throughout this chapter is the fact that the procoagulant and anti- c

coagulant properties of blood are delicately balanced by a complex system of cofactors and <

inhibitors. The thrombomodulin/protein C anticoagulant pathway is an essential anti- >9

coagulant system. As thrombin is generated at sites of vascular injury, it activates and J

aggregates platelets and clots fibrinogen. It also binds to the endothelial membrane protein «

thrombomodulin. Upon binding to thrombomodulin, thrombin takes on anticoagulant |

properties by activating protein C. The activated protein C (APC) cleaves and inactivates @

factor Va and Villa in the presence of protein S. This endothelial-based anticoagulant %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HYPERCOAGULABLE STATES AND THROMBOSIS 171

system allows blood to clot while maintaining intravascular fluidity. Defects in this anti-
coagulant pathway can provoke thrombosis; indeed, protein C and protein S deficiencies,
discussed previously, are associated with an increased risk of thrombosis in heterozygotes.

A family history of thrombotic events is frequently obtained in young adults with
venous thrombosis; however, the inherited deficiencies in anticoagulant proteins, such as
protein C and protein S, are found in only about 5% of patients (85). APC resistance is
another risk factor for venous thrombosis that is frequently found in these patients. It is,
in fact, the most common genetic risk factor for venous thrombosis described to date (86).
It is caused by a point mutation in the factor V gene, which causes increased resistance to
the anticoagulant effect of activated protein C. This defect was discovered in 1993 by
Dahlbach et al. and has been named factor V Leiden (85).

Dahlbach originally postulated that defect in the protein C pathway interfered with the
anticoagulant action of APC. He devised assays to test this possibility, in which the clotting
time of blood was measured in the presence and absence of exogenous APC. In the normal
response, the clotting time was prolonged in the presence of APC because of the inactivation
of factors Va and Villa. A defect was detected as a failure of prolongation of the clotting
time resulting from resistance to added APC. Dahlbach showed that this test detects an
autosomal dominant trait associated with thrombosis. Further work done by Bertina and
his group demonstrated that the phenotype of APC resistance is associated with a he-
terozygous or homozygous single point mutation in the factor V gene, which predicts the
synthesis of a factor V molecule that is not properly inactivated by APC (factor V Leiden)
(86). Other data confirming these results were published by Zoller and Dahlbach who
studied 50 Swedish families with inherited APC resistance (87). They found that the specific
point mutation in the factor V gene was present in 47 of 50 families. In their study, by age 33
years, 20% of the heterozygous and 40% of the homozygous patients had had manifes-
tations of venous thrombosis.

Factor V Leiden is present in 3-7% of all Caucasians but is more rare in other ethnic
populations (88). It is present in up to 20% of unselected patients with DVT and confers a
5- to 10-fold increased risk of thrombosis in heterozygotes; homozygotes have a 50- to
100-fold increased risk (88).

The laboratory diagnosis is made by measuring the responsiveness of plasma to APC as
the ratio of two activated partial thromboplastin times, one in the presence of APC and one
in its absence. The APC sensitivity ratio is normalized to the ratio obtained with reference
plasma. Resistance to APC is defined by an APC sensitivity ratio of <0.84 (86). A more
recent way of identifying this factor V resistance to activated protein C is by a direct assay for
the factor V molecule, which is resistant to inactivation by APC (factor V Leiden). The
question then arises what can be done about these point mutations, which cause factor V to
be resistant to activated protein C. It is clear that this is a major risk factor for
thromboembolic disease; however, the majority of patients with these mutant proteins will ■§

not suffer thrombosis. The risks of lifelong anticoagulation therapy in an asymptomatic g

patient must be weighed against the benefit of preventing infrequent but devastating »

thrombotic attacks. At this point, it would be a logical course of action to treat those c

patients who have already suffered thrombotic attacks with long-term warfarin therapy. <

Jj

IX. HYPERHOMOCYSTEINEMIA I

Hyperhomocysteinemia is a thiol-containing amino acid derived from the metabolism of |

methionine. Remethylation of homocysteine with a methyl group from methyltetrahydro- @

folate (MTHF) reproduces methionine. This reaction is B 12 -dependent and is catalyzed by %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

172 TOWNE

methyltetrahydro folate reductase (MTHFR). Homocysteine can also be metabolized via
transulferation to cysteine, a reaction that requires cystathionine synthase and B 6 cofactor.
Homozygous deficiency of cystathionine synthase produces classic homocystinurea, in
which levels of homocysteine in the blood and urine are quite elevated. This defect is asso-
ciated with early onset of vascular disease and venous thrombosis; however it is quite rare.

Attention has recently been turned to evaluating other factors that might contribute to
elevated levels of homocysteine in the plasma and determining their association with
thrombosis. Both a defect in MTHFR or cystathionine synthase can be associated with
increased homocysteine levels in the blood. Elevated levels of homocysteine have been
found to be present in 10-25% of patients with venous thrombosis, which is approx-
imately 2.5 times the number of control patients with increased homocysteine levels
(89,90). In one multicenter trial, homocysteine levels were followed in 264 patients with
documented DVT after their 3-month course of oral anticoagulants was stopped. Of these
patients, 25% had increased homocysteine levels, and these patients had a 19% recurrence
of DVT after 2 years, in comparison to a 6.3% recurrence rate in those patients without
increased homocysteine levels (90). It appears that elevated homocysteine levels increase
the risk of venous thrombosis up to fourfold (90).

Normalization of homocysteine levels can be accomplished by giving folic acid with or
without B 12 . It remains to be proven, however, that the normalization of homocysteine
levels confers any benefit. Treatment, therefore, is unclear. In asymptomatic patients,
vitamin supplementation is probably reasonable. In patients who have had DVT, anti-
coagulation with warfarin or low-molecular-weight heparin should be undertaken, but the
time course of this therapy is uncertain. There is evidence that these patients are at high
risk for recurrent DVT; for them, lifelong anticoagulation should be considered.

X. UNEXPLAINED THROMBOSIS-GUIDELINE FOR IDENTIFYING
HYPERCOAGULABLE PATIENTS

A thorough patient history remains the most important means of identifying patients with
potential hypercoagulable disorders. Patients should be asked about previously unexplained
thromboses experienced by themselves or by family members. Patients with hypercoagu-
lable syndromes will often report episodes of thrombophlebitis in early adulthood. Of
particular importance are those episodes of thrombophlebitis without any contributing
factors for their development (e.g., long leg fractures, prolonged immobilization, bed rest
because of illness). Hypercoagulable disorders become even more significant in patients
with recurrent episodes of thrombophlebitis. Likewise, a history of arterial thrombosis,
especially if the episodes occurred at a young age, is an indicator of a coagulation disorder.
Eldrup-Jorgensen et al. (91) found a 30% incidence of coagulation abnormalities in patients .

below 51 years of age undergoing vascular reconstruction. Abnormal clotting syndromes |

noted were protein S deficiency, protein C deficiency, presence of lupus-like anticoagulants, 8

and plasminogen deficiency. The incidence of arterial graft thrombosis in hypercoagulable J

patients was 20% at 30 days, which is markedly increased from what one would expect from
this type of vascular reconstruction.

A. Clinical Presentation

With experience, one has a sense for what kinds of reconstructions should work and has
some expectations concerning the types of problems that can occur. Likewise, one devel-

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

c

I

HYPERCOAGULABLE STATES AND THROMBOSIS 173

ops a feel for what are typical presentations of atherosclerotic occlusive disease. Unusual
or unexplained thromboses (e.g., a thrombosed suprarenal aorta, upper extremity
thrombosis, or a total tibial artery occlusion in a patient who is neither diabetic nor has
any evidence of atherosclerotic occlusive disease elsewhere) should alert the surgeon to
consider a hypercoagulable disorder as the cause. Unusual x-ray findings — in particular,
occlusions seen in young patients or in one extremity when the other extremity has no
evidence of any disease — should trigger an investigation of the coagulation system.

The role of screening for hypercoagulable states in vascular surgery patients is difficult
to ascertain. Donaldson et al. (92) found a 9.5% overall incidence of vascular surgery
patients with abnormal test results indicating potential hypercoagulability. The three most
common entities demonstrated were heparin-induced platelet aggregation, lupus-like
anticoagulants, and protein C deficiency. The incidence of infrainguinal graft occlusion
within 30 days was 27% in the hypercoagulable group compared with 1.6% in the
noncoagulable group. Currently we do not perform routine screening to detect patients
with hypercoagulable states. We depend on patient history and clinical evaluation to
identify those patients who may be hypercoagulable, which is probably more cost-effective
and efficient than routine screening.

The most difficult experience for a vascular surgeon is dealing with unexplained
thrombosis that occurs intraoperatively. Often this occurs during late evening or nighttime
hours, when support from the coagulation laboratory is not available. The first step, if
indeed heparin has been given, is to determine whether there is clotting in the operative
field, which would indicate an AT III deficiency, since AT III is essential for heparin's
anticoagulant effect. The anesthesiologist should then test the heparin effect by determin-
ing partial thromboplastin time or by performing one of the other variety of tests to
measure heparin anticoagulation. The next step is to obtain a platelet count. If it is higher
than 100,000/mm 3 and the activated clotting time (ATC) is not prolonged, the problem is
presumed to be the antithrombin system. The patient is then given two units of fresh
frozen plasma, with two units given every 12 h for 5 days. AT III deficiency is usually
confirmed the next day, with tests done on blood drawn the day before the administration
of fresh frozen plasma. Patients with AT III deficiency are maintained on long-term
warfarin therapy. If the platelet count is less than 100,000/mm , we presume that heparin-
induced platelet aggregation has developed. The patient's history should be carefully
examined to try to document the prior administration of heparin. At this time we
administer a 50-mL bolus dose of dextran and continue dextran therapy at 25 mL/h.
The heparin is reversed with protamine, and platelet aggregation abnormality is confirmed
in the morning. Warfarin treatment is continued for 3 weeks to 6 months.

If the platelet count is greater than 100,000/mm 3 and the ACT is prolonged, we
presume that the patient has some other sort of hypercoagulable state, which includes
fibrinolytic abnormalities as well as potential problems with protein C, protein S, and ■§

lupus-like anticoagulants. In these patients we institute continuous heparin therapy both g

intraoperatively and postoperatively and give them two units of fresh frozen plasma. Fresh »

frozen plasma is "shotgun" therapy for a wide variety of coagulation abnormalities. c

In the operating room, before the institution of any therapy, blood should be drawn <

for coagulation tests; it should be kept in mind that many of these tests (e.g., plasminogen >9

electrophoresis and determination of protein C, protein S, and lupus-like anticoagulant) J

are quite involved, sometimes taking days to a week at some centers. However, if the blood °

is properly handled, spun down, and frozen, the tests can be done routinely. Our policy is |

to repeat all abnormal tests in 5 days. One of the problems in diagnosing coagulation @

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

174 TOWNE

abnormalities accurately is that in the process of clotting, clotting factors can be consumed
and abnormalities may be the result of clotting and not the cause of it. For all factors that
still demonstrate abnormal values at 5-7 days, tests are repeated at 1 month. Patients who
have persistently abnormal values are then labeled truly hypercoagulable.

McDaniel et al. (93) have noted the change in coagulation factors with operation.
They determined that AT III levels fell on the third postoperative day and subsequently
returned to normal by 1 week postoperatively. AT III declined from a mean preoperative
level of 110-71% on the third postoperative day. This value returned to normal by the
seventh postoperative day, when it was 95% or at the normal. activity level. This variability
demonstrates the dynamic aspect of the clotting system and points to the danger of
attaching significance to just one isolated laboratory finding. In most patients who sustain
complications because of hypercoagulable states, warfarin therapy is instituted in the
perioperative and postoperative period. In patients with heparin-induced platelet aggre-
gation, therapy can usually be stopped after 3 months; however, we have recommended
prolonged administration in patients with protein C or S deficiency, AT III deficiency, and
plasminogen abnormalities because of the risk of recurrent thrombosis.

REFERENCES

1. Babcock RB, Dumper CW, Scharfman WB. Heparin-induced immune thrombocytopenia. N
Engl J Med 1976; 295:237-241.

2. Baird RA, Convery RF. Arterial thromboembolism in patients receiving systemic heparin
therapy. J Bone Joint Surg 1977; 59:1061-1064.

3. Bell WR, Romasulo PA, Alving BM, et al. Thrombocytopenia occurring during the
administration of heparin. Ann Intern Med 1976; 87:155-160.

4. Fratantoni JC, Pollet R, Gralnick HR. Heparin-induced thrombocytopenia: Confirmation of
diagnosis with in vitro methods. Blood 1975; 45:395^101.

5. Nelson JC, Lerner RG, Goldstein R, et al. Heparin-induced thrombocytopenia. Arch Intern
Med 1978; 138:548-552.

6. Warkentin TE, Levine MN, Hirsh J, et al. Heparin-induced thrombocytopenia in patients
treated with low molecular weight heparin or unfractionated heparin. N Engl J Med 1995;
332:1330-1335.

7. Towne JB, Bernhard VM. Hussey C, et al. White clot syndrome. Arch Surg 1979; 114:372-
377.

8. Rhodes GR, Dixon RH, Silver D. Heparin-induced thrombocytopenia. Ann Surg 1977;
186:752-758.

9. Silver D. Heparin-induced thrombocytopenia Semin. Vase Surg 1988; 1:228.

10. Laster J, Ckrit D, Walder N, et al. The heparin-induced thrombocytopenia syndrome update.
Surgery 1987; 102:763.

1 1 . Kapsch DN, Adelstein EH, Rhodes GR, et al. Heparin-induced thrombocytopenia, thrombosis,

and hemorrhage. Surgery 1979; 86:148-154. 1

12. Laster J, Silver D. Heparin-coated catheters and heparin-induced thrombocytopenia. J Vase <S
Surg 1988; 7:667-672. |

13. Rosenthanl F. Risk factors for venous thrombosis: Prevalence, risk and interactum. Semin °|
Hematol 1997; 34:171-187. ^

14. Silver D, Kapsch DN, Tsoi EKM. Heparin-induced thrombocytopenia, thrombosis, and hemor- «
rhage. Ann Surg 1983; 198:301-306. J

15. Laster J, Cikrit D, Walker N, et al. The heparin-induced thrombocytopenia syndrome: An q
update. Surgery 1987; 102:763-770. |

16. Laster J, Elfrink R, Silver D. Reexposure to heparin of patients with heparin-associated 2
antibodies. J Vase Surg 1989; 9:677-682. 2

I
§

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HYPERCOAGULABLE STATES AND THROMBOSIS 175

17. Kappa JR, Fisher CA, Berkowitz HD, et al. Heparin-induced platelet activation in sixteen
surgical patients: Diagnosis and management. J Vase Surg 1987; 5:101-109.

18. Makhoul RG, Greenberg CS, McCann RL. Heparin-associated thrombocytopenia and throm-
bosis: A serious clinical problem and potential solution. J Vase Surg 1986; 4:522-528.

19. Sobel M, Adelman B, Szaboles S, et al. Surgical management of heparin-associated
thrombocytopenia. J Vase Surg 1988; 8:395^101.

20. Latham P, Revelis AF, Joshi GP, Di Maio JM, Jessen ME. Use of recombinant hirudin in
patients with heparin-induced thrombocytopenia with thrombosis requiring cardiopulmonary
bypass. Anesthesia 2000; 92:263-269.

21. Cole CW, Bormanis J. Ancrod: A practical alternative to heparin. J Vase Surg 1988; 8:59-
63.

22. Abildguard U, Fagerhol MK, Egeberg O. Comparison of progressive antithrombin activity
and the concentration of three thrombin inhibitors in human plasma. Scand J Clin Lab Invest
1970; 26:349-354.

23. Seegers WH. Antithrombin III: Theory and clinical applications. Am J Clin Pathol 1978;
69:367-374.

24. Eseberg O. Inherited antithrombin deficiency causing thrombophilia. Thromb Diath Haemorrh
1963; 13:516-530.

25. Brozovic M, Stirling U, Hamlyn AN. Thrombotic tendency and probable antithrombin III
deficiency. Thromb Haemost 1978; 39:778-779.

26. Mackie M, Bennett B, Ogstron D, et al. Familial thrombosis: Inherited deficiency of anti-
thrombin III. Br Med J 1978; 1:136-138.

27. Marciniak E, Farley CH, DeSimone PA. Familial thrombosis due to antithrombin III defi-
ciency. Blood 1974; 43:219-231.

28. Sorenson PJ, Dyerburg J, Strotterson E, et al. Familial functional antithrombin III defi-
ciency. Scand J Haematol 1980; 24:105-109.

29. Collen D, Schetz J, DeCock F, et al. Metabolism of antithrombin III (heparin cofactor) in man:
Effects of venous thrombosis and of heparin administration. Eur J Clin Invest 1977; 7: 27-35.

30. Odeguard OR, Abildguard U. Antithombin III: Critical review of assay methods. Significance
of variations in health and disease. Haemostasis 1978; 7:127-134.

3 1 . Chan V, Chan TK, Wong V, et al. The determination of antithrombin III by radioimmuno-
assay and its clinical application. Br J Haematol 1979; 41:563-572.

32. Sas G, Blasko G, Banghogyi D, et al. Abnormal antithrombin III (antithrombin Budapest) as a
cause of familial thrombophilia. Thromb Diath Haemorrh 1974; 32:105-115.

33. Flinn WR, McDaniel MD, Yao JST, et al. Antithrombin III deficiency as a reflection of
dynamic protein metabolism in patients undergoing vascular reconstruction. I Vase Surg 1984;
1:888-895.

34. Towne JB, Bernhard VM, Hussey C, et al. Antithrombin deficiency — A cause of unexplained
thrombosis in vascular surgery. Surgery 1981; 89:735-742.

35. Bick R, Kaplan H. Syndrome of thrombosis and hypercoagulability. Med Clin North Am
1998; 82:409-458.

36. Bick RL, Pergram M. Syndrome of hypercoagualablity and thrombosis. Semin Thromb

Haemost 1994; 20(1):109. 1

37. Lynch DM, Leff, Howe SE. Preoperative AT-III values and clinical postoperative thrombosis: <S
A comparison of three antithrombin III assays. Thromb Haemost 1984; 52:42-44. js

38. Sagar S, Stamatakis JD, Thomas DP, et al. Oral contraceptives, antithrombin III activity and °|
postoperative deep vein thrombosis. Lancet 1976; 1:509-511. ^

39. Conrad J, Lecompte T, Horellou MH, et al. Antithrombin III in patients treated with >9
subcutaneous or intravenous heparin. Thromb Res 1981; 22:507-511. J

40. Tengborn L, Bergqvist D. Surgery in patients with congenital antithrombin III deficiency. Acta Q
Chir Scand 1988; 154:179-188. |

41. Salem HH, Mitchell CA, Firkin BG Current views on pathophysiology and investigations of 2
thrombotic disorders. Am J Haematol 1987; 25:463^474. 2

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

176 TOWNE

42. Towne JB. Hypercoaguable states. Semin Vase Surg 1988; 1(4):201-215.

43. Wiman B, Ljungberg B, Chmielewska J, et al. The role of the fibrinolytic system in deep vein
thrombosis. J Lab Clin Med 1985; 105:265-270.

44. Castellino FJ, Powell JR. Human plasminogen. Methods Enzymol 1981; 80:365-378.

45. Wallen P, Wiman B. Characterization of human plasminogen. Biochem Biophys Acta 1972;
257:122-134.

46. Mullertz S. Fibrinolysis: An overview. Semin Thromb Haermost 1988; 10:1-5.

47. Summaria L. Arzadon P, Bernabe P, et al. Studies on the isolation of the multiple molecular
forms of human plasminogen and plasmin by isoelectric focusing methods. J Biol Chem 1972;
247:469 1-4702.

48. Aoki N, Moroi M, Sakata Y, et al. Abnormal plasminogen — A hereditary molecular abnor-
mality found in patients with recurrent thrombosis. J Clin Invest 1978; 61:1186-1195.

49. Kazama M, Tohura C, Suzuki Z, et al. Abnormal plasminogen — A cause of recurrent throm-
bosis. Thromb Res 1981; 21:517-522.

50. Soria J, Soria C, Bertarnd O, et al. Plasminogen Paris I: Congenital abnormal plasminogen
and its incidence in thrombosis. Thromb Res 1983; 32:229-238.

51. Scharrer IM, Wohl RC, Hach V, et al. Investigation of congential abnormal plasminogen,
Frankfurt I and its relationshiop to thrombosis. Thromb Haemostas 1986; 55:396-401.

52. Liu Y, Lyons RM, McDonagh J. Plasminogen San Antonio: An abnormal plasminogen with
more cathodic migration, decrease activation and associated thrombosis. Thromb Haemost
1988; 59:49-53.

53. Ikemoto S, Sakata Y, Aoki N. Genetic polymorphism of human plasminogen in a human
population. Hum Hered 1987; 32:296-297.

54. Towne JB, Bandyk DF, Hussey CV, et al. Abnormal plasminogen: A genetically determined
cause of hypercoagulability. J Vase Surg 1984; 1:896-902.

55. Towne JB, Hussey CV, Bandyk DF. Abnormalities of the fibrinolytic system as a cause of
upper extremity ischemia. J Vase Surg 1988; 7:661-666.

56. Wiman B. The role of the fibrinolytic system in thrombotic disease. Acta Med Scand 1986; 715
(suppl): 169-171.

57. Chmielewska J, Ranby M, Wiman B. Evidence of a rapid inhibitor to tissue plasminogen
activator in plasma. Thromb Res 1983; 31:427^437.

58. Kruithof EKO, Tran-Thang C, Ransijn A, et al. Demonstration of a fast-acting inhibitor of
plasminogen activators in human plasma. Blood 1984; 64:907-913.

59. Hamsten A, Wiman B, deFaire U, et al. Increased plasma levels of a rapid inhibitor of tissue
plasminogen activator in young survivors of myocardial infarction. N Engl J Med 1985; 313:
1557-1563.

60. Wiman B, Lujungberg B, Chmielweska J, et al. The role of the fibrinolytic system in deep
venous thrombosis. J Lab Clin Med 1985; 105:265-270.

61. Wiman B, Chimielewska J, Ranby M. Inactivation of tissue plasminogen activator in plasma.
J Biol Chem 1984; 259:3644-3647.

62. Marlar RA. Protein C in thromboembolic disease. Semin Thromb Hemost 1985; 11:387-393.

63. Clouse LH, Comp PC. The regulation of hemostasis: The protein C system. N Engl J Med

1986; 314:1298-1303. 1

64. Stenflo J. Structure and function of protein C. Semin Thromb Hemost 1984; 10:109-121. <S

65. Kazmier FJ. Thromboembolism, coumarin necrosis, and protein C. Mayo Clin Proc 1985; s
60:673-674. Jf

66. Peterson CE, Kwaan HC. Current concepts of warfarin therapy. Arch Intern Med 1986; ^
146:581-584. |

67. Branson HE, Kate J, Marble R, et al. Inherited protein C deficiency and coumarin-responsive ^
chronic relapsing purpura fulminans in a newborn infant. Lancet 1983; 2:1165-1168. q

68. Salem HH, Mitchell CH, Firkin BG. Current views on pathophysiology and investigations of |
thrombotic disorders. Am J Hematol 1987; 25:463^174. 2

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

HYPERCOAGULABLE STATES AND THROMBOSIS 177

69. Broekmans AW, Veltkamp JJ, Bertina RM. Congenital protein C deficiency and venous
thromboembolism: A study of three Dutch families. N Engl J Med 1983; 390:340-344.

70. Griffen JG, Evan B, Zimmerman TS, et al. Deficiency of protein C in thrombotic disease. J Clin
Invest 1981; 68:1370-1373.

71. Tollefson DFJ, Friedman KD, Marlar RA, et al. Protein C deficiency: A cause of unusual or
unexplained thrombosis. Arch Surg 1988; 123:881-884.

72. Green D, Ganger DR, Blei AT. Protein C deficiency in splanchnic venous thrombosis. Am J
Med 1987; 82:1171-1173.

73. Coller BS, Owen J, Jesty J, et al. Deficiency of plasma protein S, protein C, or antithrombin
III and arterial thrombosis. Atherosclerosis 1987; 7:456^462.

74. Comp PC, Esmon CT. Recurrent venous thromboembolism in patients with a partial deficiency
of protein S. N Engl J Med 1984; 311:1526-1528.

75. Schwarz HP, Fischer M, Hopmeir P, et al. Plasma protein S deficiency in familial thrombotic
disease. Blood 1984; 64:1297-1300.

76. Rodgers GM, Shurman MA. Congenital thrombotic disorders. Am J Hematol 1986; 21:419-
430.

77. Gladson CL, Griffen JH, Hach V, et al. The incidence of protein C and protein S deficiency in
139 young thrombotic patients. Thrombosis 1985; 66:350a.

78. Espinoza LR, Hartmann RC. Significance of the lupus anticoagulant. Am J Hematol 1986;
22:331-337.

79. Tabechnik-Schor NF, Lipton SA. Association of lupus-like anticoagulant and nonvasculitic
cerebral infarction. Arch Neurol 1986; 43:851-852.

80. Shi W, Kriis SA, Chong BH, et al. Prevalence of lupus anticoagulant and anticardiolipin
antibodies in a healthy population. Aust NZ J Med 1990; 20:231-236.

81. Duhrsen U, Brittinger G Lupus anticoagulant associated syndrome in benign and malignant
systemic disease. Klin Wochenschr 1987; 65:818-822.

82. Aim SS, Kalunian K, Rosove M, et al. Postoperative thrombotic complications in patients
with lupus anticoagulant: Increased risk after vascular procedure. J Vase Surg 1988; 7:749-756.

83. Greenfield LJ. Lupus-like anticoagulants and thrombosis. J Vase Surg 1988; 7:818-819.

84. Tsakiris DA, Settas L, Makris PE, et al. Lupus anticoagulant-antiphospholipid antibodies and
thrombophilia: Relation to protein-C and protein-S thrombomodulin. J Rheumatol 1990; 17:
785-789.

85. Svensson PJ, Dahlbach B. Resistance to activated protein C as a basis for venous thrombosis.
N Engl J Med 1994; 330:517-521.

86. Bertino RM, Koelemon BP, Kosta, et al. Mutation in blood coagulation factor V associated
with resistance to activated protein C. Nature 1994; 369:64.

87. Zeller B, Svensson PJ, Xuhua H, Dahlbach B. Identification of the same factor V gene mutation
in 47 out of 50 thrombosis-prone families with inherited resistance to activated protein C. J Clin
Invest 1994; 94:2521.

88. Rosenthaul F. Risk factors for venous thrombosis: Prevalence, risk and interaction. Semin
Hematol 1997; 34:171-187.

89. Bos G, Den Heijer M. Hyperhomocysteinemia and venous thrombosis. Semin Thromb Hemost

1998; 24:387-391. 1

90. Elchinger Stumpflen A, Hirschl M, et al. Hyperhomocysteinemia is a risk factor of recurrent <S
venous thromboembolism. Thromb Haemost 1998; 80:566-569. js

91. Eldrup-Jorgensen J, Flanigan DP, Brace 1, et al. Hypercoagulable states and lower limb °|
ischemia in young adults. J Vase Surg 1989; 9:334-341. ^

92. Donaldson MC, Weinberg DS, Belkin M, et al. Screening for hypercoagulable states in a >9
vascular surgery practice: A preliminary study. J Vase Surg 1990; 11:825-831. J

93. McDaniel MD, Pearce WH, Yao JST, et al. Sequential changes in coagulation and platelet q
function following femoro-tibial bypass. J Vase Surg 1984; 1:261-268. |

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

8

Complications and Failures of Anticoagulant
and Antithrombotic Therapy

John R. Hoch

University of Wisconsin Medical School, Madison, Wisconsin, U.S.A.

The development of effective anticoagulation strategies was critical to the establishment of
vascular surgery as a specialty. Today, we are fortunate to have established anticoagulant
agents such as heparin and warfarin, which have proven effective for the prevention and
treatment of most thromboembolic disorders. In addition, there are numerous exciting
new pharmacological options currently undergoing clinical trials; these may ultimately
improve on the effectiveness of current agents and their safety profiles. Vascular surgeons
must be knowledgeable about the pharmacology of each agent, its indications for use, and
how the effect of the agent is monitored. Most failures in anticoagulant therapy arise from
improper choice of agent or administration of insufficient or excessive amounts of the
anticoagulant. The major complication of anticoagulation therapy is hemorrhage; how-
ever, other, less common adverse effects may affect the survival of life and/or limb.

The surgeon's choice of anticoagulant or antithrombotic agent is dependent upon the
nature of the thrombus. Major classes of anticoagulants and antithrombotics include (a)
heparins, which induce inhibition of activated coagulation proteins by their interaction
with the natural anticoagulant antithrombin (AT, formerly known as antithrombin III);
(b) vitamin K antagonists, such as warfarin (Coumadin); (c) direct thrombin inhibitors,
such as lepirudin, bivalirudin, and argatroban; (d) factor Xa inhibitors, such as fondapa-
rinux; and (e) platelet function inhibitors, which include aspirin, the thienopyridines 1

(ticlopidine and clopidogrel), and the platelet glycoprotein Ilb/IIIa (GPIIb/IIIa) receptor §

inhibitors. Each class has a different mode of action and indication for use in the treatment j§

or the prophylaxis of thromboembolic disease. Jfj

I
I. UNFRACTIONATED HEPARIN J

Unfractionated heparin (UFH) is the anticoagulant of choice for the management of most °

acute thromboembolic disorders because of its rapid anticoagulant effect when adminis- J

©

179 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

180 HOCH

tered intravenously. UFH is an effective prophylaxis for patients at high risk for deep
venous thrombosis (DVT) as well as for the treatment of venous thrombosis and
pulmonary embolism (PE). Intravenous UFH is indicated for patients undergoing cardiac
surgery and vascular surgery as well as coronary and peripheral artery angioplasty. Low-
molecular-weight heparins (LMWHs), which were introduced over 25 years ago, have
proven to be safe and offer similar or superior efficacy compared to UFH in the manage-
ment and prophylaxis of thromboembolism. LMWHs offer the advantages of ease of ad-
ministration, improved bioavailability, and predictable anticoagulation.

A. Pharmacology

UFH is a mixture of straight-chain glycosaminoglycan sulfate esters with molecular weights
ranging from 3000 to 30,000 Da. UFH has a mean molecular weight of 15,000 Da. and
contains 10-90 saccharides per molecule (1). Beef lung and porcine intestinal mucosa are the
traditional sources for commercial UFH. They have different ratios of high- and low-
molecular-weight fractions, with differing anticoagulant activities. Because of this variation
in activity, heparin is dispensed in international units (IU) rather than by weight.

Heparins anticoagulant effect is dependent on a plasma cofactor, the proteinase inhib-
itor antithrombin (AT). Heparin binds to and potentates the activity of AT via a specific
glucosamine unit contained within a pentasaccharide sequence (1,2). Only one-third of
UFH molecules contain this specific pentasaccharide sequence. The remaining two-thirds of
heparin molecules in UFH preparations contain minimal anticoagulant activity (3). At
higher than usual clinical doses of heparin, both fractions of heparin with high and low
affinity for antithrombin potentiate the antithrombin effects of a second serine protease
inhibitor, heparin cofactor II (4).

The major mechanism responsible for the anticoagulant effect of UFH is the action of
the heparin-AT complex, which inactivates activated coagulation proteins thrombin (factor
Ha), factor IXa, factor Xa, factor XIa, and factor Xlla. Thrombin and factor Xa are most
sensitive to the effects of the heparin-AT complex. Heparin combines with AT in a 1:1
stoichiometric ratio and produces a conformational change in AT that fully activates its
serine protease-inhibitory site. Heparin molecules with molecular weights greater than
5000 Da, composed of sequences at least 18 saccharides long, form a ternary complex
composed of heparin, AT, and an activated coagulation factor (thrombin, factor IXa or
factor XIa) (5). Heparin molecules with molecular weights less than 5,000 Da, composed of
sequences less than 18 saccharides long, cannot bind to thrombin (1,2,5). These small
fragments, containing a high-affinity pentasaccharide sequence, form a binary complex with
AT capable of inactivating factor Xa and factor Xlla (2,5). Heparin may disassociate from
the complex and catalyze other thrombin-AT interactions once the AT has neutralized an
activated coagulation factor (6). The success of low-dose heparin for venous thromboem- ■g

bolic prophylaxis has been attributed to the inhibition of thrombin production by the &

heparin-AT complex inhibition of factor Xa. Just 1 jig of the binary heparin-AT complex is a

able to inhibit 32 U of factor Xa, which is equivalent to the inhibition to 1600 U of thrombin. c

In contrast, 1000 \xg of AT would be required to directly inactive a similar amount of <j

thrombin without the presence of heparin (6). >3

UFH has also been shown to induce secretion of tissue factor pathway inhibitor by 4j

endothelial cells (7). Tissue factor pathway inhibitor (TFPI) inhibits factor Vila-tissue Q

factor (TF)-induced factor X activation, leading to decreased thrombin generation (7). |

S

©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 181

Decreased thrombin production by this pathway and by thrombin inactivation by the
heparin-AT complex leads to the inhibition of thrombin-induced activation of factor V
and factor VIII, potentiating the antithrombotic effect of heparin (8).

Heparin has been shown to bind to platelets and impede their function, causing a
prolongation of bleeding time in humans (9). The higher-molecular-weight heparin frac-
tions that have a lower affinity for AT seem to be primarily responsible for heparin's effect
on platelet function (10). Heparin's antihemostatic effects may be secondary to its inter-
action with platelets as well as its anticoagulation effects.

Intravenous infusion of heparin is necessary to achieve an immediate anticoagulant
effect, as its availability is decreased by the subcutaneous route (11). After intravenous in-
jection, heparin rapidly binds to the endothelium, macrophages, and UFH-binding plasma
proteins (12). Heparin's affinity for plasma proteins is partly responsible for the variability
of the anticoagulant response to UFH. The mechanism of heparin clearance is complex. At
low and therapeutic doses of heparin, clearance is achieved by binding to receptors on
endothelial cells and macrophages where depolymerization occurs. This rapid, saturable
mechanism of clearance is responsible for the nonlinear anticoagulant response seen with
UFH in the therapeutic range. A second slower, unsaturable mechanism of UFH clearance
is kidney-dependent. This mechanism of heparin clearance primarily acts at very high
plasma concentrations of heparin. The average half-life of circulating heparin in the thera-
peutic range is approximately 90 min. Renal failure does not affect the anticoagulant half-
life or clearance of UFH when it is administered in a therapeutic range. Hepatic insufficiency
has no effect on heparin clearance but may affect anticoagulant activity by reducing the
availability of AT and other clotting factors.

Several methods are available to monitor the anticoagulant effect of heparin. The acti-
vated partial thromboplastin tine (APTT) is most widely used, with maintenance of the
APTT in the range of 1.5-2 times the control values associated with inhibition of intra-
vascular coagulation without excessive risk of hemorrhage (13). The ATT is sensitive to
the inhibitory effects of heparin on thrombin, factor Xa, and factor IXa. The APTT should
be measured 4-6 h after an initial bolus dose of heparin, the hourly intravenous dose
of UFH being adjusted accordingly. The activated clotting time (ACT) may also be used
and has the advantage of a linear response to increasing doses of heparin (21). Main-
tenance of the ACT in the range of 150-200 s provides adequate levels of anticoagula-
tion in the treatment of DVT or acute pulmonary embolism (13); ACT in the range of
200-250 s is recommended for arterial surgery and/or peripheral or coronary endovas-
cular intervention (13).

B. Clinical Application

Multiple studies have documented the effectiveness of low-dose heparin (5000 U subcuta-
neously 2 h before surgery and every 8-12 h postoperatively until the patient is ambu-
latory) in reducing the incidence of postoperative DVT and fatal pulmonary embolism. In

a review of randomized trials involving more than 15,000 surgical patients, perioperative c

prophylaxis with UFH was shown to significantly reduce the incidence of DVT by ap- <j

proximately 60%, the incidence of pulmonary embolism (PE) by 47%, and total mortality >9

in the series by 21% compared with control patients (14). 4j

For patients with DVT and PE, the goal of anticoagulation is to prevent clot pro- 2

pagation and new clot formation. Today, LMWH is the anticoagulant of choice for the |

S
©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

182 HOCH

majority of patients with confirmed DVT and PE (15). Extensive iliofemoral DVT is our
primary indication for using intravenous UFH in the management of patients with venous
thromboembolism. Although multiple clinical trials comparing UFH with LMWH in the
treatment of venous thromboembolism have shown similar efficacy and safety, patients
with extensive iliofemoral DVT were usually excluded from these trials because of the
large thrombus burden and high levels of activated clotting factors. We recommend a
large heparin bolus of 200 U/kg of body weight. The heparin bolus is followed by con-
tinuous intravenous infusion of heparin that is adjusted frequently to maintain the APTT
at twice the control value. Continuous intravenous heparin infusion is maintained at
therapeutic levels for 7-10 days in this subgroup of patients. Oral anticoagulation with
warfarin should be started on the first day of unfractionated heparin treatment. Heparin
infusion should be continued until the prothrombin time international normalized ratio
(INR) has been prolonged into the therapeutic range for at least 4 days (16). UFH can be
easily reversed with protamine and has a shorter plasma half-life than LMWH, making it
the anticoagulant of choice for patients preparing to undergo surgical or endovascular
procedures who require pre- and postoperative anticoagulation.

Intravenous UFH is used by vascular surgeons on a daily basis during arterial recon-
structive surgery as well as in performing percutaneous endovascular procedures. Our
bias at this time is to give arterial reconstructive patients an intravenous bolus of 100 IU/
kg of UFH immediately prior to arterial clamping or balloon angioplasty. Adequacy of
anticoagulation is assessed by hourly ACTs. If the ACT falls below 200 s, an additional
1000-U bolus of heparin is administered.

Heparin is used in the management of acute arterial thromboembolic disease to
prevent thrombus propagation and to permit collaterals to develop. Whether maintenance
of intravenous UFH is necessary following successful thrombolectomy is controversial
(17). Our practice is to maintain heparin infusion for 3-4 days following surgery and then
to convert to warfarin anticoagulation for those patients identified to have mural throm-
bus on echocardiogram or atrial fibrillation.

C. Complications

The most common complication of heparin therapy is hemorrhage, which may vary
from mild mucosal oozing or hematuria to extensive intracranial, gastrointestinal, retro-
peritoneal, or urinary bleeding. The risk of hemorrhage from prophylactic low-dose
subcutaneous heparin therapy is small. The incidence of wound hematomas in patients
placed on prophylactic heparin therapy postoperatively is less than 15% if 15,000 IU or less
of heparin is administered daily (18). The incidence of hemorrhage requiring transfusion
in these postoperative patients is less than 4% (18). There is a 6-10% risk of bleeding
complications when anticoagulation is in the therapeutic range (19). The incidence of
hemorrhage, however, may approach 50% in patients with renal failure, underlying hemo- ■g

static defects, or thrombocytopenia (19). Patients with severe hypertension, ongoing bleed- &

ing, recent neurosurgical operation, or those undergoing percutaneous endovascular a

procedures are at heightened risk. With close monitoring of the APTT (range of 1.5-2 -c

times control), patients without these risk factors have a small risk of major bleeding <j

complications. >3

Hemorrhage that is not life-threatening is best managed by discontinuation of hepa- 4j

rin therapy. If bleeding continues, heparin may be neutralized with protamine sulfate. The 2

amount of protamine sulfate required can be calculated from results of a protamine |

titration test or the ACT. If the tests cannot be done, 1-1.5 mg of protamine is usually ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 183

required to neutralize 100 U of heparin. Portions of the calculated dose, 30-50%, are
given slowly intravenously to reduce the risk of hypotension, bradycardia, and peripheral
vasodilatation. Excessive administration of protamine should be avoided, as protamine
may also act as an anticoagulant via its interaction with platelets and serum proteins (19).

A second potentially life- or limb-threatening complication of heparin is the develop-
ment of heparin-induced thrombocytopenia (HIT) syndrome. Approximately 1-5% of
patients receiving heparin develop heparin-associated antiplatelet antibodies (HAAbs)
(21,22). HIT antibodies bind to the antigen composed of a complex of heparin and platelet
factor 4 (PF4). The heparin-PF4-IgG immune complexes bind to platelet Fella receptors
leading to platelet aggregation. Thrombin is generated, possibly as a consequence of
antibody damage leading to expression of tissue factor (12,21). The result is the potential
for devastating arterial and venous thrombotic complications.

The development of HIT occurs after 4-5 days of UFH therapy in patients exposed to
heparin for the first time; thrombocytopenia may develop as early as 24 h in patients who
have been reexposed to heparin who have had recent exposure to heparin and have
circulating HAAbs (23). Patients characteristically exhibit a drop in platelet count greater
than 50% from their baseline, platelets often falling to less than 50,000/mm 3 (23).
Importantly however, 10-15% of patients will have a platelet counts that remain above
1 50,000/mm . In addition to monitoring the platelet count, patients who demonstrate
increasing resistance to anticoagulation with UFH, or new or progressive hemorrhagic or
thrombotic complications while on heparin should be screened for HAAbs.

HIT has been associated with a 23% mortality and a 61% morbidity (24). However,
with early recognition and treatment, mortality and morbidity can be reduced to 12 and
22.5% respectively (22). Once patients are suspected of HIT, all heparin therapy must be
discontinued; this includes even "minor" sources of heparin, including intravenous flushes
and heparin-coated catheters. Patients should be administered an appropriate platelet
function inhibiting agent (aspirin or clopidogrel) pending diagnosis of HIT. The diagnosis
of HIT should be confirmed by either a functional test or an antigen assay. Available
functional tests include the serotonin release assay as well as heparin-induced platelet
aggregation assays (25,26). The heparin-PF4 enzyme linked immunosorbent assay is an
antigen assay in which patient's serum is tested for the presence of antibodies to heparin-
PF4. This assay results in a very low number of false-negative results; however, its
specificity is lower than that of the functional assays (27,28).

Management of patients with ongoing anticoagulation requirements who have HAAbs
is difficult. In centers that have access to heparin-induced platelet aggregation assays, one
can test LMWHs against the patient's antibodies to look for cross-reactivity. Cross-
reactivity rate with LMWH varies from 20-61% depending on which LMWH is tested
(28,29). If all LMWHs cross-react with the patient's antibodies or if a platelet aggregation
assay is not available at your hospital, then the recommended management of HIT in ■g

patients who require anticoagulation is with intravenous direct thrombin inhibitors (30). &

Sensitivity reactions consisting of bronchiole constriction, lack of lacrimation, or a

urticaria may occur in 2-5% of the patients receiving heparin. Anaphylaxis with -c

circulatory collapse is a rare complication of heparin administration. <j

Long-term administration of heparin may be associated with alopecia and osteopo- >9

rosis. The incidence of alopecia is higher in patients who are also receiving warfarin. 4j

Hair growth usually resumes once heparin administration is discontinued. Heparin is 2

known to suppress osteoclast formation and to activate osteoblasts that may promote |

bone loss (12). Osteoporosis and pathological fractures of the vertebral column and ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

184 HOCH

long bones have been reported in patients receiving heparin for more than 6 months in
dosages exceeding 10,000 U/day (31).

D. Failures

Most failures of heparin therapy are iatrogenic and are related to (a) the use of heparin in
patients with contraindications, (b) the administration of insufficient amounts of heparin,
(c) not beginning the prophylactic heparin regimen early enough, and (d) and on rare
occasions the development of sensitivities to heparin or a congenital or acquired AT
deficiency. Prophylactic regimens consisting of low-dose UFH must be initiated either
before or concomitantly with the event placing the patient at risk for thromboembolic
complications. Low-dose subcutaneous UFH has shown equal efficacy as a prophylactic
agent for the prevention of venous thromboembolism even in high-risk general surgery
patients compared to LMWHs (32). However, low-dose UFH is not recommended as a
venous prophylactic agent in patients undergoing major orthopedic surgery, those with
acute spinal cord injury, or in patients with multiple trauma (32). LMWH is the
prophylactic agent of choice for these patients unless heparin is contraindicated (32).

Patients who fail to achieve therapeutic levels of anticoagulation with increasing
amounts of heparin or require a higher than average doses of heparin to prolong the APTT
into the therapeutic range are designated "heparin resistant." Surgeons should become
concerned if the daily dose of intravenous UFH exceeds 35,000 U per 24/h (33). Possible
causes of heparin resistance include the presence of AT deficiency, the development of
HIT, and elevations of factor VIII, fibrinogen, and PF4 (57-62). Factor VIII and fibrinogen
levels have been shown to be elevated during many acute illnesses and during pregnancy.
Increased levels of factor VIII act to limit the response of the APTT to heparin; however;
the in vivo antithrombotic effect is not diminished (33). Thus patients proven to have ele-
vated factor VIII levels should have their heparin anticoagulation monitored by mea-
suring anti-Xa activity (33). It is recommended that for patients who require greater than
35,000 U of UFH per 24 h, the dose should be adjusted to maintain the anti-Xa levels of
0.35-0.70 IU/mL (12,33).

Heparin has limited anticoagulant activity in patients with congenital or acquired AT
deficiencies. AT concentrations may decrease during heparin therapy, with up to a 12%
decrease 4 h after initiation of therapy and a 33% decrease with continued therapy. Pa-
tients with preexisting AT deficiency may experience even greater decreases in circulating
in AT concentrations with prolonged heparin administration (34). Fresh frozen plasma or
cryoprecipitate can be utilized to replenish deficient levels of AT and allow continued use
of UFH. Alternatively, intravenous direct thrombin inhibitors can be utilized if continued
anticoagulation is necessary in patients with heparin resistance.

II. LOW-MOLECULAR-WEIGHTS HEPARINS S

LMWHs reached the clinical arena in the late 1970s with the promise of several clinical -c

advantages over UFH, including superior bioavailability when administered subcutane- <j

ously, equivalent or lower incidence of bleeding complications, lower risk of HIT; pre- >9

dictable dose response, and fixed-dose administration without monitoring (35). LMWH M

has clearly shifted the paradigm for care of patients with venous thromboembolism. Its Q

rapid adoption by clinicians is primarily because of its pharmacokinetic advantages |

over UFH. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY

185

A. Pharmacology

Commercial LMWHs are prepared by acidic hydrolysis, esterification, enzymatic depolym-
erization, or fractionation of UFH (36). LMWHs have an average molecular weight of
4500-6000 Da. Over half of the saccharide units in LMW preparations are less than 18 U
long. These smaller fragments contain the high affinity pentasaccharide sequence and are
able to catalyze the inhibition of factor Xa by AT, but they cannot bind AT and thrombin
(1,5). LMWHs have a relatively higher anti-factor Xa and lower anti-factor Ha activity
compared to UFH. Commercial LMWH differ in their Xa:IIa affinity ratios. Like UFH,
LMWH can induce the secretion of tissue factor pathway inhibitor by vascular endothelial
cells, whose action reduces the procoagulant activity of the TF-factor Vila complex (37).

LMWHs have clear pharmacokinetic advantages over UFH preparations. LMWHs
have superior bioavailability when administered subcutaneously and a longer half-life
compared to UFH (Table 1). LMWH also has a more predictable dose response allowing
clinicians to administer weight adjusted doses of LMWH without laboratory monitoring
(39). The greater bioavailability and more predictable dose response of LMWH is
attributed to their decreased affinity in binding to plasma proteins, macrophages, and
endothelial cells (40). Compared with UFH, LMWHs have reduced binding to platelets
and PF4 which explains the lower incidence of HIT associated with LMWHs (41).

LMWH is usually administered as a fixed dose without monitoring. Patients with
renal insufficiency, and who are morbidly obese or pregnant, however, require factor Xa
monitoring. The most commonly employed laboratory assay for assuring adequacy of
anticoagulation with LMWHs is the chromagenic anti-Xa assay (42). When LMWH is
administered once daily, an anti-Xa assay should be performed 4 h after the initial dose
with a therapeutic range between 1.0 and 2.0 IU/mL (42). When LMWH is administered
twice daily, an anti-Xa level ranging from 0.6 to 1.0 IU/mL is recommended (42).

LMWHs are cleared by the kidneys in a dose-independent fashion (12). Because of the
renal clearance of LMWHs, it is recommended that anti-Xa assays be utilized in patients

Table 1 A Comparison of the Pharmacokinetic Differences Between Low-Molecular-Weight
Heparins (LMWHs) and Unfractionated Heparin (UFH)

Characteristic

UFH

LMWH

Molecular weight
Plasma half-life
Anti-Xa:anti-IIa activity
Platelet inhibition
Reversal of anticoagulation
Clearance

Administered
Laboratory monitoring

3,000-30,000/Da

1-2 h

1:1

Yes

Protamine

Endothelial cell and

Macrophage binding
Renal (at high doses)
Intravenous and

subcutaneous
APTT or ACT

4,000-6,000/Da

4-6 h

2:1^1:1

Less than UFH

Protamine less effective

Renal (dose-independent)

Subcutaneous

- Fixed dose

- Serum anti-factor Xa a

i
2

I

■c

a Serum anti-factor Xa level monitoring recommended for patients with renal insufficiency, preg-
nancy, and obesity.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

186

HOCH

with renal insufficiency. This is prudent, as many of the clinical trials that led to approval
by the U.S. Food and Drug Administration (FDA) of the LMWHs excluded patients with
renal failure.

B. Clinical Application

In the United States, the LMWHs dalteparin and enoxaparin have been approved for use
as prophylactic agents for the prevention of venous thromboembolism in patients under-
going abdominal surgery and orthopedic surgery (Table 2). The ability of LMWH and
low-dose UFH to prevent DVT in general surgery patients has been studied in numerous
large trials and analyzed by metanalysis (43). LMWH proved to be efficacious at pre-
venting DVT in general surgery patients (43,44). Metanalyses revealed that there is a clear
dose-response effect of LMWH on bleeding complications (44). There is more bleeding
with LMWH if doses of greater than 3400 IU of anti-Xa are given daily, in comparison to
low-dose UFH, in a dose of 5000 IU bid or tid. Equivalent bleeding risk is evident if
LMWH is administered at less than 3400 U of anti-Xa daily. Cost-effectiveness analyses
have been performed for patients undergoing abdominal surgery comparing the cost of
LMWHs with low-dose UFH. The authors concluded that in North America prophylaxis
with lose-dose UFH was equally efficacious and more economical than LMWH (43).

Subcutaneous LMWH is rapidly replacing intravenous UFH for the initial treatment
of patients with uncomplicated venous thromboembolism. Metanalyses have found that
unmonitored, weight-adjusted, subcutaneous LMWH is as least as effective as intravenous
UFH in preventing recurrent venous thromboembolism (45,46). Early metanalyses, pub-
lished soon after FDA approval, concluded that the treatment of venous thrombo-
embolism by LMWH was safer and more effective than treatment by UFH (47). Recent

Table 2 FDA-Approved Indications for LMWH Therapy

LMWH

Indication

Dose (Subcutaneous)

Dalteparin DVT prophylaxis

Abdominal surgery
Higher-risk abdominal

surgery; Hip replacement
Treatment of unstable angina/
non-Q-wave MI
Enoxaparin DVT prophylaxis

Abdominal surgery

Hip and knee replacement

Outpatient DVT treatment,

without PE
Inpatient DVT treatment,

with or without PE
Tinzaparin Treatment of DVT, with or

without PE

2500 IU anti-factor Xa q 24 h
5000 IU anti-factor Xa q 24 h

120 IU/kg anti-factor Xa q 24 h

40 mg a q 24 h

40 mg q 24 h; initiate 12(±3) h

preop or 30 mg q 12 h
1 mg/kg q 12 h

1.5 mg/kg q 24 h or 1 mg /kg q 12 h

175 IU anti-factor Xa q 24 h

i
2

I

■c

Abbreviations: LMWH, low-molecular-weight heparin; MI, myocardial infarction; DVT, deep venous

thrombosis; PE, pulmonary embolism.

a Enoxaparin I mg =100 anti-factor Xa IU.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 187

metanalyses however, demonstrate a decrease in overall mortality with LMWH but only a
trend toward superiority of LMWH over UFH in preventing recurrent thromboembo-
lism. Gould et al. analyzed 1 1 randomized, controlled trials designed to compare the safety
and efficacy of LMWH compared to UFH for the treatment of acute DVT (45). Throm-
boembolic events occurred in 5.4% of all patients treated with UFH, compared to 4.6% of
patients treated with LMWH. Recurrent thromboembolic events were less common in
patients who received LMWH, but the difference was not statistically significant (absolute
risk reduction of 0.88%). Importantly, LMWHs reduced mortality rates over 3-6 months of
follow-up, compared with the UFH patients (29% relative risk reduction). Dolovich et al.
utilized metanalysis to examine 13 studies in which patients were randomized to receive
either LMWHs or intravenous UFH for the treatment of acute venous thromboembolism
(46). The incidence of recurrent venous thrombosis or pulmonary embolism was similar
between UFH and LMWHs groups, but patients treated with LMWH experienced a 24%
reduction in the risk of total mortality compared to patients treated with intravenous UFH.
Once-daily therapy with LMWH was found to be as safe and effective as twice-daily dosing
of LMWH.

While management of DVT with LMWHs has become the standard of care across the
United States, clinicians have been slower to adopt LMWH therapy for the management
of patients with PE. In the early trials demonstrating efficacy of LMWH for the manage-
ment of venous thrombosis, patients with PE were either excluded or represented in very
small numbers. Recently, LMWH has been evaluated in the treatment of submassive
pulmonary embolism in randomized trials (48,49). Simonneau et al. reported the results of
a randomized study of 612 patients with symptomatic pulmonary embolism who received
either subcutaneous LMWH (once-daily in a fixed dose), or adjusted dose, or intravenous
UFH followed by at least 3 months of oral anticoagulation therapy (49). The investigators
examined the outcomes of recurrent thromboembolism and death at 8 and 90 days from
the initiation of therapy and found no differences between groups. Hull et al. conducted a
double-blind, randomized trial comparing LMWH with intravenous heparin treatment in
patients with documented proximal DVT who were found on subsequent perfusion lung
scan to have a high probability of pulmonary embolism (48). No patient in the LMWH
group experienced recurrent venous thromboembolism, compared to 6.8% of patients in
the UFH group (p = 0.009). The authors concluded that once-daily subcutaneous LMWH
treatment was no less effective and probably more effective than use of dose-adjusted intra-
venous UFH for preventing recurrent venous thromboembolism in patients with nonhemo-
dynamically compromising PE and associated proximal DVT. There are no prospective
data demonstrating the efficacy and safety of LMWH treatment for patients who are hemo-
dynamically unstable with massive PE. Therefore it is our practice to manage these patients
in the hospital with very closely monitored intravenous UFH.

Although not proven, many obstetricians have a strong sense that pregnant women ■g

are at higher risk of developing venous thromboembolism than nonpregnant women. &

Women with thrombophilic disorders such as deficiencies of antithrombin, protein C, or a

protein S have an approximately eightfold increased risk of venous thromboembolism c

during pregnancy (50). Sixty percent of one series of women who developed venous throm- <j

boembolism during pregnancy tested positive for factor V Leiden (51). Therefore, with >9

patients with known thrombophilia, consideration should given for the use of prophylactic 41

LMWH, plus postpartum anticoagulation (51). For patients with multiple prior episodes 2

of venous thromboembolism or women already receiving long-term anticoagulation prior |

to pregnancy, adjusted-dose LMWH is recommended, followed by resumption of long- @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

188 HOCH

term oral anticoagulation therapy postpartum. For patients who develop venous throm-
boembolism during pregnancy, adjusted-dose LMWH is recommended. Anticoagulation
with warfarin is contraindicated, as warfarin crosses the placenta. There is strong evidence
that LMWHs do not cross the placenta and are safe for the fetus (51). The pharmaco-
konetics of LMWH change in pregnancy, resulting in a shorter plasma half-life and larger
volume distribution. Therefore, monitoring of anti-factor Xa levels is necessary (50,51).

Clinicians are now beginning to investigate the potential role for LMWH in patients
undergoing arterial revascularization and/or percutaneous transluminal angioplasty. His-
torically, on an anecdotal basis, vascular surgeons have used intravenous UFH intra-
operatively and have utilized postoperative intravenous UFH in patients who have
undergone difficult distal reconstructive procedures. Edmondson et al. compared the effect
of postoperative daily injection of LMWH versus a combination of aspirin and dipyrida-
mole every 8 h for 3 months following femoral-to-popliteal bypass grafting (52). The ran-
domization occurred approximately 1 week postoperatively with all patients receiving
LMWH during the first week after surgery. At 12 months, graft patency was 78% in the
LMWH group, compared to 64% in the aspirin-with-dipyridamole group. Significant im-
provement in graft patency was seen in the LMWH subgroup of patients undergoing
limb salvage surgery, while no benefit was appreciated in patients presenting with claudi-
cation. Simama et al. enrolled 201 consecutive patients scheduled for femorodistal recon-
structive surgery in an open randomized trial comparing intraoperative and postoperative
LMWH with UFH (53). Intraoperatively, the LMWH group received enoxaparin, 75 IU/kg
anti-Xa intravenously, while the UFH group received 50 IU/kg, also intravenously. Post-
operatively, patients received subcutaneous administration of enoxaparin ( 75 IU/kg anti-
Xa) or UFH (150 IU/kg) beginning 8 h after the initial intravenous dose and then every
12 h for 10 days. Graft thrombosis occurred by 10 days in 8% of patients in the LMWH
group, compared with 22% of patients in the UFH group (p — 0.009). Major hemorrhages
occurred in 12% of patients in each group and there was no difference in mortality rates
between groups. Hingorani et al. retrospectively identified 169 patients who postoperatively
received intravenous UFH and 161 patients who receive enoxaparin 1 mg/kg every 12 h as a
bridge to adjusted dose warfarin therapy (54). Both groups of patients received intrave-
nous UFH for 24 h after surgery. There was no standardization of the length of post-
operative LMWH prior to conversion to warfarin. The authors found no difference in the
incidence of postoperative complications except for an increased incidence of return to
surgery for graft thrombosis, failing grafts, and debridement in patients who received UFH.
The authors caution the reader not to draw the conclusion that these retrospective data
suggest a decreased incidence of graft thrombosis with LMWH (54).

Plaque rupture leads to tissue factor expression and subsequent activation in a coagu-
lation cascade and generation of factor Xa (55). As LMWHs target factor Xa to a far greater
extent than thrombin, their use has been extensively investigated for the management of •§

patients with unstable angina/non-Q-wave myocardial infarction (MI). Metanalyses have g

shown the superiority of LMWH over placebo in the setting of unstable angina and recently a

LMWHs have been found to be superior to UFH in this patient population (56). Several c

clinicians have investigated the role of LMWH as the sole anticoagulant during percuta- <j

neous coronary intervention (PCI) (57). Further prospective trials will be necessary to >9

determine the optimal level of anticoagulation; however, similar safety and efficacy out- 4j

comes with LMWH compared with UFH have been shown with target anti-Xa levels higher 2

than 0.5 IU/mL (57). 1

S

©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 189

The use of LMWHs for the management of patients undergoing peripheral percuta-
neous transluminal angioplasty is only now being investigated (58). Schweizer et al.
randomized 172 patients who had undergone iliac or superficial femoral artery angioplasty
with subsequent extensive dissections to receive either UFH or LMWH for a 7-day period
after angioplasty, followed by a 6 month course of aspirin (58). For this group, no signifi-
cant treatment-related differences in the degree of restenosis were found at 3 weeks, and
3 and 6 months postprocedurally. However, when angioplasty was performed in the
superficial femoral artery, the degree of restenosis was significantly lower in the LMWH
group compared to the UFH group at all three time points. Further prospective evaluation
of the use of LMWH in patients undergoing peripheral angioplasty is warranted.

C. Complications and Failures

Similar to UFH, the most common complication of LMWH is hemorrhage. When
LMWH is used for prophylaxis of venous thrombosis, risk of bleeding is small and com-
parable to low-dose heparin (59). Recent metanalyses also demonstrate comparable risk
between LMWHs and UFH for both major and minor bleeding in patients being treated
for venous thrombosis and pulmonary embolism (45,46,48). Overdosing patients with
LMWH is problematic because protamine is less effective at neutralizing the antithrombin
activity of LMWH compared to its effect on UFH (60). Protamine does not neutralize all
the Xa-inhibiting activity of LMWH even at protamine/heparin ratios of more than five.
Resistance of a specific LMWH to protamine neutralization is a function of not only the
LMWHs molecular size but also their degree of sulfonation (60).

In 1997, the FDA issued a public health advisory in order to alert physicians of the
increase risk of spinal and epidural hematoma associate with the use of LMWHs (61). At
that time, 43 patients in the United States had developed perispinal hematoma, with over
half the patients suffering significant neurological impairment. The 6th ACCP Consensus
Conference on Antithrombotic Therapy made several recommendations in an attempt to
improve the safety of neuroaxial in patients receiving LMWH (32). Some of the recom-
mendations include the following: (a) avoid LMWH if patients have known bleeding dis-
order, (b) delay spinal needle insertion 8-12 h after LMWH, (c) avoid DVT prophylaxis
with LMWH if there is a "bloody tap," (d) remove epidural catheter when anticoagulant
effect is at minimum, and (e) delay LMWH prophylaxis for 2 h after needle withdrawal (32).

LMWHs is less likely than UFH to cause HIT antibody formation (41). HIT antibodies
bind to the antigen composed of a complex of heparin and PF4. Because 12-14 saccharide
units are necessary to form the antigenic complex with PF4, LMW molecules greater than
4000 Da can cause HIT. LMWHs are generally not a treatment option for patients devel-
oping HIT secondary to UFH, unless platelet aggregometry is available to test patient serum
against specific LMWHs. The management of HIT for patients receiving LMWH is iden- ■g

tical to the treatment algorithms developed for patients receiving UFH. &

Patients receiving either heparin or LMWH for prolonged periods are at risk for a

developing heparin-induced osteoporosis. The risk of heparin-induced osteoporosis is re- c

lated to the length of exposure. The indications for the use of LMWH have been expanding <j

and long-term use of LMWH is indicated for patients with recurrent venous thromboemb- >3

olism while they are adequately anticoagulated with oral anticoagulant therapy (62,63). 41

There is clinical evidence that LMWHs carry a lower risk of osteoporosis than UFH as well 2

as rat data suggesting that LMWH causes less osteopenia than UFH (64). |

S

©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

190 HOCH

As with UFH, failures of LMWH therapy are usually iatrogenic. Clinicians need to be
aware that fixed-dose administration of LMWH may at times result in inadequate anti-
coagulation in selective patients. Patients, who receive LMWH during pregnancy require
periodic monitoring of anti-Xa levels because the volume of distribution for LMWH
changes as the pregnancy progresses. Options include weight-adjusted dosing or the weekly
performance of serum anti-factor Xa levels (51). For morbidly obese patients, the use
of weight-adjusted dosing is recommended. Patients with renal failure should have their
anti-Xa levels monitored because of the renal clearance of LMWH. In this group of pa-
tients, we recommend more frequent monitoring of anti-Xa levels than in the morbidly
obese patient.

III. VITAMIN K ANTAGONISTS
A. Pharmacology

Warfarin, first synthesized in 1944 at the University of Wisconsin, is the most popular oral
anticoagulant. Gastrointestinal absorption of warfarin is complete, with peak plasma
concentrations being 2-12 h after a single oral dose. Warfarin is bound (97%) to albumin
and has a circulating half-life of 36-40 h. Oral anticoagulants are principally metabolized
by the microsomal fraction of the hepatacyte; metabolites are excreted in the urine.

Warfarin interferes with the action of vitamin K in the synthesis of clotting factors II,
VII, IX, and X by the liver. Vitamin K is a cofactor in the reaction that converts glutamyl
residues of clotting factor precursors to the carboxyglutamyl residues necessary for the
binding of calcium. Patients receiving warfarin produce antigenically similar clotting fac-
tors that do not have procoagulant activity because of their abnormal calcium-binding
characteristics (65).

Anticoagulation with warfarin is dependent on the reduction of the concentrations of
all affected clotting factors and may take 3-5 days to achieve. The time required to reach
therapeutic levels of anticoagulation is affected by the rate of "turnover" of the clotting
factors, which is directly related to their circulating half-lives. Factor VII and IX
zymogens have half-lives of 6-24 h respectively, whereas prothrombin (II) has a half-life
of approximately 96 h (66). Early reduction of factor VII and IX zymogens results in an
anticoagulant effect reflected by prolongation of the prothrombin time (PT). This early
laboratory anticoagulant effect does not translate into an in vivo antithrombotic effect.
The antithrombotic effect of warfarin is tied to the reduction of prothrombin, which
usually occurs 4 or 5 days after the initiation of warfarin therapy. This is the basis for
overlapping heparin or LMWH with warfarin until the INR has been prolonged into the
therapeutic range for at least 4 days. Warfarin also inhibits carboxylation of the natural
anticoagulant proteins C and S. Protein C's half-life is only 4-6 h; thus those are reduced ■§

early and warfarin therapy, creating the potential for transient procoagulant state. Thus g

these issues — the early drop in protein C activity, and the delayed reduction in pro- a

thrombin levels — support the use of a maintainance dose of warfarin rather than a loading c

dose during the initiation in therapy (67). <j

The biological effects of warfarin may be potentiated by hepatic insufficiency, malnu- >9

trition, or hypermetabolic states such as fever or hyperthyroidism. Many common ther- 4j

apeutic agents may alter the response of the coagulation system to warfarin due to decreased 2

absorption from the gastrointestinal tract, displacement of the drug from its binding site on |

albumin, increases in the rate at which it is metabolized by the liver, decreased vitamin K @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY

191

availability, or decreases or increases in the plasma half-lives of the affected clotting factors
(Table 3).

Warfarin is known to cross the placental barrier and therefore should not be given
during pregnancy. During the first trimester, warfarin has a known teratrogenic effect.
LMWH is the anticoagulant of choice during pregnancy.

B. Clinical Application

Warfarin has been shown to be an effective prophylactic agent for the prevention of venous
thrombosis after hip surgery and major general surgery at a target INR of 2.0-3.0 (68).
Although the risk of major hemorrhage in this range of INR is low, the use of warfarin for
venous thrombosis prophylaxis is reserved for patients of very high risk because of its
complexity of administration.

Randomized trials have demonstrated that warfarin is effective at preventing recur-
rent venous thromboembolism (VTE) when administered after a course of either UFH or
LMWH for acute VTE (68). Warfarin should be started once adequate anticoagulation
is obtained with either UFH or LMWH; heparin should be continued for at least 4 days
after achieving an INR in the range of 2.0-3.0.

Warfarin therapy is indicated for at least 3 months for patients with proximal DVT or
pulmonary embolism if the risk factors leading to the VTE have been corrected. In patients
with so-called idiopathic proximal DVT in which no clear risk factors have been identified,
warfarin therapy is continued for 6 months. Lifelong warfarin therapy is indicated for
patients with recurrent idiopathic venous thromboembolism; antiphospholipid antibody
syndrome; deficiencies of protein C, protein S, or antithrombin; and in those with VTE
associated with the homozygous factor V Leiden gene type (63,69). Ridker et al. recently
reported the early termination of a randomized trial of patients with idiopathic VTE who
were assigned to either placebo or low-intensity warfarin (target INR, 1.5-2.0) after first
completing full-dose warfarin anticoagulation for a median of 6.5 months (69). The study's
endpoints were recurrent venous thromboembolism, major hemorrhage, and death. The
study was terminated early after 508 patients had been randomized and followed for a

Table 3 Drug Interactions Affecting Vitamin K Antagonist Activity

Potentiate

Inhibit

Acetaminophen

Acetylsalicylic acid

Alcohol

Amiodarone

Anabolic steroids

Ciprofloxacin

Clofibrate

Cotrimoxazole

Disulfiram

Erythromycin

Fluconazole

Itraconazole

Lovastatin

Metronidazole

Micronazole

Nalidixic acid

Norfloxacin

Ofloxacin

Omeprazole

Phenylbutazone

Phenytoln

Piroxicam

Propafenone

Propranolol

Quinidine

Sulfinpyrazone

Tamoxifen

Tetracycline

Barbiturates

Carbamazeprine

Cholestyramine

Chlordiazepoxide

Cyclosporine

Dicloxacillin

Griseofulvin

Nafcillin

Rifampin

Sucralfate

Trazodone

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

192 HOCH

mean of 2.1 years. Patients treated with low-intensity warfarin had a 64% reduction in the
risk of recurrent VTE compared to the placebo group (p< 0.001). Low-intensity warfarin
was associated with a 48% reduction in the composite endpoint of recurrent VTE, major
hemorrhage, or death. Thus there is a role for long-term low-intensity warfarin therapy.

Following heart valve replacement, patients benefit from lifelong anticoagulation to
decrease the risk of systemic embolization. The American College of Chest Physicians
1998 guidelines recommend an INR of 2.5-3.5 for most patients with mechanical pros-
thetic valves and 2.0-3.0 for those with bioprosthetic valves (68). Warfarin therapy has
also proven effective for decreasing the risk of stroke in patients with chronic atrial
fibrillation. A pooled review of five trials that addressed anticoagulant therapy for the
prevention of stroke in atrial fibrillation patients treated with warfarin showed a 69%
reduction in the risk of stroke (70). Patients who are able to continue taking their warfarin
on a daily basis benefited from an 80% stroke risk reduction.

Retrospective reviews have implied that patients undergoing peripheral arterial bypass
procedures may benefit from long-term warfarin therapy (71). The effect of warfarin plus
aspirin (WASA) versus aspirin alone (ASA) on peripheral artery bypass patency rates and
patient mortality and morbidity was investigated in a multicenter, prospective, random-
ized trial (72). Patients were randomized to receive aspirin (325 mg/day) versus WASA
(target INR 1.4-2.8). The data for patients undergoing vein bypass was analyzed sepa-
rately from those undergoing prosthetic bypass. Patency rates in patients undergoing vein
bypass were unaffected. In the prosthetic bypass group, there was no significant difference
in patency rate in patients receiving an 8-mm bypass; however, there was a significant
improvement in patency rates in patients in the WASA group who received 6 mm femoral-
popliteal bypasses compared to the ASA group (5-year assisted primary patency; 71.4%
WASA, 57.9% ASA group, p — 0.02). However, the mortality rate for all patients in the
WASA group was significantly higher (31.8%) than that for patients in the ASA group
(23.0%; p — 0.0001). Warfarin did not provide any greater benefit than ASA for the risk
of cerebral events, myocardial infarction, or thromboembolic events. Major hemorrhagic
complications were more common in the WASA group than the ASA group. The authors
conclude that low-dosage warfarin therapy may provide some additional patency benefit
for patients who undergo femoropopliteal prosthetic bypass, but at the added cost of an
increased risk of mortality and hemorrhagic events.

C. Complications

Hemorrhage is the most common complication of warfarin therapy. Frequency of hemor-
rhage with warfarin varies widely but is a function of the level of anticoagulation. Patients
treated with warfarin to prevent recurrent venous thromboembolism with a target INR of
2.0-3.0 achieve similar efficacy to patients whose target INR is adjusted to 3.0-4.5, but
with significantly less bleeding complications (total bleeding 4.3%, INR 2.0-3.0; 22.4%,
INR 3.0-4.5; p — 0.015) (73). When bleeding complications occur, warfarin should be

discontinued. The effects of warfarin may be reversed within 24 h by intravenous admin- -c

istration of 20 mg of vitamin K. Life-threatening hemorrhage is best managed with rapid <j

reversal using infusions of fresh frozen plasma with or without administration of vitamin >9

K. A prothrombin complex concentrate (PCC) has recently been shown to be more effective %

than vitamin K treatment in rapidly correcting increased INR levels in patients receiving 2

warfarin (74). The use of PCC without vitamin K may, however, result in a repeated increase |

of INR, leading investigators to recommend concomitant vitamin K use with PCC. ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 193

Other than hemorrhage, the adverse effects of warfarin are rare; they include alopecia,
dermatitis, fever, nausea, diarrhea, abdominal cramping, and hypersensitivity reactions. A
rare complication of warfarin therapy is extensive dermal gangrene, the risk of which is
increased if loading-dose regimens are used for initiating therapy. This occurs early after
warfarin administration when protein C levels are decreased while the intrinsic coagu-
lation pathway remains intact; patients with congenital or acquired protein C deficiency
are therefore at greater risk for this complication. The skin in the thigh, breast, and
buttocks is most often involved. Simultaneous heparin administration at the beginning of
warfarin treatment prevents this complication (75).

D. Failures

Failure with warfarin therapy — that is, the inability to inhibit coagulation — is most
frequently caused by the use of insufficient amounts of warfarin. Individuals have been
identified with an inherited resistance to warfarin, which causes them to require 5- to 20-
fold higher than average amounts of warfarin to achieve an acceptable INR. Because the
plasma warfarin level required to achieve anticoagulation is elevated, it is likely that there
is altered affinity of the receptor for warfarin (76). Variation in the dose response to
warfarin may also be secondary to a common mutation in the gene coding for one of the
common cytochrome P450 enzymes (2C9) responsible for the oxydative metabolism of the
warfarin S isomer (77). As previously described, drugs and dietary factors can interfere
with the effectiveness of warfarin therapy. Increased consumption of foods high in vitamin
K or vitamin K containing supplements has been shown to reduce the anticoagulant
response to warfarin.

Antiphospholipid antibody syndrome is an acquired autoimmune disorder associated
with both venous and arterial thromboses (50,51,63). While the precise mechanism leading
to the hypercoagulable state remains unclear, patients are noted to have elevated levels of
lupus anticoagulants and anticardiolipin antibodies (50,51,63). Intravenous UFH is not a
good choice for management of these patients, as lupus anticoagulants cause prolongation
of the APTT. There is growing evidence that many patients with antiphospholipid
thrombosis syndrome will develop recurrent thromboembolic episodes even in the face
of therapeutic levels of anticoagulation with warfarin. Current recommendations are to
manage patients with LMWH concomitantly with a platelet function inhibiting drug such
as aspirin or clopidogrel (63).

IV. DIRECT THROMBIN INHIBITORS

Direct thrombin inhibitors are molecules that interact with thrombin and act to block its

interaction with substrates. This is in contrast to the indirect thrombin inhibitors UFH ■g

and LMWH, which require the cofactor antithrombin in order to inhibit thrombin's &

activity. The development of direct thrombin inhibitors parallels the increasing recogni- a

tion by clinicians of the heparin-induced thrombocytopenia syndrome. The FDA has ap- c

proved two of the commercially available direct thrombin inhibitors as the anticoagulants <j

of choice for patients with HIT who require continued anticoagulation. A second reason >9

for the development of direct thrombin inhibitors pertains to the inability of the heparin- 4j

antithrombin complex to deactivate thrombin-bound to fibrin clots. This may be impor- 2

tant clinically in the setting of acute coronary syndromes, where heparin may not be able |

to inactivate active thrombin bound to fibrin. The third commercially available direct @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

194 HOCH

thrombin inhibitor, bivalirudin, is indicated for use as an anticoagulant in patients with
unstable angina undergoing percutaneous transluminal coronary angioplasty with the
concomitant use of ASA. All three of the commercially available direct thrombin
inhibitors are administered intravenously. Ximelagatran is an orally active direct thrombin
inhibitor currently under investigation. The pharamacokinetics, clinical use, and compli-
cations associated with these direct thrombin inhibitors are reviewed below.

A. Lepirudin

Hirudin is a potent direct inhibitor of thrombin that was originally isolated from the
salivary glands of the medicinal leach Hirudo medicinalis . Lepirudin and bivalirudin are
recombinant proteins derived from the structure of hirudin. Lepirudin is a recombinant
hirudin derived from yeast cells; it is a polypeptide composed of 65 amino acids and is
identical to natural hirudin except for several amino acid substitutions. Hirudin binds
noncovalently to thrombin's active site, yet the strength of this bond is such that hirudin is
only very slowly reversible. Lepirudin has a plasma half-life of 40 min after intravenous
administration and approximately 120 min after subcutaneous injection (78). After
intravenous infusion, peak levels occur rapidly; then lepirudin is cleared by the kidneys,
with clearance proportional to the glomerular filtration rate. Thus dose adjustment is
recommended based on the patient's creatinine clearance.

Lepirudin has been shown to be an effective treatment for HIT based on prospective
trials (79). Patients with normal renal function are given a bolus dose of 0.4 mg/kg followed
by an infusion of 0.1-0.15 mg/kg/h to maintain the APTT at 1.5-2.5 times normal control
values. APTT values are known to increase in a dose-dependent fashion. Some investigators
have noted poor linearity in reproducibility of the APTT for monitoring of direct thrombin
inhibitors. The ecarin clotting time (ECT) is currently under clinical investigation for
monitoring patients receiving direct thrombin inhibitors. Ecarin, a snake venom enzyme,
converts prothrombin to meizothrombin, which is then neutralized by hirudin resulting in a
dose-dependent prolongation of clotting time. The ECT has been shown to have a linear
correlation with plasma hirudin concentrations whereas both the ACT and APTT have poor
correlations (80). Lepirudin has been shown in anecdotal series to be an effective alternative
to heparin during cardiopulmonary bypass in patients with HIT (81).

Metanalysis of two prospective trials that investigated the use of intravenous lepi-
rudin to manage patients with HIT demonstrated a clear advantage of lepirudin therapy
over historical controls. Lepirudin-treated patients had significantly lower incidences of
the combined endpoints of death, limb amputation, and thromboembolic complications
compared to historic controls (p = 0.004) (82). The primary complication of lepirudin
treatment is hemorrhage. The total incidence of bleeding was 42% at 35 days in the lepi-
rudin-treated group compared to 23.6% among historical controls (p — 0.001). Common ■g
sites of hemorrhage include intracranial, retroperitoneal, and gastrointestinal bleeding. &
Although lepirudin does not cross-react with HIT antibodies, antihirudin antibodies de- a
velop in between 56 and 74% of patients treated for longer than 5 days (81). Hirudin- c
antihirudin antibody complexes, while not affecting activity, prolongs the renal clearance <j
of lepirudin, which can potentiate its anticoagulation effect. While prospective trials have >9
shown decreased thromboembolic complications in patients with HIT receiving parental 41
lepirudin compared to historic controls, lepirudin unfortunately is not completely pro- 2
tective in this setting. Of 113 patients evaluated, 6.2% of the patients underwent limb |
amputation, 10.6% experienced a new thromboembolic complication, and the mortality ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 195

rate was 9.7% (82). The major limitation of lepirudin treatment is that it is contra-
indicated in patients with significant renal insufficiency. Whereas lepirudin may have a
slight effect on the INR when warfarin is initiated, its effect is not as pronounced as seen
with argatroban. Thus transition to warfarin therapy is less complicated with lepirudin
than with argatroban.

B. Bivalirudin

Bivalirudin is a recombinant protein based on hirudin which, unlike hirudin, produces
only a transient inhibition of the active site of thrombin. Once bivalirudin is bound to
thrombin, it is converted into a lower-affinity inhibitor (78). The onset of action is rapid
after intravenous administration, with a peak response seen within 15 min. Bivalirudin has
a plasma half-life of 25 min, with clearance by renal mechanisms and proteolytic cleavage.
Patients with renal insufficiency require adjustment in the rate of administration of bivali-
rudin; patients with glomerular filtration rates less than 30 mL/min require a reduction in
dose of approximately 68% (83). Bivalirudin inhibits both free and clot-bound thrombin.
The manufacturer's recommendations are to monitor bivalirudin's anticoagulation affect
using APTT, as plasma concentrations of bivalirudin correlated with APTT at all levels of
renal function (83).

Clinical trials have demonstrated that bivalirudin is as least as effective as high-dose
heparin when combined with ASA at preventing ischemic complications in patients under-
going percutaneous transluminal angioplasty for unstable angina (84). In the study that
led to FDA approval, patients received what is now the recommended clinical dose of
bivalirudin prior to angioplasty. The recommended dose is an intravenous bolus of 1 mg/
kg followed by a continuous infusion of 2.5 mg/kg/h for 4 h, with an additional 0.2 mg/
kg/h for up to 20 h if needed. Patients in the heparin group in this double-blind, random-
ized trial received a high-dose heparin regimen consisting of a bolus dose of 175 IU/kg
followed by an 18- to 24-h infusion at a rate of 15 IU/kg/h (84). ACTs were measured at 5
and 45 min after administration of the heparin bolus, with ACTs kept greater than 350 s
by periodic rebolusing with heparin. The primary endpoint of the study was in-hospital
death, MI, rapid clinical deterioration, or abrupt vessel closure. There was not a significant
difference in the incidence of the primary endpoint between groups; however, there was a
lower incidence of bleeding (3.8% bivalirudin vs. 9.8% heparin; p < 0.001). In the
bivalirudin group, patients who presented with postinfarction angia who were randomized
to bivalirudin therapy did demonstrate a lower incidence of the primary endpoint (9.1 vs.
14.2%, p = 0.04) and a lower incidence of bleeding (3% vs. 11.1%,/? < 0.001) compared
with the group treated with high-dose heparin. There was no difference, however, in the
cumulative rate of death between groups at 6 months. These data later contributed to the
FDA's decision to approve the use of bivalirudin when combined with ASA in patients
with unstable angia or postinfarction angina undergoing coronary angioplasty. Bivalir- e

udin has shown a benefit in preventing DVT in patients undergoing orthopedic surgery s

and as an adjunct to streptokinase in patients with acute myocardial infarction, although H,

these are not approved indications (78). =j

4

C. Argatroban |

Argatroban is a small, synthetic molecule derived from arginine that binds reversibly and 2

specifically to the catalytic domain of thrombin (85). Argatroban has the ability, because of §

its small size, to be an effective inhibitor of thrombin bound to surfaces as well as in solution. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

196 HOCH

Argatroban is about 50% protein-bound and 50% free in plasma and has an elimination
half-life of 39-51 min (86). Argatroban is metabolized in the liver and is not cleared by the
kidney, therefore renal function does not affect argatroban's pharmacokinetics (85, 86). Like
lepirudin, argatroban does not cross-react with heparin-associated antibodies; however,
unlike lepirudin, drug-specific antibodies to argatroban have not been identified. Argatro-
ban is indicated in the United States for the treatment and prophylaxis of HIT.

A prospective, historically controlled study examined the efficacy of argatroban to
reduce the endpoints of death, amputation, and new thromboembolic events compared to
historic controls (86). Patients received 2 jig/kg/min of argatroban and had the rate of
administration adjusted to maintain the APTT at 1.5-3 times baseline levels. The com-
posite endpoint was significantly reduced in the argatroban-treated patients versus control
patients with HIT (25.6 vs. 38.8%, p = 0.014). The incidence of major bleeding episodes
was not different between patient groups. Argatroban has also been approved in the
United States for use during percutaneous interventions in patients with HIT. In this
setting, argatroban is given as an intravenous bolus of 350 Hg/kg followed by continuous
infusion at 1 5-40 |ig/kg/min. The infusion rate was adjusted to maintain a target ACT of
300-450 s. Lewis et al. reported a series of 91 HIT patients who underwent 112 percuta-
neous coronary interventions using intravenous argatroban (87). All patients received
aspirin (325 mg) 2-24 h before angioplasty. Sheaths were removed no sooner than 2 h after
cessation of argatroban and when the ACT was less than 160 s. Satisfactory outcome of
the procedure was attained in 94.5% of the group, and 97.8% achieved adequate anti-
coagulation. The major bleeding rate was 1.1%, with a minor bleeding rate of 32%. All
patients remained free of the major acute complication of death, emergent coronary by-
pass graft surgery, and Q-wave MI. There are favorable anecdotal reports in the literature
reporting successful outcomes in patients undergoing cardiopulmonary bypass with
argatroban anticoagulation (88). Although this is not an approved indication for the
use of argatroban in the United States, the authors cite the advantage of argatroban over
lepirudin for patients with renal insufficiency and a history of HIT requiring cardiopul-
monary bypass.

Hemorrhage is the major complication of all direct thrombin inhibitors. There have
been no direct comparisons between agents. There is no antidote to the anticoagulation
effects of the direct thrombin inhibitors. Other commonly reported complications of aga-
troban use include diarrhea in 11% of patients, pain in 9%, and rash in 2% (87). Among
direct thrombin inhibitors, argatroban alone can cause significant prolongation of the INR.
The INR is affected by the argatroban concentration and the degree of warfarin-induced
factor activity depletion. Thus it is difficult to transition patients requiring prolonged anti-
coagulation from argatroban to warfarin.

D. Ximelagatran/Melagatran -g

Melagatran is a dipeptide direct thrombin inhibitor that binds to the active site of throm- |

bin in a competitive yet reversible manner (89). Melagatran can be administered intra- j§

venously or subcutaneously, but it is not well absorbed orally. Ximelagatran is an oral 2

direct thrombin inhibitor currently under investigation for the prevention and treatment of ^

thromboembolism. Ximelagatran can be taken by mouth and is then converted to mela- "jl

gatran the active agent. Peak plasma concentrations of melagatran are attained 2 h after H

ximelagatran oral administration. Plasma half-life ranges from 2.5 to 3.5 h (90). Melaga- °

tran is not metabolized and is cleared primarily by the kidneys; plasma half-life is doubled J

in patients with renal failure. o

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 197

Laboratory monitoring does not appear necessary for the clinical use of ximelagatran.
The pharmacokinetics, pharmacodynamics, and clinical effects of oral ximelagatran were
analyzed in patients with PE and DVT (89). Patients received a fixed-dose of 48 mg oral
ximelagatran twice daily for 6-9 days. APTT correlated with plasma melagatran concen-
trations, peaking at approximately two times the baseline APTT. Clinical symptoms
improved in all 12 patients and no deaths or severe bleeding occurred. Thus fixed-dose
ximelagatran demonstrated reproducible pharmacokinetics and pharmacodynamics in
patients with PE without routine coagulation monitoring.

A randomized, double-blind study examined the dose-response of subcutaneous
melagatran followed by oral ximelagatran as thromboprophylaxis in patients undergoing
total hip or knee replacement. Efficacy and safety were compared to those of dalteparin
(91). A significant dose-dependent decrease in venous thromboembolism was seen with
melagatran/ximelagatran. The incidence of venous thromboembolism was significantly
lower in the highest-dose melagatran/ximelagatran group compared to dalteparin (15.1 vs.
28.2%,/) < 0.0001). Although ximelagatran remains in clinical trials and its safety profile
and complication risks are still to be defined, it appears to have future promise for the
management of patients with thromboembolism as an oral direct thrombin inhibitor that
does not require laboratory monitoring.

V. FACTOR Xa INHIBITORS

Fondaparinux sodium is a synthetic pentasaccharide that is an indirect inhibitor of factor
Xa. Fondaparinux is an analogue to the pentasaccharide sequence found on small heparin
and LMWH molecules ( < 18 saccharides) that acts to catalyze the inhibition of factor Xa
by antithrombin (92). By inhibiting factor Xa selectively, fondaparinux inhibits thrombin
generation without a direct effect on thrombin activity. Unlike UFH and LMWH, fond-
aparinux does not cause the release of TFPI from endothelial cells, and there is minimal
binding to plasma proteins. Fondaparinux bioavailability is excellent when administered
subcutaneously, with peak concentrations 2 h postdosing. Its plasma half-life is 14-20 h.
There is no metabolism of the drug prior to renal excretion. This pharmacokinetic profile
allows for once-daily subcutaneous administration with no need for laboratory monitor-
ing or dose adjustment in the majority of patients.

In December 2001, the FDA approved the use of fondaparinux sodium for reducing
the risk of VTE after orthopedic surgery for hip fracture, hip replacement, and knee re-
placement. Four multicenter, randomized, double-blind trials in patients undergoing elec-
tive hip replacement, extensive knee surgery, and surgery for hip fracture have compared
the efficacy and safety of fondaparinux to that of enoxaparin for the prevention of VTE
(93). Patients received either once-daily subcutaneous injections of fondaparinux (2.5 mg)
beginning 6 h after surgery or subcutaneous enoxaparin according to approved protocols 13

for the prevention of VTE (93). Metanalysis revealed that fondaparinux significantly |

reduced the incidence of VTE by day 11 (6.8%) compared to enoxaparin (13.7%, p < j§

0.001). Although fondaparinux achieved an overall 55% reduction in the risk of VTE 2

disturbingly, major bleeding occurred more frequently than with the enoxaperin regimens ^

(p — 0.008). Subsequent analyses indicate that administration of fondaparinux earlier "jl

than 6 h postoperatively is associated with an increased incidence of major bleeding (94). H

Fondaparinux is contraindicated in patients with active bleeding, thrombocytopenia, body °

weight less than 50 kg, and severe renal insufficiency because of an increased risk of major J

bleeding in these groups. Concern over the perceived increased risk of bleeding compli- o

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

198 HOCH

cations with fondaparinux led to the evaluation of strategies to reverse its anticoagulation
effects. A placebo-controlled randomized trial examined the ability of recombinant factor
Vila (rFVIIa) to reverse the effects of fondaparinux in healthy adults (95). Prolongation of
the APTT and PT caused by fondaparinux was reversed with a single intravenous bolus of
rFVIIa (90 jig/kg). Thrombin-generation time returned to normal by 6 h after rFVIIa
injection. These data suggest the potential for the use of rFVIIa to reverse the anti-
coagulation effect of fondaparinux in the event of major bleeding complications.

VI. PLATELET FUNCTION INHIBITORS
A. Pharmacology

1. Aspirin
Aspirin (ASA) produces an irreversible inactivation of the cyclooxygenase (COX)

activity of COX-1 and COX-2 by acetylating critical serine residues, blocking access of
substrate to the enzymes' active site. Inactivation of COX activity interferes with platelet
aggregation due to the inhibition of thromboxane A 2 (TXA 2 ). TXA 2 induces platelet
aggregation and vasoconstriction. Platelets are anucleate, thus inhibition of TXA 2 is
permanent. ASA is 50- to 100-fold more potent at inhibiting platelet-associated COX-1
than monocyte-associated COX-2 (96). TXA 2 is primarily derived from the action of
COX-1. The plasma half-life of aspirin is 15-20 min; thus ASA targets COX-1.
Considering that only 10% of circulating platelets are replaced daily, once-a-day dosing
with aspirin is effective at blocking TXA 2 .

2. Cilostazol

Cilostazol is a quinolinone derivative that is a potent phosphodiesterase type III
inhibitor (97). Phosphodiesterase III inhibition causes elevation of intracellular cAMP
levels in platelets and vascular smooth muscle cells, resulting in vasodilatation and
inhibition of platelet aggregation. In addition, cilostazol has been reported to inhibit
smooth muscle proliferation and to increase levels of HDL cholesterol (97). Cilostazol also
inhibits adenosine uptake, leading to increased cAMP in platelets and smooth muscle but
decreased cAMP in the heart. Thus the effect of adenosine uptake inhibition on the heart
may in part offset the increase in cAMP and the cardiotonic effect attributed to phospho-
diesterase III inhibition. Although cilostazol's approved use in the United States is for the
management of claudication, it is also an effective platelet inhibitor. Clinicians must be
cognizant of this, as many claudicators are already on platelet inhibitors. In healthy
volunteers, after a single 100-mg dose of cilostazol, peak plasma levels occurred at 3.6 h; the
maximum inhibition of platelet aggregation was 31.1% at 6 h (98).

3. Thienopyridines
The thienopyridines — ticlopidine and clopidogrel — are structurally related potent s

platelet inhibitors. They are adenosine diphosphate (ADP) receptor antagonists, which H,

inhibit ADP induced platelet aggregation without affecting the arachidonic acid metabolism =j

(99). Both are metabolized in the liver to active metabolites that cause irreversible alterations g

in ADP receptors. Ticlopidine and clopidogrel cause dose- and time-dependent inhibition £

in platelet aggregation. After the initiation of therapy, bleeding times reach a peak of 1.5-2 j|

times baseline values over a period of 3-7 days. Recovery of platelet function occurs slowly, "g

over about the same length of time it took to achieve platelet inhibition. When combined, g

ASA and ticlopidine or clopidogrel act synergistically to inhibit platelet aggregation. 8

t

u

£

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 199

4. Platelet Glycoprotein llb/llla Inhibitors

Regardless of the stimulus leading to platelet activation, the platelet glycoprotein lib/
Ilia (GPIIb/IIIa) receptors act as the final common step leading to platelet aggregation.
Fibrinogen and von Willebrand factor bind to the GPIIb/IIIa receptor, leading to platelet
aggregation and thrombus formation. Glycoprotein Ilb/IIIa inhibitors compete with fi-
brinogen and von Willebrand factor to occupy the receptor (100). Currently three GPIIb/
Ilia inhibitors are available: abciximab, tirofiban, and eptifbatide. The goal of therapy is
to achieve greater than 80% receptor blockade, which will effectively abolish platelet
aggregation.

Abciximab is a monoclonal antibody fragment to the GPIIb/IIIa receptor (101). It is a
large molecule that binds to the receptor rapidly and tightly, accounting for the short 30-
min plasma half-life yet prolonged dissociation from the receptor (40 min). A bolus dose
of 0.25 mg/kg blocks greater than 80% of receptors, prolonging the bleeding time; bleed-
ing times return to normal values after 12 h. After the bolus, an infusion of abciximab at
0.125 |ig/kg/min can effectively inhibit platelet aggregation for 12 h. Platelet transfusion
can reverse the effect of abciximab on platelets. Dosage adjustment for renal insufficiency
is not necessary for abciximab.

Tirofiban is a small nonpeptide molecule; it is a reversible antagonist of fibrinogen
binding to GPIIb/IIIa receptor (102). Eptifibatide is a synthetic cyclic heptapeptide that is
also a competitive inhibitor of GPIIb/IIIa receptor (103). Both of these agents have very
rapid rates of dissociation from the GPIIb/IIIa receptor (10-20 s), so that platelet
inhibition is dependent on the plasma drug level. Once drug is stopped, bleeding times
return to normal within 2-4 h (104). Tirofiban and eptifibatide require dose adjustments
for patients with renal insufficiency.

B. Clinical Applications

Antiplatelet therapy has been proven to decrease adverse cardiovascular events in patients
with atherosclerosis. Metanalysis by the Antithrombotic Trialists' Collaboration has shown
that antiplatelet therapy in patients at high risk — defined as having previous stroke/TIA,
previous MI or acute MI — results in a 25% proportional reduction in serious vascular
events (vascular death, MI, and stroke) (105). ASA was the primary antiplatelet therapy in
the majority of patients. An earlier metanalysis had not shown a significant benefit for ASA
in patients with peripheral arterial disease (PAD) (106). There was a proportional 23%
reduction in serious vascular events in the 9214 patients with PAD in the most recent
metanalysis, which included trials with other antiplatelet agents such as clopidogrel. The
thienopyridines have proven effective agents for patients with peripheral arterial disease.
The Swedish Ticlopidine Multicentre Study (STIMS) demonstrated a 29.1% lower mortal-
ity in the ticlopidine group compared to placebo (107). The Clopidogrel versus Aspirin in
Patients at Risk of Ischaemic Events (CAPRIE) trial compared clopidogrel to aspirin in e

patients with symptomatic atherosclerosis regardless of symptom location (coronary, e

cerebral, or peripheral) (108). Based on the first occurrence of ischemic stroke, MI, or H,

vascular death, patients treated with clopidogrel showed a relative risk reduction of 8.7% a

over and above the 25% reduction currently accepted with ASA. However in patients with g

PAD, there was a 24% risk reduction in serious vascular events in the clopidogrel group sf

compared to ASA. Cilostazol and the GPIIb/IIIa inhibitors have not been studied in 3

secondary prevention trials. "g

The management of acute stroke with platelet inhibitors and/or anticoagulants is g

appealing in concept (i.e., to decrease the risk of ongoing thrombosis), yet the con- 8

t
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

200 HOCH

sequences of hemorrhagic complications can be devastating. Metanalysis reveals that ASA
therapy results in four fewer nonfatal strokes and five fewer vascular deaths per 1000 pa-
tients presenting with acute stroke (105). Accepting that there is an increased risk of hemor-
rhage with ASA, patients benefit from ASA in the setting of acute stroke if computed
tomography reveals the absence of intracranial hemorrhage. Metanalysis of clinical trials
evaluating the addition of anticoagulation strategies to ASA during acute stroke reveals a
significant increase in mortality and rate of intracranial hemorrhage in the anticoagulation
group (109). GPIIb/IIIa inhibitors are under investigation for use in the management of
acute ischemic stroke. Preliminary data from a dosing study imply a trend toward improved
outcomes versus placebo (110).

Platelet inhibitors decrease the risk of bypass graft occlusion in patients undergoing
peripheral artery reconstructive surgery. The Antithrombotic Trialists' Collaboration met-
analysis of trials designed to assess the effect of platelet inhibitors on the fate of arterial or
graft patency found that the platelet inhibitor groups had a significantly lower rate of graft
occlusion (16%) than controls (25%, p<0. 0001) (111). The majority of trials used ASA as
their platelet inhibitor. While clopidogrel's affect on graft patency has not been evaluated,
ticlopidine has been prospectively evaluated in patients undergoing femoropopliteal or
femorotibial saphenous-vein bypass grafts. A total of 243 patients were randomly assigned
to receive either ticlopidine (250 mg twice a day) or matching placebo for 2 years. The 2-year
cumulative patency rate was 82 % in the ticlopidine group and 63 % in the placebo group
(p = 0.002) (112).

Platelet inhibitors clearly benefit patients experiencing acute coronary syndromes and
improve outcomes following percutaneous transluminal coronary angioplasty (PTCA)
and stenting. The Clopidogrel in Unstable Angina to Prevent Recurrent Ischaemic Events
Trial (CURE) randomized over 12,500 patients to receive in a double-blind manner, either
placebo or clopidogrel (113). All patients received ASA (75-325 mg daily), while the
clopidogrel arm received an initial 300-mg loading dose of clopidogrel followed by a 75-
mg daily dose. Compared to ASA alone, the combination of ASA and clopidogrel reduced
the composite risk of cardiovascular death, stroke, and MI, with a relative risk reduction
of 20% (p< 0.001). The protective effect of clopidogrel was apparent within 2 h of
initiation of therapy, and this benefit was consistent at 30 days and after long-term
treatment. The Clopidogrel for the Reduction of Events During Observation (CREDO)
trial evaluated the effect of ASA and clopidogrel dual therapy following PTCA and stent
placement in a randomized, double-blind trial enrolling 2116 patients (114). Patients
received either a preprocedure bolus of 300 mg clopidogrel followed by 75 mg daily for 12
months or a placebo bolus and clopidogrel for 28 days. At 1 year, long-term clopidogrel
therapy was associated with a 26.9% relative reduction in the combined risk of death, MI,
or stroke (p = 0.02) The risk of major bleeding at 1 year increased in the long-term group,
but not significantly (8.8 vs. 6.7%, p = 0.07). -g

Cilostazol combined with ASA has been studied as a platelet inhibitor for the &

prevention of coronary stent occlusion and restenosis (115). When cilostazol was compared a

to ticlopidine, there were no differences in the rate of early stent occlusion, but the incidence c

of long-term restenosis was reduced with cilostazol (13%) compared to ticlopidine (31%, <j

p<0.05). A recent metanalysis of 19 randomized, placebo-controlled trials was done to >9

evaluate the impact of intravenous antagonists of the GPIIb/IIIa receptor on the survival of 41

patients undergoing PTCA and stenting (116). Mortality and Mis were significantly reduced 2

in the GPIIb/IIIa group at 30 days and 6 months. Major bleeding was increased only in the |

trials that continued intravenous heparin after the procedure. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 201

The role of platelet inhibitors during peripheral arterial angioplasty/stent placement has
been studied in only a limited fashion. Applying the data from the coronary system to the
peripheral arteries, clinicians have adopted an antithrombotic strategy of periprocedural use
of ASA/clopidogrel and heparin. We recommend 325 mg of ASA at least 24 h prior to the
procedure and clopidogrel 75 mg daily for 3-5 days prior to the procedure or 300 mg of
clopidogrel at least 6 h preoperatively. During the intervention, heparin is given to maintain
an activated clotting time greater than 200 s. Use of GPIIb/IIIa inhibitors in high-risk
patients undergoing carotid stent placement has been compared, in a nonrandomized
fashion, to intravenous UFH in lower-risk patients and found to decrease the incidence
of ischemic stroke (3 vs. 12%, n.s.) (117). However, 5% of the GPIIb/IIIa-treated patients
suffered intracranial hemorrhage. GPIIb/IIIa inhibitors all appear to have the ability to
disaggregate acute platelet-rich thrombi (118). We have used GPIIb/IIIA inhibitors
selectively to treat peripheral artery dissection during angioplasty procedures and have
witnessed disaggregation of acute thrombi. Controlled clinical trials are warranted.

C. Complications and Failures

All platelet inhibitors place patients at risk for developing bleeding complications. A
decision must be weighed for each patient, balancing the patient's risk of thromboembolic
event with the risk of the platelet inhibitor. The antithrombotic effect of ASA appears to
be independent of the dose (75 to 1300 mg) (104). The recent Antiplatelet Trialists' Col-
laboration metanalysis of patients with vascular disease demonstrated that a low dose of
ASA (75-160 mg) offers the same protection against thrombotic events as higher doses,
yet with a lower risk of gastrointestinal hemorrhage (105). The higher a patient's risk of
cardiovascular complications, the more favorable is ASA's safety profile. Ticlopidine and
clopidogrel have a potential similar to that of ASA to cause bleeding complications in pa-
tients with cardiovascular disease, but they are less likely to cause gastrointestinal hemor-
rhage (119). In the CURE trial, the overall risk of major bleeding was increased in the
clopidogrel-plus-ASA group compared to the placebo (ASA) group (3.7 vs. 2.7%, p =
0.001), but the incidence of life-threatening bleeding (2.2 vs. 1.8 %, p = 0.13) or hemor-
rhagic strokes (0.1 vs. 0.1%) was similar (113). As clopidogrel has gained wider acceptance,
many surgeons have anecdotally reported their experience with prolonged needle-hole
bleeding. Hongo et al. prospectively compared patients undergoing coronary artery bypass
graft surgery whom had clopidogrel exposure within 7 days prior to surgery to those patients
never exposed to clopidogrel (120). The clopidogrel patients had a 10-fold higher rate of
reoperation for bleeding and significantly increased chest tube output and as well as greater
transfusion requirements.

Along with the hope that GPIIb/IIIa receptor inhibitors would improve patient out-
comes in interventions with a high likelihood of thromboembolism has come concern ■g
regarding their potential for significant bleeding complications. The EPIC trial first dem- &
onstrated that GPIIb/IIIa therapy combined with traditional antithrombotic therapy (ASA a
and UFH) decreased ischemic events in patients undergoing PTC A (121). However, the inci- -c
dence of major bleeding was increased in the abciximab treated patients (14%) compared to <j
those receiving placebo (7%, p = 0.001). All the GPIIb/IIIa inhibitors share this heightened -9
bleeding risk; bleeding risk can be decreased with early cessation of intravenous UFH, 4j
careful sheath management, and the use of reduced, weight-adjusted doses of heparin (116). Q

Gastrointestinal side effects are frequent with ASA. The gastrointestinal toxicity of |

ASA is dose-related (122). The overall relative risk of upper gastrointestinal complications @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

202 HOCH

associated with ASA is between 2.2 and 3.1% (122). Diarrhea, nausea, and vomiting can
occur in up to 30-50% of patients using ticlopidine (99). Patients using clopidogrel expe-
rience far fewer gastrointestinal side effects than those using ticlopidine. Patient compli-
cations in the CAPRIE trial included rash (0.26%), diarrhea (0.23%), upper gastrointes-
tinal discomfort (0.97%), and gastrointestinal hemorrhage (0.52%) (108).

Ticlopidine and clopidogrel have both been reported to cause thrombocytopenia and
thrombotic thrombocytopenic purpura (99,123). Drug withdrawal usually reverses TTP;
periodic monitoring of platelet counts is recommended at the initiation of therapy. Ticlo-
pidine has also been associated with neutropenia and aplastic anemia, while neutropenia is
rare with clopidogrel (104). Neutropenia is the most serious side effect of ticlopidine treat-
ment, occurring in 2. 1 % of patients (99). Cases usually develop during the first 3 months
of therapy; thus complete blood counts and platelet counts are recommended every 2 weeks
during this period.

Thrombocytopenia is a significant complication of GPIIb/IIIa receptor inhibitor ther-
apy. Abciximab can cause drops in platelet counts to <50,000/pL in 1-2% of patients, with
counts dropping 0.5-1.0% in the first 2 h after the start of treatment (104). Thrombocyto-
penia is reversible with withdrawal of therapy, but this may take several days. Both tiro-
fiban and eptifibatide can cause equally dramatic reversible thrombocytopenia; in their
case, the mechanism appears immune-mediated (104,124).

Headache is a common side effect of cilostazol, occurring in 34% of cilostazol-treated
patients compared to 14% of placebo-treated patients. Discontinuation of drug because of
headache occurred in 3.7% of cilostazol-treated patients compared to 1.3% of placebo-
treated patients (125). Other common side effects included palpitations in 10% and diar-
rhea in 19%. When the FDA approved cilostazol for clinical use, a prior history of heart
failure was a contraindication. This recommendation was based on the increased mortality
rate seen in patients with heart failure using older phosphodiesterase inhibitors. Pratt
pooled the safety data from eight phase III clinical trials evaluating cilostazol (125). He
reported that the cardiovascular morbidity and all-cause mortality was 6.5% for cilostazol
(100 mg bid) compared to 7.7% for placebo groups.

Most failures of platelet function inhibitors are related to inappropriate clinical
application, incorrect dosage, or delayed initiation of therapy. Future improvements to
facilitate the monitoring of platelet function, hopefully at the bedside, will likely improve
our clinical outcomes.

REFERENCES

1 . Rosenberg RD, Bauer KA. The heparin-antithrombin system: A natural anticoagulant mecha-
nism. In: Colman RW, Hirsh J, Marder VI, et al, eds. Hemostasis and Thrombosis: Basic
Principles and Clinical Practice. 3rd ed. Philadelphia: Lippincott, 1994:837-860. 1

2. Casu B, Oreste P, Torri G, Zoppetti G, Choay I, Lorneau IC, Petitou M, Sinay P. The g
structure of heparin oligosaccharide fragments with high anti-(factor Xa) activity containing js
the minimal antithrombin Ill-binding sequence. Biochem I 1981; 97:599-609. 2

3. Lam LH, Silbert IE, Rosenberg RD. The separation of active and inactive forms of heparin. ^
Biochem Biophys Res Commun 1976; 69:570-577. >9

4. Tollefsen DM, Majerus DW, Blank MK. Heparin cofactor II: Purification and properties of a jjj
heparin-dependent inhibitor of thrombin in human plasma. I Biol Chem 1982; 257:2162-2169. q

5. Choay J. Structure and activity of heparin and its fragments: an overview. Semin Thromb |
Hemost 1989; 15:359-364. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 203

6. Haas S, Bluemel G. An objective evaluation of the clinical potential of low molecular weight
heparins in the prevention of thromboembolism. Semin Thromb Hemost 1989; 15:424-434.

7. Lupu C, Poulsen E, Roquefeuil S, Westmuckett AD, Kakkar VV, Lupu F. Cellular effects of
heparin on the production and release of tissue factor pathway inhibitor in human endothelial
cells. Arterioscler Thromb Vase Biol 1999; 19:2251-2262.

8. Ofosu FA, Sie P, Modi GJ, Fernandez F, Buchanan MR, Blajchman MA, Boneu B. The
inhibition of thrombin-dependent feedback reactions is critical to the expression of anti-
coagulant effects of heparin. Biochem J 1987; 243:579-588.

9. Ockelford PA, Carter CJ, Cerskus A, Smith CA, Hirsh J. Comparison of the in vivo hemor-
rhagic and antithrombotic effects of a low antithrombin III affinity heparin fraction. Thromb
Res 1982; 27:679-690.

10. Salzman E W, Rosenberg RD, Smith MH, Lindon JN, Favreau L. Effect of heparin and heparin
fractions on platelet aggregation. J Clin Invest 1980; 65:64-73.

1 1 . Hull RD, Raskob GE, Hirsh J, Jay RM, Leclerc JR, Geerts WH, Rosenbloom D, Sackett DL,
Anderson C, Harrison L. Continuous intravenous heparin compared with intermittent sub-
cutaneous heparin in the initial treatment of proximal-vein thrombosis. N Engl J Med 1986;
315:1109-1114.

12. Hirsh J, Warkentin TE, Shaughnessy, Anand SS, Halperin JL, Raschke R, Granger C, Ohman
EM, Dalen JE. Heparin and low-molecular-weight heparin. Mechanisms of action, pharmaco-
kinetics, dosing, monitoring, efficacy, and safety. Chest 2001; 119:64S-94S.

13. Tolleson TR, O'Shea JC, Bittl JA, Hillegass WB, Williams KA, Levine G, Harrington RA,
Tcheng JE. Relationship between Heparin anticoagulation and clinical outcomes in coronary
stent intervention: observations from the ESPRIT trial. J Am Coll Cardiol 2003; 41:386-393.

14. Collins R, Scrimgeour A, Yusuf S, Peto R. Reduction in fatal pulmonary embolism and venous
thrombosis by perioperative administration of subcutaneous heparin. Overview of results of
randomized trials in general, orthopedic, and urologic surgery. N Engl J Med 1988; 318:1 162-
1173.

15. Hirsh J, Lee AY. How we diagnose and treat deep vein thrombosis. Blood 2002; 99:3102-31 10.

16. Hirsh J, Dalen JE, Anderson DR, Poller L, Bussey H, Ansell J, Deylon D. Oral antico-
agulants: Mechanism of action, clinical effectiveness, and optimal therapeutic range. Chest
2001; 119:8S-21S.

17. Forbes TL, DeRose G, Harris KA. Is long-term anticoagulation after acute thromboem-
bolic limb ischemia always necessary? Can J Surg 2002; 45:337-340.

18. Kakkar VV. Prevention of venous thromboembolism. Clin Hematol 1981; 10:543-583.

19. Kapsch DN, Silver D. Complications in failures of anticoagulation therapy. In: Bernhard VM,
Towne JB, eds. Complications in Vascular Surgery. New York: Grune and Stratton, 1985:405-
419.

20. Calaitges JG, Silver D. Antithrombotic therapy. In: Rutherford RB, et al., eds. Vascular
Surgery 5th ed. Philadelphia: Saunders, 2000.

21. Aster RH. Heparin-induced thrombocytopenia and thrombosis. N Engl J Med 1995; 332: 1374—
1376.

22. Laster J, Cikrit D, Walter N, Silver D. The heparin-induced thrombocytopenia syndrome: an

update. Surgery 1987; 102:763-770. 1

23. Mureebe L, Silver D. Heparin-induced thrombocytopenia: pathophysiology and manage- <S
ment. Vase Endovasc Surg 2002; 36:163-170. 1

24. Silver D, Kapsch DN, Tsoi EDM. Heparin-induced thrombocytopenia, thrombosis and 2
hemorrhage. Ann Surg 1983; 198:301-306. ^

25. Greinacher A, Amiral J, Dummel V, Vissac A, Kiefel V, Mueller-Eckhardt C. Laboratory >9
diagnosis of heparin-associated thrombocytopenia and comparison of platelet aggregation jjj
tests, heparin-induced platelet activation tests and platelet factor 4/heparin enzyme-linked q
immunosorbent. Transfusion 1994; 34:381-385. |

26. Pouplard C, Amirral J, Boug J, Laporte-Simitsiclis S, Delahouse B, Gruel Y. Decision analysis 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

204 HOCH

for use of platelet aggregation test common carbon 14-0 tonin release assay, and heparin-
platelet factor 4 enzyme-linked immunosorbent assay for diagnosis of heparin-induced throm-
bocytopenia. Am J Clin Pathol 1999; 111:700-706.

27. Lindhoff-Last E, Gerdsen F, Ackermann H, Bauersach SR. Determination of heparin-
plate factor 4-IgG antibodies improves diagnosis of heparin-induced thrombocytopenia. Br J
Haematol 2001; 113:886-897.

28. Kikta MJ, Keller MP, Humphrey PW, Silver D. Can low molecular weight heparins and
heparinoids be safely given to patients with heparin-induced thrombocytopenia syndrome?
Surgery 1993; 114:705-710.

29. Slocum MM, Adams JG Jr, Teel R, Spadone DP, Silver D. Use of enoxaparin in patients
with the heparin-induced thrombocytopenia syndrome. J Vase Surg 1996; 23:839-843.

30. Kaplan KL, Francis CW. Direct thrombin inhibitors. Semin Hematol 2002; 39:187-196.

31. Jaffe MD, Willis PW. Multiple fractures associated with long-term sodium heparin therapy.
JAMA 1965; 193:152-154.

32. Gerrts WH, Heit JA, Clagett GP, Pineo GF, Colwell CW, Anderson FA, Wheeler HB. Pre-
vention of venous thromboembolism. Chest 2001; 119: 132S— 175S.

33. Levine MN, Hirsh J, Gent M, Turpie AG, Cruickshank M, Weitz J, Anderson D, Johnson M.
A randomized trial comparing activated thromboplastin time with heparin assay in patients
with acute venous thromboembolism requiring large daily doses of heparin. Arch Intern Med
1994; 154:49-56.

34. Thomas DP. Heparin. Clin Hematol 1981; 1:443^58.

35. Hirsh J. Low-molecular-weight heparin: A review of the results of recent studies of the
treatment of venous thromboembolism and unstable angina. J Circ 1998; 98:1575-1582.

36. Nielsen JI, Ostergaard P. Chemistry of heparin and low molecular weight heparin. Acta Chir
Scand Suppl 1988; 534:52-56.

37. Altman R, Scazziota A, Rouvier J. Efficacy of unfractionated heparin, low molecular weight
heparin, and both combined for releasing total and free tissue factor pathway inhibitor.
Haemostasis 1998; 28:229-235.

38. Wheitz JI. Low-molecular-weight heparins. N Engl J Med 1997; 337:688-698.

39. Levine M, Gent N, Hirsh J, Leclerc J, Anderson D, Weitz J, Ginsberg J, Turpie AG, Demers C,
Kovacs M. A comparison of low-molecular- weight administered primarily at home with
unfractionated heparin administered in the hospital for proximal-deep vein thrombosis. N Engl
J Med 1996; 334:677-681.

40. Young E, Wells P, Holloway S, Weitz J, Hirsh S. Ex-vivo and in-vitro evidence that low
molecular weight heparins exibit less binding to plasma proteins than unfractionated heparin.
Thromb Haemost 1994; 71:300-304.

41. Warkentin TE, Levine MN, Hirsh J, Horsewood P, Roberts RS, Gent M, Kelton JG.
Heparin-induced thrombocytopenia in patients treated with low molecular weight heparin or
unfractionated heparin. N Engl J Med 1995; 332:1330-1335.

42. Laposata M, Green D, Van Cott EM, Barrowcliffe TW, Goodnight SH, Sosolik RC. The
clinical use in laboratory monitoring of low molecular weight heparin, danaparoid, hirudin
and related compounds, and argatroban: College of American Pathologists Conference X XXI

on laboratory monitoring of anticoagulant therapy. Arch Pathol Lab Med 1998; 122:799-807. j>

43. Etchells E, McLeod RS, Geerts W, Barton P, Desky AS. Economic analysis of low-dose g
heparin versus the low-molecular-weight heparin enoxaparin for prevention of venous js
thromboembolism after colorectal surgery. Arch Intern Med 1999; 159:1221-1228. 2

44. Koch A, Boughes S, Ziegler S, Dinkel H, Daures JP, Victor N. Low-molecular-weight heparin **
and unfractionated heparin in thrombosis prophylaxis after major surgical intervention: >9
update of previous meta-analyses. Br J Surg 1997; 84:750-759. jjj

45. Gould MK, Dembitzer AD, Doyle RL, Hastie TJ, Garber AM. Low-molecular-weight hepa- q
rins compared with unfractionated heparin for treatment of acute deep vein thrombosis. |
A meta-analysis of randomized, control trials. Ann Intern Med 1999; 130:800-809. 2

46. Dolovich LR, Ginsberg JS, Douketis JD, Holbrook AM, Cheah G A meta-analysis com- ^

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 205

paring low-molecular-weight heparins with unfractionated heparin in the treatment of venous
thromboembolism. Examining some unanswered questions regarding location of treatment,
product type, and dosing frequency. Arch Intern Med 2000; 160:181-188.

47. Siragusa S, Cosmi B, Piovella F, Hirsh J, Ginsbert JS. Low-molecular-weight heparins and
unfractionated heparin in the treatment of patients with acute venous thromboembolism:
Results of a meta-analysis. Am J Med 1996; 100:269-277.

48. Hull RD, Raskobe GE, Brant RF, Pineo GF, Elliot G, Stein PD, Gottschalk A, Valentine
KA, Mah AF. Low-molecular- weight heparin versus heparin in the treatment of patients with
pulmonary embolism. American-Canadian Thrombosis Study Group. Arch Intern Med 2000;
160:229-236.

49. Simonneau G, Sors H, Charbonnier B, Page Y, Laaban JP, Azarian R, Laurent M, Hirsch
JL, Ferrari E, Bossom JL, Mottier D, Beau B. A comparison of low-molecular-weight hepa-
rin with unfractionated heparin for acute pulmonary embolism. THESEE Study Group.
Tinzaparine ou heparine standard: evaluations dans l'envolie pulmonaire. N Engl J Med
1997; 337:663-669.

50. Friederich PW, Sanson BJ, Simioni P, et al. Frequency of pregnancy related venous thrombo-
embolism in anticoagulant factor deficient women: implications before prophylaxis. Arch
Intern Med 1996; 125:955-960.

51. Ginsberg JS, Grer I, Hirsh J. Use of antithrombotic agents during pregnancy. Chest 2001;
119:122S-131S.

52. Edmondson RA, Cohen AT, Das SK, Wagner MB, Kakkar VV. Low molecular weight
heparin versus aspirin and Dipyridamole after femoropopliteal bypass grafting. [Published
erratum appears in Lancet 1994; 344: 1307]. Lancet 1994; 344:914-918.

53. Simama CM, Gigou F, 111 P. Low molecular weight heparin versus unfractionated heparin in
femorodistal reconstruction surgery: a multi-center open randomized trial. Enoxart Study
Group. Ann Vase Surg 1995; 9(suppl):S45-S53.

54. Hingorani A, Gramse C, Ascher E. Anticoagulation with enoxaparin versus intravenous
unfractionated heparin in postoperative vascular surgery patients. J Vase Surg 2002; 36:341-
345. Comment in: J Vase Surg 2003; 37:700-701. Author replies 701.

55. Ardissino D, Merlini PA, Aarlens R, Coppola R, Bramucci E, Lucreziotti S, Repetto A,
Fetiveau R, Mannucci PM. Tissue factor in human coronary atherosclerotic plaques. Clin
Chim Acta 2000; 291:235-240.

56. Antman EM, Cohen M, Radley D, McCabe C, Rush J, Premmereur J, Braunwald E. Assess-
ment of the treatment effect of Enoxaparin for unstable angina/non-Q-wave myocardial
infarction: TIMI 11 B-ESSENCE meta-analysis. Circulation 1999; 100:1602-1608.

57. Wong GC, Giugliano RP, Antman EM. Use of low molecular weight heparins in the manage-
ment of acute coronary artery syndromes and percutaneous coronary intervention. JAMA
2003;289:331-342.

58. Schweizer J, Muller A, Forkmann L, Hellner G, Kirch W. Potential use of low-molecular-
weight heparin to prevent restenosis in patients with extensive wall damage following periph-
eral angioplasty. Angioplasty 2001; 52:659-669.

59. Nurmohamed MT, Rosendaal FR, Buller HR, Dekker E, Hommes DW, Vandenbroucke

JP, Briet E. Low-molecular-weight heparin versus standard heparin in general and orthopedic j>

surgery: a meta-analysis. Lancet 1992; 340:152-156. g

60. Crowther MA, Berry LR, Monagle PT, Chan AK. Mechanisms responsible for the failure of js
protamine to inactivate low-molecular-weight heparin. Br J Haematol 2002; 116:178-186. £

61. Lumpkin MN. FDA Public Health Advisory. Anesthesiology 1998; 88:27A-28A. ^

62. Luk C, Wells PS, Anderson D, Covacs MJ. Extended outpatient therapy of low molecular >9
weight heparin for the treatment of recurrent venous thromboembolism despite warfarin jjj
therapy. Am J Med 2001; 1 1 1:270-273. q

63. Bick RL. Antiphospholipid thrombosis syndromes. Hematol Oncol Clin North Am 2003; |
17:115-147. S

64. Muir JM, Hirsch J, Weitz JI, Andrew M, Young E, Shaughnessy SG A histomorphometric ^

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

206 HOCH

comparison of the effects of heparin in low molecular weight heparin on low cancellous in
bone in rats. Blood 1997; 89:3236-3242.

65. Suttie JW. Oral anticoagulant therapy: The biosynthetic basis. Semin Hematol 1977; 14:365-
374.

66. Zivelin A, Rao VM, Rapaport SI. Mechanisms of the anticoagulant effect of warfarin as
evaluated in rabbits by selective depression of individual procoagulant vitamin-K dependent
clotting factors. J Clin Invest 1993; 92:2131-2140.

67. Harrison L, lohnston N, Massicotte MP, Crowther M, Moffat K, Hirsh J. Comparison of 5-mg
and 10-mg loading doses in initiation of warfarin therapy. Ann Intern Med 1997; 126:133-136.

68. Hirsch I, Fuster V. Guide to anticoagulant therapy. Part 2: Oral anticoagulants. American
Heart Association. Circulation 1994; 89:1469-1480.

69. Ridker PM, Goldhaber SZ, Danielson A, Rosenberg Y, Evy CS, Deitcher SR, Cushman M,
Moll S, Kessler CM, Elliott CG, Paulson R, Wong T, Bauer KA, Schwartz BA, Miletich IP,
Bounameaux H, Glinn RJ. Long-term, low-intensity warfarin therapy for the prevention of
recurrent venous thromboembolism. N Engl I Med 2003; 348:1425-1434. E pub 2003 Feb 24.
www.nengm.org.

70. Albers GW, Sherman DG, Gress DR, Paulseth IE, Petersen P. Stroke prevention in non-
valvular atrial fibrillation: a review of prospective randomized trials. Ann Neurol 1991; 30:511-
518.

71. Flinn WR, Arohrer MI, Yao IS, McCarthy WI III, Fahey VA, Bergan S3. Improved long-
term patency of infragenicular polytetraflouroethylene grafts. I Vase Surg 1998; 7:685-
690.

72. lohnson WC, Williford WO, and members of the Department of Veteran's Affairs Coop-
erative Study #362. Benefits, morbidity, and mortality associated with long-term administra-
tion of oral anticoagulant therapy to patients with peripheral arterial bypass procedures: a
prospective randomized study. I Vase Surg 2002; 35:413-421.

73. Hull R, Hirsch I, lay R, Carter C, England C, Gent M, Turpie AG, McLoughlin D, Dodd P,
Thomas M, Raskob G, Ockelford P. Different intensities of oral anticoagulant therapy in the
treatment of proximal-vein thrombosis. N Engl I Med 1982; 307:1676-1681.

74. Yasaka M, Sakata T, Minematsu K, Naritomi H. Correction of INR by prothrombin com-
plex concentrate and vitamin K in patients with warfarin related hemorrhagic complication.
Thromb Res 2002; 108:25-30.

75. Chan YC, Valenti D, Mansfield AO, Stansby G. Warfarin induced skin necrosis. Br I Surg
2000; 87:266-272.

76. Alving BM, Strickler MP, Knight RD, Barr CF, Berenberg IL, Peck CC. Hereditary warfarin
resistance. Arch Intern Med 1985; 145:499-501.

77. Mannucci PM. Genetic control of anticoagulation. Lancet 1999; 353:688-689.

78. Kaplan KL, Francis CW. Direct thrombin inhibitors. Semin Hematol 2002; 39:187-196.

79. Greinacher A, Volpel H, lanssens U, Hach-Wunderle V, Kemkes-Matthes B, Eichler P,
Mueller- Velten HG, Potzsch B, for the HIT Investigators Group. Recombinant hirudin (lepa-
rudin) provide safe and effective anticoagulation in patients with heparin-induced thrombo-
cytopenia — a prospective study. Circulation 1999; 99:73-80.

80. Potzsch B, Madlener K, Seelig C, Reiss CF, Greinacher A, Muller-Berghaus G. Monitoring of g
small r-hirudin anticoagulation during cardiopulmonary bypass, assessment of the whole g
blood ecarin clotting time. Thromb Haemost 1997; 77:920-925. js

8 1 . Liu H, Fleming N W, Moore PG Anticoagulation for patients with heparin-induced thrombo- 2
cytopenia using recombinant hirudin during cardiopulmonary bypass. I Clin Anesth 2002; **
14:452-455. ■§

82. Greinacher A, Eichler P, Lubenow N, Kwasny H, Luz M. Heparin-induced thrombocyto- jjj
penia with thromboembolic complications: meta-analysis of 2 prospective trials to assess the q
value of parental treatment with leparudin and its therapeutic aPTT range. Blood 2000; 96: |
846-851. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 207

83. Robson R. The use of bivalirudin in patients with renal impairment. J Invas Cardiol 2000;
12(suppl F):33-36.

84. Bittl JA, Storny J, Brinker JA, Ahmed WH, Meckel CR, Chaitman BR, Maraganore J,
Deutsch E, Adelman B, for the Hirulog Angioplasty Study Investigators. Treatment with
bivalirudin (hirulog) as compared with heparin during coronary angioplasty for unstable or
postinfarction angina. N Engl J Med 1995; 333:764-769.

85. Bush L. Argatroban, a selective potent thrombin inhibitor. Cardiovasc Drug Rev 1991; 9:247-
263.

86. Lewis BE, Wallis DE, Berkowitz SD, Matthai WH, Fareed J, Walenga JM, Bartholomew J,
Sham R, Learner RG, Ziegler ZR, Rustagi PK, Jang IK, Rifkin SD, Moran J, Hursting MJ,
Kelton JG; for the ARG-9 1 1 Study Investigators. Argatroban anticoagulant therapy in patients
with heparin-induced thrombocytopenia. Circulation 2001; 103:1838-1843.

87. Lewis BE, Matthai WH, Cohen M, Moses JW, Hursting MJ, Leya F, for the ARG-216/310/
311 Study Investigators. Argatroban anticoagulation during percutaneous coronary
intervention in patients with heparin-induced thrombocytopenia. Cathet Cardiovasc Intervent
2002; 57:177-184.

88. Lubenow N, Selleng S, Wollert HG, Eichler P, Mullejans B, Greinacher A. Heparin-induced
thrombocytopenia and pulmonary bypass: perioperative argatroban use. Ann Thorac Surg
2003; 75:577-579.

89. Wahlander K, Lapidus L, Olsson CG, Thuresson A, Eriksson UG, Larson G, Eriksson H.
Pharmacokinetics, pharmacodynamics and clinical effects of the oral direct thrombin inhibitor
ximelagatran in acute treatment of patients with pulmonary embolism and deep vein
thrombosis. Thromb Res 2002; 107:93-99.

90. Eriksson H, Eriksson UG, Frison L, Hansson PO, Held P, Holmstrom M, Hagg A, Jonsson T,
Lapidus L, Leijd B, Stockelberg D, Safwenberg U, Taghavi A, Thorsen M. Pharmacokinetics
and pharmacodynamics of melagatran, a novel synthetic LMW thrombin inhibitor, in patients
with acute DVT. Thromb Haemost Mar 1999; 81:358-363.

91. Eriksson BI, Bergqvist D, Kalebo P, Dahl OE, Lindbratt S, Bylock A, Frison L, Eriksson
UG, Welin L, Gustafsson D. Melagatran for Thrombin inhibition in orthopaedic surgery.
Ximelagatran and melagatran compared with dalteparin for prevention of venous thrombo-
embolism after total hip or knee replacement: the METHRO II randomised trial. Lancet 2002;
360:1441-1447.

92. Walenga JM, Jeske WP, Samama MM, Frapaise FX, Bick RL, Fareed J. Fondaparinux: a
synthetic heparin pentasaccharide as a new antithrombotic agent. Expert Opin Invest Drugs
2002; 11:397-407.

93. Turpie AG, Bauer KA, Eriksson BI, Lassen MR. Fondaparinux vs enoxaparin for the pre-
vention of venous thromboembolism in major orthopedic surgery: a meta-analysis of 4 ran-
domized double-blind studies. Arch Intern Med 2002; 162:1833-1840.

94. Kwong LM, Muntz JE. Thromboprophylaxis dosing: the relationship between timing of first
administration, efficacy, and safety. Am J Orthop 2002; 31(suppl):16-20.

95. Bijsterveld NR, Moons AH, Boekholdt SM, van Aken BE, Fennema H, Peters RJ, Meijers
JC, Buller HR, Levi M. Ability of recombinant factor Vila to reverse the anticoagulant

effect of the pentasaccharide fondaparinux in healthy volunteers. Circulation 2002; 106: j>

2550-2554. |

96. Cipollone F, Patrignani P, Greco A, Panara MR, Padovano R, Cuccurullo F, Patrono C, s
Rebuzzi AG, Liuzzo G, Quaranta G, Maseri A. Differential suppression of thromboxane bio- 2
synthesis by indopufen and aspirin in patients with unstable angina. Circulation 1997; 96: 1 109- ^
1116. ■§

97. Liu Y, Shakur Y, Yoshitake M, Kambayashi Ji J. Cilostazol (pletal): a dual inhibitor of cyclic ^
nucleotide phosphodiesterase type 3 and adenosine uptake. Cardiovasc Drug Rev 2001; 19:369- q
386. |

98. Woo SK, Kang WK, Kwon KI. Pharmacokinetic and pharmacodynamic modeling of the 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

208 HOCH

antiplatelet and cardiovascular effects of cilostazol in healthy humans. Clin Pharmacol Ther
2002; 71:246-252.
99. Quinn MJ, Fitzgerald DJ. Ticlopidine and clopidogrel. Circulation 1999; 100:1667-1672.

100. Topol EJ, Byzova TV, Plow EF. Platelet GPIIb-IIIa blockers. Lancet 1999; 353:227-231.

101. Tcheng JE, Ellis SG, George BS, Kereiakes DJ, Kleiman NS, Talley JD, Wang AL, Weisman
HF, Califf RM, Topol EJ. Pharmacodynamics of chimeric glycoprotein Ilb/IIIa integrin
antiplatelet antibody Fab 7E3 in high-risk coronary angioplasty. Circulation 1994; 90:1757-
1764.

102. Kereiakes DJ, Kleiman NS, Ambrose J, Cohen M, Rodriguez S, Palabrica T, Herrmann HC,
Sutton JM, Weaver WD, McKee DB, Fitzpatrick V, Sax FL. Randomized, double-blind,
placebo-controlled dose-ranging study of tirofiban (MK-383) platelet Ilb/IIIa blockade in
high risk patients undergoing coronary angioplasty. J Am Coll Cardiol 1996; 27:536-542.

103. O'Shea JC, Buller CE, Cantor WJ, Chandler AB, Cohen EA, Cohen DJ, Gilchrist IC,
Kleiman NS, Labinaz M, Madan M, Hafley GE, Califf RM, Kitt MM, Strony J, Tcheng JE.
ESPRIT Investigators. Long-term efficacy of platelet glycoprotein Ilb/IIIa integrin blockade
with eptifibatide in coronary stent intervention. JAMA 2002; 287:618-621.

104. Patrono C, Coller B, Dalen JE, FitzGerald GA, Fuster V, Gent M, Hirsh J, Roth G Platelet-
active drugs: The relationships among dose, effectiveness, and side effects. Chest 2001; 119
(suppl 1): 39-63.

105. Antithrombotic Trialists' Collaboration. Collaborative meta-analysis of randomised trials of
antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk
patients. Br Med J 2002; 324:71-86.

106. Collaborative overview of randomised trials of antiplatelet therapy: I. Prevention of death,
myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of
patients. Antiplatelet Trialists' Collaboration. Br Med J 1994; 308:81-106.

107. Janzon L. The STIMS trial: the ticlopidine experience and its clinical applications. Swedish
Ticlopidine Multicenter Study. Vase Med 1996; 1:141-143.

108. CAPRIE Steering Committee. A randomised, blinded, trial of clopidogrel versus aspirin in
patients at risk of ischaemic events (CAPRIE). Lancet 1996; 348:1329-1339.

109. Berge E, Sandercock P. Anticoagulants versus antiplatelet agents for acute ischaemic stroke.
Cochrane Database Syst Rev 2002; 4:CD003242.

110. The Abciximab in Ischemic Stroke Investigators. Abciximab in acute ischemic stroke: a ran-
domized, double-blind, placebo-controlled, dose-escalation study. Stroke 2000; 31:601-609.

111. Antiplatelet Trialists' Collaboration. Collaborative overview of randomised trials of anti-
platelet therapy: II. Maintenance of vascular graft or arterial patency by antiplatelet therapy.
Br Med J 1994; 308:159-168.

1 12. Becquemin JP. Effect of ticlopidine on the long-term patency of saphenous-vein bypass grafts
in the legs. Etude de la ticlopidine apres pontage femoro-poplite and the association
Universitaire de Recherche en Chirurgie. N Engl J Med 1997; 337:1726-1731.

113. Yusuf S, Zhao F, Mehta SR, Chrolavicius S, Tognoni G, Fox KK. The Clopidogrel in
Unstable Angina to Prevent Recurrent Events Trial Investigators. Effects of clopidogrel in
addition to aspirin in patients with acute coronary syndromes without ST-segment elevation.

N Engl J Med 2001; 345:494-502. Erratum in: N Engl J Med 2001; 345:1506, 1716. 1

114. Steinhubl SR, Berger PB, Mann JT, Fry ET, DeLago A, Wilmer C, Topol EJ, CREDO |
Investigators. Clopidogrel for the reduction of events during observation. Early and sustained js
dual oral antiplatelet therapy following percutaneous coronary intervention: a randomized 2
controlled trial. JAMA 2002; 288:241 1-2420. Erratum in JAMA 2003; 289:987. ^

115. Kamishirado H, Inoue T, Mizoguchi K, Uchida T, Nakata T, Sakuma M, Takayanagi K, «
Morooka S. Randomized comparison of cilostazol versus ticlopidine hydrochloride for anti- ^
platelet therapy after coronary stent implantation for prevention of late restenosis. Am Heart q
J 2002; 144:303-308. |

116. Karvouni E, Katritsis DG, Ioannidis JP. Intravenous glycoprotein Ilb/IIIa receptor antag- 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AND FAILURES OF THERAPY 209

onists reduce mortality after percutaneous coronary interventions. J Am Coll Cardiol 2003;
41:26-32.

117. Qureshi AI, Suri MF, Ali Z, Kim SH, Lanzino G, Fessler RD, Ringer AJ, Guterman LR,
Hopkins LN. Carotid angioplasty and stent placement: a prospective analysis of perioperative
complications and impact of intravenously administered abciximab. Neurosurgery 2002;
50:466-473. Discussion 473-475.

118. Moser M, Bertram U, Peter K, Bode C, Ruef J. Abciximab, eptifibatide, and tirofiban exhibit:
dose-dependent potencies to dissolve platelet aggregates. J Cardiovasc Pharmacol 2003;
41:586-592.

119. Hankey GJ, Sudlow CL, Dunbabin DW. Thienopyridine derivatives (ticlopidine, clopidogrel)
versus aspirin for preventing stroke and other serious vascular events in high vascular risk
patients. Cochrane Database Syst Rev 2000; 2:CD001246.

120. Hongo RH, Ley J, Dick SE, Yee RR. The effect of clopidogrel in combination with aspirin
when given before coronary artery bypass grafting. J Am Coll Cardiol 2002; 40:231-237.

121. Use of a monoclonal antibody directed against the platelet glycoprotein Ilb/IIIa receptor in
high-risk coronary angioplasty. The EPIC Investigation. N Engl J Med 1994; 330:956-961.

122. Garcia Rodriguez LA, Hernandez-Diaz S, de Abajo FJ. Association between aspirin and
upper gastrointestinal complications: systematic review of epidemiologic studies. Br J Clin
Pharmacol 2001; 52:563-571.

123. Nara W, Ashley I, Rosner F. Thrombotic thrombocytopenic purpura associated with
clopidogrel administration: case report and brief review. Am J Med Sci 2001; 322:170-172.

124. Rezkalla SH, Hayes JJ, Curtis BR, Aster RH. Eptifibatide-induced acute profound thrombo-
cytopenia presenting as refractory hypotension. Catheter Cardiovasc Intervent 2003; 58:76-79.

125. Pratt CM. Analysis of the cilostazol safety database. Am J Cardiol 2001; 87:28D-33D.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

Gastrointestinal and Visceral Ischemic Complications
of Aortic Reconstruction

Daniel J. Reddy and Hector M. Dourron*

Henry Ford Hospital, Detroit, Michigan, U.S.A.

Gastrointestinal complications following aortic reconstructive surgery are well recognized
and are associated with a potentially high mortality rate. Contemporary practice, with ad-
vances in pre- and postoperative care as well as improvements in operative technique, have
contributed to a progressive decline in morbidity and mortality after aortic reconstruction
when compared to the past three decades. Nonetheless, maintenance of a low rate of mor-
bidity challenges the vascular surgeon's ability to identify and adopt new strategies to limit
perioperative complications. Although an array of gastrointestinal complications (Table 1)
have been reported following aortic reconstruction, severe morbidity leading to prolonged
hospitalization and death can often be avoided by prompt recognition and expert man-
agement. An understanding of the setting in which these complications occur and the
influence of patient comorbidities are key. As many as 50% of patients who have gas-
trointestinal complications require operative intervention, with reported mortality rates
ranging from 16-67% (1-3).

Although a number of risk factors for gastrointestinal complications have been iden-
tified, splanchnic bed arterial hypoperfusion appears to be the most common final pathol-
ogical mechanism implicated in most cases. The inferior mesenteric artery (IMA) is
commonly divided during aortic reconstruction. The loss of the IMA, reduced hypogastric
arterial flow, or hindgut atheroembolization have been implicated in the pathogenesis of g

ischemic colitis following aortic reconstruction (3). Moreover, atheroemboli may affect the §

small bowel, kidney, and spinal cord as well as the hindgut. Additional factors that in- j§

crease the risk of ischemic colitis are listed in Table 2. Jfj

Mounting evidence suggests that intestinal ischemia is important in the development j

of irreversible shock and multiple organ system failure (4). To prevent these complications, "jl

! Current affiliation: Vascular Surgical Associates, P.C., Austell, Georgia, U.S.A.

211 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

212

REDDY and DOURRON

Table 1 Gastrointestinal Complications Following
Aortic Reconstruction

Paralytic ileus

Upper gastrointestinal bleed

Gastritis

Duodenal ulcer

Gastric ulcer

Clostridium difficile enterocolitis

Acute cholecystitis

Mechanical obstruction

Chylous ascites

Pancreatitis

Colon ischemia

Small bowel ischemia

Visceral ischemia

the vascular surgeon must provide for adequate colonic and pelvic arterial blood supply
following aortic reconstruction, along with the many additional requirements of these com-
plex procedures.

I. CLASSIFICATION

Gastrointestinal complications of a functional and ordinarily self-limiting nature occur
frequently following aortic reconstruction. Paralytic ileus of the stomach and colon lasting
36-48 h following transperitoneal repair of the aorta is common and seldomly leads to
major morbidity or prolonged hospitalization. Bowel manipulation, extensive lysis of
adhesions, and hematoma formation may contribute to this problem; particularly in
patients with the autonomic neuropathic comorbidity of diabetes mellitus. Postoperative
pancreatitis is a less common but potentially more severe gastrointestinal complication
following aortic reconstruction. Elevated amylase and lipase levels support the diagnosis,
although clinical manifestations of pancreatitis are unusual. Vigorous retraction and more
proximal dissection, particularly at the paraceliac level, are considered to play a role in the
pathogenesis. The most serious gastrointestinal complication following aortic reconstruc-
tion remains ischemic colitis.

The foregut, midgut, hindgut, and pelvis are endowed with a rich collateral arterial
bed that, under normal circumstances, protects both solid and hollow viscera from

Table 2 Predisposing Factor for the Development of
Ischemic Colitis Following Aortic Reconstruction

Improper IMA ligation

Loss of IMA-hypogastric blood flow

Ruptured aneurysm

Perioperative hypotension

Retractor trauma

Inadequate development of collaterals

IMA to SMA flow in the meandering mesenteric artery

i
2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF AORTIC RECONSTRUCTION 213

ischemia when axial perfusion is interrupted by disease or operative manipulation.
However, patients with aneurysmal or occlusive processes often have diseased collaterals
or lack a complete collateral pattern, putting them at greater risk of visceral ischemia.
Although the colon or small bowel may be affected, the colon is far more commonly
involved. Colonic ischemia following aortic reconstruction is typically secondary to
arterial occlusion or atheroembolism. Venous compromise of the colon following aortic
reconstruction has been reported but seldom encountered without low flow or hemato-
logical disease, particularly in the setting of intra-abdominal sepsis. Colonic ischemia is
more commonly associated with aneurysmal rather than arterial occlusive disease owing to
the richer collateral beds that develop over time with chronic occlusive disease (5).

Greenwald et al. have provided a practical classification of colonic ischemia (CI) that
stratifies specific conditions resulting from ischemic injury to the colon as reversible or
irreversible (6). These classes can then be subcategorized further as (a) reversible ischemic
colonopathy (submucosal or intramural hemorrhage), (b) reversible or transient ischemic
colitis, (c) chronic ulcerative ischemic colitis, (d) ischemic colonic stricture, (e) colonic
gangrene, and (f) fulminant universal ischemic colitis. Subcategories a and b are limited to
the mucosa. Subcategory c extends to the muscularis and is not reversible. Subcategories e
and f are transmural and result in perforation. More than 60% of reported cases involve
transmural ischemia. The most common anatomical site for all types of colonic ischemia
following aortic reconstruction is the sigmoid colon after inferior mesenteric artery liga-
tion (7). No doubt other factors can play a contributing role in the pathogenesis (Table 2).

II. ANATOMY AND PATHOPHYSIOLOGY

The inferior mesenteric artery (IMA) and its branches serve to bridge the midgut, hindgut,
and pelvis via collateral communications with the superior mesenteric artery (SMA) and
the hypogastric arterial circulation. Among these three circuits, the IMA is the main axial
blood supply to the left colon. The SMA and IMA circuits communicate via the me-
andering artery and the marginal artery of Drummond, both of which originate from the
left branch of the middle colic artery and terminate in the left colic artery or IMA. The
meandering artery can typically be found within the mesenteric pedicle and is at risk of
injury during retraction for exposure. The marginal artery of Drummond is rather
constant along the transverse and descending colons. However, in 5% of individuals, it
is lacking in the region of the ascending colon; in 20%, it is absent in the sigmoid colon;
and it exists with even greater inconsistency at the rectosigmoid junction (8). Finally, the
hypogastric arteries, via the lower rectal branches, provide a bridge between the systemic
and visceral circulations. This can further collateralize the left colon by way of the superior
rectal branch of the IMA.

The incision and operative approach selected for aortic reconstruction is individu- ■g

alized for each patient. Most authors describe the approach to the infrarenal aorta via a &

transperitoneal midline incision, although reports are accumulating detailing certain pa- a

tient factors that favor a retroperitoneal approach to reduce gastrointestinal complications c

(9). In a 1987 nonrandomized comparison of the retroperitoneal and transabdominal <j

approaches from Sicard et al., 6% of patients with transperitoneal repair were found to -9

have a prolonged ileus (more than 96 h), while none were noted in the retroperitoneal 4j

group (10). In a later prospective randomized trial reported by these same authors, 10% of 2

patients with aortic repair by the transabdominal approach experienced prolonged ileus |

and 6% had small bowel obstructions, of which more than 50% required operative in- @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

214 REDDY and DOURRON

tervention (11). Furthermore, no patients in the retroperitoneal approach group required
reoperation for gastrointestinal complications.

Colonic ischemia becomes clinically evident in a variety of stages of severity. In the
mildest cases, mucosal ischemia may lead to an inflammatory reaction of the submucosa,
with mucosal sloughing and eventual complete healing. A moderate ischemic injury can
extend to the muscularis mucosa, and these cases generally heal with some degree of
stricture. The most severe cases of ischemia involve transmural necrosis with perforation
of the bowel wall. The mortality rate of colonic infarction after aortic reconstruction has
been reported as high as 80-100% (12). Therefore knowledge of and provision for the
anatomical collateralization pathways supplying the viscera is of vital importance during
aortic reconstruction. Avoidance of injury to collateral pathways or reimplantation or
bypass of vital arterial branches is key to the prevention of intestinal ischemia and low-
ering of morbidity following aortic reconstructive surgery.

III. CLINICAL MANIFESTATIONS AND DIAGNOSIS

A high index of suspicion will facilitate early diagnosis of postoperative colonic ischemia
and avoid potentially fatal complications. Clinical evidence of ischemic colitis is seen in
approximately 2% of elective infrarenal aortic reconstructions (13-15) and in up to 32%
of patients who survive repair of a ruptured aortic aneurysm (16). Prospective studies with
colonoscopy suggest ischemic mucosal changes in 7-35% of patients undergoing elective
aortic procedures (17) and up to 60% of patients with ruptured aortic aneurysms (18).
Clearly, not all of these patients manifest clinical signs of colonic ischemia; therefore it is
the vascular surgeon's challenge to identify the subgroup of patients with insufficient col-
lateral circulation to maintain colon viability. The classic clinical presentation for colonic
ischemia is bloody diarrhea in the early postoperative period. Nonetheless, only 29-36%
of patients with documented ischemic colitis will exhibit this feature, making it an
insensitive clinical predictor (19,20). Despite its relatively low incidence, bloody diarrhea,
when present, is a compelling clinical sign that often predicts transmural colonic ischemia
(16). Other clinical signs suggestive of colonic ischemia are listed in Table 3. Clinical sus-
picion is confirmed by colonoscopy, which should be performed at the bedside of any
patient suspected of ischemic colitis. Repeated studies over a 1- to 2-day interval can help
clarify resolution or progression of this disease process. Passage of the colonoscope to 40
cm from the anal verge is usually sufficient to detect ischemic colitis in up to 95% of pa-
tients (16), as isolated ischemia of the right or transverse colon with no left colonic

Table 3 Clinical Signs of Ischemic Colitis

Fever
Anuria

Hypotension <90 mmHg -a

Abdominal pain <j

Abdominal distention S

Elevated WBC count thrombocytopenia jj

Postoperative fluid requirements >5L q

Packed red blood cells >6U fj

Lactic acidemia s

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF AORTIC RECONSTRUCTION 215

involvement is rare (18,20). Early mucosal changes seen on colonoscopy include circum-
ferential petechial hemorrhage and edema. More advanced cases may demonstrate
pseudomembranes, erosions, and ulcers. But even in these cases, in the absence of
peritoneal signs, conservative management such as bowel rest, fluid administration, and
parenteral nutrition may suffice, with the most common sequela being colonic stricture.
Beyond these changes, however, the colon will appear yellowish-green, necrotic, and
noncontractile. At this point colonic perforation should be considered imminent and
operative intervention is advised.

Another method for predicting colonic ischemia is transluminal pH (pH t ) mucosal
monitoring (21). The theory behind pH t is that anaerobic metabolism caused by
insufficient oxygen delivery to the gut can be detected by a low pH of the gastric or
colonic mucosa. Studies have shown that sigmoid acidosis (pHt below 7.1) signals an early
warning and, if not corrected within 2 h, is predictive of major morbidity (22). Barium
contrast studies have been used to document late sequelae of ischemic colitis but are of no
use for diagnosis in the immediate postoperative period. In summary, frequent reexami-
nation of the abdomen, repeated colonoscopy, and monitoring of hemodynamic param-
eters are required when ischemic colitis is suspected. We have evaluated other modalities,
such as fluorescein injection with Wood's lamp examination, pulse oximetry, and Doppler
ultrasound in controlled experimental studies of small bowel ischemia (23). However; none
of these modalities have been confirmed by clinical trials of hindgut ischemia.

IV. ISCHEMIC COLITIS FOLLOWING STENT-GRAFT REPAIR
OF ABDOMINAL AORTIC ANEURYSMS

Several large series have reported no cases of ischemic colitis following endovascular repair
of abdominal aortic aneurysms (AAAs) (24-26). Sandison et al. reported one case of
sigmoid necrosis following endovascular repair of an infrarenal aortic aneurysm and
attributed the complication to embolization of a patent internal iliac artery (27). Jaeger et
al. reported a second case of ischemic colitis following endorepair, which, on postmortem,
was attributed to cholesterol emboli (28). Kalliafas et al. reported a partial-thickness colonic
ischemia after stent-graft placement with inadvertent exclusion of an internal iliac artery
during the procedure (29). Kalliafas identified 12 additional patients in his series who had
internal iliac exclusion but did not develop colonic ischemia. Miahe et al. reported two
episodes colonic ischemia after endovascular aortic repair (30). The first patient was found
to have mild mucosal changes after exclusion of a patent inferior mesenteric artery with
preservation of both internal iliac arteries. The second patient, who had previously under-
gone a Whipple procedure and pelvic radiation therapy for prostate cancer, required
colectomy. Finally, Marin et al. reported a case of mild ischemic colitis following stent
graft repair of an iliac aneurysm in which the contralateral iliac artery was occluded during 13

the procedure (31). In a recent study, Rhee et al. prospectively followed 49 patients who |

underwent iliac artery coverage with or without coil embolization of one or both internal J

iliac arteries during aortic endovascular reconstruction (32). They reported that internal ;e

iliac arteries can be covered by extending the graft across the internal iliac orifice into the **

external iliac artery with minimal adverse consequences in patients who have common iliac &

arteries unsuited for deployment of endografts. They further recommend that if bilateral M

internal iliac arteries orificial openings are to be covered or occluded, one of these should be 2

revascularized if possible. From these reports the calculated incidence of ischemic colitis I

following aortic stent graft repair is 1.4%. This low reported incidence of colonic ischemia ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

216

REDDY and DOURRON

following stent graft placement awaits confirmation, as the inferior mesenteric artery is
found to be patent in a large number of elective open aortic repair, procedures, and stent
graft repair occludes direct flow into the inferior mesenteric artery. Risk factors for the
development of colonic ischemia after aortic stent graft placement include significant
superior mesenteric artery occlusive disease and inadequate pelvic collateral perfusion.
Additionally, as endovascular repair of the aorta becomes more common, further risk
factors for colonic ischemia will likely be identified. As experience is gained, the magnitude
of the potential problem will come into sharper focus and future clinical trials should prompt
the development of practice guidelines.

V. PREVENTION

Preoperative evaluation of the blood supply to the colon as well as intraoperative
assessment of colonic perfusion by Doppler ultrasound or pH t may help to identify those
patients who would benefit from mesenteric reconstruction. In order to minimize the
morbidity and mortality associated with ischemic colitis, a high index of suspicion will
assist in making an earlier diagnosis. Proper ligation of the IMA, avoidance of perioper-
ative hypotension, careful placement of retractors, and finally reimplatation of an IMA

i

2

I

■c

Figure 1 Mesenteric angiogram. Retrograde filling of superior mesenteric artery through a
meandering mesenteric artery from the inferior mesenteric artery. (SMA, superior mesenteric artery;
MA, meandering mesenteric artery.)

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

u

COMPLICATIONS OF AORTIC RECONSTRUCTION 217

with inadequate collateral development should greatly reduce the incidence of this dev-
astating complication.

VI. ARTERIOGRAPHY

Aortography has been used for many years in the preoperative evaluation of patients
undergoing aortic reconstruction. It is particularly indicated in patients with a history or
physical signs of intestinal angina. Visualization of a meandering mesenteric artery may
alert the clinician to significant superior or inferior mesenteric arterial disease (Fig. 1).
Reversal of flow in the meandering mesenteric artery from the inferior to the superior
mesenteric artery warrants further study by way of a lateral aortic view to document su-
perior mesenteric artery stenosis as, these patients are at considerable without associated
superior mesenteric artery reconstruction or reimplatation of the inferior mesenteric ar-
tery, these patients are at considerable risk of small bowel ischemia.

VII. OPERATIVE TECHNIQUES AND TREATMENT

One common factor in most cases of ischemic colitis is ligation of the inferior mesenteric
artery. Inappropriate or improper ligation may be the most important pitfall for surgeons
to avoid. The inferior mesenteric artery begins to branch shortly after its origin from the
aorta. A proper ligation involves dissection of the artery to within 2 cm of its origin and
suture ligation at this point, thus avoiding interruption of important collaterals. An
alternative and perhaps better technique is IMA ligation at its orifice from within the
aneurysmal sac. Many diseased aortas contain either intramural thrombus or unstable
mural plague, which can easily be dislodged and embolize. Broadcast of this debris can
lead to ischemic events in multiple organs and is avoided through judicious handling
during dissection. In dealing with a ruptured aortic aneurysm, the surgeon faces additional
challenges. Often the patient is found to have a large hematoma within the mesentery. This
should not be evacuated, as doing so can cause disruption of vital collateral pathways.
Interruption of collateral pathways may also occur with overvigorous retraction. Ante-
grade restoration of flow to the hypogastric arteries or through other pathways preserves
the pelvic collateral circulation and minimizes the development of ischemic colitis. Clinical
prediction of the viability of bowel is unreliable; the surgeon is advised to rely on more
objective data to determine whether the bowel has adequate collateral perfusion without
restoration of the inferior mesenteric artery. Several tests of bowel perfusion have been
described, including inferior mesenteric artery stump pressure, radioisotope scanning, flu-
orescein dye, colonic serosal photoplethysmography, Doppler interrogation in the colonic
mesentery, and tonometric measurement of colonic mucosal pH. However, these methods
are often time-consuming, they require experience, and are some are currently exper- •§

imental and therefore impractical in the clinical setting. Additionally, most techniques g

determine adequacy of colonic collateral blood flow in the operating room when overall a

perfusion is optimized and do not predict maintenance of flow in the postoperative period, c

when inadequate resuscitation can often a cause hypotension and induce ischemia. The <j

advantages of Doppler interrogation in the colonic mesentery and tonometric measure- >9

ment of colonic mucosal pH are that the Doppler is available and easy to use and that the 4j

tonometric measurement can be continued well into the postoperative period. 2

Hypotension and hypoperfusion in the operative as well as postoperative period are |

major causes of morbidity following aortic reconstruction. The colonic collateral circu- @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

218 REDDY and DOURRON

lation is susceptible to periods of hypotension once the inferior mesenteric artery has been
ligated. Monitoring of the patient's overall fluid status using central venous and pulmo-
nary wedge pressures is necessary in critically ill and high-risk patients.

An alternative approach to the prevention of colonic ischemia after aortic recon-
struction is to reimplant all patent inferior mesenteric arteries. This methodology has been
practiced by the University of Florida, where 151 patients undergoing aortic reconstruc-
tion from 1986-1989 had patent mesenteric arteries reimplanted. No patients suffered
colonic ischemia requiring operative intervention, and there were no deaths attributable to
colonic ischemia (33).

VIM. SUMMARY

Gastrointestinal complications following standard or endovascular techniques of aortic
reconstruction continue to be an important problem. In view of the potential severity of
these complications following aortic reconstruction, early identification of patients at risk
is important. The early diagnosis of ischemic colitis hinges on a high index of suspicion
and prompt action to optimize the patient's hemodynamic status. Once the diagnosis of
ischemic colitis is entertained, it is the surgeon's responsibility to confirm this either
through prompt fiberoptic colonoscopy or tonometric measurement of colonic mucosal
pH. For patients with mild to moderate ischemic changes, conservative management with
fluid resuscitation and close monitoring is appropriate. However, in more severe cases,
aggressive management, including early colonic resection, is the best method by which to
lower the mortality due to these devastating complications.

REFERENCES

1 . Christenson JT, Schmuziger M, Maurice J, Simonet F, Velebit V. Gastrointestinal complications
after coronary artery bypass grafting. J Thorac Cardiovasc Surg 1994; 108:899-906.

2. Mercado PD, Farid H, O'Connell TX, Sintek C, Pfeffer T, Khonsari S. Gastrointestinal com-
plications associated with cardiopulmonary bypass proceduces. Am Surg 1994; 60:789-792.

3. Ernst CB. Prevention of intestinal ischemia following abdominal aortic reconstruction. Surgery
1983; 96:102-106.

4. Antonsson J, Fiddian-Green RG. The role of the gut in shock and multiple system organ failure.
Eur J Surg 1991; 157:3-12.

5. Ernst CB. Intestinal ischemia following abdominal aortic reconstruction. In: Bernard VM,
Town IB, eds. Complications in Vascular Surgery. New York: Grune & Stratton, 1985:325-350.

6. Greenwald DA, Brandt LI, Reinus IF. Ischemic bowel disease in the elderly. Gastroenterol
Clin North Am 2001; 30(2):445-473.

7. Farkas IC, Calvo-Verjat N, Laurain C, et al. Acute colorectal ischemia after aortic surgery: j>
Pathophysiology and prognostic criteria. Ann Vase Surg 1992; 6:11. g

8. Steward IA, Rankin FW. Blood supply of the large intestine:Its surgical considerations. Arch js
Surg 1933; 26:843. Jf

9. Shepard AD, Tollefson DF, Reddy DJ, Evans IR, Elliott IP, Smith RF, Ernst CB. Left flank ^
retroperitoneal exposure: A technical aid to complex aortic reconstruction. I Vase Surg 1991; «
14:283-291. J

10. Sicard GA, Freeman MB, VanderWoude IC, Anderson CB. Comparison between the trans- q

abdominal and retroperitoneal approach for reconstruction of the infrarenal abdominal aorta. |

I Vase Surg 1987; 5:19-27. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF AORTIC RECONSTRUCTION 219

1 1 . Sicard GA, Reilly JM, Rubin BG, Thompson RW, Allen BT, Flye MW, Schechtman KB, Young-
Beyer P, Weiss C, Anderson CB. Transabdominal versus retroperitoneal incision for abdominal
aortic surgery: Report of a prospective randomized trial. J Vase Surg 1995; 21:174-183.

12. Kim MW, Hurdahl SA, Dang CR, McNamara JJ, Staehly CJ, Whelan TJ. Ischemic colitis after
aortic aneurysmectomy. Am J Surg 1983; 145:392-394.

13. Johnson WC, Nasbeth DC. Visceral infaction following aortic surgery. Ann Surg 1963; 86:65-73.

14. Young JR, Humphries AW, deWolf VG, LeFevre FA. Complications of abdominal aortic
surgery: II. Intestinal ischemia. Arch Surg 1963; 86:51-59.

15. Papadopoulos CD, Mancini HW, Marino WM Jr. Ischemic necrosis of the colon following
aortic aneurysmectomy. J Cardiovasc Surg 1974; 15:494-500.

16. Bandyk DF, Florence MG, Johansen KH. Colon ischemia accompanying ruptured abdominal
aortic aneurysm. J Surg Res 1981; 30:297-303.

17. Fry PD. Colonic ischemia after aortic reconstruction. Can J Surg 1988; 31:162-164.

18. Hagihara PF, Ernst C, Griffen WO. Incidence of ischemic colitis following abdominal aortic
reconstruction. Surg Gynecol Obstet 1979; 149:571-573.

19. Longo WE, Lee TC, Barnett MG, et al. Ischemic colitis complicating abdominal aortic aneurysm
surgery in the U.S. veteran. J Surg Res 1996; 60:351-354.

20. Bjorck M, Bergqvist D, Troeng T. Incidence and clinical presentation of bowel ischemia after
aortoiliac surgery — 2930 operations from a population-based registry in Sweden. Eur J Vase
Endovasc Surg 1996; 12:139-144.

21. Schiedler MG, Cutler BS, Fiddian-Green RG. Sigmoid intramural pH for prediction of ischemic
colitis during aortic surgery. A comparison with risk factors and inferior mesenteric artery
pressures. Arch Surg 1987; 122:881-886.

22. Bjorck M, Hedberg B. Early detection of major complications after abdominal aortic surgery:
Predictive value of sigmoid colon and gastric intramucosal pH monitoring. Br J Surg 1994;
81:25-30.

23. Tollefson FJ, Wright DJ, Reddy DJ, Kintanar EB. Intraoperative determination of intestinal
viability by pulse oximetry. Ann Vase Surg 1995; 9:357-360.

24. Makaroun MS. The AnCure endografting system: An update. J Vase Surg 2000; 33(S): 129-134.

25. Matsumara J, Katzen BT, Hollier LH, et al. Update on the bifurcated Excluder endoprosthesis:
Phase I results. J Vase Surg 2001; 33:150-153.

26. Zarins CK, White RA, Hodgson KJ, et al. for the AneuRx Clinical Investigators. Endoleak as
a predictor of outcome following endovascular aneurysm repair: AneuRx multicenter clinical
trial. J Vase Surg 2000; 32:90-107.

27. Sandison AJ, Edmondson RA, Panayiotopoulos YP, et al. Fatal colonic ischemia after stent
graft for aortic aneurysm. Eur J Vase Endovasc Surg 1997; 13:219-220.

28. Jaeger HJ, Mathias KD, Gissler HM, et al. Rectum and sigmoid colon necrosis due to
cholesterol embolization after implantation of an aortic stent graft. J Vase Intervent Radiol
1999; 10:751-755.

29. Kalliafas S, Albertini JN, Macierewicz J, et al. Incidence and treatment of intraoperative tech-
nical problems during endovascular repair of complex abdominal aortic aneurysms. J Vase Surg
2000;31:1185-1192.

30. Mialhe C, Amicabile C, Becquemin JP. Endovascular treatment of infrarenal abdominal aneu- j>
rysm by the stentor system: Preliminary results of 79 cases. Stentor Retrospective Study Group. <S
J Vase Surg 1997; 26:199-209. |

31. Marin ML, Veith FJ, Lyon RT, et al. Transfemoral endovascular repair of iliac aneurysms. Am 2
J Surg 1995; 170:179-182. ^

32. Rhee RY, Muluk SC, Tzeng E, Carroll N, Makaroun MS. Can the internal iliac artery be «
safely covered during endovascular repair of abdominal aortic and iliac artery aneurysms? Ann jjj
Vase Surg 2002; 16:29-36. ' q

33. Burress Welborn M III, Seeger JM. Prevention and management of sigmoid and pelvic ischemia |
associated with aortic surgery. Sem Vase Surg 2001; 14:255-265. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

10

Spinal Cord Ischemia

Alfio Carroccio and Nicholas J. Morrissey

Mount Sinai School of Medicine, New York, New York, U.S.A.

Larry H. Hollier

Louisiana State University Health Sciences Center School of Medicine,
New Orleans, Louisiana, U.S.A.

Spinal cord ischemia with a resulting neurological deficit remains a devastating compli-
cation following thoracoabdominal aortic surgery. It has been reported to occur in 0-40%
of patients undergoing aortic surgery, depending upon the type of pathology present as
well as circumstances present at the time of operation (1-8). Clinical presentation can
range from somatosensory loss to complete flaccid paralysis immediately following sur-
gery or delayed presentation up to several weeks postoperatively.

An appreciation of the anatomy as well as the mechanism of ischemic injury has led to
various strategies attempting to reduce the impact of the ischemia. Such measures include
intercostal artery reimplantation, cerebrospinal fluid drainage, hypothermia techniques,
distal perfusion techniques, pharmacotherapy, and mechanisms of ischemic precondition-
ing. Although these adjuncts have been evaluated in multiple human as well as animal
research protocols and appear promising, the problem has yet to be resolved. This chapter
focuses on determinants of spinal cord ischemia as well as the various measures utilized in
avoiding the resulting neurological deficit.

I. ANATOMY 1

The predominant blood supply to the spinal cord derives from the longitudinally oriented S

spinal arteries. The single anterior spinal artery supplies the gray matter as well as the =s

anterior white matter. The paired posterior spinal arteries supply the dorsal columns as J

well as the posterior white matter (9-11). Ischemia from interruption of the anterior spinal g

artery may result in paralysis, sphincter dysfunction, and decreased pain and temperature |j

sensation, while vibratory sensation and proprioreception may be affected by interruption "g

of the posterior circulation. 2

221

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

222

CARROCCIO et al.

Anterior Spinal Artery

Posterior Spinal Arteries

Figure 1 Spinal cord cross-sectional diagram demonstrating single anterior and double posterior
spinal arteries with contribution from segmental arteries.

The posterior spinal arteries are more continuous; they are supplied by 12 dorsal
medullary arteries (Fig. 1). The anterior spinal artery may have a more segmental or dis-
continuous path, with contribution from radicular arteries from the thoracic and lumbar
regions (11,12). The most prominent of the radicular arteries, the artery of Adamkiewicz,
can originate anywhere from T8 to L3; however, it is most often in the region of T9-T12
(13). Additional collateral blood supply is provided by the internal thoracic, long thoracic,
intercostal, scapular, and proximal radicular arteries (14). In the lower aortic region,
collateral blood flow from lumbar, iliolumbar, and lateral sacral arteries may prove critical
and therefore require preservation of the hypogastric circulation. The lower incidence of
neurological consequences following aortic coarctation surgery (0.41%) as compared to
traumatic aortic rupture (19%) is a good demonstration of the protective benefit of
collaterals in the more chronic disease states (15).

II. PHYSIOLOGY OF ISCHEMIA

Neuronal tissue damage may result from two phases of ischemic insult. The first or initial
phase of acute ischemia results from interruption of neural tissue perfusion. The second
insult occurs following reperfusion of the neural tissue with the inherent byproducts of
reperfusion injury.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

SPINAL CORD ISCHEMIA 223

A. Acute Ischemia

The acute ischemic phase can result in neural infarct if the duration and severity of
ischemia is prolonged (16). Like other cells, neurons depend on high-energy substrate to
function. In particular, the gray matter, which relies on a microvascular blood supply, has
increased cellular and mitochondrial density, requiring greater oxygen concentration.
Therefore this is the first area of neuronal death under ischemic conditions (17).
Diminished energy substrate delivery results in decreased cellular activity as well as an
acidic environment following anaerobic energy utilization (18,19). In this anoxic environ-
ment, depolarization of neurons releases excitatory amino acids such as glutamate and
aspartate (20), which have been retrieved at elevated levels in the cerebrospinal fluid of
patient with ischemic spinal cord injury following aneurysm repair (21). These excitatory
messengers bind and activate N-methyl D-aspartate (NMDA) and non-NMDA receptors,
with a resulting influx of calcium and activation of second messengers and production of
prostaglandins and thromboxanes. In concert, these mediators induce constriction of
smooth muscle cells and increased resistance, with further ischemia. This cycle is initiated
by the release of excitatory amino acids, with a resulting positive feedback that can result
in cell death beyond the area of critical ischemia.

B. Reperfusion Injury

In the second phase of reperfusion injury, with the reintroduction of oxygen where there is
already preexisting vasoconstriction, tissue swelling, and metabolites, oxygen radicals are
formed. These by-products destroy cells via lipid peroxidation, with membrane breakdown
and further release of excitatory amino acids (17). Reperfusion also introduces inflamma-
tory cells, such as activated leukocytes, which adhere to the microvasculature and release
cytotoxic mediators (22). The cellular damage induced by this reperfusion process functions
as debris in further occluding the microvasculature and propogating the ischemic insult.

III. DETERMINANTS OF SPINAL CORD ISCHEMIA

A. Aortic Pathology

As the spinal cord blood supply is often segmental and dependent on a contribution from

collateral arteries, the need for extensive aortic replacement requires interruption of an

increasing number of branch vessels providing spinal cord perfusion and thus a higher risk

of ischemia (Fig. 2). The more extensive type II thoracoabdominal aneurysms have been

reported to have a 3 1 % paraplegia rate, whereas in the less extensive type IV aneurysms,

paraplegia occurred in only 4% (6,7). In patients with aortic dissection, the rate rose to

40% among type II aneurysms. The study also identified a poorer neurological outcome

for thoracic aneurysms as compared to abdominal aneurysms. Aortic dissection has been -a

associated with an increased risk of neurological injury. While dissection is often reported |

as a risk factor for spinal cord injury, a more recent study of patients undergoing J

thoracoabdominal aortic aneurysm repair using contemporary methods identified the -g

acute dissections as more predictive of neurological injury (23). §

4

B. Aortic Cross-Clamp Time |

The length of aortic cross-clamp time of is often cited as a critical factor in determining «

risk of paraplegia (2,24-27), where the risk rises linearly as the time progresses beyond |

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

224

CARROCCIO et al.

Figure 2 Intraoperative photographs following repair of a type II aortic aneurysms.

20-30 min of aortic occlusion (Fig. 3). The impact of clamp time has been disputed by
some (28,29); the suggestion is that it should be viewed with consideration of other co-
existing variables such as distal perfusion, operative experience, and adjunctive techniques.
Spinal cord protective measures, such as cerebrospinal fluid drainage and distal aortic
perfusion, may decrease the risk of spinal cord ischemia at a given cross-clamp time as
compared to the same clamp time without the utilization of these adjunctive measures.

C. Proximal Hypertension

Cerebral perfusion pressure equals the mean arterial pressure minus the cerebrospinal fluid
(CSF) pressure. Maintaining adequate cerebral perfusion, therefore, may be influenced by
fluctuation in these variables. Placement of a proximal aortic clamp can interfere with the
autoregulatory response, with resulting fluctuations in cerebral and spinal blood flow. In
an animal study, lowering of proximal aortic pressure caused a significant decrease in CSF
oxygenation; restoration of mean proximal aortic pressure caused a recovery of CSF ox-
ygen tension (30).

Control of proximal hypertension following aortic cross-clamping maintains auto-
regulation in the coronary or cerebral circulation. Proximal hypertension, however, can
cause an increase in CSF pressure, and use of nitroprusside can result in increased CSF
production. A diminished mean arterial pressure with an unchanged or possibly increased
CSF pressure may result in decreased cerebral perfusion pressure of the distal cord.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

SPINAL CORD ISCHEMIA

225

4 A

1.0

~— CroM~clamp

CSFD+DAP

S"

o.e

,''

s

* .

?

f /

r

/ /

£

/ /

o 0.4

/' f

5

s* y

s

s s

0.2

y /

n ft

-^-^^

0.0

•

\ 1 I 1 L I

20

40

60

80

100

120

Aortic Clamp Time (min)

Figure 3 Graph demonstrating association between paraplegia risk and aortic cross-clamp time in
patients undergoing repair of type I and II thoracoabdominal aneurysms. The influence of cerebral
spinal fluid drainage and distal aortic perfusion is also represented. (From Ref. 2.)

IV. DETECTION OF SPINAL CORD ISCHEMIA

An impetus for intraoperative monitoring of spinal cord ischemia is that measures may be
taken to help counteract the ischemic insult and avoid deleterious consequences. One
method of detection involves evaluating somatosensory or motor evoked potentials (SEPs
or MEPs). By stimulating a peripheral nerve in the lower extremity such as the posterior
tibial nerve and measuring the cortical response over time as a tracing, one can follow the
parameters of latency and amplitude. Increasing latency and decreasing amplitude can
then serve as a marker for neuronal ischemia (31,32). With increasing levels of ischemia,
the tracing can disappear. This can be beneficial only if the determined ischemia can be
reversed. If, during aortic exclusion, one identifies SEPs with increasing latency and
decreasing amplitude, one should consider using a distal perfusion mechanism. While
utilizing distal perfusion, changing SEPs may signify that excluded collaterals are critical;
therefore the patient may benefit from intercostal reimplantation. Also, fading SEPs with
low distal perfusion pressures should encourage measures to increase distal perfusion
pressure or a more expeditious completion of an anastomosis (32).

Criticism of the use of SEPs points to the fact that they represent the cord at the
posterior column level, influenced by peripheral nerve involvement, and reports have
noted paraplegia despite a lack of change in the SEPs (33). Despite the enthusiasm of
some, comparison of outcomes resulting from the use of SEP detection versus a clamp-
and-sew technique and no SEPs had no bearing on neurological outcome (6). Given the

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

226

CARROCCIO et al.

criticism, a more sensitive detection of the anterior horn cells by MEPs has been in-
vestigated (34). Use of epidural anesthesia or stimulation of the motor cortex can be
followed by measuring peripheral motor nerve action potential response (35), with a
decrease in amplitude greater than 25% below baseline serving as an indicator of ischemia.
Problems with seizure activity (36) and the need for low-level neuromuscular blockade are
difficulties occurring at this level of detection.

In a study comparing the utilization of motor evoked potentials (MEPs) with
somatosensory evoked potentials (SEPs) during aortic surgery, MEPs proved more
beneficial (37). In comparison to monitoring with MEPs, SEPs showed delayed detection
of ischemia and a high rate of false-positive results.

V. PREVENTION

A. Intercostal Artery Reimplantation

Prevention of spinal cord ischemia by the reimplantation of intercostal arteries using a
Carrel patch or interposition graft during aortic aneurysm repair is a measure that has
supporters as well as detractors (Fig. 4). Evidence for the benefit of this technique may rely
first on the basic concept of providing an adequate blood supply to the spinal column,

I
•5

Figure 4 Intraoperative photos following thoracoabdominal aortic aneurysm repair. A. Arrow
indicates intercostal artery reimplantation via a carrel patch. B. Arrow indicates intercostal artery
reimplantation via an interposition graft.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

SPINAL CORD ISCHEMIA 227

which may be dependent on collateral blood flow from the intercostal arteries. Animal
studies utilizing a baboon model have identified a spinal cord blood flow of 15-20 mL/
lOOg of tissue per minute at baseline. Following a cross-clamp time of 60 min, the flow had
dropped to 1.5 mL/lOOg of tissue per minute. No evidence of ischemic injury was noted
when the blood supply was maintained >10 mL/100 g/min, while paraplegia resulted from
flow <4mL/100 g/min (38).

The ability of intercostal reimplantation to supply adequate spinal cord blood flow,
thus avoiding ischemic consequences, is based on observed changes in SEPs following
selective reimplantation of various excluded intercostal branches (39,40).

The decision for intercostal reimplantation has also been described based upon eval-
uation of preoperative angiography (41-43). Arteries deemed critical because of their size
and location may require reattachment.

Although studies evaluating paraplegia following aortic replacement have been
reported to reveal the benefit of reimplantation (44,45), others have found the end result
to show no difference based upon whether or not intercostals are reimplanted (6,46). A
more accurate explanation probably relies on the status of the intercostals during intra-
operative assessment. Cases with unidentified or occluded intercostals have been described
as less likely to result in paraplegia. Also, a true assessment of the benefits of such a
measure is often difficult to describe given the individual circumstances during surgery as
well as the effect that other adjunctive measures may have.

More recently, a greater appreciation for the contribution of collateral and intercostal
blood supply has emerged from the endovascular treatment of aortic aneurysms (47). In
our experience, paraplegia following endovascular thoracic aortic aneurysm repair
occurred in 3 of 72 patients (4.2%). After excluding patients who had previous or
concommitant open AAA repair, we observed a paraplegia rate of 0%. The neurodeficits
may result from interruption of critical intercostal arteries, which become more evident
with abdominal aortic replacement and or loss of collateral flow from lumbar arteries.

