Endovascular Repair of Abdominal Aortic Aneurysms

Abstract

Objective

The impact of co-morbid conditions on early and late clinical outcomes after endovascular treatment of abdominal aortic aneurysm (AAA) was assessed in concurrent cohorts of patients stratified with respect to risk for intervention.

Summary Background Data

As a minimally invasive strategy for the treatment of AAA, endovascular repair has been embraced with enthusiasm for all prospective patients who are suitable anatomical candidates because of the promise of achieving a durable result with a reduced risk of perioperative morbidity and mortality.

Methods

From April 1994 to March 2001, endovascular AAA repair was performed in 236 patients using commercially available systems. A subset of patients considered at increased risk for intervention (n = 123) were categorized, as such, based on a preexisting history of ischemic coronary artery disease, with documentation of myocardial infarction (60%) or congestive heart failure (35%), or due to the presence of chronic obstructive disease (21%), liver disease, or malignancy.

Results

Perioperative mortality (30-day) was 6.5% in the increased-risk patients as compared to 1.8% among those classified as low risk (P = NS). There was no difference between groups in age (74 ± 9 years vs. 72 ± 6 years; mean ± SD), surgical time (235 ± 95 minutes vs. 219 ± 84 minutes), blood loss (457 ± 432 mL vs. 351 ± 273 mL), postoperative hospital stay (4.8 ± 3.4 days vs. 4.0 ± 3.9 days), or days in the ICU (1.3 ± 1.8 days vs. 0.5 ± 1.6 days). Patients at increased risk of intervention had larger aneurysms than low-risk patients (59 ± 13 mm vs. 51 ± 14 mm;P < .05). Stent grafts were successfully implanted in 116 (95%) increased-risk versus 107 (95%) low-risk patients (P = NS). Conversion rates to open operative repair were similar in increased-risk and low-risk groups at 3% and 5%, respectively. The initial endoleak rate was 22% versus 20%, based on the first CT performed (either at discharge or 1 month;P = NS). To date, increased-risk patients have been followed for 17.4 ± 15 months and low-risk patients for 16.3 ± 14 months. Kaplan-Meier analysis for cumulative patient survival demonstrated a reduced probability of survival among those patients initially classified as at increased risk for intervention (P < .05, Mantel-Cox test). Both cohorts had similar two-year primary and secondary clinical success rates of approximately 75% and 80%, respectively.

Conclusions

Early and late clinical outcomes are comparable after endovascular repair of AAA, regardless of risk-stratification. Notably, 2 years after endovascular repair, at least one in five patients was classified as a clinical failure. Given the need for close life-long surveillance and the continued uncertainty associated with clinical outcome, caution is dictated in advocating endovascular treatment for the patient who is otherwise considered an ideal candidate for standard open surgical repair.

With the advent of an endovascular treatment option for the abdominal aortic aneurysm (AAA), defining an appropriate strategy for the referral of patients to either open or endovascular repair remains a complex clinical endeavor. For example, patients who were otherwise appropriate surgical candidates for standard open repair have populated most, if not all, industry-sponsored clinical trials conducted in the United States. ^1–3 Among these patients, a significant reduction in hospital stay has been demonstrated, with early return to preoperative levels of activity. Enthusiasm for endovascular treatment for the patient at low risk has also been coupled with the proposition that endovascular therapy provides an ideal approach for patients in whom standard operative repair carries an increased risk of perioperative morbidity and mortality. ⁴ Indeed, endovascular treatment has increased the proportion of patients now referred for AAA repair by providing therapy for patients who have been deemed inoperable because of the presence of significant comorbid conditions. Nonetheless, the widespread advocacy of endovascular grafting as a preferred option to open surgery for potentially all anatomically suitable patients continues during a period when most studies have reported outcomes that are largely confined to early intervals after intervention.

We recently reported the clinical experience with endovascular AAA repair at our institution. ⁵ This updated report reviews our mid-term experience with endovascular AAA repair over a 7-year period by examining early and late clinical outcome in concurrent cohorts of patients stratified either as patients at low risk, who would otherwise be considered ideal open surgical candidates, or as those who are at increased risk for intervention. In these two groups of patients, we assessed perioperative morbidity and mortality, technical success, and late clinical success rates and patient survival.

