Endocrine Abstracts

Introduction: Several studies have suggested an association of thyrotropin (TSH) and free thyroxine (FT4) levels within the reference range with morbidity and mortality in the general population. Low free triiodothyronine (FT3) has also been associated with a poor prognosis in several conditions. However, the association between FT3 levels within the reference range and the risk of mortality in the general population remains uncertain.

Methods: We evaluated the association between FT3 levels and mortality in 6672 adults in the National Health and Nutrition Examination Survey (NHANES) 2001–2002, 2007–2008, and 2009–2010 cycles. Thyroid function was assessed at baseline. The mortality and cause of death were prospectively evaluated until December 31, 2011. Patients with a prior history of thyroid disease, patients receiving thyroid-related drugs, and those with TSH, FT4 or FT3 outside the reference range were excluded. We categorized participants according to FT3 tertiles. We used Cox proportional hazard models to estimate hazard ratios (HR) for all-cause mortality, cardiovascular mortality, cancer-related mortality and mortality from other causes. The models were fitted unadjusted and with adjustment for potential confounders: age, gender, race, BMI, smoking, education, annual salary, diabetes mellitus, hypertension, dyslipidemia, chronic kidney disease, previous cardiovascular disease, history of cancer, FT4 and TSH levels.

Results: During the follow-up period (28901 person-years, median follow-up: 45 months), a total of 353 deaths occurred (81 of cardiovascular disease, 97 of cancer and 175 of other causes). Compared to participants with FT3 levels in the upper tertile (3.3–3.9 pg/ml), participants with FT3 levels in the lower tertile (2.5–3.0 pg/ml) had higher all-cause mortality (HR 3.06 (1.71–5.45), P<0.001), cardiovascular mortality (HR 11.55 (3.92–34.04), P<0.001) and mortality from other causes (HR 2.90 (1.52–5.52), P=0.002). In the adjusted analysis, this association was not significant for all-cause mortality (HR 1.25 (0.66–2.38), P=0.479), cancer mortality (HR 1.09 (0.41–2.90), P=0.851) and mortality from other causes (HR 0.98 (0.41–2.38), P=0.969). On the other hand, compared to the highest FT3 tertile, the lowest FT3 tertile was associated with higher cardiovascular mortality, even after adjusting for confounders (HR 6.23 (1.66–23.37), P=0.008).

Conclusion: Low levels of FT3 within the reference range independently predict cardiovascular mortality. Our results suggest that FT3 levels may contribute to the stratification of cardiovascular risk in the general population.