Abstract
OBJECTIVE: To evaluate the association between the TyG index and DR while critically assessing the predictive performance of this composite index. Specifically, we aimed to quantify the independent association between TyG and DR in different subgroups and determine the practical utility of TyG for clinical DR prediction. METHODS: This cross-sectional study included 1,761 patients with T2DM visiting the National Metabolic Management Center. Biochemical parameters and results from fundus screening examinations were collected. Associations between the TyG index and DR were examined using multivariate logistic regression, smooth curve modeling, and stratified subgroup analyses. Receiver operating characteristic (ROC) curve analysis was conducted to evaluate the predictive performance of the TyG index. RESULTS: The mean TyG index was significantly higher among patients with DR compared to those without DR. In multivariate regression, each one-unit increase in the TyG index was associated with a 1.57-fold higher likelihood of DR. Stratified analysis indicated that this association was more pronounced in female patients, younger patients, and those with a shorter disease duration. Smooth curve analysis demonstrated a nonlinear relationship, with a notable increase in DR risk when the TyG index exceeded 9.5. ROC curve analysis indicated an area under the curve of 0.5523 for the TyG index in predicting DR, with a specificity of 92.4% and a sensitivity of 28.2% at a cutoff value of 10.24. CONCLUSION: Although the TyG index is statistically associated with DR (OR = 1.57, p<0.0001), its extremely poor discriminatory ability (AUC = 0.5523, sensitivity=28.16%) renders it clinically unsuitable for DR screening, diagnosis. These findings suggest that predicting DR requires multi-parametric assessment incorporating metabolic, endocrine, inflammatory, and vascular markers, rather than reliance on a single composite index.