Assessing the relationship between perfluoroalkyl substances, thyroid hormones and binding proteins in pregnant women; a longitudinal mixed effects approach.

The mechanisms involved in thyroid homeostasis are complex, and perfluoroalkyl substances (PFASs) have been indicated to interfere at several levels in this endocrine system. Disruption of the maternal thyroid homeostasis during early pregnancy is of particular concern, where subclinical changes in maternal thyroid hormones (THs) may affect embryonic and foetal development. The present study investigated associations between THs, thyroid binding proteins (TH-BPs) and PFAS concentrations in pregnant women from Northern Norway. Women participating in The Northern Norway Mother-and-Child contaminant Cohort Study (MISA) donated a blood sample at three visits related to their pregnancy and postpartum period (during the second trimester, 3 days and 6 weeks after delivery) in the period 2007-2009. Participants were assigned to quartiles according to PFAS concentrations during the second trimester and mixed effects linear models were used to investigate potential associations between PFASs and repeated measurements of THs, TH-BPs, thyroxin binding capacity and thyroid peroxidase antibodies (anti-TPOs). Women within the highest perfluorooctane sulfonate (PFOS) quartile had 24% higher mean concentrations of thyroid stimulating hormone (TSH) compared to the first quartile at all sampling points. Women within the highest quartiles of perfluorodecanoate (PFDA) had 4% lower mean concentrations of triiodothyronine (T3) and women within the highest quartile of perfluoroundecanoate (PFUnDA) had 3% lower mean concentrations of free triiodothyronine (FT3). Further, the difference in concentrations and the changes between three time points were the same for the PFAS quartiles. Thyroxin binding capacity was associated with all the THs and TH-BPs, and was selected as a holistic adjustment for individual changes in TH homeostasis during pregnancy. Finally, adjusting for maternal iodine status did not influence the model predictions. Findings in the present study suggest modifications of TH homeostasis by PFASs in a background exposed maternal population. The variation in levels of THs between PFAS quartiles was within normal reference ranges and may not be of clinical significance in the pregnant woman. However, subtle individual changes in maternal THs may have significant consequences for foetal health.

The Northern Norway Mother-and-Child Contaminant Cohort (MISA) Study: PCA analyses of environmental contaminants in maternal sera and dietary intake in early pregnancy.

BACKGROUND: Although predictors of contaminants in serum or whole blood are usually examined by chemical groups (e.g., POPs, toxic and/or essential elements; dietary sources), principal component analysis (PCA) permits consideration of both individual substances and combined variables. OBJECTIVES: Our study had two primary objectives: (i) Characterize the sources and predictors of a suite of eight PCBs, four organochlorine (OC) pesticides, five essential and five toxic elements in serum and/or whole blood of pregnant women recruited as part of the Mother-and-Child Contaminant Cohort Study conducted in Northern Norway (The MISA study); and (ii) determine the influence of personal and social characteristics on both dietary and contaminant factors. METHODS: Recruitment and sampling started in May 2007 and continued for the next 31 months until December 2009. Blood/serum samples were collected during the 2nd trimester (mean: 18.2 weeks, range 9.0-36.0). A validated questionnaire was administered to obtain personal information. The samples were analysed by established laboratories employing verified methods and reference standards. PCA involved Varimax rotation, and significant predictors (p≤0.05) in linear regression models were included in the multivariable linear regression analysis. RESULTS: When considering all the contaminants, three prominent PCA axes stood out with prominent loadings of: all POPs; arsenic, selenium and mercury; and cadmium and lead. Respectively, in the multivariate models the following were predictors: maternal age, parity and consumption of freshwater fish and land-based wild animals; marine fish; cigarette smoking, dietary PCA axes reflecting consumption of grains and cereals, and food items involving hunting. PCA of only the POPs separated them into two axes that, in terms of recently published findings, could be understood to reflect longitudinal trends and their relative contributions to summed POPs. CONCLUSIONS: The linear combinations of variables generated by PCA identified prominent dietary sources of OC groups and of prominent toxic elements and highlighted the importance of maternal characteristics.