National health and nutrition examination survey III thyroid-stimulating hormone (TSH)-thyroperoxidase antibody relationships demonstrate that TSH upper reference limits may be skewed by occult thyroid dysfunction

Context: The setting of the TSH upper reference limit impacts the diagnosis of mild hypothyroidism and is currently controversial. Objective: Our objective was to evaluate factors influencing the TSH reference range. Design: Nonpregnant subjects aged 12 yr and older from National Health and Nutrition Examination Survey III were used to study the relationships between TSH, thyroid peroxidase antibodies ( TPOAb), and thyroglobulin antibodies in different ethnic groups. Results: TPOAb prevalence was lowest (< 3%) when TSH was between 0.1 and 1.5 mIU/liter in women and between 0.1 and 2.0 mIU/ liter in men and progressively increased to above 50% when TSH exceeded 20 mIU/ liter. TSH reference range parameters ( 2.5th, 50th, and 97.5th percentiles) were analyzed according to thyroid antibody status, race/ethnicity, and age for the 14,202 subjects made up of non-Hispanic Blacks ( B), non-Hispanic whites ( W), and Mexican-Americans ( M) who did not report thyroid disease or taking thyroid-altering medications and whose total T-4 was within the reference range. For each age group of each ethnicity, the inclusion of antibody-positive subjects increased TSH medians and upper limits ( 97.5th percentiles). The TSH upper limit was lower for the entire B cohort vs. W or M. However, this difference was lost when age cohorts with a similar prevalence of TPOAb ( B age 40-49 yr vs. W and M age 20-29 yr) were compared. Conclusions: Ethnic differences in TSH were not present when populations with the same relative frequency of thyroid antibodies were compared. TSH upper reference limits may be skewed by TPOAb-negative individuals with occult autoimmune thyroid dysfunction.