The win ratio for evaluating edoxaban vs dalteparin for cancer-associated venous thromboembolism: an analysis of the randomized Hokusai Venous Thromboembolism Cancer trial

Background The Hokusai Venous Thromboembolism (VTE) Cancer trial demonstrated that edoxaban was noninferior to dalteparin for the treatment of cancer-associated venous VTE. Objectives We reanalyzed the trial using the win ratio, an approach that evaluates a composite of outcomes in a hierarchical order. Methods Forty-nine thrombosis experts ranked 10 outcomes in order of clinical importance from all-cause death (most important) to clinically relevant nonmajor bleeding (least important). We performed unmatched pairwise comparisons between participants on edoxaban and those on dalteparin at 6- and 12-month follow-up. Within each pair, edoxaban was assigned a win, loss, or tie according to the hierarchy of outcomes. We calculated the win ratio (total wins divided by total losses among edoxaban patients), with more wins than losses indicating the benefit of edoxaban, and the win difference (total wins minus total losses). Results Among 273 528 pairs (522 × 524 participants), edoxaban was associated with a win in 34.9%, a loss in 38.5%, and a tie in 26.6%. The win ratio was 0.91 (95% CI, 0.76-1.08), with a win difference of −3.55% (95% CI, −9.9% to 2.9%) at 12 months. The win ratio remained unchanged at 6 months (0.91; 95% CI, 0.75-1.11). The findings were consistent with a hierarchy of only death, recurrent VTE, and major bleeding (win ratio, 0.92; 95% CI, 0.76-1.11), or when replacing all-cause death with VTE-related death or fatal bleeding (win ratio, 0.83; 95% CI, 0.65-1.06). Conclusion We observed no significant difference between edoxaban and dalteparin for the treatment of cancer-associated VTE when using the win ratio approach with a hierarchy of 10 prioritized outcomes.