Redefining postoperative surveillance after endovascular aneurysm repair: Recommendations based on 5-year follow-up in the US Zenith multicenter trial
Article Outline
Introduction
Recommended postoperative surveillance after endovascular aneurysm repair (EVAR) includes serial contrast-enhanced CT scans. The cumulative deleterious effect on renal function, radiation exposure, and significant cost of this surveillance regimen are all problematic. However, there are scant data to support modulation of current post-EVAR surveillance regimens.
Methods
The study comprised patients who underwent EVAR as part of the prospective multicenter pivotal (phase II) and continued-access (phase III) US Zenith Endovascular (Cook, Bloomington, Ind) graft trials. A core lab prospectively recorded patient data. A composite aneurysm-related morbidity (ARM) variable was calculated to include aneurysm rupture, open conversion, any secondary intervention, limb thrombosis, migration, renal morbidity, or aneurysm-related death. The long-term freedom from ARM as a function of the presence or cumulative absence of any endoleak at 1, 6, and 12 months was analyzed. The potential additive predictive utility of aneurysm sac shrinkage (≥5 mm) was assessed at 12 months. The instructions for use for aortic neck anatomy (≥15 mm length, 18 to 28 mm diameter, ≤60° angulation) were followed.
Results
EVAR was done in 739 patients (mean follow-up, 29.9 ± 17.1 months). Freedom from endoleak at 1 month was highly predictive (P < .001) of reduced ARM: freedom from ARM was 92.3%, 89.8%, 85.2%, 83.1% and 83.1 % at 1, 2, 3, 4, and 5 years, respectively, in patients without endoleak (83.1%) and 75%, 67.1%, 61.5%, 55.9%, and 55.9% in patients with endoleak (16.9%). Cumulative absence of endoleak at 1 year (77.6%) was associated with 94%, 91.5%, 88.1%, 85.8%, and 85.8% 1- to 5-year freedom from ARM vs 73.3%, 66.7%, 56.6%, 52.5%, and 52.5% in patients with endoleak ≤1 year (22.4%), P < .001. In patients without endoleak at 12 months, the subsequent risk of any ARM was 8.2% (5-year risk, 14.2%; 1-year risk, 6.0%). In patients with significant sac shrinkage (≥5 mm) and cumulative absence of endoleak at 12 months, the subsequent risk of an ARM was 5.3% (5-year risk, 11.1%; 1-year risk, 5.8%).
Conclusions
Absence of endoleak at 30 and 365 days predicted greatly improved long-term freedom from ARM compared with early endoleak. A new EVAR surveillance regimen is recommended that modulates the intensity and frequency of postoperative imaging based on these early outcomes. In patients without early endoleak, the 6-month surveillance is eliminated, and aortic ultrasound is suggested for long-term surveillance >1 year. In most patients, this reduced surveillance regimen would be appropriate and could improve patient safety by reducing the cumulative deleterious effects of intravenous contrast and radiation exposure while also reducing health care costs. These subjective recommendations would be ideally validated in a randomized, prospective trial.
The predictive power of early outcome metrics on long-term results after endovascular aneurysm repair (EVAR) of abdominal aortic aneurysm (AAA) is unclear. The results of prior studies from single-device multicenter trials and large European multiple-device experiences have been conflicting.1, 2 The reasons are likely multifactorial, including differing methods and intensity of surveillance, use of core laboratory vs investigator reported data, variations in length of follow-up, and finally, the use of various endografts. Does the presence of an early endoleak after EVAR have significant prognostic value in long-term outcome? This clinical question is a potentially important one, because it could have implications regarding the level of aggressiveness required for long-term surveillance for these patients.
