Prenatal Exposures in the Maternal Birthing Population of Arizona: An Examination of Ambient Pesticides, Neighborhood Deprivation, and Gestational Diabetes Mellitus

Abstract

Background: Gestational diabetes mellitus (GDM) incidence has been rising for pregnant mothers over the last decade nationally and worldwide, posing a substantial threat to public health resources. The estimated prevalence of GDM is 7.8% in the United States (U.S). GDM is a glucose intolerance disorder that is diagnosed during pregnancy. GDM accounts for the majority of diabetes cases that occur during pregnancy (~90%), and usually subsides immediately after childbirth. However, recent meta-analyses suggest that mothers with prior GDM are at a greater risk of developing cardiometabolic diseases in the future. Compared to pregnant women with normal patterns of glycemia, women who develop GDM have an 8.3-fold increased risk of developing Type 2 Diabetes Mellitus (T2DM), with over half progressing to T2DM within 10 years after delivery.1,2 The reproductive population in Arizona residing in rural areas, tribal lands or near the United States-Mexico border remain at the highest risk for adverse birth outcomes, including GDM. Complicating GDM risk is the potential effect of joint exposures from maternal occupational work and residential exposures to neighborhood deprivation and agricultural pesticides during pregnancy. Studies have shown that living in a socially and economically deprived neighborhood can negatively influence the reproductive health trajectory of pregnant mothers, and children3. The aims of this dissertation research are to 1) determine magnitude of relationships between GDM and maternal exposures to occupational agricultural work, 2) examine risk of GDM and neighborhood deprivation in the birthing population of Arizona, 3) investigate the associations of selected agricultural pesticides (organophosphates, pyrethroid, and carbamates) on GDM. Methods: This research conducted a series of secondary data analyses using population-based datasets with geographic linking at the census tract level. The data comprise Arizona pesticide use reports, U.S. Census data, and birth records from 2006-2020 in Arizona. These data sources were linked to form the AzPEARS (Arizona Prenatal Environment And Reproductive outcomes Study). For Aim 1, using 2006-2013 birth data with parental usual occupation available neonates were identified as belonging to an agricultural or non-agricultural household to estimate the risk of macrosomia (>4,000 grams), postterm birth (>41 weeks), low birth weight (LBW <2,500 grams), pre-term birth (PTB <37 weeks), large-for-gestational-age (LGA), small-for-gestational-age (SGA), and 5 min-APGAR (<7). In Aim 2, we used 2010 U.S. Census tract sociodemographic data to create a composite neighborhood deprivation index (NDI) score and linked to 2014 to 2020 birth certificate data by maternal address to investigate the role of neighborhood deprivation on GDM diagnosis. Lastly, for Aim 3, birth certificate data for 2006-2020 were geographically linked with state pesticide application data. Using multivariable logistic regression modeling, adjusted odds ratios (aORs) and 95% confidence intervals (CIs) were estimated to estimate associations of GDM with ambient pesticide exposures using single pollutant models for organophosphate, pyrethroids, and carbamates. Results: Aim 1: Using the 2006-2013 data, positive associations of GDM were identified for occupational exposure to agricultural farm work on abnormal fetal growth. Newborns of agricultural households (n=6,371) had a higher risk of macrosomia (aOR 1.15, 95% CI: 1.05, 1.26), large for gestational age (aOR 1.12, 95% CI: 1.03, 1.22), postterm birth (aOR 1.20, 95% CI: 1.09, 1.33), and low 5-min APGAR (aOR 1.39, 95% CI: 1.07, 1.81), whereas low birth weight (aOR: 0.85, 95% CI: 0.76, 0.96) and preterm birth (aOR: 0.82, 95% CI: 0.74, 0.92) were inversely related. Aim 2: Between 2014 to 2020, the statewide prevalence of GDM in Arizona was 7.8%, with prevalence by county and neighborhoods ranging from 4% to 12%. Mothers living in higher NDI score quartiles had greater odds of GDM compared to those living in areas with the lowest levels of deprivation (p-trend <0.0001). Mothers living in the area with the greatest deprivation had 25% greater odds of GDM compared to mothers living in areas with the lowest levels of deprivation (95% CI: 1.21, 1.30). Our results revealed positive associations for several specific pyrethroids. In particular, bifenthrin, cyfluthrin, esfenvalerate, and permethrin were associated with elevated GDM risk. These effect estimates ranged from 1.15 to 2.39. However, exposure to the overall PYR class during early pregnancy was close to null (aOR=1.04, 95%CI: 0.94, 1.12). Any OP exposure during the third trimester increased GDM risk by 10% (95%CI: 1.00, 1.21), which was driven by the specific OPs bensulide, chlorpyrifos, tribuphos, and ethephon. Even after covariate adjustment, the effects of these exposures on GDM remained. We found no significant associations with prenatal ambient CAR exposures. Conclusion: In summary, this research contributes to the limited epidemiological studies on the increased maternal risk to GDM, particularly around neighborhood deprivation and pesticide exposures. Despite increasing concern about co-exposures of chemicals toxicants and neighborhood-level stressors and lifelong cardiometabolic maternal risk, few studies have examined their potential associations. This dissertation is an effort to contextualize one pregnancy complication and the co-occurrence of chemical contaminants and non-chemical stressors among vulnerable birthing populations that often spatially cluster to shape health outcomes.