Proefschrift buck

Innovations in Abdominal Aneurysm Repair Dominique B. Buck

Innovations in Abdominal Aneurysm Repair

Dominique Babette Buck

Innovations in Abdominal Aneurysm Repair Thesis, University of Utrecht, The Netherlands The work presented in this thesis was supported by a grant from Association Leatare, which is gratefully acknowledged. Krijnen Medical Innovations B.V., Stichting Michael-van Vloten and Chirurgisch Fonds UMC Utrecht provided sponsorship for this thesis. Financial support by the Dutch Heart Foundation for the publication of the thesis is gratefully acknowledged. Cover-design and layout: Auke Douma and Valentijn de Jong Printed by: Gildeprint – The Netherlands ISBN/EAN 978-94-6108-972-4 Š D.B. Buck, 2015. All rights reserved.

Innovations in Abdominal Aneurysm Repair

Proefschrift

Ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof. dr. G.J. van der Zwaan, ingevolge het besluit van het college van promoties, In het openbaar te verdedigen op maandag 11 mei 2015 des middags te 14.30 uur

door

Dominique Babette Buck

Geboren op 14 juli 1987 te Heemstede Nederland

Promotores Prof. Dr. F.L. Moll Co-promotores Dr. M.L. Schermerhorn Dr. J.A. van Herwaarden

For my Dad,

CONTENTS Introduction Chapter 1

General Introduction and outline of the thesis

PART I

Surgical techniques and insights in AAA management

Chapter 2

Endovascular treatment of abdominal aortic aneurysms. Dominique B. Buck, Joost A. van Herwaarden, Marc L. Schermerhorn, Frans L. Moll. Nat Rev Cardiol 2014 Feb;11(2):112-23.

Chapter 3

Long-term Outcomes for Endovascular vs. Open Repair of Abdominal Aortic Aneurysms. Marc L. Schermerhorn, Dominique B. Buck, A. James O’Malley, Thomas Curran, Lawrence Zaborski, John C. McCallum, Jeremy Darling, Bruce E. Landon. Accepted N Engl J Med.

Chapter 4

Risk factors and consequences of persistent type II endoleaks. Dominique B. Buck, Ruby C. Lo, Jeremy Herrmann, Allen D. Hamdan, Mark Wyers, Virendra I. Patel, Mark Fillinger, Marc L. Schermerhorn. Revisions J Vasc Surg.

Chapter 5

Ambulatory Endovascular Abdominal Aortic Aneurysm Repair in a National and State Database. Thomas Curran, Sarah Carlson, Dominique B. Buck, John C. McCallum, Jeremy Darling, Michelle Martin, Mark Wyers, Marc L. Schermerhorn. Submitted J Vasc Surg.

Chapter 6

Comparison of Endovascular Stent Grafts for Abdominal Aortic Aneurysm Repair in Medicare Patients. Dominique B. Buck, Bruce E. Landon, A. James O’Malley, Sara L. Zettervall, Peter A. Soden, Lawrence Zaborski, Klaas H.J. Ultee, Marc L. Schermerhorn.

Chapter 7

Percutaneous versus Femoral Cutdown Access for Endovascular Aneurysm Repair. Dominique B. Buck, Eleonora G. Karthaus, Peter A. Soden, Klaas H.J. Ultee, Joost A. van Herwaarden, Frans L. Moll, Marc L. Schermerhorn. J Vasc Surg. 2015 Mar 28.

Chapter 8

Comparing the Transperitoneal vs. Retroperitoneal Surgical Approach Open Abdominal Aortic Aneurysm Repair. Dominique B. Buck, Sara L. Zettervall, Klaas H.J. Ultee, Peter A. Soden, Jeremy Darling, Joost A. van Herwaarden, Mark Wyers, Marc L. Schermerhorn. Submitted J Am Coll Surg.

Chapter 9

The Impact of Endovascular Repair on Specialties Performing Abdominal Aortic Aneurysm Repair. Rob Hurks, Klaas H.J. Ultee, Dominique B. Buck, George S. DaSilva, Peter A. Soden, Joost A. van Herwaarden, Hence J.M. Verhagen, Marc L. Schermerhorn. Accepted J Vasc Surg.

PART II Evolution of Abdominal Aneurysm Repair Chapter 10

The impact of Endovascular Treatment on Isolated Iliac Artery Aneurysm Treatment and Mortality Dominique B. Buck, Rodney P. Bensley, Jeremy Darling, Thomas Curran, John C. McCallum, Frans L. Moll, Joost A. van Herwaarden, Marc L. Schermerhorn. Accepted J Vasc Surg.

Chapter 11 Isolated Renal Artery Aneurysms: Management and Outcomes in the Endovascular Era. Dominique B. Buck, Thomas Curran, John C. McCallum, Jeremy Darling, Rishi Mamtani, Joost A. van Herwaarden, Frans L. Moll, Marc L. Schermerhorn. Revisions J Vasc Surg.

PART III

Optimal Health Care in Vascular Surgery

Chapter 12

Online Patient Resources for Cardiovascular Surgery: An Analysis of Readability. Dominique B. Buck, Pieter G. Koolen, Jeremy Darling, Christina R. Vargas, Rishi Mamtani, Duane S. Pinto, Bernard T. Lee, Marc L. Schermerhorn. Submitted J Am Coll Cardiol.

Discussion Chapter 13

Optimal Health care in vascular surgery Summary, general discussion and future perspectives

APPENDICES Review Committee Word of Thanks Curriculum Vitae

CHAPTER 1

General Introduction and outline of the thesis

Chapter 1

Introduction

An aneurysm - derived from the ancient Greek word ανεύρυσμα (widening) - is a focal dilatation in the wall of a blood vessel. An abdominal aortic aneurysm (AAA) occurs when a weak area of the abdominal aorta expands beyond its safety margin. When ruptured, AAAs are the tenth leading cause of death in men over the age of 55 in the United States; further, mortality rates of ruptured AAA exceed 80%.1, 2 Until the early 2000s, elective open AAA repair to prevent rupture was performed with a mortality of approximately 5%. In recent years, however, endovascular aneurysm repair (EVAR) has been introduced and approved as a minimally invasive alternative to standard open surgical repair.1, 3 Different from open techniques, EVAR involves transfemoral insertion of a stent graft made of fabric and self-expanding metal stents that attaches to a segment of normal aorta below the renal arteries and again to the normal iliac arteries below, thereby excluding the aneurysm from the circulation internally (Figure 1). The use of EVAR has steadily grown, increasing from 56% of all elective AAA repairs performed in the US Medicare population in 2005 to 77% by 2008.4, 5 As EVAR has become widely accepted as a safe technique, a large variety of

Figure 1. Diagram of open and endovascular AAA repair. Reprinted from Schermerhorn et al. Endovascular vs Open Repair of Abdominal Aortic Aneurysms in the Medicare Population. N Engl J Med 2008; 358: 464-74.

12

General Introduction and outline of the thesis commercially available stent grafts have been introduced. Interestingly, no clear evidence supports one specific treatment regimen over another; due to this debate, specific practice patterns have been shown to vary between institutions.6 Not only is there disparity in the implementation of different graft types during surgery, but there is also variation in endovascular surgical techniques. Compared to femoral cutdown access, percutaneous access further minimizes invasiveness and has also been shown to have high technical success rates. So far, however, there has not yet been a large-scale study that is generalizable to all relevant patients and practices. Despite the minimally invasive treatment option, open surgery repair remains necessary in certain circumstances or in individual patients. Even further, extensive discussion continues about the optimal established open approach: transperitoneal or retroperitoneal.7-10 Data regarding the long-term efficacy of open treatment have remained conflicting, while randomized trials and observational studies have demonstrated lower perioperative morbidity and mortality with endovascular AAA treatment. The survival benefit of EVAR, however, is not sustained. 11-17 Aneurysmal degeneration is not limited to the aorta: although less common, these problems have been known to exist in the iliac and renal arteries as well. Iliac artery aneurysms may appear in isolation, comprising approximately 2% of all abdominal aneurysms.18-21 Isolated iliac artery aneurysms are uncommon, frequently asymptomatic, and most often discovered incidentally.19, 22, 23 The majority of isolated iliac aneurysms (70%) occur in the common iliac artery, while 20% and 10%are found in the internal and external iliac arteries, respectively.24 With an estimated incidence between 0.1-0.3%,25-28 renal artery aneurysms have also proven to be another form of isolated abdominal aneurysms; Unlike iliac arteries, however, the indications for surgical intervention of renal artery aneurysms - whether by open or endovascular techniques - remain controversial.29 The technical success and safety of endovascular treatment for both isolated iliac- and isolated renal artery aneurysms have been demonstrated.18, 19, 30, 31 However, at this stage, little is known about the impact of the minimally invasive endovascular therapy on mortality and outcomes over time. Beyond sharpening surgical techniques and improving healthcare practices, optimal patient care may also be obtained by enhancing the delivery and dissemination of vascular healthcare information. As the majority of patients use the internet,34, 35 providing accessible and appropriate resources is essential for patient access to health information as a means to make the most valuable decisions with regard to vascular health care.

Aims and Outlines

This thesis consists of three parts. Part I focuses on the evaluation of surgical techniques and the management of AAA repair, highlighting insights in the performance of endovascular repair, the use of different stent grafts, and insights 13

1

Chapter 1 in the execution of open repair. Additionally, outcome comparisons are made and temporal trends are shown as a means to enhance understanding of the differences between open and endovascular repair of AAA. Chapter 2 provides a general overview of EVAR, with a review of literature mainly focused on clinical evidence for the benefits, the indications for repair, and the new stent grafts that are being developed for EVAR. Chapter 3 contains a comparison between long-term outcomes of endovascular versus open repair to evaluate complications and late survival. In this chapter, changes over time are shown to identify progress in this evolving technology. Type 2 endoleaks are common after EVAR, but their clinical significance remains controversial. Chapter 4 seeks to determine risk factors for type 2 endoleaks and associations with additional adverse outcomes. Endovascular AAA repair increases in popularity worldwide, yet the cost effectiveness of EVAR as compared to open repair is still questionable. In hopes to address these concerns, the safety and cost of ambulatory EVAR as compared to inpatient EVAR is analyzed in Chapter 5. As a means to evaluate differences in postoperative and late outcomes, Chapter 6 offers a comparison of the most commonly used stent grafts. Chapter 7 and Chapter 8 address the operative techniques of endovascular and open repair, respectively. More specifically, Chapter 7 compares percutaneous access to femoral cutdown access, while Chapter 8 compares the transperitoneal approach to the retroperitoneal open approach. In Chapter 9, the impact of various specialties performing endovascular AAA repair is analyzed and discussed, as the distribution of the type of physician has changed over time. The aim of Part II is to demonstrate the evolution and epidemiologic trends in management for abdominal aneurysm repair. Chapter 10 focuses on the impact of endovascular repair on isolated iliac aneurysm repair, while Chapter 11 evaluates isolated renal aneurysm repair. Both studies compare mortality and morbidity outcomes between endovascular and open repair. Part III represents an overall evaluation of how cardiovascular health information is received by patients. In Chapter 12, reading levels of online cardiovascular resources are examined to increase awareness of patient literacy amongst physicians. Disparities of results are displayed to better guide patients and physicians to specific websites with health information on cardiovascular disease. Finally, the content of this thesis is summarized, discussed, and future perspectives are presented in Chapter 13.

14

General Introduction and outline of the thesis

References 1. Schermerhorn ML CJ. “Abdominal Aortic Aneurysm” in Vascular Surgery. Philadelphia, PA: WB Saunders Company. 2005:1408-52. 2. Sachs T, Schermerhorn M. Ruptured abdominal aortic aneurysm. Minerva Chir. 2010;65(3):303-17. 3. Parodi JC, Palmaz JC, Barone HD. Transfemoral intraluminal graft implantation for abdominal aortic aneurysms. Annals of vascular surgery. 1991;5(6):491-9. 4. Giles KA, Pomposelli F, Hamdan A, Wyers M, Jhaveri A, Schermerhorn ML. Decrease in total aneurysm-related deaths in the era of endovascular aneurysm repair. J Vasc Surg. 2009;49(3):543-50; discussion 50-1. 5. Schermerhorn ML, Bensley RP, Giles KA, Hurks R, O’Malley A J, Cotterill P, et al. Changes in abdominal aortic aneurysm rupture and short-term mortality, 1995-2008: a retrospective observational study. Annals of surgery. 2012;256(4):651-8. 6. van Marrewijk CJ, Leurs LJ, Vallabhaneni SR, Harris PL, Buth J, Laheij RJ, et al. Risk-adjusted outcome analysis of endovascular abdominal aortic aneurysm repair in a large population: how do stent-grafts compare? J Endovasc Ther. 2005;12(4):417-29. 7. Arko FR SS, Zarins CK. Repair of Infrarenal Abdominal Aortic Aneurysms. ACS Surgery: Principles and Practice. 2007. 8. Cambria RP, Brewster DC, Abbott WM, Freehan M, Megerman J, LaMuraglia G, et al. Transperitoneal versus retroperitoneal approach for aortic reconstruction: a randomized prospective study. J Vasc Surg. 1990;11(2):314-24; discussion 24-5. 9. Mehta M, Roddy SP, Darling RC, 3rd, Ozsvath KJ, Kreienberg PB, Paty PS, et al. Infrarenal abdominal aortic aneurysm repair via endovascular versus open retroperitoneal approach. Annals of vascular surgery. 2005;19(3):3748. 10. Sicard GA, Reilly JM, Rubin BG, Thompson RW, Allen BT, Flye MW, et al. Transabdominal versus retroperitoneal incision for abdominal aortic surgery: report of a prospective randomized trial. J Vasc Surg. 1995;21(2):174-81; discussion 81-3. 11. Prinssen M, Verhoeven EL, Buth J, Cuypers PW, van Sambeek MR, Balm R, et al. A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med. 2004;351(16):1607-18. 12. participants Et. Endovascular aneurysm repair versus open repair in patients with abdominal aortic aneurysm (EVAR trial 1): randomised controlled trial. Lancet. 2005;365(9478):2179-86. 13. Lederle FA, Freischlag JA, Kyriakides TC, Padberg FT, Jr., Matsumura JS, Kohler TR, et al. Outcomes following endovascular vs open repair of abdominal aortic aneurysm: a randomized trial. JAMA : the journal of the American Medical Association. 2009;302(14):1535-42. 14. De Bruin JL, Baas AF, Buth J, Prinssen M, Verhoeven EL, Cuypers PW, et 15

1

Chapter 1 al. Long-term outcome of open or endovascular repair of abdominal aortic aneurysm. N Engl J Med. 2010;362(20):1881-9. 15. Lederle FA, Freischlag JA, Kyriakides TC, Matsumura JS, Padberg FT, Jr., Kohler TR, et al. Long-term comparison of endovascular and open repair of abdominal aortic aneurysm. N Engl J Med. 2012;367(21):1988-97. 16. Jackson RS, Chang DC, Freischlag JA. Comparison of long-term survival after open vs endovascular repair of intact abdominal aortic aneurysm among Medicare beneficiaries. JAMA : the journal of the American Medical Association. 2012;307(15):1621-8. 17. Greenhalgh RM, Brown LC, Powell JT, Thompson SG, Epstein D, Sculpher MJ. Endovascular versus open repair of abdominal aortic aneurysm. N Engl J Med. 2010;362(20):1863-71. 18. Huang Y, Gloviczki P, Duncan AA, Kalra M, Hoskin TL, Oderich GS, et al. Common iliac artery aneurysm: expansion rate and results of open surgical and endovascular repair. J Vasc Surg. 2008;47(6):1203-10; discussion 10-1. 19. Patel NV, Long GW, Cheema ZF, Rimar K, Brown OW, Shanley CJ. Open vs. endovascular repair of isolated iliac artery aneurysms: A 12-year experience. J Vasc Surg. 2009;49(5):1147-53. 20. Chaer RA, Barbato JE, Lin SC, Zenati M, Kent KC, McKinsey JF. Isolated iliac artery aneurysms: a contemporary comparison of endovascular and open repair. J Vasc Surg. 2008;47(4):708-13. 21. Sandhu RS, Pipinos, II. Isolated iliac artery aneurysms. Seminars in vascular surgery. 2005;18(4):209-15. 22. Scheinert D, Schroder M, Steinkamp H, Ludwig J, Biamino G. Treatment of iliac artery aneurysms by percutaneous implantation of stent grafts. Circulation. 2000;102(19 Suppl 3):III253-8. 23. Chemelli A, Hugl B, Klocker J, Thauerer M, Strasak A, Jaschke W, et al. Endovascular repair of isolated iliac artery aneurysms. J Endovasc Ther. 2010;17(4):492-503. 24. Krupski WC, Selzman CH, Floridia R, Strecker PK, Nehler MR, Whitehill TA. Contemporary management of isolated iliac aneurysms. J Vasc Surg. 1998;28(1):1-11; discussion -3. 25. Hageman JH, Smith RF, Szilagyi E, Elliott JP. Aneurysms of the renal artery: problems of prognosis and surgical management. Surgery. 1978;84(4):56372. 26. Henriksson C, Bjorkerud S, Nilson AE, Pettersson S. Natural history of renal artery aneurysm elucidated by repeated angiography and pathoanatomical studies. European urology. 1985;11(4):244-8. 27. Stanley JC, Rhodes EL, Gewertz BL, Chang CY, Walter JF, Fry WJ. Renal artery aneurysms. Significance of macroaneurysms exclusive of dissections and fibrodysplastic mural dilations. Arch Surg. 1975;110(11):1327-33. 28. Tham G, Ekelund L, Herrlin K, Lindstedt EL, Olin T, Bergentz SE. Renal artery aneurysms. Natural history and prognosis. Annals of surgery. 1983;197(3):348-52. 29. Henke PK, Cardneau JD, Welling TH, 3rd, Upchurch GR, Jr., Wakefield TW, Jacobs LA, et al. Renal artery aneurysms: a 35-year clinical experience 16

General Introduction and outline of the thesis with 252 aneurysms in 168 patients. Annals of surgery. 2001;234(4):454-62; discussion 62-3. 30. Zhang Z, Yang M, Song L, Tong X, Zou Y. Endovascular treatment of renal artery aneurysms and renal arteriovenous fistulas. J Vasc Surg. 2013;57(3):765-70. 31. Klein GE, Szolar DH, Breinl E, Raith J, Schreyer HH. Endovascular treatment of renal artery aneurysms with conventional non-detachable microcoils and Guglielmi detachable coils. British journal of urology. 1997;79(6):852-60. 32. Shiraev T, Condous MG. Incidence and outcomes of ruptured abdominal aortic aneurysms in rural and urban Australia. ANZ journal of surgery. 2013. 33. Mani K, Lees T, Beiles B, Jensen LP, Venermo M, Simo G, et al. Treatment of abdominal aortic aneurysm in nine countries 2005-2009: a vascunet report. European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery. 2011;42(5):598-607. 34. Hesse BW, Nelson DE, Kreps GL, Croyle RT, Arora NK, Rimer BK, et al. Trust and sources of health information: the impact of the Internet and its implications for health care providers: findings from the first Health Information National Trends Survey. Archives of internal medicine. 2005;165(22):261824. 35. Health Fact Sheet, Pew Research Internet Project. http://www.pewinternet. org/fact-shets/health-fact-sheet/; January 2014.

17

1

CHAPTER 2

Endovascular treatment of abdominal aortic aneurysms Dominique B. Buck, Joost A. van Herwaarden, Marc L. Schermerhorn and Frans L. Moll. Nat Rev Cardiol 2014 Feb;11(2):112-23.

Chapter 2

Abstract

Patients with abdominal aortic aneurysms (AAAs) are usually treated with endovascular aneurysm repair (EVAR), which has become the standard of care in many hospitals for patients with suitable anatomy. Clinical evidence indicates that EVAR is associated with superior perioperative outcomes and similar long-term survival compared with open repair. Since the randomized, controlled trials that provided this evidence were conducted, however, the stent graft technology for infrarenal AAA has been further developed. Improvements include profile downsizing, optimization of sealing and fixation, and the use of low porosity fabrics. In addition, imaging techniques have improved, enabling better preoperative planning, stent graft placement, and postoperative surveillance. Also in the past few years, fenestrated and branched stent grafts have increasingly been used to manage anatomically challenging aneurysms, and experiments with off-label use of stent grafts have been performed to treat patients deemed unfit or unsuitable for other treatment strategies. Overall, the indications for endovascular management of AAA are expanding to include increasingly complex and anatomically challenging aneurysms. Ongoing studies and optimization of imaging, in addition to technological refinement of stent grafts, will hopefully continue to broaden the utilization of EVAR.

20

Endovascular treatment of abdominal aortic aneurysms

Introduction

The concept of endovascular aneurysm repair (EVAR) was first reported by Volodos et al.1,2 in 1986 and by Parodi and colleagues in 1991.3 Since that time, EVAR has become widely accepted as a safe technique for the treatment of abdominal aortic aneurysm (AAA).4,5 Early application of the endovascular technique was for the treatment of old patients with substantial comorbid illness who were unfit for open repair. However, the indications for this minimally invasive technique have expanded considerably since the procedure was first used. By 2003, EVAR was used for nearly half of all elective AAA repairs in the USA.6 Currently, the majority of patients with anatomically suitable infrarenal AAA are treated with EVAR, rather than with open repair.7 Indeed, a retrospective analysis of Medicare beneficiaries undergoing AAA repair in the USA showed that EVAR accounted for 77% of intact AAA repairs in 2008.8 In this Review, we provide both a contemporary reappraisal of evidence-based practice and an evaluation of promising new strategies in the endovascular management of AAAs. The role of imaging in the management of AAA, as well as new, innovative techniques and off-label use of stent grafts are discussed. This Review has a particular focus on indications for intervention and operative technique for patients who are anatomically unsuitable for standard EVAR. Treatment of patients with juxtarenal or suprarenal aneurysms is covered, and an overview of the premise of the evolving nature of endovascular surgery is provided.

Key Points • • • • • •

Endovascular aneurysm repair (EVAR), rather than open repair, is currently the treatment of choice for most patients with an anatomically suitable infrarenal abdominal aortic aneurysm (AAA) Clinical evidence-based research shows a lower perioperative morbidity and mortality, and similar long-term survival, for EVAR compared with open repair of suitable infrarenal AAAs The indications for endovascular management of AAA are expanding to include increasingly complex and anatomically challenging aneurysms Challenging anatomy might require the use of fenestrated and branched stent grafts, chimney grafts, or the sandwich technique Future directions for stent grafts include fenestrated and branched offthe-shelf stent grafts, multilayer stents, endoanchor systems, and sac-anchoring endoprostheses Stent graft technology for infrarenal AAA continues to evolve, with profile downsizing, optimization of sealing and fixation, and the use of fabrics with reduced porosity

21

2

Chapter 2

Incidence of AAA

The risk factors for AAA have been well described and include advanced age, cardiovascular disease, family history of AAA, smoking, and male sex.9–11 Although the incidence of AAA is fourfold higher in men than in women, AAA rupture occurs at smaller aortic diameters in women, and they present for care at a more-advanced age than men.12 Moreover, large-scale studies of females with AAA have demonstrated poorer outcomes than those experienced by men.12–14 These findings indicate a difference in the natural history of disease between men and women. In the 1980s and 1990s, an increase in overall AAA mortality was observed.15–18 However, epidemiological reports from the 2000s have shown the incidence of AAA in developed countries to be decreasing.19,20 Moreover, in an American study of 338,278 Medicare beneficiaries undergoing AAA repair from 1995 to 2008, the total number of AAA ruptures decreased from 6,535 to 3,298 over the study period, yet the proportion of ruptures undergoing repair changed only slightly (from 70% to 65%).8 Interestingly, the overall number of annual deaths from both intact and ruptured AAAs has decreased significantly since the introduction of EVAR.8,21 This reduction in ruptured AAA has been multifactorial, with valuable contributions made by improved screening programmes, decreased prevalence of smoking, improved cardiovascular risk prevention, and an increase in elective surgery in patients aged >75 years.22,23

Imaging in AAA management

Imaging has a very important role in the management of AAA. Several modalities are available for clinicians assessing patients with this condition. Ultrasonography The most-commonly used technique for screening and surveillance of patients with AAA is ultrasonography (Figure 1a, b). This imaging modality has a high sensitivity and specificity, is inexpensive and noninvasive, and does not subject the patient to ionizing radiation.24,25 The disadvantages of this technique are that it is operator-dependent and provides less-accurate diameter measurements than other available imaging modalities. Therefore, unlike computed tomography angiography (CTA; discussed below), ultrasonography is not suitable for use in preoperative work-up for AAA surgery.26 Ultrasonography is mainly used for monitoring aneurysm growth and discovering postoperative endoleaks.27 Contrast-enhanced ultrasonography, which involves the use of microbubble contrast agents, can be used safely in patients with impaired renal function. This imaging modality has been investigated as an alternative to CTA in the surveillance of patients after EVAR, and seems to provide accurate diagnostic information.28–30

22

Endovascular treatment of abdominal aortic aneurysms CTA Whereas conventional angiography has largely fallen out of favour as a firstline imaging modality in the management of AAA, spiral CTA, with its lessinvasive approach, has been demonstrated to be the best imaging technique for both preoperative patient assessment and postoperative aortic stent graft surveillance.31,32 The disadvantages of CTA are the radiation burden and the use of potentially nephrotoxic contrast agents.26 Nevertheless, CTA is fast and provides all necessary detailed anatomic information, including 3D visualization (Figure 1c,d).33–35 3D CTA currently also provides dynamic image capability, known as time-resolved or 4D CTA. These dynamic images can be used to assess variations in aortic diameter related to the pulsatile nature of the cardiac cycle. The critical nature of this variation was demonstrated in 2009 by Van Keulen and colleagues, who concluded in their systematic review that the proximal stent-graft-landing zone (AAA neck diameter) differs significantly in systole and diastole.36 a

b

e

f

c

d

g

h

Figure 1. Imaging modalities used in the management of patients with AAA. a | Ultrasonography showing a transverse view of the distal aorta showing the diameter of the vessel. b | Ultrasonography showing a transverse view showing the aortic bifurcation; left and right iliac artery. c | Preoperative and d | postoperative 3D CTA reconstruction for a patient with AAA. e,f | Two postoperative 3D MRA reconstructions for a patient with AAA. g | IVUS demonstrating incomplete expansion of a stent graft in an AAA. Reprinted from Lee, J. T. & White, R. A. Basics of intravascular ultrasound: an essential tool for the endovascular surgeon. Semin. Vasc. Surg. 17, 110–118, Copyright 2004, with permission from Elsevier. h | IVUS demonstrating standard measurement of the aorta at the level of the renal arteries. Reprinted from Marrocco, C. J. et al. Intravascular ultrasound. Semin. Vasc. Surg. 25, 144–152, Copyright 2012, with permission from Elsevier. Abbreviations: AAA, abdominal aortic aneurysm; CTA, computed tomography angiography; IVUS, intravascular ultrasound; MRA, magnetic resonance angiography.

23

2

Chapter 2 MRA Magnetic resonance angiography (MRA) is another tool at the disposal of the vascular surgeon for preoperative assessment of patients with AAA (Figure 1e, f). Benefits of MRA include high-resolution soft-tissue contrast, and possible demonstration of arterial wall movement and flow quantification. In addition, MRA can be used to evaluate both the vascular lumen and its wall.37 Unlike CTA, MRA is obtained without the use of iodinated contrast material or radiation.38 Furthermore, MRA has higher sensitivity than CTA for the detection of endoleaks in patients after EVAR, particularly type 2 endoleaks.39 However, MRA is more expensive and more time consuming than CTA, does not visualize calcium as well as CTA, and is less comfortable than CTA for the patient (indeed, this modality is contraindicated in patients with claustrophobia or certain metallic implantations). Like CTA, MRA can generate dynamic images. Dynamic, time-resolved MRAs were first described in 1996,40 and can be used to detect changes in vascular flow direction over time. The 4D MRA images are considered to be of most value in identifying endoleaks, as the inflow and outflow source can be well observed.40–42 IVUS Intravascular ultrasound (IVUS; Figure 1g, h) is also a part of the imaging armamentarium of the vascular surgeon, providing a 360° cross-sectional view of the vessel being evaluated. Maximum aortic diameter, varying with the cardiac cycle, can be evaluated in real time. IVUS can also be used to limit contrast administration in patients with renal insufficiency, as this technique is occasionally used as an adjunct without contrast injection. Despite these advantages, IVUS is susceptible to operator-dependent error, as accurate measurements require centring of the probe within the lumen, a task that becomes increasingly difficult with larger aortic diameter and longer procedure duration. Moreover, IVUS requires percutaneous vascular access, unlike noninvasive MRA or CTA. IVUS is currently primarily used for research purposes.43,44

Clinical evidence for benefits of EVAR

Comparison of EVAR with open surgical repair of AAA has been the subject of several well-designed randomized, controlled trials as well as various observational studies of the ‘real world’ setting. Randomized, controlled trials Accruing 351 patients between 2000 and 2003, the DREAM (Dutch Randomized Endovascular Aneurysm Management) trial demonstrated 30-day mortality of 1.2% and 4.6% for EVAR and open repair, respectively (P = 0.10).45 Additionally, the EVAR cohort was found to have a shorter hospital stay than the cohort who underwent open repair (6 days versus 13 days, P <0.001) and lower incidence of moderate or severe systemic complications over 30 days (11.7% versus 26.4%, P <0.001).45 The study showed a combined rate of severe complications and operative mortality of 4.7% for the EVAR group compared with 9.8% for the open-repair group (P = 0.10).45 Of note, however, subsequent longer-term follow 24

Endovascular treatment of abdominal aortic aneurysms up of this trial revealed that the perioperative survival advantage of EVAR over open repair was not sustained after the first postoperative year.46 After 6 years of follow-up, the rates of survival remained similar in the EVAR and open-repair groups (68.9% and 69.9%, respectively).47 Additionally, the long-term rates of secondary interventions were significantly higher for EVAR than for open repair over the 6-year follow-up period (29.6% versus 18.1%, P = 0.03).47 The most-common secondary interventions in the EVAR group were stent-graftrelated interventions, whereas the most-common procedure in the open-repair group was abdominal incisional hernia repair.47 The UK EVAR 1 (UK Endovascular Aneurysm Repair 1) trial,48 involving 1,082 patients treated between 1999 and 2003, also demonstrated a clear short-term 6-year follow-up period (29.6% versus 18.1%, P = 0.03).4 The most-common secondary interventions in the EVAR group were stent-graft-related interventions, whereas the most-common procedure in the open-repair group was abdominal incisional hernia repair.47 The UK EVAR 1 (UK Endovascular Aneurysm Repair 1) trial,48 involving 1,082 patients treated between 1999 and 2003, also demonstrated a clear short-term survival benefit for EVAR compared with open surgery. Among patients who were candidates for either EVAR or open repair, EVAR was associated with lower rates of 30-day operative mortality than open repair (1.7% versus 4.7%, P = 0.009).48 Median operative time (180 min versus 200 min, P <0.0001) and length of hospital stay (7 days versus 12 days, P <0.0001) were also lower in the EVAR group.48 Perioperative mortality was lower with EVAR,48 and lower disease-specific mortality was noted in the EVAR group at the 4-year follow-up (4% versus 7% in the open-repair group, P = 0.04); however, at the 4-year followup, no difference in all-cause mortality (26% versus 29%, P = 0.46) was observed between the two groups.49 Moreover, the early aneurysm-related mortality benefit with EVAR was counteracted by higher aneurysm-related mortality in the EVAR group than the open repair group after 4 years (2.1% versus 0.4%, P = 0.05, in patients followed up for 4–8 years).50 The lack of difference in all-cause mortality persisted throughout the 8 years of follow up.50 Additionally, cost analyses showed higher costs for the EVAR group (mean costs £15,303 versus £12,284 for open-repair), and a greater number of secondary interventions occurred in the EVAR group than in the open-repair group (5.1% versus 1.7%, P <0.001) over the long-term follow-up period.50 The UK EVAR 2 trial investigators compared survival in patients unfit for open repair (n = 338) who were randomly assigned to either EVAR or no intervention.51–53 Of note, the randomization in both the UK EVAR trials was determined by the surgeon at the local level. The EVAR 2 trial did not demonstrate a survival benefit for elective EVAR compared with no intervention in this frail patient population.51–53 However, notably, both UK EVAR trials were based on intentionto-treat analysis, with significant patient crossover between groups allowing for potentially biased results. In EVAR 2, more than one-quarter of patients assigned to no intervention for their aneurysm underwent aneurysm repair, of which 30% received surgery because of patient preference.51,53 This limitation 25

2

Chapter 2 and the lengthy delay between randomization and surgery in EVAR 2 (median 51 days51) have made it challenging for clinicians to accept the conclusion that EVAR is not worthwhile in high-risk patients. The USA OVER (USA Open Versus Endovascular Repair) trial investigators assigned 881 patients to either EVAR or open repair between 2002 and 2008.54 This multicentre, randomized trial showed a significant peri-operative benefit for EVAR, with lower mortality in the EVAR group than in the open-repair group at 30 days (0.5% versus 3.0%, P = 0.004) as well as a shorter operative time (2.9 h versus 3.7 h, P <0.001) and length of hospital stay (3 days versus 7 days, P <0.001) for the EVAR group.54 However, EVAR did not confer a survival benefit at the 2-year follow-up (mortality 7.0% versus 9.8% for open repair, P = 0.13).54 Furthermore, rates of other outcomes also converged by longer-term follow-up: no differences were observed at 2 years in quality of life or secondary therapeutic procedures.54 Although the study population generally demonstrated similar long-term outcomes at the mostrecent follow-up reported (up to 9 years; mean 5.2 years),55 subgroup analysis did show significantly increased long-term survival for patients aged <70 years undergoing EVAR compared with those assigned to open repair (HR 0.65, 95% CI 0.43–0.98, P = 0.04).55 From 2003 to 2008, the French ACE (Anevrysme de l’aorta abdominale, Chirurgie versus Endoprothese) investigators randomly assigned 316 patients to EVAR or open surgery.56 No difference in survival was observed between the two patient groups at a median of follow-up of 3 years (EVAR 11.3% versus open repair 8%, P = NS).56 However, significant differences were seen with respect to long-term vascular reintervention rates (EVAR 16% versus open repair 2.7%, P <0.0001).56 Like the UK EVAR 1 and 2 trials, this study followed an intention-to-treat analysis with apparent cross-over between study groups.56 Observational studies To understand the comparative effectiveness of various interventions, inquiry should be extended beyond randomized, controlled trials of ideal populations to study interventions in ‘real world’ settings using observational data. Although observational studies that rely on administrative data are subject to coding errors, they provide invaluable data about ‘real world’ outcomes outside of the tightly controlled parameters of randomized, controlled trials. In 2008, Schermerhorn and colleagues studied a cohort of 61,598 Medicare beneficiaries undergoing open or endovascular repair in the USA from 2001 to 2004.57 EVAR was associated with lower rates of perioperative death (1.2% versus 4.8%, P <0.001) and various major medical and surgical complications than open repair.57 The early benefit from endovascular repair persisted for more than 3 years, after which the survival rates associated with the two procedures converged.57 The lower early mortality with EVAR was more apparent with increasing age; for patients aged ≥85 years, the difference in mortality between the two groups was 8.5%, whereas the difference was 2.1% for patients aged 26

Endovascular treatment of abdominal aortic aneurysms 67–69 years.57 This observational study was the first to look at a broad set of procedure-related interventions. Although patients who under went EVAR were more likely to undergo AAA-related interventions during a 4-year follow-up period (9.0% versus 1.7% with open repair, P <0.001), this finding was partially offset by a reduced likelihood of hospitalizations without surgery for a diagnosis of bowel obstruction or abdominal-wall hernia (8.1% versus 14.2%, P <0.001) and laparotomy-related interventions (4.1% versus 9.7%, P <0.001).57 As in prior studies, EVAR was associated with a shorter length of stay than open repair (3.4 days versus 9.3 days, P <0.001) in this large observational study.57 Owing to the selection of patients aged >65 years, the cohort might not be representative of the general population; however, the size of the study probably overcomes any selection bias. Importantly, the results were comparable to the randomized, controlled trials discussed above, demonstrating that the findings from these aforementioned randomized trials can be generalized. In 2011, Schanzer et al. reported their findings from a study of CTA scans of 10,228 patients who underwent EVAR for AAA from 1999 to 2008 and were included in the M2S, Inc. database.58 This retrospective observational study showed a 41% incidence of sac enlargement 5 years after EVAR, a rate that increased over time.58 The rate of sac enlargement was significantly increased in patients who underwent EVAR from 2004 to 2008 compared with patients who underwent EVAR between 1999 and 2003.58 The overall incidence of any endoleak was 32%, with the majority manifesting during the first postoperative year after EVAR.58 In a letter to the Editor of Circulation, however, Mark Fillinger explained that the M2S, Inc. database is not a consecutive, observational series, and that a large percentage of patients with normal anatomy is purposely excluded.59 In most cases, post-operative CT scans are only submitted to M2S, Inc. when the aneurysm is not shrinking, which leads to a selection bias with limited available follow-up data. As the study investigators had access to measurements, but not to the CT images, the evaluation of the primary pathology is questionable. In addition, sac expansion was not calculated from the first postoperative CT, which inflates the expansion rate relative to standard reporting methods.

Procedural improvements over time

The studies discussed above were conducted over the past 10 years and include the largest, most-cutting-edge trials performed to date. Notably, however, the findings related to outcomes after EVAR have improved over time. Indeed, for the most-recent trial with detailed follow-up—the USA OVER trial—investigators reported low mortality and low reintervention rates, confirming continued improvement of EVAR. Endovascular device technology has been subject to strenuous research and development efforts over the past 20 years. Since the aforementioned studies were undertaken, a number of stent grafts have been developed and improved. 27

2

Chapter 2 Imaging technology has also improved over time, with resultant improvement in patient assessment. Preoperative CTA scanning has evolved to enable thinner slices, improving the accuracy of measurements, and the development of 3D software packages has made measuring and planning of EVAR easier. In addition, the fluoroscopy equipment used for EVAR has improved over time. In many hospitals, mobile C-arms are replaced by fixed X-ray systems in hybrid operating rooms, which provide better image quality. Additionally, physicians themselves have gained experience and revised their criteria for identifying suitable candidates for EVAR. These improvements have likely led to better outcomes now compared with >10 years ago. The clinical evidence accrued thus far gives an insight into surgical outcome rates; however, a number of unsolved issues remain. As the endovascular technique continues to improve, problems related to endoleaks, late device failures, life-long imaging, and late rupture risk must continue to be optimized. All the factors discussed above contribute to the evolving nature of endovascular surgery. This evolution necessitates continued research to evaluate outcomes on an ongoing basis.

Indications for EVAR

The required anatomical criteria listed in the manufacturers’ instructions for use for commercially available AAA endovascular devices approved by the FDA are shown in Table 1.12,60 A commonly used definition of AAA is a maximum aortic diameter of ≥3 cm. For aneurysms <5 cm, a surgical intervention is not indicated; however, the patient should be kept under surveillance. In Europe, as well as in the USA, an aneurysm diameter of ≥5 cm is generally considered an indication for aneurysm repair,26 although early surgical repair of aneurysms <5.5 cm has been shown to offer no advantage over surveillance.61 As discussed above, AAAs rupture at smaller diameters in women than in men and, therefore, women might benefit from early treatment.12 The proportion of patients treated with EVAR has increased over time, with an estimated 77% of intact AAA repairs being undertaken via EVAR in the USA in 2008.8 The manufacturers’ instructions for use for most stent grafts have not changed greatly in the past decade, but many other factors might contribute to the increased use of EVAR: diffusion of new technology from high-volume, early adopting centres to more-remote, lower-volume centres; increased operator experience with the capabilities and limitations of the technology; changing referral patterns; lower-profile and more-flexible devices allowing treatment through small or tortuous iliac arteries; and more-frequent treatment of patients who do not have anatomical criteria deemed appropriate for the procedure in the manufacturers’ instructions for use.

28

Endovascular treatment of abdominal aortic aneurysms Table 1. Anatomical criteria from the instructions for use for AAA endovascular devices approved by the FDA Endovascular device

Neck Year of FDA diameter approval (mm)

Neck length (mm)

Neck Iliac neck angulation length (°) (mm)

Iliac neck diameter (mm)

Ancure™ (EndoVascular Technologies, Inc., USA)*

1999

18–26

≥15

NS

≥20

<13.5

AneuRx®(Medtronic Vascular, Inc., USA)

1999

18–25

≥10‡

≤45

NS

NS

Excluder®(W. L. Gore & Associates, Inc., USA)

2002

19–26

≥15

≤60

≥10

10–18.5

Zenith®(Cook Medical Technologies, USA)

2003

18–28

≥15

≤60

≥15

10–20

Low-permeability Excluder®(W. L. Gore & Associates, Inc., USA)

2004

19–26

≥15

≤60

≥10

10–18.5

Powerlink®(Endologix, Inc., USA)

2004

18–26

≥15

≤60

≥15

8–18

Enlarged-neck Zenith®(Cook Medical Technologies, USA)

2006

18–32

≥15

≤60

≥15

10–20

Talent®(Medtronic Vascular, Inc., USA)

2008

18–32

≥10

≤60

≥15

8–22

Enlarged-neck Powerlink®(Endologix Inc., USA)

2009

18–32

≥15

≤60

≥15

10–23

Enlarged-neck Excluder®(W. L. Gore & Associates, Inc., USA)

2009

19–29

≥15

≤60

≥10

10–18.5

Endurant®(Medtronic Vascular, Inc., USA)

2010

19–32

≥10

≤60

≥15

8–25

Ovation®(Trivascular, Inc, USA)

2011

15.5–30

≥7

45–60

≥10

8–20

Fenestrated Zenith®(Cook Medical Technologies, USA)

2012

19–31

≥4

<45

>30

7–21

Aorfix®(Lombard Medical, UK)

2013

19–29

≥15

≤90

≥15

9–19

*Device discontinued in 2003. ‡Changed to ≥15 mm in 2003. Abbreviations: AAA, abdominal aortic aneurysm; NS, not specified.

New devices and future developments

As EVAR became widely accepted as a safe technique, some physicians began experimenting with off-label use of stent grafts to treat patients deemed unfit for open surgery, but who were also not candidates for standard EVAR. In line with this experimentation, several manufacturers have developed new stent graft 29

2

Chapter 2 systems that—although their effectiveness is not yet proven— might enable expansion of EVAR technology to many of those AAAs that are not anatomically suitable for current endografts with proven efficacy. Other devices have been developed in an attempt to improve outcomes in patients undergoing the procedure. Some of the new stent graft systems and devices are currently used widely, and others are associated with fewer evidence-based data and have not yet achieved broad utilization.

Figure 2.The endoanchor system, which can be used for transmural fixation of an aortic stent graft to the aortic wall at its landing zones

Many approaches to treating aortic aneurysms exist, and a wide variety of devices are available. No clear evidence supports one specific treatment regimen over another and, therefore, specific practice patterns vary between institutions. Our approach is to treat patients with sufficient proximal landing zone at the aortic neck with EVAR in a traditional fashion, using a stent graft of the surgeon’s choosing. For patients with anatomy that does not provide an infrarenal aortic neck for fixation and sealing of a traditional stent graft, the next step depends on the suitability of the patient for open repair and the urgency of the clinical situation. Patients who elect to undergo AAA repair can be treated with an FDA-approved fenestrated or branched stent graft (described below), and emergent cases might be considered for management with off-label use of chimney or sandwich techniques (also described below). If the aortic diameter is large at the level of the target vessels, a branched stent graft can be used. As increasingly complex aortic pathology is treated via endovascular means, innovation has been driven through the use of novel techniques and devices customized in consultation with manufacturers. The surgeon must be judicious in the utilization of novel techniques outside the manufacturers’ instructions for use, to ensure that they are employed with the utmost regard for patient safety. The systems and devices discussed below are divided into those that have been approved by the FDA to be used safely, and those that have been more-recently 30

Endovascular treatment of abdominal aortic aneurysms developed and are in need of further research before device approval. These devices and techniques require advanced endovascular skills and have not been evaluated on a large scale. Although early reports support the use of these techniques in selected patients, limited data are available on their long-term use, and open repair remains the treatment choice in suitable candidates. However, within the setting of a clinical trial, in the hands of experienced physicians investigating promising devices, or for patients who could not otherwise be treated, some of the alternative techniques described below might be considered. New devices approved by the FDA Endostapling In response to concerns about EVAR treatment failure secondary to endoleak and stent graft migration, endostaplers (Figure 2) have been developed to improve durability of the graft by fixation of the graft at its landing zones, and ultimately decrease the need for reintervention. In a prospective, single-centre study published in 2008, the feasibility of endostapler use was evaluated, and 20 of 29 endostaples were found to have been placed successfully.62 Failures occurred in the setting of heavy vessel calcification, where the stapler failed to penetrate the tissue and retracted into the delivery device. In a midterm report published 2 years after the initial analysis, the investigators reported no device failures, migrations, endoleaks, conversions, or secondary procedures during follow-up (mean 18.2 months).63 In a more-recent study of 11 patients with primary stent graft failure secondary to distal stent migration, use of a similar device, called an ‘endoanchor’, with or without additional extender cuffs, resulted in no recurrent migration over a mean 10 months of follow-up, although two patients did require reintervention for persistent endoleaks.64 Overall, early studies have demonstrated endostapler technology to be safe and feasible for primary fixation of stent grafts or management of stent migration. The device received FDA approval in 2011.

Figure 3. A fenestrated stent graft. This type of stent graft overcomes the problem of an insufficient infrarenal neck for stent graft implantation in patients with juxtarenal or suprarenal aortic aneurysms, whilst preserving renal and/or mesenteric perfusion.

Figure 4. A chimney graft. Similarly to the fenestrated stent graft, the chimney graft preserves the renal and/or mesenteric perfusion. Notably, however, the chimney graft is cheaper than the fenestrated stent graft, and has broader applicability for emergent cases because customization is not needed.

31

2

Chapter 2 Fenestrated stent grafts Although EVAR has become the most-commonly per- formed treatment for patients with infrarenal AAA, a considerable number of patients with AAA (15–20%) have more-proximal aneurysms, including in juxtarenal or suprarenal positions.65,66 EVAR was not previously considered a treatment option for patients with these challenging aneurysm configurations, because of the absence of a sufficient infrarenal neck for stent graft implantation. This problem was overcome through the introduction of the first fenestrated and branched stent grafts in 1999.67–69 Since then, technology in this area has advanced rapidly, and fenestrated stent grafts were granted FDA approval in 2012.60 Fenestrated grafts are now available for use above the renal arteries (Figure 3), with technology available for customization at the level of the individual patient. Compared with repair of infrarenal AAA, open repair of juxtarenal AAA is characterized by more-extensive visceral mobilization to achieve adequate exposure of the abdominal aorta. This requirement can lead to prolonged suprarenal aortic clamping and the need for visceral vessel revascularization. Accordingly, a prolonged period of renal ischaemia can result from open repair of juxtarenal AAA, with an associated risk of postoperative acute kidney injury and even dependence on dialysis.66 Through a similar mechanism, bowel ischaemia secondary to temporary occlusion of the superior mesenteric artery can also occur in this setting, resulting in catastrophic complications such as multiorgan failure and death.70,71 In a 2010 meta-analysis that included 21 studies of patients undergoing open repair of nonruptured juxtarenal AAA, the pooled perioperative mortality was 2.9% and the rate of newonset haemodialysis was 3.3%.66 Fenestrated endovascular repairs have been shown to compare favourably to open surgery, with low operative morbidity and mortality.72 In early reports of the use of fenestrated and branched stent grafts, patency rates of the visceral or renal arteries were approximately 95%.73–75 In a study of 650 patients undergoing branched or fenestrated AAA repair, longterm durability was favourable, with 89% freedom from secondary intervention at the 3-year follow-up.76 Subsequent smaller studies and reviews of the literature have further demonstrated these benefits. In a multicentre study of 134 patients treated with fenestrated stent grafts from 2004 to 2009, 30-day mortality was 2% and mortality during a median follow-up of 15 months was 9%.77 No ruptured aneurysms were reported.77 In another study of 100 patients with short-necked or juxtarenal AAA, who were treated with fenestrated EVAR from 2001 to 2009, 30-day mortality was 1% and, again, no ruptured aneurysms were reported.78 A 2012 systematic review that included nine observational studies involving 629 patients demonstrated that EVAR using fenestrated stent grafts was associated with low 30-day mortality (2.1%).79 In this analysis, the pooled estimates for branch-vessel patency, technical success, and reinterventions at a mean followup of 15–25 months were 93.2%, 90.4%, and 17.8%, respectively.79 Compared with standard EVAR, a disadvantage of EVAR using fenestrated stent grafts is that these grafts often require a high degree of customization, which increases cost and makes the technique prohibitively expensive for emergent cases. In addition, technical considerations such as severely angulated necks, aortic thrombus in the area of target vessels, and small, calcified target vessels might 32

Endovascular treatment of abdominal aortic aneurysms also prohibit the use of fenestrated grafts. Despite these limitations, fenestrated and branched stent grafts have increased the applicability of EVAR to patients who would otherwise be subject to highly morbid open-repair procedures. According to a study involving 26 consulting physicians located across the UK, who reviewed 192 cases of EVAR using fenestrated stent grafts, consensus on indications for EVAR using fenestrated stent grafts occurs in almost 70% of cases.80 The reviewing physicians agreed on suitability for EVAR using fenestrated stent grafts in areas considered to be associated with moderate risk from open repair and the need for suprarenal clamping. EVAR using fenestrated stent grafts was, on the other hand, less likely to be considered indicated in patients aged ≥85 years with aneurysms ≤6 cm and in cases with short-necked infrarenal aortic aneurysms.80 Ideally, a prospective, randomized, controlled trial comparing EVAR using fenestrated stent grafts with open repair should be performed. Devices and techniques not yet FDA approved The chimney technique An alternative to the fenestrated stent graft for patients with challenging anatomy is the chimney graft (Figure 4), which also preserves the renal and/ or mesenteric perfusion. The chimney graft is cheaper than the fenestrated stent graft, and has broader applicability for emergent cases because of the lack of need for customization.81,82 Similarly, patients ineligible for EVAR using fenestrated stent grafts because of anatomic exclusion criteria can potentially be treated with a chimney graft. Covered or uncovered chimney stent grafts can be deployed via brachial or axillary access while aortic stent grafts are deployed via traditional femoral access.83 A systematic review involving data for 75 patients who had undergone EVAR using a chimney procedure (96 chimney grafts used) demonstrated a technical success rate of 98.9%, although operative techniques varied considerably between studies.84 During follow-up that ranged from 2 days to 54 months, three deaths occurred as late complications and three chimney grafts occluded, necessitating two reinterventions.84 Long-term follow-up is not yet available, and FDA approval and CE mark have not yet been attained. However, the chimney procedure might one day offer an acceptable endovascular treatment option for patients with juxtarenal or suprarenal AAA, particularly those who are unfit for open surgery in an emergency setting or when fenestrated grafts are unavailable.81,82,84–86 The sandwich technique The sandwich technique involves the use of stents placed in the aortic branches, like chimneys, that are ‘sandwiched’ between two aortic stent grafts (Figure 5a).87–89 This configuration permits a stable position of coaxial grafts between two aortic stents and is an attractive alternative for patients with anatomy not amenable to conventional EVAR. The technique allows the revascularization of up to four side branches, as opposed to the two permitted by the chimney technique. Notably, however, the revascularization of four side branches using 33

2

Chapter 2 a double two-chimney approach (Figure 5b) has been reported in the past few years.89,90 Although long-term outcomes are not yet available for either the sandwich technique or the double two-chimney approach, and these strategies are performed off-label, they hold potential for patients with difficult anatomy, especially given that they do not require customization. Notably, however, these techniques have only been used in situations not listed in the manufacturers’ instructions for use. a

b

Figure 5. Endovascular aneurysm repair strategies that enable the revascularization of up to four aortic side branches. a | In the sandwich technique, which can be used in patients not deemed suitable for conventional endovascular aneurysm repair, stents placed in the aortic branches are ‘sandwiched’ between two aortic stent grafts. b | The double two-chimney approach is an alternative to the sandwich technique. Of note, long-term outcomes are not yet available for the sandwich technique or the double two-chimney technique, and both techniques have only been used in situations not listed in the manufacturers’ instructions for use.

Off-the-shelf fenestrated stent grafts Currently, fenestrated and branched stent grafts must be customized for individual patients, a process that requires meticulous preoperative planning. Such customization can lead to substantial treatment delay, with risk of rupture in the intervening period. Such delays preclude the use of the technology in the acute setting. To allow more patients to be treated in the acute setting, standard off-the-shelf fenestrated and branched (Figure 6) stent grafts have been developed for treatment of patients with complex AAA. The T-branch stent graft is an off-the- shelf device that can be used for thoracoabdominal AAA treatment, whereas the P-branch off-the-shelf device can be used for aneurysms that limit just below the superior mesenteric artery. During the period 2008–2010, Chuter et al. studied 28 patients who underwent AAA repair using custom- made or standard off-the-shelf fenestrated stent grafts (n = 14 for each group).91 No cases of stent migration, aneurysm dilatation, rupture, or perioperative death 34

Endovascular treatment of abdominal aortic aneurysms were reported for either group.91 In another study conducted by Kitagawa and colleagues,92 16 patients were recruited in 2011–2012 for treatment with an off-the- shelf P-branch device. The treatment was technically successful in all patients, taking into consideration that one patient had an occluded renal artery during follow-up, which was successfully recanalized.92 In 2012, Resch et al. reported on the successful treatment of seven patients with juxtarenal and suprarenal aneurysms using fenestrated off-the-shelf devices; no deaths and no endoleaks occurred during the 30-day follow-up period.93 Off-the-shelf fenestrated and branched stent grafts have the potential to reduce the need for preoperative graft customization and might be a solution for patients with AAA not suitable for traditional EVAR, particularly in the emergent setting. As limited clinical evidence is currently available, however, more research is needed before widespread implementation of this technique.

Figure 6. An off-the-shelf branched stent graft. These ‘ready to go’ stent grafts have been developed for treatment of patients with complex abdominal aortic aneurysm in the acute setting, as stent customization can lead to substantial treatment delay. a

b

c

d

Figure 7. The multilayer stent graft. a | Blood flow through a saccular aortic aneurysm. b | A saccular aortic aneurysm with an increased flow velocity. c | A saccular aortic aneurysm treated with a multilayer stent, which decreases the flow velocity into the aneurysm. d | The blood flow though the multilayer stent as the aortic aneurysm is thrombosed.

