Abstract
KEY POINTS
Age-related changes in thyrotropin (TSH) and free thyroxine (FT4) are generally small, with the dominant pattern within individuals being stable thyroid function rather than gradual change over time.
Any pattern of change in TSH and FT4 concentrations is associated with differential all-cause mortality risk.
SUMMARY
Background
Current evidence on the natural course of thyroid function tests across the life span is conflicting with regard to clinical implications, and research designs often included in cross-sectional analyses. This study, 1 a large individual participant data (IPD) meta-analysis including prospective population-based cohorts, examined age-related thyroid function test changes both cross-sectionally and using longitudinal data and the association with all-cause mortality.
Methods
Included were prospective population-based cohorts with repeated thyroid function tests and mortality data for adults. Data were collected from 31 cohorts with measurements between 2011 and 2022. Exclusion criteria were use of thyroid-altering medication, known thyroid disease, and pregnancy. The iodine status of cohorts was assessed using urinary iodine concentrations or, if unavailable, using time- and region-specific reports. Mixed-effect models estimated longitudinal changes in thyrotropin (TSH) and free thyroxine (FT4), categorized into quintiles of change (increasing, stable, decreasing). Linear mixed models were used for analyses, with the cohort as a random intercept and the individual as a random intercept in cases of longitudinal data. Cox models were used to evaluate associations between combined TSH/FT4 change patterns and all-cause mortality.
Results
There were 137,488 participants from 31 cohorts included, predominantly based in Europe and the United States. In cross-sectional analyses, the mean age-related variation in thyroid function was modest from 18 to 100 years (increase of 0.21 mIU/L), and most individuals maintained relatively stable TSH and FT4 levels over time. In longitudinal analyses, TSH increased by 0.61 mIU/L in females and 0.99 mIU/L in males across the life span in iodine-sufficient regions and 0.85 mIU/L for female participants and 0.99 mIU/L for male participants in iodine-insufficient regions. On average, smokers had lower TSH concentrations, whereas participants with a BMI over 30 had slightly higher TSH concentrations. Population distribution variability of TSH was larger in participants over 65 years of age. FT4 concentrations were higher with older age in cross-sectional analyses, while in longitudinal analyses, the change differed by iodine status, with an increase in average FT4 concentrations in iodine-sufficient regions and a decrease in iodine-insufficient regions. Any pattern of change in combined TSH and FT4 longitudinal analyses was associated with increased all-cause mortality (hazard ratios varied from 1.80 to 2.45 depending on pattern).
Conclusions
In a large international IPD meta-analysis, increasing age was associated with relatively small average changes in thyroid function tests, and individual trajectories were relatively stable. Any individual change in thyroid function pattern was associated with differential all-cause mortality risk. Future research could refine reference intervals and explore underlying mechanisms as well as clinical utility for identifying persons at elevated risk based on thyroid function trajectories.
COMMENTARY
This large IPD meta-analysis by Xu and colleagues provides important new data on the evolution of thyroid function across adulthood and aging and contributes to a long-standing debate in thyroidology: whether higher TSH concentrations in older individuals primarily reflect physiological aging or occult thyroid disease. Major strengths of the study include the harmonized analysis of multiple prospective cohorts, the large sample size, and the availability of repeated thyroid function measurements, allowing assessment of individual trajectories over time. Limitations include the lack of T3 measurements and incomplete data on potentially important confounders such as comorbidity and lifestyle factors, which limit the mechanistic interpretation of the observed associations.
The authors demonstrate that average age-related changes in TSH and FT4 are relatively modest and that most individuals maintain stable thyroid function throughout adult life, while variation in TSH distribution and the TSH upper reference limit was higher for the cohort over 65 years of age, supporting earlier cross-sectional analysis.2,3 These findings are particularly relevant in light of current recommendations by the European and American Thyroid Associations, which both recommend broader reference intervals for individuals over 70 years of age.4,5 The present study provides a nuanced understanding of age-related changes by showing that stable thyroid function rather than gradual change is the dominant individual pattern. This suggests that the higher upper TSH reference limits observed in elderly populations may reflect a subgroup of individuals experiencing changes related to comorbidity, environmental exposures, or evolving thyroid disease rather than a universal physiological shift affecting all older adults equally.
Another important result is the evaluation of combined longitudinal changes in TSH and FT4 in relation to mortality. Any change in trajectory was associated with higher all-cause mortality risk, irrespective of direction. These associations should be interpreted cautiously and do not imply causality; however, the changes in thyroid hormone concentrations in older individuals may represent markers of underlying illness, inflammation, or cardiovascular disease rather than direct causes of adverse outcomes. This could also (in part) be the cause of the negative results of randomized controlled trials assessing the clinical benefit of treatment of high TSH in the elderly.6,7
It remains uncertain whether elevated TSH concentrations in older adults reflect primary thyroid disease or underlying comorbidity. The current study highlights the uncertainties regarding thyroid function and aging and supports that interpretation should be individualized rather than based solely on age-adjusted reference intervals. From a clinical perspective, a slightly elevated TSH concentration should neither automatically prompt treatment nor simply be dismissed as a consequence of normal aging. Repeat testing, ideally compared with prior baseline measurements, remains the most reliable approach to differentiate persistent thyroid dysfunction from transient thyroid function test fluctuations.