B. CSF Drainage

As mentioned earlier, cerebral or spinal cord perfusion pressure depends on the difference
between mean arterial pressure (MAP) and CSF pressure (CSFP). Regulating a balance
requires control not only of arterial pressure but also of of the CSF pressure. Regulation
of proximal blood pressure following placement of an aortic cross clamp is necessary to
avoid deleterious myocardial and cerebral effects. This may diminish distal aortic
perfusion with a relative increase in CSF pressure, which can be regulated by drainage
of the CSF fluid. In a study evaluating the benefit of both distal aortic perfusion as well as
CSF drainage, the combined adjuncts demonstrated an improved neurological outcome
with repair of thoracic and thoracoabdominal aortic aneurysm (48). With numerous ani-
mal studies describing the benefit of CSF drainage, a prospective randomized study in
humans identified no difference in neurological outcome among patients undergoing CSF
drainage compared to those who did not (26). Supporters of this technique criticize the J>

study for having limited the volume of CSF drainage as well as the drainage itself to only ^

the intraoperative time period. More recent randomized studies have shown a significant J

decrease in paraplegia among patients undergoing CSF drainage (2,49). Interestingly, jf

cases of delayed spinal cord ischemia presenting weeks after aortic aneurysm surgery have g

shown some benefit from CSF drainage. While the mechanism remains unclear, the benefit "|

of CSF drainage is recognized (50). 2

t

a

s

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

228 CARROCCIO et al.

C. Hypothermia

The principle that hypothermia diminishes the consequences of spinal cord ischemia is
based upon a reduction in the decline of glucose and ATP stores as well as a reduction in
the release of excitatory amino acids (18). Methods of achieving hypothermia have
included surface cooling, perfusion of cooled blood or saline into the distal occluded
aorta, cold perfusion of the subarachnoid space, and profound hypothermia with bypass
(8). Generalized or surface cooling risks systemic complications, including coagulopathy;
therefore it may be preferable to utilize hypothermia of the paraspinal space. In studies
using rabbits, satisfactory spinal cord protection during aortic occlusion can be achieved
at moderate regional hypothermia (51). A human study comparing epidural cooling to a
previous cohort suggests a benefit of epidural cooling (52). Alternative approaches have
included localized hypothermia by perfusion of exposed collaterals (53). Unfortunately,
use of epidural cooling in humans has resulted in a small incidence of brainstem infarction,
which is obviously quite worrisome.

D. Distal Aortic Perfusion

Various methods of improving distal aortic flow have been utilized. The benefit of sustaining
lumbar cord flow and the support it provides to excluded segments of thoracic aorta is
unclear (54). Methods of distal perfusion have included both full and partial bypass.

Full cardiopulmonary bypass can provide a more controlled distal perfusion as well as
hypothermia; however, the degree of anticoagulation does impose the increased risk of
bleeding complications in extensive aortic replacements (15,25,55). The introduction of
heparin-bonded shunts (56) were meant to circumvent this problem; however, the small
cross-sectional area of the shunt conduits (35) can result in a significant gradient to the
distal aorta, with a diminished cardiac output (14). In addition, use of shunts also poses
inherent problems with dislodgment and embolization (57).

Partial bypass can provide distal perfusion and control of proximal hypertension
without cardiopulmonary arrest. Proximal access can be achieved at the levels of the
pulmonary veins, left atrium, or proximal aorta. Distal perfusion can be introduced to the
distal aorta and femoral arteries. Additional selective perfusion cannulas can be intro-
duced to the visceral and renal branches when the aortic lumen is exposed, thereby
possibly mitigating coexistent visceral and renal ischemia (Fig. 5). The clear benefit of such
distal perfusion techniques has yet to be determined (14,25,29). Distal perfusion can also
be addressed by extra-anatomic bypass. Preoperative construction of axillofemoral bypass
can provide distal flow during aortic exclusion without the need for an external pump or
anticoagulation beyond that which is necessary for the aortic replacement (15,29,58).

Cardiopulmonary bypass increases circulating levels of cytokines (59-61). We believe
that less cytokine production will result from the use of an accessory graft for perfusion of ■§

the visceral segment as opposed to the pump. A technique we have lately utilized in the |

treatment of type I, II, and III thoracoabdominal aneurysms consists of the incorporation a

of an accessory bifurcated graft originating from the side of the proximal main graft, with c

one limb being sutured initially to the left iliac artery. Following the creation of the <

proximal aortic anastamosis, a cross-clamp is applied to the main graft, distal to the origin &

of the accessory graft. This allows blood to flow through the new proximal anastamosis J

into the accessory graft and out the two limbs of this accessory graft. One limb of the °

accessory graft anastamosed to the common iliac artery will provide pelvic and lower |

extremity blood flow. The second limb provides flow through selective perfusion catheters @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

SPINAL CORD ISCHEMIA

229

Figure 5 Intraoperative photo during thoracoabdominal aortic aneurysm repair. Arrows indicate
orifice of visceral and renal arteries with catheters providing selective perfusion during repair.

to the renal and visceral vessels. Visceral and pelvic ischemia is limited to the time required
for creation of the proximal anastamosis. This accesory graft provides distal perfusion
while the intercostal, visceral, renal, and distal aortic reconstructions are performed.

E. Selective Visceral Perfusion

Changes in plasma cytokine concentrations were related to the duration of visceral ischemia
and the frequency rate of postoperative, single, or multiple-system organ dysfunction in a
study examining plasma proinflammatory cytokines after abdominal and thoracoabdomi-
nal aortic aneurysm (TAAA) repair (62). In a rabbit model, we identified an association of
viscerally derived cytokines and spinal cord ischemia (63). With interruption of aortic and
visceral flow, the deleterious effects on the spinal cord were more pronounced than with
aortic interruption alone. Currently we are investigating this relationship in a population of
patients undergoing thoracoabdominal aneurysm repair. The benefit of selective visceral
perfusion during thoracoabdominal aneurysm repair may improve outcome (64). The exact
role it may have on spinal cord ischemia has yet to be determined.

I
•5

F. Pharmacotherapy

The benefit of delineating the physiological mechanism behind spinal cord ischemia is that
medicinal intervention at multiple levels may mitigate the consequences of spinal ischemia.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

230 CARROCCIO et al.

While being investigated predominantly in animal models, areas of interest have included
blockage of calcium influx (16), control of vasoconstriction (65), control of platelet
aggregation and prostenoid release (66), blocking the release or downstream effects of
excitatory amino acids (24,67-69), the decrease of reperfusion injury (17,70,71), inhibiting
neutrophil activation after the transient ischemia with activated protein C (72), membrane
stabilization and control of apoptosis with steroids (73), and the benefit of growth factors
(74,75) as well as ischemic preconditioning (76).

Future means of spinal cord protection may include gene delivery of neurotrophic
growth factors. Investigation into the possible protective effect of the glial cell line-de-
rived neurotrophic factor via adenovirus-mediated gene delivery on transient spinal cord
ischemia in rabbits has revealed improvement in the survival of motor neurons (77).

VI. SUMMARY

Investigation into the basic elements responsible for spinal cord ischemia has provided a
better understanding of this complex process. Despite the many human as well as animal
trials to prevent this dreaded complication, no one distinguishing measure seems most
beneficial. In addition to extensive surgical experience in aortic surgery, the best current
approach appears to be the utilization of multiple adjunctive procedures as determined
appropriate on individual bases, with the hope of interrupting the deleterious ischemic
cascade at several levels.

REFERENCES

1. Coselli JS, LeMaire SA. Left heart bypass reduces paraplegia rates after thoracoabdominal
aortic aneurysm repair. Ann Thorac Surg 1999; 67:1931-1934.

2. San HJ, Hess KR, Randel M, Iliopoulos DC, Baldwin JC, Mootha RK, et al. Cerebrospinal
fluid drainage and distal aortic perfusion: Reducing neurologic complications in repair of
thoracoabdominal aortic aneurysms types II and I. J Vase Surg 1996; 23:223-229.

3. Cambria RP, Davison K, Carter C, Brewster DC, Chang Y, Clark KA, et al. Epidural cooling
for spinal cord protection during thoracoabdominal aneurysm repair: a five-year experience. J
Vase Surg 2000; 31:1093-1102.

4. Katz NM, Blackstone EH, Kirklin JW, Karg RB. Incremental risk factors for spinal cord
injury following operation for acute traumatic transection. J Thorac Cardiovasc Surg 1981;
81:669-674.

5. lacobs MI, de Mol BA, Elenbaas T, Mess WH, Kalkman CI, Schurink GW, Mochtar B.
Spinal cord blood supply in patients with thoracoabdominal aortic aneurysms. I Vase Surg
2002; 35(l):30-37.

6. Crawford E, Crawford J, Safi H, et al. Thoracoabdominal aortic aneurysm: Preoperative and
intraoperative factors determining immediate and long term results of operation in 605 pa- j>
tients. I Vase Surg 1986; 3:389-404. |

7. Svensson LG, Crawford ES, Hess KR, Coselli IS, Safi HI. Experience with 1509 patients js
undergoing thoracoabdominal aortic operations. J Vase Surg 1993; 17(2):357-370. °|

8. Kouchoukos N. Rokkas C. Descending thoracic and thoracoabdominal aortic surgery for **
aneurysm or dissection: how do we minimize the risk of spinal cord injury. Semin Thoracic «
Cardiovasc Surg 1993; 5:47-54. J

9. Svensson L, Klepp P, Hinder R. Spinal cord anatomy of the baboon: comparison with man q
and implications for spinal cord blood flow during thoracic aortic cross clamping. S Afr I Surg |
1986; 24:32-34. S

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

SPINAL CORD ISCHEMIA 231

10. Piccone A, Green R, Ricotta J, et al. Spinal cord ischemia following operations on the ab-
dominal aorta. J Vase Surg 1986; 3:94-103.

11. Szilagyi D, Hageman J, Smith R, Elliot J. Spinal cord damage in surgery of the abdominal
aorta. Surgery 1978; 83:38-56.

12. Gharagozloo F, Larson J, Dausmann M, et al. Spinal cord protection during surgical procedures
and the descending thoracic and thoracoabdominal aorta. Chest 1996; 109:799-809.

13. Adamkiewicz A. Die Blutgefasse des menschlichen Ruckernmarkes: I. Theil die Gefasse der
Ruckenmarksubstanz. Sitzungsb Akad Wissensch Wien Math Naturw Klass 1882; 84:469.

14. Svensson L, Loop F. Prevention of spinal cord ischemia in aortic surgery. In: Yao JST, ed.
Arterial Surgery. New York: Grune & Stratton, 1988:273-285.

15. von Oppell U, Dunne T, De Groot M, Zilla P. Spinal cord protection in the absence of
collateral circulation: meta analysis of mortality and paraplegia. J Card Surg 1994; 9:685-691.

16. Kwun BD, Vacanti F. Mild hypothermia protects against irriversible damage during prolonged
spinal cord ischemia. J Surg Res 1995; 59:780-782.

17. Hall E, Wolf D. A pharmacologic analysis of the pathophysiologic mechanisms of post trau-
matic spinal cord ischemia. J Neurosurg 1986; 64:951-961.

18. Allen B, Davis C, Osborne D, Karl I. Spinal cord ischemia and reperfusion metabolism: the
effect of hypothermia. J Vase Surg 1994; 19:332-340.

19. Drenger B, Parker S, Frank S, Beattie C. Changes in cerebral spinal fluid pressure and lactate
concentrations during thoracoabdominal aortic aneurysm surgery. Anesthesiology 1997; 86:
41-47.

20. Taira Y, Marsala M. Effect of proximal arterial perfusion pressure on function, spinal cord
blood flow, and histopathologic changes after increasing intervals of aortic occlusion in the rat.
Stroke 1996; 27:1850-1858.

21. Brock MV, Redmond JM, Ishiwa S, Johnston MV, Baumgartner WA, Laschinger JC,
Williams GM. Clinical markers in CSF for determining neurologic deficits after thoracoabdom-
inal aortic aneurysm repairs. Ann Thorac Surg 1997; 64(4):999-1003.

22. Clark W, Walsh C, Briley D, Brace C. Neutrophil adhesion in central nervous system ischemia
in rabbits. Brain Behav Immun 1993; 7:63-69.

23. Coselli JS, LeMaire SA, de Figueiredo LP, Kirby RP. Paraplegia after thoracoabdominal
aortic aneurysm repair: Is dissection a risk factor? Ann Thorac Surg 1997; 63(l):28-36.

24. Rothman S, Olney J. Glutamate and the pathophysiology of hypoxic-ischemic brain damage.
Ann Neurol 1986; 19:105-111.

25. Livesay L, Cooley D, Ventemiglia R, et al. Surgical experience in descending thoracic aneu-
rysmectomy with and without adjuncts to avoid ischemia. Ann Thorac Surg 1985; 39:37-46.

26. Crawford E, Svensson L, Hess K, et al. A prospective randomized study of cerebral spinal fluid
drainage to prevent paraplegia after high risk surgery on the thoracoabdominal aorta. J Vase
Surg 1990; 13:36-46.

27. Katz N, Blackstone E, Kirklin J, et al. Incrimental risk factors for spinal cord injury following
operation for acute traumatic aortic transection. J Thorac Cardiovasc Surg 1981; 81:669-
674.

28. Hollier L, Symmonds J, Pairolero P, et al. Thoracoabdominal aortic aneurysm repair: analysis

of postoperative morbidity. Arch Surg 1988; 123:871-875. 1

29. Crawford E, Rubio P. Reappraisal of adjuncts to avoid ischemia in the treatment of aneurysms <S
of the descending aorta. J Thorac Cardiovasc Surg 1973; 66:693-704. js

30. Hellberg A, Tulga Ulus A, Christiansson L, Bergqvist D, Thelin S, Karacagil S. Influence of °|
low proximal aortic pressure on spinal cord oxygenation in experimental thoracic aortic oc- **
elusion. J Cardiovasc Surg (Torino) 2001; 42(2):227-231. |.

31. Cunningham JJ, Laschinger J, Merlin H, et al. Measurement of spinal cord ischemia during J
operations upon the thoracic aorta. Ann Surg 1982; 196:285-296. q

32. Marini C, Cunningham J. Issues surrounding spinal cord protection. In: Karp R, Laks H, |

Wechsler A, eds. Advances in cardiac surgery. St. Louis: Mosby-Year Book, 1993:89-107. 2

©

!

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

232 CARROCCIO et al.

33. Lesser R, Raudzens P, Luder H, et al. Postoperative neurologic deficits may occur despite
unchanged intraoperative somatosensory evoked potentials. Ann Neurol 1986; 19:22-25.

34. Agnew W, McCreery D. Considerations of safety in the use of extracranial stimulation for
motor-evoked potentials. Neurosurgery 1987; 20:143-147.

35. de Haan P, Kalkman C, de Mol B, et al. Efficacy of transcranial motor evoked myogenic
potentials to detect spinal cord ischemia during operations for thoracoabdominal aneurysms. J
Thorac Cardiovasc Surg 1997; 113:87-101.

36. Shenaq S, Svensson L. Paraplegia following aortic surgery. J Cardiothorc Vase Anesth 1993;
7:81-94.

37. Meylaerts SA, Jacobs MJ, van Iterson V, De Haan P, Kalkman CJ. Comparison of trans-
cranial motor evoked potentials and somatosensory evoked potentials during thoracoabdo-
minal aortic aneurysm repair. Ann Surg 1999; 230(6):742-749.

38. Svensson L. Rickards E, Coull A, et al. Relationship of spinal cord blood flow to vascular
anatomy during thoracic aortic cross-clamping and shunting. J Thorac Cardiovasc Surg 1986;
91:71-78.

39. Ueda T, Shimizu H, Mori A, Kashima I, Moro K, Kawada S. Selective perfusion of segmental
arteries in patients undergoing thoracoabdominal aortic surgery. Ann Thorac Surg 2000;
70(l):38-43.

40. Meylaerts SA, De Haan P, Kalkman CJ, Jaspers J, Vanicky I, Jacobs MJ. Prevention of
paraplegia in pigs by selective segmental artery perfusion during aortic cross-clamping. J Vase
Surg 2000; 32(1): 160- 170.

41. Savader S, Williams G, Trerotola S, et al. Preoperative spinal artery localization and its
relationship to postoperative neurologic complications. Radiology 1993; 189:165-171.

42. Bachet J, Guilmet D, Rosier J, et al. Protection of the spinal cord during surgery of thoraco-
abdominal aortic aneurysms. Eur J Cardiothorac Surg 1996; 10:817-825.

43. Kieffer E, Richard T, Chiras J, et al. Preoperative spinal cord arteriography in aneurysmal
disease of the descending thoracic and thoracoabdominal aorta: preliminary results in 45
patients. Ann Vase Surg 1989; 3:34-46.

44. Shiiya N, Yasuda K, Matsui Y, et al. Spinal cord protection during thoracoabdominal aneu-
rysm repair: Result of selective reconstruction of the critical segmental arteries guided by
evoked spinal cord potential monitoring. J Vase Surg 1995; 21:970-975.

45. Svensson L, Hess K, Coselli J, Safi H. influence of segmental arteries, extent, and aortofemoral
bypass on postoperative paraplegia after thoracoabdominal aortic operation. J Vase Surg 1994;
20:255-262.

46. Schepens M, Boezeman E, Hamerlijnk R, et al. Somatosensory evoked potentials during ex-
clusion and reperfusion of critical aortic segments in thoracoabdominal aortic aneurysm
surgery. J Card Surg 1994; 9:692-702.

47. Gravereaux EC, Faries PL, Burks JA, Latessa V, Spielvogel D, Hollier LH, Marin ML. Risk of
spinal cord ischemia after endograft repair of thoracic aortic aneurysms. J Vase Surg 2001;
34(6):997-1003.

48. Estrera AL, Miller CC III, Huynh TT, Porat E, Safi HJ. Neurologic outcome after thoracic and
thoracoabdominal aortic aneurysm repair. Ann Thorac Surg 2001; 72(4):1225-1231.

49. Coselli JS, Lemaire SA, Koksoy C, Schmittling ZC, Curling PE. Cerebrospinal fluid drainage j>
reduces paraplegia after thoracoabdominal aortic aneurysm repair: results of a randomized <S
clinical trial. J Vase Surg 2002; 35(4):63 1-639. I

50. Azizzadeh A, Huynh TT, Miller CC III, Safi HJ. Reversal of twice-delayed neurologic deficits °|
with cerebrospinal fluid drainage after thoracoabdominal aneurysm repair: a case report and "•
plea for a national database collection. J Vase Surg 2000; 31(3):592-598. «

51. Martelli E, Cho JS, Mozes G, Gloviczki P. Epidural cooling for the prevention of ischemic jjj
injury to the spinal cord during aortic occlusion in a rabbit model: determination of the optimal Q
temperature. J Vase Surg 2002; 35(3):547-553. |

52. Cambria RP, Davison JK, Carter C, Brewster DC, Chang Y, Clark KA, Atamian S. Epidural 2

t

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

SPINAL CORD ISCHEMIA 233

cooling for spinal cord protection during thoracoabdominal aneurysm repair: a five-year ex-
perience. J Vase Surg 2000; 31(6): 1093- 1802.

53. Coles J, Wilson G, Sima A, et al. Intraoperative management of thoracic aortic aneurysms:
Experimental evaluation of perfusion cooling of the spinal cord. J Thorac Cardiovasc Surg
1983; 85:292-299.

54. Svensson L, Patel V, Robinson M, et al. Influence of preservation or perfusion of intra-
operatively identified spinal cord blood supply on spinal motor evoked potentials and para-
plegia after aortic surgery. J Vase Surg 1991; 13:355-365.

55. Jex R, Schaff H, Piehler J, et al. Early and late results following repair of dissection of the
descending aorta. J Vase Surg 1986; 3:226-237.

56. Gott V. Heparinazed shunts for thoracic vascular operations (editorial). Ann Thorac Surg
1972; 14:219.

57. Crawford E, Fenstermacher J, Richardson W, Sandiford F. Reappraisal of adjuncts to avoid
ischemia in treatment of thoracic aortic aneurysms. Surg 1970; 67:182.

58. Molina J, Cogordan J, Einzigs, et al. Adequacy of ascending aorta-descending aorta shunt
during cross clamping of the thoracic aorta for the prevention of spinal cord ischemia. J Thorac
Cardiovasc Surg 1985; 90:126-136.

59. Chenoweth DE, Cooper SW, Hugh TE, Stewart RW, Blackstone EH, Kirklin JW.
Complement activation during cardiopulmonary bypass: Evidence for generation of C3a and
C5a anaphylotoxins. N Engl I Med 1981; 304:497-503.

60. Bruins P, te Velthuis H, Yazdanbakhsh AP, et al. Activation of the complement system during
and after cardiopulmonary bypass surgery: postsurgery activation involves C-reactive protein
and is associated with postoperative arrhythmia. Circulation 1997; 96:3542-3548.

61. Paparella D, Yau TM, Young E. Cardiopulmonary bypass induced inflammation: patho-
physiology and treatment. An update. Eur J Cardiothorac Surg 2002; 21(2):232-244.

62. Welborn MB, Oldenburg HS, Hess PJ, et al. The relationship between visceral ischemia,
proinflammatory cytokines, and organ injury in patients undergoing thoracoabdominal aortic
aneurysm repair. Crit Care Med 2000; 28(9):3 191 -3 197.

63. Morrissey NJ, Kantonen I, Liu H, et al. The effect of mesenteric ischemia/reperfusion on spinal
cord injury following transient aortic occlusion in rabbits. I Endovasc Ther 2002; 9(suppl 2):
1144-1150.

64. Cambria RP, Davison JK, Giglia JS, Gertler JP. Mesenteric shunting decreases visceral ische-
mia during thoracoabdominal aneurysm repair. J Vase Surg 1998; 27(4):745-749.

65. Svensson L, Von Ritter C, Groeneveld H, et al. Cross-clamping of the thoracic aorta: influence of
aortic shunts, laminectomy, papaverine, calcium channel blockers, allopurinol, and superoxide
dismutase on spinal cord blood flow and paraplegia in baboons. Ann Surg 1986; 204:38-47.

66. Lapchak PA, Araujo DM, Song D, Zivin JA. Neuroprotection by the selective cyclooxygenase-
2 inhibitor SC-236 results in improvements in behavioral deficits induced by reversible spinal
cord ischemia. Stroke 2001; 32(5):1220-1225.

67. Kanellopoulos GK, Xu XM, Hsu CY, Lu X, Sundt TM, Kouchoukos NT. White matter injury
in spinal cord ischemia: protection by AMPA/kainate glutamate receptor antagonism. Stroke
2000; 31(8):1945-1952.

68. Lang-Lazdunski L, Heurteaux C, Mignon A, Mantz J, Widmann C, Desmonts J, Lazdunski j>
M. Ischemic spinal cord injury induced by aortic cross-clamping: prevention by riluzole. Eur J <S
Cardiothorac Surg 2000; 1 8(2): 1 74- 1 8 1 . I

69. Lang-Lazdunski L, Heurteaux C, Dupont H, Widmann C, Lazdunski M. Prevention of ische- °|
mic spinal cord injury: comparative effects of magnesium sulfate and riluzole. J Vase Surg 2000; ^
32(1):179-189. |.

70. Rahman A, Ustundag B, Burma O, Ozercan IH, Erol FS. Neuroprotective effect of regional jjj
carnitine on spinal cord ischemia-reperfusion injury. Eur J Cardiothorac Surg 2001; 20(1):65- q
70. |

71. de Haan P, Vanicky I, Jacobs MJ, Bakker O, Lips J, Meylaerts SA, Kalkman CL Effect of 2

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

234 CARROCCIO et al.

ischemic pretreatment on heat shock protein 72, neurologic outcome, and histopathologic
outcome in a rabbit model of spinal cord ischemia. J Thorac Cardiovasc Surg 2000; 120(3):
513-519.

72. Hirose K, Okajima K, Taoka Y, Uchiba M, Tagami H, Nakano K, Utoh J, Okabe H, Kita-
mura N. Activated protein C reduces the ischemia/reperfusion-induced spinal cord injury in
rats by inhibiting neutrophil activation. Ann Surg 2000; 232(2):272-280.

73. Kanellopoulos GK, Kato H, Wu Y, Dougenis D, Mackey M, Hsu CY, Kouchoukos NT.
Neuronal cell death in the ischemic spinal cord: the effect of methylprednisolone. Ann Thorac
Surg 1997; 64(5): 1279- 1286.

74. Nakao Y, Otani H, Yamamura T, Hattori R, Osako M, Imamura H. Insulin-like growth factor 1
prevents neuronal cell death and paraplegia in the rabbit model of spinal cord ischemia. J Thorac
Cardiovasc Surg 2001; 122(1):136-143.

75. Bowes M, Tuszynski MH, Conner J, Zivin JA, Continuous intrathecal fluid infusions elevate
nerve growth factor levels and prevent functional deficits after spinal cord ischemia. Brain Res
2000;17;883(2):178-183.

76. Zvara DA, Colonna DM, Deal DD, Vernon JC, Gowda M, Lundell JC. Ischemic pre-
conditioning reduces neurologic injury in a rat model of spinal cord ischemia. Ann Thorac Surg
1999; 68(3):874-880.

77. Sakurai M, Abe K, Hayashi T, Setoguchi Y, Yaginuma G, Meguro T, Tabayashi K. Ade-
novirus-mediated glial cell line-derived neurotrophic factor gene delivery reduces motor neuron
injury after transient spinal cord ischemia in rabbits. J Thorac Cardiovasc Surg 2000;
120(6):1148-1157.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

11

Impotence Following Aortic Surgery

Richard Kempczinski

University of Cincinnati, Cincinnati, Ohio, U.S.A.

Although Leriche and Morel (1) first described the association between impotence and
distal aortic occlusion, Harris and Jepson (2) were the first to report erectile dysfunction as
a complication of aortic surgery. These studies first called our attention to vasculogenic
impotence, but a true appreciation of the magnitude of this problem awaited more refined
diagnostic techniques and a better understanding of the mechanism of erection.

Impotence, the inability to achieve or maintain an erection adequate for satisfactory
coitus, must be distinguished from retrograde ejaculation, which is primarily a neurogenic
disorder in which bladder neck closure does not occur and semen is deposited into the
bladder. In such cases the patient is still able to complete coitus and achieve orgasm.
Although impotence, by the above definition, appears to be a disorder limited to men,
women with aortoiliac arterial obstructive disease may complain of insufficient vaginal
lubrication and loss of orgasm (3). However, this is a much less common problem, because
female genital sensation depends in great part on the integrity of the somatic pudendal
nerves and their efferent sensory fibers. These are situated deep within the pelvis and are
protected by the thick layer of endopelvic fascia. Furthermore, collateral arterial blood
supply to the female sexual organs is quite extensive. As a result, female organic "im- ■a

potence" is extremely uncommon (4). §

This chapter describes the physiology of erection as we currently understand it and a

explores how the normal interplay of neural and vascular elements can be disturbed by the c

development of arterial occlusive disease or by surgical attempts to correct it. Because <

patients with erectile dysfunction often consult vascular surgeons, the various diagnostic &

techniques that may be necessary in these patients are emphasized. Finally, those technical J

modifications that should be employed during aortic surgery to prevent iatrogenic «

impotence, as well as the most effective current treatment for established impotence, are |

discussed. 5

235

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

236

KEMPCZINSKI

I. PHYSIOLOGY OF ERECTION

In the 40 years since Leriche and Morel (1) first described the association between aor-
toiliac arterial occlusive disease and impotence, medical research has broadened our
understanding of the complex interplay between psychological, hormonal, neurological,
and vascular elements that is required to achieve an adequate erection.

Satisfactory male sexual function presupposes anatomically normal male genitalia, an
appropriate hormonal milieu, intact nerve and blood supply to the genitalia, and appro-
priate physical and/or psychic stimulation. Absence of any one of these elements or
moderate dysfunction in several of them may result in impotence.

A. Neurophysiology

The precise neurophysiological basis for erection remains unknown. The afferent and
efferent neural pathways that appear to be involved in erection are depicted in Figure 1.
Thoracolumbar sympathetic nerves (TI2-L4) are believed to be important in mediating
psychogenic erections, which can occur even in patients with complete sacral cord de-
structions (5). However, younger individuals undergoing bilateral radical retroperitoneal
node dissection, in which both sympathetic nerve chains are usually removed, rarely be-
come impotent (6). Therefore sacral efferent (parasympathetic) outflow appears capable of
mediating both psychogenic and reflex erections. Clearly, bilateral resection of the TI 2 -LI
sympathetic ganglia can result in retrograde ejaculation. However, this should not be
confused with erectile dysfunction.

Thoracolumbar
(enter

Sympathetic
nerves (T12 L4I

PSYCHIC STIMULI

Pudendal nerves
(afferent stimuli!

Reflex stimuli
Exteroceotive
Interoceptive

INTERNAL
PUDENDAL

^ Parasympathetic
nerves IS2-4)
(~)~ Nervi engentes

-— VASODILATION <- ERECTION

Figure 1 Diagrammatic representation of the neural pathways involved in penile erection. (From
Ref. 40.)

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

IMPOTENCE AFTER AORTIC SURGERY 237

Based on current neurophysiological research, erection appears to develop as a result
of neural transmissions that reach the genitalia via the pelvic parasympathetic nerves.
Destruction of the parasympathetic outflow from the sacral cord will cause impotence.
Pelvic operations — such as radical prostatectomy or abdominal perineal resection of the
rectum, in which parasympathetic nerve damage often occurs — have been associated with
postoperative impotence in 70-100% of cases (7). The final common pathway for this
hemodynamic control appears to be the short adrenergic nerves within the penis.

The central nervous system (CNS) loci that initiate erection have not been precisely
identified. However, since these signals must reach the genitalia via the spinal cord, injury
or transection of the cord may result in impotence. Reflex erections are possible in a high
percentage of patients with lesions of the upper spinal cord, but the level of injury largely
determines the preservation of erectile potency. These reflex erections appear to require the
integrity of the afferent pudendal nerves, since pudendal neurectomies in such patients
result in impotence.

Most drugs that produce impotence do so by their actions on these neurophysiological
pathways. However, it is difficult to determine whether their actions are peripheral or
central. Ganglion-blocking agents, such as hexamethonium, are a well-known cause of im-
potence and ejaculation disturbances. Propranolol frequently causes impotence when
administered in doses greater than 200 mg/day. Drugs such as reserpine, a-methyldopa,
and tricyclic antidepressants probably produce impotence by their action on the CNS.

In summary, normal erectile function appears to involve both pelvic parasympathetic
nerves and penile corporal short adrenergic receptors. Although both a-adrenergic and (3-
adrenergic receptors are present within the penis, a-adrenergic receptors are believed to
predominate in a 10:1 ratio. In addition, recent studies have suggested that vasoactive
intestinal polypeptide, either alone (8) or in synergy with a- adrenergic blockade (9) or
acetylcholine (10), may be responsible for erection. Thus erection can no longer be con-
sidered to be a purely cholinergic event and acetylcholine is not the final neurotransmitter.
Nevertheless, many questions remain regarding the neurophysiology of erection.

B. Penile Blood Supply

When neural pathways are intact, the ability to achieve an erection is largely determined
by the adequacy of arterial inflow. The blood supply of the penis arises from the internal
pudendal artery, which is one of the terminal branches of the internal iliac artery. The
paired internal pudendal arteries enter the male perineum through the lesser sciatic fo-
ramina. Each of the internal pudendal arteries, in turn, gives rise to a dorsal penile artery,
a more laterally placed deep artery of the penis, which supplies the corpus cavernosum,
and a bulbourethral artery, which supplies the corpus spongiosum (Fig. 2). Terminal
branches of the penile arteries and the penile vessels themselves appear to communicate ■§

with the cavernous spaces via structures previously called polsters, Ebner pads, or cous- |

sinets (11). a

Recent investigations in animal models and human volunteers have settled some of the c

long-standing controversies regarding the precise sequence of events in erection (12). In the <

flaccid state, the arterioles are constricted and the venous sinusoids contracted. Together, >9

they exert maximal resistance against arterial flow, thus allowing only a small amount J

of nutrient blood to enter the corpora. The venules in the periphery of the corpora run «

between the adjacent sinusoidal walls, whereas the larger intermediary venules traverse the |

sinusoidal wall and tunica albuginea for some distance before exiting as the emissary veins. @

While the sinusoids are contracted, these venules drain freely to extrapenile veins. %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

238

KEMPCZINSKI

Obturator
artery

Inferior rectal
artery

,. Common
iliac artery

/ , > , Inferior gluteal

K ^->^

^"'2

internal

- pudendal
f artery

-- —-zrt-

V '

>2

Figure 2 The major blood supply to the penis is from the deep and dorsal penile arteries and the
urethral artery. These are branches of the internal pudendal artery, which in turn is branch of the
internal iliac artery. 1. Dorsal artery of the penis; 2. deep artery of the penis; 3. perineal artery;
4. urethral artery. (From Ref. 4.)

During erection, the smooth muscles of the sinusoids and arterioles relax, which in turn
increases sinusoidal compliance and causes a maximal decrease in peripheral resistance. This
results in an immediate increase in arterial flow and filling of the sinusoids. The resulting
dilation of the arterial tree not only allows blood to enter rapidly but also permits
transmission of approximately 80% of the arterial systolic pressure to the sinusoidal spaces
(vascular or full erection phase). Subsequent contraction of the bulbocavernous and
ischiocavernous muscles either spontaneously or reflexly compresses the proximal corpora
and culminates in cavernosal rigidity, with further engorgement of the glans penis as seen
during intercourse (skeletal muscle or rigid erection phase). In the full erection phase, mean
pressure in the corpora cavernosa is approximately 90 to 100 mmHg. In the rigid erection
phase, compression of the blood-distended corpora can increase the intracavernous pressure
well above arterial systolic pressure.

This proposed sequence of events is further supported by the work of Newman et al.
(13), who infused the pudendal arteries of human cadavers at a pressure of 200 mmHg but
were unable to produce a normal erection. Subsequently, direct infusion of the corpora
cavernosa at rates ranging from 20 to 50 mL/min resulted in a normal erection. Once an
erection was obtained, it was possible to maintain turgidity with decreased infusion rates.
These results are similar to those of studies by Michal and Pospichal (14), who dem-
onstrated that a mean infusion rate of 90 mL/min directly into the corpora was initially

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I
•5

I

IMPOTENCE AFTER AORTIC SURGERY

239

necessary to produce erections in normal human subjects. But once an erection was
achieved, a maintenance flow rate of 62 mL/min was satisfactory to maintain erection.

By injecting microspheres into the internal pudendal arteries in cadavers, investigators
have confirmed the presence of arteriovenous shunts measuring 1 p.m in diameter (13).
They also noted that occlusion of the dorsal vein of the penis failed to result in an erection.
Thus erection appears to occur as the result of preferential redirection of increased arterial
flow into the corporal spaces and active venoconstriction is apparently unnecessary.

C. Psychic Influences

Although recent work has emphasized the frequent organic nature of postoperative im-
potence, the contribution of psychogenic factors should not be lightly dismissed. Following
major surgical procedures, the patient and his sexual partner may be concerned that
resumption of normal sexual activity could be potentially harmful, thus resulting in
decreased libido and functional impotence. Even if such subconscious fears alone may be
inadequate to cause erectile dysfunction, they may be contributory in the presence of
marginal penile perfusion. When initial attempts at resumption of normal sexual activity in
the postoperative period meet with failure, a reactive depression may result, which can
prolong the problem. If appropriate neurological and vascular causes of impotence have
been excluded postoperatively, complete evaluation of the patient and, if possible, his sexual
partner by a concerned and knowledgeable psychiatrist may be helpful.

IMPOTENCE

Nocturnal Penile
Plethysmography

M-

PSYCHOGENIC IMPOTENCE

ORGANIC IMPOTENCE

££l

|Vasctilar Lab Studies!

Plasma Testosterone

{+]_

acave

M

I nt rabave rnous C y stometrography

Papaverine (X2) & Sacral Latency Testing

(-) N EUROGENIC IMPOTENCE

HORMONAL INSUFFICIENCY

Arteriography

Further
Work- up

VASCUIOOENIC IMPOTENCE

I
•5

Figure 3 Algorithm suggesting a diagnostic approach to the impotent patient (see text for details).
(From Ref. 40.)

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

240 KEMPCZINSKI

II. DIAGNOSIS

Since so many factors can result in erectile dysfunction, a multimodal diagnostic approach
to the problem is essential (Fig. 3). Even in those cases where the underlying defect
cannot be directly corrected, confirmation of the organic nature of the patient's impotence
is vital in preventing the emotional havoc that this problem can wreak on his personal life.
Furthermore, once the diagnosis of organic impotence is established, the patient can be
referred for a penile implant if appropriate correction of the specific problem is impossible.

III. HISTORY

Although many of the barriers that previously precluded frank discussion of erectile dys-
function have fallen, some patients are still reluctant to broach this problem with their
physician. Since 70-80% of patients with aortoiliac arterial occlusive disease in some series
have been impotent (15,16) and as many as 50% of diabetic patients under the age of 40 may
be similarly disabled (17), vascular surgeons must be prepared to initiate such discussions
with their patients. This is especially important preoperatively, not only to permit mod-
ification of the operation to relieve impotence when possible but also to document that the
condition antedated the surgical procedure.

A careful and detailed history of the patient's sexual dysfunction may suggest its
etiology. Organic impotence is typically of gradual onset and results in complete inability to
achieve erection. It is not partner-specific, and masturbatory and morning erections are
absent. The onset of the patient's symptoms cannot be related to any identifiable emotional
stress and libido is typically retained. By contrast, psychogenic impotence may be rapid in
onset, frequently within less than 1 month, and may be intermittent in pattern. Partner
specificity may be present, and erection can be achieved during masturbation. Morning
erections occur and the onset of the patient's symptoms can frequently be related to an
identifiable emotional stress. The presence of normal sexual drive may be quite variable.

In those patients with known organic impotence, certain historical features help to
differentiate those with neurogenic versus vasculogenic impotence. Patients suffering from
neurogenic impotence are usually unable to achieve erections at all and may have
decreased testicular sensation on palpation. Ejaculation with masturbation in such pa-
tients is generally absent. On the other hand, patients with vasculogenic impotence may be
able to achieve an erection temporarily, but it is short-lived. Testicular pain on palpation is
normal and masturbatory ejaculations are present. In patients with external iliac artery
occlusion, the ipsilateral internal iliac artery may be the major collateral blood supply to
the lower extremity. Some of these patients may report that they are able to achieve a
satisfactory erection during foreplay; however, when thrusting is initiated, the penis
becomes flaccid and coitus impossible. Presumably the increased demand for blood by
the buttock and thigh muscles during active coitus shunts blood away from the genitalia,
causing loss of erection (16).

A. Nocturnal Penile Tumescence

This study is based on the observation that sexually potent men can regularly have
erections during the rapid-eye-movement (REM) phase of sleep (18). The complete
absence of tumescence during an adequate sleep study is strong evidence of organic im-
potence. Unfortunately, the failure of erection is often qualitative rather than complete,
and it has been difficult to standardize the quality of erections. Such studies are difficult to
perform properly and are best carried out on an inpatient basis in specially equipped sleep

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I
•5

I

IMPOTENCE AFTER AORTIC SURGERY

241

laboratories (19). Changes in penile circumference are monitored by means of mercury-
filled strain gauges and/or video camera characterizations of the quality of the erection.
The documentation of normal erections during REM sleep clearly establishes the psy-
chogenic basis of the patient's erectile dysfunction and allows appropriate therapy.

B. Noninvasive Vascular Testing

Canning et al. (20) first emphasized that vascular insufficiency of the pelvic vessels, even in
the presence of normal femoral pulses, could result in impotence. They attempted to identify
such patients by palpating penile pulses and performing impedance plethysmography.
Subsequently, other investigators assessed penile blood flow using mercury strain-gauge
plethsmography, spectrographic or ultrasonic measurement of penile systolic pressure, and
pulse volume recordings (21-24). When such studies are abnormal and the possibility of
vasculogenic impotence is likely, more traditional noninvasive tests to exclude aortoiliac
arterial occlusive disease should be performed. Kempczinski (24) studied 134 patients using
the Doppler velocity meter to measure penile systolic pressure. This, in turn, was divided by
brachial systolic pressure to obtain a penile brachial index (PBI). Pulse volume waveforms
(PVW), or penile volume change with each cardiac cycle, were also recorded. The influence
of both sexual function and patient age on each of these parameters was then determined.
Age exerted a deleterious influence on all variables of penile blood flow independent of
the status of sexual potency. Patients under the age of 40 had a mean PBI of 0.99, compared
with a PBI of 0.74 for equally potent men over the age of 40. By contrast, impotent men over
the age of 40 had a mean PBI of 0.58 (Fig. 4).

1.2

-

1.1

1.0

"

<

' 0.99

09

0.8

401

n-9}

C

0.7

I

■ 0.7'

"to

0.6
0.5

0)

40(n=29

CL.

0.4
0.3
0.2

1

0.58

40(n = 63)

I
•5

Potent

Impotent

Figure 4 Distribution of penile brachial index (mean + SD) in patients by age and sexual potency.
(Data from Ref. 24.)

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

242 KEMPCZINSKI

The PVW of patients under the age of 40 was of good to fair quality, and no poor-
quality waveforms were observed. With increasing age and sexual dysfunction, a greater
percentage of patients had poor-quality waveforms, but this difference was not statistically
significant. The validity of the findings has been confirmed by numerous investigators,
who have emphasized the importance of this type of testing in the evaluation of patients
with erectile dysfunction (25,26). However, the diagnosis of vasculogenic impotence can-
not be established solely on the basis of such noninvasive measurements. Although mean
PBIs differed significantly between the three groups, there was a great deal of overlap, and
several patients were fully potent despite PBIs less than 0.6. Other researchers have
similarly confirmed a lack of correlation between PBI and the degree of erectile dys-
function (16). Although a low PBI is not sufficient to establish the diagnosis of vascu-
logenic impotence, the finding of a PBI greater than 0.8 confirms the adequacy of penile
blood flow and suggests that vasculogenic impotence is extremely unlikely.

DePalma et al. (27) have placed greater emphasis on the diagnostic importance of the
penile PVW. Using a pneumoplethysmographic cuff containing a pressure transducer,
they inflated the cuff to mean arterial pressure and recorded waveforms on a polygraph.
Waveform amplitude greater than 6 mm and a systolic upstroke rate of 4-6 mm at a speed
chart of 25 mm/s were considered normal. Marked flattening of the waveforms with delayed
upstroke greater than 6 mm and rounded waveforms were considered abnormal. Al-
though the investigators noted certain borderline categories in which diagnosis is equivo-
cal, the technique was particularly helpful in cases where the PBI was between 0.6 and 0.7.

Recently, Lue et al. (28) reported the use of duplex scanning in the evaluation of
vasculogenic impotence. Using a high-resolution 10 MHz ultrasound probe, they were able
to clearly visualize the cavernosa arteries, dorsal veins, tunica albuginea, corpora cavernosa,
and corpus spongiosum. The diameter of the arterial lumina, the thickness of the arterial
walls, and the quality of their pulsations were assessed before and after papaverine injection.
Pulsed Doppler was then used to study the blood flow through each of the penile arteries.
Since this test can be performed only on the vessels distal to the pubis, further visualization
of the pelvic vasculature with internal pudendal arteriography was required when ultra-
sonography suggested arterial disease.

C. Neurological Testing

Since there are no direct measures of the neural pathways involved in erection, indirect
measures must be employed. Fortunately, the autonomic pathways involved in micturition
and erection are similar and cystometrography with measurement of bladder capacity
and residual urine can be used as an indirect measure of penile innervation, assuming that
involvement of the appropriate pelvic nerves is reflected by abnormalities in both

areas. Using this technique, Ellenberg (17) confirmed neuropathy in 82% of impotent dia- ■§

betic subjects. g

The bulbocavernous reflex may be quantified by indirect measurement of pudendal »

nerve velocity (sacral latency testing). Since this examination requires electrical stimulation c

of the penile skin with simultaneous electromyographic recording of the response in the <

bulbocavernous muscle, it must generally be performed with the patient under general >9

anesthesia. The technique has been modified by using surface-mounted perineal electrodes, J

thus making measurement of somatosensory evoked potentials (SEPs) from the dorsal pe- «

nile and posterior tibial nerves more comfortable. Values that are three standard de- |

viations above the mean are considered abnormal (27). @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

IMPOTENCE AFTER AORTIC SURGERY 243

D. Intracavernous Papaverine Injection

This technique is useful in differentiating vasculogenic from psychogenic impotence (12).
However, it cannot distinguish psychogenic erectile dysfunction from neurogenic or
hormonal impotence. It should be used only to supplement a careful history and physical
examination, not to supplant it.

In patients with a penis of average size, 60 mg of papaverine diluted with 2-5 mL of
normal saline solution is injected into the corpus cavernosum. A rubber band is wrapped
tightly around the base of the penis before the injection to ensure that most of the drug
remains in the corpus, and it is left in place for 2 min after injection. The dose of papav-
erine may need to be adjusted in patients with an unusually large or small penis. In patients
suspected of neurogenic impotence, an initial test dose of 15 mg of papaverine should be
used, since they are prone to suffer priapism.

After the rubber band is removed, the patient is asked to stand so as to increase
venous pressure in the pelvis and further reduce the entry of papaverine into the systemic
circulation. If a full erection develops within 10 min and it lasts more than 30 min, the
arterial, venous, and sinusoidal mechanisms can be assumed to be normal and vasculo-
genic impotence can be excluded. However, since a full erection may not develop in a
nervous or anxious patient under the conditions of testing, a poor response does not in-
fallibly confirm vasculogenic impotence (12). When two or more injections fail to produce
an erection, an angiogram should be considered.

E. Angiography

The pelvic vasculature can be visualized using standard angiographic techniques with
appropriate oblique projections. This should be the first procedure performed when large-
vessel arterial occlusive disease or aortic aneurysm is suspected. Patency of the internal
iliac artery on each side should be determined and the presence of significant lesions
should be noted. Unfortunately, arteriographic findings correlate poorly with the patient's
erectile function. In one study where these were compared, 23% of potent men undergoing
aortic operation were noted to have bilateral iliac artery occlusions and an additional 36%
had unilateral occlusion (29). This is not surprising, since routine angiograms rarely
provide complete definition of the distal penile vasculature and cannot assess the adequacy
of collateral blood flow around arterial occlusive lesions.

When no flow-reducing lesions are identified in the hypogastric arteries or their major
branches, selective cannulation of the internal pudendal artery with the patient positioned
in the appropriate degree of obliquity may be necessary. Since selective cannulation of this
vessel may be difficult and the subsequent injection of dye painful, such studies are usually
performed with the patient under epidural anesthesia (30). Intra-arterial vasodilators
administered before the injection of contrast material are important for improving ■§

visualization of the penile arteries (Fig. 5). |

Corpus cavernosography, which can usually be performed under local anesthesia, may as

be used in the assessment of patients with erectile dysfunction, which is thought to be sec- c

ondary to venous outflow problems. However, pulsed Doppler sonography should routinely <

be performed in such patients, since only those with a normal sinusoidal system and arterial >9

tree will have a good arterial response to papaverine. If a patient does not achieve a full J

erection, the problem can be attributed to abnormal venous channels rather than to sinus- «

oidal fibrosis. In addition, patients with congenital or acquired chordee may require |

cavernosography (31). @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

244

KEMPCZINSKI

Greater sciatic
notch

Giutea! superior;
artery

Internal i'iat artery

ischiopudenoal trunk

Exit of *
Akock's
canal

srtiai spine '

;«hiai inferior

gluteal 1 artery ^ ^ t Jj^ ^f /~* . ^ Dorsal penile artery

Deep cavernosal artery

Pemle bulb and artery-
Branch o* superficial perineal
artery and scrotal arteries

Figure 5 Idealized normal subselective angiogram. Note filling of dorsal and deep penile arteries.
The arteries of the corpus spongiosum are not visualized. (From Ref. 27.)

F. Prevention

In order to ensure preservation of erectile function, surgical correction of aortoiliac
occlusive disease must accomplish the following objectives: minimal disturbance of genital
autonomic function, maintenance of adequate pelvic blood flow, and successful revascu-
larization of the ischemic extremity. DePalma (32) has popularized a nerve-sparing ap-
proach to the infrarenal aorta that emphasizes approaching the abdominal aorta along its
right lateral aspect, minimal division of longitudinal periaortic tissues to the left of the
infrarenal aorta, avoidance of dissection at the base of the inferior mesenteric artery, and
sparing of the nerve plexus that crosses the left common iliac artery (Fig. 6). Using such a
nerve-sparing approach, several surgeons have achieved a notable reduction in post-
operative impotence (29,33-35).

Although the findings on preoperative angiograms correlate poorly with erectile func-
tion, preservation of adequate perfusion into at least one hypogastric artery appears to be
a vital component of all operations that are successful in minimizing iatrogenic erectile
dysfunction. When possible, direct antegrade perfusion of the internal iliac artery should
be ensured. This may require thromboendarterectomy of the hypogastric artery orifice
when appropriate. If both external iliac arteries are occluded or stenotic and a bypass into
the common femoral arteries is anticipated, proximal aortic anastomosis should be
performed in an end-to-side fashion, since retrograde perfusion of the internal iliac artery
would be impossible in such circumstances and significant reduction of pelvic blood flow
would be likely. When proximal disease is so extensive that thromboendarterectomy is
impractical and preoperative noninvasive testing has confirmed decreased penile perfu-
sion, simple aortofemoral grafting may not always restore a pelvic collateral blood flow
adequate to relieve vasculogenic impotence. Preoperative recognition of such cases is not
easy. However, when the probability seems likely, the surgeon should consider reimplant-
ing the hypogastric artery into one limb of the aortofemoral graft or adding a jump graft
into the distal hypogastric artery (36).

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I
•5

I

IMPOTENCE AFTER AORTIC SURGERY

245

^ Intermesentenc

plexus

Inferior mesenteric

nerve

Superior
hypogastric plexus

Figure 6 Diagrammatic representation of the autonomic nerve supply to the penis (A) and
suggested modification of the surgical approach to the aorta during resection of abdominal aortic
aneurysm (B) to minimize damage to these structures. (From Ref. 35.)

In patients with unilateral iliac artery occlusive disease, the objectives of nerve-sparing
extremity revascularization and increased hypogastric artery perfusion may all be accom-
plished using femorofemoral bypass. Several investigators have confirmed the success of this
procedure in improving penile blood flow and restoring erection (25,37). Femorofemoral
bypass is especially appropriate for young, sexually active men with unilateral disease, since
it avoids the necessity for any periaortic dissection.

When direct aortic reconstruction is necessary, it is important to avoid flushing
atheromatous debris down the hypogastric artery (29). Operative techniques should be
modified to adequately back-bleed the hypogastric arteries before completion of anasto-
moses. Unfortunately, emergent aortic surgery, such as the resection of a ruptured abdom-
inal aortic aneurysm, rarely allows time for the careful anatomic dissection necessary to
avoid nerve damage, and the incidence of iatragenic impotence is accordingly higher (38).
Although rarely under the control of the vascular surgeon, emergent operations should be
avoided whenever possible.

The impact of endovascular revascularization or endovascular graft implantation in
aortoiliac revascularization on the incidence of iatrogenic impotence has not yet been
evaluated. However, the absense of direct dissection near the parasympathetic nerve plexus
surrounding the iliac arteries should have a favorable impact on reducing iatrogenic
neurogenic impotence. Furthermore, the ability to approach and dilate vessels more distally
located within the pelvis may make it possible to revascularize vessels previously unap-
proachable by traditional, direct vascular surgical techniques.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

246 KEMPCZINSKI

G. Treatment

Once organic impotence is confirmed, a reasonably precise etiological diagnosis is essential
before initiating appropriate therapy. If arteriograms confirm occlusion of the proximal
pelvic arteries and the measured PBI is less than 0.6, bypass into a distal patent branch of the
hypogastric artery on at least one side should be considered. If the PBI is less than 0.6 but no
large-vessel occlusive lesions are identified, selective angiograms of the internal pudendal
artery may document more distal occlusive lesions. If deep patent penile or dorsal penile
arteries can be confirmed, consideration should be given to direct revascularization of the
penis using the inferior epigastric artery and microvascular anastomosis into one of these
vessels. Although the long-term durability of such procedures has not been documented,
initial success has been reported in approximately 70% of such procedures. The recent
popularity of an endovascular approach to previously inaccessible vessels and balloon di-
lation, with or without the implantation of an endovascular stent, offers the potential for yet
another alternative approach to these problems.

When neurogenic impotence appears likely, patients may be managed by teaching them
to perform self-injection of intracorporal papaverine or by implantation of a suitable penile
prosthesis. The recently proven effectiveness of sildenafil citrate (Viagra®) in treating pa-
tients with psychogenic or neurogenic impotence suggests that this form of therapy should
be seriously considered in patients with postoperative impotence, especially if vascular
laboratory testing documents adequate penile blood flow. Furthermore, since the side effects
of Viagra® are minimal, it may be more appropriate to try an empiric course of therapy with
this medication when neurogenic impotence is suspected rather than resorting to the
complex diagnostic testing, i.e., cystometrography and/or sacral latency testing, which
would be required to confirm the diagnosis of neurogenic impotence. Similarly, even if
noninvasive vascular laboratory studies suggest borderline penile blood flow, an empiric
trial of Viagra® is so much simpler than complex arteriographic studies requiring selective
cannulation of the hypogastric arteries that it seems reasonable to consider such a course of
action prior to embarking on expensive, potentially harmful diagnostic testing.

In hypertensive patients who require ganglion-blocking drugs to control their disease
and in whom impotence develops secondary to such medications, alternate forms of treat-
ment should be found. If the hypertension is secondary to renal artery stenosis, renal re-
vascularization should be considered.

H. Results

In a collected series of 138 patients undergoing aortic reconstruction using standard vas-
cular surgical techniques, Flanigan et al. (29) documented a 25% incidence of iatrogenic
impotence. By minimizing periaortic dissection and emphasizing a nerve-sparing tech-
nique — along with efforts to ensure perfusion of at least one hypogastric artery during the ■a
arterial reconstruction — they completely eliminated iatrogenic impotence. Furthermore, |
retrograde ejaculation was reduced from 43% in the collected series to only 3% in the con- a
trol group. c
Since 80% of patients who present with aortoiliac arterial occlusive disease already <
have significant erectile dysfunction, careful planning of the operative reconstruction to >9
ensure hypogastric artery perfusion is essential if this symptom is to be relieved. Excluding J
patients with diabetes mellitus, in whom neurogenic impotence is most likely, relief of «
preoperative erectile dysfunction can be anticipated in 30% of patients so afflicted (15,29). |
Half of the patients who regained potency following revascularization were noted to have @
bilateral iliac artery occlusion on the preoperative arteriogram (29). %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

IMPOTENCE AFTER AORTIC SURGERY 247

IV. CONCLUSION

A mulitmodal diagnostic approach to the impotent patient is essential to ensure precise
etiological diagnosis and appropriate therapy. Furthermore, an understanding of the
multiple factors involved in achieving a normal erection is essential if potency is to be
preserved during the course of direct aortic surgery. Since nearly 25% of patients under-
going direct aortic reconstruction will suffer iatrogenic erectile dysfunction if appropriate
technical modifications are not employed, the problem of iatrogenic impotence is not
inconsequential.

When direct aortoiliac revascularization is necessary, appropriate nerve-sparing dis-
section of the infrarenal aorta and preservation of hypogastric perfusion can virtually
eliminate the postoperative development of erectile dysfunction and maximize the chance
for improvement in preoperative impotence. If the patient is an appropriate candidate for an
endovascular treatment of their arterial lesions, the danger of postoperative neurogenic
impotence can be virtually eliminated and the likelihood of improvement in preoperative
impotence maximized if preoperative noninvasive testing shows a low PBI and the proposed
endovascular procedure can restore flow into at least one of the hypogastric arteries.

When postoperative impotence occurs despite every precaution, an attentive and
objective approach can offer much comfort to the patient and his sexual partner. In most
cases, an empirical trial of Viagra® seems appropriate. In only a small percentage of such
cases will additional revascularization be necessary.

REFERENCES

1. Leriche R, Morel A. The syndrome of thrombotic obliteration of the aortic bifurcation. Ann
Surg 1948; 127:193.

2. Harris JD, Jepson RP. Aorto-iliac stenosis: A comparison of two procedures. Aust NZ J Surg
1965; 3:211.

3. DePalma RG, Kedia K, Persky L. Vascular operations for preservation of sexual function. In:
Bergan JJ, Yao JST, eds. Surgery of the Aorta and Its Body Branches. Orlando, FL: Grune &
Stratton, 1979.

4. Queral LA, Flinn WR, Bergan JJ, Yao JST. Sexual function and aortic surgery. In: Bergan JJ,
Yao JST, eds. Surgery of the Aorta and Its Body Branches. Orlando, FL: Grune & Stratton,
1979.

5. Weiss HD. The physiology of human penile erection. Ann Intern Med 1972; 76:793.

6. Kedia KR, Markland C, Fraley EE. Sexual function following high retroperitoneal lym-
phadenectomy. J Urol 1975; 114:237.

7. Drane RJ, Siroky MB. Neurophysiology of erection. Urol Clin North Am 1981; 8:91.

8. Willis EM, Ottesen B, Wagner G, Sundler F, Fahrenkrug J. Vasoactive intestinal polypeptide
(VIP) as a possible neurotransmitter involved in penile erection. Acta Physiol Scand 1 98 1 ; 1 1 3: 545 .

9. Adaikan PG, Kottegoda SR, Ratnam SS. Is vasoactive intestinal polypeptide the principal
transmitter involved in human penile erection? J Urol 1986; 135:638.

10. Benson GS. Penile erection: In search of a neurotransmitter. World J Urol 1983; 1:209. 1)

1 1 . Conti G. L'erection due penis humain et ses bases morphologic vascularies. Acta Anat (Basel) =5
1952; 14-17. g

12. Lue TF, Tanagho EA. Physiology of erection and pharmacological management of impotence. g
J Urol 1987; 137:829. 1

13. Newman HF, Northrup JD, Devlin J. Mechanism of human penile erection. Invest Urol 1964; •%
1:350. |

14. Michal V, Pospichal J. Phalloarteriography in the diagnosis of erectile impotence. World J Surg ©
1978; 2:329. ■&

I

s

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

248 KEMPCZINSKI

15. May AG, DeWesse JA, Rob CG. Changes in sexual function following operation on the
abdominal aorta. Surgery 1969; 65:41.

16. Nath RL, Menzoian JO, Kaplan KH, McMillian TN, Siroky MB, Krane RJ. The multi-
disciplinary approach to vasculogenic impotence. Surgery 1981; 89:124.

17. Ellenberg M. Impotence in diabetes: The neurologic factor. Ann Intern Med 1971; 75:213.

18. Bohlen JG. Sleep erection monitoring in the evaluation of male erectile failure. Urol Clin North
Am 1981; 8:119.

19. Karacan I, Salis PJ, Ware JC, Dervent B, Williams RL, Scott FB, Attia SL, Beutler LE. Nocturnal
penile tumescence and diagnosis in diabetic impotence. Am J Psychiatry 1978; 135:191.

20. Canning JR, Bowers LM, Lloyd FA, Cottrell TLC. Genital vascular insufficiency and impotence.
Surg Forum 1963; 14:298.

21. Abelson D. Diagnostic value of the penile pulse and blood pressure: A Doppler study of im-
potence in diabetics. J Urol 1975; 111:636.

22. Britt DB, Kemmerer WT, Robison JR. Penile blood flow determination by mercury strain
gauge plethysmography. Invest Urol 1971; 8:673.

23. Gaskell P. The importance of penile blood pressure in cases of impotence. Can Med Assoc J
1971; 105:1047.

24. Kempczinski RF. Role of the vascular diagnostic laboratory in the evaluation of male impotence.
Am J Surg 1979; 138:278.

25. Merchant RF Jr, DePalma RG. Effects of femorofemoral grafts on postoperative sexual
function: Correlation with penile pulse volume recordings. Surgery 1981; 90:962.

26. Queral LA, Whitehouse WM, Flinn WR, Zarins CK, Bergan JJ, Yao JST. Pelvic hemodynamics
after aortoiliac reconstruction. Surgery 1979; 86:799.

27. DePalma RG, Emsellem HA, Edwards CM, Druy EM, Shultz SW, Miller HC, Bergsrud D. A
screening sequence for vasculogenic impotence. J Vase Surg 1987; 5:228.

28. Lue TF, Hricak H, Marick KW, Tangaho EA. Vasclogenic impotence evaluated by high-
resolution ultrasonography and pulsed Doppler spectrum analysis. Radiology 1985; 155:777.

29. Flanigan DP, Schuler JJ, Keifer T, Schwartz JA, Lim LT. Elimination of iatrogenic impotence
and improvement of sexual function after aorto-iliac revascularization. Arch Surg 1982; 1 17:544.

30. Ginestie JF, Romieu A. Radiologic Exploration of Impotence. The Hague: Martinus Nijhoff,
1978.

31. Fitzpatrick T. The corpus carvenosum intercommunicating venous drainage system. J Urol
1975; 113:494.

32. DePalma RG. Impotence in vascular disease: Relationship of vascular surgery. Br J Surg 1982;
69:514.

33. DePalma RG, Levine SB, Feldman S. Preservation of erectile function after aortoiliac recon-
struction. Arch Surg 1978; 113:958.

34. Miles JR, Miles DG, Johnson G. Aortoiliac operations and sexual dysfunction. Arch Surg
1982; 117:1177.

35. WeinsteinMH, MachlederHI. Sexual function after aorto-iliac surgery. Ann Surg 1975; 181:787.

36. Biller A, Dagher FJ, Queral LA. Surgical correction of vasculogenic impotence in a patient
after bilateral renal transplantation. Surgery 1982; 91:108.

37. Schuler JJ, Gray B, Flanigan DP, Williams LR. Increased penile perfusion and reversal of j>
vasculogenic impotence following femorofemoral bypass. Br J Surg 1982; 67-69. <S

38. Flanigan DP, Pratt DG, Goodreau JJ, Burnham SJ, Yao JST, Bergan JJ. Hemodynamic and js
angiographic guidelines in selection of patients for femorofemoral bypass. Arch Surg 1978; °|
113:1257. ^

39. Michal V, Kramar R, Hejhal L. Revascularization procedure of the cavernous bodies. In: «
Zorgomotti AW, Rossi G, eds. Vasculogenic Impotence Proceedings of the First International J
Conference on Corpus Cavernosum Revascularization. Springfield, IL: Charles C. Thomas, a
1980. |

40. Kempczinski FR, Birinyi LK. Impotence following aortic surgery. In: Bernhard VM, Towne 2
JB, eds. Complications in Vascular Surgery. 2d ed. Philadelphia: Saunders, 1985:311-324. «

I
§

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

12

Complications Following Reconstructions of the
Pararenal Aorta and Its Branches

Kenneth J. Cherry, Jr.

Mayo Clinic, Rochester, Minnesota, U.S.A.

I. INTRODUCTION

Operations upon the juxta- and suprarenal aorta and the visceral branches of the upper
abdominal aorta are performed much less frequently than reconstructions confined to the
infrarenal aorta and its pelvic and extremity branches. There is evidence that more
juxtarenal aneurysms are being diagnosed and repaired in absolute numbers (1). The
sophistication of modern imaging modalities allows more detailed and easier assessment
of the paravisceral aorta than in years past, and there is more of a trend to longitudinal
follow-up by vascular surgeons now than previously. In addition, endovascular treatment of
infrarenal abdominal aortic aneurysms is currently performed for anywhere from 40-80%
of patients undergoing aneurysm repair at referral centers. A larger percentage of the aortic
aneurysms now repaired conventionally are juxta- or suprarenal. In the recent series from
the Cleveland Clinic, the repair of juxtarenal aneurysms accounted for 10.8% of their total
repairs in 1995 and 31.7% in 2000 (2). This relative increase in juxtarenal reconstructions
reflects the active endovascular practice of those surgeons for the treatment of infrarenal
aortic aneurysms.

Furthermore, failed or failing stent-graft repairs of the infrarenal aorta are frequently
necessarily managed with open repair: suprarenal, supramesenteric, or supraceliac clamping ■g

is often required. &

Endovascular angioplasty and stenting of renal artery lesions has supplanted conven- a

tional surgery as the major interventional modality to treat renovascular hypertension for c

both atherosclerotic and fibromuscular dysplastic lesions around the country. Nonetheless, <j

there exist renal artery lesions that are not well treated by endovascular techniques: >9

atherosclerosis or fibromuscular dysplasia involving major branch vessels, atherosclerotic 41

lesions of extended length, extremely distal lesions or branch vessel lesions, eccentric calcific 2

renal artery plaque, dense aortic calcific disease extending into the renal artery, multiple |

small stenotic renal arteries not suitable for stenting, aneurysms involving large branch @

249 I

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

250 CHERRY

vessels or in distal locations, and complications of endovascular therapy, such as throm-
bosis, dissection, or perforation. Not all of these require open repair, but many do.

Angioplasty and stenting of the mesenteric arteries has its proponents and enthusiasts,
some of whom claim that open surgery is passe. Short segment stenoses of the superior
mesenteric artery are well treated by this modality. Stenoses of the celiac artery are more
difficult to treat because of the early branching of that vessel, its often short main trunk, and
the overlying median arcuate ligament. A "bald aorta" with no demonstrable celiac or
superior mesenteric artery ostia, as well as long occlusions of the superior mesenteric artery
with reconstitution of the vessel at the infrapancreatic position, are poorly treated by
endovascular technique. Furthermore, there are no good data concerning medium- or long-
range follow-up for angioplasty and stenting, and that lack of data must temper the
enthusiasm for this new modality. Kasirajan et al. at the Cleveland Clinic had a higher
recurrence at 3 years in their patients undergoing endovascular repair than in those having
open reconstruction; they felt good-risk patients should be offered open repair as a first
option (3). In our series of 98 patients, we identified a subset of poor-risk patients greater
than 70 years of age who might have been better served by endovascular repair than by open
surgery; we felt that the majority were better treated with conventional reconstruction (4).
As endovascular technology improves, undoubtedly the percentage of patients with chronic
mesenteric ischemia needing conventional operations will decrease. At present, however, the
ability to perform conventional operation for chronic mesenteric ischemia remains a
necessary part of our education and skill.

Endovascular repair of aortic aneurysmal disease involving the juxta- and suprarenal
aorta, or the necessity of providing safe attachment sites proximal to the renal arteries, is in
its nascent period. Surgical investigators such as John Anderson in Adelaide, Timothy
Chuter in San Francisco, and Larry Hollier and Michael Marin in New York are leading the
way, making ingenious advances in applying endovascular techniques to the upper
abdominal aorta and its branches (5-8). For most patients, however, juxta- and suprarenal
aortic aneurysmal and occlusive disease remain open surgical problems.

Reconstructions of the para visceral aorta and its visceral branches may be daunting not
only because of the relative rarity with which they are performed but also because of the
anatomical milieu. There is close proximity, indeed intimacy, of the upper abdominal aorta
and its branches with the diaphragm, esophagus, stomach, vagus nerves, spleen, liver, biliary
tree, pancreas, proximal small bowel, adrenal glands, and kidneys. Dissection, exposure,
and repair are all the more challenging. The choice of incision, choice of approach, site of
clamp placement, and proper sequence of clamping and unclamping all contribute heavily to
the success or failure of these operations. The need to clamp above the renal arteries or above
the mesenteric vessel(s) adds an element of ischemic insult to the solid organs, especially the
kidneys, not present with infrarenal aortic repairs. Juxtarenal clamping is also a potential
source of renal atheroemboli. Upper abdominal aortic clamping adds further cardiac stress.
Pulmonary function may likewise be compromised.

II. OPERATIONS ON THE JUXTA- AND SUPRARENAL AORTA

Operations on the juxta- and suprarenal aorta may be performed utilizing various incisions,
approaches, and clamp sites. Classically, type IV thoracoabdominal aortic aneurysm
repairs, as well as endarterectomy of the paravisceral aorta and/or mesenteric vessels, were
approached through low-lying thoracoabdominal or thoracoretroperitoneal incisions.
Those approaches, involving not only abdominal incisions but also incisions of the
diaphragm and thoracic cavity were and are occasioned by increased pulmonary and

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i
2

I

■c

I

u

COMPLICATIONS AFTER AORTIC RECONSTRUCTION 251

wound problems (9). There are patients, however, who still require the more extensive
thoracoretroperitoneal method. Extremely thin or asthenic persons with very narrow costal
margins are poorly served by laparotomy and medial visceral rotation, as the ipsilateral
lower rib cage prevents the surgeon from obtaining a proper orientation to deal with the
upper extent of the aorta or with the mesenteric vessels. At the other extreme, morbidly
obese patients or extremely large patients may require thoracoretroperitoneal exposure
simply for adequate and comfortable handling of the arteries.

Those anatomical considerations aside, most surgeons dealing with these problems
prefer, if possible, to use incisions confined to the abdomen, combined with transperitoneal
approaches, retroperitoneal approaches, or medial visceral rotation. Those incisions may be
midline, transverse, oblique, or combinations thereof.

The operative results for pararenal aneurysm repair at large referral centers are good to
excellent, with mortality ranging from to 15.4% and averaging approximately 5%
(Table 1) (2,10-19). In a small series, not included in Table 1, Tordoiret al. had no mortality
in their 15 patients with pararenal aortic aneurysms (20). Similarly, Schneider et al., in their
series of 23 patients, reported no mortality (17). Qvarfordt et al. had a 1.3% mortality for 77
patients, an even more salutary figure when one considers that 70% of those patients had
concomitant renal artery reconstructions (1 1). Allen and colleagues in a large series from the
Washington University-Barnes Hospital had a mortality of 1.5% in their patients and
likewise reconstructed the renal arteries (15). Jean-Claude et al., continuing the work from
the University of California-San Francisco, reported a mortality of 5.8% in 257 patients
(19). The most recent report is from Sarac and colleagues at the Cleveland Clinic, concerning
138 patients with juxtrarenal aneurysms, has a very similar mortality of 5.1%. This was
statistically different from the results of their open infrarenal aortic reconstructions (2.8%)
(2).

In two other series, not included in Table 1 because of the range of patients analyzed,
Shepard et al., at the New England Medical Center, used an extended left flank incision and
retroperitoneal exposure in 23 high-risk patients with abdominal aortic aneurysms, 14 of
whom required suprarenal and supraceliac clamping for pararenal occlusive and aneu-
rysmal problems. There was only one death (4%) (21). Shepard and coworkers, continuing
their work at Henry Ford Hospital, reported the same approach for 85 patients with
complex occlusive and aneurysmal aortic problems. The elective mortality in their series was
2.4% (22).

The two large series with the best mortality, those of Qvarfordt et al. and Allen et al., are
remarkable not only for the low mortality reported but also for the surgeons' willingness and
ability to clamp at the supramesenteric level in addition to the suprarenal and supraceliac
levels. Neither study found any increase in morbidity or mortality with that approach
(11,15).

The abdominal incisions used are, in the main, surgeons' choices. Body habitus, •§

previous operation, comorbidities, (e.g., pulmonary disease), and the segment of aorta g

needing repair all impact upon those decisions. Exposure may be infracolic transperitoneal a

for juxta- and even suprarenal aortic reconstructions and for renal artery repairs. It may be c

transperitoneal through the lesser omentum for mesenteric reconstructions. It may be <j

retroperitoneal. Darling et al. report few if any problems dealing with even the right renal >3

artery utilizing a retroperitoneal approach (23). That view is shared by the group from 4j

Henry Ford (22). For more proximal repairs, medial visceral rotation allows exposure of the 2

entire abdominal aorta; this is our preferred approach for upper abdominal operations. The |

left kidney may be mobilized anteriorly or left in situ, depending on the work to be done and ©

the anatomy. If superior mesenteric artery reconstruction is to be performed, the kidney is %

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

252

CHERRY

Table 1 Literature Summary' 1

AAA

Cross-

No. of

Type 11

Clamp

Mortality

Baseline

Study

cases

(no.)

Level

Rate

Elevated Cr

Crawford

101

JR 88

SC 93

7.9%

18.8%

et al. (10),

SR0

SSMAO

1986

RAOD 13

SR 8

Qvarfordt.

77

JR22

SC 13

1.3%

54.5%

(11), 1986

SR24
RAOD 31

SSMA 17
SR45

Green et al.

52

JR29

SC 30

15.4%

ND

(12), 1989 b

SR?
RAOD?

SSMAO
SR22

Poulias et al.

38

JR 32

SCO

5.3%

15.8%

(13), 1992

SR0
RAOD 6

SSMAO
SR 38

Breekwoldt

39

Unclear

SC8

2.6%

ND

et al. (14),

SSMA 2

1992 c

SR25

Allen et al.

65

JR24

SC27

1.5%

20.0%

(15), 1993 d

SR 15
RAOD 7

SSMA 12
SR26

Nypaver et al.

53

JR41

SC21

3.8%

17.0%

(16), 1993

SR6
RAOD 6

SSMA 4
SR28

Schneider et

23

SR23

SC23

0.0%

ND

al (17),

SR0

SSMAO

1997

RAOD

SR0

Faggioli et al.

50

JR 39

SC 8

7.0%

10.0%

(18), 1998 e

SR6

SSMAO

(elective

RAOD 5

SR42

only)

Jean-Claude

257

JR 122

SR42

5.8%

31.1%

et al. (19),

SR 58

SC 33

1999

RAOD 77

SSMA 48

Sarac et al.

138

JRA 138

SC43

5.1%

15.4%

(2), 2002

SR95

New- Elevated

Transient Onset Cr at

Cr Rise Dialysis Discharge

15.8% 7.9% ND

23.0% 2.5% 13.0%

ND 11.5% ND

23.7% 13.2% 13.2%

28.2% 2.6% 12.8%

12.3% 3.1% 3.1%

22.6% 5.7% 7.5%

26.1% 0.0% 0.0%

ND ND

0.0%

30.4% 7.0% 10.5%

19.6% 28.3% 5.8%

Abbreviations: AAA, abdominal aortic aneurysm; Cr, creatinine level; JR, juxtarenal; SR, suprarenal; RAOD, renal

artery occlusive disease; SC, superceliac; SSMA, supra-superior mesenteric artery; ND, no clear data.

a Best estimate of categorization University of California-San Francisco pararenal abdominal aortic aneurysm groups.

Very difficult to put the data in this study into this format; also suprarenal clamp group contained 1 1 patients who
initially had an infrarenal clamp.
c Only 33 abdominal aortic aneurysms matched University of California-San Francisco categories.

Only 46 abdominal aortic aneurysms matched University of California-San Francisco categories.
e Includes seven ruptured pararenal abdominal aortic aneurysms.
Source: Ref. 19.

i
2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

COMPLICATIONS AFTER AORTIC RECONSTRUCTION 253

better left in situ. Injury to the spleen and kidneys is more likely with upper retroperitoneal
exposure and medial visceral rotation than with a transperitoneal approach (9,23).

All three methods of exposure — transperitoneal, retroperitoneal, and medical visceral
rotation — offer excellent choices for the correct patient populations. All are subject to
complication. Thoracoabdominal incisions and oblique flank incisions, in distinction to
classic laparotomy incisions, are subject to flank herniation or laxity of the flank muscles
because of division of the nerves. There is also the potential problem of costochondral pain
or instability (9,22).

Of considerable import and debate is the choice of the proximal aortic clamp site. It is on
this issue that success or failure hinges. Crawford preferred supraceliac clamping and used it
almost exclusively (10). As stated previously, the surgeons at UCSF and at Washington
University-Barnes applied clamps at suprarenal, supramesenteric and supraceliac sites
(11,15). Green et al. from the University of Rochester have recommended against suprarenal
clamps (12). That group reported a higher mortality with suprarenal clamps, 32%, in
comparison to a 3% mortality for supraceliac clamping and a sevenfold increase in frank
renal failure with suprarenal clamping. Of note, 1 1 patients included in that suprarenal
group were initially clamped at an infrarenal level and the clamp level changed during the
course of their operations. That fact underscores the absolute necessity for clear and precise
judgment in choosing the clamp site and technical excellence in exposing the proper aortic
segment in a thorough, gentle manner. Renal artery atheroembolization is the most feared
sequela of dissection and exposure and of improper clamp placement; it was felt by Green et
al. to account for their poor results with suprarenal clamping. They felt that nonaneurysmal
atherosclerosis in the pararenal area was the culprit. Of note again, the authors stated: "The
placement of the proximal clamp was dependent on the operative findings [italics added]." I
would maintain that the route of exposure and the choice of proximal clamp site is depend-
ent on appropriate and accurate preoperative imaging and assessment rather than on oper-
ative findings. Nonetheless, the concern of Green et al. for atheroembolization is valid and is
reinforced in an unexpected way by the most recent finding from Jean-Claude et al. from the
University of California-San Francisco, in which 6 of their 15 deaths (40%) were attributed
to visceral ischemia or infarction. Those authors felt that atheroembolization was the
mechanism of injury in 5 of those 6 patients (19). Parenthetically, they have not experienced
this particular complication in the last 10 years (Louis Messina, personal communication).
Further, and in contradistinction to Green's findings, they found that atheroembolization
was more likely with supraceliac clamping or supramesenteric clamping than with supra-
renal clamping (9.1, 6.3, and 0.6% respectively). Similarly, Sarac and colleagues from the
Cleveland Clinic found supraceliac clamping less safe than suprarenal clamping (2).

Cardiac disease, of course, remains a primary cause of perioperative morbidity and
mortality for all vascular patients. Preoperative screening with physiological and anatom-
ical tests; appropriate medical or interventional management; and maximization of intra- •§
operative cardiac status with appropriate pharmacological, anesthetic, and monitoring g
methods are all paramount to achieving excellent results (24). This chapter does not address a
that issue further with the exception of cardiac function and site of clamp placement. c

%

III. COMPLICATIONS %

I
In most series, renal insufficiency is the leading complication of juxtarenal aortic operations; Q

it is certainly the most analyzed and the most feared, as evidenced by Table 1. It was the |

leading complication in the majority of the series cited with transient rises of serum @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

254 CHERRY

creatinine in 12.3-30.4% of patients and led to the need for new dialysis in 0-28.3% of
patients (2,10-19). Preoperative renal dysfunction is felt by most authors to be the most
accurate indicator of postoperative decline in renal function (2,15,19), but other authors
have not been able to demonstrate that association (14,16). Nypaver et al. felt that
postoperative renal insufficiency was more likely when concomitant renal artery operations
were necessary or when there was a major intraoperative complication (16). Allen and
colleagues at Washington University-Barnes Hospital, on the other hand, felt that attention
to renal artery disease improved their results (15). Authors such as Green et al. have
implicated suprarenal clamp placement and embolization. Whereas others, such as Allen et
al., have found no correlation between clamp placement and complications (12,15). Jean-
Claude et al. and Sarac and colleagues at the Cleveland Clinic have implicated supraceliac
clamping as the more hazardous maneuver (2,19). Jean-Claude and colleagues also found
statistical correlation with cross-clamp time. As 75% of their patients improved, they felt
that acute tubular necrosis secondary to cross-clamp-induced ischemia — and not atheroem-
bolization — was the disease mechanism (19). The group from Barnes found no correlation
between cross-clamp time and renal insufficiency. They felt that accurate aortic cross-
clamping, attention to renovascular disease, and cold perfusion of the renal arteries obviated
the deleterious ischemic effects of cross-clamping (15).

Shepard et al. and Poulios and colleagues also use hypothermic renal perfusion, as does
the group at Barnes (13,15,22). We also prefer its use. Some surgeons at our institution use a
continuous drip, while others perfuse the kidneys with boluses of ice-cold heparinized saline
on a periodic basis, such as every 15 min.

Kashyap et al. at Massachusettes General Hospital described renal failure in 183 pa-
tients following thoracoabdominal aortic surgery; of these, 29 had type IV thoracoabdom-
inal aortic aneurysms, 17 had suprarenal aortic aneurysms and 12 underwent renal or
mesenteric reconstructions (25). Acute renal failure was associated with a preoperative
creatinine of 1.5 or higher and a cross-clamp time exceeding 100 min; the complication of
acute renal failure was associated with a 10-fold increase in mortality. One-quarter of their
patients with preoperative renal insufficiency developed renal failure, but one must re-
member that this group of 185 patients included 125 patients with type I, II, or III thoraco-
abdominal aortic aneurysms — groups that are at higher risk for this complication and for
death. However, Sarac and colleagues, on the same note, found that supraceliac clamping
was the sole predictor of postoperative mortality and that supraceliac clamping, diabetes,
and preoperative renal insufficiency were the predictors of postoperative renal insufficiency
(2).

Renal dysfunction may result from prolonged ischemia, atheroembolization, renal
artery occlusive disease, artery or graft thrombosis, hypovolemia, hypoperfusion, hyper-
perfusion, shock, or multisystem organ failure, with its cytokine response (26). Acute renal
failure is exacerbated by proximal aortic repair, preoperative renal insufficiency, intra- •§

operative complications, and comorbidities. Renal dysfunction is also related to age, cardiac g

dysfunction, and diabetes. Cherr and Hansen feel that the recovery of renal function is a

dependent on preoperative renal function, age, and cross-clamp time (27). -c

When all these studies with their diverse findings are considered, it would seem that <

renal ischemic issues relate primarily to the patient's preexistent renal and cardiac status, the &

presence of renovascular disease, placement of the cross-clamp, clamping sequence, cross- 4j

clamp time, and the patient's hemodynamic status intraoperatively. Hydration, before 2

angiography and in the preoperative period as well as intraoperatively, with determination |

and maintenance of the patient's best hemodynamic and cardiac function, and appropriate @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AFTER AORTIC RECONSTRUCTION 255

fluid and blood replacement during the operation are paramount in lessening the chances of
renal insufficiency. Pulmonary artery catheters should be used in all patients undergoing
upper abdominal and visceral reconstructions. Transesophageal echocardiography may be
added to those patients with known cardiac disease. Lasix and mannitol are used. Mannitol
is given not only for its osmotic diuresis but also to decrease renovascular resistance and
provide free radical scavenging. It also increases the glomerular filtration rate during
periods of hypoperfusion (27). Vasodilators are given prior to clamping. Most surgeons,
ourselves included, favor the use of intraoperative dopamine in low doses for its dopamin-
ergic type 1 effects. Fenoldopam, if available, may be a better drug, as it has no dopa-
minergic type 2 effects and is known to be protective of the kidney. Heparin is, of course,
used for these repairs. With careful dissection and exposure, judicious clamp placement, and
an expeditious, well-planned reconstruction, the ischemic insults may be minimized.
Further, if prolonged renal ischemic times are anticipated, ice slush may be placed about
the kidney. Cold perfusate is used by many authors when possible, as stated above. The
visceral vessels may be clamped prior to applying the aortic clamp so as to minimize
the chances of atheroembolization, either temporarily if the clamp is juxtarenal or for the
formal reconstruction.

In regard to intraoperative cardiac function, there may be real benefit in applying a
supramesenteric rather than a supraceliac clamp if the chosen approach and the aortic and
branch vessel anatomy will allow. Jean-Claude et al. have pointed out that they had only one
death from myocardial infarction (MI) in their series of 257 patients (19). However, 5.8% of
their patients sustained an MI. In analyzing the previous papers on pararenal aneurysm
repair, they found that MI accounted for 32% of the deaths in those series. In addition to the
usual preoperative cardiac assessment and intraoperative management, those authors felt
that vasodilation before placement of the aortic cross clamp was important and helped
reduced cardiac-related deaths and that the limited use of supraceliac cross-clamping,
confined to 13% of their patients, was beneficial to the group. The increased stress on the
heart from supraceliac clamping is mitigated in part by the application of a supramesenteric
clamp instead, allowing afterload reduction through the mesenteric circulation (28).

Pulmonary dysfunction is another frequent and serious complication of surgery
involving the upper abdominal aorta or aortic branch. Messina of the University of Cali-
fornia-San Francisco feels it is the greatest risk facing these patients with pararenal
aneurysms (Louis Messina, personal communication). In Reilly et al.'s review from that
institution of their 108 operations (87 elective) utilizing medial visceral rotation, 31% of the
patients had postoperative pulmonary problems and respiratory failure (9). In the follow-up
study by Jean-Claude et al., pulmonary complications — respiratory insufficiency and
pneumonia — were the most frequent complications, occurring in 14.4% of patients. Pneu-
monia was the leading complication in the series of Breckwoldt et al. (14), and ventilator
dependence for greater than 48 h was the number one complication in the series from •§

Washington University (15). Nypaver et al. reported that it was the primary complication in g

their patients (16). In our review of 98 patients undergoing open revascularization for a

chronic mesenteric ischemia, the number one complication was ventilator dependence or c

tracheostomy; it occurred in 6 patients (4). <j

Pulmonary insufficiency is much more likely with a thoracoabdominal incision than it is >9

with a purely abdominal approach (9). Patients known preoperatively to have chronic 4j

obstructive pulmonary disease (COPD) may be pretreated with mechanical pulmonary 2

toilet and/or steroids if the latter have been shown to be helpful on preoperative pulmonary |

function testing. Transverse abdominal incisions, postoperative epidural analgesia, and ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

256 CHERRY

retroperitoneal exposures are all thought to be of benefit to patients with COPD and are
used preferentially for patients with known pulmonary compromise at our institution.

Gastrointestinal complications after upper abdominal aortic surgery occur less fre-
quently than either renal or pulmonary complications. One notable exception is documented
in the report from Jean-Claude et al. in their review of 257 patients undergoing repair of
juxtarenal aneurysms, suprarenal aneurysms, or aneurysms associated with renal artery
occlusive disease (19). Visceral ischemia or infarction was the cause of death in 6 (40%) of
the 15 patients dying in the postoperative period. Atheroembolization was felt to be the
mechanism of injury in 5 and chronic, unoperated mesenteric occlusive disease caused the
other death. Valentine et al., at Parkland, studied the problem of gastrointestinal compli-
cations in 120 patients following aortic surgery (29). All of their patients were approached
transperitoneally. Gastrointestinal complications occurred in 25 (21%), the most frequent
problem being ileus requiring replacement of a nasogastric tube. The occurrence of gastro-
intestinal complications was associated with intraoperative complications, greater than
normal blood loss, and the necessity for more fluid resuscitation. There was an increased
prevalence of pulmonary and renal complications in these same patients.

That same group looked at pancreatitis following aortic reconstructions (30). They
found that 1.8% of their patients undergoing aortic reconstructions of all types experienced
pancreatitis. None died from pancreatitis or its complications per se, and those authors felt
this to be a rare, self-limited, and seldom serious complication in their patients following
aortic surgery. The great majority of their patients were clamped infrarenally. Four patients
did develop complications of pancreatitis, with multiple system organ failure in three and
pseudocyst in one. All four of those patients had undergone suprarenal or higher clamping. In
keeping with this fact, Reilly at al. reported pancreatitis in 6 (6.8%) of their 88 patients
undergoing medial visceral rotation and suprarenal clamping, and 2 of those 6 died of
pancreatitis or its complications (9). It is probable that the combination of ischemia and
local trauma attendant on medial visceral rotation (from dissection, mobilization, and
retraction) is responsible for that difference. Great care must be taken with the pancreas no
matter what approach is used.

Splenic injury with its usual sequela, splenectomy, is reported for both retroperitoneal
exposure and medial visceral rotation (9,23). Anecdotally, it occurs occasionally with a
transperitoneal approach but not nearly so frequently.

End-organ failure, including liver failure, has been reported with supraceliac clamping
(31,32). It relates directly to ischemia time. Patients with hepatic dysfunction or a history of
cirrhosis or hepatitis are particularly susceptible to ischemic liver failure. Selective mesen-
teric shunting is appropriate for patients with types I to III thoracoabdominal aortic aneu-
rysms but is not entirely applicable for patients with pararenal aneurysms (33). Ballard has
developed a trifurcated graft technique, with separate grafts to branch vessels in addition
to the aortic replacement graft, to reduce ischemic time (34). His principle may be useful •§

for patients with known liver or renal dysfunction if prolonged supraceliac clamping is g

anticipated. a

Pseudoaneurysm, or anastomotic aneurysm, is the most common long-term problem c

with aortic grafts, occurring in approximately 4-10% of patients (35). Its exact prevalence is <j

not known. When an aggressive follow-up protocol utilizing new imaging modalities is >9

utilized, recurrence may be higher than has been suspected. %

Patients having had thoracic aneurysms repaired are much more likely to develop a Q

subsequent true or false aneurysm than patients whose first operation was for infrarenal |

aortic aneurysm (35). Whether patients with juxta- or suprarenal aneurysms are more prone @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AFTER AORTIC RECONSTRUCTION 257

to develop pseudoaneurysms than patients with infrarenal aneurysms is not known.
Obviously, follow-up for grafts with proximal anastomoses above the level of the renal
arteries requires serial computed tomography rather than ultrasound.

Hallett et al., in a population-based study of 307 patients from Olmsted County,
Minnesota, having abdominal aortic aneurysm repairs, found graft-related complications in
29 (9.4%) (36). Their study extended out to 36 years with an average follow-up of 5.8 years.
Anastomotic pseudoaneurysms occurred in 3%, graft thrombosis in 2%, graft enteric
fistulas in 1.6%, infection in 1.3% and other complications in lesser numbers. Biancari et al.
from Finland performed a retrospective study of 208 abdominal aortic aneurysm patients
operated upon at one hospital and found that there were late graft complications in a very
similar 15.4% of their patients (37). Para-aortic anastomic aneurysms developed 2.9%,
more distal aneurysms in 8.7%, and bilateral or recurring aneurysms in 3.4%. Their follow-
up ranged from 0.1-21.7 years with a median of 8.0 years.

The repair of recurrent juxtarenal aneurysms is attended by higher morbidity and
mortality than is primary reconstruction (35).

IV. CONCLUSIONS

Successful operative management of the pararenal aorta and the visceral branches requires
thoughtful preoperative planning and meticulous operative technique to reduce the inherent
morbidity and mortality to the minimum. Accurate preoperative assessment of the quality
of the aorta, especially of the extent and pattern of atherosclerosis in the paravisceral area on
the back wall, is probably the single most important step in providing a safe and effective
operation upon the pararenal aorta. Lateral views of the aorta, by either conventional
angiography or high-resolution reconstructed computed tomography, are mandatory. In
closing, I can do no better than to quote as follows:

1. Careful consideration of the route of exposure, location of the proximal aortic
clamp, and the preservation of renal function with renal hypothermia and with
the repair of significant renal artery lesions will result in minimal morbidity and
mortality in patients requiring surgery for juxtarenal with suprarenal abdominal
aortic aneurysms (15).

2. Accurate preoperative assessment of the amount of disease in the aorta at the
planned level of cross-clamping, correct selection of the optimal cross-clamp level,
selection of the optimal approach for the needed exposure, and following the
proper clamping and declamping sequence for the aorta and visceral branches are
all important factors in reducing the incidence rate of atheroembolization (19).

3. It is the patient whose cross-clamp level is incorrectly chosen who has the highest

likelihood of significant complications, usually atheroembolic and often fatal (19). -o

&

2

REFERENCES I

1. Taylor SM, Mills JL, Fujitani RM. The juxtarenal abdominal aortic aneurysm. Arch Surg u
1994; 129:734-737. |

2. Sarac TP, Clair DG, Hertzer NR, et al. Contemporary results of juxtarenal aneurysm repair. J S
Vase Surg 2002; 36:1104-1111. ^

3. Kasirajan K, O'Hara PG, Gray BH, et al. Chronic mesenteric ischemia: Open surgery versus I
percutaneous angioplasty and stenting. J Vase Surg 2001; 33:63-67. q

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

258 CHERRY

4. Park WM, Cherry KJ Jr, Chua HK, et al. Current results of open revascularization for chronic
mesenteric ischemia: A standard for comparison. J Vase Surg 2002; 35:853-859.

5. Anderson JL, Berce M, Hartley DE. Endoluminal aortic grafting with renal and superior
mesenteric artery incorporation by graft fenestration. J Endovasc Ther 2001; 8(1):3-15.

6. Faruqi RM, Chuter TA, Reilly LM, et al. Endovascular repair of abdominal aortic aneurysm
using a pararenal fenestrated stent-graft. J Endovasc Surg 1999; 6(4):354-358.

7. Chuter TA, Gordon RL, Reilly LM, et al. An endovascular system for thoracoabdominal
aortic aneurysm repair. J Endovasc Ther 2001; 8(l):25-33.

8. Burks JA, Faries PL, Gravereaux EC, et al. Endovascular repair of abdominal aortic aneu-
rysms: Stent-graft fixation across the visceral arteries. J Vase Surg 2002; 35:109-113.

9. Reilly LM, Ramos TK, Murray SP, et al. Optimal exposure of the proximal abdominal aorta:
A critical appraisal of transabdominal medial visceral rotation. J Vase Surg 1994; 19:375-390.

10. Crawford ES, Beckett WC, Greer MS. Juxtarenal infrarenal abdominal aortic aneurysm.
Special diagnostic and therapeutic considerations. Ann Surg 1986; 203:661-670.

11. Qvarfordt PG, Stoney RJ, Reilly LM, et al. Management of pararenal aneurysm of the
abdominal aorta. J Vase Surg 1986; 3:84-93.

12. Green RM, Ricotta JJ, Ouriel K, DeWeese JA. Results of supraceliac aortic clamping in the
difficult elective resection of infrarenal abdominal aortic aneurysm. J Vase Surg 1989; 9:124-134.

13. Poulias GE, Doundoulakis N, Skoutas B, et al. Juxtarenal aortic aneurysmectomy. J
Cardiovasc Surg 1992; 3:324-330.

14. Breckwoldt WL, Mackey WC, Belkin M, O'Donnell TJ Jr. The effect of suprarenal cross-
clamping on abdominal aortic aneurysm repair. Arch Surg 1992; 127:520-524.

15. Allen BT, Anderson CB, Rubin BG, et al. Preservation of renal function in juxtarenal and
suprarenal abdominal aortic aneurysm repair. J Vase Surg 1993; 17:984-959.

16. Nypaver TJ, Shepard AD, Reddy DJ, et al. Repair of pararenal abdominal aortic aneurysms.
An analysis of operative management. Arch Surg 1993; 128:803-813.

17. Schneider JR, Gottner RJ, Golan JF. Supraceliac versus infrarenal aortic cross-clamp for
repair of nonruptured infrarenal and juxtarenal abdominal aortic aneurysm. Cardiovasc Surg
1997; 5:279-285.

18. Faggioli G, Stella A, Freyrie A, et al. Early and long-term results in the surgical treatment of
juxtarenal and pararenal aortic aneurysms. Eur J Vase Endovasc Surg 1998; 15:205-211.

19. Jean-Claude JM, Reilly LM, Stoney RJ, Messina LM. Pararenal aortic aneurysms: The future
of open aortic aneurysm repair. J Vase Surg 1999; 29:902-912.

20. Tordoir JHM, van de Pavoordt HDWM, Eikelboom BC, et al. Thoraco-abdominal aortic
approach for the treatment of pararenal aneurysm. Neth J Med 1988; 40:1-5.

21. Shepard AD, Scott GR, Mackey WC, et al. Retroperitoneal approach to high-risk abdominal
aortic aneurysms. Arch Surg 1986; 121:444—448.

22. Shepard AD, Tollefson DF, Reddey DJ, et al. Left flank retroperitoneal exposure: A technical
aid to complex aortic reconstruction. J Vase Surg 1991; 14:283-291.

23. Darling RC III, Shah DM, Chang BB, et al. Retroperitoneal approach for bilateral renal and
visceral artery revascularization. Am J of Surg 1994; 168(2): 148—151 .

24. Elmore JR, Hallett JW Jr, Gibbons RJ, et al. Myocardial revascularization before abdominal

aortic aneurysmorrhaphy: Effect of coronary angioplasty. Mayo Clin Proc 1993; 68:637-641. j>

25. Kashyap VS, Cambria RP, Davison JK, L'ltalien GJ. Renal failure after thoracoabdominal <S
aortic Surgery. J Vase Surg 1997; 26:949-957. 1

26. Welborn BM, Oldenburg HS, Hess PJ, et al. The relationship between visceral ischemia, 2
proinflammatory cytokines, and organ injury in patients undergoing thoracoabdominal aortic **
aneurysm repair. Crit Care Med 2000; 28:3191-3197. &

27. Cherr SG, Hansen KJ. Renal complications with aortic surgery. Semin Vase Surg 2001; jjj
14(4):245-254. ' " " q

28. Roizen MF, Beaupre PN, Alpert RA, et al. Monitoring with two-dimensional transesophageal |
echocardiography: Comparison of myocardial function in patients undergoing supraceliac, 2
suprarenal-infraceliac, or infrarenal aortic occlusion. J Vase Surg 1984; 1:300-305. ^

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS AFTER AORTIC RECONSTRUCTION 259

29. Valentine RJ, Hagina RT, Jackson RM, et al. Gastrointestinal complications after aortic
surgery. J Vase Surg 1998; 28:404^112.

30. Burkey SH, Valentine RJ, Jackson MR, et al. Acute pancreatitis after abdominal vascular
surgery. J Am Coll Surg 2000; 191:373-380.

31. Harward TRS, Brooks DL, Flynn TC, Seeger JM. J Vase Surg 1993; 18:459-469.

32. Harward TRS, Welborne MB III, Martin TD, et al. Visceral ischemia and organ dysfunction
after thoracoabdominal aortic aneurysm repair. A clinical and cost analysis. Ann of Surg 1996;
223(6):729-736.

33. Cambria RP, Davison JK, Giglia JS, Gertler JP. Mesenteric shunting decreases visceral ische-
mia during thoracoabdominal aneurysm repair. J Vase Surg 1998; 27:745-749.

34. Ballard JL. Thoracoabdominal aortic aneurysm repair with sequential visceral perfusion: A
technical note. Ann Vase Surg 1999; 13:216-221.

35. Cherry KJ. Techniques in the management of recurrent aortic aneurysm. In: Yao JST, Pearce
WH, eds. Aneurysms: New Findings and Treatments. Norwalk, CT: Appleton & Langee,
1994:249-258.

36. Hallett JW Jr, Marshall DM, Petterson TM, et al. Graft-related complications after abdominal
aortic aneurysm repair: Reassurance from a 36-year- population-based experience. J Vase Surg
1997; 25:277-286.

37. Biancari F, Ylonen K, Anttila V, et al. Durability of open repair of infrarenal abdominal aortic
aneurysm: A 15-year follow-up study. J Vase Surg 2002; 35:87-93.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

13

Complications of Modern Renal Revascularization

Jeffry D. Cardneau and Louis M. Messina

University of California, San Francisco, San Francisco, California, U.S.A.

Renovascular hypertension due to renal artery occlusive disease is the most common cause
of secondary hypertension (1,2). Its prevalence in both the general population and the
hypertensive population is not known precisely, with estimates ranging from 0.18 (3) to
5% (4,5) of the hypertensive population to 3% of the general population (6). Renal artery
occlusive disease may be responsible for up to 16% of patients with end-stage renal
disease, and its prevalence appears to be increasing (7). The vast majority of renal artery
occlusive disease causing hypertension is due to atherosclerosis, ranging from 67 to 97.8%
(8-12).

The benefits of renal revascularization for renovascular hypertension have been
known for several decades. A study from the Mayo Clinic published 30 years ago showed
that patients who underwent surgical management of hypertension had better long-term
survival than did those treated by drug therapy alone. A total of 84% of the surgical group
but only 66% of the drug therapy group were alive at a follow-up of 7-14 years (13).
Modern-day renal revascularization includes percutaneous renal angioplasty alone, per-
cutaneous angioplasty with stenting, and traditional open surgical techniques. All tech-
niques have associated complications. These complications can, for simplicity, be placed
into three categories. These are errors in patient selection, periprocedural complications,
and late complications. Late complications should include not only technical problems but
also failure to maintain clinical benefit for which the procedure was performed. However, ^

documentation of the durability of improved control of hypertension and preservation of %

renal function is beyond the scope of this chapter. Therefore this topic is not dealt with £

here. §

■c

I. PATIENT SELECTION B "

I
Renal artery stenosis does not inevitably lead to renovascular hypertension. Renal artery 2

stenosis is present in up to 4% of unselected patients and up to 16% of hypertensive |

patients undergoing aortography (14). Indeed, renal artery stenosis may be present in up @

261 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

262 CARDNEAU and MESSINA

to 40% of new cases of end-stage renal disease in the elderly (15). Its mere presence does
not establish cause and effect. One study discovered that 45-60% of autopsies of both
normotensive and hypertensive patients showed the presence of renal artery stenosis (15).
Thus, it is imperative that an appropriate algorithm be employed to identify patients for
revascularization. A recent prospective trial failed to show the superiority of angioplasty
over best-drug therapy in the management of what was believed to be renovascular
hypertension (16). This failure of renal angioplasty in the management may have been due
to poor patient selection. Improper patient selection for renal revascularization can expose
patients to serious, life-threatening complications. Clinical clues that renal artery stenosis
may be the cause of hypertension include abrupt onset of hypertension, particularly at the
extremes of age; recently accelerated or malignant hypertension; hypertension refractory
to medication (usually at least three medications); unexplained azotemia in the setting of
hypertension; azotemia induced by angiotensin-converting enzyme (ACE) inhibitors; or
hypertension in a patient with diffuse atherosclerotic disease (coronary, carotid, or periph-
eral) (1,17). Renovascular hypertension is usually severe, such that there is a low like-
lihood that a patient with a diastolic blood pressure below 95 mmHg has this process (6).
Additionally, an episode of pulmonary edema in a patient with poorly controlled hyper-
tension and renal insufficiency should prompt investigation for the diagnosis of bilateral
renal artery stenosis (18).

Some studies have found demographic differences in the prevalence of renovascular
hypertension, with a low occurrence in African Americans. In one study of 45 patients
with new end-stage renal disease, all 10 patients with renal artery stenosis (RAS) were
white while none were African- American (19). Data from the Health Care Finance
Administration (HCFA) show that renovascular disease progressing to renal failure is
present more commonly in whites than in African Americans at a ratio of 14:1 (20). While
these differences may be present in the select population of patients with RAS progressing
to end-stage renal disease, the most recent literature suggests that RAS itself shows no
ethnic predilection (21).

Arteriography has traditionally been the "gold standard" for the diagnosis of RAS.
Certain arteriographic findings are present in renovascular hypertension. These include
the presence of collateral vessels and a systolic pressure gradient across the stenosis of
at least 10 mmHg (2). It is generally agreed that an angiographic stenosis of about 70%
is hemodynamically significant (22). This degree of stenosis corresponds to an approx-
imate 40% drop of renal perfusion pressure, which is necessary to cause a decrease in
the glomerular filtration rate (GFR) (15). However, several noninvasive studies have
been used successfully to select patients who might benefit from revascularization.
Duplex ultrasonography has proven sensitive, with numerous groups reporting sensi-
tivity of greater that 93% (23,24). Duplex criteria for stenosis are listed elsewhere (23).
A ratio of peak systolic velocities in the renal artery and aorta of at least 3.5:1 and a ■g

peak systolic renal artery velocity of at least 180 cm/s strongly suggests a stenosis of &

60% or more. Ultrasonography may be even more useful in determining patients who a

will benefit from revascularization when using the resistive index (RI) of at least 0.8 as -c

an exclusionary criterion (25). The latter suggests intrinsic renal disease as the cause of <j

renal insufficiency. >3

Measurement of plasma renin activity after ACE inhibition is another method to 4j

diagnose clinically significant RAS. Baseline renin activity is measured, followed by the 2

administration of 25 mg of captopril. Plasma renin activity (PRA) is measured again at 60 |

min. Accordingly, a post-captopril PRA of greater than 12 ng/mL/h, an increase in PRA @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF RENAL REVASCULARIZATION 263

of more than 10 ng/mL/h, or an increase of 150% from baseline was highly suggestive of
renovascular hypertension (26). The sensitivity and specificity was 100 and 95%, respec-
tively (4). Others have not shown such reliable results. Captopril renal scintigraphy may be
a more sensitive study (27). Renal scintigraphy takes advantage of the decreased
glomerular filtration rate of radiotracer after the administration of ACE inhibitor. Diag-
nostic criteria for significant RAS include a delayed time of maximal activity of radio-
tracer greater than 1 1 min after captopril, asymmetry of peak activity of each kidney,
marked cortical retention of radionuclide after captopril, and marked reduction in the cal-
culated GFR of the kidney (4). Sensitivity ranges from 90 to 93%, while specificity ranges
from about 93 to 100%. Since this test takes advantage of differential radionuclide activity
due to the stenotic renal artery, its positive predictive value decreases in the setting of bi-
lateral RAS and in patients with moderate renal insufficiency.

Magnetic resonance angiography (MRA), although considerably more expensive than
ultrasound, may become the noninvasive procedure of choice for the diagnosis of reno-
vascular occlusive disease and preprocedure planning. Recent series cite sensitivities of
93-100% and specificities of 88-100% when both 3D gadolinium-enhanced and 3D phase-
contrast sequences are used (28). Another advantage of MRA for this purpose is the ability
to image the aorta and its branches in order to define options for open surgical treatment.

II. PERIPROCEDURAL COMPLICATIONS
A. Percutaneous Procedures

When Gruntzig introduced percutaneous transluminal angioplasty (PTA) of the renal
arteries in 1978, a new treatment modality for renovascular hypertension was established.
While angioplasty alone has been highly successful for fibromuscular disease (1), it has led
to more disappointing results for atherosclerotic lesions, which are far more common. For
this reason, angioplasty followed by stent deployment is now more common. Although
renal angioplasty and/or stenting is successful technically in the large majority of patients,
its overall effectiveness in the management of renovascular hypertension has been
questioned. As mentioned previously, the only such prospective randomized trial failed
to show renal angioplasty to be more effective than best drug therapy in the management
of hypertension (16). Complications for both angioplasty alone and angioplasty with
stenting are addressed together below (Table 1), with a few exceptions. Fundamental to
the process of transluminal angioplasty is cracking the intimal and medial layers of the
artery. Thus, by definition, there will be a dissection at the end of a technically successful
procedure. It is only when the arterial wall does not remain compressed that this dissection
become flow-limiting and becomes a complication. Usually, when an arterial (flow-limit-
ing) dissection or occlusion occurs, a guidewire is still in place. This complication can al-
most always be remedied with placement of a stent. Some would therefore argue that these
are not complications. The need for unplanned stent placement for dissection or occlusion

occurs in 1-3% of renal PTA procedures (5, 29-31). Additionally, PTA can fail technically c

due to elastic recoil of the vessels, requiring stenting to reduce the stenosis. This is espe- <j

daily common in atherosclerotic ostial lesions occurring in 12-38% of lesions (17,30,32). >9

Rarely, rigid, calcified stenoses are encountered that cannot be dilated by the angioplasty 41

balloon. Other complications in percutaneous procedures are common to both PTA alone Q

and stenting; they are therefore described here collectively. Several studies have shown that |

the rate of total complications is not significantly different from PTA alone and PTA with @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

264 CARDNEAU and MESSINA

Table 1 Approximate Cited Ranges of Complications After
Percutaneous Transluminal Renal Revascularization

Overall

4-50 %

"Minor"

4.5-48

"Major"

5-17

Local

Hematoma

4-21

Retroperitoneal hematoma

0-6.8

Pseudoaneurysm

0-9.5

Occlusion of access vessel

<1

Femoral arteriorvenous fistula

0-1.6

Renal

Initial technical failure

<2

Artery dissection

0-2.4

Artery/stent occlusion

0-4.8

Artery pseudo/aneurysm

0-1.5

Transient renal failure

3.4-21

Prolonged azotemia

2

Stent maldeployment

0.2-2.4

Artery perforation/extravasation

0-6.9

Perinephric hematoma

<1.1

Cholesterol atheroemboli

1-8

Need for emergency nephrectomy

<1

Systemic

Hemorrhage requiring transfusion

0-9.5

Sepsis

<1

Aortic dissection

Rare

Stroke

0-3.3

Myocardial infarction

0-9

Distal emboli (eg: limbs or other organs)

0.1-1.1

Mortality

1

stenting, with incidences ranging from 4.4-50% (5,11,14,29-31,33-40). A recent met-
analysis comparing 644 PTA patients with 678 stented patients showed an overall com-
plication rate of 13% for PTA and 11% for stenting (41). Many have subdivided these
rates into "major" and "minor" complications, minor complications being exemplified by
groin hematomas and major complications including stroke and renal infarction. Re-
ported minor complication rates range from about 4.5-48%, while major complication
rates occur in about 5-17% of reported series. Complications have also been previously
grouped into three categories: local, renal-related, and systemic (42). This classification is g

maintained here. |

1
1 . Local Complications .g>

These complications are related to the percutaneous access that is required for these <j

procedures. Some of these complications may be minimized using ultrasound guidance for >9

puncture as well as the use of micropuncture techniques with small-gauge needles and 4j

coaxial dilating systems. Q

Hematoma is often the most common complication listed in series related to trans- |

luminal techniques. This usually occurs after the removal of the sheath or catheter. The @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF RENAL REVASCULARIZATION 265

frequency of this complication can be reduced by using diligent pinpoint digital pressure
on the artery entry site. The use of sandbags and diffuse pressure, as with a large pack of
gauze, are to be discouraged. The reported incidence is about 4-21%.

In addition to local hematoma formation, retroperitoneal hematomas can occur when
the femoral artery is accessed. This often is due to a puncture that is inadvertently high, in
the external iliac artery rather than the common femoral artery. This inadvertent puncture
thus lies above the inguinal ligament and poses difficulty in compression at the end of the
procedure. The incidence of retroperitoneal hematoma may be as high as 6.8%. Pseu-
doaneurysms can occur at the puncture site, which may require surgical repair. However,
smaller, stable pseudoaneurysms can be managed conservatively or with ultrasound-
guided compression and thrombin injection. The reported incidence of pseudoaneurysm
ranges from to 10%.

Occlusion of the access artery is possible, especially in this setting of patients who have
atherosclerosis of the punctured vessel. This can occur in the femoral artery, the iliac
artery in which the access sheath is positioned, or the brachial artery when an upper
extremity approach is used. Occlusion rates are reduced in modern series with the use of
periprocedural heparinization and smaller delivery catheters. Stents of the size appropriate
for renal vessels can almost always be delivered on a 5F catheter or smaller and delivered
through 5F or 6F sheaths. The incidence of occlusion is less than 1%.

Acute ischemia of a lower limb — or arm in the case of a brachial artery access — can be
due to thrombosis of the access vessel, as discussed above, but it also may due to vessel
dissection or emboli to the limb. Although rare, acute limb ischemia occurred in 4. 1 % of
patients in one study (43).

2. Renal-Related Complications

These complications are those distant to the percutaneous access site and are related to
wire, catheter, or stent manipulation. Dissection of the renal artery is due either to the bal-
loon angioplasty performed (see above) or to a subintimal placement of the guidewire tip
during cannulation of the vessel. This dissection may be of no hemodynamic significance
but can propagate to cause segmental renal infarction or even complete artery occlusion.
Propagation of the dissection can also proceed proximally (Fig. 1). With the more routine
use of stents, dissection is less common now than in series of PTA alone. Series report an
incidence of dissection with stenting ranging from zero to 2.4% (11,30,31,33,37,43).

Renal artery occlusion can be the result of artery dissection or acute thrombosis due
to manipulation, cracking of the atherosclerotic plaque, or the stent placement itself.
Systemic heparin is usually given during the procedure; thus acute occlusion is rare.
Occlusion occurs in zero-4.8% of cases (30,34,35,37,43). Where it has occurred, it has
sometimes been successfully treated with thrombolytics, whereas in other instances an
urgent surgical repair was required (37). ■g

Another renal-related complication is impairment of renal function. This can be due &

to multiple mechanisms. Impaired renal function can manifest itself as a postprocedure a

rise in serum creatinine, sometimes progressing to dialysis-dependent renal failure. -c

Fortunately, dialysis dependent renal failure is usually transient. Impaired renal function <j

occurs in 3.4-21% of patients (14,30,33,35). In Leertouwer's metanalysis, procedure- >9

related renal failure occurred in 34 of 799 treated arteries (4.2%) (41). Postulated 4j

mechanisms for postprocedural renal failure include contrast-induced injury, sudden onset 2

of high pressure within the glomeruli, development of reactive oxygen species after |

revascularization, and renal atheroemboli from traversing the lesion. Eosinophilia greater @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

266

CARDNEAU and MESSINA

(A)

i

2

I

■c

(B)

I

U

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF RENAL REVASCULARIZATION 267

than 5% of a peripheral blood smear may indicate that significant cholesterol emboli have
occurred (44). Large randomized trials are required to determine whether distal embolic
protection during PTA and stenting actually has an impact on outcomes and on long-term
preservation of renal function.

Late restenosis is a common problem after stenting using current technology. Reste-
nosis is due to myointimal hyperplasia or progression of atherosclerotic disease. However,
myointimal hyperplasia is more likely, as stent restenosis occurs within a relatively short
time span. Some authors have claimed that most restenosis, if it occurs, will take place
within the first year (31). However, others have found this not to be true (37). Stent reste-
nosis rates range from 9.3 to 44% at 6 months-1 year (11,14,17,30,31,37,40,45) to 1 1.4—
55% at 2 years (45,46). Leertouwer's metanalysis found a mean stent restenosis rate of
17% in a follow-up of 6-29 months (41). Primary assisted patency, during close sur-
veillance, is approximately 80-100% (14,30,34,35,37) (Fig. 2). Lederman's review empha-
sizes the importance of vessel size in restenosis, with rates (at a mean of 303 days) of 36.0,
15.8, and 6.5% for vessels less than 4.5 mm, 4.5-6 mm, and larger than 6 mm, respectively
(11). Stent restenosis may be reduced with such technological advances as drug-eluting
stents, which are the subject of several ongoing trials.

3. Systemic Complications

With the knowledge that the majority of these patients have a diffuse form of atheroscle-
rosis, it is not surprising that systemic complications occur. Myocardial infarction, stroke,
sepsis, and emboli to other vessels all occur, but uncommonly in the percutaneous pro-
cedure. Emboli to other vessels, when they do occur, most commonly travel to the lower
extremities, manifesting themselves as claudication, the blue toe syndrome, or frankly is-
chemic digits. Other sites of emboli have included the superior mesenteric artery.

Death has occurred as a result of transluminal procedures, but this is rare. Causes of
death include massive hemorrhage from puncture sites; profound systemic complications
such as massive myocardial infarction, stroke, or bowel infarction; and consequences of
acute renal failure. Thirty-day mortality rates are as high as 3.8% (14,29,33,34). The mean
mortality rate is 1% (41). A preoperative serum creatinine of more than 1.5 mg/dL is a
strong independent risk factor of death in the years after stenting (33). This risk is five
times higher than that of patients with creatinine below 1.5 mg/dL and most likely rep-
resents the severity of associated comorbidities (17). This is reinforced by studies showing
that patients on hemodialysis with end-stage renal disease due to RAS have a poorer sur-
vival than patients with end-stage renal disease due to other causes (47).

B. Surgical Procedures

Surgical revascularization for renovascular hypertension dates back to 1952, when Wylie
performed the first renal artery endarterectomy (48,49). Since that time, techniques of
renal revascularization have evolved, along with the rest of vascular surgery. A wide vari- §

I

■c

I

Figure 1 A. A 55 year-old woman undergoing angiography for claudication. She also had two- ,-,-

drug hypertension. A left renal artery ostial stenosis was noted, and PTA was performed with a 6- S

mm balloon. B. After the PTA. A dissection is noted propagating proximally in the aorta. This !3

dissection progressed up to the left subclavian artery. As this occurred prior to the advent of stents, I

this patient was managed on an aggressive oral antihypertensive regimen. @

I
u

£

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

268

CARDNEAU and MESSINA

(A)

(B)

(C)

i

2

I

■c

Figure 2 A. 74-year-old man with an occluded right renal artery and a stenosis of left renal artery.
B. The same patient after successful placement of a 6-mm Palmaz stent. C. Restenosis noted 4
months after stent deployment. This was diagnosed by ultrasonography ordered in response to a
new rise in serum creatinine. D. Successful reopening of stent restenosis with repeat balloon
angioplasty.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF RENAL REVASCULARIZATION 269

(D)

Figure 2 Continued.

ety of surgical approaches have been used in this setting based on the etiology of the RAS,
patient comorbidities, the need for concomitant operations, and even regional geographic
preferences. The occurrence of open surgical revascularization has declined, while percuta-
neous interventions have markedly risen. Based on charges submitted to Medicare from
1995 to 2001, open renovascular operations have declined by 45%, while stenting has
increased over 350% (50). Nonetheless, surgical repair will continue to have a role, par-
ticularly in younger patients, after the failure of stenting and after complications encoun-
tered because of stenting. The durability of transaortic renal endarterectomy and aortorenal
bypass has been well established. For good-risk patients and patients who require con-
comitant aortic reconstruction, these will remain the procedures of choice.

Overall complications range from 9.1 to 37.9% (9,10,12,17,29,35,40,48,51-56). Com-
plications related to open renovascular repair can be categorized, analogously to percuta-
neous procedures, into renal-related and systemic (Table 2). Further, special consideration
must be given to the type of operation performed and its attendant complications.

1 . Renal-Related Complications

When all types of renal revascularization operations are examined collectively, the early e

revascularization failure rate is approximately 5% (57) (Fig. 3). The need for transient e

postoperative dialysis ranges from zero to 10.2% (52,58). Long-term permanent dialysis is H,

significantly dependent on the degree of preoperative renal function. In a recent study, a

Tsoukas showed an outcome difference between the groups of patients with serum creatinine g

less than 2.0 mg/dL and those with creatinine greater than 2.0 mg/dL. Long-term dialysis for sf

these two groups was 2.3 versus 12%, respectively (52). 1

Aortorenal Bypass. Aortorenal bypass can be performed with a variety of conduits. |

Autogenous vein, hypogastric artery, and prosthetic all have been used with essentially ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

270 CARDNEAU and MESSINA

Table 2 Approximate Cited Ranges of Complications After Open
Surgical Renal Revascularization

Overall

9.1-37.9%

Renal

Distal intimal flap requiring repair

4-5

Artery occlusion

1-3.4

Transient worsening of renal function

4.6^13

requiring dialysis

= 5

Prolonged azotemia

1.5-12

Early graft failure

2.1-17.6

Late graft failure

1.8-18

Systemic

Postoperative hemorrhage

1.5-3.4

Multisystem organ failure

4.6

Sepsis

<1

Respiratory failure or pneumonia

0.9-25

Stroke

0-3.3

Myocardial infarction

2-8.6

Distal emboli (eg: limbs or other 01

•gans)

2.2-6.9

Limb ischemia

4.5

Mortality overall

0-7.1

Mortality (isolated renal revascularization)

0-7.0

Mortality (with concomitant aortic

repair)

1-9.2

similar long-term results (10,57). A key reason for the long-term success of these grafts is
that 20% of the total cardiac output goes to the renal bed, which is also of low vascular
resistance. Early graft failure occurs in approximately 5% of modern series at most and
almost certainly represents technical problems of the graft (12,59). Examples include
inadequate spatulation of the anastomosis, purse-stringing the anastomosis, uncorrected
intimal flaps, and placement of an excessively long graft, leading to kinking (59). Bypass
with saphenous vein can be complicated by aneurysmal degeneration. In Stanley's series of
100 consecutive saphenous aortorenal grafts, 6 of 74 grafts (8.1%>) examined late had
aneurysmal expansion (60). However, the large majority — in some series, all (59) — of these
grafts with aneurysmal dilatation occurred in children, with this problem being very rare
in the adult patient. Thus, this conduit is generally not considered an appropriate revas-
cularization technique in children (60-62).

Late graft stenosis can be heralded by worsening of previously improved renal
function or recurrence of hypertension. However, late graft stenosis can also be asympto-
matic. Conversely, clinical deterioration can occur without graft failure or stenosis (51). j>

Nonetheless, noted clinical deterioration of glomerular filtration rate or hypertension g

I

■c

I

Figure 3 A. MRA of a 55-year-old woman after thoraco-bi-iliac bypass. Note a graft-to-renal ,-,-

artery bypass arising from the main body of the aortic graft as well as a separate bypass from the S

right iliac limb to a second renal artery. B. MRA of the same patient approximately 6 months later. 2

There is now occlusion of the bypass graft arising from the aortic graft to the main renal artery. C. I

Aortography confirms the MRA finding of the occluded renal bypass. @

I

§•
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF RENAL REVASCULARIZATION

271

(A)

(B)

i

2

I

■c

(C)

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

272 CARDNEAU and MESSINA

control should prompt investigation of the graft with duplex ultrasound, MRA, or
angiography. Late graft stenosis occurs in 1.3-19% of grafts (9,52,53,59,60). Graft
stenosis can often be successfully treated by PTA (Fig. 4).

Transaortic Renal Endarterectomy . Although endarterectomy has been described
being performed through an arteriotomy of the renal artery itself, the preferred method
now is through either a longitudinal or transverse aortotomy, in which a diffuse
atherosclerotic plaque can potentially be excised from the renal arteries, the paravisceral
aorta, and mesenteric arteries if necessary.

Endarterectomy can be performed on occluded arteries provided that there is reconsti-
tution of the main renal artery (55). Additionally, endarterectomy has been successfully
performed after restenosis of bilateral renal artery stenting (63). Conversion from endar-
terectomy to another procedure, such as direct reimplantation of the artery or bypass graft,
occurs because of aortic or arterial fragility after endarterectomy. This conversion rate is
approximately 6.5% (55). An intimal flap in the endpoint of the endarterectomized renal
artery has been noted to occur in almost 50% of arteries when surveyed by duplex ultra-
sonography, but this manifests itself as a hemodynamic problem in only about 4% of
vessels (48), as has been confirmed in other studies (55,64). When necessary, a transverse
renal arteriotomy just distal to the flap is sufficient to resect the flap and restore normal
renal flow.

Nonanatomic Renal Artery Bypass. Nonanatomical bypass for renovascular hyper-
tension is becoming more common (17). This may be intuitive in light of studies showing
that the mean age of patients is increasing, as is the prevalence of atherosclerosis of the
diffuse type (9). These operations include bypass grafts arising from the hepatic, splenic, or
mesenteric vessels as well as the iliac arteries. However, it is important to remember that
these patients may have concomitant disease of the origins of their visceral vessels,
precluding the possibility of splenorenal or hepatorenal bypasses. Advantages of iliorenal
bypasses include the ability to avoid aortic cross-clamping and thus avoiding increases in
afterload in patients with impaired cardiac function. Additionally, clamping of a diffusely
calcified aorta can be obviated. Iliorenal bypasses can be performed with either an
autogenous or a prosthetic conduit. A recent review of 323 renal artery revascularizations
showed no difference in early or long-term patency between aortorenal and nonanatomic
bypass (57), with 5-year patency of 88.7 versus 82.1%, respectively.

2. Systemic Complications

Other complications cited include many of those expected after a major intra-abdominal
procedure, such as postoperative hemorrhage, multisystem organ failure, sepsis, respira-
tory failure, pneumonia, stroke, and myocardial infarction.

Thirty-day mortality ranges from zero to 7.9% for all procedures combined (9,48,51-

53,55,57,58,64). There does not appear to be a consensus on whether there is a significant "g

difference in mortality between those patients receiving isolated renal revascularization &

and those also receiving a concomitant aortic repair (10,12,57,58). a

One recurrent significant risk factor for perioperative mortality is poor preoperative c

renal function. One study of 73 patients divided them into those with creatinine below 2.0 <j

mg/dL and those with higher levels; these groups had surgery-related mortalities of 4.4 -9

versus 14% respectively (52). Similar analyses of other series confirmed this difference, 41

using 3.0 mg/dL as the dividing line (58). Hansen and colleagues found that 22 of 26 late 2

deaths in a group of 200 patients followed over the long term (up to 58 months) were |

patients who had preoperative levels of serum creatinine greater than 2.0 mg/dL. This was @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF RENAL REVASCULARIZATION

273

(A)

(B)

Figure 4 A. 58-year-old woman 6 months after placement of an aortorenal bypass using the hypo-
gastric artery. A stenosis of the distal anastomosis is now evident. B. The same patient after successful
PTA of the distal anastomotic stenosis.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

274 CARDNEAU and MESSINA

the only preoperative risk factor statistically associated with mortality (53). Additionally,
20 of these 22 patients did not have a postoperative improvement in renal function.

The survival of such patients in general is poor and reflects the severe systemic
comorbidities. Actuarial survival after renal revascularization ranges from 61 to 81% at 5
years (8,56,58,65,66). Cardiovascular causes of death, not surprisingly, are most common
(53,58), followed by stroke, renal failure, and malignancy (8).

REFERENCES

1. Messina LM, Stanley JC. Renal artery fibrodysplasia and renovascular hypertension. In:
Rutherford RB, eds. Vascular Surgery. Philadelphia: Saunders, 2000:1650-1664.

2. Upchurch GR Jr, Stanley JC. Renal artery occlusive disease. In: Greenfield LJ, Mulholland
MW, Oldham KT, Zelenock GB, Lillemoe KD, eds. Surgery: Scientific Principles and Prac-
tice. Philadelphia: Lippincott Williams & Wilkins, 2001:1708-1724.

3. Tucker RM, Labarthe DR. Frequency of surgical treatment for hypertension in adults at the
Mayo Clinic from 1973 through 1975. Mayo Clin Proc 1977; 52:549-555.

4. Nally JV Jr. Provocative captopril testing in the diagnosis of renovascular hypertension. Urol
Clin North Am 1994; 21:227-234.

5. Xue F, Bettmann MA, Langdon DR, Wivell WA. Outcome and cost comparison of per-
cutaneous transluminal renal angioplasty, renal arterial stent placement, and renal arterial
bypass grafting. Radiology 1999; 212:378-384.

6. Hansen KJ, Wong JM. Aortorenal bypass for renovascular hypertension in adults. In: Stanley
JC, Ernst CB, eds. Current Therapy in Vascular Surgery. St. Louis: Mosby, 2001:735-741.

7. Fatica RA, Port FK, Young EW. Incidence trends and mortality in end-stage renal disease
attributed to renovascular disease in the United States. Am J Kidney Dis 2001; 37:1184-1190.

8. Steinbach F, Novick AC, Campbell S, Dykstra D. Long-term survival after surgical revas-
cularization for atherosclerotic renal artery disease. J Urol 1997; 158:38-41.

9. Novick AC, Ziegelbaum M, Vidt DG, Gifford RW Jr, Pohl MA, Goormastic M. Trends in
surgical revascularization for renal artery disease. Ten years' experience. JAMA 1987; 257:498-
501.

10. Lawrie GM, Morris GC Jr, Glaeser DH, DeBakey ME. Renovascular reconstruction: factors
affecting long-term prognosis in 919 patients followed up to 31 years. Am J Cardiol 1989; 63:
1085-1092.

11. Lederman RJ, Mendelsohn FO, Santos R, Phillips HR, Stack RS, Crowley JJ. Primary renal
artery stenting: Characteristics and outcomes after 363 procedures. Am Heart J 2001; 142:314-
323.

12. Bredenberg CE, Sampson LN, Ray FS, Cormier RA, Heintz S, Eldrup-Jorgensen J. Changing
patterns in surgery for chronic renal artery occlusive diseases. J Vase Surg 1992; 15:1018-1023.

13. Hunt JC, Strong CG. Renovascular hypertension. Mechanisms, natural history and treatment.
Am J Cardiol 1973; 32:562-574.

14. MacLeod M, Taylor AD, Baxter G, Harden P, Briggs D, Moss J, Semple PF, Connell JM,
Dominiczak AF. Renal artery stenosis managed by Palmaz stent insertion: technical and clin- j>
ical outcome. J Hypertens 1995; 13:1791-1795. |

15. Textor SC, Wilcox CS. Renal artery stenosis: a common, treatable cause of renal failure? Annu js
Rev Med 2001; 52:421^142. Jf

16. van Jaarsveld BC, Krijnen P, Pieterman H, Derkx FH, Deinum J, Postma CT, Dees A, ^
Woittiez AJ, Bartelink AK, Man in 't Veld AJ, Schalekamp MA. The effect of balloon 3
angioplasty on hypertension in atherosclerotic renal-artery stenosis. Dutch Renal Artery ^
Stenosis Intervention Cooperative Study Group. N Engl J Med 2000; 342:1007-1014. q

17. Safian RD, Textor SC. Renal-artery stenosis. N Engl J Med 2001; 344:431^142. |

18. Messina LM, Zelenock GB, Yao KA, Stanley JC. Renal revascularization for recurrent 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF RENAL REVASCULARIZATION 275

pulmonary edema in patients with poorly controlled hypertension and renal insufficiency: a
distinct subgroup of patients with arteriosclerotic renal artery occlusive disease. J Vase Surg
1992; 15:73-80.

19. Appel RG, Bleyer AJ, Reavis S, Hansen KJ. Renovascular disease in older patients beginning
renal replacement therapy. Kidney Int 1995; 48:171-176.

20. IV. Patient characteristics at the start of ESRD: data from the HCFA medical evidence form.
Am J Kidney Dis 1999; 34:S63-S73.

21. Hansen KJ, Edwards MS, Craven TE, Cherr GS, Jackson SA, Appel RG, Burke GL,
Dean RH. Prevalence of renovascular disease in the elderly: a population-based study. J
Vase Surg 2002; 36:443^151.

22. Sacks D, Rundback JH, Martin LG. Renal angioplasty/stent placement and hypertension in
the year 2000. J Vase Interv Radiol 2000; 11:949-953.

23. Zierler RE. Duplex scanning for renal arterial occlusive disease. In: Stanley JC, Ernst CB, eds.
Current Therapy in Vascular Surgery. St. Louis: Mosby, 2001:717-722.

24. Olin JW, Piedmonte MR, Young JR, DeAnna S, Grubb M, Childs MB. The utility of duplex
ultrasound scanning of the renal arteries for diagnosing significant renal artery stenosis. Ann
Intern Med 1995; 122:833-838.

25. Radermacher J, Chavan A, Bleck J, Vitzthum A, Stoess B, Gebel MJ, Galanski M, Koch KM,
Haller H. Use of Doppler ultrasonography to predict the outcome of therapy for renal-artery
stenosis. N Engl J Med 2001; 344:410-417.

26. Muller FB, Sealey JE, Case DB, Atlas SA, Pickering TG, Pecker MS, Preibisz JJ, Laragh JH.
The captopril test for identifying renovascular disease in hypertensive patients. Am J Med 1986;
80:633-644.

27. Elliott WJ, Martin WB, Murphy MB. Comparison of two noninvasive screening tests for
renovascular hypertension. Arch Intern Med 1993; 153:755-764.

28. Prince MR, Dong Q, Schoenberg SO. Magnetic resonance angiographic diagnosis of reno-
vascular disease. In: Stanley JC, Ernst CB, eds. Current Therapy in Vascular Surgery. St.
Louis: Mosby, 2001:723-728.

29. Libertino JA, Beckmann CF. Surgery and percutaneous angioplasty in the management of
renovascular hypertension. Urol Clin North Am 1994; 21:235-243.

30. van de Ven PJ, Kaatee R, Beutler JJ, Beek FJ, Woittiez AJ, Buskens E, Koomans HA,
Mali WP. Arterial stenting and balloon angioplasty in ostial atherosclerotic renovascular
disease: a randomised trial. Lancet 1999; 353:282-286.

31. Blum U, Krumme B, Flugel P, Gabelmann A, Lehnert T, Buitrago-Tellez C, Schollmeyer P,
Langer M. Treatment of ostial renal-artery stenoses with vascular endoprostheses after
unsuccessful balloon angioplasty. N Engl J Med 1997; 336:459-465.

32. Canzanello VJ, Millan VG, Spiegel JE, Ponce PS, Kopelman RI, Madias NE. Percutaneous
transluminal renal angioplasty in management of atherosclerotic renovascular hypertension:
results in 100 patients. Hypertension 1989; 13:163-172.

33. Dorros G, Jaff M, Mathiak L, Dorros II, Lowe A, Murphy K, He T. Four-year follow-up of
Palmaz-Schatz stent revascularization as treatment for atherosclerotic renal artery stenosis.
Circulation 1998; 98:642-647.

34. Yutan E, Glickerman DJ, Caps MT, Hatsukami T, Harley JD, Kohler TR, Davies MG. 1
Percutaneous transluminal revascularization for renal artery stenosis: Veterans Affairs Puget g
Sound Health Care System experience. J Vase Surg 2001; 34:685-693. js

35. Weibull H, Bergqvist D, Bergentz SE, Jonsson K, Hulthen L, Manhem P. Percutaneous 2
transluminal renal angioplasty versus surgical reconstruction of atherosclerotic renal artery ■"■
stenosis: a prospective randomized study. J Vase Surg 1993; 18:841-850. «

36. Plouin PF, Chatellier G, Darne B, Raynaud A. Blood pressure outcome of angioplasty in jjj
atherosclerotic renal artery stenosis: a randomized trial. Essai Multicentrique Medicaments vs q
Angioplastie (EMMA) Study Group. Hypertension 1998; 31:823-829. |

37. Beutler JJ, Van Ampting JM, van de Ven PJ, Koomans HA, Beek FJ, Woittiez AJ, Mali WP. 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

276 CARDNEAU and MESSINA

Long-term effects of arterial stenting on kidney function for patients with ostial atherosclerotic
renal artery stenosis and renal insufficiency. J Am Soc Nephrol 2001; 12:1475-1481.

38. Tuttle KR, Chouinard RF, Webber JT, Dahlstrom LR, Short RA, Henneberry KJ, Dunham
LA, Raabe RD. Treatment of atherosclerotic ostial renal artery stenosis with the intravascular
stent. Am J Kidney Dis 1998; 32:611-622.

39. Ramsay LE, Waller PC. Blood pressure response to percutaneous transluminal angioplasty
for renovascular hypertension: an overview of published series. Br Med J 1990; 300:569-
572.

40. Harden PN, MacLeod MJ, Rodger RS, Baxter GM, Connell JM, Dominiczak AF, Junor BJ,
Briggs JD, Moss JG. Effect of renal-artery stenting on progression of renovascular renal
failure. Lancet 1997; 349:1133-1136.

41. Leertouwer TC, Gussenhoven EJ, Bosch JL, van Jaarsveld BC, van Dijk LC, Deinum J, Man
in 't Veld AJ. Stent placement for renal arterial stenosis: where do we stand? A meta- analysis.
Radiology 2000; 216:78-85.

42. Stanson AW. Complications of transluminal angioplasty of renal arteries. In: Bernhard VM,
Towne JB, eds. Complications in Vascular Surgery. Orlando, FL: Grune & Stratton, 1985:247-
257.

43. Bush RL, Najibi S, MacDonald MJ, Lin PH, Chaikof EL, Martin LG, Lumsden AB. Endo-
vascular revascularization of renal artery stenosis: technical and clinical results. J Vase Surg
2001; 33:1041-1049.

44. Dejani H, Eisen TD, Finkelstein FO. Revascularization of renal artery stenosis in patients with
renal insufficiency. Am J Kidney Dis 2000; 36:752-758.

45. Tunis MJ, Zierler RE, Glickerman DJ, Bergelin RO, Cantwell-Gab K, Strandness DE Jr.
Results of percutaneous transluminal angioplasty for atherosclerotic renal artery stenosis: a
follow-up study with duplex ultrasonography. J Vase Surg 1997; 25:46-54.

46. Henry M, Amor M, Henry I, Ethevenot G, Tzvetanov K, Courvoisier A, Mentre B, Chati Z.
Stents in the treatment of renal artery stenosis: long-term follow-up. J Endovasc Surg 1999;
6:42-51.

47. Mailloux LU, Napolitano B, Bellucci AG, Vernace M, Wilkes BM, Mossey RT. Renal vascular
disease causing end-stage renal disease, incidence, clinical correlates, and outcomes: a 20-year
clinical experience. Am J Kidney Dis 1994; 24:622-629.

48. Stoney RJ, Messina LM, Goldstone J, Reilly LM. Renal endarterectomy through the
transected aorta: a new technique for combined aortorenal atherosclerosis — A preliminary
report. J Vase Surg 1989; 9:224-233.

49. Wylie EJ, Perloff DL, Stoney RJ. Autogenous tissue revascularization technics in surgery for
renovascular hypertension. Ann Surg 1969; 170:416-428.

50. Rosenfield K. Renal Artery Disease — Rationale for Invasive Management and Consideration
Regarding Distal Embolic Protection. Washington, D.C.: Transcatheter Cardiovascular Thera-
peutics, 2002.

51. Cambria RP, Brewster DC, L'ltalien GJ, Gertler JP, Abbott WM, LaMuraglia GM, Moncure
AC, Vignati J, Bazari H, Fang LT, Atamian S. Renal artery reconstruction for the preservation
of renal function. J Vase Surg 1996; 24:371-380.

52. Tsoukas AI, Hertzer NR, Mascha EJ, O'Hara PJ, Krajewski LP, Beven EG. Simultaneous j>
aortic replacement and renal artery revascularization: the influence of preoperative renal <S
function on early risk and late outcome. J Vase Surg 2001; 34:1041-1049. s

53. Hansen KJ, Starr SM, Sands RE, Burkart JM, Plonk GW Jr, Dean RH. Contemporary surgical g
management of renovascular disease. J Vase Surg 1992; 16:319-330. ^

54. Libertino JA, Bosco PJ, Ying CY, Breslin DJ, Woods BO, Tsapatsaris NP, Swinton NW Jr. &
Renal revascularization to preserve and restore renal function. J Urol 1992; 147:1485-1487. ^

55. Clair DG, Belkin M, Whittemore AD, Mannick JA, Donaldson MC. Safety and efficacy of q
transaortic renal endarterectomy as an adjunct to aortic surgery. J Vase Surg 1995; 21:926-933. |

56. Chaikof EL, Smith RB III, Salam AA, Dodson TF, Lumsden AB, Kosinski AS, Coyle KA, 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF RENAL REVASCULARIZATION 277

Allen RC. Ischemic nephropathy and concomitant aortic disease: a ten-year experience. J Vase
Surg 1994; 19:135-146.

57. Cambria RP, Brewster DC, L'ltalien GJ, Moncure A, Darling RC Jr, Gertler JP, La Muraglia
GM, Atamian S, Abbott WM. The durability of different reconstructive techniques for
atherosclerotic renal artery disease. J Vase Surg 1994; 20:76-85.

58. Hallett JW Jr, Fowl R, O'Brien PC, Bernatz PE, Pairolero PC, Cherry KJ Jr, Hollier LH.
Renovascular operations in patients with chronic renal insufficiency: do the benefits justify
the risks? J Vase Surg 1987; 5:622-627.

59. Dean RH. Complications of renal revascularization. In: Bernhard VM, Towne JB, eds. Com-
plications in Vascular Surgery. Orlando, FL: Grune & Stratton, 1985:229-246.

60. Stanley JC, Ernst CB, Fry WJ. Fate of 100 aortorenal vein grafts: characteristics of late graft
expansion, aneurysmal dilatation, and stenosis. Surgery 1973; 74:931-944.

61. Berkowitz HD, O'Neill J A Jr. Renovascular hypertension in children. Surgical repair with
special reference to the use of reinforced vein grafts. J Vase Surg 1989; 9:46-55.

62. Stoney RJ, De Luccia N, Ehrenfeld WK, Wylie EJ. Aortorenal arterial autografts. Arch Surg
1981; 116:1416-1422.

63. Pak LK, Kerlan RK, Mully TW, Messina LM. Successful bilateral transaortic renal endar-
terectomy after failed renal artery angioplasty and stenting: a case report. J Vase Surg 2002;
35:808-810.

64. Dougherty MJ, Hallett JW Jr, Naessens JM, Bower TC, Cherry KJ, Gloviczki P, James EM.
Optimizing technical success of renal revascularization: the impact of intraoperative color-flow
duplex ultrasonography. J Vase Surg 1993; 17:849-856.

65. Wollenweber J, Sheps SG, Davis GD. Clinical course of atherosclerotic renovascular disease.
Am J Cardiol 1968; 21:60-71.

66. Ernst CB, Stanley JC, Marshall FF, Fry WJ. Renal revascularization for arteriosclerotic
renovascular hypertension: prognostic implications of focal renal arterial vs. overt generalized
arteriosclerosis. Surgery 1973; 73:859-867.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

14

The Diagnosis and Management of Aortic Bifurcation
Graft Limb Occlusions

Mark T. Eginton and Robert A. Cambria

Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

Aortic bifurcation grafts remain among the most successful and durable procedures
performed by vascular surgeons. Five-year patency rates in excess of 80% have been
widely reported in grafts placed for aortoiliac occlusive disease (1,2), with even higher
patency rates reported in grafts placed for aneurysmal disease (3). However, thrombosis of
an aortic graft limb represents the most common complication following reconstruction of
the abdominal aorta (4). This chapter focuses on the incidence, etiology, diagnosis, and
treatment of graft limb thrombosis.

I. INCIDENCE

Patency of aortobifemoral bypass grafts in the perioperative period is excellent, with
reported thrombosis of 2% or less (2). Patency rates decrease steadily over the follow-up
period, with a limb occlusion rate of approximately 4% per year, resulting in 10-year
patency from 66 to 78% (1,5,6). Table 1 illustrates not only the durability of these
procedures but also the continued risk for graft limb thrombosis 5-10 years following
implantation. Therefore thrombosis of an aortic graft limb continues to represent a
common clinical problem in spite of a recent trend away from aortic reconstruction for
aortoiliac occlusive disease (7). ■§

The incidence of graft thrombosis is influenced by risk factors that range beyond the |

technical and anatomic considerations of the graft itself. Smoking has been demonstrated as

to be associated with failure of aortofemoral reconstruction (3), and the thrombosis rate c

has been correlated with the amount of smoke exposure (8,9). Tibial vessel occlusive <

disease, both alone and in tandem with profunda femoris occlusive disease, has been as- &

sociated with aortic graft limb occlusion (10). As noted above, grafts placed for aneurys- J

mal disease have higher patency rates than those placed for occlusive disease (11). All of °

these factors are likely related to more advanced atherosclerotic occlusive disease, result- |

ing in compromised graft outflow over longer follow-up periods. @

279

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

280

EGINTON and CAMBRIA

Table 1 Patency of Aortic Bifurcation Grafts

Author

Number of Grafts

Patency Rate (%)

Follow-up (years)

Najafi (12)
Malone (10)

Martinez (1)

Brewster (6)

Nevelsteen (5)

601
180

355

464

869

92

82
66

88
78
88
74
74
70

3.5

5
10

5
10

5

10
10
15

II. ETIOLOGY

Traditionally, aortic graft limb occlusions have been classified as those occurring early in
the postoperative period or at later points during follow-up. Early thrombosis is defined as
occurring within 30 days of the primary procedure. As noted above, early thrombosis is
quite rare, occurring after 2% or less of aortic reconstructions (10,12), and early throm-
boses typically represents 10% or less of the cases in large series of thrombosed limbs
(13) (Table 2). Conversely, the majority of graft limb occlusions occur later in the follow-
up period, with the average interval from the primary procedure ranging from 28 to 60
months (4,14).

The etiology of graft occlusion, regardless of temporal relation to the original
procedure, must fall into one of three generic categories: inadequate inflow, inadequate
outflow, or problems within the graft itself. The most common problem is inadequate
outflow, usually due to progression of distal disease over time. Table 2 details the pre-
ponderance of outflow occlusion as an etiology for graft limb failure in several series.

A. Early Failure

Early graft occlusions, occurring within 30 days of the primary procedure, are usually due
to a technical error resulting in a mechanical problem within the graft or at one of the
anastomoses or to a judgmental error resulting in inadequate inflow or outflow. There
is no substitute for thoughtful preoperative planning and meticulous surgical technique
in the prevention of early graft thrombosis. This includes high-resolution preoperative
angiography to demonstrate the extent of occlusive disease in the entire abdominal aorta
as well as the common, superficial, and deep femoral arteries. Appropriate oblique views
must be obtained to delineate potential stenoses in the runoff vessels that could compro-
mise flow and result in graft occlusion. If concern exists for aortic intraluminal thrombus
or extensive calcification, which may complicate aortic cross-clamping, consideration
should be given to preoperative computed tomography to assist in planning the operative
approach.

Tailoring, flushing, tunneling, and suturing of the prosthetic graft are all potential
sources of technical error that may result in early postoperative thrombosis. If the com-
mon trunk of the graft is left too long, the iliac limb may kink at its origin, resulting in
a flow-limiting stenosis (Fig. 1). Thrombus that accumulates in the graft at the time of
placement must be carefully evacuated prior to clamp release. External compression by the

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

AORTIC BIFURCATION GRAFT LIMB OCCLUSIONS

281

-j

o

<

U

a
id

£

CM

Q)

.Q
TO

BJj

o
>.

o

o

X!

Ki

o

a

s

|

o

CD

c/;

a>

<s.

T3

^_

C

C

03

>,

;-^

■—

nj

•—

c

a

^H

J3

C3

C

S

o

c

crt

cfi

_2

G

U

Q

C

s

o

'7}

=

5

O

c

o

o

n

'7}

>

u

u

o

c

'—
d

o

5j

'crt

>>

J2

O

5

u

c

'—

o

■—

y:

tu

X>

B

3

Z

■—

c

J2

3

<

m fN m

V vo

^H O on

cc « en

we ^ ^ j^

«" ^ \ u

(N o N ©"-

(N M3 On.

y Cd X

£ *

CQ CQ W

-c

13

a"

c

c

£ «

3 3

u to

S3 c3

c3 o

O o

3 u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

282

EGINTON and CAMBRIA

(A)

I

•5

(B)

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC BIFURCATION GRAFT LIMB OCCLUSIONS 283

inguinal ligament has been described and can be avoided by careful examination of the
tunnel into the retroperitoneum from the groin incision. Twisting of the graft limb as it is
tunneled from the abdomen must be avoided and will obviously lead to decreased patency
if unnoticed. Finally, improper placement of anastomotic sutures or creation of unfavor-
able anastomotic angles can lead to early thrombosis.

Infrequently, early graft occlusion may occur in the absence of an identifiable cause.
Disorders of blood coagulation have been implicated in this process (14). Certainly, given
the relatively high frequency of hypercoagulable states in patients with vascular disease,
this must be a part of the differential diagnosis. Prosthetic grafts are thrombogenic by their
nature, and some patients may be prone to thrombosis of these grafts even in the absence
of a distinct flow-limiting lesion. Perioperative fluid shifts and alterations in blood pressure
and coagulability may contribute to early graft thrombosis in this setting.

B. Late Failure

The vast majority of late graft failures are due to compromised outflow. This may be due
to intimal hyperplasia at the distal anastomosis in the intermediate time periods, up to 2 or
3 years following the primary procedure. More commonly, gradual progression of
atherosclerotic occlusive disease results in compromise of the outflow vessels and eventual
thrombosis of the graft. This may take the form of superficial femoral artery occlusion
where this vessel had been previously patent, progressive profunda femoris disease where
this had been the primary outflow vessel, or occlusion of an infrainguinal bypass that had
been contributing to graft outflow. As the flow through the graft becomes diminished,
progressive deposition of luminal thrombus leads to thickening of the graft neointima.
Eventually, flow is reduced to a point where thrombosis of the graft occurs.

Less common causes of late graft failure include thrombosis of an anastomotic
pseudoaneurysm, infection of the bypass conduit, embolism from a more central source,
or progression of aortic disease above the conduit, resulting in compromised inflow.
Although problems with inflow can lead to unilateral limb thrombosis, this type of
problem more commonly results in occlusion of the entire aortic reconstruction. With the
increased utilization of intravascular therapy — including coronary angioplasty and stent-
ing — access to the vascular space is frequently gained through previous aortic bifurcation
grafts. In this instance, catheter disruption of graft pseudointima may lead to thrombosis
of a graft that had otherwise been functioning well.

III. DIAGNOSIS AND EVALUATION

Acute occlusion of an aortic graft limb typically presents with the abrupt onset of lower

extremity ischemic symptoms. The majority of patients will have limb-threatening

ischemia with rest pain, requiring urgent intervention (4,15). Less frequently, patients 13

may simply present complaining of decreased exercise tolerance, or with an occult limb |

occlusion. An absent femoral pulse usually confirms the diagnosis, and a diminution in j§

high thigh pressure or ankle brachial index supports the presumptive diagnosis if there is 2

Figure 1 Anteroposterior (A) and left anterior oblique (B) views of an aortogram that demon- 3

strates kinking of both graft limbs. This aortoiliac graft had been placed 6 years earlier for 2

aneurysmal disease. Excessive graft length has resulted in limb kinking, which is a potential cause of I

graft limb occlusion. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

284 EGINTON and CAMBRIA

doubt. As the diagnostic workup continues, heparin anticoagulation should be instituted
to prevent thrombus propagation and maintain patency in the outflow vessels.

A minority of patients will present without detectable blood flow in the extremity
and with a cadaveric foot. Further diagnostic evaluation is contraindicated in this situa-
tion, and the patient should be explored on an emergency basis, with restoration of in-
flow and evaluation of outflow performed intraoperatively. More commonly, patients
will present with an ischemic but viable extremity. In this instance, angiography can pro-
vide important information to assist in planning reconstructive intervention and should
be obtained.

The aortic anastomosis and more proximal aorta should be visualized using biplanar
high-resolution images to exclude anastomotic abnormalities or compromised inflow. The
contralateral iliac or femoral limb should be examined for intrinsic stenoses, anastomotic
strictures, or distal emboli. Finally, the outflow vessels in the ischemic extremity should be
visualized if possible, looking for occlusive disease distal to the groin that may need to be
addressed. It should be noted that even if no outflow vessels can be visualized in the
ischemic extremity, exploration of the groin frequently reveals a suitable profunda femoris
vessel to provide outflow for the graft revision. Computed tomography can occasionally
be useful, particularly when intraluminal aortic thrombus, anastomotic pseudoaneurysms,
or graft infection is suspected.

Given the likelihood that operative intervention will be required to revise the graft,
medical comorbidities should be evaluated, and their management optimized as the
diagnostic studies are obtained. The approach to graft revision may be tailored based
on these factors, avoiding abdominal exploration and "redo" aortic dissection in patients
with prohibitive risk profiles.

IV. MANAGEMENT OF THE OCCLUDED LIMB

Intervention for graft limb occlusion should be individualized to the clinical scenario.
Options in the management of these patients include observation without intervention,
simple thrombectomy, thrombectomy with outflow reconstruction, thrombolysis, redo
aortofemoral reconstruction, or extra-anatomic inflow reconstruction. Factors that influ-
ence decision making in this situation include but are not limited to the degree of extremity
ischemia, adequacy of contralateral inflow, interval since original inflow reconstruction,
interval since graft limb occlusion, angiographic findings, and the general medical con-
dition of the patient. It is well recognized that reoperative intervention for aortic graft
failure is technically demanding; therefore all of the above options should be considered,
as each may have advantages or disadvantages in any given patient.

As noted above, a minority of patients with graft limb occlusion may have this found
incidentally and be without symptoms or may present with a chronic history of decreased |

exercise tolerance in the absence of limb-threatening ischemia. While reintervention to <S

relieve claudication may be warranted, nonoperative management is preferable in certain ^

patients with significant comorbidity and in asymptomatic patients. a

4

A. Restoration of Inflow |

The majority of patients will be able to be treated by retrograde graft thrombectomy with «

outflow reconstruction. However, the surgeon must be prepared to find an alternate source |

of inflow or to perform more distal revascularization, as the situation dictates. With this in @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC BIFURCATION GRAFT LIMB OCCLUSIONS 285

mind, the patient should be widely prepped and draped to include the entire chest,
abdomen, both groins, and the involved lower extremity. General anesthesia is preferred,
although, in extenuating circumstances, thrombectomy under local anesthesia may be
possible.

The previous groin incision is reopened, and the thrombosed graft limb, common
femoral artery, superficial femoral and profunda femoris arteries are dissected and
isolated. Heparin should be administered to elevate the activated clotting time above
200-250 s if this drug had not already been initiated preoperatively. A transverse incision
in the graft limb close to the femoral anastomosis allows for exploration of the runoff
vessels. If necessary, thrombectomy catheters can be passed distally to retrieve thrombotic
material and achieve back-bleeding. If the orifices of the runoff vessels are involved with
occlusive disease, reconstruction to improve outflow is necessary (see below).

Thrombectomy catheters are then passed proximally. This should be done serially,
inserting the catheter 5 cm initially, and then 10 cm, and so on. Thrombotic material is
removed sequentially in this fashion to avoid pushing the thrombus over the bifurcation of
the graft and causing an embolus in the contralateral limb. Pressure on the contralateral
femoral pulse, occluding flow in the opposite limb while the thrombectomy is being
performed, also serves to prevent embolism. Once the thrombotic material is cleared,
pulsatile inflow should be restored. In early postoperative limb thrombosis, careful
examination for technical defects within the graft or at the anastomoses should follow.
If no cause for failure is identified, the graft can be closed and the patient managed with
short-term anticoagulation.

In all cases of late graft limb occlusion or if any doubt exists as to the adequacy of
inflow or the presence of residual thrombus or technical defect, further investigation is
warranted after initial thrombectomy. Surgeons comfortable with angioscopy have ad-
vocated the use of this instrument for evaluation of the graft limb following thrombectomy
(16). More commonly, retrograde injection of contrast material from the groin for angiog-
raphic analysis is performed. Intraluminal filling defects in the graft limb represent either
residual thrombus or graft pseudointima. Specialized graft thrombectomy catheters are
available and should be used in this situation. These catheters have wire loops instead of a
balloon at the tip to scrape material from the wall of the graft. In the past, ring strippers
have been used over an occlusion balloon more proximally (17), and this technique may
still be useful in selected instances. However, the graft thrombectomy catheters have been
quite effective in removing adherent thrombus (Fig. 2). Angiography to confirm adequate
clearance of the graft limb should be obtained once inflow is restored.

If adequate inflow cannot be obtained with thrombectomy catheters, alternative
sources of inflow must be considered; these include redo aortofemoral reconstruction or
extra-anatomic reconstruction (cross-femoral or axillofemoral bypass). In the absence of
stenotic lesions in the contralateral limb, a cross-femoral graft is frequently the best ■a

option. Primary and secondary 5-year patency rates of 54 and 84%, respectively, have |

been reported for femorofemoral crossover grafts used to treat aortic graft limb occlusion, a

with a limb salvage rate of 84% (18). Revision or replacement of the entire prosthesis is c

generally required for lesions at the proximal anastomosis or in the aorta above, pre- <

eluding establishment of adequate inflow (Fig. 3). This was required in less than 3% of >9

aortobifemoral bypass procedures reviewed by Szilagyi (19). In those patients with pro- J

hibitive risk profiles, axillobifemoral bypass is a viable alternative to aortic reoperation. «

Thrombolytic agents have been used successfully to reestablish patency of a throm- |

bosed aortofemoral graft limb. As in other situations where this type of treatment is @

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

286

EGINTON and CAMBRIA

(A)

(C)

I

•5

(B)

(D)

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC BIFURCATION GRAFT LIMB OCCLUSIONS

287

Figure 3 Adherent thrombus at the proximal anastomosis of an aortobifemoral bypass graft. This
patient was referred after her aortic graft had occluded and had been treated with thrombolytic
therapy. The intraluminal filling defect persisted and was treated by revision of her proximal anas-
tomosis to the more proximal infrarenal aorta.

employed, thrombolytic agents require more time to restore inflow than direct operative
thrombectomy; typically, additional procedures will be required to correct the underlying
cause of graft limb occlusion. In one series of 19 patients (20), restoration of inflow was
achieved in every case with urokinase, with an average infusion time of 32 h (range 12-60
h). Additional treatment was required in 16 cases, the majority of which required
conventional surgical repair. Occasionally, focal stenoses above the inguinal ligament
may be treatable with endovascular techniques (20,21). Because of the time required for
thrombolysis, this approach should not be considered when the involved extremity is

Figure 2 The tools of the trade. (A.) Syntel balloon thrombectomy catheter. (B.) Fogarty adherent
clot catheter for use in grafts or arteries. Wire spiral is covered with latex to minimize intimal injury.
(C.) Fogarty graft thrombectomy catheter. Exposed wire hoops for use in removal of neointima and
thrombus from luminal surface of prosthetic grafts. Not intended for use in native vessels. (D.)
Endarterectomy loop. May be used coaxially with balloon occlusion catheter to facilitate removal of
densely adherent thrombotic plug or graft neointima.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

288

EGINTON and CAMBRIA

profoundly ischemic. Given that the majority of patients will require operative revision,
many have argued effectively that operative thrombectomy is preferred in most instances
(4,14). Thrombolytic therapy may be quite helpful in selected circumstances where the
etiology of recurrent graft limb thrombosis remains elusive.

B. Outflow Revision

As noted above, the majority of late graft failures result from diminished outflow due to
anastomotic stricture or progression of distal disease. Patency of graft limb thrombectomy
in this instance is directly related to either correction or bypass of stenotic disease in the
groin or lower extremity (Fig. 4). Preoperative angiography may identify significant
femoropopliteal occlusive disease, but intraoperative evaluation of the femoral artery
branches is mandatory to assure adequate graft outflow.

The profunda femoris (deep femoral) artery is the primary outflow vessel in the
majority of patients with graft limb occlusion. The orifice of this vessel should be explored
at the time of thrombectomy and should admit a 4-mm dilator. Additional assessment
could include intraoperative angiography or gentle passage of embolectomy catheters. A

I
•5

Figure 4 Right femoral anastomosis of an aortobifemoral graft 9 years following implantation.
Stenosis at the origin of profunda femoris runoff ultimately resulted in occlusion of this graft limb.
Thrombectomy with profundaplasty was performed, but the graft limb reoccluded within several
months. Repeat thrombectomy with adjunctive femoropopliteal bypass resulted in continued pa-
tency of the graft limb.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC BIFURCATION GRAFT LIMB OCCLUSIONS

289

profunda vessel length of 20-25 cm has been suggested to be adequate to support graft
limb outflow. Stenosis at the origin of the profunda vessel should be treated with
profundaplasty. Exposure of the vessel followed by longitudinal arteriotomy, endarter-
ectomy, and patch angioplasty is usually possible and will provide adequate outflow in the
majority of cases. While some authors have advocated the benefits of autogenous tissue
patch angioplasty of the profunda vessel (22), the distal anastomosis can usually be
fashioned to incorporate the profunda femoris angioplasty with equivalent results.
Alternatively, a stenosis at the origin of the deep femoral artery can be bypassed by
moving the distal anastomosis farther down this vessel (23).

In the absence of an adequate deep femoral artery, aortic graft limb outflow will
require adjunctive bypass in the lower extremity. In this instance, either preoperative or
intraoperative angiography is mandatory to identify a suitable target vessel in continuity
with the infrageniculate arteries. Infrainguinal bypass was required in 10-15% of patients
following aortic reconstruction in one series (24). While distal bypass is required in the
minority of patients with initial graft limb occlusion, this rate increases in those who
present with recurrent limb occlusion (4).

V. RESULTS

Morbidity and mortality following operations for occluded aortofemoral bypass limbs
are low. Mortality rates ranging from 2 to 5% have been quoted in the literature
(4,13,14). Local complications include groin wound complications, graft infection,

1 00

80

I

5.

I

60

40

20

_L

J_

3 4 5 6 7

Years following reoperation

10

I
•5

Figure 5 Primary (circles) and secondary (triangles) patency following graft limb thrombectomy in
110 patients. The utility of repeated thrombectomy for graft limb reocclusion is evidenced by the
dramatic improvement in secondary patency. (Adapted from Ref. 4.)

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

290 EGINTON and CAMBRIA

pseudoaneurysm, and embolism to the ipsilateral or contralateral extremity. Systemic
complications are typical of the vascular patient population, with cardiovascular events
predominating. Taken together, complication rate for graft limb occlusion is on the order
of 10-15%.

Operations to restore patency for an occluded aortic graft limb are quite successful.
Primary patency rates following thrombectomy and outflow revision are 70 and 47% at 1
and 5 years, respectively (4). Repeat thrombectomy is valuable in the event of reocclusion,
with 5-year patency rates approaching 75% following aggressive rethrombectomy with
adjunctive procedures when necessary (Fig. 5). Many authors have emphasized the benefit
of repeated operations for recurrent graft limb occlusion, with this type of diligence
resulting in enhanced secondary patency.

VI. SUMMARY

Occlusion of an aortic graft limb represents the most common complication following
reconstruction for aortoiliac occlusive disease. Perioperative failure represents a minority
of these patients ( < 10%), and usually results from a technical or judgmental error. The
vast majority of limb occlusions occur from 2 to 5 years following graft implantation and
are associated with atherosclerotic deterioration of the outflow vessels. These patients
typically present with limb-threatening ischemia or claudication, which is more severe than
the symptoms that had prompted the original procedure. Preoperative imaging is desirable
to rule out unsuspected inflow restriction and evaluate potential outflow stenoses. While
thrombolytic therapy has been employed successfully, most patients go on to require
operative intervention. Therefore direct operative thrombectomy with concomitant out-
flow revision represents the most expeditious and efficient way to deal with this problem. A
minority of patients will require alternative inflow procedures or adjunctive distal bypass
in the lower extremity. Aggressive management of limb occlusions with reexploration for
recurrent thrombosis has resulted in a secondary 5-year patency approaching 75% after
graft limb occlusion.

REFERENCES

1. Martinez BD, Hertzer NR, Beven EG. Influence of distal arterial occlusive disease on prognosis
following aortobifemoral bypass. Surgery 1980; 88:795-805.

2. Poulias GE, Polemis L, Skoutas B, Doundoulakis N, Papaioannou K, Ershaid B, Sendekeya S.
Bilateral aortofemoral bypass in the presence of aorto-iliac occlusive disease and factors
determining results. J Cardiovasc Surg 1985; 26:527-538.

3. Wray R, DePalma RG, Hubay CH. Late occlusion of aortofemoral bypass grafts: influence of

cigarette smoking. Surgery 1971; 70:969-973. ■o

4. Brewster DC, Meier GH III, Darling RC, Moncure AC, LaMuraglia GM, Abbott WM. Re- |
operation for aortofemoral graft limb occlusion: optimal methods and long-term results. J Vase S
Surg 1987; 5:363-374. 1

5. Nevelsteen A, Suy R, Daenen W, Boel A, Stalpaert G. Aortofemoral grafting: factors influencing a
late results. Surgery 1980; 88:642-653. 6

6. Brewster DC, Darling RC. Optimal methods of aortoiliac reconstruction. Surgery 1978; 84:739- Tj
748. |

7. Brewster DC. Current controversies in the management of aortoiliac occlusive disease. J Vase 2
Surg 1997; 25:365-379. I

8. Robicsek F, Daugherty HK, Mullen DC, Masters TN, Narbay D, Sanger PW. The effect of §

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC BIFURCATION GRAFT LIMB OCCLUSIONS 291

continued cigarette smoking on the patency of synthetic vascular grafts in Leriche syndrome. J
Throac Cardiovasc Surg 1975; 70:107-112.
9. Myers KA, King RB, Scott DF, Johnson N, Morris PJ. The effect of smoking on the late patency
of arterial reconstructions in the legs. Br J Surg 1978; 65:267-271.

10. Malone JM, Moore WS, Goldstone J. The natural history of bilateral aortofemoral bypass grafts
for ischemia of the lower extremities. Arch Surg 1975; 110:1300-1306.

11. Hallett JW, Marshall DM, Petterson TM. Gray DT, Bower TC, Cherry KJ, Gloviczki P,
Pairolero PC. Graft related complications after abdominal aortic aneurysm repair: reassurance
from a 36-year population-based experience. J Vase Surg 1997; 25:277-286.

12. Najafi H, Dye WS, Javid H, Hunter JA, Goldin MD, Serry C, Julian OC. Late thrombosis
affecting one limb of aortic bifurcation graft. Arch Surg 1975; 110:409-412.

13. Bernhard VM, Ray LI, Towne JB. The reoperation of choice for aortofemoral graft occlusion.
Surgery 1977; 82:867-874.

14. Erodes LS, Bernhard VM, Berman SS. Aortofemoral graft occlusion: strategy and timing of
reoperation. Cardiovasc Surg 1995; 3:277-283.

15. Lyons JH, Weismann RE. Surgical management of late closure of aortofemoral reconstruction
grafts. N Engl J Med 1968; 278:1035-1037.

16. Towne JB, Bernhard VM. Technique of intraoperative endoscopic evaluation of occluded aorto-
femoral grafts following thrombectomy. Surg Gynecol Obstet 1979; 148:87-89.

17. Ernst CB, Dougherty ME. Removal of a thrombotic plug from an occluded limb of an aorto-
femoral graft. Arch Surg 1978; 1 13:301-302.

18. Nolan KD, Benjamin ME, Murphy TJ, Pearce WH, McCarthy WJ, Yao JST, Flinn WR.
Femorofemoral bypass for aortofemoral graft limb occlusion: a ten-year experience. J Vase Surg
1994; 19:851-857.

19. Szilagyi DE, Elliot JP, Smith RF, Reddy DJ, McPharlin M. A thirty-year survey of the
reconstructive surgical treatment of aortoiliac occlusive disease. J Vase Surg 1986; 3:421-436.

20. Enron B, Reigner B, Lescalie F, l'Hoste P, Peret M, Chevalier JM. In situ thrombolysis for late
occlusion of suprafemoral prosthetic grafts. Ann Vase Surg 1993; 7:270-274.

2 1 . Mitchell SE, Kadir S, Kaufman SL, Chang R, Williams GM, Kan JS, White RI Jr. Percutaneous
transluminal angioplasty of aortic graft stenoses. Radiology 1983; 149:439-444.

22. Malone JM, Goldstone J, Moore WS. Autogenous profundaplasty: the key to long-term patency
in secondary repair of aortofemoral graft occlusion. Ann Surg 1978; 188:817-823.

23. Ouriel K, DeWeese JA, Ricotta JJ, Green RM. Revascularization of the distal profunda femoris
artery in the reconstructive treatment of aortoiliac occlusive disease. J Vase Surg 1987; 6:217-
220.

24. Baird RJ, Feldman P, Miles JT, Madras PM, Gurry JF. Subsequent downstream repair after
aorto-iliac and aorto-femoral bypass operations. Surgery 1977; 82:785-793.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

15

Problems Related to Extra-Anatomic Bypass — Including

Axillofemoral, Femorofemoral, Obturator, and

Thoracofemoral Bypasses

Kyle Mueller and William H. Pearce

Northwestern University, Chicago, Illinois, U.S.A.

The novel concept of extra-anatomic bypass appeared more than four decades ago and
this procedure has now become a widely used and accepted method of revascularization.
It is mainly performed for patients who, due to comorbid conditions, are poor surgical
risks for standard revascularization procedures or for those who have had a complication
related to a previous bypass, such as an infected aortic graft or aortoduodenal fistula.
The four types of extra-anatomic bypass commonly used include axillofemoral, femoro-
femoral, obturator, and thoracofemoral grafts. Because the majority of patients under-
going these procedures are poor surgical risks, any complication can be catastrophic. For
this reason it is important that the vascular surgeon be aware of the complications
associated with these procedures, so as to avoid complications when possible and to
intervene in a timely manner when they do occur. This chapter reviews the diagnosis and
management of complications commonly associated with extra-anatomic bypass proce-
dures.

I. AXILLOFEMORAL BYPASS

The axillofemoral bypass graft was introduced by Blaisdell and colleagues in the early |

1960s (1,2). It has become the procedure of choice for revascularization in patients with a

aortoiliac occlusion who are found to be at an unacceptably high surgical risk for trans- c

abdominal aortic bypass. Axillofemoral bypass is also an integral technique for the man- <

agement of patients with an infected aortic graft or aortoenteric fistula. In their review of >9

916 axillofemoral grafts reported in the literature, Bunt and Moore found a 1.6% inci- J

dence of complications (15 of 916) (3). Over the past several decades, as the use of this °

technique has dramatically increased, a large number of studies and case reports dealing |

with the specific and unique complications seen with this type of extra-anatomic bypass @

293

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

294 MUELLER and PEARCE

have emerged. These include proximal anastomotic disruption, upper extremity thrombo-
embolic events, graft infection, and several others, which are discussed in this chapter.

A. Proximal Anastomotic Disruption

Disruption of the axillary anastomosis was initially thought to be a sporadic occurrence,
but a study by Taylor et al. reports that 5% of patients (10 of 202) undergoing axillo-
femoral bypass developed this serious complication (4). It occurs primarily in the early
postoperative period, as demonstrated by the Taylor study, with disruption seen over a
range of 1-46 days postoperatively. A study by White et al. demonstrates disruption at 13—
30 days after the procedure (5). The causes of axillary anastomotic disruption include
infection, technical error, and mechanical stress (6,7). The associated clinical findings in-
clude axillary pain, expanding hematoma causing brachial plexus injury, and pseudoaneu-
rysm formation. Axillary anastomotic disruption is most commonly related to mechanical
stress and, more specifically, to overexertion or extreme upper extremity movements. This
complication related to exertional disruption has been reported by several authors and
termed axillary pullout syndrome (5). Daar and Finch describe how full abduction of the
upper extremity and maximal lateral flexion of the spine contralateral to the site of mea-
surement can lead to an increase in the length of the axillofemoral graft pathway of 4-12
cm (8). Another contributing factor to disruption of the axillary anastomosis is the wide-
spread use of an external ring-reinforced polytetrafluoroethylene (PTFE) graft, which is
very inelastic compared to the previously used Dacron grafts.

The key to avoiding proximal anastomotic disruption is correct anatomical placement
of the axillary anastomosis. Blaisdell and Hall (1), in their first report on this technique in
1963, clearly stated that the anastomosis should be placed medial to the pectoralis minor
tendon on the first part of the axillary artery, a statement that they reemphasized in a 1985
study (9). This first part of the axillary artery has no branches, is less mobile, and should
be the only site used for anastomosis. It is not necessary to divide the pectoralis minor
muscle in order to expose the first part of the axillary artery; doing so only increases the
risk of anastomotic complications. Another technical point is to perform the axillofemoral
bypass procedure with the upper extremity abducted, which creates some redundancy in
the graft, permitting the pathway length to increase as the arm goes through a full range of
motion. Landry et al. described a lateral tunneled approach with an anterior anastomosis
to the axillary artery (Fig. 1) (10). The management of axillary anastomotic disruption
involves obtaining proximal control of the subclavian artery using a supraclavicular ap-
proach or balloon occlusion. Because simple repair of disruptions has been shown to lead
to secondary disruption, repair should be performed by placing a new interposed graft
with reanastomosis to the existing graft. In addition, wound cultures should be sent to

make sure that infection was not a contributing factor to the disruption.

P

B. Upper Extremity Thromboembolic Events J

Upper extremity thromboembolism is a rare complication of axillofemoral bypass that can c

occur early or late in the postoperative course; in the majority of cases, however, it is <

associated with axillofemoral graft occlusion. A 1995 study by McLafferty et al. found &

that upper extremity thromboembolism occurred in 2.7% of all patients after axillofe- J

moral bypass and in 25% of patients with occluded grafts (11). These events occurred as «

early as 26 days and as late as 7 years postoperatively. Thrombosis of the donor axillary |

artery has been reported, but the incidence is very low (12). The common finding in several @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PROBLEMS RELATED TO EXTRA-ANATOMIC BYPASS

295

Figure 1 Alternative technique for the axillary anastomosis for an axillobifemoral bypass graft.
This technique allows for redundancy along the course of the subclavian and axillary arteries for
movement of the upper extremity. (From Ref. 10.)

cases of axillary artery thrombosis has been downward traction on the donor artery by the
bypass graft, leading to a sharp angulation and subsequent thrombosis (13,14). Late severe
ischemia of the upper extremity after axillofemoral bypass is rare, but there have been
several reported cases of thromboembolic events after the procedure that caused severe
ischemia, requiring amputation (15). If the site of the axillary anastomosis is appropriate,
thrombosis seldom presents with upper extremity ischemia (Fig. 2). The first part of the
axillary artery is devoid of significant branches, and all other areas proximal and distal to
this segment have large collateral branches. If thrombosis occurs in a correctly placed
axillary anastomosis, the collaterals will perfuse the arm; but if the anastomosis is placed
in a segment with collaterals and thrombosis occurs, the collateral flow is interrupted and
perfusion of the arm may be jeopardized. Management of axillary artery thrombosis
should involve immediate exploration of the proximal anastomosis and repair of the
axillary artery rather than anticoagulation or an attempt to reach the thrombosis from a
distal approach.

In addition to thrombosis of the axillofemoral graft at the axillary anastomosis, there
are reports of thrombosis occurring in the body of the graft due to compression from

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

296

MUELLER and PEARCE

Figure 2 Axillopopliteal graft with excessive tension at the proximal anastomosis. Over time, the
axillary-subclavian artery elongates to present with kinking and potential sources of distal embo-
lization.

external body weight and kinking or twisting of the body of the graft, leading to graft
occlusion. The effect of body-weight compression has been debated, as Jarowenko et al.
found no hemodynamic changes after external body-weight compression, while Cavallaro
et al. demonstrated changes in graft hemodynamics using ankle pressure index and pulse
volume recording while patients were lying on the side of the bypass (16,17). Because the
use of external ring-reinforced polytetrefluoroethylene (PTFE) grafts has increased,
compression by body weight is seldom the cause of graft thrombosis. Redundancy in
the graft has been shown to contribute to twisting or kinking of the body of the graft
leading to stenosis and occlusion (Fig. 3). This is exaggerated in elderly patients who are in
a seated position and may be developing mild kyphosis from degenerative changes of the
spine. This complication demonstrates the balance required to provide some redundancy,
allowing for increases in the length of the graft pathway for arm movement, yet avoiding
excessive redundancy which could lead to thrombosis. Treatment is simply resection of the
redundant graft with end to end anastomosis.

Distal embolization of the upper extremity following axillofemoral bypass is an
extremely rare complication and primarily occurs after graft thrombosis. Bandyk et al.
reported four episodes of distal embolic events and described the source of the emboli as
the blind stump of the proximal graft limb that had remained patent (18). Embolization to
the brachial artery can cause severe, limb-threatening ischemia. Management of this
complication varies from brachial embolectomy to prophylactic detachment of the
proximal graft and patch angioplasty of the axillary artery in patients who do not need
immediate revision of their occluded bypass graft.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PROBLEMS RELATED TO EXTRA-ANATOMIC BYPASS

297

Figure 3 Lateral photograph of a tortuous axillofemoral bypass. Occasionally, such tortuosity is
severe, producing flow-limiting lesions.

C. Axillofemoral Graft Infection

Infection of the axillofemoral graft is a rare but serious complication. In their review of
55 axillopopliteal bypass grafts, Ascher et al. found a 3.6% incidence of infection (19).
Marston et al. demonstrated that infection of axillofemoral grafts is associated with a
perioperative mortality rate of 22%, and 57% of survivors required amputation (20).
Infection often occurs in the groin area, but midgraft infection has also been observed.
Because of the excellent incorporation of PTFE graft by surrounding tissue, removal of
the entire graft is seldom needed. Most graft infections can be managed with local
debridement, systemic antibiotic therapy, and aggressive local wound care (Fig. 4A and
B). If necessary, the infected segment can be removed and replaced with an interposed
autogenous graft. Should the infection necessitate removal of the entire graft, an alternative
reconstruction is a bypass from the descending thoracic aorta to the femoral artery, which
is described further on in this chapter.

D. Perigraft Seroma

Perigraft seroma is a rare complication that has been reported to occur in various bypass
procedures using Dacron or PTFE. A study by Buche et al. describes three cases of
perigraft seroma in a total of 123 axillofemoral bypass grafts (21). The clinical findings of
perigraft seroma are a nonpulsatile, painless mass overlying a graft, which must be
distinguished from a pseudoaneurysm. The etiology of perigraft seroma is unclear, but its
formation is related to inadequate incorporation of the graft by the surrounding
connective tissue. Histologically, the seroma is confined within a pseudocapsule formed
by a thin fibrous membrane. The pseudocapsule usually contains a clear, sterile serous

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

298

MUELLER and PEARCE

Figure 4 A. Infected axillofemoral bypass. The patient had previously undergone removal of the
axillofemoral bypass with a stump of the graft material left. B. An arteriogram demonstrating the
residual stump of the prosthetic material, which had become infected. The artery was repaired with
an interposition vein graft.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PROBLEMS RELATED TO EXTRA-ANATOMIC BYPASS 299

fluid, but gelatinous material has been reported (22,23). A 1994 study by Ahn et al. re-
ported the identification of a fibroblast growth inhibition factor from the sera of patients
with perigraft seromas (24). Chronic peri graft seroma often becomes apparent at 4-8 weeks
after graft implantation. Diagnosis can be made by duplex scan, computed tomography, or
direct aspiration of the seroma. Some authors advocate observation of perigraft seromas,
but standard treatment is excision with new graft replacement. Since perigraft seromas
are often related to graft biofilm infections, special culture techniques are necessary for the
diagnosis.

