Non-linear associations of the triglyceride-glucose index with thyroid function in a large Chinese population: a cross-sectional study.

BACKGROUND: Previous studies have identified the triglyceride-glucose (TyG) index as a practical surrogate marker of insulin resistance and have suggested its association with several metabolic disorders. However, the association between the TyG index and thyroid function parameters has not yet been fully clarified. The present study was conducted to examine this association in Chinese adults from the general population. METHODS: This cross-sectional study included 5,189 adults who underwent health examinations. Multivariable linear regression was performed to investigate the associations between the TyG index and thyroid function parameters, including free triiodothyronine (FT3), free thyroxine (FT4), and thyroid-stimulating hormone (TSH). Potential non-linear dose-response associations were explored by using restricted cubic spline (RCS) analyses, while threshold effects were evaluated using piecewise linear regression. Subgroup analyses were conducted to examine whether the associations varied across subgroups, and sensitivity analyses were performed in strictly euthyroid participants. RESULTS: After full adjustment for potential confounders, the TyG index showed no clear linear association with FT3, whereas it was inversely associated with FT4. Non-linearity analyses indicated a significant inverted U-shaped association of the TyG index with FT3 (P for non-linearity < 0.001), the TyG index was positively associated with FT3 below the inflection point of 8.311 and negatively associated with FT3 above this point. In contrast, piecewise linear regression suggested a threshold effect for FT4 (P for likelihood ratio test = 0.016), with the TyG index showing a positive association below 7.591 and a negative association above this point. Additionally, the TyG index was positively associated with lnTSH after full adjustment. Subgroup analyses suggested potential effect modification for FT3 by sex, FT4 by thyroid status, and TSH by age. The non-linear patterns for FT3 and FT4 remained generally similar in sensitivity analyses restricted to strictly euthyroid participants. CONCLUSION: This study observed complex linear and non-linear associations between the TyG index and thyroid function parameters. These findings suggest a potential metabolic-thyroid association that warrants further evaluation in longitudinal studies.