Phenols, phthalates and thyroid hormone levels during pregnancy; relying on toxicological data and Adverse Outcome Pathways to inform epidemiological analysis

Abstract

BACKGROUND AND AIM: Results of studies characterizing associations between exposure to phenols, phthalates and thyroid hormone concentrations during pregnancy are inconsistent. Given the high number of exposures and outcomes usually considered, we hypothesis that part of the inconsistency results from chance findings. METHODS: We included 437 pregnant women from the SEPAGES cohort with thyroid stimulating hormone (TSH), protein-bound and free thyroxine (T4) and triiodothyronine (T3) quantified in maternal sera. We assessed 12 phenols,13 phthalate and two 1,2-cyclohexane dicarboxylic acid diisononyl ester (DINCH) metabolites in a pool of 21 spot urine samples collected during pregnancy. We relied on an existing thyroid adverse outcome pathway (AOP) network and in vitro high-throughput screening results from the ToxCast database to select assayed chemicals interacting with well-known and putative molecular initiating events relevant to thyroid hormone disruption (e.g., thyroperoxidase inhibition). We then employed adjusted linear regressions to study associations between these selected chemicals, and hormone concentrations. RESULTS:Compared to agnostic approaches, our a priori selection allowed to reduce the number of tests performed by 20%. Among the selected chemicals, three (mono-6-hydroxy-propylheptyl phthalate (OH-MPHP), di-2-ethylhexyl phthalate (∑DEHP), bisphenol A) were negatively associated with TSH (p-values ≤ 0.09). The associations observed with bisphenol A and ∑DEHP were modified by maternal iodine status. We also observed a U-shape association between triclosan and TSH and negative associations between propylparaben (-1.4% (95%CI: -2.7; 0.0) for each doubling in urinary concentrations), monobenzyl phthalate (5.3% (95%CI: -10.12; -0.14) and the T3/T4 ratio. CONCLUSIONS:Relying on repeated urine samples, we highlighted negative associations between several phenols and phthalate metabolites and TSH, as well as with the T3/T4 ratio. This hypothesis driven methodology incorporating molecular pathway tools provides a promising approach for exposome studies where there are often large numbers of substances and for which agnostic statistical approaches increase the likelihood of false positive findings. KEYWORDS: Phenols, phthalates, DiNCH, thyroid hormones, pregnancy