Air Pollution and Pregnancy Complications in Complex Urban Environments: Risks, Heterogeneity, and Mechanisms

妊娠并发症（如妊娠期糖尿病[GDM]、妊娠期高血压[GHTN]和妊娠前 子痫和子痫[PE/E]是围产期发病率和死亡率的主要原因。研究已经研究 空气污染与妊娠并发症的关系，但有重大局限性，包括1）依赖 出生证明或账单/索赔数据，其中的信息可能缺失或有效性有疑问，无法确定 结局和合并症; 2）对其他因素的影响混杂和/或改变的考虑有限 环境因素; 3）由于缺乏居住地址历史导致的空气污染暴露错误分类; 4） 关注单个空气污染物而不是空气污染物混合物; 5）缺乏对 按暴露时间和地点、产妇状况、其他环境因素分列的空气污染风险， 结果的子类型; 6）缺乏对将孕产妇共病与 结果。我们建议进行一项为期4年的研究，以解决这些局限性，并进一步了解 空气污染物对妊娠并发症的影响我们将利用最先进的时空空气污染 建模和新的统计方法，检查个人和复合暴露概况， 一个纵向（从受孕前到产后）妊娠队列，约40万例单胎妊娠， 2008-2018年导致妊娠20周后活产或胎儿死亡的前瞻性记录 凯泽电子健康记录（EHR）中的高质量临床数据和居住地址 永久南加州成员在8个南加州县。主要结局是GHTN， PE/E，GDM。我们将估算个人水平的空气污染物暴露（颗粒物及其成分 和交通相关的污染物使用复杂的时空模型），天气（空气温度，相对湿度）， 湿度、压力）和建筑环境测量（绿化、步行便利性、噪音、社区资源） 根据预先记录的母亲地址我们将研究的协变量包括孕产妇 合并症;既往妊娠史和结局;个体和背景社会经济地位 (SES)指标;怀孕期间的就业和工作分类;自我报告的体力活动， smoking.将使用补充统计方法来评估暴露于空气混合物的影响 污染物，同时考虑共同暴露于天气，建筑环境和SES。我们将检验效果 SES，母体因素和其他环境暴露的修饰，以及 母亲因素对空气污染和结果之间的关联。我们将阐明怀孕的风险 暴露于空气污染的并发症、SES引起的风险异质性、母体条件和其他 环境因素，潜在的潜在机制，易感亚群，和时间窗口 易感性确定可改变的环境风险因素和高风险亚群可能有助于 设计有针对性的干预措施，以减少这些风险和相关的不利的孕产妇和胎儿的结果。

Pregnancy complications (e.g. gestational diabetes [GDM], gestational hypertension [GHTN], and pre- eclampsia & eclampsia [PE/E] are major causes of perinatal morbidity and mortality. Studies have examined associations of air pollution with pregnancy complications, but have major limitations, including 1) reliance on birth certificates or billing/claims data where information may be missing or of questionable validity to ascertain outcomes and co-morbidities; 2) limited consideration of effect confounding and/or modification by other environmental factors; 3) air pollution exposure misclassification due to lack of residential address history; 4) focus on individual air pollutants rather than air pollutant mixtures; 5) lack of focus on the heterogeneity of the risk from air pollution by time and place of exposures, maternal conditions, other environmental factors, and sub-types of outcomes; 6) lack of understanding of the mediation pathways linking maternal co-morbidity with outcomes. We propose a 4-year study to address these limitations and advance knowledge of the impact of air pollutant mixture on pregnancy complications. We will leverage state-of-the-art spatiotemporal air pollution modeling and novel statistical methods that examine both individual and composite exposure profiles with a longitudinal (pre-conception through postpartum) pregnancy cohort of ~400,000 singleton pregnancies in 2008-2018 that result in a live birth or fetal death after 20 weeks gestation that have prospectively-recorded high quality clinical data and residential addresses from the electronic health record (EHR) of Kaiser Permanente Southern California members in 8 southern California counties. Primary outcomes are GHTN, PE/E, GDM. We will estimate individual-level air pollutant exposures (particulate matter and its composition and traffic-related pollutants using sophisticated spatiotemporal models), weather (air temperature, relative humidity, pressure), and built environment measures (greenness, walkability, noise, neighborhood resources) based on prospectively-recorded maternal addresses. Covariates we will examine include maternal comorbidities; history of previous pregnancies and outcomes; individual and contextual socioeconomic status (SES) indicators; employment during pregnancy and job classification; self-reported physical activity and smoking. Complementary statistical methods will be used to evaluate effects of exposure to a mixture of air pollutants while accounting for co-exposure to weather, built environment, and SES. We will examine effect modifications by SES, maternal factors, and other environmental exposures, and the potential mediating role of maternal factors on associations between air pollution and the outcomes. We will elucidate risk of pregnancy complications from air pollution exposure, heterogeneity of risk due to SES, maternal conditions, and other environmental factors, potential underlying mechanisms, susceptible sub-populations, and time windows of susceptibility. Identification of modifiable environmental risk factors and high-risk subpopulation may help to design targeted interventions to reduce these risks and associated adverse maternal and fetal outcomes.