In the almost 20 years since water pollution with toxic fluorinated chemicals, or PFAS, erupted as a public health issue, research has found impacts from exposure to ever-lower levels. Yet there are still no national, legally enforceable drinking water standards for any of the hundreds of PFAS compounds currently in use.

Drawing on the best available science and emerging evidence of harm from the entire class of these chemicals,[*]EWG is proposing drinking water and cleanup standards for all PFAS chemicals as a group. To fully protect the health of children and other especially vulnerable populations, our proposed standards are 70 times lower than the Environmental Protection Agency’s drinking water advisory levels for the two most notorious chemicals in the class, PFOA and PFOS.

Americans are exposed to PFAS in many ways: not only drinking water but also food and food packaging, a wide array of consumer products, household dust and even airborne PFAS fumes. A number of studies predict that food is a major source of exposure to PFAS, but significant uncertainty remains about exactly how people are being exposed. Ultimately, drinking water is a direct source of exposure that can be reduced through regulation, and a standard for drinking water should consider ongoing exposures from food and other sources.

According to the Centers for Disease Control and Prevention’s measurements of PFAS in our bodies, the average American’s combined exposure is equal to drinking water with 14 parts per trillion, or ppt, of PFOA and 36 ppt of PFOS, daily for the past few years. Those exposure levels are 14 times higher and 36 times higher, respectively, than EWG’s recommended standard for all PFAS.

PFOA, formerly used to make DuPont’s Teflon, and PFOS, formerly in 3M’s Scotchgard and firefighting foam, have been phased out in the U.S. under pressure from the EPA, although they remain widespread in drinking water. They have been linked to cancer, birth defects, thyroid disease, weakened childhood immunity and other health problems.

In 2005 and 2009, laboratory tests commissioned by Environmental Working Group,
Commonweal and Rachel’s Network found PFOA, PFOS and other PFAS chemicals in umbilical cord samples from 20 babies born in the U.S. These tests exposed the shocking truth that American babies are born pre-polluted with PFAS and other toxic chemicals, which can pass from mothers to fetuses through the umbilical cord.

In 2016, the EPA issued non-binding drinking water advisory levels for PFOA, PFOS or their combined level, of 70 ppt. The EPA has identified more than 600 chemically similar PFAS compounds in active commercial use but has set no legal limits or health advisories for those chemicals in water, air or consumer products.

The pervasiveness of these chemicals makes it imperative that states not wait on the federal government’s toothless “action plan” for regulating PFAS. To spur more state action and inspire federal leadership, EWG scientists developed health guidelines to apply to the entire group of PFAS chemicals in drinking water; they can also be used as cleanup standards for PFAS in groundwater and at other contaminated sites. The cleanup standard, consistent with the policies of the federal Superfund law, is based on the potential for PFAS chemicals to migrate through groundwater and contaminate drinking water.

EWG’s Health-Based Standards for PFAS

  • PFAS in drinking water: 1 ppt as the total concentration for the sum of all PFAS.
  • PFAS in groundwater and cleanup of contaminated sites: 1 ppt as the total concentration for the sum of all PFAS.

In 2013, Dr. Phillipe Grandjean and Esben Budtz-Jórgensen, Ph.D., published a study showing decreased vaccine response in children that correlated with their body burden levels of PFOA and PFOS. The authors used these epidemiological data to recommend a limit of 1 ppt in water for PFOA or PFOS. A peer reviewer for the study noted that if other sources of PFAS exposure were considered, this would point to a recommended safe limit even lower than 1 ppt.

EWG used these findings, as well as studies showing effects on mammary gland development in rodents, as the basis for our human health benchmark of 1 ppt for PFOA and PFOS in drinking water. We extended that 1 ppt standard to the other PFAS chemicals detected in public water systems tested under the EPA’s Unregulated Contaminant Monitoring Rule 3 program, including PFHpA, PFHxS, PFNA and PFBS. At our proposed limit of 1 ppt, drinking water would be expected to contribute up to 0.1 ppb PFOA or 0.2 ppb PFOS.

As we show in the table below, a drinking water concentration at or near 1 ppt is also necessary for full protection from increases in cholesterol and liver and testicular cancers. The significant and mounting body of research on the shared health harms and toxicity concerns for PFAS chemicals other than PFOA and PFOS – such as GenX, a replacement for PFOA the EPA says is nearly as toxic as its predecessor – supports the application of 1 ppt as a health-based standard for the total concentration of the sum of all PFAS in water.

