Serum TSH and total T4 in the United States population and their association with participant characteristics: National Health and Nutrition Examination Survey (NHANES 1999-2002).

OBJECTIVE: Describe thyrotropin (TSH) and thyroxine (T4) levels in the U.S. population and their association with selected participant characteristics.
DESIGN: Secondary analysis of data from the National Health and Nutrition Examination Survey (NHANES) collected from 4392 participants, reflecting 222 million individuals, during 1999-2002.
RESULTS: Hypothyroidism prevalence (TSH > 4.5 mIU/L) in the general population was 3.7%, and hyperthyroidism prevalence (TSH < 0.1 mIU/L) was 0.5%. Among women of reproductive age (12-49 years), hypothyroidism prevalence was 3.1%. Individuals aged 80 years and older had five times greater odds for hypothyroidism compared to 12- to 49-year-olds (adjusted odds ratio [OR] = 5.0, p = 0.0002). ORs were adjusted for sex, race, annual income, pregnancy status, and usage of thyroid-related medications (levothyroxine/thyroid, estrogen, androgen, lithium, and amiodarone). Compared to non-Hispanic whites, non-Hispanic blacks had a lower risk for hypothyroidism (OR = 0.46, p = 0.04) and a higher risk for hyperthyroidism (OR = 3.18, p = 0.0005), while Mexican Americans had the same risk as non-Hispanic whites for hypothyroidism, but a higher risk for hyperthyroidism (OR = 1.98, p = 0.04). Among those taking levothyroxine or desiccated thyroid, the adjusted risk for either hypothyroidism (OR = 4.0, p = 0.0001) or hyperthyroidism (OR = 11.4, p = 4 x 10(-9)) was elevated.
CONCLUSIONS: Associations with known factors such as age, race, and sex were confirmed using this data set. Understanding the prevalence of abnormal thyroid tests among reproductive-aged women informs decisions about screening in this population. The finding that individuals on thyroid hormone replacement medication often remain hypothyroid or become hyperthyroid underscores the importance of monitoring.