Predicting individualized treatment effects of prasugrel vs ticagrelor for acute coronary syndrome: a post hoc analysis of a randomized clinical trial

Importance Among patients with acute coronary syndrome (ACS), prasugrel has shown superior efficacy as a P2Y12 receptor inhibitor in dual antiplatelet therapy compared to ticagrelor. However, it is not well understood, which patients benefit most from prasugrel without increased risk of bleeding. Objective To identify patients that benefit most from prasugrel compared to ticagrelor. Design, Setting, and Participants: We performed a secondary analysis of the ISAR-REACT 5 trial, a randomized controlled trial that enrolled 4,018 patients with ACS between 2013 and 2018. We used a Cox proportional hazards model, incorporating interaction terms between treatment assignment and baseline covariates and validated with 10-fold cross-validation, to predict individualized treatment effects (ITEs) for prasugrel versus ticagrelor with respect to efficacy. To ensure that enhanced efficacy did not come at the expense of safety, we also assessed the association between predicted ITEs and major bleeding. Follow-up was 12 months. Exposures Prasugrel vs ticagrelor. Main Outcomes and Measures: Primary end point was efficacy defined as a composite outcome of all-cause mortality, myocardial infarction, or stroke. Major bleeding as safety end point. Results Patients were stratified based on their predicted ITEs, i.e., their predicted benefit from prasugrel compared to ticagrelor. Hazard ratios (HRs) for the primary end point were 0.58 (95% confidence interval [CI], 0.41-0.81) in the third of patients predicted to derive most benefit from prasugrel, 0.85 (95% CI, 0.52-1.36) in the third of patients predicted to derive intermediate benefit from prasugrel, and 0.90 (95% CI, 0.62-1.31) in the third of patients predicted to derive least benefit from prasugrel. Corresponding HRs for the safety end point were 0.70 (95% CI, 0.43-1.13), 1.30 (95% CI, 0.71-2.40), and 0.65 (95% CI, 0.40-1.06), respectively, indicating that the enhanced efficacy with prasugrel was not accompanied by an increased risk of major bleeding. Conclusions and Relevance Our findings suggest that ITE estimation can effectively identify ACS patients who are most likely to benefit from prasugrel—with improved efficacy and without an elevated bleeding risk—thus supporting a personalized approach to antiplatelet therapy in ACS management.