Any nonadherence to instructions for use predicts graft-related adverse events in patients undergoing elective endovascular aneurysm repair.

BACKGROUND: A variety of devices exist for endovascular aneurysm repair (EVAR). Device-specific instructions for use (IFU) detail anatomic constraints to application and deployment of devices and are developed from rigorous bench testing. Nonadherence to IFU occurs frequently to avoid open surgery. The purpose of this study was to determine if IFU violations are associated with increased risk of graft-related adverse events (GRAEs) during follow-up. METHODS: This multicenter retrospective observational study included patients undergoing elective endovascular repair for abdominal aneurysmal disease with three different devices. Demographics, anatomic data, and follow-up data were collected on patients from 2005 to 2014. IFU violations were device specific and included neck diameter, length, and angulation and iliac artery diameter and length. GRAEs included a composite outcome of reintervention, migration, endoleak (type II excluded), rupture, limb occlusion, sac growth, and aneurysm-related mortality during the follow-up period. Kaplan-Meier survival and Cox proportional hazards modeling were performed. Any IFU violations as well as neck-specific IFU violations were analyzed. RESULTS: In 461 patients undergoing EVAR, 43.8% had at least one IFU violation. Patients with IFU violations were more likely to have peripheral vascular disease (12.4% vs 7.3%) and were less likely to be male (78.7% vs 90.3%). The most frequent IFU violations included diameter deviations of the neck (15.2%) and of the iliac artery (21.4%). Overall, the GRAE rate was 12.8%. Median follow-up time was 1.9 and 2.1 years for patients with and without an IFU violation, respectively. Kaplan-Meier survival revealed a significant association between the presence of an IFU violation and GRAEs (log-rank, P = .031). When adjusted for clinical variable through Cox hazard modeling, the association remained significant (hazard ratio 1.8; 95% confidence interval, 1.05-3.1). When neck-specific violations were considered independently, Kaplan-Meier survival (log-rank, P = .003) and Cox modeling (hazard ratio, 2.2; 95% confidence interval, 1.2-4.0) revealed a significant association between neck-specific IFU violation and GRAEs. CONCLUSIONS: A total of 43.8% of patients undergoing EVAR had a device-specific IFU violation, indicating that implanters are pushing the boundaries of device capabilities. Our study identified that any IFU violation was significantly associated with GRAEs over time. Caution should be applied to patients being considered for EVAR when IFU deviations exist.