Bleeding events among new starters and switchers to dabigatran compared with warfarin in atrial fibrillation.

The bleeding risk among patients with atrial fibrillation is higher early after initiating therapy with vitamin K antagonists (VKAs). Evidence is limited on how prior VKA experience affects bleeding risk when initiating novel oral anticoagulant therapy. We investigated this among patients with atrial fibrillation initiating dabigatran therapy. By using nationwide Danish prescription and patient registries, we identified 11,315 first-time dabigatran users with atrial fibrillation. Warfarin controls were matched in a 2:1 ratio according to VKA experience status. The average follow-up time was 13 months. Across the 6 combinations of treatment (dabigatran 110 mg, dabigatran 150 mg, and warfarin) and VKA experience status (naive or experienced), VKA-naïve warfarin initiators had the highest rate of any bleeding event. Cox regressions adjusted for baseline characteristics showed reductions relative to this group ranging from 19% for VKA-experienced dabigatran 110 mg users (hazard ratio [HR], 0.81; 95% confidence interval [CI], 0.66-1.00) to 41% for VKA-experienced dabigatran 150 mg users (HR, 0.59; 95% CI, 0.46-0.75). Among switchers to dabigatran from warfarin, when comparing with warfarin-persisting users, the rate of any bleeding was nonsignificantly decreased for switchers to dabigatran 150 mg (HR, 0.80; 95% CI, 0.62-1.03) but not for switchers to dabigatran 110 mg (HR, 1.12; 95% CI, 0.90-1.41). Results for major bleeding were similar. Crude rates of fatal, intracranial, and gastrointestinal bleeding were low. VKA-naïve warfarin initiators had the highest overall bleeding rate. We found no evidence of marked excess of overall bleeding events when comparing dabigatran with warfarin users, irrespective of prior VKA experience.