C8 Science Panel + DuPont Washington Works, what we learned

The lede

The C8 Science Panel, Kyle Steenland, Tony Fletcher and David Savitz, was created in 2005 as part of the settlement of Leach v. E.I. du Pont de Nemours and Co., the class action brought on behalf of roughly 80,000 residents of the Mid-Ohio River Valley whose drinking water had been contaminated by perfluorooctanoic acid (PFOA, internally nicknamed "C8") discharged from DuPont's Washington Works plant near Parkersburg, West Virginia. After seven years of work and a $33 million budget paid by DuPont, the panel concluded that PFOA exposure was "probably linked" to six conditions: kidney cancer, testicular cancer, ulcerative colitis, thyroid disease, pregnancy-induced hypertension (including preeclampsia) and high cholesterol (hypercholesterolaemia). Those six findings, published 2011–2012 and underpinned by a cohort of about 69,000 people, are still the empirical backbone behind the US Environmental Protection Agency's April 2024 enforceable drinking-water standard for PFOA.

The Washington Works plant, Parkersburg, West Virginia

DuPont's Washington Works site sits on a bend of the Ohio River about six miles south of Parkersburg, West Virginia. The plant has manufactured fluoropolymer resins, principally polytetrafluoroethylene (PTFE, sold as Teflon), since the early 1950s. PTFE is made by emulsion polymerisation of tetrafluoroethylene monomer, and the process required a surfactant to keep the polymer particles suspended in water. For decades that surfactant was ammonium perfluorooctanoate, the ammonium salt of PFOA, known inside DuPont and 3M by the shorthand "C8," for its eight-carbon backbone.

One distinction matters here and is often blurred in popular accounts: 3M, not DuPont, was the original manufacturer of PFOA. 3M's Cottage Grove, Minnesota plant produced PFOA from the late 1940s onward using electrochemical fluorination and sold it to DuPont as a raw material. DuPont was the dominant US user of PFOA, deploying it at Washington Works (and at sites in New Jersey and the Netherlands) as a processing aid. 3M voluntarily phased out PFOA production in 2002; DuPont took over manufacture for its own use and ultimately spun the fluorochemicals business out as Chemours in 2015. For practical purposes, the Parkersburg contamination plume is a Washington Works problem, but the chemical itself has 3M's fingerprints on the upstream chemistry.

Washington Works released PFOA to air, to the Ohio River, and to unlined landfills and digestion ponds on and adjacent to the property. Over time the airborne and water releases migrated into the source aquifers and public water systems on both sides of the river, most notably Little Hocking Water Association (Ohio), the City of Belpre (Ohio), Tuppers Plains–Chester (Ohio), Lubeck Public Service District (West Virginia), Mason County PSD (West Virginia) and the private wells around them. The 2005 consent decree DuPont signed with the EPA over Washington Works covered, among other things, PFOA emissions limits, residential drinking-water sampling, and ultimately a $16.5 million civil penalty for failure to report human health and environmental information under the Toxic Substances Control Act.

The Tennant lawsuit and the Leach class action

Wilbur Tennant was a cattle farmer in Lubeck, West Virginia, whose land bordered a DuPont-owned property the company called "Dry Run Landfill." From the mid-1990s onward, Tennant's cattle began dying in strange numbers, with blackened teeth, organ tumours and behavioural abnormalities. He filmed the deaths, kept records, and eventually wrote to a Cincinnati corporate-defence lawyer named Robert Bilott, whose grandmother had known Tennant's family. Bilott, despite working at a firm that ordinarily represented chemical companies, took the case in 1999.

Through discovery on the Tennant matter, and a 1998 EPA submission Bilott obtained that mentioned a chemical called "PFOA", Bilott pieced together that the Tennant herd had been drinking water and grazing on land downstream of a creek into which DuPont had been depositing waste containing C8. In August 2001 Bilott filed Leach v. E.I. du Pont de Nemours and Co., No. 01-C-608, in the Circuit Court of Wood County, West Virginia, on behalf of a putative class of people who had been exposed to PFOA via their drinking water. The named plaintiff, Joseph Kiger, was a Lubeck PSD customer. The class was certified in 2002 and ultimately covered residents served by six contaminated water districts in the Mid-Ohio Valley, plus people on documented contaminated private wells in the same area, approximately 80,000 individuals.

(For clarity: Robert Bilott is the attorney; Wilbur and Earl Tennant were the cattle farmers whose dying herd surfaced the contamination. The name overlap with unrelated public figures is coincidental and frequently muddled in second-hand retellings.)

