Thyroid function and life expectancy with and without noncommunicable diseases: A population-based study

Variations in thyroid function within reference ranges are associated with increased risk of diseases and death. However, the impact of thyroid function on life expectancy (LE) with and without noncommunicable diseases (NCDs) remains unknown. We therefore aimed to investigate the association of thyroid function with total LE and LE with and without NCD among euthyroid individuals.

The study was embedded in the Rotterdam Study, a prospective population-based study carried out in the Netherlands. In total, 7,644 participants without known thyroid disease and with thyroid-stimulating hormone (TSH) and free thyroxine (FT4) levels within reference ranges were eligible. NCDs were defined as presence of cardiovascular disease, diabetes mellitus type 2, or cancer. We used the demographic tool of multistate life tables to calculate LE estimates at the age of 50 years, using prevalence, incidence rates, and hazard ratios for three transitions (healthy to NCD, healthy to death, and NCD to death). The total LE and LE with and without NCD among TSH and FT4 tertiles were calculated separately in men and women. Analyses were adjusted for sociodemographic and cardiovascular risk factors. The mean (standard deviation) age of the participants was 64.5 (9.7) years, and 52.3% were women. Over a median follow-up of 8 years (interquartile range 2.7-9.9 years), 1,396 incident NCD events and 1,422 deaths occurred. Compared with those in the lowest TSH tertile, men and women in the highest TSH tertile were expected to live 1.5 years (95% confidence interval [CI] 0.8-2.3, p < 0.001) and 1.5 years (CI 0.8-2.2, p < 0.001) longer, respectively, of which 1.4 years (CI 0.5-2.3, p = 0.002) and 1.3 years (CI 0.3-2.1, p = 0.004) with NCD. Compared with those in the lowest FT4 tertile, the difference in LE for men and women in the highest FT4 tertile was -3.7 years (CI -5.1 to -2.2, p < 0.001) and -3.3 years (CI -4.7 to -1.9, p < 0.001), respectively, of which -1.8 years (CI -3.1 to -0.7, p = 0.003) and -2.0 years (CI -3.4 to -0.7, p = 0.003) without NCD. A limitation of the study is the observational design. Thus, the possibility of residual confounding cannot be entirely ruled out.

In this study, we found that people with low-normal thyroid function (i.e., highest tertile of TSH and lowest tertile of FT4 reference ranges) are expected to live more years with and without NCD than those with high-normal thyroid function (i.e., lowest tertile of TSH and highest tertile of FT4 reference ranges). These findings provide support for a re-evaluation of the current reference ranges of thyroid function.