Fenestrated and Branched Endovascular Aortic Repair and Mortality at Hospitals Without Investigational Device Trials.

Abstract

Importance: Fenestrated and branched endovascular aortic repairs (F/BEVAR) have been adopted by many centers. However, national trends of F/BEVAR use remain unclear, particularly at sites who perform them without an US Food and Drug Adminstration (FDA)-approved investigational device exemption (IDE).

Objective: To quantify the use of F/BEVAR in the US and to determine if mortality was different at IDE vs non-IDE sites.

Design, setting, and participants: This retrospective cohort study examined 100% fee-for-service Medicare claims data from 2016 to 2023. Participants were patients who underwent endovascular treatment of the visceral aorta incorporating 2 or more visceral artery endoprostheses. Hospitals with vs without an IDE were identified using hospitals' Employer Identification Number as a time varying exposure.

Exposure: F/BEVAR.

Main outcomes and measures: Trends in the center-level F/BEVAR case volume stratified by IDE status were assessed using cumulative incidence curves. Mortality outcomes at 30 days and 3 years were compared using Kaplan-Meier methods and Cox proportional hazards models with adjustment for baseline patient characteristics.

Results: From 2016 to 2023, 8017 patients were treated with F/BEVAR at 549 hospitals. The median (IQR) age was 75.8 (71.3-80.8) years; 5795 patients (72.3%) were male and 2222 (27.7%) female. A total of 2226 F/BEVAR (27.8%) were performed at 22 hospitals with an IDE. The number of patients treated with F/BEVAR increased from 771 in 2016 to 1251 in 2023. The median (IQR) annual case volume per hospital was significantly higher at IDE sites (22.3 [11.0-30.4] vs 1.2 [1.0-2.0] cases/y; P < .001); 18 IDE sites (90.0%) and 20 non-IDE sites (3.7%) completed 9 or more cases per year. The 30-day mortality (3.0% vs 4.9%) but not 3-year mortality (26.0% vs 27.1%) was lower for patients treated at hospitals with vs without an IDE. After risk adjustment, both 30-day (odds ratio, 0.47; 95% CI, 0.32-0.69) and midterm mortality (hazard ratio, 0.81; 95% CI, 0.69-0.95) were lower for patients treated at IDE sites.

Conclusions and relevance: The use of F/BEVAR is increasing across the United States, with the majority of cases being performed outside of IDE studies and at low-volume centers. F/BEVAR performed at non-IDE centers are associated with higher adjusted 30-day and midterm mortality. Transparent outcome reporting and identification of process measures from IDE sites may help achieve more equity in patient outcomes.