Secondary Endovascular and Conversion Procedures for Failed Endovascular Abdominal Aortic Aneurysm Repair: Can We Still Be Optimistic?

Patients and Methods

From January 1997 until December 2005, 625 patients (92 women and 533 men) underwent elective EVAR, and their operative data were collected prospectively after approval from the Scientific and Ethical Committee of our institution. Preoperative evaluation included in all AAA patients a contrast-enhanced computed tomographic (CT) scan with 3 mm cuts. Patients with evidence of unsuitable anatomy for EVAR and those with long life expectancy or stated preference for open repair were treated with conventional open AAA repair techniques. Exclusion criteria for EVAR were a short (< 12 mm) or wide (> 30 mm) proximal infrarenal aortic neck, angulation of the proximal neck exceeding 60°, excessive neck thrombus and/or calcification, and severe tortuosity or circumferential atherosclerotic narrowing of the iliac vessels. Primary EVAR failure occurred in eight (1.3%) patients, who underwent primary abdominal conversion and were excluded from the study. The mortality rate of primary conversions was 25% (n = 2). One of these patients had complete iliac rupture and excessive hemorrhagic shock and died on the first postconversion day from multiple organ failure. The second patient also died on the first postconversion day from acute myocardial infarction after primary EVAR failure owing to proximal misplacement of the stent graft above the renal arteries.

The mean age of 617 patients with primary EVAR success was 69.4 ± 7.6 years, and the mean maximum AAA diameter was 53.5 ± 6.4 mm. The mean proximal neck length and diameter were 13.1 ± 1.6 mm and 27.4 ± 1.5 mm, respectively, and the mean proximal neck angle was 42.5 ± 11.6° (Table 1).

All EVARs were performed using general endotracheal anesthesia, by vascular surgeons in an operating room, under fluoroscopy with a portable C-arm with angiographic and roadmapping capabilities. Bifurcated devices were used in 533 (86.4%) patients and aortouni-iliac stent grafts in 73 (11.8%) patients, and in 11 cases (of the first 3 years), we used tube aortoaortic stent grafts. In 467 (75.7%) cases, still commercially available stent grafts were used, whereas in 150 patients (24.3%), the stent graft used has been withdrawn from the market. Table 2 shows the different stent graft devices used through the 9 years of the study.

The mean follow-up was 46.7 ± 11.2 months (range 1–94 months). The patients were placed under a strict surveillance protocol, and routine follow-up included clinical examination, abdominal plain radiography, and contrast-enhanced CT (including delayed scanning) at 1, 6, 12, 18, and 24 months and annually thereafter. Angiography or urgent CT was performed in selected cases of suspected post-EVAR-related complications. Detection of complications and early or late EVAR failure was based on definitions of the Committee for Reporting Standards for Endovascular Repair ⁷ and on the classification of endoleaks and endotension proposed by Veith and colleagues. ⁸

Results

The main reasons for secondary procedures were proximal migration and stent graft kinking in 60 and 59 cases, respectively. Kaplan-Meier analysis showed that the cumulative freedom rates of stent graft migration and kinking were 90.6% and 91.3%, respectively, at 48 months of follow-up. Additionally, the cumulative rate of aneurysm-related death in Kaplan-Meier analysis was 0.96% for the entire follow-up period (Figure 1).

OPEN IN VIEWER

Figure 1.

Cumulative freedom from stent graft migration in the 617 patients with primary endovascular aneurysm repair (EVAR) success and aneurysm-related death rate in 625 EVAR patients (Kaplan-Meier analysis).

All patients treated with secondary endovascular procedures or with peripheral interventions experienced an uncomplicated and short-duration postinterventional hospital stay (mean ± SD = 3.6 ± 0.8 days). The perioperative mortality of these patients was 0%.

