Bleeding risk comparison between direct oral anticoagulants at doses approved for atrial fibrillation and aspirin: systematic review, meta-analysis and meta-regression

Abstract

Background: A considerable proportion of patients with atrial fibrillation (AF) are still treated with aspirin despite current guidelines due to presumed favorable safety. Aim: We performed a systematic review and meta-analysis of bleeding outcomes in randomized controlled trials (RCTs) comparing direct oral anticoagulants (DOACs) at doses approved for AF vs. aspirin. Methods: We searched PubMed and Scopus for phase-III RCTs of DOACs at AF-approved doses vs. aspirin. Outcomes assessed were major-, intracranial-, gastrointestinal-, clinically-relevant-non-major- and fatal bleeding. We performed two subgroup analyses: one per patient population i.e. those at high risk of arterial or venous thromboembolism, and one per DOAC. We also performed a meta-regression to assess the association with patient age. Results: In 4 eligible trials (20,440 patients) comparing DOACs vs. aspirin, the ORs were: 1.52 (95%CI: 0.91–2.53) for major bleeding in patients at high risk of arterial thromboembolism and 1.55 (95%CI:0.99-2.45, relative-risk-increase:55%, absolute-risk-increase:0.6%, number-needed-to-harm:170) in the overall analysis; 1.39 (95%CI:0.62–3.14) for intracranial bleeding in patients at high risk of arterial thromboembolism which was similar for the overall analysis; 1.27 (95%CI: 0.84–1.92) for gastrointestinal bleeding in patients at high risk of arterial thromboembolism and 1.26 (95%CI:0.86-1.85) in the overall analysis. Patient age was not a predictor of the magnitude of ORs for all bleeding outcomes. Conclusion: The present meta-analysis does not support the use of aspirin over DOACs in AF. Accordingly, the level of evidence of the related recommendations should be upgraded, which in turn may reduce further the proportion of AF patients treated with antiplatelets. © 2020 European Federation of Internal Medicine