Prenatal exposure to endocrine disrupting chemical mixtures and infant birth weight: A Bayesian analysis using kernel machine regression

Environ Res. 2021 Apr:195:110749. doi: 10.1016/j.envres.2021.110749. Epub 2021 Jan 17.


Background: Pregnant women are regularly exposed to a multitude of endocrine disrupting chemicals (EDCs). EDC exposures, both individually and as mixtures, may affect fetal growth. The relationship of EDC mixtures with infant birth weight, however, remains poorly understood. We examined the relations between prenatal exposure to EDC mixtures and infant birth weight.

Methods: We used data from the Maternal-Infant Research on Environmental Chemicals (MIREC) Study, a pan-Canadian cohort of 1857 pregnant women enrolled between 2008 and 2011. We quantified twenty-one chemical concentrations from five EDC classes, including organochlorine compounds (OCs), metals, perfluoroalkyl substances (PFAS), phenols and phthalate metabolites that were detected in >70% of urine or blood samples collected during the first trimester. In our primary analysis, we used Bayesian kernel machine regression (BKMR) models to assess variable importance, explore EDC mixture effects, and identify any interactions among EDCs. Our secondary analysis used traditional linear regression to compare the results with those of BKMR and to quantify the changes in mean birth weight in relation to prenatal EDC exposures.

Results: We found evidence that mixtures of OCs and metals were associated with monotonic decreases in mean birth weight across the whole range of exposure. trans-Nonachlor from the OC mixture and lead (Pb) from the metal mixture had the greatest impact on birth weight. Our linear regression analysis corroborated the BKMR results and found that a 2-fold increase in trans-nonachlor and Pb concentrations reduced mean birth weight by -38 g (95% confidence interval (CI): -67, -10) and -39 g (95% CI: -69, -9), respectively. A sex-specific association for OC mixture was observed among female infants. PFAS, phenols and phthalates were not associated with birth weight. No interactions were observed among the EDCs.

Conclusions: Using BKMR, we observed that both OC and metal mixtures were associated with decreased birth weight in the MIREC Study. trans-Nonachlor from the OC mixture and Pb from the metal mixture contributed most to the adverse effects.

Keywords: Bayesian kernel machine regression (BKMR); Birth weight; Chemical mixtures; Endocrine disrupting chemicals; Fetal growth.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Bayes Theorem
  • Birth Weight
  • Canada
  • Endocrine Disruptors* / toxicity
  • Environmental Pollutants* / toxicity
  • Female
  • Humans
  • Infant
  • Male
  • Maternal Exposure / adverse effects
  • Pregnancy
  • Prenatal Exposure Delayed Effects* / chemically induced
  • Prenatal Exposure Delayed Effects* / epidemiology


  • Endocrine Disruptors
  • Environmental Pollutants