JHH f United States
Environmental Protection
R mAgency
EPA-815-R-24-004
FINAL
Maximum Contaminant Level Goals (MCLGs) for Three Individual
Per- and Polyfluoroalkyl Substances (PFAS) and a
Mixture of Four PFAS
Individual MCLGs for Three Per- and Polyfluoroalkyl Substances (PFAS):
• HFPO-DA
• PFNA
• PFHxS
Mixture MCLG for Mixtures of Four PFAS:
• HFPO-DA
• PFNA
• PFHxS
• PFBS
-------
FINAL
2024
Maximum Contaminant Level Goals (MCLGs) for Three Individual Per- and
Polyfluoroalkyl Substances (PFAS) and a Mixture of Four PFAS
Prepared by:
U.S. Environmental Protection Agency
Office of Water (4304T)
Office of Science and Technology
Health and Ecological Criteria Division
Washington, DC 20460
EPA Document Number: EPA-815-R-24-004
April 2024
-------
FINAL
2024
Notices
This document has been reviewed in accordance with EPA policy and approved for publication.
This document provides a summary of information used to develop the final individual MCLGs
for HFPO-DA1, PFNA, and PFHxS and a final MCLG for mixtures of HFPO-DA, PFNA,
PFHxS, and/or PFBS. Molecular formulas and Chemical Abstract Service registry numbers
(CASRNs) for these four PFAS are as follows:
• HFPO-DA (CeFnOf; CASRN 122499-17-6)
• PFNA (C9F17CO2"; CASRN 72007-68-2)
• PFHxS (C6F13SO3-; CASRN 108427-53-8)
• PFBS (C4F9SO3-; CASRN 45187-15-3)
These PFAS may exist in multiple forms, such as isomers or associated salts, and each form may
have a separate CASRN or no CASRN at all. Additionally, these compounds have various names
under different classification systems. However, at environmentally relevant pHs, these PFAS
are expected to dissociate in water to their anionic (negatively charged) forms. For instance,
HFPO-DA is an anionic molecule which has an ammonium salt (CASRN 62037-80-3), a
conjugate acid (CASRN 13252-13-6), a potassium salt (CASRN 67118-55-2), and an acyl
fluoride precursor (CASRN 2062-98-8), among other variations. At environmentally relevant
pHs these all dissociate into the propanoate/anion form (CASRN 122499-17-6). Each PFAS
listed has multiple variants with differing chemical connectivity, but the same molecular
composition (known as isomers). Commonly, the isomeric composition of PFAS is categorized
as 'linear,' consisting of an unbranched alkyl chain, or 'branched,' encompassing a potentially
diverse group of molecules including at least one, but potentially more, offshoots from the linear
molecule. While broadly similar, isomeric molecules may have differences in chemical
properties. The final National Primary Drinking Water Regulation for PFAS covers all salts,
isomers, precursors, and derivatives of the chemicals listed, including derivatives other than the
anionic form which might be created or identified.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
1-The EPA notes that HFPO-DA is used in a processing aid technology developed by DuPont to make fluoropolymers without
using perfluorooctanoic acid (PFOA). The chemicals associated with this process are commonly known as GenX Chemicals and
the term is often used interchangeably for HFPO-DA.
1
-------
FINAL
2024
Authors, Contributors, and Reviewers
Prepared by
Carlye Austin, Ph.D., DABT
Czarina Cooper, M.P.H.
Colleen Flaherty, M.S.
Brittany Jacobs, Ph.D.
Casey Lindberg, Ph.D.
Contributors
Brandi Echols, Ph.D.
Alexis Lan, M.P.H.
EPA Technical Review
Office of Chemical Safety and Pollution Prevention
Office of Children's Health Protection
Office of Land and Emergency Management
Office of Research and Development
Executive Direction
Elizabeth Behl
Eric Burneson, P.E.
Formatting by
Tetra Tech, Inc
ICF
11
-------
FINAL 2024
Contents
Figures vii
Tables viii
Abbreviations and Acronyms x
1 Introduction and Background 1-1
1.1 Purpose 1-1
1.2 Occurrence and Co-Occurrence of PFAS in Drinking Water 1-2
1.3 Dose Additivity for PFAS Mixtures 1-2
1.3.1 Overview of Scientific Support 1-2
1.3.1.1 Science Advisory Board Support 1-5
1.3.2 Toxicological Similarity of PFHxS, PFNA, HFPO-DA, and PFBS 1-5
1.3.2.1 Background on Concept of "Toxicological Similarity" 1-5
1.3.2.2 Overview of Scientific Support 1-6
1.3.2.3 Science Advisory Board Support 1-10
1.3.3 Summary 1-11
1.4 General Hazard Index (HI) Approach for PFAS Mixtures 1-11
1.4.1 B ackground/Overvi ew 1-11
1.4.2 Consideration of Mixtures Assessment Approaches and Selection of General HI
Approach 1-12
1.4.2.1 Science Advisory Board Support 1-16
1.5 Establishment of Individual MCLGs for PFHxS, HFPO-DA, PFNA, and/or PFBS.. 1-17
1.6 Overview of Individual MCLG and Mixture Hazard Index (HI) MCLG Approaches... 1-
18
2 Calculating the Health-Based Water Concentrations for HFPO-DA, PFBS, PFNA, and
PFHxS for the HI MCLG 2-1
2.1 HFPO-DA 2-1
2.1.1 Toxicity 2-1
2.1.2 Exposure Factor 2-2
2.1.3 Relative Source Contribution 2-3
2.1.4 Derivation of HFPO-DA HBWC 2-3
2.2 PFBS 2-4
2.2.1 Toxicity 2-4
2.2.2 Exposure Factor 2-5
2.2.3 Relative Source Contribution 2-6
2.2.4 Derivation of PFBS HBWC 2-6
2.3 PFNA 2-7
2.3.1 Toxicity 2-8
iii
-------
FINAL 2024
2.3.2 Exposure Factor 2-8
2.3.3 Relative Source Contribution 2-9
2.3.4 Derivation of PFNA HBWC 2-9
2.4 PFHxS 2-10
2.4.1 Toxicity 2-11
2.4.2 Exposure Factor 2-12
2.4.3 Relative Source Contribution 2-12
2.4.4 Derivation of PFHxS HBWC 2-12
3 Derivation of MCLGs 3-1
3.1 Individual MCLGs for HFPO-DA, PFNA, and PFHxS 3-1
3.2 PFAS Mixtures Hazard Index MCLG 3-1
4 References 4-1
Appendix A. HFPO-DA: Summary of Occurrence in Water and Detailed Relative Source
Contribution A-l
A. 1. Occurrence in Water A-l
A. 1.1. Ground Water A-l
A. 1.2. Surface Water A-l
A.2. RSC for HFPO-DA, Literature Search and Screening Methodology A-8
A.2.1. Literature Search and Screening A-9
A.2.2. Additional Screening A-10
A.3. Summary of Potential Exposure Sources of HFPO-DA Other than Water A-l 1
A.3.1.Dietary Sources A-11
A.3.2. Food Contact Materials A-12
A.3.3. Consumer Products A-12
A.3.4. Indoor Dust A-12
A.3.5. Air A-13
A.3.6. Soil A-13
A.3.7. Sediment A-14
A.4. Recommended RSC A-15
Appendix B. PFBS: Summary of Occurrence in Water and Detailed Relative Source
Contribution B-l
B.l. Occurrence in Water B-l
B. 1.1. Ground Water B-l
B.l.2. Surface Water B-6
B.2. RSC for PFBS, Literature Search and Screening Methodology B-15
B.2.1. Literature Search and Screening B-16
B.2.2. Additional Screening B-18
B.3. Summary of Potential Exposure Sources of PFBS Other than Water B-19
B.3.1. Food B-19
iv
-------
FINAL 2024
B.3.2. Food Contact Materials B-27
B.3.3. Consumer Products B-28
B.3.4. Indoor Dust B-30
B.3.5. Air B-34
B.3.5.1. Indoor Air B-34
B.3.5.2. Ambient Air B-34
B.3.6. Soil B-35
B.4. Recommended RSC B-38
Appendix C. PFNA: Summary of Occurrence in Water and Detailed Relative Source
Contribution C-l
C.l. Occurrence in Water C-l
C. 1.1. Groundwater C-l
C.l.2. Surface Water C-14
C.2. RSC for PFNA, Literature Search and Screening Methodology C-34
C.2.1. Literature Search and Screening C-35
C.2.2. Additional Screening C-36
C.3. Summary of Potential Exposure Sources of PFNA Other than Water C-37
C.3.1. Dietary Sources C-37
C.3.1.1. Seafood C-37
C.3.1.2. Other Food Types C-50
C.3.2. Food Contact Materials C-74
C.3.3. Consumer Products C-78
C.3.4. Indoor Dust C-85
( .3.5. Air C-93
C.3.5.1. Indoor Air C-93
C.3.5.2. Ambient Air C-95
C.3.6. Soil C-99
C.3.7. Sediment C-l 12
C.4. Recommended RSC C-l 12
Appendix D. PFHxS: Summary of Occurrence in Water and Detailed Relative Source
Contribution D-l
D. 1. Occurrence in Water D-l
D. 1.1. Groundwater D-l
D. 1.2. Surface Water D-l3
D.2. RSC for PFHxS, Literature Search and Screening Methodology D-33
D.2.1. Literature Search and Screening D-34
D.2.2. Additional Screening D-35
D.3. Summary of Potential Exposure Sources of PFHxS Other than Water D-36
D.3.1.Dietary Sources D-36
v
-------
FINAL
2024
D.3.1.1. Seafood D-36
4.1.1.1 Other Food Sources D-48
D.3.2. Food Contact Materials D-78
D.3.3. Consumer Products D-82
D.3.4. Indoor Dust D-89
D.3.5. Air D-98
D.3.5.1. Indoor Air D-98
D.3.5.2. Ambient Air D-102
D.3.6. Soil I)-106
D.3.7. Sediment D-119
D.4. Recommended RSC D-119
vi
-------
FINAL 2024
Figures
Figure 1-1. Flow chart for evaluating chemical mixtures using component-based additive
methods. (Reproduction of Figure 2-1 from USEPA, 2024a) 1-6
vii
-------
FINAL
2024
Tables
Table 1-1. Affected health outcomes in animal toxicological and/or epidemiological studies
for the four PFAS included in the HI MCLG (adapted from Table 6-7 in
USEPA, 2024d) 1-7
Table 1-2. Affected health endpoints in animal toxicity studies for the four PFAS included
in the HI MCLG (adapted from Table 4 in US OSTP, 2023) 1-8
Table 1-3. Specific Endpoints Affected by One or More of the Four PFAS Included in the
HI MCLG 1-8
Table 2-1. EPA Exposure Factors for Drinking Water Intake for Different Candidate
Sensitive Populations or Life Stages, Based on the Critical Effect and Study for
HFPO-DA 2-2
Table 2-2. HFPO-DA HBWC - Input Parameters and Value 2-4
Table 2-3. EPA Exposure Factors for Drinking Water Intake for Different Candidate
Sensitive Populations or Life Stages, Based on the Critical Effect and Study for
PFBS 2-6
Table 2-4. PFBS HBWC - Input Parameters and Value 2-7
Table 2-5. EPA Exposure Factors for Drinking Water Intake for Different Candidate
Sensitive Populations and Life Stages, Based on the Critical Effect and Study
for PFNA 2-9
Table 2-6. PFNA HBWC - Input Parameters and Value 2-10
Table 2-7. PFHxS HBWC - Input Parameters and Value 2-13
Table 3-1. Individual MCLGs 3-1
Table A-l. Compilation of Studies Describing HFPO-DA Occurrence in Surface Water A-4
Table A-2. Populations, Exposures, Comparators, and Outcomes (PECO) Criteria A-10
Table B-l. Compilation of Studies Describing PFBS Occurrence in Groundwater B-2
Table B-2. Compilation of Studies Describing PFBS Occurrence in Surface Water B-8
Table B-3. Populations, Exposures, Comparators, and Outcomes (PECO) Criteria B-l8
Table B-4. Compilation of Studies Describing PFBS Occurrence in Food B-22
Table B-5. Compilation of Studies Describing PFBS Occurrence in Indoor Dust B-31
Table B-6. Compilation of Studies Describing PFBS Occurrence in Soil B-36
Table C-l. Summary of Studies Reporting the Occurrence of PFNA in Groundwater C-5
Table C-2. Summary of Studies Reporting the Occurrence of PFNA in Surface Water C-19
Table C-3. Populations, Exposures, Comparators, and Outcomes (PECO) Criteria C-36
Table C-4. Summary of PFNA Data in Seafood C-41
Table C-5. Summary of PFNA Data in Other Food C-54
-------
FINAL 2024
Table C-6. Summary of Studies Reporting the Occurrence of PFNA in Food Contact
Materials C-75
Table C-7. Summary of PFNA Consumer Product Data C-80
Table C-8. Summary of PFNA Indoor Dust Data C-88
Table C-9. Summary of Studies Reporting the Occurrence of PFNA in Indoor Air C-94
Table C-10. Summary of Peer-Reviewed Studies Reporting the Occurrence of PFNA in
Ambient Air C-96
Table C-ll. Summary ofPFNADatain Soil C-102
Table D-l. Studies Reporting Occurrence of PFHxS in Groundwater D-3
Table D-2. Studies Reporting PFHxS Occurrence in Surface Water D-18
Table D-3. Populations, Exposures, Comparators, and Outcomes (PECO) Criteria D-35
Table D-4. Summary of PFHxS Data in Seafood D-39
Table D-5. Summary of PFHxS Data in Other Food D-51
Table D-6. Studies Reporting PFHxS Occurrence in Food Contact Materials D-80
Table D-7. Summary of PFHxS Consumer Product Data D-85
Table D-8. Summary of PFHxS Indoor Dust Data D-92
Table D-9. Summary of PFHxS in Indoor Air D-100
Table D-10. Summary of the Occurrence of PFHxS in Ambient Air D-103
Table D-l 1. Summary of PFHxS Data in Soil D-l09
IX
-------
FINAL
2024
Abbreviations and Acronyms
AFFF
aqueous film-forming
FCM
food contact materials
foam
FDA
U.S. Food and Drug
AMAP
Arctic Monitoring and
Administration
Assessment Programme
FIFRA
Federal Insecticide,
AOP
adverse outcome pathway
Fungicide, and
AT SDR
Agency for Toxic
Rodenticide Act
Substances and Disease
GCA
groundwater
Registry
contamination area
BAF
bioaccumulation factor
GenX chemicals
hexafluoropropy 1 ene
BDL
below the detection limit
oxide dimer acid (HFPO-
BMD
benchmark dose
DA) and HFPO-DA
CASRN
Chemical Abstract
ammonium salt
Service registry number
GD
gestational day
CDC
Centers for Disease
g/L
grams per liter
Control and Prevention
HA
health advisory
CERCLA
Comprehensive
Environmental Response,
HBWC
health-based water
concentration
Compensation, and
HED
human equivalent dose
Liability Act
HDPE
high-density polyethylene
CTEPP
Children's Total
Exposure to Persistent
HFPO
hexafluoropropy 1 ene
oxide
Pesticides and Other
HI
hazard index
Persistent Organic
Pollutants
HQ
hazard quotient
DA
dose addition
IA
integrated addition
IRIS
Integrated Risk
Information System
DF
detection frequency
DWI-BW
body weight-adjusted
drinking water intake
K+PFBS
potassium
perfluorobutane sulfonate
DWTP
drinking water treatment
plant
L/kg/day
liters per kilogram body
weight per day
DWR
durable water repellent
LAS
land application site
E
duration-relevant
LB
lower bound
exposure
LOAEL
lowest-observed-adverse-
EEE
electrical and electronic
effect level
equipment
LOD
limit of detection
EFSA
European Food Safety
Authority
U.S. Environmental
Protection Agency
LOQ
limit of quantitation
EPA
MAMA
Methods Advancement
for Milk Analysis
FAO
Food and Agriculture
Organization Area
MCLG
Maximum Contaminant
Level Goal
MDL
method detection limit
x
-------
FINAL
2024
MF
modifying factor
PECO
mg/kg/day
milligrams per kilogram
body weight per day
mg/L
milligrams per liter
PFAA
MOA
mode of action
PFAS
MRL
minimal risk level
mRNA
messenger ribonucleic
acid
PFBS
MSW
municipal solid waste
PFC
MQL
method quantification
limit
PFCA
NAS
National Academy of
PFHpA
Sciences
PFHxA
NCDHHS
North Carolina
Department of Health and
PFHxS
Human Services
PFNA
NCDEQ
North Carolina
PFOA
Department of
PFOS
Environmental Quality
ng/g dw
naanograms per gram dry
Pg/g
weight
Pg/L
ng/kg
nanograms per kilogram
pg/m3
ng/L
nanograms per liter
NHANES
National Health and
PHG
Nutrition Examination
PND
Survey
POD
NO A A
National Oceanic and
POP
Atmospheric
Administration
PPARa
NOAEL
no-ob served-adverse-
effect level
ppt
PSA
NPDWR
National Primary
Drinking Water
Regulation
PWS
NRSA
National Rivers and
RA
Streams Assessment
RfD
NTP
National Toxicology
RfV
Program
RPF
OECD
Organisation for
Economic Co-operation
RSC
and Development
SAB
ORD
Office of Research and
Development
SDWA
Population, Exposure,
Comparator, and
Outcome
perfluoroalkyl acids
per- and polyfluoroalkyl
substances
perfluorobutanesulfonic
acid
perfluorochemi cal s
perfluoroalkyl carboxylic
acids
perfluoroheptanoic acid
perfluorohexanoic acid
perfluorohexanesulfonic
acid
perfluorononanoic acid
perfluorooctanoic acid
perfluorooctanesulfonic
acid
picograms per gram
picograms per liter
picograms per cubic
meter
provisional health goal
postnatal day
point of departure
persistent organic
pollutants
peroxisome proliferator-
activated receptor alpha
parts per trillion
prostate-specific antigen
public water system
response addition
reference dose
reference value
relative potency factor
relative source
contribution
Science Advisory Board
Safe Drinking Water Act
XI
-------
FINAL
SLEA
Screening Level
Exposure Assessment
SPM
suspended particulate
matter
TOFMS
time-of-flight mass
spectrometry
TOSHI
target-organ-specific
hazard indices
TRI
Toxics Release Inventory
TSCATS
Toxic Substances Control
Act Test Submissions
TTD
target organ toxicity dose
UB
upper bound
UC MR
Unregulated Contaminant
Monitoring Rule
UCMR3
third Unregulated
Contaminant Monitoring
Rule
UF
uncertainty factor
2024
UFa
interspecies uncertainty
factor
UFd
database uncertainty
factor
UFh
human interindividual
variability uncertainty
factor
UFs
extrapolation from
sub chroni c-to-chroni c
exposure duration
uncertainty factor
Mg/kg
micrograms per kilogram
^g/L
micrograms per liter
|ig/m2
micrograms per square
meter
wos
Web of Science
WWTP
wastewater treatment
plant
Xll
-------
FINAL
2024
1 Introduction and Background
1.1 Purpose
Section 1412(a)(3) of the Safe Drinking Water Act (SDWA) requires the Administrator of the
U.S. Environmental Protection Agency (EPA) to propose a Maximum Contaminant Level Goal
(MCLG) simultaneously with the National Primary Drinking Water Regulation (NPDWR). The
MCLG is set, as defined in Section 1412(b)(4)(A), at "the level at which no known or anticipated
adverse effects on the health of persons occur and which allows an adequate margin of safety."
The MCLG incorporates a margin of safety to reflect scientific uncertainty and, in some cases,
the particular susceptibility of some groups (e.g., children) within the general population.
Consistent with SDWA 1412(b)(3)(C)(i)(V), in developing the MCLG, the EPA considers "the
effects of the contaminant on the general population and on groups within the general population
such as infants, children, pregnant women, the elderly, individuals with a history of serious
illness, or other subpopulations that are identified as likely to be at greater risk of adverse health
effects due to exposure to contaminants in drinking water than the general population." Other
factors considered in determining MCLGs include health effects data for drinking water
contaminants and potential sources of exposure other than drinking water. MCLGs are not
regulatory levels and are not enforceable.
The purpose of this document is to provide a summary of the health effects and exposure
information and analyses and to describe the derivation of the EPA's final MCLGs for the
following per- and polyfluoroalkyl substances (PFAS), for which the EPA is finalizing a
NPDWR: hexafluoropropylene oxide dimer acid (HFPO-DA) (also known as GenX chemicals)2,
perfluorononanoic acid (PFNA), perfluorohexanesulfonic acid (PFHxS), and perfluorobutane
sulfonic acid (PFBS).3 The EPA is finalizing individual MCLGs for HFPO-DA, PFNA, and
PFHxS. The EPA is also finalizing a PFAS mixture MCLG for mixtures of two or more of four
PFAS—HFPO-DA, PFNA, PFHxS, and PFBS—that accounts for dose-additive health effects
when these PFAS co-occur in drinking water. The PFAS mixture MCLG is based on a hazard
index (HI) approach, a commonly used component-based mixtures risk assessment method (see
Section 1.4 and USEPA, 2024a). This document summarizes key elements (e.g., reference doses
(RfDs)) from recently published, peer-reviewed, publicly available human health toxicity
assessments for HFPO-DA (USEPA, 2021c), PFBS (USEPA, 2021d), PFNA (ATSDR, 2021),
and PFHxS (ATSDR, 2021) that the EPA used to develop MCLGs for HFPO-DA, PFNA, and
PFHxS and an MCLG for mixtures of two or more of these PFAS plus PFBS. The MCLG
represents the level below which adverse health effects over a lifetime of exposure are not
expected to occur, including for sensitive populations and life stages, and with an adequate
margin of safety. This document is not intended to be an exhaustive description of all health
effects or modeled endpoints (i.e., human health toxicity assessment) nor is it a drinking water
health advisory (HA).
2The EPA notes that HFPO-DA is used in a processing aid technology developed by DuPont to make fluoropolymers without
using PFOA. The chemicals associated with this process are commonly known as GenX chemicals and the term is often used
interchangeably for HFPO-DA along with its ammonium salt.
3Note: The EPA is also finalizing individual MCLGs for two other PFAS: perfluorooctanoic acid (PFOA) and
perfluorooctanesulfonic acid (PFOS) (see USEPA, 2024c)).
1-1
-------
FINAL
2024
1.2 Occurrence and Co-Occurrence of PFAS in Drinking Water
Improved analytical monitoring and detection methods have enabled detection of the occurrence
and co-occurrence of multiple PFAS in drinking water, ambient surface waters, aquatic
organisms, and other environmental media (see Appendices A through D and USEPA, 2024d).
The two PFAS perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS) have
historically been target analytes, and the focus of many environmental monitoring studies. More
recent monitoring studies, however, have focused on additional PFAS via advanced analytical
instruments/methods and nontargeted analysis (De Silva et al., 2021; McCord et al., 2020;
McCord and Strynar, 2019).
The EPA uses the Unregulated Contaminant Monitoring Rule (UCMR) to collect occurrence
data nationwide for contaminants that are suspected to be present in drinking water. Between
2013 and 2015, the EPA's third UCMR (UCMR 3) required all large public water systems
(PWSs) (each serving more than 10,000 people) and a statistically selected, nationally
representative sample of 800 small PWSs (each serving 10,000 people or fewer) to monitor for
30 unregulated contaminants in drinking water, including PFNA, PFHxS, and PFBS. In addition
to the UCMR 3 data collection, many states have undertaken more recent efforts to monitor for
PFAS in both source and finished drinking water using newer analytical methods and reflecting
lower reporting limits than those in UCMR 3. These results and other peer-reviewed studies
show continued PFAS occurrence and co-occurrence in multiple geographic locations (USEPA,
2024b; Cadwallader et al., 2022). These data also show certain PFAS (including PFNA, PFHxS,
and PFBS) measured at lower concentrations and significantly greater frequencies than were
measured under UCMR 3. Additionally, these state monitoring data include results for HFPO-
DA (which was not included in the suite of PFAS analyzed in UCMR 3) and demonstrate HFPO-
DA occurrence (and co-occurrence with other PFAS) in drinking water. From 2023-2025,
monitoring data for 29 PFAS including HFPO-DA, PFBS, PFNA, and PFHxS are being
collected under UCMR 5. These drinking water occurrence and co-occurrence data for HFPO-
DA, PFNA, PFHxS, PFBS, as well as additional PFAS are detailed in the EPA's PFAS
Occurrence and Contaminant Background Support Document for the Final PFAS NPDWR
(USEPA, 2024b).
1.3 Dose Additivity for PFAS Mixtures
1.3.1 Overview of Scientific Support
Dose additivity means that when two or more chemicals (in this case, PFHxS, PFNA, HFPO-
DA, and/or PFBS) exist in one mixture, the risk of adverse health effects following exposure to
the mixture is equal to the sum of the individual doses or concentrations scaled for potency
(USEPA, 2000a). Studies with PFAS and other classes of chemicals support the health-protective
conclusion that toxicologically similar chemicals (i.e., those that elicit similar observed adverse
effects following individual exposure, even if at different exposure levels) should be assumed to
act in a dose-additive manner when present in a mixture unless data demonstrate otherwise.
Experimental data demonstrate that PFAS elicit similar adverse health effects on several of the
same biological systems and functions including thyroid hormone signaling, lipid synthesis and
metabolism, development, and immune and liver function. Thus, exposure to these PFAS, at
doses that individually would not likely result in adverse health effects, when combined in a
mixture may pose health risks.
1-2
-------
FINAL
2024
Numerous published studies across multiple chemical classes, biological effects, and study
designs support a dose-additive mixture assessment approach for PFAS because they
demonstrate that experimentally observed responses to exposure to PFAS mixtures and other
chemical mixtures are consistent with modeled predictions of dose additivity (see the EPA's
Framework for Estimating Noncancer Health Risks Associated with Mixtures of Per- and
Polyfluoroalkyl Substances (PFAS) (hereafter "PFAS Mixtures Framework;" (USEPA, 2024a)).
Since the EPA's draft PFAS Mixtures Framework underwent SAB review in 2021-2022, new
studies from the EPA and others have provided robust evidence of combined toxicity of PFAS in
mixtures, corroborating and confirming earlier findings (USEPA, 2024a; e.g., Conley et al.,
2023; Conley et al., 2022). Additionally, the National Academies of Sciences, Engineering, and
Medicine (NASEM, 2022) recently recommended that clinicians apply an additive approach for
evaluating patient levels of PFAS currently measured in the National Health and Nutrition
Examination Survey (NHANES) in order to protect human health from additive effects from
PFAS co-exposure.
Data from in vivo studies that rigorously tested accuracy of Dose Additivity (DA), Integrated
Addition (IA), and Response Additivity (RA) model predictions of mixtures with components
that disrupted the same pathways (i.e., were toxicologically similar) demonstrated that DA
models provided predictions that were better than or equal to IA and RA predictions of the
observed mixture effects (Section 3.2 in USEPA, 2024a). In some circumstances the different
additivity models provide highly similar predictions of mixture effects and thus are essentially
equally effective. In situations where the models provide very different predictions, experimental
data have demonstrated that DA-based models consistently provide more accurate predictions of
observed mixture effects than RA or IA. This strongly supports the use of dose additivity as the
default method for estimating mixture effects of compounds that are toxicologically similar. The
National Academy of Sciences (NAS) conclusions on phthalates (and related chemicals) (NRC,
2008) and systematic reviews of the published literature (USEPA, 2024a; Martin et al., 2021;
Boobis et al., 2011) support dose additivity as the default model for estimating mixture effects in
some circumstances, even when the mixtures included chemicals with diverse MO As (but the
same target organs/effects). Systematic reviews of mixture studies with chemical classes other
than PFAS also indicate that departures from dose additivity are uncommon and rarely exceed
minor deviations (~2-fold) from predictions based on dose additivity (Martin et al., 2021; Boobis
et al., 2011). Boobis et al. (2011) examined literature from 1990 to 2008 that discussed synergy
in mammalian test systems, with an emphasis on "low dose" studies. They found that of the 11
available studies with synergy data that reported the magnitude of the difference between the
dose-additive estimates of toxicity and observed toxicity, six studies reported magnitudes of
synergy that were generally small, and the authors concluded that deviations from dose additivity
at low doses were not common. Additionally, Martin et al. (2021) reviewed more than 1,200
mixture studies and concluded that there was little evidence for synergy (greater than additive
effects) or antagonism (less than additive effects) among chemicals in mixtures, and that dose
additivity should be considered as the default model. This supports the health-protective
conclusion that a mixture of PFHxS, PFNA, HFPO-DA, and/or PFBS should be assumed to act
in a dose-additive manner unless data demonstrate otherwise.
Although some available in vitro studies do not provide conclusive evidence of dose additivity
for PFAS mixtures, their results also do not justify drawing a conclusion other than dose
additivity. For example, a study on PFAS cytotoxicity in a human liver cell line (Ojo et al., 2020)
1-3
-------
FINAL
2024
reported synergistic effects of mixtures of perfluoroalkyl acids (PFAAs; a type of PFAS)
compared to a dose addition model, but also reported evidence of antagonistic effects. Other in
vitro studies that have assessed PFAS mixture-based effects do not report these results; that is,
they do not offer strong evidence for synergistic or antagonistic effects, particularly at
environmentally relevant concentrations. For example, Wolf et al. (2014) evaluated in vitro
PPARa activation and reported that effects seen following exposure to combinations of different
PFAS were consistent with dose additivity in the lower tested concentration ranges. Wolf et al.
(2014) also reported slightly greater than additive effects at higher test concentrations
(approximately 500 parts per billion to over 800 parts per million); however, in environmental
media such as drinking water, PFAS are not likely to occur at these higher concentrations (e.g.,
see USEPA, 2024b). Carr et al. (2013) reported slightly less than additive effects for in vitro
PPARa activation of binary mixtures of PFAAs including PFOA, PFNA, PFOS, and PFHxS.
Addicks et al. (2023) evaluated mRNA transcription in primary human liver spheroids exposed
to seven different PFAS mixtures and found that all tested mixtures produced effects that were
consistent with effects predicted using dose addition. To summarize, the available in vitro data
do not support a conclusion other than dose additivity for PFAS mixtures.
Available in vivo data on this subject similarly support dose additivity. Two studies with PFAS
mixtures in zebrafish reported no indications of synergy (Menger et al., 2020; Ding et al., 2013).
Additionally, recent EPA Office of Research and Development (ORD) studies provide robust
evidence that PFAS behave in a dose-additive manner (Gray et al., 2024; Conley et al., 2023;
Conley et al., 2022). For example, results of a developmental toxicity study of exposure to PFOA
and PFOS mixtures in rats showed that the observed results for almost all tested endpoints were
consistent with dose additivity (Conley et al., 2022). Likewise, a rat developmental study of a
PFAS mixture of PFOS, HFPO-DA, and Nafion byproduct 2 (an emerging
polyfluoroethersulfonic acid compound recently detected in human serum (Kotlarz et al., 2020))
found that multiple tested endpoints in both parental females and offspring conformed to dose
additivity and no endpoints demonstrated synergy (Conley et al., 2023).
Additionally, as described in the final PFAS Mixtures Framework (USEPA, 2024a), over the
past two decades, many in vivo experimental animal studies have been published in which
toxicity of chemical mixtures has been systematically evaluated (e.g., Conley et al., 2023;
Conley et al., 2022; Martin et al., 2021; Hass et al., 2017; Howdeshell et al., 2015; Moser et al.,
2012; Rider et al., 2010; Kortenkamp and Haas, 2009; Rider et al., 2009; Rider et al., 2008;
Crofton et al., 2005; Moser et al., 2005; Walker et al., 2005; Gennings et al., 2004; Altenburger
et al., 2000). These studies span different chemical classes, proposed MOAs, and health
outcomes, but they generally show that chemicals in mixtures typically act dose additively. Even
when mixture components with different MOAs/adverse outcome pathways (AOPs) are
combined, they induce toxic effects consistent with dose additivity (Rider et al., 2009). This
concept was further articulated in the National Research Council's 2008 report Phthalates and
cumulative risk assessment: The tasks ahead (NRC, 2008), wherein that expert panel provided
significant evidence that mixture components that elicit similar adverse health effects
individually will demonstrate dose additivity when combined in a mixture, regardless of
similarity in MOA.
This evidence base supports the longstanding recommendation in EPA chemical mixtures
guidance for dose additivity as a default approach for evaluation of mixture toxicity (USEPA,
1-4
-------
FINAL
2024
2000a, 1986). This position is further supported and articulated in the newly published EPA Risk
Assessment Forum's Advances in Dose Addition for Chemical Mixtures: A White Paper
(USEPA, 2023b).
1.3.1.1 Science A dvisory Board Support
The EPA's conclusions regarding dose additivity of PFAS were supported by the SAB during its
2021-2022 review of the EPA's draft Framework for Estimating Noncancer Health Risks
Associated with Mixtures of Per- and Polyfluoroalkyl Substances. The EPA directly asked the
SAB for feedback on PFAS dose additivity as part of its review of technical materials supporting
development of the PFAS MCLG and NPDWR. Specifically, the EPA asked the SAB to,
"[pjlease comment on the appropriateness of this approach for a component-based mixture
evaluation of PFAS under an assumption of dose additivity" (USEPA SAB, 2022). The SAB
strongly supported the scientific soundness of this approach when evaluating PFAS and
concurred that it was a health protective conclusion. For example, the SAB said:
"The SAB supports dose additivity based on a common outcome, instead of a common
mode of action as a health protective default assumption and does not propose another
default approach." (USEPA SAB, 2022)
".. .The information included in the draft framework supports the conclusion that
toxicological interactions of chemical mixtures are frequently additive or close to
additive. It also supports the conclusion that dose additivity is a public health protective
assumption that typically does not underestimate the toxicity of a mixture..." (USEPA
SAB, 2022)
"The SAB Panel agrees with use of the default assumption of dose additivity when
evaluating PFAS mixtures that have similar effects and concludes that this assumption is
health protective." (USEPA SAB, 2022)
".. .dose additivity can provide an estimate of composite effects." (USEPA SAB, 2022)
While the SAB also noted that there remain some questions about PFAS interaction in mixtures
(USEPA SAB, 2022), the available data justify an approach that accounts for PFAS dose
additivity. As described above, studies that have assessed PFAS mixture-based effects do not
provide support for a conclusion other than dose additivity (i.e., they do not offer strong evidence
for synergistic/antagonistic effects) (USEPA, 2024a).
1.3.2 Toxicological Similarity of PFHxS, PFNA, HFPO-DA, and
PFBS
1.3.2.1 Background on Concept of "Toxicological Similarity"
This concept and application of dose additivity for "toxicologically similar components" in
mixtures assessment is consistent with EPA mixtures guidance (USEPA, 2000a, 1986) and the
EPA Risk Assessment Forum's Advances in Dose Addition for Chemical Mixtures: A White
Paper (USEPA, 2023b). Specifically, the EPA's Supplementary Guidance for Conducting
Health Risk Assessment of Chemical Mixtures (USEPA, 2000a) notes that although the shared
MOA metric for application of dose addition is optimal, MOA data are not always available and
that toxicological similarity in the context of mixtures risk assessment can be based on adverse
1-5
-------
FINAL
2024
effects observed at the organ or system level (USEPA, 2000a). This concept is further described
in the EPA Risk Assessment Forum's Advances in Dose Addition for Chemical Mixtures: A
White Paper (USEPA, 2023b): "The primary criterion for choosing between dose addition and
response addition methods is toxicological similarity among the chemicals in the mixture
[(USEPA, 2000a)]. "Toxicological similarity" is used here as an overarching concept with a
wide range of specificity across levels of biological organization, allowing simil arity judgments
to be tailored to both the specific goals of the mixture risk assessment and the availability of
hazard and dose-response information across components." Unless there are available data that
suggest deviation(s) from dose additivity, mixture chemicals that are "toxicologically similar"
(e.g., same/similar effect or profile of effect[s], regardless of differences in potencies)
prototypically behave dose additively. This concept is depicted in Figure 1-1 below, which
shows that dose additivity is the logical default approach for "toxicologically similar"
components and that component-based mixture assessment approaches including hazard index
(HI), relative potency factor (RPF), and mixture-benchmark dose (Mixture-BMD) are options for
mixture assessment in such cases (see Section 1.5).
Notes:
Modification of Figure 4-3bin USEPA (2007). BMD = benchmark dose; HI = hazard index; HQ = hazard quotient; MOE = margin of exposure; RPF = relative potency
factor, TOSHI = target organ-specific hazard index.
Component-based methods selection is based on the relevant evidence supporting toxicological similarity (dose addition) or toxicological independence (response
addition or effect summation). Integrated addition methods are reserved for mixtures of component chem icals that demonstrate a profile of both toxicological
similarity and independence.
Figure 1-1. Flow chart for evaluating chemical mixtures using component-based additive
methods. (Reproduction of Figure 2-1 from USEPA, 2024a).
1.3.2.2 Overview of Scientific Support
The EPA's approach is to evaluate risks from exposure to mixtures of PFAS based on similar
adverse health effects (but with differing potencies for effect(s)) of the individual PFAS mixture
components, rather than similar MOA. MOA describes key changes in cellular or molecular
1-6
-------
FINAL
2024
events that may cause functional or structural changes that lead to adverse health effects and can
be a useful metric by which risk can be assessed. It is considered a key determinant of chemical
toxicity, and chemicals can often be classified by their type of toxicity pathway(s) or MOA(s).
PFAS are an emerging chemical class, and MOA data are limited or entirely lacking for many
PFAS. Although similarities among some PFAS have been shown at the level of molecular and
cellular perturbations, no conserved MO As have been identified across PFAS for noncancer
health effects assessed thus far. Therefore, the EPA's approach for assessing risks of PFAS
mixtures is based on the conclusion that PFAS that are "toxicologically similar"—that is, elicit
the same or similar adverse health effects (but with differing potencies for effect(s))—will
produce dose-additive effects from co-exposures (see USEPA, 2024a).
Available epidemiological and animal toxicological data demonstrate that exposure to each of
these four PFAS (PFHxS, PFNA, HFPO-DA, and PFBS) is associated with many of the same or
similar adverse health endpoints and outcomes, and thus they are "toxicologically similar" (see
Table 1-1, Table 1-2, and Table 1-3 below). Further, these four PFAS are well-studied PFAS for
which the EPA or ATSDR has developed human health assessments and toxicity reference
values (i.e., reference doses (RfDs), minimal risk levels (MRLs)). Available animal toxicological
and/or epidemiological studies demonstrate that PFHxS, PFNA, HFPO-DA, and PFBS are
documented to affect at least five (5) of the same major health outcomes: lipids, developmental,
immune, endocrine, and hematologic (Table 1-1). Similarly, according to the 2023 Interagency
PFAS Report to Congress (US OSTP, 2023), available animal toxicological data show that
PFHxS, PFNA, HFPO-DA, and PFBS significantly affect at least eight (8) of the same major
health effect domains: body weight, respiratory, hepatic, renal, endocrine, immunological,
reproductive, and developmental (Table 1-2). Furthermore, numerous in vivo and in vitro studies
demonstrate that these four PFAS share many common health effects across diverse health
outcome categories (e.g., developmental, immunological, and endocrine), and that they induce
some of the same effects at the molecular level along biological pathways. Table 1-3 below
shows specific endpoints shared across these four PFAS, including toxicologically relevant
molecular perturbations (in vitro), and health effects (in vivo) from oral repeated-dose studies in
rats and/or mice (note that this table is a summary of select studies for illustrative purposes and
should not be construed to represent a systematic review or MOA analysis).
Table 1-1. Affected health outcomes in animal toxicological and/or epidemiological studies
for the four PFAS included in the HI MCLG (adapted from Table 6-7 in USEPA, 2024d).
Health Outcome
HFPO-DA
PFNA
PFHxS
PFBS
Lipids
X
X
X
X
Developmental
X
X
X
X
Hepatic
X
X
X
-
Immune
X
X
X
X
Endocrine
X
X
X
X
Renal
X
-
-
X
Hematologic
X
X
X
X
Notes: (X) Health outcome examined, evidence of association; (-) health outcome examined, no evidence of association.
1-7
-------
FINAL
2024
Table 1-2. Affected health endpoints in animal toxicity studies for the four PFAS included
in the HI MCLG (adapted from Table 4 in US OSTP, 2023).
Health Endpoint
HFPO-DA
PFNA
PFHxS
PFBS
Body weight
X
X
X
X
Respiratory
X
X
X
X
Cardiovascular
X
X
Gastrointestinal
X
X
X
Hematological
X
X
X
Musculoskeletal
X
X
Hepatic
X
X
X
X
Renal
X
X
X
X
Dermal
X
Ocular
X
X
Endocrine
X
X
X
X
Immunological
X
X
X
X
Neurological
X
X
X
Reproductive
X
X
X
X
Developmental
X
X
X
X
Other noncancer
X
X
Notes: (X) Health outcome examined, evidence of association.
Table 1-3. Specific Endpoints Affected by One or More of the Four PFAS Included in the
HI MCLG.
Endpoint
HFPO-DA
PFNA
PFHxS
PFBS
Molecular/Cellular Perturbations
PPAR alpha
binding/activation
Evans et al.
(2022); Nielsen
et al. (2021)
Evans et al.
(2022); Nielsen
et al. (2021);
Rosenmai et al.
(2018); Wolf et
al. (2012)
Evans et al.
(2022); Nielsen
et al. (2021);
Rosenmai et al.
(2018); Wolf et
al. (2012)
Evans et al.
(2022);
Rosenmai et al.
(2018); Wolf et
al. (2012)
PPAR gamma
binding/activation
Evans et al.
(2022); Houck et
al. (2021)
Evans et al.
(2022); Houck et
al. (2021)
Evans et al.
(2022); Houck et
al. (2021)
Evans et al.
(2022)
Liver gene induction
(PPAR signaling
pathway)
Conley et al.
(2019); Blake et
al. (2022)
NTP (2019b);
Rosen et al.
(2017); Rosen et
al. (2013)
NTP (2019c);
Rosen et al.
(2017); Rosen et
al. (2013); Chang
et al. (2018)
NTP (2019c);
Rosen et al.
(2013)
Liver gene induction
(CAR signaling
pathway)
-
NTP (2019b)
NTP (2019c)
NTP (2019c)
1-8
-------
FINAL
2024
Endpoint
HFPO-DA
PFNA
PFHxS
PFBS
Serum bile
salts/acids
(increased)
DuPont (2010c)
NTP (2019b)
-
NTP (2019c)
Serum globulin
(reduced)
DuPont (2009);
DuPont
(2008b);
DuPont (2008a)
NTP (2019b)
NTP (2019c)
NTP (2019c)
Serum
albumin:globulin
(increased)
DuPont (2009);
DuPont
(2008b);
DuPont (2008a)
NTP (2019b)
NTP (2019c);
Butenhoff et al.
(2009)
NTP (2019c)
Health Effects
Serum lipids
(reduced cholesterol
and/or triglycerides)
DuPont
(2008b);
DuPont (2009);
DuPont (2008a)
NTP (2019b)
NTP (2019c);
Chang et al.
(2018);
Butenhoff et al.
(2009)
NTP (2019c)
Serum liver enzymes
(increased ALT,
AST, and/or ALKP)
DuPont
(2008b);
DuPont (2010c)
NTP (2019b)
-
NTP (2019c)
Serum thyroid
hormones (reduced
T4, T3)
Conley et al.
(2019)
NTP (2019b)
NTP (2019c);
Gilbert et al.
(2021)
NTP (2019c)
Liver weight
(increased)
DuPont
(2008b); Blake
et al. (2020);
Conley et al.
(2021); Conley
et al. (2019);
DuPont (2009);
Rushing et al.
(2017); DuPont
(2008a)
NTP (2019b);
Das et al.
(2015)
NTP (2019c);
Chang et al.
(2018)
NTP (2019c);
Lieder et al.
(2009b)
Liver histopathology
(nonneoplastic
effects)
DuPont
(2008b);
DuPont
(2010c);
DuPont
NTP (2019b)
Chang et al.
(2018); NTP
(2019c)
NTP (2019c)
1-9
-------
FINAL
2024
Endpoint
HFPO-DA
PFNA
PFHxS
PFBS
(2008a); NTP
(2019a)
Thymus weight
(reduced)
DuPont (2009)
NTP (2019b)
-
NTP (2019c)
Spleen weight
(reduced)
-
NTP (2019b)
-
NTP (2019c);
Lieder et al.
(2009a)
Kidney weight
(increased)
DuPont (2009);
DuPont (2008b)
NTP (2019b)
NTP (2019c)
NTP (2019c)
Reduced fetal/pup
bodyweight
Conley et al.
(2021); DuPont
(2010a);
DuPont (2010b)
Das et al.
(2015)
-
Feng et al.
(2017)
Reduced fetal/pup
survival
Conley et al.
(2021)
Das et al.
(2015)
Chang et al.
(2018)
-
Reduced adult
bodyweight
DuPont (2013)
NTP (2019b)
-
NTP (2019c);
Lieder et al.
(2009a)
Overt toxicity
(lethality)
DuPont (2009)
NTP (2019b)
-
NTP (2019c)
Note: (-) indicates no statistically significant effect reported by study authors of cited studies at dose levels and dose interval
used and/or effect not measured in cited studies.
1.3.2.3 Science A dvisory Board Support
The SAB strongly supported the EPA's decision to focus on similarity of adverse health effects
rather than similarity of MO A to assess risks of exposure to PFAS mixtures during its 2021-
2022 review of the EPA's draft PFAS Mixtures Framework. Specifically, the EPA asked the
SAB, "If common toxicity endpoint/health effect is not considered an optimal similarity domain
for those PFAS with limited or no available MOA-type data, please provide specific alternative
methodologies for integrating such chemicals into a component-based mixture evaluation(s)"
(USEPA SAB, 2022). The SAB strongly supported the EPA's approach of using a similar
toxicity endpoint/health effect instead of a common MOA as a default approach for evaluating
mixtures of PFAS using dose additivity and did not recommend an alternative methodology. The
SAB panel stated that:
"The Panel agreed with use of a similar toxicity endpoint/health effect instead of a
common MOA as a default approach for evaluating mixtures of PFAS. This approach
makes sense because multiple physiological systems and multiple MO As can contribute
1-10
-------
FINAL
2024
to a common health outcome. Human function is based on an integrated system of
systems and not on single molecular changes as the sole drivers of any health outcome.
The Panel concluded that rather than the common MO A, as presented in the EPA draft
mixtures document, common physiological outcomes should be the defining position"
(USEPA SAB, 2022).
"Furthermore, many PFAS, including the four used in the examples in the draft EPA
mixtures document and others, elicit effects on multiple biological pathways that have
common adverse outcomes in several biological systems (e.g., hepatic, thyroid, lipid
synthesis and metabolism, developmental and immune toxicities)" (USEPA SAB, 2022).
1.3.3 Summary
The available scientific evidence supports the conclusion that PFAS that elicit similar adverse
health effects following individual exposure (even if with differing potencies for effect(s))
should be assumed to act in a dose-additive manner when in a mixture unless data demonstrate
otherwise. This means that individual PFAS, each at doses that are not anticipated to result in
adverse health effects, when combined in a mixture may result in adverse health effects. (For a
more complete discussion of the evidence supporting dose additivity as the default approach for
assessing mixtures of PFAS, see the final PFAS Mixtures Framework (USEPA, 2024a)). The
EPA's conclusions regarding PFAS dose additivity were supported by the SAB during its review
of the EPA's draft PFAS Mixtures Framework (USEPA SAB, 2022) and are consistent with
longstanding agency chemical mixtures guidance (USEPA, 2000a, 1986) and a recent EPA white
paper (USEPA, 2023b). The SAB also strongly supported the EPA's default assumption of dose
additivity in the absence of other information and the EPA's approach of using similar toxicity
endpoints/health effects instead of a common MO A for evaluating mixtures of PFAS (USEPA
SAB, 2022). This approach of basing the concept of toxicological similarity on same/similar
adverse effects in the absence of adequate MOA information is also consistent with the EPA's
chemical mixtures guidance (USEPA, 2000a, 1986) and the EPA Risk Assessment Forum's
Advances in Dose Addition for Chemical Mixtures: A White Paper (USEPA, 2023b).
1.4 General Hazard Index (HI) Approach for PFAS Mixtures
1.4.1 Background/Overview
The EPA's final determination that mixtures of the four PFAS "may have an adverse effect on
the health of persons" is based on the health-protective conclusion that chemicals that are
toxicologically similar (i.e., have similar observed adverse health effects, regardless of potency
differences) following individual exposure should be assumed to act in a dose-additive manner
when in a mixture unless data demonstrate otherwise (see Section 1.3 and USEPA, 2024a). This
means that where drinking water contains any combination of two or more of the four PFAS that
are the subject of the NPDWR—PFHxS, PFNA, HFPO-DA, and PFBS—the hazard associated
with each PFAS in the mixture must be added together to determine whether the mixture exceeds
a level of public health concern.
The SDWA requires the agency to establish a health-based MCLG set at "a level at which no
known or anticipated adverse effects on the health of persons occur and which allows for an
adequate margin of safety." The MCLG "incorporates a margin of safety to reflect scientific
1-11
-------
FINAL
2024
uncertainty and, in some cases, the particular susceptibility of some groups (e.g., children) within
the general population" (see S. Rep. No. 169, 104th Cong., 1st Sess. (1995) at 3). In the context
of this NPDWR, the general HI is the approach used to determine if a mixture of two or more of
four PFAS in drinking water—PFHxS, PFNA, HFPO-DA, and PFBS—exceeds the level of
health concern with a margin of safety. A general HI equal to 1 is the MCLG for any mixture of
these four PFAS.
Based on the scientific record, each of these four PFAS has a health-based water concentration
(HBWC), which is set at the level below which adverse effects are not anticipated to occur and
allows for an adequate margin of safety (see Section 2). The general HI approach accounts for
the measured drinking water concentration of each of the four PFAS in the mixture, and the
toxicity (represented by the HBWC) of each of the four PFAS. The general HI is derived by first
calculating the ratio of the measured concentration of each of the four PFAS to its toxicity (the
HBWC) to yield a "hazard quotient" (HQ) for each of the four PFAS. HQs are then added
together to account for the dose-additive health concerns that these PFAS present. Adding the
four HQs together yields the general HI. If the general HI exceeds 1, then the hazard from the
combined amounts of the four PFAS present together in drinking water exceeds a level of public
health concern.
The EPA has determined that in the context of SDWA, the general HI is an appropriate
methodology for determining the level at and below which there are no known or anticipated
adverse human health effects with an adequate margin of safety with respect to certain PFAS
mixtures in drinking water. The general HI approach is the most practical approach for
establishing an MCLG for PFAS mixtures that meets the statutory requirements outlined in
Section 1412(b)(1)(A) of SDWA. As noted above, the general HI assesses the exposure level of
each component PFAS relative to its HBWC, which is based on the most sensitive known
adverse health effect (based on the weight of evidence) and considers sensitive population(s) and
life stage(s) as well as potential exposure sources beyond drinking water. The general HI also
accounts for dose-additive health concerns by summing the hazard contributions from each
mixture component. In this way, the general HI approach ensures that mixtures of two or more of
these four PFAS are not exceeding the level below which there are no known or anticipated
adverse health effects and allows for an adequate margin of safety.
1.4.2 Consideration of Mixtures Assessment Approaches
and Selection of General Hi Approach
In selecting an approach to develop the MCLG for mixtures of two or more of four PFAS—
PFHxS, PFNA, HFPO-DA, and PFBS—the EPA followed its Guidelines for the Health Risk
Assessment of Chemical Mixtures (USEPA, 1986), Supplementary Guidance for Conducting
Health Risk Assessment of Chemical Mixtures (USEPA, 2000a), and Risk Assessment Guidance
for Superfund (e.g., USEPA, 1991b). As described below, the EPA first considered whether data
were available for the whole mixture or a "sufficiently similar" mixture, per agency guidance
(USEPA, 2000a, 1986), and then considered several mixture component-based assessment
methods (USEPA, 2024a), ultimately selecting the general HI approach for PFAS mixture
MCLG derivation.
1-12
-------
FINAL
2024
The EPA's guidance documents (USEPA, 2000a, 1991b, 1986) propose a hierarchy of mixtures
assessment approaches, with the preferred approach being evaluation of health risk using hazard
and dose-response data for a specific whole mixture of concern, or alternatively, a "sufficiently
similar" mixture. Whole-mixture data are rare; there are often many chemical combinations and
proportions in the environment (e.g., parent chemicals, metabolites, and/or abiotic degradants),
introducing a level of complexity that complicates evaluation and characterization. The
exponential diversity of PFAS co-occurring in different combinations and proportions makes
whole-mixture evaluations complex and unfeasible. Due to differing fate and transport
properties, biotic (metabolism) and abiotic (degradation) processes, pH, ultraviolet radiation,
media temperature, and so on, chemicals commonly co-occur in the environment in an array of
parent species, metabolites, and/or degradants, making characterization and evaluation of any
given mixture complicated. In controlled experimental study designs, whole mixtures can be
assembled with defined component membership and proportions, but the relevance of toxicity
associated with exposure to a defined mixture in a laboratory setting may not be translatable to
environmental mixtures of different component combinations and proportions across time and
space in environmental media. The complexities associated with the diversity of PFAS co-
occurring in different component proportions (see USEPA, 2024b) make evaluating each unique
whole mixture of PFAS intractable. This is why component-based mixture assessment
approaches are considered particularly useful and appropriate for addressing human exposure(s)
to mixtures of PFAS (see Sections 5-7 in USEPA, 2024a). For a more detailed discussion on
whole-mixture and component-based approaches for PFAS risk assessment, please see the final
PFAS Mixtures Framework (USEPA, 2024a).
The EPA considered several component-based assessment approaches to develop an MCLG for
mixtures of PFHxS, PFNA, HFPO-DA, and/or PFBS under an assumption of dose additivity,
including the general HI, the target organ-specific HI (TOSHI), the Relative Potency Factor
(RPF) approach, and the mixture-benchmark dose (mixture-BMD) approach (USEPA, 2024a).
As part of the technical support materials for the PFAS NPDWR, the EPA's draft PFAS
Mixtures Framework (USEPA, 2021f) was submitted to the SAB for expert review. The SAB
supported the EPA's proposed component-based approaches under the assumption of dose
additivity in the absence of information to support a conclusion other than dose additivity (see
Section 1.3). Following the SAB review, the EPA addressed the SAB's recommendations. Then,
the EPA solicited public comment on the draft PFAS Mixtures Framework as part of the
proposed NPDWR (88 FR 18638; USEPA (2023c)). The EPA evaluated potential component-
based mixture assessment options, and ultimately proposed using the general HI approach as the
most appropriate option based on available data and consistent with the statutory definition of
MCLG.
The EPA first considered a "whole mixture" approach. Although use of data from whole
mixtures or "sufficiently similar mixtures" is ideal in a theoretical sense, it is not practical,
possible, or necessary for evaluating mixtures of PFAS in drinking water. Instead, the EPA is
using the general HI approach, a longstanding component-based mixtures assessment approach
which was endorsed by the SAB in the context of assessing risk associated with exposure to
PFAS mixtures in drinking water (USEPA SAB, 2022), as discussed below. The goal of this
component-based mixtures assessment approach is to approximate what the whole-mixture
toxicity would be if the whole mixture could be tested and relies on toxicity information for each
individual component in a mixture (USEPA, 2000a). A whole-mixture approach for regulating
1-13
-------
FINAL
2024
mixtures of these four PFAS in drinking water is not possible because it would entail developing
a single toxicity reference value (e.g., RfD) for one specific mixture of PFHxS, PFNA, HFPO-
DA, and PFBS with defined proportions of each PFAS. Toxicity studies are typically conducted
with only one test substance to isolate that particular substance's effects on test organisms, and
whole-mixture data are exceedingly rare. There are no known whole-mixture studies for PFHxS,
PFNA, HFPO-DA, and PFBS, and even if they were available, a toxicity reference value derived
from such a study (i.e., a single RfD for a specific mixture of these four PFAS) would only be
directly applicable to that specific mixture. Thus, a more flexible approach is necessary—one
that considers the potential for the four PFAS to co-occur in different combinations and at
different concentrations across time and space. The general HI approach affords this flexibility;
the general HI indicates risk from exposure to a mixture and is useful to ensure a health-
protective MCLG for PFAS mixtures that can be spatially and/or temporally variable. Given the
variability of PFAS occurrence in drinking water across the nation (USEPA, 2024b), the general
HI allows the EPA to regulate mixtures of these PFAS in drinking water by taking into account
site-specific data at each PWS. HQs for the four different PFAS are expected to differ depending
on the actual measured concentrations of each of the four PFAS at each PWS. The general HI
approach thus allows for flexibility beyond a one-size-fits-all approach and is tailored to address
risk at each PWS. Furthermore, the EPA's application of the general HI approach accounts for
the dose additivity that was the basis for the EPA's final determination to regulate mixtures of
two or more of these PFAS.
The EPA considered the two main types of HI approaches: 1) the general HI, which allows for
component chemicals in the mixture to have different health effects or endpoints as the basis for
their toxicity reference values (e.g., RfDs, minimal risk levels), and 2) the TOSHI, which relies
on toxicity reference values based on the same specific target organ or system effects (e.g.,
effects on the liver or thyroid; effects on developmental or reproductive systems) (USEPA,
2000a). The general HI approach uses the most health-protective RfD (or minimal risk level)
available for each mixture component, irrespective of whether the RfDs for all mixture
components are based on effects in the same target organs or systems. These "overall" RfDs (as
they are sometimes called) are protective of all other adverse health effects because they are
based on the most sensitive known endpoints as supported by the weight of the evidence. As a
result, this approach is protective of all types of toxicity/adverse effects, and thus ensures that the
MCLG is the level at and below which there are no known or anticipated adverse human health
effects with an adequate margin of safety with respect to certain PFAS mixtures in drinking
water.
The TOSHI produces a less health protective indicator of risk than the general HI because the
basis for the mixture component toxicity reference values has been limited to a specific target
organ or system effect, which may occur at higher exposure levels than other effects (i.e., be a
less sensitive endpoint). In other words, a TOSHI may not be health protective compared to the
general HI if available data for a mixture component show effects in other organs at lower
exposure levels compared to the critical effect observed in the target organ used for the TOSHI.
Additionally, since a TOSHI relies on toxicity reference values aggregated for the same specific
target organ or system endpoint/effect, an absence or lack of data on the specific target organ or
system endpoint/effect for a mixture component may result in that component not being
adequately accounted for in this approach (thus, underestimating health risk of the mixture). A
TOSHI can only be derived for those PFAS for which the same target organ or system
1-14
-------
FINAL
2024
endpoint/effect-specific RfDs have been calculated. For example, a TOSHI based on changes in
thyroid effects illustrates why the target organ-specific approach underestimates risk in the
context of these four PFAS in drinking water. To develop a thyroid effects-based TOSHI for
mixtures of these four PFAS, only those PFAS with chronic toxicity reference values based on
thyroid effects—PFHxS (MRL) and PFBS (RfD)—would be included in the TOSHI calculation;
HFPO-DA and PFNA have chronic toxicity reference values based on other effects (i.e., liver
and developmental effects, respectively) and thus would not be included in a thyroid effects-
based TOSHI. Although thyroid effects are not the basis for the RfDs for HFPO-DA and PFNA,
studies have shown that these two PFAS significantly affect the thyroid; for example, both have
been shown to significantly affect serum thyroid hormone levels (reduced T4, T3) (Conley et al.,
2019; NTP, 2019b). According to the Interagency Report to Congress on PFAS, "Multiple
studies on diverse species (developing rodents and fish) suggest that some PFAS (e.g., PFOS,
PFOA, PFNA, GenXchemicals, PFHxS, PFDA, PFBA, PFBS, PFHxA) interfere with thyroid
hormone signaling pathways and thyroid homeostasis through various mechanisms, including
regulation of hepatic glucuronidation enzymes and deiodinases in the thyroid gland" (emphasis
added, US OSTP, 2023). Therefore, a thyroid-specific HI that excluded HFPO-DA and PFNA
would underestimate the dose additivity concerns for thyroid effects from the total mixture.
Many PFAS have data gaps in epidemiological or animal toxicological dose-response
information for multiple types of health effects, thus limiting derivation of target organ-specific
toxicity reference values; target organ-specific toxicity reference values for the same target for
all four PFAS are not currently available for PFHxS, PFNA, HFPO-DA, and PFBS. The EPA's
guidance recognizes the potential for organ- or system-specific data gaps and supports use of
overall RfDs in a general HI approach, stating, "The target organ toxicity dose (TTD) is not a
commonly evaluated measure and currently there is no official EPA activity deriving these
values, as there is for the RfD and RfC" ... "Because of their much wider availability than TTDs,
standardized development process including peer review, and official stature, the RfD and RfC
are recommended for use in the default procedure for the HI" (USEPA, 2000a). Even if target
organ-specific toxicity reference values (TTDs) were available for PFHxS, PFNA, HFPO-DA,
and PFBS, the general HI approach would still be more appropriate for this specific application
because it is protective of all adverse health effects rather than just those associated with a
specific organ or system, consistent with the statutory definition of MCLG.
Although these four PFAS elicit many of the same adverse health effects, the most sensitive
known endpoint for each of the four PFAS is different, and thus the toxicity reference values
used to calculate the HBWCs in the general HI approach are different. Epidemiological and/or
experimental animal studies have demonstrated that exposure to PFHxS, PFNA, HFPO-DA, and
PFBS individually is associated with many of the same observed adverse health effects (e.g.,
effects on lipids, as well as developmental, immune, endocrine, and hematologic endpoints; see
Section 1.3), but with differing potencies for effect(s). In other words, two or more PFAS may
elicit the same adverse effects, but at different exposure levels; for example, liver effects are
associated with all four PFAS (PFHxS, PFNA, HFPO-DA, and PFBS) but HFPO-DA is the only
one of the four for which liver effects represent the most sensitive known endpoint and serve as
the basis for its toxicity reference value (i.e., RfD). The fact that the toxicity reference values
(i.e., RfDs or MRLs) for the four PFAS are based on different health endpoints does not mean
that the four PFAS are not toxicologically similar; rather, it means that based on the available
data, the most sensitive endpoint currently known is different for each of these PFAS. The
1-15
-------
FINAL
2024
general HI approach uses the most health-protective toxicity reference value available for each of
the four PFAS to derive HBWCs, irrespective of whether they are based on effects in the same
target organs or systems. Since each RfD (or MRL) is based on the most sensitive known
endpoint based on the weight of evidence (i.e., toxicity reference value selection is not limited to
a specific organ or system), this approach is protective of all other adverse health effects. This
approach of allowing for component chemicals in the mixture to have different health effects or
endpoints as the basis for their toxicity reference values is consistent with EPA guidance (see
examples in USEPA, 2000a; USEPA, 1991b) and was supported by SAB (see Section 1.4.2.1).
The general HI is a well-established methodology that has been used for several decades in at
least one other regulatory context to account for dose additivity in mixtures assessments. The
EPA routinely uses the HI approach to consider the risks from multiple contaminants of concern
in the Remedial Investigations and Feasibility Studies for cleanup sites on the Superfund
National Priorities List under the Comprehensive Environmental Response, Compensation, and
Liability Act (CERCLA). Noncarcinogenic effects are summed to provide an HI that is
compared to an acceptable index, generally 1. This approach assumes dose additivity in the
absence of information on a specific mixture. These assessments of hazards from multiple
chemical exposures are important factors to help inform the selection of remedies that are
ultimately captured in the Superfund Records of Decision.
1.4.2.1 Science A dvisory Board Support
The EPA directly asked the SAB about the utility and scientific defensibility of the general HI
approach (in addition to other methods, including TOSHI) during the SAB's 2021-2022 review
of the EPA's draft Framework for Estimating Noncancer Health Risks Associated with Mixtures
of PFAS. Specifically, the EPA asked the SAB to "Please provide specific feedback on whether
the HI approach is a reasonable methodology for indicating potential risk associated with
mixtures of PFAS. If not, please provide an alternative;" and "Please provide specific feedback
on whether the proposed HI methodologies in the framework are scientifically supported for
PFAS mixture risk assessment" (USEPA SAB, 2022). In its report (USEPA SAB, 2022), the
SAB stated its support for the general HI approach:
"The HI methodology is a reasonable approach for estimating the potential aggregate
health hazards associated with the occurrence of chemical mixtures in environmental
media. The HI is an approach based on dose additivity (DA) that has been validated and
used by the EPA.... This approach is mathematically straightforward and may readily
identify mixtures of potential toxicological concern, as well as identify chemicals that
drive the toxicity within a given mixture." (USEPA SAB, 2022).
"In general, the screening level Hazard Index (HI) approach, in which Reference Values
(RfVs) for the mixture components are used regardless of the effect on which the RfVs
are based, is appropriate for initial screening of whether exposure to a mixture of PFAS
poses a potential risk that should be further evaluated. Toxicological studies to inform
human health risk assessment are lacking for most members of the large class of PFAS,
and mixtures of PFAS that commonly occur in environmental media, overall. For these
reasons, the HI methodology is a reasonable approach for estimating the potential
aggregate health hazards associated with the occurrence of chemical mixtures in
environmental media. The HI is an approach based on dose additivity (DA) that has been
1-16
-------
FINAL
2024
validated and used by the EPA. The HI does not provide quantitative risk estimates (i.e.,
probabilities) for mixtures, nor does it provide an estimate of the magnitude of a specific
toxicity. This approach is mathematically straightforward and may readily identify
mixtures of potential toxicological concern, as well as identify chemicals that drive the
toxicity within a given mixture." (USEPA SAB, 2022).
The SAB recognized the need for regulatory agencies to make decisions in the face of
uncertainty to reduce exposures to PFAS. The SAB stated,
"Given the agency's desire to support fit-for-purpose approaches, not every PFAS
mixture scenario will be one that warrants a tiered or hierarchical approach. In some
instances, an HI or target-organ-specific hazard indices (TOSHI) might provide enough
information for decision-making about PFAS (or other chemicals) contamination in
drinking water (or other media). Tiered approaches that require increasingly complex
information before reaching a final decision point can be extremely challenging for data-
poor chemicals such as PFAS. Data gaps identified in a such tiered methodologies could
result in a bottleneck through which these chemicals may never emerge..(USEPA
SAB, 2022).
1.5 Establishment of Individual MCLGs for PFHxS, HFPO-DA,
PFNA, and/or PFBS
The EPA has determined that sufficient information is available to satisfy the statutory
requirements for individual regulation of PFHxS, HFPO-DA, and PFNA (in addition to PFOA
and PFOS). To support this determination, the EPA carefully examined the health effects
information from available peer-reviewed final human health assessments as well as published
studies, reviewed PFAS drinking water occurrence data collected as part of the UCMR 3 and
state-led monitoring efforts, and considered public comments received. The EPA finds that oral
exposure to PFHxS, HFPO-DA, or PFNA individually may lead to adverse health effects in
humans; that each of these three PFAS have a substantial likelihood of occurring in finished
drinking water with a frequency and at levels of public health concern; and that, in the sole
judgment of the Administrator, regulation of PFHxS, HFPO-DA, and PFNA individually
presents a meaningful opportunity for health risk reductions for persons served by PWSs.
The agency is deferring the final individual regulatory determination for PFBS to further
consider whether occurrence information supports a finding that there is substantial likelihood
that PFBS will individually occur in PWSs and at a level of public health concern. Therefore, no
individual MCLG for PFBS is being established at this time. However, when evaluating PFBS in
mixture combinations with PFHxS, PFNA, and/or HFPO-DA, the EPA has determined that
based on the best available information, PFBS does meet all three statutory criteria for regulation
when a part of these mixtures, including that it is anticipated to have dose-additive adverse health
effects; there is a substantial likelihood of its co-occurrence in combinations with PFHxS, PFNA,
and/or HFPO-DA with a frequency and at levels of public health concern; and that there is a
meaningful opportunity for health risk reduction by regulating mixture combinations of these
four PFAS (USEPA, 2023c). Therefore, although the agency is deferring the individual final
regulatory determination for PFBS, PFBS is included in the final determination to regulate
mixture combinations containing two or more of PFHxS, PFNA, HFPO-DA, and PFBS. The
1-17
-------
FINAL
2024
establishment of individual MCLGs for PFHxS, PFNA, and HFPO-DA as well as an HI MCLG
for mixtures of two or more of PFHxS, PFNA, HFPO-DA, and PFBS addresses potential health
risks related to individual PFAS exposure as well as dose additive adverse health effects from
exposure to mixtures of two or more of these four PFAS.
1.6 Overview of Individual MCLG and Mixture Hazard Index (HI)
MCLG Approaches
To establish an MCLG for an individual contaminant, the EPA assesses the peer-reviewed
science examining cancer and noncancer health effects associated with oral exposure to the
contaminant. For contaminants determined to be known or likely human carcinogens (USEPA,
2005) with a linear carcinogenic MOA (i.e., where there is a proportional relationship between
dose and carcinogenicity at low concentrations) or for which there is insufficient information to
determine that a carcinogen has a threshold dose below which no carcinogenic effects have been
observed, the EPA has a longstanding practice of establishing the MCLG at zero (see USEPA
(1998); USEPA (2000c); USEPA (2001); see S. Rep. No. 169, 104th Cong., 1st Sess. (1995) at
3). For contaminants determined to be known or likely human carcinogens but with a nonlinear
carcinogenic MOA,4 contaminants that are designated as having suggestive evidence of
carcinogenic potential in humans (USEPA, 2005), and noncarcinogenic contaminants, the EPA
typically establishes the MCLG based on a noncancer toxicity reference value (RfV) that
represents the best available science (e.g., EPA RfD or ATSDR MRL).
