PFOA Epidemiology and Cancer Toxicity

RoseyFields evaluates the human epidemiological evidence for Perfluorooctanoic Acid (PFOA) as follows.

1. Study Populations and Data Sources

The epidemiological database for PFOA is unusually large and diverse for an environmental chemical, comprising several hundred published studies drawn from at least four distinct categories of study populations. Each population type carries its own exposure profile, strengths, and limitations, and the consistency of findings across these different populations is a central pillar of the causal case for PFOA.

Occupational cohorts consist of workers at PFOA manufacturing or processing facilities, principally the DuPont Washington Works plant in Parkersburg, West Virginia, and the 3M facility in Minnesota. These cohorts experience the highest measured serum concentrations, at times exceeding 1,000 ng/mL, and have been followed for cancer and non-cancer mortality and morbidity over decades. Because exposure is occupational and well above background, these studies offer statistical power to detect effects that might be too subtle to observe in the general population, though they are also more susceptible to healthy-worker survivor bias and limited by small numbers of specific disease outcomes.

Community or contamination-site cohorts are drawn from residents living near PFOA-emitting facilities, most notably the C8 Health Project, which enrolled more than 69,000 residents of six contaminated water districts in Ohio and West Virginia who had consumed drinking water impacted by releases from the Washington Works facility. Members of this cohort had serum PFOA concentrations markedly elevated above general-population levels (means in the low-to-mid tens of ng/mL, and considerably higher for some subgroups), providing a bridge between occupational and general-population exposure ranges. The C8 Health Project’s size and the fact that PFOA was overwhelmingly the dominant background contaminant in this population (rather than a mixture of correlated PFAS) make it particularly valuable for isolating PFOA-specific effects.

General-population biomonitoring studies rely primarily on the National Health and Nutrition Examination Survey (NHANES), an ongoing, nationally representative survey of the U.S. population conducted by the Centers for Disease Control and Prevention that has measured serum PFOA since 1999. NHANES-based analyses allow investigators to examine associations at exposure levels representative of ordinary background exposure (population medians typically in the low single-digit ng/mL range), which is critical for establishing that effects are not confined to unusually high occupational or contamination-site exposures.

Birth cohorts and prospective pregnancy cohorts, including the Faroe Islands birth cohorts, the Danish National Birth Cohort, and similar cohorts in Norway, Greenland, and elsewhere, enroll pregnant women and follow their children prospectively, often for a decade or more, measuring PFOA in maternal and/or cord serum and correlating it with subsequent child health outcomes such as vaccine antibody response, birth weight, and growth. Because exposure is measured before the outcome occurs, these prospective designs substantially reduce the risk of reverse causation that affects cross-sectional analyses.

The use of multiple, diverse populations strengthens causal inference in several complementary ways. First, replication of an association across occupational, community, and general-population settings with widely differing exposure magnitudes and co-exposure profiles reduces the likelihood that any single finding is a chance result or an artifact of a particular population’s demographics or confounding structure. Second, because different populations are exposed to different mixtures of co-occurring PFAS and other contaminants, consistent findings across populations with different co-exposure profiles (for example, the C8 Health Project population, in which PFOA exposure predominated over other PFAS) help rule out confounding by correlated chemicals as the explanation for observed associations. Third, the combination of retrospective/cross-sectional general-population data with prospective birth-cohort data allows investigators to address temporality and reduce concern about reverse causation. Finally, studies spanning life stages (prenatal, childhood, adult) permit assessment of whether effects are stage-specific (e.g., developmental immunotoxicity), which is relevant both to biological plausibility and to identification of susceptible subpopulations.

2. Overview of Human Health Outcomes Investigated, by Body System

Consistent with the U.S. EPA’s 2024 systematic review of PFOA (which screened over 450 epidemiological studies) and supporting analyses in the ATSDR Toxicological Profile for Perfluoroalkyls, human epidemiological investigation of PFOA has spanned essentially every organ system. The evidence can be organized into three tiers.