Surveillance regimens for EVAR that are the current standard-of-care were derived empirically from early multicenter trials and codified in the instructions for use (IFU) for the devices. Because long-term data were not available at that time, these surveillance regimens were not data-driven. Current recommended EVAR surveillance regimens include serial contrasted computed tomography (CT) scans and four-view plain abdominal radiographs (KUBs) at 1, 6, and 12 months and yearly thereafter. The cumulative contrast load from these CT scans is worrisome for its deleterious effect on renal function.3, 4 The potential carcinogenic effects of the cumulative radiation dose for patients is more difficult to quantify but still troubling.5, 6 Finally, the cost associated with current EVAR surveillance regimens is significant, comprising 30% to 35% of the total costs of EVAR follow-up during a 5-year period.7
The purpose of this study was to examine the correlation of early endoleaks with the long-term outcome in patients treated in the Zenith (Cook Inc, Bloomington, Ind) United States (US) multicenter trial, assessing possible correlation with early endoleak. On the basis of these data, alteration of current surveillance regimens may be appropriate.
Methods
This study is 5-year retrospective analysis of prospectively collected multicenter data from the US Zenith endovascular AAA pivotal (phase II) and continued-access (phase III) trials. The patients were enrolled between January 26, 2000, and June 18, 2003, and the follow-up was current to January 10, 2007. All data presented were prospectively collected by a core lab facility, the details of which have been previously described,8, 9 as has been a comprehensive list of inclusion/exclusion criteria.9 Aortic neck requirements included a minimum aortic neck length of ≥15mm, a diameter of ≤28 mm and ≥18 mm, with a ≤10% change in diameter in the first 15 mm of the neck. Angulation of the aortic neck to the aneurysm sac or the suprarenal aorta was limited to ≤60° and ≤45°, respectively.
The surveillance regimen included four-view KUBs and intravenous (IV) contrast- and noncontrast-enhanced CT scanning at 1, 6, and 12 months, and yearly thereafter. The study protocol mandated a CT slice thickness of ≤3 mm. All CT scan measurements and endoleak assessments were performed by the blinded core lab facility. Endoleak assessment was possible only if both noncontrast and contrast films were available.
Definitions
An aneurysm-related morbidity (ARM) variable was calculated to use as a composite end point to characterize patients who had any adverse outcome after EVAR. The ARM included aneurysm-related death, AAA rupture, open conversion, any secondary intervention, limb thrombosis, migration, renal morbidity, bowel fistula, or pseudoaneurysm. Significant migration was defined according to the Society for Vascular Surgery reporting standards10 as ≥10 mm caudal movement or any movement requiring treatment. Renal morbidity was defined as renal failure requiring dialysis. Early endoleak was defined as one seen at the 30-day CT scan.
Statistics
Data were managed by MED Institute (West Lafayette, Ind) using an Interactive Data Entry System (IDES) database (Corel Paradox tables, Corel Corp, Ottawa, Ontario, Canada). Data were analyzed using SAS 8 software (SAS Institute Inc, Cary, NC). Kaplan-Meier life-table analysis was used to estimate ARM rates at 1, 2, 3, 4, and 5 years of follow-up in patients with and without any endoleak. The log-rank test was used to test for a difference in the freedom from morbidity functions among the patient cohorts.
Results
A total of 739 patients underwent EVAR in the US Zenith endovascular trials, 352 in the pivotal (phase II) trial and 387 in continued-access (phase III) trial. The mean follow-up for the entire cohort was 29.9 ± 17.1 months. A small number of patients who did not have the requisite imaging were censored from the study, and a total of 714 patients were analyzed in this report.
Table I provides the 1-, 3-, and 5-year calculated risk of adverse outcomes in the entire patient cohort, those with endoleak at 30 days, and those with cumulative absence of endoleak at 30 days and 1 year. The presence of endoleak at 30 days (16.9%, 121 of 714) was associated with a significantly increased need for secondary procedures at 5 years compared with no endoleak at 30 days (42.4% vs 14.6%, P < .001). Nonsignificant differences were noted in the 5-year risk of aneurysm-related death (5.5% vs 1.2%, P = .08) and open conversion (5.6% vs 0.6%, P = .16) in patients with and without an endoleak at 30 days, respectively.