An MCLG that is based on noncancer effects is designed to be protective of noncancer effects
over a lifetime of exposure with an adequate margin of safety, including for sensitive populations
and life stages, consistent with SDWA 1412(b)(3)(C)(i)(V) and 1412(b)(4)(A). The inputs for a
noncancer MCLG include an oral noncancer toxicity RfV (e.g., RfD or MRL), body weight-
adjusted drinking water intake (DWI-BW), and a relative source contribution (RSC), as
presented in Equation 1:
MCLG = (°ral RfV) * RSC (Eqn. 1)
V DWI-BW) v M '
Where:
RfV = Chronic toxicity reference value (EPA RfD or ATSDR MRL). An RfD is an
estimate of a daily exposure to the human population (including sensitive populations)
that is likely to be without an appreciable risk of deleterious effects during a lifetime
(USEPA, 2002). An MRL is an estimate of the daily human exposure to a hazardous
substance that is likely to be without appreciable risk of adverse noncancer health effects
over a specified duration of exposure (ATSDR, 2021).
DWI-BW = Body weight-adjusted drinking water intake - an exposure factor for the
90th percentile body weight-adjusted drinking water intake for the identified population
or life stage, in units of liters of water consumed per kilogram body weight per day
4 A carcinogen with a nonlinear MOA is a chemical agent for which the associated cancer response does not
increase in direct proportion to the exposure level and for which there is scientific evidence demonstrating a
threshold level of exposure below which there is no appreciable cancer risk (USEPA, 2005).
1-18
-------
FINAL
2024
(L/kg/day). The DWI-BW considers both direct and indirect consumption of drinking
water (indirect water consumption encompasses water added in the preparation of foods
or beverages, such as tea or coffee). Chapter 3 of the EPA's Exposure Factors Handbook
(USEPA, 2019b) provides the most up-to-date DWI-BWs for various populations or life
stages within the U.S. general population based on publicly available, peer-reviewed data
such as from the National Health and Nutrition Examination Survey (NHANES).
RSC = Relative source contribution - the percentage of the total exposure attributed to
drinking water sources (USEPA, 2000b), with the remainder of the exposure allocated to
all other routes or sources. The purpose of the RSC is to ensure that the level of a
contaminant allowed by one criterion (e.g., MCLG), when combined with other identified
sources of exposure common to the population and contaminant of concern, will not
result in exposures that exceed the RfD. The RSC is derived by applying the Exposure
Decision Tree approach (USEPA, 2000b).
The EPA's approach to DWI-BW selection includes a step to identify the sensitive population(s)
or life stage(s) (i.e., those that may be more susceptible or sensitive to a chemical exposure) by
considering the available data for the contaminant, including the adverse health effects observed
in the toxicity study on which the RfD/minimal risk level was based (known as the critical effect
within the critical or principal study). Although data gaps can complicate identification of the
most sensitive population (e.g., not all windows or life stages of exposure and/or health
outcomes may have been assessed in available studies), the critical effect and point of departure
(POD) that form the basis for the RfD (or minimal risk level) can provide some information
about sensitive populations because the critical effect is typically observed at the lowest tested
dose among the available data. Evaluation of the critical study, including the exposure window,
may identify a sensitive population or life stage (e.g., pregnant women, formula-fed infants,
lactating women). In such cases, the EPA can select the corresponding DWI-BW for that
sensitive population or life stage from the Exposure Factors Handbook (USEPA, 2019b). DWI-
BWs in the Exposure Factors Handbook are based on information from publicly available, peer-
reviewed studies, and were updated in 2019. In the absence of information indicating a sensitive
population or life stage, the DWI-BW corresponding to the general population may be selected.
Following this approach, the EPA selected appropriate DWI-BWs for each of the four PFAS
included in the HI MCLG (see Section 2). The EPA did consider infants as a sensitive life stage
for all four PFAS; however, the agency did not select the infant DWI-BW because the exposure
intervals of the critical studies supporting the chronic toxicity reference values did not
correspond to infants. Instead, the exposure intervals were relevant to other sensitive target
populations (i.e., lactating women or women of childbearing age) or the general population.
The EPA applies an RSC to account for potential aggregate risk from exposure routes and
exposure pathways other than oral ingestion of drinking water to ensure that an individual's total
exposure to a contaminant does not exceed the daily exposure associated with toxicity (i.e.,
threshold level or RfD). Application of the RSC in this context is consistent with EPA methods
(USEPA, 2000b) and long-standing EPA practice for establishing drinking water MCLGs and
NPDWRs (e.g., see USEPA, 2010, 2004; USEPA, 1989). The RSC represents the proportion of
an individual's total exposure to a contaminant that is attributed to drinking water ingestion
(directly or indirectly in beverages like coffee, tea, or soup, as well as from dietary items
prepared with drinking water) relative to other exposure pathways. The remainder of the
exposure equal to the RfD (or MRL) is allocated to other potential exposure sources (USEPA,
1-19
-------
FINAL
2024
2000b). The purpose of the RSC is to ensure that the level of a contaminant (e.g., MCLG) in
drinking water, when combined with other identified potential sources of exposure for the
population of concern, will not result in total exposures that exceed the RfD (or MRL) (USEPA,
2000b). This ensures that the MCLG under SDWA meets the statutory requirement that it is a
level of a contaminant in drinking water at or below which no known or anticipated adverse
effects on human health occur and allowing an adequate margin of safety.
To determine the RSCs for the four PFAS, the agency assessed the available scientific literature
on potential sources of human exposure other than drinking water (see Section 2). The EPA
conducted literature searches and reviews for each of the four PFAS to identify potential sources
of exposure and physicochemical properties that may influence occurrence in environmental
media (see Appendices A-D). Considering this exposure information, the EPA followed its
longstanding, peer-reviewed Exposure Decision Tree Approach in EPA's Methodology for
Deriving Ambient Water Quality Criteria for the Protection of Human Health (USEPA, 2000b)
to determine the RSC for each PFAS. The EPA carefully evaluated studies that included
information on potential exposure to these four PFAS (PFHxS, PFNA, HFPO-DA, and PFBS)
via sources other than drinking water, such as food, soil, sediment, and air. For each of the four
PFAS, the findings indicated that there are significant known or potential uses/sources of
exposure beyond drinking water ingestion (e.g., food, indoor dust) (USEPA, 2000b), but that
data are insufficient to allow for quantitative characterization of the different exposure sources
(Box 8A in USEPA, 2000b). The EPA's Exposure Decision Tree approach states that when there
are insufficient environmental and/or exposure data to permit quantitative derivation of the RSC,
the recommended RSC for the general population is 20 percent (Box 8B in USEPA, 2000b). This
means that 20 percent of the exposure equal to the RfD is allocated to drinking water, and the
remaining 80 percent is attributed to all other potential exposure sources.
In the general HI approach, a HQ is calculated as the ratio of human exposure (E) to a health-
based RfV for each mixture component chemical (i) (USEPA, 1986). The HI involves the use of
RfVs for each PFAS mixture component (in this case, PFHxS, HFPO-DA, PFNA, and/or PFBS)
which are based on the most sensitive known health outcomes and which are expected to be
protective of all other adverse health effects observed after exposure to the individual PFAS.
This approach, which protects against all adverse effects and not just a single adverse
outcome/effect, is a conservative risk indicator and appropriate for MCLG derivation. The HI is
dimensionless, so in the HI formula, E and the RfV must be in the same units (Equation 2). For
example, if E is the oral intake rate (milligrams per kilogram per day (mg/kg/day)), then the RfV
could be the RfD or MRL, which have the same units. Alternatively, the exposure metric can be
a media-specific metric such as a measured water concentration (e.g., nanograms per liter or
ng/L) and the RfV can be an HBWC (e.g., ng/L). The component chemical HQs are then
summed across the mixture to yield the HI (Equation 2). A mixture HI exceeding 1 indicates
potential risk for a given environmental medium or site. The HI provides an indication of: (1)
concern for the overall mixture and (2) potential driver PFAS (i.e., those PFAS mixture
components with high(er) HQs). For a detailed discussion of PFAS dose additivity and the HI
approach, see the PFAS Mixtures Framework (USEPA, 2024a).
n n E
(Eqn. 2)
1-20
-------
FINAL
2024
Where:
HI = Hazard index
HQ, = Hazard quotient for chemical i
Ei = Exposure, i.e., dose (mg/kg/day) or occurrence concentration, such as in drinking
water (in milligrams per liter or mg/L), for chemical i
RfVi = Reference value (e.g., oral RfD or MRL) (mg/kg/day), or corresponding HBWC;
e.g., an MCLG for chemical i (in mg/L)
The HBWCs/MCLGs are based on the best available science and data collected by accepted
methods (see Section III in the USEPA, 2023c). Specifically, peer-reviewed, publicly available
toxicity assessments are available for HFPO-DA (USEPA, 2021c), PFBS (USEPA, 2021d),
PFNA (ATSDR, 2021), and PFHxS (ATSDR, 2021) that provide the oral toxicity values (i.e.,
RfD or MRL) used to calculate the HBWCs; the EPA selected the corresponding DWI-BW for
the relevant sensitive population or life stage from the Exposure Factors Handbook (USEPA,
2019b) based on the best available, peer-reviewed science from publicly available, peer-
reviewed studies taking into account the relevant sensitive population(s) or life stage(s); and the
RSCs are based on the best available, peer-reviewed science or best available methods taking
into account the relevant sensitive population(s) or life stage(s) (USEPA, 2000b).
1-21
-------
FINAL
2024
2 Calculating the Health-Based Water
Concentrations for HFPO-DA, PFBS, PFNA, and
PFHxS for the HI MCLG
2.1 HFPO-DA
HFPO-DA is a shorter-chain PFAS that was intended to be a replacement for the longer-chained
PFOA. In water, HFPO-DA and its various salts dissociate to form the HFPO-DA anion
(HFPO—) as a common analyte.
The HBWC for HFPO-DA that the agency is using in the HI MCLG was derived in part from
information from the agency's 2021 human health toxicity assessment of HFPO-DA (specifically
the chronic RfD of 3E-06 mg/kg/day) (USEPA, 2021c). Summaries of key information from the
HFPO-DA toxicity assessment of (i.e., information about the RfD), as well as information about
the DWI-BW and RSC that were used to derive the HBWC for HFPO-DA, are presented in the
following sections. Based on this information, an HBWC of 10 ng/L for HFPO-DA is used in the
HI MCLG for mixtures of two or more of HFPO-DA, PFBS, PFNA, and PFHxS (see Section
3.2).
2.1.1 Toxicity
The HBWC for HFPO-DA is derived from a chronic RfD that is based on liver effects
(specifically, a constellation of liver lesions including cytoplasmic alteration, single-cell and
focal necrosis, and apoptosis) observed in parental female mice following oral exposure to
HFPO-DA from pre-mating through lactation (53-64 days) (USEPA, 2021c).
As described in the EPA's human health toxicity assessment of HFPO-DA, oral toxicity studies
in rodents exposed to HFPO-DA report a range of adverse effects. Repeated-dose oral exposure
of rats and mice resulted in liver toxicity (e.g., increased relative liver weight, hepatocellular
hypertrophy, apoptosis, and single-cell/focal necrosis), kidney toxicity (e.g., increased relative
kidney weight), immune system effects (e.g., antibody suppression), hematological effects (e.g.,
decreased red blood cell count, hemoglobin, and hematocrit), reproductive/developmental effects
(e.g., increased number of early deliveries, placental lesions, changes in maternal gestational
weight gain, and delays in genital development in offspring), and cancer (e.g., liver and
pancreatic tumors) (USEPA, 2021c).
The most sensitive noncancer effects observed among the available data were the adverse effects
on liver, which were observed in both male and female mice and rats across a range of exposure
durations and dose levels, including the lowest tested dose levels and shortest exposure durations
(USEPA, 2021c). Noncancer liver effects formed the basis for the chronic RfD of 3E-06
mg/kg/day, which the EPA used to derive the HBWC for HFPO-DA. As described in the HFPO-
DA toxicity assessment, to develop the chronic RfD for HFPO-DA, the EPA derived a human
equivalent dose (HED) of 0.01 mg/kg/day from a no-observed-adverse-effect level (NOAEL) of
0.1 mg/kg/day for liver effects observed in the identified critical study (an oral
2-1
-------
FINAL
2024
reproductive/developmental toxicity study in mice (DuPont, 2010b)). The EPA then applied a
composite uncertainty factor (UF) of 3,000 (i.e., 10x for intraspecies variability (UFh), 3x for
interspecies differences (UFa), 10x for extrapolation from a subchronic-to-chronic dosing
duration (UFs), and 10x for database deficiencies (UFd)) to yield the chronic RfD (USEPA,
2021c).
The EPA determined that there is suggestive evidence of carcinogenic potential following oral
exposure to HFPO-DA in humans, but the available data are insufficient to derive a cancer risk
concentration in water for HFPO-DA (USEPA, 2021c).
2.1.2 Exposure Factor
To select an appropriate DWI-BW for use in derivation of the HBWC for HFPO-DA, the EPA
considered the HFPO-DA exposure interval used in the oral reproductive/developmental toxicity
study in mice that was the basis for chronic RfD derivation (the critical study). In this study,
parental female mice were dosed from pre-mating through lactation, corresponding to three
potentially sensitive human adult life stages that may represent critical windows of exposure for
HFPO-DA: women of childbearing age (13 to < 50 years), pregnant women, and lactating
women (Table 3-63 in USEPA, 2019b). Of these three, the DWI-BW for lactating women
(0.0469 L/kg/day) is the highest (see Table 2-1), and therefore anticipated to be protective of the
other two sensitive life stages. Therefore, the EPA used the DWI-BW for lactating women to
calculate the HBWC for HFPO-DA in the HI MCLG.
Table 2-1. EPA Exposure Factors for Drinking Water Intake for Different Candidate
Sensitive Populations or Life Stages, Based on the Critical Effect and Study for HFPO-DA.
Population
DWI-BW
(L/kg bw-day)
Description of Exposure
Metric
Source
Women of
childbearing age
0.0354
90th percentile direct and
indirect consumption of
community water, consumer-
only two-day average, 13 to
<50 years.
2019 Exposure Factors
Handbook Chapter 3,
Table 3-63, NHANES
2005-2010 (USEPA,
2019b)
Pregnant women
0.0333
90th percentile direct and
indirect consumption of
community water, consumer-
only two-day average.
2019 Exposure Factors
Handbook Chapter 3,
Table 3-63, NHANES
2005-2010 (USEPA,
2019b)
Lactating women
0.0469
90th percentile direct and
indirect consumption of
community water, consumer-
only two-day average.
2019 Exposure Factors
Handbook Chapter 3,
Table 3-63, NHANES
2005-2010a (USEPA,
2019b)
2-2
-------
FINAL
2024
Notes'. L/kg bw-day = liters of water consumed per kilogram body weight per day; DWI-BW = body weight-adjusted drinking
water intake; NHANES = National Health and Nutrition Examination Survey. The DWI-BW used to calculate the HFPO-DA
HBWC is in bold.
a Estimates are less statistically reliable based on guidance published in the Joint Policy on Variance Estimation and Statistical
Reporting Standards on NHANES III and CSFII Reports: HNIS/NCHS Analytical Working Group Recommendations (NCHS,
1993).
The EPA conducted literature searches and reviews for HFPO-DA to identify potential sources
of exposure and physicochemical properties that may influence occurrence in environmental
media. Based on the physical properties, detected levels, and limited available exposure
information for HFPO-DA, multiple non-drinking water sources (e.g., foods, indoor dust, air,
soil, and sediment) are potential exposure sources (see Appendix A). Following the EPA's
Exposure Decision Tree approach (USEPA, 2000b), potential sources other than drinking water
ingestion were identified, but the available information is limited and does not allow for the
quantitative characterization of the relative levels of exposure among these different sources.
Thus, the EPA used an RSC of 0.20 for HFPO-DA. This means that 20% of the RfD is attributed
to drinking water, and the remaining 80% is attributed to all other potential exposure sources. As
explained above (Section 1.6), applying the RSC ensures that an individual's total exposure to
HFPO-DA from all sources does not exceed the daily exposure associated with the RfD,
consistent with agency methodology (USEPA, 2000b) and longstanding practice for establishing
drinking water MCLGs and NPDWRs.
2.1.3 Relative Source Contribution
2.1.4 Derivation of HFPO-DA HBWC
The HBWC for HFPO-DA is calculated as follows and summarized in Table 2-2:
= 10 — or parts per trillion (ppt)
Li
2-3
-------
FINAL
2024
Table 2-2. HFPO-DA HBWC - Input Parameters and Value
Parameter Value Units Source
Final RfD based on liver effects (constellation of liver
lesions as defined by the NTP Pathology Working Group) in
parental female mice exposed to HFPO-DA by gavage from
pre-mating through lactation (53-64 days) (USEPA, 2021c;
DuPont, 2010b).
90th percentile 2-day average, consumer-only estimate of
t,.,7T , /, ,1 combined direct and indirect community water ingestion for
DWI-BW 0.0469 L/kg/day , . .. , 10+ ^ , ° c -mr.
° lactating women (age 13 to <50 years) based on 2005-2010
NHANES data (USEPA, 2019b).
Based on a review of the available scientific literature on
HFPO-DA, potential exposure routes and sources exist but
RSC 0.2 N/A the available information is limited and does not allow for
the quantitative characterization of the relative levels of
exposure among these different sources (see Appendix A).
HFPO-DA HBWC = 0.00001 mg/L or 10 ppt
Notes: RfD = reference dose; DWI-BW = body weight-adjusted drinking water intake; HBWC = health-based water
concentration; HFPO-DA = hexafluoropropylene oxide dimer acid; N/A = not applicable; NHANES = National Health and
Nutrition Examination Survey; NTP = National Toxicology Program; RSC = relative source contribution.
oral°RfD 3E"°6
2.2 PFBS
PFBS and its potassium salt (K+PFBS) are shorter-chain PFAS that were developed as "safer"
replacements for the longer-chained PFOS (USEPA, 2021d). In water, K+PFBS dissociates to the
deprotonated anionic form of PFBS (PFBS-) and the K+ cation at environmental pH levels (pH
4-9). These three PFBS chemical forms are referred to collectively as PFBS.
The HBWC that the agency is using for the HI MCLG was derived from information in the
agency's 2021 human health toxicity assessment for PFBS, specifically the chronic RfD of 3E-
04 mg/kg/day based on thyroid effects observed in newborn mice born to mothers that had been
orally exposed to PFBS throughout gestation (USEPA, 2021d). Summaries of key information
from the PFBS toxicity assessment (i.e., information about the RfD), as well as information
about the DWI-BW and RSC that were used to derive the HBWC for PFBS, are presented in the
following sections. Based on this information, an HBWC of 2,000 ng/L for PFBS is used in the
HI MCLG for mixtures of two or more of HFPO-DA, PFBS, PFNA, and PFHxS (see Section
3.2).
2.2.1 Toxicity
The HBWC for PFBS was derived using a chronic oral RfD based on thyroid effects seen in an
oral toxicity study in mice (USEPA, 2021d).
The EPA's human health toxicity assessment for PFBS (USEPA, 202Id) considered all publicly
available human, animal, and mechanistic studies of PFBS exposure and effects. The assessment
2-4
-------
FINAL
2024
identified associations between PFBS exposure and thyroid, developmental, and kidney effects
based on studies in animals. The limited evidence for thyroid or kidney effects in human studies
was equivocal, and no studies evaluating developmental effects of PFBS in humans were
available. Human and animal studies evaluated other health effects following PFBS exposure
including effects on the reproductive system, liver, and lipid and lipoprotein homeostasis, but the
evidence did not support clear associations between exposure and effect (USEPA, 202Id).
The most sensitive noncancer effect observed was an adverse effect on the thyroid (i.e.,
decreased serum total thyroxine) seen in newborn mice (postnatal day (PND) 1) born to mothers
that had been orally exposed to K+PFBS throughout gestation (USEPA, 2021d; Feng et al.,
2017). This critical effect was the basis for the chronic RfD of 3E-04 mg/kg/day which the EPA
used to derive the HBWC for PFBS (USEPA, 2021d). As described in the PFBS toxicity
assessment, to develop the chronic RfD for PFBS,5 the EPA derived an HED of 0.095 mg/kg/day
from BMD modeling of the critical effect in mice. The EPA then applied a composite UF of 300
(i.e., 10x for UFh, 3x for UFa, and 10x for UFd) to yield the chronic RfD (USEPA, 2021d). The
EPA did not apply an additional UF to adjust for subchronic-to-chronic duration (i.e., UFs)
because the critical effects were observed during a developmental life stage6 (USEPA, 2002).
There were no human or animal studies identified that evaluated the potential carcinogenicity of
PFBS (USEPA, 202Id).
2.2.2 Exposure Factor
To select an appropriate DWI-BW for use in deriving the HBWC, the EPA considered the PFBS
exposure interval used in the developmental toxicity study in mice that was the basis for chronic
RfD derivation. In this study, pregnant mice were exposed throughout gestation, which is
relevant to two human adult life stages: women of childbearing age (13 to < 50 years) who may
be or become pregnant, and pregnant women and their developing embryo or fetus (Table 3-63
in USEPA, 2019b). Of these two, the EPA selected the DWI-BW for women of childbearing age
(0.0354 L/kg/day) to derive the HBWC for PFBS because it is higher and therefore more health
protective (see Table 2-3).
5 Data for K+PFBS were used to derive the chronic RfD for the free acid (PFBS), resulting in the same value (3E-
04 mg/kg/day), after adjusting for differences in molecular weight between K+ PFBS (338.19) and PFBS (300.10)
(USEPA, 202Id).
6 As stated in USEPA (2002), ".. .This is because it is assumed that most endpoints of developmental toxicity can be
caused by a single exposure. If, however, developmental effects are more sensitive than those seen after longer-term
exposures, then even the chronic RfD/RfC should be based on such effects to reduce the risk of potential greater
sensitivity in children. Because the standard studies currently conducted for developmental toxicity involve repeated
exposures, data are not often available on which endpoints may be induced by acute, subacute, subchronic, or
chronic dosing regimens and, therefore, on which should be used in setting various duration reference values."
2-5
-------
FINAL
2024
Table 2-3. EPA Exposure Factors for Drinking Water Intake for Different Candidate
Sensitive Populations or Life Stages, Based on the Critical Effect and Study for PFBS.
Population
DWI-BW
(L/kg bw-day)
Description of Exposure
Metric
Source
Women of
childbearing age
0.0354
90th percentile direct and
indirect consumption of
community water, consumer-
only two-day average, 13 to
<50 years.
2019 Exposure Factors
Handbook Chapter 3,
Table 3-63, NHANES
2005-2010 (USEPA,
2019b)
Pregnant women
0.0333
90th percentile direct and
indirect consumption of
community water, consumer-
only two-day average.
2019 Exposure Factors
Handbook Chapter 3,
Table 3-63, NHANES
2005-2010 (USEPA,
2019b)
Notes'. L/kg bw-day = liters of water consumed per kilogram body weight per day; DWI-BW = body weight-adjusted drinking
water intake; NHANES = National Health and Nutrition Examination Survey. The DWI-BW used to calculate the PFBS HBWC
is in bold.
2.2.3 Relative Source Contribution
The EPA conducted literature searches and reviews for PFBS to identify potential sources of
exposure and physicochemical properties that may influence occurrence in environmental media.
Based on the physical properties, detected levels, and available exposure information for PFBS,
multiple non-drinking water sources (seafood including fish and shellfish, other foods, indoor
air, and some consumer products) are potentially significant exposure sources (see Appendix B).
Following the EPA's Exposure Decision Tree approach (USEPA, 2000b), potential sources other
than drinking water ingestion were identified, but the available information is limited and does
not allow for the quantitative characterization of the relative levels of exposure among these
different sources. Thus, the EPA used an RSC of 0.20 for PFBS. This means that 20% of the RfD
is attributed to drinking water, and the remaining 80% is attributed to all other potential exposure
sources. As explained above (Section 1.6), applying the RSC ensures that an individual's total
exposure to PFBS from all sources does not exceed the daily exposure associated with the RfD,
consistent with agency methodology (USEPA, 2000b) and longstanding practice for establishing
drinking water MCLGs and NPDWRs.
2.2.4 Derivation of PFBS HB WC
The HBWC for PFBS is calculated as follows and summarized in Table 2-4:
/ RfD \
PFBS HBWC = -7777-777 * RSC
VDWI-BW /
/0.0003 1 \
= U0.2
V°-0354ii7dW
2-6
-------
FINAL
2024
mg / mg\
= 0.0017 —— (rounded to 0.002 ——J
Lj ^ Lj '
= 2,000 — or ppt
Li
Table 2-4. PFBS HBWC - Input Parameters and Value
Parameter
Value
Units
Source
Chronic
RfD
3E-04
Final RfD based on critical effect of decreased serum total
mg/kg/day thyroxine in newborn mice after gestational exposure
(USEPA, 202Id; Feng et al., 2017).
DWI-BW
0.0354
L/kg/day
90th percentile 2-day average, consumer-only estimate of
combined direct and indirect community water ingestion for
women of childbearing age (13 to <50 years) based on
2005-2010 NHANES data (USEPA, 2019b).
RSC
0.2
N/A
Based on a review of the available scientific literature on
PFBS, potential exposure routes and sources exist but the
available information is limited and does not allow for the
quantitative characterization of the relative levels of exposure
among these different sources (see Appendix B).
PFBS HBWC = 0.002 mg/L or 2,000 ppt
Note: RiD = reference dose; DWI-BW = body weight-adjusted drinking water intake; N/A = not applicable;
NHANES = National Health and Nutrition Examination Survey; HBWC = health-based water concentration;
PFBS = perfluorobutanesulfonic acid; RSC = relative source contribution.
23 PFNA
PFNA has been used as a processing aid in the production of fluoropolymers, primarily
polyvinylidene fluoride, which is a plastic designed to be temperature resistant and chemically
nonreactive (USEPA, 2020b; NJSWQI, 2017; Prevedouros et al., 2006). PFNA has been used
since the 1950's in a wide variety of industrial and consumer products. It has also been used in
aqueous film-forming foam (AFFF) for fire suppression (USEPA, 2020b; Laitinen et al., 2014).
ATSDR has published a toxicological profile for a group of PFAS including PFNA and has
developed an intermediate-duration oral MRL for PFNA (ATSDR, 2021). The EPA's derived
HBWC for PFNA (described below) is based on the ATSDR MRL (ATSDR, 2021), a DWI-BW
(selected by the EPA) that corresponds to this MRL, and an RSC determined by the EPA. There
is no published EPA human health toxicity assessment for PFNA at the time of this writing;
however, the EPA's Integrated Risk Information System (IRIS) program is developing a human
health toxicity assessment for PFNA, which is expected to be finalized in 2024 (USEPA, 2023a,
202le). The EPA's IRIS assessment will use systematic review methods to evaluate the
2-7
-------
FINAL
2024
epidemiological and toxicological literature for PFNA, including consideration of relevant
mechanistic evidence (USEPA, 202le).
2.3.1 Toxicity
The HBWC for PFNA is based on an ATSDR intermediate-duration oral MRL that was based on
developmental effects seen in mice after oral PFNA exposure (ATSDR, 2021).
Studies of oral PFNA exposure in rodents have reported adverse effects on the liver,
development, and reproductive and immune systems (ATSDR, 2021). The most sensitive
noncancer effects and basis for the ATSDR intermediate-duration oral MRL were decreased
body weight gain and impaired development (i.e., delayed eye opening, preputial separation, and
vaginal opening) in mice born to mothers that were treated with PFNA from gestational days
(GDs) 1-17 (with presumed continued indirect exposure of offspring via lactation), and
monitoring until PND 287 (ATSDR, 2021). The study reporting these effects (Das et al., 2015)
was selected by ATSDR as the principal study for MRL derivation. To derive the MRL, an HED
of 0.001 mg/kg/day was calculated from the NOAEL of 1 mg/kg/day identified in the study.
ATSDR applied a total UF of 30 (i.e., 10x for UFh and 3 x for UFa) and a modifying factor (MF)
of 10x for database deficiencies to account for the small number/limited scope of studies
examining PFNA toxicity following intermediate-duration exposure. The resulting intermediate-
duration oral MRL was 3E-06 mg/kg/day (ATSDR, 2021). The EPA did not apply an additional
UFs to calculate the HBWC because the critical effects were observed during a developmental
life stage6 (USEPA, 2002). Toxicological assessments from other sources (e.g., states) report
toxicity reference values based on animal studies for PFNA that are in the same range, providing
additional support (USEPA, 2021e).
The carcinogenic potential of PFNA has been examined in three epidemiological studies. No
consistent associations between serum PFNA levels and breast cancer or prostate cancer were
found (ATSDR, 2021). The EPA has not yet completed a final evaluation and classification of
the carcinogenicity of PFNA.
2.3.2 Exposure Factor
Based on the life stages of exposure in the principal study from which the intermediate-duration
MRL was derived (i.e., directly to maternal animals during gestation, and indirectly to offspring
during gestation and lactation), the EPA identified three potentially sensitive life stages that may
represent critical windows of exposure for PFNA: women of childbearing age (13 to <50 years),
pregnant women, and lactating women (Table 3-63 in USEPA, 2019b). The DWI-BW for
lactating women (0.0469 L/kg/day; 90th percentile direct and indirect consumption of
community water, consumer-only 2-day average) was selected to calculate the HBWC for PFNA
because it is the highest of the three DWI-BWs and is anticipated to be protective of the other
two sensitive life stages (see Table 2-5).
2-8
-------
FINAL
2024
Table 2-5. EPA Exposure Factors for Drinking Water Intake for Different Candidate
Sensitive Populations and Life Stages, Based on the Critical Effect and Study for PFNA.
Population
DWI-BW
(L/kg bw-day)
Description of Exposure
Metric
Source
Women of
childbearing age
0.0354
90th percentile direct and
indirect consumption of
community water, consumer-
only two-day average, 13 to
<50 years.
2019 Exposure Factors
Handbook Chapter 3,
Table 3-63, NHANES
2005-2010 (USEPA,
2019b)
Pregnant women
0.0333
90th percentile direct and
indirect consumption of
community water, consumer-
only two-day average.
2019 Exposure Factors
Handbook Chapter 3,
Table 3-63, NHANES
2005-2010 (USEPA,
2019b)
Lactating women
0.0469
90th percentile direct and
indirect consumption of
community water, consumer-
only two-day average.
2019 Exposure Factors
Handbook Chapter 3,
Table 3-63, NHANES
2005-2010a (USEPA,
2019b)
Notes'. L/kg bw-day = liters of water consumed per kilogram body weight per day; DWI-BW = body weight-adjusted drinking
water intake; NHANES = National Health and Nutrition Examination Survey. The DWI-BW used to calculate the PFNA
HBWC is in bold.
2.3.3 Relative Source Contribution
The EPA conducted literature searches and reviews for PFNA to identify potential sources of
exposure and physicochemical properties that may influence occurrence in environmental media.
Based on the physical properties, detected levels, and available exposure information for PFNA,
multiple non-drinking water sources (fish and shellfish, non-fish food, some consumer products,
indoor dust, and air) are potentially significant exposure sources (see Appendix C). Following
the Exposure Decision Tree approach (USEPA, 2000b), potential sources other than drinking
water ingestion were identified but the available information is limited and does not allow for the
quantitative characterization of the relative levels of exposure among these different sources.
Thus, the EPA used an RSC of 0.20 for PFNA. This means that 20% of the RfV is attributed to
drinking water, and the remaining 80% is attributed to all other potential exposure sources. As
explained above (Section 1.6), applying the RSC ensures that an individual's total exposure to
PFNA from all sources does not exceed the daily exposure associated with the RfV, consistent
with agency methodology (USEPA, 2000b) and longstanding practice for establishing drinking
water MCLGs and NPDWRs.
2.3.4 Derivation of PFNA HBWC
The HBWC for PFNA is calculated as follows and summarized in Table 2-6:
2-9
-------
FINAL
2024
PFNA HBWC
/0.000003
VDWI-BW/
DWI-BW
mg X
kg/day \
* RSC
\ 0.0469-j—It— /
\ kg/day /
* 0.2
mg / mg\
= 0.000014 -p (rounded to 0.00001 -p)
J_j ^ J_j '
= 0.01 ^
Li
ng
= 10 — or ppt
Li
Table 2-6. PFNA HBWC - Input Parameters and Value
Parameter
Value
Units
Source
RfV
3E-06a
mg/kg/day
Based on decreased body weight gain and delayed eye
opening, preputial separation, and vaginal opening in
mouse offspring after gestational and presumed
lactational exposure (ATSDR, 2021; Das et al., 2015).
DWI-BW
0.0469
L/kg/day
90th percentile 2-day average, consumer-only estimate
of combined direct and indirect community water
ingestion for lactating women (13 to <50 years) based
on 2005-2010 NHANES data (USEPA, 2019b).
RSC
0.2
N/A
Based on a review of the current scientific literature on
PFNA, potential exposure routes and sources exist but
the available information is limited and does not allow
for the quantitative characterization of the relative
levels of exposure among these different sources (see
Appendix C).
PFNA HBWC = 0.00001 mg/L or 10 ppt
Notes: RfV = chronic toxicity reference value; DWI-BW = body weight-adjusted drinking water intake; N/A = not applicable;
NHANES = National Health and Nutrition Examination Survey; PFNA = perfluorononanoic acid; RSC = relative source
contribution.
aNote that ATSDR MRLs and EPA RfDs are not identical (e.g., intermediate-duration MRL vs. chronic RfD; the EPA and
ATSDR may apply different uncertainty /modifying factors) and are developed for different purposes. In this case, the EPA did
not apply an additional UF s to calculate the HBWC for PFNA because the critical effect is identified in a developmental
population (USEPA, 2002).
2.4 PFHxS
PFHxS has been used in laboratory applications and as a raw material or a precursor for the
manufacture of PFAS/perfluoroalkyl sulfonate-based products, though production of PFHxS in
2-10
-------
FINAL
2024
the United States was phased out by its major manufacturer in 2002 (Sigma-Aldrich, 2014 as
cited in NCBI, 2022; Backe et al., 2013; Buck et al., 2011; OECD, 2011). PFHxS has also been
used in firefighting foam and carpet treatment solutions, and it has been used as a stain and water
repellant (Garcia and Harbison, 2015 as cited in NCBI, 2022).
ATSDR has published a toxicological profile for a group of PFAS including PFHxS and has
calculated an intermediate-duration oral MRL for PFHxS (ATSDR, 2021). The EPA's derived
HBWC for PFHxS (described below) is based on the ATSDR MRL (ATSDR, 2021), a DWI-BW
(selected by the EPA) that corresponds to the MRL, and an RSC determined by the EPA. There
is no published EPA human health toxicity assessment for PFHxS at the time of this writing;
however, the EPA's IRIS program is developing a human health toxicity assessment for PFHxS,
which is expected to be finalized in 2024 (USEPA, 2023a, 2021e). The EPA's IRIS assessment
will use systematic review methods to evaluate the epidemiological and toxicological literature
for PFHxS, including consideration of relevant mechanistic evidence (USEPA, 2021e).
2.4.1 Toxicity
The HBWC for PFHxS is derived using an ATSDR intermediate-duration oral MRL based on
thyroid effects seen in male rats after oral PFHxS exposure (ATSDR, 2021).
Toxicity studies of oral PFHxS exposure to animals have reported health effects on the liver,
thyroid, and development (ATSDR, 2021). The most sensitive noncancer effect observed was
thyroid follicular epithelial hypertrophy/hyperplasia in parental male rats that had been exposed
for 42-44 days, identified in the principal developmental toxicity study selected by ATSDR
(NOAEL of 1 mg/kg/day for this effect) (ATSDR, 2021; Butenhoff et al., 2009). This critical
effect was the basis for the ATSDR intermediate-duration oral MRL which the EPA used to
derive the HBWC for PFHxS. An HED of 0.0047 mg/kg/day was calculated from the NOAEL of
1 mg/kg/day identified in the principal study. ATSDR applied a total UF of 30 (i.e., 10x for UFh
and 3x for UFa) and an MF of 10x for database deficiencies to yield an intermediate-duration
oral MRL of 2E-05 mg/kg/day (ATSDR, 2021). To calculate the HBWC, the EPA applied an
additional UFs of 10, per agency guidelines (USEPA, 2002), because the effect is not in a
developmental population (i.e., thyroid follicular epithelial hypertrophy/hyperplasia in parental
male rats). The resulting adjusted chronic reference value is 2E-06 mg/kg/day. Toxicological
assessments from other sources (e.g., states) report toxicity reference values based on animal
studies for PFHxS that are in the same range, providing additional support (USEPA, 202 le).
The carcinogenic potential of PFHxS has been examined in four epidemiological studies
(ATSDR, 2021). Bonefeld-J0rgensen et al. (2014) reported a significant negative correlation
between serum PFHxS levels (mean concentration 1.2 ng/mL) and breast cancer risk among
Danish women. However, a study in Greenland found a significant, positive association between
high serum levels of PFHxS and breast cancer risk (Wiels0e et al., 2017). The median serum
PFHxS concentration among cases in that study was 2.52 ng/mL and serum levels ranged from
0.19 ng/mL to 23.40 ng/mL (Wiels0e et al., 2017). Hardell et al. (2014) found a statistically
significant interaction between above-median PFHxS concentrations and increased risk for
prostate cancer among men with genetics as a risk factor (first-degree relative). Prostate-specific
antigen (PSA) levels were not associated with serum PFHxS levels (mean concentration
3.38 ng/mL) in men 20-49 or 50-69 years of age (Ducatman et al., 2015). The EPA has not yet
completed a final evaluation and classification of the carcinogenicity of PFHxS.
2-11
-------
FINAL
2024
2.4.2 Exposure Factor
No sensitive population or life stage was identified for DWI-BW selection for PFHxS because
the critical effect on which the ATSDR MRL was based (thyroid alterations) was observed in
adult male rats. Since this exposure life stage does not correspond to a sensitive population or
life stage, a DWI-BW for adults within the general population (0.034 L/kg/day; 90th percentile
direct and indirect consumption of community water, consumer-only 2-day average, adults 21
years and older) was selected for HBWC derivation (USEPA, 2019b).
The EPA conducted literature searches and reviews for PFHxS to identify potential sources of
exposure and physicochemical properties that may influence occurrence in environmental media.
Based on the physical properties, detected levels, and available exposure information for PFHxS,
multiple non-drinking water sources (fish and shellfish, non-fish food, some consumer products,
indoor dust, and soil) are potentially significant exposure sources (see Appendix D). Following
the Exposure Decision Tree approach (USEPA, 2000b), potential sources other than drinking
water ingestion were identified but the available information is limited and does not allow for the
quantitative characterization of the relative levels of exposure among these different sources.
Thus, the EPA used an RSC of 0.20 for PFHxS. This means that 20% of the RfV is attributed to
drinking water, and the remaining 80% is attributed to all other potential exposure sources. As
explained above (Section 1.6), applying the RSC ensures that an individual's total exposure to
PFHxS from all sources does not exceed the daily exposure associated with the RfV, consistent
with agency methodology (USEPA, 2000b) and longstanding practice for establishing drinking
water MCLGs and NPDWRs.
2.4.3 Relative Source Contribution
2.4.4 Derivation of PFHxS HB WC
The HBWC for PFHxS is calculated as follows and summarized in Table 2-7:
= 0.000012 —2
J_j
(rounded to 0.00001
= 10 — or ppt
Li
2-12
-------
FINAL
2024
Table 2-7. PFHxS HBWC - Input Parameters and Value
Parameter Value Units
RfV 2E-06a mg/kg/day
DWI-BW 0.034 L/kg/day
RSC 0.2 N/A
Source
Based on thyroid follicular epithelial
hypertrophy/hyperplasia in parental male rats
(exposed 42-44 days) (ATSDR, 2021;
Butenhoff et al., 2009).
90th percentile 2-day average, consumer-only
estimate of combined direct and indirect
community water ingestion for adults
21 years and older based on 2005-2010
NHANES data (USEPA, 2019b).
Based on a review of the current scientific
literature on PFHxS, potential exposure
routes and sources exist but the available
information is limited and does not allow for
the quantitative characterization of the
relative levels of exposure among these
different sources (see Appendix D).
PFHxS HBWC = 0.00001 mg/L or 10 ppt
Notes: RfV = chronic toxicity reference value; DWI-BW = body weight-adjusted drinking water intake; N/A = not applicable;
NHANES = National Health and Nutrition Examination Survey; RSC = relative source contribution; PFHxS =
perfluorohexanesulfonic acid.
aNote that ATSDR MRLs and EPA RfDs are not identical (e.g., intermediate-duration MRL vs. chronic RfD; the EPA and
ATSDR may apply different uncertainty /modifying factors) and are developed for different purposes. The EPA applied an
additional UF of 10 to the ATSDR MRL for PFHxS to account for subchronic-to-chronic duration (i.e., UFs) yielding a chronic
reference value of 2E-06 mg/kg/day, which was used to calculate the HBWC for PFHxS (USEPA, 2002).
2-13
-------
FINAL
2024
3 Derivation of MCLGs
3.1 Individual MCLGs for HFPO-DA, PFNA, and PFHxS
The EPA is setting the individual MCLGs for HFPO-DA, PFHxS, and PFNA at 10 ng/L based
on their respective RfVs, DWI-BWs, and RSCs (see Section 2, specifically Sections 2.1, 2.3, and
2.4). The MCLG for each of these PFAS is set, as defined in Section 1412(b)(4)(A), at "the level
at which no known or anticipated adverse effects on the health of persons occur and which
allows an adequate margin of safety." Each of these MCLGs (shown in Table 3-1 below) is set at
the same level as the HBWCs derived in Section 2.
Table 3-1. Individual MCLGs
PFAS
MCLG (ng/L or ppt)
HFPO-DA
10
PFHxS
10
PFNA
10
Notes: HFPO-DA = hexafluoropropylene oxide dimer acid; MCLG = maximum contaminant level goal; PFAS = per- and
polyfluoroalkyl substances; PFHxS = perfluorohexanesulfonic acid; PFNA = perfluorononanoic acid; ng/L = nanograms per
liter; ppt = parts per trillion.
3.2 PFAS Mixtures Hazard Index MCLG
In consideration of the known toxic effects, potential dose additivity, and occurrence and likely
co-occurrence of these PFAS in drinking water, the EPA is finalizing an HI of 1 (unitless) as the
MCLG for any mixture of two or more of PFHxS, PFNA, HFPO-DA, and PFBS. As described in
Section 1.6, a mixture HI can be calculated when HBWCs for a set of PFAS are available or can
be calculated. HQs are calculated for each of the mixture components by dividing the measured
component PFAS concentration in water (e.g., expressed as ng/L) by the relevant HBWC (e.g.,
expressed as ng/L), as shown in the equation below. Component HQs are summed across the
PFAS mixture to yield the HI MCLG. A PFAS mixture HI MCLG greater than 1 (i.e., rounded to
one significant digit) indicates an exceedance of the health-protective level and indicates
potential human health risk for noncancer effects from oral exposure to the PFAS mixture in
water. For more details, please see USEPA (2024a). The final PFAS mixture HI MCLG for
mixtures of two or more of HFPO-DA, PFBS, PFNA, and PFHxS is calculated as follows:
HI MCLG = () + (J£E£h4) + + (j££E£»4) = ,
\[HFPO-DAhbwc\J \[PFBSmbWc\'' \\PFNAhbwc\J V[P FHxShbWc\J
hi MCLG = + (J££££«il) + + = !
V [10 ng/L] J \[2000 ng/L\J \ [10 ng/L] J \ [10 ng/L] J
Where:
3-1
-------
FINAL
2024
[PFASng/L] = the measured component PFAS concentration in water (in ng/L) and
[PFAShbwc] = the HBWC of a component PFAS.
In summary, although current weight of evidence suggests that PFAS vary in their precise
structure and function, exposure to different PFAS can result in the same or similar adverse
health effects; as a result, PFAS co-exposures are likely to result in dose-additive effects and
therefore the conclusion of dose additivity is appropriate (see Section 1.3 and also USEPA,
2024a). While individual PFAS can pose a potential risk to human health if the exposure level
exceeds the chemical-specific toxicity reference value (RfD or MRL) (i.e., individual PFAS
HQ >1.00), mixtures of PFAS at lower individual PFAS concentrations can result in dose-
additive adverse health effects. For example, if the individual HQs for PFHxS, HFPO-DA,
PFNA, and PFBS were each 0.90, that would indicate that the measured concentration of each
individual PFAS in drinking water is below the level of appreciable risk for each individual
PFAS (recall that an RfV, such as an oral RfD, represents an estimate at which no appreciable
risk of deleterious effects exists). However, the overall HI for that mixture would be 4 (i.e., 3.6,
sum of four HQs of 0.90, rounded to 4), indicating risk. Thus, setting an MCLG based on the
concentration of an individual PFAS without considering the potential dose-additive effects from
other PFAS in a mixture would not provide a sufficiently protective MCLG with an adequate
margin of safety. To account for dose-additive noncancer effects associated with co-occurring
PFAS to protect against health impacts from multi-chemical exposures to mixtures of two or
more of PFHxS, HFPO-DA, PFNA, and PFBS, the agency is finalizing use of the HI approach
for the MCLG for mixtures of two or more of these four PFAS. Consistent with the statutory
requirement under 1412(b)(4)(A) of SDWA, establishing the MCLG for mixtures of two or more
of PFHxS, HFPO-DA, PFNA, and PFBS at an HI equal to 1 ensures that the MCLG is set at a
level at which there are no known or anticipated adverse effects on the health of persons and
which ensures an adequate margin of safety.
3-2
-------
FINAL
2024
4 References
3M. (2007). Remedial Investigation Report. Phase 2. Fluorochemical (FC) Data Assessment
Report for the Cottage Grove, MN Site. St. Paul, MN: 3M Corporate Toxicology.
3M. (2012). Interim Report 36: Analysis of PFBS, PFHS, and PFOS in Groundwater Samples
Collected from Decatur, AL, 3rd Quarter 2012. GLP10-01-02-36. (Docket EPA-HQ-
2004-0112-0155). St. Paul, MN: 3M Environmental Health and Safety Operations.
Abdallah, MA; Wemken, N; Drage, DS; Tlustos, C; Cellarius, C; Cleere, K, et al. (2020).
Concentrations of perfluoroalkyl substances in human milk from Ireland: Implications for
adult and nursing infant exposure. Chemosphere 246: 125724.
https://doi.oi^ 10 101 '/j.chemosphert' _« l l _ _ I
Addicks, GC; Rowan-Carroll, A; Reardon, AJF; Leingartner, K; Williams, A; Meier, MJ, et al.
(2023). Per- and polyfluoroalkyl substances (PFAS) in mixtures show additive effects on
transcriptomic points of departure in human liver spheroids. Toxicological Sciences 194:
38-52. https://doi.ore/10.1093/toxsci/kfad044
Ahrens, L; Felizeter, S; Sturm, R; Xie, Z; Ebinghaus, R. (2009a). Polyfluorinated compounds in
waste water treatment plant effluents and surface waters along the River Elbe, Germany.
Marine Pollution Bulletin 58: 1326-1333.
http://dx.doi.org/10.1016/i .Marpotbut.2009.04.028
Ahrens, L; Plassmann, M; Xie, Z; Ebinghaus, R. (2009b). Determination of polyfluoroalkyl
compounds in water and suspended particulate matter in the river Elbe and North Sea,
Germany. Frontiers of Environmental Science & Engineering in China 3: 152-170.
http://dx.doi.on -009-0021-8
Ahrens, L; Shoeib, M; Harner, To; Lee, S; Guo, Ru; Reiner, EJ. (2011). Wastewater treatment
plant and landfills as sources of polyfluoroalkyl compounds to the atmosphere.
Environmental Science and Technology 45: 8098-8105.
http://dx.doi.org/10J021/eslO; I ;
Altenburger, R; Backhaus, T; Boedeker, W; Faust, M; Scholze, M; Grimme, LH. (2000).
Predictability of the toxicity of multiple chemical mixtures to Vibrio fischeri: Mixtures
composed of similarly acting chemicals. Environmental toxicology and chemistry 19:
2341-2347. htfps://doi.org/10 J 002/etc. 5620190926
Anderson, RH; Long, GC; Porter, RC; Anderson, JK. (2016). Occurrence of select perfluoroalkyl
substances at U.S. Air Force aqueous film-forming foam release sites other than fire-
training areas: Field-validation of critical fate and transport properties. Chemosphere 150:
678-685. https://doi.on 10 101 i.chemosphere..01 01 01 I
Antignac, JP; Veyrand, B; Kadar, H; Marchand, P; Oleko, A; Le Bizec, B; Vandentorren, S.
(2013). Occurrence of perfluorinated alkylated substances in breast milk of French
women and relation with socio-demographical and clinical parameters: results of the
ELFE pilot study. Chemosphere 91: 802-808.
https://d0i.0rg/l 0 101 '/i.chemosphert' .01 '< 01 088
Appleman, TD; Higgins, CP; Quinones, O; Vanderford, BJ; Kolstad, C; Zeigler-Holady, JC;
Dickenson, ER. (2014). Treatment of poly- and perfluoroalkyl substances in U.S. full-
scale water treatment systems. Water Research 51: 246-255.
http://dx.doi.org/10.1016/i.watres.20'1 < 10 067
4-1
-------
FINAL
2024
Arp, HP; Slinde, GA. (2018). PFBS in the Environment: Monitoring and Physical-Chemical
Data Related to the Environmental Distribution of Perfluorobutanesulfonic Acid. (Report
M-l 122|2018). Norwegian Environmental Agency.
https://www.mili odirektoratet.no/ globalassets/publikasi oner/M 1122/M1122.pdf
ASTSWMO (Association of State and Territorial Solid Waste Management Officials). (2015).
Perfluorinated Chemicals (PFCS): Perfluorooctanoic Acid (PFOA) & Perfluorooctane
Sulfonate (PFOS) Information Paper.
ATSDR (Agency for Toxic Substances and Disease Registry). (2021). Toxicological Profile for
Perfluoroalkyls. Atlanta, GA: U.S. Department of Health and Human Services, Public
Health Service, Agency for Toxic Substances and Disease Registry.
https://wwwn.cdc.gov/TSP/ToxProfiles/ToxProfiles.aspx7i id=237
Bach, C; Dauchy, X; Boiteux, V; Colin, A; Hemard, J; Sagres, V, et al. (2017). The impact of
two fluoropolymer manufacturing facilities on downstream contamination of a river and
drinking water resources with per- and polyfluoroalkyl substances. Environmental
Science and Pollution Research 24: 4916-4925. http://dx.doi.o
8243-3
Backe, WJ; Day, TC; Field, JA. (2013). Zwitterionic, cationic, and anionic fluorinated chemicals
in aqueous film forming foam formulations and groundwater from U.S. military bases by
nonaqueous large-volume injection HPLC-MS/MS. Environmental science & technology
47: 5226-5234. htfps://doi. >l/es3034999
Barber, JL; Berger, U; Chaemfa, C; Huber, S; Jahnke, A; Temme, C; Jones, KC. (2007).
Analysis of per- and polyfluorinated alkyl substances in air samples from Northwest
Europe. Journal of Environmental Monitoring 9: 530-541.
http://dx.doi.oo >701417a
Barbosa, V; Maulvault, AL; Alves, RN; Kwadijk, C; Kotterman, M; Tediosi, A, et al. (2018).
Effects of steaming on contaminants of emerging concern levels in seafood. Food and
Chemical Toxicology 118: 490-504. http://dx.doi.oi ^.fct.2018.05.047
Barreca, S; Busetto, M; Colzani, L; Clerici, L; Marchesi, V; Tremolada, L, et al. (2020).
Hyphenated high performance liquid chromatography-tandem mass spectrometry
techniques for the determination of perfluorinated alkylated substances in Lombardia
region in Italy, profile levels and assessment: One year of monitoring activities during
2018. MDPI 7. http://dx.doi.ore/10.3390/separations7010017
Becanova, J; Melymuk, L; Vojta, S; Komprdova, K; Klanova, J. (2016). Screening for
perfluoroalkyl acids in consumer products, building materials and wastes. Chemosphere
164: 322-329. https://doi.Qi ^ 10 101 >/i.chemosphere.2016.0S I I J
Beesoon, S; Genuis, SJ; Benskin, JP; Martin, JW. (2012). Exceptionally high serum
concentrations of perfluorohexanesulfonate in a Canadian family are linked to home
carpet treatment applications. Environmental Science and Technology 46: 12960-12967.
http://dx.doi.org/10J02 l/es3034654
Beser, MI; Pardo, O; Beltran, J; Yusa, V. (2011). Determination of per- and polyfluorinated
substances in airborne particulate matter by microwave-assisted extraction and liquid
chromatography-tandem mass spectrometry. Journal of Chromatography A 1218: 4847-
4855. httj) ://dx.doi.org/10.1016/j .chroma.2011.04.082
Beser, MI; Pardo, O; Beltran, J; Yusa, V. (2019). Determination of 21 perfluoroalkyl substances
and organophosphorus compounds in breast milk by liquid chromatography coupled to
4-2
-------
FINAL
2024
orbitrap high-resolution mass spectrometry. Analytica Chimica Acta 1049: 123-132.
https://doi.oi\ 10 101 >/i.aca.20 i s 10 0 > '<
Bhavsar, SP; Zhang, X; Guo, R; Braekevelt, E; Petro, S; Gandhi, N, et al. (2014). Cooking fish is
not effective in reducing exposure to perfluoroalkyl and polyfluoroalkyl substances.
Environment international 66: 107-114.
https://dot.or e/https://doi.org/.10 101 5/i.envint.2Q I i 01 0 J i
Binnington, MJ; Lei, YD; Pokiak, L; Pokiak, J; Ostertag, SK; Loseto, LL, et al. (2017). Effects
of preparation on nutrient and environmental contaminant levels in Arctic beluga whale
(Delphinapterus leucas) traditional foods. Environmental Science: Processes & Impacts
19: 1000-1015. http://dx.doi.org/10.1039/C7EM 3
Blaine, AC; Rich, CD; Hundal, LS; Lau, C; Mills, MA; Harris, KM; Higgins, CP. (2013).
Uptake of perfluoroalkyl acids into edible crops via land applied biosolids: field and
greenhouse studies. Environmental science & technology 47: 14062-14069.
https://doi.ore/10.1021/es403094q
Blaine, AC; Rich, CD; Sedlacko, EM; Hundal, LS; Kumar, K; Lau, C, et al. (2014).
Perfluoroalkyl acid distribution in various plant compartments of edible crops grown in
biosolids-amended soils. Environmental science & technology 48: 7858-7865.
https://doi.ore/10.1021/es500016s
Blake, BE; Cope, HA; Hall, SM; Keys, RD; Mahler, BW; McCord, J, et al. (2020). Evaluation of
Maternal, Embryo, and Placental Effects in CD-I Mice following Gestational Exposure
to Perfluorooctanoic Acid (PFOA) or Hexafluoropropylene Oxide Dimer Acid (HFPO-
DA or GenX). Environmental health perspectives 128: 27006.
https://doi.ore/10.1289/EHP6233
Blake, BE; Miller, CN; Nguyen, H; Chappell, VA; Phan, TP; Phadke, DP, et al. (2022).
Transcriptional pathways linked to fetal and maternal hepatic dysfunction caused by
gestational exposure to perfluorooctanoic acid (PFOA) or hexafluoropropylene oxide-
dimer acid (HFPO-DA or GenX) in CD-I mice. Ecotoxicol Environ Saf 248: 114314.
https://doi.oi ^ 10 101 '/i.ecoenv.2022 I I l'< I I
Boiteux, V; Dauchy, X; Bach, C; Colin, A; Hemard, J; Sagres, V, et al. (2017). Concentrations
and patterns of perfluoroalkyl and polyfluoroalkyl substances in a river and three
drinking water treatment plants near and far from a major production source. Science of
the Total Environment 583: 393-400. http://dx.doi.ore/10 101o/i.scitotem .01 01 0 °
Boiteux, V; Dauchy, X; Rosin, C; Munoz, JF. (2012). National screening study on 10
perfluorinated compounds in raw and treated tap water in France. Archives of
environmental contamination and toxicology 63: 1-12. http://dx.doi.ore/10.1007/s00244~
012-9754-7
Bonefeld-J0rgensen, EC; Long, M; Fredslund, SO; Bossi, R; Olsen, J. (2014). Breast cancer risk
after exposure to perfluorinated compounds in Danish women: a case-control study
nested in the Danish National Birth Cohort. Cancer Causes & Control : CCC 25: 1439-
1448. https://doi.or
Boobis, A; Budinsky, R; Collie, S; Crofton, K; Embry, M; Felter, S, et al. (2011). Critical
analysis of literature on low-dose synergy for use in screening chemical mixtures for risk
assessment. Critical reviews in toxicology 41: 369-383.
https://doi.ore/10.3109/10408444.2010.543655
Boone, JS; Guan, B; Vigo, C; Boone, T; Byrne, C; Ferrario, J. (2014). A method for the analysis
of perfluorinated compounds in environmental and drinking waters and the determination
4-3
-------
FINAL
2024
of their lowest concentration minimal reporting levels. Journal of Chromatography A
1345: 68-77. http://dx.doi.ore/ chroMa.2014.04.001
Boone, JS; Vigo, C; Boone, T; Byrne, C; Ferrario, J; Benson, R, et al. (2019). Per- and
polyfluoroalkyl substances in source and treated drinking waters of the United States.
Science of the Total Environment 653: 359-369.
http://dx.doi.on 10 101 i.scitotenv.l< /i.reprotox.200'" 01 00 I
Byrne, S; Seguinot-Medina, S; Miller, P; Waghiyi, V; von Hippel, FA; Buck, CL; Carpenter,
DO. (2017). Exposure to polybrominated diphenyl ethers and perfluoroalkyl substances
in a remote population of Alaska Natives. Environmental pollution (Barking, Essex :
1987) 231: 387-395. https://doi.oi lenvpol.2017.08.020
Cabrerizo, A; Muir, DCG; De Silva, AO; Wang, X; Lamoureux, SF; Lafreniere, MJ. (2018).
Legacy and emerging persistent organic pollutants (POPs) in terrestrial compartments in
the high Arctic: Sorption and secondary sources. Environmental Science and Technology
52: 14187-14197. http://dx.doi.ore/10.1021/acs.est.8b0
Cadwallader, A; Greene, A; Holsinger, H; Lan, A; Messner, M; Simic, M; Albert, R. (2022). A
Bayesian hierarchical model for estimating national PFAS drinking water occurrence.
AWWA Water Science 4: el284. https://doi.ore ws2.1284
Cariou, R; Veyrand, B; Yamada, A; Berrebi, A; Zalko, D; Durand, S, et al. (2015).
Perfluoroalkyl acid (PFAA) levels and profiles in breast milk, maternal and cord serum of
French women and their newborns. Environment international 84: 71-81.
https://doi.ore/10 101 >/i.envint.201^0 Oil
Carlsson, P; Herzke, D; Kallenborn, R. (2014). Polychlorinated biphenyls (PCBs),
polybrominated diphenyl ethers (PBDEs) and perfluorinated alkylated substances
4-4
-------
FINAL
2024
(PFASs) in traditional seafood items from western Greenland. Environmental Science
and Pollution Research 21: 4741-4750. http://dx.doi.on x
Carr, CK; Watkins, AM; Wolf, CJ; Abbott, BD; Lau, C; Gennings, C. (2013). Testing for
departures from additivity in mixtures of perfluoroalkyl acids (PFAAs). Toxicology 306:
169-175.
Chang, S; Butenhoff, JL; Parker, GA; Coder, PS; Zitzow, JD; Krisko, RM, et al. (2018).
Reproductive and developmental toxicity of potassium perfluorohexanesulfonate in CD-I
mice. Reproductive Toxicology 78: 150-168.
https://doi.org/l i\ .reprotox.2018.04.007
Chiesa, LM; Nobile, M; Ceriani, F; Malandra, R; Arioli, F; Panseri, S. (2019). Risk
characterisation from the presence of environmental contaminants and antibiotic residues
in wild and farmed salmon from different FAO zones. Food additives & contaminants
Part A, Chemistry, analysis, control, exposure & risk assessment 36: 152-162.
https://doi.org/10.1080/19440049.2Q18.1563723
Ciofi, L; Renai, L; Rossini, D; Ancillotti, C; Falai, A; Fibbi, D, et al. (2018). Applicability of the
direct injection liquid chromatographic tandem mass spectrometric analytical approach to
the sub-ng L"1 determination of perfluoro-alkyl acids in waste, surface, ground and
drinking water samples. Talanta 176: 412-421.
http://dx.doi.on .talanta.2017.08.052
Conley, JM; Lambright, CS; Evans, N; Farraj, AK; Smoot, J; Grindstaff, RD, et al. (2023). Dose
additive maternal and offspring effects of oral maternal exposure to a mixture of three
PFAS (HFPO-DA, NBP2, PFOS) during pregnancy in the Sprague-Dawley rat. The
Science of the total environment 892: 164609.
https://d0i.0rg/l 0.1016/i .scitotenv.20:
Conley, JM; Lambright, CS; Evans, N; McCord, J; Strynar, MJ; Hill, D, et al. (2021).
Hexafluoropropylene oxide-dimer acid (HFPO-DA or GenX) alters maternal and fetal
glucose and lipid metabolism and produces neonatal mortality, low birthweight, and
hepatomegaly in the Sprague-Dawley rat. Environment international 146: 106204.
https://doi.o /i .envint.2020.106204
Conley, JM; Lambright, CS; Evans, N; Medlock-Kakaley, E; Dixon, A; Hill, D, et al. (2022).
Cumulative maternal and neonatal effects of combined exposure to a mixture of
perfluorooctanoic acid (PFOA) and perfluorooctane sulfonic acid (PFOS) during
pregnancy in the Sprague-Dawley rat. Environment international 170: 107631.
https://d0i.0rg/l 0.1016/i .envint.2022.107631
Conley, JM; Lambright, CS; Evans, N; Strynar, MJ; McCord, J; Mclntyre, BS, et al. (2019).
Adverse Maternal, Fetal, and Postnatal Effects of Hexafluoropropylene Oxide Dimer
Acid (GenX) from Oral Gestational Exposure in Sprague-Dawley Rats. Environmental
health perspectives 127: 37008. https://dot.org/10.1289/EHP4372
Croes, K; Colles, A; Koppen, G; Govarts, E; Bruckers, L; Van de Mieroop, E, et al. (2012).
Persistent organic pollutants (POPs) in human milk: a biomonitoring study in rural areas
of Flanders (Belgium). Chemosphere 89: 988-994.
https://doi.Org/10.1016/i.chemosphere.2012.06.058
Crofton, KM; Craft, ES; Hedge, JM; Gennings, C; Simmons, JE; Carchman, RA, et al. (2005).
Thyroid-Hormone-Disrupting Chemicals: Evidence for Dose-Dependent Additivity or
Synergism. Environmental health perspectives 113: 1549-1554.
https://doi.org/10.1289/ehp.8195
4-5
-------
FINAL
2024
D'Eon J, C; Crozier, PW; Furdui, VI; Reiner, EJ; Libelo, EL; Mabury, SA. (2009).
Perfluorinated phosphonic acids in Canadian surface waters and wastewater treatment
plant effluent: discovery of a new class of perfluorinated acids. Environmental toxicology
and chemistry 28: 2101-2107. https://doi.ore/10.1897/09-048.1
D'Hollander, W; Herzke, D; Huber, S; Hajslova, J; Pulkrabova, J; Brambilla, G, et al. (2015).
Occurrence of perfluorinated alkylated substances in cereals, salt, sweets and fruit items
collected in four European countries. Chemosphere 129: 179-185.
http://dx.doi.on 10 101 i.chemosphere. JO I i 10 01 I
D'Hollander, W; Roosens, L; Covaci, A; Cornelis, C; Reynders, H; van Campenhout, K, et al.
(2010). Brominated flame retardants and perfluorinated compounds in indoor dust from
homes and offices in Flanders, Belgium. Chemosphere 81: 478-487.
http://dx.doi.on 10 101 t.chemosphere..010 0 0 I '<
D'Ambro, EL; Pye, HOT; Bash, JO; Bowyer, J; Allen, C; Efstathiou, C, et al. (2021).
Characterizing the Air Emissions, Transport, and Deposition of Per- and Polyfluoroalkyl
Substances from a Fluoropolymer Manufacturing Facility. Environmental science &
technology 55: 862-870. https://doi.ore/10.1021/acs.est.0c06580
Das, KP; Grey, BE; Rosen, MB; Wood, CR; Tatum-Gibbs, KR; Zehr, RD, et al. (2015).
Developmental toxicity of perfluorononanoic acid in mice. Reproductive Toxicology 51:
133-144. https://doi.oi^ 10 101 /i.reprotox.JO II IJ 01J
Dauchy, X; Boiteux, V; Bach, C; Rosin, C; Munoz, JF. (2017). Per- and polyfluoroalkyl
substances in firefighting foam concentrates and water samples collected near sites
impacted by the use of these foams. Chemosphere 183: 53-61.
http://dx.doi.ore/10.1016/i .chemosphere.. 01 0^016
Dauchy, X; Boiteux, V; Colin, A; Hemard, J; Bach, C; Rosin, C; Munoz, JF. (2019). Deep
seepage of per- and polyfluoroalkyl substances through the soil of a firefighter training
site and subsequent groundwater contamination. Chemosphere 214: 729-737.
http://dx.doi.ore/10.1016/i .chemosphere.. 01 \ 10 00 '<
Dauchy, X; Boiteux, V; Rosin, C; Munoz, JF. (2012). Relationship between industrial discharges
and contamination of raw water resources by perfluorinated compounds. Part I: Case
study of a fluoropolymer manufacturing plant. Bulletin of Environmental Contamination
and Toxicology 89: 525-530. http://dx.doi.ore/10.1007/s0012g ^x
de la Torre, A; Navarro, I; Sanz, P; Martinez, MLA. (2019). Occurrence and human exposure
assessment of perfluorinated substances in house dust from three European countries.
Science of the Total Environment 685: 308-314.
http://dx.doi.ore/10.1016/i.scitotenv.l0 r" 0 I '<
De Silva, AO; Armitage, JM; Bruton, TA; Dassuncao, C; Heiger-Bernays, W; Hu, XC, et al.
(2021). PFAS Exposure Pathways for Humans and Wildlife: A Synthesis of Current
Knowledge and Key Gaps in Understanding. Environmental toxicology and chemistry
40: 631-657. https://doi.ore/10.1002/etc.4935
De Silva, AO; Spencer, C; Scott, BF; Backus, S; Muir, DC. (2011). Detection of a cyclic
perfluorinated acid, perfluoroethylcyclohexane sulfonate, in the Great Lakes of North
America. Environmental science & technology 45: 8060-8066.
https ://doi. ore/10.1021 /es20013 5 c
Delinsky, AD; Strynar, MJ; McCann, PJ; Varns, JL; McMillan, L; Nakayama, SF; Lindstrom,
AB. (2010). Geographical distribution of perfluorinated compounds in fish from
4-6
-------
FINAL
2024
Minnesota lakes and rivers. Environmental science & technology 44: 2549-2554.
https://doi.org/10.1021/es9Q3777s
Dellatte, E; Brambilla, G; De Filippis, SP; di Domenico, A; Pulkrabova, J; Eschauzier, C, et al.
(2013). Occurrence of selected perfluorinated alkyl acids in lunch meals served at school
canteens in Italy and their relevance for children's intake. Food additives & contaminants
Part A, Chemistry, analysis, control, exposure & risk assessment 30: 1590-1597.
https://doi.org/l 0.1080/19440049.2013.813648
Ding, G; Zhang, J; Chen, Y; Wang, L; Wang, M; Xiong, D; Sun, Y. (2013). Combined effects of
PFOS and PFOA on zebrafish (Danio rerio) embryos. Archives of environmental
contamination and toxicology 64: 668-675. https://doi.Hv 10 100 i00J II 0 i J 9864-2
Domingo, JL; Jogsten, IE; Eriksson, U; Martorell, I; Perello, G; Nadal, M; van Bavel, B. (2012).
Human dietary exposure to perfluoroalkyl substances in Catalonia, Spain. Temporal
trend. Food Chemistry 135: 1575-1582.
http://dx.doi.oo .foodcheM.2012.06.054
DRBC (Delaware River Basin Commission). (2013). Contaminants of Emerging Concern in the
Tidal Delaware River: Pilot Monitoring Survey 2007-2009. Delaware River Basin
Commission, http://www.ni.gov/drbc/librarv/docurnents/contarriinants~of~errierging~
concern Aug if
Dreyer, A; Thuens, S; Kirchgeorg, T; Radke, M. (2012). Ombrotrophic peat bogs are not suited
as natural archives to investigate the historical atmospheric deposition of perfluoroalkyl
substances. Environmental Science and Technology 46: 7512-7519.
http://dx.doi.org/10J02 i/es20 II ^
Dreyer, A; Weinberg, I; Temme, C; Ebinghaus, R. (2009). Polyfluorinated compounds in the
atmosphere of the Atlantic and Southern Oceans: evidence for a global distribution.
Environmental science & technology 43: 6507-6514. https://doi.org/10.1021/es9010465
Ducatman, A; Zhang, J; Fan, H. (2015). Prostate-specific antigen and perfluoroalkyl acids in the
C8 health study population. Journal of Occupational and Environmental Medicine 57:
111-114. https://doi.orj 97/iom.0000000000000319
DuPont. (2008a). DuPont-24447: A 28-day oral (gavage) toxicity study of H-28397 in rats with a
28-day recovery. Volume 1 USEPAHERO Database 4221045
DuPont. (2008b). DuPont-24459: A 28-day oral (gavage) toxicity study of H-28397 in mice with
a 28-day recovery. Volume 1. USEPA HERO Database 4221051.
