The US AneuRx Clinical Trial: 6-year clinical update 2002

The US AneuRx Clinical Trial: 6-year clinical update 2002 Christopher K. Zarins, MD, for the AneuRx Clinical Investigators,* Stanford, Calif A total of 1193 patients with infrarenal abdominal aortic aneurysms were treated with the AneuRx Stent Graft System at 19 US investigational centers from 1996 to 1999. This report summarizes clinical data collected and analyzed as of August 28, 2002. There have been 10 late (>30 days) aneurysm ruptures, 8 late (>30 days) aneurysm-related deaths, and 38 late (>30 days) surgical conversions, including 8 for rupture. Kaplan-Meier analyses of the primary outcome measures at 4 years indicate the following: a freedom from rupture rate of 98.4%; a freedom from surgical conversion rate of 90.4%; a freedom from aneurysm-related death rate of 96.9%; and a probability of survival rate, based on all-cause mortality, of 62.4%. Secondary outcome measures at 4 years include stent graft patency in 96.4%, endoleak in 13.9%, aneurysm enlargement in 11.5% and stent graft migration in 9.5% of patients. These results provide evidence that the AneuRx Stent Graft System continues to be a safe and effective treatment option for appropriately selected patients with infrarenal abdominal aneurysms. (J Vasc Surg 2003;37:904-8.) The US AneuRx Clinical Trial (IDE No. G960016) was designed to evaluate the safety and effectiveness of the AneuRx Stent Graft System (Medtronic AVE, Santa Rosa, Calif) for the treatment of selected patients with infrarenal abdominal aortic aneurysms. Continuing long-term follow-up of all patients treated during the course of the trial is being carried out. Details of the study design, patient eligibility and 4-year results have been previously published. 1,2 The purpose of this report is to provide an update on the primary outcome measures for all patients treated during the course of the clinical trial. THE ANEURX CLINICAL TRIAL The US AneuRx clinical trial was conducted in three phases at 19 US investigational centers from 1996 to 1999. The Phase I study consisted of 40 patients treated from 1996 to 1997. The Phase II study included 424 patients treated with the AneuRx stent graft and 66 control patients treated with open surgical repair from 1997 to 1998. The Phase III study included 639 patients treated from 1998 to 1999. Patient selection criteria included an abdominal aortic aneurysm with an infrarenal neck length of at least 10 mm and a neck diameter between 18 and 26 mm with a maximum distal iliac diameter of 15 mm. During the course of the trial, 90 patients who did not meet the inclusion criteria were treated with the AneuRx stent graft in a high-risk study arm. These patients are included in the total of 1193 From Stanford University Medical Center, Stanford, California. *See Appendix 1 for list of principal investigators and clinical sites. Competition of interest: CKZ has served as a consultant to Medtronic AVE and has owned shares of Medtronic, Inc. Reprint requests: Christopher K. Zarins, MD, Stanford University Medical Center, Division of Vascular Surgery, 300 Pasteur Drive, H3642, Stanford, CA 94305-5642 (e-mail: zarins@stanford.edu). Copyright 2003 by The Society for Vascular Surgery and The American Association for Vascular Surgery. 0741-5214/2003/$30.00 0 doi:10.1067/mva.2003.311 904 patients treated during the trial and are considered in the long-term analysis. All patients in Phase I and 134 patients in Phase II (174 patients total) received a device design that incorporated a unibody stiff stent in the aortic body of the device. The remaining 1019 patients were implanted with a device design that incorporated a multisegmented flexible aortic body. This flexible device configuration is that which is currently marketed. This data analysis includes all 1193 patients implanted with both stiff-body and flexible devices. Clinical data and follow-up records were analyzed through August 28, 2002, 2 years and 11 months following the end of patient enrollment on September 30, 1999. More than 50% of all patients in the trial were enrolled during the last year of the trial and data at the 3-year time point are not yet available for the majority of patients. Thus 1-year data are available for 92% of all patients, 2-year data for 75% of patients, 3-year data for 43% of patients, and 4-year data for 24% of patients. A total of 120 patients (10%) have been lost to follow-up and have been terminated from the study. See Appendix 2 (online only) for additional information. PRIMARY OUTCOME MEASURES A total of 15 of the 1193 patients experienced rupture (1.3%) (Table I). Two patients with intraoperative rupture died during the initial stent graft procedure. Three patients with postoperative ( 30 days) rupture underwent surgical conversion with survival of one patient. Ten patients (0.8%) had late ( 30 days) aneurysm rupture. Eight patients with late rupture underwent emergent open surgical repair with survival of 5 patients; 2 patients ruptured and died without attempted surgical repair. Overall rupture mortality was 9 of 15 patients (60%). A total of 53 patients have undergone surgical conversion to open aneurysm repair, with 11 intraoperative conversions, 4 postoperative ( 30 days) conversions, and 38 late ( 30 days) conversions. Thirty of the 38 late surgical conversions were elective with 3 deaths

