Development and Evaluation of a Clinical Prediction Model for Acute Kidney Injury Risk in Elderly Patients with Heart Failure with Reduced Ejection Fraction
TL;DR
Clinicians can utilize this model to optimize clinical decision-making based on individualized patient characteristics, enabling personalized risk stratification for therapeutic interventions and prognostic evaluations.
Abstract
Background: Acute kidney injury (AKI) is a significant complication among patients with heart failure with reduced ejection fraction (HFrEF). Once AKI occurs, therapeutic options are limited to supportive care, underscoring the clinical importance of early identification of patients at risk. This study sought to identify independent predictors of AKI development in geriatric HFrEF patients and subsequently establish a clinically applicable bedside risk quantification. Method: A total of 1,496 elderly patients (≥60 years) diagnosed with HFrEF at Guangdong Provincial People’s Hospital from January 2010 to December 2024 were enrolled according to predefined inclusion/exclusion criteria and stratified into an AKI group (n = 300) and a non-AKI group (n = 1196). Relevant parameters were screened using LASSO (Least Absolute Shrinkage and Selection Operator) regression, and risk factors for AKI in HFrEF patients were identified through univariate and multivariate logistic regression analyses. A nomogram was developed based on multivariate logistic regression results, accompanied by a corresponding heatmap. The predictive accuracy of the nomogram was evaluated using receiver operating characteristic (ROC) curves and calibration plots, while its clinical utility was demonstrated through decision curve analysis (DCA). Result: This study identified four independent predictors of AKI in patients with HFrEF, including N-terminal pro-B-type natriuretic peptide (NT-proBNP), uric acid (UA), CRP (C-reaction protein) and urea. A nomogram was developed based on these factors. The evaluation results demonstrated that the model exhibited good diagnostic accuracy, with an area under the receiver operating characteristic curve (AUC) of 0.721 (95% CI: 0.658–0.785). The calibration curve and DCA further indicated that the model possesses favorable clinical applicability. Conclusion: This study developed and validated a nomogram model for predicting AKI in patients with HFrEF. Clinicians can utilize this model to optimize clinical decision-making based on individualized patient characteristics, enabling personalized risk stratification for therapeutic interventions and prognostic evaluations.