E. Structural Failure

The majority of axillofemoral bypasses are performed using external ring-reinforced PTFE
grafts, which offer greater strength and durability than knitted Dacron. Despite this
advantage, several authors have reported material failure of PTFE grafts, leading to graft
disruption or pseudoaneurysm (25). We as well as other authors have observed disruption
of PTFE grafts with suture lines remaining intact (26). According to White et al. (5), more
than 40 such cases have been recorded, suggesting that this problem is not as rare as
previously thought. Another serious complication related to the mechanical failure of
PTFE grafts is the formation of pseudoaneurysms in the body of the bypass graft. There
have been documented cases of pseudoaneurysms developing in axillofemoral grafts after
direct trauma, but a 1993 report by Piazza et al. describes a nonanastomotic pseudo-
aneurysm occurring without any trauma (27). Another report describes multiple aneurysms
in an axillofemoral graft with coagulopathy (28). The management of such complica-
tions from structural failure includes prompt surgical exploration with replacement of the
affected portion of the graft using a short segment interposed graft.

II. FEMOROFEMORAL GRAFT

Since the technique was introduced by Oudot and Beaconfield and Freeman and Leeds in
the early 1950s and popularized by Vetto in the 1960s, the femorofemoral graft has been
shown to have good long-term patency and is now considered a permanent reconstructive
procedure (29-31). The femorofemoral graft is an acceptable alternative bypass to aorto-
femoral reconstruction when unilateral iliac artery occlusion is present; it has also been
used for the management of aortofemoral graft limb occlusion (32). Initially the procedure
was reserved for high-risk older patients, but it is now probably the procedure of choice in
the young adult when donor iliac arterial occlusive disease is minimal and avoidance of
sexual dysfunction is important.

Early complications of femorofemoral bypass grafts are similar to those associated
with other bypass procedures; they include hemorrhage, thrombosis, hematoma, and local ■§

wound complications. A complication unique to femorofemoral bypass is penile swelling g

or scrotal hematoma related to the tunneling of the graft. These result from attempting »

to pass the graft retrofascially with disruption of the venous plexus in Retzius's space. As c

in the case of axillofemoral bypass, the graft pathway is extremely important for femoro- <

femoral bypass. Kinking of the graft at the proximal or distal anastamosis, leading to >9

stenosis or thrombosis, can occur if care is not taken in aligning the graft with both the J

donor and the recipient arteries. To avoid this complication, the graft should be placed «

in the subcutaneous tunnel first. Then the arteriotomies and subsequent anastomoses |

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

300 MUELLER and PEARCE

are made based on the angle of the graft with the donor and recipient arteries and on
whether the anastomosis is to be constructed in a C or S manner.

Late complications include the formation of anastomotic aneurysm, distal emboliza-
tion, and development of the steal phenomenon. With the increased use of PTFE grafts, it
is now uncommon to encounter an anastomotic aneurysm in femorofemoral grafts.
Thrombosis in a femorofemoral graft and distal embolization to the superficial femoral
or popliteal artery have been reported (33). Donor limb steal has been observed, especially
when the graft is performed for intermittent claudication. The decrease of donor limb
ankle pressure seen in our series was caused by either unrecognized iliac disease or
multilevel disease (34). For patients with intermittent claudication, ankle pressure of the
donor limb must be evaluated by treadmill exercise testing. A decrease in ankle pressure
after exercise of the donor limb increases the likelihood of the steal phenomenon if a
femorofemoral graft is performed. Other hemodynamic tests — such as duplex scanning
and intra-arterial pressure measurement of the femoral artery — may also help to detect
hemodynamically significant iliac artery disease. When, after femorofemoral bypass, a
large femoral/brachial pressure gradient correlating to steal is measured, one alternative
treatment is percutaneous transluminal angioplasty of the common iliac artery (35).
Treatment of superficial groin infections after femorofemoral bypass involves systemic
antibiotics and local wound care. Infection of the femorofemoral bypass is a serious
complication, as demonstrated by a 1995 study by de Virgilio et al. describing a mortality
rate of 20% and an amputation rate of 10% (36). Treatment requires prompt graft
removal and reconstitution of blood flow by another route.

III. OBTURATOR BYPASS

Obturator bypass is often performed in patients with an infected groin or dense scar tissue
as a result of multiple groin dissections or postradiation changes. The procedure first
described by Shaw and Baue is now regarded as acceptable for bypass (37).

Complications of the procedure include injury to the obturator artery, obturator
nerve, or the genitourinary tract. The original description of the procedure called
for passing the tunneling device from above after identification of the obturator foramen.
Such a maneuver may be difficult and injury to the obturator nerve or artery during
dissection is a real possibility. Injury to the obturator nerve is manifest by pain radiating
from the groin to the medial aspect of the knee. Paresthesia and hyperesthesia may
also occur. Motor dysfunction produces a wide-based gait that results from adductor
muscle weakness.

Electromyography or obturator nerve blocks are useful to establish the diagnosis (38).
Injury of the obturator artery may result in a retroperitoneal hematoma or excessive
blood loss. The urological complications that have been reported are perforation of ■§

the bladder and transection of the ipsilateral ureter (39). We have observed bladder injury g

in 1 of 14 patients who underwent this procedure. This patient had a history of pelvic as

surgery followed by radiation therapy and should not have been considered for this type c

of bypass. <

To simplify this procedure, we favor passing the tunneling device from below (40,41). >9

After an incision is made just below the adductor longus tendon in the midthigh, the J

tunnel is created below the adductor longus and magnus muscles with the leg abducted «

and externally rotated. Next, a large DeBakey clamp is passed upward from below and |

directed toward the obturator foramen. At this time the operator's hand is placed over the @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PROBLEMS RELATED TO EXTRA-ANATOMIC BYPASS 301

obturator foramen and the membrane penetrated. The graft then is drawn from above into
the thigh to be anastomosed to the popliteal artery.

One of the interesting complications of obturator bypass is interval gangrene of the
thigh. Unless there are sufficient collateral pathways connecting the popliteal and tibial
trunks with the profunda femoris, proximal myonecrosis may occur because the obturator
graft bypasses the profunda femoris artery at the groin. This complication can be avoided
by including the profunda femoris artery in the reconstruction.

IV. THORACOFEMORAL BYPASS

Extra-anatomic bypass from the descending thoracic aorta to the femoral artery was first
introduced by Blaisdell and coworkers in 1961 (42). Since its development, the procedure
has been used by many surgeons for various indications. We use this type of bypass in
patients for the following reasons: (a) conversion of existing or recently failed axillopop-
liteal or axillofemoral bypasses originally placed for septic abdominal aortic indications
(infected aortic grafts or aortoduodenal fistulas), (b) avoidance of redissection of the
retroperitoneum in patients after many other complex operations or infections, or (c)
multiple failed aortofemoral bypasses. In each instance the infrarenal abdominal aorta is
relatively inaccessible to reoperation (43). The procedure is rather simple and the
descending thoracic aorta is approached through a posterolateral thoracotomy. The
graft is then tunneled in a transdiaphragmatic retroperitoneal anterior axillary line to
the left femoral artery. A femorofemoral bypass is then added to complete the
revascularization. The technique has not been widely used; therefore the incidence of
complications is unknown. In our series the 4-year patency rate of the thoracic aorta to
femoral artery bypass was 100%, with a single graft failing at 49 months (44). There
were 5 total complications in 21 patients, including a fracture of the spleen during
tunneling requiring splenectomy and a bladder injury with retrofascial tunneling
during the femorofemoral bypass portion of the procedure. The remaining complications
were a minor vein injury during tunneling and reoperation for persistent thoracic
bleeding in another patient. Another patient developed a late complication of thoracic
aorta infection from an ascending groin infection after two failed lower extremity
bypass procedures. A recent study by Passman et al. describes a 5-year primary patency
rate for bypass procedures from the descending thoracic aorta to the iliofemoral artery
of 79%, supporting a more widespread use of this durable procedure for primary
revascularization (45).

V. CONCLUSION

Extra-anatomic bypass grafts are useful and often provide an effective means to restore ■§

blood flow in special conditions and in a unique population of patients. Although some of g

these grafts may be considered temporary, enough data have now been accumulated to »

demonstrate that femorofemoral grafts may be considered permanent reconstructions and c

the other grafts to be more durable than anticipated. Similarly, the frequent use of external <

ring-reinforced PTFE grafts may provide patients with a much better patency rate and >9

fewer graft complications than previously reported. Extra-anatomic bypass represents an ^

advance in techniques in vascular surgery. With the rather low incidence of complications «

and an aging population, these procedures should remain viable and important alternate |

surgical techniques for revascularization. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

302 MUELLER and PEARCE

REFERENCES

1. Blaisdell FW, Hall AD. Axillary-femoral artery bypass for lower extremity ischemia. Surgery
1963; 54:563-568.

2. Louw JH. Splenic to femoral and axillary to femoral bypass grafts in diffuse arteriosclerotic
occlusive disease. Lancet 1963; 1:401-402.

3. Bunt TJ, Moore W. Optimal proximal anastomosis/tunnel for axillofemoral grafts. J Vase Surg
1986; 3:673-676.

4. Taylor LM Jr, Park TC, Edwards JM, Yeager RA, McConnell DC, Moneta GA, Porer JM,
McConnell DC, Moneta GA, Porter JM. Acute disruption of polytetrafluoroethylene grafts
adjacent to axillary anastomoses: a complication of axillogemoral grafting. J Vase Surg 1994;
20:520-526.

5. White GH, Donayre CE, Williams RA, White RA, Stabile BE, Wilson SE. Exertional disrup-
tion of axillofemoral graft anastomosis — "The axillary pull-out syndrome". Arch Surg 1990;
125:625-627.

6. Yeager RA, Taylor LM Jr. Axillary artery anastomosis to avoid axillofemoral bypass dis-
ruption. Semin Vase Surg 2000; 13:74-76.

7. Sullivan LP, Davidson PG, D'Anna JA, Sithian N. Disruption of the proximal anastomosis of
axillofemoral grafts: Two case reports. J Vase Surg 1989; 10:190-192.

8. Darr AS, Finch DRA. Graft avulsion: An unreported complication of axillofemoral bypass
grafts. Br J Surg 1978; 65:442^146.

9. Blaisdell FW. Late axillary thrombosis in patients with occluded axillary-femoral bypass grafts.
J Vase Surg 1985; 2:925.

10. Landry GJ, Moneta GL, Taylor LM, Porter JM. Axillobifemoral bypass. Ann Vase Surg 2000;
14:296-305.

11. McLafferty RB, Taylor LM Jr, Moneta GL, Yeager RA, Edwards JM, Porter JM. Upper ex-
tremity thromboembolism caused by occlusion of axillofemoral grafts. Am J Surg 1995; 169:
492-495.

12. Kempczinski R, Penn I. Upper extremity complication of axillofemoral grafts. Am J Surg 1978;
136:209-211.

13. Rashleigh-Belcher HJC, Newcombe JF. Axillary artery thrombosis: A complication of axillo-
femoral bypass grafts. Surgery 1987; 101:373-375.

14. Farina C, Schultz RD, Feldhaus RJ. Late upper limb acute ischemia in a patient with an
occluded axillofemoral bypass graft. J Cardiovasc Surg 1990; 31:178-181.

15. Hartman AR, Fried KS, Khalil I, Riles TS. Late axillary artery thrombosis in patients with
occluded axillary-femoral bypass grafts. J Vase Surg 1985; 2:285-287.

16. Jarowenko MV, Buchbinder D, Shah DM. Effect of external pressure on axillo-femoral bypass
grafts. Arch Surg 1981; 193:274-276.

17. Cavallaro A, Sclacca V, di Marzo LD, Bove S, Mingoli A. The effect of body weight
compression on axillo-femoral bypass patency. J Cardiovasc Surg 1988; 29:476^179.

18. Bandyk DF, Thiele BG, Radke HM. Upper extremity embolus secondary to axillofemoral
bypass grafts. Arch Surg 1983; 118:673-676.

19. Ascher E, Veith FJ, Gupta S. Axillopopliteal bypass grafting: Indications, late results, and 1
determinants of long-term patency. J Vase Surg 1989; 10:285-291. |

20. Marston WA, Risley GL, Criado E, Burnham SJ, Keagy BA. Management of failed and j§
infected axillofemoral grafts. J Vase Surg 1994; 20:357-365. j?

21. Buche M, Schoevaerdts JC, Jaumin P, Ponlot R, Chalant CH. Perigraft seroma following <
axillofemoral bypass: Report of three cases. Ann Vase Surg 1986; 1:374-377. Jq

22. Blumberg RM, Gelfand ML, Dale WA. Perigraft seromas complicating arterial grafts. J §
Cardiovasc Surg 1983; 24:372. g

23. Borreor E, Doscher W. Chronic perigraft seromas in PTFE grafts. J Cardiovasc Surg 1988; "|
29:46-49. J

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PROBLEMS RELATED TO EXTRA-ANATOMIC BYPASS 303

24. Aim SS, Williams DE, Thye DA, Cheng KQ, Lee DA. The isolation of a fibroblast growth
inhibitor associated with perigraft seroma. J Vase Surg 1994; 20:202-208.

25. Friedman SG, Long KC, Scher LA. Axillofemoral bypass graft fracture. Ann Vase Surg 1996;
10:490-492.

26. Brophy CM, Quist WC, Kwolek C, LoGerfo FW. Disruption of proximal axillobifemoral
bypass graft anastomosis. J Vase Surg 1992; 15:218-220.

27. Piazza D, Ameli FM, von Schroeder HP, Lossing A. Nonanastomotic pseudoaneurysm of ex-
panded polytetrafluoroethylene axillofemoral bypass graft. J Vase Surg 1993; 17:777-779.

28. Okadome SMK, Onohara T, Yamamura S, et al. Recurrent multiple aneurysms in an
axillofemoral graft with coagulopathy. Acta Chir Scand 1990; 156:571-573.

29. Oudot J, Beaconfield P. Thrombosis of the aortic bifurcation treated by resection and
homograft replacement. Arch Surg 1953; 66:365-374.

30. Freeman NE, Leeds FH. Operations of large arteries: Application of recent advances. Calif
Med 1952; 77:229-233.

31. Vetto RM. The femoro-femoral shunt: An appraisal. Am J Surg 1966; 112:162.

32. Nolan KD, Benjamin ME, Murphy TJ, Pearce WH, McCarthy WJ, Yao JST, Flinn WR.
Femorofemoral bypass for aortofemoral graft limb occlusion: A ten year experience. J Vase
Surg 1994; 19:851-857.

33. Seeger JM, Kwab-Gatt CS, Lazarus HM, Albo D. Embolic and occlusive complications from
thrombosed femorofemoral grafts. J Cardiovasc Surg 1980; 21:547-558.

34. Harris JP, Flinn WR, Rudo ND, Bergan JJ, Yao JS. Assessment of donor limb hemodynamics
in femorofemoral bypass for claudication. Surgery 1981; 90:764-773.

35. Gupta SK, Veith FJ, Kram HB, Wengerter KA. Significance and management of inflow
gradients unexpectedly generated after femorofemoral, femoropopliteal, and femoroinfrapop-
liteal bypass grafting. J Vase Surg 1990; 12:278-283.

36. de Virgilio C, Cherry KJ Jr, Gloviczki P, Naessens J, Bower T, Hallett J, Pairolero P. Infected
lower extremity extra-anatomic bypass grafts: Management of a serious complication in high-
risk patients. Ann Vase Surg 1995; 9:459^466.

37. Shaw RS, Baue AE. Management of sepsis complicating arterial reconstructive procedures.
Surgery 1962; 53:75-76.

38. Sheiner NM, Sigman H. Stilman A. An unusual complication of obturator foramen arterial
bypass. J Caardiovasc Surg 1969; 10:324.

39. Pearce WH, Ricco JB, Yao JST. Modified technique of obturator bypass in failed or infected
grafts. Ann Surg 1983; 197:344-347.

40. Pearce WH, McCarthy WJ, Flinn WR, Yao JST. Obturator foramen bypass. In: Bergan JJ,
Yao JST, eds. Techniques in Arterial Surgery. Orlando, FL: Saunders, 1990:367-371.

41. Rudich M, Gutierrez IZ, Gage AA. Obturator foramen bypass in the management of infected
vascular prostheses. Am J Surg 1979; 137:657-660.

42. Blaisdell FW, DeMattei GA, Gauder PJ. Extraperitoneal thoracic aorta to femoral bypass
graft as replacement for an infected aortic bifurcation prosthesis. Am J Surg 1961; 102:583-585.

43. McCarthy WJ, Rubin JR, Flinn WR, Williams LR, Bergan JJ, Yao JS. Descending thoracic
aorta-to-femoral artery bypass. Arch Surg 1986; 121:681-688.

44. McCarthy WJ, Mesh CL, McMillan WD, Flinn WR, Pearce WH, Yao JST. Descending 1
thoracic aorta-to-femoral artery bypass: Ten years' experience with a durable procedure. J Vase <S
Surg 1993; 17:336-347. |

45. Passman MA, Farber MA, Criado E, Marston WA, Burnham SJ, Keagy BA. Descending ||
thoracic aorta to iliofemoral artery bypass grafting: A role for primary revascularization for **
aortoiliac occlusive disease. J Vase Surg 1999; 29:249-258. >9

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

16

Vascular Graft Infections: Epidemiology,
Microbiology, Pathogenesis, and Prevention

John M. Draus, Jr.

University of Louisville School of Medicine, Louisville, Kentucky, U.S.A.

Thomas M. Bergamini

University of Louisville School of Medicine and Surgical Care Associates,
Louisville, Kentucky, U.S.A.

The development of biomaterials to replace or bypass diseased arterial segments has
revolutionized the management of arterial disease. Although the search for the ideal
conduit continues, present-day grafts have proven to be durable alternatives with accept-
able patency for the treatment of both aneurysmal and occlusive disease in the peripheral
circulation. Their usage has allowed successful revascularization operations in numerous
patients who otherwise would have suffered loss of life or limb. Vascular surgical graft
infection is among the most feared complications that the vascular surgeon faces, often
resulting in prolonged hospitalization, multiple operations, and removal of the graft with
resulting organ failure, amputation, and death. This chapter focuses on the epidemiology
of vascular graft infection, reviews the most common pathogens, explores the theories of
pathogenesis, and suggests practical strategies for the prevention of this ominous compli-
cation.

I. EPIDEMIOLOGY 1

&

Infection of a vascular prosthesis is a relatively uncommon complication of vascular a

surgery. The reported incidence of infection involving synthetic vascular grafts ranges -c

from 1 to 6% with an average of 2.1% (Table 1). This variability can be partially explained <j

by differences in duration of postoperative follow-up, type of graft material and method of >9

construction, use of antibiotic prophylaxis, and virulence of the infecting pathogens. The 4j

actual incidence may be higher, since many graft infections do not become clinically 2

evident until years after implantation. The indication for intervention and the implanta- |

tion site have been shown to influence the incidence of vascular graft infection. Infection is @

305 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

306 DRAUS and BERGAMINI

Table 1 Incidence of Vascular Graft Infections

Number of

Study

Number of cases

infections (%)

Hoffert et al., 1965 (1)

201

12 (6.0)

Lindenauer et al., 1967 (2)

890

12(1.3)

Conn et al., 1970 (3)

435

22 (5.1)

Szilagyi et al., 1972 (4)

3,397

40 (1.2)

Goldstone and Moore, 1974 (5)

566

27 (4.8)

lameison et al., 1975 (6)

664

15 (2.3)

Liekweg and Greenfield, 1976 (7)

859

22 (2.6)

Yasher et al., 1978 (8)

590

32 (5.4)

Edwards et al., 1987 (9)

2,614

24 (0.92)

Fletcher et al., 1991 (10)

322

11(3.4)

Total

10,538

217 (2.1)

most common in grafts placed in the inguinal region or in superficial locations, possibly
associated with increased bacterial colonization and contamination with the patient's skin
flora at this site (10).

Vascular graft infections are commonly associated with operative events that lead to
bacterial contamination of the graft and patient characteristics that impair host defenses
(Table 2). Greater than 90% of vascular patients have one or more risk factors for the
development of a graft infection at the time of their operation. Graft infection is more likely
to occur after emergent procedures, such as the repair of a ruptured abdominal aortic
aneurysm. Breaks in sterile surgical technique and improper sterilization of grafts or
instruments are obvious sources of contamination. Prolonged preoperative hospitalization
increases the patient's risk of becoming colonized by more virulent bacterial species, which
are frequently resistant to conventional antibiotic therapy. Infected skin ulcers, gangrenous
toes, and other remote infection sites represent potential sources of graft sepsis. Operative
time greater than 4 h has been linked to intraoperative bacterial seeding. Concomitant
biliary, bowel, or urological procedures introduce additional sources of potential contam-
ination that may jeopardize the clean vascular bed into which the graft is placed. Infection
rates after reoperative vascular procedures for hematoma or graft thrombosis reflect the
frequency of arterial wall and wound colonization. Bile fistula, anastomosis breakdown,
and other postoperative complications often result in colonization of the vascular graft.
Patients with impaired immune function due to malnutrition, malignancy, or autoimmune
disease will have more difficulty fighting infection. Similarly, the administration of
medications such as steroids or immunosuppressive chemotherapy will alter the patient's •§

immunological competence. g

Infection of a vascular graft is an ominous complication whose outcome is often worse a

than the natural history of the vascular problem that led to implantation. Despite aggressive c

antibiotic administration and surgical treatment, overall mortality rates remain between 10 <j

and 50% and overall amputation rates between 15 and 60% (9-11). Infected aortic grafts >3

have the highest mortality rates (40-75%) due to hemorrhage, sepsis, and complications of 4j

multiple operative procedures (7,11,12). When systemic sepsis is the presenting symptom, 2

the prognosis is particularly grim. In contrast, femoropopliteal graft infections have a lower |

mortality (10-25%) (7,12). However, the amputation rate approaches 80% in some studies, @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

VASCULAR GRAFT INFECTIONS 307

Table 2 Risk Factors for Graft Infection

Contamination of the graft

Emergency surgery

Prolonged preoperative hospital stay

Remote infection

Faulty sterile technique

Extended operating time

Simultaneous gastrointestinal procedure

Reoperative vascular procedure

Postoperative superficial wound infection

Altered host defenses
Local factors

Biomaterial properties

Bacterial adherence

Slime production

Biofilm proliferation

Macrophage suppression

Cytokine production

Fibroblast inhibition

Collagen hyperplasia
Systemic factors

Malnutrition

Malignancy

Diabetes mellitus

Autoimmune disease

Chronic renal failure

Leukopenia

Chemotherapy

Corticosteroid administration

especially when infected grafts present with sepsis or anastomotic bleeding (13). Although
prosthetic vascular graft infection is uncommon, the threat of amputation, loss of organ
function, or death from this complication justifies the concern of the vascular surgeon and
the use of aggressive prophylaxis.

II. BACTERIOLOGY

Virtually any micro-organism is capable of infecting a synthetic graft. Initial studies ■g

considered graft infection to be an early complication of vascular surgery and regarded &

Staphylococcus aureus as the most prevalent pathogen (4,7,14,15). Refinements in arterial a

grafting techniques, prosthetic materials, and antibiotic usage have influenced the fre- -c

quency and nature of graft infections. Since the 1970s, coagulase-negative staphylococci

gram-negative bacteria, and polymicrobial infections have become significant (Table 3)

Surgeons have become cognizant of the possibility of microbiological sampling errors

especially in late-appearing infections, when low numbers of bacteria are present (16)

Gram-negative bacteria — such as Escherichia coli and Pseudomonas, Klebsiella, Proteus

and Enterohacter species — are particularly virulent. The incidence of anastomotic dehis- @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

308

DRAUS and BERGAMINI

Table 3 Bacteriology of Prosthetic Vascular Graft Infections: Data from 1305 Collected Cases

Number of Cases (%)

Pathogen

AEF

AI

AF

FD

FA

ICS

Total

397

86

460

251

55

56

1305

16(4)

7(8)

124 (27)

70 (28)

12 (22)

28 (50)

257 (19)

8(2)

22 (26)

120 (26)

28 (11)

14 (25)

1 1 (20)

203 (16)

75 (19)

3(3)

46 (10)

28 (11)

1(2)

—

153 (12)

71 (18)

18 (21)

55 (12)

18(7)

1(2)

—

163 (12)

12(3)

8(9)

28(6)

40 (16)

8(14)

—

96(7)

20(5)

4(5)

23(5)

5(2)

1(2)

6(10)

59(5)

32 (8)

8(9)

9(2)

18(7)

2(4)

—

69(5)

32 (8)

1(1)

14(3)

5(2)

—

—

52(4)

16(4)

1(1)

18(4)

18(7)

1(2)

—

54(4)

20(5)

8(9)

9(2)

5(2)

—

—

42(3)

12(3)

—

5(1)

3(1)

2(4)

—

22(2)

4(1)

—

5(1)

5(2)

—

—

14(1)

12(3)

3(3)

18(4)

15(2)

—

—

48 (4)

71 (18)

11 (13)

9(2)

5(2)

9(16)

1 1 (20)

116(9)

Staphylococcus aureus
Staphylococcus epidermidis
Streptococcus species
Escherichia coli
Pseudomonas species
Klebsiella species
Enterococcus species
Bacteroides species
Proteus species
Enter ohacter species
Candida species
Serratia species
Other species
No growth culture

Abbreviations: AEF = aortoenteric fistula or erosion; AI = aortoiliac, aortofemoral, or aortic tube
graft; AF = aortobifemoral or iliofemoral graft; FD = femoropopliteal, femorotibial, axillofem-
oral, or femorofemoral graft; TA = thoracic aorta graft; ICS = innominate, carotid, or subclavian
bypass graft or carotid patch following endarterectomy.

cence and artery rupture is high and is due to the organisms' ability to produce destructive
endotoxins (e.g., elastase and alkaline protease) that act to compromise the structural
integrity of the vessel wall. (17,18). Fungal infections due to Candida, Mycobacterium, and
Aspergillus species are rare and are typically seen in patients with previously established
fungal infections or severe immunosuppression.

Early-appearing graft infections occur within the first 4 months following vascular
bypass surgery and are associated with virulent pathogens. Patients present with classic
signs of graft sepsis, such as fever, leukocytosis, and bacteremia. The pathogens are easily
identified by cultures of blood or perigraft tissues. S. aureus continues to be the most
prevalent pathogen. Coagulase-positive strains produce hemolysis and toxins to leuko-
cytes that provoke an intense local and systemic host response and permit early
recognition of the infectious complications. Gram-negative bacteria are also implicated
in early graft infection. Pseudomonas aeruginosa infection is most commonly associated
with anastomotic bleeding. Graft healing complications, such as graft-enteric erosion or
fistula, typically involve infection with gram-negative enteric bacteria and can develop in
both the early and late postoperative periods.

Late-appearing infections are most frequently the result of graft colonization by Staphy-
lococcus epidermidis or other coagulase-negative staphylococci (19,20). These indolent in-
fections manifest themselves months to years after implantation and have replaced early
graft infections as the most common presentation in vascular patients (19,21). Coagulase-
negative staphylococci, organisms of low-virulence are normal inhabitants of the skin flora.
Bacteria are sequestered on prosthetic surfaces and survive within an adherent biofilm (22).
This bacteria-laden surface biofilm is composed of coalescing microcolonies enclosed in an

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i
2

I

■c

I

u

VASCULAR GRAFT INFECTIONS 309

extracellular nutrient glycocalyx produced by the organisms (23). Initially, the bacterial bio-
film creates a symbiotic infection that is localized to the prosthetic surface. Signs of graft
sepsis (e.g., fever, leukocytosis) are absent. If the bacterial biofilm is recognized by host
defenses, local inflammation of the perigraft tissue and adjacent artery ensues. Gram's stain
of perigraft fluid shows only white blood cells, and routine cultures of blood and perigraft
tissue or fluid are sterile. Bacterial biofilm infections present as a wound-healing compli-
cation, such as anastomotic aneurysm, perigraft abscess, graft-enteric fistula, or graft-
cutaneous sinus tract.

III. PATHOGENESIS

A. Bacterial Seeding of Biomaterial Surfaces

Any process that exposes the prosthetic conduit to microorganisms via direct, lymphatic,
or hematogenous routes can result in graft colonization and subsequent infection.
Biomaterial surfaces can become seeded with bacteria or fungi during the implantation
procedure, during postoperative wound-healing complications, or at any time thereafter
during occurrences of transient bacteremia.

Direct contamination of the vascular graft during arterial reconstruction results from
breaks in aseptic technique by the operative team. Improper sterilization of grafts and sur-
gical instruments are obvious sources of contamination. Contact with the skin and epi-
dermal appendages exposes the prosthetic graft to the patient's endogenous flora. It is not
surprising that an increased incidence of graft infections is observed when grafts are placed
in superficial locations (e.g., femoral, popliteal, and axillary anastomoses), where skin con-
tact is likely. Intestinal bag effluents and injury to or opening of the intestinal or urinary
tract are other sources of endogenous flora that can compromise graft placement. Dis-
eased artery walls (e.g., atherosclerotic plaque or aneurysmal thrombus) are a frequently
unrecognized source of bacteria, especially coagulase-negative staphylococci. Infected
lymph nodes and lymph channels are other important sources of direct contamination.
Transection of the lymphatics proximal to a septic focus bathes the implanted graft in a
bacterial inoculum that has been shown to contribute to the development of acute graft
infection in canine models (24). Femoral anastomoses are particularly vulnerable to this
type of contamination, especially when ischemic foot ulcers or gangrene is present in the
lower extremity. A concomitantly performed surgical procedure (e.g., gastrointestinal or
biliary tract) provides additional sources of potential contamination to the prosthetic graft.

The risk of prosthetic vascular graft contamination via a direct route continues into
the postoperative period. If the surgical wound does not promptly develop a fibrin seal
following the revascularization, the underlying graft is susceptible to direct extension from
initially trivial superficial wound problems. With persistent wound drainage, a septic focus ■g

can develop in ischemic or injured tissues and progress by deep extension to involve the &

vascular prosthesis (25). Hematomas and lymphoceles create an area of dead space a

around the graft that is favorable to bacterial growth, unfavorable to host defenses, and c

inhibitory to graft incorporation by perigraft tissues. Patients that undergo graft revision <j

for failed revascularization are repeatedly exposed to potential sources of graft infection. >9

These patients commonly harbor bacteria within scar tissue and lymphoceles and on the 4j

surfaces of previously implanted prosthetic grafts and suture materials. In explanted 2

prosthetic material, micro-organisms were cultured from 90% of grafts associated with |

anastomotic aneurysm and from 69% of thrombosed grafts (26). Furthermore, multiple @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

310 DRAUS and BERGAMINI

operations cause additional trauma to the tissues and increase the risk of poor wound
healing.

Bacterial seeding of the graft via a hematogenous route is an uncommon but
potentially important mechanism of graft infection. Experimentally, a single intravenous
infusion of 107 colony-forming units (CFUs) of S. aureus produces positive graft cultures
in 100% of animals up to 1 month following graft placement (27). During the periopera-
tive period, hematogenous seeding of the graft may occur via transient bacteremias from
sources such as intravenous catheters, an infected urinary tract, or remote tissue infection
(e.g., pneumonia or infected foot ulcers). Experimentally, parenteral antibiotic therapy
reduces the incidence of graft infection from bacteremia and is the basis of culture-
specific antibiotic therapy in patients with known infection at remote sites. The suscepti-
bility of graft to contamination from hematogenous sources is influenced by the integrity
of the cellular lining that covers biomaterial surfaces. The risk of contamination is high-
est immediately after implantation and gradually decreases as the luminal pseudointimal
lining develops and matures over time (28). The graft remains susceptible to hematoge-
nous seeding long after implantation; infection from bacteremia has been documented
beyond 1 year after implantation. Transient bacteremia in an immunocompromised pa-
tient may account for graft infections occurring years after the original revascularization.

B. Microcolony Formation and Bacterial Biofilms

Once micro-organisms seed the prosthetic surface, the ability of bacteria to produce a
vascular graft infection depends on adhesion, colonization, and biofilm formation. The
biomaterial infection persists because of the failure of the host immune system to kill
adherent bacteria and the surrounding inflammation and cellular damage that inhibits
tissue incorporation.

The physical properties and chemical composition of the vascular conduit significantly
influence the ability of bacteria to adhere to the biomaterial surface. Autogenous grafts,
such as the saphenous vein, have the lowest susceptibility to infection. In the absence of
immediate postoperative infection, these grafts exhibit long-term patency and represent
the best natural replacement for the original conduit. The two most common synthetic
materials used are external velour-knitted Dacron and expanded polytetrafiuoroethylene
(ePTFE). Dacron is composed of multiple interlocking strands. The porous nature of the
material favors bacterial sequestration. Bacterial strains have been shown to have greater
affinity for Dacron than for ePTFE (22,29,30). The extrusion process in the manufacturing
of ePTFE significantly decreases the porosity of the graft material. ePTFE is the material
of choice when prosthetic graft is used in patients at risk for bacteremia.

Bacterial adherence to the vascular prosthesis depends on the cell wall and the growth
characteristics of the bacteria species. Bacteria adhere by means of a mass of tangled fibers "g

of polysaccharides that extend from the bacterial cell surface and form a felt-like &

"glycocalyx" surrounding an individual cell or colony of cells (31). In a competitive natu- a

ral environment, the glycocalyx influences the selection and protection of a particular kind c

of bacteria from all others in the population. The prevalence of staphylococcal graft in- <j

fections can be partially explained by the increased adherence of gram-positive bacteria >3

to prosthetic surfaces. Under experimental conditions, Staphylococcus species have been 4j

shown to adhere to synthetic vascular graft materials in as high as 1000 times greater 2

number than do gram-negative bacteria (32). Similarly, S. aureus has been demonstrated |

to adhere to suture material in as high as 1000 times greater number than E. coli (33). The @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

VASCULAR GRAFT INFECTIONS 311

adherence of staphylococci is postulated to be enhanced by specific capsular adhesions
that potentiate micro-organism attachment and colonization. Antibodies to these specific
cell surface glycoproteins have been developed, and their application to graft surfaces can
inhibit adherence of adhesin-producing strains.

The pathogenesis of late-appearing graft infections is the result of prosthesis colo-
nization by low-virulence microorganisms. Some strains of coagulase-negative staph-
ylococci, such as S. epidermidis, adhere to the graft surface and survive in a bacteria-
laden biofilm. Unlike S. aureus and gram-negative bacteria, S. epidermidis grows only
within the biofilm, does not produce toxins or products capable of producing tissue
autolysis, and does not invade the surrounding tissues in the presence of host defenses (25).
The surface biofilm is a complex structure composed of coalescing microcolonies enclosed
in an extracellular nutrient glycocalyx that is produced by the organisms (23). The biofilm
plays an important role in promoting persistence of the infection by protecting it from host
defense mechanisms, entrapping nutrients from the environment, and resisting antimicro-
bial penetration (21,34,35).

C. Activation of the Immune System

Graft infection is a complex process involving activation of the host immune system by
coinflammatory stimuli produced by both the micro-organisms and the vascular biomate-
rials. The inflammatory process attempts to localize the infection but results in micro-
vascular disruption, cellular death, tissue necrosis, and neutrophils at the interface between
the surrounding tissues and the biomaterial surface. The virulence of the infecting pathogen
determines the extent of perigraft inflammation and tissue injury, but even indolent infec-
tions can produce tissue autolysis that results in vessel wall or anastomotic disruption and
hemorrhage. The clinical effects of graft infection manifest themselves as a broad spectrum
of signs including graft sepsis, localized perigraft abscess, anastomotic pseudoaneurysm,
graft cutaneous sinus tract, or graft-enteric erosion or fistula (aortoduodenal fistula).

Biomaterial infection is characterized by a lack of normal tissue incorporation. The
monocyte, or tissue macrophage, is crucial for directing this activity via specific growth
factors and enzymes (36). Using a mouse model of S. epidermidis vascular graft infection,
we have shown that vascular graft infection does suppress the expression of local macro-
phage la, which is associated with a local elevation of the proinflammatory cytokines and
a lack of normal healing (37,38). The overproduction of tumor necrosis factor alpha
(TNF-a) and interleukin 1 (IL-1) has been shown to cause tissue damage. TNF-a is a
major monocyte product that enhances the expression of neutrophil complement receptor.
The release of IL-1 has been shown to correlate with the extent of fibrous capsule forma-
tion. Moreover, TNF-a and IL-1 are known to stimulate collagenases, which would favor
nonhealing. These two cytokines may mediate the local tissue damage associated with ■g

chronic bacterial biofilm infections. Persistent growth of S. epidermidis in the biomaterial's &

surface biofilm results in a chronic inflammatory process and cellular damage to the sur- a

rounding tissues, inhibiting incorporation of the perigraft tissue into the biomaterial and c

resulting in the formation of a perigraft cavity and abscess. This process is mediated by the <j

host's immune response, producing autolysis of the periprosthetic tissues. The immune >9

response results in poor graft incorporation, with surrounding tissue inflammation and 4j

a perigraft cavity containing an exudate of many polymorphonuclear leukocytes (39-41). 2

Further progression of the inflammation in the perigraft tissues can result in a graft- |

cutaneous sinus tract or graft-enteric fistula, depending on the site of the graft. The chronic @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

312 DRAUS and BERGAMINI

inflammation mediated by the host's immune system can also result in an adjacent arteritis,
producing a decreased anastomotic tensile strength and subsequent pseudoaneurysm for-
mation.

In the mouse model of subcutaneous vascular graft infection, we have shown that
biomaterial infection is associated with fibroblast inhibition that is independent of
proinflammatory cytokines (42). The mediators of fibroblast inhibition using this mouse
model were bacterial cell products. Effective removal of this inhibitor of fibroblast
proliferation has been reported by plasmapheresis and also by removal of the implanted
biomaterial. In our rabbit model of vascular graft aortic infection, we have shown that the
perigraft fluid as well as serum from animals with infected grafts as compared to sterile
grafts inhibits fibroblast proliferation.

The prosthetic graft materials invoke a foreign-body reaction that produces an acidic,
ischemic microenvironment conductive to bacterial biofilm proliferation. Unlike autoge-
nous grafts, implanted synthetic biomaterials never develop vascular connections with the
surrounding tissue. This prevents host immune defenses and antibiotics from exerting their
maximal effect on infecting organisms. Furthermore, a complete cellular pseudointima
never develops on the luminal surfaces of biomaterials. Possible mechanisms include a
lack of fibroblast hyperplasia and collagen production. Failure of prosthetic grafts to
develop a protective endothelial lining renders them susceptible to late infection by
bacterial seeding.

IV. PREVENTION

The prevention of prosthetic graft infection is a critical concept in vascular surgery.
Prophylactic measures must be taken preoperatively, intraoperatively, and postoperatively
to help reduce the risk of this dreaded complication.

A. Preoperative Measures

A prolonged preoperative hospital stay should be avoided so as to reduce the patient's risk
of becoming colonized with nosocomial bacterial strains that are resistant to commonly
used antibiotics. Within 5 days of admission to a hospital, a patient's skin flora may al-
ready have become resistant to several antibiotics (43). Preoperative angiography should
be performed more than 1 week or less than 24 h prior to graft implantation. When trans-
femoral angiography is completed between 1 and 7 days before surgery, the incidence of
wound infection has been shown to increase (44).

Since the patient's own flora is the primary source of graft contamination, special
consideration should be given to skin preparation. Preoperative bathing with an antiseptic
soap may help reduce the number of bacterial colonies present (45,46). Shaving the skin ■g

increases the incidence of wound infections when compared the use of a depilatory cream &

(46). If shaving is preferred, it should be delayed until the patient's arrival in the operative a

suite. Prior to the incision, the skin should be thoroughly cleansed with povidone-iodine, c

chlorhexidine, or both (46,47). <j

The standard administration of prophylactic antibiotics has been shown to decrease >3

the occurrence of wound infections that potentially lead to vascular graft infections (26). 41

The choice of antibiotic should be determined by local bacterial prevalence and sensitiv- 2

ities, remembering that S. aureus, S. epidermidis, and gram-negative organisms are the |

most prevalent pathogens. The first dose of the antibiotic should be administered prior to @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

VASCULAR GRAFT INFECTIONS 313

the skin incision; subsequent doses should be given at regular intervals during the
operation to maintain tissue levels above the minimal bactericidal concentration for
expected pathogens. Alterations in the dosing regimen may be required during the
procedure based on the antibiotic's elimination and volume of distribution. It may become
necessary to increase the dosage or frequency of administration when the operative time
exceeds 4 h or if the patient experiences extreme changes in blood volume, fluid ad-
ministration, or renal perfusion. Cefazolin sodium, 1 g given intravenously 1 h before sur-
gery; 1 g given every 2-3 h during surgery; and then 1 g given every 8 h postoperatively for
three doses will achieve excellent tissue levels in the majority of patients undergoing pros-
thetic graft implantation. When methicillin-resistant S. aureus is cultured on the patient's
skin or is a known pathogen in hospitalized patients, a single dose of vancomycin hydro-
chloride, 1 g given intravenously may be added 1 h before surgery. For patients with aller-
gies to penicillin or the cephalosporins, parenteral vancomycin (1 g) plus gentamicin (1.5
mg/kg), administered 1 h before surgery and then repeated every 8-12 h for three addi-
tional doses is also an appropriate prophylactic antibiotic regimen. Patients undergoing
vascular graft implantation who have coexisting infections of the leg or another remote
site should be placed on culture-specific antibiotics. Some vascular centers continue pro-
phylactic antibiotics for 3-5 days in patients considered to be at high risk for transient
bacteremia or until all central intravenous catheters have been removed. However, there is
no solid evidence to substantiate the continued administration of prophylactic antibiotics
for more than two or three postoperative doses.

B. Intraoperative Measures

The importance of sound judgment and meticulous surgical technique is illustrated by the
observation that up to 60% of vascular graft infections are associated with a preventable
operative complication (5,6,48). When feasible, autologous vein should be used because it
is associated with a reduced risk of infection and with less serious consequences should
infection occur (49,50). Reconstructive procedures should avoid the groin region if
possible, because groin wounds are the most prone to infection. The use of iodine-
impregnated plastic drapes or antibiotic-soaked towels protects the graft from contact
with potentially contaminating sources outside the operative field. Gentle handling of
tissues, meticulous hemostasis to prevent hematoma formation, and closure of the incision
in multiple layers to eliminate dead space are important technical caveats for reducing the
risk of wound-healing problems and subsequent infections. The application of topical
antibiotics prior to wound closure may offer some benefit by increasing the antibiotic
concentration in the tissues surrounding the graft.

Simultaneous gastrointestinal procedures are best avoided during vascular graft
implantation to prevent contamination with enteric pathogens. If an inadvertent enter-
otomy should occur during celiotomy, the patient's incision should be closed and the
arterial reconstruction rescheduled for a second operation in a few days, with planned

implantation of an antibiotic-impregnated prosthesis. One possible exception to the c

admonition against simultaneous gastrointestinal procedures is cholecystectomy for <j

asymptomatic cholelithiasis, which can be safely performed in patients undergoing aortic >3

graft implantation. An 18% incidence of postoperative acute cholecystitis has been re- 4j

ported in patients with cholelithiasis after elective repair of an abdominal aortic aneurysm 2

(51). The risk of aortic graft infection after concomitant cholecystectomy remains low, but |

an increased complication rate has been reported in a large retrospective study of the @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

314 DRAUS and BERGAMINI

Mayo Clinic experience. The surgeon should proceed with gallbladder removal only after
the aortic graft has been implanted and the retroperitoneum completely closed.

C. Postoperative Measures

The patient's urinary catheter and intravenous lines should be removed as soon as possible
to eliminate potential sources of bacteremia. Early recognition and aggressive treatment of
postoperative wound infections are essential in minimizing the risk of extension to the
underlying graft. Patients with prosthetic vascular grafts should be fully educated as to the
potential risks of bacteremia following interventional procedures. Hematogenous seeding
is a continuing risk for as long as the graft remains in place. Antibiotic prophylaxis is
recommended prior to dental work, angiography, cystoscopy, and colonoscopy to prevent
bacterial colonization and subsequent late graft infection.

REFERENCES

1. Hoffert PW, Gensler S, Haimovici H. Infection complicating arterial grafts. Arch Surg 1965;
90:427-435.

2. Lindenauer SM, Fry WS, Schaub G, Wild D. The use of antibiotics in the prevention of
vascular graft infections. Surgery 1967; 62:487-492.

3. Conn JH, Hardy JD, Chavez CM, Fain WR. Infected arterial grafts: Experience in 22 cases
with emphasis on unusual bacteria and techniques. Ann Surg 1970; 171:704-712.

4. Szilagyi DE, Smith RE, Elliott JP, Vrandecic MP. Infection in arterial reconstruction with
synthetic grafts. Ann Surg 1972; 176:321-333.

5. Goldstone J, Moore WS. Infection in vascular prostheses: Clinical manifestations and surgical
management. Am J Surg 1974; 128:225-233.

6. Jamieson GC, DeWeese JA, Rob CG. Infected arterial grafts. Ann Surg 1975; 181:850-852.

7. Liekweg WG, Greenfield LJ. Vascular prosthetic infections: Collected experience and results of
treatment. Surgery 1977; 81:335-342.

8. Yashar JJ, Wevman AK, Burnard RJ, Yashar J. Survival and limb salvage in patients with
infected arterial prostheses. Am J Surg 1978; 135:499-504.

9. Edwards WH, Martin RS, Jenkins JM, Edwards WH, Mulherin JL. Primary graft infections. J
Vase Surg 1987; 6:235-239.

10. Fletcher JP, Dryden M, Sorrell FC. Infection of vascular prostheses. Aust NZ J Surg 1991;
61:432-435.

11. Seeger JM, Back MR, Albright JL, Carlton LM, Harward TRS, Kubulis MS, Flynn TC,
Huber FS. Influence of patient characteristics and treatment options on outcome of patients
with prosthetic aortic graft infection. Ann Vase Surg 1999; 13:413-420.

12. Bunt TJ. Synthetic vascular graft infections: I. Graft infections. Surgery 1983; 93:733-746.

13. O'Brien T, Collin J. Prosthetic vascular graft infection. Br J Surg 1992; 79:1262-1267.

14. Javid H, Julian OC, Dye WS, Hunter JA. Complications of abdominal aortic grafts. Arch Surg j>
1962; 85:650-662. ' |

15. Fry WS, Lindenauer SM. Infection complicating the use of plastic arterial implants. Arch Surg js
1967; 94:600-609. Jf

16. Kwaan JHM, Connolly JE. Successful management of prosthetic graft infection with con- ^
tinuous povidone-iodine irrigation. Arch Surg 1981; 116:716-720. «

17. Calligaro KD, Westcott CJ, Buckley RM, Savarese RP, DeLaurentis DA. Infrainguinal J
anastomotic arterial graft infections treated by selective graft preservation. Ann Surg 1992; q
216:74-79. |

18. Calligaro KD, Veith FJ, Gupta SK, Ascer E, Dietzek AM, Franco CD, Wengerter KR. A 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

VASCULAR GRAFT INFECTIONS 315

modified method of management of prosthetic graft infections involving an anastomosis to the
common femoral artery. J Vase Surg 1990; 11:485-492.

19. Bandyk DF, Berni GA, Thiele BL, Towne JB. Aortofemoral graft infection due to Staphy-
lococcus epidermidis. Arch Surg 1984; 119:102-108.

20. Santini C, Baiocchi P, Venditti M, Brandimarte C, Tarasi A, Rizzo L, Speziale F, Fiorani P,
Serra P. Aorto-femoral graft infections: A clinical and microbiological analysis. J Infect 1993;
27:17-26.

21. Dougherty SH, Simmons RL. Infections in bionic man: the pathobiology of infection in pros-
thetic devices. Curr Probl Surg 1982; 119:268-319.

22. Bergamini TM, Bandyk DF, Govostis D, Keabnick HW, Towne JB. Infection of vascular
prosthesis caused by bacterial biofilms. J Vase Surg 1988; 7:21-30.

23. Schmitt DD, Bandyk DF, Pequet AJ, Malangoni MA, Towne JB. Mucin production by
Staphylococcus epidermidis: A virulence factor promoting adherence to vascular grafts. Arch
Surg 1986; 121:89-95.

24. Rubin JR, Malone JM, Goldstone J. The role of the lymphatic system in acute arterial pros-
thetic graft infections. J Vase Surg 1985; 2:92-98.

25. Bergamini TM. Vascular prosthesis infection caused by bacterial biofilms. Semin Vase Surg
1990; 3:101-109.

26. Bandyk DF. Vascular graft infections: epidemiology, microbiology, pathogenesis and
prevention. In: Bernhard VM, Towne JB, eds. Complications in Vascular Surgery. St. Louis:
Quality Medical Publishing, 1991:223-234.

27. Roon AJ, Malone JM, Moore WS, Bean B, Campagna G. Bacteremic infectability: A function
of vascular graft material and design. J Surg Res 1977; 22:489^198.

28. Malone JM, Moore WS, Campagna G, Bean B. Bacteremic infectability of vascular grafts: The
influence of pseudointimal integrity and duration of graft function. Surgery 1975; 78:211-216.

29. Schmitt DD, Bandyk DF, Pequet AJ, Towne JB. Bacterial adherence to vascular prostheses: A
determinant of graft infectivity. J Vase Surg 1986; 3:732-740.

30. Sugarman B. In vitro adherence of bacteria to prosthetic vascular grafts. Infection 1982; 10:
9-12.

31. Costerton JW, Geesey GG, Cheng K-J. How bacteria stick. Sci Am 1978; 238:86-95.

32. Bergamini TM, Bandyk DF, Govostis D, Vetsch R, Towne JB. Identification of Staphylococcus
epidermidis vascular graft infections: A comparison of culture techniques. J Vase Surg 1989;
9:665-670.

33. Chih-Chang C, Williams DF. Effects of physical configuration and chemical structure of
materials on bacterial adhesion. Am J Surg 1984; 174:197-204.

34. McAuley CE, Steed DL, Webster MW. Bacteria presence in aortic thrombus at elective
aneurysm resection: Is it clinically significant? Am J Surg 1984; 147:322-324.

35. Costerton JW, Irvin RT, Cheng K-J. The bacterial glycocalyx in nature and disease. Annu Rev
Microbiol 1981; 35:299-334.

36. Ziegler-Heitbrock HW, Strobel M, Kieper D, Fingerle G, Schlunck T, Petersmann I, Ellwart J,
Blumenstein M, Haas JG. Differential expression of cytokines in human blood monocyte
subpopulations. Blood 1992; 79:503-511.

37. Henke PK, Bergamini TM, Garrison JR, Brittian KR, Peyton JR, Lam TM. Staphylococcus \
epidermidis graft infection is associated with locally suppressed MHC-II and elevated MAC-1 g
expression. Arch Surg 1997; 132:894-902. ,1

38. Henke PK, Bergamini TM, Brittian KR, Polk HC Jr. Prostaglandin E2 modulates monocyte 2
MCH-II(Ia) suppression in biomaterial infection. J Surg Res 1997; 69:372-378. ^

39. Steenfoos HH, Hunt TK, Scheuenstuhl H, Goodson WH. Selective effects of tumor necrosis >9
factor-alpha on wound healing in rats. Surgery 1989; 106:171-176. jjj

40. Knighton DR, Fiegel VD. The macrophage: Effector cell wound repair. Perspect Shock Res q
1989; 21:217-226. |

41. Barbul A. Immune aspects of wound repair. Clin Plast Surg 1990; 17:433-442. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

316 DRAUS and BERGAMINI

42. Henke PK, Bergamini TM, Watson AL, Brittian KR, Powell DW, Peyton JC. Bacterial
products primarily mediate fibroblast inhibition in biomaterial infection. J Surg Res 1998;
74:17-22.

43. Levy M, Schmitt DD, Edmistone CE. Sequential analysis of staphylococcal colonization of
body surfaces of patients undergoing vascular surgery. J Clin Microbiol 1990; 28:664-669.

44. Landreneau MD, Raju S. Infections after elective bypass surgery for lower limb ischemia: The
influence of pre-operative transcutaneous arteriography. Surgery 1981; 90:956-961.

45. Cruse PJ, Foord R. A five-year prospective study of 23,649 surgical wounds. Arch Surg 1973;
107:206-209.

46. Cruse PJ, Foord R. A ten-year prospective study of 62939 wounds. The epidemiology of wound
infection. Surg Clin North Am 1980; 60:27^10.

47. Berry AR, Watt B, Goldacre MJ, Thomson JWW, McNair TJ. A comparison of the use of
povidone-iodine and chlorhexidine in the prophylaxis of postoperative wound infection. J
Hosp Infect 1982; 3:55-63.

48. Lorentzen JE, Nielsen OM, Arendrup H, Kimose HH, Bille S, Anderson J, Jensen CH,
Jacobsen F, Roder OC. Vascular graft infection: An analysis of sixty-two graft infections in
2411 consecutively implanted synthetic vascular grafts. Surgery 1985; 98:81-86.

49. Johnson JA, Cogbill TH, Strutt PJ, Gundersen AL. Wound complications after infra-inguinal
bypass: Classification, predisposing factors, and management. Arch Surg 1988; 123:859-862.

50. Newington D, Houghton PWJ, Baird RN, Horrocks M. Groin wound infections after arterial
surgery. Br J Surg 1991; 78:617-619.

51. Calligaro KE, Veith FJ. Surgery of the infected aortic graft. In: Bergan JJ, Yao JST, eds. Aortic
Surgery. Philadelphia: Saunders, 1989:485-496.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

17

Aortic Graft Infections

G. Patrick Clagett

University of Texas Southwestern Medical Center, Dallas, Texas, U.S.A.

Aortic graft infections are among the most challenging and taxing problems encountered
by vascular surgeons. Patients with these infections are often elderly, frail, and severely ill
with multiple medical comorbidities; they are poorly equipped to tolerate the extensive,
complex operations usually required to treat the problem. Complete resection and excision
of all infected graft material and debridement of vascular structures are usually necessary
to eradicate infection. Immediate restoration of flow to critical vascular beds by alternate
anatomical routes or with in situ replacements that minimize the risk of recurrent infection
challenge the skill and ingenuity of the vascular surgeon. Despite a great deal of progress
in the treatment of aortic graft infections, morbidity and mortality remain higher than in
any other vascular condition (1-3).

I. PATHOGENESIS

Vascular grafts are foreign bodies that can be primarily infected by contamination at the
time of placement or secondarily infected after implantation by hematogenous, lymphatic,
or contiguous spread. The overall incidence of clinically overt graft infection varies ac-
cording to anatomical site. Aortic grafts confined to the abdominal or thoracic cavity
rarely become infected; the incidence ranges from 0.5 to 2% (2). The incidence is higher,

from 2 to 6%, when distal anastomotic sites are at the femoral level (4). ■§

Several features of the femoral area predispose to infectious complications. The groin |

is difficult to clean and incisions placed in the groin are prone to infection and healing a

problems. Groin incisions that extend obliquely across the inguinal crease tend to gape, c

and in obese patients they lie buried in moist skin folds. Furthermore, superficial inguinal <

lymph nodes are usually transected during exposure of the common femoral artery; if they >9

are not ligated, they will bathe a vascular graft in lymphatic fluid that may contain bac- J

teria. Potential sources of graft contamination in this circumstance include open, infected °

ischemic ulcers of a lower extremity, gangrenous toes, and wounds in any other area |

drained by the inguinal lymphatics, such as the perineum and perianal area. Another @

317

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

318 CLAGETT

factor implicating the groin wound in the etiology of vascular graft infections is transient
local ischemia during placement of the graft, which may render the wound more sus-
ceptible to infection.

The majority of vascular graft infections are initiated at the time of operation (2,3,5).
Although direct proof of this is difficult to obtain, the prevalence of Staphylococcus epi-
dermidis among offending organisms suggests that skin contamination with the patient's
own flora is an important mechanism (6,7). S. epidermidis can often be cultured from the
gloves of the surgical team, and grafts may be contaminated by this source during place-
ment (8). The presence of S. aureus and other nosocomially acquired bacteria is also com-
mon and points to other environmental sources of contamination. These include intestinal
flora when the gastrointestinal tract is entered or when operations such as cholecystectomy
are performed at the time of vascular reconstruction. Laminated thrombus lining the walls
of aneurysms has been implicated as a source of contamination and, when cultured, yields
bacteria in about 10% of specimens (9,10). S. epidermidis is the most common isolate.
Postoperative sources of aortic graft infection include wound complications, urinary tract
infections, and invasive line sepsis. Early and late hematogenous seeding of grafts can
occur during transient bacteremia associated with remote infections or dental procedures
(11).

Although bacteria cause most aortic graft infections, other, less common micro-
organisms such as fungi, mycoplasmas, and mycobacteria have been encountered. S.
epidermidis is the most common pathogen reported in modern series and outnumbers S.
aureus infections two to one. Gram-negative and polymicrobial infections are increasingly
being encountered but remain less prevalent than gram-positive infections. In many
instances, negative cultures are reported despite convincing local evidence of infection,
including nonincorporated graft material surrounded by grossly purulent fluid (12). These
cases are most likely caused by S. epidermidis or other low-virulence organisms that are
exposed to perioperative antibiotics at the time of sampling and require fastidious micro-
biological techniques for growth. Sonication of graft material, growth in tryptic soy broth,
and prolonged incubation for several days have been reported to increase the yield of
cultures positive for S. epidermidis (13).

Methicillin-resistant S. aureus (MRSA) has emerged as a serious and prevalent patho-
gen in some recent series tracking the epidemiology of vascular infections (14). These micro-
organisms can infect native arteries, destroy vascular tissue, and be difficult to eradicate
(15). MRSA appear to be particularly virulent, leading to poor outcomes. In one recent
series of 55 MRSA graft infections, 55% of patients died or underwent amputation (14).

The presence of a foreign body, such as an implanted device, increases the risk of
infection. Early investigations documented that it takes only 10~ S. aureus organisms to
cause an abscess at the site of a suture but 10 6 organisms to cause an infection in normal
skin. The vulnerability of foreign materials to infection involves physicochemical proper- ■§

ties of the material, impairment of host defenses, and special properties of the bacteria |

themselves that facilitate their growth in the presence of a biomaterial (16). The biological a

reaction to an implanted vascular graft comprises an acute inflammatory response in the c

early stages that progresses to formation of a fibrous capsule or tissue ingrowth. Neu- <

trophils rapidly become associated with any implanted biomaterial in vivo, become pre- >9

maturely activated by contact with the material, and rapidly lose the capacity to become J

activated in response to subsequent stimuli, such as the presence of bacteria. Neutrophils «

in contact with biomaterials rapidly lose their ability to produce superoxide and other re- |

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC GRAFT INFECTIONS 319

active oxygen species and become relatively impotent in their microbicidal activity (17,18).
Thus the biomaterial acts as a massive "decoy" that prevents the neutrophils from re-
sponding normally to bacteria in the microenvironment. In addition, neutrophil products
released in these circumstances may promote dysfunction of new neutrophils entering the
microenvironment (19).

Vascular graft materials may vary in their susceptibility to infection by different mi-
cro-organisms. Highly textured or rough-surfaced biomaterials, such as textiles made of
Dacron (woven or knitted), are more prone to bacterial adherence than smooth-surfaced
biomaterials, such as expanded polytetrafluoroefhylene (ePTFE) or polyurethane (20). In
vivo, the adherence of platelets, plasma proteins, and other blood constituents and varying
conditions of shear may dramatically alter the responses of different biomaterials to micro-
organisms, and all biomaterials remain susceptible to infection (21,22).

The principal organism responsible for infections of all implanted medical devices,
including vascular grafts, is S. epidermidis. This organism is a ubiquitous skin commensal
with of relatively slow growth and low virulence. It causes chronic infections with local
manifestations and little or no systemic toxicity. Pivotal in the pathogenesis of Staph-
ylococcus epidermidis infection is the production of multilayered biofilms composed of
exopolysaccharides, usually referred to as "slime." The elaboration of biofilms takes place
following the adherence of S. epidermidis to biomaterials and usually occurs when orga-
nisms adhere to one another in microcolonies (23). Adherence of organisms to both poly-
mer surfaces and each other (cell-cell adhesion) is mediated by capsular polysaccharide
adhesins (23,24). Mutant bacteria that do not produce adhesins lack cell-cell adhesions
and do not produce biofilms (25). Once elaborated, biofilms form a protective shield that
allows continued bacterial growth in relatively hostile environments. Bacterial nutrients
and metabolic wastes freely traverse the polysaccharide biofilm, but antibiotics do not.
Biofilms also alter inflammatory changes, impair host defenses, and promote tenacious
adherence of microbial colonies to the biomaterial (26). S. epidermidis infections tend to be
persistent and refractory to antibiotics; therefore the implant must be removed to clear the
infection.

Once established, bacterial infection spreads throughout a vascular graft and even-
tually involves anastomotic sites. The eventual destruction of vascular tissue leads to the
formation of an anastomotic false aneurysm. The first manifestation of a vascular graft
infection is often an anastomotic false aneurysm or its most frequent complication, graft
thrombosis. When the false aneurysm involves the aortic anastomosis, rupture into the
duodenum may occur and produce an aortoduodenal fistula with catastrophic hemor-
rhage. Although all micro-organisms producing vascular graft infections are associated
with false aneurysms, they vary in their propensity to destroy vascular tissue. Gram-nega-
tive organisms — such as Pseudomonas aeruginosa, Proteus species, and Escherichia coli —
are particularly notorious for their ability to digest vascular tissue (27). These organisms ■a

elaborate elastase and alkaline protease, which break down elastin, collagen, fibronectin, |

and fibrin. In addition to causing vascular disruption and the formation of false aneu- a

rysms, many bacteria produce substances that are highly thrombogenic and can induce c

thrombosis that may be the first manifestation of a vascular graft infection. <

Aortic stent grafts would appear to be uniquely predisposed to infection. Tissue in- >9

growth from surrounding tissues is generally absent (28), and lack of tissue incorporation J

into prosthetic interstices is widely acknowledged to be a permissive condition for prosthe- «

tic infections. In addition, stent grafts are usually surrounded by luminal thrombus, which |

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

320 CLAGETT

may harbor micro-organisms. Despite these theoretical considerations, acute aortic stent
graft infections have been reported infrequently. Environmental sources of infection from
gloved hands and the operating field are reduced because most commercial stent grafts are
jacketed in sterile packages until deployed. In cases of acute stent graft infection, bac-
teremia was clearly documented (29). Infection may emerge as a significant problem on
longer follow-up. In one series with a mean follow-up of 49 months, the incidence of in-
fection of stent grafts requiring removal was an alarming 3.3% (30).

II. CLINICAL PRESENTATION

The clinical presentation of aortic graft infections can be protean and subtle, making the
diagnosis difficult. The tempo and severity of the clinical manifestations often depend on
the micro-organism. A patient whose infection is caused by a virulent organism — such as
S. aureus, P. aeruginosa, and E. coli — presents with systemic signs of sepsis. As an exam-
ple, a patient with a vascular graft who has persistent fever, chills, and an elevated white
blood cell count with a left shift should be suspected of having a vascular graft infection.
Virulent micro-organisms also tend to cause earlier manifestations of infection, with the
interval between implantation of the graft and diagnosis of infection being months. Very
early graft infections, diagnosed within weeks of implantation, are often associated with
wound complications that involve vascular grafts by contiguous spread.

In contrast, patients with graft infection caused by a low-virulence organism, such as
S. epidermidis, present later, often years after placement (7). Systemic signs and symptoms
are usually mild or absent. These patients most often present with local manifestations,
such as a chronic groin sinus that discharges small amounts of purulent material, a chronic
wound infection with exposed graft, femoral anastomotic false aneurysm, or aortofemoral
bypass limb thrombosis. They may have low-grade fever and mild constitutional symp-
toms, but overt signs of sepsis are absent. The white blood cell count is usually normal or
only mildly elevated, but the erythrocyte sedimentation rate is often elevated. A patient
presenting with a femoral anastomotic false aneurysm or limb thrombosis who has an
elevated erythrocyte sedimentation rate should be suspected of having a graft infection.

Patients presenting with massive gastrointestinal hemorrhage from an aortoduodenal
or aortoenteric fistula have frequently had lesser episodes of bleeding hours to days before
the major episode. These are often referred to as "herald" or "sentinel" episodes of bleed-
ing and offer a window of opportunity for the diagnosis and management prior to the
onset of exsanguinating hemorrhage. Any patient with an aortic graft who has an episode
of upper or lower gastrointestinal bleeding should be suspected of having an underlying
aortoenteric fistula, and an expeditious workup is important. Chronic gastrointestinal
bleeding can also occur in patients with an aortoenteric fistula but is more often associated
with an enteric erosion. This condition, often referred to as a "graft-enteric erosion," |

differs from aortoenteric fistula in that the body or limb of the aortic graft erodes into a

bowel and the aortic suture line is not involved. This produces chronic bleeding from the H,

eroded bowel mucosa, analogous to bleeding from an ulcer, and patients may present with =j

chronic anemia. The diagnosis should be suspected in a patient with an aortic graft who «

has anemia, stool positive for occult blood, and fever. g

Hydroureteronephrosis may also be the first manifestation of an aortic graft infection. «

This can develop if the ureter becomes obstructed as a result of perigraft inflammation and "g

may be bilateral or unilateral, depending on the extent of infection. g

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC GRAFT INFECTIONS 321

III. DIAGNOSIS

Because the manifestations of aortic graft infections are so varied and subtle and the
consequences of a missed diagnosis may be lethal, imaging tests are important (31). The
types of imaging and other diagnostic tests used are based on the clinical presentation.
Computed tomography (CT) has been the mainstay of diagnostic imaging for a suspected
aortic graft infection. CT findings suggestive of infection include ectopic gas, peripros-
thetic fluid, loss of tissue planes, perigraft inflammatory changes, thickening of adjacent
bowel, hydroureteronephrosis, and anastomotic false aneurysm (32). These findings are
most specific and useful for late infections. During the immediate perioperative period
following implantation, perigraft fluid, air, and inflammatory changes may persist for 2-3
months. After 3 months, postoperative hematoma and gas should resolve and tissue planes
return to normal (33).

Magnetic resonance imaging (MRI) has provided an alternative to CT for cross-
sectional imaging. In addition to demonstrating the same features seen on CT (peripros-
thetic air, fluid, and structural abnormalities), MRI is particularly helpful in assessing
periprosthetic inflammatory changes. These changes are high intensity signals on T2-
weighted images in the tissues surrounding the prosthesis and accurately portray tissue
edema (34). Such images can be particularly helpful in assessing the extent of infection,
which may determine the operative approach. For example, in a patient with an infection
localized to a single distal limb of an aortobifemoral bypass, removal of the entire pros-
thesis may not be required for adequate treatment of the infection.

Radionuclide scanning has also been used in the diagnosis of vascular prosthetic
infections. Scintigraphy with the use of autologous white blood cells labeled with indium
111 ( 'i) is the most common technique currently used, although the use of white cells
labeled with gallium 67, technetium, and other isotopes has been reported (35,36). In ad-
dition, scintigraphy based on labeled human immunoglobulin G has been used and may
be more sensitive than scintigraphy with white cells (37). A problem with all scintigraphic
methods in diagnosing vascular graft infections is the lack of specificity caused by uptake
in other organs or tissues that may be contiguous. In addition, faint or no uptake in the
presence of limited or low-virulence infection can result in false-negative results. Scintig-
raphy is most helpful when occult prosthetic infection is suspected. An example would be a
patient with an aortic graft presenting with a fever of unknown origin or a complex of
other nonspecific symptoms in whom white blood cell scintigraphy identifies the graft as
the source.

Arteriography is of limited usefulness in the diagnosis of vascular graft infection but it
may, on occasion, demonstrate an aortic false aneurysm or even leakage of contrast into
the bowel lumen, which is pathognomonic for an aortoenteric fistula. Arteriography is
helpful in planning reconstruction after removal of the prosthesis and is most useful in late
infection, when the vascular anatomy may have been altered by progressive occlusive di- |

sease. CT angiography may replace conventional angiography and provide additional a

information obtained by conventional cross-sectional CT imaging. H,

In patients presenting with gastrointestinal bleeding and suspected aortoenteric fis- =j

tula, complete upper gastrointestinal endoscopy with visualization of the third and fourth j

portions of the duodenum, the most common sites of fistula, is necessary. Even if this g

study is incomplete, with inability to visualize the distal duodenum or the finding of gas- «

trointestinal lesions such as chronic peptic ulcer that are not actively bleeding, an aorto- "g

enteric fistula may still be present. Continued unexplained bleeding mandates operative g

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

322 CLAGETT

exploration to rule out aortoenteric fistula. At the time of operation, the duodenum,
proximal jejunum, and any other bowel in contact with an aortic graft must be dissected
free to make or exclude this diagnosis.

IV. PREVENTION OF AORTIC GRAFT INFECTIONS

The benefit of short-term antibiotic prophylaxis in preventing wound infections after
vascular surgery has been demonstrated in randomized trials (38-40). Most often, a first-
generation cephalosporin is administered intravenously shortly before operation, during
operation if blood loss is extensive or the operation is prolonged, and 2 h after operation.
Some evidence suggests that a more prolonged course for up to 4-5 days after operation or
until all invasive lines are removed may provide additional protection (41). In circum-
stances where patients have infected lower extremity ischemic lesions, culture-specific anti-
biotics should be administered perioperatively. Also, the use of more specific prophylactic
antibiotic therapy should be considered in hospital settings where certain organisms are
prevalent, especially when exposure is increased by prolonged preoperative hospitalization.

Attention to intraoperative factors is also important in preventing aortic graft in-
fections. Reoperations and emergency operations are especially prone to wound infections
and present additional risks. Meticulous attention to hemostasis and avoidance of wound
hematomas and seromas that can become secondarily infected are important surgical goals
that are often difficult to achieve in patients anticoagulated during the operation who are
also being treated with antiplatelet agents. If possible, these agents should be discontinued
one week prior to operation. Ligation and control of femoral lymphatics are also impor-
tant technical features in preventing aortic graft infections. Electrocautery of lymphatic
tissue leads to coagulation necrosis of lymphatic vessels but does not prevent extravasation
of lymph fluid. Fibrin glue applied to groin wounds has been shown to decrease lymph
drainage and groin would complications (42).

Patients undergoing aortic operations are prone to intraoperative hypothermia; this
condition has been shown to impair neutrophil function and increase the incidence of
postoperative wound infection (43). Maintenance of normal body temperature should be
the goal during major vascular operations. Additional procedures on the gastrointestinal
or biliary tract that may result in intraoperative contamination of an aortic graft should be
avoided unless the additional procedure is deemed necessary to avoid life-threatening
postoperative complications. Hematogenous seeding of a vascular graft is a continuing
risk for as long as the graft is in place. Dental work, procedures on the gastrointestinal and
genitourinary tracts, and angiographic procedures should be carried out under the pro-
tection of prophylactic antibiotics.

■6

V. TREATMENT 1

The primary goals of treatment are to save life and limb, and these are best accomplished S

by eradicating infection and maintaining adequate circulation to portions of the body =j

perfused by the infected aortic graft. Secondary goals include minimizing morbidity, re- g

storating of normal function, and maintaining of long-term function without the need for g

reintervention and risk of amputation. «

These goals are best achieved by the removal of all infected graft material and vascular "g

tissues combined with appropriate arterial reconstruction. The currently favored methods g

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC GRAFT INFECTIONS

323

a

o

<

TO

a3 C3

J_j

o

<L>

Q

o

IT)

O

r-i

c

a

S

o

a

:/:

a

c3

!-

d

■Si

C3

t3

5 o

« S

o id

04

cd BO

■ a

aj d

g "S

3 Oh

z

o

■ — I u">

o
o

'ON

-r

>,

,

o

©
©

cd

5Jj

1

|

mo cn

1 |

!-

J-l

o

r-

rn o

1>

G

m

lO

■•c

o

c\

od oo "^

Sh

£

o

o

rn

-—

IN -H

IN — i

m —<

r- —i

-f

IN —

<g o 8 2 a -c a
5 ft cs mi 2 "K 2
&-i "E.,2 S p B

c^c^Vh^^OUhO

a. S B -S 2 U 5 a.
c a a

c

Ph

c
c

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

324 CLAGETT

of arterial reconstruction for aortic graft infection include extra-anatomic bypass (44-57)
and in situ replacement using autogenous superficial femoral popliteal veins (58-61),
arterial allografts (62-70), or vascular prostheses often treated or soaked in antibiotic so-
lutions (71-75). Pooled outcome data from contemporary series reported since 1985 are
presented in Table 1. Direct comparisons in attempting to judge the relative success of
these approaches is difficult from these data because of the heterogeneity of patients with
varying severity of illness and comorbidities among reported series. All of these ap-
proaches are valid and have utility depending upon patient-specific characteristics and
circumstances. It is a mistake to think that a single surgical approach is applicable to all
patients with this condition. These complicated patients with varying levels of illness
severity require individualized attention.

VI. EXTRA-ANATOMIC BYPASS

Extra-anatomic bypass usually involving axillofemoral bypass is an excellent choice for
infected aortoiliac reconstructions in which femoral sites are free of sepsis and the arterial
runoff is good. It is also possibly less of a physiological insult in comparison to other
procedures, particularly when the operations can be staged with extra-anatomic bypass
preceding removal of the infected aortic prosthesis by a period of days (45). This approach
has the advantage of preserving lower extremity blood flow during removal of the aortic
prosthesis, thus minimizing lower extremity ischemia time.

Unfortunately, extra-anatomic bypasses have limited durability in patients with mul-
tilevel occlusive disease and poor runoff. Most patients with infected aortic grafts have
aortobifemoral bypasses, and extra-anatomic bypass in such patients usually requires bi-
lateral axillofemoral procedures with distal anastomoses to diseased and small, deep fem-
oral or popliteal arteries. These are disadvantaged reconstructions with poor long-term
patency despite the use of antithrombotic agents. They are prone to sudden thrombotic
occlusion without warning and amputation rates are high, even with thrombectomy and
multiple revisions. In one large series, one-third of patients required major amputation
during long-term follow-up (46). In addition, reinfection of extra-anatomic bypass grafts
occurs in 10-20% of patients, and this condition is often lethal. A final problem with
extra-anatomic bypass is continuing infection at the site of aortic closure or the aortic
"stump." Although an infrequent occurrence (less than 10%), aortic stump blowout is
almost always fatal.

VII. IN SITU REPLACEMENT WITH SUPERFICIAL

FEMOROPOPLITEAL VEINS

|

Dissatisfaction with the long-term patency of extra-anatomic bypass led to the develop- g

ment of in situ autogenous vein reconstruction (58-60). Early experiences were with greater »

saphenous veins, but this procedure rapidly evolved to the use of superficial femo- c

ropopliteal veins because of their large caliber and superior patency (58). This procedure <

has been referred to as creation of a "neoaortoiliac" system (NAIS) procedure. This recon- &

struction is most applicable in patients with extensive occlusive disease and poor runoff, a J

circumstance where an autogenous venous reconstruction would have better patency than «

a prosthetic graft bypass. The situation is analogous to the superior patency of vein grafts |

in comparison to prosthetic conduits in the performance of femoropopliteal and distal @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC GRAFT INFECTIONS 325

bypasses. This advantage has been realized in excellent 5-year cumulative patency rates for
NAIS reconstructions of 85% for primary patency and 100% for secondary/assisted pa-
tency (60). Long-term amputation rates have been reported to be correspondingly low.

Although the technical details of the NAIS procedure have been previously published
(60,76,77), some features merit emphasis. Duplex ultrasound imaging of the lower ex-
tremity's deep and superficial veins is essential in preoperative planning. Duplex vein map-
ping allows preoperative determination of diameter and length of available superficial
femoropopliteal veins. Important findings that may alter the operative plan include deep
venous thrombosis involving the superficial femoropopliteal veins, recanalization changes,
and congenital absence or unusually small superficial femoropopliteal veins. In our expe-
rience, these findings are fortunately limited to one side. In situations where the superficial
femoropopliteal vein is incomplete, absent, or unusually small (less than 4-5 mm in diam-
eter), a dominant deep femoral vein is often present. This large vein courses posteriorly
through the thigh to connect with the popliteal vein and can be used as a venous autograft
for the NAIS procedure. Duplex vein mapping of the greater saphenous vein will also give
information that may be useful if concomitant femoropopliteal/distal or visceral/renal re-
construction is required.

In order to minimize body exposure and lower extremity ischemia, the operation is
sequenced as follows: (a) dissect superficial femoropopliteal veins and leave in situ until
needed; (b) isolate femoral vessels; (c) enter abdomen and obtain aortic control; (d) re-
move infected aortic graft, and (e) perform reconstruction with superficial femoropopliteal
veins. In the initial phases of the operation, it is useful to use two surgical teams to dissect
the veins and isolate the femoral vessels.

The sartorius muscle is reflected medially and posteriorly and the subsartorial canal is
opened in the midthigh. The saphenous nerve is vulnerable during this dissection, and
excessive traction or unplanned division of this nerve can result in annoying postoperative
medial leg neuralgia. Care must be taken in mobilizing the branches of the superficial
femoral artery, especially around the adductor hiatus. These branches may represent
important collaterals to distal arterial beds. Interruption of these when the superficial
femoral artery is occluded may result in critical and unanticipated distal ischemia after
completion of the proximal reconstruction.

The superficial femoropopliteal vein has multiple large and small branches. The larger
ones are doubly ligated with 2-0 and 3-0 silk. Very large side branches are suture-ligated.
The importance of secure branch ligature cannot be overemphasized. A "popped" tie can
result in exsanguinating hemorrhage. The walls of the superficial femoropopliteal vein
tend to be thin and tenuous near the origins of side branches. Torn branches can be
frustrating to repair and require tedious closure with 7-0 polypropylene sutures. It is best
to avoid this with careful and patient dissection of all branches. In addition, placement of
branch ligatures should be close and contiguous to the vein wall. This is different from ■§

dissection of the greater saphenous vein, where emphasis is placed on securing ligatures |

slightly away from the vein wall so as not to constrict the vein. The superficial a

femoropopliteal vein is large and can easily tolerate close ligatures, which are helpful in c

preventing tears from the thin wall near the branch origins. <

Dissection is carried proximally to where the superficial femoral vein joins the deep >9

femoral vein to form the common femoral vein. The deep femoral vein is readily identified J

as a large-caliber vessel penetrating deep through the fascia on the floor of Hunter's canal. «

The dissection is then carried distally with division of the adductor tendons to open the |

adductor hiatus. The distal dissection is carried to the level of the knee and, if necessary, @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

326 CLAGETT

the dissection can be continued below the knee for an additional few centimeters of
autograft.

The next portion of the operation involves isolation of the femoral vessels and distal
limbs of the aortobifemoral bypass graft. In most patients, this can be accomplished by
extending the vein harvest incision along the lateral border of the sartorius muscle to its
attachment site at the anterosuperior iliac spine. In thin patients, the entire common fe-
moral artery can be exposed by reflecting the sartorius medially. On occasion, dissection
medial to the sartorius is helpful to complete this exposure. This approach obviates the
need to dissect through the old femoral incisions.

The abdomen is then entered through the old incision or through a left flank
retroperitoneal approach. The latter is particularly helpful in avoiding tedious adhesions
and facilitates obtaining aortic control. This approach also facilitates suprarenal or supra-
celiac control if the proximal anastomosis is close to the renal arteries. The vein grafts are
then prepared and the lengths required are determined by measuring the distance from the
proximal anastomosis to the femoral levels on both sides. The proximal superficial femoral
vein is divided flush with the deep femoral vein and oversewn with 5-0 polypropylene
continuous sutures. This allows unimpeded flow from the deep femoral to the common
femoral vein and minimizes the possibility of thrombus forming in a residual venous cul-
de-sac of the superficial femoral vein. The vein graft usually has three to four large valves
that are easily identified and ablated using a valvulotome or directly excised by tempo-
rarily everting the vein graft. The nonreversed configuration will allow placement of the
larger end of the vein at the proximal aortic anastomosis.

Following systemic heparinization, cross clamps are placed on the proximal aorta and
the distal limbs of the aortic graft. The body of the graft is removed and the femoral limbs
are left in place while the proximal anastomosis to the superficial femoral vein is per-
formed. Leaving the femoral limbs in place during this period cuts down on blood loss that
typically occurs when the femoral limbs are extracted from their tunnels.

Several configurations using superficial femoral popliteal vein grafts are possible and
allow flexibility in performing the NAIS reconstruction (Fig. 1). The proximal end of the
superficial femoral vein is frequently (1.5 cm in diameter and can easily be anastomosed to
a normal aorta. Standard, continuous polypropylene (4-0) suture technique is used, taking
care to make slightly more advancement on the aorta than the venous autograft because of
the greater circumference of the aorta (Fig. 2). Larger aortas and greater size mismatches
are often encountered and require different anastomotic techniques. Plication of the distal
aorta may be performed to reduce the diameter of the aorta at the anastomosis (Fig. 3).
Both superficial femoral popliteal vein grafts can be joined together in a "pantaloon" con-
figuration (Fig. 4). This technique essentially doubles the circumference of the vein graft's
proximal anastomosis. Another technique to increase circumference is shown in Figure 5.

Following completion of the proximal anastomosis, the old graft limbs are carefully ■§

removed. Most often, superficial femoral popliteal vein grafts are placed in the old tunnels s

because it is difficult to fashion new tunnels through the scarred retroperitoneum. The »

superficial femoral popliteal vein grafts are often larger than the tunnels; when this occurs, c

careful proximal and distal finger dilation can be useful to prevent luminal compromise of <

the vein graft. Care must be taken when pulling vein grafts through the tunnels, as side >9

branch ligatures may become dislodged. To avoid this problem, vein grafts are passed non- J

distended. Q

Following operation, antibiotic coverage is continued for 5-7 days. Antibiotics are |

modified as culture results isolate organisms sensitive to specific antibiotics. In patients @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC GRAFT INFECTIONS

327

r

Figure 1 A. An aortofemoral bypass is fashioned on the left side with a superficial femoropopliteal
vein graft. Another vein graft is used to cross over to the right femoral region. B. Instead of a
crossover femoral bypass, the right limb is connected to the aorta-left femoral bypass using an end-
to-side anastomosis. In both A and B, a shorter length of superficial femoropopliteal vein graft is
required for the right side. C. In this example, infection was limited to the femoral region and the
distal aortofemoral limb was replaced with a superficial femoropopliteal vein graft. Care must be
taken to prevent cross contamination and the vein graft routed lateral to the infected area or via a
transobturator route. D. In this circumstance, an aortic-left common iliac bypass is fashioned from
one-third to one-half of a superficial femoropopliteal vein graft and the right limb is fashioned from
the remainder of the vein graft. This technique permits the harvesting of only a single femoro-
popliteal vein graft. E. In some circumstances, it is easier to approach the paraceliac aorta via a
retroperitoneal approach and to use this area for the proximal anastomosis, as shown in this figure.
Although the distal anastomosis is to the common iliac artery in this case, an additional length su-
perficial femoropopliteal vein graft can be harvested, if necessary, to reach the femoral area.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

328

CLAGETT

Figure 2 When the aorta is of normal size, the proximal end of the superficial femoropopliteal vein
graft is anastomosed end to end.

I
•5

Figure 3 Simple plication of a portion of the aorta can be performed to allow comfortable end-to-
end anastomosis when the aorta is either larger than normal or the vein graft is of comparatively
small size.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC GRAFT INFECTIONS

329

Figure 4 In this example, two superficial femoropopliteal vein grafts are joined together ("pan-
taloon" technique) and sewn end to end to the aorta. This effectively doubles the circumstance of the
vein graft and is used to accommodate very large aortas.

I
•5

Figure 5 In this case, a wedge-shaped portion of vein is incorporated into the proximal end to
increase the circumstance of the superficial femoropopliteal vein graft.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

330 CLAGETT

who are severely immunocompromised, prolonged antibiotic therapy for 4-6 weeks may
sometimes be necessary. Intermittent pneumatic compression plus low-dose subcutaneous
heparin (5000 U q 8-12 h) are used for prophylaxis of venous thromboembolism. Most
patients develop venous thrombosis in the residual popliteal vein segment or "stump."
Aggressive prophylaxis may prevent propagation into calf veins. Full anticoagulation for
this limited venous thrombosis is unnecessary, since — because of the absence of the super-
ficial femoral popliteal vein — proximal extension and pulmonary embolism are unlikely.
Patients are seen every 3 months on an outpatient basis for the first year following op-
eration. Noninvasive vascular testing includes ankle/brachial pressure indices and com-
plete graft duplex examination. Surveillance is directed at detecting vein graft and anas-
tomotic stenoses as well as progression of distal disease.

The principal disadvantage of the NAIS reconstruction is that it is a technically de-
manding, a long procedure. The mean operative time is approximately 8 h. The lower
extremity ischemic time is longer than with other approaches but can be minimized by
using a two-team approach and carefully sequencing the procedure to shorten aortic cross-
clamp time.

Acute venous hypertension following harvest of the superficial femoral popliteal vein
can contribute to the development of lower extremity compartment syndromes. Leg fas-
ciotomy is required in approximately 25% of patients. Preexisting lower extremity is-
chemia, prolonged aortic cross-clamp times, and absence of the ipsilateral saphenous vein
are risk factors for the development of a compartment syndrome. Prophylactic four-
compartment fasciotomy should be considered when these risk factors are present.

Long-term lower extremity venous morbidity is also a potential drawback to harvest-
ing the superficial femoral popliteal veins. However, venous morbidity has been surpris-
ingly infrequent and mild (59,60). Approximately 30% of patients will have transient
lower extremity swelling that requires compression stockings. This usually resolves within
a period of weeks to months after operation; then compression stockings are no longer
necessary. The benign course following removal of the superficial femoral popliteal vein is
due to several compensating mechanisms (78). First, the junction of the deep and common
femoral veins is carefully preserved after disconnecting the proximal superficial femoral
vein, thus allowing unimpeded drainage via the deep venous system. Second, there are
other anatomical collateral connections between the remaining distal popliteal vein and
the deep venous system; many of these collaterals enlarge to accommodate the increase in
volume flow following removal of the superficial femoral popliteal vein. Finally, the valves
in the tibial veins and collateral circuits remain functional, such that distal venous reflux
does not occur.

A final concern is that the placement of superfemoral popliteal veins in an infected
field might lead to reinfection and disruption. This has been rare and experience with this
approach has documented that these vein grafts resist gram-positive, gram-negative, and ■§

fungal infections. Long-term aneurysmal degeneration has been studied up to 10 years |

after placement of these vein grafts, and the incidence of this problem has been <1%. a

VIII. IN SITU REPLACEMENT WITH ALLOGRAFT AND
ANTIBIOTIC-TREATED PROSTHETIC GRAFTS

In situ allograft replacement with varying degrees of success has been reported. Most
authors prefer to use cryopreserved rather than fresh allografts because they increase the
availability of suitable conduits for emergency use, blood compatibility can be matched,

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

■c

I

AORTIC GRAFT INFECTIONS 331

and long-term storage allows the allografts to be judged relatively safe from viral trans-
mission by observing outcomes in recipients of organs from the same donor (69,70). In
addition, in contrast to refrigerated or freshly implanted allografts, cryopreserved arterial
allografts may be relatively inert immunologically. Optimal cryopreservation methods
for arterial allografts are not well delineated (68). Current cryopreservation protocols
usually recommend rate-controlled freezing and storage at very low temperatures in liquid
nitrogen vapor, mainly as a means of achieving long-term preservation of functional en-
dothelium and smooth muscle cells. However, preservation of cellular viability may be un-
necessary for performance as a large-caliber replacement graft, and viable endothelial and
smooth muscle cells may elicit immune responses that could be deleterious. In addition,
some authors feel that current cryopreservation protocols result in making arterial allo-
grafts brittle and predisposed to graft dilation and rupture (69).

Degenerative changes in allografts are the major drawback of this technique. Nu-
merous studies on arterial allograft rejection have identified the sequence of events in
arterial wall immune injury and response that progressively leads to graft dilation and
rupture. Acute and delayed allograft disruption has been reported and is a distinct limi-
tation of using allografts in infected fields (64,70,79).

Technical refinements in the use of cryopreserved allografts have been reported to
reduce the incidence of early and late complications (69). These include use of allografts of
appropriate length to prevent tension at anastomoses, careful through-and-through li-
gature of side branches (simple ligation of side branches may lead to rupture at these sites),
and circumferential anastomotic reinforcement with allograft strips. In addition, aggres-
sive retroperitoneal drainage and prolonged antibiotic and antifungal therapy has been
recommended to prevent infection of allografts (69).

Replacement of the infected aortic graft with a new synthetic graft has also been re-
ported (71-75). Most often, the new aortic graft is soaked in an antibiotic solution prior to
implantation. It is recommended that a gelatin-sealed polyester graft be soaked in a ri-
fampin solution of 60 mg/mL for this purpose (74,75). This approach is most often suc-
cessful with limited infections of low virulence following aggressive debridement of all
infected vascular and surrounding tissues to create a clean field. Despite this, the potential
for reinfection is a serious drawback, and patients treated in this manner require close and
vigilant follow-up with frequent imaging studies such as CT scanning or MRI. They are
also usually treated with lifelong oral antibiotics.

In situ prosthetic and allograft reconstructions may have their greatest utility in very
ill and unstable patients and also in those with actively bleeding aortoenteric fistulas. Ex-
peditious in situ replacement in such cases may be lifesaving. Under these circumstances,
the procedure may be used as a "bridge" procedure, with definitive reconstruction (extra-
anatomic or NAIS) carried out at a later date, when the patient has been rendered fit for
such a reconstruction.

IX. ALTERNATIVE APPROACHES TO REMOVING THE ENTIRE I

AORTIC GRAFT |

Conservative approaches that do not involve removal of the infected aortic graft have also ■jj.

been reported (80-83). These are based on aggressive drainage and debridement of in- jj

fected tissues; intensive, culture-specific antibiotic therapy; meticulous wound care to 2

achieve coverage of exposed prosthetic material; and coverage of exposed prosthetic ma- I

terial with muscle flaps. The most appropriate use of these conservative approaches is ©

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

332 CLAGETT

when infection is extracavitary and limited in extent, systemic signs of sepsis are absent,
the infecting organisms are of low virulence, and anastomotic sites are uninvolved (83). As
with in situ prosthetic replacement, these patients need close follow-up and indefinite oral
antibiotic treatment. A conservative approach may be the only option in some frail and
desperately ill patients who would not be able to tolerate aortic graft removal.

With infections that involve only one limb of an aortobifemoral bypass graft, re-
section of the limb is usually performed (13,84-87). Revascularization is often carried out
via obturator bypass or other reconstructions performed in clean fields. Autogenous su-
perficial femoropopliteal or greater saphenous vein grafts have also been used for this
purpose (87). It is important that the extent of infection be assessed with imaging studies
as well as direct visual inspection. In the case of unilateral femoral infection of an aorto-
femoral bypass, the general approach is to begin the operation by inspection of the intra-
abdominal portion of the prosthesis. If the infection grossly involves the main body of the
bifurcated prosthesis, complete removal is necessary. If the suspected limb is well incor-
porated and free of gross infection, division of the limb, closure of the tunnel, and ob-
turator or other extra-anatomic bypass is performed. The final portion of this operation is
to remove the infected limb from below, taking care to prevent cross-contamination of
other freshly placed incisions that have been closed.

X. CONCLUSIONS

There are multiple operative management strategies that are appropriate for the treatment
of infected aortic grafts. All have advantages and disadvantages that must be taken into
account in dealing with individual patients. Extra-anatomic bypass is a relatively straight-
forward procedure, can be staged, and may be physiologically less stressful than others.
However, thrombectomy and revision are often required and long-term patency is only
fair. The long-term amputation rates are high and anticoagulation is often used to main-
tain patency. In addition, reinfection of the prosthetic extra-anatomic bypass and aortic
stump blowout are of concern. In situ replacement with superficial femoropopliteal vein
grafts provides the best long-term patency and durability. Amputation rates are low and
indefinite antithrombotic and antibiotic therapies are unnecessary. However, the proce-
dure is long and complex; it can be associated with significant lower extremity ischemia
times. Leg fasciotomy is also necessary in about one-quarter of these patients. In situ al-
lograft replacement is expeditious, but reinfection, allograft aneurysmal and occlusive de-
terioration, and limited availability make this option less attractive. In situ prosthetic
replacement is also expeditious, but its use is limited to low-grade, nonvirulent infections
involving only part of the prosthesis. In addition, antibiotic therapy of indefinite duration
is usually required and the potential for reinfection is always present. All of these man-
agement options are appropriate in specific circumstances; their judicious use will lead to
improved outcomes.

REFERENCES

1. Balas P. An overview of aortofemoral graft infection. Eur J Vase Endovasc Surg 1997; 14(suppl
A):3-4.

2. Kearney RA, Eisen HJ, Wolf JE. Nonvalvular infections of the cardiovascular system. Ann
Intern Med 1994; 121:219-230.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC GRAFT INFECTIONS 333

3. O'Brien T, Collin J. Prosthetic vascular graft infection. Br J Surg 1992; 79:1262-1267.

4. Lorentzen JE, Nielsen OM, Arendrup H, et al. Vascular graft infection: An analysis of sixty-two
graft infections in 2411 consecutively implanted synthetic vascular grafts. Surgery 1985; 98:81-
86.

5. Seabrook GR. Pathobiology of graft infections. Semin Vase Surg 1990; 3:81-88.

6. Bandyk DF, Berni GA, Thiele BL, Towne JB. Aortofemoral graft infection due to Staph-
ylococcus epidermidis. Arch Surg 1984; 119:102-108.

7. Jones L, Braithwaite BD, Davies B, Heather BP, Earnshaw JJ. Mechanism of late prosthetic
vascular graft infection. Cardiovasc Surg 1997; 5:486-489.

8. Zdanowski Z, Danielsson G, Jonung T, Norgren L, Ribbe E, Thorne J, Kamme C, Schalen C.
Intraoperative contamination of synthetic vascular grafts. Effect of glove change before graft
implantation. A prospective randomized study. Eur J Vase Endovasc Surg 2000; 19:283-287.

9. Schwartz JA, Powell TW, Burnham SJ, Johnson G Jr. Culture of abdominal aortic aneurysm
contents. Arch Surg 1987; 122:777-780.

10. Ernst CB, Campbell HC, Daugherty ME, et al. Incidence and significance of intra-operative
bacterial cultures during abdominal aortic aneurysmectomy. Ann Surg 1977; 185:626-633.

11. Hayrinen-Immonen R, Ikonen TS, Lepantalo M, Lindgren L, Lindqvist C. Oral health of pa-
tients scheduled for elective abdominal aortic correction with prosthesis. Eur J Vase Endovasc
Surg 2000; 19:294-298.

12. Padberg FT Jr, Smith SM, Eng RHK. Accuracy of disincorporation for identification of
vascular graft infection. Arch Surg 1995; 130:183-187.

13. Bandyk DF, Bergamini TM, Kinney EV, Seabrook GR, Towne JB. In situ replacement of
vascular prostheses infected by bacterial biofilms. J Vase Surg 1991; 13:575-583.

14. Naylor AR, Hayes PD, Darke S, on behalf of the Joint Vascular Research Group. A
prospective audit of complex wound and graft infections in Great Britain and Ireland: The
emergence of MRSA. Eur J Vase Endovasc Surg 2001; 21:289-294.

15. Nasim A, Thompson MM, Naylor AR, Bell PRF, London NJM. The impact of MRSA on
vascular surgery. Eur J Vase Endovasc Surg 2001; 22:211-214.

16. Merritt K, Hitchins VM, Neale AR. Tissue colonization from implantable biomaterials with
low numbers of bacteria. J Biomed Mater Res 1999; 44:261-265.

17. Kaplan SS, Basford RE, Jeong MH, Simmons RL. Mechanisms of biomaterial-induced su-
peroxide release by neutrophils. J Biomed Mater Res 1994; 28:377-386.

18. Kaplan SS, Basford RE, Jeong MH, Simmons RL. Biomaterial-neutrophil interactions: Dys-
regulation of oxidative functions of fresh neutrophils induced by prior neutrophil-biomaterial
interaction. B Biomed Mater Res 1996; 30:67-75.

19. Kaplan SS, Heine RP, Simmons RL. Defensins impair phagocytic killing by neutrophils in
biomaterial-related infection. Infect Immun 1999; 67:1640-1645.

20. Brunstedt MR, Sapatnekar S, Rubin KR, Kieswetter KM, Ziats NP, Merritt K, Anderson JM.
Bacterial/blood/material interactions: I. Injected and preseeded slime-forming Staphylococcus
epidermidis in flowing blood with biomaterials. J Biomed Mater Res 1995; 29:455-466.

21. Wang I, Anderson JM, Jacobs MR, Marchant RE. Adhesion of Staphylococcus epidermidis to
biomedical polymers: contributions of surface thermodynamics and hemodynamic shear

conditions. J Biomed Mater Res 1995; 29:485-493. 1

22. Shive MS, Hasan SM, Anderson JM. Shear stress effects on bacterial adhesion, leukocyte <S
adhesion, and leukocyte oxidative capacity on a polyetherurethane. J Biomed Mater Res 1999; s
46:511-519. Jf

23. Veenstra GC, Cremers FFM, van Dijk H, Fleer A. Ultrastructural organization and regulation ^
of a biomaterial adhesion of Staphylococcus epidermidis. J Bacteriol 1996; 178:537-541. >9

24. Mack D, Riedewald J, Rohde H, et al. Essential functional role of the polysaccharide inter- J
cellular adhesion of Staphylococcus epidermidis in hemagglutination. Infect Immun 1999; a
67:1004-1008. |

25. Rupp ME, Ulphani JS, Fey PD, Bartscht K, Mack D. Characterization of the importance of S

©

!

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

334 CLAGETT

polysaccharide intercellular adhesin/hemagglutinin of Staphylococcus epidermidis in the path-
ogenesis of biomaterial-based infection in a mouse foreign body infection model. Infect Immun
199; 67:2627-2632.

26. Henke PK, Bergamini TM, Watson AL, Brittian KR, Powell DW, Peyton JC. Bacterial pro-
ducts primarily mediate fibroblast inhibition in biomaterial infection. J Surg Res 1998; 74:17-
22.

27. Geary KJ, Tomkiewicz ZM, Harrison HN, et al. Differential effects of a gram-negative and a
gram-positive infection on autogenous and prosthetic grafts. J Vase Surg 1990; 11:339-347.

28. McArthur C, Teodorescu V, Eisen L, et al. Histopathologic analysis of endovascular stent
grafts from patients with aortic aneurysms: Does healing occur? J Vase Surg 2001; 33:733-738.

29. Jackson MR, Joiner DR, Clagett GP. Excision and autogenous revascularization of an infected
aortic stent graft resulting from a urinary tract Infection. J Vase Surg 2002; 36:622-624.

30. Schlensak C, Doenst T, Hauer M, et al. Serious complications that require surgical inter-
ventions after endoluminal stent-graft placement for the treatment of infrarenal aortic aneu-
rysms. J Vase Surg 2001; 34:198-203.

3 1 . Modrall JG, Clagett GP. The role of imaging techniques in evaluating possible graft infections.
Semin Vase Surg 1999; 12:339-347.

32. Low RN, Wall SD, Jeffrey RB, et al. Aortoenteric fistula and perigraft infection: evaluation
with CT. Radiology 1990; 175:157-162.

33. Qvafordt PG, Reilly LM, Mark AS, et al. Computerized tomographic assessment of graft
incorporation after reconstruction. Am J Surg 1985; 150:227-231.

34. Auffermann W, Olofsson PA, Rabahie GN, et al. Incorporation versus infection of retro-
peritoneal aortic grafts: MR imaging features. Radiology 1989; 172:359-362.

35. Brunner MC, Mitchell RS, Baldwin JC, et al. Prosthetic graft infection: Limitations of indium
white blood cell scanning. J Vase Surg 1986; 3:42—48.

36. Fiorani P. Speziale F, Rizzo L, et al. Detection of aortic graft infection with leukocytes labeled
with technetium 99m-hexametazime. J Vase Surg 1993; 17:87-96.

37. LaMuraglia GM, Fischman AJ, Strauss HW, et al. Utility of the indium Ill-labeled human
immunoglobulin G scan for the detection of focal vascular graft infection. J Vase Surg 1989;
10:20-28.

38. Kaiser A, Kaiser B, Clayson KR, Mulherin JL, et al. Antibiotic prophylaxis in vascular sur-
gery. Ann Surg 1978; 188:283-288.

39. Pitt HA, Postier RG, MacGowan WAL, et al. Prophylactic antibiotics in vascular surgery:
Topical, systemic, or both? Ann Surg 1980; 192:356-364.

40. Hasselgren P, Ivarsson L, Risberg B, Seeman T. Effects of prophylactic antibiotics in vascular
surgery: A prospective, randomized, double-blind study. Ann Surg 1984; 200:86-92.

41. Hall JC, Christiansen KJ, Goodman M, et al. Duration of antimicrobial prophylaxis in
vascular surgery. Am J Surg 1998; 175:87-90.

42. Giovannacci L, Renggli JC, Eugster T, et al. Reduction of groin lymphatic complications by
application of fibrin glue: Preliminary results of a randomized study. Ann Vase Surg 2001;
15:182-185.

43. Kurz A, Sessler DL, Lenhardt R. Study of Wound Infection and Temperature Group.
Perioperative normothermia to reduce the incidence of surgical wound infection and shorten j>
hospitalization. N Engl J Med 1996; 334:1209-1215. |

44. O'Hara PJ, Hertzer NR, Beven EG, Krajewski LP. Surgical management of infected ab- js
dominal aortic grafts: Review of a 25-year experience. J Vase Surg 1986; 2:725-731. °|

45. Reilly LM, Stoney RJ, Goldstone J, Ehrenfeld WK. Improved management of aortic graft ^
infection: The influence of operation sequence and staging. Vase Surg 1987; 5:421^-31. >9

46. Quinones-Baldrich WJ, Hernandez JJ, Moore WS. Long-term results following surgical man- J
agement of aortic graft infection. Arch Surg 1991; 126:507-511. q

47. Ricotta JJ, Faggioli GL, Stella A, et al. Total excision and extra-anatomic bypass for aortic |
graft infection. Am J Surg 1991; 162:145-149. S

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

AORTIC GRAFT INFECTIONS 335

48. Leather RP, Darling RC III, Chang BB, Shah DM. Retroperitoneal in-line aortic bypass for
treatment of infected infrarenal aortic grafts. Surg Gynecol Obstet 1992; 175:491-494.

49. Olah A, Vogt M, Laske A, Carrell T, Bauer E, Turina M. Axillo-femoral bypass and simul-
taneous removal of the aorto-femoral vascular infection site: Is the procedure safe? Eur J Vase
Surg 1992; 6:252-254.

50. Bacourt F, Koskas F, and the French University Association for Research in Surgery.
Axillobifemoral bypass and aortic exclusion for vascular septic lesions: A multicenter retro-
spective study of 98 cases. Ann Vase Surg 1992; 6:119-126.

51. Lehnert T, Gruber HP, Maeder N, Allenberg JR. Management of primary aortic graft in-
fection by extra-anatomic bypass reconstruction. Eur J Vase Surg,, 307-701.

52. Sharp WJ, Hoballah JJ, Mohan CR, et al. The management of the infected aortic prosthesis: A
current decade of experience. J Vase Surg 1994; 19:844-850.

53. Kuestner LM, Reilly LM, Jicha DL, Ehrenfeld WK, Goldstone J, Stoney RJ. Secondary
aortoenteric fistula: Contemporary outcome with use of extraanatomic bypass and infected
graft excision. J Vase Surg 1995; 21:184-196.

54. Hannon RJ, Wolfe JHN, Mansfield AO. Aortic prosthetic infection: 50 patients treated by
radical or local surgery. Br J Surg 1996; 83:654-658.

55. Schmitt DD, Seabrook GR, Bandyk DF, Towne JB. Graft excision and extra-anatomic
revascularization: The treatment of choice for the septic aortic prosthesis. J Cardiovasc Surg
1990; 31:327-332.

56. Bunt TJ. Vascular graft infections: A personal experience. Cardiovasc Surg 1993; 1:489^492.

57. Yeager RA, Taylor LM, Moneta GL, Edwards JM, Nicoloff AD, McConnell DB, Porter JM.
Improved results with conventional management of infrarenal aortic infection. J Vase Surg
1999; 30:76-83.

58. Clagett GP, Bowers BL, Lopez- Viego MA, Rossi MB, Valentine RJ, Myers SI, Chervu A.
Creation of a neo-aortoiliac system from lower extremity deep and superficial veins. Ann Surg
1993; 218:239-249.

59. Nevelsteen A, Lacroix H, Suy R. Autogenous reconstruction with the lower extremity deep
veins: An alternative treatment of prosthetic infection after reconstructive surgery for aortoiliac
disease. J Vase Surg 1995; 22:129-134.

60. Clagett GP, Valentine RJ, Hagino RT. Autogenous aortoiliac/femoral reconstruction from
superficial femoral-popliteal veins: Feasibility and durability. J Vase Surg 1997; 25:255-270.

61. Gorden LL, Hagino RT, Jackson MR, Modrall JG, Valentine RJ, Clagett GP. Complex
aortofemoral prosthetic infections — The role of autogenous superficial femoropopliteal vein
reconstruction. Arch Surg 1999; 134:615-621.

62. Kieffer E, Bahnini A, Koskas F, Ruotolo C, LeBlevec D, Plissonnier D. In situ allograft
replacement of infected infrarenal aortic prosthetic grafts: results in forty-three patients. J Vase
Surg 1993; 17:349-356.

63. Vogt PR, Pfammatter T, Schlumph R, et al. In situ repair of aortobronchial, aortoesophageal,
and aortoenteric fistulae with cryopreserved aortic homografts. J Vase Surg 1997; 26:11-17.

64. Ruotolo C, Plissonnier D, Bahnini A, Koskas F, Kieffer E. In situ arterial allografts: a
new treatment for aortic prosthetic infection. Eur J Vase Endovasc Surg 1997; 14(suppl A): 102-

107. 1

65. Nevelsteen A, Feryn T, Lacroix H, Suy R, Goffin Y. Experience with cryopreserved arterial <S
allografts in the treatment of prosthetic graft infections. Cardiovasc Surg 1998; 4:378-383. s

66. Chiesa R, Astore S, Piccolo G, Melissano G, et al. Fresh and cryopreserved arterial homografts °|
in the treatment of prosthetic graft infections: Experience of the Italian Collaborative Vascular ^
Homograft Group. Ann Vase Surg 1998; 12:457-462. |

67. Verhelst R, Lacroix V, Vraux H, Lavigne JP, Vandamme H, Limet R, Nevelsteen A, Bellens B, ^
Vasseur MA, Wozniak B, Goffin Y. Use of cryopreserved arterial homografts for management q
of infected prosthetic grafts: a multicentric study. Ann Vase Surg 2000; 14:602-607. |

68. Leseche G, Castier Y, Petit MD, Bertrand P, Kitzis M, Mussot S, Besnard M, Cerceau O. 2

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

336 CLAGETT

Long-term results of cryopreserved arterial allograft reconstruction in infected prosthetic grafts
and mycotic aneurysms of the abdominal aorta. J Vase Surg 2001; 34:616-622.

69. Vogt PR, Brunner-LaRocca HP, Lachat M, Ruef C, Turina MI. Technical details with the use
of cryopreserved arterial allografts for aortic infection: Influence on early and midterm mor-
tality. J Vase Surg 2002; 35:80-86.

70. Noel AA, Gloviczki P, Cherry KJ Jr, San H, Goldstone J, Morasch MD. Johansen KH,
members of the United States Cryopreserved Aortic Allograft Registry. Abdominal aortic
reconstruction in infected fields: Early results of the United States Cryopreserved Aortic Al-
lograft Registry. J Vase Surg 2002; 35:847-852.

71. Walker WE, Cooley DA, Duncan JM, Hallman GL, Ott DA, Reul GJ. The management of
aortoduodenal fistula by in situ replacement of the infected abdominal aortic graft. Ann Surg
1987; 205:727-732.

72. Speziale F, Rizzo L, Sbarigia E, et al. Bacterial and clinical criteria relating to the outcome of
patients undergoing in situ replacement of infected abdominal aortic grafts. Eur J Vase Endo-
vasc Surg 1997; 13:127-133.

73. Hayes PD, Nasim A, London NJM, Sayers RD, Barrie WW, Bell PRF, Naylor AR. In situ
replacement of infected aortic grafts with rifampicin-bonded prostheses: The Leicester expe-
rience (1992 to 1998). J Vase Surg 1999; 30:92-98.

74. Young RM, Cherry KJ Jr, Davis PM, Gloviczki P, Bower TC, Panneton JM, Hallett JW Jr.
The results of in situ prosthetic replacement for infected aortic grafts. Am J Surg 1999; 178:
136-140.

75. Bandyk DF, Novotney ML, Back MR, et al. Expanded application of in situ replacement for
prosthetic graft infection. J Vase Surg 2001; 34:411-420.

76. Clagett GP. Treatment of aortic graft infection. In: Ernst CB Stanley JC, eds. Current Therapy
in Vascular Surgery. 4th ed. Philadelphia: Mosby-Year Book, 2001:422-428.

77. Seidel SA, Modrall JG, Jackson MR, Valentine RJ, Clagett GP. The superficial femoral-
popliteal vein graft: A reliable conduit for large caliber arterial and venous reconstructions.
Perspect Vase Surg Endovasc Ther 2001; 14(l):57-80.

78. Wells JK, Hagino RT, Bargmann KM, Jackson MR, Valentine RJ. Kakish HB, Clagett GP.
Venous morbidity after superficial femoral-popliteal vein harvest. J Vase Surg 1999; 29:282-
291.

79. Koskas F, Plissonnier D, Bahnini A, Ruotolo C, Kieffer E. In situ arterial allografting for
aortoiliac graft infection: A 6-year experience. Cardiovasc Surg 1996; 4:495-499.

80. Calligaro KD, Veith FJ, Schwartz ML, et al. Selective preservation of infected prosthetic
arterial grafts. Analysis of a 20-year experience with 120 extracavitary-infected grafts. Ann
Surg 1994; 220:461^471.

81. Morris GE, Friend PJ, Vassallo DJ, Farrington M, Leapman S, Quick CRG. Antibiotic
irrigation and conservative surgery for major aortic graft infection. J Vase Surg 1994; 20:88-95.

82. Belair M, Soulez G, Oliva VL, et al. Aortic graft infection: The value of percutaneous drainage.
Am J Radiol 1998; 171:119-124.

83. Calligaro KD. Veith FJ. Graft preserving methods for managing aortofemoral prosthetic graft
infection. Eur J Vase Endovasc Surg 1997; 14(suppl A):38-42.

84. Becquemin JP, Qvarfordt P, Kron J, et al. Aortic graft infection: Is there a place for partial j>
graft removal? Eur J Vase Endovasc Surg 1997; 14(suppl A):53-58. <S

85. Miller JH. Partial replacement of an infected arterial graft by a new prosthetic polytetrafluoro- js
ethylene segment: A new therapeutic option. J Vase Surg 1993; 17:546-558. °|

86. Towne JB, Seabrook GR, Bandyk D, Freischlag JA, Edmiston CE. In situ replacement of ^
arterial prosthesis infected by bacterial biofilms: Long-term follow-up. J Vase Surg 1994; 19: >9
226-235. |

87. Sladen JG, Chen JC, Reid JDS. An aggressive local approach to vascular graft infection. Am J q
Surg 1998; 176:222-225. |

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

18

Detection and Management of Failing Autogenous Grafts

Jonathan B. Towne

Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

A significant decline in primary patency of autogenous grafts invariably occurs over time
because of the development of fibrointimal hyperplasia of the vein conduit or anastomotic
sites and the progression of atherosclerotic disease of the native arteries. The incidence of
lesions that threaten the long-term patency of both in situ and reversed saphenous vein
bypasses ranges from 20 to 26% (1-3). Anatomical and hemodynamic alternations that
require revision will develop in approximately 5% of vein bypasses per year; the majority of
these will occur with the first 2 years after bypass (4). The failure to detect and correct lesions
that threaten bypass patency before graft thrombosis significantly decreases long-term
patency of the autogenous graft. The long-term patency for revisions of autogenous grafts
after thrombosis is dismal. The 3-year patency rate of revision of thrombosed in situ and
reversed saphenous vein grafts ranges from 22 to 47% in recent reports (1,5,6).

By contrast, the result of revision on hemodynamically failing but patent conduits are
excellent: secondary patency rates are equivalent to those for bypasses that never undergo
revision (1). Secondary procedures on patent grafts normalize the hemodynamics at the
revision site and are associated with a low incidence of restenosis; long-term patency rates
range from 80 to 93% (1,7-9). The excellent durability of secondary procedures on the
patent but hemodynamically abnormal bypass coupled with the dismal results of revision of
the thrombosed bypass underscores the significant impact of monitoring graft hemody-
namics with a surveillance protocol and elective bypass revision. This chapter discusses the -6
definition and detection of autogenous graft failure by intraoperative and postoperative |
surveillance and details the surgical principles and results of management of the failing |
autogenous graft. |j>

%

I. ETIOLOGY OF AUTOGENOUS GRAFT FAILURE %

I
The etiology of autogenous graft failure differs according to the time of occurrence. «

Perioperatively (up to 30 days), autogenous graft failure most commonly is caused by |

technical errors, an inadequate inflow or outflow artery, or an inadequate vein. From 30 @

337 I

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

338 TOWNE

days to 2 years postoperatively, stenosis of the vein conduit or the anastomotic sites caused
by fibrointimal hyperplasia is the most common cause of graft failure. Autogenous graft
failure occurring more than 2 years postoperatively is likely a result of atherosclerotic
disease progression of the inflow or outflow artery. Lesions associated with a diameter-
reducing stenosis of the graft or native artery, regardless of the cause, can be identified
reliably by abnormal graft hemodynamics. Occasionally autogenous graft thrombosis
occurs despite normal hemodynamics because of thromboembolism, anastomotic pseudo-
aneurysm formation, aneurysmal degeneration of the conduit, infection, or a hypercoagu-
lable state.

II. PERIOPERATIVE FAILURE

Perioperative failure of the autogenous graft is related to technical errors and the quality of
the arterial and venous systems used for bypass. Technical errors in constructing the
proximal and distal anastomoses and preparing the conduit can occur. Dissecting and
clamping the inflow and outflow arteries can raise intimal flaps, which can create luminal
stenoses and/or thromboses. Errors in performing the proximal and distal anastomoses can
result in suture line bleeding, pseudoaneurysm formation, or suture line stricture. Likewise,
proximal occlusive disease in the aortoiliac segment can cause graft failure. Technical
imperfections of the vein conduit include luminal thrombus, platelet aggregates, graft
torsion, kinking of the graft, graft entrapment, inadequate ligation of venous branches, and
vein injury from dissection or valve ablation.

The quality of the vein and artery used for the bypass is important for success of the
autogenous graft. A poor-quality vein (i.e., one with sclerotic segments, varicosities, or a
small diameter) can predispose to technical errors in handling the venous conduit and result
in graft thrombosis. Sclerotic vein segments can have a thrombogenic flow surface and can
also result in significant luminal stenosis, both of which can result in graft failure. Veins less
than 2 mm in diameter are inadequate for long bypasses with both the in situ and reverse
techniques. Inadequate arterial outflow caused by florid tibial and pedal occlusive disease
can cause high outflow resistance, resulting in low graft flow and subsequent graft failure.

III. POSTOPERATIVE FAILURE

Beyond the perioperative period, the two main disease processes that result in autogenous
graft failure are fibrointimal hyperplasia of the vein conduit or anastomotic sites and the
progression of atherosclerotic disease of the inflow or outflow vessel. From 1 to 24 months
postoperatively, fibrointimal hyperplasia is the most common cause of vein graft stenosis;

the incidence is 5% per year. Approximately half of the lesions causing hemodynamic failure ■a

are located in the vein conduit itself. The remaining sites involve the proximal or distal |

anastomosis or the adjacent inflow or outflow artery. Lesions such as anastomotic stenoses, a

valve leaflet fibrosis, and focal or extensive fibrotic stenoses of the vein result in significant c

diameter-reducing stenosis with time. Beyond 2 years, the incidence of significant graft <

stenoses decreases to 1-3% per year (10). During this time interval, atherosclerotic disease >9

progression is the most common cause of late graft failure. Atherosclerosis can progress in J

the native arterial inflow and outflow vessels, resulting in significant diameter-reducing «

stenosis and a decrease in blood flow. Atherosclerosis also can develop in the vein bypass |

itself, resulting in aneurysmal dilatation or stenosis of the vein graft and native arteries, @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

FAILING AUTOGENOUS GRAFTS 339

which yields hemodynamic abnormalities that can be detected and corrected during post-
operative surveillance (11,12).

IV. GRAFT SURVEILLANCE PROTOCOL

The ideal autogenous graft surveillance protocol should be noninvasive, associated with
minimal complications, and highly accurate and reliable in identifying graft lesions that
threaten patency. Based on the time course of occurrence of graft failure, surveillance of the
autogenous grafts should be most frequent during the perioperative period and the first 2
years postoperatively. The graft surveillance techniques should be applicable intraoper-
atively, immediately postoperatively, and throughout the postoperative follow-up period.
The results should be reproducible and correlate with one another on serial examinations.
Follow-up of autogenous grafts based on the recurrence of symptoms of limb ischemia is
inadequate in identifying grafts at risk for thrombosis. In studies of the in situ saphenous
vein bypass, Bandyk and associates (7,10) have documented that only one-third of patients
had recurrence of symptoms despite noninvasive vascular laboratory studies indicating
hemodynamic graft failure. In a study of reverse saphenous vein grafts, Minh Chau et al. (2)
documented that only 29% of limbs with low graft flow velocities (less than 45 cm/s) had
recurrence of symptoms alone. This would result in up to two-thirds of patients with
autogenous grafts at risk for thrombosis not being identified. Clinical and hemodynamic
assessment with use of noninvasive vascular laboratory testing is essential for the post-
operative surveillance protocol. Graft failure is indicated by the following:

Recurrence of symptoms of limb ischemia
Low peak systolic graft flow velocity (<45 cm/s)
Decrease of flow velocity >30 cm/s
Ankle/brachial index >0.15 on serial examinations

Surveillance protocols for autogenous grafts include hemodynamic monitoring of the
arterial systolic pressure of the limb and the blood-flow velocities of the venous conduit and
native arteries. Ankle systolic pressure measurements correlate with the clinical symptoms
of limb ischemia and are predictive of healing of ulcers or amputation sites, but they are
unable to localize the obstructive lesion to the inflow conduit or outflow. Also, because of
the high rate of lower limb bypass required by diabetic patients, the presence of incompres-
sible calf vessels makes measurement of ankle pressures impossible. Analysis of blood-flow
velocity with the use of a duplex scanner can evaluate the venous conduit, anastomotic sites,
and adjacent arteries for hemodynamic abnormalities of flow. Unlike limb pressures, the
duplex scan can reliably identify the location and anatomy of the lesions by ultrasound and
quantitate the severity of flow disturbance and stenosis. Systolic pressure measurement and ■§

analysis of blood-flow velocity are complementary tests that can aid in the prediction of g

recurrence of symptoms or graft thromboses, both should be used during postoperative »

surveillance of the autogenous graft (10,13) (Table 1). c

Hemodynamic surveillance should be used intraoperatively, perioperatively, and dur- <

ing the postoperative follow-up period for both in situ and reversed autogenous grafts >9

(1,2,7,10). B-mode imaging has been used to assess completed arterial reconstructions, and ^

intravascular defects can be detected by this method. Advances in technology have provided «

instrumentation that combines the advantages of Doppler spectral analysis for hemody- |

namic assessment and B-mode imaging to determine vessel wall integrity. The further @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

340 TOWNE

Table 1 Duplex Scan Classification of Severity of Stenosis
Classification Velocity Waveform Analysis

< 20% DR No increase in peak systolic velocity compared

with adjacent proximal segment spectral

broadening during systole
20-50% DR >30% increase in peak systolic velocity

compared with adjacent proximal segment;

spectral broadening during systole and diastole
50-75 % DR >100% increase peak systolic velocity

(Vp > 125 cm/s) compared with adjacent

proximal segment, end-diastolic velocity

<100 cm/s: spectral broadening during systole

and diastole
>75% DR >100% increase in peak systolic velocity

(Vp > 125 cm/s) compared with adjacent

proximal segment end-diastolic velocity

>100 cm/s: spectral broadening during systole

and diastole

Abbreviations: DR, diameter reduction; Vp, velocity peak.

addition of color-coded imaging greatly facilitates vessel imaging, making this the ideal
method for intraoperative evaluation of arterial reconstructions.

V. TECHNIQUE

Duplex examinations use color coding to display velocity spectra within the B-mode image
(14). Intraoperative techniques are identical, whether gray-scale or color-coded imaging is
employed, although color allows for more rapid vessel identification and interrogation. This
technique can be used for cerebrovascular; peripheral arterial, and visceral vessels and re-
quires a high-frequency transducer, because it is applied directly to the reconstructed vessel.
The duplex scan is performed and recorded on videotape after closure of the arteriotomy
and restoration of blood flow. A sterile plastic sleeve is filled at one end with acoustic gel or
saline before the transducer is placed inside. The vascular technologist is present to make
necessary adjustments in Doppler angle assignment, sample volume placement, and color
parameters in order to obtain the optimal image and accurate spectral display.

The examination begins with the location of patient arteriovenous fistulas (AVFs). With
the duplex probe placed on the proximal graft, the distal graft is occluded with finger
pressure. Persistence of forward flow anywhere along the conduit indicates an AVF between ■§

the point of the probe and the point of finger occlusion. The duplex probe then is passed |

down the length of the graft and the persistence of forward flow proximal to the AVF is a

noted. Once past the fistula, the forward flow diminishes. When the fistulas are ligated, there c

should be no forward flow with distal finger pressure occlusion. The proximal and distal <

anastomoses, the inflow and outflow arteries, and the flow along the entire graft conduit — >9

especially at the valve incision sites — are then evaluated. Anatomical and technical defects J

are identified and quantitated according to the extent of spectral broadening and increase of «

peak systolic frequency. Mild to moderate flow disturbances are associated with spectral |

broadening during systole and peak systolic velocity (Va > 125 cm/s). Severe flow dis- @

turbances are associated with spectral broadening during systole and diastole and peak %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

FAILING AUTOGENOUS GRAFTS 341

systolic velocity greater than 125 cm/s and end diastolic velocity > 100 cm/s. With the in situ
technique, partial valve lysis is associated with severe flow disturbances just distal to the
valve site. Sclerotic vein segments, graft torsion, platelet aggregates, and anastomotic
stenosis are often associated with severe flow disturbances at the site and just distal to the
abnormal graft segment. The location, length, and severity of stenosis can be identified
precisely with intraoperative spectral analysis.

Postoperative graft surveillance should be performed before hospital discharge using
limb blood pressure measurement and duplex scanning. Systolic pressure measurements are
obtained by measuring the ankle/brachial index (ABI); duplex scanning of proximal and
distal anastomoses, inflow and outflow arteries, and the entire autogenous conduit can be
performed. Velocity waveform analysis can be performed by placing the sample volume in
the center of the vessel lumen at the desired Doppler beam angle. Calculation of the graft
flow velocities at designated sites with normal arterial flow patterns, where the vessel diam-
eter does not vary, is necessary for reliable and reproducible results from one examination to
the next. Flow velocity should be measured with the patient in the supine position, limbs
slightly rotated externally, and the knee bent and at rest.

Vein graft lesions detected in this time period include arteriovenous fistulas, residual
intact valve leaflets, and graft or anastomotic stenosis. Arteriovenous fistulas are recognized
by proximally located high diastolic flows, increased flow turbulence at the site, and
decreased or absent diastolic flow distal to the site of the fistula. Stenotic lesions of the
graft conduit or anastomotic site are identified by increases in spectral broadening and peak
systolic and diastolic velocities at or just distal to the site. In addition to identifying
hemodynamically significant lesions predisposing to early graft failure, early duplex
scanning provides a baseline for subsequent graft surveillance studies. Serial evaluations
with postoperative hemodynamic surveillance should be performed at 6 weeks postoper-
atively, every 3 months for the first year, and every 6 months thereafter. Each examination
should include limb pressure measurements and determination of peak systolic flow
velocities at the middle and distal graft; the results should be compared with prior studies.
Autogenous grafts that exhibit abnormal hemodynamics (low graft flow velocities, >50%
graft stenoses, significant decrease in peak systolic frequency on serial examinations greater
than 30 cm/s, or decrease inABI greater than 0.15) should be evaluated by complete duplex
scanning of the entire graft and by arteriography.

More recently, studies have been conducted to determine whether ongoing surveillance
needs to be done for the life of the conduit. In the long-term evaluation of 462 saphenous
vein in situ bypasses performed over a 13-year period, 30% of the grafts required at least one
revision (11). Even grafts that have exhibited good hemodynamics for up to 24 months are at
risk for developing abnormalities that could lead to graft failure. Of the initial graft revisions
in this study, 18% occurred after 24 months. Because of the increasing incidence of athero-
sclerosis in the inflow and outflow vessels with long-term follow-up, a greater percentage of ■§
revisions involve inflow and outflow vessels as opposed to the conduit itself. Of the revisions g
performed after 24 months, 68% were to the conduit, compared with 85% in the earlier time as
period. c

As the follow-up becomes longer, degenerative changes can develop in the conduit itself. <

Previous work from our institution has demonstrated that more than 50% of vein bypass >9

conduits followed for at least 5 years demonstrated evidence of atherosclerotic degeneration ^

(12). Often these changes represent areas of intimal thickening, but in the significant portion «

the disease has progressed to form focal point stenosis caused by atherosclerosis. The |

likelihood of developing graft-threatening lesions is even greater in conduits that have @

previously been revised or have hemodynamic abnormalities, but some conduits that have |j,

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

342

TOWNE

been free of abnormalities for the first 2 years will go on to require revision. By recognizing
that conduits which have previously required revision are more prone to develop secondary
degenerative processes, the surveillance of these conduits can be more focused.

Other authors have suggested that if the conduit has normal hemodynamics in the early
perioperative period, the chances that problems may arise are such that further surveillance
may not be warranted (15). Our study reveals that of the 67 graft revisions performed after
24 months, 37 were to previously revised conduits, but 30 were to vein grafts that had
required no previous revision (11). Conduits that are hemodynamically normal beyond 2
years evolved lesions at a rate significant enough to warrant ongoing surveillance. The
average incidence of primary graft failure was 10% of the number of grafts remaining patent
at each yearly time interval beyond 24 months. If vascular surgeons want to optimize long-
term graft patency, surveillance must continue for the life of the conduit. Other authors
whose series consist of primarily reverse vein bypasses have drawn similar conclusions (16).

The failure to identify and revise a graft with a hemodynamically significant lesion
before thrombosis is associated with poor patency. Long-term patency rates are improved
markedly for autogenous grafts revised while patent, compared with long-term patency
rates for revision of thrombosed bypass conduits. The secondary patency rates for bypasses
patent at the time of revision are equivalent to those for bypasses that never undergo
revision. Revision of the thrombosed reversed or in situ saphenous vein bypass uniformly
results in poor long-term patency. In a study of 1 09 autogenous grafts, Whittemore et al. ( 1 7)
reported that revision of the patent but failing autogenous graft yielded an 85% 5-year
patency rate, compared with a 37% 5-year patency rate for revision of the thrombosed
bypass. In a study of 95 bypasses that underwent revision, Bergamini et al. (1) found a
significant decline in the secondary patency for bypasses thrombosed at the time of revision
(47%) compared with that for bypasses patent at the time of revision (93%)(Fig. 1).

s

Patent

100

80

60

% Patent

40

20

(72) (71) (70) (67) (62) ( 56 ) , ,

— a a » « B x '

-a 1_

-! 1_

(46)

-1 1 j 93%

(23)

(19)

J I I I I

(16) (11) (10) (10) (9) L

\

(8)
Thrombosed

47%

0-1 1-3 3-6 6-12 12-18 18-24 24-30 30-36

Months

Figure 1 Secondary patency of in situ saphenous vein bypasses that were patent at time of revision
(open circles) was significantly higher (j> < 0.0000005) than that of bypasses that were thrombosed at
time of revision (closed circles).

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

FAILING AUTOGENOUS GRAFTS

343

Table 2 Incidence of Technical Errors with In Situ Saphenous
Vein Bypass: First Half of Series vs. Later Half 11

Technical error

51-1985

1986-1990

Valvulotome injury
Anastomotic stenosis

20

5

9

Bypass torsion
Residual valve leaflet
Total b

4

4

33

(18%)

9 (5%)

a In the first half of the series (1981-1985), there were 179 cases; in
the second half (1986-1990), there were 182 cases.
b Total number of technical errors for the first half compared with
the second half is significantly different/) = 0.0001(z 2 analysis).

Compulsive bypass surveillance can identify the failing autogenous graft and permit elective
revision, a practice that has significant impact on long-term patency.

The superior long-term patency rates of both in situ and reversed autogenous grafts also
were related directly to improved surgical technique and the experience of the vascular
surgeon (1,18,19). Increasing surgical experience was associated with a significant decrease
in the number of technical errors encountered with vein preparation (1) (Table 2). Increasing
surgical experience was also associated with a significant improvement in the secondary
patency rate (92%) compared with the secondary patency rate of the first half of the series
(80%o) at 3 years (Fig. 2). Taylor et al. (19) have also attributed the recent improvement in
long-term patency of reversed autogenous grafts to improvements in surgical technique and

% Patent

S

1986-1990

100

(182) (171)
B 1 a .

(158)

(139)

(116)

(85) (51) (40)

< 1 ™ (168)

(162)

-U.

1 L

....I., . u I

80

(150)

(134)

< 124 > (116) (109)

\

\ 1981-1985

60
it

40

20

92%
60%

I
•5

0-1 1-3 3-6 6-12 12-18 18-24 24-30 30-36

Months

Figure 2 Secondary patency of in situ saphenous vein bypasses performed by the more experienced
surgeon from 1986 to 1990 (open circles) was significantly higher (p < 0.02) than that of bypasses in
the earlier years from 1981 to 1985 (closed circles).

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

344 TOWNE

increasing surgical experience. Long-term patency rates of the in situ and reversed
autogenous grafts are improved with careful operative technique, meticulous postoperative
graft surveillance, and increasing surgical experience.

VI. LONG-TERM CHANGES IN AUTOGENOUS GRAFTS

Autogenous grafts that are in place for a long time are vulnerable to the same atherosclerotic
changes as the native arteries, although in overall timing they take less time to develop. In
a study of 72 lower extremity vein grafts that functioned and were patent from 4.6 to
21.6 years, the median being 6.6 years, we found that only 43% were normal (12). In
addition, nearly 1 in 5 grafts harbored a lesion that was felt to pose a threat to continued
graft patency.

Atherosclerotic degeneration of saphenous vein grafts was first described in 1947 when a
femoral interposition graft that had been in place for 22 years was removed and found to
contain atheromatous plaques. In 1973, Szilagyi et al. reported their experience with a lower
extremity saphenous vein graft that had been monitored with arteriography (20). They
described eight different morphological findings in grafts of varying ages. Several of these
were related to surgical technique, including suture stenosis caused by tying side branches
too closely, long venous side branch stumps, and traumatic stenosis caused by clamps. They
also described the changes that occur, including intimal thickening, myointimal hyperplasia
at valve sites, atherosclerotic irregularity, and aneurysmal dilatation. In our study,
autogenous grafts were examined at least 4.6 years after construction. The early post-
operative changes that Szilagyi et al. described were not detectable. However, we did find
three distinct atherosclerotic abnormalities: wall plaque, aneurysmal dilatation, and discrete
stenosis.

The most prevalent finding was wall plaque, which was present in all of the abnormal
grafts, although this was frequently overshadowed by more impressive stenosis or aneu-
rysms. Typically, plaques were several centimeters long, multicentric, echogenic, and slightly
raised from the normal wall. These are mild forms of atherosclerotic degeneration. In the
series of Szilagyi et al., atherosclerotic changes developed at approximately 45 months.
Atkinson et al. looked at coronary artery saphenous vein grafts during autopsy and found
atherosclerotic changes in 21% of the grafts that had been placed an average of 62 months
(21). However, when DeWeese and Rob monitored long-term grafts with the use of
arteriography, they noted atherosclerotic changes in only 3 of 18 patients studied after 5
years, and two grafts did not develop changes until after 10 years (22). We use color duplex
ultrasonography to study the grafts and have been able to visualize changes in the arterial
wall, such as thickening and wall plaque, that are not necessarily seen on contrast
arteriography, which defines only the column of flowing blood.

I

VII. EFFECT OF SITE OF DISTAL ANASTOMOSIS 1

I
In contrast to the patency for in situ saphenous vein bypass, the long-term patency for the c

reversed vein graft was significantly decreased for infrapopliteal bypasses, long bypasses <

performed with vein less than 3 mm in diameter, and reversed vein grafts in diabetic patients. >9

Taylor and coworkers (19) reported a significant decline in primary patency for autogenous J

reversed grafts to the infrapopliteal arteries (69%) compared with that for below-knee °

popliteal arteries (80%). The secondary patency for infrapopliteal arteries (77%) was also |

decreased compared with that for below-knee popliteal artery grafts (86%). Rutherford et @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

FAILING AUTOGENOUS GRAFTS 345

al. (23) also showed a decrease in the cumulative patency rate at 3 years for reversed vein
grafts (63%) compared with that for in situ (88%) vein bypasses to the tibial outflow artery.

VIM. EFFECT OF VEIN GRAFT DIAMETER AND DIABETES
MELLITUS

The long-term patency of reversed autogenous grafts with vein diameters less than 3 mm is
decreased significantly compared with that for reversed autogenous grafts with greater vein
diameters (24). The in situ autogenous graft is less affected by vein size (1,18), but 2 mm is the
lowest diameter of usable vein.

Long-term patency and limb salvage are adversely affected by the presence of diabetes in
patients with reversed autogenous grafts (9,19,25). On the other hand, in situ autogenous
graft long-term patency and limb salvage are not significantly different for the diabetic
patient compared with those for the nondiabetic patient (1).

IX. OPERATIVE MANAGEMENT OF THE PATENT BUT FAILING
AUTOGENOUS GRAFT

The goals of intraoperative modification or postoperative revision of the patent but failing
autogenous graft should be correction of the anatomic abnormality, restoration of normal
hemodynamics of the autogenous graft, and maintenance of long-term bypass patency. The
principles and techniques of treatment of a graft stenosis are the same for the in situ and
reversed saphenous vein grafts. Graft revisions require a clear understanding of the cause,
location, and extent of the lesion (7). Secondary procedures require meticulous dissection
and technical precision. Regional and general anesthesia is preferred over local anesthesia to
perform the revision with minimal patient discomfort. The use of scalpel dissection is
essential to expose the autogenous graft without vein injury. After dissection of the abnor-
mal segment of the autogenous graft, intraoperative spectral analysis can be performed to
confirm the precise location and extent of the lesion. After exposure and control of bleeding
points, the patients are heparinized systemically. Secondary procedures on the distal graft,
distal anastomotic site, or native outflow artery are enhanced by the use of a pneumatic
tourniquet as a substitute for vascular clamps for proximal and distal control in the scarred
tissue planes (26). The limb is exsanguinated by leg elevation and the use of an elastic wrap
before inflation of the tourniquet (40 to 50 mmHg above brachial systolic pressure). The
secondary procedure can then be performed without the necessity to dissect the proximal
and distal conduits circumferentially for control and without the need to work around the
vascular clamps. Just before completion of the revision, the tourniquet should be deflated to
confirm the presence of back-bleeding from the distal conduit. Intraoperative spectral
analysis and arteriography are performed routinely to document the restoration of normal
graft hemodynamics and anatomic configuration.

A. Modification of the Failing Autogenous Graft During the Primary
Procedure

During the placement of the primary in situ or reversed autogenous graft, intraoperative
modification of the venous conduit is performed to correct technical errors or a poor-quality
vein. The key to successful treatment is recognition and correction of the lesion before graft
thrombosis occurs. The intraoperative occurrence of the low-flow state can be caused by

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

346 TOWNE

problems with the inflow artery, vein conduit, anastomotic site, or out flow artery. Significant
stenosis of the inflow artery can be identified by intraoperative measurement of the systolic
pressure by insertion of an 1 8- to 20-gauge needle in the common femoral artery connected
to the arterial line monitor. A significant gradient (greater than 20 mmHg) at rest or after
injection of 20 mg of papaverine between the inflow artery pressure and the brachial
systolic pressure confirms the presence of a poor inflow artery. Improved inflow can be
achieved by a proximal aortic bypass, extra-anatomic bypass, or intraoperative balloon
angioplasty if a short segmental lesion is present.

The low-flow state caused by lesions of the vein conduit or anastomotic sites can be
detected by complete survey of the graft from the proximal to the distal anastomosis with
high-frequency spectral analysis. Technical errors or poor-quality vein segments are
associated uniformly with increased peak systolic frequency and spectral broadening.
Anastomotic stenoses are caused by technical errors, such as missing the endothelium,
incorporation of adventitial tissue in the intraluminal flow surface, poorly placed stitches
resulting in suture line stricture or bleeding, and intimal flaps or raised plaques. The distal
anastomotic stricture usually involves the toe of the end-to-side anastomosis. This is best
corrected by reconstructing the entire anastomosis. Endarterectomy of the native artery to
correct intimal flaps or atherosclerotic plaques is sometimes necessary. With the reversed
vein graft, meticulous technique is mandatory in anastomosing the large end of the reversed
vein to the smaller outflow artery.

Intraoperative modification of the venous conduit for both in situ and reversed
autogenous grafts is performed to correct a poor-quality vein or technical error. Sclerotic
veins have been treated by vein patching of the sclerotic segment or resection of sclerotic
segment and replacement with a translocated vein segment from another source. If the
endothelium of the sclerotic segment is abnormal and an alternate vein source of adequate
diameter and quality is available, resection and replacement of the sclerotic vein segment is
preferred. Varicosities of the vein are treated with plication of the wall or a partial resection
and vein patch angioplasty. Small-diameter vein segment or previously used segments of
vein in the leg should be replaced with an alternate source of vein. The translocated vein
segments used for interposition replacement should be of good quality, with a thin wall, a
diameter greater than 2 mm, and a glistening endothelial flow surface. Intraoperative
modification caused by technical failure of the in situ bypass is most commonly performed
for retained valve leaflets or AVFs. The retained valve leaflet can be incised simply by
reinserting the valvulotome through a side branch or directly into the vein conduit via an 18-
gauge needle puncture hole. Sclerotic valves prompt valve excision under direct vision
through a longitudinal venotomy. The venotomy is closed primarily if it this technically
achievable without creating luminal stenosis or, if this is not possible, with a vein patch.
Arteriovenous fistulas are treated simply by ligation. Injured segments of vein during dis-
section are repaired simply with lateral venorrhaphy, vein patch angioplasty, resection and ■§
primary anastomosis, or resection and interposition grafting. The key to the treatment |
of this technical problem is to resect all abnormal endothelium and injured vein wall and a
then reconstruct the bypass graft without creating luminal stenosis or increased tension on c
the anastomosis. Vein conduit torsion is best treated by transecting the conduit, untwisting <
it, and performing a primary reanastomosis. Minor kinks or twists of less the 90 degrees >9
in the bypass are treated with vein patch angioplasty across the twisted segment. Graft J
entrapment is detected by the presence of normal graft velocities with the knee flexed, but the °
development of a low-flow state with absence of diastolic flow with the knee extended is seen |
in patients who spend significant amounts of time in wheelchairs. Treatment of this @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

FAILING AUTOGENOUS GRAFTS 347

complication includes a myotomy of the tendons and muscles in which the vein bypass is
entrapped. The formation of platelet aggregates along the endothelial flow surface caused by
heparin-induced platelet aggregation should be treated with reversal of the heparin, low-
molecular- weight dextran drip, removal of the platelet aggregates by means of longitudinal
venotomy, and subsequent primary or vein patch closure of the venotomy. Intraoperative
thrombosis of the vein graft without the presence of a poor inflow artery, technical errors,
or poor-quality vein should prompt a search for evidence of a hypercoagulable state (anti-
thrombin III deficiency, heparin-induce platelet aggregation, abnormal plasminogen) or a
poor outflow artery. Conduits with a poor outflow artery have a low-flow state and an
absence of diastolic forward flow caused by high outflow resistance (27). The best manage-
ment is a translocated vein sequential graft or jump graft from the vein bypass to an alternate
outflow artery, if available, to increase the outflow and decrease the resistance. Following
the secondary graft procedures, intraoperative pulsed Doppler spectral analysis and
arteriography should be repeated to make sure that there is normalization of the bypassed
hemodynamic and anatomical abnormalities.

X. SECONDARY PROCEDURES FOR REVISION OF THE FAILING
AUTOGENOUS GRAFT

During postoperative surveillance, operative management of the failing but patent autog-
enous graft depends on the anatomy and location of the lesion. Stenoses of the graft conduit
are secondary to fibrointimal hyperplasia at the anastomotic sites, valve sites or areas of
intraoperative vein injury. The areas of graft stenosis can vary from focal to extensive,
affecting a long segment of the graft conduit (Fig. 3). Focal stenosis in the in situ saphenous
vein bypass occurs at fibrotic valve sites. The formation of lesions caused by fibrointimal
hyperplasia usually is associated with technical errors or injury in preparing the venous
conduit for bypass. If focal stenosis develops in the vein conduit, sufficient length usually is
present to permit resection of the lesion and primary end-to-end anastomosis to restore
bypass patency. The amount of redundant graft is less for reversed autogenous grafts,
making reanastomosis an infrequent option. If there is inadequate vein length or the stenosis
is immediately adjacent to an anastomotic site, the focal stenosis is treated by a longitudinal
venotomy across the length of the stenosis and vein patch angioplasty reconstruction.
Stenoses of the proximal and distal anastomoses are treated with a vein patch angioplasty
extending distally onto the native artery. A third treatment option for focal graft stenosis is
percutaneous transluminal angioplasty (PTA). PTA is especially suitable for treatment of
focal stenosis in high-risk patients or in the native arteries proximal or distal to the graft.

More extensive lesions include long sclerotic vein segments, multiple stenotic lesions,
and fibrointimal hyperplasia or atherosclerosis involving a long segment of the distal
anastomosis or outflow artery. Long or multiple stenotic lesions of the vein conduit are ■§

treated by vein patch angioplasty or resection and interposition of a translocated vein |

segment if an alternate, good-quality vein of similar diameter is available. Long stenotic as

lesions of the distal anastomosis are treated with vein patch angioplasty across the stenotic c

segment. Extensive lesions of the distal vein conduit and outflow artery are treated with a <

jump graft (extension of the graft to the same outflow artery) or a sequential graft (translo- >9

cated vein segment to a different outflow artery) using translocated vein (greater saphenous, J

lesser saphenous, cephalic). The technical success of secondary procedures is evaluated by «

intraoperative spectral analysis and angiography to ensure achievement of normal graft |

hemodynamics and anatomic reconstruction. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

348

TOWNE

Valve Site

Figure 3 A. Focal stenotic lesions of the failing autogenous graft detected during postoperative
surveillance were caused by fibrointimal hyperplasia. Valve site stenosis was treated by excision and
end-to-end reanastomosis. Vein conduit stenosis was treated with percutaneous balloon angioplasty.
Anastomotic stenosis was treated by vein patch angioplasty. B. Extensive stenotic lesions of the failing
autogenous graft detected during postoperative surveillance. Segmental stenotic lesions of the vein
conduit and distal anastomosis caused by fibrointimal hyperplasia were treated with resection and
interposition translocated vein graft. Extensive disease of the outflow artery caused by progression of
atherosclerotic disease was treated with sequential jump grafting to the more distal outflow artery site.

XI. PERIOPERATIVE RESULTS OF MANAGEMENT OF THE PATENT
BUT FAILING AUTOGENOUS GRAFT

The early patency (30 days) of the autogenous graft is excellent if intraoperative normal
hemodynamics and arteriography are achieved with the revascularization procedure (20). In
a study of in situ saphenous vein bypasses, the 30-day patency for the 83 bypasses in the
series with normal intraoperative hemodynamics was 100%. Initially, 77 of the 83 in situ
saphenous vein grafts had normal intraoperative hemodynamics with peak systolic velocity
of greater than 40 cm/s and biphasic (hyperemic flow) waveform. Six (7%) of the bypasses
had low peak systolic blood flow velocity (less than 40 cm/s) and an absence of hyperemic
flow, a predictor of early failure of the in situ bypass (29). Correction of the underlying
causes (AVF, intact valve leaflet, poor-quality vein segment) resulted in early bypass patency
with normal hemodynamics after modification in these six bypasses. The short-term patency
was dependent on intraoperative identification and correction of the lesions causing flow

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

FAILING AUTOGENOUS GRAFTS 349

abnormalities (28). The determination of normal intraoperative hemodynamics of the
autogenous graft is a highly accurate predictor of early patency following lower extremity
revascularization. The primary and secondary patencies for both in situ and reversed
autogenous grafts are decreased when intraoperative modification of the vein conduit is
required to complete the lower extremity arterial reconstruction (1,19).

XII. FACTORS THAT AFFECT THE LONG-TERM PATENCY RATE
OF AUTOGENOUS GRAFTS

The use of postoperative surveillance protocol of in situ and reversed autogenous grafts
allows the study of the pathophysiology of autogenous graft failure and also the analysis of
factors that adversely affect long-term patency. The determinant factors that significantly
affect long-term patency for both the in situ and reversed saphenous vein grafts are
modification of the venous conduit at the initial operation to correct a poor-quality vein
with patch angioplasty or vein interposition, technical error caused by injury to the vein
during valve ablation, failure to revise the bypass before thrombosis during the surveillance
period, and experience of the vascular surgeon in placing the autogenous graft. The long-
term patency of autogenous grafts is decreased significantly when the ipsilateral saphenous
vein is not adequate and requires some modification. In a recent large series of reversed
saphenous vein grafts, Taylor and colleagues (19) found that the ipsilateral greater
saphenous vein was inadequate in 45% of the cases. The ipsilateral saphenous vein was
inadequate because the vein was removed previously, the available vein was to short, the
available vein was to small or the vein contained sclerotic segments. Techniques practiced in
performing the reversed vein graft in patients with inadequate ipsilateral saphenous vein
were to use a more distal inflow artery, performed a venovenous anastomosis after removal
of the abnormal segment, or use another source of vein — contralateral greater saphenous
vein, cephalic vein, or lesser saphenous vein. The primary patency of the reversed vein grafts
that had inadequate ipsilateral saphenous vein was 68%, a significant decrease compared
with the 80% primary patency of those grafts that had adequate ipsilateral saphenous veins.
The secondary patency of the grafts with inadequate veins requiring a modification
technique was 77%, which also was decreased compared with the 84% patency of the
reversed vein grafts with nonmodified conduits. Similarly, in a recent large series of in situ
saphenous vein bypasses (1), the primary and secondary patency rates also were decreased
significantly for those bypassed veins undergoing modification at the time of bypass. The
need to modify the vein conduit because of a technical failure or inadequate vein was
associated with a significant increase in the incidence of late-appearing bypass stenosis and
revision. The bypass stenoses were most commonly caused by the occurrence and pro-
gression of fibrointimal hyperplasia resulting from the injurious effects of the bypass ■a
modification (Fig. 4). The primary (50%) and secondary (72%) patencies of the in situ |
saphenous vein bypasses that required modification were decreased significantly compared as
with the primary (70%) and secondary (84%) patency rates of bypasses that did not undergo c
modification at the time of the bypass procedure. This increase in incidence of graft stenoses, <
bypass revision, and decline in primary and secondary patencies was believed to be caused >9
by the occurrence of fibrointimal hyperplasia associated with the injurious effects of the J
bypass modification. The long-term patency of these conduits is dependent on the °
meticulous postoperative surveillance protocol identifying correctable lesions before throm- |
bosis, permitting elective revision of patent conduits. Reversed and in situ autogenous grafts @

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

350

TOWNE

14%

Modified Conduit

p < 0.003
p < 0.009

vs. no modification

No Modification

□ No Revision

mother Bypass Revisions

■ Revision Conduit Stenosis

Figure 4 Modification to correct an inadequate vein or technical failure during the in situ saphenous
vein bypass procedure was associated with a significant increase in number of bypasses in which graft
stenoses developed and that required revision during the postoperative surveillance period. Late-
appearing stenoses occurred in 12 of 86 modified conduits (14%) compared with only 12 of 269 in situ
bypasses (4.5%) not undergoing modification at the time initial modification. During the post-
operative surveillance period, revision of the bypasses with modified conduits (34 of 86, 40%)
was significantly more frequent than that of bypasses with no modification of conduits (60 of 269,
22%).

that have modified conduits mandate close postoperative surveillance to identify the
development of lesions that threaten graft patency.

For the in situ saphenous vein bypasses, severe atherosclerotic disease of the common
femoral artery was associated with a decline in primary patency but did not significantly
alter secondary patency (1). Severe disease of the common femoral artery necessitated
endarterectomy, replacement with an interposition prosthetic or vein graft, or closure with a
patch angioplasty in order to perform the proximal anastomosis. Bypasses originating from
a reconstructed inflow artery had a significant increase in the number of revisions performed
to correct anastomotic or outflow artery stenosis compared to bypasses with no recon-
struction of the inflow artery (Fig. 5). With postoperative surveillance and elective revision,
the secondary patency was not decreased significantly for the bypasses originating from a
reconstructed inflow artery.

Long-term patency of the failing autogenous graft can be maintained if revision is
performed before graft thrombosis occurs. Attempts to treat focal short-segment stenoses of
the autogenous vein conduit by total excision of the short segment of diseased vein and
primary end-to-end reanastomosis were highly successful, with no occurrence of late re-
stenosis or bypass failure. The treatment of graft stenosis with vein patch angioplasty also
resulted in excellent long-term patency (greater than 80%) (7,17) but was associated with an
increased incidence of late restenosis at the revision site itself. In a recent study Bandyk et al.
(7) demonstrated that 8 of 31 stenotic lesions (24%) treated by vein patch angioplasty re-
sulted in restenosis; most occurred more than 3 months after the secondary procedure. PTA
of focal vein graft stenosis also was associated with early bypass patency (8); however,
recurrent stenosis did occur in at least half of the cases (6,7,30,31). Because of the high
incidence of recurrent graft stenosis, PTA should be used primarily for high-risk patients
with a failing autogenous graft caused by focal lesion or for the management of athero-
sclerotic lesions in the native arteries. Restenosis of vein patch angioplasty or PTA site was

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

FAILING AUTOGENOUS GRAFTS

351

16%

Reconstructed Inflow Artery

n.s.

p < 0.05

vs. no reconstruction

21%

No Reconstruction

□ No Revision
1 Other Bypass Revisions
■ Revision Outflow Artery/
Anastomotic stenosis

Figure 5 In situ saphenous vein bypass revision performed to correct anastomotic or outflow artery
stenosis was significantly more frequent for grafts with reconstructed inflow arteries (7 of 44, 16%)
than for bypasses originating from nonreconstructed inflow arteries (14 of 312, 4.5%). Revision of
graft stenosis was not significantly different for bypasses with reconstruction of the inflow arteries (15
of 44, 34%) and those without (78 of 310, 25%).

best treated by resection of the diseased, stenotic segment of vein and replacement with a
translocated interposition vein graft.

Secondary procedures used to treat extensive lesions of the graft, distal anastomosis, or
outflow artery also result in maintenance of long-term patency. In a recent study by Bandyk
et al. (7), recurrent stenosis and graft failure occurred less frequently after resection and
interposition grafting compared with a sequential or jump graft procedure. Only 1 of 17 of
the interposition grafts had a recurrent stenosis or jump graft failure. However, 11 of 21
sequential or jump grafting procedures were associated with recurrent stenosis or autoge-
nous graft failure. The difference in patency was partially explained by the fact that the
interposition grafts were placed for flbrointimal hyperplasia of a graft that was resectable
and reconstructible, but the sequential or jump grafts were placed primarily for progression
of atherosclerotic disease of the outflow artery. The bypass of the failing graft caused
by progression of atherosclerotic disease of the outflow arteries with sequential or jump
grafting was the least durable procedure in 5 of 21 grafts that eventually failed.

Graft revision uniformly resulted in the restoration of normal hemodynamics of the
autogenous graft. The ABIs and the peak systolic velocities returned to comparable levels
that were present for the graft before the occurrence of the graft stenosis. The ABI was
usually in the range of 0.6 before the secondary procedure and increased to greater than 0.9
following revision in 85% of the lower extremities. The peak systolic velocity was less than
45cm/s in 845 of the autogenous grafts in this series and uniformly increased to a level
comparable to the initial postoperative studies (5,7).

I
•5

XIII. CONCLUSION

Long-term patency of the autogenous vein graft hinges on the detection and correction of
the failing bypass before thrombosis. The development of flbrointimal hyperplasia of the
graft or progression of atherosclerosis of the native arteries can result in diameter-reducing

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

352 TOWNE

stenoses that threaten bypass patency. The need for intraoperative reconstruction of the
inflow artery or for modification of the vein conduit during a bypass procedure because of a
technical error or poor-quality vein significantly increases the risk of late-appearing stenoses
and the need for bypass revision. The long-term patency of the autogenous vein grafts that
develop hemodynamically significant lesions (low graft flow velocity, >75% stenosis) and
threaten bypass patency is dependent on the identification and correction of the lesions
before thrombosis. The secondary patency rates for the revision of the patent but
hemodynamically failing autogenous graft have been equal to those of bypasses that do
not undergo revisions of the thrombosed autogenous graft has been uniformly poor (1,5).
The success of graft revision is dependent on excision of the lesion, use of autogenous tissue
for reconstruction, restoration of normal hemodynamics, and continued postoperative
surveillance. Secondary procedures to cored the failing graft result in excellent long-term
patency greater than 85% at 5 years (7,12). Secondary procedures should restore normal
graft and hemodynamics with the excision of the pathoanatomical lesion and the use of
normal autogenous tissue for reconstruction.

REFERENCES

1. Bergamini TM, Towne JB, Bandy DF, et al. Experience with in situ saphenous vein bypasses
from 1981-1989: determinant factors of long-term patency. J Vase Surg 1991; 13:137-147.

2. Minh-Chau L, Friedman EI, Figg-Hoblyn L, et al. Decreased graft flow is a reliable predictor of
impending failure of reversed vein grafts. J Vase Technol 1988; 12:133.

3. Sladen JG, Gilmour JL. Vein graft stenosis. Am J Surg 1981; 141:549.

4. Bandyk DF, Kaebnick HW, Stewart GW, et al. Durability of the in situ vein arterial bypass: a
comparison of primary and secondary patency. J Vase Surg 1987; 5:526.

5. Belkin M, Donaldson MC, Whittemore AD, et al. Observations on the use of thrombolytic
agents fro thrombotic occlusion of infrainguinal vein grafts. J Vase Surg 1990; 11:299.

6. Cohen JR, Mannick JA, Couch NP, et al. Recognition and management of impending vein-graft
failure. Arch Surg 1986; 121:758.

7. Bandyk DF, Bergamini TM, Towne JB, et al. Durability of vein graft revision: the outcome of
secondary procedures. J Vase Surg 1991; 13:200-210.

8. Berkowitz HD, Hobbs CL, Roberts B, et al. Value of routine vascular laboratory studies to
identify vein graft stenosis. Surgery 1081; 90:971.

9. Veith FJ, Weiser RK, Gupta SK, et al. Diagnosis and management of failing lower extremity
arterial reconstructions prior to graft occlusion. J Cardiovasc Surg 1984; 25:381.

10. Bandyk DF, Schmitt DD, Seabrook GR, et al. Monitoring patency of in situ saphenous vein
bypasses: the impact of a surveillance protocol and elective revision. J Vase Surg 1989; 9:286.

1 1. Erickson CA, Towne JB, Seabrook GR, et al. Ongoing vascular laboratory surveillance is es-
sential to maximize long-term in situ saphenous vein bypass patency. J Vase Surg 1996; 23:18-27.

12. Reifsnyder T, Towne JB, et al. Biologic characteristics of long-term autogenous vein grafts — A

dynamic evaluation. J Vase Surg 1993; 71:207-217. |

13. Green RM, McNamara J, Ouriel K\, et al. Comparison of infrainguinal graft surveillance tech- 8
niques. J Vase Surg 1990; 11:207. 1

14. Cato R, Bandyk DF, et al. Duplex scanning after carotid reconstruction: A comparison of ■$
interoperative and postoperative results. J Vase Tech 1991; 15:61-65. §

15. Mills JL, Bandyk DF, Gahton V, Esses GE. The origin of infrainguinal vein graft stenosis: A ^
prospective study based on duplex surveillance. J Vase Surg 1995; 21:16-25. jj

16. Passman MA, Monetta GL, Naylor ME, Taylor LM, et al. Do normal early color flow duplex «
surveillance examination results of infrainguinal vein graft preclude the need for late graft ■§
revision. J Vase Surg 1995; 22:476-484. J

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

FAILING AUTOGENOUS GRAFTS 353

17. Whittemore AD, Clowes AW, Couch NP, et al. Secondary femoropopliteal reconstruction. Ann
Surg 1981; 193:35.

18. Leather RP, Shah DM, Chang BB, et al. Resurrection of the in situ saphenous vein bypass 1000
cases later. Ann Surg 1988, 208-435.

19. Taylor LM, Edwards JM, Porter JM. Present status of reversed vein bypass grafting: five-year
results of a modern series. J Vase Surg 1990; 11:193.

20. Szilagyi DE, Elliott JB, et al. Biologic fate of autogenous vein implants as arterial substitutes.
Ann Surg 1973; 178:775-784.

21. Atkinson JB, Forman MB, et al. Morphologic changes in long-term saphenous vein bypass
grafts. Chest 1985; 88:341-348.

22. DeWeise JA, Rob CG. Autogenous venous grafts ten years later. Surgery 1978; 82:775-784.

23. Rutherford RB, Jones DN, Bergentz SE, et al. Factors affecting the patency of infrainguinal
bypass. J Vase Surg 1988; 8:236.

24. Wengerter KR, Veith FJ, Gupta SK, et al. Influence of vein size (diameter) on infrapopliteal
reversed vein graft patency. J Vase Surg 1990; 1 1:525.

25. Cutler BS, Thompson JE, Kleinsasser JG, et al. Autogenous saphenous vein femoropopliteal
bypass: analysis of 298 cases. Surgery 1976; 79:325.

26. Bernhard VM, Boren CH, Towne JB. Pneumatic tourniquet as a substitute for vascular clamps in
distal bypass surgery. Surgery 1980; 87:709.

27. BandykDF. Postoperative surveillance of infrainguinal bypass. Surg Clin North Am 1990; 70:71.

28. Schmitt DD, Seabrook GR, Bandyk DF, et al. Early patency of in situ saphenous vein bypasses
as determined by intraoperative velocity waveform analysis. Ann Vase Surg 1990; 4:270.

29. Bandyk DF, Kaebnick HW, Bergamini TM, et al. Hemodynamics of in situ saphenous vein
arterial bypass. Arch Surg 1988; 123:477.

30. Greenspan B, Pillari G, Schulman ML, et al. Percutaneous transluminal angioplasty of stenotic,
deep vein arterial bypass grafts. Arch Surg 1985; 120:492.

31. Sheridan J, Thompson J, Gazzard S, et al. The role of transluminal angioplasty in the manage-
ment of femoro-distal graft stenosis. Br J Radiol Suppl 1989; 62:564.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

19

An Approach to Treatment of Infrainguinal
Graft Occlusions

Lloyd M. Taylor, Jr., Gregory J. Landry, and Gregory L. Moneta

Oregon Health & Science University, Portland, Oregon, U.S.A.

Management of patients with infrainguinal bypass graft occlusions is an inevitable part of
all vascular surgery practices. Included within this category are a wide-ranging group of
clinical scenarios. Graft occlusions may occur in the immediate postoperative period or
years later. The occlusion may result in acute, severe limb-threatening ischemia, or be com-
pletely asymptomatic and discovered by accident at the time of examinations for other
causes. Despite their fabled objectivity, most surgeons have considerable emotional/ego
involvement in their work — their natural response to graft occlusions is nearly always to
try to restore patency to the existing system by the most efficient method possible. In the
opinion of the authors, this is rarely if ever the correct response.

This chapter describes the approach to graft occlusions followed by the vascular surgery
service at Oregon Health & Sciences University. Due to the referral nature of this practice,
we have treated more than 20 cases of infrainguinal graft occlusions per year for the past
17 years. This large clinical volume combined with the experience of others reported in the
literature has allowed the development of an approach to graft occlusions that, in the opin-
ion of the authors, results in maximal patient survival, limb salvage, and long-term graft
patency. The following discussion is divided into (a) treatment of acute postoperative oc-
clusions and (b) treatment of those that occur after the immediate postoperative period,
because the management of these two clinical scenarios differs somewhat. Before discussing |

the treatment of graft occlusions, a few remarks regarding prevention are in order. js

£

■c

I. PREVENTION OF POSTOPERATIVE GRAFT OCCLUSION "

I
A. The Decision to Operate <3

A very large percentage of the patients presenting to our service for treatment of infrain- |

guinal bypass graft occlusions were originally operated on for claudication. For fortunate @

355 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

356 TAYLOR et al.

patients, bypass graft occlusion results in a return to the original symptomatic state, but
this is not always the case. A disturbingly large number of patients with graft occlusions
develop more severe ischemia than was present at the time of the original operation. Thus
patients who had a very low likelihood of ever developing limb-threatening ischemia from
their original disease process are suddenly placed into this category as a complication of
treatment, which, in retrospect, they (and their surgeons) naturally wish they had never
undertaken. This disturbing and dangerous tendency is especially true when the original
bypass graft was placed using prosthetic material (1).

It is not overstating the case to say that the best way to avoid progression to limb-
threatening ischemia in a patient with claudication is to not operate for the claudication,
especially using a prosthetic graft. Surgery for claudication should be approached very
cautiously and only when fully informed patients clearly understand that the most signi-
ficant risk to their limb is probably from the treatment and not from the disease.

B. Arteriography

It is axiomatic that successful bypass grafting requires unobstructed arterial inflow and a
site of distal anastomosis to a distal vessel of the best possible quality with the best possible
unobstructed outflow bed. The sites of proximal and distal anastomosis are best chosen
using high-quality arteriography. In our practice, more than half of the patients referred
for bypass grafting with previously obtained arteriograms undergo repeat arteriography
prior to surgery so as to better delineate anastomotic sites. Intraoperative "exploration" is
unreliable for choosing anastomotic sites. Widely patent vessels that are perfectly suitable
for anastomosis may be deceptively calcified and needlessly rejected; pliable, apparently
suitable arteries may be severely diseased intraluminally.

C. Conduit

The most effective means of preventing postoperative bypass graft occlusion is to
construct the bypass conduit from autogenous vein. Intact, good-quality greater saphe-
nous vein is the best available conduit, but lesser saphenous, arm, and deep leg veins are all
satisfactory and are all superior to prosthetic grafts, even when multiple segments must be
anastomosed together to form conduits of adequate length. Two techniques are of great
assistance in maximizing the number of grafts that can be placed using autogenous vein.
The first is the use of duplex scan vein mapping to identify which autogenous veins are the
most likely adequate conduits. The second is using multiple operative teams to facilitate
the multiple steps required in complex "redo" bypass surgery. A single operative team of
surgeon and an assistant (faculty attending and resident in our practice) can nearly always
complete a first-time tibial bypass using intact ipsilateral greater saphenous vein within 3-
4 h. Make the operation a redo, with a need to harvest three segments of vein from both g

arms to create a conduit of adequate length, and the time required may exceed twice |

that — really not an acceptable length operation for an elderly patient with multisystem s

comorbidities. In this dilemma lies the origin of many a prosthetic graft. On the other jS

hand, two or three operating teams working simultaneously can easily complete such d

complex operations within the same time required for first-time surgery. In our opinion, g

the advantages of using autogenous conduit are sufficiently great that surgeons who are |j

unable to muster the necessary manpower for multiple operating teams should seriously s

consider referring complex redo cases to medical center services that can. M

Q

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

INFRAINGUINAL GRAFT OCCLUSIONS 357

D. Confirmation of Technical Success

Adequacy of inflow, conduit, proximal and distal anastomoses, and outflow vessels should
be confirmed by objective means prior to closing wounds and/or leaving the operating
room. At a minimum, improved ankle continuous-wave Doppler signals that respond
appropriately to temporary graft occlusion and release should be confirmed. Intraoper-
ative duplex scanning or operative completion arteriography are more cumbersome but
provide more detailed and anatomical information. Abnormalities should be explained
and corrected before leaving the operating room.

E. Pharmacological Management

All patients with atherosclerotic disease should be on regular aspirin therapy or its equiv-
alent. This should be continued perioperatively. The authors add perioperative heparin
anticoagulation and postoperative warfarin for patients with documented hypercoagulable
disorders. The most common of these is the presence of anticardiolipin antibodies, which
may be found in as many as one-third of patients requiring redo bypass surgery (2). Peri-
operative heparin therapy results in an increased incidence of postoperative wound he-
matomas requiring reoperation for drainage. This is a reasonable exchange for improved
graft patency.

F. Postoperative Graft Surveillance

Stenotic lesions develop in 20-30% of autogenous vein grafts, the majority within the first
postoperative year (3). These lesions can be reliably detected by duplex scanning. If un-
corrected, at least 50% of such stenoses will result in graft occlusion within 12 months of
their detection (4). Vein grafts in which stenoses are detected and corrected have excellent
long-term patency (5). The authors obtain duplex scans of autogenous vein grafts every 3
months for the first postoperative year and every 6 months thereafter. The majority of
detected lesions are repaired surgically; a few with especially favorable characteristics can
be treated by balloon angioplasty. Duplex scanning has not been shown to be useful in
preventing stenoses of prosthetic infrainguinal grafts.

G. Other Factors

A recent review from our service identified continued cigarette smoking and poor com-
pliance with postoperative graft surveillance as the most important modifiable risk factors
associated with graft occlusion (6). The message seems clear. Patients need support and
counseling, especially about the importance of smoking cessation, and this must be re-
peated and reinforced at every visit. The importance of the graft surveillance studies must
be emphasized clearly. Patients who fail to keep appointments should be contacted; the
risk incurred by missing surveillance examinations must be explained.

II. TREATMENT OF ACUTE POSTOPERATIVE GRAFT OCCLUSIONS

For the purpose of this chapter, acute postoperative graft occlusions are those that occur
prior to the patient's discharge from the hospital during which the bypass procedure was
performed. During this interval, occlusions are usually detected promptly and can be
treated immediately with a reasonable expectation that patency can be restored and that

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

358 TAYLOR et al.

long-term patency will be acceptable, which is almost never true once the patient has
been discharged.

A. Initial Management

Acute postoperative graft occlusions result in the return of ankle brachial pressure indices
(ABI) to preoperative levels or below and recurrence of preoperative ischemic symptoms.
If the indication for the bypass was claudication, there may be no symptoms in a bed-
confined hospital patient. Any decrease in ABI from immediate postoperative values must
be explained. In some patients, emergency duplex scanning or arteriography may be neces-
sary to determine whether grafts are occluded or patent, with another explanation (prox-
imal stenosis, graft stenosis, runoff occlusion, etc.) for the reduced ABI. Once diagnosed,
the most appropriate response to immediate postoperative graft occlusions is full heparin-
ization followed by an immediate return to the operating room. Of course there may be
compelling reasons not to follow this course. Patients' conditions may change markedly
postoperatively. Myocardial infarction, pneumonia, or other acute conditions may pre-
clude early reoperation, and other conditions — such as gastrointestinal bleeding — may
preclude anticoagulation. If immediate reoperation is contraindicated, it is extremely un-
likely that the original bypass conduit can be salvaged. This may be a reasonable and ap-
propriate price for delay when dictated by patient condition.

B. Conduct of the Operation

In addition to the operated extremity, the surgical field should include a source of additional
vein conduit sufficient to replace or to extend the original graft. The patient should be
placed on an operating table that will accommodate fluoroscopy of the extremity's entire
arterial tree, from the aortic bifurcation to the toes. Full heparin anticoagulation should
be maintained until the cause of the graft occlusion has been determined and corrected. It
is helpful to monitor the dose of heparin using the activated clotting time (ACT).

The first step is to open the incisions over the proximal and distal anastomoses and
determine the cause of the occlusion. A normal pulse in the inflow artery/proximal graft
rules out inflow obstruction as the cause. Liquid blood in the hood of the distal anas-
tomosis similarly rules out distal obstruction. Hard thrombus in either location points to a
cause at the site where it is found. In the absence of a problem with the proximal or distal
arteries/anastomoses, a problem with the vein graft must be assumed.

Catheter thrombectomy is difficult in the immediate postoperative period. The venous
valves, whether lysed for in situ grafts or reversed, trap adjacent thrombus, which is highly
resistant to removal by balloon catheter and can serve as the nidus for further thrombosis.
The authors prefer to open and/or detach the distal vein graft anastomosis. If the occlu-
sion is very fresh, the thrombus is often ejected from the graft simply by the incoming ■g
arterial pressure. Intraoperative arteriography can then be performed to locate the site of &
the obstruction that produced the occlusion. If the graft remains obstructed after the distal a
anastomosis is detached, it is best to extract it from the tunnel and express the remaining c
thrombus by gentle digital pressure. The graft can then be inspected for sclerotic/stenotic <j
segments. If none are discovered and this is confirmed by intraoperative arteriography, >9
extrinsic compression in the tunnel can be assumed to be the cause of the occlusion. Re- 4j
tunneling and reanastomosis of the distal end of the graft will solve the problem. 2

Sclerotic/stenotic segments of vein must be excised and replaced. Unsuspected or un- |

detected proximal or distal occlusive disease must be repaired or bypassed by graft exten- @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

INFRAINGUINAL GRAFT OCCLUSIONS 359

sion. Technically unsatisfactory anastomoses should never be the cause of graft occlusion;
they should have been detected by the measures used to ensure technical success at the
time of the first operation.

Regardless of the cause of the graft occlusion and the method chosen for its correc-
tion, operative completion arteriography should conclude operations performed to correct
acute graft occlusions. Once the final reconstruction has been proven to be technically
perfect, the authors prefer to continue heparin anticoagulation for several days postoper-
atively to prevent early rethrombosis due to the highly thrombogenic inner surface of the
graft from which thrombus has been removed. An increased incidence of wound hema-
tomas requiring drainage is an inescapable result of this approach.

C. Rethrombosis

Bypass grafts that reocclude after the operative steps described above have been taken will
not remain patent after another operation in which the thrombus is reremoved (1). If a
different operation (different anastomotic sites, new conduit) is possible, it is acceptable to
proceed with this taking the patient's condition into account. If not, there is little to be
gained from repeated and increasingly futile attempts to make a flawed system work.

III. TREATMENT OF LATE GRAFT OCCLUSIONS

Four courses of action are possible in response to graft occlusions that occur following hos-
pital discharge. These include no treatment, percutaneous endovascular treatment (lytic
therapy with correction of stenoses by angioplasty and/or stenting), operative graft throm-
bectomy with correction of stenoses, and reoperation with a new graft. The factors that
govern decision making among these options include the severity of ischemia that exists
following graft occlusion, the patient's need for a patent graft, and the likelihood that the
planned intervention will remain patent. Three easily quantitated objective factors are rel-
evant to the long term outcome of graft occlusion treatment. These are patient survival, limb
salvage, and graft patency. Other factors are also very important, but some are subjective
and inherently difficult to quantitate and some are important only transiently. These include
patient inconvenience, pain, physical impairment, and long-term functional status.

A. Severity of Ischemia

Bypass grafts performed for mild to moderate claudication may become occluded and not
produce any symptoms in sedentary patients, particularly if years have passed and the
patient's activity level, for which the surgery was originally performed, has declined. Most

grafts performed for severe claudication and/or limb-threatening ischemia produce clear- ■g

cut ischemic symptoms when occlusion occurs. Initially the degree of ischemia that is &

present may be severe, with no detectable circulation and neuromuscular impairment in a

the distal extremity. These prominent symptoms may lead to a mistaken impression that c

the limb is acutely threatened and that a true emergency exists, in which case revascula- <j

rization would have to be accomplished within a few hours to avoid amputation. Ex- >9

perienced vascular surgeons recognize that this is rarely the case, as first demonstrated by 4j

Blaisdell (7). For most patients with acute graft occlusions, the initial severe symptoms 2

improve rapidly and adequate neuromuscular function returns, allowing for a deliberate, |

elective approach to treatment. The authors treat acute bypass graft occlusions with hospi- ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

360 TAYLOR et al.

talization, bed rest, and anticoagulation with intravenous unfractionated heparin. Only
when ischemic symptoms fail to respond to these measures is urgent or emergent revascu-
larization considered. Thankfully such patients are rare.

B. No Treatment

When the initial indication for bypass grafting was claudication, no treatment is frequently
an option for the management of graft occlusion. Patients may not wish to have further
invasive treatment for claudication symptoms that have lessened in importance since the
time of the bypass. No treatment is a particularly attractive option if the initial operation
was performed using prosthetic conduit and sources of autogenous vein are limited. Some
patients with diabetes may have required bypass grafting for relatively mild occlusive dis-
ease in order to assist with the healing of neuropathic/infectious ulcers. Once the ulcers
have healed, graft occlusion may be well tolerated. Obviously, decisions for no treatment
must be individualized. It is reasonable to assume that the ischemic symptoms existing at
the time of graft occlusion are the most severe that will occur. Some spontaneous im-
provement can be anticipated in nearly all patients. It is reasonable to observe patients for
this improvement and to postpone the decision about repeat revascularization if the
absolute need for revascularization is not initially clear.

C. Percutaneous Treatment

Infusion of thrombolytic agents into thrombosed bypass grafts frequently results in res-
toration of patency and resolution of ischemic symptoms. This fact is extraordinarily se-
ductive. The bypass graft is thrombosed, and the leg is ischemic; lytic therapy is applied,
the graft is patent, and the leg is no longer ischemic. Patients are much relieved, emergency
surgery is avoided, and — perhaps most seductive — the surgeon's ego is assuaged. What
was lost has been regained and the graft is once again patent. Add to this the potential to
reveal by lysis the stenosis that led to the thrombosis and to correct this by percutaneous
means — for example angioplasty and stenting — and there is a possibility that this clinical
catastrophe can be efficiently and effectively managed by a single trip to the interventional
suite, with hospitalization required at most overnight if at all.

Since thrombolytic treatment for thrombosed infrainguinal grafts was first described
in 1981, a mass of evidence has accumulated that this ideal scenario rarely if ever occurs.
There have been many advances in thrombolytic therapy: better drugs, improved tech-
nology, more rational dosing, etc. Each has resulted in increases in the percentage of
patients in whom lysis can be accomplished and in the speed and safety with which it can
be done. Despite these advances, lytic therapy remains dangerous. No large series is free of
occasional deaths from intracerebral hemorrhage, and less lethal bleeding complications
remain common. The main problem with the lytic approach to graft occlusion, however, •§

has to do with disappointing long-term patency. Multiple series accumulated over the past g

two decades, examples of which are summarized in Table 1 , indicate that fewer than half a

of the grafts to which patency has been restored by lysis remain patent for as long as 1 c

year, with or without adjunctive correction of underlying stenoses. A single prospective <j

randomized clinical trial (the STILE study) compared lytic therapy to best surgery for >9

infrainguinal graft occlusion. This study was stopped because of superiority of the surgical 4j

group, even though many operations were thrombectomies, or far from ideal surgery (8). 2

Proponents of lytic therapy acknowledge this flaw but point out that at least lysis |

relieves the acute ischemia, so that more definitive elective treatments can be carried out in @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

INFRAINGUINAL GRAFT OCCLUSIONS

361

Table 1 Results of Lytic Therapy for Infrainguinal Bypass Graft Occlusion

Author and year (ref)

Graor, 1985 (13)
Graor, 1988 (14)
Gardiner, 1989 (15)
Belkin, 1990 (16)
Faggioli, 1994 (17)
Comerota, 1996 (8)

a Long-term patency = patency by life table at 1 year postthrombolysis.

b Patency of grafts in which stenosis was not corrected/patency in those with corrected stenosis.

c Mortality at 1 year postlysis.

n

Initial success

;%)

Mortality

Long-term patei

24

15 (62%)

not given

33

29 (88%)

20%

72

50 (69%)

1.4%

18%/75% b

35

22 (60%)

37%

51

23 (45%)

1.9%

39%

78

37 (47%)

6% c

27%

appropriate patients. If the acute ischemia of graft occlusion were truly immediately limb-
threatening, this would be an advantage indeed. But in fact this is rarely the case. Nearly
all patients with graft occlusion can be managed by hospitalization, bed rest, antico-
agulation, elective arteriography, and revascularization without incurring the considerable
expense and risk of an initial episode of lytic therapy.