DuPont. (2009). DuPont-17751-1026: A 90-Day Oral (Gavage) Toxicity Study of H-28548 in
Rats with a 28-Day Recovery. . USEPA HERO Database 4222142.
DuPont. (2010a). DuPont-18405-841: An oral (gavage) prenatal developmental toxicity study of
H-28548 in rats. OECD Guideline 414. EI du Pont de Nemours and Company USEPA
HERO ID: 4222145.
DuPont. (2010b). DuPont-18405-1037: An oral (gavage) reproduction/developmental toxicity
screening study of H-28548 in mice. EI du Pont de Nemours and Company USEPA
HERO ID: 4222148.
DuPont. (2010c). DuPont-18405-1307: H-28548: Subchronic Toxicity 90-Day Gavage Study in
Mice. OECD Guideline 408. USEPA HERO Database 4229663.
DuPont. (2013). DuPont-18405-1238. H-28548: Combined chronic toxicity/oncogenicity study
2-year oral gavage study in rats. USEPA HERO Database 4222150
4-7
-------
FINAL
2024
Eberle, D; Ball, R; Boving, TB. (2017). Impact of ISCO Treatment on PFAA Co-Contaminants
at a Former Fire Training Area. Environmental science & technology 51: 5127-5136.
https://doi.ore/10.1021/acs.est.6b06591
ECHA (European Chemical Agency). (2019). Support Document for Identification of
Perfluorobutane Sulfonic Acid and its Salts as Substances of Very High Concern Because
of Their Hazardous Properties Which Cause Probable Serious Effects to Human Health
and the Environment Which Give Rise to An Equivalent Level of Concern to Those of
CMR and PBT/vPvB Substances (Article 57F).
https://echa.eiiropa.eii/documents/10162/891ab33d-d263-cc4b-0f2d-d84cfb7f424a
EFSA. (2012). Perfluoroalkylated substances in food: Occurrence and dietary exposure. EFSA
Journal 10: 2743.
Ericson, I; Marti-Cid, R; Nadal, M; Van Bavel, B; Lindstrom, G; Domingo, JL. (2008a). Human
exposure to perfluorinated chemicals through the diet: Intake of perfluorinated
compounds in foods from the Catalan (Spain) Market. Journal of agricultural and food
chemistry 56: 1787-1794. http://dx.doi.org/10.1021,
Ericson, I; Nadal, M; van Bavel, B; Lindstrom, G; Domingo, JL. (2008b). Levels of
perfluorochemicals in water samples from Catalonia, Spain: Is drinking water a
significant contribution to human exposure? Environmental Science and Pollution
Research 15: 614-619. htfp://dx.dot.c 38-0040-1
Eriksson, U; Karrman, A; Rotander, A; Mikkelsen, B; Dam, M. (2013). Perfluoroalkyl
substances (PFASs) in food and water from Faroe Islands. Environmental Science and
Pollution Research 20: 7940-7948. http://dx.doi.org 30-3
Eschauzier, C; Beerendonk, E; Scholte-Veenendaal, P; De Voogt, P. (2012). Impact of treatment
processes on the removal of perfluoroalkyl acids from the drinking water production
chain. Environmental Science and Technology 46: 1708-1715.
http://dx.doi. org/10.1021 /es201662b
Eschauzier, C; Hoppe, M; Schlummer, M; de Voogt, P. (2013). Presence and sources of
anthropogenic perfluoroalkyl acids in high-consumption tap-water based beverages.
Chemosphere 90: 36-41. http://dx.doi.oo .chemosphere.2012.06.070
Evans, N; Conley, JM; Cardon, M; Hartig, P; Medlock-Kakaley, E; Gray, LE, Jr. (2022). In vitro
activity of a panel of per- and polyfluoroalkyl substances (PFAS), fatty acids, and
pharmaceuticals in peroxisome proliferator-activated receptor (PPAR) alpha, PPAR
gamma, and estrogen receptor assays. Toxicol Appl Pharmacol 449: 116136.
https://doi.oi^ 10 101 >/i.taap.202, II l'<
Falandysz, J; Taniyasu, S; Gulkowska, A; Yamashita, N; Schulte-Oehlmann, U. (2006). Is fish a
major source of fluorinated surfactants and repellents in humans living on the Baltic
Coast? Environmental Science and Technology 40: 748-751.
http://dx.doi.oo 21/es051799m
Favreau, P; Poncioni-Rothlisberger, C; Place, BJ; Bouchex-Bellomie, H; Weber, A; Tremp, J, et
al. (2016). Multianalyte profiling of per- and polyfluoroalkyl substances (PFASs) in
liquid commercial products. Chemosphere 171: 491-501.
http://dx.doi.on 10 101 t.chemosphere.JOl>- II IJ
Feng, X; Cao, X; Zhao, S; Wang, X; Hua, X; Chen, L; Chen, L. (2017). Exposure of Pregnant
Mice to Perfluorobutanesulfonate Causes Hypothyroxinemia and Developmental
Abnormalities in Female Offspring. Toxicological Sciences 155: 409-419.
https://doi.org/l 0.1093/toxsci/kfw219
4-8
-------
FINAL
2024
Feng, X; Chen, X; Yang, Y; Yang, L; Zhu, Y; Shan, G, et al. (2021). External and internal
human exposure to PFOA and HFPOs around a mega fluorochemical industrial park,
China: Differences and implications. Environment international 157: 106824.
https://doi.org/l 0.1016/i .envint.2021.106824
Filipovic, M; Woldegiorgis, A; Norstrom, K; Bibi, M; Lindberg, M; Osteras, AH. (2015).
Historical usage of aqueous film forming foam: A case study of the widespread
distribution of perfluoroalkyl acids from a military airport to groundwater, lakes, soils
and fish. Chemosphere 129: 39-45. http://dx.doi.ore/10 J 016/i.chemosphere... 01 i 0l> 005
Fraser, AJ; Webster, TF; Watkins, DJ; Strynar, MJ; Kato, K; Calafat, AM, et al. (2013).
Polyfluorinated compounds in dust from homes, offices, and vehicles as predictors of
concentrations in office workers' serum. Environment international 60: 128-136.
https://doi.o /i. envint.2013.08.012
Fromme, H; Schlummer, M; Moller, A; Gruber, L; Wolz, G; Ungewiss, J, et al. (2007). Exposure
of an adult population to perfluorinated substances using duplicate diet portions and
biomonitoring data. Environmental science & technology 41: 7928-7933.
http s: //dot. or g/10.1021 /est rt
Furdui, VI; Stock, NL; Ellis, DA; Butt, CM; Whittle, DM; Crozier, PW, et al. (2007). Spatial
distribution of perfluoroalkyl contaminants in lake trout from the Great Lakes.
Environmental science & technology 41: 1554-1559. https://doi.org/10.1021/es0620484
Galloway, JE; Moreno, AVP; Lindstrom, AB; Strynar, MJ; Newton, S; May, AA; Weavers, LK.
(2020). Evidence of Air Dispersion: HFPO-DA and PFOA in Ohio and West Virginia
Surface Water and Soil near a Fluoropolymer Production Facility. Environmental science
& technology 54: 7175-7184. https://doi.org/10J02 l/acs.est.9b07384
Garcia, JS; Harbison, RD. (2015). Chapter 69: Perfluoralkyl Compounds. In RD Harbison; MM
Bourgeois; GT Johnson (Eds.), Hamilton & Hardy's Industrial Toxicology (pp. 689-
718). New York, NY: John Wiley & Sons, Inc.
Gebbink, WA; Glynn, A; Darnerud, PO; Berger, U. (2015). Perfluoroalkyl acids and their
precursors in Swedish food: The relative importance of direct and indirect dietary
exposure. Environmental Pollution 198: 108-115.
http://dx.doi.on 10 101 t.envpol JO! I i: 022
Gebbink, WA; van Asseldonk, L; van Leeuwen, SPJ. (2017). Presence of emerging per- and
polyfluoroalkyl substances (PFASs) in river and drinking water near a fluorochemical
production plant in the Netherlands. Environmental Science and Technology 51: 11057-
11065. http://dx.doi.org/10.1021 /acs.est.7b02488
Gellrich, V; Brunn, H; Stahl, T. (2013). Perfluoroalkyl and polyfluoroalkyl substances (PFASs)
in mineral water and tap water. Journal of Environmental Science and Health, Part A:
Toxic/Hazardous Substances & Environmental Engineering 48: 129-135.
http://dx.doi.on 10 1080/10934529.201 r I ; i
Gennings, C; Carter, WH, Jr.; Carney, EW; Charles, GD; Gollapudi, BB; Carchman, RA. (2004).
A Novel Flexible Approach for Evaluating Fixed Ratio Mixtures of Full and Partial
Agonists. Toxicological Sciences 80: 134-150. https://doi.( 3/toxsci/kfhl34
Genualdi, S; Jeong, N; deJager, L; Begley, T. (2017). Investigation into perfluoroalkyl
substances (PFASs) in a cranberry bog: method development and sampling results. Food
additives & contaminants Part A, Chemistry, analysis, control, exposure & risk
assessment 34: 2181-2189. https://doi.org/10.1080/19440049.^
4-9
-------
FINAL
2024
Geosyntec. (2019). Offsite Human Health Screening Level Exposure Assessment (SLEA) of
Table 3+ PFAS. Chemours Fayetteville Works. (Geosyntec Project Number TR0795).
www.deq.nc.eov/genx/consentorder/paraeraph-16/fw-cap4ina appendix-
f/download
Ghelli, E; Tondo, MT; Zironi, E; Pagliuca, G; Sirri, F; Gazzotti, T. (2019). Preliminary
monitoring of the presence of perfluoroalkyl substances in Italian eggs from different
breeding systems. Italian journal of food safety 8: 7702.
https://doi.ore/10.4081/iifs.2019.7702
Gilbert, ME; o'Shaughnessy, KL; Thomas, SE; Riutta, C; Wood, CR; Smith, A, et al. (2021).
Thyroid Disruptors: Extrathyroidal Sites of Chemical Action and Neurodevelopmental
Outcome-an Examination Using Triclosan and Perfluorohexane Sulfonate (PFHxS).
Toxicological sciences : an official journal of the Society of Toxicology.
https://doi.ore/10.1093/toxsci/kfab080
Giovanoulis, G; Nguyen, MA; Arwidsson, M; Langer, S; Vestergren, R; Lagerqvist, A. (2019).
Reduction of hazardous chemicals in Swedish preschool dust through article substitution
actions. Environment international 130: 104921.
http://dx.doi.ore/10.1016/i.envint-O I" 10 I
Gliige, J; Scheringer, M; Cousins, IT; DeWitt, JC; Goldenman, G; Herzke, D, et al. (2020). An
overview of the uses of per- and polyfluoroalkyl substances (PFAS). Environmental
Science: Processes & Impacts 22: 2345-2373. http://dx.doi.ore )0EM00291G
Gobelius, L; Hedlund, J; Diirig, W; Troger, R; Lilja, K; Wiberg, K; Ahrens, L. (2018). Per- and
polyfluoroalkyl substances in Swedish groundwater and surface water: Implications for
environmental quality standards and drinking water guidelines. Environmental Science
and Technology 52: 4340-4349. http://dx.doi.ore/10.102 l/acs.est.7b05718
Goosey, E; Harrad, S. (2012). Perfluoroalkyl substances in UK indoor and outdoor air: Spatial
and seasonal variation, and implications for human exposure. Environment international
45: 86-90. http://dx.doi.ore/10.1016/i .envint JO I J 0 I 00
Gray, LE; Conley, JM; Bursian, SJ. (2024). Dose addition models accurately predict the subacute
effects of a mixture of perfluorooctane sulfonate and perfluorooctanoic acid on japanese
quail (Coturnix japonica) chick mortality. Environmental toxicology and chemistry 43:
97-104. https://doi.ore/10.1002/et •a
Gremmel, C; Fromel, T; Knepper, TP. (2016). Systematic determination of perfluoroalkyl and
polyfluoroalkyl substances (PFASs) in outdoor jackets. Chemosphere 160: 173-180.
http://dx.doi.ore/10.1016/i .chemosphere.. 01 0 0 I '<
Groffen, T; Eens, M; Bervoets, L. (2019). Do concentrations of perfluoroalkylated acids
(PFAAs) in isopods reflect concentrations in soil and songbirds? A study using a distance
gradient from a fluorochemical plant. Science of the Total Environment 657: 111-123.
http://dx.doi.ore/10.1016/i .scitotenv.2018.12.072
Grannestad, R; Vazquez, BP; Arukwe, A; Jaspers, VLB; Jenssen, BM; Karimi, M, et al. (2019).
Levels, patterns, and biomagnification potential of perfluoroalkyl substances in a
terrestrial food chain in a Nordic skiing area. Environmental Science and Technology 53:
13390-13397. http://dx.doi.ore/10.102 1 /acs.est.9b02533
Gyllenhammar, I; Berger, U; Sundstrom, M; McCleaf, P; Euren, K; Eriksson, S, et al. (2015).
Influence of contaminated drinking water on perfluoroalkyl acid levels in human serum -
A case study from Uppsala, Sweden. Environmental research 140: 673-683.
http://dx.doi.ore/10.1016/i .envres JO I ^ O^ 01 •>
4-10
-------
FINAL
2024
Hale, SE; Arp, HP; Slinde, GA; Wade, EJ; Bj0rseth, K; Breedveld, GD, et al. (2017). Sorbent
amendment as a remediation strategy to reduce PFAS mobility and leaching in a
contaminated sandy soil from a Norwegian firefighting training facility. Chemosphere
171: 9-18. http://dx.doi.org/10.1016/i .chemosphere.l „ < <
Hansen, S; Vestergren, R; Herzke, D; Melhus, M; Evenset, A; Hanssen, L, et al. (2016).
Exposure to per- and polyfluoroalkyl substances through the consumption of fish from
lakes affected by aqueous film-forming foam emissions - A combined epidemiological
and exposure modeling approach. The SAMINOR 2 Clinical Study. Environment
international 94: 272-282. http://dx.doi.o '/i.envint.2016.05.030
Hardell, E; Karrman, A; van Bavel, B; Bao, J; Carlberg, M; Hardell, L. (2014). Case-control
study on perfluorinated alkyl acids (PFAAs) and the risk of prostate cancer. Environment
international 63: 35-39. https://doi.orv 10 101 | ^nvint.201 '< 10 005
Harrad, S; Drage, DS; Sharkey, M; Berresheim, H. (2020). Perfluoroalkyl substances and
brominated flame retardants in landfill-related air, soil, and groundwater from Ireland.
Science of the Total Environment 705: 135834.
http://dx.doi.on 10 101 i.scitotenv.20l'> L«s>34
Harrad, S; Wemken, N; Drage, DS; Abdallah, MAE; Coggins, AM. (2019). Perfluoroalkyl
substances in drinking water, indoor air and dust from Ireland: Implications for human
exposure. Environmental Science and Technology 53: 13449-13457.
http://dx.doi.ore/10.102 l/acs.est.9b04604
Hass, U; Christiansen, S; Axelstad, M; Scholze, M; Boberg, J. (2017). Combined exposure to
low doses of pesticides causes decreased birth weights in rats. Reproductive Toxicology
72: 97-105. https://doi.org/10.1016/i.reprotox.2017.05.004
Haug, LS; Huber, S; Schlabach, M; Becher, G; Thomsen, C. (2011). Investigation on per- and
polyfluorinated compounds in paired samples of house dust and indoor air from
Norwegian homes. Environmental Science and Technology 45: 7991-7998.
http://dx.doi.org/10J02 l/es!03456h
Herzke, D; Huber, S; Bervoets, L; D'Hollander, W; Hajslova, J; Pulkrabova, J, et al. (2013).
Perfluorinated alkylated substances in vegetables collected in four European countries;
occurrence and human exposure estimations. Environmental Science and Pollution
Research 20: 7930-7939. http://dx.doi.or -013-1777-8
Herzke, D; Posner, S; Olsson, E. (2009). Survey, Screening and Analysis of PFCs in Consumer
Products. (Project report 09/47). Sweden: SwerealVF.
Heydebreck, F; Tang, J; Xie, Z; Ebinghaus, R. (2015). Alternative and legacy perfluoroalkyl
substances: Differences between European and Chinese river/estuary systems.
Environmental Science and Technology 49: 8386-8395.
http://dx.doi.org/10.102 l/acs.est.5b01648
Hlouskova, V; Hradkova, P; Poustka, J; Brambilla, G; De Filipps, SP; D'Hollander, W, et al.
(2013). Occurrence of perfluoroalkyl substances (PFASs) in various food items of animal
origin collected in four European countries. Food Additives & Contaminants: Part A,
Chemistry, Analysis, Control, Exposure & Risk Assessment 30: 1918-1932.
http://dx.doi.on 10 1080/19440049.201 '< s ;?585
Hztisseter, A; Pfaff, A; Breedveld, GD. (2019). Leaching and transport of PFAS from aqueous
film-forming foam (AFFF) in the unsaturated soil at a firefighting training facility under
cold climatic conditions. Journal of Contaminant Hydrology 222: 112-122.
http://dx.doi.on 10 101 i.iconhyd.2011' OJ 010
4-11
-------
FINAL
2024
Holder, C; DeLuca, N; Luh, J; Alexander, P; Minucci, JM; Vallero, DA, et al. (2023). Systematic
Evidence Mapping of Potential Exposure Pathways for Per- and Polyfluoroalkyl
Substances Based on Measured Occurrence in Multiple Media. Environmental science &
technology 57: 5107-5116. https://doi.ore/10.1021/acs.est.2c07185
Holzer, J; Goen, T; Just, P; Reupert, R; Rauchfuss, K; Kraft, M, et al. (2011). Perfluorinated
compounds in fish and blood of anglers at Lake Mohne, Sauerland area, Germany.
Environmental Science and Technology 45: 8046-8052.
http://dx.doi.ore/10J02 i/ev 10 I > ° I
Houck, KA; Patlewicz, G; Richard, AM; Williams, AJ; Shobair, MA; Smeltz, M, et al. (2021).
Bioactivity profiling of per- and polyfluoroalkyl substances (PFAS) identifies potential
toxicity pathways related to molecular structure. Toxicology 457: 152789.
https://doi.oi ^ 10 101 '/i.tox.2^. I I . "9
Howdeshell, KL; Rider, CV; Wilson, VS; Furr, JR; Lambright, CR; Gray, LE, Jr. (2015). Dose
Addition Models Based on Biologically Relevant Reductions in Fetal Testosterone
Accurately Predict Postnatal Reproductive Tract Alterations by a Phthalate Mixture in
Rats. Toxicological Sciences 148: 488-502. https://doi.ore/ 3/toxsci/kfvl96
Huber, S; Haug, LS; Schlabach, M. (2011). Per- and polyfluorinated compounds in house dust
and indoor air from northern Norway - a pilot study. Chemosphere 84: 1686-1693.
http://dx.doi.on 10 101 i.chemosphere. JO I I 0 I 0 ^
Joerss, H; Apel, C; Ebinghaus, R. (2019). Emerging per- and polyfluoroalkyl substances
(PFASs) in surface water and sediment of the North and Baltic Seas. Sci Total Environ
686: 360-369. https://doi.c >/i.scitotenv.2019.05.363
Joerss, H; Xie, Z; Wagner, CC; von Appen, WJ; Sunderland, EM; Ebinghaus, R. (2020).
Transport of Legacy Perfluoroalkyl Substances and the Replacement Compound HFPO-
DA through the Atlantic Gateway to the Arctic Ocean-Is the Arctic a Sink or a Source?
Environmental science & technology 54: 9958-9967.
https://doi.ore/10.1021/acs.est.0c00228
Jogsten, IE; Nadal, M; van Bavel, B; Lindstrom, G; Domingo, JL. (2012). Per- and
polyfluorinated compounds (PFCs) in house dust and indoor air in Catalonia, Spain:
Implications for human exposure. Environment international 39: 172-180.
http://dx.doi. ore/10.1016/j. envint
Jogsten, IE; Perello, G; Llebaria, X; Bigas, E; Marti-Cid, R; Karrman, A; Domingo, JL. (2009).
Exposure to perfluorinated compounds in Catalonia, Spain, through consumption of
various raw and cooked foodstuffs, including packaged food. Food and Chemical
Toxicology 47: 1577-1583. http://dx.doi.ore/10.1016/j lct.2009.04.004
Johansson, JH; Berger, U; Vestergren, R; Cousins, IT; Bignert, A; Glynn, A; Darnerud, PO.
(2014). Temporal trends (1999-2010) of perfluoroalkyl acids in commonly consumed
food items. Environmental Pollution 188: 102-108.
http://dx.doi.ore/10.1016/j .envpol JO I I 01 026
Jorundsdottir, H; Halldorsson, TI; Gunnlaugsdottir, H. (2014). PFAAs in fish and other seafood
products from Icelandic waters. Journal of Environmental and Public Health 2014:
573607. http://dx.doi.ore/10.1155/2014/573607
Jurado-Sanchez, B; Ballesteros, E; Gallego, M. (2013). Semiautomated solid-phase extraction
followed by derivatisation and gas chromatography-mass spectrometry for determination
of perfluoroalkyl acids in water. Journal of Chromatography A 1318: 65-71.
http s: //dot. or e/http s: //doi. or e/ .10 101 )/i .chroma... 01'< 10 0 H
4-12
-------
FINAL
2024
Karaskova, P; Venier, M; Melymuk, L; Becanova, J; Vojta, S; Prokes, R, et al. (2016).
Perfluorinated alkyl substances (PFASs) in household dust in Central Europe and North
America. Environment international 94: 315-324.
https://doi.o 7/07-659.1
Koponen, J; Airaksinen, R; Hallikainen, A; Vuorinen, PJ; Mannio, J; Kiviranta, H. (2015).
Perfluoroalkyl acids in various edible Baltic, freshwater, and farmed fish in Finland.
Chemosphere 129: 186-191. https://doi.c >/i.chemosphere.2014.08.077
Kortenkamp, A; Haas, U. (2009). Expert workshop on combination effects of chemicals.
Hornbsek, Denmark. Odense, Denmark: Ministry of Environment of Denmark,
Environmental Protection Agency.
Kotlarz, N; McCord, J; Collier, D; Lea, CS; Strynar, M; Lindstrom, AB, et al. (2020).
Measurement of Novel, Drinking Water-Associated PFAS in Blood from Adults and
Children in Wilmington, North Carolina. Environmental health perspectives 128: 077005.
https://doi.ore/10.1289/EHP6837
Kotthoff, M; Miiller, J; Jiirling, H; Schlummer, M; Fiedler, D. (2015). Perfluoroalkyl and
polyfluoroalkyl substances in consumer products. Environmental science and pollution
research international 22: 14546-14559. https://doij '2-7
Kovarova, J; Marsalek, P; Blahova, J; Jurcikova, J; Kasikova, B; Svobodova, Z. (2012).
Occurrence of perfluoroalkyl substances in fish and water from the Svitava and Svratka
4-13
-------
FINAL
2024
rivers, Czech Republic. Bulletin of Environmental Contamination and Toxicology 88:
456-460. http://dx.doi.org/ 10128-011-0484-8
Krippner, J; Brunn, H; Falk, S; Georgii, S; Schubert, S; Stahl, T. (2014). Effects of chain length
and pH on the uptake and distribution of perfluoroalkyl substances in maize (Zea mays).
Chemosphere 94: 85-90. https://doi.org/10.1016/i chemosphere .01 '< 0° 01 s
Krippner, J; Falk, S; Brunn, H; Georgii, S; Schubert, S; Stahl, T. (2015). Accumulation
Potentials of Perfluoroalkyl Carboxylic Acids (PFCAs) and Perfluoroalkyl Sulfonic
Acids (PFSAs) in Maize (Zea mays). Journal of agricultural and food chemistry 63:
3646-3653. https://doi.org/10.102 l/acs.iafc.5bO(V IJ
Kubwabo, C; Kosarac, I; Lalonde, K. (2013). Determination of selected perfluorinated
compounds and polyfluoroalkyl phosphate surfactants in human milk. Chemosphere 91:
771-777. https://d oi.org/https://doi.org/10.1016/i .chemosphere. _ 01 '< OJ 01 I
Kubwabo, C; Stewart, B; Zhu, J; Marro, L. (2005). Occurrence of perfluorosulfonates and other
perfluorochemicals in dust from selected homes in the city of Ottawa, Canada. Journal of
Environmental Monitoring 7: 1074-1078. http://dx.dou^ 10 \0 '9/b5Q7'
Kuklenyik, Z; Reich, JA; Tully, JS; Needham, LL; Calafat, AM. (2004). Automated solid-phase
extraction and measurement of perfluorinated organic acids and amides in human serum
and milk. Environmental science & technology 38: 3698-3704.
https://doi.org/10.1021/es040332ii
Labadie, P; Chevreuil, M. (2011). Biogeochemical dynamics of perfluorinated alkyl acids and
sulfonates in the River Seine (Paris, France) under contrasting hydrological conditions.
Environmental Pollution 159: 3634-3639. http://dx.doi.org/10.1016/i .envpol.-01 I 0 028
Laitinen, JA; Koponen, J; Koikkalainen, J; Kiviranta, H. (2014). Firefighters' exposure to
perfluoroalkyl acids and 2-butoxyethanol present in firefighting foams. Toxicology letters
231: 227-232. htfps://doi.c >/i.toxlet.2014.09.007
Lankova, D; Lacina, O; Pulkrabova, J; Hajslova, J. (2013). The determination of perfluoroalkyl
substances, brominated flame retardants and their metabolites in human breast milk and
infant formula. Talanta 117: 318-325. http://dx.doij10 101 | uLmta.101 < OS.040
Lasier, PJ; Washington, JW; Hassan, SM; Jenkins, TM. (2011). Perfluorinated chemicals in
surface waters and sediments from northwest Georgia, USA, and their bioaccumulation
in Lumbriculus variegatus. Environmental toxicology and chemistry 30: 2194-2201.
http://dx.doi.oo Q2/etc.622
Lee, DG; Roehrdanz, PR; Feraud, M; Ervin, J; Anumol, T; Jia, A, et al. (2015). Wastewater
compounds in urban shallow groundwater wells correspond to exfiltration probabilities of
nearby sewers. Water Research 85: 467-475.
http://dx.doi.oo .wattes.2015.08.048
Lescord, GL; Kidd, KA; De Silva, AO; Williamson, M; Spencer, C; Wang, X; Muir, DC. (2015).
Perfluorinated and polyfluorinated compounds in lake food webs from the Canadian high
Arctic. Environmental Science and Technology 49: 2694-2702.
http://dx.doi.oo 21/es5048649
Li, J; He, J; Niu, Z; Zhang, Y. (2020b). Legacy per- and polyfluoroalkyl substances (PFASs) and
alternatives (short-chain analogues, F-53B, GenX and FC-98) in residential soils of
China: Present implications of replacing legacy PFASs. Environment international 135:
105419. http://dx.doi.org/10 101 >/j.envint.2QI" 10 5419
Li, Y; Feng, X; Zhou, J; Zhu, L. (2020). Occurrence and source apportionment of novel and
legacy poly/perfluoroalkyl substances in Hai River basin in China using receptor models
4-14
-------
FINAL
2024
and isomeric fingerprints. Water Research 168: 115145.
http://dx.doi.on 10 101 i.watres.20'1 ° I I I I ^
Li, Y; Feng, X; Zhou, J; Zhu, L. (2020a). Occurrence and source apportionment of novel and
legacy poly/perfluoroalkyl substances in Hai River basin in China using receptor models
and isomeric fingerprints. Water Research 168: 115145.
https://doi.oi \ 10 101 >/i.watres.20 rs I I I I ^
Li, Y; Yao, J; Zhang, J; Pan, Y; Dai, J; Ji, C; Tang, J. (2022). First Report on the
Bioaccumulation and Trophic Transfer of Perfluoroalkyl Ether Carboxylic Acids in
Estuarine Food Web. Environmental science & technology 56: 6046-6055.
https://doi.ore/10.1021/acs.est.lc00965
Lieder, PH; Chang, SC; York, RG; Butenhoff, JL. (2009a). Toxicological evaluation of
potassium perfluorobutanesulfonate in a 90-day oral gavage study with Sprague-Dawley
rats. Toxicology 255: 45-52. https://doi.org/)0 101 )/i.tox.2008 10 002
Lieder, PH; York, RG; Hakes, DC; Chang, SC; Butenhoff, JL. (2009b). A two-generation oral
gavage reproduction study with potassium perfluorobutanesulfonate (K+PFBS) in
Sprague Dawley rats. Toxicology 259: 33-45. https://doi.oi /j.tax.2009.01.027
Lindstrom, AB; Strynar, MJ; Delinsky, AD; Nakayama, SF; McMillan, L; Libelo, EL, et al.
(2011). Application of WWTP biosolids and resulting perfluorinated compound
contamination of surface and well water in Decatur, Alabama, USA. Environmental
science & technology 45: 8015-8021. https://doi.org/10J02 l/es!039425
Liu, X; Guo, Z; Krebs, KA; Pope, RH; Roache, NF. (2014). Concentrations and trends of
perfluorinated chemicals in potential indoor sources from 2007 through 2011 in the US.
Chemosphere 98: 51-57. https://doi.org/10.1016/i chemospher^' .01'< 10 00 I
Llorca, M; Farre, M; Pico, Y; Miiller, J; Knepper, TP; Barcelo, D. (2012). Analysis of
perfluoroalkyl substances in waters from Germany and Spain. Science of the Total
Environment 431: 139-150. http://dx.doi.org 10 101 i.scitotenv.201 J 0^ 01 I
Llorca, M; Farre, M; Pico, Y; Teijon, ML; Alvarez, JG; Barcelo, D. (2010). Infant exposure of
perfluorinated compounds: levels in breast milk and commercial baby food. Environment
international 36: 584-592. https://doi.org/10.1016/i envint.i.010 0 I 01
Loos, R; Locoro, G; Comero, S; Contini, S; Schwesig, D; Werres, F, et al. (2010). Pan-European
survey on the occurrence of selected polar organic persistent pollutants in ground water.
Water Research 44: 4115-4126. http://dx.doi.org/10.1016/i .watres.2010.05.032
Loos, R; Tavazzi, S; Mariani, G; Suurkuusk, G; Paracchini, B; Umlauf, G. (2017). Analysis of
emerging organic contaminants in water, fish and suspended particulate matter (SPM) in
the Joint Danube Survey using solid-phase extraction followed by UHPLC-MS-MS and
GC-MS analysis. Science of the Total Environment 607-608: 1201-1212.
http://dx.doi.on 10 101 i.scitotenv.101 0 0'<°
Lorenzo, M; Campo, J; Farre, M; Perez, F; Pico, Y; Barcelo, D. (2015). Perfluoroalkyl
substances in the Ebro and Guadalquivir river basins (Spain). Science of the Total
Environment 540: 191-199. http://dx.doi.org/10.1016/i .scitotenv.20I 0 OS-*
Makey, CM; Webster, TF; Martin, JW; Shoeib, M; Harner, T; Dix-Cooper, L; Webster, GM.
(2017). Airborne Precursors Predict Maternal Serum Perfluoroalkyl Acid Concentrations.
Environmental science & technology 51: 7667-7675.
https://doi.org/10.1021/acs.est.7b00615
Martin, O; Scholze, M; Ermler, S; McPhie, J; Bopp, SK; Kienzler, A, et al. (2021). Ten years of
research on synergisms and antagonisms in chemical mixtures: A systematic review and
4-15
-------
FINAL
2024
quantitative reappraisal of mixture studies. Environment international 146: 106206.
https://doi.o /i.chroma.l^M << I
Moser, VC; Casey, M; Hamm, A; Carter, WH, Jr.; Simmons, JE; Gennings, C. (2005).
Neurotoxicological and statistical analyses of a mixture of five organophosphorus
pesticides using a ray design. Toxicological sciences : an official journal of the Society of
Toxicology 86: 101-115. https://doi.ore/10.1093/toxsci/kfi 163
4-16
-------
FINAL
2024
Moser, VC; Padilla, S; Simmons, JE; Haber, LT; Hertzberg, RC. (2012). Impact of chemical
proportions on the acute neurotoxicity of a mixture of seven carbamates in preweanling
and adult rats. Toxicological sciences : an official journal of the Society of Toxicology
129: 126-134. https://doi.ore/10.1093/toxsci/kfsl90
Motas Guzman, M; Clementini, C; Perez-Carceles, MD; Jimenez Rejon, S; Cascone, A;
Martellini, T, et al. (2016). Perfluorinated carboxylic acids in human breast milk from
Spain and estimation of infant's daily intake. The Science of the total environment 544:
595-600. https://doi.o: /i.scitotenv.2015.11.059
Munoz, G; Fechner, LC; Geneste, E; Pardon, P; Budzinski, H; Labadie, P. (2016). Spatio-
temporal dynamics of per and polyfluoroalkyl substances (PFASs) and transfer to
periphytic biofilm in an urban river: Case-study on the River Seine. Environmental
Science and Pollution Research 25: 23574-23582. http://dx.doi.of!
8
Munoz, G; Fechner, LC; Geneste, E; Pardon, P; Budzinski, H; Labadie, P. (2018). Spatio-
temporal dynamics of per and polyfluoroalkyl substances (PFASs) and transfer to
periphytic biofilm in an urban river: Case-study on the River Seine. Environmental
Science and Pollution Research 25: 23574-23582. http://dx.doi.of!
8
Mussabek, D; Ahrens, L; Persson, KM; Berndtsson, R. (2019). Temporal trends and sediment-
water partitioning of per- and polyfluoroalkyl substances (PFAS) in lake sediment.
Chemosphere 227: 624-629. http://dx.doi.on 10 101 i.chemosphere _0 r" 0 I 0 I
Nakayama, S; Strynar, MJ; Helfant, L; Egeghy, P; Ye, X; Lindstrom, AB. (2007). Perfluorinated
compounds in the Cape Fear Drainage Basin in North Carolina. Environmental science &
technology 41: 5271-5276. https://doi.org/10.1021/es070792v
Nakayama, SF; Strynar, MJ; Reiner, JL; Delinsky, AD; Lindstrom, AB. (2010). Determination
of perfluorinated compounds in the Upper Mississippi River Basin. Environmental
Science and Technology 44: 4103-4109. http://dx.doi.ore/10.1021/esl00382z
NASEM (National Academies of Sciences Engineering and Medicine). (2022). Guidance on
PFAS Exposure, Testing, and Clinical Follow-Up. Washington, DC: The National
Academies Press, https://doi.org/10.17226/26156
NCBI (National Center for Biotechnology Information). (2022a). PubChem Compound
Summary for CID 67734, Perfluorohexanesulfonic acid. Bethesda, MD: U.S. National
Library of Medicine, National Center for Biotechnology Information.
https://piibchem.ncbi.nlm.nih.eov/compoimd/Perfliiorohexanesiilfonic-acid
NCBI (National Center for Biotechnology Information). (2022b). PubChem Compound
Summary for CID 67821, Perfluorononanoic Acid. Bethesda, MD: U.S. National Library
of Medicine, National Center for Biotechnology Information.
https://piibchem.ncbi.nlm.nih.eov/compoiind/Perfluorononanoic-acid
NCDEQ (North Carolina Department of Environmental Quality). (2017). 60-Day Notice of
Intent to Suspend Permit NC0003573. Raleigh, NC: The Chemours Company,
Fayetteville Works. https://ftles.nc.eov/ncdeq/GenX/60-
Day%20Notice%20of%20Intent%20~%20€hemours.pdf
NCDEQ (North Carolina Department of Environmental Quality). (2018a). 60-Day Notice of
Intent to Modify Air Quality Permit No. 03735T43. Raleigh, NC: The Chemours
Company, Fayetteville Works.
4-17
-------
FINAL
2024
https://files.nc.gOv/ncdeq/GenX/2 pri!6 Letter to Ch em ours DAQ FINAL signe
d.pdf
NCDEQ (North Carolina Department of Environmental Quality). (2018b). Groundwater and
Rainwater Sample Testing Results. Raleigh, NC: North Carolina Department of
Environmental Quality.
https://files.nc.gov/ncdeq/GenX/Rainwater%20and%20Groundwater%20Test%20Results
%2020180405.pdf
NCDEQ (North Carolina Department of Environmental Quality). (2018c). North Carolina
Department of Environmental Quality and North Carolina Department of Health and
Human Services: Secretaries' Science Advisory Board Review of the North Carolina
Drinking Water Provisional Health Goal for GenX. Raleigh, NC: North Carolina
Department of Environmental Quality, https://files.nc. gov/ncdeq/Gen XJ S AB/ S AB -
GenX-Report-draft-08-29-2018.pdf
NCHS (National Center for Health Statistics). (1993). Joint policy on variance estimation and
statistical reporting standards on NHANES III and CSFII reports: HNIS/NCHS Analytic
Working Group Recommendations. Human Nutrition Information Service
(HNIS)/Analytic Working Group. Riverdale, MD: Agricultural Research Service, Survey
Systems/Food Consumption Laboratory.
Newsted, JL; Holem, R; Hohenstein, G; Lange, C; Ellefson, M; Reagen, W; Wolf, S. (2017).
Spatial and temporal trends of poly- and perfluoroalkyl substances in fish fillets and
water collected from pool 2 of the Upper Mississippi River. Environmental toxicology
and chemistry 36: 3 138-3 147. http://dx.doi.org/ ;.3891
Newton, S; McMahen, R; Stoeckel, JA; Chislock, M; Lindstrom, A; Strynar, M. (2017). Novel
polyfluorinated compounds identified using high resolution mass spectrometry
downstream of manufacturing facilities near Decatur, Alabama. Environmental Science
and Technology 51: 1544-1552. http://dx.doi.org/10.1021/acs.est.6b05330
Nickerson, A; Maizel, AC; Kulkarni, PR; Adamson, DT; Kornuc, JJ; Higgins, CP. (2020).
Enhanced Extraction of AFFF-Associated PFASs from Source Zone Soils.
Environmental science & technology 54: 4952-4962.
https://doi.org/10.1021/acs.est.0cQ0792
Nielsen, G; Heiger-Bernays, WJ; Schlezinger, JJ; Webster, TF. (2021). Predicting the effects of
per- and polyfluoroalkyl substance mixtures on peroxisome proliferator-activated
receptor alpha activity in vitro. Toxicology 465: 153024.
https://doi.oi ^ 10 101 »/i.tox.20J \ \ I
NJSWQI (New Jersey Drinking Water Quality Institute). (2017). Public review draft: Health-
based maximum contaminant level support document: Perfluorooctane sulfonate (PFOS)
(CAS #: 1763-23-1; Chemical Formula: C8HF1703S). NJDWQI Health Effects
Subcommittee. https://cswab.org/wp~content/uploads/2018/09/New~Jersey~Scientific~
Docum entati on ~ Supportin g~ PF OS-13 -ppt.pdf
NOAA (National Oceanic and Atmospheric Administration). (2022). NOAA's National Status
and Trends Data Page. NCCOS Data Collections. Silver Spring, MD: U.S. Department of
Commerce, National Oceanic and Atmospheric Administration, National Centers for
Coastal Ocean Science, http s: //products. coastal science.noaa. gov/nsandt data/ data.aspx
Noorlander, CW; van Leeuwen, SP; Te Biesebeek, JD; Mengelers, MJ; Zeilmaker, MJ. (2011).
Levels of perfluorinated compounds in food and dietary intake of PFOS and PFOA in the
4-18
-------
FINAL
2024
Netherlands. Journal of agricultural and food chemistry 59: 7496-7505.
http://dx.doi.oo fl 04943p
Norwegian Environment Agency. (2018). Investigation of Sources to PFHxS in the
Environment. Final Report. Oslo, Norway: Norwegian Environment Agency.
NRC (National Research Council). (2008). Phthalates and Cumulative Risk Assessment: The
Tasks Ahead. Washington, DC: The National Academies Press.
https://doi.org/10.17226/12528
NTP. (2019a). FINAL REPORT on the Pathology Peer Review of Liver Findings from H-28548:
Subchronic Toxicity 90-Day Gavage Study in Mice (Project ID: Dupont-18405-1307)
and An Oral (Gavage) Reproduction/Developmental Toxicity Screening Study of H-
28548 in Mice (Study Number Wil-189225) (Study Sponsor Number: Dupont-18405-
1037).
NTP. (2019b). NTP Technical Report on the toxicity studies of perfluoroalkyl carboxylates
(perfluorohexanoic acid, perfluorooctanoic acid, perfluorononanoic acid, and
perfluorodecanoic acid) administered by gavage to Sprague-Dawley (HSD: Sprague-
Dawley SD) rats. NTP Tox 97 ISSN: 2378-8992.
NTP. (2019c). NTP Technical Report on the toxicity studies of perfluoroalkyl sulfonates
(perfluorobutane sulfonic acid, perfluorohexane sulfonate potassium salt, and
perfluorooctane sulfonic acid) administered by gavage to Sprague-Dawley (HSD:
Sprague-Dawley SD) rats. NTP Tox 96 ISSN: 2378-8992.
Nyberg, E; Awad, R; Bignert, A; Ek, C; Sallsten, G; Benskin, JP. (2018). Inter-individual, inter-
city, and temporal trends of per- and polyfluoroalkyl substances in human milk from
Swedish mothers between 1972 and 2016. Environmental science Processes & impacts
20: 1136-1147. https://doi. 8em00174i
OECD (Organisation for Economic Co-operation and Development). (2011). PFCS: Outcome of
the 2009 Survey - Survey on the Production, Use and Release of PFOS, PFAS, PFOA,
PFCA, Their Related Substances and Products/Mixtures Containing These Substances.
Series on Risk Management No. 24. Paris, France: Organisation for Economic Co-
operation and Development.
Ojo, AF; Peng, C; Ng, JC. (2020). Combined effects and toxicological interactions of
perfluoroalkyl and polyfluoroalkyl substances mixtures in human liver cells (HepG2).
Environmental pollution (Barking, Essex : 1987) 263: 114182.
http s: //dot. or e/10.1016/i. envpol. 2 2
Padilla-Sanchez, JA; Haug, LS. (2016). A fast and sensitive method for the simultaneous
analysis of a wide range of per- and polyfluoroalkyl substances in indoor dust using on-
line solid phase extraction-ultrahigh performance liquid chromatography-time-of-flight-
mass spectrometry. Journal of Chromatography A 1445: 36-45.
http://dx.doi.oo .chroma.2016.03.058
Pan, Y; Zhang, H; Cui, Q; Sheng, N; Yeung, LWY; Guo, Y, et al. (2017). First report on the
occurrence and bioaccumulation of hexafluoropropylene oxide trimer acid: An emerging
concern. Environmental Science and Technology 51: 9553-9560.
http://dx.doi.oo 21/acs.est.7b02259
Pan, Y; Zhang, H; Cui, Q; Sheng, N; Yeung, LWY; Sun, Y, et al. (2018). Worldwide distribution
of novel perfluoroether carboxylic and sulfonic acids in surface water. Environmental
Science and Technology 52: 7621-7629. http://dx.doi.ore/10.1021/acs.est.8b00829
4-19
-------
FINAL
2024
Papadopoulou, E; Poothong, S; Koekkoek, J; Lucattini, L; Padilla-Sanchez, JA; Haugen, M, et
al. (2017). Estimating human exposure to perfluoroalkyl acids via solid food and drinks:
Implementation and comparison of different dietary assessment methods. Environmental
research 158: 269-276. http://dx.doi.cn^ 10 101 5/i.envres.2QI 0 Oil
Perez, F; Llorca, M; Kock-Schulmeyer, M; Skrbic, B; Oliveira, LS; da Boit Martinello, K, et al.
(2014). Assessment of perfluoroalkyl substances in food items at global scale.
Environmental research 135: 181-189. http://dx.doi.org/10.1016/j.envres.JO I i 08.004
Petre, MA; Genereux, DP; Koropeckyj-Cox, L; Knappe, DRU; Duboscq, S; Gilmore, TE;
Hopkins, ZR. (2021). Per- and Polyfluoroalkyl Substance (PFAS) Transport from
Groundwater to Streams near a PFAS Manufacturing Facility in North Carolina, USA.
Environmental science & technology 55: 5848-5856.
https://doi.ore/10.1021/acs.est.0c07978
Post, GB; Louis, JB; Lippincott, RL; Procopio, NA. (2013). Occurrence of perfluorinated
compounds in raw water from New Jersey public drinking water systems. Environmental
science & technology 47: 13266-13275. https://dob >402884x
Pratt, I; Anderson, W; Crowley, D; Daly, S; Evans, R; Fernandes, A, et al. (2013). Brominated
and fluorinated organic pollutants in the breast milk of first-time Irish mothers: is there a
relationship to levels in food? Food Additives & Contaminants: Part A 30: 1788-1798.
https://doi.org/10.1080/19440049.2013.822569
Prevedouros, K; Cousins, IT; Buck, RC; Korzeniowski, SH. (2006). Sources, fate and transport
of perfluorocarboxylates. Environmental science & technology 40: 32-44.
http s: //doi. or g/10.1021 /est
Pritchett, JR; Rinsky, JL; Dittman, B; Christensen, A; Langley, R; Moore, Z, et al. (2019). Notes
from the field: Targeted biomonitoring for GenX and other per- and polyfluoroalkyl
substances following detection of drinking water contamination - North Carolina, 2018.
MMWR Morbidity and Mortality Weekly Report 68: 647-648.
http://dx.doi.org/10.155S5/mmwr.mm6829a4
Procopio, NA; Karl, R; Goodrow, SM; Maggio, J; Louis, JB; Atherholt, TB. (2017). Occurrence
and source identification of perfluoroalkyl acids (PFAAs) in the Metedeconk River
Watershed, New Jersey. Environmental Science and Pollution Research 24: 27125-
27135. http://dx.doi.on 10 100 v I I > "n - 01 0/.09-3
Quinones, O; Snyder, SA. (2009). Occurrence of perfluoroalkyl carboxylates and sulfonates in
drinking water utilities and related waters from the United States. Environmental science
& technology 43: 9089-9095. https://doi.org. :i/es9024707
Rider, CV; Furr, J; Wilson, VS; Gray, LE, Jr. (2008). A mixture of seven antiandrogens induces
reproductive malformations in rats. International journal of andrology 31: 249-262.
https://doi.Org/10.l 111/j.1365-2605.2 -859.x
Rider, CV; Furr, JR; Wilson, VS; Gray, LE, Jr. (2010). Cumulative effects of in utero
administration of mixtures of reproductive toxicants that disrupt common target tissues
via diverse mechanisms of toxicity. International journal of andrology 33: 443-462.
https://doi.o 35.2009.01049.x
Rider, CV; Wilson, VS; Howdeshell, KL; Hotchkiss, AK; Furr, JR; Lambright, CR; Gray, LE,
Jr. (2009). Cumulative effects of in utero administration of mixtures of "antiandrogens"
on male rat reproductive development. Toxicologic pathology 37: 100-113.
https://doi.o 70192623308329478
4-20
-------
FINAL
2024
Riviere, G; Jean, J; Gorecki, S; Hulin, M; Kolf-Clauw, M; Feidt, C, et al. (2019). Dietary
exposure to perfluoroalkyl acids, brominated flame retardants and health risk assessment
in the French infant total diet study. Food and Chemical Toxicology 131: 110561.
http://dx.doi.on 10 101 tlet..0 r" 0 008
Rosen, MB; Das, KP; Rooney, J; Abbott, B; Lau, C; Corton, JC. (2017). PPARalpha-
independent transcriptional targets of perfluoroalkyl acids revealed by transcript
profiling. Toxicology 387: 95-107. https://doi.oi^ 10 101 ^tax.201 0^ 01 '<
Rosen, MB; Das, KP; Wood, CR; Wolf, CJ; Abbott, BD; Lau, C. (2013). Evaluation of
perfluoroalkyl acid activity using primary mouse and human hepatocytes. Toxicology
308: 129-137. https://doi.oiy 10 101 >/i.tox.20l '< 0'« 01 I
Rosenmai, AK; Ahrens, L; le Godec, T; Lundqvist, J; Oskarsson, A. (2018). Relationship
between peroxisome proliferator-activated receptor alpha activity and cellular
concentration of 14 perfluoroalkyl substances in HepG2 cells. J Appl Toxicol 38: 219-
226. https://doi.ore. 10 1002/iai > ^ I ^
Rostkowski, P; Taniyasu, S; Yamashita, N; Falandysz, JJ; Zegarowski, L; Chojnacka, A, et al.
(2009). Survey of perfluorinated compounds (PFCs) in surface waters of Poland. Journal
of Environmental Science and Health, Part A: Toxic/Hazardous Substances &
Environmental Engineering 44: 1518-1527.
http://dx.doi.ore/i0J080/10934520903263330
Rushing, BR; Hu, Q; Franklin, JN; McMahen, R; Dagnino, S; Higgins, CP, et al. (2017).
Evaluation of the immunomodulatory effects of 2,3,3,3-tetrafluoro-2-
(heptafluoropropoxy)-propanoate in C57BL/6 mice. Toxicological sciences : an official
journal of the Society of Toxicology 156: 179-189. https://doi.oi ^ 10 10' '3/toxsci/kfu _ * I
Sadia, M; Yeung, LWY; Fiedler, H. (2020). Trace level analyses of selected perfluoroalkyl acids
in food: Method development and data generation. Environmental pollution (Barking,
Essex : 1987) 263: 113721. https://doi.ore/10.1016/i.envpoi J0rs II '< J I
Sammut, G; Sinagra, E; Sapiano, M; Helmus, R; de Voogt, P. (2019). Perfluoroalkyl substances
in the Maltese environment - (II) sediments, soils and groundwater. Science of the Total
Environment 682: 180-189. http://dx.doi.ore 10 101 i.scitotenv.20 I" 0 i 103
Schaider, LA; Balan, SA; Blum, A; Andrews, DQ; Strynar, MJ; Dickinson, ME, et al. (2017).
Fluorinated Compounds in U.S. Fast Food Packaging. Environmental science &
technology letters 4: 105-111. https://doi.ore/10.102 i/acs.estlett.6b00435
Schecter, A; Colacino, J; Haffner, D; Patel, K; Opel, M; Papke, O; Birnbaum, L. (2010).
Perfluorinated compounds, polychlorinated biphenyls, and organochlorine pesticide
contamination in composite food samples from Dallas, Texas, USA. Environmental
health perspectives 118: 796-802. https://doi.ore/10.1289/ehp.0901347
Scher, DP; Kelly, JE; Huset, CA; Barry, KM; Hoffbeck, RW; Yingling, VL; Messing, RB.
(2018). Occurrence of perfluoroalkyl substances (PFAS) in garden produce at homes with
a history of PFAS-contaminated drinking water. Chemosphere 196: 548-555.
https://doi.ore/10 101 '/i.ehemosphen' JO I IJ I °
Scher, DP; Kelly, JE; Huset, CA; Barry, KM; Yingling, VL. (2019). Does soil track-in contribute
to house dust concentrations of perfluoroalkyl acids (PFAAs) in areas affected by soil or
water contamination? Journal of exposure science & environmental epidemiology 29:
218-226. https://doi.ore/10.1038A 11 '< 0-01 s 0101
Schlummer, M; Solch, C; Meisel, T; Still, M; Gruber, L; Wolz, G. (2015). Emission of
perfluoroalkyl carboxylic acids (PFCA) from heated surfaces made of
4-21
-------
FINAL
2024
polytetrafluoroethylene (PTFE) applied in food contact materials and consumer products.
Chemosphere 129: 46-53. https://doi.ore/10 101 i chemosphere.201 I I I 0 o
Schultes, L; Vestergren, R; Volkova, K; Westberg, E; Jacobson, T; Benskin, JP. (2018). Per- and
polyfluoroalkyl substances and fluorine mass balance in cosmetic products from the
Swedish market: Implications for environmental emissions and human exposure.
Environmental Science: Processes & Impacts 20: 1680-1690.
http://dx.doi.org/10.1039/c8em00368h
Scordo, CVA; Checchini, L; Renai, L; Orlandini, S; Bruzzoniti, MC; Fibbi, D, et al. (2020).
Optimization and validation of a method based on QuEChERS extraction and liquid
chromatographic-tandem mass spectrometric analysis for the determination of
perfluoroalkyl acids in strawberry and olive fruits, as model crops with different matrix
characteristics. Journal of Chromatography A 1621: 461038.
http://dx.doi.org/10 J 016/j .chroMa.2020.46103 8
Semerad, J; Hatasova, N; Grasserova, A; Cerna, T; Filipova, A; Hanc, A, et al. (2020). Screening
for 32 per- and polyfluoroalkyl substances (PFAS) including GenX in sludges from 43
WWTPs located in the Czech Republic - Evaluation of potential accumulation in
vegetables after application of biosolids. Chemosphere 261: 128018.
https://doi.ou 10 101 /i.chemosphere.20JO I JN018
Shafique, U; Schulze, S; Slawik, C; Bohme, A; Paschke, A; Schiiurmann, G. (2017).
Perfluoroalkyl acids in aqueous samples from Germany and Kenya. Environmental
Science and Pollution Research 24: 11031-11043. http://dx.doi.on
1
Sigma-Aldrich. (2014). Material Safety Data Sheet for Tridecafluorohexane-1-sulfonic acid
potassium salt, Product Number 50929. Version 3.6. Revision Date 12/22/2014.
https://www.sigmaaldrich.com/safety-center.html
Simcik, MF; Dorweiler, KJ. (2005). Ratio of perfluorochemical concentrations as a tracer of
atmospheric deposition to surface waters. Environmental science & technology 39: 8678-
8683. https://doi.org/10.1021/esQ-
Sinclair, E; Kim, SK; Akinleye, HB; Kannan, K. (2007). Quantitation of gas-phase
perfluoroalkyl surfactants and fluorotelomer alcohols released from nonstick cookware
and microwave popcorn bags. Environmental science & technology 41: 1180-1185.
https://doi.org/10.1021/es062377w
Sinclair, E; Mayack, DT; Roblee, K; Yamashita, N; Kannan, K. (2006). Occurrence of
perfluoroalkyl surfactants in water, fish, and birds from New York State. Archives of
environmental contamination and toxicology 50: 398-410.
https://doi.org/10.1007/sQ0244-0C
Skaar, JS; Raider, EM; Lyche, JL; Ahrens, L; Kallenborn, R. (2019). Elucidation of
contamination sources for poly- and perfluoroalkyl substances (PFASs) on Svalbard
(Norwegian Arctic). Environmental Science and Pollution Research 26: 7356-7363.
http://dx.doi.ore 10 100 v I I > m Ols-2162-4
Song, X; Vestergren, R; Shi, Y; Huang, J; Cai, Y. (2018). Emissions, transport, and fate of
emerging per- and polyfluoroalkyl substances from one of the major fluoropolymer
manufacturing facilities in China. Environmental Science and Technology 52: 9694-
9703. http://dx.doij ;s.est.7b06657
Stahl, LL; Snyder, BD; Olsen, AR; Kincaid, TM; Wathen, JB; McCarty, HB. (2014).
Perfluorinated compounds in fish from U.S. urban rivers and the Great Lakes. The
4-22
-------
FINAL
2024
Science of the total environment 499: 185-195.
https://doi.org/l 0.1016/i .scitotenv.20
Steele, M; Griffith, C; Duran, C. (2018). Monthly variations in perfluorinated compound
concentrations in groundwater. Toxics 6: 56. http://dx.doi.ore/10.3390/toxics6030056
Still, M; Schlummer, M; Gruber, L; Fiedler, D; Wolz, G. (2013). Impact of Industrial Production
and Packaging Processes on the Concentration of Per- and Polyfluorinated Compounds in
Milk and Dairy Products. Journal of agricultural and food chemistry 61: 9052-9062.
httos ://doi. org/10.1021 /i F402013 7
Stock, NL; Furdui, VI; Muir, DC; Mabury, SA. (2007). Perfluoroalkyl contaminants in the
Canadian Arctic: evidence of atmospheric transport and local contamination.
Environmental science & technology 41: 3529-3536. https://doi.ore/10.1021/es062709x
Strakova, J; Schneider, J; Cingotti, N. (2021). Throwaway Packaging, Forever Chemicals:
European wide survey of PFAS in disposable food packaging and tableware.
https://arnika.org/en/publications/throwawav-packaging-forever-chemicals-european-
wide-survev-of-pfas-in-disposable-food-packaging-and-tableware
Strynar, M; Dagnino, S; McMahen, R; Liang, S; Lindstrom, A; Andersen, E, et al. (2015).
Identification of Novel Perfluoroalkyl Ether Carboxylic Acids (PFECAs) and Sulfonic
Acids (PFESAs) in Natural Waters Using Accurate Mass Time-of-Flight Mass
Spectrometry (TOFMS). Environmental science & technology 49: 11622-11630.
https://doi.org/l 0.1021/acs.est.5b(
Strynar, MJ; Lindstrom, AB. (2008). Perfluorinated compounds in house dust from Ohio and
North Carolina, USA. Environmental science & technology 42: 3751-3756.
https://doi.ore/10.1021/es7032058
Subedi, B; Codru, N; Dziewulski, DM; Wilson, LR; Xue, J; Yun, S, et al. (2015). A pilot study
on the assessment of trace organic contaminants including pharmaceuticals and personal
care products from on-site wastewater treatment systems along Skaneateles Lake in New
York State, USA. Water Research 72: 28-39.
http://dx.doi.org 10 101 t.watres .01 I 10 0 1°
Sun, M; Arevalo, E; Strynar, M; Lindstrom, A; Richardson, M; Kearns, B, et al. (2016). Legacy
and emerging perfluoroalkyl substances are important drinking water contaminants in the
Cape Fear River watershed of North Carolina. Environmental science & technology
letters 3: 415-419. http://dx.doi.ore/10.1021/acs.estlett.6b00398
Sundstrom, M; Ehresman, DJ; Bignert, A; Butenhoff, JL; Olsen, GW; Chang, SC; Bergman, A.
(2011). A temporal trend study (1972-2008) of perfluorooctanesulfonate,
perfluorohexanesulfonate, and perfluorooctanoate in pooled human milk samples from
Stockholm, Sweden. Environment international 37: 178-183.
https://doi.o /j.envint.2010.08.014
Surma, M; Piskula, M; Wiczkowski, W; Zielinski, H. (2017). The perfluoroalkyl carboxylic
acids (PFCAs) and perfluoroalkane sulfonates (PFSAs) contamination level in spices. Eur
Food Res Technol 243: 297-307. http://dx.doi.ore/10.1007/s00
Surma, M; Wiczkowski, W; Zielinski, H; Cieslik, E. (2015). Determination of Selected
Perfluorinated Acids (PFCAs) and Perfluorinated Sulfonates (PFASs) in Food Contact
Materials Using LC-MS/MS. Packaging Technology and Science 28: 789-799.
https://onlinelibrarY.wileY.com/doi/abs/10.1002/pts.214Q
Sznajder-Katarzynska, K; Surma, M; Cieslik, E; Wiczkowski, W. (2018). The perfluoroalkyl
substances (PFASs) contamination of fruits and vegetables. Food Additives &
4-23
-------
FINAL
2024
Contaminants: Part A, Chemistry, Analysis, Control, Exposure & Risk Assessment 35:
1776-1786. http://dx.doi.org/10.1080/19440049.2018.1502477
Sznajder-Katarzynska, K; Surma, M; Wiczkowski, W; Cieslik, E. (2019). The perfluoroalkyl
substance (PFAS) contamination level in milk and milk products in Poland. International
Dairy Journal 96: 73-84. http://dx.doi.org/10.1016/i.idairvi.2019.04.008
Tipton, JJ; Guillette, LJ, Jr.; Lovelace, S; Parrott, BB; Rainwater, TR; Reiner, JL. (2017).
Analysis of PFAAs in American alligators part 1: Concentrations in alligators harvested
for consumption during South Carolina public hunts. Journal of environmental sciences
(China) 61: 24-30. https://doi.org/i 0. i 0 i 6/i .ies.101 0 ¦> 0 I ¦>.
Tittlemier, SA; Pepper, K; Seymour, C; Moisey, J; Bronson, R; Cao, XL; Dabeka, RW. (2007).
Dietary exposure of Canadians to perfluorinated carboxylates and perfluorooctane
sulfonate via consumption of meat, fish, fast foods, and food items prepared in their
packaging. Journal of agricultural and food chemistry 55: 3203-3210.
https ://doi. org/10.1021 /j f063¦4045
US FDA (U.S. Food and Drug Administration). (2018). Analytical Results for PFAS in 2018
Produce Sampling (Parts Per Trillion). Silver Spring, MD: U.S. Department of Health
and Human Services, U.S. Food and Drug Administration.
https://www.fda.gov/media/127848/download
US FDA (U.S. Food and Drug Administration). (2020a). Analytical Results for PFAS in 2019
Total Diet Study Sampling (Parts Per Trillion)—Dataset 1. Silver Spring, MD: U.S.
Department of Health and Human Services, U.S. Food and Drug Administration.
https://www.fda.gov/media/127852/download
US FDA (U.S. Food and Drug Administration). (2020b). Analytical Results for PFAS in 2019
Total Diet Study Sampling (Parts Per Trillion)—Dataset 2. Silver Spring, MD: U.S.
Department of Health and Human Services, U.S. Food and Drug Administration.
https ://www.fda. gov/media/13 3 693/download
US FDA (U.S. Food and Drug Administration). (2021a). Analytical Results for PFAS in 2021
Total Diet Study Sampling (Parts Per Trillion)—Dataset 3. Silver Spring, MD: U.S.
Department of Health and Human Services, U.S. Food and Drug Administration.
https://www.fda.gov/media/150338/download
US FDA (U.S. Food and Drug Administration). (2021b). Analytical Results for PFAS in 2021
Total Diet Study Sampling (Parts Per Trillion)—Dataset 4. Silver Spring, MD: U.S.
Department of Health and Human Services, U.S. Food and Drug Administration.
https://www.fda.gov/medi: lownload
US FDA (U.S. Food and Drug Administration). (2021c). Analytical Results for PFAS in 2018-
2021 Dairy Farm Sampling (Parts Per Trillion). Silver Spring, MD: U.S. Department of
Health and Human Services, U.S. Food and Drug Administration.
https://www.fda.gov/media/127850/download
US FDA (U.S. Food and Drug Administration). (2022a). Update on FDA's Continuing Efforts to
Understand and Reduce Exposure to PFAS from Foods. Silver Spring, MD: U.S.
Department of Health and Human Services, U.S. Food and Drug Administration.
https://www.fda.gov/food/cfsan-constitiient-iipdates/iipdate-fdas-continuing-eftbrts-
understand-and-reduce-exposure-pfas-foods
US FDA (U.S. Food and Drug Administration). (2022b). Analytical Results for PFAS in 2022
Total Diet Sampling (Parts Per Trillion)—Dataset 5. Silver Spring, MD: U.S. Department
4-24
-------
FINAL
2024
of Health and Human Services, U.S. Food and Drug Administration.
https://www.fda.gOv/media/l 56351/download
US FDA (U.S. Food and Drug Administration). (2022c). Analytical Results of Testing Food for
PFAS from Environmental Contamination. Silver Spring, MD: U.S. Department of
Health and Human Services, U.S. Food and Drug Administration.
https://www.fda.gov/food/chemical-contaminants-food/analvtical-results-testing-food-
pfas-environmental-contamination
US GPO (U.S. Government Publishing Office). (2022). Electronic Code of Federal Regulations:
40 CFR 180. Washington, DC: U.S. Government Publishing Office.
https://www.ecfr.gOv/current/title-40/chapter-I/subchapter-E/p ?toc=l
US OSTP (United States Office of Science and Technology Policy). (2023). Per- and
Polyfluoroalkyl Substances (PFAS) Report, A Report by the Joint Subcommittee on
Environment, Innovation, and Public Health.
https://www.whitehouse.gov/wpcontent/uploads/2023/03/QSTP-March-2023-
PFASReport.pdf
USEPA (U.S. Environmental Protection Agency). (1986). Guidelines for the Health Risk
Assessment of Chemical Mixtures. (EPA/630/R-98/002). Washington, DC: EPA Risk
Assessment Forum, https://www.epa.gov/risk/guidelines-health-risk-assessment-
chemical-m ixtures
USEPA (U.S. Environmental Protection Agency). (1989). Notice of Availability of Drinking
Water Health Advisories for pesticides. (54 FR 7599). Washington, DC: Federal
Register.
USEPA (U.S. Environmental Protection Agency). (1991b). Risk Assessment Guidance for
Superfund, Vol 1: Human Health Evaluation Manual (Part B, Development of Risk-
Based Preliminary Remediation Goals). (EPA/540/R-92/003). Washington, DC: U.S.
Environmental Protection Agency, Emergency and Remedial Response.
https://www.epa.gov/risk/risk-assessment-giiidance-siiperfimd-rags-part-b
USEPA (U.S. Environmental Protection Agency). (1998). National Primary Drinking Water
Regulations: Disinfectants and Disinfection Byproducts. Federal Register. 63 FR 69390-
69476. December 16, 1998.
USEPA (U.S. Environmental Protection Agency). (2000a). Supplementary Guidance for
Conducting Health Risk Assessment of Chemical Mixtures. (EPA/630/R-00/002).
Washington, DC: EPA, Risk Assessment Forum Technical Panel.
https://cfpub. epa.gov/ncea/risk/recordisplav. cfm?deid=20533
USEPA (U.S. Environmental Protection Agency). (2000b). Methodology for Deriving Ambient
Water Quality Criteria for the Protection of Human Health. (EPA-822-B-00-004).
Washington, DC: EPA, Office of Water, https://www.epa.gov/sites/default/files/2018-
10/documents/methodology-wqc-protection-hh-2000.pdf
USEPA (U.S. Environmental Protection Agency). (2000c). National Primary Drinking Water
Regulations; Radionuclides; Final Rule. Federal Register. 65 FR 76708. December 7,
2000.
USEPA (U.S. Environmental Protection Agency). (2001). National Primary Drinking Water
Regulations; Arsenic and Clarifications to Compliance and New Source Contaminants
Monitoring. Federal Register. FR 28342. July 19, 2001.
USEPA (U.S. Environmental Protection Agency). (2002). A Review of the Reference Dose and
Reference Concentration Process. (EPA/630/P-02/002F). Washington, DC: EPA, Risk
4-25
-------
FINAL
2024
Assessment Forum, https://www.epa.gov/sites/default/files/20 locuments/rfd-
final.pdf
USEPA (U.S. Environmental Protection Agency). (2004). Drinking Water Health Advisory for
Manganese. (EPA-822-R-04-003). Washington, DC: U.S. Environmental Protection
Agency, https://www.epa.eov/sites/default/files/
09/documents/supp< v i i,iagnesedwrepoit 0.pdf
USEPA (U.S. Environmental Protection Agency). (2005). Guidelines for Carcinogen Risk
Assessment. (EPA/630/P-03/001F). Washington, DC: EPA Risk Assessment Forum.
https://www.epa.eov/sites/defaiilt/files/2013-09/documents/cancer guidelines final 3-
25-05.pdf
USEPA (U.S. Environmental Protection Agency). (2009a). The National Study of Chemical
Residues in Lake Fish Tissue. (EPA-823-R-09-006). Washington, DC: U.S.
Environmental Protection Agency, Office of Water, Office of Science and Technology.
https://nepis.epa.eov/Exe/ZvPDF.cei/P1005P2Z.PDF7Dockey DF
USEPA (U.S. Environmental Protection Agency). (2009b). Consent Order and Determinations
Supporting Consent Order for Premanufacture Notice Numbers: P-08-508 and P-08-509.
Washington, DC: U.S. Environmental Protection Agency, Office of Pollution Prevention
and Toxics.
https://chemview.epa.eov/chemview/proxy?filename=sanitized consent ord 8 050
8c.pdf
USEPA (U.S. Environmental Protection Agency). (2010). Fluoride: Exposure and Relative
Source Contribution Analysis. (EPA 820-R-10-015). Washington, DC: U.S.
Environmental Protection Agency, https://www.epa.gov/sites/defaiilt/files/2019-
03/documents/fluoride-exposure-relative~report..pdf
USEPA (U.S. Environmental Protection Agency). (2016a). Drinking Water Health Advisory for
Perfluorooctanoic Acid (PFOA). (EPA-822-R-16-005). Washington, DC: EPA Office of
Water, https://www.epa.eov/sites/production/files/2
05/docum ents/pfoa h ealth advi sorv final -pi ain. p df
USEPA (U.S. Environmental Protection Agency). (2019b). Update for Chapter 3 of the Exposure
Factors Handbook: Ingestion of Water and Other Select Liquids. (EPA/600/R-18/259F).
Washington, DC: EPA Office of Research and Development.
https://www.epa.eov/sites/defaiilt/files/2019-02/documents/efh - chapter 3 update.pdf
USEPA (U.S. Environmental Protection Agency). (2020a). National Aquatic Resource Surveys.
National Rivers and Streams Assessment 2013-2014. http://www.epa.gov/national-
aquatic-resource-survevs/data-national-aquatic-resource-survevs
USEPA (U.S. Environmental Protection Agency). (2020b). Systematic Review Protocol for the
PFBA, PFHxA, PFHxS, PFNA, and PFDA (anionic and acid forms) IRIS Assessments.
