Evidence is inconsistent regarding which windows of PM2.5 exposure are critical for adverse perinatal outcomes. We investigated associations between timing of gestational PM2.5 exposure and perinatal outcomes. Participants included 19,108 mother-infant dyads from 51 sites of the Environmental influences on Child Health Outcomes (ECHO) cohort. Repeated measures of PM2.5 exposure were included based on high-resolution spatiotemporal models for trimesters 1-3, early first trimester (≤14 days), and late first trimester (70-92 days). We estimated associations of PM2.5 exposure (per 5 μg/m3 increase) and continuous outcomes (gestational age at birth [GA] and birthweight for gestational age z-scores [BWZs]) using generalized estimating equation (GEE) models for linear regression. Poisson regression via GEE was used to estimate risk ratios (RRs) of PM2.5 exposure (per 5 μg/m3 increase) with binary outcomes (preterm birth [PTB], <37 completed weeks of gestation), and term small for gestational age [SGA], <10th percentile). We explored effect modification by participants' characteristics. In fully adjusted models, early 1st trimester PM2.5 exposure was associated with lower BWZ (β = -0.03, 95 % CI -0.06, -0.001); association with term SGA was RR = 1.06, 95 % CI 0.99, 1.13. Results were mostly null for other windows of gestational exposure. When stratified by sex, early pregnancy PM2.5 exposure and lower BWZ associations were observed among females, but not males. Suggestive evidence indicates that associations of PM2.5 exposure with GA, PTB risk, and term SGA risk may vary by maternal race and ethnicity. Our results suggest that policies and practices that reduce the risks of PM2.5 exposure, particularly in pre-conception and early pregnancy, may improve perinatal outcomes.
Keywords: Birth weight; Environmental pollutants; Gestational age; Infant; Particulate matter; Pregnancy; Premature birth.
Copyright © 2026 The Authors. Published by Elsevier Inc. All rights reserved.