Per- and polyfluoroalkyl substances exposure science: current knowledge, information needs, future directions

Abstract

Background:

Per- and polyfluoroalkyl substances have been documented at all spatial scales with concerns of adverse ecological and human health effects. Human exposures and relative pathway contributions depend on the specific population, their exposure scenarios, and pathways of local sources.

Objectives:

Provide a narrative overview of (1) current per- and polyfluoroalkyl substances knowledge for sources, concentrations, and exposures; (2) critical per- and polyfluoroalkyl substances exposure information gaps and needs, and (3) United States Environmental Protection Agency’s strategies and action plans in collaboration with other federal, industrial, and academic partners.

Methods:

A literature review was conducted for per- and polyfluoroalkyl substances (primarily perfluorooctane sulfonate and perfluorooctanoic acid) compounds in blood, water, soil, house dust, indoor and outdoor air, consumer products, food, and fish, as well as per- and polyfluoroalkyl substances exposure modeling.

Results:

Large variability exists in measured per- and polyfluoroalkyl substances environmental concentrations and human exposures. Literature indicated that ingestion of food (“background”), drinking water (“contaminated” scenarios), and house dust (for children) are main pathways for perfluorooctane sulfonate and perfluorooctanoic acid.

Discussion:

Needs for addressing critical data gaps are identified. More information is available on long-chain per- and polyfluoroalkyl substances than for replacement and emerging compounds. A large-scale research effort by the United States Environmental Protection Agency and other federal agencies is underway for a better understanding of per- and polyfluoroalkyl substances exposures.

Introduction

During the early 1950s, per- and polyfluoroalkyl substances (PFAS), which include many types of anthropogenic chemicals, were introduced into consumer products because of their unique physiochemical properties (ATSDR 2017). Since then, PFAS have been released into the atmosphere while waste PFAS materials have been released into waters (McCord and Strynar, 2019). For the past six decades, PFAS have been used in various consumer products for multiple purposes such as surface coating materials used for non-stick cookware (ACS 2020), aqueous film forming foam (AFFF) for firefighting (Hoisater et al. 2019), and food packaging for food bags and wrappers (Schaider et al. 2017). As stable and persistent materials, PFAS were detected ubiquitously worldwide across different media in the environment (Vedagiri et al. 2018). PFAS have gained attention due to its adverse health and environmental effects (Steenland et al. 2020). Because of persistency, bioaccumulation, toxicity properties of PFAS, the 3M Company, which is the sole U.S. manufacturer of perfluorooctane sulfonate (PFOS), has voluntarily stopped the production of PFOS in 2002; by voluntary decision, 8 major industry producers have stated that they have reduced the production of designated perfluorinated carboxylic acid (PFCA) including perfluorooctanoic acid (PFOA), longer-chain PFCA, and related precursors by 95% by early 2010 and eliminated the designated PFCA emissions by 2015 (ITRC 2020). During that period, alternate or replacement PFAS (e.g., short-chain PFAS) have emerged in place of legacy long-chain (i.e., long carbon chain lengths include perfluorinated carboxylic acids ≥ C₇ and sulfonic acids ≥ C₆) PFAS into the U.S. markets. To date, over 5000 PFAS have been manufactured, and the physicochemical and toxicity properties of many of the PFAS remain to be discovered (USEPA 2019a). Due to their ubiquitous usage and persistence in the environment, people around the world have been exposed to different PFAS (Sunderland et al. 2018). PFAS have the potential to bioaccumulate in humans from low exposure in the general population (Steenland et al. 2020). Evidence indicates that continued exposure to certain PFAS such as PFOA or PFOS above specific levels may cause adverse health effects (USEPA 2016a, 2017a). PFAS may also pose potential health risks such as increased testicular and kidney cancer (Steenland et al. 2020) and be linked to decreased fertility (Bach et al. 2016), diagnosed high cholesterol (Eriksen et al. 2013), and thyroid problems (Caron-Beaudoin et al. 2019). PFAS contamination of drinking water has been one of the major concerns and between 2013 and 2015 the U.S. EPA required monitoring of drinking water for six PFAS, PFOS, PFOA, perfluorononanoic acid (PFNA), perfluorohexanesulfonic acid (PFHxS), perfluoroheptanoic acid (PFHpA), and perfluorobutanesulfonic acid (PFBS) as a part of the Third Unregulated Contaminant Monitoring Rule (UCMR 3) (USEPA 2017b). Also, in 2009 the U.S. EPA added PFOA and PFOS to the third draft drinking water contaminant candidate list (compounds not subjected to any national primary drinking water regulations). In 2016 the U.S. EPA issued a health advisory for drinking water with individual or combined levels of 70 parts per trillion (ppt) for PFOA and PFOS (USEPA 2016b). Multiple states have also established more stringent drinking water standards or guidance values for PFOA and PFOS (Cordner et al. 2019). The U.S. EPA is committed to following the Maximum Contamination Level (MCL) rulemaking process as established by the Safe Drinking Water Act (SDWA). The U.S. EPA has proposed a regulatory determination for PFOA and PFOS in drinking water and is considering nationwide drinking water monitoring for PFOA, PFOS, and other PFAS under the fifth UCMR monitoring cycle in 2023-2025 (USEPA 2019b, 2020a). Further, the U.S. EPA is also gathering and evaluating information to determine if regulation is appropriate for other chemicals in the PFAS family (USEPA 2019c).

Much material in this paper is based on a presentation at the February 5-6, 2018 Federal PFAS Information Exchange. Also, during 2018, the U.S. EPA hosted a National Leadership Summit (Washington, D.C.) to tackle PFAS issues in the environment (USEPA 2018a). In that summit over 220 participants were present, comprising senior officials from 40 states, 3 tribes, Guam, Northern Marianas Islands, 13 federal agencies, congressional staff, and dozens of associations, industry groups, and non-governmental organizations (USEPA 2018b). Some of the purposes of the summit were: (1) to share information on ongoing efforts to monitor and characterize risks from PFAS; (2) discuss specific near-term and long-term actions, beyond those already underway, that are needed to address challenges currently faced by states and local communities; and (3) discuss risk communication strategies to address public concerns with PFAS. The U.S. EPA also reviewed approximately 120K public comments, which established the important input for the PFAS Action Plan. Following a series of public Community Engagement Meetings, in February 2019 U.S. EPA released a PFAS Action Plan (USEPA 2019b; updated USEPA 2020a) and some of the details pertinent to this paper are provided in subsequent subsections.

