PFOA: What the C8 Science Panel Found About This Chemical

The PFAS overview covers the broader class. The endocrine disruptors guide covers why persistent synthetic chemicals in the food chain matter more than the food chain lets on.

One hedging note before the research. The C8 Science Panel's findings are framed as 'probable link' rather than 'causes.' That phrase is doing real scientific work, not legal hedging — it reflects the specific evidentiary standard the Panel was contractually required to meet, and it appears in their peer-reviewed publications. In late 2023, however, a separate body — the International Agency for Research on Cancer — moved past that hedge for two specific endpoints, upgrading PFOA to Group 1 carcinogen in IARC Monograph Volume 135. That's the strongest language the international cancer research community has for any chemical.

PFOA is the common name for perfluorooctanoic acid — a synthetic surfactant built around an eight-carbon backbone where every hydrogen atom has been replaced by a fluorine. The chain length is why it was nicknamed C8Common industry and litigation nickname for PFOA — the 'C' for carbon, the '8' for the eight-carbon backbone of the molecule. in internal DuPont documents, and why the court-ordered epidemiological study that followed the Parkersburg exposure took the same name. The carbon-fluorine bond is one of the strongest single bonds in organic chemistry — the reason PFOA works at all (it repels water, oil and grease) and the reason it doesn't meaningfully degrade in soil, rivers or living things.

3M began producing PFOA at its Cottage Grove, Minnesota facility in 1947, using the Simons electrochemical fluorination process, and continued until 2002. DuPont began purchasing PFOA from 3M in 1951 to use as a processing aid in the manufacture of PTFE, the fluoropolymer marketed under the Teflon brand. PTFE-coated cookware entered the consumer market in 1956, when Marc Grégoire's Tefal company launched the first commercial non-stick frying pan in Sarcelles, France. Between 1951 and 2003, DuPont's Washington Works plant in Parkersburg, West Virginia released more than 1.7 million pounds of PFOA released to air, water and soil by DuPont Washington Works 1951-2003 of PFOA to air, water and soil — discharging into the Ohio River, into landfills, and into the drinking water of six downstream water districts.

The human biological half-life of PFOA — how long it takes your body to clear half of a given exposure — is unusually long for a small molecule. Twenty-six retired fluorochemical production workers at 3M's plants, followed over five years, showed a geometric mean serum elimination half-life of 3.5 years half the PFOA in your blood today will still be there in 2030 for PFOA (95% CI 3.0-4.1) Olsen et al. 2007. A 3.5-year half-life means that if you stopped all exposure today, you'd still be carrying roughly one seventh of your current body burden a decade from now.

Industry toxicology on PFOA began early. In 1961, DuPont's own Haskell Laboratories toxicology chief Dorothy Hood documented that Teflon-related materials enlarged the livers of rats at low doses and recommended they be 'handled with extreme care.' Gaber, Bero and Woodruff (2023, Annals of Global Health) traced this and later internal findings through the document production forced by Bilott's litigation Gaber Bero Woodruff 2023. Twenty years after the 1961 rat study, in 1981, 3M shared new rat data with DuPont showing that PFOA caused fetal eye defects in offspring of exposed mothers. DuPont responded by surveying its own female workers on the Washington Works Teflon line — and found that of seven live births to exposed women during 1979-1981, two of seven DuPont Washington Works live births with facial birth defects 1979-1981 had facial birth defects: one infant with a single nostril and eye abnormalities, one with eye and tear-duct defects. DuPont reassigned women of childbearing age off the C8 exposure jobs and did not disclose the human birth-defect data to the EPA.

Internal DuPont decision-making on PFOA is documented through the company's own meeting records. In May 1984, executives met at corporate headquarters to review four options for C8 use: stop using it entirely, continue with pollution-control investment, reduce production, or continue with no abatement. Meeting notes, later produced in litigation, record that 'none of the options developed are economically attractive.' DuPont chose the last option — continue with no abatement — and subsequently increased production. A separate Haskell Laboratories internal science summary in 1988 compiled what the company already knew about PFOA's 'dangerous qualities,' citing in-house studies from 1969 through 1981 and confirming liver and eye effects. The 1988 document sat inside the company. It was not published.

Bilott's breakthrough came in 1999, when he filed a federal suit on behalf of the Tennant family and attached a discovery demand for every DuPont internal document referencing C8. The resulting production ran into tens of thousands of pages — by early 2002, Bilott's team had reviewed roughly 185,000 pages of DuPont and 3M internal documents on C8 reviewed by Bilott team by early 2002 of DuPont and 3M documents — including the 1961 rat study, the 1981 pregnancy survey and the 1984 corporate decision. The Leach class action followed in August 2001, settling in 2004-2005 for more than 300 million dollars: cash to the class, a medical screening programme, water filtration installation for the six affected water districts, and — the part that mattered for science — an independent epidemiological study funded by DuPont but run by scientists the court appointed. None of this was a surprise inside DuPont. The surprise was that someone outside the building finally saw the memos.

The C8 Science Panel — Kyle Steenland (Emory University), Tony Fletcher (London School of Hygiene and Tropical Medicine) and David Savitz (then Mount Sinai, later Brown University) — ran from 2005 and delivered its probable-link findings by late 2012. The baseline was the C8 Health Project described by Frisbee et al. (2009, Environmental Health Perspectives), which enrolled 69,030 residents of the six affected water districts — an 80% response rate from the exposed community residents of the six affected water districts across 2005-2006 — an 80% response rate from the exposed community Frisbee et al. 2009. Mean serum PFOA in the enrolled cohort was roughly five times the contemporary US general population reference. The Panel's contractual definition of 'probable link' was explicit: more likely than not, based on consistent evidence across multiple study designs.

