Applying and advancing modern approaches for studying the joint impacts of environmental chemicals on pregnancy outcomes

Preterm birth is a significant public health challenge due to increasing rates over time, as well as serious consequences for infant mortality, childhood morbidity, and economic costs to society. Conditions that contribute to preterm birth remain unclear, though an influence by environmental chemical exposures is

suspected but poorly understood. Establishing links with common environmental chemicals could have huge public health impact since many exposures could be modifiable through remediation, policies or other interventions. Low birth weight and fetal growth restriction also represent a great public health challenge, as

they too have increased in recent decades and may be influenced by modifiable exposures to environmental chemicals. This project proposes to leverage the established Boston Lifecodes cohort study (N~4,000) of risk factors for preterm birth and other adverse pregnancy outcomes, with a focus on exposure to mixtures of

commonly-encountered chemicals. We propose to select 1,000 singleton births from Lifecodes with detailed information and samples collected at multiple times during pregnancy. We will then utilize state-of-the-art methods to estimate biomarkers of exposure to chemical mixtures (per- and polyfluorinated substances

[PFAS], phthalates and phthalate replacement chemicals, polycyclic aromatic hydrocarbons [PAH], and metals/metalloids), in addition to intermediate biomarkers of effect and repeated ultrasound measures of fetal growth, in order to provide much needed human data on environmental and other predictors of pregnancy

outcomes and insights on the biological pathways involved. Results from our preliminary work show that oxidative stress may be an important link between exposure and outcome that needs to be explored in more depth using the proposed pathway-specific biomarkers. To accomplish our aims we will develop innovative

statistical and machine learning approaches for analyzing mixtures and mediation with high-dimensional mediator sets, with the goal of improving our ability to discover and define these relationships. Finally, a study sub-aim is to identify conditions/activities contributing to high exposures that can inform exposure reduction

strategies. The expected outcomes of this study are new and much needed information on the magnitude, sources, and impacts of exposure to commonly encountered chemicals, both individually and in combination, among pregnant women, and innovative methods for identifying relevant biological pathways and assessing

health impacts from exposure to mixtures. Our findings will have a significant impact on public health given widespread exposure to the target chemicals, the growing need to identify environmental agents that adversely impact pregnancy, and the need to discover contributors to the high rates of preterm birth in the U.S. and

beyond that could be prevented. Our study will also provide new information on the role of oxidative stress in adverse pregnancy outcomes which may inform future therapeutic or preventative interventions, and contribute new statistical and machine learning methods for investigating mixtures, mediation, and birth outcomes.