Comparison of HAS-BLED and ORBIT bleeding risk scores in atrial fibrillation patients treated with non-vitamin K antagonist oral anticoagulants: a report from the ESC-EHRA EORP-AF General Long-Term Registry

Abstract Aims Bleeding risk assessment is recommended in guidelines for the management of atrial fibrillation (AF). The HAS-BLED score was proposed prior to non-vitamin K antagonist oral anticoagulants (NOACs) and it has been suggested that the ORBIT score may be superior in predicting bleeds in NOAC users. We aimed to compare the HAS-BLED and ORBIT scores in contemporary AF patients treated with NOACs. Methods and results We analysed patients enrolled in the ESC-EHRA EORP-AF (EURObservational Research Programme in AF) General Long-Term Registry. HAS-BLED and ORBIT scores were computed based on original schemes. The primary outcome was the occurrence of major bleeding (MB). A total of 3018 patients (median age 70; 39.6% females) were included: median [interquartile range (IQR)] HAS-BLED and ORBIT scores were 1 [1–2] and 1 [0–2], respectively; 356 (11.8%) patients were at high risk for MB using HAS-BLED (≥3) and 123 (4.1%) using ORBIT (≥4). Overall, 60 (2.0%) MB events were recorded, with an incidence of 1.1 per 100 patient-years. Both HAS-BLED and ORBIT were associated with outcome, modestly predicting MB [area under the curve (AUC) 0.653, 95% confidence interval (CI) 0.593–0.714 and AUC 0.601, 95% CI 0.526–0.677, respectively]. Calibration plots showed that both scores were poorly calibrated, particularly the ORBIT score, which showed consistent poorer calibration. Time-dependent reclassification analysis showed a trend towards incorrect lower risk reclassification using ORBIT compared with HAS-BLED. Conclusion In this real-life contemporary cohort of AF patients treated with NOACs, the ORBIT score did not provide reclassification improvement, showing even poorer calibration compared with HAS-BLED. Our findings do not support the preferential use of ORBIT in NOAC-treated AF patients.