Discussion

Dr Charles Sternbergh, III (New Orleans, LA). We have a great deal of good data on the natural history of small AAA. In the ADAM VA cooperative trial, the rupture rate of AAA between 4.0 and 5.4 cm in the observational group was 0.6% per year. With an operative mortality of 2.1% in the open group, there were no differences in long-term patient outcome between observation and open surgery after 4.9 years. Importantly, the majority of observational patients with AAA of less than 5 cm did not go on to require repair in this time interval. Can endografting improve on this rather benign natural history of small AAA? Based on the VA cooperative data, I’m skeptical. But there is no level I evidence directly addressing this controversy. To help definitively answer this question, Medtronic has recently begun the PIVOTAL trial, which is designed to enroll 1700 patients with AAA of less than 5 cm, randomized to AneuRx treatment or observation.

Dr Zarins and colleagues have demonstrated that when compared with small AAA, patients with large AAA have much poorer long-term outcomes after EVAR. While this is not new information, it bears repeating. It is well known that patients with larger AAA have, on average, more unfavorable aortic neck anatomy. However, when groups are controlled for variable anatomy and other cohort differences, AAA size remains as independent risk factor for poorer outcome. This was first demonstrated from the EUROSTAR database and has been confirmed today.

How can this be explained? My hypothesis is a simple one. If there is late endograft failure in a large AAA, the risk of subsequent rupture is significant. However, if there is endograft failure in a small AAA that already has a negligible risk of rupture, the incidence of aneurysm-related problems is low.

Thus, knowing why endografts fail is crucial, especially if they are being used to treat patients with a high risk of rupture. In The Cleveland Clinic report of disparate outcome of small vs large AAAs, there was a fourfold increase in type I endoleaks and a threefold increase in migration in AAA greater than 5.5 cm (Ouriel et al, J Vasc Surg 2003:37:1206). The EUROSTAR report demonstrated similar findings (Peppelenbosch et al, J Vasc Surg 2004;39:288-97). It is perplexing, therefore, that in the current study, there were no differences in these outcome measures. This leads me to my first question:

Dr Christopher Zarins. In response to the question of whether endografting can improve on the natural history of small AAA, I would like to remind you that the UK and ADAM small aneurysm trials reported a low risk of rupture in selected, good risk patients who were closely monitored with ultrasound surveillance every 3–6 months and promptly referred for surgery if the aneurysm enlarged or became symptomatic. Despite this, 1% of the small aneurysms ruptured each year—that is 5% in 5 years. Thus ultrasound surveillance does not appear to have been a successful strategy, particularly when rupture mortality was 90% and when 75% of the surveillance patients ultimately required surgery. In contrast to surveillance of small aneurysms, endovascular repair was very successful in our experience of small aneurysms <5 cm. There were no ruptures over 5 years, only one aneurysm-related death and only two surgical conversions. Clearly the size difference in definition of small aneurysms is a critical factor in this difference and this is one of the major points of this study. Aneurysms 5.5 cm in diameter are average in size, not small.

I agree with Dr. Sternberg’s comments regarding large aneurysms. Results are not as good as for small aneurysms and the risk of late failure and rupture is higher. Since large aneurysms are more likely to rupture if they are untreated, late device failures are a greater concern in patients with large aneurysms. We did not note a relationship between endoleak and enlargement in this study, but this does not rule out this possibility. However, greater risk does not mean that patients with large aneurysms should not be treated with EVAR. Despite higher operative risk and less favorable anatomy, early results for large aneurysms are similar to smaller AAAs with differences appearing only after 4 years. The cumulative aneurysm related death rate for large aneurysms at 5 years is still only 8%, which is considerably below the expected mortality from rupture of non-treated large aneurysms.

With regard to migration, there is debate as to whether Kaplan-Meier analysis is the preferred way to report migration since migration is not a fixed endpoint, such as rupture or death, and can start and stop. Kaplan-Meier analysis will give higher estimates of migration and we have reported both Kaplan-Meier estimates as well as migration prevalence rates at specific time points in our previous migration analysis. In this study we focused on primary outcome measures and reported migration, endoleak and enlargement as prevalence rates at each annual follow up time, since all are variable end-points.

Finally, Dr. Sternbergh asked about the increase in relative risk of rupture and aneurysm related death after 4 years in large AAA patients. Further follow up is needed to determine whether this risk will continue to rise, but this finding highlights the importance of life-long image based surveillance of patients with aneurysms >6 cm following endovascular repair. At the same time, the absence of late rupture in patients with small aneurysms raises the question of whether patients with aneurysms <5 cm need less intensive follow up and imaging following successful endovascular repair. Perhaps this is where simple ultrasound surveillance will find a place. These questions deserve further study. Our findings suggest that answers to these questions may be easier to find by differentiation aneurysm size groups into small (<5 cm), medium (5-6 cm) and large (>6 cm).