Resident and fellow experiences after the introduction of endovascular aneurysm repair for abdominal aortic aneurysm

Transcription

1 EDUCATION CORNER From the New England Society for Vascular Surgery Resident and fellow experiences after the introduction of endovascular aneurysm repair for abdominal aortic aneurysm Teviah Sachs, MD, a Marc Schermerhorn, MD, a Frank Pomposelli, MD, a Philip Cotterill, PhD, b James O Malley, PhD, c and Bruce Landon, MD, c Boston, Mass; and Baltimore, Md Objectives: This study assessed trends in open and endovascular repair (EVAR) of intact and ruptured abdominal aortic aneurysm (AAA) in the Medicare population and evaluated recent trends in AAA repair at vascular fellowship training programs. Methods: We identified all Medicare beneficiaries with a diagnosis of AAA who underwent repair or had a primary diagnosis of rupture ( ). Cohorts were compared by type of repair (open vs EVAR) and presentation (intact vs ruptured AAA). Demographics of age, sex, and race were evaluated. We used unique hospital identifier codes to compare trends and 30-day mortality between hospitals that participate in vascular surgery fellowship training and those that do not. American Council on Graduate Medical Education data, only available for the years 1999 to 2008, were further used to better understand the changes in number of EVAR and open repairs of AAA performed each year for vascular fellows and general surgery residents, over time. Results: We identified 449,122 patients (76% men), with 376,355 intact AAAs (84%) and 72,767 ruptured AAAs (16%). Mean age was 75.1 years. Use of EVAR for intact AAA rose to from 35% in 2001 to 63% in 2005 and comprised 78% of repairs by During the same period, the number of ruptured AAAs decreased by 40% overall, with nonoperative ruptured AAAs decreasing by 29% and EVAR increasing to 31% of rupture repairs. Hospitals training vascular fellows were quicker to adopt EVAR (2-year lag time) for intact AAA and had higher rates of EVAR for ruptured AAA (41.1% vs 29.2%; P.001) than did hospitals without fellows. Mortality rates for open repairs of intact (4.0% vs 5.0%; P.01) and ruptured AAA (34.1% vs 41.0%; P.031) were lower at fellowship hospitals. The average number of open AAA repairs performed by vascular fellows dropped 50% (44.1 to 21.6/year) from 1999 to Conclusions: Contrary to the expectation of a plateau, use of EVAR for intact AAA continues to rise at fellowship and nonfellowship hospitals. Use of EVAR for rupture is being used more often at fellowship programs. The decline in open repairs performed by vascular fellows, and at fellowship and non-fellowship hospitals, may have important implications for future attending experience. (J Vasc Surg 2011;54:881-8.) From the Division of Vascular and Endovascular Surgery, Beth Israel Deaconess Medical Center, Boston a ; the Centers for Medicare and Medicaid Services, Baltimore b ; and the Department of Health Care Policy, Harvard Medical School, Boston. c Competition of interest: Dr Schermerhorn has been paid a consulting fee by WL Gore, Medtronic, Endologix, and Nellix. Presented at the Thirty-seventh Annual Meeting of the New England Society for Vascular Surgery, September 24-26, 2010, Rockport, Me. Correspondence: Marc Schermerhorn, 110 Francis St, 5th Fl, Boston, Mass ( mscherm@bidmc.harvard.edu). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a competition of interest /$36.00 Copyright 2011 Published by Elsevier Inc. on behalf of the Society for Vascular Surgery. doi: /j.jvs Since its advent in 1991, 1 endovascular aortic aneurysm repair (EVAR) has become a mainstay of treatment for abdominal aortic aneurysm (AAA) requiring operative intervention. 2-4 Previous studies, such as the Comparison of Endovascular Aneurysm Repair with Open Repair in Patients with Abdominal Aortic Aneurysm (EVAR 1) trial, have shown its effectiveness in reducing perioperative mortality, 3,5,6 and this difference is even more pronounced in older patients. 7,8 Still, for patients deemed unsuitable for open repair (EVAR trial 2), perioperative mortality with EVAR was nearly 9%, demonstrating that utility of EVAR has a limit. 9 Recent national data national data show that the percentage of AAA repairs being performed via EVAR has increased dramatically over time. 2-4,6 As recently as 2001, over two-thirds of AAA repairs were performed using open repair. 10 Since that time, the number of open repairs has decreased by 61%. 2,4 Previous research shows that outcomes of open repair are strongly related to surgeon and institutional volume Therefore, the effect of the decrease in open repairs on trainee experience has been of 881

