Application of the win ratio method to clinical trials in cerebrovascular disease

Abstract

Objective  To introduce the concepts and principles of the win ratio method and to analyze it in the context of a case study of a clinical trial in cerebrovascular disease. Methods  Based on the study of clopidogrel with aspirin in high risk patients with acute non disabling cerebrovascular events 2, and key prognostic factors, the outcome events were defined sequentially as ① time to death within 90 d, ② time to recurrence of ischemic stroke within 90 d, ③ time to moderate-to-severe hemorrhage within 90 d. Using clopidogrel combined with aspirin as the reference group, the winning ratio (R_w) of ticagrelor combined with aspirin was analyzed by the win ratio method, and the 95% confidence interval (CI) of R_w was estimated by the Bootstrap method and compared with the hazard ratio (HR) calculated by the competing risk model. Results  When only fatal events were considered, the win ratio method suggested that the ticagrelor group was significantly better than the clopidogrel group, R_w=2.00 (95%CI:1.52-2.47), and after stepwise inclusion of  ischemic stroke and moderate-to-severe hemorrhage recurrence, the win ratio method yielded a value of 1.29(95% CI:1.25-1.57), and the HR value from Fine and Gray competing risk regression was 0.78 (95% CI:0.65-0.95), both of which indicated that the efficacy of the ticagrelor group was superior to that of the clopidogrel group. Conclusion  The win ratio method can be used to analyze clinical trials with composite endpoints after prioritizing multiple outcome variables, showing the advantages of win ratio and its promising application in cerebrovascular disease research.