The analytical methods for PFAS chemicals in water are advancing rapidly.

A few years ago, 1 ppt was at or below the detection limit commonly reported, but as analytical techniques advance, detection methods for lower concentrations are becoming available.

Chemical Test subjects Study Uncertainty factor Recommended target exposure limit Notes
PFOA Rodents EPA and NIEHS study of delayed mammary gland developmentin mice. 30-fold uncertainty factor applied for the derivation of the Reference Dose, or RfD. 0.8 ppt Reference Dose published by New Jersey. According to the state’s assessment, the target human serum level is below current exposure levels, and any additional exposure would be a concern.
PFOA Rodents 3M study of testicular cancer in rats. Cancer potency factor based on a lower one-sided confidence limit on the benchmark dose (BMDL) for 5% tumor incidence. 14 ppt, corresponding to 1-in-1-million cancer risk level Calculated by New Jersey
PFOA Humans Decreased vaccine antibodyresponse study of 431 children. 10-fold uncertainty factor applied to the BMDL. 1 ppt if water only source of exposure BDML represents the concentration that elicits a response in 5% of the test subjects.
PFOA Humans Increase in cholesterol levels in 47,000 people in two different studies, one in West Virginiaand one in Denmark. 0 uncertainty factor applied to the BDML. 3 ppt Value calculated from EFSA review using standard drinking water assumptions.
PFOS Rodents 3M study of liver tumors in female rats. Cancer potency factor from linear slope of exposure groups. 3.3 ppt corresponding to one-in-one-million cancer risk level Calculated by New Jersey.
PFOS Rodents Immune suppressionfollowing 60 days of exposure in male rats. 30-fold uncertainty factor. 1.313 ppt Calculated by New Jersey. If New Jersey had determined that the cancer risk data was uninformative or inadequate, as EPA determined, it would have applied an additional uncertainty factor of 10 to the immune risk calculation, resulting in a non-cancer target value of 1.3 ppt.
PFOS Humans Decreased vaccine antibodyresponse in 431 children. 10-fold uncertainty factor applied to the BMDL, 1 ppt if water only source of exposure BDML represents the concentration that elicits a response in 5% of the test subjects.
PFOS Humans Increase in cholesterol levels in 49,000 people in three different studies, one in West Virginia, one across the whole U.S., and one in Denmark. 0 uncertainty factor applied to the BDML. 6.5 ppt Value of 6.5 calculated from EFSAfindings using standard drinking water assumptions.

In the Centers for Disease Control’s 2015-2016 national biomonitoring results, the geometric mean concentration of PFOA in Americans’ blood serum was 1.56 parts per billion, or ppb, and for PFOS was 4.72 ppb. It is unlikely that drinking water is the only source of exposure to PFAS but if it were, judging from blood levels, the average American is currently drinking approximately two liters a day of water with 14 ppt PFOA and 36 ppt PFOS.[2] At our proposed limit of 1 ppt, drinking water would be expected to contribute up to 0.1 ppb PFOA or 0.2 ppb PFOS.

Researchers at the Natural Resources Defense Council, a nonprofit organization, recently proposed a maximum drinking water limit of 2 ppt for the combined concentrations of PFOA, PFOS, PFNA, PFHxS, and 5 ppt for GenX. The 2 ppt concentration limit is based on the most common detection limits reported by commercial laboratories that test water for PFAS.

NRDC’s report notes that carbon filtration can remove these chemicals to levels lower than the detection limit of 2 ppt. NRDC also proposed a Maximum Contaminant Level Goal of zero for all PFAS, a number based on the potential for PFOA to cause cancer. NRDC used the studies reviewed by scientists for the state of  New Jersey, the federal Agency for Toxic Substances and Disease Registry and the EPA to calculate safe drinking water exposure levels of 0.01 ppt for PFOA, 0.002 ppt for PFOS, 0.3 ppt for PFNA, 2 ppt for PFHxS and 0.2 ppt for GenX.

