UC Berkeley

Recent studies suggest that phthalates, parabens, and other phenols found in personal care products and plastics may be related to immunologic diseases. This dissertation focuses on in utero exposure to these chemicals and the development of asthma, aeroallergies, and eczema in childhood, as well as lung function and relative concentrations of immune system biomarkers in childhood. We explored measurements of immune system functioning and atopic disease in children in the Center for the Health Assessment of Mothers and Children in Salinas (CHAMACOS) study, a prospective cohort of pregnant women and their children, from age six months to seven years and describes data collection and data consistency across ages. We determined classifications of cases of probable asthma, aeroallergies, and eczema at age seven in this population and found that, at age seven, 36 children out of 353 with relevant data (10%) had probable asthma, 87 children of 339 with relevant data (26%) had aeroallergies, and 23 children of 338 with relevant data (7%) had eczema. Lung function measurements were similar to national averages, and cytokine measurements changed with age in expected patterns.

We analyzed associations of prenatal urinary concentrations of high molecular weight phthalates and bisphenol A, commonly found in some plastics, with probable asthma, aeroallergies, eczema, and spirometry at age seven, and with blood cytokine measurements at ages two, five, and seven. Logistic and linear regressions were conducted for 392 children, and Bayesian Model Averaging (BMA) was used to identify a select number of additional personal care product and plasticizing chemicals to be controlled for in the models, in order to account for confounding by joint chemical exposure to all phthalates, parabens, and phenols studied in this dissertation. We found that concentrations of monocarboxyisooctyl phthalate (MCOP) were associated with increased odds of having probable asthma (OR: 1.54, 95% CI: 1.12, 2.12) and poorer lung function (RR for forced expiratory volume in one second [FEV1]: -0.09, 95% CI: -0.15, -0.03; RR for forced expiratory flow from 25–75% of FVC [FEF25-75%]: -6.98, 95% CI: -10.95, -2.84). MCOP (OR: 1.28, 95% CI: 1.02, 1.62), MCPP (OR: 1.36, 95% CI: 1.07, 1.73), and BPA (OR: 1.35, 95% CI: 1.06, 1.73) were associated with aeroallergies, in crude models and models adjusting for demographic factors, but not in models adjusting further for additional chemical concentrations.

We also examined associations between prenatal urinary concentrations of low molecular weight phthalates, parabens, triclosan, 2,4-dichlorophenol, 2,5-dichlorophenol, and benzophenone-3, all commonly found in personal care products, and the outcomes described above. Similar methods were used, including selection of additional chemical covariates using BMA. We found that concentrations of propyl paraben (OR: 0.84, 95% CI: 0.73, 0.97) and 2,5-dichlorophenol (OR: 0.63, 95% CI: 0.47, 0.86) were associated with decreased odds of having probable asthma. We also found that concentrations of monoethyl phthalate (MEP) were associated with lower spirometry measurements (RR for FEV1: -0.03, 95% CI: -0.07, 0.00; RR for FEF25-75%: -3.23, 95% CI: -5.87, -0.52), but that concentrations of mono-n-butyl phthalate (MBP) were associated with higher spirometry measurements (RR for FEV1: 0.06, 95% CI: 0.01, 0.11; RR for FEF25-75%: 4.29, 95% CI: 0.04, 8.71). We also found that MBP concentrations were longitudinally associated with higher T helper cell 2 percentage (Th2%) across ages two, five, and seven (RR: 8.40, 95% CI: 1.95, 15.26) and that mono-isobutyl phthalate (MiBP) concentrations were longitudinally associated with lower T helper 1:T helper 2 cell ratio (RR: -6.19, 95% CI: -11.70, -0.32).

We explored novel methods to analyze joint exposure to the chemicals studied in this dissertation, as people are typically exposed to mixtures of chemicals, rather than to a single chemical. We chose to analyze probable asthma and allergies, but not eczema, with these methods because eczema showed no associations in our regression analyses above. We also included only FEV1 out of the four spirometry measures, though results for other spirometry measures were similar. We chose not to include cytokine measurements because associations in the above analyses were not consistent and would likely not yield results useful for the demonstration of these new methods. Bayesian Profile Regression was used to cluster children into groups based on their patterns of joint prenatal exposure to phthalates, parabens, and other phenols. Clusters were then evaluated for their relationship to probable asthma, aeroallergies, and FEV1. We found that participants clustered into seven exposure groups, characterized by high and low levels of chemicals consistent with expected patterns of product use. However, these clusters were not related to the three immune outcomes. Bayesian Kernel Machine Regression (BKMR) was used to determine associations of each chemical, in the context of joint exposure to the other chemicals, with probable asthma, aeroallergies, and FEV1. We found that BKMR associations were similar to those found via logistic and linear regressions, and that strong associations of some chemicals had additive effects on associations of other chemicals.

We found evidence that in utero urinary concentrations of high molecular weight phthalates appear to be associated with asthma and adverse respiratory function in childhood, and possibly with aeroallergies in childhood. In utero urinary concentrations of low molecular weight phthalates appear to be associated with altered cytokine levels in early childhood, and concentrations of parabens and other phenols show limited or no association with respiratory or immune function in childhood.