Round 1 responses from the panel were fairly mixed on a preference for a definition of PFAS. We are hoping to work towards consensus on this important issue. A table of example PFAS has been assembled and pertinent information gathered (see link below). Please refer to this when answering this question. Note that the OECD 2021 definition forces many compounds used in medicine and agriculture, like Prozac and fipronil to be considered PFAS. Given this information, which of the following definitions (listed in order of broadest definition to narrowest) is most applicable for defining PFAS that are of interest for drinking water exposures?

Answer Explanations

• USEPA CCL5/TSCA reporting: Per- and polyfluoroalkyl substances or PFAS, for the purpose of this part, means any chemical substance or mixture that structurally contains the unit R-(CF2)-C(F)(R′)R″. Both the CF2 and CF moieties are saturated carbons. None of the R groups (R, R′ or R″) can be hydrogen

  Expert 6

  I had previously preferred the Buck et al 2011 or OECD definitions, but the US EPA CCL5/TSCA definition now seems more likely to capture more of the PFAS of interest. It is also simpler.

• USEPA CCL5/TSCA reporting: Per- and polyfluoroalkyl substances or PFAS, for the purpose of this part, means any chemical substance or mixture that structurally contains the unit R-(CF2)-C(F)(R′)R″. Both the CF2 and CF moieties are saturated carbons. None of the R groups (R, R′ or R″) can be hydrogen

  Expert 3

  This reviewer is not a chemist. The advantage of using the US EPA TSCA definition is that presumably there will be reporting data that will be consistent with the aim of defining PFAS in drinking water.

• ITRC 2021: PFAS include only fluorinated aliphatic (carbon chain) substances. PFAS do not include fluorinated compounds that contain aromatic (carbon ring) features in their structures (for example, active pharmaceutical ingredients, crop protection agents, or chlorofluorocarbons [refrigerants])

  Expert 9

  The ITRC is an organization that reflects a US perspective and the interests of states within the US. Although its definition is more limiting than some of the others listed, it provides a balanced definition that is based on technical knowledge and regulatory decision making and still is protective of human and environmental health.

• Expert 11

  Three points. First, the choice of the definition of PFAS within the context of drinking water standards is not decided based only on technical findings; as a result, each of the options could be acceptable under different policies. Second, whatever is selected as the basis for a definition, the requirement that the compound be sufficiently mobile in surface and ground water that they will reach drinking water systems will eliminate some PFAS from consideration (perfluorinated polymers, polymers with perfluorinated side chains, and oligomers). Third, standard setting for drinking water required the consideration of the risks of any compound including pesticides and drugs. As a result, there may be no harm in including chemicals like Prozac and Fipronil in the PFAS definition when it is applied to drinking water standards.

• OECD 2021: fluorinated substances that contain at least one fully fluorinated methyl or methylene carbon atom (without any H/Cl/Br/I atom attached to it), i.e. with a few noted exceptions, any chemical with at least a perfluorinated methyl group (–CF3) or a perfluorinated methylene group (–CF2–) is a PFAS

  Expert 4

  It’s not clear to me why you would not include Prozac etc. in the initial definition. For some problems (e.g. examining ecological effects), it may be appropriate to assess these substances, regardless of their intended users human drugs. They are easy to remove during problem formulation if not relevant. The definition should be broader, not narrower.

• OECD 2021: fluorinated substances that contain at least one fully fluorinated methyl or methylene carbon atom (without any H/Cl/Br/I atom attached to it), i.e. with a few noted exceptions, any chemical with at least a perfluorinated methyl group (–CF3) or a perfluorinated methylene group (–CF2–) is a PFAS

  Expert 1

  There are pros and cons for each of these definitions. A con for the ITRC definition is that it excludes chemicals with an aromatic ring (which would exclude chemicals like Perfluoroethylcyclohexanesulfonate (PFECHS)). The Buck definition would exclude fluorotelomers because some C atoms are not bound to F (e.g., 8:2 Fluorotelomer phosphate monoester (8:2 MonoPAP)). While it is true that the OECD definition could include other chemicals such as prozac (which probably do not share the same mechanism of action as legacy PFAS), I wonder if they co-occur in drinking water in a way that would significantly impact risk assessment. I would therefore stick to the broad definition, assuming unrelated organofluorines (drugs/pesticides) do not contribute much to drinking water total organofluorine load.