METHODS

Patient Selection

Data for 236 consecutive patients undergoing elective endovascular AAA repair at Emory University Hospital (Atlanta, GA) were retrospectively collected from April 1994 through March 2001. An endovascular program was initially instituted at Emory University as part of an investigator-sponsored, investigational trial (Endovascular Technologies, Inc, Menlo Park, Calif/Guidant, Inc, Indianapolis, ID). This program expanded in 1999 to include a second investigational device (Excluder, WL Gore and Associates, Inc, Flagstaff, AZ). We have also used the AneuRx (Medtronic, Inc, Sunnyvale, CA) endograft system after its approval by the Food and Drug Administration for commercial use in September 1999. During the study period, implanted endografts included the EVT/Guidant endograft (n = 150), the AneuRx stent graft system (n = 58), and the Excluder endograft (n = 28). The EVT/Guidant endografts included tube (n = 26), bifurcated (n = 109), and aortoiliac endografts combined with a femorofemoral bypass graft (n = 15). The Gore endografts were all phase II devices.

Patients were considered at increased risk for intervention if there was 1) documentation of previous myocardial infarction (MI) or congestive heart failure; 2) significant respiratory disease as demonstrated by a forced expiratory volume in 1 second of < 1 liter/min or a requirement for home oxygen therapy; 3) chronic liver disease with documented cirrhosis or portal hypertension, or; 4) the presence of concurrent or recent malignancy. Of note, all patients underwent preoperative cardiac risk assessment that included dobutamine echocardiography or persantine thallium scanning.

Endograft Implantation

All endovascular AAA repairs were performed in a standard operating room environment with complete angiographic capability by a team of vascular surgeons and interventional radiologists. The techniques of transfemoral endovascular AAA prosthesis implantation have been described previously. ^1–4,6 Fluoroscopic guidance (OEC 9600, OEC Medical Systems, Inc, Thousand Oaks, CA) was used for placement of the endoprosthesis, and most of the procedures were performed with the patients under general anesthesia. All patients underwent systemic anticoagulation with 100 U/kg heparin. Postimplantation aortography was performed to assess graft positioning, vessel patency, periprosthetic leakage, and graft limb stenosis. Type I endoleaks, (leakage around the proximal or distal attachment site) were treated during the operation with additional endovascular measures. Type II endoleaks (those through retrograde lumbar or inferior mesenteric arteries) were observed and monitored with serial CT scans. At the discretion of the attending physician, this type of endoleak was treated with coil embolization of the patent collateral pathway.

Clinical Follow-up

Contrast-enhanced CT was performed either in the immediate postoperative period or within 1 month of endograft placement. Additional imaging studies including CT, duplex ultrasound scanning, and plain abdominal x-ray evaluation were performed at 6 months, 12 months, and then annually thereafter. If an endoleak was visualized, more frequent surveillance imaging was performed as clinically indicated.

Definitions

All perioperative complications are described. However, major morbidity was defined as any complication that resulted in an increase in hospital stay, a secondary surgery, or a significant disability. The definitions of technical success, clinical success, and continuing success as described by the Society for Vascular Surgery/International Society for Cardiovascular Surgery (SVS/ISCVS) Ad Hoc Committee on Reporting Standards for Endovascular AAA Repair were used. ⁷ In brief, 30-day technical success was defined on an intent-to-treat basis as successful endograft deployment without death, need for standard aortic reconstruction for 30 days, or evidence of persistent (>48 hours) endoleak. Clinical success was inclusive of those patients who at 6 months after implantation had spontaneously sealed a persistent endoleak and had demonstrated no evidence of aneurysm enlargement. Secondary clinical success was used if additional endovascular techniques were required to seal an endoleak. Continuing success was defined as the maintenance of both clinical and technical success without evidence of graft thrombosis, infection, endoleak, or aneurysm expansion of greater than 0.5 cm. Any late graft complication that was successfully treated by an endovascular technique was classified as a secondary continuing success. Other outcomes analyzed included successful graft deployment irrespective of the presence or absence of endoleak, surgical time, operative blood loss, duration of stay in an intensive care unit, length of hospital stay, and patient survival.

Statistical Analysis

Descriptive data are expressed as mean ± SD. Continuous variables were compared with the use of the Student t-test. Nominal variables were analyzed by contingency tables. The Kaplan-Meier method with Mantel-Cox (log-rank) posthoc analysis was used to determine success and survival rates. P < .05 was considered statistically significant. An SAS statistical package was used for analysis (Version 5.0, Abacus Concepts, Berkeley, CA).