By the end of the Panel's work, six conditions had met that bar. Barry, Winquist and Steenland (2013, Environmental Health Perspectives) worked with 32,254 cohort participants and 2,507 validated incident cancers across 21 cancer types, and reported elevated hazard ratios in the highest cumulative-exposure quartile: HR 1.58 kidney cancer top cumulative PFOA quartile vs lowest, Barry 2013 for kidney cancer and HR 3.17 testicular cancer top cumulative PFOA quartile vs lowest, Barry 2013 for testicular cancer, both relative to the lowest exposure quartile Barry Winquist Steenland 2013. Steenland, Fletcher and Savitz's earlier 2010 synthesis review had concluded that the evidence was still insufficient to draw firm conclusions at that point Steenland Fletcher Savitz 2010 — the strong findings on kidney and testicular cancer came later, in the Barry 2013 analysis and the Panel's formal probable-link determinations that same year. 'Probable link' is the Panel's own language. It reflects consistent associations across multiple analyses but stops short of claiming a fully proven causal relationship — in chemical epidemiology, that's a strong finding.

A complementary geographic analysis of 18 cancer types across 13 Ohio and West Virginia counties using 1996-2005 incidence data Vieira et al. 2013 reported positive associations with kidney, testicular, prostate and ovarian cancers and non-Hodgkin lymphoma in the most highly exposed residents. Breast and pancreatic cancers did not show associations. A deliberate note on what the Panel did not find: several conditions sometimes linked to PFAS in popular reporting — including diabetes, infertility, breast cancer — did not meet the 'probable link' threshold in this cohort. The strongest scientific statement about PFOA health effects is the Panel's statement. Claims stronger than that should be treated as hypothesis, not finding — except for the two cancers IARC has now elevated to Group 1.

The short answer: the production is phased out, the molecule isn't. The EPA's 2010/2015 PFOA Stewardship Program launched in 2006 with voluntary commitments from eight fluorochemical manufacturers (3M, DuPont, Arkema, Asahi, Ciba, Clariant, Daikin and Solvay Solexis) to achieve a 95% reduction by 2010 and complete elimination of PFOA from US production and related long-chain compounds by the end of 2015. That phase-out happened on schedule. The replacement chemistries — particularly HFPO-DA, marketed as GenX and produced by Chemours at the Fayetteville Works plant on the Cape Fear River in North Carolina — took over the same applications. The story of regrettable substitution is covered in the BPA-free article and applies here nearly verbatim.

Environmental persistence means legacy PFOA is still moving through the hydrological cycle, still accumulating in fish, still in food packaging and imports — and, because of the 3.5-year human half-life, still in your blood, slowly declining. The US EPA finalised a national drinking water standard in April 2024 setting enforceable maximum contaminant levels of 4 parts per trillion EPA 2024 PFOA and PFOS MCL for PFOA and PFOS individually, with monitoring required by 2027 and compliance by 2031 (originally 2029, extended in May 2025) EPA 2024 PFAS NPDWR. The agency estimated 6-10% of US public water systems — several thousand utilities — would initially exceed one or more of the new limits. That's a large number of water districts doing filter installations on an aggressive timeline.

In November 2023, IARC's Working Group reclassified PFOA from Group 2B (possibly carcinogenic) to Group 1 carcinogenic to humans in Monograph Volume 135, based primarily on the renal and testicular cancer evidence from the C8 cohort and replicated in other occupational studies. That is the same classification IARC uses for asbestos, benzene and tobacco smoke. It does not mean PFOA is as dangerous as those substances — IARC classifies the quality of evidence, not the magnitude of risk. But it does mean the evidence for human carcinogenicity is now considered sufficient. The Panel's 'probable link' has been replaced, for two specific cancers, by a stronger scientific statement.

For the average person in 2026, active PFOA exposure comes mostly from drinking water (if you live near a historical manufacturing or AFFF site), legacy cookware, and imported consumer goods. For background exposure from the food chain, individual-level mitigation is limited — this is what the regulation exists for. For the exposures you can control, the practical interventions are narrow.

The thing to remember about PFOA, more than any other chemical in this library, is that the science was present inside industry labs for decades before it reached a regulator's desk. The 1961 rat study was known. The 1981 worker pregnancy survey was known. The 1984 corporate decision to keep using the chemical was documented in a meeting. What changed was not the science. What changed was that a farmer's dying cattle in Parkersburg forced the documents into a courtroom, which forced the numbers into peer review, which forced regulators to catch up to what industry had already written down.

PFOA is the most-studied member of a family that is still being charted. The research base is now strong enough for the IARC classification to move from 'possibly carcinogenic' to 'carcinogenic to humans' on the strength of kidney and testicular cancer evidence alone. The specific molecule is phased out of US production. None of that means it's gone — the environmental persistence is the whole story, and a 3.5-year biological half-life means population serum levels will be dropping for decades rather than years. The practical advice is narrow: filter your water if you live near a site, replace legacy cookware if it's scratched, and read the drinking water reports your local utility is now required to publish. The broader advice is on the PFAS class profile, which covers the substitutes that are still in use in 2026.