The authors reserve lytic therapy for graft occlusions that occur in patients known to
have no further possibilities for reconstruction or for those with documented hyper-
coagulable states who have previously had graft thromboses in the absence of stenosis. In
actual practice, such patients are rare.

D. Graft Thrombectomy

Effective thrombectomy of vein grafts using balloon catheters is very difficult. The reasons
include the varying caliber of the vein, the fact that most autogenous graft thromboses
result from fixed stenoses, and the tendency of thrombus to adhere very tightly to the
inflamed intimal surface. In contrast, thrombectomy of prosthetic grafts is usually easily
accomplished. Unfortunately numerous studies — examples of which are listed in Table 2 —
have shown that thrombectomy is rarely if ever followed by durable long-term patency.
Even when the culprit stenoses are discovered and corrected at the time of thrombectomy,
all studies have shown disappointing patency. Despite these facts, the temptation to re-
store patency to the existing system through the use of thrombectomy with revision when
appropriate is strong. It is best resisted. The authors do not use thrombectomy to treat
infrainguinal graft thrombosis that occurs after the immediate postoperative period.
Available information indicates that patient survival, limb salvage, and long-term recon-

Table 2 Results of Thrombectomy for Infrainguinal Bypass Graft Occlusions

Author and year (ref)

n

Initial success

Mortality Long-term patency"

Ascer, 1987 (18)
Graor, 1988 (13)
Lombardi, 2000 (19)

128
38
21

Not given
42%
Not given

5.6% 20-40% b
5% 18%
3% 47%

i

2

I

■c

a Long-term patency = patency at 1 year.

b Difference in patency depending on site of distal anastomosis.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

362 TAYLOR et al.

struction patency are best maximized by elective reoperation with a new autogenous vein
graft.

E. Elective Reoperation with a New Autogeneous Vein Graft

Patients with infrainguinal graft thrombosis have varying degrees of ischemia, some of
which is acute. The authors decide whether emergency hospitalization is needed based
upon three patient factors: the presence of ischemic rest pain, absent ankle Doppler sig-
nals, and neuromuscular dysfunction. Any of these is an indication for immediate hospi-
talization, bed rest, and heparin anticoagulation. Patients with lesser degrees of ischemia
can be scheduled for elective hospitalization.

Once hospitalized and anticoagulated, most patients find that their ischemic symptoms
improve rapidly. Return of adible ankle Doppler signals within a day or two is common.
During this time, patients can be carefully assessed for surgery and medically evaluated and
stabilized. Duplex scan vein mapping can be performed to accurately delineate available
sources of autogenous conduit.

The authors prefer to delay arteriography for an interval of at least 2-3 days from the
time of the acute graft occlusion. This allows initial ischemic vasospasm to resolve and
full collateral development. Arteriograms performed immediately after acute occlusions are
frequently noninformative; those performed after an appropriate interval are much more
likely to reveal satisfactory distal bypass targets.

Once the arteriogram has been obtained, it is possible to plan in detailed fashion an
operation to revascularize the ischemic limb using autogenous vein. The need to use arm
vein conduits arises frequently, as does the need for anastomosis of multiple venous seg-
ments to create conduits of adequate length. The common femoral artery has frequently
been seriously compromised by a combination of disease and multiple previous surgeries.
Common femoral excision and interposition prosthetic grafting is an excellent and durable
solution for this problem (9).

These reoperative procedures are frequently extensive, involving multiple operative
sites and extremities and difficult redissections of previously operated areas. This type of
operation is ideally suited to a multiple operative team approach. Indeed, some of the
more extensive procedures cannot be accomplished within reasonable time limits without
multiple operating teams.

Table 3 Results of Treatment of Infrainguinal Graft Occlusions with Reoperation with New
Autogenous Vein Grafting

Author and Year (ret)

n

Initial Success

Mortality

Long-Term Patency"

Brewster, 1983 (11)

19

NA b

NA

63% c

1

Ascer, 1987 (18)

27

100%

NA

39-48% d

1

Edwards, 1990 (20)

103

98%

NA

62%/71% e

&

DeFrang, 1994 (10)

85

100%

3.5%

80%

•a

■c

Biancari, 2000 (21)

30

100%

6%

44%

§

a Long term patency = patency at 3 years.

b N A = not available.

c Five-year patency.

d Two-thirds of patients had prosthetic grafts.

= Primary/secondary patency at 3 years.

I

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

INFRAINGUINAL GRAFT OCCLUSIONS 363

As seen from the examples listed in Table 3, the results achieved in this difficult patient
population compare very favorably in terms of morbidity, mortality, and patency with
those reported for thrombolysis and/or thrombectomy (Tables 1 and 2). This is true even
in patients presenting after failure of more than two previous attempts at bypass grafting.
DeFrang and coauthors were able to achieve 80% primary patency and 70% limb salvage
at 3 years using repeat autogenous vein grafting in this highly selected patient group (10).

Other vascular surgery units with extensive experience in the management of occluded
infrainguinal graft have reached the same conclusion regarding optimal management.
Veith and coworkers from Montefiore in New York (11) and Brewster and colleagues
from Boston (12) have also found that reoperation with a new autogenous vein graft is the
treatment with the best outcome.

IV. CONCLUSIONS

Based upon the data summarized in Tables 1, 2, and 3 and a considerable personal prac-
tice experience, the authors have developed an approach to lower extremity graft occlusion
that first emphasizes prevention. Indications for surgery should be conservative. The vast
majority of claudicants are best managed nonoperatively. Surgery requires careful plan-
ning. Detailed arteriography is necessary. Autogenous vein is always the best conduit. The
technical result of surgery should be objectively confirmed in the operating room. Post-
operative duplex surveillance with correction of detected lesions maximizes patency.

Immediate postoperative occlusions are best managed by immediate reoperation, with
a systematic approach to the detection of the underlying cause and its correction. Intra-
operative imaging is essential for these procedures.

Despite the temptation to restore the patency to bypass grafts that occlude after the
immediate postoperative period, the long-term patency results achieved do not justify this
approach to the treatment of occlusions. The great majority of such patients are best man-
aged by hospitalization and heparin anticoagulation, followed by elective arteriography
and elective reoperation using a new autogenous vein bypass.

REFERENCES

1. Robinson KD, Sato DT, Gregory RT, Gayle RG, DeMasi RJ, Parent FN III, Wheeler JR.
Long-term outcome after early infrainguinal graft failure. J Vase Surg 1997; 26:425-438.

2. Taylor LM Jr, Chitwood RW, Dalman RL, Sexton G, Goodnight SH, Porter JM. Anti-
phospholipid antibodies in vascular surgery patients: a cross-sectional Study. Ann Surg 1994;
220:544-551.

3. Passman MA, Moneta GL, Nehler MR, Taylor LM Jr, Edwards JM, Yeager RA, McConnell

DB, Porter JM. Do normal early color flow duplex surveillance examinations of infrainguinal •o

vein grafts preclude the need for late graft revision? J Vase Surg 1995; 22:476-484. §

4. Mattos MA, Van Bemmelen PS, Hodgson KJ, Ramsey DE, Barkmeier LD, Sumner DS. Does 8
correction of stenoses identified with color duplex scanning improve infrainguinal graft pat- •§)
ency? J Vase Surg 1993; 17:54-66. |

5. Nehler MR, Moneta GL, Yeager RA, Edwards JM, Taylor LM Jr, Porter JM. Results of y
surgical revision of threatened reversed infrainguinal vein grafts. J Vase Surg 1994; 20:558-565. g

6. Giswold MA, Moneta GL, Landry GL, Yeager RA, Edwards JM, Taylor LM Jr. Modifiable 3
risk factors associated with infrainguinal vein graft occlusion. J Vase Surg 2003; 37:47-53. !3

7. Blaisdell FW, Steele M, Allen R. Management of acute lower extremity arterial ischemia due to I
embolism and thrombosis. Surgery 1978; 84:822-834. |

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

364 TAYLOR et al.

8. Comerota AJ, Weaver FA, Hosking JD, Froehlich J, Folander H, Sussman B, Rosenfield K.
Results of a prospective randomized trial of surgery versus thrombolysis for occluded lower
extremity bypass grafts. Am J Surg 1996; 172:105-112.

9. Nehler MR, Taylor LM Jr, Lee RW, Moneta GL, Porter JM. Interposition grafting for
reoperation on the common femoral artery. J Vase Surg 1998; 28:37-44.

10. DeFrang RD, Edward JM, Moneta GL, Yeager RA, Taylor LM Jr, Porter JM. Repeat leg
bypass following multiple prior bypass failures. J Vase Surg 1994; 19:258-278.

11. Veith FJ, Ascer E, Gupta SK, et al. Management of the occluded and failing PTFE graft. Acta
Chir Scand 1987; 538:117-124.

12. Brewster DC, LaSalle AJ, Robinson JG, et al. Femoropopliteal graft failures: Clinical con-
sequences and success of secondary procedures. Arch Surg 1983; 118:1043-1047.

13. Graor RA, Risius B, Denny KM, Young JR, Beven EG, Hertzer NR, Ruschhaupt WF III,
O'Hara PJ, Geisinger MA, Zelch MG. Local thrombolysis in the treatment of thrombosed
arteries, bypass grafts, and arteriovenous fistulas. J Vase Surg 1985; 2:406-414.

14. Graor RA, Risuis G, Young JR, Lucas FV, Beven EG, Hertzer NR, Krajewski LP, O'Hara PJ,
Olin J, Ruschhaupt WE. Thrombolysis of peripheral arterial bypass grafts: Surgical throm-
bectomy compared with thrombolysis. J Vase Surg 1988; 7:347-355.

15. Gardiner GA, Harrington DP, Koltun W, Whittemore A, Mannick J A, Levin DC. Salvage of
occluded arterial bypass grafts by means of thrombolysis. J Vase Surg 1989; 9:426-431.

16. Belkin M, Donaldson MC, Whittemore AD, Polak JF, Grassi CJ, Harrington DP, Mannick
JA. Observations on the use of thrombolytic agents for thrombotic occlusion of infrainguinal
vein grafts. J Vase Surg 1990; 11:289-296.

17. Faggioli GL, Peer RM, Pedrini L, Di Paola MD, Upson JA, D'Addato M, Ricotta JJ. Failure
of thrombolytic therapy to improve long-term vascular patency. J Vase Surg 1994; 19:289-297.

18. Ascer E, Collier P, Gupta SK, Veith FJ. Reoperation for polytetrafluoroethylene bypass fail-
ure: The importance of distal outflow site and operative technique in determining outcome. J
Vase Surg 1987; 5:298-310.

19. Lombardi JV, Dougherty MJ, Calligaro KD, Compbell FJ, Schindler N, Raviola C. Predictors
of outcome when reoperating for early infrainguinal bypass occlusion. Ann Vase Surg 2000;
14:350-355.

20. Edwards JM, Taylor LM Jr, Porter JM. Treatment of failed lower extremity bypass grafts with
new autogenous vein grafting. J Vase Surg 1990; 11:132-145.

21. Biancari F, Railo M, Lundin J, Alback A, Kantonen 1, Lehtola A, Lepantalo M. Redo bypass
surgery to the infrapopliteal arteries for critical leg ischaemia. Eur J Vase Endovasc Surg 2000;
21:137-142.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

20

Wound Complications Following Vascular
Reconstructive Surgery

David R. Lorelli

St. John Hospital and Medical Center, Detroit, Michigan, U.S.A.

Jonathan B. Towne

Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

I. INTRODUCTION

Wound healing problems are the most frustrating and dangerous possible complications
associated with vascular reconstructive surgery. Wound healing complications are related
to age, advanced atherosclerotic disease, tissue ischemia, lengthy incisions, and the multi-
ple comorbid medical conditions that routinely exist in the vascular patient population.
However, the significance of wound complications in patients undergoing vascular recon-
struction has received only minor emphasis in the literature. Fortunately, most wound
complications are superficial and are not limb-nor graft-threatening, causing only incon-
venience and additional expense to the patient. In the most severe cases, wound have the
potential to involve the underlying graft, either prosthetic or autogenous, which may often
lead to hemorrhage or graft occlusion with resultant loss of life or limb. This chapter reviews
the classification, incidence, and etiological factors responsible for wound complications in
different anatomical locations as well as methods of prevention and treatment, including
alternative operative techniques, that may have a favorable impact upon the frequency of e

wound complications following vascular reconstructive surgery. 2

I

■c

II. CLASSIFICATION SCHEMES y

Two classification schemes for wound complications have been proposed over the past %

30 years. The original clinical classification of infection was put forth by Szilagyi and 2

associates (1) in a review of over 3300 arterial reconstructions over a 20-year period. |

Wounds were classified into three grades in accordance with the depth of involvement. @

365 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

366 LORELLI and TOWNE

Grade I infections involved the dermis only. Infections extending into the subcutaneous
region without graft involvement were classified as grade II. The most serious infections,
or grade III, involved the arterial graft itself. This classification scheme had significant
practical implications as the management strategies were more complex and the risk of
limb or life loss significant with grade III infections.

A second scheme to classify wound complications following infrainguinal arterial
bypass grafting was proposed by Johnson et al. (2) in 1988. They further divided Szilagyi
grade II wound complications into two groups based on the presence or absence of cul-
ture-proven wound infection. Erythema or seroma formation without wound edge sep-
aration encompassed class 1 wounds. Class 2 wounds were those with ischemic necrosis
of the wound edge without infection. Class 3 wounds included ischemic necrosis of the
wound edge with overt infection. Open, infected wounds with exposed graft comprised
class 4 wounds.

Wound complications following vascular reconstruction and following saphenous
vein harvest for coronary artery bypass account for the vast majority of wound problems
faced by the vascular surgeon. Wound healing is a complex process. Many factors con-
tribute to this process and any alteration in one or more of these factors can influence the
success of primary wound healing.

III. ETIOLOGICAL FACTORS
A. Preoperative Factors

Wound healing is a complex sequence of events with a multitude of factors, which may
influence its ultimate outcome. Optimization of the preoperative factors affecting wound
healing provides the best foundation for a successful outcome. Probably no other condi-
tion has been investigated as much as diabetes in its relation to wound healing. Poorly
controlled diabetics may have defects in leukocyte chemotaxis and phagocytosis, which
reduces the ability to heal wounds and fight infection (3). This defect can be improved with
appropriate insulin therapy that maintains blood glucose at normal levels (4). Despite
some conflicting data, numerous studies have demonstrated diabetes to be a risk factor for
wound complications (5-11). Careful control of blood glucose may help to lessen the im-
pact of diabetes on wound healing.

Patients with chronic renal insufficiency or end-stage renal disease are at high risk for
wound healing difficulties (12,13). This is likely due to a significant association with
uremia, diabetes, anemia, malnutrition, and altered immune function, resulting in rates of
perioperative wound complications are as high as 29-54% (14-16). In patients with end-
stage renal disease undergoing infrainguinal arterial bypass, the amputation rate — despite
a functioning graft — due to wound complications or progressive necrosis is 14% (13).
Because of concern about this devastating problem, some have advocated anatomical 13

tunneling of all infrainguinal grafts in renal patients with the avoidance of an in situ §

bypass (12). _§

Treatment with anti-inflammatory steroids or other immunosuppressive agents can 2

hinder wound healing. This is likely mediated through a reduction in transforming growth ^

factor beta (TGF-fJ), which can be countered by administration of vitamin A (17). Anemia "5

may also play an important role. Taylor et al. studied the effects on wound healing in rats H

with compensated oligemia. They found that the effect of anemia on wound healing is °

more pronounced in skin than in muscle tissue and that this is due to a specific increase in J

collagen turnover, with the increase in reabsorption exceeding that in synthesis (18). In @

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

WOUND COMPLICATIONS 367

addition, malnutrition correlates with an increased incidence of wound complications. In a
study of 79 patients undergoing a variety of vascular surgical operations, Casey et al.
correlated the immune and nutritional status of patients with the development of sig-
nificant wound complications (11). A comprehensive nutritional assessment — including
anthropometric measurements, serological testing, neutrophil functional analysis, and
cutaneous assessment of delayed hypersensitivity — was performed on all patients. Those
patients with albumin levels less than 3 g/dL or those with serum transferrin levels below
150 mg/dL had significantly more wound complications than the rest of the cohort. It is
certainly advisable for any elective procedure to optimize a patient's nutritional status in
order to reduce operative morbidity and mortality. To evaluate the multiple factors that
predisposed to wound infections in our series, we evaluated 126 consecutive patients who
underwent in situ bypass, and found that early graft revision (<4 days) and the presence of
a lymph leak significantly increased the risk for post-operative wound infection (19).
However, factors such as age, race, diabetes, duration of operation, and presence of
gangrene or ulceration did not significantly influence the incidence of infectious compli-
cations in that series. Wengrovitz and coworkers retrospectively studied 163 subcutaneous
saphenous vein bypasses and found on regression analysis that chronic steroid use,
ipsilateral ulceration, and pedal bypasses predicted an increased incidence of wound
infection (6). They also identified female gender, diabetes, use of continuous incisions, and
procedures for limb salvage as factors associated with wound complications in their group
of patients.

The natural history of graft infection depends partly on the timing of presentation,
which has a widely variable interval between implantation and recognition of the infection.
Multiple studies have indicated that wound and graft infections tend to occur early. In
Lorentzen's study, 85% of graft infections occurred in the first 30 days (20). Likewise,
Liekweg et al. reported that 85% of groin wound infections in their series presented within
5 weeks of the initial operation (21). However, graft infection may not become clinically
evident for months to years after placement. Early graft infections are usually easily iden-
tified due to associated wound complications and signs of systemic inflammation. Graft
infections that are present in a delayed fashion, on the other hand, usually do not present
with signs of sepsis and are associated with more nonspecific symptoms.

Morbidity and mortality associated with graft infection depends not only on the
timing of presentation but also on microbiology, graft location, and method of treatment.
Unrecognized or inadequately treated infrainguinal graft infection has a mortality rate
ranging from to 22% and results in amputation in between 8 and 53% of cases, with one
series reporting an amputation rate of 79% (1,22,23).

B. Intraoperative Management

The conduct of the operation itself has a significant impact on primary wound healing and 13

the incidence of wound complications. Meticulous sterile technique begins at the scrub §

sink and does not end until the dressings are applied. During this variable length of time, j§

the surgeon must not only concentrate on the procedure at hand but also be ever vigilant £

toward issues that may ultimately influence the balance of wound healing. There is no ^

doubt that the stakes are high when it comes to managing the risk of infection in vascular "jl

surgery, especially with the use of prosthetic material. As an adjunct to meticulous surgical 3

technique, the administration of prophylactic antibiotics has helped to limit the incidence °

of wound infections. Multiple retrospective studies have demonstrated this efficacy. In J

addition, two prospective, randomized, and blinded studies have also shown that the use o

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

368 LORELLI and TOWNE

of cephalosporins results in a highly significant decrease in infection rates compared to
placebo (24,25). Another intraoperative concern is maintaining normothermia. All
anesthetics tend to cause hypothermia by decreasing heat production and causing vaso-
dilatation (26). Evidence suggests that intraoperative normothermia may decrease wound
infection rates by as much as two-thirds (27). High ambient room temperature and forced-
air warming seem to be the two most efficacious methods for maintaining normal body
temperature (28).

Avoiding the creation of skin flaps is paramount to avoid eventual ischemia of the
undermined skin (5,29). This is especially important in harvesting the saphenous vein in the
thigh because of its unpredictable and deep location (30). Routine preoperative vein map-
ping to evaluate the quality of vein and determine its exact location helps to avoid large
dissection flaps (25). Intraoperative ultrasound can also aid in determining saphenous vein
location. Lack of adequate hemostasis can also lead to considerable problems, especially if
large potential dead spaces are created due to poorly placed saphenectomy incisions. If
any question remains in regards to adequate hemostasis following saphenectomy for cor-
onary artery bypass, the wound should be closed only after heparin reversal. The type of
conduit used for lower extremity bypasses can influence the incidence of postoperative
wound problems. Prospective, randomized data have shown autogenous vein to be su-
perior to prosthetic grafts when considering long-term patency for infrapopliteal bypasses
(31). Additionally, the use of autologous tissue may reduce the potential for graft in-
fection. Johnson et al. (2) noted a statistically significant difference in the incidence of
wound complications when comparing PTFE bypasses (43%) to autogenous vein (27%)
grafts. Lorentzen et al. reviewed over 2400 arterial synthetic reconstructions over a 4-year
period with an incidence of graft infection of 2.6%. However, local wound complications
such as infection, necrosis, and hematoma were predisposing factors in half of these cases
(20).

There is no incision longer than that required for saphenous vein harvesting, and
studies have demonstrated that a reduced morbidity can be achieved with the use of in-
terrupted incisions rather than a continuous incision. Schwartz et al. analyzed 93 patients
for wound complications after in situ bypass with an overall incidence of 33%. However,
only 20.5% of patients with interrupted incisions developed wound problems compared to
42.5% of patients with a continuous incision. This proved to be a significant difference
(32). Likewise, a review of 163 autogenous bypass grafts by Wengrovitz et al. revealed a
wound complication incidence of 27.7% for a continuous incision versus 9.6% for in-
terrupted incisions with an overall incidence of 17% (6). Prospective, randomized, multi-
center data comparing reversed and in situ grafts revealed no difference in the incidence of
wound complications in 125 patients (33).

No general consensus exists in regards to the best method of skin closure. Often the
choice is guided by surgeon preference. The only prospective, randomized data are from •§

Angelini et al., in which they compared the outcome of saphenectomy incisions in 113 g

patients undergoing coronary artery bypass grafting (34). Four different skin closure a

techniques were used, including continuous nylon vertical mattress suture, continuous c

subcuticular absorbable suture, skin staples, and adhesive sutureless skin closure. Two <

independent observers evaluated all wounds at 5, 10, and 45 days after operation. The &

incidence of established wound infection was 4.5% overall; however, no patient developed 4j

an infection in wounds closed with a continuous subcuticular absorbable suture. In addi- 2

tion, this method of skin closure produced the best cosmetic results and therefore should |

be the method of choice for closure of saphenectomy wounds. ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

WOUND COMPLICATIONS 369

C. Postoperative Care

Routine postoperative principals of wound care certainly apply to vascular patients.
Sterile operative dressings should be left untouched for at least 48 h to allow for wound
epithelialization. Any part of a wound that drains serous fluid should continue to be
cleansed daily and dressed with sterile technique. Some would also advocate prophylactic
antibiotic coverage in addition. All groin wounds, especially those in obese patients,
should continue to have wound dressings to avoid direct skin-to-skin contact from a large
pannus. Another aspect unique to vascular surgery is managing lower limb edema fol-
lowing distal arterial bypass grafting. This is a common problem occurring in 40-100% of
patients after bypass (35,36). Up to a 40% increase in leg volume has been shown to occur
in the subcutaneous tissue via computed tomography (CT) and magnetic resonance imag-
ing (MRI) (37). The control of this dependent edema is important in minimizing wound
complications following distal bypass (10). The etiology of this edema is incompletely
understood but is likely due to a multitude of factors, which may contribute to edema
formation to varying degrees in different patients. These include deep venous thrombosis,
lymphatic disruption, increased capillary filtration, and the generation of oxygen-derived
free radicals during reperfusion (37). While the use of antioxidants and lymphatic-sparing
incisions has been advocated, the best results are achieved with leg elevation and com-
pressive stockings to reduce postoperative leg edema (37).

IV. IMPACT

Wound complications have a significant impact on the overall outcome of patients under-
going lower extremity bypass surgery. A recent review of 112 consecutive infrainguinal
bypass grafts by Nicoloff et al. from Portland was performed to determine how often an
ideal result is actually achieved with bypass surgery for limb salvage (38). An ideal result
was defined as an uncomplicated operation, elimination of ischemia, prompt wound
healing, and rapid return to premorbid functional status without recurrence or repeat
surgery for wound complications or to maintain graft patency. While clinically important
palliation was frequently achieved (5-year graft patency and limb salvage of 77 and 87%),
ideal results occurred in only 14% of patients. Wound complications occurred in 24% of
patients, with operative and ischemic wounds requiring a mean of 4.2 months to heal.
This — along with other operative complications, graft patency, limb salvage, survival,
functional status, recurrent ischemia, and the need for repeat operations — contributed
significantly to the low number of ideal results in this study. Additionally, the data under-
score the point that all patients undergoing infrainguinal bypass require ongoing surveil-
lance and treatment to achieve optimal results. Indeed, another study, which was a
prospective evaluation of 1 19 patients with 156 infrainguinal incisions, revealed that distal
wound complications incurred additional expense related to reoperation, extended hospi- e

talization or rehospitalization, and rehabilitation or visiting nurse services (39). s

V. ANATOMICAL CONSIDERATIONS AND INCIDENCE
A. Cervical Incisions

Differing anatomical areas in vascular reconstructive surgery carry very different concerns
and implications for wound complications. Cervical incisions for carotid endarterectomy
rarely develop problems with wound healing. This is likely due to an exceptional blood

•a

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

370 LORELLI and TOWNE

supply in this area and the clean nature of the case. Thompson et al. noted an incidence of
wound infection of 0.09% in 1140 carotid endarterectomies (40). Infected wounds should
be incised and drained, along with consideration given to replacing any prosthetic patch
with an autogenous vein if the vascular reconstruction is involved in the infection.

B. Abdominal Approach

There is much debate as well as conflicting data about the best method with which to
approach abdominal aortoiliac procedures (41-43). Those favoring the retroperitoneal
approach claim this to be a more physiological route and thus preferable to the more
widely applied midline approach (41^43). However, the only prospective, randomized
study, by Cambria et al., enrolled 59 transperitoneal and 54 retroperitoneal approaches to
elective aortic reconstructions (44). A multitude of parameters were evaluated, including
incidence of wound complications, cross-clamp times, crystalloid and transfusion require-
ments, respiratory morbidity, return of gastrointestinal function, and duration of hospital
stay. They were unable to demonstrate an advantage for the retroperitoneal approach in
regard to any of these factors; thus it should not be adopted as a preferred technique for
routine aortic reconstruction.

C. Groin Incisions

Access to the common femoral, superficial femoral, and profunda femoris arteries is a
mainstay in vascular reconstruction. These vessels often provide outflow for aortoiliac
procedures and are universally used to provide inflow for lower extremity bypasses. In-
cisions in this area can be hazardous given its proximity to the perineum and the potential
for direct contamination from both urine and stool. This part of the body tends to remain
moist from lack of hygiene and contact from an overlying pannus. Studies have shown
value to preoperative cleansing and the use of a povidone-iodine surgical scrub to reduce
the incidence of wound infections (7,19). Patients and surgeons must be vigilant in regard
to wound hygiene to avoid problems before they arise. Kent et al. noted a 10% incidence
of wound complications in 77 isolated groin incisions (39).

Another unique aspect of groin wounds is the propensity for development of
lymphatic complications due to the rich lymphatic network of the femoral triangle. These
are uncommon but potentially serious complications of femoral reconstruction, partic-
ularly if a vascular prosthesis is involved. A review of the Henry Ford Hospital experience
by Tyndall et al. identified 41 lymphatic complications (28 lymphocutaneous fistulas and
13 lymphoceles) in 2679 arterial reconstructions over a 15-year period, for an incidence of
1.2% per incision (40). Interestingly, the incidence varied with the type of procedure and
whether the procedure was a reoperation. Aortobifemoral bypass for aneurysmal disease
in a previously operated groin had the highest incidence (8.1%) of lymphatic complica- •§

tions, followed by an isolated femoral procedure in a previously operated groin (5.3%). g

The lowest frequency was found in patients undergoing a femoropopliteal/tibial bypass for a

the first time (0.5%). Aggressive treatment of lymphatic fistulas provided the best results. -c

There were no wound or graft infections in patients treated with operation; in them, re- <j

solution of the fistula occurred 2 1/2 times sooner than in those treated conservatively. >9

Length of hospital stay, time to resolution, and incidence of wound complications did not 4j

vary between those treated operatively or with aspiration and in those with lymphoceles 2

treated conservatively. A similar incidence of lymphatic complications with improved out- |

comes with operative therapy has been shown in other studies (46). ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

WOUND COMPLICATIONS 371

Drainage of the leg wound, especially the groin, early in the postoperative period is
often the result of divided lymphatics that have not sealed. These complications, a com-
mon problem in patients who have undergone kidney transplantation, are presumably
caused by an increase in lymph drainage from placement of the donor kidney in the lower
quadrant. The problems caused by a lymph leak were documented in a series of 126 con-
secutive patients reported by Reifsnyder et al., who underwent in situ bypasses of the lower
extremity (19). Risk-factor analysis demonstrated that the development of a postoperative
lymph leak was significantly related to the subsequent wound infection. Rubin et al. dem-
onstrated, in an animal model, that lymphatics contaminated with bacteria resulted in
positive blood and graft cultures (47). Experimentally, transection of lymphatics at the
graft site in the presence of a distal infection leads to significantly more graft infections
than does lymphatic ligation and exclusion. Lymphatic bacterial transport contributed to
the graft infection both from direct seeding and from transmission of bacteria to the
blood, leading to seeding of the graft.

Because of the relationship of lymph leak with subsequent significant wound infec-
tions, these patients should be treated aggressively. Our initial plan is to paint the wound
with povidone-iodine (Betadine) and apply a tight compressive dressing. Antibiotics, usu-
ally cephalosporins, are given and the patient is placed on bed rest. If the wound continues
to drain for greater than 72 h, the patient should be returned to the operating room and
the wound explored. The offending lymphatic can often be identified and suture-ligated. A
subcutaneous drain, well separated from the arterial prosthesis, is then brought out
through a separate stab hole. Placement of the drain allows the skin incision to heal. This
technique is generally successful in controlling the wound drainage and, more importantly,
prevents secondary infection of the lymphatic cavity.

1 . Lymphocele

Patients who develop lymphoceles following groin surgery are followed expectantly. If the
lymphoceles increase in size with time, they should be operatively explored and treated as
noted above with regard to the leaking wound. Likewise, if they communicate with the
groin wound and begin to leak, they should also be explored. Small to moderate-size
lymphoceles that are away from the incision and do not involve the graft can be followed.
Many times these will resolve slowly over time. Lymphoceles also can be treated op-
eratively if they become large or uncomfortable for the patient; a lymphocele that distends
the groin wound should be treated operatively as well. The injection of isosulfan blue into
the foot prior to the operation helps to identify the lymphatic channels that feed the
lymphocele. The use of duplex ultrasound can clearly identify the lymphocele and de-
termine whether it is adjacent to the prosthesis. Lymphoceles that are in close proximity to
the vascular prosthesis are best drained surgically in order to prevent any possible sec-
ondary infection of the vascular graft. When the lymphocele is well separated from the "g
vascular prosthesis, sclerotherapy using powdered tetracycline can be used. Powdered &
tetracycline is mixed with sterile saline and injected into the lymphocele; this will often a
sclerose the lymphocele (48). The most important factor in dealing with lymphatic c
problems of the groin is prevention, which may be achieved by doing meticulous dissection <j
with ligation of any lymphatic channels noted during vascular exposure. In particular, if a >9
lymph node is inadvertently bisected during the dissection, both halves should be suture- jjj
ligated. 2
With the number of endovascular aortic stent graft procedures increasing over the last |
few years, wound problems with the femoral cutdown have been scrutinized. Oblique @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

372 LORELLI and TOWNE

groin incisions have become a popular method of femoral access (49). In a study of 98
consecutive stent graft patients, infectious and lymphatic complications developed in 5 of
176 groin incisions (2.8%) (43). This compares favorably with the results of other studies
employing the standard vertical incision for both endovascular and conventional proce-
dures and is thus advocated as the preferred technique for femoral access (19,39,50,51).

D. Lower Extremity Incisions

Lower leg incisions account for a significant proportion of wound complications; they and
tend to be the most difficult to heal and take the longest to do so. These wounds also cause
additional anxiety to those using the in situ technique, in which even superficial wounds
may place the underlying subcutaneous graft at risk for exposure and infection. Multiple
studies have shown an incidence of wound complications for lower extremity bypass in the
range of 13-44% (19,39,52,53). Independent predictors of subsequent wound problems
include obesity (39) age, early graft revision, and lymph leak as well as ipsilateral limb
ulcer and pedal bypasses. Of the deep grade III infections that involve the bypass graft, the
best results to achieve limb salvage are found with aggressive therapy involving surgical
debridement and soft tissue or muscle flap coverage in order to preserve the bypass graft
(19,54).

With this in mind, many have investigated the usefulness of the endoscope in perform-
ing lower extremity bypasses. Two relatively small, retrospective studies have compared
conventional saphenectomy to endoscopic saphenous vein harvest in the performance of
reversed vein bypasses. Illig et al. showed a difference in all wound complications (21 and
51%) between endoscopic and conventional techniques (55). Serious wound problems
(Szilagyi class II or III) also differed (25 and 14%) between the two groups. No difference
existed in the short-term (30-day) graft patency, or the average length of stay for all pa-
tients. Robbins et al. found no difference in wound complications or short-term patency.
In addition, endoscopic vein harvest was associated with an increased operative time but
a shorter length of hospital stay (56).

Endovascular assisted in situ bypass grafting has been advocated to lessen the in-
cidence of operative wound problems. This technique utilizes a coaxial catheter emboliza-
tion system for intraoperative coil embolization of the vein's side branches. This obviates
the need for one long or multiple short incisions for vein preparation. A prospective,
randomized trial of 97 in situ bypasses by van Dijk et al. revealed significantly less overall
wound complications in the endovascular-assisted group (34%) compared to the conven-
tional group (72%) (57). However, the number of Szilagyi grade III wounds was equal in
both groups. While 1-year patency rates were similar, 14% of the endovascular-assisted
group had to have skin incisions performed for technical failure, which included branch
ligation due to inability to cannulate side branches or because the greater saphenous vein •§

was too small in diameter to accept the coaxial catheter. In addition, 20 of the 47 endo- g

vascular-assisted patients required postoperative intervention for retained arteriovenous a

fistulas, while only 4 of the 50 conventional patients required such an intervention. A c

report by Cikrit et al. found similar results in regard to wound complication rates and an <j

increased number of retained arteriovenous fistulas in the endovascular-assisted patients &

(58). |

Over half a million coronary artery bypass graft (CABG) procedures are performed in 2

the United States each year. This produces a significant number of saphenectomy wounds, |

many of which lead to evaluation by vascular surgeons because of wound breakdown. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

WOUND COMPLICATIONS 373

While the potential for exposure of an underlying bypass graft does not exist in these
patients, wound problems still pose a significant morbidity for a large number of patients,
which may result in prolonged length of stay, increased hospital cost, additional surgical
procedures, and possibly limb loss. The reported incidence of saphenectomy problems
after CABG ranges from 1 to 35% (58-61). One of the largest series, by Paletta et al.,
retrospectively reviewed 3525 CABG procedures over a 10-year period, finding wound
complications in 145 patients (4.1%), 23 of whom (0.65%) required additional surgical
interventions, including 1 1 vascular bypass procedures and 5 amputations (62).

Endoscopic vein harvesting has been used to help reduce the incidence of wound
complications. The data in the published literature are mixed in regards to the usefulness
of endoscopic vein harvesting for CABG. The results of three prospective studies have
been published within the last 5 years. The first, by Allen et al., is a prospective, ran-
domized trial of 112 patients undergoing elective CABG. They found a significant re-
duction in leg wound complications (4 vs. 19%) between the endoscopic and traditional
technique, with a 5.6% conversion rate to an open saphenectomy harvest (63). Hayward et
al. prospectively randomized 100 patients to endoscopic or open vein harvest. No sig-
nificant difference in wound complications or length of hospital stay between the two
groups was noted (64). Another prospective study of 60 patients compared endoscopic
harvest to the skin bridge harvesting technique (65). While harvesting times were shorter
with the endoscopic approach, this group had a wound complication rate of 32%, which
compared to only 3% in the skin bridge group. Even excluding wound hematomas, the
endoscopic group still had a wound complication rate of 13%.

VI. WOUND INFECTION

Once wound complications have occurred, they must be treated in an appropriate and
timely manner in order to prevent further morbidity. A multitude of treatment modalities
exist for wound care; however, basic surgical principles of wound care take precedence and
all other modalities are likely to fail if these principles are not upheld. The most important
of these is operative debridement of all devitalized tissue. Necrotic tissue provides an
excellent nidus for further infection and will significantly hinder any other attempts at
wound healing by other modalities. In order for a wound to heal, viable tissue is needed in
order to foster an environment for the development of granulation tissue. In the most
severe cases, soft tissue or muscle flap coverage may be necessary in order to provide even
viable tissue to promote healing and, in some instances, to provide soft tissue coverage for
an underlying vascular graft (54,66). All wounds should be appropriately dressed to pro-
tect them while healing, and wound care applications or changes should be performed in a
clean or sterile fashion depending on the particular wound. The use of antibiotics is an
adjunctive measure to the treatment of wound infection, especially after vascular bypass ■g

operations in which perfusion has been improved dramatically. Any localized collection or &

abscess needs open drainage. a

The wound care market has been expanding significantly in recent years. Most health -c

care companies have some sort of wound care product on the market and available for use <j

by any physician, nurse, or wound care specialist. Wound care ointments include papain- >9

urea debriding ointment for the removal of necrotic tissue, which can also be combined 4j

with a chlorophyllin copper complex to aid in deodorizing. Other gels include iodine- 2

containing hydrophilic beads for the reduction of microbial load and exudate absorption. |

While not strictly a wound complication, many of the patients undergoing lower extremity ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

374 LORELLI and TOWNE

bypass have nonhealing diabetic ulcers. The treatment of these hard-to-heal wounds has
been assisted by the use of becaplermin, a recombinant human platelet-derived growth
factor for topical administration. The biological activity of this agent includes promoting
the chemotactic recruitment and proliferation of cells involved in wound repair and pro-
moting the formation of granulation tissue.

Treatment options also include different mechanical devices. A subatmospheric pres-
sure dressing has been approved for wound treatment in the United States since 1995.
These dressings use negative pressure to create a suction force that promotes wound
healing in two ways. First, it enables wound drainage by removing excess interstitial fluid.
Second, it transmits mechanical forces to the surrounding tissues, with resultant defor-
mation of the extracellular matrix and cells. This encourages rapid ingrowth of granu-
lation tissue, allowing wounds to close more quickly than with other forms of therapy
(67,68). This method of wound treatment has been extensively studied in animal models
and been shown to create an environment that promotes wound healing by increasing
blood flow levels and rates of granulation tissue formation while also decreasing bacterial
counts (69).

Given that tissue hypoxia is a major component in wound complications, theoretically
hyperbaric oxygen therapy (HBO), which transiently increases the partial pressure of
oxygen in the plasma, should aid in wound healing. However, the published data on the
benefit of HBO in wound healing are inconclusive. Most studies are small, retrospective,
and nonrandomized, with several potential sources of bias (70). Indeed, Ciaravino et al.
reviewed 54 patients with lower extremity wounds, all of whom underwent HBO therapy
for an average of 30 treatments (71). The average cost for the HBO treatments alone was
$14,000. A total of 80% of patients showed no improvement, 11% showed some improve-
ment, and none were completely healed. Complications of HBO occurred in 63% of pa-
tients; most commonly barotrauma to the ears occurred in 43%, as a result of which the
vast majority required myringotomy tubes. Other complications included cardiac arrhyth-
mias, stroke, and seizure. Without strong prospective, randomized, placebo-controlled
data, it is difficult to justify any benefit to HBO therapy in the treatment of wound prob-
lems given the current data, the cost and the incidence of complications.

In conclusion, wound complications following vascular surgery constitute a significant
problem that, if not addressed in a timely and appropriate manner, may lead to significant
morbidity and mortality with compromise of any underlying vascular graft. The treatment
must be aggressive, with prompt exploration of wounds, wound debridement, and cover-
age of exposed vascular structures with viable autogenous tissue is the cornerstone of
effective treatment. Delay results in an increase of limb loss and mortality secondary to
graft complication of anastomotic bleeding and thrombosis.

REFERENCES 1

8

1. Szilagyi DE, Smith RF, Elliott JP, Vrandecic MP. Infection in arterial reconstruction with js

synthetic grafts. Ann Surg 1972; 176:321-333. Jf

2. Johnson JA, Cogbill TH, Strutt PJ, Gundersen AL. Wound complications after infrainguinal ^
bypass. Arch Surg 1988; 123:859-862. ■§

3. Goodson WH III, Hunt TK. Wound healing in experimental diabetes mellitus: importance of ^
early insulin therapy. Surg Forum 1978; 29:95-98. q

4. Bagdade JD, Root RK, Bulger RJ. Impaired leukocyte function in patients with poorly con- |
trolled diabetes. Diabetes 1974; 23:9-15. 2

5. Utley JR, Thomason ME, Wallace DJ, Mutch DW, Staton L, Brown V, Wilde CM, Bell MS. §

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

WOUND COMPLICATIONS 375

Preoperative correlates of impaired wound healing after saphenous vein excision. J Thorac
Cardiovasc Surg 1989; 98:147-149.

6. Wengrovitz M, Atnip RG, Gifford RR, Neumyer MM, Heitjian DF, Thiele BL. Wound
complications of autogenous subcutaneous infrainguinal arterial bypass surgery: predisposing
factors and management. J Vase Surg 1990; 11:156-163.

7. Cruse PJ, Foord R. A prospective study of 23,649 surgical wounds. Arch Surg 1973; 107:206-210.

8. Verta MJ, Gross WS, van Bellen B, Yao JS, Bergan JJ. Forefoot perfusion pressure and minor
amputation for gangrene. Surg 1976; 80:729-734.

9. Edwards WH, Martin RS, Jenkins JM, Edwards WH Sr, Mulherin JL. Primary graft infec-
tions. J Vase Surg 1987; 6:235-239.

10. Robison JG, Ross JP, Brothers TE, Elliott BM. Distal wound complications following pedal
bypass: analysis of risk factors. Ann Vase Surg 1995; 9:53-59.

11. Casey J, Flinn WR, Yao JS, Fahey V, Pawlowski J, Bergan JJ. Correlation of immune and
nutritional status with wound complications in patients undergoing vascular operations. Surg
1983; 93:822-827.

12. Blankensteijn JD, Gertier JP, Peterson MJ, Brewster DC, Cambria RP, LaMuraglia GM,
Abbott WM. Avoiding infrainguinal bypass wound complications in patients with chronic re-
nal insufficiency: the role of the anatomic plane. Eur J Vase Endovasc Surg 1996; 11:98-104.

13. Dovgan PS, Shepard AD, Nypaver TJ. Critical limb ischemia in patients with end stage renal
disease: do long-term results justify an aggressive surgical approach? Perspect Vase Surg 1999;
12:81-92.

14. Johnson BL, Glickman MH, Bandyk DF, Esses GE. Failure of foot salvage in patients with
end-stage renal disease after surgical revascularization. J Vase Surg 1995; 22:280-286.

15. Wasserman RJ, Saroyan RM, Rice JC, Kerstein MD. Infrainguinal revascularization for limb
salvage in patients with end stage renal disease. S Med J 1991; 84:190-192.

16. Simsir SA, Cabellon A, Kohlman-Trigoboff D, Smith BM. Factors influencing limb salvage
and survival after amputation and revascularization in patients with end stage renal disease.
Am J Surg 1995; 170:113-117.

17. Hunt TK, Hopf HW. Wound healing and infection: what surgeons and anesthesiologists can
do. Surg Clin North Am 1997; 77:587-606.

18. Taylor DE, Whamond JS, Penhallow JE. Effects of hemorrhage on wound strength and fi-
broblast function. Br J Surg 1987; 74:316-319.

19. Reifsnyder T, Bandyk D, Seabrook G, Kinney E, Towne JB. Wound complications of the in
situ saphenous vein bypass technique. J Vase Surg 1992; 15:843-850.

20. Lorentzen JE, Nielson OM, Arendrup H, Kimose HH, Bille S, Andersen J, Jensen CH,
Jacobsen F, Roder OC. Vascular graft infection: an analysis of sixty-two graft infections in
2411 consecutively implanted synthetic vascular grafts. Surgery 1985; 98:81-86.

21. Liekweg WG Jr, Greenfield LJ. Vascular prosthetic infections: collected experience and results
of treatment. Surgery 1977; 81:335-342.

22. Calligaro KD, Westcott CJ, Buckley RM, Savarese RP, DeLaurentis DA. Infrainguinal
anastomotic arterial graft infections treated by selective graft preservation. Ann Surg 1992;

216:74-79. .

23. Kilta MJ, Goodson SF, Bishara RA, Meyer JP, Schuler JJ, Flanigan DP. Mortality and limb 1
loss with infected infrainguinal bypass. J Vase Surg 1987; 5:566-571. g

24. Kaiser AB, Clayson KR, Mulherin JL Jr, Roach AC, Allen TR, Edwards WH, Dale WA. ,|
Antibiotic prophylaxis in vascular surgery. Ann Surg 1978; 188:283-289. 2

25. Pitt HA, Postier RG, MacGowan WA, Frank LW, Surmak AJ, Sitzman JV, Bouchier-Hayes ^

J g

D. Prophylactic antibiotics in vascular surgery: topical, systemic, or both? Ann Surg 1980; «

192:356-364. J

26. Matsukawa T, Sessler DI, Sessler AM, Schroeder M, Ozaki M, Kurz A, Cheng C. Heat flow and q
distribution during induction of general anesthesia. Anesthesiology 1995; 82:662-673. |

27. Kurz A, Sessler DI, Lenhardt R. Perioperative normothermia to reduce the incidence of 2
surgical wound infection and shorten hospitalization. N Engl J Med 1996; 334:1209-1215. ^

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

376 LORELLI and TOWNE

28. Kurz A, Kurz M, Poeschl G, Faryniak B, Redl G, Hackl W. Forced air warming maintains intra-
operative normothermia better than circulating water mattresses. Anesth Analg 1993; 77:89-95.

29. DeLaria GA, Hunter JA, Goldin MD, Serry C, Javid H, Najafi H. Leg wound complications
associated with coronary revascularization. J Thorac Cardiovasc Surg 1981; 81:403^107.

30. Ruoff BA, Cranley JJ, Hannan LA, Aseffa N, Karkow WS, Stedje KG, Cranley RD. Real-time
duplex ultrasound mapping of the greater saphenous vein before in situ infrainguinal
revascularization. J Vase Surg 1987; 6:107-113.

31. Veith FJ, Gupta SK, Ascer E, White-Flores S, Samson RH, Scher LA, Towne JB, Bernhard
VM, Bonier P, Flinn WR, Astelford P, Yao JS, Bergan JJ. Six-year prospective multicenter
randomized comparison of autologous saphenous vein and expanded polytetrafluoroethylene
grafts in infrainguinal arterial reconstructions. J Vase Surg 1986; 3:104-114.

32. Schwartz ME, Harrington EB, Schanzer H. Wound complications after in situ bypass. J Vase
Surg 1988; 7:802-807.

33. Wengerter KR, Veith FJ, Gupta SK, Goldsmith J, Farrell E, Harris PL, Moore D, Shanik G.
Prospective randomized multicenter comparison of in situ and reversed vein infrapopliteal
bypasses. J Vase Surg 1991; 13:189-199.

34. Angelini GD, Butchart EG, Armistead SH, Breckenridge IM. Comparative study of leg wound
skin closure in coronary artery bypass graft operations. Thorax 1984; 39:942-945.

35. AbuRahma AF, Wooddruff BA, Lucente FC. Edema after femoropopliteal bypass surgery:
lymphatic and venous theories of causation. J Vase Surg 1990; 11:461-467.

36. Eickhoff JH, Engell HC. Local regulation of blood flow and the occurrence of edema after
arterial reconstruction of the lower limbs. Ann Surg 1982; 195:474-478.

37. Soong CV, Barros D'Sa AA. Lower limb oedema following distal arterial bypass grafting. Eur
J Vase Endovasc Surg 1998; 16:465-471.

38. Nicoloff AD, Taylor LM, McLafferty RB, Moneta GL, Porter JM. Patient recovery after
infrainguinal bypass grafting for limb salvage. J Vase Surg 1998; 27:256-266.

39. Kent CK, Bartek S, Kuntz KM, Anninos E, Skillman JJ. Prospective study of wound com-
plications in continuous infrainguinal incisions after lower limb arterial reconstruction: in-
cidence, risk factors, and cost. Surgery 1996; 119:378-383.

40. Thompson JE. Complications of carotid endarterectomy and their prevention. World J Surg
1979; 3:155-165.

41. Sicard GA, Freeman MB, VanderWoude JC, Anderson CB. Comparison between the trans-
abdominal and retroperitoneal approach for reconstruction of the infrarenal abdominal aorta. J
Vase Surg 1987; 5:19-27.

42. Honig MP, Mason RA, Giron F. Wound complications of the retroperitoneal approach to the
aorta and iliac vessels. J Vase Surg 1992; 15:28-34.

43. Leather RP, Shah DM, Kaufmann JL, Fitzgerald KM, Chang BB, Feustel PJ. Comparative
analysis of retroperitoneal and transperitoneal aortic replacement for aneurysm. Surg Gynecol
Obstet 1989; 168:387-393.

44. Cambria RP, Brewster DC, Abbott WM, Freehan M, Megerman J, LaMuraglia G, Wilson R,
Wilson D, Teplick R, Davison JK. Transperitoneal versus retroperitoneal approach for aortic
reconstruction: a randomized prospective study. J Vase Surg 1990; 11:314-325.

45. Tyndall SH, Shepard AD, Wilczewski JM, Reddy DJ, Elliott JP, Ernst CB. Groin lymphatic 1
complications after arterial reconstruction. J Vase Surg 1994; 19:858-864. <S

46. Roberts JR, Walters GK, Zenilman ME, Jones CE. Groin lymphorrhea complicating revas- js
cularization involving the femoral vessels. Am J Surg 1993; 165:341-344. 2

47. Rubin JR, Malone JM, Goldstone J. The role of lymphatic system in acute arterial prosthetic ^
graft infections. J Vase Surg 1985; 2:92-98. ■§

48. Cannon L, Walker AJ. Sclerotherapy of a wound lymphocele using tetracycline. European J ^
Vase Endovas Surg 1997; 14(6):505. q

49. Caiati JM, Kaplan D, Gitlitz D, Hollier LH, Marin ML. The value of the oblique groin incision |
for femoral artery access during endovascular procedures. Ann Vase Surg 2000; 14:248-253. 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

WOUND COMPLICATIONS 377

50. Moore WS, Rutherford RB. Transfemoral endovascular repair of abdominal aortic aneurysm:
results of the North American EVT phase 1 trial. J Vase Surg 1996; 23:543-553.

51. Chuter TA, Reilly LM, Stoney RJ, Messina LM. Femoral artery exposure for endovascular
aneurysm repair through oblique incisions. J Endovasc Surg 1998; 5:259-260.

52. Dalman RL, Abbruzzese T, Bushnik T, Harris EJ. Open saphenectomy complications fol-
lowing lower extremity revascularization. Cardiovasc Surg 2000; 8:51-57.

53. Treiman GS, Copland S, Yellin AE, Lawrence PF, McNamara RM, Treiman RL. Wound
infections involving infrainguinal autogenous vein grafts: a current evaluation of factors de-
termining successful graft preservation. J Vase Surg 2001; 33:948-958.

54. Tukiainen E, Biancari F, Lepantalo M. Deep infection of infrapopliteal autogenous vein grafts —
Immediate use of muscle flaps in leg salvage. J Vase Surg 1998; 28:611-616.

55. Illig KA, Rhodes JM, Sternbach Y, Shortell CK, Davies MG, Green RM. Reduction in wound
morbidity rates following endoscopic saphenous vein harvest. Ann Vase Surg 2001; 15:104-109.

56. Robbins MR, Hutchinson SA, Helmer SD. Endoscopic saphenous vein harvest in infrainguinal
bypass surgery. Am J Surg 1998; 176:586-590.

57. van Dijk LC, van Urk H, du Bois NA, Yo TI, Koning J, Jansen WB, Wittens CH. A new
"closed" in situ vein bypass technique results in a reduced wound complication rate. Eur J Vase
Endovasc Surg 1995; 10:162-167.

58. Cikrit DF, Fiore NF, Dalsing MC, Lalka SG, Sawchuk AP, Ladd AP, Dodson S. A comparison
of endovascular assisted and conventional in situ bypass grafts. Ann Vase Surg 1995; 9:37-43.

59. Galbraith GF, Pica-Furey W. A retrospective comparative study of open and endoscopic
saphenous vein harvesting. J Endovasc Ther 2000; 7:460^168.

60. L'Ecuyer PB, Murphy D, Little JR, Fraser VJ. The epidemiology of chest and leg wound in-
fections following cardiothoracic surgery. Clin Infec Dis 1996; 22:424-429.

61. Slaughter MS, Olson MM, Lee JT Jr, Ward HB. A fifteen-year wound surveillance study after
coronary artery bypass. Ann Thorac Surg 1993; 56:1063-1068.

62. Paletta CE, Huang DB, Fiore AC, Swartz MT, Rilloraza FL, Gardner JE. Major leg wound
complications after saphenous vein harvest for coronary revascularization. Ann Thorac Surg
2000; 70:492-497.

63. Allen KB, Griffith GL, Heimansohn DA, Robison RJ, Matheny RG, Schier JJ, Fitzgerald EB,
Shaar CJ. Endoscopic versus traditional saphenous vein harvesting: a prospective, randomized
trial. Ann Thorac Surg 1998; 66:26-32.

64. Hayward TZ III, Hey LA, Newman LL, Duhaylongsod FG, Hayward KA, Lowe JE, Smith
PK. Endoscopic versus open saphenous vein harvest: the effect on postoperative outcomes.
Ann Thorac Surg 1999; 68:2107-2111.

65. Horvath KD, Gray D, Benton L, Hill J, Swanstrom LL. Operative outcomes of minimally
invasive vein harvest. Am J Surg 1998; 175:391-395.

66. Patterson MA, Cambria RA, Seabrook GR, Towne JB. Rotational Myoplasty: Treatment
Techniques for Covering Exposed or Infected Vascular Structures Associated with Com-
plicated Inguinal Wounds [abstr], Milwaukee, WI: Medical College of Wisconsin, 2002.

67. Evans D, Land L. Topical negative pressure for treating chronic wounds: a systematic review.
Br J Plastic Surg 2001; 54:238-242.

68. Argenta LC, Morykwas MJ. Vacuum-assisted closure: a new method for wound control and j>
treatment: clinical experience. Ann Plast Surg 1997; 38:563-577. <S

69. Morykwas MJ, Argenta LC, Shelton-Brown EI, McGuirt W. Vacuum-assisted closure: a new js
method for wound control and treatment: animal studies and basic foundation. Ann Plast Surg 2
1997; 38:553-562. «!

70. Wunderlich RP, Peters EJ, Lavery LA. Systemic hyperbaric oxygen therapy: lower extremity —
wound healing and the diabetic foot. Diabetes Care 2000; 23:1551-1555. jjj

71. Ciaravino ME, Friedell ML, Kammerlocher TC. Is hyperbaric oxygen a useful adjunct in the q
management of problem lower extremity wounds? Ann Vase Surg 1996; 10:558-562. |

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

21

Complications in the Management of the Diabetic Foot

Gary R. Seabrook

Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

In 2000, the prevalence of diabetes in the United States was estimated to be 17 million,
with 5.9 million of these cases undiagnosed. Of all American citizens over the age of 20,
some 8.6% have diabetes mellitus; but in the age group over 65 years, 20.1% have the
diagnosis. It is estimated that 60-70% of people with diabetes have some form of neu-
ropathy, leading in large part to impaired sensation of the extremities. In combination
with peripheral vascular disease, these nervous system changes place the feet of diabetic
patients at significant risk for injury and tissue loss. From 1997 to 1999, the Centers for
Disease Control recorded 82,000 nontraumatic lower extremity amputations performed on
patients with diabetes mellitus. It is estimated that $44 billion is spent annually in direct
medical expenses for the treatment of diabetes in the United States and that indirect
costs — including long-term disability, time lost from work, and premature mortality —
account for another $54 billion spent in association with this disease process (1). Aside
from the financial costs associated with diabetes, there are significant physical and
emotional costs associated with the possibility of lower extremity limb loss. Because of
this significant health care risk, complications in the management of the diabetic foot re-
quire particular attention.

For patients with diabetes mellitus, foot ulcers and foot infections are themselves
complications of the disease process. The triad of atherosclerotic vascular disease, periph-
eral neuropathy, and a propensity to develop polymicrobial infection places the patient ■g
with diabetes at constant risk for injury to the integument, soft tissue infection, osteo- &
myelitis, chronic ulceration, derangement of the skeletal architecture, and amputation. a
Mechanical trauma that would result in only a trivial injury to an individual with a nor- c
mal arterial system plagues the patient with diabetes with the risk of cellulitis, ulcerating <j
wounds, necrotizing infection, systemic sepsis, gangrene, and in some cases death. The >9
severity of the patient's diabetes, duration of the disease process, precision of glucose con- 4j
trol, or daily insulin requirements are not good predictors of which patients will be at 2
greater risk for foot complications. Every patient with diabetes is advised to be compulsive |
about wearing well-fitting footwear and inspecting the feet on a daily basis to detect any @

379 I

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

380 SEABROOK

sign of minor injury or neurotrophic trauma. Minor abrasions or a superficial blister that
would immediately be recognized by a patient with normal peripheral sensation can easily
be ignored by a diabetic patient because the pain is not appreciated; or, owing to visual
impairment related to diabetic retinopathy, the patient may simply be unable to see the
soft tissue defect.

Because the etiology of diabetic foot wounds is so closely related to microangiopathy
and the resulting chronic ischemia to the muscles, nerves, and bones of the foot, the
treatment of these lesions falls within the diagnostic and therapeutic purview of the vas-
cular surgeon. Even though many of the clinical presentations do not require arterial
revascularization, the expertise of a vascular specialist is appropriate in their management.

The chronic metabolic derangement of hyperglycemia from inappropriate insulin
secretion results in the common pathway of inadequate perfusion to the surface tissues,
intrinsic muscles, neural tissue, and osseous structures of the foot. The most obvious clin-
ical presentation of this chronic ischemia is tissue breakdown in a spectrum beginning with
superficial loss of the dermis and progressing to bullae, full-thickness skin loss, a pen-
etrating ulcer, destruction of underlying bone, or gangrene. Vascular occlusive disease in
diabetic patients results in arterial obliteration primarily in the vessels at the level of the
tibia, forefoot, and digits. It is not uncommon for the diabetic population to have normal
aortoiliac segments and even a patent femoral popliteal system, although arterial wall
calcification may be present at these levels. Not only does the distribution of the occlusive
process follow a unique pattern in the diabetic patient but there is a difference in the
structure of the blood vessel that is different from atherosclerotic changes due to cardio-
vascular risk factors such as hypertension, hypercholesterolemia, and cigarette smoking.
Arterial wall abnormalities are associated with calcification adjacent to the internal elastic
lamina. At the arteriolar level, histopathology reveals thickening of intima at the basement
membrane. The changes in the capillary walls may reveal a patchy distribution, with
normal capillaries interspersed with those that are diseased. Because these findings are
present in patients who may not yet have been diagnosed with clinical diabetes, it suggests
that the thickening of the basement membrane may be due to a genetic predisposition
rather than simply altered glucose metabolism (2).

While surface injuries from arterial insufficiency may be obvious, ischemic injury to
sensory and motor nerves results in more subtle insults. The lack of normal sensation allows
direct trauma to the foot to go unnoticed. It is not uncommon to hear that the patient has
walked for an entire day with a pebble in his or her shoe or even a nail protruding through
the soul of a shoe or boot into the foot. Were the patient to have a normal sensory system, a
foreign or penetrating object or would be promptly recognized and removed. Even poorly
fitting shoes may cause significant trauma to the foot. If a patient has normal sensation,
improperly fitting shoes are accommodated by subtle changes in weight and posture so as
to offload the irritating focus and prevent injury. In addition to sensory denervation to the "g

integument of the foot, diabetic neuropathy deprives the intrinsic muscles of the foot of &

normal motor innervation. This leads to disruption of the fine balance of the flexor and a

extensor mechanisms that provide for the usual precise motor function that results in a c

normal gait. The neuropathic dennervation of the intrinsic muscles leads to abnormalities <j

in the foot architecture characterized and by extensor subluxation of the toes, protruding >9

plantar prominence of the metatarsal heads, proximal migration of the metatarsal fat pads, 41

and dominance of the action of the toe flexors over the extensors. With dislocation of the 2

metatarsophalangeal joint, the metatarsal heads are no longer protected by the mechanical |

levers of the phalanx bones and the full weight of the patient's body becomes focused along @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

MANAGEMENT OF THE DIABETIC FOOT 381

a narrow strike point during ambulation. Because conventional footwear is designed for
normal foot anatomy, the patient's shoes may actually play a role in provoking physical
trauma to the surface of the foot. An ulceration will begin with local inflammation,
frequently followed by the formation of a callus, within which a foreign body or bacteria
may be sequestered. This closed-space environment provides a medium for bacterial
growth resulting in abscess formation within the dermis or, in more serious cases,
penetrating into the underlying a soft tissue. With further trauma, the callus breaks down,
exposing the tissue defect. Because the etiology of the process frequently relates to
abnormal forces applied to the underlying bone, the first presentation may involve not
only tissue loss but also injury from osteomyelitis of the underlying phalanx or metatarsal
(Fig. 1).

When chronic neuropathy leads to failure of the ligamentous structures between the
metatarsal and tarsal bones, there is eventual collapse of the ankle mechanism. This or-
thopedic defect, termed Charcot foot, manifests initially as loss of the plantar arch; how-
ever, in the absence of protection from the intrinsic musculature, the disorganized tarsal
bones may erode though the plantar surface of the foot (Fig. 2).

Polymicrobial infection completes the triad that plagues the foot of the diabetic
patient. Infections in diabetic patients pose a more serious risk to the patient's limb be-
cause they occur in tissue that is poorly perfused and because the clinical signs of the
infection may not be recognized in a timely fashion due to masking by the peripheral neu-
ropathy. When there is a break in the integument, and invasive infection may spread ra-
pidly because the inflammatory process is not perceived. The hyperglycemic state enhances
the environment for an infection to flourish. Intracellular bactericidal activity of leuko-
cytes is diminished in the hyperglycemic state. Tissue glycosylation further impairs wound
healing by increasing the activity of collagenases, thereby reducing the structural collagen
content at the site of injury and disrupting the normal reparative mechanisms. Dysfunc-
tional phagocytes further enhance an environment conducive to a synergistic microbial

Figure 1 Neurotrophic ulcer on plantar surface of first metatarsal head with exposed bone in-
volved with osteomyelitis.

i
2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

382 SEABROOK

Figure 2 Charcot foot with large plantar ulcer at site of collapsed forefoot.

infection. Polymicrobial infections with aerobes and anaerobes result in increased poten-
tial for tissue destruction. The combined virulence of multiple organisms in the wound
exceeds the individual destructive capabilities of any of the microbes acting alone.
"Synergistic virulence" results in the rapid progression of foot infections, and dictates
in the need for broad-spectrum antimicrobial therapy in their treatment. The microbial
populations recovered in culture from diabetic foot infections includes Gram negative and
positive aerobes, and gram-negative and positive anaerobes. (Table 1). Peptostreptococci
are recovered in significant frequency from deep soft tissue cultures obtained from patients
with diabetic foot infections, with only Bacteroides species being recovered more fre-
quently from anerobic cultures. Peptostreptococcus magnus has been identified as the
single most common clinical isolate from this bacterial group in diabetic foot infections.
These bacteria produce a potent collagenase that is responsible for skin and subcutaneous
tissue destruction, and this enzymatic activity results in the rapid progress of tissue de-
struction associated with diabetic foot infections. The ability to destroy tissue and pro-
liferate in an anaerobic environment further enhances the ability of these organisms to
invade the deep anatomical spaces of the foot (3).

I. ANTIBIOTIC THERAPY

For patients with serious diabetic foot infections, surgical intervention to drain abscess
cavities, remove devitalized tissue, and excise any septic focus is the primary treatment
objective. Antibiotics serve as adjunctive therapy to eradicate micro-organisms in the soft
tissue adjacent to the site of the infection. Antibiotics do not kill bacteria in the soft tissue

that have already been devitalized, nor are they effective in reaching infected material -c

without adequate arterial perfusion. Bacteria that have collected in the ascending lym- <j

phatic system may be effectively eradicated by antibiotic therapy. >9

The pathobiology of infections in patients with diabetes must be considered in pre- 41

scribing antibiotic therapy for associated soft tissue infections. In addition to a significant 2

microbial presence, there may be injury to the microenvironment already contaminated with |

foreign bodies or tissue debris. Secondary host factors, including tissue ischemia and a @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

MANAGEMENT OF THE DIABETIC FOOT

383

Table 1 Microbial Organisms Identified in Culture from Diabetic
Foot Infections-Prevalence of Isolate Recovered in a Population of 246
Patients

Prevalence

Gram-negative aerobes

Proteus species

60%

Escherichia coli

30%

Klebsiella species

25%

Pseudomonas aeruginosa

30%

Pseudomonas species

45%

Morganella morganii

10%

Enterohacter species

20%

Acinetohacter species

<5%

Citrohacter species

<5%

Gram-positive aerobes

Enterococcus faecalis

40%

Enterococcus faecium

5%

Staphylococcus aureus

30%

Staphylococcus epidermidis

50%

Streptococcus milleri group

35%

Streptococcus agalactiae

10%

Micrococcus species

20%

Corynehacterium species

10%

Miscellaneous coagulase-negative staphylococci

25%

Gram-negative anaerobes

Bacteroides fragilis group

70%

Bacteroides fragilis

45%

Porphyromonas species

5%

Fusobacterium species

25%

Prevotella species

15%

Veillonella parvula

<5%

Bifidobacterium species

<5%

Gram-positive anaerobes

Peptostreptococcus manus

50%

Peptostreptococcus species

35%

Actinomyces species

5%

Propionibacterium

< 10%

Clostridium perfringens

< 10%

Clostridium species

<5%

Source: Ref. 13.

compromised immune system, further complicate the environment in which antimicrobial
activity must occur. In this milieu, the infection has the potential to be enhanced by micro-
bial synergism, where the cooperative interaction of two or more bacterial species produces a
result not achieved by the individual bacterial alone. In mixed soft tissue infections, several
major microbial populations flourish. Anaerobes including Enterobacteriaceae, Staph-
ylococcus, Streptococcus, and Enterococcus coexist with anaerobes including Bacteroides,
Fusobacterium, Peptostreptococcus, and Clostridium, creating a population that contribute

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

384 SEABROOK

to microbial synergism. In the presence of this virulent microenvironment, there is both local
and systemic anergy, further disrupting anatomical integrity of the soft tissue and creating a
host for further bacterial proliferation. Conditions contributing to microbial proliferation
include metabolic abnormalities, not the least of which is hyperglycemia. The immune
system, which may already be compromised, is further suppressed in the presence of a
synergistic infection.

The selection of antibiotic therapy must take into account this unique collection of
conditions occurring in the foot of the diabetic patient. For this reason, antibiotics that
would be appropriate for any subset of the identified infecting bacteria may be inadequate
in this synergistic, polymicrobial infection. Because these infections tend to be invasive,
destructive to the soft tissue, and rapid in their progression, even empirical antibiotic ther-
apy must utilize a broad-spectrum regimen.

Semisynthetic penicillins with a beta-lactamase inhibitor provide a suitable broad-
spectrum antibiotic regimen for patients presenting with diabetic foot infections. Pseu-
domonas, Enterobacter, and Serratia all possess a gene for inducible beta-lactamase
production, which quickly hydrolyzes the beta-lactam ring of common penicillins including
ampicillin, amoxicillin, ticarcillin, piperacillin, rendering them inert and ineffective. Adding
the beta-lactamase inhibitor to the penicillin substrate results in a therapeutic compound
that prevents the beta-lactamase from degrading the beta-lactam agent, which can then go
on to inhibit cell wall synthesis in the micro-organism. Beta-lactamase inhibitors include
clavulanic acid, sulbactam, and tazobactam. With the use of these agents, there is the risk of
inducing of beta-lactamase activity within the infection; when new beta-lactam antibiotics
are introduced into clinical use, some previously unrecognized beta-lactamases with the
capacity to destroy the activity of the compound are identified.

Quinolones are ideal compounds for the treatment of diabetic foot infections. They
act by inhibiting DNA gyrase, the enzyme that mediates supercoiling of the bacterial DNA
into compacted domains capable of fitting within the confines of the bacterial wall. In-
hibition of DNA gyrase causes relaxation of the supercoiled DNA, thereby terminating
chromosomal replication, interfering with bacterial cell division, and blocking gene
expression. The antimicrobial spectrum of these agents varies greatly depending on the
specific generation of the drug. Second-generation quinolones (e.g., ciprofloxacin) exhibit
excellent gram-negative facultative activity but are not effective alone in treating diabetic
foot infections because they demonstrate little activity against the gram-negative anaero-
bic populations. Therefore these agents typically require the addition of an antianaerobic
compound to provide effective coverage against the anaerobic microbial flora encountered
in the diabetic foot. Fourth-generation quinolones, in which a fluorine atom is added to
the bicyclic aromatic core, provide excellent aerobic and anaerobic coverage against both
gram-positive and gram-negative microbial pathogens. This therapy is attractive because a
single agent with a pharmacokinetic profile that often allows for once-a-day dosing may be ■g

employed effectively in treating these infections. Because tissue penetration is excellent, &

patients may be started on an intravenous dosing schedule while hospitalized and then a

discharged home on an oral formulation of the drug. c

Traditional empirical therapy for polymicrobial infections has utilized prescription of <j

an aminoglycoside, which functions by inhibiting the synthesis of bacterial cell proteins. >9

Noteworthy are the significant potentially toxic side effects associated with this class of 4j

drugs. In the diabetic patient, the most serious complication is antibiotic-introduced neph- 2

rotoxicity, which is particularly hazardous in a patient population already likely to have |

impaired renal function due to diabetic nephrosclerosis. Despite potentially increased safety @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

MANAGEMENT OF THE DIABETIC FOOT 385

by careful pharmacokinetic dosing, the risk of renal failure in diabetic patients precludes the
use of aminoglycosides for the treatment of foot infections.

Although vancomycin is particularly effective against gram-positive bacteria through
its inhibition of cell-wall synthesis, the increasing amount of bacterial resistance that has
developed due to its extensive use in the hospital setting dictates that it should be ad-
ministered only in carefully selected patients (4). This drug should not be utilized as part of
an empirical therapeutic regimen. It does remain as an appropriate antibiotic for meth-
icillin-resistant Staphylococcus aureus. With the emergence of vancomycin-resistant En-
terococcus (VRE), linezolid has become an appropriate therapy (5). As with vancomycin,
this agent should be used with considerable discretion in an effort to control the emergence
of new resistant organisms, particularly in the nosocomial setting (6).

II. WOUND CARE

Equally important to antibiotic therapy as an adjunct to effective control of a diabetic foot
infection is appropriate care of the wound. Any break in the integument represents a
wound that may require attention both to control the infection and to promote effective
healing. Wound care is provided by removing the devitalized tissue, enhancing the normal
reparative functions of the skin, and providing protection against further insult from
mechanical or bacterial forces. Factors that impede wound healing include pressure, a dry
environment, the presence of necrotic tissue, ischemia, and inadequate nutrition.

Foot wounds should be protected from extrinsic pressure, with the realization that
neurotrophic trauma may, in fact, have been the initial etiology of the wound. Any wound
should be covered with a clean, bulky dressing. Gauze pads should be inserted between the
toes and a protective pad placed over the heel. In a compromised limb, the weight of the
foot against the bed sheets and mattress may provide an impediment to wound healing or
even cause further tissue injury.

A moist wound environment is one of the most important factors to promote healing
of a soft tissue defect. While it is true that a warm, moist milieu is also advantageous for
bacterial proliferation, effective wound treatment requires an environment conducive to
tissue repair and epithelialization. Simply covering a wound is helpful in that it protects
the site of wound healing from external trauma, and a wound dressing in that capacity
mimics the function of the epidermis. Initial research supporting this concept demon-
strated that experimentally induced blister wounds healed more rapidly when the blister
roofs were left intact (7). This concept of a "biological dressing" leads to the recom-
mendation of an occlusive dressing to both protect the wound and provide a constant
degree of moisture. This may best be provided by hydrocolloid dressings (Replicare, Smith
& Nephew; Thinsite and Transorbent, B. Braun Medical), which contain hydrophilic col-
loidal particles comprising gelatin or pectin formulated in an adhesive mass. Fluid in the "g
wound is absorbed by capillary action, which occurs with slow swelling of the particles. &
Although the absorptive capacity of the hydrocolloid is significant, dressing changes may a
be required more frequently in the early phases of wound healing, when more exudate is c
being produced. Although hydrocolloid dressings are occlusive, skin maceration is pre- <j
vented because of the absorptive qualities of the hydrophilic particles. The pliable nature >9
of the hydrocolloid dressing allows it to conform to the irregular surfaces adjacent to a 4j
foot wound. When a hydrocolloid dressing is removed, a viscous, gel-like material is evi- 2
dent, the result of decomposition of the adhesive mass. The clinician is cautioned not to |
confuse this material with a purulent infection. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

386 SEABROOK

In the absence of an occlusive dressing, hydrogels (SoloSite, Smith & Nephew; Curasol,
Healthpoint) may be used to create a moist environment. These agents, which contain a
significant percentage of water, have a complex lattice that traps a dispersion medium,
typically a cross-linked polymer. Because they possess high specific heat, these dressings feel
cool to the touch and are perceived as soothing by the patient. Lidocaine gel may be
incorporated with a hydrogel dressing to provide further analgesia for treatment of a
sensitive wound. Although clearly superior to a simple saline-soaked gauze dressing, use of a
hydrogel requires vigilant observation of the wound to prevent desiccation.

During the early inflammatory phase of wound healing, the coagulation cascade is
activated to deposit fibrin, which is polymerized to stabilize the wound. In the normal
course of wound healing, inflammation is followed by a process of granulation, in which a
loose matrix infiltrated by macrophages, fibroblasts, and capillary endothelial cells
appears in the open wound. Collagen, fibronectin, and hyaluronic acid add substance to
the wound. When this granulation process deviates from its natural balance, there may be
a proliferation of exudative materials, including tenacious fibrin strands that are actually
counterproductive to wound closure. In such a case, which is more likely to occur when
the wound edges are not in close proximity, debridement is required. The goal of de-
bridement is to remove devitalized exudative products and expose healthy granulation
tissue to allow epithelial cell migration. For the removal of gross material, mechanical
debridement with scissors or scalpel is appropriate. As a rule, this is a painless procedure
and should be performed whenever there is evidence of fibrinous exudate or devitalized
tissue on the surface of a wound. Fibrin, an important constituent the initial phase of
wound healing, can become counterproductive if a mechanical barrier develops over the
wound surface. An alternative to mechanical debridement is the use of an enzymatic agent
for chemical debridement. Proteolytic products (Collagenase Santyl, Smith & Nephew)
contain a collagenase that breaks down necrotic tissue without destroying the viable
granulation surface. Another product (Accuzyme, Healthpoint) uses papain (derived from
the Carica papaya) combined with urea, which is effective in breaking protein cross links of
wound debris. As the urea disrupts cross links, the proteins on the wound surface unfold
and thereby become susceptible to the enzymatic effect of the papain. When the proteins
are digested into small peptide fragments, the material can be washed from the wound
surface. Wound gels are also employed to rid a wound surface of excess fibrin deposition.
One product (Panafil, Healthpoint) contains a chlororphyllin copper complex that inhibits
fibrin formation while enhancing the structural integrity of the collagen matrix of the
granulation surface. The chlorophyll in the product is also an effective agent to reduce
odor emitted from the wound surface. Silver compounds have been proven effective in
augmenting wound healing, primarily by their bactericidal activity in reducing microbial
colonization on a granulating wound. Silver sulfadiazine cream is particularly effective in
controlling the growth of Pseudomonas on the wound surface. ■g

Wounds with a clean, granulated surface and well-perfused epithelial margins may &

benefit from the use of growth factors administered topically. Belcaplermin (Regranex, a

Ortho-McNeil) is a recombinant human platelet-derived growth factor that, when applied c

in daily to the surface of diabetic foot wounds, has demonstrated an increased incidence of <

complete healing in a shorter period of time (8). The product requires refrigeration and is &

expensive, but the cost may be justified when the patient requires a shorter period of 4j

wound therapy and returns more promptly to a functional status. 2

Wound closure may also be enhanced with mechanical maneuvers. The application of |

a subatmospheric pressure dressing to a foot wound provides a number of features that ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

MANAGEMENT OF THE DIABETIC FOOT

387

promotes wound closure. A foam sponge placed on the surface of an open wound and
connected by plastic tubing to a vacuum pump (Vacuum Assisted Closure, Kinetic Con-
cepts, Inc.) is effective in promoting the formation of granulation tissue (Fig. 3A and B).
The vacuum dressing augments wound contracture through the application of a constant,
localized negative pressure. The apparatus is also effective in removing interstitial wound
fluid and necrotic debris and thereby decreasing the buildup of devitalized exudative
products that act to impede wound closure. In addition to the mechanical benefits of this
process, the dressing provides a closed, moist wound-healing environment. Because the
dressing is usually left in place for a 48-h period, wound care is simplified, allowing more
efficient use of resources in the outpatient setting (9-11).

(A)

(B)

i

2

I

■c

Figure 3 (A) Foot ulcer covered with granulation tissue; (B) Foot ulcer treated with vacuum
closure device. Ulcer is covered with foam sponge secured with occlusive dressing incorporated with
vacuum pump tubing.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

388 SEABROOK

III. CASE MANAGEMENT

A. Disordered Architecture — No Infection

The feet of patients with diabetic peripheral neuropathy will be at risk for complications
from their disordered foot architecture. In some cases, the sensory deficit will lead to
neurotrophic ulceration; while these lesions result in disruption of the skin envelope, they
do not always represent a clinical infection. In the case of a superficial wound and in the
absence of swelling, erythema, and purulent discharge or wound exudate, the patient does
not require treatment for a soft tissue infection. The abnormal configuration of the foot
skeleton requires adaptive footwear to offload abnormal strike points and weight-bearing
surfaces that can result in continued soft tissue trauma.

A typical presentation involves superficial ulceration between the digits (Fig. 4). These
lesions frequently occur due to simple compression of the toes within the toe box of the
shoe. Diabetic patients are particularly at risk for developing these lesions when they are
outfitted with new shoes. Their neuropathy will not allow them to notice that the shoes do
not fit properly or to appreciate where the stiff fabric of the shoe is creating abnormal
contact with the foot. Even this type of minor trauma, which would quickly be obvious to
a person with normal sensation, will go unrecognized and lead to tissue breakdown. Dia-
betic patients must be cautioned to inspect their feet daily and, when using new footwear,
more than once a day in order to prevent this complication. When a superficial ulcer forms,
the offending pressure must be relieved. The wound should be treated with a hydrocolloid
dressing to provide a proper wound-healing environment. In the presence of a normal ar-
terial perfusion pressure, new epithelialization should occur.

B. Chronic Neurotrophic Ulcer — Callus — Sinus Tract

The patient with diabetic neuropathy may have a chronic callus, usually occurring over the
first or fifth metatarsal head. This is likely to be due to pressure resulting from an ab-
normal gait due to the lack of normal sensation of the foot striking the floor. In the
absence of protection of the skin surface, a thickened layer of epidermis forms a callus at
the site. Eventually this thickened area of cornified skin becomes vulnerable to bacterial
invasion, either from breakdown at the interface of the callus and the normal skin or due

i

2

I

■c

Figure 4 Superficial ulceration created from pressure of toes compacted in poorly fitting shoe.

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

u

MANAGEMENT OF THE DIABETIC FOOT

389

Figure 5 Chronic callus with subcutaneous abscess and sinus tract.

to some trivial injury to the foot. As the callus separates from the normal tissue, a space
forms that may sequester fluid, leading to chronic drainage (Fig. 5). Reduction of the
callus and the use of corrective footwear may eliminate the source of the abnormal
pressure and allow soft tissue healing. If the space beneath the callus is not well drained,
conditions exist for the formation of an abscess. Because of the neuropathy, it is unlikely
that the patient will sense of the normal inflammatory response. Recognition of the prob-
lem may be prompted by the drainage of foul smelling fluid or swelling and erythema
associated with the infection. This condition represents a surgical emergency. The wound
should be explored under regional anesthesia and broad-spectrum antibiotics prescribed.
In the operating room, the callus should be unroofed and the wound widely debrided.
Usually the amount of underlying soft tissue destruction will be more extensive than might
be anticipated from the initial appearance of the wound (Fig. 6). After surgical debride-
ment, the wound should be treated with hydrogel dressings. The site should be inspected at

i

2

I

■c

Figure 6 Site of chronic callus after surgical debridement.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

390 SEABROOK

least every 12 h to determine that all of the infection has been drained. The presence of
purulent fluid entering the wound when pressure is applied to the margins dictates that the
patient should be returned to the operating room for more extensive drainage. With
adequate arterial perfusion and a proper wound-healing environment, these wounds will
usually close by secondary intention. When the wound has completely healed and the foot
has a stable configuration, the patient should be referred for a custom orthotic with a total-
contact insert to prevent further soft tissue injury.

C. Web Space Infection

A soft tissue ulcer and/or associated gangrene occurring in the web space of the patient
with diabetes must be treated as a potentially serious, deep soft tissue infection (Fig. 7).
Even though the patient may report that the condition has been present for only a short
time, there may be a considerable amount of tissue necrosis beneath the surface of the
wound. Radiographs of the foot should be obtained to detect the presence of gas tracking
into the plantar space (Fig. 8). This type of wound requires urgent exploration in the
operating room, and broad-spectrum antibiotics should be administered. Commonly, the
infection will have extended to the joint space between the phalanx and the metatarsal and
will usually involve the soft tissue of both of the toes adjacent to the web space. It will not
be uncommon for the process to have extended to adjacent metatarsophalangeal joints,
where chronic neuropathy and altered foot architecture will have allowed disruption of the
planes normally separating these structures. The involved toes should be amputated and
the adjacent metatarsal heads resected so that no cartilaginous surface remains exposed.
The cartilage has no blood supply and will not allow granulation or wound healing.

It is not uncommon for the bacterial contamination to extend along the tracks of the
tendinous attachments to the phalanxes, creating a plantar space infection. This area must
be opened to allow resection of the devitalized tissue and permit drainage, which is usually
are accomplished by opening the plantar space below the level of the metatarsals.
However, at times the infection may be so extensive that the entire tract must be opened,

Figure 7 Web space infection with gangrene and invasive soft tissue infection.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

MANAGEMENT OF THE DIABETIC FOOT

391

Figure 8 Radiograph of foot with severe soft tissue infection. Gas permeates the tissue and tracks
into the plantar space.

onto both of the plantar and dorsal surfaces (Fig. 9). This creates a wedged defect in the
foot. If there is viable tissue to form an adequate walking surface, and the soft tissue can
be resected back to healthy margins, foot amputation should not proceed. After all of the
infection has cleared, it is possible to reapproximate the defect and allow the remainder of
the wound to heal by secondary intention (Fig. 10). This preserves a viable limb, providing
a better alternative than below-knee amputation and lifetime use of a prosthesis.

i

2

I

■c

Figure 9 Surgical excision of complex soft tissue infection extending through the foot. The second
metatarsal has been resected.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

392

SEABROOK

Figure 10 Following resolution of the infection, plantar surface is reconstructed to create a
walking surface.

D. Maggots

On occasion a patient will present for evaluation with an open foot wound infested with
maggots (Fig. 11). Such a condition obviously represents an aspect of poor personal
hygiene, because the surface of the wound at some time would have been contaminated by
the larvae, usually transmitted by a housefly. The maggots themselves are not actually
harmful to the patient and, conversely, may be effective in cleansing the wound (12).
However, their presence is usually untenable in a hospital situation, since they have not
been prescribed as part of a medical application and are perceived as representing active
wound contamination. Simple cleansing techniques are usually ineffective in removing the

i

2

I

■c

Figure 1 1 Maggots on surface of a chronic foot ulcer with gangrene.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

MANAGEMENT OF THE DIABETIC FOOT

393

creatures. However, they are rendered lifeless by the application of a topical spray an-
esthetic and can then be removed using simple surgical instruments.

E. Tissue Loss and Gangrene

A diabetic foot infection may result from an ischemic site in one toe. Although one might
think this due to neglect, in fact a polymicrobial infection in a diabetic patient can proceed
very rapidly and destructively. At the time of presentation, there is not only dermal gan-
grene but also necrotic changes to the underlying soft tissue and frequently disintegration
of the bone (Fig. 12 A and B). An infection of this magnitude represents a true surgical
emergency as the patient is at risk for ascending infection and limb loss. Immediate

(A)

i

2

I

■c

(B)

Figure 12 (A) Diabetic foot infection with extensive tissue loss and destruction of the second digit;
(B) Gangrene and soft tissue infection involves the plantar surface over the second metatarsal.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

394 SEABROOK

operative intervention should be undertaken. Usually a general anesthetic is preferred if
not contraindicated for cardiac or respiratory comorbidities, so that the scope of the re-
section will not be hindered by a limited regional anesthetic block. The surgeon and pa-
tient must be prepared to consider ankle disarticulation if the infection extends through
the plantar space to the level of the ankle joint. However, it is possible that the infection
will have been identified early enough that a local resection can be accomplished. The
surgical incision should follow the margins of the devitalized tissue. Any exposed bone
should be removed, and, with use of a bone rongeur, the metatarsals should be resected
deep to the surface of the soft tissue plane. Exposed tendons should be divided and al-
lowed to retract. The goal of the initial operation is to resect all infected material; attempts
to preserve marginal anatomic structures for future reconstruction should be avoided.
Scraping the soft tissue with a surgical curette is an effective method to differentiate viable
from dead tissue. This maneuver will remove fragments of soft tissue that are unlikely to
survive. At the conclusion of the surgical excision, the wound should be irrigated with
saline solution (Fig. 13). A power irrigating system using up to 3L of fluid may be em-
ployed. The wound should then be treated with hydrogel dressings and inspected fre-
quently for evidence of residual infection.

In the presence of a normal blood supply, granulation tissue should form on the
surface of the wound after a week to 10 days (Fig. 14). When there is an extensive area of
exposed soft tissue, a split-thickness skin graft is effective in providing wound closure
(Fig. 15). If there is a suspicion of surface bacterial colonization, quantitative cultures can
be obtained to avoid the complication of placing a skin graft on an infected surface. Col-
ony counts of less than 10 5 micro-organisms per cubic centimeter are usually considered
acceptable to proceed with skin grafting.

F. Failed Transmetatarsal Amputation

A diabetic foot infection may lead to loss of all of the digits on a patient's foot, requiring
open transmetatarsal amputation. It is not uncommon for this more extensive degree of
tissue loss to be associated with poor arterial perfusion of the distal foot. Following the
normal principles of treating a surgical infection, all of the devitalized tissue is debrided.

i

2

I

■c

Figure 13 Forefoot immediately following wide debridement of extensive diabetic foot infection.

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

u

MANAGEMENT OF THE DIABETIC FOOT

395

Figure 14 Surgical wound after 5 days with further debridement and amputation of the residual
fifth digit.

Once it is evident that the wound is clean, the metatarsal bones are further resected and a
plantar flap is fashioned to provide soft tissue coverage to the terminal aspect of the foot.
Although at the time of the procedure the tissue may appear viable, there may be evidence
of inadequate wound healing several days or sometimes even weeks after the procedure.
This may be manifest by separation of the wound margins, wound edge necrosis, or even
recurrent infection (Fig. 16). Such a patient should be returned to the operating room for
debridement of the wound, which is a good candidate for vacuum-assisted closure. The
vacuum-assisted closure device is effective in promoting granulation tissue, removing
wound exudate, promoting wound edge contracture, and providing a proper environment
for secondary wound closure (Fig. 17). Frequently diabetic patients facing this clinical
problem are not candidates for arterial revascularization due to the lack of target vessels in
the lower aspect of the extremity, following the typical pattern of distal small vessel disease

i

2

I

■c

Figure 15 Wound after split thickness skin grafting. Wound edges have contracted.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

396

SEABROOK

Figure 16 Transmetatarsal amputation on foot with chronic ischemia with wound edge necrosis
and fibrinous exudate covering the wound surface.

in this population. Although the patient may require several months of wound care and
often more than one return visit to the operating room, persistence will frequently result in
limb salvage.

G. Ascending Infection

The most serious diabetic foot complication involves an ascending infection originating with
involvement of a significant part of the foot. This surgical emergency requires urgent at-
tention. Frequently these infections cause a necrotizing fasciitis that quickly spreads up the
leg and can result in a true life-threatening condition. Patients should be moved promptly to
the operating room and informed consent should be obtained for the possibility of open
below-knee amputation. In the diabetic population, it is not unusual for these patients to be

i

2

I

■c

Figure 17 Transmetatarsal wound after treatment with vacuum dressing. Granulation tissue
covers wound surface. Viable wound edges have contracted.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

MANAGEMENT OF THE DIABETIC FOOT

397

Figure 18 Ankle disarticulation for severe foot infection. Surgical clamps control the anterior and
posterior tibial vessels.

metabolically unstable, with significant hyperglycemia and electrolyte abnormalities. When
it is evident that there is no hope of salvaging the patient's foot, the surgeon should proceed
to perform an ankle disarticulation. This procedure can be executed expeditiously with an
incision that courses just distal to the lateral and medial malleoli, dividing the Achilles
tendon posteriorly, and completing the incision across the dorsum of the foot. From this
approach, the ankle joint space is easily entered and the foot can be removed without
dividing any bone, requiring the incision of a minimal amount of soft tissue with a relatively
limited blood supply (Fig. 18). Usually the infection will have been confined to the foot, and
after a few days of antibiotic therapy, the residual soft tissue contamination will have been
eradicated and the patient may be returned to the operating room for formal below-knee
amputation. On occasion, however, there will be evidence of continued purulent drainage

i

2

I

■c

Figure 19 Level of ascending soft tissue infection is demonstrated with sequential transverse

incisions.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

398

SEABROOK

Figure 20 Guillotine amputation.

from the ankle disarticulation site. To determine the level of involvement, sequential
transverse incisions should be performed above the level of the ankle until there is evidence
of viable tissue free of infection (Fig. 19). At this level a guillotine amputation should be
performed (Fig. 20). Secondary closure is performed when the patient is clinically stable and
the wound free of residual infection.

REFERENCES

10.

Centers for Disease Control and Prevention. National diabetes fact sheet: General information

and national estimates on diabetes in the United States, 2000. Atlanta, GA: U.S. Department

of Health and Human Services, Centers for Disease Control and Prevention, 2002.

Siperstein MD, Unger RH, Madison LL. Studies of muscle capillary basement membranes in

normal subjects, diabetics and prediabetic patients. J Clin Invest 1968; 47:1973.

Krepel CJ, Gohr CM, Edmiston CE. Anaerobic pathogenesis: Collagenase production by Pep-

tostreptococcus magnus and its relationship to site of infection. J Infect Dis 1991; 163:1148.

Centers for Disease Control and Prevention. Staphylococcus aureus resistant to vancomycin —

United States, 2002. MMWR 2002; 51:557-565.

Murray BE. Drug therapy — Resistant enterococcal infections. N Engl J Med 2000; 342(10):

710-721.

Wilson P, Andrews JA, Charlesworth R, Walesby R, Singer M, Farrell DJ, Robbins M. Line-

zolid resistance in clinical isolates of Staphylococcus aureus. J Antimicrob Chemother 2003;

51:186-188.

Winter GD. Formation of scab and the rate of epithelialization on superficial wounds in the

skin of the domestic pig. Nature 1962; 193:292-294.

Glover JL, Weingarten MS, Buchbinder DS, Poucher RL, Deitrick GA III, Fylling CP. A 4-

year outcome-based retrospective study of wound healing and limb salvage in patients with

chronic wounds. Adv Wound Care 1997; 10(l):33-38.

Argenta LC, Morykwas MJ. Vacuum assisted closure: a new method for wound control and

treatment: Clinical experience. Ann Plast Surg 1997; 38(6):563-577.

DeFranzo AJ, Argenta LC, Marks MW, Molnar JA, David LR, Webb LX, Ward WG,

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

MANAGEMENT OF THE DIABETIC FOOT 399

Teasdall RG. The use of vacuum-assisted closure therapy for the treatment of lower-extremity
wounds with exposed bone. Plast Reconst Surg 2001; 108(5):1 184-1 191.

1 1 . Kloth LC. 5 questions and answers about negative pressure wound therapy. Adv Skin Wound
Care 2002; 15(5):226-229.

12. Sherman RA. Maggot versus conservative debridement therapy for the treatment of pressue
ulcers. Wound Repair Regeneration 2002; 10(4):208-214.

13. Bacterial Isolate Reference Bank, Surgical Microbiology Research Laboratory. Department of
Surgery. Milwaukee: Medical College of Wisconsin, 2002.

i
2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

22

Complications of Lower Extremity Amputation

Kenneth E. Mclntyre, Jr.

University of Nevada School of Medicine, Las Vegas, Nevada, U.S.A.

I. INTRODUCTION

Lower extremity amputation is one of the oldest and most commonly performed surgical
procedures and yet has undergone very few modifications since its inception. Despite
advances in limb-salvage surgery, lower extremity amputation is still commonly required
as an end result of the progression of arterial occlusive disease or failed arterial recon-
struction. Furthermore, the increasing prevalence of diabetes has led to the appearance of
more patients with foot complications leading to eventual lower extremity amputation. In
1993, some 98,000 lower extremity amputations were performed in nonfederal acute care
hospitals (1).

Since lower extremity amputation is an ablative procedure and often regarded as a
"simple" one it is often delegated to interns in surgical training programs who may per-
form the operation without appropriate supervision. According to one Veterans Admin-
istration (VA) study, 75% of amputations were performed with a resident as the primary
surgeon (2). Despite the fact that the procedure itself is not technically difficult, com-
plications can and do occur. Moreover, a high surgical mortality rate is associated with ■g
major lower extremity amputation; this is often underappreciated by patients as well &
as physicians unfamiliar with amputation surgery. The most common chronic ailments a
leading to lower extremity amputation are diabetes mellitus and peripheral arterial occlu- c
sive disease (PAOD) (2-5). In a recent review of amputations performed at VA hospitals <j
over a 10-year period, diabetes accounted for 62.9% and PAOD for 23.6% (2). -9

The complications and impressive operative mortality seen with lower extremity 41

amputation are a reflection of the medical comorbidities of this high-risk group of patients. 2

This chapter identifies and discusses the complications that may commonly be encountered |

following major (above-knee or below-knee) lower extremity amputation. @

401 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

402 MCINTYRE

II. MORTALITY

One of the most striking characteristics of lower extremity amputation is the associated
high operative mortality rate. Selection bias in this elderly group of patients with co-
morbid diseases accounts for the high operative mortality (Table 1). The most common
chronic illness in the amputee population is diabetes, accounting for more than half of the
nontraumatic amputations performed yearly (2-5). For example, amputation rather than
revascularization may be selected as the appropriate alternative in an elderly patient who
does not ambulate because of chronic pain or hemiplegia from prior stroke. An operative
mortality rate of 4-16% for below-knee and 12-40% for above-knee amputations has
been well documented (3,5-7). The cause of death is almost always related to the significant
cardiovascular disease that is always present in patients with limb-threatening ischemia
of the lower extremity (3,5,8). Fully two-thirds of perioperative amputation deaths occur as
a consequence of myocardial infarction, congestive heart failure, or stroke (3). Even with
the associated high operative mortality, extensive preoperative cardiac evaluations are
usually unnecessary unless there is also severe ventricular dysfunction, unstable angina, or
recent myocardial infarction.

III. WOUND HEALING FAILURE

The most worrisome complication following amputation at any level is failure of the
amputation stump to heal primarily. Healing failure results in prolonged hospitalization
and delayed rehabilitation; it may even require amputation revision to a more proximal
level. Primary healing occurs in direct relation to the level of amputation — i.e., the more
proximal the level of amputation, the higher the rate of primary healing. However, the
rehabilitation potential of an amputee with a healed below-knee amputation stump far
exceeds that of one with an above-knee stump (9-1 1). For this reason, below-knee ampu-
tations should always be considered as a first alternative in patients who require major
lower extremity amputation (Table 2).

When the amputation level is selected based on a physical examination alone, primary
healing can be anticipated in 80-85% of below-knee and 85-90% of above-knee ampu-
tations (3,12). Failure of the suture line to heal primarily may be attributed to stump
infection, hematoma, or poor nutritional state prior to amputation. However, as a general
rule, the amputation site will heal primarily as long as there is enough blood flow to ensure

Table 1 Preexisting Disease in Patients Undergoing
Lower Extremity Amputation

Disease Percent

Diabetes mellitus 60-82

Ischemic heart disease 26-77

Cerebrovascular disease 20-25

Hypertension 1 5-70

Chronic obstructive pulmonary disease 5-20

Smoking 50-100

Renal failure 10

Source: Data from Refs. 3, 4, 7, 10, 11, 19, 33, and 37.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

LOWER EXTREMITY AMPUTATION 403

Table 2 Complications Occurring After Lower
Extremity Amputation

Complication Percent

Failure to heal 3-28

Infection 12-28

Flexion contracture 1-3

Stump pain/phantom limb pain 5-30

Deep vein thrombosis 5-40

Source: Data from Refs. 3, 5, 7, 8, 12, 13, 19, 22,
28, 29, 33, 35, and 36.

healing at the skin level selected for amputation. Most commonly, failure of the ampu-
tation site to heal primarily often occurs when an inappropriate level for amputation is
chosen without regard to skin perfusion at that level. For this reason, several preopera-
tive tests have been developed to accurately assess the skin perfusion at any given level,
thereby improving the rate of primary healing. These tests range in scope from deter-
mining Doppler-derived systolic ankle pressures to predicting skin blood flow using nu-
clear isotopes (3,13).

Doppler-derived systolic ankle pressures >30 mmHg predict below-knee amputation
in 94% of patients (3,13). This test can easily be performed at the bedside and is highly
reproducible. More involved preoperative tests can be performed using laser Doppler
velocimetry, determining the washout of isotopic agents (Xenon 133) from the skin (14),
or recording the transcutaneous P value at a selected amputation level. The accuracy of
these tests in predicting successful amputation healing varies between 92 and 100% (3,
13,14). One must be aware, however, that systemic as well as local factors in the region
selected for amputation level testing may reduce the reliability of the technetium P test.
Oxygen content and cardiac output are systemic factors and local edema and/or cellulitis
local factors that may adversely influence the reliability of the technetium P test (15).

Eneroth and Persson examined the risk factors for failure of wound healing in 177
consecutive patients undergoing amputation for ischemia. Absence of gangrene as well as a
hemoglobin level >12 g predicted a higher rate of failure. In this review, age, sex diabetes,
level of amputation, previous vascular surgery, smoking, preoperative blood pressure,
serum creatinine, erythrocyte sedimentation rate, blood glucose, and fever had no correla-
tion with ability of the amputation wound to heal (16).

Low serum albumin (<3.5 g) or total lymphocyte counts < 1500/mm3 were both shown
by Dickhaut et al. to adversely affect the healing ability of the amputee (17). The pre-
operative hematocrit may also have an effect on wound healing of amputation stumps. ■§
Hansen et al. documented a higher incidence of wound complications following below-knee g
amputation when the hematocrit was >40% (18). It has generally been accepted that a
patients with diabetes are more prone to wound complications. However, there is really no c
compelling scientific evidence that a difference in healing between those amputees with or <j
without diabetes exists (7,19,20). >S

Finally, the specific technique used for below-knee amputations does not matter as 41

long as standard surgical principles are utilized — i.e., meticulous technique, care not to Q

leave devitalized tissue, strict hemostasis, and atraumatic tissue approximation without |

undue tension. Ruckley et al. compared the results of skew flaps versus long posterior flaps ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

404 MCINTYRE

in a multicenter trial examining the outcomes after below-knee amputation. They found
no significant difference in outcomes based on each surgical technique employed (21).

IV. INFECTION

Infectious gangrene alone rather than ischemia accounts for a significant number of am-
putations each year, especially among patients afflicted with diabetes mellitus. Unfortu-
nately, when infection occurs in an amputation stump, revision of the stump to a higher
level is often required. The incidence of infection in amputation stumps ranges from 12
to 28% and is correlated directly with the indication for amputation (7,19,20,22). When
the amputation is performed following an unsalvageable foot infection, it is not surprising
that the incidence of stump infection is four times greater than that when no foot infection
exists (3). In the case of infectious gangrene of the foot, deep muscle as well as the lym-
phatics that drain the leg are filled with bacteria (23). Performing amputation with primary
closure in this setting will often lead to infection of the stump and the need for subsequent
revision. Mclntyre et al. were the first to describe the benefit of two-stage amputation for
infectious gangrene (20). This retrospective review documented a lower incidence of stump
infection and need for revision in a group that underwent ankle guillotine amputation
followed by definitive amputation at the below-knee level than in a group that underwent
definitive below-knee amputation in one stage. In a related study, Desai et al. documented
similar outcomes (24). In a prospective study designed to assess the benefit of two-stage
amputation, Fisher et al. confirmed the benefits of two-stage amputation (23).

The use of prophylactic antibiotics was also found to be of benefit in reducing the risk
of infection in the amputation stump. Sonne-Holm et al. compared preoperative cefoxitin
to placebo in patients undergoing amputation for ischemia (25). They reported a wound
infection rate of 16.9% in the antibiotic group compared to 38.7% in the placebo group.
Norlin et al. employed prophylactic cefotaxime in 19 patients prior to amputation and
compared it to placebo. They reported similar beneficial results, with 83% achieving heal-
ing in the antibiotic group compared to only 59% in the placebo group (26).

Rubin et al. documented an increased incidence of infection in the stump after below-
knee amputation when a thrombosed prosthetic graft was left in place during the amputa-
tion procedure (27). When a thrombosed prosthetic graft is encountered during amputation,
this complication can be avoided by retracting it the graft, carefully, transecting and
allowing it to retract well away from the stump closure. A segment of the most distal
remaining graft should be taken and sent for bacterial culture to ensure that no portion of
retained graft is infected at the time of amputation.

V. PHANTOM LIMB PAIN

1
Chronic stump pain that occurs following lower extremity amputation can be difficult to g

treat. The incidence of chronic stump pain or phantom limb pain following major lower a

extremity amputation is between 5 and 30 (22,28). The etiology of chronic pain in the -c

amputation stump is often unclear but may pose a significant problem that limits rehabili- <j

tation. The patient who develop, phantom limb pain following amputation may be so >9

debilitated that he or she will not be able to walk with a prosthesis despite successful 4j

healing of the amputation stump. Phantom limb pain must be differentiated from a 2

"phantom sensation," which is a feeling that the amputated limb is still present. True |

phantom limb pain is often described as "burning, aching, squeezing, and knife-like" often @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

LOWER EXTREMITY AMPUTATION 405

resembling the pain that the patient may have experienced prior to the amputation (3,
28,29). Phantom pain is persistent and can be exacerbated by proximal or remote stimuli.
Moreover, phantom limb pain can be improved by changing the somatic input (28-30).
Any stump pain that persists beyond 6 months following the amputation will be extra-
ordinarily difficult to treat and will likely never resolve (30). Although there is no consensus
on the best treatment for phantom limb pain, tricyclic antidepressants have been used suc-
cessfully (28-30).

VI. FLEXION CONTRACTURES

Although it is not uncommon for flexion contractures to occur following lower extremity
amputation, proper postoperative care will almost always avoid this problem. Unless care-
ful attention is directed to prevention, flexion contractures following lower extremity
amputation can be directly responsible for failure of rehabilitation. Flexion contractures
should not be regarded as an insignificant problem, because without proper attention,
successful rehabilitation cannot occur despite an otherwise well-healed amputation stump.
Flexion contractures are due to disuse of a joint, because of major lower extremity ampu-
tation, and are more common in elderly amputees. The incidence of flexion contractures
ranges from 1 to 3% and is more common in patients above 80 years of age and in those who
have had a prior stroke on the same side as the amputation (3,7,19,31). The unopposed
muscles responsible for hip and knee flexion after above-knee and below-knee amputation,
respectively, cause these contractures. Strengthening and range-of-motion exercises direc-
ted by physical therapy preoperatively help to prepare the amputee for the physical
requirements he or she will face following successful amputation. At the time of surgery,
placing the below-knee amputation stump in a well-padded protective rigid dressing with
the knee in slight flexion will help prevent flexion contractures of the knee (7,19,31). The
above-knee amputee will tend to develop flexion contractures of the hip joint. Placing the
patient in a prone position at regular intervals during the day will help to prevent flexion
contractures of the hip. Standard daily physical therapy with range-of-motion exercises will
also help to strengthen the patient and reduce the incidence of flexion contractures.

VII. INABILITY TO AMBULATE

Failure to ambulate with a prosthetic limb following an otherwise successful lower
extremity amputation is attributable to several factors. First and most important is the
patient's preamputation ambulatory status (3,5,7,32,33). Clearly, if a patient who is unable
to walk undergoes lower extremity amputation, one cannot realistically expect that he or
she will be able to walk following a successful lower amputation. All too often, patients

who have been crippled by stroke or are disabled from severe arthritis require ampu- •§

tation and cannot successfully be rehabilitated, since they were unable to walk prior to the g

amputation. In addition, any cognitive impairment will have a negative influence on the a

ability to successfully complete rehabilitation. Similarly, if the preamputation medical c

condition is severe (e.g., congestive heart failure), the amputee will not usually be able to <j

successfully rehabilitate, since more energy is needed to ambulate with a prothetic leg >3

(5,32,33). That is, the amputee with limited cardiac reserve may not be able to generate the 4j

cardiac energy required to ambulate with a prosthetic leg. In these circumstances, it is really 2

the patient's preamputation physical condition that determines his or her ability to achieve |

ambulation following lower extremity amputation. It is important to advise the patient @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

406 MCINTYRE

undergoing amputation of the potential to fail rehabilitation based on the significance of
comorbid illnesses.

VIM. DEEP VENOUS THROMBOSIS/PULMONARY EMBOLISM

The risk of deep venous thrombosis (DVT) and/or pulmonary embolism following lower
extremity, amputation is not insignificant. However, DVT is often overlooked because
swelling and pain, symptoms of DVT, often accompany lower extremity amputation. Of
even greater concern is that DVT involving major veins may produce no symptoms.
Therefore a high index of suspicion is needed to diagnose and treat DVT while reducing
the risk of pulmonary embolism. Amputees are at increased risk over the conventional
postoperative patient for several reasons. First, prolonged bed rest prior to and following
amputation poses the risk of stasis. Second, vascular surgery patients are known to have
an increased risk of hypercoagulability (34). Third, there is interruption (endothelial
injury) of the major veins during the normal amputation procedure. It should come as no
surprise, therefore, that the incidence of DVT occurring after lower extremity amputation
varies between 5 and 40% (22,35). Yeager et al., in a prospective review of patients under-
going lower extremity amputations, reported no difference in the incidence of DVT be-
tween those who underwent above-knee versus below-knee amputations (35). Moreover, if
DVT was present, it was diagnosed in the majority of patients (67%) preoperatively. If
routine duplex venous scans is performed on amputees in the perioperative period, the
incidence may well be higher. Fortunately, the risk of pulmonary embolism is significantly
less than the risk of DVT. The incidence of pulmonary embolism following lower extrem-
ity amputation ranges between 1 and 5% (3). Certainly, if DVT is detected preoperatively,
conventional intravenous heparin therapy should be used for several days prior to ampu-
tation to allow for some stability of the thrombus on the vein wall.

IX. LONG-TERM OUTLOOK

For elderly patients who have required lower extremity amputation as a consequence of
diabetes and/or peripheral arterial disease, the long-term prognosis is not good. The mean
survival of a lower extremity amputee is between 2 and 5 years (36). In patients with
diabetes, the outlook for contralateral limb loss is even worse. Following a lower extremity
amputation, 28-51% of the patients with diabetes will require amputation of the remain-
ing leg within 5 years (37). A 5-year mortality rate of between 39 and 68% can be anti-
cipated following lower extremity amputation in the patient with diabetes (37).

X. SUMMARY AND RECOMMENDATIONS 1

&

Lower extremity amputation is associated with a significant risk of perioperative mortality a

and complications. Those patients who undergo amputation have significant comorbid c

medical conditions that contribute to the risk of mortality and complications. Patients <j

with diabetes and PAOD make up the greatest cohort of patients who undergo am- >9

putation of the lower extremity, and these risk factors are often responsible for the 4j

complications that may occur. Appropriate preoperative planning may help to reduce the 2

complication rate. If an amputation is required to treat infectious gangrene, two-stage |

amputation should be performed to reduce the risk of stump infection, which may require ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

LOWER EXTREMITY AMPUTATION 407

revision to a higher level. When amputation is performed for ischemia, prophylactic anti-
biotics help to reduce the perioperative infection rate. Healing at the below-knee level
should be the goal for any amputation that occurs above the ankle. If there is doubt
whether an amputation will heal at a given level, several tests are available to evaluate skin
perfusion. Appropriate preoperative physical therapy will help with rehabilitation out-
comes. Attention to detail and appropriate pre- and postoperative care will help to avoid
complications and improve rehabilitation potential following lower extremity amputation.

REFERENCES

1. Leonard JA Jr, Meier RH. Upper and lower extremity prosthetics. In: DeLisa JA, ed. Reha-
bilitation Medicine: Principles and Practice. 3rd ed. Philadelphia: Lippincott-Raven, 1998.

2. Mayfield JA, Reiber GE, Maynard C, et al. Trends in lower limb amputation in the Veterans
Health Administration, 1989-1998. J Rehabil Res Devel 2000; 37.

3. Malone JM, Ballard JL. Complications of lower extremity amputation. In: Bernhard VM,
Towne JB, eds. Complications in Vascular Surgery. St Louis: Quality Medical Publishing,
1991:313-329.

4. Most RS, Sinnock P. The epidemiology of lower extremity amputations in diabetic individuals.
Diabetes Care 1983; 6:87.

5. Coletta EM. Care of the elderly patient with lower extremity amputation. J Am Board Fam
Pract 2000; 1323-34.

6. Berardi RS, Keonin Y. Amputations in peripheral vascular occlusive disease. Am J Surg 1978;
135:231-234.

7. Malone JM, Moore WS, Goldstone J, et al. Therapeutic and economic impact of a modern
amputation program. Ann Surg 1979; 189:798-802.

8. Yekutiel M, Brooks ME, Ohry A, et al. The prevalence of hypertension, ischemic heart
disease, and diabetes in traumatic spinal cord injured patients and amputees. Paraplegia 1989;
27:57.

9. Evans WE, Hayes JP, Vermilion BD. Rehabilitation of the bilateral amputee. J Vase Surg 1987;
5:589.

10. Roon AJ, Moore WS, Goldstone J. Below-knee amputation. A modern approach. Am J Surg
1977; 134:153.

11. Steinberg FU, Sunwool 1, Roettger RF. Prosthetic rehabilitation of geriatric amputee patients:
A follow-up study. Arch Phys Med Rehabil 1985; 66:742.

12. Burgess EM, Matsen FA, Wyss CR, et al. Segmental transcutaneous measurements of P 0i in
patients requiring below the knee amputation for peripheral vascular insufficiency. J Bone Joint
Surg 1982; 64:378.

13. Durham JR. Lower extremity amputation levels: Indications, methods of determining appro-
priate level, technique, and prognosis. In: Rutherford RB, ed. Vascular Surgery. 3rd ed.
Philadelphia: Saunders, 1989.

14. Malone JM, Leal JM, Moore WS, et al. The "gold standard" for amputation level selection: •o
Xenon-133 clearance. J Surg Res 1981; 30:449. §

15. Lalka SG, Malone JM, Anderson GG, et al. Transcutaneous oxygen and carbon dioxide 8
pressure monitoring to determine severity of limb ischemia and to predict surgical outcome. •§)
J Vase Surg 1988; 7:507. |

16. Eneroth M, Persson BM. Risk factors for failed healing in amputation for vascular disease: y
A prospective, consecutive study of 177 cases. Acta Orthop Scand 1993; 369. g

17. Dickhaut SC, DeLee JC, Pate CP. Nutritional status: Importance in predicting wound-healing S
after amputation. J Bone Joint Surg 1984; 66A:71. °

18. Hansen ES, Wethelund JD, Skajaa K. Hemoglobin and hematocrit as risk factors in below- I
knee amputations for incipient gangrene. Arch Orthop Trauma 1988;10792. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

408 MCINTYRE

19. Malone JM, Moore WS, Leal JM, et al. Rehabilitation for lower extremity amputation. Arch
Surg 1981; 116:93.

20. Mclntyre KE, Bailey SA, Malone JM, et al. Guillotine amputation in the treatment of non-
salvageable lower extremity infections. Arch Surg 1984; 1 19:450^453.

21. Ruckley CV, Stonebridge PA, Prescott RJ. Skewflap versus long posterior flap in below-knee
amputations: Multicenter trial. J Vase Surg 1991; 13:423.

22. Gottschalk FA, Fisher DF Jr. Complications of amputation. In: Rutherford RB, ed. Vascular
Surgery. 5th ed. Philadelphia: Saunders, 2000.

23. Fisher DF Jr, Clagett GP, Fry RE, et al. One-stage versus two-stage amputation for wet
gangrene of the lower extremity: A randomized study. J Vase Surg 1988; 8:428-433.

24. Desai Y, Robbs JV, Keenan JP. Staged transtibial amputations for septic peripheral lesions due
to ischemia. Br J Surg 1986; 73:392.

25. Sonne-Holm S, Boeckstyns M, Menck H, et al. Prophylactic antibiotics in amputation of the
lower extremity for ischemia. A placebo-controlled, randomized trial of cefoxitin. J Bone Joint
Surg 1985; 67:800-803.

26. Norlin R, Fryden A, Nilsson L, et al. Short-term prophylaxis reduces the failure rate in lower
limb amputations. Acta Orthop Scand 1991; 62:509.

27. Rubin JR, Yao JST, Thompson RG, et al. Management of infection of major amputation
stump following failed femoro-distal grafts. Surgery 1985; 98:810.

28. Sherman RA, Sherman CJ, Parker L. Chronic phantom and stump pain among American
veterans. Results of a survery. Pain 1984; 18:83.

29. Jensen TS, Brebs B, Nielsen J, et al. Immediate and long-term phantom limb pain in amputees:
Incidence, clinical characteristics and relationships to preamputation limb pain. Pain 1985; 21:
267.

30. Esquenazi A, Meier RH III. Rehabilitation in limb deficiency. Arch Phys Med Rehabil 1996;
77(suppl):18.

31. Mooney V, Harvey JP Jr, McBride E, et al. Comparison of postoperative stump management:
Plaster vs. soft dressings. J Bone Joint Surg 1971; 53A:241.

32. Cutson TM, Bongiorni DR. Rehabilitation of the older lower limb amputee: A brief review.
J Am Geriatr Soc 1996; 44:1388.

33. Leung EC-C, Rush PJ, Devlin M. Predicting prosthetic rehabilitation outcome in lower limb
amputee patients with the functional independence measure. Arch Phys Med Rehabil 1996;
77:605.

34. Taylor LM Jr, Chitwood RW, Dalman RL, et al. Antiphospholipid antibodies in vascular
surgery patients: A cross-sectional study. Ann Surg 1994; 220:544.

35. Yeager RA, Moneta GL, Edwards JM, et al. Deep vein thrombosis associated with lower
extremity amputation. J Vase Surg 1995; 22:612.

36. Pernot HF, de Witte LP, Lindeman E, et al. Daily functioning of the lower extremity amputee:
an overview of the literature. Clin Rehabil 1997; 11:93-106.

37. National Diabetes Data Group. Diabetes in America. 2d ed. Bethesda, MD: National Institutes
of Health, 1995.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

23

Complications of Vascular Access

Mark B. Adams, Christopher P. Johnson, and Allan M. Roza

Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

Currently more than 280,000 patients are maintained on chronic hemodialysis in the United
States. Approximately 27,571 other patients undergo acute dialysis or hemofiltration in the
treatment of a variety of conditions ranging from acute renal failure to drug overdose (1,2).
Obtaining vascular access, temporary or permanent, in order to institute treatment in these
patients requires the implementation of prosthetic devices, specific knowledge, and careful
monitoring. Many of the complications related to vascular access are specific to the access
technique, device, and site.

I. TEMPORARY ACCESS

Temporary access techniques use external prosthetic devices, such as Mahurkar catheters,
that are not intended to be used on a permanent basis. With the advent of these temporary
venous access devices, previous techniques, such as Scribner shunts and their variants,
have fallen into disuse.

The most common form of temporary vascular access currently in use is the percuta-
neous subclavian or femoral dialysis catheter. Femoral catheters are not used as often as
subclavian catheters because they are associated with a high rate of infection and throm-
bosis and are difficult to immobilize. Femoral catheters are often placed at the beginning of
each dialysis treatment and removed at the end or after several treatments. Their use may
lead to iliofemoral venous thrombosis, with its attendant risk of pulmonary embolism, and
may also be the cause of significant retroperitoneal hemorrhage. The major advantage of

using femoral catheters is avoidance of the risks associated with subclavian and jugular c

punctures. Iliofemoral thrombosis may become particularly problematic if and when the <j

patient comes to renal transplantation. >9

Subclavian venous catheters remain a popular choice for temporary access. Subcla- %

vian catheters are introduced using standard subclavian technique over a guidewire. These Q

are usually double-lumen catheters with the two lumens separated by 3-4 cm. The double- |

lumen catheter allows the patient to be dialyzed using the standard two-needle machine @

409 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

410 ADAMS et al.

technique. Single-lumen catheters require a different technique, which generally gives a
lower clearance rate and causes an increased recirculation effect. Because subclavian
temporary (Mahurkar) catheters are associated with a high incidence of subclavian vein
stenosis or occlusion, they should be avoided whenever possible. A better choice is an
internal jugular catheter directed into the superior vena cava. One disadvantage of this
placement is that the catheter protrudes near the patient's ear, making it annoying and
difficult to dress.

Temporary dialysis catheters are available in a variety of models but have several
characteristics in common. They must be of sufficient caliber to support the high flow rates
required for adequate hemodialysis (greater than 200 mL/min) and rigid enough to prevent
collapse when drawing pressures become high (3). The most commonly used temporary
catheter is the Quinton-Mahurkar polyurethane double-lumen catheter (Quinton Instru-
ment Co., Seattle, WA).

With careful nursing and patient management, these catheters may be left in place for
4-6 weeks, which is usually sufficient time to construct a permanent fistula.

There has been an increasing incidence of patients starting hemodialysis using a central
venous catheter. In addition, a large percentage of patients are dependent on temporary
catheters while obtaining permanent access at various times during their dialysis course
(USRDS data).

Because Mahurkar catheters do not have a Dacron cuff, which functions as a barrier to
bacterial colonization, they are associated with a high infection rate if left in place for more
than a short time. Cuffed catheters (see below) provide longer and more reliable central
venous access.

As the importance of reliable vascular access has become well accepted, a number of
major efforts to promote a rational approach to the placement and maintenance of such
access have been made. Notable have been the DOQI Guidelines of the National Kidney
Foundation. This document stands as the best reference to vascular access in patients with
end-stage renal disease (see clinical practice guidelines at www.kidney.org).

A. Complications

1. Infection

Infection represents the most common complication of temporary vascular access and re-
quires the removal or replacement of the catheter. Infection will eventually occur in most
percutaneously placed temporary access catheters. The clinical impact of infection related
to the catheter can be minimized by early removal whenever signs of fever, pericatheter
exudate, or erythema appear. In situations where the access site is of great clinical im-
portance and another access site would be difficult to obtain, replacement of the catheter
over a guidewire and careful observation with antibiotic coverage may allow continued use ■§

of the site for at least a short time. g

As an additional barrier to infection, some dialysis catheters are manufactured with a a

proximal cuff (e.g., PermCath; Quinton Instrument Co., Seattle, WA). The cuff becomes c

incorporated into the subcutaneous tissue and retards the migration of bacteria along the <j

catheter surface. The incidence of infection appears to be reduced with this modification, but >9

these catheters remain prone to thrombosis and failure as a result of fibrin sheath formation 4j

at the tip. In general, their use is limited to patients in whom all other routes for permanent 2

vascular access have been exhausted or in those patients in whom permanent access will not |

be usable for 6-8 weeks. The presence of a cuffed catheter lowers the risk of infection and ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR ACCESS 411

usually provides reliable access at least until a permanent access site can be developed. These
catheters have also been advocated for permanent use. In our hands, they have not worked
well in this role. Other reports are more optimistic (4,5). However, they do not achieve flow
rates or dialysis adequacy equal to that of peripheral access (6).

Cuffed subclavian dialysis catheters are often placed percutaneously using large dila-
tors and peel-away sheaths. The size (20F) and rigidity of these sheaths are troublesome,
however. Disruption of the subclavian veins and/or the superior vena cava have occasion-
ally occurred. The safest way to place such catheters is using ultrasound imaging for initial
placement with the additional use of fluoroscopy. Patients with a listing of multiple previous
catheters are at significantly increased risk of complications and should have placement in a
well-equipped interventional suite.

Internal jugular placement of both uncuffed and cuffed central catheters for dialysis
has significantly lowered the incidence of central venous stenosis and/or occlusion. At pres-
ent, most patients have a previous history of central venous catheters and up to 30% will
have either central venous stenosis or occlusion (7). Because of this, further central cath-
eter placement is best done with access to both ultrasound and angiography.

The way in which a subclavian dialysis catheter is cared for and flushed will, in many
cases, determine its useful life span. Temporary subclavian catheters should be handled in
an aseptic manner and flushed routinely with high-dose heparin solution (5000 U/mL). Care
must be taken to aspirate the residual heparin before using the catheter for any purpose
other than dialysis. Often the patient or a relative can handle flushing of the catheter, but in
most cases qualified medical personnel should be responsible for the dressing change, since
this requires knowledge and practice of sterile technique.

Recently, the LifeSite Hemodialysis Access System has been introduced as an alter-
native to cuffed catheters. The LifeSite is an implantable port (one or two valves), that can be
accessed by subcutaneous puncture with a 14-gauge needle. In initial trials, this device was
associated with lower infection rates than percutaneous catheters. The luminal diameter is
slightly larger than that of conventional catheters, which also increases the effective flow rate
on dialysis. However, this has raised concerns regarding central vein thrombosis (8,9).

2. Bleeding

The surgeon performing vascular access will undoubtedly become involved in situations
where uremic patients experience excessive and prolonged bleeding. Most commonly, this
occurs following percutaneous placement of access catheters.

Bleeding is more likely with temporary hemodialysis catheters of the uncuffed type
than with other central venous catheters because of their size and rigidity and the manip-
ulation that inevitably occurs. Bleeding can be minimized by careful handling of the catheter
during dialysis. When bleeding occurs, it can usually be controlled by local pressure and
correction of any coagulation abnormalities. ■g

Uremic patients have a well-known tendency toward increased bleeding (10,11). The &

precise derangements are not entirely understood but predominantly relate to primary a

hemostasis (i.e., platelet-vessel interaction and platelet aggregation). The single best test -c

that quantitatively measures this platelet dysfunction is the skin bleeding time. Other <j

hemostatic defects that have been identified include abnormally increased production of >9

prostaglandin I 2 by the uremic vessel wall, decreased production of platelet thromboxane 4j

A 2 , changes in the von Willebrand factor (vWf) molecule, and decreased platelet factor III. 2

Presumably, the platelet dysfunction in uremia allows improved patency of vascular |

access. Observations that support this theory include the well-known propensity of fistulas ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

412 ADAMS et al.

to clot following successful renal transplantation and the generally accepted difficulty of
maintaining vascular access in preuremic individuals.

Platelet transfusions will have little beneficial effect in these circumstances since, in the
uremic milieu, normal platelets acquire the characteristics of uremic platelets. Cryopreci-
pitate is a plasma derivative and is a good source of factor VIII and vWf. Infusion of
cryoprecipitate shortens the bleeding time of uremic patients, with the peak effect seen
4-6 h after infusion.

Desmopressin (DDAVP) is a synthetic derivative of an antidiuretic hormone and acts
to increase plasma concentrations of vWf by releasing it from endothelial storage sites
(12). Administration of DDAVP intravenously (0.3 /kg) temporarily corrects prolonged
bleeding time and has a more rapid onset of action (peak effect, 1 h) than cryoprecipitate.
DDAVP can also be given intranasally or subcutaneously at 10 times the intravenous
dose.

Conjugated estrogens can normalize bleeding time for 3-10 days in patients with
chronic renal failure. The mechanism of this effect is unknown but may relate again to an
alteration of vWf. A single oral dose of estrogen (Premarin 25 mg) or intravenous doses
(3 mg/kg) divided over 5 consecutive days may be effective and long-lasting (13).

Whenever bleeding occurs that is not easily controlled by pressure or is of high vol-
ume, the possibility of major vessel injury and/or coagulopathy should be considered. A
bleeding time, prothrombin time, and partial thromboplastin time should be obtained and
corrected with appropriate therapy. Most patients on dialysis have a significant qualitative
platelet defect. Even if platelet counts are adequate, DDAVP or cryoprecipitate will often
reduce bleeding that is otherwise difficult to control. Bleeding that occurs following dial-
ysis or catheter flushing is often related to the effect of heparin and can be corrected by
protamine. Heparinized patients usually respond to protamine. As in other surgical sit-
uations, however, it is unwise and unsafe to attribute ongoing bleeding to coagulopathy
alone. If bleeding fails to come under control and the patient is in jeopardy, operation or
interventional radiological approaches are indicated.

Major hemorrhage related to vessel injury is rare and most often occurs as a com-
plication of venous perforation. Catheter removal, application of pressure, and correction
of coagulation defects is the usual treatment. If the bleeding is vigorous and the catheter has
been in place for some time, the possibility that injury or erosion into an artery has occurred
should be considered. Several units of blood can easily be lost into the pleural space or
retroperitoneum with few outward signs. Rarely, these patients require surgery to control
arterial or venous lacerations. Operation should not be delayed if bleeding cannot be readily
controlled.

3. Pneumothorax/Hemothorax

When the procedure is performed by experienced physicians hemothorax and pneumo- ■g

thorax are relatively uncommon complications of subclavian placement of temporary &

access (14). When it is performed by those with little experience or in patients with central a

venous problems from previous catheters, it is doubly dangerous because these compli- c

cations are not only more likely to occur but also are often not immediately recognized. A <j

postplacement chest x-ray should be obtained both to locate the catheter and to rule out >9

hemorrhage or pneumothorax. These complications should be suspected whenever a pa- 41

tient complains of shortness of breath and/or when hypotension develops after catheter 2

placement. The postplacement film should be carefully analyzed to make sure that the |

catheter has not transgressed the vessel wall into the pleural space. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR ACCESS 413

The diagnosis can be difficult, since many of these catheters are placed just before the
initiation of dialysis and patients may experience brief episodes of hypotension during
hemodialysis as a result of reactions to materials in the dialysis system or to fluid shifts.
The chance of a major hemorrhage is also increased in this setting because the patient is
often systemically anticoagulated during dialysis. When hemothorax or pneumothorax is
suspected, it should be treated by removal of the catheter and placement of a chest tube.
X-ray film documentation should not delay appropriate treatment.

Catheter perforation of the superior vena cava is best treated by placing the patient in
an upright position, inserting a chest tube, and removing the catheter. Hypotension should
be rapidly corrected before placing the patient in an upright position, since this positional
change can be dangerous in the hypovolemic patient.

4. Thrombosis

Catheter Thrombosis. Thrombosis of the catheter is a frequent problem encountered
in patients with temporary central venous access. It can be treated by aggressive flush-
ing with normal saline solution; enzymatic digestion with streptokinase, urokinase, or
tissue plasminogen activator (TPA); or reestablishment of a lumen with a large guidewire.
If these measures fail, the catheter can often be stripped angiographically via a femoral
approach using a looped stripping device. If this is unsuccessful, the catheter will need
replacement.

Subclavian Thrombosis. Acute subclavian thrombosis is not uncommon in patients
with temporary access because of the size and mechanical properties of these catheters. If
subclavian thrombosis/stenosis occurs, the catheter should be removed and another route
of dialysis used. The administration of anticoagulation therapy (heparin) in this setting is
controversial but should be considered. Pulmonary embolism in dialysis patients appears
to be extremely rare. Patients with massively swollen arms may benefit symptomatically
from a short course of heparin.

Replacement of the subclavian catheter on the opposite side may be unwise in such
situations, since occlusion of both subclavian veins may lead to a superior vena cava syn-
drome. Also, bilateral subclavian thrombosis severely limits future choices for the site
of permanent access but unfortunately may be unavoidable. When repeated thromboses
occur, a primary hypercoagulable state should be suspected (e.g., antithrombin III defi-
ciency, antiphospholipid syndrome, or protein C or S deficiency) (15). Such individuals
frequently have associated heavy proteinuria (16). Antithrombin deficiencies can be effec-
tively treated by administering fresh frozen plasma at the time of access placement and until
conversion to long-term warfarin (Coumadin) therapy is complete.

With the increased use of subclavian dialysis catheters for initial vascular access in the
patient presenting for dialysis, the incidence of subclavian thrombosis has increased, and "g

may be as high as 30%. Creation of an arteriovenous fistula (AVF) distal to an occluded &

subclavian vein usually results in marked edema of the extremity and the eventual necessity a

of fistula takedown. Since subclavian dialysis catheters have been used previously in many -c

patients currently referred for permanent access placement, the access surgeon must doc- <j

ument that the subclavian vein is patent before proceeding with fistula formation. Ver- >9

ification of patency is most easily accomplished with duplex ultrasonography and can also 4j

be obtained with venography. The latter is more accurate than ultrasound, which has only 2

an 80% sensitivity for proximal central vein obstruction (17). In the current milieu of ready |

placement of subclavian access, the importance of this step cannot be overemphasized. ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

414 ADAMS et al.

5. Pulmonary Embolism

Pulmonary embolism appears to be rare in uremic patents because of their platelet dys-
function, but it should be suspected in any patient with a long-standing indwelling central
venous or femoral catheter who experiences symptoms such as dyspnea or chest pain.
Pulmonary embolism should also be suspected when thrombectomy of venous access cath-
eters or a polytetrafluoroethylene (PTFE) graft is associated with dyspnea, hemoptysis, or
deteriorating pulmonary function. The small size of these emboli makes death from a single
embolus unlikely. However, repeated thrombectomy declotting could potentially produce a
volume of embolus sufficient to produce chronic pulmonary hypertension.

II. PERIPHERAL ACCESS

Permanent vascular access is any type of access used frequently and repeatedly for an
extended or indefinite period of time. Forms of permanent access include both autogenous
AVFs and fistulas constructed with prosthetic material (PTFE, umbilical vein, Dacron) (18).

An autogenous AVF should be constructed even in patients undergoing peritoneal
dialysis. Patients on peritoneal dialysis may have episodes of peritonitis or other complica-
tion, that temporarily interrupt their treatment. The average time patients who are started
on peritoneal dialysis are able to be adequately managed via this modality is usually not more
than a year or two. For most patients, peritoneal dialysis will not be their long-term means of
renal replacement therapy. Because reliable autogenous access becomes increasingly difficult
with time, all patients presenting for dialysis should be evaluated for autogenous access
placement unless they can rapidly proceed to living donor transplantation. In addition, since
patients with end-stage renal disease (ESRD) are experiencing increased longevity, cephalic
veins are frequently used for intravenous therapy, usually resulting in thrombosis. The best
time for fistula placement is when the patient initially presents with renal failure.

With more patients in renal failure presenting without adequate vessels available for
autogenous fistula construction, the use of synthetic material has increased (19). The most
common synthetic materials currently in use for access are PTFE (Gore-Tex, W. L. Gore
and Associates, Inc., Elkton, MD; Impra, Impra, Inc., Tempe, AZ), tetrafluoroethylene
(Medtronic, Minneapolis, MN; etc.), and tanned human umbilical vein (Biograft). PTFE
is overwhelmingly the most commonly used material. Dacron has not proved useful
because of the associated high rate of infection and pseudoaneurysm formation.

Theses materials are readily available and most vascular surgeons have experience with
them in peripheral vascular reconstruction. Such fistulas have the advantage of being large
and therefore easily palpated and cannulated. They can also be used soon after placement
(20,21). This is particularly useful when a patient with advanced uremia is referred for urgent
access, which has increasingly become the case. Synthetic grafts have their own significant
and somewhat unique problems, however (22). E

s

3

A. Choice of Access Location and Type — Autogenous g>

Permanent autogenous access is preferably placed at the wrist in the nondominant arm. <j

This allows use of the dominant arm during dialysis and may lessen the chance of trauma >9

to the fistula during daily activities. Unfortunately, in many patients, the cephalic vein at 4j

one or both wrists is unavailable for use, usually as the result of injudicious placement of 2

intravenous lines by the time the patients are seen by a vascular surgeon. In these patients, |

obtaining autogenous vascular access becomes a challenge. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR ACCESS 415

If neither wrist has a patent cephalic vein, the antecubital space is the next best site. An
AVF at this site is constructed between the brachial artery and the cephalic vein or between
the brachial artery and a connecting branch, such as the median antecubital vein. It is
important to disconnect the new fistula from the deep venous system by ligation of deep
veins going into the forearm to prevent the subsequent development of venous hyper-
tension. This preferentially directs blood up the upper arm cephalic vein, which becomes the
cannulation site.

Fistulas constructed between the brachial artery and the basilic vein or another deep
vein are generally not successful, since dialysis personnel cannot consistently gain access to
the deep venous system. In addition, because the vein is adjacent to the brachial artery,
cannulation carries a risk of arterial injury.

Occasionally, the basilic vein in the upper arm can be transposed superficially by
dissecting it out for a distance of 10-15 cm and then bringing it into a superficial position,
where it is easy to cannulate. This procedure may lead to wound complications (seroma,
hematoma, arm swelling) because of the extent of dissection required and the rich lymphatic
network around the brachial and axillary arteries.

While upper arm cephalic vein fistulas have the advantages of high flow and earlier
maturity, they may be difficult to cannulate reliably in obese patients.

When there are no usable veins in the arms for fistulas, a decision must be made
whether to continue to pursue autogenous access or to construct access with prosthetic
material. Although it is almost always easier to use prosthetic grafts, their useful life is not
as good as that of autogenous access (23). Prosthetic AVFs generally last half as long and
require twice as many interventions as autogenous AVFs.

Autogenous access occasionally can be constructed using the saphenous vein. It can be
mobilized completely to the knee and then looped subcutaneously on the anterior surface
of the thigh and anastomosed to the superficial femoral artery just distal to the bifurcation
of the common femoral artery. Alternatively, in a thin patient, the saphenous vein can be
left in situ and anastomosed end to side to the distal superficial femoral artery at the level
of the adductor canal. It is unnecessary to ligate venous branches, since their presence only
increases the venous outflow from the fistula. This particular procedure can easily be
accomplished with the patient under local anesthesia and provides good length for easily
accessible cannulation sites in a thin patient. The saphenous vein can also be harvested and
transplanted to the arm for use as autogenous fistula graft material.

B. Prosthetic Arteriovenous Fistula

In the event that an autogenous fistula cannot be constructed, most surgeons prefer PTFE
for prosthetic material. PTFE has generally outperformed most other prosthetic materials
in terms of long-term patency. Many different sizes and shapes are currently available,
including straight, tapered, stepped, and externally reinforced grafts as well as grafts with
a hooded configuration at the venous end. No single form of PTFE has demonstrated

clear superiority over others (24-26). Thicker-walled grafts are generally preferable be- c

cause of their increased durability with repeated cannulations. <j

Forearm constructions include straight grafts (radial artery to antecubital vein) and >3

loops (brachial artery looping back into antecubital vein). Loop grafts have the theoretical 4j

advantage of higher flow because of the larger artery on which they are based. However, 2

this is a theoretical consideration, and loop grafts have not been proved to be superior |

over straight grafts. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

416 ADAMS et al.

Straight grafts have the advantage that, in the event of infection, graft removal is safer,
since the radial artery can frequently be ligated without placing the hand in jeopardy.
Ligation of the brachial artery will often result in tissue loss, especially in diabetic patients
with extensive upper extremity peripheral vascular disease.

Upper arm loop prosthetic fistulas are based on the more proximal brachial artery and
generally return into the basilic or deep venous system. Sites lower in the forearm should be
selected first, since, with each successive access placement, the vein distal to the arterio-
venous anastomosis is ligated to prevent venous hypertension. Therefore an upper arm loop
generally precludes further attempts distally in the same arm.

When an upper arm loop PTFE fistula is being created, it is generally easier to perform
both venous and arterial anastomoses at the same level (through the same incision medially
on the upper arm). We usually do this with the venous anastomosis directed toward the
heart and the arterial anastomosis situated deep to it. There may be some advantage in
performing end-to-end venous anastomoses, although this has not been proved.

In general, it is unwise to cross joints with prosthetic graft material. Prolonged flexion
at the joint, such as that which occurs during sleep, may lead to graft occlusion. However,
use of spiral-wrapped reinforced PTFE can help to prevent kinking. Unfortunately, spiral-
reinforced PTFE has an increased tendency to erode through the skin especially if placed
too superficially or in thin elderly patients.

If access cannot be obtained in either upper extremity, more exotic maneuvers will
need to be undertaken. It behooves the surgeon performing access to frequently remind his
medical colleagues of the importance of preserving cephalic veins in those patients with
renal disease, because once the cephalic vein is lost, obtaining good autogenous access
becomes difficult.

C. Complications

1 . Problems Associated with Cannulation

Because the technique for cannulation of synthetic grafts is different from that of au-
togenous AVFs, dialysis units with less experienced personnel have a high rate of perigraft
hematomas, pseudoaneurysm formation, and graft infection.

Perigraft hematoma results from cannulation with the needle entering parallel to the
graft and lacerating a portion of the graft wall. The result may be external bleeding or
hematoma. Graft infection occurs when less than ideal aseptic technique is used during
cannulation and dialysis. Proper technique involves puncturing the graft perpendicular to
the long axis of the graft and then changing the angle of the needle once its tip has entered
the graft lumen. It is important to avoid repeated punctures at the same site, since ex-
tensive damage to the graft wall will eventually result in pseudoaneurysm formation and/
or localized infection. The failure to rotate sites has increased during the last decade in the -o

United States. |

Perigraft hematomas should be treated conservatively unless signs of infection appear; £

if infection occurs, incision and drainage are required. Infected perigraft hematomas usually !«

necessitate graft revision or removal. It is wise to avoid the use of a fistula with a large ^

perigraft hematoma; temporary access should be obtained until the hematoma has resolved. g

I

D. Maintenance/Surveillance <3

Once a fistula has been successfully constructed, the surgeon has an ongoing responsibility §

to the patient to ensure that the fistula lasts as long as possible. To maximize fistula ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR ACCESS 417

longevity, the surgeon must consider three important points. First, the new fistula should
not be used until it has had sufficient time to mature. Maturation of autogenous fistulas
usually takes 2-3 months. In many cases the fistula can be easily palpated and cannulated
before this time; however, the vein wall requires 2-3 months (DOQI) to sufficiently enlarge
and arterialize. Early use frequently results in hematoma formation around the fistula,
which may progress to scarring and stenosis and markedly shortens the useful life of the
fistula. Fistulas constructed with prosthetic material usually require 10-14 days for suf-
ficient incorporation of the foreign material. Premature puncture of the graft commonly
leads to perigraft hematoma formation and a greater likelihood of infection.

Second, and whenever possible, the surgeon should ensure that dialysis personnel use
proper aseptic technique at the time of cannulation and that they rotate puncture sites.
Continual cannulation of a site leads to degeneration of the vessel or prosthetic graft wall,
with pseudoaneurysm formation and/or infection. This is a matter of education and
communication, which is the surgeon's as well as the nephrologist's responsibility. It is
also important to avoid cannulation in the vicinity of the anastomotic sites to prevent
cutting a suture and causing disruption.

Third, the dialysis nurses and technicians should be informed of the need to contact
the surgeon whenever a problem with the fistula occurs. Early detection of changes in the
hemodynamic characteristics of the fistula — such as high venous pressures, which fre-
quently precede thrombosis — is vital. Correction of the abnormality before the fistula clots
will markedly improve the useful life of the access site.

E. Thrombosis

Thrombosis is the most common problem with hemodialysis access. The majority of pa-
tients on hemodialysis will need revision or thrombectomy of their access at some time or
other. Once thrombosis occurs, performance of a fistulogram is useless. Early surgical
exploration or declotting by percutaneous intervention yields the highest salvage rates. At
exploration, a careful search will usually determine the cause of thrombosis. Unless the
cause is discovered and the problem corrected, thrombectomy will be unsuccessful in re-
establishing a usable fistula. This cannot be overemphasized.

In most cases, the existing graft or fistula can be salvaged with thrombectomy and
revision. AVFs that have been clotted longer than 24-48 h have a lower rate of salvage,
probably because of the intimal damage that occurs when clot is present in the vessel lumen.
Even in patients who present late following thrombosis of an AVF, unless overt phlebitis is
evident, an attempt at thrombectomy or reconstruction should usually be made. Interven-
tional approaches to declotting of autogenous AVFs (especially radiocephalic ones ) are not
often successful. Therefore, a long-standing autogenous AVF that occludes should gen-
erally be approached surgically. ■g

Thrombosis of synthetic AVFs occurs frequently enough that many surgeons view it &

as a routine part of the management of fistulas made with synthetic material. In most cases, a

such grafts can be successfully thrombectomized if both arterial and venous connecting c

vessels are patent and thrombectomy is accomplished before clot has propagated and <j

occluded the proximal venous runoff or before extensive intimal damage has occurred in >9

the runoff vessel. Recently placed grafts are easy to thrombectomize. Older grafts may 41

accumulate thick layers of pseudointima, which is difficult to remove. When this is the 2

situation, it usually is better to excise and replace the graft or to place another fistula |

elsewhere. ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

418 ADAMS et al.

It is important to remove all thrombus from the arterial and venous ends of the fistula.
Visualization of the arterial end can be difficult once blood flow is restored. One way to
obtain a good look at the arterial anastomosis is to temporarily occlude arterial flow with
a proximal sterile tourniquet, which allows complete inspection and calibration of the
arterial anastomosis with a minimal amount of blood loss. Calibration can be performed
with a balloon catheter or coronary dilator. Both anastomoses should be directly vi-
sualized. When there is less than a clear, strong, palpable thrill proximally in the fistula, an
operative fistulogram should be obtained.

The most common cause of graft thrombosis with synthetic material is stenosis at or
proximal to the venous anastomosis. This probably represents a form of pseudointimal
hyperplasia often seen in PTFE peripheral arterial reconstructions. Correction requires
angioplasty, bypass, or movement of the anastomosis to another outflow vein.

Assessment of what constitutes adequate venous runoff is probably the single most
difficult decision in creating a permanent venous access. Calibration of the outflow tract
with coronary artery dilators is useful. A critical property of the outflow vein is elasticity.
Free passage of anything smaller than a No. 3 coronary dilator is probably insufficient. If
the graft or vein has a strong pulse but no thrill, venous outflow obstruction is present.

The interventional radiologist has gained a significant role in the management of ac-
cess problems. This has had both positive and negative effects. On the positive side, the
current ability to percutaneously perform thrombectomy, diagnostic angiography, and/or
angioplasty with or without stenting has prolonged the life of many access sites. Negative
effects include cost as compared to surgical thrombectomy/revision and lack of planning
that incorporates important aspects of the patient's history and prognosis. We utilize a
system in which the patient is first evaluated by a surgeon, who decides the best course of
action for the patient based on past history, physical examination, performance in dialysis,
etc. Repeated interventional thrombectomies are rarely justified.

III. USE OF DIALYSIS HISTORY AND PHYSICAL EXAMINATION

A. Arterial or Inflow Stenosis

Arterial or inflow stenosis may be diagnosed by noting that before occlusion the fistula was
not providing adequate inflow into the dialysis machine (frequent "negative pressures").
On physical examination before occlusion, the fistula usually appears somewhat collapsed
or has a weak pulse when occluded downstream. Such inflow stenoses are most common in
radial artery-based fistulas. Treatment usually involves revision of the fistula with anas-
tomosis to a more proximal area of the radial artery. If there is no associated venous
stricturing, the cephalic vein can be thrombectomized and will provide excellent outflow.
Because the vein is already arterialized, such a fistula can be used immediately following ^

revision. g

3

B. Venous or Outflow Stenosis g>

This situation is diagnosed by noting that before occlusion, high venous (or "return") <j

pressures were present. In autogenous fistulas, venous stenoses are usually caused by >9

premature cannulation of an inadequately arterialized vein or by repeated punctures at the 4j

same site. Occasionally it is possible to revise the anastomosis or bypass the area of stenosis. 2

However, in many cases the stenoses are multiple and cannot be bypassed; therefore the |

fistula must be moved to another site. In PTFE fistulas, neointimal hyperplasia at the distal ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR ACCESS 419

venous anastomosis is the common cause of stenosis; it is best treated by patch angioplasty
or revision with bypass of the area of stenosis. Alternatively, balloon angioplasty and
stenting can be used.

C. Hypotension

Patients undergoing hemodialysis frequently experience episodes of hypotension due to the
removal of extracellular fluid during dialysis or occasionally from sepsis. AVFs may
thrombose during these events. If the fistula was functioning well before thrombosis,
thrombectomy is frequently successful. As noted earlier, it is wise to carefully evaluate both
arterial and venous limbs at the time of thrombectomy to rule out mechanical obstruction
as a predisposing factor. Hypotension on dialysis is usually due to aggressive fluid removal.
This sometimes is the result of the patient having gained significant weight between dialysis
treatments. It is more often a problem in small patients who are noncompliant with fluid
restrictions and following the treatment occurring after a 2-day gap (e.g., Monday for a
patient on M-W-F dialysis).

D. Hypercoagulable State

Hypercoagulable states are not uncommon in dialysis patients and should be considered in
two situations: (a) when a well-functioning fistula suddenly thromboses and cannot be
salvaged even though the thrombectomy or revision seems adequate intraoperatively and
(b) when a patient on chronic hemodialysis presents with a history of repeated unsuc-
cessful attempts at obtaining vascular access by experienced vascular surgeons.

Any patient in whom a technically adequate AVF has failed to remain patent should
be suspected of having a hypercoagulable state. "Technically adequate" is defined as the
operative documentation of a palpable thrill or postoperative presence of a clearly audible
bruit in the AVF and/or intraoperative measurement of adequate blood flows. If an ad-
equate thrill and bruit are present but are subsequently lost in the early postoperative
period, the patient should be evaluated for a coagulation abnormality. An "adequate thrill
or bruit" should be present throughout the entire cardiac cycle, indicating sufficient ar-
terial inflow and low outflow resistance.

The two most common abnormalities resulting in hypercoagulability and access throm-
bosis are antithrombin III deficiency and heparin-induced platelet aggregation. Patients
with antithrombin deficiency can usually be managed by perioperative infusion of fresh
frozen plasma to supply the missing factors, followed by warfarin therapy. Patients with
heparin-induced platelet aggregation often have a history of frequent blockage of dialysis
coils by clots. If hemodialysis is the only means for maintenance dialysis, these patients can
usually be dialyzed without the use of heparin; however, this situation presents difficulties
for most dialysis units. Many of these patients will eventually require peritoneal dialysis. ^

Most patients with a hypercoagulable state will need long-term warfarin anticoagulation. E

Additionally, patients with repeated episodes of graft thrombosis of unknown origin are 8

best served by chronic anticoagulation. %

E. Erythropoietin

Because erythropoietin is now administered to most hemodialysis patients, they are usually
not as anemic as before its availability. Anemia contributes to impaired coagulation in
uremic patients and thus promotes patency of AVFs (27). Vigorous correction of anemia
using erythropoietin therapy may therefore serve to potentiate thrombosis. However, it has

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

420 ADAMS et al.

been our experience that patients on erythropoietin therapy in whom the fistula thromboses
have a technical abnormality that explains the incident.

F. Pseudoaneurysm Formation

Pseudoaneurysms occur commonly in AVFs that have been used for vascular access over a
prolonged period. They occur in both autogenous and prosthetic fistulas and can result
from improper technique in graft cannulation, failure to rotate sites, and the nonhealing
nature of prosthetic material. If pseudoaneurysms erode through overlying skin, the result
may be life-threatening hemorrhage. The risk of erosion is best judged by whether the skin
over the pseudoaneurysm is fixed or movable. Skin that is firmly fixed and thin should
prompt corrective action. Once this condition is diagnosed, it can be corrected by by-
passing or replacing the involved segment with new material or, when the pseudoaneurysm
involves an anastomosis, by reconstructing that portion of the AVF.

Occasionally, a pseudoaneurysm is so large and complex that it requires sacrificing
the fistula. The principal risks of pseudoaneurysm are continued enlargement with throm-
bosis, rupture, or distal embolization. Dialysis personnel should be cautioned to avoid
repeated cannulation of areas of pseudoaneurysm because of the risk of skin breakdown and
rupture.

G. Infection

Infection can occur in any fistula but is more common with vascular access using synthetic
graft material or in lower extremity AVFs, particularly in obese patients. It is usually
related to lack of proper aseptic technique at the time of cannulation (28). Repeated
cannulations at the same site and poor sterile technique (e.g., poor skin preparation,
touching the prepared site with an ungloved finger) also place the fistula at higher risk for
infection. Such infections can occasionally be managed by excision of the involved
segment if the graft has been well incorporated before onset of the infection and the
infected area does not involve an anastomosis to a host vessel.

More often, infection in prosthetic material involves one or both anastomoses and the
whole graft must be removed. Before a wound with a suspected prosthetic graft infection is
opened, a proximal tourniquet should be readily available. At the time of removal, a small
rim (2-3 mm) of prosthetic material may be left attached to the artery to provide a secure
closure. It is technically difficult to excise all the PTFE and close the arteriotomy without
occluding it because of the arterial wall retention, which occurs after the PTFE-to-artery
anastomosis. Infected wounds are best left open and allowed to heal secondarily. However,
it is a good idea to cover the arterial closure with soft tissue.

Infection that occurs in an autogenous fistula is usually the result of a secondary
infection of a hematoma surrounding the fistula caused by repeated cannulation of the "g

same segment of vein. Because of the autogenous nature of the infection, these can often &

be treated with antibiotics and local care unless hemorrhage is threatened or has occurred. a

When prosthetic materials are inserted, perioperative coverage with systemic anti- -c

biotics is recommended. A single dose of an antibiotic with gram-positive coverage (first- <j

generation cephalosporin, nafcillin, or vancomycin) is probably sufficient. The surgeon >9

should try to anticipate when prosthetic material might be used so that antibiotics can be 4j

administered before the skin is incised. 2

Infection involving a major artery, such as the brachial artery at its bifurcation or the |

superficial femoral artery, may place a limb in jeopardy. With synthetic grafts placed in the ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR ACCESS 421

groin, the arterial anastomosis should be made to the superficial femoral artery and not to
the common femoral artery. In this way, if graft infection should occur and ligation of the
donor vessel becomes necessary, it may be done with a lesser chance of limb compromise.
In general, it is better to avoid placing any prosthetic material in the groin for vascular
access if alternate sites are available in the upper body.

H. Venous Hypertension

Venous hypertension results from either high flow into a venous bed distal to the AVF and/
or proximal venous occlusion (29). Occasionally both conditions may be present. Patients
usually present with redness and edema of a hand or arm, and stasis ulcers may actually
develop. Other signs of venous stasis include rapid capillary filling time, cyanosis, brawny
edema, and ecchymosis. If the AVF is placed at the wrist, there prominent pulsatile veins are
usually palpable through the edema on the dorsum of the hand. When an antecubital fistula
is the source of venous hypertension, patients more commonly present with edema of the
hand and forearm and a vigorous thrill at the site of the fistula extending distally into the
forearm.

Attention to detail at the time of initial access placement can usually avoid post-
operative venous hypertension. Distal venous branches from the fistula should be ligated
unless such branches allow for the filling of vessels that will be needed for access. The most
common example of this situation is the lateral ulnar vein on the dorsum of the forearm,
which in some cases will fill through a branch off the distal cephalic vein proximal to the
fistula. In the case of an antecubital AVF, the distal cephalic vein extending to the forearm
may provide excellent sites for cannulation and should be preserved. However, venous
branches going deep in to the forearm and proximally along any major arteries should
be ligated. Even though there is no apparent flow at the time of fistula construction, as
veins dilate, venous valves may become incompetent, resulting in distal flow and venous
hypertension.

If a patient with a previously functioning AVF suddenly presents with venous hyper-
tension, deep venous occlusion should be suspected. This serious situation can be resolved
only by takedown of the fistula on the affected side. If the fistula is patent at the time, it
may be possible to secure new access elsewhere before taking down the fistula causing
venous hypertension. Delay in takedown in this situation may put the patient at risk for
subsequent skin necrosis.

I. Fistulogram

A fistulogram is a radiographic contrast study of the AVF to define arterial and venous

anatomy. Although noninvasive techniques such as color duplex ultrasound scanning can

provide similar information, a fistulogram provides the precise anatomical information that ■g

the surgeon requires. A fistulogram may play an important role in diagnosing and treating &

venous hypertension and the failing fistula. It aids in both locating venous branches sup- a

plying the distal venous bed and documenting the presence or absence of proximal deep c

venous occlusion. <j

y
The technique is simple and can be done without formal arteriography. A needle is «

placed into a vein connecting with the AVF or into the prosthetic graft itself and a blood 4j

pressure cuff is placed high on the upper arm and inflated above systolic pressure. Intra- Q

vascular contrast (usually full-strength) is injected into the needle, thus filling both veins and |

arteries supplying the fistula. More than a single view may be necessary to demonstrate the ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

422 ADAMS et al.

relevant anatomy. The fistulogram is rarely of use in the evaluation of a thrombosed fistula
and its use in this situation should be abandoned.

Immediately following thrombectomy and/or revision, the surgeon may wish to
perform an intraoperative arteriogram. This often proves useful because unsuspected
stenoses may be identified and corrected before the next episode of thrombosis occurs. It
is rarely necessary to perform angiography. Venography is occasionally of use in the
evaluation of veins preoperatively in obese patients in whom physical examination alone is
inconclusive.

Many dialysis units now monitor blood flows ultrasonically and send patients with
lower blood flows for study by fistulogram. Hopefully, this practice will identify problem-
atic AVFs before occlusion occurs, allowing correction of abnormalities.

J. Steal Syndromes

Steal syndromes occur most often in elderly and/or diabetic patients with advanced
peripheral atherosclerosis involving the arteries of the upper extremities. Patients at high
risk usually have skin changes typical of diabetes and at operation have heavily calcified
atherosclerosis of even 2- to 3- mm vessels. These patients can usually be identified
preoperatively by careful palpation of the radial artery and/or by the examination of x-ray
films of the forearms for vascular calcifications. Attempts to obtain vascular access for
hemodialysis are often unsuccessful in these patients. However, some individuals with
calcified vessels can still undergo successful fistula placement. A more precise approach is
the use of vascular laboratory measurements of finger pressures on both sides. The side
with the least severe disease can be chosen for placement of the fistula. In extremities with
finger pressures below 90 mmHg, the results will predictably be poor. Any fistula placed in
a patient with arterial insufficiency of the limb places distal tissue at risk; such placement
should be avoided. Some patients may not have an acceptable site for vascular access and
will need to have peritoneal dialysis or dialysis via a central catheter.

If the access has already been placed and ischemic changes of the hand or fingers
develop, the fistula should be taken down urgently. Once ischemic neuropathy or gangrene
develops, many of these patients will eventually require amputation of fingers and, on
occasion, a hand. Attempts at downsizing a fistula in this situation are usually unsuccess-
ful and prolong the period of ischemia.

IV. THE FAILING FISTULA

Many episodes of access failure can be prevented if the surgeon, nephrologist, and dialysis
personnel work in concert to identify and correct access problems before thrombosis

occurs. This same principle has been shown to apply in peripheral arterial bypass surgery ■g

(30). The role of noninvasive evaluation of access function has been unclear. Early studies &

suggested that routine Doppler evaluation of AVFs was useful in predicting which AVFs a

were at risk for problems. These efforts were short-lived because of lack of reimbursement c

for this surveillance. Venous pressure measurements during hemodialysis became the main <j

surveillance technique used by dialysis personnel. Recently there has been increased use of >9

ultrasonic measurements of access blood flow to predict thrombosis (31,32). This has the 4j

advantage of quantitating a change over time for a given patient. It requires relatively little 2

time and effort and can serve to identify a patient's risk for access failure. Timely referral |

to a vascular surgeon should be made. Often the problem can be diagnosed by history and ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR ACCESS 423

physical examination. If not, a fistulogram should be performed. If the fistulogram iden-
tifies a problem amenable to angioplasty/stenting, it can be performed at the same sitting.
If not, plans can be made for surgical revision or new AVF placement. Access issues should
be viewed as urgent and dealt with accordingly. Delay often complicates both the problem
and the solution.

Vascular surgeons are frequently called to evaluate patients with "failing fistulas,"
that is, AVFs that previously provided adequate sites for cannulation and flow on dialysis
for prolonged periods of time but have become difficult to cannulate or are not providing
enough flow for adequate dialysis. In these situations it is important to speak directly with
the dialysis nurse or technician to determine the exact nature of the problem, in addition
to performing a careful physical examination, since a number of different situations may
exist to cause this problem. With a dialysis history and physical examination, a fistulo-
gram is not necessary to formulate a diagnosis. However, it may be useful for operative
planning.

A. High Venous Pressure

The AVF may have developed high venous resistance to the reinfusion of blood. For this
reason, extra pressure is required to reinfuse blood into the patient, resulting in ultra-
filtration. A rising venous pressure over a period of weeks to months should be a clue that
venous outflow obstruction is developing and revision should be undertaken before the
access fails completely.

A fistulogram to delineate the site of stenosis is often useful in this setting. If the area
of venous stenosis is single and located near a radiocephalic anastomosis, it will be
possible, in most cases, to reconstruct the fistula proximal to the stenosis. This will usually
provide a number of sites for cannulation. If the area causing the problem is some distance
from the anastomosis and is too proximal to bypass readily, it may be possible to perform
an angioplasty or resect the area of stenosis. Often there are several stenotic areas with
surrounding scar, and a new AVF will need to be reconstructed at another site.

Venous stenosis occurring proximal to the venous anastomosis of a PTFE graft may
occur at one site or multiple sites. The single-site stenosis can be corrected by bypassing it
or performing patch angioplasty. Balloon angioplasty of venous stenoses is frequently
performed but the long-term success rates are not good (25% patency at 1 year) (33,34).
Problems on the venous side of a fistula are usually more serious than arterial problems in
terms of maintaining adequate vascular access in the same general area.

B. Poor Arterial Inflow

The second type of problem occurs when there is insufficient or failing arterial inflow to

the dialysis machine. Useful access requires arterial inflow of at least 200 mL/min. ■g

Problems with arterial inflow present with collapse of the arterial segment of the dialysis &

circuit because the inflow pressure becomes negative with respect to the dialysis pump. a

Here, as in the case when there is a high venous pressure, a fistulogram will usually provide -c

useful information regarding the location of the problem. <j

The radiologist should be informed that special attention must to be paid to the >3

arterial segment of the fistula. Occasionally additional films will be required to demon- 4j

strate the full arterial anatomy apart from overlying veins. Unless the patient is diabetic, Q

the stenosis most often occurs in the vicinity of the anastomosis and can be corrected by |

reconstructing the fistula proximal to the original site. ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

424 ADAMS et al.

If the problem is extensive atherosclerosis, it will usually be necessary to create a new
fistula at another site. The hemodynamic effects of sequential stenoses are greater than the
severity of any single stenosis would suggest (35).

C. Difficulty in Cannulation

The final problem that occurs in the dialysis unit and causes a "failing fistula" is increased
difficulty with cannulation. This is more common in units with less experienced personnel
or with an uncooperative or obese patient. It is important to carefully examine the
involved extremity. Not infrequently, the problem is progressive scarring around a site
that has been used for an extended period of time. There are often additional viable sites
proximal to the favored one.

If, in fact, the fistula is patent but cannot be reliably cannulated, a new one will have
to be constructed, usually at a different site. A fistula that can be cannulated only with
difficulty by the most experienced nurse or technician is inadequate, regardless of its
hemodynamic characteristics once cannulated. Before another fistula is constructed, a
concerted attempt should be made to use the existing access site. The performance of a
fistulogram may be useful, followed by a cannulation demonstration of the access site by
the surgeon or nephrologist and/or their marking of new sites for dialysis personnel.

D. Congestive Heart Failure

Congestive heart failure is a rare complication of an AVF created for hemodialysis access
(36). A well-developed AVF may obligate as much as 20% of the cardiac output. Rarely,
flow through a large antecubital or femoral loop fistula can develop to greater than 2 L/
min. If the patient is small or has significant organic heart disease, this can result in high-
output cardiac failure. The diagnosis is missed only when it is not considered. It is
suggested by a drop in heart rate with temporary occlusion of the fistula. The diagnosis
can be more difficult in patients receiving beta blockers in whom the heart rate response to
fistula occlusion is muted. Thermodilution cardiac output measurements may be useful in
selected situations.

If vascular access is still required, two alternatives exist; either the fistula may be
ligated after usable access (either temporary or permanent) is obtained or the flow through
the existing access site may be decreased by downsizing the fistula. This can be done by
direct reconstruction of the anastomosis or by banding the outflow vein to a smaller
diameter. It is difficult to adequately judge the degree to which the fistula can be
diminished without causing thrombosis; this is one situation in which Doppler flow
probes may be of use. The best intraoperative information supporting adequate diminu-
tion of flow is a significant decrease in heart rate.

E. Takedown of Arteriovenous Fistulas I

If the AVF is no longer needed (resolution of renal failure, permanent peritoneal dialysis, a

successful renal transplant), takedown of the AVF may be indicated. This may be done c

either for cosmetic reasons or because of the possibility of hemorrhage from minor <j

trauma. In addition, there is some evidence that closure of AVFs may be beneficial in >9

reducing myocardial workload and left ventricular hypertrophy (37). 4j

In general, fistula takedown following renal transplantation should not be considered 2

until renal function has been stable for at least 1 year without significant complications |

relating to immunosuppression or recurrent renal disease. A patient's history of vascular @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR ACCESS 425

access should be considered in deciding whether to remove a fistula. Patients in whom it has
been difficult to obtain and maintain vascular access should usually not have the fistula
removed.

AVF takedown can easily be done as an outpatient procedure with the patient under
local anesthesia. A pneumatic tourniquet should be considered as a safety precaution, but
it should be deflated unless needed. Ligation of outflow veins will accomplish takedown.
Arterial flow to the hand should be preserved by reconstruction or preservation of the
artery. Alternatively, venovenous reanastomosis and arterial reconstruction may preserve
the vein and artery for future use and requires little extra time or effort.

Following fistula takedown or ligation, the large venous conduit becomes relatively
stagnant and thrombosis with rather pronounced phlebitis frequently occurs. This can be
lessened by elastic compression dressings placed at the time of fistula takedown. The treat-
ment for this condition is symptomatic and the patient should be reassured that the prob-
lem will resolve.

F. Preservation of Cephalic Veins

In any patient with abnormal renal function or a disease process likely to lead to renal
failure, all cephalic veins should be carefully preserved. This is often difficult to accom-
plish, since the cephalic veins are the most readily accessible intravenous sites. Often, by
the time the surgeon has been consulted regarding vascular access, both cephalic veins are
thrombosed because of ill-advised placement of intravenous sites. Obtaining an AVF in
this situation becomes more difficult. It is important to constantly remind those taking care
of patients with any renal abnormality or with the potential for it that the need for quality
permanent vascular access requires preservation of all cephalic veins.

G. Neurological Sequelae

The patient should be informed that the position of the radial cutaneous nerve puts it at
risk for injury during fistula formation or takedown. Injury results in a small area of de-
creased sensation on the back of the thumb. Careful preservation of the radial cutaneous
nerve at operation will minimize the risk of this complication. This nerve is more often
injured during takedown than during formation of AVF's, probably because of the scar
tissue created by the original operation through which the takedown dissection must
proceed.

Rarely, a patient with marked distal arterial insufficiency due to the AVF will present
with palsy of the radial or median nerve. Nerve conduction studies will demonstrate abnor-
malities that cannot be related to the same level for both nerves. When nerve palsy occurs,
the fistula must be taken down urgently and another form or site of dialysis undertaken.

I

H. Carpal Tunnel Syndrome |

&

For unknown reasons, carpal tunnel syndrome occurs more commonly in hemodialysis a

patients than in the general population. Since the syndrome occurs as commonly on the c

side without the fistula, it is probably not related to its presence. <j

Treatment of carpal tunnel syndrome in a wrist with a functioning fistula requires >3

planning. Since this syndrome is best treated surgically under tourniquet hemostasis, 4j

attention must be given to preservation of the fistula. We handle this problem by injecting 2

the fistula with heparin before inflating the tourniquet. Using this technique, we have not |

lost a fistula even with up to 20 min of tourniquet occlusion. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

426 ADAMS et al.

V. CONCLUSION

Complications common to all areas of vascular surgery occur frequently in vascular access
patients. Morbidity rarely involves limb loss, as it does in peripheral reconstructive
surgery, but it represents significant problems nevertheless. Because these vessels or grafts
are routinely cannulated, sometimes under less than ideal conditions, infection and
pseudoaneurysm formation are not common.

The usable life of any vascular access depends to a large extent on how it is handled by
personnel in the dialysis unit. The vascular surgeon's responsibility to the patient continues
beyond the point at which usable vascular access is obtained. It is important to continually
remind dialysis personnel of the rules in handling and cannulating AVFs and to remind
physicians caring for these patients not to use any cephalic veins for intravenous sites. A
well-constructed and cared for autogenous fistula can last for many years.

REFERENCES

1. www.usrds.org.

2. Shrier RW, Gottschalk CW. Diseases of the Kidney. 4th ed. Boston: Little, Brown, 1988:1526.

3. Bour ES, Weaver AS, Yang HC, GifFord RRM. Experience with the double lumen Silastic
catheter for hemoaccess. Surg Gynecol Obstet 1990; 17(l):33-39.

4. Schwab ST, Buller GI, McCann RL, Bollinger RL, Stickel DL. Prospective evaluation of a
Dacron cuffed hemodialysis catheter for prolonged use. Am J Kidney Dis 1988; 11:166-169.

5. Mendes RR, Farber MA, Marston WA, Dinwiddie LC, Keagy BA, Burnham SJ. Prediction of
wrist arteriovenous fistula maturation with preoperative vein mapping with ultrasonography.
Vase Surg 2002; 36(3):460-463.

6. Atherikul K, Schwab SJ, Conlon PJ. Adequacy of hemodialysis with cuffed central-vein
catheters. Nephrol Dial Transplant 1998; 13(3):745-749.

7. Hirsch DJ, Bergen P, Jindal KK. Polyurethane catheters for long-term hemodialysis access.
Artif Organs 1997; 21(5):349-354.

8. Schwab SJ, Weiss MA, Rushton F, Ross JP, Kapoian T, Yegge J, Rosenblatt M, Reese WJ,
Soundararajan R, Pedan A, Moran JA. Multicenter clinical trial results with the LifeSite
hemodialysis access system. Kidney Int 2002; 62(3): 1026-1033.

9. Beathard GA, Posen GA. Initial clinical results with the LifeSite Hemodialysis Access System.
Kidney Int 2000; 58(5):222 1-2227.

10. Castaldo PA. Homeostasis and kidney disease. In: Tatnoff DO, Forbes CD, eds. Disorders of
Hemostasis. Orlando, FL: Grune and Stratton, 1984:473-483.

11. Livio M, Benigni A, Remuzzi G. Coagulation abnormalities in uremia. Semin Nephrol 1985;
5:82-90.

12. Mannucci PM, Remuzzi G, Pusiner F, Lombardi R, Valsecchi C, Mecca G, Zimmerman TS.
Deamino-8-D-arginine vasopressin shortens the bleeding time in uremia. N Engl J Med 1983;
308:8. '

13. Liu KY, Kosfeld RE, Marcum SG. Treatment of uremic bleeding with conjugated oestrogen. £
Lancet 1984; 2:887-890. 1

14. Kappes S, Towne JB, Adams MB, Kauffman HM, Maierhofer W. Perforation of the superior 2
vena cava: a complication of subclavian dialysis. JAMA 1983; 249:2232-2233. '°

15. Joseph RE, Radhakrishnan J, Appel GB. Antiphospholipid antibody syndrome and renal dis- ^
ease. Curr Opin Nephrol Hypertens 2001; 10(2): 175-181 . &

16. Kauffman HM, Elborn GA, Adams MB, Hussey CV. Hypercoagulability: A cause of vascular jjj
access failure. Proc Clin Dialysis Transplant Forum 1979; 9:28. Q

17. Passman MA, Criado E, Farber MA, Risley GL, Burnham CB, Marston WA, Burnham SJ, |
Keagy BA. J Vase Surg 1998; 28(5):869-875. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF VASCULAR ACCESS 427

18. Giacchino JL, Geis P, Buckingham JM, Vertumo VL, Bansal VK. Vascular access: Long-term
results, new techniques. Arch Surg 1979; 114:403-409.

19. Konner K. Vascular access in the 21st century. J Nephrol 2002; 15(suppl 6):S28-S32.

20. Anderson CB, Etheredge EE, Sicard GA. One hundred polytetrafluoroethylene vascular access
grafts. Dialysis Transplant 1980; 9.

21. Tellis VA, Kohlberg WE Bhat DJ, Driscoll B, Veith FJ. Expanded polytetrafluoroethylene
graft fistula for chronic hemodialysis. Ann Surg 1979; 189:101-105.

22. Morgan AP, Dammin GJ, Lazarus JM. Failure modes in secondary vascular access for
hemodialysis. Am Soc Artif Int Organs 1978; 1:44-52.

23. Johnson CP, Zhu YR, Matt C, Pelz C, Roza AM, Adams MB. Prognostic value of intra-
operative blood flow measurements in vascular access surgery. Surgery 1998; 124(4):729-737.
Discussion 737-738.

24. Glickman MH, Stokes GK, Ross JR, Schuman ED, Sternbergh WC III, Lindberg JS, Money
SM, Lorber MI. Multicenter evaluation of a polytetrafluoroethylene vascular access graft as
compared with the expanded polytetrafluoroethylene vascular access graft in hemodialysis
applications. J Vase Surg 2001; 34(3):465-472. Discussion 472-473.

25. Almonacid PJ, Pallares EC, Rodriguez AQ, Valdes JS, Rueda Orgaz JA, Polo JR. Comparative
study of use of Diastat versus standard wall PTFE grafts in upper arm hemodialysis access.
Ann Vase Surg 14(6):659-662.

26. Lemson MS, Tordoir JH, van Det RJ, Welten RJ, Burger H, Estourgie RJ, Stroecken HJ,
Leunissen KM. Effects of a venous cuff at the anastomosis of polytetrafluoroethylene grafts for
hemodialysis vascular access. J Vase Surg 2000; 32(6): 11 55-1 163.

27. Livio M, Gotti R, Marchesi D, Mecca de Gaetano G Uremic bleeding: Not of anemia and
beneficial effect of red cell transfusions. Lancet 1982; 2:1013-1015.

28. Appel GB. Vascular access infections with long-term hemodialysis. Arch Intern Med 1978;
138:1609-1610.

29. Wilson SE. Complications of vascular access procedures. In: Wilson SE, Owens MI, eds. Vas-
cular Access Surgery. Chicago: Year Book, 1980:185-207.

30. Bandyk DF, Schmitt DD, Seabrook GR, Adams MB, Towne JB. Monitoring functional
patency of in situ saphenous vein bypass: The impact of a surveillance protocol and elective
revision. J Vase Surg 1989; 9:286-296.

31. Krivitski NM, Gantela S. Access flow measurement as a predictor of hemodialysis graft
thrombosis: making clinical decisions. Semin Dialysis 2001; 14(3): 181185.

32. Steuer RR, Miller DR, Zhang S, Bell DA, Leypoldt JK. Noninvasive transcutaneous deter-
mination of access blood flow rate. Kidney Int 2001; 60(1):284-291.

33. Lombardi JV, Dougherty MJ, Veitia N, Somal J, Calligaro KD. A comparison of patch
angioplasty and stenting for axillary venous stenoses of thrombosed hemodialysis grafts. Vase
Endovasc Surg 2002; 36(3):223-229.

34. Clark TW, Hirsch DA, Jindal KJ, Veugelers PJ, LeBlanc J. Outcome and prognostic factors of
restenosis after percutaneous treatment of native hemodialysis fistulas. J Vase Intery Radiol
2002; 13(l):51-59.

35. Beckmann CF, Levin DC, Kubicka RA, Henschke CI. The effect of sequential arterial stenoses

on flow and pressure. Radiology 1981; 140:655-658. g

36. Fee HJ, Levisman JE, Doud RB, Golding AL. High output congestive failure from femoral <S
arteriovenous shunts for vascular access. Ann Surg 1976; 183:321-323. s

37. Unger P, Wissing KM, de Pauw L, Neubauer J, van de Borne P. Reduction of left ventricular 2
diameter and mass after surgical arteriovenous fistula closure in renal transplant recipients. ■"■
Transplantation 2002; 74(l):39-73. 4

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

24

Complications of Thoracic Outlet Surgery

David Rigberg

UCLA Medical Center, Los Angeles, California, U.S.A.

Julie Freischlag

Johns Hopkins School of Medicine, Baltimore, Maryland, U.S.A.

I. INTRODUCTION

Perhaps no disorder is as vexing to the vascular surgeon as the thoracic outlet syndrome
(TOS). Although the venous (Paget-Schroetter) and arterial sequelae of thoracic outlet
compression have clear objective signs, the more common neurogenic form requires a
clinical diagnosis. There is considerable controversy regarding this diagnosis, and edito-
rials by physicians experienced in treating these patients can be found warning clinicians
of both over- and underdiagnosing of neurogenic TOS (1,2).

As discussed further on, TOS stems from the compression of several important struc-
tures traversing the thoracic outlet. Over the years, operative treatment has evolved from
more radical therapies, such as bilateral claviculectomies, to the currently practiced ap-
proaches: transaxillary first rib resection and supraclavicular scalenectomy with or without
first rib resection. The complications of these two procedures are the focus of this chapter,
and these procedures are performed in a similar fashion for whichever form of TOS is being
treated. In addition, recurrent TOS is considered an operative complication for the purposes
of this review.

