FDA finds surprisingly high levels of PFAS in certain foods – including chocolate cake

[Update: FDA has published a webpage on PFAS and released the data for the studies discussed in this blog].

Tom Neltner, J.D.Chemicals Policy Director and Maricel Maffini, Ph.D., Consultant

As reported by the Associated Press today, at a conference last week in Helsinki, Finland, the U.S. Food and Drug Administration (FDA) presented the results of three studies it conducted of 16 per- and poly-fluoroalkyl substances (PFAS) in various foods. A friend who attended the conference sent us photos of the poster. The results for samples of meat and chocolate cake purchased by the agency in October 2017 as part of its ongoing Total Diet Study (TDS) jumped out at us as surprisingly high and worth further investigation:

  • 17,640 parts per trillion (ppt) of perfluoro-n-pentanoic acid (PFPeA) in chocolate cake with icing. These levels suggest that the cake was contaminated from the intentional use of the chemical to greaseproof paper that contacted the cake rather than from an environmental source. We cannot find any evidence that FDA ever reviewed the safety of PFPeA as a food contact substance – meaning the manufacturer may have secretly designated it as Generally Recognized as Safe (GRAS). We also found little evidence – good or bad – of the health risks posed by this PFAS. We have reached out to FDA to learn more, but as of this blog posting the agency has not yet responded. This chemical was also found in chocolate milk at 154 ppt.
  • Nearly half (10 of 21) meat samples had quantifiable levels of perfluoroctanesulfonate (PFOS) with concentrations ranging from 134 ppt in a frankfurter to 865 ppt in tilapia. Unlike the chemical in chocolate cake, PFOS has been extensively studied because of widespread environmental contamination, especially around the facilities in Alabama and Minnesota where it was previously produced. It is associated with increased cholesterol, thyroid disease, testicular cancer, and decreased birth weight. While comparisons are complicated, the PFOS levels found in some of these meats were far greater than the 70 ppt health advisory set by the Environmental Protection Agency (EPA) for drinking water in May 2016. Two years later, the Centers for Disease Control and Prevention (CDC) proposed limits that are almost 7 times more protective than EPA’s, partly because more recent studies indicate the chemical may undermine the effectiveness of vaccines. Production of PFOS in the United States reportedly ended in 2002, though it is still made overseas and may have been imported paper. In 2016, FDA removed its approval to greaseproof paper with PFOS.

FDA’s poster also showed testing results from food produced around two PFAS contaminated areas. FDA found most of the 16 PFAS at varying levels measured in produce sold in farmer’s markets downstream of a PFAS production facility in the Eastern U.S. – presumably Chemours’ plant in North Carolina. The highest produce sample had 1,200 ppt and was purchased within 10 miles downstream of the production plant and short-chain PFAS were prevalent.

The other contaminated area was a dairy farm near an air force base in New Mexico. FDA found many of the 16 PFAS in the water and silage used to feed the cows but PFOS was the most prevalent among a few PFAS measured in the milk with levels higher than 5000 ppt. The agency also detected several PFAS in cheese produced by the farm in lower amounts than the milk. Many of the PFAS are likely from aqueous film forming foam (AFFF) used to fight fire and conduct firefighting training at the Air Force base.

What about that chocolate cake with icing?

The 17,640 ppt result was from three samples of chocolate cake mixed together that FDA staff bought from stores in three randomly selected cities in October 2017 in the mid-Atlantic. FDA has not yet identified the cities, but defines the mid-Atlantic as Tennessee, Kentucky, Ohio, West Virginia, Virginia, Maryland, Delaware, and the District of Columbia. Since it was the only baked good with a quantifiable level of PFAS, it may not be representative of other cakes or baked goods.

If the PFPeA contamination was from only one of the three cakes mixed together, then the levels could be almost 53,000 ppt. See Figure 1. In addition, a store may use the PFPeA-greaseproofed paper to prepare most of their baked goods. Given that shoppers are typically store- and brand-loyal, their cumulative exposure may be significant.

In addition, little is known about the PFPeA and its health effects. With five fully fluorinated carbons in a chain, it is considered a short-chain PFAS. But it is similar to the eight-carbon variant called PFOA that has been well-studied and shown to pose significant health effects similar to PFOS.

Why do we think PFPeA in the cake was from an intentional use not approved by FDA?

Chocolate cake with icing is a mixture of many ingredients – outside of flour and oil, no ingredient dominates. So, if the PFPeA comes from only one of ingredient, the levels would have to be extremely high.  For that reason, we think it is much more likely that the chemical got into the cake from grease-proofed paper used by the bakery.

From our review of more than 20 notifications submitted by PFAS-manufacturers to FDA describing the use of their PFAS-variant to greaseproof paper, it would not be surprising to find these levels in food. So we think these levels are likely from contact with paper treated with PFPeA.

The disturbing part is that none of the 14 different PFASs currently allowed by FDA to greaseproof paper products are PFPeA. They do not appear to be made from the chemical or expected to break down into it either. Therefore, if the use was intentional rather than simply environmental contamination, then the PFAS manufacturer and company making the treated paper must have self-certified it as GRAS. This is a loophole in the 1958 law establishing the food additive regulatory program that FDA has allowed industry to exploit. EDF and other public health advocates think that this practice violates the Federal Food Drug and Cosmetic Act and are in the middle of a legal challenge to the agency’s rule. The lawsuit was filed two years ago. In April 2019, we asked the court to declare a summary judgment in our favor.

