Machine learning model based on routine blood and biochemical parameters for early diagnosis of diabetic kidney disease - read online

Aug 3, 2026

Upload by Aurelia

Machine learning enables earlier identification of diabetic kidney disease (DKD), yet clinical early detection remains difficult because standard biomarkers have limited reliability and sensitivity. This study develops and validates machine learning models using routine blood and biochemical parameters for early DKD prediction. A retrospective cohort was combined with external validation, comparing seven algorithms and using SHAP for feature importance. Logistic regression delivered the best performance and highlighted TyG, HbA1c, and globulin as key predictors for cost-effective screening.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages14

Reading Time35 min

AccessFree to download

Background

Methods

Results

Conclusions

Introduction

TYPE Original Research PUBLISHED 28 January 2026 DOI 10.3389/fendo.2026.1720574 OPEN ACCESS EDITED BY Cem Haymana, University of Health Sciences, Türkiye REVIEWED BY Ashwin Dhakal, The University of Missouri, United States Sheng Ding, The Central Hospital of Wuhan, China *CORRESPONDENCE Jun-jie Gao [wnmcjykjj@163.com](wnmcjykjj@163.com)[ ](wnmcjykjj@163.com)Ruo-xue Cao [caoruoxue9490@163.com](caoruoxue9490@163.com) †These authors have contributed equally to this work RECEIVED 08 October 2025 REVISED 28 December 2025 ACCEPTED 05 January 2026 PUBLISHED 28 January 2026 CITATION Yong W, Peng D-d, Ye K, Gao J-jand Cao R-x (2026) Machine learning model based on routine blood and biochemical parameters for early diagnosis of diabetic kidney disease. Front. Endocrinol. 17:1720574 . doi: 10.3389/fendo.2026.1720574 COPYRIGHT © 2026 Yong, Peng, Ye, Gao and Cao. This isan open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Machine learning model based on routine blood and biochemical parameters for early diagnosis of diabetic kidney disease Wei Yong1†, Dan-dan Peng 1†, Kai Ye1, Jun-jie Gao 1* and Ruo-xue Cao 2,3* 1 Department of Clinical Laboratory, The Second Afﬁliated Hospital of Wannan Medical College, Wuhu, China, 2 Department of Laboratory Medicine, The Second People ’s Hospital of Lianyungang, Lianyungang, China, 3 Department of Laboratory Medicine, The Second People’s Hospital of Lianyungang Afﬁliated with Kangda College of Nanjing Medical University, Lianyungang, China Background: Diabetic kidney disease (DKD) is the leading cause of end-stage renal disease globally, yet early diagnosis remains challenging due to conventional biomarker limitations, including UACR variability and reduced eGFR sensitivity. While machine learning shows promise in diabetes prediction, its application to early DKD identiﬁcation using routine parameters remains underexplored. This study aimed to develop and validate machine learning models incorporating routine blood and biochemical parameters for early DKD prediction. Methods: This retrospective study analyzed 3,114 diabetic patients from the Second Afﬁliated Hospital of Wannan Medical College (EDN1) and 1,496 patients from NHANES 2005-2018 (EDN2) for external validation. Early DKD was deﬁned as UACR 30–300 mg/g with eGFR ≥60 ml/min/1 .73m² . Seven machine learning algorithms were compared. Feature importance was assessed using SHAP framework, and Mendelian randomization explored causal relationships. Results: Among 3,114 patients, 1,333 (42 . 8%) had early DKD. Logistic regression achieved optimal performance (AUC = 0 . 689, sensitivity=40 .5%, speciﬁcity=81 .3%) . Top predictors included triglyceride-glucose index (TyG), gender, creatinine, globulin, and age. External validation conﬁrmed signiﬁcant associations for HbA1c, globulin, TyG, and neutrophil-to-albumin ratio. Conclusions: The machine learning model successfully identiﬁed early DKD using routine parameters, with TyG index, HbA1c, and globulin as key predictors, demonstrating potential as a cost-effective screening tool. KEYWORDS diabetic kidney disease, early diagnosis, machine learning, risk prediction, routine blood parameters Frontiers in Endocrinology 01 [frontiersin.org](frontiersin.org) 1 Introduction Diabetic kidney disease (DKD), the most common and severe microvascular complication of diabetes, has become the leading cause of end-stage renal disease (ESRD) worldwide (1) . Its prevalence continues to rise alongside the global diabetes epidemic, imposing a substantial burden on patient

FAQs

How is early diabetic kidney disease defined in this study?

Early DKD is defined as UACR 30–300 mg/g with eGFR ≥ 60 ml/min/1.73m².

Which machine learning model performed best?

Logistic regression achieved the optimal performance (AUC 0.689) with reported sensitivity and specificity values.

What predictors were identified as most important?

Top predictors included the triglyceride-glucose index (TyG), gender, creatinine, globulin, and age, with external validation supporting HbA1c, globulin, TyG, and neutrophil-to-albumin ratio.

Aug 3, 2026

Upload by Aurelia