The easier, the better: age, creatinine, ejection fraction score for operative mortality risk stratification in a series of 29,659 patients undergoing elective cardiac surgery

Abstract

Objective: Age, preoperative creatinine value, and ejection fraction are easily arranged in the age, creatinine, ejection fraction score to predict operative mortality in elective cardiac operations, as recently shown. We validate the age, creatinine, ejection fraction score in a large multicentric study.

Methods: We analyzed 29,659 consecutive patients who underwent elective cardiac operations in 14 Italian institutions during the period from 2004 to 2009. The operative (30-day) mortality rate was recorded for the entire population and for subgroups of patients based on the risk distribution. The predicted mortality was assessed using the additive and logistic European System for Cardiac Operative Risk Evaluations, and the age, creatinine, ejection fraction score. Accuracy and clinical performance of the different models were tested.

Results: The observed mortality rate was 2.77% (95% confidence interval, 2.59-2.96). The predicted mortality rate was 2.84% (95% confidence interval, 2.79-2.88) for the age, creatinine, ejection fraction score (not significantly different from the observed rate), 6.26% for the additive European System for Cardiac Operative Risk Evaluation, and 9.67% for the logistic European System for Cardiac Operative Risk Evaluation (both significantly overestimated). For all deciles of risk distribution, the European System for Cardiac Operative Risk Evaluation significantly overestimated mortality risk; the age, creatinine, ejection fraction score slightly overestimated the mortality risk in very low-risk patients and significantly underestimated the mortality risk in very high-risk patients, correctly estimating the risk in 7 of 10 deciles. The accuracy of the age, creatinine, ejection fraction score was acceptable (area under the curve of 0.702). In a separate analysis, this value increased to 0.74 by excluding centers that reported no operative mortality. These values were similar or worse for the European System for Cardiac Operative Risk Evaluation.

Conclusions: The age, creatinine, ejection fraction score provides an accuracy level comparable to that of the European System for Cardiac Operative Risk Evaluation, with far superior clinical performance.