Surveillance after endovascular abdominal aortic aneurysm repair (EVAR) is considered mandatory but there is wide heterogeneity in surveillance strategies among centres. The purpose of this study was to compare two different follow-up schemes with the aim of finding optimal, but still safe timing of patient categorization for infrequent long-term surveillance.

Patients were treated electively with standard EVAR in one academic centre during 2005 - 2015 with three different devices (Zenith by Cook, Excluder by Gore and Endurant by Medtronic). The data were reviewed retrospectively and collected from case histories and surveillance images up to the end of 2020. Patients' CTAs at 30-day and at two years were analysed for freedom from any endoleak and, additionally, at two-years for aneurysm shrinkage ≥ 5mm. The reinterventions during the surveillance were collected. We then analysed the sensitivity and specificity of both 30-day and 2-year CTA findings to predict reintervention requiring complication during the long-term follow-up.

Altogether 435 patients we treated for AAA with EVAR during the study period. 30-day CTA was available for 424 patients (97.5%) and two-year CTA for 344 patients (79.1%). In 30-day CTA an endoleak was found in 25.9% (n=85) of the cases (type I n=10, type II n=75). In two-year CTA an endoleak was found in 20.9% (n=72) cases (type I n=9, type II n=63). At least 5mm shrinkage was seen in 192 cases (55.8%) at two-years (Figure 1). The sensitivity of 30-day CTA to identify reintervention requiring complication (based on endoleak finding) during the follow-up was 44.6%, the specificity 80.4% and negative predictive value (NPV) 81.1%. Corresponding figures at two-years were 57.3%, 88.9% and 88.2%. When combining both the freedom from any endoleak and sac shrinkage ≥5mm at two-year CTA, the figures were 90.6%, 65.6% and 96.9%, respectively. Further, there were device specific differences. The type II endoleak was most common at 30-days in patients treated with Excluder stent graft (35.3%) and the lowest with Endurant (4%) (p<0.001). Sac shrinkage ≥5mm at two years was seen most common in the patients treated with Zenith stent graft (59.7%) and most seldom with Excluder (43.9%) (p=0.04). In long-term surveillance 82.7% of the patients treated with Endurant stent graft required no re-interventions compared to 76.2% of the patients treated with Zenith stent graft (p=0.25).

Patient categorization into infrequent surveillance at two years seems to be safer compared to categorization at 30-days. According to our study the categorization at two years based on both endoleak and sac shrinkage is more robust with high sensitivity and specificity thus supporting the infrequent surveillance. Further, maybe in cases with uncomplicated early surveillance and uncomplicated CTA with shrinking AAA at two-years further surveillance could be even abstained as only small number (3.2%, n=6) of the patients require reinterventions in long-term surveillance. In long-term surveillance the re-intervention rates are similar between devices despite the difference in endoleak and sac shrinkage rate.