Using the ORACLE Risk Score to Assess Hemorrhagic Risk in Patients with Acute Coronary Syndrome and Atrial Fibrillation

Aim. To study the prognostic value of the ORACLE risk score for assessing the risk of bleeding in patients with acute coronary syndrome (ACS) undergoing anticoagulants for atrial fibrillation using the combined database of the ORACLE II and RECORD 3 registers. Material and methods. This analysis included patients with ACS from 2 observational studies: ORACLE II (ObseRvation after Acute Coronary syndrome for deveLopment of trEatment options; n=1803) and the RECORD-3 register (n=2370). In total, the database included 4173 patients, of which 246 (6.08%) received oral anticoagulants for atrial fibrillation. The mean age of patients was 64.7 +/- 11.9 years, 2493 (59.7%) were men. Hemorrhagic risk was assessed using the ORACLE, CRUSADE, ORBIT, and HAS-BLED risk score. Results. Patients receiving anticoagulant therapy were older (69.9 +/- 11.3 years and 64.0 +/- 12.2 years, p<0.001). Among these patients there was a larger proportion of women, and a smaller proportion of patients with ACS with ST elevation, they were more likely to have chronic heart failure, chronic kidney disease, history of stroke. Among patients receiving anticoagulants and included in the ORACLE study, the frequency of percutaneous coronary intervention was higher than in patients included in the RECORD study. In the joint database, 71 significant bleeding was recorded during the hospitalization period - 64 (1.7%) in patients without anticoagulants and 7 (2.8%) among patients taking anticoagulants (p=0.06). Over 6 months, among patients who did not receive anticoagulants, there were 97 cases of bleeding (in 2.6% of patients), in the group of patients receiving anticoagulants - 12 cases of bleeding (4.9%) - the differences in frequency were significant (p=0.029). The ORACLE risk score had the greatest prognostic value (area under the ROC curve 0.874 +/- 0.0416, sensitivity 82.7%, specificity 79.1%). The predictive value of the HAS-BLED risk score was slightly lower (area under the ROC curve 0.710 +/- 0.0360, sensitivity 63.2%, specificity 56.8%). The value of the CRUSADE risk score (area under the ROC curve 0.612 +/- 0.0269, sensitivity 53.7%, specificity 59.5%) and ORBIT risk score (area under the ROC curve 0.606 +/- 0.0457, sensitivity 62.5%, specificity 58.3%) were lower (p<0.001 for all scales). Conclusion. The use of the ORACLE bleeding risk score can be recommended for patients with ACS requiring anticoagulant therapy.