Clinical performances of EuroSCORE II risk stratification model in the Serbian cardiac surgical population: a single centre validation study including 10,048 patients

Background/Aim. The EuroSCORE II has recently been developed with an idea to provide better accuracy in prediction of perioperative mortality in the patients who underwent open heart surgery. The aim of this study was to validate clinical performances of the European System for Cardiac Operative Risk Evaluation (EuroSCORE) II risk stratification model in the Serbian adult cardiac surgical population undergoing open heart surgery. Methods. The Euro- SCORE II values on 10,048 consecutive patients undergoing major adult cardiac surgery from 1st January 2012 to 31st March 2017, were prospectively calculated and entered the institutional database. The discriminative power of the model was tested by calculating the area under the receiver operating characteristic curve (AUC). The calibration of the model was assessed by the Hosmer-Lemeshow (H-L) statistics and the observed to expected (O/E) mortality ratio. The patients with the EuroSCORE II values of 0.5-2.50%, > 2.50-6.50%), and > 6.50% were defined to be at low, moderate, and high perioperative risk, respectively. Results. The observed in-hospital mortality was 3.86% (388 of 10,048) and the mean predicted mortality by the Euro- SCORE II was 3.61%. The discriminatory power was very good for the entire cohort as well as for all subgroups [coronary, valve(s), combined (coronary plus valve), aortic and other] of performed cardiac procedures (all AUCs > 0.75). The H-L test confirmed good calibration only for category other cardiac procedures. The O/E mortality ratio confirmed good calibration for the whole sample [O/E ratio 1.07, 95% confidence interval (CI) 0.96-1.18] and for all subgroups of performed cardiac procedures, excluding significant underprediction of mortality for aortic surgery (O/E ratio 1.64; 95% CI 1.31-1.97). The EuroSCORE II overestimated perioperative risk in a low and underestimated perioperative risk in a high risk group, with acceptable discrimination (both AUCs = 0.72). On the contrary, the O/E mortality ratio confirmed good calibration for all three subcategories of high risk group. Conclusion. The results of our study confirmed acceptable overall performances of the EuroSCORE II risk stratification model in terms of discrimination and the accuracy of model when applied to the contemporary Serbian cardiac surgical cohort undergoing open heart surgery at our Institute.