Threshold effect and age interaction of TyG index on diabetes incidence in normolipidemic population: a multicenter cohort study.

Background: Although the triglyceride-glucose (TyG) index-diabetes association has been widely studied, its relationship in normolipidemic populations remains poorly understood. Methods: A retrospective cohort of 60,103 normolipidemic Chinese adults was included from routine health screening programs conducted across 32 healthcare institutions in China. Data collection included demographic characteristics, anthropometric measurements, serum biochemical parameters, smoking and alcohol consumption history, and family history of diabetes. We employed multivariable Cox regression, restricted cubic spline analysis, threshold effect analysis, stratified analysis, and interaction tests to comprehensively assess the association between the TyG index and incident diabetes. Results: Multivariable-adjusted Cox regression revealed a robust positive association between the TyG index and incident diabetes in normolipidemic subjects (HR: 10.10, 95% CI: 7.94-12.84, P < 0.001). Restricted cubic spline analysis detected a nonlinear relationship, with a critical threshold at TyG ≥ 8.53, beyond which diabetes risk increased exponentially (HR: 51.84, 95% CI: 24.83-108.24, P < 0.001). Despite consistent findings across subgroups, a significant interaction with age was detected (P for interaction < 0.05). Conclusions: In normolipidemic individuals, the TyG index demonstrated a nonlinear positive association with diabetes risk, particularly above 8.53. It can serve as an early warning signal for diabetes risk in normolipidemic individuals, facilitating personalized prevention strategies for diabetes prevention and control.