Risk factors for congenital hypothyroidism: an investigation of infant's birth weight, ethnicity, and gender in California, 1990-1998

Background: Approximately 85% of primary congenital hypothyroidism (CH) is sporadic and due to malformations of the thyroid gland. Past studies have reported an increased birth weight among infants with CH. We have attempted to replicate and expand these observations, examining the association between different birth weight categories and CH stratified by infant's sex. We have also examined the prevalence of CH by mother's age and infant's ethnicity, gender, and year of birth.

Methods: A cross-sectional study was conducted on 5, 049,185 infants screened by the statewide California Newborn Screening Program between 1990 and 1998, an estimated 98.6% of all newborns in the state. Dried blood spots from a heel stick were assayed for thyroxine (T4), and presumptive positives had follow-up assays of thyroid-stimulating hormone (TSH) to determine definite positives.

Results: A total of 1,806 cases of CH were identified. The following findings are unlikely to be due to chance. Compared with infants with birth weights of 3,000-3,499 g, infants weighing <2,000 g and those weighing >/=4,500 g had a twofold or greater increase in the prevalence of CH. This was not explained as a result of confounding by the infant's ethnicity or gender. Compared with whites, elevated prevalence rates were found in most ethnic groups, which include the following: Hispanics, Chinese, Vietnamese, Asian Indians, Filipinos, Middle Easterners, and Hawaiians. As reported previously, black infants had about one-third the prevalence rate of whites. We also observed the frequently described female preponderance of CH. The female excess was maintained at all birth weights, however it varied by infant's ethnicity. Trends in the prevalence of CH were not associated with mother's age or with the time interval between 1990 and 1998.

Conclusions: We observed an increased risk of CH in both low-birth-weight (<2,000-g) and macrosomic (>/=4,500-g) infants. This U-shaped association has not been described in past studies. We have also expanded the previously described ethnic differences in CH risk to include ethnic groups not previously studied. The unique pattern of CH occurrence suggests that further studies to define modifiable risk factors may be useful.