EuroSCORE vs. EuroSCORE II vs. society of thoracic surgeons risk algorithm

Introduction: This study presents a validation series for EuroSCORE II compared with the previous additive and logistic EuroSCORE and the Society of Thoracic Surgeons risk prediction algorithm. Patients and methods: Clinical data of 2004 patients undergoing isolated coronary artery bypass surgery between 2006 and 2010 were retrospectively collected and individual expected risks of death were calculated by all 3 risk prediction algorithms. Performance of these risk algorithms was evaluated in terms of discrimination and calibration. Results: There were 76 deaths (3.8%) among 2004 patients. The mean EuroSCORE II predicted mortality was 3.72% ± 5.11%, additive EuroSCORE was 4.35% ± 3.65% and logistic EuroSCORE was 6.41% ± 10.06%. The additive EuroSCORE was better than EuroSCORE II in terms of both discrimination and calibration (C-statistic 0.866 and Hosmer-Lemeshow p value 0.230 vs. C-statistic 0.836 and Hosmer-Lemeshow p value 0.013 for EuroSCORE II). In a subset of 380 patients, we compared EuroSCORE II with the Society of Thoracic Surgeons risk prediction. Actual mortality was 2.89%. Predicted mortality by EuroSCORE II was 4.27% ± 5.22% and Society of Thoracic Surgeons risk prediction was 2.30% ± 4.16%. The area under the curve was 0.759 for EuroSCORE II and 0.898 for the Society of Thoracic Surgeons risk prediction, whereas the Hosmer-Lemeshow p value was 0.267 for EuroSCORE II and 0.981 for Society of Thoracic Surgeons risk prediction. Conclusion: The Society of Thoracic Surgeons risk prediction algorithm is a better risk assessment tool compared to additive and logistic EuroSCORE and EuroSCORE II in Pakistani patients.