Evaluation of estimation of physiologic ability and surgical stress to predict in-hospital mortality in cardiac surgery.

PURPOSE: Prediction of postoperative risk in cardiac surgery is important for cardiac surgeons and anesthesiologists. We generated a prediction rule for elective digestive surgery, designated as Estimation of Physiologic Ability and Surgical Stress (E-PASS). This study was undertaken to evaluate the accuracy of E-PASS in predicting postoperative risk in cardiac surgery.
METHODS: We retrospectively collected data from patients who underwent elective cardiac surgery at a low-volume center (N = 291) and at a high-volume center (N = 784). Data were collected based on the variables required by E-PASS, the European system for cardiac operative risk evaluation (EuroSCORE), and the Ontario Province Risk Score (OPRS). Calibration and discrimination were assessed by the Hosmer-Lemeshow test and the area under the receiver operating characteristic curve (AUC), respectively. The ratio of observed-to-estimated in-hospital mortality rates (OE ratio) was defined as a measure of quality.
RESULTS: In-hospital mortality rates were 7.6% at the low-volume center and 1.3% at the high-volume center, accounting for an overall mortality rate of 3.0%. AUC values to detect in-hospital mortality were 0.88 for E-PASS, 0.77 for EuroSCORE, and 0.71 for OPRS. Hosmer-Lemeshow analysis showed a good calibration in all models (P = 0.81 for E-PASS, P = 0.49 for EuroSCORE, and P = 0.94 for OPRS). OE ratios for the low-volume center were 0.83 for E-PASS, 0.70 for EuroSCORE, and 0.83 for OPRS, whereas those for the high-volume center were 0.26 for E-PASS, 0.14 for EuroSCORE, and 0.27 for OPRS.
CONCLUSIONS: E-PASS may accurately predict postoperative risk in cardiac surgery. Because the variables are different between cardiac-specific models and E-PASS, patients' risks can be double-checked by cardiac surgeons using cardiac-specific models and by anesthesiologists using E-PASS.