421-P: Development and Validation of a Risk Prediction Model for Diabetic Kidney Disease in Overweight/Obese Patients with Type 2 Diabetes Mellitus

Introduction and Objective: To develop and validate a risk prediction model for diabetic kidney disease (DKD) in overweight/obese patients with type 2 diabetes mellitus (T2DM) in Guangdong Province, China.

Methods: The study enrolled 1,453 overweight/obese patients with T2DM who were treated at the Department of Endocrinology and Metabolism, the Third Affiliated Hospital of Sun Yat-sen University between January 2021 and December 2023. The data was randomly assigned to form a training set comprising 70% of the patients and a validation set containing the remaining 30%, which included the non-DKD group (n=1,214) and the DKD group (n=239). We developed a predictive nomogram using multivariate logistic regression, incorporating variables from Least Absolute Shrinkage and Selection Operator (LASSO) models. The model's performance was assessed using C-index, calibration plots, and decision curve analysis for both internal and external validation, ensuring its discriminative power, calibration, and clinical utility.

Results: The prevalence of DKD was 16.4%, and the risk prediction model included five predictors, namely, diabetes duration, systolic blood pressure(SBP), albumin(ALB), total cholesterol(TC), and cystatin C(CysC). ALB was a protective factor for DKD in overweight/obese patients with T2DM [OR=0.900, (0.8610.940), P< 0.001]. The model showed high discrimination ability and the area under the curve (AUC) was 0.783 (95% CI: 0.7430.822, P < 0.001). It showed superb calibration against actual DKD rates, with a c-index of 0.783 for internal and 0.763 for external validation. Decision curve analysis also indicated that this new nomogram is clinically useful.

Conclusion: ALB might be a protective factor for DKD in overweight/obese patients with T2DM. The risk prediction model is a useful tool for predicting DKD in overweight/obese patients with T2DM at an early stage, and facilitate prompt and effective intervention.

Y. Zhang: None. H. Li: None. X. Hu: None. F. Xu: None. M. Cai: None. Y. Chen: None.