Table I. Rate of morbidities from Kaplan-Meier analysis
| Morbidity | 1 year, % (No.)a | 3 year, % (No.) | 5 year, % (No.) |
|---|---|---|---|
| Aneurysm-related death | |||
 Entire cohort | 1.9(736) | 2.8(497) | 2.8(155) |
| Endoleak at 30 days | 1.7(121) | 5.5(83) | 5.5(21)b |
| No endoleak at 30 days | 0.9(593) | 1.2(406) | 1.2(133) |
| No endoleak at 1 year | 0.9(554) | 1.3(381) | 1.3(126) |
| Open conversion | |||
| Entire cohort | 0.7(736) | 1.0(497) | 1.7(155) |
| Endoleak at 30 days | 0.9(121) | 0.9(83) | 5.6(21)c |
| No endoleak at 30 days | 0.2(593) | 0.6(406) | 0.6(133) |
| No endoleak at 1 year | 0.2(554) | 0.6(381) | 0.6(126) |
| Migration | |||
| Entire Cohort | 0(625) | 0.4(478) | 0.4(152) |
| Endoleak at 30 days | 0(109) | 0(83) | 0(21) |
| No endoleak at 30 days | 0(513) | 0.5(393) | 0.5(130) |
| No endoleak at 1 year | 0(482) | 0.6(368) | 0.6(123) |
| Rupture | |||
| Entire cohort | 0.1(736) | 0.1(497) | 0.1(155) |
| Endoleak at 30 days | 0(121) | 0(83) | 0(21) |
| No endoleak at 30 days | 0.2(593) | 0.2(406) | 0.2(133) |
| No endoleak at 1 year | 0.2(554) | 0.2(381) | 0.2(126) |
| Secondary interventions | |||
| Entire cohort | 9.4(734) | 16.4(439) | 19.0(124) |
| Endoleak at 30 days | 23.7(121) | 36.7(57) | 42.4(10)d |
| No endoleak at 30 days | 6.5(591) | 12.4(376) | 14.6(114) |
| No endoleak at 1 year | 4.7(552) | 9.3(357) | 11.7(111)e |
| Limb thrombosis | |||
| Entire cohort | 1.9(734) | 2.6(486) | 2.6(152) |
| Endoleak at 30 days | 0(121) | 0(83) | 0(21) |
| No endoleak at 30 days | 2.4(591) | 3.1(395) | 3.1(130) |
| No endoleak at 1 year | 2.5(552) | 3.4(370) | 3.4(123) |
| Renal morbidity | |||
| Entire cohort | 1.1(736) | 1.5(494) | 1.5(155) |
| Endoleak at 30 days | 0.9(121) | 0.9(83) | 0.9(21) |
| No endoleak at 30 days | 0.8(593) | 1.3(404) | 1.3(133) |
| No endoleak at 1 year | 0.9(554) | 1.4(379) | 1.4(126) |
aData in parenthesis are patients at risk; |
bP = .08; |
cP = .16; |
dP < .001; |
eP < .001 vs cohort without endoleak at 30 days. |
Patients with absence of early (30-day) endoleak (83.1%, 593 of 714), had a significant reduction (P < .001) in ARM events at 1, 2, 3, 4, and 5 years compared with those patients with any endoleak at 30 days (16.9%, 121 of 714; Fig 1). These differences were seen by 1 year, as those patients without early endoleak had a 1-year freedom from ARM of 92.3 % compared with 75.0% in those patients with early endoleak. These significant differences in occurrence rates of ARM between the two cohorts were magnified with long-term follow-up. At 5 years, the freedom from ARM in those with an early endoleak fell to 55.9%, significantly less than the 83.1% freedom from ARM in patients without early endoleak.
Fig 1.
Kaplan-Meier plots of freedom from aneurysm-related morbidity in patients with (solid line) or without endoleak (dashed line) during the first 30 days.
