Risk factors and logistic regression-based nomogram prediction model for diabetic retinopathy in type 2 diabetes mellitus
- 1Department of Ophthalmology, The First Affiliated Hospital of Ningbo University
Abstract
Introduction: Diabetic Retinopathy (DR) is a major cause of irreversible vision loss in Type 2 Diabetes Mellitus (T2DM). Early identification of high-risk patients using readily available clinical tools is crucial. This study aimed to analyze risk factors for DR in T2DM patients and develop a practical nomogram prediction model based on common clinical and laboratory parameters.
Methods: A retrospective study was conducted on 379 T2DM patients (64 with DR) from a single center. LASSO regression (with the penalty λ chosen by the one-standard-error rule, λ.1se) was used for variable selection; the selected variables were entered into multivariable logistic regression and only those that remained independently significant were retained to construct the nomogram. Model performance was evaluated by discrimination (Area Under the Curve, AUC), calibration (Hosmer-Lemeshow test), and clinical utility (Decision Curve Analysis). Only internal validation (1000-sample bootstrapping) was performed; no external validation cohort was available.
Results: Five independent risk factors were identified: longer disease duration (OR = 1.27 per year), hypertension (OR = 3.56), elevated fasting plasma glucose (OR = 1.86 per mmol/L), higher glycated hemoglobin (OR = 6.38 per %), and hypertriglyceridemia (triglycerides ≥1.7 mmol/L, OR = 6.37). The nomogram showed good discrimination (AUC 0.922, 95% CI 0.886-0.954; bootstrap optimism-corrected AUC 0.914), good calibration (Hosmer-Lemeshow P = 0.739; bootstrap-corrected calibration slope 0.93), and a positive net benefit on decision curve analysis.
Discussion: This nomogram, incorporating five easily obtainable clinical variables, is an effective tool for predicting individual DR risk. Its practicality makes it suitable for initial risk stratification in primary care settings, enabling timely ophthalmology referrals. As a single-center model developed with a limited number of DR events and without external validation, its generalizability should be interpreted with caution and confirmed in larger, multi-center, externally validated cohorts.
Related
- Blood biomarkers and penalized logistic regression for the prediction of complicated appendicitis in children: a retrospective cohort study
- Rates of nonproliferative diabetic retinopathy (NPDR) regression among patients with very mild or mild NPDR in an Indian Health Service teleophthalmology program, 2015 – 2020
- Early postoperative prediction of gastrointestinal bleeding after surgery for acute type A aortic dissection: a five-variable logistic regression model benchmarked against machine learning.
- Development and Internal Validation of a Risk Prediction Model for Visual Impairment Progression in Elderly Patients with Diabetic Retinopathy: A Single-Center Retrospective Study.
- Development and internal validation of a prediction model for new-onset macular oedema following vitrectomy for diabetic retinopathy: a single-centre retrospective analysis
BibTeX
@article{tian2026risk,
title = {{Risk factors and logistic regression-based nomogram prediction model for diabetic retinopathy in type 2 diabetes mellitus}},
author = {Chuan Tian and Li Wang and Hong Pan and Lu Yang and Xi Chen},
journal = {Frontiers in Endocrinology},
year = {2026},
month = nov,
doi = {10.3389/fendo.2026.1886207},
url = {https://www.frontiersin.org/articles/10.3389/fendo.2026.1886207/full},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Risk factors and logistic regression-based nomogram prediction model for diabetic retinopathy in type 2 diabetes mellitus
AU - Chuan Tian
AU - Li Wang
AU - Hong Pan
AU - Lu Yang
AU - Xi Chen
PY - 2026
DA - 2026/11/16
JO - Frontiers in Endocrinology
DO - 10.3389/fendo.2026.1886207
UR - https://www.frontiersin.org/articles/10.3389/fendo.2026.1886207/full
AB - Introduction: Diabetic Retinopathy (DR) is a major cause of irreversible vision loss in Type 2 Diabetes Mellitus (T2DM). Early identification of high-risk patients using readily available clinical tools is crucial. This study aimed to analyze risk factors for DR in T2DM patients and develop a practical nomogram prediction model based on common clinical and laboratory parameters. Methods: A retrospective study was conducted on 379 T2DM patients (64 with DR) from a single center. LASSO regression (with the penalty λ chosen by the one-standard-error rule, λ.1se) was used for variable selection; the selected variables were entered into multivariable logistic regression and only those that remained independently significant were retained to construct the nomogram. Model performance was evaluated by discrimination (Area Under the Curve, AUC), calibration (Hosmer-Lemeshow test), and clinical utility (Decision Curve Analysis). Only internal validation (1000-sample bootstrapping) was performed; no external validation cohort was available. Results: Five independent risk factors were identified: longer disease duration (OR = 1.27 per year), hypertension (OR = 3.56), elevated fasting plasma glucose (OR = 1.86 per mmol/L), higher glycated hemoglobin (OR = 6.38 per %), and hypertriglyceridemia (triglycerides ≥1.7 mmol/L, OR = 6.37). The nomogram showed good discrimination (AUC 0.922, 95% CI 0.886-0.954; bootstrap optimism-corrected AUC 0.914), good calibration (Hosmer-Lemeshow P = 0.739; bootstrap-corrected calibration slope 0.93), and a positive net benefit on decision curve analysis. Discussion: This nomogram, incorporating five easily obtainable clinical variables, is an effective tool for predicting individual DR risk. Its practicality makes it suitable for initial risk stratification in primary care settings, enabling timely ophthalmology referrals. As a single-center model developed with a limited number of DR events and without external validation, its generalizability should be interpreted with caution and confirmed in larger, multi-center, externally validated cohorts.
SN - 1664-2392
ER -