ABC-AF-Stroke score predicts thromboembolism in non-anticoagulated patients following successful atrial fibrillation ablation: a report from the Chinese Atrial Fibrillation Registry

BACKGROUND: The age, biomarkers, and clinical history (ABC)-atrial fibrillation (AF)-Stroke score have been proposed to refine stroke risk stratification, beyond what clinical risk scores such as the CHA(2)DS(2)-VASc score can offer. This study aimed to identify risk factors associated with thromboembolism and evaluate the performance of the ABC-AF-Stroke score in predicting thromboembolism in non-anticoagulated AF patients following successful ablations. METHODS: A total of 2692 patients who underwent successful ablations with discontinued anticoagulation after a 3-month blanking period in the Chinese Atrial Fibrillation Registry (CAFR) between 2013 and 2019 were included. Cox regression analysis was conducted to present the association of risk factors with thromboembolism risk. The ABC-AF-Stroke score was evaluated in terms of discrimination, including concordance index (C-index), net reclassification improvement (NRI) and integrated discrimination improvement (IDI), clinical utilization by decision curve analysis (DCA), and calibration by comparing the predicted risk with the observed annualized event rate. RESULTS: After a median follow-up of 3.5 years, 64 patients experienced thromboembolism events. Age, prior history of stroke/transient ischemic attack (TIA), high-sensitivity cardiac troponin T (cTnT-hs), and N-terminal pro-B-type natriuretic peptide (NT-proBNP) were independently associated with thromboembolism risk. The ABC-AF-Stroke score performed statistically significantly better than the CHA(2)DS(2)-VASc score in terms of C-index (0.67, 95% confidence interval [CI]: 0.59–0.74 vs. 0.60, 95% CI: 0.52–0.67, P = 0.030) and reclassification capacity. The DCA implied that the ABC-AF-Stroke score could identify more thromboembolism events without increasing the false positive rate compared to the CHA(2)DS(2)-VASc score. The calibration curve showed that the ABC-AF-Stroke score was well calibrated in this population. CONCLUSIONS: In this real-world study enrolling non-anticoagulated AF patients following successful ablations, age, prior history of stroke/TIA, level of NT-proBNP, and cTnT-hs were independently associated with an increased risk of thromboembolism. The ABC-AF-Stroke score was well-calibrated and statistically significantly outperformed the CHA(2)DS(2)-VASc score in predicting thromboembolism risk.