Abstract
INTRODUCTION: Risk prediction models, such as The Society of Thoracic Surgeons (STS) risk score and the European System for Cardiac Operative Risk Evaluation II (EuroSCORE II), are recommended for assessing operative mortality in coronary artery bypass grafting (CABG). However, their performance is questionable in Brazil. OBJECTIVE: To assess the performance of the STS score and EuroSCORE II in isolated CABG at a Brazilian reference center. METHODS: Observationaland prospective study including 438 patients undergoing isolated CABG from May 2022-May 2023 at the Instituto Dante Pazzanese de Cardiologia. Observed mortality was compared with predicted mortality (STS score and EuroSCORE II) by discrimination (area under the curve [AUC]) and calibration (observed/expected ratio [O/E]) in the total sample and subgroups of stable coronary artery disease (CAD) and acute coronary syndrome (ACS). RESULTS: Observed mortality was 4.3% (n=19) and estimated at 1.21% and 2.74% by STS and EuroSCORE II, respectively. STS (AUC=0.646; 95% confidence interva [CI] 0.760-0.532) and EuroSCORE II (AUC=0.697; 95% CI 0.802-0.593) presented poor discrimination. Calibration was absent for the North American mode (P<0.05) and reasonable for the European model (O/E=1.59, P=0.056). In the subgroups, EuroSCORE II had AUC of 0.616 (95% CI 0.752-0.480) and 0.826 (95% CI 0.991-0.661), while STS had AUC of 0.467 (95% CI 0.622-0.312) and 0.855 (95% CI 1.0-0.706) in ACS and CAD patients, respectively, demonstrating good score performance in stable patients. CONCLUSION: The predictive models did not perform optimally in the total sample, but the EuroSCORE was superior, especially in elective stable patients, where accuracy was satisfactory.