TAVR versus surveillance in patients with moderate aortic stenosis and heart failure: a conversion-censored analysis of the TAVR UNLOAD trial

Abstract Background The neutral TAVR UNLOAD trial randomized patients with heart failure with reduced ejection fraction and moderate aortic stenosis (AS) to transcatheter aortic valve replacement (TAVR) or clinical AS surveillance (CASS). Frequent conversion from CASS to TAVR, primarily due to AS progression, potentially reduced the benefit of TAVR vs CASS. Aims To identify conversion predictors and evaluate outcomes censoring CASS patients at conversion. Methods and results Patients eligible for balloon-expandable transfemoral TAVR were 1:1 randomized to TAVR or CASS. The hierarchical primary endpoint comprised all-cause death, disabling stroke, heart failure hospitalization/equivalents, and quality of life change. Analyses were performed to longest follow-up, censoring CASS patients at TAVR. Baseline predictors of conversion were identified using cause-specific multivariable Cox regression. Between 2017 and 2022, 178 patients (77 ± 9 years; 21% female; 56% New York Heart Association ≥ III) were randomized (TAVR n ampa#thinsp;89; CASS n ampa#thinsp;89). During a median follow-up of 690 days, 38 CASS patients (43%) underwent TAVR after a median of 366 days. After censoring, the primary endpoint favoured TAVR over CASS 50% vs 30% wins; 20% ties; win ratio: 1.68 (95% CI: 1.16–2.43); P = .006. Adjusting for age, sex, atrial fibrillation, and left ventricular ejection fraction, the mean aortic valve gradient was the only independent conversion predictor HR per mmHg: 1.09 (95% CI: 1.02–1.16); P = .015. Conclusions In patients with heart failure with reduced ejection fraction and moderate AS, TAVR conferred benefit over CASS when censoring CASS patients at conversion. Although caution is warranted given the observational nature of the comparison, these findings suggest that these patients may derive clinical benefit from TAVR, warranting confirmation in a larger trial.