Use of the win ratio approach to assess outcomes in the DapaTAVI trial.

Abstract

Introduction and objectives: The win ratio (WR) approach is used to assess composite endpoints in a hierarchical fashion. This novel method offers an excellent opportunity to assess the robustness of the findings yielded by landmark trials, such as the DapaTAVI trial.

Methods: We applied the WR method to evaluate the treatment effect of dapagliflozin in hierarchically ordered clinical outcomes. Several combinations of outcomes were tested, including time-to-event, binary, and continuous endpoints.

Results: The WR of the original primary endpoint was 1.36 (95%CI, 1.03-1.78; P=.028), comparable to the reciprocal of the original hazard ratio (1/HR, 1.38; 95%CI, 1.06-1.81). The win difference was 4.84% (95%CI, 0.55-9.12), confirming consistent findings in terms of absolute effect. Alternative combinations of the primary outcome with different prioritization of its components yielded similar treatment effects and statistical significance. Ignoring a time-to-event approach and including recurrent events did not substantially affect treatment efficacy and its statistical significance. In contrast, the inclusion of the total length of stay for heart failure hospitalizations in the hierarchy shifted the point estimate toward the null. Including New York Heart Association functional class improved the precision of the estimate (WR=1.31; 95%CI, 1.09-1.56; P=.003). Conversely, including quality of life through Kansas City Cardiomyopathy Questionnaire comparisons shifted the overall estimate toward the null (WR=1.10; 95%CI, 0.94-1.30; P=.236).

Conclusions: The WR approach is a solid method to assess treatment efficacy. We observed consistent findings using this approach in the DapaTAVI trial.