Abstract
BACKGROUND AND AIMS: Acquired resistance to thyroid hormone appears to exist in the general population. We aimed to evaluate the association between indices of thyroid hormone sensitivity and non-alcoholic fatty liver disease (NAFLD), and made stratified analyses by diabetic status. METHODS: We included 26,413 participants from a health screening program and 8,246 hospitalized patients with type 2 diabetes. Thyroid Feedback Quantile-based Index (TFQI), thyroid stimulating hormone index (TSHI) and thyrotroph thyroxine resistance index (TT4RI) were calculated. Advanced fibrosis risk was determined using the FIB-4 score. Multivariate logistic regression analysis was performed. RESULTS: TFQI was associated with an increased risk of NAFLD in patients with diabetes (fourth quartile vs. first quartile: odds ratio [OR]=1.39 and 1.82 in hospitalized and non-hospitalized patients, respectively, both P<0.001) but not non-diabetic participants (OR=0.94, P=0.40). Further adjustment for the homeostasis model assessment of insulin resistance generated similar findings in diabetes (OR=1.27, P=0.025). The TFQI-associated NAFLD risk increase in diabetic patients was confined to NAFLD with low probability of advanced fibrosis (OR 1.42, P=0.001), but not those with intermediate-to-high probability (OR=0.86, P=0.23). Also, TFQI was associated with a significantly lower risk for advanced fibrosis in the diabetic at-risk patients (OR=0.62, P=0.005) but not those non-diabetic at-risk participants, independent of the presence of NAFLD. The association was less significant for TT4RI and TSHI. CONCLUSIONS: Impaired sensitivity to thyroid hormone was associated with an increased risk of developing NAFLD but a reduced risk of advanced fibrosis limited to diabetic individuals. Our findings suggest stratified studies of NAFLD based on diabetic status are needed in the future.