External validation of AF-BLEED for predicting major bleeding and for tailoring NOAC dose in AF patients: A post hoc...

AF-BLEED, a simple bleeding risk classifier, was found to predict major bleeding (MB) in patients with atrial fibrillation (AF) and identify AF patients at high risk of MB who might potentially benefit from a lower direct oral anticoagulant dose. This post hoc study aimed to externally validate these findings in the ENGAGE AF-TIMI 48 (Effective aNticoaGulation with factor Xa next Generation in Atrial Fibrillation-Thrombolysis in Myocardial Infarction study 48) trial.

The ENGAGE AF-TIMI 48 trial randomized AF patients to higher-dose edoxaban regimen (HDER 60/30 mg) versus lower-dose edoxaban regimen (LDER 30/15 mg), with prespecified dose reduction criteria. AF-BLEED was calculated in the modified intention-to-treat cohort (n = 21,026 patients) used for primary outcome analysis. Annualized event rates and hazard ratios (HRs) were obtained for the primary composite outcome (PCO) and its single components (MB, ischemic stroke/systemic embolism and death) to compare LDER 30 mg with HDER 60 mg in both AF-BLEED classes.

AF-BLEED classified 2882 patients (13.7 %) as high-risk, characterized by a two- to three-fold higher MB risk than AF-BLEED classified low-risk patients. AF-BLEED classified high-risk patients randomized to LDER 30 mg demonstrated a 3.3 % reduction in MB at the cost of a 0.5 % increase in ischemic stroke/systemic embolism. LDER 30 mg resulted in a 3.1 % reduction of PCO compared to HDER 60 mg (HR of 0.81; 95%CI 0.65-1.01). Additional to existing dose reduction criteria, another 6 % of patients could potentially benefit of this dose adjustment strategy.

AF-BLEED could identify AF patients to be at high risk of major bleeding. Our findings support the hypothesis that LDER 30 mg might provide a reasonable option in AF patients with legitimate bleeding concerns.