Predictive factors for bleeding during treatment with rivaroxaban and warfarin in Japanese patients with atrial fibrillation - Subgroup analysis of J-ROCKET AF

Background: Results from the J-ROCKET AF study revealed that rivaroxaban was non-inferior to warfarin with respect to the principal safety outcomes in patients with non-valvular atrial fibrillation. This subgroup analysis evaluated whether non-major clinically relevant bleeding (NMCRB) could be a predictive factor for major bleeding (MB). Other predictive factors for MB were also obtained in both rivaroxaban and warfarin treatment groups. Methods: The temporal incidence of MB was compared between the rivaroxaban and warfarin treatment groups. Assessment was made whether MB events were often preceded by NMCRB. Univariate and multivariate analyses were carried out to identify any independent predictive factors for MB in both treatment groups. Results: The incidences of MB and NMCRB were 18.04% (138/639 patients) in the rivaroxaban arm, and 16.42% in the warfarin arm (124/639 patients). NMCRB preceded MB in only four patients in each treatment group (rivaroxaban: 4/117 and warfarin: 4/98). Multivariate analysis identified predictive factors for bleeding events: anemia with warfarin treatment and concomitant use of antiplatelet agents with rivaroxaban treatment. Conclusions: Results from this subgroup analysis, particularly the fact that there was no repeated or sequential pattern between NMCRB and MB occurrences in both treatment groups, suggests that NMCRB might not be a predictive factor for MB. On the contrary, anemia and concomitant use of antiplatelet therapy were likely predictive factors for bleeding with warfarin and rivaroxaban treatment, respectively. (C) 2016 Japanese College of Cardiology. Published by Elsevier Ltd. All rights reserved.