Outcomes with strong, consistent, quantitatively modeled evidence — those for which EPA’s systematic review concluded the evidence “indicates” (i.e., is sufficient to support that PFOA exposure is likely to cause the effect in humans under relevant exposure conditions), and which EPA used to derive quantitative toxicity values — are: hepatic injury (elevated serum liver enzymes, principally alanine aminotransferase, ALT); immune suppression (reduced antibody response to childhood vaccination); cardiovascular/lipid effects (elevated serum total cholesterol and LDL cholesterol); and adverse developmental effects (reduced birth weight and related measures of fetal growth restriction). Kidney (renal cell) cancer and testicular cancer are also supported by strong, consistent, replicated evidence and underlie EPA’s classification of PFOA as “Likely to Be Carcinogenic to Humans” and the C8 Science Panel’s independent “probable link” determinations for both cancers.

Outcomes with suggestive but inconsistent evidence include: breast cancer (associations observed in some studies, particularly within genetically or hormonally susceptible subgroups, but null or inverse in others); thyroid disease and thyroid hormone perturbation; adverse reproductive outcomes in women, including preeclampsia, gestational hypertension, altered pubertal timing, and earlier menopause; adverse reproductive outcomes in men, including altered testosterone levels and semen quality parameters; non-malignant kidney function measures (reduced estimated glomerular filtration rate, elevated uric acid); increased infectious disease incidence in children (a biologically coherent companion finding to the vaccine-response data, but less consistently observed); ulcerative colitis; and ALT elevations in children (positive in girls but inconsistent in boys across studies).

Outcomes investigated with null or largely inadequate findings for a causal association include: most solid cancers other than kidney and testicular (colorectal, bladder, prostate, lung, and, in most well-designed studies, liver cancer); neurodevelopmental and neurobehavioral outcomes in children; non-Hodgkin lymphoma and other hematologic malignancies; musculoskeletal effects; and general markers of hematologic toxicity (white blood cell counts) beyond the immune-specific findings above. For several of these outcomes the epidemiological database is comparatively small, and available studies are frequently of low methodological confidence, such that the evidence is better characterized as inadequate to determine an association than as affirmatively null.

3. Strong-Evidence Outcomes: Key Studies, Exposure Metrics, and Effect Sizes

Hepatic injury. The most extensively replicated finding in the PFOA epidemiological literature is a positive, dose-related association between serum PFOA and serum ALT, a standard clinical biomarker of hepatocellular injury. In a large mid-Ohio Valley community cohort drawn from the C8 Health Project, modeled cumulative PFOA exposure was associated with increased ALT and an exposure-response gradient across exposure quintiles, with an odds ratio for elevated ALT of 1.16 (95% CI: 1.02, 1.33) comparing the highest to the lowest quintile. This finding has been replicated in general-population NHANES analyses and in an independent Swedish longitudinal cohort in which within-person changes in PFOA over time tracked with changes in liver enzymes, as well as in a Chinese occupational/community cohort. Positive associations with ALT (and, less consistently, AST) have also been reported in adolescent girls in a national biomonitoring study. Supporting evidence includes findings of altered liver histology metrics in a pediatric cohort, non-alcoholic fatty liver disease severity markers in obese adults with elevated PFOA and other persistent organic pollutants, and elevated mortality from liver disease/cirrhosis in a highly exposed occupational cohort. Median/geometric-mean serum PFOA concentrations in the informative studies range from roughly 1.3 ng/mL (general population) to occupational levels orders of magnitude higher, and the direction and approximate magnitude of the ALT association is consistent across this range, which EPA identified as a key basis for its “evidence indicates” hepatic hazard conclusion and for deriving a candidate reference dose.