JOURNAL OF VASCULAR SURGERY Volume 37, Number 4 Zarins 905 Table I. Primary outcome measures Intraoperative 30 days* 30 days Total Aneurysm rupture 2/1193 (0.2%) 3/1193 (0.3%) 10/1193 15/1193 Conversion to surgical repair 11/1193 (0.9%) 4/1193 (0.3%) 38/1193 53/1193 Aneurysm-related death 0 22/1193 (1.8%) 8/1193 30/1193 Death (all causes) 0 22/1193 (1.8%) 228/1193 250/1193 *Postoperative, but within 30 days. Patients within this group are currently at a variety of follow-up periods; therefore, rates are not given for these columns. Please refer to Kaplan-Meier estimates for specific projections for patients 30 days post procedure. Aneurysm-related death is defined as any death occurring within 30 days of the initial treatment, a rupture, a conversion, or any other secondary stent graft procedure. Table II. Primary causes of surgical conversion Cause Endoleak w/aaa enlargement 18 Rupture 11 Migration/displacement of modular component 11 Failure to access 7 AAA enlargement 2 Endoleak 1 Other (pseudoaneurysm, aortic fistula, aneurysm 3 pressing on colon) Total 53 (10% elective conversion mortality); 8 late conversions were for rupture with 3 deaths (38% mortality for conversion following rupture). The most common causes of surgical conversion were endoleak with aneurysm enlargement (34%), followed by rupture (21%), migration or displacement of modular components (21%), and failure to introduce the device (13%) (Table II).There were 22 perioperative ( 30 days) aneurysm-related deaths (1.8%) of which 3 were following rupture and 1 following surgical conversion without rupture. There were 8 late ( 30 days) aneurysm-related deaths, of which 4 were following rupture, 3 following elective surgical conversion, and one 3 weeks following a secondary stent graft procedure for limb occlusion. There were 22 perioperative deaths and 228 late deaths, of which 220 were unrelated to the aneurysm or its treatment (Table I). KAPLAN-MEIER ANALYSES The freedom from all-cause rupture at 1 year is 99.5%, at 2 years, 98.5%, at 3 years, 98.4%, and at 4 years, 98.4% (Fig 1). The freedom from surgical conversion at 1 year is 98.5%, at 2 years, 96.9%, at 3 years, 94.2%, and at 4 years, 90.4% (Fig 2). The freedom from aneurysm-related death at 1 year is 98.2%, at 2 years, 97.5%, at 3 years, 96.9%, and at 4 years, 96.9% (Fig 3). The survival rate (based on all-cause mortality) at 1 year is 91.8%, at 2 years, 83.1 %, at 3 years, 76.6%, and at 4 years, 62.4% (Fig 4). Data related to Figures 1 through 4 are in Appendix 2 (online only). SECONDARY OUTCOME MEASURES At 4 years, primary stent graft patency was 96.4%, endoleak was present in 13.9% of patients, aneurysm enlargement 5 mm was present in 11.5% of patients, and stent graft migration was noted in 9.5% of patients (Table III). OTHER OUTCOME MEASURES Core laboratory analysis of imaging studies in Phase II patients revealed issues involving device integrity (stent fracture or suture pop) in only a small number of patients (Appendix 2, online only). Analysis of 115 stent grafts explanted at surgery or autopsy for the worldwide experience (1996-2002) revealed stent element fractures in 0.16% of all available stent elements, mostly from metal fatigue (2.1 fatigue fractures per implant). Fabric abrasion and weave separation resulting in holes 0.2 mm were seen in 29 of 81 explants with an average of 1.3 holes per explant. Suture breaks have been observed in 1.5% of all sutures explanted. The clinical significance of stent fractures, graft wear, and suture breaks is unclear and is discussed in more detail in Appendix 2 (online only). DISCUSSION More than 6 years have elapsed since the first AneuRx stent graft was implanted, and 3 years have elapsed since the last patient was enrolled and treated in the clinical trial. Although complete data are thus far available on less than 50% of trial patients at 3 years, meaningful conclusions regarding the primary endpoints can be drawn by using Kaplan-Meier analysis. At 4 years, freedom from rupture rate is 98.4%, freedom from aneurysm-related death rate is 96.9%, freedom from surgical conversion is 90.4%, and survival rate, based on all-cause mortality, is 62.4%. It should be noted that this patient cohort includes the initial patient experience with endovascular aneurysm repair for most of the 19 clinical centers in the trial and thus includes their primary learning curve. Furthermore, several important features, which are critical for long-term success using the AneuRx stent graft, were discovered after completion of patient enrollment in the trial. The first published reports of aneurysm ruptures

906 Zarins JOURNAL OF VASCULAR SURGERY April 2003 Table III. Secondary outcome measures Predischarge 1 month 6 months Stent graft patency 1105/1106 (99.9%) 1036/1041 (99.5%) 972/978 (99.4%) Endoleak 306/1103 (27.7%) 147/1056 (13.9%) 135/987 (13.7%) Aneurysm enlargement* N/A 48/523 (9.2%) 41/710 (5.8%) Stent graft migration N/A 3/947 (0.3%) 10/920 (1.1%) *This endpoint is defined as any increase 5mm as compared to predischarge measurement. Fig 1. Kaplan-Meier curve for freedom from rupture. Fig 2. Kaplan-Meier curve for freedom from surgical conversion.

