Influence of Oversizing on Outcome in Thoracic Endovascular Aortic Repair

Transcription

1 738 J ENDOVASC THER 2013;20: CLINICAL INVESTIGATION Influence of Oversizing on Outcome in Thoracic Endovascular Aortic Repair Jip L. Tolenaar, MD 1,2 ; Frederik H.W. Jonker, MD, PhD 3 ; Frans L. Moll, MD, PhD 2 ; Joost van Herwaarden, MD, PhD 2 ; Mark D. Morasch, MD 4 ; Michel S. Makaroun, MD 5 ; and Santi Trimarchi, MD, PhD 1 on behalf of the GORE TAG Device Investigators 1 Thoracic Aorta Research Center, Policlinico San Donato, IRCCS, Milan, Italy. 2 University Medical Center Utrecht, The Netherlands. 3 Department of Surgery, Maasstad Hospital, Rotterdam, The Netherlands. 4 Department of Vascular Surgery, Division of Vascular Surgery, Northwestern University, Chicago, Illinois, USA. 5 Division of Vascular Surgery, UPMC Hospitals, Pittsburgh, Pennsylvania, USA. Purpose: To investigate the influence of stent-graft oversizing on device-related complications after thoracic endovascular aortic repair (TEVAR) for thoracic aortic aneurysm (TAA). Methods: The study cohort was composed of patients enrolled in 4 clinical trials of the TAG thoracic stent-graft. A total of 337 TAA patients (222 men; mean age 72 years) treated in these trials had sufficient data for analysis of oversizing and post-procedure mortality and complications, such as endoleak, migration, rupture, and reinterventions. Mean oversizing at the proximal landing zone was 14.6% (range 3.4% to 39.7%). Patients were stratified based on the percentage of oversizing:,10% (n¼85, group 1), 10% 20% (n¼188, group 2), and.20% (n¼64, group 3). Results: Patients in group 1 had significantly larger preoperative proximal aortic diameters (32.6 vs vs mm, respectively; p,0.001) and neck lengths (6.9 vs. 5.8 vs. 5.2 cm (p¼0.035). Overall, type I endoleak was the most frequent complication during the first 30 days of follow-up (35, 10.4%), but the incidences did not differ among the 3 groups (10.6% vs. 11.2% vs. 7.8%, respectively; p¼0.809). Over a mean follow-up of months, there were no significant differences in the occurrence of device-related complications among the groups, though the incidence of type I endoleaks was lower in group 2 (9.4% vs. 3.2% vs. 7.8%, respectively; p¼0.073). Cox proportional hazards modeling showed no difference in the time to type I endoleak among oversizing groups [group 1 vs. 2: HR 1.24, 95% CI 0.65 to 2.36 (p¼0.509) and group 3 vs. 2: HR 1.24, 95% CI 0.60 to 2.60 (p¼0.562)]. Conclusion: The percentage of oversizing did not significantly affect the incidence of devicerelated complications after TEVAR for TAA. Although oversizing may enhance the radial force and help maintain a good proximal seal, additional oversizing seemed not to improve the overall outcome in this analysis. The current guidelines regarding stent-graft oversizing for TAA seem appropriate, though the correct percentage remains to be determined J Endovasc Ther. 2013;20: Key words: thoracic aortic aneurysm, thoracic endovascular aortic aneurysm repair, stentgraft, oversizing, device-related complications, endoleak, migration, branch vessel occlusion, rupture, mortality, reintervention This study was supported by W.L. Gore & Associates, Flagstaff, Arizona. Frans Moll is a consultant to Best Doctors and Medtronic Inc. Mark Morasch is a consultant to W.L. Gore & Associates and King Pharmaceuticals. Michel Makaroun is a consultant to W.L. Gore & Associates, Medtronic Inc., and Cordis, a Johnson & Johnson company. The other authors declare no association with any individual, company, or organization having a vested interest in the subject matter/products mentioned in this article. Corresponding author: Santi Trimarchi, MD, Policlinico San Donato IRCCS, Department of Cardiovascular Surgery, Thoracic Aorta Research Center, University of Milano, Via Morandi 30, San Donato Milanese, Italy. santi. trimarchi@unimi.it, santi.trimarchi@grupposandonato.it Q 2013 by the INTERNATIONAL SOCIETY OF ENDOVASCULAR SPECIALISTS Available at