Immune suppression. The strongest and most biologically coherent evidence for an immune effect of PFOA is reduced antibody response to routine childhood vaccination, particularly against tetanus and diphtheria. In the Faroe Islands birth cohorts, prenatal and childhood serum PFOA concentrations were inversely associated with antibody titers against diphtheria and tetanus measured at multiple ages, with associations reaching statistical significance at several time points. A follow-up analysis using multi-pollutant modeling to adjust for co-occurring PFAS found that the association with PFOA was not materially attenuated after accounting for PFOS, PFHxS, and PFNA, addressing a key confounding concern for this cohort. Similar reductions in antibody response following prenatal PFOA exposure were reported in a Norwegian birth cohort for rubella antibody, and diminished response to influenza vaccination was observed in highly exposed adults in the C8 Health Project — a population in which confounding by co-occurring PFAS is of less concern because PFOA exposure predominated. Additional supporting studies include a German cohort of one-year-olds, NHANES-based analyses of adolescent rubella and mumps antibody, a Faroese cohort followed to age 28, and a study in Guinea-Bissau children examining measles antibody response. Reported effect sizes in the Faroe Islands cohorts correspond to roughly a 10–20% reduction in antibody concentration per doubling of PFOA exposure, at median serum concentrations as low as approximately 1.1 ng/mL. EPA identified decreased anti-tetanus and anti-diphtheria antibody concentration in children as one of the co-critical effects underlying its final oral reference dose for PFOA.

Cardiovascular / serum lipid effects. A consistent, positive association between serum PFOA and total and LDL cholesterol has been observed across dozens of studies of adults in the general population, occupational cohorts, and contamination-site communities. In the C8 Health Project community, a study of more than 46,000 adults reported a clear positive, monotonic association between serum PFOA and total cholesterol and LDL cholesterol across exposure quartiles. A related analysis subsequently linked modeled cumulative PFOA exposure to increased risk of high cholesterol as well as coronary artery disease and hypertension in the same community and worker cohorts. In the general U.S. population, an analysis using multiple cycles of NHANES data (2003–2014) confirmed a positive association between serum PFOA and total cholesterol — an analysis that EPA relied upon (alongside the Faroese vaccine-response data and a Swedish birth-cohort study of birth weight) as one of three co-critical effects supporting its final reference dose. Findings for triglycerides have been less consistent than for total and LDL cholesterol. The association is observed across a wide range of exposure levels — from the general-population median of roughly 3–4 ng/mL to the substantially higher levels in the C8 community and occupational cohorts — and is corroborated by coherent findings of elevated blood pressure in some of the same cohorts.

Developmental effects (birth weight/fetal growth). Reduced birth weight associated with maternal PFOA exposure is among the most replicated findings in the developmental toxicology literature for PFOA. In an early and influential analysis within the Danish National Birth Cohort, higher maternal serum PFOA was associated with lower birth weight and increased odds of low birth weight after adjustment for potential confounders. This finding has since been replicated across multiple additional cohorts and summarized in a formal meta-analysis with bias analysis, and further supported by a Swedish cohort study that EPA selected as a co-critical, high-confidence study underlying its reference dose derivation. Across the epidemiological database, roughly two-thirds of informative studies reported some deficit in birth weight or a related fetal growth restriction measure (e.g., decreased birth length, head circumference), with the direction of effect notably consistent even where individual point estimates were imprecise. Median maternal serum PFOA concentrations in the informative studies generally range from about 1.1 to 5.2 ng/mL — squarely within the range of ordinary background exposure in pregnant populations at the time of sampling — supporting the conclusion that this effect is not confined to unusually high-exposure settings.

Cancer (kidney and testicular). The epidemiological cancer evidence for PFOA is strongest, and most consistent, for renal cell carcinoma (kidney cancer) and testicular cancer. Within the C8 Health Project population, one study reported elevated kidney cancer risk in the most highly contaminated water districts (adjusted odds ratios as high as 2.0, 95% CI: 1.3, 3.1), and a related analysis of the same cohort reported an elevated (though not statistically significant) hazard ratio for kidney cancer with increasing cumulative PFOA exposure. In the West Virginia occupational cohort, a mortality study found a statistically significant excess of kidney cancer mortality in the highest quartile of modeled cumulative PFOA exposure. More recently, a large, medium-confidence case-control study nested within the National Cancer Institute’s PLCO cancer screening trial found a statistically significant, dose-dependent increase in renal cell carcinoma risk with pre-diagnostic serum PFOA (odds ratio 2.63, 95% CI: 1.33, 5.20, highest versus lowest quartile; odds ratio 1.71 per doubling of exposure), an association that remained significant after adjustment for co-occurring PFAS. A critical review and meta-analysis of the epidemiological literature estimated a 16% increase in kidney tumor risk per 10 ng/mL increase in serum PFOA. For testicular cancer, studies of the C8 Health Project population reported statistically significant elevated risk with increasing PFOA exposure, and the International Agency for Research on Cancer’s independent review concluded that this association was “considered credible and unlikely to be explained by bias and confounding,” while cautioning that estimates were based on small case numbers. These findings, developed independently across the C8 community study and the West Virginia occupational cohort, provided the basis for the C8 Science Panel’s formal “probable link” determinations for both kidney and testicular cancer, and for EPA’s ultimate classification of PFOA as carcinogenic (discussed further below).