Acute secondary abdominal conversions showed an intraoperative mortality rate of 20%, but the postoperative hospital stay was uncomplicated, and the morbidity rate of these patients was 0%. Conversely, elective secondary abdominal conversions showed a low mortality rate of 8.8% but were linked to significant morbidity. Nearly 65% of these patients experienced postconversion life-threatening morbid conditions requiring aggressive medical or surgical management with prolonged intensive care and total hospital stay. In six patients, the conversion surgery resulted in acute renal insufficiency, with three requiring permanent renal dialysis. Two patients survived after acute postoperative myocardial infarction, and four patients had a prolonged hospital stay owing to pulmonary infection and respiratory insufficiency. Two patients with occlusion of both hypogastric arteries showed evidence of ischemic colitis and gluteal ischemia. One of these patients required left hemicolectomy, whereas the second patient had colonoscopic evidence of partial mucosal ischemia and was treated conservatively with broad-spectrum antibiotics and parenteral nutrition. Control colonoscopy in the first month showed complete resolution of mucosal ischemia in this patient. Six patients developed large retroperitoneal hematoma requiring drainage, and one had distal embolization requiring thromboembolectomy on the third postconversion day. Finally, one patient experienced acute nonlithiasic cholecystitis and was treated with urgent cholecystectomy without further complications. The mean postoperative hospital stay of patients undergoing secondary abdominal conversion was 12.6 ± 4.8 days.

Bivariate (ANOVA) analysis of anatomic suitability for EVAR data showed significant differences in the maximum diameter of the AAAs and the length and width of the proximal necks between group A patients and groups B and C patients (p < .001; see Table 1). Conversely, proximal neck angle had a similar distribution and no significant differences among patients of the three groups (p = 1.0; see Table 1). Data analysis among patients undergoing secondary procedures (group B vs group C) showed no significant differences in all studied parameters (p > .05; see Table 1). Additionally, commercial withdrawal of the stent graft was a significant factor for secondary procedures (group A vs B and group A vs C, p < .001; see Table 1). However, comparison of commercial availability between patients of groups B and C showed no significant difference (p = .697; see Table 1). Multivariate logistic regression analysis revealed that stent graft withdrawal from the market and AAAs' maximum diameter were strong determinants for secondary procedures. Furthermore, the proximal neck width and length had a significant but less powerful correlation with secondary procedures (Table 4).

Discussion

Criticism of EVAR is focused on the frequent need for repeated interventions and open conversions with significant morbidity and mortality rates, as well as the increased postplacement cost. ^1,10 Secondary procedures significantly affect the long-term outcome and survival EVAR rates, but only a few reports have addressed the clinical impact of secondary procedures and especially of conversion surgery for failed EVAR.

Life table analysis of European Collaborators on Stent-graft Techniques for Abdominal Aortic Aneurysm Repair (EUROSTAR) data reveals freedom from secondary interventions to be 42% at 72 months. ¹¹ Recently, the EVAR Lifeline Registry showed that among 2,664 patients treated with EVAR, 583 (21.8%) patients required a secondary procedure. Endovascular interventions were performed in 487 (18.2%) of these cases, and 96 (3.6%) patients were treated with open conversion. ¹² The incidence of secondary procedures in our series is almost identical to the reported results of the EVAR Lifeline Registry, but the reported incidence of secondary procedures is greater in many large institutional series. ^13,14 However, the comparison of published results from different EVAR institutional series has limited value owing to differences in the used stent grafts, in the patients' selection criteria, in the quality and duration of follow-up, and in the criteria and decision for repeated intervention. ^15,16

All patients of our series met the anatomic inclusion criteria for EVAR, but data analysis revealed that patients who experienced a secondary procedure had aneurysms slightly greater in diameter. Additionally, these patients most frequently had a wider and/or shorter proximal aortic neck. The impact of these factors on EVAR's clinical outcomes has been well recognized. ¹⁷ Patient selection for EVAR and use of a proper stent graft according to indications of use remain important issues. Our series included AAAs with a short proximal neck length (12–15 mm), but nearly 60% of these patients were > 3 according to the American Society of Anesthesiologists classification, and treatment with EVAR resulted in a low aneurysm-related death rate, despite the high incidence of secondary interventions. The commercial availability of the EVAR stent graft also plays an important role. In 41% of our secondary abdominal conversions and in 36% of our secondary endovascular or peripheral procedures, the used stent graft had already been withdrawn from the market. Multivariate analysis revealed that stent graft commercial availability was the most powerful factor correlated with repeated intervention (odds ratio [OR] 2.792) followed by AAA maximum diameter (OR 2.199). Statistical analysis failed to show any significant correlation between secondary procedures and different stent graft types or timing of EVAR failure.