Patients with cumulative absence of endoleak at 6 months (79.4%, 567 of 714) and 12 months (77.6%, 554 of 714) had an identical 5-year freedom from ARM of 85.8%, a difference of only 2.7% improvement from the 30-day group (Fig 2). Thus, surveillance for continued absence of endoleak at 6 and 12 months did not greatly alter the predicted 5-year freedom from ARM. In the 343 patients (59.7%) with cumulative absence of endoleak at 1 year and significant (≥5 mm) aneurysm sac shrinkage, the 5-year freedom from ARM rose to 88.9% (Fig 3).
Fig 2.
Kaplan-Meier plots of freedom from aneurysm-related morbidity in patients with no endoleaks up to 30 (dotted line), 180 (dotted-dashed line), and 365 (dashed line) days.
Fig 3.
Kaplan-Meier plots of freedom from aneurysm-related morbidity patients with (solid line) endoleaks and no sac shrinkage or those without endoleak (dashed line) and with sac shrinkage in the first year.
Fig 4 demonstrates the freedom from ARM in those without (74.4%, 531 of 714), or with endoleak (25.6%, 183 of 714) at any time interval during the study. The 5-year freedom from ARM in patients without endoleak at any time in the study was 89.5% vs 49.8% in those with endoleak (P < .001). In patients without endoleak, 75% of ARM events (52 of 69) occurred within the first 12 months (Table II).
Fig 4.
Kaplan-Meier plots of freedom from aneurysm-related morbidity in patient groups with absence (solid line) or presence (dashed line) of endoleak through the study.
Table II. Aneurysm-related morbidity by time for patients without endoleak throughout the study
| Type of morbidity | 0-30 days, % (No.)a | 31-180 days, % (No.) | 181-365 days, % (No.) | >365 days, % (No.) |
|---|---|---|---|---|
| Patients at risk | 531 | 531 | 531 | 531 |
| Patients with ≥1 event | 2.82(15) | 2.45(13) | 1.13(6) | 2.64(14) |
| Related death | 0.19(1) | 0.75(4) | 0.0(0) | 0.38(2) |
| Rupture | 0.0(0) | 0.0(0) | 0.19(1) | 0.0(0) |
| Conversion | 0.0(0) | 0.0(0) | 0.19(1) | 0.38(2) |
| Secondary intervention | ||||
| Other | 1.13(6) | 0.56(3) | 0.56(3) | 0.56(3) |
| Leak | 0.0(0) | 0.19(1) | 0.0(0) | 0.56(3) |
| Occlusion | 1.51(8) | 0.56(3) | 0.38(2) | 0.19(1) |
| Limb thrombosis | 1.69(9) | 0.56(3) | 0.38(2) | 0.75(4) |
| Renal | 0.19(1) | 0.75(4) | 0.0(0) | 0.38(2) |
| Pseudoaneurysm | 0.0(0) | 0.0(0) | 0.0(0) | 0.0(0) |
| Bowel fistula | 0.0(0) | 0.0(0) | 0.0(0) | 0.0(0) |
aData in parenthesis are the number of events. |
In patients with no endoleak at 12 months, the subsequent risk of any ARM was 8.2% (5-year risk, 14.2%; 1-year risk, 6.0%; Fig 2). In patients with significant sac shrinkage (≥5 mm) and cumulative absence of endoleak at 12 months, the subsequent risk of an ARM was just 5.3% (5-year risk, 11.1%; 1-year risk, 5.8%; Fig 3).
Discussion
There are several important observations from these data:
Absence of early endoleak
At 30 days, 593 of the 714 patients (83.1%) treated with Zenith endografts had no endoleak, and this cumulative absence of endoleak continued to 1 year in 554 (77.6%). Thus, most patients treated with the Zenith endograft would be candidates for a reduced surveillance regimen. The 6-month surveillance studies in these patients did not correlate with any difference in 5-year ARM compared with the 12-month data, so it seems reasonable to delete this imaging completely if there is no early (30-day) endoleak. These findings are consistent with those of a recently presented single-center experience.11 Those patients with a cumulative absence of endoleak at 1 year had a projected 5-year risk of ARM of 14.2%.
