Early Sac Behavior Can Predict Endoleak Resolution in Patients With Early Type II Endoleak

Endoleak is a known complication of endovascular aortic repair (EVAR). The purpose of this study was to investigate the natural behavior of endoleaks based on initial aneurysm sac behavior. We analyzed 1631 patients who underwent EVAR for abdominal aortic aneurysm at a single center between 2001 and 2020. Patients were stratified into no leak or leak groups and then further grouped by sac behavior: expansion, stabilization, or regression. Endoleaks were confirmed on computed tomography scan, magnetic resonance imaging, or angiography. Statistical analysis was performed using χ2 and odds ratios. Among 1631 patients undergoing EVAR, 296 (18.2%) had type II endoleak at first follow-up. Of these 296, 125 (42.2%) had sac regression, 121 (40.9%) sac expansion, and 50 (16.9%) stable sacs. At a mean follow-up of 52.9 ± 46.9 months, regression group demonstrated 50 (54.9%) endoleak resolution, 27 (29.6%) showed persistent type II, 2 (2.2%) developed late type I, and 12 (13.2%) had an unspecified leak. The expansion group (mean follow-up of 58.7 ± 42.3 months) had 24 (19.8%) endoleak resolution, 64 (52.9%) with persistent type II endoleak, 10 (8.3%) developed late type I endoleak, and 23 (19.0%) had unspecified endoleaks. The stable group (mean follow-up of 27.9 ± 34.0 months) showed 12 (24%) with leak resolution, 32 (64%) persistent type II leak, 2 (4%) late type I, and 4 (8%) with a late unspecified leak. Regression was associated with fewer developments of late endoleak than nonregression (45% vs 76%; P < .001). Nonregression was 8.33-fold more likely to develop late type I endoleak, 4.94-fold more likely to have persistent type II endoleak, and 3.13-fold more likely to develop late unspecified endoleak compared with regression (Table). Regression was 4.59-fold more likely to experience spontaneous endoleak resolution compared with nonregression (Table). EVAR patients with early type II endoleak and sac regression were less likely to develop late type I or unspecified endoleaks and were more likely to undergo spontaneous endoleak resolution. This study confirms that sac regression is the most important prognostic metric for successful EVAR even in patients with early type II endoleak.TableOdds ratio analysis of the presence of late type I to III endoleak across sac behavior groupsTII leak, regressionTII leak, stabilizationTII leak, expansionOdds ratio (95% CI)P valueType I late endoleak22108.33 (1.76-39.54).008Type II late endoleak2732644.94 (2.70-9.04)<.0001Unspecified late endoleak124233.13 (1.40-6.98).006Resolution of endoleak5012240.22 (.13-.38)<.0001CI, Confidence interval; TII, type II.Boldface entries indicate statistical significance. Open table in a new tab