RESULTS

Between April 1994 and March 2001, elective endovascular repair of infrarenal AAA was carried out on 236 patients, with 123 (52%) procedures conducted in patients classified at increased risk and 113 (48%) procedures performed in patients considered low risk for major morbidity or mortality. The incidence of comorbid conditions among patients deemed at increased risk for intervention is presented in Table 1. Patient and procedural characteristics for these two groups are summarized in Table 2, and the types of endografts implanted are described in Table 3.

Table 1. CHARACTERISTICS DEFINING PATIENTS AT INCREASED RISK FOR INTERVENTION (n = 123)^*

* Patients may have had more than one factor increasing the risk of intervention.

† Chronic obstructive pulmonary disease documented by pulmonary function testing with a forced expiratory volume in 1 second ≤1 L/min or the need for home oxygen therapy.

†† Child’s class B.

§ Primary lung cancer (n = 5), metastatic colon cancer (n = 2; Duke’s stage D), laryngeal cancer (n = 1), transitional cell carcinoma of the bladder (n = 1).

Table 2. COMPARISON OF PATIENT SUBGROUPS UNDERGOING ENDOVASCULAR AAA REPAIR

Mean ± SD.

NS, No statistical significance; ICU, intensive care unit.

* Includes only patients having successful endograft deployment.

Table 3. TYPES OF ENDOGRAFTS IMPLANTED

* Aortoiliac endograft performed in conjunction with contralateral common iliac artery occlusion and femorofemoral cross-over graft.

† No significant difference when analyzed by Fischer’s exact test.

Notably, cardiac disease was a major indication for the categorization of patients at increased risk for intervention. To obtain a more precise determination of the severity of cardiac disease in our population, additional risk stratification of patients was performed with the SVS/ISCVS Cardiac Grading System. ⁸ In brief, cardiac status is graded with a 0 to 3 flat scale where grade 0 indicates a patient with no symptoms and a normal electrocardiogram (ECG); grade 1 is used for a patient with no symptoms and a history of a remote MI (> 6 months), occult MI by ECG, or fixed defect on dipyridamole thallium or similar scan; grade 2 is used for the patient with stable angina, the presence of a significant reversible perfusion defect on dipyridamole thallium scan, ejection fraction of 25% to 45%, controlled ectopy/arrhythmia, or compensated congestive heart failure; and grade 3 is used for patients with unstable angina, ejection fraction of less than 25%, symptomatic or poorly controlled ectopy/arrhythmia, poorly compensated or recurrent congestive heart failure, or MI within 6 months. Patients in the study classified as low risk (n = 113) had an SVS/ISCVS cardiac score of 0, whereas patients categorized at increased risk solely by a history of cardiac disease had a score of 1.82 ± 0.53 (n = 71). Of note, most patients with cardiac disease had a score of 2 (58%; 41 of 71) or 3 (8%; 6 of 71).

Technical and Clinical Success

Endovascular stent graft deployment was successful in 116 (95%) of 123 of patients at increased risk and in 107 (95%) of 113 patients at low risk, with conversion rates of 2.4% and 5.3%, respectively. No intraoperative deaths occurred. Intraoperative conversions to open repair and/or aborted procedures all occurred during attempted implantation of EVT/Guidant endografts, except one case of attempted AneuRx endograft placement. These technical failures were not clustered during any given time period. In the increased-risk group, there were three immediate conversions to open repair and four aborted procedures. The sole case of AneuRx endograft conversion occurred when a contralateral catheter was caught in the nitinol strut and was unable to be removed. In the second case of immediate conversion, the distal attachment hooks of an EVT/Guidant tube graft became caught on the aortic bifurcation and were unable to be released. In the third case, a device twist was not resolvable with endoluminal techniques. Two aborted procedures occurred in patients with tortuous, heavily calcified iliac arteries. One patient subsequently died of progressive congestive heart failure several weeks after hospital discharge, while the other patient declined open repair. The third and fourth aborted procedures were also related to an inability to access the aneurysm. The third patient declined open repair and subsequently had a fatal aneurysm rupture, and the fourth patient died 6 months later. The cause of death in this patient was not determined. A late conversion also occurred in this group at 30 months. A patient who underwent implantation with the original EGS (EVT, Inc.) system had attachment system failure in the form of a hook fracture. This was recognized because of the presence of a persistent endoleak and aneurysm enlargement.

The results for the low-risk group were similar, with six conversions. Two were related to iliac artery injury and two to the inability to access the aneurysm because of narrowed and calcified iliac arteries. Two cases of EVT/Guidant device malfunction occurred during deployment. In all six cases that required conversion, successful open repair was performed without postoperative complications. Two late conversions occurred in the low-risk group, one a consequence of a hook fracture identified at 26 months and the other of a graft infection at 2 months.

