Safety and effectiveness of apixaban in comparison to warfarin in patients with nonvalvular atrial fibrillation: a propensity-matched analysis from Japanese administrative claims data

Objective: To investigate the risk of bleeding events and stroke/systemic embolism (SE) among Japanese patients with nonvalvular atrial fibrillation (NVAF), focusing on the initial dosage of apixaban and patient age. Methods: This retrospective cohort study used de-identified electronic health records based claims data from 314 acute-care hospitals in Japan. NVAF patients newly initiated on warfarin or apixaban, with no prescription during the 180-day blanking period, were eligible. Patients were allocated to receive warfarin or 5 or 2.5 mg twice daily (BID) apixaban. One-to-one propensity-score matching was used to balance patient characteristics between apixaban and warfarin. Results: Among 31,006 eligible patients, 11,972 matched pairs were identified for apixaban versus warfarin. Mean age ± standard deviation was 77.7 ± 10.0 and 77.6 ± 10.0 years and CHADS₂ score was 2.2 ± 1.4 and 2.2 ± 1.4 for warfarin and apixaban, respectively. In the apixaban cohort, 39.4% of patients received the standard dose (5 mg BID) and 60.6% received the reduced dose (2.5 mg BID). Incidence rates (events per 100 person-years) of major bleeding, any bleeding and stroke/SE were 3.7, 23.1 and 3.1, and 2.5, 18.6 and 2.0 for warfarin and apixaban cohorts, respectively. Apixaban was associated with a significantly lower risk of any bleeding (hazard ratio [HR] 0.809, 95% confidence interval [CI] 0.731–0.895; p <.001), major bleeding (HR 0.655, 95% CI 0.505–0.849; p =.001) and stroke/SE (HR 0.637, 95% CI 0.478–0.850; p =.002). Conclusions: Our observational data from clinical practice broadly confirms the safety and efficacy results of pivotal randomized controlled trials of apixaban for stroke prevention among NVAF patients.