The leukocyte glucose index: a novel inflammatory-glucose biomarker for prevalent diabetic retinopathy and its supplementary predictive capacity in individuals with type 2 diabetes mellitus.
Huang S, Wang L, et al. • Frontiers in endocrinology • 2026
Elevated Leukocyte Glucose Index (LGI) was independently associated with prevalent diabetic retinopathy and showed a positive linear dose-response relationship, providing incremental discriminative value beyond traditional clinical factors in patients with type 2 diabetes mellitus.
Key Findings
Results
Higher LGI was independently associated with prevalent diabetic retinopathy in patients with type 2 diabetes mellitus.
Adjusted OR = 1.340 (95% CI: 1.056–1.701, P = 0.016) for higher LGI category
Per standard deviation increase in LGI: adjusted OR = 1.166 (95% CI: 1.029–1.320)
Highest versus lowest tertile comparison: adjusted OR = 1.378 (95% CI: 1.028–1.848)
Association was identified after multivariable logistic regression adjusting for traditional clinical factors
Study included 1,727 hospitalized patients with T2DM: 661 with DR and 1,066 without DR
Results
The association between LGI and prevalent diabetic retinopathy followed a positive linear dose-response relationship without evidence of nonlinearity.
Restricted cubic spline analysis was used to assess the dose-response relationship
No statistically significant nonlinear trend was detected
Trend analysis confirmed a consistent positive linear association across the range of LGI values
Both trend and spline analyses were performed as part of the statistical assessment
Results
A nomogram incorporating LGI and variables selected by LASSO regression achieved an AUC of 0.727 for predicting prevalent diabetic retinopathy.
AUC = 0.727 (95% CI: 0.703–0.752)
Optimism-corrected AUC after bootstrap internal validation = 0.719
Variables were selected using LASSO regression followed by multivariable logistic regression
Model performance was evaluated by discrimination, calibration, bootstrap internal validation, reclassification metrics, and decision curve analysis (DCA)
Results
Adding LGI to the prediction model did not produce a statistically significant improvement in AUC, but reclassification metrics and decision curve analysis indicated incremental benefit.
AUC with versus without LGI: 0.727 vs. 0.725, P = 0.198 (not statistically significant)
Net Reclassification Improvement (NRI) indicated improved discrimination after adding LGI
Integrated Discrimination Improvement (IDI) also indicated improved discrimination
Decision curve analysis (DCA) showed greater net clinical benefit across a range of threshold probabilities when LGI was included
Results
Subgroup analyses were performed to assess the consistency of the association between LGI and prevalent diabetic retinopathy across different patient subgroups.
Subgroup, trend, and restricted cubic spline analyses were all conducted
The study population consisted of 1,727 hospitalized patients with T2DM
Similar results were observed across analyses using different LGI parameterizations (continuous per SD, tertile comparison, and higher category)
The authors noted that further external validation and prospective studies are warranted
Background
The Leukocyte Glucose Index is described as a biomarker reflecting both systemic inflammation and glycemic burden, with prior associations reported in cardiovascular and metabolic diseases.
Prior literature has shown promising predictive value of LGI in cardiovascular and metabolic diseases
Its association with prevalent diabetic retinopathy had not been previously established
The authors characterize LGI as 'a simple laboratory marker' that may be useful for identifying prevalent DR
What This Means
This research suggests that a combined blood test measure called the Leukocyte Glucose Index (LGI) — which takes into account both white blood cell count (a marker of inflammation) and blood glucose levels — is linked to the presence of diabetic retinopathy (eye damage caused by diabetes) in people with type 2 diabetes. The study examined 1,727 hospitalized patients with type 2 diabetes, of whom 661 had diabetic retinopathy. People with higher LGI values were roughly 34% more likely to have diabetic retinopathy compared to those with lower values, and this relationship was consistent and linear — meaning the higher the LGI, the greater the association with retinopathy.
The researchers also built a prediction tool (called a nomogram) using LGI alongside other clinical factors, which achieved reasonable accuracy in identifying who had diabetic retinopathy (AUC of 0.727). While adding LGI to the model did not dramatically improve accuracy on one measure (AUC), other statistical tests (NRI and IDI) and a clinical usefulness analysis (decision curve analysis) suggested that including LGI does provide additional value in distinguishing patients with and without diabetic retinopathy beyond what traditional clinical information alone can offer.
This research suggests that LGI, being derived from routine blood test results, could serve as a practical and accessible screening marker to help identify diabetic patients who may have retinopathy. However, the authors note that this was a single-center study of hospitalized patients, and that external validation in other populations and prospective (forward-looking) studies are needed before the LGI can be broadly recommended as a clinical tool.
Huang S, Wang L, Zhang C, Zheng R, Sun X, Li J, et al.. (2026). The leukocyte glucose index: a novel inflammatory-glucose biomarker for prevalent diabetic retinopathy and its supplementary predictive capacity in individuals with type 2 diabetes mellitus.. Frontiers in endocrinology. https://doi.org/10.3389/fendo.2026.1959346