The EVAR trial 2 (EVAR-2) is the only randomized clinical trial evaluating the efficacy of EVAR in patients deemed unfit for open surgery due to significant comorobidities and resulting physiologic high-risk. The study was a multicenter trial conducted in the United Kingdom with 197 patients randomized to EVAR and 207 patients randomized to no-intervention with a primary endpoint of all-cause mortality. There were also secondary endpoints of aneurysm-related mortality, graft-related complications and total hospital cost [4, 5]. Enrolling physicians were given guidelines for patient enrollment (Table 8.3), but the final decision was left to the discretion of the treating physician [10]. This study had a 30-day perioperative mortality rate of 7.3 %. There was no difference in all cause mortality between the study groups with a total mortality of 21.0 deaths per 100 person years in the EVAR group and 22.1 deaths per 100 person years in the no-interventions group (P = 0.97). However there was a difference in the aneurysm related mortality of 3.6 deaths per 100 person years in the EVAR group and 7.3 deaths per 100 person years in the no-intervention group (P = 0.02). Over the 8 year study period 158 graft related complications occurred in 97 patients with 66 reinterventions in 55 patients. During the study period the mean cost of aneurysm related procedures was $22,687 for the EVAR group and $7,821 in the no-intervention group [4]. There was no difference in QOL between the two groups observed in the midterm analysis [5].

Though this is a randomized clinical trial, a number of aspects of the study’s design and analysis introduce bias and reduce the quality of the evidence. First, the intervention arm of the group was compared to a group with no intervention. With no standard therapy for the non-EVAR group, there is risk of bias in the intention to treat design of this study. Indeed, in the no intervention arm [4], there was significant crossover with 33.8 % of patients undergoing aneurysm repair; 12 patients undergoing open repair and 35 patients undergoing EVAR. Second, the 7.3 % perioperative mortality exceeds that of other studies evaluating EVAR in the elective setting. This may be in part due to the 18 patients randomized to EVAR that did not undergo an intervention [4]. Additionally, in the EVAR arm, 4 patients underwent open aneurysm repair, 2 for a ruptured aneurysm. Finally, there were no strict inclusion or exclusion criteria beyond recommended guidelines and the treating physician decided suitability for enrollment [5, 10].

The failure of this study to demonstrate EVAR to be protective against aneurysm related mortality could relate to an underpowered study, high incidence of aneurysm repair in the no treatment arm and, potentially, the use of older endograft technology. Their conclusion of EVAR not being effective in patients deemed unfit for open surgery may not be justified because of these confounding factors [4].

In a retrospective review using the Vascular Study Group of New England (VSGNE) database, De Martino et al. stratified patients undergoing EVAR for intact AAAs <6.5 cm into those patients deemed fit and unfit for open surgery. As with EVAR-2, the treating physician made the decision regarding fitness for open surgery. The cohort consisted of EVAR performed in 1344 patients fit for open surgery and 309 unfit for open surgery due to physiologic high-risk. Patients in the unfit group were older and had more heart disease, chronic obstructive pulmonary disease and larger aneurysms than the fit group. They observed a statistically significant difference in all cause perioperative morbidity with rates of 3.7 % in the fit for open surgery group and 12.5 % in the unfit group. They also observed statistically higher rates of cardiovascular complications, respiratory complications, need for vasopressors and intestinal ischemia in the unfit group. Though there was a slight statistically significant increase in number of ICU days for the unfit patients, they did not report an increased total length of stay. Despite having more complications in the unfit group, they do not appear to have affected the overall hospital course. The perioperative morality rate of 0.3 % in the fit for surgery group and 0.7 % in the unfit group was not statistically significant [3].

The long-term results from the VSGNE cohort showed that patients undergoing EVAR deemed unfit for open surgery did worse than those who were fit for open surgery, as one would expect. At 1, 3 and 5 years, patients fit for open surgery had a survival of 96 %, 89 % and 80 % respectively, compared to survivals of 93 %, 73 % and 61 % for those deemed unfit. As part of their analysis, they divided the unfit patients into patients meeting any criteria in the suggested guidelines for inclusion in EVAR-2 (Table 8.3) and those without any criteria for EVAR-2. The patients in the VSGNE cohort meeting the EVAR-2 guidelines had an even worse 5 year survival rate at 46 % compared to the fit patients at 80 % and those without any EVAR-2 criteria at 71 % (Fig. 8.1) [3]. Because it only evaluated patients undergoing intervention, this study was not able to determine the relative benefit of EVAR compared to no intervention. However, this study does demonstrate that, for patients undergoing EVAR, survival is worse in those with significant medical comorbidities.

In another nonrandomized trial, Hynes et al. did not observe the same results seen in EVAR-2. In this study of high-risk patients with AAA, they prospectively followed 66 patients undergoing EVAR and 44 treated with medical therapy at a single institution. Patients were administered treatment based on the clinical setting as determined by the treating physician. Their observed survival at 4 years was 78.8 % for EVAR compared to 27.9 % in the medical therapy group. In this study, survival without aneurysm related mortality was 97.7 % in the EVAR group compared to 66.8 % in the medical therapy group. Only advanced age and aneurysm size were risks factors for aneurysm related mortality. The EVAR patients had significantly higher QOL when compared to the medical treatment group at 4 years. The QOL metric they used also included a length of life component, so the increased mortality in the medical treatment group largely accounts for the observed difference. Only 5.5 % of the EVAR group required reintervention at 4 years, a much lower rate than was seen in EVAR-2. Patients assigned to treatment based on clinical judgment rather than explicit patient factors introduces a large potential for bias [6]. While the study cohorts likely differed in severity of both physiologic and anatomic risk, it does demonstrate that survival and quality of life is worse in aneurysm patients that do not undergo EVAR. Additionally, the surgeons participating in this study were generally able to identify those highest risk patients and avoid subjecting them to a potentially unnecessary procedure.