Multilayer stent For management of complex aneurysms involving aortic branches, a new multilayer self-expanding stent technology (Figure 7) has been developed. The technology has flow-diverting capabilities that can exclude the aneurysm while 35

2

Chapter 2 preserving collateral circulation via a series of interconnected, braided layers. Similar to the flow dynamics seen in saccular aneurysms with clinically relevant thrombosis, turbulent flow through the thrombotic segment creates a pressure gradient such that flow is drawn into the aortic branch. In a similar fashion, this multilayered stent graft could achieve diversion of flow into the vital visceral branches involved. Given that this technology has been developed over the past 5 years, however, little is known about its reliability and effectivness.94–96 In 2012, a systematic review involving data from 12 reports of patients (n = 35; 38 aneurysms) treated with flow-diverting stents was published; three of these patients had been treated for AAA.96 Each of these three cases achieved technical success in stent deployment, aneurysm thrombosis and shrinkage, and patency of branch vessels.96 Further investigation for flow-diverting stents in the use of AAA management is ongoing. Although this off-the-shelf stent might potentially be used in patients with AAA not amenable to treatment via other previously described endovascular or open methods, the technique has not undergone the rigors of clinical testing and the effectiveness remains to be proven. A sac-anchoring endoprosthesis Unlike open AAA repair, EVAR leaves the aneurysm sac itself untreated, which allows the possibility of persistent blood flow (type 2 endoleak) and stent migration. The sac-anchoring endoprosthesis (Figure 8) is an endoluminal device that eliminates the endoleak space via obliteration of the aneurysm sac. This technology utilizes two femorally inserted stent grafts with polymer-filled endobags on the outside of the stents. The polymer- filled endobags enlarge to a pressure preset to support the stents within the aneurysm sac. Such technology might be employed in the setting of an infrarenal AAA with adverse anatomic features.97,98 In a prospective clinical trial published in 2011, perioperative aneurysm-related mortality was 2.9% among 34 patients undergoing AAA with sac anchoring.98 One patient had a second- ary procedure because of an endoleak. CT surveillance for >2 years revealed no change in aneurysm size, no new endoleaks, and no change in stent graft position.98 The sac-anchoring endoprosthesis has been available in Europe since 2012, but is still awaiting FDA approval. Other changes to stent graft design Developments in stent grafts have facilitated their deployment through smaller access sites than was previously possible, and with improved sealing and fixation in an ever-increasing cohort of patients. In addition, fabrics with reduced porosity have been developed to enable a rapid reduction in aneurysm size.99 Reduced stent graft profile and other improvements have thus far been achieved without adversely impacting the durability and sustainability of the stent graft itself.100 Low-profile stents with a delivery system diameter of 14 French have received CE Mark approval. To optimize sealing and fixation, low-profile stent graft systems have been designed with the use of transmural fixation with endoanchors,101,102 or with polymer-filled sealing rings,100 to improve stability in placement. Also, single low-profile introducer sheaths combined with a thin-walled, low-porosity 36

Endovascular treatment of abdominal aortic aneurysms fabric have been developed to reduce aneurysm size.103

2

Thrombus

Endoframes supporting

Figure 8. The sac-anchoring system, consisting of two femorally inserted stent grafts with polymer-filled endobags on the outside of the stent. The polymer-filled endobags enlarge to a pressure preset to support the stents within the aneurysm sac. The technology eliminates the endoleak space via obliteration of the aortic aneurysm sac.

Conclusions

Over the past 20 years, endovascular technology has propelled EVAR from an obscure technology with limited applicability to being the standard of care in an increasing number of AAA indications. EVAR utilization has broadened so dramatically that up to 77% of patients are now treated with endovascular repair.8 EVAR is not only becoming the first choice of treatment for patients with infrarenal AAAs, but is also gaining traction for the management of increasingly challenging anatomy. Interestingly, the survival benefit in terms of 30-day operative mortality for EVAR compared with open repair remains most convincing in patients with few or no comorbidities.7 As indications expand and devices evolve, ongoing research is critical to ensure effective and evidence-based utilization of this promising technology.

37

Chapter 2

References 1. 2. 3. 4.

5.

6. 7. 8. 9.

10.

11. 12. 13. 14. 15. 38

Volodos, N. L., Shekhanin, V. E., Karpovich, I. P., Troian, V. I. & Gur’ev, I. A. A self-fixing synthetic blood vessel endoprosthesis [Russian]. Vestn. Khir. Im. I. I. Grek. 137, 123–125 (1986). Volodos, N. L. Historical perspective: the first steps in endovascular aortic repair: how it all began. J. Endovasc. Ther. 20 (Suppl. 1), I3–I23 (2013). Parodi, J. C., Palmaz, J. C. & Barone, H. D. Transfemoral intraluminal graft implantation for abdominal aortic aneurysms. Ann. Vasc. Surg. 5, 491–499 (1991). Eisenack, M., Umscheid, T., Tessarek, J., Torsello, G. F. & Torsello, G. B. Percutaneous endovascular aortic aneurysm repair: a prospective evaluation of safety, efficiency, and risk factors. J. Endovasc. Ther. 16, 708–713 (2009). Drury, D., Michaels, J. A., Jones, L. & Ayiku, L. Systematic review of recent evidence for the safety and efficacy of elective endovascular repair in the management of infrarenal abdominal aortic aneurysm. Br. J. Surg. 92, 937–946 (2005). Nowygrod, R. et al. Trends, complications, and mortality in peripheral vascular surgery. J. Vasc. Surg. 43, 205–216 (2006). Brown, L. C., Greenhalgh, R. M., Howell, S., Powell, J. T. & Thompson, S. G. Patient fitness and survival after abdominal aortic aneurysm repair in patients from the UK EVAR trials. Br. J. Surg. 94, 709–716 (2007). Schermerhorn, M. L. et al., Changes in abdominal aortic aneurysm rupture and short- term mortality, 1995–2008: a retrospective observational study. Ann. Surg. 256, 651–658 (2012). Cornuz, J., Sidoti Pinto, C., Tevaearai, H. & Egger, M. Risk factors for asymptomatic abdominal aortic aneurysm: systematic review and metaanalysis of population-based screening studies. Eur. J. Public Health 14, 343–349 (2004). Lederle, F. A. et al. The aneurysm detection and management study screening program: validation cohort and final results. Aneurysm Detection and Management Veterans Affairs Cooperative Study Investigators. Arch. Intern. Med. 160, 1425–1430 (2000). Powell, J. T. & Greenhalgh, R. M. Clinical practice. Small abdominal aortic aneurysms. N. Engl. J. Med. 348, 1895–1901 (2003). Lo, R. C. et al. Gender differences in abdominal aortic aneurysm presentation, repair, and mortality in the Vascular Study Group of New England. J. Vasc. Surg. 57, 1261–1268 (2013). Mofidi, R. et al. Influence of sex on expansion rate of abdominal aortic aneurysms. Br. J. Surg. 94, 310–314 (2007). Egorova, N. N. et al. Effect of gender on long-term survival after abdominal aortic aneurysm repair based on results from the Medicare national database. J. Vasc. Surg. 54, 1–12 (2011). Semmens, J. B., Norman, P. E., Lawrence-Brown, M. M., Bass, A. J. &

Endovascular treatment of abdominal aortic aneurysms

16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

Holman, C. D. Population-based record linkage study of the incidence of abdominal aortic aneurysm in Western Australia in 1985–1994. Br. J. Surg. 85, 648–652 (1998). Fowkes, F. G., Macintyre, C. C. & Ruckley, C. V. Increasing incidence of aortic aneurysms in England and Wales. BMJ 298, 33–35 (1989). Filipovic, M. et al. Trends in mortality and hospital admission rates for abdominal aortic aneurysm in England and Wales, 1979–1999. Br. J. Surg. 92, 968–975 (2005). Acosta, S. et al. Increasing incidence of ruptured abdominal aortic aneurysm: a population-based study. J. Vasc. Surg. 44, 237–243 (2006). Sandiford, P., Mosquera, D. & Bramley, D. Trends in incidence and mortality from abdominal aortic aneurysm in New Zealand. Br. J. Surg. 98, 645–651 (2011). Norman, P. E., Spilsbury, K. & Semmens, J. B. Falling rates of hospitalization and mortality from abdominal aortic aneurysms in Australia. J. Vasc. Surg. 53, 274–277 (2011). Giles, K. A. et al. Decrease in total aneurysm- related deaths in the era of endovascular aneurysm repair. J. Vasc. Surg. 49, 543–550 (2009). Anjum, A. & Powell, J. T. Is the incidence of abdominal aortic aneurysm declining in the 21st century? Mortality and hospital admissions for England & Wales and Scotland. Eur. J. Vasc. Endovasc. Surg. 43, 161–166 (2012). Anjum, A., von Allman, R., Greenhalgh, R. & Powell, J. T. Explaining the decrease in mortality from abdominal aortic aneurysm rupture. Br. J. Surg. 99, 637–645 (2012). Wilmink, A. B., Forshaw, M., Quick, C. R., Hubbard, C. S. & Day, N. E. Accuracy of serial screening for abdominal aortic aneurysms by ultrasound. J. Med. Screen. 9, 125–127 (2002). Rudarakanchana, N. & Powell, J. T. Advances in imaging and surveillance of AAA: when, how, how often? Prog. Cardiovasc. Dis. 56, 7–12 (2013). Moll, F. L. et al. Management of abdominal aortic aneurysms clinical practice guidelines of the European society for vascular surgery. Eur. J. Vasc. Endovasc. Surg. 41 (Suppl. 1), S1–S58 (2011). Beeman, B. R. et al. Duplex ultrasound factors predicting persistent type II endoleak and increasing AAA sac diameter after EVAR. J. Vasc. Surg. 52, 1147–1152 (2010). Iezzi, R. et al. Contrast-enhanced ultrasound versus color duplex ultrasound imaging in the follow-up of patients after endovascular abdominal aortic aneurysm repair. J. Vasc. Surg. 49, 552–560 (2009). Ten Bosch, J. A. et al. Contrast-enhanced ultrasound versus computed tomographic angiography for surveillance of endovascular abdominal aortic aneurysm repair. J. Vasc. Interv. Radiol. 21, 638–643 (2010). Wilson, S. R., Greenbaum, L. D. & Goldberg, B. B. Contrast-enhanced ultrasound: what is the evidence and what are the obstacles? AJR Am. J. Roentgenol. 193, 55–60 (2009). Sun, Z. Helical CT angiography of fenestrated stent grafting of abdominal aortic aneurysms. Biomed. Imaging. Interv. J. 5, e3 (2009). 39

2

Chapter 2 32. Rydberg, J. et al. Endovascular repair of abdominal aortic aneurysms: assessment with multislice CT. AJR Am. J. Roentgenol. 177, 607–614 (2001). 33. Sun, Z. Multislice CT angiography in abdominal aortic aneurysm treated with endovascular stent grafts: evaluation of 2D and 3D visualisations. Biomed. Imaging Interv. J. 3, e20 (2007). 34. Sun, Z. H. Abdominal aortic aneurysm: treatment options, image visualizations and follow-up procedures. J. Geriatr. Cardiol. 9, 49–60 (2012). 35. van Keulen, J. W., Moll, F. L. & van Herwaarden, J. A. Tips and techniques for optimal stent graft placement in angulated aneurysm necks. J. Vasc. Surg. 52, 1081–1086 (2010). Vasc. Endovasc. Surg. 38, 586–596 (2009). 36. van Keulen, J. W., van Prehn, J., Prokop, M., Moll, F. L. & van Herwaarden, J. A. Dynamics of the aorta before and after endovascular aneurysm repair: a systematic review. Eur. J. 37. Neschis, D. G. et al. The role of magnetic resonance angiography for endoprosthetic design. J. Vasc. Surg. 33, 488–494 (2001). 38. Upchurch, G. R. Jr & Criado, E. (eds) Aortic Aneurysms: Pathogenesis and Treatment (Humana Press, 2009). 39. Habets, J. et al. Magnetic resonance imaging is more sensitive than computed tomography angiography for the detection of endoleaks after endovascular abdominal aortic aneurysm repair: a systematic review. Eur. J. Vasc. Endovasc. Surg. 45, 340–350 (2013). 40. Korosec, F. R., Frayne, R., Grist, T. M. & Mistretta, C. A. Time-resolved contrast- enhanced 3D MR angiography. Magn. Reson. Med. 36, 345–351 (1996). 41. Cohen, E. I. et al. Time-resolved MR angiography for the classification of endoleaks after endovascular aneurysm repair. J. Magn. Reson. Imaging 27, 500–503 (2008). 42. Lookstein, R. A., Goldman, J., Pukin, L. & Marin, M. L. Time-resolved magnetic resonance angiography as a noninvasive method to characterize endoleaks: initial results compared with conventional angiography. J. Vasc. Surg. 39, 27–33 (2004). 43. Lee, J. T. & White, R. A. Basics of intravascular ultrasound: an essential tool for the endovascular surgeon. Semin. Vasc. Surg. 17, 110–118 (2004). 44. Marrocco, C. J. et al. Intravascular ultrasound. Semin. Vasc. Surg. 25, 144–152 (2012). 45. Prinssen, M. et al. A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N. Engl. J. Med. 351, 1607–1618 (2004). 46. Blankensteijn, J. D. et al. Two-year outcomes after conventional or endovascular repair of abdominal aortic aneurysms. N. Engl. J. Med. 52, 2398–2405 (2005). 47. De Bruin, J. L. et al. Long-term outcome of open or endovascular repair of abdominal aortic aneurysm. N. Engl. J. Med. 362, 1881–1889 (2010). 48. Greenhalgh, R. M. et al. Comparison of endovascular aneurysm repair 40

Endovascular treatment of abdominal aortic aneurysms

49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60.

61. 62. 63.

with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day operative mortality results: randomised controlled trial. Lancet 364, 843–848 (2004). EVAR trial participants. Endovascular aneurysm repair versus open repair in patients with abdominal aortic aneurysm (EVAR trial 1): randomised controlled trial. Lancet 365, 2179–2186 (2005). United Kingdom EVAR Trial Investigators. Endovascular versus open repair of abdominal aortic aneurysm. N. Engl. J. Med. 362, 1863–1871 (2010). EVAR trial participants. Endovascular aneurysm repair and outcome in patients unfit for open repair of abdominal aortic aneurysm (EVAR trial 2): randomised controlled trial. Lancet 365, 2187–2192 (2005). Brown, L. C. et al. The UK EndoVascular Aneurysm Repair (EVAR) trials: randomised trials of EVAR versus standard therapy. Health Technol. Assess. 16, 1–218 (2012). United Kingdom EVAR Trial Investigators. Endovascular repair of aortic aneurysm in patients physically ineligible for open repair. N. Engl. J. Med. 362, 1872–1880 (2010). Lederle, F. A. et al. Outcomes following endovascular vs open repair of abdominal aortic aneurysm: a randomized trial. JAMA 302, 1535–1542 (2009). Lederle, F. A. et al. Long-term comparison of endovascular and open repair of abdominal aortic aneurysm. N. Engl. J. Med. 367, 1988–1997 (2012). Becquemin, J. P. et al., A randomized controlled trial of endovascular aneurysm repair versus open surgery for abdominal aortic aneurysms in low- to moderate-risk patients. J. Vasc. Surg. 53, 1167–1173 (2011). Schermerhorn, M. L. et al. Endovascular vs. open repair of abdominal aortic aneurysms in the Medicare population. N. Engl. J. Med. 358, 464– 474 (2008). Schanzer, A. et al. Predictors of abdominal aortic aneurysm sac enlargement after endovascular repair. Circulation 123, 2848–2855 (2011). Fillinger, M. Letter by Fillinger regarding article, “Predictors of abdominal aortic aneurysm sac enlargement after endovascular repair”. Circulation 125, e341 (2012). US Food and Drug Administration. Medical devices: device approvals and clearances [online], http://www.fda.gov/MedicalDevices/ ProductsandMedicalProcedures/ DeviceApprovalsandClearances/default. htm (2013). Cao, P. et al., Comparison of surveillance versus aortic endografting for small aneurysm repair (CAESAR): results from a randomised trial. Eur. J. Vasc. Endovasc. Surg, 41, 13–25 (2011). Donas, K. P., Kafetzakis, A., Umsheid, T., Tessarek, J. & Torsello, G. Vascular endostapling: new concept for endovascular fixation of aortic stent-grafts. J. Endovasc. Ther. 15, 499–503 (2008). Donas, K. P. & Torsello, G. Midterm results of the Anson Refix endostapling fixation system for aortic stent-grafts. J. Endovasc. Ther. 17, 320–323 (2010). 41

2

Chapter 2 64. Avci, M. et al. The use of endoanchors in repair EVAR cases to improve proximal endograft fixation. J. Cardiovasc. Surg. (Torino) 53, 419–426 (2012). 65. Taylor, S. M., Mills, J. L. & Fujitani, R. M. The juxtarenal abdominal aortic aneurysm. A more common problem than previously realized? Arch. Surg. 129, 734–737 (1994). 66. Jongkind, V. et al. Juxtarenal aortic aneurysm repair. J. Vasc. Surg. 52, 760–767 (2010). 67. Faruqi, R. M. et al. Endovascular repair of abdominal aortic aneurysm using a pararenal fenestrated stent-graft. J. Endovasc. Surg. 6, 354–358 (1999). 68. Browne, T. F. et al. A fenestrated covered suprarenal aortic stent. Eur. J. Vasc. Endovasc. Surg. 18, 445–449 (1999). 69. Anderson, J. L., Berce, M. & Hartley, D. E. Endoluminal aortic grafting with renal and superior mesenteric artery incorporation by graft fenestration. J. Endovasc. Ther. 8, 3–15 (2001). 70. Nypaver, T. J. et al. Repair of pararenal abdominal aortic aneurysms. An analysis of operative management. Arch. Surg. 128, 803–811 (1993). 71. Sarac, T. P. et al. Contemporary results of juxtarenal aneurysm repair. J. Vasc. Surg. 36, 1104–1111 (2002). 72. Katsargyris, A., Oikonomou, K., Klonaris, C., Töpel, I. & Verhoeven, E. L. Comparison of outcomes with open, fenestrated, and chimney graft repair of juxtarenal aneurysms: are we ready for a paradigm shift? J. Endovasc. Ther. 20, 159–169 (2013). 73. Scurr, J. R. et al. Fenestrated endovascular repair for juxtarenal aortic aneurysm. Br. J. Surg. 95, 326–332 (2008). 74. Bicknell, C. D. et al., Treatment of complex aneurysmal disease with fenestrated and branched stent grafts. Eur. J. Vasc. Endovasc. Surg. 37, 175–181 (2009). 75. O’Neill, S. et al. A prospective analysis of fenestrated endovascular grafting: intermediate- term outcomes. Eur. J. Vasc. Endovasc. Surg. 32, 115–123 (2006). 76. Mastracci, T. M., Greenberg, R. K., Eagleton, M. J. & Hernandex, A. V. Durability of branches in branched and fenestrated endografts. J. Vasc. Surg. 57, 926–933 (2013). 77. Amiot, S. et al., Fenestrated endovascular grafting: the French multicentre experience. Eur. J. Vasc. Endovasc. Surg. 39, 537–544 (2010). 78. Verhoeven, E. L. et al. Fenestrated stent grafting for short-necked and juxtarenal abdominal aortic aneurysm: an 8-year single-centre experience. Eur. J. Vasc. Endovasc. Surg. 39, 529–536 (2010). 79. Linsen, M. A., Jongkind, V., Nio, D., Hoksbergen, A. W. & Wisselink, W. Pararenal aortic aneurysm repair using fenestrated endografts. J. Vasc. Surg. 56, 238–246 (2012). 80. Cross, J. et al. Indications for fenestrated endovascular aneurysm repair. Br. J. Surg. 99, 217–224 (2012). 81. Ohrlander, T. et al. The chimney graft: a technique for preserving or rescuing 42

Endovascular treatment of abdominal aortic aneurysms 82.

83. 84. 85. 86. 87. 88.

89.

90.

91. 92. 93. 94. 95. 96.

aortic branch vessels in stent-graft sealing zones. J. Endovasc. Ther. 15, 427–432 (2008). Patel, R. P., Katsargyris, A., Verhoeven, E. L., Adam, D. J. & Hardman, J. A. Endovascular aortic aneurysm repair with chimney and snorkel grafts: indications, techniques and results. Cardiovasc. Intervent. Radiol. 36, 1443–1451 (2013). Criado, F. J. Chimney grafts and bare stents: aortic branch preservation revisited. J. Endovasc. Ther. 14, 823–824 (2007) Tolenaar, J. L. et al. The chimney graft, a systematic review. Ann. Vasc. Surg. 26, 1030–1038 (2012). Tolenaar, J. L., Zandvoort, H. J., Moll, F. L. & van Herwaarden, J. A. Technical considerations and results of chimney grafts for the treatment of juxtarenal aneurysms. J. Vasc. Surg. 58, 607–615 (2013). Bruen, K. J., Feezor, R. J., Daniels, M. J., Beck, A. W. & Lee, W. A. Endovascular chimney technique versus open repair of juxtarenal and suprarenal aneurysms. J. Vasc. Surg. 53, 895–904 (2011). Allaqaband, S., Kumar, A. & Bajwa, T. A novel technique of aortomonoiliac AAA repair in patients with a single patent iliac artery: a “stent-graft sandwich”. J. Endovasc. Ther. 11, 550–552 (2004). Lobato, A. C. Sandwich technique for aortoiliac aneurysms extending to the internal iliac artery or isolated common/internal iliac artery aneurysms: a new endovascular approach to preserve pelvic circulation. J. Endovasc. Ther. 18, 106–111 (2011). Kolvenbach, R. R., Yoshida, R., Pinter, L., Zhu, Y. & Lin, F. Urgent endovascular treatment of thoraco-abdominal aneurysms using a sandwich technique and chimney grafts—a technical description. Eur. J. Vasc. Endovasc. Surg. 41, 54–60 (2011). Tolenaar, J. L., Zandvoort, H. J., Hazenberg, C. E., Moll, F. L. & van Herwaarden, J. A. The double two-chimney technique for complete renovisceral revascularization in a suprarenal aneurysm. J. Vasc. Surg. 58, 478–481 (2013). Chuter, T. A., Hiramoto, J. S., Park, K. H. & Reilly, L. M. The transition from custom-made to standardized multibranched thoracoabdominal aortic stent grafts. J. Vasc. Surg. 54, 660–667 (2011). Kitagawa, A., Greenberg, R. K., Eagleton, M. J. & Mastracci, T. M. Zenith p-branch standard fenestrated endovascular graft for juxtarenal abdominal aortic aneurysms. J. Vasc. Surg. 58, 291–300 (2013). Resch, T. A. et al. Development of off-the-shelf stent grafts for juxtarenal abdominal aortic aneurysms. Eur. J. Vasc. Endovasc. Surg. 43, 655–660 (2012). Henry, M. et al. Treatment of renal artery aneurysm with the multilayer stent. J. Endovasc. Ther. 15, 231–236 (2008). Zhang, Y. X., Lu, Q. S. & Jing, Z. P. Multilayer stents, a new progress in the endovascular treatment of aneurysms. Chin. Med. J. (Engl.) 126, 536–541 (2013). Sfyroeras, G. S. et al. Flow-diverting stents for the treatment of arterial 43

2

Chapter 2 aneurysms. J. Vasc. Surg. 56, 839–846 (2012). 97. Donayre, C. E. et al. Initial clinical experience with a sac-anchoring endoprosthesis for aortic aneurysm repair. J. Vasc. Surg. 53, 574–582 (2011). 98. Krievins, D. K. et al. EVAR using the Nellix Sac- anchoring endoprosthesis: treatment of favourable and adverse anatomy. Eur. J. Vasc. Endovasc. Surg. 42, 38–46 (2011). 99. Chuter, T. A. Stent-graft design: the good, the bad and the ugly. Cardiovasc. Surg. 10, 7–13 (2002). 100. Mehta, M. et al. One-year outcomes from an international study of the Ovation Abdominal Stent Graft System for endovascular aneurysm repair. J. Vasc. Surg. http://dx.doi.org/10.1016/j.jvs.2013.06.065 101. Perdikides, T. et al. Primary endoanchoring in the endovascular repair of abdominal aortic aneurysms with an unfavorable neck. J. Endovasc. Ther. 19, 707–715 (2012). 102. Kasprzak, P., Pfister, K., Janotta, M. & Kopp, R. EndoAnchor placement in thoracic and thoracoabdominal stent-grafts to repair complications of nonalignment. J. Endovasc. Ther. 20, 471–480 (2013). 103. Endologix. AFX™ endovascular AAA system (bifurcated stent graft models and accessory models): instructions for use [online], http://www.endologix. com/pdf/AFX%20StentGraft&Access%20English%20C00541RevA%20 -%20PW%20protected.pdf (2010).

44

Endovascular treatment of abdominal aortic aneurysms

2

45

CHAPTER 3

Long-term Outcomes for Endovascular vs. Open Repair of Abdominal Aortic Aneurysms

Marc L. Schermerhorn, Dominique B. Buck, A. James O’Malley, Thomas Curran, Lawrence Zaborski, John C. McCallum, Jeremy Darling, Bruce E. Landon. Accepted N Engl J Med.

Chapter 3

Abstract Background Randomized trials and observational studies have demonstrated lower perioperative morbidity and mortality with endovascular compared to open repair for abdominal aortic aneurysm, but the survival benefit is not sustained. Methods We studied perioperative and long-term survival, reinterventions, and complications after endovascular and open repair of abdominal aortic aneurysm in propensity-score-matched cohorts of Medicare beneficiaries undergoing repair from 2001 to 2008, with follow-up through 2009. Results There were 39,966 matched patients undergoing open or endovascular repair in each cohort. Overall perioperative mortality was 1.6% for endovascular repair and 5.2% for open repair (p<.001). From 2001 to 2008, perioperative mortality with endovascular repair decreased by 0.8% (p=0.001) and with open repair decreased by 0.6% (p=0.013), and conversion to open repair dropped from 2.2 % to 0.3% (P<.001). Survival was significantly better after endovascular repair through the first 3 years of follow-up after which time the survival probabilities were indistinguishable. At 8 years, however, rupture was seen in 5.4% of patients after endovascular repair and in 1.4% after open repair (P<.001). Conclusions Compared to open repair, endovascular repair has a substantial early survival advantage that is gradually eroded over time. Late rupture after endovascular repair is a concern that merits further study. Endovascular repair outcomes are improving over time.

48

Long-term Outcomes of A Repair

Introduction

The use of endovascular repair of abdominal aortic aneurysms (AAAs) continues to increase, accounting for 78% of all intact repairs by 2010.1,2 Randomized controlled trials comparing endovascular to open repair generally have shown a perioperative benefit of endovascular over open repair.3-5 Long-term survival, however, is similar for the two approaches.6-9 As long-term data accumulate concerns have arisen about increased late failure of endovascular repair leading to rupture and higher reintervention rates. In previous analyses using Medicare data, which accounts for over 83% of abdominal aortic aneurysm repairs nationally,10 we replicated these findings in usual clinical practice, but also found that at 4-year follow-up there were increased aneurysm related reinterventions in the endovascular repair group that were offset by increased laparotomy-related complications in the open repair group. Long-term data from the randomized OVER trial confirmed this finding,7 while long-term data derived from EVAR-1 and DREAM did not report all laparotomyrelated complications.6,9 Importantly, our prior analysis did not account for prior laparotomy, which might influence treatment choice.11 In this study, we compare long-term outcomes for up to 8 years after endovascular versus open repair using propensity-score matched cohorts from Medicare to evaluate late survival and complications, accounting for prior laparotomy. We also examine whether event rates have changed over time as practitioners gain experience with this evolving technology.

Methods Patients We identified all traditional Medicare beneficiaries who underwent elective AAA repair during the period from January 1, 2001-December 31, 2008. Patients were included if they were continuously enrolled in traditional Medicare parts A and B for at least two years prior to repair, had a discharge diagnosis of AAA, and underwent open or endovascular repair. We excluded all patients with ruptured AAA, thoracic aneurysms, thoracoabdominal aortic aneurysms, or aortic dissections. In addition we excluded those who underwent visceral or renal bypass (Appendix List). Beneficiaries who enrolled in Medicare Advantage during the follow-up period were censored from the analyses of complications and reinterventions, because subsequent claims data were not available. To improve coding accuracy for repair type, we examined physicians’ claims corresponding to the hospitalization. In cases in which the codes from the hospital and physician conflicted, we assigned procedures on the basis of the physicians’ claims. Our study was approved by the institutional review board at Harvard Medical School. 49

3

Chapter 3 Creating matched cohorts To control for the non-random assignment of patients, we constructed logistic regression models predicting the likelihood of endovascular repair (the propensity score), and matched patients in each cohort by this score. We used as explanatory variables all available demographic and clinical characteristics for beneficiaries available from the prior two years, not including diagnoses from the index admission.12 We also identified prior abdominal operations during the previous two years, classified as a 3-level categorical variable based on the likelihood of late complications (laparoscopic, minor open, major open), as well as prior retroperitoneal and hernia operations, which also confer an increased risk of subsequent complications.13 We measured the rates of coexisting conditions using a version of the Elixhauser algorithm adapted to include diagnoses from the outpatient setting.14,15 The propensity score models also included an indicator variable for year to control for time. To ensure close matches, we required that the estimated log-odds scores predicting endovascular repair for matched pairs be within 0.60 SD of each other, which ensures removal of approximately 90% of bias in estimates of effects due to differences in observed covariates. 16,17 Outcomes Mortality Perioperative mortality was defined as death during the index admission or within 30 days after surgery. Long-term mortality included all deaths during the followup period, which is available from the Beneficiary Summary File. Perioperative Outcomes and Complications We identified perioperative surgical complications (e.g. conversion from endovascular to open repair, return to the operating room, etc.) and medical complications (e.g. myocardial infarction, pneumonia) using ICD-9-CM diagnostic and complication codes, as well as physicians’ current-procedural-terminology (CPT) codes (Appendix). We also recorded length of stay and whether patients were discharged home . Postoperative Outcomes and Reinterventions We identified all hospitalizations and outpatient interventions occurring after repair that potentially were related to the AAA, including hospitalizations for rupture, major reinterventions (e.g. open repair of the aneurysm/pseudoaneurysm, repair of graft–enteric fistula or graft infection), and minor reinterventions (e.g., stent– graft extension, embolization, aortic or iliac angioplasty, graft thrombectomy). We also identified laparotomy-related complications requiring procedures, (e.g. lysis of adhesions, bowel resection, and repair of abdominal-wall hernia) and readmissions for bowel obstruction without operation. Statistical analyses We compared the characteristics of the unmatched cohorts using chi-square and t-tests as appropriate. To account for the dependence of the matched pairs, post50

Long-term Outcomes of A Repair matching differences were tested with McNemar’s test for categorical variables and paired t-tests for continuous variables. We estimated the association between the initial treatment strategy and the rate of the outcomes of interest for the matched pairs and determined the significance of the differences using McNemar’s test. Rates of survival, freedom from rupture, and reintervention related to AAA were estimated with Kaplan–Meier life-table methods, and comparisons were made with log-rank analysis. To evaluate changes in event-rates over time we analyzed the perioperative and 2-year post-operative outcomes for each year from 2001-2008 using individual-level multivariable models that controlled for clinical and demographic characteristics that might have changed over time. We also performed separate survival analyses for the 2005-2008 period and the 2001-2004 period. Previously, we and others theorized about a differential “survival of the fittest” phenomena whereby the risk of death conditional on survival to that point (the hazard ratio) reverses direction (i.e., changes from favoring one procedure to favoring the other) based on the characteristics of those who survive for specific times periods. In this case, we hypothesized that OPEN survivors might have a decreased risk of death once they had survived the insult of the surgery. We therefore used an adaptation of the Cox model that allows the effect of endovascular repair (vs. open repair) to change over time using change-points based on empirical analyses.18 More details are available in the Statistical Appendix. To estimate the overall survival benefit of endovascular repair in the presence of time-varying treatment effects (or hazards), we compute the Restricted Mean Survival Time (RMST) – the total amount of time over a given follow-up period that a patient with given characteristics is expected to survive.19 The RMST (or its proportional equivalent obtained by dividing by the followup time) may be plotted against the follow-up time to reveal how the net (or aggregate) advantage of endovascular repair relative to open changes with the length of follow-up, thereby summarizing the overall risk of death for any follow-up time of interest. Together the interval-specific treatment effects and the difference in RMST due to endovascular repair allow changes in risk over time and their impact on survival through a given period of follow-up to be detected and quantified.

Results

We identified 128,598 patients, age 67 and above, who underwent elective AAA repair during the 2001–2008 period, including endovascular repair in 79,463 patients, and open repair in 49,135. Baseline characteristics and coexisting conditions, before and after propensity score matching are shown in Table 1. After matching there were 79,932 patients with 39,966 in each group with no substantial differences between the two cohorts. Perioperative outcomes Perioperative mortality for endovascular repair was 1.6% and 5.2% for open 51

3

Chapter 3 Table 1: Baseline Characteristics of Medicare Beneficiaries Undergoing Endovascular Repair (EVAR) or Open Repair of Abdominal Aortic Aneurysms in 2001-2008, before and after Propensity Score Matching Unmatched Cohort Variable

EVAR (n=79,463)

Open (n=49,135)

Matched Cohort P Value

EVAR (n=39,966)

Open (n=39,966)

P Value

Year 2001

5,423 (6.8)

9,748 (19.8)

<.001 5,150 (12.9)

5,166 (12.9)

0.866

2002

6,768 (8.5)

8,139 (16.6)

<.001 5,860 (14.7)

5,858 (14.7)

0.984

2003

7,456 (9.4)

7,213 (14.7)

<.001 5,800 (14.5)

5,839 (14.6)

0.696

2004

9.327 (11.7)

6,812 (13.9)

<.001 6,073 (15.2)

6,107 (15.3)

0.738

2005

11,493 (14.5) 5,821 (11.9)

<.001 5,584 (14.0)

5,663 (14.2)

0.422

2006

12,541 (15.8) 4,668 (9.5)

<.001 4,588 (11.5)

4,623 (11.6)

0.698

2007

13,049 (16.4) 2,984 (7.6)

<.001 3,825 (9.6)

3,731 (9.3)

0.256

2008

13,406 (16.9) 2,984 (6.1)

<.001 3,086 (7.7)

2,979 (7.5)

0.153

Male sex

65,446 (82.4) 36,303 (73.9) <.001 31,047 (77.7) 31,012 (77.6) 0.766

Race or ethnic group * Other

997 (1.3)

White

75,824 (95.4) 46,801 (95.3) 0.158 38,120 (95.4) 38,093 (95.3) 0.650

Black

2,212 (2.8)

1,421 (2.9)

0.255 1,119 (2.8)

1,138 (2.9)

0.685

Hispanic

430 (0.5)

281 (0.6)

0.47

224 (0.6)

0.887

76.8

75.2

<.001 75.7

75.5

<.001

67-69

9,465 (11.9)

7,777 (15.8)

<.001 5,851 (14.6)

5,968 (14.9)

0.244

70-74

20,692 (26.0) 15,712 (32.0) <.001 12,120 (30.3) 12,192 (30.5) 0.580

75-79

23,307 (29.3) 15,145 (30.8) <.001 12,468 (31.2) 12,297 (30.8) 0.191

80-84

17,440 (22.0) 8,128 (16.5)

<.001 7,284 (18.2)

7,217 (18.1)

0.539

85+

8,559 (10.8)

2,373 (4.8)

<.001 2,243 (5.6)

2,292 (5.7)

0.454

Urgent admission

2,974 (3.7)

3,451 (7.0)

<.001 2,071 (5.2)

1,916 (4.8)

0.012

Prior AAA diagnosis

59,587 (75.0) 31,152 (63.4) <.001 27,534 (68.9) 27, 724 (69.4) 0.146

Mean age (years)

632 (1.3)

0.623 500 (1.3)

227 (0.6)

511 (1.3)

0.728

Age category (years)

Coexisting conditions MI within 6 mo *

1,208 (1.5)

795 (1.6)

0.169 629 (1.6)

654 (1.6)

0.482

MI within 6-18 mo *

6,096 (7.7)

3,093 (6.3)

<.001 2,709 (6.8)

2,686 (6.7)

0.746

3,969 (8.1)

<.001 3,467 (8.7)

3,439 (8.6)

0.725

Valvular heart disease 8,240 (10.4)

52

Long-term Outcomes of A Repair Congestive heart failure

11,573 (14.6) 5,240 (10.7)

<.001 4,671 (11.7)

4,630 (11.6)

0.651

Peripheral vascular disease

15,187 (19.1) 9,887 (20.1)

<.001 7,897 (19.8)

7,767 (19.4)

0.247

Neurovascular disease

10,901 (13.7) 7,020 (14.3)

0.004 5,562 (13.9)

5,544 (13.9)

0.854

Hypertension

52,138 (65.6) 30,429 (61.9) <.001 25,257 (63.2) 25,137 (62.9) 0.379

Diabetes

15,287 (19.2) 7,215 (14.7)

Chronic obstructive pulmonary disease

22,980 (28.9) 13,407 (27.3) <.001 11,117 (27.8) 11,102 (27.8) 0.906

Renal failure

5,946 (7.5)

2,482 (5.1)

<.001 2,250 (5.6)

2,197 (5.5)

0.413

End-stage renal disease

508 (0.6)

148 (0.3)

<.001 165 (0.4)

146 (0.4)

0.280

History of cancer

16,119 (20.3) 7,590 (15.5)

<.001 6,710 (16.8)

6,762 (16.9)

0.623

Obesity

2,179 (2.7)

<.001 811 (2.0)

787 (2.0)

0.544

854 (1.7)

<.001 6,427 (16.1)

6,373 (16.0)

0.603

* MI = Myocardial Infarction. Race or ethnic group was self-reported

repair (relative risk (RR) 3.2; 95% confidence interval (CI) 3.0 to 3.5; P<.001). This mortality benefit was seen for all age groups and across the entire time period (Appendix Table S1). Endovascular repair patients also showed lower rates of medical and surgical perioperative complications (e.g. pneumonia 4% v. 13%, p<.001), were more likely to be discharged home (95% v. 83%, p<.001), and had a shorter length of stay (3.5 v. 9.8, p<.001). (Appendix Table S1) Long-term outcomes Long-term survival is shown in Figure 1. The early survival benefit after endovascular repair persisted for approximately 3 years, after which the estimated survival curves became indistinguishable. When comparing the results over 2005-2008 with those over 2001-2004, overall survival was better in the later period, but the findings were otherwise similar. (Figure 2) The survival hazards, however, were not proportional. Endovascular repair dominates strongly for 30 days (HR 0.32; CI 0.29 to 0.35; P<.001), continues dominating for the next 60 days (HR 0.64; CI 0.58 to 0.71; P<.001) after which endovascular repair has a slightly higher instantaneous risk (hazard) until year 4 (HR 1.17; CI 1.13 to 1.21; P<.001), after which the hazard is statistically significantly different at the 0.05 level but not meaningfully different in practical/ clinical terms (HR 1.05; CI 1.00 to 1.09; P=0.031). Because early survival confers a greater advantage in terms of the total amount of time patients are expected to survive, the restricted mean survival analyses reveal that at 4 years, patients in the endovascular cohort survive an average of 12.4 days longer (95% CI 9.0 to 53

3

Chapter 3 Table 2. Late Outcomes after Endovascular and Open Repair Year 1

Year 2

Variable

Endovascu- Open Endovascu- Open lar Repair Repair P Value lar Repair Repair P Value N=36, 835 N=35,627 N=31,659 N=31,161

Death

3,131 (7.8%)

4,339 (10.9%)

<.001

5,543 (14.1%)

6,257 (15.8%)

<.001

Rupture

304 (0.8%)

167 (0.5%)

<.001

487 (1.4%)

247 (0.7%)

<.001

Any aneurysm-related Intervention

1,497 (4.0%)

254 (0.7%)

<.001

2,405 (6.7%)

393 (1.1%)

<.001

Major reintervention

103 (0.3%)

100 (0.3%)

0.975

175 (0.5%)

124 (0.3%)

0.003

Minor reintervention

1,422 (3.8%)

161 (0.4%)

<.001

2,279 (6.4%)

283 (0.8%)

<.001

Embolizaton

631 (1.7%)

28 (0.1%) <.001

1,113 (3.2%)

56 (0.2%) <.001

AAA Hospitalization without reintervention

60 (0.2%)

20 (0.1%) <.001

103 (0.3%)

25 (0.1%) <.001

Laparotomy-related reintervention

522 (1.4%)

1,816 (5.1%)

<.001

903 (2.6%)

3,024 (8.8%)

<.001

Repair of an abdominal-wall hernia

198 (0.5%)

1,304 (3.7%)

<.001

352 (1.0%)

2,268 (6.6%)

<.001

Lysis of adhesions without bowel resection

55 (0.1%)

232 (0.6%)

<.001

101 (0.3%)

366 (1.1%)

<.001

Bowel resection

304 (0.8%)

371 (1.0%)

0.002

524 (1.5%)

606 (1.8%)

0.004

Laparotomy related hospitalization without intervention

1,026 (2.7%)

1,723 (4.7%)

<.001

1,758 (4.9%)

2,728 (7.8%)

<.001

Aneurysm or Laparotomy-relat- 1,989 ed reintervention (5.3%)

2,045 (5.7%)

0.017

3,238 (9.1%)

3,365 (9.8%)

<.001

Aneurysm or Laparotomy related hospitalization without intervention

1,100 (2.9%)

1,772 (4.9%)

<.001

1,884 (5.3%)

2,796 (8.0%)

<.001

Rupture, AAA- or laparotomy-realted reintervention or hospitalization without intervention

2,293 (6.1%)

2,213 (6.1%)

0.768

3,666 (10.3%)

3,591 (10.4%)

0.535

54

Long-term Outcomes of A Repair

Year 5

Year 8

Endovascu- Open lar Repair Repair N=14,894 N=15,011

Endovascu- Open P Value lar Repair Repair N=2,176 N=2,237

11,705 (34.4%)

11,842 (34.2%)

0.567

14,548 (54.9%)

14,681 (54.7%)

0.758

820 (3.0%)

326 (1.1%)

<.001

962 (5.4%)

353 (1.4%)

<.001

3,816 (13.3%)

651 (2.4%)

<.001

4,165 (18.8%)

754 (3.7%)

<.001

333 (1.3%)

168 (0.6%)

<.001

392 (2.3%)

186 (0.8%)

<.001

3,609 (12.6%)

509 (1.9%)

<.001

3,924 (17.5%)

597 (3.1%)

<.001

1,727 (6.0%)

125 (0.5%)

<.001

1,857 (8.0%)

161 (1.0%)

<.001

199 (0.7%)

45 (0.2%)

<.001

233 (1.2%)

55 (0.3%)

<.001

1,509 (5.3%)

4,182 (14.2%)

<.001

1,695 (8.2%)

4,427 (17.7%)

<.001

562 (2.0%)

2,980 (9.9%)

<.001

610 (2.7%)

3,070 (11.2%)

<.001

198 (0.8%)

586 (2.1%)

<.001

238 (1.4%)

654 (3.1%)

<.001

901 (3.2%)

1,048 (3.9%)

<.001

1,035 (5.2%)

1,199 (6.0%)

0.008

3,083 (11,1%)

4,303 (15.2%)

<.001

3,510 (17.3%)

4,805 (22.2%)

<.001

5,138 (17.8%)

4,710 (16.0%)

<.001

5,614 (25.1%)

5,034 (20.6%)

<.001

3,287 (11.7%)

4,371 (15.4%)

<.001

3,710 (17.9%)

4,846 (22,0%)

<.001

5,755 (19.8%)

5,007 (17.0%)

<.001

6,279 (27.8%)

5,355 (21.8%)

<.001

P Value

3

55

Chapter 3 Panel A

Panel B

Survival of Endovascular vs. Open Repair AAA, All Ages 99% Conf. Int.

1.0

1.0

0.9

0.8

Probability of Survival

Probability of Survival

0.9

0.7

0.6

EVAR Open

0.8

0.7

0.6

0.5

0.5

0.4 0

39,966

39,966

1

36,835

35,627

2

31,659

31,161

3

26,227

26,132

4

20,580

20,708

5

14,894

15,011

6

9,693

9,791

7

5,562

5,626

0

8

2,176

1

2

3

4

2,237

Survival Time (Years)

Survival Time (Years)

Figure 1. Survival of Patients Undergoing Endovascular Repair or Open Repair of Abdominal Aortic Aneurysms. Data are shown for all patients (Panel A) and for Time Periods 2001-2004 and 20052008 (Panel B).

Panel A

Panel B

1.0

1.0

0.9

0.9

0.8

0.8

0.7

0.7

0.6

0.6

EVAR

0.5

0.5 0

39,966

39,966

1

33,573

32,495

2

26,896

26,386

3

20,820

20,970

4

15,273

15,772

Time (Years)

5

10,370

10,869

6

6,353

6,783

7

3,455

3,768

8

1,286

1,427

0 22,886 / 17,080 22,973 / 16,993

1 19,131 / 14,065 18,712/ 13,485

2 12,213 / 9,844 12,128 / 9,622

3 6,982 / 6,071 7,120 / 6,209

4 2,705 / 2,717 2,922 / 2,894

Time (Years)

Figure 2. Rupture, Aneurysm or Laparotomy-related Reintervention of Patients Undergoing Endovascular Repair or Open Repair of Abdominal Aortic Aneurysms. Data are shown for all patients (Panel A) and for Time Periods 2001-2004 and 2005-2008 (Panel B).

56

10.80% 11.20%

Readmission in 30 days after discharge

0.90% 9.60%

Re-operation for bleeding

Readmission in 30 days after discharge 10

5.70%

Death (all ages)

Length of stay (mean)

2004

2005

2007

Endovascular Repair

2006

2008

3.46

9.30%

1.60%

Open Repair

3.66

2.10%

0.50%

1.60%

3.43

9.70%

0.90%

0.50%

1.20% 0.40% 0.50% 9.70% 3.33

0.40% 0.90% 9.40% 3.34

3.41

9.40%

0.60%

0.50%

1.60%

Open Repair

1.60%

1.50%

3.32

9.40%

0.30%

0.20%

1.40%

0.013

9.64

9.63

9.82

9.93

9.89

<.001

9.99

1.20%

5.10%

9.86

0.80%

1.10%

5.00%

0.236

5.10%

5.50%

10.40% 10.20% 10.30% 10.20% 10.60% 10.30% 10.50%

1.10%

5.20%

0.313

1.00%

5.80%

<.001

<.001

<.001

<.001

0.001

0.013

0.001

<.001

0.129

P Value (InterP Value action (linear between trend) EVAR and Open)

1.20%

1.00%

6.10%

N=5,166 N=5,858 N=5,839 N=6,107 N=5,663 N=4,623 N=3,731 N=2,979

3.52

2.20%

Conversion to open repair

0.70%

0.80%

Re-operation for bleeding

2.10%

2.20%

Length of stay (mean)

2003

Endovascular Repair

2002

N=5,150 N=5,860 N=5,800 N=6,073 N=5,584 N=4,588 N=3,825 N=3,086

2001

Death (all ages)

Perioperative outcomes

Table 3. Time Trend for Perioperative and Postoperative Outcomes

Long-term Outcomes of A Repair

57

3

58 2002

N=5,015 845 (14.4%) 86 (1.6%) 394 (7.3%) 33 (0.6%) 370 (6.9%) 191 (3.6%) 19 (0.4%) 130 (2.4%) 55 (1.0%) 16 (0.3%) 74 (1.4%) 275 (5.1%) 511 (9.4%)

2001

N=4,312 838 (16.3%) 59 (1.3%) 372 (7.9%) 22 (0.5%) 358 (7.6%) 194 (4.2%) 19 (0.4%) 118 (2.5%) 48 (1.0%) 17 (0.4%) 63 (1.4%) 253 (5.4%) 487 (10.4%)

Death

Rupture

Any aneurysm-related Intervention

Major reintervention

Minor reintervention

Embolizaton

AAA Hospitalization without reintervention

Laparotomy-related reintervention

Repair of an abdominal-wall hernia

Lysis of adhesions without bowel resection

Bowel resection

Laparotomy related hospitalization without intervention

Aneurysm or Laparotomy-related reintervention

Postoperative outcomes at two years

2005

N=5,287

N=4,840

Endovascular Repair

2004

N=3,957

2006

N=3,267

2007

490 (9.2%)

301 (5.7%)

86 (1.6%)

14 (0.3%)

55 (1.0%)

148 (2.8%)

13 (0.2%)

175 (3.3%)

337 (6.3%)

27 (0.5%)

354 (6.6%)

82 (1.5%)

470 (8.5%)

280 (5.1%)

82 (1.5%)

16 (0.3%)

54 (1.0%)

143 (2.6%)

15 (0.3%)

172 (3.1%)

326 (5.9%)

21 (0.4%)

342 (6.2%)

72 (1.3%)

473 (9.5%)

233 (4.6%)

81 (1.6%)

15 (0.3%)

50 (1.0%)

132 (2.7%)

13 (0.3%)

141 (2.9%)

336 (6.7%)

23 (0.5%)

351 (7.0%)

69 (1.4%)

330 (8.1%)

168 (4.1%)

50 (1.2%)

9 (0.2%)

42 (1.0%)

94 (2.3%)

11 (0.3%)

111 (2.8%)

228 (5.6%)

19 (0.5%)

243 (6.0%)

50 (1.2%)

311 (9.1%)

162 (4.8%)

60 (1.8%)

10 (0.3%)

34 (1.0%)

97 (2.9%)

7 (0.2%)

82 (2.5%)

205 (6.0%)

18 (0.5%)

221 (6.5%)

40 (1.2%)

822 (14.2%) 786 (13.0%) 744 (13.3%) 631 (13.7%) 558 (14.6%)

N=4,978

2003

Continuation Table 3. Time Trend for Perioperative and Postoperative Outcomes

2008

<.001

<.001

0.604

0.179

0.320

0.550

0.033

<.001

<.001

0.424

<.001

0.175

<.001

P Value (linear trend)

0.001

0.001

0.265

0.267

0.499

0.533

0.568

0.576

0.007

0.280

0.002

0.250

0.014

P Value (Interaction between EVAR and Open)

Chapter 3

0.666 0.923

278 (8.5%)

35 (0.7%) 58 (1.1%) 18 (0.3%) 41 (0.8%) 13 (0.3%) 5 (0.1%) 421 (8.1%) 332 (6.4%) 51 (1.0%) 67 (1.3%) 393 (7.5%) 473 (9.1%)

33 (0.7%) 42 (0.9%) 16 (0.3%) 29 (0.6%) 7 (0.2%) 4 (0.1%) 406 (8.9%) 309 (6.8%) 56 (1.2%) 71 (1.6%) 364 (7.9%) 444 (9.7%) 370 (8.1%)

476 (10.4%) 507 (9.8%) 567 (11.0%) 558 (10.4%) 520 (10.6%) 397 (10.0%) 352 (10.9%)

Rupture

Any aneurysm-related Intervention

Major reintervention

Minor reintervention

Embolizaton

AAA Hospitalization without reintervention

Laparotomy-related reintervention

Repair of an abdominal-wall hernia

Lysis of adhesions without bowel resection

Bowel resection

Laparotomy related hospitalization without intervention

Aneurysm or Laparotomy-related reintervention

Aneurysm or Laparotomy related hospitalization without intervention

Rupture, AAA- or laparotomy-realted reintervention or hospitalization without intervention

399 (7.6%)

940 (16.1%)

871 (16.8%)

38 (1.0%) 7 (0.2%) 2 (0.1%) 332 (8.4%) 246 (6.3%) 46 (1.2%)

58 (1.2%) 10 (0.2%) 4 (0.1%) 428 (8.8%) 319 (6.6%) 46 (1.0%) 93 (1.9%) 377 (7.6%) 493 (10.1%) 395 (8.0%)

6 (0.1%) 476 (8.9%)

64 (1.2%) 103 (1.9%) 475 (8.8%) 524 (9.8%) 489 (9.0%)

391 (7.5%)

535 (10.2%)

379 (7.3%)

100 (1.9%)

44 (0.9%)

365 (7.1%)

480 (9.3%)

3 (0.1%)

9 (0.2%)

36 (0.7%)

339 (64%)

4 (0.1%)

44 (0.8%)

12 (0.2%)

319 (7.9%)

373 (9.4%)

307 (7.6%)

72 (1.8%)

13 (0.3%)

20 (0.4%)

18 (0.3%)

333 (10.3%)

278 (8.5%)

60 (1.9%)

35 (1.1%)

236 (7.4%)

306 (9.5%)

1 (0.0%)

3 (0.1%)

18 (0.6%)

17 (0.5%)

33 (1.0%)

0.650

0.592

0.030

0.845

0.498

0.845

0.097

0.512

0.212

0.474

0.112

50 (1.2%)

55 (1.0%)

25 (0.8%)

74 (1.5%)

5.3 (1.0%)

35 (0.6%)

0.674

43 (0.8%)

27 (0.7%)

N=3,156

29 (0.6%)

N=3,902

0.241

N=4,821

948 (16.2%) 934 (15.3%) 842 (14.9%) 721 (15.6%) 576 (15.4%)

N=5,173

<.001

Death

N=4,891

343 (10.1%)

N=4,918

376 (9.2%)

N=4,295

Open Repair

524 (10.5%)

540 (9.7%)

<.001

561 (10.5%)

173 (5.1%)

590 (10.9%)

182 (4.4%)

546 (11.6%)

Rupture, AAA- or laparotomy-realted reintervention or hospitalization without intervention

252 (5.0%)

302 (5.4%)

319 (6.0%)

295 (5.5%)

271 (5.8%)

Aneurysm or Laparotomy related hospitalization without intervention

<.001

0.001

Long-term Outcomes of A Repair

59

3

Chapter 3 15.6, p<.0001), which remains significantly different through 7 years of follow-up (8.1 days, 95% CI 1.5 to 14.4, p=0.016). (Appendix Figure S2) These analyses suggest there is a substantial, immediate benefit of endovascular repair that endures for a considerable period of time despite the fact that patients who receive open repair and survive through 90-days of follow-up have a slightly lower risk of death over the next several years conditional on surviving to 90days. We studied the risk of a variety of secondary endpoints with death accounted for as a censoring event. Rupture was seen in 3.0% of endovascular repair patients vs. 1.1% of open repair patients at 5 years and in 5.4% vs. 1.4% at 8 years of follow up (P<.001). (Table 2) After 8 years, aneurysm-related reinterventions were more common after endovascular than open repair (18.8% vs. 3.7%, P<.001), including relatively infrequent major reinterventions (2.3% vs. 0.8%, P<.001) and minor reinterventions (17.5% vs. 3.1%, P<.001). Laparotomyrelated reinterventions were more common after open repair (17.7% vs. 8.2%, P<.001), the majority of which were abdominal wall hernia repairs. At 8 years, readmission for bowel obstruction without surgery was also more common in the open repair group (22.2% vs. 17.3%, P<.001). The overall rate of intervention for the combination of aneurysm-related or laparotomy related complications was 25.1% in the endovascular repair group versus 20.6% in the open repair group (P<.001). Note that not all laparotomy related readmissions and re-interventions we capture are unequivocally related to the index AAA repair (none should be referable to an index endovascular repair) and that the important finding is the difference in these rates between endovascular and open surgery patients. Time Trends: Peri-Operative and two-year Outcomes Endovascular perioperative mortality decreased from 2.2% to 1.4% (p=0.001), while open repair mortality decreased from 5.7% to 5.1% (p=0.013). (Table 3) Conversion to open repair dropped significantly from 2.2% to 0.3% (P<.001). Re-operation for bleeding decreased in the endovascular repair group from 0.8% to 0.2% (P<.001) and readmission in 30 days improved for endovascular repair from 10.8% to 9.4% (P=<.001), but these did not change for open repair. Two-year mortality after endovascular repair decreased from 16.3% (2001 procedures) to 14.6% (2007 procedures, p<.001), but did not change significantly after open repair (16.8% vs. 15.4%, P=0.24). (Table 3) There also was a decrease in total reinterventions at two years after endovascular repair over time (10.4% to 9.1%, P<.001), driven by a decrease in minor reinterventions, primarily coil embolization.

