A Novel Risk Stratification System for Predicting In-Hospital Mortality Following Coronary Artery Bypass Grafting Surgery with Impaired Left Ventricular Ejection Fraction.

Abstract

Coronary artery disease (CAD) is the most common cause of heart failure (HF), and impaired ejection fraction (EF<50%) is a crucial precursor to HF. Coronary artery bypass grafting (CABG) is an effective surgical solution to CAD-related HF. In light of the high risk of cardiac surgery, appropriate scores for groups of patients are of great importance. We aimed to establish a novel score to predict in-hospital mortality for impaired EF patients undergoing CABG. Clinical information of 1,976 consecutive CABG patients with EF<50% was collected from January 2012 to December 2017. A novel system was developed using the logistic regression model to predict in-hospital mortality among patients with EF<50% who were to undergo CABG. The scoring system was named PGLANCE, which is short for seven identified risk factors, including previous cardiac surgery, gender, load of surgery, aortic surgery, NYHA stage, creatinine, and EF. AUC statistic was used to test discrimination of the model, and the calibration of this model was assessed by the Hosmer-lemeshow (HL) statistic. We also evaluated the applicability of PGLANCE to predict in-hospital mortality by comparing the 95% CI of expected mortality to the observed one. Results were compared with the European Risk System in Cardiac Operations (EuroSCORE), EuroSCORE II, and Sino System for Coronary Operative Risk Evaluation (SinoSCORE). By comparing with EuroSCORE, EuroSCORE II and SinoSCORE, PGLANCE was well calibrated (HL P = 0.311) and demonstrated powerful discrimination (AUC=0.846) in prediction of in-hospital mortality among impaired EF CABG patients. Furthermore, the 95% CI of mortality estimated by PGLANCE was closest to the observed value. PGLANCE is better with predicting in-hospital mortality than EuroSCORE, EuroSCORE II, and SinoSCORE for Chinese impaired EF CABG patients.