Date of Award

Fall 12-2018

Degree Name

Doctor of Philosophy (PhD)



Second Advisor


Third Advisor



The evidence from studies of maternal ambient air pollutant exposure (including O3, fine particulate matter (PM2.5) and PM2.5 metal constituents) and stillbirth is limited and equivocal. There is also limited evidence of the acute effect of temperature increases in the week preceding delivery and stillbirth. In the first investigation of ambient air pollution, temperature and stillbirth among a large cohort of women in Harris County, Texas, we obtained birth and fetal death records from 2008-2013 and estimated weekly residential levels of apparent temperature and ozone (O3), fine particulate matter (PM2.5), nitrogen dioxide (NO2) and PM2.5 metals from the local monitoring network.

In one study, we used Cox survival models to examine associations between time-varying O3 exposure and stillbirth with adjustment for PM2.5 and NO2 and further evaluated effect measure modification by race/ethnicity and length of gestation. In another study, we used a case-control design and applied conditional logistic regression to examine associations between PM2.5 mass and PM2.5 metals by assigning controls (live births) the same truncated exposure period as the gestational length of their matched case. Finally, we used a case-crossover design with symmetric bidirectional sampling of control periods to investigate the association of temperature (lag days 1 through 6) and all-cause stillbirths, as well as stillbirths caused by placental abruptions, in conditional logistic regression models that adjusted for air pollutant exposure.

We found a 9% (95% confidence interval (CI): 1%, 18%) increased stillbirth risk associated with a 3.61 parts per billion increase in O3 exposure. The risk was higher among women with pregnancies <37 gestational weeks (hazard ratio (HR) =1.13, 95% CI: 1.04, 1.23) as compared to women with pregnancies of longer gestation (HR=1.05, 95% CI: 0.87, 1.27) and among Hispanic women (HR=1.14, 95% CI: 1.02, 1.27).

We observed null associations between PM2.5 metal exposure (modeled as a continuous variable) and stillbirth and slightly increased risks for mothers with exposure to high (≥95th percentile) levels of PM2.5 mass. There were also elevated risks (21% to 36%) associated with high exposures (≥95th percentile) to metal constituents of PM2.5 (Al, Cr, Cu, Fe, Pb, Mn, Ni, Se, Ti and Zn), with the strongest association observed for Ni (OR = 1.36, 95% confidence interval (CI): 1.03, 1.79).

We found that a 10°F increase in apparent temperature was positively associated with stillbirths caused by placental abruption on lag days 1, 2 and 3. The risk was highest (OR= 1.29, 95% confidence interval (CI): 1.02, 1.63, per 10°F increase) when exposure was examined as a moving average over lag days 1-3. We observed a slightly elevated risk (OR= 1.06, 95% CI: 0.99, 1.13, per 10°F increase) for all-cause stillbirth on lag day 1 and no evidence of an association on subsequent lag days. Our findings contribute to the literature on air pollution and temperature exposures during pregnancy and stillbirth, particularly concerning methods of comparing different risk periods based on gestational length.