2 J ENDOVASC THER STENT-GRAFT OVERSIZING IN TEVAR ;20: Tolenaar et al. Thoracic endovascular aortic repair (TEVAR) of descending thoracic aortic aneurysm (TAA) has been widely adopted in patients with suitable anatomy as a superior treatment compared to open surgical repair. 1 4 However, these procedures are associated with complications related to aortic wall fixation of the stent-graft, such as neck dilation, device migration, and endoleaks, which may require secondary interventions. 5 Therefore, stentgraft fixation, especially at the site of the proximal landing zone, is considered the Achilles heel of endovascular repairs in the aorta, and the exact mechanisms regarding this fixation are still poorly understood. Multiple factors have been identified that may influence the fixation and seal, including characteristics of the aorta, such as neck angulation, neck length, and presence of thrombus or calcification. 6,7 In addition, characteristics of the device, including oversizing, radial force, presence or absence of active fixation (e.g., barbs), and features of the surgical procedure (e.g., correct placement and balloon molding), may influence fixation. 8 Based on experiences of endovascular aneurysm repair (EVAR) in the abdominal aorta, stent-graft oversizing is essential, as it enhances the radial force of the device against the aortic wall, improving fixation and sealing. 9,10 On the other hand, excessive oversizing in EVAR is associated with an increased risk of complications, as it may lead to infolding of the graft or dilatation of the aortic neck. 11 A systematic review of the influence of oversizing on the outcome and complications in EVAR demonstrated that 10% to 20% oversizing, which corresponds with the Instructions of Use (IFU) of most manufacturers, offers the best results. 10 Interestingly, complications following TEVAR are greater than after EVAR, which could be the result of a difference in dynamics and forces working on these grafts Hemodynamic studies showed that the dilatation of the aorta during the cardiac cycle is larger in the thoracic compared to the abdominal aorta. 14,15 Therefore, the optimal measure of oversizing in TEVAR might be different from the percentage used for abdominal aortic stent-grafts. The purpose of this study was to investigate the influence of stent-graft oversizing on Figure 1 Flow chart describing acquisition of patients for an analysis of the impact of stent-graft oversizing on device-related complications in TEVAR. device-related complications in TEVAR, thus elucidating the optimal sizing strategy for TEVAR patients. Study Design METHODS The study cohort was composed of patients enrolled in 4 clinical trials of the TAG thoracic stent-graft (W.L. Gore & Associates, Flagstaff, AZ, USA). 1 4 All participating sites had Institutional Review Board approval before enrollment of study subjects; patient inclusion and device sizing criteria have been previously described. 4 Preoperative computed tomography (CT) was used for measurement of the aortic diameter in the most proximal 2 cm of the aortic neck. Aortic measurements and sizing were performed by physicians at the different study sites, which were regularly verified by a core laboratory. The mean of the 3 diameter measurements was subtracted from the nominal device diameter and then divided by the mean diameter to calculate the degree of oversizing. Postoperative imaging included chest radiographs and CT scans at 1, 6, and 12 months and annually thereafter. Site-reported data were used for the current analysis rather than core laboratory data, which was inconsistently available across the studies. Only patients with a saccular or fusiform atherosclerotic descending TAA were eligible; patients with any other aortic diseases or ruptured aneurysms were excluded (Fig. 1). After review of these data from the

3 740 STENT-GRAFT OVERSIZING IN TEVAR J ENDOVASC THER Tolenaar et al. 2013;20: TABLE 1 Demographics and Pre-Treatment Medical History Oversizing Groups,10% (n¼85) 10% 20% (n¼188).20% (n¼64) p Age, y Male gender 59 (69.4%) 126 (67.0%) 37 (57.8%) Mean height, cm Mean weight, kg Body surface area, units History of Cigarette smoking 73 (85.9%) 155 (82.5%) 51 (79.7%) Coronary artery disease 49 (57.7%) 97 (51.6%) 22 (34.4%) Cardiac arrhythmia 24 (28.2%) 54 (28.7%) 9 (14.1%) PAD 15 (17.7%) 23 (12.2%) 14 (21.9%) Symptomatic aneurysm 18 (21.2%) 51 (27.1%) 13 (20.3%) Other aneurysms 27 (31.8%) 66 (35.1%) 26 (40.6%) Stroke 10 (11.8%) 21 (11.2%) 9 (14.1%) COPD 33 (38.8%) 85 (45.2%) 18 (28.1%) ASA I 5 (5.9%) 3 (1.6%) 2 (3.1%) II 9 (10.6%) 22 (11.7%) 8 (12.5%) III 48 (56.5%) 119 (63.3%) 39 (60.9%) IV 23 (27.1%) 44 (23.4%) 15 (23.4%) Continuous data are presented as the means 6 standard deviation; categorical data are given as the counts (percentage). PAD: peripheral artery disease, COPD: chronic obstructive pulmonary disease, ASA: American Society of Anesthesiologists. 