The 30-day technical success rates as defined by the SVS/ISCVS reporting standards were 73% for the increased-risk group and 78% for the low-risk group (P = NS). At 1 month after implantation, 25 (20.3%) patients at increased risk and 21 (19.6%) at low risk had endoleaks detected by CT imaging. These results remained essentially unchanged at 6 months, with clinical success rates at 6 months of 83% for the patients at increased risk and 80% for the low-risk cohort. Thirteen patients at increased risk and five at low risk had spontaneous sealing of their endoleaks. All remaining endoleaks were observed during this period, and no further intervention was taken in this regard. A statistically significant relationship between Type I and II endoleaks and graft type was not detected. Continuing primary and secondary success as defined by the SVS/ISCVS reporting standards are represented in Figure 1 and were 73.5% ± 10.2% and 76.5% ± 9.3% for increased-risk and low-risk groups, respectively, at 24 months. If the definition of clinical success is revised to exclude the presence of a Type II endoleak, Kaplan-Meier analysis revealed clinical success rates at 24 months of 76.2% ± 19.60% and 82.3% ± 11.50% for increased-risk and low-risk groups, respectively.

Figure 1. Primary (A) and secondary (B) continuing success rates for low-risk (▴) and increased-risk (•) groups presented by Kaplan-Meier method. Statistically significant differences in success rates were not observed.

Adjunctive endovascular techniques were used in both groups to facilitate graft implantation and aneurysm exclusion. In 11 patients with increased risk, one or both limbs of a bifurcated graft had intraluminal stents placed for fabric folds observed with either intravascular ultrasound scanning or angiography at the time of endograft deployment. Intraluminal stents were also placed in 19 patients at low risk. Internal iliac arteries were unilaterally embolized in 13 patients (seven at high risk, six at low risk) for the exclusion of ectatic or aneurysmal common iliac arteries. Iliac artery dissection was noted in one patient in each study group at the time of graft implantation and was treated successfully in both cases with stent coverage.

Complications

The perioperative complication rate was 17.5% and 15.0% in the increased- and low-risk groups, respectively (Table 4). All wound infections were superficial and successfully treated on an outpatient basis with local wound care and antibiotic therapy. Two patients developed acute renal failure necessitated hemodialysis. Overall, major morbidity necessitating an increase in hospital stay or significant disability occurred in 4% (5 of 123) of patients at increased risk and 6% (7 of 113) of patients at low risk.

Table 4. PERIOPERATIVE (30-DAY) COMPLICATIONS

* Twisting of one limb of a nonsupported bifurcated graft required treatment with a femorofemoral crossover graft at the time of stent graft implantation.

Follow-up

Follow-up data were complete for all patients, with a mean follow-up interval of 17.4 ± 15 months for patients at high risk and 16.3 ± 14 months for the low-risk group. No patient was lost to follow-up. The perioperative (30-day) mortality rates were 6.5% and 1.8% for the increased-risk and low-risk groups, respectively (P = .2013, Fisher exact test). Eight perioperative deaths occurred in the group at increased risk for intervention. One death occurred in a patient who required conversion from endovascular repair to open repair and one in a patient who had an aborted procedure and severe coronary artery disease. The third death occurred in a patient who had a successful endovascular repair without evidence of postoperative endoleak. A malignant arrhythmia was the presumed cause of death. Two deaths occurred due to postoperative myocardial ischemia. One patient with severe chronic obstructive pulmonary disease and emphysema developed pneumonia postoperatively. He developed adult respiratory distress syndrome that eventually contributed to his death. The seventh patient developed acute renal failure and pneumonia postoperatively and died two weeks following the endovascular aneurysm repair. The eighth patient died of severe heart failure after hospital discharge. Endograft deployment had been successful in this patient, and no endoleak had been detected by CT scanning at the time of discharge. Fifteen other patients died during the follow-up period.

In the low-risk group, two perioperative and nine late deaths occurred. One death was due to intraoperative hemorrhage and another death occurred due to presumed postoperative pulmonary embolism. Kaplan-Meier cumulative survival curves are shown in Figure 2. The two-year mortality rates were 26.5% ± 8.1% and 14.2% ± 7.5% for the increased- and low-risk groups, respectively. A significant difference between the two patient cohorts was noted by the Mantel-Cox (log-rank) test (P = .035). None of the reported late deaths in our series were related to the initial endovascular procedure, device failure, or late aneurysm rupture.