JOURNAL OF VASCULAR SURGERY Volume 37, Number 4 Zarins 907 1 year 2 years 3 years 4 years 945/953 (99.2%) 762/773 (98.6%) 443/451 (98.2%) 132/137 (96.4%) 132/951 (13.9%) 129/772 (16.7%) 63/451 (14.0%) 19/137 (13.9%) 48/689 (7.0%) 46/551 (8.3%) 39/346 (11.3%) 15/130 (11.5%) 20/904 (2.2%) 39/732 (5.3%) 32/449 (7.1%) 13/137 (9.5%) Fig 3. Kaplan-Meier curve for freedom from aneurysm-related death. Fig 4. Kaplan-Meier curve for freedom from death.

908 Zarins JOURNAL OF VASCULAR SURGERY April 2003 following AneuRx stent graft repair appeared in 2000 and highlighted the importance of stent graft fixation to the infrarenal aortic neck, iliac arteries, and bifurcation junction gate, rather than to endoleak as a primary cause of failure. 3 Further detailed analysis of morphologic predictors of late adverse events demonstrated that preoperative aortic neck length and angulation, degree of device oversizing, and length of proximal and distal seal zones were critical to long-term success. 4 This information became known and was provided to clinical investigators after completion of the trial and was included in training sessions for new users of the device after approval by the Food and Drug Administration. Updated Instructions For Use for the AneuRx stent graft now indicate that neck length should be greater than 1 cm with avoidance of neck angulation 45 degrees and that a 15-mm proximal and a 25-mm distal (iliac) fixation seal zone should be achieved. It is not yet known whether this new information, along with improvements in intraoperative imaging and implantation technique, which became available after completion of the trial, will result in a reduction in the rate of late adverse events noted in this trial. REFERENCES 1. Zarins CK, White RA, Schwarten D, Kinney K, Diethrich EB, Hodgson KJ, et al, for the Investigators of the Medtronic AneuRx Multicenter Clinical Trial. AneuRx stent graft vs. open surgical repair of abdominal aortic aneurysm multicenter prospective clinical trial. J Vasc Surg 1999; 29:292-308. 2. Zarins CK, White RA, Moll FL, Crabtree T, Bloch DA, Hodgson KH, et al. The AneuRx stent graft: four-year results and worldwide experience 2000. J Vasc Surg 2001;33:S135-45. 3. Zarins CK, White RA, Fogarty TJ. Aneurysm rupture after endovascular repair using the AneuRx stent graft. J Vasc Surg 2000;31:960-70. 4. Fillinger M. Predictors of Adverse Outcomes Based on Analysis of AneuRx Patients. Presented at the meeting of the Society of Vascular Surgeons in Baltimore, MD in June 2001. In press. Submitted Jan 27, 2003; accepted Feb 14, 2003. Please see related article by Dorothy B. Abel on pages 902-3. Appendix I. List of AneuRx investigators and clinical sites Investigator Rodney A. White, MD Matthew Jung, MD Christopher K. Zarins, MD Marian F. McNamara, MD Jon S. Matsumura, MD Kenneth Ouriel, MD Samuel R. Money, MD Robert M. Bersin, MD Kim J. Hodgson, MD John R. Laird, MD Edward B. Diethrich, MD William M. Moore, Jr, MD Mark H. Wholey, MD Zvonimir Krajcer, MD Mark F. Fillinger, MD Ted R. Kohler, MD John A. Kern, MD William McMillan, MD Larry A. Scher, MD Clinical site Harbor UCLA Medical Center Baptist East Hospital Stanford University Medical Center St Vincent s Hospital Northwestern University Medical Center Cleveland Clinic Foundation Oschner Clinic Sanger Clinic, Pa SIU School of Medicine/Memorial Hospital Washington Hospital Center Arizona Heart Institute South Carolina Pittsburgh Vascular Institute St Luke s Episcopal Hospital Dartmouth-Hitchcock Medical Center University of Washington VA Puget Sound Health Care University of Virginia Minneapolis Heart Institute North Shore Long Island Jewish Health System