The 2005 settlement, what it actually agreed to

The Leach settlement, approved by the court in February 2005, did not assess damages and did not concede liability for any specific disease. Instead, it did something procedurally unusual: it converted what would otherwise have been a tort case into a long-running epidemiology project.

DuPont agreed to fund three things. First, an immediate cash payment of roughly $70 million to the class, intended in part to underwrite the science. Second, the installation and operation of granular activated carbon treatment on the six affected public water systems, so that future PFOA exposure would be reduced. Third, and this is the part that mattered for the rest of the world, DuPont agreed to fund an independent C8 Science Panel of three epidemiologists, chosen jointly by the parties, with a defined task: to evaluate whether there was a "probable link" between PFOA exposure and any human disease in the exposed population. Until that panel reported, the class members released DuPont from no personal-injury claims; once the panel reported, any disease the panel found a "probable link" for would entitle class members to pursue individual personal-injury actions, and DuPont further agreed to fund a medical monitoring programme for those conditions.

The three panellists were Dr Kyle Steenland (Emory University, formerly NIOSH), Dr Tony Fletcher (London School of Hygiene & Tropical Medicine) and Dr David Savitz (then at Mount Sinai, later Brown). Their work ran from 2005 to 2013 and ultimately cost DuPont approximately $33 million.

The C8 Health Project, about 69,000 residents

To give the panel data to analyse, the settlement created the C8 Health Project: a mass enrolment of class members who agreed to give a blood sample, complete a detailed health and exposure questionnaire, and allow follow-up. Enrolment ran from August 2005 through July 2006 in storefront clinics across the six water-district service areas. Roughly 69,000 people participated, a participation rate well over 80% of the eligible exposed population, and the largest community-scale environmental cohort ever assembled around a single chemical.

The design and methods were published by Frisbee and colleagues in Environmental Health Perspectives in 2009 ("The C8 Health Project: Design, Methods, and Participants," 117(12): 1873–1882). Each participant's serum PFOA concentration was measured; the geometric mean was about 32.9 ng/mL, more than seven times the contemporary US background (around 4 ng/mL in NHANES). The cohort also contained a long upper tail, thousands of participants above 100 ng/mL, and a meaningful number above 500 ng/mL, driven by length of residency in the most-contaminated water districts and by occupational exposure for the subset who had worked at Washington Works.

Because participants gave both a current blood measurement and a residential and water-source history going back years, the panel was able to reconstruct individual cumulative PFOA exposure with unusual precision, using pharmacokinetic models that account for PFOA's long human serum half-life (approximately 2 to 4 years for PFOA in adults, per ATSDR's 2021 Toxicological Profile for Perfluoroalkyls). That historical exposure reconstruction is what allowed the panel to conduct dose–response analyses, rather than just compare exposed-versus-unexposed groups. The methodological logic is laid out in Steenland, Fletcher and Savitz, "Epidemiologic Evidence on the Health Effects of Perfluorooctanoic Acid (PFOA)," Environmental Health Perspectives 118(8): 1100–1108 (2010), which the panel published as a peer-reviewed review of the broader PFOA evidence base in parallel with its own work.

The six "Probable Link" findings

Between October 2011 and October 2012 the C8 Science Panel issued a series of "Probable Link Evaluation" reports, each published on c8sciencepanel.org and each accompanied or followed by peer-reviewed papers in the epidemiology literature. Of the more than 50 health endpoints the panel evaluated, six met the panel's threshold for a "probable link" to PFOA exposure. The other endpoints, including breast cancer, prostate cancer, type 2 diabetes, asthma, autoimmune diseases other than ulcerative colitis, and a long list of birth outcomes other than PIH, were classified as "no probable link" on the evidence available to the panel at that time.

Year	Condition	Lead author / key paper	Summary effect
Oct 2011	Hypercholesterolaemia high cholesterol	Steenland et al., Am J Epidemiol 2009; panel report Oct 2011	Monotonic increase in total and LDL cholesterol across PFOA quartiles in adults and children.
Dec 2011	Pregnancy-induced hypertension including preeclampsia	Stein, Savitz et al., Am J Epidemiol 2009; panel report Dec 2011	Modest dose–response elevation in PIH/preeclampsia with higher modelled PFOA serum.
Apr 2012	Kidney cancer renal cell carcinoma	Barry, Winquist, Steenland, Environ Health Perspect 2013	Hazard ratio rising with cumulative PFOA serum; strongest of the cancer signals.
Apr 2012	Testicular cancer	Barry, Winquist, Steenland, Environ Health Perspect 2013	Elevated risk in the highest cumulative-exposure quartile; small absolute numbers.
Jul 2012	Thyroid disease	Winquist & Steenland, Environ Health Perspect 2014	Increased risk of thyroid disease (functional + structural) with higher exposure.
Oct 2012	Ulcerative colitis	Steenland et al., Environ Health Perspect 2013	Roughly doubled risk in the highest exposure quartile vs. lowest; clear dose–response.