As shown in Fig. 1 from the National Research Council (NRC) Exposure Science in the 21^st Century (ES21) report, there are many elements of exposure science. The backbone of human exposure science is measurements and methods development so that environmental sources can be evaluated and addressed (USEPA 2019d). Environmental source characterization, including occurrence in various environmental media, is a fundamental initial step and basis for the development of models that can fill measurement gaps and simulate various mitigation scenarios. This is a narrative paper covering current knowledge and information gaps on PFAS sources, environmental concentrations, and human exposures and doses.

Fig. 1.

Core elements of exposure science (adapted from NRC 2012)

PFAS sources include industrial sites, fire training/fighting facilities, landfills, wastewater treatment plants, consumer products, and food items (Sunderland et al. 2018). Exposure pathways are dependent on the type of PFAS source (Fig. 2). Exposures can occur through drinking contaminated water; ingestion of contaminated food, non-dietary ingestion, inhalation by breathing contaminated outdoor air from industrial or manufacturing sources, or indoor air contaminated from outdoor air or indoor consumer products, and dermal exposure (Egeghy and Lorber 2011). However, precursors to perfluoroalkyl acids (PFAA) (PreFAA), such as dipolyfluoroalkyl phosphates (diPAP), fluorotelomer alcohols (FTOH), perfluorooctyl sulfonamides (FOSA), and sulfonamidoethanols (FOSE), can be bio-transformed to PFAA, for example, which may also be a source of exposure (Makey et al. 2017). Fig. 2 adapted from Oliaei et al. (2013) illustrates that PFAS exposure pathways reflect a complex system requiring multimedia (e.g., air, water, and soil) thinking for systems-level understanding and risk management solutions.

Fig. 2.

The pathways for PFAS exposure to humans (adapted from Oliaei et al. 2013)

There are many exposure questions related to PFAS occurrence, sources, concentrations, multimedia exposures, relative source, and pathway contributions. Regarding environmental occurrence of PFAS in various media, key questions include the following: What methods are needed for comprehensive, reliable measurements? What is the variability in chemical mixtures over time and space? Regarding relative source contributions to environmental concentrations of PFAS, what are the sources, determinants, and “hot spots” of environmental occurrence? Regarding population distributions of exposure for priority PFAS: What are the sources and determinants and scenarios of exposure? How does human exposure to PFAS vary in time, space, demographics, consumer practices? What populations and life stages are vulnerable to high-level exposure? Regarding relative exposure pathway contributions for multimedia PFAS exposures: How do they vary for key scenarios and populations of interest? How can this information be used to inform risk mitigation/communication efforts and to focus future data collection efforts? Since measurements cannot be made everywhere and all the time, fate and transport modeling of PFAS helps to fill in the data gaps to increase our understanding of the life cycle of PFAS. This research brings in additional questions to help measurements as well as modeling efforts. Important aspects include but not limited to are: how are wide-ranging PFAS distributed across various environmental media? What are physiochemical properties of PFAS and how they differ from each other? How will PFAS modeling uncertainties impact the fate and transport simulations?

Of the several PFAS, PFOA and PFOS were widely studied, some data exist for all media though these data are not spatially and temporally collocated. However, larger data gaps exist for other PFAS. Our objectives were to focus on PFOA and PFOS with respect to: (1) sources, concentrations, exposures; (2) critical exposure information gaps and needs; and (3) ongoing and future scientific research to aid multi-media exposure modeling.

Methods

A series of web searches, reference citation checks, and an advanced Web of Science database search were conducted for literature published between 2003 and July 2020 that potentially contain PFAS, PFOA, PFOS, perfluorinated compound (PFC), and/or perfluoroalkoxy alkanes (PFA) environmental media concentration data or exposure modeling pertaining to human exposure in the U.S., Canada, Europe, and Asia.

The below Web of Science Query Logic were used to perform a literature survey:

((((TS=((PFOA OR PFOS OR PFAS OR PFA OR PFC OR PFCs OR "perfluorooctane sulfonate" OR "perfluorooctanoic acid" OR "perfluorinated compounds" OR "perfluor*" OR "perfluoroalkyl") Near/20 (level* OR concentration* OR exposure* OR bioavailab*) Near/20 (soil* OR *water* OR air OR dust* OR food* OR blood OR atmospher* OR "PM₁₀" OR "PM_2.5" OR TSP OR sediment OR diet* OR vegetable* OR fruit* OR "well water" OR "ground water" OR "drinking water" OR environment OR tap OR aerosol*)))))) AND LANGUAGE: (English)

Inclusion criteria used are: Contains PFAS, PFC, PFA, PFOA, and/or PFOS concentration data for blood, water, food, dust, air, soil/sediment, and/or consumer products that pertain to direct human exposure or near-field assessments/studies. Peer reviewed articles/literature published between 2003 and July 2020 are considered in this study. Data included in this study are from North America and Europe while data from Asia only for food are included (this exception has been made due to the scarcity of data in this medium).

Exclusion criteria: Experimental PFAS, PFC, PFA, PFOA, and/or PFOS data were excluded. Asian (except for food), African, Oceana, South American, and Antarctic data were not considered. Non-direct, fate and transport, and far-field assessments/studies are not included. If no digital or physical version of literature were available, then all such articles were excluded. Note that over 500 peer reviewed articles have been identified in this literature search, while only a subset of those articles is cited in this article by selecting the latest article for a type of research area.

Results

This section is organized in a logical way instead of a traditional way and it starts with the end-point of exposure, which is humans and then human exposure, dominant pathways of exposure in USA which are drinking water, soil, consumer products, house dust, indoor and outdoor air, and food.