The main causes for repeated interventions include EVAR technique flaws and stent graft pitfalls. Further progression of the aneurysmal disease appears to be the etiology for a minority of these cases. ^18–20 In our series, recognized graft-related failure was the reason for late acute or elective abdominal conversions in 11 (31%) patients. Furthermore, the impact of graft generation–related failure in 139 secondary procedures remains unknown but is possibly higher. In contrast to the previously reported 3.1% incidence of endotension, this occurred in 0.8% of our cases, probably owing to extensive radiologic imaging and accurate endoleak detection with delayed CT. ^21–23 According to Greenberg and colleagues, the proximal migration is significant when it exceeds 10 mm, ⁹ but we treated all proximal migrations at 5 mm, even cases without type I endoleak because these patients had a short (12–15 mm) proximal neck and further migration would be dangerous.

Despite previous reports, true proximal neck enlargement owing to further aneurysmal disease progression was the reason for secondary procedures in only 0.2% of our patients. ^24,25 The precise proximal landing and accurate iliac fixation of the graft probably reduced the possibility of late migration. ²⁶ We believe that some of the previously reported late migration cases due to proximal neck enlargement might be related to graft or technique pitfalls and that the true incidence of further aneurysmal proximal neck expansion is actually very low.

In our series of 139 secondary procedures, 84 patients (60.5%) were treated with merely endovascular intervention and 16 patients (11.5%) were treated with peripheral (crossover bypass) surgery. All of these less invasive procedures shared an uncomplicated and short hospital stay, with 0% morbidity and mortality rates and durable results during the entire follow-up period. In our protocol of EVAR failure management, the secondary endovascular intervention and peripheral procedure are the first treatment options in acute or elective cases. Our indications for secondary abdominal conversions were ruptures in unstable patients, thrombosis of the stent graft's main body, and extreme tortuosity of iliac limbs.

Suprarenal or juxtarenal proximal clamping was not necessary in any of our secondary abdominal conversions, even in cases with removal of the suprarenal fixation stent graft part. Gentle stent graft withdrawal with application of an infrarenal clamp was effective in all cases. This was possible since there was a sufficient length (≅10 mm) of proximal neck in all cases and the majority of the removed stent grafts were without barbs and hooks. In three cases with solid stent graft proximal fixation with hooks, we used the previously described inclusion technique with incorporation of the suprarenal stent graft part in the anastomotic suture line, avoiding the hazardous suprarenal clamping. ^27–30 Following this technique, renal failure occurred in six of our patients. The suprarenal aortic dissection and clamping offers better surgical field and safer removal of the stent graft but is related to high mortality rates, and we avoid using it. Additionally, techniques using balloon suprarenal aortic occlusion via femoral or brachial access have been described to extend our ability for proximal control without suprarenal clamping even in more demanding cases. ²⁰

The mortality of acute late conversion in our patients was as high as 20%, but despite the severe comorbidity, the mortality rate of elective abdominal conversions decreased to below 9%. Finally, the total mortality rate of all 139 secondary procedures was 2.9%, whereas the overall EVAR-related mortality rate was 0.6% among the 617 EVAR-treated patients and below 1%, including the 8 cases with primary EVAR failure (625 patients).

Reported EVAR studies, including our series, conventionally compare results from mid- and long-term follow-up with those with old- and new-generation devices, which can bias the results and exclude firm conclusions. Using currently available devices, however, many of the graft-related problems have been solved. Contemporary stent grafts offer secure proximal landing in the proximal neck, adequate contralateral limb overlapping and sealing, and better balance between device rigidity and flexibility.

Conclusions

Conversion surgery and acute or elective removal of stent grafts still carry high morbidity and mortality rates, and this emphasizes the value of selecting proper patients and stent grafts for EVAR. Our data support the position that adequate EVAR technique, close follow-up, and early detection of failed EVAR can result in excellent primary and secondary EVAR success rates, with very low mortality rates. We conclude that this justifies optimism concerning the primary and secondary EVAR results. Longer follow-up studies with inclusion of merely last-generation stent grafts, however, are necessary to draw robust conclusions about the safety, efficacy, and durability of EVAR and the lack of secondary endovascular or conversion procedures.