• OECD 2021: fluorinated substances that contain at least one fully fluorinated methyl or methylene carbon atom (without any H/Cl/Br/I atom attached to it), i.e. with a few noted exceptions, any chemical with at least a perfluorinated methyl group (–CF3) or a perfluorinated methylene group (–CF2–) is a PFAS

  Expert 10

  It's unclear if this question refers to grouping PFAS for drinking water thresholds. As it's not stated what the purpose of the grouping is, I have to assume that this refers to the broader question of the definition of PFAS. In my view, the definition of PFAS should only cover those synthetic fluorinated organic substances that are extremely environmentally persistent (i.e. following the EU REACH definition of persistence, i.e. those substances that are persistent themselves or have persistent transformation products). The OECD definition is not ideal, in this regard, because some substances that meet this structural definition of PFAS are mineralizable (Cousins et al., 2020; Singh and Papanastasiou, 2021).
  However, I would prefer to have a broad "chemistry" definition and then exclude substances for regulation that are proven to be mineralizable, or that have essential functions (e.g. in medicine and agriculture). The OECD report that published the new OECD PFAS definition also states this; i.e. that regulatory definitions (subgroupings) of PFAS will need to be devised for individual regulatory purposes.

  If we take, for example, the drinking water thresholds, using the OECD PFAS definition would be nonsensical given the lack of exposure and toxicity data for most PFAS, and also the knowledge that many PFAS are highly volatile or insoluble and would never be present in water.

  Cousins, I.T.; DeWitt, J.C.; Glüge, J.; Goldenman, G.; Herzke, D.; Lohmann, R.; Ng, C.A.; Scheringer, M.; Wang, Z. (2020) The High Persistence of PFAS is Sufficient for their Management as a Chemical Class. Environ Sci-Process Impacts, 22, 2307-2312.

  Singh, R.R.; Papanastasiou, D.K. (2021) Comment on “Scientific Basis for Managing PFAS as a Chemical Class. Environ. Sci. Technol. Lett., 8, 2, 192–194

• Buck et al. 2011: highly fluorinated aliphatic substances that contain 1 or more C atoms on which all the H substituents … have been replaced by F atoms, in such a manner that they contain the perfluoroalkyl moiety CnF2n+1–

  Expert 2

  Buck et al. provide adequate justification for listing of a range of PFAS in that category. Only progress that can made is that developing subgroups of PFAS that are in commerce. As mentioned above such subgroups of PFAS should be based on chemical structure/functional groups or mode/mechanism of toxicity. Many pharmaceuticals and agricultural chemicals with one or two fluorine attached to an aliphatic or aromatic moiety cannot be included with this category of PFAS and that will only hinder making progress to develop risk assessment. An aliphatic chain with a polar functional group would be a typical PFAS; if some other types are included adequate justification needs to be provided.

• OECD 2021: fluorinated substances that contain at least one fully fluorinated methyl or methylene carbon atom (without any H/Cl/Br/I atom attached to it), i.e. with a few noted exceptions, any chemical with at least a perfluorinated methyl group (–CF3) or a perfluorinated methylene group (–CF2–) is a PFAS

  Expert 8

  These definitions simply define the very broad class of chemicals considered to be PFAS. None actually define the specific PFAS to be considered for the purposes of drinking water criteria. So whether a definition “forces” compounds like Prozac is not important in choosing a specific definition for the class. I would choose the OECD 2021 definition as it has been vetted internationally and more importantly the OECD document provides excellent examples of the proper use and the misuse of the term PFAS and the subgroups of PFAS.

• OECD 2021: fluorinated substances that contain at least one fully fluorinated methyl or methylene carbon atom (without any H/Cl/Br/I atom attached to it), i.e. with a few noted exceptions, any chemical with at least a perfluorinated methyl group (–CF3) or a perfluorinated methylene group (–CF2–) is a PFAS

  Expert 5

  The assignment of chemicals to PFAS as a whole (the "big picture use of the acronym") or to subgroups of PFAS has different impact when considering (1) regulation and (2) risk assessment. When focusing on risk assessment via cumulative risk methods, the inclusion of chemicals in the "big group" or a specific subgroup impacts risk. Keeping in mind that risk=+ hazard X exposure, then exposure is critical. Regardless whether a given chemical (chemical structure) is included in the group, if there is no exposure, then there is no risk, and grouping is moot. SO, it will be important to remain mindful of the reason (problem formulation) why we are considering grouping. This (my) thoughts come from a center on drinking water risk assessment, and an expectation, naïve or not, that fipronil and prozac may not be anticipated co-contaminants. Additional consideration, if they are expected to be co-contaminants, might be given to the expected toxicities associated with fipronil and prozac - relative to those expected from "traditional" PFAS chemicals. But it seems, from a risk assessment perspective, that there isn't any harm in including those chemicals in the PFAS definition.

• Expert 7

  None of the above.
  For drinking water all the longer chain compounds can be eliminated. The above have their value in academic, regulatory or other context but for the purpose, it does not need a general definition since only very compounds need to be addressed (and not the many thousand or even millions of compounds that would fall into the OECD 2021 definition)