(EPA/63 5/R-2 0/131). Washington, DC: EPA Office of Research and Development.
https://ordspub.epa.eov/ords/eims/eimscomm.eetfile7p download id=541666
USEPA (U.S. Environmental Protection Agency). (2021c). Human Health Toxicity Values for
Hexafluoropropylene Oxide (HFPO) Dimer Acid and Its Ammonium Salt (CASRN
13252-13-6 and CASRN 62037-80-3). Also Known as "GenX Chemicals.". (EPA/822/R-
21/010). Washington, DC: EPA Office of Water, Health and Ecological Criteria Division.
https://www.epa.eov/chemical-research/hiiman-health4oxicity-assessments-eenx-
chemicals
4-26
-------
FINAL
2024
USEPA (U.S. Environmental Protection Agency). (2021d). Human Health Toxicity Values for
Perfluorobutane Sulfonic Acid (CASRN 375-73-5) and Related Compound Potassium
Perfluorobutane Sulfonate (CASRN 29420-49-3). (EPA/600/R-20/345F). Washington,
DC: EPA Office of Research and Development.
http s: //cfpub. epa. gov/n cea/ri sk/recordi spl ay. cfm ? dei d=3 5 08 8 8
USEPA (U.S. Environmental Protection Agency). (2021e). Systematic Review Protocol for the
PFBA, PFHxA, PFHxS, PFNA, and PFDA (anionic and acid forms) IRIS Assessments.
(EPA/63 5/R-2 0/131). Washington, DC: EPA Office of Research and Development.
https://ordspub.epa.eov/ords/eims/eimscomm.eetfile7p download id=541666
USEPA (U.S. Environmental Protection Agency). (2021f). Draft Framework for Estimating
Noncancer Health Risks Associated with Mixtures of Per- and Polyfluoroalkyl
Substances (PFAS). (EPA-822-D-21-003). Washington, DC: U.S. Environmental
Protection Agency.
https://sab.epa.eov/ords/sab/f?n iOO i s 10 '< I I ^ 11 v?88: * * * * * M * .601
USEPA (U.S. Environmental Protection Agency). (202lg). National Aquatic Resource Surveys.
National Coastal Condition Assessment 2015. http://www.epa.eov/national-aquatic-
resource-survevs/data-national-aquatic-resource-survevs
USEPA (U.S. Environmental Protection Agency). (2022a). Per- and Polyfluoroalkyl Substances
(PFAS) in Pesticide and Other Packaging. Washington, DC: EPA, Office of Pesticides
Program, https://www.epa.eov/pesticides/pfas-packaeine
USEPA (U.S. Environmental Protection Agency). (2022b). Drinking Water Health Advisory:
Perfluorobutane Sulfonic Acid (CASRN 375-73-5) and Related Compound Potassium
Perfluorobutane Sulfonate (CASRN 29420-49-3). (EPA/822/R-22/006). Washington,
DC: EPA, Office of Water, Health and Ecological Criteria Division.
https://www.epa.eov/sdwa/drinkine-water-health-advisories-eenx-chemicals-and-pfb
USEPA (U.S. Environmental Protection Agency). (2022c). Toxics Release Inventory (TRI)
Explorer - Release Reports: Release Chemical Report Page. (2020 Updated dataset
(released October 2021)) (Internet database).
https://enviro.epa.eov/triexplorer/ ase. chemical
USEPA (U.S. Environmental Protection Agency). (2022d). Initial List of Hazardous Air
Pollutants with Modifications. Research Triangle Park, NC: EPA Air Toxics Assessment
Group, https://www.epa.gov/haps/initial-list-hazardous-air-pollutants-
modifications#mods
USEPA (U.S. Environmental Protection Agency). (2022e). EPA's CompTox Chemicals
Dashboard Ammonium Perfluoro-2-Methyl-3-Oxahexanoate 62037-80-3.
https://comptox.epa.eov/dashboard/chemical/details/DTXSID40108559
USEPA (U.S. Environmental Protection Agency). (2022f). EPA's CompTox Chemicals
Dashboard for Perfluoro-2-Methyl-3-Oxahexanoic Acid 13252-13-6.
https://comptox.epa.eov/dashboard/chemical/details/DTXSID401i
USEPA (U.S. Environmental Protection Agency). (2023a). EPA Office of Research and
Development Integrated Risk Information System (IRIS) Program Outlook.
www.epa.eov/sYStem/files/documents/2023-10/iris proeram outlook oct2023.pdf
USEPA (U.S. Environmental Protection Agency). (2023b). Advances in dose addition for
chemical mixtures: A white paper. Washington, DC: U.S. Environmental Protection
Agency. https://assessments.epa.eov/risk/document/&deid:
4-27
-------
FINAL
2024
USEPA (U.S. Environmental Protection Agency). (2023c). PFAS National Primary Drinking
Water Regulation Rulemaking. Final Rule. (RIN 2040-AG18). Washington, DC:
Submitted for publication in the Federal Register April 2024.
USEPA (U.S. Environmental Protection Agency). (2024a). Framework for Estimating
Noncancer Health Risks Associated with Mixtures of Per- and Polyfluoroalkyl
Substances (PFAS). (EPA-815-R-24-003). Washington, DC: EPA, Office of Water and
Office of Research and Development.
USEPA (U.S. Environmental Protection Agency). (2024b). Per- and Polyfluoroalkyl Substances
(PFAS) Occurrence & Contaminant Background Support Document for the Final PFAS
National Primary Drinking Water Regulation. (EPA-815-R-24-013). Washington, DC:
EPA, Office of Groundwater and Drinking Water.
USEPA (U.S. Environmental Protection Agency). (2024c). OW Final Maximum Contaminant
Level Goals (MCLGs) for Perfluorooctanoic Acid (PFOA) and Perfluorooctane Sulfonic
Acid (PFOS) in Drinking Water. (EPA-815-R-24-010). Washington, DC: EPA, Office of
Water.
USEPA (U.S. Environmental Protection Agency). (2024d). Economic analysis for the final per-
and polyfluoroalkyl substances national primary drinking water regulation appendices.
(815R24002). Washington, DC: U.S. Environmental Protection Agency.
USEPA SAB (U.S. Environmental Protection Agency Science Advisory Board). (2022). Review
of EPA's Analysis to Support EPA's National Primary Drinking Water Rulemaking for
PFAS. (EPA-SAB-22-008). Washington, DC.
https:// sab. epa. gov/ords/ sab/f?p= 100:1 \ I i90947S * - * * - M * _601#report
Valsecchi, S; Rusconi, M; Mazzoni, M; Viviano, G; Pagnotta, R; Zaghi, C, et al. (2015).
Occurrence and sources of perfluoroalkyl acids in Italian river basins. Chemosphere 129:
126-134. http://dx.doi.org/10.1016/j .chemosphere.201 SO OSS
van der Veen, I; Hanning, AC; Stare, A; Leonards, PEG; de Boer, J; Weiss, JM. (2020). The
effect of weathering on per- and polyfluoroalkyl substances (PFASs) from durable water
repellent (DWR) clothing. Chemosphere 249: 126100.
http://dx.doi.on 10 101 t.chemosphere.2020.126100
Vassiliadou, I; Costopoulou, D; Kalogeropoulos, N; Karavoltsos, S; Sakellari, A; Zafeiraki, E, et
al. (2015). Levels of perfluorinated compounds in raw and cooked Mediterranean finfish
and shellfish. Chemosphere 127: 117-126.
http://dx.doi.org/10.1016/i.chemosphere.l I l.
Vavrous, A; Vapenka, L; Sosnovcova, J; Kejlova, K; Vrbik, K; Jirova, D. (2016). Method for
analysis of 68 organic contaminants in food contact paper using gas and liquid
chromatography coupled with tandem mass spectrometry. Food Control 60: 221-229.
https://doi.oi ^ 10 101 >/j .foodcont _ 0! ^ 0 0 l'<
Veillette, J; Muir, DCG; Antoniades, D; Small, JM; Spencer, C; Loewen, TN, et al. (2012).
Perfluorinated chemicals in meromictic lakes on the northern coast of Ellesmere Island,
High Arctic Canada. Arctic 65: 245-256.
Venkatesan, AK; Halden, RU. (2013). National inventory of perfluoroalkyl substances in
archived U.S. biosolids from the 2001 EPA National Sewage Sludge Survey. Journal of
Hazardous Materials 252-253: 413-418.
https://dot.org/https://doi.org/)0 101 3/j.jhazmat..01 '< 0 '< 01
4-28
-------
FINAL
2024
Venkatesan, AK; Halden, RU. (2014). Loss and in situ production of perfluoroalkyl chemicals in
outdoor biosolids-soil mesocosms. Environmental research 132: 321-327.
https://doi.oi\ 10 101 >/i.envres.2PI I 0 I 0J I
Vestergren, R; Berger, U; Glynn, A; Cousins, IT. (2012). Dietary exposure to perfluoroalkyl
acids for the Swedish population in 1999, 2005 and 2010. Environment international 49:
120-127. https://doi.ore/10.1016/i.envint.2012.C
Vestergren, R; Herzke, D; Wang, T; Cousins, IT. (2015). Are imported consumer products an
important diffuse source of PFASs to the Norwegian environment? Environmental
Pollution 198: 223-230. http://dx.doi.org/10J016/ienvpol.JO I I I J 0M
Vestergren, R; Orata, F; Berger, U; Cousins, IT. (2013). Bioaccumulation of perfluoroalkyl acids
in dairy cows in a naturally contaminated environment. Environmental Science and
Pollution Research 20: 7959-7969. https://doi.oi
Villaverde-de-Saa, E; Racamonde, I; Quintana, JB; Rodil, R; Cela, R. (2012). Ion-pair sorptive
extraction of perfluorinated compounds from water with low-cost polymeric materials:
Polyethersulfone vs polydimethylsiloxane. Analytica Chimica Acta 740: 50-57.
https://dot.or g/https://doi.org/)0 101 3/j.aca.2Q IJ 0 >.027
von Ehrenstein, OS; Fenton, SE; Kato, K; Kuklenyik, Z; Calafat, AM; Hines, EP. (2009).
Polyfluoroalkyl chemicals in the serum and milk of breastfeeding women. Reproductive
toxicology (Elmsford, NY) 27: 239-245. https://doi.o: /i.reprotox.2009.03.001
Vughs, D; Baken, KA; Dingemans, MML; de Voogt, P. (2019). The determination of two
emerging perfluoroalkyl substances and related halogenated sulfonic acids and their
significance for the drinking water supply chain. Environmental Science: Processes &
Impacts 21: 1899-1907. http://dx.doi.org/10.1039/c9em00393b
Wagner, A; Raue, B; Brauch, HJ; Worch, E; Lange, FT. (2013). Determination of adsorbable
organic fluorine from aqueous environmental samples by adsorption to polystyrene-
divinylbenzene based activated carbon and combustion ion chromatography. Journal of
Chromatography A 1295: 82-89. http://dx.doi.org/10.1016/i.chroma.201 > 0 I 0 ^ I
Walker, NJ; Crockett, PW; Nyska, A; Brix, AE; Jokinen, MP; Sells, DM, et al. (2005). Dose-
additive carcinogenicity of a defined mixture of "dioxin-like compounds". Environmental
health perspectives 113: 43-48. https://doi.org/10.1289/ehp.7351
Wang, Q; Tsui, M; Ruan, Y; Lin, H; Zhao, Z; Ku, J, et al. (2019a). Occurrence and distribution
of per- and polyfluoroalkyl substances (PFASs) in the seawater and sediment of the
South China sea coastal region. Chemosphere 231: 468-477.
https://doi.oi \ 10 101 '/i.chemosphert' _^ I ^ I .
Wiels0e, M; Kern, P; Bonefeld-J0rgensen, EC. (2017). Serum levels of environmental pollutants
is a risk factor for breast cancer in Inuit: a case control study. Environmental Health 16:
56. https://doi.oi^ 10 I Is >/sll's iO 01 0. 9-6
Wilkinson, JL; Hooda, PS; Swinden, J; Barker, J; Barton, S. (2017). Spatial distribution of
organic contaminants in three rivers of Southern England bound to suspended particulate
material and dissolved in water. Science of the Total Environment 593-594: 487-497.
http://dx.doi.org/10 J 016/i.scitotenv.l01 0 '< I
Winkens, K; Giovanoulis, G; Koponen, J; Vestergren, R; Berger, U; Karvonen, AM, et al.
(2018). Perfluoroalkyl acids and their precursors in floor dust of children's bedrooms -
Implications for indoor exposure. Environment international 119: 493-502.
http://dx.doi.org/10 J 016/i.envint _ 01 \ 0 009
4-29
-------
FINAL
2024
Wolf, CJ; Rider, CV; Lau, C; Abbott, BD. (2014). Evaluating the additivity of perfluoroalkyl
acids in binary combinations on peroxisome proliferator-activated receptor-a activation.
Toxicology 316: 43-54. https://doi.org/10 J 016/i .tox JO l'< i J 00 J
Wolf, CJ; Schmid, JE; Lau, C; Abbott, BD. (2012). Activation of mouse and human peroxisome
proliferator-activated receptor-alpha (PPARalpha) by perfluoroalkyl acids (PFAAs):
further investigation of C4-C12 compounds. Reproductive Toxicology 33: 546-551.
https://doi.on 10 101 >/j.reprotox JO i I 0'> 009
Wong, F; Shoeib, M; Katsoyiannis, A; Eckhardt, S; Stohl, A; Bohlin-Nizzetto, P, et al. (2018).
Assessing temporal trends and source regions of per- and polyfluoroalkyl substances
(PFASs) in air under the Arctic Monitoring and Assessment Programme (AMAP).
Atmospheric Environment 172: 65-73. https://doi.on 10 101 /i.atmosem JO I I0 0JN
Wu, XM; Bennett, DH; Calafat, AM; Kato, K; Strynar, M; Andersen, E, et al. (2014). Serum
concentrations of perfluorinated compounds (PFC) among selected populations of
children and adults in California. Environmental research 136C: 264-273.
http://dx.doi.org/10.1016/i .envres.2014.09.026
Yamada, A; Bemrah, N; Veyrand, B; Pollono, C; Merlo, M; Desvignes, V, et al. (2014).
Perfluoroalkyl acid contamination and polyunsaturated fatty acid composition of French
freshwater and marine fishes. Journal of agricultural and food chemistry 62: 7593-7603.
http://dx.doi.org/10J02 l/jt-01 I I >|
Ye, X; Strynar, MJ; Nakayama, SF; Varns, J; Helfant, L; Lazorchak, J; Lindstrom, AB. (2008).
Perfluorinated compounds in whole fish homogenates from the Ohio, Missouri, and
Upper Mississippi Rivers, USA. Environmental pollution (Barking, Essex : 1987) 156:
1227-1232. https://doi.org/10.1016/i envpol.20Qs 0 '<01 I
Yeung, LWY; Stadey, C; Mabury, SA. (2017). Simultaneous analysis of perfluoroalkyl and
polyfluoroalkyl substances including ultrashort-chain C2 and C3 compounds in rain and
river water samples by ultra performance convergence chromatography. Journal of
Chromatography A 1522: 78-85. http://dx.doi.org/10.1016/i.chroma.201 0° 0 1°
Yoo, H; Washington, JW; Jenkins, TM; Ellington, JJ. (2011). Quantitative determination of
perfluorochemicals and fluorotelomer alcohols in plants from biosolid-amended fields
using LC/MS/MS and GC/MS. Environmental science & technology 45: 7985-7990.
https://doi.org/10.1021/esl02972m
Young, W; Wiggins, S; Limm, W; Fisher, CM; DeJager, L; Genualdi, S. (2022). Analysis of Per-
and Poly(fluoroalkyl) Substances (PFASs) in Highly Consumed Seafood Products from
U.S. Markets. Journal of agricultural and food chemistry 70: 13545-13553.
https://doi.org/lQ.lQ21/acs.iafc.2cQ4673
Young, WM; South, P; Begley, TH; Diachenko, GW; Noonan, GO. (2012). Determination of
perfluorochemicals in cow's milk using liquid chromatography-tandem mass
spectrometry. Journal of agricultural and food chemistry 60: 1652-1658.
http s: //dot. or g/10.1021 /i £2045 65x
Young, WM; South, P; Begley, TH; Noonan, GO. (2013). Determination of perfluorochemicals
in fish and shellfish using liquid chromatography-tandem mass spectrometry. Journal of
agricultural and food chemistry 61:11166-11172. https://doi.org/10.1021/if403935g
Zafeiraki, E; Costopoulou, D; Vassiliadou, I; Bakeas, E; Leondiadis, L. (2014). Determination of
perfluorinated compounds (PFCs) in various foodstuff packaging materials used in the
Greek market. Chemosphere 94: 169-176.
https://doi.o A .chemosphere.2013.09.092
4-30
-------
FINAL
2024
Zafeiraki, E; Costopoulou, D; Vassiliadou, I; Leondiadis, L; Dassenakis, E; Hoogenboom, R;
van Leeuwen, SPJ. (2016a). Perfluoroalkylated substances (PFASs) in home and
commercially produced chicken eggs from the Netherlands and Greece. Chemosphere
144: 2106-2112. http://dx.doi.org/10. 1016/j.chemosphere.201 ^ 10 iOs
Zafeiraki, E; Vassiliadou, I; Costopoulou, D; Leondiadis, L; Schafft, HA; Hoogenboom, R; van
Leeuwen, SPJ. (2016b). Perfluoroalkylated substances in edible livers of farm animals,
including depuration behaviour in young sheep fed with contaminated grass.
Chemosphere 156: 280-285. http://dx.doi.org/10.1016/i.chemosphere..01 0^ 003
Zhang, X; Lohmann, R; Dassuncao, C; Hu, XC; Weber, AK; Vecitis, CD; Sunderland, EM.
(2016). Source attribution of poly- and perfluoroalkyl substances (PFASs) in surface
waters from Rhode Island and the New York Metropolitan Area. Environmental science
& technology letters 3: 316-321. https://doi.< l/acs.estlett.6b00255
Zhao, Z; Xie, Z; Tang, J; Sturm, R; Chen, Y; Zhang, G; Ebinghaus, R. (2015). Seasonal
variations and spatial distributions of perfluoroalkyl substances in the rivers Elbe and
lower Weser and the North Sea. Chemosphere 129: 118-125.
http://dx.doi.on 10 101 t.chemosphere..01 I 0 '< 0 SO
Zheng, G; Boor, BE; Schreder, E; Salamova, A. (2020). Indoor exposure to per- and
polyfluoroalkyl substances (PFAS) in the childcare environment. Environmental
pollution (Barking, Essex : 1987) 258: 113714.
https://doi.org/l 0.1016/i .envpol.2
Zhu, H; Kannan, K. (2019). Distribution and partitioning of perfluoroalkyl carboxylic acids in
surface soil, plants, and earthworms at a contaminated site. The Science of the total environment
647: 954-961. https://doi.org/10.1016/i .scitoten v.2018.08.051
4-31
-------
FINAL
2024
Appendix A. HFPO-DA: Summary of Occurrence
in Water and Detailed Relative Source
Contribution
A.l. Occurrence in Water
HFPO-DA can enter the aquatic environment through industrial discharges, runoff into surface
water, and leaching into groundwater from soil and landfills (USEPA, 2021c). HFPO-DA is
water-soluble, with solubilities of greater than 751 grams per liter (g/L) and greater than 739 g/L
for HFPO-DA and its ammonium salt, respectively, at 20°C (USEPA, 2021c). Volatilization
from water surfaces is expected to be an important fate process for HFPO-DA and its ammonium
salt (USEPA, 2021c).
A. 1.1. Ground Water
Petre et al. (2021) quantified the mass transfer of PFAS, including HFPO-DA, from
contaminated groundwater to five tributaries of the Cape Fear River. All sampling sites were
located within 5 km of a manufacturing plant known known to be a major source of PFAS
contamination. HFPO-DA and another fluoroether (perfluoro-2-[perfluoromethoxy] propanoic
acid) together accounted for 61% of the total quantified PFAS. The study authors calculated that
approximately 32 kg/year of PFAS is discharged from contaminated groundwater to the five
tributaries. These data indicate that the discharge of contaminated groundwater has led to long-
term contamination from HFPO-DA in surface water and could lead to subsequent impacts on
downstream drinking water (Petre et al., 2021).
A. 1.2. Surface Water
Chemours has reported that HFPO-DA has been discharged into the Cape Fear River for several
decades as a byproduct of other manufacturing processes (NCDEQ, 2017). An exposure
assessment of the Chemours Fayetteville Works Facility located in Bladen County, North
Carolina evaluated HFPO-DA in surface water data collected between July 2017 and October
2019 at four locations in the Cape Fear River (one upstream, one facility-adjacent location, and
two downstream), one pond located in the facility site, and one offsite pond (Geosyntec, 2019).
HFPO-DA was detected in surface water at the six sampled locations where mean (range)
concentrations were 5 ng/L (n=l), 23.37 ng/L (2.1-160 ng/L; n=26), 133.6 ng/L (8.6-580 ng/L;
n= 17), 16.38 ng/L (2.14-76 ng/L; n=79), 800 ng/L (730-940 ng/L; n=4), and 303.3 ng/L (290-
310 ng/L; n=3), respectively (Geosyntec, 2019). Additionally, several other studies evaluated the
occurrence of HFPO-DA in surface waters in North America (see Table A-l). As noted in the
EPA's human health toxicity assessment for HFPO-DA (USEPA, 2021c), HFPO-DA was first
detected in North Carolina's Cape Fear River and its tributaries in the summer of 2012 (Pritchett
et al., 2019; Strynar et al., 2015). Since that finding, U.S. studies of surface waters, some of
which are source waters for PWSs, have reported results of sampling efforts from contaminated
areas near the Cape Fear River (McCord et al., 2018; Sun et al., 2016) and in Ohio and West
Virginia (Galloway et al., 2020).
A-l
-------
FINAL
2024
In studies of the Cape Fear River basin by McCord et al. (2018) and Sun et al. (2016), surface
water concentrations of HFPO-DA ranged from below the North Carolina Department of Health
and Human Services (NCDHHS) provisional health goal (PHG) of 140 ng/L to a maximum level
of 4,560 ng/L. Sun et al. (2016) analyzed surface water from two sites upstream of a drinking
water treatment plant (DWTP) and one site downstream. They reported a median HFPO-DA
concentration of 304 ng/L with a maximum of 4,560 ng/L in the source water of the plant.
HFPO-DA levels did not exceed the quantitation limit (10 ng/L) at the two upstream locations. In
source water samples collected from the Cape Fear River near a DWTP downstream of a
fluorochemical manufacturer, McCord et al. (2018) reported initial HFPO-DA concentrations of
approximately 700 ng/L. After the manufacturer diverted waste stream emissions from one of its
manufacturing lines, the measured concentrations decreased to levels below the NCDHHS PHG
(140 ng/L).
In Ohio and West Virginia, Galloway et al. (2020) sampled rivers and streams located upstream,
downstream, and downwind to the north and northeast of the Chemours Washington Works
facility outside Parkersburg, West Virginia. The downwind sampling was intended to explore
potential airborne deposition. Some of the downstream sampling sites were in the vicinity of
landfills. Reported levels of HFPO-DA in these waters ranged from non-detectable levels to a
maximum of 227 ng/L. The highest HFPO-DA concentrations were measured downwind of the
facility (i.e., to the northeast). The study observed an exponentially declining trend of HFPO-DA
concentrations in surface water with distance from the facility in this direction and attributed its
occurrence in surface water to air dispersion of emissions from the facility. The most distant site
where HFPO-DA was detected was 24 km north of the facility.
In one study of sites located in highly industrialized commercial waterways (authors did not
indicate whether sampling sites were in the vicinity of known PFAS point sources), Pan et al.
(2018) detected HFPO-DA in 100% of samples from sites in the Delaware River (n=12),
reporting median and maximum concentrations of 2.02 ng/L and 8.75 ng/L, respectively, in
surface waters.
Globally, HFPO-DA occurrence has been reported in surface waters from Germany (Pan et al.,
2018; Heydebreck et al., 2015), China (Li et al., 2020a; Pan et al., 2018; Song et al., 2018; Pan et
al., 2017; Heydebreck et al., 2015), the Netherlands (Pan et al., 2018; Gebbink et al., 2017;
Heydebreck et al., 2015), the United Kingdom (Pan et al., 2018), South Korea (Pan et al., 2018),
and Sweden (Pan et al., 2018). HFPO-DA was also detected with a mean concentration of 30
picograms per liter (pg/L; 0.030 ng/L) in Artie seawater samples, suggesting long-range transport
(Joerss et al., 2020).
In one study of surface water collected from industrialized areas in Europe (authors did not
indicate whether sampling sites were in the vicinity of known PFAS point sources), Pan et al.
(2018) reported HFPO-DA detections in 100% of samples from the Thames River in the United
Kingdom (n=6 sites), the Rhine River in Germany and the Netherlands (n=20 sites), and the
Malaren Lake in Sweden (n=10 sites). Across these three river systems, median HFPO-DA
concentrations ranged from 0.90 to 1.38 ng/L and the highest concentration detected was 2.68
ng/L. In another study, Heydebreck et al. (2015) detected HFPO-DA at 17% of sampling
locations on the industrialized non-estuarine reaches of the Rhine River, with a maximum
concentration of 86.08 ng/L; however, HFPO-DA was not detected at locations on the Elbe
River.
A-2
-------
FINAL
2024
Gebbink et al. (2017) evaluated surface water samples upstream and downstream of a
fluorochemical production plant in the Netherlands and reported only one of three samples
upstream of the plant with detectable HFPO-DA concentrations (22 ng/L; method quantification
limit [MQL] = 0.2 ng/L). Downstream of the fluorochemical plant, HFPO-DA was detected in
100% of samples, with a mean concentration of 178 ng/L and a range of 1.7 to 812 ng/L. Vughs
et al. (2019) analyzed surface water from 11 water suppliers in the Netherlands and Belgium,
some of which were in the vicinity of a fluoropolymer manufacturing plant. The authors reported
HFPO-DA detections in 77% of surface water samples (n=13) with a mean concentration of 2.2
ng/L and a maximum of 10.2 ng/L; however, only three samples in the study had HFPO-DA
concentrations exceeding 1 ng/L.
Of the five studies conducted in China, one study evaluated surface water samples from an
industrialized region (authors did not indicate whether sampling sites were in the vicinity of
known PFAS point sources) (Pan et al., 2018), one study evaluated surface water river and
reservoir samples in an industrialized river basin with potential PFAS point sources (Li et al.,
2020a), and three studies examined samples from sites along the Xiaoqing river at locations
upstream, downstream, or in the vicinity of known PFAS sources (Song et al., 2018; Pan et al.,
2017; Heydebreck et al., 2015). HFPO-DA was detected in freshwater systems sampled in all
five studies, though HFPO-DA concentrations appeared to be positively correlated with
proximity to known PFAS point sources. Song et al. (2018), Pan et al. (2017), and Heydebreck et
al. (2015) sampled sites in the Xiaoqing River system, including one of its tributaries, nearby a
known fluoropolymer production facility. These three studies reported maximum HFPO-DA
concentrations of 9,350, 2,060, and 3,060 ng/L, respectively. HFPO-DA concentrations in
samples collected upstream of the facility did not exceed 3.64 ng/L. Other Chinese freshwater
systems evaluated in the other two studies (Li et al., 2020a; Pan et al., 2018) generally reported
maximum concentrations like those from the upstream Xiaoqing River system sites (< 10.3
ng/L), except for one site in Tai Lake which was reported to have a maximum HFPO-DA
concentration of 143 ng/L. Similarly, in a study that sampled an industrialized river in South
Korea (authors did not report whether sampling sites were in the vicinity of known PFAS point
sources), HFPO-DA was found in 100% of samples and the maximum concentration found was
2.49 ng/L (Pan et al., 2018).
A-3
-------
FINAL
2024
Table A-l. Compilation of Studies Describing HFPO-DA Occurrence in Surface Water
Study
Location
Site Details
Results
North America
Sun et al.
(2016)
United States
(North Carolina,
Cape Fear River
Basin)
Source waters of three
community drinking water
treatment plants, two
upstream and one
downstream of a PFAS
manufacturing plant (LOQ
= 10 ng/L)
Community A (upstream):
DF 0%
Community B (upstream):
DF NR, median (range) =
ND (ND-10 ng/L)
Community C
(downstream): DF NR,
mean = 631 ng/L, median
(range) = 304 (55-4,560)
ng/L
McCord et al.
(2018)
United States
(North Carolina,
Cape Fear River
Basin)
Source water of a drinking
water treatment plant near
the industrial waste outfall
of a fluorochemical
manufacturer, before and
after the manufacturer
diverted a waste stream
(exact values NR,
estimated values from
Figure 3)
Before waste diversion
(estimated): DF NR,
measured concentration = ~
>700 ng/L
After waste diversion
(estimated): DRNR,
measured concentration = <
140 ng/L
Galloway et al.
(2020)
United States (Ohio
and West Virginia,
Ohio River Basin)
Rivers and tributaries
located upstream,
downstream, and
downwind of a
fluoropolymer production
facility; some sample
locations potentially
impacted by local landfills
DF = 21/24 unique sites
with detections > LOQ,
mediana (range) = 46.7
(ND-227) ng/L
Europe
Gebbink et al.
(2017)
The Netherlands
Upstream and downstream
of the Dordrecht
fluorochemical production
plant; two control sites
Control sites: DF 0%
Upstream of plant (n=3):
DFa 33%, point = 22 ng/L
Downstream of plant
(n=13): DF 100%, meana
(range) = 178 (1.7-812)
ng/L
(MQL = 0.2)
A-4
-------
FINAL
2024
Study
Location
Site Details
Results
Vughs et al.
(2019)
The Netherlands
and Belgium
Thirteen surface water
samples collected from
eleven water suppliers,
some near a
fluoropolymer
manufacturing plant. The
study did not map the
distribution of reported
concentrations by
geographic location or
with respect to distance
from the fluoropolymer
manufacturing plant.
DF 77%, mean (range) = 2.2
(ND-10.2) ng/L (LOQ = 0.2
ng/L)
Asia
Pan et al. (2017)
China (Xiaoqing
River and tributary)
Upstream and downstream
of a fluoropolymer
production plant in an
industrialized region
Upstream of plant in the
Xiaoqing River (n=6): DFa
100%, mediana (range) =
2.10(1.61-3.64) ng/L
Tributary directly receiving
plant effluent (n=4): DFa
100%), mediana (range) =
1,855 (2.34-2,060) ng/L
Downstream of plant in the
Xiaoqing River receiving
tributary waters (n=8): DFa
100%), mediana (range) =
311 (118-960) ng/L
Song et al.
(2018)
China (Xiaoqing
River)
Near the Dongyue group
industrial park, including
a fluoropolymer
production plant
DF NR, mean, median
(range) = 519, 36.7 (
-------
FINAL
2024
Study
Location
Site Details
Results
Heydebreck et
al. (2015)
Germany (Elbe and
Rhine Rivers), the
Netherlands (Rhine-
Meuse delta)
All sampling locations in
industrialized areas
Rhine River (n=23): DFa
17%, range = ND-86.08
ng/L
Elbe River (n=22): DF 0%
China (Xiaoqing
River)
Some sampling locations
were downstream of
PFAS point sources
Xiaoqing River (n=20): DFa
65%, range = ND-3,060
ng/L
Pan et al. (2018)
United States
(Delaware River)
Sampling sites along
industrialized river
systems that were not
proximate to known point
sources of PFAS from
fluorochemical facilities
Delaware River (n=12): DF
100%), mean, median
(range) = 3.32, 2.02 (0.78-
8.75) ng/L
United Kingdom
(Thames River),
Germany and the
Netherlands (Rhine
River), Sweden
(Malaren Lake)
Sampling sites along
industrialized river
systems that were not
proximate to known point
sources of PFAS from
fluorochemical facilities
Thames River (n=6): DF
100%), mean, median
(range) = 1.12, 1.10(0.70-
1.58) ng/L
Rhine River (n=20): DF
100%o, mean, median
(range) = 0.99, 0.90 (0.59-
1.98) ng/L
Malaren Lake (n=10): DF
100%o, mean, median
(range) = 1.47, 1.38 (0.88-
2.68) ng/L
A-6
-------
FINAL
2024
Study
Location
Site Details
Results
South Korea (Han
River), China (Liao,
Huai, Yellow,
Yangtze, and Pearl
Rivers; Chao and
Tai Lakes)
Sampling sites along
industrialized river
systems that were not
proximate to known point
sources of PFAS from
fluorochemical facilities
Han River (n=6): DF 100%,
mean, median (range) =
1.38, 1.16 (0.78-2.49) ng/L
Liao River (n=6): DF 100%,
mean, median (range) =
1.44, 0.88 (0.62-4.51) ng/L
Huai River (n=9): DF 100%,
mean, median (range) =
1.66, 1.40 (0.83-3.62) ng/L
Yellow River (n=15): DF
67%), mean, median (range)
= 1.01, 1.30 (< LOQ-1.74)
ng/L
Yangtze River (n=35): DF
94%), mean, median (range)
= 0.73, 0.67 (< LOQ-1.54)
ng/L
Pearl River (n=13): DF
100%o, mean, median
(range) = 1.51, 0.70 (0.21-
10.3) ng/L
Chao Lake (n=13): DF
100%o, mean, median
(range) = 1.92, 1.81 (0.93-
3.32) ng/L
Tai Lake (n=15): DF 100%,
mean, median (range) =
14.0, 0.77 (0.38-143.7)
ng/L
(LOQ = 0.05 ng/L; MDL =
0.38 ng/L)
All locations
Sampling sites were not
proximate to known point
sources of any
fluorochemical facilities
All locations (n=160): DF
96%o, mean, median (range)
= 2.55, 0.95 (0.18-144)
ng/L
(LOQ = 0.05 ng/L; MDL =
0.38 ng/L)
Notes:
DF = detection frequency; LOQ = limit of quantification; ND = not detected.; ng/L = nanograms per liter; NR = not reported;
MQL = method quantification limit; MDL = method detection limit.
a The DF, median, and/or mean was not reported in the study and was calculated in this synthesis. Mean values were only
calculated if DF = 100%.
A-7
-------
FINAL
2024
b The DF in Li et al. (2020a) was reported as 82.5% in the main article. The DF of 80% shown in this table is based on the
supporting information data, which show only 32/40 samples with data > MDL.
c The Xiaoqing River results reported in Heydebreck et al. (2015) included samples from Laizhou Bay. The EPA considered
freshwater samples only.
A.2. RSC for HFPO-DA, Literature Search and Screening
Methodology
The EPA applies an RSC to the RfD when calculating an MCLG based on noncancer effects or
for carcinogens that are known to act through a nonlinear mode of action to account for the
fraction of an individual's total exposure allocated to drinking water (USEPA, 2000b). The EPA
emphasizes that the purpose of the RSC is to ensure that the level of a chemical allowed by a
criterion (e.g., the MCLG for drinking water) or multiple criteria, when combined with other
identified sources of exposure (e.g., diet, ambient and indoor air) common to the population of
concern, will not result in exposures that exceed the RfD. In other words, the RSC is the portion
of total daily exposure equal to the RfD that is attributed to drinking water ingestion (directly or
indirectly in beverages like coffee tea or soup, as well as from transfer to dietary items prepared
with drinking water) relative to other exposure sources; the remainder of the exposure equal to
the RfD is allocated to other potential exposure sources. For example, if for a particular
chemical, drinking water were to represent half of total exposure and diet were to represent the
other half, then the drinking water contribution (or RSC) would be 50%. The EPA considers any
potentially significant exposure source when deriving the RSC.
The RSC is derived by applying the Exposure Decision Tree approach published in the EPA's
Methodology for Deriving Ambient Water Quality Criteria for the Protection of Human Health
(USEPA, 2000b). The Exposure Decision Tree approach allows flexibility in the RfD
apportionment among sources of exposure and considers several characteristics of the
contaminant of interest, including the adequacy of available exposure data, levels of the
contaminant in relevant sources or media of exposure, and regulatory agendas (i.e., whether there
are multiple health-based criteria or regulatory standards for the contaminant). The RSC is
developed to reflect the exposure to the U.S. general population or a sensitive population within
the U.S. general population and may be derived qualitatively or quantitatively, depending on the
available data.
A quantitative RSC determination first requires "data for the chemical in question...
representative of each source/medium of exposure and... relevant to the identified population(s)"
(USEPA, 2000b). The term "data" in this context is defined as ambient sampling measurements
in the media of exposure, not internal human biomonitoring metrics. More specifically, the data
must adequately characterize exposure distributions including the central tendency and high-end
exposure levels for each source and 95% confidence intervals for these terms (USEPA, 2000b).
Frequently, an adequate level of detail is not available to support a quantitative RSC derivation.
When adequate quantitative data are not available, the agency relies on the qualitative
alternatives of the Exposure Decision Tree approach. A qualitatively-derived RSC is an estimate
that incorporates data and policy considerations and thus, is sometimes referred to as a "default"
RSC (USEPA, 2000b). Both the quantitative and qualitative approaches recommend a "ceiling"
RSC of 80%) and a "floor" RSC of 20% to account for uncertainties including unknown sources
A-8
-------
FINAL
2024
of exposure, changes to exposure characteristics over time, and data inadequacies (USEPA,
2000b).
In cases in which there is a lack of sufficient data describing environmental monitoring results
and/or exposure intake, the Exposure Decision Tree approach results in a recommended RSC of
20%. In the case of MCLG development, this means that 20% of the exposure equal to the RfD
is allocated to drinking water and the remaining 80% is reserved for other potential sources, such
as diet, air, consumer products, etc. This 20% RSC value can be replaced if sufficient data are
available to develop a scientifically defensible alternative value. If scientific data demonstrating
that sources and routes of exposure other than drinking water are not anticipated for a specific
pollutant, the RSC can be raised as high as 80% based on the available data, allowing the
remaining 20% for other potential sources (USEPA, 2000b). Applying a lower RSC (e.g., 20%)
is a more conservative approach to public health and results in a lower MCLG.
A.2.1. Literature Search and Screening
In support of the EPA's human health toxicity assessment for HFPO-DA (USEPA, 2021c),
literature searches were conducted of four databases (PubMed, Toxline, Web of Science (WOS),
and Toxic Substances Control Act Test Submissions (TSCATS)) to identify publicly available
literature using CASRN, synonyms, and additional relevant search strings (see USEPA (2021c)
for details). Due to the limited search results, additional databases were searched for information
on physicochemical properties, health effects, toxicokinetics, and mechanism of action. The
initial date-unlimited database searches were conducted in July 2017 and January/February 2018,
with updates completed in February 2019, October 2019, and March 2020. In addition,
information on toxicokinetics; acute, short-term, subchronic, and chronic toxicity; developmental
and reproductive toxicity; neurotoxicity; immunotoxicity; genotoxicity; and cancer in animals
was submitted with premanufacture notices to the EPA by DuPont/Chemours, the manufacturer
of HFPO-DA, as required under the Toxic Substances Control Act pursuant to a consent order
(USEPA, 2009b) or reporting requirements (15 U.S.C. § 2607.8(e)). The results of the literature
searches of publicly available sources and submitted studies from DuPont/Chemours are
available through the EPA's Health & Environmental Resource Online website at
https://hero.epa.eov/hero/index.cfm/proiect/paee/proiect id/2627.
The HFPO-DA literature search results and all studies submitted by DuPont/Chemours were
imported into SWIFT-Review (Sciome, LLC, Research Triangle Park, NC) and filtered through
the Evidence Stream tags to identify human studies and non-human (i.e., those not identified as
being in humans) studies. Studies identified as human studies were further categorized into seven
major PFAS pathway categories (Cleaning Products, Clothing, Environmental Media, Food
Packaging, Home Products/Articles/Materials, Personal Care Products, and Specialty Products)
as well as an additional category for Human Exposure Measures. Non-human studies were
grouped into the same seven major PFAS pathway categories, except that the Environmental
Media category did not include soil, wastewater, or landfill.
Application of the SWIFT-Review tags identified 52 studies for title and abstract screening. An
additional three references were identified through gray literature sources that were included to
supplement the search results. Title and abstract screening to determine relevancy followed the
populations, exposures, comparators, and outcomes (PECO) criteria in Table A-2:
A-9
-------
FINAL
2024
Table A-2. Populations, Exposures, Comparators, and Outcomes (PECO) Criteria
PECO Element
Inclusion Criteria
Population
Adults (including women of childbearing age) and/or children in the general
populations from any country
Exposure
Primary data from peer-reviewed studies collected in any of the following
media: ambient air, consumer products, drinking water, dust, food, food
packaging, groundwater, human blood/serum/urine, indoor air, landfill,
sediment, soil, surface water (freshwater), wastewater/biosolids/sludge
Comparator
Not applicable
Outcome
Measured concentrations of HFPO-DA (or measured emissions from food
packaging and consumer products only)
The title and abstract of each study were independently screened for relevance by two screeners
using litstream™. A study was included as relevant if it was unclear from the title and abstract
whether it met the inclusion criteria. When two screeners did not agree if a study should be
included or excluded, a third reviewer was consulted to make a final decision. The title and
abstract screening resulted in 24 studies tagged as relevant (i.e., data on occurrence of HFPO-DA
in one of the media of interest were presented in the study) that were further screened with full-
text review using the same inclusion criteria. Of these 24 studies, 4 contain only human
biomonitoring data and are not discussed further here. Based on full-text review, 15 studies were
identified as relevant and are summarized below. At the full-text review stage, two additional
studies were identified as only containing biomonitoring data.
A.2.2. Additional Screening
To supplement the primary literature database, the EPA also searched the following publicly
available gray literature sources in February 2022 for information related to relative exposure of
HFPO-DA for all potentially relevant routes of exposure (oral, inhalation, dermal) and exposure
pathways relevant to humans:
• USEPA (2021c) Human Health Toxicity Values for Hexajluoropropylene Oxide
(HFPO) Dimer Acid and Its Ammonium Salt (CASRN13252-13-6 and CASRN
62037-80-3) Also Known as "GenXChemicals
• AT SDR's Toxicological Profiles.,
• Centers for Disease Control and Prevention's (CDC's) national reports on human
exposures to environmental chemicals;
• EPA's CompTox Chemicals Dashboard;
• EPA's fish tissue studies;
• EPA's Toxics Release Inventory;
• Relevant documents submitted under the Toxic Substances Control Act and relevant
reports from the EPA's Office of Chemical Safety and Pollution Prevention;
A-10
-------
FINAL
2024
• U.S. Food and Drug Administration's (FDA's) Total Diet Studies and other similar
publications from FDA, U.S. Department of Agriculture, and Health Canada;
• National Oceanic and Atmospheric Administration's (NOAA's) National Centers for
Coastal Ocean Science data collections;
• National Science Foundation direct and indirect food and/or certified drinking water
additives;
• PubChem compound summaries;
• Relevant sources identified in the relative source contribution discussions (Section 5)
of the EPA's Proposed Approaches to the Derivation of a Draft Maximum
Contaminant Level Goal for Perfluorooctanoic Acid (PFOA)/Perfluorooctane
Sulfonic Acid (PFOS) in Drinking Watery and
• Additional sources, as needed.
The EPA has included available information from these gray literature sources for HFPO-DA
relevant to its uses, chemical and physical properties, and for occurrence in ambient or indoor
air, foods (including fish and shellfish), soil, dust, and consumer products. The EPA has also
included available information specific to HFPO-DA below on any regulations that may restrict
HFPO-DA levels in media (e.g., water quality standards, air quality standards, food tolerance
levels).
The EPA incorporated six references (Li et al., 2022; Burkhard, 2021; Feng et al., 2021;
Strakova et al., 2021; Semerad et al., 2020; Geosyntec, 2019) that were not identified in the
EPA's RSC literature search strategy; these references were provided by Chemours as part of
their outreach to the EPA on uses and sources for HFPO-DA in April 2022 and/or through the
public comment period for the proposed PFAS NPDWR in 2023.
A3. Summary of Potential Exposure Sources of HFPO-DA Other
than Water
A.3.1. Dietary Sources
HFPO-DA was included in a suite of individual PFAS selected as part of PFAS-targeted
reexaminations of samples collected for the U.S. Food and Drug Administration's (FDA's) Total
Diet Study (US FDA, 2022b, 2022a, 2021b, 2021a, 2020b, 2020a); however, it was not detected
in any of the food samples tested. It should be noted that FDA indicated that the sample sizes
were limited and that the results should not be used to draw definitive conclusions about PFAS
levels or presence in the general food supply (US FDA, 2022c). HFPO-DA was not detected in
cow milk samples collected from a farm with groundwater known to be contaminated with
PFAS; however, it was detected in produce (collard greens, cabbage) collected from an area near
a PFAS production plant in FDA studies of the potential exposure to the U.S. population to
PFAS (US FDA, 2021c, 2018). HFPO-DA was detected at low levels in 14% of vegetable
garden crops (endive, beets, celery, lettuce, and tomatoes) grown near a PFAS manufacturing
facility in the Netherlands (NCDEQ, 2018c; Mengelers et al., 2018). An exposure assessment of
the Chemours Fayetteville Works Facility located in Bladen County, North Carolina evaluated
HFPO-DA in fish fillet tissue samples collected between July and September 2019 at five
locations within the Cape Fear River (Geosyntec, 2019). HFPO-DA was detected in three
A-ll
-------
FINAL
2024
samples of largemouth bass from the Bladen Bluffs location which is approximately eight miles
downstream of the Fayetteville Works Facility at concentrations of 68,000 ng/kg, 54,000 ng/kg,
and 24,000 ng/kg (Geosyntec, 2019). HFPO-DA was not detected in fish (largemouth bass,
catfish, bluegill, or redbreast sunfish) collected from the other sites located upstream, adjacent to
the site, 4 miles downstream, or in an on-site pond.
Feng et al. (2021) measured HFPO-DA in food samples collected from up to ten home gardens
or farms in villages within 15 km of a large fluoropolymer facility located on the Dongzhulong
River in Shandong Province, China. The authors detected HFPO-DA in wheat (mean
concentration: 5.53 nanograms per gram dry weight [ng/g dw]; range: 2.27-9.19 ng/g dw;
detection frequency [DF] 100%), maize (mean concentration: 1.17 ng/g dw; range: not detected
(ND)-1.94 ng/g dw; DF 80%), and vegetable samples (mean concentration: 20.1 ng/g dw; range:
ND-67.2 ng/g dw; DF 82%). In fish collected at two sites along the Dongzhulong River, HFPO-
DA was detected at concentrations of 43.9 and 3.23 ng/g dw at sites approximately 3 km and 15
km downstream of the fluoropolymer facility, respectively. HFPO-DA was not found in eggs
(home-produced and store-bought), store-bought meat or seafood, or milk from domestic goats
(Feng et al., 2021). Except for the fish sampled at two sites, the study did not report HFPO-DA
concentrations in food according to sampling location or proximity to the fluoropolymer facility.
HFPO-DA was not a target chemical in the EPA's National Lake Fish Tissue Study or the EPA's
2015 Great Lakes Human Health Fish Fillet Tissue Study, nor in the EPA's 2008-2009 or 2013-
2014 National Rivers and Streams Assessment studies (USEPA, 2021e, 2020a; Stahl et al., 2014;
USEPA, 2009a). HFPO-DA was detected in a redear sunfish fillet composite sample collected
from a privately-owned lake near a PFAS manufacturing facility in North Carolina at a
concentration of 270 nanograms per kilogram (ng/kg) (wet weight tissue) (USEPA, 2021c;
NCDEQ, 2018c). HFPO-DA was not included in the NOAA's National Centers for Coastal
Ocean Science, National Status and Trends Data (NOAA, 2022). Burkhard (2021) identified a
single bioaccumulation factor (BAF) in muscle tissue/fillet reported in the literature of 4.07 L/kg
wet weight (reported as a logBAF of 0.61 L/kg).
A3.2. Food Contact Materials
In an analysis performed at the Department of Food Analysis and Nutrition of the University of
Chemistry and Technology in Prague, Czech Republic, HFPO-DA was not detected in 42
samples of disposable food packaging and tableware purchased from six different European
countries between May and December 2020 (LOQ =1.7 mg/kg) (Strakova et al., 2021).
A3.3. Consumer Products
Although no specific studies on the occurrence of HFPO-DA in consumer products were
identified, DuPont began transitioning to GenX processing aid technology in 2009 to work
toward eliminating long-chain PFAS as part of the company's commitment under the 2010/2015
PFOA Stewardship Program (USEPA, 2021c). It is unknown if HFPO-DA in consumer products
have increased as a result of this transition.
A3.4. indoor Dust
Feng et al. (2021) detected HFPO-DA in indoor dust samples taken from homes from 10 villages
within 15 km of a large fluoropolymer facility in Shandong Province, China, at concentrations
A-12
-------
FINAL
2024
ranging from ND to 841 ng/g (mean concentration 159 ng/g; DF 72%). Contaminated dust was
found in homes as far as 15 km from the fluoropolymer facility and HFPO-DA concentrations
were highest in homes nearest to the facility. Although only one study on the occurrence of
HFPO-DA in indoor dust was identified, other PFAS have been detected in indoor dust and on
window films (ATSDR, 2021).
A3.5. Air
PFAS have been released to air from wastewater treatment plants, waste incinerators, and
landfills (USEPA, 2016a). HFPO-DA could be transported in the vapor phase or with
particulates (USEPA, 2021c). When released to air or volatilized from water, HFPO-DA is stable
and short- and long-range transport has occurred (D'Ambro et al., 2021; Galloway et al., 2020).
Galloway et al. (2020) analyzed HFPO-DA concentrations in soil samples downwind of and
surface water samples upstream of the Chemours Washington Works facility outside of
Parkersburg, West Virginia, and results suggest atmospheric transport of HFPO-DA emissions.
Additionally, a study that modeled the atmospheric transport of a PFAS mixture containing
HFPO-DA from a fluoropolymer manufacturing facility in North Carolina (D'Ambro et al.,
2021) predicted that only 2.5% of total HFPO-DA (consisting of HFPO-DA and HFPO-DA
fluoride) would be deposited within 150 km of the facility (USEPA, 2021c).
HFPO-DA is persistent in air (half-lives longer than 6 months), and not readily broken down by
biodegradation, direct photolysis, or hydrolysis (USEPA, 2021c). In the vapor phase, HFPO-DA
is expected to undergo hydroxyl radical-catalyzed indirect photolysis slowly, with a predicted
average hydroxylation rate of 8.50 x 10-13 cubic centimeters (cm3)/molecule - second (USEPA,
2022f, 2022e, 2021c). Based on a measured vapor pressure of 2.7 mm Hg at 20°C for HFPO-
DA, volatilization is expected to be an important fate process for this chemical (USEPA, 2021c).
EPA's Toxics Release Inventory reported release data for HFPO-DA in 2020 (USEPA, 2022c).
HFPO-DA is not listed as a hazardous air pollutant (USEPA, 2022d).
HFPO-DA has been identified in air emissions. North Carolina Department of Environmental
Quality (NCDEQ) estimates for the Chemours Fayetteville Works plant, located in the North
Carolina Cape Fear watershed, indicate that annual emissions of HFPO-DA could have exceeded
2,700 pounds per year during the reporting period (2017-2018) (NCDEQ, 2018a). Rainwater
samples collected within a seven-mile radius of this facility were reported to have detectable
levels of HFPO-DA (NCDEQ, 2018b), with the highest concentration of 810 ng/L found in a
rainwater sample collected five miles from the facility. The three samples collected seven miles
from the plant had HFPO-DA concentrations ranging from 45.3 to 60.3 ng/L (NCDEQ, 2018b).
A3.6. Soil
When HFPO-DA is deposited on or applied to soil, it is expected to run off into surface waters or
rapidly leach to groundwater (USEPA, 2021c). PFAS can also be taken up from contaminated
soil by plants (ATSDR, 2021). No specific studies on the occurrence of HFPO-DA in biosolids
were identified. An exposure assessment of the Chemours Fayetteville Works Facility located in
Bladen County, North Carolina evaluated HFPO-DA in soil samples collected between July and
September 2019 at twelve offsite locations (Geosyntec, 2019). HFPO-DA was detected in two of
four surface soil samples (0 to 0.5 ft depth) collected within a 2.5 km radius of the facility at
concentrations of 2,600 ng/kg and 360 ng/kg. HFPO-DA was also detected in two of four
A-13
-------
FINAL
2024
subsurface soil samples (4 to 4.5 ft depth) collected within a 2.5 km radius of the facility at
concentrations of 430 ng/kg and 590 ng/kg, and in one of four subsurface samples (excluding
duplicates) collected within a 5 km radius to the facility at a concentration of 400 ng/kg
(Geosyntec, 2019). HFPO-DA was not detected in surface or subsurface soil samples collected in
a 10 km radius to the facility.
Two peer-reviewed studies reported HFPO-DA concentrations in soil. In the United States,
Galloway et al. (2020) analyzed 13 soil samples for HFPO-DA at locations in Ohio and West
Virginia that were upstream and downwind of the Chemours Washington Works facility in order
to evaluate HFPO-DA contamination due to atmospheric deposition. HFPO-DA was detected in
5 out of 13 samples, with a maximum concentration of 8.14 ng/g dw. In China, Li et al. (2020b)
collected and analyzed residential soil samples throughout the country from 31 provincial-level
administrative regions (consisting of 26 provinces, 4 municipalities, and 1 special administrative
region). HFPO-DA was detected in 40.5% of soil samples at concentrations up to 967 picograms
per gram (pg/g) dw and a mean level of 19.1 pg/g dw. PFOA was detected in these soils more
frequently (96.6%) and at higher mean levels (354 pg/g dw), leading the authors to conclude that
HFPO-DA consumption was still limited at the national scale of China, despite its use as a PFOA
replacement.
One study measured concentrations of HFPO-DA in and/or on grass and leaves collected from
sites various distances from a fluoropolymer manufacturing plant in the Netherlands (Brandsma
et al., 2019). HFPO-DA concentrations ranged from 86 ng/g in leaves from a site closest to the
plant to ND furthest from the plant. A similar pattern was observed for grass samples, except the
maximum HFPO-DA concentration was lower (27 ng/g). The study authors note that it hadn't
rained for five days prior to sample collection.
Semerad et al. (2020) investigated occurrence of HFPO-DA in sewage sludge from 43 facilities
in the Czech Republic. HFPO-DA was detected in 7 of 43 samples at concentrations ranging
from 0.3 to 1.2 ng/g dw. The authors raised concerns about the agricultural use of sludge
containing PFAS for growing crops.
A3.7. Sediment
HFPO-DA is expected to remain in water and exhibit low partitioning to sediment (USEPA,
2021c). One study evaluated the occurrence of HFPO-DA in sediments from the North and
Baltic Seas in Europe, and reported that HFPO-DA was not detected in any of the 24 sediment
samples taken in the North and Baltic Seas in the vicinity of Germany (Joerss et al., 2019). An
additional four studies analyzed sediments in China (Li et al., 2022; Li et al., 2020b; Wang et al.,
2019a; Song et al., 2018). Of the four studies, Wang et al. (2019a) analyzed sediment from the
South China Sea coastal region in the area of the highly industrialized Pearl River Delta and
reported that HFPO-DA was below the LOQ in all 53 samples. Li et al. (2020) analyzed 20
sediment samples from eight rivers and three reservoirs in the Hai River Basin in the vicinity of
several industrialized areas. HFPO-DA was reportedly detected at minimal levels, but the authors
did not report actual concentrations. Song et al. (2018) analyzed concentrations of HFPO-DA in
24 sediment samples from the Xiaoqing River in the vicinity of a fluoropolymer production
facility. The study reported a maximum HFPO-DA concentration in sediment of 22.3 ng/g dw,
with median and mean levels below the LOQ. Li et al. (2022) also analyzed sediment samples
A-14
-------
FINAL
2024
from five sites of the Xiaoqing River estuary and reported a mean HFPO-DA concentration of
0.23 ng/g dw.
A.4. Recommended RSC
The EPA followed the Exposure Decision Tree approach to determine the RSC for HFPO-DA
(USEPA, 2000b). The EPA first identified three potential populations of concern (Box 1):
pregnant women, lactating women, and women of childbearing age (see Section 2.1.2).
However, limited information was available regarding specific exposure of these populations to
HFPO-DA in different environmental media. The EPA considered exposures in the general U.S.
population as likely being applicable to these two populations. Second, the EPA identified
several relevant HFPO-DA exposures and pathways (Box 2), including dietary consumption,
incidental oral consumption via dust and soil or dermal exposure via soil and dust, and inhalation
exposure via ambient air. Several of these may be potentially significant exposure sources. Third,
the EPA determined that there was inadequate quantitative data to describe the central tendencies
and high-end estimates for all of the potentially significant sources (Box 3). For example, a study
from China indicates that exposure via dust may be a significant pathway for HFPO-DA. At the
time of the literature search, the EPA was unable to identify studies assessing HFPO-DA
concentrations in dust samples from the U.S. and therefore, the agency does not have adequate
quantitative data to describe the central tendency and high-end estimate of exposure for this
potentially significant source in the U.S. population. However, the agency determined there were
sufficient data, physical/chemical property information, fate and transport information, and/or
generalized information available to characterize the likelihood of exposure to relevant sources
(Box 4). Notably, based on the studies summarized in the sections above, there are significant
known or potential uses/sources of HFPO-DA other than drinking water (Box 6), though there is
not information available on each source to make a characterization of exposure (Box 8A). For
example, the EPA identified physico-chemical properties of HFPO-DA indicating that
volatilization may be an important fate process for this chemical. There is evidence in the
literature of atmospheric transport of HFPO-DA and occurrence in rainwater. However,
monitoring data describing the occurrence of HFPO-DA in ambient air is limited to a single
report from an area located nearby a known point source of this chemical. The levels of HFPO-
DA in ambient air at this source may not be representative of the U.S. as a whole. Therefore, it is
not possible to determine whether ambient air can be considered a major or minor contributor to
total HFPO-DA exposure in the U.S. general population. Similarly, it is not possible to determine
whether the other potentially significant exposure sources such as vegetables and soil should be
considered major or minor contributors to total HFPO-DA exposure. Given these considerations,
following recommendations of the Exposure Decision Tree (USEPA, 2000b), the EPA
recommends an RSC of 20% (0.20) for HFPO-DA.
A-15
-------
FINAL
2024
Appendix B. PFBS: Summary of Occurrence in
Water and Detailed Relative Source Contribution
B.l. Occurrence in Water
PFBS can enter the aquatic environment through releases from manufacturing sites, industrial
uses, fire/crash training areas, and wastewater treatment facilities, as well as from the use of
contaminated biosolids (USEPA, 2021d; ATSDR, 2021). PFBS has a high solubility in water
(52.6 g/L at 22.5-24 °C for the potassium salt) and high mobility in the environment (log Koc
1.2 to 2.7) (ECHA, 2019). A literature review of physicochemical properties and environmental
monitoring data for PFBS by the Norwegian Environment Agency determined that volatilization
from water is negligible, but that the presence of PFBS in ambient air can result from direct
emissions or transport of droplets in contaminated water (Arp and Slinde, 2018). PFBS has been
found in rain as well as in snow/ice in the Arctic and Antarctic (Arp and Slinde, 2018).
The EPA collected information about PFBS occurrence in water (described below). To better
understand PFBS sources and occurrence patterns in water, this section includes studies
conducted within and outside the United States. Overall, studies that analyzed water from sites
receiving inputs from or in proximity to known sources of PFAS (as reported by study authors)
did not provide a consistent pattern of detection; increased PFBS detection frequencies (DFs) or
concentrations were not only observed in studies of sites with known sources of PFAS
contamination from point sources. Specifically, DFs of 0% were reported at some sites with
known, suspected, or historic PFAS contamination, and DFs of 100% were reported at some sites
with no known point sources of PFAS contamination. However, the maximum reported PFBS
concentrations in groundwater and surface water were measured at sites with known PFAS
contamination from AFFF usage (Anderson et al., 2016).
B.t.t. Ground Water
Several studies evaluated the occurrence of PFBS in raw groundwater in the United States or
Europe (see Table B-l). Most of the available studies sampled from groundwaters known or
suspected to be contaminated with PFAS through various sources, as reported by the study
authors. Importantly, some of these groundwaters are known to be used as input sources for
PWSs.
The EPA identified four U.S. studies assessing PFBS concentrations in groundwater at sites
known to be contaminated with PFAS from the use of AFFF (Steele et al., 2018; Eberle et al.,
2017; Anderson et al., 2016; Moody et al., 2003). Of the three studies that reported PFBS
detections, two reported DFs of 78.26% and 100% (Eberle et al., 2017; Anderson et al., 2016);
the third study did not report a PFBS DF across sample sites but indicated a range of PFBS
concentrations (ND-48 ng/L) (Steele et al., 2018). The fourth study, which analyzed
groundwater from the decommissioned Wurtsmith Air Force Base, did not detect PFBS at any of
the ten sites sampled, though other PFAS were detected (Moody et al., 2003). However, a case
study published by the Association of State and Territorial Solid Waste Management Officials
reported quantifiable levels of PFBS in four of seven samples tested from the Wurtsmith Air
B-l
-------
FINAL
2024
Force Base; one site sampled directly below the fire training area was reported to have a PFBS
concentration of 4,100 ng/L (ASTSWMO, 2015).
Additionally, PFBS has been detected at concentrations ranging from 0.00211 ng/L to 0.0261
ng/L in groundwater wells (100% well DF) at a site near the 3M Cottage Grove
perfluorochemical manufacturing facility in Minnesota (ATSDR, 2021; 3M, 2007). Lee et al.
(2015) evaluated urban shallow groundwater contaminated by wastewater effluent discharge and
reported a DF of 20% (1 of 5 shallow sites) and a maximum PFBS level of 36.3 ng/L. In
contrast, Procopio et al. (2017) collected groundwater from 17 sampling sites (53 total across all
water types sampled), some of which were located downstream of an industrial facility that used
materials containing PFOA. PFBS was not detected in groundwater collected from any of the
sampling locations. Post et al. (2013) assessed raw water from PWS intakes in New Jersey; these
intake locations were selected to represent New Jersey geographically and they were not
necessarily associated with any known PFAS release. PFBS was detected pre-treatment in 1 of
18 systems at a concentration of 6 ng/L (minimum reporting level = 5 ng/L). Lindstrom et al.
(2011) analyzed water from 13 wells intended for uses other than drinking water (e.g., livestock,
watering gardens) in areas impacted by up to 12 years of field applications of biosolids
contaminated by a fluoropolymer manufacturer. PFBS was detected in three of the wells (mean
concentration 10.3 ng/L; range: ND-76.6 ng/L).
Of the 10 identified studies conducted in Europe, seven studies evaluated groundwater samples
from sites with known or suspected PFAS releases associated with AFFF use, fluorochemical
manufacturing, or other potential emission sources including landfill/waste disposal sites, skiing
areas, or areas of unspecific industries that use PFAS in manufacturing (e.g., metal plating)
(Dauchy et al., 2019; Fteisseter et al., 2019; Gobelius et al., 2018; Dauchy et al., 2017;
Gyllenhammar et al., 2015; Wagner et al., 2013; Dauchy et al., 2012). All of these studies
reported PFBS detections in at least one sample or site, though only two studies (both conducted
in the vicinity of areas with known AFFF usage) reported PFBS concentrations >100 ng/L
(Dauchy et al., 2019; Gyllenhammar et al., 2015). The remaining three studies of the 10
identified did not provide information on whether there were potential sources of PFAS at the
sampling locations or were designed to be regionally, nationally, or internationally representative
(Barreca et al., 2020; Boiteux et al., 2012; Loos et al., 2010). At these sites, PFBS was detected
infrequently (DFs 4 to 18%) with a maximum concentration of 25 ng/L across the three studies.
Table B-l. Compilation of Studies Describing PFBS Occurrence in Groundwater
Study
Location
Site Details
Results
North America
Lee et al. (2015)
United States
(California)
Samples from 5 urban
shallow groundwater
wells with wastewater
contamination
DFa 20%, range = ND-36.3
ng/L
B-2
-------
FINAL
2024
Study
Location
Site Details
Results
Appleman et al.
(2014)
United States
(New Jersey)
Samples from 5 New
Jersey groundwater source
waters for PWSs impacted
by upstream wastewater
effluent discharge
DFa 100%, meana (range) =
2.4 (0.43-3.7) ng/L
Post et al. (2013)
United States
(New Jersey)
Raw water from 18 public
drinking water system
groundwater intakes
DF 6%, range = ND-6 ng/L
Steele et al.
(2018)
United States
(Alaska)
Military base
contaminated with PFAS
from AFFF use (4 wells
sampled once per month
for 8 months)
DFa NR, range = ND-48
ng/L
Eberle et al.
(2017)
United States
(Joint Base
Langley-Eustis,
VA)
Former fire training site,
site characterization and
pretreatment groundwater
samples
Site characterization: DF
100%, meana (range) =
3,700 (1,100-13,000) ng/L
(10 wells)
Pretreatment: DF 100%,
meana (range) = 3,400
(1,200-5,000) ng/L (5 wells,
2 laboratory samples/well)
Anderson et al.
(2016)
United States
(national)
Ten active U.S. Air Force
installations with historic
AFFF release
DF 78.26%, median of
detects (range) = 200 (ND-
110,000) ng/L
Moody et al.
(2003)
United States
(Oscoda, MI)
Groundwater plume at
former Wurtsmith Air
Force Base; firefighting
training area active from
1952 to 1993
DF 0%
Procopio et al.
(2017)
United States
(New Jersey)
Samples collected from
temporary wells in a small
area of an
industrial/business park
located within the
Metedeconk River
Watershed
DF 0%
B-3
-------
FINAL
2024
Study
Location
Site Details
Results
Lindstrom et al.
(2011)
United States
(Alabama)
Samples from 13 wells
used for purposes aside
from drinking water (e.g.,
livestock, watering
gardens, washing), located
in areas with historical
land application of
fluorochemical industry-
impacted biosolids
DFa 23%, mean (range) =
10.3 (ND-76.6) ng/L
Europe
Barreca et al.
(2020)
Italy (Lombardia
region)
Groundwater sampling
stations representative of
region
DF 18%a, concentrations
NR
Boiteux et al.
(2012)
France (national)
Raw water from 2
sampling campaigns of
DWTPs, some sites
possibly affected by
industrial or commercial
releases
DF 4%, range = ND-9 ng/L
Loos et al.
(2010)
23 European
countries
Monitoring stations were
not necessarily
representative of
surrounding area or
contaminated
DF 15.2%, range = ND-25
ng/L
Gobelius et al.
(2018)
Sweden (national)
Sampling locations
selected based on potential
vicinity of PFAS hot spots
and importance as a
drinking water source area
DF 26%a (triplicate samples
removed), range = ND-22
ng/L
Dauchy et al.
(2012)
France
(unspecified)
Raw water from 2 DWTPs
supplied by alluvial wells;
DWTPs located 15 km
downstream of
fluorochemical
manufacturing facility
DFa 40%, range = ND-4
ng/L
B-4
-------
FINAL
2024
Study
Location
Site Details
Results
Fteisseter et al.
(2019)
Norway
(unspecified)
Samples from 19 sampling
campaigns of 5 pumping
wells placed to intercept a
groundwater
contamination plume
originating from a
firefighting training
facility that ceased usage
of PFAS- and
fluorotelomer-based AFFF
15 years prior
Detections reported but DF
and concentrations not
provided
Dauchy et al.
(2019)
France
(unspecified)
Samples collected over 2
campaigns from 6 areas
(13 monitoring wells) of a
firefighter training site
DFa 77%, range = ND-750
ng/L
Dauchy et al.
(2017)
France
(unspecified)
Samples collected near 3
sites (A, C, D) impacted
by the use of AFFF. Site
A results describe 1
sampling location with 2
sampling events. Site C
results describe a single
sampling location and
event. Site D results
describe 5 sampling
locations, each with a
single sampling event
Site A: DFa 100% meana = 8
ng/L
Site C: point = 6 ng/L
Site D: DFa 20%, range =
ND-59 ng/L
B-5
-------
FINAL
2024
Study
Location
Site Details
Results
Gyllenhammar et
al. (2015)
Sweden (Uppsala)
Samples from local
aquifers extracted by 21
production wells, 6
observation wells or 1
private well located in the
vicinity of a potential
AFFF point source
(military airport). Results
for all well sites were not
provided.
Site 1 (production well): DF
0% (n = NR)
Site 3 (observation wells):
DF 100%, median =100
ng/L (n = 3)
Site 5 (observation well):
DF 0% (n = NR)
Site 6 (production well): DF
0% (n = NR)
Site 7 (observation well):
DF 100%, median = 35 ng/L
(n = 3)
Site 8 (production well):
DFa 91%), median =13 ng/L
(n= 103)
Site 10 (production well):
DFa 2%, median = ND (n =
50)
Wagner et al.
(2013)
Germany
(unspecified)
Samples (n = 3) taken
downstream from a site
contaminated by AFFF
from firefighting activities
DFa 100%), concentrations
NR
Notes: AFFF = aqueous film-forming foam; DF = detection frequency; DWTP = drinking water treatment plant; km =
kilometer; ND = not detected; ng/L = nanogram per liter; PFAA = perfluoroalkyl acid; PFAS = per- and polyfluoroalkyl
substances; NR = not reported; WWTP = wastewater treatment plant.
aThe DF and/or mean was calculated using point data. Means were calculated only when DF = 100%.
B.1.2. Surface Water
Studies evaluating the occurrence of PFBS in surface water in North America or Europe are
summarized in Table B-2. Broadly, studies either targeted surface waters used as drinking water
sources, surface waters known to be contaminated with PFAS (as reported by the study authors),
or surface waters over a relatively large geographic area (i.e., statewide) with some or no known
point sources of PFAS.
