RE: Submission of Perchlorate Paper with Conditions

Previous epidemiological studies provide conflicting evidence as to whether environmental perchlorate exposure can affect levels of circulating thyroid hormones in the general population. We investigated the statistical relationships between biomarkers of perchlorate exposure and serum thyroid hormone levels in 2007–2008 National Health and Nutrition Evaluation Survey (NHANES) subjects. Generalized additive mixed models (GAMMs) were developed to estimate the relationships between T3 and T4 levels and creatinine-adjusted urinary perchlorate excretion. The models included covariates related to gender, age, ethnicity, income, smoking status, prescription medications, and biomarkers of exposures to other goitrogenic ions and phthalate ester metabolites. Where necessary, relationships between hormone levels and covariates were represented as nonlinear smoothed terms. The effect of the hypothalamic–pituitary–thyroid (HPT) axis on serum hormone levels was taken into account by including a term for thyroid stimulating hormone (TSH) in the models. Regression coefficients for perchlorate were significant and negative in GAMMs predicting total T4 and free T3 levels in males, females, and for the entire cohort when phthalate ester biomarkers and other covariates were included. Coefficients for perchlorate were also significant and negative in regressions predicting free T4 levels in males and in the entire study population. The consistency of these results suggests that HPT axis controls do not completely compensate for small changes in thyroid hormone levels associated with perchlorate and phthalate ester exposures. & 2012 Elsevier Inc. All rights reserved.