Edoxaban is an oral, reversible, direct factor Xa inhibitor in phase III clinical development for the prevention of stroke in atrial fibrillation (AF). A phase II study was undertaken to evaluate the safety and efficacy of edoxaban in Asian patients with non-valvular AF with CHADS2 score ≥1. In a multicentre, active-controlled, double-blind edoxaban and open-label warfarin, parallel-group study, a total of 235 patients from four Asian countries were randomly assigned to edoxaban 30 mg qd, 60 mg qd or warfarin dose adjusted to international normalised ratio of 2-3 for three months. The primary endpoint was the incidence of centrally adjudicated all bleeding events (major, clinically relevant non-major and minor). Secondary endpoints included thromboembolic events, biomarkers of thrombus formation and all adverse events (AEs). The incidence of all bleeding events (95% CI) was 20.3% (12.9, 30.4) for edoxaban 30 mg, 23.8% (15.8, 34.1) for edoxaban 60 mg, and 29.3% (20.2, 40.4) for warfarin. A subgroup analysis suggested low body weight (≤60 kg) may affect the incidence of bleeding events with edoxaban. The incidence of study drug-related AEs was 22% for edoxaban 30 mg, 29% for edoxaban 60 mg and 33% for warfarin. No thromboembolic events occurred in any treatment group. In conclusion, this phase II study found a trend for a reduction in the incidence of all bleeding events in Asian AF patients with edoxaban 30 mg and 60 mg compared with warfarin. Adverse events were similar between the edoxaban 60-mg and warfarin groups and were lower with the edoxaban 30-mg group.