APPENDIX 2 (ONLINE ONLY). DECEMBER 2002 CLINICAL UPDATE VOL I: THE ANEURX STENT GRAFT SYSTEM This report was prepared by Medtronic AVE and approved by the FDA for distribution to physicians using the AneuRx Stent Graft System. It has been edited to comply with the format requirements of the Journal. The report is divided into four sections. The first section contains clinical endpoint data from the entire Medtronic AVE US IDE cohort, which includes 1193 patients and data to 4 years. These data include safety endpoints (patient mortality, rupture, surgical conversion) as well as key endovascular measures (endoleak, aneurysm size, migration, patency, and device integrity). The second section summarizes information regarding aneurysm rupture, surgical conversion, and mortality from our combined global commercial experience. The third section summarizes the results from our explant analyses, and the fourth section provides summary comments. The data in this report include information collected and analyzed as of August 28, 2002. Section I: Clinical trial data update A total of 1193 patients were treated with the AneuRx Stent Graft System at 19 US investigational centers from 1996 to 1999. These patients were enrolled in three study phases. Phase I, consisting of 40 patients, enrolled patients from June 1996 to April 1997. Phase II began in April 1997 and ended in September 1998, with 424 patients inclusion criteria were enrolled in a high-risk arm of the trial. These patients are included in the long-term analysis. Sixty-six patients were also enrolled in the control arm of the study, receiving standard surgical treatment for abdominal aortic aneurysm (AAA). These patients were evaluated for morbidity and mortality and compared with 416 Phase II patients for premarket approval. Beyond the 1-year follow-up period for these 66 patients, we have only actively collected mortality status. All patients in Phase I and 134 patients in Phase II (174 patients total) received a device design that incorporated a unibody stiff stent in the aortic body of the device. The remaining 1019 patients were implanted with a device design that incorporated a multisegmented flexible aortic body. This flexible device configuration is that which is currently marketed. The data below include all 1193 patients implanted with both stiff-body and flexible devices and include patient data through August 28, 2002. One patient is not included in the rupture, survival, and aneurysm-related death analyses because this patient received a conversion device. The conversion device was approved as a treatment arm of the investigational device exemption (IDE) and is intended for use with the bifurcated device. Table I provides a summary of clinical results relating to rates of rupture, conversion, aneurysm-related death, and all-cause mortality. Along with the intraoperative, less than or equal to 30 days, and greater than 30 days results, Kaplan-Meier summaries are provided for these endpoints for each year of follow-up, out to 4 years. A Table I. Rupture, conversion, and death Intraoperative 30 days* 30 days n Total n Kaplan- Meier summaries 1-year Kaplan- Meier 2-year Kaplan- Meier 3-year Kaplan- Meier 4-year Kaplan- Meier Aneurysm rupture 2/1193 (0.17%) 3/1193 (0.25%) Conversion to surgical repair 11/1193 (0.92%) 4/1193 (0.34%) Aneurysm-related 0 22/1193 death (1.8%) Death (all-cause) 0 22/1193 (1.8%) 10/1193 15/1193 Freedom from aneurysm Rupture 38/1193 53/1193 Freedom from conversion 8/1193 30/1193 Freedom from aneurysm-related death 228/1193 250/1193 Probability of survival (based on all-cause mortality) 99.5% 98.5% 98.4% 98.4% 98.5% 96.9% 94.2% 90.4% 98.2% 97.5% 96.9% 96.9% 91.8% 83.1% 76.6% 62.4% *Postoperative, but within 30 days. Patients within this group are currently at a variety of follow-up periods; therefore, rates are not given for these columns. Please refer to Kaplan-Meier estimates for specific projections for patients 30 days postprocedure. Aneurysm-related death is defined as any death occurring within 30 days of the initial treatment, a rupture, a conversion, or any other secondary stent graft procedure. enrolled and treated and one patient enrolled and not treated (the patient experienced a myocardial infarction after enrollment and prior to the procedure and did not receive the stent graft). Phase III enrolled 639 patients beginning in August 1998 and ending on September 30, 1999. Ninety additional patients not meeting the trial s breakdown of the number of patients followed through each follow-up period for the clinical data presented in this report is provided in Table XII. Tables II and III outline the primary causes of conversion and the primary categories of aneurysm-related death provided in Table I.

