SAT-447 Thyroid Function, Cardiovascular Disease, and Mortality: A Mediation Analysis

Abstract

Background: Subclinical hypothyroidism is a common clinical entity among the United States (U.S) adults and has been associated with increased risk of cardiovascular disease (CVD) and mortality in some studies. However, the mediation effect of CVD from elevated serum thyroid stimulating hormone (TSH) to mortality has not yet been well established or sufficiently quantified. In this study, we aimed to elucidate the extent to which subclinical hypothyroidism or high-normal thyroid stimulating hormone [TSH] concentrations (i.e. upper normative-range TSH concentrations) contribute to mortality through its effect on CVD among U.S. adults. Methods: This study relies on the U.S. representative samples of 9,020 adults enrolled in the National Health and Nutrition Examination Surveys (NHANES) 2001-2002, 2007-2012 and their mortality data through 2015. We employed Cox proportional hazards regression models to investigate associations between the TSH concentration categories (subclinical hypothyroidism or tertiles of serum TSH concentrations within the reference range) and all-cause mortality. Utilizing mediation analysis within the counterfactual framework, we estimated the mediation effect of CVD on the association between TSH and all-cause mortality. Results: The median duration of follow-up for mortality ascertainment was 7.3 (interquartile range, 5.4–8.3) years, during which 435 deaths from all causes were identified. Subjects with TSH in the subclinical hypothyroidism and the high-normal TSH tertile concentrations were associated with increased all-cause mortality (subclinical hypothyroidism: hazard ratio [HR], 1.90; 95% confidence interval [CI], 1.14–3.19; high-normal TSH: HR, 1.36; 95% CI, 1.07–1.73) compared with the middle-normal TSH tertile concentration. CVD mediated 14.3% and 5.9% of the effects of subclinical hypothyroidism and high-normal TSH on all-cause mortality, respectively, with the CVD mediation effect being most pronounced in women and subjects ≥60 years old. No mediation through CVD was found for low-normal TSH levels and all-cause mortality. Conclusion: CVD mediated the effects of subclinical hypothyroid and high-normal TSH concentrations on all-cause mortality in the U.S. general population. These findings may have potential implications for treatment, and early and more aggressive CVD screening for such at risk populations. Further studies are needed to examine the clinical impact of targeted therapy towards mid-normal TSH concentration.