Development and validation of a risk predictive model for incident chronic kidney disease in prediabetic patients based on time-dependent Cox regression

To develop and validate a risk prediction model for incident chronic kidney disease (CKD) in individuals with prediabetes by integrating clinical, biochemical, and demographic parameters to facilitate early prevention strategies. A retrospective study was conducted on prediabetic patients from the outpatient and inpatient departments of Zhongshan Hospital, Fudan University, between January 1, 2014, and November 5, 2024. Laboratory test results and imaging examination data were collected. Incident chronic kidney disease was defined as the study endpoint. A time-dependent Cox regression model was employed to construct a risk prediction model for incident CKD in prediabetic patients, followed by internal validation and performance evaluation. During the research process, DeepSeek-R1 was utilized for feature extraction from clinical data. A total of 13,966 participants were included. Cox regression analysis and five-fold cross-validation were used for model development and internal validation, respectively. Uric acid disorders, hypertension, anemia, glucose-lowering medication (GLM) class count, use of renal-protective GLM, use of insulin, age (per 10 years), left ventricular ejection fraction (LVEF), use of novel oral anticoagulants (NOACs), low high-density lipoprotein (HDL) cholesterol, and carotid intima–media thickness (IMT), and gender were incorporated into the model. Five-fold cross-validation based on the time-dependent Cox model demonstrated time-dependent area under the curve (AUC) values of 0.818(95%CI:0.765,0.863), 0.815(95%CI:0.791,0.838), 0.801(95%CI:0.782,0.819), 0.782(95%CI: 0.764,0.797), 0.744(95%CI: 0.723,0.758), and 0.685(95%CI 0.659,0.705) at 1, 3, 5, 7, 9, and near 10 years, respectively. The developed model demonstrated moderate but stable discriminative ability and satisfactory calibration over short- and intermediate-term follow-up, and it should serve as an early risk-screening reference rather than a definitive clinical decision-making tool. ChiCTR2400089463, 2024-09-09.