4 clinical trials, a total of 337 patients (222 men; mean age 72 years) were selected for this analysis based on the ability to identify the most proximal device implanted (in the case of multiple devices). Mean oversizing was 14.6%, with a broad range ( 3.4% to 39.7%). Based on site-reported measurements, 23 of the 337 subjects had oversizing values,5%, of which 16 were between 0% (n¼2) and 5%; 5 had negative values. The patients were divided into 3 different cohorts based on the degree of oversizing:,10%, 10% 20%, and.20%. Group 1 consisted of 85 patients (59 men; mean age 72.6 years), group 2 of 188 patients (126 men; mean age 71.3 years), and group 3 of 64 patients (37 men; mean age 73.0 years). Baseline characteristics, history, and comorbidities of the groups are reported in Table 1. Statistical Analyses The outcomes assessed included mortality, endoleak, device migration, branch vessel occlusion, lumen obstruction, material failure, increased sac diameter, aneurysm rupture, other complications, and aneurysm-related reinterventions. Categorical demographic and treatment variables were analyzed using a Fisher exact or chi-square test, and continuous variables were analyzed using ANOVA or a Kruskal-Wallis test. Differences achieving p,0.05 were considered significant. Kaplan- Meier plots were constructed to estimate the time to first device-related complication as well as the time to first type I endoleak; differences between oversizing groups were assessed using the log-rank test. Time to first type I endoleak was analyzed using a Cox proportional hazards model based on a propensity score constructed using ordinal logistic regression to compensate for the small number of events; results of the model are reported as the hazard ratios (HR) with 95% confidence interval (CI). Variables were chosen for the propensity score if they were significantly associated with the outcome (p,0.2) and also significantly associated with

4 J ENDOVASC THER STENT-GRAFT OVERSIZING IN TEVAR ;20: Tolenaar et al. TABLE 2 Preoperative Computed Tomography Measurements Oversizing Groups,10% 10% 20%.20% p Maximum aneurysm diameter, mm (n¼59) (n¼154) (n¼51) Proximal LZ diameter, mm (n¼85) (n¼188) (n¼64), Distal LZ diameter, mm (n¼85) (n¼188) (n¼64) Proximal neck length, cm (n¼85) (n¼187) (n¼61) Distal neck length, cm (n¼85) (n¼187) (n¼64) Aneurysm length, cm (n¼85) (n¼187) (n¼64) Data are presented as the means 6 standard deviation (sample size). LZ: landing zone. the oversizing group variable (p,0.3); variables with excessive missing data or diameter variables already used in the oversizing calculation were excluded. Statistical analysis was performed using SAS software (version 9.2; SAS Institute, Cary, NC, USA). RESULTS Anatomical Features and Perioperative Management The mean diameter and length of the proximal landing zone (Table 2) were significantly larger in the,10% group 1 compared to groups 2 and 3: 32.6 vs vs mm (p,0.001) and 6.9 vs. 5.8 vs. 5.2 cm (p¼0.035), respectively. Patients in the 10% 20% group 2 had a slightly larger maximum aneurysm diameter (61.0 mm, p¼0.096). No differences were observed among the perioperative variables of the 3 groups (Table 3), although group 2 patients more frequently had a left subclavian bypass or transposition (4.7% vs. 9.6% vs. 1.6% for groups 1, 2, and 3, respectively; p¼0.063). Device-Related Complications and Follow-up Type I endoleak was the most frequent complication (n¼35, 10.4%) in the 30-day follow-up interval (Table 4), but the incidence of type I endoleak did not differ significantly TABLE 3 Outcomes After TEVAR for the Oversizing Groups Oversizing Groups,10% 10% 20%.20% p Proximal stent-graft diameter, mm (n¼85) (n¼188) (n¼64) Procedure duration, min (n¼85) (n¼180) (n¼62) Blood loss, ml (n¼82) (n¼184) (n¼62) LSA bypass/transposition 4 (4.7%) 18 (9.6%) 1 (1.6%) Embolization 5 (5.9%) 6 (3.2%) 1 (1.6%) Conduit use 10 (11.8%) 27 (14.4%) 8 (12.5%) Number of devices implanted (n¼85) (n¼188) (n¼64) Hospital stay, d (n¼84) (n¼188) (n¼64) Time in ICU, d (n¼26) (n¼75) (n¼19) Return to normal activities, d (n¼69) (n¼170) (n¼58) Follow-up, mo (n¼85) (n¼188) (n¼64) Continuous data are presented as the means 6 standard deviation (sample size); categorical data are given as the counts (percentage). LSA: left subclavian artery, ICU: intensive care unit.

