Dioxins vs. PFAS: Science and Policy Challenges

Abstract

Background:

Dioxin-like chemicals are a group of ubiquitous environmental toxicants that received intense attention in the last two decades of the 20th century. Through extensive mechanistic research and validation, the global community has agreed upon a regulatory strategy for these chemicals that centers on their common additive activation of a single receptor. Applying these regulations has led to decreased exposure in most populations studied. As dioxin-like chemicals moved out of the limelight, research and media attention has turned to other concerning contaminants, including per- and polyfluoroalkyl substances (PFAS). During the 20th century, PFAS were also being quietly emitted into the environment, but only in the last 20 years have we realized the serious threat they pose to health. There is active debate about how to appropriately classify and regulate the thousands of known PFAS and finding a solution for these “forever chemicals” is of the utmost urgency.

Objectives:

Here, we compare important features of dioxin-like chemicals and PFAS, including the history, mechanism of action, and effective upstream regulatory strategies, with the objective of gleaning insight from the past to improve strategies for addressing PFAS.

Discussion:

The differences between these two chemical classes means that regulatory strategies for dioxin-like chemicals will not be appropriate for PFAS. PFAS exert toxicity by both receptor-based and nonreceptor-based mechanisms, which complicates mixtures evaluation and stymies efforts to develop inexpensive assays that accurately capture toxicity. Furthermore, dioxin-like chemicals were unwanted byproducts, but PFAS are useful and valuable, which has led to intense resistance against efforts to restrict their production. Nonetheless, useful lessons can be drawn from dioxin-like chemicals and applied to PFAS, including eliminating nonessential production of new PFAS and proactive investment in environmental remediation to address their extraordinarily long environmental persistence. https://doi.org/10.1289/EHP14449

Introduction

The Stockholm Convention on Persistent Organic Pollutants (POPs) was enacted in 2004 as an international response to concern over some of the world’s most harmful chemicals.¹ This framework targets compounds that are not only toxic but are persistent in the environment and bioaccumulate through trophic levels and serves as the basis for the argument that these chemicals are especially deserving of reduction or elimination. At its inception, 12 chemicals, commonly known as the “dirty dozen,” were chosen to be the focus of the Convention. Periodically, new POPs have been added, reflecting both a growing understanding of our exposures and a growing number of industrial chemicals being added to the market.

Among the original “dirty dozen” were 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and its structurally related cousins, the polychlorinated dibenzo-p-dioxins (PCDDs) and the polychlorinated dibenzofurans (PCDFs). TCDD is often cited as the most toxic anthropogenic chemical ever created.^2–4 This toxicity, in combination with its extraordinary environmental persistence, secured its position as a top priority for the Stockholm Convention from its inception.² The PCDDs and PCDFs were unintentionally generated in large quantities beginning in the 1940s and peaking in the 1960s through the early 1980s and are notoriously persistent in the environment.⁵ The bulk of these chemicals were produced as unwanted byproducts from waste incineration, metal smelting and refining, and the manufacturing of certain useful chlorinated chemicals. In humans and wildlife, these chemicals exert a range of toxic effects, including effects on the immune, endocrine, cardiovascular, neural, and reproductive systems; developmental processes; and cancer.³ TCDD can exert acute toxicity and has been involved in several high-profile poisonings, both accidental, as in the 1976 disaster at Seveso, Italy,⁶ and intentional, as in the 2004 assassination attempt on former Ukrainian president Viktor Yushchenko.⁷ From an environmental health perspective, the concern over these chemicals has been lower-level, chronic exposures for the general population⁸ as well as exposures for some highly contaminated communities.^9–11

Regulation of the dioxin-like chemicals has been a resounding success in terms of reducing emissions; today, emissions of PCDDs and PCDFs are a mere fraction of what they were at their peak,¹² and concentration of dioxin-like compounds in human serum has also decreased.^8,13 Unfortunately, dioxins are slow to degrade in the environment, and because of their high toxicity, we believe even low levels in serum represent concerning levels of exposure for populations worldwide.^13,14