2 882 Sachs et al JOURNAL OF VASCULAR SURGERY September 2011 growing interest and concern during the past decade. Between 2001 and 2007, the number of open aortic aneurysm procedures performed per vascular trainee decreased by almost 22%, whereas general surgery residents experienced a decrease in all operative aneurysms endovascular and open repairs between 1997 and The ability of future vascular surgeons to safely and effectively perform open aneurysm repairs is, therefore, of genuine concern. To better understand these trends, we used data from the Medicare program, which pays for 70% of U.S. AAA repairs, and data on trainee procedure volume, to describe current trends in AAA repair for intact and ruptured aneurysms and to further elucidate the potential effects on both future resident and fellow training and attending experience. METHODS Patient identification. We used comprehensive data from the Medicare program to identify all AAA repairs or rupture that occurred during the time period 1995 to Cases included open and endovascular repair, as well as intact and ruptured aneurysms. Using International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) diagnosis and procedure codes, we identified all Medicare beneficiaries with Part A coverage at admission, who were aged 65 years. These patients were then stratified by diagnosis code (abdominal aneurysm without mention of rupture) for intact AAA or code (abdominal aneurysm, ruptured) for ruptured AAA. These groups were further categorized as those who underwent open repair by code (resection of abdominal aorta with replacement) or (aortoiliac femoral bypass) or endovascular repair by code (endovascular implantation of graft) of their AAA. The code identifying endovascular repair was introduced in October For prior years, the code (insertion of non-drug-eluting peripheral vessel stent) combined with a principal diagnosis code of AAA as the reason for admission, was used to identify patients undergoing endovascular repair of AAA. Patients with the ICD-9 diagnosis code of ruptured AAA (441.3), but who did not have an associated procedure code, were identified as not having undergone repair of their ruptured AAA. We next categorized patients into four groups: (1) AAA alone, (2) AAA with thoracic involvement, (3) AAA site unspecified, and (4) AAA with visceral involvement. We further defined those with a concomitant diagnosis of aortic dissection based on diagnosis codes to (dissection of aorta: unspecified, thoracic, abdominal, thoracoabdominal) or procedure code (fenestration of dissecting aneurysm of thoracic aorta). Aneurysm involvement of the thoracic aorta was defined based on diagnosis codes (thoracic aneurysm, ruptured) or (thoracoabdominal aneurysm, ruptured) for ruptured aneurysm, or codes (thoracic aneurysm without mention of rupture) or (thoracoabdominal aneurysm, without mention of rupture) for intact aneurysm, or by procedure codes (resection of thoracic vessel with replacement) or (endovascular implantation of graft in thoracic aorta) for repair of the thoracic aorta. The AAA site unspecified group was identified as those without dissection or thoracoabdominal codes as above, and then further separated as ruptured by code (aortic aneurysm of unspecified site, ruptured) or intact by code (aortic aneurysm of unspecified site without mention of rupture). Finally, the visceral group was identified as those who had the code (resection of vessel with replacement of abdominal arteries: celiac, gastric, hepatic, iliac, mesenteric, renal, splenic, umbilical, excluding the abdominal aorta) or code (endarterectomy of abdominal arteries: celiac, gastric, hepatic, iliac, mesenteric, renal, splenic, umbilical, excluding the abdominal aorta). Those patients who did not have any of the subgroup classifications aortic dissection, AAA with thoracic involvement, AAA site unspecified, and AAA with visceral involvement were identified as strictly AAA, ruptured or intact, and were the primary population of study. Patient characteristics. Beneficiary