Heavy Metal Index
Ask the IndexAboutIndex Pro

Chemical mixtures and birth weight: comparison of multipollutant models in the Maternal-Infant Research on Environmental Chemicals (MIREC) study

Cite this pageVersion history

Overview

Complete-case analysis of 1,127 mother-infant pairs from the Maternal-Infant Research on Environmental Chemicals (MIREC) study examining associations between a 46-chemical prenatal mixture and birth weight z-scores. Four statistical methods were compared: principal component analysis (PCA), k-means clustering, WQS regression, and BKMR. The chemical panel included arsenic species, mercury, lead, cadmium, plus phthalate metabolites, bisphenols, triclosan, organophosphate pesticides, fluoride, polybrominated diphenyl ethers, PCBs, organochlorine pesticides, cotinine, and PFAS.

Key numbers

n=1,127 mother-infant pairs, MIREC study, Canada.

Findings by method:

  • PCA: significant inverse association between birth weight z-score and a PCB-loaded component (beta=negative 0.035, 95% CI: negative 0.068, negative 0.002).
  • k-means clustering: cluster characterized by high arsenic, mercury, lead, chlordane, and PCBs was inversely associated with birth weight z-score (beta=negative 0.171, 95% CI: negative 0.337, negative 0.005).
  • WQS regression: inverse association with overall mixture index (beta=negative 0.065, 95% CI: negative 0.171, 0.04; not statistically significant); main drivers Aroclor 1260, chlordane, glyphosate, PCB180.
  • BKMR: birth weight z-score 0.054 (95% CI: negative 0.100, 0.209) when all chemicals at 25th percentile versus medians; decreased to negative 0.04 (95% CI: negative 0.219, 0.14) when all chemicals at 75th percentile versus medians.

Chemicals most consistently associated with decreased birth weight across all four methods: chlordane, organochlorine insecticides, Aroclor 1260, dimethylarsinic acid, lead, PCB170, PCB180.

Sex stratification: inverse associations tended to be larger in magnitude in female infants.

PCA component with high PCBs (negative 0.035): significant for female infants. PCA component with arsenic and cadmium (PC4): inverse association in female infants (beta=negative 0.075, 95% CI: negative 0.138, negative 0.011).

Evidence Fitness

Context only. Cross-sectional outcome assessment (birth weight) in a prospective Canadian cohort. The 46-chemical panel means that the effect attributable specifically to metals cannot be isolated from co-occurring organochlorines, PCBs, PFAS, and pesticides. The consistent appearance of dimethylarsinic acid, lead, and cadmium across methods supports their role in the mixture but does not quantify their independent contribution from this analysis. No food concentration data.

Methods (brief)

MIREC study, Canada. Prenatal biomarkers measured in maternal blood and urine. Birth weight z-score from Canadian reference charts. PCA (unsupervised), k-means clustering (unsupervised), WQS (supervised), BKMR (supervised) compared. Complete-case analysis (n=1,127 of full cohort). No imputation for outcome; multiple imputation for exposure covariates below LOD.

Implications

Certification: The consistent appearance of lead and arsenic species (dimethylarsinic acid) across all four methods as contributors to lower birth weight z-scores in a prenatal cohort is relevant to the infant-safety rationale for HMT&C.

Courses: Exceptional for demonstrating that method choice matters in mixture analysis and that convergence across methods strengthens inference.

App: No food concentration data.

Microbiome: Not applicable.

Update history

No substantive edit history is available in this build. The full commit record is available in git.

Referenced by