Discussion

In this large national study, we found that the early survival benefit from endovascular repair persisted for almost 3 years, after which survival was similar in both groups. Overall late complications when considering both reinterventions 60

Long-term Outcomes of A Repair and laparotomy-related procedures and admissions for bowel obstruction were similar (though slightly favoring open repair), but we now show for the first time that endovascular repair outcomes including perioperative mortality and reintervention rates are improving over time. Nonetheless, late rupture after endovascular repair, occurring in over 5% of patients by 8 years, is alarming and warrants continued follow up. Data from previous randomized trials demonstrate that the early advantage with endovascular repair was lost after 1 to 2 years (DREAM and EVAR1) or 3 years (OVER).6,7,9 The OVER trial also showed improved survival after endovascular repair among patients younger than 70 years of age, but not for those older than 70. In contrast, in our substantially larger analysis that was not restricted to ideal patients who would qualify for a clinical trial, we find an early survival advantage with endovascular repair for all ages and this benefit increases with age, and lasts for approximately 3 years. When measured in terms of expected time survived or “area under the curve�, the survival advantage of endovascular repair is estimated to persist through 7 years. Over up to 8 years of follow-up, we also find continued evidence of higher rates of aneurysm-related reinterventions after endovascular repair and offsetting higher rates of laparotomy-related reinterventions after open repair. Overall reinterventions were higher in the endovascular group. The EVAR-1 and DREAM trials showed higher reintervention rates after endovascular repair, but EVAR-1 did not account for laparotomy-related reinterventions, while DREAM included only hernia repairs. 6,9 The recent report of long-term results from the OVER trial did include hernia repair or readmission and reintervention for bowel obstruction and found that total reintervention rates over time were similar for both groups. Our analysis suggests that OVER may have been underpowered to detect a small difference in total reinterventions, yet our findings suggest that this difference is small and may not be clinically significant. 7 The overall rate of ruptures, reinterventions, or readmission is significantly higher with open repair at 1 and 2 years; however, the rate for endovascular repair became significantly higher after 5 years. Endovascular repair also has a significantly higher rate of late aneurysm rupture than open repair with 5.4% of survivors having rupture at 8 years. Randomized trials and other studies suggest that late rupture after endovascular repair may be increasing. 9 We confirm this finding and believe that more data are needed on the late risk of rupture after AAA repair, as AAAs are repaired to prevent precisely that occurrence. 20 We found improvements in perioperative mortality and conversion to open repair as well as two-year mortality and reinterventions after endovascular repair over time. The decline in reinterventions seems to be driven by a decrease in minor reinterventions, primarily coil embolization, which likely represents a more conservative attitude towards type 2 (side branch) endoleak management. The decline in perioperative mortality likely represents increased familiarity with the 61

3

Chapter 3 procedure and improvements in endografts over time. It is unlikely, however, that this improvement is driven by improved patient selection as the vast majority of patients are now being treated using endovascular repair and mortality rates for open repair improved over this time period as well. Our analyses using Medicare data are subject to several limitations including that the data are observational, subject to potential coding error, and lack anatomic detail (e.g. aneurysm diameter, calcification, iliac involvement, infrarenal neck anatomy etc.) and some clinical details (e.g. smoking, anemia, infection, medications etc.) which may be important determinants of patient selection and outcomes. These limitations are somewhat offset by several strengths including the large study size, the incorporation of physician CPT codes, determination of coexisting conditions from prior encounters, and the use of propensity-score matching. 11 After propensity-score matching, there were no clinically significant differences between the two cohorts in known risk factors for adverse events after AAA repair. Nonetheless, although our list of confounders was extensive, propensity analyses cannot account for selection bias related to unmeasured characteristics. Our analysis confirms lower perioperative mortality and complication rates with endovascular repair compared to open repair. The early survival benefit in terms of probability of survival persists for almost three years, after which survival is similar, while the total amount of time patients are expected to survive (restricted mean survival) remains higher for the endovascular repair group through 7 years of follow-up. Higher rates of reinterventions related to AAA were found in the endovascular repair group, which were partially balanced by a higher rate of laparotomy-related complications in the open repair group. Perioperative mortality and conversion to open repair as well as two-year mortality and reinterventions are improving over time with endovascular repair but late rupture after endovascular repair is a concern and warrants further follow up.

62

Long-term Outcomes of A Repair

References 1. Schermerhorn ML, Bensley RP, Giles KA, et al. Changes in abdominal aortic aneurysm rupture and short-term mortality, 1995-2008: a retrospective observational study. Annals of surgery 2012;256:651-8. 2. Dua A, Kuy S, Lee CJ, Upchurch GR, Jr., Desai SS. Epidemiology of aortic aneurysm repair in the United States from 2000 to 2010. J Vasc Surg 2014. 3. Prinssen M, Verhoeven EL, Buth J, et al. A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med 2004;351:1607-18. 4. participants Et. Endovascular aneurysm repair versus open repair in patients with abdominal aortic aneurysm (EVAR trial 1): randomised controlled trial. Lancet 2005;365:2179-86. 5. Lederle FA, Freischlag JA, Kyriakides TC, et al. Outcomes following endovascular vs open repair of abdominal aortic aneurysm: a randomized trial. JAMA : the journal of the American Medical Association 2009;302:153542. 6. De Bruin JL, Baas AF, Buth J, et al. Long-term outcome of open or endovascular repair of abdominal aortic aneurys. N Engl J Med 2010;362:1881-9. 7. Lederle FA, Freischlag JA, Kyriakides TC, et al. Long-term comparison of endovascular and open repair of abdominal aortic aneurysm. N Engl J Med 2012;367:1988-97. 8. Jackson RS, Chang DC, Freischlag JA. Comparison of long-term survival after open vs endovascular repair of intact abdominal aortic aneurysm among Medicare beneficiaries. JAMA : the journal of the American Medical Association 2012;307:1621-8. 9. Greenhalgh RM, Brown LC, Powell JT, Thompson SG, Epstein D, Sculpher MJ. Endovascular versus open repair of abdominal aortic aneurysm. N Engl J Med 2010;362:1863-71. 10. Dominique B. Buck JD, Jack Cronenwett et al. Substantial Regional Variation Exists in Patient Selection and Treatment of Abdominal Aortic Aneurysms in the United States. Abstract accepted for presentation at New England Society for Vascular Surgery annual meeting, Boston MA 2014. 11. Schermerhorn ML, O’Malley AJ, Jhaveri A, Cotterill P, Pomposelli F, Landon BE. Endovascular vs. open repair of abdominal aortic aneurysms in the Medicare population. N Engl J Med 2008;358:464-74. 12. Zhang JX, Iwashyna TJ, Christakis NA. The performance of different lookback periods and sources of information for Charlson comorbidity adjustment in Medicare claims. Medical care 1999;37:1128-39. 13. Bensley RP, Schermerhorn ML, Hurks R, et al. Risk of late-onset adhesions and incisional hernia repairs after surgery. Journal of the American College of Surgeons 2013;216:1159-67, 67 e1-12. 14. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Medical care 1998;36:8-27. 15. Baldwin LM, Klabunde CN, Green P, Barlow W, Wright G. In search of the 63

3

Chapter 3 perfect comorbidity measure for use with administrative claims data: does it exist? Medical care 2006;44:745-53. 16. Gu XS RP. Comparison of Multivariate Matching Methods: Structures, Distances, and Algorithms. Journal of Computational and Graphical Statistics 1993;2. 17. Cochran WG RD. Controlling bias in observational studies: a review. . SankhyÄ : The Indian Journal of Statistics, Series A 1973;35. 18. T.M.Therneau PMG. Modeling Survival Data; extending the Cox Model. New York: Springer; 2000. 19. Andersen PK, Hansen MG, Klein JP. Regression analysis of restricted mean survival time based on pseudo-observations. Lifetime data analysis 2004;10:335-50. 20. Schermerhorn ML, Giles KA, Sachs T, et al. Defining perioperative mortality after open and endovascular aortic aneurysm repair in the US Medicare population. Journal of the American College of Surgeons 2011;212:349-55.

64

Long-term Outcomes of A Repair

Appendix Methodological Appendix HMO enrollment Medicare beneficiaries enrolled in an HMO subsequent to their surgery made up 3.2% in 2001, 3.5% in 2002, 3.1% in 2003, 3.2% in 2004, 2.4% in 2005, 1.2% in 2006, 0.6% in 2007 and 0.2% in 2008. Ultimately, 1,911 total patients, including 975 from the endovascular group and 936 from the open repair group enrolled in Medicare Advantage at some time during the follow up period. Endovascular treatment was equally distributed by year. Statistical Appendix Modeling Long Term Outcomes and Cumulative Benefits As noted in the text, we used an adaptation of the Cox Proportional Hazards method to model the long term outcomes of endovascular versus open repair. Previously, we and others have theorized about the differential survival of the fittest phenomena whereby the risk of death conditional on survival to that point (the hazard ratio) reverses direction (i.e., changes from favoring one procedure to favoring the other procedure). In the case of endovascular and open repair the argument is that following an initial period of follow-up the surviving patients the open arm will on average be healthier (in unmeasured ways) in order to have endured the more invasive procedure and so have a better prognosis going forward than the surviving patients in the less invasive endovascular repair arm, resulting in an increase in the hazard-ratio of endovascular to open repair from less than 1 to greater than 1. Under such reasoning, the proportional hazards assumption of the basic Cox survival analysis method is violated. However, we may extend the Cox regression model to allow the effect of endovascular repair (vs. open repair) to change across time.17 To determine a suitable form of time dependence, we initially allowed separate effects of endovascular repair for the perioperative period, the next 60-days, the remainder of the first year, and over each year of follow-up thereafter. We found extensive heterogeneity in the effect of endovascular repair in the perioperative period to that over the next 60days and in-turn to the remainder of follow-up. However, the hazard ratio varied minimally from 90-days post procedure to 4-years post-procedure, and then varied minimally from 4 through 8 years post-procedure. Therefore, we settled on an extended Cox regression model that allowed the effect of endovascular repair to vary and thus accommodate non-proportionality of hazards at 30-days, 90-days, and 4-years post-procedure. The model is estimated by setting up our analysis data set in the counting process form (essentially defining multiple timeintervals for each patient based on the above break-points) of survival analysis and including an endovascular repair by time-period interaction effect in the specification of the model. We used the R statistical package to estimate the model. Having estimated the extended Cox regression model, one is faced with a 65

3

Chapter 3 dilemma on how best to report the results from the analysis. While the time-period specific estimates of endovascular repair enable the change in the hazard-ratio of death to be compared across time and thus the above differential-survival theory to be tested, one cannot simply add the estimates together to obtain an overall measure of survival benefit (or not) of endovascular repair. To overcome this problem, we compute the Restricted Mean Survival Time (RMST) – the total amount of time over a specified period of follow-up that a patient with given characteristics is expected to survive.18 A graphical interpretation of the RMST is that it equals the area beneath the survival curve. Early survival confers a greater advantage in terms of RMST than late survival (the reasoning is the same as that for why an early investment yields more Panel A

Panel B

Survival of Endovascular vs. Open Repair AAA, Ages 67 − 74

Survival of Endovascular vs. Open Repair AAA, Ages 75 − 84 99% Conf. Int.

1.0

1.0

0.8

0.8

Probability of Survival

Probability of Survival

99% Conf. Int.

0.6

0.4

EVAR Open

0.2

0.6

0.4

EVAR Open

0.2 17,971 18,160

16,878 16,854

14,728 14,859

12,383 12,563

9,831 10,073

7,345 7,428

4,862 4,867

2,829 2,844

1,101 1,170

19,752 19,514

0.0

18,065 17,012

15,429 14,838

12,680 12,402

9,898 9,753

6,988 6,984

4,493 4,550

2,542 2,580

992 992

0.0 0

1

2

3

4

5

6

7

8

9

0

1

2

3

Survival Time (Years)

4

5

6

7

8

9

Survival Time (Years)

Panel C Survival of Endovascular vs. Open Repair AAA, Ages 85 plus 99% Conf. Int.

1.0

Probability of Survival

0.8

0.6

0.4 EVAR Open

0.2 2,243 2,292

1,893 1,765

1,502 1,466

1,166 1,170

851 883

562 603

343 378

192 204

64 84

4

5

6

7

8

0.0 0

1

2

3

9

Survival Time (Years)

Figure S1. Survival of Patients Undergoing Endovascular Repair or Open Repair of Abdominal Aortic Aneurysms, According to Age. Data are shown for those 67 to 74 years of age (Panel A), those 75 to 84 years of age (Panel B), and those 85 years of age or older (Panel C)

66

Long-term Outcomes of A Repair

16 14 4

6

8

10

12

3

2

Additional Area Under the Curve for EVAR (days)

18

20

interest than a later investment because it has a longer period of time to grow) and so when assessing the overall effect of endovascular repair (vs. open repair) the comparison is incomplete if only the likelihood of survival for a sub-interval of the follow-up time-period of interest is presented. The RMST on the other hand aggregates the probability a patient with given characteristics survives each interval of time into an overall measure of expected survival. The RMST (or its proportional equivalent obtained by dividing the RMST by the follow-up time) may be plotted against follow-up time to reveal how the aggregate advantage of endovascular repair relative to open repair changes with the length of follow-up, thereby summarizing and presenting the overall risk of death across all follow-up times of interest.

0

5.3 * (p<.0001)

0

1

9.5 (p<.0001)

2

12.0 (p<.0001)

3

12.4 (p<.0001)

11.6 (p<.0001)

10.1 (p=0.0004)

4

5

6

8.2 5.6 (p=0.0159) (p=0.2179)

7

8

Years after Abdominal Aortic Aneurysm Repair

Figure S2. Survival Benefit for Endovascular Repair over Time (Additional Area under the Curve) * net days of survival advantage of endovascular compared to open repair

The above described behavior of the effect of endovascular repair is seen in the RMSTs at the end of each interval; the difference in the RMST increases from 0.2 days at 30 days of follow-up to 12.4 days at time 4 years and then reduces to 5.6 days by 8-years. In summary, the analysis presented here suggests there is a substantial, immediate benefit of endovascular repair that endures for a considerable period of time despite the fact that patients who receive open and survive through 90-days of follow-up have a slightly more optimistic risk-profile from that time forward. 67

Chapter 3 Table S1. Perioperative Outcomes after Endovascular Repair or Open Repair Perioperative Outcome

Endovascular Repair

Open Repair P Value

Relative Risk (95% CI)

Unmatched Cohort

N = 79,463

N = 49,135

Death (all ages)

1,270 (1.6)

2,561 (5.2)

Matched Cohort

N=39,966

N=39,966

All ages

651 (1.6)

2,094 (5.2)

<.001

3.22 (2.95-3.51)

67-69 yr

51 (0.9)

168 (2.8)

<.001

3.23 (2.37-4.41)

<.001

3.26 (3.05-3.48)

Death (% of patients)

70-74 yr

138 (1.1)

424 (3.5)

<.001

3.05 (2.52-3.70)

75-79 yr

191 (1.5)

648 (5.3)

<.001

3.44 (2.93-4.03)

80-84 yr

187 (2.6)

564 (7.8)

<.001

3.04 (2.59-3.58)

≼85 yr

84 (3.7)

290 (12.7)

<.001

3.38 (2.67-4.28)

Medical complications (% of patients) Myocardial Infarction

1,013 (2.5)

2,064 (5.2)

<.001

2.04 (1.89-2.19)

Pneumonia

1,522 (3.8)

5,139 (12.9)

<.001

3.38 (3.19-3.57)

Acute Renal Failure

1,726 (4.3)

4,531 (11.3)

<.001

2.63 (2.49-2.77)

Hemodialysis

174 (0.4)

244 (0.6)

<.001

1.40 (1.16-1.70)

Deep-vein Thrombosis

362 (0.9)

718 (1.8)

<.001

1.98 (1.75-2.25)

232 (0.6)

454 (1.1)

<.001

1.96 (1.67-2.29)

Surgical complications (% of patients) Re-operation for Bleeding Tracheostomy

78 (0.2)

636 (1.6)

<.001

8.15 (6.45-10.31)

Embolectomy

446 (1.1)

659 (1.7)

<.001

1.48 (1.31-1.66)

Conversion to Open Repair

445 (1.1)

Mesenteric Ischemia

232 (0.6)

853 (2.1)

<.001

3.68 (3.18-4.25)

Major Amputation

15 (0.0)

43 (0.1)

<.001

2.87 (1.59-5.16)

Lysis of adhesions without resection

17 (0.0 )

460 (1.2)

<.001

27.06 (16.6843.90)

Bowel Resection (Small)

43 (0.1)

136 (0.3)

<.001

3.16 (2.25-4.46)

Bowel Resection (Large)

114 (0.3)

349 (0.9)

<.001

3.06 (2.48-3.78)

Complications related to laparotomy

68

Long-term Outcomes of A Repair Ileus or bowel obstruction w/o resection Mean length of hospital stay (days)

1,071 (2.7)

6,422 (16.1)

<.001

3.5±5.3

9.8±8.9

<.001

2

7

<.001

37,517 (95.0)

31,692 (83.2)

<.001

6.00 (5.63-6.39)

Median length of hospital stay (days) Discharged home (% of survivors) All ages

0.88 (0.87-0.88)

67-69 yr

5,703 (98.0)

5,418 (93.1)

<.001

0.95 (0.94-0.96)

70-74 yr

11,589 (97.2)

10,586 (89.7)

<.001

0.92 (0.92-0.93)

75-79 yr

11,733 (95.2)

9,637 (82.3)

<.001

0.87 (0.86-0.87)

80-84 yr

6,539 (91.4)

4,820 (71.8)

<.001

0.79 (0.77-0.80)

≥85 yr

1,853 (85.0)

1,231 (60.3)

<.001

0.71 (0.68-0.74)

4,278 (10.7)

4,395 (11.0)

0.183

1.03 (0.99-1.07)

Readmission in 30 days after Discharge

3

* Plus-minus values are means ± SD.

69

Chapter 3 Table S2. Baseline Characteristics of Medicare Beneficiaries Undergoing Endovascular Repair of Abdominal Aortic Aneurysms in 2001-2008, who were Propensity Score Matched vs. Unmatched Matched Cohort

Unmatched Cohort

N=39,966

N=39,497

P Value

2001

5,150 (12.9%)

273 (0.7%)

<.001

2002

5,860 (14.7%)

908 (2.3%)

<.001

2003

5,800 (14.5)

1,656 (4.2%)

<.001

2004

6,073 (15.2%)

3,254 (8.2%)

<.001

2005

5,584(14.0%)

5,909 (15.0%)

<.001

2006

4,588 (11.5%)

7,953 (20.1%)

<.001

2007

3,825 (9.6%)

9,224 (23.4%)

<.001

2008

3,086 (7.7%)

10,320 (26.1%)

<.001

Male sex

31,047 (77.7%)

34,399 (87.1%)

<.001

Variable Year

Race or ethnic group * Other

500 (1.3%)

497 (1.3%)

0.937

White

38,120 (95.4%)

37,704 (95.5%)

0.861

Black

1,119 (2.8%)

1,093 (2.8%)

0.968

227 (0.6%)

203 (0.5%)

0.482

75.7

78.0

<.001

67-69

5,851 (14.6%)

3,614 (9.2%)

<.001

70-74

12,120 (30.3%)

8,572 (21.7%)

<.001

75-79

12,468 (31.2%)

10,839 (27.4%)

<.001

80-84

7,284 (18.2%)

10,156 (25.7%)

<.001

85+

2,243 (5.6%)

6,316 (16.0%)

<.001

Hispanic Mean age (years) Age category (years)

Urgent admission Prior AAA diagnosis

2,071 (5.2%)

903 (2.3%)

<.001

27,534 (68.9%)

32,053 (81.2%)

<.001

Coexisting conditions MI within 6 mo *

70

629 (1.6%)

579 (1.5%)

0.214

MI within 6-18 mo *

2,709 (6.8%)

3,387 (8.6%)

<.001

Valvular heart disease

3,467 (8.7%)

4,773 (12.2%)

<.001

Congestive heart failure

4,671 (11.7%)

6,902 (17.5%)

<.001

Peripheral vascular disease

7,897 (19.8%)

7,290 (18.5%)

<.001

Long-term Outcomes of A Repair Neurovascular disease

5,562 (13.9%)

5,339 (13.5%)

0.102

Hypertension

25,257 (63.2%)

26,881 (68.1%)

<.001

Diabetes

6,427 (16.1%)

8,860 (22.4%)

<.001

Chronic Pulmonary Disease

11,117 (27.8%)

11,863 (30.0%)

<.001

2,250 (5.6%)

3,696 (9.4%)

<.001

End-stage renal disease

Renal failure

165 (0.4%)

343 (0.9%)

<.001

History of cancer

6,710 (16.8%)

9,409 (23.8%)

<.001

811 (2.0%)

1,368 (3.5%)

<.001

Obesity

* MI = Myocardial Infarction. Race or ethnic group was self-reported

3

71

Chapter 3 Table S3. Perioperative Outcomes after Endovascular Repair or Open Repair, who were Propensity Score Matched vs. Unmatched Matched N = 39,966

Unmatched N = 39,497

P Value

Relative Risk (95% CI)

651 (1.6%)

619 (1.6%)

0.488

1.04 (0.93-1.16)

67-69 yr

51 (0.9%)

29 (0.8%)

0.721

1.09 (0.69-1.71)

70-74 yr

138 (1.1%)

73 (0.9%)

0.043

1.34 (1.01-1.77)

75-79 yr

191 (1.5%)

148 (1.4%)

0.290

1.12 (0.91-1.39)

80-84 yr

187 (2.6%)

177 (1.7%)

<.001

1.47 (1.20-1.81)

≼85 yr

84 (3.7%)

192 (3.0%)

0.104

1.23 (0.96-1.58)

Myocardial Infarction

1,013 (2.5%)

1,007 (2.6%)

0.894

0.99 (0.91-1.08)

Pneumonia

1,522 (3.8%)

1,570 (4.0%)

0.224

0.96 (0.89-1.03)

Acute Renal Failure

Perioperative Outcome Death (% of patients) All ages

Medical complications (% of patients)

1,726 (4.3%)

2,049 (5.2%)

<.001

0.83 (0.78-0.89)

Hemodialysis

174 (0.4%)

243 (0.6%)

0.001

0.71 (0.58-0.86)

Deep-vein Thrombosis

362 (0.9%)

484 (1.2%)

<.001

0.74 (0.65-0.85)

Re-operation for Bleeding

232 (0.6%)

172 (0.4%)

0.004

1.33 (1.09-1.62)

Tracheostomy

78 (0.2%)

66 (0.2%)

0.352

1.17 (0.84-1.62)

Surgical complications (% of patients)

Embolectomy

446 (1.1%)

365 (0.9%)

0.007

1.21 (1.05-1.39)

Conversion to Open Repair

445 (1/1%)

168 (0.4%)

<.001

2.62 (2.19-3.12)

Mesenteric Ischemia

232(0.6%)

243 (0.6%)

0.525

0.94 (0.79-1.13)

Major Amputation

15 (0.0%)

4 (0.0%)

0.013

3.71 (1.23-11.17)

Lysis of adhesions without resection

17 (0.0%)

20 (0.1%)

0.597

0.84 (0.44-1.60)

Bowel Resection (Small)

43(0.1%)

24 (0.1%)

0.023

1.77 (1.07-2.92)

Bowel Resection (Large)

114 (0.3%)

110 (0.3%)

0.858

1.02 (0.79-1.33)

1,071 (2.7%)

977 (2.5%)

0.067

1.08 (0.99-1.18)

Complications related to laparotomy

Ileus or bowel obstruction w/o resection

72

Long-term Outcomes of A Repair

Discharged home (% of survivors)

N=39,506

N=39,104

37,517 (95.0%) 36,707 (93.9%)

<.001

1.01 (1.01-1.02)

67-69 yr

5,703 (98.0%)

3,536 (98.3%)

0.349

1.00 (0.99-1.00)

70-74 yr

11,689 (97.2%)

8,299 (97.3%)

0.866

1.00 (0.99-1.00)

75-79 yr

11,733 (95.2%) 10,286 (95.8%)

0.020

0.99 (0.99-1.00)

80-84 yr

6,539 (91.4%)

9,240 (92.1%)

0.128

0.99 (0.98-1.00)

≥85 yr

1,853 (85.0%)

5,346 (86.2%)

0.140

0.99 (0.97-1.01)

4,278 (10.7%)

4,423 (11.2%)

0.026

0.96 (0.92-0.99)

All ages

Readmission in 30 days after Discharge

* Plus-minus values are means ± SD.

3

73

Chapter 3 Table S4. Late Outcomes after Endovascular and Open Repair, who were Propensity Score Matched vs. Unmatched Year 1 Variable Death

Matched N=36,835

Year 2

Unmatched N=35,967

P Value

Matched N=31,659

Unmatched N=24,256

P Value

3,131

7.8%

3,521

8.9%

<.001

304

0.8%

308

0.8%

0.636

487

1.4%

459

1.4%

0.840

1,497

4.0%

1,415

3.9%

0.385

2,405

6.7%

2,149

6.5%

0.195

Major reintervention

103

0.3%

84

0.2%

0.232

175

0.5%

119

0.4%

0.004

Minor reintervention

1,422

3.8%

1,354

3.7%

0.516

2,279

6.4%

2,067

6.3%

0.481

Embolizaton

631

1.7%

602

1.7%

0.719

1,113

3.2%

1,036

3.2%

0.634

AAA Hospitalization without reintervention

60

0.2%

49

0.1%

0.351

103

0.3%

69

0.2%

0.023

Laparotomy-related reintervention

522

1.4%

568

1.5%

0.070

903

2.6%

878

2.7%

0.326

Repair of an abdominal-wall hernia

198

0.5%

208

0.6%

0.452

351

1.0%

311

0.9%

0.450

Lysis of adhesions without bowel resection

55

0.1%

86

0.2%

0.006

101

0.3%

126

0.4%

0.036

Bowel resection

304

0.8%

314

0.9%

0.471

524

1.5%

513

0.6%

0.307

Laparotomy related hospitalization without intervention

1,026

2.7%

1,045

2.8%

0.527

1,758

4.9%

1,673

5.1%

0.374

Aneurysm or Laparotomy-related reintervention

1,989

5.3%

1,948

5.3%

0.912

3,238

9.1%

2,964

9.0%

0.557

Aneurysm or Laparotomy related hospitalization without intervention

2,293

6.1%

2,226

6.1%

0.837

3,666 10.3% 3,336 10.0% 0.331

Rupture Any aneurysm-related Intervention

Rupture, AAA- or laparotomy-realted reintervention 2,243 or hospitalization without intervention

74

6.0%

2,180 5.9%

0.872

5,543 14.1% 6,091 16.4% <.001

3,593

10.1% 3,284

9.9%

0.464

Long-term Outcomes of A Repair

Year 5 Matched N=14,899

Year 8

Unmatched N=3,211

P Value

Matched N=2,178

Unmatched N=34

P Value

11,705 34.4% 10,150 41.9% <.001 14,548 54.9% 10,708 71.1% <.001 820

3.0%

674

3.6%

0.003

3,816 13.3% 2,868 13.0% 0.455 333

1.3%

196

1.2%

0.473

962

5.4%

692

5.8%

0.612

4,165 18.8% 2,898 16.8% 0.186 392

2.3%

203

1.9%

0.274

3,609 12.6% 2,747 12.4% 0.649

3,924 17.5% 2,773 15.8% 0.228

1,727

6.0%

1,442

6.8%

0.002

1,857

8.0%

1,452

7.8%

0.794

199

0.7%

113

0.6%

0.166

233

1.2%

115

0.8%

0.051

1,509

5.3%

1,180

5.6%

0.258

1,695

8.2%

1,207

8.4%

0.748

562

2.0%

405

1.8%

0.313

610

2.7%

416

2.6%

0.846

198

0.8%

165

0.8%

0.538

238

1.4%

169

1.2%

0.326

901

3.2%

720

3.6%

0.051

1,035

5.2%

740

5.9%

0.378

3,083 11.1%

2,408 12.2% 0.003

3,510 17.3%

2,466 17.9% 0.655

5,138 17.8% 3,932 17.8% 0.943

5,614 25.1% 3,985 22.9% 0.044

5,755 19.8% 4,402 19.8% 0.855

6,279 27.8% 4,462 26.2% 0.239

5,643

19.5% 4,335

19.5% 0.990

6,164

3

27.4% 4,396

26.0% 0.260

75

CHAPTER 4

Risk factors and consequences of persistent type II endoleaks. Dominique B. Buck, Ruby C. Lo, Jeremy Herrmann, Allen D. Hamdan, Mark Wyers, Virendra I. Patel, Mark Fillinger, Marc L. Schermerhorn. Revisions J Vasc Surg.

Chapter 4

Abstract Objectives Type 2 endoleaks are common after endovascular aneurysm repair (EVAR) but their clinical significance remains controversial. We determined risk factors for type 2 endoleak and associations with adverse outcomes. Methods We performed a retrospective study using the Vascular Study Group of New England abdominal aortic aneurysm (AAA) database, where we identified all EVARs. Patients were subdivided into two groups: 1) those with no endoleak or transient type 2 endoleak and 2) persistent type 2 endoleak or new type 2 endoleak (no endoleak at completion of case). Patients with other endoleak types and follow-up shorter than 6 months were excluded. Multivariable analysis was used to evaluate predictors of persistent or new type 2 endoleaks. Kaplan-Meier and Cox regression analysis were used to evaluate predictors of reintervention and survival. Results 2367 EVAR patients had information on endoleaks; 1977 (84%) were in group 1, of which 79% had no endoleak at all and 21% were transient endoleaks that resolved at follow-up. The other 390 (16%) were in group 2, of which 31% had a persistent leak and 69% had a new leak at follow-up that was not seen at the time of surgery. Group 2 was older (mean age 75 vs. 73 years, P<.001), and less likely to have COPD (24% vs. 34%, P<.001) or elevated creatinine levels (2.6% vs. 5.3%, P=0.027). Coil embolization of one or both hypogastric arteries was associated with a higher rate of persistent type 2 endoleaks (12 vs. 8%, P=0.024), as was distal graft extension (12% vs. 8%, P=0.008). In multivariable analysis, COPD (OR 0.7, 95% CI 0.5-0.9, P=0.017) was protective against persistent type 2 endoleak, while hypogastric artery coil embolization (OR 1.5, 95% CI 1.0-2.2, P=0.044), distal graft extension (OR 1.6, 95% CI 1.1-2.3, P=0.025) and age ≼ 80 (OR 2.7, 95% CI 1.4-5.3, P=0.004) were predictive. Graft type was also associated with endoleak development. Persistent type 2 endoleaks were predictive of post-discharge reintervention (OR 15.3, 95% CI 9.7-24.3, P<.001), however not predictive of long-term survival (OR 1.1, 95%CI 0.9-1.6, P=0.477). Conclusion Persistent type 2 endoleak is associated with hypogastric artery coil embolization, distal graft extension, older age, the absence of COPD, and graft type, but not with aneurysm size. Persistent type 2 endoleaks are associated with an increased risk of reinterventions, but not rupture or survival. This reinforces the need for continued surveillance of patients with persistent type 2 endoleaks and the importance of follow-up to detect new type 2 endoleaks over time.

78

Type II endoleaks

Introduction

Endovascular aneurysm repair (EVAR) of abdominal aortic aneurysms (AAA) is associated with significantly higher short-term survival rates when compared to open AAA repair, but equivalent long-term survival rates.1-5 Though EVAR imparts an early survival benefit, this benefit is not sustained and EVAR is associated with more aneurysm-related reinterventions than open repair.6 Endoleaks are the most common complication of EVAR and are a frequent indication for reintervention. While Type 1 and Type 3 endoleaks necessitate reintervention and repair, the clinical significance of type 2 endoleaks remains controversial. Most type 2 endoleaks resolve spontaneously and the one-year post-operative prevalence ranges from 1-10%.7-9 However, there is evidence that persistent type 2 endoleaks are associated with an increased risk of adverse outcomes (sac enlargement, aneurysm rupture, need for reintervention, conversion to open repair).9, 10 Results from the EUROSTAR registry suggest that type 2 endoleaks are associated with aneurysmal growth and reintervention, but not with rupture or conversion to open repair.10 A study of 832 EVAR patients found persistent type 2 endoleaks accounting for 38% of secondary reinterventions.11 Other studies have shown no such relationships between type 2 endoleaks and adverse outcomes,12, 13 but these studies may have been insufficiently powered. Current recommendations suggest intervention for type 2 endoleaks in the presence of aneurysmal growth and/or endoleak persistence; however the exact criteria for reintervention remain undefined.7, 8, 14 Some physicians argue for aggressive management, such as prophylactic embolization of patent vessels communicating with the endoleak, while others endorse conservative monitoring approaches.9, 15, 16 Nevertheless, prophylactic embolization has mixed success in preventing type 2 endoleaks.17, 18 Our goal was to study the incidence of type 2 endoleaks, both transient and persistent, and to assess adverse outcomes, including post-operative complications and reinterventions. In addition, we wished to identify independent predictors of persistent and delayed type 2 endoleaks, and to determine whether persistent type 2 endoleaks were associated with late reinterventions and survival.

Methods

We undertook a retrospective analysis of the Vascular Surgery Group of New England (VSGNE) database, which is a multidisciplinary quality improvement collaborative between both academic and non-academic hospitals. More detailed information can be found at www.vascularweb.org/regionalgroups/ vsgne/pages/home.aspx. We used data from January 2003 through December 2014 and identified all patients undergoing EVAR. Patients were subdivided into two categories: 1) those who either had no endoleak at any time or those with a transient type 2 endoleak (defined as an endoleak at completion of the case 79

4

Chapter 4

Figure 1. Consort diagram of inclusion- and exclusion criteria for this study.

that was not seen at follow-up) and 2) persistent type 2 endoleaks (defined as an endoleak at completion of case that was also detected at follow-up) or those who developed a new type 2 endoleak (defined as no endoleak at completion of the case that developed an endoleak at any point postoperatively). Patients who underwent reinterventions for a type 2 endoleak were grouped into the latter category. Patients who developed Type 1 or Type 3 endoleaks at any time, or those without follow-up endoleak information were excluded from the analysis. Patients with a follow-up time shorter than 6 months were also excluded from our analysis. (Figure 1) We analyzed potential predictors of persistent type 2 endoleak including baseline demographics and comorbidities and operative factors including hypogastric artery coverage, both intentional (planned prior to procedure to treat distal aneurysm extent) and unintentional (inadvertent extension of graft not necessary to treat distal aneurysm extent). In addition, both preoperative and intraoperative hypogastric artery coiling were analyzed. Presence of an iliac aneurysm was defined by an iliac diameter > 1.5cm, using the maximum diameter of either the common or internal iliac artery. We also evaluated the influence of graft type on persistent type 2 endoleak, but limited this to only those grafts with at least 100 implants. 80

Type II endoleaks Patient demographics, comorbidities, perioperative details, and outcomes are reported as proportions of the total. We compared categorical variables between outcome subgroups using the χ2 and Fisher’s exact tests. The means of continuous variables were compared using student’s t-test. Multivariable logistic regression using purposeful selection19 was used to determine independent predictors of type 2 endoleak development. Reintervention rates and survival were compared between the two subgroups using Kaplan-Meier analysis, survival curves were compared using the log-rank test, and multivariable predictors of reintervention and survival were identified using Cox regression modeling. All statistical analyses were performed using SPSS statistical software version 20 (IBM Corp, Armonk, NY). Table I. Comparison of baseline characteristics and comorbidities between patients with no/transient type 2 endoleaks vs. patients with persistent/new type 2 endoleaks. No/Transient Persistent/New Type II Endoleak Type II Endoleak N=1977

N=390

P-value

73± 8.1

75±8.1

<.001

Male

82%

79%

0.157

White race

97%

97%

0.876

Smoking history

87%

80%

<.001

Hypertension

85%

82%

0.078

Diabetes Mellitus

19%

19%

0.947

Coronary Artery Disease

33%

32%

0.479

CABG/PCI

30%

28%

0.264

Congestive Heart Failure

8.4%

9.2%

0.613

COPD

34%

24%

<.001

Dialysis-dependent

0.8%

0.5%

0.538

Creatinine, >1.8 mg/dL

5.3%

2.6%

0.027

Aspirin

74%

70%

0.087

Statin

71%

69%

0.344

Plavix

7.4%

7.7%

0.862

Age (mean in yrs ± SD)

4

Preoperative Medication

CABG/PCI, coronary artery bypass graft/percutaneous coronary intervention; COPD, chronic obstructive pulmonary disease;

81

Chapter 4

Results Baseline Characteristics At the time of analysis, 2,367 patients in the VSGNE undergoing EVAR had information on endoleaks at both the end of the procedure as well as long-term follow-up. Group 1 included 1977 patients (84% of the total) with no or transient endoleak. Of these, 1560 (79%) had no endoleaks and 417 (21%) were transient endoleaks that had resolved spontaneously at follow-up. Group 2 consisted of 390 patients (16% of the total); of these, 120 (31%) had a persistent leak and 270 (69%) developed a new leak that was not seen at the time of surgery (Figure 1). There were 36 patients undergoing an intervention for type 2 endoleak. The baseline demographics and comorbidities of patients in groups 1 and 2 are shown in Table I. Patients in group 2 were older, less likely to have had high creatinine levels (>1.8) preoperatively, and less likely to have chronic obstructive pulmonary disease (COPD), or a history of smoking. Anatomic Characteristics and Surgical History Across all anatomical measured characteristics, no significant differences were found between the two groups (Table II). The mean aneurysm diameter was similar for both groups and there was no difference in iliac aneurysm involvement. Aortic rupture and prior aortic surgery did not differ across groups. Intraoperative Details Procedural details are shown in Table III. While coverage of the hypogastric artery and graft configuration (AUI vs. bifurcated) had no association with persistent type 2 endoleaks, hypogastric coil embolization was associated with more persistent type 2 endoleaks (12% vs. 8%, P=0.024). Of the 390 patients in group 2, 12% had a graft extension during surgery, while only 8% of those in group 1 had an extension (P=0.008). Table IV shows proportions of type 2 endoleak development at the procedure and at follow-up per graft type. Complications and Reinterventions At follow-up, data on sac growth were available for 104 patients, where 46% of patients with a persistent/new type 2 endoleak and 6% of patients with no/ transient type 2 endoleak had sac growth (P<.001). (Table V) Conversion to open repair was more common in patients with a persistent or new type 2 endoleak (0.8% vs. 0.2%, P=0.024). For the 2,330 patients with available followup data on any reintervention, those with persistent type 2 endoleaks were more likely to undergo reinterventions (18.6% vs. 1.5%), which was significant by Kaplan Meier (P<.001) (Figure 2). Among patients for whom long-term data were available regarding graft migration or symptoms/rupture there were no significant differences in the rates of these complications. Survival of patients with persistent type 2 endoleaks did not differ from those with no/transient type 2 endoleaks (Figure 3).

82

Type II endoleaks Table II. Aortic surgical history and anatomic characteristics. No/Transient Persistent/New Type II Endoleak Type II Endoleak

Aneurysm Diameter (mean, mm ± SD)

N=1977

N=390

P-value

57.0±24.2

57.7±11.1

0.615 0.412

Iliac Aneurysm None

77%

76%

Unilateral

13%

11%

Bilateral

11%

13% 0.775

Prior Aortic Surgery None

98%

98%

AAA

1.1%

0.5%

SAAA

0.2%

0.3%

Bypass

0.3%

0.5%

Other

0.8%

0.5%

EVAR

0.2%

0.0%

Rupture

2.4%

2.8%

0.605

AAA, infrarenal abdominal aortic aneurysm repair; SAAA, suprarenal abdominal aortic aneurysm repair Table III. Intraoperative details. No/Transient Type 2 Endoleak

Persistent/New Type 2 Endoleak

(N=1977)

(N=390)

%

%

P-value

0.755

Graft Configuration Aorto-bi-iliac

94.0%

94.9%

Aorto-uni-iliac

4.7%

3.8%

Aorto-aortic

1.3%

1.3%

Hypogastric Artery Coiling

8.3%

12.0%

0.024

Hypogastric Artery Coverage

12.4%

14.4%

0.287

Graft Extension

7.7%

12.0%

0.008

* Hypogastric artery coiling and graft extension data were missing for 183 patients

83

4

Chapter 4 Table IV. Development of persistent or new type 2 endoleaks at procedure and at follow-up by graft type. Type II Endoleak at procedure Type II Endoleak at Follow-up N

total

%

N

total

%

AneurRx

86

257

33.5%

27

257

10.5%

Endurant

61

194

31.4%

17

194

8.8%

Excluder Low-Permeability

175

747

23.4%

167

747

22.4%

Powerlink

60

289

20.8%

23

289

8.0%

Talent

61

147

41.5%

8

147

5.4%

Zenith

124

537

23.1%

76

537

14.2%

Grafts with fewer than 100 implants were not included in this analysis (Ancure, Excluder Original, MEGS/VI, Cordis, Aorfix, Zenith Low Profile, AFX, Endurant II, Zenith Flex, Other)

100

No/Transient Endoleak New/Persistent Endoleak

Freedom from reintervention [%]

80

60

P<.001 40

20

0

0

1

2

3

4

5

6

7

8

9

10

Years

Figure 2. Freedom from reinterventions among patients with no/transient type 2 endoleaks vs. patients with persistent/new type 2 endoleaks.

84

Type II endoleaks Table V. Complications and reinterventions at Follow-up. Data Available For

No/Transient Persistent/New Type 2 Endoleak Type 2 Endoleak

P-value

#

%

%

Any Reintervention

2,330

1.5%

18.6%

<.001

Conversion to Open Repair

2,298

0.2%

0.8%

0.024

Sac Growth

104

5.9%

45.7%

<.001

Graft Migration

103

11.4%

2.9%

0.082

Symptoms or Rupture

101

0.0%

4.5%

0.210

Table VI. Multivariable predictors of persistent/new type 2 endoleaks. Variable

Odds Ratio

95% CI

P-value

-

-

-

60-69

1.5

0.7-2.9

0.271

70-79

1.7

0.9-3.3

0.119

≼ 80

2.7

1.4-5.3

0.004

History of Smoking

0.7

0.5-1.0

0.076

COPD

0.7

0.5-0.9

0.017

Preop ASA

0.8

0.6-1.1

0.202

Hypertension

0.8

0.6-1.0

0.093

Creatinine, >1.8 mg/dL

0.5

0.3-1.0

0.059

Hypogastric coiling (preop or intraop)

1.5

1.0-2.2

0.044

Graft extension

1.6

1.1-2.3

0.025

AneuRx

0.8

0.5-1.3

0.357

Endurant

0.5

0.3-0.9

0.026

Excluder Low Permeability

1.5

1.1-2.0

0.011

Powerlink (Intuitrak)

0.5

0.3-0.9

0.012

Talent

0.6

0.3-1.1

0.119

Zenith

ref

ref

ref

Age <60

4

Graftype

85

Chapter 4 Table VII. Multivariable predictors of reintervention and survival Variable

Odds Ratio

95% CI

P-value

1.0

0.6-1.8

0.899

Smoking History

1.1

0.6-1.9

0.814

Creatinine, >1.8 mg/dL

0.6

0.2-2.5

0.454

COPD

1.2

0.8-2.0

0.394

Persistent/New Type 2 Endoleak

15.3

9.7-24.3

<.001

-

-

-

60-69

1.4

0.6-3.2

0.424

70-79

2.5

1.1-5.4

0.025

≼ 80

3.4

1.5-7.5

0.003

Female

1.1

0.8-1.5

0.551

Not Home Discharge

1.6

1.0-2.4

0.032

CHF

1.4

1.0-2.1

0.078

COPD

1.6

1.2-2.1

<.001

Creatinine, >1.8 mg/dL

1.7

1.1-2.8

0.031

Persistent/New Type 2 Endoleak

1.1

0.9-1.6

0.477

Predictors of Post-discharge Reintervention Female

Predictors of Mortality Age, years <60

86

Type II endoleaks Multivariable Predictors Multivariable predictors of persistent type 2 endoleaks were hypogastric artery coil embolization, distal graft extension, and age 80 or older, while COPD was protective. After accounting for these characteristics, the Powerlink (OR 0.5, 95% CI 0.3-0.9, P=0.012) and Endurant (OR 0.5, 95% CI 0.3-0.9, P=0.026) (which are both no longer available) were protective of persistent or new type 2 endoleaks, while the Excluder with its lower permeability fabric was predictive of persistent or new type 2 endoleaks (OR 1.5, 95% CI 1.1-2.0, P=0.011). (Table VI) Persistent type 2 endoleaks were predictive of post-discharge reinterventions (OR 15.3, 95% CI 9.7-24.3, P<.001), however not predictive of mortality (Table VII). Predictors of mortality were COPD, older age, creatinine level >1.8 mg/ dL, and discharge to an institutional facility. There was insufficient power to determine independent predictors of conversion to open repair.

No/Transient Endoleak New/Persistent Endoleak

100

Survival [%]

90

80

4

70

P=0.299

60

50

0

1

2

3

4

5

6

7

8

9

10

Years

Figure 3. Survival of patients with no/transient type 2 endoleaks vs. patients with persistent/new type 2 endoleaks.

87

Chapter 4

Discussion

In this large study of 2367 patients who underwent EVAR, we found that patients with persistent type 2 endoleaks are associated with hypogastric coil embolization, distal graft extension, the absence of COPD, age 80 and older, and graft type. Patients with persistent type 2 endoleaks are more likely to have sac expansion and to undergo reintervention during follow-up. The EUROSTAR registry showed that coil embolization of side branches increased the risk of type 2 endoleak. They showed that blocking one or two hypogastric arteries was predictive on univariate analysis, but not on multivariable analysis.10 In addition, Conrad et al. found an increased risk for secondary intervention in patients who had coil embolization of any vessel, but did not find an association of hypogastric embolization with persistent type 2 endoleaks.11 The results from our study provide a possible connection between these two prior findings, suggesting that the increased risk of reintervention, seen in patients with coil embolization, may stem from an increased risk of developing persistent type 2 endoleaks. A possible explanation for the increased risk of developing persistent type 2 endoleaks among patients who receive hypogastric artery coiling is that these patients likely have more extensive aneurysms that extend into the common iliac arteries. Although the presence of iliac aneurysms showed no correlation with persistent type 2 endoleaks, some iliac aneurysms might have a sufficient distal neck that does not require extension into the external iliac artery. Increased blood flow through lumbar arteries and the inferior mesenteric artery (IMA) may be enhanced to provide flow to the pelvis, thus contributing to persistent lumbar and IMA type 2 endoleaks. Coverage without coiling likely preserves some hypogastric branches that would otherwise be lost with embolization, which may reduce the need for enhanced flow from lumbar and IMA branches. In contrast, intraoperative coverage of the hypogastric arteries, whether intentional or not, was not associated with persistent type 2 endoleak development. In our analyses the Powerlink and Endurant stent grafts had lower rates of new or persistent type 2 endoleaks. The Endurant stent graft however is no longer available and was enhanced in 2013 into the Endurant II with a lower profile, additional limb lengths and an aorto-uni-iliac option. The number of patients treated with the Endurant II in our study was too small to draw any conclusions. The Powerlink is also no longer available in the US as it was replaced by the AFX device. Powerlink has a longer main body and external billowing fabric that should occupy more space thereby potentially decreasing type 2 endoleaks. However, the Zenith graft, which also has a longer main body was not associated with as low a rate of type 2 endoleak. The low-permeability Excluder was predictive of type 2 endoleaks. The original Excluder was associated with sac expansion despite an absence of radiographic endoleak, which led to an introduction of a new graft with lower permeability in 2004. This stent graft had the lowest profile for years and may have allowed an increase in treatment of those with peripheral arterial disease and women, which may have an impact on development of type 88

Type II endoleaks 2 endoleak, but we are unable to analyze this further. Our study demonstrates that the development of persistent type 2 endoleaks correlates with an increase in sac growth at follow up. This is consistent with a previous report, which found sac enlargement in 54.5% of patients with persistent type 2 endoleaks during their median follow-up period of 28.7 months.9 In our analyses, 46% of patients with persistent type 2 endoleaks had sac growth. Along with the risk of sac growth, we found that patients with persistent type 2 endoleaks were more likely to undergo a reinterventions. Buth et al. also found an increased risk for intervention and aneurysmal growth, without a change in rupture risk or survival.20 Jones et al. showed similar findings, but their results also revealed an increased risk of aneurysm rupture, which our study did not confirm.9 Our analysis includes data with a 1-year mean follow-up period, and thus limits our ability to assess the risk of late rupture. The overall increased risk of reintervention reinforces the need for continued surveillance of patients with persistent type 2 endoleaks and if necessary, intervening for sac expansion. Our study revealed no relationship between aneurysm size and persistent type 2 endoleaks. This is consistent with some earlier studies,9, 10 while other reports have demonstrated that the risk of developing type 2 endoleaks is increased with larger aneurysm diameter.21, 22 Aneurysm sac diameter was also found in one study to be an independent predictor of secondary reintervention.11 Our results, however, do not demonstrate this associated risk when accounting for persistent type 2 endoleak. Others have noted an association of persistent type 2 endoleak with and an increasing numbers of patent aortic branches. 21, 22 However, these data are not collected in the VSGNE dataset. Another significant predictor of type 2 endoleaks in our analyses is increased patient age, which is consistent with previous studies.10, 21 Van Marrewjik et al. similarly showed that patients with persistent endoleaks were two years older than patients without endoleaks.10 An explanation for this trend has yet to be proposed. Also, gender did not affect the incidence of persistent type 2 endoleak in our analyses.23 Recent work by Dubois et al. also showed that the incidence of endoleaks did not vary significantly across gender, despite differences in anatomic characteristics of aneurysms. 24 Earlier studies have reported an association between COPD and spontaneous resolution of type 2 endoleaks,21 which is confirmed in our study. This phenomenon may be explained by an increased blood viscosity in COPD leading to increased thrombus formation or atherosclerosis in vessels.10, 21 We also demonstrated on univariate analysis that smoking was associated with a lower rate of persistent type 2 endoleak, which has been seen in previous studies. 10, 20, 25 However, as smoking is the primary risk factor for COPD, these two variables may be collinear and in our study smoking was no longer significant after accounting for COPD in multivariable analysis.

89

4

Chapter 4 To limit the number of reinterventions that occur post-EVAR, further research is needed to explain the association between aneurysms that require hypogastric branch coiling and an increased risk of persistent type 2 endoleak. Although our study has a large number of patients from multiple institutions, an important caveat is that it is not a randomized controlled trial. This study is also limited by our lack of information on all potentially relevant anatomy, e.g. thrombus burden and patency of side branches. Importantly, long-term follow-up data for several outcome variables are limited (e.g. sac growth, graft migration, conversion to open, and rupture). It is also possible that there is reporting bias, in that those performing re-intervention or noting sac growth may be more likely to carefully inspect for and document the presence of a type 2 endoleak. However, to our knowledge, our study represents the largest American series and may be broadly generalizable due to both academic and community hospitals that the VSGNE incorporates.

Conclusion

Type II endoleaks that persist or develop during follow-up are associated with future sac growth and reintervention. Persistent type 2 endoleaks are predicted by coil embolization of hypogastric arteries, distal graft extension, older age, the absence of COPD, and graft type; allowing heightened awareness in those patients at increased risk. Future analysis with improved follow-up and modern grafts will improve our understanding of persistent type 2 endoleak.

