Development and Validation of a Cardiovascular Disease Risk Prediction Model for Patients with Non-Dialysis-Dependent Chronic Kidney Diseases Based on the Nomogram.

Abstract

Introduction: Most chronic kidney disease (CKD) patients experience cardiovascular issues before commencing renal replacement therapy. An accuracy prediction model is helpful for physicians to assess cardiovascular prognoses in each individual and to provide insights on how to outline individualized lines of therapy.

Method: This study enrolled 1,138 participants with non-dialysis-dependent chronic kidney disease (NDD-CKD). Following a proportion of 7:3, patients were randomly assigned to training and validation cohorts. The relevant predictors of cardiovascular events were screened using the least absolute shrinkage and selection operator (Lasso) regression. The area under the receiver operating characteristic curve (AUC) and the calibration curve with 1,000 bootstrap resamples were used to assess the nomogram's performance. Tests on the discrimination of the prediction model used Kaplan-Meier (KM) curve.

Results: After screening all the predictors by lasso regression, the five remaining ones (albumin, estimated glomerular filtration rate, etiology of CKD, cardiovascular disease history, and age) were used to construct the prediction model. The AUCs of 1 year, 2 years, and 3 years were 0.81 (95% CI = 0.75-0.87), 0.80 (95% CI = 0.75-0.86), and 0.80 (95% CI = 0.73-0.86), respectively. The calibration curve and the KM curve showed good prediction features, and the external validation also had a good prediction performance (AUCs of 1, 2, and 3 years were 0.77, 0.84, and 0.82, respectively).

Conclusion: We successfully developed a novel nomogram that has decent prediction performance and can be used for assessing the probability of cardiovascular events in patients with NDD-CKD, displaying valuable potential for clinical application.