With regard to the operative treatment of other consequences of TOS, complications ■a

tend to mirror these procedures in other, non-TOS settings. Examples of these include |

resection of subclavian aneurysms, lysis and later venoplasty of axillo-subclavian throm- a

bosis, or even the need for upper extremity sympathectomy. The one caveat is that the long- c

term success of most of these procedures depends on adequate decompression of the inciting <

TOS, without which recurrence or even primary treatment failure can be expected. >9

Most patients with neurogenic thoracic outlet syndrome never require operative J

intervention, as physical therapy tends to have good results in this setting. Physical therapy «

programs designed to open the anatomical confines of the thoracic outlet were originally |

described by Peet, who also coined the term TOS in 1956 (3). These programs are designed @

429

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

430 RIGBERG and FREISCHLAG

to relax muscle groups that tighten the thoracic outlet while conditioning those that open it.
Aligne and Barral further described a program in which the trapezius, levator scapulae, and
sternocleidomastoid muscles are strengthened and the middle scalene, subclavian, and
pectoralis muscles are relaxed (4). These goals can be met by many different protocols,
usually with a combination of supervised and at-home exercises. Complications from such
treatment are minimal, although improperly performed physical therapy can lead to
worsening of neurogenic TOS symptoms. The injection of botulinum toxin (Botox) provides
another nonoperative intervention for TOS, although there is little experience with this
modality's long-term results. Even with the considerable number of patients helped by
conservative therapies, there are several tertiary referral centers with large series of TOS
patients who have remained symptomatic.

II. ANATOMY

The limited space and large number of important structures that must traverse the neck
and chest areas on their way to the arm make the thoracic outlet an area like no other in
the body. Although there are any number of anatomical anomalies that predispose to or
directly cause compression of the neural, venous, and arterial structures within its con-
fines, the normal anatomy itself does not leave much room for stress positioning. Any of
the structures within the thoracic outlet can be injured operatively, so a thorough appre-
ciation of the region's anatomy is required before these procedures are performed.

Definitions may vary from author to author, but it is generally accepted that the
thoracic outlet is the area from the edge of the first rib extending medially to the upper
mediastinum and superiorly to the fifth cervical nerve. The clavicle and subclavian muscles
can be pictured as forming a roof, while the superior surface of the first rib forms the floor
(Fig. 1). Machleder's description of the thoracic outlet as a triangle with its apex pointed

First Rib ^^. / jj

'C

Figure 1 A schematic representation of the fulcrum and "scissoring" effect of the first rib and the
clavicle is shown. As can be seen from the diagram, decompression can be accomplished via removal
of either of these structures.

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

COMPLICATIONS OF THORACIC OUTLET SURGERY

431

toward the manubrium is helpful in visualizing the three-dimensional orientation of the
structures as well as the dynamic changes that can lead to injury (5). In this model, the
clavicle and its underlying subclavian muscle and tendon form the superior limb, while the
base is the first thoracic rib.

Although most cases of TOS are neurogenic, almost any structure that travels through
the thoracic outlet can be involved with the disease or, as previously mentioned, injured
when treating the disorder. Moving from medially to laterally, one first encounters the
exiting of the subclavian vein, usually positioned adjacent to the region where the first rib
and claivicular head fuse to form a fibrocartilagenous joint with the manubrium. Immedi-
ately lateral to the vein is the anterior scalene muscle, which inserts onto a prominence on
the first rib. Lateral to this site is the subclavian artery, so that the anterior scalene muscle
lies between the subclavian artery and vein, with the artery deep, lateral and somewhat
cephalad. The brachial plexus is the next structure encountered (Fig. 2). The C4-C6 roots
are superiorly oriented, and the C7-T1 roots inferiorly. Posterior and lateral to the plexus,
there is a generally rather broad attachment of the middle scalene to the first rib. This is an
area of particular importance during operative decompression of the thoracic outlet, for it is
here that the long thoracic nerve can be inadvertently injured as it travels to the serratus
anterior muscle (Fig. 3).

Other structures encountered in the thoracic outlet include the phrenic and dorsal
scapular nerves, the stellate ganglion, the thoracic duct, and the cupola of the lung. The
phrenic nerve lies between the prescalene fat pad and the anterior scalene muscle (Fig. 4).

middle scalene muscle

overlying clavicle

subclavian artery

anterior scalene muscle

i costoclavicularligament

subclavian vein

I
•5

Figure 2 The anatomy of the thoracic outlet involves the passing of many important structures in
close proximity on their way to the upper extremity and chest. Particular note should be made of the
relationships between the major neurovascular structures and the scalene muscles.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

432

RIGBERG and FREISCHLAG

long thoracic nerve

middle scalene muscle

subclavian artery

first rib

Figure 3 This diagram, an anteroposterior view of the thoracic outlet with the clavicle and
subclavius muscle removed, also shows the proximity of the long thoracic nerve and suggests how
injury to this structure can occur. Note should also be made of the broad insertions of the anterior
and middle scalene muscles.

Compression of this structure does not generally occur, but it can be injured during
supraclavicular approaches and must be left intact while the underlying scalene muscle is
dissected. The dorsal scapular nerve comes off the brachial plexus on its way to innervate the
medially inserting muscles of the scapula (rhomboids and levator scapulae). It is usually
neither involved nor encountered. The stellate ganglion is found along the sympathetic
chain. This structure can be involved in compression, and occasionally a cervicothoracic
sympathectomy is part of the treatment plan for TOS. The thoracic duct may be
encountered if a left supraclavicular approach is undertaken, and care must be taken not
to injure it or to ligate it if injury occurs. Finally, one must watch for pleural injury in any
approach to TOS and be prepared to evacuate pneumothoraces when indicated.

I
•5

III. OPERATIONS

As previously described, there are two primary operations for decompressing the thoracic
outlet, transaxillary first rib resection and supraclavicular scalenectomy with or without first

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF THORACIC OUTLET SURGERY 433

Pre scalene fat pad
Scalenus anticus m

Subclavian a.

Figure 4 This surgical view represents the anatomy as seen via the supraclavicular route for
scalenectomy with or without 1st rib resection. The prescalene fat pad must be divided with caution,
as the phrenic nerve runs directly on top of the underlying anterior scalene muscle. The internal
jugular vein typically serves as the median extent of the dissection. This exposure affords better
access to the brachial plexus, particularly when neurolysis is to be included as part of the
decompressive procedure.

rib resection. Various groups favor one operation over the other, but it appears they are
similarly effective. Most surgeons treating recurrent TOS use the route not previously used,
so that clinicians with a sizable TOS practice are versed in both approaches. Complications
of thoracic outlet decompression do differ significantly based on the approach used, as
would be expected based on the differences in the anatomy one encounters. These differences
are pointed out below where appropriate.

IV. COMPLICATIONS
A. Operative Injuries

Any structure encountered in the dissection for first rib resection is a potential site of injury.
The catastrophic complications of brachial plexus, subclavian artery, and subclavian vein

injury occur infrequently. Concern about brachial plexus injury dates back to the early days ■a

of first rib resection. Particularly in the neurology literature, this has been a controversial |

topic. For a number of years there were only a few case reports of plexus injuries. However, a

Dale in 1982 published the results of a survey of thoracic surgeons performing first rib c

resections and discovered 273 injuries, 19% of which were permanent (6). While this study <

certainly suggested underreporting of this complication, reports of plexus injuries remained >9

low, although not at the almost negligible rate that had been accepted. J

What is most notable about the publications that followed is that no mention is made Q

of the incidence of this injury. Wilbourn in 1988 reported on eight patients with plexus |

injuries form the Cleveland Clinic. They underwent first rib resection at some time between @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

434 RIGBERG and FREISCHLAG

1974 and 1983, but the total number of procedures performed is not given (7). Likewise,
Horowitz reported on four cases of plexus injury, but no figures are given regarding the
number of procedures performed from which these patients were taken (8). The same can
be said of the four cases reported by Cherington in 1986 (9). In several large recent surgical
series, most notably that of Roos, the incidence of such injuries is very low (0-2%). In the
experience of the University of California-Los Angeles (UCLA), there have been no such
injuries. It is fair to say that plexus injuries certainly do occur but that the risk of the injury
must be weighed against the possible benefits of the procedure.

Another issue regarding plexus injuries is that of retraction. It is more than likely that
most of the neurological injuries were related to stretching of the perineurium, with resultant
ischemia. During the transaxillary approach, a considerable amount of stretch is applied to
the arm; most surgeons relieve the traction intermittently to allow blood flow. This was not
necessarily the case when an assistant was retracting. Specialized retraction systems, such as
the Machleder retractor, allow for periods of extremity relaxation with easy repositioning to
continue the procedure.

The incidence of major arterial or venous injury during first rib resection is also difficult
to determine. It is clear that these injuries can and do occur and that they demand immediate
attention. In a review of 2445 cases, Roos reported only 3 instances of major injuries of this
nature (0.12%) (10). In all 3 cases, the patients had full recovery. Delayed bleeding, usually
from a small subclavian branch or intercostal artery, is also seen. Roos reported 7 cases in
the same series (0.2%). The patients also had complete recovery. In Green's review of 136
patients, there were no major vascular injuries (1 1). UCLA reported one such injury in their
series over 10 years of operations. If the artery is injured during surgery via the transaxillary
route, a supraclavicular incision must be made to allow for proximal control of the vessel.
Major venous injuries can usually be addressed through the transaxillary incision and
repaired. If better exposure is needed, the medial third of the clavicle can be removed to aid
in visualization. Rarely if ever does the vein need to be ligated.

Injuries to other nerves occur, particularly the long thoracic. Roos reported a 0.12%
incidence, with two of the patients having complete recoveries and one lost to follow-up.
In Sharp's series of 36 patients, there was one such injury (12). Most of these tend to be
temporary, but a permanently winged scapula can occur. Treatment of these injuries with
nerve grafting can sometimes be successful. Injuries to the phrenic nerve are not common
and are more associated with the anterior approach, particularly with reoperative cases.
Most of these injuries result in temporary, subclinical diaphragmatic paralysis, although
complete division with permanent injury is possible.

The most common nerve injury is not a true complication but a by-product of the
operation; division of intercostal brachial cutaneous branches leading to cutaneous numb-
ness. This occurs to some extent in most patients, not unlike that which occurs with axillary
dissection for other disease states. It is usually well tolerated and resolves. ■§

Reports of patients with postoperative causalgia or other pains are also difficult to |

place into clinical perspective. In most series, they are unusual. Significant causalgia is a

usually attributed to brachial plexus injury and thus should parallel the incidence of that c

injury. In the Washington State workers' compensation study of patients with neurogenic <

TOS, 6% of patients were reported to have causalgia and 13% had "other pains." It is not >9

clear what these represented (13). Green's series of 136 cases had 3 patients with some J

form of postoperative vasospasm. Finally, postoperative Horner's syndrome occurs in «

from 0.5 to 2% of patients in most series. In almost all reported cases, it is self-limited. |

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF THORACIC OUTLET SURGERY 435

Entry into the pleural space occurs in as many as 30% of cases. This is usually recog-
nized and easily evacuated at the time of operation without the need for a chest tube. The
highest reported incidence of postoperative pneumothorax is 5%.

As mentioned previously, there are reports of injuries to all of the structures encoun-
tered during first rib resection. Thus, supraclavicular approaches can lead to thoracic duct
or even recurrent laryngeal nerve injury, although these injuries are rare. The risk of compli-
cations is increased with reoperative surgery. Many structures tend to be adherent in a par-
ticular pattern during these procedures, for example, the subclavian artery to the anterior
scalene muscle. Care must be taken in these cases to identify all structures adequately.

B. Treatment for Recurrent TOS

Recurrence of TOS following operative intervention is not uncommon. Published series
have reported rates as low 2.2%, but most are on the range of 15-20% (14-17). Defining
recurrence in this situation is frequently difficult, because it is not always clear that the
patient's symptoms ever improved. Scalene block has utility as a predictor of surgical
outcome for neurogenic TOS, with relaxation of the anterior scalene muscle approximating
the decompression achieved with first rib resection/scalenectomy. Patients are given a series
of injections of either lidocaine or saline and then pain with provocative maneuvers is
assessed [generally the elevated arm stress test (EAST)]. Machleder and colleagues reported
on 122 patients in whom this technique was used and found a 90% positive predictive value
for correlation with the clinical diagnosis of TOS (18). In addition, for patients undergoing
first rib resection for TOS, those with a positive scalene block had a much greater chance of a
good outcome (94%) than those with a negative preoperative scalene block (50%). This test
can be positive with other disorders, particularly radiculopathies, but is a useful adjunct to
not only the diagnosis of TOS but also in gauging the likelihood of surgical benefit. This can
reduce the number of recurrent or, more accurately, unimproved cases of TOS following
surgery.

It is also of paramount importance that patients have realistic expectations before
surgery. A discussion of the risks should include the possibility of no improvement, wors-
ening of the symptoms, or later recurrence. Armed with this information, the patient is in a
much better position to make an informed decision.

Recurrent symptoms tend to be similar to the original complaints, with paresthesias of
the hand and pain of the neck and shoulder being the most common manifestations. The
etiology of recurrence is usually not clear, although postoperative scarring is considered
one of the main culprits. In this setting, there is generally a relatively long asymptomatic
period (1-2 years) before the symptoms return. Several studies have looked at the
implications of a long posterior stump to the first rib, but there is little correlation with
return of symptoms. Other reported etiologies include middle scalene reattachment, ■§

calcified rib masses, and missed cervical ribs or cervical rib stumps. Some have drawn a |

distinction between a spontaneous recurrence, attributable to scar, and a recurrence sec- a

ondary to a traumatic insult. In the latter situation, the patient again tends to have the c

original symptoms, despite the fact that the thoracic outlet has already been decom- <

pressed. A whiplash type injury frequently occurs in this scenario. >5

The workup for recurrence is essentially the same as for untreated disease. Special J

emphasis should be placed on ruling out other causes, as a percentage of patients failing «

treatment will have done so on the basis of a faulty diagnosis. Iatrogenic injury to the plexus, |

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

436

RIGBERG and FREISCHLAG

carpal tunnel syndrome, tendinitis, cervical arthritis or spine injury must be sought before
treatment for the recurrence is started. A conservative plan is initially undertaken, although
the overwhelmingly positive response seen with physical therapy in TOS patients who avoid
the original operation is not reproduced here. It is worth noting that intense postoperative
physical therapy, if properly prescribed and performed, is thought to decrease the incidence
of recurrent symptoms after both initial and "redo" operations. If conservative methods fail
in treating recurrent symptoms, reoperation is considered, although it should be noted that
only a 50% improvement rate is quoted for these patients. In addition, at least a year should
be allowed to pass before surgical intervention is again performed.

Although clinical practices vary, many surgeons have adopted an algorithm for
reoperation (Fig. 5). If the original operation was via a transaxillary approach, a supra-
clavicular approach is taken. Care is taken to identify any remaining first rib and to resect
it all the way to the transverse process. If the patient's original operation did not include
first rib resection, it is removed now. Most surgeons also add some form of neurolysis to
the reoperation, whereby the scar around the nerves is carefully removed. Some surgeons

Minimum of 6 weeks

Physical Therapy

Strengthen: trapezius. Jevator scapulae.
sternoc leidomastoi d .

Relax: middle scalene. subclavius,
pectoral is muscles.

(m proved Ergonomics

Bolulinum injection
with electrophysiologic/
fluoroscopic may reduce
symptoms and allow longer
trial of conservative
therapy.

Improvement

Mo Improvement

Incorporate new ergonomics into wrjrt< h etc..
Physical therapy as needed.

Improved

Transaxillary first rib resection

Mo improvement or recurrent
symptoms

Consider supraclavicular sea lonec-to my

I

•5

Figure 5 This algorithm incorporates the important modalities for treating TOS. This includes
treating recurrence of symptoms following operative decompression. Note should be made that
physical therapy should be utilized following operation and can decrease the incidence of recurrent
symptoms. Physical therapy should also be considered for treating recurrent symptoms before
further operation is undertaken.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF THORACIC OUTLET SURGERY 437

remove the middle scalene during the course of reexploration. If the supraclavicular
approach was already used, the transaxillary is utilized for the second operation. Again,
care is taken to remove the entire first rib, any remaining attachments and scar. Neurolysis
can be performed via this route, with some advocating that it be done over the supra-
clavicular route, particularly when ulnar symptoms predominate.

Several methods have been tried to prevent the formation of scar tissue and adhesions
following thoracic outlet decompression, particularly in the face of reoperation. After the
nerve roots are cleared, they are usually covered with the overlying adipose tissue, although
the benefit of this technique has never been documented. Similarly, care is taken to replace
the scalene fat pad, but this does not appear to influence the formation of scar. Attempts to
control scarring with the administration of exogenous agents, including steroids, have also
been disappointing. Sheets of polytetrafluoroethylene (PTFE) have been used to cover the
nerves, but this technique has been abandoned by most and probably leads to additional scar
formation. The use of hyaluronic acid gels showed some promise several years ago, but it is
not clear whether these products are particularly helpful and they are not approved by the
U.S. Food and Drug Administration for this use.

Brief mention should be made here of the treatment of bilateral TOS. For neurogenic
disease, the initial procedure is performed on the most symptomatic side. Only 10% of
patients require bilateral operations for this disorder, and the second side should be treated
no sooner than 12 months following the first. Patients who have bilateral operations within
less than 12 months will have difficulty regaining strength and stability of the neck and
shoulders and take a longer time to recuperate from the second operation. Even more rarely
do patients have venous compression on the opposite side. If this is demonstrated by
provocative noninvasive testing, first rib decompression can be done prophylacticaly no
sooner than 12 months later. Bilateral arterial TOS is practically never seen.

V. CONCLUSIONS

It has been demonstrated that decompression of the thoracic outlet can be safely
accomplished in most patients with TOS. However, two factors dictate that surgeons
performing these procedures be well trained in them and strive to minimize morbidity. The
first is the abundance of critical structures in a confined space and the devastating
consequences of harming them. The second is the subjective nature of neurogenic TOS
and the importance of not replacing a patient's symptoms with a different problem, a
situation that is fortunately not common. For a controversial clinical entity such as TOS,
it is difficult to objectively demonstrate posttreatment improvement. However, it is not
difficult to demonstrate many of the potential complications of therapy. Careful patient
selection and meticulous technique are needed to ensure that the risk-benefit ratio of these
procedures continues to justify operative intervention for neurogenic TOS in particular. 1

REFERENCES I

1. Wilbourn AJ. Thoracic Outlet Syndrome is overdiagnosed. Muscle Nerve 1999; 22:130-136. g

2. Roos DB. Thoracic Outlet Syndrome is underdiagnosed. Muscle Nerve 1999; 22:126-129. Tj

3. Peet RM, Hendricksen JD, Anderson TP, et al. Thoracic outlet syndrome: Evaluation of the 3
therapeutic exercise program. Mayo Clin Proc 1956; 31:281-287. 2

4. Aligne C, Barral X. Rehabilitation of patients with thoracic outlet syndrome. Ann Vase Surg I
1992; 6:381-389. |

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

438 RIGBERG and FREISCHLAG

5. Machleder HI. Vascular Disorders of the Upper Extremity. 3d ed. Mt Kisco, NY: Futura Press,
1999.

6. Dale WA. Thoracic outlet compression syndrome. Arch Surg 1982; 164:149-153.

7. Wilbourn AJ. Thoracic outlet syndrome surgery causing severe brachial plexopathy. Muscle
Nerve 1988; 11:66-74.

8. Horowitx SH. Brachial plexus injuries with causalgia resulting from transaxillary rib resection.
Arch Surg 1985; 120:1189-1191.

9. Cherington M, Happer I, Machanic B, et al. Surgery for thoracic outlet syndrome may be
hazardous to your health. Muscle and Nerve 1986; 9:632-634.

10. Roos DB. Thoracic outlet nerve compression. In: Rutherford RB, ed. Vascular Surgery. 3d ed.
Philadelphia: Saunders, 1989:858-875.

11. Green RM, McNamara J, Ouriel K. Long-term follow-up after thoracic outlet decompression:
An analysis of factors determining outcome. J Vase Surg 1991; 14:739-746.

12. Sharp WJ, Nowak LR, Zamani T, et al. Long-term follow-up and patient satisfaction after
surgery for thoracic outlet syndrome. Ann Vase Surg 2001; 15:32-36.

13. Franklin GM, Fulton-Kehoe D, Bradley C, et al. Outcome of surgery for thoracic outlet
syndrome in Washington State workers' compensation. Neurology 2000; 54:1252-1257.

14. Lindgren KA, Leino E, Lepantalo M, et al. Recurrent thoracic outlet syndrome after first rib
resection. Arch Phy Med Rehabil 1991; 72:208-210.

15. Roos DB. Recurrent thoracic outlet syndrome after first rib resection. Acta Chir Belg 1980;
79:363-372.

16. Sessions RT. Recurrent thoracic outlet syndrome: Causes and treatment. South Med J 1982;
75:1453-1461.

17. Sanders RJ, Monsour JW, Gerber FG, et al. Scalenectomy versus first rib resection for treat-
ment of the thoracic outlet syndrome. Surgery 1979; 85:109-121.

18. Jordan SE, Machleder HI. Diagnosis of thoracic outlet syndrome using electrophysiologically
guided anterior scalene blocks. Ann Vase Surg 1998; 12:260-264.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

25

Stroke as a Complication of Noncerebrovascular Surgery

Sukru Dilege

Istanbul Medical Faculty, Istanbul, Turkey

Matthew I. Foley and Gregory L. Moneta

Oregon Health & Science University, Portland, Oregon, U.S.A.

Although improvements in perioperative care have resulted in the ability to perform
complex operations in elderly, high-risk patients, serious complications can occur. One
such complication is perioperative stroke. Most strokes occur 3-30 days following a
surgical procedure. Perioperative stroke complicating noncerebrovascular surgery is
uncommon, but it can be associated with devastating disability and high mortality (1).
With the exception of death, long-term neurological disability following postoperative
stroke is undoubtedly an elderly patient's most feared perioperative complication (2).
Patients who have suffered a perioperative stroke can present a formidable social and
financial burden to the family and community. All surgeons who operate on elderly, high-
risk patients must be familiar with the risk factors for stroke, the incidence of perioperative
stroke associated with various procedures, and variables that can be modified to reduce the
risk of perioperative stroke.

I. STROKE DEMOGRAPHICS

Stroke is the third leading cause of death in the United States. In 1999, 167,366 people died "g

of stroke-related causes (3). Each year, about 600,000 people suffer a stroke, and 82% of &

these strokes are initial events. Persons with hemorrhagic strokes have a higher 30-day a

mortality than those with ischemic strokes (4). Among patients with a first ischemic stroke, -c

the stroke itself is the most common cause of early death (5). <j

A marked decrease in stroke mortality occurred in the United States, Canada, and >3

western Europe during the 1970s and early 1980s. Stroke mortality, however, actually 4j

increased in eastern European countries (6-8). Japan and Finland, especially, have been 2

very successful in reducing their stroke mortality rates (6). In the United States, the stroke- |

related death rate decreased 13.0%, but the actual number of stroke-related deaths increased ©

439 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

440 DILEGE et al.

8.6% from 1989 to 1999 (3). Although the decline in stroke mortality rates in the United
States is encouraging and the lowest in the industrialized world, recent evidence suggests
that stroke rates have stabilized (3,4,6,9).

In addition to mortality, stroke remains a major source of disability. In 1999, more than
1,100,000 Americans who suffered strokes were affected with functional limitations, many
with moderate to severe permanent neurological deficits (3).

II. ETIOLOGY OF STROKE

Vascular events were the cause of more than 90% of acute neurological deficits in the late
1970s (10) and continue to be so today. In the National Heart Lung and Blood Institute's
Atherosclerosis Risk in Communities cohort, causes of stroke in middle-aged adults were
reported as ischemic in 83%, intracerebral hemorrhage in 10%, and subarachnoid hemor-
rhage in 7% (4). In a study by Petty et al. (11), the underlying cause of ischemic stroke was
cervical or intracranial large vessel atherosclerosis with stenosis in 16%, cardioembolic in
29%, lacunar in 16%, and other in 3%. The cause was uncertain in 36% of patients.
Principal etiologies of stroke in young patients are extracranial arterial dissection, cardi-
oembolism, premature atherosclerosis, hematological and immunological diseases,
migraines, and drug abuse (12). Heavy consumption of alcohol predisposes to both
hemorrhagic and nonhemorrhagic stroke (13). Less than half of thromboembolic events
appear to originate from the extracranial cerebrovascular system and thus are potentially
amenable to preoperative surgical correction (10,14).

III. RISK FACTORS FOR STROKE

Sacco et al. published a consensus in 1997 entitled "Risk Factors in Stroke"(15). This and
other reports have concluded that risk factors for stroke have not changed for the last 20
years. Nonmodifiable and modifiable risk factors have been identified and are listed in
Table 1.

A. Nonmodifiable Risk Factors

Age is the single most important nonmodifiable risk factor for stroke. Stroke risk increases

with age. Stroke is more common in men than women except for in the very old. In the

United States, approximately 1% of people between ages 65 and 75 die from a stroke

annually. Some 22% of men and 25% of women die within the first year of having a stroke.

In 1999, a total of 38.5 and 61.5% of deaths from stroke were in men and women,

respectively (3). In the Framingham Study, a family history of stroke was associated with

an increased stroke risk (16). Age-standardized mortality rates for all stroke subtypes are -o

higher in blacks than in whites (17). Blacks have a 38% greater risk of first stroke than whites |

(4). Stroke incidence and mortality rates vary dramatically from one population or racial |

group to another. However, some of these risks for stroke may be related to environmental g>

factors or inherited risk factors other than race. «*

4

B. Modifiable Risk Factors J

One of the most important modifiable risk factors for stroke is hypertension, even at 2

borderline levels. The risk of stroke seems to be directly related to the magnitude of blood |

pressure elevation (18,19). An analysis of seven studies has estimated the stroke risk from @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

STROKE AS A COMPLICATION OF SURGERY

441

Table 1 Risk Factors for Stroke

Non-modifiable

Age, male sex, race

Modifiable (potentially)

Hypertension
Cardiac disease

Aortic arch plaque

Diabetes

Low HDL a

Elevated homocysteine

Cigarette smoking

Alcohol abuse

Cocaine

Obesity

Physical inactivity

Atrial fibrillation
Cardiac valve abnormalities
Coronary artery disease
Left ventricular hypertrophy
Congestive heart failure

High-density lipoprotein.

the lowest to the highest blood pressure ranges to increase about 10-fold (15). Isolated
systolic hypertension in the elderly may engender less risk of stroke than sustained diastolic
hypertension (20). About half of those who have a first stroke have blood pressures higher
than 160/95 mmHg (3). Women do not appear to tolerate hypertension any better than men.
A higher percentage of men have high blood pressure until age 55, when the prevalence
increases in women and surpasses that in men (3). In women taking oral contraceptives,
hypertension is two to three times more common than in age-matched controls. In a
metanalysis of 17 treatment trials of hypertension throughout the world involving nearly
50,000 patients, there was a 38% reduction in all strokes and a 40% reduction in fatal stroke
in subjects treated for their hypertension (21). It appears that optimal prevention of late-life
stroke will likely require control of midlife blood pressure (22).

Cardiac disease, especially atrial fibrillation (AF), has a definite effect on stroke rates.
It is estimated that almost half of all cardioembolic strokes occur in the setting of AF. The
incidence and prevalence of AF increase with age. A study from Denmark suggests that
patients with an in-hospital diagnosis of AF have an increased risk of stroke that is
greatest during the first year after discharge (23). In the Framingham Study, patients with
nonvalvular AF had 3-5 times the risk of stroke (24). To prevent stroke in AF, it is advised
to offer warfarin to patients and reserve aspirin for young subjects at low risk of stroke
and those with contraindications to warfarin use (25). Cardiac valve abnormalities
(especially mitral stenosis), coronary artery disease, left ventricular hypertrophy, and
congestive heart failure are additional cardiac risk factors for stroke.

Proximal atherosclerotic plaque of the aorta is also associated with an increased risk
of ischemic stroke in the elderly. One study suggests that aortic plaques are significantly
more frequent in men than in women (26).

Diabetes mellitus is another risk factor for stroke. However, other risk factors for
stroke such as hypertension, obesity, and hyperlipidemia are prevalent in patients with

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

442 DILEGE et al.

diabetes (15). Stroke risks from hypertension and diabetes mellitus are difficult to separate
and probably have joint effects on the development of atherosclerosis in the cerebrovas-
cular system.

Although there is a positive correlation between total cholesterol and the extent of
extracranial carotid atherosclerosis, analysis of the EUROSTROKE project could not
identify an association between total cholesterol with fatal or nonfatal, hemorrhagic, or
ischemic strokes (27,28). HDL cholesterol was inversely related to stroke in men but not in
women (28). Increased HDL levels are associated with a reduced risk of ischemic stroke in
the elderly (29).

Elevated levels of homocysteine have been found to be associated with an increased risk
of carotid artery disease and stroke. Recently, the British Regional Heart Study showed a
strong, independent, graded relationship between homocysteine levels and stroke risk
among middle-aged men (30). Randomized, controlled trials of homocysteine-lowering
interventions are in progress to determine whether stroke incidence can be reduced (31).

Significant cigarette smoking increases the risk of stroke twofold. There is a dose-
response relationship between cigarette smoking and the extent of extracranial cerebro-
vascular atherosclerosis (32). Moderate consumption of alcohol may be protective against
cardiovascular disease, including stroke. Heavy alcohol consumption, however, increases
the risk of intracranial hemorrhage (33). Cocaine abuse is also a major social and human
health problem associated with stroke (34).

The independent effects of oral contraceptives are unclear. According to most studies,
low-dose oral contraceptives (< 50 ug estrogen) do not increase risk of stroke (35,36). In
the Heart and Estrogen-progestin Replacement Study (HERS), hormone therapy with
conjugated equine estrogen and progestin had no significant effect on the risk of stroke
among postmenopausal women with coronary artery disease (37).

Obesity and physical inactivity are associated with an increased stroke risk. The role of
lifestyle modification in the primary prevention of stroke cannot be overemphasized (38).

C. Carotid Artery Disease

Asymptomatic carotid artery atherosclerosis is common; it is often discovered incidentally
or found in a patient with a cervical bruit. The prevalence of carotid stenosis ranges from
0.5% in people under 60 years old up to 10% in those over age 80 (39). Although the
prevalence is high, the degree of stenosis in most is not severe (40). Patients of advanced
age, and those with histories of cigarette smoking, hypertension, and low HDL have a
much higher incidence carotid artery stenosis (41,42).

The prevalence of a carotid bruit is 5% in persons more than 50 years old. However,
less than a quarter of patients with a carotid bruit have greater than 50% internal carotid
artery stenosis. A carotid bruit has a sensitivity of 63-76% and a specificity of 61-76% for
predicting > 70% carotid artery stenosis (43).

Carotid ulceration has been implicated as a risk factor for stroke, especially when the

ulcerations are large. This may be true even without concomitant severe stenosis (44,45). -c

However, the added risk of ulceration, if any, is not firmly established. Certain plaque char- <j

acteristics, such as echolucency, have also been implicated as a risk factor for stroke (46). >3

The risk of stroke among patients with asymptomatic carotid artery stenosis is relatively 4j

low. Some 45% of strokes in patients with stenosis of 60-99% are attributable to lacunar 2

infarcts or cardioembolism (47). Several different randomized trials have compared aspirin |

therapy with endarterectomy in patients with asymptomatic carotid artery stenosis (48-5 1). @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

STROKE AS A COMPLICATION OF SURGERY 443

The Asymptomatic Carotid Atherosclerosis Study (ACAS) randomized 1662 patients with
asymptomatic carotid stenosis of 60% or greater to medical management or medical
management plus carotid endarterectomy (51). This study suggested a benefit of surgery,
reducing the risk of ipsilateral stroke or death at 5 years from 11 to 5.1%. The absolute
reduction in major stroke and death over 5 years in ACAS was insignificant. The benefit of
endarterectomy for asymptomatic carotid stenosis remains controversial at present (52).

D. Transient Ischemic Attacks

A transient ischemic attack (TIA) is a temporary focal neurological deficit caused by a brief
interruption of local cerebral blood flow. A significant percentage of TIAs occur in the
absence of detectable significant extracranial carotid artery disease. Patients with TIAs have
significantly more hypertension, cardiac disease, and diabetes than age-matched controls
(53). One study demonstrated a 7% annual stroke risk for patients experiencing TIAs (54).
Recent or multiple TIAs have a higher short-term risk for ischemic stroke than a remote
TIA, and the same may be true for "crescendo" TIAs (15,55). The risk of stroke associated
with TIAs is highest in the first year following TIA.

E. Previous Stroke

Patients with prior strokes are at much higher risk for a new cerebral infarction than
patients without prior strokes (3). There is a 5-20% annual recurrence risk for patients
experiencing stroke (3,53,56).

At least three prospective randomized clinical trials comparing carotid endarterectomy
versus medical management for the treatment of symptomatic carotid artery stenosis have
been completed (54,57,58). The North American Symptomatic Carotid Endarterectomy
Trial (NASCET) demonstrated a major advantage of endarterectomy plus medical manage-
ment over medical management alone in patients with greater than 70% stenosis and
neurological symptoms referable to the stenosed artery (54). The cumulative risk of
ipsilateral stroke at 2 years in patients with 70-99% stenosis was 26% in the medical group
and 9% in the surgical group. Endarterectomy was also beneficial in symptomatic patients
with 50-69% stenosis, but the absolute risk reduction (6.5%) was not as pronounced (59).
The European Carotid Surgery Trial (ECST) and Veterans Affairs Cooperative Sympto-
matic Trial showed similar results (57,58).

IV. STROKE IN NONCEREBROVASCULAR SURGERY

The incidence and mechanism of stroke in patients undergoing noncerebrovascular surgery
depends on the primary operative procedure.

1

P

A. Etiology of Perioperative Stroke J

The pathogenesis of perioperative stroke is often not known with certainty. The currently c

recognized ischemic stroke mechanisms are decreased perfusion, embolism, and throm- <j

bosis. Lacunar infarction (infarct size <1.5 cm) may be associated with all three >9

mechanisms. 41

The mechanism of perioperative stroke is predominantly embolic, although hypoper- 2

fusion may play a role (60). Acute hypotension in the setting of chronic hypertension has |

been implicated as a cause of stroke (61). A patient who undergoes a difficult procedure @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

444 DILEGE et al.

associated with significant intraoperative blood loss and hypotension followed by a post-
operative neurological deficit will generally have the stroke ascribed to intraoperative
hypotension. However, such patients are unusual. Severe alteration in blood pressure in
direct proximity to stroke occurrence is an infrequent finding in perioperative cerebral
infarction. Reviews of patients with perioperative cerebral infarction following aortoiliac
procedures indicate that neurological events almost always occur after a variable lucid
interval, not intraoperatively (1,62,63). Hart and Hindman (64) identified 12 perioperative
strokes associated with 24,500 general surgical procedures. Although intraoperative hypo-
tension was frequent, the onset of the neurological deficit was intraoperative in only 17% of
the cases; the deficit clearly occurred postoperatively in 83%. Larsen et al. (65) found that 6
of 2463 patients suffered perioperative stroke after noncardiac, noncerebrovascular surgery;
all 6 of these did so late in the postoperative period, between 5 and 26 days after surgery.
Gerraty et al. (66) feel that cerebrovascular hemodynamics are important in that patients
with severe symptomatic carotid disease may be at high risk for perioperative watershed
infarction.

One explanation of the infrequent association of intraoperative hypotension with
perioperative stroke is that agents used for general anesthesia may have cerebral protective
effects. Furthermore, most cases of intraoperative hypotension are transient and may be of
little overall hemodynamic significance. Hypotension in the postoperative recovery period
can, however, be associated with stroke (67).

Embolism to the brain is often cardiac in origin. Commonly recognized cardiac
sources of embolism include atrial fibrillation, sinoatrial disorder, recent acute myocardial
infarction (AMI), subacute bacterial endocarditis, cardiac tumors, and valvular disorders
affecting both native and artificial valves (68). Hart and Hindman (64) identified cardiac
embolization as the underlying mechanism of stroke in 42% of their noncardiac, non-
cerebrovascular perioperative stroke patients. Larsen et al. (65) and Parikh et al. (69)
emphasized the role of atrial fibrillation in perioperative stroke. Platelet emboli from
atheromatous arterial plaques have been suggested as a cause of perioperative stroke as
well (64,67).

The role, if any, of a perioperative hypercoagulable state in the occurrence of
perioperative cerebral infarction is unknown. Recent evidence suggests that perioperative
alterations in blood rheology and transient postoperative hypercoagulable states may
contribute to the risk of postoperative stroke. Such speculation is consistent with the
observation that perioperative stroke is nearly always thrombotic or embolic in origin.

Blood rheology appears chronically altered in patients with previous stroke, TIAs, or
significant risk factors for cerebrovascular disease (70-72). Whole blood and plasma
viscosities are significantly elevated in such patients as compared with normal controls.
The effect appears to be independent of the hematocrit and primarily associated with
elevated plasma fibrinogen and a decreased albumin/globulin ratio (73). ■§

The postoperative state is also associated with a number of alterations in the g

coagulation system. Gibbs et al. (74) compared postoperative changes in procoagulant, a

anticoagulant, and antifibrinolytic factors in patients undergoing peripheral vascular c

surgery. There was an increase in plasma fibrinogen and factor VIII levels, whereas <j

antithrombin III and protein C levels decreased. The results suggest that peripheral >3

vascular procedures are associated with an increased potential for thrombosis due to 4j

increases in procoagulant factors. Specific acquired hypercoagulable conditions such as 2

antiphospholipid antibodies, which clearly predispose surgical patients to postoperative |

thrombotic complications, are being recognized with increasing frequency (75). @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

STROKE AS A COMPLICATION OF SURGERY 445

One can postulate that the increase in perioperative blood viscosity and coagulation
factor abnormalities following vascular and general surgical procedures may predispose
patients to thrombotic or embolic stroke.

B. Cardiac Surgery

Although stroke rates have declined with the improvements in surgical techniques and
advances in myocardial protection, stroke remains a major complication of cardiac surgery.
The majority of these strokes result in temporary, minor cognitive and psychiatric
dysfunction. In a prospective study of perioperative stroke in cardiac surgery, less than
5% of neurological deficits consisted of a central focal deficit (76). In experienced centers,
the contemporary stroke risk for patients undergoing coronary artery bypass grafting
(CABG) is reported to be less than 2%. Estimates of the perioperative stroke risk of CABG
in the presence of hemodynamically significant carotid stenosis generally vary from 6 to 16%
(77). The risk of stroke associated with cardiac surgery clearly increases with patient age.
Age > 75 is associated with an increased risk of nonembolic stroke after myocardial
revascularization (76,78,79). The risk of stroke in this group reaches a disturbing incidence
of 3.2-10% (80-83). In addition to previous stroke or transient ischemic attack, peripheral
vascular disease, emergency operation, diabetes, and a left ventricular ejection fraction
<40% are independent predictors of stroke after cardiac surgery (84,85). Duration of
cardiopulmonary bypass in excess of 2 h increases the probability of neurological damage
(78,86). Engelman et al. found that perioperative TIA or stroke during the cardiac operation
was associated with postoperative low cardiac output and atrial fibrillation (80). Finally,
Hogue et al. (87) reported that female gender was associated with a 6.9-fold increased risk of
early stroke and a 1.7-fold increased risk of delayed stroke.

The incidence of stroke during valvular surgery has decreased with time, despite the
increased prevalence of risk factors. The overall risk of perioperative stroke with valve
replacement is two to three times greater than the risk with coronary artery bypass.
Patients with single valve replacement have a lower incidence than combined procedures
with CABG (1.2 vs. 7.6%) (88). Causes of stroke in valve surgery are atherosclerotic
emboli, septic emboli, and shock (79).

Early studies have mainly focused on intraoperative events, but symptoms may
develop later in the postoperative period (88). Some 70% of strokes associated with
cardiac surgery occur intraoperatively; the remaining 30% develop more than 24 h post-
operatively (88-90). Intraoperative events are attributed most frequently to technical prob-
lems with aortic clamping or cannulation, inadequate venting of the heart of air and debris
at the completion of the reconstruction, and occasionally hypoperfusion associated with
cardiac bypass. Borger et al. have suggested that macroemboli from the ascending aorta
are the predominant cause of stroke during coronary bypass surgery (79). •§

The role of extracranial cerebrovascular disease in the etiology of stroke associated with g

cardiac surgery is not clear at present. Generally, detectable carotid artery stenosis coexists a

with significant coronary artery disease in 15% of coronary disease patients. Schwartz et al. -c

found carotid atherosclerosis to be a risk factor for hemispheric stroke in patients under- <j

going cardiopulmonary bypass (91). The risk of hemispheric stroke in patients with >9

unilateral 80-99% stenosis, bilateral 50-99% stenosis, or unilateral occlusion with con- 41

tralateral 50% or greater stenosis was 5.3%. Scotnicki et al. (92) reported that patients with 2

as asymptomatic carotid bruit can safely undergo coronary artery surgery. In the group of |

patients without preoperative neurological symptoms, postoperative neurological deficits @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

446 DILEGE et al.

were rarely caused by carotid occlusive disease. Patients with asymptomatic bruits can be
safely screened with ultrasonic carotid duplex scanning and do not require arteriography
prior to cardiopulmonary bypass (93).

C. Noncerebrovascular Vascular Surgery

The increased mortality observed among claudicants is most often a consequence of
atherosclerotic disease in other vascular beds. Myocardial infarction and cerebrovascular
accidents are the cause of death in at least 50% of patients with peripheral vascular disease.
About 20% of vascular surgery patients have a cervical bruit (94). Using continuous-wave
Doppler examinations, Hennerici et al. (95) found a 32.8% incidence of vertebral or carotid
artery disease (> 50% stenosis) in 325 patients undergoing peripheral vascular surgery. In
264 patients with severe coronary artery disease documented by angiography, these
investigators found similar carotid disease in only 6.8%. Barnes and Marszalek (96) detected
> 50% carotid stenosis in 17.2% of 116 patients undergoing peripheral vascular recon-
struction and in 10.6% of 198 patients undergoing coronary bypass.

Plate and Hollier et al. (97) found 8.2% late mortality from stroke in 1112 patients
operated upon for abdominal aortic aneurysms. The incidence of cerebrovascular accident
(CVA) was 4.2 and 9.5% within 5 and 10 years, respectively. Advanced age, hypertension,
and heart disease carried a greater risk for this complication.

Despite the prevalence of extracranial vascular disease and associated risk factors for
stroke in peripheral vascular surgery patients, the overall incidence of perioperative stroke in
noncerebrovascular vascular surgery patients is surprisingly low. Perioperative strokes
occur in around 1 % of peripheral vascular surgery patients, as shown in Table 2. Harris et al.
(98) found only 13 perioperative neurological events associated with 1390 (0.9%) non-
carotid vascular surgery procedures. There were 2 TIAs, 10 anterior circulation strokes, and
1 posterior circulation stroke. The neurological deficit developed in the immediate post-
operative period in 3 1 %, more than 4 h but less than 72 h postoperatively in 54%, and within
3-14 days postoperatively in 15%. Of these strokes, 27% were fatal. These investigators
established that new anterior circulation strokes in vascular surgical patients tended to be
associated with intra-abdominal procedures, perioperative hypotension, and the presence of

Table 2 Incidence of Perioperative Strokes Following
Noncerebrovascular Surgery

Procedure

Patients

Stroke (%)

Cardiac valve replacement

1,343

4.1

CABG

41,769

1.5

Elective AAA

2,210

0.7

Ruptured AAA

437

1.3

Aortofemoral bypass

2,464

0.8

Infrainguinal bypass

2,246

0.5

Thoracoabdominal aneurysm

706

1.8

Other"

26,336

0.06

a Includes general surgical and subspecialty procedures
(orthopedic, urologic).

Abbreviations: CABG, coronary artery bypass grafting; AAA,
abdominal aortic aneurysm.

i
2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

STROKE AS A COMPLICATION OF SURGERY 447

a greater than or equal to 50% ipsilateral internal carotid artery stenosis. Many vascular
surgery patients are hypertensive, have suffered a previous stroke, or have significant cardiac
dysfunction, all of which are risk factors for stroke independent of carotid artery disease.
Preoperative knowledge and correction of modifiable risk factors can potentially reduce the
incidence of perioperative stroke (99).

D. General Surgery

Perioperative stroke as a complication of general surgery is infrequent, with a reported
incidence between 0.08 and 2.9% (69,100). Landercasper et al. (67) reported no perioper-
ative strokes in 7517 consecutive patients without a prior history of stroke who underwent
nonneurosurgical, noncardiac, and noncerebrovascular procedures while under general
anesthesia. The incidence of perioperative stroke in 173 patients with a previous history of
stroke was 2.9%. In another study, it was found that 2.1 % of 279 patients with a history of
cerebrovascular disease developed a new stroke in the perioperative period (65).

Ischemic strokes after general surgery most commonly occur after an asymptomatic
interval (69,65). Parikh et al. (69), in a study involving 24,641 patients, found that majority
of perioperative strokes occurred late in the postoperative period. The average time to the
occurrence of perioperative stroke is 7 days (2). Patients with previous cerebrovascular
disease, atrial fibrillation, hypertension, advanced age, or atherosclerosis were found to have
an increased risk. Limburg et al. (100) described three major risk factors for perioperative
stroke: prior CVA, chronic obstructive pulmonary disease, and peripheral arterial disease.

V. MINIMIZING PERIOPERATIVE STROKE

Despite the low incidence of perioperative stroke in noncerebrovascular surgery, certain
steps may decrease the risk of perioperative stroke.

A. Patient Selection

An important factor in reducing the incidence of perioperative stroke is careful patient
selection. A prior neurological event, carotid artery stenosis, diabetes mellitus, and
advanced age have been found to increase susceptibility to perioperative stroke in many
studies. By the avoidance of all but essential surgery in stroke-prone patients (patients of
advanced age, especially greater than 75 years and those with a history of recent myocardial
infarction, previous stroke, atrial fibrillation, or known thrombotic tendencies), the
incidence of perioperative stroke can likely be reduced. Stroke is an important complication
in patients with acute myocardial infarction, occurring in 1-3% of all infarctions. Most
strokes occur in the first weeks after the infarction, but the risk for stroke remains for an
indefinite time. Therefore all major surgical interventions should be postponed as long as
possible in such patients. -g

Preoperative optimization of the patient by identifying and treating coexisting medical g

conditions such as coronary artery disease, hypertension, and atrial fibrillation is very a

important for reducing perioperative stroke (2). Preoperative patients with atrial fibrillation c

should be considered for conversion to sinus rhythm or for anticoagulation (69). Anesthesia <j

cj

and surgery are best delayed for 4-6 weeks after an acute CVA. -

I

B. Intraoperative Management a

Although most perioperative strokes do not occur intraoperatively, some obviously do. §

Certainly in cardiac surgery, extreme care must be exercised in manipulating the ascending @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

448 DILEGE et al.

and transverse aorta, and the heart must be well flushed of air and debris before circulation is
restored. Careful attention should be given to maintaining proper mean arterial pressure
during cardiac bypass, and pump times should be kept as short as possible (78,86).

Intraoperatively, the noncardiac surgery patient should be maintained as close to his
or her resting physiological state as possible. Both mean arterial pressure and arterial
carbon dioxide should be maintained within the patient's usual range. Blood pressure
control and optimal cerebral oxygen delivery are crucial factors to prevent stroke. Wide
variations may alter cerebral autoregulation and increase susceptibility to stroke. Blood
glucose levels should be maintained at normoglycemic levels, because hyperglycemia
exacerbates ischemic neurological damage.

Choice of anesthetic technique and agents may be of importance in high-risk patients.
The major volatile anesthetic agents produce cerebral vasodilatation and decrease cerebral
oxygen consumption. These agents can cause stealing of blood from peri-infarcted areas in
patients with recent cerebral infarctions (2). Anesthetic agents or techniques may influence
the incidence of perioperative stroke, but results from prospective controlled trials are not
available. Hypocapnia and hypercapnia should be avoided (2).

C. Prophylactic Carotid Endarterectomy

Current data suggest that the > 80% asymptomatic lesions have the highest potential to
result in a neurological event during follow-up (Table 3). However, even if prophylactic
carotid endarterectomy were to be offered to all patients with asymptomatic carotid artery
stenosis, as many as one-third of strokes in this population would remain unavoidable
because of their occurrence in a different vascular territory. Moreover, some additional
strokes would result from surgery itself (106). The absolute reduction in major stroke and
death due to surgery over 5 years in ACAS was small and surgery may result in other
nonfatal complications (cranial nerve injury, myocardial infarction). In addition, the low
complication rate of the ACAS-selected surgeons is not likely to reflect the typical risk of
endarterectomy in the community. Therefore the optimal management of high-grade
asymptomatic carotid disease prior to noncerebrovascular surgery still remains question-
able.

In 1998, Benavante et al. (39) reported a metanalysis of six randomized clinical trials.
The results of this metanalysis suggest that carotid endarterectomy should not be routinely
recommended for unselected patients with asymptomatic carotid stenosis despite the
substantial reduction in the risk of ipsilateral stroke by surgery. The incidence of ipsilateral
stroke was relatively low in those patients who did not undergo the operation; hence the

Table 3 Annual Stroke Risks in Patients with >80% -o

Asymptomatic Internal Carotid Artery Stenosis E

8

I

■c

Author Stroke risk (%)

Cambers and N orris (101) 5.5

Bogousslavsky et al. (102) 4.2

Henneirci et al. (103) 8.1

Moneta et al. (104) 12; 4 a

Caracci et al. (105) 9

a First year, 12%; second year; 4%.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

STROKE AS A COMPLICATION OF SURGERY 449

benefit of carotid endarterectomy will remain small until high-risk subgroups are
identified.

It seems reasonable to perform carotid endarterectomy before another operative
procedure under certain circumstances. First, in patients with symptomatic carotid lesions,
the performance of carotid endarterectomy can be justified on the basis of the carotid disease
alone. Second, the patient should be a good operative risk and the operating surgeon should
have a record of a low complication rate with carotid artery surgery. Third, the performance
of carotid endarterectomy should not place the patient at significantly increased risk
resulting from the inherent delay in performing the other indicated operative procedure.

D. Combined Carotid/Coronary Surgery

There is a close relationship between carotid and coronary artery disease. O'Donnell et al.
(107) pointed out that 66% of patients undergoing carotid artery surgery had clinical
evidence of coronary artery disease. Hertzer et al. (108) reported that coronary artery
bypass grafting (CABG) might be indicated in as many as 37% of carotid endarterectomy
patients. In the presence of significant carotid artery stenosis, the stroke rate of patients
undergoing CABG varies from 6 to 16% (109). It is generally accepted that combined
carotid endarterectomy and CABG procedures have higher stroke and death rates than
isolated procedures. This is hardly surprising, since each operation carries an inherent risk
of stroke. Brow et al. (110). found that risk factors are more prevalent in the combined
endarterectomy and CABG group. Some investigators suggest that the difference between
isolated and combined procedures is not significant if the carotid disease is asymptomatic
(111). Patients with significant bilateral carotid artery disease have a worse stroke risk
whether undergoing combined or isolated procedures (1 12,1 13). A metanalysis of 1 1 series
determined that patients with unilateral carotid artery stenosis have a 6.9% stroke risk for
combined procedures compared to 12.7% for patients with bilateral disease. The highest-
risk group included patients with an occluded carotid and contralateral stenosis. These
patients experienced a 29% stroke rate with combined procedures (114).

The authors of a retrospective German study of 313 simultaneous carotid endarter-
ectomy and myocardial revascularization procedures concluded that combined procedures
are justified. The risk of myocardial infarction, stroke, or mortality was not significantly
different than reported with isolated procedures (115). Khaitan et al. (116) reported 121
consecutive combined operations with a perioperative neurological event rate of 5.8%. Two
patients had transient ischemic attacks. The procedure-related mortality rate was 5.8%.

In summary, the available data suggest that for a majority of patients, combining
CABG with carotid endarterectomy will not result in a substantial improvement in the
incidence of stroke following CABG. No conclusions can be drawn for patients with
severe bilateral carotid artery stenosis or those with symptomatic carotid disease who
require urgent coronary revascularization. Combining CABG and carotid artery surgical 13

procedures under these circumstances may be appropriate in centers with adequate |

expertise in both procedures. j§

VI. CONCLUSION

Perioperative stroke following noncerebrovascular surgery is an infrequent problem. Its
incidence is about 4% for cardiac valvular surgery, 1-2% for CABG, and 0.2-0.5% for
other surgical procedures. The risk of perioperative stroke is increased in older patients,

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

450 DILEGE et al.

patients with prior stroke, and those with multiple risk factors for stroke. With the exception
of cardiac procedures, perioperative strokes occur mainly in the postoperative period. There
does not appear to be a consistent relationship between perioperative stroke and carotid
bruit or asymptomatic carotid stenosis, although only a few patients with very high grade
carotid stenosis have been studied. Prophylactic carotid endarterectomy in preparation for
another procedure should be performed only when the extent of the carotid disease itself is
adequate justification for carotid surgery and when delaying the primary procedure does not
endanger the patient.

REFERENCES

1. Carney WI, Stewart WB, De Pinto DJ, Mucha SC, Roberts B. Carotid bruit as a risk factor in
aortoiliac reconstruction. Surgery 1977; 81:567-570.

2. Kam PCA, Calcroft RM. Peri-operative stroke in general surgical patients. Anaesthesia 1997;
52:879-883.

3. American Heart Association. 2002 Heart and Stroke Statistical Update. Dallas: American
Heart Association, 2001.

4. Rosamond WD, Folsom AR, Chambless LE, Wang CH, McGovern PG, Howard G, Copper
LS, Shahar E. Stroke incidence and survival among middle-aged adults: 9-year follow-up of
the Atherosclerosis Risk in Communities (ARIC) cohort. Stroke 1999; 30:736-743.

5. Hartmann A, Rundek T, Mast H, Paik MC, Boden-Albala B, Mohr JP, Sacco RL. Mortality
and causes of death after first ischemic stroke: The Northern Manhattan Stroke Study.
Neurology 2001; 57:2000-2005.

6. Sarti C, Rastenyte D, Cepaitis Z, Tuomilehto J. International trends in mortality from stroke,
1968 to 1994. Stroke 2000; 31:1588-1601.

7. Mihalka L, Smolanka V, Bulecza B, Mulesa S, Bereczki D. A population study of stroke in
West Ukraine: Incidence, stroke services, and 30-day case fatality. Stroke 2001; 32:2227-2231.

8. Stegmayr B, Vinogradova T, Malyutin S, Peltonen M, Nikitin Y, Asplund K. Widening gap
of stroke east and west. Eight years trends in occurrence and risk factors in Russia and
Sweden. Stroke 2000; 31:2-8.

9. Howard G, Howard WJ, Katholi C, Oli MK, Huston S. Decline in US stroke mortality: An
analysis of temporal patterns by sex, race, and geographic region. Stroke 2001; 32:2213-2220.

10. Mohr JP, Caplan LR, Melski JW, Goldstein RJ, Duncan GW, Kistler JP, Pessin MS, Bleich
HL. The Harvard Cooperative Stroke Registry: A prospective registry. Neurology 1978; 28:
754-762.

11. Petty GW, Brown RD Jr, Whisnant JP, Sicks JD, O'Fallon WM, Wiebers DO. Ischemic
stroke subtypes: A population-based study of incidence and risk factors. Stroke 1999; 30:
2513-2516.

12. Martin PJ, Enevoldson TP, Humphrey PR. Causes of ischemic stroke in the young. Postgrad
Med J 1997; 73:8-16.

13. Gill JS, Shipley MJ, Tsementzis SA, Hornby RS, Gill SK, Hitchcock ER, Beevers DG ^
Alcohol consumption-a risk factor for hemorrhagic and non-hemorrhagic stroke. Am J Med £
1991; 90:489-497. 1

14. Heyman A, Fields WS, Keating RD. Joint study of extracranial arterial occlusion. VI Rapid .2
differences in hospitalized patients with ischemic stroke. JAMA 1972; 222:285-289. .5

15. Sacco RL, Benjamin EJ, Broderick JP, Dyken M, Easton JD, Feinberg WM, Goldstein LB, j
Gorelick PB, Howard G, Kittner SJ, Manolio TA, Whisnant JP, Wolf PA. Risk factors. ^
Panel. Stroke 1997; 28:1507-1517. I

16. Kiel DK, Wolf PA, Couples LA, Beiger AS, Myers RH. Familial aggregation of stroke: The 2
Framingham Study. Stroke 1993; 24:1366-1371. |

17. Ayala C, Greenlund KJ, Croft JB, Keenan NL, Donehoo RS, Giles WH, Kittner SJ, Marks >■

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

STROKE AS A COMPLICATION OF SURGERY 451

JS. Racial/ethnic disparities in mortality by stroke subtype in the United States, 1995-1998.
Am J Epidemiol 2001; 154:1057-1063.

18. Kannel WB, Wolf PA, Verter J, McNamara PM. Epidemiologic assessment of the role of
blood pressure in stroke: The Framingham Study. JAMA 1970; 214:301-310.

19. Garraway WM, Whisnant JP. The changing pattern of hypertension and the declining in-
cidence of stroke 1987; 258:214-217.

20. Perry HM, Smith WM, Mc Donald RH, Black D, Cutler JA, Furberg CD, Greenlick MR,
Kuller LH, Schnaper HW, Schoenberger JA. Morbidity and mortality in the systolic
hypertension in the elderly program (SHEP) pilot study. Stroke 1989; 20:4-13.

21 . MacMahon S, Rodgers A. The epidemiological association between blood pressure and stroke:
Implications for primary and secondary prevention. Hypertens Res 1994; 17(suppl l):23-32.

22. Seshadri S, Wolf PA, Beiser A, Vasan RS, Wilson PW, Kase CS, Kelly-Hayes M, Kannel WB,
D'Agostino RB. Elevated midlife blood pressure increases stroke risk in elderly persons: The
Framingham Study. Arch Intern Med 2001; 161:2343-2350.

23. Frost L, Ingham G, Johnson S, Muller H, Hosted S. Incident stroke after discharge from the
hospital with a diagnosis of atria fibrillation. Am J Med 2000; 108:36-40.

24. Wolf PA, Abbott RD, Kennel WB. Atria fibrillation as an independent risk factor for stroke:
The Framingham Study. Stroke 1991; 22:983-988.

25. American College of Physicians. Guidelines for medical treatment for stroke prevention. Ann
Intern Med 1994; 121:54-55.

26. Di Tullio MR, Sacco RL, Savoia MT, Sciacca RR, Homma S. Gender differences in the risk
of ischemic stroke associated with aortic atheromas. Stroke 2000; 31:2623-2627.

27. Heiss G, Sharrett AR, Barnes R, Chambless LE, Szklo M, Alzola C. Carotid atherosclerosis
measured by B-mode ultrasound in populations: Associations with cardiovascular risk factors
in the ARIC study. Am J Epidemiol 1991; 134:250-256.

28. Bots ML, Elwood PC, Nikitin Y, Salonen JT, Freire de Concalves A, Inzitari D, Sivenius J,
Benetou V, Tuomilehto J, Koudstaal PJ, Grobbee DE. Total and HDL cholesterol and risk of
stroke. EUROSTROKE: A collaborative study among research centers in Europe. J
Epidemiol Commun Health 2002; 56(suppl 1): 19-24.

29. Sacco RL, Benson RT, Kargman DE, Boden-Albala B, Tuck C, Lin IF, Cheng JF, Paik MC,
Shea S, Berglund L. High-density lipoprotein cholesterol and ischemic stroke in the elderly:
The Northern Manhattan Stroke Study. JAMA 2000; 285:2729-2735.

30. Perry IJ, Refsum H, Morris RW, Ebrahim SB, Ueland PM, Shaper AG. Prospective study of
serum total homocysteine concentration and risk of stroke in middle-aged British men. Lancet
1995; 346:1395-1398.

31. Perry IJ. Homocysteine and risk of stroke. J Cardiovasc Risk 1999; 6:235-240.

32. Wolf PA, D'Agostino RB, Kannel WB, Bonita R, Belanger AJ. Cigarette smoking as a risk
factor for stroke: The Framingham Study. JAMA 1988; 259:1025-1029.

33. Gorelick PB. Does alcohol prevent or cause stroke? Cerebrovascular Diseases 1995; 5:379.

34. Kelly MA, Gorelick PB, Mirza D. The role of drugs in the etiology of stroke. Clin Neuro-
pharmacol 1992; 15:249-275.

35. Petitti DB, Sidney S, Bernstein A, Wolf S, Quesenberry C, Ziel HK. Stroke in users of low-
dose oral contraceptives. N Engl J Med 1996; 335:8-15. 1

36. Leblanc ES, Laws A. Benefits and risks of third-generation oral contraceptives. J Gen Intern g
Med 1999; 14:625-632. ,|

37. Simon JA, Hsia J, Cauley JA, Richards C, Harris F, Fong J, Barrett-Connor E, Hulley SB. 2
Postmenopausal hormone therapy and risk of stroke: The Heart and Estrogen-progestin **
Replacement Study (HERS). Circulation 2001; 103:620-622. &

38. Paganini-Hill A, Perez Barreto M. Stroke risk in older men and women: Aspirin, estrogen, jjj
exercise, vitamins, and other factors. J Gend Specif Med 2001; 4:18-28. q

39. Benavente O, Moher D, Pham B. Carotid endarterectomy for asymptomatic carotid artery |
stenosis: A meta-analysis. Br Med J 1998; 317:1477-1480. 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

452 DILEGE et al.

40. Pujia A, Rubba P, Spencer MP. Prevalence of extracranial carotid artery disease detectable by
echo-Doppler in an elderly population. Stroke 1996; 23:818-822.

41 . Fine-Edelstein JS, Wolfe PA, O'Leary DH, Poehlman H, Belanger AJ, Kase CS, D'Agustino
RB. Precursors of extracranial carotid atherosclerosis in the Framingham Study. Neurology
1994;44:1046-1050.

42. Prati P, Vanuzzo D, Casaroli M, Di Chiara A, Di Basi F, Feruglio GA, Toubol PJ. Prevalence
and determinants of carotid atherosclerosis in a general population. Stroke 1992; 23:1705-
1711.

43. Sauve JS, Lauoacis A, Ostbye T, Feagan B, Sackett DL. Does this patient have a clinically
important carotid bruit? JAMA 1993; 270:2843-2845.

44. Dixon S, Pais SO, Raviola C, Gomes A, Machleder HI, Baker JD, Busuttil RW, Barker WF,
Moore WS. Natural history of nonstenotic asymptomatic ulcerative lesions of the carotid
artery. Arch Surg 1982; 117:1493-1498.

45. Moore WS, Boren C, Malone JM, Roon AJ, Eisenberg R, Goldstone J, Mani R. Natural
history of nonstenotic asymptomatic ulcerative lesions of the carotid artery. Arch Surg 1978;
113:1352-1359.

46. Liapis CD, Kakisis JD, Kostakis AG. Carotid stenosis: Factors affecting symptomatology.
Stroke 2001; 32:2782-2786.

47. Inzitari D, Eliasziw M, Gates P, Sharpe BL, Chan RK, Meldrum HE, Barnett HJ. The causes
and risk of stroke in patients with asymptomatic internal-carotid-artery stenosis. North Amer-
ican Symptomatic Carotid Endarterectomy Trial Collaborators. N Engl J Med 2000; 342: 1 693-
1700.

48. Mayo Asymptomatic Carotid Artery Study Group. Results of a randomized controlled trial
of carotid endarterectomy for asymptomatic carotid stenosis. Mayo Clin Proc 1992; 67:513-
518.

49. The CASANOVA Study Group. Carotid surgery versus medical therapy in asymptomatic
carotid stenosis. Stroke 1991; 22:1229-1235.

50. Hobson RW, Weiss DG, Fields WS, Goldstone J, Moore WS, Towne JB, Wright CB. Efficacy
of carotid endarterectomy for asymptomatic carotid stenosis. The Veterans Affairs Coop-
erative Study Group. N Engl J Med 1993; 328:221-227.

51. Asymptomatic Carotid Atherosclerosis Study Group. Carotid endarterectomy for patients
with asymptomatic internal carotid artery stenosis. JAMA 1995; 273:1421-1428.

52. The European Carotid Surgery Trialists Collaborative Group. Risk of stroke in the dis-
tribution of an asymptomatic carotid artery. Lancet 1995; 345:209-212.

53. Dyken ML, Wolf PA, Barnett HJ. Risk factors in stroke: A statement for physicians by the
Subcommittee on Risk Factors and Stroke of the Stroke Council. Stroke 1984; 15:1105-1111.

54. North American Symptomatic Carotid Endarterectomy Trial Collaborators. Beneficial effect
of carotid endarterectomy in symptomatic patients with high-grade carotid stenosis. N Engl J
Med 1991; 325:445^153.

55. Estol CJ, Pessin MS. Anticoagulation: Is there still a role in atherothrombotic stroke. Stroke
1990; 21:820-824.

56. Sacco RL, Wolf PA, Kennel WB, McNamara PM. Survival and recurrence following stroke:

The Framingham study. Stroke 1982; 13:290. 1

57. ECSTCG. MRC European Carotid Surgery Trial: Interim results for symptomatic patients g
with severe (70-99%) or with (0-29%) carotid stenosis. Lancet 1991; 337:1235-1243. J

58. Mayberg MR, Wilson ES, Yatsu F, et al. Carotid endarterectomy and prevention of cerebral 2
ischemia in symptomatic carotid stenosis. JAMA 1991; 266:3289-3294. ^

59. North American Symptomatic Carotid Endarterectomy Trial Collaborators: Benefit of ca- >9
rotid endarterectomy in patients with symptomatic moderate or severe stenosis. N Engl J Med jjj
1998; 339:1415-1425. q

60. Barbut D, Grassineau D, Lis E, Heier L, Hartman GS, Isom OW. Posterior distribution of |
infarcts in strokes related to cardiac operations. Ann Thorac Surg 1998; 65:1656-1659. 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

STROKE AS A COMPLICATION OF SURGERY 453

61. Kelly RE, Kovacks AG. Mechanism of in-hospital cerebral ischemia. Stroke 1986; 17:430-433.

62. Barnes RW, Liebman PR, Marszalek PB, Kirk CL, Goldman MH. The natural history of
aymptomatic carotid disease in patients undergoing cardiovascular surgery. Surgery 1981;
90:1075-1081.

63. Bernhard VM. Discussion of Turnipseed WD, Berkoff HA, Belzer FO, Postoperative stroke
cardiac and peripheral vascular disease. Ann Surg 1980; 192:367.

64. Hart R, Hindman B. Mechanisms of perioperative cerebral infarction. Stroke 1982; 13:766-773.

65. Larsen SF, Zaric D, Boysen G. Postoperative cerebrovascular accidents in general surgery.
Acta Anaesthesiol Scand 1988; 32:698-701.

66. Gerraty RP, Gilford EJ, Gates PC. Watershed cerebral infarction associated with peri-
operative hypotension. Clin Exp Neurol 1993; 30:82-89.

67. Landercasper J, Merz BJ, Cogbill TH, Strutt PJ, Cochrane RH, Olson RA, Hutter RD.
Perioperative stroke in 173 consecutive patients with a past history of stroke. Arch Surg 1990;
125:986-989.

68. Mohr JP, Albers GW, Amarenco P, Babikian VL, Biller J, Brey RL, Coull B, Easton JD,
Gomez CD, Helgason CM, Kase CS, Pullicino PM, Turpie AGG Etiology of stroke. Stroke
1997; 28:1501-1506.

69. Parikh S, Cohen JR. Perioperative stroke after general surgical procedures. N Y State J Med
1993; 93:162-165.

70. Coull BM, Beamer N, de Garmo P, Sexton G, Nordt F, Knox R, Seaman GV. Chronic blood
hyperviscosity in subjects with acute stroke, transient ischemic attacks and risk factors for
stroke. Stroke 1991; 22:162-168.

71. Tanahashi N, Gotoh F, Tomita M, Shinohara T, Terayama Y, Mihara B, Ohta K, Nara M.
Enhanced erythrocyte aggregability in occlusive cerebrovascular disease. Stroke 1989; 20:
1202-1207.

72. Antonova N, Velcheva I. Hemorheological disturbances and characteristic parameters in
patients with cerebrovascular disease. Clin Hemorheol Microcirc 1999; 21:405-408.

73. Beamer N, Coull BM, Sexton G, de Garmo P, Knox R, Seaman G Fibrinogen and the
albumin-globulin ratio in recurrent stroke. Stroke 1993; 24:1133-1139.

74. Gibbs NM, Crawford GP, Michalopoulos N. A comparison of postoperative thrombotic
potential following abdominal aortic surgery, carotid endarterectomy, and femoro-popliteal
bypass. Anaesth Intens Care 1996; 24:11-14.

75. Coull BM, Clark WM. Abnormalities of hemostasis in ischemic stroke. Med Clin North Am
1993; 77:77-94.

76. Roach GW, Kanchuger M, Mangano CM, Newman M, Nussmeier N, Wolman R, Aggarval A,
Marschall K, Graham SH, Ley C. Adverse cerebral outcomes after coronary bypass surgery.
Multicenter study of Perioperative Ischemia Research Group and the Ischemia Research and
Education Foundation Investigators. N Engl J Med 1996; 335:1857-1863.

77. Faggioli GI, Curl GR, Ricotti JJ. The role of carotid screening before coronary artery bypass.
J Vase Surg 1990; 12:724.

78. Pompilio G, Lotto AA, Agrifoglio M, Antona C, Alamanni F, Spirito R, Biglioli P. Non-
embolic predictors of stroke risk in coronary artery bypass patients. World J Surg 1999; 23:

657-663. 1

79. Borger MA, Ivanov J, Weisel RD, Rao V, Peniston CM. Stroke during coronary bypass <S
surgery: Principal role of cerebral macroemboli. Eur J Cardiothorac Surg 2001; 19:627-632. js

80. Engelman DT, Cohn LH, Rizzo RJ. Incidence and predictors of tias and strokes following 2
coronary artery bypass grafting: Report and collective review. Heart Surg Forum 1999; 2:242- **
245. ■§

81. Gardner TJ, Horneffer PF, Manolio TA, Hoff SJ, Pearson TA. Major stroke after coronary ^
artery bypass surgery: Changing magnitude of the problem. J Vase Surg 1986; 3:684-689. q

82. Horneffer PF, Gardner TJ, Manolio TA. The effects of age on outcome after coronary bypass |
surgery. Circulation 1987; 76(suppl V):V6-V12. 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

454 DILEGE et al.

83. Rao V, Christakis GT, Weisel RD, Ivanov J, Peniston CM, Ikonomidis JS, Shirai T. Risk
factors for stroke following coronary artery bypass surgery. J Card Surg 1995; 10(suppl 4):
468-474.

84. Borger MA, Ivanon J, Weisel RD, Peniston CM, Mickleborough LL, Rambaldini G, Cohen
G, Rao V, Findel CM, David TE. Decreasing incidence of stroke during valvular surgery.
Circulation 1998; 98(suppl 19):II137 II143.

85. Salasidis GC, Latter DA, Steinmetz OK, Blair JF, Graham AM. Carotid artery duplex
scanning in preoperative assessment for coronary artery revascularization: The association
between peripheral vascular disease, carotid artery stenosis, and stroke. J Vase Surg 1995; 21:
154-160.

86. Libman RB, Wirkowski E, Neystat M, Barr W, Gelb S, Graver M. Stroke associated with
cardiac surgery. Determinants, timing, and stroke subtypes. Arch Neurol 1997; 54:83-87.

87. Hogue CW Jr, Murphy SF, Schechtman KB, Davila-Roman VG. Risk factors for early or
delayed stroke after cardiac surgery. Circulation 1999; 100:642-647.

88. Ahlgren E, Aren C. Cerebral complications after coronary artery bypass and heart valve
surgery: Risk factors and onset of symptoms. J Cardiothorac Vase Anesth 1998; 12:270-273.

89. Reed GL, Singer DE, Pickard EH, DeSanctis RW. Stroke following coronary artery bypass
surgery: A case control estimate of the risk from carotid bruits. N Engl J Med 1988; 319:1246-
1250.

90. Gonzales-Scarano F, Hurtig HI. Neurologic complications of coronary artery bypass graf-
ting: Case control study. Neurology 1981; 31:1032-1035.

91. Schwartz LB, Bridgman AH, Kieffer RW, Wilcox RA, McCann RL, Tawil MP, Scott SM.
Asymptomatic carotid artery stenosis and stroke in patients undergoing cardiopulmonary
bypass. J Vase Surg 1995; 21:146-153.

92. Skotnicki SH, Schulte BP, Leyten QH, Tacke TJ, Arntz IE. Asymptomatic carotid bruit in
patients who undergo coronary artery surgery. Eur J Cardiothorac Surg 1987; 1:11-15.

93. Ivey TD, Strandness E, Williams DB, Langlois Y, Misbach GA, Kruse AP. Management of
patients with carotid bruit undergoing cardiopulmonary bypass. J Thorac Cardiovasc Surg
1984;87:183-189.

94. Hart RG, Easton JD. Management of cervical bruits and carotid stenosis in preoperative
patients. Stroke 1982; 14:290-297.

95. Hennerici M, Aulich A, Sandmann W, Freund HJ. Incidence of asymptomatic extracranial
arterial disease. Stroke 1981; 12:750-758.

96. Barnes RW, Marszalek PB. Asymptomatic carotid disease in the cardiovascular surgical
patients: Is prophylactic endarterectomy necessary? Stroke 1981; 12:497-500.

97. Plate G, Hollier LH, O'Brien PC, Pairolero PC, Cherry KJ. Late cerebrovascular accidents
after repair of abdominal aortic aneurysms. Acta Chir Scand 1988; 154:25-29.

98. Harris EJ Ir, Moneta GL, Yeager RA, Taylor LM Jr, Porter JM. Neurologic deficits fol-
lowing noncarotid vascular surgery. Am J Surg 1992; 163:537-540.

99. Van den Brande P, Vanhandenhove I. The vascular surgical patient: problems in the man-
agement of coronary and cerebrovascular risk. Acta Chir Belg 1988; 88:359-362.

100. Limburg M, Wijdicks EF, Li H. Ischemic stroke after surgical procedures: Clinical features,
neuroimaging, and risk factors. Neurology 1998; 50:895-901. j>

101 . Chambers BR, Norris JW. Outcome in patients with asymptomatic neck bruits. N Engl J Med g
1986; 315:860-865. ,|

102. Bogousslavsky J, Despland PA, Regli F. Asymptomatic tight stenosis of the internal carotid £
artery: Long-term prognosis. Neurology 1986; 36:861-863. ^

103. Hennerici M, Hulshhomer H-B, Hefter H, et al. Natural history of asymptomatic extracranial «
arterial disease: Results of a long term prospective study. Brain 1987; 110:777-791. jjj

104. Moneta GL, Taylor DC, Nicholls SC, et al. Operative versus nonoperative management of q
asymptomatic high-grade internal carotid artery stenosis: Improved results with endarter- |
ectomy. Stroke 1987; 18:1005-1010. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

STROKE AS A COMPLICATION OF SURGERY 455

105. Caracci BF, Zukowski AJ, Hurley JJ, et al. Asymptomatic severe carotid stenosis. J Vase Surg
1989; 9:361-366.

106. Warlow C. Endarterectomy for asymptomatic carotid stenosis? Lancet 1995; 345:1254-1256.

107. O'Donnell TF Jr, Callow AD, Willet C, Payne D, Cleveland RJ. The impact of coronary
artery disease on carotid endarterectomy. Ann Surg 1983; 198:705-712.

108. Hertzer NR, Loop FD, Beven EG, et al. Coronary angiography in 506 patients with extra-
cranial cerebrovascular disease. Arch Intern Med 1985; 145:849-852.

109. Faggioli GI, Curl GR, Ricotti JJ. The role of carotid screening before coronary artery bypass.
J Vase Surg 1990; 12:724.

1 10. Brow TD, Kakkar VV, Pepper JR, Das SK. Toward a rational management of concomitant
carotid and coronary artery disease. J Cardiovasc Surg 1999; 40:837-844.

111. Terramani TT, Rowe VL, Hood DB, Eton D, Nuno IN, Yu H, Yellin AE, Starnes VA, Weaver
FA. Combined carotid endarterectomy and coronary artery bypass grafting in asymptomatic
carotid artery stenosis. Am Surg 1998; 64:993-997.

1 12. Hertzer NR, Loop FD, Beven EG, O'Hara PJ, Krajewski LP. Surgical staging for simultaneous
coronary and carotid disease: A study including prospective randomization. J Vase Surg 1989;
9:455-463.

113. Rizzo RJ, Whittemore AD, Couper GS, Donaldson MC, Aranki SF, Collins JJ Jr, Mannick
J A, Cohn LH. Combined carotid and coronary revascularization: The preferred approach to
the severe vasculopath. Ann Thorac Surg 1992; 54:1099-2108.

1 14. Brener BJ, Brief DK, Alpert J, Goldenkranz RJ, Parsonnet V. The risk of stroke in patients
with asymptomatic carotid stenosis undergoing cardiac surgery: A follow-up study. J Vase
Surg 1987; 5:269-279.

115. Evagelopoulos N, Trenz MT, Beckmann A, Krian A. Simultaneous carotid endarterectomy
and coronary artery bypass grafting in 313 patients. Cardiovasc Surg 2000; 8(1):3 1 — 40.

116. Khaitan L, Sutter FP, Goldman SM, Chamogeorgakis T, Wertan MA, Priest BP, Whitlark
JD. Simultaneous carotid endarterectomy and coronary revascularization. Ann Thorac Surg
2000; 69:421-424.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

26

Complications of Repair of the Supra-Aortic
Trunks and the Vertebral Arteries

Jeffery B. Dattilo

Vanderbilt University Medical Center, Nashville, Tennessee, U.S.A.

Richard P. Cambria

Massachusetts General Hospital and Harvard Medical School, Boston,
Massachusetts, U.S.A.

I. INTRODUCTION

The supra-aortic trunks (SATs) are defined as those arteries that originate from the aortic
arch and course through the mediastinum, terminating just proximal to the carotid bifur-
cation or the origin of the vertebral arteries. Atherosclerotic occlusive disease of the SATs
or vertebral arteries may lead to a variety of cerebral or peripheral symptoms. Patients
may demonstrate symptoms of either carotid or posterior circulation ischemia or upper
extremity vascular insufficiency. Hypoperfusion and microembolic phenomena are mecha-
nisms by which symptoms are generated. Such symptoms occur across a spectrum of pa-
tient subsets, including the arteritities, among which Takayasu arterititis is more frequent
in the younger population.

The anatomical locations of these lesions vary depending upon the vessel involved.
The vast majority of SATs and vertebral lesions are diagnosed incidentally in conjunction
with a workup for more distal circulatory pathology, specifically the internal carotid. "g

Consequently, little is known about the natural history of morphology of these lesions. &

Stenosis of these arteries usually occurs at the origin of the vessel and often involves more a

than one artery. The Joint Study of Extracranial Arterial Occlusion reported the specific c

anatomical locations and incidence of lesions in patients with symptoms (1) (see Table 1). <j

The frequency of SAT lesions is relatively lower than that of carotid bifurcation >9

lesions. Further, patients with supra-aortic trunk lesions are generally asymptomatic at the 4j

time of presentation. Even if the patient has symptoms of arm claudication or vertebro- 2

basilar ischemia, nonoperative approaches are often chosen. Indications for open surgical |

repair of lesions of the SAT include flow-limiting stenosis greater than 70% with mor- @

457 I

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

458

DATTILO and CAMBRIA

Table 1 Number of Lesions of SATs from \t
Studied

Patients

Location

Number of lesions

Left subclavian

Innominate or right subclavian

Right common carotid

Left common carotid

Right vertebral

Left vertebral

129

73
27
28
34

27

Source: Ref. 1.

phology of the plaque consisting of ulceration or surface irregularities and symptoms
relating to the lesion such as distal embolization, preocclusive lesions of greater than 90%,
or unique circumstances such as a subclavian artery that feeds an internal mammary graft.
A relative indication for reconstruction of the vertebral artery is to treat vertebrobasilar
ischemia. Of course, it is often difficult for the clinician to determine that the symptoms
are secondary to a radiographically significant lesion of the vertebral artery. Further, radio-
graphic subclavian steal is relatively common, but clinical sequelae are distinctly uncommon
in these patients.

A major consideration in planning arterial reconstruction is whether to perform an
anatomical reconstruction or extra-anatomical bypass. Generally, anatomical reconstruc-
tions are reserved for the younger, lower-risk patient who has an innominate artery lesion
or multiple lesions. Further, the anatomical approach is prudent in a patient undergoing
coronary artery bypass grafting where use of the mammary artery distal to the SAT lesion
is planned. Extra-anatomical repairs are often reserved for the elderly patient in whom a
thoracotomy or sternotomy would present too high a risk or for those who have had pre-
vious transsternal procedures. This chapter discusses the potential complications encoun-
tered perioperatively as well as issues of graft patency as they pertain to open surgical
reconstruction.

II. EXTRA-ANATOMICAL REPAIRS OF THE SUPRA-AORTIC
TRUNKS

Extra-anatomical repairs circumvent the normal human arterial anatomy. There are many
described variations for bypassing these lesions (Table 2). The carotid-to-subclavian by-

Table 2 Extra- Anatomic
Reconstructions

Carotid to subclavian
Subclavian to carotid
Carotid to carotid
Axilloaxillary

Carotid subclavian transposition
Subclavian carotid transposition
Subclavian to subclavian
Femoral to subclavian

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

REPAIR OF ARTERIES 459

pass was originally described by Lyons in 1957 and popularized in the 1970s, and it became
the standard of care for proximal subclavian lesions (2). The cervical approach causes less
morbidity than an anatomical reconstruction. In modern surgical series, extra-anatomical
reconstructions are generally reserved for high-risk patients with formidable cardiopul-
monary risk factors that prevent transthoracic repair. Flow-dynamic advantages to these
bypass grafts are short conduits in relatively large-bore, high-flow systems; historically,
they perform well.

A. Technical Considerations

Technical complications of exposure of the supraclavicular carotid and subclavian arteries
depend on a firm understanding of the anatomical relationships of the structures encoun-
tered during the dissection. Extrathoracic exposure of the subclavian artery starts with a
supraclavicular incision centered over the clavicular head of the sternocleidomastoid
muscle, which is then divided with impunity. Next, the prescalene fat pad is divided, with
care taken to ligate all lymphatic channels. Dissection in the region of the thoracic duct on
the left is to be avoided. The phrenic nerve, which typically lies on the anterior portion of
the scalene fascia, must be identified and preserved, usually by medial retraction. Next, the
anterior scalene muscle is divided, exposing the subclavian artery, which lies beneath. An
understanding of the relationship of the brachial plexus as it exits laterally is also bene-
ficial. The subclavian artery is then dissected proximally and distally, encountering the
thyrocervical trunk, internal mammary artery, and vertebral artery medially and superi-
orly. The carotid artery can be mobilized from the carotid sheath using the same incision.
Care must be taken in mobilizing the proximal carotid artery because the cervical sym-
pathetic chain lies posterior to the verebral artery.

Most subclavian carotid or carotid subclavian bypasses are performed end to side,
complemented by in-line ligation if there is a truly ulcerative lesions proximally, because
of the risk of distal embolization. Further, early in the 1970s, the saphenous vein was used
as the preferred conduit. However, because of problems such as mismatches in caliber,
kinking, or compression of the vein, prosthetic materials were explored. Comparative
studies have demonstrated impressive patency of prosthetics in this position, mainly
because it is a relatively short and high-flow graft. It is currently used nearly exclusively
in this position.

B. Perioperative Complications

1 . Mortality

Mortality from extra-anatomic cervical repairs is low, ranging from 0-3% (3-6). The most

common causes of perioperative death in these high-risk patients are myocardial infarction ■g

or stroke. &

Since myocardial infarction is an important cause of death, consideration of cardiac a

risk stratification is appropriate, as in any vascular operation. Considerable controversy c

remains regarding what testing modalities are optimal for assessing coronary risk and <j

whether such tests should be applied routinely. Perhaps the most prudent method is a >9

selective approach to preoperative testing with dipyridamole-thallium imaging in patients 4j

undergoing vascular surgery, based on the presence of certain clinical markers of coronary 2

artery disease such as overall functional status, age of 70 or greater, diabetes, Q wave on |

electrocardiography, history of angina, and ventricular ectopy — all of which have been @

shown to be independent predictors of perioperative ischemic events (7,8). %

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

460 DATTILO and CAMBRIA

2. Stroke

The incidence of stroke has been reported to range from to 3.3% (3,4). The mechanism
of stroke is thromboembolism or ischemia. The role of intraoperative maneuvers to reduce
the risk of stroke should be considered. Electroencephalographic monitoring, especially
during carotid artery bypasses, should be considered to assist the clinician in determining
candidates for intraoperative shunting. Basic vascular surgical tasks such as thorough
antegrade and retrograde arterial flushing prior to completion of the anastomosis cannot
be overemphasized.

3. Nerve Injury

Nerve injury is another complication related to extra-anatomical cervical repairs. The vagus
nerve and its recurrent nerve branch are of particular risk of injury upon manipulation of the
carotid sheath or the phrenic nerve during the anterior scalene maneuvers. Palsy of the
recurrent and phrenic nerves ranges from to 3% (4,5). Due to its relative clinical insig-
nificance, it is probable that temporary palsy of the phrenic nerve is higher in incidence
than has been reported. To avoid injury to the nerves involves avoidance of manual mani-
pulation with instruments or vessel loops. Identification of these nerves without manipu-
lation is often the best policy in carrying out the dissection or in the application of retractors
for exposure.

Injury to the cords of the brachial plexus should also be considered in operating near
the anterior scalene muscle for exposure of the subclavian artery. The reported risk of
injury is typically quite low. It is only when the dissection is carried out too laterally or
superiorly that the plexus is in danger of injury.

4. Lymphatic Injury

Lymphatic injury producing lymphocele or lymph leak ranges from to 7% (3-1 1). This is
the most common wound complication in operations conducted in the supraclavicular
space. The mechanism is related unrecognized injury to the thoracic duct or the multitude
of lymphatics associated with the scalene fad pad. Attention to detail with deliberate
ligation of all lymphatic tissue during the original dissection is key to the prevention of
these often troublesome injuries. We favor division under direct vision and ligation of the
entirety of the scalene fat pad to minimize the risk of lymphatic injury. Reexploration with
ligation of the source of the lymph leak may occasionally be required. Methylene blue can
be used to detail the anatomy of the lymph leak for precise ligation.

5. Tunneling Complications

Certain extra-anatomical bypasses require tunneling of the graft in both subcutaneous and
occasionally retropharyngeal tissue planes. Tunneling for an axillary-axillary bypass re-
quires a subcutaneous tunnel. Routing of a carotid-carotid bypass requires a tunnel under
the strap muscles. Crossing of the midline (i.e., trachea and sternum) can result in skin 1

erosion and infectious complications (11). Additionally, these routes complicate subse- |

quent surgical therapies such as tracheostomy, median sternotomy, and subsequent arch s

reconstructions. 2

6. Patency s
Long-term patency varies with the extra-anatomical procedure undertaken as well as the 41
chosen conduit (Table 3). Carotid subclavian bypass is well tolerated and has the best 2
results of the extrathoracic approaches. Recently, AbuRahma from West Virginia |
reported a primary patency rate of 96 and 92% at 5 and 10 years in 51 patients (3). This ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

REPAIR OF ARTERIES

461

Table 3 Five- Year Patency of
Carotid Subclavian Bypass

Transposition

100%

PTFE

95%

Dacron

84%

Saphenous vein

65%

Source: Ref. 4.

study used only PTFE for conduit. Five-year patency rates for carotid subclavian bypass
were 95% for bypass grafts using PTFE in the review by UCLA (4). Interestingly, Law
and colleagues further reported that Dacron had a patency rate of 84% and saphenous
vein graft of 65% in the carotid subclavian bypass position. Patency rates for 124 patients
reviewed from the University of Arkansas was 95%> at both 5 and 10 years, with 35%> of
grafts being constructed from PTFE and the remainder composed of Dacron (5). Finally,
Perler reviewed the Johns Hopkins experience and found the 5-year patency of PTFE-
constructed grafts to be 92 and the 8-year patency at 83% (6) (see Table 4).

Transposition appears to be superior in terms of patency even to bypass. The benefits
seem to include both reduced morbidity and greater patency. Subclavian transposition in
178 procedures was reported by Edwards and coworkers (12). The mortality rate asso-
ciated with the isolated subclavian-carotid transposition was 1.1%, and all but one artery

Table 4 Death, Stroke, and Patency Rates of Extra-anatomic Bypass

Number of

Mortality

Stroke

5-year

Series, year (Ref.)

patients

(percent)

(percent)

patency

Abu Rahma, 2000 (3)

51

96

Carotid-subclavian bypass

with PTFE

Law et al., 1995 (4)

60

1.7

3.3

87.5

Carotid-subclavian and

subclavian-carotid

transposition and with

differing conduits

Vitti et al., 1994 (5)

124

0.8

95

Carotid-subclavian with

Dacron or PTFE

Perler and Williams, 1990 (6)

31

3.2

3.2

92

Carotid-subclavian with

PTFE, Dacron vein,

transpostion

Edwards et al., 1994 (12)

178

1.1

1

99 a

Subclavian-carotid

transposition

Mingoli et al., 1999 (11)

61

1.6

86.5

Axilloaxillary bypass

i

2

I

■c

1 46 months.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

462 DATTILO and CAMBRIA

remained patent after a mean follow-up of 46 months. Schardey and colleagues reviewed
108 patients who underwent subclavian artery transposition; they reported no strokes and
patency of 100% at 70 months (13). Van der Vliet found patency differences in trans-
position compared to bypass to be 100% for transposition at 5 and 10 years compared to
bypass patency of 62 and 52%, respectively (9). The risk of perioperative stroke maybe less
with transposition than with bypass. Indeed, Kretschmer published a 5.3% stroke rate
with bypass and a rate of zero with transposition (10). Disadvantages to transposition
include preservation of the vertebral artery because of its medial location. Additionally,
bypass grafting can preserve the internal mammary artery for future potential coronary
artery bypass surgery.

Other extra-anatomical surgical strategies have less favorable patency rates. Indeed,
axilloaxillary bypas has poor long-term results. Criado reported axilloaxillary bypass to
have the worst patency of any extra-anatomical bypass (14). Mingoli and colleagues
reported somewhat more favorable results, with 5-year patency rates nearing 86% (11).
Probably the worst patency is related to the femoral-to-axillary bypass (15). Clearly, the
retrograde flow and length of the bypass contribute to these poor results.

III. ASCENDING AORTIC OPERATIONS— INNOMINATE ARTERY
ATHEROSCLEROSIS

Innominate artery atherosclerosis is uncommon. In a comprehensive review of nearly 2000
operations on the aortic arch branches, Wylie and associates found that lesions of the
innominate artery occured in only 7.5% of the reported cases (16). Most patients undergoing
innominate artery reconstruction, however, have multiple supra-aortic lesions. Indeed, it
has been reported that 61-84% of patients have multiple arch lesions (1,17,18). Trans-
thoracic reconstructions are the preferred treatment in Takayasu's arteritis, radiation arte-
ritis, recurrent disease, and multiple-level disease. These axial reconstructions offer better
inflow for multiple reconstruction if needed for multiple lesions.

A. Technical Considerations

Axial reconstruction may be accomplished by bypass grafting directly from the ascending
aorta performing an endarterectomy. Endarterectomy is usually done at the aortic arch
origin of the innominate artery and carried on to the right subclavian and/or the right
common carotid artery. Endarterectomy is limited by the safe placement of the proximal
clamp onto the often diseased aorta and the proximity to the left common carotid artery.
Caution must be exercised in performing endarterectomy because of its technical hazards.
Most vascular surgeons have abandoned endarterectomy in favor of bypass grafting.

Transthoracic bypass of the innominate artery or its branches begins with a median
sternotomy and extension of the incision into the right subclavian fossa, depending on the
extent of disease. After isolation of the proximal aorta and the distal target artery, a partial

occlusion clamp is placed as proximal on the ascending aorta as possible. In our evaluation c

scheme of clamping strategies, we evaluate the ascending aorta for calcium and thrombus <j

with either transesophageal echocardiography or spiral computed tomography. Dacron >9

or externally supported PTFE is used if the bypass is a single one. Bifurcated Dacron grafts 4j

are used for multiple bypasses. The proximal anastomosis should be positioned on the right 2

side of the very proximal intrapericardial ascending aorta to prevent compromise of the |

graft by closure of the sternum. Further, the graft should not be overstretched, as this may @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

REPAIR OF ARTERIES 463

cause fibrosis and later kinking of the graft with movement of the head and neck. Distal
anastomoses are created in one of two ways. If the disease proximally is thought to create
atheroemboli, the surgeon may consider transecting the diseased artery at it origin and
creating an end-to-end anastomosis with the graft. If thromboembolism is not a threat
proximally, an end-to-side anastomosis can be created. Additionally, care and attention
must be given to protecting the right vagus nerve during these manipulations.

Care must be exercised in tunneling the graft. The graft should not obstruct the
innominate artery or internal jugular vein. In most circumstances, the graft should not be
tunneled underneath the veins. Placing the graft ventral to the veins will avoid venous com-
pression or obstruction. Another described technique is to ligate and divide the brachioce-
phalic vein. Although permanent arm swelling has been described, this is usually a transient
problem that subsides over time.

Concomitant coronary artery disease occurs frequently in this patient population
(18,19). It is often prudent to evaluate the patient for coronary artery disease angiograph-
ically prior to performing a median sternotomy. Recently, Takach and colleagues from the
Texas Heart Institute reported a perioperative mortality of only 3.2% when direct revas-
cularization of the SATs was performed along with coronary artery bypass grafting (20).
Our practice is to routinely do at least a dipyridamole-thalliwm study if we do a sternotomy
and then, if positive, to consider coronary angiography.

B. Mortality

Interest in ascending axial reconstructions of the innominate artery decreased in the 1970s
because of the high mortality rate following those procedures (21). However, since the
inception of the modern era of critical care and refined surgical techniques, the mortality
rate for transthoracic procedures has diminished substantially. Mortality rate in a sam-
pling of large contemporary series ranges from to 14.7% (17-19,22-25) (see Table 5).

C. Stroke

Perioperative stroke ranges from to 8% in the latest series (see Table 5). Frequently,
these patients will have multiple other lesions, including carotid bifurcation lesions. We
commonly use electroencephalographic control to detect global ischemia. It is often tech-
nically possible to shunt through the graft. In our experience with routine electroenceph-
alographic monitoring, ischemic changes with clamping are seldom noted. This can be best
explained by the external-to-internal carotid artery pathway above the clamps.

Table 5 Perioperative Stroke and Mortality Rates of Transthoracic Repair of Supra-aortic Trunks

Author, year (Ref.) Number of patients Stroke (percent) Mortality (percent)

Berguer, 1998 (18) 100 8 8

Kieffer, 1995 (19) 148 5.4 5.4

Cherry, 1989 (17) 26 3.8

Brewster, 1985 (25) 29 6.9 3.4

Zelenock, 1985 (22) 17 5.9

Crawford, 1983 (24) 43 5.5 4.7

Vogt, 1982 (23) 34 14.7

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

464 DATTILO and CAMBRIA

D. Venous Problems

As mentioned previously, careful placement of the graft in relation to the brachiocephalic
vein is critical in avoiding compression of the vein. Our stance is to place the graft over the
vein. We have had no venous compression or thrombosis related to this technique.

E. Nerve Injury

The surgeon should have an understanding of the location of the vagus and recurrent
laryngeal nerves as they course through the mediastinum. Additionally, if more distal
exposure of the subclavian artery is necessary, the phrenic nerve must be identified and
preserved.

F. Sternotomy Complications

Sternotomy issues that should be considered include graft kinking after the bypass, kink-
ing of venous structures after the closure, and the periodically encountered sternal wound
infections. The surgeon should be keenly aware of the right of domain of the mediastinum
in constructing the bypass and prior to sternotomy closure. As discussed previously,
reducing the amount of tissue in the mediastinum is crucial in preventing overcrowding in
the mediastinum. Perhaps resecting the proximal diseased segment of the innominate
artery would be helpful to get the graft to lie properly. Again, we routinely do end-to-end
anastomosis as opposed to end-to-side, partially to help avoid these and other problems.

G. Patency

Direct transthoracic reconstruction of the supra-aortic trucks results in excellent patency
rates (see Table 5). The best assessment of the modern surgical experience and long-term
results of transthoracic anatomical reconstruction have been reported by the groups in
Detroit and Paris (18,19). Primary patency in the series by Kieffer and collegues was 98.4
at 5 years and 96.3% at 10 years. Berguer and coworkers reported 94% and 88% patency
rates at 5 and 10 years, respectively. In comparing endarterectomy to direct reconstruc-
tion, no difference in patency was found (17,25,26). This operation, however, is performed
less often than anatomical reconstruction.

IV. DIRECT VERTEBRAL ARTERY RECONSTRUCTION

Surgical treatment of vertebrobasilar insufficiency continues to generate a great deal of
debate in the literature. Most agree, however, that if clinical manifestations and definitive
objective imaging studies both confirm the diagnosis, surgical options should be consid-
ered. Depending on the level of disease, three open surgical options have been described:
transposition of the vertebral to the common carotid artery, vertebral artery endarter- -o

ectomy, or — for cervical segment disease — bony decompression. While large surgical series |

such those of Berguer et al. are noted, primary vertebral artery reconstruction is seldom |

required (27,28). Furthermore, while most respect Ramon Berguer's work, the fact is that .g>

direct surgical reconstruction is (in the hands of most) confined to the vertebral artery <i

origin. s

I

A. Technical Considerations <3

The vertebral artery is approached through a standard supraclavicular incision. The §

sternocleidomastoid muscle laterally is divided. The anterior scalene muscle is likewise @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

REPAIR OF ARTERIES 465

divided. Care is taken to avoid injury to the phrenic nerve, which is retracted laterally. If
the thoracic duct is identified it is also ligated. The common carotid artery is mobilized
from the carotid sheath. The origin of the vertebral artery is identified and ligated. The
vertebral artery is then carefully mobilized and reimplanted into either the CCA or some-
times described a nondiseased external carotid artery. Other options include an interpo-
sition vein graft from the subclavian artery to the vertebral artery. Of note is the rather
fragile nature to the vertebral artery. The wall of the vertebral artery is on the order of
one-third the thickness of the saphenous vein; therefore clamping must be done carefully
and untoward stretching avoided. Technical considerations of bypass or reconstruction of
the distal vertebral artery are beyond the scope of this chapter.

B. Mortality/Stroke

The perioperative mortality rate of vertebral artery reconstruction is low, at around 0.6-
3.2% (27). More recent reports claim lower mortality rates, near 0.6%, and state that this
is due to modern anesthetic regimens and improved patient selection (28). The perioper-
ative stroke rate of vertebral artery reconstruction ranges from 1.9 to 4.1% (27,28).

C. Other Complications

The left vertebral artery origin lies in close proximity to the thoracic duct. Often, if iden-
tified, the duct can be ligated to prevent the occurrence of a lymphocele. Local compli-
cations such as Horner's syndrome are due to manipulation or injury of the lower cervical
sympathetics and are usually transient.

D. Patency

Patency rates of vertebral artery transposition are usually quite satisfactory, ranging from
90 to 97% at 5 years (27,28). Occlusion of these reconstructions is usually exclusively in
patients in whom vertebral-to-carotid transposition is not possible and interposition vein
graft is necessary.

REFERENCES

1. Fields WS, Lemak NA. Joint study of extracranial arterial occlusion. JAMA 1972; 222:1139.

2. Lyons C, Galbraith G. Surgical treatment of atherosclerotic occlusion of the internal carotid
artery. Ann Surg 1957; 146:484-487.

3. AbuRahma AF, Robinson PA, Jennings TG. Carotid subclavian bypass grafting with PTFE
grafts for symptomatic subclavian artery stenosis or occlusion: A 20 year experience. J Vase
Surg 2000; 32:411-419.

4. Law MM, Colburn MD, Moore WS, Quinones-Baldrich WJ, Machleder HI, Gelabert HA. -o
Carotid subclavian bypass for brachiocephalic occlusive disease: Choice of conduit and long §
term follow-up. Stroke 1995; 26:1565-1571. 8

5. Vitti MJ, Thompson BW, Read RC, et al. Carotid-subclavian bypass: A twenty-two year expe- •§,
rience. J Vase Surg 1994; 20:41 1^18. |

6. Perler BA, Williams GM. Carotid-subclavian bypass-A decade of experience. J Vase Surg 1990; u
12:716. "5

7. Eagle KA, Coley CM, Newell JB, et al. Combining clinical and thallium data optimizes S
preoperative assessment of cardiac risk before major vascular surgery. Ann Intern Med 1989; 2
110:859-866. 1

8. Cambria RP, Brewster DC, Abbott WM, L'ltalien GJ, Megerman JJ, LaMuraglia GM, |

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

466 DATTILO and CAMBRIA

Moncure AC, Zelt DT, Eagle K. The impact of selective use of dipyridamole-thallium scans
and surgical factors on the current morbidity of aortic surgery. J Vase Surg 1992; 15(1):43— 50.
9. Van der Vliet JA, Palamba HW, Scharn DM, et al. Arterial reconstruction for subclavian
obstructive disease: A comparison of extrathoracic procedures. Eur J Vase Endovasc Surg 1995;
9:454-458.

10. Kretschmer G, Teleky B, Marosi L, et al. Obliterations of the proximal subclavian artery: To
bypass or to anastomose? J Cardiol Surg 1991; 32:334-339.

11. Mingoli A, Sapienza P, Feldhaus RJ, Bartoli S, Palombi M, de Marzo L, Cavallaro A. Long-
term results and outcomes of crossover axillo-axillary bypass grafting: A 24 year experience.
J Vase Surg 1999; 29:894-901.

12. Edwards WH, Tapper SS, Edwards WH Sr, Mulherin JL, Martin RS, Jenkins JM. Subclavian
revascularization: A quarter-century experience. Ann Surg 1994; 219:673.

13. Schardey HM, Meyer G, Rau HG, et al. Subclavian caroted transposition: An analysis of a
clinical series and a review of the literature. Eur J Vase Endovasc Surg 1996; 12:43 1 — 436.

14. Criado FJ. Extrathoracic management of aortic arch syndrome. Br J Surg 1982; 69(suppl):45-
51.

15. Sproul G. Femoral-axillary bypass for cerebral vascular insufficiency. Arch Surg 1971; 103:746-
747.

16. Wylie EJ, Effeney DJ. Surgery of the aortic arch branches and vertebral arteries. Surg Clin
North Am 1979; 59:669-680.

17. Cherry KJ Jr, McCullough JL, Hallet JW Jr, Pairolero P. Technical principles of direct in-
nominate artery revascularization: A comparison of endarterectomy and bypass grafts. J Vase
Surg 1989; 9:718-724.

18. Berguer R, Flynn LM, Kline RA, Caplan L. Transthoracic repair of innominate and common
carotid artery disease: immediate and long-term outcome for 100 consecutive surgical recon-
structions. J Vase Surg 1998; 27(1):34-41.

19. Kieffer E, Sabatier J, Koskas F, Bahnini A. Atherosclerotic innominate artery occlusive disease:
Early and long term results of surgical reconstruction. J Vase Surg 1995; 21:326-337.

20. Takach TJ, Reul GJ Jr, Cooley DA, et al. Concomitant occlusive disease of the coronary
arteries and great vessels. Ann Thorac Surg 1998; 65:79-84.

21. Crawford ES, DeBakey ME, Morris GC, Howell JF. Surgical treatment of occlusion of the
innominate, common carotid, and subclavian arteries: A 10-year experience. Surgery 1969; 65:
17-31.

22. Zelenock GB, Cronenwett JL, Graham LM, et al. Brachiocephalic arterial occlusions and
stenosis. Arch Surg 1985; 120:370-376.

23. Vogt DP, Hertzer NR, O'Hara PJ, Beven EG. Brachiocephalic arterial reconstruction. Ann
Surg 1982; 196:541-552.

24. Crawford ES, Stowe CL, Powers RW Jr. Occlusion of the innominate, common carotid and
subclavian arteries: Long-term results of surgical treatment. Surgery 1983; 94:781.

25. Brewster DC, Moncure AC, Darling RC, et al. Innominate artery lesions: Problems en-
countered and lessons learned. J Vase Surg 1985; 2:99-112.

26. Reul GJ, Jacobs MJHM, Gregoric ID, et al. Innominate artery occlusive disease: Surgical
approach and long-term results. J Vase Surg 1991; 14:405^112. j>

27. Berguer R, Morasch MD, Kline RA. A review of 100 consecutive reconstructions of the distal g
vertebral artery for embolic and haemodynamic disease. J Vase Surg 1998; 27(5):852-859. js

28. Berguer R, Flynn LM, Kline RA, Carplan L. Surgical reconstruction of the extracranial £
vertebral artery: Management and outcome. J Vase Surg 2000; 31(1 Pt 1):9-18. ^

I

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

27

Prevention of Transient Ischemic Attacks and Acute

Strokes After Carotid Endarterectomy : A Critique of

Techniques for Cerebrovascular Protection During

Carotid Endarterectomy

William H. Baker and Maureen Sheehan

Loyola University Medical Center, Maywood, Illinois, U.S.A.

The purpose of carotid endarterectomy is the prevention of stroke. The role of the opera-
tion has been well established by prospective, randomized studies in both symptomatic
and asymptomatic patients (1-3). Newer antiplatelet agents, while effective, have not
supplanted operation. Carotid dilatation and stenting is currently undergoing scrutiny in a
prospective study sponsored by the National Institutes of Health.

If carotid endarterectomy is to remain the "gold standard" of treatment, stroke and
death rates associated with operation must be kept to a minimum. Many Centers of Ex-
cellence report stroke and mortality rates of 1-3% following carotid endarterectomy.
Older statistics of Medicare patients gathered by the Rand Corporation suggest an almost
10% combined stroke and mortality rate (4). The American Heart Association suggests
that an operation-related stroke and death rate of 2-5% is acceptable in symptomatic
patients (5). It is this author's opinion that the stroke and death rate ought to be below 3%
in all circumstances. Thus it behooves all vascular surgeons to use the safest methods for
cerebrovascular protection during operation. Although perioperative stroke rates may
also be related to such factors as patient selection, anesthesia, and avoidance of cardiac e

complications, this chapter concentrates on those operative techniques that may correlate e

with intraoperative stroke. ■§,

■c

I. TEMPORARY INDWELLING SHUNT B "

I
During the performance of a carotid endarterectomy, vascular clamps by necessity are Q

applied to the common, external, and internal carotid arteries, resulting in cessation of |

flow in the ipsilateral internal carotid artery and a reduction of flow in the ipsilateral ©

467 I

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

468 BAKER and SHEEHAN

cerebral hemisphere. If the collateral cerebral circulation is adequate, no neurological
sequelae result from this temporary reduction of flow. If the cerebral collateral circulation
is inadequate, the patient is at risk for intraoperative stroke. In an effort to avoid stroke in
patients in whom the cerebral collateral flow is inadequate, surgeons have used a tem-
porary indwelling shunt (6). A variety of plastic tubes can be used to shunt blood from the
common carotid artery to the distal internal carotid artery during the performance of
endarterectomy. Whereas this technique does increase ipsilateral hemispheric flow during
operation, the shunt itself may "snowplow" debris from the atheromatous intima into the
ipsilateral cerebral circulation, causing hemispheric emboli. If air is inadvertently intro-
duced during insertion of a shunt, air emboli may result. Finally, the shunt itself is an
encumbrance during the performance of the endarterectomy. It can be easier to perform
the operation without the presence of a plastic tube in the middle of the artery. Although
the shunt does not preclude the performance of an adequate operation and, in fact, in
some operations is easy to work around, it is simpler not to have it in the way. Further-
more, in some patients with a long tail of atheroma in whom the carotid bifurcation is
situated quite high in the neck underneath the angle of the mandible, the operation is
challenging even without the shunt in place. Use of a temporary indwelling shunt in this
situation is almost impossible. To protect the brain during the performance of carotid end-
arterectomy, some surgeons usually use a temporary indwelling shunt (6), most surgeons
use a shunt selectively for all the reasons enumerated above (7), and relatively few sur-
geons entirely avoid using a shunt (8).

Surgeons who always use a shunt find it is the easiest way to perform the operation,
arguing that the labyrinth of data concerning selective shunt usage is inconsistent. Thus,
rather than sift through this maze of statistics to determine indications for shunt usage, they
use the shunt in every case. These surgeons suggest that with consistent shunt usage, they
become more proficient and accustomed to the shunt's presence in the operating field, and
they do not find it an impediment to the performance of the endarterectomy. They have
developed techniques for its use even in the presence of a long atheroma located unusually
high up in the neck. In these patients, carotid endarterectomy is performed expeditiously
in the internal carotid artery, the distal endpoint is seen directly and tacked if necessary,
and finally the distal end of the shunt is inserted under direct vision, taking care not to
disturb the distal endpoint. Insertion time may be somewhat prolonged in this group; the
total time of the decreased ipsilateral hemispheric blood flow is reduced markedly
compared with that in patients in whom a shunt is not used. With a shunt in place,
closure of the arteriotomy is said to be facilitated. However, most vascular surgeons
currently use a patch angioplasty technique; thus compromise of the internal carotid artery
is a rare occurrence. Closure after an eversion carotid endarterectomy is not facilitated by
a shunt.

A. Criteria for Use I

Most surgeons find that it is technically more pleasing to perform the operation without a a

shunt. Thus surgeons who recognize the occasional need for a shunt base their decisions c

for selective use on a variety of criteria. "<

One of the earliest criteria surgeons used was the observation of the back-bleeding «

from the internal carotid artery. Pulsatile blood flow that shot across the operative field 41

clearly was adequate, whereas flow that seeped from the distal end of the internal carotid 2

artery was inadequate. The exact distinction between adequate and inadequate is difficult |

to quantify using this method. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

PREVENTION OF TIAs AND STROKES 469

In an effort to quantify this back-bleeding, stump or distal internal carotid artery back
pressures were measured. This method was first used by Moore and Hall (9) at the San
Francisco Veterans Administration Hospital and reported in 1969. Their initial and sub-
sequent experiences with the measurement of stump pressure in patients undergoing
operation using local anesthesia indicated that if the stump pressure was greater than 25
mmHg, the operation could be performed without risk of intraoperative neurological
deficit. Those patients with a reduced stump pressure were at an increased risk of intra-
operative stroke. In the same city, the University of California at San Francisco group
suggested, on the basis of their experience, that the safe level of stump pressure was 50
mmHg, not 25 mmHg (7). This latter pressure has become the accepted norm in this
country. Most surgeons who practice this technique use mean pressure rather than systolic
pressure.

This technique has its detractors also. Hobson et al. (10) reported that test occlusion
will result in neurological symptoms even with a back pressure greater than 50 mmHg.
Connelly and coworkers (11) have reported similar experiences. It would be interesting to
know whether these authors tested the adequacy of carotid occlusion as reported by
Archie (12). During measurement of stump pressure, Archie occludes the internal carotid
artery with the pickups. If indeed the stump pressure does not become a straight line at
zero, he repositions the clamp on either the external or common carotid artery to ensure
that each is occluded.

Electroencephalography has also been used successfully to determine which patients
require a temporary indwelling shunt. One champion of this technique is Callow (13) and
his group at Tufts University Medical Center. The electroencephalogram is unquestion-
ably accurate in detecting subtle abnormalities as they relate to cerebral circulation. Sur-
geons who use this technique to select patients for shunt usage find that up to one-third
require a shunt. This number is quite high considering the number of patients in whom
a shunt is used while under local anesthesia and the observance of neurological deficits
(14). Thus it must be concluded that the electroencephalogram is hypersensitive to subtle
changes resulting from decreased perfusion. Surgeons using this technique undoubtedly
will use the shunt more often, but they should feel secure that their patients are carefully
selected.

There are newer electroencephalograms in use that are easier for surgeons to interpret.
They compress the tracings and interpret wave heights in numbers. Excellent results with
this technique in carotid endarterectomy have been reported (15).

Many surgeons use local anesthesia in an awake patient to determine whether a shunt
is indicated. A field block is used to anesthetize the neck, a routine exposure is performed,
and the carotid artery is clamped. If indeed the patient exhibits a neurological deficit
during the 3-5 min of test clamping, the clamps are removed and preparation is made for
the use of a temporary indwelling shunt. Surgeons practicing this technique report that ■g

shunt usage is between 5 and 10%. Excellent results have been reported with the tech- &

nique, but detractors point out that it is not suitable for all patients and that general a

anesthetic agents, although they are myocardial depressants, also depress the cerebral -c

metabolic rate and may in themselves confer safety to the operation. <j

There are relatively few surgeons who never use a shunt. Initially our group opposed >3

the use of shunts. We reasoned that the data were not precise in selecting who should and 4j

should not have a temporary indwelling shunt, and we elected not to use them. During 2

that time we measured stump pressures in all patients. Interestingly, we came to the con- |

elusion that in patients with contralateral carotid occlusion and a stump pressure below @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

470 BAKER and SHEEHAN

50 mmHg, the stroke rate was sufficiently high (11%) to strongly consider the use of a
temporary indwelling shunt (16). In a subsequent publication, our group reported an
increased stroke risk in patients with a stump pressure below 25 mmHg regardless of the
status of the contralateral carotid artery (17). These studies form the basis of our current
pattern of selective shunt usage.

In the discussion of our initial results, we reasoned that if we had used a shunt in those
patients with a low stump pressure and a contralateral carotid occlusion, the stroke rate in
this subset would have been reduced from 11 to 2%. With the use of a shunt, seven
patients who indeed had a stroke might have been spared this neurological deficit. Overall,
this would have reduced the incidence of permanent neurological deficit from 2 to 1.1%.
This small reduction in stroke, although significant, especially for those seven patients,
underscores that there is a preeminent role for other factors in the production of oper-
ation-related stroke.

II. AVOIDANCE OF INTRAOPERATIVE EMBOLIZATION

What are these other factors? Most surgeons agree that they are the avoidance of intra-
operative embolization and the performance of a technically perfect operation. The inci-
dence of intraoperative embolization is increased if the surgeon is rough in handling the
carotid bifurcation before the application of appropriate vascular clamps. Evans, of Co-
lumbus, Ohio, suggests that those surgeons who obtain superior results with local anesthesia
may do so because local anesthesia forces them to be gentle during the performance of the
operation (personal communication, 1981). We make a special effort not to dissect the
carotid bulb until after the appropriate clamps have been placed to avoid the possibility of
emboli. The incidence of embolization of air and atherosclerotic debris can be increased by
the use of a temporary indwelling shunt unless the shunt is inserted perfectly.

III. THE TECHNICALLY PERFECT OPERATION

A technically superior operation can be performed with or without shunt usage. To assess
our technical results in the operating room, we have increasingly used B-mode imaging.
We find this technique to be superior to contrast radiology because of its ease of use and
avoidance of the theatrics that sometimes accompany intraoperative arteriography.
Initially we found several defects that required correction. Because of the use of this tech-
nique, we have routinely carried the proximal extent of endarterectomy more caudad in
the common carotid artery. This has resulted in a proximal shelf of atheroma that is less
prominent. This monitoring technique, like operative arteriography and imaging of distal
bypass grafts, forces the surgeon to be extremely meticulous, knowing that his or her oper-
ative result will be scrutinized immediately. This protects against the possibility of early
postoperative occlusion.

IV. CORRECTION OF NEUROLOGICAL DEFICITS

What if the patient has neurological deficit in the immediate postoperative period? Is there
a "golden period" during which the surgeon can boldly act before such a neurological
deficit becomes permanent? If the stroke is caused by intraoperative ischemia, clearly
nothing can be done. If the stroke is the result of intraoperative embolization, any surgical
actions taken postoperatively are akin to "closing the barn door after the horse has gone."

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i
2

I

■c

I

u

PREVENTION OF TIAs AND STROKES 471

Thus the surgeon can positively affect a postoperative stroke only if indeed there is a
mechanical blockage in the internal carotid artery at the operative site.

In prior years, all patients who awoke in the operating room with a neurological
deficit were reexplored. We assumed that the patient with an internal carotid artery
occlusion would benefit from this algorithm. Kwaan and colleagues have established that
a prompt restoration of flow results in total reversal of a neurological deficit in selected
patients (18). In all likelihood the duration of this golden period is 1-2 h. In some patients
it is clearly much shorter. This policy is still used in many institutions and cannot either be
faulted or proven to be efficacious (19).

At Loyola we have used intraoperative duplex to assess our technical result since 1988.
During that time we have evolved a different therapeutic algorithm for the patient who has
a stroke after carotid endarterectomy (20). Those patients who awaken in the operating
room, sometimes within 15 min of a normal intraoperative duplex, are rarely reexplored.
Of the 15 patients in our series who awoke from surgery with a related deficit, 5 were
reexplored and all had patent internal carotid arteries. Nine of the remaining 10 had a
duplex-proven patent internal carotid artery. One patient without lateralizing signs was
found to have extensive thrombosis at postmortem.

However, those patients who awaken neurologically intact and then later, after a lucid
period, exhibit a related neurological deficit are assumed to have an internal carotid artery
occlusion and thus are returned to the operating room. We had 10 patients in this cate-
gory; 6 were reexplored, and all had thrombosed internal carotid arteries. Of these 6
patients, 5 improved after thrombectomy. One of the patients who was not reexplored had
an observed embolus during the original intraoperative duplex; a post operative duplex at
the time of a neurological deficit revealed a perfectly normal internal carotid artery. The
second patient had a transient ischemic attack and a duplex-proven normal internal caro-
tid artery. Two additional patients had delayed recognition of their deficit (greater than
2 h) as well as a normal postoperative duplex and were not reexplored. The only patient of
these 10 who did not improve neurologically was a man who was reexplored and had flow
restored within 1 h of onset of his neurological deficit.

At reexploration, the surgeon often feels a pulse in the operating area. Our bias is that
these arteries ought to be opened or at the very least be interrogated with intraoperative
duplex. The pulse at the operative site may be associated with a distal thrombus. If a
technical cause for the thrombosis cannot be found, it is our habit to reclose the arteries
with a vein patch. Occasionally the artery is replaced with either vein or polytetrefluoro-
ethylene.

Patients who exhibit a postoperative transient ischemic attack with rapid normal-
ization of their neurological function pose a special problem. Most patients do not have a
preoperative arteriogram; their surgery is based solely on duplex ultrasonography. Many
patients do not have preoperative computed tomography or magnetic resonance imaging.
Thus, such patients need a complete workup to ensure that the operative site is pristine
and that other unrecognized pathology is not present.

A. Neurological Rescue

Many patients who undergo exploration for stroke after carotid endarterectomy will have
a relatively clean endarterectomy site. Many may have had intracranial embolization. In
prior years, emboli were treated with heparin, and sometimes these patients were switched
to warfarin. At the present time invasive neuroradiologists are able to traverse the
endarterectomy site and identify thrombi in the anterior and/or middle cerebral arteries.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

472 BAKER and SHEEHAN

Either by mechanically passing a catheter through the clot or direct administration of
thrombolytic therapy, these clots can often be dissolved or broken up to restore flow to the
affected brain. Thus it is our current practice to call neuroradiology when we begin an
exploration for a stroke after carotid endarterectomy. If, at the end of the procedure, we
suspect that the patient has had an embolism and this same patient awakens with a
significant neurological deficit, the patient is transported directly to the neuroradiology
suite from the operating room. A repeat arteriogram is then performed; if a distal embolus
is identified, the embolus is treated as outlined above.

Direct intervention for an intracerebral embolus is debatable in neurological circles
and is clearly not the standard of care for a post-carotid endarterectomy embolus in 2002.
However, neurorescue as described above may be the most promising therapy on the
horizon. Fortunately this situation is rare. We have only embarked on this therapy in one
patient to date and cannot comment on its efficacy.

V. CONCLUSION

The main methods used to avoid stroke during carotid endarterectomy are the use of an
indwelling shunt during the performance of the operation to prevent intraoperative
ischemia, avoidance of intraoperative embolization by gentle operative techniques, and
performance of a technically perfect operation that prevents early postoperative occlusion.
A variety of methods may safely predict who should and should not require a temporary
indwelling shunt. Surgical technique that avoids trauma to the carotid bifurcation must be
learned but cannot be monitored. Intraoperative arteriography, B-mode imaging, and
other imaging techniques are of great assistance in detecting defects that could lead to
postoperative stroke and are recommended especially for those centers whose existing
morbidity and mortality rates are judged to be excessive. If indeed postoperative stroke
occurs, especially after a period during which the patient was neurologically intact, imme-
diate reexploration of a totally thrombosed carotid and reestablishment of cerebral blood
flow may result in total or partial restoration of function.

REFERENCES

1 . North American Symptomatic Carotid Endarterectomy Trial Collaborators. Beneficial effect of
carotid endarterectomy in symptomatic patients with high-grade carotid stenosis. N Engl J
Med 325(7):445-451.

2. North American Symptomatic Carotid Endarterectomy Trial Collaborators. Benefit of carotid
endarterectomy in patients with symptomatic moderate or severe stenosis. N Engl J Med
339(20): 1415-1425.

3. Executive Committee for Asymptomatic Carotid Artery Stenosis. Endarterectomy for asymp- •o
tomatic carotid artery stenosis. JAMA 1995; 273(18):1421-1428. |

4. Winslow CM, Solomon DH, Chassin MR, et al. The appropriateness of carotid endarter- 2
ectomy. N Engl J Med 1988; 318:721-727. t

5. Beebe HG, Clagett P, DeWeese JA, Moore W, et al. Assessing risk associated with carotid =s
endarterectomy. Stroke 1989; 20:314-315. g

6. Thompson JE, Talkington CM. Carotid endarterectomy. Ann Surg 1976; 184:1-15. ,-,-

7. Hays RJ, Levinson SA, Wylie EJ. Intraoperative measurement of carotid back pressure as a S
guide to operative management for carotid endarterectomy. Surgery 1972; 72:593. 2

8. Ott DA, Cooley DA, Chapa L, et al. Carotid endarterectomy without temporary intraluminal I
shunt: Study of 309 consecutive operations. Ann Surg 1980; 191:708-714. |

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

PREVENTION OF TIAs AND STROKES 473

9. Moore WS, Hall AD. Carotid artery back pressure: A test of cerebral tolerance to temporary
carotid occlusion. Arch Surg 1969; 99:702-710.

10. Hobson RW, Wright CB, Jublett JW, et al. Carotid artery back pressure and endarterectomy
under regional anesthesia. Arch Surg 1974; 109:682.

11. Connelly JE, Kwaan JH, Stemmer EA. Improved results with carotid endarterectomy. Ann
Surg 1977; 186:334.

12. Archie JP. Technique and clinical results of carotid back pressure to determine selective
shunting during carotid endarterectomy. J Vase Surg 1991; 13:319-327.

13. Callow AD. The Leriche Memorial Lecture. J Cardiovasc Surg 1980; 21:641-658.

14. Imparato AM, Ramirez A, Riles T, et al. Cerebral protection in carotid surgery. Arch Surg
1982; 117:1073-1078.

15. Tempelhoff R, Modica PA, Grubb RL, et al. Selective shunting during carotid endarterectomy
based on two-channel computerized electroencephalographic compressed spectral array anal-
ysis. Neurosurgery 1989; 24:339-344.

16. Baker WH, Littooy FN, Hayes AC, et al. Carotid endarterectomy without a shunt: The control
series. J Vase Surg 1984; 1:50-56.

17. Littooy FN, Halstuk KS, Mamdani M, et al. Factors influencing morbidity of carotid end-
arterectomy without a shunt. Am Surg 1984; 50:350-353.

18. Kwaan JH, Connolly JE, Sharefkin JB. Successful management of early stroke after carotid
endarterectomy. Ann Surg 1979; 190:676-678.

19. Edwards WH, Jenkins JM, Edwards WH Sr, et al. Prevention of stroke during carotid end-
arterectomy. Am Surg 1988; 54:125-128.

20. Sheehan MK, Littooy FN, Greisler HP, et al. The effect of intraoperative duplex upon the
treatment of post carotid endarterectomy stroke. Presented at the 59th annual meeting of the
Central Surgical Association, March 9, 2002.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

28

Nonstroke Complications of Carotid Endarterectomy

Caron Rockman and Thomas S. Riles

New York University School of Medicine, New York, New York, U.S.A.

The benefits of carotid endarterectomy (CEA) over the medical treatment of patients with
both symptomatic and asymptomatic carotid stenosis have been confirmed by random-
ized, prospective clinical trials (1,2). However, the advantage of surgical therapy is
achieved only if the complications of carotid surgery are maintained at an extremely
low level. Particularly in asymptomatic patients, in whom the margin of benefit in stroke
prevention is less remarkable, reducing the incidence of post-CEA complications is critical.
The most commonly discussed and analyzed complication of CEA is, of course, stroke.
However, a variety of other complications can occur and can cause considerable morbidity
in both the short and long term. This chapter reviews the nonstroke complications that can
compromise the outcome of CEA.

The general complications of CEA can be divided into neurological and nonneuro-
logical complications. Neurological complications include cranial nerve injury, the cerebral
hyperperfusion syndrome, and ischemic stroke. Nonneurological complications of CEA
include perioperative myocardial infarction and other cardiac complications; cardiopul-
monary events as a complication of vascular surgical procedures in general are covered in
another chapter of this book. However, perioperative hemodynamic instability consisting of
either postoperative hyper- or hypotension is specific to CEA as opposed to other vascular
surgical procedures and is discussed below. Other nonneurological complications of CEA
include wound complications as well as complications related to patch material, hematoma, ■&

and postoperative infection. Finally, the development of recurrent carotid stenosis can be %

considered a late complication of CEA; this topic is addressed elsewhere in this book. £

■a

■c

I. NEUROLOGICAL COMPLICATIONS !

s

A. Cranial Nerve Injury |

Although not a frank cerebral infarction, the rare permanent, severe cranial nerve injury 2

following CEA can have equally devastating consequences to a patient's quality of life. |

Cranial nerve injury following CEA is well documented in the literature. However, it is @

475 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

476 ROCKMAN and RILES

usually a transient neuropraxia: the mechanism of injury is thought to be most often
related to stretch, retraction, or clamping of the involved nerve rather than outright
transection of the structure (3). Given the familiarity of most carotid surgeons with the
anatomy of the neck, severe cranial nerve injuries most often occur in unusual circum-
stances such as reoperative carotid surgery or surgery involving unusually high lesions (4).
The reported incidence of cranial nerve dysfunction following CEA varies widely from
approximately 3-23% in representative series (5-16); the discrepancy in the reported
incidence of this complication most likely depends on the exact methodology by which
cranial nerve dysfunction is defined and diagnosed. The true incidence of clinically
significant cranial nerve injuries is more difficult to ascertain and is likely much lower.

The cranial and cervical nerves at risk during CEA include the following: the
hypoglossal, vagus, recurrent laryngeal, marginal mandibular, superior laryngeal, glosso-
pharyngeal, spinal accessory, transverse cervical, and greater auricular nerves. Severe
injuries to the hypoglossal, vagus, recurrent laryngeal, or glossopharyngeal nerves have
the potential to result in distressing clinical consequences. Severe injury to the hypoglossal
nerve can result in tongue clumsiness and biting, dysarthria, and impaired mastication and
deglutition (3). Vagal and/or recurrent laryngeal nerve injury can result in vocal cord
paralysis, hoarseness, or airway obstruction when bilateral. Glossopharyngeal injury most
often results from unusually high dissection during CEA, especially if the posterior belly of
the digastric muscle is divided (3). Symptoms of glossopharyngeal injury can range from
dysphagia to recurrent aspiration, respiratory failure, and malnutrition. Tracheostomy
and feeding jejunostomy may rarely be required.

Injury to the spinal accessory nerve is rare, given its typical location away from the
field of the routine CEA. Injury to the superior laryngeal branch of the vagus can result in
subtle voice changes that may be troublesome to singers or public speakers (3). Injuries to
the transverse cervical and greater auricular nerves produce generally benign sensory
losses. Injury to the marginal mandibular branch of the facial nerve is usually transient
and of mainly cosmetic concern.

Several recent prospective studies have specifically examined the incidence and etiology
of cranial nerve injuries following CEA. Maroulis, et al. (14) prospectively evaluated 269
operations. They found a 5.6% incidence of documented injury, including unilateral vocal
cord paralysis (2.6%), hypoglossal palsy (3.3%), glossopharyngeal injury (0.7%), and
marginal mandibular palsy (0.4%). All patients showed improvement within a few weeks,
and none had residual disability at times ranging from 2 weeks to 14 months. Ballotta, et al.
(15) prospectively reviewed 200 consecutive CEAs. They report a 12.5% incidence of
injuries, with 5.5% hypoglossal, 4% recurrent laryngeal, 1% superior laryngeal, 1%
marginal mandibular, and 1% greater auricular nerve injuries. All nerve dysfunctions were
transient; two patients with recurrent laryngeal nerve dysfunction had prolonged but full
recoveries within 37 months. Forssell et al. (16) prospectively studied 689 operations. •§

Injuries were found in 11.4% of cases, including hypoglossal (10.7%), recurrent laryngeal g

(1.2%), glossopharyngeal (0.3%), and superior laryngeal (0.3%) injuries. One hypoglossal a

and one recurrent nerve injury were permanent. Nerve injury was more frequent in c

operations performed with a shunt (p — 0.05), with patch closure (p = 0.01), and by a <j

junior surgeon (p — 0.05). Finally, AbuRahma et al. (4), in a prospective study of 89 carotid >9

reoperations, reported a 21% rate of cranial nerve injury, with 3 permanent injuries. 4j

In summary, cranial nerve dysfunction following CEA is relatively common, but most 2

injuries are transient and do not have great clinical significance. Treatment when required, |

is expectant and supportive. In the case of planned bilateral CEAs, care must be taken to @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF CAROTID ENDARTERECTOMY 477

document vocal cord recovery prior to proceeding with contralateral surgery. Severe
cranial nerve injuries can best be avoided by excellent knowledge of the variations in
cranial and cervical nerve anatomy, attention to detail in dissection and retractor place-
ment, and meticulous placement of vascular clamps.

B. Cerebral Hyperperfusion Syndrome

The cerebral hyperperfusion syndrome with resulting intracerebral hemorrhage (ICH) is
one of the most feared complications by surgeons who perform CEAs. Although relatively
infrequent, this complication can have tremendous and often fatal sequelae, and it remains
a significant cause of neurological morbidity following CEA. ICH following cerebral
revascularization was recognized as early as 1964 by Wylie et al. (17) and was found to be
a common and catastrophic complication in patients who underwent CEA or carotid
thrombectomy for an acute stroke (18). In 1984, Bernstein, et al. (19) reported the first
CEA patient with the classic symptoms of postoperative unilateral head, face, and eye
pain; the patient subsequently developed seizures and delayed ICH and then died.
Cerebral hyperperfusion from lack of vascular autoregulation was proposed as the
mechanism of this hemorrhage, which occurred in the absence of other, previously
identified risk factors (18,19). The common denominator in the pathophysiology of the
syndrome appears to be reactive hyperemia. In its less severe forms, hyperperfusion can
result in mild cerebral edema, headache, and seizures. When an abnormal, hyperperfused
vessel ruptures, ICH results (18).

The true incidence of the cerebral hyperperfusion syndrome following CEA is
unknown. In its mild form, it is probably more common than clinically recognized. The
reported incidence from several large series ranges from 0.4 to 2% (18). In the authors'
institution, an evaluation of 1500 CEAs performed between 1975 and 1984 identified 11
patients with ICH, for an incidence of 0.7% (20). However, the mortality rate among these
patients was 36%. A more recent review of 2024 CEAs performed between 1985 and 1997
identified only 5 patients with intracranial hemorrhage (0.3%). However, ICH still
accounted for 5 of 38 perioperative neurological deficits that occurred after CEA during
this period (13.2%) (18). An additional report by Ouriel et al. (21) revealed an incidence of
0.75% in 1471 patients during a 6-year period, with massive hemorrhage and death
occurring in 4 cases. Considering the severe morbidity of this particular complication, it is
critically important to understand which patients are at increased risk.

The most commonly identified factors that predispose a patient to the development of
the cerebral hyperperfusion syndrome include a history of stroke, especially when recent;
relief of a severely stenotic lesion (>90%); severe intraoperative and postoperative hyper-
tension; anticoagulant use; severe chronic cerebral ischemia; and perhaps occlusion of the
contralateral carotid artery (18,20). In the report by Ouriel et al. (21), five factors were •§

associated with a statistically increased risk of intracranial bleeding: increased age, a history g

of hypertension, a high degree of ipsilateral stenosis, a high degree of contralateral stenosis, a

and contralateral carotid occlusion. Of course, one or more of these recognized risk factors c

are often present in many patients undergoing CEA. Clinically evident hyperperfusion does <j

not occur in the majority of patients in whom one or even more of these risks exist. >9

However, awareness of the early symptoms of the syndrome is important in these cases. 4j

In the authors' series (18), most postoperative neurological deficits related to 2

thrombosis or embolization occurred within the first 4 h after operation. In marked |

contrast, three of five intracranial hemorrhage occurred on postoperative day 4 or later. @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

478 ROCKMAN and RILES

These patients presented with headaches, seizures, progressive obtundation, or hemi-
paresis. It appears that the symptoms of cerebral hyperperfusion are more likely to occur
somewhat later in the postoperative course than deficits associated with thromboembo-
lization and are also more likely to present with either headache or seizures than with
hemiparesis alone. The diagnosis of the syndrome is often clinical and rests heavily on the
surgeon's suspicion in a patient with appropriate risk factors. Computed tomograpy (CT)
scanning to evaluate for hemorrhage or edema is the test of choice in any patient who
presents with severe frontoparietal or orbital pain or seizure activity. Electroencephalog-
raphy may reveal lateralizing epileptiform discharges or frank seizure activity.

Prophylaxis is difficult because it would require predicting which patients will sustain
such an event. However, in patients thought to be at increased risk, certain measures are
reasonable: strict blood pressure control, judicious use of anticoagulants and antiplatelet
agents, and close monitoring of the neurological status are indicated. In patients with
headaches only, simple analgesia may suffice. In patients with more severe symptoms,
antihypertensives and anticonvulsant medications may be warranted. If cerebral edema is
significant, diuretics and anti-inflammatory medications may be utilized. If petechial or
small cerebral hemorrhages occur, the above measures will often suffice (21). However, if
massive hemorrhage occurs, neurosurgical intervention may become necessary. However,
in all reported series, the prognosis of massive ICH after CEA is exceedingly poor (18).

II. NONNEUROLOGICAL COMPLICATIONS
A. Vein Patch Rupture

Patch angioplasty reconstruction of the carotid artery following CEA is an accepted
technique and may reduce the incidence of postoperative thrombosis, embolization, and
recurrent carotid artery stenosis. However, the ideal material for patch reconstruction
remains to be defined. Certainly the greater saphenous vein has the advantages of being
autologous, readily available, and easy to use. However, a rare but potentially life-
threatening complication of its use is rupture of the vein patch following CEA. Its reported
incidence in the literature ranges from to 4.0% (22-27). Manifestations of vein patch
rupture in the early postoperative period obviously include massive hemorrhage into the
neck with airway compromise, respiratory arrest, and possible death.

Yamamoto et al. (25) reported 5 postoperative vein ruptures among 2888 CEAs
(0.17%). All ruptures occurred within 4 days of surgery, including 2 during the first 24 h.
All patients were found to have intact suture lines and tears in the middle of the grafts.
Two patients died. Tawes et al. (26) reported a combined experience of 1760 operations, in
which vein patch rupture occurred in 13 patients (0.7%). In 12 of these, the saphenous vein
was harvested from the ankle. All ruptures occurred from a split in the center of the vein
patch. Four patients died (30.7%), and three had strokes but survived. Riles et al. (27)
reported a series of 2275 CEAs; in 3 patients out of 75 in whom the vein patch had been

harvested from the ankle, rupture of the patch occurred. In all cases, reoperations revealed -c

necrosis of the central portion of the vein with no evidence of infection. O'Hara et al. (23) <j

reported 8 postoperative ruptures after 1691 CEAs in which saphenous vein patch angio- >9

plasty was used (0.5%). In each case, the vein had been harvested distal to the knee. Of the 4j

patients in whom rupture occurred, 29% either died or sustained strokes. 2

Based on the above studies as well as his own experience, Archie prospectively |

established diameter criteria for the use of the saphenous vein as a patch material (22). @

In a 7-year prospective study of 614 CEAs, using only greater saphenous veins with a %

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF CAROTID ENDARTERECTOMY 479

distended diameter of greater than 3.5 mm, no ruptures occurred. This compared
favorably with our own prior experience with 3 patch ruptures in 239 cases when no vein
diameter criteria were utilized (p — 0.03).

In summary, although saphenous vein is an excellent patch material, its results may be
compromised by the devastating complication of early patch rupture. This complication
often results in death or severe morbidity. The incidence of vein patch rupture can probably
be reduced or nearly eliminated by using greater saphenous vein harvested from the thigh
or using only vessels greater than 3.5 mm in diameter when vein must be harvested from
below the knee. Late aneurysmal degeneration of vein patching for CEA, with the potential
for thrombus formation and thromboembolization, has been reported in the literature as
well (28).

B. Postoperative Hematoma

Wound hematoma following CEA has not been extensively studied specifically. However,
because of the risk of airway compromise with hemorrhage into the neck, this complica-
tion can be life-threatening. The incidence of wound hematoma reported in the literature
varies from 1 to 4.5% (29-32). Risk factors reported include perioperative hypertension,
lack of reversal of intraoperative heparinization with protamine sulfate, and postoperative
resumption of anticoagulation with heparin and/or warfarin for treatment or prophylaxis
of arterial or venous thromboembolization.

Treiman et al. (29) reported a study on the influence of protamine use following CEA
on postoperative stroke and wound hematoma. A review of 697 operations was per-
formed, in which 328 patients received protamine and 369 did not based on the practices
and judgment of the individual operating surgeons. The incidence of stroke was similar
between the groups. Thirty patients (4.3%) experienced a postoperative wound hematoma
requiring reoperation. The incidence of wound hematoma was 1.8% in patients given
protamine and 6.5% in patients not given protamine, and this difference was statistically
significant (p — 0.004). In this series, all patients were intubated without difficulty.
However, if intubation is difficult or impossible, the authors recommend opening the
wound using local anesthesia prior to intubation and reoperation. No complications
occurred in this series from evacuation of a wound hematoma.

The successful management of wound hematoma following CEA centers on immedi-
ate recognition of the complication and return to the operating room for hematoma
evacuation. Even a hematoma that initially appears benign can rapidly progress to stridor
and respiratory compromise. Although these hematomas are most often related to
seemingly minor venous bleeding or diffuse bleeding from residual heparin effect, their
consequences can be severe if not treated expeditiously.

C. Infectious Complications I

CD

While vein has long been considered the optimal standard for arteriovascular bypasses a

elsewhere in the arterial tree, concerns have arisen regarding the use of saphenous vein for -c

carotid reconstruction. Specifically, reports of late aneurysmal dilatation and patch <!

rupture, as described previously, have led to investigations for alternative acceptable patch 3

materials. Harvest site morbidity and the need for venous conduits for subsequent coronary ,jjj

or lower extremity revascularizations have also been cited as reasons to consider prosthetic Q

patch options. §

Prosthetic patches such as polytetrafluoroethylene (PTFE) and polyester are attractive q

alternatives to vein in that they are readily available, technically easy to use, and may %

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

480 ROCKMAN and RILES

reduce the risk of patch rupture or aneurysmal dilatation. While several studies have
shown that knitted polyester has been comparable to vein in rates of restenosis, potential
drawbacks to prosthetic materials include possible increases in recurrent stenosis,
increased thrombogenicity, and infection.

While potential infectious complications remain a significant concern, there have been
very few reported series of infected prosthetic patch infections, and their true incidence is
unknown. Several recent case series of carotid prosthetic patch infections have addressed
this issue (33-35). Naylor et al. (34) report that 8 of 936 CEAs performed with prosthetic
patch material developed a patch infection (0.85%). Responsible organisms included
primarily Staphylococcus and Streptococcus species. Surgical repair consisted of carotid
ligation in 3 cases and reconstruction with autologous material in 5; 1 patient suffered a
disabling postoperative stroke following reoperations, and 2 had transient cranial nerve
injuries. Rizzo et al. (35) reported an additional 8 patients with infected Dacron patches.
Of these cases from their own institution, 6 represented 1.8% of 340 synthetic patches
utilized. As in the prior series, gram-positive organisms predominated; in this series, all
carotid arteries were reconstructed with autologous material. Again, two temporary
cranial nerve injuries occurred.

The authors' institution is in the process of completing a report of 10 patients who
required reoperations and management of postoperative Dacron patch infections follow-
ing CEA (33). Half of these patients presented early with cellulitis and abscess formation
within several weeks following the original operation. However, the remaining half pre-
sented with draining sinus tracts 1-2 years following surgery. All patients had their carotid
arteries reconstructed with autologous material, with no significant morbidity. In all three
of these reports (33-35), all patients remain free of infection following surgical manage-
ment and removal of the patch.

Infection of a prosthetic patch is a rare but serious complication following CEA.
However, it can be successfully managed with an acceptably low complication rate.
Although infection of prosthetic patches is probably more common that infection of
autologous patch material, the latter has in fact been reported (36).

D. Perioperative Blood Pressure Instability

Blood pressure instability following CEA may contribute to morbidity, increase length of
hospital stay, increase costs, and occasionally result in mortality (37). Carotid barorecep-
tor dysfunction and impaired cerebral autoregulation following manipulation of the
carotid bulb and carotid sinus nerve have been implicated in the etiology of both hyper-
and hypotension post-CEA.

Reported risk factors for postoperative hypertension include having undergone an
eversion endarterectomy (38) and having undergone surgery under general as opposed to •§

regional anesthesia (39). Postoperative hypertension in particular has been shown to be g

associated with increased postoperative stroke and death and also with postoperative a

cardiac complications (40). In a report by Wong et al. (40), independent risk factors for c

postoperative hypertension included angiographic intracranial carotid stenosis greater <j

than 50%, cardiac dysrhythmia, preoperative systolic blood pressure greater than 160 >3

mmHg, neurological instability, and renal insufficiency. Postoperative hypotension and %

bradycardia did not correlate with perioperative outcome. 2

Management of post-CEA patients must include aggressive treatment of hemody- |

namic instability in order to prevent secondary complications. In addition to potential ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

COMPLICATIONS OF CAROTID ENDARTERECTOMY 481

cardiac complications, severe hypotension may theoretically predispose toward thrombo-
sis of the arterial reconstruction, and severe hypertension can increase the chance of
cerebral hyperperfusion and intracranial hemorrhage.

III. CONCLUSIONS

In addition to stroke, CEA has the potential to result in a wide variety of neurological and
nonneurological complications that have the potential to cause significant morbidity or
mortality. In order for CEA to achieve success in stroke prevention, potential perioper-
ative complications other than stroke must also be appropriately appreciated, recognized,
and managed in order to minimize their occurrence and the resulting morbidity.

REFERENCES

1. The Executive Committee for the Asymptomatic Carotid Atherosclerosis Study. Endarter-
ectomy for asymptomatic carotid artery stenosis. JAMA 1995; 273:1421-1428.

2. North American Symptomatic Carotid Endarterectomy Trial Collaborators. Beneficial effect of
carotid endarterectomy in symptomatic patients with high-grade carotid stenosis. N Engl J
Med 1991; 325:445-453.

3. Schauger MD, Fontenelle LJ, Solomon JW, Hanson TL. Cranial/cervical nerve dysfunction
after carotid endarterectomy. J Vase Surg 1997; 25:481^487.

4. AbuRahma AF, Chouerie MA. Cranial and cervical nerve injuries after repeat carotid
endarterectomy. J Vase Surg 2000; 32:649-654.

5. Maniglia AJ, Han AP. Cranial nerve injuries following carotid endarterectomy: An analysis of
336 procedures. Head Neck 1991; 13:121-124.

6. Hertzer NR, Feldman BJ, Beven EG, Tucker HM. A prospective study of the incidence of injury
to the cranial nerves during carotid endarterectomy. Surg Gynecol Obstet 1980; 151:781-784.

7. Knight FW, Yeager RM, Morris DM. Cranial nerve injuries during carotid endarterectomy.
Am J Surg 1987; 154:529-532.

8. Verta MJ Jr, Applebaum EL, McClusky DA, Yao JST, Bergan JJ. Cranial nerve injury during
carotid endarterectomy. Ann Surg 1977; 185:192-195.

9. Dehn TCB, Taylor GW. Cranial and cervical nerve damage associated with carotid
endarterectomy. Br J Surg 1983; 70:365-368.

10. Massey EW, Heyman A, Utley C, Haynes C, Fuchs J. Cranial nerve paralysis following carotid
endarterectomy. Stroke 1984; 15:157-159.

11. Lusby RJ, Wylie EJ. Complications of carotid endarterectomy. Surg Clin North Am 1983;
63:1293-1301.

12. Rodgers W, Root HD. Cranial nerve injuries after carotid endarterectomy. South Med J 1988;
81:1006-1009.

13. Evans WE, Mendelowit DS, Liapis DS, Wolf V, Florence CL. Motor speech deficit following •o
carotid endarterectomy. Ann Surg 1982; 196:461^463. §

14. Maroulis J, Karkanevatos A, Papakostas K, Giling-Smith GL, McCormick MS, Harris PL. 8
Cranial nerve dysfunction following carotid endarterectomy. Int Angiol 2000; 19:237-241. •§)

15. Ballotta E, Da Giau G, Renon L, Name S, et al. Cranial and cervical nerve injuries after a
carotid endarterectomy: A prospective study. Surgery 1999; 125:85-91. y

16. Forssell C, Kitzing P, Bergqvist D. Cranial nerve injuries after carotid artery surgery. A *"
prospective study of 663 operations. Eur J Vase Endovasc Surg 1995; 10:445-449. S

17. Wylie EJ, Hein MF, Adame JE. Intracranial hemorrhage following surgical revascularization !3
for treatment of acute strokes. J Neurosurg 1964; 21:212-215. I

18. Rockman CB, Riles TS. Cerebral hyperperfusion syndrome after carotid endarterectomy. In: @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

482 ROCKMAN and RILES

Ernst CB, Stanley JC, eds. Current Therapy in Vascular Surgery. 4th ed. St. Louis: Mosby,
2001:69-71.

19. Bernstein M, Fleming JFR, Deck JHN. Cerebral hyperperfusion after CEA: A cause of
cerebral hemorrhage. Neurosurgery 1984; 15:50-56.

20. Pomposelli FB, Lamparello PJ, Riles TS, et al. Intracranial hemorrhage after CEA. J Vase Surg
1988; 7:248-255.

21. Ouriel K, Shortell CK, Illig KA, et al. Intracerebral hemorrhage after carotid endarterectomy:
Incidence, contribution to neurologic morbidity, and predictive factors. I Vase Surg 1999;
29:82-89.

22. Archie IP. Carotid endarterectomy saphenous vein patch rupture revisited: Selective use on the
basis of vein diameter. I Vase Surg 1996; 24:346-352.

23. O'Hara PI, Hertzer NR, Krajewski LP, Beven EG. Saphenous vein patch rupture after carotid
endarterectomy. I Vase Surg 1992; 15:504-509.

24. Lawhorne TW Ir, Brooks HB, Cunningham IM. Five hundred consecutive carotid endar-
terectomies: emphasis on vein patch closure. Cardiovasc Surg 1997; 5:141-144.

25. Yamamoto Y, Piepgras DG, Marsh WR, Meyer FB. Complications resulting from saphenous
vein patch graft after carotid endarterectomy. Neurosurgery 1996; 39:670-676.

26. Tawes RL Ir, Treiman RL. Vein patch rupture after carotid endarterectomy: A survey of the
Western Vascular Society members. Ann Vase Surg 1991; 5:71-73.

27. Riles TS, Lamparello PI, Giangola G, Imparato AM. Rupture of the vein patch a rare com-
plication of carotid endarterectomy. Surgery 1991; 107:10-12.

28. Rockman CB, Riles TS, Landis R, Lamparello PI, Giangola G, Adelman MA, lacobowitz GR.
Redo carotid surgery: An analysis of materials and configurations used in reoperative carotid
surgery and their effects on perioperative stroke and subsequent recurrent stenosis. I Vase Surg
1999; 29:72-81.

29. Treiman RL, Cossman DV, Foran RF, Levin PM, Cohen IL, Wagner WH. The influence of
neutralizing heparin after carotid endarterectomy on postoperative stroke and wound hema-
toma. I Vase Surg 1990; 12:440-446.

30. Kunkel IM, Gomez ER, Spebar MI, et al. Wound hematomas after carotid endarterectomy.
Am I Surg 1984; 148:844-847.

31. Welling RE, Ramadas HS, Gansmuller KJ. Cervical wound hematoma after carotid endar-
terectomy. Ann Vase Surg 1989; 3:229-231.

32. Oiler DW, Welch H. Complications of carotid endarterectomy. A military hospital experience.
Am Surg 1986; 52:479^184.

33. Rockman CB, Wu WT, Domenig C, et al. Postoperative infection associated with polyester
patch angioplasty following carotid endarterectomy. I Vase Surg 2003; 38:251-256.

34. Naylor AR, Payne D, London NIM, et al. Prosthetic patch infection after carotid endar-
terectomy. Eur J Vase Endovasc Surg 2002; 23:11-16.

35. Rizzo A, Hertzer NR, O'Hara PI, Krajewski LP, Beven EG. Dacron carotid patch infection: A
report of eight cases. Dacron carotid patch infection: a report of eight cases. I Vase Surg 2000;
32:602-606.

36. Motte S, Wautrecht IC, Bellens B, Vincent G, Dereume IP, Delcour C. Infected false aneurysm
following carotid endarterectomy with vein patch angioplasty. I Cardiovasc Surg 1987; 28:734- j>
736.

37. Nowak LR, Corson ID. Blood pressure instability after carotid endarterectomy. In: Ernst CB, js
Stanley IC, eds. Current Therapy in Vascular Surgery. 4th ed. St. Louis: Mosby, 2001:71-73. 2

38. Mehta M, Rahmani O, Dietzek AM, et al. Eversion technique increases the risk for post- ■"■
carotid endarterectomy hypertension. I Vase Surg 2001; 34:839-845. >9

39. Sternbach Y, Illig KA, Zhang R, et al. Hemodynamic benefits of regional anesthesia for carotid jjj
endarterectomy. I Vase Surg 2002; 35:333-339. q

40. Wong IH, Findlay IM, Suarez-Almazor ME. Hemodynamic instability after carotid endarter- |
ectomy: Risk factors and association with operative complications. Neurosurgery 1997; 41:35- 2
43. 1

I

u

s

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

29

Radiation Exposure and Contrast Toxicity

Evan C. Lipsitz, Frank J. Veith, and Takao Ohki

Monteftore Medical Center and the Albert Einstein College of Medicine,
Bronx, New York, U.S.A.

I. RADIATION EXPOSURE

A. Introduction

Endovascular repair of aortoiliac aneurysms has gained widespread acceptance and the
number of investigational devices as well as devices approved by the U.S. Food and Drug
Administration for this procedure is growing. Other endovascular procedures for the
treatment of such entities as aortoiliac occlusive disease and renal artery stenosis are also
being employed more frequently. It has been estimated that the vast majority of
abdominal aortic aneurysms are amenable to treatment with an endovascular graft and
that, in the near future, 40-70% of all vascular interventions will be performed by an
endovascular method, including an increasing number of peripheral and cerebrovascular
interventions (1). These procedures require the use of digital cine-fluoroscopy, which
exposes both the patient and staff to ionizing radiation.

B. Units of Measurement

Several different measures are used in the evaluation of radiation exposure. Absorbed dose

is the energy delivered to an organ divided by the mass of the organ, expressed in grays

(Gy). Equivalent dose is the average absorbed dose in an organ or tissue multiplied by a

radiation weighting factor, expressed in sieverts (Sv). Since radiation used in medicine

generally has a weighting factor of 1, the absorbed dose and the equivalent dose are

considered equal. Total effective dose (TED) is the sum of the equivalent doses in all |j)

tissues and organs multiplied by a tissue weighting factor for each organ or tissue. It is this ^

value that is used to evaluate total body exposure (2). g

I

C. Biological Effects £

The biological effects of radiation can be divided into two types, deterministic and §

stochastic (2). Deterministic effects are observed only when many cells in an organ or ©

483 I

I

s

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

484

LIPSITZ et al.

tissue are killed by virtue of a dose that is above a given threshold. Stochastic effects are
due to radiation-induced injury to the DNA of a single cell. In this setting there is no
threshold below which the risk is eliminated. However, the probability of an effect
occurring is small. Stochastic effects may be somatic, affecting somatic cells, or hereditary,
affecting germ cells. It is these stochastic effects that are of special concern, because there is
no threshold below which their occurrence can be prevented.

There are many sources of radiation present within the environment, both naturally
occurring and man-made. These are collectively referred to as background radiation.
Radon gas constitutes the single most important source of naturally occurring external
background radiation, followed closely by solar, cosmic, and galactic radiation (3).
Natural atmospheric radiation and radionuclides also contribute to background radiation,
as do terrestrial nuclides. Nuclear reactors and nuclear weapons testing also add to
background radiation.

Radiation exposure is cumulative and its effects are permanent. The total exposure for
an individual performing fluoroscopic procedures is the sum of his or her exposure during
these procedures plus the background exposure as well as any incidental medical
exposure — e.g., diagnostic chest x-rays — incurred by the individual. In the United States,
the average person receives approximately 3.5 millisieverts (mSv) per year in background
exposure. This dose increases with altitude, doubling at every 2000 m. Other local effects,
such as radionuclides in the soil, can affect the background radiation significantly. Table 1
highlights the currently recommended dose limits for both the occupational and civilian
settings.

D. Specific Recommendations

1. General Principles

Fluoroscopy Time. Radiation exposure is proportional to total fluoroscopy time.
Therefore the most effective way to reduce exposure to both the patient and staff during
endovascular procedures is to reduce the total fluoroscopy time. Several steps can be taken
toward this end. Catheter-guidewire exchanges with a stable wire position do not need to
be visualized in their entirety. When repositioning the field of interest either by moving the
table or the C-arm, the desired position should be estimated and then fine-tuned under
fluoroscopy rather than imaged along the entire course. This is also true when obtaining
oblique or angled projections. In performing cine-acquisition, each screening should be
carefully planned and have a specific objective. Poorly planned runs add no information to

Table 1 Yearly Recommended Dose Limits

Dose limit (mSv/year)

Application

Occupational

Effective dose

20

Equivalent dose

lens of eye

150

skin

500

hands and feet

500

Public

15

50

Source: From Ref. 2.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

RADIATION EXPOSURE AND CONTRAST TOXICITY

485

the procedure and increase exposure, contrast load, and operative time. For example, a
subtraction run over the upper abdomen without holding the respirations — either by
anesthesia in the intubated patient or voluntarily in the awake patient — is likely to
produce a useless image. All individuals involved in the procedure must be constantly
aware of when the fluoroscope is on. When the fluoroscope is on, they must evaluate
whether or not essential information is being obtained or if it is necessary for the technical
performance of the procedure. Simply measuring the fluoroscopy time on a routine basis
may increase awareness enough to reduce overall fluoroscopy time. Because radiation is
not detectable by any of the five senses, coupling the fluoroscopy "on" setting to a
detectable signal facilitates control of exposure. Hough et al. found that the use of audible
dose-sensitive radiation monitors led to a significant reduction in exposure to the staff
wearing them (4).

Distance from Source. The next most effective way to reduce exposure is to increase
the distance from the source. The exposure to the operator resulting from scatter decreases
with the square of the distance from the source. This is known as the inverse square law.
There is a substantial drop in scattered radiation to an operator when he or she moves to
between 30 and 50 cm from the scatter source (5,6). For most endovascular interventions,
the working distance from the source is largely fixed by the distance between the area of
interest and the arterial access site (Fig. 1). The radiation dose to the operator during
cardiac interventions has been shown to increase by 1.5-2.6 times when the operator
moves from the femoral to the subclavian position (7). Kuwayama found that radiation to
the operator was increased by approximately two to three times when a transcarotid versus

I
•5

Figure 1 Showing fixed working distance from the sheath to the area of interest, in this case the
abdominal aorta.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

486 LIPSITZ et al.

a transfemoral route was used for neuroradiological procedures (8). In this same study, the
transcarotid approach led to a tenfold increase in exposure to the hands due to proximity
to the beam.

Endovascular aortoiliac aneurysm repair requires prolonged imaging over the abdo-
men and pelvis. Penetration of these tissues requires more energy and results in a
significantly higher exposure than imaging the periphery (9).

2. Use of the Fluoroscope and Patient Positioning

The radiation exposure of the operator is proportional to that of the patient. Therefore
reducing the patient exposure will also reduce the operator exposure. Several methods can
be used to achieve these ends. The beam should be positioned under the patient — i.e.,
posteroanterior imaging (Fig. 2A). This will decrease scatter as well as exposure of the
operator's hand. Placing the beam in the anteroposterior position (source anterior to
patient, image intensifier posterior to patient, patient supine) results in approximately four
times greater exposure to the operator's head, neck, and upper extremities (Fig. 2B) (5).
Not only is the exposure higher but these areas are far more difficult to shield than the area
below the waist. Obtaining oblique views will also impact on the scattered radiation dose.
The right anterior oblique view will result in significantly more scatter to an operator
standing on the patient's left than the left anterior oblique view. The reverse is true when
the operator stands on the patient's right (10).

The image intensifier should be positioned as close as possible to the patient. This
reduces the amount of scatter by allowing for a lower entrance exposure. This positioning
also results in a sharper image (Fig. 3A and B). Pulse-mode fluoroscopy at rates of 15-30
frames per second or lower greatly reduces exposure as compared to continuous mode.

A larger image intensifier mode requires less radiation than a smaller one. The
radiation dose approximately doubles with each successively smaller image intensifier
setting — i.e., increasing the magnification (1 1). Large image intensifier sizes should be used
whenever possible. Avoid excessive use of high-level or cine-fluoroscopy mode. This mode
should be used only for essential acquisitions.

The amount of radiation produced by the fluoroscope is dependent on the energy used
to generate the beam. The factors determining this are the milliamperes (mA) and kilovolts
(kV). The mA setting controls the number of photons produced (1 1). Low mA produces a
mottled image, which can be eliminated by increasing the mA at the cost of higher
radiation. The kV control determines the penetration of the beam and image contrast. For
most fluoroscopic units, the mA and kV settings are determined by an automatic bright-
ness control, which sets the values using feedback from the image obtained. However,
where these are not set, the use higher kV and lower mA techniques will reduce exposure
while not greatly affecting image quality. One study found that increasing the fluoroscopy
voltage from 75 to 96 kV decreased the entrance dose by 50% (12). ■§

There are factors intrinsic to the fluoroscopic unit itself that affect the radiation dose. |

These include the design and manufacture of the unit. Mehlmen et al. found that deep and a

shallow unprotected collar exposure as well as eye exposure were increased by at least 1.5 c

times when using an OEC 9600 as compared to a Philips BV 29 (6). There was also a <

substantial increase in deep and shallow unprotected waist exposure. These differences are >9

likely accounted for by the increased mA generated by the OEC 9600 (3.3 mA/69 kV) as J

compared to the Philips (2.7 mA/72 kV). Newer models include a "low-dose" button, «

which allows the operator to reduce the exposure without significantly compromising |

image quality. In another study, Watson et al. found a statistically significant difference @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

RADIATION EXPOSURE AND CONTRAST TOXICITY

487

Figure 2 A. Posteroanterior imaging showing the majority of scatter directed at the level of the
patient and below. B. Anteroposterior imaging showing the majority of scatter directed at the level
of the patient and above.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

488

LIPSITZ et al.

Figure 3 A. Image intensifier located close to the patient. Less energy is required for tissue
penetration because scatter is reduced, resulting in a clearer image with decreased dose. B. Image
intensifier located far away from the patient. More energy is required for tissue penetration because of
increased scatter, resulting in increased exposure to obtain adequate image quality.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

RADIATION EXPOSURE AND CONTRAST TOXICITY 489

between two wall-mounted units that used different imaging technologies (13). A General
Electric LU-C MPX/L500 PULSCAN 17178 Video Processor using pulsed progressive
fluoroscopy showed a 45% higher dose per case than a Philips DCI-S Poly-Diagnostic using
digital imaging technology. This difference was largely due to differences in the techniques
used for image acquisition, since progressive pulsed fluoroscopy generally reduces radiation
exposure. Finally, a heavier patient will require greater radiation energy to penetrate the
tissues, with a consequent increase in radiation exposure to the patient and staff. We have
found increased doses in heavier patients, although the amount is difficult to quantify
because of the variability in the amount of high-level fluoroscopy used in each case.

Although the collimation of all fluoroscopic units is regulated by federal law, the ratio
of the field of view to the total exposed area is not 1:1. Granger et al. found that the percent
difference between the total exposed area and the field of view may be quite significant
even though the fluoroscopic unit is in compliance with federal regulations (14). They
evaluated 18 fluoroscopic units from different manufacturers and of different ages and
found that only 67% of the units met federal compliance standards. For units not in
compliance, the measured difference between the total exposed area and the field of view
ranged from 22 to 48%. For units in compliance, the difference ranged from 5 to 32%.
This excess exposed area provides no additional clinical information, increases the
radiation doses to the patient and staff, and reduces image contrast and quality. After
the units were serviced, a 40% average reduction in beam area was achieved and 100% of
the units met compliance standards.

Although automatic collimation is part of all current systems, reducing the field size
by using manual collimation will greatly decrease exposure and has the added benefit of
enhancing image quality by reducing stray radiation. Lindsay et al. found that by
collimating the field of image during radiofrequency catheter ablation procedures, the
radiation dose to the patient and staff was reduced by 40% (7).

Antiscatter grids mounted in front of the input screen decrease the amount of scatter
reaching the image intensifier and improve image quality by doing so. They also greatly
increase both the required radiation to obtain a satisfactory image and the backscatter to
the patient and staff (15). Removal of these grids can reduce the radiation dose by factor of
2-4, but with some loss of resolution. This is not the case during pediatric procedures,
where grids can and should be removed without loss of image quality (15).

The fluoroscope should undergo at least biannual inspection and calibration, as
required by law. More frequent quality-control checks are probably in order. If the unit
requires service and any components are replaced, the fluoroscope should be recalibrated.

3. Radiological Protection

Protective barriers must be readily available and used consistently. The most important of

these is the lead apron. These are generally available in 0.5- and 0.25-mm thicknesses. In ■§

optimal circumstances, the 0.5-mm thickness has the ability to attenuate 98-99.5% of the g

radiation dose, while the 0.25-mm thickness attenuates approximately 96% of the dose »

(11,16). Deterioration of the apron's lead lining occurs with use and is increased by rough c

handling or improper storage. Aprons should undergo periodic screening and replacement <

if inadequate protection is found. Many aprons are not of the wraparound type and as >9

such do not provide circumferential protection. Scattered radiation from the sides may J

produce unprotected exposure. A thyroid collar and "protective" glasses are essential. °

These glasses are highly variable in the amount of protection afforded and allow for |

anywhere from 3-98% transmission of the radioactive beam (17). The greatest protective @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

490 LIPSITZ et al.

effect is obtained with glasses containing lead. Glasses at the lower end of this spectrum
may provide protection against ultraviolet (UV) but not against ionizing radiation. Also of
note is that a significant amount of the ocular exposure, up to 21%, is the result of scatter
from the operator's head (17). Depending on the head position of the operator during the
procedure, side shields or wraparound configurations are necessary to provide adequate
protection. A lead acrylic shield, which can either be ceiling-mounted or on a mobile floor
stand, should be placed between the operator and patient to further reduce exposure. Eye
radiation can be reduced by a factor of 20-35 with the use of a ceiling-suspended lead glass
shield (7,10). Lead-lined gloves also help to reduce exposure but can be cumbersome.
Because back-scattered radiation is more intense than forward-scattered radiation, and
because with the C-arm in the posteroanterior orientation the greatest exposure due to
scatter occurs from under the table, we use a lead drape suspended from the operating
table on the operator's side to reduce this exposure (18). Use of this additional shield
eliminates a significant amount of this scatter (5).

4. Role of Experience

Certain endovascular procedures may be quite complex and require lengthy fluoroscopy
times. This may be the case especially at tertiary referral centers, which generally have
affiliated training programs. In a study of radiation exposure during cardiology fellowship
training, Watson et al. found a statistically significant increase in exposure for cases done
in the first versus the second year of fellowship (13). This difference was largely accounted
for by an increase in fluoroscopy time but not cine time, reflecting an increased time for the
less experienced operators to position the catheters. These results have implications for
fellowship training programs, where the teaching of less experienced operators will result
in increased radiation exposure for the patients and staff alike. The needs of training must
be balanced against the increased fluoroscopy times and resulting exposure.

5. Patient and Staff Monitoring

The use of radiation badges by all persons working with fluoroscopy is mandatory and
required by law. The position of the badges is important. A badge must be worn at waist
level under the lead apron. Additional badges should also be worn on the collar to monitor
the head dose and to aid in calculating the total effective dose, since there is a large and
variable difference between the over- and under-lead doses (19). Ring badges are also
advisable. Waist and collar badges should be worn toward the operator's left side in
working on the patient's right side and toward the operator's right side in working on the
patient's left side — i.e., the badge should face the source directly. Ring badges should be
worn on the hand most likely to be exposed. A self-retaining device to stabilize the sheaths
may also reduce exposure. Monitoring of all at-risk body positions is essential, since doses
to the fingers of the dominant hand have been shown not to correlate with doses estimated ■§

by shoulder badges in interventionalists performing percutaneous drainage procedures g

(20). Although the use of badges is mandated, it is the responsibility of the individual to »

wear them and of the institution to have a monitoring program with feedback to the c

exposed individuals in place. <

In general, the patient is exposed only once. Many patients undergoing these pro- >9

cedures are in an older age range and are less likely to suffer from potential malignancies. J

However, because of the long screening times, patients should be warned about transient «

skin erythema, which may present up to several weeks following the procedure, and other |

skin conditions. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

RADIATION EXPOSURE AND CONTRAST TOXICITY 491

In one large prospective study of interventional radiologists, Marx et al. found that the
only variable correlating with the over-lead collar dose was number of procedures performed
per year and the only variable correlating with the under-lead waist dose was the thickness of
the lead apron (0.5 vs. 1 mm) (19). This study also included a questionnaire that inquired
about the practice habits of the interventional radiologists involved. Nearly half of the
respondents reported rarely or never wearing their radiation badges. Half of the respondents
either had exceeded or did not know whether they had exceeded monthly or quarterly
occupational dose limits at some time within the past year. With regard to protection habits,
30% rarely or never wore a thyroid shield, 73% rarely or never wore lead glasses, 70% rarely
or never used a ceiling-mounted lead shield, and 83% rarely or never wore leaded gloves.
These results indicate that there can be significant complacency even among the most at-risk
population of physicians who have had substantial background and education in radiation
safety and physics.

We reviewed our own radiation exposure incurred during 47 endovascular aortic and/
or iliac aneurysm repairs performed over a 1-year period (21). We did not include other
fluoroscopic procedures we routinely perform, such as diagnostic or completion angiog-
raphy, iliac or renal artery angioplasty and stenting, fluoroscopically assisted thromboem-
bolectomy, and inferior vena cave filter placement.

Each of three surgeons wore three radiation dosimeters (Landauer, Inc., Glenwood,
IL), as follows: (a) on the waist under the lead apron, (b) on the waist outside the lead
apron, and (c) on the collar outside the thyroid shield. A ring dosimeter was worn on the
ring finger of the left hand by each surgeon. Additional badges were placed around the
operating room to estimate the exposure to the scrub and circulating nurses. Patient
entrance doses were calculated using the fluoroscopic energies and positions recorded
during each case. Total effective doses (TEDs) were calculated and compared to standards
established by the International Commission on Radiological Protection (ICRP) (2).

Yearly TEDs for the surgeons (under lead) ranged from 5 to 8% of the ICRP occu-
pational exposure limit. Outside lead doses for all surgeons approximated the recom-
mended occupational limit. Ring and calculated eye doses ranged from 1 to 5% ICRP
occupational exposure limits. Lead aprons attenuated 85-91% of the dose. Patient
entrance doses averaged 360 mSv per case (range 120-860). Outside lead exposure to
the scrub and circulating nurses were 4 and 2% of ICRP occupational limits respectively.

Our results suggested that a team of surgeons could perform 386 h of fluoroscopy per
year, or 587 endovascular aortoiliac aneurysm repairs per year, and remain within occu-
pational exposure limits. This does not take into account other endovascular procedures
performed by the surgeons, which would reduce these figures accordingly.

6. Additional Equipment to Help Reduce Exposure

Several available devices not directly related to the fluoroscopic unit are helpful in re- ■§

ducing the total exposure. A floating table simplifies positional changes and reduces the |

need to constantly adjust the fluoroscope. A power injector (Contrast injection system, a

Acist Medical Systems, Minneapolis, MN) ensures that an adequate volume of contrast is c

delivered, which maximizes image quality and reduces the need for multiple screening <

runs. This is especially important in imaging the thoracic or abdominal aorta and it >9

branches. A power injector may actually reduce the overall contrast required to perform J

the procedure by eliminating the need for multiple image acquisitions. An equally im- «

portant benefit is that a power injector allows the operator to increase his or her distance |

from the source. The same effect can be achieved by adding extension tubing to the @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

492 LIPSITZ et al.

catheter injection port during manual injection technique. The tabletop should be
maximally radiolucent. The equipment used (stent grafts, guidewires, catheters) should
be well marked with radio-opaque indicators that are easily visualized, such that one does
not have to strain or increase the image intensifier size to see them.

Noninvasive vascular imaging techniques such as duplex Doppler and intravascular
ultrasound contribute information that can aid in the performance of endovascular pro-
cedures and lessen the fluoroscopic time. Often these modalities are complementary; the
information gained from duplex, for example, can be used to limit the contrast load as well
as both fluoroscopy and procedural time. These modalities, however, may not always
provide the same anatomical detail as angiography. Finally, marking appropriate land-
marks on the screen with an erasable pen allows one to work under regular fluoroscopy
rather than road mapping, which may lead to increased exposure.

7. Summary

It is critical to remember that radiation exposure is cumulative and the effects permanent.
The major factors increasing exposure include increased fluoroscopy time and the
proximity of the surgeon to the operative field. The maximum allowable occupational
and civilian radiation exposure doses as defined by the various regulatory agencies have
been lowered with time. It is likely that with increasing knowledge concerning the effects of
radiation, this trend will continue. We recommend keeping exposure to less than 10-20%
of established occupational limits. All centers performing endovascular procedures should
actively monitor their effective doses and educate personnel in methods for reducing
exposure.

II. CONTRAST TOXICITY

A. Introduction

The recent increasing number and complexity of endovascular interventions being
performed means that patients are being exposed to radiological contrast agents in greater
numbers and with larger doses. There has been much improvement in the production of
contrast agents since the first use of an iodinated contrast agent 70 years ago. This has led
to a reduction in the overall toxicity of these agents (22). Adverse reactions to contrast
agents vary widely in their severity and clinical presentation, from minimal, transient
elevations in serum creatinine to major anaphylactic reactions.

B. Types of Reaction

1 . Systemic

Many systemic reactions begin with isolated endothelial, hematological, or cardiac toxicity ■§

and may progress to a systemic response. Although immunological responses occur, they |

are not felt to be traditional allergic reactions for two major reasons (23): (a) antibodies to a

contrast agents and additives have been identified in only a few cases and (b) repeat c

contrast administration leads to a response in less than 50% of cases. Systemic reactions to <

contrast agents may be mild, moderate, or severe. Mild reactions include pain, heat, >9

itching, and pallor. Moderate reactions include hypotension and wheezing. Severe re- J

actions include unresponsiveness, arrhythmias, and cardiac arrest. Many minor reactions «

to contrast agents are related to the osmolality of the agent. Specifically, pain is greatly |

reduced with the use of low-osmolality agents. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

RADIATION EXPOSURE AND CONTRAST TOXICITY 493

The occurrence of severe reactions is more difficult to predict. Several risk factors are
associated with the development of systemic reactions, including severe allergies, active
asthma, and cardiac disease. Shellfish allergies have traditionally been thought to be
closely associated with contrast sensitivity, but this has not been shown to be a stronger
risk factor than the presence of any other severe allergy (24). Additionally, although a
prior contrast reaction does increase the likelihood of a subsequent reaction, it does not
help to predict the severity of such a reaction (23).

The treatment of systemic contrast reactions is dictated by the nature and severity of
the episode. Vasovagal reactions can be treated with Trendelenburg positioning and fluids.
Respiratory reactions may resolve spontaneously but should be treated to prevent
progression to a more severe reaction. Adrenergic inhalers are the first line of treatment,
which can be followed by intramuscular or intravenous epinephrine if these local measures
fail. Cardiopulmonary reactions may require full cardiopulmonary resuscitation.

2. Hematological

There are a myriad of effects on the vessel wall and hematological system, all of which
appear to be insignificant when contrast agents are administered at the doses generally
required for routine imaging studies (23). For example, while early studies using high
doses of contrast with stagnant flow showed endothelial damage, these results were not
borne out in studies using more clinically relevant doses in the presence of blood flow (25-
28). Additionally, the incidence of venous thrombosis in the setting of high-volume, high-
concentration, slow infusions with high-osmolality nonionic agents has been reported to
be as high as 30%; however, with a reduction in contrast osmolality and alteration of
technique, this figure can be reduced sixfold (29-32). For arterial injections, these concerns
are not as relevant, since these agents are cleared rapidly by high flow and hence reduced
exposure time to the contrast. In fact, with ionic agents there seems to be a mild, transient
anticoagulant effect, which once again is not clinically significant (33). Any significant
thrombotic or anticoagulation effects of these agents are far outweighed by the technique
used, the state of the vessels being evaluated, and the presence of the catheters and
guidewires used in the procedure.

3. Cardiac

As is the case for hematological effects, cardiac toxicity is generally not significant at
clinically relevant doses; the osmolality of the contrast agent employed is the most
important factor determining toxicity. These effects are a more pronounced during the
performance of cardiac catheterization that during peripheral angiography. For example,
injection of contrast into the left ventricle leads to a fall in heart rate and blood pressure
with a slight increase in diastolic pressure, all of which return to baseline within 30 s to 3
min. Similar but far less notable effects are noted with the injection of contrast into the ■a

periphery. Although the results vary depending on the model used, studies with isolated |

heart perfusion show that almost all contrast agents suppress both myocardial and as

coronary artery contractility and that the effects increase with the osmolality of the c

contrast agent (23). Another possible cardiac effect is the development of electromechan- <

ical dissociation — which can mimic an anaphylactic response — due to a rapid drop in the >9

serum calcium concentration. This can occur during angiography, because many contrast J

agents contain calcium chelating agents that are used to stabilize the solution (34). Newer «

agents have a reduced amount of such compounds and therefore pose a reduced risk of |

this complication. @

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

494 LIPSITZ et al.

While perhaps not strictly considered a contrast reaction, fluid overload, especially in
patients with congestive heart failure, must be guarded against. Contrast agents can be a
significant volume load both in terms of pure volume as well as the osmotic activity of the
agents. This is yet another reason to minimize contrast load whenever possible. Thus
respiratory difficulty during an angiographic examination may represent a reaction to
contrast or simply fluid overload (35).

4. Nephrotoxicity

Contrast nephrotoxicity is the third leading cause of acute renal failure in hospitalized
patients and has been reported to occur in up to 6% of unselected patient populations and
up to 50% of high-risk populations receiving radiological contrast agents (36-39). Because
many of the patients undergoing these interventions fall into the high-risk category due to
the underlying nature of their disorder, the incidence of contrast nephrotoxicity is
significant (40).

