Comparison Of Three Cardiac Surgery Risk Evaluation Systems For In-hospital Mortality In Jiangsu Patients Undergoing Isolated Coronary Artery Bypass Grafting

Objectives:To compare the predictive efficiency of postoperative mortality with EuroSCORE Ⅱ (European System for Cardiac Operative Risk Evaluation Ⅱ), STS (Society of Thoracic Surgeons) risk evaluation system and SinoSCORE (Sino-System for Coronary Operative Risk Evaluation) in Chinese patients from Jiangsu province undergoing isolated CABG (Coronary Artery Bypass Grafting).Methods:Perioperative data from the department of cardiothoracic surgery, first affiliated hospital of Nanjing Medical University were collated from the period January2004to December2013, in which1047consecutive patients who underwent isolated Off-Pump CABG (OPCABG) were predicted mortalities online to the EuroSCORE Ⅱ, STS risk evaluation system and SinoSCORE. Average mortality of each system was calculated. Four teams were separated by predicted mortalities in each system:Ⅰ (0-1.99%), Ⅱ (2-3.99%), Ⅲ (4-5.99%), Ⅳ (>6%). The average predictive mortalities in each team were calculated and compared to the observed ones. Stratified by SinoSCORE score, patients were also divided into three groups of low (score≤1), moderate (score2-5) and high risk (score≥6). Calibration and discrimination in total and three groups were assessed by Hosmer-Lemeshow (H-L) test and the area under the receiver operating characteristic curve (AUC) respectively to evaluate the predictive efficiency of three systems.Results:Observed mortality was2.29%(24in1047) overall, EuroSCORE Ⅱ gave a more proximal average result (2.02%) compared to SinoSCORE (1.01%) and STS system (1.13%). For total patients, three systems all showed good calibrations with P value of H-L higher than0.05; in discrimination assessment, SinoSCORE showed available with AUC higher than0.7while EuroSCORE II and STS showed good with AUC higher than0.75. For the separated four teams, three systems all underestimated mortalities in team Ⅲ and Ⅳ with P<0.05while all predicted well in team I. In team Ⅱ, only EuroSCORE Ⅱ predicted well and the other two systems underestimated mortalities with P<0.05. For the divided three risk groups, three systems all showed good calibrations (H-L:P>0.05) in low and high risk group; in moderate risk group, only STS system performed poor (H-L:P=0.028). In discrimination assessment, three systems all performed well in high risk group with AUC higher than0.75and EuroSCORE II showed the best (AUC:0.807,95%CI:0.695-0.920); low and moderate risk group showed no significant in discrimination calculation by the three systems.Conclusion:Three systems all could be applied in risk evaluation pre-operation for Jiangsu CABG patients. EuroSCORE II achieved the most accurate result compared to the other two systems. SinoSCORE needed constant update to adapt to the progression of surgical technology in the future.