Comparative efficacy and safety of novel oral anticoagulants in patients with atrial fibrillation: A network meta-analysis with the adjustment for the possible bias from open label studies.

Background: This study was designed to compare efficacy and safety among novel oral anticoagulants (NOACs), which have not been directly compared in randomized control trials to date.

Method: We performed network meta-analyses of randomized control trials in preventing thromboembolic events and major bleeding in patients with atrial fibrillation. PubMed, Embase, and the Cochrane Database of Systematic Reviews for published studies and various registries of clinical trials for unpublished studies were searched for 2002-2013. All phase III randomized controlled trials (RCTs) of NOACs (apixaban, edoxaban, dabigatran, rivaroxaban), idraparinux, and ximelagatran were reviewed.

Results: A systematic literature search identified nine phase III RCTs for primary analyses. The efficacy of each NOAC was similar with respect to our primary composite endpoint following adjustment for open label designs [odds ratios (ORs) versus vitamin K antagonists: apixaban 0.79; dabigatran 150mg 0.77; edoxaban 60mg 0.87; rivaroxaban 0.86] except for dabigatran 110mg and edoxaban 30mg. Apixaban and edoxaban 30mg and 60mg had significantly fewer major bleeding events than dabigatran 150mg, ricvaroxaban, and vitamin K antagonists. All NOACs were similar in reducing secondary endpoints with the exception of dabigatran 110mg and 150mg which were associated with a significantly greater incidence of myocardial infarction compared to apixaban, edoxaban 60mg, and rivaroxaban.

Conclusions: Our indirect comparison with adjustment for study design suggests that the efficacy of the examined NOACs is similar across drugs, but that some differences in safety and risk of myocardial infarction exist, and that open label study designs appear to overestimate safety and treatment efficacy. Differences in study design should be taken into account in the interpretation of results from RCTs of NOACs.