demographic characteristics were obtained from the Medicare Provider Analysis and Review (MedPAR) files and included age, sex, and race (white, black, or other). Identifying fellowship hospitals. Using data available from the American Council on Graduate Medical Education (ACGME), we identified all vascular surgery fellowship programs in the United States. We then examined data available from each fellowship program to identify hospitals that participated in the operative training of vascular fellows and linked our data using unique hospital identifiers available through Centers for Medicare and Medicaid Services (CMS). We then identified patients who were treated at hospitals that participated in vascular surgery fellowship programs. Veterans Affairs hospitals are not included in CMS data. Trainee procedure volume. We identified the number of procedures done per year by vascular fellows and general surgery residents for all years available through the ACGME (1999 to 2008). Data regarding these resident and fellow case logs were obtained directly from the ACGME. 15 We focused on the following measures, which were available consistently during the 10-year period: mean AAA procedures per year, total numbers of AAA procedures per year, as well as the total number of residents, fellows, and ACGME-accredited residency and fellowship programs per year. Outcomes. Perioperative mortality was compared between fellowship and nonfellowship programs, and was defined as in-hospital death or death 30 days of the procedure. Date of death was identified from the MedPAR files. Statistical analysis. Our data are largely descriptive. We used 2 analysis for comparisons between categoric variables, and significance was established at P.05. Continuous variables were analyzed using the t test or

3 JOURNAL OF VASCULAR SURGERY Volume 54, Number 3 Sachs et al 883 Table I. Number of patients undergoing abdominal aortic aneurysm (AAA) repair in the Medicare population per year: Year Code for AAA only Code for AAA and Intact Ruptured Dissection TAA Visceral bypass Unspecified Total , , , , , , , , , , , , , , , , , , , , , , , , , , , ,753 Total 376,355 72, , , ,316 TAA, Thoracoabdominal aneurysm. Wilcoxon rank sum test, depending on the normality of the distribution. Time trends were analyzed using a 2 test of trend, and significance was achieved at P.01. All statistical analyses were performed using SAS 9.2 software (SAS Institute, Cary, NC). RESULTS We identified 502,316 Medicare beneficiaries from 1995 to 2008 who underwent repair of an aortic aneurysm, of which 449,122 (89.4%) were strictly AAA repairs (Table I). The annual number of patients with an AAA-related repair increased from 30,658 to 36,459 (18.9%) during the the 14-year period. For patients identified as strictly AAA, there were 376,355 intact AAAs (84%) and 72,767 ruptured AAAs (16%). AAA patients. Mean age for patients undergoing AAA repair alone was 75.1 years, and 76% were male (Table II). Patients with ruptured AAA who did not undergo an operation were older (80.2 years) than those undergoing intact repairs (open, 74.1 years; EVAR, 75.7 years) and rupture repairs (open, 75.9 years; EVAR, 77.1 years), and a greater proportion of nonoperative ruptures were in women (37.6% vs 24%; P.01). Race was predominantly white for all groups. The use of EVAR for intact AAA rose from 0 in 1995 to 25,192 (78% of repairs) by 2008 (Fig 1). At the same time, open repairs decreased from 23,968 to 7190 cases, representing a 70.1% decrease in volume of repairs. A greater proportion of EVAR patients were men (80.1% vs 74.8%; P.03). Over time the percentage of men in the open repair group dropped, from 76.6% in 1995 to 69.7% in For EVAR, the percentage of patients undergoing repair who were male increased from 73.1% in 1996 to 79.8% in 2008, reaching a peak of 81.9% in There was no significant difference between patients undergoing open repair or EVAR in age at time of operation or for race, either aggregate or over time. Ruptures, both operative and nonoperative, decreased during the 14-year period by 40% (6690 to 4077; P.01; Fig 2). In total, 68.5% of ruptured AAAs were repaired, with a decrease over time from 70.3% in 1995 to 65.6% in Ruptures undergoing open repair decreased from 4700 to 1839 (69% of those undergoing repair), whereas those undergoing