Application of age- and sex-specific reference intervals for thyroid-stimulating hormone and free thyroxine in evaluating incidence and trends of thyroid dysfunctions : a population-based cohort study

Abstract

Background: Diagnosis of thyroid dysfunctions heavily relies on the biochemical tests for thyroid-stimulating hormone (TSH) and free thyroxine (FT4). Conventional laboratory-specific reference intervals (RIs) for TSH and FT4 are uniform across adults and ignore age-, sex-, and time-related variations, which may lead to misdiagnosis of thyroid dysfunction. This study aimed to establish the age- and sex-specific RIs for TSH and FT4, and to evaluate their implications on the incidence and secular trends of thyroid dysfunctions. Methods: Using a population-based electronic health record database in Hong Kong, we identified individuals without thyroid-related disorders who had valid TSH (N = 2,111,661) and FT4 (N = 825,522) measurements from 2006 to 2019. After harmonizing data across multiple institutions, we derived age- and sex-specific lower and upper reference limits (LRLs and URLs) for TSH and FT4 using the 2.5th and 97.5th percentiles of the biomarker distribution in seven age groups for both females and males. In comparison with the laboratory-specific RIs, we reclassified the thyroid status of the cohort individuals. We also compared the standardized incidence rates and secular trends in thyroid dysfunctions defined by the age- and sex-specific RIs and the conventional RIs. Results: TSH and FT4 distributions varied by age, sex, and calendar year. Age- and sex-specific RIs for TSH and FT4 were significantly wider among elderly individuals (≥ 70 years) compared to the youngest adults (18–29 years). Among those with both TSH and FT4 measurements, 30.78% were classified as having thyroid dysfunctions using laboratory-specific RIs, whereas applying age- and sex-specific RIs reduced this proportion to 18.57%. The use of age- and sex-specific RIs was associated with lower observed incidence rates and revealed estimates of incidence and secular trends that are likely closer to the underlying disease burden. Conclusions: Age- and sex-specific RIs could mitigate misdiagnosis of thyroid dysfunctions, thereby improving diagnostic accuracy and enabling more precise monitoring of secular trends. The secular trends and joinpoints identified for the age- and sex-specific URLs and LRLs of TSH and FT4 from 2006 to 2019 indicate that regular RI review is needed to ensure their continued clinical relevance. Future validation in geographically and ethnically diverse populations is warranted.