Assessing the Individual and Interactive Associations of Pre- and Postnatal Air Pollution Exposures and Maternal Nutrition During Pregnancy with Child Blood Pressure: A Prospective Study in a Community-based Birth Cohort

Yu Ni | 2020

Advisor: Catherine Karr

Research Area(s): Environmental & Occupational Health, Maternal & Child Health, Social Determinants of Health


Background: Limited data suggest that the intrauterine environment and early life experience may contribute to increased child blood pressure, with effects possibly extending into adulthood. The individual and interactive effects of air pollution exposures and maternal nutrition during pregnancy upon child blood pressure are unclear. Methods: In this dissertation, we conducted two separate prospective studies with three aims to investigate these associations in the CANDLE study, a socioeconomically diverse birth cohort in the Memphis, Tennessee. Mother-child dyads were included if they had available residential address histories, pregnancy nutrition data (either the Healthy Eating Index 2010 [HEI], determined by a second trimester Food Frequency Questionnaires, or plasma folate measurement in mid- to late pregnancy), and a valid child blood pressure measurement at age 4-6 years. We calculated the systolic (SBP) and diastolic blood pressure (DBP) percentile incorporating sex, age and height, and categorized children as high blood pressure (HBP) (≥ 90th percentile) or normal. In the first study, we assessed the associations between air pollution exposures and child blood pressure (Aim 1) and determined whether the associations would be modified by maternal nutrition (Aim 3). Effect modifications by child sex and maternal race were also examined. NO2 and PM2.5 estimates in both pre- and postnatal windows were obtained from annual national models and spatio-temporal models respectively. We fit multivariate Poisson and linear regressions to estimate the exposure-outcome associations, and quantified multiplicative joint effects with maternal nutrition, child sex and maternal race using interaction terms. In the second study, we tested the hypothesis that better pregnancy nutrition status characterized by higher HEI and plasma folate would be related to lower child blood pressure percentiles and reduced risk of HBP (Aim 2). Linear, Poisson and Generalized Additive Models were used with adjustments for socio-demographics, anthropometric measurements, behavioral factors, maternal stress and child nutrition. Interactions of maternal nutritional measurements with child sex, maternal race, pre-pregnancy maternal overweight or obesity, maternal smoking and breastfeeding for the associations of interest were explored. Results: In the first study with 822 mother-child dyads, mean PM2.5 and NO2 in the prenatal period were 10.8 µg/m3 and 10.0 ppb, respectively, and 9.9 µg/m3 and 8.8 ppb from birth to age 4 birthday. 29.2% of the children were classified as HBP, largely driven by isolated diastolic HBP. There was an estimated 13.6 percentile increase in SBP (Beta: 13.6, 95%CI: 3.7, 23.5) and an 8.0 percentile increase in DBP (Beta: 8.0, 95%CI: 0.7, 15.3) for each 2 µg/m3 increase of PM2.5 in the 2nd trimester. PM2.5 averaged over the full prenatal period was only associated with higher DBP percentiles (Beta: 11.6, 95%CI: 3.0, 20.2). The adverse effects were more pronounced in children whose mother had low-quality diet or possible folate deficiency during pregnancy, female children and children with a mother self-identified as Black. We detected no association of NO2, road proximity and postnatal PM2.5 with any outcomes. In the second study with 846 mother-child dyads, mean HEI and folate were 60.0 (SD: 11.32) and 23.1 (SD: 11.1) nmol/L respectively. 29.6% of the children were defined as HBP based on measurements at one visit. Maternal HEI and plasma folate were not associated with child BP percentiles and HBP in the full cohort. The conclusion remained the same after adjustments for potential mediators and in sensitivity analyses. Among mothers self-identified as White, there was an inverse relationship between maternal HEI and child SBP percentile (Beta: -0.45, 95%CI: -0.80, -0.09). Maternal HEI 59 and above was associated with reduced HBP in girls (IRR: 0.57, 95% CI: 0.34, 0.96). A moderate non-linear relationship was suggested for maternal plasma folate and child SBP percentile in women with pre-pregnancy overweight or obesity. It was also indicated that child nutrition may confound the non-linear maternal folate-child SBP associations. We found no evidence of modified associations by maternal smoking and breastfeeding. Conclusion: The findings from this dissertation add to the limited evidence that higher prenatal PM2.5 exposure, particularly in the 2nd trimester, is associated with elevated early childhood blood pressure. The results also suggest that healthy maternal nutrition may potentially ameliorate the adverse cardiovascular effects of air pollutants. However, no independent association between maternal nutrition during pregnancy and childhood blood pressure was detected. This dissertation contributes to the evolving scientific inquiry regarding developmental origins of disease. The evidence generated from this dissertation can be used to inform regulatory policy on acceptable air pollution levels, and to raise interest for further studies to understand the role of pregnancy dietary interventions in child health.