Figure 2. Cumulative survival rates for low-risk (▴) and increased-risk (•) groups presented by Kaplan-Meier method. A statistically significant difference in survival in favor of the low-risk group was noted by the log-rank test (P < .035).

DISCUSSION

The introduction of endovascular grafting was a milestone in the treatment of patients with AAA in that it provided a treatment option for those patients with large aneurysms who had been deemed inoperable because of the presence of significant medical comorbidities. ⁴ In the extension of this technology to all patients with aneurysmal disease, clinical investigations have confirmed that compared with open surgery, an early benefit in quality of life can be achieved, as it relates to reducing hospital stay and recovery period. ^1,2 Nonetheless, even minimally invasive interventions may be associated with an adverse early outcome, and the presumption that an endovascular approach reduces perioperative mortality in patients at low risk compared with the results of standard surgery remains unproven. Moreover, an early benefit in quality of life may be offset by a lower level of late clinical success that carries with it a requirement for more intensive long-term surveillance, increased rates of reintervention, and higher costs and psychological stress. Thus, in advocating endovascular treatment for patients who are at low risk for operative repair, a critical analysis of late outcomes is required.

In our retrospective analysis, patients classified at low risk for intervention with accepted clinical and laboratory criteria exhibited a 30-day mortality rate of 1.8% after endovascular intervention. This result is of particular interest in the context of a recent review of open aortic surgery performed on 856 patients at our institution between 1986 and 1996. ⁹ The in-hospital mortality rate was 1.3%, with a major complication rate of 15.9%. Thus, although the data generated by these two distinct reviews at the Emory University Hospital are not strictly comparable, our experience suggests that in the patient at low risk, endovascular treatment of the infrarenal AAA is not associated with a reduction in perioperative mortality compared with standard surgical repair.

Many reports, nevertheless, confirm that endovascular strategies do offer unique advantages among those patients whose comorbid conditions increase the risk of major complications including death. For example, May et al. ¹⁰ compared outcomes of patients treated concurrently with either open or endovascular repair. Although more than 40% of patients treated with endografts had been declined open repair because of comorbid illness, no significant difference in perioperative mortality rates or long-term survival was observed. In addition, Chuter et al. ¹¹ observed a 30-day mortality rate of 1.7% in their review of patients treated by endovascular approaches that had otherwise been refused conventional AAA repair. In their patient population, coronary artery disease was present in 81%, congestive heart failure in 34%, and respiratory disease in 49%. These reports compare favorably with published studies of conventional open aneurysmectomy in patients at high risk that have been associated with mortality rates of up to 40%. ^12–14

Our review does reemphasize, however, that conversion to an open repair and the aborting of an endograft procedure may not be well tolerated among those patients with significant comorbidities. This is consistent with results reported by May et al., ^10,15 who have noted mortality rates of 18% to 43% when primary conversion was required for patients considered at prohibitive risk for standard surgery. It is our view that both the prolonged anesthesia time and the blood loss incurred during a preliminary attempt at endovascular repair before conversion are important contributing factors to these poor results. Therefore, a cautious approach should be adopted in recommending endovascular repair for the patient at high risk in the presence of anatomical constraints, which might reduce the potential for successful endograft deployment.

With SVS/ISCVS-recommended reporting standards, 30-day primary technical success, 6-month clinical success, and 24-month primary and secondary continuing clinical success rates were all approximately 75% in both study subgroups. Our 30-day primary technical success rate is similar to the 77% rate reported by Zarins et al. ² for 190 patients treated as part of the multicenter Medtronic AneuRx stent graft trial. Likewise, our 24-month success rate is comparable to that recently reported by May et al. ¹⁶ for second-generation endovascular prostheses used in 148 patients. Thus, although these results are encouraging and will undoubtedly improve in coming years, the success of endovascular repair remains uncertain in a significant proportion of patients. Two years after endograft implantation, 25% of all patients were classified as failures with the SVS/ISCVS reporting standards definition. ⁷ Therefore, in advocating an aggressive approach for endovascular intervention in the patient who is an otherwise ideal surgical candidate, it is also important to recognize that significant limitations to endovascular repair remain. Moreover, the impact of this failure rate on increasing costs and reducing patient quality of life is probably significant but admittedly was not defined in this report.