Clinical Course, Pathogenesis, and Risk Factors. Although nearly all patients receiv-
ing radiographic contrast experience a small, transient decreases in glomerular filtration, the
exact definition of contrast nephropathy varies. It is generally defined as a rise in the serum
creatinine to >25% or >0.5 mg/dL above baseline in the absence of other inciting events
(41). The elevation in creatinine typically occurs within 24^48 h of contrast administration.
The peak increase occurs between 3 and 5 days, with a return to baseline by 7-10 days. In the
majority of cases the renal failure is nonoliguric and entirely reversible. The urinalysis may
be unremarkable but is usually consistent with acute tubular necrosis, showing coarse
granular casts and renal tubular epithelial cells. A small percentage of patients may present
with more severe renal failure, with symptoms manifesting within 24 h of contrast
administration. Less than 1% of patients with contrast nephropathy will require dialysis,
but in this setting the mortality is approximately 30% (42). Lactic acidosis resulting from
contrast nephropathy in diabetic patients who take the oral hypoglycemic medication
metformin is a rare complication (43). It has been recommended by some that patients
discontinue this medication 24 h prior to any contrast study, while others have suggested
that patients not restart the medication until 48 h following the procedure, and then only if
there is no evidence of nephrotoxicity (41).

Contrast agents are excreted by glomerular filtration alone without undergoing tubular
resorption or secretion. Contrast nephrotoxicity occurs due a combination of alterations in
renal hemodynamics and direct tubular epithelial cell toxicity. There is a biphasic response
to the administration of contrast in the renal vasculature, with an initial brief period
of vasodilatation followed by a more prolonged but variable period of vasoconstriction.
These changes are most likely mediated by alterations in nitric oxide, endothelin,
adenosine, or prostaglandin metabolism (41). Direct cellular toxicity is inferred from in
vitro and in vivo studies showing increased excretion of enzymes and low-molecular-weight ■§

proteins in the urine in addition to pathological changes seen on histology (44). |

The most important risk factor for the development of contrast nephrotoxicity is the a

presence of preexisting renal insufficiency (45,46). Patients with a creatinine level of >1.5 c

mg/dL have up to a 21 -fold increase risk of contrast nephropathy compared to patients <

with a normal creatinine level. There are several other important factors that may increase >9

the risk for the development of contrast nephropathy including dehydration, diabetes J

mellitus, contrast volume, congestive heart failure, and concurrent exposure to other ne- «

phrotoxic agents (Table 2). Many of these factors act synergistically. For example, diabetic |

patients without preexisting renal insufficiency do not seem to be at increased risk of @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

RADIATION EXPOSURE AND CONTRAST TOXICITY 495

Table 2 Risk Factors for Contrast Nephropathy

Preexisting renal insufficiency

Volume of contrast

Intravascular volume depletion

Diabetes mellitus

Concurrent administration of other nephrotoxic agents

Repeat contrast procedures

Congestive heart failure

Abnormal liver function

Multiple myeloma

Nephrosis

Nonsteroidal anti-inflammatory drugs

Angiotensin-converting enzyme inhibitors

contrast nephropathy, while those with renal insufficiency are at higher risk than in the
presence of renal insufficiency alone (45,46). One study of patients undergoing cardiac
catheterization found that the risk of renal failure ranged from 1 to 100% as the number of
risk factors increased (47).

Prevention. The principles in the prevention of contrast nephrotoxicity include
selection of the appropriate diagnostic and/or therapeutic modality, preprocedural
correction of risk factors, ensuring adequate hydration, eliminating or reducing any
additional nephrotoxic agents, limiting the amount of contrast administered, and close
follow-up of serum creatinine postprocedurally.

All patients undergoing contrast procedures should receive adequate hydration either
orally or intravenously at the time of the procedure, such that positive fluid balance and
high urine output are achieved. Patients with preexisting renal insufficiency should be
hydrated well before and after these procedures. Patients with cardiopulmonary dysfunc-
tion must be monitored closely while receiving hydration. There are many different pro-
tocols for hydration, but in general the use of 0.45% saline at 1—1.5 mL/Kg/h beginning
1-2 h prior to the procedure and continuing for up to 6 h (outpatients) or 24 h (inpatients)
are acceptable. The urine volumes should be approximately 0.75-1.25 mL/h (40,41,48-50).

The use of low-osmolality nonionic contrast media has also been advocated for the
prevention of contrast nephropathy. The impetus for the development of these agents was
the high incidence of side effects associated with the earlier agents. However, their high
cost raised concern for their routine use. Additionally, large studies have shown a signi-
ficant risk reduction only in patients with preexisting renal dysfunction (51,52). Because of
these factors, low-osmolality nonionic contrast agents may best be reserved for use in •g

patients with underlying renal dysfunction, especially diabetics and patients with potential s

hemodynamic instability. a

Diuretics have been proposed as a method to reduce contrast nephropathy based on c

the theoretical advantage that loop diuretics (e.g., furosemide) might decrease medullary <

oxygen consumption and hence the potential for ischemic injury during contrast studies J3

(53,54). Mannitol, alone or in conjunction with loop diuretics, has also been proposed to J

reduce contrast nephropathy. In one large study, patients with creatinine >1.6 mg/dL Q

undergoing cardiac catheterization were randomized to three groups. One group received |

hydration alone, one received hydration plus mannitol, and one received hydration plus «

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

496 LIPSITZ et al.

furosemide. A significant increase in creatinine was noted in 11% of the patients who
received hydration alone, 28% of the patients receiving hydration plus mannitol, and 40%
of the patients who received hydration plus furosemide (50). In another study, a similar
cohort of patients was treated with hydration alone versus a regimen of hydration,
mannitol, furosemide, and renal-dose dopamine. In this study, urine output was replaced
with saline to control for dehydration effects. There was no significant difference between
the groups in terms of increase in creatinine; however, there was a higher incidence of
contrast nephropathy in patients with a lower (<150 mL/h) as opposed to a higher urine
output, suggesting some protective effect of maintaining a good urine output (55). One
additional study did find that the risk of contrast nephropathy was reduced by mannitol in
nondiabetic azotemic patients (56). Presently there is no evidence to suggest that the use of
either mannitol or diuretics reduces the risks of contrast nephropathy; they may, in fact,
increase the risk by their tendency to produce negative fluid balance and dehydration.

Based on studies suggesting that reactive oxygen species may play a major role in the
pathogenesis of contrast nephropathy, the antioxidant V-acetylcysteine has been eval-
uated for a potential protective effect. Two recent studies, one in patients undergoing
contrast computed tomography scanning and one in patients undergoing cardiac catheter-
ization, both showed some reduction of contrast nephropathy (57,58).

Another approach to the prevention of contrast nephropathy involves the use of
agents that selectively increase renal blood flow. Low-dose dopamine is one such agent.
Despite its theoretical advantages, studies using low-dose dopamine have failed to show a
clear-cut benefit in terms of preventing contrast nephropathy or in providing a protective
effect once contrast nephropathy has occurred (56,59). In one of these studies, diabetic
patients actually had a higher rate of contrast nephropathy with dopamine, while non-
diabetic patients had a lower rate (56). These results plus other studies suggest that dopa-
mine may be of some benefit in preventing contrast nephropathy in nondiabetic patients,
although its routine use has not been advocated (60,61). It is not recommended for
prevention of contrast nephropathy in diabetic patients (40,41).

Fenoldopam is a selective dopamine- 1 (DA-1) receptor agonist approved for the
treatment of systemic hypertension. It is given parenterally but has no stimulatory effect
on dopamine-2 (DA-2) or adrenergic receptors (both cause vasoconstriction), as does
dopamine at higher doses (40). Fenoldopam has been found to increase renal blood flow
and improve glomerular filtration rate while reducing both systolic and diastolic blood
pressure in hypertensive patients (62-66). Blood pressure reductions are only mild in nor-
motensive patients. Fenoldopam also prevents shunting of blood flow from the medulla to
the cortex, permitting maintenance of medullary oxygenation. Several recent studies have
shown significant reductions in the incidence of contrast nephropathy with the use of
fenoldopam (67-69).

Various other agents have been used in an attempt to reduce the incidence of contrast ■a

nephropathy. Atrial natriuretic peptide is another agent that increases renal blood flow |

and was found to reduce contrast nephropathy in animal studies. However, in a recent as

randomized, double-blind, placebo-controlled study, atrial natriuretic peptide was not c

shown to decrease the incidence of contrast nephropathy, even across subgroup analysis <

for diabetic patients (70). Theophylline was hypothesized to have a role in the reduction of >9

contrast nephropathy, since it is an adenosine antagonist. Adenosine is thought to J

contribute to the renal vasoconstriction seen in contrast nephropathy. Some studies have «

shown benefit with the use of theophylline, while others have not. In the absence of pros- |

pective trails, theophylline has not been recommended. The use of calcium channel @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

RADIATION EXPOSURE AND CONTRAST TOXICITY 497

blockers, which have been demonstrated to attenuate the reductions in renal blood flow
associated with contrast administration in laboratory studies, have similarly not been
shown to have a significant effect in clinical studies. Endothelin antagonists have also not
shown benefit in clinical studies.

C. Summary

As is the case with any medication and/or imaging modality, the best way to reduce
toxicity is to use the lowest possible volume and concentration of the agent while still
obtaining an adequate diagnostic or therapeutic result. Many different agents are under-
going evaluation for their role in reducing contrast nephropathy, but more long-term
studies are needed to clarify their efficacy. At the present time, maintenance of adequate
hydration is the most important factor in preventing contrast nephropathy.

REFERENCES

1. Veith FJ. Presidential address: Charles Darwin and vascular surgery. J Vase Surg 1997;
25(1):8-18.

2. The International Committee on Radiological Protection. Radiological protection and safety in
medicine. A report of the International Commission on Radiological Protection. Ann ICRP
1996; 26(2):1^17.

3. National Council on Radiation Protection and Measurements. Ionizing Radiation Exposure of
the Population of the United States. NCRP Report No. 93, 1987. Bethesda, MD: NCRP, 1987.

4. Hough DM, Brady A, Stevenson GW. Audible radiation monitors: The value in reducing
radiation exposure to fluoroscopy personnel. AJR 1993; 160:407^408.

5. Boone JM, Levin DC. Radiation exposure to angiographers under different fluroscopic
imaging conditions. Radiology 1991; 180:861-865.

6. Mehlman CT, DiPasquale TG. Radiation exposure to the orthopaedic surgical team during
fluoroscopy: "How far away is far enough?" J Orthop Trauma 1997; 1 1(6):392 398.

7. Lindsay BD, Eichling JO, Ambos HD, Cain ME. Radiation exposure to patients and medical
personnel during radio frequency catheter ablation for supraventricular tachycardia. Am J
Cardiol 1992; 70:218-223.

8. Kuwayama N, Takaku A, Endo S, Nishljima M, Kamei T. Radiation exposure in endovascular
surgery of the head and neck. Am J Neuroradiol 1994; 15:1801-1808.

9. Ramalanjaona GR, Pearce WH, Ritenour ER. Radiation exposure risk to the surgeon during
operative angiography. J Vase Surg 1986; 4(3):224-228.

10. Pratt TA, Shaw AJ. Factors affecting the radiation dose to the lens of the eye during cardiac -a
catheterization procedures. Br J Radiol 1993; 66:346-350. %

11. Aldridge HE, Chisholm RJ, Dragatakis L, Roy L. Radiation safety in the cardiac cathe- £
terization laboratory. Can J Cardiol 1997; 13(5):459^67. t

12. Heyd RL, Kopecky KK, Sherman S, Lehman GA, Stockberger SM. Radiation exposure to s
patients and personnel during interventional ERCP at a teaching institution. Gastrointest u
Endosc 1996; 44:287-292. g

13. Watson LE, Riggs MW, Bourland PD. Radiation exposure during cardiology fellowship a
training. Health Phys 1997; 73(4):690-693. °

14. Granger WE, Bednarek DR, Rudin S. Primary beam exposure outside the fluoroscopic field of |
view. Med Phys 1997; 24(5):703-707. g

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

498 LIPSITZ et al.

15. Coakley KS, Ratcliffe J, Masel J. Measurement of radiation dose received by the hands and
thyroid of staff performing gridless fluoroscopic procedures in children. Br J Radiol 1997; 70:
933-936.

16. Kicken PJ, Bos AJJ. Effectiveness of lead aprons in vascular radiology: Results of clinical
measurements. Radiology 1995; 197:473^478.

17. Cousin AJ, Lawdahl RB, Chakraborty DP, Koehler RE. The case for radioprotective eyewear/
facewear: Practical implications and suggestions. Invest Radiol 1987; 22:688-692.

18. Lo NN, Goh SS, Khong KS. Radiation dosage from use of the image intensifier in orthopaedic
surgery. Singapore Med J 1996; 37:69-71.

19. Marx MV, Niklason L, Mauger EA. Occupational radiation exposure to interventional radio-
logists: A prospective study. J Vase Intervent Radiol 1992; 3:597-606.

20. Vehmas T, Tikkanen H. Measuring radiation exposure during percutaneous drainages: Can
shoulder dosimeters be used to estimate finger doses? Br J Radiol 1992; 65(779): 1007-1010.

21. Lipsitz EC, Veith FJ, Ohki T, et al. Does the endovascular repair of aortoiliac aneurysms pose
a radiation safety hazard to vascular surgeons? J Vase Surg 2000; 32:704-710.

22. Osborne ED, Sutherland CG, Scholl AJ, Rowntree LD. Roentgenography of the urinary tract
during excretion of sodium iodine. JAMA 1923; 80:368-373.

23. Bettmann MA. Physiologic effects and systemic reactions. In: Baum S, ed. Abrams' Angio-
graphy. 4th ed. Philadelphia: Lippincott Williams & Wilkins, 1996:22-33.

24. Lasser EC, Berry CC, Talner LB, et al. Pre-treatment with corticosteroids to alleviate reactions
to intravenous contrast material. New Engl J Med 1987; 317:845-849.

25. Merseredu WA, Robertson HR. Observations on venous endothelial injury following the
injection of various radiographic contrast media in the rat. J Neurosurg 1961; 18:289-294.

26. Laerum F. Cytotoxic effects of six angiographic contrast media on human endothelium in
culture. Acta Radiol 1987; 28:99-105.

27. Morgan DML, Bettmann MA. Effects of x-ray contrast media and radiation on human vas-
cular endothelial cells in vitro. Cardiovasc Intervent Radiol 1989; 12:154-160.

28. Thiesea B, Muetzer W. Effects of angiographic contrast media on venous endothelium of rabbits.
Invest Radiol 1990; 25:121-126.

29. Albrechtsson U, Olsson C-G. Thrombotic side effects of lower limb phlebography. Lancet
1976; 1:723-724.

30. Laeram F, Holm HA. Postphlebographic thrombosis: A double blind study with methylgluc-
amine metrizoate and metrizamide. Radiology 1981; 140:651-654.

31. Bettmann MA, Salzman EW, Rosenthal D, et al. Reduction of venous thrombosis
complicating phlebography. AJR 1980; 134:1169-1172.

32. Bettmann MA, Robbins A, Braun SD, Wetzner S. Dunnick NR, Finkelstein J. Comparison of
the diagnostic efficacy, tolerance and complication rates of a nonionic and an ionic contrast
agent for leg phlebography. Radiology 1987; 165:113-116.

33. Mamoa JF, Hoppensteadt MS, Fareed J, Moncada R. Biochemical evidence for a relative lack
of inhibition of thrombin formation by nonionic contrast media. Radiology 1991; 179:399-
402.

34. Bourdillon PD, Bettmann MA, McCracken S, Poole-Wilson PA, Grossman W. Effects of a

new ionic and a conventional ionic contrast agent on coronary sinus ionized calcium and left j>

ventricular hemodynamics in dogs. J Am Coll Cardiol 1985; 6:845-853. <S

35. Rosen RJ. Angiography: General techniques, complications, and risk management. In: Taveras js
JM, et al., eds. Radiology. Vol. 2. Philadelphia: Lippincott Williams & Wilkins, 2002: 2-4. Jf

36. Hou SH, Bushinsky DA, Wish JB, Cohen JJ, Harrington JT. Hospital-acquired renal ^
insufficiency: A prospective study. Am J Med 1983; 74:243-248. «

37. Davidson CJ, Hlatky M, Morris KG, et al. Cardiovascular and renal toxicity of a nonionic ^
radiographic contrast agent after cardiac catheterization: A prospective trial. Ann Intern Med q
1989; 110:119-124. |

38. Lautin EM, Freeman NJ, Schoenfeld AH, et al. Radiocontrast-associated renal dysfunction: 2
Incidence and risk factors. AJR 1991; 157:49-58. 2

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

RADIATION EXPOSURE AND CONTRAST TOXICITY 499

39. Weisberg LS, Kurnik PB, Kurnik BRC. Risk of radiocontrast nephropathy in patients with
and without diabetes mellitus. Kidney Int 1994; 45:259-265.

40. Waybill MM, Waybill PN. Contrast media-induced nephrotoxicity: Identification of patients at
risk and algorithms for prevention. J Vase Intervent Radiol 2001; 12(l):3-9.

41. Murphy SW, Barrett BJ, Parfrey PS. Contrast nephropathy. J Am Soc Nephrol 2000; 11(1):
177-182.

42. McCullough PA, Wolyn R, Rocher LL, Levin RN, O'Neill WW. Acute renal failure after
coronary intervention: Incidence, risk factors, and relationship to mortality. Am J Med 1997;
103:368-375.

43. Thomsen HS, Morcos SK. Contrast media and metformin: Guidelines to diminish the risk of
lactic acidosis in non-insulin-dependent diabetics after administration of contrast media. ESUR
Contrast Media Safety Committee. Eur Radiol 1999; 9:738-740.

44. Humes HD, Hunt DA, White MD. Direct toxic effect of the radiocontrast agent diatrizoate on
renal proximal tubule cells. Am J Physiol 1987; 252:F246-F255.

45. Parfrey PS, Griffiths SM, Barrett BJ, Paul MD, Genge M, Withers J, Farid N, McManamon
PJ. Contrast-material induced renal failure in patients with diabetes mellitus, renal in-
sufficiency, or both: A prospective controlled study. N Engl J Med 1989; 320:143-149.

46. Rudnick MR. Berns JS, Cohen RM, Goldfarb S. Contrast-media associated nephrotoxicity.
Semin Nephrol 1997; 17:15-26.

47. Rich MW, Crecelius CA. Incidence, risk factors, and clinical course of acute renal insufficiency
after cardiac catheterization in patients 70 years of age or older. Arch Intern Med 1995; 150:
1237-1242.

48. Eisenberg RL, Bank WO, Hedgock MW. Renal failure after major angiography can be avoided
with hydration. AJR 1981; 136:859-861.

49. Taylor AJ, Hotchkiss D, Morse RW. McCabe J. PREPARED: Preparation for Angiography
in Renal Dysfunction. A randomized trial of inpatient vs outpatient hydration protocols for
cardiac catheterization in mild-to-moderate renal dysfunction. Chest 1998; 114:1570-4574.

50. Solomon RC, Werner C, Mann D, D'Elia J, Silva P. Effects of saline, mannitol, and furosemide
on acute decreases in renal function induced by radiocontrast agents. N Engl J Med 1994;
331:1416-1420.

5 1 . Barrett BJ, Carlisle EJ. Meta-analysis of the relative nephrotoxicity of high- and low-osmolality
iodinated contrast media. Radiology 1993; 188:171-178.

52. Rudnick MR, Goldfarb S, Wexler L, Ludbrook PA, Murphy MJ, Halpern EF, Hill JA,
Winniford M, Cohen MB, VanFossen DB. Nephrotoxicity of ionic and nonionic contrast in
1196 patients: A randomized trial. Kidney Int 1995; 47:254-261.

53. Heyman SN, Brezis M, Epstein FH, Spokes K, Silva P, Rosen S. Early renal medullary hypoxic
injury from radiocontrast and indomethacin. Kidney Int 1991; 40:632-642.

54. Barrett BJ. Contrast nephrotoxicity. J Am Soc Nephrol 1994; 5:125-137.

55. Stevens MA, McCullough PA, Tobin KJ, Speck JP, Westveer DC, Guido-Allen DA, Timmis
GC, O'Neill WW. A prospective randomised trial of prevention measures in patients at high
risk for contrast nephropathy. J Am Coll Cardiol 1999; 33:403-411.

56. Weisberg LS, Kurnik PB, Kurnik BRC. Dopamine and renal blood flow in radiocontrast-

induced nephropathy in humans. Renal Failure 1993; 15:61-68. g

57. Tepel M, Van Der Giet M, Schwarzfeld C, Laufer U, Liermann D, Zidek W. Prevention of <S
radiographic-contrast-agent-induced reductions in renal function by acetylcysteine. New Engl J js
Med 2000; 343:180-184. Jf

58. Diaz-Sandoval LJ, Kosowsky BD, Losordo DW. Acetylcysteine to prevent angiography- ^
related renal tissue injury (the APART trial). Am J Cardiol 2002; 89:356-358; 2000; 343:180- 4
184. |

59. Abizaid AS, Clark CE, Mintz GS, Dosa S, Popma JJ, Pichard AD, Satler LF, Harvey M, Kent q
KM, Leon MB. Effects of dopamine and aminophylline on contrast-induced acute renal failure |
after coronary angioplasty in patients with preexisting renal insufficiency. Am J Cardiol 1999; 2
83:260-263. 2

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

500 LIPSITZ et al.

60. Hans SS, Hans BA, Dhillon R, Dmuchowski C, Glover J. Effect of dopamine on renal function
after arteriography in patients with pre-existing renal insufficiency. Am Surg 1998; 64:432-436.

61. Hall KA, Wong RW, Hunter GC, Camazine BM, Rappaport WA, Smyth SH, Bull DA,
Mclntyre KE, Bernhard VM, Misiorowski RL. Contrast-induced nephrotoxicity: The effects
of vasodilator therapy. J Surg Res 1992; 53:317-320.

62. Barkis GL, Lass NA, Glock D. Renal hemodynamics in radiocontrast medium-induced renal
dysfunction: A role for dopamine-1 receptors. Kidney Int 1999; 56:206-210.

63. Panacek EA, Bednarczyk EM, Dunbar LM, Foulke GE, Holsclaw TL. Randomized, pro-
spective trial of fenoldopam vs. sodium nitroprusside in the treatment of acute severe hyper-
tension. Acad Emerg Med 1995; 2:959-965.

64. Post JB, Frishman WH. Fenoldopam: A new dopamine agonist for the treatment of hyper-
tensive urgencies and emergencies. J Clin Pharmacol 1998; 38:2-13.

65. Singer I, Epstein M. Potential of dopamine A-l agonists in the management of acute renal
failure. Am J Kidney Dis 1998; 31:743-755.

66. Mathur VS, Ellis D, Fellmann J, Luther RR. Therapeutics for hypertensive urgencies and
emergencies-fenoldopam: A novel systemic and renal vasodilator. Cardiovasc Intervent Radiol
1998; 1:43-53.

67. Madyoon H. Clinical experience with the use of fenoldopam for prevention of radiocontrast
nephropathy in high-risk patients. Rev Cardiovasc Med 2001; 2(suppl 1):S26-S30.

68. Tumlin JA, Wang A, Murray PT, Mathur VS. Fenoldopam mesylate blocks reductions in renal
plasma flow after radiocontrast dye infusion: A pilot trial in the prevention of contrast ne-
phropathy. Am Heart J 2002; 143:894-903.

69. Kini AS, Mitre CA, Kim M, Kamran M, Reich D, Sharma SK. A protocol for prevention of
radiologic contrast nephropathy during percutaneous coronary intervention: Effect of selective
dopamine receptor agonist fenoldopam. Cathet Cardiovasc Intervent 2002; 55:169-173.

70. Kurnik BRC, Allgren RL, Genter FC, Solomon RJ, Bates ER, Weisberg LS. Prospective study
of atrial natriuretic peptide for the prevention of radiocontrast-induced nephropathy. Am J
Kidney Dis 1998; 31:674-680.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

30

Complications in Peripheral Thrombolysis

Kenneth Ouriel

The Cleveland Clinic Foundation, Cleveland, Ohio, U.S.A.

The principal goal of thrombolytic therapy is to dissolve intravascular thrombus. Throm-
bolytic agents are remarkably effective in accomplishing this goal, and they can do so in
a minimally invasive fashion. Through percutaneous means alone, recanalization of an
occluded bypass graft or artery can be achieved within a few hours and the thrombus can
be completely dissolved over the course of 12-48 h (1). Culprit stenotic lesions become
readily apparent after successful thrombolytic therapy; these lesions must be addressed
to diminish the risk of reocclusion (2). The unmasked lesions can frequently be repaired
percutaneously. Even when open surgical revascularization is necessary, it can often be
performed on an elective basis, allowing ample time for patient preparation. Thrombo-
lytic therapy can be employed to clear thrombus from small vessels that are inaccessi-
ble to standard balloon catheter thrombectomy, sometimes identifying target vessels for
an operative bypass procedure (Fig. 1) (3). Clinical trials have proven thrombolytic ther-
apy to be effective for the treatment of acute arterial occlusion, resulting in a reduction
in mortality (4), limb loss (5), length of hospital stay (6), and the need for open surgical
intervention (7).

But successful thrombolytic therapy cannot be separated from its major complication,
the inextricable risk of life-threatening hemorrhage. In fact, other complications pale in
comparison to bleeding, and the association between thrombolytic therapy and hemor- ■§

rhage is the sole factor that has precluded the use of thrombolysis to an even broader |

number of patients (8). To understand the association between thrombolytic therapy and as

hemorrhage, it is beneficial to first consider the biochemistry underlying the mechanism of c

action of thrombolytic agents. The fundamental physiology of dissolving (a) pathological <

intravascular thrombi and (b) desirable plugs sealing remote sites where vascular integrity >9

has been lost is virtually identical. Herein lies the failure of thrombolytic therapy: how J

does one effect dissolution of "bad" thrombus without causing distant hemorrhage due to °

the dissolution of "good" thrombus sealing small vascular defects in the gastrointestinal |

tract, at the catheter insertion site, or, most importantly, within the calvarium? @

501

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

502

OURIEL

Figure 1 Thrombolytic therapy can clear thrombus from small vessels inaccessible to balloon
catheter thrombectomy. A. This patient had no identifiable vessels below the popliteal, even on delayed
views. B. Following thrombolysis, a lateral plantar artery was identified in this patient. C. A saphenous
vein popliteal-plantar artery bypass was now possible.

I. BIOCHEMISTRY OF THROMBOLYTIC AGENTS

The sum and substance of intravascular thrombus is fibrin, a polymerized meshwork of
cross-linked protein interspersed with abundant platelets and occasional red and white
blood cells. The pharmacologic goal of thrombolytic therapy is to digest this meshwork,
breaking the bonds on which the integrity of the thrombus depends. In this manner, the
thrombolytic agents convert the highly insoluble fibrin-platelet mass into microscopic frag-
ments of degenerated thrombus and, preferably, soluble breakdown products that wash
into the circulation and are cleared by the kidney and liver.

It is important to emphasize that all clinically available thrombolytic agents do not
degrade thrombus directly. Rather, each is a "plasminogen activator." As such, they do
not directly degrade fibrinogen, and without plasminogen they are inert. The thrombolytic
agents comprise trypsin-like serine proteases that have high specific activity directed at the
cleavage of a single peptide bond in the plasminogen zymogen, converting it to plasmin
(9). Importantly, plasmin is the active molecule that cleaves fibrin polymer to cause the
dissolution of thrombus. Milstone first recognized the importance of plasminogen in 1941,
when it was noted that clots formed with highly purified fibrinogen and thrombin were not
lysed by streptococcal fibrinolysin unless a small amount of human serum (plasminogen)

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS IN PERIPHERAL THROMBOLYSIS

503

was added (10). Recognizing this direct role of plasminogen, early investigators attempted
to dissolve occluding thrombi with the administration of exogenous plasmin (11). Free
plasmin administered intravenously is ineffective as a thrombolytic agent, accounting for
the failure of these attempts. Effective thrombolysis was achieved only when fibrin-bound
plasminogen was converted to active plasmin at the site of the thrombus (9).

The dependence of fibrinolysis on adequate circulating levels of plasminogen is best
illustrated by studies of the fibrinolytic potential of blood drawn from patients receiving
intravenous of thrombolytic agents after acute myocardial infarction (12). Blood obtained
soon after the start of the administration of these agents displayed a great degree of in
vitro fibrinolytic potential. Aliquots of plasma drawn from the patients and then added to
radiolabeled clots in test tubes produced rapid dissolution of the clots. By contrast, similar
aliquots drawn from patients after 20 min of thrombolysis had considerably less throm-
bolytic potential. The explanation for this observation relates to the amount of plasmi-
nogen present in the blood. Prolonged thrombolysis consumed all of the endogenous
plasminogen and, despite continued administration of the thrombolytic agent, no further
clot lysis was possible.

II. SAFETY OF THROMBOLYTIC AGENTS IN CLINICAL TRIALS

There are few well-designed clinical comparisons of different thrombolytic agents for the
treatment of peripheral arterial occlusion. By contrast, the literature is replete with a broad
spectrum of in vitro studies and retrospective clinical trials, most pointing to improved
efficacy and safety of urokinase and alteplase over streptokinase (13-16). In an analysis of
data collected in a prospective, single-institution registry at the Cleveland Clinic Founda-
tion, urokinase demonstrated a diminished rate of bleeding complications compared with
alteplase (Table 1) (17).

There have been two prospective, randomized comparisons of urokinase and alte-
plase. Neither was blinded. Meyerovitz and associates from the Brigham and Women's
Hospital randomized 32 patients with peripheral arterial or bypass graft occlusions of less
than 90 days duration to alteplase (10 mg bolus, 5 mg/h to a maximum of 24 h) or uro-
kinase (60,000 IU bolus, 4,000 IU/min for 2 h, 2000 IU/min for 2 h, then 1000 IU/min to a
maximum of 24 h total administration) (18). There was significantly greater systemic fibrin-
ogen degradation in the alteplase group (p = 0.01), indicating that the fibrin specificity of

Table 1 Relative Rate of Complications in 627 Patients Treated with Urokinase or Alteplase
Between 1990 and 1998 at the Cleveland Clinic Foundation

Urokinase

Alteplase

Event

N = 483

TV = 144

p Value

Bleeding requiring transfusion

12%

23%

0.004

Hematoma at catheter insertion site

22%

44%

<0.001

False aneurysm

1.7%

2.8%

Not significant

Intracranial bleeding

0.6%

2.8%

0.03

Death

2.7%

4.2%

Not significant

I

•5

Source: Ref. 17.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

504 OURIEL

Table 2 Complications Associated with Three Escalating Doses of Prourokinase (Initiated
at 2, 4, or 8 mg/h) Versus Urokinase (Initiated at 240,000 IU/h) in the PURPOSE Trial

Site

2.0 mg

4.0 mg

8.0 mg

Urokinase

Epistaxis

0.0%

0.0%

3.8%

0.0%

Gastrointestinal

0.0%

0.0%

1.9%

0.0%

Gums

0.0%

0.0%

9.6%

0.0%

Intracranial

0.0%

0.0%

0.0%

0.0%

Hematuria

6.6%

7.3%

9.6%

8.3%

Retroperitoneum

1.6%

0.0%

3.8%

1.7%

Source: Ref. 22.

alteplase was lost at this dosing regimen. Alteplase patients achieved more rapid initial
thrombolysis, but efficacy was identical in the two groups by 24 h. The trade-off to more
rapid thrombolysis was a trend toward a higher rate of bleeding complications in the
alteplase-treated patients.

The second randomized comparison of urokinase and alteplase was the STILE trial, a
three-armed multicenter comparison of urokinase (250,000 IU bolus, 4000 IU/min for 4 h,
then 2000 IU/min for up to 36 h), alteplase (0.05 to 0.1 mg/kg/h for up to 12 h) and primary
operation (19). There was one intracranial hemorrhage in the urokinase group (0.9%) and
two in the alteplase group (1.5%, no significant difference). Although actual rates of overall
bleeding complications and efficacy were not reported for the two thrombolysis groups, the
authors remarked that there were no significant differences detected in any of the outcome
variables. In a subsequent "reanalysis" of the STILE data, reported in 1999, the frequency
of complete clot lysis was similar with urokinase and alteplase at the time of the early
arteriographic study (A.J. Comerota, personal communication). These recent data suggests
that the rate of thrombolysis may be quite similar, in direct opposition to the popularly held
view that alteplase is a much more rapidly acting agent.

A multicenter, blinded trial compared the results of thrombolysis with urokinase
versus recombinant urokinase in 300 patients with peripheral arterial occlusion (20). These
data were never published. No significant differences were noted between the two agents.
A North American multicenter trial compared three different doses of prourokinase to
urokinase in 241 patients with lower extremity arterial occlusions of less than 14 days
duration (21). While the higher prourokinase dose was associated with slightly greater
percentage of patients with complete (>95%) clot lysis at 8 h, there was a mild increase in
the rate of bleeding complications compared with either the urokinase or the lower-dose
prourokinase groups. The fibrinogen levels fell in the higher prourokinase group, suggest-
ing that fibrin specificity was lost at the higher dose regimens for this compound (Table 2).

III. PREVENTING HEMORRHAGE DURING THROMBOLYSIS:
POTENTIAL FOR THE FUTURE

There are several methods to consider in the quest to diminish the rate of distant hemor-
rhagic complications with thrombolytic therapy. First, one might strive to limit the "leak"
of active thrombolytic agent into the systemic circulation. In fact this was the primary
reason for considering the local, catheter-directed route for low-dose thrombolysis in the

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I
•5

I

COMPLICATIONS IN PERIPHERAL THROMBOLYSIS

505

1970s and 1980s (22,23). Clearly, the intrathrombus administration of thrombolytic agents
is associated with improved efficacy (24). Whether increased safety has been achieved with
this now ubiquitous method, however, remains to be established (Fig. 2).

The use of "fibrin-specific" agents was investigated to decrease the risk of distant
bleeding (25). Fibrin specificity is quantified by the ratio of fibrin breakdown to fibrino-
gen breakdown. A fibrin-specific agent degrades fibrin but not fibrinogen. This property is
accomplished through a variety of mechanisms. Alteplase was the first "designer" throm-
bolytic agent with fibrin-specific properties (26). Alteplase binds to fibrin but not to fibrin-
ogen and attains plasminogenolytic activity only after binding. Tenectaplase, a modified
form of alteplase, has even greater fibrin specificity, with very little reduction in fibrino-
gen concentration during therapy (27). Nevertheless, fibrin specificity has not resulted
in a reduction in the risk of bleeding. The agents, in fact, may be associated with a greater

I
•5

Figure 2 Thrombolysis-induced bleeding from the suture holes in a recently constructed poly-
tetrafluoroethylene femoropopliteal bypass graft.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

506 OURIEL

risk of distant hemorrhage, possibly related to the production of an intermediary fibrin
breakdown product, fragment X (28). Fragment X is readily incorporated into existing
distant thrombi, rendering the thrombi fragile and more susceptible to subsequent throm-
bolytic dissolution and hemorrhage. Interestingly, urokinase and streptokinase do not
appear to generate significant quantities of fragment X during thrombolysis due to rapid
conversion to smaller breakdown products. This mechanism might explain the putative
lower incidence of distant bleeding with streptokinase and urokinase versus alteplase (29).

The use of concomitant systemic heparinization has been considered to reduce the
rate of bleeding complications during thrombolytic therapy (30). The TOPAS trial was
initially begun using full-dose systemic heparinization (7). The rate of intracranial bleed-
ing decreased substantially when therapeutic heparinization was deleted from the proto-
col. Similar results were observed when heparin was decreased in the prourokinase trials
for stroke (31). Recently, during peripheral arterial thrombolysis, a trend has emerged
restricting heparin administration to an infusion rate of several hundred units per hour or
less (32). Further, experience from the cardiology literature suggests that bleeding compli-
cations bear a greater relationship to heparin anticoagulation than to pharmacological
thrombolysis (29).

The use of mechanical thrombectomy devices as adjuvants to pharmacological throm-
bolysis holds the potential to lower the rate of bleeding complications (33-35). Recogniz-
ing the link between duration of administration and hemorrhage, bleeding complications
might be lower if mechanical adjuvants diminish thrombolytic exposure (36). As well, a
new mechanical thrombectomy device utilizes proximal and distal balloon occlusion to
limit the distribution of thrombolytic agent to the thrombosed vascular segment. Whether
these mechanical adjuvants will result in a significant reduction in the risk of bleeding
remains unproved; a critical analysis is not possible without a prospective trial.

Novel agents may diminish the risk of distant bleeding. For example, the agent alfime-
prase, an analogue of fibrolase, is a direct fibrinolytic with activity that is independent
of plasminogen (37). Circulating alpha2 macroglobulin inactivates alfimeprase. Systemic
fibrinolysis will not occur as long as the dose of alfimeprase is kept below a threshold
determined by available alpha2 macroglobulin. Like alfimeprase, plasmin analogues hold
promise as alternative thrombolytic agents with a decreased risk of systemic bleeding.
Plasmin, while ineffective when administered systemically, retains its activity when ad-
ministered directly into the thrombus (38). As well, plasmin analogues may diminish the
rate of distant bleeding as a result of systemic inactivation by circulating alpha2 anti-
plasmin (Fig. 3).

IV. NONHEMORRHAGIC COMPLICATIONS OF THROMBOLYTIC

AGENTS |

There are a wide variety of nonhemorrhagic complications associated with thrombolysis as

(39). Streptokinase is associated with allergic reactions in a significant percentage of c

patients, especially in those who have recently experienced a streptococcal infection and in <

those who undergo retreatment with the agent. Urokinase may be accompanied with >9

rigors, especially in patients for whom the agent is administered in a large bolus dose. The J

target vessel or bypass graft may rethrombosis during thrombolytic infusion, a compli- «

cation that occurs in approximately 8% of patients and seems to be reduced with heparin |

anticoagulation (21,30). The patient may sustain distal embolization of partially dissolved @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS IN PERIPHERAL THROMBOLYSIS

507

□ Thrombolysis
D Surgery

Figure 3 Complications in the STILE trial (19). Major morbidity was diminished in the surgery
group, a finding that was related to a greater frequency of "ongoing ischemia" in the thrombolysis
group. This did not translate into differences in the more important endpoints of death and major
amputation.

thrombus, a problem that is noted in 10-20% of patients and one that appears to be
diminished in frequency when a glycoprotein Ilb/IIIa inhibitor is added to the thrombo-
lytic regimen (21,40). Of importance, the amputation rate is higher when either rethrom-
bosis or distal embolization develop during thrombolytic infusion (41). Last, technical
complications may occur during percutaneous thrombolytic therapy, including catheter
or wire perforation of the arterial wall (1%), arterial dissection (2%), pericatheter throm-
bosis (3%), and false aneurysm formation (1-3%) (17,21).

V. SUMMARY AND CONCLUSIONS

Thrombolytic therapy is associated with significant benefits to the patient, including (a) the
ability to achieve restoration of blood flow in a minimally invasive fashion, (b) the iden-
tification of the culprit lesion responsible for the occlusive event, and (c) clearance of
thrombus from small vessels inaccessible to balloon catheters. These benefits have cul-
minated in a reduction in the rate of amputation (19) and death (1) for patients with acute
limb ischemia. Complications, however, continue to occur at a relatively high frequency.
Hemorrhage is the most frequent and clinically significant complication of thrombolytic
therapy. The rate of hemorrhage can be diminished by avoiding therapeutic heparin-
ization, decreasing the dose of thrombolytic agent, limiting the period of exposure, and
prudent patient selection. Even with attention to these and other measures, bleeding,
distal embolization, and rethrombosis will continue to occur with some frequency. These
issues should be considered in presenting the therapeutic options to a patient, carefully

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

508 OURIEL

weighing the risks versus the potential benefits prior to embarking on a particular treat-
ment regimen.

REFERENCES

1. Ouriel K, Shortell CK, DeWeese JA, Green RM, Francis CW, Azodo MV, et al. A comparison
of thrombolytic therapy with operative revascularization in the initial treatment of acute
peripheral arterial ischemia. J Vase Surg 1994; 19(6):1021 1030.

2. Sullivan KL, Gardiner GAJ, Kandarpa K, Bonn J, Shapiro MJ, Carabasi RA, et al. Efficacy of
thrombolysis in infrainguinal bypass grafts. Circulation 1991; 83(suppl 2):99— 105.

3. Garcia R, Saroyan RM, Senkowsky J, Smith F, Kerstein M. Intraoperative intra-arterial
urokinase infusion as an adjunct to Fogarty catheter embolectomy in acute arterial occlusion.
Surg Gynecol Obstet 1990; 1 71(3)201-205.

4. Ouriel K, Shortell CK, Azodo MV, Guiterrez OH, Marder VJ. Acute peripheral arterial
occlusion: predictors of success in catheter-directed thrombolytic therapy. Radiology 1994; 193
(2):561-566.

5. Comerota AJ, Weaver FA, Hosking JD, Froehlich J, Folander H, Sussman B, et al. Results of
a prospective, randomized trial of surgery versus thrombolysis for occluded lower extremity
bypass grafts. Am J Surg 1996; 172(2): 105-1 12.

6. Weaver FA, Comerota AJ, Youngblood M, Froehlich J, Hosking JD, Papanicolaou G Sur-
gical revascularization versus thrombolysis for nonembolic lower extremity native artery
occlusions: results of a prospective randomized trial. The STILE Investigators. Surgery versus
Thrombolysis for Ischemia of the Lower Extremity. J Vase Surg 1996; 24(4): 5 13-521.

7. Ouriel K, Veith FJ, Sasahara AA. A comparison of recombinant urokinase with vascular sur-
gery as initial treatment for acute arterial occlusion of the legs. N Engl J Med 1998; 338:1105-
1111.

8. Ricotta JJ, Green RM, DeWeese JA. Use and limitations of thrombolytic therapy in the
treatment of peripheral arterial ischemia: results of a multi-institutional questionnaire. J Vase
Surg 1987; 6(l):45-50.

9. Alkjaersig N, Fletcher AP, Sherry S. The mechanism of clot dissolution by plasmin. J Clin
Invest 1959; 38:1086.

10. Milstone H. A factor in normal human blood which participates in streptococcal fibrinolysis.
J Immunol 1941; 42:116.

11. Cliffton EE. The use of plasmin in humans. Ann NY Acad Sci 1957; 68:209-229.

12. Onundarson PT, Haraldsson HM, Bergmann L, Francis CW, Marder VJ. Plasminogen deple-
tion during streptokinase treatment or two-chain urokinase incubation correlates with de-
creased clot lysability ex vivo and in vitro. Thromb Haemost 1993; 70(6):998-1004.

13. van Breda A, Robison JC, Feldman L, Waltman AC, Brewster DC, Abbott WM, et al. Local
thrombolysis in the treatment of arterial graft occlusions. J Vase Surg 1984; 1(1): 103 1 12.

14. Ouriel K, Welch EL, Shortell CK, Geary K, Fiore WM, Cimino C. Comparison of strepto-
kinase, urokinase, and recombinant tissue plasminogen activator in an in vitro model of venous
thrombolysis. J Vase Surg 1995; 22(5):593-597. 1

15. Fox D, Ouriel K, Green RM, Stoughton J, Riggs P, Cimino C. Thrombolysis with prouro- <S
kinase versus urokinase: an in vitro comparison. J Vase Surg 1996; 23(4):657-666. js

16. McNamara TO, Fischer JR. Thrombolysis of peripheral arterial and graft occlusions: Improved °|
results using high-dose urokinase. AJR 1985; 144:769-775. <

17. Ouriel K, Gray BH, Clair DG, Olin JW. Complications associated with the use of urokinase and «
recombinant tissue plasminogen activator for catheter-directed peripheral arterial and venous J
thrombolysis. Vase Intervent Radiol 2000; 1 1 :295-298. q

18. Meyerovitz M, Goldhaber SZ, Reagan K, Polak JF, Kandarpa K, Grassi CJ, et al. Recom- |
binant tissue-type plasminogen activator versus urokinase in peripheral arterial and graft 2
occlusions: a randomized trial. Radiology 1990; 175:75-78. «

I
§

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS IN PERIPHERAL THROMBOLYSIS 509

19. Results of a prospective randomized trial evaluating surgery versus thrombolysis for ischemia
of the lower extremity. The STILE trial. Ann Surg 1994; 220(3):251-266.

20. Abbott Laboratories Venture Group. A comparison of urokinase and recombinant urokinase
in the treatment of peripheral arterial occlusion. Data on file. Abbott Park, IL: Abbott Lab-
oratories, 1994.

21. Ouriel K, Kandarpa K, Schuerr DM, Hultquist M, Hodkinson G, Wallin B. Prourokinase
versus urokinase for recanalization of peripheral occlusions, safety and efficacy: the PURPOSE
trial. J Vase Intervent Radiol 1999; 10(8): 1083-1094.

22. Dotter CT, Rosch J, Seaman AJ. Selective clot lysis with low-dose streptokinase. Radiology
1974; 111:31-37.

23. Katzen BT, Edwards KC, Albert AS, van Breda A. Low-dose direct fibrinolysis in peripheral
vascular disease. J Vase Surg 1984; l(5):718-722.

24. Graor RA, Risius B, Denny KM, Young JR, Beven EG, Hertzer NR, et al. Local thrombolysis
in the treatment of thrombosed arteries, bypass grafts, and arteriovenous fistulas. J Vase Surg
1985; 2(3):406-414.

25. Hoylaerts M, Rijken DC, Lijnen HR, Collen D. Kinetics of the activation of plasminogen by
human tissue plasminogen activator: role of fibrin. J Biol Chem 1982; 257:2912.

26. Lijnen HR. Van Hoef B, De Cock F, Collen D. Effect of fibrin-like stimulators on the activa-
tion of plasminogen by tissue-type plasminogen activator (t-PA): Studies with active site muta-
genized plasminogen and plasmin resistant t-PA. Thromb Haemost 1990; 64:61.

27. Cannon CP, Gibson CM, McCabe CH, Adgey AA, Schweiger MJ, Sequeira RF, et al. TNK-
tissue plasminogen activator compared with front-loaded alteplase in acute myocardial in-
farction: Results of the TIMI 10B trial. Thrombolysis in Myocardial Infarction (TIMI) 10B
Investigators. Circulation 1998; 98(25):2805-2814.

28. Owen J, Friedman KD, Grossman BA, Wilkins C, Berke AD, Powers ER. Quantitation of
fragment X formation during thrombolytic therapy with streptokinase and tissue plasminogen
activator. J Clin Invest 1987; 79(6): 1642-1647.

29. The GUSTO Investigators. An international randomized trial comparing four thrombolytic
therapies for acute myocardial infarction. N Engl J Med 1993; 329:673-682.

30. Ouriel K, Katzen B, Mewissen MW, Flick P, Clair DG, Benenati J, McNamara TO, Gibbens
D. Initial experience with reteplase in the treatment of peripheral arterial and venous occlusion.

I Vase Intervent Radiol 2000; 11:849-854.

31. del Zoppo GI, Higashida RT, Furlan AJ, Pessin MS, Rowley HA, Gent M. PROACT: a phase

II randomized trial of recombinant prourokinase by direct arterial delivery in acute middle
cerebral artery stroke. Stroke 1999; 29:4-11.

32. McNamara TO, Dong P, Chen J, Quinn B, Gomes A, Goodwin S, et al. Bleeding compli-
cations associated with the use of rt-PA versus r-PA for peripheral arterial and venous throm-
boembolic occlusions. Tech Vase Intervent Radiol 2001; 4(2):92-98.

33. Kasirajan K, Haskal ZJ, Ouriel K. The use of mechanical thrombectomy devices in the man-
agement of acute peripheral arterial occlusive disease. J Vase Intervent Radiol 2001; 12(4): 405-
411.

34. Greenberg R, Ouriel K, Srivastava S, Shortell C, Ivancev K, Waldman D, Illig KA, Green RM.

Mechanical versus chemical thrombolysis: an in vitro differentiation of thrombolytic media- j>

nisms. J Vase Intervent Radiol 2000; 1 1:199-205. |

35. Silva JA, Ramee SR, Collins TJ, Jenkins JS, Lansky AJ, Ansel GM, et al. Rheolytic throm- js
bectomy in the treatment of acute limb-threatening ischemia: immediate results and six-month °|
follow-up of the multicenter AngioJet registry. Possis Peripheral AngioJet Study AngioJet **
Investigators. Cathet Cardiovasc Diagn 1998; 45(4):386-393. |.

36. Ansel GM, George BS, Botti CF, McNamara TO, Jenkins JS, Ramee SR, et al. Rheolytic J
thrombectomy in the management of limb ischemia: 30-day results from a multicenter registry. q
J Endovasc Ther 2002; 9(4):395-402. |

37. Ahmed NK, Tennant KD, Markland FS, Lacz JP. Biochemical characteristics of fibrolase, S
a fibrinolytic protease from snake venom. Haemostasis 1990; 20(3): 147-1 54. «

I
§

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

510 OURIEL

38. Marder VJ, Stewart D. Towards safer thrombolytic therapy. Semin Hematol 2002; 39(3):206-
216.

39. Sharma GVRK, Cella G, Parisi AF, Sasahara AA. Thrombolytic therapy. N Engl J Med 1982;
306:1268-1272.

40. Ouriel K. The RELAX Trial. Presented at the TCT Conference, Washington DC, Sept. 2002.

41. Galland RB, Earnshaw JJ, Baird RN, Lonsdale RJ, Hopkinson BR, Giddings AE, et al. Acute
limb deterioration during intra-arterial thrombolysis. Br J Surg 1993; 80(9): 1 1 18 1 120.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

31

Complications of Sclerotherapy

John J. Bergan

University of California, San Diego, School of Medicine,
San Diego, California, and Uniformed Services University of the Health Sciences,

Bethescla, Maryland, U.S.A.

Mitchel P. Goldman

University of California, San Diego, School of Medicine,
San Diego, California, U.S.A.

I. INTRODUCTION

Venous insufficiency manifests itself in different ways. Its appearance varies from simple
telangiectatic blemishes to severe chronic leg ulcer in an edematous, pigmented leg. Symp-
toms also vary from the essentially asymptomatic limb to that with disabling postexercise
pain. Manifestations of venous insufficiency also vary and include those conditions — such
as telangiectasias, varicose veins, and axial incompetence — that are usually well treated. In
contrast, severe chronic venous insufficiency may be refractory to treatment. For example,
the severely damaged postthrombotic limb manifests segmental occlusion in combination
with universal reflux.

It is clear that treatment of varicose veins, primary venous insufficiency, and severe
chronic venous insufficiency will involve sclerotherapy to a greater extent in the near fu-
ture than in the recent past. Seventy years ago, sclerotherapy dominated treatment of ve-
nous insufficiency; however, through the 1950s, controlled trials comparing sclerotherapy ■a
with surgery revealed comparable short-term results but a much higher rate of recurrence |
after sclerotherapy than after surgery (1). Therefore interest in primary sclerotherapy de- a
clined and surgery became dominant. c

Various disciplines including internal medicine, family practice, dermatology, general <

surgery, and vascular surgery have treated venous insufficiency to a greater or lesser ex- &

tent. However, there is widely accepted and reasonable agreement among the various dis- J

ciplines that treatment of large varicose veins and saphenous venous insufficiency should «

be treated surgically. It is small veins that might benefit from sclerotherapy. Dermatol- |

ogists, for example, have taken up ambulatory phlebectomy for larger veins because of the @

511

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

512 BERGAN and GOLDMAN

influence of European dermatologists who invented specialized hook techniques for re-
moving varicosities (2).

However, new stimuli are reviving sclerotherapy. One of these is the wide availability
of ultrasound technology, which is used to guide sclerotherapeutic injections. Although
some adverse sequelae have been seen with ultrasound-guided sclerotherapy, this has not
deterred investigators (3). Large-vein sclerotherapy is liable to be associated with adverse
reactions such as superficial thrombophlebitis, and hyperpigmentation of the skin; there-
fore surgery has remained the mainstay of treatment.

Recently a number of techniques have been developed to create sclerosant foam,
which totally displaces the blood column and allows undiluted sclerosant agents to act on
the endothelium and vein wall to produce both endothelial damage and venoconstriction
(4). It is anticipated that sclerotherapy using foam will be increasingly popular and will
largely replace surgery in the near future. This is truly minimally invasive and is perceived
by patients and physicians alike as being very effective treatment.

The fundamental principle of sclerotherapy is obliteration of the lumen of the vein so
that no blood can flow through it (5). Injection sclerotherapy has been used in the past for
all sizes of varicose veins, ranging from protuberant saccular varicosities to minute telan-
giectasias. Gradually, principles have been developed that allow the successful perform-
ance of sclerotherapy.

II. PRINCIPLES OF SCLEROTHERAPY

In general, the first principle is the smaller the vessel to be destroyed, the greater the success
of sclerotherapy. A second principle is that varicose veins treated by sclerotherapy will
recur if axial reflux through the long and short saphenous veins is not controlled first.
Following these principles relegates the sclerotherapy of varicose veins to a distant second
place. It is most applicable in varicose veins that are persistent or recurrent following am-
bulatory surgery and varicose veins in the aged or infirm when these veins are symptomatic.

The fundamental mechanism of action of sclerotherapy is fibrotic obliteration of the
vein lumen. The detergent sclerosants in use today denude the endothelium. If blood is
present in the vein lumen, thrombus may form, and this thrombus may defeat the objectives
of complete fibrosis. Such postinjection thrombosis is best referred to in the presence of the
patient as "trapped blood." It is the most common side effect of sclerotherapy and is a
causative factor in the development of postinjection pigmentation, which is discussed below.
Veins larger than 1 mm and those raised above the skin surface are particularly prone to
thrombosis. When this occurs, the vein becomes black, raised, and tender. In veins larger
than 1 mm, a #18 needle can be introduced to allow external expression of the trapped blood.

One of the most common complications of sclerotherapy is postinjection hyper-
pigmentation (6). This is a product of the inflammatory process that obliterates the in- ■§
jected veins. There is no treatment for postinjection hyperpigmentation, but 90% of this |
will disappear within a year. Other side effects, including skin ulceration and postinflam- a
matory telangiectatic matting, are less commonly seen. c

For most injections, a 3- to 5- mL plastic syringe can be used, and this should be <

coupled to a 30-gauge hypodermic needle. The syringes are filled from vials of 3-5% >9

solutions and diluted to appropriate strength. For sodium tetradecyl sulfate, this is 0.25% J

for telangiectasias and 0.5% for varices 1-3 mm in diameter. Equivalent strengths of other «

available sclerosants can be calculated. |

There is much dogma surrounding the practice of sclerotherapy, most of it untrue. For @

example, there is no truth in the statement that crossing the legs brings on telangiectasias. H,

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF SCLEROTHERAPY 513

Similarly, the instruction for patients to walk around the clinic for 15-30 min and then
come back for an evaluation is unnecessary. Most practitioners of the art of sclerotherapy
believe that external compression is essential to the process, and a small but growing
number of physicians believe that compression after the injection of telangiectasias is to-
tally unnecessary.

Injection sclerotherapy does hold a rightful place in the treatment of venous insuf-
ficiency, but the rule still holds that the smaller the vessel, the more adaptable it is to
sclerotherapy.

III. SCLEROTHERAPY COMPLICATIONS

As with any therapeutic technique, sclerotherapy carries with it a number of potential
adverse sequelae and complications (Table 1). In addition to the previously mentioned
cutaneous pigmentation, there can be edema of the injected extremity, pain with injection
of certain sclerosing solutions, localized urticaria over injected sites, blisters or folliculitis
caused by postsclerosis compression, recurrence of previously treated vessels, stress-related
problems, and localized hirsutism. Relatively rare complications include localized cuta-
neous necrosis, systemic allergic reactions, clinically significant thrombophlebitis of the
injected vessel, arterial injection with resultant necrosis, deep venous thrombosis, nerve
damage, compartment syndrome, and air emboli. This chapter addresses the pathophysi-
ology of some of these reactions, methods for reducing their incidence, and treatment of
their occurrence.

A. Hyperpigmentation

This is a common occurrence after sclerotherapy of veins of all sizes; it occurs in ap-
proximately 10-80% of patients and can follow the use of any of the sclerosant agents. Its
incidence appears to depend on technique and on the concentration of the sclerosant. The

Table 1 Complications of
Sclerotherapy

Thrombosis, trapping

Post injection hyperpigmentation

Skin ulceration

Telangiectatic matting

Edema

Pain during injection

Localized urticaria 1

Blisters |

Vein recurrence a

Localized hirsutism c

Localized skin necrosis <

Systemic allergy 3

Superficial thrombophlebitis Sj

Arterial injection g

Deep venous thrombosis fj

Compartment syndrome g

Air emboli ®

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

514

BERGAN and GOLDMAN

brown color, due entirely to hemosiderin deposition, may be linear or punctate and is
reported to persist longer than 1 year in only 6% of patients (Fig. 1).

Reduction of pigmentation may be achieved by the use of lower concentrations of the
sclerosant, by treating the source of proximal reflux first in order to reduce intravascular
pressure, and by minimizing injection pressure. This last change can be accomplished by
exerting less pressure on the plunger of the syringe and by using a larger syringe (2- to
3- mL syringes are recommended). The use of postsclerotherapy compression with a 30- to
40- mm graduated medical compression stocking has been associated with the develop-
ment of less pigmentation.

Some patients are predisposed to the development of pigmentation, although the reas-
ons for this are poorly understood (7). Certain medications (such as minocycline) may lead
to the formation of a blue-gray discoloration in the skin following sclerotherapy. Certain
stages of the menstrual cycle are associated with increased vessel fragility, which may lead
to increased pigmentation.

I
•5

Figure 1 Linear brown streaks shown in this photograph are the product of intravascular
thrombosis in large telangiectasias. (From Ref. 16.)

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF SCLEROTHERAPY 515

Removal of postsclerotherapy thrombi may decrease the incidence and intensity of
pigmentation. The size of the thrombus may initially be minimized through the application
of strong compression over treated veins, especially if the veins are >4 mm in diameter.
Removal of thrombi may be easily accomplished by first locating the tender, often ery-
thematous "lumpy" areas, infiltrating the tissue with lidocaine, and then using an 18-
gauge needle or #11 blade or lancet to make a small puncture. The thrombus is then
expressed with manual pressure. Thereafter compression may be reapplied for up to 3 days
to decrease further thrombus accumulation.

Pigmentation usually lasts from 6 to 12 months. If it persists, one should search for a
vessel with persistent reflux into the area. Bleaching agents are usually ineffective. Ex-
foliants (trichloroacetic acid) may hasten the resolution of pigmentation by decreasing the
overlying cutaneous pigmentation, but they carry a risk of scarring or permanent hypo- or
hyperpigmentation. Laser treatment has been shown to be effective in many cases.

B. Swelling

Multiple factors are responsible for swelling (edema) after sclerotherapy. Edema is most
common when varicose veins or telangiectatic veins below the ankle are treated, especially
when volumes greater than 1 mL are injected into the ankle or foot or when higher con-
centrations of sclerosing agents are utilized. It may be caused directly by the application of
nongradient compression following sclerotherapy. Edema may be significantly reduced by
the use of graduated compression stockings following sclerotherapy. Edema is self-limiting
and generally resolves within several days to several months.

C. Telangiectatic Matting

The new appearance of previously unnoticed fine red telangiectatic veins occurs in many
patients after sclerotherapy or surgery for varicose veins or telangiectatic veins (Fig. 2).
These are sometimes referred to as flares, telangiectatic matting, blushing, or postscler-
otherapy neovascularization. The reported incidence varies from 5 to 75%. Probable risk
factors include obesity, use of estrogen-containing hormones, pregnancy, and a family
history of telangiectatic veins. Excessive postsclerotherapy inflammation may also predis-
pose to the development of matting.

New vessels can occur in 2-3 days or may not be visible for several weeks following
treatment, and they can develop on any part of the leg. Generally, telangiectatic matting
resolves in 3-12 months without any specific treatment. For persistent matting, one should
first attempt to locate a source of reflux into the area, usually a feeding reticular vein,
although matting may be a sign of underlying saphenous insufficiency. If matting persists,
sclerotherapy treatment with a different sclerosant or pulsed-dye laser treatment may be

effective. ^

g

D. Pain 1

I

Several variables may be altered to minimize the pain felt by the patient during sclero- c

therapy. The smallest needle possible should be used. If one inserts a needle multiple times, <

the needle should be changed frequently. Needles with an acute angle on the bevel and >9

those coated with silicone generally cause in less pain. The preprocedure application of a ^

topical anesthetic cream (EMLA) may be helpful in diminishing the pain with needle «

insertion. The cream must be applied in a thick layer 2 h before the procedure in order to |

be effective. @

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

516

BERGAN and GOLDMAN

Figure 2 Telangiectatic matting is a common and disturbing complication of sclerotherapy. It
consists of fine red intracutaneous vessels that are difficult to eradicate.

Choose the least painful sclerosing agent and dilute it with normal saline to the lowest
effective concentration. Hypertonic agents frequently cause severe burning and muscle
cramping, although this may be lessened with slower injection (8). Dilute sotradecyl or
polidocanol in any concentration usually result in the least pain after injection. Burning or
stinging pain resulting from any sclerosant may be diminished by firm pressure over the
affected area or by rubbing the area with alcohol.

Aching in the legs is common for several hours to several days following sclerother-
apy. This may be relieved by having the patient walk briskly and by the application of a
30- to 40- mmHg graduated medical compression stocking immediately following treat-
ment. Aching that does not respond to these measures may indicate the presence of deep
venous thrombosis.

E. Localized Urticaria

Localized urticaria occurs after injection of all sclerosing agents. It usually lasts less than
30 min and is probably the result of endothelial irritation. As it can occur with undiluted
hypertonic saline, it should not be confused with a systemic allergic response to the scle-
rosant. Bothersome itching can be diminished by applying topical steroids immediately
after injection and limiting the volume of additional injections.

I
•5

F. Folliculitis

Occlusion of any hairy area by bandaging can promote the development of folliculitis,
especially if the area becomes moist with perspiration. Treatment consists of removal of
the dressing or compression and application of an antibacterial soap or topical antibiotic

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF SCLEROTHERAPY

517

gel. The folliculitis usually disappears within a few days. If the itching is bothersome, a
topical steroid preparation may be used. Systemic antibiotics will rarely be necessary.

G. Localized Hirsutism

Hypertrichosis overlying a vein treated with sclerotherapy may occur, possibly due to
improved cutaneous oxygenation or increased vascularity in the area. It may develop one
or more months after treatment and is generally self-limiting.

H. Cutaneous Necrosis

Ulceration of the skin may occur with the injection of any sclerosing agent even under
ideal circumstances and does not necessarily imply physician error (Fig. 3). It is thought to
be the result of extravasation of a sclerosing solution into the perivascular tissues. Injec-
tion into a dermal arteriole is more likely (Fig. 4). Otherwise it may be due to an arterio-
venous anastomosis, severe vasospasm of the vessel, or excessive cutaneous pressure
created by compression techniques.

Direct extravasation may be due to poor technique, leakage of the sclerosant through
holes made in the vein from multiple prior injections or through-and-through perforation
of the vein, tracking the solution into the tissues as the needle is withdrawn, or a "blow-
out" of a vessel due to the endothelial necrosis that may occur when strong sclerosants are
used.

I
•5

Figure 3 Cutaneous necrosis as shown here is a significant complication of sclerotherapy. It is a
misunderstood and difficult-to-treat complication.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

518

BERGAN and GOLDMAN

<• "./* '>

V

- >rS

Figure 4 In this biopsy of an area of cutaneous necrosis, Goldman has shown a thrombosed
arteriole in the base of the ulceration. (From Ref. 17.)

Some sclerosing agents are more caustic to the tissues — such as hypertonic saline,
dextrose, or polyiodinated iodine — although cutaneous necrosis has been reported with all
sclerosing solutions. If significant extravasation is suspected, the area should be infiltrated
with normal saline in an attempt to dilute the sclerosant. Ten times the volume of ex-
travasated solution must be used in order to dilute the sclerosant effectively. Infiltration of
the area with 250 U of hydraluronidase may also be helpful due to the accelerated dilution,
cellular stabilization, and wound repair properties of this agent (9).

Approximately 4% of telangiectatic veins are associated with a dermal arteriole. Thus,
injection into an arteriovenous anastomosis is probably the most common cause of
cutaneous ulcerations following sclerotherapy (Fig. 5). Reduction of the volume injected
in each site along with a reduction of the pressure exerted on the plunger during injection
may help to reduce this complication.

Rarely after injection, an immediate porcelain-white appearance to the skin is noted at
the site of injection. A hemorrhagic bulla may form over this area within 2-48 h and later
progress to an ulcer (Fig. 6). It is thought that this represents arterial spasm. In an attempt
to reverse the spasm, vigorous massage may be directed to the area. To further reduce the
likelihood of ulceration, one can rub a small amount of 2% nitroglycerin ointment into the
area of blanching. This is usually successful in preventing an ulcer.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF SCLEROTHERAPY 519

Precapillary
Sphincter

Arteriole

Venule

Thermoregulatory
A-V Shunt

Figure 5 This diagram of a cutaneous thermoregulatory shunt explains why a true intravenous
injection of sclerosant can be distributed by long-standing venous hypertension.

Excessive compression of the skin overlying the treated vein may produce tissue an-
oxia with the development of cutaneous ulceration. This is primarily found in patients
with underlying arterial occlusive disease and most often occurs on the foot. Using in-
elastic compression, which does not exert any pressure on the foot or leg when the patient
is supine, or using two 20- to 30- mmHg stockings during the day when the patient is
upright and removing one of them at nighttime may prevent this complication.

Since most of these ulcerations are less than 4 mm in diameter, primary healing usu-
ally results in an acceptable scar. Larger ulcerations may be excised and closed for faster
healing and a smaller scar. Application of hydrocolloid or hydrophilic dressings may re-
duce the pain associated with the ulcer and speed the rate of healing. Most ulcers will heal
within 4 weeks to 4 months.

I. Systemic Allergic Reactions

Allergic reactions may occur following sclerotherapy with any of the sclerosants, even
hypertonic saline. Minor reactions such as urticaria are easily treated with an oral anti-
histamine such as diphenhydramine (Benadryl) 25 to 50 mg PO or hydroxyzine (Atarax)
10 to 25 mg PO. If the reaction does not subside readily, a short course of prednisone 40-
60 mg PO q.d. for one week in conjunction with the antihistamine is helpful.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

520

BERGAN and GOLDMAN

Figure 6 This area of tissue necrosis developed following ultrasound-guided injection of 2 mL of
sclerosant foam equivalent to 0.5% STD.

Because of the possibility of angioedema or bronchospasm, each patient with an
allergic reaction should be examined for stridor and wheezing by auscultation. Supine and
seated pulse and blood pressure should be measured to rule out orthostatic changes,
hypotension, or tachycardia that might result from the vasodilation that precedes ana-
phylactic shock. If any of these more serious allergic reactions occur, an intravenous line
should be started immediately. For angioedema with stridor, the patient should be given

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF SCLEROTHERAPY 521

Benadryl IV or IM and corticosteroids IV. A laryngoscope and endotracheal tube should
be available. For bronchospasm, antihistamines and corticosteroids should be adminis-
tered intravenously. A bronchodilator should be given either as an inhaler or nebulizer. If
anaphylaxis is present, large volumes of intravenous fluids should be administered, along
with 0.3-0.5 mL of epinephrine 1:1000 subcutaneously every 20-30 min as needed, up to
three doses. If the reaction is feared to be life-threatening, 5 mL of epinephrine 1:10,000
should be administered intravenously every 5-10 min as needed. Obviously, patients with
serious allergic reactions should be transferred to a hospital and observed for at least 6-
12 h due to the biphasic nature of allergic reactions.

In addition to these general allergic reactions, there are point tenderness toxic re-
actions that may occur with all of the popular sclerosing agents. The physician should be
familiar with these reactions and treat them if they occur.

J. Superficial Thrombophlebitis

Superficial thrombophlebitis may appear 1-3 weeks after injection as a tender, erythem-
atous induration over the injected vein. This complication is observed less often if
compression is maintained for an adequate period of time following sclerotherapy. The
compression should be applied over the treated veins and even more proximally, as the
sclerosing action may spread proximal to the injection site. When thrombophlebitis oc-
curs, the thrombus should be evacuated and adequate compression and frequent ambu-
lation should be maintained until the pain and inflammation subside. Aspirin or other
nonsteroidal anti-inflammatory medication may be helpful in limiting both the inflam-
mation and the pain. The concomitant presence of a deep venous thrombosis must be
considered.

K. Arterial Injection

The most feared complication in sclerotherapy is inadvertent injection into an artery.
Fortunately, this complication is very rare. Arterial injection of a sclerosing agent pro-
duces a sludge embolus that obstructs small arteries and the microcirculation. Experiments
demonstrate little effect on the artery itself (10). The most common location for arterial
injection is the posterior or medial malleolar region. Less common areas include the
proximal thigh and popliteal fossa (the regions of the saphenofemoral junction and the
saphenopopliteal junction). After intra-arterial injection, the patient will usually though
not always note immediate pain. A dusky cyanotic hue may be noted at first, followed by
pallor. Sloughing of the superficial tissues may result in significant scarring or even ampu-
tation of a limb. Occlusion of small arteries may lead to the development of a compart-
ment syndrome with nerve damage and even paralysis. ■§
Arterial injection is a true sclerotherapy emergency (11). The extent of cutaneous ne- |
crosis is usually related to the amount of solution injected (12). Efforts to treat this com- a
plications are usually unsatisfactory but should be made. Practical treatments include c
periarterial infiltration with 1 mL of procaine 3%, which will form a complex with <
sotradecyl and render it inactive. The affected area should be cooled with ice packs to >9
minimize tissue anoxia. Immediate heparinization (continued for 6 days) and the admin- J
istration of intravenous 10% dextran, 500 mL for 3 days, is recommended. Thrombolysis °
should be considered if no contraindication exists. Finally, use of prazosin, hydralazine, or |
nifedipine PO for 30 days should be considered. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

522 BERGAN and GOLDMAN

L. Deep Venous Thrombosis

Pulmonary embolism and deep venous thrombosis are diagnosed very rarely after scle-
rotherapy (13). However, most cases of thrombosis and embolization go unnoticed, so the
incidence is greater than statistics demonstrate (14). Sclerotherapy may affect all three
components of the Virchow triad: Endothelial damage, vascular stasis, and changes in
coagulability (15). Thus efforts to reduce the introduction of the sclerosant agent or a
thrombus into the deep venous system should be a high priority. The volume of sclerosing
agent per injection site should be limited in order to reduce the chance of its entering the
deep venous system at a concentration sufficient to cause damage. Blood flow in the deep
venous system must be rapidly stimulated with compression and muscle movement in an
attempt to dilute any sclerosing agent that might be present. For this reason, some phy-
sicians ask patients to flex their ankles periodically during treatment. Patients should be
asked to ambulate immediately and frequently after treatment. Postsclerotherapy com-
pression may also reduce the incidence of thrombosis if applied in a graduated fashion.
Sclerotherapy should be undertaken cautiously in patients with previous deep venous
thrombosis or in those with a hypercoagulable state.

M. Nerve Damage

Because of their close proximity to the long and short saphenous veins, the saphenous and
sural nerves may be damaged inadvertently during sclerotherapy. Injection into a nerve is
very painful and may cause anesthesia and sometimes permanent interruption of nerve
function. An area of paresthesia following sclerotherapy may result from perivascular
inflammation extending from the treated vein to nearby superficial nerves. Nonsteroidal
anti-inflammatory medications or high-potency topical corticosteroids may hasten reso-
lution of this problem, which may take 3-6 months.

REFERENCES

1. Hobbs JT. The treatment of varicose veins: A random trial of injection compression therapy
versus surgery. Br J Surg 1968; 55:777-780.

2. Neumann HAM, De Roos KP, Veraart JCJM. Muller's ambulatory phlebectomy and com-
pression. Dermatol Surg 1998; 24:471-474.

3. Bergan JJ, Weiss RA, Goldman MP. Extensive tissue necrosis following high-concentration
sclerotherapy for varicose veins. Dermatol Surg 2000; 26:535-542.

4. Tessari L, Cavezzi A, Frullini A. Preliminary experience with a new sclerosing foam in the
treatment of varicose veins. Dermatol Surg 2001; 27:58-60.

5. Bergan JJ. Varicose veins: Treatment by surgery and sclerotherapy. In: Rutherford RB, ed.
Vascular Surgery. 5th ed. Saunders, 2000:2007-2021.

6. Georgiev M. Post sclerotherapy hyperpigmentations: A one-year followup. J Dermatol Surg
Oncol 1990; 16:608-610.

7. Lupo ML. Sclerotherapy: Review of results and complications in 200 patients. J Dermatol Surg

s

Oncol 1989; 15:214-219. 1

8. Thibault P, Wlodarczky J. Post sclerotherapy hyperpigmentation: The role of ferritin levels and
effectiveness of treatment with the copper vapor laser. J Dermatol Surg Oncol 1992; 18:47-52.

9. McCoy S, Evans A, Spurrier N. Sclerotherapy for leg telangiectasias: A blinded comparative
trial of polidocanol and hypertonic saline. Dermatol Surg 1999; 25:371-386.

10. Zimmet SE. The prevention of cutaneous necrosis following extravasation of hypertonic saline
and sodium tetradecyl sulfate. J Dermatol Surg Oncol 1993; 19:641-646.

11. MacGowan WAL, Holland PDJ, Browne HI, Byrnes DP. The local effects of intraarterial

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

c

I

COMPLICATIONS OF SCLEROTHERAPY 523

injections of sodium tetradecyl sulphate (STD) 3%: An experimental study. Br J Surg 1972;
59:101-104.

12. Natali J, Farman T. Implications medico-legales au cours du traitement sclerosant des varices.
J Malad Vase 1996; 21:227-232.

13. Bergan JJ, Weiss RA, Goldman MP. Extensive tissue necrosis following high-concentration
sclerotherapy for varicose veins. Dermatol Surg 2000; 26:535-542.

14. Yamaki T, Nozaki M, Sasaki K. Acute massive pulmonary embolism following high ligation
combined with compression sclerotherapy for varicose veins: Report of a case. Dermatol Surg
1999; 25:321-325.

15. Feied CF. Deep venous thrombosis: Fhe risks of sclerotherapy in hypercoagulable states. Sem
Dermatol 1993; 12:135-149.

16. Weiss RA, Feied CF, Weiss MA. Vein Diagnosis and Treatment: A Comprehensive Approach.
New York: McGraw Hill Medical Publishing Division, 2001.

17. Goldman MP, Bergan JJ, eds. Sclerotherapy Treatment of Varicose and Telangiectatic Leg
Veins. 3rd ed. Mosby: St Louis, 2001.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

32

Complications of Subfascial Endoscopic Perforator Vein
Surgery and Minimally Invasive Vein Harvests

Peter Gloviczki and Manju Kalra

Mayo Clinic, Rochester, Minnesota, U.S.A.

Minimally invasive endoscopic procedures have been used for almost two decades for
surgical treatment of both venous and arterial pathology. The primary aim of minimally
invasive techniques is to reduce complications related to open surgical techniques. En-
doscopic procedures are used to accelerate functional recovery, achieve a better cosmetic
result, and enhance patient satisfaction.

Subfascial endoscopic perforator vein surgery (SEPS) was introduced in the mid-1980s
by Hauer to treat incompetent perforating veins in patients with varicosity, advanced
chronic venous insufficiency, and venous ulcers (1). Endoscopic techniques permitted in-
terruption of incompetent perforating veins in the calf under direct vision with the help of
an endoscope introduced through a small incision made proximal to the area of lipo-
dermatosclerosis and venous ulcer. This technique was rapidly adopted and refined by
several groups; in the past decade, SEPS has emerged as an effective minimally invasive
technique to interrupt incompetent perforating veins (2-21).

Endoscopic vein harvesting (EVH) techniques to remove the great saphenous vein ■§

(GSV) were developed to decrease wound complications and improve patient satisfaction |

in those who undergo coronary artery bypass grafting (CABG) or lower extremity revas- a

cularization for critical limb ischemia. Experience using different endoscopic instrumen- c

tations to harvest the GSV has rapidly increased in the past decade (22-37). <

In this chapter first we discuss current techniques and the associated complications of >9

SEPS. We also describe the most frequently used techniques of EVH, discuss complica- J

tions, and compare those with complications observed after traditional open vein harvest- «

ing techniques. |

525

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

526 GLOVICZKI and KALRA

I. SUBFASCIAL ENDOSCOPIC PERFORATOR VEIN SURGERY
A. Surgical Techniques

Two main techniques for SEPS have been developed. The first is a single-scope technique,
a refinement of the original work of Hauer (1), Fischer (2), and Jugenheimer (3) with fur-
ther development by Bergan and his team (8,14,17) and by Wittens and Pierik (6,10,11). It
uses a single scope with channels for the camera and working instruments, which some-
times makes visualization and dissection in the same plane difficult (Fig 1). Improvement
in instrumentation for this technique now allows for carbon dioxide insufflation into the
subfascial plane.

The second technique, using standard laparoscopic instrumentation, was introduced in
the United States by O'Donnell (4) and developed simultaneously by Conrad in Australia
(5) and by our team at the Mayo Clinic (7,16,20,21). This two-port technique employs one
port for the camera and a separate port for instruments, thereby making it easier to work in
the subfascial space. First the limb is exsanguinated with an Esmarque bandage and a thigh
tourniquet is inflated to 300 mmHg to provide a bloodless field (Fig. 2A). A 10-mm endo-
scopic port is placed in the medial aspect of the calf 10 cm distal to the tibial tuberosity, 2 cm
medial to the anterior edge of the tibia. Balloon dissection is used to widen the subfascial
space and facilitate access after port placement (Fig 2B). The distal 5-mm port is now placed
halfway between the first port and the ankle (about 10-12 cm apart), about 5-7 cm posterior
to the anterior edge of the tibia, under direct visualization with the camera (Fig. 2C). Car-
bon dioxide is insufflated into the subfascial space and pressure is maintained at 30 mmHg
to improve visualization and access to the perforators. Using laparoscopic scissors inserted
through the 5-mm port, loose connective tissue between the calf muscles and the superficial
fascia is sharply divided.

I
•5

Figure 1 Olympus endoscope for the subfascial perforating vein interruption. The scope can be
used with or without C0 2 insufflation. It has an 85-degree field of view and the outer sheath is either
16 or 22 mm in diameter. The working channel is 6 by 8.5 mm, with a working length of 20 cm.
(From Ref. 43.)

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

PERFORATOR VEIN SURGERY AND VEIN HARVESTS

527

B

Figure 2 Two port technique of SEPS. A. A thigh tourniquet inflated to 300 mmHg is used to
create a bloodless field. B. Balloon dissection is used to widen the subfascial space. C. SEPS is
performed using two ports: a 10-mm camera port and a 5- or 10-mm distal port inserted under video
control. Carbon dioxide is insufflated through the camera port into the subfascial space to a pressure
of 30 mmHg to improve visualization and access to perforators. D. The subfascial space is widely
explored from the medial border of the tibia to the posterior midline and down to the level of the
ankle, and all perforators are interrupted using clips or harmonic scalpel. E. A paratibial fasciotomy
is routinely performed to identify perforators in the deep posterior compartment. (From Ref. 44.)

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

528

GLOVICZKI and KALRA

D

Figure 2 Continued.

The subfascial space is explored from the medial border of the tibia to the posterior
midline and down to the level of the ankle. All perforators encountered are divided either
with the harmonic scalpel or sharply between clips (Fig. 2D). A paratibial fasciotomy is next
made by incising the fascia of the posterior deep compartment (Fig. 2E). The Cockett II and
Cockett III perforators are located frequently within an intermuscular septum, and this has
to be incised before identification and division of the perforators can be accomplished. The
medial insertion of the soleus muscle on the tibia is also exposed proximally as high as
possible to visualize proximal paratibial perforators. By rotating the ports cephalad and
continuing the dissection up to the level of the knee, the more proximal perforators can also

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PERFORATOR VEIN SURGERY AND VEIN HARVESTS 529

E

Figure 2 Continued

be divided, although in our practice we seldom perform this part of the procedure. While the
paratibial fasciotomy can aid in distal exposure, reaching the retromalleolar Cockett I per-
forator endoscopically safely is seldom possible.

After completion of SEPS, the instruments and ports are removed, the CO2 is manually
expressed from the limb and the tourniquet is deflated. For postoperative pain control,
20 mL of 0.5% bupivacaine (Marcain) solution is instilled into the subfascial space. Stab
avulsion of varicosities in addition to high ligation and stripping of the great and/or small
saphenous vein or radiofrequency closure of the saphenous vein is performed as needed.
The wounds are closed and the limb is wrapped with an elastic bandage. Elevation is main-
tained at 30 degrees postoperatively for 3 hr, after which ambulation is permitted. The
patients are usually discharged on the same day or occasionally next morning following
overnight observation.

B. Complications

SEPS may lead to systemic or local, nonvascular and vascular complications.

1 . Systemic Complications

Patient selection is important to prevent systemic complications and this procedure should TJ

not be done in those at high risk for general or epidural anesthesia or if unfit for surgical |

treatment. Using these criteria, mortality following SEPS is exceedingly rare. No deaths was a

reported in the North American SEPS registry and no pulmonary embolism was noted c

within 30 days in 148 patients who underwent 155 SEPS procedures (18,19). In our expe- <

rience with over 1 50 SEPS procedures performed at our institution using general or epidural •3

anesthesia, no death or pulmonary embolism was noted. One patient with known protein C 3

deficiency with a history of multiple episodes of deep venous thrombosis developed recur- °

rent popliteal vein thrombosis 2 months after the operation (20). Iafrati and O'Donnell I

reported on 51 SEPS procedures, with no deaths or cardiac, thromboembolic, or other ©

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

530 GLOVICZKI and KALRA

systemic complications (15). To prevent deep vein thrombosis and pulmonary embolism, we
give 3000 U of heparin before placement of the tourniquet for those patients who have
underlying coagulation abnormality and postthrombotic syndrome. These patients also
receive low-molecular-weight heparin postoperatively until full ambulation is resumed. In
all others, leg elevation, early ambulation and elastic compression alone are used for
thrombosis prophylaxis.

2. Local Complications

Local complications develop in 5-8% of the patients. Nonvascular local complications
include cellulitis, wound dehiscence or infection, seroma or lymphocele, postoperative pain
usually related to subfascial hematoma, saphenous nerve neuralgia, and injury to the tibial
nerve. Vascular complications include thrombophlebitis, injury to the posterior tibial
vessels, and deep venous thrombosis.

In the 148 registry patients, cellulitis occurred in 4 (3%) and wound infections in 8 (5%);
five at port entry sites and three at other incisions (18). Wound infections after 30 days
developed in 2 patients, one of whom also had early infection. Therefore the overall inci-
dence of wound infection was at least 6% (9 of 148) (18). Of the 9 patients with wound
infections, 7 had active ulcers at the time of the SEPS procedure. However, we failed to
identify any clinical factors that predicted the development of wound infection in these
patients (Table 1). Of the 10 patients with saphenous neuralgia, 9 underwent stripping of
the GSV (8 from groin to knee, 1 from groin to ankle). Only avulsion of varicose veins was
identified as a clinical variable significantly associated with saphenous neuralgia (Table 1). It
is important, however, to avoid nerve injury at the time the proximal port is placed through
the fascia.

A roll-on tourniquet caused skin necrosis in one patient (18), and this tourniquet is not
used at our institution. A thigh tourniquet should be released after 90 min, and it should
always be padded with gauze to protect the skin from pressure necrosis. No complication
has been reported with pneumatic tourniquets, where constant pressure (200-300 mmHg)
can be maintained and monitored during the procedure.

The registry reported superficial thrombophlebitis in 5 patients, of which 3 occurred
within 30 days. Because all direct medial perforator veins join the paired posterior tibial
veins (38), the potential for injury of the neurovascular bundle at that level exists. Gesels-
chap et al. reported tibial nerve injury in one patient and injury to the posterior tibial artery
in another (39). Dissection of the low medial perforators (Cockett II) should be done cau-
tiously, since many times these perforators are located in the deep posterior compartment
and the fascia overlying this compartment has to be incised to gain control of these im-
portant perforators (38). Interruption of all perforators must be done as close to the super-
ficial fascia as possible to avoid injury to deep structures. As emphasized by Wittens and
his team, one must be absolutely sure at this level that the divided structure penetrates the
superficial fascia before clipping or division is done (39). Otherwise one can inadvertently
injure the posterior tibial vessels or the tibial nerve.

s

A prospective randomized study between SEPS and open perforator ligation was per- |j>

formed in Holland and found less complications and shorter hospitalization following
SEPS (11). The incidence of wound infections after open exploration was 53%, compared
to 0% in the endoscopic group (/;< 0.001). Patients in the open group needed longer
hospital stays (mean, 7 days; range, 3-39 days) than patients in the endoscopic group (mean,
4 days; range, 2-6 days; p — 0.001).

With appropriate patient selection, adherence to the guidelines of endoscopic surgical
techniques and thrombosis prophylaxis SEPS has a low complication rate and similar

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

c

I

PERFORATOR VEIN SURGERY AND VEIN HARVESTS

531

Table 1 Association of Clinical Variables and Early Surgical Complications in the NASEPS
Registry

Clinical variables

Wound infection Saphenous neuralgia

(n = 8) p Value 11 (n = 6) p Value"

Ulcer

Present

Absent
Saphenous vein stripping

Done

Not done
High ligation

Yes

No
Tourniquet time

> 60 min

< 60 min
Endoscopes used

One

Multiple
C0 2 insufflation

Yes

No
Varicose vein avulsion

Yes

No
Diabetes

Yes

No :

Gender
Female
Male

6/79 (7.6%)
2/29 (6.9%)

4/66 (6.1%)
4/42 (9.5%)

1/10 (10.0%)
7/98 (7.1%)

1/22 (4.6%)
3/59 (5.1%)

4/55 (7.3%)
4/53 (7.6%)

3/52 (5.8%)
5/56 (8.9%)

4/53 (7.6%)
4/55 (7.3%)

0/4 (0%)

;/ioo (8.0%)

5/49 (10.2%)
3/59 (5.1%)

1.0

0.709

0.553

1.0

1.0

0.718

1.0

0.464

3/79 (3.8%)
3/29 (10.3%)

6/66(9.1%)
0/42 (0%)

0/10 (0%)
6/98 (6.1%)

0/22 (0%)
6/59 (10.2%)

4/55 (7.3%)
2/53 (3.9%)

2/52 (3.9%)
4/56(7.1%)

6/53 (11.3%)
0/55 (0%)

0/4 (0%)
6/100 (6.0%)

5/49 (10.2%)
1/59(1.7%)

0.340

0.080

1.0

0.182

0.679

0.680

0.012

0.09

Fisher's exact test.

long-term clinical outcome as open interruption of incompetent perforators (1 1). Although
foam sclerotherapy is emerging as potentially acceptable minimally invasive alternative
therapy, that procedure is not without complications and the recurrence rate is not yet
known. The minimally invasive SEPS remains the treatment of choice for perforator
interruption in 2003.

II. ENDOSCOPIC SAPHENOUS VEIN HARVESTING
A. Surgical Techniques

Several instruments are available for endoscopic saphenous vein harvesting (EVH), includ-
ing instrumentation of Ethicon Endo-Surgery Inc. (Cincinnati, OH) and the Vaso View
System (Guidant Cardiac and Vascular Surgery, Menlo Park, CA), among others. Our
technique using the Ethicon instrumentation for EVH has been previously described (27).
The principal instrument is a disposable subcutaneous retractor, which incorporates a 5-mm

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

532

GLOVICZKI and KALRA

straight or a 30-degree angled endoscopic camera for viewing. This retractor is available in
two different sizes. The endoscopic instruments include a pigtail dissector (a modified Mayo
vein stripper) (Fig. 3), an endoscopic clip applier (5-mm Allport Clip applier) and endo-
scopic scissors (Ethicon Endo-Surgery, Inc., Cincinnati, OH). A 2.5 cm long incision is first
made in the groin crease, the saphenofemoral junction is identified and all tributaries of the
GSV are ligated proximally and distally. A subcutaneous tunnel is developed caudally over
the anterior surface of the GSV with the aid of the endoscopic retractor containing the 5-mm
videoscope (Fig. 3). The retractor is advanced distally to dissect and retract the tissues along
the anterior surface of the vein. A gentle upward pressure has to be applied on the retractor
to avoid injury to the saphenous vein by stripping its adventitia. All tributaries are clipped
distally and divided. Only few very large tributaries need to be double clipped before
division. The GSV must not be held under great tension during dissection to avoid
adventitial and endothelial damage.

Once the limits of the retractor are reached (about 30 cm), the pigtail dissector is in-
serted to free up the vein circumferentially (Fig. 4). Additional lateral and any posterior
tributaries are dissected, clipped, and divided (Fig. 5). Since clips are placed only laterally
on the tributaries, the pigtail dissector can be advanced over the GSV without danger of
vein injury. Nevertheless, smaller tributaries are frequently avulsed during dissection, and
this renders in situ bypass rather difficult to perform because of bleeding in the tunnel
through these sites.

When the entire length of the exposed vein is mobilized, additional distal incisions are
made to insert the scope for further dissections. A longer incision at the knee level is fre-
quently needed because of multiple genicular tributaries at this level. The adjacent saphe-
nous nerve should be protected at the knee and the calf during vein dissection.

I
•5

Figure 3 Instruments for endoscopic vein harvest (Ethicon Endosurgery, Inc., Cincinnati, OH).
Left. Endoscopic dissector/retractor with a port for a 5- by 300-mm videoscopic lens. Right. A
pigtail dissector (a modified Mayo vein stripper). (From Ref. 45.)

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PERFORATOR VEIN SURGERY AND VEIN HARVESTS

533

I
•5

B

Figure 4 A and B. After the side branches are clipped with an endoscopic clip applier on the side of
the body, they are divided with endoscopic scissors on the side of the vein. Venospasm and
competent valves minimize blood loss. (From Ref. 45.)

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

534

GLOVICZKI and KALRA

MAYO

Figure 5 The anterior surface of the saphenous vein has been dissected using an endoscopic dis-
sector. The vein is encircled with the open circle of the pigtail dissector, gently dissected from sur-
rounding soft tissue, and the branches identified. (From Ref. 45.)

Upon completion of the vein dissection, high ligation and division of the proximal
tributaries is performed, the vein is divided and clipped distally, and is removed. The vein
is distended with heparinized saline/papaverine solution (our vein solution contains 1000
U of heparin sodium and 30 mg of papaverine in 400 mL of normal saline). Inflation over
arterial pressure (120 mmHg) is avoided by using a syringe with a pressure gauge. Pre-
viously transected tributaries are ligated with 3-0 or 4-0 silk ties. Any tear in the vessel is
repaired with 7-0 monofilament polypropylene sutures.

The technique of the Vaso View (Guidant Cardiac and Vascular Surgery, Menlo Park,
CA) was described by several authors, including Bitondo et al. (37). This group uses a 2- to
2.5-cm transverse incision made just above the knee, where the vein is identified first. With
the aid of carbon dioxide insufflation, a dissection cannula is used to isolate the vein and its
surrounding branches. Once isolation of the vein is achieved, a second endoscopic instru-
ment is used through which scissors connected to a bipolar cautery is inserted to cauterize
and cut the tributary branches. A 1- to 1.5-cm incision is made in the groin to perform high
ligation and division of the proximal saphenous vein and an additional 1-cm incision is made

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PERFORATOR VEIN SURGERY AND VEIN HARVESTS 535

in the lower leg to expose the end of the vein for ligation and division. Metal clips can also be
used to clip distal tributaries of the vein. After hemostasis is confirmed, the tunnel is irrigated
with antibiotic solution and the incision closed with one or two interrupted, absorbable
subcutaneous stitches and a running, 4-0 absorbable suture for the subcuticular layer. A
drain is used only occasionally in obese patients or those with bleeding tendency. If the EVH
was performed for CABG, the leg is wrapped immediately after skin closure with an elastic
bandage that is used for at least 2 days.

B. Complications

Since EVH is done in patients who undergo lower limb revascularization or CABG,
systemic complications are usually the results of the primary operation. Local complica-
tions of EVH are listed in Table 2 and range from wound dehiscence, infection, lympho-
cele, and lymphorrhea to wound necrosis, graft failure and even limb loss.

There is increasing evidence that EVH is associated with a reduced rate of wound com-
plications. Most reports emphasize improved patient comfort, early mobility, decreased
length of hospital stay, and superior cosmetic results (22-37). In early reports, Lumsden et
al. observed only 3 wound complications following endoscopic removal of the GSV in 30
limbs. Two of these patients had cautery burns and one patient had injury to the saphenous
vein (23). In a prospective series of 68 consecutive GSV harvests for lower extremity revas-
cularization, Jordan et al. reported a wound complication rate of 8.8% (22). No graft failure
or, more importantly, no graft harvest-related wound complications were noted, saphenous
neuralgia appeared to decrease and hospitalization was also reduced (22). In a recent study,
Alcocer and Jordan reported results in 185 operations using EVH (40). Only 16 patients
(8.6%) had minor wound complications, including drainage, erythema, hematoma, and
mild inflammation. Eleven patients had prolonged hospitalization or readmission because
of wound complications. All complications were managed easily without further compli-
cations or wound dehiscence. The authors concluded that pain and perioperative compli-
cations were reduced after EVH compared to their own experience with open harvesting of
the saphenous vein.

Table 2 Complications of Great
Saphenous Vein Harvest

Wound infection

Cellulitis

Lymphangitis

Skin necrosis

Wound dehiscence a

Lymphorrhea «

Lymphocele a

Seroma c

Hematoma <

Abscess M

Saphenous neuralgia §

Paresthesia g

Graft failure "§

Limb amputation g

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

536 GLOVICZKI and KALRA

In a prospective nonrandomized study Bitondo et al. studied wound complications
and compared results of EVH in 92 patients to those with open vein harvesting (OVH)
in 133 (37). Wound complications (dehiscence, drainage for greater than 2 weeks post-
operatively, cellulitis, hematoma, and seroma/lymphocele) were significantly less in the
endoscopic vein harvest group (9 of 133, 6.8%) versus the open vein harvest group (26 of
92, 28.3%). The 9 complications in the EVH group included seroma in 6 and dehiscence,
lymph drainage and cellulites in one patient each. By multivariable analysis with logistic
regression, the open vein harvest technique was the only risk factor for postoperative leg
wound complication (relative risk 4.0). These results were confirmed in prospective
randomized trials by Allen et al. (28) and in another by Puskas et al. (31), who observed
more drainage from leg incisions at hospital discharge in the open harvest group (34%)
versus EVH group (8%; p = 0.001), but more ecchymosis in the EVH group. This study
did not find reduced leg incision pain in the EVH group, and there was no statistically
significant difference in pain or in the quality of life measured at any point in time. In
addition, there was no difference between groups in readmission to hospital, administra-
tion of antibiotics, or incidence of leg infection. Mean hospital charges for the EVH group
were approximately $1500 greater than for OVH; this difference however, did not reach
statistical significance. These authors concluded that EVH is a safe, reliable, and cost-
neutral method for saphenous vein harvest, but commented that the best indication for
EVH may be in patients who are at increased risk for wound infection and in those for
whom cosmesis is a major concern.

In another prospective randomized trial, Schurr et al. recently compared EVH to tra-
ditional open harvesting technique for patients who underwent CABG. In this trial, 140
CABG patients were randomized into two groups, 80 EVH and 60 open vein harvest (OVH).
There was a 7% conversion of EVH patients to OVH. EVH time was significantly longer
than OVH time (45 +/— 6.2 min vs. 31.1 +/— 6.5 min). However, morbidity was signifi-
cantly lower, with reduced pain and better cosmetic results following EVH. Bleeding com-
plications were less frequent and no local infections or wound complications were observed
in the EVH group versus 1 1 (18%) cases in the OVH group. Two OVH cases (3.6%) were
readmitted for wound debridement. These authors concluded EVH is a safe and efficient
technique for patients who require CABG. While these studies performed in patients who
underwent EVH for CABG are convincing, EVH continues to be avoided by many vascular
surgeons. Increased cost, increased OR time, and the possibility of injury of a long conduit
for infrainguinal bypass are continuing concerns that have prevented widespread acceptance
of this technique by vascular surgeons who perform limb salvage procedures.

The pivotal issue that must be considered is the risk of endothelial injury and
subsequent intimal hyperplasia in these patients, leading to failure of the grafts. There
are obvious concerns that traction during adventitial dissection and manipulation of the
vein will result in intimal trauma and secondary changes affecting long-term patency. ■§

Several studies have investigated the effects of minimally invasive GSV harvest on g

morphology and function of the veins and failed to show a deleterious effect. Meldrum- »

Hanna demonstrated preservation of intima by scanning electron microscopy (SEM) and c

observed a 93% graft patency rate at 10 days after surgery (25). In another histological <

study of vein samples performed using light microscopy and SEM, preservation of >9

endothelial architecture, cell cohesion, and intercellular junctional gaps was observed J

(41). Functional studies from the Mayo Clinic of porcine venous endothelium confirmed, °

that endothelial release of vasoactive substances after endoscopic harvesting is similar to |

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PERFORATOR VEIN SURGERY AND VEIN HARVESTS 537

that after the traditional extended incision technique, and microscopy confirmed similar
histology and integrity (27).

Vein injuries requiring vein patch angioplasty occurred in two patients in the series of
Jordan et al. (22) One of the veins thrombosed at 5 months after operation. In that series,
primary patency rate of femoropopliteal bypasses was 63% and secondary patency rate
was 83% at a mean follow-up of 7.9 months. Long-term follow-up with 5 year results by
the same group was recently reported (40,42). The 1-, 3-, and 5-year secondary patency
rates of the infrainguinal vein grafts were 85, 74, and 68%. Of the 30 failed grafts, 7 (4%)
failed in the first month related to inadequate runoff (4), cardiac instability (2), and an
additional surgical procedure (1). Twenty-three late graft failures were salvaged in all but
one patient. Five-year limb salvage was 89%.

With improvement in technology and decrease in cost, EVH may regain popularity
among vascular surgeons and will be used even more frequently by our cardiac colleagues.
Available evidence supports decreased wound complications with EVH compared to open
harvesting techniques in patients who undergo CABG, and long-term graft patency and
limb salvage in those who underwent EVH for lower extremity bypass does not appear to
be adversely affected. Long-term randomized study of infrainguinal reconstructions with
EVH and open techniques will still be required to provide the final answer.

Prevention of systemic and local complications of SEPS and EVH requires careful pa-
tient selection, adherence to open and minimally invasive surgical techniques, and appro-
priate postoperative care. Early recognition and correct management of complications will
result in improved outcome and will avert potentially severe secondary complications,
disability, limb loss, or even death. These minimally invasive procedures are here to stay, and
with progress in technology, they have an excellent chance to be used more frequently in the
future in cardiovascular surgery.

REFERENCES

1. Hauer G. Endoscopic subfascial discussion of perforating veins — Preliminary report. [German].
Vasa 1985; 14:59-61.

2. Fischer R. Surgical treatment of varicose veins: Endoscopic treatment of incompetent Cockett
veins. Phlebologie 1989; 1040-1041.

3. Jugenheimer M. Junginger T. Endoscopic subfascial sectioning of incompetent perforating
veins in treatment of primary varicosis. World J Surg 1992; 16:971-975.

4. O'Donnell TF. Surgical treatment of incompetent communicating veins. In: Atlas of Venous
Surgery. Philadelphia: Saunders, 1992:111-124.

5. Conrad P. Endoscopic exploration of the subfascial space of the lower leg with perforator
interruption using laparoscopic equipment: A preliminary report. Phlebology 1994; 9:154-157.

6. Wittens CH, Pierik RG, van Urk H. The surgical treatment of incompetent perforating veins. ^
Eur J Vase Endovasc Surg 1995; 9:19-23. |

7. Gloviczki P, Cambria RA, Rhee RY, Canton LG, McKusick MA. Surgical technique and £
preliminary results of endoscopic subfascial division of perforating veins. J Vase Surg 1996; %
23:517-523. |

8. Bergan JJ, Murray J, Greason K. Subfascial endoscopic perforator vein surgery: A preliminary j
report. Ann Vase Surg 1996; 10:211-219. fj.

9. Gloviczki P, Bergan JJ, Menawat SS, Hobson RW, Kistner RL, Lawrence PF, et al. Safety, |
feasibility, and early efficacy of subfascial endoscopic perforator surgery: A preliminary report 2
from the North American registry. J Vase Surg 1997; 25:94-105. §

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

538 GLOVICZKI and KALRA

10. Pierik EG, Toonder IM, van Urk H, Wittens CH. Validation of duplex ultrasonography in
detecting competent and incompetent perforating veins in patients with venous ulceration of
the lower leg. J Vase Surg 1997; 26:49-52.

1 1. Pierik EG, van Urk H, Hop WC, Wittens CH. Endoscopic versus open subfascial division of
incompetent perforating veins in the treatment of venous leg ulceration: A randomized trial.
J Vase Surg 1997; 26:1049-1054.

12. Gloviczki P, Bergan JJ, Menawat SS. Hobson RW, Kistner RL, Lawrence PF, et al. Safety,
feasibility, and early efficacy of subfascial endoscopic perforator surgery: A preliminary report
from the North American registry. J Vase Surg 1997; 25:94-105.

13. Stuart WP, Adam DJ, Bradbury AW, Ruckley CV. Subfascial endoscopic perforator surgery
is associated with significantly less morbidity and shorter hospital stay than open operation
(Linton's procedure). Br J Surg 1997; 84:1364-1365.

14. Sparks SR, Ballard JL, Bergan JJ, Killeen JD. Early benefits of subfascial endoscopic per-
forator surgery (SEPS) in healing venous ulcers. Ann Vase Surg 1997; 11:367-373.

1 5. Iafrati MD, Welch HJ, O'Donnell TF Jr. Subfascial endoscopic perforator ligation: An analysis
of early clinical outcomes and cost. J Vase Surg 1997; 25:995-1000.

16. Rhodes JM, Gloviczki P, Canton LG, Rooke T, Lewis BD, Lindsey JR. Factors affecting
clinical outcome following endoscopic perforator vein ablation. J Vase Surg 1998; 176:162-167.

17. Murray JD, Bergan JJ, Riffenburgh RH. Development of open-scope subfascial perforating
vein surgery: Lessons learned from the first 67 patients. Ann Vase Surg 1999; 199:372-377.

18. Gloviczki P, Bergan JJ, Rhodes JM, Canton LG, Harmsen S, Ilstrup DM. Mid-term results of
endoscopic perforator vein interruption for chronic venous insufficiency: Lessons learned from
the North American subfascial endoscopic perforator surgery registry. The North American
Study Group. J Vase Surg 1999; 29:489-502.

19. Gloviczki P, Bergan JJ, Rhodes JM, Canton LG, Harmsen S, Ilstrup DM. Mid-term results of
endoscopic perforator vein interruption for chronic venous insufficiency: Lessons learned from
the North American subfascial endoscopic perforator surgery registry. The North American
Study Group. J Vase Surg 1999; 29:489-502.

20. Kalra M, Gloviczki PM, Noel AA, et al. Subfascial endoscopic perforator vein surgery (SEPS)
in patients with post-thrombotic venous insufficiency — Is it justified? Vase Endovasc Surg 2002;
36:41-50.

21. Kalra M, Gloviczki P. Subfascial endoscopic perforator vein surgery: Who benefits? Semin
Vase Surg 2002; 15:39-49.

22. Jordan WD, Voellinger DC, Schroeder PT, McDowell HA. Video-assisted saphenous vein
harvest: The evolution of a new technique. J Vase Surg 1997; 26:405^114.

23. Lumsden AB, Eaves FF, Ofenloch JC, Jordan WD. Subcutaneous, video-assisted saphenous
vein harvest: Report of the first 30 cases. Cardiovasc Surg 1996; 4:771-776.

24. Cable DG, Dearani JA. Endoscopic saphenous vein harvesting: minimally invasive video-
assisted saphenectomy. Ann Thorac Surg 1997; 64:1183-1185.

25. Meldrum-Hanna W, Ross D, Johnson D, Deal C. An improved technique for long saphenous
vein harvesting for coronary revascularizaiton. Ann Thorac Surg 1986; 42:90-92.

26. Cable DG, Dearani JA, Pfeifer EA, Daly RC, Schaff HV. Minimally invasive saphenous vein
harvesting: endothelial integrity and early clinical results. Ann Thorac Surg 1998; 66:139- j>
140. |

27. Cho JS, Gloviczki P. Techniques for harvesting the GSV. In: Whittemore AD, ed. Advances in js
Vascular Surgery. St. Louis: Mosby, 1998:171-140. Jf

28. Allen KB, Griffith GL, Heimansohn DA, et al. Endoscopic versus traditional saphenous vein ^
harvesting: A prospective randomized trial. Ann Thorac Surg 1998; 66:26-32. >9

29. Crouch JD, O'Hair DP, Keuler JP, et al. Open versus endoscopic saphenous vein harvesting: J
Wound complications and vein quality. Ann Thorac Surg 1999; 68:1513-1516. q

30. Kan CD, Luo CY, Yang YJ, et al. Endoscopic saphenous vein harvest decreases leg wound |
complication in coronary artery bypass grafting patients. J Card Surg 1999; 14:157-162. S

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

PERFORATOR VEIN SURGERY AND VEIN HARVESTS 539

3 1 . Puskas JD, Wright CE, Miller PIC, et al. A randomized trial of endoscopic versus open saphenous
vein harvest in coronary bypass surgery. Ann Thorac Surg 1999; 68:1509-1512.

32. Vitali RM, Reddy RC, Molinaro PJ, et al. Hemodynamic effects of carbon dioxide insufflation
during endoscopic vein harvesting. Ann Thorac Surg 2000; 70:1098-1099.

33. Meyer DM, Rogers TE, Jessen ME, et al. Histologic evidence of the safety of endoscopic
saphenous vein graft preparation. Ann Thorac Surg 2000; 70:487^491.

34. Alwari SJ, Samee M, Raju R, et al. Intercellular and vascular cell adhesion molecule levels in
endoscopic and open saphenous vein harvesting for coronary artery bypass surgery. Heart Surg
Forum 2000; 3:241-245.

35. Fabricius AM, Diegeler A, Doll N, et al. Minimally invasive saphenous vein harvesting tech-
niques: Morphology and postoperative outcome. Ann Thorac Surg 2000; 70:473^478.

36. Patel AN, Hebeler RF, Hamman BL, Hunnicutt C, Williams M, Liu L, Wood RE. Prospective
analysis of endoscopic vein harvesting. Am J Surg 2001; 182(6):716-719.

37. Bitondo JM, Daggett WM, Torchiana DF, Akins CW, Hilgenberg AD, Vlahakes GJ, Madsen
JC, MacGillivray TE, Agnihotri AK. Endoscopic versus open saphenous vein harvest: a com-
parison of postoperative wound complications. Ann Thorac Surg 2002; 73(2):523-528.

38. Mozes G, Gloviczki P, Menawat SS, Fisher DR, Carmichael SW, Kadar A. Surgical anatomy
for endoscopic subfasical division of perforating veins. J Vase Surg 1996; 24:800-808.

39. Geselschap JH, van Gent WB, Wittens CH. Complications in subfascial endoscopic perforating
vein surgery: A report of two cases. J Vase Surg 2001; 33:1108-1110.

40. Alcocer F, Jordan WD. Long-term follow-up of endoscopically harvested vein graft. In: Pearce
WE, Matsumura JS, Yao STJ, eds. Trends in Vascular Surgery. Chicago: Precept Press, 2002:
293-300.

41. Dimitri WR, West IE, Williams BT. A quick and atraumatic method of autologous vein har-
vesting using the subcutaneous extraluminal dissector. J Cardiovasc Surg 1987; 28:103-111.

42. Jordan WD Jr, Alcocer F, Voellinger DC, Wirthlin DJ. The durability of endoscopic saphenous
vein grafts: A 5-year observational study. J Vase Surg 2001; 34:434-439.

43. Bergan JJ, Ballard JL, Sparks S. In: Gloviczki P, Bergan JJ, eds. Atlas of Endoscopic Perforator
Vein Surgery. London: Springer-Verlag, 1998:141-149.

44. Gloviczki P, Canton LG, Cambria RA, Rhee RY. Subfascial endoscopic perforator vein surgery
with gas insufflation. In: Gloviczki P, Bergan JJ, eds. Atlas of Endoscopic Perforator Vein
Surgery. London: Springer-Verlag, 1998:125-138.

45. Cho JS, Gloviczki P. Techniques for harvesting the greater saphenous vein. In: Whittemore
AD, ed. Advanced in Vascular Surgery. St. Louis: Mosby, 1998:171-180.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

33

Complications of Venous Endovascular Lysis
and Stenting (Iliac, Subclavian)

Peter Neglen

River Oaks Hospital, Jackson, Mississippi, U.S.A.

Indications for percutaneous venous endovascular procedures are expanding and tech-
niques for venous interventions are developing rapidly. This chapter will deal mainly with
complications of procedures performed for acute or chronic obstruction of the venous
outflow of the upper and lower extremities. Complications related to endovascular pro-
cedures for malfunctioning dialysis accesses and interruption of the inferior vena cava are
discussed in Chapters 34 and 35, respectively.

The most frequent complications of endovascular venous procedures are related to
cannulation of the vessel, as with arterial percutaneous interventions. In addition, admin-
istration of lytic agents through inserted catheters may lead to local or systemic bleeding
complications, which mitigate potential benefits. Inadvertent events related to the venous
balloon angioplasty are infrequent. Contrary to angioplasty of the arteries, clinical rupture
of veins has to our knowledge not been reported. Dissection does not occur in veins,
but embolization secondary to balloon dilation in chronic venous disease or after outflow
thrombolysis, although extremely rare, may be of clinical importance in individual patients.
Complications may also be related to the stent. Unlike arterial obstruction, a stent is always ■a

required in treating venous obstruction because the fibrous and elastic venous lesion often |

recoils after dilation alone. After the insertion, there is a potential risk for migration of the as

stent proximally or distally, fracture of the stent material, in-stent restenosis, and early or c

late occlusion. Patency rates of stents placed in the venous system may vary depending upon <

the anatomic location, etiology of the obstruction, development rate of neointimal hyper- >9

plasia, magnitude of venous inflow, and presence of concomitant diseases. The significance J

of many of these factors for the patency of stents placed in the venous system is not fully «

known and warrants further investigation. |

541

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

542 NEGLEN

I. COMPLICATIONS RELATED TO VENOUS CANNULATION

Complications related to cannulation of the vein are inherent in all percutaneous proce-
dures. Although the formation of hematomas or pseudoaneurysms and development of
traumatic arteriovenous fistulas are rare (<0.5%) (1), the frequency is higher with increasing
use of larger percutaneous instruments and more intensive anticoagulation. The rate of
hematoma formation appears to be less frequent in venous than in arterial interventions,
probably owing to the lower venous pressure and the ease with which compression can
control venous cannulation-site bleeding. We have found that the insertion of collagen plugs
(VasoSeal, Datascope Corp., Montvale, NJ) at the end of the procedure effectively controls
the bleeding at the venous cannulation site within a minute in the majority of limbs without
later rebleeding and obviates prolonged manual compression.

Formation of traumatic arteriovenous fistulas or pseudoaneurysms occurs when the
artery beside the vein is inadvertently injured during cannulation. Ultrasound-guided can-
nulation, either by ultrasound-tipped cannulas or special ultrasound probes, makes multiple
blind attempts unnecessary, puncture of the posterior wall is avoided, and the risk of arterial
injury is substantially decreased. Trauma to the artery is rare when ultrasound-guided
cannulation of the proximal femoral or popliteal veins is performed, but it may still occur
despite ultrasound guidance when it is necessary to access the distal femoral vein, which is
often hidden behind the artery at this location.

Both pseudoaneurysms and traumatic arteriovenous fistulas can be diagnosed by du-
plex Doppler ultrasound scan with high accuracy. Spontaneous thrombosis may occur in as
many as one-third of pseudoaneurysms. If the pseudoaneurysm persists, the next step is to
attempt ultrasound-guided manual compression or compression by the ultrasound probe
(2,3). Closure of the pseudoaneurysm is successful in the majority of cases since the inflow
channel of the pseudoaneurysm is very narrow and easily compressed. Compression time is
relatively long, averaging 40 min; the pressure is painful for the patient; and larger aneu-
rysms (> 4 cm in diameter) are more difficult to close. A more expeditious and efficient
method is injection of thrombin under guidance of duplex scanning (4,5). The pseudoaneu-
rysm closes within minutes. The rate of inadvertent arterial injection in one study was 4%
(4). Intra-arterial injection must be avoided, since it results in limb-threatening ischemia. If
these interventions are unsuccessful, the patients must undergo surgery, usually with simple
closure of the arterial defect.

The majority of traumatic arteriovenous fistulas are closed by open surgery. Short stent
grafts may be inserted intra-arterially to close the connecting tract. Closure is easily
achieved, but the long term patency has not been assessed.

II. COMPLICATIONS RELATED TO LYSIS OF DEEP

VEIN THROMBOSIS |

The current standard of treatment for deep venous thrombosis remains anticoagulation »

therapy with intravenous unfractionated or subcutaneous low-molecular-weight heparin, c

followed by oral warfarin. It is clear, however, that anticoagulation therapy does not <

actively promote fibrinolysis. Consequently, this method leaves clot behind, which under- >9

goes retraction and organization and results in remaining obstruction and valve damage. J

Only 20-30% of iliac veins completely recanalize spontaneously following thrombosis, «

while the remaining veins recanalize partially and develop varying degrees of collateraliza- |

tion (6,7). When an early complete lysis of the clot can be achieved, improved patency of the @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

VENOUS ENDOVASCULAR LYSIS AND STENTING 543

thrombosed vein and better preservation of the valve function are observed (8-10). This is
the rationale behind actively attempting to dissolve a deep venous thrombus. Although no
randomized prospective studies have been performed, thrombolysis of acute deep venous
thrombosis (DVT) involving the iliofemoral and axillary-subclavian outflows is an attrac-
tive form of therapy. Results are similar to those achieved with surgical thrombectomy, a
procedure that is scorned in the United States but still practiced in parts of Europe (11). With
the introduction of the catheter-directed thrombolytic (CDT) techniques reported by Semba
and Dake in 1994 (12), a better degree of lysis has been obtained than with the previously
used systemic administration. By application of high concentrations of thrombolytic agent
to the clot itself, the lysis rates have been improved and the duration of treatment reduced.
The complication rate appears to be lower in CDT than that in systemic administration.

At present, the most common approach to the iliofemoral vein segment is ultrasound-
guided cannulation of the ipsilateral popliteal vein with the patient in the prone position.
When the popliteal vein is occluded, the ipsilateral posterior tibial vein is cannulated. The
internal jugular access or contralateral femoral approach has now been abandoned because
of the difficulty of pushing guidewires and catheters safely through a femoral thrombosis in a
retrograde direction against the direction of the venous valves. A sheath is inserted and
venography is performed, visualizing the distal extent of the clot. The thrombus is then
crossed by a guidewire and catheter, through which the proximal extent is shown by repeat
contrast dye injection. A multiside-hole catheter is then inserted into the thrombus, covering
its entire extent if possible. A thrombolytic agent is then infused continuously or in a pulse-
spray fashion. Concomitantly, heparin is infused intravenously. The extent of thrombolysis
is assessed every 12 h with venography performed through the sheath. It may be necessary to
reposition the infusion catheter into remaining thrombus or sometimes to gently macerate
the thrombus with a balloon catheter. Typically, the thrombolysis is continued to complete
lysis or terminated when no discernible progress is shown on consecutive venograms unless
the lysis is interrupted because of complications. The duration of infusion is often 24-48 h.
Heparin infusion is continued until therapeutic anticoagulation is achieved by oral warfarin
administration.

General complications of lytic therapy are outlined in Chapter 30. Two major reports
regarding specific complications of CDT of iliofemoral vein thrombosis have been pub-
lished. A review by Grossman and McPherson of 15 studies, including 263 patients who had
CDT using urokinase (234 patients) and rt-PA (29 patients) for iliofemoral thrombosis (13).
This study was compiled before publication of the multicenter National Registry study by
Mewissen et al., detailing the results of 473 patients treated by CDT with urokinase (8).
Metanalysis in the Grossman-McPherson study (13) found an overall major bleeding rate
(requiring transfusion) in 4.9% (13 of 263 patients). This rate is comparable to that reported
for conservative treatment with unfractionated heparin, approximately 5% (14). Almost
half the bleeding patients, however, were given rt-PA, resulting in bleeding rates for those ■§

treated by urokinase and rt-PA of 2.9% (7 of 234) and 21% (7 of 29), respectively. The 1

mortality was 0.4 % (1 of 263, owing to cardiac ischemia). Symptomatic pulmonary embo- a

lism (PE) was observed in 0.8% (2 of 263), but none was fatal. c

The prospective National Registry study (8) revealed a higher rate (11%) of major <

bleeding complications (4% at cannulation site, 1% spontaneous retroperitoneal bleeding, >9

3% gastrointestinal, genitourinary or musculoskeletal sites, 3% unknown sites) and a 16% J

rate of minor bleeding events (the majority at the cannulation site). Intracranial bleeding «

occurred in two patients (<1%, one fatal) and 1% had symptomatic pulmonary embolism |

(one fatal). Thus the mortality rate was the same in the two studies, i.e., 0.4%. The com- @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

544 NEGLEN

plication rate may appear high but should be compared to anticoagulation treatment alone,
which also carries risk. Major bleeding risk requiring blood transfusion appears higher with
urokinase CDT and even higher with rt-PA than with unfractionated heparin and certainly
than with low-molecular-weight heparin (0.8-2.4%) (15). The rates of symptomatic and
fatal pulmonary embolism are, however, higher with anticoagulation therapy alone (7.9 and
1.3%, respectively) (13).

Spontaneous intracranial hemorrhage is the most devastating and dreaded complica-
tion of thrombolytic therapy. Although it was not reported in the compiled study, it was
observed in the national registry in < 1 % . Since urokinase was abruptly removed from the
market in the United States in 1999, alteplase and reteplase replaced it and are now the
preferred drugs for noncoronary thrombolysis. Initially, no intracranial bleedings or strokes
were reported in the pilot studies. Rates of 0.8 and 1.4%, respectively, have been reported
with reteplase treatment for acute myocardial infarction (14). The optimal dose and the
effects of concomitant use of heparin in lysis of iliofemoral DVT were not fully defined
initially, which probably led to an increased rate of bleeding complications, including intra-
cerebral hemorrhage. It appears that the rate of bleeding complications is related to the
dosage (16). The continuous evaluation of adequate dosage has led to a substantial
reduction in the dosage of both rt-PA and concomitant heparin. The initial recommenda-
tion, for example, was to administer reteplase at 1 U/h to treat venous occlusion (16). With
increased experience, most centers using this agent have successively decreased the dosage.
In our service we now use <0.4 U/h of reteplase for CDT of deep vein thrombosis with
apparently the same efficacy but fewer bleeding events. A similar development has occurred
with alteplase. Just as we are learning how to use rt-PA, urokinase is apparently making a
comeback in the United States. Its future role is not known at this time.

Bleeding complications related to the catheter insertion site can be minimized by careful
needle access of the vein under ultrasound guidance to avoid inadvertent puncture of the
adjacent vessels, multiple attempts, and puncture through the posterior wall. The systemic
effect of the thrombolytic agents, which results in distant bleeding, is more difficult to avoid.
It is important that an optimal dose with maximal efficacy and minimal bleeding rate be
found for each thrombolytic agent and concomitant heparin. The patient must be metic-
ulously supervised and frequently assessed for bleeding, including measurements of regular
hemoglobin levels, hematocrit, platelet count, prothrombin time and partial prothrombin
time. PTT is used to monitor intravenous unfractionated heparin infusion, but there is no
reliable marker for the thrombolytic agent. The current recommendation is that the blood
fibrinogen level be measured to assess the lytic state. When the level falls to 100-1 50 mg/dL,
the dose may be reduced; below the level of 100 mg/dL, it may be prudent to stop the infusion
altogether. Unfortunately, major bleeding complications do not appear to be directly related
to the fibrinogen level (16-18). A low level does not necessarily lead to increased bleeding,
and a major bleeding event may occur despite normal or near-normal fibrinogen levels. ■§

Therefore clinical surveillance is crucial. In addition, careful selection of patients for CDT g

treatment is important. It appears that elderly patients, especially those with arterial hyper- »

tension, are more prone to develop cerebral hemorrhage. The patient must also have a c

reasonable life expectancy, since the main impetus to perform CDT for deep venous throm- <

bosis is the avoidance of future postthrombotic disease. >9

CDT for iliofemoral venous thrombosis appears to be a safe procedure with good ^

short-term results when applied in patients with relatively acute clot formation (< 10 °

days). If the benefits are to surpass the risks, careful assessment of the thrombotic disease |

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

VENOUS ENDOVASCULAR LYSIS AND STENTING 545

per se, appropriate dosage of the lytic agent administered, and evaluation of the patient's
concomitant diseases are essential.

III. COMPLICATIONS RELATED TO STENTING OF ILIOFEMORAL
VENOUS OUTFLOW CHANNEL

Experience from stenting of the venous system is collected mainly from treatment of large
vein obstruction. As compared to more proximal veins, results of balloon angioplasty and
stenting of femoropopliteal or brachioaxillary veins have not been encouraging. Both
location and etiology of the obstruction appear to be important. The most commonly seen
acute obstruction of the iliofemoral outflow channel is acute deep veinous thrombosis. Even
after open venous thrombectomy or catheter-directed thrombolysis (CDT), a chronic ve-
nous stenosis is often "uncovered." The rate of rethrombosis is apparently less if this
stenosis is controlled (8). The most commonly seen chronic obstruction is that due to
postthrombotic disease. Following thrombosis that has been conservatively treated with
anticoagulation alone, as many as 70-80% of iliac veins remain only partly recanalized or
occluded, with varying degrees of collateral formation (6,7). Less frequently, benign or
malignant masses, retroperitoneal fibrosis, iatrogenic injury, irradiation, cysts, and aneu-
rysms may also cause blockage of the iliac vein. A "primary" nonthrombotic iliac vein
obstruction [May-Thurner syndrome (19) or iliac compression syndrome (20)] has also been
described and is probably more common than is recognized.

Balloon angioplasty alone of the iliofemoral vein was found to be insufficient early in
the development of this technique. This procedure leads to immediate complete recoil in the
majority of limbs and results in early restenosis. Stenting is therefore advised in all cases (21—
24). Depending on the location of the obstruction, ultrasound-guided access to the vein
below the obstruction is obtained. A guidewire is inserted, followed by a sheath, through
which catheters, balloons, and stents are inserted. The guidewire must cross the obstruction
in order for stenting to proceed, which may be difficult when complete obstruction is
present. The details of the technique are described elsewhere (24-26). At least in the medium
term, there appears to be good symptom relief following the intervention (27).

The nonthrombotic complication rate related to the endovascular intervention is
presently minimal and comprises mostly cannulation site hematoma, although a few cases
of retroperitoneal hematoma requiring blood transfusions have been described (24,28). As
previously described, the utilization of ultrasound-guided cannulation and closure with
collagen plugs have largely abolished these problems, reducing the nonthrombotic compli-
cation rate from 3 to <1%. The mortality has been zero. Iliac vein balloon dilation and
stenting of an iliofemoral and caval vein obstruction is a safe, minimally invasive method
with a low complication rate and no mortality. ■§

Data regarding early rethrombosis (<30 days) after iliofemoral stenting are sparse. The g

rate was found to be 11 % in the Creighton University experience following stenting after »

thrombolysis of an acute DVT (29) and 15% in the National Registry study (8). The early c

thrombosis rate was found to be lower (4%) when stenting was performed for chronic, <

nonmalignant iliac vein obstruction without preceding thrombolysis (25). All of the latter >9

postoperative obstructions occurred in patients with chronic postthrombotic disease (8% J

thrombosis rate), while none occurred in nonposthrombotic limbs (0% thrombosis rate) °

with primary disease. Early failures appear more common with stents placed across |

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

546 NEGLEN

complete occlusions than across stenoses (30). The occlusions appeared to be related to
limbs with severe nonyielding stenoses or complete obstructions that could not be fully
dilated. Thrombolysis of the newly formed clot may be attempted in initially technically
successful limbs to reveal and treat unknown additional obstructions. Midterm patency
apparently depends on the etiology of the obstruction, but results are generally encouraging.
One-year primary patency after CDT of an acute DVT and iliofemoral stenting is 74-80%
(8,29). Primary, assisted-primary, and secondary patency rates at 3 years after stenting for
chronic nonmalignant obstructions were 75, 92, and 93%, respectively. Similarly, the rates
at 32 months for limbs with nonthrombotic (primary) etiology of the obstruction was
significantly better than for postthrombotic limbs (84 and 73%, 100 and 86%, and 100 and
86%, respectively) (25,31). Typically, the stented limbs that occluded were characterized by
postthrombotic etiology, stent lengths >13 cm, and stent placement below the inguinal
ligament. At the time of initial balloon angioplasty and stenting, the majority of the iliac
veins were either completely occluded or had tight, nonyielding stenoses. The reason for late
occlusions is probably multifactorial but appears to result mainly from a combination of
development of in-stent restenosis and recurrent attacks of thrombosis in a patient with
chronic postthrombotic disease. Iliofemoral occlusion caused by malignant pelvic tumors
has poorer patency rates, with a 1-3 year secondary patency rate of 64-68% (32,33).
Thrombolysis may be attempted to disobliterate the occluded stent vein and, if possible,
treat underlying restenosis.

In-stent restenosis is a serious long-term complication, which may result in recur-
rence of symptoms and potential occlusion of the stent; further interventions may be
required (Figs. 1 and 2). The nature of in-stent restenosis, rate of development, preven-
tion, and treatment for stents placed in the venous vasculature are poorly understood.
Symptomatic restenosis was observed in 9% of limbs 6-12 months after initial treatment in
the Creighton University experience (29) and responded well after angioplasty. Based on
venographic follow-up and survival analysis by the Kaplan-Meier method, only 23 % of
limbs remained hyperplasia-free at 3 years follow-up (31). However, most stents with
hyperplasia had only minimal development. At the same time interval, 61 and 15% of limbs
had >20 and >50% diameter reduction, respectively. Probably a stenosis does not become
symptomatic until it exceeds 50% narrowing. Increased hyperplasia formation has been
observed in stents crossing thrombotic obstructions, in long stents extending below the
inguinal ligament, and in the presence of hypercoagulable states. The observed in-stent
restenosis may be due to true neointimal hyperplasia or a thrombotic lining of the stent. In
the majority of narrowed stents, balloon angioplasty reveals a very tough stenosis, which is
often nonyielding and largely recoiling, suggesting true neointimal hyperplasia formation.
Despite this observation, balloon angioplasty alone will often improve outflow sufficiently
to result in improvement of the patients' symptoms. The efficacy of covering the in-stent
restenosis with an additional stent inside the previously inserted stent has not been ■§

adequately assessed. The role of drug therapy — e.g., with clopidogrel, placement of drug- s

eluting stents (now under development), or primary vascular brachytherapy of stents as

inserted in the venous system — is unknown. Experience from treatment of in-stent restenosis c

of coronary and noncoronary arterial stents cannot necessarily be extrapolated to the <

venous system. >5

Short migration of Wallstent endoprostheses placed over short and tight venous ste- ^

nosis of the left common iliac vein at the vessel crossing close to the junction of the IVC (iliac °

vein compression syndrome) has been described (24,34). When the end of the stent was |

placed just at the confluence or slightly into the IVC, antegrade migration toward the @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

VENOUS ENDOVASCULAR LYSIS AND STENTING

547

Figure 1 In-stent restenosis revealed by a transfemoral venogram in two patients with stents
placed in the iliac veins because of postthrombotic obstruction. The arrows outline the outer limits of
the Wallstent.

IVC was observed in 3 of 18 limbs (17%) and retrograde migration with recurrence of ste-
nosis was found in 9 of 25 limbs (36%). The stent appears to be "squeezed" proximally or
more commonly distally by the external compression (Fig. 3). This may be a result of the
spiral configuration and the inherent, somewhat weaker radial force of the self-expanding
Wallstent. The "recurrence" of the previously treated stenosis was easily corrected by
placement of an additional stent. To avoid this migration, we have suggested that the stent
should be placed well into the IVC in this type of stenosis when Wallstent endoprostheses
are used. This IVC placement raises concern for risk of occlusion of the contralateral iliac
vein in the long-term, although the stent does not appear to significantly impair the flow
from the contralateral limb. The few cases of contralateral limb DVT observed appear to be
caused by recurrent attacks of thrombosis. Longer follow-up is necessary, however, to fully
assess outcome.

The Stanford group found that, over a 5-year period, major misplacement or migration
occurred in 2.5% of 801 noncoronary vascular stent placements, the majority in the venous

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

548

NEGLEN

Figure 2 An image obtained by intravascular ultrasound (IVUS), distinctly showing the in-stent
restenosis. Seen is the remaining cross area lumen (A) and the stent area (B); the center dark circle is
the IVUS catheter. Commonly, the in-stent stenosis shown on IVUS is more severe than the venogram
suggests.

system (35). "Lost" Palmaz stents could be managed by using a balloon catheter to repo-
sition the stent, either in the intended vessel (not considered as misplaced) or in a stable
alternate location (13 of 16). If the stent could be withdrawn into the sheath, it was removed
percutaneously (2 of 16). One stent was surgically removed from the right ventricle. Eleven
Wallstents were managed primarily with use of different snare techniques, well described in
detail in the report. Nine stents were removed percutaneously and two were removed sur-
gically after being repositioned percutaneously in the femoral vein and artery, respectively.
Care was taken to avoid injury to the vessel wall during the procedure and the patients were
fully anticoagulated. Investigations were performed in 13 patients, 2-65 months later indi-
cating full patency after retrieval. To avoid gross misplacement/migration during inser-
tion of venous stents and to facilitate retrieval, it is vital to maintain wire access across
the stent throughout the entire procedure and to place the wire through the IVC and SVC
to prevent stent migration into the heart and pulmonary arteries. The stent and balloon
size should be carefully chosen to avoid undersizing the stent. A large stent (14—16 mm
diameter) is recommended for the iliofemoral venous segment. The vein seems to accept

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

VENOUS ENDOVASCULAR LYSIS AND STENTING

549

Figure 3 Two examples of recurrent stenosis (arrows) proximal to the stents previously inserted in
the iliac vein. These stents were inserted flush with the IVC and appear to have been squeezed distally
by a severe iliocaval junction stenosis (From Ref. 24).

extensive dilation without clinical rupture in contrast to the artery. No clinical rupture, of
the vein has been reported to date, even when a total occlusion is recanalized and dilated
up to 14-16 mm width.

Fracture of the inserted stent may occur owing to external pressure or metal fatigue
of the stent. To our knowledge, fracture has not been described for stents placed below the
inguinal ligament in the lower extremity venous outflow tract. There is concern that move-
ment of the hip may compress a stent placed under the inguinal ligament. We have some-
times found it difficult to fully expand a stent under the inguinal ligament, but we have not
observed any "crushed" stent in that position so far. A higher in-stent restenosis rate has
been observed when the stent reaches below the inguinal ligament. It is, however, rare to
have a stent placed for a localized stenosis under the ligament. When placed in this position,
the stent is usually a distal extension of a proximal iliac stent. The problems with these long
stents may be related less to external compression and more to the significant over-
representation of postthrombotic etiology of the obstruction in these limbs (31).

There is concern that the enlarged uterus during pregnancy may compress an IVC
filter placed in the infrarenal position (36). Similarly, a stent in the iliocaval vein could be
compressed. In one report, a Palmaz stent placed in the iliocaval junction was crushed
during pregnancy and resulted in thrombosis (34). The extent of this problem is not known.
Placement of stent in women of childbearing age to treat iliac compression syndrome
is increasingly performed. Some caution may be justified until further data have been
collected.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

550 NEGLEN

Complications may occur due to the perioperative anticoagulation. Currently, we do
not routinely anticoagulate the patient during diagnostic procedures, including intra-
vascular ultrasound. All patients receive 2500 U of dalteparin, subcutaneously before sur-
gery. When it is decided to continue with balloon angioplasty and stenting during the
procedure, 5000 U of heparin and 30 mg of ketorolac are administered intravenously. Post-
operatively, a foot compression device is applied, and dalteparin (2500 U) is given sub-
cutaneously. During postoperative hospitalization, an additional dose of dalteparin (5000 U
SQ) and a ketorolac injection are given the morning before discharge. Low-dose aspirin
(81 mg PO) daily is started immediately after surgery and continued. Most patients have no
additional anticoagulation. Only patients already on warfarin preoperatively are given
warfarin postoperatively. These are a minority of patients with prior recurrent DVT and/or
thrombophilia, which makes lifelong anticoagulation necessary. When warfarin is discon-
tinued prior to surgery, dalteparin 5000 U is injected subcutaneously during the days
warfarin is discontinued. This regimen appears to result in minimal bleeding problems and
still prevent early thrombosis.

Venous balloon angioplasty and stenting of the IVC and iliac veins appears to be a
safe, relatively simple, and efficient method to treat iliocaval vein obstruction in the mid-
term. An immediate or late failure of the procedure does not preclude later open surgery to
correct the obstruction. Further studies are necessary to elucidate long-term clinical results
and potential risks.

IV. COMPLICATIONS RELATED TO STENTING OF THE
SUBCLAVIAN VEIN

Management of venous stenosis of the upper extremity owing to complications of place-
ment of central venous access catheters or related to arteriovenous fistulas for dialysis
access are described in Chapter 34. Acute obstruction of the venous outflow by a DVT
results in sudden symptoms with swelling, stasis, and pain. Patients may also complain of
intermittent swelling, bluish discoloration and pain on exertion owing to a slowly devel-
oping stenosis. The unique anatomical situation in the thoracic outlet makes treatment of
venous obstruction of the upper extremity radically different from that of the lower extrem-
ity. Venous outflow obstruction in the arm is usually associated with subclavian vein com-
pression in the thoracic outlet. Infrequently, it is caused by external compression or direct
ingrowth by malignant and benign tumors. The entire neurovascular bundle may be com-
pressed to varying degrees between the first rib and the clavicle. Abnormal soft tissue,
muscles, and bands of fascia may also contribute. When a venous stenosis in this area is
only balloon-dilated or stented, the vein will still be subjected to extrinsic compression
between the clavicle and the first rib, and to flexion forces when the arm is abducted. This ■§

may lead to structural failure, restenosis, and occlusion. |

With acute nonaccess related subclavian DVT, a catheter-directed thrombolysis is rec- a

ommended in a fashion similar to that for acute DVT in the lower extremity. Complications c

due to thrombolysis are the same as described above. A successful lysis will often reveal an <

underlying stenosis due to thoracic outlet compression of the vein. It appears logical to >9

alleviate the compression by removing the first rib (37,38). However, there is often a J

residual significant venous stenosis observed after release, even after early venous balloon «

angioplasty has been performed. This residual stenosis can then be corrected by insertion of |

a stent to ensure patency and prevent early rethrombosis. Although there is an increasing @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

VENOUS ENDOVASCULAR LYSIS AND STENTING 551

consensus that upper limbs with subclavian vein stenosis should be treated by first-rib
resection, the issues of timing of balloon dilation of any remaining stenosis and whether a
stent should be routinely inserted are still controversial (39,40). There is strong evidence,
however, that venous balloon angioplasty alone does not always suffice in the thoracic
outflow channel and that stent insertion is necessary (41). A stent should never be inserted
into the thoracic outlet without decompressive surgery. The stent will inevitably fracture
or be compressed by the strong extrinsic forces and ultimately fail (Fig. 4) (41-44). The
decompression appears to prevent deleterious external stent compression but may not
decrease the possibility of stent fracture due to metal fatigue following repeated stent
flexion. Therefore the use of a rigid stent in the thoracic outlet is not advised. We favor
the flexible, self-expanding type of stent. These stents do not appear to fracture because of
repeated arm and neck movement so long as decompressive surgery has been performed.
Providing thoracic outlet decompression has been performed, a remaining subclavian
venous stenosis of less than 50% is apparently acceptable. Kreienberg et al. (39) have
reported that in all patients with less than 50% residual stenosis after percutaneous trans-
luminal angioplasty (PTA), the subclavian vein was patent and the patient asymptomatic
at a mean follow-up of 4 years. Conversely, almost one-third of stented patients (5 of 14)
had occluded the stent during a similar follow-up. The stents were inserted in the subcla-
vian vein because of a residual stenosis of more than 50% after PTA. This result could be
blamed in part on the presence of thrombophilia in three patients and extensive brachio-
axillary-subclavian venous involvement in all patients. The long-term in-stent restenosis
rate has not been established and consequences of intimal reaction to the stents are un-
known. On the other hand, patients with >50% stenosis remaining after release and bal-
loon angioplasty appear to be at increased risk for rethrombosis. This finding, coupled
with the fact that prognosis of a stent thrombosis appears not to be no worse than throm-
bosis of the vein itself, causes us to favor the use of stents in residual stenosis of >50% of
the subclavian vein as measured by intravascular ultrasound providing that a first-rib re-
section has been performed.

V. COMPLICATIONS RELATED TO STENTING OF THE SUPERIOR
VENA CAVA

By far the most common cause of SVC syndrome is malignancy (85-97% of cases) (45).
The combination of thrombolysis, angioplasty, and stent insertion is the current "method
of choice" for malignant obstruction, with a good technical success rate, primary patency
rates of 50-100%, and improvement in secondary patency by subsequent procedures. The
important observation is that all patients were symptom-free within 48 h of treatment and

90% remained symptom-free until death (usually within 1 year) (46). When treatment is ■§

performed for a benign lesion (mainly catheter — related obstructions), a secondary pat- g

ency rate is 85% with relatively short follow-up (mean 17 months) in these patients with »

longer life expectancy (47). Although stents placed inside the thorax should be free from c

external forces, "crushing" of stents placed in the left brachiocephalic vein between the <

manubrium and the aortic arch has been described (48). Surgical bypass may be a better >9

alternative for patients with benign obstruction. J

Complications are similar to those described with previous venous procedures (a rate «

of 3.2% minor and 7.8% major complications) (49). Most complications are related to |

restenosis, early or late occlusion, stent displacement, and problems inherent in anti- @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

•5

Figure 4 A 28-year-old woman with a subclavian vein thrombosis undergoing thrombolysis reveal-
ing a stenosis, which was successfully balloon-dilated and stented (a). No surgical decompression was
performed. Control venogram 1 year later demonstrated a 50% stenosis (b). No additional treatment
was performed. Follow-up with venogram another year later showed stent fracture and recurrent total
occlusion by thrombosis (c)(From Ref. 44).

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

VENOUS ENDOVASCULAR LYSIS AND STENTING 553

coagulation and thrombolysis. Complications specific to the SVC procedure include the
infrequently seen mediastinal hematoma, hemopericardium with cardiac tamponade, and
pulmonary edema following the increased venous return after disobliteration. If the patient
collapses after the procedure, immediate ultrasound or thoracic CT/CT pulmonary angi-
ography is recommended to detect these complications.

Undoubtedly, venous stenting appears to be a promising treatment, but some caveats
are prudent. In comparison to coronary and arterial angioplasty and stenting, there is little
information regarding the development of neointimal hyperplasia, long-term patency, and
potential long-term risks with stent placement in the venous system. The technology is
relatively recent; thus the follow-up period is limited. Monitoring for several more years is
necessary to assess the efficacy and safety of this therapeutic modality in venous disease.

REFERENCES

1. Fellmeth BD, Roberts AC, Bookstein JJ, Freischlag JA, Forsythe JR, Buckner NK, Hye RJ.
Postangiographic femoral artery injuries: Nonsurgical repair with US-guided compression.
Radiology 1991; 178:671-675.

2. Dean SM, Olin JW, Piedmonte M, Grubb M, Young JR. Ultrasound-guided compression clo-
sure of postcatherization pseudoaneurysm during concurrent anticoagulation: A review of
seventy-seven patients. J Vase Surg 1996; 23:28-35.

3. Sorrell KA, Feinberg RL, Wheeler JR, Gregory RT, Snyder SO, Gayle RG, Parent NF III.
Color-flow duplex-directed manual occlusion of femoral false aneurysms. J Vase Surg 1993;
17:571-577.

4. Khoury M, Alanna R, Greene K, Rama K, Colaiuta E, Flynn L, Berg R. Duplex scanning-
guided thrombin injection for the treatment of iatrogenic pseudoaneurysms. J Vase Surg 2002;
35:517-521.

5. Liau C-S, Ho F-M, Chen M-F, Lee Y-T. Treatment of iatrogenic femoral artery pseudo-
aneurysm with percutaneous thrombin injection. J Vase Surg 1997; 26:18-23.

6. Mavor GE, Galloway JMD. Iliofemoral venous thrombosis: Pathological considerations and
surgical management. Br J Surg 1969;56:45-59.

7. Plate G, Akesson H, Einarsson E, Ohlin P, Eklof B. Long-term results of venous thrombectomy
combined with a temporary arterio-venous fistula. Eur J Vase Surg 1990; 4: 483-489.

8. Mewissen MW, Seabrook GR, Meissner MH, Cynamon J, Labrapoulos N, Haughton SH.
Catheter-directed thrombolysis for lower extremity deep venous thrombosis: Report of a
national multicenter registry. Radiology 1999; 211:39-49.

9. Comerota AJ, Aldridge SC. Thrombolytic therapy for acute deep vein thrombosis. Semin Vase
Surg 1992; 5:76-81.

10. Meissner MH, Manzo RA, Bergelin RO, Markel A, Strandness DE Jr. Deep venous in-
sufficiency: The relationship between lysis and subsequent reflux. J Vase Surg 1993; 18:596-

608. ,j

1 1 . Eklof B, Kistner RL. Is there a role for thrombectomy in iliofemoral venous thrombosis? Semin §
Vase Surg 1996; 9:34-^5. 8

12. Semba CP, Dake MD. Iliofemoral deep venous thrombosis: Aggressive therapy with catheter- •§,
directed thrombolysis. Radiology 1994; 191:487-494. a

13. McPherson S, Grossman C. Safety and efficacy of catheter-directed thrombolysis for iliofemoral u
venous thrombosis. AJR 1999; 172:667-672. §

14. INJECT Study Group. Randomised, double-blind comparison of reteplase double-bolus 3
administration with streptokinase in acute myocardial infarction (INJECT): Trial to investigate 2
equivalence. Lancet 1995; 346:329-336. 1

15. Siragusa S, Cosmi B, Piovella F, Hirsh J, Ginsberg JS. Low-molecular-weight heparins and ©

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

554 NEGLEN

unfractionated heparin in the treatment of patients with acute venous thromboembolism:
Results of a meta-analysis. Am J Med 1996; 100:269-277.

16. Ouriel K, Katzen B, Mewissen M, Flick P, Clair DG, Benenati J, McNamara TO, Gibbens D.
Reteplase in the treatment of peripheral arterial and venous occlusions: A pilot study. J Vase
Intervent Radiol 2000; 11:849-854.

17. Ouriel K, Veith FJ, Sasahara AA. A comparison of recombinant urokinase with vascular surgery
as initial treatment for acute arterial occlusion of the legs. N Engl J Med 1998; 338: 1105-1111.

18. Ouriel K, Kandarpa K, Schuerr DM, Hultquist M, Hodkinson G, Wallin B. Prourokinase vs.
urokinase for revanalization of peripheral occlusions, safety and efficacy: The PURPOSE Trial.
J Vase Intervent Radiol 1999; 10:1083-1091.

19. May R, Thurner J. The cause of the predominantly sinistral occurrence of thrombosis of the
pelvic veins. Angiology 1957; 8:419^128.

20. Cockett FB, Thomas ML. The iliac compression syndrome. Br J Surg 1965; 52:816-821.

21. Neglen P, Al-Hassan HKh, Endrys J, Nazzal MMS, Christenson JT. Eklof B. Iliofemoral
venous thrombectomy followed by percutaneous closure of the temporary arteriovenous
fistula. Surgery 1991; 110:493-499.

22. Wisselink W, Money SR, Becker MO, Rice KL, Ramee SR, White CJ, Kazmier FJ, Hollier LH.
Comparison of operative reconstruction and percutaneous balloon dilatation for central venous
obstruction. Am J Surg 1993; 166:200-205.

23. Marzo KP, Schwartz R, Glanz S. Early restenosis following percutaneous transluminal balloon
angioplasty for the treatment of the superior vena caval syndrome due to pacemaker-induced
stenosis. Cathet Cardiovasc Diagn 1995; 36:128-131.

24. Neglen P, Raju S. Balloon dilation and stenting of chronic iliac vein obstruction: Technical
aspects and early clinical outcome. J Endovasc Ther 2000; 7:79-91.

25. Neglen P, Berry MA, Raju S. Endovascular surgery in the treatment of chronic primary and
post-thrombotic iliac vein obstruction. Eur J Vase Endovasc Surg 2000; 20:560-571.

26. O'Sullivan GJ, Semba CP, Bittner CA, Kee ST, Razavi MK, Sze DY, Dake MD. Endovascular
management of iliac vein compression (May-Thurner) syndrome. J Vase Intervent Radiol 2000;
11:823-836.

27. Raju S, Owen S Jr, Neglen P. The clinical impact of iliac venous stents in the management of
chronic venous insufficiency. J Vase Surg 2002; 35:8-15.

28. Hurst DR, Forauer AR, Bloom JR, Greenfield LJ, Wakefield TW. Diagnosis and endovascular
treatment of iliocaval compression syndrome. J Vase Surg 2001; 34:106-113.

29. Thorpe PE. Endovascular therapy for chronic venous obstruction. In: Ballard JL, Bergan JJ,
eds. Chronic Venous Insufficiency. New York: Springer, 1999: 179-219.

30. Nazarian GK, Austin WR, Wegryn SA, Bjarnason H, Stackhouse DJ, Castaneda-Zuniga WR,
Hunter DW. Venous recanalization by metallic stents after failure of balloon angioplasty or
surgery: Four-year experience. Cardiovasc Intervent Radiol 1996; 19:227-233.

31. Neglen P, Raju S. In-stent restenosis in stents placed in the lower extremity venous outflow
tract. J Vase Surg 2003. In print.

32. Nazarian GK, Bjarnason H, Dietz CA Jr, Bernades CA, Hunter DW. Iliofemoral venous ste-
nosis: Effectiveness of treatment with metallic endovascular stents. Radiology 1996; 200:193-

199. 1

33. Carlson JW, Nazarian GK, Hartenbach E, Carter JR, Dusenbery KE, Fowler JM, Hunter <S
DW, Adcock LL, Twiggs LB, Carson LF. Management of pelvic venous stenosis with Jj
intravascular steel stents. Gynecol Oncol 1995; 56:362-369. £

34. Juhan C, Hartung O, Alimi Y, Barthelmy P, Valerio N, Portier F. Treatment of nonmalignant ^
obstructive iliocaval lesions by stent placement: Mid-term results. Ann Vase Surg 2001; 15:227- >9
232. |

35. Slonim SM, Dake MD, Razavi MK, Kee ST, Samuels SL, Rhee JS, Semba CP. Management &
of misplaced or migrated endovascular stents. J Vase Intervent Radiol 1999; 10:851-859. |

36. Greenfield LJ, Proctor MC. Suprarenal filter placement. J Vase Surg 1998; 28:432-438. 2

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

VENOUS ENDOVASCULAR LYSIS AND STENTING 555

37. Molina JE. Surgery for effort thrombosis of the subclavian vein. J Cardiovasc Surg 1992;
103:341-346.

38. Molina JE. Letter to the editor. J Vase Surg 2001; 33:662-663.

39. Kreienberg PB, Chang BB, Darling RC HI, Roddy SP, Paty PS, Lloyd WE, Cohen D,
Stainken B, Shah DM. Long-term results in patients treated with thrombolysis, thoracic inlet
decompression, and subclavian vein stenting for Paget-Schroetter syndrome. J Vase Surg 2001;
33(suppl 2):S100-S105.

40. Rutherford RB. Primary subclavian-axillary vein thrombosis: The relative roles of
thrombolysis, percutaneous angioplasty, stents, and surgery. Sem Vase Surg 1998; 11:91-95.

41. Meier GH, Pollak JS, Rosenblatt M, Dickey KW, Gusberg RJ. Initial experience with venous
stents in exertional axillary-subclavian vein thrombosis. J Vase Surg 1996; 24:974-983.

42. Phipp LH, Scott DJA, Kessel D, Robertson I. Subclavian stents and stent-grafts: Cause for
concern? J Endovasc Surg 1999; 6:223-226.

43. Dowling R, Mitchell P, Cox GS, Thomson KR. Complication of a venous Wallstent. Australas
Radiol 1999; 43:246-248.

44. Maintz D, Landwehr P, Gawenda M, Lackner K. Failure of Wallstents in the subclavian vein
due to stent damage. Clin Imaging 2001; 25:133-137.

45. Schindler N, Vogelzang RL. Superior vena cava syndrome: Experience with endovascular
stents and surgical therapy. Surg Clin North Am 1999; 79:683-694.

46. Nicholson AA, Ettles DF, Arnold A, Greenstone M, Dyet JF. Treatment of malignant superior
vena cava obstruction. Metal stents or radiation therapy. J Vase Intervent Radiol 1997; 8:781-
788.

47. Kee ST, Kinoshita L, Razavi MK, Nyman UR, Semba CP, Dake MD. Superior vena cava
syndrome: Treatment with catheter-directed thrombolysis and endovascular stent placement.
Radiology 1998; 187-193.

48. Hammer F, Becker D, Gofette, Mathurin P. Crushed stents in benign left brachiocephalic vein
stenosis. J Vase Surg 2000; 32:392-396.

49. Brant J, Peebles C, Kalra P, Odurny A. Hemopericardium after superior vena cava stenting for
malignant SVC obstruction: The importance of contrast-enhanced CT in the assessment of
postprocedural collapse. Cardiovasc Intervent Radiol 2001; 24:353-355.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

34

Complications of Endovascular Intervention for AV

Access Grafts

Abigail Falk

Mount Sinai Medical Center, New York, New York, U.S.A.

I. INTRODUCTION

The incidence of treated end-stage renal disease (ESRD) in the United States is 180 per
million and it continues to rise at a rate of 7.8% per year (1). A majority of patents with
ESRD require hemodialysis for survival. In such patients, the creation and maintenance of
an arteriovenous (AV) access are two of the most difficult issues associated with their
hemodialysis treatment. Hemodialysis vascular access dysfunction is a major clinical prob-
lem. Although current options for vascular access include an arteriovenous fistula (AVF),
an arteriovenous graft (AVG), a venous catheter, or a totally subcutaneous catheter-based
access system, approximately 70% of permanent dialysis vascular access in the United
States is a polytetrafluoroethylene (PTFE) AVG. In general, such AVGs have dismal
unassisted primary patency rates — 50% at 1 year and 25% at 2 years — and require
interventions for maintenance of adequate flow (2). Endovascular interventions, including
balloon or percutaneous angioplasty (PTA) (intraluminal balloon dilatation), stent deploy-
ment (placement of a self-expanding or balloon expandable stent), pharmacological
thrombolysis (catheter-directed infusion of thrombolytic agents), mechanical thrombolysis
and thrombectomy (fragmentation, maceration, or mobilization of thrombus by mechan-
ical means or devices), and pharmacomechanical thrombolysis (mechanical disruption of ■§
thrombus by fragmentation, maceration, or mobilization accompanied by catheter-direc- g
ted administration of an agent that results in pharmacological thrombolysis) are the current as
standards of practice for AV access maintenance. c
The natural history of all AVG is the gradual development of intimal hyperplastic <
stenoses, which will ultimately lead to AVG failure. Some 80% of all vascular access >9
dysfunction in PTFE grafts is due to venous stenosis and thrombosis formation at the graft- J
vein anastomosis. The stenosis occurs as a result of venous neointimal hyperplasia, which is °
composed of smooth muscle cells and characterized by significant angiogenesis. While |
many local therapeutic approaches — including radiation therapy, gene therapy, and use of @

557

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

558 FALK

coated stents and polymeric drug delivery systems — are being investigated to combat this
problem, balloon angioplasty (with or without lysis/thrombectomy when needed) and stent
placement remain the current standard of care. The National Kidney Foundation's
Diseases Outcomes Quality Initiative (K-DOQI) recommends surgery for correction of
graft degeneration, infection, and pseudoaneurysms and endovascular intervention for
correction of stenoses (3). Each of these interventions is itself associated with varying
complications that require specific active management. This chapter focuses on endovas-
cular interventions and their associated complications.

II. ENDOVASCULAR INTERVENTIONS

Intervention is mandated whenever stenosis becomes hemodynamically significant — i.e.,
when there is a >50% reduction of vessel or graft diameter along with abnormal changes in
hemodynamic or clinical parameters (3). The K-DOQI guidelines recommend that a
patient's AVG be routinely monitored for stenosis. Increased static or dynamic blood
pressure, decreased blood flow, increased access recirculation, or swollen extremity are all
abnormal changes indicative of possible underlying stenosis. If left untreated, the decreas-
ing pressures within stenotic vessels become conducive to thrombus formation, progressive
graft occlusion, and compromised dialysis. Balloon angioplasty is a well-recognized
intervention for the maintenance of graft patency (4). When angioplasty remains unsuc-
cessful or is required more than two times within a 3-month period, surgical revision should
be attempted provided that the patient is a candidate for surgery (3). In cases where surgery
is contraindicated or not possible, intragraft stent placement may be an option for
improving patency. In patients with underlying thrombus, additional treatment directed
toward the mechanical, pharmacological, or pharmacomechanical lysis of the thrombus is
necessary to increase the duration of graft patency. Guideline 10 of the K-DOQI states that
prospective surveillance of AVG for hemodynamically significant stenosis, when combined
with correction, improves patency and decreases the incidence of thrombosis (3). Thus,
maintenance of hemodialysis access requires both prospective surveillance and need-based
intervention.

III. COMPLICATIONS OF ENDOVASCULAR INTERVENTIONS

A team approach is best for the successful management of complications associated with
endovascular interventions. The concerted input of nephrologists, interventional radiolo-
gists, vascular and transplant surgeons, fellows, nurses, and physicians' assistants is im-
portant in achieving the overall goal of vascular access preservation.

A. General Complications |

General allergic reactions to radiographic contrast media and fluid overload, with possible »

progression to congestive heart failure (CHF), are two overall complications to be cognizant c

of in this patient population regardless of the type of endovascular intervention. A history of <

moderate or severe reactions to contrast media or of asthma is considered an important risk >9

factor for generalized contrast media reactions (5). Patients with such risk factors may be J

prophylactically treated with corticosteroids at least 1 1 h prior to the procedure and should «

preferably receive nonionic contrast media. Some physicians additionally use oral antihista- |

mine treatment. In patients at high risk for reaction, a resuscitation team should be available @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOVASCULAR INTERVENTION FOR ACCESS GRAFTS 559

at the time of the procedure. In those cases where an immediate need for intervention
precludes steroid premedication, gadolinium may be used as a contrast agent, albeit with
some compromise in resolution. Because of the risk of carbon dioxide reflux and arterial
CO2 embolism in the cerebral circulation, CO2 is contraindicated for arterial use above the
diaphragm and should not be used in upper extremity grafts. There is a potential for serious
consequences with the use of CO2. In one case, the use of CO2 contrast during the declotting
procedure of an upper arm AVG led to a CO2 embolism of the vertebral artery and
subsequent major stroke (A. Novick, personal communication, 1999).

Patients with ESRD requiring hemodialysis typically have poor cardiopulmonary
function, and fluid overload during the interventional procedure may lead to increased
cardiac burden and CHF. Monitoring of fluid delivery during endovascular interventions is
thus critical in reducing cardiac complications.

B. The Nature of the Stenotic Lesion and Its Influence
on Interventional Complications

Stenotic lesions in AVGs are typically concentric, intimal thickenings of the vein wall
consisting of smooth muscle and extracellular matrix components. Although they can
occur anywhere along the outflow circuit, stenotic lesions are more common at venous
anastomoses (6). They are generally more difficult to treat than atherosclerotic lesions
because of their aggressive development and greater resistance to balloon dilatation.

For treatment to be successful, it should be instituted early — i.e., before progression to
thrombosis (7). When they are instituted later, interventions should restore patency as well
address any underlying problem of thrombosis, since the risk for rethrombosis is generally
90% or greater. Anatomical success is defined as a less than 30% residual diameter stenosis
(8). For stenosis with AVG thrombosis, restoration of flow combined with a less than 30%
maximal residual stenosis should be considered successful. In all cases, restoration of
hemodynamic and flow parameters is a measure of treatment or clinical success and helps
prolong the duration of patency of AVGs. Prospective monitoring of graft function is
recommended to facilitate early detection and intervention of stenosis in AVGs, since grafts
treated with percutaneous transluminal angioplasty (PTA) alone have been shown to have
a longer duration of patency than those treated with a combination of thrombolysis plus
PTA.

C. Percutaneous Transluminal Angioplasty

The PTA of venous stenoses related to AVGs is well accepted, with technical success rates of

85-97%. While recurrent stenosis requiring repeat PTAs may be an issue, the rates of

restenosis do not appear to be greater than with surgical intervention. The 6-month

unassisted patency rate following PTA is between 40 and 50% . The intervention is relatively §

safe, with complications occurring in 2-4% of cases (9). |

I
1. Complications .g>

In general, percutaneous procedures are outpatient procedures, with reduced morbidity, <

reduced postprocedural pain, and little or no wound edema. However, since graft stenoses >9

are intimal hyperplastic lesions of muscle and matrix material that form a concentric, ^

focal thickening of the vein wall, high pressures are often needed to inflate the angioplasty «

balloon completely. These high inflation pressures are causative of a majority of com- |

plications associated with PTA intervention. First, the tremendous forces required to @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

560 FALK

dilate stenotic lesions may directly cause a tear or perforation in the vascular wall. Second,
vascular ruptures may also result from bursting of the angioplasty balloon during the
dilatation procedure. As the balloon bursts, a fluid jet is released almost instantaneously
with the burst. Since the balloon is tightly apposed to the endovascular wall at the time of
balloon rupture, the pressurized fluid is directed at the adjacent vascular wall, causing a
rupture of that wall. Fortunately, such complications are uncommon. Reported rates of
incidence range from 0.7 to 4.5% (9).

In some cases, PTA-associated complications may become serious and lead to emergent
crises. The Standards of Practice Committee of the Society of Cardiovascular and Interven-
tional Radiology (SCVIR), in their Quality Improvement Guidelines for Percutaneous
Management of the Thrombosed or Dysfunctional Dialysis Access, recommends a thresh-
old rate of 0.5% for major complications associated with vascular rupture or perforation
(10). A major complication was defined in these guidelines as a perforation or rupture
requiring blood transfusion or emergent surgery or one leading to limb-threatening
ischemia.

When venous rupture occurs during PTA intervention, it does not necessarily result in
a poor outcome. Treatment can result in a resolution of the adverse event and restoration
of graft function. The classic response in the acute management of angioplasty-induced
venous ruptures is an immediate manual compression of the injured site. Alternatively,
repositioning of the angioplasty balloon across the injured segment and reinflation of
balloon to tamponade the bleeding may be effective in "tacking up" the vascular wall. In
the past, an unsuccessful outcome with these two maneuvers led to an abandonment of the
procedure or to an immediate occlusion of the rupture to stop uncontrollable bleeding. In
recent years, however, reports indicate that deployment of metallic endovascular stents
across the area of extravasation may be useful in such cases to treat the vascular rupture
and provide patency for the access graft (11,12). Indeed, the rupture of outflow vein or
venous anastomotic stricture from balloon angioplasty is typically treated by tacking up
the vein with repeat angioplasty or stent placement. When all such management options
fail, an alternative is to occlude the affected graft, reconstruct a new AVG, and insert a
new hemodialysis catheter for interim hemodialysis.

The treatment of arterial lesions is no different from standard arterial balloon
angioplasty. Rupture of artery from a balloon angioplasty is likewise treated by tacking
up the artery with repeat balloon angioplasty or surgery. A patient with an arterial rupture
that remains uncontrolled following manual maneuvers, for example, can be controlled
with surgical ligation of graft and placement of an arterial bypass graft (13).

D. Endovascular Stents

The role of endovascular stent placement in the management of AVG stenosis is not ■§

completely clear. The unassisted patency of stents in hemodialysis access is not different g

from that following PTA except in elastic stenoses (3,14). Thus, stents may be used to a

salvage additional AVGs but may not provide any additional advantage in terms of c

prolonging patency duration. They should be reserved for surgically inaccessible stenoses <

that fail PTA or used in patients who have limited remaining access sites. When deployed, >9

they must never overlap major side veins and obviate the potential for the creation of J

future access sites (Fig. 1A and B). Q

Stent deployment appears to be safe, with procedural success rates of 96-100% |

reported with both Brescia-Cimino and polytetrafluoroefhylene (PTFE) grafts (15,16). @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOVASCULAR INTERVENTION FOR ACCESS GRAFTS

561

B

Figure 1 Case study of an immediate venous anastomotic recoil refractory to PTA and with an
expanding hematoma at the sheath entry site due to outflow obstruction; it was treated with stent
placement over the vein/graft anastomosis. The patient was a 40-year-old, HIV-positive male who
had undergone nephrectomy for renal cell carcinoma. He had a history of a failed left radiocephalic
AVF and a left forearm loop AVG that had been declotted several times and surgically revised. The
patient presented with a left-upper-arm AVG that was placed 3 months earlier and had thrombosed.
The AVG was declotted using 1.2 U of reteplase and 3000 U of heparin (22). A 7-mm high-pressure
angioplasty balloon was used to treat the venous anastomotic stricture. However, immediate recoil
was noted. An 8-mm high-pressure angioplasty balloon was then used to treat the venous anastomotic
stricture. Immediate recoil was also noted with the larger balloon. This case was complicated by an
expanding hematoma at sheath entry site (Fig. 1A). Finally, immediate patency of the graft/venous
junction was established by placement of an 8 mm Wallstent. No further extravasation was seen after
antegrade flow was reestablished (Fig. IB). Return of 'thrill' to the AVG was noted and the patient
underwent successful hemodialysis following the procedure.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

562 FALK

Both Wallstent and intragraft metallic stents have been used safely in hemodialysis grafts.
No consistent complications have been associated with their use although there is a
potential for stent migration. Stent-related pseudoaneurysm has also been described in the
literature (15). Stent migration can be treated with either retrieval (if possible) and
repositioning or placement of an additional stent to treat the underlying lesion. As with
PTA, restenosis may occur and necessitate repeat procedures. Causes of recurrence include
intimal hyperplasia in or near the stent, stent slippage, and remote stenoses (15).

Several authors have documented that restenosis remains a significant problem when
initial balloon angioplasty is employed to treat venous stenotic disease caused by neo-
intimal hyperplasia (6,17,18). In such patients, insertion of a stent may be indicated when
stenoses of graft venous anastomoses have recurred twice in less than 6 months and has
been shown to increase mean restenosis interval (17). A variety of stents have been used for
this purpose, including Wallstent, Craggstent, and Passager stent. More recently, flexible
self-expanding stents have been used in clinical trials. However, neointimal hyperplasia can
still recur through the stent interstices and at the ends of such stents (19). Preliminary
results from a multicenter phase 1 trial of a PTFE stent for hemodialysis graft venous
anastomotic stenoses show that the PTFE encapsulated stent graft is safe and provides
promising patency results for treatment of venous anastomotic stenoses (20). Indicated for
the treatment of tracheobronchial strictures, the Viabahn endoprosthesis is constructed
from an expandable PTFE tubular lining with a helical nitinol (a flexible alloy) reinforce-
ment exoskeleton. Although not currently indicated for vascular applications, the flexibility
and radial strength of this device are being exploited for salvaging stenotic AVGs (21) (Fig.
2A to D). Further studies are needed to determine whether their use can replace traditional
surgical patch angioplasty and help combat the long-term complication of balloon
angioplasty-namely, restenosis.

E. Thrombus Removal

Removal of any occlusive thrombus is an integral part of the management of AVG-related
complications. This can be achieved by thrombolysis (pharmacological or pharmaco-
mechanical) or thrombectomy. There are a variety of percutaneous thrombectomy tech-
niques, including suction thrombectomy, balloon thrombectomy, clot maceration, and
mechanical thrombectomy. Thrombolysis, likewise, can be achieved via a percutaneous
thrombolytic device, pulse-spray pharmacomechanical thrombolysis, or lyse-and-wait tech-
niques.

1 . Complications

Complications reported in various series of patients undergoing hemodialysis graft
thrombectomy include major and minor bleeding, venous rupture, arterial embolism,

acute respiratory arrest, contrast reaction, sepsis or infection, and death (Table 1). In a ■a

recent retrospective analysis of 935 thrombectomy procedures (74% mechanical device- |

mediated) that compared different thrombectomy techniques and devices, an overall com- a

plication rate of 3.3% was found, with the most commonly used interventions of throm- c

bectomy — percutaneous thrombolytic device, Amplatz thrombectomy device, AngioJet, <

Oasis — having comparable rates of complications (9). The complications were similar to >9

those reported in earlier series. The SCVIR Quality Improvement guidelines suggest J

threshold rates of 0.5% for major bleeding, venous rupture, acute respiratory arrest, «

and death and a threshold of 2% for arterial emboli (Fig. 3 A to D) (10). Note that the most |

common complication in these studies was venous rupture, induced by PTA-related @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOVASCULAR INTERVENTION FOR ACCESS GRAFTS

563

B

Figure 2 Case study of restenoses at the venous anastomosis and within several sites in the outflow
vein (to the level of the axillary vein) in an AVG that was treated three times previously with balloon
angioplasty; the poor angioplasty result was subsequently treated with a PTFE stent-graft using
endovascular techniques. The patient was a 52-year-old HIV-positive male with a 6-month-old left
upper arm AVG. At 3.5 months, the patient presented with high venous pressures and was referred
for intervention. At that time, three sites of focal stenoses — one at the venous anastomosis, one in
the high brachial vein, and one in the axillary vein — were identified and treated with a 9-mm
angioplasty balloon. The patient presented again at 6 months with a thrombosed AVG. Following
declotting, balloon angioplasty was performed once again using a 9-mm angioplasty balloon. The
same three sites of stenosis were identified and the AVG rethrombosed in 2 days. After discussion
with the vascular surgeon and nephrologist, it was decided to place a PTFE stent-graft (V1ABAHN
Endoprosthesis, W.L. Gore and Associates, Flagstaff, AZ). The AVG was declotted again and the
outflow vein was found to be irregular and sclerotic from the level of the graft/vein anastomosis
through the axillary vein following balloon angioplasty with a 9-mm balloon (Figs. 2A and B). A
7-mm by 15-cm Viabahn stent-graft extension was placed and dilated to 7 mm, reestablishing
excellent flow across the diseased venous segment (Fig. 2C). A magnified view of the stent-graft is
shown in Figure 2D.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

564

FALK

"""«i»»i?

Figure 2 Continued.

procedures that were performed in conjunction with thrombectomy or thrombolysis
and not directly related to the latter. The cardiopulmonary complications is likely related
to dislodgement of arterial plugs and subsequent embolization of thrombotic material to
pulmonary arteries, a risk that would be greater if the volume of the thrombosed plug
exceeded the diameter of the pulmonary vasculature. Although, in most cases, the size
of the thrombosed plugs are reported to be small and safe during pulmonary circulation,

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOVASCULAR INTERVENTION FOR ACCESS GRAFTS

565

Table 1 Comparison of the Complications Rates Reported in Published Series
Versus Threshold Levels Recommended by the SCVIR Quality Improvement
Guidelines (10)

Complication

Published series

SCVIR guidelines

Major bleeding

Minor bleeding

Venous rupture/dissection

Arterial emboli

Acute respiratory arrest

Contrast reaction

Sepsis, infection

Death

0.7-1.7 %
2.8-11.1 %
1.2-7.0 %
1.2-9.3 %
1.6-2.3 %
0.7-1.4 %
1.2-2.0 %
0-1.5%

0.5%

0.5%

2.%

0.5%

0.5%

Source: Ref. 9.

the long-term consequences of such silent emboli remain uncertain and care should be
taken to minimize embolization. Patients with severe cardiopulmonary disease may be at
high risk for complications during thrombectomy, and this should be a contraindication
for that procedure. Finally, residual stenosis may be expected in >85% of cases of
thrombosis. In all cases, the access should be evaluated with a fistulogram at the end of the
procedure and any residual stenosis corrected by angioplasty or surgery.

Thrombolysis can be complicated by pulmonary embolism, cerebral embolism (in the
presence of a patent foramen ovale), septic emboli, or local bleeding. The last commonly

Figure 3 Embolization of thrombus into distal brachial artery, treated with Fogarty balloon
catheter thrombectomy. The patient was an 87-year-old hypertensive male with a 7-month-old
left forearm loop AVG. This was the patient's first access site. The AVG was previously declotted
at 4 months after placement, with no complications. At 7 months, the patient presented with a
rethrombosed AVG. A standard declotting procedure was performed using 1 .2 U reteplase and 3000
U heparin. Initially, the distal brachial, ulnar, and radial arteries were noted to be patent (Fig. 3A).
The procedure was complicated by embolization of thrombus into the distal brachial artery, ob-
structing flow to the radial and ulnar arteries (Fig. 3B). The thrombus was retrieved with a 3F over-
the-wire Fogarty balloon catheter (Baxter Healthcare Corporation, Deerfield, IL) (Fig. 3C). As seen
here, patency of distal brachial, radial, and ulnar arteries was reestablished (Fig. 3D).

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

566

FALK

B

D

I

•5

Figure 3 Continued.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOVASCULAR INTERVENTION FOR ACCESS GRAFTS 567

occurs from old puncture sites and can be conservatively treated with manual compres-
sion. Because the lytic agent is administered into a closed space and not into the patient's
circulation, there is minimal risk of systemic bleeding complications. Embolization to
radial and ulnar arteries may be treated with a Fogarty balloon and heparinization (21).
Anticoagulation may be discontinued following confirmation of normal pulses.

IV. SUMMARY

Complications associated with interventions that salvage dysfunctioning or stenotic
hemodialysis grafts are typically related to the type of intervention used and the presence
or absence of thrombus within the affected region of the graft. Overall, nonthrombosed
grafts have better outcomes than thrombosed grafts. Balloon angioplasty or PTA is a well-
accepted intervention for the correction of stenosis. The most common complication
associated with PTA intervention of stenotic AVGs is venous rupture, caused largely by
the high pressures needed to dilate vessels against the strong intimal hyperplastic walls.
Fortunately, the incidence of such complications is low and most complications can be
managed with manual compression, repositioning of the balloon and prolonged inflation,
stent deployment, or possibly surgical revision. In patients in whom lesions have progressed
to thrombosis, additional intervention must resolve the clot. In general, complication rates
associated with pharmacomechanical thrombolysis are slightly higher than those with
mechanical thrombectomy. Complications associated with various thrombectomy proce-
dures and devices include major and minor bleeding, venous rupture, arterial embolism,
acute respiratory arrest, contrast reaction, sepsis or infection, and death. Caution should
be exercised in using thrombolytic agents, and thrombectomies should not be performed
in patients with severe cardiopulmonary disease. In patients who receive stents as
endovascular interventions, complication rates are low. Restenosis appears to be the major
problem, necessitating repeat procedures. In summary, endovascular interventions can be
safe provided that there is appropriate patient selection and proper technique. Such
interventions contribute to prolongation of the long-term patency of vascular access
grafts.

REFERENCES

1. Sidawy AN, Gray R, Besarab A, Henry M, Ascher E, Silva M Jr, Miller A, Scher L, Trerotola
S, Gregory RT, Rutherford RB, Kent KC. Recommended standards for reports dealing with
arteriovenous hemodialysis accesses. J Vase Surg 2002; 35(3):603— 610.

2. Roy-Chaudhry P. Hemodialysis access dysfunction from bedside to bench to bedside. Vascular
Access for Hemodialysis VIII. The Eighth Biannual Symposium, Rancho Mirage, CA, May 9-

10, 2002. «

3. NKF-K/DOQI Clinical Practice Guidelines for Vascular Access: update 2000. Am I Kidney |
Dis 2001; 37(1 suppl 1):S137-S181. |

4. Beathard GA. Percutaneous angioplasty for the treatment of venous stenosis: A nephrologists ■a

view. Semin Dial 1995; 8:166-170.

Morcos SK, Thomsen HS, Webb JA. Contrast Media Safety Committee of the European

Society of Urogenital Radiology. Prevention of generalized reactions to contrast media: A

consensus report and guidelines. Eur Radiol 2001; 1 1(9): 1720— 1728.

Kanterman RY, Vesely TM, Pilgram TK, Guy BW, Wndus DW, Picus D. Dialysis access

grafts: anatomic location of venous stenosis and results of angioplasty. Radiology 1995;

195(1):135-139.

c

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

568 FALK

7. Katz SG, Kohl RD. The percutaneous treatment of angioaccess graft complications. Am J
Surg 1995; 170(3):238-242.

8. Gray RJ, Sacks D, Martin LG. Trerotola SO and the members of the Technology Assessment
Committee. Reporting standards for percutaneous interventions in dialysis access. J Vase
Intervent Radiol 1999; 10:1405-1415.

9. Vesely TM. Complications related to percutaneous thrombectomy of hemodialysis grafts.
Manuscript submitted to AJKD.

10. Aruny JE, Lewis CA, Cardella JF, Cole PE, Davis A, Drooz AT, Grassi CJ, Gray RJ, Husted
JW, Jones MT, McCowan TC, Meranze SG, Van Moore A, Neithamer CD, Oglevie SB,
Omary RA, Patel NH, Rholl KS, Roberts AC, Sacks D, Sanchez O, Silverstein MI, Singh H,
Swan TW, Towbin RB, Trerotola SO, Bakal CW, for the Standards of Practice Committee of
the Society of Cardiovascular & Interventional Radiology. Quality improvement guidelines for
percutaneous management of the thrombosed or dysfunctional dialysis access. J Vase Intervent
Radiol 1999; 10:491-498.

11. Rundback JH, Leonardo RF, Poplausky MR, Rozenblit G. Venous rupture complicating
hemodialysis access angioplasty: Percutaneous treatment and outcomes in seven patients. Am J
Roentgenol 1998; 171(4): 108 1-1084.

12. Welber A, Schur I, Sofocleous CT, Cooper SG, Patel RI, Peck SH. Endovascular stent
placement for angioplasty-induced venous rupture related to the treatment of hemodialysis
grafts. J Vase Intervent Radiol 1999; 10(5):547-551 (Comment in: J Vase Intervent Radiol 1999;
10(8):1 135-1 136.

13. Falk A, Mitty H, Guller J, Teoderescu V, Uribarri J, Vassalotti J. Thrombolysis of clotted
hemodialysis grafts with tissue-type plasminogen activator. J Vase Intervent Radiol 2001;
12:305-311.

14. Patel RI, Peck SH, Cooper SG, Epstein DM, Sofocleous CT, Schur I, Falk A. Patency of
Wallstents placed across the venous anastomosis of hemodialysis grafts after percutaneous
recanalization. Radiology 1998; 209:365-370.

1 5. Gray RJ, Horton KM, Dolmatch BL, Rundback JH, Anaise D, Aquino AO, Currier CB, Light
JA, Sasaki TM. Use of Wallstents for hemodialysis access-related venous stenoses and occlu-
sions untreatable with balloon angioplasty. Radiology 1995; 195(2):479-484.

16. Zaleski GX, Funaki B, Rosenblum J, Theoharis J, Leef J. Metallic stents deployed in synthetic
arteriovenous hemodialysis grafts. Am J Roentgenol 2001; 176(6):1515-1519.

