Abstract
INTRODUCTION: Although studies have found positive associations between exposure to PM2.5 and preterm birth, distinguishing between spontaneous preterm birth (sPTB) and iatrogenic preterm birth (iPTB) was a challenge in previous research. This study examined associations between total PM2.5 and PM2.5 constituent exposure and sPTB. METHODS: This is a retrospective cohort study from 2008 to 2018 of singleton live births within a large health care system in southern California, United States. Daily total PM2.5 concentrations and monthly data on five PM2.5 constituents (sulfate, nitrate, ammonium, organic matter, and black carbon) were obtained. The average concentrations of total PM2.5 and constituents were calculated over the pregnancy and by trimester. A novel natural language processing algorithm was used to identify sPTB in medical records. Discrete-time survival models were used to estimate the associations of total PM2.5 and constituents with sPTB. Effect modifiers included maternal race/ethnicity, educational attainment, household income, and green space. RESULTS: There were 19,341 (4.7%) sPTBs among 409,037 births. We observed significant associations of sPTB with PM2.5, black carbon, nitrate, and sulfate. The second trimester was the most susceptible window. Significantly higher associations with PM2.5 were observed among mothers with lower educational attainment, lower income, and less green space exposure. CONCLUSIONS/IMPLICATIONS: Maternal exposures to PM2.5 and specific PM2.5 constituents were associated with an increased risk of sPTB. Mothers with lower socioeconomic status were vulnerable, whereas green space was a protective effect modifier.