Heavy Metal Index
Ask the IndexAboutIndex Pro

Effects of maternal arsenic exposure on birth outcomes using harmonized data across three birth cohorts

Cite this pageVersion history

Summary

This 2026 harmonised multi-cohort study analysed associations between maternal urinary arsenic (measured during pregnancy) and birth outcomes (birthweight, length, head circumference, gestational age, preterm birth, small or large for gestational age) in 3,222 women from three US birth cohorts: NBCS (Navajo Nation, New Mexico), NHBCS (New Hampshire), and PROTECT (Puerto Rico). Total urinary arsenic and arsenic speciation were measured; cohort-specific harmonisation was applied using standardised exposure and outcome definitions. No consistent statistically significant association between maternal arsenic exposure and birth outcomes was found across all three cohorts.

Key numbers

  • Geometric mean total urinary arsenic (gMean; pooled across cohorts): 6.03 ug/L (gSD 3.2)
  • By cohort: PROTECT 9.66 ug/L (highest; Puerto Rico); NHBCS 5.48 ug/L; NBCS 4.80 ug/L
  • Comparison: NHANES women aged 14-45: gMean 6.8 ug/L (gSD 1.04)
  • Speciation (NBCS and NHBCS): inorganic arsenic (iAs), DMA, MMA measured; arsenic sum = iAs + DMA + MMA used as sensitivity analysis
  • Birth outcome associations: effect estimates ranged from -6.50 g (PROTECT) to +23.26 g (NBCS) for birthweight per unit log total arsenic (As); not statistically significant or consistent in direction
  • Heterogeneity: I2 = 79.96% for iAs sum and preterm birth, indicating high between-cohort variability

Methods

NBCS: anion-exchange HPLC-ICP-MS (Dartmouth Trace Element Analysis Core) for urinary speciation. PROTECT and NHBCS: ICP-MS (CDC NCEH DLS) for total and speciated urinary arsenic. Urinary dilution correction: specific gravity (PROTECT, NHBCS) or creatinine (NBCS). Harmonisation via secure online ETL framework with standardised exposure and outcome definitions. Meta-analysis of cohort-specific estimates.

Implications

The absence of consistent statistically significant associations is consistent with the broader epidemiological literature on low-to-moderate arsenic exposure and birth outcomes, which shows mixed results. The high between-cohort heterogeneity (I2 near 80%) illustrates that pooled analyses of heterogeneous populations can obscure real cohort-specific effects. For exposure context: US pregnant women in this study had total urinary As geometric means within the range of the general NHANES female population, indicating moderate rather than high exposure. The PROTECT Puerto Rico cohort showed the highest arsenic levels, consistent with known drinking water arsenic issues in parts of Puerto Rico.

Update history

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

Referenced by