Validation of EuroSCORE II in Patients Undergoing Coronary Artery Bypass Surgery

Background. The European System for Cardiac Operative Risk Evaluation II (EuroSCORE II) has been recently developed to improve the performance of the original EuroSCORE. Herein we evaluated its discriminatory ability in predicting the immediate and late outcome after coronary artery bypass grafting (CABG). Methods. Complete data on 1,027 patients who underwent isolated CABG were available for validation of EuroSCORE II and to compare its discriminatory ability with the original EuroSCORE and its Finnish modified version. Results. EuroSCORE II performed somewhat better (area under the curve [AUC] 0.852, Brier score 0.031) than the original logistic EuroSCORE (AUC 0.838, Brier score 0.034) and its Finnish modified version (AUC 0.825, Brier score 0.034) in predicting operative mortality. The overall expected-to-observed operative mortality ratio for the original logistic EuroSCORE was 1.8, for its Finnish modified version was 0.6, and for EuroSCORE II was 1.2. EuroSCORE II showed expected-to-observed ratios ranging from 1.05 to 1.17 in its highest third quintiles. The best cutoff of EuroSCORE II in predicting operative postoperative mortality was 10% (21.5% vs 1.6%, p < 0.0001; sensitivity 91.5%, specificity 60.5%, negative predictive value 98.4%, accuracy of 90.3%). The EuroSCORE II was predictive of de novo dialysis (AUC 0.805), prolonged use of inotropes (AUC 0.748), and intensive care unit stay 5 days or greater (AUC 0.793). The risk of late mortality significantly increased across increasing quintiles of EuroSCORE II (p < 0.0001). Conclusions. The EuroSCORE II performs better than its original version in predicting operative mortality and morbidity after isolated CABG. Its ability to predict 30-day mortality in high-risk patients is of particular importance. The EuroSCORE II is also a good predictor of late postoperative survival. (Ann Thorac Surg 2012;93:1930-5) (C) 2012 by The Society of Thoracic Surgeons