Patients With Atrial Fibrillation Taking Nonsteroidal Anti-Inflammatory Drugs and Oral Anticoagulants in the ARISTOTLE Trial

Background: The use of nonsteroidal anti-inflammatory drugs (NSAIDs) with oral anticoagulants has been associated with an increased risk of bleeding. We investigated the risk of bleeding and major cardiovascular outcomes in patients with atrial fibrillation taking NSAIDs and apixaban or warfarin. Methods: The ARISTOTLE trial (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation; n=18 201) compared apixaban with warfarin in patients with atrial fibrillation at an increased risk of stroke. Patients in ARISTOTLE without severe renal (creatine clearance <= 30 mL/min) or liver disease were included in this analysis (n=17 423). NSAID use at baseline, NSAID use during the trial (incident NSAID use), and never users were described. The primary outcome was major bleeding. Secondary outcomes included clinically relevant nonmajor bleeding, gastrointestinal bleeding, heart failure hospitalization, stroke or systemic embolism, and all-cause mortality. NSAID use during the trial, and the interaction between randomized treatment, was analyzed using time-dependent Cox proportional hazards models. Results: Those with baseline NSAID use (n=832 [4.8%]), incident NSAID use (n=2185 [13.2%]), and never users were similar in median age (age [25th, 75th]; 70 [64, 77] versus 70 [63, 75] versus 70 [62, 76]). Those with NSAID use at baseline and incident NSAID use were more likely to have a history of bleeding than never users (24.5% versus 21.0% versus 15.6%, respectively). During a median follow-up (25th, 75th) of 1.8 (1.4, 2.3) years and when excluding those taking NSAID at baseline, we found that incident NSAID use was associated with an increased risk of major bleeding (hazard ratio [HR], 1.61 [95% CI, 1.11-2.33]) and clinically relevant nonmajor bleeding (HR, 1.70 [95% CI, 1.16-2.48]), but not gastrointestinal bleeding. No significant interaction was observed between NSAID use and randomized treatment for any outcome. Conclusions: A substantial number of patients in the ARISTOTLE trial took NSAIDs. Incident NSAID use was associated with major and clinically relevant nonmajor bleeding, but not with gastrointestinal bleeding. The safety and efficacy of apixaban versus warfarin appeared not significantly to be altered by NSAID use. This study warrants more investigation of the effect of NSAIDs on the outcomes of patients treated with apixaban.