Natural history of thyroid function in ageing: an individual participant data analysis of 137 488 participants from 31 prospective cohort studies

Abstract

BACKGROUND: Evidence regarding thyroid function changes with ageing remains inconsistent and the implications of potential changes are unclear. We aimed to investigate ageing-related thyroid function changes and their associations with mortality. METHODS: In this individual participant data (IPD) analysis, prospective population-based cohorts were eligible for inclusion when data on thyroid function measurements and mortality were available in individuals aged 18 years and older. Eligible datasets were identified through a systematic search of PubMed. We excluded cohorts of participants with only thyroid disease or thyroid-altering medications, or pregnant individuals. We requested data from all eligible cohorts that agreed to participate in the study. Linear mixed models were used to investigate associations between age and thyroid function, stratified for sex and regional iodine status. Annual changes in thyroid-stimulating hormone (TSH) and free thyroxine (FT4) were estimated per individual and categorised into quintiles, with the highest and lowest quintiles defined as increasing and decreasing, respectively, and the rest as stable. Patterns of thyroid function change were identified based on combined TSH and FT4 evolution. We used cohort-stratified Cox models to assess associations between changing patterns and all-cause mortality. This study is registered with PROSPERO, CRD42023408086. FINDINGS: In this IPD analysis, we analysed data collected between Jan 1, 2011, and Oct 13, 2022, from 31 cohorts across Europe (n=19), the USA (n=5), Asia (n=3), Brazil (n=2), and Australia (n=2; 137 488 participants; 68 322 [49·7%] were female and 69 166 [50·3%] were male; median age 60 years [range 18-106]). Cross-sectionally, older age was associated with higher TSH in iodine-sufficient regions and with lower TSH in iodine-insufficient regions. Longitudinal analyses showed that TSH increased with increasing age regardless of iodine status. The overall increase in TSH from age 18 years to 100 years was 0·61 mIU/L (0·52 SD) for female participants and 0·99 mIU/L (0·76) for male participants from iodine-sufficient regions. Greater variability in population distribution and longitudinal TSH changes was observed in adults aged 65 years or older. Higher FT4 with older age was suggested cross-sectionally, but longitudinally FT4 increased in iodine-sufficient regions and decreased in iodine-insufficient regions. Compared with stable thyroid function, all changing patterns were associated with increased all-cause mortality: hazard ratios of 1·80 (95% CI 1·57-2·06) for increasing TSH with stable or decreasing FT4; 2·45 (2·01-2·97) for increasing TSH and increasing FT4; 2·45 (1·99-3·01) for decreasing TSH with decreasing FT4; and 1·94 (1·68-2·24) for decreasing TSH with stable or increasing FT4. INTERPRETATION: Ageing-related changes in thyroid function varied by sex and iodine status. Most individuals had stable thyroid function during ageing with a slight increase in TSH, although older adults displayed greater variability. Patterns of changing thyroid function were associated with an increased all-cause mortality risk, warranting further exploration of the underlying mechanisms and clinical management.None.