Evaluation of the updated ABC-AF-bleeding score 2.0 in patients with atrial fibrillation treated with a direct oral anticoagulant or warfarin

BACKGROUND: Oral anticoagulation (OAC) reduces stroke in patients with atrial fibrillation (AF), but increases bleeding. OBJECTIVES: This study aimed to evaluate an updated version of the Age, Biomarkers, and Clinical history of bleeding in AF (ABC-AF)-bleeding score (2.0) including consideration of OAC type (direct oral anticoagulant [DOAC] or warfarin) and compare its performance with other bleeding risk scores in 25 962 patients from the COMBINE AF cohort. METHODS: The COMBINE AF biomarker cohort contains individual participant data from patients with AF enrolled in 3 pivotal randomized trials comparing DOACs with warfarin. The biomarkers in the ABC-AF-bleeding score (growth differentiation factor 15, hemoglobin, and troponin-T) were analyzed in baseline samples. The biomarker-based ABC-AF-bleeding score was updated (version 2.0) by incorporating OAC type into the model (DOAC or warfarin). Discrimination was assessed by Harrell C-index and compared with clinically based bleeding scores; HAS-BLED (Hypertension, Abnormal renal/liver function, Stroke, Bleeding history or predisposition, Labile INR, Elderly, Drugs/alcohol), DOAC, and ORBIT (Older age, Reduced haemoglobin/haematocrit or history of anaemia, Bleeding history, Insufficient renal function, Treatment with antiplatelet agents). RESULTS: During follow-up, 1321 patients (5.1%) had an International Society on Thrombosis and Haemostasis major bleeding event, including 480 gastrointestinal, and 248 intracranial hemorrhages. The ABC-AF-bleeding 2.0 risk score showed better discrimination and calibration than the original version and provided superior discrimination than clinical risk scores for all outcomes. The ABC-AF-bleeding score 2.0 C-indices for major bleeding were 0.69 (95% CI, 0.68-0.71); gastrointestinal bleeding, 0.72 (95% CI, 0.69-0.74); and intracranial bleeding, 0.66 (95% CI, 0.63-0.70). The ABC-AF-bleeding score 2.0 also provided consistent superior discrimination in clinically relevant subgroups. CONCLUSION: The updated ABC-AF-bleeding score 2.0 provided better discrimination and calibration for the risk of major bleeding than clinical risk scores, which was consistent across multiple subgroups. These findings support the utility of the ABC-AF-bleeding score for advancing precision medicine in AF.