Abstract
AIMS: This study aimed to investigate the potential non-linear relationships between thyroid hormones and thyroid-stimulating hormone levels on glycemic control levels. METHODS: A retrospective analysis of electronic medical records was performed on patients with T2DM who received treatment at a tertiary care hospital in Chengdu, Sichuan Province, between 2018 and 2023. RCS regression and threshold effect analyses were employed to assess potential nonlinear associations among THs, TSH, and glycemic control. RESULTS: Data from a total of 1413 patients were included in the analysis. RCS regression revealed a significant non-linear association between FT3 and HbA1c (p for nonlinearity < 0.05). Threshold effect analysis demonstrated no statistically significant inflection point for FT3 in relation to HbA1c (FT3 < 5.92 pmol/L, β = 0.008, 95% CI (-0.122, 0.281); FT3 > 5.92 pmol/L, β = -0.109, 95% CI (-0.261, 0.042)). FT4 exhibited a significant non-linear relationship with HbA1c (p for nonlinearity < 0.05), identifying an inflection point at 14.82 pmol/L. Below this threshold, each 1 pmol/L increase in FT4 was associated with a 0.263% elevation in HbA1c (β = 0.263, 95% CI 0.189-0.337). Similarly, TSH demonstrated a non-linear association with HbA1c (p for nonlinearity < 0.05), with an inflection point identified at 5.53 mIU/L. When TSH was below 5.53 mIU/L, each 1 mIU/L increase was associated with a 0.179% reduction of HbA1c (β = -0.179, 95% CI -0.281 to -0.076). CONCLUSION: Nonlinear associations were observed between thyroid function markers (FT3, FT4, and TSH) and HbA1c levels in patients with T2DM. These findings provide novel evidence for understanding thyroid-glucose metabolic interactions.