Creation and assessment of a predictive model for the risk of diabetic kidney disease in individuals with type 2 diabetes mellitus: a study utilizing hospital data platforms

By
ZhanLin Zhang
GuoXia Ma
MeiFang Ma
TianRong Guo
Yu Zhao
XiuSheng Cheng
Zhonglin Yan
ZhengMing Yin
Gangyi Wang
YongDong An
June 30, 2026

Frontiers In Endocrinology

Overview

This study developed a nomogram to predict diabetic kidney disease (DKD) risk in patients with type 2 diabetes mellitus (T2DM) using clinical data.

Background

Diabetic kidney disease (DKD) is a significant complication of type 2 diabetes mellitus (T2DM) and a leading cause of end-stage renal disease (ESRD). Early identification of patients at high risk for DKD is crucial for timely intervention, as approximately 20%-40% of T2DM patients may develop this condition. Current screening methods have limitations.

Data Highlights

Data Point	Value
Patients with DKD	5,019 (21.68%)
Patients without DKD	18,133 (78.32%)
AUC (Training Set)	0.773 (95% CI: 0.764-0.782)
AUC (Temporal Validation Set)	0.758 (95% CI: 0.743-0.774)

Key Findings

LASSO regression identified nine predictors for DKD: hypertension, diabetes duration, HbA1c, eGFR, urine protein, serum creatinine, uric acid, total cholesterol, and homocysteine.
The nomogram showed moderate discrimination with an AUC of 0.773 in the training set.
In temporal validation, the AUC remained at 0.758.
Calibration curves indicated good agreement between predicted and observed probabilities.

Clinical Implications

The developed nomogram can serve as a decision-support tool for clinicians to identify patients with T2DM at high risk for DKD.

Conclusion

This study presents a predictive model for DKD risk in T2DM patients. Further multicenter external validation is necessary before widespread implementation.