Even though dioxins are no longer in the spotlight, we argue lessons learned from this group of chemicals can help us tackle more contemporary challenges. The toxic effects of per- and polyfluorinated alkyl substances (PFAS), often referred to as the “forever chemicals,” are one of today’s most difficult environmental health problems. PFAS are an enormous class of fluorinated organic chemicals that are highly persistent in the environment¹⁵ and within humans,¹⁶ and several have been added as additional POPs to the Stockholm Convention.^17–19 First synthesized in the 1930s²⁰ and initially lauded as an innovative breakthrough in a wide range of industries, PFAS are raising alarms for human health globally because they are ubiquitous in the environment²¹ and strongly linked in laboratory and epidemiological studies to toxic effects.^22–24 Like dioxins, PFAS were produced and emitted extensively before any hint of health problems was widely publicized.²⁵ However, unlike dioxins, which were merely unwanted contaminants from uncontrolled combustion and certain industrial processes, PFAS are intentionally produced, highly useful, and valuable.^26,27 They have applications ranging from surfactants and chemical engineering processes to consumer products like nonstick items.²⁶ Their physiochemical properties that make them so useful, such as their stability and simultaneous hydrophobicity and lipophobicity,²⁶ also produce complex interactions in biota^28–30 and cause them to defy regulatory strategies.³¹ Recently, final health advisories from US EPA for two PFAS (PFOA and PFOS) were set at values below current analytical limits of detection and quantitation for routine water monitoring, which has brought additional public attention to this class of compounds.³² Indeed, new information on PFAS and decisions regarding their control and use are currently arising rapidly.

Though there are obvious differences between dioxins and PFAS, including their chemical structure and toxic mechanisms of action, we argue there are striking similarities that warrant a thoughtful comparison between these two groups. Both are extremely persistent in the environment, and both were emitted at high levels decades before the public and the environmental health research community became aware of their negative health effects. These similar characteristics suggest that we can learn from the dioxin-like chemicals to help us tackle the unprecedented challenge of PFAS.

Discussion

Definitions of the Chemical Classes

The boundaries of the category “dioxin-like compounds” have been cleanly defined since the 1990s.^33,34 An internationally harmonized regulatory approach from the World Health Organization set key criteria for inclusion.³⁴ One criterion is structural similarity – all these chemicals are structurally related to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD, or simply “dioxin”) in that they all have a dibenzo-p-dioxin, dibenzofuran, or a biphenyl core surrounded by a varying number and position of halogen atoms. The term “dioxin-like” chemicals also has a mechanistic component—to be included, these chemicals must bind to and activate the aryl hydrocarbon receptor (AhR).^34,35 The group includes 7 PCDDs, 10 PCDFs, and 12 polychlorinated biphenyls (PCBs), as well as the brominated congeners. Here, we focus on the unintentionally created PCDDs and PCDFs rather than the mostly synthetic PCBs, which were banned in the United States in 1976 with the passage of the Toxic Substances Control Act.^36,37

Although the dioxin-like chemical class is neatly defined, there is no universal agreement on which chemicals are included within the boundaries of the term PFAS. The first proposal for a standard definition of PFAS was made in 2011 by Buck et al.³⁸ They defined PFAS as aliphatic substances containing the ${\mathrm{C}}_{\mathrm{n}}{\mathrm{F}}_{2\mathrm{n}+1}$ moiety—molecules containing at least one carbon atom whose hydrogens have all been replaced by fluorine that terminate in a -CF3 group, and that also have a variable functional group. This classic definition is still widely referenced, though subsequent influential publications have offered alternative definitions.^16,26,39 In 2018, the Organisation for Economic Co-operation and Development (OECD) reported PFAS lacking the terminal -CF3 group,³⁹ and in 2021 the OECD published a definition that included chemicals containing either a perfluorinated methyl group or a perfluorinated methylene group.⁴⁰ There is also debate about whether the definition should be expanded to include additional fluorinated compounds, such as those with an aromatic component or compounds which are polyfluorinated and contain one or more other halogens (I, Br, or Cl).¹⁶ In 2021, the OECD tallied more than 4,700 PFAS with an assigned CAS number that may be on the global market,⁴⁰ but the number grows when we expand the scope to include degradants and impurities that are themselves PFAS. For instance, a recent tally by the US EPA’s CompTox Chemistry Dashboard reports more than 14,000 PFAS.⁴¹ And as analytical chemistry techniques continue to innovate, new PFAS are discovered all the time.⁴² In this article, when we refer to all PFAS as a class, we will reference the 2021 OECD definition; however, we recognize that stakeholders are having active conversations about which definition of PFAS is best. Nevertheless, for any of these definitions, the number of PFAS dwarfs the dioxin-like chemicals, which is one reason PFAS pose such a problem.