PFAS in Human Blood

Multimedia, multipathway exposures over time have led to widely detected amounts of perfluoroalkyl substances in available biomonitoring studies. In Centers for Disease Control and Prevention (CDC)’s National Health and Nutrition Examination Survey (NHANES) (2013-2014), multiple PFAS such as PFC, PFOS, PFNA, and PFHxS were detected in blood of more than 99% of 2165 U.S. individuals (older than 12 years). NHANES and other studies showed that levels of PFOA and PFOS and other common PFAS but not all were generally decreasing in the blood of the general population from 1999 to 2014 (e.g., CDC 2019). Hurley et al. (2018) reported serum levels of 10 long-chain PFAS from more than 1200 California women during 2011-2015. The results showed that nearly all measured PFAS except PFHxS declined ~10%-20% per year, the phase-out of many PFAS had contributed to the reduced human exposures while non-declining PFHxS may indicate that environmental exposures are not going down. In a cross-sectional study in El Paso county, Barton et al. (2020) found that of the five PFAS, the median PFHxS concentrations in adults’ serum was about 12 times as high as the U.S. national average; their analysis indicated that the dominant PFAS exposure route may be through water ingestion and PFAS occurrence in the drinking water may come from AFFF used at the nearby Air Force Base. In another local cross-sectional study, Graber et al. (2019) found that PFNA concentration in serum in 192 residents of Paulsboro, New Jersey (NJ) to be about 285% higher than that in U.S. national average due to consumption of community water. Measured concentrations of 15 PFAS in home tap water during 1989-1990 for 225 participants at the national scale who consumed 8 or more cups of water per day were used by Hu et al. (2019) to estimate modeled plasma concentrations for comparison with measurements for a subset of 110 participants; it was found that for PFAS human exposure, the default relative source contribution (RSC) of drinking water (20%) used in risk assessment compared well with their modeling studies.

The half-lives of short- and long-chain PFAS in human blood varied and the longer-chain PFAS tend to have longer half-lives; the serum half-life of short-chain PFAS such as PFBS was around 44 d while long-chain PFAS such as PFOA and PFOS may stay in serum for a longer time; the serum half-live of PFOA and PFOS were around 1.77 and 2.93 years, respectively (Xu et al. 2020). In a longitudinal study encompassing several years from 1999-2011, Ding et al. (2020) found that serum concentrations of n-PFOA, n-PFOS, and sm-PFOS (legacy PFAS) decreased significantly, while serum concentration of PFNA (emerging PFAS) increased but these temporal trends in PFAS concentrations are not uniform across ethnic groups. In another longitudinal study, Kim et al. (2020) studied nine PFAS in serum samples collected in 2009–2016 from 450 Northern California mothers when their child was 2–5 years old with adjustment for characteristics that may affect maternal concentrations. They found that the PFOA, PFOS and PFHxS concentrations decreased over the study period while PFNA and PFDA showed mixed temporal trends. Lin et al. (2020) used linear mixed models and hazard models to estimate longitudinal associations between baseline PFAS and rate of blood pressure changes as well as the risk of developing hypertension. Overall, it was found that there were modest and mostly null associations of plasma PFAS concentrations with blood pressure and hypertension. Using NHANES data for the period 2013-2014, Calafat et al. (2019) found that the replacement PFAS-PFBA and perfluorohexanoic acid (PFHxA) were detected in urine and the 90^th percentile urine concentrations were 0.1 and 0.3 mg/L, respectively. Their results suggested that biomonitoring of short-chain PFAS in general population using urine was not supported. In another biomonitoring study of western New York State (ethnic) residents, Savadattia et al. (2019) observed higher geometric means for serum PFOS and other contaminants compared to the 2013-2014 NHANES reference levels.

Moreover, humans were also likely exposed to many PFAS that are not detected by the standard analytical methods, one study indicated the occurrence of some poorly understood PFAS [e.g., perfluoro-3,5,7,9-tetraoxadecanoicacid (PFO4DA)] in serum of residents from Wilmington, NC (Kotlarz et al. 2020). Thus, continuation of exposures to PFAS calls for continued biomonitoring to obtain additional data on exposure pathways. Fig. 3 shows PFOA blood levels in 3M and Dupont workers; the C8 health studies in West Virginia (WV) and Ohio (OH); residents of Hoosick Falls, New York (NY); and the U.S. population. It illustrates that some populations are more prone to be exposed to PFAS than others and PFAS blood levels depend on population, source proximity, and contact with contaminated media (NYSDH 2016).

Fig. 3.

PFAS in human blood from multimedia exposures over time (Adapted from NYSDH 2016)

Analyzing multimedia exposures can inform risk management and risk communication. For example, U.S. EPA’s drinking water health advisory for PFOA is calculated using a RSC of 20%, and other PFOA exposure sources such as dust, diet, and air make up the remaining 80% of the Reference Dose (RfD) (20 ng/kg/day RfD (tolerable daily intake)) (USEPA 2016a). Furthermore, in Superfund and other programs such as Resource Conservation and Recovery Act (RCRA) corrective action, site-related media in addition to drinking water are considered, including exposures to children from incidental ingestion of contaminated soil (ATSDR 2019). Analyzing multimedia exposures can also help focus data collection efforts on the most critical sources, media, and pathways for a given population and scenario.

From the ongoing evaluation of toxicity information for several PFAS by the U.S. EPA, it is not clear whether exposure to structurally similar PFAS results in similar health effects (Kotthoff and Bucking 2018). Literature indicated that long-chain PFAS, in general, may cause greater toxicity in humans than shorter-chain PFAS (e.g., Eschauzier et al. 2012), though the toxicities of short-chain PFAS have generally been less well studied (Danish EPA 2015).

