Exposure-adjusted event rates for evaluating bioprosthetic valve failure after transcatheter aortic valve implantation in the presence of competing mortality.

BACKGROUND: Bioprosthetic valve failure (BVF) is a key clinical endpoint for evaluating valve durability after transcatheter aortic valve implantation (TAVI). However, in populations with substantial competing mortality or heterogeneous follow-up durations, the estimated incidence of BVF may vary according to evaluation metric. We evaluated how different analytic approaches influence estimates of the long-term BVF incidence. METHODS: We analyzed data from a multicenter registry of patients undergoing TAVI with balloon-expandable valves (SAPIEN XT or SAPIEN 3). Outcomes of interest were all-cause death and BVF, defined according to VARC-3 criteria. Incidence of BVF was evaluated using both cumulative incidence analysis with competing-risk methodology and exposure-adjusted event rates per 1000 patient-years. Valve generations (SAPIEN XT vs SAPIEN 3) were examined as a representative comparison. RESULTS: Follow-up duration differed substantially between the two valve groups, with longer exposure in the SAPIEN XT. Although the cumulative incidence of BVF was higher with SAPIEN XT, death was the predominant competing event at the 8-year horizon. After accounting for exposure time, the overall BVF rate was significantly higher with SAPIEN XT than with SAPIEN 3 (13.1 vs. 5.7 events per 1000 patient-years, p < 0.01). Most BVF events were related to nonstructural mechanism, particularly paravalvular leakage and prosthetic valve endocarditis. CONCLUSIONS: In the presence of heterogeneous follow-up durations and competing mortality, cumulative incidence-based estimates may not fully characterize valve-related outcomes. Exposure-adjusted event rates provide complementary information for comparing BVF across valve generations. These findings highlight the importance of selecting appropriate analytic approaches when interpreting BVF after TAVI.