Long-term anticoagulation is associated with type II endoleaks and failure of sac regression after endovascular aneurysm repair

Within the context of endovascular aneurysm repair (EVAR), the role of anticoagulation therapy on endoleak development and subsequent reintervention is unclear with conflicting data in the literature. The hypothesis of this study is that long-term anticoagulation is associated with persistent type II endoleaks and failure of sac regression in patients undergoing endoluminal repair of intact infrarenal aortic aneurysm.

Retrospective cohort abstracted from the Vascular Quality Initiative index hospitalization and long-term follow-up datasets for EVAR (2003-2017) were included in the analysis. Patients not taking aspirin preoperatively and postoperatively were excluded. Patients taking anticoagulation and aspirin concomitantly (treatment) after the index procedure were compared against patients taking aspirin alone (control). Anticoagulation included warfarin and novel oral anticoagulants, including factor Xa inhibitors and direct thrombin inhibitors. One-to-one greedy matching using propensity scores was implemented to match patients. The primary end points were failure of aneurysm sac regression, sac expansion, risk of endoleak, and reintervention rate for endoleak at follow-up. Sac regression was defined as a decrease of at least 5 mm and sac expansion was defined as an increase of at least 5 mm.

There were 9004 patients who received ASA alone and 332 patients who received ASA and anticoagulation. Propensity scores were used to create 301 matching pairs to account for differences in baseline characteristics and comorbidities, including but not limited to age, sex, smoking, coronary artery disease, heart failure, and chronic kidney disease between the treatment and control groups. After adjusting for covariables anticoagulation use was independently associated with a significantly decreased abdominal aortic aneurysm sac regression (41.59% vs 58.41%; P = .001), but no statistically significant difference in sac expansion with long-term anticoagulation use (9.7% vs 4.9%; P = .056). There was increased risk of type II endoleaks (11.96% vs 6.31%; P = .023; relative risk, 1.89; 95% confidence interval, 1.11-3.23; P = .016), but no significant differences in type I, III, or indeterminate endoleaks. There was no statistical difference in 2-year reintervention rates (4.32% vs 2.66%; hazard ratio, 1.43; 95% confidence interval, 0.55-3.77; P = .461). There were no differences in any primary outcome between warfarin and novel oral anticoagulants.

These data demonstrate that long-term aspirin plus anticoagulation use is associated with a lack of aortic sac reduction and persistent type II endoleak, but not an increased risk for subsequent reintervention. Because prior studies have demonstrated that sac regression is a correlate of survival, these findings associating regression failure suggest a potential therapeutic failure for patients undergoing EVAR who also require long-term anticoagulation therapy. Although not a contraindication, long-term anticoagulation should be considered when counseling patients with a surgical indication aortic aneurysm.