Risk of major bleeding in patients with non-valvular atrial fibrillation treated with oral anticoagulants: a systematic review of real-world observational studies.

Objective: To conduct a systematic review of real-world (RWD) studies comparing the risk of major bleeding (MB) among patients with non-valvular atrial fibrillation (NVAF) on direct oral anticoagulants (DOACs) or warfarin.
Methods: MEDLINE, Embase, NHS-EED, and EconLit were searched for RWD studies published between January 2003 and November 2016 comparing MB risk among DOACs and warfarin. Proceedings of clinical conferences from 2012 to 2016 were reviewed.
Results: A total of 4218 citations were identified, 26 of which met eligibility criteria. Most studies were retrospective analyses of administrative claims databases and patient registries (n = 23 of 26); about half were based in the United States (n = 15). Apixaban showed a significantly lower risk of MB versus warfarin in all eight included studies. MB risk was either significantly lower (n = 9 of 16) or not significantly different (n = 7 of 16) between dabigatran and warfarin; there was no significant difference between rivaroxaban and warfarin in all seven included studies. The risk was significantly lower with apixaban versus rivaroxaban (n = 7 of 7) but not significantly different from dabigatran (n = 6 of 7). MB risk was significantly lower (n = 3 of 4) or not significantly different (n = 1 of 4) with dabigatran versus rivaroxaban. No evidence was identified for edoxaban.
Conclusion: DOACs were associated with similar or lower risks of MB versus warfarin. A lower MB risk was consistently observed for apixaban, but less consistently for dabigatran; MB risk was similar between rivaroxaban and warfarin. Among DOACs, the risk of MB with apixaban was consistently lower than with rivaroxaban, but similar to dabigatran.