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

Background The Leukocyte Glucose Index (LGI) reflects systemic inflammation and glycemic burden and has shown promising predictive value in cardiovascular and metabolic diseases. However, its association with prevalent diabetic retinopathy (DR) remains unclear. We investigated the association and dose–response relationship between LGI and prevalent DR in patients with T2DM and its incremental value beyond traditional clinical factors. Methods 1, 727 hospitalized patients with T2DM, including 661 with DR and 1, 066 without DR. Multivariable logistic regression, subgroup, trend, and restricted cubic spline analyses were performed to assess the association between LGI and prevalent DR. Variables selected by LASSO regression and multivariable logistic regression were used to develop a nomogram. Model performance was evaluated by discrimination, calibration, bootstrap internal validation, reclassification metrics, and decision curve analysis (DCA). Results Higher LGI was independently associated with prevalent DR (adjusted OR = 1.340, 95% CI: 1.056–1.701, P  = 0.016). Similar results were observed per standard deviation increase (adjusted OR = 1.166, 95% CI: 1.029–1.320) and in the highest versus lowest tertile (adjusted OR = 1.378, 95% CI: 1.028–1.848). A positive linear association was identified without evidence of nonlinearity. The nomogram achieved an AUC of 0.727 (95% CI: 0.703–0.752), with an optimism-corrected AUC of 0.719. Although the improvement in AUC after adding LGI was not statistically significant (0.727 vs. 0.725, P  = 0.198), NRI and IDI indicated improved discrimination, and DCA showed greater net clinical benefit across a range of threshold probabilities. Conclusions Elevated LGI was independently associated with prevalent DR and showed a positive linear dose–response relationship. LGI may provide incremental discriminative value beyond traditional clinical factors and serve as a simple laboratory marker for identifying prevalent DR. Further external validation and prospective studies are warranted.