EVAR grew from 0 to 834 (31% of those undergoing repair). Nonoperative ruptures dropped from 1990 to 1404, and their proportion of total ruptures rose from 29.7% (n 6690 total ruptures) to 34.4% (n 4077 total ruptures). Repairs performed at fellowship vs nonfellowship hospitals. Between 1995 and 2008, 14% of total AAA repairs (14.5% of intact AAA, 10% of ruptured AAA) were performed at fellowship hospitals. Fellowship hospitals adopted EVAR for intact AAA repair more quickly than nonfellowship hospitals, achieving 50% EVAR for intact repairs more than 2 years before nonfellowship hospitals (Fig 3). By 2008, however, the percentage of repairs using EVAR was similar between groups. A greater proportion of ruptured AAAs were repaired using EVAR at fellowship hospitals in 2008 compared with nonfellowship hospitals (41.1% vs 29.2%; P.001). During the 14-year period, the aggregate mortality rate for intact AAA repair was significantly lower for fellowship hospitals for open repair (4.0% vs 5.0%; P.01) but not for EVAR (1.6% vs 1.6%; P.87). Mortality for EVAR decreased for both fellowship and nonfellowship hospitals over time (Fig 4). Mortality for open rupture repairs (35.6% vs 44.6%; P.03) and EVAR rupture repairs (23.3% vs 33.6%; P.02) were both significantly lower in fellowship hospitals. Trainee volume: 1999 to 2008 (ACGME). The number of vascular fellows accepted to fellowship each year increased from 97 in 1999 to 116 in Similarly, the number of fellowship programs increased from 83 to 90 during the same period. The number of general surgery residency programs decreased from 252 to 224

4 884 Sachs et al JOURNAL OF VASCULAR SURGERY September 2011 Table II. Characteristics of Medicare patients undergoing abdominal aortic aneurysm (AAA) repair: Demographics Age Male sex N Mean age P n % P All patients 449, , Non-op rupture 22, , Open EVAR Open EVAR Group N Mean age n Mean age n % n % Intact AAA repair 232, , , , Operative rupture 46, , Subgroups Dissection Thoracoabdominal 19, Visceral bypass Unspecified 21, , Race White Black Other race Demographics n % P n % P n % P All patients 421, %.5 10, %.1 Non-op rupture 21, % % % Open EVAR Open EVAR Open EVAR Group n % n % N % n % n % n % Intact AAA repair 219, , %.81 Operative rupture 43, %.39 Subgroups Dissection %.48 Thoracoabdominal 14, %.51 Visceral bypass %.45 Unspecified 20, %.79 EVAR, Endovascular aneurysm repair during the 10-year period, but the number of general surgery residents increased from 989 to The average number of open aortic aneurysm repairs performed by vascular fellows between 1999 and 2008 decreased by 50%, from (73% elective infrarenal, 9% ruptured infrarenal, 18% suprarenal) to (66%, 18%, 16%, respectively). During the same period, the average number of EVAR by vascular fellows increased almost threefold, from repairs to repairs, reaching a peak of repairs in 2006 (Fig 5, A). General surgery residents in their chief year performed an average of open aneurysm repairs in 2008, down from in In 2008, chief residents averaged EVARs (Fig 5, B). DISCUSSION Despite the expectation of a plateau, 16,17 the number of AAAs being repaired with EVAR continues to rise. In 2008, 78% of AAAs in the Medicare population were repaired with EVAR. Moreover, nearly one-third of ruptured AAAs were repaired with EVAR. Although hospitals identified as those that train vascular fellows were quicker to adopt EVAR, hospitals without fellows likely in the community setting have similarly adopted EVAR as the standard of care for most elective AAA repairs. As a consequence, the number of open repairs performed by vascular surgery fellows and, in particular, general surgery residents, has fallen dramatically. The mortality rates for ruptured AAA have also decreased at both fellowship and nonfellowship hospitals, while at the same time the number of ruptured AAAs has decreased. Hospitals identified as training vascular fellows were earlier to adopt EVAR for both intact and ruptured AAA. These hospitals also show improved mortality rates for open AAA repair, intact and ruptured, as we might expect with high-volume centers. 7 EVAR has become the standard adopted for elective repair of AAA in fellowship training hospitals as well as in those without fellows, and the mortality rates of EVAR for intact and ruptured aneurysms are similar at both types of hospital.