Zhang et al. (2016) identified major sources of surface water PFAS contamination by collecting
samples from 37 rivers and estuaries in the northeastern United States (metropolitan New York
area and Rhode Island). PFBS was detected at 82% of sites and the range of PFBS
concentrations was ND to 6.2 ng/L. Appleman et al. (2014) collected samples of surface water
that were impacted by wastewater effluent discharge in several states. PFBS was detected in 64%
of samples from 11 sites with a range of PFBS concentrations from ND to 47 ng/L. Several other
studies from North America (four from the United States and two from Canada) evaluated
B-6
-------
FINAL
2024
surface waters from sites for which authors did not indicate whether sites were associated with
any specific, known PFAS releases (Pan et al., 2018; Yeung et al., 2017; Subedi et al., 2015;
Veillette et al., 2012; Nakayama et al., 2010). Nakayama et al. (2010) also collected samples
across several states, but no specific source of PFAS was identified. The DF in the Nakayama et
al. (2010) study was 43% with median and maximum PFBS levels of 0.71 and 84.1 ng/L,
respectively. Pan et al. (2018) sampled surface water sites in the Delaware River and reported a
100% DF, though PFBS levels were relatively low (0.52 to 4.20 ng/L); Yeung et al. (2017)
reported results for a creek (PFBS concentration of 0.02 ng/L) and a river (no PFBS detected) in
Canada. Veillette et al. (2012) analyzed surface water from an Arctic lake and detected PFBS at
concentrations ranging from 0.011 to 0.024 ng/L. Subedi et al. (2015) evaluated lake water
potentially impacted by septic effluent from adjacent residential properties, and detected PFBS in
only one sample at a concentration of 0.26 ng/L.
Additional available studies assessed surface water samples at U.S. sites contaminated with
PFAS from nearby PFAS manufacturing facilities (ATSDR, 2021; Galloway et al., 2020;
Newsted et al., 2017; Newton et al., 2017) or facilities that manufacture products containing
PFAS (Procopio et al., 2017; Zhang et al., 2016; Lasier et al., 2011). A few of these studies
identified potential point sources of PFAS contamination, including industrial facilities (e.g.,
textile mills, metal plating/coating facilities), airports, landfills, and wastewater treatment plants
(WWTPs) (Galloway et al., 2020; Zhang et al., 2016). Among these sites, DFs (0 to 100%) and
PFBS levels (ND to 336 ng/L) varied. In general, DFs that ranged from 0 to 3% were associated
with samples collected upstream of PFAS point sources, and higher DFs (up to 100%) and PFBS
concentrations were associated with samples collected downstream of point sources. An
additional study (Lindstrom et al., 2011) sampled pond and stream surface water in areas
impacted by up to 12 years of field applications of biosolids contaminated by a fluoropolymer
manufacturer, and the maximum and mean PFBS concentrations were 208 and 26.3 ng/L,
respectively.
Another group of studies from the United States evaluated sites known to be contaminated from
military installations with known or presumed AFFF use (Anderson et al., 2016; Post et al.,
2013; Nakayama et al., 2007). The highest PFBS levels reported among these available studies
were from Anderson et al. (2016) who performed a national study of 40 AFFF-impacted sites
across 10 military installations and reported a maximum PFBS concentration of 317,000 ng/L.
Lescord et al. (2015) examined PFAS levels in Meretta Lake, a Canadian lake contaminated with
runoff from an airport and military base, which are likely sources of PFAS from AFFF use. The
authors reported a 70-fold higher mean PFBS concentration for the contaminated lake versus a
control lake. In addition to AFFF, Nakayama et al. (2007) identified industrial sources, including
metal-plating facilities and textile and paper production, as contributing to the total PFAS
contamination in North Carolina's Cape Fear River Basin. Nakayama et al. (2007) reported a
PFBS DF of 17%) and PFBS concentrations ranging from ND to 9.41 ng/L at these sites.
The EPA identified additional studies evaluating surface water samples from sites in Europe with
known or suspected PFAS releases associated with AFFF use (Mussabek et al., 2019; Gobelius
et al., 2018; Dauchy et al., 2017) or fluorochemical manufacturing (Bach et al., 2017; Boiteux et
al., 2017; Gebbink et al., 2017; Valsecchi et al., 2015). PFBS levels were comparable at the
AFFF-impacted sites (< 300 ng/L overall). Of the four study sites potentially contaminated based
on proximity to fluorochemical manufacturing sites, two (from studies conducted in France) did
not have PFBS detections (Bach et al., 2017; Boiteux et al., 2017). PFBS levels were low at most
B-7
-------
FINAL
2024
sampling locations of the remaining two studies (up to approximately 30 ng/L) except for the site
in River Brenta in Italy (maximum PFBS concentration of 1,666 ng/L) which is also impacted by
nearby textile and tannery manufacturers (Valsecchi et al., 2015).
Eight studies in Europe evaluated areas close to urban areas, commercial activities, or industrial
activities (e.g., textile manufacturing) (Lorenzo et al., 2015; Zhao et al., 2015; Boiteux et al.,
2012; Eschauzier et al., 2012; Rostkowski et al., 2009) and/or wastewater effluent discharges
(Wilkinson et al., 2017; Lorenzo et al., 2015; Labadie and Chevreuil, 2011; Moller et al., 2010).
Among these sites, DFs varied (0 to 100%) and PFBS levels were < 250 ng/L overall.
Ten studies conducted in Europe evaluated sites with no known fluorochemical point source of
contamination (Barreca et al., 2020; Pan et al., 2018; Loos et al., 2017; Shafique et al., 2017;
Munoz et al., 2016; Eriksson et al., 2013; Wagner et al., 2013; Ahrens et al., 2009b; Ahrens et
al., 2009a; Ericson et al., 2008b). Pan et al. (2018) analyzed surface water from sites in the
United Kingdom (Thames River), Germany and the Netherlands (Rhine River), and Sweden
(Malaren Lake). None of the sites sampled were proximate to known sources of PFAS, but PFBS
was detected in all three water bodies. Concentrations of PFBS ranged from 0.46 to 146 ng/L;
the highest level (146 ng/L) was detected in the Rhine River and was more than 20 times greater
than any maximum level found in the other water bodies. In the remaining nine studies, reported
PFBS levels ranged from ND to 26 ng/L, except for one study in Italy that reported a PFBS DF
of 39% and levels in the |ig/L range at three out of 52 locations within the same river basin:
Legnano (16,000 ng/L), Rho (15,000 ng/L), and Pero (3,400 ng/L) (Barreca et al., 2020).
Table B-2. Compilation of Studies Describing PFBS Occurrence in Surface Water
Study
Location
Site Details
PFBS Results
North America
Yeung et al.
(2017)
Canada (Ontario;
Mimico Creek, Rouge
River)
Two water samples at
each of the sites
Mimico Creek: point = 0.020
ng/L
Rouge River: DF 0%
Subedi et al.
(2015)
United States (New
York; Skaneateles Lake)
Lake water along the
shoreline of residences
that use an enhanced
treatment unit for onsite
wastewater treatment
DFa 4% (n=28); single detection
value = 0.26 ng/L
Appleman et al.
(2014)
United States
(Wisconsin, Oklahoma,
Alaska, California,
Alabama, Colorado,
Ohio, Nevada,
Minnesota, New Jersey)
Raw surface waters from
11 sites, some impacted
by upstream wastewater
effluent discharge
DFa 64% (n=25); range = ND-
47 ng/L
(MRL = 0.3)
Veillette et al.
(2012)
Canada (Ellesmere
Island, Nunavut)
A lake near the northwest
coast with no known
sources of PFAS
DFa 100%, mean (range) =
0.016 (0.011-0.024) ng/L
B-8
-------
FINAL
2024
Study
Location
Site Details
PFBS Results
Nakayama et al.
(2010)
United States (Illinois,
Iowa, Minnesota,
Missouri, Wisconsin;
Upper Mississippi River
Basin and Missouri
River Basin)
88 sampling sites from
tributaries and streams
DF 43%, median (range) = 0.71
(ND-84.1) ng/L
Galloway et al.
(2020)
United States (Ohio and
West Virginia; Ohio
River Basin)
Rivers and tributaries 58
km upstream to 130 km
downwind of a
fluoropolymer production
facility, some sample
locations potentially
impacted by local
landfills
DF NR, range3 = ND-28.0 ng/L
Newsted et al.
(2017)
United States
(Minnesota; Upper
Mississippi River Pool
2)
Upstream and
downstream of 3M
Cottage Grove facility
outfall, which is a source
ofPFAS
Upstream: DFa 3%, point = 4.2
ng/L
Downstream: DFa 67%, range =
ND-336.0 ng/L
Procopio et al.
(2017)
United States (New
Jersey; Metedeconk
River Watershed)
Downstream of suspected
illicit discharge to soil
and groundwater from a
manufacturer of industrial
fabrics, composites, and
elastomers that use or
produce products
containing PFAAs
DFa 5%, range = ND-100 ng/L
Newton et al.
(2017)
United States (Decatur,
Alabama; Tennessee
River)
6 sites upstream and 3
sites downstream of
fluorochemical
manufacturing facilities
Upstream: DF 0%
Downstream: DFa 100%, mean3
(range) = 69 (10-160) ng/L
Zhang et al.
(2016)
United States (Rhode
Island, New York
Metropolitan Region)
Rivers and creeks, some
sampling locations
downstream from
industrial activities,
airport, textile mills, and
WWTP. PFAS are used
for water resistant coating
in textiles.
DFa 85%, range = ND-6.181
ng/L
B-9
-------
FINAL
2024
Study
Location
Site Details
PFBS Results
Lescord et al.
(2015)
Canada (Resolute Bay,
Nunavut)
One lake (Meretta)
contaminated with runoff
from an airport, which is
a known source ofPFAS;
one control lake (9 Mile)
Meretta: DF NR, mean = 4.9
ng/L
9 Mile: DF NR, mean = 0.07
ng/L
Lasier et al.
(2011)
United States (Georgia;
Coosa River watershed)
Upstream (sites 1 and 2)
and downstream (sites 3-
8) of a land-application
site where effluents from
carpet manufacturers
(suspected of producing
wastewaters containing
perfluorinated chemicals)
are processed at a WWTP
and the treated WWTP
effluent is sprayed onto
the site. Site 4 was
downstream of a
manufacturing facility for
latex and polyurethane
backing material.
Upstream
Sites 1 and 2: DF 0%
Downstream
Site 3: DF NR, mean = 205
ng/L
Site 4: DF NR, mean = 260
ng/L
Site 5: DF NR, mean =125
ng/L
Site 6: DF NR, mean = 134
ng/L
Site 7: DF NR, mean =122
ng/L
Site 8: DF NR, mean =105
ng/L
Anderson et al.
(2016)
United States (national)
Ten U.S. Air Force
installations with historic
AFFF release
DF 80.00%, median (range) =
106 (ND-317,000) ng/L
Post et al.
(2013)
United States (New
Jersey)
6 rivers and 6 reservoirs
from public drinking
water system intakes,
some sites may include
nearby small industrial
park and civil-military
airport
DF 17%, range = ND-6 ng/L
Nakayama et al.
(2007)
United States (North
Carolina; Cape Fear
River Basin)
80 sampling sites in river
basin; some sites near
industrial areas and Fort
Bragg and Pope Air Force
Base with suspected use
of AFFF at the Air Force
Base
DF 62%, mean (range) = 2.58
(ND-9.41) ng/L
B-10
-------
FINAL
2024
Study
Location
Site Details
PFBS Results
Lindstrom et al.
(2011)
United States (Alabama)
32 surface water samples
(ponds and streams) from
areas with historical land
application of
fluorochemical industry-
impacted biosolids
DFa 63%, range = ND-208
ng/L
Bradley et al.
(2020)
United States (Lake
Michigan)
Untreated Lake Michigan
water from treatment
plant intake (4 sites)
DF 29%, range = ND-0.5 ng/L
Europe
Barreca et al.
(2020)
Italy (Lombardia
Region)
Rivers and streams with
no known fluorochemical
sources
DFa 39%, range = ND-16,000
ng/L
Loos et al.
(2017)
Austria, Bulgaria,
Croatia, Moldova,
Romania, Serbia,
Slovakia (Danube River
and tributaries)
Some sampling locations
downstream of major
cities
DF 94%, mean (range) = 1.6
(ND-3.7) ng/L
Wilkinson et al.
(2017)
England (Greater
London and southern
England; Hogsmill
River, Chertsey Bourne
River, Blackwater River)
50 m upstream and 250 m
and 1,000 m downstream
from WWTP effluent
outfalls
Upstream: DF NR, mean = 20.4
ng/L
Downstream 250 m: DF NR,
mean = 40.3 ng/L
Downstream 1,000 m: DFNR,
mean = 41.1 ng/L
Shafique et al.
(2017)
Germany (Leipzig,
PleiBe-Elster River,
Saale River, and Elbe
River)
Sampling sites were not
proximate to known point
sources of any
fluorochemical facilities
PleiBe-Elster: DF NR, mean =
1.2 ng/L
Saale: DF NR, mean = 7.5 ng/L
Elbe: DF NR, mean = 4.3 ng/L
Munoz et al.
(2016)
France (Seine River)
Two sites downstream of
Greater Paris and one site
unaffected by the Greater
Paris region
DF 70%, range = ND-3.1 ng/L
B-ll
-------
FINAL
2024
Study
Location
Site Details
PFBS Results
Lorenzo et al.
(2015)
Spain (Guadalquivir
River Basin, Ebro River
Basin)
Guadalquivir sampling
locations included
downstream ofWWTPs,
near industrial areas, near
a military camp, or
through major cities;
Ebro sampling locations
included nearby ski
resorts and downstream
of WWTP and industrial
areas
Guadalquivir: DF 8%, mean
(range) = 10.1 (ND-228.3) ng/L
Ebro: DF 0%
Zhao et al.
(2015)
Germany (Elbe River
and lower Weser River)
Some sampling sites near
Hamburg city and
industrial plants
Elbe: DF 100%, mean (range) =
7.4 (0.24-238) ng/L
Weser: DF 100%, mean (range)
= 1.41 (0.75-1.85) ng/L
Eriksson et al.
(2013)
Denmark (Faroe Islands)
Lakes Leitisvatn,
Havnardal, Kornvatn, and
A Myranar with no
known point sources of
any fluorochemical
facilities
Leitisvatn: DF 0%
Havnardal Lake: DF 0%
Kornvatn Lake: DF 0%
A Myranar: DF 0%
Wagner et al.
(2013)
Germany (Rhine River)
Sampling sites were not
proximate to known point
sources of any
fluorochemical facilities
DFa 100%, meanb (rangeb) =18
(9-26) ng/L
Boiteux et al.
(2012)
France (national)
Rivers; some locations
may have upstream
industrial sources
DF 1%, range = ND-5 ng/L
Eschauzier et al.
(2012)
The Netherlands
(Amsterdam; Lek Canal,
tributary of Rhine River)
Downstream of an
industrial point source in
the German part of the
Lower Rhine
DFa 100%, mean (range) = 35
(31-42) ng/L
Labadie and
Chevreuil
(2011)
France (Paris; River
Seine)
Urban stretch of the River
Seine during a flood
cycle, sampling location
under the influence of
two urban WWTPs and
two major combined
sewer overflow outfalls
DF 100%, mean (range) = 1.3
(0.6-2.6) ng/L
B-12
-------
FINAL
2024
Study
Location
Site Details
PFBS Results
Moller et al.
(2010)
Germany (Rhine River
watershed)
Upstream and
downstream of
Leverkusen, where
effluent of a WWTP
treating industrial
wastewater was
discharged; other major
rivers and tributaries
Rhine upstream Leverkusen:
DF 100%, mean (range) = 3.19
(0.59-6.58) ng/L
Rhine downstream Leverkusen:
DF 100%, mean (range) = 45.4
(15.0-118) ng/L
River Ruhr: DF 100%, mean
(range) = 7.08 (2.87-11.4) ng/L
River Moehne: point = 31.1
ng/L
Other tributaries: DF 100%,
mean (range) = 2.84 (0.22-
6.82) ng/L
Ahrens et al.
(2009b)
Germany (Elbe River)
Sampling sites in
Hamburg city (sites 16-
18) and from Laurenburg
to Hamburg (sites 19-24)
Hamburg:
Dissolved: DFa 100%, mean
(range) =1.6 (1.1-2.5) ng/L
Laurenburg to Hamburg:
Dissolved: DFa 100%, mean
(range) = 1.1 (0.53-1.5) ng/L
Ahrens et al.
(2009a)
Germany (Elbe River)
Sampling locations 53 to
122 km (sites 1 to 9)c
upstream of estuary
mouth of Elbe River
DF NR; range of mean (for
different locations) = 1.8-3.4
ng/L
Rostkowski et
al. (2009)
Poland (national)
Rivers, lakes, and streams
in northern and southern
Poland, some southern
locations near chemical
industrial activities
North: DFa 60%, range = ND-
10 ng/L
South: DFa 73%, range = ND-
16.0 ng/L
Ericson et al.
(2008b)
Spain (Tarragona
Province; Ebro River,
Francoli River, Cortiella
River)
Sampling sites were not
proximate to known point
sources of any
fluorochemical facilities
Ebro site 1: DF 0%
Ebro site 2: DF 0%
Francoli: DF 0%
Cortiella: DF 0%
Bach et al.
(2017)
France (southern)
Upstream and
downstream from
discharge point that
receives wastewater from
an industrial site with two
fluoropolymer
manufacturing facilities
Upstream: DF 0%
Downstream: DF 0%
B-13
-------
FINAL
2024
Study
Location
Site Details
PFBS Results
Boiteux et al.
(2017)
France (northern)
River samples from
upstream and downstream
of an industrial WWTP
that processes raw
sewage from
fluorochemical
manufacturing facility
Upstream: DF 0%
Downstream: DF 0%
Gebbink et al.
(2017)
The Netherlands
(Dordrecht)
Upstream and
downstream of Dordrecht
fluorochemical
production plant; two
control sites
Control sites: DFa 100%, mean3
(range) =17 (12-22) ng/L
Upstream: DFa 100%, mean3
(range) = 19.7 (18-21) ng/L
Downstream: DFa 100%, mean3
(range) = 21 (16-27) ng/L
Valsecchi et al.
(2015)
Italy (Po River Basin,
Brenta River Basin,
Adige River Basin,
Tevere River Basin, and
Arno River Basin)
Two river basins (Po and
Brenta) which receive
discharges from
two chemical plants that
produce fluorinated
polymers and
intermediates; three river
basins (Adige, Tevere,
Arno) with no known
point sources of any
fluorochemical facilities
Po: DF3 56%, range =ND-30.4
ng/L
Brenta: DF3 100%, mean3
(range) = 707 (23.1-1,666)
ng/L
Adige: DF3 20%, range = ND-
4.3 ng/L
Tevere: DF 0%
Arno: DF3 58%, range = ND-
31.4 ng/L
Mussabek et al.
(2019)
Sweden (Lulea)
Samples from lake and
pond near a firefighting
training facility at the
Norrbotten Air Force
Wing known to use
PFAS-containing AFFF
Lake: DF NR, mean = 200 ng/L
Pond: DF NR, mean =150 ng/L
Gobelius et al.
(2018)
Sweden (national)
Sampling locations
selected based on
potential vicinity of
PFAS hot spots and
importance as a drinking
water source area, some
sites include firefighting
training sites at airfields
and military areas
DF3 29%, range = ND-299
ng/L
B-14
-------
FINAL
2024
Study
Location
Site Details
PFBS Results
Dauchy et al.
(2017)
France (unspecified)
Samples collected near 3
sites (B, C, D) impacted
by the use of firefighting
foams
Site B: DF 0%
Site C: DF 0%
Site D: DFa 30%, range = ND-
138 ng/L
Multiple Continents
Pan et al. (2018)
United States (Delaware
River)
Sampling sites were not
proximate to known point
sources of any
fluorochemical facilities
DFa 100%, mean (range) = 2.19
(0.52-4.20) ng/L
United Kingdom
(Thames River)
Sampling sites were not
proximate to known point
sources of any
fluorochemical facilities
DFa 100%, mean (range) = 5.06
(3.26-6.75) ng/L
Germany and the
Netherlands (Rhine
River)
Sampling sites were not
proximate to known point
sources of any
fluorochemical facilities
DFa 100%, mean (range) = 21.9
(0.46-146) ng/L
Sweden (Malaren Lake)
Sampling sites were not
proximate to known point
sources of any
fluorochemical facilities
DFa 100%, mean (range) = 1.43
(0.75-1.92) ng/L
Notes: AFFF = aqueous film-forming foam; DF = detection frequency; km = kilometer; m = meter; ND = not detected; ng/L =
nanogram per liter; NR = not reported; PFAA = perfluoroalkyl acid; PFAS = per- and polyfluoroalkyl substances; WWTP =
wastewater treatment plant; (ig/L = microgram per liter.
a The DF and/or mean was not reported in the study and was calculated in this synthesis. Means were calculated only when DF
= 100%.
b For Wagner et al. (2013), PFBS concentrations were calculated using the fluorine concentrations reported in Table 4 from the
study.
c Freshwater locations determined as sites with conductivity <1.5 milliSiemens/cm.
B.2. RSC for PFBS, Literature Search and Screening
Methodology
The EPA applies an RSC to the RfD when calculating an MCLG based on noncancer effects or
for carcinogens that are known to act through a nonlinear mode of action to account for the
fraction of an individual's total exposure allocated to drinking water (USEPA, 2000b). The EPA
emphasizes that the purpose of the RSC is to ensure that the level of a chemical allowed by a
criterion (e.g., the MCLG for drinking water) or multiple criteria, when combined with other
identified sources of exposure (e.g., diet, ambient and indoor air) common to the population of
concern, will not result in exposures that exceed the RfD. In other words, the RSC is the portion
of total daily exposure equal to the RfD that is attributed to drinking water ingestion (directly or
indirectly in beverages like coffee tea or soup, as well as from transfer to dietary items prepared
B-15
-------
FINAL
2024
with drinking water) relative to other exposure sources; the remainder of the exposure equal to
the RfD is allocated to other potential exposure sources. For example, if for a particular
chemical, drinking water were to represent half of total exposure and diet were to represent the
other half, then the drinking water contribution (or RSC) would be 50%. The EPA considers any
potentially significant exposure source when deriving the RSC.
The RSC is derived by applying the Exposure Decision Tree approach published in the EPA's
Methodology for Deriving Ambient Water Quality Criteria for the Protection of Human Health
(USEPA, 2000b). The Exposure Decision Tree approach allows flexibility in the RfD
apportionment among sources of exposure and considers several characteristics of the
contaminant of interest, including the adequacy of available exposure data, levels of the
contaminant in relevant sources or media of exposure, and regulatory agendas (i.e., whether there
are multiple health-based criteria or regulatory standards for the contaminant). The RSC is
developed to reflect the exposure to the U.S. general population or a sensitive population within
the U.S. general population and may be derived qualitatively or quantitatively, depending on the
available data.
A quantitative RSC determination first requires "data for the chemical in question...
representative of each source/medium of exposure and... relevant to the identified population(s)"
(USEPA, 2000b). The term "data" in this context is defined as ambient sampling measurements
in the media of exposure, not internal human biomonitoring metrics. More specifically, the data
must adequately characterize exposure distributions including the central tendency and high-end
exposure levels for each source and 95% confidence intervals for these terms (USEPA, 2000b).
Frequently, an adequate level of detail is not available to support a quantitative RSC derivation.
When adequate quantitative data are not available, the agency relies on the qualitative
alternatives of the Exposure Decision Tree approach. A qualitatively-derived RSC is an estimate
that incorporates data and policy considerations and thus, is sometimes referred to as a "default"
RSC (USEPA, 2000b). Both the quantitative and qualitative approaches recommend a "ceiling"
RSC of 80%) and a "floor" RSC of 20% to account for uncertainties including unknown sources
of exposure, changes to exposure characteristics over time, and data inadequacies (USEPA,
2000b).
In cases in which there is a lack of sufficient data describing environmental monitoring results
and/or exposure intake, the Exposure Decision Tree approach results in a recommended RSC of
20%. In the case of MCLG development, this means that 20% of the exposure equal to the RfD
is allocated to drinking water and the remaining 80% is reserved for other potential sources, such
as diet, air, consumer products, etc. This 20% RSC value can be replaced if sufficient data are
available to develop a scientifically defensible alternative value. If scientific data demonstrating
that sources and routes of exposure other than drinking water are not anticipated for a specific
pollutant, the RSC can be raised as high as 80% based on the available data, allowing the
remaining 20% for other potential sources (USEPA, 2000b). Applying a lower RSC (e.g., 20%)
is a more conservative approach to public health and results in a lower MCLG.
B.2.1. Literature Search and Screening
In 2020, the EPA conducted a literature search to evaluate evidence for pathways of human
exposure to eight PFAS (PFOA, PFOS, PFBA, PFBS, PFDA, perfluorohexanoic acid (PFHxA),
PFHxS, and PFNA) (Holder et al., 2023). This search was not date limited and spanned the
B-16
-------
FINAL
2024
information collected across the Web of Science (WOS), PubMed, and ToxNet/ToxLine (now
ProQuest) databases. The results of the PFBS literature search of publicly available sources are
available through the EPA's Health & Environmental Resource Online website at
https://hero.epa.eov/hero/index.cfm/proiect/paee/proiect id/2610.
The 654 literature search results for PFBS were imported into SWIFT-Review (Sciome, LLC,
Research Triangle Park, NC) and filtered through the Evidence Stream tags to identify human
studies and nonhuman (i.e., those not identified as human) studies. Studies identified as human
studies were further categorized into seven major PFAS pathway categories (Cleaning Products,
Clothing, Environmental Media, Food Packaging, Home Products/Articles/Materials, Personal
Care Products, and Specialty Products) as well as an additional category for Human Exposure
Measures. Nonhuman studies were grouped into the same seven major PFAS pathway
categories, except that the Environmental Media category did not include soil, wastewater, or
landfill. Only studies published between 2003 and 2020 were considered. Application of the
SWIFT-Review tags identified 343 peer-reviewed papers matching these criteria for PFBS.
The 343 articles were screened to identify studies reporting measured occurrence of PFBS in
human matrices and media commonly related to human exposure (human blood/serum/urine,
drinking water, food, food contact materials, consumer products, indoor dust, indoor and ambient
air, and soil). For this synthesis, additional screening was conducted to identify studies relevant
to surface water (freshwater only) and groundwater using a keyword7 search for water terms.
Following the Populations, Exposures, Comparators, and Outcomes (PECO) criteria outlined in
Table B-3, the title and abstract of each study were independently screened for relevance by two
screeners using litstream™. A study was included as relevant if it was unclear from the title and
abstract whether it met the inclusion criteria. When two screeners did not agree whether a study
should be included or excluded, a third reviewer was consulted to make a final decision. The title
and abstract screening resulted in 191 unique studies being tagged as relevant (i.e., having data
on occurrence of PFBS in exposure media of interest) that were further screened with full-text
review using the same inclusion criteria. After additional review of the evidence collected by
Holder et al. (2023), 87 studies originally identified for other PFAS also contained information
relevant to PFBS. Based on full-text review, 147 studies were identified as having relevant,
extractable data for PFBS from the United States, Canada, or Europe for environmental media,
not including studies with only human biomonitoring data. Of these 147 studies, 130 were
identified from Holder et al. (2023), where primary data were extracted into a comprehensive
evidence database. Parameters of interest included sampling dates and locations, numbers of
collection sites and participants, analytical methods, limits of detection and detection
frequencies, and occurrence statistics. Seventeen of the 147 studies were identified in this
synthesis as containing primary data on only surface water and/or groundwater.
The evidence database of Holder et al. (2023) additionally identified 18 studies for which the
main article was not available for review. As part of this synthesis, 17 of the 18 studies could be
retrieved. An additional three peer-reviewed references were identified through gray literature
sources that were included to supplement the search results. The combined 20 studies underwent
7 Keyword list: water, aquifer, direct water, freshwater, fresh water, groundwater, ground water, indirect water, lake,
meltwater, melt water, natural water, overland flow, recreation water, recreational water, river, riverine water,
riverwater, river water, springwater, spring water, stream, surface water, total water, water supply
B-17
-------
FINAL 2024
full-text screening using the inclusion criteria in Table B-3. Based on full-text review, four
studies were identified as relevant.
Table B-3. Populations, Exposures, Comparators, and Outcomes (PECO) Criteria
PECO Element Inclusion Criteria
Population Adults and/or children in the general population and populations in the
vicinity of PFAS point sources from the United States, Canada, or Europe
Exposure Primary data from peer-reviewed studies collected in any of the following
media: ambient air, consumer products, drinking water, dust, food, food
packaging, groundwater51, human blood/serum/urine, indoor air, landfill,
sediment, soil, surface water51 (freshwater), wastewater/biosolids/sludge
Comparator Not applicable
Outcome Measured concentrations of PFBS (or measured emissions from food
packaging and consumer products only)
Notes: PFBS = perfluorobutane sulfonic acid
a Surface water and groundwater were not included as relevant media in Holder et al. (2023). Studies were re-screened for these
two media in this synthesis.
Using the screening results from the evidence database and this synthesis, a total of 151 peer-
reviewed studies were identified as relevant.
B.2.2. Additional Screening
The EPA also searched the following publicly available gray literature sources for information
related to relative exposure of PFBS for all potentially relevant routes of exposure (oral,
inhalation, dermal) and exposure pathways relevant to humans:
• USEPA (202 Id). Human Health Toxicity Values for Perfluorobutane Sulfonic Acid
(CASRN 375-73-5) and Related Compound Potassium Perfluorobutane Sulfonate
(CASRN 29420-49-3);
• ATSDR's Toxicological Profiles;
• CDC's national reports on human exposures to environmental chemicals;
• The EPA's CompTox Chemicals Dashboard;
• The EPA's fish tissue studies;
• The EPA's Toxics Release Inventory;
• Relevant documents submitted under the Toxic Substances Control Act and relevant
reports from the EPA's Office of Chemical Safety and Pollution Prevention;
• U.S. Food and Drug Administration's (FDA's) Total Diet Studies and other similar
publications from FDA, U.S. Department of Agriculture, and Health Canada;
• National Oceanic and Atmospheric Administration's (NOAA's) National Centers for
Coastal Ocean Science data collections;
• National Science Foundation direct and indirect food and/or certified drinking water
additives;
• Throwaway Packaging, Forever Chemicals: European wide survey of PFAS in
disposable food packaging and tableware (Strakova et al., 2021);
B-18
-------
FINAL
2024
• PubChem compound summaries;
• Relevant sources identified in the relative source contribution discussions (Section 5) of
the EPA's Proposed Approaches to the Derivation of a Draft Maximum Contaminant
Level Goal for Perfluorooctanoic Acid (PFOA)/Perfluorooctane Sulfonic Acid (PFOS) in
Drinking Water; and
• Additional sources, as needed.
The EPA has included available information from these gray literature sources for PFBS relevant
to its uses, chemical and physical properties, and for occurrence in ambient or indoor air, foods
(including fish and shellfish), soil, dust, and consumer products. The EPA has also included
available information specific to PFBS below on any regulations that may restrict PFBS levels in
media (e.g., water quality standards, air quality standards, food tolerance levels).
B.3. Summary of Potential Exposure Sources of PFBS Other
than Water
B.3.1. Food
PFBS was included in a suite of individual PFAS selected as part of PF AS-targeted
reexaminations of samples collected for the U.S. Food and Drug Administration's (FDA's) Total
Diet Study (US FDA, 2022b, 2022a, 2021b, 2021a, 2020b, 2020a); however, it was not detected
in any of the food samples tested. It should be noted that FDA indicated that the sample sizes
were limited and that the results should not be used to draw definitive conclusions about PFAS
levels or presence in the general food supply (US FDA, 2022c). PFBS was detected in cow milk
samples collected from a farm with groundwater known to be contaminated with PFAS, as well
as in produce (collard greens) collected from an area near a PFAS production plant, in FDA
studies of the potential exposure of the U.S. population to PFAS (US FDA, 2021c, 2018).
Maximum residue levels for PFBS were not found in the Global MRL Database (Bryant Christie
Inc, 2022).
In addition to efforts by FDA, peer-reviewed studies conducted in North America, Europe, and
across multiple continents analyzed PFBS in food items obtained from home, recreational, or
commercial sources (see Table B-4). Food types evaluated include fruits and vegetables, grains,
meat, seafood, dairy, and fats/other (e.g., eggs, spices, and oils), with seafood showing the
highest levels of PFBS detected. PFBS was not detected in any of the eight studies that analyzed
human milk for PFAS (not shown in Table B-4)—one in the United States (von Ehrenstein et al.,
2009) and seven in Europe (Abdallah et al., 2020; Nyberg et al., 2018; Cariou et al., 2015;
Lankova et al., 2013; Beser et al., 2011; Karrman et al., 2010; Karrman et al., 2007).
Of the studies conducted in North America, four U.S. studies (Scher et al., 2018; Byrne et al.,
2017; Blaine et al., 2014; Schecter et al., 2010) found PFBS in at least one food item. Locations
and food sources varied in these studies. In Schecter et al. (2010), PFBS was detected in cod
samples but not in any of the other foods collected from Texas grocery stores. Scher et al. (2018)
detected PFBS in plant parts (leaf and stem samples) analyzed from garden produce collected at
homes in Minnesota within a GCA impacted by a former 3M PFAS production facility (PFBS
concentrations ranged from ND to 0.065 nanograms per gram [ng/g]). The authors suggested that
B-19
-------
FINAL
2024
the PFBS detections in plant parts were likely associated with PFAS present in irrigation water
that had accumulated in produce. Blaine et al. (2014) found PFBS in radish, celery, tomato, and
peas that were grown in soil amended with industrially impacted biosolids. They also found
PFBS in these crops grown in soil that had received municipal biosolid applications over 20
years. In unamended control soil samples, PFBS was only detected in radish root with an average
value of 22.36 ng/g (Blaine et al., 2014). In a similar study conducted by Blaine et al. (2013),
PFBS was found in lettuce, tomato, and corn grown in industrially impacted biosolids-amended
soils in greenhouses. Young et al. (2012) analyzed 61 raw and retail milk samples from 17 states
for PFAS, but PFBS was not detected.
Based on the available data collected to date, seafood (including fish and shellfish) has been
found to contain the highest concentrations of PFBS out of all food types examined. Burkhard
(2021) identified 16 studies reporting BAFs for PFBS and calculated a median (standard
deviation) bioaccumulation factor (BAF) in muscle tissue/fillet of 22.39 ± 6.92 L/kg wet weight
(reported as a logBAF of 1.35 ± 0.84 L/kg). Several large-scale sampling efforts have been
conducted by the EPA and other agencies to determine PFAS levels in fish. In the EPA's 2013-
2014 National Rivers and Streams Assessment (NRSA), PFBS was detected at concentrations
between the quantitation limit (1 ng/g) and the method detection limit (0.1 ng/g) at 0.571 ng/g in
a largemouth bass fish fillet sample collected from Big Black River, Mississippi; 0.475 ng/g in a
smallmouth bass fillet composite collected from Connecticut River, New Hampshire; and 0.148
ng/g in a walleye fillet composite collected from Chenango River, New York (USEPA, 2020a).
Notably, PFBS was not detected in any fish species sampled in the 2008-2009 NRSA (Stahl et
al., 2014). PFBS was also detected at a concentration of 0.36 ng/g in a smallmouth bass fillet
composite collected from Lake Erie, New York in the EPA's 2015 Great Lakes Human Health
Fish Fillet Tissue Study (USEPA, 2021g). PFBS has been detected in Irish pompano, silver
porgy, grey snapper, and eastern oyster from the St. Lucie Estuary in the National Oceanic and
Atmospheric Administration's (NOAA's) National Centers for Coastal Ocean Science, National
Status and Trends Data (NOAA, 2022). PFBS was not a target chemical in the EPA's National
Lake Fish Tissue Study (USEPA, 2009a).
Several peer-reviewed publications that examined PFBS concentrations in fish and shellfish are
also available. As mentioned previously, Schecter et al. (2010) detected PFBS in cod samples.
Mean PFBS levels in cod from this study (0.12 ng/g wet weight [ww]) were much lower than
maximum levels detected in Alaska blackfish obtained from the Suqi River, Alaska in remote
locations upstream and downstream of a former (unnamed) defense site (59.2 ng/g ww) (Byrne
et al., 2017). In this study, blackfish were considered sentinel species but are not among the
traditional fish consumed in the area. The authors noted that the presence of PFAS in fish from
remote sites is suggestive of atmospheric deposition. In two additional studies from North
America, PFBS was not detected in samples of farmed and wild-caught seafood (Chiesa et al.,
2019; Young et al., 2013).
The European Food Safety Authority (EFSA) reported the presence of PFBS in various food and
drink items, including fruits, vegetables, cheese, and bottled water (EFSA, 2012). For average
adult consumers, the estimated exposure ranges for PFBS were 0.03-1.89 nanograms per
kilogram body weight per day (ng/kg bw-day) (minimum) to 0.10-3.72 ng/kg bw-day
(maximum) (EFSA, 2012). Of the studies conducted in Europe, 12 found PFBS in at least one
food type (Table B-4). Eight of the 12 studies included food samples obtained solely from
B-20
-------
FINAL
2024
markets (Scordo et al., 2020; Sznajder-Katarzynska et al., 2019; Surma et al., 2017; D'Hollander
et al., 2015; Perez et al., 2014; Eschauzier et al., 2013; Hlouskova et al., 2013; Domingo et al.,
2012). Across studies, PFBS detections were found in seafood; other animal products such as
meat, dairy, and eggs; fruits and vegetables; tap water-based beverages such as coffee; sweets;
and spices.
Papadopoulou et al. (2017) analyzed duplicate diet samples with PFBS detected in only one solid
food sample (ND-0.001 ng/g; DF 2%; food category unspecified). Eriksson et al. (2013)
evaluated foods that were farmed or freshly caught in the Faroe Islands, and only detected PFBS
in cow milk (0.019 ng/g ww) and packaged dairy milk (0.017 ng/g ww) samples among the
products analyzed. In eight of the European studies where PFBS was not detected, foods were
primarily obtained from commercial sources, but wild-caught seafood was also included.
Two of the 12 European studies examined both market-bought and fresh-caught fish, and PFBS
was detected in seafood from both sources (Vassiliadou et al., 2015; Yamada et al., 2014).
Yamada et al. (2014) found higher PFBS in fresh-caught river fish samples (0.16 ng/g ww
maximum) versus fresh or frozen market samples (0.03 ng/g ww maximum) in France.
Vassiliadou et al. (2015) detected PFBS in raw shrimp (from Greek markets) but did not detect
PFBS in either fried shrimp, raw hake (from Greek fishing sites), or fried hake.
In summary, in Europe and North America, PFBS has been detected in multiple food types,
including fruits, vegetables, meats, seafoods, and other fats. Several large-scale fish tissue
sampling efforts conducted by the EPA and others indicate that fish consumption may be an
important PFBS exposure source. Future large-scale sampling efforts by FDA and others may
help to similarly elucidate PFBS concentrations in other food types. Although several U.S.
studies have evaluated PFBS in meats, fats/oils, fruits, vegetables, and other non-seafood food
types, many of these sampling efforts were localized to specific cities or markets and/or used
relatively small sample sizes. Broader-scale sampling efforts will be helpful in determining the
general levels of PFBS contamination in these food types, as well as the impact of known PFAS
contamination sources on PFBS concentrations in foods.
B-21
-------
FINAL
2024
Table B-4. Com
lilation of Studies Describing PFBS Occurrence in Food
Study
Location and Source
Food Types
Results
North America
Schecter et al.
(2010)
United States (Texas)
Grocery stores
Dairy, fruits and
vegetables,
grain, meat,
seafood,
fats/other
Cod: DF NR, mean = 0.12 ng/g ww
ND in salmon, canned sardines,
canned tuna, fresh catfish fillet,
frozen fish sticks, tilapia, cheeses
(American, mozzarella, Colby,
cheddar, Swiss, provolone, and
Monterey jack), butter, cream
cheese, frozen yogurt, ice cream,
whole milk, whole milk yogurt,
potatoes, apples, cereals, bacon,
canned chili, ham, hamburger, roast
beef, sausages, sliced chicken breast,
sliced turkey, canola oil, margarine,
olive oil, peanut butter, eggs
Byrne et al.
(2017)
United States (Alaska)
Upstream/downstream of
former defense site (Suqi
River)
Seafood
Blackfish: DF 48%, range = ND-
59.2 ng/g ww
Highest concentration was upstream
Seller et al.
(2018)
United States (Minnesota)
Home gardens
Near former 3M PFAS
production facility, homes
within and outside a GCA
Fruits and
vegetables
Within GCA:
Leaf: DF 6%, max = 0.061 ng/g
Stem: DF 4%, max = 0.065 ng/g
ND in floret, fruit, root, seed
Outside GCA: ND
Blaine et al.
(2014)
United States (Midwestern)
Greenhouse study,
unamended controls
Fruits and
vegetables
Radish root: DF NR, mean = 22.36
ng/g
ND in celery shoot, pea fruit
Blaine et al.
(2013)
United States (Midwestern)
Greenhouse and field
studies, unamended controls
Fruits and
vegetables, grain
ND in corn, lettuce, tomato in
unamended soil.
Young et al.
(2013)
United States (Maryland,
Mississippi, Tennessee,
Florida, New York, Texas,
Washington, D.C.)
Retail markets
Seafood
ND in crab, shrimp, striped bass,
farm raised catfish, farm raised
salmon
Young et al.
(2012)
United States (17 states)
Retail markets
Dairy
ND in retail cow's milk
B-22
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Europe
Domingo et al.
(2012)
Spain (Catalonia)
Local markets, small stores,
supermarkets, big grocery
stores
12 food
categories
Vegetables: DF NR, mean = 0.013
ng/g fw
Fish and seafood: DF NR, mean =
0.054 ng/g fw
ND in meat and meat products,
tubers, fruits, eggs, milk, dairy
products, cereals, pulses, industrial
bakery, oils
Perez et al.
(2014)
Serbia (Belgrade and Novi
Sad), Spain (Barcelona,
Girona, and Madrid)
Various supermarkets and
retail stores
8 food
categories
Categories included cereals, pulses
and starchy roots, tree-nuts, oil crops
and vegetable oils, vegetables and
fruits, meat and meat products, milk,
animal fats, dairy products, and eggs,
fish and seafood, and others such as
candies or coffee
Spain: DF 3.2%, range = ND-13
ng/g (primarily fish, oils)
Serbia: DF 5.2%, range = ND-
0.460 ng/g (primarily meat and
meat products, cereals)
D'Hollander et al.
(2015)
Belgium, Czech Republic,
Italy, Norway
PERFOOD study; items
from 3 national retail stores
of different brands and
countries of origin
Fruit, cereals,
sweets, salt
Sweets: DFa 25%, range = ND-
0.0016 ng/g
Fruit: DFa 19%, range = ND-0.067
ng/g
ND in cereals, salt
Hlouskova et al.
(2013)
Belgium, Czech Republic,
Italy, Norway
Several national
supermarkets
Pooled
milk/dairy
products, meat,
fish, hen eggs
DF 5%, mean (range) = 0.00975
(0.006-0.012) ng/g
Eriksson et al.
(2013)
Denmark
Farm, dairy farm, fish from
Faroe Shelf area
Dairy, fruits and
vegetables,
seafood
Milk:
Farmer (Havnardal): point =
0.019 ng/g ww
Diary (Faroe Island): point =
0.017 ng/g ww; ND or NQ in 4
samples
ND in yogurt, creme fraiche,
potatoes, farmed salmon, wild-
caught cod, wild-caught saithe
B-23
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Sznajder-
Katarzynska et al.
(2019)
Poland
Markets
Dairy
All dairy: sum PFBS = 0.04 ng/g
Butter: range = 0.01-0.02 ng/g
ND in camembert-type cheese,
cottage cheese, milk, natural yogurt,
sour cream, kefir (bonny clabber)
Yamada et al.
(2014)
France
Freshwater fish from 6 major
French rivers; fresh and
frozen fish from markets
Seafood
Freshwater fish: DF NR, range =
0.06-0.16 ng/g ww
Fresh or frozen fish: DF NR, range
= 0.02-0.03 ng/g ww
Vassiliadou et al.
(2015)
Greece
Local fish markets,
mariculture farm, fishing
sites
Seafood
Hake: raw mean = 0.45 ng/g ww,
fried mean = 0.83 ng/g ww
Shrimp: raw mean = 1.37 ng/g ww
ND in raw, fried, and grilled
anchovy, bogue, picarel, sand smelt,
sardine, squid, striped mullet, raw
and fried mussel, fried shrimp, and
grilled hake
Eschauzier et al.
(2013)
The Netherlands
(Amsterdam)
Cafes, universities,
supermarkets
Fats/other
Brewed coffee (manual): mean
(range) = 1.6 (1.3-2.0) ng/L
Brewed coffee (machine): mean
(range) = 2.9 (ND-9.8) ng/L
Cola: mean (range) = 7.9 (ND-12)
ng/L
Surma et al.
(2017)
Spain, Slovakia
Source NR
Fats/other
Spices: ND-1.01 ng/g
Spain:
Detected in anise, star anise, fennel,
coriander, cinnamon, peppermint,
parsley, thyme, laurel, cumin, and
oregano
ND in white pepper, cardamon,
clove, nutmeg, allspice, vanilla,
ginger, garlic, black paper, and hot
pepper (mild and hot)
Slovakia: ND in anise, star anise,
white pepper, fennel, cardamom,
clove, coriander, nutmeg, allspice,
cinnamon, vanilla, and ginger
B-24
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Papadopoulou et
al. (2017)
Norway
A-TEAM project: food and
drinks collected by
participants as duplicate diet
samples
Solid foods (11
food categories),
liquid foods (5
drinks)
Solid foods (unspecific food
category): DF 2%, range = ND-
0.001 ng/g
ND in liquid foods (coffee, tea and
cocoa, milk, water, alcoholic
beverages and soft drinks)
Scordo et al.
(2020)
Italy
Supermarkets
Fruits
Olives: DFa 100%, mean3 (range) =
0.294 (0.185-0.403) ng/g dw
ND in strawberries
Ericson et al.
(2008a)
Spain
Local markets, large
supermarkets, grocery stores
18 food
categories
ND in all categories: veal, pork,
chicken, lamb, white fish, seafood,
tinned fish, blue fish, whole milk,
semi-skimmed milk, dairy products,
vegetables, pulses, cereals, fruits, oil,
margarine, and eggs
Noorlander et al.
(2011)
The Netherlands
Several Dutch retail store
chains with nationwide
coverage
15 food
categories
ND in all categories: flour, fatty fish,
lean fish, pork, eggs, crustaceans,
bakery products, vegetables/fruit,
cheese, beef, chicken/poultry, butter,
milk, vegetable oil, and industrial oil
Jogsten et al.
(2009)
Spain (Catalonia)
Local markets, large
supermarkets, grocery stores
Fruits and
vegetables,
meat, seafood,
fats/other
ND in lettuce, raw, cooked, and fried
meat (veal, pork, and chicken), fried
chicken nuggets, black pudding,
lamb liver, pate of pork liver, foie
gras of duck, "Frankfurt" sausages,
home-made marinated salmon, and
common salt
Sznajder-
Katarzynska et al.
(2018)
Poland
Markets
Fruits and
vegetables
ND in apples, bananas, cherries,
lemons, oranges, strawberries,
beetroots, carrots, tomatoes,
potatoes, and white cabbage
Falandysz et al.
(2006)
Poland
Gulf of Gdansk, Baltic Sea
south coast
Meat, seafood
ND in eider duck, cod
Barbosa et al.
(2018)
Belgium, France, the
Netherlands, Portugal
Various markets
Seafood
ND in raw and steamed fish (P.
platessa, M. australis, M. capenis, K.
pelamis, and M. edulis)
B-25
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Holzer et al.
(2011)
Germany
Fish from Lake Mohne and
river Mohne, contaminated
with PFCs from use of
polluted soil conditioner on
agricultural lands; retail
trade, wholesale trade,
supermarkets, and producers
Seafood
Lake Mohne /River Mohne: ND in
cisco, eel, perch, pike, and roach
Trade/markets: ND in eel,
pike/perch, and trout
Jorundsdottir et
al. (2014)
Iceland
Collected during biannual
scientific surveys,
commercially produced
Seafood
ND in anglerfish, Atlantic cod, blue
whiting, lemon sole, ling, lumpfish,
plaice, and pollock
Riviere et al.
(2019)
France
Based on results of national
consumption survey
Seafood,
fats/other
ND in infant food, vegetables, non-
alcoholic beverages, dairy-based
desserts, milk, mixed dishes, fish,
ultra-fresh dairy products, meat,
poultry and game
Lankova etal.
(2013)
Czech Republic
Retail market
Fats/other
ND in infant formula
Zafeiraki et al.
(2016a)
Greece, the Netherlands
Home and commercially
produced
Fats/other
ND in chicken eggs
Gebbink et al.
(2015)
Sweden
Major grocery chain stores,
market basket samples
12 food
categories
ND in all categories: dairy products,
meat products, fats, pastries, fish
products, egg, cereal products,
vegetables, fruit, potatoes, sugar and
sweets, soft drinks
Herzke et al.
(2013)
Belgium, Czech Republic,
Italy, Norway
PERFOOD study: items
from 3 national retail stores
of different brands per
location
Vegetables
ND for all vegetables
Zafeiraki et al.
(2016b)
The Netherlands
Local markets and
slaughterhouses
Meat
ND for horse, sheep, cow, pig, and
chicken liver
B-26
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Multiple Continents
Chiesa etal.
(2019)
United States (Pacific
Ocean)
Wholesale fish market
Seafood
ND in wild-caught salmon
Canada
Wholesale fish market
Seafood
ND in wild-caught salmon
Norway
Wholesale fish market
Seafood
ND in farm salmon
Scotland
Wholesale fish market
Seafood
ND in wild-caught and farm salmon
Notes: DF = detection frequency; dw = dry weight; fw = fresh weight; GCA = groundwater contamination area; ND = not
detected; ng/g = nanogram per gram; ng/L = nanogram per liter; NR = not reported; PFAS = per- and polyfluoroalkyl
substances; NQ = not quantified; (ig/L = microgram per liter; ww = wet weight.
Bold indicates detected levels of PFBS in food.
a The DF and/or mean was not reported in the study and was calculated in this synthesis. Means were calculated only when DF =
100%.
B.3.2. Food Contact Materia Is
PFBS is not authorized for use in food packaging in the United States; however, PFBS has been
detected in food packaging materials in the few available studies that investigate this potential
route of exposure (USEPA, 202 Id; AT SDR, 2021). In one report from the United States, PFBS
was detected in fast-food packaging (7/20 samples) although the concentrations detected were
not reported (Schaider et al., 2017). Additionally, in an analysis performed at the Department of
Food Analysis and Nutrition of the University of Chemistry and Technology in Prague, Czech
Republic, PFBS was not detected in 42 samples of disposable food packaging and tableware
purchased from six different European countries between May and December 2020 (LOQ =1.7
mg/kg) (Strakova et al., 2021).
The EPA identified five peer-reviewed studies in Europe (conducted in Poland, Norway, Greece,
Czech Republic, and Germany) analyzed the occurrence of PFBS in food packaging or food
contact materials (FCMs), such as baking papers and fast-food boxes and wrappers. Surma et al.
(2015) measured levels of 10 perfluorinated compounds in three different brands of common
FCMs commercially available in Poland, including wrapping papers (n = 3), breakfast bags (n =
3), baking papers (n = 3), and roasting bags (n = 3). PFBS was detected in one brand of baking
paper at 0.02 picograms per square centimeter (pg/cm2), but PFBS was not detected at or below
the LOQ in all other FCMs. Vestergren et al. (2015) analyzed paper plates (n = 2), paper cups (n
= 1), baking covers (n = 1), and baking molds (n = 1) purchased from retail stores in Troms0 and
Trondheim, Norway. PFBS was detected in one paper plate at 6.9 pg/cm2.
The remaining three studies did not detect PFBS in FCMs. Zafeiraki et al. (2014) analyzed
FCMs made of paper, paperboard, or aluminum foil collected from a Greek market. PFBS was
B-27
-------
FINAL
2024
not detected in any of the samples of beverage cups (n = 8), ice cream cups (n = 1), fast-food
paper boxes (n = 8), fast-food wrappers (n = 6), paper materials for baking (n = 2), microwave
bags (n = 3), and aluminum foil bags/wrappers (n = 14). The study concluded that the use of
perfluorinated compound alternatives such as fluorophosphates and fluorinated polyethers in the
local manufacturing process potentially explains the low levels of other PFAS (i.e.,
perfluorobutanoic acid [PFBA], perfluorohexanoic acid [PFHxA], perfluoroheptanoic acid
[PFHpA], perfluorononanoic acid [PFNA], perfluorodecanoic acid [PFDA], and
perfluorododecanoic acid [PFDoDA]) detected in the sampled FCMs. Vavrous et al. (2016)
analyzed 15 samples of paper FCMs acquired from a market in the Czech Republic. FCMs
included paper packages of wheat flour (n = 2), paper bags for bakery products (n = 2), sheets of
paper for food packaging in food stores (n = 2), cardboard boxes for packaging of various
foodstuffs (n = 3), coated bakery release papers for oven baking at temperatures up to 220°C (n =
3), and paper filters for coffee preparation (n = 3). PFBS was not detected in any samples.
Kotthoff et al. (2015) analyzed 82 samples for perfluoroalkane sulfonate (PFSA) and
perfluoroalkyl carboxylic acid (PFCA) compounds in 10 consumer products including individual
paper-based FCMs (n = 33) from local retailers in Germany in 2010. PFBS was not detected in
paper-based FCMs.
Overall, the single available studies conducted in the United States indicate that PFBS may be
present in food packaging materials; however, further research is needed to understand which
packaging materials generally contain PFBS at the highest concentrations and with the greatest
frequency. There are also uncertainties related to data gaps on topics that may influence whether
food packaging is a significant PFBS exposure source in humans, including differences in
transfer efficiency from different packaging types directly to humans or indirectly through
foodstuffs.
B.3.3. Consumer Products
Several studies examined a range of consumer products and found multiple PFAS, including
PFBS, at various levels (van der Veen et al., 2020; Zheng et al., 2020; Schultes et al., 2018;
Becanova et al., 2016; Favreau et al., 2016; Gremmel et al., 2016; Kotthoff et al., 2015;
Vestergren et al., 2015; Liu et al., 2014). Two of the studies collected consumer products in the
United States, five purchased consumer products in Europe, and two studies did not report the
purchase location(s) of the consumer products that were tested.
Zheng et al. (2020) determined the occurrence of ionic and neutral PFAS in items collected from
childcare environments in the United States. Nap mats (n = 26; 20 polyurethane foam, 6 vinyl
cover samples) were collected from seven Seattle childcare centers. PFBS was detected in 5% of
nap mat samples at a maximum concentration of 0.04 ng/g. Liu et al. (2014) analyzed the
occurrence of PFAS in commonly used consumer products (carpet, commercial carpet-care
liquids, household carpet/fabric-care liquids, treated apparel, treated home textiles, treated non-
woven medical garments, floor waxes, membranes for apparel, and thread-seal ant tapes)
purchased from retail outlets in the United States. PFBS was detected in 100% of commercial
carpet/fabric-care liquids samples (n = 2) at concentrations of 45.8 and 89.6 ng/g, in 75% of
household carpet/fabric-care liquids and foams samples (n = 4) at concentrations up to 911 ng/g,
in one treated apparel samples (n = 2) at a concentration of 2 ng/g, in the single treated floor wax
and stone/wood sealant sample (143 ng/g, n = 2), and in the single apparel membrane sample
B-28
-------
FINAL
2024
(30.7 ng/g, n = 2). PFBS was not detected in treated home textile and upholstery (n = 2) or
thread-sealant tapes and pastes (n = 2).
van der Veen et al. (2020) examined the effects of weathering on PFAS content in durable water-
repellent clothing collected from six suppliers in Sweden (1 pair of outdoor trousers, 7 jackets, 4
fabrics for outdoor clothes, 1 pair of outdoor overalls). Two pieces of each of the 13 fabrics were
cut. One piece of each fabric was exposed to elevated ultraviolet radiation, humidity, and
temperature in an aging device for 300 hours (assumed lifespan of outdoor clothing); the other
was not aged. Both pieces of each fabric were analyzed for ionic PFAS (including PFBS) and
volatile PFAS. In general, aging of outdoor clothing resulted in increased perfluoroalkylated acid
(PFAA) levels of 5-fold or more. For 8 of 13 fabrics, PFBS was not detected before or after
aging. For three fabrics, PFBS was detected before and after aging, increasing approximately 3-
to 14-fold in the aged fabric (i.e., from 43 to 140 micrograms per square meter [|ig/m2], 45 to
350 |ig/m2, and 9.6 to 130 |ig/m2 respectively for the 3 fabrics). For the remaining two fabrics,
PFBS was not detected prior to aging but was detected afterward at concentrations of 0.57 and
1.7 |ig/m2, respectively. The authors noted that possible explanations for this could be
weathering of precursor compounds (e.g., fluorotelomer alcohols) to PFAAs such as PFBS or
increased extractability due to weathering.
Kotthoff et al. (2015) analyzed 82 samples for PFSA and PFCA compounds in outdoor textiles
(n = 3), gloves (n = 3), carpets (n = 6), cleaning agents (n = 6), impregnating sprays (n = 3),
leather (n = 13), wood glue (n = 1), ski wax (n = 13), and awning cloth (n = 1). Individual
samples were bought from local retailers or collected by coworkers of the involved institutes or
local clubs in Germany. The age of the samples ranged from a few years to decades. PFBS was
detected in outdoor textiles (level not provided), carpet samples (up to 26.8 (j,g/m2), ski wax
samples (up to 3.1 micrograms per kilogram [[j,g/kg]), leather samples (up to 120 (J,g/kg), and
gloves (up to 2 (J,g/kg). Favreau et al. (2016) analyzed the occurrence of 41 PFAS in a wide
variety of liquid products (n = 132 consumer products, 194 total products), including
impregnating agents, lubricants, cleansers, polishes, AFFFs, and other industrial products
purchased from stores and supermarkets in Switzerland. PFBS was not detected in impregnation
products (n = 60), cleansers (n = 24), or polishes (n = 18). PFBS was detected in 13% of a
miscellaneous category of products (n = 23) that included foam-suppressing agents for the
chromium industry, paints, ski wax, inks, and tanning substances, with mean and maximum
concentrations of 998 and 2,992 parts per million (ppm), respectively (median = ND).
The remaining two European studies from Norway (Vestergren et al., 2015) and Sweden
(Schultes et al., 2018) did not detect PFBS in the consumer products analyzed. Vestergren et al.
(2015) analyzed furniture textile, carpet, and clothing samples (n = 40) purchased from retail
stores in Troms0 and Trondheim, Norway, while Schultes et al. (2018) determined levels of 39
PFAS in 31 cosmetic products collected in Sweden. Both studies found measurable
concentrations of at least one PFAS; however, PFBS was not detected in any of the samples.
Of the two studies for which purchase location(s) were not specified, Gremmel et al. (2016)
determined levels of 23 PFAS in 16 new outdoor jackets since it has been shown that outdoor
jackets emit PFAS to the air as well as into water during washing. The jackets were selected
based on factors such as fabric and origin of production (primarily Asia, with some origins not
specified). PFBS (concentration of 0.51 |ig/m2) was only detected in one large hardshell jacket
made of 100% polyester that was polyurethane-coated and finished with Teflon® (production
B-29
-------
FINAL
2024
origin unknown). Becanova et al. (2016) analyzed 126 samples of (1) household equipment
(textiles, floor coverings, electrical and electronic equipment (EEE), and plastics); (2) building
materials (oriented strand board, other composite wood and wood, insulation materials, mounting
and sealant foam, facade materials, polystyrene, air conditioner components); (3) car interior
materials; and (4) wastes of electrical and electronic equipment (WEEE) for 15 target PFAS,
including PFBS. The condition (new versus used) and production year of the samples varied; the
production year ranged from 1981 to 2010. The origin(s) of production were not specified. PFBS
was detected in 31/55, 9/54, 7/10, and 6/7 household equipment, building materials, car interior,
and WEEE samples, respectively. The highest level was 11.4 |ig/kg found in a used 1999 screen
associated with WEEE.
In summary, in the few studies available from North America and Europe, PFBS was detected in
a wide range of consumer products including clothing, household textiles and products,
children's products, and commercial/industrial products. However, there is some uncertainty in
these results as the number and types of products tested in each study were often limited in terms
of sample size. While there is evidence indicating PFBS exposure may occur through the use of
or contact with consumer products, more research is needed to understand the DF and
concentrations of PFBS that occur in specific products, as well as how the concentrations of
PFBS change in these products with age or weathering.
B.3.4. indoor Dust
Dust ingestion may be an important exposure source of PFAS including PFBS (ATSDR, 2021),
though it should be noted that dust exposure may also occur via inhalation and dermal routes.
The EPA identified several studies conducted in the United States, Canada, various countries in
Europe, and across multiple continents that analyzed PFBS in dust of indoor environments
(primarily in homes, but also schools, childcare facilities, offices, and vehicles; see Table B-5).
Most of the studies sampled dust from areas not associated with any known PFAS activity or
release. PFBS concentrations in dust measured in these studies ranged from ND to 170 ng/g with
three exceptions: two studies (Kato et al., 2009; Strynar and Lindstrom, 2008) reported
maximum PFBS concentrations > 1,000 ng/g in dust from homes and daycare centers, and a third
study (Huber et al., 2011) reported a PFBS concentration of 1,089 ng/g in dust from a storage
room that had been used to store "highly contaminated PFC [polyfluorinated compounds]
samples and technical mixtures for several years."
Of the two available studies that measured PFBS in dust from vehicles, one (in the United States)
detected no PFBS (Fraser et al., 2013) and the other (in Ireland) reported a DF of 75% and PFBS
concentrations ranging from ND to 170 ng/g (Harrad et al., 2019).
One U.S. study, Scher et al. (2019) evaluated indoor dust from 19 homes in Minnesota within a
GCA impacted by the former 3M PFAS production facility. House dust samples were collected
from both interior living rooms and entry ways to the yard. The DFs for PFBS were 16% and
11% for living rooms and entry ways, respectively, and a maximum PFBS concentration of 58
ng/g was reported for both locations.
Haug et al. (2011) indicated that house dust concentrations are likely influenced by a number of
factors related to the building (e.g., size, age, floor space, flooring type, ventilation); the
residents or occupants (e.g., number of people, housekeeping practices, consumer habits such as
B-30
-------
FINAL
2024
buying new or used products); and the presence and use of certain products (e.g., carpeting,
carpet or furniture stain-protective coatings, waterproofing sprays, cleaning agents, kitchen
utensils, clothing, shoes, cosmetics, insecticides, electronic devices). In addition, the extent and
use of the products affects the distribution patterns of PFAS in dust of these buildings.
At this time, there is uncertainty regarding the extent of human exposure to PFBS through indoor
dust compared with other exposure pathways.
Table B-5. Compilation of Studies Describing PFBS Occurrence in Indoor Dust
Study
Location
Site Details
Results
North America
Zheng et al.
(2020)
United States
(Seattle,
Washington and
West Lafayette,
Indiana)
Childcare facilities
(20 samples from 7
facilities in Seattle
and 1 in West
Lafayette)
DF 90%, mean (range) = 0.34
(ND-0.86) ng/g
Byrne et al.
(2017)
United States (St.
Lawrence Island,
Alaska)
Homes (49)
DF 16%, median = ND; 95th
percentile = 1.76 ng/g
Fraser et al.
(2013)
United States
(Boston,
Massachusetts)
Homes (30); offices
(31); vehicles (13)
Homes: DF 3% (single
detection), range = ND-4.98
ng/g
Offices: DF 10%, range = ND-
12.0 ng/g
Vehicles: DF 0%
Knobeloch et al.
(2012)
United States
(Great Lakes Basin,
Wisconsin)
Homes (39)
DF 59%), median (range) = 1.8
(ND-31) ng/g
Strynar and
Lindstrom (2008)
United States
(Cities in North
Carolina and Ohio)
Homes (102) and
daycare centers (10);
samples had been
collected in 2000-
2001 during EPA's
Children's Total
Exposure to
Persistent Pesticides
and Other Persistent
Organic Pollutants
(CTEPP) study
DF 33%), mean (range) = 41.7
(ND-1,150) ng/g
B-31
-------
FINAL
2024
Study
Location
Site Details
Results
Scher et al.
(2019)
United States (Twin
Cities metropolitan
region, Minnesota)
Near former 3M
PFAS production
facility; 19 homes
within the GCA
Entry way: DF 11%, median
(range) = ND (ND-58 ng/g)
Living room: DF 16%, median
(range) = ND (ND-58 ng/g)
Kubwabo et al.
(2005)
Canada (Ottawa)
Homes (67)
DF 0%
Europe
de la Torre et al.
(2019)
Spain (unspecified),
Belgium
(unspecified), Italy
(unspecified)
Homes (65)
Spain: DF 52%, median (range)
= 0.70 (ND-12.0) ng/g
Belgium: DF 27%, median
(range) = 0.40 (ND-56.7) ng/g
Italy: DF 18%, median (range)
= 0.40 (ND-11.6) ng/g
Harrad et al.
(2019)
Ireland (Dublin,
Galway, and
Limerick counties)
Homes (32); offices
(33); cars (31);
classrooms (32)
Homes: DF 81%, mean (range)
= 17 (ND-110) ng/g
Offices: DF 88%, mean (range)
= 19 (ND-98) ng/g
Cars: DF 75%, mean (range) =
12 (ND-170) ng/g
Classrooms: DF 97%, mean
(range) =17 (ND-49) ng/g
Giovanoulis et al.
(2019)
Sweden
(Stockholm)
Preschool s (20)
DF 0%
Winkens et al.
(2018)
Finland (Kuopio)
Homes (63
children's bedrooms)
DF 12.7%), median (range) =
ND (ND-13.5) ng/g
Padilla-Sanchez
and Haug (2016)
Norway (Oslo)
Homes (7)
DF 14% (single detection),
range = ND-3 ng/g
Jogsten et al.
(2012)
Spain (Catalonia)
Homes (10)
DF 60%), range = ND-6.5 ng/g
Haug et al. (2011)
Norway (Oslo)
Homes (41)
DF 22%), mean (range) =1.3
(0.17-9.8) ng/g
B-32
-------
FINAL
2024
Study
Location
Site Details
Results
Huber et al.
(2011)
Norway (Troms0)
Homes (7; carpet,
bedroom, sofa); one
office; one storage
room that had been
used for storage of
"highly
contaminated PFC
[polyfluorinated
compounds] samples
and technical
mixtures for several
years"
All homes: DF NR, median =
1.1 ng/g
Living room: DFa 57%, range =
ND-10.6 ng/g
Carpet, bedroom, sofa: DF 0%
Office: point = 3.8 ng/g
Storage room: point = 1,089
ng/g
D'Hollander et al.
(2010)
Belgium (Flanders)
Homes (45); offices
(10)
Homes: DF 47%, median = 0
ng/g dw
Offices: DF NR, median = 0.2
ng/g dw
Multiple Continents
Kato et al. (2009)
United States
(Atlanta, Georgia),
Germany
(unspecified),
United Kingdom
(unspecified),
Australia
(unspecified)
Homes (39)
DF 92.3%, median (range) =
359 (ND-7,718) ng/g
Karaskova et al.
(2016)
United States
(unspecified)
Homes (14)
DF 60%), mean (range) =1.4
(ND-2.6) ng/g
Canada
(unspecified)
Homes (15)
DF 55%o, mean (range) =1.6
(ND-5.8) ng/g
Czech Republic
(unspecified)
Homes (12)
DF 37.5%o, mean (range) = 3.6
(ND-14.4) ng/g
Notes: DF = detection frequency; GCA = groundwater contamination area; ND = not detected; ng/g = nanogram per gram; NR
= not reported; dw = dry weight
a The DF and/or mean was not reported in the study and was calculated in this synthesis. Means were calculated only when DF
= 100%.
B-33
-------
FINAL
2024
B.3.S. Air
PFAS have been released to air from WWTPs, waste incinerators, and landfills (USEPA, 2016a).
ATSDR (2021) noted that PFAS have been detected in particulates and in the vapor phase in air
and can be transported long distances via the atmosphere; they have been detected at low
concentrations in areas as remote as the Arctic and ocean waters. However, EPA's Toxic Release
Inventory did not report release data for PFBS in 2020 (USEPA, 2022c). In addition, PFBS is not
listed as a hazardous air pollutant (USEPA, 2022d).
B. 3.5.1. Indoor Air
No studies from the U.S. reporting levels of PFBS in indoor air were identified from the peerr-
reviewed or gray literature. However, the EPA identified studies from Europe that are
summarized below. These three studies were conducted in Norway (Barber et al., 2007), Spain
(Jogsten et al., 2012), and Ireland (Harrad et al., 2019).
In Norway, neutral and ionic PFAS were analyzed in four indoor air samples collected from
homes in Troms0 (Barber et al., 2007). PFBS levels were below the limit of quantitation. The
authors noted that measurable amounts of other ionic PFAS were found in indoor air samples,
but levels were not significantly elevated above levels in outdoor air. In Spain, Jogsten et al.
(2012) collected indoor air samples (n = 10) from selected homes in Catalonia and evaluated
levels of 27 perfluorinated chemicals (PFCs). PFBS was not detected.
In Ireland, Harrad et al. (2019) measured eight target PFAS in air from cars (n = 31), home living
rooms (n = 34), offices (n = 34), and school classrooms (n = 28). PFBS was detected in all four
indoor microenvironments, at DFs of 53%, 90%, 41%, and 54% in samples from homes, cars,
offices, and classrooms, respectively. The mean (maximum) concentrations were 22 (270)
picograms per cubic meter (pg/m3) in homes, 54 (264) pg/m3 in cars, 37 (313) pg/m3 in offices,
and 36 (202) pg/m3 in classrooms.
