Association of maternal ozone exposure with term low birth weight and susceptible window identification

Background: Ozone pollution keeps deteriorating in the context of climate change. Maternal ozone exposure may be associated with low birth weight (LBW), but the results are still inconsistent. The identification of the critical exposure windows, a specific period of particular susceptibility during pregnancy, remains unresolved. We aimed to evaluate whether ozone exposure was associated with term LBW and further identify the susceptible exposure windows. Methods: A retrospective cohort study was conducted in Guangzhou, a megacity in the most populous and economically developed city clusters in China. We included 444,096 singleton live births between January 2015 and July 2017. From 11 fixed stations, we collected daily 1-h maximum and 8-h maximum moving average ozone level (O-3-1 h and O-3-8 h) and calculated exposures for each participant based on their district of residence during pregnancy. We used traditional Logistic regression to estimate the trimester-specific association between ozone exposure and term LBW, and further estimated monthly- and weekly association by distributed lag models (DLMs) with Logistic regression. Odds ratios (ORs) and 95% confidence intervals (CIs) of term LBW were calculated for an interquartile range (IQR) increase in ozone exposure. Stratified analyses and heterogeneity tests were conducted by maternal age and infant sex. Results: The incidence of term LBW was 1.9%. During the study period, the mean O-3-1 h and O-3-8 h levels were 112.6 mu g/m(3) and 84.5 mu g/m(3), respectively. Increased O-3-1 h (IQR: 90 mu g/m(3)) and O-3-8 h (73 mu g/m(3)) exposure during the second trimester were associated with increased risk of term LBW. At a monthly level, the term LBW risk was associated with O-3-1 h exposure during the 4th-6th month and O-3-8 h exposure during the 6th month. By estimating the weekly-specific association, we observed that critical exposure windows were the 15th-26th gestational weeks for O-3-1 h, and the 20th-26th weeks for O-3-8 h, respectively. Estimated ORs and 95% CIs ranged from 1.012 (1.000, 1.024) to 1.023 (1.007, 1.039). When examined by subgroups, the effects were present among women >= 35 years or < 25 years old and those with female babies. Conclusions: This study provides compelling evidence that exposure to O-3 was associated with increased risk of term LBW, and gestational weeks 15th-26th was found to be particularly susceptible. These findings provide a research basis for further mechanism examination, public health interventions, and targeted environmental policy-making.