Development and Validation of a Prognostic Model for Patients with Heart Failure with reduced Ejection Fraction

Abstract

Prognostic predictive model for patients with heart failure with reduced ejection fraction (HFrEF) is scarce. This study aimed to develop a prognostic model for HFrEF patients. This retrospective cohort enrolled 211 eligible patients with HFrEF, the median follow-up of this cohort was 16 months, the one-year mortality rate was 26.1%(55/211). Cox regression showed that age, history of coronary artery disease (CAD), glucose (Glu), and the use of beta-blockers were independent predictors of the occurrence of all-cause mortality in patients with HFrEF after discharge. The cohort was divided into the development set (N = 120) and the validation set (N = 91) in a ratio of 6:4. Least absolute shrinkage and selection operator (LASSO) regression and cox regression screened out 4 variables for the development of the model, including age, sex, serum creatinine (sCr) and use of beta-blockers). A nomogram was constructed, it has a C indexes 0.76 (95%CI 0.67–0.84), AUC of ROC curve of 1 year mortality of the model was 0.79 (95%CI 0.68–0.89), indicating a good differentiation of the model. Calibration plot of the model was drawn and revealed a good calibration. The DCA plots showed that when the threshold probability was between 3 ~ 77% (development group), the model adds benefit to the patients. This model is sure to cast some light on clinical medicine for HFrEF.