Thyroid Function Reference Intervals by Age, Sex, and Race A Cross-Sectional Study

Background: Current clinical practice uses a one-size-fits-all approach to define reference intervals for the results of diagnostic tests about thyroid function. This approach does not recognize subgroup differences according to age, sex, or race. Objective: To identify age-, sex-, and race-specific reference intervals for the common diagnostic tests that measure thyroid function and to examine how these new reference intervals reclassify persons into disease categories when compared with current reference intervals. Design: Cross-sectional analysis. Setting: Data from the U.S. NHANES (National Health and Nutrition Examination Survey) supplemented with data from a multicenter Chinese study. Participants: A nationally representative sample from NHANES aged 20 years or older (n = 8308) supplemented with a Chinese database of routine health checkups from 49 hospitals in 10 provinces aged 18 years or older (n = 314 302). Measurements: The thyroid function reference interval was defined as the interval of diagnostic indicator levels from the 2.5th (lower limit) to the 97.5th (upper limit) percentile by age, sex, and race subgroups. Results: In 8308 NHANES participants, the 97.5th percentile levels of thyroid-stimulating hormone (TSH) increased with age, whereas total triiodothyronine (TT3) levels declined with age and total thyroxine (TT4) levels were stable across different ages. Women had higher TT4 levels, and White participants had higher TSH levels. Using current reference intervals, the prevalence of subclinical hypothyroidism increased from 2.4% for ages 20 to 29 years to 5.9% for ages 70 years and older. In contrast, using age-, sex-, and race-specific reference intervals reclassified 48.5% of persons with subclinical hypothyroidism as normal, especially women and White participants, and reclassified 31.2% of persons with subclinical hyperthyroidism as normal, especially women, Black participants, and Hispanic participants. When compared with the findings from U.S. participants, many of the findings from 314 302 Chinese participants were similar. Limitation: Cross-sectional data; sample size limitations for subgroup. Conclusion: These findings should help establish more accurate reference intervals for thyroid diseases and facilitate development of a consensus about how to define and manage those diseases.