Human Exposure Pathway Analyses for PFAS

Multiple studies in the published literature have applied a scenario-based risk assessment exposure modeling approach to examine relative contributions to PFAS blood levels through different exposure pathways. This involves combining measured concentrations in environmental media with exposure/dose factors and biomonitoring data to estimate aggregated exposures and pathway apportionment for different scenarios (low vs. high exposures, children vs. adults). There seems to be scientific consensus about dominant PFOS/PFOA exposure pathways (e.g., Lorber and Egeghy 2011). From these studies, it was found that ingestion from water, food, and dust is greater than dermal and inhalation exposure. As documented in the following sections, from a national perspective, the main pathway is expected to be from food ingestion for background or “typical scenarios” and drinking water ingestion from “contaminated scenarios”. Drinking water exposure is dominant for populations near sources of contaminated drinking water while inhalation may also be a potential dominant factor for populations near industrial sites. Long-term exposure to low levels of PFAS in drinking water may be a major exposure source for non-industrial regions. Ingestion of contaminated house dust is also considered an important source of exposure for children. Inhalation exposure for PFAS was reported as the main pathway for adults in these and other studies.

Egeghy and Lorber (2011) and Lorber and Egeghy (2011) presented multimedia exposure analyses results for PFOS and PFOA using a probabilistic approach. They also discussed various uncertainties such as the simple pharmacokinetic (PK) model component, limited food concentrations, absence of soil pathways, and assumptions about indoor vs. outdoor air concentrations. They found that food ingestion appears to be the primary route of exposure in the general population while contributions via specific pathways vary by several orders of magnitude with considerable overlap. Note that soil ingestion was not included because of lack of data, not necessarily because that exposure pathway is not important.

Gebbink et al. (2015) estimated daily exposures from direct and indirect (i.e., precursor) intake of PFOS and PFCA, for the general adult population in Sweden. They found that exposures were highest for PFOS and PFOA, followed by PFHxA, perfluorodecanoic acid (PFDA) while lower for pentafluorobenzoic acid (PFBA), and perfluorododecanoic acid (PFDoDA). For PFOS, dietary intake was dominant for all scenarios; for PFCAs, the dominant pathway depended on chain length and scenario. They also reported uncertainties including lack of precursor concentrations, uptake and biotransformation factors, and improved analytical methods. PFOS exposure estimates with updated data were lower than earlier ones, in line with temporal trend monitoring studies and improved analytical methods. The U.S. EPA is synthesizing data on exposure from various environmental media and evaluating health impacts of additional PFAS (USEPA 2019b).

PFAS in Drinking Water

In the UCMR 3 study, some PFAS were detected at concentrations greater than minimum reporting levels (MRL) (10-90 ppt) in 198 out of 4920 public water system (PWS), serving approximately 16 M people; 599 samples had PFAS above or equal to MRL (USEPA, 2020b). Out of 4920 PWS (small and large) 63 (1.3%) reported one or more UCMR 3 results for PFOA and PFOS above U.S. EPA’s health advisory value 70 ppt (Table 1). Other PFAS are being detected in addition to PFOS and PFOA (USEPA 2020b). PFAS have been detected in drinking water in proximity to PFAS manufacturing and use, and AFFF are significantly associated with PFAS in drinking water (e.g., Banzhaf et al. 2017).

Table 1.

Compounds of PFAS in drinking water (USEPA 2019e)

Analyte	% of PWSs with measurements ≥ MRL	Minimum of results ≥ MRL (ng/L)	Median of results ≥ MRL (ng/L)	Maximum of results ≥ MRL (ng/L)
PFOA	2.38	20	30	349
PFOS	1.93	40	60	7000
PFBS	0.16	90	170	370
PFHpA	1.75	10	20	410
PFHxS	1.12	30	70	1600
PFNA	0.28	22	32	56

^aUCMR 3 included all large PWSs and a nationally representative subset of small PWSs

The U.S. EPA and United States Geological Survey (USGS) collaborated on a study measuring contaminants of emerging concern in source and treated drinking water (Glassmeyer et al. 2017). In this independent measurement study, six of the drinking water treatment plants sampled by the UCMR 3 happened to be on two large rivers. The first had a PFAS mixture dominated by PFOA; the second river had mostly PFBA (Boone et al. 2019). For each river, the sampling points were separated in both time (6 to 12 months between samplings) and space (100 to 1000 km separated the locations), further illustrating PFAS’ environmental persistence. Further, they also found that in all 50 water samples collected across the U.S., the median summed concentrations of 17 PFAS were about 21.4 ng/L in the source water while it was about 19.5 ng/L in treated drinking water. This indicated minor effectiveness of the water treatment to remove PFAS at those facilities. In another study, Sun et al. (2016) studied the contaminants of Cape Fear River watershed in North Carolina (NC) and documented the presence of legacy and emerging PFAS. Of the 127 samplings, 57 exceeded the U.S. EPA’s health advisory level for PFOA (70 ppt) while a replacement compound, GenX, the ammonium salt of hexafluoropropylene oxide dimer acid (HFPO-DA), alternative for PFOA was detected at mean concentration of 631 ppt for 37 samples. Three other unknown compounds exhibited peaks exceeding 9000 ppt. At the single drinking water treatment plant (DWTP) site, analyzed samples indicated that the water treatment methods (coagulation, ozonation, biofiltration and disinfection) used at the plant were ineffective for removing perfluoroalkyl ether carboxylic acids (PFECA). In fact, for some of the PEFCA, concentrations in the treated water have increased as compared to the source water, potentially due to oxidation of precursor compounds in the source water. Further, other analyses of measurements indicated that conventional drinking water treatment (e.g., coagulation, ozonation, biofiltration and disinfection techniques) is not effective for removing PFAS. New treatment methods and technologies are being explored to remove PFAS in contaminated drinking water. The U.S. EPA’s 2008-2010 National Rivers and Streams Assessment (NRSA) provided PFAS data in 164 urban river sites; 2010 Great Lakes Study, 157 sites (Stahl et al. 2014). Of the 10 PFAS tested, maximum observed concentrations were 127 and 80 ng/g in urban river samples and Great Lakes samples, respectively. Stahl et al. (2014) noted that most of these samples (over about 90%) were collected from eastern U.S. sites indicating sparse data for western U.S. sites. Goodrow et al. (2020) studied the PFAS presence in measurements made at 11 targeted waterbodies in NJ. Shorter chain PFAS tended to be dominant in surface water while longer chain PFAS dominated in fish and sediments. The sum of all PFAS in water samples ranged from 22.9 to 279.5 ng/L while in sediments summed concentrations ranged from below detection levels to 30.9 ng/g.