4. Study Quality Considerations

Confounding by co-occurring PFAS. Because PFOA is one of dozens of structurally related PFAS that frequently co-occur in contaminated water and in human serum, a recurring methodological challenge is distinguishing PFOA-specific effects from effects attributable to correlated PFAS (e.g., PFOS, PFHxS, PFNA) or to PFOA precursor compounds that metabolize in the body. Investigators have addressed this concern in several ways: through multi-pollutant regression models that adjust for co-exposures (as demonstrated in a re-analysis of the Faroese vaccine-response data, which found no material attenuation of the PFOA association after adjustment for other PFAS), and through reliance on populations such as the C8 Health Project community, in which PFOA was overwhelmingly the dominant PFAS exposure, reducing the plausibility that observed associations are attributable to correlated co-exposures. EPA’s own review concluded that available multi-pollutant analyses did not provide evidence that PFOA associations are fully attributable to confounding by co-occurring PFAS, though it acknowledged this as an area of ongoing uncertainty given the nascent state of statistical methods for chemical mixtures.

Reverse causation. For outcomes related to renal function in particular, reverse causation is a recognized methodological concern: because the kidney is a primary route of PFOA elimination, reduced kidney function could itself cause serum PFOA to rise (rather than PFOA exposure causing reduced kidney function), potentially generating a spurious inverse association between PFOA and glomerular filtration rate. A national biomonitoring analysis tested for this possibility directly and found the inverse association between PFOA and estimated glomerular filtration rate remained (and was, if anything, stronger) among adults without chronic kidney disease, arguing against reverse causation as the explanation. By contrast, a study of children and adolescents in the C8 Health Project found no association between kidney function and PFOA levels estimated from residential history years before enrollment, suggesting that the cross-sectional association observed at the time of enrollment could reflect reverse causation. This divergence illustrates why renal function measures remain in the “suggestive but inconsistent” tier rather than the “strong evidence” tier, and underscores the importance of prospective, rather than purely cross-sectional, exposure assessment for outcomes where reverse causation is biologically plausible. Reverse causation is also a recognized concern for at least one thyroid cancer study, where the authors themselves noted that thyroid pathology could alter PFAS metabolism and thereby distort the exposure-outcome relationship.

Exposure misclassification. Because PFOA has a long serum half-life (generally estimated at several years in humans), a single serum measurement is thought to reasonably reflect exposure over a period of years, which mitigates (but does not eliminate) concern about misclassification in studies relying on one-time measurements. Misclassification concerns are more substantial for studies that model historical exposure from residential history and environmental monitoring data (as in several C8 Health Project analyses) rather than measuring serum directly, and for outcomes with long latency (e.g., cancer), where a single contemporaneous serum measurement may poorly represent cumulative lifetime exposure or exposure during an etiologically relevant window. EPA specifically noted that declining background serum PFOA levels in the U.S. population since the early 2000s complicate the interpretation of studies using recent NHANES cycles as representative of historical peak exposure, and for its own dose-response modeling purposes selected exposure windows more likely to capture historical peak U.S. exposure.

Co-exposure to other chemicals and confounding by socioeconomic and demographic factors. Some populations studied — for example, those near industrial or firefighting-foam-impacted sites — may have elevated co-exposure to other environmental contaminants (e.g., polychlorinated biphenyls, methylmercury), which could confound associations if not addressed. In addition, a subset of studies has been criticized for incomplete adjustment for socioeconomic status, race/ethnicity, or lifestyle factors (diet, smoking, obesity) that could independently affect both PFOA body burden and the health outcome under study; EPA acknowledged this as a persisting data gap warranting further research in racially and ethnically diverse populations, while noting that observed associations for several outcomes (ALT, cholesterol, antibody response, birth weight, kidney cancer) have been replicated across populations with very different confounding structures, which increases confidence that residual confounding alone does not explain the pattern of findings.

