FDA must abandon its flawed assumptions when reviewing safety of approved PFAS uses in food

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

[pullquote]

All the PFAS uses allowed by FDA that we reviewed had estimated exposures exceeding the most protective minimal risk level for PFOS proposed by the Centers for Disease Control and Prevention (CDC).

[/pullquote]In its June 2019 release of a webpage dedicated to per- and poly-fluoroalkyl substances (PFAS) in food, FDA stated that it is “reviewing the limited authorized uses of PFAS in food contact applications.” As we mentioned in a previous blog, we were pleased to see FDA’s public position on PFAS but we highlighted three major concerns that could impact the ongoing safety review and questioned the conclusion that all is fine. In this blog, we discuss the implications of FDA’s statements on its review of 62 authorized PFAS uses in contact with food and make recommendations to the agency as it proceeds with this promising effort.

FDA authorized uses are mixtures of PFAS

According to FDA’s Food Contact Substance Notification (FCN) Inventory,[1] in the past 20 years, the agency has allowed 69 PFAS notices submitted by 19 different companies to become “effective” – which essentially is an approval by FDA. Of these, three companies have voluntarily ceased the use of seven FCNs because they involved long-chain PFAS[2] – thus we are considering these seven inactive. This leaves 62 active FCNs[3] for mixtures of PFAS of mostly unknown composition, various lengths of fully-fluorinated carbon chains, and concentrations of PFAS.

The vast majority of the 62 active FCNs include polymers that the agency has determined have little ability to be absorbed by the body and, therefore, pose little risk to human health. However, as we noted in a previous blog, impurities – such as monomers, oligomers, and degradation byproducts –present in a polymer product can be absorbed by the body and could affect various organs.

FDA should be using reference doses for PFOA/PFOS to inform its safety assessment of other PFAS

In its webpage, FDA stated that “[f]or PFAS found in food, the FDA currently uses the U.S. Environmental Protection Agency’s [EPA] reference doses (RfD) for PFOA[4] and PFOS[5] of 0.02 µg/kg bw/day[6] as the most appropriate toxicity reference value (TRV).” The RfD is the lowest exposure determined to be safe and is commonly based on toxicology information for individual chemicals. We took FDA’s statement to mean that the agency’s safety assessment consisted of treating PFAS as a class and applying the RfD to each member of the PFAS family and, possibly, collectively for the combined exposure to all PFAS in the diet. .

After checking with FDA, we learned that the RfD would only apply to long-chain PFAS for which it tested in food, and did not explain what RfD it used for the short-chain PFAS. However, because the agency lacks an alternative RfD for short-chain chemicals, and these are the most common PFAS covered by the 62 active FCNs, we think the agency should use the most protective RfD – the one proposed by CDC as its minimal risk level (MRL) in its PFOS safety review, that is an order of magnitude lower than EPA’s.

With both EPA’s RfD and CDC’s proposed RfD in mind, we looked closely at FDA’s response to our 2017 FOIA for 31 FCNs that we considered significant at the time. We think this approach is appropriate because most of the FCNs covered a mixture of PFAS and the toxicology information was insufficient to identity “safe doses.”

We found:

• 15 FCNs where the exposure estimates for at least one PFAS covered by the notice exceeded the 2016 EPA RfD for PFOA and PFOS; and
• All FCNs contained unredacted estimated exposures for the listed PFAS that exceeded the CDC 2018 RfD^[7] for PFOS.

Our analysis does not include the combined exposure to all the PFAS covered by the FCNs. It also does not adjust for the existing contamination of the food supply, drinking water, and the human body. If these were considered – as FDA is required to do under the law – we would anticipate that all exposure to PFAS in all active FCNs would likely be above EPA’s RfD and therefore considered unsafe – meaning that there is no longer “a reasonable certainty in the minds of competent scientists that the substance is not harmful under the conditions of its intended use.”[8]

Recommendations for FDA as it reviews the existing authorizations

As FDA proceeds with its review of authorized PFAS in FCNs, we recommend that the agency:

• Notify PFAS manufacturers that FDA has determined that all 62 active PFAS FCNs are no longer effective. Based on our review of FDA’s responses to our 2017 FOIA for 31 FCNs, if the agency were to apply the most protective RfD to PFAS (i.e., CDC’s) as required by 21 CFR 170.18 and consider other sources of PFAS in the diet, none of the 31 FCNs should be allowed as food contact substances. We do not expect a different outcome for the remaining 31 FCNs if they used a similar approach that we saw in the FOIA responses. Therefore, consistent with 21 CFR § 170.105, FDA should notify companies with PFAS FCNs that it has determined their FCNs are no longer effective and give them an opportunity to make their case that the uses are safe. If a company fails to adequately respond in a timely manner, FDA should make its decision official with a Federal Register notice.
• Ensure that companies provide all relevant data as required by law. In a previous blog, we highlighted studies that Daikin improperly withheld from the agency. It is important that FDA have all relevant information so it can make an informed decision on the safety of a chemical or mixture of chemicals. In addition to toxicity, data on bioaccumulation in the body are critical for short-chain PFAS. We have previously reported on FDA’s own research showing potential bioaccumulation which increases health risks of short-chain PFAS. The issue is important because the agency’s method of assessing safety assumes that chemicals in the diet do not accumulate in the human body – which is not the case for many PFAS.
• Test foods for mixtures of PFAS in the active FCNs. In the three food studies the agency conducted, it tested for environmental contamination from 16 PFAS. However, when performing chemical safety assessments, FDA has the statutory obligation to consider the cumulative effects of chemically- and pharmacologically-related substances in the diet.[9] Therefore, it should also test for those PFAS it has allowed to be used in contact with food to ensure its collective exposure estimates are accurate and representative of the PFAS in the diet. To do this, the agency should utilize the most sensitive analytical methods – or develop them if there aren’t any available – that can measure PFAS at the parts per trillion levels in the diet.
• Require companies to provide comprehensive estimates of environmental releases of their products. Previously, we have highlighted that most FCNs provided little to no information on the estimated environmental releases from the manufacture, processing, use, and disposal of these PFAS. However, four FCNs – two from Chemours and two from Daikin – did provide estimated releases from a typical paper mill applying their PFAS mixture to paper. The estimates ranged from 95 to 225 pounds per day into the water and nine times more into biosolids – which the companies assumed would be burned for energy recovery. Given the extent of environmental contamination from PFAS, we encourage the agency to require the companies with authorized PFAS uses in food packaging to provide more rigorous and comprehensive estimates of environmental releases consistent with its obligations under the National Environmental Policy Act. This information should then be made publicly available without a FOIA.
• Demand that the companies provide samples of the products for each of the FCNs. The FCNs are not for single pure chemicals. Rather, each is for a mixture of PFAS. FDA scientists made that clear in a 2013 publication when they tested two PFAS food contact substances provided by Ciba Specialty Chemicals. They also showed that the same mixture of PFAS that was present in food packaging paper had migrated into food. The agency needs samples of the products it approved to better match the mixture of PFAS found in food to the FCNs. This would help identifying the source of contamination and avoid the situation such as the chocolate cake with icing, where the PFAS known as PFPeA was found in extremely high levels, but its source remains unknown. In addition, FDA should share the products’ samples with other federal and state agencies to allow for them to more accurately identify sources of environmental contamination.

Summary

As we said in the previous blog, it is a positive step that FDA made public its position on PFAS food contamination, released the three studies and committed to review the safety of intentional uses of PFAS authorized for use in food packaging. However, we are concerned that from our examination of PFAS included in 31 FCNs, FDA’s estimated exposures already exceed the most protective reference dose provided by the CDC for PFAS as a class. For the agency to be successful in protecting the public from ongoing exposures to these hazardous chemicals, it needs to begin the process of revoking the effectiveness of all PFAS FCNs, and it needs to act quickly if the companies do not provide compelling evidence that the uses are safe.

[1] As of July 31, 2019.

[2] Long-chain PFAS are those with eight or more fully –fluorinated carbons in a chain.

[3] Some of the FCNs address different or expanded uses of the same PFAS mixture.

[4] Perfluorooctanoic acid – a PFAS with seven fully-fluorinated carbons.

[5] Perfluoroctanesulfonate – a PFAS with eight fully-fluorinated carbons.

[6] µg/kg bw/day means micrograms of the chemicals per kilogram of body weight per day. FDA uses 60 kilograms or 132 pounds as representative weight for adults.

[7] CDC uses the term minimal risk level (MRL) and defines it as ‘An estimate of daily human exposure to a hazardous substance that is likely to be without an appreciable risk of adverse noncancer health effects over a specified route and duration of exposure.”

[8] See 21 USC § 348 and 21 CFR 170.3.

[9] 21 USC § 348