LAUSR

BACKGROUND Associations between ambient air pollution and stillbirth have recently been explored, but most studies have focused on long-term (trimester or gestational averages) rather than short-term (within one week) air pollution exposures. OBJECTIVE To evaluate whether short-term exposures to criteria air pollutants are associated with increased risk of stillbirth. METHODS Using air pollution and fetal death certificate data from 1999 to 2009, we assessed associations between acute prenatal air pollution exposure and stillbirth in California. In a time-stratified case-crossover study, we analyzed single day and/or cumulative average days (up to a 6 day lag) of exposure to fine (PM2.5) and coarse particles (PM10-2.5), ozone (O3), nitrogen dioxide (NO2), sulfur dioxide (SO2), and carbon monoxide (CO) for mothers estimated to reside within 10 km of a pollution monitor based on reported zip code. We also examined potential confounding by apparent temperature or co-pollutants, and effect modification by maternal demographic factors, fetal sex, gestational age, and cause of stillbirth. RESULTS Stillbirth cases in the primary analyses ranged between 1,203 and 13,018, depending on the pollutant. For an IQR increase in SO2 (lag 4), O3 (lag 4), and PM10-2.5 (lag 2), we found a 2.8% (95% confidence interval (CI) 0.2%, 5.5%), 5.8% (95% CI 1.6%, 10.1%), and 6.1% (95% CI 0.1%, 12.4%) increase in the odds of stillbirth, respectively. Additional adjustment by apparent temperature had little effect on the SO2 association but slightly attenuated O3 (adjusted % change: 4.2 (95% CI -0.2%, 8.9%) and PM10-2.5 (5.7 (95% CI -1.1%, 13.0%)) associations, while other co-pollutants had minimal impact. Associations were observed specifically for stillbirths from obstetric complications and in women with higher educational attainment. CONCLUSIONS This study provides evidence for associations between prenatal short-term air pollution exposure, specifically SO2, O3 and PM10-2.5, and stillbirth in California and warrants replication of findings in other settings.