Endocrine Abstracts

Background: Recently many cross-sectional and a few longitudinal studies reported that thyroid hormones even within normal ranges are associated with measures of adiposity. We aim to assess if changes in TSH and Free T4 (FT4) over 10 years of follow-up would be associated with changes in BMI and waist circumference or risk of general and abdominal obesity.

Methods: 4179 participants of Tehran Thyroid Study(TTS) who had attended at baseline(1999–2001) and at three subsequent follow-ups every 4 years(up to 2011) were enrolled. At baseline, participants with serum TSH < 0.1 or >10 mU/l, BMI < 18.5 kg/m², pregnancy or history of thyroid medication or surgery and missing data were excluded. Finally, data of 2317 subjects remained for the study analysis. The median follow up time was 9.7y. Body weight, waist circumference were measured and serum concentrations of FT4 and TSH were assayed at baseline and three follow-ups. To account for within-subject correlation, the generalized estimating equation (GEE) was usedto assess the association between a1SD change in main exposures(CE. TSH and CE.FT4) and changes in BMI; calculated scores of CE.TSH and CE.FT4 were included in models as a time varying exposure.

Results: Mean age of the study population was 43 ± 13y, of whom 59.7% were women. Cumulative excess of TSH or FT4 were not associated with increased incidence of obesity, overweight and abdominal obesity. However, in over weights, after adjusting for age, sex, smoking, education, physical activity and HOMA-IR, for each standard deviation increase in the CE.FT4(β = −0.23,95% CI : –0.34, –0.11) and CE.TSH(β = 0.13,95% CI : 0.03, 0.22), BMI was changed negatively and positively, respectively[sd(CE.TSH) = 19.27, sd(CE.FT4) = 1.72)]. In GEE analysis, however, a one unit increase in TSH was associated with 0.02 increase in BMI in the total population(β = 0.02, 95% CI : 0.008, 0.03), an increase which remained significant only in women; a one unit increase in FT4 was significantly associated with 1.7 decrease in BMI(β = −1.7.95% CI : –1.9, –1.4) in the total population which remained significant in both sexes after adjusting for age, sex, smoking, physical activity, education and HOMA-IR. After entering both exposures of TSH and FT4 to the model, only FT4 independent of TSH was inversely associated with BMI. The one unit decrease in FT4 was also associated with increase in WC(β = −3.88, 95% CI : –5.09, –2.68) in both genders.

Conclusion: Cumulative effects in thyroid hormones over time indicated no risk for general or abdominal obesity. However TSH alterations in women and FT4 in both sexes were positively and negatively associated with BMI, respectively. FT4 variations could negatively affect WC in men and women independent of insulin resistance and even TSH.