90

Type II endoleaks

References 1. Prinssen M, Verhoeven EL, Buth J, Cuypers PW, van Sambeek MR, Balm R, et al. A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med. 2004;351(16):1607-18. 2. Lederle FA, Freischlag JA, Kyriakides TC, Padberg FT, Jr., Matsumura JS, Kohler TR, et al. Outcomes following endovascular vs open repair of abdominal aortic aneurysm: a randomized trial. Jama. 2009;302(14):153542. 3. Blankensteijn JD, de Jong SE, Prinssen M, van der Ham AC, Buth J, van Sterkenburg SM, et al. Two-year outcomes after conventional or endovascular repair of abdominal aortic aneurysms. N Engl J Med. 2005;352(23):2398405. 4. Greenhalgh RM, Brown LC, Powell JT, Thompson SG, Epstein D, Sculpher MJ. Endovascular versus open repair of abdominal aortic aneurysm. N Engl J Med. 2010;362(20):1863-71. 5. De Bruin JL, Baas AF, Buth J, Prinssen M, Verhoeven EL, Cuypers PW, et al. Long-term outcome of open or endovascular repair of abdominal aortic aneurysm. N Engl J Med. 2010;362(20):1881-9. 6. Lederle FA, Freischlag JA, Kyriakides TC, Matsumura JS, Padberg FT, Jr., Kohler TR, et al. Long-term comparison of endovascular and open repair of abdominal aortic aneurysm. N Engl J Med. 2012;367(21):1988-97. 7. Silverberg D, Baril DT, Ellozy SH, Carroccio A, Greyrose SE, Lookstein RA, et al. An 8-year experience with type II endoleaks: natural history suggests selective intervention is a safe approach. Journal of vascular surgery. 2006;44(3):453-9. 8. Steinmetz E, Rubin BG, Sanchez LA, Choi ET, Geraghty PJ, Baty J, et al. Type II endoleak after endovascular abdominal aortic aneurysm repair: a conservative approach with selective intervention is safe and cost-effective. Journal of vascular surgery. 2004;39(2):306-13. 9. Jones JE, Atkins MD, Brewster DC, Chung TK, Kwolek CJ, LaMuraglia GM, et al. Persistent type 2 endoleak after endovascular repair of abdominal aortic aneurysm is associated with adverse late outcomes. Journal of vascular surgery. 2007;46(1):1-8. 10. van Marrewijk CJ, Fransen G, Laheij RJ, Harris PL, Buth J. Is a type II endoleak after EVAR a harbinger of risk? Causes and outcome of open conversion and aneurysm rupture during follow-up. Eur J Vasc Endovasc Surg. 2004;27(2):128-37. 11. Conrad MF, Adams AB, Guest JM, Paruchuri V, Brewster DC, LaMuraglia GM, et al. Secondary intervention after endovascular abdominal aortic aneurysm repair. Ann Surg. 2009;250(3):383-9. 12. Timaran CH, Ohki T, Rhee SJ, Veith FJ, Gargiulo NJ, 3rd, Toriumi H, et al. Predicting aneurysm enlargement in patients with persistent type II endoleaks. Journal of vascular surgery. 2004;39(6):1157-62. 13. Tuerff SN, Rockman CB, Lamparello PJ, Adelman MA, Jacobowitz GR, 91

4

Chapter 4

Gagne PJ, et al. Are type II (branch vessel) endoleaks really benign? Ann Vasc Surg. 2002;16(1):50-4. 14. Rhee SJ, Ohki T, Veith FJ, Kurvers H. Current status of management of type II endoleaks after endovascular repair of abdominal aortic aneurysms. Ann Vasc Surg. 2003;17(3):335-44. 15. Baum RA, Carpenter JP, Stavropoulous SW, Fairman RM. Diagnosis and management of type 2 endoleaks after endovascular aneurysm repair. Tech Vasc Interv Radiol. 2001;4(4):222-6. 16. Rubin BG, Marine L, Parodi JC. An algorithm for diagnosis and treatment of type II endoleaks and endotension after endovascular aneurysm repair. Perspect Vasc Surg Endovasc Ther. 2005;17(2):167-72. 17. Gould DA, McWilliams R, Edwards RD, Martin J, White D, Joekes E, et al. Aortic side branch embolization before endovascular aneurysm repair: incidence of type II endoleak. J Vasc Interv Radiol. 2001;12(3):337-41. 18. Axelrod DJ, Lookstein RA, Guller J, Nowakowski FS, Ellozy S, Carroccio A, et al. Inferior mesenteric artery embolization before endovascular aneurysm repair: technique and initial results. J Vasc Interv Radiol. 2004;15(11):12637. 19. Bursac Z, Gauss CH, Williams DK, Hosmer DW. Purposeful selection of variables in logistic regression. Source Code Biol Med. 2008;3:17. 20. Buth J, Harris PL, van Marrewijk C, Fransen G. The significance and management of different types of endoleaks. Semin Vasc Surg. 2003;16(2):95102. 21. Abularrage CJ, Crawford RS, Conrad MF, Lee H, Kwolek CJ, Brewster DC, et al. Preoperative variables predict persistent type 2 endoleak after endovascular aneurysm repair. J Vasc Surg. 2010;52(1):19-24. 22. Brountzos E, Karagiannis G, Panagiotou I, Tzavara C, Efstathopoulos E, Kelekis N. Risk factors for the development of persistent type II endoleaks after endovascular repair of infrarenal abdominal aortic aneurysms. Diagn Interv Radiol. 2012;18(3):307-13. 23. Ouriel K, Greenberg RK, Clair DG, O’Hara P J, Srivastava SD, Lyden SP, et al. Endovascular aneurysm repair: gender-specific results. Journal of vascular surgery. 2003;38(1):93-8. 24. Dubois L, Novick TV, Harris JR, Derose G, Forbes TL. Outcomes after endovascular abdominal aortic aneurysm repair are equivalent between genders despite anatomic differences in women. Journal of vascular surgery. 2013;57(2):382-9 e1. 25. Koole D, Moll FL, Buth J, Hobo R, Zandvoort H, Pasterkamp G, et al. The influence of smoking on endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2012;55(6):1581-6.

92

Type II endoleaks

4

93

CHAPTER 5

Ambulatory Endovascular Abdominal Aortic Aneurysm Repair in a National and State Database Thomas Curran, Sarah Carlson, Dominique B. Buck, John C. McCallum, Jeremy Darling, Michelle Martin, Mark Wyers, Marc L. Schermerhorn. Submitted J Vasc Surg.

Chapter 5

Abstract Objectives Length of stay after AAA repair has decreased with endovascular repair (EVAR) though reports of ambulatory EVAR (aEVAR) are limited. We compared the safety and cost of aEVAR versus inpatient EVAR (iEVAR) using clinical and administrative databases. Methods All patients undergoing non-emergent EVAR in the NSQIP from 2005-2012 were identified. Patients with non-death, same-day discharge after EVAR had aEVAR while others had iEVAR. Demographics and outcomes were compared. Charges for aEVAR and iEVAR (1-2 day stay only) were compared using the Florida Inpatient and Ambulatory Surgery Databases from 2006-2011. Results Among 16,420 EVAR procedures in NSQIP, there were 53 aEVAR and 16,367 iEVAR. aEVAR patients were younger (69 vs. 74 years; P=.003), more often female (32% vs. 18%; P=.011) and less likely to have creatinine > 1.2g/dL (14% vs. 28%; P=.027). Femoral cut-down rate did not differ between aEVAR and iEVAR (45 vs. 50%; P=.495). Pre-discharge complications after aEVAR included DVT (1) and transfusion (2). Post-discharge complications included transfusion (2), return to the OR (2) and DVT (1). Administrative data showed median total charges of $67,000 for aEVAR (N=66) as compared to $90,600 for iEVAR (N=9,013). Conclusions Though aEVAR sample size is limited, this multi-center, national database shows aEVAR to have acceptable morbidity in select patients. Further studies may delineate aEVAR candidate selection criteria and explore differential resource utilization between aEVAR and iEVAR.

96

Ambulatory Endovascular A Repair

Introduction

Endovascular AAA repair (EVAR) now accounts for over three quarters of intact abdominal aortic aneurysm (AAA) repair in the United States,1 yet the cost effectiveness of EVAR as compared to open AAA repair remains controversial.2-4 Perioperative costs for EVAR are largely dominated by endograft device cost while those of open AAA repair relate to increased hospital stay and ICU utilization, shown to be significantly longer than EVAR across a number of randomized controlled trials (RCT’s).5-11 However, despite shorter length of stay for EVAR as compared to open repair, a review of six RCT’s still showed mean length of stay following EVAR to range from 3 to 6 days. Further decrease in length of stay after EVAR, if safe and feasible, may offer a means by which to increase cost effectiveness for this procedure that has already been so widely adopted. Looking outside the purview of vascular surgery, cardiac catheterization, now routinely performed in the ambulatory setting, was previously thought to be unsafe for outpatients until a 1988 prospective randomized trial by Block et al showed not only equivalent clinical outcomes but also substantial cost savings for ambulatory catheterization.12 Cochrane reviews of laparoscopic cholecystectomy, a procedure once requiring hospital admission, found no significant clinical differences for patients undergoing ambulatory laparoscopic cholecystectomy as compared to those with an overnight stay.13, 14 Similarly, ambulatory endovascular lower extremity arterial interventions have been shown to be safe and cost-effective in a number of single center studies, even for procedures for those with critical limb ischemia.15-17 With preliminary reports from the Veteran’s Administration and Europe demonstrating the feasibility of same day discharge after EVAR with high patient satisfaction,18, 19 ambulatory EVAR for select patients may offer improved cost effectiveness while still maintaining patient safety. However, despite the success demonstrated at these select single centers in the performance of ambulatory EVAR, evidence on its safety, utilization and costs in the U.S. are limited. In our study, we aim to evaluate the safety and cost of ambulatory EVAR as compared to inpatient EVAR using large clinical and administrative databases.

Methods Data Sources For the clinical component of our study, we utilized data from the National Surgical Quality Improvement Program (NSQIP) database from 2005 to 2012. NSQIP aggregates prospectively collected, clinical data from over 350 centers which contribute cases based upon overall volume through an algorithm designed to select a representative mix of qualifying cases. A detailed set of demographic, comorbidity, and perioperative information is collected through the use of proprietary software and medical record review by specially trained clinical nurse reviewers. A similarly detailed list of postoperative outcomes within thirty days of surgery, including post-hospitalization information, are collected 97

5

Chapter 5 from hospital records, clinic visits, and follow-up phone contact, with each participating institution required to maintain a follow up rate of 80% or better in order to maintain membership. The NSQIP data are subject to annual auditing and the reliability of accurate data acquisition has improved with each year.20 Further details regarding data collection and quality control are available through the NSQIP User’s Guide.21 Importantly, the NSQIP data set does not allow for the evaluation of trends in the utilization or cost of aEVAR given that (1) NSQIP participation is voluntary and not necessarily representative of trends outside of member institutions, (2) NSQIP cases are de-identified with respect to the contributing institution, (3) the number of NSQIP institutions has grown each year since from 121 in 2005 to 374 sites in 2012, and (4) NSQIP does not include cost or charge data.21 Further, the Nationwide Inpatient Sample, a 20% stratified sample of inpatient discharges from United States non-federal hospitals used to track trends in surgical practice,22-24 is also invalid for the evaluation of aEVAR usage as it, by definition, excludes ambulatory procedures. Thus, for the cost and utilization aspect of our study, we used the State Inpatient Database (SID) and State Ambulatory Surgery Database (SASD). The SID and SASD are administrative databases compiled as a part of the Healthcare Cost and Utilization Project (HCUP) under the auspices of the Agency for Healthcare Research and Quality (AHRQ). The SID is designed to capture hospital discharge data for all inpatient surgical procedures while the SASD is designed to capture all discharge data for ambulatory surgical procedures performed at both hospitals and, in certain states, freestanding surgery sites. Through the use of hospital billing data, diagnosis codes and procedural codes, these databases capture information on patient demographics, comorbidities and procedures. In order to avoid regional variation in hospital and freestanding surgery center charges, we restricted analysis to the Florida SID and SASD from 2006 to 2011, the most populous and high volume state for which detailed CPT code procedural data was available. Further detailed information may be found at http://www.hcup-us.ahrq.gov. As this study contained only de-identified data without any protected health information, the study is not considered human research and therefore not subject to institutional review board approval or patient consent. Patients For the purpose of clinical comparisons, patients undergoing ambulatory EVAR were identified via query of the NSQIP Participant User Files from 2005 to 2012 using the following Current Procedural Terminology (CPT) codes: 34800, 34802, 34803, 34804 and 34805. These relate to the performance of EVAR in the setting of infrarenal AAA or abdominal aortic dissection with each code corresponding to a different endograft configuration. For the purposes of our study, those undergoing EVAR with discharge on the day of procedure were considered to undergo ambulatory EVAR (aEVAR). Patients discharged on the day of surgery as a result of mortality were excluded from the ambulatory group. Those with a length of stay beyond the day of surgery were considered to undergo inpatient EVAR (iEVAR). Baseline patient demographics, comorbidities, operative details and postoperative course were extracted from the database. For hospital charge and utilization comparisons, aEVAR cases were extracted 98

Ambulatory Endovascular A Repair from the Florida SASD from 2006 to 2011. Cases were identified using the CPT codes as listed above. As these cases were necessarily ambulatory related to their inclusion in the SASD, all SASD cases were considered to undergo aEVAR. Inpatient EVAR cases were taken from the Florida SID from 2006 to 2011. Rather than the CPT codes used for procedure identification in the SASD and NSQIP, the SID employs International Classification of Diseases, 9th Revision (ICD-9) procedure codes to identify procedures. All cases with an ICD-9 code of 39.71, corresponding to “endovascular implantation of graft in abdominal aorta,” were selected for inclusion. In attempting to identify those patients undergoing routine, elective iEVAR, we restricted charge analysis to those with a length of stay of one or two days only. Patients experiencing a death event were also excluded. Additionally, EVAR patients in the SID with a non-fatal discharge and length of stay less than one day were considered as a separate aEVAR group. Outcomes Our primary clinical outcome measures included incidence of NSQIP-defined complications within thirty days of surgery, stratified by pre- and post-discharge occurrence. A list of NSQIP-defined complications and their specific definitions may be found in the NSQIP User Guide.21 In 2011, the NSQIP introduced a variable for readmission within 30-days of surgery to any hospital, including nonNSQIP hospitals, as determined by medical record review and direct patient contact. For this reason readmission analysis was restricted to 2011 and 2012 only. The accuracy of NSQIP readmission data was compared to that of physician chart review and administrative data and found to be excellent.25 We also provide a descriptive comparison of those patients undergoing aEVAR and iEVAR with respect to patient demographics, comorbidities and operative details. Trends in the utilization of aEVAR were assessed by recording the number of cases performed each year in the state of Florida from 2006 to 2011. Total hospital or surgical center charges associated with the episode of care were also recorded and compared between aEVAR and routine iEVAR. Statistical Analysis All analyses were conducted using IBM SPSS Statistics version 21.0.0 for Macintosh (IBM Corp., Armonk, NY). Categorical variables were analyzed using the chi-square or Fisher’s exact test where appropriate. Continuous variables were compared using two-tailed independent samples t-test. Test for trend over time was assessed using the chi-square test for trend. Cases missing data for any given parameter were eliminated from consideration for the purposes of bivariate analysis. Throughout all analyses, statistical significance was determined by a criterion of p < .05.

Results Demographics/Clinical Details - NSQIP Compared to patients who underwent iEVAR, those undergoing aEVAR were younger (69 vs. 74 years, p < 0.01), and more likely to be female (32.1% vs. 17.9%, p = 0.01), (Table I). The aEVAR population were less likely to have 99

5

Chapter 5 history of prior percutaneous coronary intervention (5.3% vs. 21.0%, p = 0.02) or baseline creatinine > 1.2 (14.0% vs. 27.9%, p = 0.03). There was a trend toward lower prevalence of COPD, obesity, and ASA class IV status among aEVAR patients, though these did not reach statistical significance in this population. No patient undergoing aEVAR had history of MI or was dialysis dependent. There was no difference in operative time (146 minutes for aEVAR vs. 151 minutes for iEVAR, p = 0.66), endograft type used, or surgeon specialty (Table II). The rate of groin cut-down was similar between groups (45.3% for aEVAR vs. 50.1% for iEVAR, p = 0.50). Table I: Demographics and Comorbidities of Patients Undergoing EVAR in the 2005-2012 NSQIP All (N = 16,420)

Ambulatory (N = 53)

Non-Ambulatory (N = 16,367)

P-Value

73.8 (8.5)

69.0 (11.3)

73.9 (8.5)

0.003

Female

2,934 (17.9)

17 (32.1)

2,916 (17.9)

0.011

Diabetes Mellitus

2,561 (15.6)

9 (17.0)

2,552 (15.6)

0.707

Smoker

4,862 (29.6)

20 (37.7)

4,842 (29.6)

0.227

History or Myocardial Infarction

115 (0.9)

0

115 (0.9)

1.000

Prior Cardiac Surgery

2,924 (23.0)

6 (15.8)

2,918 (23.0)

0.340

Prior Percutaneous Coronary Intervention

2,662 (20.9)

2 (5.3)

2,660 (21.0)

0.015

Chronic Obstructive Pulmonary Disease

3,098 (18.9)

6 (11.3)

3,092 (18.9)

0.216

144 (0.9)

0

144 (0.9)

1.000

Creatinine > 1.2 g/dL

4,430 (27.8)

7 (14.0)

4,423 (27.9)

0.027

BMI > 30 kg/m2

5,173 (31.7)

10 (18.9)

5,163 (31.8)

0.057

ASA Class 4

2,991 (18.2)

5 (9.4)

2,986 (18.3)

0.109

Dependent Functional Status

392 (2.4)

0

392 (2.4)

0.640

History of Extremity Revascularization or Amputation

731 (5.7)

3 (7.9)

728 (5.7)

0.479

Age, years; Mean (SD)

Dialysis Dependent

Postoperative Outcomes - NSQIP Postoperative complication rates were similar between groups. Among patients undergoing aEVAR, 4 of 53 (8%) experienced complications, 2 of whom had complications prior to discharge (both with bleeding/transfusion), and 2 postdischarge (1 patient with deep vein thrombosis and bleeding/transfusion; 1 patient 100

Ambulatory Endovascular A Repair with bleeding/transfusion only). Two patients in the aEVAR group required return to the operating room within 30 days although indication for reoperation was not available. Overall complication rate among iEVAR patients was 200/1637 (12%), which was not significantly higher than that of the aEVAR cohort (p = 0.40). There was no difference in rate of readmission (4% vs. 8%, p = 0.72) or return Table II: Operative Details for Patients Undergoing EVAR in the 2005-2012 NSQIP Overall N (%) OR Time, minutes; Mean (SD)

151.6 (72.1)

Ambulatory Non-Ambulatory N (%) N (%) 146.2 (91.5)

150.6 (72.1)

Procedure

P-Value 0.661 0.339

Aorto-aortic tube prosthesis (34800)

1,418 (8.6)

6 (11.3)

1,412 (8.6)

Modular bifurcated (1 limb) (34802)

8,024 (48.9)

31 (58.5)

7,993 (48.8)

Modular bifurcated (2 limbs) (34803)

5,021 (30.6)

13 (24.5)

5,008 (30.6)

Unibody bifurcated (34804)

1,317 (8.0)

3 (5.7)

1,314 (8.0)

640 (3.9)

0

640 (3.9)

Aorto-uniiliac (34805) Surgeon Specialty

0.200

Vascular

15,846 (98.6)

48 (90.6)

15,798 (96.5)

General

509 (3.1)

5 (9.4)

504 (3.1)

Cardiac

33 (0.2)

0

33 (0.2)

Thoracic

30 (0.2)

0

30 (0.2)

20

5

# aEVAR

15 10 5 0

2006

2007

2008

2009

2010

2011

Figure 1: Number of Ambulatory EVAR Performed in Florida SASD by Year

101

Chapter 5 Total Charges by Year Total Charge (USD)

200000

aEVAR (SASD) aEVAR (SID) iEVAR

150000 100000 50000

20 11

20 10

20 09

20 08

20 07

20 06

0

Year Figure 2: Total Hospital Charges for Ambulatory vs. Inpatient EVAR in Florida SASD and SID Table III: Postoperative Outcomes in Patients Undergoing EVAR in the 2005-2012 NSQIP Overall N (%) Length of Postoperative Stay, days; Mean (SD)

Ambulatory Non-Ambulatory N (%) N (%)

P-Value

2.5 (3.8)

0

2.5 (4)

<0.001

2,262 (13.8)

4 (8)

2,001 (12)

.401

Pre Discharge

1,756 (10.7)

2 (4)

1,372 (8)

.320

Post Discharge

714 (4.3)

2 (4)

712 (4)

1.000

168 (1.0)

0

168 (1.0)

1.000

In House

97 (0.6)

x

97 (0.6)

1.000

Post Discharge

54 (0.3)

0

54 (0.3)

1.000

Readmission

205/2,495 (8.3)

1/29 (4)

443/5,640 (8)

.721

Return to OR

612 (3.7)

2 (4)

610 (4)

1.000

Any Complication

Mortality

102

Ambulatory Endovascular A Repair Charges by Department

Charge (USD)

6000

aEVAR (SASD) aEVAR (SID) iEVAR

4000

2000

A

ne

st

he

si a R ec ov er y

gy

b

lo

La

io ad R

I Ph CU ar m ac y

R

oo

m

0

Figure 3A: Charge Breakdown for Inpatient vs. Ambulatory EVAR in Florida SASD and SID for General Charges Charges for OR and Supplies

Charge (USD)

80000

p < 0.001

60000 40000

aEVAR (SASD) aEVAR (SID) iEVAR

p < 0.001

20000

s lie Su

pp

O

R

0

Figure 3B: Charge Breakdown for Inpatient vs. Ambulatory EVAR in Florida SASD and SID for Operating Room and Device Costs

to OR (4% vs. 4%) in aEVAR versus iEVAR patients, respectively. Thirty-day mortality rate was zero among aEVAR patients compared to 1% among iEVAR patients (p = 1.000), (Table III). Utilization and Charges – SASD/SID The total number of aEVAR identified in the Florida SASD from 2006-2011 was 66 while the Florida SID contained 15,988 total EVAR. Of those within the SID, 65 were aEVAR, 9,013 were non-fatal iEVAR with length of stay equal to one or two days and 6,910 were iEVAR with either mortality or non-routine length of stay. The number of aEVAR performed in the SASD by year are shown in Figure 1. Overall charges by year generally increased over the study period for both aEVAR and iEVAR (Figure 2). Mean charges associated with aEVAR in the SASD were less than those of iEVAR with routine stay ($66,938 vs. $98,872; P<0.001) and aEVAR in the SID ($66,938 vs. $90,633; P<.001). Moreover, charges between aEVAR in the SID and iEVAR failed to demonstrate a difference ($90,633 vs. 98,872; P=0.146). Itemized charges between groups are shown in Figure 3. Compared to iEVAR, aEVAR in the SASD had lower charges for both supplies 103

5

Chapter 5 Charge by Length of Stay Total Charge (USD)

150000

p < 0.001

100000

50000

2

1

r) nt C en te ie

(In pa t 0

0

(A m bu la to ry

C en te r)

0

Length of Stay (Days) Figure 4: Charges by Length of Stay for Ambulatory vs. Inpatient EVAR in Florida SASD and SID

and the operating room. Figure 4 demonstrates institutional charges stratified by length of stay for each group. The charges within the SID were similar regardless of length of stay.

Discussion

In this first study of a national clinical database investigating patient selection and outcomes related to aEVAR, we found that aEVAR was performed with morbidity similar to that of iEVAR in NSQIP participating centers though patients in the aEVAR group were younger and generally healthier than those in the inpatient cohort. However, utilization of aEVAR within Florida remains exceedingly low, accounting for less than one percent of all EVAR (N = 123/16,054). Charges for aEVAR in the SASD were, on average, approximately one third less than those of iEVAR. Yet, interestingly, charges for aEVAR in the SID, while less than those of iEVAR, were significantly greater than aEVAR in the SASD. Thus, the same clinical service of aEVAR with ambulatory status in the SASD produced charges, on average, 26% less than the same clinical service of aEVAR with inpatient status in the SID, despite a discharge on day of surgery. Regarding the perioperative clinical outcomes following aEVAR, the NSQIP series compares favorably to prior reports from Lachat and Dosluoglu.18, 19 Lachat and colleagues published 100 cases of ambulatory EVAR performed at two European centers with excellent clinical results, low readmission rate (4%), substantial cost savings and high patient satisfaction. Dosluogu et al from the Veteran’s Administration reported the first U.S. series on ambulatory EVAR with 104

Ambulatory Endovascular A Repair 21 patients successfully undergoing the procedure with acceptable results. No post-discharge deaths occurred in either series. Thirty-day readmission rate for the NSQIP aEVAR group was 4% (N = 1/29) which was similar to those reported by Lachat et al (4%; N = 4/100) and Dosluoglu and colleagues (3%; N = 1/29). Notably, the readmission rate for the iEVAR NSQIP cohort was 8%. Differences did exist between the NSQIP group and these previously published reports in certain regards, however. Intraoperatively, approximately half of the NSQIP aEVAR group (45.3%) received cut-down vascular access while Dosluoglu et al specifically excluded cut-down patients from same day discharge and Lachat and colleagues performed cut-down access on only 12% of patients. Preoperatively, selection for aEVAR candidacy was well defined in these previously reported series with specific criteria set out a priori. Such criteria permitted Dosluoglu et al to perform one third of percutaneous EVAR with same day discharge even among patients deemed eligible for aEVAR. The criteria included favorable anatomy, age < 80 years, good performance status, normal renal function, absence of congestive heart failure or oxygen dependent COPD, residence < 50 miles from hospital, and a companion to stay with them on the night of surgery. Lachat et al defined their aEVAR exclusion criteria as any acute medical condition or hemodialysis dependence though did not note the proportion of patients receiving EVAR who did not meet those. As NSQIP data availability precludes insight toward the surgeon’s thought process with respect to preoperative selection for aEVAR, further study will be required to best assess candidacy outside of these select centers and/or apply the criteria utilized by Lachat or Dosluoglu to a new clinical setting. Ambulatory EVAR, in both the SASD and the SID, had charges significantly less than routine stay inpatient EVAR. This was similar to the findings of Lachat and colleagues who also demonstrated decreased cost for ambulatory EVAR as compared to inpatient EVAR in 21 pairs of inpatient/ambulatory patients matched for graft type and year of treatment (€13,732 vs. €15,903; p = .05). With the exception of physician fees and materials costs, the Lachat study demonstrated decreased costs for each component of treatment in the ambulatory group as compared to the inpatient group (e.g. management costs, nurse fees, etc.). However, while our data point out the striking charge disparity between aEVAR and iEVAR, it is important to note that charges for aEVAR in the SASD were also less than those for aEVAR in the SID. This is in accordance with data on lower extremity peripheral arterial interventions which have also demonstrated lower costs in the ambulatory setting as compared to the hospital setting.15, 16 Similarly, ambulatory laparoscopic cholecystectomy was found to have lower charges in freestanding surgical centers as compared to hospitals.26 Yet, as none of the aEVAR in the SASD were performed in freestanding facilities, but rather in hospitals, the charge disparity between aEVAR in the SASD and aEVAR in the SID suggests that charges may be related to specific billing conventions (DRG vs. non-DRG) rather than a tangible difference in true healthcare resource utilization. Further study at the institutional level would likely be required to ascertain the exact nature of this discrepancy whereby different charges are assessed for the same clinical service in the same setting based on admission status. The institutional experience of the Dartmouth group may provide insight in regard 105

5

Chapter 5 to the potential magnitude of cost savings with ambulatory EVAR. Stone and colleagues showed inpatient EVAR costs to create a net mean total technical EVAR associated operating margin loss of over $4,000 per procedure when performed for Medicare beneficiaries.27 Importantly, this report was generated by an institution with a mean EVAR length of stay of 1.7 days and a cost profile in the lowest quartile of reported costs for Diagnosis Related Group 238, that which relates to EVAR. Itemized cost analysis found that over half of total technical costs were related to the endograft itself. Operating room costs accounted for 17% of total costs while inpatient bed costs represented 7% of the total. Hospital overhead was responsible for 12% of costs. Thus, with an already low mean length of stay and costs in the lowest quartile of reporting centers, the Dartmouth group still demonstrated a significant per procedure operating loss for performance of EVAR. While performance of EVAR in the ambulatory setting may provide cost savings for operating room, institutional overhead and inpatient bed fees, these incremental improvements should not undermine the need for more competitive device costs. The findings of this study must be interpreted within the context of its retrospective design. Regarding the clinical data derived from the NSQIP database, these institutions represent a self-selected group of hospitals with a demonstrated interest in quality improvement. Moreover, as the data generated from the NSQIP is de-identified with respect to center, it is possible that the aEVAR sample relates to a relatively small number of highly specialized centers. As such, these data may be non-representative of a broader group of institutions and should be generalized with caution. Additionally, we are unable to discern the number of cases planned for aEVAR that were subsequently admitted to the hospital secondary to physician concern and/or complications. However, despite this, our study demonstrates feasibility in regard to performance of aEVAR in a select subset of those undergoing the procedure. Finally, our data pertain to EVAR related institutional charges and not reimbursement. Whereas the NIS provides cost-to-charge ratios for various institutions and procedures, these data are unavailable for the SASD and SID. Thus, charge based comparisons were necessary though it is possible that the charges presented may not reflect true payer or hospital cost. In conclusion, our data show aEVAR to be clinically safe and feasible in select patients though only representing a minute fraction of overall EVAR being performed. Further, as prior data has shown, such a practice should take place within the framework of a well-defined treatment algorithm to determine appropriate patient candidacy for aEVAR. When performed in highly selected patients, aEVAR may be cost-effective relative to iEVAR yet the specific means by which these cost savings are achieved (i.e. decrease in resource utilization vs. billing practice) will require institutional level study to better elucidate.

106

Ambulatory Endovascular A Repair

References 1. Schermerhorn ML, Bensley RP, Giles KA, Hurks R, O’Malley A J, Cotterill P, et al. Changes in abdominal aortic aneurysm rupture and short-term mortality, 1995-2008: a retrospective observational study. Annals of surgery. 2012;256(4):651-8. 2. Epstein D, Sculpher MJ, Powell JT, Thompson SG, Brown LC, Greenhalgh RM. Long-term cost-effectiveness analysis of endovascular versus open repair for abdominal aortic aneurysm based on four randomized clinical trials. The British journal of surgery. 2014;101(6):623-31. 3. Blackhouse G, Hopkins R, Bowen JM, De Rose G, Novick T, Tarride JE, et al. A cost-effectiveness model comparing endovascular repair to open surgical repair of abdominal aortic aneurysms in Canada. Value in health : the journal of the International Society for Pharmacoeconomics and Outcomes Research. 2009;12(2):245-52. 4. Lederle FA, Stroupe KT, Open Versus Endovascular Repair Veterans Affairs Cooperative Study G. Cost-effectiveness at two years in the VA Open Versus Endovascular Repair Trial. European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery. 2012;44(6):543-8. 5. Dangas G, O’Connor D, Firwana B, Brar S, Ellozy S, Vouyouka A, et al. Open versus endovascular stent graft repair of abdominal aortic aneurysms: a meta-analysis of randomized trials. JACC Cardiovascular interventions. 2012;5(10):1071-80. 6. Becquemin JP, Pillet JC, Lescalie F, Sapoval M, Goueffic Y, Lermusiaux P, et al. A randomized controlled trial of endovascular aneurysm repair versus open surgery for abdominal aortic aneurysms in low- to moderate-risk patients. J Vasc Surg. 2011;53(5):1167-73 e1. 7. De Bruin JL, Baas AF, Buth J, Prinssen M, Verhoeven EL, Cuypers PW, et al. Long-term outcome of open or endovascular repair of abdominal aortic aneurysm. N Engl J Med. 2010;362(20):1881-9. 8. Greenhalgh RM, Brown LC, Powell JT, Thompson SG, Epstein D, Sculpher MJ. Endovascular versus open repair of abdominal aortic aneurysm. N Engl J Med. 2010;362(20):1863-71. 9. Lottman PE, Laheij RJ, Cuypers PW, Bender M, Buth J. Health-related quality of life outcomes following elective open or endovascular AAA repair: a randomized controlled trial. J Endovasc Ther. 2004;11(3):323-9. 10. Lederle FA, Freischlag JA, Kyriakides TC, Padberg FT, Jr., Matsumura JS, Kohler TR, et al. Outcomes following endovascular vs open repair of abdominal aortic aneurysm: a randomized trial. JAMA : the journal of the American Medical Association. 2009;302(14):1535-42. 11. Soulez G, Therasse E, Monfared AA, Blair JF, Choiniere M, Elkouri S, et al. Pain and quality of life assessment after endovascular versus open repair of abdominal aortic aneurysms in patients at low risk. J Vasc Interv Radiol. 2005;16(8):1093-100. 107

5

Chapter 5 12. Block PC, Ockene I, Goldberg RJ, Butterly J, Block EH, Degon C, et al. A prospective randomized trial of outpatient versus inpatient cardiac catheterization. N Engl J Med. 1988;319(19):1251-5. 13. Vaughan J, Nagendran M, Cooper J, Davidson BR, Gurusamy KS. Anaesthetic regimens for day-procedure laparoscopic cholecystectomy. Cochrane Database Syst Rev. 2014;1:CD009784. 14. Vaughan J, Gurusamy KS, Davidson BR. Day-surgery versus overnight stay surgery for laparoscopic cholecystectomy. Cochrane Database Syst Rev. 2013;7:CD006798. 15. Shindelman LE, Ninnul GB, Curtiss SI, Konigsberg SF. Ambulatory endovascular surgery: cost advantage and factors influencing its safe performance. Journal of endovascular surgery : the official journal of the International Society for Endovascular Surgery. 1999;6(2):160-7. 16. O’Brien-Irr MS, Harris LM, Dosluoglu HH, Dayton M, Dryjski ML. Lower extremity endovascular interventions: can we improve cost-efficiency? J Vasc Surg. 2008;47(5):982-7; discussion 7. 17. Zayed HA, Fassiadis N, Jones KG, Edmondson RD, Edmonds ME, Evans DR, et al. Day-case angioplasty in diabetic patients with critical ischemia. International angiology : a journal of the International Union of Angiology. 2008;27(3):232-8. 18. Lachat ML, Pecoraro F, Mayer D, Guillet C, Glenck M, Rancic Z, et al. Outpatient endovascular aortic aneurysm repair: experience in 100 consecutive patients. Ann Surg. 2013;258(5):754-8; discussion 8-9. 19. Dosluoglu HH, Lall P, Blochle R, Harris LM, Dryjski ML. Ambulatory percutaneous endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2013. 20. Shiloach M, Frencher SK, Jr., Steeger JE, Rowell KS, Bartzokis K, Tomeh MG, et al. Toward robust information: data quality and inter-rater reliability in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg. 2010;210(1):6-16. 21. ACS NSQIP User Guide for the 2012 Participant Use Data File. 2013; Available from: http://site.acsnsqip.org/participant-use-data-file/. 22. Timaran CH, Rosero EB, Smith ST, Valentine RJ, Modrall JG, Clagett GP. Trends and outcomes of concurrent carotid revascularization and coronary bypass. J Vasc Surg. 2008;48(2):355-60; discussion 60-1. 23. Schermerhorn ML, Giles KA, Hamdan AD, Wyers MC, Pomposelli FB. Mesenteric revascularization: management and outcomes in the United States, 1988-2006. J Vasc Surg. 2009;50(2):341-8 e1. 24. LaPar DJ, Bhamidipati CM, Mery CM, Stukenborg GJ, Jones DR, Schirmer BD, et al. Primary payer status affects mortality for major surgical operations. Ann Surg. 2010;252(3):544-50; discussion 50-1. 25. Sellers MM, Merkow RP, Halverson A, Hinami K, Kelz RR, Bentrem DJ, et al. Validation of new readmission data in the american college of surgeons National Surgical Quality Improvement Program. J Am Coll Surg. 2013;216(3):420-7. 26. Paquette IM, Smink D, Finlayson SR. Outpatient cholecystectomy at 108

Ambulatory Endovascular A Repair hospitals versus freestanding ambulatory surgical centers. J Am Coll Surg. 2008;206(2):301-5. 27. Stone DH, Horvath AJ, Goodney PP, Rzucidlo EM, Nolan BW, Walsh DB, et al. The financial implications of endovascular aneurysm repair in the cost containment era. J Vasc Surg. 2014;59(2):283-90 e1.

5

109

CHAPTER 6

Comparison of Endovascular stent grafts in Patients with Abdominal Aortic Aneurysms in the Medicare Population. Dominique B. Buck, Bruce E. Landon, A. James O’Malley, Sara L. Zettervall, Peter A. Soden, Lawrence Zaborski, Klaas H.J. Ultee, Marc L. Schermerhorn.

Chapter 6

Abstract Background Previous studies have demonstrated the safety and efficacy of multiple stent grafts for endovascular repair of abdominal aortic aneurysm (AAA), however comparative data is limited. Medicare data allow identification of stent grafts by current procedural terminology (CPT) codes that distinguish bifurcated stent grafts as either single docking limb, two docking limb, or unibody. The aim of this study is to compare outcomes between these formations of stent grafts. Methods Medicare beneficiaries undergoing endovascular repair of AAA from 2005-2008 with the single docking limb (AneuRx/Excluder), two docking limb (Zenith) and unibody (Powerlink) bifurcated stent grafts were identified. Propensity score weighting was utilized to account for differences in patient characteristics between the different graft formations. Perioperative outcomes including mortality, length of stay, and post-operative complications; as well as 4-year rates of survival, rupture, and reintervention were evaluated. Results 46,171 Medicare beneficiaries were identified including 32,909 (71%) single docking limb (AneuRx/Excluder), 11,002 (24%) two docking limb (Zenith) and 2,260 (5%) unibody (Powerlink) bifurcated stent grafts. After propensity score weighting, there were no significant differences in patients’ baseline and clinical demographics among groups. Zenith had higher perioperative mortality (1.7% vs. 1.3%, P=.001), renal failure (6.0% vs 4.7%, P<.001), mesenteric ischemia (0.7 vs. 0.4%, P=.001), and a longer length of stay (3.4 days vs. 3.0 days, P<.001) compared to patients with AneuRx/Excluder. Powerlink had higher rates of renal failure (5.9% vs. 4.7%, P<.001), mesenteric ischemia (1.0% vs. 0.4%, P<.001), and conversion to open repair (0.7% vs. 0.1%, P<.001) compared to AneuRx/ Excluder. At 4 years, mortality remained higher with Zenith than AneuRx/Excluder (30% vs. 29%, P=.047). However, Zenith had fewer conversions to open repair (0.6% vs. 0.9%, P=.033), and aneurysm related reinterventions (10% vs. 12%, P=.003), mainly driven by extension cuffs (2.9% vs. 4.1%, P<.001). At 4 years, Powerlink had more aneurysm related interventions (15% vs. 12%, P=.002), driven by extension cuffs (7.8% vs. 4.1%, P<.001) compared to AneuRx/ Excluder, but Powerlink had fewer conversions to open repair (0.4% vs. 0.9%, P=.022). Late rupture did not differ between groups. Conclusions Compared to AneuRx/Excluder, from 2005-2008, Zenith had higher perioperative mortality but fewer reinterventions including conversion, while Powerlink had more perioperative complications and aneurysm-related reinterventions, but fewer conversions to open repair. Despite technologic changes over time and our inability to distinguish AneuRx from Excluder, this study highlights the need for further studies comparing the effectiveness of different stent grafts. 112

Comparison of Endovascular Stent Grafts

Introduction

Endovascular aneurysm repair (EVAR) has become the standard of care in many hospitals for patients with abdominal aortic aneurysms (AAA).1, 2 Clinical evidence suggests that EVAR is associated with lower morbidity and mortality than open repair; however EVAR may lead to an increase in reinterventions. Moreover, late rupture and conversion to open repair remain a concern following EVAR.2-5 Previous literature has demonstrated the safety and feasibility of using commercially available stent grafts.2-4, 6, 7 Postmarket surveillance, however, has highlighted issues with each of these grafts leading to changes in their design. There are little data comparing the effectiveness of the various stent grafts to help guide a physician’s choice of one stent graft over another. The Eurostar collaborators compared six stent grafts to the Vanguard stent graft. This study found the Excluder to be the only graft to prevent rupture and death compared to Vanguard.8 Their study, however, included stent grafts that are currently off the market or that were never available in the United States. Other studies addressing this issue are small and therefore likely underpowered to detect meaningful differences between grafts.9-12 In addition, to the extent that there are tradeoffs related to perioperative versus longer terms outcomes and costs, these should also be identified. In this study, we compare the most commonly used stent grafts in the United States with data from Medicare Beneficiaries, to evaluate differences in perioperative and 4-year outcomes.

Methods Patients All Medicare beneficiaries who underwent endovascular repair of an intact abdominal aortic aneurysm between January 1st 2001 and December 31st 2008 were identified. Only patients continuously enrolled in Medicare parts A and B for at least two years prior to their repair were included. Patients with ruptured abdominal aortic aneurysm, thoracic aneurysms, thoracoabdominal aortic aneurysms, and aortic dissections, were excluded (Supplemental Appendix). Beneficiaries who enrolled in health maintenance organizations at any time during the follow-up period were censored from the analyses of complications and reinterventions due to the lack of subsequent claims data. Survival data were available for all patients. This study was approved by the institutional review board at Harvard Medical School. Stent grafts Stent graft type was determined from CPT codes, including 34804 for the unibody bifurcated (Powerlink), 34802 for the bifurcated single docking limb stent graft (AneuRx/Excluder), and 34803 and 0001T for the bifurcated two docking limb stent graft (Zenith). (Table I) During the study period, these graft formations represented a limited number of grafts such that the Cook Zenith was 113

6

114 1999 2002 2004 2003 2006 2004

Excluder® (W. L. Gore & Associates, Inc., USA)

Low-permeability Excluder® (W. L. Gore & Associates, Inc., USA)

Zenith® (Cook Medical Technologies, USA)

Enlarged-neck Zenith® (Cook Medical Technologies, USA)

Powerlink® (Endologix, Inc., USA)

18–26

18–32

18–28

19–26

19–26

18–25

Year of Neck FDA ap- Diameter proval (mm)

AneuRx® (Medtronic Vascular, Inc., USA)

Endovascular Device

≥15

≥15

≥15

≥15

≥15

≥10‡

Neck Length (mm)

≤60

≤60

≤60

≤60

≤60

≤45

Neck Angulation (°)

≥15

≥15

≥15

≥10

≥10

NS

8–18

10–20

10–20

10–18.5

10–18.5

NS

Iliac neck Iliac neck length diameter (mm) (mm)

Table I. Commercial endovascular stent grafts available in the U.S.

34804

34803

34803

34802

34802

34802

ICD-9 code

Unibody bifurcated

Two docking limb

Two docking limb

Single docking limb

Single docking limb

Single docking limb

Stent graft

Chapter 6

Comparison of Endovascular Stent Grafts the only bifurcated graft with two docking limbs, the Endologix Powerlink was the only unibody bifurcated graft. The bifurcated single docking limb formation represented both the Medtronic AneuRx and the Gore Excluder, but we could not distinguish between these two. Outcomes Perioperative surgical complications (e.g. conversion from endovascular to open repair, return to the operating room, etc.) and medical complications (e.g. myocardial infarction, pneumonia) were identified using ICD-9-CM diagnostic and complication codes, as well as physicians’ current-procedural-terminology (CPT) codes using both hospital and physician claims (Supplemental Appendix). Discharge disposition (home or facility) and length of stay for each patient was recorded. Mortality was identified from the Medicare Beneficiary Summery File. Perioperative mortality was defined as death during the index admission or within 30 days after surgery and long-term mortality included all deaths during the available follow-up period. Subsequent hospitalizations and outpatient interventions occurring after repair related to AAA, including hospitalizations for rupture, major reinterventions (e.g. open repair of the aneurysm/pseudoaneurysm, repair of a graft–enteric fistula or graft infection), and minor reinterventions (e.g., stent–graft extension, coil embolization, aortic or iliac angioplasty, graft thrombectomy) also were identified from both hospital and physician claims. Statistical Analyses The characteristics of the unmatched cohorts were compared using chi-square tests and t-tests (or F-tests for simultaneous comparisons across all treatment groups) as appropriate. To account for differences in patients characteristics and other predictors between the sub-samples of patients we used a propensity score weighted approach. We computed the estimated propensity of a patient with given characteristics of receiving one of the types of stents relative to the other stents on the market at that time. The fitted generalized logistic regression model estimated three probabilities (corresponding to receiving the Zenith, Powerlink, and the AneuRx or Excluder stents) that are constrained to sum to one. The propensity to receive a given stent was equalized or balanced across the groups of patients by weighting observations by the inverse of the estimated probability of receiving that stent. With three or more treatment groups, inverse probability weighting has an advantage compared to pairwise propensity score matching of ensuring that estimates pertain to the same population. Specifically, estimated effects of each stent represent the average treatment effect (ATE) over all the patients in the sample. Analogous to propensity score methods for two treatment groups, balancing the distribution of a propensity score across three or more treatment groups is expected to balance the distribution of the predictors in the propensity score model. To address residual imbalances in the distribution of the predictors across the samples of patients receiving each stent, we added the square of any continuous predictors lacking balance to account for nonlinearities and the cross-products of any predictors that were substantially correlated. 115

6

Chapter 6 To check whether the resulting distributions of the covariates were balanced among the groups receiving each stent, weighted counterparts of the chi-square and t-tests were performed. In most analyses the AneuRx/Excluder stents were treated as the base or comparison stent group and we separately tested whether the distribution of each predictor varied between the AneuRx/Excluder, the Zenith, and Powerlink stent groups. Tests that compared all three groups simultaneously by using chi-square tests involving higher degrees-of-freedom and F-tests for continuous predictors also were conducted as appropriate. We estimated the association between the initial treatment strategy and the rate of in-hospital death, reintervention, perioperative complications, and long-term complications by computing weighted event rates or hazard-ratios. This allowed us to evaluate the statistical significance of the differences by using methods that treat the propensity score weights as design weights as opposed to frequency weights (analogous to survey analysis). Each of the outcomes was treated as an event. For perioperative outcomes, tests of proportions were used to test for simple differences. The single docking limb (AneuRx/Excluder) served as the comparison group for both the other groups. To evaluate the effects of trends over time we added time as a covariate and used logistic regression models with both the treatment indicator and its interaction with time as predictors. In sensitivity analyses we added patient clinical and demographic characteristics as additional predictors to these models. Because censoring is a concern for outcome events evaluated over longerterm follow-up, we used propensity score weighted Cox proportional hazard models to compare the risk of the outcome event between the treatment groups. The proportional hazard assumption is reasonable for comparisons between different stents as their risk profiles tend to track over time, unlike comparisons of interventional (e.g., endovascular repair) procedures versus open surgical repair, which often have vastly different risk profiles over the perioperative period compared to subsequent follow-up. In analyses of time-to-death, the end of follow up was the only source of censoring. In analyses of nonfatal events, death was a censoring event as was subsequent enrollment in a Medicare Advantage health plan. In a sensitivity analysis we also estimated competing risk models of time to a given non-fatal event or death (i.e., whichever occurred first). The estimated effects obtained from Cox proportional hazard models are loghazard ratios. To convert these to fixed-time event probabilities, we estimated a logistic regression model in which the unit of analysis is whether an individual experienced the outcome event in a given month. The probability of the outcome event occurring by month t was computed by multiplying the fitted survival probabilities from the earlier months together, emulating the calculations performed when evaluating Kaplan-Meier tests. This procedure allowed both the event probability and its associated standard error (thus, confidence intervals and p-values) to be evaluated directly from the estimated model. A separate analysis was performed for the 2001-2003 period to demonstrate how 1st generation commercial stent grafts performed and is included in the supplementary Appendix. Although only two stents were compared over this earlier time frame, for consistency and comparability with our primary results 116

Comparison of Endovascular Stent Grafts over the period 2005-2008 we used propensity score weighting as opposed to matching. (Supplemental Appendix)

Results

We studied We studied 46,171 Medicare Beneficiaries from 2005 to 2008, of whom 32,909 (71%) received AneuRx/Excluder, 11,002 (24%) received Zenith and 2,260 (5%) received Powerlink. Before propensity score weighting there were significant differences between the three cohorts, including age, admission type, prior AAA diagnosis, hypertension, diabetes and renal failure. (Table II) After propensity score weighting, there were no significant differences in patients’ baseline demographic and clinical characteristics among the three stent grafts Perioperative Outcomes For 2005 to 2008, patients with a Zenith graft had a significantly higher rates of perioperative mortality (1.7% vs. 1.3%, p=0.001), acute renal failure (6.0% vs. 4.7%, p<.001), mesenteric ischemia (0.7% vs. 0.4%, p=0.001), embolectomy (1.1% vs. 0.8%, p=0.002), MI’s (2.8% vs. 2.2%, p<.001), and pneumonia (4.6% vs. 3.7%, p<.001) compared to patients with an AneuRx or Excluder. (Table III) Additionally, length of stay (3.4 days vs. 3.0 days, p<.001) and readmissions within 30days (12% vs. 10%, p<.001) were higher/greater in patients with a Zenith graft compared to patients with AneuRx/Excluder. Patients with a Zenith graft were also less likely to be discharged home compared to AneuRx/Excluder (94% vs. 95%, p<0.001). Compared to patients with AneuRx/Excluder, the use of the Powerlink graft had an increase in observed mortality with a similar effect size to that of Zenith, however the confidence interval was broad and had less power due to a smaller numbers of Powerlink cases. Powerlink patients had more conversions to open repair (0.7% vs. 0.1%, p<.001), acute renal failure (5.9% vs. 4.7%, p=0.019), mesenteric ischemia (1.0% vs. 0.4%, p<.001), and embolectomy (1.4% vs. 0.8%, p=0.005), compared to AneuRx/Excluder patients. Additionally, length of stay was was greater in the Powerlink group. 4-year Outcomes Mortality at 4-year follow-up remained significantly higher in patients with Zenith than with AneuRx/Excluder (30% vs. 29%, p=0.047). (Table IVa) In contrast, AAA related reinterventions (10% vs. 12%, p=0.003) and conversions to open repair (0.6% vs. 0.9%, p=0.033) were less common in Zenith stent grafts. The majority of reinterventions were minor (10% vs. 11%, p=0.004), and the difference was primarily driven by decreased utilization of extension cuffs (2.9% vs. 4.1%, p<.001) in the Zenith group. At 4 years follow-up, patients with Powerlink had fewer conversions to open repair (0.4% vs. 0.8%, p=0.022), but more AAArelated interventions (15% vs. 12%, p=0.002) compared to AneuRx/Excluder. This was again due to minor reinterventions (14% vs 11%, p<.001), primarily an increased utilization of extension cuffs (7.8% vs. 4.1%, p<.001). (Table IVb) Rupture and overall major reinterventions did not differ between groups at 4-year follow-up. 117

6

118 8123 (24.7) 8541 (26.0) 26684 (81.1)

2007

2008

Male Sex

625 (1.9) 76.9

Other

Mean Age

MI past 6 months

Coexisting Conditions

Prior AAA

452 (1.4)

24258 (73.7)

1316 (4.0)

927 (2.8)

Black

Urgent Admission

31357 (95.3)

White

Race

8393 (25.5)

2006

139 (1.3)

8281 (75.3)

477 (4.3)

77.2

193 (1.8)

304 (2.8)

10505 (95.5)

9297 (84.5)

2929 (26.6)

3101 (28.2)

2672 (24.3)

2300 (20.9)

32909 (%)

N

7852 (23.9)

Bifurcated Double

Bifurcated Single

2005

Year

Variable

34 (1.5)

1689 (74.7)

62 (2.7)

77.0

42 (1.9)

64 (2.8)

2154 (95.3)

1832 (81.1)

803 (35.5)

653 (28.9)

477 (21.1)

327 (14.5)

11002 (%)

Unibody Bifurcated

No Weight

0.562

0.004

0.002

<.001

0.623

0.954

0.693

<.001

<.001

<.001

<.001

<.001

2260 (%)

P value

446 (1.4)

24400 (74.1)

1323 (4.0)

76.9

615 (1.9)

922 (2.8)

31372 (95.3)

26952 (81.9)

8745 (26.6)

8468 (25.7)

8229 (25.0)

7467 (22.7)

Bifurcated Single

150 (1.4)

8161 (74.2)

444 (4.0)

76.9

208 (1.9)

309 (2.8)

10485 (95.3)

9017 (82.0)

2921 (26.5)

2839 (25.8)

2759 (25.1)

2484 (22.6)

32909 (%)

Bifurcated Double

32 (1.4)

1672 (74.0)

92 (4.1)

77.0

43 (1.9)

66 (2.9)

2151 (95.2)

1829 (80.9)

592 (26.2)

580 (25.7)

558 (24.7)

530 (23.5)

11002 (%)

Unibody Bifurcated

Balanced Weights*

0.983

0.983

0.996

0.870

0.988

0.943

0.953

0.544

0.922

0.984

0.939

0.727

2260 (%)

P value

Table II. Baseline Characteristics of Medicare Beneficiaries Undergoing Endovascular Repair (EVAR) of Abdominal Aortic Aneurysms in 2005-2008, before and after Propensity Weighting

Chapter 6

2324 (21.1) 3160 (28.7) 1018 (9.3) 79 (0.7)

6211 (18.9) 4571 (13.9) 22273 (67.7) 6616 (20.1) 9569 (29.1) 2838 (8.6) 213 (0.6)

Peripheral Vascular Disorder

NeuroVascular Disease

Hypertension

Diabetes

Chronic Pulmonary Disease

2256 (20.5) 329 (3.0)

6601 (20.1) 972 (3.0)

1534 (13.9)

2078 (18.9)

1567 (14.2)

64 (2.8)

489 (21.6)

16 (0.7)

232 (10.3)

704 (31.2)

492 (21.8)

1598 (70.7)

326 (14.4)

481 (21.3)

308 (13.6)

227 (10.0)

191 (8.5)

0.920

0.143

0.710

0.007

0.068

0.019

0.004

0.776

0.018

0.736

0.770

0.098

* Defined as 1/(prob(k)) for treatment k.

