Background: Volunteer firefighters are a critical public safety population who comprise 65% of the U.S. fire service, yet they are an understudied occupational group. Firefighters, including volunteers, are occupationally exposed to multiple toxicants, including per- and polyfluoroalkyl substances (PFAS). PFAS are a class of thousands of chemicals of public health and environmental concern due to their widespread occurrence, use, and persistence. PFAS exposures can vary within and between populations. PFAS are also associated with several adverse health outcomes, including dyslipidemia. Dyslipidemia is a causal risk factor for cardiovascular disease (CVD). This is of relevance to the fire service as cardiac events are the primary cause of line-of-duty deaths among all firefighters. However, volunteer firefighters’ PFAS exposures have not been well characterized. Nor have the ways in which volunteer firefighters’ PFAS exposures may affect their blood lipid levels. Objective and Specific Aims: This objective of this dissertation was to investigate variation in and predictors of serum PFAS exposures and associations between serum PFAS and lipids among volunteer firefighters enrolled in the Cancer Assessment and Prevention Study (CAPS), a partner of the national Fire Fighter Cancer Cohort Study (FFCCS). Three specific aims supported this objective: (1) Assess the distribution and predictors of PFAS exposure in a nation-wide cohort of U.S. volunteer firefighters; (2) Characterize serum lipid profiles and examine associations between serum PFAS and dyslipidemia in a nation-wide cohort of U.S. volunteer firefighters; (3) Examine variation in PFAS serum profiles in a cohort of New Jersey (NJ) volunteer firefighters and explore the utility of state vs. national comparison groups in biomonitoring research. Methods: This dissertation included volunteer firefighters enrolled in CAPS from 9 U.S. states. Cross sectional measures collected at study enrollment were used in all three aims. Sociodemographic, health, fire service, and occupational data were collected using an electronic survey. Blood was also collected at study enrollment, from which 12 PFAS and 7 lipid measures were quantified in serum. To contextualize volunteer firefighters’ serum PFAS exposures, data from the 2017 – March 2020 cycle of the National Health and Examination Survey (NHANES) were used to construct a nationally representative reference population to compare serum PFAS measures with CAPS in Study 1 and Study 3. In Study 3, data from the New Jersey (NJ) Department of Health’s NJ Biomonitoring Study (NJBM) were also used to construct state- and county-level reference populations. In Study 1, PFAS detection frequencies, average concentrations, and distributions were characterized overall and by regional groups among 550 active CAPS volunteer firefighters. Multivariable generalized estimating equations were used to characterize associations between fire service, sociodemographic, health, and occupational predictors of interest and serum PFAS. In Study 2, associations between each of the 12 serum PFAS measures and each serum lipid measure were characterized in 354 active volunteer firefighters under age 50 using general linear models. The effects of PFAS mixtures on lipids were assessed using Bayesian Kernel Machine Regression and quantile g-computation. In Study 3, 210 male CAPS volunteer firefighters from six NJ counties were included. Geometric means (GMs), detection frequencies (DFs), and 95% confidence intervals (CIs) for six serum PFAS were calculated in each population (CAPS, NHANES, and NJBM). For the CAPS-NJBM analyses, NJ was divided into five county-based regions to examine geographic differences in serum PFAS profiles. Results: In Study 1, CAPS volunteer firefighters’ average serum PFOS, PFOA, PFHxS, PFDA, and PFUnDA concentrations were lower than NHANES 2017-March 2020 non-Hispanic white adult males, however, the temporal differences between CAPS and NHANES serum collection may have resulted in higher NHANES concentrations than may be expected in a temporally comparable reference population. CAPS participants’ serum PFAS profiles varied geographically and by sex. Stronger associations with serum PFAS were observed with sociodemographic and health factors, with less strong associations with occupational and fire service factors. In Study 2, total cholesterol (TC), low-density lipoprotein, triglycerides, and non-high-density lipoprotein were positively associated with PFOA, PFOS, PFNA, PFHxS, and PFDA. A doubling of serum PFOA was associated with a 6.19 mg/dL (95% CI: 1.69 – 10.68) increase in TC. Associations were less precise between these and other PFAS compounds with other lipid measures. Mixture models were generally consistent with single-pollutant models. In Study 3, average PFOA and PFNA concentrations were slightly higher among NJ CAPS than NHANES non-Hispanic white males but were similar or lower than NJBM males. At the county group level, CAPS participants had similar or lower serum PFAS concentrations as NJBM except for one north-NJ county group. However, the temporal differences between serum collection dates for NHANES, NJBM, and CAPS may be driving these findings. Distinct advantages and limitations of different reference groups used in biomonitoring studies include temporal, geographic and study design factors. These factors may substantially impact the interpretation of biomonitoring results across populations. Conclusion: The findings from this dissertation contribute to the literature on volunteer firefighters, an understudied population who comprise 65% of the U.S. fire service. The results provide evidence that, on average, volunteer firefighters may have similar exposures to several PFAS compared to national and state-level populations. However, the results point to a likely multifaceted PFAS exposure experience that may differ according to individual and population level factors. Strong associations between serum PFAS and several lipid measures indicate that tailored interventions to reduce PFAS exposure and improve cardiometabolic health may benefit this critical public service population.
Katherine A Lubina (Thu,) studied this question.