Table II. Primary causes of conversion Cause Endoleak w/aaa expansion 18 Rupture 11 Migration/displacement of modular component 11 Failure to access 7 AAA expansion 2 Endoleak 1 Other (pseudoaneurysm, aortic fistula, aneurysm pressing 3 on colon) Total 53 Table III. Primary categories of aneurysm-related death* Category Within 30 days of initial treatment 18 Within 30 days of rupture 4 Within 30 days of surgical conversion 7 Within 30 days of secondary stent graft procedure 1 Total 30 *Aneurysm-related death is defined as any death occurring within 30 days of the initial treatment, a rupture, a conversion, or any other secondary stent graft procedure. Kaplan-Meier analyses of device performance during the US IDE trial are provided in Tables IV through VII. Kaplan-Meier curves for this data analysis are shown in Figs 1 through 4 (print). Please note that there are too few patients at this time to draw any valid statistical conclusions for the 5-year probability. Ruptures: Kaplan-Meier analysis. As demonstrated in Table IV, patients treated with the AneuRx stent graft have a 4-year freedom from rupture rate of 98.4%. Surgical conversion: Kaplan-Meier analysis. As demonstrated in Table V, patients treated with the AneuRx stent graft have a 4-year freedom from surgical conversion rate of 90.4%. Aneurysm-related death: Kaplan-Meier analysis. Aneurysm-related death is defined as any death occurring within 30 days of the initial treatment, a rupture, a conversion, or any other secondary stent graft procedure. As demonstrated in Table VI, patients treated with the AneuRx stent graft have a 4-year freedom from aneurysmrelated death rate of 96.9%. Mortality (all cause): Kaplan-Meier analysis. As demonstrated in Table VII, patients treated with the AneuRx stent graft have a 4-year probability of survival (based on all-cause mortality) of 62.4%. Other outcome measures. The data presented in Table VIII contain hospital data as defined by the investigating physician. The decrease in denominators over time is mostly attributed to patients not having reached that follow-up period. The decrease could also be attributed to the data point not having been collected at the visit or the patient missing the follow-up visit. n n Core lab (Phase II patients)/hospital observations for device integrity. As part of the clinical study and routine follow-up, data were analyzed for device integrity. This endpoint was based on core lab determination of the device displaying a stent fracture or (probable/possible) suture pop. A total of 75 clinical trial patients had one or more device integrity reports. Forty-six Phase II patients had one or more device integrity reports as demonstrated by the core lab; seven patients did not have core lab reports but had noted hospital findings for possible stent graft integrity issues. Twenty-two Phase III patients had hospital reports only, as core lab evaluation was not required for Phase III. Core lab evaluations were based on review of plain abdominal radiographs for mostly Phase II patients (protocol requirement). The core lab device integrity report observations include the following: 39 unconfirmed or possible suture pop (also described as step off ); 8 endoleaks (of questionable origin or existence); 3 appearances of migration of the stent graft; 1 note of separation of cranial position of the graft; 1 note of iliac limb separation; 4 possible stent graft fractures (with no fracture noted by hospital assessment); and 8 reports of cannot determine device integrity. Individual patients could have more than one observation. Of the 46 patients with core lab evaluations, 40 patients had hospital reports of no device integrity issue noted, and 6 patients had possible graft integrity issues recorded as yes on the hospital case report forms (3 reports of graft wall defect, tear, or disruption ; 1 report of attachment system fracture ; 1 report of possible material defect in the right collateral limb ; and 1 report of suture pop). One patient with a report of graft wall defect underwent surgical conversion for endoleak and AAA expansion, and one patient had additional iliac limbs implanted to treat a persistent endoleak. None of the remaining four patients with both a core lab and hospital integrity concern had had an additional procedure of any kind. Of the total patients with a device integrity report of any kind, six patients underwent successful surgical conversions. Two of these conversions were elective. Of the four remaining patients, two patients did not have confirmed loss of device integrity by either hospital or core lab findings. During the successful surgical conversion procedure, one of the patients exhibited some blood flow through the stent graft, which had migrated due to lack of oversizing per the instructions for use. The aortic neck was observed to be thrombosed and calcified, which likely had a deleterious effect on the proximal seal. No stent ring fractures were observed. Some weave separation, minimal graft abrasion, and suture breaks were observed in the explanted device. This device did not contain the tighter fabric weave contained in the currently commercialized product. The second patient was considered to have stent graft integrity issues as evidenced by hospital assessment and an endoleak of questionable origin as noted by the core lab report. At 2-year follow-up, this patient presented with

Table IV. Ruptures: Summary of Kaplan-Meier analysis* Treatment to 6 months 6 months to 1 year 1 year to 2 years 2 years to 3 years 3 years to 4 years No. at risk 1193 1116 1002 714 292 No. of events 5 1 8 1 0 No. of censored 72 113 280 421 251 Cumulative censored 72 185 465 886 1137 Kaplan-Meier estimate 0.996 0.995 0.985 0.984 0.984 Standard error 0.0019 0.0021 0.0039 0.0044 0.0044 *For additional information on the number of patients followed at these time intervals, please refer to Table XII. Number of patients at risk at the beginning of interval (based on actual implantation duration). Patients are censored because their last follow-up has not reached the end of the time interval, due to all-cause death, due to surgical conversion, or because they are lost to follow-up. The total censored for all time intervals up to and including that specific time interval. Estimate made at end of time interval. One patient is not included. This patient received a conversion device. Table V. Surgical conversion: Summary of Kaplan-Meier analysis* Treatment to 6 months 6 months to 1 year 1 year to 2 years 2 years to 3 years 3 years to 4 years No. at risk 1193 1105 998 712 293 No. of events 16 2 14 14 6 No. censored 