In 2018, the U.S. EPA expanded the drinking water Method 537 for inclusion of the GenX and additional PFAS and the new expanded method is Method 537.1 (Shoemaker and Tettenhorst 2018). The U.S. EPA has also developed a new method for additional short-chain PFAS not considered in the Method 537.1, the newly validated Method 533 effectively complements Method 537.1 (USEPA 2019f). At present, validated U.S. EPA drinking water measurements methods are available for 29 PFAS and more are in the developmental phase. However, validated analytical methods for the vast number of exposure-relevant permutations considering media and chemical are lacking. For UCMR 5, EPA has proposed expanding the number of PFAS monitored to a total of 29 and with lower minimum reporting levels (MRLs) than for UCMR 3, thereby expanding our knowledge of PFAS occurrence in U.S. drinking water. These monitoring results will improve understanding of the frequency and concentration of many PFAS occurrence in treated U.S. drinking water.

PFAS in Soil

Long- and short-chain PFAS present in soil and biosolids (a residual waste from wastewater treatment plants) can be taken up into agricultural or aquatic-based food products. Literature (Washington et al. 2010) indicated that commercial PFAS precursors can degrade in soil settings with degradation half-lives of about 1 to 2 years, forming recalcitrant PFAS degradation products. Consequently, plants and macro-invertebrates take up PFAS from soils and sediments, and thus soil contamination can contribute to exposure through base components of the food chain or through partitioning into surface and ground water.

Exposure is also possible through soil and dust ingestion, which will vary according to age (lifestage) in young children. Soil concentrations will vary depending on location and can vary by about or more than four orders of magnitude; human exposures are therefore likely to be highly variable, depending on location and lifestages (Table 2). Furthermore, PFAS detection does not necessarily imply a local source and published literature provide data for a very limited set of PFAS based on a limited set of types of sites. Agencies are now compiling available data on types of sites and concentrations of PFAS found at different types of sites, including different types of industrial sites. Data are being collected by the U.S. EPA for some of the replacement and emerging PFAS for use in toxicity studies as well as modeling studies (USEPA 2019b).

Table 2.

PFAS contaminants in soil

Data Type	Compounds	Contamination route	Levels	Citation
Survey: 60 soils, 6 countries	PFOA & PFOS	Air deposition (proposed)	0.124 ng/g and 0.472 ng/g; All U.S. samples tested positive (n=10)	Strynar et al. 2012
Survey: 62 surface soils samples U.S., several islands, all continents	Range of PFCAs, PFSAs & precursors	Air deposition (proposed)	< 0.02 to 14.3 ng/g, with PFOS and PFOA most frequently detected	Rankin et al. 2016
Industrial (data limited, highly variable)	PFOA PFOS	Industrial discharge	Up to 48 ng/g Up to 10 ng/g	Zereitalabad et al. 2013
Biosolids	PFOS	Soil amendment, Land application	2-480 ng/g	Sepulvado et al. 2011
Biosolids	Range of PFCAs, PFSAs & precursors	Soil amendment, Land application	Up to 5500 ng/g	Washington et al. 2010
AFFF sites	PFAS PFOA PFOS	Fire training activity & runoff	Often > 1000 ng/g 21 ng/g 2400 ng/g	Houtz et al. 2013
AFFF: 40 U.S. Air Force release sites	15 PFAAs & PFOSA precursors	Fire training activity & runoff	Median 53 ng/g; max 9700 ng/g. PFOS, PFHxS, PFOA, PFOS at highest concs.	Anderson et al. 2016

For sites contaminated by aqueous film forming foam (AFFF) (Houtz et al. 2013), PFAS soil screening levels for various exposure scenarios can be calculated using RfD for PFOA and PFOS. These can be used in U.S. EPA’s standard equations to calculate soil screening levels for various exposure scenarios including residential, industrial, construction, and recreation. A soil screening level can also be calculated based on protection of groundwater (Simon 2019).

PFAS in Consumer Products

Consumer products containing PFAS are a potential source of human exposure, for example, to children playing on carpets treated with stain protectants. Hand-to-mouth transfer from surfaces treated with PFAS-containing stain protectants, such as carpets, is thought to be most significant for infants and toddlers (ATSDR 2017a). While PFOA, PFOS and other long-chain PFAS usage has been decreasing, other fluorochemicals are still found in a wide variety of consumer products (e.g., Kotthoff et al. 2014). Such products can include stain resistant coatings used on carpets and upholstery, water resistant clothing, grease-resistant paper, food packaging, nonstick cookware, cleaning products, personal care products, cosmetics, paints, varnishes, and sealants (ATSDR 2017a).

U.S. EPA’s PFOA Stewardship Program was launched in 2006 in cooperation with 8 major leading companies, with the goal of eliminating PFOA and long-chain PFAS from emissions and products by 2015 (USEPA 2020c). Liu et al. (2012) reported PFCAs presence in 116 consumer products/articles produced in 19 countries in 2007-2011 aiming to assess the human exposures to PFAS. The observed trends of measured PFCAs indicated the following results produced by the PFOA Stewardship Program: (1) the consumer articles treated with fluorinated (PFOA/PFOS) chemicals have been less available; (2) PFCA content has decreased substantially for most of the consumer articles; and (3) although PFOS is still being used in the market, the alternatives such as PFBS-C4 are being used for some products (Liu et al. 2012). All participating companies state that they have met the program goals while some companies not part of the Program continued to produce or import and use PFOA and other long-chain PFAS that were phased out under the program (USEPA 2020d).