There is some evidence from European studies indicating PFBS exposure via indoor air.
However, further research is needed to understand the DF and concentrations of PFBS that occur
in indoor environments in the United States.
B.3.5.2. Ambient Air
Similar to studies on indoor PFBS air concentrations, no studies from the U.S. reporting levels of
PFBS in ambient air were identified from the peer-reviewed or gray literature. Four studies
conducted across Europe (Harrad et al., 2020; Jogsten et al., 2012; Beser et al., 2011; Barber et
al., 2007) and one study conducted in Canada (Ahrens et al., 2011) analyzed ambient air samples
for PFBS. Two of the studies (Harrad et al., 2020; Barber et al., 2007) found detectable levels of
PFBS in outdoor air. Barber et al. (2007) collected air samples from four field sites in Europe
(one semirural site [Hazelrigg] and one urban site [Manchester] in the United Kingdom, one
rural site from Ireland, and one rural site from Norway) for analysis of neutral and ionic PFAS.
The study authors did not indicate whether any of the sites had a history of local PFAS-related
activities (e.g., AFFF usage, PFAS manufacturing or use). PFBS was detected in the particle
phase of outdoor air samples during one of the two sampling events in Manchester at 2.2 pg/m3
and one of the two sampling events in Hazelrigg at 2.6 pg/m3. PFBS was not detected above the
method quantification limit at the Ireland and Norway sites. Harrad et al. (2020) measured PFBS
B-34
-------
FINAL
2024
in air near 10 Irish municipal solid waste landfills located in non-industrial areas. Air samples
were collected upwind and downwind of each landfill. PFBS was detected in more than 20% of
the samples, with mean concentrations (ranges) at downwind and upwind locations of 0.50 (<
0.15-1.4) pg/m3 and 0.34 (< 0.15-1.2) pg/m3, respectively. Beser et al. (2011) and Jogsten et al.
(2012) did not detect PFBS in ambient air samples in Spain. Beser et al. (2011) analyzed fine
airborne particulate matter (PM2.5) in air samples collected from five stations located in Alicante
province, Spain (3 residential, 1 rural, 1 industrial) to determine levels of 12 ionic PFAS. PFBS
was below the method quantification limit at all five locations. Jogsten et al. (2012) did not
detect PFBS in ambient air samples collected outside homes in Catalonia, Spain.
In the one study identified from North America, Ahrens et al. (2011) determined levels of PFAS
in air around a WWTP and two landfill sites in Canada. PFBS was not detected in any sample
above the method detection limit.
PFBS has been detected in Artie air in one study, with a DF of 66% and mean concentration of
0.1 pg/m3 (Arp and Slinde, 2018; Wong et al., 2018).
As with exposure to PFBS via indoor air, there is some evidence from European studies
indicating PFBS is present in some ambient air samples. Further research is needed to understand
the DF and concentrations of PFBS that occur in ambient environments in the United States.
B.3.6. Soil
PFBS can be released into soil from manufacturing facilities, industrial uses, fire/crash training
sites, and biosolids containing PFBS (USEPA, 2021d; ATSDR, 2021). The EPA identified 16
studies that evaluated the occurrence of PFBS and other PFAS in soil, conducted in the United
States, Canada, or Europe (see Table B-6). Two U.S. studies and two Canadian studies
(Cabrerizo et al., 2018; Venkatesan and Halden, 2014; Blaine et al., 2013; Dreyer et al., 2012)
were conducted in areas not reported to be associated with any known PFAS release or were
experimental studies conducted at research facilities. At these sites, PFBS levels were low (<
0.10 ng/g) or below detection limits in non-amended or control soils. Two U.S. studies by Scher
et al. (2019; 2018) evaluated soils at homes in Minnesota within and outside of a GCA impacted
by a former 3M PFAS production facility; for sites within the GCA, one of the studies reported a
DF of 10%) and a 90th percentile PFBS concentration of 0.02 ng/g, and the other reported a DF of
9% and a maximum PFBS concentration of 0.017 ng/g. For sites outside of the GCA, the DF was
17% and the maximum PFBS concentration was 0.031 ng/g. Three U.S. studies and one
Canadian study analyzed soils potentially impacted by AFFF used to fight fires—one at U.S. Air
Force installations with historic AFFF use (Anderson et al., 2016), two at former fire training
sites (Nickerson et al., 2020; Eberle et al., 2017), and another at the site of a train derailment and
fire in Canada (Mejia-Avendano et al., 2017). In these four studies, DFs ranged from 35 to
100%). PFBS concentrations in the study of the U.S. Air Force installations ranged from ND-79
ng/g, and PFBS concentrations ranged from ND-58.44 ng/g at one fire training site (Nickerson et
al., 2020). The study of the other fire training site measured PFBS pre-treatment (0.61-0.6.4
ng/g) and post-treatment (0.07-0.83 ng/g) (Eberle et al., 2017). The DFs and range of PFBS
concentrations measured in soils at the site of the train derailment were 75%> DF and ND-3.15
ng/g, respectively, for the AFFF run-off area (measured in 2013, the year of accident) and 36%>
DF and ND-1.25 ng/g, respectively, at the burn site and adjacent area (measured in 2015)
(Mejia-Avendano et al., 2017).
B-35
-------
FINAL
2024
Of the six European studies, one study (Harrad et al., 2020) analyzed soil samples collected
upwind and downwind of 10 municipal solid waste landfills in Ireland and found PFBS levels to
be higher in soils from downwind locations. Based on the overall study findings, however, the
authors concluded there was no discernible impact of the landfills on concentrations of PFAS in
soil surrounding these facilities. Grannestad et al. (2019) investigated soils from a skiing area in
Norway to elucidate exposure routes of PFAS into the environment from ski products, such as
ski waxes. The authors found no significant difference in mean total PFAS in soil samples from
the Granasen skiing area and the Jonsvatnet reference area but noted that the skiing area samples
were dominated by long-chain PFAS (C8-C14; > 70%) and the reference area samples were
dominated by short-chain PFAS (> 60%), which included PFBS. A study in Belgium (Groffen et
al., 2019) evaluated soils collected at a 3M fluorochemical plant in Antwerp and at four sites
located at increasing distances from the plant. PFBS levels were elevated at the plant site and
decreased with increasing distance from the plant. The other three studies analyzed soil samples
from areas near firefighting training sites in Norway and France, and reported PFBS
concentrations varying from ND to 101 ng/g dry weight (Dauchy et al., 2019; Skaar et al., 2019;
Hale etal., 2017).
A U.S. study of biosolid samples from 94 WWTPs across 32 states and the District of Columbia
detected PFBS in 60% of samples at a mean concentration (range) of 3.4 (2.5-4.8) ng/g
(Venkatesan and Halden, 2013). As mentioned, PFBS has been detected in drinking water wells,
food types, and plant samples from soils or fields that have received biosolids applications that
were industrially impacted (Blaine et al., 2014; Blaine et al., 2013; Lindstrom et al., 2011).
In summary, results of some available studies suggest that proximity to a PFAS production
facility or a site with historical AFFF use or firefighting is correlated with increased PFBS soil
concentrations compared to soil from sites not known to be impacted by PFAS. However, few
available studies examined PFBS concentrations in soils not known to have nearby sources of
PFBS. Additional research is needed that quantifies ambient levels of PFBS in soils in the United
States.
Table B-6. Compilation of Studies Describing PFBS Occurrence in Soil
Study
Location
Site Details
Results
North America
Venkatesan and
Halden (2014)
United States
(Baltimore,
Maryland)
Control (nonamended)
soil from Beltsville
Agricultural Research
Center
DF 0%
Blaine et al. (2013)
United States
(Midwestern)
Urban and rural full-
scale field study control
(nonamended) soil
Urban control: DF NR, mean =
0.10 ng/g
Rural control: DF NR, mean = ND
Scheretal. (2019)
United States (Twin
Cities metropolitan
region, Minnesota)
Near former 3M PFAS
production facility,
homes within a GCA
DF 10%, median (p90) = ND
(0.02) ng/g
B-36
-------
FINAL
2024
Study
Location
Site Details
Results
Scheretal. (2018)
United States (Twin
Cities metropolitan
region, Minnesota)
Near former 3M PFAS
production facility,
homes within and
outside a GCA
Within GCA: DF 9%, median
(range) = ND (ND-0.17 ng/g)
Outside GCA: DF 17%, median
(range) = ND (ND-0.031 ng/g)
Anderson et al.
(2016)
United States
(unspecified)
Ten U.S. Air Force
installations with
historic AFFF release,
surface and subsurface
soils
Surface soil: DF 35%, median
(range) = 0.775 (ND-52.0) ng/g
Subsurface soil: DF 35%, median
(range) = 1.30 (ND-79.0) ng/g
Eberle et al. (2017)
United States (Joint
Base Langley-
Eustis, Virginia)
Firefighting training
site, pre- and
posttreatment
Pretreatment: DF 60%, range =
0.61-6.4 ng/g
Posttreatment: DF 100%, range =
0.07-0.83 ng/g
Nickerson et al.
(2020)
United States
(unspecified)
Two AFFF-impacted
soil cores from former
fire-training areas
Core E: DFa 91%, range = ND-
27.37 ng/g dw
Core F: DF 100%, range = 0.13—
58.44 ng/g dw
Cabrerizo et al.
(2018)
Canada (Melville
and Cornwallis
Islands)
Catchment areas of
lakes
DF 100%, mean3 (range) = 0.0024
(0.0004-0.0083) ng/g dw
Dreyer et al. (2012)
Canada (Ottawa,
Ontario)
Mer Bleue Bog Peat
samples (core samples)
Detected once at 0.071 ng/g in
1973 sample and not considered
for further evaluation
Mejia-Avendano et
al. (2017)
Canada (Lac-
Megantic, Quebec)
Site of 2013 Lac-
Megantic train accident
(oil and AFFF runoff
area [sampled 2013],
burn site and adjacent
area [sampled 2015])
Background: DF NR, mean =
0.035 ng/g dw
2013: DF 75%, mean range =ND-
3.15 ng/g dw
2015: DF 36%, mean range = ND-
1.25 ng/g dw
Europe
Harrad et al. (2020)
Ireland (multiple
cities)
10 landfills, samples
collected upwind and
downwind
Downwind: DF NR, mean (range)
= 0.0059 (ND-0.044) ng/g dw
Upwind: DF NR, mean (range) =
0.0011 (ND-0.0029) ng/g dw
Gronnestad et al.
(2019)
Norway (Granasen,
Jonsvatnet)
Granasen (skiing area);
Jonsvatnet (reference
site)
Skiing area: DF 0%b
Reference area: DF 70%, mean
(range) = 0.0093 (ND-0.0385 ng/g
dw)
B-37
-------
FINAL
2024
Study
Location
Site Details
Results
Groffen et al.
(2019)
Belgium (Antwerp)
3M perfluorochemical
plant and 4 sites with
increasing distance
from plant
Plant: DF 92%, mean (range) =
7.84 (ND-33) ng/g dw
Vlietbos (1 km from plant): DF
90%, mean (range) = 2.79 (ND-
7.04) ng/g dw
2.3 km, 3 km, 11 km from plant:
DF 0%
Dauchy et al.
(2019)
France (unspecified)
Firefighting training
site, samples collected
in 6 areas collected up
to 15-m depth; in areas
2 and 6, foams used
more intensely and/or
before concrete slab
was built
Areas 1, 3, 4, and 5 combined: DFa
0-10%, range = ND-7 ng/g dw,
across all depths
Area 2: DFa 35%, range = ND-82
ng/g dw, across all depths
Area 6: DFa 55%, range = ND-101
ng/g dw, across all depths
Skaaretal. (2019)
Norway (Ny-
Alesund)
Research facility near
firefighting training site
Background: DF 0%
Contaminated: DF 100%, mean3
(range) = 4.9 (2.64-7.13) ng/g dw
Hale et al. (2017)
Norway
(Gardermoen)
Firefighting training
site
DF 0%
Notes: AFFF = aqueous film-forming foam; DF = detection frequency; dw = dry weight; GCA = groundwater contamination
area; km = kilometer; ND = not detected; ng/g = nanogram per gram; NR = not reported; PFAS = per- and polyfluoroalkyl
substances; p90 = 90th percentile
a The DF and/or mean was not reported in the study and was calculated in this synthesis. Means were calculated only when DF
= 100%.
b Grannestad et al. (2019) reported a DF = 10% but a range, mean, and standard deviation of < LOQ.
B.4. Recommended RSC
The EPA followed the Exposure Decision Tree approach to determine the RSC for PFBS
(USEPA, 2000b). The EPA first identified two potential populations of concern (Box 1):
pregnant women and their developing fetuses and women of childbearing age (see Section 2.2.2).
However, limited information was available regarding specific exposure of these populations to
PFBS in different environmental media. The EPA considered exposures in the general U.S.
population as likely being applicable to these two populations. Second, the EPA identified
several relevant PFBS exposures and pathways (Box 2), including dietary consumption,
incidental oral consumption via dust, consumer products, and soil or dermal exposure via soil,
consumer products, and dust, and inhalation exposure via indoor or ambient air. Several of these
may be potentially significant exposure sources. Third, the EPA determined that there was
inadequate quantitative data to describe the central tendencies and high-end estimates for all of
the potentially significant sources (Box 3). For example, studies from Canada and Europe
indicate that indoor and ambient air may be a significant source of exposure to PFBS. At the time
of the literature search, the EPA was unable to identify studies assessing PFBS concentrations in
B-38
-------
FINAL
2024
indoor or ambient air samples from the U.S. and therefore, the agency does not have adequate
quantitative data to describe the central tendency and high-end estimate of exposure for this
potentially significant source in the U.S. population. However, the agency determined there were
sufficient data, physical/chemical property information, fate and transport information, and/or
generalized information available to characterize the likelihood of exposure to relevant sources
(Box 4). Notably, based on the studies summarized in the sections above, there are significant
known or potential uses/sources of PFBS other than drinking water (Box 6), though there is not
information available on each source to make a characterization of exposure (Box 8A). For
example, there are several studies from the U.S. indicating that PFBS may occur in dust sampled
from various microenvironments (e.g., homes, offices, daycare centers, vehicles). However, the
majority of studies sampled in only one location and few studies examined dust samples outside
of the home (e.g., one study from the U.S. assessed PFBS occurrence in dust sampled from
vehicles). Additionally, though several studies from around the U.S. measured PFBS
concentrations in dust from houses, the detection frequencies in these studies varied widely
(from 3% to 59%) and may be a result of uncertainties including home characteristics, behaviors
of the residents, and the presence or absence of PFBS-containing materials or products (Haug et
al., 2011). Therefore, it is not possible to determine whether dust can be considered a major or
minor contributor to total PFBS exposure. Similarly, it is not possible to determine whether the
other potentially significant exposure sources such as seafood and consumer products should be
considered major or minor contributors to total PFBS exposure. Given these considerations,
following recommendations of the Exposure Decision Tree (USEPA, 2000b), the EPA
recommends an RSC of 20% (0.20) for PFBS.
B-39
-------
FINAL
2024
Appendix C. PFNA: Summary of Occurrence in
Water and Detailed Relative Source Contribution
C.l. Occurrence in Water
The use and production of PFNA could result in its release to the aquatic environment through
various waste streams (NCBI, 2022b). PFNA has an estimated water solubility of 62.5 |ig/L
(6.25 x 10 2 mg/L) at 25°C and when released to surface water, it is expected to adsorb to
suspended solids and sediment (NCBI, 2022b). Volatilization from water surfaces is not
expected to be an important fate process for PFNA (NCBI, 2022b).
C.l.l. Groundwater
Several peer-reviewed studies were identified that examined PFNA occurrence in groundwater
sources from the U.S. In three studies, sampling locations included sites known or suspected to
contain PFAS but not related to AFFF use (Procopio et al., 2017; Post et al., 2013; Lindstrom et
al., 2011). Procopio et al. (2017) evaluated groundwater samples from a small area of an
industrial/business park located within the South Branch Metedeconk River watershed. Two
sampling events were conducted as part of a source trackdown study to identify potential sources
of PFAS contamination after elevated PFOA levels were discovered at a raw surface water intake
of the Brick Township Municipal Utilities Authority. Samples were collected following the
installation of 16 temporary monitoring wells by the New Jersey Geological and Water Survey or
a contract driller during August 2013 and June and July 2014. PFNA was detected in 32% of
samples (n = 19), with concentrations ranging from < 5 to 63 ng/L; a mean concentration was not
reported. The maximum PFNA level found (63 ng/L) occurred at a well located in the middle of
the industrial/business park where the highest PFAAs were detected in the study. Based on the
results of all PFAAs analyzed, the authors concluded a strong likelihood of a groundwater plume
of PFAS contamination resulting from the suspected illicit discharge of liquid waste to soil and
groundwater from a manufacturer of industrial fabrics, composites, and elastomers that use or
produce products containing PFAAs. Post et al. (2013) evaluated raw groundwater samples from
public drinking water system intakes in two sampling campaigns. Between August 2009 and
February 2010, groundwater samples from 18 drinking water systems were obtained from 1
confined well (sunk into an aquifer located between two impermeable strata) and 17 unconfined
wells in the upper unconfined aquifer that were chosen to represent New Jersey geographically.
The sampled locations included one site with a nearby industrial facility that previously used
large quantities of PFNA. PFNA was found in 5 of 18 samples with concentrations ranging from
not detected (ND) to 96 ng/L. The maximum PFNA concentration was at the site with the nearby
industrial facility. Sampling was also conducted in 2010-2013 from unconfined wells of two
additional public drinking water systems with groundwater known to be contaminated by PFOA.
Four wells were sampled at the first system and one well at the second system. PFNA was
detected in all five wells. Sample detection frequencies were not reported but concentrations
ranged from ND to 16 ng/L across the four wells in the first system and from 24 to 72 ng/L in the
second system. Lindstrom et al. (2011) analyzed well water samples from 13 wells used for
livestock, watering gardens, and washing in Decatur, Alabama. The samples were collected in
C-l
-------
FINAL
2024
February 2009 from farms that had applied PFC-contaminated biosolids to local agricultural
fields as a soil amendment for at least 12 years. PFNA was below the LOQ (10 ng/L) in all 13
wells.
Three studies from the U.S. evaluated groundwater potentially impacted by wastewater (Boone
et al., 2019; Appleman et al., 2014; Quinones and Snyder, 2009). In the first study, Boone et al.
(2019), evaluated 17 PFAS in source and treated waters collected in 2010-2012. Of the three
groundwater sources evaluated, PFNA was detected in two out of three samples at levels of 1.25
and 0.156 ng/L. In Appleman et al. (2014), authors assessed source water from five utilities in
New Jersey from November 2011 to September 2012, the majority of which were selected
because they were either known from previous monitoring or expected to contain detectable
PFAS because they were impacted by upstream wastewater effluent discharge. PFNA was
detected in six of seven groundwater samples. The study did not report an average value for
PFNA, but concentrations ranged from below the method RL (0.5 ng/L) to 47 ng/L. In the third
study, Quinones and Snyder (2009) examined levels of eight PFAS at two sites in Nevada that
were highly impacted from treated wastewater. Mean PFNA levels at the two sites were 6.9 and
5.7 ng/L at sites 1 (n = 7) and 2 (n = 8), respectively.
The remaining three U.S. studies identified addressed sites of current and/or historic use of
AFFFs (Steele et al., 2018; Eberle et al., 2017; Anderson et al., 2016). Anderson et al. (2016)
assessed 40 sites across 10 active Air Force installations throughout the continental United States
and Alaska between March and September 2014. Installations were included if there was known
historic AFFF release in the period 1970-1990. It is assumed that the measured PFAS profiles at
these sites reflect the net effect of several decades of all applicable environmental processes. The
selected sites were not related to former fire training areas and were characterized according to
volume of AFFF release—low (n = 24), medium (n = 100), and high (n = 25). Across all sites,
the PFNA detection frequency was 46.38% and the median concentration at sites with detectable
levels was 105 ng/L. PFNA was detected only at low- and medium-volume release sites with
detection frequencies of 37.5% and 40.6%, respectively, and mean concentrations of 300 and
900 ng/L, respectively. Authors noted that given PFNA is not present in 3M AFFF formulations,
there may be some degree of telomer-based AFFF contamination. Steele et al. (2018)
investigated a contaminated military base in Alaska and former Pease Air Force Base (the latter
being historical, secondary data). Authors reported the primary source of contamination for the
Alaska military base to be from prior legacy AFFF use and wells were selected for sampling
based on historical data that indicated PFOS and PFOA contamination. Well samples at the
Alaska base were collected monthly from July 2016 to March 2017 to determine if monthly
variations in PFAS concentrations existed. For four wells, PFNA was detected one to three times
during the monthly sampling at concentrations ranging from 0.91 to 6.6 ng/L. PFNA was not
detected in any of the eight monthly timepoints in two other wells. A seventh well was only
sampled in July 2016 and reported a PFNA concentration of 1.3 ng/L. The authors found that
PFAS concentrations did not vary significantly on the scale of weeks or months. Eberle et al.
(2017) collected groundwater samples first in April and December 2012 as part of a
screening/site characterization analysis. Additional samples were collected in 2013-2014 before
and after a pilot scale field test at a former fire training site at Joint Base Langley-Eustis,
Virginia. Monthly fire training activities were conducted at the site from 1968 to 1980 and
irregular fire training activities continued until 1990. Of the data reported, samples collected for
site characterization showed PFNA was detected in all wells (seven deep, three shallow)
C-2
-------
FINAL
2024
sampled. PFNA was also detected in all pre-treatment samples (n = 5), ranging from
approximately 100 to 1,500 ng/L.
The EPA also identified studies from Canada and Europe reporting the occurrence of PFNA in
groundwater, which are briefly summarized below. Detailed results from each study are
presented in Table C-l. PFNA was detected in groundwater samples collected in 2010 from the
Highland Creek watershed in Canada at concentrations ranging from 0.071 ng/L to 0.54 ng/L
(Meyer et al., 2011, as cited in NCBI, 2022). In this study, the authors reported that none of the
sampling sites receive water that is impacted by known PFAS point sources (Meyer et al., 2011).
In Europe, PFNA investigations in groundwater were conducted in France (Dauchy et al., 2019;
Bach et al., 2017; Boiteux et al., 2017; Dauchy et al., 2017; Gellrich et al., 2013; Boiteux et al.,
2012; Dauchy et al., 2012), Germany (Gellrich et al., 2013), Ireland (Harrad et al., 2020), Italy
(Barreca et al., 2020; Ciofi et al., 2018; Gellrich et al., 2013), Malta (Sammut et al., 2019),
Norway (Fteisseter et al., 2019), Spain (Jurado-Sanchez et al., 2013; Llorca et al., 2012) and
Sweden (Gobelius et al., 2018; Gyllenhammar et al., 2015). Loos et al. (2010) also collected
groundwater from 164 groundwater monitoring stations of participating European Union
laboratories, within 23 countries, in the fall of 2008. They detected PFNA in 15.2% of these
collected samples with concentrations up to 10 ng/L. Gellrich et al. (2013) also collected samples
in multiple countries along the Rhine River, collecting both river filtrate and combined
groundwater and percolated water from the Rhine riverbed in Germany, France, and Italy. In this
campaign, they did not detect PFNA from the Rhine River or riverbed groundwater.
In the studies conducted in countries along the Mediterranean Sea, there was little-to-no detected
PFNA in groundwater sources. In Spain, no PFNA was detected in well water samples collected
by Jurado-Sanchez et al. (2013) in the southeast of the country nor by Llorca et al. (2012) in
Barcelona. In the Lombardia region of Italy, Barreca et al. (2020) detected PFNA in 3% of 130
collected groundwater samples across 57 sampling stations in 2018. Ciofi et al. (2018) collected
groundwater samples at 12 locations across Tuscany, including Siena, Florence, and Prato. One
grab sample was collected at each of these 12 locations, detecting PFNA at each with
concentrations between <0.26-4.8 ng/L (Ciofi et al., 2018). Sammut et al. (2019) collected
groundwater from ten boreholes across the island country in 2015-2016, which were sites used
by the Malta Water Services Corporation for both water extraction and quality analysis sampling.
Across these ten boreholes, they detected PFNA in one borehole at a concentration of 0.90 ng/L
(Sammut et al., 2019).
The EPA identified a number of studies reporting PFNA measurements within France (Dauchy
et al., 2019; Bach et al., 2017; Boiteux et al., 2017; Dauchy et al., 2017; Gellrich et al., 2013;
Boiteux et al., 2012; Dauchy et al., 2012). Boiteux et al. (2012) analyzed raw water from
drinking water treatment plants distributed across 100 French departments, representing
approximately 20% of the national water supply flow in 2009 and 2010. In their first sampling
campaign in 2009, they detected PFNA in 6% of collected samples, with a maximum
concentration of 14 ng/L. In their second sampling campaign in 2010, they did not detect PFNA
at a limit of detection of 1.3 ng/L (Boiteux et al., 2012). In 2013, Bach et al. (2017) and Boiteux
et al. (2017) evaluated the PFNA concentration in alluvial wells that are influent groundwater to
drinking water treatment plants in southern and northern France, respectively. Both Bach et al.
(2017) and Boiteux et al. (2017) sampled groundwater downstream from industrial sites which
produce fluoropolymers and fluorotelomer-based products. In southern France, Bach et al.
(2017) detected FPNA in 86-100%) of collected samples, with concentrations from <4 pg/m3 (the
C-3
-------
FINAL
2024
limit of quantification) to 37 pg/m3. Alternatively, in northern France, Boiteux et al. (2017) did
not detect PFNA in any of the sampled alluvial wells.
A number of studies focused specifically on investigating sites where PFAS-containing material
had been heavily used. In Norway, Fteisseter et al. (2019) analyzed groundwater at a firefighting
training site that extensively used PFAS-containing materials until their ban in 2011. Five
monitoring wells were sampled in 2016, totaling 19 sampling campaigns, which detected a mean
concentration of approximately 950 ng/L in the monitoring wells at the site. Dauchy et al. (2019)
investigated a similar firefighting training facility with heavy PFAS use in France in 2015-2016
but detected no groundwater contamination. Neighboring Norway, in Sweden, Gyllenhammar et
al. (2015) sampled monitoring, private, and production wells for four drinking water treatment
plants downstream of a military airport with firefighting training activities. Unlike the findings of
Fteisseter et al. (2019) in Norway, they found no detectable PFNA across the wells they sampled
in 2012-2014 (Gyllenhammar et al., 2015). Gobelius et al. (2018) also sampled at "PFAS hot
spots" (e.g., firefighter training sites, sewage treatment plants, landfills) across Sweden,
detecting PFNA in 27% of samples with concentrations between <0.08-66 ng/L. In Ireland,
Harrad et al. (2020) collected groundwater samples from boreholes downgradient from ten
municipal solid waste landfills across the country, which accepted municipal waste, non-
hazardous industrial waste, construction and demolition waste, and biomedical waste. Across
these sites, they detected PFNA in 10% of groundwater samples, with concentrations ranging
from <0.1-0.22 ng/L.
Dauchy et al. (2012) sampled raw water from monitoring groundwater wells at a fluoropolymer
manufacturing plant and at two drinking water treatment plants downstream of the
manufacturing plant and many other domestic and industrial activities. At the fluoropolymer
manufacturing plant, three of the four sampled wells reported detectable levels of PFNA from
21-724 ng/L. Downstream, at the drinking water treatment plants, five of five sampled
monitoring wells contained detectable PFNA ranging from 13-35 ng/L (Dauchy et al., 2012).
Dauchy et al. (2017) also conducted sampling campaigns in late 2014 through early 2015,
investigating a previously operated oil refinery, a military airport, and a training center for
firefighters. These sites were selected due to their heavy use of fluorosurfactant-based foams,
and samples were collected from groundwater monitoring wells. PFNA was detected in the oil
storage depot, between 11-12 ng/L in October 2014 and March 2015. During this period, no
PFNA was detected in the military airport or firefighter training center groundwater (limit of
quantification=4 ng/L) (Dauchy et al., 2017).
C-4
-------
FINAL
2024
Table C-l. Summary of Studies Reporting the Occurrence of PFNA in Groundwater
Study
Location
Site Details
Results
United States
Procopio etal. (2017)
United States (New Jersey)
Groundwater from an industrial/business park located
within the South Branch Metedeconk River watershed,
where there was suspected illicit discharge to soil and
groundwater from a manufacturer of industrial fabrics,
composites, and elastomers that use or produce
products containing PFAAs. Samples were collected
following the installation of 16 temporary monitoring
wells by the NJ Geological and Water Survey or a
contract driller during August 2013 (sampling event
#7) and June-July 2014 (sampling event #8). Samples
were taken from the upper 1.5 m (5 ft) of the water
table from each well, except for one "profile well" in
which samples were collected at three different depths
(3.7-4.6, 6.7-7.6, and 10.7-11.6 m below grade; 12-
15,22-25, and 35-38 ft below grade, respectively).
n = 19, DFa 32%, range = <5-63 ng/L
(minimum reporting level = 5 ng/L)
Post et al. (2013)
United States (New Jersey)
Raw water collected from public drinking water
system intakes in two sampling campaigns. In the first
sampling campaign, samples from 18 drinking water
systems were collected between August 2009 and
February 2010 from 1 confined well (sunk into an
aquifer located between two impermeable strata) and
17 unconfined wells in the upper unconfmed aquifer;
sites were chosen to represent NJ geographically and
included 1 site with a nearby industrial facility that
previously used large quantities of PFNA (site 5). In
the second sampling campaign, samples from two
drinking water systems (PWS-A and PWS-B) were
collected in 2010-2013 from five unconfmed wells.
Groundwater at these two systems were known to be
contaminated by PFOA.
1st sampling campaign:
n = 18, DF 28%, range = ND-96 ng/L
2nd sampling campaign:
PWS-A, WF1A: n = 5, DF NR, range =
ND-6 ng/L
PWS-A, WF1B: n = 4, DF NR, range = ND-
12 ng/L
PWS-A, WF2A: n = 9, DF NR, range =
ND-16 ng/L
PWS-A, WF2B: n = 9, DF NR, range = ND-
7 ng/L
PWS-B: n = 8, DF NR, range = 24-72 ng/L
(minimum reporting level = 5 ng/L)
Lindstrom et al. (2011)
United States (Decatur,
Alabama)
Thirteen samples collected in February 2009 from 13
wells located on farms with historical land application
of PFC-contaminated biosolids to local agricultural
fields between 1995 and 2008. Biosolids obtained
from local municipal WWTP where sources
discharging to the WWTP included facilities involved
in the production and use of fluoropolymers,
n = 13, DF (frequency of quantification) 0%
(LOQ = 10 ng/L)
C-5
-------
FINAL
2024
Study
Location
Site Details
Results
fluorocarbon fibers, polymers, polymer films, and
resins.
Boone et al. (2019)
United States (unspecified)
Three groundwater sites used as source waters for
three DWTPs, collected in 2010-2012; some locations
with known or suspected sources of wastewater in the
source water, but study did not differentiate which
locations had known or suspected sources.
n = 3, DFa 67%, range = ND-1.25 ng/L
(LCMRL = 0.094 ng/L)
Appleman et al. (2014)
United States (New Jersey)
Groundwater source water for five DWTPs, sampled
November 2011 to September 2012. Majority of the
utilities were selected because they were either known
from previous monitoring or expected based on their
source waters to contain detectable PFAS (i.e.,
impacted by upstream wastewater effluent discharge).
Two sites were sampled twice and three sites were
sampled only once.
n = 7, DFa 86%, range =
-------
FINAL
2024
Study
Location
Site Details
Results
samples were collected from existing monitoring wells
and temporary monitoring wells installed with direct
push technology.
*Non-detects were substituted with Vi the
reporting limit
Steele et al. (2018)
United States (Alaska)
Monthly samples collected from a military installation
during July 2016-March 2017; six wells from around
the installation were sampled each month, along with
a seventh well that was only sampled in July 2016.
PFAS contamination predominately from prior legacy
AFFF use. Wells selected based on historical sample
data indicating PFAS contamination.
Data for July, August, September, October,
November, December, January, and February,
respectively:
Well A: 1, ND, ND, ND, ND, ND, ND, ND
ng/L
Well B: 1.2, ND, ND, ND, ND, ND, ND, ND
ng/L
Well D: 1.3 ng/L (no values provided for other
months)
Well E: ND, ND, ND, ND, ND, ND, ND, ND
ng/L
Well F: ND, ND, ND, ND, ND, ND, ND, ND
ng/L
DK: 4.8, ND, ND, ND, 6.4, 6.6, ND, ND ng/L
FG: 0.91, 0.91, ND, ND, ND, ND, ND, ND
ng/L
(method detection limit not reported)
Eberle et al. (2017)
United States (Joint Base
Langley-Eustis, Virginia)
Pilot testing area in former fire training area (Training
Site 15) at Joint Base Langley-Eustis where monthly
fire training activities were conducted from 1968 to
1980 in a zigzag pattern burn pit. Facility was
abandoned in 1980 but irregular fire training activities
using an above-ground germed burn pit continued
until 1990. Groundwater samples collected for
screening/site characterization (April and December
2012), and for pre- (April 2013) and post- (October
2013 and February 2014) in situ chemical oxidation
treatment using a peroxone activated persulfate
(OxyZone) technology. Treatment was conducted in
Test Cell 1 over 113 days (April through August
2013). Pre-treatment samples were collected from 14
wells screened in the deep zone, and 3 wells screened
in the shallow zone. Post-treatment samples were
collected from the same wells as the pre-treatment
samples with an additional three wells (two shallow,
one deep) sampled. Wells EC-1, EC-2, EC-3, EC-4,1-
Screening/site characterization:
EC-1 (deep, sentry): 100 ng/L
EC-2 (deep, sentry): 600 ng/L
1-1 (deep): 700 ng/L
1-2 (deep, sentry): 400 ng/L
1-4 (deep): 900 ng/L
1-5 (shallow): 100 ng/L
1-6 (shallow): 200 ng/L
MW-2904 (deep): 100 ng/L
U-16D (deep): 1,700 ng/L
U-16S (shallow): 200 ng/L
Pre-treatment (values reported for two
different laboratories):
EC-2: 900; 500 ng/L
EC-3: 1,500; 700 ng/L
1-1: 200; 300 ng/L
1-2: 200; 100 ng/L
1-4: 1,700; 800 ng/L
C-7
-------
FINAL
2024
Study
Location
Site Details
Results
2, and 1-3 were sentry wells to monitor the possible
migration of oxidants and contaminants outside Test
Cell 1. No PFNA data reported for post-treatment
samples.
(LOQ not reported)
Europe
Bachetal. (2017)
France (southern)
Samples were collected from alluvial wells that
provide source water for two DWTPs. The two
DWTPs are located on both sides of a river, ~15 km
downstream from an industrial site where two
facilities produce fluoropolymers; the industrial site
discharges its effluents at three points along a river.
The alluvial wells are located along the river, with
wells for the first DWTP (DWTP A) located on the
left shore and alluvial wells for the second DWTP
(DWTP B) located on the right shore, on an island
formed by a backwater. Sample collection occurred in
April, July, October, and December 2013.
Alluvial wells for DWTP A:
April 2013: n = 7, DFa 86%, range = <4-25
ng/L
July 2013: n = 7, DFa 86%, range = <4-25
ng/L
October 2013: n = 7, DFa 86%), range = <4-
37 ng/L
December 2013: n = 7, DFa 86%o, range =
<4-30 ng/L
Alluvial wells for DWTP B:
April 2013: n = 8, DFa 100%, meana (range)
= 8.13 (4-17) ng/L
July 2013: n = 8, DFa 88%, range = <4-15
ng/L
October 2013: n = 7, DFa 100%), meana
(range) = 9.43 (5-15) ng/L
December 2013: n = 8, DFa 100%o, meana
(range) = 7.38 (4-10) ng/L
(LOQ = 4 ng/L)
*DF represents frequency of quantification
Boiteux et al. (2017)
France (northern)
Samples were collected in four sampling campaigns
(May, July, October, and December 2013) from
alluvial wells that provide source water for two
DWTPs. The two DWTPs (A and B) are located
downstream of an industrial WWTP that processes
raw sewage from a facility that manufactures
fluorotelomer-based products and side-chain-
fluorinated polymers used in firefighting foams and
stain repellents.
DWTP A is located 15 km downstream from the
WWTP and is supplied by five alluvial wells. DWTP
B is located 20 km downstream of the WWTP and is
supplied by four alluvial wells.
DWTP A:
May 2013: n = 5, DF 0%
July 2013: n = 5, DF 0%
October 2013: n = 5, DF 0%o
December 2013: n = 5, DF 0%o
DWTP B:
May 2013: n = 4, DF 0%
July 2013: n = 4, DF 0%
October 2013: n = 4, DF 0%o
December 2013: n = 4, DF 0%o
(LOQ = 4 ng/L)
*DF represents frequency of quantification
Dauchy et al. (2012)
France (unspecified)
Raw water sampled in June 2010 from four
monitoring wells at a fluoropolymer manufacturing
Fluoropolymer manufacturing plant:
C-8
-------
FINAL
2024
Study
Location
Site Details
Results
plant (PI3, PI4, PI5, P01). Groundwater flowed from
well P14 to P01 and well PI 5 is nearest to the
polyvinylidene fluoride production area.
Raw water resources also collected from two DWTPs
(five sampling sites - DWA-1, DWA-2, DWA-3,
DWA-4, DWB-1); the first DWTP (DWA) is supplied
by four alluvial wells, and the second DWTP (DWB)
is supplied by one alluvial well. The two DWTPs are
located on both sides of a river, 15 km downstream of
fluorochemical manufacturing facility. The river
receives wastewater from many domestic and
industrial activities.
PI 3: not quantifiable due to dilution or
matrix effects
P14: n = NR, DF NR, 21 ng/L
P15: n = NR, DF NR, 342 ng/L
P01: n = NR, DF NR, 724 ng/L
DWA:
DWA-1: n = NR, DF NR, 13 ng/L
DWA-2: n = NR, DF NR, 35 ng/L
DWA-3: n = NR, DF NR, 30 ng/L
DWA-4: n = NR, DF NR, 33 ng/L
DWB:
DWB-1: n = NR, DF NR, 21 ng/L
(LOQ = 4 ng/L)
* Study did not indicate whether
concentrations reported were point values or
means
Harrad et al. (2020)
Ireland (multiple cities)
Groundwater samples collected between November
2018 and January 2019 from ten municipal solid waste
landfills at two sampling points down-gradient from
the main body of each landfill. Each sampling point
consisted of a borehole leading down to water
reservoirs at a minimum depth of 5 m below ground
level. Waste accepted by the landfills included:
municipal solid waste, industrial (non-hazardous)
waste, construction and demolition, and biomedical
waste.
n = 10, DFa 10%, range = <0.1-0.22 ng/L
(LOD = <0.1 ng/L)
*Non-detects replaced by Vi LOD
Gobelius et al. (2018)
Sweden (national)
Sampling conducted between May and August 2015,
with the majority in July and a few samples in
September and November 2015. Samples were
collected in 21 regional counties by the County
Administration Boards. Sampling locations selected
based on potential vicinity of PFAS hot spots (i.e., fire
training sites, unspecific industry, sewage treatment
plant effluent, landfill/waste disposal, skiing, and
urban areas) and/or importance as a drinking water
source. Sample numbers varied for each county and
sampling sites were spread unevenly across Sweden.
n = 161, DFa 27%, range = <0.08-66 ng/L
(method detection limit = 0.084 ng/L)
C-9
-------
FINAL
2024
Study
Location
Site Details
Results
Boiteux et al. (2012)
France (national)
Raw water from DWTPs distributed across 100
French departments to represent -20% of the national
water supply flow; samples collected during two
sampling campaigns in July-September 2009 (first
campaign) and June 2010 (second campaign - focused
on sites from first sampling campaign that had PFC
levels >LOQ). Some sites possibly affected by
industrial or commercial releases.
Overall: n = 196, DF (frequency of
quantification) 3%, maximum =14 ng/L
1st Sampling Campaign
n = 163, DF 6%, mean, median (maximum)
= <1, <1 (14) ng/L
2nd Sampling Campaign
n = 33; results not reported
(LOD =1.3 ng/L, LOQ = 4 ng/L)
Loos et al. (2010)
23 European countries
Groundwater collected from 164 groundwater
monitoring stations of participating European Union
Member State laboratories during an 8-week window
in Fall 2008. There were no strict selection criteria for
the sampling sites such as "representative" or
"contaminated". Most monitoring stations were
"official" monitoring stations also used for drinking
water abstraction.
n = 164, DF 15.2%, mean, median (maximum)
= 0,0 (10) ng/L
(LOD = 0.4 ng/L)
Dauchy et al. (2019)
France (unspecified)
Samples collected in two sampling campaigns in and
around site where fluoro surfactant-based foams have
been used extensively. From 1969 to 1984, the site
was an oil refinery, with the exact location of the
firefighting training area, frequency of training
sessions, and history of firefighting training activities
unknown. From 1987 to date, it has been a large
training area for firefighters. First sampling campaign
collected 13 samples from 9 monitoring wells and 4
springs in June 2015. Second sampling campaign
collected from four monitoring wells in October 2016.
Monitoring wells MW-1 to MW-5 were located
upgradient from the firefighter training site around a
landfill site. Monitoring well MW-11 and springs SW-
A, SW-B, and SW-D located downgradient from the
landfill or firefighter training site but not in the
direction of groundwater flow. Monitoring wells MW-
6 to MW-13 and spring SW-C were located
downgradient from the firefighter training site in the
direction of groundwater flow.
Upgradient:
Monitoring wells: n = 5, DF 0%
Downgradient but not in the direction of
groundwater flow:
Monitoring wells: n = 1, DF 0%
Spring water: n = 3, DF 0%
Downgradient in the direction of groundwater
flow:
Monitoring wells: n = 7, DF 0%
Spring water: n = 1, DF 0%
(LOQ = 4 ng/L)
Lfoisaster et al. (2019)
Norway (unspecified)
Firefighting training site with an airport that
extensively used AFFF containing PFOS since the
early 1990s until 2001 when it was replaced by
fluorotelomer containing AFFF. All PFAS containing
n = 19, DF NR, mean* = 950 ng/L
(LOD/LOQ not reported)
*Mean estimated from Figure 4b in the paper
C-10
-------
FINAL
2024
Study
Location
Site Details
Results
firefighting foams was banned at the airport in 2011.
Groundwater samples collected in 2016 at five
pumping wells installed down gradient of the site to
intercept and pump and treat the plume spreading from
the firefighting training site. A total of 19 sampling
campaigns were performed.
Dauchy et al. (2017)
France (unspecified)
Samples collected in the vicinity of three sites (A, C,
D) where fluorosurfactant-based foams are or were
being heavily used. Site A is an oil storage depot
located in a river port. In June 1987, a large explosion
occurred in the depot and the fire was extinguished by
applying a large amount of fluorosurfactant-based
foams. Two groundwater samples were collected in
October 2014 and March 2015 from a monitoring well
located in the center of the depot. The water table lies
2.5 - 3.5 m below the ground.
Site C is a military airport, with the exact location of
the training area, frequency of the training sessions,
and history of the firefighting training activities
unknown. The well supplying the DWTP was sampled
in March 2015.
Site D is a training center for firefighters. From 1969
to 1984, the site was an oil refinery. Starting in 1987,
the site became a training area for firefighters, with
exercises carried out directly on the soil. From the
1990s, some exercise areas were covered with
concrete. In November 2014, groundwater samples
were collected from five springs.
Site A:
October 2014: n = 1, point =11 ng/L
March 2015: n = 1, point = 12 ng/L
Site C: n = 1,DF0%
Site D: n = 5, DFa 0%
(LOQ = 4 ng/L)
Gyllenhammar et al. (2015)
Sweden (Uppsala)
Three observation well sites (Tuna backar: n = 3
wells; Svartbacken: n = 1 well, Libroback: n = 2
wells;) were sampled from September 2012 to January
2013.
Four DWTP production well sites (Storvad: n = 9
wells; Galgbacken: n = 1 well; Stadstradgarden and
Kronasen: n = 6 wells; Sunnersta: n = 5 wells) were
sampled from July 2012 to February 2014.
One private well (Klastorp) was sampled in September
2012.
Observation wells:
Tuna backar: n = 3, DF 0%
Libroback: n = 4, DF 0%
Svartbacken: n = 3, DF 0%
Production wells:
Storvad: n = 12, DF 0%
Galgbacken: n = 7, DF 0%
Stadstradgarden and Kronasen: n = 103, DF
0%
Sunnersta: n = 50, DF 0%
Private well:
Klastorp: n = 1, DF 0%
C-ll
-------
FINAL
2024
Study
Location
Site Details
Results
All wells located downstream of a military airport
with firefighting training activities up to the year
2003. It is not known when the usage of AFFF started.
(method detection limit =10 ng/L)
Barreca et al. (2020)
Italy (Lombardia region)
Fifty-seven groundwater sampling stations throughout
the region. Samples collected in 2018.
n = 130, DFa 3%, range NR
(LOQ = 1 ng/L)
Sammut et al. (2019)
Malta
Groundwater collected from ten boreholes at different
areas on the island during November and December
2015 and January 2016. Collection sites were the most
commonly used extraction sites by the Malta Water
Services Corporation for water extraction as well as
for sampling for water quality analysis.
n = 10, DF 10%, range = ND-0.90 ng/L
(LOD = 0.02 ng/L; LOQ = 0.04 ng/L)
Ciofietal. (2018)
Italy (Tuscany)
Groundwater samples were collected at 12 locations.
Sampling year was not reported. Each sample was
collected from the phreatic layer with a mean depth
between 10-75 m:
GW-1: Siena, 10 m
GW-2, GW-3, GW-4: Florence, 15 m
GW-5: Prato, 75 m
GW-6: Prato, 71 m
GW-7: Prato, 70 m
GW-8: Prato, 61 m
GW-9: Florence, 17 m
GW-10, GW-11, GW-12: Florence, 10 m
GW-1: n= 1, point = <0.28 ng/L
GW-2: n = 1, point = <0.28 ng/L
GW-3: n = 1, point = <0.28 ng/L
GW-4: n = 1, point = <0.29 ng/L
GW-5: n = 1, point = <0.33 ng/L
GW-6: n = 1, point = <0.46 ng/L
GW-7: n = 1, point = 3.2 ng/L
GW-8: n = 1, point =1.3 ng/L
GW-9: n = 1, point = 4.8 ng/L
GW-10: n = 1, point = <0.27 ng/L
GW-11: n= 1, point = <0.26 ng/L
GW-12: n = 1, point = <0.27 ng/L
(MDL = 0.26-0.46 ng/L)
*method detection limit varied by sample and
was provided for 9 of 12 samples
Gellrich et al. (2013)
Germany (Hesse); France; Italy
Untreated water samples for preparation of mineral
water included seven from Hesse, three from France,
and four from Italy. The supplying waterworks obtain
their untreated water either from Rhine river filtrate, a
mixture of ground water and percolation water from
the Rhine riverbed, drawn from wells 30-50 m deep or
from wells in their closer vicinity. Sampling year not
reported.
n = 14, DF (frequency of quantification) 0%
(LOQ = 1 ng/L)
C-12
-------
FINAL
2024
Study
Location
Site Details
Results
Jurado-Sanchez et al. (2013)
Spain (southeast)
Well water samples were collected and analyzed using
a newly developed analytical method. Authors did not
report sampling details or sampling year.
n = NR, DF 0%
(LOD = 0.1 ng/L)
Llorca et al. (2012)
Spain (Barcelona)
Well water samples from two different sites were
collected from the North of Barcelona metropolitan
area in 2011.
n = 2, DF 0%
(method LOD = 1.9; method LOQ = 6.3 ng/L)
Notes: AFFF = aqueous film-forming foam; DF = detection frequency; DWTP = drinking water treatment plant; ft = feet; m = meter; ND = not detected; ng/L = nanogram per
liter; PFAA = perfluoroalkyl acid; PFAS = per- and polyfluoroalkyl substances; NR = not reported; LOD = limit of detection; LOQ = limit of quantification; LCMRL = lowest
concentration minimum reporting level; WWTP = wastewater treatment plant.
a The DF and/or mean was calculated using point data. Means were calculated only when DF = 100%.
C-13
-------
FINAL
2024
C.1.2. Surface Water
The EPA identified many studies that reported on the occurrence of PFNA in surface water in
both the U.S. and internationally. Overall, most U.S. studies reported PFNA detected in at least
one surface water sample site in each study. Concentrations of PFNA in 12 remote and urban
Minnesota surface water samples, including samples collected from Lake Michigan, ranged
from <0.3 ng/L to 3.1 ng/L (<0.0003 |ig/L to 0.0031 |ig/L) (Simcik and Dorweiler, 2005, as cited
in ATSDR, 2021). PFNA was detected in 38% of eight surface water samples from U.S. streams
in the Great Lakes basin collected during 1994 to 2000 at concentrations between 0.03 ng/L and
0.4 ng/L (0.00003 |ig/L and 0.0004 |ig/L) (Klecka et al., 2010, as cited in NCBI, 2022). PFNA
was detected in six locations in the Delaware River at concentrations ranging from 1.65 ng/L to
976 ng/L (0.00165 |ig/L to 0.976 |ig/L) in 2007 to 2009 (DRBC, 2013, as cited in ATSDR,
2021).
Three studies investigated surface water upstream and downstream of fluoropolymer facilities,
with some sites also downstream of other potential PFAS sources (e.g., landfills, WWTPs)
(Galloway et al., 2020; Newsted et al., 2017; Newton et al., 2017). Galloway et al. (2020)
assessed several rivers and tributaries along the Ohio River in three sampling trips in 2016. The
sampling sites ranged from upstream, downstream, and north/northeast of a fluoropolymer
facility and known PFAS containing landfills. In June 2016, samples were collected on a 188 km
stretch of the Ohio River, from 130 km downstream to 58 km upstream of the facility, and
tributaries that pass near known PFAS-containing landfills. In July 2016, samples were collected
from lakes, rivers, and creeks to the north and northeast of the facility as far as 16 km downwind.
The December 2016 trip expanded the collection radius to more than 48 km downwind to the
north and northeast of the facility. PFNA was detected in 92% of samples (n = 26) in June 2016,
however all detects were below the LOQ (10 ng/L). From the second sampling trip, PFNA was
not detected in any sample in July 2016 (n = 25). Finally, in December 2016, PFNA was
detected at levels above the LOQ in one sample at 24.2 ng/L, detected but below the LOQ in 31
samples, and not detected in 8 samples. In Newsted et al. (2017), surface water samples were
collected in August 2011 from a 3-mile section of the Upper Mississippi River: ten sampling
reaches (three samples each) in an area between Ford Dam (between Minneapolis and St. Paul)
and Hastings Dam (near Hastings) and which had been subject to 10-15 years of actions to
reduce PFAS contamination from 3M Cottage Grove plant and other commercial/industrial
entities. PFNA was detected in one sample from reach 10, immediately downstream the 3M
Cottage Grove facility outfall, at a concentration of 2.0 ng/L. PFNA in all other samples was
below the LOQ (2.0 ng/L). Newton et al. (2017) investigated surface water upstream and
downstream of facilities that manufactured or used fluorinated materials along the Tennessee
River near Decatur, Alabama. Six sampling sites were located upstream of the manufacturing
facilities and three sites were downstream. Among the upstream sites, three were also upstream
of a WWTP. All samples were collected in October 2015. PFNA was below the LOQ (10 ng/L)
in all nine samples from the nine different sampling sites.
In four U.S. studies, sampling locations included surface waters potentially impacted by current
and/or historic use of AFFFs (Anderson et al., 2016; Post et al., 2013; Nakayama et al., 2010;
Nakayama et al., 2007). Anderson et al. (2016) assessed 40 sites across 10 active Air Force
installations throughout the continental United States and Alaska between March and September
2014. Installations were included if there was known historic AFFF release in the period 1970-
C-14
-------
FINAL
2024
1990. It is assumed that the measured PFAS profiles at these sites reflect the net effect of several
decades of all applicable environmental processes. The selected sites were not related to former
fire training areas and were characterized according to volume of AFFF release—low (n = 2),
medium (n = 32), and high (n = 2). PFNA was detected only at medium-volume release sites
(11.4% detection frequency; mean concentration of 15,400 ng/L). Across all 36 sites, the
detection frequency was 36.00% with a median concentration of detects of 96 ng/L. Authors
noted that given PFNA is not present in 3M AFFF formulations, there may be some degree of
telomer-based AFFF contamination. Post et al. (2013) evaluated raw surface water samples from
12 public drinking water system intakes collected between August 2009 and February 2010. Six
rivers and six reservoirs, including two reservoirs in Atlantic County near a civil-military airport
with possible AFFF use, were selected to represent New Jersey geographically. PFNA was
below the minimum RL (5 ng/L) in all six river samples. In reservoir samples, PFNA was
detected in 67% of samples (n = 6) at a maximum level of 19 ng/L. At the two reservoir sites
near the civil-military airport, PFNA was below the minimum RL (5 ng/L) at one, and the other
found PFNA at 5 ng/L. Two studies from Nakayama et al. (2010; 2007) assessed surface water
samples from the Upper Mississippi River, Missouri River, and Cape Fear River Basins. In
Nakayama et al. (2010), a large-scale evaluation of the Upper Mississippi River Basin and
portion of the Missouri River Basin was conducted to provide preliminary PFC data given the
importance of the two basins in supplying drinking water. Between the two basins, 173 samples
were collected across 88 sampling sites in March-August 2008 by several different agencies—
Minnesota Pollution Agency, Wisconsin Department of Natural Resources, Illinois
Environmental Protection Agency, and the EPA Region 7 Water Quality Monitoring Team.
Overall, the detection frequency of PFNA was 87% with a median concentration of 0.71 ng/L.
Authors reported higher PFC concentrations adjacent to chemical manufacturers, downstream of
WWTPs receiving waste from those types of manufacturers, and near an airport with historic use
of firefighting foams. In Nakayama et al. (2007), one hundred surface water samples were taken
from 80 sites selected to reflect water quality throughout the basin. PFNA was detected in 89.9%
of samples with mean and median concentrations of 33.6 and 5.70 ng/L, respectively. The
highest concentrations were found in the middle reaches of the Cape Fear River and its two
major tributaries. The authors noted possible sources of PFCs to the basin included firefighting
foam from nearby air force bases and commercial/industrial facilities.
Three studies conducted in the U.S. examined surface water near or downstream of land
application sites where PFC-contaminated WWTP effluent or biosolids were applied (Lasier et
al., 2011; Lindstrom et al., 2011; Konwick et al., 2008). In Lindstrom et al. (2011), authors
analyzed surface water samples from ponds and streams in Decatur, Alabama. The samples were
collected in February 2009 from farms that had applied PFC-contaminated biosolids to local
agricultural fields as a soil amendment for at least 12 years. The biosolids were obtained from a
local municipal WWTP where authors noted that sources discharging to the WWTP included
facilities involved in the production and use of fluoropolymers, fluorocarbon fibers, polymers,
polymer films, and resins, although specific sources could not be characterized. PFNA was
detected in 28% of samples (n = 32), with levels ranging from below the LOQ (10 ng/L) to 286
ng/L. The remaining two studies (Lasier et al., 2011; Konwick et al., 2008) evaluated surface
water upstream and downstream of a land application site (LAS) in Georgia, where treated
WWTP effluent was sprayed. The WWTP processed effluents from multiple carpet
manufacturers who were reported to use significant quantities of PFCs. Lasier et al. (2011)
sampled along the Conasauga, Oostanaula, and Coosa Rivers during summer 2008; samples
C-15
-------
FINAL
2024
included two sites upstream (sites 1 and 2) and six sites downstream (sites 3-8) of the LAS.
Additionally, site 2 was downstream of a local airport, and site 4 was downstream of a
manufacturing facility of latex and polyurethane back material—inputs for the carpet
manufacturers. PFNA was below the MDL (0.005 ng/g) at both of the sites upstream. Mean
concentrations for sites 3-8 were 35, 44, 17, 21, 20, and 21 ng/L, respectively. Authors reported
highest concentrations downstream of the land application and backing-material sites and then
decreased concentrations increasingly downstream as a result of dilution. Konwick et al. (2008)
sampled from the Conasauga River near Dalton, Georgia in March 2006 at one location
upstream, two downstream, and one at the LAS. Sampling was also conducted in January 2005 at
one freshwater location in the Altamaha River, a river remote from the carpet industry, and from
four different streams and ponds located in Dalton, Georgia. PFNA was detected at all sites.
Mean PFNA concentrations for the four sites along the Conasauga River were 32.8, 201.6, 369,
and 284 ng/L for the upstream, LAS site, and two downstream locations, respectively (n = 5 at
each site), with a pattern of increasing concentration with distance downstream of the LAS
before a decrease in concentration at the final site. Authors suggested sorption to sediments,
particularly organic carbon, as a possible reason for the decrease in PFNA concentration at the
final site. At the single freshwater site in the Altamaha River, PFNA was detected in one of three
samples at a concentration of 0.6 ng/L. The range of PFNA concentrations in ponds and streams
near Dalton were: 11.1-12.2, 40.6-41.0, 4.8-6.3, and 2.1-2.5 ng/L for sites 1, 2, 3, and 4,
respectively (n = 2 at each site).
Three studies evaluated surface water potentially impacted by wastewater (Boone et al., 2019;
Subedi et al., 2015; Appleman et al., 2014). Boone et al. (2019) evaluated 17 PFAS in source and
treated waters collected in 2010-2012. Authors attempted to select locations with known or
suspected sources of wastewater in the source water, but ultimately the site selection was
dependent upon the willingness of DWTPs to participate. The study did not differentiate which
locations had known or suspected sources. Of the 22 surface water sources evaluated (16 river
and 6 lake/reservoir), PFNA was detected in all samples (n = 22), with a mean concentration of
2.93 ng/L. Subedi et al. (2015) collected 28 lake water samples from 3 sampling events in
August-September 2012 and four sampling events in May-September 2013 from Skaneateles
Lake. Sites were selected to be along the shoreline of homes that use an enhanced treatment unit
for onsite wastewater treatment. Wastewater effluents were identified as a source of
contamination to the lake. PFNA was detected in 57% of samples with mean and median
concentrations of 0.36 and 0.26 ng/L, respectively. Appleman et al. (2014) assessed source water
from 11 utilities in Alaska, Alabama, Colorado, Nevada, New Jersey, Ohio, Oklahoma, and
Wisconsin from August 2011 to May 2012, the majority of which were selected because they
were either known from previous monitoring or expected to contain detectable PFAS because
they were impacted by upstream wastewater effluent discharge. Authors evaluated the utilities
and their effectiveness for removing PFAS. The study did not report an average concentration for
PFNA, but PFNA was detected in 14 of 25 samples (from 7 of 11 utilities) with a maximum
concentration of 5.7 ng/L.
In two studies, surface water samples were collected from locations with potential sources of
PFAS that were not related to AFFF use (Procopio et al., 2017; Zhang et al., 2016). Procopio et
al. (2017) evaluated samples collected between September 2011 and July 2014 from the
Metedeconk River. Eight sampling events were conducted as part of a source trackdown study to
identify potential sources of PFAS contamination after elevated PFOA levels were discovered at
C-16
-------
FINAL
2024
a raw surface water intake of the Brick Township Municipal Utilities Authority. In all 56
samples, PFNA was below the minimum laboratory RL (5 ng/L). Zhang et al. (2016) conducted
analyses to determine major sources of surface water PFAS contamination. Freshwater sample
collection sites included 22 sites in the state of Rhode Island (sampled June 2014) and 6 sites in
the New York Metropolitan Area (sampled October 2014). Surface water sites were creeks and
rivers in urban and rural locations. PFNA was detected at all sites with a mean of 1.914 ng/L (n =
28). Authors identified potential PFAS sources at these sites to be metal coating plating; paint,
coating, adhesive manufacturing; paper manufacturing; petroleum coal products manufacturing;
printing activity; printing ink manufacturing; semiconductor manufacturing; sewage treatment;
textile mills; waste management including landfills, and airports.
Of the remaining four U.S.-based studies (Boone et al., 2014; Quinones and Snyder, 2009)|(Pan
et al., 2018; Kim and Kannan, 2007), Boone et al. (2014) analyzed PFCs in water samples
collected from the surface of the Mississippi River at a low flow level (2.95 ft) in September
2010 and a high flow level (8.32 ft) in June 2009. PFNA levels were 1.800 and 3.30 ng/L at low
and high flow levels, respectively. In Quinones and Snyder (2009), surface water samples from
the Boulder Basin, Hoover Dam, and the lower Colorado River were collected in 2008. Mean
PFNA levels at all sites were below the method RL (1.0 ng/L). Kim and Kannan (2007) sampled
two urban lakes in Albany, New York during five sampling trips from February-November
2006. The lakes, Washington Park Lake and Rensselaer Lake, are located in downtown Albany
and receive surface runoff from nearby roadways and residential areas during stormwater runoff.
PFNA was detected in Washington Park Lake (n = 6) at mean and median concentrations of 1.99
and 2.14 ng/L, respectively, and in Rensselaer Lake (n = 5) at mean and median concentrations
of 1.35 and 1.45 ng/L, respectively. Overall, PFNA was detected in 81.8% of the 11 total
samples. Finally, in a multicontinental study, Pan et al. (2018) assessed surface water samples
from several countries including the United States (Delaware River), United Kingdom (Thames
River), Germany and the Netherlands (Rhine River), and Sweden (Malaren Lake). Twelve
samples were collected in September-December 2016 along the Delaware River that spanned
seven cities—Trenton, Bristol, Philadelphia, Chester, Delaware, Smyrna, and Frederica. Authors
noted that all sampling sites were along the main stream of the rivers and not proximate to
known point sources of any fluorochemical facilities. PFNA was detected in all samples from the
Delaware River with a mean concentration of 2.51 ng/L and were similar to levels found in the
Thames River.
The EPA also identified studies from Canada and Europe reporting the occurrence of PFNA in
surface water, which are briefly summarized below. Detailed results from each study are
presented in Table C-2. Most Canadian and European studies reported PFNA detected in at least
one surface water sample site in each study. Concentrations of PFNA in creek and river samples
measured throughout Canada ranged from <125 pg/L to 3,000 pg/L (D'Eon J et al., 2009 as cited
in NCBI, 2022). Also, PFNA concentrations ranged from 0.80 ng/L to 2.4 ng/L in surface water
samples collected from Highland Creek watershed, Canada in 2010 (Meyer et al., 2011 as cited
in NCBI, 2022). Concentrations of PFNA in lake water samples collected from four lakes on
Cornwallis Island, Canada from 2003 to 2005 ranged from not detected to 6.1 ng/L (Stock et al.,
2007 as cited in NCBI, 2022).
Several studies in Europe sampled surface water from sites in proximity to fluoropolymer
facilities or in locations with current or past AFFF usage. Four studies investigated surface water
C-17
-------
FINAL
2024
in proximity to fluoropolymer facilities. One study in France (Boiteux et al., 2017) reported no
detections, but the other three, from France, the Netherlands, and Italy, respectively reported at
least one site or sampling event with detections from 0.49 ng/L to 2,637 ng/L (Bach et al., 2017;
Gebbink et al., 2017; Valsecchi et al., 2015). In the study by Bach et al. (2017), the maximum
level detected of 2,637 ng/L was from a sampling area in a river downstream from two facilities
that produce fluoropolymers. Four additional studies included sampling locations potentially
impacted by current or past AFFF usage (Mussabek et al., 2019; Skaar et al., 2019; Gobelius et
al., 2018; Dauchy et al., 2017). The reported PFNA detection frequencies ranged from 0% to
89% with varying maximum levels from 1.81 ng/L to 26 ng/L.
Many European and Canadian studies sampled sites where there was no known PFAS
manufacturing or use of PFAS-heavy materials (e.g., AFFF). In some cases, PFNA was not
detected or concentrations were not reported (Barreca et al., 2020; Yeung et al., 2017; Lescord et
al., 2015; Jurado-Sanchez et al., 2013; Villaverde-de-Saa et al., 2012; Moller et al., 2010), and in
other cases there was 100% (or near 100%) detection frequency of PFNA (Zhao et al., 2015;
Eschauzier et al., 2012; Kovarova et al., 2012; Labadie and Chevreuil, 2011; Ahrens et al.,
2009a). Maximum PFNA concentrations reported among all PFNA detections in these studies
ranged from low (e.g., 0.18 ng/L in Zhao et al. (2015)) to very high (e.g., 8,100 ng/L in
Kovarova et al. (2012)). Several studies with high maximum detections of PFNA used sampling
locations near or potentially near wastewater treatment plants or other industrial activity
(Wilkinson et al., 2017; Lorenzo et al., 2015; Boiteux et al., 2012; Llorca et al., 2012). The
remaining studies did not report details on the area surrounding sampling locations and how
nearby activities may have impacted the results (Ciofi et al., 2018; Munoz et al., 2018; Loos et
al., 2017; Shafique et al., 2017; Eriksson et al., 2013; Veillette et al., 2012; Ahrens et al., 2009b;
Rostkowski et al., 2009; Ericson et al., 2008b); the detected concentrations in these studies
ranged from 0.057 ng/L to 9.89 ng/L.
C-18
-------
FINAL
2024
Table C-2. Summary of Studies Reporting the Occurrence of PFNA in Surface Water
Study
Location
Site Details
Results
United States
Galloway et al. (2020)
United States (Ohio; West
Virginia)
Rivers and tributaries near a fluoropolymer facility
sampled throughout three trips on June, July, and
December 2016. In June 2016, samples were collected
on a 188 km stretch of the Ohio River, from 130 km
downstream to 58 km upstream of the facility, and
tributaries that pass near known PFAS-containing
landfills. In July 2016, samples were collected from
lakes, rivers, and creeks to the north and northeast of
the facility as far as 16 km downwind. The December
2016 trip expanded the collection radius to more than
48 km downwind to the north and northeast of the
facility.
June 2016:
n = 26; DFa 92%*
*PFNA was detected but below the LOQ in
24 samples, and ND in 2 samples
July 2016:
n = 25; DFa 0%
December 2016:
n = 40; DFa 80%*, range =
-------
FINAL
2024
Study
Location
Site Details
Results
not specifically controlled for in the site selection
process; active remedies had not been applied at any
of the sites selected. Approximately ten samples were
collected between March and September 2014 at each
site; sites were grouped according to volume of AFFF
release—low-volume typically had a single AFFF
release, medium-volume had one to five releases, and
high-volume had multiple releases. Surface water
sample locations included engineered storm water
channels, engineered AFFF ponds, and natural
streams.
n = 2, DF 0%
(median reporting limit =17 ng/L)
*Median calculated using quantified
detections
*Non-detects were substituted with Vi the
reporting limit
Post et al. (2013)
United States (New Jersey)
Raw water collected from 12 public drinking water
system intakes between August 2009 and February
2010 from 6 rivers and 6 reservoirs. Sites were chosen
to represent NJ geographically and included two
reservoir sites near a civil-military airport with
possible AFFF use.
Overall: n = 12, DF 33%, range = ND-19
ng/L
Rivers: n = 6, DF 0%
Reservoirs: n = 6, DF 67%, range = <5-19
ng/L
(minimum reporting limit = 5 ng/L)
Nakayama et al. (2010)
United States (Illinois; Iowa;
Minnesota; Missouri;
Wisconsin)
Eighty-eight sampling sites collected between March
and August 2008 from tributaries and streams in the
Upper Mississippi River Basin and a portion of the
Missouri River Basin. Samples were collected by the
Minnesota Pollution Agency, Wisconsin Department
of Natural Resources, Illinois Environmental
Protection Agency, and U.S. EPA Region 7 Water
Quality Monitoring Team. Each agency selected
sampling sites with the intention of providing
preliminary PFC data to the individual regions.
Sampling sites included locations adjacent to chemical
manufacturers, downstream of WWTPs receiving
waste from those types of manufacturers, and near an
airport with historic use of firefighting foams.
n = 173, DF 87%), median (range) = 0.71
(ND-72.9) ng/L
(LOD = 0.02 ng/L)
*ND data points were substituted with
LOD/sqrt(2) = 0.014 ng/L
Nakayama et al. (2007)
United States (North Carolina)
Eighty sampling sites in river basin during spring
2006. The sites were selected to reflect water quality
throughout the basin. Possible sources of PFCs include
use of firefighting foam from Fort Bragg and Pope Air
Force Base, metal-plating facilities, textile, and paper
production, and other industries.
n = 100, DF 89.9%o, mean, GM, median
(range) = 33.6, 9.73, 5.70 (
-------
FINAL
2024
Study
Location
Site Details
Results
agricultural fields between 1995 and 2008. Biosolids
obtained from local municipal WWTP where sources
discharging to the WWTP included facilities involved
in the production and use of fluoropolymers,
fluorocarbon fibers, polymers, polymer films, and
resins.
Lasieretal. (2011)
United States (Georgia)
Upstream (sites 1 and 2) and downstream (sites 3-8)
of a land application site were sampled along the
Conasauga, Oostanaula, and Coosa Rivers during
summer 2008, where effluents from carpet
manufacturers (suspected of producing wastewaters
containing perfluorinated chemicals) are processed at
a WWTP and the treated WWTP effluent is sprayed
onto the site. Additionally, site 2 was downstream of a
local airport and site 4 was downstream of a
manufacturing facility for latex and polyurethane
backing material.