PFAS assessments published by New JerseyATSDR, the EPAthe European Food Safety Authority, the German Human Biomonitoring Commission and other peer-reviewed studies show strong scientific support for a drinking water standard of 1 ppt for PFOA or PFOS. Research on the health effects of exposure to PFAS – in particular to the harms caused by low-dose exposure – is ongoing, consistently pointing to previously unanticipated health harms.

Some of the health endpoints most sensitive to PFOA and PFOS exposure are cancer and impacts to the liver, immune system, endocrine sytem and impacts on development and reproduction. The available studies of other PFAS chemicals, including PFNA, PFHxS, PFDeA, PFDoA, PFUA, PFBA, PDBA, GenX and PFHxA, indicate they all affect one if not more of these same health endpoints. Because of the lack of test data, it is impossible to determine conclusively the most sensitive health endpoints for each or a safe exposure level different from PFOA or PFOS.

For these reasons, a strong standard for the entire class of PFAS chemicals is essential for public health.

Why PFAS chemicals must be regulated as a class

The EPA has already set a precedent for regulation of chemicals as groups and categories, and PFAS strongly fits this precedent. EPA’s 2015 Significant New Use Rule for PFOA- and PFOA-related chemicals provided a definition for a category of a subgroup of the so-called long-chain PFAS chemicals.[3] In light of its review of toxicity tests of replacement chemicals, EWG believes this category should be expanded to include all PFAS substances, both for assessment and regulatory purposes.

PFAS chemicals never occur alone but rather in complex mixtures within products, the environment and people. The EPA recently tested 25 public water systems and found PFAS in every one, with most having 10 or more of the 17 PFAS they tested for.

The PFAS chemicals that have been extensively studied are harmful to nearly every human organ, physiological system and life stage. According to EPA’s 2016 drinking water advisory fact sheet, exposure to PFOA and PFOS can cause “developmental impacts during pregnancy and to infants, cancer, liver damage, harm to the immune system, damage to the thyroid and other effects including changes in cholesterol.” In a draft assessment of GenX, the short-chain replacement chemical for PFOA, EPA used a study that showed the chemical was even more potent in toxicity tests. The vast majority of PFAS in commerce lack even basic safety testing data.

The lack of adequate safety data makes it impossible to exonerate a particular PFAS currently in commerce or in tests and research. Even for PFBS, one of the most thoroughly studied replacements, the EPA found a lack of tests for developmental and immunological toxicity; a lack of a chronic mouse study and full two-generation reproductive toxicity study; and a lack of human epidemiological studies. These are the kinds of studies that could determine whether these chemicals are hazardous in the same ways as PFOA and PFOS.

The sheer number of PFAS chemicals makes a full assessment of each individual substance neither practical nor possible. There are thousands of known PFAS chemicals and more than 600 that EPA has classified as being active in commercial use.

A chemical-by-chemical regulatory approach is not currently achievable without an overhaul of U.S. chemical regulation. In the absence of adequate safety data, all members of the group should be considered equally toxic until proven otherwise. People are regularly exposed to multiple PFAS chemicals, and at the current pace, it could take decades or longer to review and regulate even the most common PFAS.

Drinking water standards

The safe level of PFAS in drinking water, particularly PFOA, has been the subject of regulatory scrutiny and active research for the past 20 years. The more we learn about PFOA, PFOS and other PFAS chemicals, the more concerns mount about their health effects. When we consider studies of human exposure to PFAS, the safe exposure level decreases to 1 ppt part in water or lower. In their proposed legal limit for PFOA, New Jersey state scientists said, “New Jersey scientists in their proposed drinking water limit for PFOA expressed this concern that any level of exposure from water could harm health.” They continue:

 . . . PFOA is associated with several human health effects, some with evidence supporting multiple criteria for causality, within the general population exposure range even without additional exposure from drinking water. Furthermore, the Reference Dose (RfD) for mammary gland effects in mice is below the average exposure level in the general population, and other toxicological effects occurred at similarly low doses in animal studies. Therefore, any additional exposure from drinking water may potentially pose some risk of health effects. For this reason, it cannot be concluded that lifetime exposure to a certain drinking water concentration, no matter how low, is protective of sensitive subpopulations with a margin of exposure. [Emphasis added.]

EWG’s proposed health-based drinking water standard of 1 ppt for the sum of all PFAS chemicals would minimize the potential for additional harm from drinking water.

