The tradeoff in safety versus efficacy in substituting a non–vitamin K antagonist oral anticoagulant for a vitamin K antagonist (VKA) in the stented atrial fibrillation patient has not been quantitatively evaluated.

Based on summary data from the PIONEER AF-PCI and RE-DUAL PCI trials, 4 antithrombotic regimens were compared with VKA-based triple therapy: (1) rivaroxaban (riva) 15 mg daily + P2Y₁₂ inhibitor, (2) riva 2.5 mg twice daily + P2Y₁₂ inhibitor + aspirin, (3) dabigatran (dabi) 110 mg twice daily + P2Y₁₂ inhibitor, and (4) dabi 150 mg twice daily + P2Y₁₂ inhibitor. A bivariate model with a noninferiority margin of 1.38 was used to simultaneously assess safety and efficacy. The safety end point was major or clinically relevant nonmajor bleeding by International Society on Thrombosis and Haemostasis definitions. The efficacy end point was a thromboembolic event (myocardial infarction, stroke, or systemic embolism), death, or urgent revascularization. The bivariate outcome, a measure of risk difference in the net clinical outcome, was compared between antithrombotic regimens.

All 4 non–vitamin K antagonist oral anticoagulant regimens were superior in bleeding and noninferior in efficacy compared with triple therapy with VKA. Riva 15 mg daily and 2.5 mg twice daily were associated with bivariate combined risk reductions of 5.6% (2.3%-8.8%) and 5.5% (2.1%-8.7%), respectively, and dabi 110 mg twice daily and 150 mg twice daily reduced the bivariate risk by 3.8% (0.5%-7.0%) and 6.3% (2.4%-9.8%), respectively.

A bivariate analysis that simultaneously characterizes both risk and benefit demonstrates that riva- and dabi-based regimens were both favorable over VKA plus dual antiplatelet therapy among patients with atrial fibrillation undergoing PCI.