Development of a Risk Score for Long-term Survival and MACCE After Coronary Artery Bypass Grafting Surgery

Abstract

To develop risk scoring models predicting long-term survival and major adverse cardiovascular and cerebrovascular events (MACCE) including myocardial infarction and stroke after coronary artery bypass grafting (CABG). All-consecutive 4,821 patients undergoing isolated CABG at Lankenau between 01/2005-07/2021 were included. MACCE was defined as all-cause mortality+myocardial infarction (MI)+stroke. Variable selection for both outcomes was obtained using a double selection logit Lasso with adaptive selection. Models performance was internally evaluated by calibration and accuracy using bootstrap cross-validation. Mortality and MACCE were compared among patients split into three groups based on the predicted risk scores for all-cause mortality and MACCE. An external validation of our database was performed with 665 patients from the University of Brescia, Italy. Pre-operative risk predictors were found to be predictors for all-cause mortality and MACCE. In addition, being of African American ethnicity is a significant predictor for MACCE after isolated CABG. The AUC which measures the discrimination of the models were 80.4%, 79.1%, 81.3%, and 79.2% for mortality at 1,2,3, and 5 years follow-up. The AUC for MACCE were 75%, 72.5%, 73,8%, and 72.7% at 1, 2, 3, and 5-years follow-up. For external validation, AUC for all-cause mortality and MACCE at 1, 2, 3, and 5 years was 73.7%, 70.8%, 68.7%, and 72.2% and 72.3%, 68.2%, 65.6%, 69.6%, respectively. The Advanced (AD) Coronary Risk Score for All-Cause Mortality and MACCE provide good discrimination of long-term mortality and MACCE after isolated CABG. External validation observed a more AUC's above 70%.