Challenges in comparing the non-vitamin K antagonist oral anticoagulants for atrial fibrillation-related stroke prevention

The aim of this review is to provide context for meta-analyses interpreting data from phase III stroke prevention studies of non-vitamin K antagonist oral anticoagulants (NOACs) in patients with non-valvular atrial fibrillation (NVAF). Differences between the four phase III NOAC stroke prevention studies in patients with NVAF (ROCKET AF, ARISTOTLE, RE-LY, and ENGAGE AF), their potential impact on outcomes, and inter-study differences were evaluated, as well as the potential role of real-world evidence in evaluating NOACs in this setting. Study design differences included blinding strategy, dose-reduction options, and transition from blinded treatment to standard of care. There were small but relevant variations in the definition of AF used (RE-LY used the least precise definition); patient risk profiles (ROCKET AF patients had the highest risk); the primary safety outcome (a composite of major bleeding and clinically relevant non-major bleeding events in ROCKET AF vs. major bleeding in the others); and the definitions of stroke, major bleeding, and clinically relevant non-major bleeding events. In real-world studies, methodological variations and biases are amplified, making cross-study comparisons and meta-analyses problematic. Because of these methodological differences, meta-analyses of phase III studies need to be robust, and if outcomes of the reference (warfarin-treated) arms differ significantly, the basis of the meta-analysis is not strong. These key issues need to be taken into consideration for direct comparisons across studies, and for the interpretation of meta-analytic data.