Because there was a 6.0% risk of ARM in the first 12 months in this cohort, the subsequent predicted risk of ARM to 5 years was 8.2%. In those patients with cumulative absence of endoleak at 12 months and significant AAA shrinkage, the subsequent risk of ARM at 5 years was 5.3%. In these groups, the low risk of subsequent events suggested that a reduced surveillance regimen is reasonable. The finding that 10.5% of patients are projected to experience some form of ARM at 5 years despite the cumulative absence of endoleak is perhaps surprising. It is possible that an endoleak could have been missed, but the objective, blinded interpretation of the images by a core lab facility, as in this study, would minimize this occurrence.
Although the primary emphasis of surveillance has been to identify endoleak and potential aneurysm sac enlargement, it is clear from these data that the absence of these markers does not confer complete protection from ARM. Some ARM is not predictable by or correlated with endoleak, including limb thrombosis requiring a secondary procedure, renal morbidity, and endograft infection requiring open conversion. These were all exceptionally uncommon events in this study (Table II), but together they contributed to a large percentage of ARM. Although surveillance for endoleak is important and has been the primary focus of follow-up, it does not appear to be the sine qua non. The etiology of ARM is multifactorial and is likely not totally preventable or predictable with any imaging regimen.
Presence of early endoleak
Endoleak at 30 days was uncommon in this large patient cohort, occurring in 16.9% (121 of 714) of the treated subjects. An endoleak at this early time interval correlated with a 44.1% risk of ARM over 5 years. Most of these endoleaks were type II.9 These data confirm prior reports that early type II leaks are not necessarily benign. These conclusions are consistent with reports from the European Collaborators on Stent/Graft Techniques for Aortic Aneurysm Repair (EUROSTAR) database (n = 3595).12 In a study where a variety of endografts were used, patients with type II endoleaks were found to have an increased composite adverse event rate (AAA growth, transfemoral or transabdominal reintervention) of 55% vs 15% after 3 years.12 The long-term results from the current study underscore that these patients should have intensified surveillance.
New surveillance regimen
On the basis of these data and 5-year outcomes, we propose a modified surveillance regimen for patients undergoing EVAR with the Zenith endograft (Fig 5). All patients should have a contrasted CT scan and four-view KUBs at 30 days postprocedure. If there is any endoleak, or less than one stent component or iliac overlap, an intensified surveillance should be undertaken, consisting of serial IV contrast-enhanced CT scans at a minimum time interval of every 6 to 12 months. Four-view KUBs should also be considered. Subsequent treatment should be performed as indicated.
Fig 5.
New surveillance protocol for Zenith endografts. CT, computed tomography; IV, intravenous; KUB, plain abdominal radiographs; U/S, ultrasound.
In patients with no early (30-day) endoleak and satisfactory component/iliac overlap, a reduced surveillance regimen is appropriate. No 6-month imaging is recommended. At the 12-month follow-up, a contrast-enhanced CT scan is recommended. If there is no endoleak and a stable or shrinking aneurysm sac, a yearly aortic ultrasound examination is then suggested. If the patient's body habitus or other technical issues preclude an adequate examination by ultrasound, then a noncontrast CT scan should be substituted. An increasing aneurysm sac or new endoleak should prompt more intensive imaging and treatment as appropriate.
This suggested new surveillance regimen is predicated on the low risk of ARM in patients without early endoleak and the ability of aortic ultrasound imaging to detect evidence of significant aneurysm sac repressurization, as manifest by an increasing sac size or endoleak, or both. The low risk of ARM is strongly supported by the results from this large patient cohort, the prospective multicenter data collection, and blinded interpretation of the imaging. However, the recommendation of this new surveillance regimen must be considered subjective, because no data are available demonstrating long-term outcomes in patients followed up with such an altered surveillance regimen. Other investigators have also recently suggested changes in surveillance after EVAR, which include noncontrast CT scans or ultrasound imaging for long-term surveillance.13, 14 Ideally, these recommendations should be validated with a prospective randomized trial comparing the different surveillance regimens.