17. Turmel-Rodrigues L, Pengloan J, Blanchier D, Abaza M, Birmele B, Haillot O, Blanchard D.
Insufficient dialysis shunt: Improved long-term patency rates with close hemodynamic moni-
toring, repeated percutaneous balloon angioplasty, and stent placement. Radiology 1993;
187:273-278.

18. Beathard GA. Percutaneous transvenous angioplasty in the treatment of vascular access ste-
nosis. Kidney Int 1992; 42:1390-1397.

19. Schurmann K, Vorwerk D, Kulisch A, Rosenbaum C, Biesterfeld S, Gunther RW. Puncture of
stents implanted into veins and arateriovenous fistulas: an experimental study. Cardiovasc
Intervent Radiol 1995; 18(6):383-390.

20. Vesely T, Dammers R, Planken R, Pouls K, van Det R, Burger H, van de Sande F, Tordoir J.
Multicenter phase I results of a PTFE stent graft for hemodialysis graft venous anastomotic j>
stenoses. Vascular Access for Hemodialysis VIII. The Eighth Biannual Symposium, Rancho <S
Mirage, CA, May 9-10 2002. g

21. Belville J, Borzatta M. Covered stents for treatment of AV graft stenosis. Vascular Access for °|
Hemodialysis VIII. The Eighth Biannual Symposium, Rancho Mirage, CA, May 9-10, 2002. ^

22. Falk A, Guller J, Nowakowski FS, Mitty H, Teoderscu V, Uribarri J, Vassalotti J. Reteplase in >9
the treatment of thrombosed hemodialysis grafts. J Vase Interv Radiol 2001; 12:1257-1262. S

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

35

Complications of Vena Cava Filters

Enrico Ascher, Anil Hingorani, and William R. Yorkovich

Maimonides Medical Center, Brooklyn, New York, U.S.A.

I. INTRODUCTION

The impact of pulmonary embolism (PE) on patient morbidity is significant. A study of
5000 autopsies reported that 55% had evidence of gross PE and that 18% of all deaths
were directly attributable to PE (1). It has been estimated that PE occurs in as many as
630,000 patients annually in the United States, and the diagnosis of pulmonary embolism
will have been missed in 71% (450,000) of these patients (2,3). Those fortunate enough to
survive an 11% (69,300 deaths) first-hour mortality rate are then subjected to a 23.8%
first-year mortality (150,000 deaths) (4). With an increasing population (general and aged),
these numbers can only be expected to increase.

Lower extremity deep venous thromboses (LEDVT) accounts for 88-92% of pulmo-
nary emboli, and upper extremity deep venous thromboses (UEDVT) account for the
remaining 8-12% (5,6). It has also been shown that PE occurs in up to 30-50% of patients
with LEDVT of iliac origin who are not anticoagulated. Additionally, when compared to
LEDVT, UEDVT is associated with higher morbidity and mortality rates (7), and there is
evidence that the incidence of PE of UEDVT origin is rising due to the increased use of
central venous catheters (8).

II. INDICATIONS |

Anticoagulation has been long proven and continues to be the most effective way of a

preventing of pulmonary embolism (9,10). However, many occasions arise where this c

protocol is not appropriate; an alternative treatment must therefore be made available to <

those patients at risk for developing PE. In the presence of deep venous thrombosis (DVT) >9

in patients for whom anticoagulation has either failed or is contraindicated or who have J

sustained or are at risk for a complication of anticoagulation (i.e., hemorrhage, thrombo- «

cytopenia), the implantation of one or more filtration devices in either or both the inferior |

and/or superior vena cava to prevent pulmonary embolism is indicated. Filtration is also @

569

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

570

ASCHER et al.

indicated in the presence of large free-floating vein thrombi in the inferior vena cava or
iliac arteries. In those patients in whom there exists either a history or a high risk for the
development of DVT, prophylactic implantation of one or more filtration devices may be
indicated (11,12). This includes high-risk patients who will be undergoing hip or knee
replacement surgery, or gastric bypass for morbid obesity as well as those who are
pregnant, in whom anticoagulation is contraindicated.

Prophylactic placement of one or more vena cava filters is also indicated in situations
of high-risk trauma, such as a spinal cord injury with an associated severe head injury and
complex pelvic fractures or multiple long bone fractures. In patients with DVT, expanded
indications for the placement of caval filters may include the presence of cor pulmonale, or
metastatic disease, syncope in elderly persons, and or the risk of sustaining falls.

As the use and indications of vena cava filters increases, so does the frequency of
associated complications. It is imperative that physicians performing filter placement be
familiar with the myriad of potential complications related to the use of these filters. This
chapter provides an overview of those complications.

III. DEVICES IN USE TODAY

The most widely placed vena cava filters are the Greenfield (titanium and "over-the-wire"
stainless steel), the Bird's Nest, the Vena Tech-LGM, and the Simon-Nitinol (Table 1).
Significant data are available in the literature for indications and techniques of placing
these devices. Three more recent devices are the Gunther-Tulip, Cordis TrapEase, and
Vena Tech LP; for all of which limited experience has been published.

While all of these transvenous filters are effective in preventing pulmonary embolism
in the majority of patients, each device has its own characteristics, advantages, disadvan-
tages, and complication rate. Considerations for selection of the appropriate filtration
device will include patency, luminal diameter of the vena cava, introducer sheath size,
selection of the venous access site, anatomical variations affecting ease of placement, filter
stability and migration tendencies, efficacy in clot trapping, rate of vena cava and access
vein thromboses, PE recurrence rate, and artifact producing ferromagnetic properties that
may affect future magnetic resonance imaging (MRI) examinations. The techniques
necessary to deploy these devices into optimal position successfully has been previously
described in detail elsewhere (13-15). It can also be noted that there is increasing
experience with the successful placement of superior vena cava filters (16-18).

Table 1 Comparison of Key Vena Cava Filter Attributes

Filler

Titanium
Greenfield

OTW SS

Greenfield

Bird's Nest

Simon
Nilinol

Vena
Tech LGM

Vena
Tech LP

Gunther
Tulip

Trap
Ease

Introducer size

15

15

14

9

13.6

9F 8.5

8

s

OD diameter (F)

•a

Insertion site

13.1-28

9.6-23

7.4-23

11.5-31

16.7-36

-

6

thrombosis (%)

Filter migration (%)

-

11.0

12.0

1.7

18.4

-

£

DVT (%)

22.7

5.9

36

8 9

-

3

-

I
Q

Maximum caval

28

28

40

28

28

30 30

30

diameter (mm)

IVC thrombosis (%)

1-9

4

3-8

4-25

4.5-30

-

4.5

1

PE recurrence (%)

3-5

3-5

2.7-4.1

<l^t.8

2-3.8

-

s

MRI

Safe

Sale

Not recommended

Safe

Safe

Safe Safe

Sale

©

I

s

Marcel Dekker, Inc.

Ifl

270 Madison Avenue. New York, New York 1 00 1 6

■au,

COMPLICATIONS OF VENA CAVA FILTERS 571

IV. COMPLICATIONS

A. Filter Misplacements and Migration

Correct placement of the filter requires careful preoperative planning and meticulous
technique. The use of contrast venography to identify the optimal discharge site is essential
and will reduce the possibility of misplacing a filter. Intraoperative venography may reveal
errant catheter placement, as into the mesenteric vein through the portal system (Fig. 1A).
Identification of the malposition and redirection of the catheter for ideal filter deployment
from its delivery system may then result in proper placement of the filter into the vena cava
(Fig. IB). Failure to correctly identify and verify catheter positioning prior to filter
discharge has resulted in discharge of the filter into the aorta and necessitated use of a
second, correctly placed filter (Fig. 2).

Rapid yet precise extrusion of the filter reduces the incidence of leg asymmetry,
incomplete opening, and filter tilting and lessens the chance of "caudal drop." Even so,
misplacement of the filter at operation may occur, either through inadequate identification
of the discharge site, suboptimal technique, or inappropriate device selection. A filter that
is tilted or incompletely expanded is at increased risk for migration (Fig. 3). Accidental
guideline manipulations or central line placement have also contributed to the dislodge -
ment of previously successfully placed filters (19,20). Migrations are most frequently
observed in the immediate postoperative period but may also occur many years later.
Either fractured components or the entire device may migrate (21).

Filters may be inadvertently ejected into the right atrium at the time of the procedure,
and subsequent proximal migrations have been reported to the right ventricle and the
pulmonary artery in addition to the right atrium. Intracardiac migration of a Greenfield
filter, with stenting open of the tricuspid valve and wide-open regurgitation has also been
reported (22).

B. Perforation of the Vena Cava Wall

The incidence of erosion or perforation of the inferior vena cava (IVC) wall by vena cava
filters, injuring adjacent retroperitoneal and abdominal structures, is rare, and symptoms
are reported infrequently by patients with these complications (23). Perforations with or
without filter migration are often a result of filter angulation or tilting, and filter struts
may put adjacent structures in danger of penetration. Perforations of the vena cava by a
filter may extend into the abdominal aorta (24-27). Temporary vena cava filters may also
be culpable, and incorporation into the caval wall has been reported (28).

C. Gastrointestinal Complications

Besides bleeding, other serious complications of vena cava filter migration and perforation

are possible, some extremely urgent. Gastrointestinal complications include small bowel |

volvulus (29) and duodenal perforation. Duodenal-caval fistulas may be caused by a

migrating filters and early abdominal surgery after filter placement and may present as H,

upper gastrointestinal bleeding when bleeding ulcers are induced by local strut trauma =j

(30,31). i

I

D. Guidewire Mishaps <3

Guidewire-related mishaps are potential complications of inferior vena cava filters, and it §

has been postulated that they may be underreported (32-34). An entrapped guidewire may ©

be impossible to remove following placement of the filter, and endoscopic or operative li

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

572

ASCHER et al.

B

Figure 1 A. Intraoperative venography reveals catheter placement into the mesenteric vein through
the portal system. Misplacement was identified and the catheter was repositioned. B. Correctly placed
TrapEase filter in the inferior vena cava.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VENA CAVA FILTERS

573

Figure 2 Filter misplaced into the aorta. Also shown is the second filter, placed correctly into the
vena cava.

I
•5

Figure 3 Incomplete expansion of a Greenfield filter.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

574

ASCHER et al.

intervention may be warranted. The guidewire may become entangled and fluoroscopic
examination will reveal entanglement of the J-tip guidewire in the IVC filter (Fig. 4).

E. Recurrent Embolism

Pulmonary embolism, despite the presence of a filter and adequate anticoagulation, may
recur (35,36). In patients with existing thromboembolic disease, recurrent DVT is not an
unexpected event. When possible, anticoagulation is used in conjunction with the filter to
treat existing DVT, reduce the progression of thrombus, and potentially reduce subse-
quent complications. However, anticoagulation does not seem to reduce the rate of
recurrent DVT (37). Paradoxical cerebral embolism must be considered in patients with
DVT who have new-onset neurological deficits even in the presence of a caval filter (38—
40).

F. Renal Complications

Due to the risk of renal vein thrombosis, IVC filter placement above the renal veins may
not be appropriate in advanced-stage cancer patients who have a single functional kidney,
renal insufficiency, or prior renal vein thrombosis. Suprarenal filter placement should be
performed only after analysis of predicted survival, after detailed discussions with the
patient, and — most importantly — after renal function evaluation (41). Obstructive urop-
athy due to a pelviureteric obstruction resulting from vena cava wall perforation may
occur (42), and penetration by an IVC filter may also cause ureteral injury (43) or
symptomatic hydronephrosis (44).

G. Other Miscellaneous Complications

Less frequently reported complications of vena cava filter mishaps have included
penetration of a vertebral body by a filter strut (45) and phlegmasia cerulea dolens (46).

I
•5

Figure 4 Guidewire entrapped in filter.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VENA CAVA FILTERS

575

V. PREVENTING AND CORRECTING VENA CAVA FILTER
COMPLICATIONS

Imaging before, during, and after a procedure can help to reduce or prevent some
potential problems. A preprocedural venacavogram is required for measuring the vein's
diameter, which is necessary for proper device selection. Pertinent anatomical variations,
such as a megacava or dual inferior vena cavas, can be also be identified, as can the
presence of IVC thrombosis. In the case of a superior vena cava (SVC) procedure, a
venogram is performed to assess the innominate vein, in which placement of the filter is
critical because of the vein's short length. Postprocedural venacavograms will assist in the
assessment of filter placement and degree of tilting or asymmetry. Continuous fluoroscopy
whenever a filter is being placed from the right internal jugular to the IVC to prevent
puncture of the right atrium should be used.

Techniques for minimizing the effects of misplaced or migrated vena cava filters are
varied and unique for each situation. Successful operative extractions utilizing cardiopul-
monary bypass and incision for intracardiac misplaced filters or for those that have
migrated to the heart have been reported (47). A vena cava wrapping technique for
bleeding from a vena cava perforation, rather than surgical removal of the filter, has been
performed because of the still present risk for recurrent PE (48).

I
•5

Figure 5 Migrated Greenfield filter into the right inferior pulmonary artery.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

576 ASCHER et al.

Emergent laparotomy may be required to retrieve a broken filter wire projecting into
the duodenum. The offending wire/strut/hook may be transected and the vessel wall
closed, leaving remaining filter in place if not otherwise contraindicated.

Several techniques have been successfully applied to correct guidewire mishaps, which
may often be corrected by the introduction of a snare-tip catheter alone or in conjunction
with myocardial biopsy forceps (49). Successful filter retrieval has also been achieved by
using a combination of a guidewire and a snare (50). Several techniques of percutaneous
retrievals of vena cava filters have also been described with minimal risk, including a "rail
and reins" technique (51) and a transfemoral retrieval (52).

Because of the risk for arrhythmias and valvular insufficiency, intracardiac filter
retrievals need not be attempted if no other indications are present. Not all retrieval
attempts are successful. In one series, attempts at removal with a wire loop and sheath in
two cases failed and resulted in the migration of one filter to the right inferior pulmonary
artery (53) (Fig. 5).

The balloon of a 6F Fogarty catheter was successfully used to dilate the distal legs in
an incomplete filter expansion, adequately expanding the filter base to inhibit the clots and
prevent possible migration (54). An additional filter, either above or below a partially
expanded filter as appropriate, may also be considered to complete the objective of clot
entrapment. Such placement could also conceivably reduce the severity of a migration of
an incompletely expanded filter.

VI. CONCLUSION

With careful preoperative planning, proper patient and device selection, and meticulous
operative technique, the placement of vena cava filters by a properly trained surgeon is
most often a straightforward procedure with minimal adverse consequences. However, the
procedure is not without risk of complications and the observations presented here
underscore that fact.

In the presence of a suspected complication, computed tomography, aortography, and
ascending cavography may be used to demonstrate inferior vena cava penetration by filter
struts into the infrarenal aorta. Scanning electron microscopy can reveal structural defects
or corrosion in retrieved filters. Energy-dispersive radiographic analysis may demonstrate
impurities in the metal composition.

Recent studies have found that duplex ultrasound-directed IVC filter placement is
safe, cost-effective, and convenient and that intravascular ultrasound is a more accurate
method, than contrast venography for localizing renal veins and measuring vena cava
diameter (55,56). Use of this technology can help to limit the incidence of misplaced filters
and reduce the chance of subsequent migrations due to suboptimal operative technique.

The need to monitor patients with IVC filters over the long term, preferably using ■§

computed tomography, should be considered in those patients in whom misplacement, g

migration, or other filter abnormalities have either occurred or are at risk for occurring. »

Complications are rare, but the physician must remain aware of their potential appearance c

in order to minimize their incidence. <

REFERENCES 1

9

1. Havig O. Deep vein thrombosis and pulmonary embolism. An autopsy study with multiple |

regression analysis of possible risk factors. Acta Chir Scand Suppl 1977; 478:1-120. 2

©

!

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VENA CAVA FILTERS 577

2. Braverman SJ, Battey PM, Smith RB III. Vena caval interruption. Am Surg 1992; 58(3): 188-1 92.

3. Alpert JS, Dalen JE. Epidemiology and natural history of venous thromboembolism. Prog
Cardiovasc Dis 1994; 36:417-422.

4. Juni JE, Abass A. Lung scanning in the diagnosis of pulmonary embolism: The emperor re-
dressed. Semin Nucl Med 1991; 21:282-296.

5. Horattas MC, Wright DJ, Fenton AH, Evans DM, Oddi MA, Kamienski RW, Shields EF.
Changing concepts of deep venous thrombosis of the upper extremity — Report of a series and
review of the literature. Surgery 1988; 104(3):561-567.

6. Gloviczki P, Kazmier FJ, Hollier LH. Axillary-subclavian venous occlusion: The morbidity of a
nonlethal disease. J Vase Surg 1986; 4(4):333-337.

7. Hingorani A, Ascher E, Hanson J, Scheinman M, Yorkovich W, Lorenson E, DePippo P, Salles-
Cunha S. Upper extremity versus lower extremity deep venous thrombosis. Am J Surg 1997;
174(2):214-217.

8. Monreal M, Lafoz E, Ruiz J, Vails R, Alastrue A. Upper-extremity deep venous thrombosis and
pulmonary embolism. A prospective study. Chest 1991; 99(2):280-283.

9. Alexander P, Giangola G. Deep venous thrombosis and pulmonary embolism: Diagnosis,
prophylaxis, and treatment. Ann Vase Surg 1999; 13(3):318-327.

10. Arcasoy SM, Kreit JW. Thrombolytic therapy of pulmonary embolism: A comprehensive review
of current evidence. Chest 1999; 1 15(6): 1695-1707.

1 1 . Langan EM III, Miller RS, Casey WJ III, Carsten CG, Graham RM, Taylor SM. Prophylactic
inferior vena cava filters in trauma patients at high risk: Follow-up examination and risk/benefit
assessement. J Vase Surg 1999; 30(3):484-490.

12. Hingorani A, Ascher E, Ward M, Mazzariol F, Gunduz Y, Ramsey PJ, Yorkovich W. Combined
upper and lower extremity deep venous thrombosis. Cardiovasc Surg 2001; 9(5):472^177.

13. Ascher E, Hingorani AP, Yorkovich WR. Vena cava filter placement. In: Cameron J, ed. Current
Surgical Therapy. 7th ed. St Louis: Mosby, 2001.

14. Ascher E, Hingorani AP, Yorkovich WR. Inferior vena cava filter placement. In: Ahn S, Moore
WS, eds. Endovascular Surgery. 3rd ed. Philadelphia: Saunders, 2000.

15. Yorkovich WR, Ascher E. Vena cava filter placement. In: Ahn S, Moore WS, eds. Endovascular
Surgery For Venous Disease: Handbook of Endovascular Surgery. 3rd ed. Sec XI. Philadelphia:
Karger-Landes, 1999.

16. Ascher E, Hingorani AP, Tsemekhin B, Yorkovich WR, Gunduz Y. Lessons learned from a 6-
year clinical experience with superior vena cava Greenfield filters. J Vase Surg 2000; 32(5):881-
887.

17. Ascher E, Hingorani A, Mazzariol F, Jacob T, Yorkovich WR, Gade P. Clinical experience with
superior vena cava Greenfield filters. J Endovasc Surg 1999; 6(4):365-369.

18. Ascer E, Yorkovich WR. Superior vena cava Greenfield filters: Indications, techniques and
results. Eur Phlebol Digest, 1996.

19. Granke K, Abraham FM, McDowell DE. Vena cava filter disruption and central migration due
to accidental guidewire manipulation: A case report. Ann Vase Surg 1996; 10(l):49-53.

20. Rogers F, Lawler C. Dislodgement of an inferior vena cava filter during central line placement in
an ICU patient: A case report. Injury 2001; 32(10)787-788.

21. Ferreiro C, Abad-Cervero JF. Lopez-Pino MA. Fracture and migration of an Antheor vena cava j>
filter. J Vase Intervent Radiol 1996; 7(1): 149-150. |

22. James KV, Sobolewski AP, Lohr JM, Welling RE. Tricuspid insufficiency after intracardiac s
migration of a Greenfield filter: Case report and review of the literature. J Vase Surg 1996; °|
24(3):494^198. ^

23. Feezor RJ, Huber TS, Welborn MB III, Schell SR. Duodenal perforation with an inferior vena >9
cava filter: An unusual cause of abdominal pain. J Vase Surg 2002; 35(5): 1-3. J

24. Dabbagh A, Chakfe N, Kretz JG, Demri B, Nicolini P, Fuentes C, Mettauer B, Epailly E, Muster q

D, Eisenmann B. Late complication of a Greenfield filter associating caudal migration and |

perforation of the abdominal aorta by a ruptured strut. J Vase Surg 1995; 22(2): 182-187. S

©

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

578 ASCHER et al.

25. Seita J, Sakakibara Y, Jikuya T, Shigeta O, Nakata H, Tsunoda H, Mitsui T. Surgical man-
agement of a penetrated Greenfield inferior vena cava filter. Thorac Cardiovasc Surg 2001; 49(4):
243-244.

26. Bochicchio GV, Scalea TM. Acute caval perforation by an inferior vena cava filter in a mul-
titrauma patient: Hemostatic control with a new surgical hemostat. J Trauma 2001; 51(5): 991-
992; discussion, 993.

27. Bochicchio GV, Scalea TM, Adams R. Acute caval perforation by an inferior vena cava filter in a
multitrauma patient: Hemostatic control with a new surgical hemostat. J Trauma 2001; 51(5):
991-992; discussion 993.

28. Burbridge BE, Walker DR, Millward SF. Incorporation of the Gunther temporary inferior vena
cava filter into the caval wall. J Vase Interv Radiol 1996; 7(2):289-290.

29. Lok SY, Adkins J, Asch M. Caval perforation by a Greenfield filter resulting in small-bowel
volvulus. J Vase Interv Radiol 1996; 7(l):95-97.

30. al Zahrani HA. Bird's nest inferior vena caval filter migration into the duodenum: A rare cause of
upper gastrointestinal bleeding. J Endovasc Surg 1995; 2(4):372-375.

31. Guillem PG, Binot D, Dupuy-Cuny J, Laberenne JE, Lesage J, Triboulet JP, Chambon JP.
Duodenocaval fistula: A life-threatening condition of various origins. J Vase Surg 2001; 33(3):
643-645.

32. Loehr SP, Hamilton C, Dyer R. Retrieval of entrapped guide wire in an IVC filter facilitated with
use of a myocardial biopsy forceps and snare device. J Vase Intervent Radiol 2001; 12(9): 11 16-
1119.

33. Sing RF, Adrales G, Baek S, Kelley MJ. Guidewire incidents with inferior vena cava filters. J Am
Osteopath Assoc 2001; 101(4):231-233.

34. Duong MH, Jensen WA, Kirsch CM, Wehner JH, Kagawa FT. An unusual complication during
central venous catheter placement. J Clin Anesth 2001; 13(2): 13 1-1 32.

35. Vandemergel X. Extensive vena cava thrombosis and massive pulmonary embolism despite the
presence of a vena cava filter and an optimal anticoagulant. Rev Med Liege 200 1 ; 56( 1 2):807-808.

36. Barreras JR, Agarwal DM, Maximin ST, Friedman A. Recurrent pulmonary embolism despite
the use of a Greenfield filter. Clin Nucl Med 2001; 26(12): 1040-1041.

37. Greenfield LJ, Proctor MC. Recurrent thromboembolism in patients with vena cava filters. J
Vase Surg 2001; 33(3):5 10-514.

38. Kinney TB, Rose SC, Lim GW, Auger WR. Fatal paradoxic embolism occurring during IVC
filter insertion in a patient with chronic pulmonary thromboembolic disease. J Vase Intervent
Radiol 2001; 12(6):770-772.

39. Ionita C, Giglio P, Isayev E, Alberico R, Pullicino P. Paradoxical brain embolism from thrombus
associated with vena caval filter in a patient with cancer. J Neuroimaging 2002; 12( 1 ):69 71 .

40. Georgopoulos SE, Chronopoulos A, Dervisis KI, Arvanitis DP. Paradoxical embolism. An old
but, paradoxically, under-estimated problem. J Cardiovasc Surg (Torino) 2001; 42(5):675-677.

41. Marcy PY, Magne N, Frenay M, Bruneton JN. Renal failure secondary to thrombotic com-
plications of suprarenal inferior vena cava filter in cancer patients. Cardiovasc Intervent Radiol
2001; 24(4):257-259.

42. Flanagan D, Creasy T, Chataway F, Kerr D. Caval umbrella causing obstructive uropathy.

Postgrad Med J 1996; 72(846):235-237. 1

43. Goldman HB, Hanna K, Dmochowski RR. Ureteral injury secondary to an inferior vena caval <S
filter. J Urol 1996; 156(5):1763. I

44. Berger BD, Jafri SZ, Konczalski M. Symptomatic hydronephrosis caused by inferior vena cava °|
penetration by a Greenfield filter. J Vase Intervent Radiol 1996; 7(1):99 101. ^

45. Wambeek ND, Frazer CK, Kumar A. Penetration of a vertebral body by a limb of the Greenfield >9
filter. Australas Radiol 1996; 40(3):364-366. J

46. Phlegmasia cerulea dolens: A complication of use of the filter in the vena cava. J Bone Joint Surg q
Am 1995; 77(1 1): 1783; Comment, J Bone Joint Surg Am 1995;77(3):452-454. |

47. Defraigne JO, Vahdat O, Lacroix H, Limet R. Proximal migration of vena caval filters: Report of S

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF VENA CAVA FILTERS 579

two cases with operative retrieval. Ann Vase Surg 1995; 9(6):571— 575; Comment, Ann Vase Surg
1997;11(1):106.

48. Seita J, Sakakibara Y, Jikuya T, Shigeta O, Nakata H, Tsunoda H, Mitsui T. Surgical man-
agement of a penetrated Greenfield inferior vena cava filter. Thorac Cardiovasc Surg 2001; 49(4):
243-244.

49. Loehr SP, Hamilton C, Dyer R. Retrieval of entrapped guide wire in an IVC filter facilitated with
use of a myocardial biopsy forceps and snare device. J Vase Intervent Radiol 2001; 12(9): 1 1 16-
1119.

50. Lin M, Soo TB, Horn LC. Successful retrieval of infected Gunther Tulip IVC filter. J Vase
Intervent Radiol 2000; 1 1(10): 1 341 1343.

51. Liddell RP, Spinosa DJ, Matsumoto AH, Angle JF, Hagspiel KD. Guidewire entrapment in a
Greenfield IVC Filter: "Rail and reins technique." Clin Radiol 2000; 55(1 1):878— 881.

52. Salamipour H, Rivitz SM, Kaufman J A. Percutaneous transfemoral retrieval of a partially
deployed Simon-Nitinol filter misplaced into the ascending lumbar vein. J Vase Intervent Radiol
1996;7(6):917-919.

53. Gelbfish GA, Ascer E. Intracardiac and intrapulmonary Greenfield filters: A long-term follow-
up. J Vase Surg 1991; 14(5):614-617.

54. Danikas D, Constantinopoulous GS, Stratoulias C, Ginalis EM. Use of a Fogarty catheter to
open an incompletely expanded Vena Tech-LGM vena cava filter — A case report. Angiology
2001; 52(4):283-286.

55. Conners MS III, Becker S, Guzman RJ, Passman MA, Pierce R, Kelly T, Naslund TC. Duplex
scan-directed placement of inferior vena cava filters: A five-year institutional experience. J Vase
Surg 2002; 35(2):286-291.

56. Ashley DW, Gamblin TC, Burch ST, Solis MM. Accurate deployment of vena cava filters:
Comparison of intravascular ultrasound and contrast venography. J Trauma 2001; 50(6):975-
981.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

36

Complications of PercutaneousTreatment
of Arteriovenous Malformations

Robert J. Rosen and Thomas Maldonado

New York University Medical Center, New York, New York, U.S.A.

Congenital vascular malformations are among the most difficult lesions to treat in all of
vascular surgery. Even in the most experienced hands, the results of treatment are often
more palliative than curative, and the risk of significant complication is always present. The
risks cannot be eliminated but can be minimized by carefully defining the nature of the
lesion, educating the patient and family as to the natural history, and determining the type of
treatment indicated, if any. Many of the complications of treatment in these cases are due to
initial misdiagnosis — a common problem in this confusing group of disorders. Some of these
lesions follow a benign clinical course, while others can produce life-threatening complica-
tions and still others involute spontaneously and completely in childhood. It behooves the
clinician caring for these patients to correctly identify the problem; only then can an
appropriate risk-benefit decision be made regarding therapy.

This chapter deals primarily with complications of percutaneous treatment of vascular
malformations. These can be divided into complications related to technique, those related
to specific anatomic regions, and complications related to specific embolic agents or de-
vices. Prior to discussing these issues, it is necessary to briefly review the basic types of
vascular anomalies in order to avoid the most fundamental mistake of misdiagnosis.

I. TYPES OF VASCULAR ANOMALIES I

There are many systems of classification of vascular anomlies, most unnecessarily complex H,

and exhaustive for the purpose of this chapter. We have found that a simple division of =j

these lesions into four major types is sufficient to make clinical decisions regarding the g

natural history of the condition and the type of treatment required. This classification is £

derived from the work of Mulliken, Folkman, and Glowacki (1-3) and includes: j|

1 . Hemangioma §

2. Arteriovenous fistula @

581 f

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

582

ROSEN and MALDONADO

3. Arteriovenous malformation

4. Venous and lymphatic malformation

Hemangiomas are not malformations but benign vascular tumors encountered at
birth or in early infancy. They may occur on the skin, where they present as the classic
"strawberry birthmark," or elsewhere in the body, including the viscera (Fig. 1A and B).
They may be single or multiple and are much more common in females than males. From
a clinical point of view, the most important aspect of these lesions is their natural tendency
to involute spontaneously during childhood, so that only a minority of these patients will
require specific treatment. Often, education and reassurance of the parents are the only
interventions required. There are unusual cases where treatment is required during the
proliferative phase, particularly where the lesion involves respiratory or digestive struc-
tures or interferes with visual development. Rarely, extensive hepatic hemangiomas may
cause high-output congestive heart failure and require urgent intervention (4,5).

The vast majority of arteriovenous fistulas (AVF) are acquired, although congenital
types do exist, primarily in the central nervous system. An AVF, by definition, consists of
a direct communication between artery and vein, generally resulting in high flow with
secondary findings of arterial and venous dilation, venous hypertension, and often distal
ischemia due to the proximal steal phenomenon. It has long been known that these lesions
are curable, but only by directly interrupting or isolating the fistula itself. This principle is
still violated on a regular basis by both vascular surgeons and interventional radiologists
who occlude proximal feeding vessels using either ligation or embolic devices. Both will
invariably result in persistence of the fistula with almost instantaneous recruitment of col-
lateral feeding vessels, producing a lesion that is much more difficult to treat (Fig. 2). Some
of these initially simple fistulas recruit so many new collaterals that they are indistinguish-

B

> V -

**%

5

- ■

i

Figure 1 Misdiagnosis of vascular lesions can lead to serious errors in treatment. On a single
physical examination, many biologically distinct lesions can have a superficially similar appearance.
A. This 10-month-old child has an extensive hemangioma over the buttock and leg, which is starting
to show early signs of spontaneous involution. No specific treatment is required. B. Another child
with a similar appearing lesion on the foot. This is a venous malformation, which will grow with the
child and will never show spontaneous involution.

i
2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

TREATMENT OF ARTERIOVENOUS MALFORMATIONS 583

Figure 2 This 48-year-old carpenter sustained a penetrating injury to the palm 4 years previously,
resulting in an arteriovenous fistula. Rather than occluding the fistula itself, the radial artery was
ligated, leading to recruitment of numerous collaterals and a series of secondary procedures,
eventually resulting in amputation of a significant portion of the hand. This angiogram demonstrates
profuse collateral development resulting in a lesion as difficult to treat as a complex congenital
malformation. This fundamental error in treatment can result from proximal embolization as well
as ligation.

able from complex congenital malformations. As in most surgical scenarios, the best
chance of cure comes with the initial procedure, which must be directed at obliteration of
the fistula itself.

True arteriovenous malformations (AVM) are functionally similar to fistulas, in the
effect of a shunt between artery and vein, although the communication is more complex and
represents a continuum from microfistulous connections just above capillary level to
macrofistolous types. These are congenital lesions that are present at birth by definition
and generally grow at the same rate as the individual. They may or may not cause symp-
toms or even be detected, but they never involute spontaneously. As some of these lesions in
children can resemble hemangiomas, the potential for confusion in terms of prognosis and
treatment planning is evident. Experienced observers can usually distinguish between the j>

two, but some children must be followed over time before a correct diagnosis can be made e

confidently. AVMs are notoriously difficult to treat due to their complex blood supply and H,

tendency to recur; since these are slow-growing lesions, there should be no rush to treat, and s

a careful risk-benefit analysis should be made in conjunction with the patient and family. J

Lesions may require intervention due to mass effect, distal ischemia, venous hypertension, jf

growth disturbance, or, more rarely, hemorrhage or high-output states. The latter two 1

situations, while often the major concern of parents and referring physicians, are actually fj

quite uncommon, and prophylactic treatment is not warranted. §

o

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

584

ROSEN and MALDONADO

Treatment failures and complications in patients with AVMs are similar to those
outlined above for AVFs. That is, treatment must be directed at eradicating the actual
arteriovenous communication, rather than occluding or ligating feeding vessels, which will
only result in rapid collateral recruitment. It is obviously much more difficult to interrupt a
complex network of feeding vessels than a simple fistula, particularly when there may be
no clear delineation between those vessels feeding the AVM and those required for nutri-
tive blood flow to a region. It is this complexity that makes many interventions palliative
rather than curative. Only the most localized and accessible AVMs are amenable to com-
plete surgical resection. Partial resections are often complicated by hemorrhage and rapid
recurrence, often with exacerbation of the patient's clinical symptoms. Similarly, surgical
"skeletonization," meticulous ligation of all feeding vessels in the region, is rarely effective
in the long term and often sacrifices transvascular access, making subsequent treatment
more difficult (6-8) (Fig. 3A and B).

Venous and lymphatic malformations present clinically in a variety of ways, depending
on the anatomical location and depth of the lesion. Superficial venous malformations have a
typical bluish discoloration, are soft and compressible, and demonstrate no pulsation or
bruits on physical examination. If in an extremity, they may show dramatic enlargement
when placed in a dependent position. While these lesions are usually painless, patients may
complain of periodic pain and tenderness due to spontaneous thrombosis, or they may
experience discomfort due to distension of the lesion after activity or in certain positions.
Klippel-Trenaunay syndrome is one of the most common venous anomalies encountered
clinically, generally presenting as venous varicosities involving a single extremity, usually the
leg (Fig. 4A and B). A significant number of patients with this syndrome have aplasia or
hypoplasia of the deep venous system, such that essentially all of the venous drainage from
the extremity is via the superficial veins. Venous stripping in these patients can cause a
dramatic worsening of symptoms. Any patient with purely unilateral varicosities should

i

2

I

■c

Figure 3 Surgical "skeletonization" is another form of proximal vessel occlusion that is generally
ineffective in treating vascular malformations. This 22-year-old patient with a high-flow AVM of the
left pelvis and thigh underwent an extensive skeletonization procedure 4 months previously. After
studies for recurrent symptoms, the angiogram shows successful ligation of all right iliac and
femoral branches (A) but resupply of the lesion via multiple enlarged middle sacral and lumbar
collaterals (B).

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

TREATMENT OF ARTERIOVENOUS MALFORMATIONS

585

Figure 4 Klippel-Trenaunay syndrome (^4) is one of the most commonly encountered venous
malformations. The distinctive triad includes unilateral varicose veins, hypertrophy of the involved
extremity, and cutaneous pigmented lesions in the same area. A serious potential error is to strip or
sclerose the varicose veins without initially performing a venogram, as a significant percentage of
these patients will have aplasia or hypoplasia of the deep venous system (B). Occlusion or removal of
the superficial veins may result in acute venous insufficiency.

therefore be investigated with detailed contrast venography prior to any treatment. In ex-
treme cases, the malformation may extend to the pelvis or retroperitoneum. Lymphatic mal-
formations are less common and range from cutaneous lesions with vesicle formation to
large soft tissue masses like the cystic hygroma. One of the distinctive features of lymphatic
malformations is their tendency to recurrent infection (9,10).

II. EMBOLIZATION TECHNIQUES

Many techniques have been described for the percutaneous treatment of vascular malfor-
mations. The two major approaches are transcatheter embolization, used in lesions with a
significant arterial component, and direct puncture techniques, mainly used in low-flow
venous or lymphatic malformations. Transcatheter treatment involves percutaneous ac-
cess to the arterial system, selective catheterization of arterial branches feeding the malfor-
mation, and the injection of occluding devices or agents. Complications may occur related
to the catheterization procedure itself, the anatomical region involved, or the specific em-
bolic device or agent used.

Complications related to arterial access and selective catheterization should be un-
usual in experienced hands but can occur, particularly in pediatric patients. Problems can
include vasopasm, dissection of an artery, and arterial thrombosis, either at the entry site
or in arterial branches being selectively catheterized during the procedure. Entry site prob-
lems can be minimized by routinely using vascular sheaths, which protect the vessel during
catheter manipulation, especially when multiple catheter exchanges may be required. The
pulse should always be checked after gaining vascular access; if it is no longer palpable,

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

586 ROSEN and MALDONADO

antispasm agents should be administered (e.g., intra-arterial nitroglycerine) and the pa-
tient should be heparinized. This is a problem encountered most often in pediatric patients
and young females. The pulse will generally return following sheath removal but may be
delayed for several hours, during which time the extremity must be observed closely. Man-
agement of the patient whose pulse does not return is controversial — a problem, again,
seen mostly in infants and children. While these patients will rarely show signs of ischemia,
there is literature documenting the possibility of future growth disturbance and leg-length
discrepancy (11). The increased risk of arterial injury in very young patients should be
taken into consideration in making the decision on the timing of intervention.

The actual embolization procedure may be complicated by occluding a feeding vessel
too proximally or too distally or occluding the wrong vessel altogether. Proximal occlusion
of arterial feeders to an AVM, while simple to perform, virtually always results in long-
term failure. This is entirely analogous to the proximal ligation of an artery supplying an
AVF; collateral vessels will be recruited almost immediately, leading to a recurrence with a
much more complex blood supply which is much more refractory to subsequent treatment
(12). Just as the goal in surgically treating an AVF is closure of the fistula itself, the goal of
embolization in an AVM must be penetration and eradication of the nidus. This goal often
cannot be completely achieved in practice, but at the very least, proximal occlusions are to
be avoided. Thus, embolic devices such as coils, detachable balloons, or large pledgets of
Gelfoam have no role in the treatment of complex lesions other than as protective devices
for normal branches in the region.

Conversely, embolic materials that are too small or fine can cause other problems. If the
material is smaller than the arteriovenous connections, it will shunt through to the venous
circulation and ultimately to the lungs (13-15). The otherwise healthy lung actually has a
remarkably high tolerance for this type of embolization, and the patient will rarely dem-
onstrate any signs, symptoms, or changes in pulmonary function studies. Nevertheless, this
should be avoided by choosing the appropriate embolic agent. Materials that are extremely
penetrating, such as Gelfoam powder, may cause damage to normal tissues if they enter
normal parts of the arterial circulation. Nontarget embolization may occur through a vari-
ety of mechanisms, including reflux of the embolic material, catheter recoil, device mig-
ration, insecure catheter position, and failure of the agent to be completely extruded from
the catheter tip. Experience, the use of coaxial microcatheter systems, and scrupulous tech-
nique can reduce but not eliminate this type of complication. Depending on the location of
the nontarget embolization, the clinical result may range from completely unapparent to an
ischemic digit and even to stroke (16). Large embolic devices such as coils can often be
retrieved from a nontarget vessel using either interventional or surgical techniques, but most
of the agents used in AVM treatment are not retrievable.

Following aggressive embolization of especially large arteriovenous malformations,
blood flow may be redistributed toward other circulation beds. The sudden increased "g

pressure in these previously less perfused vessels may result in rupture and significant &

hemorrhage. This relatively unusual but potentially devastating complication is common a

to all embolic agents in treating large AVMs. -c

III. SPECIFIC EMBOLIC AGENTS AND PROBLEMS RELATED §

TO THEIR USE |

A variety of embolic agents have been employed in the treatment of vascular malforma- I

tions over the past 30 years. Some of these agents remain in common use, others have been ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

TREATMENT OF ARTERIOVENOUS MALFORMATIONS 587

abandoned, and still others remain investigational. Some of the features and problems of
these agents are considered below.

A. Autologous Tissue or Clot

Among the earliest agents described for the intentional occlusion of blood vessels, these
materials are no longer in clinical use due to the unpredictable location and duration of
vascular occlusion (17).

B. Gelfoam

Gelfoam is available in a variety of preparations, from sponges to powder. While these are
still used as embolic agents in some institutions for the control of gastrointestinal bleeding
and traumatic hemorrhage, the have little use in treating AVMs due to their variable
duration of occlusion. Gelfoam powder can be extremely dangerous when injected intra-
arterially, as the small particle size often leads to tissue necrosis (18).

C. Particles and Microspheres

Two of the most common embolic agents in use are polyvinyl alcohol (PVA) particles and
microspheres. These agents are available in graded sized from 50-1000 um and are injected
as a suspension, generally in radiographic contrast. They are thus quite simple to use and can
be injected through almost any catheter or microcatheter. The major difference between
these two agents is that standard PVA particles are irregularly shaped, being ground
mechanically from a sponge material, while the microspheres are smoothly spherical. Both
of these agents are non-resorbable, but PVA particles are often associated with early
recanalization and AVM recurrence, as the irregular particles are incorporated into vessel
walls and new channels are formed (19,20). The experience with microspheres is too limited
at this point in determining whether the long-term results will be superior, although this
should theoretically be true. In addition, new preparations of spherical PVA have recently
been introduced.

Depending on the vascular bed being treated, the risks of particulate embolization
primarily relate to improper sizing of the particles or spheres. Particles that are smaller
than the AV communication will obviously be shunted through the lesion and lodge in the
pulmonary circulation. Since the particles and spheres are themselves radiolucent, this
complication may not be immediately apparent. Although experience has shown that
inadvertent pulmonary embolization is remarkably well tolerated in the normal lung, large
amounts of embolic material or smaller amounts in patients with previously compromised
pulmonary function may have significant consequences (21). Extremely small particles
(less than 50 urn) that reach normal nontarget vessels may also cause significant ischemic
complications, particularly in the gastrointestinal tract (22). Particles that are too large to -o

enter the nidus of the malformation will act as proximal occlusions, resulting in early |

recurrence of the lesion. %

D. Detachable Balloons

These devices consist of latex or silicone balloons with a self-sealing valve that are mounted
on a microcatheter and are available (in certain countries) in a variety of shapes and sizes.
They offer both distinct advantages and disadvantages. They can be flow-guided fluoro-
scopically to a high-flow lesion and then inflated and deflated until optimal positioning is
attained, at which point they are detached by applying traction to the microcatheter. They

•a

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

588 ROSEN and MALDONADO

are used primarily to treat high-flow fistula-type lesions such as pulmonary AVMs and
carotid-cavernous fistulas. The major disadvantages of this device are its complexity, cost,
and the occasional premature deflation of the balloon, which can result in reopening of the
fistula and possible paradoxical embolization (23). At the present time they are available
only on an investigational basis in the United States.

E. Embolization Coils

Coils have been used for transcatheter embolization for over 20 years. They consist of
stainless steel or platinum and are supplied in straightened form inside a cartridge; they are
pushed through the catheter with a wire or injected with a push of saline, and reassume their
coiled shape as they exit the catheter tip. Most coils have fibers attached to increase their
occlusive properties. They are available in a wide range of sizes and shapes, including micro-
coils, which can be passed through microcatheters. More sophisticated coils have recently
been introduced that have specialized properties, including extreme flexibility, allowing the
coil to behave almost like a liquid (useful in aneurysms), and complex detachment systems
that allow for precise positioning prior to releasing the device. Standard coils are relatively
inexpensive and easy to use but are not suitable for AVM embolization, as they are equiva-
lent to a vessel ligation and recurrence is virtually guaranteed while access has been sacri-
ficed. In certain anatomical situations, proximal coil occlusion may actually be performed to
protect normal branches from distal liquid or particulate embolization. Probably the
commonest complication associated with coil embolization is misplacement or migration
of the device; this can be minimized by ensuring a stable catheter position prior to extruding
the coil as well as using coaxial catheter systems. When the catheter tip is near the ostium of a
vessel, the extruding coil may push the catheter tip backward and allow the coil to embolize
distally into the normal circulation. Miniature snares are now widely available that usually
but not always allow a misplaced coil to be retrieved transluminally.

F. Absolute Ethanol

Ethanol has been used in both arterial and venous malformations, both via transcatheter
injection and direct puncture techniques (21,24,25). Essentially a sclerosing agent, ethanol
causes intravascular thrombosis and endothelial damage, leading to occlusion, inflamma-
tion, and fibrosis. When ethanol is used in venous lesions by direct injection, these properties
are desirable, but when it is injected intra-arterially, the risk of complications is considerable.
Its direct tissue toxicity can result in sloughing, nerve damage, and mucosal ulceration or
perforation (26,27) (Fig. 5A and B). Acute systemic reactions to ethanol may also occur (see
Sec. IV. E, below). Since ethanol is nonviscous and radiolucent, extreme care must be
exercised to control its distribution and avoid normal structures.

G. Liquid Adhesives I

Acrylic adhesives were first introduced 30 years ago for the surgical management of solid- a

organ trauma, particularly the liver. They were subsequently used to occlude blood vessels -c

supplying vascular malformations, although these agents have been available only inter- <j

mittently in the United States. The only agent of this type currently available is N-butyl >9

cyanoacrylate (NBCA), which received approval from the U.S. Food and Drug Admin- 4j

istration in 2000 for use in central nervous system lesions. The agent is supplied in liquid 2

form and is generally mixed with an oily radiographic contrast agent (ethiodol) that pro- |

vides radiopacity and slows the polymerization time, which would otherwise be virtually ©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

TREATMENT OF ARTERIOVENOUS MALFORMATIONS

589

Figure 5 Pelvic angiogram of 32-year-old woman with right pelvic discomfort, which was found to
be caused by a high-flow AVM supplied by the right hypogastric artery (A). Repeat angiography
following intra-arterial absolute ethanol embolization (B) shows marked reduction in flow to the
lesion. The patient developed colon and bladder perforations requiring colostomy and partial bladder
resection as well as permanent nerve injury with foot drop and chronic leg pain. Ethanol is extremely
tissue-toxic and may be associated with severe complications when injected intra-arterially.

Figure 6 A 35-year-old male with a high-flow arteriovenous malformation in the plantar aspect of
the foot, causing steal, with resultant pain and chronic ulceration of the toes (A). Repeat study (B)
following superselective embolization with TY-butyl cyanoacrylate (NBCA) tissue adhesive shows
marked reduction in flow to the nidus with improved distal flow. The patient's pain and ulceration
resolved within 3 weeks. Although there is a residual nidus, the clinical problem was successfully
resolved. NBCA adhesive has little or no tissue toxicity and can be used safely in almost any part of
the circulation.

i
2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

590 ROSEN and MALDONADO

instantaneous. NBCA polymerizes on contact with any ionic media, including blood, tis-
sue, saline, and contrast agents. Its use therefore requires considerable training and ex-
perience as well as scrupulous attention to technique. The theoretical advantage of this
type of agent is its ability to penetrate the nidus of an AVM and form a cast, preventing
ingrowth of new collaterals. Complete obliteration of the nidus is, in fact, rarely achieved,
but a significant reduction in flow can generally be accomplished (Fig. 6A and B). We have
used this agent as our primary embolic material for high-flow AVMs for over 20 years and
have been impressed with the clinical results and relative freedom from adverse effects
(28,29).

There has been ongoing concern regarding possible long-term toxicity or even carci-
nogenicity with this agent, but there has never been any report of such an effect in humans
despite many years of use. Problems and complications with these adhesives have been
primarily mechanical and technical in nature, including nontarget embolization and gluing
a catheter to the vessel wall. This latter event, while frequently mentioned, rarely occurs in
clinical use and is nearly always associated with intracranial lesions where extremely small,
fragile microcatheters are used; these have less tensile strength than the glue cast (30). A
potential limitation of these cyanoacrylates is the possibility for dissolution and subse-
quent recanalization of the AVM on long-term follow-up. Resorption and recanalization
appears to be more likely if the AVM nidus is incompletely casted with adhesive (31,32).

IV. COMPLICATIONS ASSOCIATED WITH SPECIFIC ANATOMIC
REGIONS

Certain anatomic regions are associated with potential complications from percutaneous
treatment of vascular malformations due to the nature of the vascular bed or the end organ
involved. Neuroembolization procedures obviously carry significant risk, but these are
beyond the scope of this chapter and are not discussed here.

A. Pulmonary AVM

Most pulmonary AVMs are asymptomatic and constitute one of the few lesions that are

treated prophylactically to eliminate the risk of paradoxical embolization, which may result

in stroke or brain abscess (33,34). Complications related to embolization of pulmonary

AVMs consist of those involving the lung and those related to paradoxical embolization

during the procedure itself. Pulmonary complications related to embolization are unusual

due to the dual blood supply of the lung, so that pulmonary infarction is uncommon. When

this does occur, it is manifest as pleuritic pain, infiltrate, atelectasis, and pleural effusion.

These problems are nearly always self-limited, resolving in days. Passage of the embolic

material or device through the malformation is a much more serious risk, presenting the e

possibility of stroke — the very problem the procedure was intended to protect against. This e

event is fortunately rare and usually associated with improper sizing of the coil or balloon, H,

premature balloon detachment, or premature balloon deflation. Interestingly, the only a

patient in whom we encountered this complication had none of the above problems during g

the procedure but awoke the morning following the procedure with a stroke (unpublished sf

data). The hypothesis was that, after balloon occlusion, thrombus accumulating in the pul- g

monary artery branch proximal to the occluded AVM was swept through another smaller, fj

angiographically undetected AVM. g

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

TREATMENT OF ARTERIOVENOUS MALFORMATIONS 591

B. Gastrointestinal Tract

Other than in patients with Rendu-Osler-Weber syndrome, most vascular malformations
in the gastrointestinal (GI) tract are angiodysplasias found in the colon, generally in older
patients with intermittent bleeding. While most of the GI tract is so well collateralized that
the risk of ischemia is slight, the colon is sparsely vascularized and presents a significant
risk of mucosal ischemia with subsequent stricture or even infarction following emboliza-
tion (22,35,36). The appropriate level of occlusion is controversial, as the occlusion must
be distal enough to stop the bleeding but not so distal that there is insufficient collateral
flow to maintain viability. NBCA and PVA particles may be used safely, and this is one of
the few locations where microcoils may be an acceptable alternative. Ethanol should never
be used intra-arterially in the GI tract.

An unusual complication that we have encountered twice is portal vein thombosis
following embolization of a high-flow AVM in the small intestine with shunting into the
portal system. Both of these patients presented with symptoms of severe portal hyper-
tension (variceal bleeding, intractable ascites) necessitating treatment. Both were treated
with superselective NBCA adhesive with an excellent angiographic result and no loss of
the agent into the portal system. Both did well initially, but within a week they thrombosed
their entire portal system and subsequently expired (unpublished data).

C. Liver, Spleen, and Kidney

As in the lung, the dual blood supply of the liver generally permits safe embolization of
vascular malformations. The risk of ischemia is increased in the presence of severe portal
hypertension as well as portal vein and biliary obstruction (37).

The spleen is a rare location for vascular malformations but presents a risk that is
unusual elsewhere in the body following embolization — the complication of infarction
followed by infection with abscess formation (38). For this reason, any patient undergoing
splenic embolization should be placed on prophylactic intravenous antibiotics, which
should be continued for at least a week following the procedure. Another potential com-
plication is pancreatitis or pancreatic infarction due to embolic material that has entered
the small pancreatic branches originating from the splenic artery (39). This complication
has occurred primarily when liquid agents have been used. Renal vascular malformations
tend to be of two main types. The first is the small angiomatous lesion associated with
intermittent gross hematuria; it is usually found in the right kidney of adult females. These
lesions are supplied by small intrarenal end vessels; therefore they can be completely cured
by embolization. However, a small renal infarction will generally occur, with the patient
experiencing flank pain and fever (40,41) (Fig. 7A). This "postembolization syndrome" is
virtually always self-limited and resolves in a matter of days. The second type is the large
congenital intrarenal fistula, which may present with flank pain, hematuria, a bruit, and
occasionally a high-output state. Interestingly, this lesion also appears to favor the right g

kidney and is often misdiagnosed on imaging studies, because of the massive dilation of the |

draining veins, as a giant intrarenal aneurysm. This lesion is also curable with embolization, s

but the patient may experience considerable postoperative discomfort due to the throm- jS

bosis of the draining veins. d

D. Extremities |

Q

Extremity lesions can be among the most difficult malformations to treat, with a "g

significant risk of complications (27,42,43). This is due to the difficulty in separating the g

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

592

ROSEN and MALDONADO

Figure 7 Vascular malformations involving "end vessels" such as the kidney (^4) or digits (B) are
extremely difficult to treat without causing ischemia of normal tissues. In the kidney, the segmental
infarction that would follow embolization of this small AVM, causing hematuria, would be well
tolerated, while in the finger (5), embolization carries a real risk of tissue loss.

vessels supplying the malformation from those necessary for maintaining distal viability.
This is particularly true in patients who have undergone previous surgical ligations (as
many have) or embolizations (Fig. 7A and B). The more distal the lesion, the higher the
difficulty in terms of both the catherization technique and preserving distal flow (Fig. 7B).
Thus hand and foot lesions are particularly refractory to achieving a satisfactory result
(28). The end-vessel digital circulation is also unforgiving of any nontarget embolization,
presenting the real risk of loss of the digit. Infants and young children appear to have a
better chance of reperfusion than adult patients in this situation.

Ulceration is not uncommon in high-flow extremity malformations; it is related to
both tissue ischemia from a steal phenomenon as well as venous hypertension. Particularly
in previously treated patients, it can sometimes be diffcult to determine which mechanism
is predominantly responsible for the ulceration. If ischemia is the primary mechanism,
further embolization may actually worsen the clinical problem, while venous hypertension
will be improved by embolization, which reduces the shunt.

E. Venous Malformations

Venous malformations range from focal cavernous lesions to diffuse intramuscular lesions
to anomalous venous channels such, as those seen in Klippel-Trenaunay syndrome. By
definition, they show little or no arteriovenous shunting on angiography. Embolization of
the arterial branches supplying the area of these venous lesions has been tried in the past,
and — not surprisingly — has had no beneficial effect. As with any vascular malformation,
treatment is indicated only for significant symptoms or complications; the mere presence
of a venous lesion, once its nature has been determined, is not an indication for treatment.
Symptoms may include the presence of a soft tissue mass, pain — particularly after activity
or when the lesion is in a dependent position, ulceration, secondary signs of venous hyper-
tension, and, rarely, bleeding.

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i
2

I

■c

I

u

TREATMENT OF ARTERIOVENOUS MALFORMATIONS 593

Treatment options (aside from the usual conservative measures) include surgical re-
section and direct embolization. Surgical resection is often more difficult than anticipated,
as the visible or palpable component of the lesion may only be the "tip of the iceberg,"
with extension into deep tissues and surrounding structures. Detailed imaging studies are
mandatory prior to attempted resection, with magnetic resonance imaging (MRI) studies
providing the most accurate depiction of the extent of the lesion. Only the most localized
of these malformations are suitable for surgical resection, and recurrence after surgery is
common (44).

Direct embolization of venous malformations is a form of sclerotherapy. Using radio-
logical guidance [ultrasound, fluoroscopy, computed tomography, or MRI], the lesion is
entered directly with a sheathed needle. Contrast material is then injected to outline the
lesion and determine the approximate volume, with specific attention to the route of
venous drainage. Some cavernous lesions are almost completely isolated from the normal
deep veins in the region, while others show large, relatively rapid communication with the
systemic venous circulation. Particularly when there is significant communication with
normal veins, the lesion must be isolated, using either tourniquet or cuff compression or
direct compression of the draining vein. When a lesion is in an extremity, the deep venous
system is generally flushed continuously with a heparinized saline solution to reduce the
risk of deep venous thrombosis from the sclerosing agent. The commonest sclerosing
agents are absolute ethanol and sotradecol solution, both of which are quite tissue-toxic.
Leakage of the agent from the lesion out to a skin or mucosal surface may result in ul-
ceration; the entry tract is generally injected with a collagen suspension as the catheter is
withdrawn to reduce this risk. Venous malformations in the calf or forearm may be asso-
ciated with compartment syndromes following embolization. Inflammation and resulting
edema can be reduced by administering steroids before and after treatment. Furthermore,
in the case of extensive lesions, embolizations should be staged and performed slowly.

Major complications have been reported during ethanol embolization when a sig-
nificant amount of the agent escapes into the venous circulation (25). The most severe is
cardiovascular collapse, which may occur with little warning and is thought to be related
to an acute direct vasospastic response of the central pulmonary vasculature to the etha-
nol. In some cases, a large bolus of ethanol soaked thrombus may also embolize centrally
and cause a similar clinical picture. While in our institution virtually all of these pro-
cedures are performed under general anesthesia with close physiological monitoring, some
authors have advocated routine placement of Swan-Ganz catheters to monitor pressures
throughout the procedure (25). We have found that using radiographic contrast, fluoro-
scopic monitoring, and limiting the total volume of ethanol used in a single procedure
(generally 0.5 mL/kg maximum) makes this type of complication exceedingly rare.

V. PEDIATRIC ISSUES 1

&

A significant number of patients being diagnosed and treated for vascular malformations a

will be in the pediatric age range. In this group of patients, a team approach to the prob- c

lem is particularly important. The need for intervention and its timing is often an issue. <j

The first critical step is making an accurate diagnosis, which may sometimes be quite >3

difficult in infants and children; in them, some hemangiomas and vascular malformations 41

have a similar appearance. Many of these situations can be clarified on imaging studies, Q

while others may need to be followed over a period of time. Hemangiomas will typically |

show the classic pattern of proliferation followed by involution, while vascular malfor- @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

594 ROSEN and MALDONADO

mations tend to grow at the same rate as the child. Unless there are severe functional
disturbances related to the lesion (impaired feeding, breathing, or vision), a rush to treat-
ment is rarely necessary. While the vast majority of hemangiomas will resolve without spe-
cific treatment, some authorities now advocate more aggressive early intervention for
lesions that are cosmetically disfiguring in order to avoid adverse effects on psychosocial
development (45). Timing may also be critical in some extremity vascular malformations
that affect limb length, either accelerating or retarding normal bone growth. This problem
can be encountered both in venous lesions (Klippel-Trenaunay type) and arterial lesions,
specifically high-flow lesions involving the epiphyseal plate. The limb-length discrepancy is
usually treated by epiphyseodesis in the venous type, and superselective embolization in
the arterial type. Obviously, a careful orthopedic assessment of expected bone growth is
critical in the timing of intervention to minimize the ultimate limb-length discrepancy.

VI. SUMMARY

Congenital vascular anomalies constitute a complex and disparate group of conditions.
Some of these are benign incidental findings that require no treatment, some will resolve
spontaneously, and some are life-threatening conditions requiring urgent intervention.
Caring for these patients requires familiarity with this range of conditions and the ability to
make an accurate diagnosis. Probably the most common complications are related to initial
misdiagnosis followed by misguided attempts at therapy. A multidisciplinary approach is
strongly recommended to achieve optimum clinical results. Percutaneous management of
these lesions has assumed a primary role in the care of these patients, either as stand-alone
therapy or in conjunction with surgery. The two primary approaches are transcatheter
embolization, generally employed in the treatment of high flow arteriovenous malforma-
tions, and direct embolization, where a sclerosing agent is directly injected into the lesion.
Complications of percutaneous procedures are uncommon but include those related to
faulty technique, inherent problems with many of the embolic devices and agents, and the
complex physiology of vascular malformations.

REFERENCES

1. Mulliken JB, Zetter BR, Folkman J. In vitro characteristics of endothelium from hemangiomas
and vascular malformations. Surgery 1982; 92(2):348-353.

2. Mulliken JB, Glowacki J. Hemangiomas and vascular malformations in infants and children: A
classification based on endothelial characteristics. Plast Reconstr Surg 1982; 69(3):412^122.

3. Mulliken JB, Glowacki J. Classification of pediatric vascular lesions. Plast Reconstr Surg 1982;
70(1):120-121.

4. Linderkamp O, Hopner F, Klose H, et al. Solitary hepatic hemangioma in a newborn infant
complicated by cardiac failure, consumption coagulopathy, microangiopathic hemolytic ane- j>
mia, and obstructive jaundice. Case report and review of the literature. Eur J Pediatr 1976; <S
124(l):23-29. ,|

5. Berdon WE, Baker DH. Giant hepatic hemangioma with cardiac failure in the newborn infant. 2
Value of high-dosage intravenous urography and umbilical angiography. Radiology 1969; **
92(7): 1523-1 528. ■§

6. Griffin JM, Vasconez LO, Schatten WE. Congenital arteriovenous malformations of the upper jjj
extremity. Plast Reconstr Surg 1978; 62(l):49-58. q

7. Olcott Ct, Newton TH, Stoney RJ, Ehrenfeld WK. Intra-arterial embolization in the man- |
agement of arteriovenous malformations. Surgery 1976; 79(1):3-12. 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

TREATMENT OF ARTERIOVENOUS MALFORMATIONS 595

8. Trout HH, III. Management of patients with hemangiomas and arteriovenous malformations.
Surg Clin North Am 1986; 66(2):333-338.

9. Padwa BL, Hayward PG, Ferraro NF, Mulliken JB. Cervicofacial lymphatic malformation:
Clinical course, surgical intervention, and pathogenesis of skeletal hypertrophy. Plast Reconstr
Surg 1995; 95(6):951-960.

10. Raveh E, de Jong AL, Taylor GP, Forte V. Prognostic factors in the treatment of lymphatic
malformations. Arch Otolaryngol Head Neck Surg 1997; 123(10): 1061-1065.

11. Perry MO. Iatrogenic injuries of arteries in infants. Surg Gynecol Obstet 1983; 1 57(5):41 5—418.

12. Szilagyi DE, Smith RF, Elliott JP, Hageman JH. Congenital arteriovenous anomalies of the
limbs. Arch Surg 1976; 111(4):423^129.

13. Carapiet DA, Stevens JE. Pulmonary embolism following embolization of an arteriovenous
malformation. Paediatr Anaesth 1996; 6(6):491-494.

14. Coard K, Silver MD, Perkins G, et al. Isobutyl-2-cyanoacrylate pulmonary emboli associated
with occlusive embolotherapy of cerebral arteriovenous malformations. Histopathology 1984;
8(6):9 17-926.

15. Kjellin IB, Boechat MI, Vinuela F, et al. Pulmonary emboli following therapeutic embolization
of cerebral arteriovenous malformations in children. Pediatr Radiol 2000; 30(4):279-283.

16. Purdy PD, Batjer HH, Risser RC, Samson D. Arteriovenous malformations of the brain:
choosing embolic materials to enhance safety and ease of excision. J Neurosurg 1992; 77(2): 212-
217.

17. Barth KH, Strandberg JD, White RI Jr. Long term follow-up of transcatheter embolization
with autologous clot, oxycel and gelfoam in domestic swine. Invest Radiol 1977; 12(3):273-280.

18. Nakano H, Igawa M. Complication after embolization of internal iliac artery by gelatin sponge
powder. Hiroshima J Med Sci 1986; 35(l):21-25.

19. Standard SC, Guterman LR, Chavis TD, Hopkins LN. Delayed recanalization of a cerebral
arteriovenous malformation following angiographic obliteration with polyvinyl alcohol em-
bolization. Surg Neurol 1995; 44(2): 109-1 12; discussion 112-113.

20. Sorimachi T, Koike T, Takeuchi S, et al. Embolization of cerebral arteriovenous malfor-
mations achieved with polyvinyl alcohol particles: Angiographic reappearance and complica-
tions. AJNR 1999; 20(7):1323-1328.

21. Yakes WF, Haas DK, Parker SH, et al. Symptomatic vascular malformations: Ethanol em-
bolotherapy. Radiology 1989; 170(3 Pt 2): 1059-1066.

22. Rosenkrantz H, Bookstein JJ, Rosen RJ, et al. Postembolic colonic infarction. Radiology 1982;
142(1):47-51.

23. DeSouza NM, Reidy JF. Embolization with detachable balloons — Applications outside the
head. Clin Radiol 1992; 46(3): 170-175.

24. Yakes WF, Luethke JM, Parker SH, et al. Ethanol embolization of vascular malformations.
Radiographics 1990; 10(5):787-796.

25. Yakes WF, Rossi P, Odink H. How I do it. Arteriovenous malformation management. Car-
diovasc Intervent Radiol 1996; 19(2):65-71.

26. Yakes WF, Luethke JM, Merland JJ, et al. Ethanol embolization of arteriovenous fistulas: A
primary mode of therapy. J Vase Intervent Radiol 1990; l(l):89-96.

27. Dickey KW, Pollak JS, Meier GH III, et al. Management of large high-flow arteriovenous j>
malformations of the shoulder and upper extremity with transcatheter embolotherapy. J Vase <S
Intervent Radiol 1995; 6(5):765-773. ,1

28. Sofocleous CT, Rosen RJ, Raskin K, et al. Congenital vascular malformations in the hand and £
forearm. J Endovasc Ther 2001; 8(5):484-494. ^

29. Jacobowitz GR, Rosen RJ, Rockman CB, et al. Transcatheter embolization of complex pelvic «
vascular malformations: Results and long-term follow-up. J Vase Surg 2001; 33(l):51-55. jjj

30. Pollak J, White RI. The use of cyanoacrylate adhesives in peripheral embolization. J Vase q
Intervent Radiol 2001; 12:907-913. |

31. Vinters HV, Lundie MJ, Kaufmann JC. Long-term pathological follow-up of cerebral arterio- 2

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

596 ROSEN and MALDONADO

venous malformations treated by embolization with bucrylate. N Engl J Med 1986; 314(8):
477-483.

32. Rao VR, Mandalam KR, Gupta AK, et al. Dissolution of isobutyl 2-cyanoacrylate on long-
term follow-up. AJNR 1989; 10(1): 1 35 141 .

33. Faughnan ME, Lui YW, Wirth JA, et al. Diffuse pulmonary arteriovenous malformations:
Characteristics and prognosis. Chest 2000; 1 17(1):3 1—38.

34. Brydon HL, Akinwunmi J, Selway R, Ul-Haq I. Brain abscesses associated with pulmonary
arteriovenous malformations. Br J Neurosurg 1999; 13(3):265-269.

35. Mitty HA, Efremidis S, Keller RJ. Colonic stricture after transcatheter embolization for
diverticular bleeding. AJR 1979; 133(3):519-521.

36. Hemingway AP, Allison DJ. Colonic embolisation: Useful but caution required. Gut 1998; 43(1):
4-5.

37. Schwartz RA, Teitelbaum GP, Katz MD, Pentecost MJ. Effectiveness of transcatheter embo-
lization in the control of hepatic vascular injuries. J Vase Intervent Radiol 1993; 4(3):359-365.

38. Trojanowski JQ, Harrist TJ, Athanasoulis CA, Greenfield AJ. Hepatic and splenic infarctions:
Complications of therapeutic transcatheter embolization. Am J Surg 1980; 139(2):272-277.

39. Raat H, Stockx L, De Meester X, et al. Percutaneous embolization of a splenic arteriovenous
fistula related to acute necrotizing pancreatitis. Eur Radiol 1999; 9(4):753.

40. Clouse ME, Levin DC, Desautels RE. Transcatheter embolotherapy for congenital renal
arteriovenous malformations. Long-term follow-up. Urology 1983; 22(4):360-365.

41. Nakamura H, Uchida H, Kuroda C, et al. Renal arteriovenous malformations: Transcatheter
embolization and follow-up. AJR 1981; 1 37(1): 1 13— 1 16.

42. Gomes AS, Busuttil RW, Baker JD, et al. Congenital arteriovenous malformations. The role of
transcatheter arterial embolization. Arch Surg 1983; 11 8(7): 8 17-825.

43. White RI Jr, Pollak J, Persing J, et al. Long-term outcome of embolotherapy and surgery for
high-flow extremity arteriovenous malformations. J Vase Intervent Radiol 2000; 1 1(10): 1285—
1295.

44. Noel AA, Gloviczki P, Cherry KJ Jr, et al. Surgical treatment of venous malformations in
Klippel-Trenaunay syndrome. J Vase Surg 2000; 32(5):840-847.

45. Demiri EC, Pelissier P, Genin-Etcheberry T, et al. Treatment of facial haemangiomas: The
present status of surgery. Br J Plast Surg 2001; 54(8):665-674.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

37

Endovascular Complications of Angioplasty and Stenting

Gary M. Ansel

Riverside Methodist Hospital, Columbus, Ohio, U.S.A.

As in every therapeutic procedure, endovascular procedures such as angioplasty and
stenting carry an inherent risk of complication to the patient. Though with the proper
training these complications can usually be successfully managed endovascularly, improper
management may lead to emergency surgery, limb loss, functional disability, and death. It is
paramount that physician operators have the proper training and ability to anticipate,
recognize, and treat complications as they arise during endovascular procedures.

In a recent review (1), the occurrence of a major morbidity such as myocardial infarc-
tion, renal failure, and stroke from elective endovascular procedures has been reported in up
to 2.3% of patients. An amputation rate of 0.6% was also seen, though only in patients with
preexisting critical limb ischemia.

The incidence of major complications at the site of intervention during balloon angio-
plasty procedures varies widely depending on the type of procedure and patient popula-
tion. Prior to the widespread utilization of stents, a large review by Gardiner et al. (2)
reported major complications at the angioplasty site in approximately 3% of procedures.
Though the addition of stent placement has historically shown an increased complication
rate, this most likely represented a learning curve phenomenon combined with the large
French size of the early stent equipment (3,4). More recent randomized data comparing ■g

balloon angioplasty to stenting in the femoral artery appears to show decreased compli- &

cations with stent utilization (5). Complex lesions such as occlusions, associated with a

longer procedure times and frequent equipment manipulations, may predispose to higher c

rates of complication than those seen in the treatment of simple stenotic lesions (6-9). <j

Limb-salvage patients have also shown a higher procedural complication rate than have >9

claudicants (10). When endovascular complications do occur, over 86% are usually 4j

evident in the angiographic suite and almost all are evident within 5 h postprocedure Q

(11,12). As in all procedures, it is proper training and anticipation of the potential for a |

complication that will be most likely to prevent it. @

597 I

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

598 ANSEL

I. PROCEDURE SITE COMPLICATIONS

Acute vascular complications at the endovascular procedure site include arterial perfo-
ration, dissection, thrombosis, spasm, side branch occlusion, and equipment failure.
Subacute complications include aneurysm formation, infection, and arteriovenous fistula
formation (13-16).

II. ARTERIAL PERFORATION

Arterial perforation can occur both at the site of balloon angioplasty and distally from the
guidewire. Perforation has been reported in 0-2.3% of patients (10,17,18). Guidewire
perforations are usually related to the use of hydrophilic wires and failure to visualize the
distal wire during equipment manipulation. Perforations at the angioplasty site appear to
be related to the presence of significant vessel calcification, tortuosity of the vessel, high
balloon inflation pressures, and utilization of a balloon that is larger than the native arte-
rial vessel. Lowering the risk of arterial perforation at the angioplasty site can be accom-
plished by undersizing the angioplasty balloon for the first arterial dilation or utilizing
intravascular ultrasound for assessment of arterial size and extent of calcification. Eval-
uating the patient's symptoms of discomfort during balloon inflation, which is related to
stretching of the vessel's adventitie, may also also decrease the risk of perforation during
balloon angioplasty and stenting. Discomfort should signal to the operator that they have
reached the maximum safe dilation size has been reached and should not be exceeded, as
otherwise arterial rupture may occur.

The treatment of distal guidewire perforations is usually conservative. However, in a
few vascular beds — such as the deep pelvic and renal areas — small vessel perforations may
lead to life-threatening hemorrhage. In these cases transcatheter coil embolization will
usually control the hemorrhage (Fig. 1) (19). Vascular perforations at the site of balloon
angioplasty may be controlled with prolonged balloon inflation while anticoagulation is
reversed. However, more significant perforations may require the placement of covered
vascular stents (Fig. 2) or surgical repair (16,20). The choice of treatment is often influ-
enced by the vessel size and location, extent of perforation, and patient's comorbidities.
Regardless of the perforation size, the initial treatment is to reinflate the angioplasty
balloon to control the hemorrhage. The operator should then evaluate the angiogram to
assess the extent of perforation, presence of complicating arterial side branches, and
procedural options. Small perforations are often successfully treated simply by inflating
the angioplasty balloon to the lowest pressure that will allow for sealing of the perforation.
Systemic anticoagulation should be reversed. After a balloon inflation time of approx-
imately 15 min, repeat angiography is completed. Once the perforation is sealed, the
patient can usually be safely observed until discharge. Large perforations or vascular ■g

ruptures appear to be extremely rare. However, when a rupture occurs in a large vessel, &

life-threatening hemorrhage can occur. For example, if a rupture should occur during a

angioplasty of an iliac artery, successful treatment requires control of the hemorrhage by c

balloon reinflation. Placement of a second balloon into the aorta or proximal iliac artery <j

from another vascular access site is often required. This second balloon is placed proximal >3

to the rupture site to control the hemorrhage that may occur during definitive repair at the 4j

original site of intervention. After control of the hemorrhage is obtained, the use of a 2

covered stent is often successful in sealing the perforation. The covered stent should be |

long enough to fully cover the arterial perforation. Sealing of the perforation with a stent- @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

B

5

f
J

w

*****

f

1

■

i

2

I

■c

Figure 1 A. Angiogram demonstrating contrast extravasation from a guidewire perforation of a
renal artery branch vessel. B. Repeat angiogram after thrombotic coil placement.

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

u

600

ANSEL

B

i

2

I

■c

Figure 2 A. Right iliac artery perforation with contrast extravasation. B. Repeat angiogram after
placement of two covered stents.

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

u

ANGIOPLASTY AND STENTING

601

Figure 3 A. Extensive femoral artery guidewire perforation. B. Selective femoral artery angiogram
after placement of a Nitinol stent.

graft may not occur if severe eccentric calcification is present or if a significant branch
vessel such as the hypogastric artery arises from within the area of perforation. Large side
branch vessels located at the site of perforation may lead to delayed bleeding as collaterals
dilate. After stent-graft placement, delayed thrombosis of the stent-graft is a concern if it is
placed in smaller vessels, as in the femoropopliteal regions (21,22). Very large perforations
or those that cannot be successfully treated by either prolonged balloon inflation or a

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

602

ANSEL

B

Figure 3 Continued.

covered stent will require surgical repair. Continued control of the hemorrhage should be
maintained by ongoing balloon inflation. The balloon is deflated only after the affected
artery has been surgically exposed in the operating suite. Various balloon expandable and
self-expanding covered stents may be utilized. Currently however, a covered stent
approved for arterial perforation by the U.S. Food and Drug Administration (FDA) is
available only for the coronary vasculature Covered stents used in the peripheral
vasculature are utilized "off label."

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ANGIOPLASTY AND STENTING

603

Figure 4 A. Carotid stent with intraprocedural thrombus formation. B. Repeat angiogram after
treatment with an intravenous lib/Ilia antiplatelet agent.

III. ARTERIAL DISSECTION

Endovascular stents have diminished the clinical impact of even large intimal dissections
(Fig. 3). Since stents can effectively treat a dissection, most clinically relevant dissections
occur either during guide wire placement, or from unrecognized and untreated dissection
after wire removal. Wire dissections that cannot be crossed from the original acess site
can often be crossed and treated from the opposite end of the dissection. Unrecognized

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

604

ANSEL

B

Figure 4 Continued.

wire dissections that go from true lumen into false lumen and back into true lumen after
an important branch vessel are also important causes of complication from dissection.
This true-false-true phenomenon most frequently occurs during the treatment of total
occlusions.

IV. ARTERIAL THROMBOSIS

Arterial thrombosis during angioplasty or stenting is usually due to inadequate anti-
coagulation or antiplatelet therapy, inadequate balloon angioplasty result, dissection, or
unrecognized hypercoagulable condition. Thrombosis may also occur due to blood stasis if
the endovascular device is large and obstructs arterial flow in the setting of inadequate
anticoagulation. An activated clotting time (ACT) should be obtained after heparin

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ANGIOPLASTY AND STENTING 605

administration and during long cases every 60 min. The size of the endovascular equipment
relevant to the arterial vessel and the restriction of arterial flow as well as the length of the
procedure determine what constitutes an adequate ACT.

The prevention of subacute thrombosis has been studied best in the coronary
intervention literature. Previous aggressive anticoagulation regimens utilized in early
coronary artery stenting usually consisted of aspirin, dypyridamole, dextran, heparin,
and warfarin. These regimens were associated with considerable bleeding complications,
especially at the sheath insertion sites, at rates of 3-6% (23). However, subacute
thrombosis rates of over 7% were still seen (24). In the largest study to date, 1652 cardiac
patients were randomized to aspirin plus warfarin, aspirin plus ticlopidine, or asprin alone
after coronary stenting. The aspirin-plus-ticlopidine group had a significantly lower
subacute thrombosis rate (0.6%) than either of the other groups (both 2.4%) (25). It
would appear prudent to recommend the use of aspirin plus ticlopidine or clopidogrel for
peripheral angioplasty/stenting procedures to decrease the risk of subacute thrombosis.
The formation of visible thrombus (Fig. 4), usually platelet-mediated, during an endo-
vascular procedure should prompt the consideration of an intravenous Ilb/IIIa antipla-
telet agent such abciximab (26).

V. ARTERIAL SPASM

Arterial spasm rarely leads to a serious complication during angioplasty. However, severe
symptoms of pain may occur. The occurrence of arterial vessel spasm appears to be more
common in younger patients (18). It has also been our experience that small vessels or
vessels supplying an organ are more prone to spasm than the large vessels supplying the
limbs. Though prolonged spasm could lead to thrombosis, the major risk of spasm is often
the difficulty with differentiating it from dissection (Fig. 5). Spasm should be prevented by
routinely administering a vasodilator agent (nitroglycerine, papaverine, verapamil) prior

Figure 5 A. Left renal artery angiogram. B. Severe arterial spasm after balloon angioplasty.
C. Final angiogram after stent placement and intra-arterial nitroglycerin.

i
2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

606

ANSEL

B

Figure 5 Continued.

to vessel manipulation. Occasionally, a guidewire may have to be removed to distinguish
refractory spasm from dissection.

i

2

I

■c

VI. EQUIPMENT FAILURE

Another source of procedural complication is related to device failure. This includes angio-
plasty balloon rupture prior to stent expansion, stent embolization, guidewire fracture, and
catheter fracture.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ANGIOPLASTY AND STENTING

607

Figure 5 Continued.

Angioplasty balloon rupture prior to stent expansion has decreased as technology
has improved. Most balloon rupture currently seen is due to the utilization of a balloon
that is longer than the stent (Fig. 6). During early balloon inflation on the ends, the stent
edges may lead to balloon puncture. Another source of balloon rupture is arterial calci-
fication. A small number of balloons are defective from the time of manufacturing; these
should be noted when negative pressure is applied during balloon preparation. Negative
pressure should always be repeated after hand-mounting stents on to a previously inflated
balloon.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

608

ANSEL

Figure 6 Fluoroscopy demonstrating balloon rupture before stent expansion. Note the excess
balloon material beyond the ends of the stent.

When a balloon rupture occurs prior to stent expansion, several strategies may allow
for successful completion of stent deployment. Rapid inflation of the balloon with hand
pressure will often expand the stent enough to allow for balloon removal and replacement
with a new balloon. However, occasionally hand inflation is not adequate and another
method of balloon inflation, by connecting the balloon port to the contrast power injector
filled with saline, will be necessary for adequate stent expansion. An injection larger than

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ANGIOPLASTY AND STENTING

609

the balloon capacity at a flow rate of approximately 5 mL/s is usually successful. Other
strategies include the following: if a 0.035 wire based system has been used, replacing the
0.035 wire with a 0.014/0.018 wire to utilized a coronary balloon for initial stent expansion,
bringing the sheath or guide catheter to the end of the stent for support while te balloon is
removed, or advancing the balloon out of the stent and advancing a coronary wire and
balloon beside the original balloon shaft and enlarging the stent to allow for balloon
removal.

VII. DEVICE EMBOLIZATION

There are multiple ways for whole devices such as stents or parts of devices such as wires
and catheters, to embolize and complicate an angioplasty or stent procedure. Balloon
material, wires, or ends of catheters can become embedded on the ends of the stent and be
torn free (Fig. 7). Rarely is surgical exploration and removal the proper method of dealing

i

2

I

■c

Figure 7 A. Fluoroscopy demonstrating the tip of an angiographic catheter that has been sheared
off at the end of a renal artery stent. B. Fluoroscopy after placement of a 0.014 guidewire and
coronary balloon to allow for removal of the catheter tip to the abdominal aorta. C. Fluoroscopy
showing the catheter tip secured against the wall of the iliac artery by a balloon expandable stent.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

B

i

2

I

■c

Figure 7 Continued.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ANGIOPLASTY AND STENTING

611

with these events. The basis of treatment is controlling the embolized material for removal
or trapping the material within another stent to allow for endothelialization. There are a
number of devices and maneuvers for capturing loose stents or fractured catheters or wires
(27). Vascular snare (Fig. 8) and grasping forceps are the most commonly used tools for
capturing debris. The decision must then be made as to whether the debris can be removed
without snagging on the arterial wall or the extravascular tissue. If the embolized material
is flexible and smooth, such as a wire or catheter, it can usually be removed. However,
stents that have become embolized should first be captured and then expanded in a more
proximal vessel such as the iliac artery or trapped against the side of the aorta with a large
self-expanding stent (Fig. 9).

i

2

I

■c

Figure 8 Vascular snare being utilized to secure a guidewire.

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

612

ANSEL

S

Figure 9 Schematic of a self-expanding stent being utilized to secure a embolized, nonexpanded
stent against the arterial wall.

VIM. CONCLUSION

Though uncommon, complications of endovascular angioplasty and stenting may have
devastating outcomes that can threaten limb and life. Adequate awareness of the com-
plications inherent to specific vascular beds, as well as the ability and training to promptly
recognize and treat complications as they arise, will allow the endovascular specialist to
achieve excellent clinical outcomes.

ACKNOWLEDGMENT

I would like to thank Dr. Mark H. Wholey, UPMC Shadyside, Pittsburgh, PA, and Dr.
Mark Burket, Medical College of Toledo, Toledo, OH, for graciously allowing the use of
case examples.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ANGIOPLASTY AND STENTING 613

REFERENCES

1. Axisa B, Fishwick G, Bolia A, et al. Complications following peripheral angioplasty. Ann R
Coll Surg Engl 2002; 84:39-42.

2. Gardiner GA Jr, Meyerovitz MF, Stokes KR, et al. Complications of transluminal angioplasty.
Radiology 1986; 159:201-208.

3. Henry M, Amor M, Ethevenor G, Henry I, Amicabile C, Beron R, et al. Palmaz stent
placement in iliac and femoropopliteal arteries: Primary and secondary patency in 310 patients
with 2-4 year follow-up. Radiology 1995; 197:167-174.

4. Gray BH, Olin JW. Limitations of percutaneous transluminal angioplasty with stenting for
femoropopliteal arterial occlusive disease. Semin Vase Surg 1997; 10:8-16.

5. Food and Drug Administration IntraCoil data: Cardiovascular and Radiological Health Ad-
visory Board.

6. Tegtmeyer CJ, Hartwell GD, Selby JB, et al. Results and complications of angioplasty in
aortoiliac disease. Circulation 1991; 83(suppl I):I-53-I-56.

7. Jorgensen B, Skovgaard N, Norgard J, et al. Percutaneous transluminal angioplasty in 226 iliac
artery stenoses: Role of the superficial femoral artery for clinical success. Vasa 1992; 21:382-
386.

8. Blum U, Gabelmann A, Redecker M, et al. Percutaneous recanalization of iliac artery
occlusions: results of a prospective study. Radiology 1993; 189:536-540.

9. Gupta AK, Ravimandalam K, Rao VR, et al. Percutaneous balloon angioplasty for arte-
riosclerosis obliterans: Long term results. In: Yao JST, Perarce WH, eds. Techniques in Vascular
Surgery. Philadelphia: Saunders, 1992:329-345.

10. Matsi PJ, Manninen HI. Complications of lower-limb percutaneous transluminal angioplasty:
A prospective analysis of 410 procedures on 295 consecutive patients. Cardiovasc Intervent
Radiol 1998; 21:361-366.

11. Burns BJ, Phillips AJ, Gox A, et al. The timing and frequency of complications after peripheral
percutaneous transluminal angioplasty and iliac stenting: Is a change from inpatient to
outpatient therapy feasible? Cardiovasc Intervent Radiol 2000; 23:452-456.

12. Kruse JR, Cragg AH. Safety of short stay observation after peripheral vascular intervention. J
Vase Intervent Radiol 2000; 11:45-49.

13. Bortslap A, Lampmann L. Balloon rupture and arteriovenous communication: A rare com-
plication of transluminal angioplasty. Vasa 1993; 22:352-354.

14. Hunter D, Simmons R, Hulbert J. Antibiotics for radiology interventional procedures. Radi-
ology 1988; 166:572-573.

15. Paddon AJ, Nicholson AA, Eltles DF, et al. Long-term follow-up of percutaneous balloon
angioplasty in adult aortic coarctation. Cardiovasc Intervent Radiol 2000; 23:364-367.

16. Scheinert D, Ludwig J, Steinkamp HJ, et al. Treatment of cath induced iliac artery injuries with
self-expanding endografts. J Endovasc Ther 2000; 7:213-220.

17. Lederman RJ, Mendelsohn FO, Santos R, et al. Primary renal artery stenting: Characteristics
and outcomes after 363 procedures. Am Heart J 2001; 142:314-323.

18. Morris CS, Bonnevie GJ, Najarian KE. Nonsurgical treatment of acute iatrogenic renal artery ^
injuries occurring after renal artery angioplasty and stenting. Am J Roentgenol 2001; 177:1353- E
1357. 1

19. Oltaenu B, Oltaenu C, Borelli C. Embolization of a perforation of a cortical renal artery H,
occurring during percutatneous renal angioplasty. Eur Radiol 2000; 10:1357.

20. Ragg JL, Biamino G. Perforations in recanalization of arterial occlusions of the femoro-
popliteal area. Zentralbl Chir 2000; 125:34-41.

21. Beregi JP, Prat A, Willoteaux S, et al. Covered stents in the treatment of peripheral arterial
aneurysms: Procedural results and mid term follow-up. Cardiovasc Intervent Radiol 1999;
22:13-19.

I

u

■c

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

614 ANSEL

22. Bauermeisto G. Endovascular stent-grafting in the treatment of superficial femoral artery
occlusive disease. J Endovasc Ther 2001; 8:315-320.

23. Schatz RA, Bairn DS, Leon M, et al. Clinical experience with the Palmaz-Schatz coronary
stent: Initial results of a multicenter study. Circulation 1991; 83:148-161.

24. Karrillon GJ, Morice MC, Benveniste E, et al. Intracoronary stent implantation without ultra-
sound guidance and with replacement of conventional anticoagulation by antiplatelet therapy:
30-day clinical outcome of the French Multicenter Registry. Circulation 1996; 94:1519-1527.

25. Zidar JP. Rationale for low-molecular weight heparin in coronary stenting. Am Heart J 1997;
134(suppl):S81-S87.

26. Tong FC, Cloft HJ, Joseph GS, et al. Abciximab rescue in acute carotid stent thrombosis. Am J
of Neuroradiol 2000; 21:1750-1752.

27. Bartorelli AL, Fobbiocchif, Montarsi F, et al. Successful transcatheter management of Palmaz
stent embolization after superior vena cava stenting. Cathet Cardiovasc Diagn 1995; 34:162-
166.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

38

Complications of Carotid Stenting

Ramtin Agah

University of Utah, Salt Lake City, Utah, U.S.A.

Patricia Gum and Jay S. Yadav

The Cleveland Clinic Foundation, Cleveland, Ohio, U.S.A.

I. INTRODUCTION

Stenting of the carotid artery is rapidly making the transition from an investigational tool
to the primary treatment modality for symptomatic and asymptomatic carotid lesions. A
key impetus behind this drive has been the significant reduction in the periprocedural
complication rate associated with this technique. The most significant sequela of carotid
stenting is the development of neurological deficit, described as small reversible deficits to
frank large strokes. With more experience, the development of dedicated angioplasty
equipment for the carotid bed and technological advances — including the development
and utilization of embolic protection devices — the periprocedural complication rate
(death, major and minor stroke) of 5-9% reported in the early series has been reduced
to 2-3% currently (1-15). This significant reduction of neurological events associated with
carotid artery stenting (CAS) is best understood by an in-depth review of the nature and
causes of these complications.

II. PROCEDURAL COMPLICATIONS ™

A. Access-Related I

I
1 . Access .g 1

Early in the carotid stenting experience, the direct antegrade carotid approach was used in <

some patients to establish access (4,16). Even though this approach simplified the ability to >9

access and intervene in the lesion, it had significant limitations, most significantly compres- ^

sion hematoma and a nidus for thrombus formation close to the stent. With the evolution of «

guide catheter technology specifically for percutaneous carotid intervention, the direct |

carotid access site has by and large been abandoned for the retrograde femoral approach. @

615 I

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

616

AGAH et al.

Furthermore, the refinement in equipment has caused the access-size requirement to
decrease over the past decade, with standard 9F sheath giving way to a 6F sheath or 8F
guides for carotid stenting using the femoral approach. The initial experience with these
larger sheaths was tainted with more frequent vascular complications, including arterio-
venous fistulas, large hematomas, and retroperitoneal bleeding; but these have decreased
in frequency and severity with the adoption of the smaller sheaths (4,5).

2. Guide Catheters

One of the major influences in the technical success of carotid stenting has been the evolution
of the guide catheter. Indeed, the most common reason for "technical failure" early on was
the inability to deliver a guide across a tortuous aortic arch. These technical limitations
resulted in both lower success rates (5-15% failure rate) and also more complications due to
plaque embolization and or dissection in the common carotid in the process of delivering the
guide (1,4,15,17). The development of new guide catheters has minimized these issues, with
recent series reporting a 98-100% technical success rate with minimal to no procedural
complications related to guide delivery by the experienced operator (3,8,9, 1 8) (Figs. 1 and 2).

B. Lesion-Related Complications

1. Cerebral Ischemia

Most patients tolerate transient occlusion of the blood flow to the respective hemisphere
very well. This is especially true with present protocols of short balloon inflation during
percutaneous transluminal angioplasty (PTA) and stenting. However, reports of episodes
of loss of consciousness, seizure, and or transient ischemic attack (TIA) during balloon
inflation still range from 1.5 to 6.2% (1,5,6). These events appear to be more common in
patients with an incomplete circle of Willis and/or multiple stenotic and/or occlusive
lesions in the contralateral supra-aortic territory (2) (Fig. 3). No long-term sequelae from
these transient episodes of ischemia have been reported.

Figure 1 Guide catheters for carotid intervention.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF CAROTID STENTING

617

Stiff Angled Glide Wire

Cordis H1 Guide: 8 FR

JR4:5FR, 125cmorVITEK

Cook 6 FR Sheath
Figure 2 Telescoping diagnostic catheters through guide catheters.

2. Bradycardia and Hypotension

This is a relatively common phenomenon, reported in as many as half of all the patients
undergoing CAS (5,19,20). The mechanism of this phenomenon is stimulation of the vagal
nerve via the glossopharyngeal afferent fibers emanating from the carotid baroreceptors.
Patients with previous carotid endarterectomy (CEA), who have commonly lost these
neuroreceptors, are less susceptible to this reflex during carotid stenting for restenotic
lesions. Furthermore, it has been suggested that this reflex is more common with inflation

I
•5

Right Internal Carotid Artery

Left Internal Carotid Artery

Figure 3 Severe bilateral carotid stenosis. Treatment of right internal carotid artery resulted in
transient seizure and neurological deficit in patient.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

618

AGAH et al.

of the balloon at the carotid bifurcation or the ostium of the internal carotid and
utilization of larger balloons (20,21) (Fig. 4).

Early on in the CAS experience, transvenous pacemakers were routinely used to ob-
viate these episodes of bradycardia (1). In the more recent series, atropine has been used
primarily to treat these patients. Most of these episodes are self-limited and as such require
no further treatment beyond atropine. However, in 10-19% of the patients, there can
episodes of prolonged hypotension requiring pressor therapy beyond the initial 12 h (19,
20). In only one reported case has there been a need for placement of a permanent pace-
maker after CAS in a patient with preexisting sick sinus syndrome (5). Again, even though
most of these episodes are self-limited and relatively transient, the operator must be vi-
gilant for such episodes and ready to treat them in patients with multiple supra-aortic
lesions, who may be more susceptible to the consequences of cerebral ischemia with tran-
sient hypotension.

3. Acute Vessel Closure

These episodes, not unlike interventions in the coronary bed, are secondary to spasms,
thrombosis, and dissections. Even though infrequent, the consequence of acute vessel
closure in the carotid vascular bed can be catastrophic due to the relatively short period of
ischemia that brain parenchyma can tolerate prior to irreversible injury. With routine
stenting of the carotid lesion, better conformity of stent design to carotid architecture, and
more optimal antiplatelet and antithrombotic regimen, the frequency of acute vessel clo-
sure has been significantly reduced, from 5 to 7% in the initial experience to below 1% in
the current series (2,4,5,18).

4. Embolization

With the development of self-expanding stent technology for carotid applications, guide
catheters, and optimization of the antiplatelet regimen, the number of neurological com-

I
•5

Msm PenatKm

Figure 4 Rhythm and pressure recordings during postdilatation of a carotid artery stent.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF CAROTID STENTING

619

plications due to carotid artery dissection, vasospasm, and stent thrombosis has declined;
in the current era, distal embolization remains the major cause of neurological sequelae
with CAS (Fig. 5).

It is estimated that routine PTA and stenting is associated with a 5% rate of embo-
lization in most vascular beds. However, even though this rate of embolization may be
acceptable in other vascular beds, the clinical sequelae of these events in the carotid
vasculature is much more troublesome due to the nature of the target organ. Vascular
surgeons have long recognized that patients with atherosclerosis often harbor a friable and
unstable plaque at the carotid bifurcation, poised to shower the brain distally with micro-
emboli (22,23). Several studies using transcranial Doppler have reported episodes of
embolization during revascularization (22-26). In a study of 301 patients undergoing
CEA, Ackerstaff et al. demonstrated that although microembolization is noticed in the
majority (69%) of the patients, the overall rate of clinical sequelae in this cohort was only
5.7% (23). A smaller study of 39 patients undergoing CAS has also revealed a similar
incidence of embolization to be operative with CAS, especially during balloon inflation and
stent deployment (25).

At best, these noninvasive measures for microembolization appear sensitive but not
specific for predicting clinical neurological events; as in all these studies, the incidence of
clinical neurological sequelae is much lower than the rate of embolization (23,25). As such,
the suggestion has been made that microembolization may have a threshold effect. Hence,
even though the majority of patients may have embolization with a routine carotid
procedure, it is only when these events reach a certain threshold that they may cause clinical
sequelae. In favor of this hypothesis are data showing that there is a direct correlation be-
tween the frequency of embolization and the size of emboli and the rate of clinical neuro-
logical events (27).

Regardless of its exact pathophysiological mechanism, there has been a concerted ef-
fort to prevent and reduce the risk of embolization with carotid stenting, Initial phar-
macological attempts included use of glycoprotein IIB-IIIa (GPIIB-IIIa) antagonist.
However, this approach has had mixed results (28,29). In cases where intraprocedural
embolization is associated with signs of a neurological event, the efficacy of intra-arterial

Middle

Cerebral

Artery

Velocity

crn/sec

Inflated Deflated
Balloon Balloon

Figure 5 Transcranial Doppler recording of the middle cerebral artery during carotid angioplasty.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

620

AGAH et al.

Table 1 The Results of Series Before and After Utilization of Emboli Protection Devices in CAS

Without

jrotection

With protection

Study

Stroke

Stroke/Death

Stroke

Stroke/Death

Al-Mubarak

(18)

1% (n = 162)

2%

Reimers (8)

1.2% (n = 84)

1.2%

Tybler (27)

2.0% (n = 54)

2%

Henry (7)

2.2% (n = 184)

2.7%

Vitek (10)

6.0% (n = 350)

7.0% +

Wholey (6)

4.2% (n = 4757)

5.1%

Diethrich (4)

6.4% (n = 110)

7.4%

fibrinolytic agents has not been demonstrated. In small case series there is no evidence of
clinical benefit with such an approach, likely due to the plaque-like rather thrombotic
nature of the emboli (30). In turn, mechanical means to prevent and "capture" these
embolic particles have come of age. The number of these prevention devices is rapidly
expanding; some that are currently in clinical use include the PercuSurge balloon wire,
filter wire EX, the Angiogaurd basket device, and Neuro shield filters (7,8,18).

The results of the few published series using a variety of these distal protection devices
have consistently reported a stroke rate (minor and major) of 1.2-2.0%, which is less than
half the event rates commonly reported prior to the utilization of these devices (4.2-60%),
as shown in Table 1 (7,8,18). However, the ability of these devices to reduce neurological
sequelae of carotid stenting can only be conclusively determined when the results of
current trials become available. Furthermore, the superiority of one device over another is
hard to assess at the present time, but devices that do not totally occlude cerebral blood
flow during deployment seem to be associated with higher procedural success rate, as
patients with bilateral disease and an incomplete circle of Willis do not tolerate transient
occlusion of the carotid artery (Fig. 6).

Figure 6 Two examples of an Angioguard filter with a large amount of embolic debris
resulted in slow flow through the filter during a carotid artery stenting procedure.

that

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF CAROTID STENTING 621

In addition, these prevention devices may not completely eliminate embolic events,
since the embolization via external carotid to internal carotids and collaterals is possible in
patients with severe stenosis, as suggested in at least one case report (31).

III. POSTPROCEDURAL

A. Stent Thrombosis

The incidence of stent thrombosis with carotid stenting has been declining since the early
published series (5,15,26,32). The technical changes leading to this improvement can be
summarized as follows.

1 . Stent Apposition and Type

Early in the experience, most stents used in the carotid setting were hand-mounted bal-
loon-expandable Palmaz stents. Although these stents offered radial strength, they were
more susceptible to incomplete deployment and malposition, especially in the presence of
calcification, tissue rigidity, and significant variations in diameter between the internal and
common carotid arteries. These mechanical issues produced a substrate with higher
subacute stent thrombosis rates (33). With the introduction of self-expanding nitinol stent
technology, the ability to adequately appose stents in most if not all cases has significantly
improved.

2. Anticoagulation Regimen

Early experience with carotid stenting preceded the current regimen of clopidogrel or tic-
lodopine to prevent subacute stent thrombosis. As such, the early experience with carotid
intervention, like that with coronary intervention, suffered from episodes of subacute stent
thrombosis due to lack of an optimized anticoagulation regimen. The present regimen of
clopidogrel or ticlodopine dosing followed by a maintenance dose for a minimum of 1
month has significantly reduced such events.

B. Stent Deformation or Collapse

The Palmaz balloon-expandable stent had an incidence of stent deformation or collapse
ranging from 8 to 16% (5,33-35). Equally alarming was the fact that most patients were
asymptomatic and that these findings could not be discerned by noninvasive means such as
ultrasound (33). This mechanical failure of the stent was attributed to external compres-
sion and plastic deformation of the stainless steel structure. The transition to self-
expanding stents has overcome this problem, and balloon-expandable stents are no longer
used in the carotid bifurcation (Fig. 7).

C. Neck Pain

Up to 4% of patients have developed neck pain postprocedure due to stretching of the

carotid adventitia; furthermore there have also been reports of pain in the posterior aspect %

of the scalp, likely secondary to irritation of the greater auricular nerve (1). Most these ^

episodes are self-limited and require only oral analgesics until the symptom subsides. S

I

D. Hyperperfusion Cerebral Hemorrhage a

Cerebral hyperperfusion syndrome is a recognized complication of CEA (36,37). There §

have been several case and small series reports of this complication also occurring with ©

t

a

s

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

622 AG AH et al.

Figure 7 A crushed balloon-expandable Palmaz stent in the internal carotid artery and evidence of
luminal compromise secondary to the stent deformity.

carotid stenting (38-40). By alleviating a high-grade symptomatic stenotic lesion, hyper-
perfusion may occur as a result of a sudden, rapid increase in cerebral blood flow. This
transient cerebral hyperemia can lead to headache, vomiting, arterial hypertension,
confusion, seizures, focal neurological deficits, and subarachnoid hemorrhage. The exact
predisposition is not known, but some have suggested risk factors including lesions with a
significant pressure gradient across the stenosis with poor distal flow to the ipsilateral
hemisphere, poor collateral blood flow, contralateral carotid occlusion, perioperative
hypertension, and the use of anticoagulant and antiplatelet agents. The worst complica-
tion associated with this syndrome is cerebral hemorrhage, occurring in less than 0.5% of
the cases, based on a small series (40).

IV. LATE EVENTS

Remarkably there has been a minimal incidence of late events associated with carotid
stenting. Early on with the predominant use of Palmaz stents, late events (TIA, death,
stroke) beyond the initial 30 days were reported in 2.5% of the patients, possibly as a
direct consequence of the frequent crush injuries with this stent (4,32). In more recent
series with self-expanding stents, this complication has been reduced to 1.0% (6).

The rate of restenosis ranges from 2 to 5%. This rate is significantly lower than that in
other vascular beds, likely due to the more focal nature of the lesion (bifurcation) and the ■a

large size of the arterial lumen (3-6). Most cases of restenosis are amenable to repeat |

angioplasty with good results, and there have even been case reports of gamma radiation a

therapy for refractory in-stent restenosis in this setting (41). c

V. CONCLUSIONS f

The dramatic rapid progress in CAS in the last decade has transformed this investigational «

tool into a viable treatment strategy. This transformation has been driven by technical and |

technological improvements in CAS, which, in turn, have resulted in the reduction of @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

COMPLICATIONS OF CAROTID STENTING 623

complications associated with this procedure. Ultimately results of ongoing trials will
allow objective assessment of the safety and efficacy of this technique in everyday practice.

REFERENCES

1. Wholey MH, et al. Endovascular stents for carotid artery occlusive disease. J Endovasc Surg
1997; 4(4):326-338.

2. Theron JG, et al. Carotid artery stenosis: Treatment with protected balloon angioplasty and stent
placement. Radiology 1996; 201(3):627-636.

3. Bonaldi G. Angioplasty and stenting of the cervical carotid bifurcation: Report of a 4-year series.
Neuroradiology 2002; 44(2):164-174.

4. Diethrich EB, Ndiaye M, Reid DB. Stenting in the carotid artery: Initial experience in 110
patients. J Endovasc Surg 1996; 3(l):42-62.

5. Yadav JS, et al. Elective stenting of the extracranial carotid arteries. Circulation 1997; 95(2):376-
381.

6. Wholey MH, et al. Global experience in cervical carotid artery stent placement. Catheter Car-
diovasc Intervent 2000; 50(2):160-167.

7. Henry M, et al. Benefits of cerebral protection during carotid stenting with the PercuSurge
GuardWire system: Midterm results. J Endovasc Ther 2002; 9(1): 1-13.

8. Reimers B, et al. Cerebral protection with filter devices during carotid artery stenting. Circulation
2001; 104(1):12-15.

9. Cremonesi A, Castriota F. Efficacy of a nitinol filter device in the prevention of embolic events
during carotid interventions. J Endovasc Ther 2002; 9(2): 155-1 59.

10. Vitek J, Iyer S, Roubin G. Carotid stenting in 350 vessels: Problems faced and solved. J Invas
Cardiol 1998; 10(5):3 11-314.

11. Laird JS, Pompa J. Procedural results and early clinical outcomes after carotid stent-supported
angioplasty in high-risk patients. J Am Coll Cardiol 1997; 29(suppl):362A.

12. Iyer SR, Dorros G. Clinical significance of neurological events associated with carotid stenting.
J Am Col Cardiol 1997; 29(suppl):362A.

13. Henry MA, Henry I. Endovascular treatment of atherosclerotic stenosis of the internal carotid.
J Vase Intervent Radiol 1998; 9(suppl):162.

14. Henry M, Amor M, Henry I. Endovascular treatment of atherosclerotic stenosis of the internal
carotid. J Am Coll Cardiol 1997; 29(suppl):221a.

15. Iyer SSR, Yadav GS, Diethrisch EB. Angioplasty and stenting for extracranial carotid stenosis:
Multicenter experience. Circulation 1996; 94(8):I-58.

16. Bergeron P, et al. Percutaneous stenting of the internal carotid artery: The European CAST I
Study. Carotid Artery Stent Trial. J Endovasc Surg 1999; 6(2):155-159.

17. Dietz A, et al. Endovascular treatment of symptomatic carotid stenosis using stent placement:
Long-term follow-up of patients with a balanced surgical risk/benefit ratio. Stroke 2001; 32(8):
1855-1859.

18. Al-Mubarak N, et al. Multicenter evaluation of carotid artery stenting with a filter protection
system. J Am Coll Cardiol 2002; 39(5):841-846.

19. Gray WWH, Barret D. Hemodynamic consequences of carotid stenting. Circulation 1997; 96

20. Mendelsohn F, Weissman N, Crowley J. Hypotension associated with carotid stenting
(HAWCS). Circulation 1997; 96(suppl):I-307.

21. Sivaguru ASG, Venables J. Blood pressure changes after carotid endarterectomy and angio-
plasty. Br J Surg 1997; 84:562-578.

22. Gaunt ME, et al. Clinical relevance of intraoperative embolization detected by transcranial
Doppler ultrasonography during carotid endarterectomy: A prospective study of 100 patients. Br
J Surg 1994; 81(10): 1435-1439.

s

(suppl):I-284. 1

c

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

624 AGAH et al.

23. Ackerstaff RG, et al. The significance of microemboli detection by means of transcranial Doppler
ultrasonography monitoring in carotid endarterectomy. J Vase Surg 1995; 21(6):963-969.

24. Markus HS, et al. Carotid angioplasty. Detection of embolic signals during and after the
procedure. Stroke 1994; 25(12):2403-2406.

25. Al-Mubarak N, et al. Effect of the distal-balloon protection system on microembolization during
carotid stenting. Circulation 2001; 104(17): 1999-2002.

26. McCleary AJ, et al. Cerebral haemodynamics and embolization during carotid angioplasty in
high-risk patients. Br J Surg 1998; 85(6):77 1-774.

27. Tubler T, et al. Balloon-protected carotid artery stenting: relationship of periprocedural neuro-
logical complications with the size of particulate debris. Circulation 2001; 104(23):2791-2796.

28. Kapadia SR, et al. Initial experience of platelet glycoprotein Ilb/IIIa inhibition with abciximab
during carotid stenting: a safe and effective adjunctive therapy. Stroke 2001; 32(10):2328-2332.

29. Hofmann R, et al. Abciximab bolus injection does not reduce cerebral ischemic complications of
elective carotid artery stenting: a randomized study. Stroke 2002; 33(3):725-727.

30. Wholey MH, et al. Management of neurological complications of carotid artery stenting. J
Endovasc Ther 2001; 8(4):341-353.

31. Al-Mubarak N, et al. Embolization via collateral circulation during carotid stenting with the
distal balloon protection system. J Endovasc Ther 2001; 8(4):354-357.

32. Roubin GS, et al. Carotid stent-supported angioplasty: a neurovascular intervention to prevent
stroke. Am J Cardiol 1996; 78(3A):8-12.

33. Mathur A, et al. Palmaz stent compression in patients following carotid artery stenting. Cathet
Cardiovasc Diagn 1997; 41(2): 137-140.

34. Mathur AR, Dorros G, Iyer SJ. Palmaz stent collapse in patinets following carotid artery
stenting. J Am Coll Cardiol 1997; 29(suppl):363a.

35. Johnson SP, et al. Stent deformation and intimal hyperplasia complicating treatment of a post-
carotid endarterectomy intimal flap with a Palmaz stent. J Vase Surg 1997; 25(4):764-768.

36. Jorgensen LG, Schroeder TV. Transcranial Doppler for detection of cerebral ischaemia during
carotid endarterectomy. Eur J Vase Surg 1992; 6(2): 142-147.

37. Magee TR, Davies AH, Horrocks M. Transcranial Doppler evaluation of cerebral hyper-
perfusion syndrome after carotid endarterectomy. Eur J Vase Surg 1994; 8(1): 104-106.

38. McCabe DJ, Brown MM, Clifton A. Fatal cerebral reperfusion hemorrhage after carotid
stenting. Stroke 1999; 30(ll):2483-2486.

39. Chamorro A, et al. A case of cerebral hemorrhage early after carotid stenting. Stroke 2000;
31(3):792-793.

40. Meyers PM, et al. Cerebral hyperperfusion syndrome after percutaneous transluminal stenting of
the craniocervical arteries. Neurosurgery 2000; 47(2):335-343; discussion 343-345.

41. Chan AW, et al. Carotid brachytherapy for in-stent restenosis. Catheter Cardiovasc Intervent
2003; 58(l):86-92.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

39

Endovascular Access Complications

Mark A. Farber and Robert Mendes

University of North Carolina, Chapel Hill, North Carolina, U.S.A.

Since receiving approval by the U.S. Food and Drug Administration in 1999, endovascular
treatment of abdominal aortic aneurysms has been increasingly performed, especially in
high-risk patients. Despite this new approach, a significant number of patients are excluded
for various reasons, including unacceptable anatomy and challenging access issues (1-4).
Current delivery systems are designed with outer diameters in excess of 18 F (Table 1) for
introduction of the primary device, and despite the recognized need for smaller, more
flexible delivery catheters (5), they remain unavailable except in clinical trials and design
stages. Therefore the current stiff introducer systems may result in access complications that
need to be addressed during or after endovascular device implantation. Many times these
complications are directly related to the extent of atherosclerotic occlusive disease within the
iliac system. These complications fall into three categories: access failure, vessel compro-
mise, and hemodynamic compromise; they result in conversion to open repair in 1-4% of
procedures (6-8).

I. ACCESS FAILURE

Access complications may occur in as many as 20-30% of all implantations (9-11). Many
of these events can be avoided by careful preoperative planning and patient selection. Iliac
arteries less than 7.5 mm in diameter, which may be more difficult to traverse (12), are more
common in women (13,14). Fairman (10) has reported that the incidence of these
complications increases as the case difficulty and anatomic complexity increases. However,

adjuvant maneuvers and procedures can, in many instances, result in endovascular salvage -c

(10,15). Three characteristics that should be identified during planning to help minimize <j

complications include (a) the caliber of access vessels (common iliac, external iliac, and >3

common femoral); (b) the degree, location, extent and circumferential nature of calcium 4j

present; and (c) vessel tortuosity. Critical assessment of the iliac vessels should be under- 2

taken with noncontrasted and contrast-enhanced computed tomography (CT) scans. By |

comparing these images, the degree of calcification can best be appreciated. Three-dimen- @

625 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

626 FARBER and MENDES

Table 1 Catheter Sizes for Endo vascular
Device Delivery

Device

Outer diameter (F) a

Vanguard

20.5

Ancure

23.5

AneuRx

21

Talent

24-27

Excluder

18

Zenith

21

MEGS b

18

a 3F = 1 mm.

b Montefiore Endovascular Graft System.

sional reconstructions of CT scans or angiography can also be extremely useful in further
evaluating the extent of aortoiliac occlusive disease and tortuosity of the vessels. Vessel
tortuosity often compounds access issues even when stiffer wires — such as an Amplatz
Super Stiff (Boston Scientific) or Lunderquist (Cook) — are employed.

Access issues may also be device-specific. Knowledge of device- specific characteristics
such as flexibility, tip shape, hydrophilic nature, and delivery system profile is an integral
part of becoming proficient at device insertion without complications. It is rare, however,
that a single iliac vessel or device characteristic leads to an access failure. More commonly,
the cause is a combination of factors acting synergistically to prevent device insertion —
factors that were not appreciated at the time of initial evaluation.

When access failure occurs, there are several options to facilitate device insertion. These
include dilatation, angioplasty, wire manipulation, conduits, and extra-anatomic bypass.
One must keep in mind, however, that the goal is to completely exclude the aneurysm while
providing adequate inflow for the extremity. In an attempt to limit the need for unnecessary
associated procedures, the larger, straighter iliac vessel should be used as the primary site for
deployment.

A. Stenosis Management

Access vessel stenosis is best addressed during endovascular prosthesis implantation. If
angioplasty is attempted prior to stent-graft insertion, it may be difficult to obtain access
across the lesion. In addition, dissections or complications may arise from the angioplasty
itself, preventing future access and deployment of an endoprosthesis. This philosophy also
applies to the use of stents, which may decrease the functional luminal diameter of the vessel
or become dislodged during deployment of the device. Our approach is to treat iliac artery
lesions with sequential dilatation with Coons (Cook) hydrophilic dilators at the beginning
of the procedure to facilitate passage of the endoprosthesis and/or introducer sheaths. This

is typically performed by gently advancing a 16F dilator through the lesion and then -c

sequentially inserting larger dilators until the desired size is reached (usually 22-24F). Renal <

fascial dilators have been used in some cases; however, they are stiffer and constructed with a &

more abrupt tapered profile, rendering them less effective. Once the vessel has been 4j

successfully dilated, the device is inserted. After device deployment, angioplasty, if neces- Q

sary, can be undertaken in a more protected environment with the stent-graft in place. g

When recoil of the stenotic lesion exists, angioplasty may be necessary. When this is @

performed, vessel integrity and strength are compromised and can lead to vessel avulsion H,

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ENDOVASCULAR ACCESS 627

with removal of the sheath or delivery catheter. This occurs more frequently with iliac
bifurcation lesions resulting in external iliac avulsion. Vessel disruption can occur in per-
forming these dilatations or after device insertion, even though extreme care is taken. For
this reason, detailed completion angiograms of the entire iliac system should be performed
with small (<8F) sheaths in place to evaluate for residual hemodynamic lesions, dissection,
or vessel rupture. Larger sheaths may occlude ruptures or mask a dissection that could be
present. In addition, access wires should be the last intravascular component removed from
the patient prior to closure of the femoral vessels. It is customary to leave the wire in place for
several minutes after sheath removal and to monitor the patient for signs of bleeding from an
undiagnosed access vessel complication.

When residual problems of dissection or stenosis exist, placement of a stent may be
necessary to ensure stent-graft patency and viability of the extremity. If extravasation of
contrast exists, then a covered stent or modular stent-graft extension should be deployed to
prevent ongoing hemorrhage (see Sec. Ill, below).

B. Tortuosity

Vessel tortuosity is best negotiated by employing stiffer wires, as mentioned above. It is
helpful to provide slow countertension on the wire while inserting the stent-graft. This
reduces the potential for wire kinking, typically encountered with tortuous vessels. How-
ever, it should be noted that even though the vessel straightens with the use of stiff wires, the
vessel will return to its original conformation after the wire is removed. This may influence
exact endoprosthesis deployment and positioning. When tortuosity is felt to be the primary
cause of access failure, several options exist to circumvent it. Many times insertion of an
introducer sheath will alter the anatomy of the vessel and allow for passage of the device. If
this is unsuccessful, Yano (16) has reported that digital dissection and straightening of the
iliac artery has been successful in device insertion and reduces the risk of vessel rupture.
Closed iliac endarterectomy with or without endoluminal bypass, brachofemoral access
("body-flossing"), or an iliac conduit can also be utilized to facilitate device insertion (16-
18). Each of these options carries its own risks.

C. Iliac Conduits

The use of iliac conduits has been previously described (16,19) and is typically reserved to
facilitate device insertion in patients who have both severe occlusive disease and tortuosity.
Retroperitoneal exposure of the iliac vessel can be accomplished under regional anesthesia if
necessary. After exposure of the common and external iliac vessels an 8- or 10-mm prosthetic
graft is sewn onto the iliac artery above the level of the critical stenosis or tortuosity.
Generally Dacron is preferred for the prosthetic material, since it can more easily be secured
around the device or introducer sheath. Because of pelvic anatomy, insertion of the device ^

directly through the retroperitoneal incision is difficult and involves maneuvering at £

awkward angles. To overcome this problem, the conduit is typically routed to the groin £

and brought out under the inguinal ligament. In this way it can also be used as an iliofemoral li

bypass if desired. Care must be taken in traversing the anastomosis to avoid disruption, and =5

it is generally preferred to place a large introducer sheath at the onset of device deployment J

to prevent excess manipulation of the anastomosis. jj

Q

D. Brachial Artery Catheterization |

The brachial approach is an alternative option. When vessels cannot be traversed in a @

retrograde fashion, a wire and protection catheter can be placed from the left brachial %

I

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

628 FARBER and MENDES

approach (17). The wire is typically snared and brought out of the ipsilateral common
femoral artery. By placing the wire under tension at both ends, the endoprosthesis is inserted
more easily through tortuous vessels. While the brachial approach allows insertion of
devices that may otherwise never have been placed, it is not without risks. Left brachial
artery trauma and avulsion have been reported, and it also subjects the patient to a higher
risk of stroke, although there is no report of this in the literature associated with endo-
vascular aneurysm repair (20).

II. VESSEL COMPROMISE

Vessel rupture can occur after attempted angioplasty or subsequent to device insertion.
Avulsion of an iliac vessel is associated with larger introducers and introducer withdrawal.
In inserting these large devices, it is important to use fluoroscopic guidance of the iliac vessels
to avoid inadvertent perforation. When resistance is encountered, forcing the device should
be avoided. If avulsion or rupture occurs, it is important to maintain wire access across the
region of concern. Immediate availability of aortic and large iliac balloons is a necessary part
of safe endovascular abdominal aortic aneurysm (AAA) procedures. Over-the-wire balloon
insertion is used to convert an emergent situation into a controlled one. The balloon should
be centered across the injury and dilute contrast should be used for balloon inflation. This
allows for visualization of the balloon placement while still allowing for a rapid deflation of
the balloon. Hypogastric artery avulsion is rare but typically requires exposure and either
ligation or reimplantation.

Once control of the situation has been obtained, management can proceed with careful
consideration to the sizes and types of devices needed. For vessel disruption, either modular
stent-graft components or covered stents (Wallgrafts, Boston Scientific; or ViaBahn, W. L.
Gore & Associates) can be used. Wallgrafts seem better suited for more tortuous vessels and
external iliac lesions. When an avulsion occurs, the status of the hypogastric artery and
potential for common iliac seal must be evaluated. Of note, hypogastric artery avulsion may
be difficult to control with simple balloon tamponade. Often, placement of the primary
stent-graft may exclude the injured vessel segment. If an adequate seal has occurred above
the perforation, the stent-graft can be extended down to the common femoral vessels and an
endoluminal anastomosis performed. If this is not feasible, the vessels can be exposed with a
retroperitoneal incision and bypass grafting performed as needed. When a bypass graft
cannot be performed due to a severely damaged or aneurysmal vessel, then ligation with a
femorofemoral bypass can be employed to reestablish limb blood flow.

Advanced techniques for endovascular placement of a common iliac occlusion device
and crossover femorofemoral bypass exist; however this option should be reserved for the
more experience user until commercially available occlusion device are available.

1

III. HEMODYNAMIC COMPROMISE 1

Although hemodynamic compromise can occur anywhere within the limbs of an endovas- c

cular endoprosthesis, it arises most commonly within the native iliac vessels. Unsupported <j

devices may kink within tortuous vessels and require placement of intragraft stents. The >9

reported use intragraft stents for unsupported devices approaches 30% (21,22). For sup- 41

ported devices, molding and angioplasty should be performed when necessary. If needed, Q

bilateral pull-through or groin pressures can be measured to help determine whether a |

hemodynamic lesion exists. In addition, multiple views of the iliac vein can help to localize @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ENDOVASCULAR ACCESS 629

residual stenosis. Sometimes limitation of flow is caused by dissections distal to the
endograft. It is important to evaluate the iliac vessels without large sheaths in place to
avoid obscuring an iliac vessel complication. Dissections and vessel stenosis that are
resistant to angioplasty are typically best treated with stent placement. Choice of stent type
is determined by the length of the lesion, vessel tortuosity, and vessel size.

A. Common Femoral Artery Disease

Most patients with aortic aneurysms have associated atherosclerotic disease of the common
femoral vessels, which predisposes them to injury during endovascular aneurysm repair (23).
Injury to the femoral artery can occur in many of patients from repeated insertion,
withdrawal, and deployment of devices and sheaths. Disease is often worse at the femoral
bifurcation than at the level of the inguinal ligament. For this reason a high exposure of the
common femoral artery as it emerges from underneath the inguinal ligament is preferred. As
a result, femoral artery reconstructions may be necessary, requiring standard vascular
techniques. For device insertion, the choice of a longitudinal versus transverse arteriotomy is
dictated by the extent of atherosclerotic disease present. For vessels in this region that are
heavily diseased, a longitudinal arteriotomy is preferred, so that patch closure with or
without endarterectomy can be performed more easily. Focal dissections of the common
femoral vessel occur occasionally, and careful inspection of the posterior wall of the vessel
should be performed prior to closure. If necessary, tacking stitches should be placed or
endarterectomy performed to ensure adequate lower extremity blood flow. In rare cases of
severe disease or concomitant femoral aneurysmal disease, interpostion grafting can be used
for reconstruction. This approach, instead of routine or patch closure, may also be necessary
when large or multiple components are inserted through the femoral vessels. Prosthetic
material is often used for these conditions; however, autologous vein is also an option.

When inflow or common femoral repairs are performed, it is advisable to follow
patients closely for vessel occlusion or stenosis during the postoperative period. May has
reported a 9% incidence of acute lower extremity ischemia associated with endovascular
aneurysm repair (24). Acceleration times (25) and duplex inspection are an invaluable tool
in this region and can help detect problems before a clinical emergency develops.

Percutaneous approaches for contralateral limb management have been described (26)
with complications. Complications include vessel occlusion or disruption, embolic compli-
cations, and pseudoaneurysm formation. These may occur at the completion of the
procedure or during follow-up. In light of the minimal complication rate from common
femoral artery exposure, we prefer to perform bilateral femoral artery exposures to avoid
these complications. As newer smaller devices are introduced, this philosophy may need to
be reevaluated.

Embolic complication can occur from the access vessels; however, this is rare. Decreased ^

distal perfusion is more often the result of a dissection or common femoral artery £

complication. If embolic complications do occur, lower extremity angiography and throm- £

bectomy can be performed at the completion of the procedure. Adequate heparinization IL

should be maintained to avoid thrombotic complications during the procedure. =j

I
IV. CONCLUSION I

Access complications usually involve a combination of iliac calcification, tortuosity, and §

stenosis, which can be identified with preoperative imaging. While the presence of any one of @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

630 FARBER and MENDES

these aspects can easily be overcome, their combination can lead to technical failure and
misadventures during the procedure. Carefully planned endovascular techniques should be
employed to allow for safe device insertion. High-quality angiography and analysis of the
iliac vessels allows for early recognition of these complications. Safe repair can often be
achieved using endovascular techniques; however, conversion to open surgical intervention
may occasionally be required.

REFERENCES

1 . Blum U, Voshage G, Lammer J, Beyersdorf F, Tollner D, Kretschmer G, Spillner G, Polterauer
P, Nagel G, Holzenbein T. Endoluminal stent-grafts for infrarenal abdominal aortic aneurysms.
N Engl J Med 1997; 336:13-20.

2. Collin J. Transluminal aortic aneurysm replacement. Lancet 1995; 346:457-458.

3. Moore WS. The role of endovascular grafting technique in the treatment of infrarenal abdominal
aortic aneurysm. Cardiovasc Surg 1995; 3:109-114.

4. Armon MP, Yusuf SW, Latief K, Whitaker SC, Gregson RH, Wenham PW, Hopkinson BR.
Anatomical suitability of abdominal aortic aneurysms for endovascular repair. Br J Surg 1997;
84:178-180.

5. Carpenter JP, Baum RA, Barker CF, Golden MA, Mitchell ME, Velazquez OC, Fairman RM.
Impact of exclusion criteria on patient selection for endovascular abdominal aortic aneurysm
repair. J Vase Surg 2001; 34:1050-1054.

6. Buth J, Laheij RJ. Early complications and endoleaks after endovascular abdominal aortic
aneurysm repair: Report of a multicenter study. J Vase Surg 2000; 31:134-146.

7. Stelter W, Umscheid T, Ziegler P. Three-year experience with modular stent-graft devices for
endovascular AAA treatment. J Endovasc Surg 1997; 4:362-369.

8. Jacobowitz GR, Lee AM, Riles TS. Immediate and late explantation of endovascular aortic
grafts: The endovascular technologies experience. J Vase Surg 1999; 29:309-316.

9. Zarins CK, White RA, Schwarten D, Kinney E, Diethrich EB, Hodgson KJ, Fogarty TJ.
AneuRx stent graft versus open surgical repair of abdominal aortic aneurysms: Multicenter
prospective clinical trial. J Vase Surg 1999; 29:292-305; discussion, 306-308.

10. Fairman RM, Velazquez O, Baum R, Carpenter J, Golden MA, Pyeron A, Criado F, Barker C.
Endovascular repair of aortic aneurysms: Critical events and adjunctive procedures. J Vase
Surg 2001; 33:1226-1232.

11. Wolf YG, Fogarty TJ, Olcott CI, Hill BB, Harris EJ, Mitchell RS, Miller DC, Dalman RL,
Zarins CK. Endovascular repair of abdominal aortic aneurysms: Eligibility rate and impact on
the rate of open repair. J Vase Surg 2000; 32:519-523.

12. Naslund TC, Edwards WH Jr, Neuzil DF, Martin RS III, Snyder SO Jr, Mulherin JL Jr, Failor
M, McPherson K. Technical complications of endovascular abdominal aortic aneurysm repair. J
Vase Surg 1997; 26:502-509; discussion 509-510.

13. Fleischmann D, Hastie TJ, Dannegger FC, Paik DS, Tillich M, Zarins CK, Rubin GD.
Quantitative determination of age-related geometric changes in the normal abdominal aorta. J •o
Vase Surg 2001; 33:97-105. §

14. Velazquez OC, Larson RA, Baum RA, Carpenter JP, Golden MA, Mitchell ME, Pyeron A, |
Barker CF, Fairman RM. Gender-related differences in infrarenal aortic aneurysm morphologic •§)
features: Issues relevant to Ancure and Talent endografts. J Vase Surg 2001; 33:S77-S84. =s

1 5. Chuter TA, Reilly LM, Kerlan RK, Sawhney R, Canto CJ, Ring EJ, Messina LM. Endovascular d
repair of abdominal aortic aneurysm: Getting out of trouble. Cardiovasc Surg 1998; 6:232-239. ,-,-

16. Yano OJ, Faries PL, Morrissey N, Teodorescu V, Hollier LH, Marin ML. Ancillary techniques S
to facilitate endovascular repair of aortic aneurysms. J Vase Surg 2001; 34:69-75. °

17. Criado FJ, Wilson EP, Abul-Khoudoud O, Barker C, Carpenter J, Fairman R. Brachial artery I

©

I
u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOVASCULAR ACCESS 631

catheterization to facilitate endovascular grafting of abdominal aortic aneurysm: Safety and
rationale. J Vase Surg 2000; 32:1137-1141.

18. Wain RA, Lyon RT, Veith FJ, Marin ML, Ohki T, Suggs WA, Lipsitz E. Alternative techniques
for management of distal anastomoses of aortofemoral and iliofemoral endovascular grafts. J
Vase Surg 2000; 32:307-314.

19. White GH, May J, McGahan T, Yu W, Waugh RC, Stephen MS, Harris JP. Historic control
comparison of outcome for matched groups of patients undergoing endoluminal versus open
repair of abdominal aortic aneurysms. J Vase Surg 1996; 23:201-211; discussion 211-212.

20. Lin PH, Bush RL, Weiss VJ, Dodson TF, Chaikof EL, Lumsden AB. Subclavian artery dis-
ruption resulting from endovascular intervention: Treatment options. J Vase Surg 2000; 32:607-
611.

2 1 . Amesur NB, Zajko AB, Orons PD, Makaroun MS. Endovascular treatment of iliac limb stenoses
or occlusions in 31 patients treated with the ancure endograft. J Vase Intervent Radiol 2000;
11:421-428.

22. Parent FN GV III, Meier GH III, et al. Endograft limb occlusion and stenosis after Ancure
endovascular abdominal aneurysm repair. J Vase Surg 2002; 35:686-690.

23. Henretta JP, Karch LA, Hodgson KJ, Mattos MA, Ramsey DE, McLafferty R, Sumner DS.
Special iliac artery considerations during aneurysm endografting. Am J Surg 1999; 178:212-
218.

24. May J, White GH, Waugh R, Stephen MS, Chaufour X, Yu W, Harris JP. Adverse events after
endoluminal repair of abdominal aortic aneurysms: A comparison during two successive
periods of time. J Vase Surg 1999; 29:32-37; discussion 38-39.

25. Burnham SJ, Jaques P, Burnham CB. Noninvasive detection of iliac artery stenosis in the
presence of superficial femoral artery obstruction. J Vase Surg 1992; 16:445-451; discussion 452.

26. Traul DK, Clair DG, Gray B, O'Hara PJ, Ouriel K. Percutaneous endovascular repair of
infrarenal abdominal aortic aneurysms: A feasibility study. J Vase Surg 2000; 32:770-776.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

40

Device Failure

Tikva S. Jacobs and Michael L. Marin

Mount Sinai School of Medicine, New York, New York, U.S.A.

Larry H. Hollier

Louisiana State University Health Sciences Center School of Medicine,
New Orleans, Louisiana, U.S.A.

I. INTRODUCTION

Parodi performed the first endovascular repair of an abdominal aortic aneurysm over a
decade ago (1). More than 25,000 aortic stent grafts have since been deployed worldwide,
and preliminary results have been promising. However further follow-up and close moni-
toring to determine the long-term safety and efficacy of these devices has been recommended
(2-6). Problems with deployment, stent-graft migration, endoleak, material failure, and
aneurysm rupture have all been reported (7-11). Many of these problems were seen with
first-generation stent grafts, suggesting parallel learning curves between the surgeons, device
engineers, and manufacturers. Technical and mechanical device problems have been ad-
dressed and individual implants improved. However, new problems continue to be dis-
covered as patients with second-generation stent grafts approach midterm follow-up. As an
increasing number of explanted grafts become available for analysis, new and device-specific
material failure is being identified.

Device fatigue remains one of the most concerning modes for potential procedure ■§

failure, encompassing the breakdown of the intrinsic mechanical parts of the stent graft. It g

is often difficult to identify device fatigue, as patients are typically asymptomatic at the as

time of presentation. Many of the first identified stent fractures were initially recognized c

within explanted stent-graft devices that had been removed for evidence of aneurysm <

expansion or recovered at autopsy. The challenges of identifying material failure have >9

made a true understanding of the magnitude of the problem difficult. Moreover the clinical J

significance of many identified failures is unknown. Q

The purpose of this chapter is to familiarize the reader with the more common modes |

of material failure associated with individual devices and to dicuss the frequency, cause, @

633

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

634 JACOBS et al.

and significance of such material failures, as we have learned about these through our in-
stitution's own experience.

Multiple devices have been implanted in patients worldwide; even though some are no
longer available for clinical use, it is important to educate oneself about them so that their
failure can be recognized early and monitored closely. Prior to discussing endovascular
stent-graft failure, this chapter first offers a short review of the techniques used for
endovascular repair and a description of the actual devices used, so as to help the reader
understand where and why these devices fail.

II. BACKGROUND

A. Techniques for Endovascular AAA and TAA Repair

The four techniques used to deploy endovascular stent grafts in the repair of abdominal
aortic aneurysms (AAAs) are described in detail elsewhere (12-17). In brief, they are
aortoaortic, aorto-uni-iliac, bifurcated-modular, and bifurcated-unibody (Fig. 1A to E).
Although the aortoaortic tube graft is seldom used for primary AAA repair today, it is
used to repair thoracic aortic aneurysms (TAAs). The technique chosen for such repairs is
dependent on the anatomy of the aorta and the iliac arteries as well as the patient's clinical
presentation and its relevance to specific protocols. A variety of devices are deployed using
these four techniques. Some of these devices are no longer available for clinical use, while
others are currently under investigation by the U.S. Food and Drug Administration
(FDA) and still others are being developed. What these devices have in common is the
combination of a metallic stent and a material graft. However, how they differ helps to
explain not only why some become subject to fatigue but also where and how.

B. Devices for AAA and TAA Repair

1. Individually Fabricated Endovascular Graft

The first endovascular stent grafts were "homemade" devices, fabricated by individual vas-
cular surgeons. The first abdominal aortic aneurysm treated with an endovascular stent-
graft repair was constructed by Parodi using a custom-made device comprising a Palmaz
stent attached to a polyester tubular graft (Barone Industries, Buenos Aires, Argentina) (18).
Similar endovascular stent-graft devices were individually constructed throughout the world
using components designed for other purposes, including the Chuter device, Sydney endo-
vascular graft and the Volodos/Kharkov Institute Device (19,20). At our institution, a mod-
ified Parodi endovascular AAA device was fabricated. It was constructed from a thin-walled,
funnel-shaped polytetrafluoroethylene (PTFE) graft sutured to a stainless steel Palmaz stent
(Cordis Endovascular, a Johnson & Johnson Company) (Fig. 2). The Palmaz stents used for
device anchoring are balloon-expandable and made of 316L stainless steel (20). g

2. Commercially Fabricated Endovascular Devices J

As endovascular repair of aortic aneurysms began to gain acceptance in the vascular c

community, commercially fabricated stent-graft devices were developed and FDA trials <

began. Discussing the components needed for the "ideal" endovascular stent graft seems >9

relatively straightforward; however, turning those ideas into reality is more challenging ^

(21,22). Through the years, multiple devices were introduced for experimental use and «

individual devices underwent numerous revisions. The first stent graft to undergo clinical |

trials in the United States was the EVT/Guidant Endovascular stent graft. @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

DEVICE FAILURE

635

Figure 1 Artist illustration of techniques for endovascular repair of abdominal aortic aneurysms,
a. Aorto-aortic tube graft, b. Aortouni-iliac reconstruction. An aortouni-common or external iliac
graft is inserted. A femorofemoral bypass combined with a contralateral common iliac artery oc-
cluder is used to complete the reconstruction, c. Endovascular bifurcated aortic graft reconstruction.
Two techniques are suggested, i. Modular bifurcated reconstruction employing an aortic body and
contralateral limb inserted separately; ii. Unibody bifurcated reconstruction employing a single
endograft and a cross femoral wire to bring one limb to the contralateral extremity.

I
•5

Figure 2 A Parodi-type handmade balloon expandable stent grafting system. This graft is com-
posed of a Palmaz balloon expandable stent sutured to a tapered polytetrafmoroethylene graft.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

636

JACOBS et al.

«rt ss *c; ~

es r tta ; '—

.

£

Figure 3 The Guidant Ancure endovascular grafting system, a. The first generation device
produced by the Endovascular Technologies Corporation was composed of a polyester woven graft
with an attachment apparatus at either end. The attachment devices are constructed, of elgiloy metal
bent to form Z configured hooks, b. The device was later modified into its current form employing a
varied attachment system with the same principle of Z configured metal coupled with penetrating
hooks for aortic fixation. Loose polyester fibers were left on the outside of the revised graft to
stimulate endoluminal sealing, c. The initial EVT graft permitted the fixation hooks to be laser
welded onto the Z-configured stent spring, d. Separation of the hooks from the Z stent resulted in a
design modification where by each hook was sutured onto the surface of the graft directly (see also
Fig. 10). e. The original EVT hook had an acute angle at its junction point with the main hook shaft,
f. Fractures occurred at this angle necessitating a more gradual curve in the hook structure.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

DEVICE FAILURE

637

EVT ( Endovascular Technologies ) / ' Guidant- Ancure Device. The EVT stent graft com-
prises a polyester fabric vascular graft hand-sewn to self-expanding Z-configured elgiloy
(cobalt-chromium and nickel) metal stents at the proximal and distal attachment sites
(Fig. 3). The aortic attachment portion of this device has eight angled hooks to aid in the
fixation of the graft to the aortic wall. In the initial model, eight individual shanks with
hooks at the distal end were laser-welded to the zigzag stent rings (Fig. 3C). Hook
fractures were encountered during the phase 2 trial, prompting modification of this design.
Presently there are four V-shaped shanks containing two hooks each, which are sewn to
the polyester graft (Fig. 3D). The shape of the hook has been modified as well to incor-
porate a gradual angle rather than a sharp right angle, which is prone to fracture (15,23,
24) (Fig. 3E and F). This device is available in bifurcated, aorto-uni-iliac, and tubular
configurations.

Vanguard Device. The Vanguard (Boston Scientific Inc., Nadile, MA) endovascular
stent graft is a modification of the Stentor stent-graft system (25). It is modular graft
available for tube or bifurcated repairs. It comprises a flexible, self-expanding prosthesis
constructed of a nitinol wire frame coverd by a thin-walled woven polyester fabric. The
Vanguard stent is composed of independent rows of zigzag nitinol wires that are held
together by polypropylene suture ties. These ties are located at the apex of the metal bends,
allowing movement of the individual nitinol stent rows to accommodate for arterial
angulation. The graft is a seamless low-porosity polyester fabric attached at the proximal
and distal ends of the device by polyester sutures (Fig. 4). There are small barbs on the
proximal end of the graft to assist in anchoring the device to the aortic wall (2,6,26).

I
•5

Figure 4 The Boston Scientific Vanguard endovascular grafting system, a. The endoskeleton of the
Vanguard system is composed of Z configured nitinol wire fixed at the apices by polypropylene
sutures. An aortic body and ipsilateral limb are displayed with a contralateral iliac stent limb. b. The
nitinol skeleton is covered with a thin walled polyester graft in the complete prosthesis for clini-
cal use.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

638

JACOBS et al.

Talent Device. The Talent (Medtronic AVE/World Medical Manufacturing, Minn.,
MN) device is composed of a series of self-expanding serpentine nitinol stents, several of
which are connected to a longitudinal nitinol bar. Each stent is sutured over its entire
length to a polyester graft. The fabric material is fixed to the outside of the nitinol stents in
the aortic body, while the stents are on the outside of the graft, on the limbs. The proximal
aortic fixation device can be an uncovered nitinol stent, allowing for transrenal fixation
(Fig. 5). The Talent endovascular stent-graft system has a configuration designed for tube,
aorto-uni-iliac, or bifurcated graft use and can be manufactured in custom-made sizes de-
signed on the basis of computed tomographic (CT) or angiographic measurements (26,27).

The AneuRx Device. The AneurRx (Medtronics) endovascular device is a self-
expanding modular bifurcated stent-graft system (Fig. 6). The modular component
consists of a thin-walled noncrimped woven polyester graft supported by suture-fixed
individual 1-cm nitinol rings, forming an exoskeleton. The nitinol stents are laser-cut out
of nitinol hypotubing and sewn to the graft with polyester sutures (4,20).

Gore Device. The Gore Excluder (W.L. Gore Associated, Inc.) is a modular system
with a self-expanding Nitinol exoskeleton attached to an ePTFE graft internally (Fig. 7).
Angled wired barbs are attached to the proximal end of the main device to assist in
anchoring the device to the aortic wall. Polyethelene tape attaches the ePTFE graft to the
nitinol exoskeleton to avoid sutures or suture holes in the graft material (5,20).

Figure 5 The Talent endovascular grafting system, a. The Talent endovascular tube graft is
composed of Z configured nitinol stents covered on the external surface with a polyester graft. Each
Z shaped stent is sutured completely to the surface of the fabric material with suture material, b. The
Talent modular bifurcated stent graft system. A main aortic body containing an ipsilateral limb is
joined with a contralateral limb to reconstruct the aortic bifurcation. Note in the inset the
longitudinal bar present in the aortic body seen with trans-illuminated imaging, c. Aortouni-iliac
stent graft system. A tapered Talent endograft is displayed along with a single complete polyester
encapsulated Z stent, which is used for the contralateral common iliac occlusion.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

DEVICE FAILURE

639

Figure 6 The AneuRx endovascular graft. The AneuRx graft is composed of individual rings of
nitinol laser cut into nitinol hypotubing. Each of the Z rings is sutured to the fabric graft by a
blanket stitch. A modular bifurcated device is demonstrated.

The Cordis LP Device. The modular Cordis LP bifurcated endvascular stent graft con-
sists of a nitinol stent sutured within a seamless woven polyester bifurcated graft (Fig. 8A).
The senitinol stents are laser-cut from seamless nitinol hypotubing and heat-treated to
achieve proper thermal properties. The nitinol is attached to the fabric using multiple
polyester point and blanket sutures. The transrenal attachment device has broad open
spaces that permit stent-graft fixation across the renal arteries, and six to eight fixation barbs
angled caudad at its base at the proximal end of the graft fabric to penetrate the aortic wall

I
•5

Figure 7 The W.L. Gore TAG™ endovascular graft for thoracic aortic aneurysms. This endo-
vascular graft is constructed of expanded polytetrafluoroethylene fabric with external Z configured
nitinol wire support. The nitinol wire frame is fixed to the outer surface of the polytetrafluoro-
ethylene graft by means of ePTFE membrane.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

640

JACOBS et al.

jam . H

Pi W

M

\ y V v s b J

l * K

■ \ f \ / \ /

7 V |

8d

Figure 8 The Cordis Quantum LP™ bifurcated endvascular stentgraft system, a. The Quantum
LPTM is comprised of nitinol stents sutured within a seamless woven polyester bifurcated graft. The
nitinol stents are laser cut from seamless nitinol hypotubing and heat-treated to achieve proper
thermal material properties. The nitinol is attached to the fabric using multiple polyester point and
blanket sutures, b. The transrenal attachment device has broad open spaces that permit stentgraft
fixation across the renal arteries, c. Six to eight fixation barbs angled caudad at its base at the
proximal end of the graft fabric to assist in aortic fixation, d. Along the length of the stump the graft
has circumferential crimps or fabric hinges that form five discrete pockets for fixation of the
sinusoidal nitinol Z-stents. This configuration allows flexibility without kinking and allows the legs
to absorb cyclic motion in the unstented areas.

below the renal arteries and firmly anchor the stent graft in position (Fig. 8B,C). There are
two 5-cm equal-length stumps at the end of the aortic body stent into which the iliac leg
prostheses are inserted and permanently fixed. Along the length of the stump, the graft has
circumferential crimps or fabric hinges that form five discrete pockets for fixation of the
short sinusoidal nitinol Z stents. This configuration allows flexibility without kinking and
allows the legs to absorb cyclic motion in the unstented areas (Fig. 8D). The iliac limbs are
designed to overlap with in the stump legs of the aortic prosthesis a minimum of 2 cm and a
maximum of 5 cm (28).

Thoracic Aortic Stentgrafts. Two thoracic stent grafts are commonly used, the Talent
device (Medtronic AVE/World Medical Manufacturing, Minn., MN) and the TAG (W.L.
Gore Associates, Inc.). The Talent stentgraft is a tubular device that is identical in design
to the one previously described (see above). The TAG consists of self-expanding Z-shaped
nitinol stents covered internally by ePTFE material. The stent members are fixed to the
graft without sutures by means of a continuous ePTFE membrane. Both designs have a
longitudinal bar running the length of the stent graft. Multiple overlapping stent grafts
may be used with either device to achieve adequate length to exclude an aneurysm (29-31).

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

DEVICE FAILURE 641

C. Evaluation of Material Failure

Material failure has been a problem since early first-generation stent-graft trial. In 1996,
Moore reported metal fractures in the EVT device. The patient had a persistent endoleak
12 months postimplantation and underwent open repair. At exploration, proximal and
distal hook fractures were found. Review of this patient's follow-up plain films revealed
evidence of hook fracture at the 6-month follow-up. This fracture led to a retrospective
review of all plain films at that institution. Nine patients were found to have fractures in
one or more of the metal components of the attachment system. Of these, only one patient
went on to require open conversion 18 months after initial implantation (24).

In 1997, Blum et al. and Mialhe et al. reported early results of the Stentor endo-
vascular stent graft. During this time, "leaks" caused by tears in the polyester graft along
its suture line were found at angiography. While some of these leaks sealed spontaneously,
others required secondary procedures and were corrected with a second stent graft (25,32).
Both of these first-generation stent-graft devices underwent redesign. EVT was acquired
by Guidant, who redesigned the metal fixation device as described earlier. Boston Scien-
tific took over Stentor and introduced a low-water-permeability polyester fabric that did
not require suturing. Ancure and Vanguard were, respectively the new models introduced,
in hopes that they would succeed where their predecessors failed. Although material fail-
ure was discovered by persistent endoleaks, these findings prompted a closer surveillance
of stent-graft devices and closer attention to design, since material failure was now a
known entity that could be identified on plain film x-rays.

III. MODES OF FAILURE

Over the last 10 years, reports of device failure have been scattered throughout the
literature in the form of case reports or follow-up results of FDA trials (6,9,33-35). These
isolated reports prompted our institution to review its own experience with endovascular
stent grafts and device fatigue.

686 patients underwent endovascular aortic aneurysm repair (over a ten-year period)
and were followed prospectively in a database. Of these, 404 had a full complement of
follow-up analyses for review and formed the study set for our investigation. A total of 60
patients (15%) of this subset demonstrated stentgraft fatigue as identified by x-ray studies
or the analysis of an explanted stentgraft device. Forty-nine (81%) of those patients had
been treated for an AAA, while 11(19%) had a TAA repair. Patients were variably fol-
lowed at 1, 3, 6, and 12 months and annually thereafter with physical exam, plain film
abdominal or chest x-rays, duplex ultrasonography, and spiral CT scans. Of the 60 pa-
tients with stentgraft failure, 55 were identified with x-ray analysis and 5 were found on
inspection of explanted stent grafts (36). ■§

The 60 fatigued stent grafts in this study were distributed among seven different graft g

types (Table 1). The average time to fatigue for all devices combined was 19 months (range »

1-48 months), the average follow-up since fatigue identification being 8 months. A marked c

variation in the time to failure was seen depending upon the type of fatigue, — i.e., suture <

fracture, graft holes, or stent fractures (Table 2). Of the 60 patients followed with graft >9

fatigue, 1 1 expired during follow-up and 1 was lost to folow-up 2 years after his fatigue was ^

identified. Of the 11 deaths, 3 were device-related and 1 patient died on post-op day 15 °

secondary to complications suffered after an open surgical conversion. A second patient's |

death was device-related, caused by an aneurysmal rupture in the presence of a type I @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

642

JACOBS et al.

-

SO

3

3

o

I

a
c?5

3

a>

.Q

o

£>

o

.s

to

'so

c

C^

7Z

i-t

d

i

2

G

Gi)

•1

5

EI)

C3

=1

■—

tu

O

>

<

<£

<D

SJJ

o

c

^

OJ

^— ^

1

u

=1

v»

bO

OJ

•pj

CJJ

rt

^

ti-J

'—

0>

>

<

3

u

eg

<£

1^

-

^

El)

1

a

O

o

ed

H

<h

T3

tu

X

>
to

u

T3

a

C

o

J5

H

&

J

o

l-i

t$

o

Ph

a

£

a

s

60

(U

>,

O

Ih

Q

3

a

&

-

GO

[»

a

ft

m

CS

CS

~6

' —

tM

'—

G

c

3

a

^

O

c

<y:

O

ON

CS

, — .

, — .

r<~<

1

^c

J=

CO

OS

co

T3

co

xj

ja

£

C

-=

a

c

a

a

c

G

c

O

c

s

a

S

S

£

n ^ 'yQ Tf n

m ^ n ^-<

,Q CO CO CO CO

« ^ ^ ^ ^
£0000

m 00 (N o

mo r-H m ^H ,-H

SS

s?

(N

10

-f

r-

; '. ' '

■

~~

©^ o^ s

IO ON ^H

00 ^ (N

O O

o o

no r^ ^- o on 00 o
cs m no cs m t-H ^~

ON

NO

=

o
ri

7

CO CO CO

ri

O m 't
-— < m o

(N "^ GO

c3

lis

> H S

a

G

a

O

H

o

H

a
o

5
I

SI)

o

a

I
a

o

o

J3

2? Dh

s *

c<3 a

u — <

•a 3
^ a

•a a

S .Si

& ^

O aj
so so

s .s

-a "3
pq w

u

b °
a o

ft CS

o ©
o so

a §"

so ^

tU NO

S m

BO «

s ^

•3 ■ ■

!>5

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

DEVICE FAILURE

643

Table 2 Time to Diagnosis of Device Fatigue Based on Failure Type

Device

Type of fracture

Number of failures

Average time to fatigue (range)

AAA

Vanguard

Suture disruption

14

25 months (3-48)

Graft hole

2

Intraoperative and 36 months

Talent

Metal fracture

23

13 months (1-31)

Graft hole

1

16 months

AneuRx

Metal fracture

3

10 months (1-24)

MEGs

Metal fracture

5

38 months (33-48)

EVT

Metal fracture

1

8 months

TAA

Gore

Metal fracture

7

24 months (3-38)

Talent

Metal fracture

4

9.5 months (1-24)

endoleak. The stent fracture was in the proximal region of the graft, while the endoleak
occurred at the distal attachment site. The endoleak was not amenable to endovascular ex-
clusion and the patient was not a candidate for open repair. The third patient died 3 1 months
after stent-graft implantation from a ruptured aneurysm. The patient had been noncom-
pliant after his 12-month follow-up and, in the interim, developed a type I endoleak that
became symptomatic at the time of rupture. The remaining 8 patients died of causes
unrelated to their aortic aneurysms (Table 3). Excluding the one early postoperative death,
the average time to expiration was 29 months (range of 1 7-52 months) postimplantation and

Table 3 Mortality of Patients with Stent-Graft Device Fatigue

Time from operating

room device

Time from

Device-

Patient

Cause

implantation

fracture

Device"

related

1

Congestive heart failure

27 months

4 months

Vanguard

No

2

Emphysema

34 months

26 months

Vanguard

No

3

Colon cancer

24 months

9 months

Vanguard

No

4

Lung cancer

30 months

26 months

Vanguard

No

5

Post-op from open
surgical conversion

POD #5

NA

Vanguard

Yes

6

Myocardial infarction

17 months

12 months

Vanguard

No

7

Cardiac arrest

23 months

20 months

Vanguard

No

8

Cardiac arrythmia

52 months

4 months

MEG's

No

9

Prostate cancer/Ml

26 months

2 months

Talent

No

10

Ruptured AAA

31 months

NA

Talent

Yes

11

Rupture of type I
Endoleak* 3

23 months

NA

Talent-TAA

Yes

Abbreviations: OR, POD, post-operative day.

a Vanguard devices were implanted 2 years prior to the Talent device and therefore have had longer
follow-up.
Leak diagnosed prior to rupture; however, patient was not a candidate for open surgery.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

644

JACOBS et al.

13 months (2-26 months) after fatigue was first identified. The remaining patients continue
to be followed for clinical sequelae of their stent-graft fatigue (36).

A. Stentgraft Failure Analyzed by Device

Of the 60 patients with device fatigue, 16 had Vanguard AAA stent grafts inserted (Table 4);
9 failures occurred in bifurcated grafts and 7 in aortic tube grafts. There were 14 suture
disruptions, 5 proximal row separations, and 9 body separations. Two explanted devices
were found to have multiple wear holes through the graft fabric (Fig. 9) (36).

Twenty-four material failures occurred in Talent stentgrafts. Sixteen Talent patients
had bifurcated grafts while 6 fatigues occurred in aorto-uni grafts and two in a tube graft.
Fatigue was recognized within several different regions of the endovascular graft devices.
Twenty-three patients had fractured stents. Fourteen occurred along the longitudinal bar
of the graft (7 in the aortic body proximal bar, 1 in the aortic body distal bar, and 6 in an
iliac limb) (Fig. 10A to C). Nine fractures were detected in the serpentine nitinol wire of
the Talent device (4 in the body and 5 within the proximal transrenal stent) (Fig. 10D and
E). One patient had a wear hole detected in an explanted prosthesis at the site of graft-to-
stent fixation (Fig. 10F) (36).

The nine remaining fatigued stent grafts occurred in three different device designs, the
MEGS, the EVT, and the AneuRx. Metallic fractures occurred in 5 patients with MEGS
stent grafts, all after at least 33 months of follow-up. Four occurred within the proximal
diamond row portion of the Palmaz stainless steel stent in the setting of maximal stent
dilatation (Fig. 11) and one in the second row. The EVT device was noted to have distal
hooks and proximal shank fractures approximately 6 months after insertion (Fig. 12).
Three patients had fractures of their AneurRx device (Fig. 13). Metal fatigue within

Table 4 Failure Mode Analysis by Device

Device

Number failed

Location

Number

AAA

49

Talent

24

Graft hole

1

Longitudinal bar

14

Aortic body, proximal

7

Aortic body, distal

1

Iliac

6

Proximal transrenal stent

5

Z-shaped body stent

4

Vanguard

16

Row separation

5

Body separation

9

Graft hole

2

EVT

1

Hooks and shanks

1

MEGS

5

Top row of stent

4

Second row

1

AneuRx

3

Stent fracture

3

TAA

11

Gore

7

Isolated longitudinal bar

3

Longitudinal bar and Z stent

2

Z stent

2

Talent

4

Longitudinal bar

1

Z stent

3

I

•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

DEVICE FAILURE

645

Figure 9 Clinical examples of fatigue in the Boston Scientific Vanguard endovascular graft. A 69
year old man had a tube graft inserted for the repair of an abdominal aortic aneurysm, a. Proximal
row separation was defined on this graft along with a distal endoleak secondary to retraction of the
prosthesis, b. The graft was explanted and a conventional repair was done at 36 months. Arrow
points to the site of the proximal row separation, c. High-powered magnification of the graft
depicted in 9b demonstrates fabric fatigue and wear holes on the prosthesis surface, d. A 79-year-old
man had an endovascular tube graft placed for the repair of an AAA. On surveillance abdominal x-
rays suture fractures in the body of the graft are detected (arrow). This patient remains in a
surveillance program (From Ref. 36. Reprinted with permission.).

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

646

JACOBS et al.

Figure 10 Clinical examples of fatigue in the Talent endovascular system, a. A proximal
longitudinal bar fracture in a Talent endovascular graft (arrow), b. A distal aortic longitudinal bar
fracture (arrow), c. An ipsilateral limb longitudinal bar fracture in a Talent endovascular graft, d. A
midbody Z-stent fracture (arrow), e. A proximal transrenal stent fracture in a Talent endovascular
graft. Note the fracture has occurred adjacent to the site of the nitinol wire crimp, which contains the
two ends of the Z-configured stent, f. Following a persistent Type I endoleak a 70 year old man had his
endovascular graft explanted and a conventional repair completed. The explanted graft demonstrated
signs of graft wear with frayed fabric yarns and the creation of a defined hole (arrow). (From Ref. 36.
Reprinted with permission.).

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

DEVICE FAILURE

647

Figure 10 Continued.

I
•5

Figure 11 The Palmaz balloon expandable proximal stent attachment device of the handmade
Parodi-Palmaz system. Note a fracture of the proximal diamond row (arrow) (From Ref. 36. Re-
printed with permission.).

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

648

JACOBS et al.

Figure 12 Attachment system of the EVT endovascular grafting system. Note fracture of the hook
shank at the site of its laser weld on the Z-configured attachment system (arrow) (From Ref. 36.
Reprinted with permission.).

I
•5

Figure 13 An abdominal X-ray of a patient who underwent AneuRx stentgrafting of an
abdominal aortic aneurysm. Arrow denotes site of a fracture in one of the nitinol rings on the stent
strut (From Ref. 36. Reprinted with permission.).

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

DEVICE FAILURE

649

AneuRx grafts was extremely difficult to identify conclusively on plain film examination
secondary to overlapping metal densities of the stent and the relatively close diamond
pattern (36).

Eleven patients with thoracic aortic stent-graft devices were identified that had device
material failure. Seven Gore TAG prostheses were found to have stent fractures and four
Talent thoracic grafts demonstrated evidence of metal fatigue. One TAG graft had a strut
fracture seen on chest x-ray films 3 months postimplantation and subsequently developed
a second fracture of a longitudinal nitinol bar detected at 32 months. Two patients (one
Talent and one TAG) presented with a sudden onset of pain and were diagnosed with
enlarging aneurysms and new type I endoleaks on CT scan. The patient who had the TAG
prosthesis underwent open conversion and device explantation (Fig. 14). The patient with
the Talent graft was not a candidate for open surgery and ultimately died of aneurysmal
rupture. The remaining 8 patients had stent fractures detected in the longitudinal bar (3)
and z stents (5) of the stent graft (36).

B. Metallic Fracture

The 42 metallic stent fractures analyzed at our institution occurred in devices fabricated by
six different manufacturers. Thirty-six fractures were documented in superelastic Nitinol
stents, 5 fractures in stainless steel, devices and 1 in an elgiloy stent. Of the 686 devices
implanted in this investigation, 493 (72%) were fabricated from nitinol, 24% were stainless
steel, and 4% were elgiloy.

I
•5

Figure 14 A high resolution explant radiograph of a W.L. Gore TAG stentgraft that was used to
repair a thoracic aortic aneurysm. This patient presented 32 months after endovascular repair of a
thoracic aneurysm with a new endoleak. The endoprosthesis was explanted and a conventional
reconstruction was successfully accomplished. Explanted examination of the endograft using a plain
film x-ray demonstrates Z-configured wire fractures (arrow-heads) and a longitudinal bar fracture
(full arrow) (From Ref. 36. Reprinted with permission.).

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

650

JACOBS et al.

As mentioned previously, elgiloy metal fractures were first reported in the early EVT
devices. Eight of the 39 patients enrolled in that trial were found to have hook fractures
(24). This device was withdrawn from clinical use and has since been redesigned to the
improved Ancure/Guidant device. Even though the hooks were remodeled to decrease the
stress on the metal, Najibi et al. recently reported two cases of delayed isolated hook
fractures that occurred at 36 months postimplantation in the proximal attachment system
of the Ancure device (34). Both patients were asymptomatic and the clinical relevance of
their fracture for the device as a whole has yet to be determined. Although the EVT device at
our institution had similar fractures to those reported in the literature, there was no evidence
of metallic fracture in those devices implanted using the remodeled Ancure system. How-
ever, the mean follow-up for Ancure patients at our institution was only 13 months.

Although stress fatigue has been suggested as the likely cause of metallic fracture in the
EVT and later Ancure devices, other forms of failure including metal corrosion have been
postulated as the cause for the nitinol fractures in early explanted Stentor grafts (35,37).
Heintz et al. found evidence of corrosion on scanning electron microscope (SEM) studies of
Stentor explants, with more severe irregularities detected in those stents implanted for
longer time periods. The explanted stents reviewed in our study failed to show significant
signs of corrosion, even in high-risk regions of the metal, and did reveal a relatively uniform
surface oxide layer, which may provide resistance to corrosion (Fig. 15A to D). The lack of

Figure 15 Scanning electron microscopy (SEM) of explanted endovascular grafts for AAA. a. The
Vanguard endovascular graft has platinum wire wound around the proximal and distal portions of
the attachment system to improve x-ray visualization (x40). b. Despite the suspected increased
potential for metal fatigue and corrosion at the interface of these two metals no such fatigue was
found in the nitinol in this 36 month old implant (200 x) c. A SEM of a Talent stent explanted 15
months after insertion. The site depicts the crimped region of the two ends of the metal bar. No
significant wear or pitting was detected on the surface of the nitinol stent at this presumably high risk
area for fatigue (x40). d. High powered (4K) magnification of the surface of a Talent nitinol spring
after explantation. A rough appearing, uniform oxide layer is discerned (From Ref. 36. Reprinted
with permission.).

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

DEVICE FAILURE 651

corrosion on the explanted devices in this study compared to the results of the earlier
Stentor models may reflect an improved understanding of nitinol processing. Surface
treatments such as electropolishing (38), TiN annealing (39), heat treatments, and nitric
acid passivation (40) have all been shown to improve the corrosion resistance of nitinol by
enhancing the formation of a thin, uniform oxide layer on the surface of the metal (41).

With improvements in corrosion resistance and the absence of significant evidence of
corrosion identified on the newer stent grafts explanted at our institution, other causes of
metallic fracture should be considered. Studies have shown that structural changes in the
morphology of the stent graft, such as foreshortening of the aneurysm after exclusion (42),
or the impact of aortic pulsations and the associated cyclic loading during systole and
diastole (43), may result in accelerated endovascular stent graft fatigue. Harris et al. found
kinking responsible for complications such as limb occlusion and migration or separation
of the modular components in the Stentor/Vanguard device (42). Although the inves-
tigation by Harris and associates did not show evidence of metallic fracture, one can
postulate that those structural changes that occur secondary to tortuosity of the aorta and
cyclic loading could ultimately result in fracture. Umschied et al. reported kinking in
stentgrafts resulting in similar complications with the Stentor/Vanguard devices. That
study also reported metallic fracture in the longitudinal bar of Talent stentgraft devices
(43). At our institution, 18 fractures occurred in the longitudinal nitinol bars of Talent or
Gore stentgrafts. These events were associated with tortuosity of the implant vessels with
presumed increased stress across the nitinol wire.

Microcracks, which are material irregularities produced commonly during metal laser
cutting, together with the pulsatile aorta, may ultimately result in surface disruption and
crack propagation. At the present time limited in vivo information exists on fatigue crack
propagation for endovascular stents. The majority of studies have been done on larger
implants for orthopedic applications or heart valve replacements (44). These applications
experience very different loads, and endovascular stents are manufactured with much
thinner metal strut widths, approximating 250 |im. At these tolerances, even if the material
used has a high threshold for crack propagation, there is a very small distance in the stent
structure for the crack to propagate prior to fracture (45). In our review the metallic
fractures in the diamond row portion of the Palmaz stainless steel stents, or in the zizzag
portion of the nitinol ring stents, could have been due to a combination of the causes
mentioned above, which emphasizes the importance of further studies in vivo of fatigue
crack propagation for endovascular stent grafts.

C. Fabric Fatigue

The polyester material used in stent grafts today is similar to the material used in con-
ventional AAA repairs. Degradation of the polyester fabric is known to take place after ■§
10-20 years following conventional surgical implantation (46). However, the fabric wear g
seen in implanted endovascular stents is reported much earlier. Fabric fatigue can be »
accelerated when mechanical stresses combine with the extrinsic forces and electro- c
chemical properties of the intravascular environment. Fabric fatigue was one of the first <
reported causes of aneurysmal rupture and stent-graft failure in the earlier Stentor graft &
(6). External abrasions of the fabric on the metal stent as well as frank graft holes have J
been described in the literature (37,47). Both can be related to a combination of pulsatile «
flow of the aorta and the geometric characteristics of the stent itself. Movement between |
the stent graft and the arterial wall can cause fabric wear and surface breakdown, resulting ©
in external abrasions (47). The graft holes are probably caused by micromotion of the 1,

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

652 JACOBS et al.

individual stents against the polyester fabric, causing abrasion of the material and a likely
type III endoleak (6). At our institution, 5 explanted stent grafts had evidence of fabric fa-
tigue: 2 in Vanguard grafts, 2 in a Talent AAA stent graft, and the last in a Gore TAG.
In the Stentor and Vanguard devices, polypropylene ties fasten the individual stent
rows to each other. The stent itself is attached to the graft only at the proximal and distal
ends, allowing for continuous motion between the stent and the graft. Fewer examples of
graft fenestrations have been reported with the Talent design, which could be secondary to
complete fixation of the graft by sutures to the stent, thereby decreasing the amount of
movement between the graft and the metallic nitinol stents. Material fatigue can also occur
during manufacturing and packaging. In our study, one patient was found to have a type
III endoleak immediately after deployment; on explantation fabric fatigue was seen. This
can occur when the stent graft is drawn into the catheter from the opposite direction from
which it will be deployed. Then the angled points of the stents can press firmly against the
graft fabric, potentially causing a small tear or hole (36).

D. Suture Breakage

The polypropylene sutures used to assemble the Stentor/Vanguard device may also be
subject to fatigue and contribute to device failure. Micromotion of the nitinol endoskeleton
secondary to the pulsatile flow of the aorta causes friction and wear of the sutures, with
ultimate suture fracture and stent row separation. As of January 2001, there has been a 21%
prevalence of row separation reported in the Vanguard model (48). Riepe et al. found that
the most movement took place in the large frames of the body middle ring in the Stentor
model (7). The twisting motion often caused by micromovements and circulation can lead to
wear and finally rupture of the sutures of the body of the frame. Of the 14 suture disruptions
that were identified at our institution, 5 were row separations and 9 occurred in the body,
corroborating Riepe's hypothesis of increased motion in the body of the stent.

Fracture of sutures linking individual stents may also play a role in secondary graft
failure. As more sutures break along the length of a graft, the longitudinal strength of the
graft is decreased, causing instability of the stent graft as a whole. This results in kinking
and morphological changes in the stent graft that can ultimately lead to separation from
the aorta and device migration (7).

IV. ETIOLOGY OF FAILURE

Device fatigue is not the result of one single event. Most likely a combination of causes

results in ultimate device fatigue. Besides intrinsic properties of the materials used for

stent-graft fabrication — i.e., strength of a metal, corrosion resistance, or flexibility of a

fabric — and the extrinsic hostile environment of the native aorta, other factors can play a tj

role in the ultimate fatigue/failure of a device. |

A. Endoleaks .g>

Endoleaks and their clinical application have been discussed at length in the literature (49- <

54). In brief, there are four types of endoleaks that have been described. A type I endoleak >9

involves leakage at the stentgraft attachment sites, either the distal or proximal end. A type J

II endoleak results from retrograde filling of the sac from arteries arising from the °

aneurysm. A type III endoleak is a disruption of the stent-graft device either from graft |

tears or modular separation, and a type IV endoleak refers to leakage from a porous stent @

graft. It seems obvious that increased pressure of the aneurysmal sac by lack of complete %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

DEVICE FAILURE

653

exclusion could potentially be a cause of rupture; however, the interaction of endoleaks
and device failure is less clear.

In the literature, endoleaks are reported to occur in 10-44% of endovascular repairs
(50). At our institution, the overall endoleak rate is approximately 35%. In our study
population, 24 (40%) of the 60 patients with stent-graft device fatigue were diagnosed with
an endoleak at some point after graft implantation (Table 5) Device fatigue was identified
prior to endoleak in 5 patients (21%). One patient had proximal and distal metallic
fractures in a first generation EVT stent graft and the second had a row separation
secondary to suture fracture of a Vanguard device, which later developed a type I endoleak
from dislodgment of the distal left limb. The third patient was noted to have a body
separation of a Vanguard device at his 17 month x-ray, and 7 months later was noted to
have a small endoleak. One year later, at his 36-month follow-up, an angiogram was done,
which noted distal and proximal dislodgement as well as an increase in aneurysmal size,
and the patient was treated with a Talent stentgraft. The fourth patient was noted to have
a body separation at his 3-month follow-up. Nine months later, a CT scan revealed a type
I distal endoleak and enlargement of the aneurysmal sac. The patient was treated with an
endovascular limb with resolution of the endoleak. The last patient was noted at 6-month
follow-up to have a stent fracture of the third row of metallic stents in the AneuRx stent
graft and 18 months later developed a type II endoleak. In 4 of the 5 cases where an
endoleak developed after identification of device fatigue the 2 events were unrelated (36).

Table 5 Incidence of Endoleaks in Grafts with Device Fatigue

Device

Number of
endoleaks

Time of diagnosis in relation to fatigue

Before

After

Same time

Management

AAA
Vanguard

Talent

EVT
MEGS

1
1

AneuRx

2

TAA

Gore

2

Talent

2

Endovascular

repair (5)
Open repair (2)
Observation (2)
Embolization of type

II endoleak (2)
Open repair (3)
Resolved

spontaneously (2)
Observation (1)
MEGS A-I-F (1)
Endovascular

repair (1)
Endovascular

repair (1)
Observation (1)

Open repair (1)
Observation (1)
Refused surgery, died

from rupture (1)
Observation (1)

I
•5

Source: Ref. 36. Copyright© Elsevier, Inc. Reprinted with permission.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

654 JACOBS et al.

However; the contrary may not be true. Eight patients (33%) developed stent-graft
fatigue after the endoleak was diagnosed, and 11 fatigued grafts (46%) were identified at
the same time as the endoleak (36). Endoleaks have been shown to maintain pressure and
turbulent flow in the aneurysmal sac, leading to eventual enlargement of the aneurysm and
risk of rupture (3,49). Residual aneurysmal sac pressure may be communicated to the stent
graft itself in the form of increased device pulsatility, which can lead to metal fracture,
suture disruption, and fabric wear, resulting in material failure. Although type I and III
endoleaks are usually treated immediately to prevent aneurysmal expansion and rupture,
type II endoleaks are more common and are usually followed (55,56). In our study, six of
the endoleaks that were diagnosed prior to device fatigue were type II endoleaks. All six
were observed on average 20 months (range 6-32) prior to diagnosis of the stent-graft
fracture. One can hypothesize that although a type II endoleak may not generate enough
pressure to cause aneurysmal dilatation and eventual rupture, there is enough pulsatile
flow to expose the stent graft to excess micromotion, which, in turn, could lead to device
failure in metal fracture, suture disruption, or fabric wear.

Several authors have shown that even aneurysms that do not show evidence of
endoleak still have pressure transmitted through the mural thrombus or the aortic wall,
which may contribute to aneurysmal rupture — a concept known as endotension (57,58).
This transmitted pressure can also be responsible for micromotion between the stentgraft
components, thus increasing the likelihood of abrasion, corrosion, or fabric wear.

V. CLINICAL SIGNIFICANCE

Although fractures were found in 60 patients with endovascular stent grafts, the majority
have been asymptomatic and have not required secondary intervention. It appears that
endoleaks and aneurysm size are more important than material failure in determining the
risk of rupture. While three patients developed symptomatic aneurysms in the setting of
stent fractures, these events did not appear to be related to the fatigued stents. The one
device related death was in a patient who had an immediate type III leak (graft hole) after
implantation and died after open conversion as a consequence of that leak. Since concern
regarding stent erosion through the fabric of stent grafts remains, we recommend that if a
patient is asymptomatic and there is no evidence of aneurysm enlargement, rupture, or a
type I or III endoleak, observation of the device for fatigue is acceptable in the setting of
increased graft surveillance.

VI. CONCLUSION

Endovascular grafts have been used clinically to treat aortic aneurysms for over 10 years.

A significant growth in our understanding of the failure modes of the materials used to ■§

fabricate these devices as well as respect for the relatively harsh environment into which g

they must function has led to the development of improved stent-graft designs. However, »

these devices may still experience metal fatigue, fabric fatigue, and ultimately device c

fatigue. Improved manufacturing practices can decrease the frequency of these events; <

however, the clinical implications of many forms of device fatigue remains to be defined. &

I

REFERENCES 1

Parodi JC, Palmaz JC, Barone HD. Transfemoral intraluminal graft implantation for abdom-
inal aortic aneurysms. Ann Vase Surg 1991; 5:491^499.

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I

DEVICE FAILURE 655

2. Becquemin JP, Lapie V, Favre JP, Rousseau H. Mid-term results of a second generation
bifurcated endovascular graft for abdominal aneurysm repair: The French Vanguard trial. J
Vase Surg 1999; 30:209-218.

3. Zarins CK, White RA, Hodgson KJ, Schwarten D, Fogarty TJ. Endoleak as a predictor of
outcome after endovascular aneurysm repair: AneurRx multicenter clinical trial. J Vase Surg
2000; 32:90-107.

4. Zarins CK, White RA, Moll FL, Crabtree T, Bloch DA, Hodgdon KJ, et al. The AneuRx stent
graft: Four-year results and worldwide experience 2000. J Vase Surg 2001; 33:S135-S145.

5. Bush RL, Lumsden AB, Dodson TF, Salam AA, Weiss VJ, Smith RB, et al. Mid-term results
after endovascular repair of the abdominal aortic aneurysm. J Vase Surg 2001; 33:S70-S76.

6. Beebe HG, Cronenwett JL, Katzen BT, Brewster DC, Green RM. Results of an aortic endo-
graft trial: Impact of device failure beyond 12 months. J Vase Surg 2001; 33:S55-S63.

7. Riepe G, Heilberger P, Umschield T, Chakfe N, Raithel D, Stelter W, Morlock M, Kretz JG,
Schroder A, Imig H. Frame dislocation of body middle rigs in endovascular stent tube grafts.
Eur J Vase Endovasc Surg 1999; 17:28-34.

8. Bohm T, Soldner J, Rott A, Kaiser WA. Perigraft leak of an aortic stent graft due to material
fatigue. AJR 1999; 172:1355-1357.

9. Norgren L, Jernby B, Engellau L. Aortoenteric fistula caused by a ruptured stent-graft: A case
report. J Endovasc Surg 1998; 5:269-272.

10. Maleux G, Rousseau H, Otal P, Colombier D, Glock Y, Joffre F. Modular component
separation and reperfusion of abdominal aortic aneurysm sac after endovascular repair of the
abdominal aortic aneurysm: A case report. J Vase Surg 1998; 28:349-352.

11. Holzenbein TJ, Kretschmer G, Thrunher S, Schoder M, Aslim E, Lammer J, Polterauer P.
Midterm durability of abdominal aortic aneurysm endograft repair: A word of caution. J Vase
Surg 2001; 33:S46-S54.

12. Parodi JC, Marin ML, Veith FJ. Transfemoral, endovascular stented graft repair of an ab-
dominal aortic aneurysm. Arch Surg 1995; 130:549-552.

13. Marin ML, Hollier LH, Avrahami R, Parsons R. Varying strategies for endovascular repair of
abdominal and iliac artery aneurysms. Surg Clin North Am 1998; 78(4):63 1-645.

14. Ohki T, Veith FJ, Sanchez LA, Marin ML, Cynamon J, Parodi JC. Varying strategies and
devices for endovascular repair of abdominal aortic aneurysms. Semin Vase Surg 1997; 10(4):
242-256.

15. Moore WS. The EVT tube and bifurcated endograft systems: Technical considerations and
clinical summary. J Endovasc Surg 1997; 4:182-194.

16. Dake MD, Miller C, Semba CP, Mitchell S, Walker PJ, Liddell RP. Transluminal placement of
endovascular stent-grafts for the treatment of descending thoracic aortic aneurysm. N Engl J
Med 1994; 331:1729-1734.

17. Temudom T, D'Ayala M, Marin ML, Hollier LH, Parsons R, Teodorescu V, Mitty H, Ahn J,
Falk A, Kahn R, Griepp R. Endovascular grafts in the treatment of thoracic aortic aneurysms
and pseudoaneurysms. Ann Vase Surg 2000; 14:230-238.

18. Parodi JC, Barone A, Piraino R, Schonholz C. Endovascular treatment of abdominal aortic

aneurysms: Lessons learned. J Endovasc Surg 1997; 4:102-110. E

19. Chuter TA, Green RM, Ouriel K, Fiore WM, Deweese JA. Transfemoral aortic graft place- a
ment. J Vase Surg 1993; 18:185-197. |

20. Faries PL. Endovascular grafts for the treatment of abdominal aortic aneurysm. In: Marin a
ML, Hollier LH, eds. Endovascular Grafting: Advanced Treatment for Vascular Disease. j
Armonk, NY: Futura, 2000. "5

21. Allen RC, White RA, Zarins CK, Fogarty TJ. What are the characteristics of the ideal endo- 3
vascular graft for abdominal aortic aneurysm exclusion? J Endovasc Surg 1997; 4:195-202. 2

22. Chuter TAM. Stent-graft design: The good, bad and the ugly. Cardiovasc Surg 2002; 10:7-13. I

23. Broeders IAMJ, Blankensteijn JD, Wever JJ, Eikelboom BC. Mid-term fixation stability of the ©
endovascular technologies endograft. Eur J Vase Endovasc Surg 1999; 18:300-307. %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

656 JACOBS et al.

24. Moore WS, Rutherford RB. Transfemoral endovascular repair of abdominal aortic aneurysm:
Results of the North American EVT phase 1 trial. J Vase Surg 1996; 23:543-553.

