LAUSR

Objective The purpose of this study was to evaluate the association between triglyceride glucose (TyG) index and new-onset diabetes under different glycemic states and to compare the predictive value of TyG−related parameters, obesity indices, and lipid ratios for new-onset diabetes. Methods Data were collected from the China Health and Retirement Longitudinal Study (CHARLS), consisting of 6,258 participants aged ≥45 years. Participants were grouped according to their glycemic states. Cox proportional hazards models and restricted cubic spline regression were used to explore the association between TyG index and diabetes. Cox proportional hazard models were applied to confirm the predictive value of the optimal marker. Receiver operating characteristic (ROC) curves were used to compare the predictive value. Results TyG index was positively correlated with the risk of diabetes (hazard ratio (HR), 1.75; 95% confidence interval (CI), 1.56–1.97), and the linear association existed (p < 0.001). The highest correlation with diabetes was visceral adiposity index (VAI) (HR, 2.04; 95% CI, 1.44–2.90) in normal fasting glucose (NFG) group and TyG-body mass index (TyG-BMI) (HR, 2.53; 95% CI, 1.97–3.26) in impaired fasting glucose (IFG) group. The largest area under curve (AUC) was observed in TyG-waist-to-height ratio (TyG-WHtR) in the NFG group (AUC, 0.613; 95% CI, 0.527–0.700), and TyG-BMI had the highest AUC in the IFG group (AUC, 0.643; 95% CI, 0.601–0.685). Conclusion The association between TyG index and new-onset diabetes was positive and linear. TyG-WHtR was a clinically effective marker for identifying the risks of diabetes in the NFG group and TyG-BMI was an effective marker to predict diabetes in the IFG group.