Eliminating the 6-month surveillance entirely and changing the yearly surveillance to an aortic ultrasound scan in those qualifying patients as above could have several salutary effects. Most patients would not require the cumulative IV contrast load and radiation exposure inherent in serial CT scanning. Prior studies have demonstrated a decline in renal function in EVAR patients. Interestingly, there have been no differences in devices with or without suprarenal fixation.15 The cumulative contrast load from the CT scans has been implicated for the deterioration in renal function.3, 4 The potential carcinogenic effects of the cumulative radiation exposure from CT scanning is troublesome and becoming better understood.5, 6
Finally, this new surveillance paradigm has the potential for significant cost savings. Costs of imaging constitute 30% to 35% of the total postimplant costs during the 5-year period after EVAR.7 Aortic ultrasound imaging is less expensive and, most importantly, has zero potential morbidity.
The four-view KUBs were a part of the original surveillance regimen to identify failure in the metallic support system (stent fracture, barb separation, suprarenal stent detachment) and assess adequacy of component overlap. Stent fracture and barb separation were very uncommon in this trial16 and, most importantly, had no discernible clinical impact. Barb separation could theoretically weaken the active proximal fixation and thus increase the risk of migration. However, only two episodes of migration were observed in this large patient cohort, giving a 5-year freedom from migration of 99.6%.16 Two episodes of detachment of the suprarenal stent from the main body were noted early in the Zenith experience, but this has not occurred since the suture connections were reinforced in 2002.
Confirmation of adequate component and iliac overlap is important initially. As such, a set of four-view KUBs was retained in the initial 30-day imaging. Indeed, the single AAA rupture after EVAR in this entire cohort occurred in a patient with inadequate overlap of a limb into the iliac artery.9 The risk of subsequent late component separation has been very rare, occurring in only three patients.16 Although an initial four-view KUB is still considered prudent, the rarity of clinically significant problems on serial studies suggests that this imaging could be omitted for most patients. In patients with any combination of a very large AAA or significant iliac or aortic tortuosity, greater stress on the components may occur as the sac shrinks, and continued use of KUB surveillance may be prudent.
The risk of subsequent ARM in the reduced surveillance group is low, but it would be inappropriate to perform no long-term surveillance at all. Although some of these ARM events are not related to endoleak, some clearly are and should be readily identified by an enlarging aneurysm sac or endoleak, or both, seen on ultrasound. In the Zenith study, there was a 12% to 15 % risk of late endoleak occurring when none was detected at 30 days.16 Most of these endoleaks were type II in nature. The ultimate clinical impact of these late endoleaks is uncertain, but it is prudent to continue some sort of longitudinal imaging in all patients undergoing EVAR.
Limitations
The EVAR surveillance recommendations are based on data from a single device. As such, they must be considered device-specific and not be used with other devices. Serial CT scanning is the only reliable method of detecting endograft migration and should continue to be an integral part of surveillance regimens for devices with a significant risk of this failure mode.17, 18, 19
These long-term data are from patients who had stringent aortic neck inclusion criteria. Patients treated with Zenith endografts who are outside the device IFU, particularly those with disadvantaged aortic necks, may not have the same long-term results. As such, use of these new surveillance regimens cannot be recommended in these patients.
Five-year follow-up was not available for the entire cohort because the original study design only included 2 years of follow-up. Exhaustive efforts were made to obtain consent from patients in the pivotal arm to extend their follow-up to 5 years. The continued-access patients did not have follow-up after 2 years in the trial. To minimize any bias in the long-term results, Kaplan-Meier analysis was used to properly estimate occurrence rates, taking into account the censored patients.
