Prenatal airshed pollutants and preterm birth in an observational birth cohort study in Detroit, Michigan, USA
Cassidy-Bushrow AE, Burmeister C, Lamerato L, Lemke LD, Mathieu M, O'Leary BF, Sperone FG, Straughen JK, and Reiners JJ, Jr. Prenatal airshed pollutants and preterm birth in an observational birth cohort study in Detroit, Michigan, USA. Environ Res 2020; 189:109845.
Detroit, Michigan, currently has the highest preterm birth (PTB) rate of large cities in the United States. Disproportionate exposure to ambient air pollutants, including particulate matter ≤2.5 μm (PM(2.5)), PM ≤ 10 μm (PM(10)), nitrogen dioxide (NO(2)) and benzene, toluene, ethylbenzene, and xylenes (BTEX) may contribute to PTB. Our objective was to examine the association of airshed pollutants with PTB in Detroit, MI. The Geospatial Determinants of Health Outcomes Consortium (GeoDHOC) study collected air pollution measurements at 68 sites in Detroit in September 2008 and June 2009. GeoDHOC data were coupled with 2008-2010 Michigan Air Sampling Network measurements in Detroit to develop monthly ambient air pollution estimates at a spatial density of 300 m(2). Using delivery records from two urban hospitals, we established a retrospective birth cohort of births by Detroit women occurring from June 2008 to May 2010. Estimates of air pollutant exposure throughout pregnancy were assigned to maternal address at delivery. Our analytic sample size included 7961 births; 891 (11.2%) were PTB. After covariate adjustment, PM(10) (P = 0.003) and BTEX (P < 0.001), but not PM(2.5) (P = 0.376) or NO(2) (P = 0.582), were statistically significantly associated with PTB. In adjusted models, for every 5-unit increase in PM(10) there was a 1.21 times higher odds of PTB (95% CI 1.07, 1.38) and for every 5-unit increase in BTEX there was a 1.54 times higher odds of PTB (95% CI 1.25, 1.89). Consistent with previous studies, higher PM(10) was associated with PTB. We also found novel evidence that higher airshed BTEX is associated with PTB. Future studies confirming these associations and examining direct measures of exposure are needed.