Abstract
BACKGROUND: Preterm birth (PTB; <37 weeks completed gestation) is associated with exposure to air pollution, though variability in association magnitude and direction across exposure windows exists. We evaluated associations between weekly gestational exposure to fine particulate matter (PM(2.5)), nitrogen dioxide (NO(2)), and ozone (O(3)) with PTB in a North Carolina Birth Cohort from 2003 to 2015 (N = 1,367,517). METHODS: Daily average PM(2.5) and daily 8-hour maximum NO(2) concentration estimates were obtained from a hybrid ensemble model with a spatial resolution of 1 km(2). Daily 8-hour maximum census tract-level concentration estimates for O(3) were obtained from the EPA's Fused Air Quality Surface Using Downscaling model. Air pollutant concentrations were linked by census tract to residential address at delivery and averaged across each week of pregnancy. Modified Poisson regression models with robust errors were used to estimate risk differences (RD [95% confidence intervals (CI)]) for an interquartile range increase in pollutants per 10,000 births, adjusted for potential confounders. RESULTS: Associations were similar in magnitude across weeks. We observed positive associations for PM(2.5) and O(3) exposures, but generally null associations with NO(2). RDs ranged from 15 (95% CI = 11, 18) to 32 (27, 37) per 10,000 births for PM(2.5); from -7 (-14, -1) to 0 (-5, 4) for NO(2); and from 4 (1, 7) to 13 (10, 16) for O(3). CONCLUSION: Our results show that increased PM(2.5) exposure is associated with an increased risk of PTB across gestational weeks, and these associations persist in multipollutant models with NO(2) and/or O(3).