Mediating effect analysis of visceral adiposity index on free triiodothyronine to free thyroxine ratio and non-alcoholic fatty liver disease in euthyroid population

BACKGROUND: The association between free triiodothyronine/free thyroxine (FT3/FT4) and non-alcoholic fatty liver disease (NAFLD) in euthyroid subjects is unclear. In addition, few studies have explored whether VAI mediates the association between FT3/FT4 ratio and NAFLD in the euthyroid population. We aimed to analyze the mediating effect of VAI on the FT3/FT4 ratio and NAFLD risk in the euthyroid population. METHODS: This cross-sectional study included 7 946 annual health examinees from the Health Examination Center, Hebei General Hospital, from January to December 2020. The basic information and biochemical parameters, as well as calculated FT3/FT4 ratio and VAI were collected. NAFLD was diagnosed according to abdominal ultrasonography. The fibrosis score for NAFLD positive subjects (NFS) was calculated to reflect the extent of liver fibrosis. The risk of NAFLD was analyzed by quartiles of FT3/FT4 ratio (Q1-Q4 quartiles) and VAI (V1-V4 quartiles), respectively. Pearson correlation analysis was performed to investigate the correlation between FT3/FT4 ratio and VAI. Multivariate logistic regression analysis was applied to analyze the effect of FT3/FT4 ratio and VAI on NAFLD and NFS status. Bootstrap was conducted to explore whether VAI mediated the association between FT3/FT4 ratio and NAFLD. RESULTS: Of the 7 946 participants, 2 810 (35.36%) had NAFLD and 5 136 (64.64%) did not. Pearson correlation analysis indicated that FT3/FT4 ratio was positively associated with VAI (P<0.05). Multivariate logistic regression analysis indicated that compared to the Q1 group, the risk of NAFLD significantly increased in Q3 group [OR=1.255, 95%CI (1.011, 1.559)] and Q4 group [OR=1.553, 95%CI (1.252, 1.926)](P<0.05). Compared to the V1 group, the risk of NAFLD notably increased in V2 group [OR=1.584, 95%CI (1.205, 2.083)], V3 group [OR=2.386, 95%CI (1.778, 3.202)] and V4 group [OR=4.104, 95%CI (2.835, 5.939)] (P<0.01). There was no relevance between FT3/FT4 ratio, VAI and NFS status. Mediating effect analysis showed that FT3/FT4 ratio significantly directly influenced NAFLD prevalence [β=3.7029, 95%CI (2.9583, 4.4474)], and VAI partly mediated the indirect effect of the FT3/FT4 ratio on NAFLD prevalence [β=2.7649, 95%CI (2.2347, 3.3466)], and the mediating effect accounted for 42.75% of the total effects. CONCLUSION: Both FT3/FT4 ratio and VAI were predictors of NAFLD, and VAI partly mediated the indirect effect of the FT3/FT4 ratio on NAFLD prevalence in the euthyroid population.