5. Independent Expert Panel and Agency Weight-of-Evidence Conclusions

Several independent scientific bodies have reached formal weight-of-evidence conclusions regarding PFOA and human disease based substantially on the epidemiological record described above. The C8 Science Panel — a group of three independent epidemiologists appointed pursuant to a class-action settlement to evaluate the C8 Health Project data — concluded that a “probable link” exists between PFOA exposure and both kidney cancer and testicular cancer, as well as high cholesterol, thyroid disease, ulcerative colitis, and pregnancy-induced hypertension, based on its multi-year, independent epidemiological investigation of the Ohio/West Virginia community. The International Agency for Research on Cancer (IARC), in its 2016 review, concluded that the epidemiological evidence for testicular cancer was “credible and unlikely to be explained by bias and confounding.” Most recently, the U.S. EPA’s 2024 final toxicity assessment, developed under the IRIS program and subjected to external peer review by the EPA Science Advisory Board, concluded — based on an integration of epidemiological, animal toxicological, and mechanistic evidence — that PFOA is “Likely to Be Carcinogenic to Humans,” and separately concluded that the human epidemiological evidence “indicates” (the Agency’s second-highest confidence tier, reflecting a judgment that the evidence is sufficient to conclude the exposure is “likely” to cause the effect in humans) that PFOA causes hepatic injury, immune suppression, adverse cardiovascular/lipid effects, and developmental toxicity. The California Environmental Protection Agency reached a similar conclusion regarding kidney and testicular cancer. These weight-of-evidence determinations are notable in that they were each reached through structured, transparent, and independently reviewed processes that explicitly weighed study quality, consistency, exposure-response relationships, biological plausibility, and alternative explanations such as confounding and reverse causation.

6. Use of Epidemiological Evidence to Support Causation Opinions

Epidemiological evidence of the kind summarized above forms one of the two principal evidentiary pillars — alongside toxicological and mechanistic evidence in experimental animal and in vitro systems — that support both general and specific causation opinions in PFOA litigation. For general causation, the consistency of findings (particularly for hepatic injury, immune suppression, dyslipidemia, developmental effects, and kidney/testicular cancer) across occupational, community, and general-population studies conducted by independent research groups in different countries, using different study designs and exposure assessment methods, satisfies core elements of the Bradford Hill considerations traditionally used to evaluate causal inference in epidemiology — including consistency, strength and dose-response gradient of association, temporality (particularly in the prospective birth-cohort literature), and biological plausibility (which is independently supported by coherent mechanistic and animal toxicological findings, such as PPARα-mediated hepatic effects and altered lipid homeostasis in rodent models). The convergence of this epidemiological record with formal weight-of-evidence determinations by independent expert bodies (the C8 Science Panel, IARC, and U.S. EPA) provides additional support for a reasoned general causation opinion that PFOA exposure, at biologically relevant serum concentrations, is capable of causing the strong-evidence health outcomes identified above in the general population.

For specific causation in an individual case, this epidemiological literature provides the quantitative building blocks — dose-response relationships, relative risk estimates, and characterized exposure ranges — that allow an expert to evaluate whether a particular plaintiff’s individual exposure history (as reflected in serum PFOA measurements or reconstructed historical exposure) falls within ranges associated with increased risk in the epidemiological studies, and to characterize the magnitude of any resulting increase in risk or contribution to the plaintiff’s diagnosed condition, typically in conjunction with the plaintiff’s individual medical and exposure history, differential diagnosis, and consideration of alternative causes. Because the epidemiological database for PFOA is unusually mature relative to many other environmental chemicals — spanning decades, multiple continents, and diverse study designs, and having been subjected to structured, independent weight-of-evidence review by regulatory and scientific bodies — it provides an unusually well-developed foundation for both general and specific causation analysis relative to many other toxic tort contexts.

Bibliography