In addition to the sheer number of chemicals in the group, PFAS are subcategorized along many axes, with each grouping strategy lending itself to different applications.³¹ We can, for example, consider short-chain vs. long-chain PFAS, which differ in the length of their carbon chain. Long-chain PFAS (generally containing 8 for the carboxylic acids and 6 for the sulfonic acids), are the earlier, “legacy” PFAS and have longer biological half-lives, whereas short-chain PFAS are a more recent innovation and have shorter biological half-lives.³⁸ There are also the ultrashort chain PFAS⁴³ such as trifluoroacetic acid and pentafluoropropionic acid. We can also categorize PFAS based on how they were made: Mixtures made with electrochemical fluorination (ECF) tend to contain isomers with branched carbon chains, whereas products of telomerization have linear carbon chains.³⁸ These structural characteristics may enable identification of the manufacturer of PFAS samples from the environment.⁴⁴ Researchers and risk assessors also group PFAS based on whether they are perfluorinated (fully fluorinated) or poly fluorinated (incompletely fluorinated); by their functional group (with the carboxylic and sulfonic acids getting the most attention); whether they are the original products, degradation byproducts, or production byproducts; what they degrade into (with some terminal degradation products being of greater concern); polymers or nonpolymers; by their volatility; by their pKa; and so on.¹⁶ An excellent summary of these categories can be found in the Interstate Technology and Regulatory Council’s technical resources on PFAS.¹⁶ Although each method of subcategorization has uses, most are limited when it comes to regulating PFAS for their toxicity because each subgroup contains significant toxicological heterogeneity.¹⁵

We argue the class definitions of both dioxin-like chemicals and PFAS inform and are informed by the toxic mechanisms of action and the regulatory approaches used. For dioxin-like chemicals, a neat definition is firmly cemented in the World Health Organization’s toxic equivalencies (TEQ) approach from the 1990s.³⁴ However, the story is still unfolding for PFAS, because a consensus definition has not yet been reached. The unclear boundaries of the term PFAS pose serious communication roadblocks for groups working on regulatory strategies. To gain further insight, we will next examine the timelines, mechanisms, and regulatory strategies of these two groups.

Timeline of Dioxins and PFAS

Many anthropogenic toxicants found in the environment have timelines that fit a general pattern described by the Interstate Technology and Regulatory Council.¹⁶ In this paradigm, the chemical of interest is first produced in enormous quantities with low concern for potential toxicity, either intentionally (as for PFAS) or unintentionally as a contaminant of some large-scale industrial processes (as for PCDDs and PCDFs). Then, awareness of potential health effects by researchers and society develops in subsequent decades as concerning evidence emerges. Around this time, we also see efforts from governments, industry, and others to reduce or eliminate the chemical. The sequence of events for PFAS, from initial production to the extensive health concerns and regulatory efforts unfolding today, has parallels to the story of the dioxin-like chemicals.

After the beginning of anthropogenic emissions of dioxin-like chemicals in the 1930s and 1940s, uncontrolled municipal waste incineration, as well as emission from industrial processes, contributed to their rapid growth and their peak in the 1960s through 1980s.^5,45 In addition to industrial combustion, other important sources were metal smelting and refining, production of chlorinated biocides, and production of paper and pulp products. Dioxin-like chemicals experienced a boom in attention from the public sphere and research communities in the 1970s and 1980s due to high-profile exposures, including rice oil poisoning events in Japan and Taiwan,⁴⁶ exposure to Agent Orange during the Vietnam War,^47,48 soil contamination in Times Beach, Missouri,⁴⁹ and an industrial explosion in Seveso, Italy in 1976.⁵⁰ Focused research efforts rapidly generated toxicological data and culminated in development of a cohesive, globally standardized approach to defining toxicity of mixtures of dioxin-like chemicals.³⁴ Attention from the public sphere drove efforts to curb emissions of PCDDs and PCDFs, and regulatory actions played a key role in their successful control in the United States.¹² Between 1987 and 2000, there was a 90% reduction in emissions of PCDDs/PCDFs to all media in the United States, in large part due to regulation of large-scale combustion by the Clean Air Act Amendments of 1990.¹² Actions were taken to decrease emissions from other sources as well.^12,51 In the 1980s and 1990s, the US EPA canceled registrations of several pesticides that were contaminated by PCDDs and PCDFs by virtue of their production process.¹² And in 1998, the US EPA published a final rule limiting toxic chemicals, including dioxins and furans, in effluent from paper and pulp facilities.⁵²

