Trends in oral anticoagulant choice for acute stroke patients with nonvalvular atrial fibrillation in Japan: The SAMURAI-NVAF Study

Background: Large clinical trials are lack of data on non-vitamin K antagonist oral anticoagulants for acute stroke patients. Aim: To evaluate the choice of oral anticoagulants at acute hospital discharge in stroke patients with nonvalvular atrial fibrillation and clarify the underlying characteristics potentially affecting that choice using the multicenter Stroke Acute Management with Urgent Risk-factor Assessment and Improvement-NVAF registry (ClinicalTrials.gov NCT01581502). Method: The study included 1192 acute ischemic stroke/transient ischemic attack patients with nonvalvular atrial fibrillation (527 women, 77·7±9·9 years old) between September 2011 and March 2014, during which three nonvitamin K antagonist oral anticoagulant oral anticoagulants were approved for clinical use. Oral anticoagulant choice at hospital discharge (median 23-day stay) was assessed. Results: Warfarin was chosen for 650 patients, dabigatran for 203, rivaroxaban for 238, and apixaban for 25. Over the three 10-month observation periods, patients taking warfarin gradually decreased to 46·5% and those taking nonvitamin K antagonist oral anticoagulants increased to 48·0%. As compared with warfarin users, patients taking nonvitamin K antagonist oral anticoagulants included more men, were younger, more frequently had small infarcts, and had lower scores for poststroke CHADS₂, CHA₂DS₂-VASc, and HAS-BLED, admission National Institutes of Health stroke scale, and discharge modified Rankin Scale. Nonvitamin K antagonist oral anticoagulants were started at a median of four-days after stroke onset without early intracranial hemorrhage. Patients starting nonvitamin K antagonist oral anticoagulants earlier had smaller infarcts and lower scores for the admission National Institutes of Health stroke scale and the discharge modified Rankin Scale than those starting later. Choice of nonvitamin K antagonist oral anticoagulants was independently associated with 20-day or shorter hospitalization (OR 2·46, 95% CI 1·87-3·24). Conclusions: Warfarin use at acute hospital discharge was still common in the initial years after approval of nonvitamin K antagonist oral anticoagulants, although nonvitamin K antagonist oral anticoagulant users increased gradually. The index stroke was milder and ischemia-risk indices were lower in nonvitamin K antagonist oral anticoagulant users than in warfarin users. Early initiation of nonvitamin K antagonist oral anticoagulants seemed safe.