Date of Award

9-1-2023

Language

English

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

College/School/Department

Department of Environmental Health Sciences

Dissertation/Thesis Chair

Erin M Bell

Committee Members

Steven Forand, Sanghamitra Savadatti, Xiaobo Romeiko, Elizabeth Lewis-Michl, Stephen Shost

Keywords

determinants, exposure assessment, maternal health, perfluoroalkyl, predictors

Subject Categories

Environmental Health

Abstract

Background: Per- and polyfluoroalkyl substance exposure has been implicated as a factor contributing to the development of many adverse health outcomes across populations of all ages. Despite this, the factors associated with PFAS concentrations in humans are not well understood. Most studies evaluating factors associated with PFAS concentrations consider a small number of variables. Differences in the factors associated with PFAS levels in populations of high exposure versus low or background exposure (i.e., the general population) have not been considered. Advancing our understanding on this topic is necessary to better understand the serum elimination half-lives and toxicokinetics of PFAS. Additionally, it may aid public health practitioners in identifying vulnerable groups so appropriate guidance to medical professionals and regulation to reduce exposure can be developed. This study aimed to identify factors associated with PFAS levels in three populations: highly-exposed (i.e., the communities of Hoosick Falls and Newburgh, New York), pregnant mothers and neonates, and the general U.S. population. A secondary goal was to evaluate associations between serum PFOA and PFOS levels and health outcomes among biomonitoring participants in two highly-exposed communities.Methods: Based on the number of covariates collected and the population size, appropriate statistical approaches were employed. To identify factors associated with PFOA and PFOS levels among Hoosick Falls and Newburgh biomonitoring participants, respectively, we employed stepwise regression models. Logistic regression models adjusted for age group and sex were employed to evaluate associations between serum PFOA and PFOS among biomonitoring participants in Hoosick Falls and Newburgh, New York, respectively. Given the body of literature on determinants of PFAS concentrations in mothers, we conducted a systematic review to summarize the current state of the science and gaps in the literature, as well as offer suggestions for future research. Finally, we evaluate factors associated with legacy PFAS concentrations among a large population of the National Health and Nutrition Examination Survey (NHANES) participants across nine survey cycles, employing elastic net regression models to identify factors best explaining interindividual variance in PFAS levels. Results: Factors associated with PFAS concentrations varied by population. Age, sex, and sample collection date were identified as factors associated with PFAS levels among Hoosick Falls and Newburgh biomonitoring participants, with some differences in other factors that may be partially explained by contrasting community characteristics and discrepancies in questionnaire content between the communities. Among pregnant mothers, parity, breastfeeding history, maternal race and country of origin, and household income had the strongest and most consistent evidence to support their roles as determinants of legacy PFAS concentrations. When evaluating exogenous factors, demographic characteristics, socioeconomic factors, and general measures of health and healthcare access best explained interindividual variance in legacy PFAS concentrations. Evaluating endogenous factors, body measures, environmental contaminant concentrations, and complete blood count and standard biochemistry results were important factors associated with PFAS concentrations. Conclusions: Factors associated with PFAS concentrations varied by population and level of exposure, and clinical implications and implications for risk assessment may vary by population. Given the limitations of cross-sectional studies, especially for evaluating factors associated with PFAS concentrations in females given their additional routes of elimination compared to men, well-designed longitudinal studies with rich covariate data are needed to further advance our understanding on this topic.

Download

Share

COinS