AusSCORE II in predicting 30-day mortality after isolated coronary artery bypass grafting in Australia and New Zealand

Objectives: To update the Australian System for Cardiac Operative Risk Evaluation (AusSCORE) model for operative estimation of 30-day mortality risk after isolated coronary artery bypass grafting in the Australian population. Methods: Data were collected by the Australian and New Zealand Society of Cardiac and Thoracic Surgeons registry from 2001 to 2011 in 25 hospitals. A total of 31,250 patients underwent isolated coronary artery bypass grafting and the outcome was 30-day mortality. A total of 2154 (6.9%) patients had 1 or multiple missing values. Missing values were estimated assuming missing completely at random and logistic regression with a generalized estimating equation was used to address within-hospital variance. Bootstrapping methods were used to construct and validate the updated model (AusSCORE II). Also the model was validated on an out-of-creation sample of 4700 patients who underwent bypass surgery in 2012. Results: The average age of the patients was 65.6 +/- 12.9 years and 78.6% were male. Thirteen variables were selected in the updated model. The bootstrap discrimination and calibration of the AusSCORE II was very good (receiver operating characteristics [ROC], 82.0%; slope calibration, 0.987). The overall observed/AusSCORE II predicted mortality was 1.63% compared with the original AusSCORE predicted mortality of 1.01%. The validation of the AusSCORE II on the out-of-sample data also showed a high performance of the model (ROC, 84.5%; Hosmer-Lemoshow P value, .7654). Conclusions: The AusSCORE II model provides improved prediction of 30-day mortality and successfully stratifies patient risk. The model will be useful to improve the preoperative consultation regarding risk stratification in terms of 30-day mortality.