25. Mialhe C, Amicabile C, Becquemin JP. Endovascular treatment of infrarenal abdominal an-
eurysms by the Stentor system: Preliminary results of 79 cases. J Vase Surg 1997; 26:199-209.

26. May J, White GH, Harris JP. Devices for aortic aneurysm repair. Surg Clin North Am 1999;
79(3):507-527.

27. Criado FJ, Wilson EP, Fairman RM, Abdul-Khoudoud O, Wellons E. Update on the Talent
aortic stent-graft: A preliminary report from United States phase I and II trials. J Vase Surg
2001; 33:S146-S149.

28. Brener BJ, Faries P, Connelly T, Sefranek V, Hertz S, Kirksey L, Hollier L, Marin M. An in
situ adjustable endovascular graft for the treatment of abdominal aortic aneurysms. J Vase
Surg 2002; 35:114-119.

29. Mitchell RS, Miller DC, Dake MD. Stent graft repair of thoracic aortic aneurysms. Semin Vase
Surg 1997; 10:257-271.

30. Mitchell RS, Dake MD, Semba CP, Fogarty TJ, Zarins CK, Liddell RP, et al. Endovascular
stent graft repair of thoracic aortic aneurysms. J Thorac Cardiovasc Surg 1996; 11:1054-1062.

31. D'Ayala M. Endovascular grafting for thoracic aortic aneurysms. In: Marin ML, Hollier LH,
eds. Endovascular Grafting: Advance Treatment for Vascular Disease. Armonk, NY: Futura,
2000.

32. Blum U, Voshage G, Lammer J, Beyersdorf F, Tollner D, Kretschmer G, Spillner G,
Polterauer P, Nagel G, Holzenbein T, Thurnher S, Langer M. Endoluminal stent-grafts for
infrarenal aortic aneurysms. N Engl J Med 1997; 336:13-20.

33. Breek JC, Hamming JF, Lohle PNM, Lampmann LEH, Van Berge Henegouwen DP. Spon-
taneous perforation of an aortic endoprosthesis. Eur J Vase Endovasc Surg 1999; 18:174-175.

34. Najibi S, Steinberg J, Katzen BT, Zemel G, Lin PH, Weiss VJ, Lumsden AB, Chaikof EL.
Detection of isolated hook fractures 36 months after implantation of the Ancure endograft: A
cautionary note. J Vase Surg 2001; 34:353-356.

35. Heintz C, Riepe G, Birken L, Kaiser E, Chakfe N, Morlock M, Delling G, Imig H. Corroded
nitinol wires in explanted aortic endografts: An important mechanism of failure? J Endovasc
Ther 2001; 8:248-253.

36. Jacobs TS, Won J, Gravereaux EC, Faries PL, Morrissey N, Teodorescu VJ, Hollier LH,
Marin ML. Mechanical failure of prosthetic human implants: A ten-year experience with aortic
stent graft devices. J Vase Surg 2003; 37:16-26.

37. Guidoin R, Marios Y, Douville Y, King MW, Castonguay M, Traore A, Formichi M, Staxrud
LE, Norgen L, Bergeron P, Becquemin JP, Egana JM, Harris PL. First generation aortic
endografts: Analysis of explanted Stentor devices from the EUROSTAR registry. J Endovasc
Ther 2000; 7:105-122.

38. Trepanier C, Leung TK, Tabrizian M, Yahia L, Bienvenu JG, Tanguay JF, Piron DL, Bilo-
deau L. Preliminary investigations of the effects of surface treatments biological response to
shape memory NiTi stents. J Biomed Mater Res (Appl Biomater) 1999; 48:165-171.

39. Starosvetsky E, Gotman I. Corrosion behavior of titanium nitride coated Ni-Ti shape memory
surgical alloy. Biomaterials 2001; 22:1853-1859.

40. Trepanier C, Tabrizian M, Yahia L, Bilodeau L, Piron DL. Effect of modification of oxide layer j>
on NiTi stent corrosion resistance. J Biomed Mater Res (Appl Biomater) 1998; 43:433-440. <S

41. Duerig TW, Pelton AR, Stockel D. An overview of nitinol medical applications. Mater Sci Eng js
A273-A275 1999; 5:149-160. f

42. Harris P, Brennan J, Martin J, Gould D, Bakaran A, Gilling-Smith G, Buth J, Gevers E, White <.

... "

D. Longitudinal aneurysm shrinkage following endovascular aortic: A source of intermediate «

and late complications. J Endovasc Surg 1999; 6:11-16. jjj

43. Umschied T, Stelter WJ. Time-related alterations in shape, position, and structure of self- q

expanding modular aortic stent-grafts: A 4 year single center follow-up. J Endovasc Surg 1999; |

6:17-32. S

©

I

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

DEVICE FAILURE 657

44. Teoh SH. Fatigue of biomaterials: A review. Int J Fatigue 2000; 22:825-837.

45. McKelvey AL, Ritchie RO. Fatigue-crack propagation in nitinol, a shape-memory and su-
perelastic endovascular stent material. J Biomed Mater Res 1999; 47:301-308.

46. Riepe G, Loos J, Imig H, Schroder A, Schneider E, Peterman J, et al. Long-term in vivo al-
terations of polyester vascular grafts in humans. Eur J Vase Endovasc Surg 1997; 13(6):540-
548.

47. Alimi YS, Chakfe N, Rivoal E, Slimane KK, Valerio N, Riepe G, Kretz JG, Juhan C. Rupture
of an abdominal aortic aneurysm after endovascular graft placement and aneurysm size re-
duction. J Vase Surg 1998; 28:178-183.

48. Device Alert. Vanguard endoprosthesis. Upper stent row separation with or without nitinol
wire fracture 2001. Available from: http://www. medical-devices. gov. uk/da2001(01). htm.

49. Wain RA, Marin ML, Ohki T, Sanchez LA, Lyon RT, Rozenblit A, Suggs WD, Yuan JG,
Veith FJ. Endoleaks after endovascular graft treatment of aortic aneurysms: Classification, risk
factors, and outcome. J Vase Surg 1998; 27:69-80.

50. White GH, Yu W, May J, Chaufour X, Stephens MS. Endoleak as a complication of
endoluminal grafting of abdominal aortic aneurysms: Classification, incidence, diagnosis and
management. J Endovasc Surg 1997; 4:152-168.

51. Schurink GWH, Aarts NJM, van Baalen JM, Shultze Kool LJ, van Bockel JH. Experimental
study of the influence of endoleak size on pressure in the aneurysm sac and the consequences of
thrombosis. Br J Surg 2000; 87:71-78.

52. White GH, May J, Petrasek P, Waugh R, Stephen M, Harris J. Endotension: An explanation
for continued AAA growth after successful endoluminal repair. J Endovasc Surg 1999; 6:308-
315.

53. Schurink GWH, Aarts NJM, Wilde J, van Baalen JM, Chuter TAM, Schultze Kool LJ, van
Bockel JH. Endoleakage after stent-graft treatment of abdominal aneurysm: Implications on
pressure and imaging — An in vitro study. J Vase Surg 1998; 28:234-241.

54. Parodi JC, Berguer R, Ferreira LM, La Mura R, Schermerhorn ML. Intra-aneurysmal
pressure after incomplete endovascular exclusion. J Vase Surg 2001; 33:909-914.

55. Deaton DH, Makaroun MS, Fairman RM. Endoleak: Predictive value for anueryms growth at
3 years. Ann Vase Surg 2002; 16:37-42.

56. Resch F, Ivancev K, Lindh M, Nyman U, Brunkwall J, Malina M, Lindblad B. Persistant
collateral perfusion of abdominal aortic aneurysm after endovascular repair does not lead to
progressive change in aneurysm diameter. J Vase Surg 1998; 28:242-249.

57. Gilling-Smith G, Brennan J, Harris P, Bakaran A, Gould D, McWilliams R. Endotension after
edovascular aneurysm repair: Definition, classification, and strategies for surveillance and
intervention. J Endovasc Surg 1999; 6:305-307.

58. Kato N, Shimono F, Hirano F, Mizumoto F, Suzuki F, Ishida M, Fujii H, Yada I, Fakeda K.
Aneurysm expansion after stent-graft placement in the absence of endoleak. J Vase Intervent
Radiol 2002; 13:321-326.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

41

Endoleak

Hugh G. Beebe

Jobst Vascular Center, Toledo, Ohio, and Dartmouth-Hitchcock Medical Center,
Hanover, New Hampshire, U.S.A.

I. INTRODUCTION

The understanding of endoleaks and their management is critically important in the endo-
vascular repair of abdominal aortic aneurysm (AAA). The goal of all elective AAA
treatments is to protect the patient from the life-threatening potential of aneurysm rupture
and resultant hemorrhage. Thus an endovascular prosthesis can eliminate the potential for
such hemorrhage if it succeeds in achieving total exclusion of circulation within the aneu-
rysm. This dichotomous view of success, either the AAA has no circulation within the sac
or it has circulation, was held to be the primary criterion of treatment success in the early
days of endografts.

However, another way of thinking about endograft success derives from the accumu-
lating experience of the past decade. Restated, it can be said that the goal of all elective
AAA treatments is to provide the best prevention of aneurysm-related death. This is a
more sophisticated concept that involves including risk of the treatment itself and
inclusion of a range of late problems that may occur with either conventional open sur-
gery or endografts. Viewing the goal of endovascular AAA treatment in this way dimin-
ishes the impact of endoleak per se on evaluating results. This view seems appropriate now ■§
because of the still emerging understanding of different types of endoleaks with unequal s
potential effects, differences among device types in relation to endoleak and the complex »
influence of imaging methods used to identify them. c

Some experts have the opinion that endoleak is a poor predictor of aneurysm growth <

even though it is statistically associated with AAA enlargement. Absence of endoleak is >9

not a reliable predictor of aneurysm shrinkage since many apparently excluded AAA re- J

main unchanged in size after stent grafting. It has recently been stated that while endoleak «

is a risk factor for aneurysm enlargement, it cannot be used as an endpoint for effective |

endovascular aneurysm treatment (1). @

659

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

660 BEEBE

This chapter summarizes what is presently understood about endoleaks, emphasizes
prevention or avoidance, and discusses endoleak management. It also indicates where
incomplete understanding and controversial issues remain. Endovascular aneurysm treat-
ment is still a young endeavor having come into general use less than a decade ago, and it
is expected that added experience will produce changes that clarify and improve endoleak
prevention and management.

II. ENDOLEAK TYPES

The term endoleak was only first introduced in 1996, when the leadership group in Sidney,
Australia, at the Royal Prince Alfred Hospital, wrote a letter to the editor suggesting its
use to differentiate this type of AAA exclusion failure from hemorrhage due to AAA
rupture (2). At that time, endoleaks were still being lumped together regardless of cause,
but not long thereafter, the same group suggested differentiating endoleaks in two
publications that formed the basis for our current classification scheme (3,4). At the same
time, opinion articles appeared that opened the discussion about the clinical implication of
the various types of endoleak (5). The following classification combines a description of
endoleak types and summarizes their significance or potential significance in clinical terms.

A. Type I

This endoleak is defined simply as flow into the aneurysm sac around either the proximal
attachment zone or distal attachment, whether a straight graft in the aorta or either of two
limb attachments within the iliac arteries. Proximal type I endoleak is a clear failure of
endografting with implications for ongoing AAA rupture risk. The systolic blood pressure
in the sac from this type of endoleak is always assumed to be equal to systemic levels.
Distal type I endoleak is perhaps less threatening to the success of endograft treatment in
some special cases, but not clearly so. Recognition of type I endoleak during the operative
procedure requires immediate additional steps to eliminate it. Endoleak detection and
management are taken up below (Fig. 1).

B. Type II

This is the most complex and controversial endoleak type. The definition of type II is
circulation within the sac from aortoiliac anatomical branches. These most commonly are
lumbar and inferior mesenteric arteries, but type II endoleaks can also arise from accessory
renal arteries, horseshoe kidney arteries, internal iliac arteries, and occasional aberrant
pelvic branches of the iliolumbar, uterine, or middle sacral vessels.

It is probably going to be necessary in the future to subclassify type II endoleaks,
because including them all together in a single category has resulted in conflicting reports
about their behavior and significance. The ideal classification would be to stratify them
according to their physiology, but data on flow and pressure, presently requiring direct
measurement, are not often available. At present it seems useful to attempt to classify type .g>

II endoleaks according to whether they appear to have separate anatomical inflow and §

outflow tracts. It is probably true that lumbar branch endoleaks not associated with M

another vessel will usually occlude spontaneously (Fig. 2). M

Q

C. Type III 1

This is also a very serious, life-threatening type of endoleak that should prompt immediate ©

treatment. It is defined as a direct endoleak arising from loss of physical integrity of the li

I

s

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ENDOLEAK

661

Figure 1 Type I endoleak, shown both proximally and distally, occurs when there is inadequate
sealing at the stent-graft attachment zones.

endograft prosthesis. There are several ways in which this can arise. One way is through
perforation of the prosthesis fabric by an adjacent metallic stent. This is caused by relative
motion between these two parts of the endograft caused by a variety of influences that
include inherent looseness of the fabric, lack of fabric strength, and odd angulation of the
endograft usually associated with AAA sac shrinkage that presents an apex of stent metal
in direct contact with fabric. Another is through loss of attachment between modular
components such that they separate where an iliac limb contralateral to the main endograft
trunk has been inserted. This type of endoleak typically occurs late in follow-up after
endografting. It is for this reason that it appears to represent such a special danger to the
patient, as discussed below (Fig. 3).

D. Type IV

This is a somewhat vague type of endoleak that is device-specific in its occurrence and
unclear in its several implications. It is defined as a direct endoleak occurring through
porous fabric of the intact endoprosthesis. This kind of bleeding is commonly seen in
conventional open surgery, especially where knitted fabric grafts are used. The practice of
preclotting is a common remedy and various types of coated grafts that address this type
of problem are commercially available.

In endovascular therapy there is uniquely a requirement to use arteriography right
after endograft deployment to determine the procedure's success. Thus one significance of

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

662

BEEBE

Figure 2 Type II endoleak can arise from various branches of the aortoiliac segment, most com-
monly from lumbar arteries and the inferior mesenteric artery.

type IV endoleak is that it may make detection of a more serious endoleak unclear or lead
to wrong assumptions in interpreting completion arteriography during the endograft pro-
cedure. Some investigators have questioned whether such porous fabric might be capable
of transmitting pressure within the sac even after flow has ceased from thrombus forming
in the fabric interstices (Fig. 4).

E. Type V

This type of endoleak is not really an endoleak at all in the strict sense because it is not
associated with blood flow. Type V is defined as "endotension," a name proposed by Gil-
ling-Smith and his colleagues (6) to express the important concept that even though blood
flow might not exist, a thrombus is fully capable of transmitting pressure into the AAA sac.
There are both clinical and experimental data to support this concept, and the topic serves to
stimulate fundamental and skeptical reflection on how much we actually know about the
behavior of excluded AAA. The great variable in the literature addressing endotension is the
quality of images used to rule out endoleak in AAA that enlarge after stent grafting. Meier et
al. (7) found only 2 of 17 cases with AAA expansion among 658 receiving Ancure stent grafts
that did not show endoleak in a clinical trial setting with core lab review. They marshaled an
argument that the concept is flawed and only represents missed endoleak. Several authors
have reported AAA sac expansion following endografting with repeated lack of apparent
flow on follow-up contrast imaging (8-10). Zarins et al. (11) even reported post-stent-graft

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

I
•5

I

ENDOLEAK

663

Figure 3 Type III endoleak, dangerously associated with rapid aneurysm expansion and rupture,
is shown here caused by fabric erosion from physical contact with a metal stent (right) and by
disconnection of the left iliac limb of a modular bifurcation endograft.

rupture in five patients who did not show evidence of endoleak, as have others (12).
Experimentally, the Montefiore group demonstrated conclusively that endoleaks, produced
in a prosthetic in vivo model and subsequently excluded by coil embolization proven to stop
flow, still result in a pressurized sac at almost systemic levels (13).

III. DIAGNOSIS OF ENDOLEAK
A. Intraprocedural

Assessment of endoleak at the completion of an aortic stent graft procedure is almost ex-
clusively done by arteriography unless there is a strong contraindication to contrast use.
Completion arteriography should always involve use of a power injector of contrast, con-
current use of cine loop and frame-by-frame review, and — as some have recommended —
multiple arteriograms taken at several intervals along the endograft length (14). The use of
steep oblique views will sometimes add important information about endoleak source. A
slightly prolonged imaging run will allow the late filling of a type II endoleak by iliolumbar
branches that may occur several seconds after contrast has left the endograft (Fig. 5 A and B).
Unfortunately, in many institutions, the practice of completion arteriography is limited
to a single anteroposterior (AP) injection. The result of this may be the discovery on the first

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

664

BEEBE

Figure 4 Type IV endoleak, illustrated schematically here, through interstices of the graft fabric in
certain specific endograft types may make assignment of endoleak source difficult on completion
arteriography.

postoperative computed tomography (CT) images of a posterior midline type I endoleak,
which was probably present from the beginning and might have been resolved by adjunctive
measures at the time if more careful completion arteriography had identified it. If an
endograft procedure has been long and difficult, the combination of a tired team feeling the
lead apron's weight and a higher than usual cumulative fluoroscopy time may limit interest
in detailed imaging. But it is often the difficult case that is more liable to have the important
type I endoleak. A disciplined approach to completion arteriography will yield benefits. As
mentioned above, the presence of a type IV endoleak from porous fabric or other
construction details of certain stent graft types may produce confusing findings on
completion arteriography. Reversing heparin and waiting for an interval to repeat the
arteriogram may be helpful.

Some centers have investigated the use of intraoperative pressure monitoring by
placing small catheters into the AAA sac alongside the stent graft so as to be able to mea-

I
•5

Figure 5 A. This arteriogram was obtained at the end of a short run time upon completion of an
aortic endograft. The full length of the endograft and renal and mesenteric branches can be seen. No
endoleak is apparent. B. This arteriogram was obtained in the same patient shown in Figure 5A
using the same position and power injection of contrast but with a longer run time. The iliolumbar
artery (ILA) filling a type II endoleak is now apparent (arrows).

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOLEAK

665

B

I

•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

666 BEEBE

sure pressure changes upon completion of the procedure and, in some cases for a short
interval postoperatively, before withdrawing the catheter. Stelter and Umscheid (15) were
pioneers in this physiological approach to assessing whether a satisfactory endpoint of the
endograft procedure had been achieved and reported summary information on 201 pa-
tients. But they concluded that correlation between pressure changes observed during
stent-graft deployment and the "fate of the aneurysm over time" was not known. Thus the
use of pressure monitoring was considered to be an adjunct to endoleak detection by ar-
teriography. This type of monitoring was studied by Bell and colleagues (16), who found
that AAA sac pressure monitoring operatively and for 24 h postoperatively showed
evidence of higher pressures and pulsatility in 5 of 15 patients with endoleaks. However,
these endoleaks were also seen on arteriography. Given the uncertain meaning of the pres-
sure data, the likelihood that significant endoleak can be found with careful arteriography
and the theoretical added risk of complications such as infection, it would seem that pres-
sure monitoring, at least by this method, would not be advisable outside of a clinical trial.

B. Postoperative Follow-Up

The diagnosis of endoleak during follow-up is presently made exclusively through various
types of imaging, CT scanning most commonly, duplex ultrasound, arteriography when
the source is unclear or when a therapeutic attempt by endovascular occlusion is made,
and, far less commonly, by magnetic resonance arteriography.

C. CT Scan

The quality of imaging is a large variable in determining the presence of endoleak and
the incidence of its occurrence. CT scanning done with thick acquisition protocols that
allow collimation greater than 5 mm or with poor timing of contrast can fail to show
evidence of an endoleak that is found to be present on closely timed arteriography. The
ideal CT scan for endograft follow-up is done with 1:1.5 pitch, 3-5 mm collimation, and
timing to ensure that the 150 mL of intravenously injected contrast arrives in the
aortoiliac segment at the right time. This matter of the right time for detecting endoleak
is not quite the same as for preoperative evaluation of morphology. It has been observed
that delayed CT will show some endoleaks that are missed without an interval of time,
allowing low-flow volumes to be seen. Usually these are type II endoleaks, as types I and
III come directly from the aortic flow lumen. When the aorta is heavily calcified or
contains unusual calcified projections into the lumen, it may be important to obtain a CT
image acquisition before and after contrast injection, allowing one to distinguish between
calcified vascular tissue and endoleak. It may also be difficult on occasion to decide
about a small or subtle endoleak in the presence of prominent metallic flare artifact on
the CT image. The quality of CT scans for endograft follow-up requires high standards ■a

of excellence in technique and interpretation, as with any diagnostic test being used to |

rule out a potentially serious finding. a

Three-dimensional (3D) postprocessing of CT scans for follow-up assessment of aortic c

stent grafts is a large subject beyond the scope of this chapter. However, in discussing <

endoleak, one aspect ought to be mentioned — the use of volume measurement of the >9

infrarenal aorta. This procedure, from the lowermost renal artery to the aortic bifurcation, J

can be readily done using postprocessing of CT scan data, enabling serial comparison of «

an indicator of aneurysm size change that is more sensitive than diameter (B Kritpracha |

and H Beebe, unpublished data) (17). @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOLEAK

667

White and colleagues (18) demonstrated that significant variation in measuring AAA
size occurs with variation in CT scan technique and concluded that volumetric analysis is
very useful in the follow-up of patients whose AAA do not shrink. When the CT images
fail to show contrast within the sac but AAA volume is increasing, an undetected endoleak
should be suspected. This may present an indication for further diagnostic imaging by
arteriography. Another use of volume measurement from 3D postprocessing that may be
useful in following type II endoleaks is to measure the volume of the endoleak separately
from that of the AAA. Anecdotal evidence from our imaging laboratory suggests that
declining endoleak volume predicts eventual spontaneous closure (Fig. 6).

The assignment of endoleak type by CT scan is probably inaccurate in as many as
20% of cases, although well-controlled data on this matter are limited. In one report
specifically related to endoleak type, Parent et al. (19) detected the source artery in all of 36
patients with type II endoleak using color duplex ultrasound but in only 7 of the same
patient cohort using CT scan. There is a tendency for observers to make a value judgment
about endoleaks from CT scan appearance, but there are no physiological data to be
found on these images. Thus comments sometimes seen in published reports about
"major" or "minor" endoleaks based on the size of the contrast-filled space in the aorta
and its contrast density are wholly theoretical.

6.5 cc

I
•5

0.5 cc

Figure 6 A series of CT scans (top) showing evidence of endoleak (arrows) obtained over a total
span of 12 months. Below the CT scan are views of a 3D model made from the CT scan showing the
endoleak (open arrow). Serial calculations of the type II endoleak volume (numbers below model)
showed progressive decrease and the endoleak closed spontaneously. (Medical Media Systems, West
Lebanon, NH.)

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

668

BEEBE

D. Duplex Ultrasound Scanning

The appeal of color-flow duplex scanning for follow-up evaluation of stent grafts comes
from its low cost, patient acceptance, and risk-free repeatability. The drawbacks, however,
are significant. As with all ultrasound examinations, there is an influence of operator
variability, and this is heightened in abdominal studies. Additionally, body habitus,
intestinal gas, and the lower resolution of the required low-frequency transducers may
all adversely affect quality. There is also the need for patient preparation for the
examination, which constitutes a nuisance factor (Fig. 7).

McLafferty et al. (20) reported correlation between duplex and CT scans in 79 patients
after stent graft, including 7 with endoleak, and concluded that it was an accurate test.
They also summarized a complex literature from other centers evaluating the role of
duplex scanning. One of those reports from highly experienced French authors empha-
sized the accuracy of ultrasound in detecting endoleak in a study of 54 patients without
and 35 with endoleak by CT (21). There was poor correlation between ultrasound and CT
determination of AAA diameter. Wolf et al. (22) found discordant results in 8% of 100
endograft patients with 1 1 endoleaks seen on CT but not on duplex. They concluded that
duplex scanning, if of assured high quality, was comparable to CT scanning in practical
clinical value since those not seen by duplex scan were not associated with AAA
expansion. The general value of such a conclusion will need to be validated by additional
larger studies. The use of ultrasound contrast agents has been shown to enhance the value
of ultrasound in stent graft follow-up (23,24) (Fig. 8).

I
•5

Figure 7 A normal duplex ultrasound color-flow image showing the aneurysm sac (arrowheads)
and the stent graft (SG) within it. There was no color-flow evidence of endoleak throughout the
entire aortoiliac segment.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOLEAK

669

Figure 8 A color-flow ultrasound image, printed in gray scale, showing flow within stent graft
(SG) and a type II endoleak (E) elsewhere within the AAA sac.

E. Arteriography

Standard contrast arteriography is used for endoleak evaluation when the source is
unclear and a therapeutic maneuver is undertaken to stop the endoleak. Examples include
expanding AAA without demonstrated endoleak or after determining that an endoleak is
present but its type is not clear. The clinical setting is an important part of judging the
need for arteriography. If an AAA is stable in size and the endoleak is highly likely to be a
type II or an apparently small endoleak from a distal type I, observation without being
entirely certain of the source may be appropriate for an interval of time. However, when
an AAA is enlarging after stent-grafting and the endoleak source is presumed to be a type
II, an arteriogram to establish this conclusively is indicated and may provide an
opportunity for concurrent treatment as discussed below (Fig. 9A and B).

Technical details are of value in getting the most information from this type of
examination and most often involve thoughtful projection angles to reveal endoleak
origins and use of balloon occlusion to isolate or augment small sites of endoleak blood
flow. For example, Matsumura et al. (25) were able to demonstrate microleaks arising
from an endograft, that were not clearly shown on routine CT views by using balloon
occlusion arteriography.

F. Magnetic Resonance Arteriography (MRA)

This imaging method (MRA) varies quite widely in quality from one center to the next
largely because of variation in equipment, especially the software used to process magnetic
resonance images and local expertise in its use. However, gadolinium-enhanced MRA has
been shown to be of value in endoleak detection, and it does not confuse calcium with
contrast in the AAA sac, as occasionally happens in CT scans (26). MRA also offers

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

670

BEEBE

I
•5

B

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOLEAK 671

benefit in patients with mild chronic renal insufficiency who should not receive large
volumes of contrast for CT scanning.

G. Adjuncts to Imaging Methods

The usefulness of physical findings in endograft follow-up has seldom been emphasized. In
patients whose AAA can be palpated, detection of pulsatility by simple but skilled
examination should not be dismissed in this era of highly technical imaging. The finding
of a quiet AAA after stent-grafting is reassuring, and a change in later follow-up to an
AAA that has a definite pulsatile quality is a valid index of concern. Some very expe-
rienced clinicians have emphasized the importance of physical findings in deciding which
patients with endoleaks may require treatment (27).

An interesting and creative approach to endoleak detection was suggested by the work
of Serino and colleagues (28), who explored changes in D-dimer levels, a fibrin degradation
product, after endografting in a well-controlled study of 74 patients. They found highly
significantly increased D-dimer levels in those with type I endoleak and concluded that this
may prove to be a useful marker for fixation problems after endovascular AAA repair.

H. Pseudoendoleak

Endoleak evidence is mostly derived from imaging data. The appearance of an area of
radiodensity within the treated AAA sac that is outside the endoprostheses and of in-
creased density compared to the rest of the sac is a fair description of the criteria for
diagnosis of endoleak. However, there are at least three ways in which CT scans can meet
these criteria and yet not be associated with endoleak. One is specific to endografts con-
structed with the fabric on the outside of the stent. If the fabric bulges out or away from
the underlying stent during contrast CT, it can have the false appearance of endoleak. This
misleading finding is most often seen in follow-up images of the Vanguard stent graft
(Boston Scientific Corp., Natick MA), which is no longer used. However, many patients
with this endograft remain in follow-up. Another source of confusion may result from odd
calcifications in the AAA sac, sometimes quite far into the lumen and perhaps from old
areas of dissection. If the available images are not duplicated before and after contrast
injection, that additional step may be needed to resolve the question on subsequent exams.
Last, the persistence of contrast trapped in the AAA during the insertion of the endograft
has been reported as a source of pseudoendoleak (29).

IV. TIME OF ENDOLEAK OCCURENCE

If one considers all endoleaks regardless of type, they are most common in the immediate

postoperative period, as might be imagined. But if one considers persistent endoleak, a

more significant endpoint descriptor, then endoleaks continue to accumulate during §

follow-up. The reasons for this are many and include the timing and quality of imaging, |

characteristics of specific endograft types, morphology changes in excluded AAA, and 3

device structural failures. 5

Figure 9 A. This CT scan image clearly shows an endoleak, but does not identify the source which ^

remained unclear after reviewing the entire scan. B. The patient whose CT image was seen in Figure &

9A had an arteriogram with selective injections done to localize the endoleak source. As seen here, a ©

type II endoleak is proven to be arising from a focal structural defect in the endograft (arrow). %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

672 BEEBE

An example is the increasing endoleak problem during an endograft trial reported by
the Vanguard IDE investigators (30). These results were presented as presence of per-
sistent endoleak by follow-up interval. In that way, inspection of a life table provides the
trend of endoleak development or resolution and patterns that may be revealed by the
shape of the curve. It is also simplifies results reporting because of the problem of indi-
vidual patients whose endoleaks seal and recur.

The Eurostar registry, which collates data from 87 European centers, showed in-
creasing endoleaks over length of follow-up in data on 2463 patients who received stent
grafts of eight types with 171 (6.9%) endoleaks detected at 1 month and 317 occurring
later as new-onset endoleaks (31). The Eurostar analysis was primarily in the form of clin-
ical effects that could reasonably be associated with endoleak, such as increase in AAA
size, late rupture, and, less clearly, secondary intervention. These clinical endpoints were
correlated with endoleak status and defined by three groups: (a) type II endoleak, (b) type
I or type III endoleak, and (c) no direct imaging evidence of endoleak. Analysis of effect on
AAA growth showed surprising enlargement of AAA in group A, which caught up with
group B by 36 months. Unfortunately, the Eurostar analysis did not separate early and
late new-onset endoleaks.

Reports on type II endoleaks both early and late are conflicting in many respects.
Resch and colleagues (32) observed a decrease in AAA size in endoleak-free patients, but
those with type II endoleak showed no statistically significant change in aortic diameter on
follow-up, and this has been found repeatedly in other studies (33,34). Also, AAA with
type II endoleaks have been observed to shrink (35) and, on occasion, to expand and
rupture (36,37). Some authors have concluded that this variable behavior may be ex-
plained by time of occurrence in addition to the probable differences in blood pressure
between sources of type II endoleaks.

It has long been noted anecdotally that late-onset type III endoleaks usually cause
dramatically rapid AAA expansion, leading to speculation that this reflects atrophy of the
aortic wall during the interval of sac depressurization between stent-graft exclusion and
endoleak development. Wolf et al. (38,39) reported interesting observations on AAA
diameter in four patients with late-onset endoleak, some increasing by 16 mm over 20 days
and 10 mm over 10 days. Ohki and associates (40) reported the seriousness of eight type I
and III endoleaks occurring 10-55 months after stent-grafting that required prompt open
conversion or secondary stent graft. Carpenter et al. (41) found endoleak treatment to be
significantly the most common indication for readmission after endografting in a popu-
lation of 337 patients, with only a 71% readmission-free survival at one year.

In summary, review of diverse and often conflicting literature on the subject of
endoleak occurrence time suggests that type I endoleak is a serious threat and especially
dangerous if it is proximal or occurs late. Type II is not predictable in behavior and must be
observed for its effect on AAA size. Type III usually occurs late and suddenly, though not ■§

always. Its presence requires immediate intervention. Type IV should be restricted to |

periprocedural time interval by definition. Type V or endotension is discussed separately as

below but is best thought of in the same way as type II because it is a diagnosis of exclusion. c

V. AVOIDING ENDOLEAK f

Even though endoleak is of limited value as a criterion for judging endograft success, «

absolute exclusion of blood flow within an AAA is a desirable outcome and thus makes |

endoleak prevention a goal (42). The main focus of prevention has been case selection @

t

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOLEAK 673

based on anatomical factors. It should be realized that case selection is driven by
consideration of device limitations and therefore will evolve and change in response to
the characteristics of new stent-graft devices. Perhaps the influence of difference among
various stent-graft devices is the main reason for the lack of agreement in analyses of risk
factors for endoleak and a wide range of early postoperative occurrences.

Proximal type I endoleak, generally considered a clear failure of stent-grafting, has
usually been thought to be associated with adverse case selection involving anatomical
features of the proximal infrarenal attachment zone, such as severe angulation, short
length, conical shape, luminal thrombus, or dense calcification. Conflicting reports
disclose lack of agreement on all of these factors. Neck angulation was associated with
proximal type I endoleak in a report of results in 184 patients from Albertini et al. (43).
However, the Eurostar data from 2146 endografts showed no correlation with aortic neck
angulation (44). That same report did show that neck length was a highly significant
(p — 0.0001) risk factor for proximal type I endoleak. But another report from four
endograft centers showed no effect on endoleak of neck length less than 10 mm in 55
patients who received suprarenal stent grafts, thus suggesting the possible influence of
device characteristics (45). Petrik and Moore (46) studied 100 endograft patients with a
surprisingly high 44% incidence of endoleak and concluded that they were unable to
demonstrate endoleak predictive value to anatomical factors that included neck angula-
tion, neck thrombus or calcification, and number of patent lumbar arteries or the inferior
mesenteric artery. They did not examine neck length, however.

Type II endoleaks have generally been thought to be predicted by a larger number of
patent lumbar arteries, particularly when associated with a patent inferior mesenteric
artery, but not all reports have found that correlation (47,48). This led to investigation of
preventive branch embolization prior to endografting, which proved unsuccessful in a
study involving 25 of 72 patients but illustrated the difficulties of this technically involved
method (49). Another approach has been to prevent type II endoleak using insertion of
thrombogenic material into the sac. Walker and colleagues (50) injected an absorbable
gelatin sponge material into AAA sacs during stent-graft procedures when aortic side
branches were seen on arteriography. Forty-eight of 93 (52%) patients had the test
treatment. Excluding 15 patients with type I endoleaks, no patient treated with the sponges
or those who did not show patent branch vessels and did not receive sponge injection had
evidence of endoleak on median follow-up of 4 months. This "broad brushstrokes"
approach to eliminating type II endoleaks has much theoretical appeal because it could
form the basis for a selective approach to follow-up that might simplify the process, make
duplex scan more useful, and reduce the overall cost of surveillance. Type III and IV
endoleaks are fundamentally a design or device characteristics problem that presents
management challenges for clinicians. These complications may be preventable by better
device selection when problems become evident in retrospective analyses.

The complexity of endoleaks is increased by including type II endoleaks together with all
other types in some reports of endograft outcomes. In the analysis of endoleak significance
provided by Zarins et al. (34), the conclusion was reached that "The presence or absence
of endoleak on CT scan before hospital discharge does not appear to predict patient sur-
vival or aneurysm rupture rate after endovascular aneurysm repair." Many theoretical
reasons could be advanced to account for this surprising result: a relatively small study

s

VI. MANAGEMENT OF ENDOLEAKS |

'C

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

674 BEEBE

cohort, a short follow-up interval (only 61% had CT images available for analysis at 1
year), differences in follow-up imaging protocols among institutions, and the definition of
primary success endpoints that also had an aggregating effect. The understanding of
whether different types of endoleak are also different in their implications for the patient is
unlikely to be improved by treating them equally.

Among experienced clinicians, consensus was recently expressed that "type I and type
III endoleaks will have serious consequences, even if sealing appears to have occurred"
(51). They also believed that based on current data, most but certainly not all type II
endoleaks appear to have a benign course; they frequently seal but probably should be
separated into different types. It would be of obvious value to identify type II endoleaks
destined to produce continued AAA enlargement. But current methods depend on
anatomical differences as a surrogate for direct information about physiological param-
eters such as peak pressure, pulse pressure, and mean pressure.

The proposed anatomical basis for classifying type II endoleaks is to separate those
with only a single vessel source (no outflow and somewhat similar conceptually to pseu-
doaneurysm) from those with inflow and outflow. The implications of this for the im-
portance of pressure levels and flow volumes is unclear, but many believe that type II
endoleaks with named vessel inflow and outflow are more likely to result in sac enlarge-
ment than the common "lumbar only" endoleak that often disappears. This does not align
well with the observations made by several investigators that small endoleaks with only an
inflow source may produce high pressures within the AAA sac. Schurink et al. (52) found
that even a small endoleak causes considerable pressure in the aneurysm sac, and this was
independent of endoleak size. In their in vitro experiment, systolic pressure was consid-
erably reduced, but diastolic pressure was similar to that in the sac or even higher than in
the systemic circulation because thrombus could cause a valve phenomenon. Parodi and
colleagues (53) used an in vitro model to provide data of compelling interest by showing
that small endoleaks could produce sac pressures higher than systemic levels but also went
on to show the effect of providing an outflow vessel. When both inflow and outflow were
simulated, there was a flow volume-related fall in pressure, and mean sac pressure became
systemically lower than systemic.

It would be desirable to base the management of endoleak on observed physiological
data, and perhaps that goal can be achieved when implantable pressure-sensing devices are
available for guidance. But presently the exercise of clinical judgment about whether to
intervene and what to do is based on anatomical and device imaging.

A. Type I

Most practitioners of experience act to remedy a proximal type I endoleak as soon as it is

observed, and its presence is a clear indication for intervention. When the endoleak is ■$

recognized during the primary procedure, a variety of methods from forceful balloon |

dilation to insertion of additional stents or stent-graft extensions have been successfully s

employed (54,55). The use of large-diameter balloon-expanded stents, such as the Palmaz c

(Cordis Endovascular, Warren, NJ), has proven to be a successful adjunct to close a type I <!

endoleak and is probably used more often than the few reports in the literature suggest (56). J3

Short-length large-diameter stent grafts, so-called extender cuffs, have been provided by J

device manufacturers and are frequently used to resolve proximal type I endoleaks. It has a

been observed, however, that short-term success does not always yield long-term success. A |

well-documented example of proximal stent extension failure serves to heighten awareness g

that no controlled observations on late results of proximal cuff extensions have yet ■£

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOLEAK 675

appeared (39). Disturbing observational evidence of proximal extender cuff use to manage
primary endograft migration with failure in three of six cases has been reported (57).

The primary onset of a late proximal type I endoleak almost always is associated with
stent-graft migration. The cause of the migration should be sought, since this will be the
primary determinant of appropriate treatment. Three general causes for late migration are
(a) enlargement of the aortic attachment zone; (b) traction on the upper endograft by a
change in AAA morphology, causing severe angulation; and (c) loss of frictional attach-
ment from changes in the device such as stent fracture.

Distal type I endoleak — nowadays virtually limited to iliac limbs, since straight aortic
endograft use is rare — is approached in a similar way with use of increasingly involved
measures from simple balloon angioplasty to stents and stent-graft extensions. The blood
pressure implications of any endoleak communicating directly with the aortic lumen are
important, but the decision to manage a small distal type I endoleak that persists after
various measures, initially by observation, may sometimes be appropriate depending on
the procedural circumstances.

B. Type II

Assigning an endoleak observed by CT the classification of type II is sometimes found to be
wrong on subsequent arteriography. Type II endoleak that appears to be limited to lumbar
artery branches does not represent a proven threat to endograft treatment success unless
later follow-up shows further AAA enlargement. Therefore, most experienced clinicians
who elect to observe an endoleak thought to be type II at the completion of the endograft
procedure use routine follow-up imaging, anticipating occlusion without further interven-
tion. However, an exception to this may exist in the growing use of endografts for emergency
treatment of ruptured AAA. The presence of hemorrhage presents a blood loss problem that
is separate from the small potential for late AAA expansion related to type II endoleak.

When the source includes inflow and outflow vessels, a direct approach by endovas-
cular catheterization with insertion of coils to induce thrombosis is often used and has been
well described by Baum and colleagues (58) at the University of Pennsylvania who later
changed their approach. After observing late endoleak recurrence, they adopted a direct
translumbar approach, which has been more successful. Improved late results of endoleak
occlusion in 33 cases of proven type II endoleak but without data on AAA sac size change
allowed them to conclude that transarterial (in contrast to direct translumbar) coil
embolization was "ineffective and should not be performed" (59). This approach must
be confirmed and accompanied by data on late AAA size effects. Also, it has not been
compared with laparoscopic clipping of endoleak source vessels, a technique described only
in anecdotal reports. This report of late failure of transarterial coil insertion correlates with
the observations made experimentally by Marty and colleagues (60) at Montefiore Medical •g

Center, who showed that proven coil occlusion of experimental endoleaks in a canine model §

failed to reduce sac pressure. They concluded that coil embolization failed to interrupt |

pressure transmission to the aneurysmal wall and therefore may be not a reliable manage- c

ment option for endoleaks. It may also be true that the material used for branch vessel <!

occlusion is an important variable and that coils are not the best for this purpose. Martin et J3

al. (61) used a liquid embolic agent known as Onyx in six patents with documented J

occlusion and early modest AAA sac shrinkage in five. The injection of thrombin into Q

lumbar branches has been discouraged because of neurological complication risk when this §

compound extends into remote blood vessels (59). The need for chronic anticoagulation g

with warfarin has been shown not to influence the occurrence of type II endoleak but to ■£

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

676

BEEBE

B

I

•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOLEAK 677

interfere with endoleak resolution and AAA shrinkage (62). Thus, if stopping antico-
agulation is an option, at least for an interval of time, it should be considered.

C. Type III

The loss of endograft integrity usually occurs late and is a serious matter. The dis-
connection of modular components of a bifurcation endograft is usually due to either
operator error in placing the limb into the aortic trunk for too short a distance or when
morphology change puts physical stress on the junction of the two parts of the endograft.
Often this can be remedied by getting a guidewire across the space and into both distracted
components and using this as a basis for insertion of an additional stent graft (63). But
angulation and thrombosis may make this impossible. Alternatives to consider include
maneuvers to achieve an aorto-uni-iliac configuration together with femorofemoral bypass
and contralateral occlusion of the disconnected limb or direct open repair using the
proximal portion of the endograft as inflow to the disconnected side. Employment of a
preplaced balloon through the patent intact limb facilitates this procedure. Other type III
endoleaks arising from a stent graft in good position but with loss of fabric integrity due to
stent perforation have been treated successfully by insertion of an entirely new endograft
within the failed one (Fig. 10A and B).

Although endovascular repair of late endoleaks is usually possible, one should review
whether the original decision to attempt endograft repair of an aneurysm was appropriate
in the first place. If the patient's anatomical features are very adverse it may be that the
endograft failure cannot be treated by further endografting and is best managed by
conversion to conventional open surgery (64,65).

VII. A VIEW TOWARD THE FUTURE OF ENDOLEAK

Despite opinions to the contrary, most clinicians remain vigilant and concerned about
endoleak. Yet the practical difficulty and cost of endlessly pursuing a series of contrast CT
scans is recognized by all as undesirable. It seems easy to predict that alternative imaging
methods such as duplex ultrasound, perhaps together with plain abdominal x-rays, and
criteria for selecting which patients need more intensive imaging follow-up protocols will
be an active subject of clinical investigation in the near term.

It may be that use of computerized postprocessing yielding sophisticated 3D images
and sensitive measurements of volume, angulation, and diameter change that are auto-
matically processed and made available by Internet access with graphic plots of change
that are now becoming available (Medical Media Systems, West Lebanon, NH) will make
follow-up easier. A broad database resulting from that approach could be used to identify
categories that benefit from specific types of follow-up or early intervention to prevent clin-
ical events from endoleaks rather than waiting to react to them after the fact. 1

Investigation of various pressure-sensing devices to be inserted into the AAA sac at |

the time of stent-grafting or as a part of the endograft is being pursued. If these come into j§

•c

I

Figure 10 A. These two plain abdominal x-rays of an aortic straight endograft taken 12 months «*

apart demonstrate upward migration of the distal end of the endograft after expansion of the distal g

aortic attachment zone. The border of the L5 vertebral body is marked for reference (L5). A late ■%

occuring distal type I endoleak resulted. B. The patient was treated by a secondary endograft in J

a bifurcated configuration (arrows) inserted through the original straight endograft (arrowhead), ©

which resulted in sealing. %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

678 BEEBE

clinical application, validation work to understand the effects, if any, of variable thrombus
patterns within the sac over time will be important. Possibly a semiautomatic method of
evaluating endoleak status will result. However, in this regard, it would seem that such a
pressure-measuring device would be analogous to the smoke detector fire alarm that tells
you when a problem has arisen but does not anticipate its occurrence.

Adjunctive means of attaching an endoluminal prosthesis to the vascular wall have
been described, suggesting the possibility that methods such as endovascular stapling or
other physical attachments could be used in the future management of endoleak (66).

In an effort to simplify the concern over endoleaks of all types, it has been proposed
that a preemptive approach by injecting an agent that will solidify the AAA sac as an
adjunct to endografting would beg the question of endoleak significance (67). This in-
triguing idea should provide a stimulus to creative thinking but will require extensive work
to define materials and indications. Still, given the complex and unsettled nature of en-
doleak management today, there is great appeal to an approach that prevents them or pre-
vents their adverse effect.

ACKNOWLEDGMENT

The author gratefully acknowledges the medical illustrations contributed by Steven S.
Gale, M.D.

REFERENCES

1. Deaton DH, Makaroun MS, Fairman RM. Endoleak: Predictive value for aneurysm growth at
3 years. Ann Vase Surg 2002; 16:37-42.

2. White GH, Yu W, May J. Endoleak — A proposed new terminology to describe incomplete
aneurysm exclusion by an endoluminal graft. J Endovasc Surg 1996; 3:124-125.

3. White GH, May J, Waugh RC, Yu W. Type I and type II endoleaks: A more useful classification
for reporting results of endoluminal AAA repair. J Endovasc Surg 1998; 5:189-191.

4. White GH, May J, Waugh RC, Chaufour X, Yu W. Type III and type IV endoleak: Toward a
complete definition of blood flow in the sac after endoluminal AAA repair. J Endovasc Surg
1998; 5:305-309.

5. Beebe HG, Bernhard VM, Parodi JC, White GH. Leaks after endovascular therapy for an-
eurysm: Detection and classification. J Endovasc Surg 1996; 3:445^148.

6. Gilling-Smith G, Brennan J, Harris P, Bakran A, Gould D, McWilliams R. Endotension after
endovascular aneurysm repair: Definition, classification, and strategies for surveillance and
intervention. J Endovasc Surg 1999; 6:305-307.

7. Meier GH, Parker FM, Godziachvili V, Demasi RJ, Parent FN, Gayle RG Endotension after
endovascular aneurysm repair: The Ancure experience. J Vase Surg 2001; 34:421-426.

8. White GH, May J, Petrasek P, Waugh R, Stephen M, Harris J. Endotension: An explanation |
for continued AAA growth after successful endoluminal repair. J Endovasc Surg 1999; 6:308- |
315. |

9. Baum RA, Carpenter JP, Cope C, Golden MA, Velazquez OC, Neschis DG, Mitchell ME, jf
Barker CF, Fairman RM. Aneurysm sac pressure measurements after endovascular repair of <
abdominal aortic aneurysms. J Vase Surg 2001; 33:32^-1. a

10. Gilling-Smith GL, Martin J, Sudhindran S, Gould DA, McWilliams RG, Bakran A, Brennan J"
JA, Harris PL. Freedom from endoleak after endovascular aneurysm repair does not equal q
treatment success. Eur J Vase Endovasc Surg 2000; 19:421-425. 1

11. Zarins CK, White RA, Fogarty TJ. Aneurysm rupture after endovascular repair using the 2
AneuRx stent graft. J Vase Surg 2000; 31:960-970. 2

1
§

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOLEAK 679

12. Bade MA, Ohki T, Cynamon J, Veith FJ. Hypogastric artery aneurysm rupture after endo-
vascular graft exclusion with shrinkage of the aneurysm: Significance of endotension from a
"virtual," or thrombosed type II endoleak. J Vase Surg 2001; 33:1271-1274.

13. Mehta M, Ohki T, Veith FJ, Lipsitz EC. All sealed endoleaks are not the same: A treatment
strategy based on an ex- vivo analysis. Eur J Vase Endovasc Surg 2001; 21:541-544.

14. White GH, Yu W, May J, Chaufour X, Stephen MS. Endoleak as a complication of
endoluminal grafting of abdominal aortic aneurysms: Classification, incidence, diagnosis, and
management. J Endovasc Surg 1997; 4:152-168.

15. Stelter W, Umscheid T, Ziegler P. Three-year experience with modular stent-graft devices for
endovascular AAA treatment. J Endovasc Surg 1997; 4:362-369.

16. Treharne GD, Loftus IM, Thompson MM, Lennard N, Smith J, Fishwick G, Bell PR. Quality
control during endovascular aneurysm repair: Monitoring aneurysmal sac pressure and super-
ficial femoral artery flow velocity. J Endovasc Surg 1999; 6:239-245.

17. Wever JJ, Blankensteijn JD, Th M Mali WP, Eikelboom BC. Maximal aneurysm diameter
follow-up is inadequate after endovascular abdominal aortic aneurysm repair. Eur J Vase
Endovasc Surg 2000; 20:177-182.

18. White RA, Donayre CE, Walot I, Woody J, Kim N, Kopchok GE. Computed tomography
assessment of abdominal aortic aneurysm morphology after endograft exclusion. J Vase Surg
2001; 33(2 suppl):Sl-S10.

19. Parent FN, Meier GH, Godziachvili V, LeSar CJ, Parker FM, Carter KA, Gayle RG, De-
Masi RJ, Marcinczyk MJ, Gregory RT. The incidence and natural history of type I and II
endoleak: A 5-year follow-up assessment with color duplex ultrasound scan. J Vase Surg
2002; 35:474-481.

20. McLafferty RB, McCrary BS, Mattos MA, Karch LA, Ramsey DE, Solis MM, Hodgson KJ.
The use of color-flow duplex scan for the detection of endoleaks. J Vase Surg 2002; 36:100-104.

21. d'Audiffret A, Desgranges P, Kobeiter DH, Becquemin JP. Follow-up evaluation of endo-
luminally treated abdominal aortic aneurysms with duplex ultrasonography: Validation with
computed tomography. J Vase Surg 2001; 33:42-50.

22. Wolf YG, Johnson BL, Hill BB, Rubin GD, Fogarty TJ, Zarins CK. Duplex ultrasound
scanning versus computed tomographic angiography for postoperative evaluation of endovas-
cular abdominal aortic aneurysm repair. J Vase Surg 2000; 32:1142-1148.

23. McWilliams RG, Martin J, White D, Gould DA, Harris PL, Fear SC, Brennan J, Gilling-
Smith GL, Bakran A, Rowlands PC. Use of contrast-enhanced ultrasound in follow-up after
endovascular aortic aneurysm repair. J Vase Intervent Radiol 1999; 10:1107-1114.

24. Heilberger P, Schunn C, Ritter W, Weber S, Raithel D. Postoperative color flow duplex
scanning in aortic endografting. J Endovasc Surg 1997; 4:262-271.

25. Matsumura JS, Ryu RK, Ouriel K. Identification and implications of transgraft microleaks
after endovascular repair of aortic aneurysms. J Vase Surg 2001; 34:190-197.

26. Haulon S, Lions C, McFadden EP, Koussa M, Gaxotte V, Halna P, Beregi JP. Prospective
evaluation of magnetic resonance imaging after endovascular treatment of infrarenal aortic
aneurysms. Eur J Vase Endovasc Surg 2001; 22:62-69.

27. Greenberg R, Green R. A clinical perspective on the management of endoleaks after abdominal

aortic endovascular aneurysm repair. J Vase Surg 2000; 31:836-837. j>

28. Serino F, Abeni D, Galvagni E, Sardella SG, Scuro A, Ferrari M, Ciarafoni I, Silvestri L, <S
Fusco A. Noninvasive diagnosis of incomplete endovascular aneurysm repair: D-dimer assay to js
detect type I endoleaks and nonshrinking aneurysms. J Endovasc Ther 2002; 9:90-97. °|

29. Lee WA, Rubin GD, Johnson BL, Arko F, Fogarty TJ, Zarins CK. "Pseudoendoleak"— ^
Residual intrasaccular contrast after endovascular stent-graft repair. J Endovasc Ther 2002; >9
9:119-123. |

30. Beebe HG, Cronenwett JL, Katzen BT, Brewster DC, Green RM. Results of an aortic endo- q
graft trial: Impact of device failure beyond 12 months. J Vase Surg 2001 Feb; 33(2 suppl):S55- |
S63. S

31. van Marrewijk C, Buth J, Harris PL, Norgren L, Nevelsteen A, Wyatt MG Significance of «

I
§

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

680 BEEBE

endoleaks after endovascular repair of abdominal aortic aneurysms: The EUROSTAR ex-
perience. J Vase Surg 2002; 35:461-473.

32. Resch T, Ivancev K, Lindh M, Nyman U, Brunkwall J, Malina M, Lindblad B. Persistent
collateral perfusion of abdominal aortic aneurysm after endovascular repair does not lead to
progressive change in aneurysm diameter. J Vase Surg 1998; 28:242-249.

33. Tuerff SN, Rockman CB, Lamparello PJ, Adelman MA, Jacobowitz GR, Gagne PJ, Nal-
bandian MM, Weiswasser J, Landis R, Rosen RJ, Riles TS. Are type II (branch vessel) endo-
leaks really benign? Ann Vase Surg 2002; 16:50-54.

34. Zarins CK, White RA, Hodgson KJ, Schwarten D, Fogarty TJ. Endoleak as a predictor of
outcome after endovascular aneurysm repair: AneuRx multicenter clinical trial. J Vase Surg
2000; 32:90-107.

35. Schunn CD, Krauss M, Heilberger P, Ritter W, Raithel D. Aortic aneurysm size and graft
behavior after endovascular stent-grafting: Clinical experiences and observations over 3 years. J
Endovasc Ther 2000; 7:167-176.

36. Hinchliffe RJ, Singh-Ranger R, Davidson IR, Hopkinson BR. Rupture of an abdominal aortic
aneurysm secondary to type II endoleak. Eur J Vase Endovasc Surg 2001; 22:563-565.

37. White RA, Walot I, Donayre CE, Woody J, Kopchok GE. Failed AAA endograft exclusion
due to type II endoleak: Explant analysis. J Endovasc Ther 2001; 8:254-261.

38. Wolf YG, Hill BB, Rubin GD, Fogarty TJ, Zarins CK. Rate of change in abdominal aortic
aneurysm diameter after endovascular repair. J Vase Surg 2000; 32:108-115.

39. Wolf YG, Hill BB, Fogarty TJ, Cipriano PR, Zarins CK. Late endoleak after endovascular
repair of an abdominal aortic aneurysm with multiple proximal extender cuffs. J Vase Surg
2002; 35:580-583.

40. Ohki T, Veith FJ, Shaw P, Lipsitz E, Suggs WD, Wain RA. Bade M, Mehta M, Cayne N,
Cynamon J, Valldares J, McKay J. Increasing incidence of midterm and long-term compli-
cations after endovascular graft repair of abdominal aortic aneurysms: A note of caution based
on a 9-year experience. Ann Surg 2001; 234:323-334.

41 . Carpenter JP, Baum RA, Barker CF, Golden MA, Velazquez OC, Mitchell ME, Fairman RM.
Durability of benefits of endovascular versus conventional abdominal aortic aneurysm repair.
J Vase Surg 2002; 35:222-228.

42. Lee WA, Wolf YG, Fogarty TJ, Zarins CK. Does complete aneurysm exclusion ensure long-
term success after endovascular repair? J Endovasc Ther 2000; 7:494-500.

43. Albertini J, Kalliafas S, Travis S, Yusuf SW, Macierewicz J A, Whitaker SC, Elmarasy NM,
Hopkinson BR. Anatomical risk factors for proximal perigraft endoleak and graft migration
following endovascular repair of abdominal aortic aneurysms. Eur J Vase Endovasc Surg 2000;
19:308-312.

44. Mohan IV, Laheij RJ, Harris PL. Risk factors for endoleak and the evidence for stent-graft
oversizing in patients undergoing endovascular aneurysm repair. Eur J Vase Endovasc Surg
2001; 21:344-349.

45. Greenberg R, Fairman R, Srivastava S, Criado F, Green R. Endovascular grafting in patients
with short proximal necks: An analysis of short-term results. Cardiovasc Surg 2000; 8:350-354.

46. Petrik PV, Moore WS. Endoleaks following endovascular repair of abdominal aortic aneu-

rysm: The predictive value of preoperative anatomic factors — A review of 100 cases. J Vase g

Surg 2001; 33:739-744. |

47. Arko FR, Rubin GD, Johnson BL, Hill BB, Fogarty TJ, Zarins CK. Type-II endoleaks |
following endovascular AAA repair: Preoperative predictors and long-term effects. J Endovasc °|
Ther 2001; 8:503-510. ^

48. Gorich J. Rilinger N. Sokiranski R, Soldner J, Kaiser W, Kramer S, Ermis C, Schutz A, «
Sunder-Plassmann L, Pamler R. Endoleaks after endovascular repair of aortic aneurysm: Are J
they predictable?— Initial results. Radiology 2001; 218:477-480. q

49. Gould DA, McWilliams R, Edwards RD, Martin J, White D, Joekes E, Rowlands PC, |
Brennan J, Gilling-Smith G, Harris PL. Aortic side branch embolization before endovascular S
aneurysm repair: Incidence of type II endoleak. J Vase Intervent Radiol 2001; 12:337-341. «

I
§

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOLEAK 681

50. Walker SR, Macierewicz J, Hopkinson BR. Endovascular AAA repair: Prevention of side
branch endoleaks with thrombogenic sponge. J Endovasc Surg 1999; 6:350-353.

51. Veith FJ, Baum RA, Ohki T, Amor M, Adiseshiah M, Blankensteijn JD, Buth J, Chuter TA,
Fairman RM, Gilling-Smith G, Harris PL, Hodgson KJ, Hopkinson BR, Ivancev K, Katzen
BT, Lawrence-Brown M, Meier GH, Malina M, Makaroun MS, Parodi JC, Richter GM,
Rubin GD, Stelter WJ, White GH, White RA, Wisselink W, Zarins CK. Nature and signif-
icance of endoleaks and endotension: Summary of opinions expressed at an international con-
ference. J Vase Surg 2002; 35:1029-1035.

52. Schurink GW, Aarts NJ, Wilde J, van Baalen JM, Chuter TA, Schultze Kool LJ, van Bockel
JH. Endoleakage after stent-graft treatment of abdominal aneurysm: Implications on pressure
and imaging — an in vitro study. J Vase Surg 1998; 28:234-241.

53. Parodi JC, Berguer R, Ferreira LM, La Mura R, Schermerhorn ML. Intra-aneurysmal pres-
sure after incomplete endovascular exclusion. J Vase Surg 2001; 34:909-914.

54. Fairman RM, Velazquez O, Baum R, Carpenter J, Golden MA, Pyeron A, Criado F, Barker C.
Endovascular repair of aortic aneurysms: Critical events and adjunctive procedures. J Vase
Surg 2001; 33:1226-1232.

55. Chuter TA, Reilly LM, Kerlan RK, Sawhney R, Canto CJ, Ring EJ, Messina LM. Endovascu-
lar repair of abdominal aortic aneurysm: Getting out of trouble. Cardiovasc Surg 1998; 6:232-
239.

56. Dias NV, Resch T, Malina M, Lindblad B, Ivancev K. Intraoperative proximal endoleaks dur-
ing AAA stent-graft repair: Evaluation of risk factors and treatment with Palmaz stents. J
Endovasc Ther 2001; 8:268-273.

57. Cao P, Verzini F, Zannetti S, De Rango P, Parlani G, Lupattelli L, Maselli A. Device mi-
gration after endoluminal abdominal aortic aneurysm repair: Analysis of 1 1 3 cases with a
minimum follow-up period of 2 years. J Vase Surg 2002; 35:229-235.

58. Baum RA, Carpenter JP, Tuite CM, Velazquez OC, Soulen MC, Barker CF, Golden MA,
Pyeron AM, Fairman RM. Diagnosis and treatment of inferior mesenteric arterial endoleaks
after endovascular repair of abdominal aortic aneurysms. Radiology 2000; 215:409-413.

59. Baum RA, Carpenter JP, Golden MA, Velazquez OC, Clark TW, Stavropoulos SW, Cope C,
Fairman RM, Stavropoulous SW. Treatment of type 2 endoleaks after endovascular repair of
abdominal aortic aneurysms: Comparison of transarterial and translumbar techniques. J Vase
Surg 2002; 35:23-29.

60. Marty B, Sanchez LA, Ohki T, Wain RA, Faries PL, Cynamon J, Marin ML, Veith FJ. Endoleak
after endovascular graft repair of experimental aortic aneurysms: Does coil embolization with
angiographic "seal" lower intraaneurysmal pressure? J Vase Surg 1998; 27:454-461.

61. Martin ML, Dolmatch BL, Fry PD, Machan LS. Treatment of type II endoleaks with Onyx. J
Vase Intervent Radiol 2001; 12:629-632.

62. Fairman RM, Carpenter JP, Baum RA, Larson RA, Golden MA, Barker CF, Mitchell ME,
Velazquez OC. Potential impact of therapeutic warfarin treatment on type II endoleaks and sac
shrinkage rates on midterm follow-up examination. J Vase Surg 2002; 35:679-685.

63. Holzenbein TJ, Kretschmer G, Thurnher S, Schoder M, Aslim E, Lammer J, Polterauer P.
Midterm durability of abdominal aortic aneurysm endograft repair: A word of caution. J Vase

Surg 2001; 33(2 suppl):S46-S54. 1

64. May J, White GH, Waugh R, Petrasek P, Chaufour X, Arulchelvam M, Stephen MS, Harris |
JP. Life-table analysis of primary and assisted success following endoluminal repair of ab- js
dominal aortic aneurysms: The role of supplementary endovascular intervention in improving °|
outcome. Eur J Vase Endovasc Surg 2000; 19:648-655. ^

65. May J, White GH, Harris JP. Early and late conversion from endoluminal to open repair. >9
Semin Vase Surg 1999; 12(3):207-214. J

66. Trout HH III, Tanner HM. A new vascular Endostaple: A technical description. J Vase Surg q
2001; 34:565-568. |

67. Fry PD, Martin M, Machan L. Endoleaks and the need for a paradigm shift. J Endovasc Ther S

2000; 7:521.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

42

Complications Following Endovascular
Thoracic Aortic Aneurysm Repair

Alfio Carroccio and Sharif H. Ellozy

Mount Sinai School of Medicine, New York, New York, U.S.A.

Larry H. Hollier

Louisiana State University Health Sciences Center School of Medicine,
New Orleans, Louisiana, U.S.A.

Complications associated with endovascular stent-graft repair of thoracic aneurysms
include anatomical as well as device-related problems. The consequences related to
anatomy include issues with access during and following endograft insertion, endoleaks
resulting from poor stent-graft fixation or persistent collateral flow within the excluded
segment, ischemia following interruption of excluded aortic branch vessels, and the host
inflammatory response to the endograft (1-10). Device-related complications include stent
fractures as well fabric breakdown, which may ultimately result in device-related endoleak
(11). These endovascular-related complications are the focus of this chapter.

I. ANATOMICALLY RELATED COMPLICATIONS
A. Access Vessels

Successful endovascular repair of aortic aneurysms requires access arteries, that will allow ■§

safe passage of the endograft device. The iliac arteries, which are the more commonly used g

access vessels, are often plagued by vessel tortuosity and calcified obstructive occlusive »

disease. A combination of these variables may, in their worst form, result in vessel dis- c

ruption, thrombosis, or procedure termination due to failed access. Despite successful <

deployment, arterial repair may be necessary in nearly a third of such cases (10). This >9

problem becomes even more pronounced with thoracic aortic aneurysms due to the larger- -|

diameter devices required for thoracic aneurysms as compared to abdominal aneurysms. Q

Endograft delivery systems used in thoracic aneurysm repair can range in diameter from 22 |

to 27F; therefore the avoidance of vessel injury calls for thorough preoperative evaluation. @

683

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

684

CARROCCIO et al.

With preoperative knowledge of the underlying access vessel anatomy (Fig. 1), cases
with prohibitive anatomy can be approached by adjunctive measures to enhance success,
such as balloon angioplasty, as well as the utilization of accessory conduits (12). Balloon
angioplasty can dilate a restrictive arterial lesion to a diameter that can allow passage of
the endograft delivery system. Similarly, with multiple occlusive lesions and/or tortuosity,
a more prudent approach may be to place an accessory conduit of either polytetrafluoro-
ethylene (PTFE) or Dacron to the common iliac artery or distal aorta through which the
device can be inserted, thus bypassing the complex diseased area. Following the procedure,
the conduit can be removed.

B. Endoleaks

Endoleaks, a complication specific to endovascular therapy, is the persistent blood flow
within the aneurysmal sac following endograft placement. Endoleaks can be due to poor
exclusion of blood flow at the proximal and distal fixation sites (type I), collateral flow
between arterial branches within the excluded aneurysmal sac (type II), leaks at junction
sites between sequential grafts, or device failure (type III) (Fig. 2).

Risk for type I endoleak is increased with severe angulation at aortic neck fixation sites,
where the currently available, relatively stiff devices may not adequately exclude blood flow
on placement as well as risk subsequent migration with delayed endoleak (13). The transition
from the aortic arch to the descending thoracic aorta results in angulation, which — when
combined with a short segment of normal aorta in which fixation is attempted — may prove
problematic. These situations may be approached by increasing the area of fixation by laying

I
•5

Figure 1 Angiogram in a patient with an aortic aneurysm demonstrating the presence of severe
occlusive disease within the iliac arteries.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOVASCULAR ANEURYSM REPAIR 685

One week

' 1

■s
y

K

;

s*> •

-^^-^

^^LT^fc

^^^^^p

■

^fc^ k \

'

>

*

w

! Three
months

\ 4

m

J >

\.

V^

Figure 2 Chest CT scan following endovascular repair of thoracic aneurysm, (upper left) Type
II endoleak evident in the early postoperative period (arrow), (lower right) CT scan repeated with no
endoleak identified.

the device across the origin of the left subclavian. Ischemic consequences following inten-
tional sacrifice of left subclavian blood flow by endograft deployment across its origin do
exist. Alternatively, to avoid the ischemic consequences of subclavian interruption, a
transposition of the left subclavian to the common carotid artery or using a conduit for
carotid-subclavian bypass may provide a means of ensuring successful exclusion of the
aneurysm from luminal blood flow and avoiding a type I endoleak (14,15).

Type II endoleaks resulting from the development of collateral channels devoloping
between excluded intercostal or bronchial arteries are likely to thrombose (1,2). Manage-
ment of type II endoleaks that fail to thrombose on delayed follow-up is considered in the
context of aneurysmal size changes. Although there is no general consensus in this regard
we feel that aneurysm growth does necessitate intervention to thrombose these collateral
channels, while aneurysms that remain stable are observed. Data to support intervention
despite the lack of growth remains to be determined.

C. Ischemia

Following endograft exclusion of thoracic aneurysms, intercostal arteries within the ■a

excluded segment are likely to thrombose. The clinical consequences following loss of |

flow within these arteries may range from a benign and clinically silent event to spinal cord as

ischemia with clinically evident paralysis/paresis. c

Paraplegia following endovascular treatment of thoracic aortic aneurysms has been <

reported in the literature to occur at a rate of 0% — 12% (2-4,10,16,17). When it is &

combined with concomitant or previous open abdominal aneurysm repair, there appears J

to be an increased risk of paraplegia. In our experience, patients undergoing a repair Q

limited to endovascular means with no previous or concurrent aortic surgery had no |

paraplegia (16). It seems important, therefore, to discuss the possible mechanisms that @

may lead to spinal cord ischemia in endovascular repair. %

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

686 CARROCCIO et al.

We feel that four mechanisms may explain spinal cord ischemia during or following
endovascular repair of thoracic aortic aneurysm:

1 . Paraplegia that is immediately evident following occlusion of critical intercostal
arteries.

2. Despite sacrifice of intercostal vessels, collateral arterial blood supply maintains
spinal cord perfusion in the resting state. Spinal cord ischemia, however, may
develop with episodes of hypotension or subsequent loss of collaterals.

3. Coverage of critical intercostal arteries does not result in their occlusion; instead,
they are evident as endoleaks. Subsequently intercostal arteries can thrombose
with resolution of endoleak, presenting with spinal cord compromise.

4. Visceral ischemia from abdominal repair or visceral embolization results
in cytokine release, inducing secondary cord ischemia from a "no reflow"
phenomenon.

The lower incidence of spinal cord ischemia during endovascular repair of thoracic
aortic aneurysms places emphasis on the role of intercostal arteries and other collateral
blood supplies. In a report from Griepp et al., routine intraoperative somatosensory
evoked potentials were utilized during thoracoabdominal aortic aneurysm repair (18).
They observed no acute changes after serial temporary occlusion of segmental vessels. In
their study, with no reattached intercostal or lumbar arteries in any patients, an overall
paraplegia rate of 2% was achieved. The absence of evidence of spinal cord ischemia
following occlusion of intersegmental arteries during aneurysm resection questions the
essential integrity of the "artery of Adamkiewicz" in spinal cord function.

They support there is a functionally continuous anterior spinal artery stretching from
the foramen magnum to the cauda equina, with multiple inputs throughout its length, and
that no single segmental input is absolutely required for maintenance of spinal cord
integrity. Instead, with multiple contributions into the anterior spinal artery, serial
sacrifice of segmental vessels can occur without spinal cord ischemia. They identified a
risk of paraplegia when the number of sacrificed intercostals was greater than 10 — a
scenario more likely with the more extensive type II aneurysms.

The severity of the spinal cord ischemia following endograft deployment can be
dependent upon the extent of existing collateral circulation. Where there is adequate
collateral preservation after endografting, we do not expect to see clinical evidence of
ischemia. If collaterals are absent and critical intercostal arteries are covered, an ischemic
event occurs (Fig. 3). If the existing intercostal or lumbar artery collateral supply is
marginal, a tenuous cord perfusion more vulnerable to any postoperative hemodynamic
insult results. There may exist an incomplete or intermediate cord ischemia in the regional
distribution of the excluded intercostal arteries secondary to marginal collateralization.
This may present as a delayed-onset neurological deficit in the endograft patients, as the in-

vulnerable cord is more sensitive to decreases in spinal artery perfusion pressure caused by a

postoperative hemodynamic compromise or delayed thrombosis of previously patent yet H,

covered intercostal arteries. =j

While intercostal thrombosis can, in its more deleterious form, result in paraplegia, we g

have noticed a syndrome of back pain following endovascular thoracic aneurysm repair g

that we propose is also a consequence of intercostal artery thrombosis. We identified a «

proportion of our patients complaining of paraspinal back pain 24-72 h following "g

endovascular repair of their thoracic aneurysms. This pain persists for 2-14 days without g

any other clinically evident events. In these instances, we identified occlusion of previously 8

1
§

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOVASCULAR ANEURYSM REPAIR

687

Figure 3 MRA of the thoracic aorta (sagittal view) illustrating a single dominant spinal artery
originating from the aneurysmal segment.

patent intercostal arteries when these patients were evaluated with computed tomography
(CT) or magnetic resonance angiography (MRA). Paraspinus muscle ischemia from
intercostal thrombosis is our proposed etiology of this ischemia. To more accurately
quantify these findings, we have instituted a protocol utilizing preoperative MRA to
identify the intercostal and spinal artery blood supply in patients undergoing thoracic and
thoracoabdominal aneurysm repair. Following their aneurysm repairs, reevaluation of

I
•5

Figure 4 CT scan of the chest following thoracic aneurysm stent-grafting demonstrating the pre-
sense of a reactive effusion.

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

688

CARROCCIO et al.

their intercostal and spinal artery blood supply will be evaluated for association with
clinical sequelae ranging from paraspinous pain to paraplegia.

D. Inflammatory Response

Reports of a host response to the presence of the endografts as a postimplanatation syn-
drome has been reported to occur in 43-78% of patients (8,9). The clinical consequences
of this phenomenon include pain, fever, leukocytosis, elevated sedimentation rates, and ele-
vated c-reactive protein.

An inflammatory response specific to thoracic aneurysm repairs has been the develop-
ment of a pleural effusion not present prior to endograft insertion (Fig. 4). We have
witnessed this in 10% of our patient population. In its more severe form, it can result in
respiratory compromise requiring repeated thoracentesis. Interestingly, despite the lack of
evidence of aneurysmal leak or rupture in any of these patients, thoracentesis yields a
serosanguinous effusion.

II. DEVICE-RELATED COMPLICATIONS

As with any device, trials and follow-up investigations provide evidence of durability.
Stent grafts are no exception to this rule. They are composed of both a metallic frame and
a synthetic graft, and repeated stress on these elements can ultimately result in metal
fracture and fabric erosion. Concerns in this regard, of course, depend on whether these
forces will result in a type III endoleak and thus treatment failure and risk of rupture
(Fig. 5).

I
•5

Figure 5 Fluoroscopic image demonstrating stent fracture (arrow).

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

ENDOVASCULAR ANEURYSM REPAIR 689

A. Stent Fracture

In a review by Jacobs et al. reporting device failure rates in endovascular aortic aneurysm
repair, a 15% device failure rate as determined by radiographic imaging and or pathologic
specimens was identified (11). Of these failures, 23% were suture disruption, 5% were
fabric failure, and 72% were metal stent fractures. Approximately 20% of thoracic
aneurysms resulted in device failures, all due to metallic stent fractures. This is concerning,
as any device failure can potentially result in treatment failure, with patients requiring
repeated intervention to exclude the aneurysm sac.

III. SUMMARY

Complications following endovascular repair of thoracic aortic aneurysms are both
anatomical and device-related. This suggests a more crucial role for adequate preoperative
anatomical assessment as well as modification of devices. Restrictive access vessels,
complex aneurysmal disease, and critical collateral blood supply to the spinal cord are
areas that need further characterization. Similarly, the creation of devices with increased
durability, lower profiles, and increased flexibility may prove successful in overcoming
some of the limitations we currently face.

REFERENCES

1. Thompson CS, Gaxotte VD, Rodriguez JA, et al. Endoluminal stent grafting of the thoracic
aorta: initial experience with the Gore Excluder. J Vase Surg 2002; 35(6): 1 163 — 1 170.

2. Heijman RH, Deblier IG, Moll FL, et al. Endovascular stent-grafting for descending thoracic
aortic aneurysms. Eur J of Cardiothorac Surg 2002; 21:5-9.

3. Taylor PR, Gaines PA, McGuinness CL, Cleveland TJ, Beard JD, Cooper G, Reidy JF.
Thoracic aortic stent grafts — Early experience from two centers using commercially available
devices. Eur J Vase Endovasc Surg 2001; 22(l):70-76.

4. Dake MD, Miller DC, Mitchell RS, et al. The "first generation" of endovascular stent-grafts
for patients with aneurysms of the descending thoracic aorta. J Thorac Cardiovasc Surg 1998;
116:689-704.

5. Dake MD, Kato N, Mitchell RS, et al. Endovascular stent-graft placement for the treatment of
acute aortic dissection. N Engl J Med 1999; 340:1546-1552.

6. Mitchell RS, Miller DC, Dake MD. Endovascular stent graft repair of thoracic aortic aneu-
rysms. Semin Vase Surg 1997; 10:257-271.

7. Mitchell RS. Endovascular stent-graft repair of thoracic aortic aneurysms. Semin Thorac Car-
diovasc Surg 1997; 9:257-268.

8. Won JY, Lee DY, Shim WH, et al. Elective endovascular treatment of descending thoracic
aortic aneurysms and chronic dissections with stent-grafts. J Vase Intervent Radiol 2001;
12(5):575-582. ™

9. Nienaber CA, Fattori R, Lund G, et al. Non surgical reconstruction of thoracic aortic |
dissection by stent-graft placement. N Engl J Med 1999; 340:1539-1545. 8

10. White RA, Donayre CE, Walot I, et al. Endovascular exclusion of descending thoracic aortic %
aneurysms and chronic dissections: initial clinical results with the AneurRx device. J Vase Surg a
2001; 33:927-934. j

11. Jacobs T, Won J, Faries P, et al. Device failure following aortic stent grafting. 56th Annual g
Meeting for the Society of Vascular Surgery, Boston, MA, June 9-12, 2002. s

12. Yano OJ, Faries PL, Morrissey N, Teodorescu V, Hollier LH. Marin ML. Ancillary techniques 2
to facilitate endovascular repair of aortic aneurysms. J Vase Surg 2001; 34(l):69-75. |

13. Resch T, Koul B, Dias NV, et al. Changes in aneurysm morphology and stent-graft con- 5

1

a

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

690 CARROCCIO et al.

figuration after endovascular repair of aneurysms of the descending thoracic aorta. J Thorac
Cardiovasc Surg 2001; 22(l):47-52.

14. Shigemura N, Kato M, Kuratani T, et al. New operative method for acute type B dissection:
Left carotid artery-left subclavian artery bypass combined with endovascular stent-graft im-
plantation. J Thorac Cardiovasc Surg 2000; 120(2):406-408.

1 5. Moore RD, Brandschwei F. Subclavian-to-carotid transposition and supracarotid endovascular
stent graft placement for traumatic aortic disruption. Ann Vase Surg 2001; 15(5):563 — 566.

16. Gravereaux EC, Faries PL, Burks JA, Latessa V, Spielvogel D, Hollier LH, Marin ML. Risk of
spinal cord ischemia after endograft repair of thoracic aortic aneurysms. J Vase Surg 2001;
34(6):997-1003.

17. Greenberg R, Resch T, Nyman U. Endovascular repair of descending thoracic aortic aneurysms:
An early experience with intermediate-term follow up. J Vase Surg 2000; 31:147-156.

18. Griepp RB, Ergin MA, Galla JD, Lansman S, Khan N, Quintana C. Looking for the artery of
Adamkiewicz: A quest to minimize paraplegia after operations for aneurysms of the descending
thoracic and thoracoabdominal aorta. J Thorac Cardiovasc Surg 1996; 112:1202-1215.

I
•5

I

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

43

Complications of AngiogenesisTherapy

Joshua Bernheim, Sashi Kilaru, and K. Craig Kent

New York Presbyterian Hospital-The University Hospitals of Columbia and Cornell,

New York, New York, U.S.A.

Cardiovascular disease is the major cause of death in adults in most developed and many
developing countries. The mainstay of therapy for coronary and lower extremity ischemia is
revascularization, either by catheter-based or surgical techniques. Both approaches have
been effective in reducing the morbidity associated with these disease processes. Unfortu-
nately, not all patients with coronary or peripheral ischemia are candidates for intervention.
Often, multiple comorbidities associated with diffuse atherosclerosis preclude the use of
invasive treatments. Consequently, a less invasive strategy would be a welcome adjunct to
current therapeutic alternatives.

Therapeutic angiogenesis — the promotion of new vessel growth using vascular growth
factors — is currently one of the more dynamic areas in biomedical research. Reestablishing
blood flow to an ischemic region through angiogenesis has the potential to provide a
"biological bypass" for patients with atherosclerotic occlusive disease. Studies in animals
have suggested that the exogenous administration of angiogenic factors, particularly
fibroblast growth factor (FGF) or vascular endothelial growth factor (VEGF), may aug-
ment blood flow in regions of arterial ischemia, thereby improving distal tissue perfusion.
The successful development of therapeutic angiogenesis as a minimally invasive approach to
vascular insufficiency could tremendously expand our ability to treat patients with coronary
and limb-threatening ischemia. ■g

Numerous reports have documented, in animals, the potential for a variety of cytokines &

and growth factors to induce neovascularization and enhance perfusion of ischemic tissues. a

These investigations have led to a number of pilot clinical trials designed to test the safety and -c

efficacy of therapeutic angiogenesis. In an early study, Isner at al. in 1997 reported the results <j

of a phase 1 clinical trial of VEGF in nine patients with critical limb ischemia (1). The >3

majority of these patients had either rest pain or nonhealing ulcers, and none were 41

considered to be candidates for surgical or percutaneous revascularization. Gene transfer 2

was performed on two occasions separated by 4-week intervals with intramuscular injection |

into ischemic limbs of naked plasmid DNA encoding VEGF, 65 . Successful gene expression @

691 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

692 BERNHEIM et al.

in patients was documented by an increase in serum VEGF levels. The investigators noted an
improvement in the average ankle-brachial index (ABI) at 12 weeks from 0.33 to 0.48. New
collaterals were demonstrated by both MRA and contrast radiography. Limb salvage was
achieved in 6 of 9 patients.

Isner's initial report produced a wave of enthusiasm that led to the organization of
several prospective clinical trials where VEGF and other growth factors were further tested.
In a recently completed phase II randomized, placebo-controlled trial (TRAFFIC, Chiron
Corporation), patients with intermittent claudication were randomized to placebo versus
one or two doses of intra-arterial recombinant FGF (2). A total of 192 patients were en-
rolled. A positive effect on peak walking time was observed at 90 days in one treatment arm.
At 1 80 days, however, treatment with FGF did not alter claudication severity, stair climbing,
walking speed, or walking distance. The findings of this study were generally discouraging.

In a recently completed phase I multicenter trial, Comerota et al. tested the safety and
tolerability of an increasing single dose of plasmid DNA encoding FGF (NV1FGF) in
patients with limb-threatening peripheral occlusive disease (3). Fifty-one patients were
enrolled, and doses ranging from 500 to 4000 g of NV1FGF were injected intramuscularly
into the thighs and calves of ischemic extremities. No serious adverse events were noted
during the course of this study. Clinical outcomes for the first 15 patients with 6-month
follow-up were reported. A significant decrease in rest pain was noted and the ABI increased
in all patients. Moreover, healing was observed in all 9 patients who presented with ulcers.
These encouraging results have led to the initiation of a phase II placebo-controlled trial that
is currently under way.

Multiple other trials of angiogenesis have either been completed or are currently under
way. The results of these trials have in general been mixed, and it is clear at this point that
therapeutic angiogenesis is still investigational. However, scientific investigation in this
field is advancing at a rapid pace. New growth factors, combinations of growth factors, or
new approaches including stem cell therapy are all being evaluated. Over the next few
years, advances in these areas may bring this technology to fruition.

The ability to "turn on" growth of new blood vessels can be advantageous. However,
angiogenesis can potentially produce complications, such as diabetic retinopathy and the
spread of tumors. In recent years, gene therapy — one method of applying angiogenic
proteins — has been plagued by a number of untoward events that have dampened enthu-
siasm for its use. There is clearly a need to fully understand both the benefits of these new
techniques and their potential complications. In the remainder of this chapter, we discuss the
various potential complications and concerns that have been raised regarding the use of
therapeutic angiogenesis.

I. INCREASED VASCULAR PERMEABILITY

1
Although VEGF is most widely known for its properties as a stimulant of angiogenesis, it g

was originally described as a vascular permeability factor (4). In fact, VEGF is one of the a

most effective permeability-enhancing agents yet discovered, with a rapid onset of action and c

a potency up to 50,000 times greater than that of histamine (5,6). Increased permeability <j

stimulated by VEGF can result in extravasation of plasma proteins into the extravascular >9

space, leading to fibrin deposition and edema. This phenomenon has been well documented 4j

in clinical trials. Baumgartner found — in a study where naked plasmid DNA encoding 2

VEGF was used to treat peripheral ischemia — that 24% of patients with rest pain and 60% |

S

©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ANGIOGENESIS THERAPY 693

of those with gangrene developed significant peripheral edema (7). There were no untoward
events in these patients, and in most cases the edema promptly responded to diuretic therapy.
Capillary leakage from increased vascular permeability has the potential to be a life-
threatening side effect of angiogenesis when the treatment is systemically applied. In a
murine model, Thurston et al. observed lethal consequences related to vascular perme-
ability — including diffuse tissue edema and brain swelling — with systemic adenoviral gene
transfer of VEGF (8). In human subjects, however, there have been no reports of life-
threatening edema attributable to VEGF therapy. This is likely due to the fact that, in
trials thus far, patients have been treated with localized therapy, resulting in low doses of
VEGF of short duration. In Thurston's model, the gene transfer regimen was designed to
produce circulating levels of VEGF 3-5 logs greater than those seen in any human clinical
application (1,8,9). It appears that local delivery results in systemic levels of VEGF that
are in the range of picograms per milliliter. This is due to the high affinity of VEGF for
matrix proteins. The effect on vascular permeability appears to be specific to VEGF and its
homologues. In a Miles assay, an established test to demonstrate the permeability-enhancing
effects of substances on the vascular system, other angiogenic factors, such as granulocyte-
macrophage colony-stimulating factor, transforming growth factor beta, PDGF, and FGF
were not associated with significant vascular permeability (10).

II. HYPOTENSION

Vasodilatation does occur in response to multiple growth factors, including VEGF and
FGF. This effect results in part from upregulation of nitric oxide (NO) synthase (11-13);
NO is an important mediator of angiogenesis as well as a vasodilator (14-16).

In an early study of angiogenesis in pigs, 4 of 8 animals undergoing intracoronary VEGF
administration died of refractory hypotension (17). Hypotension has also been observed in
human subjects receiving intra-arterial FGF protein. Lazarous et al. reported a case of
hypotension in a study of claudicants receiving recombinant FGF, after which the study was
modified to reduce the rate of administration of the drug (18). Unger et al. also reported a
patient who developed hypotension after intracoronary administration of 100 (ig/kg of FGF
(19). Thus far, hypotension has been confined to protocols where the angiogenic protein is
administered either intra-arterially or intravenously. Hypotension has not developed
following delivery of angiogenic agents via vectors; their gradual expression results in lower
circulating levels of growth factors. It is also likely that, even when intravascular proteins are
used as the method of delivery, refinements of dosing and rate of administration of growth
factors will lessen the potential for this complication.

III. VASCULAR MALFORMATIONS

1
Growth factors used for therapeutic angiogenesis have the potential to stimulate disorga- &

nized proliferation of normal vessels, leading to the formation of hemangiomas or other a

vascular malformations. Hemangiomas have been observed in animals when there has been -c

prolonged exposure of skeletal muscle or myocardium to high concentrations of VEGF (20- <j

22). To date, no association has been made between VEGF and vascular malformations in >9

human clinical trials with the doses and durations of therapy used. There is only a single case 41

report of telangectasia formation in a leg following VEGF gene transfer (23). Despite Q

detailed examination of autopsy specimens, amputated limbs, and explanted hearts, |

S

©

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

694 BERNHEIM et al.

vascular neoproliferation has not been demonstrated in other studies of clinical angiogenesis
(10,24-30).

IV. RETINOPATHY

The high incidence of diabetes mellitus in patients with peripheral vascular disease has
raised concern over the possibility that angiogenic growth factors might exacerbate
proliferative diabetic retinopathy, which results from the abnormal proliferation of retinal
arterioles and is exceedingly common in patients with diabetes. Levels of VEGF and FGF-
2 are increased in the neovascular membrane and ocular fluid of patients with diabetic
retinopathy (31-33). Moreover, subretinal administration of adenoviral vectors expressing
VEGF results in the development of retinopathy (34).

Thus far, in clinical trials of angiogenesis, this complication has not occurred. Close
monitoring of patients with careful, routine ophthalmological examinations has been a
required part of most angiogenesis protocols. Isner et al. reported their findings in 44 diabetic
patients undergoing VEGF gene transfer for a variety of indications. All patients underwent
detailed retinal evaluation with fluorescin before the introduction of the vector and at 3, 6,
and 12 months after gene transfer. No patient with preexisting retinopathy had progression
of his or her proliferative disease, and no patient with a normal examination prior to therapy
developed retinopathic changes. Moreover, none of the nondiabetic patients developed
retinal pathology in the first year following gene transfer (10). Thus, at this point, the risk of
developing retinopathy remains theoretical.

V. NEPHROTOXICITY

Nephrotoxicity is a side effect angiogenic therapy. Rats receiving 4 weeks of high-dose
basic FGF developed glomerular pathology that included both vacuolization of the
glomerular cells regulating filtration, and hyperplasia of Bowman's capsule epithelium
(35). Thus, renal toxicity appears to be related to a direct toxic effect of FGF, resulting in
structural changes in the glomerulus, and not as a consequence of the angiogenic
properties of FGF. Nephrotoxicity has been observed in clinical trials where FGF is the
angiogenic agent. In a recently conducted phase II trail sponsored by Scios Corporation, 5
of the 16 subjects who received FGF intravenously developed significant proteinuria,
leading to early termination of the trial. Though a worrisome occurrence, it should be
noted that patients in this trial were receiving systemic therapy with recombinant protein.
Intramuscular or even intra-arterial administration of FGF, when this growth factor is
applied through a genetic vector, can lessen systemic toxicity and result in more focused
targeting of ischemic tissues.

1

VI. NEOPLASIA 1

In many of the initial studies of angiogenesis, the focus was its important role in tumor -c

progression (36,37). As such, antiangiogenic proteins have been extensively evaluated as <j

tools for treating cancer. It follows that attempts to provoke angiogenesis for therapeutic >3

purposes has the potential to stimulate the growth or spread of preexisting tumors. This is 4j

particularly concerning since patients with vascular insufficiency are usually elderly and at 2

high risk for developing cancer. Although solid tumors require angiogenesis to grow, angi- |

ogenic proteins do not appear to directly produce neoplastic changes. FGF can stimulate @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ANGIOGENESIS THERAPY 695

proliferation of multiple cell types. VEGF can also stimulate growth of several non-
endothelial tumors (i.e., those that express VEGFR-1 and VEGFR-2 receptors) (38).
Neurolipin, a third receptor for VEGF, has also been found in many neoplastic cells.

Despite this theoretical concern, in neither animal nor human studies has there been
any definitive evidence that patients developed or had progression of cancer. In the VIVA
(Vascular Endothelial Growth Factor in Ischemia for Vascular Angiogenesis) trial (39),
where recombinant VEGF protein was used to treat myocardial ischemia, and the
TRAFFIC (Therapeutic Angiogenisis with Recombinant Fibroblast Growth Factor-2
for Intermittent Claudication) trial (2), where VEGF was used to treat claudication, the
only patients diagnosed with new neoplasms were in the control groups. Given the
potential for tumors to remain silent for extended period of time, the verdict is still out.
Large, randomized trials of angiogenesis with long-term follow-up will be required before
any final determination can be made.

VII. ATHEROSCLEROSIS

Angiogenesis proteins not only promote new vessel formation but their effects on existing
vessels may lead to the development or progression of atherosclerotic lesions. Multiple
studies have demonstrated that the development of atherosclerotic plaque is associated
with proliferation of the vasa vasorum (39-41). Treating hypercholesterolemic mice with
endostatin, a potent angiogenesis inhibitor, for 16 weeks reduced plaque development by
85% (42). Moreover, VEGF has been identified as a monocyte chemoattractant (43,44);
experiments have also shown that activated macrophages present within atherosclerotic
lesions have a profoundly deleterious effect on plaque evolution, stability, and rupture
(45).

Inhibition of angiogenesis retards plaque formation. However, it does not necessarily
follow that angiogenesis-promoting factors will accelerate atherosclerosis. Human studies
to date have not demonstrated progression of atherosclerotic disease as a complication of
angiogenisis. Balloon injury of rat and rabbit arteries followed by application of VEGF
using a variety of vectors did not result in accelerated restenosis or disease progression. On
the contrary, reduced intimal thickening and mural thrombus were observed in these
animals, likely as a result of VEGF's ability to stimulate endothelial regeneration (46-49).
Vale et al. conducted similar studies in humans. After angioplasty of diseased femoral
arteries, balloon catheters impregnated with a hydrogel coating were used to deliver naked
DNA encoding VEGF. These patients experienced no significant increase in restenosis up
to 48 months after gene transfer and showed no evidence of new lesion formation (50).

VIM. COMPLICATIONS ASSOCIATED WITH VIRAL VECTORS

1
Adenoviral vectors have been used in over half of the trials of therapeutic angiogenesis. g

Inflammation and cell toxicity are virus-related side effects that can develop regardless of the a

gene expressed. First-generation adenoviral vectors can stimulate expression of viral c

proteins in target cells, resulting in the production of cytotoxic T lymphocytes. This im- <j

mune response prevents repeated application of the same vector. In addition, high titers of >9

adenoviruses can damage cells and produce an inflammatory response that results in 4j

systemic toxicity and damage to tissues. The most widely publicized complication of gene 2

therapy to date was the death of a young man at the University of Pennsylvania who |

underwent hepatic arterial infusio of large doses of adeno viral vector. The patient's @

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

696 BERNHEIM et al.

presenting condition was a relatively mild form of ornithine transcarboxylase deficiency, a
rare hereditary metabolic disorder in which the liver is unable to process ammonia. The gene
replacement protocol was designed to treat this deficiency (51). This event prompted intense
scrutiny into the potential complications of gene therapy. The field has since rebounded from
this incident, and multiple studies are currently under way. However, issues related to viral
vectors remain, and new, less toxic and antigenic vectors are under development.

IX. MORTALITY

Patients with peripheral vascular disease have diminished survival and are prone to
developing complications related to their multiple comorbidities. This has led to difficulty
in classifying adverse events in angiogenesis trials, since it is not readily possible to
determine which events are due to the actual intervention versus the underlying disease
process. The 2-year mortality for patients with critical limb ischemia can be as high as 30-
35% (52). The small sample size and nonrandomized design of most trials performed to
date compounds this problem. Isner reported that among the 100 patients enrolled in gene
therapy trials at his institution, 9 deaths occurred over a 7-year period (10). Although none
of these deaths were directly attributable to gene therapy, the presence of comorbidities
complicates the determination of outcome.

X. CONCLUSIONS

Angiogenic agents that are currently under study have potent biological effects and the
potential to produce a wide array of complications. However, analysis of available data in
ongoing phase I and II trails is very encouraging. There is little evidence that these proteins
produce major morbidity or mortality. Once the size and number of trials increases and
more randomized studies are initiated, a better safety profile can be generated for these
angiogenic agents. The future of angiogenesis therapy lies in the development of growth
factors and methods of their delivery that allow tissues to be specifically targeted with
sufficient selectivity that systemic complications do not occur. The outcome will likely be
safe and effective new therapies for the treatment of atherosclerosis.

REFERENCES

1. Baumgartner I, Pieczek A, Manor O, Blair R, Kearney M, Walsh K, Isner JM. Constitutive
expression of ph VEGF165 after intramuscular gene transfer promotes collateral vessel
development in patients with critical limb ischemia. Circulation 1998; 97(12): 1 1 14-1 123.

2. Lederman RJ, Mendelsohn FO, Anderson RD, Saucedo JF, Tenaglia AN, Hermiller JB,

Hillegass WB, Rocha-Singh K, Moon TE, Whitehouse MJ, Annex BH. Therapeutic angiogenesis g

with recombinant fibroblast growth factor-2 for intermittent claudication (the TRAFFIC study): <S

A randomised trial. Lancet 2002; 359(9323):2053-2058. ,1

3. Comerota AI, Throm RC, Miller KA, Henry T, Chronos N, Laird J, Sequeira R, Kent KC, Jf
Bacchetta M, Goldman C, Salenius IP, Schmieder FA, Pilsudski R. Naked plasmid DNA ^
encoding fibroblast growth factor type 1 for the treatment of end-stage unreconstructible lower >9
extremity ischemia: preliminary results of a phase I trial. I Vase Surg 2002; 35(5):930-936. jjj

4. Senger DR, Galli SI, Dvorak AM, Perruzzi CA, Harvey VS, Dvorak HF. Tumor cells secrete a q
vascular permeability factor that promotes accumulation of ascites fluid. Science 1983; 219:983- |
985. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ANGIOGENESIS THERAPY 697

5. Dvorak HF, Brown LF, Detmar M, et al. Vascular permeability factor/vascular endothelial
growth factor, microvascular hyperpermeability, and angiogenesis. Am J Pathol 1995; 146:1029-
1039.

6. Brown LF, Detmar M, Claffey K, et al. Vascular permeability factor/vascular endothelial growth
factor: A multifunctional angiogenic cytokine. EXS 1997; 79:233-269.

7. Baumgartner I, Rauh G, Pieczek A, Wuensch D, Magner M, Kearney M, Schainfeld R, Isner
JM. Lower-extremity edema associated with gene transfer of naked DNA vascular endothelial
growth factor. Ann Intern Med 2000; 132:880-884.

8. Thurston G, Suri C, Smith K, McClain J, Sato TN, Yancopoulos GD, McDonald DM. Leakage-
resistant blood vessels in mice transgenically overexpressing angiopoietin-1. Science 1999;
286:2511-2514.

9. Isner JM, Baumgartner I, Rauh G, Schainfeld R, Blair R, Manor O, Razvi S, Symes JF.
Treatment of thromboangiitis obliterans (Buerger's disease) by intramuscular gene transfer
of vascular endothelial growth factor: preliminary clinical results. J Vase Surg 1998; 28:964-975.

10. Isner JM, Vale PR, Symes JF, Losordo DW. Assessment of risks associated with cardiovascular
gene therapy in human subjects. Circ Res 2001; 89:389-400.

11. van der Zee R, Murohara T, Luo Z, Zollmann F, Passed J, Lekutat C, Isner JM. Vascular
endothelial growth factor (VEGF)/vascular permeability factor (VPF) augments nitric oxide
release from quiescent rabbit and human vascular endothelium. Circulation 1997; 95: 1030-1037.

12. Cuevas P, Carceller F, Ortega S, Zazo M, Nieto I, Gimenez-Gallego G. Hypotensive activity of
fibroblast growth factor. Science 1991; 254:1208-1210.

13. Fulton D, Gratton JP, McCabe TJ, Fontana J, Fujio Y, Walsh K, Franke TF, Papapetropoulos
A, Sessa WC. Regulation of endothelium-derived nitric oxide production by the protein kinase
akt. Nature 1999; 399:597-601.

14. Murohara T, Asahara T, Silver M, Bauters C, Masuda H, Kalka C, Kearney M, Chen D, Symes
JF, Fishman MC, Huang PL, Isner JM. Nitric oxide synthase modulates angiogenesis in response
to tissue ischemia. J Clin Invest 1998; 101:2567-2578.

15. Ziche M, Morbidelli L, Choudhuri R, Zhang H-T, Donnini S, Granger HJ, Bicknell R. Nitric
oxide synthase lies downstream from vascular endothelial growth factor-induced but not
fibroblast growth factor-induced angiogenesis. J Clin Invest 1997; 99:2625-2634.

16. Babaei S, Teichert-Kuliszewska K, Monge J-C, Mohamed F, Bendeck MP, Stewart DJ. Role of
nitric oxide in the angiogenic response in vitro to basic fibroblast growth factor. Circ Res 1998;
82:1007-1015.

17. Hariawala M, Horowitz JR, Esakof D, Sheriff DD, Walter DH, Chaudhry GM, Desai V,
Keyt B, Isner JM, Symes JF. VEGF improves myocardial blood flow but produces EDRF-
mediated hypotension in porcine hearts. J Surg Res 1996; 63:77-82.

18. Lazarous DF, Unger EF, Epstein SE, Stine A, Arevalo JL, Chew EY, Quyyumi AA. Basic
fibroblast growth factor in patients with intermittent claudication: results of a phase I trial. J Am
Coll Cardiol 2000; 36:1339-1344.

19. Unger EF, Goncalves L, Epstein SE, et al. Effects of a single intracoronary injection of basic
fibroblast growth factor in stable angina pectoris. Am J Cardiol 2000; 85:1414-1419.

20. Springer ML, Chen AS, Kraft PE, Bednarski M, Blau HM. VEGF gene delivery to muscle:

Potential role of vasculogenesis in adults. Mol Cell 1998; 2:549-558. 1

21. Lee RJ, Springer ML, Blanco-Bose WE, Shaw R, Ursell PC, Blau HM. VEGF gene delivery to |
myocardium: Deleterious effects of unregulated expression. Circulation 2000; 102:898-901. s

22. Schwarz ER, Speakman MT, Patterson M, Hale SS, Isner JM, Kedes LH, Kloner RA. Jf
Evaluation of the effects of intramyocardial injection of DNA expressing vascular endothelial ^
growth factor (VEGF) in a myocardial infarction model in the rat — Angiogenesis and angioma «
formation. J Am Coll Cardiol 2000; 35:1323-1330. J

23. Isner JM, Pieczek A, Schainfeld R, Blair R, Haley L, Asahara T, Rosenfield K, Razvi S, Walsh K, q
Symes J. Clinical evidence of angiogenesis following arterial gene transfer of ph VEGF165. |
Lancet 1996; 348:370-374. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

698 BERNHEIM et al.

24. Symes JF, Losordo DW, Vale PR, Lathi K, Esakof DD, Maysky M, Isner JM. Gene therapy with
vascular endothelial growth factor for inoperable coronary artery disease: Preliminary clinical
results. Ann Thorac Surg 1999; 68:830-837.

25. Tabata H, Silver M, Isner JM. Arterial gene transfer of acidic fibroblast growth factor for
therapeutic angiogenesis in vivo: Critical role of secretion signal in use of naked DNA.
Cardiovasc Res 1997; 35:470-479.

26. Tsurumi Y, Takeshita S, Chen D, Kearney M, Rossow ST, Passed J, Horowitz JR, Symes JF.
Direct intramuscular gene transfer of naked DNA encoding vascular endothelial growth factor
augments collateral development and tissue perfusion. Circulation 1996; 94:3281-3290.

27. Witzenbichler B, Asahara T, Murohara T, Silver M, Spyridopoulos I, Magner M, Principe N,
Kearney M, Hu J-S, Isner JM. Vascular endothelial growth factor-C (VEGF-C/VEGF-2)
promotes angiogenesis in the setting of tissue ischemia. Am J Pathol 1998; 153:381-394.

28. Laitinen M, Makinen K, Mannienen H, Matsi P, Kossila M, Agrawal RS, Pakkanen T, Luom-
Viita H, Hartikainen J, Alhava E, Laakso M, Yla-Herttuala S. Adenovirus-mediated gene
transfer to lower limb artery of patients with chronic critical leg ischaemia. Hum Gene Ther 1998;
9:1481-1486.

29. Mack CA, Patel SR, Schwarz EA, Zanzonico P, Hahn RT, llercil A, Devereux RB, Goldsmith
SJ, Christian TF, Sanborn TA, Kovesdi I, Itackett N, Isom OW, Crystal RG, Rosengart TK.
Biologic bypass with the use of adenovirus-mediated gene transfer of the complementary
deoxyribonucleic acid for vascular endothelial growth factor 121 improves myocardial perfusion
and function in the ischemic porcine heart. J Thorac Cardiovasc Surg 1998; 1 15:168-176.

30. Giordano FJ, Ping P, McKirnan D, Nozaki S, DeMaria AN, Dillmann WH, Mathieu-Costello
O, Hammond HK. Intracoronary gene transfer of fibroblast growth factor-5 increases blood flow
and contractile function in an ischemic region of the heart. Nat Med 1996; 2:534-539.

31. Frank RN, Amin RH, Eliott D, Puklin JE, Abrams GW. Basic fibroblast growth factor and
vascular endothelial growth factor are present in epiretinal and choroidal neovascular
membranes. Am J Ophthalmol 1996; 122:393-403.

32. Aiello LP, Avery RL, Arrigg PG, Keyt BA, Jampel HD, Shah ST, Pasquale LR, Theme H,
Iwamoto MA, Parke JE, Nguyen MD, Aiello LM, Ferrara N, King GL. Vascular endothelial
growth factor in ocular fluids of patients with diabetic retinopathy and other retinal disorders. N
Engl J Med 1994; 331:1480-1487.

33. Adamis AP, Miller JW, Bernal M-T, D'Amico DJ, Folkman J, Yeo T-K, Yeo K-T. Increased
vascular endothelial growth factor levels in the vitreous of eyes with proliferative diabetic
retinopathy. Am J Ophthalmol 1994; 118:445-450.

34. Baffi J, Byrnes G, Chan CC, Csaky KG. Choroidal neovascularization in the rat induced by
adenovirus mediated expression of vascular endothelial growth factor. Invest Ophthalmol Vis Sci
2000;41:3582-3589.

35. Mazue G, Bertolero F, Garafano L, Brughere M, Carminati P. Experience with the preclinical
assessment of basic fibroblast growth factor. Toxicol Lett 1992; 64/65:329-338.

36. Folkman J. Tumor angiogenesis: Therapeutic implications. N Engl J Med 1971; 285:1182-1186.

37. Hanahan D, Folkman J. Patterns and emerging mechanisms of the angiogenic switch during
tumorigenesis. Cell 1996; 86:353-364.

38. Herold-Mende C, Steiner HH, Andl T, et al. Expression and functional significance of vascular j>
endothelial growth factor receptors in human tumor cells. Lab Invest 1999; 79:1573-1582. <S

39. Kwon HM, Sangiorgi G, Ritman EL, McKenna C, Holmes DR, Schwartz RS, Lerman A. js
Enhanced coronary vasa vasorum neovascularization in experimental hypercholesterolemia. J 2
Clin Invest 1998; 101:1551-1556. ^

40. Williams JK, Armstrong ML, Heistad DD. Vasa vasorum in atherosclerotic coronary arteries: «
Responses to vasoactive stimuli and regression of atherosclerosis. Circ Res 1988; 62:515-523. jjj

41. Barger AC, Beeuwkes R III, Lainey LL, Silverman KJ. Hypothesis: vasa vasorum and q
neovascularization of human coronary arteries: A possible role in the pathophysiology of |
atherosclerosis. N Engl J Med 1984; 310:175-177. S

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

ANGIOGENESIS THERAPY 699

42. Moulton KS, Heller E, Konerding MA, Flynn E, Palinski W, Folkman J. Angiogenesis inhibitors
endostatin and TNP-470 reduce intimal neovascularization and plaque growth in apolipoprotein
E-deficient mice. Circulation 1999; 99:1726-1732.

43. Barleon B, Sozzani S, Zhou D, et al. Migration of human monocytes in response to vascular
endothelial growth factor (VEGF) is mediated via the VEGF receptor flt-1. Blood 1996; 87:3336-
3343.

44. Clauss M, Gerlach M, Gerlach H, et al. Vascular permeability factor: A tumor-derived
polypeptide that induces endothelial cell and monocyte procoagulant activity and promotes
monocyte migration. J Exp Med 1990; 172:1535-1545.

45. Libby P, Geng YJ, Aikawa M, et al. Macrophages and atherosclerotic plaque stability. Curr Opin
Lipidol 1996; 7:330-335.

46. Van Belle E, Tio FO, Couffmhal T, Maillard L, Passeri J, Isner JM. Stent endothelialization:
Time course, impact of local catheter delivery, feasibility of recombinant protein administration,
and response to cytokine expedition. Circulation 1997; 95:438-448.

47. Van Belle E, Tio FO, Chen D, Maillard L, Kearney M, Isner JM. Passivation of metallic stents
following arterial gene transfer of ph VEGF165 inhibits thrombus formation and intimal
thickening. J Am Coll Cardiol 1997; 29:1371-1379.

48. Asahara T, Chen D, Tsurum Y, Kearney M, Rossow S, Passeri J, Symes J, Isner J. Accelerated
restitution of endothelial integrity and endothelium-dependent function following phVEGF165
gene transfer. Circulation 1996; 94:3291-3302.

49. Asahara T, Bauters C, Pastore CJ, Kearney M, Rossow S, Bunting S, Ferrara N, Symes JF, Isner
JM. Local delivery of vascular endothelial growth factor accelerates reendothelialization and
attenuates intimal hyperplasia in balloon-injured rat carotid artery. Circulation 1995; 91:2793-
2801.

50. Vale PR, Wuensch DI, Rauh GF, Rosenfield K, Schainfeld RM, Isner JM. Arterial gene therapy
for inhibiting restenosis in patients with claudication undergoing superficial femoral artery
angioplasty [abstr]. Circulation 1998; 98(suppl I):I-66.

5 1 . Stolberg SG FDA officials fault Penn team in gene therapy death. New York Times, December 9,
1989.

52. TransAtlantic Inter-Society Consensus (TASC). Management of peripheral arterial disease
(PAD). J Vase Surg 2000; 3FS1-S296.

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

i

2

I

■c

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

I

u

Index

AAA. see Abdominal aortic aneurysm (AAA)
Abciximab

pharmacology, 199

side effects, 202
Abdominal aortic aneurysm (AAA), 16

anastomotic pseudoaneurysms, 257

endovascular repair, 634-641

repair, 35

stent graft fatigue, 641

stent graft repair, ischemic colitis, 215-216
Abdominal aortic surgery, 27
Abdominal aortogram, aortoiliac knitted graft,

dilation, 87
Abdominal approach, wound complications,

370
Abdominal incisions, pararenal arteries,

251
ABG, 36
ABI, 341
Absolute ethanol, arteriovenous malformations

(AVM), 588
ACAS, 443

Access, carotid stents, 615-616
Access vessels, thoracic aortic aneurysm repair,

683-684
Accuzyme, diabetic foot, 386
ACE inhibitors, 57, 262
Acetylcysteine, 57
Acoustic shadowing, 8, 1 1
Activated protein C resistance, 170-171

Acute coronary syndrome, platelet inhibitors,

200
Acute ischemic injury, 54
Acute myocardial infarction (AMI), 16, 18
Acute postoperative graft occlusions,
treatment, 357-359

initial management, 358

operation, 358-359
Acute renal failure (ARF), 49

aortic surgery, 50

causes, 563

incidence with vascular surgery, 49-50

parenchymal causes, 54-56

prevention, 56-57
Acute respiratory distress syndrome (ARDS),

34
Adenosine, 23
Adenoviral vectors, angiogenesis therapy,

695-696
Adverse postoperative cardiac events

clinical risk factors, 20

clinical risk predictors, 19
African Americans, renovascular hypertension,

262
Albumin, 52

lower extremity amputations, 403
Alcohol, stroke, 440, 442
Alfimeprase, hemorrhage prevention, 506
Aliasing, 7
Allograft, aortic graft infections, 330-331

i

2

I

■c

701

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

702

INDEX

Alopecia, unfractionated heparin (UFH), 183
Alpha-antiplasmin, 162
Alpha-methyldopa, adverse effects, 237
Alteplase

clinical trials, safety, 503-504
hemorrhage prevention, 505, 506
Ambulation, lower extremity amputations,

405-406
Ambulatory ECG (Holter) monitoring, 21
American College of Cardiology /American
Heart Association (ACC/AHA) Task
Force on Practice Guidelines, 1 8
American Society of Anesthesiologists (ASA),

34
AMI, 16, 18
Aminoglycosides, 55
diabetic foot, 384
side effects, 384-385
Aminophylline, 23
Amputations

failed transmetatarsal, diabetic foot, 394-396
guillotine, diabetic foot, 398
lower extremities, ambulation inability,
405-406
Anastomotic aneurysms, 139-152
aortic grafts, 255
aortoiliac, management, 146-149
arteriogram, 144

clinical presentation and diagnosis, 142-143
etiology and pathogenesis, 139-140
femoral

dacron grafts, 93-96
management, 146
femorofemoral graft, 300
iliac, management, 147-148
incidence, 142
management, 143-151
endovascular repair, 149
operative repair, 145-146
visceral patch following TAAA surgery,
149-155
outcome, 151
prevention, 151-152
silk sutures, 118
surveillance, 143
umbilical vein grafts, 125
Anastomotic neointimal hyperplasia, expanded

PTFE grafts, 111-113
Anastomotic pannus ingrowth, 85
Anastomotic pseudoaneurysms

abdominal aortic aneurysm (AAA), 257
expanded PTFE grafts, 113-1 14

Anastomotic stenosis

dacron grafts, 99-100

failing autogenous grafts, 346

umbilical vein grafts, 125
Ancure/Guidant device, fractures, 649-651
Anesthesia, postoperative pulmonary function,

38
AneuRx device, 638, 639

failure, 644

fracture, 648
Aneurysms, see specific aneurysms
Angioedema, 520
Angiogenesis therapy

adenoviral vectors, 695-696

complications, 691-696

increased vascular permeability, 692-693

mortality, 696
Angiography, 55

carbon dioxide, 57

color power, 6

impotence, 243
Angioguard filter, 620
Angioplasty, see also specific angioplasty

arterial perforation, 598-602

arterial spasms, 605-606

arterial thrombosis, 605-606

balloon rupture, 606-609

complications, procedure site, 598

device embolization, 609-61 1

equipment failure, 606-609
Angiosarcoma, dacron grafts, 102-103
Angiotensin-converting enzyme (ACE),
inhibitors, 57

plasma renin activity (PRA), 262
Angiotensin II receptor antagonists, 57
Ankle-brachial index (ABI), failing autogenous

grafts, 341
Ankle-brachial systolic pressure, 3, 358
Ankle disarticulation, diabetic foot, 397
Antibiotics

aortic graft infections, 322, 326

diabetic foot, 382-385
Anticardiolipin antibodies, infrainguinal graft

occlusions, 357
Anticoagulants

cerebral hyperfusion syndrome, 477

deep venous thrombosis (DVT), 542

intimal hyperplasia, 76
Antiphospholipid antibodies, 169-170
Antiphospholipid antibody syndrome, 193
Antiplatelets, intimal hyperplasia, 76
Antithrombin (AT), 179

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

INDEX

703

Antithrombin deficiency, 160-162
clinical presentation, 160-161
diagnosis, 161-162
Antithrombin III. see Antithrombin (AT)
Anti-tissue-type plasminogen activator (t-PA),

166
Aorta

cross-clamp time, spinal cord ischemia, 223
distal perfusion, 59
pathology, spinal cord ischemia, 223
plication, 328
reconstruction, 58
repair, 54-56

surgery, acute renal failure, 50
Aortic aneurysm
surgery, 34-35
type II, 224
Aortic bifurcation graft limb occlusions,
279-290
diagnosis and evaluation, 283-284
etiology, 280-283
early failure, 280-283
late failure, 283
incidence, 279
management, 284-289

inflow restoration, 284-288
outflow revision, 288-289
results, 289-290
patency, 279-280
Aortic cross-clamping, 58, 59
Aortic graft infections, 317-332

allograft and antibiotic-treated prosthetic

grafts, 330-331
clinical presentation, 320
diagnosis, 321-322
extra-anatomic bypass, 324
graft removal, alternative approaches,

331-332
pathogenesis, 317-320
prevention, 322
superficial femoropopliteal veins in situ

replacement, 324-326
treatment, 322-324
Aortic grafts

anastomotic aneurysms, 255
pseudoaneurysms, 255
Aortic para-anastomotic aneurysms, dacron

grafts, 91-93
Aortic reconstruction, gastrointestinal
complications, 211-218
anatomy and pathophysiology, 213-214
arteriography, 217

[Aortic reconstruction, gastrointestinal
complications]

classification, 211-212

clinical manifestations and diagnosis,
214-215

operative techniques and treatment, 217-218

prevention, 216-217
Aortobifemoral bypass graft, 327-329

anastomosis, 288

thrombus, 287

wound complications, 370
Aortoiliac anastomotic aneurysms,

management, 146-149
Aortoiliac knitted graft, dilation, abdominal

aortogram, 87
Aortorenal bypass

complications, 269-270

stenosis, 273
Aplastic anemia, platelet inhibitors, 202
Aprons, radiation exposure, 489
ARDS, 34

ARFS. see Acute renal failure (ARF)
Argatroban, 195-196
Arm, venous outflow obstruction, 550
Arterial allografts, cryopreserved allografts,

126-127
Arterial blood gas (ABG), 36
Arterial dissection, stents, 603-605
Arterial injections, sclerotherapy, 521
Arterial insufficiency, 380
Arterial replacement, cryopreserved allografts,

126
Arterial revascularization, low molecular

weight heparin (LMWH), 188
Arterial spasms, angioplasty, 605-606
Arterial stenosis, vascular access, 418
Arterial thrombosis, angioplasty, 605-606
Arterial wall abnormalities, 380
Arteriography

acute postoperative graft occlusions, 358

aortic graft infections, 321

aortic reconstruction, 217

endoleaks, 669

infrainguinal graft occlusions, 356

renal artery stenosis (RAS), 262
Arteriovenous access grafts, endovascular
intervention, 557-567

complications, 558-567
Arteriovenous anastomosis, sclerotherapy, 518
Arteriovenous fistula (AVF), 557, 582, 583

distal, expanded PTFE grafts, 110

failing autogenous grafts, 341, 346

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

704

INDEX

[Arteriovenous fistula (AVF)]

failing fistulas, 424-425

prosthetic, peripheral vascular access,
415-416
Arteriovenous graft (AVG), 557

expanded PTFE grafts, 106

venous anastomosis, restenosis, 563-564
Arteriovenous malformations (AVM), 583,
584

children, 593-594

complications, anatomic regions, 590-593

embolization agents, 586-590

embolization techniques, 585-586

pulmonary, 590
Artery of Adamkiewicz, 221-222, 686
Artery walls, calcification, 8
ASA, 34. see also Aspirin (ASA)
Ascending infection, diabetic foot, 396-398
Aspergillus, vascular grafts, 308
Aspirin (ASA), 34, 199-200

antiphospholipid antibodies, 170

heparin-induced platelet aggregation, 159

infrainguinal graft occlusions, 357

peripheral arterial angioplasty, 201

pharmacology, 198

side effects, 201-202

stents, 201
Asymptomatic carotid artery stenosis, stroke,

442
Asymptomatic Carotid Atherosclerosis Study

(ACAS), 443
AT, 179

Atarax, 519-520
Atherectomy, 68
Atheroembolism, 55, 59
Atherosclerosis

angiogenesis therapy, 695

failing autogenous grafts, 338-339, 344

innominate artery, 462-464

stroke, 442
Atrial fibrillation

stroke, 441

warfarin, 192
Atrial natriuretic peptide, contrast toxicity,

496
Autogenous grafts, see also Failing autogenous
grafts

long-term patency rate, factors affecting,
349-351
Autogenous vein, bypass conduit, 356
Autologous tissue, arteriovenous
malformations (AVM), 587

Autologous vein graft healing, 67-69
AVF. see Arteriovenous fistula (AVF)
AVG. see Arteriovenous graft (AVG)
AVM. see Arteriovenous malformations

(AVM)
Axial reconstruction, 462
Axillary pullout syndrome, 294
Axillofemoral bypass, infected, 298
Axillofemoral bypass graft, 293-299

axillary anastomosis, 295

axillofemoral graft infection, 297

perigraft seroma, 297-299

proximal anastomotic disruption, 294

structural failure, 299

upper extremity thromboembolic events,
294-296
Axillofemoral graft infection, 297
Axillopopliteal graft, 296

Bacterial biofilms, vascular graft infections,

310-311
Bacteroides, diabetic foot, 382, 383
Balloon angioplasty
arterial healing, 69-70
iliofemoral vein, 545
intimal hyperplasia, 73
Balloon-injured baboon saphenous artery,

70
Balloons

detachable, arteriovenous malformations

(AVM), 587-588
dilatation, 68
BARI, 28

Belcaplermin, diabetic foot, 386
Benadryl, 519-520
Beta blockade, 31-32, 57
Beta-lactamase inhibitors, diabetic foot, 384
Bifurcated Dacron grafts, 462
Biological dressing, diabetic foot, 385
Bird's Nest, 570
Bisoprolol, 32
Bivalirudin, 194, 195

Bleeding, temporary vascular access, 41 1-412
Blood pressure, perioperative, carotid

endarterectomies (CEA), 480-481
Blushing, sclerotherapy, 515
B-mode imaging, failing autogenous grafts,

339-340
Boston Scientific Vanguard endovascular graft,

fatigue, 645
Bovine pericardium, 120
Bovine xenografts, 127

i
2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

INDEX

705

Brachial artery catheterization, endovascular

access complications, 627-628
Brachial plexus, 431
injury, thoracic outlet syndrome (TOS),

433-434
Bradycardia, carotid stents, 617-618
Braided Dacron sutures, 118
Brescia-Cimino grafts, stent deployment,

560-561
Bronchospasm, 520
Bypass vs. Angioplasty Revascularization

Investigation (BARI), 28

CABG. see Coronary artery bypass graft

(CABG)
Calcium channel blockers, 57

contrast toxicity, 496-497
Calculated velocity, 8
Callus, diabetic foot, 388-389
Cancer, see Neoplasia
Candida, vascular grafts, 308
CAPRIE, 199, 202
Captopril
plasma renin activity (PRA), 262
renal scintigraphy, 263
Carbon coating, expanded PTFE grafts, 107
Carbon dioxide angiography, 57
Carcinogenesis, see Neoplasia
Cardiac assessment risk assessment postures,

vascular surgery, 27-3 1
Cardiac output, kidney, 50
Cardiac risk stratification

adjunctive screening methods, 21-27
screening protocol, myocardial scintigraphy,

24
selective approach, 3 1
Cardiac screening efficacy, myocardial

scintigraphy, 26
Cardiopulmonary bypass, spinal cord ischemia

prevention, 228
Carotid angioplasty, cerebral artery, 619
Carotid artery disease, stroke, 442^143
Carotid artery stenosis, 617-618

stroke, 442, 445
Carotid baroreceptor dysfunction, carotid

endarterectomies (CEA), 480-481
Carotid bruit, stroke, 442
Carotid endarterectomies (CEA), 1 19
cerebrovascular protection techniques,
467-472
neurological deficits, 470-472
technically perfect operation, 470

[Carotid endarterectomies (CEA)]

nonneurological complications, 478-481
nonstroke complications, 475-481

neurological complications, 475-478
prophylactic, perioperative stroke,

448-449
wound complications, 370
Carotid patching, expanded PTFE grafts,

119-121
Carotid stents
access, 615-616
bradycardia, 617-618
complications, 615-623
postprocedural, 621-622
procedural, 615-621
deformation, 621
embolization, 618-620
guide catheters, 616
late events, 622
Carotid ulceration, stroke, 442
CARP, 30
Carpal tunnel syndrome, failing fistulas,

425
CASS, 28

Catheter-directed thrombolytic (CDT)
techniques, 545
deep venous thrombosis (DVT),
543-544
Catheters, see also specific catheters
retrograde graft thrombectomy, aortic
bifurcation graft limb occlusions,
285-287
thrombectomy, acute postoperative graft

occlusions, 358
thrombosis, temporary vascular access, 413
CDT, 543-545

CEA. see Carotid endarterectomies (CEA)
Cefoxitin, lower extremity amputations,

infection, 404
Celiotomy, 313
Cell-cycle inhibitors, intimal hyperplasia,

73-75
Cellulitis, subfascial endoscopic perforator vein

surgery (SEPS), 530
Cephalic veins, failing fistulas, 425
Cephalosporin, vascular access, 420
Cerebral artery, carotid angioplasty, 619
Cerebral hyperfusion syndrome, carotid
endarterectomies (CEA), 477-478
Cerebral ischemia
carotid stents, 616
cerebral hyperfusion syndrome, 477

i
2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

706

INDEX

Cerebral spinal fluid (CSF) drainage, spinal

cord ischemia prevention, 227
Cerebrovascular disease, 16
Cerebrovascular testing, 10-11
Cervical incisions, wound complications,

369-370
Charcot foot, 381,382
CHF. see Congestive heart failure (CHF)
Children, arteriovenous malformations

(AVM), 593-594
Cholecystectomy, 313
Chronic obstructive pulmonary disease

(COPD), 15, 34, 38, 255
Cigarettes, see Smoking
Cilostazol

pharmacology, 198

side effects, 202

stents, 200
Clavulanic acid, diabetic foot, 384
Clopidogrel

peripheral arterial angioplasty, 201

pharmacology, 198-199

side effects, 201-202

stents, 201
Clopidogrel in Unstable Angina to Prevent
Recurrent Ischaemic Events Trial
(CURE), 200, 201
Clopidogrel versus Aspirin in Patients at Risk
of Ischaemic Events (CAPRIE), 199,
202
Clostridium, diabetic foot, 383
Cocaine abuse, stroke, 442
Cocket perforators, 528-529
Colitis, see Ischemic colitis
Collagenase Santyl, diabetic foot, 386
Collar badges, radiation exposure, 490
Colon

ischemia

aortic reconstruction, 212-213
clinical manifestations, 214-215

paralytic ileus, 212
Color Duplex ultrasonography, 10

potential pitfalls, 7
Color-flow imaging, 5
Color power angiography, 6
Compression stockings, 330
Computed tomography (CT)

aortic graft infections, 321

dacron grafts, 91

endoleaks, 666-667

single-photon-emission, 23
Congestive heart failure (CHF), 33-34

[Congestive heart failure (CHF)]

arteriovenous access grafts, endovascular

intervention, 558
failing fistulas, 424
Conjugated estrogens, bleeding, 412
Continuous positive airway pressure (CPAP),

38
Contrast agents, 55
allergic reactions, 558
toxicity, 492-497

cardiac reactions, 493-494
hematological reactions, 493
nephrotoxicity, 494-495
prevention, 495-496
systemic reactions, 492^193
Coons (Cook) hydrophilic dilators, 626
COPD. see Chronic obstructive pulmonary

disease (COPD)
Cordis LP device, 638-639, 640
Cordis TrapEase, 570

Coronary artery bypass graft (CABG), 28, 525,
536
stroke, 445

wound complications, 372-373
Coronary artery disease, 463

peripheral vascular disease, 15-18
selective screening, 29-30
stroke, 442
Coronary Artery Revascularization for Elective

Vascular Surgery (CARP) trial, 30
Coronary Artery Surgery Study (CASS), 28
Coronary revascularization, 28
preoperative, protective effect, 29
preoperative percutaneous, perioperative
outcomes, 28
Corpus cavernosography, 243
Corticosteroids, 521
Coumadin, see Warfarin
CPAP, 38
Cranial nerve injury, carotid endarterectomies

(CEA), 475-477
Crawford's inclusion technique, 152
Creatinine, contrast toxicity, 495
Crossover grafts, expanded PTFE grafts, 122
Cryopreservation, 331
Cryopreserved allografts, 125-127
CSF drainage, spinal cord ischemia prevention,

227
CT. see Computed tomography (CT)
Cuffed catheters, infection, temporary vascular

access, 410-411
CURE, 200, 201

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

INDEX

707

Cutaneous necrosis, sclerotherapy, 517-519
Cutaneous thermoregulatory shunt,

sclerotherapy, 519
Cystometrography, 242

Dacron grafts, 84-103

aortic graft infections, 319

bacterial infections, 310

complications, 86-103

anastomotic stenosis, 99-100
aneurysms, 91-99
dilatation, 86-90
neoplasia, 102-103
ureteric obstruction, 100-102

fabrication, 84-85

healing characteristics, 85-86

vs. PTFE bifurcation and femoropopliteal
grafts, 114-116

rupture, 97-98
Dalteparin, FDA approved indications, 186t
Danaparoid

heparin-induced thrombosis, 158
DAVF, expanded PTFE grafts, 1 10
DDAVP, 412
Deep venous thrombosis (DVT), 9, 1 1

bivalirudin, 195

lower extremity amputations, 406

low molecular weight heparin (LMWH),
197

lysis, complications, 542-545

sclerotherapy, 522

subfascial endoscopic perforator vein surgery
(SEPS), 529
Dermal gangrene, 193
Desmopressin (DDAVP), bleeding, 412
Detachable balloons, arteriovenous

malformations (AVM), 587-588
Device embolization, angioplasty, 609-611
Device failure, 633-654

clinical significance, 654

etiology, 652-654

modes, 641-652
Dextrose, 518
Diabetes mellitus

aortic reconstruction, 212

contrast toxicity, 496

failing autogenous grafts, 345

late graft occlusions, 360

stroke, 441
Diabetic foot

antibiotics, 382-385

case management, 388-398

[Diabetic foot]

ascending infection, 396-398
chronic neurotrophic ulcer, 388-390
disordered architecture, 388
failed transmetatarsal amputation,

394-396
gangrene, 393-394
maggots, 392-393
web space infection, 390-391
management, 379-398
prevalence, 379
wound care, 385-387
Dialysis

cryopreserved allografts, 126
expanded PTFE grafts, 105-106, 1 1 1
history, vascular access, 418-422
temporary access catheters, 410
Digits, ulceration, diabetic foot, 388
Dilatation
balloons, 68
dacron grafts, 86-90
complications, 91
computed tomography, 91
diagnosis, 90
etiology, 88-90
management, 98-99
ultrasound, 90-91
prosthetic grafts, anastomotic aneurysms,

141
umbilical vein grafts, 124-125
Diphenhydramine (Benadryl), 519-520
Dipyridamole

antiphospholipid antibodies, 170
heparin-induced platelet aggregation,
159
Dipyridamole-thallium scintigraphy, 23, 30
Direct thrombin inhibitors, 193-197
Direct vertebral artery reconstruction,

464-465
Distal anastomosis site, failing autogenous

grafts, 344-345
Distal aortic perfusion, spinal cord ischemia

prevention, 228
Distal arteriovenous fistula (DAVF), expanded

PTFE grafts, 110
Distal embolization

femorofemoral graft, 300
upper extremity, axillofemoral bypass, 296
Distal vein cuff (DVC), expanded PTFE grafts,

110
Diuretics, contrast toxicity, 495
DNA gyrase, 384

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

708

INDEX

Dobutamine stress echocardiography (DSE),

22
Dopamine, 56-57, 58

contrast toxicity, 496
Doppler frequency shift frequency, 7
Doppler spectral analysis, failing autogenous

grafts, 339-340
Doppler ultrasound, 1

factors affecting, 4

frequency, 5
Drug abuse, stroke, 440
DSE, 22

Duplex-acquired velocity, 3
Duplex ultrasonography, 1, 2, 3, 10

color, 10

endoleaks, 668

factors affecting, 4

failing autogenous grafts, 339-340, 340-344

potential pitfalls, 7
DVC, expanded PTFE grafts, 1 10
DVT. see Deep venous thrombosis (DVT)

EAST, thoracic outlet syndrome (TOS), 435
EGFR, 71

intimal hyperplasia, 75
Electroencephalography, temporary indwelling
shunt, carotid endarterectomies (CEA),
469
Electromyography, obturator bypass, 300
Elevated arm stress test (EAST), thoracic outlet

syndrome (TOS), 435
Embolism, see also Pulmonary embolism;
Venous thromboembolism (VTE)
recurrent, vena caval filters, 574
stroke, 444
Embolization, see also Distal embolization
agents, arteriovenous malformations

(AVM), 586-590
carotid stents, 618-620
coils, arteriovenous malformations (AVM),

588
device, angioplasty, 609-611
distal

femorofemoral graft, 300
upper extremity, 296
intraoperative, carotid endarterectomies

(CEA), 470
techniques, arteriovenous malformations

(AVM), 585-586
thrombus, 565-566
venous malformations, 593
EMLA, 515

Endarterectomy, failing autogenous grafts, 346
Endoleaks, 659-678
avoiding, 672-673
device failure, 652-654
diagnosis, 663-671
arteriography, 669
CT scan, 666-667
duplex ultrasound, 668
intraprocedural, 663-666
magnetic resonance arteriography (MRA),

669-671
postoperative, 666
future, 677-678
management, 673-677
thoracic aortic aneurysm repair, 684-685
time of occurrence, 671-672
types, 660-663
Endoscopic dissector/retractor, 532
Endoscopic saphenous vein harvesting,
531-537
complications, 535-537
surgical techniques, 531-535
Endoscopic vein harvesting (EVH), 525

wound complications, 373
Endothelialization, expanded PTFE grafts,

108-109
Endovascular access complications, 625-630
access failure, 625-626
hemodynamic compromise, 628-629
stenosis, 626-627
vessel compromise, 628
Endovascular grafts

commercially fabricated, 634-640
implantation, impotence, 245
individually fabricated, 634
Endovascular revascularization, impotence,

245
Endovascular stents, complications, 559-560
End-stage renal disease (ESRD), 557
Enoxaparin, FDA approved indications, 186t
Enterobacter

diabetic foot, 384
vascular grafts, 307
Enterococcus, diabetic foot, 383
Epidermal growth factor receptor (EGFR),
71
intimal hyperplasia, 75
Erection, physiology, 236-239
neurophysiology, 236-237
penile blood supply, 237-239
psychic influences, 239
Erythropoietin, vascular access, 419

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

INDEX

709

Escherichia coli

aortic graft infections, 319, 320

vascular grafts, 307, 310
ESRD, 557

Estrogens, conjugated, bleeding, 412
EVH, 525

wound complications, 373
EVT device

failure, 644

fractures, 648, 649-651

metal fractures, 640-641
EVT/Guidant-Ancure device, 634-637, 636
Examination technique, 6-8
Exercise-stress ECG testing, 21
Expanded PTFE grafts, 103-122

carotid patching, 119-121

complications, 111-114

anastomotic neointimal hyperplasia,

111-113
anastomotic pseudoaneurysm, 113-1 14

dacron PTFE bifurcation and
femoropopliteal, 114-116

dialysis, 111

fabrication, 104

healing characteristics, 104-106

occlusive disease, 121

perigraft seroma, 116-117

small-diameter graft patency enhancement,
107-111

suture line failure, 117-119

venous system, 121-122
Extra-anatomic bypass

aortic graft infections, 324

axillofemoral bypass graft, 293-299

obturator bypass, 300-301

thoracofemoral bypass, 301
Extremities, see also Lower extremities; Upper
extremities

arteriovenous malformations (AVM),
591-592

Factor Xa, 188

Factor Xa inhibitors, 197-198

Failed transmetatarsal amputations, diabetic

foot, 394-396
Failing autogenous grafts

detection and management, 337-352

diabetes mellitus, 345

distal anastomosis site, 344-345

etiology, 337-338

graft surveillance
protocol, 339-340

[Failing autogenous grafts]
technique, 340-344

hemodynamic monitoring, 339-340

indications, 339

long-term changes, 344

management, perioperative results, 348-349

operative management, 345-347

modification during primary procedure,
345-347

perioperative failure, 338

postoperative failure, 338-339

secondary procedures, 347

vein graft diameter, 345
Failing fistulas

arteriovenous fistula takedown, 424-425

cannulation difficulty, 424

carpal tunnel syndrome, 425

cephalic veins, 425

congestive heart failure, 424

high venous pressure, 423

neurological sequelae, 425

poor arterial inflow, 423^124

vascular access, 422-427
Feet, see Diabetic foot
Femoral anastomotic aneurysms

dacron grafts, 93-96

management, 146
Femoral artery disease, endovascular access

complications, 629
Femoral catheters, temporary access, 409
Femoral triangle, lymphatic network, 370
Femorofemoral graft, 299-300
Femoropopliteal bypass graft, thrombolysis-

induced bleeding, 505
Femoropopliteal grafts, 328, 329

vs. PTFE bifurcation and dacron grafts,
114-116
Fenoldopam, 58

contrast toxicity, 496

pararenal aorta, 255
FGF-I, 691-696

intimal hyperplasia, 76
Fibrin, 502

diabetic foot, 386
Fibrinolytic system defects, 162-166

clinical presentation, 163-166

testing, 166
Fibroblast growth factor I (FGF-I), 691-696

intimal hyperplasia, 76
Fibrointimal hyperplasia, failing autogenous

grafts, 338-339, 347
Filter wire EX, 620

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

710

INDEX

Fistulas, see specific fistulas
Fistulogram, vascular access, 421-422
Flares, sclerotherapy, 515
Flexion contractures, lower extremity

amputations, 405
Floating table, radiation exposure, 491
Fluoroscope

patient positioning, radiation exposure,
486-487

radiation exposure, 484-485
Folic acid, hyperhomocysteinemia, 172
Folliculitis, sclerotherapy, 516-517
Fondaparinux sodium, 197-198
Foot ulcers, see Diabetic foot
Fractures, see Metallic fractures
Fragment X, 506
Furosemide, 56-57
Fusobacterium, diabetic foot, 383

Gadolinium, 57
Gangrene

diabetic foot, 393-394

vascular graft infections, 306
Gastrointestinal tract, arteriovenous

malformations (AVM), 591
Gelfoam, arteriovenous malformations

(AVM), 587
Gelseal, 85
Gelweave, 85

Glossopharyngeal nerve, 476
Gloves, radiation exposure, 490
Goldman cardiac risk criteria, 34
Gore Excluder, 638
Gore TAG prostheses

fabric fatigue, 651-652

failure, 644
Grafts, see also specific grafts

defects, anastomotic aneurysms, 141

late occlusions, treatment, 359-363

late stenosis, aortorenal bypass, 270

thrombectomy, late graft occlusions, 361-362
Grasping forceps, 611
Grays (Gy), 483
Greater auricular nerve, 476
Greenfield filters, 570

incomplete expansion, 573

migrated, 575f
Groin

aortic graft infections, 317-318

incisions

lymphoceles, 371-372

wound complications, 370-372

Growth factors, intimal hyperplasia, 75
Guide catheters, 617

carotid stents, 616
Guidewires

perforation, 598-602

vascular snare, 611

vena cava filters, 571-574
Guillotine amputations, diabetic foot, 398
Gunther-Tulip, 570
Gy, 483

HBO, wound infection, 374

Headache, platelet inhibitors, 202

Heart and Estrogen-Progestin Replacement

Study (HERS), 442
Heart valve replacement, warfarin, 192
Hemangiomas, 582
Hemashield graft, 85, 115-116
Hematoma

low molecular weight heparin (LMWH),

189
percutaneous transluminal angioplasty

(PTA), renal hypertension, 264
perigraft, peripheral vascular access, 416
postoperative, carotid endarterectomies

(CEA), 479
scrotal, femorofemoral graft, 299
vascular graft infections, 309
Hemodialysis catheters, bleeding, 41 1^412
Hemorrhage
argatroban, 196
intracerebral, 440
intracranial
stroke, 442

thrombolytic therapy, 544
low molecular weight heparin (LMWH), 189
platelet inhibitors, 201-202
prevention, thrombolysis, 504-506
unfractionated heparin (UFH), 182-183
warfarin, 192
Hemothorax, temporary vascular access,

412-413
Heparin, 58, 358. see also Low molecular weight
heparin (LMWH); Unfractionated
heparin (UFH)
bleeding, 412

heparin-induced thrombosis, 158
infrainguinal graft occlusions, 357
intimal hyperplasia, 76
pararenal aorta, 255
peripheral arterial angioplasty, 201
stents, 201

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

INDEX

711

Heparin bonding, expanded PTFE grafts,

107-108
Heparin-induced platelet aggregation, 158-159

vascular access, 419
Heparin-induced thrombocytopenia, 183
Heparin-induced thrombosis, 155-158
clinical presentation, 156-157
diagnosis, 157
treatment, 158
HERS, 442

Hexamethonium, adverse effects, 237
Hirsutism, sclerotherapy, 517
Hirudin, 194
Holter monitoring, 21
Homocysteine, stroke, 442
Hooded grafts, expanded PTFE grafts, 1 1 1
Horner's syndrome, thoracic outlet syndrome

(TOS), 434
Human umbilical vein (HUV) grafts, see

Umbilical vein grafts
HUV grafts, see Umbilical vein grafts
Hydrocolloid dressing, diabetic foot, 385
Hydrogels, diabetic foot, 386
Hydroureteronephrosis, 103
aortic graft infections, 320
Hydroxyzine (Atarax), 519-520
Hyperbaric oxygen therapy (HBO), wound

infection, 374
Hypercoagulable states
identification, 172-174
vascular access, 419
vascular graft thrombosis, 155-174
Hyperglycemia, 380

diabetic foot, 381,384
Hyperhomocysteinemia, 171-172
Hyperlipidemia, stroke, 441
Hyperperfusion cerebral hemorrhage, carotid

stents, 621-622
Hyperpigmentation, sclerotherapy, 512,

513-515
Hypertension
impotence, 246

proximal, spinal cord ischemia, 224
renovascular, 261

African Americans, 262
surgical revascularization, 267-274
stroke, 441
venous

superficial femoropopliteal vein harvest,

330
vascular access, 421
Hypertonic saline, 518

Hypogastric artery, avulsion, 628
Hypoglossal nerve, 476
Hypoperfusion, aortic reconstruction,

217-218
Hypotension, 55

angiogenesis therapy, 693

aortic reconstruction, 217-218

carotid stents, 617-618

stroke, 443-444

vascular access, 419
Hypothermia, spinal cord ischemia prevention,
227-228

ICA, 9

IL-1, vascular grafts, 311-312

Iliac anastomotic aneurysms, management,

147-148
Iliac arteries, 240

occlusive disease, 684
Iliac compression syndrome, 549
Iliac conduits, endovascular access

complications, 627
Iliofemoral venous outflow channel stents,

complications, 545-550
IMA. see Inferior mesenteric artery (IMA)
Impotence, 235-247

angiography, 243

diagnosis, 239-240

history, 240-246

intracavernous papaverine injection, 243

neurogenic, 240

neurological testing, 242-243

nocturnal penile tumescence, 240-241

noninvasive vascular testing, 241-242

organic, 240

prevention, 244-245

treatment, 246
Infection

anastomotic aneurysms, 140

axillofemoral bypass, 298

carotid endarterectomies (CEA), 479-480

lower extremity amputations, 404

peripheral vascular access, 416

skin ulcers, vascular graft infections, 306

temporary vascular access, 410-414

vascular access, 420-421
Inferior mesenteric artery (IMA), 211, 213-214

ligation, 217

reimplantation, 218
Inflammatory response, thoracic aortic

aneurysm repair, 688
Inflow stenosis, vascular access, 418

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

712

INDEX

Infrainguinal graft occlusions

acute postoperative graft occlusions,

treatment, 357-359
late graft occlusions, treatment, 359-363
pharmacological management, 357
postoperative graft surveillance, 357
prevention, 355-357
arteriography, 356
conduit, 356
operation, 355-356
treatment, 355-363
Innominate artery

atherosclerosis, 462-464
mortality, 463

technical considerations, 462-463
transthoracic bypass, 462
In-stent restenosis, iliofemoral venous outflow

channel stents, 546
Instrument calibration, 3
Intercostal artery reimplantation, spinal cord

ischemia prevention, 226-227
Interleukin 1 (IL-1), vascular grafts, 311-312
Intermittent positive-pressure breathing

(IPPB), 38
Internal carotid artery (ICA), 9
Interpretative pitfalls, 2, 8-9
Intimal hyperplasia, 67-77, 536, 559-560
molecular mechanisms, 70-73
therapy, 68, 73-76

antiplatelets and anticoagulants, 76
cell cycle-directed, 73-75
growth factor inhibition, 75-76
mechanical, 73
Intracavernous papaverine injection,

impotence, 243
Intracerebral hemorrhage, 440
Intracranial hemorrhage
stroke, 442

thrombolytic therapy, 544
Intraoperative embolization, carotid
endarterectomies (CEA), 470
Intravenous hydration, 57
IPPB, 38

Ipsilateral saphenous vein, 349
Ischemia, see also Cerebral ischemia; Spinal
cord ischemia
colon, aortic reconstruction, 212-213
kidney, 54

late graft occlusions, 359-360
lower extremity, 16
paraspinus muscle, 687
thoracic aortic aneurysm repair, 685-688

Ischemic colitis

abdominal aortic aneurysms stent-graft

repair, 215-216
aortic reconstruction, 212-213
clinical manifestations, 214-215
operative techniques and treatment, 217-218

Juxtarenal aortic operations, 250-253
complications, 253-257
supraceliac clamping, 254

Kidneys

arteriovenous malformations (AVM), 591

cardiac output, 50

dysfunction

categories, 52-56
diagnosis and treatment, 59-60
vascular surgery, 60-61
function, protective strategies, 56-57
impaired function, percutaneous
transluminal angioplasty (PTA)
renal hypertension, 264-265
ischemia, 54

neuroendocrine modulators, 51-52
normal function, 50-52
toxic injury, 55-56
Klebsiella, vascular grafts, 307
Klippel-Trenaunay syndrome, 584, 585, 592,

594
Knitted Dacron grafts, 84

Lactamase inhibitors, diabetic foot, 384

Lacunar infarction, 443

Lasix, pararenal aorta, 255

Late graft occlusions, treatment, 359-363

graft thrombectomy, 361-362

ischemia, 359-360

new autogeneous vein graft, 362-363

percutaneous, 360-361
Late graft stenosis, aortorenal bypass, 270
Lepirudin, 194-195

LifeSite Hemodialysis Access System, 411
Linton patches, expanded PTFE grafts, 110
Liquid adhesives, arteriovenous malformations

(AVM), 588-589
Liver

arteriovenous malformations (AVM), 591

supraceliac clamping, 255
LMWH. see Low molecular weight heparin

(LMWH)
Loop of Henle, 51
Lost Palmaz stents, 548

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

INDEX

713

Lower extremities
amputations

ambulation inability, 405^106
complications, 401^107
deep venous thrombosis (DVT), 406
flexion contractures, 405
infection, 404
long-term outlook, 406
mortality, 402

phantom limb pain, 404-405
pulmonary embolism, 406
wound healing failure, 402-404
compartment syndromes, 330
incisions, wound complications, 372-373
ischemia, 16
Low molecular weight heparin (LMWH),
184-190
clinical application, 186-189
complications and failures, 189-190
FDA approved indications, 186t
pharmacology, 185-186
Lymphatic injury, supraaortic trunks (SAT)

extra-anatomical repairs, 460
Lymphoceles

groin incisions, 371-372
vascular graft infections, 309
Lytic therapy, late graft occlusions, 360-361

Maggots, diabetic foot, 392-393
Magnetic resonance arteriography (MRA),

endoleaks, 669-671
Magnetic resonance imaging (MRI), aortic

graft infections, 321
Mahurkar catheters, temporary access, 409, 410
Malignancy, see Neoplasia
Mandibular nerve, marginal, 476
Mannitol, 56-57, 58

contrast toxicity, 495

pararenal aorta, 255
Marginal mandibular nerve, 476
Mediastinal malignancy, 121-122
MEGS stent graft, failure, 644
MEP, 224-226
Mesenteric vessels, endarterectomy, incision,

250
MET, 18

Metabolic equivalents (MET), 18
Metallic fractures, 649-651

EVT device, 640-641
Methicillin-resistant Staphylococcus aureus

aortic graft infections, 318

vancomycin, 313

Methylxanthines, 25

Microcolony formation, vascular graft

infections, 310-311
Microcracks, 651
Microspheres, arteriovenous malformations

(AVM), 587
Miller cuffs, expanded PTFE grafts, 1 10
Minocycline, 514

Motor evoked potentials (MEP), 224-226
MRA, endoleaks, 669-671
MRI, aortic graft infections, 321
Mycobacterium, vascular grafts, 308
Myocardial infarction, 459-462

percutaneous transluminal angioplasty
(PTA), renal hypertension, 267

postoperative, 17
Myocardial ischemia, silent, 18
Myocardial scintigraphy, 25

cardiac risk stratification, 24

cardiac risk stratification screening protocol,
24

cardiac screening efficacy, 26

interpretation maximization, 26
Myoglobinuria, 55

Nafcillin, vascular access, 420
NAIS, 324-326

disadvantages, 330
NASCET, 443
National Veterans Administration Surgical

Quality Improvement Program, 35
NBCA, arteriovenous malformations (AVM),

588-590
N-butylcyanoacrylate (NBCA), arteriovenous

malformations (AVM), 588-590
Neck pain, carotid stents, 621
Necrotizing fasciitis, diabetic foot, 396-397
Negative predictive values (NPV), 2
Neoaortoiliac system (NAIS), 324-326

disadvantages, 330
Neoplasia

angiogenesis therapy, 694-695

mediastinal, 121-122

SVC syndrome, 551
Nephrotoxicity, angiogenesis therapy, 694
Nerve injury

innominate artery atherosclerosis, 464

sclerotherapy, 522

supraaortic trunks (SAT), extra-anatomical
repairs, 460
Neurogenic impotence, 240
Neuro shield filters, 620

i
2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

714

INDEX

Neutropenia, platelet inhibitors, 202
New autogeneous vein graft, late graft

occlusions, 362-363
New or worsening wall motion abnormalities

(NWMA), 22
Nifedipine, 57
99mTc sestamibi, 23, 24
Nitric oxide synthase, 693
Nitroglycerin, 605

Nocturnal penile tumescence, 240-241
Nonanatomic renal artery bypass,

complications, 272
Noninvasive vascular testing, pitfalls, 1-12

classification, 2
North American Symptomatic Carotid

Endarterectomy Trial (NASCET),

443
NPV, 2
Nutritional assessment, wound complications,

367
NWMA, 22
Nyquist limit, 7

Obesity, stroke, 441

Obturator bypass, 300-301

Obturator nerve blocks, obturator bypass,

300
Occlusive disease, expanded PTFE grafts, 121
Olympus endoscope, 526
Oral contraceptives, stroke, 441
Organic impotence, 240
Osteomyelitis, diabetic foot, 381
Osteoporosis

low molecular weight heparin (LMWH),
189

unfractionated heparin (UFH), 183
Outflow stenosis, vascular access, 418^419

PAC, 32-33, 58

pararenal aorta, 255
Paclitaxel (taxol), 74
Pain, sclerotherapy, 515-516
Palmaz stents, 549, 634

crushed, 622

fracture, 647

lost, 548
Panafil, diabetic foot, 386
Pancreatitis

pararenal aorta, 255

postoperative, aortic reconstruction, 212
Pancuronium, 38-39
Papaverine, 605

Paraplegia, 685-686

Pararenal aneurysm, operative results, 251

Pararenal arteries, reconstruction,

complications, 249-257
Paraspinus muscle, ischemia, 687
Paravisceral aorta, endarterectomy, incision,

250
Parodi-type handmade balloon expandable

stent graft, 635
Particles, arteriovenous malformations (AVM),

587
Patency

direct vertebral artery reconstruction, 465
innominate artery atherosclerosis, 464
supraaortic trunks (SAT), extra-anatomical
repairs, 460^162
PCC, 192
PCI, low molecular weight heparin (LMWH),

188
PDGF, 71

intimal hyperplasia, 75
Penis

autonomic nerve supply, 245
blood supply, 237-239
revascularization, 246
swelling, femorofemoral graft, 299
Peptostreptococcus, diabetic foot, 383
Peptostreptococcus magnus, diabetic foot, 382
PercuSurge balloon wire, 620
Percutaneous coronary intervention (PCI),
low molecular weight heparin
(LMWH), 188
Percutaneous transluminal angioplasty (PTA),
10,616
anastomotic recoil, 561
bivalirudin, 195
complications, 559-560
failing autogenous graft, 347
low molecular weight heparin (LMWH),

188
platelet inhibitors, 200
renal hypertension, complications, 263-267
Percutaneous transluminal coronary

angioplasty (PTCA), noncardiac
surgery, 28
Percutaneous venous endovascular procedures,

indications, 541
Perigraft fibrous tissue, 85
Perigraft hematoma, peripheral vascular access,

416
Perigraft seroma, axillofemoral bypass,
297-299

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

INDEX

715

Perioperative blood pressure, carotid

endarterectomies (CEA), 480^81
Perioperative outcomes, preoperative
percutaneous coronary
revascularization, 28
Perioperative stroke

combined carotid/coronary surgery, 448-449
etiology, 443-445
minimizing, 447^149

intraoperative management, 447-448
patient selection, 447
prophylactic carotid endarterectomy,
448-449
Peripheral arterial angioplasty, platelet

inhibitors, 201
Peripheral arterial bypass, warfarin, 192
Peripheral arterial disease, platelet function

inhibitors, 199
Peripheral arterial Doppler studies, 6
Peripheral arterial testing, 9-10
Peripheral thrombolysis, 501-508
Peripheral vascular disease, coronary artery

disease, 15-18
Peripheral venous Doppler studies, 6
Peripheral venous testing, 11-12
PFT, 36
Phantom limb pain, lower extremity

amputations, 404-405
Photoplethysmographic techniques, 1 1
Phrenic nerve, 431
Plaque

failing autogenous grafts, 344
stroke, 441
Plasma renin activity (PRA)
ACE inhibitors, 262
captopril, 262
Plasminogen, 162-163
Plasminogen activator, 502
Platelet-derived growth factor (PDGF), 71

intimal hyperplasia, 75
Platelet function inhibitors, 198-202
clinical applications, 199-201
complications and failures, 201-202
pharmacology, 198-199
Platelet glycoprotein Ilb/IIIa inhibitors,

pharmacology, 199
Plethysmography, 1, 2
Pneumothorax, temporary vascular access,

412-413
Polyiodinated iodine, 518
Polymicrobial infection, diabetic foot, 381
Polypropylene sutures, 118-119

Polytetrafluoroethylene (PTFE)
aortic graft infections, 319
bacterial infections, 310
bifurcation grafts vs. dacron grafts and

femoropopliteal grafts, 114-116
grafts, 69

stent deployment, 559-560
structural failure, 299
Polyurethane grafts, 122-124
Polyvinyl alcohol (PVA) particles,

arteriovenous malformations (AVM),

587
Portocaval shunt, expanded PTFE grafts,

122
Positive predictive values (PPV), 2
Postoperative bypass graft occlusion,

prevention, 356
Postoperative causalgia, thoracic outlet

syndrome (TOS), 434
Postoperative hematoma, carotid

endarterectomies (CEA), 479
Postoperative myocardial infarction, 17
Postoperative pancreatitis, aortic

reconstruction, 212
Postoperative pulmonary function, anesthesia,

38
Postrenal dysfunction, 53-54
Postsclero therapy neovascularization,

sclerotherapy, 515
Power injector, radiation exposure, 491
PPV, 2
PRA

ACE inhibitors, 262
captopril, 262
Precuffed grafts, expanded PTFE grafts, 1 1 1
Prednisone, antiphospho lipid antibodies, 170
Preoperative coronary revascularization,

protective effect, 29
Preoperative percutaneous coronary

revascularization, perioperative

outcomes, 28
Prerenal dysfunction, 52-53
Pressure cuffs, 3

width, 3
PREVENT, 74
Procedural pitfalls, 2, 3-8
Profunda femoris, 288-289
Project of Ex-vivo Vein Graft Engineering via

Transfection (PREVENT), 74
Prophylactic carotid endarterectomy,

perioperative stroke, 448-449
Propranolol, adverse effects, 237

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

716

INDEX

Prosthetic arteriovenous fistula, peripheral

vascular access, 41 5^416
Prosthetic grafts

antibiotic-treated, aortic graft infections,
330-331

dilatation, anastomotic aneurysms, 141

healing, 69
Protamine, carotid endarterectomies (CEA),

479
Protective glasses, radiation exposure, 489^190
Protein C deficiency, 166-168
Protein S deficiency, 168-169
Proteus

aortic graft infections, 319

vascular grafts, 307
Prothrombin complex concentrate (PCC),

192
Proximal aortic clamping, pararenal arteries,

253
Proximal cuff, infection, temporary vascular

access, 4 1 0^4 1 1
Proximal hypertension, spinal cord ischemia,

224
Pseudoaneurysms

aortic grafts, 255

axillofemoral bypass, 299

dacron grafts, 91-93

vascular access, 420
Pseudoendoleaks, 668
Pseudomonas

diabetic foot, 384

vascular grafts, 307
Pseudomonas aeruginosa

aortic graft infections, 319, 320

vascular grafts, 308
Psychogenic impotence, 240
PTA. see Percutaneous transluminal

angioplasty (PTA)
PTCA, 28

PTFE. see Polytetrafluoroethylene (PTFE)
Pulmonary arteriovenous malformations

(AVM), 590
Pulmonary artery catheter (PAC), 32-33, 58

pararenal aorta, 255
Pulmonary dysfunction, pararenal aorta, 255
Pulmonary embolism

lower extremity amputations, 406

temporary vascular access, 414

thrombolytic therapy, 543
Pulmonary function testing (PFT), 36
Pulsatility index, 3
Pulsed Doppler sonography, 243

Pulse-mode fluoroscopy, patient positioning,
radiation exposure, 486-487

Pulse volume recordings (PVR), 3

PVA particles, arteriovenous malformations
(AVM), 587

PVR, 3

Quinolones, diabetic foot, 384

Radiation badges, 490^191
Radiation dosimeters, 491
Radiation exposure, 483-492

biological effects, 483-484

experience, 490

fluoroscopy time, 484-485

measurement units, 483

monitoring, 490-491

patient positioning, 486-487

radiological protection, 489-490

source distance, 485-486
Radicular arteries, 221-222
Radionuclide scanning, aortic graft infections,

321
Radionuclide ventriculography, 22
Rapamycin, 74
RAS, 261-262

arteriography, 262
Reactive oxygen species, contrast toxicity, 496
Recurrent embolism, vena caval filters, 574
Recurrent laryngeal nerve, 476
Reflex erection, 237
Regional renal hypothermia, 58
Renal angioplasty, 262
Renal artery bypass, nonanatomic,

complications, 272
Renal artery occlusion, percutaneous

transluminal angioplasty (PTA), renal
hypertension, 264
Renal artery stenosis (RAS), 261-262

arteriography, 262
Renal failure, thoracoabdominal aortic

surgery, 254
Renal failure fluid shifts, 49-61
Renal insufficiency, juxtarenal aortic

operations, 253-254
Renal parenchymal dysfunction, 54-56
Renal pelvis, obstruction, 53-54
Renal revascularization

complications, 261-274
periprocedural, 263-267
surgical procedures, 267-274

patient selection, 261-262

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

INDEX

717

Renal scintigraphy, captopril, 263
Renovascular hypertension, 261
African Americans, 262
surgical revascularization, 267-274
Reserpine, adverse effects, 237
Residual intact valve leaflets, failing

autogenous grafts, 341
Respiratory failure, noncardiac surgery, clinical

risk index, 35
Respiratory failure index, 36
Restenosis

in-stent, iliofemoral venous outflow channel

stents, 546
percutaneous transluminal angioplasty

(PTA), renal hypertension, 267, 268-269
superior vena cava stents, 551-553
Retained valve leaflet, failing autogenous

grafts, 346
Rethrombosis

acute postoperative graft occlusions, 359
iliofemoral venous outflow channel stents,
complications, 545-546
Retinoblastoma protein, 74
Retinopathy, angiogenesis therapy, 694
Retrograde ejaculation, 235
Retrograde graft thrombectomy, aortic
bifurcation graft limb occlusions,
284-288
Retzius's space, 299
Ring badges, radiation exposure, 490

Sacral arteries, 222
Sacral latency testing, 242
Saphenous vein

endoscopic harvesting, 531-537
harvesting, wound complications, 378
injuries, sclerotherapy, 522
in situ bypasses

correcting inadequate veins, 350-351
patency, 342, 343
surveillance, 341
technical errors, 343
ipsilateral, 349
Sartorius muscle, 325
SAT. see Supraaortic trunks (SAT)
Scalene block, thoracic outlet syndrome (TOS),

435
Scalene muscle, 43 1
Scapular nerves, 431
Scintigraphy

aortic graft infections, 321
dipyridamole-thallium, 23, 30

[Scintigraphy]

myocardial, 25

cardiac risk stratification, 24
cardiac screening efficacy, 26

renal, captopril, 263
Sclerotherapy, 511-522

complications, 513-522

principles, 512-513

systemic allergic reactions, 519-521
Sclerotic veins, failing autogenous grafts,

346
Scrotal hematomas, femorofemoral graft,

299
Segmental pressure measurements, 7
Selective screening, coronary artery disease,

29-30
SEP, 224-226, 242
SEPS. see Subfascial endoscopic perforator

vein surgery (SEPS)
Seroma

expanded PTFE grafts, 116-1 17

perigraft, axillofemoral bypass, 297-299
Serratia, diabetic foot, 384
Sestamibi, 23, 24
Shields, radiation exposure, 490
Sieverts (Sv), 483

Sildenafil citrate (Viagra), impotence, 246
Silent myocardial ischemia, 18
Silk sutures, anastomotic aneurysms, 118
Silver sulfadiazine cream, diabetic foot, 386
Simon-Nitinol, 570
Single-photon-emission computed tomography

(SPECT), 23
Sirolimus (rapamycin), 74
Skin flaps, wound complications, 378
Skin necrosis, subfascial endoscopic perforator

vein surgery (SEPS), 530
Skin ulcers, infected, vascular graft infections,

306
SMA, 213-214

angiogram, 216
SMC, injury, 68
Smoking, 38, 279

infrainguinal graft occlusions, 357

stroke, 442
Smooth muscle cells (SMC), injury, 68
Somatosensory evoked potentials (SEP),

224-226, 242
SPECT, 23
Spectral waveform, 5
Spinal accessory nerve, 476
Spinal arteries, 221

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

718

INDEX

Spinal cord

anatomy, 221-222

blood supply, 221-222
Spinal cord ischemia, 221-230, 686

detection, 224-226

determinants, 223-224

physiology, 222-223

prevention, 226-230
Spleen

arteriovenous malformations (AVM), 591

pararenal aorta, 255
Staphylococcus

anastomotic aneurysms, 140

carotid endarterectomies (CEA), 480

diabetic foot, 383
Staphylococcus aureus

aortic graft infections, 320

diabetic foot, 385

vascular graft infections, 310

vascular grafts, 307
Staphylococcus epidermidis

aortic graft infections, 318, 320

vascular grafts, 308, 311
Statistical pitfalls, 8-9

Steal phenomenon, femorofemoral graft, 300
Steal syndrome, vascular access, 422
Stellate ganglion, 431
Stenosis

duplex scan classification, 340

endovascular access complications, 626-627

failing autogenous graft, 347, 348

failing autogenous grafts, 341

outflow, vascular access, 418-419

PTFE, 562

vascular access, 418
Stent graft fatigue, 641-643
Stentor grafts

fabric fatigue, 651-652

fractures, 650-651
Stentor/Vanguard device, suture breakage,

652
Stents, 68. see also specific stents

arterial dissection, 603-605

arterial healing, 69-70

arterial perforation, 598-602

arterial thrombosis, 605-606

aspirin (ASA), 201

Brescia-Cimino grafts, 559-560

cilostazol, 200

clopidogrel, 201

complications, procedure site, 598

device embolization, 609-611

[Stents]

equipment failure, 606-609

fabric fatigue, 651-652

failure, 643-649

fracture, 549, 641

thoracic aortic aneurysm repair, 688-689

heparin, 201

platelet inhibitors, 201

securing embolized stent, 612

ticlopidine, 200
STIMS, 199

Stomach, paralytic ileus, 212
Strawberry birthmark, 582
Streptococcus

carotid endarterectomies (CEA), 480

diabetic foot, 383
Streptokinase, nonhemorrhagic complications,

506-507
Stress myocardial scintigraphy, 30
Stress scintigraphy, 25
Stroke, 439-450. see also Perioperative stroke

cardiac surgery, 445^146

cerebral hyperfusion syndrome, 477

demographics, 439-440

direct vertebral artery reconstruction, 465

etiology, 440

general surgery, 447

innominate artery atherosclerosis, 463

noncerebrovascular surgery, 443-447

noncerebrovascular vascular surgery,
446-447

percutaneous transluminal angioplasty
(PTA), renal hypertension, 267

platelet function inhibitors, 199-200

risk factors, 440-443

supraaortic trunks (SAT), extra-anatomical
repairs, 460
Subarachnoid hemorrhage, 440
Subclavian artery, 431
Subclavian dialysis catheters, infection,

temporary vascular access, 410-411
Subclavian thrombosis, temporary vascular

access, 413
Subclavian vein, 43 1

catheters, temporary access, 409-410

stents, complications, 550-551
Subfascial endoscopic perforator vein surgery
(SEPS), 525-537

complications, 529-531

techniques, 526-529
two port, 527-528
Sulbactam, diabetic foot, 384

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

INDEX

719

Superficial femoropopliteal vein, 325
harvest, venous hypertension, 330
Superficial thrombophlebitis, sclerotherapy,

521
Superior laryngeal nerve, 476
Superior mesenteric artery (SMA), 213-214

angiogram, 216
Superior vena cava, replacement, expanded

PTFE grafts, 121-122
Superior vena cava, stents, complications,

551-553
Supraaortic trunks (SAT), 457-465
extra-anatomical repairs, 458-462
perioperative complications, 459-462
technical considerations, 459
Supraceliac clamping
juxtarenal aortic operations, 254
liver failure, 255
Suprarenal aorta, operations, 250-253
Suprarenal cross clamp, 54
Suture defects, anastomotic aneurysms, 141
Sutures, expanded PTFE grafts, 117-119
Sv, 483

SVC syndrome, 551
Swedish Ticlopidine Multicentre Study

(STIMS), 199
Swelling, sclerotherapy, 515
Sympathetic nerves, erection, 236-237

TAA. see Thoracoabdominal aneurysm

(TAA)
TAG graft, fatigue, 649
Talent device, 637-638, 640
Talent endovascular system, fatigue, 646
Talent stent graft

fabric fatigue, 651-652

failure, 644
Taxol, 74

Taylor patches, expanded PTFE grafts, 110
Tazobactam, diabetic foot, 384
Tc sestamibi, 23, 24
TED, 483, 491
TEE, 33

pararenal aorta, 255
Telangiectasia, 514

Telangiectatic matting, sclerotherapy, 515
Temporary indwelling shunt, carotid

endarterectomies (CEA), 467-471
TFPI, 180-181
TGF-beta, 71

intimal hyperplasia, 76

wound complications, 366

Theophylline, contrast toxicity, 496
Thienopyridines, pharmacology, 198-199
Thoracic aortic aneurysm repair,
complications, 683-689

anatomically related, 683-684

device-related, 688-689

endoleaks, 684-685

inflammatory response, 688

ischemia, 685-688
Thoracic aortic stent grafts, 640
Thoracic duct, 431
Thoracic outlet, anatomy, 430-432
Thoracic outlet decompression, 551
Thoracic outlet surgery

anatomy, 430-432

complications, 433-437
operative injuries, 433-434

operations, 432-433
Thoracic outlet syndrome (TOS), treatment,

435-437
Thoracoabdominal aneurysm (TAA), 35, 54,
59, 229

anastomotic aneurysms, 142

endovascular repair, 634-641

incisions, 250

stent graft fatigue, 641
Thoracoabdominal aortic aneurysm, 226
Thoracoabdominal aortic surgery, renal failure,

254
Thoraco-bi-iliac bypass, 270-271
Thoracofemoral bypass, 301
Thrombectomy

catheters, acute postoperative graft
occlusions, 358

grafts, late graft occlusions, 361-362
Thrombocytopenia, platelet inhibitors,

202
Thrombolysis

complications, 565

hemorrhage prevention, 504-506
Thrombolytic agents, 543

aortic bifurcation graft limb occlusions,
285-288

biochemistry, 502-503

clinical trials, safety, 503-504

nonhemorrhagic complications, 506-507
Thrombophlebitis, subfascial endoscopic

perforator vein surgery (SEPS), 530
Thrombosis, see also Deep venous thrombosis
(DVT)

activated protein C resistance, 170-171

angioplasty, 605-606

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

720

INDEX

[Thrombosis]

antiphospholipid antibodies, 169-170

carotid stents, 621

catheters, temporary vascular access, 413

dacron grafts, 96-97

embolization, 565-566

expanded PTFE grafts, 106, 107

hyperhomocysteinemia, 171-172

intraoperative, 173

paraplegia, 686

peripheral vascular access, 4 1 7^4 1 8

plasminogen, 163-165

protein C deficiency, 166-168

protein S deficiency, 168-169

removal, 562-567
Thrombosis phenomenon, femorofemoral

graft, 300
Thrombotic thrombocytopenic purpura,

platelet inhibitors, 202
Thyroid collar, radiation exposure, 489
TIA. see Transient ischemic attacks (TIA)
Ticlopidine

pharmacology, 198-199

side effects, 201-202

stents, 200
TIMP-1, intimal hyperplasia, 76
Tinzaparin, FDA approved indications, 186t
Tirofiban, pharmacology, 199
Tissue factor pathway inhibitor (TFPI),

180-181
Tissue inhibitor of matrix metalloproteinase-1

(TIMP-1), intimal hyperplasia, 76
Tissue-type plasminogen activator (t-PA), 162,

166
TNF-alpha, vascular grafts, 311-312
Toes, ulceration, diabetic foot, 388
Tortuosity, endovascular access complications,

627
Tortuous axillofemoral bypass, 297
TOS, treatment, 435^137
Total cholesterol, stroke, 442
Total effective dose (TED), 483, 491
T-PA, 162, 166
Transaortic renal endarterectomy,

complications, 272
Transesophageal echocardiography (TEE),
33

pararenal aorta, 255
Transforming growth factor-beta (TGF-beta),
71

intimal hyperplasia, 76

wound complications, 366

Transient ischemic attacks (TIA), 443, 616

carotid endarterectomies (CEA), 471
Transthoracic bypass, innominate artery, 462
Transverse cervical nerve, 476
Treadmill, 3
Tricyclic antidepressants, adverse effects,

237
Tumor necrosis factor-alpha (TNF-alpha),

vascular grafts, 311-312
Tumors, .see Neoplasia
Tunneling complications, supraaortic trunks

(SAT), extra-anatomical repairs, 460
Tyrell cuffs, expanded PTFE grafts, 1 10

UFH. see Unfractionated heparin (UFH)
Ultrasonography, duplex (.?ee Duplex

ultrasonography)
Ultrasound beam steering, 5
Umbilical vein grafts, 124-125

anastomotic aneurysms, 125
Unfractionated heparin (UFH), 179-184

clinical application, 181-182

complications, 182-184

failure, 184

pharmacology, 180-181
Upper extremities

axillofemoral bypass, 294-296

thromboembolic events

axillofemoral bypass, 294-296
Uremia, bleeding, 411-412
Ureter, obstruction, 53-54
Ureteric obstruction, dacron grafts, 100-102
Urokinase, 162, 543

clinical trials, safety, 504

hemorrhage prevention, 506

nonhemorrhagic complications, 506-507
Urticaria, sclerotherapy, 516

Vacuum dressing, diabetic foot, 387

Vagus nerve, 476

Vancomycin

diabetic foot, 385

methicillin-resistant Staphylococcus aureus,

313
vascular access, 420
Vancomycin-resistant Enterococcus (VRE),

diabetic foot, 385
Vanguard stents, 637, 643-644
endoleaks, 653
fabric fatigue, 651-652
Varicose veins, failing autogenous grafts,
346

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6

INDEX

721

Vascugraft, 122-123
Vascular access
complications, 409-426
dialysis history, 418-422
failing fistulas, 422^127
peripheral, 414-418
autogenous, 414-415
complications, 416
maintenance, 416-417
prosthetic arteriovenous fistula,

415-416
thrombosis, 417-418
temporary, 409-414
bleeding, 411-412
hemothorax, 412^113
infection, 410-414
pneumothorax, 412-413
pulmonary embolism, 414
thrombosis, 413
Vascular anomalies, types, 581-585
Vascular brachytherapy, intimal hyperplasia,

73-75
Vascular endothelial growth factor (VEGF),

691-696
Vascular graft infections, 305-314
bacteriology, 307-309
epidemiology, 305-307
incidence, 306
pathogenesis, 309-312

biomaterial surface bacterial seeding,

309-312
immune system activation, 311-312
microcolony formation and bacterial
biofilms, 310-311
prevention, 312-314

intraoperative measures, 313-314
postoperative measures, 314
preoperative measures, 312-314
risk factors, 307
Vascular graft thrombosis, hypercoagulable

states, 155-174
Vascular malformations, angiogenesis therapy,

693-694
Vascular prostheses, 83-128
biological grafts, 124-127
bovine xenografts, 127
cryopreserved allografts, 125-127
umbilical vein, 124-125
dacron grafts, 84-103
expanded PTFE grafts, 103-122
polyurethane, 122-124
Vascular snare, guidewires, 611

Vascular surgery

cardiac assessment risk assessment postures,

27-31
cardiopulmonary complications, 15-39
clinical risk assessment, 18-21
risk-reduction techniques, 31-33
pulmonary complications, 34-39
clinical risk assessment, 34-35
preoperative evaluation, 36-37
risk reduction, 37-39
Vascular testing, goals, 12
Vasculogenic impotence, 240
Vasodilators, pararenal aorta, 255
Vectra graft, 122-123
Vein conduit torsion, failing autogenous grafts,

346
Vein cuffs and patches, expanded PTFE grafts,

109-111
Vein graft

diameter, failing autogenous grafts, 345
failure, 67-69
Vein interposition cuffs, expanded PTFE grafts,

110
Vein patch, carotid endarterectomies (CEA),

471
Vein patch angioplasty

failing autogenous graft, 347
perforator vein surgery, 537
Vein patch rupture, carotid endarterectomies

(CEA), 478-479
Veins, endoscopic harvesting, 373, 525
Velocity-calibrated string phantom, 5
Velocity waveform analysis, failing autogenous

grafts, 341
Vena cava filters

complications, 569-576

filter misplacements, 571-573
gastrointestinal, 571
guidewire, 571-574
prevention, 575-576
renal, 574

vena cava wall perforation, 571
current devices, 570
indications, 569-570
Venal caval replacement, expanded PTFE

grafts, 121-122
Vena Tech-LGM, 570
Venotomy

failing autogenous graft, 347
failing autogenous grafts, 346
Venous cannulation, complications,
542

i

2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

722

INDEX

Venous hypertension

superficial femoropopliteal vein harvest, 330

vascular access, 421
Venous malformations, arteriovenous

malformations (AVM), 592
Venous replacement, cryopreserved allografts,

126
Venous rupture, percutaneous transluminal

angioplasty (PTA), 560
Venous stenosis, vascular access, 418-419
Venous thromboembolism (VTE)

low molecular weight heparin (LM WH),
186

warfarin, 191
Verapamil, 605

Vessel closure, carotid stents, 618
Viagra, impotence, 246
Visceral patch, following TAAA surgery,

anastomotic aneurysms, 149-155
Visceral perfusion, spinal cord ischemia

prevention, 228-229
Vitamin A, 366
Vitamin K antagonists, 190-193

clinical application, 191-192

complications, 192-193

drug interactions affecting, 191

failures, 193

pharmacology, 190-191
VRE, 385
VTE. see Venous thromboembolism (VTE)

Waist badges, radiation exposure, 490
Wallstents, 548, 562

migration, 546-547
Warfarin, 173, 183

clinical application, 191-192

complications, 192-193

crossover grafts, 122

failures, 193

pharmacology, 190-191

subclavian thrombosis, temporary vascular
access, 413

thrombosis, plasminogen, 165
Whiplash, 435
W.L. Gore TAG stent graft, 639

fractures, 649
Wound care, diabetic foot, 385-387
Wound complications, 365-374

anatomical considerations, 369-373

classification, 365-366

etiology, 366

impact, 369

incidence, 369-373

intraoperative care, 367-368

postoperative care, 369

preoperative care, 366-369

wound infection, 373-374
Wound infection, 373-374
Woven Dacron grafts, 84

Ximelagatran/melagatran, 196-197

i
2

I

■c

I

u

Marcel Dekker, Inc.
270 Madison Avenue, New York, New York 1 00 1 6

About the Editors

Jonathan B. Towne is Professor of Surgery and Chief of Vascular Surgery, Medical
College of Wisconsin, Milwaukee. He received the M.D. degree (1967) from the University
of Rochester School of Medicine and Dentistry, New York. He obtained vascular surgery
training under Dr. Jesse Thompson in Dallas, Texas.

Larry H. Hollier is Professor of Surgery and Dean of the Louisiana State University
Health Sciences Center School of Medicine, New Orleans, Louisiana. He previously served
as President of the Mount Sinai Hospital and Chairman of the Department of Surgery at
the Mount Sinai School of Medicine, New York, New York. The author of more than 300
articles in medical literature, he serves on the editorial boards of 14 surgical journals.
Dr. Hollier received the M.D. degree (1968) from Louisiana State University School of
Medicine, New Orleans. He obtained vascular surgery training under Dr. Jesse Thompson
in Dallas, Texas.

i

2

I

■c

723 f

I
u

Marcel Dekker, Inc.
270 Madison Avenue. New York, New York 1 00 1 6