ABSTRACT
Infertility and spontaneous abortion (SAB) are significant public health problems, each affecting up to 20% of
reproductive-aged couples in the United States. As couples increasingly postpone childbearing to the later
reproductive years, many seek infertility treatment, which has relatively low success rates, costs an estimated
$5 billion per year, and is associated with adverse pregnancy outcomes. Therefore, identifying modifiable risk
factors for subfertility and SAB is an important public health goal. The potential effects of air pollution on risk of
subfertility and SAB are understudied. The few existing human studies have major limitations including
retrospective study design, small study size, suboptimal assessment of exposure and outcome, inadequate
control for potential confounding variables, and limited generalizability. The proposed study will assess the
relation of several air pollutants to risk of subfertility and SAB in two interrelated preconception prospective cohort
studies in North America and Denmark. Using web-based recruitment, data collection, and follow-up, we have
enrolled over 16,500 women attempting pregnancy, collectively comprising the largest preconception cohort
study ever conducted on factors related to fertility and SAB. We have collected prospective data on time-to-
pregnancy, SAB, potential confounders in both partners, and residential addresses at baseline and on bi-monthly
follow-up questionnaires for all participants, providing the foundation for accurate time-varying assessment of
exposure to air pollution throughout the preconception and early pregnancy periods. Our specific aims are to
develop spatio-temporal models of ambient PM2.5, NO2, and O3 for the etiologic time periods of interest, construct
measures of traffic and roadway density around each residence using geographic information system software,
and assign appropriate time-varying estimates of exposure. Among 18,500 women, we will examine the
association between individual measures of air pollution (PM2.5, NO2, O3 and measures of traffic density) and
fecundability, the cycle-specific probability of conception. We will evaluate the same measures of air pollution
with risk of SAB. We will also utilize state-of-the-art statistical methods, including predictive k-means clustering,
to examine the effects of pollutant mixtures, with an emphasis on speciation of PM components. Strengths of
this application include its large study size, preconception enrollment of a geographically and ethnically diverse
cohort of pregnancy planners, assessment of repeated measures of exposure during the preconception and
early pregnancy periods, excellent control for individual-level and neighborhood-level confounding via collection
of data on a wide range of covariates, and the use of multi-pollutant modeling techniques. The present application
is cost-effective in leveraging already-established cohort studies funded by the NICHD. The proposed study
addresses novel and important reproductive health outcomes that have had only limited study in relation to air
pollution. Study results are likely to inform primary prevention of infertility and SAB, and add to evidence used
by regulators to improve air quality.