5 742 STENT-GRAFT OVERSIZING IN TEVAR J ENDOVASC THER Tolenaar et al. 2013;20: TABLE 4 Device-Related Adverse Events During 30-Day and Midterm Follow-up Oversizing Groups at 30 Days Oversizing Groups at.30 Days,10% (n¼85) 10% 20% (n¼188),10% (n¼85) p,10% (n¼85) 10% 20% (n¼188),10% (n¼85) p Endoleak I 9 (10.6%) 21 (11.2%) 5 (7.8%) (9.4%) 6 (3.2%) 5 (7.8%) Ia 7 (8.2%) 16 (8.5%) 5 (7.8%) (4.7%) 4 (2.1%) 3 (4.7%) Ib 2 (2.4%) 5 (2.7%) 1 (1.6%) (4.7%) 2 (1.1%) 2 (3.1%) II 3 (3.5%) 11 (5.9%) 5 (7.8%) (5.9%) 12 (6.4%) 5 (7.8%) III 2 (2.4%) 4 (2.1%) 2 (3.1%) (0.5%) 2 (3.1%) IV Indeterminate 2 (2.4%) 8 (4.3%) 1 (1.6%) (2.4%) 5 (2.7%) 1 (1.6%) Migration 0 1 (0.5%) (1.2%) 2 (1.1%) Branch vessel occlusion 2 (2.4%) 2 (1.1%) 1 (1.6%) (1.2%) 3 (1.6%) 1 (1.6%) Lumen obstruction (1.6%) Prosthesis material failure (2.7%) 2 (3.1%) Aneurysm rupture (1.2%) 2 (1.1%) Extrusion/erosion Other device complication 2 (2.4%) 5 (2.7%) 1 (1.6%) (1.2%) 3 (1.6%) 1 (1.6%) Increased sac diameter (1.2%) 4 (2.1%) Mortality 4 (4.7%) (30.6%) 64 (34.0%) 18 (28.1%) among the 3 groups (10.6% vs. 11.2% vs. 7.8%; p¼0.809). The 30-day mortality was significantly higher in group 1 compared to the other groups (4.7% vs. 0 vs. 0; p¼0.005). In the,5% oversizing group, mortality was 13% (3/23), but none of these deaths were procedure or aneurysm-related (bowel ischemia, respiratory failure, and cardiac arrest). Over a mean follow-up of months, there were no significant differences in the occurrence of device-related complications among the groups (Fig. 2). In this period, no differences were observed in the rate of conversions (2.4% vs. 0.5% vs. 0.0%, respectively; p¼0.252) or the necessity for additional grafts (9.4% vs. 4.8% vs. 4.7%, respectively; p¼0.292). There was a nonsignificant decreased incidence of type I endoleaks in group 2(9.4% vs. 3.2% vs. 7.8% for groups 1, 2, and 3, respectively; p¼0.073). Risk Model for Type I Endoleak The propensity score was built using the following variables: number of devices implanted, pre-treatment aneurysm length, pretreatment potential proximal neck length, history of coronary artery disease, history of cancer, and history of cardiac arrhythmia. The hazard ratios for the time to type I endoleak based on the oversizing group and propensity score variables were 1.24 for group 1 vs. 2 (95% CI 0.65 to 2.36; p¼0.509) and 1.24 for group 3 vs. 2 (95% CI 0.60 to 2.60; p¼0.562), indicating no difference in timing of this event. DISCUSSION Studies concerning EVAR have shown that 10% to 20% stent-graft oversizing yields the best results. 10 Although these guidelines are also recommended in TEVAR, to our knowledge, the effect of oversizing in the descending thoracic aorta has never been studied. The thoracic aorta presents different hemodynamic and mechanical forces compared to the abdominal aorta, 13,14 which might influence outcomes for these patients. Furthermore, the angulation of the proximal thoracic aorta can lead to alterations in the distribution of the radial forces of the stent-graft, potentially compromising the proximal sealing. 6 These characteristics might increase the risk of device-related complications, as more reinter-

6 J ENDOVASC THER STENT-GRAFT OVERSIZING IN TEVAR ;20: Tolenaar et al. Figure 2 Kaplan-Meier estimates of (A) freedom from device-related events and (B) freedom from type I endoleak during follow-up. ventions are reported in TEVAR patients compared to EVAR. 12 In this study, we investigated the relationship between the measure of oversizing and complications after TEVAR for descending thoracic aortic aneurysms. The,10% oversizing group had a mean proximal neck diameter that was larger than the other groups. This can be explained by the fact that some patients had such large diameters at the proximal landing zone, which made additional oversizing impossible, as the range of stentgraft diameters was limited. To