RESUMO
AIM: The additive EuroSCORE system for predicting operative mortality of cardiac patients tends to underestimate the mortality risk of high risk patients and concomitantly to overestimate that of low risk patients. We propose a modification of stratification groups aiming at improving its precision. We also tested its ability to predict the length of postoperative mechanical ventilation of our patients. METHODS: The high risk group of the EuroSCORE system (>6 points) was divided into three additional groups (group I: 0-2 points, group II: 3-5 points, group III: 6-8 points, group IV: 9-13 points, group V: >14 points) thus producing a 5 classes system. In a group of 301 cardiac surgery patients operated on in a low volume cardiac center, we calculated the expected mortality rate for each EuroSCORE class, the calibration of the modified scoring system, the ROC and the corresponding AUC values and the relative risk of each predisposing factor used by the original EuroSCORE sytem. RESULTS: The proposed modification increased the discrimination ability of EuroSCORE in predicting mortality (Hosmer-Lemeshow P=0.78, ROC size: 0.791) and marginally affected its accuracy in predicting length of postoperative mechanical ventilation (Hosmer-Lemeshow: 0.11, ROC size: 0.711). Combined operations of CABG and valve replacement were shown to exert a statistically significant effect on mortality (odds ratio 3.85, CI: 1.15-12.87, P=0.028). CONCLUSION: The proposed modification of additive EuroSCORE can presumably increase its discrimination ability in predicting mortality of cardiac patients handled in a low volume cardiac center. The need for prolonged mechanical ventilation could be predicted with acceptable accuracy, possibly providing support in resource management.