Laboratory Test Predictors for Major Bleeding in Elderly (≥80 Years) Patients With Nonvalvular Atrial Fibrillation Treated With Edoxaban 15 mg: Sub‐Analysis of the ELDERCARE‐AF Trial

Background We investigated the predictors related to major bleeding events during treatment with edoxaban 15 mg in patients aged ≥80 years with nonvalvular atrial fibrillation and high bleeding risk, for whom standard oral anticoagulants are inappropriate, focusing on standard laboratory tests related to bleeding. Methods and Results This was a prespecified subanalysis of the on-treatment analysis set of the ELDERCARE-AF (Edoxaban Low-Dose for Elder Care Atrial Fibrillation Patients) trial. Major bleeding was the primary safety end point. The event rates were calculated according to prespecified characteristics at baseline. A total of 984 Japanese patients were randomly assigned to edoxaban 15 mg or placebo (n=492, each). During the study period, 20 and 11 major bleeding events occurred in the edoxaban and placebo groups, respectively. The adjusted analysis revealed that hemoglobin <12.3 g/dL (adjusted hazard ratio [aHR], 3.57 [95% CI, 1.10-11.55]) and prothrombin time ≥12.7 seconds; (aHR, 2.89 [95% CI, 1.05-8.02]) independently predicted major bleeding, while creatinine clearance <30 mL/min showed a tendency towards an increase in major bleeding (aHR, 2.68; 95% CI, 0.96-7.46). In patients treated with edoxaban lacking these 3 risk factors, no major bleeding occurred; major bleeding event rates increased with each risk factor. Patients with 3 risk factors were significantly more likely to have a major bleeding event at 11.05%/year (HR, 7.15 [95% CI, 1.92-26.71]). Conclusions In elderly patients with nonvalvular atrial fibrillation with high bleeding risk, baseline hemoglobin <12.3 g/dL, prothrombin time ≥12.7 seconds, and creatinine clearance <30 mL/min may predict major bleeding during treatment with edoxaban 15 mg. Registration URL: ELDERCARE-AF https://www.clinicaltrials.gov; Unique number: NCT02801669.