address this problem, larger stent-grafts have recently been introduced. In addition, because we used the mean of the 3 measurements in a landing zone, some stent-grafts in a tapered proximal landing zone may be considered relatively under- or oversized. Endoleaks at the proximal landing zone may be related to sizing of the stent-grafts in two ways: relative undersizing may result in insufficient apposition of the graft to the aortic wall, while excessive oversizing may lead to pleating of the graft material. Both result in a compromised seal. 16 We found a lower incidence of type I endoleaks in group 2 during follow-up (3.2% vs. 9.4% in group 1 and 7.8% in group 3), but this was not statistically significant. These findings correlate with results from studies in EVAR patients, which demonstrated a positive effect of oversizing between 10% and 25%. 9,10 The lower radial force, as well as the limited conformation of the Gore TAG device, might have led to an increased incidence of endoleak. These properties are also potential risk factors for proximal stent-graft collapse or infolding. No statistically significant differences were shown for the other types of endoleaks or other device-related complications. Most interestingly, the patients in the,5% oversizing group had a higher mortality, but none of these was related to the aneurysm, endoleak, or the procedure. Aortic neck dilatation can result from excessive oversizing due to an increase in radial forces, as has been described in EVAR patients. 17 Natural aortic neck dilatation can lead to complications, such as the loss of aortic wall fixation and stent-graft migration, which are more frequently present in patients with inadequate oversizing. In our study, no measurements of the neck were available during follow-up to objectify such findings, although the incidence of these complications did not differ significantly among the groups. Static CTA imaging has proven to be insufficient for adequate sizing of stent-grafts because dynamic imaging studies have shown that aortic pulsatility varies considerably among patients. 18 Increased aneurysm neck pulsatility, which could influence the measure of oversizing, cannot be analyzed using static imaging. As van Keulen et al. 15 showed, preoperative pulsatility of the proximal landing zone for abdominal stent-grafts is significantly associated with device migration after 3 years. In our study, such a

7 744 STENT-GRAFT OVERSIZING IN TEVAR J ENDOVASC THER Tolenaar et al. 2013;20: phenomenon might have influenced the outcome, but it would probably have influenced all groups equally. In addition, since most physicians currently use static imaging for sizing aneurysms and aortic necks, this study approximates daily clinical practice. Limitations This study was based on controlled prospective studies of the TAG thoracic stentgraft; therefore, these results are applicable only for this device and cannot be generalized for other stent-grafts or aortic diseases other than aneurysms. Patient selection for the initial trials was limited to strict inclusion/ exclusion criteria, so the outcomes were possibly more favorable, as patients with increased anatomical and physiological risk were excluded. In addition, because of the low number of events, the study might have been underpowered, although it represents one of the largest cohorts gathered to address this topic. Furthermore, because prospective clinical studies are typically performed in highvolume experienced centers with relatively low event rates, our results may not reflect community practice. Other important limitations are the potential unreliability of the measurement data from the different centers and the lack of availability of post-procedural CT data to analyze important anatomical features. Conclusion The percentage of oversizing did not significantly affect the incidence