Objective: To evaluate the safety of antithrombotic drugs used in patients with both atrial fibrillation (AF) and non-end-stage chronic kidney disease (NECKD). Methods: A search was performed for studies on major bleeding outcomes in patients with concurrent AF and NECKD using Medline and Cochrane databases on 19th February, 2015. Fixed-or random-effects meta-analysis was adopted for evaluating pooled effect sizes according to whether heterogeneity existed. Results: Twelve articles were included for analysis. Three studies evaluated AF patients who took warfarin vs. placebo/antiplatelet drugs in the presence of NECKD. No significant difference in major bleeding risk was observed according to the pooled analysis using the random-effects model (RR: 1.05, 95% CI: 0.74-1.36). The risk of a composite of major bleeding outcomes was reduced by 19% in patients randomized to direct oral anticoagulants (DOACs) compared to dose-adjusted warfarin from pooled data of three randomized controlled trials with regard to AF and NECKD (RR: 0.81, 95% CI: 0.75-0.88). This superiority of DOACs to warfarin maintained until the renal function was severely impaired. Conclusions: In patients with AF and NECKD, no significant increase in the incidence of major bleeding outcomes was observed in warfarin use compared with placebo/antiplatelet drugs. DOACs reduced the risk of major bleeding by 19% compared to warfarin and further data-exploration indicated that the risk did not increase as renal function deteriorated during the renal status of mild to moderate impairment. (C) 2015 Elsevier Ltd. All rights reserved.