It is notable that all reported late deaths in our series were unrelated to the initial endovascular procedure, device failure, or late aneurysm rupture. Although late survival was significantly compromised in those patients who were deemed at increased risk for intervention, 75% were alive at 2 years. These results are not unexpected, and others have reported similar late mortality rates for patients initially considered poor surgical candidates. ¹¹ Nonetheless, all of this suggests that the benefit of endovascular repair may be limited for patients who have a compromised life expectancy. In this regard, patients with a concomitant history of recent or concurrent malignant disease are a subgroup of particular interest. Of the nine patients in this group, two died of progressive cancer 12 months after endovascular AAA repair. However, the six remaining patients were alive at the time of last follow-up (12.4 ± 8.2 months). Thus, given the imprecision in predicting the risk of AAA rupture and long-term survival either in response to cancer therapy or other major medical illness, decisions to proceed with endovascular repair must be carefully individualized. In this regard, we presently advocate endovascular intervention for the patient with significant medical comorbidity only when aneurysm size is equal to or exceeds 6 cm in diameter and patient life expectancy is estimated to exceed 2 years. We believe this to be a prudent recommendation given respective annual rates of rupture of approximately 6.6% and 19% for untreated patients with 5.7-cm and 7.0-cm diameter aortic aneurysms ¹⁷ and our combined major morbidity and 30-day mortality rate of 12% for the patient at increased risk for intervention.

In summary, our analysis suggests that endovascular aneurysm repair currently remains most appropriate for those patients with large aneurysms who are otherwise prohibitive operative candidates. It is significant that endovascular grafting provides these patients with a treatment option when one was not previously available. Advocating endovascular treatment for the patient who is at low risk for standard operative intervention remains problematic. Although clinical success can be achieved in most patients, inadequate results continue to be observed in a significant portion. In deciding on a course of treatment, an informed decision on the part of the patient requires a consideration of these data and an appreciation that endovascular aortic aneurysm repair remains in a relatively early stage of development.

Discussion

Dr. Gregorio A. Sicard (St. Louis, MO): I would like to thank Dr. Dodson and Dr. Chaikof and the Emory University Division of Vascular Surgery for allowing me to discuss this excellent paper as well as for the timely copy of the manuscript to review.

Over the last 10 years since the initial description by Juan Parodi, endoluminal repair of abdominal aortic aneurysm has undergone extensive scrutiny evaluating its efficacy and safety. Despite the phase 2 FDA trials of various devices, no prospective, randomized trial comparing open to endoluminal elective repair of aneurysm has been conducted, making a scientifically valid comparison unfeasible.

In this excellent report by Dr. Chaikof and the Emory University Vascular Surgery Group they extend the scrutiny of endoluminal grafts by comparing patients of high and low risk and ask the question to whom should it be offered. The excellent results presented reflect the experience of the Emory Group in performing endovascular repair of abdominal aortic aneurysm that may not be duplicable in most clinical settings. A few questions:

Did you notice any difference in your results in your earlier 2 or 3 years of experience comparing it to the last 4 or 5 years? In other words, what do you think is the learning curve for this procedure?

Several endografts were used in your series, with the largest number being the two currently commercially available grafts, the Ancure (Guidant) and the AneuRx Medtronic. Did you find any differences between these two grafts in technical and clinical success, complications, long-term survival, in either the low or high-risk groups.

I also notice in your presentation and manuscript that advanced age was not an indicator of high-risk status. Our recently published institutional experience comparing open versus endoluminal repair for elective aneurysm repair in 470 patients, 260 performed endoluminally and 210 open, showed that although there was not a significant difference in mortality between octogenarians and the younger-than-80 groups, although it was lower in the younger-than-80 group, there was a significant difference in the perioperative complication rate being significantly higher in the octogenarians compared to those younger than 80 years of age for both open as well as endoluminally repaired aneurysms. Did you evaluate your data for results in octogenarians? If so, did you find any differences?

Because of the high incidence of abdominal aortic aneurysms in men, as well as, anatomical constraints, women tend to be recipients of endografts less frequently than men. Could you comment about what percent of your patients in both either low- or high-risk groups were women and was there a difference in complication rate and long-term survival?

Finally, I would like to get your opinion: where do you think the indications and application of this novel approach for aneurysm repair is heading? We are currently using first and second-generation endografts. As these devices become smaller and easier to use and as procedure-related complications decrease, do you think that this technology should be limited to high-risk patients or that both patient predilection as well as increased technical success will make endoluminal treatment of abdominal aneurysm the treatment of choice in anatomically suited both high and low risk patients?

I really appreciate the opportunity to comment on this paper and the privilege of the floor.

Dr. James M. Seeger (Gainesville, FL): I also want to compliment Dr. Dodson on his nice presentation and the Emory Group for sending me their paper in a timely fashion to review.

