Performance of HAS-BLED, ORBIT, PRECISE-DAPT, and PARIS risk score for predicting long-term bleeding events in patients taking an oral anticoagulant undergoing percutaneous coronary intervention

Background: No standardized tool exists for weighting bleeding risk before selecting an antithrombotic regimen in patients undergoing percutaneous coronary intervention (PCI) who require both oral anticoagulant (OAC) and antiplatelet agents. Methods: We performed PCI in 3718 consecutive patients between April 2011 and March 2017, 302 of whom were treated with both OAC and antiplatelet agents. We retrospectively evaluated the predictive performance of four major bleeding risk scores (HAS-BLED, ORBIT, PRECISE-DAPT, and PARIS score). Patients were followed for up to 3 years for bleeding events, defined as a composite of major and minor bleeding according to the Thrombolysis in Myocardial Infarction (TIMI) criteria and the Bleeding Academic Research Consortium (BARC) criteria. Results: TIMI significant bleedings (major, minor, and requiring medical attention) were seen in 90 patients (29.8%); whereas the BARC class >= 3 bleedings were seen in 53 patients (17.5%). Regarding TIMI significant bleedings, HAS-BLED, ORBIT, and PRECISE-DAPT scores equally categorized high-risk patients, but the PARIS score could not [high-risk versus non-high-risk: hazard ratio (HR), 1.74; 95% confidence interval (CI), 1.15-2.64; p = 0.01; HR, 1.63; 95% CI, 1.08-2.48; p = 0.02; HR, 1.62; 95% CI, 1.06-2.51; p = 0.03; HR, 1.05; 95% CI, 0.70-1.63; p = 0.79, respectively); regarding BARC class >= 3 bleeding, all four scores could stratify high-risk patients (high-risk versus non-high-risk: HR, 2.23; 95% CI, 1.30-3.88; p = 0.004; HR, 2.25; 95% CI,1.31-3.96; p = 0.003; HR. 3.87; 95% CI, 2.06-7.91; p < 0.0001; HR, 1.85; 95% CI, 1.04-3.47; p = 0.04, respectively). Conclusions: In patients taking an OAC undergoing PCI, HAS-BLED, ORBIT, and PRECISE-DAPT scores predicted TIMI significant bleeding events better than the PARIS score; whereas all four scores could predict BARC class >= 3 bleeding events. (C) 2018 Japanese College of Cardiology. Published by Elsevier Ltd. All rights reserved.