As a direct result of these control measures, new emissions of dioxin-like chemicals from all sources are down in the United States, and a similar trend is found in other developed nations.^45,53 Due to this dramatic decline, the historically steady, low levels of nonindustrial combustion, such as backyard burning, are now the major source of dioxin-like chemicals in developed countries.^53–55 The trend of declining dioxin emissions in the United States and Europe has translated to a decline in exposure; dioxin concentrations in serum in the United States peaked around 1970,⁴⁵ and this trend has fortunately continued into the 21st century.^8,13

In developing and more-recently developed nations, dioxin emissions are higher and may have different sources. Cao et al. found that countries with lower GDPs tend to have more emissions from geographically dispersed nonindustrial sources, such as burning of residential waste, and that these sources are more difficult to regulate.⁵⁶ Ssebugere et al. concluded that dioxin emissions in African nations stem from both industrial and nonindustrial anthropogenic sources and urged more government intervention to control these emissions.⁵⁷ Simultaneously, countries with more recent booms in industrial activity are making progress on regulating dioxin pollution, albeit on a delayed timeline.⁵⁸ Critically, data on dioxin-like chemical emissions are incomplete for many parts of the world,^8,13,57 and there is a need to fill these data gaps to inform health-protective policies.

Despite the overall trend toward decreased emissions and increased regulation, even low levels of dioxins in the environment pose a threat to human health.^13,14 And even in places where emissions have been cut, there is nontrivial exposure from food because of the extraordinary persistence of these chemicals.¹⁴ Public attention has moved on, but the extraordinary environmental persistence of dioxin-like chemicals ensures we will not soon forget the errors of decades past.

The saga of the dioxin-like compounds was characterized by uncontrolled pollution and widespread exposure before concern from the general public and extensive research efforts drove initiatives to reduce emissions.^45,59 These same features are currently playing out with PFAS.

The first synthesis of any PFAS compound occurred in 1938 when a chemist at DuPont was experimenting with fluorinated refrigerants and inadvertently created polytetrafluorethylene (PTFE), which was later marketed as Teflon.²⁰ The new substance had interesting chemical properties, including repelling oil and water and being resistant to heat, and DuPont quickly recognized the potential applications of this new substance. Mass production of PFAS started in the 1950s and increased dramatically through the rest of the 20th century,¹⁶ resulting in extensive releases into the environment.

Until the late 1990s, nonindustry research communities and the general public were not aware of the general presence of PFAS in people or of the dangers posed by PFAS.⁶⁰ However, PFAS were recognized as toxic by the companies that made them as early as 1978 as evidenced by corporate documents revealed in lawsuits after the turn of the century.^61–63 One famous lawsuit⁶⁴ in 2005 revealed massive quantities of documents that proved the manufacturer knew the dangers of PFOA but failed to notify the US EPA as required by the Toxic Substances Control Act. However, all the while, PFAS were being emitted into the environment, and populations worldwide were being exposed—through spills near chemical plants, through routine industrial waste, spraying of firefighting foams, and through use of consumer products.^65–69 In 1999, scientists working at 3M published investigations of PFAS in human sera,⁷⁰ spurring a first step toward reduction in PFAS production; between 2000 and 2002, one famous PFAS compound, PFOS, and some of its precursors were voluntarily phased out by 3M.^70,71

Today, PFAS are copiously emitted into the environment from both point and nonpoint sources. Emissions into both air and water come from industrial production of a variety of PFAS, and in some cases the contamination of drinking water can be traced to nearby PFAS production and use facilities.^72,73 Firefighting foams, or aqueous film-forming foams, used at military installations, fire training sites, and airports, are a major point source of contamination to surface and groundwater.^74,75 Other point sources of PFAS include release from suboptimal landfills into groundwater⁷⁶ and release from wastewater treatment plants.⁷⁷ There are also concerns about emission from consumer use of PFAS-containing products, including textiles, housewares, and personal care products.²¹ PFAS generally are highly mobile in the environment³⁸ and may enter the boundaries of a community or watershed via precipitation, runoff, atmospheric deposition, and so on, which then becomes a burden on the people residing in that area. In this commentary, we focus on upstream regulation—restricting the production of new PFAS—while acknowledging that even a global embargo on PFAS manufacture would not guarantee that PFAS levels in every community would stop increasing.