72 105 272 405 244 Cumulative censored 72 177 449 854 1098 Kaplan-Meier estimate 0.987 0.985 0.969 0.942 0.904 Standard error 0.0034 0.0036 0.0056 0.0089 0.0177 *For additional information on the number of patients followed at these time intervals, please refer to Table XII. Number of patients at risk at the beginning of interval (based on actual implantation duration). Patients are censored because their last follow-up has not reached the end of the time interval, due to all-cause death, due to surgical conversion, or because they are lost to follow-up. The total censored for all time intervals up to and including that specific time interval. Estimate made at end of time interval. One patient is not included. This patient received a conversion device. Table VI. Aneurysm-related death: Summary of Kaplan-Meier analysis* Treatment to 6 months 6 months to 1 year 1 year to 2 years 2 years to 3 years 3 years to 4 years No. at risk 1193 1112 999 709 291 No. of events 22 0 5 3 0 No. censored 59 113 285 415 249 Cumulative censored 59 172 457 872 1121 Kaplan-Meier estimate 0.982 0.982 0.975 0.969 0.969 Standard error 0.0039 0.0039 0.0047 0.0059 0.0059 *For additional information on the number of patients followed at these time intervals, please refer to Table XII. Number of patients at risk at the beginning of interval (based on actual implantation duration). Patients are censored because their last follow-up has not reached the end of the time interval, due to all-cause death, due to surgical conversion, or because they are lost to follow-up. The total censored for all time intervals up to and including that specific time interval. Estimate made at end of time interval. One patient is not included. This patient received a conversion device. abdominal pain radiating to the back. The aneurysm diameter had increased with evidence of a persistent endoleak. Type II endoleaks were noted at predischarge, 1-month CT scan, and 1-year CT scan, and coil embolization of the inferior mesenteric artery was performed during the follow-up period. During surgical conversion, a leak was noted in the contralateral limb. Upon explant, one fatigue fracture was noted which was not in the seal zone, and weave separation and suture breaks were also observed. The weave separation, noted on the contralateral limb of the explanted device, is likely related to the leak observed during surgical conversion. The patient was successfully

Table VII. Mortality (all cause): Summary of Kaplan-Meier analysis* Treatment to 6 months 6 months to 1 year 1 year to 2 years 2 years to 3 years 3 years to 4 years No. at risk 1193 1111 999 708 291 No. of events 61 35 87 38 24 No. censored 21 77 204 379 225 Cumulative censored 21 98 303 681 906 Kaplan-Meier estimate 0.948 0.918 0.831 0.766 0.624 Standard error 0.0064 0.0080 0.0115 0.0147 0.0305 *For additional information on the number of patients followed at these time intervals, please refer to Table XII. Number of patients at risk at the beginning of interval (based on actual implantation duration). Patients are censored because their last follow-up has not reached the end of the time interval, due to all-cause death, due to surgical conversion, or because they are lost to follow-up. The total censored for all time intervals up to and including that specific time interval. Estimate made at end of time interval. One patient is not included. This patient received a conversion device. Table VIII. Other measures (or efficacy outcomes ) Predischarge 1 month 6 months 1 year 2 years 3 years 4 years Endoleak 306/1103 (27.7%) 147/1056 (13.9%) Enlarged N/A 48/523 aneurysm* (9.2%) Prosthesis N/A 3/947 migration (0.3%) Patency 1105/1106 1036/1041 (99.9%) (99.5%) 135/987 (13.7%) 41/710 (5.8%) 10/920 (1.1%) 972/978 (99.4%) 132/951 (13.9%) 48/689 (7.0%) 20/904 (2.2%) 945/953 (99.2%) 129/772 (16.7%) 46/551 (8.3%) 39/732 (5.3%) 762/773 (98.6%) 63/451 (14.0%) 39/346 (11.3%) 32/449 (7.1%) 443/451 (98.2%) 19/137 (13.9%) 15/130 (11.5%) 13/137 (9.5%) 132/137 (96.4%) *This endpoint is defined as any increase 5mm as compared to pre-discharge measurement. converted; however, he expired 10 days later due to respiratory failure. The third patient had modular disconnection at 2 years by hospital assessment. At 28 months, the patient had a successful open repair because the proximal neck was too short to permit endovascular treatment. The explant analysis notes two fatigue fractures in the bifurcated segment and one fatigue fracture in the aortic cuff (none were in the seal zone, so they probably were not the cause of migration), an abrasion hole at the crown edge (totally obscured by tissue without evidence of leakage in the adjacent tissue), and consecutive suture breaks in the distal iliac cuff limb (not likely a cause of proximal endoleak). The fourth patient had a successful surgical conversion at 25 months. The hospital assessment at 2 years was modular disconnection. Angiogram revealed separation between the stent body and the cuff. There was evidence of a Type III endoleak, modular component disconnect, and evidence of separation of stent rings. Further assessment noted that the aortic extension cuff was implanted in an angled neck. This angle was also associated with the implant of the bifurcated segment. During explant analysis, stent fractures, graft abrasion, and suture breaks were observed. Of the remaining patients, 21 underwent an additional procedure (additional stent graft component placed, coil embolization or ligation of arteries for treatment of Type II endoleaks, or Dacron taping via aneurysmorrhaphy) for any reason. The remaining 48 patients did not require any additional procedure. Careful patient selection and appropriate follow-up are important factors for success with endovascular stent grafting. The recently updated AneuRx Stent Graft Instructions for Use outlines the following critical factors for successful clinical outcomes: Appropriate patient selection, including a 15-mm aortic seal zone and 25-mm iliac seal zones and aortic neck angulation of 45 degrees; 10% to 20% oversizing; effective device deployment in accordance with the instructions for use; appropriate and timely patient follow-up. Section II: Commercial experience/product performance update As of August 30, 2002, more than 25,000 AneuRx devices have been distributed worldwide. From this commercial experience, Medtronic AVE gathered product performance information via field reports as well as from the active product and explant analysis programs. Table IX summarizes all reports received from commercial experience with the AneuRx Stent Graft through August 30, 2002.