of device-related complications after TEVAR for TAA. Although oversizing may enhance the radial force to maintain a good proximal seal, additional oversizing seemed not to improve the overall outcome in this analysis. The current guidelines regarding stent-graft oversizing for TAA seem appropriate, though the correct percentage remains to be determined. REFERENCES 1. Bavaria JE, Appoo JJ, Makaroun MS, et al. Endovascular stent grafting versus open surgical repair of descending thoracic aortic aneurysms in low-risk patients: a multicenter comparative trial. J Thorac Cardiovasc Surg. 2007;133: Cambria RP, Crawford RS, Cho JS, et al.; GORE TAG Investigators. A multicenter clinical trial of endovascular stent graft repair of acute catastrophes of the descending thoracic aorta. J Vasc Surg. 2009;50: e Makaroun MS, Dillavou ED, Wheatley GH, et al. Five-year results of endovascular treatment with the Gore TAG device compared with open repair of thoracic aortic aneurysms. J Vasc Surg. 2008;47: Makaroun MS, Dillavou ED, Kee ST, et al. Endovascular treatment of thoracic aortic aneurysms: results of the phase II multicenter trial of the GORE TAG thoracic prosthesis. J Vasc Surg. 2005;41: Leurs LJ, Harris PL, Buth J, et al. Secondary interventions after elective endovascular repair of degenerative thoracic aortic aneurysms: results of the European Collaborators Registry (EUROSTAR). J Vasc Interv Radiol. 2007;18: Canaud L, Alric P, Laurent M, et al. Proximal fixation of thoracic stent-grafts as a function of oversizing and increasing aortic arch angulation in human cadaveric aortas. J Endovasc Ther. 2008;15: Wyss TR, Dick F, Brown LC, et al. The influence of thrombus, calcification, angulation, and tortuosity of attachment sites on the time to the first graft-related complication after endovascular aneurysm repair. J Vasc Surg. 2011; 54: Resch T, Malina M, Lindblad B, et al. The impact of stent design on proximal stent-graft fixation in the abdominal aorta: an experimental study. Eur J Vasc Endovasc Surg. 2000;20: Mohan IV, Laheij RJ, Harris PL. Risk factors for endoleak and the evidence for stent-graft oversizing in patients undergoing endovascular aneurysm repair. Eur J Vasc Endovasc Surg. 2001;21: van Prehn J, Schlösser FJ, Muhs BE, et al. Oversizing of aortic stent grafts for abdominal aneurysm repair: a systematic review of the benefits and risks. Eur J Vasc Endovasc Surg. 2009;38: Lin KK, Kratzberg JA, Raghavan ML. Role of aortic stent graft oversizing and barb characteristics on folding. J Vasc Surg. 2012;55: Leurs LJ, Bell R, Degrieck Y, et al. Endovascular treatment of thoracic aortic diseases: combined experience from the EUROSTAR and United

8 J ENDOVASC THER STENT-GRAFT OVERSIZING IN TEVAR ;20: Tolenaar et al. Kingdom Thoracic Endograft registries. J Vasc Surg. 2004;40: Figueroa CA, Taylor CA, Chiou AJ, et al. Magnitude and direction of pulsatile displacement forces acting on thoracic aortic endografts. J Endovasc Ther. 2009;16: van Prehn J, Bartels LW, Mestres G, et al. Dynamic aortic changes in patients with thoracic aortic aneurysms evaluated with electrocardiography-triggered computed tomographic angiography before and after thoracic endovascular aneurysm repair: preliminary results. Ann Vasc Surg. 2009;23: van Keulen JW, Moll FL, Barwegen GK, et al. Pulsatile distension of the proximal aneurysm neck is larger in patients with stent graft migration. Eur J Vasc Endovasc Surg. 2010; 40: Jonker FH, Schlösser FJ, Geirsson A, et al. Endograft collapse after thoracic endovascular aortic repair. J Endovasc Ther. 2010;17: Makaroun MS, Deaton DH. Is proximal aortic neck dilatation after endovascular aneurysm exclusion a cause for concern? J Vasc Surg. 2001;33:S van Keulen JW, van Prehn J, Prokop M, et al. Dynamics of the aorta before and after endovascular aneurysm repair: a systematic review. Eur J Vasc Endovasc Surg. 2009;38: