Characterization of per- and polyfluoroalkyl substances in human serum: novel methods for assessing exposure
Embargo Date
2026-08-23
OA Version
Citation
Abstract
Per and polyfluoroalkyl substances (PFAS) are a large class of human made fluorinated chemicals, some of which are persistent in the environment. The small number of PFAS studied have been shown to be mobile in the environment and are routinely measured in humans. For those PFAS, epidemiological and toxicological studies have demonstrated that low exposures are linked to negative health impacts. Understanding exposure to PFAS in human populations is critical to the evaluation of associated health effects and to parsing the relative sources of PFAS. This dissertation examines PFAS definitions, unidentified organofluorine in serum, and the contributions of drinking water and diet to exposure, enhancing our understanding of human PFAS exposures.
Definitions of PFAS that group chemicals based on their structure have been proposed that range in their specificity and inclusion. These definitions have implications for the broader suite of organofluorine chemicals, yet little research has been done to evaluate what compounds are included under different definitions of PFAS. Naturally-occurring organofluorine compounds are rare, but organic fluorine is incorporated into many pharmaceuticals. Widely used organofluorine pharmaceuticals, including anti-depressants and cholesterol lowering medications, may commonly occur in human serum and in wastewater. It is not known how PFAS definitions used for categorizing, tracking, and regulating chemicals in commercial products, the environment, and humans vary in their inclusion of organofluorine pharmaceuticals.
Thousands of PFAS exist, yet routine human biomonitoring is limited to 7–12 legacy compounds, and much less is known on the occurrence of other PFAS. Expanded testing shows additional PFAS in drinking water and food packaging beyond those that are routinely monitored. Studies using organofluorine mass balance have shown 30–70% of extractable organofluorine (EOF) in serum from China is not explained by targeted PFAS, and research into the identification of unexplained organofluorine is needed. Large amounts of unknown organofluorine have potentially important implications for human biomonitoring. Sources of unknown organofluorine may include replacement PFAS that are not yet measured, pesticides, or pharmaceuticals, the latter frequently used at relatively high doses.
People can be exposed to PFAS from multiple and varied sources. Studies in highly exposed communities have shown associations with legacy PFAS in drinking water. While 99% of the U.S. population have detectable levels in serum, most people are not documented to live near large identifiable sources that may contaminate drinking water supplies, suggesting other sources are important. Studies in Europe have estimated the dominant source of exposure in general population to be diet, but regional differences in industry and unique dietary patterns make direct extrapolation to U.S. populations difficult. PFAS can enter foods via bioaccumulation from the environment or through contact with food packaging materials that are treated with PFAS for their grease-proofing properties. Many existing dietary studies in the U.S. use data from many years ago, but shifts in PFAS production over the last 20 years render older studies less reliable for explaining current exposure patterns. Water may also contribute to exposure in the general population, but few epidemiological studies consider contributions from diet and water together. Research on multiple exposure routes is needed in order to better characterize the extent to which sources contribute to exposure.
The objectives of this dissertation were to 1) assess the implications of different PFAS definitions for inclusion of organofluorine pharmaceuticals, 2) determine to what extent organofluorine pharmaceuticals contribute to unknown organofluorine in commercial serum from U.S. blood donors, and 3) to examine associations between PFAS body burden and measures of exposure via diet and drinking water for participants in the California Regional Exposure Study.
In Chapter 2, nine PFAS definitions were used to screen a comprehensive list of the 360 organofluorine pharmaceuticals approved and used globally between 1954–2021. Definitions ranged in their inclusion of organofluorine pharmaceuticals (1%–100%). The most inclusive definitions included several top prescribed pharmaceuticals, e.g., Prozac and Lipitor. Using pharmaceuticals as an example, we showed that the specificity of a given definition and its usefulness for its intended purpose depends on the context in which the definition is applied.
Chapter 3 used combustion ion chromatography to measure EOF in commercial serum from U.S. blood donors. Using fluorine mass balance, we assessed differences in unexplained organofluorine (UOF) associated with pharmaceutical use and compared them with concentrations of organofluorine predicted based on the pharmacokinetic properties of each drug. Pharmacokinetic estimates of organofluorine attributable to pharmaceuticals ranged from 0.1 to 55.6 ng F/mL. For comparison, the geometric mean concentration of PFOS in U.S. adults from 2017–2018 as reported in NHANES was equivalent to 2.9 ng F/mL. Analysis of 44 target PFAS and EOF in samples of commercial serum (n = 20) showed the fraction of EOF not explained by Σ44 PFAS ranged from 15% to 86%. Compared to donors not taking pharmaceuticals, self-reported use of organofluorine pharmaceuticals was associated with a 0.36 ng F/mL (95% CL: −1.26 to 1.97) increase in UOF. Discrepancies between pharmacokinetic estimates and EOF may be partly explained by methods of measuring EOF. Findings from this study suggested pharmaceuticals can contribute to EOF, but large amounts of EOF remain unexplained.
Chapter 4 assessed contributions from drinking water and diet to levels of PFAS in serum collected from 700 adults living in southern and eastern California in 2018–2020. Using survey information on dietary consumption and levels of PFAS in drinking water measured in the U.S. Environmental Protection Agency’s third Unregulated Contaminant Monitoring Rule (UCMR 3), we used robust linear regression to examine the association between non-transformed serum concentrations, diet, and drinking water. Findings from this analysis suggest drinking and diet both contribute to exposure in this population. Fewer associations were observed for diet compared to epidemiological studies from many years ago. There are multiple possibilities for these unexpected results, including diminished levels in foods and corresponding reductions in serum concentrations, limitations of dietary assessment methods to measure patterns of diet, as well as potential confounding not adequately accounted for in our analysis.
This dissertation illustrates the importance of clearly defining PFAS and provides recent data on the occurrence of known and unknown organofluorines in human serum, as well as potential exposure sources in the general population. These findings can guide future human biomonitoring studies and inform source reduction strategies.