These six findings are sometimes summarised loosely in the press as "PFOA causes six diseases." That is not what the panel concluded. The panel applied a specific, settlement-defined threshold, "probable link", and was careful to label its findings accordingly. The next section explains why the wording matters.

What "probable link" actually meant

The phrase "probable link" was not a casual journalistic gloss. It was a defined term in the Leach settlement agreement, negotiated between class counsel and DuPont and approved by the court. The settlement defined it roughly as follows: a probable link exists between PFOA exposure and a given disease where, based on the weight of the available scientific evidence, it is more likely than not that there is a link, that is, that a connection is more likely than no connection, among class members. The threshold is therefore civil preponderance applied to causation in epidemiology, rather than the higher bars used in regulatory risk assessment ("sufficient evidence of carcinogenicity") or in clinical decision-making.

That framing has two consequences worth being precise about. First, a "probable link" finding under Leach is not "proof" in any colloquial sense, and the C8 Science Panel itself was scrupulous about not claiming it was. It means the panel judged a causal link more likely than not on the evidence then available, no more, no less. Second, because the threshold was contractually defined to govern access to a personal-injury cause of action, a probable-link finding had a specific legal consequence inside the Mid-Ohio Valley class: it opened the door for class members with that disease to bring individual claims against DuPont. Hundreds did. The Leach probable-link mechanism is therefore best understood as a hybrid creature: a piece of evidence-weighing dressed in contract.

For wider regulatory and public-health bodies, the C8 panel's reports were one input among several, not a verdict. The International Agency for Research on Cancer reviewed PFOA again in 2023 and, in its forthcoming IARC Monograph Vol 135 (2023), reclassified PFOA from Group 2B ("possibly carcinogenic") to Group 1 ("carcinogenic to humans"), citing renal cell carcinoma as the cancer with sufficient evidence in humans, drawing heavily on the post-C8 epidemiology including subsequent work by Shearer and colleagues at the US National Cancer Institute. ATSDR's 2021 Toxicological Profile, similarly, treats the C8 findings as a major but not solitary line of evidence.

How C8 findings shaped the 2024 EPA MCL

When the EPA finalised the first federally enforceable Maximum Contaminant Levels for six PFAS in April 2024, published at 89 Fed. Reg. 32532 (April 26, 2024), the agency set PFOA at 4.0 ng/L, with a non-enforceable Maximum Contaminant Level Goal (MCLG) of zero on the basis that PFOA is a likely human carcinogen with no safe threshold. The agency's quantitative case for that MCLG, and the lower-bound dose–response used to derive the reference dose for non-cancer effects, draws extensively on the C8-era literature. In particular, the rule's preamble and supporting Health Effects Support Document cite the C8 Science Panel findings on kidney cancer and the related epidemiology on serum cholesterol and thyroid disease as part of the weight-of-evidence for treating PFOA as a chemical for which exposure should be minimised.

The connection is most direct on kidney cancer. The C8 panel's 2012 probable-link finding for renal cell carcinoma, followed by subsequent independent cohort and case–control work, particularly Shearer et al. (2021) using the PLCO cancer screening trial, gave EPA the human evidence base needed to set a cancer-based MCLG of zero for PFOA. Without the unusually high-exposure Mid-Ohio Valley cohort, the human dose–response for renal cell carcinoma at environmentally relevant exposures would be much harder to pin down.

For context on how the 2024 rule actually operates in your drinking water system, see our explainer at EPA's April 2024 PFAS MCL. The underlying toxicology of the long-chain perfluoroalkyls discussed here, PFOA in particular, is covered in PFOA & PFOS toxicology. If you are pregnant, the lower tolerable intakes derived in part from C8 cohort work are summarised at PFAS, pregnancy and tolerable intake. To see which US communities live downstream of historic PFAS point sources comparable to Washington Works, see our PFAS groundwater hotspot map.

Sources

Last reviewed 30 June 2026. Educational only, this is not medical or legal advice. How Safe Is My Water is independent and not affiliated with the US EPA, ATSDR, NSF International, DuPont, Chemours, 3M, or any party to the Leach litigation.