Abstract
OBJECTIVE: To explore the association of diabetic retinopathy (DR) with TyG index and TyG-related parameters among the United States population. METHODS: This cross-sectional study is conducted in adults with diabetes mellitus based on the National Health and Nutrition Examination Survey (NHANES) from 2005 to 2018. Multivariate logistic regression, restricted cubic spline, trend test, receiver operating characteristic curve and subgroup analysis are adopted to uncover the association of DR with TyG index and TyG-related parameter levels in diabetics. RESULTS: An aggregate of 888 eligible participants with diabetes is included, involving 263 (29.6%) patients with DR. The participants are stratified according to the quartile of TyG index and TyG-related parameters (Q1-Q4). Following the adjustments of the confounding factors, a multivariate logistic regression analysis finds that TyG-BMI, TyG index and Q4-TyG index are significant risk factors for DR. The restricted cubic spline shows that TyG index and the DR risk of diabetes patients are proved to be U-shaped related (p for nonlinearity = 0.001). CONCLUSIONS: The triglyceride-glucose index has a U-shaped correlation with the risk of diabetic retinopathy, which has potential predictive value.