Histogram of implant duration in months (N 115). Table IX. Commercial experience/product performance update AE type US 9/28/99 to 8/30/02 International 11/97 to 8/30/02 Rupture (total) 30 13 Preoperative* 6 Perioperative 8 Postimplant 16 13 Conversion 103 38 Aneurysm-related death (any 78 11 death occurring within 30 days of the initial treatment, a rupture, a conversion, or any other secondary stent graft procedure) Number of bifurcated devices distributed 20,694 4714 *The preoperative ruptures occurred in patients who had a rupture prior to inserting the AneuRx Stent Graft. The treating physician attempted rupture treatment with the endovascular stent graft rather than open surgical repair due to the patient s pre-existing comorbidities, surgical risk, and emergent circumstance. Section III: Explant analyses Medtronic AVE has analyzed AneuRx devices that were implanted and subsequently explanted either at surgery or at autopsy. As of August 30, 2002, Medtronic AVE has completed and submitted to the FDA individual case analyses (including summary reports and individual case reports) on 115 explants from patients implanted worldwide (Table X) from 1996 to 2002. A summary of these 115 explants is provided in the following pages. Each explanted device received by Medtronic AVE is scientifically evaluated by Medtronic AVE personnel as well as outside technical experts where appropriate. The detailed findings are then evaluated in conjunction with the patientspecific clinical information to assess the potential clinical significance, if any, of the device findings. Patient anatomy and the procedural technique used by the physician are also assessed. Findings are presented to the implanting physician and appropriate regulatory bodies, including the FDA. Explant summary of observations. The Figure shows the distribution of implant duration for these 115 explants. The explants had been implanted for an average of 16.9 months. Ninety-five of 115 explants (83%) had been implanted for more than 1 month. Table XI shows the primary reasons for explant as determined by the treating or explanting physician. Perioperative implants ( 1 month) and longer-term explants ( 1 month) are given separately. Explant device observations. As part of an individual analysis of each explanted device, the integrity of the components of the stent graft is examined. This examination includes the stents, graft material, and sutures. Stent observations. Nitinol stent strut fractures have been observed in devices that have been surgically explanted as well as removed at autopsy. Fractures can be categorized as due to single event overload or fatigue. It is likely that the majority of single event overload fractures occur during the explant procedure itself. The AneuRx stent graft is designed to contain numerous and redundant stent strut elements. For example, bifurcated stent grafts contain between 728 and 1020 of these elements per device, and the observed rate of stent fractures has been very small. The rate of occurrence of stent element fractures as a percentage of available stent elements was 0.16% for the 115 explants, of which 0.13% are fatigue fractures. An average of 2.1 fatigue fractures per explant have been observed for the 115 explants. Fatigue fractures have been observed in 57 of the 115 explants, with a range of 1 to 29 fatigue fractures. These fatigue fractures appear to be clustered in the midaortic stent rings, generally outside the seal zones of the bifurcated device. The majority of fatigue fractures observed during explant analysis do not appear to be related to adverse clinical outcomes; however, it appears that severe aortic neck angulation may be associated with the development of fatigue fractures. Patient selection and follow-up in accordance with the instructions for use is recommended. Graft material and suture observations. The AneuRx stent graft includes a polyester graft material. A change in the material was implemented after December 1997 during the IDE clinical trial. This change to a tighter fabric weave reduced the porosity of the graft material. We denote this

Table X. Sources of explanted devices Number from surgical conversions Number from autopsies following rupture Number from incidental autopsy Total From IDE clinical study (of 1193 patients enrolled) 31 1 13 45 From US commercial sales (of 20,694 bifurcated 30 0 7 37 devices distributed as of August 2002) From commercial sales outside of the US (of 4714 27 0 6 33 bifurcated devices distributed as of August 2002) Totals 88 1 26 115 newer version as the Reduced Porosity Material (RPM). The original version is referred to as pre-rpm. Eighteen of the 115 devices were made with the pre-rpm material, and 97 of the 115 devices were made with the RPM material. Observations related to graft fabric include weave separation and abrasion. The RPM, with its tighter weave, exhibits less weave separation. Weave separation and abrasion occur predominantly in the ipsilateral and contralateral iliac legs. Both have also been observed, to a lesser extent, directly below the aortic seal zone. The presence of weave separation or abrasion is typically observed in areas of increased vessel tortuosity. A portion of the RPM devices could not be evaluated for weave separation and abrasion, as they had been destructively analyzed when evaluated by an outside pathology laboratory. Eighty-one of the RPM devices were evaluated