Supplementary Material for: Thyroid Function and High-Sensitivity C-Reactive Protein in Cross-Sectional Results from the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil): Effect of Adiposity and Insulin Resistance
datasetposted on 02.09.2016, 10:53 by Peixoto de Miranda É.J.F., Bittencourt M.S., Santos I.S., Lotufo P.A., Benseñor I.M.
Background: Subclinical hypothyroidism (SCH) is associated with an increased cardiovascular risk, but little information is available about its association with high-sensitivity C-reactive protein (hs-CRP). Objectives: This study aims to analyze the association between SCH and hs-CRP using baseline data from the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil). Methods: The study has a cross-sectional design. We included subjects with normal thyroid function (thyroid-stimulating hormone, TSH, 0.4-4.0 μIU/ml and normal free thyroxine, fT4, 10.3-24.45 pmol/l) and SCH (TSH >4.0 μIU/ml and normal fT4) who were evaluated for hs-CRP. We excluded individuals on medications that affect thyroid function and those who had prevalent cardiovascular disease. Multivariate linear regression evaluated hs-CRP and TSH as continuous variables, and logistic regression models assessed hs-CRP ≥19.05 nmol/l as the dependent variable and crescent quintiles of TSH as the independent variables adjusted for demographic and cardiovascular risk factors. Results: We included 12,284 subjects with a median age of 50 years (interquartile range = 45-57); 6,408 (52.2%) were female, 11,589 (94.3%) were euthyroid, and 695 (5.7%) had SCH. Bivariate analyses of participants stratified into quintiles of TSH revealed differences according to hs-CRP but not the Framingham risk score. The fifth quintile of TSH was not associated with elevated hs-CRP, odds ratio = 1.11 (95% confidence interval = 0.98-1.26), p = 0.102, in a fully adjusted logistic model, also consistent with the linear model (β = 0.024, p = 0.145). Conclusions: TSH is not associated with hs-CRP. Obesity and insulin resistance are very important confounders in the study of the association between SCH and hs-CRP.