Regardless of the legal challenge, FDA should be investigating the source of the PFPeA. If it was indeed self-certified by food companies as GRAS without the agencies’ or the public’s knowledge, then the agency needs to determine whether it is safe. As with food additives, there must be “a reasonable certainty in the minds of competent scientists that the substance is not harmful under the conditions of its intended use,” and that evaluation must consider “the cumulative effect of the substance in the diet, taking into account any chemically- or pharmacologically-related substances in such diet.” 21 CFR 170.3(i). In addition, the substance’s safety must be generally recognized by experts. Given how little is known about PFPeA and the potential cumulative effect of all PFAS in the diet, we think it is unlikely that the use is safe. We recommend that FDA determine the source of the chemical and assess whether its use is allowed.

What about PFOS in meat?

FDA found PFOS in ten of 21 samples of poultry (chicken thighs and ground turkey), red meat (beef steak, lamb chops, and frankfurters), and seafood (tilapia, cod, salmon, shrimp, and catfish). Pork is the only major type of meat that is typically tested but is not listed in the results which means that either the levels were below the limit of quantification (LOQ) or it was not tested.

As we described with cake, each result is from a composite of three product samples bought at stores in three different cities. So the individual levels could be three times higher if only one product was contaminated with PFOS. See Figure 1.

Given the extent of environmental contamination from PFOS, especially of water, we assume the PFOS came from those sources. However, it is also possible that FDA collected many of those samples from a single store that wrapped the meats in PFOS-coated or treated butcher paper. If so, this would be illegal since, as noted above, FDA removed the approval of those uses of PFOS in 2016.

As with the cake, we recommend that FDA thoroughly investigate the source of the PFOS in these meats starting with the grocery stores and moving up the supply chains.

Why only 16 types of PFAS?

FDA’s methods did not explain why it selected these 16 PFAS. They are the ones commonly tested by commercial laboratories in drinking water, although some labs test for many more. Still, it is a useful spectrum of environmental contaminants.

However, we find it worrisome that FDA did not also test for the many types of PFAS that it currently allows to be used to greaseproof paper. We strongly recommend that in future testing, the agency will expand the scope of the chemicals tested.

Are parts per trillion levels really important?

Almost 20 years ago, we learned that some forms of PFAS, such as PFOS, have been found in the blood of most humans. A few years ago, we learned certain PFAS were all too commonly found in drinking water. Now we are increasingly finding them in food. While the levels seem extraordinarily small, the health risks are not. We know that these chemicals persist in the environment and many types bioaccumulate in the human body – the half-life of PFOS in the body is 5.4 years. Ongoing exposures to many types of PFAS that bioaccumulate means that even these low levels matter because our bodies will never get rid of the chemicals – increasing the odds that our children are born pre-polluted.

What is next?

These studies, especially the Total Diet Study, are designed to explore potentially significant exposures to chemicals in food. They are not intended to comprehensively assess the marketplace. So we applaud FDA for getting the investigation started and recommend that, given these results, it reduce PFAS contamination in the food supply. Specifically, the agency should test food more extensively, including for the forms of PFAS the agency currently allows to be used; identify the sources for high levels of PFAS by investigating the supply chain; and ensure that no PFAS is used because of the GRAS loophole.

This entry was posted in FDA, Food, Public health and tagged , , , , . Bookmark the permalink. Both comments and trackbacks are currently closed.


  1. Wayne
    Posted June 3, 2019 at 1:16 pm | Permalink

    “Let them eat chocolate cake” – Melania Trump

  2. Prof. Graham F. Peas
    Posted June 14, 2019 at 2:50 pm | Permalink

    The answer to the chocolate cake source question is on the same poster. C5 acid is relatively rare in the environment and is not used in packaging in general…too volatile. Usually diPaps are used on packaging or other PFAS. The dairy that was condemned in New Mexico has the answer though. The dairy was next to a USAF base with AFFF-impacted water. The cows were drinking high levels of PFOA that was passed through to their milk. They were also eating the silage that was grown on the farm…which was very high in C5!!. Apparently wheat preferentially bio concentrates C5. This has been seen before in leafy vegetables and fruits in Colorado and Italy…and it can get into grains. It could come from AFFF laced irrigation but since there isn’t much c5 in AFFF, then it likely comes from C5 in biosolids applied to the fields. Since this cake was from the east coast, it means that there is presumably another wheat field out east that has biosolids applications or irrigation with PFAS contamination…and that wheat ended up as flour for humans.

    Note that the take away message from the FDA poster was that it seems most of the food supply is safe….but there were a few “hot spots” observed. This would be very consistent with the idea that wherever there is a contamination problem in water or biosolids there is a food security issue, and the background level of PFAS are low in general…and not coming from cake plants but from our landfills that are leaching into our foods and water.

    The Australians found egg yolks at a fire station I South Australia was giving the highest PFAS blood sera values…more than the occupational exposure to AFFF for those firefighters!

    Look at the tabular data carefully from the dairy farm and you can confirm in the published literature elsewhere…this cake mix is likely contaminated wheat.