The California Department of Toxic Substances Control (DTSC) Safer Consumer Products Program (SCP) has been evaluating PFAS in carpets, rugs, indoor upholstered furniture, their care/treatment products (DTSC 2019a). Also, DTSC is identifying the next round of proposed priority products and a draft technical document evaluating the state of the science (DTSC 2019b). The three products that were adopted as priority products are: (1) Spray polyurethane foam (SPF) systems that have unreacted diisocyanates; (2) Children’s foam padded sleeping products that has Tris (1,3-dichloro-2-propyl) phosphate (TDCPP); and (3) Paint and varnish strippers, and surface cleaners with methylene chloride, and proposed priority products include carpets and rugs, laundry detergents, graffiti removers, nail products etc., these products contain one or more PFAS that may harm humans or the environment (DTSC 2020). Multiple federal agencies and states and other countries are working to understand possible PFAS risks associated with consumer products and develop approaches to reduce exposure where possible. These studies will address an important need to better understand PFAS exposures resulting from indoor and consumer product related sources.

PFAS in House Dust

Strynar and Lindstrom (2008) reported that “perfluorinated compounds are present in-house dust at levels that may represent an important pathway for human exposure.” Ten PFAAs and 3 FTOHs in homes (n = 102) and day care centers (n = 10) were measured in OH and NC in 2000-2001. PFOS and PFOA were the most frequently detected, in over 95% of the samples at median concentrations of 201 and 142 ng/g, respectively. In one study, 20 PFAS in U.S., Canada, and the Czech Republic (n=41) were measured at lower concentrations where PFHxA and PFOS were the most frequently detected (Karaskova et al. 2016). In a small-scale study, using data collected by the Minnesota Department of Health, Scher et al. (2019) found that median concentrations of PFOA in entryways and living rooms of household were about 35 and 70 times respectively higher than soil concentrations suggesting interior sources of PFAS dust contributes predominantly. Wu et al. (2020) found that there exists a strong association between PFAS levels in carpet and dust pairs in 18 childcare centers in California and carpets can be a source as well as a sink for PFAS with mean summed PFAS concentrations of 471 ng/g while for dust it was 523 ng/g to which carpets also contributed. In another study of 37 neutral and ionic PFAS concentrations in eight childcare environments, Zheng et al. (2020) estimated the U.S. children’s exposure through dust and dermal absorption. Summed concentrations of 28 PFAS in dust ranged from 8.1 to 3700 ng/g with two neutral group PFAS (FTOH and FOSA/FOSE) dominating among all PFAS while ionic PFAS median concentrations were significantly at lower concentrations (median was 5.8 ng/g). Daily intake of neutral PFAS via dust ingestion was estimated at 0.2 ng/kg bw/day.

Karaskova et al. (2016) reported measured PFAS dust concentrations data from living rooms and bedrooms from 41 houses: 12 in the Czech Republic, 15 in Canada, and 14 in the U.S. PFHxA and PFOS were most frequently detected while in all countries, PFOA concentrations were detected at high concentrations. For these three countries median concentrations for PFOS were found to vary between 9.1 and 14.1 ng/g while for PFOA the range was 8.2 and 9.3 ng/g. The study also found that there was a shift from long- to short-chain PFAS in North American data. Winkens et al. (2018) analyzed PFAS of indoor floor dust from 65 children’s bedrooms in Finland. The results revealed that PFCA and PFOS existed in greater than half of the dust samples with PFOA having the highest median concentration of 5.26 ng/g. Moreover, polyfluoroalkyl phosphoric acid esters (PAPs) and FTOHs were the dominant groups of PFAS in dust with highest median concentrations of 53.9 ng/g and 45.7 ng/g for diPAP and FTOH, respectively.

PFAS in Indoor Air

Indoor air concentrations of PFAS may result from outdoor air or indoor sources and may be higher than outdoor air concentrations for certain PFAS. Concentrations range from 1-1000 pg/m³ for PFOA and PFOS, and 1000-10000 pg/m³ for volatile PFAS such as FTOHs (e.g., Winkens et al. 2016). Makey et al. (2017) suggests the importance of airborne PFAA precursors for PFAS exposures. House dust concentrations have been measured by the U.S. EPA and others at levels that may represent an important human exposure pathway. PFOS and PFOA were most detected, in over 95% of samples in OH and NC residences and these results are consistent with other studies. There is a paucity of recent U.S. data for emerging PFAS (McCord and Strynar 2019).

PFAS in Outdoor Air

PFAS have been detected in ambient air around the globe (Lindstrom et al. 2011) while there are very limited measurements in the U.S. There is growing concern over PFAS air releases due to long-range transport and widespread regional environmental contamination. The potential for air release is considerable across industrial production, application (e.g., chrome plating, carpet manufacturing, fabric coating), and disposal (wastewater treatment, landfill). Certain PFAS are commonly found in ambient air and elevated concentrations have been observed near emission sources (Rauert et al. 2018); thus, human exposures near emission sources still need further investigations (Lindstrom et al. 2011). PFOA and PFOS outdoor air concentrations are typically 1-10 pg/m³; however, they can be as high as 900000 pg/m³ near large manufacturing facilities (Shin et al. 2011). Atmospheric transport of PFAS may be a significant source of ecological exposure to soils and aquifers. However, for the U.S., very little information exists for PFAS ambient air concentrations. It may be possible to address this need by relying on archived PM samples using methods similar to that suggested by Lin et al. (2019).

Efforts by U.S. EPA are ongoing to incorporate PFAS information into atmospheric models to understand the potential for fate and transport of PFAS. The outcome from this effort should be very useful to model the fate and transport of PFAS in other media (e.g., water) as well as for biota exposure.