The countless routes of PFAS emission into the environment, combined with their environmental persistence, provide abundant opportunities for human exposure. PFAS have been found in human tissues across the globe, including in the blood of essentially all Americans tested.⁷⁸ Major sources of exposure in nonoccupationally exposed populations are diet, consumer products, and house dust.^16,75 For populations living with contaminated water supplies, drinking water becomes the main source of PFAS exposure.¹⁶ Critically, even more than dioxin-like compounds, PFAS are incredibly persistent within the environment,^79,80 resulting in people (and wildlife) being continually exposed, even in cases when new emissions are successfully eliminated.

The timeline of the dioxin-like chemicals has passed its heyday; current research on the matter focuses on remediation, and media coverage is only an echo of what it once was.⁸¹ But the PFAS saga is actively unfolding, with stakeholders engaging in passionate debate over how to manage these chemicals. As we shall soon see, for dioxin-like chemicals that choice was informed by their mechanism of action, which is less well-defined for PFAS.

Mechanisms of Action

When health concerns warrant regulation of a group of chemicals, regulatory decisions can be supported by an understanding of the mechanism of action. The mechanisms of both dioxin-like chemicals and PFAS have been the subjects of intense research. But although the mechanism for dioxin-like chemicals was elucidated before the end of the 20th century, fully characterized mechanisms for PFAS still elude us due to their complexity.

Dioxin-like chemicals share a common molecular initiating event: binding and activation of the aryl hydrocarbon receptor (AhR). The evidence for AhR as the mediator of toxicity of dioxin-like compounds is unequivocal.² We know that the binding affinity of a dioxin-like chemical to the AhR dictates its potency; those that bind AhR to a lesser degree require higher doses to achieve the same effect.⁸² The AhR is highly conserved among vertebrates, and although there are some differences in the toxic effects of these chemicals between species, there are notable overlapping effects, and the effects within a given test species are the same.⁸³ Furthermore, AhR-knockout mice are mostly unaffected by doses of TCDD 10-fold higher than the acutely toxic dose in wild-type mice.⁸⁴ The necessary and sufficient role of AhR in the toxic effect of dioxin-like compounds translated well to screening compounds and drawing conclusions about their toxicity based on this one metric, and we believe this facilitated an expedited timeline between discovery and successful emissions reduction compared with PFAS.

The time between identification of the AhR and the first globally harmonized regulatory strategy for dioxin-like compounds was just 21 y. The Ah receptor was first identified in 1976,⁸⁵ and the first relative potency approaches were implemented in the late 1980s.^86–88 By 1997, there were sufficient in vivo and in vitro data to establish globally standardized relative potencies for a core subset of dioxin-like compounds.³⁴ However, the same milestones for PFAS are not yet reached. The mechanistic complexity of PFAS has stymied efforts to fully characterize the mechanism of action.

Unlike dioxin-like compounds, whose toxic effects are almost all mediated by a single receptor, the toxicity of any given PFAS is governed by multiple receptor and nonreceptor mechanisms. The first identified and best understood mechanism of action of PFAS is their ability to bind to the peroxisome proliferator–activated receptors (PPAR). The PPAR family was first identified in 1990,⁸⁹ and research on mechanisms of action for PFAS did not begin until around the turn of the century; the first evidence that PFOA binds to $\text{PPAR}\mathrm{\alpha }$ was published in 1999.⁹⁰ Since then, numerous studies have investigated the ability of different PFAS to bind and activate PPARs, particularly $\text{PPAR}\mathrm{\alpha }$ and $\text{PPAR}\mathrm{\gamma }$.^91–97 In the absence of exogenous ligands, PPARs bind to specific endogenous lipid ligands, translocate to the nucleus, and function as transcription factors. Activated PPARs affect expression of genes related to metabolism, including processing of fatty acids, lipoproteins, and glucose homeostasis.⁹⁸ Drugs that bind and activate the PPARs have effects on plasma levels of triglycerides and cholesterol,⁹⁹ which has disturbing implications for the potential health effects of PFAS. Binding of PFAS to $\text{PPAR}\mathrm{\alpha }$ leads to liver tumors in mice, but liver tumors can also occur in $\text{PPAR}\mathrm{\alpha }$ knockout mice, signifying alternative mechanisms.⁹¹