5 JOURNAL OF VASCULAR SURGERY Volume 54, Number 3 Sachs et al 885 All Intact AAA Repair Open AAA Repair EVAR # Intact AAA Repairs % 69% 78% % Year Fig 1. The number of repairs of intact abdominal aortic aneurysm (AAA) performed by open and endovascular (EVAR) techniques in the Medicare population are shown from 1995 to Fig 2. Number of repairs of ruptured abdominal aortic aneurysm (AAA) performed by open or endovascular (EVAR) techniques and nonrepaired AAA in the Medicare population is shown from 1995 to Most vascular fellows are performing fewer open AAA repairs with each successive year, raising important concerns that many may not possess the foundation necessary to effectively perform an open AAA repair when they complete their fellowship. Furthermore, it has been shown that most vascular surgery fellows expect to practice at least partially in an academic center. 18 Therefore, this decrease in exposure to open aortic procedures, while important for academic centers, has an even greater impact for the rural and community centers where vascular procedures are often performed by general surgeons. The ACGME data show that the average general surgery resident is not ade-

6 886 Sachs et al JOURNAL OF VASCULAR SURGERY September 2011 Fig 3. Percentage of intact and ruptured abdominal aortic aneurysm (AAA) repairs performed using endovascular (EVAR) techniques in the Medicare population in fellowship and nonfellowship hospitals is shown from 1995 to Fig 4. Mortality rates of intact abdominal aortic aneurysm (AAA) repairs performed using endovascular (EVAR) and open techniques in the Medicare population, in fellowship and nonfellowship hospitals, are shown for 1995 to The percentages represent 2008 mortality rates.

7 JOURNAL OF VASCULAR SURGERY Volume 54, Number 3 Sachs et al 887 Fig 5. A, Average number of abdominal aortic aneurysm (AAA) repairs performed using endovascular (EVAR) and open techniques by vascular fellows by American Council of Graduate Medical Education (ACGME) reporting are shown for 1999 to B, Average number of AAA repairs (open and EVAR) by general surgery chief residents by ACGME reporting is shown for 1999 to quately trained to perform AAA repair after graduation, having performed, on average, just one to two open AAA repairs per year. As has been shown in the literature, large-volume centers demonstrate superior outcomes for ruptured and elective AAA repair. 7,12 Most of these ruptured AAA repairs are still being performed with an open technique. Yet, if the trends we see in Fig 5 continue, EVAR will likely become even more common in nonfellowship and community hospitals, and open repairs will be more commonly performed in large training hospitals. This illustrates the importance of early diagnosis of ruptured AAA and prompt transfer to an appropriate facility for operative management for these patients, where they have the best chance of survival. 19,20 Our study has certain limitations. The inability to specifically quantify or distinguish the complexity of presenting symptoms or the individual patient anatomic variations, can lead to considerable selection bias. Rödel et al 21 showed that even highly trained endovascular surgeons demonstrate substantial variability in their estimation of individual patient suitability for EVAR. Furthermore, our

8 888 Sachs et al JOURNAL OF VASCULAR SURGERY September 2011 distinction of hospitals as either fellowship or nonfellowship does not incorporate VA hospitals, where many fellows undergo a portion of their training. However, it is unlikely that this would bias our data substantially because VA patients are excluded from both groups. Finally, the use of ACGME data precludes further analysis beyond what is provided in their database and these data could be linked to specific surgical procedures to examine outcomes of care. CONCLUSIONS Use of EVAR continues to increase for both elective and ruptured AAA. Mortality for intact AAAs repaired with EVAR is lower than with open repair at both fellowship and nonfellowship hospitals. EVAR is being used for ruptured AAA at higher rates in hospitals with fellowship training. At the same time, fellowship hospitals demonstrate lower mortality rates with open AAA repair and EVAR than nonfellowship hospitals. General surgery residents are inadequately