1329

Background: Few prior studies of prenatal air pollutant exposures on birth outcomes have examined effect modification by socioeconomic factors or infant sex, nor examined critical exposure windows smaller than trimesters. We investigated associations of prenatal concentrations of fine particulate matter (PM_2.5) with birthweight and small for gestational age (SGA).
Methods: The analytic population (n=1,879) comprised of term (≥37 weeks) births in two pregnancy cohorts in the ECHO-PATHWAYS consortium (CANDLE and TIDES). PM_2.5 exposure was estimated using a fine-scale spatiotemporal model and averaged over windows that may be critical to prenatal growth and development (0-2, 10-12 weeks of gestation, and last month of pregnancy as well as each trimester. SGA was calculated as <10th percentile based on U.S. referent percentiles. We fit adjusted linear and modified Poisson regression models to estimate linear associations with birthweight and relative risk (RR) of SGA, respectively, per 2-µg/m³ PM_2.5. Interactions by maternal education, household income, neighborhood socioeconomic environment, and infant sex were explored.
Results: Overall, 2.5% of births were low birthweight (<2,500g) and 10.3% were SGA. Mean PM_2.5 pregnancy exposure was 9.8 µg/m³ (interquartile range = 2.2). Birthweight was inversely associated with PM_2.5 during 0-2 weeks (β: -25.0g; 95%CI: -46.3,-3.8), month before delivery (β: -35.5g; 95%CI: -62.1,-8.9), and in trimester 1, 2, and 3 (βs: -43.3g, -70.2g, -51.3g; 95%CIs: -85.0,-1.7; -111.4,-29.0; -90.7,-11.9; respectively), but not 10-12 weeks (β: -9.7g; 95%CI: -34.3,14.8). For SGA, we only observed associations with 0-2 weeks (RR: 1.2; 95%CI: 1.0,1.4). There was suggestive evidence associations with SGA is modified by household income, with stronger associations in lower incomes. No modification by other socioeconomic factors or sex were observed.
Conclusions: Findings indicate that prenatal PM_2.5 exposure, particularly during early and late pregnancy, is associated with fetal growth. There is suggestive evidence that the association is stronger with lower household incomes.