Additional in vivo and in vitro research has identified the activation of other receptors by PFAS, including the constitutive androstane receptor (CAR)^100,101 and the pregnane X-receptor (PXR).^101,102 Recent in silico evidence suggests that, to varying degrees, different PFAS interact with as many as 14 total nuclear receptors,¹⁰³ including the androgen receptor (AR) and estrogen receptors (ERs). These receptors vary in their expression in different tissues and in various stages of development, further complicating the PFAS mechanism of action. For example, $\text{PPAR}\mathrm{\alpha }$ is highly expressed in the liver, heart, and small intestine, in contrast to $\text{PPAR}\mathrm{\gamma }$, which is highly expressed in adipose tissue.⁹⁹ Much research is still required on downstream effects of PFAS binding to nuclear receptors. To date, despite strong association of PFAS with health effects, including impairment of the immune system and some cancers,^23,24 researchers have not drawn a clear line between any molecular initiating event of PFAS and the pathophysiology observed in epidemiological studies.²²

Furthermore, unlike the dioxin-like compounds, PFAS also have multiple putative nonreceptor-based molecular initiating events that appear to have nontrivial effects. These include interfering with lipid metabolism, partitioning into lipid bilayers, interfering with mitochondrial function, causing oxidative stress, and interfering with protein-protein interactions and cell signaling pathways by directly binding to proteins.²² It is interesting to note that several putative mechanistic components of PFAS can be attributed to their structural resemblance to fatty acids or other amphipathic linear lipids. The nuclear receptors mentioned above have endogenous ligands that are fatty acids or related lipids,¹⁰⁴ and researchers have noted the structural resemblance between fatty acids and many PFAS compounds, which have long carbon chains and polar heads.¹⁰⁵ Structural similarity of some PFAS to phospholipids has been confirmed by researchers who demonstrated partitioning into lipid bilayers.¹⁰⁶ We still need to rigorously test how these contribute to the observed cellular and organismal effects of PFAS.²²

We will next discuss how the single molecular initiating event of the dioxin-like compounds makes them amenable to an effective regulatory policy, although the multiple bases for biological effects of PFAS make it difficult to choose an ideal regulatory strategy.

Regulatory Approaches

We believe the fundamentally different properties of the dioxin-like chemicals and PFAS—physicochemical, mechanistic, and otherwise—translate to a fundamental need for different regulatory approaches. The characteristics of dioxin-like compounds we have discussed so far—their unified initiating mechanism of action and their persistence in the environment—make them suited to a special form of the relative potency factor approach called toxic equivalencies. This framework has four key criteria for inclusion as a dioxin-like chemical: a) it must be structurally related to TCDD; b) it must bind to the AhR; c) its biochemical and toxic responses are mediated by activation of the AhR; and d) it must be persistent and bioaccumulate.³⁴ For each chemical that meets all criteria, the approach assigns a toxic equivalency factor (TEF) that summarizes the toxic effects across all toxic end points and multiple test species. The toxic equivalency (TEQ) approach can be applied to dioxin-like mixtures of known composition to establish their potency in terms of an equivalent amount of pure TCDD, the most toxic of the dioxin-like chemicals.³⁴ At the inception of the TEQ method in the 1990s, the alternatives were limited; other approaches were to consider all these chemicals as equally toxic to TCDD, or on the opposite extreme, ignoring the congeners that lacked toxicological data.¹⁰⁷ The TEQ approach was therefore a rational middle ground. Initially conceived as a placeholder, the TEQ approach has proven to be a useful framework for dioxin-like chemicals. These early relative potency values have undergone only minor changes since their initial publication,^34,108,109 and the framework is still used today.^35,110 An interesting feature of the dioxin-like chemicals is that TEQ values are used to quantify emissions¹² as well as to determine the degree of contamination at a given site.¹¹¹ Thus, our mechanistic toxicological knowledge of these chemicals, which includes 17 unintentionally produced dioxins and furans, is tied to both upstream and downstream regulation.

