Ubiquitous Environmental Pollutants and Liver Outcomes: Per- and Polyfluoroalkyl Substances (PFAS) and Outdoor Air Pollution
Date of Award
Spring 2024
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Public Health
First Advisor
Vasiliou, Vasilis
Abstract
Per- and polyfluoroalkyl substances (PFAS) are widely distributed in the environment and in biosamples from different populations around the world. They have been associated with induction or progression of liver injury in numerous studies, but mechanisms of action remain unclear. Fine particulate matter (PM2.5) is another ubiquitous environmental pollutant that has been associated with liver cancer incidence and mortality in a limited number of studies. The specific aims of this dissertation were to 1) implement an AI-assisted screening tool to synthesize relevant studies on the mechanisms of liver injury induced or progressed from exposure to PFAS from a large body of literature and to assess the reliability of the screening tool; 2) investigate the association of PFAS, both independently and jointly with alcohol intake, on liver function biomarkers in a sample of the U.S. general population; 3) assess the association between PM2.5 exposure and liver cancer for the first time in a U.S. cohort with historical exposure assessment.The review conducted in the first aim followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews guidelines. PFAS of interest were those of identified research interest and being monitored by the U.S. Environmental Protection Agency. We searched 7 scientific databases for relevant articles and used citation chaining for review articles. One reviewer conducted manual screening of titles and abstracts and a second reviewer applied ASReview, a publicly accessible AI-assisted screening tool. Full texts were screened manually by both reviewers. A total of 7,211 papers were found from the manual search and N=1,465 papers were screened via ASReview. The agreement between AI-assisted and manual review was high (94.8%; Cohen’s k=0.70). The final set for data extraction included 330 papers. The majority of the studies focused on the direct effects of PFAS on liver injury, while a smaller number explored detoxification with antioxidant supplements after PFAS exposure. We identified six thematic areas of mechanisms across studies, including metabolic interaction, oxidative stress and inflammation, and epigenetic and signaling alteration. In addition to summarizing key mechanisms underlying PFAS-induced liver injury, this review demonstrated the reliability of AI-assisted reviews for efficient screening of a large body of scientific literature. The second aim employed data from the National Health and Nutrition Examination Survey (2003-2016; N=11,794) and examined the five most historically prevalent PFAS. Adjusted linear regression was used to evaluate associations between PFAS and continuous liver function biomarkers and adjusted logistic regression to assess markers dichotomized by 95th percentile. Potential interactions between PFAS and alcohol consumption and sex were evaluated via stratified analyses and conducted sub-analyses adjusting for daily alcohol intake among those with available drinking history (N=10,316). Consistent with other studies, we found a positive association between serum perfluorooctanoic acid (PFOA) and high alanine transaminase levels, with no monotonic trend (ORQ4vsQ1=1.45, CI: 0.99-2.12; p-trend=0.18), and a positive association with aspartate transaminase when modeled continuously (ORcon=1.15, CI: 1.02-1.30; p-trend=0.03). Perfluorooctane sulfonate (PFOS) and perfluorohexane sulfonate (PFHxS) showed a negative association with alkaline phosphatase (ALP), with trend observed only for PFHxS (p=0.02). PFOA was inversely but non-montonically associated with ALP (p-trend=0.10). The highest quartile of PFOS was associated with high total bilirubin (ORQ4vsQ1=1.57, CI: 1.01-2.43, p-trend=0.02). We did not find statistically significant associations between any PFAS and γ-glutamyl transpeptidase and no associations for perfluorodecanoic acid and perfluorononanoic acid. We found some suggestive evidence of interactions between PFAS and alcohol intake, particularly among heavy drinkers, highlighting the need for further assessment of this their joint effects. The third aim evaluated the association between ambient PM2.5 exposure and risk of liver cancer among 499,729 participants from the NIH-AARP Diet and Health Study, a cohort in 6 states (California, Florida, Louisiana, New Jersey, North Carolina, and Pennsylvania) and 2 metropolitan areas (Atlanta, Georgia, and Detroit, Michigan). We used spatiotemporal prediction models to estimate annual average historical PM2.5 concentrations (1980-2015) at participant enrollment addresses. The cohort was enrolled in 1995-1996 and followed up through 2017. A time-varying Cox model was used to estimate the association for liver cancer overall and the predominant histologic type, hepatocellular carcinoma (HCC), per 5 µg/m3 increase in estimated outdoor PM2.5 levels. The model employed a time-varying, 5-year average exposure, lagged 10 years prior to cancer diagnosis and was adjusted for age, sex, race/ethnicity, education level and catchment state. We also stratified the analyses by hypothesized effect modifiers. We observed a non-significantly increased risk of liver cancer associated with estimated PM2.5 concentrations (Hazard ratio [HR]=1.05 [0.96-1.14], N=1,625); associations were slightly stronger for HCC, (84% of cases; HR=1.08 [0.98-1.18]). We observed geographic variation of PM2.5 levels across states and suggestive variation in state-specific associations, although none were statistically significant (p-interaction>0.05). Participants aged 70 or older at enrollment had an increased risk of liver cancer compared with other age groups (HR=1.50 [1.01-2.23]); p-interaction=0.01) and the association was also increased among participants who did not exercise (HR=1.81 [1.22-2.70]; p-interaction=0.01). We found no evidence of effect modification by sex, smoking status, body mass index, diabetes status, or alcohol consumption (p-interaction>0.05). In this large prospective cohort, our results suggested a potential association between residential ambient PM2.5 levels and liver cancer risk, and further research would benefit from further exploration of potential heterogeneity in these associations by age and physical activity.
Recommended Citation
Ma, Xiuqi, "Ubiquitous Environmental Pollutants and Liver Outcomes: Per- and Polyfluoroalkyl Substances (PFAS) and Outdoor Air Pollution" (2024). Yale Graduate School of Arts and Sciences Dissertations. 1257.
https://elischolar.library.yale.edu/gsas_dissertations/1257