trained to perform aortic repairs upon graduation. At the same time, vascular fellows performed 50% fewer open aortic repairs in 2009 than in 1999, which may have a far-reaching impact on their future ability to perform these operations safely and effectively. AUTHOR CONTRIBUTIONS Conception and design: TS, MS, FP, PC, JO, BL Analysis and interpretation: TS, MS, PC, JO, BL Data collection: TS, PC Writing the article: TS Critical revision of the article: TS, MS, FP, JO, BL Final approval of the article: TS, MS, FP, PC, BL Statistical analysis: TS, MS, JO, BL Obtained funding: MS, FP, BL Overall responsibility: TS REFERENCES 1. Parodi JC, Palmaz JC, Barone HD. Transfemoral intraluminal graft implantation for abdominal aortic aneurysms. Ann Vasc Surg 1991;5: Giles KA, Pomposelli F, Hamdan A, Wyers M, Jhaveri A, Schermerhorn ML, et al. Decrease in total aneurysm-related deaths in the era of endovascular aneurysm repair. J Vasc Surg 2009;49: Schermerhorn ML, O Malley A, Jhaveri A, Cotterill P, Pomposelli F, Landon BE, et al. Endovascular vs. open repair of abdominal aortic aneurysms in the Medicare population. N Engl J Med 2008;358: Schwarze ML, Shen Y, Hemmerich J, Dale W. Age-related trends in utilization and outcome of open and endovascular repair for abdominal aortic aneurysm in the United States, J Vasc Surg 2009;50: e2. 5. Greenhalgh RM, Brown LC, Kwong GP, Powell JT, Thompson SG, et al. Comparison of endovascular aneurysm repair with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day operative mortality results: randomized controlled trial. Lancet 2004;364: 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: Giles KA, Hamdan AD, Pomposelli FB, Wyers MC, Dahlberg SE, Schermerhorn ML, et al. Population-based outcomes following endovascular and open repair of ruptured abdominal aortic aneurysms. J Endovasc Ther 2009;16: Giles KA, Pomposelli FB, Hamdan AD, Wyers MC, Schermerhorn ML. Comparison of open and endovascular repair of ruptured abdominal aortic aneurysms from the ACS-NSQIP J Endovasc Ther 2009;16: 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 2005;365: Nowygrod R, Egorova N, Greco G, Anderson P, Gelijns A, Moskowitz A, et al. Trends, complications, and mortality in peripheral vascular surgery. J Vasc Surg 2006;43: Birkmeyer JD, Lucas FL, Wennberg DE. Potential benefits of regionalizing major surgery in Medicare patients. Eff Clin. Practitioner 1999; 2: Birkmeyer JD, Siewers AE, Finlayson EV, Stukel TA, Lucas FL, Batista I, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002;346: Landon BL, O Malley AJ, Giles K, Cotterill P, Schermerhorn ML. Volume outcomes relationships and AAA repair. Circulation 2010;122: Schanzer A, Steppacher R, Eslami M, Arous E, Messina L, Belkin M, et al. Vascular surgery training trends from : a substantial increase in total procedure volume is driven by escalating endovascular procedure volume and stable open procedure volume. J Vasc Surg 2009;49: Accreditation Council for Graduate Medical Education. Case log statistical reports. documentation/statistical_reports.asp. Accessed April Cotroneo AR, Iezzi R, Giancristofaro D, Santoro M, Quinto F, Spigonardo F, et al. Endovascular abdominal aortic aneurysm repair: how many patients are eligible for endovascular repair? Radiol Med 2006; 111: Arko FR, Filis KA, Seidel SA, Gonzalez J, Lengle SJ, Webb R, et al. How many patients with infrarenal aneurysms are candidates for endovascular repair? The Northern California experience. J Endovasc Ther 2004;11: Hingorani AP, Ascher E, Marks N, Shiferson A, Puggioni A, Tran V, et al. 219 vascular fellows perception of the future of vascular surgery. Ann Vasc Surg 2009;23: Vogel TR, Nackman GB, Brevetti LS, Crowley JG, Bueno MM, Banavage A, et al. Resource utilization and outcomes: effect of transfer on patients with ruptured abdominal aortic aneurysms. Ann Vasc Surg 2005;19: Hames H, Forbes TL, Harris JR, Lawlor DK, DeRose G, Harris KA, et al. The effect of patient transfer on outcomes after rupture of an abdominal aortic aneurysm. Can J Surg 2007;50: Rödel SG, Geelkerken RH, van Herwaarden JA, Kunst EE, van den Berg JC, van der Palen J, et al. Consistency in endovascular aneurysm repair suitability assessment requires group decision audit. J Vasc Surg 2006;43: Submitted Oct 4, 2010; accepted Mar 1, 2011.