Phenol and Phthalate Effects on Thyroid Hormone Levels during Pregnancy: Relying on In Vitro Assays and Adverse Outcome Pathways to Inform an Epidemiological Analysis

Abstract

Background: Studies characterizing associations between phenols, phthalates and thyroid hormones during pregnancy produce inconsistent results. This divergence may be partly attributable to false positives due to multiple comparison testing of large numbers of chemicals, and measurement error as studies rely on small numbers of biospecimens despite high intra-individual variability in urinary chemical metabolite concentrations.

Objectives: This study employs a priori chemical filtering and expanded urinary biomonitoring to evaluate associations between phenol/phthalate exposures and serum thyroid hormones assessed during pregnancy.

Methods: A two-tiered approach was implemented: a) In vitro high-throughput screening results from the ToxCast/Tox21 database, as informed by a thyroid Adverse Outcome Pathway network, were evaluated to select phenols/phthalates with activity on known and putative molecular initiating events in the thyroid pathway; and b) Adjusted linear regressions were used to study associations between filtered compounds and serum thyroid hormones measured in 437 pregnant women recruited in Grenoble area (France) between 2014 and 2017. Phenol/phthalate metabolites were measured in repeated spot urine sample pools (median: 21 samples/women).

Results: The ToxCast/Tox21 screening reduced the chemical set from 16 to 13 and the associated number of statistical comparisons by 19%. Parabens were negatively associated with free triiodothyronine (T3) and the T3/T4 (total thyroxine) ratio. Monobenzyl phthalate was positively associated with total T4 and negatively with the T3/T4 ratio. Effect modification by iodine status was detected for several compounds (among them $\Sigma \text{DEHP}$ and mono-n-butyl phthalate) that were associated with some hormones among women with normal iodine levels.

Conclusion: For these chemicals, screening for compounds with an increased likelihood for thyroid-related effects and relying on repeated urine samples to assess exposures improved the overall performance of multichemical analyses of thyroid disruption. This approach may improve future evaluations of human data for the thyroid pathway with implication for fetal health and may serve as a model for evaluating other toxicity outcomes. https://doi.org/10.1289/EHP10239.

Figure 1.

Expected changes and 95% CIs in (A) TSH, (B) total T4, (C) free T4, (D) Total T3, (E) free T3, and (F) T3/T4 ratio associated with concurrently increasing quantiles of all exposure biomarkers, relative to when all concentrations are fixed at their 25th percentile. Note: Numerical value of effect estimates and 95% CIs are reported in Supplemental Material, Table S4. Analyses were adjusted for maternal age, BMI before pregnancy, education level, maternal smoking during the first trimester of pregnancy, parity, gestational age at serum collection, time of serum collection, maternal urinary iodine concentrations and selenium concentrations in sera during pregnancy. Models were also adjusted for analytical batch for all hormones but TSH for which no batch effect was detected. BMI, body mass index; CI, confidence interval; T3, triiodothyronine; T4, thyroxine; TSH, thyroid stimulating hormone.

Figure 2.

Estimated effect and 95% CI of an increase from the 25th to 75th percentile in a single biomarker concentration on T3/T4 ratio when all other exposure biomarkers are fixed at either the 25th, 50th, or 75th percentiles. Note: Numerical value of effect estimates and 95% CI are reported in Supplemental Material, Table S5. Analyses were adjusted for maternal age, BMI before pregnancy, education level, maternal smoking during the first trimester of pregnancy, parity, gestational age at serum collection, time of serum collection, maternal urinary iodine concentrations and selenium concentrations in sera during pregnancy. Models were also adjusted for analytical batch for all hormones but TSH for which no batch effect was detected. BMI, body mass index; CI, confidence interval; DEHP, di(2-ethylhexyl) phthalate, DINCH, di(isononyl)cyclohexane-1,2-dicarboxylate; DiNP, Diisononyl phthalate; ΣDEHP, molar sum of the five DEHP metabolites; ΣDINCH, molar sum of the two DINCH metabolites; ΣDiNP, molar sum of the three DiNP metabolites; MBzP, monobenzyl phthalate; MiBP, monoisobutyl phthalate; MnBP, mono-n-butyl phthalate; oh-MPHP, mono-6-hydroxy-propylheptyl phthalate; T3, triiodothyronine; T4, thyroxine.