PFAS in Food

According to the Food and Drug Administration (FDA), some PFAS were authorized to be used as food contact substances for cookware, food packaging, and food processing equipment unless health concerns were identified. Also, FDA has detected PFAS in foods grown/produced from areas contaminated with PFAS or from general food supply but without indication of human health concerns (FDA 2020). Some published data are available for retail and raw milk samples, cranberries and associated irrigation water, seafood samples, and bottled water samples (Genualdi et al. 2017; FDA 2020). Limited data exists for some food packaging materials that contain PFAS (Schaider et al. 2017). Some international studies indicated food as the major exposure pathway for the general population (Gebbink et al. 2015; Vestergren and Cousins 2013). A European Food Safety Authority (EFSA) review reported that foods such as fish meat, fruit and fruit products, and eggs and egg products were the major contributors to human exposure based on food samples obtained from 16 European countries in 2007-2018 (EFSA 2020). However, improved analytical methods are needed for food items (Vestergren and Cousins 2013) to aid in exposure studies. Wide variability in detection and concentrations are found in different countries. For example, fish and other seafood appear to be food groups in which more PFAS are detected and where concentrations are higher (Domingo and Nadal 2017). In a study of 941 adults during 1996-1999 by Lin et al (2020), it was found that meat/fish/shellfish, low-fiber and high-fat bread/cereal/rice/pasta and coffee diet was associated with higher PFAS (e.g., PFNA) plasma concentrations while diet high in vegetables, fruits, and Omega-3 rich fish were associated with lower plasma concentrations of PFOS and 2-(N-methyl-perfluorooctane sulfonamido) acetic acid (MeFOSAA). Susmann et al. (2019) using the NHANES 2003-2014 serum PFAS and dietary recall data found that eating meals from fast food/pizza restaurants and other restaurants may be associated with higher serum PFAS concentrations. Furthermore, eating popcorn was also associated with higher serum levels of four PFAS, PFOA, PFNA, PFDA, and PFOS. In 2020, the United States Department of Agriculture’s Food Safety and Inspection Service (USDA’s FSIS) will perform sampling to determine PFAS contamination in FSIS-regulated products (USDA 2020). Reports found PFOA and PFOS in muscle (≤ 20 ppb) and serum (≤ 125 ppb) in cattle grazing on a contaminated site in Alabama (Washington et al. 2010). The presence of PFAS in agricultural products indicates a need for such surveillance for food products coming out of contaminated regions.

Food packaging materials containing PFAS can include: some grease-resistant paper, fast food containers/wrappers, microwave popcorn bags, pizza boxes, and candy wrappers. Schaider et al. (2017) found that fluorinated compounds are common in food contact papers and other fast food packaging in the U.S. While much of the U.S. production of PFOS and PFOA was phased out between 2000 and 2015, these compounds are still produced in other regions of the world. In another research that provided potential evidence of exposure to PFAS through food packaging and dental floss, Boronow et al. (2019) found that PFAS in the blood of 178 California women (87 African American and 91 non-Hispanic white) differed by race with African American women having lower serum PFOA and PFHxS concentrations. A toxicity profiles review by the Agency for Toxic Substances and Disease Registry (ATSDR) reported that perfluoroalkyls have been detected in human breast milk and umbilical cord blood in multiple studies (ATSDR 2018). Vestergren et al. (2012) noted some PFAS detection limit issues in food and the difficulties of accurate measurements of PFAS in food; the research suggested that previous dietary exposures may be overestimated by an order of magnitude, and the difficulties of accurate PFAS measurements in food may be a source of uncertainty in exposure estimates.

PFAS in Fish Tissue

Fish consumption may be a source of human exposure to PFOS or long-chain perfluorocarboxylic acids. Fish in high on the food chain can accumulate high PFAS concentrations. Studies have found that PFOS, in particular, can build up (bioaccumulate) in fish. Fair et al. (2019) measured 11 PFAS in fish tissues collected from Charleston, South Carolina (SC), mean total PFAS concentrations were in the range of 12.7-33.0 ng/g wet weight in whole fish with PFOS as the dominant compounds, the results implicated that the consumption of local wild fish may cause health risks.

Moreover, Guillette et al. (2020) measured the serum concentrations of 23 PFAS in Striped Bass from the Cape Fear River as well as in “reference” Striped Bass raised in well-water under controlled aquaculture conditions; the comparison between the two Bass groups indicated the significantly higher serum PFAS concentrations in Cape Fear River Striped Bass serum with PFOA being the dominant compounds (as high as 977 ng/mL); this research also revealed the association between high PFAS concentrations of Cape Fear River Striped Bass and biomarkers of altered immune and liver function. In another study, based on the analyses of fish tissue collected between July 2016 and November 2016, Goodrow et al. (2020) found that PFOS concentrations in nearly all fish species in New Jersey area were high enough for fish consumption advisories (1.5-126.1 ng/g).

PFAS Heterogeneity in Ethnic Groups

Park et al. (2019) studied the roles of geographic location and ethnicity in PFAS exposure, this research collected data from 1302 women aged 45-56 years within U.S. and it was found that white women had higher PFOA serum concentrations compared to Chinese and African women. However Chinese and Japanese women from two different cities had higher PFNA concentrations compared to Caucasian women in those respective locations. These results indicated that PFAS exposure is highly heterogeneous among different populations. In an analysis of PFOA serum concentrations of 1030 NJ residents, Yu et al (2020) found that sex and age stratifications indicated differences in PFOA, PFNA, PFHxS, and PFOS with significantly lower concentrations in younger females (20-59) compared to older females (60-74) and males (20-74).

The U.S. EPA PFAS Action Plan

Current scientific efforts

The U.S. EPA is currently advancing the scientific foundation to understand and manage PFAS risks. A variety of research is in progress and these efforts, and short- and long-term plans are detailed in the U.S. EPA’s PFAS Action Plan (USEPA 2019b). For full details, readers are referred to USEPA (2019b; 2020a). The exposure-relevant aspects of the Action Plan appropriately focus on the development of robust methods of measurement. The development of methods is both important and made difficult by the large number of chemical forms of PFAS present and the wide range of environmental media that are contaminated. Due to their regulatory and public health significance, methods related to drinking water, food, air, soil, and sediment are priority considerations. The challenge of the potential for there to be a large number of PFAS yet to be identified in the environment is being addressed by the development and application of non-targeted analysis methods (McCord and Strynar 2019).

