A novel, comprehensive tool for predicting 30-day mortality after surgical aortic valve replacement

Abstract

Abstract OBJECTIVES We sought to develop and validate a novel risk assessment tool for the prediction of 30-day mortality after surgical aortic valve replacement incorporating a patient’s frailty. METHODS Overall, 4718 patients from the multicentre study OBSERVANT was divided into derivation (n = 3539) and validation (n = 1179) cohorts. A stepwise logistic regression procedure and a criterion based on Akaike information criteria index were used to select variables associated with 30-day mortality. The performance of the regression model was compared with that of European System for Cardiac Operative Risk Evaluation (EuroSCORE) II. RESULTS At 30 days, 90 (2.54%) and 35 (2.97%) patients died in the development and validation data sets, respectively. Age, chronic obstructive pulmonary disease, concomitant coronary revascularization, frailty stratified according to the Geriatric Status Scale, urgent procedure and estimated glomerular filtration rate were independent predictors of 30-day mortality. The estimated OBS AVR score showed higher discrimination (area under curve 0.76 vs 0.70, P < 0.001) and calibration (Hosmer–Lemeshow P = 0.847 vs P = 0.130) than the EuroSCORE II. The higher performances of the OBS AVR score were confirmed by the decision curve, net reclassification index (0.46, P = 0.011) and integrated discrimination improvement (0.02, P < 0.001) analyses. Five-year mortality increased significantly along increasing deciles of the OBS AVR score (P < 0.001). CONCLUSIONS The OBS AVR risk score showed high discrimination and calibration abilities in predicting 30-day mortality after surgical aortic valve replacement. The addition of a simplified frailty assessment into the model seems to contribute to an improved predictive ability over the EuroSCORE II. The OBS AVR risk score showed a significant association with long-term mortality.