PFAS, on the other hand, likely require a different approach. The cornerstone of the toxic equivalency approach for dioxin-like compounds is that they all agonize the AhR, but there is no evidence for an equivalent mechanism for PFAS, and this fundamentally prevents assignment of TEFs for PFAS.¹¹² However, useful strides have been made with other approaches for screening and risk assessment. For example, relative potency factors (RPFs), which are related to toxic equivalency factors, may have use for PFAS. RPFs apply to the relative potency of a particular end point, such as hepatoxicity, immunotoxicity, or developmental toxicity, and are less generalized than TEFs, which summarize all toxicity end points in a single number. Bil et al. calculated RPFs for 22 PFAS based on liver toxicity in male rats and gave examples of how this approach might be used in risk assessment.¹¹³ They acknowledge, however, that the PFAS they studied have important end points outside of liver toxicity, illustrating a major limitation of the RPF approach. For example, although PFOA is a more potent activator of $\text{PPAR}\mathrm{\alpha }$ than is PFOS, PFOS is more potent as a developmental toxicant¹¹²; this is a clear example of how the higher potency of a given PFAS for one end point does not determine a similar outcome for all end points. Analysis from another group demonstrated that the relative potency factor for PFOS relative to PFOA depended on the end point measured (such as pup survival or pup liver glycogen in rats) and ranged more than 20-fold.¹¹⁴ This characteristic makes it inappropriate to use a single RPF for all PFAS.

Although the RPF approach has its limitations, it is an example of an alternative to the chemical-by-chemical approach. Rules that impose restrictions on only one or a handful of PFAS represent progress, but we believe they are insufficient for a long-term solution. For example, in 2020 the US Food and Drug Administration (FDA) issued a voluntary phaseout of food contact substances containing a single PFAS, 6:2 FTOH.¹¹⁵ In the United States, the Unregulated Contaminant Monitoring Rules (UCMR) require monitoring of some currently unregulated chemicals in drinking water and are periodically updated. UCMR 3 included 4 PFAS (PFHpA, PFNA, PFBS, and PFHxS), and the upcoming UCMR 5 includes measurement of 29 individual PFAS in municipal drinking water.^116,117 However, with thousands of PFAS posing a potential threat, performing monitoring and toxicity testing on a chemical-by-chemical basis is impossible. Measuring and calculating the toxic equivalency factors for dioxins was a challenging but finite endeavor. The monetary and time cost of such an effort for PFAS would be astronomical.

Efforts are underway to restrict production of some PFAS subclasses and allow others to be produced. One of the largest examples of this approach has been a phaseout of long-chain PFAS in favor of shorter-chain PFAS, with the rationale that shorter-chain PFAS have shorter half-lives in biota¹⁶ and are therefore less toxic. For example, as part of the US EPA’s 2010/2015 Stewardship Program, eight major manufacturers of PFAS phased out long-chain precursors of PFOA.¹¹⁸ However, for the handful of short-chain PFAS that have been studied, biological effects similar to those of long-chain PFAS have been found.^119–122 Critically, short-chain PFAS are also highly persistent in the environment: They do not naturally degrade to non-PFAS substances. Continual, poorly reversible exposure¹²³ may mean that the shorter biological half-lives of these chemicals are not sufficient to protect human health. We believe the transition from longer to shorter-chain PFAS is an example of why regulating only subgroups of PFAS may be insufficient in the long term. There is a tendency toward invention of replacement chemicals that are untested but have similar properties. And because PFAS are intentionally created but PCDDs and PCDFs were unintentional, we argue that PFAS are more challenging to regulate because of the potential for regrettable substitution.¹²⁴

Mechanistic understanding for the thousands of PFAS in the environment currently eludes us and likely will continue to do so for some time absent paradigm-shifting innovations in toxicity testing. But we have enough data about PFAS to warrant concern for human health and to know how persistent PFAS are in the environment. Here, we argue that an important first step to getting ahead of this problem is broad, upstream restriction of PFAS production as a chemical class, an approach that has already gained significant support.^15,124,125 For dioxin-like chemicals, governments and industry worked to successfully cut emissions. We believe the same urgency should be applied to stem the production of new PFAS, whose environmental persistence is even greater.