Dr. Dodson presented results from a retrospective review of a relatively large number of patients undergoing endovascular aneurysm repair in a single institution. Furthermore, he has assessed and compared the outcomes in these patients categorized as low risk and increased risk groups, categories that seem clinically relevant to the decision-making concerning endograft repair of abdominal aortic aneurysms. The reported results are very acceptable and, as pointed out in the manuscript, very similar to other single institutional trials.

Beyond that, what have we really learned from this review? To me, there are three things: One, that attempted but unsuccessful stent graft repair of abdominal aortic aneurysms is associated with not inconsequential risk, particularly in patients with significant comorbidities. Second, that stent graft repair even in low risk patients is not risk free. And finally, that continuing clinical success defined as the percentage of patients surviving with continuing aneurysm exclusion in the Emory experience is about 80% at two years.

Like carotid endarterectomy, elective aortic aneurysm repair is a prophylactic procedure and thus to be beneficial must result in better outcomes than the natural history of the disease without treatment. It has been presumed that stent graft repair will be less morbid and thus would improve the benefit of intervention in the aneurysm disease. To date that has not been conclusively demonstrated. Therefore, given results such as those presented today, how should we currently choose the patients to whom stent graft repair of aortic aneurysm should be offered? As previously mentioned, I think that considering the patients in two different groups, low risk and high risk for open surgical repair, is valuable, as the risk-benefit ratio, and thus the decision to offer repair is different in these two groups. This raises several questions:

First, your definition of increased risk seems to me a bit broad. With even a history of MI or stable angina being included – and in fact, as noted in the manuscript, representing the majority of patients who were in the increased risk group, does this skew your results in the increased risk group and did the complications in this group of patients actually occur in the patient’s more severe comorbidities? If this is true, this to me makes it difficult to compare these results of endograft repair to previously reported results of open surgical repair inpatient judged high risk of open repair.

Second, as you point out in your manuscript, from a recent review of open surgical repair in your institution, the mortality associated with open surgical repair of infra-renal aortic aneurysms was equal to that in the low-risk patients undergoing endograft repair. However, I suspect that all your patients undergoing open repair were not without any clinical evidence of heart disease as it appeared from your manuscript that your low risk patients were. This would make the outcome of endograft repair in low-risk patients, who presumably were very good surgical candidates for open repair, less compelling when compared to the results of open surgical repair.

Finally, on a little bit more philosophical note – and this is, I think, where we all stand in trying to understand this – how do we evaluate this new approach to the treatment of aneurysm disease in two patient groups? Chris Zarins, for one, has suggested using the concept of rupture-free survival as the best method of assessing benefit after aneurysm treatment, much like we assess stroke-free survival in patients with carotid disease undergoing intervention. Based on your endograft results presented here and your open surgical reports reported in your paper, are you benefiting the high risk and the low risk patients when you use this method of comparing them? Using your numbers, my calculation suggests that you likely are benefiting the high-risk patients compared to no intervention and potentially even to open surgical repair. In contrast, surprisingly, the outcome at 2 years is essentially equivalent in the low-risk patients undergoing either endograft or open surgical repair. Calculating the rupture-free outcome in these groups, at least by my calculations, I saw only one patient in the surgical group that would have survived that would not have survived in the endograft group. I would appreciate your comments on this and your thoughts, as Dr. Sicard has suggested, on how these results are going to influence your selection of patients for endograft repair.

Dr. Raymond S. Martin, III (Nashville, TN): I enjoyed hearing the report of the Emory Group and found the outcomes similar to our own. I have a couple of additional questions:

First, did you find a difference in endoleak or other complications when using component versus unibody devices?

Second, was there a difference in the mean diameter or the complication rates in these patients and those treated by the open technique during the same time period?

Third, we found that some of our perioperative complications are related, in retrospect, to stretching the limits of the devices in terms of aortic neck diameter or iliac size, particularly small iliacs and tortuosity. Have you experienced the same thing? And do you think there is a role for endovascular repair in low risk patients using conservative criteria; that is, the best patients anatomically?

Finally, one of the disadvantages of the endograft procedure is the burdensome follow-up process. This alone is a reason to consider using the open procedure in patients with long life expectancies. Do you think that yearly CT scans are really necessary in patients after 2 or 3 years of stability?

