Since the 2000s, air pollution has generally continued to decrease in the U.S. To investigate preterm birth (PTB) risk associated with air pollutants in two consecutive pregnancies, we estimated exposures using modified Community Multiscale Air Quality models linked to the NICHD Consecutive Pregnancy Study. Electronic medical records for delivery admissions were available for 50,005 women with singleton births in 20 Utah-based hospitals between 2002-2010. We categorized whole pregnancy average exposures as high (>75th percentile), moderate (25-75) and low (<25). Modified Poisson regression estimated second pregnancy PTB risk associated with persistent high and moderate exposure, and increasing or decreasing exposure, compared to persistent low exposure. Analyses were adjusted for prior PTB, interpregnancy interval and demographic and clinical characteristics. Second pregnancy PTB risk was increased when exposure stayed high for sulfur dioxide (32%), ozone (17%), nitrogen oxides (24%), nitrogen dioxide (43%), carbon monoxide (31%) and for particles < 10 microns (29%) versus consistently low exposure. PTB risk tended to increase to a lesser extent for repeated PTB (19-21%) than for women without a prior PTB (22-79%) when exposure increased or stayed high. Area-level changes in air pollution exposure appear to have important consequences in consecutive pregnancies with increasing exposure associated with higher risk.
Keywords: air pollution; carbon monoxide; nitrogen oxides; ozone; particulate matter; pregnancy; preterm birth; sulfur dioxide.