Obesity

History of Cancer

ESRD

Renal Failure

7552 (68.6)

4633 (14.1)

Congestive Heart Failure

1160 (10.5)

3426 (10.4)

Valvular Disease

813 (7.4)

2382 (7.2)

MI 6-18 months

973 (3.0)

6667 (20.3)

218 (0.7)

2911 (8.8)

9575 (29.1)

6710 (20.4)

22396 (68.1)

4582 (13.9)

6252 (19.0)

4638 (14.1)

3430 (10.4)

2413 (7.3)

325 (3.0)

2228 (20.3)

71 (0.6)

972 (8.8)

3194 (29.0)

2267 (20.6)

7490 (68.1)

1533 (13.9)

2096 (19.1)

1554 (14.1)

1149 (10.4)

810 (7.4)

67 (3.0)

455 (20.1)

19 (0.8)

195 (8.6)

661 (29.3)

454 (20.1)

1546 (68.4)

310 (13.7)

421 (18.6)

326 (14.4)

247 (10.9)

170 (7.5)

1.000

0.989

0.700

0.937

0.976

0.829

0.946

0.958

0.891

0.918

0.779

0.954

Comparison of Endovascular Stent Grafts

6

119

Chapter 6 Table III. Perioperative Outcomes of Medicare Beneficiaries Undergoing Endovascular Repair (EVAR) of Abdominal Aortic Aneurysms in 2005-2008, for Bifurcated Single vs. Bifurcated Double and Bifurcated Single vs. Unibody Bifurcated

Variable

Bifurcated Single N=32909 (%)

Bifurcated P-value Double (Single vs N=11002 Double) (%)

Relative Unibody P-value Risk (95% Bifurcated (Single vs CI) N=2260 (%) Unibody)

Relative Risk (95% CI)

Death (all ages) - Unweighted Cohort

417 (1.3)

188 (1.7)

<.001

1.4 (1.1-1.6)

35 (1.6)

0.250

1.2 (0.9-1.7)

Death (all ages) Weighted Cohort

420 (1.3)

186 (1.7)

0.001

1.3 (1.1-1.6)

37 (1.7)

0.191

1.3 (0.9-1.8)

Myocardial Infarction

716 (2.2)

311 (2.8)

<.001

1.3 (1.1-1.5)

48 (2.1)

0.913

1.0 (0.7-1.3)

Pneumonia

1217 (3.7)

503 (4.6)

<.001

1.2 (1.1-1.4)

82 (3.6)

0.878

1.0 (0.8-1.2)

Acute Renal Failure

1530 (4.7)

655 (6.0)

<.001

1.3 (1.2-1.4)

133 (5.9)

0.019

1.3 (1.1-1.5)

Hemodialysis (New)

169 (0.5)

60 (0.5)

0.716

1.1 (0.8-1.4)

14 (0.6)

0.584

1.2 (0.7-2.1)

Deep-vein Thrombosis

417 (1.3)

164 (1.5)

0.076

1.2 (1.0-1.4)

27 (1.2)

0.802

1.0 (0.7-1.4)

Re-operation for Bleeding

122 (0.4)

48 (0.4)

0.335

1.2 (0.8-1.6)

9 (0.4)

0.957

1.0 (0.5-2.0)

Tracheostomy

43 (0.1)

13 (0.1)

0.827

0.9 (0.5-1.7)

5 (0.2)

0.205

1.9 (0.8-4.5)

Embolectomy

265 (0.8)

125 (1.1)

0.002

1.4 (1.1-1.7)

32 (1.4)

0.005

1.8 (1.2-2.5)

Conversion to Open Repair

32 (0.1)

6 (0.1)

0.149

0.5 (0.2-1.3)

15 (0.7)

<.001

7.0 (3.812.8)

Mesenteric Ischemia

146 (0.4)

77 (0.7)

0.001

1.6 (1.2-2.1)

24 (1.0)

<.001

2.3 (1.5-3.6)

Major Amputation

4 (0)

3 (0)

0.301

2.2 (0.5-9.8)

1 (0)

0.323

2.9 (0.332.0)

Lysis of Adhesions without Resection

5 (0)

4 (0)

0.211

2.3 (0.6-8.5)

1 (0.1)

0.299

3.0 (0.424.4)

Small Bowel Resection

17 (0.1)

12 (0.1)

0.036

2.2 (1.0-4.5)

1 (0.1)

0.843

1.2 (0.2-6.7)

Large Bowel Resection

77 (0.2)

30 (0.3)

0.440

1.2 (0.8-1.8)

9 (0.4)

0.273

1.6 (0.8-3.3)

Ileus or bowel obstruction w/o resection

754 (2.3)

297 (2.7)

0.017

1.2 (1.0-1.3)

59 (2.6)

0.360

1.2 (0.9-1.5)

N = 32655

N = 10877

N = 2234

Discharge Home (all 30882 (94.6) 10168 (93.5) ages)

<.001

1.0 (1.0-1.0) 2120 (94.9)

0.552

1.0 (1.0-1.0)

Readmission in 30 3383 (10.3) days after Discharge

1368 (12.4)

<.001

1.2 (1.1-1.3)

218 (9.6)

0.357

0.9 (0.8-1.1)

Length of Stay, mean 2.98 (±0.02) 3.38 (±0.05) (SE)

<.001

3.31 (±0.14)

0.018

120

Comparison of Endovascular Stent Grafts

Discussion

The The Zenith had higher mortality rates and greater length of stay compared to AneuRx/Excluder, but lower rates of aneurysm-related reinterventions (primarily extension cuffs) and conversion to open. Compared to the AneuRx/Excluder configuration, Powerlink had a non-significant increase in perioperative and late mortality and a significant increase in multiple perioperative complications. During follow-up there were more reinterventions in the Powerlink group, but fewer cases of conversion to open, compared to AneuRx/Excluder. In 1999, the Ancure stent graft and the AneuRx stent graft were the first devices to be approved by the FDA for minimal invasive treatment of AAA. Due to malfunctioning and damage to the artery wall the Ancure was removed from the market in 2003. The AneuRx by Medtronic, which is still available, is a Dacron single docking limb stent graft without fixation appendices. Concerns regarding significant device migration with the AneuRx system26-29 led to a notification issued by the FDA in 2003 stating, “the risk of late AAA-related mortality associated with AneuRx may exceed that associated with open surgery” and “the overall AAA-associated mortality from the AneuRx Stent Graft is likely to cross-over and exceed the AAA-associated mortality from open surgery at some point in time.”30 To address this concern for migration, attributed to the device’s shorter neck, the AneuRx stent graft recommended the use of a 15mm neck length that same year. Stent graft migration may increase the need for extension cuffs and conversion to open repair, which our study confirms for the single docking limb stent graft. This study, however, included both the AneuRx and Excluder in the single docking limb group, as they share the same CPT code, limiting our ability to look at the benefits of the grafts separately. The original Excluder by GORE, introduced in 2002, was associated with postoperative sac expansion13-17 despite an absence of radiographic endoleak, which was thought to be due to material permeability.16, 18-21 To address this concern, a low-permeability Excluder, still made of PTFE with fixation hooks and nitinol exoskeleton, was introduced in 2004 and led to improvement of this problem following EVAR.17,18, 22-25 This improvement allowed an increase in treatment of women and those with peripheral arterial disease, which also may have had an impact on perioperative and late survival, although we carefully match on each of these characteristics in our analyses. Approved in 2003, the Zenith by Cook is a bifurcated graft with 2 docking limbs designed with a low-permeability Dacron and a supra-renal bare stent with hooks to improve fixation and minimize migration. It has been associated with high rates of sac regression35, 36 but was found to have limbs prone to kinking in tortuous and narrow iliac arteries, due to its interrupted design.35, 37-39 To avoid the risk of limb occlusion from kinking a self-expandable stent inside the Zenith stent graft was sometimes deployed.39-41 In 2006, Cook came out with a largeneck version, to allow for treatment of patients with larger aortic neck diameters than was previously possible with existing stent grafts at the time. A potential benefit to Zenith is that it was the first graft with suprarenal stenting to prevent migration. The caveat to this is that it may be chosen for those patients with 121

6

122

8.2%

0.9%

7.8%

0.8%

3.6%

0.2%

0.1%

0.1%

0.1%

3.5%

0.2%

Death

Rupture

Any aneurysm-related Intervention

Major reintervention

Conversion to Open Repair

Axillofemoral or axillobifemoral bypass

Repair infected graft/graftenteric fistula

Minor reintervention

Endovascular AAA repair (redo)

0.1%

3.0%

0.1%

0.1%

0.1%

0.2%

3.2%

10071

30382

Number surviving to End of Year

Variable

Year 2

Year 3

Year 4

0.266

0.008

0.942

0.876

0.108

0.240

0.006

0.534

0.096

0.3%

6.1%

0.1%

0.1%

0.3%

0.4%

6.4%

1.3%

14.5%

20630

0.3%

5.4%

0.1%

0.1%

0.2%

0.3%

5.6%

1.4%

15.2%

6831

0.205

0.003

0.937

0.870

0.048

0.183

0.003

0.494

0.055

0.6%

8.5%

0.1%

0.2%

0.5%

0.7%

8.9%

2.0%

21.6%

12322

0.5%

7.4%

0.1%

0.2%

0.3%

0.5%

7.8%

2.2%

22.6%

4066

0.185

0.003

0.937

0.870

0.033

0.165

0.002

0.481

0.045

0.8%

10.9%

0.2%

0.2%

0.9%

1.1%

11.5%

2.7%

28.8%

5303

0.7%

9.6%

0.2%

0.2%

0.6%

0.9%

10.1%

2.9%

30.1%

1749

0.194

0.004

0.938

0.872

0.033

0.174

0.003

0.494

0.047

Bifurcated Bifurcated Bifurcated Bifurcated Bifurcated Bifurcated Bifurcated Bifurcated P-value P-value P-value P-Value Single Double Single Double Single Double Single Double

Year 1

Table IVa. Late Outcomes after Endovascular Repair for Bifurcated Single vs. Bifurcated Double Stent Grafts

Chapter 6

Rupture or AAA reintervention

AAA related hospital w/out reintervention

Fem-Fem. Bypass

Thrombectomy

Extension Cuff

Aortic Angioplasty

Embolization

1.4%

0.6%

0.7%

0.2%

0.5%

0.1%

3.9%

1.4%

0.7%

1.0%

0.1%

0.4%

0.1%

4.2%

0.040

0.537

0.283

0.202

0.000

0.208

0.642

7.3%

0.2%

0.6%

0.2%

1.9%

1.0%

3.0%

6.7%

0.2%

0.7%

0.3%

1.4%

0.9%

2.9%

0.024

0.457

0.257

0.167

0.000

0.185

0.588

10.0%

0.3%

0.7%

0.2%

2.9%

1.3%

4.4%

9.2%

0.4%

0.8%

0.3%

2.1%

1.1%

4.3%

0.022

0.434

0.254

0.165

0.000

0.183

0.577

12.9%

0.5%

0.8%

0.3%

4.1%

1.5%

5.8%

11.8%

0.6%

1.0%

0.4%

2.9%

1.3%

5.6%

0.027

0.440

0.273

0.185

<.001

0.200

0.589

Comparison of Endovascular Stent Grafts

6

123

124

8.2%

0.8%

7.8%

0.8%

3.6%

0.2%

0.1%

0.1%

0.1%

3.5%

0.2%

Death

Rupture

Any aneurysm-related Intervention

Major reintervention

Conversion to Open Repair

Axillofemoral or axillobifemoral bypass

Repair infected graft/graft-enteric fistula

Minor reintervention

Endovascular AAA repair (redo)

0.3%

4.6%

0.0%

0.1%

0.1%

0.1%

4.7%

2063

30382

Unibody Bifurcated BifurcatSingle ed

Number surviving to End of Year

Variable

Year 1

0.132

0.001

0.000

0.909

0.043

0.075

0.002

0.848

0.285

P-value

0.3%

6.1%

0.1%

0.1%

0.3%

0.4%

6.4%

1.3%

14.5%

20630

0.6%

8.1%

0.0%

0.1%

0.1%

0.2%

8.2%

1.4%

15.2%

1427

Unibody Bifurcated BifurcatSingle ed

Year 2

0.113

0.001

0.000

0.908

0.026

0.062

0.002

0.843

0.263

P-value

0.6%

8.5%

0.1%

0.2%

0.5%

0.7%

8.9%

2.0%

21.6%

12322

1.0%

11.1%

0.0%

0.2%

0.2%

0.4%

11.3%

2.1%

22.6%

857

Unibody Bifurcated BifurcatSingle ed

Year 3

0.107

0.001

0.000

0.907

0.022

0.058

0.001

0.842

0.255

P-value

Year 4

0.8%

10.9%

0.2%

0.2%

0.9%

1.1%

11.5%

2.7%

28.8%

5303

1.4%

14.3%

0.0%

0.2%

0.4%

0.6%

14.7%

2.8%

30.2%

372

Unibody Bifurcated BifurcatSingle ed

Table IVb. Late Outcomes after Endovascular Repair for Bifurcated Single vs. Unibody Bifurcated Stent Grafts

0.109

<.001

<.001

0.908

0.022

0.060

0.002

0.844

0.256

P-Value

Chapter 6

Rupture or AAA reintervention

AAA related hospital w/out reintervention

Fem-Fem. Bypass

Thrombectomy

Extension Cuff

Aortic Angioplasty

Embolization

1.3%

1.4%

2.0%

0.2%

0.3%

0.1%

5.3%

1.4%

0.7%

1.0%

0.1%

0.4%

0.1%

4.2%

0.003

0.607

0.440

0.479

0.000

0.002

0.481

7.3%

0.2%

0.6%

0.2%

1.9%

1.0%

3.0%

9.2%

0.1%

0.4%

0.3%

3.7%

2.0%

2.8%

0.002

0.584

0.433

0.469

0.000

0.002

0.460

10.0%

0.3%

0.7%

0.2%

2.9%

1.3%

4.4%

12.5%

0.2%

0.6%

0.3%

5.6%

2.4%

4.0%

0.002

0.579

0.433

0.468

0.000

0.002

0.456

12.9%

0.5%

0.8%

0.3%

4.1%

1.5%

5.8%

16.1%

0.4%

0.7%

0.4%

7.8%

2.8%

5.3%

0.002

0.580

0.438

0.473

<.001

0.002

0.460

Comparison of Endovascular Stent Grafts

6

125

Chapter 6 a suboptimal neck anatomy, which may increase the technical difficulty of the procedure. Subsequently, this may lead to more perioperative complications, as a result of the challenging anatomy rather than the graft itself. Unfortunately, we do not have information on physician preference for endograft choice to help sort out this question. Also, the Zenith stent graft, with a 21 French size, has a larger profile and is stiffer than other stent grafts, which could make implantation more difficult and may be associated with higher rates of perioperative complications and survival. The Powerlink by Endologix, introduced in 2004, is a unibody bifurcated expandable PTFE stent graft with a metallic endoskeleton.31, 32 It was noted to have access problems in the delivery system,33, 34 but the stent graft has undergone improvement over time.32 The Powerlink’s graft bifurcation sits on the aortic bifurcation to minimize migration; however, the proximal extension to obtain proximal seal is almost uniformly required, which may negate this potential benefit, but our current data did not allow us to adequately look at this question. Supra-renal fixation was subsequently added, potentially increasing the risk of renal failure and mesenteric ischemia. It is also hypothesized that the longer main body and external billowing fabric may decrease type 2 endoleaks, however we did not find lower rates of reintervention, including embolization, in the Power link group, but conversion to open repair was significantly lower compared to the AneuRx/Excluder stent grafts. Our study is limited by the lack of anatomic detail and stent grafts analyzed in this study differ in fabric composition, fixation patterns, porosity, and deployment, which could not be accounted for. These limitations are minimized by several strengths, including the large study size and inclusive population, the incorporation of physician CPT codes, determination of coexisting conditions from prior encounters, and the use of propensity score weighted matching, which adjusted for all measured variables of co-morbidity. However, the weighted analyses cannot account for selection bias related to unmeasured characteristics. This is the largest comparison of stent grafts available in a U.S. population to date, and further, we would like to emphasize that all endovascular devices in our study, from 2005 to 2008, are still in use and are subject to the 2nd or 3rd generation stent grafts.

Conclusion

Despite our inability to distinguish AneuRx from Gore and potential selection bias hampering our early outcome comparison, the Zenith stent graft demonstrated fewer reinterventions, but higher mortality. The Powerlink had more perioperative complications and aneurysm-related reinterventions, but fewer conversions to open repair. Further comparative analyses of different stent grafts are important, as there are likely to be differences in performance.

126

Comparison of Endovascular Stent Grafts

References 1. Drury D, Michaels JA, Jones L, Ayiku L. Systematic review of recent evidence for the safety and efficacy of elective endovascular repair in the management of infrarenal abdominal aortic aneurysm. The British journal of surgery. 2005;92(8):937-46. 2. Schermerhorn ML, O’Malley AJ, Jhaveri A, Cotterill P, Pomposelli F, Landon BE. Endovascular vs. open repair of abdominal aortic aneurysms in the Medicare population. N Engl J Med. 2008;358(5):464-74. 3. Lederle FA, Freischlag JA, Kyriakides TC, Matsumura JS, Padberg FT, Jr., Kohler TR, et al. Long-term comparison of endovascular and open repair of abdominal aortic aneurysm. N Engl J Med. 2012;367(21):1988-97. 4. De Bruin JL, Baas AF, Buth J, Prinssen M, Verhoeven EL, Cuypers PW, et al. Long-term outcome of open or endovascular repair of abdominal aortic aneurysm. N Engl J Med. 2010;362(20):1881-9. 5. Brown LC, Powell JT, Thompson SG, Epstein DM, Sculpher MJ, Greenhalgh RM. The UK EndoVascular Aneurysm Repair (EVAR) trials: randomised trials of EVAR versus standard therapy. Health technology assessment. 2012;16(9):1-218. 6. White GH, May J, McGahan T, Yu W, Waugh RC, Stephen MS, et al. Historic control comparison of outcome for matched groups of patients undergoing endoluminal versus open repair of abdominal aortic aneurysms. J Vasc Surg. 1996;23(2):201-11; discussion 11-2. 7. Zarins CK, White RA, Schwarten D, Kinney E, Diethrich EB, Hodgson KJ, et al. AneuRx stent graft versus open surgical repair of abdominal aortic aneurysms: multicenter prospective clinical trial. J Vasc Surg. 1999;29(2):292305; discussion 6-8. 8. van Marrewijk CJ, Leurs LJ, Vallabhaneni SR, Harris PL, Buth J, Laheij RJ, et al. Risk-adjusted outcome analysis of endovascular abdominal aortic aneurysm repair in a large population: how do stent-grafts compare? J Endovasc Ther. 2005;12(4):417-29. 9. Tonnessen BH, Sternbergh WC, 3rd, Money SR. Mid- and long-term device migration after endovascular abdominal aortic aneurysm repair: a comparison of AneuRx and Zenith endografts. J Vasc Surg. 2005;42(3):392400; discussion -1. 10. Broker HS, Foteh KI, Murphy EH, Davis CM, Clagett GP, Modrall JG, et al. Device-specific aneurysm sac morphology after endovascular aneurysm repair: evaluation of contemporary graft materials. J Vasc Surg. 2008;47(4):702-6; discussion 7. 11. Ouriel K, Clair DG, Greenberg RK, Lyden SP, O’Hara PJ, Sarac TP, et al. Endovascular repair of abdominal aortic aneurysms: device-specific outcome. J Vasc Surg. 2003;37(5):991-8. 12. Sheehan MK, Ouriel K, Greenberg R, McCann R, Murphy M, Fillinger M, et al. Are type II endoleaks after endovascular aneurysm repair endograft dependent? J Vasc Surg. 2006;43(4):657-61. 127

6

Chapter 6 13. Bertges DJ, Chow K, Wyers MC, Landsittel D, Frydrych AV, Stavropoulos W, et al. Abdominal aortic aneurysm size regression after endovascular repair is endograft dependent. J Vasc Surg. 2003;37(4):716-23. 14. Cho JS, Dillavou ED, Rhee RY, Makaroun MS. Late abdominal aortic aneurysm enlargement after endovascular repair with the Excluder device. J Vasc Surg. 2004;39(6):1236-41; discussion 2141-2. 15. van der Laan MJ, Prinssen M, Bertges D, Makaroun MS, Blankensteijn JD. Does the type of endograft affect AAA volume change after endovascular aneurysm repair? J Endovasc Ther. 2003;10(3):406-10. 16. Fillinger M, Excluder Bifurcated Endoprosthesis Clinical I. Three-dimensional analysis of enlarging aneurysms after endovascular abdominal aortic aneurysm repair in the Gore Excluder Pivotal clinical trial. J Vasc Surg. 2006;43(5):888-95. 17. Tanski W, 3rd, Fillinger M. Outcomes of original and low-permeability Gore Excluder endoprosthesis for endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2007;45(2):243-9. 18. Kong LS, MacMillan D, Kasirajan K, Milner R, Dodson TF, Salam AA, et al. Secondary conversion of the Gore Excluder to operative abdominal aortic aneurysm repair. J Vasc Surg. 2005;42(4):631-8. 19. Matsumura JS, Brewster DC, Makaroun MS, Naftel DC. A multicenter controlled clinical trial of open versus endovascular treatment of abdominal aortic aneurysm. J Vasc Surg. 2003;37(2):262-71. 20. Mennander A, Pimenoff G, Heikkinen M, Partio T, Zeitlin R, Salenius JP. Nonoperative approach to endotension. J Vasc Surg. 2005;42(2):194-9. 21. Risberg B, Delle M, Eriksson E, Klingenstierna H, Lonn L. Aneurysm sac hygroma: a cause of endotension. J Endovasc Ther. 2001;8(5):447-53. 22. Bush RL, Najibi S, Lin PH, Weiss VJ, MacDonald MJ, Redd DC, et al. Early experience with the bifurcated Excluder endoprosthesis for treatment of the abdominal aortic aneurysm. J Vasc Surg. 2001;34(3):497-502. 23. Bastos Goncalves F, Jairam A, Voute MT, Moelker AD, Rouwet EV, ten Raa S, et al. Clinical outcome and morphologic analysis after endovascular aneurysm repair using the Excluder endograft. J Vasc Surg. 2012;56(4):9208. 24. Curci JA, Fillinger MF, Naslund TC, Rubin BG, Excluder Bifurcated Endoprosthesis I. Clinical trial results of a modified gore excluder endograft: comparison with open repair and original device design. Annals of vascular surgery. 2007;21(3):328-38. 25. Kibbe MR, Matsumura JS, Excluder I. The Gore Excluder US multi-center trial: analysis of adverse events at 2 years. Seminars in vascular surgery. 2003;16(2):144-50. 26. Conners MS, 3rd, Sternbergh WC, 3rd, Carter G, Tonnessen BH, Yoselevitz M, Money SR. Endograft migration one to four years after endovascular abdominal aortic aneurysm repair with the AneuRx device: a cautionary note. J Vasc Surg. 2002;36(3):476-84. 27. Sampaio SM, Panneton JM, Mozes G, Andrews JC, Noel AA, Kalra M, et al. AneuRx device migration: incidence, risk factors, and consequences. Annals 128

Comparison of Endovascular Stent Grafts of vascular surgery. 2005;19(2):178-85. 28. Cao P, Verzini F, Zannetti S, De Rango P, Parlani G, Lupattelli L, et al. Device migration after endoluminal abdominal aortic aneurysm repair: analysis of 113 cases with a minimum follow-up period of 2 years. J Vasc Surg. 2002;35(2):229-35. 29. Zarins CK, Bloch DA, Crabtree T, Matsumoto AH, White RA, Fogarty TJ. Stent graft migration after endovascular aneurysm repair: importance of proximal fixation. J Vasc Surg. 2003;38(6):1264-72; discussion 72. 30. Feigal Jr D. FDA Public Health Notification: Updated Data on Mortality Associated with Medtronic AVE AneuRx速 Stent Graft System. December 2003; http://www.fda.gov/MedicalDevices/Safety/AlertsandNotices/ PublicHealthNotifications/ucm062158.htm%5D. 31. Carpenter JP, Endologix I. Midterm results of the multicenter trial of the powerlink bifurcated system for endovascular aortic aneurysm repair. J Vasc Surg. 2004;40(5):849-59. 32. Wang GJ, Carpenter JP, Endologix I. The Powerlink system for endovascular abdominal aortic aneurysm repair: six-year results. J Vasc Surg. 2008;48(3):535-45. 33. Nano G, Dalainas I, Volpe P, Casana R, Lupattelli T, Paroni G, et al. A variant deployment technique for the powerlink bifurcated endograft. J Endovasc Ther. 2005;12(6):638-41. 34. Qu L, Raithel D. Experience with the Endologix Powerlink endograft in endovascular repair of abdominal aortic aneurysms with short and angulated necks. Perspectives in vascular surgery and endovascular therapy. 2008;20(2):158-66. 35. Greenberg RK, Chuter TA, Sternbergh WC, 3rd, Fearnot NE, Zenith I. Zenith AAA endovascular graft: intermediate-term results of the US multicenter trial. J Vasc Surg. 2004;39(6):1209-18. 36. Sternbergh WC, 3rd, Money SR, Greenberg RK, Chuter TA, Zenith I. Influence of endograft oversizing on device migration, endoleak, aneurysm shrinkage, and aortic neck dilation: results from the Zenith Multicenter Trial. J Vasc Surg. 2004;39(1):20-6. 37. Alric P, Hinchliffe RJ, MacSweeney ST, Wenham PW, Whitaker SC, Hopkinson BR. The Zenith aortic stent-graft: a 5-year single-center experience. J Endovasc Ther. 2002;9(6):719-28. 38. Baum RA, Shetty SK, Carpenter JP, Soulen MC, Velazquez OC, ShlanskyGoldberg RD, et al. Limb kinking in supported and unsupported abdominal aortic stent-grafts. Journal of vascular and interventional radiology : JVIR. 2000;11(9):1165-71. 39. Sivamurthy N, Schneider DB, Reilly LM, Rapp JH, Skovobogatyy H, Chuter TA. Adjunctive primary stenting of Zenith endograft limbs during endovascular abdominal aortic aneurysm repair: implications for limb patency. J Vasc Surg. 2006;43(4):662-70. 40. Bos WT, Tielliu IF, Zeebregts CJ, Prins TR, van den Dungen JJ, Verhoeven EL. Results of endovascular abdominal aortic aneurysm repair with the Zenith stent-graft. European journal of vascular and endovascular surgery : the 129

6

Chapter 6 official journal of the European Society for Vascular Surgery. 2008;36(6):65360. 41. Oshin OA, Fisher RK, Williams LA, Brennan JA, Gilling-Smith GL, Vallabhaneni SR, et al. Adjunctive iliac stents reduce the risk of stent-graft limb occlusion following endovascular aneurysm repair with the Zenith stentgraft. J Endovasc Ther. 2010;17(1):108-14.

130

Comparison of Endovascular Stent Grafts

Supplemental Appendix

We identified 17,606 Medicare Beneficiaries from 2001 to 2003, of which 13,089 (74%) received AneuRx/Excluder and 4,517 (26%) received Powerlink. After propensity weighting, there were no significant differences in patients’ baseline demographics between the three stent grafts. (Supplemental Table I) From 2001 to 2003, perioperative mortality was 1.6% for AneuRx/Excluder and 2.2% for Powerlink (p=0.024). (Supplemental Table II) Patients with the AneuRx/ Excluder stent grafts had fewer MI’s (2.6 vs. 3.6, p=0.003) and fewer bowel obstructions (2.4% vs. 3.2%, p=0.010). There were no significant differences for pneumonia, readmission within 30 days or length of stay between groups. At 8 years of follow-up, patients with bifurcated single stent grafts showed a higher rupture rate (5.7% vs. 4.2%, p=0.007) and underwent more endovascular redo’s (4.1% vs. 2.2%, p<.001), primarily extension cuffs (9.0% vs. 6.4%, p=<.001). (Supplemental Table III)

6 131

132 13089 (%)

Variable

N

4119 (31.5) 5593 (42.7) 10992 (84.0)

2002

2003

Male Sex

321 (2.5) 200 (1.5)

Black

Other

1455 (11.1) 3576 (27.3) 4053 (31.0) 2751 (21.0)

67-69

70-74

75-79

80-84

Age

12568 (96.0)

White

Race

3377 (25.8)

2001

Year

Bifurcated Single

933 (20.7)

1340 (29.7)

1269 (28.1)

584 (12.9)

63 (1.4)

96 (2.1)

4358 (96.5)

3864 (85.5)

997 (22.1)

2005 (44.4)

1515 (33.5)

4517 (%)

Unibody Bifurcated

No Weight

0.606

0.102

0.316

0.001

0.524

0.213

0.166

0.013

<.001

<.001

<.001

P Value

2741 (20.9)

4007 (30.6)

3602 (27.5)

1515 (11.6)

195 (1.5)

311 (2.4)

952 (21.1)

1377 (30.5)

1245 (27.6)

521 (11.5)

63 (1.4)

113 (2.5)

4340 (96.1)

3824 (84.6)

11048 (84.4) 12583 (96.1)

1693 (37.5)

1573 (34.8)

1251 (27.7)

4517 (%)

Unibody Bifurcated

4900 (37.4)

4553 (34.8)

3635 (27.8)

13089 (%)

Bifurcated Single

Balanced Weights*

Supplemental Table I. Two Graft Methods, years 2001 - 2003

0.848

0.881

0.958

0.943

0.690

0.903

0.651

0.718

0.968

0.956

0.914

P Value

Chapter 6

2366 (18.1) 1702 (13.0) 7833 (59.8) 2166 (16.5) 3718 (28.4)

Peripheral Vascular Disorder

NeuroVascular Disease

Hypertension

Diabetes

Chronic Pulmonary Disease

279 (2.1)

2718 (20.8)

* Defined as 1/(prob(k)) for treatment k.

Obesity

History of Cancer

78 (0.6)

2034 (15.5)

Congestive Heart Failure

ESRD

1301 (9.9)

Valvular Disease

645 (4.9)

1113 (8.5)

MI 6-18 months

Renal Failure

248 (1.9)

10129 (77.4)

MI past 6 months

Coexisting Conditions

Prior AAA

367 (2.8)

76.6

Mean Age

Urgent Admission

1254 (9.6)

85+

113 (2.5)

928 (20.5)

15 (0.3)

187 (4.1)

1219 (27.0)

702 (15.5)

2603 (57.6)

588 (13.0)

756 (16.7)

644 (14.3)

431 (9.5)

366 (8.1)

72 (1.6)

3487 (77.2)

127 (2.8)

76.3

391 (8.7)

0.146

0.752

0.035

0.031

0.067

0.114

0.009

0.98

0.042

0.038

0.439

0.403

0.192

0.794

0.978

0.066

291 (2.2)

2703 (20.7)

69 (0.5)

618 (4.7)

3669 (28.0)

2130 (16.3)

7758 (59.3)

1697 (13.0)

2321 (17.7)

1991 (15.2)

1287 (9.8)

1100 (8.4)

99 (2.2)

924 (20.5)

22 (0.5)

210 (4.7)

1262 (27.9)

720 (15.9)

2676 (59.2)

568 (12.6)

800 (17.7)

681 (15.1)

442 (9.8)

381 (8.4)

87 (1.9)

3477 (77.0)

10119 (77.3)

239 (1.8)

124 (2.8)

76.6

421 (9.3)

367 (2.8)

76.6

1224 (9.3)

0.926

0.793

0.749

0.865

0.913

0.633

0.967

0.512

0.981

0.846

0.927

0.942

0.717

0.67

0.879

0.973

Comparison of Endovascular Stent Grafts

6

133

Chapter 6 Supplemental Table II. Perioperative Outcomes of Medicare Beneficiaries Undergoing Endovascular Repair (EVAR) of Abdominal Aortic Aneurysms in 2001-2003, for Bifurcated Single vs. Unibody Bifurcated Variable

Bifurcated Single

Unibody Bifucated

P-value

Relative Risk (95% CI)

N

13089 (%)

4517 (%)

Death (all ages) - Unweighted Cohort

214 (1.6)

96 (2.1)

0.031

1.3 (1.0-1.7)

Death (all ages) - Weighted Cohort

212 (1.6)

97 (2.2)

0.024

1.3 (1.1-1.7)

Death (67-69)

8 (0.5)

6 (1.1)

0.093

2.3 (0.8-6.6)

Death (70-74)

34 (0.9)

20 (1.6)

0.057

1.8 (1.0-3.0)

Death (75-79)

65 (1.6)

28 (2.0)

0.343

1.2 (0.8-1.9)

Death (80-84)

68 (2.5)

30 (3.2)

0.317

1.3 (0.8-1.9)

Death (85+)

37 (3.1)

13 (3.2)

0.891

1.1 (0.6-1.9)

Myocardial Infarction

345 (2.6)

161 (3.6)

0.003

1.4 (1.1-1.6)

Pneumonia

451 (3.5)

158 (3.5)

0.905

1.0 (0.9-1.2)

Acute Renal Failure

464 (3.5)

176 (3.9)

0.310

1.1 (0.9-1.3)

Hemodialysis (New)

58 (0.4)

21 (0.5)

0.930

1.0 (0.6-1.7)

Deep-vein Thrombosis

66 (0.5)

24 (0.5)

0.785

1.1 (0.7-1.7)

Re-operation for Bleeding

96 (0.7)

34 (0.8)

0.857

1.0 (0.7-1.5)

Tracheostomy

26 (0.2)

7 (0.2)

0.591

0.8 (0.3-1.8)

Embolectomy

136 (1.0)

63 (1.4)

0.063

1.3 (1.0-1.8)

Conversion to Open Repair

70 (0.5)

35 (0.8)

0.073

1.5 (1.0-2.2)

Mesenteric Ischemia

72 (0.6)

32 (0.7)

0.277

1.3 (0.9-1.9)

Major Amputation

6 (0.1)

3 (0.1)

0.750

1.3 (0.3-5.5)

Lysis of Adhesions without Resection

5 (0%)

4 (0.1)

0.273

2.1 (0.5-8.2)

Small Bowel Resection

15 (0.1)

5 (0.1)

0.930

1.0 (0.4-2.6)

Large Bowel Resection

35 (0.3)

22 (0.5)

0.037

1.8 (1.1-3.1)

Ileus or bowel obstruction w/o resection

315 (2.4)

142 (3.2)

0.010

1.3 (1.1-1.6)

12359 (95.4)

4214 (94.8)

0.092

1.0 (1.0-1.0)

Discharge Home (67-69)

1485 (98.2)

510 (98.6)

0.499

1.0 (1.0-1.0)

Discharge Home (70-74)

3488 (97.5) 1198 (97.3)

0.781

1.0 (1.0-1.0)

Discharge Home (75-79)

3816 (96.2) 1301 (96.1)

0.891

1.0 (1.0-1.0)

Discharge Home (all ages)

134

Comparison of Endovascular Stent Grafts

Discharge Home (80-84)

2525 (93.9)

850 (91.2)

0.012

1.0 (1.0-1.0)

Discharge Home (85+)

1045 (87.0)

355 (85.9)

0.615

1.0 (0.9-1.0)

Readmission in 30 days after Discharge

1423 (10.9)

521 (11.5)

0.237

1.1 (1.0-1.2)

3.36 (±0.05) 3.54 (±0.08)

0.052

Length of Stay, mean (SE)

6 135

136

8.3%

0.7%

8.6%

0.9%

4.5%

0.3%

0.1%

0.1%

0.0%

4.3%

0.3%

Death

AAA Rupture

Any aneurysm-related Intervention

Major reintervention

Conversion to Open Repair

Axillofemoral or axillobifemoral bypass

Repair infected graft/graft-enteric fistula

Minor reintervention

Endo AAA repair (redo)

0.2%

4.1%

0.0%

0.2%

0.1%

0.3%

4.3%

4112

11994

Unibody Bifurcated BifurcatSingle ed

Number surviving to End of Year

Variable

Year 1

0.039

0.426

0.939

0.227

0.833

0.873

0.435

0.036

0.301

P-value

0.5%

7.0%

0.1%

0.1%

0.3%

0.5%

7.3%

1.5%

15.9%

11024

0.3%

6.6%

0.1%

0.2%

0.3%

0.5%

7.0%

1.1%

15.2%

3823

Unibody Bifurcated BifurcatSingle ed

Year 2

0.011

0.363

0.931

0.205

0.777

0.847

0.372

0.015

0.195

P-value

1.6%

13.1%

0.2%

0.2%

1.0%

1.2%

13.8%

3.1%

38.0%

8119

0.9%

12.5%

0.2%

0.3%

0.9%

1.2%

13.2%

2.3%

36.8%

2864

Unibody Bifurcated BifurcatSingle ed

Year 5

<.001

0.308

0.916

0.178

0.714

0.813

0.316

0.005

0.097

P-value

3.2%

18.1%

0.3%

0.2%

1.9%

2.2%

19.2%

5.7%

58.6%

1448

1.9%

17.2%

0.3%

0.4%

1.8%

2.3%

18.4%

4.2%

57.0%

544

Unibody Bifurcated BifurcatSingle ed

Year 8

Supplemental Table III. Late Outcomes after Endovascular Repair for Bifurcated Single vs. Unibody Bifurcated Stent Grafts

<.001

0.331

0.917

0.185

0.726

0.821

0.339

0.007

0.091

P-value

Chapter 6

Rupture or AAA reintervention

AAA related hospital w/out reintervention

Fem-Fem. Bypass

Thrombectomy

Endo Extension Cuff

Aortic Angioplasty

Embolization

2.4%

1.2%

0.7%

0.1%

0.5%

0.2%

4.9%

2.3%

0.9%

1.0%

0.1%

0.3%

0.1%

5.3%

0.192

0.816

0.043

0.694

0.004

0.023

0.494

8.5%

0.3%

0.5%

0.2%

1.9%

1.2%

3.7%

7.9%

0.3%

0.8%

0.1%

1.3%

1.7%

4.0%

0.137

0.782

0.023

0.670

<.001

0.014

0.434

15.8%

0.6%

0.8%

0.4%

5.6%

1.7%

6.2%

14.7%

0.7%

1.3%

0.3%

3.9%

2.4%

6.6%

0.096

0.736

0.013

0.617

<.001

0.009

0.392

22.1%

1.1%

1.1%

0.5%

9.6%

2.2%

8.2%

20.6%

1.2%

1.8%

0.4%

6.7%

3.1%

8.7%

0.114

0.737

0.020

0.629

<.001

0.014

0.414

Comparison of Endovascular Stent Grafts

6

137

CHAPTER 7

Percutaneous versus Femoral Cutdown Access for Endovascular Aneurysm Repair Dominique B. Buck, Eleonora G. Karthaus, Peter A. Soden, Klaas H.J. Ultee, Joost A. van Herwaarden, Frans L. Moll, Marc L. Schermerhorn. J Vasc Surg. 2015 Mar 28.

Chapter 7

Abstract Objective Prior studies suggest that percutaneous access for endovascular abdominal aortic aneurysm repair (pEVAR) offers significant operative and post-operative benefits compared to femoral cutdown (cEVAR). National data on this topic, however, are limited. We compared patient selection and outcomes for elective pEVAR and cEVAR. Methods We identified all patients undergoing either pEVAR (bilateral percutaneous access whether successful or not) or cEVAR (at least one planned groin cutdown) for abdominal aortic aneurysms (AAA), from January 2011 to December 2013 in the Targeted Vascular dataset from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database. Emergent cases, ruptures, cases with an iliac conduit, and cases with a preoperative wound infection were excluded. Groups were compared using chi-square test or t-test or the Mann-Whitney test where appropriate. Results 4112 patients undergoing elective EVAR were identified; 3004 cEVAR (73%) and 1108 pEVAR (27%). Of all EVAR patients 26% had bilateral percutaneous access, 1.0% had attempted percutaneous access converted to cutdown (4% of pEVARs), while the remainder had a planned cutdown, 63.9% bilateral, and 9.1% unilateral. There were no significant differences in age, gender, aneurysm diameter or prior open abdominal surgery. Patients undergoing cEVAR were less likely to have congestive heart failure (1.5% vs. 2.4%, P=0.04) but more likely to undergo any concomitant procedure during surgery (32% vs. 26%, P<.01) than patients undergoing pEVAR. Postoperatively, pEVAR patients had shorter operative time (mean 135 vs. 152 minutes, P<.01), shorter length of stay (median 1 day vs. 2 days, P<.01), and fewer wound complications (2.1% vs. 1.0%, P=0.02). On multivariable analysis the only predictor of percutaneous access failure was performance of any concomitant procedure (OR 2.0, 95% CI 1.0-4.0, P=0.04). Conclusions Currently, 1 in 4 patients treated at Targeted Vascular NSQIP centers are getting pEVAR, which is associated with a high success rate, shorter operation time, shorter length of stay, and fewer wound complications compared to cEVAR.

140

Percutaneous versus Femoral Cutdown Access

Introduction

For patients with an anatomically suitable abdominal aortic aneurysm (AAA), endovascular aortic aneurysm repair (EVAR) has become the preferred choice of treatment during the past decade.1 Percutaneous access (pEVAR) further minimizes invasiveness compared to femoral cutdown access (cEVAR). A recently published American multicenter randomized trial with 151 patients in centers of excellence with one stent graft, reported high success rates in selected pEVAR patients when compared to cEVAR.2 Several small single center studies using a variety of grafts showed a reduction in total operative time2-8 and length of hospital stay.3, 6, 9, 10 Additionally, access-related complication rates were lower with pEVAR compared to cEVAR.2, 4, 6-12 Despite these promising results, the possibility of publication bias should be considered. Therefore a larger scale study, of contemporary management of AAA comparing pEVAR and cEVAR, is needed to see if the results from the prior RCT and single centers may be generalizable. We analyzed national outcomes of pEVAR and cEVAR for AAA repair. We aimed to analyze patient selection, anatomic variation and outcomes for elective pEVAR and cEVAR.

Methods Data Source We identified all patients undergoing either pEVAR (bilateral percutaneous access whether successful or not) or cEVAR (at least one planned groin cutdown) for abdominal aortic aneurysms (AAA), from January 2011 to December 2013 in the Targeted Vascular dataset from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database. This is a multiinstitutional, risk-adjusted database with 83 participating hospitals in the United States, which collects prospective clinical data of patients undergoing vascular surgery. Data are recorded on preoperative, operative and postoperative patient-level variables after the index procedure. All data collection is performed by trained clinical nurse reviewers to ensure quality. These variables being collected were chosen by vascular surgeons and specific to the index operation e.g. AAA diameter, indication for surgery and attempt at percutaneous access. Variables definitions and details of data collection are available on the ASC NSQIP website.13 NSQIP does not identify the site of surgery in any way thus precluding volume -outcome analyses as well as outcomes comparison between sites. Emergent cases and ruptures were excluded. Cases with an iliac conduit or with a preoperative open or infected wound were also excluded. As this study contained only de-identified data without any protected health information, the study is not considered human research and therefore is not subject to Institutional Review Board approval or patient consent. Clinical and outcome variables Data were collected on relevant patient demographics, including gender, age, 141

7

Chapter 7 history of prior abdominal operations, ASA (American Society of Anesthesia) classification, and aneurysm diameter. Intra-operative data were compared, including indication for surgery, anatomical details, graft type and operative time. To rule out the effect of additional interventions on the mean total operation time, we excluded patients who had a concomitant intervention or a fenestrated graft for the analyses of operative time. Post-operative outcomes were also compared, including death, rupture, bleeding requiring transfusion, reintubation, return to the operating room, surgical site infections (SSI), any wound complications and overall length of stay. Multivariable logistic regression was used to determine independent predictors of percutaneous access failure, adjusted for potential confounders. Multivariable logistic and linear regressions were used to identify predictors of operative time and length of stay. Additionally, we compared patients with attempted pEVAR converted to cutdown to those with successful pEVAR to identify possible associations with failure. To check for homogeneity within cEVAR group we compared the patients with bilateral groin cutdown to one groin cutdown within the cEVAR cohort. For obesity we used the cutoff BMI>30 kg/m2. Any wound complication includes wound dehiscence and superficial, deep and organ space surgical site infection (SSI). Statistical analyses Categorical variables were compared using chi-square test and continuous variables were compared using the Student t-test or the Mann-Whitney test where appropriate. Statistical significance was defined as P<0.05. All statistical analyses were performed using SPSS statistical software (version 20; IBM Corp, Armonk, NY).

Results

We identified a total of 4479 patients who underwent elective EVAR, of which 4112 patients remained after excluding iliac conduits (N=367, 8.2%) and prior wound infections (N=39, 0.9%). There were 3004 (73%) cEVAR and 1108 (27%) pEVAR patients. Of all cEVAR patients 88% had bilateral groin cutdown and 12% had unilateral groin cutdown. Of all pEVAR patients, 96.% had a successful bilateral percutaneous access and 4% had attempted percutaneous access converted to cutdown. pEVAR was performed in 20% in 2011, 27% in 2012, and 29% in 2013. The majority of procedures were performed by vascular surgeons (98% of total patients). Baseline characteristics Gender, age, aneurysm diameter and prior open abdominal surgery were similar for both access types. (Table I.) There was a significant difference in race/ ethnicity between the two groups, most likely due to more Asian and less white patients in the pEVAR group. More patients with congestive heart failure (2.4% vs. 1.5%, P=0.04) were seen in the pEVAR group. 142

Percutaneous versus Femoral Cutdown Access Table I: Baseline Characteristics of Patients with Abdominal Aortic Aneurysms undergoing EVAR (Percutaneous vs. Cutdown)

Variable Male gender

Percutaneous (n=1108)

Cutdown (n=3004)

P Value

Total Cohort (n=4112)

81%

81%

0.10

82%

<.01

Race or ethnic group * Other/Unknown

5.1%

6.8%

6.3%

White

84%

87%

86%

Black

7.0%

4.1%

4.9%

0%

0.1%

0.1%

Native Hawaiian/Pacific Islander

0.2%

0%

0.1%

Asian

4.2%

1.7%

2.4%

Age (mean)

74

74

American Indian/Alaska Native

0.08

74.1

0.10

Age Category (years) 18-59

6.1%

5.1%

5.3%

60-69

25%

23%

23%

70-79

42%

42%

42%

80-89

25%

29%

28%

90+

1.7%

1.5%

1.6%

Prior Open Abdominal Surgery

22%

25%

0.07

24%

ASA 4 Classification

21%

22%

0.23

22%

5.7

5.7

0.82

5.7

Congestive heart failure

2.4%

1.5%

0.04

1.8%

Hypertension

81%

81%

0.74

81%

Diabetes

16%

16%

0.82

16%

History of severe COPD

16%

19%

0.08

18%

Dialysis (pre-op)

1.4%

1.2%

0.47

1.2%

Obesity

30%

33%

0.07

32%

Aneurysm diameter (cm) Coexisting conditions

Outcomes Patients undergoing cEVAR had more concomitant procedures (32% vs. 26%, P<.01), however, hypogastric embolizations (7.6% vs. 5.8%, P=0.04) were more 143

7

Chapter 7 Table II: Intra-operative Outcomes of Patients with Abdominal Aortic Aneurysm undergoing EVAR (percutaneous vs. cutdown)

Outcomes

Percutaneous (n=1108)

Cutdown (n=3004)

P Value

Total Cohort (n=4112)

Acute conversion to open AAA repair

0.5%

0.4%

0.56

0.4%

Cutdown access Attempted percutaneous access converted to cutdown

3.6%

1%

Bilateral groin cutdown

88%

64%

One groin cutdown

12%

9%

Percutaneous Bilateral

96%

26% 0.09

Indication for surgery Diameter

88%

86%

87%

Dissection

0.2%

0.2%

0.2%

Embolization

0.3%

0.6%

0.5%

Non-ruptured symptomatic

4.1%

6.0%

5.5%

Prior endovascular intervention w/ unsatisfactory result

3.5%

2.5%

2.7%

Prior open intervention w/ unsatisfactory result

0.3%

0.5%

0.4%

Rupture with or without hypotension

0.6%

1.7%

1.7%

Thrombosis

0.6%

1.7%

1.7%

Not documented

1.5%

2.3%

2.1% <.01

Main Body Device Cook Zenith

16%

21%

20%

Cook Zenith Fenestrated

1.2%

2.9%

2.4%

Cook Zenith Renu

0.5%

1.3%

1.1%

Endologix Powerlink

9.8%

6.1%

7.1%

Gore Excluder

35%

32%

33%

Lombard Aorfix

0%

0.1%

0%

Medtronic AnueRx

0.4%

0.2%

0.2%

Medtronic Endurant

27%

31%

30%

Medtronic TALENT

0.1%

0.6%

0.5%

144

Percutaneous versus Femoral Cutdown Access

Not documented

0.7%

1.2%

1.1%

Other

7.9%

3.0%

4.3%

Trivascular Ovation

1.7%

0.7%

1.0% 0.07

Aneurysm Extent Infrarenal

85%

84%

84%

Juxtarenal

3.4%

4.8%

4.4%

Pararenal

2.0%

1.5%

1.6%

Suprarenal

3.8%

2.9%

3.1%

Type IV

0.4%

0.5%

0.5%

Not documented

5.3%

6.4%

6.3% <.01

Distal Aneurysm Extent Aortic

37%

38%

38%

Common Iliac

37%

33%

34%

External Iliac

5.2%

5.0%

5.1%

Internal Iliac

5.0%

8.0%

7.2%

Not documented

17%

16%

16%

Mean OR time (min)

135

152

<.01

148

Median OR time (min)

117

137

<.01

132

Mean OR time (all concomitant procedures excluded)

123

140

<.01

136

Median OR time (all concomitant procedures excluded)

111

128

<.01

122

Renal Stent

5.9%

7.1%

0.17

6.8%

Hypogastric Embolization

7.6%

5.8%

0.04

6.3%

Hypogastric Revascularization

3.2%

4.6%

0.05

4.2%

Lower Extremity Revascularization

2.0%

4.7%

<.01

3.9%

Iliac Branched Device

11%

14%

0.01

13%

Aortic (Bare metal) Stent

1.3%

2.5%

0.02

2.1%

Iliac (Bare metal) Stent

3.4%

3.4%

0.92

3.4%

Any Concomitant Procedure

26%

32%

<.01

30%

145

7

Chapter 7 Table III: Post-operative Outcomes of Patients with Abdominal Aortic Aneurysms undergoing EVAR (percutaneous vs. cutdown) Percutaneous (n=1108)

Cutdown (n=3004)

P Value

Death at 30 days (% of patients)

1.7%

1.4%

0.51

Death in hospital

1.1%

0.6%

0.14

Rupture

0.3%

0.1%

0.10

Myocardial Infarction

0.7%

1.3%

0.12

Cardiac Arrest requiring CPR

0.5%

0.5%

0.76

Pneumonia

0.8%

0.7%

0.62

Pulmonary Embolism

0.2%

0.2%

0.75

Progressive Renal Insufficiency

0.1%

0.4%

0.09

Acute Renal Failure

0.4%

0.5%

0.56

Urinary Tract Infection

1.2%

1.3%

0.81

Stroke

0.1%

0.4%

0.15

Deep Vein Thrombosis

0.9%

0.4%

0.05

Transfusion

9.3%

10.9%

0.13

Sepsis

0.4%

0.5%

0.56

Prolonged Intubation (> 48 hours)

0.8%

0.8%

0.88

Reintubation

1.4%

1.3%

0.65

Return to OR

3.4%

4.0%

0.43

Ischeamic colitis

0.5%

0.4%

0.65

Lower Extremity Ischemia

1.6%

1.2%

0.25

Any Wound Complication

1.0%

2.1%

0.02

Wound dehiscence

0%

0.1%

0.22

1.0%

2.0%

0.03

2.8

3.0

0.48

1

2

<.01

LOS > 2 days

26%

32%

<.01

LOS > 1 day

46%

56%

<.01

Discharged home

93%

93%

0.52

Outcomes

Medical complications (% of patients)

Surgical complications (% of patients)

Surgical Site Infection Mean length of hospital stay (no. of days) Median length of hospital stay (no. of days)

146

Percutaneous versus Femoral Cutdown Access Table IV: Baseline Characteristics of Patients with Abdominal Aortic Aneurysms undergoing EVAR (Percutaneous failure vs. percutaneous bilateral access) Variable

Percutaneous Percutaneous failed (N=40) (N=1068)

P Value

Male gender

73%

84%

0.08

White race

83%

84%

0.83

Age (mean)

75.3

73.7

0.25

Prior Open Abdominal Surgery

31%

22%

0.24

ASA 4 Classification

20%

21%

1.00

5.7

5.7

0.99

Congestive heart failure

2.5%

2.4%

1.00

Hypertension

85%

80%

0.55

Diabetes

5.0%

16%

0.07

Chronic Obstructive Pulmonary Disease

20%

16%

0.52

Dialysis (pre-op)

5.0%

1.3%

0.11

Obesity

23%

30%

0.48

Acute conversion to open AAA repair

10%

0.2%

<.01

Mean OR time (min)

219

131

<.01

Median OR time (min)

195

116

<.01

Mean OR time (all concomitant procedures excluded)

207

121

<.01

Median OR time (all concomitant procedures excluded)

191

110

<.01

Renal Stent

7.5%

5.8%

0.51

Hypogastric Embolization

5.0%

7.7%

0.76

Hypogastric Revascularization

5.0%

3.1%

0.36

Lower Extremity Revascularization

9.4%

1.8%

0.02

Iliac Branched Device

18%

10%

0.18

Aortic (Bare metal) Stent

13%

0.8%

<.01

Iliac (Bare metal) Stent

7.5%

3.3%

0.15

Any Concomitant Procedure

38%

26%

0.10

Aneurysm diameter (cm)

7

Coexisting conditions

147

Chapter 7 common in patients with pEVAR (Table II). Between groups, there were no differences in the indication for surgery, the proximal extent of the aneurysm, or the rate of acute conversion to open repair. The distal extent of the aneurysm, however, was more likely to be common or external iliac in the pEVAR group (P<.01). There were differences in main body device used between cohorts (P<.01), with slightly more Gore Excluder and Endologix Powerlink used in the pEVAR group and slightly more Cook Zenith and Medtronic Endurant use in the cEVAR group. The mean time of operation was shorter for pEVAR (135 min vs. 152 min, P<.01). After exclusion of any cases with concomitant interventions and fenestrated cases the mean total operation time remained significantly shorter (123 min vs. 141 min, P<.01) for pEVAR. Postoperatively there were more wound complications in patients undergoing cEVAR (2.1% vs. 1.0%, P=0.02), primarily superficial surgical site infection (2% vs. 1%, P=0.03). There were no other significant differences in post-operative complications (Table III). In pEVAR patients the median length of stay was significantly shorter (1 day vs. 2 days, P<.01). The proportion of patients with a length of stay longer than 2 days was significantly lower with pEVAR (46% vs. 56%, P<.01). After correcting for differences in age, race, Chronic Obstructive Pulmonary Disease (COPD), CHF, obesity and concomitant procedures, cutdown was still associated with a significantly longer operative time (16 minutes) and a 39% greater likelihood of length of stay > 2days. Comparing single groin cEVAR to bilateral groin cEVAR, we found no differences in post-operative results, other than more deep venous thrombosis (1.1% vs. 0.3%, P=0.03) in the unilateral cutdown cohort. Therefore we combined them in the cEVAR group for all analyses. The failed percutaneous access patients (N=40, 4%) showed no differences in baseline characteristics when compared to patients with successful pEVAR. Peri-operatively, however, failed pEVAR patients had more acute conversions to open repair (10% vs. 0.2%), more concomitant procedures particularly aortic stenting (13% vs. 0.8%) and lower extremity revascularization (9.4% vs. 1.8%), and as expected longer operative times (Table IV). The only multivariable predictor of percutaneous access failure was performance of any concomitant procedure (OR 2.0, 95% CI 1.0-4.0, P=0.04). After adjustment, age, CHF, COPD and obesity were not predictive of percutaneous access failure.

