Accounting for device evolution and competing risks in comparisons of TAVR and SUAVR

We read with interest the meta-analysis by Choi et al1 comparing transcatheter aortic valve replacement (TAVR) with sutureless or rapid-deployment aortic valve replacement (SUAVR). By reconstructing individual patient data from propensity score–matched cohorts, the authors provide an ambitious and informative comparison of two increasingly used valve strategies. We would like to offer several additional methodological considerations that may further refine interpretation of the findings and guide future investigations. First, the analysis aggregates data collected over a prolonged period (2007–2022), during which both TAVR and SUAVR technologies underwent substantial evolution. Within the TAVR group, outcomes differ meaningfully according to valve platform (self-expanding vs balloon-expandable) and device generation, particularly with respect to permanent pacemaker implantation and paravalvular leak. Prior studies have demonstrated higher rates of conduction disturbances with self-expanding valves2, whereas newer-generation devices show improved sealing and less paravalvular leak3. Similarly, heterogeneity exists between SUAVR systems such as PERCEVAL and INTUITY. Without stratification by valve type or generation, it remains difficult to determine whether observed outcome differences reflect procedural modality or underlying device-specific effects. Second, although propensity score matching was employed, residual confounding related to patient frailty is likely. In routine clinical practice, TAVR is frequently favored for patients with physical frailty, sarcopenia, or cognitive impairment, even when conventional surgical risk scores are comparable to those of surgical candidates. Frailty is a well-established, independent predictor of long-term mortality following aortic valve replacement4 and is often incompletely captured in administrative or registry-based datasets. Consequently, the inferior long-term survival observed in the TAVR cohort may partially reflect baseline vulnerability rather than a true disadvantage of the transcatheter approach. Third, long-term outcomes were assessed using conventional Kaplan–Meier and Cox proportional hazards methods. Given the advanced age of the study population (mean age approximately 80 years), noncardiac mortality represents a substantial competing risk. Standard survival techniques may therefore overestimate procedure-related mortality by treating competing events as censored or by focusing exclusively on all-cause death. Competing risk models, such as Fine–Gray subdistribution analyses, could provide a more nuanced estimation of valve- and procedure-specific outcomes5. Finally, although survival was reported up to 9 years, the study did not specifically evaluate valve durability, structural valve deterioration, or bioprosthetic valve failure. These endpoints are increasingly important as transcatheter valves are implanted in younger and lower-risk patients. Recent studies highlight concerns about earlier SVD in some TAVR valves compared to surgical bioprostheses6,7. Incorporation of standardized definitions for structural valve deterioration and valve failure would enhance the relevance of long-term comparisons between TAVR and SUAVR. We commend the authors for their comprehensive analysis and believe that addressing these issues in future updates or collaborative individual patient data studies would further strengthen the evidence guiding selection between TAVR and SUAVR. Ethical approval Not applicable. Consent Not applicable.