Upstream:
Sites 1 and 2: DF 0%
Downstream:
Site 3: DF NR, mean = 35 ng/L
Site 4: DF NR, mean = 44 ng/L
Site 5: DF NR, mean =17 ng/L
Site 6: DF NR, mean = 21 ng/L
Site 7: DF NR, mean = 20 ng/L
Site 8: DF NR, mean = 21 ng/L
(MDL = 0.005 ng/g, LOQ = 0.010 ng/g)
*Half of the MDL was used when measured
concentrations were below the MDL
¦"Concentrations measured in triplicate
samples
Konwick et al. (2008)
United States (Georgia)
Samples collected in March 2006 from the Conasauga
River near Dalton, a major carpet manufacturing city,
where there is high use of PFAAs in the carpet
industry. Four sites on the Conasauga River were
sampled: one location upstream, two downstream, and
one at the site of a land application system where
treated wastewater (approximately 87% industrial
source) is sprayed.
One freshwater site on the Altamaha River, a
reference site away from Dalton, was sampled in
January 2005.
Four ponds and streams in Dalton were also sampled
in January 2005.
Conasauga River:
Site 1 (upstream): n = 5, DFa 100%, mean
(range) = 32.8 (12.3-75.4) mg/L
Site 2 (at site): n = 5, DFa 100%), mean
(range) = 201.6 (136-248) mg/L
Site 3 (downstream): n = 5, DFa 100%o,
mean (range) = 369 (280^156) mg/L
Site 4 (downstream): n = 5, DFa 100%,
mean (range) = 284 (190-366) mg/L
Altamaha River:
Site 1 (freshwater): n = 3, DFa33%o, range =
<0.6-0.6 ng/L
Dalton streams/ponds:
Site 1: n = 2, DFa 100%, range = 11.1-12.2
ng/L
Site 2: n = 2, DFa 100%o, range = 40.6^11.0
ng/L
Site 3: n = 2, DFa 100%, range = 4.8-6.3
ng/L
C-21
-------
FINAL
2024
Study
Location
Site Details
Results
Site 4: n = 2, DFa 100%, range = 2.1-2.5
ng/L
(LOD = 0.6 ng/L)
Boone et al. (2019)
United States (unspecified)
Twenty-two surface waters (16 rivers and 6
lakes/reservoirs) used as source waters for 22 DWTPs,
collected in 2010-2012; some locations with known
or suspected sources of wastewater in the source
water, but study did not differentiate which locations
had known or suspected sources.
n = 22, DFa 100%, meana (range) = 2.93
(0.117—41.4) ng/L
(LCMRL = 0.094 ng/L)
Subedi et al. (2015)
United States (New York)
Lake water along the shoreline of residences that use
an enhanced treatment unit for onsite wastewater
treatment; samples were collected -40 ft from the
lakeshore about 2 ft below surface. Sampling occurred
August-September 2012 (three sampling events) and
May-September 2013 (four sampling events).
Wastewater effluents identified as source of
contamination.
n = 28, DFa 57%), mean, median (range) =
0.36, 0.26 (ND-1.21) ng/L
(LOQ = 0.2 ng/L)
*Data points
-------
FINAL
2024
Study
Location
Site Details
Results
Rhode Island: n = 22, DFa 100%, meana
(range) = 1.868 (0.104-13.986) ng/L
Urban: n = 10, DFa 100%, meana (range) =
2.576 (0.308-13.986) ng/L
Rural: n = 12, DFa 100%), meana (range) =
1.278 (0.104-7.235) ng/L
New York Metropolitan Area (all urban sites):
n = 6, DFa 100%o, meana (range) = 2.084
(0.151-6.658) ng/L
(LOD = 0.04 ng/L)
Boone et al. (2014)
United States (New Orleans,
Louisiana)
Surface samples from the Mississippi River collected
in June 2009 when the river was at a high flow level
(8.32 ft) and in September 2010 when the river was at
a low flow level (2.95 ft).
Low flow (2.95 ft): mean based on a primary
and duplicate sample = 1.800 ng/L
High flow (8.32 ft): mean based on four
replicates = 3.30 ng/L
(DL = 0.047 ng/L, LCMRL = 0.110 ng/L)
Quinones and Snyder (2009)
United States (Arizona; Nevada)
Samples collected in 2008 from three sites in Boulder
Basin, one site in Hoover Dam, and two sites from the
lower Colorado River. PFC occurrence had not been
previously determined or reported for these sites.
n = 40, DF NR
(Method RL= 1.0 ng/L)
*Mean values at all sites were Method reporting limit
Kim and Kannan (2007)
United States (Albany, New
York)
Samples collected from two urban lakes—Washington
Park and Rensselaer Lake—during five sampling trips
from February-November 2006. Both lakes are
located in downtown Albany and receive surface
runoff from nearby roadways and residential areas
during stormwater runoff.
Total: n = 11, DF 81.8%, mean, median
(range) = 1.70, 1.63 (ND-3.51)ng/L
Washington Park Lake: n = 6, DF NR, mean,
median (range) = 1.99,2.14 (0%
Resolute: n = 5, DF >0%
¦"Concentrations and summary statistics were
not reported in tables or text; Figure 2 shows
non-zero concentrations of PFNA in Meretta
C-23
-------
FINAL
2024
Study
Location
Site Details
Results
approximately 0.5 km downstream from a local
airport, where wastewater from the airport and
military base was discharged with little treatment into
the Meretta catchment from 1949 to 1998.
and Resolute. Concentrations in Char, Small,
North, and 9 Mile lakes are unclear
(method detection limit = 0.0018 ng/L)
Yeung et al. (2017)
Canada (Ontario)
River water samples were collected from Mimico
Creek and Rouge River in November 2014 and
analyzed using two methods. Reference method used
ultra performance liquid chromatograph and a newly
developed method used ultra performance
convergence chromatograph for separation.
Results presented as reference method, new
method:
Mimico: n = 1, point = <2, <5 ng/L
Rouge: n = 1, point = <2, <5 ng/L
(MLOQ for reference method not reported;
MLOQ for new method reported in Table S3
as 200 ng/L but based on Table S7, this should
likely be 2 ng/L)
*Results in Table S7 are presented in ng/L but
based on a comparison to Figure 4, Table S7
should be in ng/mL
Veillette et al. (2012)
Canada (Ellesmere Island,
Nunavut)
Lake catchment area located on the northwest coast of
the island. Surface water was collected from the center
of the lake, the littoral zone (30 m from the delta), the
delta, and lake inflow and outflow in July 2007, May
2008, and August 2008. Samples were collected at
depths of 2 m (underneath the ice cover), 10 m (the
bottom of the mixed layer), and 32 m (in the
monimolimnion).
n = 11, DFa 100%, mean (range) = 0.118
(0.057-0.192) ng/L
(method detection limit = 0.009 ng/L)
Europe
Bach et al. (2017)
France (southern)
Grab water samples were collected from six locations
along the shore of a river in April, July, October, and
December 2013. The river selected for the study
receives effluent at three points along the river from
an industrial site where two facilities produce
fluoropolymers. The first facility has been active since
the 1960s, with production including PTFE synthesis
from the beginning of the 1960s to 1985 with PFOA
as a processing aid; more recently, PVDF has been
synthesized since the early 1970s with fluorotelomer
sulfonic acid (6:2 FTSA) or PFNA as a processing aid.
The second facility, established in 2002, produced
fluoropolymers with PFOA as a processing aid until
2008 when it was replaced with PFHxA. Samples
were collected starting ~1.3 km upstream from the
industrial site and covered ~15 km of the river.
Upstream:
Sampling point #1: n = 1, point = <4 ng/L
for April, July, October, and December
2013
Downstream:
Sampling point #2: n = 1, point = 209,
2,637,15, and 12 ng/L for April, July,
October, and December 2013
Sampling point #3: n = 1, point = <4, <4,
<4, and <4 ng/L for April, July, October,
and December 2013
Sampling point #4: n = 1, point = 9, 42, <4,
and <4 ng/L for April, July, October, and
December 2013
C-24
-------
FINAL
2024
Study
Location
Site Details
Results
Samples (point #1 to #6) were collected from
upstream to downstream.
Sampling point #5: n = 1, point = <4, <4,
<4, and <4 ng/L for April, July, October,
and December 2013
Sampling point #6: n = 1, point = <4, 87, <4,
and <4 ng/L for April, July, October, and
December 2013
(LOQ = 4 ng/L)
Boiteux et al. (2017)
France (northern)
Grab water samples were collected from seven
locations along a river in May, July, October, and
December 2013. The river selected for the study
receives wastewater from an industrial WWTP that
treats raw sewage coming from a facility that
manufactures fluorotelomer-based products and side-
chain-fluorinated polymers used in firefighting foams
and stain repellents. Samples were collected starting
~1.2 km upstream of the WWTP discharge and
encompassed ~65 km of the river. Samples were
collected from upstream to downstream.
Upstream:
Sampling point #1: n = 1, DF 0% for May,
July, October, and December 2013
Downstream:
Sampling points #3,4, 5,7, 9,11: n= 1,DF
0% for May, July, October, and December
2013
(LOQ = 4 ng/L)
*DF represents frequency of quantification
Gebbink et al. (2017)
The Netherlands (Dordecht)
River water samples collected in October 2016 at sites
downstream (R1-R13) and upstream (R14-R16) of
the Dordrecht fluorochemical production plant.
Samples (R17-R18) were also collected from
different waterbodies at control sites.
Control sites: n = 2, DFa (frequency of
quantification) 100%, meana (range) = 0.9
(0.8-1.0) ng/L
Upstream: n = 3, DFa (frequency of
quantification) 100%, meana (range) = 0.75
(0.54-0.92) ng/L
Downstream: n = 13, DFa (frequency of
quantification) 100%, meana (range) = 0.67
(0.49-1.0) ng/L
(minimum quantification level = 0.03 ng/L)
Valsecchi et al. (2015)
Italy (River Basins Po, Brenta,
Adige, Tevere, and Arno)
Five river basins were sampled between 2008 and
2013. Two river basins (Po and Brenta) receive
discharges from two chemical plants that produce
fluorinated polymers and intermediates; two river
basins (Tevere and Adige) are not impacted by
relevant industrial activities; and one river basin
(Arno) has textile and tannery districts located along
parts of the river. In total, 20 rivers were sampled at
their basin closure stations. Rivers Arno, Tevere, and
Po were also sampled along the course of the river.
Po: n = 105, DFa 64%, range =
-------
FINAL
2024
Study
Location
Site Details
Results
Skaaretal. (2019)
Norway (Ny-Alesund; Lake
Linnevatnet area)
Freshwater samples were collected from Ny-Alesund
(research facility) in June 2016 and from the Lake
Linnevatnet area (background site) in March 2014 and
from April to June 2015. Surface water in Ny-Alesund
was contaminated by a local firefighting training site.
Lake Linnevatnet receives water from meltwater of
the adjacent glaciers and has few potential pollution
sources.
Ny-Alesund:
Background: n = 7, DF 0%
Contaminated: n = 3, DFb 67%, range =
<0.02-1.81 ng/L
Lake Linnevatnet:
Background: n = 20, DFb 65%, range =
<0.03-0.16 ng/L
(LOD = 0.021 ng/L; LOQ = 0.085 ng/L)
Mussabek et al. (2019)
Sweden (Lulea)
Samples from a man-made lake and pond
approximately 500 m southwest from a firefighting
training facility at the Norrbotten Air Force Wing
collected in October 2015. The training facility has
been active since 1941 and has used PFAS-containing
AFFFs in the last decades. The lake and pond lie
above a groundwater reservoir with high permeable
soil and were selected because they are isolated water
bodies receiving PFAS contamination and can
potentially impact groundwater.
Lake: n = 2, DF NR, mean = <0.5 ng/L
Pond: n = 2, DF NR, mean = <0.5 ng/L
(LOD = 0.5 ng/L)
Gobelius et al. (2018)
Sweden (national)
Sampling conducted between May and August 2015,
with the majority in July and a few samples in
September and November 2015. Samples were
collected in 21 regional counties by the County
Administration Boards. Sampling locations selected
based on potential vicinity of PFAS hot spots (i.e., fire
training sites, unspecific industry, sewage treatment
plant effluent, landfill/waste disposal, skiing, and
urban areas) and/or importance as a drinking water
source. Sample numbers varied for each county and
sampling sites were spread unevenly across Sweden.
Surface water samples collected approximately 10 cm
below the water surface.
n = 281, DFa 89%), range = <0.08-26 ng/L
(MDL = 0.084 ng/L)
*Two types of water (i.e., surface water and
recipient water [surface water]) included
Dauchy et al. (2017)
France (unspecified)
Samples collected in the vicinity of three sites (B, C,
D) where fluorosurfactant-based foams are or were
being heavily used. Site B is an international civilian
airport built in 1974. The exact location of the training
area, frequency of training sessions, and history of
firefighting training activities are unknown. In
November 2014, surface water samples were collected
in the only river running alongside the airport.
Site B: n = 5, DF 0%
Site C: n = 9, DF 0%
Site D: n = 2, DF 0%
(LOQ = 4 ng/L)
C-26
-------
FINAL
2024
Study
Location
Site Details
Results
Downstream from the airport, this river joins two
other rivers, which were also sampled.
Site C is a military airport, with the exact location of
the training area, frequency of the training sessions,
and history of the firefighting training activities
unknown. In April 2014, surface water samples were
collected in several rivers surrounding the military
base.
Site D is a training center for firefighters. From 1969
to 1984, the site was an oil refinery. Starting in 1987,
the site became a training area for firefighters, with
exercises carried out directly on the soil. From the
1990s, some exercise areas were covered with
concrete. In November 2014, two surface water
samples were collected from the river receiving
effluent from a WWTP at the site, one upstream and
one downstream of the discharge pipe.
Ciofietal. (2018)
Italy (Tuscany)
Surface water samples were collected at 13 locations.
Sampling year was not reported.
SW-1: Arno river before the "Canale Maestro della
Chiana" (Arezzo), receiving agricultural runoff and
untreated urban wastewater
SW-2: Arno river after the "Canale Maestro della
Chiana" (Arezzo)
SW-3: Arno river before entering in Florence
SW-4: Arno river after the discharge of the Florence
WWTP
SW-5: Arno river after the confluence of the Bisenzio
river
SW-6: Arno river after the city of Empoli (Florence)
SW-7: Arno river after receiving the WWTP effluent
from the leather industrial district of Santa Croce
(Pisa)
SW-8: Arno river in the proximity of the mouth (Pisa)
SW-9: Bisenzio river before the confluence with Arno
river (Florence)
SW-10: Serchio river in the proximity of the mouth
(Lucca)
SW-11: East area of the coastal lake "Massaciuccoli"
(Lucca)
SW-1: n = 1, point = <0.23 ng/L
SW-2: n = 1, point = <0.21 ng/L
SW-3: n = 1, point = <0.25 ng/L
SW-4: n = 1, point = <0.21 ng/L
SW-5: n = 1, point = <0.22 ng/L
SW-6: n = 1, point = <0.22 ng/L
SW-7: n = 1, point = <0.20 ng/L
SW-8: n = 1, point = <0.25 ng/L
SW-9: n = 1, point = 2.7 ng/L
SW-10: n = 1, point = <0.23 ng/L
SW-11: n = 1, point = <0.19 ng/L
SW-12: n = 1, point = <0.70 ng/L
SW-13: n = 1, point = <0.27 ng/L
(MDL = 0.19-0.27 ng/L; MQL = 0.70 ng/L)
*MDL/MQL varied by sample. MDL
provided for 11 of 13 samples; minimum
quantitation level provided for 1 of 13 samples
C-27
-------
FINAL
2024
Study
Location
Site Details
Results
SW-12: West area of the coastal lake "Massaciuccoli"
(Lucca)
SW-13: Central area of the artificial lake "Bilancino"
(Florence)
Munoz et al. (2018)
France (Marnay-sur-Sein;
Bougival; Triel-sur-Seine)
Surface water along the Seine River was collected
during four sampling campaigns between September
2011 and December 2012, each conducted in a
different season. For each campaign, two to four
samples were collected over a one-month period.
Three sampling sites were investigated: Marnay-sur-
Sein, located 200 km upstream from Paris, was
selected as a reference site, non-affected by the
Greater Paris region; Bougival, situated 40 km
downstream from Paris, was chosen to investigate the
impact of Greater Paris on PFAS levels; Triel-sur-
Seine, another 40 km further downstream, was
selected to assess the global influence of the Paris
urban area, including other inputs such as WWTPs.
n = 36, DF 97%, range = ND-2.0 ng/L
(LOD = 0.02-0.3 ng/L for all PFAS)
*PFNA concentrations were not reported in
text or table by site but relative abundance by
site is available in Figure 1
Wilkinson et al. (2017)
England (Greater London and
southern England)
Three rivers selected due to their accessibility,
receiving only STW effluent outfalls (i.e., no
confluence with another major river in the study area)
and accessibility of river headwater sampling for the
Hogsmill and Blackwater Rivers. Each river sampled
received inputs from at least one STW. Three STWs
discharge into the Blackwater River, one STW
discharges into the Hogsmill River, and one STW
discharges to Chertsey Bourne River. Headwaters
were evaluated for the Hogswill and Blackwater
Rivers. Water samples were collected 50 m upstream
and 250 m and 1,000 m downstream from STW
effluent outfalls. In total, samples were collected on 3-
4 separate occasions from 23 sites. Sampling dates
were not reported.
Headwaters: n = 6, DF NR, mean = 2.75 ng/L
Upstream: n = 19, DF NR, mean = 16.7 ng/L
Downstream 250 m: n = 19, DF NR, mean =
32.5 ng/L
Downstream 1,000 m: n = 19, DF NR, mean =
23.9 ng/L
(LOD = 0.23 ng/L; LOQ = 0.75 ng/L)
Lorenzo et al. (2015)
Spain (Guadalquivir River
Basin; Ebro River Basin)
Surface water was collected from the Guadalquivir
River and its main tributaries and from the Ebro River
and its main tributaries in October 2010. Guadalquivir
sampling locations included downstream of WWTPs,
near industrial areas, near a military camp, or through
major cities; Ebro sampling locations included nearby
ski resorts and downstream of WWTP and industrial
areas.
Guadalquivir: n = 24, DF 8%, mean (range) =
5.1 (6.8-116.1) ng/L
Ebro: n = 24, DF 8%, mean (range) = 0.5
(4.8-7.9) ng/L
*Minimum reported is the lowest amount
quantified
C-28
-------
FINAL
2024
Study
Location
Site Details
Results
*Mean was calculated with not detected
concentrations as zeros
(MQL = 0.4 ng/L)
Zhaoetal. (2015)
Germany (Elbe River)
Four sampling campaigns conducted in February,
April, August, and October 2011 to represent the four
seasons. Freshwater samples (sites E619 to E689 with
salinity <1 PSU) were collected at nine locations in
the river Elbe. Some sampling sites were near
Hamburg city and experienced occasional discharge of
wastewater from industrial plants,
February: n = 8, DFa 100%, meana (range) =
0.23 (0.16-0.43) ng/L
April: n = 9, DFa 100%), meana (range) = 0.18
(0.16-0.23) ng/L
August: n = 9, DFa 100%o, meana (range) =
0.24 (0.19-0.36) ng/L
October: n = 8, DFa 100%o, meana (range) =
0.15 (0.12-0.18) ng/L
(MDL = 0.03 ng/L)
Boiteux et al. (2012)
France (national)
Raw water from rivers used as source water for
DWTPs. Sites distributed across 100 French
department to represent -20% of the national water
supply flow; samples collected during two sampling
campaigns in July-September 2009 (first campaign)
and June 2010 (second campaign - focused on sites
from first sampling campaign that had PFC levels
>LOQ). Some sites possibly affected by
commercial/industrial releases.
Overall: n = 135, DF (frequency of
quantitation) 2%o, maximum = 52 ng/L
1st Sampling Campaign
n = 99, DF 5%, mean, median (maximum) =
<1, <1 (4) ng/L
2nd Sampling Campaign
n = 36, results not reported
(LOD =1.3 ng/L, LOQ = 4 ng/L)
Eschauzier et al. (2012)
The Netherlands (Amsterdam)
Intake water from the Lek canal (n = 2) was collected
in January and September 2010 to determine the
behavior of PFAAs during the drinking water
treatment processes. The Lek canal, a tributary of the
river Rhine, is the source of drinking water for the city
of Amsterdam and is downstream of an industrial
point source in the German part of the Lower Rhine.
n = 2, DFa 100%o, mean (range) = 0.6 (0.5-0.8)
ng/L
(LOQ = 0.24 ng/L)
Kovarova et al. (2012)
Czech Republic (Brno)
Seven locations in the Svitava and Svratka Rivers
upstream and downstream of Brno, a city with highly
developed chemical, engineering, textile, and food-
processing industries. A sampler was installed at each
site for 30 days twice a year (May and September
2008). Due to technical problems, samples were
produced from only four of seven sites in May and
from five of seven sites in September.
May: n = 4, DFa 100%, meana (range) = 2,735
(240-5,700) ng/L
September: n = 5, DFa 100%, meana (range) =
2,690 (170-8,100) ng/L
(LOD not reported)
Llorca et al. (2012)
Germany (Hesse), Spain
(national)
Forty-eight surface river waters were sampled in
2010-2012 (24 from Spain and 24 from Germany).
Germany: n = 24, DF 0%
C-29
-------
FINAL
2024
Study
Location
Site Details
Results
Samples from Germany were collected from
agriculturally or industrially influenced streams.
Samples from Spain were collected from the Xuquer
River Basin, Llobregat River Basin, and Ebro River
Basin.
Spain: n = 24, DF 13%, mean, median (range)
= 26, 20 (0.03-52) ng/L
(MLOQ = 0.03 ng/L)
*MLOQ reported above is from Table 3;
Table 2 reports MLOD =1.9 ng/L and MLOQ
= 6.3 ng/L
Labadie and Chevreuil (2011)
France (Paris)
Samples collected weekly in January-May 2010 in an
urban stretch of the River Seine at the Austerlitz Quay,
downtown Paris during a flood cycle. The sampling
station is under the influence of two major WWTPs
and two major combined sewer overflow outfalls.
n = 16, DF 100%, mean, median (range) = 0.5,
0.5 (0.1-1.2) ng/L
(LOQ = 0.17 ng/L)
Molleretal. (2010)
Germany (Rhine River
watershed)
Raw freshwater samples collected in September-
October 2008 along the River Rhine (stations 1-36)
and major tributaries of the River Rhine (e.g., Rivers
Neckar, Main, Rhur, stations 37^18). Along the River
Rhine, samples were taken upstream and downstream
of Leverkusen, where effluent of a WWTP treating
industrial wastewater was discharged. All samples
taken at a water depth <1 m.
n = 48, DF NR, authors noted that PFNA was
quantified but results not provided
(LOD = 0.014-1.60 ng/L for all PFAS)
Rostkowski et al. (2009)
Poland (national)
Inland surface water samples were collected at 12
locations in the southern part of Poland and 14
locations in the northern part of Poland in October and
December 2004. Inland surface waters included rivers,
lakes, and streams. The northern locations flowed
through forested, agricultural, and rural areas; these
areas are considered unpolluted with industrial
chemicals. Some southern locations were near
chemical industrial activities.
North: DFa 51%, range = <0.1-0.6 ng/L
South: n = 11, DFa 36%, range = <0.1-
0.6 ng/L
(LOQ = 0.1-0.5 ng/L)
Barreca et al. (2020)
Italy (Lombardia Region)
Fifty-two surface water sampling stations (rivers and
streams) throughout the region. Samples collected in
2018.
n = 286, DFa 6%
(LOQ = 1 ng/L)
Loos et al. (2017)
Austria, Bulgaria, Croatia,
Moldova, Romania, Serbia,
Slovakia (Danube River and
tributaries)
Samples were collected in August-September 2013
from 68 sites along a 2,581 km-stretch of the Danube
River, with 14 of the sites in the mouths of tributaries
or side arms. Three additional samples were also
collected between the Iskar and Olt tributaries, the Olt
River, and between the Siret and Prut tributaries. The
investigated tributary rivers were the Morava
(Austria/Slovakia), the Vah (Slovakia), the Drava
n = 71, DF 79%), mean, median (range) = 1.2,
1.1 (
-------
FINAL
2024
Study
Location
Site Details
Results
(Croatia), the Tisa (Serbia), the Sava (Serbia), the
Velika Morava (Serbia), the Timok (Serbia/Bulgaria),
the Iskar (Bulgaria), the Olt (Romania), the Jantra
(Bulgaria), the Russenski Lom (Bulgaria), the Arges
(Romania), the Siret (Romania) and the Prut
(Romania/Moldavia). Some sampling locations were
downstream of major cities.
Shafique et al. (2017)
Germany (River Elster; River
PleiBe; River Saale; and River
Elbe)
Surface water samples were collected from the River
Elster, River PleiBe, River Saale, and River Elbe at
the start of 2015.
Elster: n = 4, DF NR, mean = 0.96 ng/L
PleiBe: n = 2, DF NR, mean = 1.65 ng/L
Saale (Site A): n = 10, DF NR, mean = 6.57
ng/L
Saale (Site B): n = 10, DF NR, mean = 9.89
ng/L
Saale (Site C): n = 10, DF NR, mean = 0.83
ng/L
Elbe: n = 2, DF NR, mean = 0.91 ng/L
(MDL = 0.07 ng/L)
*Values extracted from SI, which provides a
more detailed breakdown of sites compared to
that reported in the main text (where Elster
and PleiBe sites were combined and Saale
sites were combined)
Eriksson et al. (2013)
Denmark (Faroe Islands)
Grab samples collected in April-May 2012 from
Lakes Leitisvatn, Havnardal, Kornvatn, and A
Myranar.
Leitisvatn: n = 1, point = 0.16 ng/L
Havnadal Lake: n = 1, point = 0.14 ng/L
Kornvatn Lake: n = 1, point = 0.22 ng/L
A Myranar: n = 1, point = 0.13 ng/L
(LOD = 0.028 ng/L)
Jurado-Sanchez et al. (2013)
Spain (southeast)
Raw water samples were collected from a reservoir
used as the source for tap water production and
analyzed using a newly developed analytical method.
Samples were collected in triplicate once a week in six
different months at the intake of two different
DWTPs.
River water samples were also collected. Authors did
not report sampling details or sampling year.
DWTP 1 intake: n = 3, DFa 0%
DWTP 2 intake: n = 3, DFa 0%
River water: n = NR, DF 0%
(LOD = 0.1 ng/L)
C-31
-------
FINAL
2024
Study
Location
Site Details
Results
Villaverde-de-Saa et al. (2012)
Spain (Santiago de Compostela;
Surface water samples were collected from Sar river in
n = 3, DF 0%
Pontevedra)
Santiago de Compostela on January 2011 and from
(LOD = 0.6 ng/L)
Lerez river in Pontevedra on March 2011.
Ahrens et al. (2009a)
Germany (Hamburg;
Nine samples collected from the river Elbe in
Hamburg:
Laurenburg)
Hamburg city (sites 16-18) and from Laurenburg to
Dissolved: n = 3, DFa 100%, mean (range) =
Hamburg (sites 19-24) in August 2006. Samples were
1.8 (1.7-2.0) ng/L
collected at a water depth of 1 m. Dissolved and
Particulate: n = 3, DFa 67%, mean (range) =
particulate phases were analyzed for each of the water
0.040 (ND-0.088) ng/L
samples.
Laurenburg to Hamburg:
Dissolved: n = 6, DFa 100%, mean (range) =
1.7 (0.6-2.1) ng/L
Particulate: n = 6, DFa 100%, mean (range)
= 0.044 (0.003-0.074) ng/L
(MDL = 0.045 ng/L for dissolved phase; 0.005
ng/L for particulate phase)
Ahrens et al. (2009b)
Germany (Elbe River)
Samples collected at 53 to 122 km (sites 1 to 9)
Site 1 (122 km): n = NR, DF NR, mean =
upstream of estuary mouth of Elbe River in June 2007.
0.73 ng/L
*Only locations with conductivity <1.5 mS/cm were
Site 2(118 km): n = NR, DF NR, mean =
assumed to be freshwater and extracted
1.1 ng/L
Site 3(115 km): n = NR, DF NR, mean =
0.7 ng/L
Site 4 (103 km): n = NR, DF NR, mean =
0.8 ng/L
Site 5 (90 km): n = NR, DF NR, mean =
0.6 ng/L
Site 6 (80 km): n = NR, DF NR, mean =
0.9 ng/L
Site 7 (74 km): n = NR, DF NR, mean =
0.7 ng/L
Site 8 (64 km): n = NR, DF NR, mean =
0.6 ng/L
Site 9 (53 km): n = NR, DF NR, mean =
0.7 ng/L
(MDL = 0.04 ng/L; MQL = 0.12 ng/L)
Ericson et al. (2008b)
Spain (Tarragona Province)
River water samples collected from the Ebro (at two
Ebro site 1: n = 1, point = 0.44 ng/L
points, Garcia and Mora), Francoli, and Cortiella
Ebro site 2: n = 1, point = 0.36 ng/L
Rivers in February 2007.
Francoli: n = 1, point = 0.64 ng/L
C-32
-------
FINAL
2024
Study
Location
Site Details
Results
Cortiella: n = 1, point = <0.42 ng/L
(LOD = 0.42 ng/L)
Multiple Continents
Pan et al. (2018)
United States (Delaware River)
Samples were collected from the Delaware River
between September and December 2016. Sampling
sites were not proximate to known point sources of
any fluorochemical facilities. Cities included Trenton,
Bristol, Philadelphia, Chester, Delaware, Smyrna, and
Frederica.
n = 12, DFa 100%, mean, median (range) =
2.51,2.36 (0.76^.81) ng/L
(MDL = 0.02 ng/L)
United Kingdom (Thames
River)
Samples were collected from the Thames River in
October 2016. Sampling sites were not proximate to
known point sources of any fluorochemical facilities.
Cities included Oxford and London.
n = 6, DFa 100%), mean, median (range) =
1.18,1.17 (0.77-1.71) ng/L
(MDL = 0.02 ng/L)
Germany and The Netherlands
(Rhine River)
Samples were collected from the Rhine River in
December 2016. Sampling sites were not proximate to
known point sources of any fluorochemical facilities.
Cities in Germany included Offenbach, Frankfurt,
Goarshausen, Rheinbrohl, Bonn, Cologne,
Leverkusen, Dormagen, Dusseldorf, Duisburg, Wesel,
and Emmerich. Cities in The Netherlands included
Arnhem, Lienden, Duurstede, Nijmegen, Wamel, and
Zaltbommel.
n = 20, DFa 100%o, mean, median (range) =
0.42,0.39 (0.09-0.67) ng/L
(MDL = 0.02 ng/L)
Sweden (Malaren Lake)
Samples were collected from Malaren Lake in
September 2016. Sampling sites were not proximate to
known point sources of any fluorochemical facilities.
Cities included Orebro and Stockholm.
n = 10, DFa 100%o, mean, median (range) =
0.54,0.54 (0.24-0.76) ng/L
(MDL = 0.02 ng/L)
Notes: AFFF = aqueous film-forming foam; DF = detection frequency; DWTP = drinking water treatment plant; ND = not detected; ng/L = nanogram per liter; PFAA =
perfluoroalkyl acid; PFAS = per- and polyfluoroalkyl substances; NR = not reported; LOD = limit of detection; LOQ = limit of quantification; LCMRL = lowest concentration
minimum reporting level; WWTP = wastewater treatment plant.
a The DF and/or mean was calculated using point data. Means were calculated only when DF = 100%.
C-33
-------
FINAL
2024
C.2. RSC for PFNA, Literature Search and Screening
Methodology
The EPA applies an RSC to the RfD when calculating an MCLG based on noncancer effects or
for carcinogens that are known to act through a nonlinear mode of action to account for the
fraction of an individual's total exposure allocated to drinking water (USEPA, 2000b). The EPA
emphasizes that the purpose of the RSC is to ensure that the level of a chemical allowed by a
criterion (e.g., the MCLG for drinking water) or multiple criteria, when combined with other
identified sources of exposure (e.g., diet, ambient and indoor air) common to the population of
concern, will not result in exposures that exceed the RfD. In other words, the RSC is the portion
of total daily exposure equal to the RfD that is attributed to drinking water ingestion (directly or
indirectly in beverages like coffee tea or soup, as well as from transfer to dietary items prepared
with drinking water) relative to other exposure sources; the remainder of the exposure equal to
the RfD is allocated to other potential exposure sources. For example, if for a particular
chemical, drinking water were to represent half of total exposure and diet were to represent the
other half, then the drinking water contribution (or RSC) would be 50%. The EPA considers any
potentially significant exposure source when deriving the RSC.
The RSC is derived by applying the Exposure Decision Tree approach published in the EPA's
Methodology for Deriving Ambient Water Quality Criteria for the Protection of Human Health
(USEPA, 2000b). The Exposure Decision Tree approach allows flexibility in the RfD
apportionment among sources of exposure and considers several characteristics of the
contaminant of interest, including the adequacy of available exposure data, levels of the
contaminant in relevant sources or media of exposure, and regulatory agendas (i.e., whether there
are multiple health-based criteria or regulatory standards for the contaminant). The RSC is
developed to reflect the exposure to the U.S. general population or a sensitive population within
the U.S. general population and may be derived qualitatively or quantitatively, depending on the
available data.
A quantitative RSC determination first requires "data for the chemical in question...
representative of each source/medium of exposure and... relevant to the identified population(s)"
(USEPA, 2000b). The term "data" in this context is defined as ambient sampling measurements
in the media of exposure, not internal human biomonitoring metrics. More specifically, the data
must adequately characterize exposure distributions including the central tendency and high-end
exposure levels for each source and 95% confidence intervals for these terms (USEPA, 2000b).
Frequently, an adequate level of detail is not available to support a quantitative RSC derivation.
When adequate quantitative data are not available, the agency relies on the qualitative
alternatives of the Exposure Decision Tree approach. A qualitatively-derived RSC is an estimate
that incorporates data and policy considerations and thus, is sometimes referred to as a "default"
RSC (USEPA, 2000b). Both the quantitative and qualitative approaches recommend a "ceiling"
RSC of 80%) and a "floor" RSC of 20% to account for uncertainties including unknown sources
of exposure, changes to exposure characteristics over time, and data inadequacies (USEPA,
2000b).
In cases in which there is a lack of sufficient data describing environmental monitoring results
and/or exposure intake, the Exposure Decision Tree approach results in a recommended RSC of
C-34
-------
FINAL
2024
20%. In the case of MCLG development, this means that 20% of the exposure equal to the RfD
is allocated to drinking water and the remaining 80% is reserved for other potential sources, such
as diet, air, consumer products, etc. This 20% RSC value can be replaced if sufficient data are
available to develop a scientifically defensible alternative value. If scientific data demonstrating
that sources and routes of exposure other than drinking water are not anticipated for a specific
pollutant, the RSC can be raised as high as 80% based on the available data, allowing the
remaining 20% for other potential sources (USEPA, 2000b). Applying a lower RSC (e.g., 20%)
is a more conservative approach to public health and results in a lower MCLG.
C.2.1. Literature Search and Screening
In 2020, the EPA conducted a literature search to evaluate evidence for pathways of human
exposure to eight PFAS chemicals (PFOA, PFOS, PFBA, PFBS, PFDA, perfluorohexanoic acid
(PFHxA), PFHxS, and PFNA) (Holder et al., 2023). This search was not date limited and
spanned the information collected across the Web of Science (WOS), PubMed, and
ToxNet/ToxLine (now ProQuest) databases. The results of the PFNA literature search of
publicly available sources are available through the EPA's Health & Environmental Resource
Online website at https://hero.epa.eov/hero/index.cfm/proiect/paee/proiect id/2633.
The 2,408 literature search results for PFNA were imported into SWIFT-Review (Sciome, LLC,
Research Triangle Park, NC) and filtered through the Evidence Stream tags to identify human
studies and nonhuman (i.e., those not identified as human) studies (Holder et al., 2023). Studies
identified as human studies were further categorized into seven major PFAS pathways (Cleaning
Products, Clothing, Environmental Media, Food Packaging, Home Products/Articles/Materials,
Personal Care Products, and Specialty Products) as well as an additional category for Human
Exposure Measures. Nonhuman studies were grouped into the same seven major PFAS pathway
categories, except that the Environmental Media category did not include soil, wastewater, or
landfill. Only studies published between 2003 and 2020 were considered. Application of the
SWIFT-Review tags identified 1,359 peer-reviewed papers matching these criteria for PFNA.
Holder et al. (2023) screened the 1,359 papers to identify studies reporting measured occurrence
of PFNA in human matrices and media commonly related to human exposure (human
blood/serum/urine, drinking water, food, food contact materials, consumer products, indoor dust,
indoor and ambient air, and soil). For this synthesis, additional screening was conducted to
identify studies relevant to surface water (freshwater only) and groundwater using a keyword8
search for water terms.
Following the Population, Exposure, Comparator, and Outcome (PECO) criteria outlined in
Table C-3, the title and abstract of each study were independently screened for relevance by two
screeners using litstream™. A study was included as relevant if it was unclear from the title and
abstract whether it met the inclusion criteria. When two screeners did not agree whether a study
should be included or excluded, a third reviewer was consulted to make a final decision. The title
and abstract screening of Holder et al. (2023) and of this synthesis resulted in 679 unique studies
being tagged as relevant (i.e., having data on occurrence of PFNA in exposure media of interest)
8 Keyword list: water, aquifer, direct water, freshwater, fresh water, groundwater, ground water, indirect water, lake,
meltwater, melt water, natural water, overland flow, recreation water, recreational water, river, riverine water,
riverwater, river water, springwater, spring water, stream, surface water, total water, water supply
C-3 5
-------
FINAL
2024
that were further screened with full-text review using the same inclusion criteria. After additional
review of the evidence collected by Holder et al. (2023), 98 studies originally identified for other
PFAS also contained information relevant to PFNA. Based on full-text review, 171 studies were
identified as having relevant, extractable data for PFNA from the United States, Canada, or
Europe for environmental media, not including studies with only human biomonitoring data. Of
these 171 studies, 156 were identified from (Holder et al., 2023), where primary data were
extracted into a comprehensive evidence database. Parameters of interest included: sampling
dates and locations, numbers of collection sites and participants, analytical methods, limits of
detection and detection frequencies, and occurrence statistics. Fifteen of the 171 studies were
identified in this synthesis as containing primary data on only surface water and/or groundwater.
The evidence database of Holder et al. (2023) additionally identified 18 studies for which the
main article was not available for review. As part of this synthesis, 17 of the 18 studies could be
retrieved. An additional three peer-reviewed references were identified through gray literature
sources that were included to supplement the search results. The combined 20 studies underwent
full-text screening using the inclusion criteria in Table C-3. Based on full-text review, five
studies were identified as relevant.
Table C-3. Populations, Exposures, Comparators, and Outcomes (PECO) Criteria
PECO Element Inclusion Criteria
Population Adults and/or children in the general population and populations in the
vicinity of PFAS point sources from the United States, Canada, or Europe
Exposure Primary data from peer-reviewed studies collected in any of the following
media: ambient air, consumer products, drinking water, dust, food, food
packaging, groundwater51, human blood/serum/urine, indoor air, landfill,
sediment, soil, surface water51 (freshwater), wastewater/biosolids/sludge
Comparator Not applicable
Outcome Measured concentrations of PFNA (or measured emissions from food
packaging and consumer products only)
Notes: PFNA = perfluorononanoic acid.
a Surface water and groundwater were not included as relevant media in Holder et al. (2023). Studies were re-screened for these
two media in this synthesis.
Using the screening results from the evidence database and this synthesis, a total of 176 studies
were identified as relevant. Forty-seven of these contained information relevant to the United
States and were summarized for this effort.
C.2.2. Additional Screening
The EPA also searched the following publicly available gray literature sources for information
related to relative exposure of PFNA for all potentially relevant routes of exposure (oral,
inhalation, dermal) and exposure pathways relevant to humans:
• ATSDR's Toxicological Profiles;
• CDC's national reports on human exposures to environmental chemicals;
• EPA's CompTox Chemicals Dashboard;
C-36
-------
FINAL
2024
• EPA's fish tissue studies;
• EPA's Toxics Release Inventory;
• Relevant documents submitted under the Toxic Substances Control Act and relevant
reports from the EPA's Office of Chemical Safety and Pollution Prevention;
• U.S. Food and Drug Administration's (FDA's) Total Diet Studies and other similar
publications from FDA, U.S. Department of Agriculture, and Health Canada;
• National Oceanic and Atmospheric Administration's (NOAA's) National Centers for
Coastal Ocean Science data collections;
• National Science Foundation direct and indirect food and/or certified drinking water
additives;
• Throwaway Packaging, Forever Chemicals: European wide survey o/PFAS in
disposable food packaging and tableware (Strakova et al., 2021);
• PubChem compound summaries;
• Relevant sources identified in the relative source contribution discussions (Section 5) of
the EPA's Proposed Approaches to the Derivation of a Draft Maximum Contaminant
Level Goal for Perfluorooctanoic Acid (PFOA)/Perfluorooctane Sulfonic Acid (PFOS) in
Drinking Water; and
• Additional sources, as needed.
The EPA has included available information from these gray literature sources for PFNA
relevant to its uses, chemical and physical properties, and for occurrence in ambient or indoor
air, foods (including fish and shellfish), soil, dust, and consumer products. The EPA has also
included available information specific to PFNA below on any regulations that may restrict
PFNA levels in media (e.g., water quality standards, air quality standards, food tolerance levels).
C.3. Summary of Potential Exposure Sources of PFNA Other
than Water
C.3.1. Dietary Sources
C.3.1.1. Seafood
PFNA was detected in 108 of 157 fish tissue composite samples collected during the EPA's
National Lake Fish Tissue Study, with a maximum concentration of 9.70 ng/g and a 50th
percentile concentration of 0.32 ng/g (Stahl et al., 2014). It was detected in one of 162 fish tissue
composite samples collected during the EPA's 2008-2009 National Rivers and Streams
Assessment (NRSA) at a concentration of 2.48 ng/g (Stahl et al., 2014). More recently, PFNA
was detected in 135 of 349 fish tissue composite samples at concentrations ranging from
0.100 ng/g to 1.910 ng/g in the EPA's 2013-2014 NRSA (USEPA, 2020a). PFNA was also
detected in 119 of 152 fish tissue composite samples at concentrations ranging from 0.12 ng/g to
9.32 ng/g in the EPA's 2015 Great Lakes Human Health Fish Fillet Tissue Study (USEPA,
2021g). In 2001, PFNA was detected at mean concentrations of 1.0 ng/g, 0.57 ng/g, 2.8 ng/g,
2.9 ng/g, and 1.1 ng/g (wet weight) in whole body homogenates of lake trout collected from
Lake Superior, Lake Michigan, Lake Huron, Lake Erie and Lake Ontario, respectively (Furdui et
al., 2007 as cited in ATSDR, 2021 and NCBI, 2022). In addition, PFNA was detected in lake
trout at concentrations of 0.70 ng/g for Lake Superior, 1.4 ng/g for Lake Huron, 2.6 ng/g for
C-37
-------
FINAL
2024
eastern Lake Erie, and 0.90 ng/g for Lake Ontario; PFNA was also detected at a concentration of
1.2 ng/g in walleye collected from western Lake Erie (ATSDR, 2021; De Silva et al., 2011).
PFNA was detected in mixtures of whole fish from the Missouri River, the Mississippi River,
and the Ohio River at concentrations of 0.43 ng/g, 0.78 ng/g, and 1.03 ng/g, respectively
(ATSDR, 2021; Ye et al., 2008). Concentrations of PFNA ranged from 0.01 ng/g to 0.73 ng/g in
capelin whole body samples, <0.09 ng/g to 1.3 ng/g in cod muscle samples, and 0.05 ng/g to
8.0 ng/g in salmon muscle samples collected from the Hudson Bay region of northeast Canada in
1999 to 2003 (NCBI, 2022b; Kelly et al., 2009). PFNA was not included in NOAA's National
Centers for Coastal Ocean Science, National Status and Trends Data (NOAA, 2022). Burkhard
(2021) identified 79 studies reporting BAFs for PFNA and calculated a median (standard
deviation) bioaccumulation factor (BAF) in muscle tissue/fillet of 144.54 ± 6.03 L/kg wet weight
(reported as a logBAF of 2.16 ± 0.78 L/kg).
Among the peer-reviewed studies identified, there was considerable variability in sampling and
analysis methodologies. Seafood products analyzed comprised a broad range of fish, crustaceans,
and mollusks which were locally caught, farmed, obtained from local seafood markets, and/or
obtained from large grocery chains. There was also considerable variability in sample
preparation which might affect the interpretation and comparability of results. Some studies
included tinned or prepared seafood while others focused only on unprepared or raw items; some
studies composited many organisms into each analysis sample while others focused on single
organism measurements; and some studies analyzed whole organisms while others measured
only muscle tissue, potentially excluding fatty tissues likely to be higher in PFNA. Results from
these studies are provided in detail in Table C-4.
Five U.S.-based studies were identified that evaluated PFNA levels in seafood (Young et al.,
2022; Chiesa et al., 2019; Byrne et al., 2017; Young et al., 2013; Schecter et al., 2010). Four of
these studies analyzed fish purchased from stores and fish markets. PFNA was detected
infrequently in samples reported in Chiesa et al. (2019), Schecter et al. (2010) and Young et al.
(2013): one of 10 samples of striped bass (1.4 ng/g) and in one of nine samples of shrimp
(1.2 ng/g), but not in samples of crab meat, catfish, clams, cod, flounder, pangasius, pollock, tuna
(including canned), salmon, scallops, tilapia, canned sardines, or frozen fish sticks. No other fish
types were sampled in these three studies, and other than canned tuna and sardines, none were
analyzed as prepared for eating. Seafood samples reported in Young et al. (2022) reported
detectable PFNA in five out of the eight types of seafood evaluated. These included canned
clams, canned tuna, cod, crab meat, and pollock (fish fillets and frozen fish sticks). No PFNA
was detected in salmon, tilapia or shrimp. Seafood packaging was also evaluated for PFAS
coatings, and it was determined the packaging did not contribute to any PFAS concentrations
observed in the study.
One study evaluated fish samples collected directly from rivers and lakes (Byrne et al., 2017). As
part of a study to assess exposure to PFNA and other PFAS among residents of two remote
Alaska Native villages on St. Lawrence Island, Byrne et al. (2017) measured PFAS
concentrations in stickleback and Alaska blackfish, resident fish used as sentinel species to detect
accumulation of PFAS in the local environment. Stickleback were collected from three locations:
Suqitughneq (Suqi) River watershed (n = 9 composite samples), Tapisaggak (Tapi) River (n = 2
composite samples), and Troutman Lake (n = 3 composite samples). Blackfish were collected
from the Suqi River (n = 29) but were not found in the other water bodies. Authors reported that
C-38
-------
FINAL
2024
the Suqi River watershed was upstream and downstream of a formerly used defense site and Tapi
River was east of a military site, however at the start of the study none of the sites were known to
be contaminated with PFAS. The sample dates were not reported. PFNA was not detected in the
blackfish samples but was detected in 100%, 56%, and 50% of stickleback samples from
Troutman Lake, Suqi River, and Tapi River, respectively, with authors noting that PFNA was the
most frequently detected PFAS in stickleback. PFNA concentrations ranged between 2.72 ng/g
and 4.13 ng/g ww at Troutman Lake, from below the limit of detection (LOD) to 1.52 ng/g ww
at Suqi River, and
-------
FINAL
2024
samples were wild caught, farmed, or purchased from local seafood markets and included seven
species of fish caught off the coast of Iceland (Jorundsdottir et al., 2014); three species of fish
caught in lakes in Norway (Hansen et al., 2016); four fish species caught in Canadian Rivers
(Bhavsar et al., 2014); rainbow trout collected from fish farms along the Swedish Baltic Sea
coast (Johansson et al., 2014); shrimp, squid, mussels, and seven species of fish caught or farmed
on the Greek coast (Vassiliadou et al., 2015); two species of fish originating from the West coast
of Greenland (Carlsson et al., 2014); five species of fish collected from a river and lake in
Germany (Holzer et al., 2011); eight species of fish collected from commercially and
recreationally important fishing areas across the Baltic Sea, a large freshwater Lake, and four
fish farming facilities in Finland (Koponen et al., 2015); cod caught along the Polish Baltic Sea
Coast (Falandysz et al., 2006); and two species of wild caught fish and one species of farmed
fish from Faroe island area of Denmark (Eriksson et al., 2013). Two studies reported sampling in
freshwater sources, one of which was likely to be contaminated with PFAS due to proximity to
an airport with known AFFF usage (Hansen et al., 2016), and the other which had nearby
industrial activities and previous monitoring results finding PFAS contamination (Bhavsar et al.,
2014), leading the authors to expect elevated PFAS concentrations in fish captured from these
sites. In both cases detectable levels of PFNA were measured in all fish sampled (Hansen et al.,
2016; Bhavsar et al., 2014). The highest reported PFNA contents from Hansen et al. (2016) was
2.39 ng/g ww in brown trout muscle tissue, while the highest reported PFNA contents reported in
Bhavsar et al. (2014) was 0.374 ng/g ww in fried Lake Trout. Among other European sites, there
were often no organisms sampled with measurable concentrations of PFNA or it was present
only in some organisms sampled.
Several studies from Europe provided PFNA measurements for seafood products purchased from
supermarkets and other retailers. These samples were not identified as originating from the
region of study and the product origin was often unknown or undisclosed. These included at least
four fish species purchased in Germany (Holzer et al., 2011); marine and freshwater fish
purchased in Southern France (Yamada et al., 2014); fish purchased in the center region of
France (Riviere et al., 2019); five species of fish purchased in Belgium, France, the Netherlands,
and Portugal (Barbosa et al., 2018); fresh and processed fish samples purchased from major
grocery store chains in Sweden (Gebbink et al., 2015; Vestergren et al., 2012); fish purchased in
the Netherlands (Noorlander et al., 2011); and a variety of seafood products purchased in Spain
(Domingo et al., 2012; Jogsten et al., 2009; Ericson et al., 2008a). Results from European market
studies were similar to those of U.S. studies, with PFNA detected infrequently among the
samples. However, several of the European seafood studies analyzed composites of all seafood
products sampled rather than individual organisms, thus the results are less precise than in the
U.S. studies.
C-40
-------
FINAL
2024
Table C-4. Summary of PFNA Occurrence in Seafood
Study
Location and Source
Seafood Type
Results
United States
Byrne et al. (2017)
United States (Alaska)
Stickleback collected from three locations on St.
Lawrence Island: Suqitughneq (Suqi) River
watershed (upstream and downstream of a formerly
used defense site), Tapisaggak (Tapi) River (located
approximately 5 km east of military site), and
Troutman Lake, a coastal lake situated adjacent to
the village of Gambell.
Alaska blackfish collected from the Suqi River but
were absent from the other water bodies.
Sampling year not reported. No sites were known to
be contaminated with PFASs at the initiation of the
study.
Stickleback and Alaska
blackfish
Strickleback:
Troutman Lake: n = 3*, DFa 100%, mean3
(range) = 3.43 (2.72-4.13) ng/g ww
Suqi River: n = 9*, DFa 56%, range =
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
*n reflects number of composite samples, each
composed of ~10 individual samples
Canada
Bhavsar et al. (2014)
Canada (Ontario)
Recreationally caught fish from four rivers - Credit
River, Thames River, Niagara River, Welland River
- in summer and fall of 2010 and 2011. Chinook
salmon were caught from Credit River, common
carp from Thames River, lake trout from Niagara
River, and walleye from Welland River. Elevated
PFASs concentrations were expected in the fish
based on nearby industrial activities or previous
monitoring work conducted by the Ontario Ministry
of Environment. Raw fish were analyzed, as well as
cooked fish using three different cooking methods
(baking, broiling, and frying).
Raw and cooked fish (chinook
salmon, common carp, lake
trout, walleye)
Chinook salmon:
Raw: n = 5, DF NR, mean = 0.067 ng/g ww
Baked: n = 5, DF NR, mean = 0.086 ng/g ww
Broiled: n = 5, DF NR, mean = 0.083 ng/g ww
Fried: n = 5, mean = 0.078 ng/g ww
Common carp:
Raw: n = 5, DF NR, mean = 0.092 ng/g ww
Baked: n = 5, DF NR, mean = 0.099 ng/g ww
Broiled: n = 5, mean = 0.105 ng/g ww
Fried: n = 5, mean = 0.101 ng/g ww
Lake trout:
Raw: n = 4, DF NR, mean = 0.298 ng/g ww
Baked: n = 4, DF NR, mean = 0.370 ng/g ww
Broiled: n = 4, mean = 0.358 ng/g ww
Fried: n = 4, mean = 0.374 ng/g ww
Walleye:
Raw: n = 5, DF NR, mean = 0.063 ng/g ww
Baked: n = 5, DF NR, mean = 0.079 ng/g ww
Broiled: n = 5, mean = 0.074 ng/g ww
Fried: n = 5, mean = 0.067 ng/g ww
(LOQ not reported)
Europe
Hansen et al. (2016)
Norway (Evenes; Skanland)
Fish were sampled from Lake Langvatnet, Lake
Lavangsvatnet, River Tarstadelva, and the reference
Lake Strandvatnet. A civilian airport (location also
shared with the Air Station of the Royal Norwegian
Air Force) is situated on a ridge between Lake
Langvatnet and Lake Lavangsvatnet. These waters
are affected by PFAS due to AFFF emissions from
the airport.Lake Lavangsvatnet drains into the river
Tarstadelva and Lake Strandvatnet is ~15 km away
from the airport with no connection to the airport
runoff. Samples of the dorsolateral muscle were
taken from 10 salmon, 10 anadromous brown trout,
12 stationary brown trout, and 3 European flounder
by local fishermen and by personnel from Sweco, an
environmental consulting company. The samples
were collected in August and September 2014.
Brown trout, European
flounder, salmon
Brown trout (stationary)
Lake Langvatnet: n = 6, DFa 100%, mean3
(range) = 1.05 (0.11-2.39) ng/g ww
Lake Lavangsvatnet: n = 5, DFa 100%, mean3
(range) = 0.7 (0.10-1.23) ng/g ww
Lake Strandvatnet (reference): n = 1, point =
0.10 ng/g ww
Brown trout (anadromous)
Lake Lavangsvatnet: n = 3, DFa 100%, mean
(range) = 0.06 (0.02-0.15) ng/g ww
River Tarstadelva: n = 5, DFa 80%, range =
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
Lake Lavangsvatnet: n = 3, DFa 100%, mean3
(range) = 0.153 (0.10-0.19) ng/g ww
Salmon (anadromous)
Lake Lavangsvatnet: n = 5, DFa 60%, range =
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
Rainbow trout: n = 10, DF 0% (<0.35 ng/g)
(LOQ = 0.18-0.39 ng/g)
Jorundsdottir et al. (2014)
Iceland
Samples were collected by the Icelandic Marine
Research Institute in March 2011 during their
biannual scientific survey. Cod and anglerfish were
caught south-west of Iceland, blue whiting was
caught south-east of Iceland, and lumpfish and
pollock were caught north-west of Iceland, while
ling, plaice, and lemon sole were caught west of
Iceland. Each fish sample consisted of a pooled
sample from the entire edible part from ten
individuals of the same species.
Anglerfish, Atlantic cod, blue
whiting, lemon sole, ling,
lumpfish, plaice, pollock
Anglerfish (n = 1), Atlantic cod (n = 2), blue whiting
(n = 2), lemon sole (n = 1), ling (n = 1), lumpfish (n
= 4), plaice (n = 1), pollock (n = 1): DF 0%
(LOD = 0.10 ng/g)
*n represents number of composite samples
Yamada et al. (2014)
France
Marine fish sampled were selected based on the fish
consumption habits of the population of four areas -
La Rochelle in Gironde-Charente Maritime Sud, Le
Havre in Normandy-Baie de Seine, Lorient in South
Brittany, and Toulon in Mediterranean-Var. Five
primary samples of fish were bought from the fish
market and/or supermarket in each region for each
species in January-April 2005.
Freshwater fish sampled were selected based on the
individual dietary consumption analysis of anglers
or their family members of the ICAR-PCB study.
Freshwater fish were collected in six major French
rivers with each river divided into three or four
section in 2008-2009. Half of the samples were
composite samples.
Freshwater fish, fresh or
frozen marine fish
Results presented for lower bound and upper bound
if LB value different from UB value
Fresh and frozen marine fish:
Total LB: n = 95, DF NR, mean (range) = 0.09
(0-0.27) ng/g ww
Total UB: n = 95, DF NR, mean (range) = 0.11
(0.04-0.27) ng/g ww
Anchovy: n = 1, LB-UB = 0-0.06 ng/g ww
Monkfish: n = 4, LB-UB = 0.09-0.11 ng/g ww
Catshark: n = 4, LB-UB = 0.06-0.08 ng/g ww
Cod: n = 4, LB-UB = 0.15 ng/g ww
Common dab: n = 4, LB-UB = 0.09-0.1 ng/g ww
Orange roughy: n = 3, LB-UB = 0.19 ng/g ww
Plaice/witch: n = 2, LB-UB = 0.16 ng/g ww
Goatfish: n = 3, LB-UB = 0.15-0.19 ng/g ww
Grenadier: n = 4, LB-UB = 0.05-0.07 ng/g ww
Gurnard: n = 1, LB-UB =0-0.04 ng/g ww
Haddock: n = 2, LB-UB = 0.17 ng/g ww
Hake: n = 4, LB-UB = 0.05-0.06 ng/g ww
C-44
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
Halibut: n = 4, LB-UB = 0.03-0.06 ng/g ww
John Dory: n = 2, LB-UB = 0.08-0.1 ng/g ww
Ling: n = 4, LB-UB = 0.03-0.07 ng/g ww
Mackerel: n = 4, LB-UB = 0.03-0.06 ng/g ww
Pollack: n = 3, LB-UB = 0.19 ng/g ww
Pout: n = 1, LB-UB = 0.12 ng/g ww
Ray: n = 4, LB-UB = 0.13 ng/g ww
Saithe: n = 4, LB-UB =0.04-0.06 ng/g ww
Salmon: n = 4, LB-UB = 0.05-0.07 ng/g ww
Sardine: n = 4, LB-UB = 0-0.07 ng/g ww
Scorpionfish: n = 1, LB-UB = 0.12 ng/g ww
Seabass: n = 4, LB-UB = 0.21-0.24 ng/g ww
Sea bream: n = 4, LB-UB = 0.07-0.09 ng/g ww
Sole: n = 4, LB-UB = 0.27 ng/g ww
Swordfish: n = 4, LB-UB = 0-0.06 ng/g ww
Tuna: n = 4, LB-UB = 0-0.04 ng/g ww
Whiting: n = 4, LB-UB = 0.07-0.1 ng/g ww
Freshwater fish:
Barbel: n = 5, LB-UB = 1.19-1.21 ng/g ww
Bleak: n = 9, LB-UB = 0.01-0.03 ng/g ww
Brown trout: n = 31, LB-UB = 0.08-0.13 ng/g ww
Chub: n = 9, LB-UB = 0.68-0.7 ng/g ww
Common carp: n = 7, LB-UB = 0.03-0.1 ng/g ww
Common roach: n = 67, LB-UB = 0.34-0.38 ng/g
WW
Minnow: n = 1, LB-UB = 0.19 ng/g ww
European eel: n = 137, LB-UB =0.12-0.26 ng/g
WW
European perch: n = 9, LB-UB = 0.12-0.13 ng/g
WW
Freshwater bream: n= 34, LB-UB =0.11-0.13
ng/g ww
Gudgeon: n = 5, LB-UB = 0.21-0.24 ng/g ww
Northern pike: n = 8, LB-UB = 0.04-0.07 ng/g
WW
White bream: n = 22, LB-UB = 0.06-0.11 ng/g
WW
Thicklip grey mullet: n = 6, LB-UB = 0.15 ng/g
WW
Wels catfish: n = 14, LB-UB = 0.08-0.1 ng/g ww
Western vairone: n = 1, LB-UB = 0-0.08 ng/g ww
Pike-perch: n = 22, LB-UB = 0.03-0.08 ng/g ww
C-45
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
(LOD = 0.007-0.95 ng/g for PFAAs other than
PFOAandPFOS)
*Lower bound (LB) scenario defined as values
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
M. edulis: North Sea, The Netherlands; France,
France
Gebbink et al. (2015)
Sweden
Food items were purchased at two major grocery
store chains in four major Swedish cities in 1999
and 2005. In 2010, sampling was limited to Uppsala
city since no systematic geographical differences in
food contamination was observed in the two earlier
market basket studies. The food items were selected
based on Swedish food and production statistics and
were not cooked before analysis. Homogenates of
fish products (fresh and frozen fillets of fish, canned
fish products, shellfish) were prepared for each
collection year by mixing food items proportionally
according to food consumption statistics. Results
were not reported for the 2005 and 2010 fish product
composite samples (only reported pooled with other
food types).
Fish
1999: n = 1, point = 0.07 ng/g
(MLOQ = 0.0003 ng/g)
*n represents number of composite samples
Vassiliadou et al. (2015)
Greece
Samples were obtained during the winter and early
spring of 2011. Finfish, squids, and shrimps were
purchased from the local fish market in Kallithea,
Athens, while mussels were obtained from a
mariculture farm within the Saronikos Gulf, Attika.
Samples were analyzed raw as well as cooked in the
ways favored in Greek cuisine (pan-fried in olive oil
and/or grilled). Quadruplicate composite samples
were created for each food type, each consisting of
four to six items of raw or cooked fish or shellfish.
Anchovy, bogue, hake,
picarel, sand smelt, sardine,
striped mullet, mussel, shrimp,
squid
Striped mullet:
Raw: n = 4, DF NR, mean = 0.60 ng/g ww
Fried: n = 4, DF NR, mean = 0.57 ng/g ww
Grilled: n = 4, DF NR, mean = 0.50 ng/g ww
Shrimp:
Raw: n = 4, DF NR, mean = 1.27 ng/g ww
Fried: n = 4, DF NR, mean = 1.52 ng/g ww
Anchovy (raw, fried, grilled), bogue (raw, fried,
grilled), hake (raw, fried, grilled), picarel (raw,
fried), sand smelt (raw, fried), sardine (raw, fried,
grilled), mussel (raw, fried), squid (raw, fried,
grilled): n = 4, DF 0%
(LOD = 0.42 ng/g; LOQ = 1.25 ng/g)
*n represents number of composite samples
Carlsson et al. (2014)
Greenland (Nuuk)
Seafood was purchased at the local fish market and
grocery shops in June 2010. All items were
originally caught in the vicinity of the Nuuk area
and/or along the West coast of Greenland and
represented the common food items consumed by
the local Inuit population.
Salmon, halibut
Raw salmon fillet: n = 6, DF 0%
Smoked salmon fillet: n = 6, DF 0%
Smoked halibut fillet: n = 6, DF 0%
(LOD = 0.014-0.224 ng/g for all PFAS)
Domingo et al. (2012)
Spain (Catalonia)
Foods purchased from 4 shops/stores of each of the
12 representative cities of Catalonia (Barcelona,
Fish and seafood (sardine,
tuna, anchovy, swordfish,
salmon, hake, red mullet, sole,
n = 2, DF NR, mean = 0.5 ng/g t\v
(LOD not reported)
*n represents number of composite samples
C-47
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
l'Hospitalet de Llobregat, Vilanova I la Geltru,
Matarr, Sabadell, Terrassa, Girona, Tarragona, Reus,
Tortosa, Lleida and Manresa) in September 2011.
Shops/stores included local markets, small stores,
supermarkets, and big grocery stores. For each food
item, two composite samples were prepared for
analysis, where each composite sample consisted of
24 individual units. Only edible parts of each food
item were included in the composites.
cuttlefish, clam, mussel, and
shrimp)
Vestergren et al. (2012)
Sweden (Malmoe, Gothenburg, Uppsala, Sundsvall)
Purchasing locations of the two largest retail chains
in Sweden were selected in each of four major
Swedish cities. All purchases were made in
spring/summer of 1999, 2005, and 2010. In 2010,
the study was limited to the largest retail chains in
Uppsala located in close vicinity to Stockholm. An
equal amount of each food group from each of the
four cities was combined into one sample pool to
provide a representative sample for the Swedish
urban population.
Fish products (fresh and
frozen fillets of fish, canned
fish products, shellfish)
1999: n = 1, point = 0.090 ng/g
2005: n = 1, point = 0.090 ng/g
2010: n = 1, point = 0.072 ng/g
(MDL = 0.0025 ng/g; MQL = 0.0063 ng/g)
*n represents number of composite samples
Noorlander et al. (2011)
The Netherlands
Fish randomly purchased from several Dutch retail
stores with nationwide coverage in November 2009.
Fish samples were ground, homogenized, and
pooled for analysis.
Fatty fish (herring, eel,
mackerel, salmon), lean fish
(cod, plaice, pollack, tuna),
crustaceans (mussels, shrimp,
crab)
Fatty fish: n = 1, point = 0.005 ng/g
Lean fish: n = 1, point = 0.077 ng/g
Crustaceans: n = 1, point = 0.058 ng/g
(LOD not reported)
*n represents number of composite samples
Jogsten et al. (2009)
Spain (Catalonia)
Fish samples purchased from local markets, large
supermarkets, and grocery stores from two different
areas of Tarragona Province, Catalonia in January
and February 2008. The cities of Tarragona and
Reus were sampled in the northern area and
L'Ametlla de Mar and Tortosa in the southern area.
For each food item, two composite samples were
analyzed (one composite for the northern area and
one for the southern area). Each composite was
formed of a minimum of six individual sub-samples
of the same product.
Marinated salmon (homemade
and packaged)
Homemade: n = 2, DF 0%
Packaged: n = 2, DF 0%
(LOD not reported)
*n represents number of composite samples
*Values of ND were replaced with l/2*LOD
Ericson et al. (2008a)
Spain
Food samples purchased from local markets, large
supermarkets, and grocery stores within Tarragona
County in July 2006. Food samples were randomly
White fish, seafood, tinned
fish, blue fish
White fish: n = 2, DF 0%
Seafood: n = 2, DF 0%
Tinned fish: n = 2, DF 0%
Blue fish: n = 2, DF 0%
C-48
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
purchased with origin source not specified. Each of
the food samples were duplicated and combined to
analyze a composite sample. Only the edible part of
each food was included in the composite samples.
Composite samples included the following:
White fish: hake, whiting blue, sea bass, monkfish
Seafood: mussel, shrimp
Tinned fish: tuna, sardine, mussel
Blue fish: salmon, sardine, and tuna
(LOD = 0.001-0.65 ng/g.fw)
*n represents number of composite samples
Johansson et al. (2014)
Sweden
Farmed rainbow trout (whole fish) were collected
from fish farms along the Swedish Baltic Sea coast
(brackish water). Only fish older than 12 months
were sampled. Samples were collected annually
from 1999 to 2010 within the Swedish National
Food Agency's official food control program.
Rainbow trout
n = 36, DF 0%
(MDL = 0.050 ng/g fw; MQL = 0.140 ng/g fw)
Multiple Continents
Chiesa et al. (2019)
United States (Pacific Ocean)
Wild-caught fish were collected at a wholesale fish
market in Milan, Italy. Sampling year was not
reported. The wild-caught salmon were from USA-
Pacific Ocean (Food and Agriculture Organization
Area 67 and 77).
Wild-caught salmon
(Oncorhynchus kisutch and
Oncorhynchus keta)
Oncorhynchus kisutch: n = 5, DF 0%
Oncorhynchus keta\ n = 2, DF 0%
(LOQ = 0.005 ng/g)
Canada
Wild-caught fish were collected at a wholesale fish
market in Milan, Italy. Sampling year was not
reported. The wild-caught salmon were from Canada
(Food and Agriculture Organization Area 67).
Wild-caught salmon
(Oncorhynchus nerka)
n = 15, DF 0%
(LOQ = 0.005 ng/g)
Norway
Farmed fish were collected at a wholesale fish
market in Milan, Italy. Sampling year was not
reported. The wild-caught salmon were from
Norway (Food and Agriculture Organization Area
27).
Farmed salmon (Salmo salar)
n = 25, DF 0%
(LOQ = 0.005 ng/g)
Scotland
Wild-caught and farmed fish were collected at a
wholesale fish market in Milan, Italy. Sampling year
was not reported. The wild-caught salmon were
from Scotland (Food and Agriculture Organization
Area 27).
Wild-caught and farmed
salmon (Salmo salar)
Wild-caught: n = 2, DF 0%
Farmed: n = 17, DF 0%
(LOQ = 0.005 ng/g)
Notes'. DF = detection frequency; ww = wet weight, LOD = limit of detection; LOQ = limit of quantitation; MDL = method detection limit; ND = not detected.
a The DF and/or mean was not reported in the study and was calculated in this synthesis. Means were calculated only when DF = 100%.
C-49
-------
FINAL
2024
C.3.1.2. Other Food Types
PFNA was included in a suite of PFAS evaluated in FDA's 2019, 2021, and 2022 Total Diet
Study Sampling (US FDA, 2022c, 2022b, 2022a, 2021b, 2021a, 2020b, 2020a); it was detected
at concentrations of 233 ng/kg (0.233 ng/g) in baked cod and 50 ng/kg (0.050 ng/g) in frozen
(oven-cooked) fish sticks or patties in 2021, but it was not detected in any of the other food
samples tested. It should be noted that FDA indicated that the sample sizes used in the PFAS
2019, 2021, and 2022 Total Diet Study Sampling were limited and that the results should not be
used to draw definitive conclusions about PFAS levels in the general food supply (US FDA,
2022c). PFNA was not detected in milk samples collected from a farm with groundwater known
to be contaminated with PFAS; however, it was detected in produce (corn) collected from an
area near a PFAS production plant in FDA studies of the potential PFAS exposure to the U.S.
population (US FDA, 2021c, 2018). PFNA was detected in beef steak in the Canadian Total Diet
studies from 1992 to 2004, but it was not detected in any of the other food samples tested
(ATSDR, 2021; Tittlemier et al., 2007).