Discussion

The use of pEVAR in the Targeted Vascular ACS NSQIP increased over the study period. Our main findings in this study of the Targeted Vascular ACS NSQIP database are a technical success rate of 96% in pEVAR patients, a significantly shorter total operation time, a shorter length of hospital stay, and fewer wound complications in patients who were treated with pEVAR. A recently published multi-center randomized controlled trial by Nelson et al, randomized 151 patients undergoing EVAR with the Endologix device to pEVAR, using either the Prostar XL or Perclose Proglide, versus standard cEVAR.2 Their 148

Percutaneous versus Femoral Cutdown Access study demonstrated a technical success rate of 96% with the use of Perclose Proglide, which was the same as our finding. A lower success rate was seen with the Prostar XL (90%). Unfortunately, we cannot tell which closure device was used in the NSQIP centers. Additionally, Nelson et al also reported significantly shorter operative time for both closure devices and a shorter length of hospital stay for pEVAR compared to cEVAR, which again was confirmed in our analysis suggesting that the results of the randomized trial are generalizable to a broader population of patients and centers using a variety of stentgrafts. The Targeted Vascular ACS NSQIP includes data from a smaller subset of the total NSQIP group of hospitals (83 of 435). This is likely a subset of hospitals with a volume of vascular procedures high enough and a commitment to quality improvement strong enough to warrant participation in this new quality improvement effort at its initiation. Therefore, it seems likely that these data may not be generalizable to all patients undergoing EVAR at all institutions by all providers. Our previous study reported a success rate of 96% in our single center experience with ultrasound scan-guided pEVAR,8 and noted an increasing success rate over time with the increased use of ultrasound scan-guided pEVAR. We found that patients with smaller access vessel diameters were more prone to have percutaneous failure. Additionally, a shorter operative time and fewer wound complications in pEVAR patients were found. Among prior published reports we found success rates for pEVAR varying from 71% to 100%.2-12, 14-16 In addition to access vessel diameter and type of closures device, femoral artery calcification, access vessel tortuousity and groin scars have all been associated with pEVAR failure.8, 17, 18 Unfortunately, this database does not provide anatomic information about femoral artery calcification, access vessel diameter and previous groin operations. However, our study demonstrates that performing a concomitant procedure is a predictor of percutaneous access failure. A significantly decreased mean total operative time2-8 and a shorter length of hospital stay3, 6, 9, 10 were noted in many previous studies in pEVAR patients when compared to cEVAR patients. Prior studies have noted wound complications ranging from 0-11%9, 14 with a rate of 5.0%2 noted in the randomized trial. In this analysis of NSQIP centers we found a significantly lower rate of wound complications in pEVAR patients. It may be expected that in the context of a randomized trial minor wound complications may be more likely to be noted compared to retrospective single center studies,19 however, trained NSQIP nurse clinical reviewers ability to detect surgical site infection has been previously demonstrated.20 Intraoperative blood loss was noted to be less with pEVAR in one report11, while other studies report similar blood loss for both groups6, 7. Our study shows that the need for intra- and postoperative transfusion is similar in both groups. Although not significant, we found more obese patients among our cEVAR cohort. It may be that surgeons find percutaneous access more difficult in obese patients. Previous studies showed conflicting results regarding the influence of obesity on pEVAR success rates. Some studies suggest an association between obesity and technical failure of pEVAR 4, 7, 18, 22, while others found that obesity had no influence on the pEVAR success rate8, 12, 14, 17. In particular, those studies employing routine use of ultrasound scan-guided pEVAR were more 149

7

Chapter 7 likely to note the lack of association of obesity with success8, 16, 23. Given the increased incidence of wound complications in obese patients, these patients could potentially benefit most from pEVAR,24 This study has several limitations, which should be noted. With the use of the Targeted Vascular NSQIP database our study design was retrospective and the cohorts were not randomized. Due to the lack of randomization, surgeon’s preference and experience likely played a roll in the choice of treatment. We had no data on how patients were selected for pEVAR and we were unable to determine when ultrasound guidance was used, which has been suggested to reduce access related complications.8, 25 We do not have data describing the type of vascular closure devices that were used. Additionally, the database does not provide information on long-term follow-up and prevented us from comparing the incidence of ileo-femoral stenosis.

Conclusion

This study demonstrates that with the use of pEVAR in elective AAA patients, high technical success rates, shorter operation time, shorter length of hospital stay, and fewer wound complications can be achieved.. Our study confirms the findings of the earlier randomized trial and single center studies, highlighting the current state of EVAR in centers that are represented in the Targeted Vascular ACS NSQIP database.

150

Percutaneous versus Femoral Cutdown Access

References 1. Schermerhorn ML, Bensley RP, Giles KA, Hurks R, O’Malley A J, Cotterill P, et al. Changes in abdominal aortic aneurysm rupture and short-term mortality, 1995-2008: a retrospective observational study. Annals of surgery. 2012;256(4):651-8. 2. Nelson PR, Kracjer Z, Kansal N, Rao V, Bianchi C, Hashemi H, et al. A multicenter, randomized, controlled trial of totally percutaneous access versus open femoral exposure for endovascular aortic aneurysm repair (the PEVAR trial). Journal of vascular surgery. 2014;59(5):1181-93. 3. Dosluoglu HH, Lall P, Blochle R, Harris LM, Dryjski ML. Ambulatory percutaneous endovascular abdominal aortic aneurysm repair. Journal of vascular surgery. 2014;59(1):58-64. 4. Lee WA, Brown MP, Nelson PR, Huber TS. Total percutaneous access for endovascular aortic aneurysm repair (“Preclose” technique). Journal of vascular surgery. 2007;45(6):1095-101. 5. Malkawi AH, Hinchliffe RJ, Holt PJ, Loftus IM, Thompson MM. Percutaneous access for endovascular aneurysm repair: a systematic review. European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery. 2010;39(6):676-82. 6. Morasch MD, Kibbe MR, Evans ME, Meadows WS, Eskandari MK, Matsumura JS, et al. Percutaneous repair of abdominal aortic aneurysm. Journal of vascular surgery. 2004;40(1):12-6. 7. Torsello GB, Kasprzak B, Klenk E, Tessarek J, Osada N, Torsello GF. Endovascular suture versus cutdown for endovascular aneurysm repair: a prospective randomized pilot study. Journal of vascular surgery. 2003;38(1):78-82. 8. Bensley RP, Hurks R, Huang Z, Pomposelli F, Hamdan A, Wyers M, et al. Ultrasound-guided percutaneous endovascular aneurysm repair success is predicted by access vessel diameter. J Vasc Surg. 2012;55(6):1554-61. 9. Jean-Baptiste E, Hassen-Khodja R, Haudebourg P, Bouillanne PJ, Declemy S, Batt M. Percutaneous closure devices for endovascular repair of infrarenal abdominal aortic aneurysms: a prospective, non-randomized comparative study. European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery. 2008;35(4):422-8. 10. McDonnell CO, Forlee MV, Dowdall JF, Colgan MP, Madhavan P, Shanik GD, et al. Percutaneous endovascular abdominal aortic aneurysm repair leads to a reduction in wound complications. Irish journal of medical science. 2008;177(1):49-52. 11. Howell M, Villareal R, Krajcer Z. Percutaneous access and closure of femoral artery access sites associated with endoluminal repair of abdominal aortic aneurysms. Journal of endovascular therapy : an official journal of the International Society of Endovascular Specialists. 2001;8(1):68-74. 12. Petronelli S, Zurlo MT, Giambersio S, Danieli L, Occhipinti M. A single-centre experience of 200 consecutive unselected patients in percutaneous EVAR. 151

7

Chapter 7 La Radiologia medica. 2014;119(11):835-41. 13. American College of Surgeons National Surgical Quality Improvement Program. http://www.site.acsnsqip.orgWeb Page. 14. Eisenack M, Umscheid T, Tessarek J, Torsello GF, Torsello GB. Percutaneous endovascular aortic aneurysm repair: a prospective evaluation of safety, efficiency, and risk factors. J Endovasc Ther. 2009;16(6):708-13. 15. Lee WA, Brown MP, Nelson PR, Huber TS, Seeger JM. Midterm outcomes of femoral arteries after percutaneous endovascular aortic repair using the Preclose technique. Journal of vascular surgery. 2008;47(5):919-23. 16. Mousa AY, Abu-Halimah S, Nanjundappa A, AbuRahma AF, Richmond BK. Current update on the status of totally percutaneous aneurysm repair. Vascular and endovascular surgery. 2013;47(6):409-14. 17. Mousa AY, Campbell JE, Broce M, Abu-Halimah S, Stone PA, Hass SM, et al. Predictors of percutaneous access failure requiring open femoral surgical conversion during endovascular aortic aneurysm repair. J Vasc Surg. 2013;58(5):1213-9. 18. Teh LG, Sieunarine K, van Schie G, Goodman MA, Lawrence-Brown M, Prendergast FJ, et al. Use of the percutaneous vascular surgery device for closure of femoral access sites during endovascular aneurysm repair: lessons from our experience. European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery. 2001;22(5):418-23. 19. Ortega G, Rhee DS, Papandria DJ, Yang J, Ibrahim AM, Shore AD, et al. An evaluation of surgical site infections by wound classification system using the ACS-NSQIP. The Journal of surgical research. 2012;174(1):33-8. 20. Snyder RA, Johnson L, Tice J, Wingo T, Williams D, Wang L, et al. Wound classification in pediatric general surgery: significant variation exists among providers. Journal of the American College of Surgeons. 2013;217(5):81926. 21. Kolluri R, Fowler B, Nandish S. Vascular access complications: diagnosis and management. Current treatment options in cardiovascular medicine. 2013;15(2):173-87. 22. Starnes BW, Andersen CA, Ronsivalle JA, Stockmaster NR, Mullenix PS, Statler JD. Totally percutaneous aortic aneurysm repair: experience and prudence. Journal of vascular surgery. 2006;43(2):270-6. 23. Arthurs ZM, Starnes BW, Sohn VY, Singh N, Andersen CA. Ultrasound-guided access improves rate of access-related complications for totally percutaneous aortic aneurysm repair. Annals of vascular surgery. 2008;22(6):736-41. 24. Falagas ME, Kompoti M. Obesity and infection. The Lancet Infectious diseases. 2006;6(7):438-46. 25. Lo RC, Fokkema MT, Curran T, Darling J, Hamdan AD, Wyers M, et al. Routine use of ultrasound-guided access reduces access site-related complications after lower extremity percutaneous revascularization. J Vasc Surg. 2014.

152

Percutaneous versus Femoral Cutdown Access

7

153

CHAPTER 8

Comparing the Transperitoneal vs. Retroperitoneal Surgical Approach Open Abdominal Aortic Aneurysm Repair. Dominique B. Buck, Sara Zettervall, Klaas H.J. Ultee, Pete Soden, Jeremy Darling, Joost A. van Herwaarden, Mark Wyers, Marc L. Schermerhorn. Submitted J Am Coll Surg.

Chapter 8

Abstract Background Retroperitoneal approach for AAA repair may offer a faster return of gastrointestinal function, fewer pulmonary complications, superior access to the suprarenal aorta, and fewer long-term reinterventions for hernia’s and bowel obstructions compared to the transperitoneal approach. We sought to compare current practices in patient selection and 30-day outcomes for transperitoneal and retroperitoneal AAA repairs. Study Design All patients undergoing elective transperitoneal or retroperitoneal surgical repair for AAA between January 2011 and December 2013 were identified in the Targeted Vascular NSQIP database. Emergent cases were excluded. Baseline characteristics, and anatomical detail were evaluated using chi-square test or t-test. Intra-operative and post-operative outcomes were evaluated among those with infrarenal or juxtarenal AAA only. Adjusted outcomes were compared using multivariable analyses. Results 1,135 patients were identified; 788 transperitoneal (69%), 347 retroperitoneal (31%). Patients with a retroperitoneal approach were more likely to be white (90% vs. 81%, p<.001), male (74% vs. 67%, p=0.018) and had more prior open abdominal surgery than the transperitoneal group (35% vs. 21%, p<.001). However, there were no significant differences in age, ASA class, aneurysm diameter or coexisting comorbidities. Evaluating infrarenal and juxtarenal aneurysms only, the retroperitoneal patients were less likely to have an infrarenal clamp location (43% vs. 68%), had more renal revascularizations (15% vs. 6%, p<.001), more visceral revascularizations (5.6% vs. 2.4%, p=0.014) and more lower extremity revascularizations (11% vs. 7%, p=0.021) compared to the transperitoneal approach. Post-operatively there were no differences in mortality, or return to OR. Transperitoneal patients had a higher rate of wound dehiscence (2.4% vs. 0.4%, p=0.045), while retroperitoneal patients had higher incidence of pneumonia (9% vs 5%, p=0.034), transfusions (77% vs. 71%, p=0.037), reintubations (11% vs. 7%, p=0.034) and a greater median length of stay (8 vs. 7 days, p=0.048). After exclusion of all concomitant procedures, only transfusions remained more common in the retroperitoneal approach (78% vs. 70%, p=0.036). Multivariable analyses showed only higher rates of reintubation in the retroperitoneal group (OR1.7, 95% CI 1.0-3.0, p=0.047). Conclusions The retroperitoneal approach is more commonly used for more proximal aneurysms and was associated with higher rates of pneumonia, reintubations, transfusions and a greater length of stay on univariate analyses. However, multivariable analysis demonstrated only higher rates of reintubation in the retroperitoneal approach. 156

Transperitoneal vs. Retroperitoneal Surgical Approach Thus perioperative results are overall similar. The long-term benefits and frequency of reinterventions remain to be proven.

8

157

Chapter 8

Introduction

Despite the development of endovascular techniques for abdominal aortic aneurysms (AAA) in recent years, the open surgical repair remains necessary to treat anatomically complex aneurysms and may be beneficial for younger patients and those unable to comply with long-term surveillance.1 Medicare data show that in 2008 open surgical repair was used in 23% of intact AAA repairs in the United States, but also highlighted higher rates of medical and surgical complications after open surgical repair.2 Additionally, lower rates of late bowel and hernia complications were demonstrated with retroperitoneal surgery compared to intraperitoneal surgery for non-aortic surgical procedures.3 The transperitoneal approach is most familiar to surgeons 4, 5 and provides extensive intra-abdominal access. The time from skin incision to aortic clamping was shown to be shorter in the transperitoneal group compared to the retroperitoneal group.6, 7 However, complications associated with this approach have been reported leading to prolonged ileus.8 In an attempt to avoid complications associated with entrance into the peritoneal cavity, the retroperitoneal approach was adopted and also offers improved access to the suprarenal and supraceliac aorta.4, 9 Several studies comparing the transperitoneal and retroperitoneal approach for AAA repair suggest that the retroperitoneal approach may result in lower rates of ileus, shortened hospital length of stay, and improved respiratory function, in relation to transperitoneal procedures.5, 6, 8, 10-14 Conversely, other studies have found no differences in respiratory complications, return of bowel function and length of stay.15, 16 This study aims to identify the demographic and anatomical differences between patients selected for elective transperitoneal versus retroperitoneal AAA repair and to assess differences in intra-operative details and perioperative mortality and complications.

Methods Data Source The cohort of patients for this study was identified using the Targeted Vascular American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database. This dataset contains prospectively collected clinical data and includes over 80 participating hospitals in the United States. Data are recorded on preoperative, operative and postoperative variables, which were chosen by vascular surgeons specific to the procedure performed. To ensure quality, data are collected by trained clinical nurse reviewers. The Targeted Vascular ACS NSQIP database contains unidentifiable patients only, therefore, Institutional Review Board approval and patient consent were waived. Details of quality control and data collection have been described extensively on the ACS NSQIP website.17 Definitions for the specific Targeted Vascular ACS NSQIP collected data points may be found in the user guide available online.18 We identified all patients undergoing either elective transperitoneal or 158

Transperitoneal vs. Retroperitoneal Surgical Approach retroperitoneal open surgical repair for AAA, from January 2011 to December 2013. Emergent cases were excluded from analyses. Outcomes We compared patients with transperitoneal surgical repair (midline and transverse incisions) to retroperitoneal repair. Data were collected on relevant patient demographics, comorbidities, operative details and postoperative course. Obesity was defined at BMI> 30kg/m2. Baseline characteristics were compared for all AAA, including infrarenal-, juxtarenal-, pararenal-, suprarenal- and Type IV thoracoabdominal aneurysms. In this database, pararenal is defined as AAAs that involve the origin of the renal arteries, and is distinct from both juxtarenal (AAAs that do not involve the renal arteries but because of proximity require clamping above the renal arteries to complete the proximal anastomosis) and suprarenal (AAAs that begin above at least one main renal artery but below the visceral segment). Renal and visceral revascularization included bypass and endarterectomy. Operative time was defined as the time from incision to skin closure. For comparison of intra-operative and post-operative outcomes, only juxtarenal and infrarenal aneurysms undergoing transperitoneal or retroperitoneal surgical repairs were included. Intra-operative anatomical details included clamp location, extent of aneurysm, extent of distal iliac involvement, and concomitant renal, visceral or lower extremity revascularization performed. Concomitant procedures were excluded in comparison of operative time, to rule out the effect of additional interventions. Any wound complication was defined as a composite variable inclusive of any surgical site infection (SSI), superficial SSI, deep SSI, organ/space SSI, or dehiscence. Statistical Analysis All analyses were conducted with IBM SPSS Statistics version 20 (IBM Corp, Armonk, NY). Categorical variables were compared using Pearson chisquare or Fisher exact test, and continuous variables were analyzed by twotailed independent samples t-test or Mann-Whitney test where appropriate. Multivariable logistic regressions were performed correcting for gender, age, race, prior abdominal surgery, concomitant procedure, aneurysm extent, and management of the inferior mesenteric artery. Statistical significance was defined as p<0.05.

Results

A cohort of 1,135 patients was identified, including 788 (69%) patients with a transperitoneal approach, and 347 (31%) with a retroperitoneal approach for repair. Patients that underwent the retroperitoneal approach were less likely to have an infrarenal aneurysm (31% vs. 61%), and more likely to have juxtarenal(41% vs. 28%), pararenal- (9% vs. 4%), suprarenal- (14% vs.5%) and type IV thoracoabdominal aneurysms (5% vs. 1%). (Table 1) The transperitoneal approach was used more often in men (74% vs. 67%, P=0.018), and less often in those of white race (81% vs. 90%, P<.001) or with 159

8

Chapter 8 Table 1. Baseline Characteristics of Patients with Abdominal Aortic Aneurysms undergoing open repair (Transperitoneal vs. Retroperitoneal approach) Variable Male gender

Transperitoneal (n=788)

Retroperitoneal (n=347)

P Value

74%

67%

0.018

White Race

81%

90%

<.001

Age (mean)

70.59

71.28

0.228

18-59

11%

7.80%

60-69

33%

33%

70-79

39%

42%

80+

17%

18%

0.313

Age Category (years)

Prior Open Abdominal Surgery

21%

35%

<.001

ASA 4 Classification

34%

36%

0.505

Aneurysm diameter (cm)

6.16

6.27

0.261

Coexisting conditions 2.20%

1.40%

0.420

Hypertension

84%

83%

0.623

Diabetes

13%

12%

0.843

History of severe COPD

20%

22%

0.484

0.80%

0.90%

0.857

27%

30%

0.294

Infrarenal

61%

31%

Juxtarenal

28%

41%

Congestive heart failure

Dialysis (pre-op) Obesity Proximal Aneurysm Extent

Pararenal

4.30%

9.20%

Supra-renal

5.20%

14%

Type IV Thoracoabdominal aneurysm

1.40%

4.90%

* COPD = Chronic Obstructive Pulmonary Disease

prior abdominal operations (21% vs. 35%, P<.001). There were no significant differences in mean age (71 vs. 71, P=0.228), ASA Class ≼ 4 (34 % vs. 36%, P=0.505), or mean aneurysm diameter (6.2 vs. 6.3, p=0.261). The same significant differences in baseline characteristics persisted when the analysis was confined to infrarenal and juxtarenal aneurysms only. Analysis of intra-operative characteristics among infrarenal and juxtarenal repairs only indicated that patients undergoing retroperitoneal repairs were more likely to have a more renal revascularization (15% vs. 6%, p<.001), lower extremity revascularization (11% vs. 7%, p=0.021), and visceral revascularization (6% vs.

160

Transperitoneal vs. Retroperitoneal Surgical Approach Table 2. Intra-operative Outcomes of Patients with Infrarenal or Juxtarenal Abdominal Aortic Aneurysms undergoing Transperitoneal vs. Retroperitoneal Open Surgical Repair Outcomes

Transperitoneal (n=702)

Retroperitoneal (n=248)

P Value

Surgical Approach 100.00%

Retroperitoneal Transperitoneal-midline

95%

Transperitoneal-transverse

5%

Aneurysm Extent Infrarenal

68%

43%

Juxtarenal

32%

57% 0.205

Indication for surgery 81%

78%

Dissection

0.90%

0.40%

Embolization

0.30%

1.60%

Non-ruptured symptomatic

8.50%

11%

Not documented

0.60%

0.80%

Prior endovascular intervention w/ unsatisfactory result

3.10%

4.40%

Prior open intervention w/ unsatisfactory result

0.60%

0.00%

Diameter

Rupture w/ or w/out hypotension

2.30%

2.00%

Thrombosis

3.30%

2.00%

4.60%

15%

Between SMA & renals

16%

28%

Above one renal

16%

17%

Infrarenal

58%

34%

6.00%

6.90%

Aortic

39%

53%

Common Iliac

44%

30%

External Iliac

4.40%

6.50%

Internal Iliac

4.10%

1.20%

Not documented

8.40%

9.30%

<.001

Proximal Clamp Location Supraceliac

Not documented Distal Extent

<.001

Management of Inferior Mesenteric Artery 11%

4.80%

6.00%

2.80%

Ligated

42%

36%

Not documented

41%

57%

Chronically occluded Implanted

161

8

Chapter 8 Renal Revascularization

6.40%

15%

<.001

Visceral (SMA & Celiac) Revascularization

2.40%

5.60%

0.014

Lower Extremity Revascularization

6.70%

11%

0.021

Abdominal, non-arterial repair or excision

3.30%

1.60%

0.175

Any Concomitant Procedure

16.20%

25.80%

0.001

Mean OR time (min)

253

264

0.196

Median OR time (min)

234

238

0.124

Mean OR time (all concomitant procedures excluded)

240

244

0.627

Median OR time (all concomitant procedures excluded)

219

227

0.248

* SMA = superior mesenteric artery, OR = operative time Table 3. Post-operative Outcomes of Patients with Infrarenal and Juxtarenal Abdominal Aortic Aneurysms undergoing Transperitoneal vs. Retroperitoneal Open Surgical Repair Outcomes

Transperitoneal Retroperitoneal (n=702) (n=248)

P Value

Odds Ratio (95% CI)

Mortality

3.80%

3.60%

0.878

0.9 (0.4-2.0)

Rupture of Aneurysm

0.70%

0.80%

0.881

1.1 (0.2-5.9)

Myocardial Infarction

2.60%

2.80%

0.827

1.1 (0.5-2.7)

Cardiac Arrest requiring CPR

2.10%

2.80%

0.537

1.3 (0.5-3.3)

Pneumonia

5.10%

8.90%

0.034

1.8 (1.0-3.1)

Pulmonary Embolism

0.40%

0.40%

0.960

0.9 (0.1-9.1)

Progressive Renal Insufficiency

1.70%

2.40%

0.481

1.4 (0.5-3.8)

Medical complications (% of patients)

Acute Renal Failure

3.40%

6.00%

0.073

1.8 (0.9-3.5)

Urinary Tract Infection

3.00%

1.60%

0.244

0.5 (0.2-1.6)

Stroke / CVA

0.40%

0.80%

0.478

1.9 (0.3-11.4)

DVT

2.10%

1.60%

0.612

0.8 (0.2-2.3)

71%

77%

0.037

1.4 (1.0-2.0)

1.30%

3.20%

0.047

2.6 (1.0-6.7)

Prolonged Intubation > 48 hours

9.10%

13%

0.061

1.5 (1.0-2.4)

Reintubation

6.70%

11%

0.034

1.7 (1.0-2.8)

Return to OR

9.50%

12%

0.254

1.3 (0.8-2.1)

Transfusion Sepsis Surgical complications (% of patients)

Any Wound Complication

5.60%

2.80%

0.085

0.5 (0.2-1.1)

Wound dehiscence

2.40%

0.40%

0.045

0.2 (0.0-1.2)

Superficial SSI

2.10%

1.60%

0.612

0.8 (0.2-2.3)

Deep SSI

0.70%

0.40%

0.579

0.6 (0.1-4.9)

162

Transperitoneal vs. Retroperitoneal Surgical Approach 1.10%

0.40%

0.303

0.4 (0.0-2.8)

Ischemic colitis

3.70%

2.40%

0.335

0.6 (0.3-1.6)

Lower Extremity Ischemia

2.40%

2.00%

0.715

0.8 (0.3-2.3)

9.6

10.5

0.173

7

8

0.048

Organ/Space SSI

Mean length of hospital stay (no. of days) Median length of hospital stay (no. of days) Length of stay ≥ 8 days

43%

52%

0.012

1.5 (1.1-1.9)

Discharged home

77%

74%

0.341

0.9 (0.6-1.2)

CPR = cardiopulmonary resuscitation, CVA = cerebrovascular accident, DVT = deep venous thrombosis, OR = operative time, SSI = surgical site infection Table 4. Adjusted Outcomes of Patients with Infrarenal and Juxtarenal Abdominal Aortic Aneurysms undergoing Transperitoneal vs. Retroperitoneal Open Surgical Repair Outcome

Odds Ratio

(95% CI)

P-value

Mortality

0.9

0.4-2.1

0.889

Myocardial Infarction

1.1

0.4-3.1

0.805

Pneumonia

1.8

1.0-3.2

0.067

Reintubation

1.7

1.0-3.0

0.047

Acute Renal Failure

1.8

0.8-3.9

0.144

Stroke/CVA

1.9

0.2-14.3

0.552

Return to the OR

0.7

0.4-1.3

0.268

Ischemic Colitis

1.4

0.5-3.9

0.514

Any Wound Complication

0.7

0.3-1.6

0.398

Wound dehiscence

0.2

0.0-1.8

0.157

Wound Infection

0.9

0.1-8.3

0.941

Length of stay ≥ 8 days

1.4

1.0-1.9

0.070

* CVA = cerebrovascular accident, OR = operative time † Variables included in the regression: male gender, age, white race, prior abdominal surgery, any concomitant procedure, aneurysm extent into the iliac artery, management of inferior mesenteric artery. The retroperitoneal approach is the designated reference category.

2%, p=0.014). (Table 2) When all concomitant procedures were combined, this rate was higher in the retroperitoneal approach (26% vs. 16%, p=0.001). There was no difference in the management of the inferior mesenteric artery, whether ligation, re-implantation or observation of chronic occlusion. Between the two cohorts, there were no differences in operative time. However, median operative time was significantly shorter for the transperitoneal approach among patients with tube grafts alone (211 vs. 235 minutes, p=0.027). 163

8

Chapter 8 Analysis of the unadjusted post-operative outcomes demonstrates that the retroperitoneal group had significantly higher rates of transfusion (77% vs. 71%, p=0.037), sepsis (3.2% vs. 1.3%, p=0.047), pneumonia (8.9% vs. 5.1%, p=0.034), reintubation (11% vs. 7%, p=0.034) and median length of stay (8 vs. 7, p=0.048). (Table 3) Additionally, length of stay >8 days was significantly higher in the retroperitoneal group (52% vs. 43%, p=0.012). The rate of wound dehiscence was higher in the transperitoneal cohort (2.4% vs. 0.4%, p=0.045). There were no significant differences in mortality (3.6% vs. 3.8%, p=0.878) or return to the operating room (12% vs. 10%, p=0.382). Following exclusion of any concomitant procedures during AAA repair, the only transfusion remained higher in the retroperitoneal group (78% vs. 70%, p=0.036). Multivariable analyses, with adjustment for concomitant procedures and baseline characteristics, show that only reintubation remains significantly higher with the retroperitoneal approach (OR 1.7, 95% CI 1.0-3.0, p=0.047). (Table 4)

Discussion

Our study found that as expected, the retroperitoneal approach was more commonly utilized for aneurysms with more proximal extent and was associated with higher rates of pneumonia, reintubations, transfusions and greater length of stay. The transperitoneal approach, however, showed higher rates of wound dehiscence. After multivariable analysis, however, most of these differences were simply related to concomitant procedures, and only reintubation rate remains significantly higher with the retroperitoneal approach. Data from previous randomized control trials have demonstrated antithetical outcomes.8, 16, 19-21 Arya et al found that retroperitoneal surgery shortens the length of stay, in a study of 35 patients, 19 while Sieunarine et al (n=100) found no differences in length of stay.16 Sicard et al, (n=145) also found that retroperitoneal surgery was associated with shorter hospital stays. Our analysis noted a greater length of stay in retroperitoneal approach patients although this was related to concomitant procedures and did not persist after adjustment for concomitant procedures and aneurysm extent. Sicard et al demonstrated that patients undergoing the retroperitoneal approach had fewer postoperative complications (including bowel obstruction, wound-, pulmonary- and cardiac complications.8 Sienaruine et al found higher rates of wound complications in the retroperitoneal group, but no differences in any other postoperative complication. Our data showed lower rates of wound dehiscence with the retroperitoneal approach and higher rates of pneumonia, reintubations, transfusions and greater length of stay. However, only reintubation was found to be significantly higher in the retroperitoneal group after accounting for other patient and procedural characteristics. Although published before 1990, several larger retrospective reviews of 299, 270 and 213 patients have also demonstrated mixed results. 22-24 Leather et al. (n=299) demonstrated a reduction of pulmonary complications and length of stay favorable for the retroperitoneal approach.22 Peck et al. (n=270) also showed a shorter length of stay with the retroperitoneal approach and more pneumonia, 164

Transperitoneal vs. Retroperitoneal Surgical Approach nasogastric tube, and atelectasis in the transperitoneal group. They did not demonstrate significant differences in wound infections or post-op myocardial infarctions.24 Sicard et al. (n=213) showed that the retroperitoneal approach was preferable with respect to blood loss and again reduced hospital stay.23 Our study demonstrates that the retroperitoneal approach is more commonly used for more proximal aneurysms, as this approach is generally considered to provide better exposure of the suprarenal abdominal aorta. Our analysis showed that concomitant procedures were more common in the retroperitoneal group. After excluding all concomitant procedures, all differences in outcomes were similar except for transfusion. In our multivariable model, which allowed inclusion of greater numbers and still accounts for concomitant procedures as well as differences in anatomic and comorbid conditions, only the reintubation rate remains significantly higher. These results suggest that concomitant procedures drive the observed increase in complications and length of stay. The use of the Targeted Vascular ACS NSQIP database is limited by the lack of randomization. The surgeon’s preference and the volume of the surgeon or center performing open AAA repair cannot be captured with this database, however likely played a roll in the choice of treatment. Additionally, the NSQIP does not have data describing ileus or bowel obstruction in the post-operative period. However, length of stay is likely a reasonable proxy for this and we did not see any difference. Finally, the database does not provide information on outcomes >30 postoperative days. As a result, we are unable to draw conclusions on the potential late effects of open surgical approach on hernia and bowel obstruction, nor on survival. Despite these limitations, our study provides a large cohort including anatomical detail and demonstrates current outcomes with open AAA repair in centers participating in the recently released Targeted Vascular ACS NSQIP.

Conclusion

The retroperitoneal surgery was more commonly used for a more proximal aneurysm extent and with concomitant renal, visceral, and lower extremity revascularization procedures. Although the retroperitoneal approach was associated with was associated with higher rates of pneumonia, reintubation, transfusion, and a longer length of stay, these were driven by concomitant procedures and after adjustment, the only significant difference was a higher rate of reintubation with the retroperitoneal approach Further long-term data regarding late survival, bowel obstruction, and hernia formation will be useful to guide treatment selection. For the time being, the choice should be driven by anatomy and surgeon preference.

165

8

Chapter 8

References 1. Shaw PM, Veith FJ, Lipsitz EC, Ohki T, Suggs WD, Mehta M, et al. Open aneurysm repair at an endovascular center: value of a modified retroperitoneal approach in patients at high risk with difficult aneurysms. J Vasc Surg. 2003;38(3):504-10. 2. Schermerhorn ML, Bensley RP, Giles KA, Hurks R, O’Malley A J, Cotterill P, et al. Changes in abdominal aortic aneurysm rupture and short-term mortality, 1995-2008: a retrospective observational study. Annals of surgery. 2012;256(4):651-8. 3. Bensley RP, Schermerhorn ML, Hurks R, Sachs T, Boyd CA, O’Malley AJ, et al. Risk of late-onset adhesions and incisional hernia repairs after surgery. Journal of the American College of Surgeons. 2013;216(6):1159-67, 67 e112. 4. Chaikof EL, Brewster DC, Dalman RL, Makaroun MS, Illig KA, Sicard GA, et al. The care of patients with an abdominal aortic aneurysm: the Society for Vascular Surgery practice guidelines. J Vasc Surg. 2009;50(4 Suppl):S2-49. 5. Quinones-Baldrich WJ, Garner C, Caswell D, Ahn SS, Gelabert HA, Machleder HI, et al. Endovascular, transperitoneal, and retroperitoneal abdominal aortic aneurysm repair: results and costs. J Vasc Surg. 1999;30(1):59-67. 6. Nakajima T, Kawazoe K, Komoda K, Sasaki T, Ohsawa S, Kamada T. Midline retroperitoneal versus midline transperitoneal approach for abdominal aortic aneurysm repair. J Vasc Surg. 2000;32(2):219-23. 7. Wachenfeld-Wahl C, Engelhardt M, Gengenbach B, Bruijnen HK, Loeprecht H, Woelfle KD. Transperitoneal versus retroperitoneal approach for treatment of infrarenal aortic aneurysms: is one superior? VASA Zeitschrift fur Gefasskrankheiten. 2004;33(2):72-6. 8. Sicard GA, Reilly JM, Rubin BG, Thompson RW, Allen BT, Flye MW, et al. Transabdominal versus retroperitoneal incision for abdominal aortic surgery: report of a prospective randomized trial. J Vasc Surg. 1995;21(2):174-81; discussion 81-3. 9. Twine CP, Lane IF, Williams IM. The retroperitoneal approach to the abdominal aorta in the endovascular era. J Vasc Surg. 2012;56(3):834-8. 10. Arko FR SS, Zarins CK. Repair of Infrarenal Abdominal Aortic Aneurysms. ACS Surgery: Principles and Practice. 2007. 11. Darling RC, 3rd, Shah DM, McClellan WR, Chang BB, Leather RP. Decreased morbidity associated with retroperitoneal exclusion treatment for abdominal aortic aneurysm. The Journal of cardiovascular surgery. 1992;33(1):65-9. 12. Borkon MJ, Zaydfudim V, Carey CD, Brophy CM, Guzman RJ, Dattilo JB. Retroperitoneal repair of abdominal aortic aneurysms offers postoperative benefits to male patients in the Veterans Affairs Health System. Annals of vascular surgery. 2010;24(6):728-32. 13. Kirby LB, Rosenthal D, Atkins CP, Brown GA, Matsuura JH, Clark MD, et al. Comparison between the transabdominal and retroperitoneal approaches for aortic reconstruction in patients at high risk. J Vasc Surg. 1999;30(3):400-5. 166

Transperitoneal vs. Retroperitoneal Surgical Approach 14. Ballard JL, Yonemoto H, Killeen JD. Cost-effective aortic exposure: a retroperitoneal experience. Annals of vascular surgery. 2000;14(1):1-5. 15. Cambria RP, Brewster DC, Abbott WM, Freehan M, Megerman J, LaMuraglia G, et al. Transperitoneal versus retroperitoneal approach for aortic reconstruction: a randomized prospective study. J Vasc Surg. 1990;11(2):314-24; discussion 24-5. 16. Sieunarine K, Lawrence-Brown MM, Goodman MA. Comparison of transperitoneal and retroperitoneal approaches for infrarenal aortic surgery: early and late results. Cardiovascular surgery. 1997;5(1):71-6. 17. American College of Surgeons National Surgical Quality Improvement Program. http://www.site.acsnsqip.orgWeb Page. 18. American College of Surgeons National Surgical Quality Improvement Program. User guide for the 2013 ACS NSQIP procedure Targeted participant use data file. http://site.acsnsqip.org/wp-content/uploads/2014/11/NSQIP. PUF_.ProcedureTargeted.UserGuide.2013.pdf. 19. Arya N, Muhammad Anees S, Lau LL, Lee B, Hannon RJ, Young IS, et al. Retroperitoneal repair of abdominal aortic aneurysm reduces bowel dysfunction. Vascular and endovascular surgery. 2009;43(3):262-70. 20. Lau LL, Gardiner KR, Martin L, Halliday MI, Hannon RJ, Lee B, et al. Extraperitoneal approach reduces neutrophil activation, systemic inflammatory response and organ dysfunctionin aortic aneurysm surgery. European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery. 2001;21(4):326-33. 21. Volta CA, Ferri E, Marangoni E, Ragazzi R, Verri M, Alvisi V, et al. Respiratory function after aortic aneurysm repair: a comparison between retroperitoneal and transperitoneal approaches. Intensive care medicine. 2003;29(8):125864. 22. Leather RP, Shah DM, Kaufman JL, Fitzgerald KM, Chang BB, Feustel PJ. Comparative analysis of retroperitoneal and transperitoneal aortic replacement for aneurysm. Surgery, gynecology & obstetrics. 1989;168(5):387-93. 23. Sicard GA, Allen BT, Munn JS, Anderson CB. Retroperitoneal versus transperitoneal approach for repair of abdominal aortic aneurysms. The Surgical clinics of North America. 1989;69(4):795-806. 24. Peck JJ, McReynolds DG, Baker DH, Eastman AB. Extraperitoneal approach for aortoiliac reconstruction of the abdominal aorta. American journal of surgery. 1986;151(5):620-3.

167

8

CHAPTER 9

The Impact of Endovascular Repair on Specialties Performing Abdominal Aortic Aneurysm Repair Rob Hurks*, Klaas H.J. Ultee*, Dominique B. Buck, George S. DaSilva, Peter A. Soden, Joost A. van Herwaarden, Hence J.M. Verhagen, Marc L. Schermerhorn. Accepted J Vasc Surg.

* These authors contributed equally

Chapter 9

Abstract Introduction Abdominal aortic aneurysm (AAA) repair has been performed by various surgical specialties for many years. Endovascular aneurysm repair (EVAR) may be a disruptive technology, impacting which specialties care for patients with AAA. Therefore, we examined the proportion of AAA repairs performed by various specialties over time in the United States and evaluated the impact of the introduction of EVAR. Methods The Nationwide Inpatient Sample, 2001-2009, was queried for intact and ruptured AAA and for open repair and EVAR. Specific procedures were used to identify vascular surgeons (VS), cardiac surgeons (CS), and general surgeons (GS) as well as interventional cardiologists (IC), and interventional radiologists (IR) for states that reported unique treating physician identifiers. Annual procedure volumes were subsequently calculated for each specialty. Results We identified 108,587 EVAR and 85,080 open AAA repairs (3,011 EVAR and 12,811 open repairs for ruptured AAA). VS performed an increasing proportion of AAA repairs over the study period (52% in 2001 to 66% in 2009, P < .001). GS and CS performed fewer repairs over the same period (25% to 17%, P < .001 and 19% to 13%, P < .001, respectively). EVAR was increasingly utilized for intact (33% to 78% of annual cases, P < .001) as well as ruptured AAA repair (5% to 28%, P < .001). The proportion of intact open repairs performed by VS increased from 52% to 65% (P < .001), while for EVAR the proportion went from 60% to 67% (P < .001). For ruptured open repairs, the proportion performed by VS increased from 37% to 53% (P < .001) and for ruptured EVAR repairs from 28% to 73% (P < .001). Compared to treatment by VS, treatment by a CS (0.55 [0.53-0.56]) and GS (0.66 [0.64-0.68]) was associated with a decreased likelihood of undergoing endovascular rather than open AAA repair. Conclusions VS are performing an increasing majority of AAA repairs, in large part driven by the increased utilization of EVAR for both intact and ruptured AAA repair. However, GS and CS still perform AAA repair. Further studies should examine the implications of these national trends on the outcome of AAA repair.

170

Specialties Performing A Repair

Introduction

During the late 20th century, surgery has become a technology driven profession.1 Since then, innovations such as endoscopic and endovascular surgery have transformed clinical medicine. Besides changing the procedure itself, these disruptive technologies have had their effect on the type of physicians performing the procedures. Percutaneous coronary intervention, for example, has diminished the proportion of coronary revascularizations performed by cardiac surgeons, while the proportion of interventional cardiologists increased dramatically with the use of this technique.2 For abdominal aortic aneurysm (AAA) repair, it is unclear how the introduction and widespread adoption of endovascular repair (EVAR) has changed the distribution of specialties performing elective and ruptured AAA repair. Prior to the introduction of EVAR, open surgical repair was the primary method of treatment. Using Medicare data, Birkmeyer et al. showed that between 1998 and 1999, prior to the widespread adoption of EVAR, vascular surgeons (VS) performed 39% of all elective AAA repairs, while cardiac and general surgeons (CS, GS) performed 33% and 28%, respectively.3 In contrast to elective AAA repair, general surgeons performed the largest proportion of ruptured AAA repairs at 39%, followed by vascular surgeons at 33% and cardiac surgeons at 29%.4 Currently, as with coronary revascularization, the endovascular approach has also led to the inclusion of nonsurgical specialties in treating patients with AAA, such as interventional cardiology (IC) and interventional radiology (IR). Since the performance of EVAR requires a specific skill set that has not been mastered by many surgeons from other specialties, we hypothesize that the proportion of surgical specialties other than VS (i.e., GS and CS) has declined, while VS, IR, and IC are responsible for an increasing number of patients due to a shift from open repair towards EVAR. The purpose of this study is to analyze how the introduction of EVAR has influenced which specialties are providing care for AAA patients for both elective and ruptured AAA repair in the United States.

Methods Database The Nationwide Inpatient Sample (NIS) is the largest national administrative database and represents a 20% sample of all payer (insured and uninsured) hospitalizations. The NIS is maintained by the Agency for Health Care Research and Quality as part of the Healthcare Cost and Utilization Project. Years 2001 to 2009 were queried using International Classification of Diseases, 9th revision (ICD-9) codes to identify patients with diagnosis codes for intact (i.e. elective, symptomatic and mycotic aneurysms) AAA (441.4) and ruptured AAA (441.3). ICD-9 coding does not distinguish infrarenal from juxtarenal or suprarenal AAA. More recent years could not be interrogated due to discontinuation of the surgeon identification variables in the NIS database after 2009.5 Patients who underwent open AAA repair (38.44, 39.25) or EVAR (39.71) were selected. Patients with procedural codes for both open repair and EVAR were considered to have undergone EVAR, as they likely represent conversions to open repair. 171

9

Chapter 9 Patients with codes for a thoracic aneurysm (441.1 or 441.2), thoracoabdominal aneurysm (441.6 or 441.7) or aortic dissection (441.00-441.03) were excluded. The primary outcome was proportional procedure volume by physician specialty over time for intact and ruptured AAA repair. We evaluated the uptake of EVAR overall and by specialty over time. Additionally, we assessed the likelihood of receiving EVAR rather than open repair by specialty. Physician specialty For AAA repair, we were interested in the following type physicians: vascular surgeons (VS), general surgeons (GS), cardiac surgeons (CS), interventional cardiologists (IC), and interventional radiologists (IR). The NIS provides unique physician identifiers per state that allow tracking of procedures performed by that physician during that specific year in that state. Of the available states, 27 provide 2 unique physician identifiers, with 22 of the 27 specifically detailing which physician performed the primary procedure (Supplemental Table III). For the remaining 5 states, the identifiers were only used when both identifiers were the same to ensure that the identified physician was the one performing the primary procedure. We composed a list of specific procedures (Supplemental Table I) that we used to determine the specialty of each physician (VS, GS, CS, IC, or IR). The top 15 procedures identified for each of the physician specialties are listed in Supplemental Table II. Similar approaches have been previously reported.6-8 Subsequently, a hierarchical model was created: each physician that performed >10 cardiac surgery procedures was labeled a CS; the remaining physicians that performed >10 interventional cardiology procedures (e.g. coronary stenting) were labeled IC; physicians with >10 interventional radiology procedures not typically performed by VS (e.g. liver biopsy, nephrostomy, etc.) were identified as IR; the remaining physicians whose procedures consisted of 75%-100% of vascular procedures with >10 in number, were classified as VS; physicians whose procedures consisted of 0%-75% vascular procedures and performed >10 general surgery procedures were classified as GS. Similar approaches have been previously described.9, 10 Two hundred and ten procedures labeled as open repairs were coded as being performed by IC or IR (0.1% of total procedures). We felt these were most likely miscoded endovascular procedures and excluded these patients from further analyses. Statistical approach Mean and standard deviation are reported for parametric data. Baseline variables were compared using Chi-square tests or t-tests, where appropriate. We examined the proportional volume of open AAA repairs and EVAR for each specialty and how this changed over the study period. Trends over time were assessed using the Cochran-Armitage test for trend. A P-value less than .05 indicates that annual procedural volumes followed a significant upward or downward (i.e., non-random) trend over time. Multivariable logistic regression analysis was conducted to examine the influence of physician specialty type on the type of procedure performed, whether open or endovascular. Analyses were considered statistically significant when P < .05. Statistical analyses were performed using SAS 9.2 software (SAS Institute, Cary, NC) and SPSS Statistics 21 (IBM Inc., Chicago, IL). 172

Specialties Performing A Repair

Results

Overall, 108,587 EVAR and 85,080 open AAA repairs were identified in the study period, of which 3,011 EVAR and 12,811 open repairs were for ruptured AAA. The annual overall volume increased from 20,134 in 2001 to 22,541 in 2009 (P 3%

4%

90% 19% 80%

13%

100%

70% 60%

17% IC + IR (P < .001)

25%

CS (P < .001)

50%

GS (P < .001)

40% 30% 20%

66%

52%

VS (P < .001)

10% 0% 2001

2002

2003

2004

2005

2006

2007

2008

2009

Figure 1. Proportion of all abdominal aortic aneurysm repairs (both open and endovascular) performed by each physician specialty from 2001-2009 in the Nationwide Inpatient Sample (totals sum to 100%)

100% 90%

19%

24%

80% 70% 60%

16% 25%

CS (P = .001)

50%

GS (P < .001)

40% 30% 20%

65%

52%

VS (P < .001)

10% 0% 2001

2002

2003

2004

2005

2006

2007

2008

2009

Figure 2A. Proportion of all open repairs for intact abdominal aortic aneurysms performed by physician specialty from 2001-2009 in the Nationwide Inpatient Sample (totals sum to 100%)

< .001). Patient and hospital characteristics are detailed in Table I. Of all AAA repairs, 61% of AAA repairs were performed by VS, 20% by GS and 16% by CS, while the remainder was performed by IC and IR (3% combined). Figure 1 illustrates changes over time for each physician specialty. VS performed an increasing proportion of AAA repairs over the study period (52% in 2001 to 173

9

Chapter 9 66% in 2009, P < .001, Supplemental Table IV). During the same period, GS and CS performed fewer repairs (25% to 17%, P < .001 and 19% to 13%, P < .001, respectively). Similarly, the absolute number of VS performing AAA repair increased with 30% over the study period, while the number of GS and CS decreased over time (46% and 30%, respectively). Intact AAA Repair With 55%, VS performed the majority of open AAA repairs (increasing from 52% to 65% from 2001 to 2009, P < .001). Over this same time period GS performed 24% of all intact open repairs (decreasing from 25% to 16%, P < .001), followed 10%

100% 90%

6% 11%

11%

80%

17%

70% 19%

IC + IR (P = .015)

60%

CS (P = .009)

50%

GS (P < .001)

40% 30%

67%

60%

VS (P < .001)

20% 10% 0% 2001

2002

2003

2004

2005

2006

2007

2008

2009

Figure 2B. Proportion of all endovascular repairs for intact abdominal aortic aneurysms performed by physician specialty from 2001-2009 in the Nationwide Inpatient Sample (totals sum to 100%)

100 78%

80 60

78%

VS (P < .001)

68%

GS (P < .001)

%

CS (P < .001)

40

36% 28%

20 0

18%

2001

2002

2003

2004

2005

2006

2007

2008

2009

Figure 3. Proportion of intact abdominal aortic aneurysms treated by endovascular repair within each specialty from 2001-2009 in the Nationwide Inpatient Sample.

174

Specialties Performing A Repair 100% 90%

14%

19%

80% 70%

33%

60% 44%

CS (P = .001)

50%

GS (P < .001)

40%

VS (P < .001)

30% 20%

53% 37%

10% 0% 2001

2002

2003

2004

2005

2006

2007

2008

2009

Figure 4A. Proportion of all open ruptured AAA repairs performed by physician specialty from 2001-2009 in the Nationwide Inpatient Sample (totals sum to 100%)

100% 90%

3%

0%

16%

80% 70% 60% 50%

IC + IR (P = .095) 54%

CS (P = .008)

40%

73%

30% 20% 10%

9

8%

18%

GS (P < .001) VS (P < .001)

28%

0% 2001

2002

2003

2004

2005

2006

2007

2008

2009

Figure 4B. Proportion of all EVARs for ruptured AAA repairs performed by physician specialty from 2001-2009 in the Nationwide Inpatient Sample (totals sum to 100%)

by CS with 22% of cases (24% to 19%, P <. 001, Figure 2A). VS also performed the majority of EVARs at 67%, (increasing from 60% to 67% from 2001 to 2009), followed by 16% performed by GS (19% to 17%, P < .001), 13% by CS (10.5% to 11.4%, P = .009) and 4% by IC and IR combined (10% to 6%, P = .015, Figure 2B). The absolute number of EVARs increased from 5,906 in 2001 (33% of the annual intact AAA repairs) to 16,252 in 2009 (78%). Over the same period, the number of open repairs decreased from 12,188 (67%) to 4,678 (22%). Consequently, EVAR has become the primary treatment method for intact AAA in all three surgical specialties (Figure 3). Since VS perform a greater proportion of endovascular procedures, the rise in EVAR utilization has in part led to VS performing an increasing majority of overall intact AAA repairs (54% in 2001 to 175

176

25.5%

90.4%

60.5%

95.5%

Female

White race

Teaching hospital

Urban location

75.1%

Large 21.0%

65.1%

26.7%

8.2%

88.2%

30.7%

90.7%

23.6%

71.9

21,625

GS

11.5%

71.5%

23.0%

5.5%

93.6%

45.1%

91.0%

23.0%

71

17,651

CS

<0.001

< .001

< .001

< .001

.051

< .001

< .001

P-value

3.0%

75.9%

15.5%

8.6%

95.4%

63.9%

91.5%

18.3%

73.7

72,489

VS

2.6%

73.5%

22.1%

4.4%

92.8%

33.8%

91.2%

16.9%

74.0

17,500

GS

2.0%

72.3%

19.5%

8.2%

93.9%

48.4%

92.2%

15.4%

72.8

14,033

CS

1.4%

77.6%

9.4%

13.0%

100%

46.3%

87.3%

16.8%

73.1

3,055

IC

Endovascular Repair

3.0%

75.6%

24.4%

0.0%

97.2%

41.0%

78.7%

17.0%

74.1

1,510

IR

< .001

< .001

< .001

< .001

< .001

< .001

< .001

P-value

EVAR

15.1%

71.8%

21.2%

7.0%

93.2%

49.7%

90.6%

24.5%

71.4

2.8%

75.1%

17.0%

7.9%

95.0%

56.2%

91.2%

17.7%

73.6

85,080 108,587

Overall

Open

EVAR, endovascular aneurysm repair; VS, vascular surgeons; GS, general surgeons; CS, cardiac surgeons; IC, interventional cardiologists; IR, interventional radiologists; P-values compare differences within the treatment groups

13.6%

17.9%

Medium

Emergent admission

7.0%

Small

Hospital Bedsize

71.3

45,804

Age (years)

Number

VS

Open Repair

< .001

< .001

< .001

< .001

< .001

< .001

< .001

P-value

Table I. Demographic and comorbidity characteristics of patients undergoing open aortic aneurysm repair or endovascular aneurysm repair per physician specialty.

Chapter 9

Specialties Performing A Repair 50

46%

VS (P < .001)

40

GS (P < .001) 30

%

25%

20 10 0

CS (P < .001)

23% 6% 5% 4% 2001

2002

2003

2004

2005

2006

2007

2008

2009

Figure 5. Proportion of ruptured abdominal aortic aneurysms treated by endovascular repair within each specialty from 2001-2009 in the Nationwide Inpatient Sample. Table II. Multivariable predictors for the likelihood of receiving EVAR (OR >1 predicts EVAR) Variable

Odds ratio

95% CI

P-value

Vascular surgeons

Reference

-

-

Cardiac surgeons

0.55

0.53 – 0.56

<.001

General surgeons

0.66

0.64 – 0.68

<.001

Emergent admission

0.14

0.14 – 0.15

<.001

Age (per 10 y)

1.42

1.40 – 1.44

<.001

Female sex

0.57

0.55 – 0.58

<.001

Non-white race

0.88

0.84 – 0.91

<.001

Year of surgery (per y)

1.33

1.32 – 1.34

<.001

Teaching hospital

1.27

1.25 – 1.30

<.001

Hospital location

1.33

1.27 – 1.40

<.001

9

Surgeon Specialty

Hospital Bed size Small

Reference

-

-

Medium

0.72

0.69 – 0.76

<.001

Large

0.98

0.94 – 1.02

<.319

66% in 2009, P < .001). Rupture AAA Repair VS performed 49% of open ruptured AAA repairs (increasing from 37% to 53%, P < .001, Figure 4A), followed by GS with 35% (44% to 33%, P < .001) and 16% by CS (19% to 14%, P = .001). With 73%, VS also carried out the majority of 177

Chapter 9 ruptured EVAR (rEVAR) (28% to 73%, P < .001, Figure 4B), while GS performed 15% (54% to 16%, P < .001) and CS 9% (18% to 8%, P = .008). IC and IR together are responsible for 3% of rEVARs (0% to 3%, P = .095). A dramatic overall increase in the utilization of rEVAR was observed (5% to 38% of the annual ruptured volume). This was most pronounced for VS, where the utilization of rEVAR went from 4% in 2001 to 46% in 2009 (P < .001, Figure 5). Likelihood of receiving EVAR Compared to treatment by VS, treatment by CS and GS was associated with a significantly lower likelihood of receiving EVAR (OR: 0.55, 95% CI: 0.53 – 0.56 for CS and OR: 0.66, 95% CI: 0.64 – 0.68 for GS, Table II). Additionally, women (OR: 0.57, 95% CI: 0.55 – 0.58) those with non-white race (OR: 0.88 95% CI: 0.84 – 0.91), and emergency admission (OR: 0.14, 95% CI: 0.14 – 0.15) were significantly less likely to undergo EVAR. Advanced age (OR: 1.42, 95% CI: 1.40 – 1.44, per 10 years), treatment in a teaching hospital (OR: 1.27, 95% CI: 1.25 – 1.30) and urban designation of the hospital (OR: 1.33, 95% CI: 1.27 – 1.40) were predictive of EVAR. Over the study period, the probability for receiving EVAR increased annually (OR: 1.33, 95% CI: 1.32 – 1.34, per year).