The cooperation between various federal agencies to collect and share PFAS information has been critically important and thus U.S. EPA is coordinating with different research teams to provide consistent and accurate information on PFAS for policy making. The information of PFAS sources and concentrations provided by the U.S. EPA are accessible online for stakeholders and local governing bodies to manage PFAS while the integration of available PFAS information can help state, tribes, and communities to easily use PFAS data for decision making (USEPA 2019b).

Furthermore, more information regarding potential exposure pathways of PFAS in different exposure scenarios can help people to take proper actions to minimize PFAS risks in the future (USEPA 2019b).

Future plans

In the long-term plans, the U.S. EPA will be collecting more data regarding PFAS released into the environment and at the same time trying to help decrease the releases of PFAS into environments such as waterbodies. The PFAS occurrence in drinking water will be more rigorously monitored and more information on PFAS concentrations in U.S. drinking water for a broader type of PFAS will be collected. Moreover, the PFAS data will be easier to access and utilize in the future. The U.S. EPA plans to develop a PFAS data inventory and share PFAS monitoring data in various environmental media such as soil, air, and water (USEPA 2019b). The U.S. EPA also plans to get an in-depth understanding of atmospheric fate and transport of PFAS using available atmospheric models (e.g., the Community Multi-scale Air Quality Model); the PFAS information will be incorporated into the U.S. EPA atmospheric models as these results may help to understand the potential PFAS exposure pathways and provide guidance to take appropriate actions for PFAS regulations (USEPA 2019b).

EPA will continue to develop and refine its non-targeted analysis methods with applications to a wider range of environmental media and encompassing a larger chemical space. Tremendous investments and advances are being made in advanced instrumentation technology (e.g., chromatography and mass spectrometry, workflow processes, and computational capabilities (McCord and Strynar 2019; Washington et al. 2020). In the future, U.S. EPA also plans to provide more information regarding a) PFAS bioaccumulation in organisms and food chains, b) species more vulnerable to PFAS exposure, and c) may determine the thresholds values for ecological toxicity (USEPA 2019b). The future collaboration between the U.S. EPA and other agencies such as the USGS, United States Army Corps of Engineers (USACE), United States Department of Defense (DoD), CDC, ATSDR, FDA, National Institute of Standards and Technology (NIST), and universities provide the basis for PFAS exposure science investigation (USEPA 2019b). The collaboration between the U.S. EPA and other agencies will also facilitate development of an exposure communication toolbox that contains multi-media PFAS information for federal, state, tribal, and local partners to share with the public of concern (USEPA 2019b). In the future, the U.S. EPA will also prioritize the PFAS research in the agriculture, provide help for agricultural community and to help ensure safe food supplies (USEPA 2020a). It is anticipated that future studies will focus more on advanced and sensitive PFAS measurement techniques as it is critical to obtain high quality PFAS measurement data.

Discussion

PFAS can co-exist in multiple environmental media and are an issue globally causing heterogenous ecological and human health issues to varying degrees. PFAS exposure is an occupational and residential issue, and a multimedia issue with important public health implications. PFAS are in human blood while exposure magnitudes and relative pathway contributions depend on population, scenario, proximity to local sources as well as remote sources. Many data/information gaps exist, in particular for emerging PFAS. Activities and needs for addressing critical data gaps include the following: literature reviews/data analyses to identify and characterize PFAS sources, media concentrations, relative contributions to human exposure; field and laboratory methods development (e.g., non-targeted analyses); field research to address data gaps, involving collaborations among federal agencies and states; data visualization and geospatial analysis of PFAS concentrations and exposures; development and application of modeling tools.

Though regional and global models exist to characterize fate and transport of PFAS in the environment, lack of various accurate input data (e.g., detailed emission sources) may pose some challenges to the utility of these model outputs for use in exposure modeling and analyses. Thus, understanding of relative source contributions for vulnerable populations and different exposure scenarios is very limited. However, ongoing efforts at the U.S. EPA to make use of the best available PFAS data to model fate and transport of PFAS are likely to help fill in the gaps in environmental measurements. The small number of multimedia modeling papers in the literature indicated that ingestion of food (“background”), drinking water (“contaminated” scenarios), house dust (for children) are main pathways for PFOS/PFOA; however, there are many uncertainties in the modeling efforts to date. Although some multimedia exposure modeling analyses on existing PFAS and precursors have been conducted (mostly international vs. U.S.), there are many uncertainties. Data gaps including environmental concentrations in water, soil, dust, air, and food as well as human activity patterns related to PFAS exposures need to be addressed for additional modeling analyses, including sensitivity and uncertainty analyses. As an interim measure, it is important to consider the underlying concentration data and exposure factors used in the different publications from different countries, as well as the underlying assumptions. It is anticipated that more modeling efforts will be in place for precursors of PFAS, existing PFAS, and emerging PFAS, a research area that will be very useful to inform risk management decisions and focus future data collection efforts.

Perhaps in the U.S. the only town that had comprehensive measurements of PFOA across the media for several years, human exposure information along with measurements of PFOA in serum, and epidemiological studies is Parkersburg, WV (Shin et al. 2011). It is hoped that the on-going inter-agency efforts will result in additional sets of comprehensive measurements at other PFAS manufacturing/impacted sites across the U.S.

It is also anticipated that PFAS exposure-related methods development, data collection, Geographic Information System (GIS) mapping, fate and transport modeling, and human exposure and dose modeling will help inform risk management/communication and help focus future field studies, and these will be linked with risk assessment, health science, and toxicity testing efforts. Research is ongoing and needed for detecting/analyzing emerging PFAS.

Some of these gaps are being addressed by the U.S. EPA and other federal agencies. Determination of toxicity of about 150 PFAS, new methods to identify PFAS in the environment, estimation of emission sources, systems-based approaches and coordination are in the works by various federal agencies to prioritize and inform focused data collection efforts; fate and transport of PFAS in the environment; health-based risk standards/guidelines/advisories; and risk assessments/communication/ mitigation efforts (USEPA 2020e).