Dr. Ian Hamilton (Chattanooga, TN): Whether it is open aneurysm repair with simultaneous renal artery revascularization or, as we see today, endograft repair, we continue to look to Emory for leadership in vascular surgery. I enjoyed the paper. A couple of questions:

With a young endograft program in Chattanooga, we have been counseled to try to apply this technology to our anatomically low-risk patients on the front end of our learning curve. I wonder if the similarities between high and low-risk patients in this group may reflect a disproportionate timing of low-risk patients that are being treated early in the learning curve of three different devices.

The other issue that is not addressed in the paper has to do with postendograft versus postopen aneurysm repair quality of life. I was wondering if the authors have any insight into those quality of life issues.

Dr. Gary Maxwell (Wilmington, NC): I have a question about what may come to be called “incidental” renal artery angioplasties. As I enter data into our vascular registry, I have noted about 30% of patients are receiving renal artery angioplasty. Whether it is “incidental” or not, I can’t be sure from the data, since I cannot identify the indications for the renal angioplasty.

My question is, how many of these patients have had concomitant angioplasty, and do you have any opinion about whether we are helping or hurting patients with regard to their hypertension? Do you have any data that speaks to that question?

Dr. Elliot L. Chaikof (Atlanta, GA): I would like to thank the discussants for their insightful comments and probing questions. Dr. Sicard inquired as to whether our technical success rates improved as our experience with endovascular devices grew. To some degree this has been the case, with a commensurate improvement in our experience in both patient selection and device deployment. Overall, technical success rates remain higher among lower profile devices, such as those produced by Medtronic and W. L. Gore, as compared to the higher profile Ancure system produced by Guidant, Inc. Nonetheless, we continue to believe that hook fixation, which is associated with the Ancure device, is an important feature for endograft stabilization. Overall, each of these devices has their own respective advantages and limitations.

Dr. Sicard requested information regarding our results among patients over 80 years of age and among women. There were 30 octogenarians in an intent-to-treat subgroup and successful deployment was achieved in 28. Significantly, major postoperative morbidity occurred in only two or 7% of these patients. Thus, a reduction in major postoperative morbidity, the capacity for patients to rapidly return to their preoperative quality of life, and a reduction in hospital stay continues to drive our interest in endovascular approaches to AAA repair. Although not demonstrated by randomized trail, we also believe that a reduction in major adverse postoperative events is inherently associated with a reduction in mortality among patients at increased risk for open repair. With regards to the distribution of male and female patients, access to an aortic aneurysm using an endovascular device is currently difficult or impossible with available commercial devices unless at least one iliac artery is at least 7 millimeters in diameter or larger. Most women with aneurysmal disease have relatively small iliac arteries and, therefore, are often excluded from consideration for endovascular repair. As a consequence, most of the patients in this report were men.

Dr. Seeger questioned whether our complications may have been disproportionately localized to a small group of patients at especially high risk for intervention and inquired as to whether rupture-free survival is an appropriate form for reporting results of endovascular AAA repair. In brief, complications in our analysis were evenly distributed among all patients. It bears reemphasis that the overall incidence of major postoperative morbidity was relatively modest, on the order of 5% among patients at increased risk for intervention, which compares to reported incidences in the literature of at least 10 to 15% for patients treated with open repair. Admittedly, reports that include rupture-free survival as a sole measure of success are limited by the inherent heterogeneity of aneurysm size within the study population and, as a consequence, varying risks of rupture from one treatment to another. Thus, use of this outcome measure alone would be comparable to an analysis of two different treatment strategies for cancer management without stratification with respect to tumor burden or overall stage of disease.

As a final comment, our indications for aneurysm intervention have not changed with the advent of endovascular approaches for repair. For the low-risk patient, intervention is recommended for aneurysms of at least 5 centimeters in diameter and for those at higher risk, treatment is recommended for aneurysms of at least 6 centimeters in diameter. Of course, the one exception to this statement is the patient with a large aneurysm and significant comorbidities who, before the introduction of endoprostheses, was not considered a treatment candidate. The utility of endografts for these patients is quite clear. In summary, it should be noted that among patients with suitable anatomy there are no absolute contraindications for endovascular AAA repair. Nonetheless the need for lifelong surveillance, secondary interventions, the risk of late device failure, and the absence of long-term data in large numbers of patients, has tempered our enthusiasm for this approach in the young patient with a long life expectancy who is otherwise at low risk for standard surgery. For the time being, the watchword associated with the implantation of an endovascular device in the low-risk patient, particularly those under 70 years of age, must remain caveat emptor. In contrast, for the patient at increased risk for standard surgery, the positive impact of endovascular repair is clear and unequivocal.