Discussion

The main finding of this study is that with the introduction of EVAR, the proportion of AAA repairs being performed by each physician specialty has changed. Vascular surgeons performed an increasing majority of both open and endovascular intact AAA repairs, while the proportion carried out by CS and GS has steadily declined. The distribution of specialties performing rAAA repair shifted from predominantly GS in the first years of the study towards VS in later years. Over the study period, EVAR has become the dominant treatment for intact AAA repair and is being utilized for an increasing number of ruptured AAA repairs as well. As EVAR is most likely performed by VS, the overall proportion of repairs done by VS –intact and ruptured– increased substantially with the widespread adoption of EVAR. Regarding intact AAA repair, our results are in line with a study by Birkmeyer et al., showing that between 1998 and 1999 intact open repairs were predominantly performed by VS. However, Birkmeyer et al. found a relatively even distribution with 39% being done by VS, 33% by CS and 28% by GS, while our results showed that VS already performed a majority of the open repairs in the early years of the study and this difference continued to increase over time. Regionalization of open AAA repairs to high-volume centers during the turn of the century is likely to have contributed to VS being increasingly responsible for AAA surgery.11 In addition, retirement of senior surgeons who were trained at a time when GS and CS customarily performed repair of abdominal aortic aneurysms and may have been less likely to obtain endovascular skills may have added to the shift towards VS. Similar to the early years of our study, previous reports show that GS performed the majority of ruptured AAA repairs at 39%, followed by VS at 33%, and CS at 29% before the introduction of EVAR.12 As EVAR became more widely used in the emergency setting, this distribution changed towards a growing proportion of emergency repairs being performed by VS. Since VS also 178

Specialties Performing A Repair performed an increasing proportion of open rAAA repairs separately from trends in endovascular repair, centralization of rAAA care is likely to have contributed to the shift from GS towards VS as well. Yet a persistent presence of GS treating open rAAA repair remained, which could be due to geographic location where the presence of VS may be lacking.13 In these areas, the emergent nature of a ruptured AAA may preclude the transfer of the patient to a center with VS necessitating immediate treatment by an available GS. The phenomenon of disruptive technologies in healthcare is not new. In coronary artery disease, the number of coronary revascularizations performed with coronary stenting rapidly increased after the first coronary stent was introduced in 1994, while the utilization of coronary bypass grafting declined.2 As a result, interventional cardiologists rather than cardiac surgeons currently perform the majority of coronary revascularizations. A similar shift was seen in vascular surgery with the introduction of carotid stenting. Before its introduction, carotid revascularization through endarterectomy was predominantly done by VS and to a lesser extent GS, CS and neurosurgeons.14 We noted that GS and CS practice included a substantial percentage of carotid endarterectomies (8.6% and 10.0% of the selected procedures we identified, respectively) (Supplemental Table II). However, we did not evaluate changes in carotid revascularization over time in this study. After FDA approval in 2004, carotid stenting is increasingly utilized with rapid adoption by not only surgeons, but also interventional radiologists and interventional cardiologists.15, 16 Currently, carotid endarterectomy use is still declining, while some patients are being treated through stenting predominantly by interventional cardiologists.17, 18 EVAR has certainly changed how AAA is being treated but contrary to the examples above, VS have only increased their role as the treating surgeon for intact and ruptured AAA Our study has several limitations that should be addressed. Since administrative databases were used, important clinical data such as anatomical information or hemodynamic status, which could influence the choice of procedure and subsequent outcomes, could not be assessed. Additionally, the difficulty of distinguishing a pre-existing comorbid condition from a post-operative complication in this dataset makes an adequate risk-adjustment model difficult. Therefore, we chose not to perform outcomes analysis using NIS. However, the NIS does afford national representation of all age groups thus making it an optimal source of epidemiologic data. Also, as the NIS database does not include the specialty of the attending physician, we employed an algorithm incorporating specialty specific procedures similar to what has been described before.7, 19, 20 Our algorithm for identifying vascular surgeons is arbitrary (75% vascular surgery cases), suggesting that some board certified VS may have been mistakenly classified as GS or vice versa. However, our methodology reflects what physicians were actually doing routinely in practice and focused on the change over time. Additionally, our algorithm generated similar distributions of physician’s specialty performing open repair in the early years of the current studied period when compared to a previously published large study using physician self identification of specialty.4 Further, procedures coded as open repairs performed by IC or IR (0.1%) were excluded from his study, as they are, most likely, miscoded endovascular procedures. Unfortunately, it is not possible to identify similar procedural coding errors for the remaining open 179

9

Chapter 9 repairs. Consequently, a very small proportion of miscoded procedures may have remained. Finally, the discontinuation of the surgeon-identifying variable after 2009 prevented the inclusion of more recent years. However, as is demonstrated in this study, the major shift from open repair to EVAR was already well established by 2009.21

Conclusions

Our results show that VS are performing an increasing majority of AAA repairs, in large part driven by the increased utilization of EVAR for both intact and ruptured AAA repair. However, GS and CS still perform AAA repair. Treatment by GS and CS, as well as emergent admission, female sex and non-white race are associated with a decreased likelihood of receiving EVAR. Advanced age, more recent year of surgery, treatment in a teaching hospital and urban designated area of the hospital increased the probability of receiving EVAR. Further studies should examine the implications of these national trends on the outcome of AAA repair.

180

Specialties Performing A Repair

References 1. Riskin DJ, Longaker MT, Gertner M, Krummel TM. Innovation in surgery: a historical perspective. Ann Surg. 2006;244(5):686-93. 2. Epstein AJ, Polsky D, Yang F, Yang L, Groeneveld PW. Coronary revascularization trends in the United States, 2001-2008. JAMA. 2011;305(17):1769-76. 3. Birkmeyer JD, Stukel TA, Siewers AE, Goodney PP, Wennberg DE, Lucas FL. Surgeon volume and operative mortality in the United States. N Engl J Med. 2003;349(22):2117-27. 4. Cronenwett JL BJ, editors. The Dartmouth atlas of vascular health care. Chicago: AHA Press; 2000. 5. Availability of Data Elements in the 1988-2012 National Inpatient Sample (NIS) [press release]. NIS Database Documentation: HCUP2013. 6. Csikesz NG, Simons JP, Tseng JF, Shah SA. Surgical specialization and operative mortality in hepato-pancreatico-biliary (HPB) surgery. J Gastrointest Surg. 2008;12(9):1534-9. 7. Schipper PH, Diggs BS, Ungerleider RM, Welke KF. The influence of surgeon specialty on outcomes in general thoracic surgery: a national sample 1996 to 2005. Ann Thorac Surg. 2009;88(5):1566-72; discussion 72-3. 8. Vogel TR, Dombrovskiy VY, Carson JL, Haser PB, Graham AM. Lower extremity angioplasty: impact of practitioner specialty and volume on practice patterns and healthcare resource utilization. J Vasc Surg. 2009;50(6):13204; discussion 4-5. 9. Dimick JB, Cowan JA, Jr., Stanley JC, Henke PK, Pronovost PJ, Upchurch GR, Jr. Surgeon specialty and provider volumes are related to outcome of intact abdominal aortic aneurysm repair in the United States. J Vasc Surg. 2003;38(4):739-44. 10. Liang P, Hurks R, Bensley RP, Hamdan A, Wyers M, Chaikof E, et al. The rise and fall of renal artery angioplasty and stenting in the United States, 1988-2009. J Vasc Surg. 2013;58(5):1331-8 e1. 11. Hill JS, McPhee JT, Messina LM, Ciocca RG, Eslami MH. Regionalization of abdominal aortic aneurysm repair: evidence of a shift to high-volume centers in the endovascular era. J Vasc Surg. 2008;48(1):29-36. 12. Cronenwett JL, Birkmeyer JD. The Dartmouth Atlas of Vascular Health Care. Cardiovasc Surg. 2000;8(6):409-10. 13. Maybury RS, Chang DC, Freischlag JA. Rural hospitals face a higher burden of ruptured abdominal aortic aneurysm and are more likely to transfer patients for emergent repair. J Am Coll Surg. 2011;212(6):1061-7. 14. Hollenbeak CS, Bowman AR, Harbaugh RE, Casale PN, Han D. The impact of surgical specialty on outcomes for carotid endarterectomy. J Surg Res. 2010;159(1):595-602. 15. Rosenfield K, Babb JD, Cates CU, Cowley MJ, Feldman T, Gallagher A, et al. Clinical competence statement on carotid stenting: training and credentialing for carotid stenting--multispecialty consensus recommendations: a report of the SCAI/SVMB/SVS Writing Committee to develop a clinical competence statement on carotid interventions. J Am Coll Cardiol. 2005;45(1):165-74. 181

9

Chapter 9 16. Gray WA. A cardiologist in the carotids. J Am Coll Cardiol. 2004;43(9):16025. 17. Nallamothu BK, Lu M, Rogers MA, Gurm HS, Birkmeyer JD. Physician specialty and carotid stenting among elderly medicare beneficiaries in the United States. Arch Intern Med. 2011;171(20):1804-10. 18. Skerritt MR, Block RC, Pearson TA, Young KC. Carotid endarterectomy and carotid artery stenting utilization trends over time. BMC Neurol. 2012;12:17. 19. Eppsteiner RW, Csikesz NG, Simons JP, Tseng JF, Shah SA. High volume and outcome after liver resection: surgeon or center? J Gastrointest Surg. 2008;12(10):1709-16; discussion 16. 20. Eslami MH, Csikesz N, Schanzer A, Messina LM. Peripheral arterial interventions: trends in market share and outcomes by specialty, 1998-2005. J Vasc Surg. 2009;50(5):1071-8. 21. Dua A, Kuy S, Lee CJ, Upchurch GR, Jr., Desai SS. Epidemiology of aortic aneurysm repair in the United States from 2000 to 2010. J Vasc Surg. 2014;59(6):1512-7.

182

Specialties Performing A Repair Supplemental Table I. Proportion of the cohort per state State

Proportion of cohort (%)

Arkansas

1.1%

Arizona

5.3%

Colorado

2.0%

Florida

17.6%

Georgia

0.4%

Iowa

2.1%

Kansas

0.7%

Kentucky

3.2%

Maryland

4.6%

Maine

0.1%

Michigan

4.2%

Minnesota

0.9%

Missouri

5.9%

Montana

0.1%

Nebraska

0.9%

New Hampshire

2.0%

New Jersey

5.8%

Nevada

0.8%

New York

14.4%

Oregon

0.6%

Pennsylvania

5.8%

Rhode Island

0.2%

South Carolina

1.4%

South Dakota

0.1%

Tennessee

5.4%

Texas

8.9%

Virginia

4.3%

Washington

1.4%

West Virginia

<0.1%

Wyoming

<0.1%

9

183

Chapter 9 Supplemental Table II. Procedures used to identify physician specialty Vascular Surgeon ICD-9

Description

38.12

Carotid endarterectomy

39.29

Peripheral vascular bypass

84.15

Below knee amputation

84.17

Above knee amputation

General Surgeon ICD-9

Description

17.11-24; 53.00-9

Hernia repair

47.01-19

Appendectomy

51.21-24

Cholecystectomy

Cardiac Surgeon ICD-9

Description

36.10-19

Coronary artery bypass grafting

35.20-28

Heart valve replacement

39.61

Cardiopulmonary bypass

Interventional Cardiologist ICD-9

Description

00.66; 36.01-02; 36.05

Percutaneous transmural coronary angioplasty

36.04

Intracoronary thrombolysis

36.06-07

Intracoronary stenting

37.21-23

Heart catheterization

Interventional Radiologist ICD-9

184

Description

33.26

Closed lung biopsy

39.1

Transjugular intrahepatic portosystemic shunt

50.11

Closed liver biopsy

55.03-04

Percutaneous nephrostomy

78.49; 81.65

Percutaneous vertebroplasty

99.25

Chemoembolization

Specialties Performing A Repair Supplemental Table III. Top 15 procedures performed per physician specialty Vascular Surgeon ICD-9

%

Description

38.12

17.6%

Carotid endarterectomy

39.29

10.0%

Peripheral vascular bypass

39.5

7.7%

Angioplasty or atherectomy of other non-coronary vessel(s)

39.71

5.2%

Endovascular implantation of graft in abdominal aorta

38.7

3.4%

Interruption of the vena cava

39.49

3.3%

Revision of anastomosis of blood vessel or vascular procedure

39.27

3.0%

Arteriovenostomy for renal dialysis

38.44

2.8%

Resection of vessel with replacement, aorta, abdominal

86.22

2.5%

Excisional debridement of wound, infection, or burn

39.25

2.4%

Aorta-iliac-femoral bypass

84.15

2.2%

Other amputation below knee

84.17

2.2%

Amputation above knee

38.95

2.1%

Venous catheterization for renal dialysis

38.18

1.8%

Endarterectomy, lower limb arteries

84.11

1.7%

Amputation of toe

9

General Surgeon ICD-9

%

Description

51.23

10.2%

Laparoscopic cholecystectomy

38.12

8.6%

Carotid endarterectomy

39.29

4.1%

Peripheral vascular bypass

47.09

3.8%

Other appendectomy

47.01

3.1%

Laparoscopic appendectomy

45.73

2.5%

Open and other right hemicolectomy

38.93

2.5%

Venous catheterization

86.22

2.4%

Excisional debridement of wound, infection, or burn

45.76

2.1%

Open and other sigmoidectomy

39.5

2.0%

Angioplasty or atherectomy of other non-coronary vessel(s)

51.22

1.9%

Cholecystectomy

39.27

1.8%

Arteriovenostomy for renal dialysis

38.7

1.7%

Interruption of vena cava

86.04

1.6%

Other incision with drainage of skin and subcutaneous tissue

54.59

1.6%

Other lysis of peritoneal adhesions

Cardiac Surgeon

185

Chapter 9 ICD-9

%

Description

38.12

10.0%

Carotid endarterectomy

36.12

9.1%

(Aorto)coronary bypass of two coronary arteries

36.13

8.9%

(Aorto)coronary bypass of three coronary arteries

36.14

4.2%

(Aorto)coronary bypass of four or more coronary arteries

36.11

3.5%

(Aorto)coronary bypass of one coronary artery

36.15

3.2%

Single internal mammary-coronary artery bypass

39.29

3.1%

Peripheral vascular bypass

35.22

2.9%

Other replacement of aortic valve

35.21

2.9%

Replacement of aortic valve with tissue graft

38.44

1.7%

Resection of vessel with replacement, aorta, abdominal

32.4

1.7%

Lobectomy of lumg

39.71

1.5%

Endovascular implantation of graft in abdominal aorta

39.5

1.3%

Angioplasty or atherectomy of other non-coronary vessel(s)

32.29

1.3%

Other local excision or destruction of lesion

35.12

1.1%

Open heart valvuloplasty of mitral valve without replacement

Interventional Cardiologist ICD-9

%

Description

0.66

21.4%

Percutaneous transmural coronary angioplasty

37.22

13.7%

Left heart cardiac catheterization

39.5

11.7%

Angioplasty or atherectomy of other non-coronary vessel(s)

36.01

11.6%

Percutaneous transmural coronary angioplasty

39.71

2.5%

Endovascular implantation of graft in abdominal aorta

36.05

2.3%

Percutaneous transmural coronary angioplasty

0.61

2.1%

Percutaneous angioplasty or atherectomy of precerebral extracranial vessel(s)

37.23

2.1%

Combined right and left heart cardiac catheterization

37.72

1.9%

Initial insertion of transvenous leads [electrodes] into atrium and ventricle

37.61

1.2%

Implant of pulsation balloon

88.72

1.1%

Diagnostic ultrasound of heart

64

1.0%

Circumcision

57.94

0.9%

Insertion of indwelling catheter

35.52

0.9%

Repair of atrial septal defect with prosthesis, closed technique

88.56

0.7%

Coronary arteriography using two catheters

Interventional Radiologist ICD-9

186

%

Description

Specialties Performing A Repair 38.9

18.3%

Venous catheterization, not elsewhere classified

39.5

11.9%

Angioplasty or atherectomy of other non-coronary vessel(s)

38.7

7.0%

Interruption of the vena cava

54.91

6.8%

Percutaneous abdominal drainage

34.91

6.1%

Thoracentesis

38.95

2.9%

Venous catheterization for renal dialysis

55.03

2.4%

Percutaneous nephrostomy without fragmentation

50.11

2.1%

Closed (percutaneous) [needle] biopsy of liver

88.41

2.1%

Arteriography of cerebral arteries

39.71

2.0%

Endovascular implantation of graft in abdominal aorta

33.26

2.0%

Closed (percutaneous) [needle] biopsy of lung

81.66

1.7%

Percutaneous vertebral augmentation

34.04

1.3%

Insertion of intercostal catheter for drainage

88.42

1.3%

Aortography

99.29

1.2%

Injection or infusion of other therapeutic or prophylactic substance

9

187

188 25.2% 19.4% 3.1%

General surgeons

Cardiac surgeons

IC + IR

24.7% 23.7%

General surgeons

Cardiac surgeons

19.4% 10.5% 10.4%

General surgeons

Cardiac surgeons

IC + IR

44.1% 19.3%

General surgeons

Cardiac surgeons

18.3%

Cardiac surgeons 0%

54.1%

General surgeon

IC + IR

27.5%

Vascular surgeons

Ruptured EVAR

36.6%

Vascular surgeons

Ruptured Open

59.6%

Vascular surgeons

Intact EVAR

51.6%

Vascular surgeons

Intact Open

52.4%

2001

Vascular surgeons

Overall

0%

10.5%

24.9%

64.6%

16.0%

34.0%

50.0%

4.1%

14.0%

18.8%

63.2%

21.5%

26.9%

51.6%

1.7%

17.9%

24.1%

56.4%

2003

7.1%

4.9%

11.0%

77.0%

12.5%

27.6%

59.9%

2.1%

8.2%

13.8%

75.9%

19.1%

22.7%

58.2%

1.1%

13.2%

18.6%

67.1%

2004

1.9%

9.6%

12.7%

75.8%

13.2%

39.0%

47.8%

3.1%

14.6%

17.2%

65.1%

21.9%

27.0%

51.2%

1.7%

17.3%

22.5%

58.6%

2005

1.2%

3.3%

14.4%

81.2%

15.4%

36.7%

47.9%

2.5%

14.7%

12.7%

70.0%

20.4%

22.8%

56.9%

1.6%

16.2%

16.9%

65.2%

2006

IC, interventional radiologists; IR, interventional radiologists

3.7%

22.4%

33.6%

40.2%

17.4%

39.0%

43.6%

4.1%

14.2%

20.1%

61.6%

21.9%

26.2%

51.9%

1.5%

18.9%

25.4%

54.3%

2002

2.8%

12.9%

9.9%

74.5%

19.4%

28.6%

52.0%

3.7%

14.0%

14.6%

67.7%

22.0%

18.2%

59.8%

2.5%

16.2%

16.1%

65.1%

2007

Supplemental Table IV. Annual proportions of AAA repairs performed per specialty

3.9%

10.4%

6.3%

79.4%

14.3%

31.4%

54.3%

5.1%

14.0%

17.0%

63.8%

24.7%

19.1%

56.3%

3.8%

16.1%

17.9%

62.2%

2008

3.4%

7.5%

15.8%

73.3%

13.9%

33.3%

52.8%

5.5%

11.3%

16.7%

66.5%

18.9%

16.0%

65.1%

4.0%

12.9%

17.2%

65.8%

2009

.095

.008

< .001

< .001

.001

< .001

< .001

.015

.009

< .001

< .001

.001

< .001

< .001

< .001

< .001

< .001

< .001

p-value

Chapter 9

Specialties Performing A Repair

9

189

CHAPTER 10

The impact of Endovascular Treatment of Isolated Iliac Artery Aneurysms Dominique B Buck, Rodney P Bensley, Jeremy Darling, Thomas Curran, John C McCallum, Frans L Moll, Joost A van Herwaarden, Marc L Schermerhorn. Accepted J Vasc Surg.

Chapter 10

Abstract Objective Isolated Iliac artery aneurysms are rare, but potentially fatal. The impact of recent trends in the utilization of endovascular iliac aneurysm repair (EVIR) on isolated iliac artery aneurysm-associated mortality is unknown. Methods We identified all patients with a primary diagnosis of iliac artery aneurysm in the NIS from 1988 to 2011. We examined trends in management (open vs. EVIR, elective and urgent) and overall isolated iliac artery aneurysm related deaths (with or without repair). We compared in-hospital mortality and complications for the subgroup of patients undergoing elective open and EVIR from 2000-2011. Results We identified 33,161 patients undergoing isolated iliac artery aneurysm repair from 1988-2011: of which there were 9,016 EVIR and 4,933 open elective repairs from 2000-2011. Total repairs increased after introduction of EVIR from 28 to 71 per 10M US population (P<.001). EVIR surpassed open repair in 2003. Total isolated iliac artery aneurysm-related deaths, due to rupture or elective repair, decreased after the introduction of EVIR (4.4 to 2.3 per 10M US population, P<.001). However, urgent admissions have not decreased over this time period (15 to 15 procedures per 10M US population, P=0.30). Among elective repairs after 2000, EVIR patients were older (72.4 vs. 69.4 years, P=0.002) and were more likely to have a history of prior MI (14.0% vs. 11.3%, P<.001) and renal failure (7.2% vs. 3.6%, P<.001). Open repair had significantly higher in-hospital mortality (1.8% vs. 0.5%, P<0.001) and complications (17.9% vs. 6.7%, P<0.001), and a longer length of stay (6.7 vs. 2.3 days, P<0.001). Conclusions Treatment of isolated iliac artery aneurysm has increased since the introduction of EVIR and is associated with lower perioperative mortality, despite a higher burden of comorbid illness. Decreasing iliac artery aneurysm-attributable inhospital deaths are likely related primarily to lower elective mortality with EVIR rather than rupture prevention.

192

Isolated Iliac Artery Aneurysms

Introduction

Comprising approximately 2% of all abdominal aneurysms,1-4 isolated iliac artery aneurysms are uncommon, are frequently asymptomatic, and are most often discovered incidentally.2, 5, 6 When ruptured, however, these aneurysms carry significant risk of mortality.2, 3, 7 Historically, open repair has been the primary treatment, however, given the pelvic location, this may be technically challenging and lead to complications. 2, 3 Therefore, treatment with endovascular techniques has become particularly appealing with good early- and mid-term results.1, 2, 8-12 Reported operative mortality rates with elective open and endovascular repair are 3-6% and 0-1%, respectively.1-4, 6, 9, 13, 14 However, the majority of these studies are from single institutions with small sample sizes The purpose of this national study is to identify epidemiologic trends in management and mortality for isolated iliac artery aneurysm and to evaluate the impact of the introduction of endovascular iliac aneurysm repair (EVIR) in the United States.

Methods Dataset All patients with an isolated iliac artery aneurysm from 1988 through 2011 were extracted from the Nationwide Inpatient Sample (NIS) database. NIS is the largest US all-payer inpatient database and has been assembled as part of the Healthcare Cost and Utilization Project (HCUP). NIS represents a 20% stratified sample of all payer (insured and uninsured) hospitalizations and represents approximately eight million hospitalizations per year. These data contain sampling weights to approximate U.S. population estimates. All of our analyses were performed using weighted data. This study contained de-identified data only without any protected health information and is therefore not subject to patient consent or Institutional Review Board approval. Patients Using International Classification of Diseases, edition 9 (ICD9) diagnosis code 442.2, we identified all patients with a primary diagnosis (reason for admission) of iliac artery aneurysm undergoing open surgical repair (ICD9 procedure codes 38.06, 38.16, 38.36, 38.46, 38.66, 38.86, 39.25, 39.52) or EVIR (39.71, 39.79, 39.90) in the Nationwide Inpatient Sample from 1988-2011. We excluded patients with concomitant diagnoses of aortic aneurysms (ICD9 diagnosis codes 441.3, 441.4, 441.9) as well as those with thoracic aneurysm and/or aortic dissection (ICD9 diagnosis codes 441.0, 441.1, 441.2, 441.5, 441.6, 441.7). Age, sex, race, and co-morbid conditions were documented, including diabetes mellitus (ICD9 diagnosis code 250.*), chronic obstructive pulmonary disease (ICD9 diagnosis codes 491.*, 492.*, 496.), congestive heart failure (ICD9 diagnosis code 428.*), prior myocardial infarction (ICD9 diagnosis code 412.), and hypertension (ICD9 diagnosis codes 401.*, 402.*, 403.*, 404.*, 405.*). Procedures were categorized as either elective or urgent. A specific ICD 9 code for ruptured iliac aneurysm 193

10

Chapter 10 does not exist. Outcomes We examined trends in total repairs and overall isolated iliac artery aneurysm related deaths per ten million citizens of the Unites States, including the percentage of EVIR vs. open repairs. We analyzed elective repairs, urgent repairs, and hospitalizations for isolated iliac artery aneurysm without repairs resulting in death. We compared in-hospital deaths and postoperative complications after elective open repair and EVIR from 2000-2011, to create contemporaneous, comparable cohorts. Our primary outcome was mortality. We also examined post-operative complications, including cardiac complications, respiratory complications, postoperative wound infection, and hematoma. We also analyzed discharge status, and length of stay. Statistical Analysis Statistical analysis was completed with SPSS statistical software (version 20; IBM Corp, Armonk, NY). When appropriate, continuous variables were compared using two-tailed independent samples t-test and categorical variable were compared using Chi-squared. A Cochran-Armitage test for linear trend was performed to assess for changes in death and repairs over time. Statistical significance was defined as p < 0.05.

Results Epidemiologic Trends We identified 33,161 patients undergoing isolated iliac artery aneurysm repair, including elective and urgent procedures, from 1988 to 2011. Endovascular repairs increased steadily over time and surpassed open repair in 2003. (Figure 1) The overall rate of repair increased after introduction of EVIR from 28 to 71 repairs per 10 million of the US population (P<.001). Total deaths, including non-operative deaths, decreased after the introduction of EVIR, from 4.4 to 2.3 deaths per 10 million of the US population (P<.001). Deaths due to isolated iliac artery aneurysm over time are demonstrated in Figure 2. Of all deaths in 2011, 73% were after open repair, despite the fact that open repair only made up only 20% of total isolated iliac artery aneurysm repairs in 2011. In-hospital mortality with open surgery and EVIR over time is shown in Figure 3. Operative mortality, for elective and urgent repairs, decreased from 13.4% to 2.4%, with an overall death rate of 1.8% in 2011. The number of urgent procedures, shown in Figure 4, has remained stable over time (from 15 to 15 procedures per 10 million of the US population, P=0.30). Elective EVIR vs. Open Repair 2000-2011 We identified 13,949 patients who underwent elective repair from 2000 to 2011 after excluding 5,426 patients who underwent urgent repair; there were 4,933 (35%) open surgical repairs and 9,016 (65%) EVIR. For elective repairs, EVIR patients were older (72.4 vs. 69.4 years, P=.002), and were more likely to have a 194

Isolated Iliac Artery Aneurysms

Figure 1. Elective and urgent procedures over time, per 10 million U.S. population.

10

Figure 2. Elective and urgent deaths over time, per 10 million US population.

195

Chapter 10

Figure 3. In-hospital mortality for all isolated iliac aneurysm repairs (elective and urgent) over time.

Figure 4. Elective and urgent procedures over time, per 10 million US population.

196

Isolated Iliac Artery Aneurysms Table I. Baseline characteristics for elective endovascular and open isolated Iliac Artery Aneurysm Repair from 2000-2011 Pre-operative Comorbidities Age Coronary Artery Disease

EVIR N=9016

Open repair N=4933

P-value

72

69

0.002

43%

37%

<.001

Hypertension

64%

57%

<.001

Dysrhythmia

4.3%

7.2%

<.001

Prior MI

14%

11%

<.001

Heart Failure

8.8%

7.3%

0.003

COPD

19%

22%

0.002

Afib

13%

13%

0.111

Acute Renal Failure

1.8%

4.8%

<.001

Chronic Renal Failure

7.2%

3.6%

<.001

CVD

2.1%

2.5%

0.202

PVD

24%

24%

0.722

Hyperlipidemia

39%

31%

<.001

Table II. Post-operative Outcomes following elective endovascular and open repair of isolated Iliac artery aneurysms from 2000-2011. EVIR N=9016

Open repair N=4933

P-value

In-hospital Mortality

0.5%

1.8%

<.001

Cardiac complications

1.2%

3.3%

<.001

Respiratory complications

1.4%

9.2%

<.001

Peripheral vascular complications

1.0%

1.4%

0.022

Wound dehiscence

0.2%

0.8%

<.001

Post-operative Outcomes

Bleeding complications

3.2%

5.7%

<.001

Infection

0.3%

1.0%

<.001

Post-operative complications

6.7%

17.9%

<.001

2.3

6.7

<.001

Length of Stay (days)

10

197

Chapter 10 history of MI (14.0% vs. 11.3%, P<.001) and CRF (7.2% vs. 3.6%, P<.001) than patients undergoing open repair. (Table 1) Patients undergoing elective open repair had higher in-hospital mortality than EVIR (1.8% vs. 0.5%, P<.001). (Table 2) For urgent procedures from 2000 to 2011 the in-hospital mortality was 7.5% for open repair vs. 1.1% for EVIR (P<.001) In addition, elective open repair had higher rates of overall postoperative complications (17.9% vs. 6.7%, P<.001): including cardiac complications (3.3% vs. 1.2%, P<.001), respiratory complications (9.2% vs. 1.4%, P<.001), wound dehiscence (0.8% vs. 0.2%, P<.001 ), and wound infections (10.% vs. 0.3%, P<.001). Patients undergoing open repair had a median length of stay of 6.7 days versus 2.3 days (P<.001) in the EVIR group.

Discussion

We studied trends over the period 1988-2011 and found that the treatment of isolated iliac artery aneurysm has dramatically increased over time since the introduction of EVIR. EVIR is currently the dominant treatment method for isolated iliac artery aneurysm and is associated with lower morbidity, mortality and shorter length of stay, despite more comorbid illness. Deaths from isolated iliac artery aneurysm have decreased over time, despite the increase in total repairs. Similar to the trend seen in abdominal aortic aneurysm (AAA) repair,15, 16 treatment of isolated IAA has shifted away from open surgical repair towards a less invasive endovascular technique. The total number of isolated iliac artery aneurysm repairs has increased with the introduction of EVIR, likely in part due to increased detection related to increased utilization of abdominal imaging over time.17 However, unlike what was seen with AAA, the increase in elective procedures may not have lead to a decrease in urgent procedures over time. This suggests that the increase in elective procedures may be primarily related to a lower threshold for intervention with availability of a less invasive treatment. The decrease in isolated iliac artery aneurysm in-hospital deaths over time is likely related to the reduction in elective mortality, similar to trends observed in AAA treatment.15, 16 However, there may not be a reduction in deaths due to rupture prevention. Unfortunately, unlike AAA, there is no specific ICD-9 code for rupture and we can only extrapolate from the number of emergent admissions. Currently, the existing guidelines recommend intervention at a diameter of >3cm for iliac artery aneurysms.18 A study by Santilli et al and a recent survey study by Williams et al. suggest that it is safe to wait until the common iliac diameter is 4cm before intervention.19, 20 However, our data do not include diameter nor do they distinguish the common iliac artery from the hypogastric iliac artery. Therefore they do not support treatment of smaller iliac aneurysms. They simply reflect the impact of current utilization trends. Prior studies have demonstrated 30-day mortality-rates ranging between 4-6% for open repair and 0-2% for endovascular repair.1, 2, 4, 21 We found comparable mortality-rates of 1.8% for open repair and 0.5% for endovascular repair; however, the NIS contains data for only in-hospital mortality instead of 30-day mortality, which was reported in the prior publications. In a study of abdominal 198

Isolated Iliac Artery Aneurysms aneurysm repair using Medicare data we found an in-hospital mortality of 4.6% and 1.1% and a 30-day mortality of 4.8% and 1.6% for open and endovascular repair respectively. Thus, 96% of 30-day deaths after open repair occurred during the initial inpatient stay compared to only 69% of EVAR deaths. We expect this overestimation of the benefit of endovascular repair using inpatient mortality alone to be present to a similar extent in isolated iliac artery aneurysm patients.22 Prior literature shows hospital stays of 2-3 days for endovascular patients, and 5-9 days for open repair patients similar to our findings.2, 3, 9 An important caveat is that our study is not a randomized controlled trial, but an observational study of administrative data, which is subject to coding errors and selection bias. To improve the comparison of open surgery and EVIR for the more recent years (2000-2011), we excluded all patients who underwent an urgent procedure since these patients were more likely to have an open repair and were more likely to have adverse outcomes. Administrative studies lack anatomic and clinical details, which could impact patient selection and outcomes. As noted, our study lacks data regarding iliac artery diameters, and we cannot distinguish hypogastric from common iliac arteries. We were unable to distinguish the precise type of endovascular repair due to lack of specificity of ICD 9 codes. Additionally, this database lacks long-term follow-up data including reinterventions, late rupture, and buttock claudication. However, late ruptures that present to the hospital would potentially be captured as another admission. Despite these limitations the strengths of this database are the large numbers with national representation and the ability to demonstrate national trends in utilization and mortality.

Conclusions

This study demonstrates that repair of isolated iliac artery aneurysm across the U.S. is increasing. Overall deaths due to isolated iliac artery aneurysms are decreasing, despite the increase in repair rates, related to the increased use of EVIR with its lower operative mortality. It remains to be seen whether there has been any decrease in deaths due to rupture prevention.

199

10

Chapter 10

References 1. Huang Y, Gloviczki P, Duncan AA, Kalra M, Hoskin TL, Oderich GS, et al. Common iliac artery aneurysm: expansion rate and results of open surgical and endovascular repair. J Vasc Surg. 2008;47(6):1203-10; discussion 101. 2. Patel NV, Long GW, Cheema ZF, Rimar K, Brown OW, Shanley CJ. Open vs. endovascular repair of isolated iliac artery aneurysms: A 12-year experience. J Vasc Surg. 2009;49(5):1147-53. 3. Chaer RA, Barbato JE, Lin SC, Zenati M, Kent KC, McKinsey JF. Isolated iliac artery aneurysms: a contemporary comparison of endovascular and open repair. J Vasc Surg. 2008;47(4):708-13. 4. Sandhu RS, Pipinos, II. Isolated iliac artery aneurysms. Seminars in vascular surgery. 2005;18(4):209-15. 5. Scheinert D, Schroder M, Steinkamp H, Ludwig J, Biamino G. Treatment of iliac artery aneurysms by percutaneous implantation of stent grafts. Circulation. 2000;102(19 Suppl 3):III253-8. 6. Chemelli A, Hugl B, Klocker J, Thauerer M, Strasak A, Jaschke W, et al. Endovascular repair of isolated iliac artery aneurysms. J Endovasc Ther. 2010;17(4):492-503. 7. Uberoi R, Tsetis D, Shrivastava V, Morgan R, Belli AM, Subcommittee on Reporting Standards for Arterial Aneurysms of The Society for Vascular S. Standard of practice for the interventional management of isolated iliac artery aneurysms. Cardiovascular and interventional radiology. 2011;34(1):3-13. 8. Boules TN, Selzer F, Stanziale SF, Chomic A, Marone LK, Dillavou ED, et al. Endovascular management of isolated iliac artery aneurysms. J Vasc Surg. 2006;44(1):29-37. 9. Pitoulias GA, Donas KP, Schulte S, Horsch S, Papadimitriou DK. Isolated iliac artery aneurysms: endovascular versus open elective repair. J Vasc Surg. 2007;46(4):648-54. 10. Stroumpouli E, Nassef A, Loosemore T, Thompson M, Morgan R, Belli AM. The endovascular management of iliac artery aneurysms. Cardiovascular and interventional radiology. 2007;30(6):1099-104. 11. Kim MD, Lee DY, Lee M, Won JY, Lee SJ, Kim IJ, et al. Single-center experience in the endovascular management of isolated iliac artery aneurysm. Acta radiologica. 2013. 12. Fossaceca R, Guzzardi G, Di Terlizzi M, Divenuto I, Cerini P, Malatesta E, et al. Long-term efficacy of endovascular treatment of isolated iliac artery aneurysms. La Radiologia medica. 2013;118(1):62-73. 13. Dorigo W, Pulli R, Troisi N, Alessi Innocenti A, Pratesi G, Azas L, et al. The treatment of isolated iliac artery aneurysm in patients with non-aneurysmal aorta. European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery. 2008;35(5):585-9. 14. Ferreira J, Canedo A, Brandao D, Maia M, Braga S, Chaparro M, et al. Isolated iliac artery aneurysms: six-year experience. Interactive cardiovascular and thoracic surgery. 2010;10(2):245-8. 15. Giles KA, Pomposelli F, Hamdan A, Wyers M, Jhaveri A, Schermerhorn 200

Isolated Iliac Artery Aneurysms ML. Decrease in total aneurysm-related deaths in the era of endovascular aneurysm repair. J Vasc Surg. 2009;49(3):543-50; discussion 50-1. 16. Schermerhorn ML, Bensley RP, Giles KA, Hurks R, O’Malley A J, Cotterill P, et al. Changes in abdominal aortic aneurysm rupture and short-term mortality, 1995-2008: a retrospective observational study. Annals of surgery. 2012;256(4):651-8. 17. Kocher KE, Meurer WJ, Fazel R, Scott PA, Krumholz HM, Nallamothu BK. National trends in use of computed tomography in the emergency department. Annals of emergency medicine. 2011;58(5):452-62 e3. 18. Moll FL, Powell JT, Fraedrich G, Verzini F, Haulon S, Waltham M, et al. Management of abdominal aortic aneurysms clinical practice guidelines of the European society for vascular surgery. European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery. 2011;41 Suppl 1:S1-S58. 19. Williams SK, Campbell WB, Earnshaw JJ. Survey of management of common iliac artery aneurysms by members of the Vascular Society of Great Britain and Ireland. Annals of the Royal College of Surgeons of England. 2014;96(2):116-20. 20. Santilli SM, Wernsing SE, Lee ES. Expansion rates and outcomes for iliac artery aneurysms. J Vasc Surg. 2000;31(1 Pt 1):114-21. 21. Antoniou GA, Nassef AH, Antoniou SA, Loh CY, Turner DR, Beard JD. Endovascular treatment of isolated internal iliac artery aneurysms. Vascular. 2011;19(6):291-300. 22. Schermerhorn ML, Giles KA, Sachs T, Bensley RP, O’Malley AJ, Cotterill P, et al. Defining perioperative mortality after open and endovascular aortic aneurysm repair in the US Medicare population. Journal of the American College of Surgeons. 2011;212(3):349-55.

201

10

CHAPTER 11

Isolated Renal Artery Aneurysms: Management and Outcomes in the Endovascular Era.

Dominique B Buck, Thomas Curran, John C McCallum, Jeremy Darling, Rishi Mamtani, Joost A van Herwaarden, Frans L Moll, Marc L Schermerhorn. Revisions J Vasc Surg.

Chapter 11

Abstract Objective Isolated renal artery aneurysms are rare and controversy remains about indications for surgical repair. Little is known about the impact of endovascular therapy on patient selection and outcomes of renal artery aneurysms. Methods We identified all patients undergoing open or endovascular repair of isolated renal artery aneurysms in the Nationwide Inpatient Sample (NIS) from 1988 to 2011 for epidemiologic analysis. Elective cases were selected from the period of 2000 to 2011, to create comparable cohorts for outcome comparison. We identified all patients with a primary diagnosis of renal artery aneurysms undergoing open surgery (reconstruction or nephrectomy) or endovascular repair (coil or stent). Patients with a concomitant aortic aneurysms or dissections were excluded. We evaluated patient characteristics, management, and in-hospital outcomes for open and endovascular repair, and we examined changes in management and outcomes over time. Results We identified 6,234 renal artery aneurysm repairs between 1988 and 2011. Total repairs increased after the introduction of endovascular repair (6.2 in 1988 to 10.5 in 2011 per 10million(M) US population, P=0.002). Endovascular repair increased from 0 in 1988 to 4.8 in 2011 per 10M US population (P<.0001). However, there was no concomitant decrease in open surgery (4.0 in 1988 to 5.6 in 2011 per 10M US population, P=0.071). From 2000-2011 there were 1,627 open and 1,082 endovascular elective repairs. Patients undergoing endovascular repair were more likely to have a history of coronary artery disease (18% vs. 11%, P<0.001), prior myocardial infarction (5.2 vs. 1.8%, P<0.001) and renal failure (7.7 vs. 3.3%, P<0.001). In-hospital mortality was 1.8% for endovascular and 0.9% for open reconstruction (P=0.037), and 5.4% for nephrectomy (P<.001 compared to all revascularization). Complication rates were 12.4% for open repair vs 10.5% for endovascular repair (P=0.134), including more cardiac (2.2 vs. 0.6%, P=0.001) and peripheral vascular complications (0.6 vs. 0.0%, P=0.014) with open repair. Open repair had a longer length of stay (6.0 vs. 4.6 days, P<0.001). After adjustment for other predictors of mortality, age (OR 1.05 per decade, 95% CI 1.0-1.1, P=0.001), heartfailure (OR 7.0, 95% CI 3.1-16.0, P<.001) and dysrhythmia (OR 5.9, 95% CI 2.0-16.8, P=0.005), Endovascular repair was still not protective (OR 1.6, 95% CI 0.8-3.2, P=0.145). Conclusion More renal artery aneurysms are being treated with the advent of endovascular techniques, without a reduction in operative mortality or a reduction in open surgery. Indications for repair of renal artery aneurysms should be evaluated.

204

Isolated Renal Artery Aneurysms

Introduction

Isolated renal artery aneurysms are rare, with an estimated incidence cited between 0.1 and 1.3% in the general population.1-4 The natural history of these aneurysms is uncertain, and therefore the indications for surgical intervention remain controversial.5, 6 The majority of patients are asymptomatic at time of discovery of a renal artery aneurysms.4, 7 The increased rate of incidental discovery is largely attributable to the use of non-invasive imaging for evaluation of other conditions.6 Potential complications of renal artery aneurysms include rupture, distal embolization, infarction, hypertension, dissection, renal failure, and arteriovenous fistula.8, 9 The mortality rate of rupture is reported to be as high as 80%.10 Generally, intervention is undertaken for aneurysm diameter > 2.0 cm, ruptured aneurysm, dissection, localized symptoms, female gender within childbearing years, and renovascular hypertension.6, 11-15 Before the advent of endovascular techniques, open repair was the conventional method of treatment, with techniques including excision with primary repair or patch angioplasty, excision with reconstruction using bypass, extracorporeal reconstruction with autotransplantation, and nephrectomy.13, 16 The less invasive alternative, endovascular repair, includes coil embolization or the use of a stent graft. The technical success and safety of the endovascular treatment for renal artery aneurysms in experiences hands have been demonstrated, with a low morbidity and mortality.17-21 The purpose of this population based retrospective study is to evaluate the impact of the introduction of endovascular treatment on total repair rates in the United States, and compare mortality and morbidity outcomes between endovascular and open surgical repair.

Methods Dataset All patients who had an isolated renal artery aneurysm in the period from 1988 to 2011 were extracted from the Nationwide Inpatient Sample (NIS). The NIS is the largest US all-payer inpatient database, and has been collected as part of the Healthcare Cost and Utilization Project (HCUP). It represents 20% of U.S. hospitalizations, and contains sampling weights to approximate total U.S. population estimates. The NIS database contains de-identified data only without any protected health information. Therefore Institutional Review Board approval and patient consent were waived. Patients Using the International Classification of Diseases, edition 9 (ICD9) code 442.1, we identified all patients with a primary diagnosis of renal artery aneurysm undergoing open or endovascular repair in the NIS from 1988 to 2011. We excluded patients with concomitant diagnoses of aortic aneurysms, as well as those with a thoracic aneurysm and/or aortic dissection (ICD9 441, 441.0, 441.1, 441.2, 441.3, 441.4, 205

11

Chapter 11 441.5, 441.6, 441.7, 441.9, 441.00, 441.01, 441.02, 441.03). The open repair cohort consisted of patients who underwent either nephrectomy (ICD9 554, 555.1, 555.2) or an open reconstruction (ICD9 380.6, 381.6, 383.6, 384.6, 386.6, 388.6, 392.4, 392.6, 395.0, 395.2, 395.5) and the endovascular cohort consisted of patients who had a coil embolization (ICD9 397.9) or a stent placement (ICD9 399.0, 397.1). We documented demographics including age, sex, race, and comorbid conditions including coronary artery disease, hypertension, dysrhythmia, prior myocardial infarction, heart failure, chronic obstructive pulmonary disease, atrial fibrillation, acute renal failure, chronic renal failure, and peripheral vascular disease. Outcomes We examined trends in management of total repairs and proportion of open vs. endovascular repairs during our entire study period (1988-2011). However, for our outcomes comparison, we limited our analysis to elective repairs during the years 2000 to 2011, to create contemporaneous comparable cohorts. We compared open surgical revascularizations to endovascular interventions and subsequently compared nephrectomy to all other repairs. Outcomes included in-hospital deaths, post-operative complications and length of stay. Multivariable logistic regression was used to adjust for other potential predictors of mortality including age, coronary artery disease, prior myocardial infarction, chronic renal failure, peripheral vascular disease, heartfailure, and dysrhythmia. Nephrectomy patients were excluded from multivariable analysis. Statistical Analysis Statistical analysis was completed using SPSS statistical software (version 20; IBM Corp, Armonk, NY). Where appropriate, continuous variables were compared using two-tailed independent sample t-test, and Chi-square or Fischer’s exact test were used for categorical variables. The Cochran-Armitage test for trend was used to determine changes over time. Statistical significance was defined as P < 0.05.

Results Epidemiologic Trends We identified 6,234 patients undergoing renal artery aneurysm repairs between 1988 and 2011, encompassing both elective and urgent procedures. Both the overall total repair rate and the endovascular repair rate were noted to increase steadily over time (total repairs from 6.2 in 1988-1990 to 10.5 in 2009-2011 per 10M US population, P=0.002 and endovascular repairs from 0 in 1988-1990 to 4.8 in 2009-2011 per 10M US population, P<.0001) shown in Figure 1. After the introduction of endovascular repairs, the rate of open repair did not decrease significantly (4.0 in 1988-1990 to 5.6 in 2009-2011 per 10M US population, P =0.071), indicating that endovascular procedures are not replacing open procedures. 206

Isolated Renal Artery Aneurysms

Figure 1: Quantitative comparison of endovascular and open repairs (including both elective and urgent procedures)

Elective Endovascular vs. Open Repair 2000-2011 There were 2,709 elective procedures from 2000 to 2011; 1,627 open repairs and 1,082 endovascular repairs. Patients undergoing endovascular repair were more likely to have a history of coronary artery disease (18 vs. 11%, P<0.001), prior myocardial infarction (5.2 vs. 1.8%, P<0.001), chronic renal failure (7.7 vs. Table I. Pre-Operative characteristics of patients undergoing elective endovascular and open repair of renal artery aneurysm from 2000-2011 Open (N=1627)

EVAR (N=1082)

P-value

57

58

0.087

Female

57%

42%

<.001

Coronary Artery Disease

11%

18%

<.001

Hypertension

61%

58%

0.168

Dysrythmia

2.4%

4.0%

0.019

Prior myocardial infarction

1.8%

5.2%

<.001

Heartfailure

5.0%

4.9%

0.922

Chronic obstructive pulmonary disease

6.2%

7.7%

0.123

Atrial fibrillation

6.6%

6.4%

0.837

Chronic Renal Failure

3.3%

7.7%

<.001

Cardiovascular disease

0.9%

1.5%

0.133

Peripheral vascular disease

14%

9.2%

<.001

Pre-operative Demographics Age

11

207

Chapter 11 Table II. In hospital outcomes of elective endovascular and open renal artery aneurysm repairs from 2000-2011 Open (N=1627)

Endo (N=1082)

P-value

In-hospital Mortality

0.9%

1.8%

0.037

Cardiac complications

2.2%

0.6%

0.001

Respiratory complications

4.6%

4.3%

0.658

Peripheral vascular complications

0.6%

0.0%

0.014

Acute renal failure

10%

6.8%

0.001

Wound dehiscence

0.3%

0.0%

0.068

Bleeding complications

5.2%

5.0%

0.842

Infection

0.9%

0.8%

0.983

Any complication

12.4%

10.5%

0.134

6.0

4.6

<.001

Length of Stay (days)

3.3%, P<0.001) and dysrhythmia (4.0 vs. 2.4%, P=0.019) and less likely to have peripheral arterial disease (9.2% vs.14, P<0.01). Age and other pre-operative comorbidities, including hypertension, heart failure, chronic pulmonary disease, and atrial fibrillation, were comparable between the two groups, as demonstrated in Table I. In-hospital mortality was significantly higher for endovascular repairs compared to all open repairs (1.8 vs. 0.9%, P=0.037). (Table II) Overall complication rates were 12.4% for open repairs vs 10.5% for endovascular repairs (P=0.134). (Table II) This included more cardiac (2.2 vs. 0.6%, P=0.001) and peripheral vascular complications (0.6 vs. 0.0%, P=0.014) with open repair. The rates of respiratory complications, wound dehiscence, bleeding complications, and infection were not significantly different between groups, as shown in Table II. Open repair was Table III. In hospital outcomes after nephrectomy versus any revascularization for renal artery aneurysm from 2000-2011 Nephrectomy (N=459)

All Repairs (N=2709)

P-value

In-hospital Mortality

5.4%

1.3%

<.001

Cardiac complications

2.4%

1.5%

0.169

Respiratory complications

19.6%

4.5%

<.001

Peripheral vascular complications

0.0%

0.3%

0.217

Wound dehiscence

3,5%

0.2%

<.001

Bleeding complications

5.4%

5.0%

0.721

Post-operative Outcomes

Infection

2.2%

0.8%

0.009

Any complication

24.0%

11.6%

<.001

7.9

5.5

<.001

Length of Stay (days)

208

Isolated Renal Artery Aneurysms associated with a longer length of stay (6.0 vs. 4.6 days, P<0.001). Nephrectomy patients had a higher in hospital mortality than the overall repair group ( 5.4% vs. 1.3% p<0.001) and length of stay was longer (8 vs. 6 days, P<.001). (Table III) Infection, wound dehiscence, respiratory complications, and total complications were higher after nephrectomy. Multivariable predictors of mortality after a renal artery aneurysm were age (OR 1.05 per decade, 95% CI 1.0-1.1, P=0.001), heartfailure (OR 7.0 , 95% CI 3.1-16.0, P<.001) and dysrhythmia (OR 5.9, 95% CI 2.0-16.8, P=0.005). After adjustment, endovascular repair was neither protective nor predictive of mortality ((OR 1.6, 95% CI 0.8-3.2, P=0.145).).

Discussion

Our study shows that there has been an increase in total repairs since the introduction of endovascular techniques, without any decrease in open surgery. Patients undergoing endovascular repair had more pre-operative cardiac comorbidities and higher in-hospital mortality compared to open repair patients, although adjusted mortality was similar after accounting for differences in baseline characteristics. However, complication rates were higher for the open repair group with a longer length of stay. Henke et al reviewed 168 patients with 252 aneurysms, where 121 patients underwent surgery. Three patients presented with ruptured renal artery aneurysm. They found no perioperative deaths or late postoperative deaths due to renal artery aneurysm surgery.6 In a retrospective analysis of forty-four renal artery aneurysm repairs from 2000 to 2012, Tsilimparis et al. reported no mortality and equivalent perioperative complication rates of 15% and 17% for open repair and endovascular repair patients respectively. Additionally, they showed that endovascular repair was associated with a significantly shorter hospitalization of 2.3 days, in comparison to 6.3 days for open repair, which was comparable to our findings.22 Zhang et al. endorse the use of endovascular repair as a first-line technique for renal artery aneurysm treatment.21 In their study of 15 patients treated with endovascular techniques, they had no peri-procedural mortality or major complications, with a technical success rate of 100%. While our analysis found fewer complications and shorter length of stay with endovascular techniques, we found no reduction in operative mortality. The Low Frequency Disease Consortium reported an analysis in support of more conservative management and surveillance rather than aggressive surgical treatment, proposing that the current recommendations for renal artery aneurysm treatment at a diameter of 2 cm may be too aggressive.23 In their retrospective study of 40 patients, they found a rate of aneurysm rupture and death to be zero over a mean 36-month follow-up period, and a very low growth rate of 0.60 Âą 0.16 mm/year.23 An analysis by Hislop et al. using the New York State inpatient database noted a significant increase in both the total number of renal artery aneurysm repairs as well as an increase in the proportion of endovascular repairs performed from 2000 to 2006.9 Our study confirms this finding with a longer time period including 209

11

Chapter 11 the pre-endovascular era and is representative of the entire United States. In their retrospective review of 215 patients undergoing renal artery aneurysm repair, endovascular therapy was also associated with a lower incidence of complications, and a significantly lower length of stay.9 Additionally, the favorable post-operative outcomes reported for the endovascular repairs were comparable to those observed in our analysis and lend support to the notion that endovascular repairs may also be economically more favorable as indicated by the lower median cost and earlier discharge. The mortality rate was 1.1% in their endovascular cohort and 3.2% in the open repair group.9 This however, was not significantly different. In our analysis of 2709 procedures, endovascular repairs had significantly higher in-hospital mortality than open repairs. This was in part attributed to the comorbidities of the patients who received an endovascular treatment compared to those in the open repair cohort. Endovascular procedures may also be associated with technique-specific complications, including renal artery dissection, postembolization syndrome and coil migration.21 While we are able to study a large sample size, our study is limited in that it is not a randomized controlled trial. It is an observational study of administrative data and is subject to coding errors. Our data lack information on aneurysm diameter and the location of the renal artery aneurysm, whether it is found in the main renal artery or the hilum of the kidney. The dataset provides in-hospital data only, with no information available regarding follow-up after discharge. Nevertheless, our study represents one of the larger American series and may be broadly generalizable due to both academic and community hospitals that the NIS incorporates.

Conclusion

This retrospective review demonstrates that more renal artery aneurysms are being treated after the introduction of endovascular techniques. Although there is evidence supporting a significantly lower rate of post-operative complications and a shorter length of stay with endovascular repair, there has not been a reduction in operative mortality nor has there been a reduction in open surgical procedures. Further evaluation of the indications of repair of isolated renal artery aneurysms, by either endovascular or open techniques, is warranted.