for weave separation and abrasion. Twenty-nine of the 81 RPM explants exhibited holes larger than 0.2 mm 2, which is the current manufacturing specification for this graft material. An average of 1.3 holes per RPM explant were observed. The holes in the RPM explants averaged 0.7mm 2 in size, with a median size of 0.5mm 2. The potential for transgraft endoleak as a result of these holes is unclear, because the process is influenced by many factors such as anticoagulation, tissue incorporation and healing, pressure gradients, and collateral flow. However, Medtronic AVE will continue to assess the potential for clinical impact as additional explants are received. Broken sutures have also been observed in explanted AneuRx stent grafts. Abrasion of the suture against the nitinol stent appears to be the primary mode of the breakage. The breakage is typically seen at the junction (stent tip to stent tip connection) stitches. Suture breaks do not appear to be clustered around a specific area of the stent graft. A sample of explanted devices was examined for suture breaks. Suture breaks have been observed in 12 of the 14 explants in this sample. An average of 24.6 breaks was observed, with a median of 22 breaks. The number of suture breaks ranged from 0 to 72 breaks. It should be noted that the AneuRx stent graft is designed to contain numerous and redundant suture attachment points. Bifurcated stent grafts contain between 1619 and 2079 sutures per device. The observed rate of suture breakage has been Table XI. Reasons for explant: Primary cause as determined by physician Reason for explant Number of occurrences: Implant duration 1 month Number of occurrences: Implant duration 1 month Number of occurrences: Total Increase in AAA Size 0 32 32 Incidental Autopsy 8 18 26 Rupture 4 13 17 Postimplant 0 13 13 At time of implant 4 0 4 Endoleaks (Type I, Type II, 0 14 14 Type III, including component separation) Implantation difficulties 7 0 7 Limb occlusion 1 5 6 Aortoenteric fistula 0 3 3 Symptomatic AAA 0 2 2 Infection 0 2 2 Migration 0 3 3 Disease progression* 0 3 3 *As defined by localized dilatation of the aortic neck very small. For instance, in this sampling of 14 explanted devices the rate of suture breaks as a percentage of total sutures is less than 1.5%. Microleaks, typically through suture attachment sites, have been observed during a small number of explant procedures. These microleaks are typically observed during the surgical conversion when anticoagulation is administered, the aneurysmal sac is then opened, and finally the thrombus adhering to the sac is disturbed. Potential clinical consequences. Potential clinical implications associated with stent strut fatigue fracture include, but are not limited to, incomplete seal, endoleak, device migration, and/or vessel perforation. Potential clinical consequences of fabric wear and/or broken sutures include microleaks, loss of column strength, endoleaks, and aneurysm enlargement. These potential clinical consequences can often be detected and treated during regular patient follow-up. The presence of device anomalies such as stent fractures, graft wear, or suture breaks does not necessarily result in adverse clinical outcomes. Some anomalies may be the result of the device adapting to tortuous or remodeling patient anatomy.

Table XII. Patients followed and follow-up periods 6 months 1 year 2 years 3 years 4 years 5 years Phase I Patients followed 36 31 25 18 8 1 Deaths* 4 7 9 12 12 14 Patients censored* 0 2 6 10 20 25 Lost/terminated 0 2 6 6 7 7 Not reaching follow-up 0 0 0 4 13 18 Phase II Patients followed 445 412 351 230 34 0 Deaths* 19 33 66 84 103 106 Patients censored* 6 25 53 156 333 364 Lost/terminated 4 13 27 43 57 62 Not reaching follow-up 2 12 26 113 276 302 Phase III Patients followed 630 556 332 43 0 0 Deaths* 38 56 108 125 130 130 Patients censored* 15 71 243 515 553 553 Lost/terminated 7 18 31 46 51 51 Not reaching follow-up 8 53 212 469 502 502 *Number of deaths and patients censored are cumulative Medtronic AVE is continuing to examine explanted devices and the environment of AAA disease to further the development of this important therapy option. If a physician has an explanted device available for analysis by Medtronic AVE, they should contact thier local Medtronic AVE representative for assistance with return of the device. Section IV: Summary The results of the US clinical trial, worldwide commercial experience, and explant analyses confirm that the AneuRx stent graft continues to be a safe and effective option for appropriately selected patients with infrarenal abdominal aortic aneurysms. PATIENTS FOLLOWED AND FOLLOW-UP PERIODS Table XII is intended to inform the reader of the number of patients in each phase of the IDE clinical study that have been followed at the various follow-up periods. The following definitions are included to enable interpretation. Patients followed: the number of patients reaching the end of the follow-up period with a study visit form. Deaths: the number of death forms received at the end of the follow-up period. Patients censored: the number of patients lost, terminated, or of unknown status. Patients lost/terminated: the number of patients with a study termination form. Patients not reaching follow-up: the number of patients in which a visit is not yet expected and the number of patients who should have a visit but don t have a visit or study termination form.