If PFAS are to be regulated broadly as a class, our previously mentioned discussion of the definition of PFAS becomes even more important. For instance, industry often cites some PFAS as deserving exceptions to a uniform approach, but they have problematic life cycles. Fluoropolymers, for example, although toxicologically inert,^126,127 require other, highly toxic PFAS for their production, and many fluoropolymers also break down during their lifecycle to release monomeric PFAS, including perfluoroalkyl carboxylic acids (PFCAs).^15,128 Thus, we argue that broad restrictions on PFAS as a class are appropriate.

A practical implementation of a classwide restriction would involve the concept of essential use. This model bans all PFAS within specified use categories, such as food contact materials, firefighting foam, and textiles, but allows other use categories that are considered critical for health or safety, such as fluoropolymers in some medical devices.¹²⁹ A successful example of essential use has been the Stockholm Convention, which bans three key PFAS (PFOA, PFOS, and PFHxS) but allows limited acceptable uses.¹

Stemming the flow of PFAS production is just part of the solution. PFAS are already ubiquitous in the environment, with many communities suffering from serious contamination. This ubiquity raises the questions of how much PFAS are acceptable in various environmental media, and if levels are too high, how will they be remediated and who will pay for it? Although the complex challenges of cleanup and remediation are not the focus of this commentary, we issue a call to action for funding cleanup of contaminated communities, as well as advancing research on remediation of PFAS.

Conclusion

TCDD and related compounds that bind the AhR are infamous, with a storied history of public concern. These chemicals act at extremely low doses; the tolerable weekly intake from food is just $2\text{\hspace{0.17em}pg}/\mathrm{kg}$ body weight.¹⁴ Even though the generation of dioxin-like chemicals has waned thanks to a successful program of regulation and emissions reduction, most dioxins produced in the 20th century are still in the environment due to their slow rate of degradation.

When the pressure to control dioxin emissions came to a head in the 1990s, governments were able to mobilize and require alterations to the processes producing these harmful waste products. But unlike dioxin-like chemicals, PFAS are extremely valuable and useful,^26,27 and in our opinion, the pushback from industry against regulating them is intense because it acutely disrupts commercial and financial interests.^130–132 Powerful litigation campaigns from chemical manufacturers,⁶⁰ combined with ambiguity and disagreement over best upstream regulatory strategies, are delaying attempts to rein in production of new PFAS. Because of their usefulness, their complex mechanisms that evade regulatory and toxicity screening strategies, and their extreme persistence in the environment, we argue that PFAS have the potential for a greater negative impact on environmental health than TCDD and the dioxin-like compounds, even though they are not as toxic on a molecule-for-molecule basis.

There are still lessons from the story of dioxins that can help us tackle the PFAS problem. The lingering effects of dioxin as microcontamination of the food supply and highly contaminated communities reveal the need for proactive funding of remediation research and allocating funds to communities for cleanup. This is especially true for PFAS because their environmental persistence surpasses dioxin-like chemicals^5,16; even if we completely stopped creating new carbon-fluorine bonds immediately, absent sufficient remediation, exposures to forever chemicals would continue indefinitely.

Furthermore, because of the serious threat of PFAS to human health, and because of myriad barriers to successfully regulate PFAS individually or as subgroups, we favor regulatory approaches that restrict production of PFAS broadly. Although some have argued that this approach overestimates the hazard of PFAS, we maintain that the severity of the problem warrants a unilateral approach to protect human health and the environment. Efforts from multiple levels of government^133,134 and from industry that target PFAS as a class are beginning to take hold, such as 3M’s commitment to eliminating production of all PFAS by 2025.¹³⁵ We hope that framing PFAS in the historical context of another group of challenging chemicals, the dioxin-like compounds, will offer an opportunity for reflection and inspire actions toward health-protective policies.

Conclusions and opinions are those of the individual authors and do not necessarily reflect the policies or views of EHP Publishing or the National Institute of Environmental Health Sciences.