210

Isolated Renal Artery Aneurysms

References 1. Hageman JH, Smith RF, Szilagyi E, Elliott JP. Aneurysms of the renal artery: problems of prognosis and surgical management. Surgery. 1978;84(4):56372. 2. Stanley JC, Rhodes EL, Gewertz BL, Chang CY, Walter JF, Fry WJ. Renal artery aneurysms. Significance of macroaneurysms exclusive of dissections and fibrodysplastic mural dilations. Arch Surg. 1975;110(11):1327-33. 3. Henriksson C, Bjorkerud S, Nilson AE, Pettersson S. Natural history of renal artery aneurysm elucidated by repeated angiography and pathoanatomical studies. European urology. 1985;11(4):244-8. 4. Tham G, Ekelund L, Herrlin K, Lindstedt EL, Olin T, Bergentz SE. Renal artery aneurysms. Natural history and prognosis. Annals of surgery. 1983;197(3):348-52. 5. English WP, Pearce JD, Craven TE, Wilson DB, Edwards MS, Ayerdi J, et al. Surgical management of renal artery aneurysms. J Vasc Surg. 2004;40(1):5360. 6. Henke PK, Cardneau JD, Welling TH, 3rd, Upchurch GR, Jr., Wakefield TW, Jacobs LA, et al. Renal artery aneurysms: a 35-year clinical experience with 252 aneurysms in 168 patients. Annals of surgery. 2001;234(4):454-62; discussion 62-3. 7. Gallagher KA, Phelan MW, Stern T, Bartlett ST. Repair of complex renal artery aneurysms by laparoscopic nephrectomy with ex vivo repair and autotransplantation. J Vasc Surg. 2008;48(6):1408-13. 8. Giulianotti PC, Bianco FM, Addeo P, Lombardi A, Coratti A, Sbrana F. Robot-assisted laparoscopic repair of renal artery aneurysms. J Vasc Surg. 2010;51(4):842-9. 9. Hislop SJ, Patel SA, Abt PL, Singh MJ, Illig KA. Therapy of renal artery aneurysms in New York State: outcomes of patients undergoing open and endovascular repair. Annals of vascular surgery. 2009;23(2):194-200. 10. Abath C, Andrade G, Cavalcanti D, Brito N, Marques R. Complex renal artery aneurysms: liquids or coils? Techniques in vascular and interventional radiology. 2007;10(4):299-307. 11. Cohen JR, Shamash FS. Ruptured renal artery aneurysms during pregnancy. J Vasc Surg. 1987;6(1):51-9. 12. Hidai H, Kinoshita Y, Murayama T, Miyai K, Matsumoto A, Ide K, et al. Rupture of renal artery aneurysm. European urology. 1985;11(4):249-53. 13. Hupp T, Allenberg JR, Post K, Roeren T, Meier M, Clorius JH. Renal artery aneurysm: surgical indications and results. European journal of vascular surgery. 1992;6(5):477-86. 14. Martin RS, 3rd, Meacham PW, Ditesheim JA, Mulherin JL, Jr., Edwards WH. Renal artery aneurysm: selective treatment for hypertension and prevention of rupture. J Vasc Surg. 1989;9(1):26-34. 15. Sorcini A, Libertino JA. Vascular reconstruction in urology. The Urologic clinics of North America. 1999;26(1):219-34, x-xi. 211

11

Chapter 11 16. Robinson WP, 3rd, Bafford R, Belkin M, Menard MT. Favorable outcomes with in situ techniques for surgical repair of complex renal artery aneurysms. J Vasc Surg. 2011;53(3):684-91. 17. Bui BT, Oliva VL, Leclerc G, Courteau M, Harel C, Plante R, et al. Renal artery aneurysm: treatment with percutaneous placement of a stent-graft. Radiology. 1995;195(1):181-2. 18. Dib M, Sedat J, Raffaelli C, Petit I, Robertson WG, Jaeger P. Endovascular treatment of a wide-neck renal artery bifurcation aneurysm. Journal of vascular and interventional radiology : JVIR. 2003;14(11):1461-4. 19. Tateno T, Kubota Y, Sasagawa I, Sawamura T, Nakada T. Successful embolization of a renal artery aneurysm with preservation of renal blood flow. International urology and nephrology. 1996;28(3):283-7. 20. Klein GE, Szolar DH, Breinl E, Raith J, Schreyer HH. Endovascular treatment of renal artery aneurysms with conventional non-detachable microcoils and Guglielmi detachable coils. British journal of urology. 1997;79(6):852-60. 21. Zhang Z, Yang M, Song L, Tong X, Zou Y. Endovascular treatment of renal artery aneurysms and renal arteriovenous fistulas. J Vasc Surg. 2013;57(3):765-70. 22. Tsilimparis N, Reeves JG, Dayama A, Perez SD, Debus ES, Ricotta JJ, 2nd. Endovascular vs open repair of renal artery aneurysms: outcomes of repair and long-term renal function. Journal of the American College of Surgeons. 2013;217(2):263-9. 23. Klausner JQ, Harlander-Locke MP, Plotnik AN, Lehrman E, DeRubertis BG, Lawrence PF. Current treatment of renal artery aneurysms may be too aggressive. J Vasc Surg. 2014;59(5):1356-61.

212

Isolated Renal Artery Aneurysms

11

213

CHAPTER 12

Online Patient Resources for Cardiovascular Surgery: An Analysis of Readability

Dominique B Buck, Pieter G Koolen, Jeremy Darling, Christina R Vargas, Rishi Mamtani, Duane S Pinto, Bernard T Lee, Marc L Schermerhorn. Submitted J Am Coll Cardiol.

Chapter 12

Abstract Background As patients increasingly obtain health information online, the American Medical Association (AMA) and National Institute of Health (NIH) recommend that medical information for patients be written at a 6th grade reading level for optimal comprehensibility. This study evaluates the reading grade levels of information found online for seven common cardiovascular diseases. Methods and Results Using www.google.com as the search engine for online patient information, we identified the top ten results and appropriate medical society websites [American Heart Association, Society for Vascular Surgery, Society of Interventional Radiology, Society of Thoracic Surgery] for the following cardiovascular disease processes: coronary artery disease, heart failure, atrial fibrillation, aortic aneurysm, carotid disease, peripheral arterial disease and varicose veins. A total of 746 links were identified and evaluated for readability using nine established tests assigning a grade level. The average website grade level for the studied websites for all the terms was 12.3 (range: 11.6-12.7). The grade levels of the average of the top search results and the average of society websites for each disease were: coronary artery disease (12.3, 13.9), heart failure (10.6, 11.9), atrial fibrillation (12.1, 11.0), aortic aneurysm (12.7, 13.3), carotid artery disease (12.6, 13.5), peripheral arterial disease (12.7, 14.0), and varicose veins (11.6, 12.2). Conclusions The readability of online patient information for cardiovascular diseases exceeds the recommended grade level. We suggest that medical societies revise their online patient information to a 6th grade reading level. Providing, appropriatelevel resources is essential for patient access to health information about these important topics.

216

Online Patient Resources in Cardiovascular Disease

Introduction

Utilizing the Internet as a source for health information in the United States continues to increase at an impressive rate. In 2014, the Pew Research Internet Project reported that 87% of American adults had used the internet (an increase of 24% from 2003)1 and that approximately 72% had searched for health information online.2 Sixty percent of those patients said the online information they found affected a decision about how they treated a condition; 38% said it helped them decide whether to see a doctor.3 The Health Information National Trends Survey (HINTS) reported that approximately 60% of patients use the Internet for health information, while only 21% consult their doctor.4 Several previous studies have suggested that increasing patient use of “supplemental” educational material positively contributes to more patient involvement, increased satisfaction, and improved health outcomes.5-10 Therefore, a great importance should be placed on ensuring that these auxiliary materials are advantageous and comprehensible. Reduced health literacy multiplies adverse outcomes among at risk populations, adversely increasing health disparities and mortality.11-15 Given these concerns, the American Medical Association (AMA) has encouraged physicians to increase awareness of patient literacy, and published their recommendation for patient information regarding health topics to be provided at or below the sixth grade level.5, 11 The National Institutes of Health (NIH) has also recommended that patient-centered health information be written at a sixth grade level as a means to enhance accessibility.16 Previous studies have shown that patients have difficulty comprehending information presented on both professional and public websites.17-22 Most of these studies, however, were conducted a number of years ago, did not generalize to real patient searches, and used only a few of the well-established readability analyses. Several tests are available for analyzing the readability of educational content, however no published data are available regarding the readability of currently available online patient resources for basic cardiovascular disorders.19, 23, 24

As the availability and use of online public health resources continue to grow, there is a crucial need to objectively evaluate the information patients are exposed to online. The objective of this study was to examine the reading level of online cardiovascular resources in a manner similar to a typical patient internet search. The aim of this study was to identify the most popular public websites as well as the professional medical society websites for cardiovascular disease and to assess their readability and compare to the average American’s reading ability.

Methods

Google searches were conducted for terms pertinent to cardiovascular disease, including “coronary artery disease,” “heart failure,” “atrial fibrillation,” “aortic aneurysm,” “carotid artery disease,” “peripheral arterial disease,” and “varicose veins”. Prior to searching, cookies, user account information, and user location 217

12

Chapter 12 were disabled to avoid result bias. All sites were accessed between April 1st 2014 and June 11th 2014. Sponsored search results were excluded. The patientcentered information presented from the first ten websites results was organized into articles by topic. Additionally, specified society websites [American Heart (Circulation Patient Page), Vascular Web, Society of Interventional Radiology, Society of Thoracic Surgery] outside of the top ten results were also included in this analysis. For analytic purposes, tables, figures, bibliographic references, videos, captions, disclaimers, acknowledgements, and embedded links were removed from the collected online information. The total number of articles obtained was 746. Readability tests were conducted to determine the grade level of each article. The analysis was performed with established testing software, Readability Table 1. Tests for Reading Grade Level Analysis Test

Qualities Assessed

Formula

Coleman-Liau

word length, sentence, length

G = ((5.88*C)/W) - ((29.5*S)/W) - 15.8

Flesch-Kincaid Grade

word, complexity, sentence, length

G = (11.8*(B/W)) + (0.39*(W/S)) -15.59

FORCAST

word, complexity, sentence, length

G = 20 - (M/10)

Fry Graph

word, complexity, sentence, length

Extract 100 word samples, Count number of sentences, Count number of syllables, Plot on the Fry graph, Fry score = average of samples

Gunning Fog

word, complexity, sentence, length

G = 0.4*(W/S+((X/W)*100))

New Dale-Chall

word, complexity, sentence, length

G = (0.0496*(W/S)) + (0.1579*(U/W)) +3.6365

New Fog Count

word, complexity, sentence, length

G = (((E + (3*X))/S) - 3)/2

Raygor Estimate

word length, sentence, length

Extract 100 word samples, Count number of sentences, Count number of words >6 letters, Plot on the Raygor graph, Raygor Estimate = average of samples

SMOG

word, complexity, sentence, length

G = 1.0430* √X + 3.1291

G=Grade level, I=Index, W=number of words, C=number of characters, S=number of sentences, B=number of syllables, M=number of monosyllabic words, X=number of complex words (>3 syllables), E=number of easy words, U=number of unfamiliar words (based on a list of 3000 common words known to average 4th grade students)

218

Online Patient Resources in Cardiovascular Disease Table 2. Average reading grade level per website, classified according to each disease. Coronary Carotid Peripheral Heart Atrial Aortic Varicose Artery Artery Arterial Mean Failure Fibrillation Aneurysm Veins Disease Disease Disease American Heart (Circulation Patient Page)

11.9

11.9

11

13.1

12.1

12.3

12.3

12.1

Society for Vascular Surgery (Vascular Web)

14

X

X

12.4

12.7

12.3

11.5

12.6

Society of Interventional Radiology

X

X

X

14.5

12.4

14.6

14

13.9

Society of Thoracic Surgery

13.9

X

X

13.2

16.8

16.8

10.9

14.3

WebMD

11.5

9.2

11.7

9.9

11.4

10.8

10.3

10.7 15.2

Wikipedia

15.4

15.5

15.1

15

15.6

15.2

14.6

NHLBI

9.6

8.3

9.7

11.4

10.5

10.3

8.7

9.8

15.5

12.2

13.2

Medicine Net

14.9

10.1

14.3

13

12.2

Medline Plus

11.1

8.3

10

9.6

10.5

9.4

8.7

9.7

CDC

9.8

X

X

13.2

11.5

12.4

12.5

11.9

Cleveland Clinic

12.1

10.1

13

14.2

13.1

10.2

12.6

12.2

Mayo Clinic

11.5

11

12.1

12.6

12

12.6

10.3

11.7

Mean Reading Grade Level

12.3

10.6

12.1

12.7

12.6

12.7

11.6

12.1

* bold numbers show averages within the top 10 search † non bold numbers show averages outside the top 10 search ‡ X = no data available

Studio Professional Edition v2012.1software (Oleander Software, Ltd, Vandalia, Ohio).24,25,29the National Institutes of Health (NIH All 746 articles were sorted, collectively analyzed, and compared to other articles within the same parent source. Ultimately, each source’s reading level was assessed using the following tests: Flesch Kincaid Grade Level, Gunning Fog Index, New Fog Count, Simple Measure of Gobbledygook (SMOG) Readability Formula, Coleman-Liau Index, FORCAST Formula, Fry Graph, Gunning Fog Index, New Dale-Chall, and Raygor Readability Estimate.25-29 The formulas with their assessed qualities are listed in Table 1. The grade level scale retrieved from the readability tests is directly correlated with the traditional grade parameters; for instance, a reading grade level of 9 describes a freshman high-school student reading level and grade 13 indicates a college freshman reading level.

219

12

Chapter 12 Figure 1 B

Coronary Artery Disease

Afib

Website

ik ip ed W ia eb M St D op Af ib N M H ed LB l I M ine ed Pl ic us H in ea e M rt N R ayo et hy th Cli n m Am Cle So ic v . H el ci ea and ety rt As Clin so ic ci at io n

W

W ik ip ed W ia eb M D N H L M ed BI M sca ed pe l M ine ed Pl ic us in M eN ay et o C lin Am Cle H ic . H vel FSo ea and M rt As Clin so ic ci at io n

Te xa s

Figure 1 C

Heart Failure

W ik ip ed i H We a ea bM rt In D st itu te Fa N m HL ily BI M Do ed ct ic or in M e. C ay net o le ve C la lin n Am d ic C .H lin ic ea rt As NL so M ci at io n

Mean Readability

Figure 1 A Doctorate Post-graduate sophomore Post-graduate freshman University senior University junior University sophomore University freshman 12th grade 11th grade 10th grade 9th grade 8th grade 7th grade 6th grade 5th grade 4th grade 3rd grade 2nd grade 1st grade Kindergarten

Website

Website

Figure 1: Boxplots demonstrating the distribution of reading grade levels of each website for “coronary artery disease”, “heart failure”, and “Atrial fibrillation ”. The dashed line demonstrates the AMA readability grade level guideline of grade 6.

Figure 2 A

Figure 2 B

Carotid Artery Disease

W ik ip ed ia Va We sc bM ul ar D W eb U C D av is M NI ed H M sca pe C ay o le ve C la lin nd ic C l C hi Circ inic ca u go lat M ion ed ic in e

Doctorate Post-graduate sophomore Post-graduate freshman University senior University junior University sophomore University freshman 12th grade 11th grade 10th grade 9th grade 8th grade 7th grade 6th grade 5th grade 4th grade 3rd grade 2nd grade 1st grade Kindergarten

W ik ip ed ia Va We sc bM ul ar D W eb M ed So lin IR e M ed Pl ic us in M eN et C ay o le ve C la lin nd ic C l C irc inic ul at io n C D C

Mean Readability

Aortic Aneurysm

Website

Website

Figure 2 C

Figure 2 D

Va M sc D ul W ar om W eb en 's H ea lth Ve ne M ed fit ic in e N M et ay o C M lin ed ic lin e Pl us

ik ip W

W eb

ed i

a

Mean Readability

Varicose Veins

U

M

ar

W ik ip ed ia Va We bM s yl an cul ar D d M W ed eb C M en ed te ic r al N ew SoI R s M Tod ay ay o C lin ic Am N M H .H e LB ea dic in I rt As e N so et ci at io n

Mean Readability

Peripheral Arterial Disease Doctorate Post-graduate sophomore Post-graduate freshman University senior University junior University sophomore University freshman 12th grade 11th grade 10th grade 9th grade 8th grade 7th grade 6th grade 5th grade 4th grade 3rd grade 2nd grade 1st grade Kindergarten

Website

Website

Figure 2: Boxplots illustrating the range of reading grade levels obtained for each website providing information pertinent to “aortic aneurysm”, “carotid artery disease”, “peripheral arterial disease”, and “varicose veins”. The dashed line demonstrates the AMA readability grade level guideline of grade 6.

220

Online Patient Resources in Cardiovascular Disease

Results

Mean reading grade level for all tests by topic were: 12.3 (range 9.6-15.4) for “coronary artery disease”, 10.6 (8.3-15.5) for “heart failure”, 12.1 (9.7-15.1) for “atrial fibrillation ”, 12.7 (9.6-15.0) for “aortic aneurysm”, 12.6 (10.5-16.8) for “carotid artery disease”, 12.7 (9.4-16.8) for “peripheral artery disease”, and 11.6 (8.7-14.6) for “varicose veins”(Table 2). For all statistical tests and cardiovascular diseases combined, the mean reading grade level is 12.1. Mean reading grade levels for the society websites, including American Heart, Society for Vascular Surgery, Society of Interventional Radiology and Society of Thoracic Surgery were 13.3, 11.9, 11.0 13.3, 13.5, 14.0, and 12.2, for ‘coronary artery disease,’ ‘heart failure,’ ‘atrial fibrillation,’ ‘aortic aneurysm,’ ‘carotid artery disease,’ ‘peripheral arterial disease,’ and ‘varicose veins’ respectively. Subsequently, for the individual society websites, the mean reading grade levels were 12.1 (11.0-13.1) for American Heart, 12.6 (11.5-14) for Society for Vascular Surgery (Vascular Web), 13.9 (12.4-14.6) for Society of Interventional Radiology, and 14.3 (10.9-16.8) for Society of Thoracic Surgery. For the individual top ten websites mean grade levels were 10.7 (9.211.7) for WebMD, 15.2 (14.6-15.6) for Wikipedia, 9.8 (8.3-11.4) for NHLBI, 13.2 (10.1-15.5) for Medicine Net, 9.7 (8.3-11.1) for Medline Plus, 11.9 (9.8-13.2) for CDC, 12.2 (10.1-14.2) for the Cleveland Clinic, and 11.7 (10.3-12.6) for the Mayo Clinic. Figures 1 and 2, which illustrate the range of reading level of each website and cardiovascular disease, display the disparities between tests for specified terms in order to better guide patients and physicians to specific websites which best approximate the recommended reading level. All results, however, exceed the guidelines for the recommended 6th grade reading level.

Discussion

Our findings demonstrate that the reading level of major cardiovascular health information websites consistently exceeds the recommended 6th grade literacy level. The finding was true for medical society patient information websites as well as other websites. Despite a broad range, these were uniformly above the AMA and NIH recommended 6th grade reading level proposed for optimal patient comprehension. This finding may have negative consequences on the delivery and dissemination of healthcare information. As the internet becomes an increasingly influential source of health information,2 it is crucial that patients have access to relevant, beneficial, and understandable information when making essential decisions regarding their medical care. These data strongly suggest that a change is needed in order to comply with the AMA and NIH reading-level recommendations and as a means to optimize patient information comprehension of publicly accessible online health information. Other studies have shown that accessing health information on the Internet is becoming a more commonly utilized source compared with actual healthcare providers themselves.30 Inadequate health literacy is being recognized as an increasing 221

12

Chapter 12 problem in American society. In fact, according to the Council of Scientific Affairs, patients with the most severe health care needs have the lowest ability to comprehend health care information pertinent to their conditions.5 Furthermore, the extent of this concern is quantified in the 1992 National Adult Literacy Survey (NALS), which found that approximately 44 million Americans are functionally illiterate, while another 50 million have only marginal literacy skills.5 This limited level of both functional literacy as well as health literacy, coupled with the high reading levels of credible online health information sources, are impediments for physicians attempting to optimize health care management for patients. Whether complementing or hindering a physician’s proposed management plan, online health information can significantly influence a patient’s medical outcome.30 The tests utilized in our study have served as the standard for measuring the reading level of different websites, incorporating an objective measure of vocabulary analysis, level of comprehension (i.e. 100% or 50-75% comprehension), character count, and indices for ease of reading an article.31 Our finding that reading levels are above average comprehension levels are supported by previous studies in a variety of medical specialties. Previous research has illustrated high reading levels of medical and surgical society websites, both in professional and public domains.32-34 Possible ways to improve this patient comprehension dilemma may involve increasing physician awareness of currently available resources and their readability in hopes that they may edit the content, guiding patients to more effective online resources, creating new websites with additional multimedia features that may aid a patient to better comprehend the presented information.32-34 In direct patient-physician encounters, patients often fail to understand the information provided,35-38 which emphasizes the importance of comprehensible online health information as confirmed by our study. In future research, we would recommend examination of online health information and the influence it has on patient health. An objective examination of the interactions between patients and physicians may be helpful to analyze the priority that patients place on different sources of information and their perceived credibility. Further research could interrogate online browsing patterns for health information and site traffic. A potential limitation of this analysis may involve the methodology necessitating organization of website information into articles. Each article was created without any embedded links, graphics, videos and disclaimers, in order to obtain a ‘representative’ reading level of the text. Removing such aspects of a website does not account for the potential impact of those features on comprehension. The number of articles in each source group also varied, which may be another confounding factor. However, all information assessed was suggestive of a substantially higher reading grade level of online sources than the recommended sixth grade level.

Conclusion

The results of this study suggest that the reading level of commonly used cardiovascular health information websites is higher than the recommended 222

Online Patient Resources in Cardiovascular Disease national average. This disparity may have negative implications on the delivery and dissemination of Healthcare information, and will be magnified as patient use of internet resources continues to increase. Further research should be directed at more quantitative examinations of the relationships between internetbased health literacy and patient decision-making on healthcare disparities and outcomes.

12 223

Chapter 12

References 1. Hesse BW, Nelson DE, Kreps GL, Croyle RT, Arora NK, Rimer BK, Viswanath K. Trust and sources of health information: The impact of the internet and its implications for health care providers: Findings from the first health information national trends survey. Archives of internal medicine. 2005;165:2618-2624 2. Health fact sheet, pew research internet project. January 2014 3. Sussannah Fox SJ. The social life of health information. June 2009 4. Kim K, Kwon N. Profile of e-patients: Analysis of their cancer informationseeking from a national survey. Journal of health communication. 2010;15:712-733 5. Health literacy: Report of the council on scientific affairs. Ad hoc committee on health literacy for the council on scientific affairs, american medical association. JAMA : the journal of the American Medical Association. 1999;281:552-557 6. Ashraf AA, Colakoglu S, Nguyen JT, Anastasopulos AJ, Ibrahim AM, Yueh JH, Lin SJ, Tobias AM, Lee BT. Patient involvement in the decision-making process improves satisfaction and quality of life in postmastectomy breast reconstruction. The Journal of surgical research. 2013;184:665-670 7. Lantz PM, Janz NK, Fagerlin A, Schwartz K, Liu L, Lakhani I, Salem B, Katz SJ. Satisfaction with surgery outcomes and the decision process in a population-based sample of women with breast cancer. Health services research. 2005;40:745-767 8. Lee BT, Chen C, Yueh JH, Nguyen MD, Lin SJ, Tobias AM. Computer-based learning module increases shared decision making in breast reconstruction. Annals of surgical oncology. 2010;17:738-743 9. Rutten LJ, Squiers L, Hesse B. Cancer-related information seeking: Hints from the 2003 health information national trends survey (hints). Journal of health communication. 2006;11 Suppl 1:147-156 10. Yueh JH, Slavin SA, Adesiyun T, Nyame TT, Gautam S, Morris DJ, Tobias AM, Lee BT. Patient satisfaction in postmastectomy breast reconstruction: A comparative evaluation of diep, tram, latissimus flap, and implant techniques. Plastic and reconstructive surgery. 2010;125:1585-1595 11. B.D. W. Health literacy: A manual for clinicians. AMA Foundation. 2003:1-49 12. Bostock S. SA. Association between low functional health literacy and mortality in older adults: Longitudinal cohort study. The BMJ. 2012;344:e1602 13. Powers BJ, Trinh JV, Bosworth HB. Can this patient read and understand written health information? JAMA : the journal of the American Medical Association. 2010;304:76-84 14. Rodriguez V, Andrade AD, Garcia-Retamero R, Anam R, Rodriguez R, Lisigurski M, Sharit J, Ruiz JG. Health literacy, numeracy, and graphical literacy among veterans in primary care and their effect on shared decision making and trust in physicians. Journal of health communication. 2013;18 Suppl 1:273-289 224

Online Patient Resources in Cardiovascular Disease 15. Sentell TL, Halpin HA. Importance of adult literacy in understanding health disparities. Journal of general internal medicine. 2006;21:862-866 16. Health NIo. How to write easy to read health materials. July 15, 2014 17. Aliu O, Chung KC. Readability of asps and asaps educational web sites: An analysis of consumer impact. Plastic and reconstructive surgery. 2010;125:1271-1278 18. Cox N, Bowmer C, Ring A. Health literacy and the provision of information to women with breast cancer. Clinical oncology. 2011;23:223-227 19. Friedman DB, Hoffman-Goetz L, Arocha JF. Health literacy and the world wide web: Comparing the readability of leading incident cancers on the internet. Medical informatics and the Internet in medicine. 2006;31:67-87 20. Hoppe IC, Ahuja NK, Ingargiola MJ, Granick MS. A survey of patient comprehension of readily accessible online educational material regarding plastic surgery procedures. Aesthetic surgery journal / the American Society for Aesthetic Plastic surgery. 2013;33:436-442 21. Misra P, Agarwal N, Kasabwala K, Hansberry DR, Setzen M, Eloy JA. Readability analysis of healthcare-oriented education resources from the american academy of facial plastic and reconstructive surgery. The Laryngoscope. 2013;123:90-96 22. Wang SW, Capo JT, Orillaza N. Readability and comprehensibility of patient education material in hand-related web sites. The Journal of hand surgery. 2009;34:1308-1315 23. Sand-Jecklin K. The impact of medical terminology on readability of patient education materials. Journal of community health nursing. 2007;24:119-129 24. Shedlosky-Shoemaker R, Sturm AC, Saleem M, Kelly KM. Tools for assessing readability and quality of health-related web sites. Journal of genetic counseling. 2009;18:49-59 25. Coleman M LT. A computer readability formula designed for machine scoring. J . Appl. Psychol. 1975;60:283-284 26. R G. The fog index after twenty years. J. Bus. Commun. 1969;6:3-13 27. Burke V GD. Determining readability : How to select and apply easy-to-use readability formulas to assess the difficulty of adult literacy materials. Adult Basic Educ. Lit. J. 2010;4:34-42 28. R F. A new readability yardstick. J. Appl. Psychol. 1948;32:221-233 29. W.H. D. Smart language: Readers, readability, and the grading of text. 2007;Online Submission 30. Hou J, Shim M. The role of provider-patient communication and trust in online sources in internet use for health-related activities. Journal of health communication. 2010;15 Suppl 3:186-199 31. Pinero-Lopez MA, Modamio P, Lastra CF, Marino EL. Readability assessment of package inserts of biological medicinal products from the european medicines agency website. Drug safety : an international journal of medical toxicology and drug experience. 2014;37:543-554 32. Bailey MA, Coughlin PA, Sohrabi S, Griffin KJ, Rashid ST, Troxler MA, Scott DJ. Quality and readability of online patient information for abdominal aortic aneurysms. J Vasc Surg. 2012;56:21-26 225

12

Chapter 12 33. Hansberry DR, Kraus C, Agarwal N, Baker SR, Gonzales SF. Health literacy in vascular and interventional radiology: A comparative analysis of online patient education resources. Cardiovascular and interventional radiology. 2014;37:1034-1040 34. San Norberto EM, Taylor J, Salvador R, Revilla A, Merino B, Vaquero C. The quality of information available on the internet about aortic aneurysm and its endovascular treatment. Revista espanola de cardiologia. 2011;64:869-875 35. Engel KG, Heisler M, Smith DM, Robinson CH, Forman JH, Ubel PA. Patient comprehension of emergency department care and instructions: Are patients aware of when they do not understand? Annals of emergency medicine. 2009;53:454-461 e415 36. Safeer RS, Keenan J. Health literacy: The gap between physicians and patients. American family physician. 2005;72:463-468 37. Smith SG, von Wagner C, McGregor LM, Curtis LM, Wilson EA, Serper M, Wolf MS. The influence of health literacy on comprehension of a colonoscopy preparation information leaflet. Diseases of the colon and rectum. 2012;55:1074-1080 38. Musso MW, Perret JN, Sanders T, Daray R, Anderson K, Lancaster M, Lim D, Jones GN. Patients’ comprehension of their emergency department encounter: A pilot study using physician observers. Annals of emergency medicine. 2014:15

226

Online Patient Resources in Cardiovascular Disease

12 227

CHAPTER 13

Summary, general discussion and future perspectives

Chapter 13

Summary, general discussion and future perspectives

Abdominal aortic aneurysm (AAA) disease is a growing healthcare burden. Consequently, given the growing population, the increasing elderly population, and the steadily rising life expectancy rates, the prevalence of AAA disease will continue to increase. Although the etiology of AAA disease is still not known, it is believed to be multifactorial and mainly degenerative. Risk factors for AAA development include hypertension, atherosclerosis, male gender, obesity, smoking, and a positive family history.1, 2 Elective repair of AAA is undertaken in order to prevent future rupture with its accompanying high mortality rate.3, 4 Initially, open surgical repair was the traditional treatment for AAA repair; however, the minimally invasive endovascular aneurysm repair (EVAR) has now become a safe and widely accepted surgical technique. The introduction of EVAR has changed the risks and benefits of elective aneurysm repair as it may be offered to patients considered unfit for open surgical repair. Surgical techniques and insights in AAA management The utilization of EVAR has broadened dramatically and has, over the years, slowly become the standard of care in an increasing number of AAA indications.5 In Chapter 2, previous literature was presented on the indications for EVAR, patient selection, the development of stent grafts, and clinical evidence for benefits of EVAR. As compared to open repair, randomized controlled trials and observational studies on endovascular repair have demonstrated lower perioperative morbidity and mortality,5-8 with the two approaches providing similar rates of long-term survival.9-12 In Chapter 3, we demonstrated that EVAR has a substantial survival benefit for approximately 3 years, after which the survival benefit gradually eroded over time. Our study further showed that outcomes after EVAR are improving over time and there was a decline in perioperative mortality. This is most likely due to increased familiarity with the procedure and improvements in stent grafts over time. The resulting decrease in reinterventions seemed to be driven by primarily coil embolizations, which may suggest a more conservative approach towards type 2 endoleak management. It is unlikely, however, that these improvements are driven by improved patient selection, as the vast majority of patients are now being treated using EVAR and mortality rates for open repair improved over this time period as well. Late rupture after endovascular repair remains alarming and merits further follow-up. Although our study in Chapter 3 includes an enormous cohort, determination of coexisting conditions from prior encounters, and the use of propensity-score matching, it lacked anatomical detail. In Chapter 4, access to intra-operative and clinical detail allowed us to look at the impact of type 2 endoleaks. We determined risk factors for and associations with type 2 endoleaks, a common complication following EVAR, where criteria for reinterventions remain undefined. Prior evidence has shown that persistent type 2 endoleaks are associated with an increased risk of adverse outcomes.13, 14 Our study confirms this finding, linking type 2 endoleaks to higher rates of sac expansion, reinterventions, and 230

Summary , general discussion and future perspectives conversion to open repair. Furthermore, we found that hypogastric artery coiling correlates with persistent type 2 endoleaks -- an outcome that suggests that these patients may have aneurysms that extend into the common iliac arteries. Despite no correlation with the presence of iliac aneurysms, these aneurysms might have a sufficient distal neck that does not require extension into external iliac arteries. Our study suggests that continued surveillance of patients with persistent type 2 endoleaks is needed, especially given the rising in popularity of EVAR. The previous Chapters have demonstrated an increase of EVAR, a lack of sustained survival benefit, and the need for postoperative imaging surveillance. Along with the steady increase of health care costs over time,15 EVAR practice sustainability has become a growing concern. Primarily due to device costs, surveillance imaging and reinterventions, the overall cost of EVAR has been reported to be significantly higher than open repair. 16-19 Based on preliminary reports, however, the feasibility of same day discharge after EVAR may offer improved cost effectiveness without loss of patient safety.20, 21 Chapter 5 identified ambulatory EVAR to be clinically safe and feasible in selected patients, though only representing a fraction of overall EVAR being performed. Chapter 5 suggests that appropriate patient candidacy for ambulatory EVAR, relative to inpatient EVAR, may be impressively cost-effective: In our study, charges for ambulatory EVAR were approximately one-third less than those of inpatient EVAR. It is important to note that, although performance of EVAR in the ambulatory setting may provide cost savings for operating room time, institutional overhead, and inpatient bed fees, these incremental improvements should not undermine the need for more competitive device costs. Not only seen in costs, competitiveness has also led to a wide variety of available stent grafts within this expanding market. Previous literature has demonstrated the safety of the use of commercially available stent grafts.5, 9, 12, 22, 23 Recently, however, issues have been raised about the performance of stent grafts resulting in changes in their design. Due to the limited research regarding comparing between stent graft types, we evaluated outcomes of the single docking limb, double docking limb, and unibody bifurcated stent graft. Chapter 6 established these stent graft comparisons, demonstrating fewer reinterventions with the Zenith stent graft, despite our inability to distinguish the AneuRx from the Excluder. The Zenith stent graft was the first graft to incorporate suprarenal stenting to prevent migration, an attribute that may be an explanation for its ability to decrease postoperative reinterventions. We further demonstrated that, as compared to the AneuRx/Excluder, the Powerlink graft has more perioperative complications, yet fewer conversions to open repair. An added potential benefit for the Powerlink graft is that its graft bifurcation sits on the aortic bifurcation and may assist in minimizing migration. Our study represents 2nd and 3rd generation stent grafts, but was limited by the impossibility to distinguish the AneuRx from the Excluder. Further comparative analyses of different stent grafts are necessary, as our study suggests impactful differences in performance. The performance and execution of the surgical procedure itself may also have an affect on the outcomes of endovascular and open surgical repair. For EVAR, 231

13

Chapter 13 percutaneous access further minimizes invasiveness, compared to the femoral cutdown access. In Chapter 7, we demonstrated a technical success rate of 96% with the use of percutaneous access, as well as several postoperative benefits, including shorter operative time, shorter length of stay, and fewer wound complications. We showed that the previously published results are generalizable to a broader population of patients, with our study including 83 centers in the United States. Despite the development of endovascular techniques for abdominal aortic aneurysms in recent years, open surgical repair remains necessary for centers that may lack the necessary resources or the appropriate skills to adopt endovascular AAA repair, as well as to treat anatomically complex aneurysms..24 Additionally, open repair may be practical for younger patients requiring longterm follow-up surveillance.25 Over the years, there has been an extensive debate about whether open surgical AAA repair should be completed by transperitoneal of retroperitoneal approach. In Chapter 8, we demonstrated that the retroperitoneal approach -- more commonly used for more proximal aneurysms -- is associated with higher rates of pneumonia, reintubations, and blood transfusions. Although after multivariable adjustment this was simply related to concomitant procedures. The retroperitoneal approach is more commonly used for more proximal aneurysms, which is likely attributed to better exposure of the suprarenal abdominal aorta, allowing a more secure proximal anastomosis. The transperitoneal approach, on the other hand showed more wound dehiscence, indicating that neither open surgical approach offers more benefits than the other. Surgeon specialty was reported to influence outcomes after open surgical repair.26, 27 The effect of the changing technology in this innovative driven profession has been seen in the varying types of physicians performing EVAR. In Chapter 9, we showed that vascular surgeons perform an increasing majority of AAA repairs in the U.S., while general surgeons and cardiac surgeons perform progressively fewer. EVAR became more widely used in the emergency setting, contributing to the growing proportion of emergency repairs being performed by vascular surgeons. Additionally, regionalization of open AAA repairs to highvolume centers over time has likely contributed to vascular surgeons becoming increasingly responsible for AAA surgery.28 A persistent presence of general surgeons treating open ruptured AAA repair remained, however, which could be due to geographic locations where the presence of a vascular surgeon may be lacking.29 Overall, AAA treatment by general- and cardiac surgeons is associated with a decreased likelihood of receiving EVAR, most likely due to the lack of resources or the appropriate skills to adopt endovascular AAA repair. Evolution of abdominal aneurysm repair Mainly due to the comparative infrequency of occurrences, the management of other isolated abdominal artery aneurysms is much less studied than that of abdominal aortic aneurysms. In Chapter 10, we studied epidemiologic trends in the management and mortality of isolated iliac artery aneurysms in the U.S., 232

Summary , general discussion and future perspectives showing that the treatment of these aneurysms have increased since the introduction of endovascular repair. Similar to the trend seen in AAA repair,30, 31 the treatment of isolated iliac artery aneurysms has shifted away from open surgical repair towards endovascular repair. However, despite the increase of treatment, the mortality has decreased over time. The increase in elective procedures did not lead to a decrease in urgent procedures, suggesting that surgeons have a lower threshold for intervention with availability of a less invasive treatment option. Similar to iliac artery aneurysms, information regarding the impact of endovascular therapy on renal artery aneurysms remains elusive. Previous studies have suggested that indications for surgical intervention for isolated renal artery aneurysms remain controversial.32, 33 Mirroring our iliac aneurysm study, Chapter 11 showed that there has been an increase in total renal artery aneurysm repairs since the introduction of endovascular techniques. There was, however, no decrease in open surgical repair. Additionally, when comparing outcomes between open and endovascular surgical repair, we found higher mortality rates amongst endovascular repair patients. After multivariable adjustment, mortality rates were equivalent in both cohorts. Overall, this could suggest that the current treatment of isolated renal artery aneurysms may be too aggressive. Optimal health care in vascular surgery Over the years, as patients have become more informed healthcare consumers and shared decision-makers, the traditional role of the physician as the sole provider of health information has progressively transformed. With the increased use of the internet,34 online health information has become an important contributor to patient involvement. In Chapter 12, the reading grade levels of internet-gathered information on cardiovascular disease were evaluated. We demonstrated that top websites on cardiovascular disease had a higher reading level than the advised 6th grade reading level.35 Surgeons are in a position to proactively assess both the health literacy and supplemental needs of their patients, and should consider providing or directing patients to appropriate level information. This can, subsequently, encourage patient participation and may positively contribute to both patient satisfaction and overall outcomes.

233

13

Chapter 13

References 1. Forsdahl SH, Singh K, Solberg S, Jacobsen BK. Risk factors for abdominal aortic aneurysms: a 7-year prospective study: the Tromso Study, 1994-2001. Circulation. 2009;119(16):2202-8. 2. Kent KC, Zwolak RM, Egorova NN, Riles TS, Manganaro A, Moskowitz AJ, et al. Analysis of risk factors for abdominal aortic aneurysm in a cohort of more than 3 million individuals. J Vasc Surg. 2010;52(3):539-48. 3. Bengtsson H, Bergqvist D. Ruptured abdominal aortic aneurysm: a population-based study. J Vasc Surg. 1993;18(1):74-80. 4. Bengtsson H, Bergqvist D, Sternby NH. Increasing prevalence of abdominal aortic aneurysms. A necropsy study. The European journal of surgery = Acta chirurgica. 1992;158(1):19-23. 5. Schermerhorn ML, O’Malley AJ, Jhaveri A, Cotterill P, Pomposelli F, Landon BE. Endovascular vs. open repair of abdominal aortic aneurysms in the Medicare population. N Engl J Med. 2008;358(5):464-74. 6. participants E. Endovascular aneurysm repair versus open repair in patients with abdominal aortic aneurysm (EVAR trial 1): randomised controlled trial. Lancet. 2005;365(9478):2179-86. 7. Prinssen M, Verhoeven EL, Buth J, Cuypers PW, van Sambeek MR, Balm R, et al. A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med. 2004;351(16):1607-18. 8. Lederle FA, Freischlag JA, Kyriakides TC, Padberg FT, Jr., Matsumura JS, Kohler TR, et al. Outcomes following endovascular vs open repair of abdominal aortic aneurysm: a randomized trial. JAMA : the journal of the American Medical Association. 2009;302(14):1535-42. 9. De Bruin JL, Baas AF, Buth J, Prinssen M, Verhoeven EL, Cuypers PW, et al. Long-term outcome of open or endovascular repair of abdominal aortic aneurysm. N Engl J Med. 2010;362(20):1881-9. 10. Greenhalgh RM, Brown LC, Powell JT, Thompson SG, Epstein D, Sculpher MJ. Endovascular versus open repair of abdominal aortic aneurysm. N Engl J Med. 2010;362(20):1863-71. 11. Jackson RS, Chang DC, Freischlag JA. Comparison of long-term survival after open vs endovascular repair of intact abdominal aortic aneurysm among Medicare beneficiaries. JAMA : the journal of the American Medical Association. 2012;307(15):1621-8. 12. Lederle FA, Freischlag JA, Kyriakides TC, Matsumura JS, Padberg FT, Jr., Kohler TR, et al. Long-term comparison of endovascular and open repair of abdominal aortic aneurysm. N Engl J Med. 2012;367(21):1988-97. 13. van Marrewijk CJ, Fransen G, Laheij RJ, Harris PL, Buth J, Collaborators E. Is a type II endoleak after EVAR a harbinger of risk? Causes and outcome of open conversion and aneurysm rupture during follow-up. European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery. 2004;27(2):128-37. 14. Jones JE, Atkins MD, Brewster DC, Chung TK, Kwolek CJ, LaMuraglia GM, 234

Summary , general discussion and future perspectives et al. Persistent type 2 endoleak after endovascular repair of abdominal aortic aneurysm is associated with adverse late outcomes. J Vasc Surg. 2007;46(1):1-8. 15. Kaplan RS, Porter ME. How to solve the cost crisis in health care. Harvard Bus Rev. 2011;89:46-61. 16. Angle N, Dorafshar AH, Moore WS, Quinones-Baldrich WJ, Gelabert HA, Ahn SS, et al. Open versus endovascular repair of abdominal aortic aneurysms: what does each really cost? Annals of vascular surgery. 2004;18(5):612-8. 17. Dryjski M, O’Brien-Irr MS, Hassett J. Hospital costs for endovascular and open repair of abdominal aortic aneurysm. Journal of the American College of Surgeons. 2003;197(1):64-70. 18. Garcia-Madrid C, Josa M, Riambau V, Mestres CA, Muntana J, Mulet J. Endovascular versus open surgical repair of abdominal aortic aneurysm: a comparison of early and intermediate results in patients suitable for both techniques. European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery. 2004;28(4):36572. 19. Prinssen M, Wixon CL, Buskens E, Blankensteijn JD. Surveillance after endovascular aneurysm repair: diagnostics, complications, and associated costs. Annals of vascular surgery. 2004;18(4):421-7. 20. Dosluoglu HH, Lall P, Blochle R, Harris LM, Dryjski ML. Ambulatory percutaneous endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2014;59(1):58-64. 21. Lachat ML, Pecoraro F, Mayer D, Guillet C, Glenck M, Rancic Z, et al. Outpatient endovascular aortic aneurysm repair: experience in 100 consecutive patients. Annals of surgery. 2013;258(5):754-8; discussion 8-9. 22. White GH, May J, McGahan T, Yu W, Waugh RC, Stephen MS, et al. Historic control comparison of outcome for matched groups of patients undergoing endoluminal versus open repair of abdominal aortic aneurysms. J Vasc Surg. 1996;23(2):201-11; discussion 11-2. 23. Zarins CK, White RA, Schwarten D, Kinney E, Diethrich EB, Hodgson KJ, et al. AneuRx stent graft versus open surgical repair of abdominal aortic aneurysms: multicenter prospective clinical trial. J Vasc Surg. 1999;29(2):292305; discussion 6-8. 24. Danjoux NM, Martin DK, Lehoux PN, Harnish JL, Shaul RZ, Bernstein M, et al. Adoption of an innovation to repair aortic aneurysms at a Canadian hospital: a qualitative case study and evaluation. BMC health services research. 2007;7:182. 25. Shaw PM, Veith FJ, Lipsitz EC, Ohki T, Suggs WD, Mehta M, et al. Open aneurysm repair at an endovascular center: value of a modified retroperitoneal approach in patients at high risk with difficult aneurysms. J Vasc Surg. 2003;38(3):504-10. 26. Cronenwett JL, Birkmeyer JD. The Dartmouth Atlas of Vascular Health Care. Cardiovascular surgery. 2000;8(6):409-10. 27. Tu JV, Austin PC, Johnston KW. The influence of surgical specialty training on the outcomes of elective abdominal aortic aneurysm surgery. J Vasc 235

13

Chapter 13 Surg. 2001;33(3):447-52. 28. Hill JS, McPhee JT, Messina LM, Ciocca RG, Eslami MH. Regionalization of abdominal aortic aneurysm repair: evidence of a shift to high-volume centers in the endovascular era. Journal of vascular surgery. 2008;48(1):29-36. 29. Maybury RS, Chang DC, Freischlag JA. Rural hospitals face a higher burden of ruptured abdominal aortic aneurysm and are more likely to transfer patients for emergent repair. Journal of the American College of Surgeons. 2011;212(6):1061-7. 30. Giles KA, Pomposelli F, Hamdan A, Wyers M, Jhaveri A, Schermerhorn ML. Decrease in total aneurysm-related deaths in the era of endovascular aneurysm repair. J Vasc Surg. 2009;49(3):543-50; discussion 50-1. 31. Schermerhorn ML, Bensley RP, Giles KA, Hurks R, O’Malley A J, Cotterill P, et al. Changes in abdominal aortic aneurysm rupture and short-term mortality, 1995-2008: a retrospective observational study. Annals of surgery. 2012;256(4):651-8. 32. English WP, Pearce JD, Craven TE, Wilson DB, Edwards MS, Ayerdi J, et al. Surgical management of renal artery aneurysms. J Vasc Surg. 2004;40(1):5360. 33. Henke PK, Cardneau JD, Welling TH, 3rd, Upchurch GR, Jr., Wakefield TW, Jacobs LA, et al. Renal artery aneurysms: a 35-year clinical experience with 252 aneurysms in 168 patients. Annals of surgery. 2001;234(4):454-62; discussion 62-3. 34. Fox S JS. The Social Life of Health Information. pewinternet.org: California HealthCare Foundation, May 2011. 35. Health NIo. How to Write Easy to Read Health Materials. http://www.nlm.nih. gov/medlineplus/etr.htmlJuly 15, 2014.

236

Review Committee Prof. dr. J.P.J. van Schaik Radiology University Medical Center Utrecht Utrecht, The Netherlands Prof. dr. W.A. van Klei Anesthesiology University Medical Center Utrecht Utrecht, The Netherlands Prof. dr. I.H.M. Borel Rinkes Surgery University Medical Center Utrecht Utrecht, The Netherlands Prof. dr. H.J.M. Verhagen Vascular Surgery Erasmus Medical Center Rotterdam, The Netherlands Prof. dr. W. Wisselink Vascular Surgery VU Medical Center Amsterdam, The Netherlands Prof.dr. P.J. van Diest Pathology University Medical Center Utrecht Utrecht, The Netherlands

Word of Thanks The time has come, this thesis is done. I would like to thank everyone who has contributed to the realization of this thesis and shared this absolutely wonderful time with me. I’d like to thank a number of people in particular. Dear Prof.dr. Moll, thank you for taking me under your wing and giving me the incredible opportunity to move to Boston. Your supervision and guidance is one of a kind, and I admire your personal involvement. You always said that working on a PhD is much more than just doing research; it is a combination of being able to work with yourself, manage your time and collaborate with others. Thank you for being my mentor during all these steps. Dear Dr. Schermerhorn, it has been an amazing time and personal enrichment to work in your lab. I look up to how you manage to be a magnificent surgeon and exceptional in scientific research at the same time. I admire how you always make time, how you prioritize, how you never give up, how you share your ideas with others and how you collaborate with many groups. You are an inspiration. Thank you for showing me the right directions, thank you for putting up with the silly Dutch girl and thank you for teaching me that ‘data are’. Dear Dr. van Herwaarden, I really enjoyed working with you over the past years, and in particular our teamwork when we wrote the Nature review. I truly appreciate your determination, your always honest feedback and your love for vascular surgery. Thank you for your mentorship and guidance during clinical work and my research years. Although I’m definitely not your first PhD-candidate, I am your first female one and hope we can continue writing articles together. Dear Professors of the Review Committee, thank you very much for the critical evaluation of this thesis. Dear Prof. van Diest, I feel extremely lucky to have met you a couple of years ago and to have you as a tutor. Although we haven’t worked with each other over the last two years, you have always been there for me and given me advice. That is priceless to me. I hope we do get to work with each other again in the future, but for now… I can’t wait to make music together again. My Colleagues in Boston; Ruby and Tommy, thanks for showing me the ropes and for all the fun times, it was short but very sweet. Jdogzzzzz, I can’t believe we’re not together in the office anymore and that I’ll have to celebrate Thanksgivings without you from now on. Thank you for being my crazy and true friend, and don’t forget that another Tough Mudder still stands. Jack, it has been a pleasure working with you and thank you for showing me a new world of wikipedia writers. Pete, you’re one of the kindest and most genuine men that I have met in my life. Thank you for always smiling and for your helping hands (please say hi to

Angus, Maeve and Collin for me). Sara, I was psyched when you joined our lab, a lady in the house. And I’m so happy it was you with your warm and vibrant personality. Thanks for everything, I’m going to miss you woman! Cheese, it has been an honor and so much fun. Thank you for your spss-tricks, your one-of-akind expressions and for all the insanity. Noor, although you came to Boston as a research-student, it felt like you were a colleague. Thanks for your hard work, your laughs and for sharing the love for chocolate. Thank you guys, you’re not just colleagues, but life-long friends. Dear Dr. Landon, I have sincerely enjoyed working with you. I admire your decisiveness, your writing-skills and your knowledge. Larry and James, thank you for the countless conference calls and all the stats-lessons, it was great working with you. Additionally, I would like to thank all the co-authors who I haven’t mentioned before. Thank you to all the attendings, residents, nurses, NP’s, researchers and assistants at the Department of Vascular Surgery at Beth Israel Deaconess Medical Center. Special thanks to Dr. LoGerfo, who taught me how to tie the perfect knot and invited me to his home for Christmas, which was absolutely lovely. Thank you Carla for sharing laughs and for being a great friend, I’m looking forward to seeing you in Holland soon. Mary, thank you so much for all your help, couldn’t have done it without you. Everyone at UMCU, thanks for making me feel like home every time I came back. And a million thanks for taking care of my dad so extremely well the past months, I can’t thank you all enough. Special thanks to Jacqueline, my fellow little pony. I’m on the lookout for a mini-pig that we can adopt. And special thanks to Luts and Kari, fellow researchers, for sharing the Boston adventure with me. Cobie and Susan, thank you for all your help and the nice chats. Thanks to Big G, Wouter, Foeke and Thomas, who, although in different cities, enjoyed life in the United States with me. I still hope to be seeing you at vascular meetings in the future. All my other incredible friends in Boston, who made my stay even more valuable, thank you. Thanks to all the players at Myopia and Billiebo, for all the great weekends, the fast polo games, and glasses of wine. Phil aka Queen B, Vicky, Sky, Norma and Wink the dink, I’ll miss you. Karates 06, thanks for the great dinners, drinks and trips, which have probably contributed very little to this book. Hatsaaa, my medicine-buddies Pien, Daf, Sjoerd, Henk, Mie, Ballo, en Allie, thanks for the endless moments I shared with you all over the world. What a fantastic group of friends. SPN, it was awesome to be in beantown together.

‘T bootje’, Anna, Marieke, Aniek, Roos en Fleur thank you for always being there for me and for visiting me wherever I go. I’m so happy we found each other 10 years ago. We’ve laughed and cried together, and although we don’t always agree, who doesn’t enjoy a good discussion?! Our recent USA/Canada trip was a memorable one and I toast to a long life ahead of us. And Anna, thank you for everything! You understand me without words, but I still have a few for you: my kidney is your kidney. Maja, kochana sissy, we grew up in Poland together and still stand side by side. I’m sure we’ll follow each other everywhere we go and we’ll be dancing and singing until we grow old and wrinkled. Thank you for always being there. Pieter, I can’t imagine what Boston would have been like without you. We shared our work, social life and many magnificent trips together. Thank you for your friendship, your humor and your advice. I still laugh my heart out when I think of Shaws’ monopoly, our home-made Old Dutch Games or gazing into the crowd after a joke. Lisa, you’re the sister I never had and my paranypmh. Thank you for everything! Although we’re quite the opposite in character, we love the same things in life. I’m extremely happy we ended up in Boston together, both doing research, and I’m sure we’ll never forget to let those extremities go wild. Joepie, I’m proud to have such a wonderful big little brother. You never stop to amaze me with everything you have up your sleeve. I love your insanity, your enthusiasm and I’m extremely glad to have you as my paranypmh. Hè? Tijn, thank you for supporting me in my career, my hobbies and my other unusual ideas. Never a dull moment right? Thank you for your help with practically all the figures in this book and with the entire lay-out. The best thing is, the long distance is over and the best is yet to come. Love you. Mom and Dad, thank you for your unconditional support and love in every step of the way. You’ve taught me how to make decisions, how to accomplish goals and how to enjoy the little things in life. You have made me who I am today. We’re having a tough year with dad being sick, but we’re going to enjoy every last bit of it together. And this one is for you Daddy!

Curriculum Vitae Auctoris

Dominique Babette Buck was born on July 14th, 1987, in Heemstede, The Netherlands. She grew up partly in Holland and partially in Poland. She studied medicine at VU University in Amsterdam and Slavic Languages and Cultures at the University of Amsterdam. In her college years she sang in a band and was an avid horse rider. During medical school, she spent months abroad in München (Germany), Poznan (Poland) and Oranjestad (Aruba) for internships and research experience. In 2012 she started as a resident at the department of Surgery at the University Medical Center in Utrecht. A year later, she moved to Boston (USA) for a 2-year research fellowship at Beth Israel Deaconess Medical Center, Harvard Medical School to work with Dr. Marc L. Schermerhorn. Her dog ‘Bruno’ moved with her to the United States and enjoyed it as much as she did. Dominique will be defending her PhD thesis ‘Innovations in Abdominal Aneurysm Repair’ at the University of Utrecht on May 11th, 2015.

Innovations in Abdominal Aneurysm Repair Dominique B. Buck