Association between thyroxine level and nonalcoholic fatty liver disease

ObjectiveTo investigate the association between thyroxine level and nonalcoholic fatty liver disease (NAFLD). MethodsA retrospective analysis was performed for 3289 subjects who underwent physical examination in PLA Rocket Force Characteristic Medical Center from July 2015 to April 2019, and according to medical history and thyroid function, they were divided into subclinical hypothyroidism group with 210 subjects and normal thyroid function group with 3079 subjects. According to the results of abdominal color Doppler ultrasound, the normal thyroid function group was divided into NAFLD group with 516 subjects and non-NAFLD group with 2563 subjects; according to body mass index (BMI), the normal thyroid function group was divided into non-obese group (BMI ＜25 kg/m2) and obese group (BMI ≥25 kg/m2); according to the age, the normal thyroid function group was divided into elderly group (age ≥60 years) and young and middle-aged group (age ＜60 years). The normal thyroid function group was typed based on age and body type. Related data were collected, including sex, age, BMI, blood pressure, waist circumference, fasting blood glucose, uric acid, total cholesterol, triglyceride, high-density lipoprotein, low-density lipoprotein, free triiodothyronine, free thyroxine, triiodothyronine, thyroxine, and thyroid stimulating hormone (TSH). The two-independent-samples t test was used for comparison of normally distributed continuous data between two groups, and the Mann-Whitney U test was used for comparison of non-normally distributed continuous data between two groups; the chi-square test used for comparison of categorical data between two groups. A multivariate logistic regression analysis was used to investigate risk factors, and the receiver operator characteristic (ROC) curve was used to analyze the cut-off values of related indices in predicting NAFLD. ResultsThe subclinical hypothyroidism group had a higher prevalence rate of NAFLD than the normal thyroid function group (22.38% vs 16.76%，χ2=4.380，P=0.036), and in the subclinical hypothyroidism group, the NAFLD patients had a higher level of TSH than the non-NAFLD patients（Z=-1.994,P=0.046）. In the subclinical hypothyroidism group, there were no significant differences in thyroid parameters between the NAFLD group and the non-NAFLD group (all P＞0.05); after stratification based on age and body type, in the obese-young and middle-aged subgroup, male sex, low free thyroxine, fasting blood glucose, and triglyceride were independent risk factors for NAFLD (odds ratio=4.729, 0.067, 1.814, and 1.717, P=0.003, 0.010, 0.011, and 0.014). The cut-off values of free thyroxine, fasting blood glucose, and triglyceride were 1.123 ng/dL, 5.15 mmol/L, and 1.02 mmol/L, respectively, in predicting NAFLD, and the area under the ROC curve was 0.832 for combined prediction. ConclusionThere is a high prevalence rate of NAFLD in the population with subclinical hypothyroidism, and when thyroid function is within the normal range, low free thyroxine is associated with the onset of NAFLD in the young and middle-aged obese people