Cleanup standards based on drinking water

Federal cleanup regulations for groundwater and other contaminated sites require reducing pollutants to a level equivalent to what is safe in drinking water. For Superfund sites, subject to the federal Comprehensive Environmental Response, Compensation, and Liability Act, that starts with a legal limit, or Maximum Contaminant Level, set by the EPA.  The use of an equivalent value may at first seem unnecessarily strict because no one would knowingly drink the water from a contaminated site, but it is necessary to ensure that the contamination does not migrate off-site and affect drinking water, agriculture or wildlife. In the context of potential PFAS site cleanups, EPA stated:

In consideration of the Agency’s goals of pollution prevention, ambient waters should not be contaminated to a level where the burden of achieving health objectives is shifted away from those responsible for pollutant discharges and placed on downstream users to bear the costs of upgraded or supplemental water treatment.

To ensure that the costs of cleaning up contamination are borne by the responsible parties, we have used this same approach in setting EWG’s proposed cleanup as equal to our drinking water standard.

States and government agencies moving to address the PFAS crisis

The EPA has indicated it is taking enforcement actions when groundwater sources that directly affect drinking water are contaminated at levels over the advisory level of 70 ppt. North Carolina, Texas, Alaska, Maine and Minnesota all have groundwater cleanup standards for PFOA or PFOS that are higher than 70 ppt, but they were set almost entirely before EPA published its health advisory in 2016.

Vermont has set an emergency groundwater standard of 20 ppt from the sum concentration of PFOA, PFOS, PFHxS and PFHpA.

Effective in March, New Jersey has set groundwater standards of 10 ppt for PFOA and 10 ppt for PFOS. These values are interim standards based on the proposed drinking water limits and, when finalized, will likely match the 14 ppt for PFOA and 13 ppt for PFOS already in effect. In March the state announced an initiative to acquire more information about past and ongoing PFAS production and release into the environment. New Jersey also demanded reimbursement for the $3 million in PFAS-related expenses already occurred and noted that the PFAS manufacturers would be responsible for future cleanup costs anticipated to run in the hundreds of millions.

The Pentagon, which has responsibility for at least 100 sites contaminated at levels exceeding the EPA’s lifetime health advisory for PFOA and PFOS, published its own calculations for PFAS levels in groundwater that can be used as drinking water. Shockingly, the Pentagon’s chosen level of 380 ppt for PFOA or PFOS is vastly in excess of the level that states and independent scientists consider safe, and more than five times greater than EPA’s health advisory. In a report to Congress, the Pentagon acknowledged that its proposed groundwater benchmark for PFAS “would create tremendous risk communication challenges . . . when explaining how the groundwater cleanup levels are protective of human health.”


From what we know of the latest science, EWG’s proposed standards for drinking water and remediation represent values that will fully protect health. These values do not include consideration of the limits in analytical techniques necessary to measure all PFAS, the likely considerable costs of cleaning up contamination or the political willpower to do so.

Calgon Carbon, a major supplier of carbon filtration systems, has tested filters and predicted they can reduce long- and most short-chain PFAS to below 1 ppt over two years of use. PFAS with carbon chains of 3 or shorter passed through the filters much more quickly. EWG has estimated that 110 million Americans could be drinking PFAS-contaminated water at a level of 2.5 ppt or greater, and it is possible that most water in the country is already contaminated above our guideline. Just because contamination is common doesn’t make it safe.

[*]We define PFAS as any manmade perfluoroalkyl or polyfluoroalkyl substance with at least one fully fluorinated carbon atom. PFAS concentration should be established by detection of total organic fluorine in water. This drinking water standard updates EWG’s previous PFAS standard, published in July 2017, of 1 ppt for each of six compounds (PFOA, PFOS, PFNA, PFHxS, PFHpA, PFBS).

[2] To convert between the drinking water concentration and a blood serum level, we used a bioconcentration factor of 114 for PFOA and 197 for PFOS as derived by EPA and published by the state of New Jersey.

[3] PFOA and PFOS are so-called long-chain chemicals because they contain eight or more carbon atoms. Their “short-chain” replacements, such as GenX, have six or fewer carbon atoms, which allows them to be eliminated more quickly from the body.


Written by Loknath Das