Conclusions
The large majority of patients undergoing EVAR with the Zenith endograft have absence of endoleak at 30 and 365 days. In these patients, a reduced surveillance regimen is recommended that may improve patient safety by reducing the cumulative deleterious effects of IV contrast and radiation exposure while also reducing health care costs. The small number of patients who have early endoleak are at higher risk for subsequent events and need more aggressive surveillance and treatment as indicated. Finally, although surveillance for endoleak has been a primary focus of follow-up, it does not appear to be the sine qua non. The etiologies of ARM are multifactorial and are likely not totally preventable or predictable with any imaging regimen.
Author contributions
We would like to thank Johnna D. Anderson, MS, for performing the statistical analysis for this study.
References
- . Endoleak as a predictor of outcome after endovascular aneurysm repair: AneuRx multicenter clinical trial. J Vasc Surg. 2000;32:90–107
- . Significance of endoleaks after endovascular repair of abdominal aortic aneurysms: the EUROSTAR experience. J Vasc Surg. 2002;35:461–473
- The impact of aortic endografts on renal function. J Vasc Surg. 2005;41:926–930
- Renal function after endovascular aortic aneurysm repair: a single-center experience with transrenal versus infrarenal fixation. J Endovasc Ther. 2007;14:130–137
- Cancer risks attributable to low doses of ionizing radiation: assessing what we really know. Proc Natl Acad Sci U S A. 2003;100:13761–13766
- . Computed tomography–an increasing source of radiation exposure. N Engl J Med. 2007;357:2277–2284
- . Long-term follow-up cost after endovascular aneurysm repair. J Vasc Surg. 2007;46:9–15
- . Influence of endograft oversizing on device migration, endoleak, aneurysm shrinkage and aortic neck dilation: results from the Zenith multi-center trial. J Vasc Surg. 2004;39:20–25
- . Zenith AAA endovascular graft: intermediate-term results of the US multi-center trial. J Vasc Surg. 2004;39:1209–1218
- Reporting standards for endovascular aortic aneurysm repair. J Vasc Surg. 2002;35:1048–1060
- . What is the clinical utility of a six-month CT in the follow-up of endovascular aneurysm repair patients? (Abstract). Baltimore, MD: Society for Vascular Surgery Meeting; 2007;
- . Is a type II endoleak after EVAR a harbinger of risk? (Causes and outcome of open conversion and aneurysm rupture during follow-up). Eur J Vasc Endovasc Surg. 2004;27:128–137
- . Long-term outcome and reintervention after endovascular abdominal aortic aneurysm repair using the Zenith stent graft. J Vasc Surg. 2007;45:461–466
- Intermediate follow-up after endovascular aneurysm repair: can we fogo CT scanning in certain patients?. Ann Vasc Surg. 2007;21:663–670
- . Endovascular aneurysm repair with suprarenal vs infrarenal fixation: a study of renal effects. J Vasc Surg. 2006;43:19–25
- Greenberg RK, Chuter TAM, Cambria RP, Sternbergh WC III, Nolte L. The Zenith endovascular graft: 5 year US multicenter trial results and analysis. J Vasc Surg [in press].
- . Mid and long term device migration after endovascular abdominal aortic aneurysm repair: a comparison of AneuRx and Zenith endografts. J Vasc Surg. 2005;42:392–401
- . Stent graft migration after endovascular aneurysm repair: importance of proximal fixation. J Vasc Surg. 2003;38:1264–1272
- Predictive factors and clinical consequences of proximal aortic neck dilatation in 230 patients undergoing abdominal aorta aneurysm repair with self-expandable stent-grafts. J Vasc Surg. 2003;37:1200–1205
Competition of interest: Dr Sternbergh, Dr Greenberg, and Dr Chuter are present or past consultants for Cook Inc, and have received research support for the Zenith multicenter trial. Dr Chuter receives royalties for patents licensed to Cook Inc.
PII: S0741-5214(08)00423-0
doi:10.1016/j.jvs.2008.02.075
© 2008 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
