Predictive performance of different bleeding risk scores in patients with atrial fibrillation and acute coronary syndrome or undergoing percutaneous coronary intervention

Abstract

Abstract Background Standard scoring system for bleeding risk assessment has not been developed in patients with atrial fibrillation (AF) and acute coronary syndrome (ACS) or undergoing percutaneous coronary intervention (PCI). The study aims to evaluate the predictive value of the HAS-BLED, ORBIT, ARTIA, PARIS, and PRECISE-DEAPT score in patients with AF and ACS/PCI who received both anticoagulant and antiplatelet therapy. Methods A total of 930 patients with AF and ACS/PCI receiving both anticoagulant and antiplatelet therapy were consecutively recruited and followed up for 1 year. The primary endpoint was defined according to the bleeding academic research consortium (BARC) criteria as major (BARC 3a, 3b, 3c, and 5) and any bleeding (BARC 2, 3a, 3b, 3c, and 5). The secondary endpoints included major or minor bleeding according to TIMI criteria. Univariate and multivariate logistic regression were performed to evaluate the association between bleeding risk scores and outcomes. Receiver operating characteristic (ROC) curves were constructed and c-statistics were calculated to evaluate the predictive performance of different bleeding risk scores. Results BARC class ≥2 bleedings occurred in 36 patients (3.9%), while BARC class ≥3 bleedings were seen in 134 patients (14.4%). The HAS-BLED, ORBIT, ARTIA, PARIS, and PRECISE-DEAPT score were significantly associated with incidences of BARC class ≥2 and ≥3 bleedings during 1-year follow-up (all p<0.05). The ROC analysis showed that the predictive performance of the five bleeding risk scores for BARC class ≥3 bleedings were low-to-moderate [c-statistic: HASBLED 0.584 (0.552–0.616), ATRIA 0.624 (0.592–0.655), ORBIT 0.623 (0.591–0.654), PARIS 0.633 (0.601–0.664), PRECISEDAPT 0.616 (0.584–0.647)]. Multivariable logistic regression indicated that previous bleeding history and hemoglobin were independent predictors of BARC class ≥3 bleedings. Compared to the HAS-BLED score, the model constructed by previous bleeding history and hemoglobin displays a significant improvement in bleeding risk prediction (0.704 vs. 0.584, p=0.013). Conclusions In patients with AF and ACS/PCI receiving both anticoagulant and antiplatelet therapy, the HAS-BLED, ORBIT, ARTIA, PARIS, and PRECISE-DEAPT score displayed low-to-moderate performance for predicting BARC class ≥3 bleedings during 1-year follow-up. More accurate scoring systems for bleeding risk evaluation should be developed for these patients in the future. Funding Acknowledgement Type of funding sources: None. Receiver operating characteristic curve