Several U.S. studies were identified that examined PFNA in breastmilk or food types other than
breastmilk (Tipton et al., 2017; Blaine et al., 2014; Blaine et al., 2013; Young et al., 2012;
Schecter et al., 2010; von Ehrenstein et al., 2009; Kuklenyik et al., 2004) (Table C-3). Few U.S.
studies analyzed foods from any one origin - only two studies sampled from store- or market-
bought meats, eggs, produce, and dairy, one studied wild alligator meat, two sampled from crops
(produce and corn grain and stover) grown in biosolids-amended soils (and also control and
municipal soils) as part of greenhouse and field studies, and two studied breastmilk.
Two studies purchased food items from stores and markets for evaluation (Young et al., 2012;
Schecter et al., 2010). Schecter et al. (2010) assessed PFNA and other PFAS in food samples
collected from five Dallas, Texas grocery stores in 2009. The origin or source of each food was
not described. Food items included meat products (bacon, canned chili, chicken breast, ground
beef, roast beef, ham, sausage, and turkey), dairy (butter, cheeses, frozen yogurt, ice cream, milk,
and yogurt), eggs, and grains (cereal), fruits and vegetables (apples, potatoes), and fats/other
(canola oil, margarine, olive oil, peanut butter). PFNA was not detected in any of the food
samples. In Young et al. (2012), cow milk was purchased from retail markets across the
continental United States representing 17 states; the sampling year was not reported. Cow milk
samples included organic milk, vitamin D added milk, and ultra-pasteurized milk. PFNA was not
detected in any of the 49 retail milk samples (method detection limit (MDL) = 0.28 ng/g).
One study investigated PFAS levels in wild meat (Tipton et al., 2017). Tipton et al. (2017)
assessed alligator tail meat that was collected during the South Carolina recreational hunting
season between September to October 2015. Tail meat samples were collected from four
different public hunt units - Southern Coastal, Middle Coastal, Midlands, and Pee Dee. PFNA
was detected in samples from all hunt units with the exception of the Midlands (n = 2), where
PFNA was not detected. Median concentrations from Southern Coastal (n = 19), Middle Coastal
(n = 17), and Pee Dee (n = 2) were 0.107 ng/g, 0.102 ng/g, and 0.117 ng/g wet mass,
respectively.
C-50
-------
FINAL
2024
Two studies by Blaine et al. (2014; 2013) evaluated PFNA in crops grown in greenhouse and
field studies. In Blaine et al. (2014), PFAS levels were measured in celery root, pea fruit, and
radish root grown in a greenhouse study with control (unamended) soil, industrially impacted
soil, and municipal soil (n = 3-5). PFNA was detected in radish root and celery shoot from all
three soils and pea fruit from only industrially impacted soil. Mean concentrations of PFNA in
radish root for the control, industrially impacted, and municipal soil were 4.79 ng/g, 26.88 ng/g,
and 5.99 ng/g, respectively. Mean concentrations of PFNA in celery shoot for the control,
industrially impacted, and municipal soil were 1.89 ng/g, 13.81 ng/g, and 1.62 ng/g, respectively.
The mean concentration of PFNA in pea fruit in the industrially impacted soil was 1.45 ng/g.
Authors noted minor cross-contamination of the control soil due to the proximity of the
unamended soil to biosolids-amended soil. In Blaine et al. (2013), authors studied the uptake of
PFAS into edible crops in both field and greenhouse studies. In the field study, PFAS levels were
measured in corn grain and corn stover grown with control (unamended), urban biosolids-
amended, and rural biosolids-amended soil (n = 3-7). Mean PFNA concentrations were below
the LOQ in both corn grain and corn stover grown in any field study plots (<0.10 ng/g for corn
grain; <0.29 ng/g for corn stover). In the greenhouse study, lettuce and tomato plants were grown
in control soil, industrially impacted soil, or municipal soil (n = 3-5). Mean PFNA
concentrations were below the LOQ (2.96 ng/g) in any tomato plants but was detected in lettuce
grown in industrially impacted soil and municipal soil at mean concentrations of 57.39 ng/g and
4.73 ng/g, respectively. PFNA was not detected above the LOQ (0.04 ng/g) in lettuce grown in
control soil. Sampling year was not reported.
The remaining two studies evaluated the occurrence of PFNA in breastmilk (von Ehrenstein et
al., 2009; Kuklenyik et al., 2004). von Ehrenstein et al. (2009) collected breastmilk samples
between December 2004 and July 2005 from women between the ages of 18 and 38 at the time
of recruitment as part of the pilot study Methods Advancement for Milk Analysis (MAMA).
Women provided milk samples at two visits - the first visit was 2-7 weeks postpartum, and the
second visit was 3-4 months postpartum. PFNA was not detected in any of the samples from the
first visit (n = 18) or second visit (n = 20). Similarly, PFNA was below the LOD (1.0 ng/mL) in
the samples reported by Kuklenyik et al. (2004). Kuklenyik et al. (2004) did not report
information on the breastmilk donors or the sampling procedure as it was unavailable; PFNA
was not detected in either of the two samples.
Results for all of the identified non-U.S. studies are presented in detail in Table C-5 and are
summarized here. Among the European studies, many collected food samples of unknown origin
from grocery stores (Scordo et al., 2020; Riviere et al., 2019; Sznajder-Katarzynska et al., 2019;
Sznajder-Katarzynska et al., 2018; Papadopoulou et al., 2017; Surma et al., 2017; D'Hollander et
al., 2015; Gebbink et al., 2015; Perez et al., 2014; Herzke et al., 2013; Hlouskova et al., 2013;
Domingo et al., 2012; Vestergren et al., 2012; Noorlander et al., 2011; Jogsten et al., 2009;
Ericson et al., 2008a). A wide variety of items were analyzed including meats and seafood, dairy,
fruits and vegetables, grains, pastries and other sweets, spices, sweeteners, and other beverages.
Of these studies, five reported no detectable PFNA in any of the items sampled (Riviere et al.,
2019; Sznajder-Katarzynska et al., 2018; Surma et al., 2017; Jogsten et al., 2009; Ericson et al.,
2008a). Among the studies that did report detectable PFNA in some food items, PFNA was
found in all major food categories examined except spices, salts, and sweeteners (sugar and
honey) though there was no apparent trend in specific food types with PFNA present in
measurable quantities.
C-51
-------
FINAL
2024
Three studies focused on PFNA in eggs. Ghelli et al. (2019) collected eggs from commercial
laying hen farms in Italy. PFNA was detected at trace levels only in 5% of the 44 eggs sampled.
Johansson et al. (2014) collected eggs from 20 to 25 commercial egg producers in Sweden every
year between 1999-2010. Ten to 12 eggs were pooled for analysis and they reported detectable
quantities of PFNA in 28% of 72 pooled samples, with a maximum observed value of 0.143 ng/g
fw. Zafeiraki et al. (2016a) collected home and commercially produced eggs in Greece and the
Netherlands. They reported detectable levels of PFNA in home produced eggs from both regions,
with a maximum observed value of 2.0 ng/g. PFNA was not detected in any of the commercially
produced eggs sampled.
Two studies focused on PFNA in milk and other dairy products. Johansson et al. (2014) collected
milk from about 10 dairy farms in Sweden every year between 1999-2010. PFNA was not
present at detectable levels in any of the samples. Still et al. (2013) collected commercially sold
dairy product samples in Germany. PFNA was present at detectable levels in one of four milk
samples, all three cheese samples and one butter sample, with a maximum observed value of
0.0094 ng/g. They reported non-detect results in all other products examined. Eriksson et al.
(2013) identified PFNA concentrations in milk from two major dairy farms, with a maximum of
0.073 ng/g ww. They identified PFNA in 75% of four additional dairy samples from Faroe
Island, but it was not detected in other samples of yogurt, creme fraiche, and potatoes.
Vestergren et al. (2013) evaluated concentrations in milk and reported a mean of 0.0023 ng/mL.
This study also evaluated concentrations in cow liver, blood, and muscle, finding the highest
concentration in liver at a mean of 0,016 ng/g.
In addition, two non-U. S. studies focused on locally caught food items of importance to
indigenous populations. Binnington et al. (2017) sampled whale blubber collected from two
beluga whales captured off the West coast of Canada. PFNA was present at measurable
concentrations in all solid samples (max = 0.1718 ng/g lipids) but was not detectable in rendered
oil. Carlsson et al. (2014) assessed wild caught seal beef, narwhal, and whale beef in Greenland
and found detectable concentrations of PFNA only in seal beef.
PFNA was also detected in some beverages. Stahl et al. (2014) measured PFNA in a selection of
Hessian, Belgian, and Bavarian beers and found measurable concentrations in 14% of 93
samples. Eschauzier et al. (2013) measured PFNA in cola and brewed coffee samples collected
from various locations in Amsterdam and found that PFNA was not present at measurable
concentrations in any sample.
Of the thirteen non-U. S. studies that evaluated the occurrence of PFNA in breastmilk, five did
not report detectable concentrations. These studies evaluated 13 women in Ontaria, Canada
(Kubwabo et al., 2013); 10 women in Spain (Karrman et al., 2010), 11 pooled samples obtained
from 109 women in Ireland (Pratt et al., 2013), 61 women in France (Cariou et al., 2015), and 20
women in Spain (Llorca et al., 2010). However, the remaining European studies did report
quantifiable concentrations of PFNA in breastmilk. Among these studies, detectable
concentrations of PFNA were reported in 17% of samples measured from 12 individual Swedish
women and 33% of nine composite samples composed of breastmilk from 25-90 Swedish
women (Karrman et al., 2007); 42.5% of breastmilk samples from 40 Belgian women (Croes et
al., 2012); 6% of breastmilk samples from 67 French women (Motas Guzman et al., 2016); 100%
of 31 composite samples composed of breastmilk from 5-116 Swedish women and 94% of
samples measured from 46 individual Swedish women (Nyberg et al., 2018); 5% of samples
C-52
-------
FINAL
2024
from 14 Spanish women (Beser et al., 2019); 69% of 16 samples from Irish women (Abdallah et
al., 2020); 48% of samples from 50 women in the Czech Republic (Lankova et al., 2013); and
2% of samples from 48 French women (Antignac et al., 2013). Abdallah et al. (2020) reported
the highest value for PFNA in breastmilk, at 0.1 ng/mL.
Several European studies focused on other sources of dietary exposures in children. Lankova et
al. (2013) measured PFNA in infant formula samples purchased at Czech retail markets and
reported a maximum observed value of 0.011 ng/g. Llorca et al. (2010) measured PFNA in baby
cereals and infant formulas purchased from reatailers in Spain. They reported PFNA present in
measureable quantities in all of the samples measured, with a maximum reported value of 0.1138
ng/g in baby cereal and 0.219 ng/g in infant formulas. Dellatte et al. (2013) measured PFNA in
ready-to-eat meals collected from nursery and primary school canteens. PFNA was present at
detectable levels in one of six samples, with a reported concentration of 0.0063 ng/g.
C-53
-------
FINAL
2024
Table C-5. Summary of PFNA Occurrence in Other Food
Study
Location and Source
Food Types
PFNA Results
United States
Schecter et al. (2010)
United States (Texas)
Food samples from five different grocery stores in
Dallas, Texas were collected in 2009. Ten individual
samples were collected for each food type and
combined to form composite samples. The
origin/source of the food samples were not reported.
Dairy; fruits and vegetables;
grains; meat; fats/other
Meat:
Hamburger: n = 1, DF 0%
Bacon: n = 1, DF 0%
Sliced turkey: n = 1, DF 0%
Sausages: n = 1, DF 0%
Ham: n= 1,DF0%
Sliced chicken breast: n = 1, DF 0%
Roast beef: n = 1, DF 0%
Canned chili: n = 1, DF 0%
Dairy and Eggs:
Butter: n = 1, DF 0%
American cheese: n = 1, DF 0%
Other cheese: n = 1, DF 0%
Whole milk: n = 1, DF 0%
Ice cream: n = 1, DF 0%
Frozen yogurt: n = 1, DF 0%
Whole milk yogurt: n = 1, DF 0%
Cream cheese: n = 1, DF 0%
Eggs: n = 1, DF 0%
Grains:
Cereals: n = 1, DF 0%
Fruits and Vegetables:
Apples: n = 1, DF 0%
Potatoes: n = 1, DF 0%
Fats and Other:
Olive oil: n = 1, DF 0%
Canola oil: n = 1, DF 0%
Margarine: n = 1, DF 0%
Peanut butter: n = 1, DF 0%
(LOD not reported for any food item)
*n reflects number of composite samples, each
composed of ~10 individual samples
Young et al. (2012)
United States (17 states)
Retail cow's milk samples were all pasteurized
whole milk, commercially available, and purchased
at retail markets across the continental United States
representing 17 states. Samples were organic milk,
Dairy
n = 49, DF 0%
(MDL = 0.28 ng/g)
C-54
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
vitamin D added milk, or ultra-pasteurized milk.
Sampling year not reported.
Tipton et al. (2017)
United States (South Carolina)
Alligator tail meat samples were collected from a
local wild game meat processer during the South
Carolina recreational hunt season between
September to October 2015. Samples were from
four different public hunt units—Southern Coastal,
Middle Coast, Midlands, and Pee Dee.
Meat
Alligator tail:
Southern coastal: n = 19, DFa 74%, median
(range) = 0.107 (<0.088-0.551) ng/g wet mass
Middle coastal: n = 17, DFa 65%, median
(range) = 0.102 (<0.073-0.553) ng/g wet mass
Pee Dee: n = 2, DFa 100%, median (range) =
0.117 (0.100-0.135) ng/g wet mass
Midlands: n = 5, DF 0%
(RL not reported)
Blaine et al. (2014)
United States (Midwest)
Crops grown in in greenhouse study with control
(unamended), industrially impacted soil, or
municipal soil. Control soil had minor cross-
contamination due to proximity to biosolids-
amended fields. Industrially impacted soil was
amended with industrially impacted biosolids, and
municipal soil was amended with municipal
biosolids for over 20 years.
Crops grown in the greenhouse study were grown
from seed in pots, which were randomly arranged
within the greenhouse. Sampling year not reported.
Fruits and vegetables
Radish root:
Control: n = 3-5, DF NR, mean = 4.79 ng/g
Industrially impacted; n = 3-5, DF NR, mean =
26.68 ng/g
Municipal: n = 3-5, DF NR, mean = 5.99 ng/g
Celery shoot:
Control: n = 3-5, DF NR, mean = 1.89 ng/g
Industrially impacted: n = 3-5, DF NR, mean =
13.81 ng/g
Municipal: n = 3-5, DF NR, mean = 1.62 ng/g
Pea fruit:
Control: n = 3-5, DF 0%
Industrially impacted: n = 3-5, DF NR, mean
1.45 ng/g
Municipal: n = 3-5, DF 0%
(LOQ = 0.07ng/g)
Blaine et al. (2013)
United States (Midwest)
Crops grown in urban and rural full-scale field study
with control (unamended) and biosolids-amended
soil. Three agricultural fields were amended (0.5x,
1 x, or 2x) with municipal biosolids. Urban biosolids
(1 x and 2*) were from a WWTP receiving both
domestic and industrial waste. Rural biosolids (0.5x)
were from a WWTP receiving domestic waste only.
Control plots were proximal to the rural and urban
amended corn grain and corn stover field sites;
sampling year not provided.
Fruits and vegetables; grains
Field study:
Corn grain:
Urban nonamended: n = 3-7, DF NR, mean =
<0.10 ng/g
Urban 1 x: n = 3-7, DF NR, mean = <0.10 ng/g
Urban 2x: n = 3-7, DF NR, mean = <0.10 ng/g
Rural nonamended: n = 3-7, DF NR, mean =
<0.10 ng/g
Rural 0.5x: n = 3-7, DF NR, mean = <0.10 ng/g
Corn stover:
Urban nonamended: n = 3-7, DF NR, mean =
<0.29 ng/g
C-55
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
Crops grown in greenhouse study with control
(nonamended) and biosolids-amended soil.
Nonamended soil obtained from a field that received
commercial fertilizers and had a similar cropping
system as the nearby municipal soil site. Municipal
soil was obtained from a reclamation site in Illinois
where municipal bioso lids were applied at
reclamation rates for 20 years, reaching the
cumulative biosolids application rate of 1,654
Mg/ha. Industrially impacted soil was created by
mixing composted biosolids from a small municipal
(but impacted by PFAA manufacturing) WWTP
with control soil on a 10% mass basis. Sampling
year not provided.
Urban 1 x: n = 3-7, DF NR, mean = <0.29 ng/g
Urban 2*: n = 3-7, DF NR, mean = <0.29 ng/g
Rural nonamended: n = 3-7, DF NR, mean =
<0.29 ng/g
Rural 0.5*: n = 3-7, DF NR, mean = <0.29 ng/g
(LOQ = 0.10 ng/g for corn grain; LOQ = 0.29 ng/g
for corn stover)
Greenhouse study:
Lettuce:
Nonamended: n = 3-5, DF NR, mean = <0.04
ng/g
Industrially impacted: n = 3-5, DF NR, mean
= 57.39 ng/g
Municipal: n = 3-5, DF NR, mean = 4.73 ng/g
Tomato:
Nonamended: n = 3-5, DF NR, mean = <2.86
ng/g
Industrially impacted: n = 3-5, DF NR, mean =
<2.86 ng/g
Municipal: n = 3-5, DF NR, mean = <2.86 ng/g
(LOQ = 0.04 ng/g for lettuce; LOQ = 2.86 ng/g for
tomato)
von Ehrenstein et al. (2009)
United States (North Carolina)
As part of the Methods Advancement for Milk
Analysis (MAMA) pilot study, 34 breastfeeding
women aged 18 to 38 years at recruitment provided
breastmilk samples at two visits. The first visit
occurred 2-7 weeks postpartum, and the second visit
occurred 3^1 months postpartum. Both visits were
between December 2004 and July 2005.
Breastmilk
Visit #1: n= 18,DF0%
Visit #2: n = 20, DF 0%
(LOQ = 0.30 ng/mL)
Kuklenyik et al. (2004)
United States (Georgia)
Authors reported that no information was provided
on the human milk donors or the sampling
procedure.
Breastmilk
n = 2, DF 0%
(LOD = 1.0 ng/mL)
Canada
Binnington et al. (2017)
Canada (Tuktoyaktuk, Northwest Territory)
Samples were collected from two beluga whales
(HI-14-06 and HI-14-11) caught offshore of
Tuktoyaktuk during summer 2014. The belugas
Meat
Beluga whale blubber - muktuk
Air dry: n = 4, DF NR, range of means =
0.0713-0.1288 ng/g lipids
Hang dry: n = 4, DF NR, range of means =
0.1013-0.150 ng/g lipids
C-56
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
were members of the Eastern Beaufort Sea
population, with a home range including respective
summering and wintering grounds in the Beaufort
and Bering Seas. The selected individuals were
harvested on the shore of nearby Hendrickson
Island. Beluga whale blubber forms the basis for two
distinct traditional food types: (i) muktuk designates
food items composed of the outer layer of blubber
and its attached skin and connective tissue, while (ii)
uqsuq designates food items derived from the inner
layers of blubber. Samples were analyzed in
duplicate at each step in the preparation process.
Boil drum: n = 4, DF NR, range of means
ll.IIS.il II.I26S ng/g lipids
Boil pot: n = 4, DF NR, range of means =
0.0617-0.079 ng/g lipids
Roast: n = 4, DF NR, range of means = 0.0813-
0.1205 ng/g lipids
Aged 2 days: n = 4, DF NR, range of means =
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
Sour cream: n = 5, DFa 60%, range = 0.03-0.05
ng/g
Camembert-type cheese (n = 5), butter (n = 5): DF
0%
(LOD = 0.006 ng/g; LOQ = 0.019 ng/g)
*Range reported for detected values
Riviere et al. (2019)
France
Samples collected between July 2011 and July 2012
in the center region of France. Food items were
selected based on the results of a national
consumption survey to obtain a representative and
general view of children's (0-3 years old) food
consumption. All analyzed samples were formed of
12 subsamples of the same food and of equal
weight. The products purchased were prepared in a
way that reflected as closely as possible what is
done in the home (preparation and cooking).
Meat; dairy; fruits and
vegetables; fats/other
Infant-specific foods:
Milk-based beverage (n = 8), cereals (n = 5), milk-
based desserts (n = 6), growing-up milk (n = 9),
soups and puree (n = 11), fruit puree (n = 4),
vegetable-based ready-to-eat meal (n = 20),
meat/fish-based ready-to-eat meal (n = 45), infant
formula (n = 28), follow-on formula (n = 33): DF
0%
Common foods:
Non-alcoholic beverages (n = 1), dairy-based
desserts (n = 1), milk (n = 1), mixed dishes (n = 1),
ultra-fresh dairy products (n = 1), meat (n = 1),
poultry and game (n = 1): DF 0%
(LOD = 0.0002-3.7 ng/g for all PFAS)
*n represents number of composite samples
Sznajder-Katarzynska et al.
(2018)
Poland
Samples were purchased in Polish markets in 2017.
Individual food items were selected among the most
frequently consumed in Poland. Vegetables
(potatoes, beetroots, carrots, white cabbage,
tomatoes) and fruits (apples, cherries, strawberries)
of Polish origin were bought in season when
naturally ripe. Bananas, lemons, and oranges were
bought after being imported to Poland. Five
different samples of each fruit or vegetable were
collected.
Fruits and vegetables
Apples, bananas, cherries, lemons, oranges,
strawberries, beetroots, carrots, tomatoes, potatoes,
and white cabbage: n = 5 for each, DF 0%
(LOD = 0.007 ng/g; LOQ = 0.014 ng/g)
Surma et al. (2017)
Spain, Slovakia
Spice samples were collected in powder form from
Spain and Slovakia. Sampling year not reported.
Fats/other
Spain:
Anise (n = 1), star anise (n = 1), white pepper (n =
1), fennel (n = 1), cardamom (n = 1), clove (n = 1),
coriander (n = 1), nutmeg (n = 1), allspice (n = 1),
cinnamon (n = 2), vanilla (n = 1), ginger (n = 1),
peppermint (n = 1), parsley (n = 1), thyme (n = 1),
laurel (n = 1), garlic (n = 1), cumin (n = 1), black
C-58
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
pepper (n = 1), mild hot pepper (n = 1), hot hot
pepper (n = 1), oregano (n = 2): DF 0%
Slovakia:
Anise (n = 1), star anise (n = 1), white pepper (n =
1), fennel (n = 1), cardamom (n = 1), clove (n = 1),
coriander (n = 1), nutmeg (n = 1), allspice (n = 1),
cinnamon (n = 1), vanilla (n = 1), ginger (n = 1):
DF 0%
(LOD = 0.03lg/g; LOQ = 0.093 ng/g)
Zafeiraki et al. (2016a)
Greece, The Netherlands
Home and commercially-produced eggs were
collected from different regions in the Netherlands
and Greece in August 2013-August 2014. Home-
produced eggs were voluntarily provided, and
commercial eggs were purchased from
supermarkets. The yolks of the same sample of eggs
were pooled, homogenized, and then analyzed.
Fats/other
Domestic eggs:
The Netherlands: n = 73, DF 18%, median
(range) = 0.9 (<0.5-2.0) ng/g
Greece: n = 45, DF 20%, median (range) = 0.8
(<0.5-l) ng/g
Commercial eggs:
The Netherlands: n = 22, DF 0%
Greece: n = 31, DF 0%
(LOD = 0.15 ng/g; LOQ = 0.5 ng/g)
*Median calculated only on the concentrations above
LOQ
Zafeiraki et al. (2016b)
The Netherlands
Samples purchased from local markets and
slaughterhouses in the Netherlands in 2014. Samples
included liver samples of horse, sheep, bovine, pig,
and chicken.
Meat
Horse: n = 19, DF 0%
Sheep: n= 18, DF 0%
Bovine: n = 22, DF 0%
Pig: n = 20, DF 0%
Chicken: n = 20, DF 0%
(LOQ = 0.5 ng/g ww)
D'Hollander et al. (2015)
Czech Republic, Belgium, Norway, Italy
The Czech Republic, Belgium, Norway, and Italy
were selected to represent eastern, western, northern,
and southern Europe. Sampling took place between
spring and summer 2010 as part of the PERFOOD
study. Individual items were randomly selected in
three national retail stores covering different brands
or countries of origin. Of each item, three to ten
single samples were combined to create a pooled
sample. Individual food items that were collected to
create pooled samples were:
Cereals: rice, wheat (white), oats, rye
Grains; fruits; fats/other
Czech Republic:
Cereals: wheat (white), oats, rye: n = 1 each, point
= <0.020 ng/g
Sweets: sugar (beet), honey: n = 1 each, point =
<0.001 ng/g
Fruits - berries: strawberries: n = 1, point = <0.001
ng/g
Fruits - citrus fruit: oranges, tangerines: n = 1,
point = <0.001 ng/g
Fruits - pipe and stone fruit:
Apples: n = 1, point = 0.002 ng/g
Pears: n = 1, points = 0.001 ng/g
Peaches: n = 1, point = <0.001 ng/g
C-59
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
Sweets: sugar (beet), sugar (cane), honey
Fruits - others and exotic fruit: melons: n = 1,
Fruits: berries - strawberries, citrus fruit -
point = <0.004 ng/g
oranges, tangerines, lemons, grapefruits, pipe
Miscellaneous: rock salt: n = 1, point = <0.001
and stone fruit - apples, pears, peaches, plums,
ng/g
others and exotic fruit - melons, grape, bananas
Italy:
Miscellaneous: "rock" salt, "marine" salt
Cereals:
Rice, maize: n = 1 each, point = <0.001 ng/g
Wheat (white): n = 1, point = <0.020 ng/g
Sweets: sugar (beet), honey: n = 1 each, point =
<0.001 ng/g
Fruits - citrus fruit: lemons: n = 1, point = <0.001
ng/g
Fruits - pipe and stone fruit:
Apples: n = 1, point = 0.016 ng/g
Pears: n = 1, point = 0.002 ng/g
Peaches: n = 1, point = 0.012 ng/g
Plums: n = 1, point = <0.001 ng/g
Fruits - others and exotic fruit:
Grapes: n = 1, point = <0.001 ng/g
Bananas: n = 1, point = 0.003 ng/g
Miscellaneous: marine salt: n = 1, point = <0.001
ng/g
Norway:
Cereals: wheat (white): n = 1, point = <0.020 ng/g
Sweets: sugar (cane), honey: n = 1 each, point =
<0.001 ng/g
Fruits - berries: strawberries: n = 1, point = <0.001
ng/g
Fruits - citrus fruit:
Oranges: n = 1, point = <0.001 ng/g
Grapefruits: n = 1, point = 0.0248 ng/g
Fruits - pipe and stone fruit: apples, pears: n = 1
each, point = <0.001 ng/g
Fruits - others and exotic fruit: melons: n = 1,
point = 0.0099 ng/g
Miscellaneous: rock salt: n = 1, point = <0.001
ng/g
Belgium:
Cereals: rice, wheat (white), wheat (dark), oats: n
= 1 each, point = <0.001 ng/g
Sweets: sugar (beet), honey: n = 1 each, point =
<0.001 ng/g
C-60
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
Fruits - berries: strawberries: n = 1, point = <0.001
ng/g
Fruits - citrus fruit: oranges, lemons: n = 1 each,
point = <0.001 ng/g
Fruits - pipe and stone fruit:
Apples: n = 1, point = 0.205 ng/g
Pears: n = 1, point = <0.001 ng/g
Plums: n = 1, point = <0.001 ng/g
Fruits - others and exotic fruit: grapes: n = 1,
point = <0.007 ng/g
(LOD = 0.001 or 0.020 ng/g)
*n represents number of composite samples
Gebbink et al. (2015)
Sweden
Food items were purchased at two major grocery
store chains in four major Swedish cities in 1999
and 2005. In 2010, sampling was limited to Uppsala
city since no systematic geographical differences in
food contamination was observed in the two earlier
market basket studies. The food items were selected
based on Swedish food and production statistics and
were not cooked before analysis. The food items
were divided into 12 groups and homogenates for
each food group were prepared by mixing food
items proportionally according to food consumption
statistics. Results by food group were not reported
for the 2005 and 2010 years. For all three sampling
years, a homogenate was prepared by mixing
proportional amounts of each food group according
to consumption data for the respective year (includes
fish samples).
Fruits and vegetables; meat;
grains; fats/other
1999:
Dairy products: n = 1, point = 0.0005 ng/g
Meat products: n = 1, point = 0.0067 ng/g
Fats: n = 1, point = 0.0037 ng/g
Pastries: n = 1, point = 0.0012 ng/g
Egg: n = 1, point = 0.024 ng/g
Fruit: n = 1, point = 0.0006 ng/g
Soft drinks: n = 1, point = 0.0005 ng/g
Cereal products: n = 1, point = <0.0003 ng/g
Vegetables: n = 1, point = <0.0003 ng/g
Potatoes: n = 1, point = <0.0003 ng/g
Sugar and sweets: n = 1, point = <0.0003 ng/g
Year pool: n = 12, point = 0.0077 ng/g
2005:
Year pool: n = 12, point = 0.016 ng/g
2010:
Year pool: n = 12, point = 0.015 ng/g
(MLOQ = 0.0003 ng/g)
*n represents number of composite samples
Carlsson et al. (2014)
Greenland (Nuuk)
Meat was purchased at the local fish market and
grocery shops in June 2010. All items were
originally caught in the vicinity of the Nuuk area
and/or along the West coast of Greenland and
represented the common food items consumed by
the local Inuit population.
Meat
Seal beef: n = 2, DFa 50%, range =
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
Perez et al. (2014)
Serbia (Belgrade and Novi Sad), Spain (Barcelona,
Girona, and Madrid)
Between September 2011 and February 2013,
samples were purchased from different supermarkets
and retail stores in representative cites around the
world, including cities in Serbia and Spain in
Europe. Foodstuffs were grouped into the following
categories: cereals; pulses and starchy roots; tree-
nuts, oil crops, and vegetable oils; vegetables and
fruits; meat and meat products; milk, animal fats,
dairy products, and eggs; fish and seafood; and other
such as candies and coffee.
Grains; fruits and vegetables;
fats/other; meat; dairy;
seafood
Spain: n = 174, DF 13.3%, mean, median (range)
= 1.175, 0.43 (ND-13) ng/g
Serbia: n = 36, DF 10.3%, mean, median (range)
= 0.208, 0.185 (ND-0.43) ng/g
(MLOD = 0.039-0.412 ng/g, depending on food
item)
*Results not reported for individual food categories
Stahl et al. (2014)
Belgium
Samples of 83 Hessian beers were obtained from the
Hessian control authority, 4 Bavarian beers were
provided by the Bavarian Health and Food Safety
Authority, and 6 Belgian beers were obtained from
Belgian retail stores. Sampling year not provided.
Fats/other
All beer: n = 93, DF (frequency of quantification)
14%, mean, median (maximum) = 0.00377,
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
study. Individual items were randomly selected in
three national retail stores covering different brands
or countries of origin. Of each item, three to ten
single samples were combined to create one pooled
sample per country. The following items were
sampled:
Root vegetables: carrots
Bulb vegetables: onions
Fruiting vegetables: tomatoes, courgettes,
cucumbers, aubergine, peppers
Brassica vegetables: cauliflower, cabbage,
broccoli
Leaf vegetables: lettuce and other salads,
spinaches, chicory, pre-packed lettuce mix, pre-
packed and minced frozen spinach
Stem vegetables: asparagus, celery, fennel,
cultivated mushrooms
Starchy root tubers: potatoes, prepacked ready-to-
cook pommes frites
Legumes, beans, dried: peas, beans
Norway: n = 17, DF NR, mean = 0.00226 ng/g
(MQL = 0.002-0.050 ng/g)
*n represents number of composite samples
* Values below the MQL were substituted with the
MQL value
Hlouskova et al. (2013)
Belgium, Czech Republic, Italy, Norway
Food products were randomly purchased in several
nationwide supermarkets in four European regions
during summer 2010. Within the sampling
campaign, the collection of at least one food item
per subcategory (meat, fish, hen eggs, milk and
dairy products, and butter) in all four countries was
acquired. Food items within each subcategory
included the following:
Meat: beef, canned pork meat, poultry, pork,
pig/bovine liver, rabbit, and/or sheep/lamb
Fish: farmed freshwater fish, farmed marine fish,
and/or seafood)
Hen eggs
Milk and dairy products: ultra-high temperature
whole cow milk, ultra-high temperature
skimmed cow milk, cheese (yellow,
Gouda/Edamer, etc.), and butter
Samples were pooled within a respective food
category but not across food groups.
Pooled milk/dairy products;
meat; fish; hen eggs
n = 50, DF 16%, mean, median (range) = 0.0295,
0.0253 (0.00503-0.0701) ng/g
(MQL = 0.005 ng/g for fish and seafood, meat, hen
eggs, and cheese; 0.002 ng/mL for milk, and 0.025
ng/g for butter)
*n represents number of pooled samples
*Results not reported for individual food groups
C-63
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
Domingo et al. (2012)
Spain (Catalonia)
Foods purchased from 4 shops/stores of each of the
12 representative cities of Catalonia (Barcelona,
l'Hospitalet de Llobregat, Vilanova I la Geltru,
Matarr, Sabadell, Terrassa, Girona, Tarragona, Reus,
Tortosa, Lleida and Manresa) in September 2011.
Shops/stores included local markets, small stores,
supermarkets, and big grocery stores. Analyzed
samples included 40 items:
Meat and meat products: veal steak, loin of pork,
chicken breast, steak of lamb, boiled ham,
"Frankfurt"-type sausage, cured ham
Vegetables and tubers: lettuce, tomato, potato,
carrot
Fresh fruits: apple, orange, banana
Milk and dairy products: whole and semi-
skimmed milk, yogurt, cheese I - low fat, cheese
II - medium fat, cheese III - extra fat
Cereals: French bread, pasta
Pulses: lentils
Industrial bakery: cookies
Eggs: hen eggs
Oils and fats: olive oil
For each food item, two composite samples were
prepared for analysis, where each composite sample
consisted of 24 individual units. Only edible parts of
each food item were included in the composites.
Meat; vegetables; grains;
fruits; dairy; fats/other
Dairy products: n = 2, DF NR, mean = 0.04 ng/g
t\v
Meat and meat products: n = 2, DF NR, mean =
<0.079 ng/g fw
Vegetables: n = 2, DF NR, mean = <0.37 ng/g fw
Tubers: n = 2, DF NR, mean = <0.095 ng/g fw
Fruits: n = 2, DF NR, mean = <0.096 ng/g fw
Eggs: n = 2, DF NR, mean = <0.1 ng/g fw
Milk: n = 2, DF NR, mean = <0.13 ng/g fw
Cereals: n = 2, DF NR, mean = <0.033 ng/g fw
Pulses: n = 2, DF NR, mean = <0.068 ng/g fw
Oils: n = 2, DF NR, mean = <0.038 ng/g fw
Industrial bakery: n = 2, DF NR, mean = <0.029
ng/g fw
(LOD not reported)
*n represents number of composite samples
Vestergren et al. (2012)
Sweden (Malmoe, Gothenburg, Uppsala, Sundsvall)
Purchasing locations of the two largest retail chains
in Sweden were selected in each of four major
Swedish cities. All purchases were made in
spring/summer of 1999, 2005, and 2010. In 2010,
the study was limited to the largest retail chains in
Uppsala located in close vicinity to Stockholm. An
equal amount of each food group from each of the
four cities was combined into one sample pool to
provide a representative sample for the Swedish
urban population.
Dairy; meat; grains; fruits;
vegetables; fats/other
Meat products:
1999: n = 1, point = 0.0071 ng/g
2005: n = 1, point = 0.0092 ng/g
2010: n = 1, point = 0.0058 ng/g (estimated)
(MDL = 0.0025 ng/g; MQL = 0.0062 ng/g)
Egg:
1999: n = 1, point = 0.0022 ng/g
2005: n = 1, point = 0.0056 ng/g (estimated)
2010: n = 1, point =
-------
FINAL
2024
Food Types
PFNA Results
(MDL = 0.0013 ng/g; MQL = 0.0032 ng/g)
Dairy products:
1999, 2005, 2010: n = 1 each year, point =
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
Noorlander et al. (2011)
The Netherlands
Food products randomly purchased from several
Dutch retail stores with nationwide coverage in
November 2009. Food samples were ground,
homogenized, and pooled for analysis. Food items
within each subcategory included the following:
Flour: whole wheat flour, flour
Pork: sausage, slice of bacon, pork chop, bacon,
minced meat rolled in bacon
Eggs: chicken eggs
Bakery products: cake, almond paste cake,
biscuits, brown spiced biscuit, pie
Vegetables/fruit: apple, orange, grape, banana,
onion, carrot, beet, chicory or leek, tomato,
cucumber, paprika, mushroom, cauliflower,
broccoli, white cabbage, red cabbage, brussel
sprout, spinach, endive, lettuce, French beans
Cheese: gouda cheese, edammer cheese, cheese
(>48% fat, less salt), cheese (>30% fat), brie
cheese
Beef: ground beef, beefburger, stewing steak,
braising steak, minced steak
Chicken/poultry: chicken leg, quarter chicken,
chicken filet, chicken burger, collared chicken
Butter: butter salt-free, salted, low-fat
Milk: half cream milk
Vegetable oil: margarine (solid/fluid), low-fat
margarine, frying fat (vegetable), frying oil
(vegetable), sunflower oil
Industrial oil: low-fat margarine, frying fat
(industrial), frying oil (industrial)
Meat; dairy; fruits and
vegetables; grains; fats/other
Butter: n = 1, point = 0.002 ng/g
Cheese: n = 1, point = 0.007 ng/g
Eggs: n = 1, point = 0.006 ng/g
Pork: n = 1, point = 0.002 ng/g
Beef: n = 1, point = 0.004 ng/g
Chicken/poultry: n = 1, point = 0.001 ng/g
Bakery products: n = 1, point = 0.001 ng/g
Vegetables/fruit: n = 1, point = 0.001 ng/g
Flour: n = 1, point = 0.015 ng/g
Milk: n = 1, <0.001 ng/g
Vegetable oil: n = 1, <0.0001 ng/g
Industrial oil: n = 1, <0.0003 ng/g
(LOD not reported)
*n represents number of composite samples
Jogsten et al. (2009)
Spain (Catalonia)
Food samples purchased from local markets, large
supermarkets, and grocery stores from two different
areas of Tarragona Province, Catalonia in January
and February 2008. The cities of Tarragona and
Reus were sampled in the northern area and
L'Ametlla de Mar and Tortosa in the southern area.
For each food item, two composite samples were
analyzed (one composite for the northern area and
one for the southern area). Each composite was
Fruits and vegetables; meat;
fats/other
Lettuce; raw, cooked, and fried meat (veal, pork, and
chicken); fried chicken nuggets; black pudding; lamb
liver; pate of pork liver; foie gras of duck;
"Frankfurt" sausages; common salt: n = 2 for each
food item, DF 0%
(LOD not reported)
*n represents number of composite samples
*Values of ND were replaced with l/2xLOD
C-66
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
formed of a minimum of six individual sub-samples
of the same product.
Ericson et al. (2008a)
Spain
Food samples purchased from local markets, large
supermarkets, and grocery stores within Tarragona
County in July 2006. Food samples were randomly
purchased with origin source not specified. Each of
the food samples were duplicated and combined to
analyze a composite sample. Composite samples
included the following:
Vegetables: lettuce, tomato, green bean, spinach
Pulses: lentils, beans, chickpeas
Cereals: rice, spaghetti, bread
Pork: sausage, hot dogs, steak, hamburger, ham
Chicken: breast, thighs, sausage
Veal: steak, hamburger
Lamb: steak
Dairy products: three different kinds of cheese,
yogurt, "petit-Swiss" creamy yogurt, cream
caramel, custard
Fruits: apple, orange, pear, banana
Oil: olive oil, sunflower oil, corn oil
Fats: margarine
Eggs
Meat; dairy; fruits; vegetables;
grains; fats/other
Vegetables (n = 2), pulses (n = 2), cereals (n = 2),
pork (n = 2), chicken (n = 2), veal (n = 2), lamb (n =
2), eggs (n = 2), dairy products (n = 2), whole milk
(n = 2), semi-skimmed milk (n = 2), fruits (n = 2),
margarine (n = 2), oil (n = 2): DF 0%
(LOD = 0.001-0.65 ng/g fw)
*n represents number of composite samples
Papadopoulou et al. (2017)
Norway
Participants of the A-TEAM project collected a
duplicate portion of all consumed foods and drinks,
prepared as for consumption, over two consecutive
weekdays. Only the samples collected in the first
day were analyzed. Sampling year not reported.
Solid foods: cereals and cereal
products, dairy products (not
milk), fish and seafood, meat
and meat products, sugar and
sugar products, fats and oils,
vegetables and nuts, fruits,
salty snacks, eggs, potatoes;
liquid foods: coffee, tea and
cocoa, milk, water, alcoholic
beverages, soft drinks
Solid foods:
n = 61, DF 2%, median (range) = 0 (0-0.001)
ng/g
(LOQ = not available)
Liquid foods:
n = 61, DF 16%, median (range) = 0 (0-0.00057)
ng/g
(LOQ = 0.00001 ng/g)
¦"Concentrations
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
forbade the use of anti-stick cookware. For each
canteen, lunch meals related to five school days
(from Monday to Friday) were weighed, pooled, and
homogenized. Beverages were not included in the
composites.
Ghellietal. (2019)
Italy
Egg samples were collected from commercial laying
hen farms in 2017. Sampling was based on
geographical origin of the eggs and rearing system
(e.g., organic, aviary system, battery cage and barn).
A total of 132 eggs were collected and eggs were
boiled. Four pools (containing three homogenized
yolks) were created for each of the following groups
(geographical origin, rearing system), for a total of
44 samples analyzed:
Group A: Pavia, barn
Group B: Verona, organic
Group C: Forli-Cesena, battery cage
Group D: Bologna, barn
Group E: Forli-Cesena, battery cage
Group F: Ravenna, aviary system
Group G: Ravenna, aviary system
Group H: Bologna, organic
Group I: Romagna, battery cage
Group L: Romagna, organic
Group M: Romagna, barn
Fats/other
Group A: n = 4, DF 0%
Group B: n = 4, DF 0%
Group C: n = 4, DFa 25%, range = ND-traces
Group D: n = 4, DFa 25%, range = ND-traces
Group E: n = 4, DF 0%
Group F: n = 4, DF 0%
Group G: n = 4, DF 0%
Group H: n = 4, DF 0%
Group I: n = 4, DF 0%
Group L: n = 4, DF 0%
Group M: n = 4, DF 0%
(LOD = 0.1 ng/g for all PFAS; LOQ = 0.25 ng/g for
all PFAS)
"Traces defined as value between LOD and LOQ
Johansson et al. (2014)
Sweden
Eggs from 20 to 25 producers were collected each
year from 1999 to 2010 within the Swedish National
Food Agency's official food control program. Each
sample consisted of a pool of 10-12 eggs from one
producer. The pooled samples comprised eggs from
both conventional and organic production.
Information on the number of organic eggs sampled
was not available.
Fresh milk was sampled from the tanks of milk
transport vehicles between 1999 and 2009 as part of
the food control program. The tanks generally
contained milk from ten dairy farms. In 2010, milk
samples were taken from the milk storage tanks on
individual dairy farms. Between 10-25 milk samples
Dairy; fats/other
Hen's eggs:
Total: n = 36, DFa 28%, range = ND-0.143 ng/g
fw
1999: n = 3, DFa 33%, range = <0.020-0.062
ng/g fw
2000: n = 3, DFa 33%, range = <0.020-0.025
ng/g fw
2001: n = 3, DFa 100%, mean3 (range) = 0.052
(0.020-0.075) ng/g fw
2002: n = 3, DFa 33%, range = <0.020-0.143
ng/g fw
2003: n = 3, DFa 33%, range = <0.020-0.087
ng/g fw
2004: n = 3, DFa 33%, range = <0.020-0.020
ng/g fw
C-68
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
were collected each year. The milk samples were
extracted in two different batches.
2005: n = 3, DF 0%
2006: n = 3, DF 0%
2007: n = 3, DFa 33%, range = <0.020-0.020
ng/g fw
2008: n = 3, DF 0%
2009: n = 3, DF 0%
2010: n = 3, DFa 33%, range = <0.020-0.026
ng/g fw
(MDL = 0.020 ng/g fw; MQL = 0.097 ng/g fw)
Cow's milk (1st batch):
Total: n = 18, DF 0%
(MDL = 0.0073 ng/g fw; MQL = 0.0156 ng/g fw)
Cow's milk (2nd batch):
Total: n = 18, DF 0%
(MDL = 0.0073 ng/g fw; MQL = 0.0156 ng/g fw)
Eriksson et al. (2013)
Denmark (Faroe Islands)
Locally produced food items sampled in 2011-2012.
Packaged dairy products were supplied by Faroe
Islands, Meginfelag BunaSarmanna - dairy products
included samples of milk, low fat (0.5%), semi-
skimmed (1.5%), yoghurt with banana and pear
(3.4% fat), low fat (0.9%) plain yoghurt, and creme
fraiche (18% fat). Yoghurt with banana and pear
was sampled from two production batches, and the
low fat plain yoghurt and creme fraiche was
sampled from one production batch. Potatoes were
sampled from two different farms.
Dairy; fruits and vegetables
Milk:
Farmer (Innan Glyvur): n = 1, point = 0.061
ng/g ww
Farmer (Havnardal): n = 1, point = 0.073 ng/g
WW
Diary (Faroe Island): n = 4, DFa 75%, range =
<0.048-0.058 ng/g ww
Yogurt (n = 3), creme fraiche (n = 1), potato (n = 2):
DF 0%
(LOD = 0.0048 ng/L for milk; LOD = 0.0014 ng/g
for dairy; LOD = 0.0017 ng/g for potato)
Still etal. (2013)
Germany
Fourteen commercially available samples of various
dairy products and raw milk were provided from a
cooperating dairy.
Dairy
Milk products:
Raw milk: n = 1, point = <0.0016 ng/g
Fresh milk: n = 1, point = <0.0016 ng/g
Fresh whole milk: n = 1, point = <0.0016 ng/g
UHT milk: n = 1, point = 0.0034 ng/g
Yogurt:
Yogurt (0.1%) fat): n= 1, point = <0.0016 ng/g
Yogurt (3.8%o fat): n = 1, point = <0.0016 ng/g
Cheese:
Semihard cheese: n = 1, point = 0.0094 ng/g
Semisoft cheese: n = 1, point = 0.0062 ng/g
Soft cheese: n = 1, point = 0.0081 ng/g
C-69
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
Other dairy products:
Whey drink: n = 1, point = <0.0016 ng/g
Butter milk: n = 1, point = <0.0016 ng/g
Cream yogurt: n = 1, point = <0.0016 ng/g
Cream: n = 1, point = <0.0016 ng/g
Butter: n = 1, point = 0.0047 ng/g (between
LOQ and LOD)
(LOD = 0.0016 ng/g)
Vestergren et al. (2013)
Sweden (Karsta)
Study was conducted at a dairy cattle farm that was
selected to represent a background contaminated
agricultural area with no known point sources of
PFAS in the proximity. The farm had no history of
sewage sludge application to the pasture land. Milk
samples were collected between November 2010
and April 2011 from a milk tank, where milk from
the entire farm is stored after milking. Muscle, liver,
and whole blood samples were obtained from five
individual cows from the slaughterhouse on two
different occasions (April and June 2011).
Meat; dairy
Milk: n = 6, DF NR, mean = 0.0023 ng/mL
Liver: n = 5, DF NR, mean = 0.016 ng/g
Blood: n = 5, DF NR, mean = 0.014 ng/g
Muscle: n = 5, DF NR, mean = 0.0045 ng/g
(MDL = 0.0012 ng/mL in milk; 0.0023 ng/g in liver;
0.0020 ng/g in blood; 0.0013 ng/g in muscle)
Falandysz et al. (2006)
Poland
Eider duck samples were collected from the Gulf of
Gdansk in the Baltic Sea (south coast of Poland) in
February 2003.
Meat
n = 16, DF NR, mean, median (range) = 0.4, 0.32
(0.3-0.9) ng/mL
(LOD not reported)
*Values reported for animal whole blood samples
Lankova et al. (2013)
Czech Republic
Breastmilk samples were obtained from 50 women
living in the Olomouc region from April to August
2010. The age of participating mothers ranged from
20 to 43 years.
Six types of infant formula from the Czech retail
market were also examined: one powdered formula
for infants, two formulas for toddlers, one formula
for babies with lactose intolerance, one formula for
premature babies, and one soya-based formula for
babies with non-milk diets. Sampling year not
provided.
Fats/other; breastmilk
Breastmilk:
n = 50, DF (frequency of quantification) 48%,
range = <0.006-0.015 ng/mL
(LOQ = 0.006 ng/mL)
Infant formula:
n = 6, DF NR, maximum = 0.011 ng/g
(LOQ = 0.009 ng/g)
Llorca et al. (2010)
Spain (Barcelona)
Breastmilk; grains; fats/other
Breastmilk:
n = 20, DF 0% (frequency of quantification)
(MLOD = 0.0035 ng/mL; MLOQ = 0.0115 ng/mL)
C-70
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
Individual breastmilk samples were collected from
20 women in Barcelona city in 2008. All samples
were collected within 40 days postpartum.
Samples of powdered infant formula (three brands)
and dry cereal baby food (two brands) were
collected from retail stores (sample year not
reported).
Baby cereals:
n = 2, DFa 100%, mean3 (range) = 0.091 (0.044-
0.138) ng/g
(MLOD = 0.017 ng/g; MLOQ = 0.0575 ng/g)
Milk infant formulas:
n = 3, DFa 100%, mean3 (range) = 0.166 (0.118-
0.219) ng/g
(MLOD = 0.012 ng/g; MLOQ = 0.039 ng/g)
Abdallah et al. (2020)
Ireland (Dublin)
Breastmilk samples obtained from mothers recruited
from breastfeeding clinics at two Irish maternity
hospitals. Mothers provided samples between 3 and
8 weeks postpartum. Mothers were up to 41 years of
age, primiparas, in good health, and exclusively
feeding one infant. Sampling year not reported.
Breastmilk
n = 16, DF 69%, mean, median (range) = 0.026,
0.014 (<0.01-0.1) ng/mL
(LOQ = 0.01 ng/mL)
*Values
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
of the Mediterranean Sea due to mining activities.
Samples were collected a few weeks postpartum.
*Range reported only for detected values
Antignac et al. (2013)
France (Seine-Saint Denis, Ardeche, Isere, Loire,
Savoie counties)
Breastmilk samples collected from mothers
participating in the ELFE pilot study. Sampling year
not reported, though all mothers gave birth in
October 2007. Mothers were contacted by phone
one month after leaving the maternity and provided
with instructions on breastmilk collection. Milk
samples could be collected during several lactation
sessions. On average, 15 aliquot samples of 10 mL
were collected for each participant and pooled into
one sample for analysis.
Breastmilk
n = 48, DF 2%, range = <0.05-0.064 ng/mL
(LOD = 0.05 ng/mL)
Croes et al. (2012)
Belgium (Flanders)
Breastfeeding mothers were recruited from 9
maternities in 24 rural communities in East and
West Flanders and Flemish Brabant in May 2009 -
June 2010. Breastmilk samples were collected
between two and eight weeks after delivery and a
subset was analyzed for perfluorinated compounds.
Breastmilk
n = 40, DF 42.5%, median (10th-90th percentile) =
-------
FINAL
2024
Study
Location and Source
Food Types
PFNA Results
collected between the fourth and fifth day after
delivery.
Pratt etal. (2013)
Ireland
Pooled breastmilk samples were collected from 109
first-time mothers at four centers across Ireland.
Sampling year not reported.
Breastmilk
n = 11, DF 0%
(LOD = 0.5-5 ng/mL for all PFAS)
*n represents number of pooled samples
Karrman et al. (2010)
Spain (Catalonia)
Breastmilk samples were collected from healthy
primiparae mothers aged 30-39 years who lived in
Tarragona County for at least the last five years.
Babies were aged 41-60 days when milk samples
were collected in 2007.
Breastmilk
n = 10, DF 0%
(LOQ = 0.03 ng/mL)
Notes'. DF = detection frequency; LOD = limit of detection; LOQ = limit of quantitation; 0.5 x, lx; or 2* = 54, 1, or 2 times the agronomic rate of biosolids application to meet
nitrogen requirements of the crop; MDL = method detection limit; NR = not reported; PFAA = perfluoroalkyl acids; RL = reporting limit; WWTP = wastewater treatment plant.
Bold indicates detected levels of PFNA in food.
a The DF and/or mean was not reported in the study and was calculated in this synthesis. Means were calculated only when DF = 100%.
C-73
-------
FINAL
2024
C.3.2. Food Contact Materials
In 2011, FDA reached a voluntary agreement with industry to remove from the market certain
PFAS grease-proofing agents used in fast food packaging. As such, the occurrence of PFNA in
fast food packaging in the U.S. may be declining over time. The EPA identified two studies
reporting PFNA in food contact materials (FCM) conducted in the U.S. Liu et al. (2014)
analyzed the occurrence of PFAS in treated food contact paper and other consumer products
purchased from local retailers and online stores in the United States between March 2007 and
September 2011. All treated food contact paper was manufactured in the United States. PFNA
was detected in 33% of samples (n = 9), with two of the detects below 10 ng/g and the third
detect at 212 ng/g. Sinclair et al. (2007) sampled microwave popcorn bags purchased in 2005 in
New York City that may have originated from international retailers. PFNA was detected in one
of three brands of microwave popcorn bags, both before and after cooking.
Several peer-reviewed studies conducted in Europe also monitored for PFNA in food contact
materials. PFNA was not detected in any FCMs including paper packaging, cardboard, coated
bakery release papers for oven baking, paper filters for coffee, microwave popcorn bags, and an
ice cream tub in two studies (Vavrous et al., 2016; Moreta and Tena, 2013). Two additional
studies had the majority of samples reported as
-------
FINAL
2024
Table C-6. Summary of Studies Reporting the Occurrence of PFNA in Food Contact Materials
Study
Location
Site Details
Results
United States
Liuetal. (2014)
United States (unspecified)
Treated food contact papers, including microwave
cooking bags, were purchased between October 2007
and September 2011 from local retailers and online
stores in the United States. All products originated
from the United States. A total of nine samples were
purchased.
N = 9, DFa 33%, range = BDL-212 ng/g
(DL not reported)
Europe
Vavrous et al. (2016)
Czech Republic
Real samples of paper FCM (11 with direct food
contact and 4 with indirect food contact) were
acquired from a market. Samples included paper
packages of wheat flour (n = 2), paper bags for bakery
products (n = 2), sheets of paper for food packaging in
food stores (n = 2), cardboard boxes for packaging of
various foodstuffs (n = 3), coated bakery release
papers for oven baking at temperatures up to 220°C (n
= 3), and paper filters for coffee preparation (n = 3).
Sampling year and country of origin for products not
reported.
N = 15, DF 0%
(LOQ = 0.0047 mg/kg)
Kotthoff et al. (2015)
Germany (Schmallenberg)
Thirty-three random samples of recent individual
paper-based FCMs collected in the first until the third
quarter of 2010 in Germany. Individual samples were
bought from local retailers or collected by coworkers
of the involved institutes. Sampled products spanned
all quality levels from entry level to cutting edge
products. The age of the samples ranged from a few
years to decades. Country of origin not reported.
"Archived" older samples of FCMs (baking paper
purchased before 2010) were collected from the staff
of the institutes. The age of these samples ranged from
a few years to decade. Country of origin not reported.
Recent samples: n = 33, DF 24%, median
(range) =
-------
FINAL
2024
Study
Location
Site Details
Results
Fraunhofer-Institut IVV. The authors estimated the
institute-acquired items were purchased between pre-
2000 to 2006. Two scenarios were tested: an
"overheating scenario" for the pans (250-370°C) and
a "normal application" scenario for 1 hour for all other
food contact materials (185-221°C). Each product was
tested 2-3 times.
Waffle irons: n = 9, DFa 22%, range =
-------
FINAL
2024
Study
Location
Site Details
Results
wrappers, and paper boxes were collected in Athens
from October to December 2012 from popular Greek
fast food chain restaurants, coffee shops, and
multiplex cinemas. Other FCMs (muffin cups, baking
papers, and microwave popcorn and rice bags) were
purchased from large supermarkets. All products
except for microwave popcorn and rice bags were
manufactured in Greece. Sampled packaging materials
included unused items and used items (i.e., contained
food products).
A microwave popcorn bag was also analyzed before
and after cooking.
Paper materials for baking: n = 2, DF 0%
Microwave bags: n = 3, DF 0%
Before cooking: n = 1, point =
-------
FINAL
2024
C.3.3. Consumer Products
Since the 1950's, PFNA has been used in industrial and consumer products, including fabric and
carpet protective coatings, paper coatings, insecticide formulations, and surfactants (NCBI,
2022b). PFNA and other long-chain PFAS are found in aqueous film-forming foams, cosmetics,
dental floss, floor polish, leather, food packaging materials, lithium batteries, ski wax, treated
apparel, work apparel for medical staff, pilots, and firefighters, and in hair treatment products
(NCBI, 2022b). Based on limited testing, PFNA has been detected in rinsates from fluorinated
high-density polyethylene (HDPE) containers used by one pesticide product supplier (USEPA,
2022a). PFNA is not a registered pesticide under the Federal Insecticide, Fungicide, and
Rodenticide Act (FIFRA), and the EPA does not set a 40 CFR Part 180 pesticide tolerance in
food and feed commodities for PFNA (US GPO, 2022). Maximum residue levels for PFNA were
not found in the Global Maximum Residue Level Database (Bryant Christie Inc, 2022).
Two studies based in the U.S. were identified that analyzed PFNA concentrations in a range of
consumer products, including children's nap mats, household carpet/fabric-care liquids, and
textiles (Zheng et al., 2020; Liu et al., 2014) (Table C-7). Of these two studies, the consumer
products evaluated are likely used by adults (e.g., floor waxes), can come into contact with both
adults and children (e.g., treated upholstery), or the user was not specified (e.g., clothing). Zheng
et al. (2020) determined the occurrence of ionic and neutral PFAS in the childcare environment
(dust and nap mats). Samples of children's nap mats were collected from seven Seattle childcare
centers (n = 26; 20 polyurethane foam, 6 vinyl cover samples). PFNA was detected in 36% of
nap mat samples with a mean concentration of 0.19 ng/g. Half of the analyzed mats were
purchased as new products and the other half were used. The authors reported that total PFAS
levels in the new versus Used mats were not significantly different. Total PFAS levels in mat
foam versus Mat covers were also similar. Based on these results, the authors suggested that
indoor air was not the major source of PFAS in mats and that PFAS in mats could be associated
with the manufacturing process. Liu et al. (2014) analyzed the occurrence of PFAS in consumer
products (including pretreated carpeting, commercial carpet-care liquids, household
carpet/fabric-care liquids, treated apparel, treated home textiles and upholstery, treated non-
woven medical garments, treated floor waxes and stone-wood sealants, membranes for apparel,
and thread-sealant tapes and pastes) purchased between March 2007 and September 2001 from
local retailers and online stores in the United States. The consumer products originated from the
United States, England, Vietnam, China, Thailand, El Salvador, Bangladesh, Dominican
Republic, Malaysia, and Indonesia. PFNA was detected in 44% of nine pretreated carpeting
samples (ranging from below the detection limit (BDL) to 236 ng/g); in 58% of 12 commercial
carpet/fabric-care liquid samples (BDL-8,860 ng/g); in 15% of 13 household carpet/fabric-care
liquid and foam samples (BDL-37.3 ng/g); in 60% of 15 treated apparel samples (BDL-
235 ng/g); in 100% of six treated home textile and upholstery samples with a mean of 42.6 ng/g;
in 56%) of nine treated non-woven medical garment samples (BDL-334 ng/g); in 88% of eight
treated floor wax and stone/wood sealant samples (BDL-2,740 ng/g); and in 75% of eight
membranes for apparel samples (BDL-12.8 ng/g). PFNA was not detected in thread-sealant
tapes and pastes (n = 6). Detection limits were not reported in the study.
Results for all of the identified non-U.S. studies are presented in detail in Table C-7 and are
summarized here. One study from Germany did not detect PFNA in cleaners and wood glue, but
among other consumer products such as nanosprays/impregnation sprays, gloves, and ski wax
C-78
-------
FINAL
2024
found medians ranging from 2.8-10.7 ng/g (Kotthoff et al., 2015). Additional consumer products
from this study found outdoor textiles, carpets, leather, and awing cloth median concentrations
ranging between
-------
FINAL
2024
Table C-7. Summary of PFNA Consumer Product Data
Study
Location
Site Details
Results
United States
Zheng et al. (2020)
United States (Seattle,
Washington)
Children's nap mat samples (n = 26, finely cut) from
seven Seattle childcare centers, including polyurethane
foam (n = 20) and vinyl cover (n = 6) samples.
Sampling year not reported.
N = 26, DF 36%, mean, median (range) =
0.19,0.11 (ND-0.65)ng/g
(MDL = 0.08 ng/g)
Liuetal. (2014)
United States (unspecified)
Consumer products commonly used indoors were
purchased between March 2007 and September 2011
from local retailers and online stores in the United
States. The samples analyzed for PFCAs included pre-
treated carpeting, commercial carpet/fabric-care
liquids, household carpet/fabric-care liquids and
foams, treated apparel, treated home textile and
upholstery (i.e., mattress pads), treated non-woven
medical garments, treated floor waxes and stone-wood
sealants, membranes for apparel, and thread-sealant
tapes and pastes. The products originated from the
United States, England, Vietnam, China, Thailand, El
Salvador, Bangladesh, Dominican Republic, Malaysia,
and Indonesia.
Pre-treated carpeting: n = 9, DFa 44%, range =
BDL-236 ng/g
Commercial carpet/fabric-care liquids: n = 12,
DFa 58%), range = BDL-8,860 ng/g
Elousehold carpet/fabric-care liquids and
foams: n = 13, DFa 15%, range = BDL-37.3
ng/g
Treated apparel: n = 15, DFa 60%, range =
BDL-235 ng/g
Treated home textile and upholstery: n = 6,
DFa 100%, meana (range) = 42.6 (3.80-213)
ng/g
Treated non-woven medical garments: n = 9,
DFa 56%o, range = BDL-334 ng/g
Treated floor waxes and stone-wood sealants:
n = 8, DF 88%o, range = BDL-2,740 ng/g
Membranes for apparel: n = 8, DFa 75%o, range
= BDL-12.8 ng/g
Thread-sealant tapes and pastes: n = 6, DFa
0%
(DL not reported)
Europe
van der Veen et al. (2020)
Sweden (unspecified)
Samples of durable water repellent outdoor clothing
collected from six suppliers from the outdoor textile
industry in Sweden (one pair of outdoor trousers, six
jackets, and six fabrics for outdoor clothes*). Each
sample was cut into two pieces - one exposed to
elevated UV radiation, humidity, and temperature for
300 hours (assumed lifespan of outdoor clothing) and
one untreated (not aged). Sampling year not reported.
Year of manufacturing not reported for nine of the 13
samples; the remaining four samples (samples 4-7)
Point values presented as before aging (n =1),
after aging (n = 1)
Sample 1 (outdoor trousers): ND, 1.3 (ig/m2
Sample 2 (fabric for jacket): 0.05 |ig/m2,
0.14 (ig/m2
Sample 3 (fabric for jacket): ND, 0.13 (ig/m2
Sample 4 (men's jacket): ND, ND
Sample 5 (men's jacket): 0.08 |ig/m2, 12
lig/m2
C-80
-------
FINAL
2024
Study
Location
Site Details
Results
reported a manufacturing year of 2012/2013. Country
of origin not reported.
*The breakdown of the 13 items of outdoor clothing is
reported differently in Section 2.2 and Table 1 of the
article. Section 2.2 reports one pair of outdoor
trousers, seven jackets, four fabrics for outdoor
clothes, and one outdoor overall. Table 1 shows one
pair of outdoor trousers, six jackets, and six fabrics for
outdoor clothes
Sample 6 (fabric for outdoor clothes): 0.29
(ig/m2, 100 (ig/m2
Sample 7 (children's jacket): ND, ND
Sample 8 (jacket): 1.1 |ig/m2, 3.5 (ig/m2
Sample 9 (fabric for outdoor clothes): ND,
0.63 (ig/m2
Sample 10 (fabric for outdoor clothes): ND,
0.15 (ig/m2
Sample 11 (fabric for outdoor clothes): ND,
0.49 (ig/m2
Sample 12 (fabric for outdoor clothes): ND,
0.11 (ig/m2
Sample 13 (fabric for outdoor clothes): ND,
ND
(LOD = 0.02-0.1 (ig/m2 for ionic PFAS)
Schultes et al. (2018)
Sweden (unspecified)
Thirty-one cosmetic products from five product
categories (moisturizing cream, foundation, eye
pencil, powder and eye shadow, shaving foam)
purchased from the Swedish market in 2016-2017.
Cosmetic products were selected based on (i) the 2015
KEMI survey which reported the most frequently
reported PFAS in cosmetic products and (ii) a
database of ingredient lists compiled by the Swedish
Society for Nature Conservation. Twenty-four
products listing nine different PFAS as active
ingredients were purchased. In addition, seven
products which did not list PFAS in their ingredients
were also purchased from the same stores as control
samples. Year of manufacture and country of origin
not reported.
Control:
Moisturizing cream: n = 1, DF 0%
Foundation: n = 3, DF 0%
Powder and eye shadow: n = 2, DF 0%
Shaving cream: n = 1, DF 0%
PFAS-containing:
Moisturizing cream: n = 6, DF 0%
Foundation: n = 6, DFa 16%, range = <3.45-
651 ng/g
Eye pencil: n = 1, DF 0%
Powder and eye shadow: n = 10, DFa 30%,
range = <3.45^17.2 ng/g
Shaving cream: n = 1, DF 0%
(LOD = 3.45 ng/g)
Favreau et al. (2016)
Switzerland (national)
Liquid consumer products, including impregnation
agents, cleansers, polishes, lubricants, miscellaneous
items, and commercial AFFFs purchased in 2012 and
2013 from stores and supermarkets throughout
Switzerland. Products were purchased from 82
different producers and were selected based on their
susceptibility to contain PFAS according to previous
screenings. Miscellaneous "other" products included
foam-suppressing agents for the chromium industry,
paints, ski wax, inks, and tanning substances. AFFFs
were divided into two sets based on the sampling
source. AFFF set 1 was derived from stock solution in
Impregnation products: n = 60, DF 5%, mean,
median (range) = 800, ND (100-1,900) ng/g
Cleansers: n = 24, DF 0%
Polishes: n = 18, DF 0%
Others: n = 23, DF 0%
AFFF set 1: n = 27, DF 70%, mean, median
(range) = 3,400,200 (100-37,900) ng/g
AFFF set 2: n = 35, DF 3%, mean, median
(range) = 200, ND (200-200) ng/g
C-81
-------
FINAL
2024
Study
Location
Site Details
Results
fire installation of industrial sites storing chemicals
and petroleum products and samples may be the result
of multiple AFFF fillings over the years (1990-2010
was the last documented filling date). AFFF set 2
came from commercially available AFFFs between
2012 and 2013 from six producers.
(LOQ = 0.5 ng/mL)
*Mean and range values only include samples
where PFNA was detected
*ND treated as 0 for median calculations
Kotthoff et al. (2015)
Germany (Schmallenberg)
Forty-nine random samples of consumer products
collected in the first until the third quarter of 2010 in
Germany, including outdoor textiles, carpets, cleaning
agents, impregnating agents, leather samples, and ski
waxes. Individual samples were bought from local
retailers or collected by coworkers of the involved
institutes or local clubs (e.g., ski waxes from local
skiing club). Sampled products spanned all quality
levels from entry level to cutting edge products. The
age of the samples ranged from a few years to
decades. Country of origin not reported.
Cleaner: n = 6, DF = 0%
Wood glue: n = 1, DF = 0%
Nanosprays and impregnation sprays: n = 3,
DF = 56%, median (maximum) = 2.8 (8.0)
ng/g
Outdoor textiles: n = 3, DF = 67%, median
(maximum) = 1.0 (8.3) (ig/m2
Carpet: n = 6, DF = 20%, median (maximum)
=
-------
FINAL
2024
Study
Location
Site Details
Results
Vestergren et al. (2015)
Norway (Tromse, Trondheim)
Samples of furniture textile (samples included baby-
related items such as baby mattress, baby blanket, and
baby bed cover), carpet, and clothing samples were
purchased from three major retail stores during
November 2012-February 2013. Sampling campaign
designed to evaluate consumer products in product
categories that were previously found to contain PFAS
residuals and that were representative of products
imported from China in large quantities. Individual
products randomly selected without prior knowledge
of surface treatment with PFAS. Outdoor clothing was
excluded. Year of manufacture not reported.
Furniture textiles: n = 27, DFa = 15%, range =
<0.005-0.097 (ig/m2
Carpets/mats: n = 9, DFa = 33%, range =
<0.005-0.077 (ig/m2
Cotton/leather clothes: n = 4, DFa = 50%,
range = <0.005-0.008 (ig/m2
(MDL = 0.005 (ig/m2)
Lfoisaster et al. (2019)
Norway (unspecified)
AFFF concentrate, containing PFOS as the main
PFAS, from the same supplier as assumed to have
been used historically at a firefighting training facility
where AFFF containing PFOS was used extensively
from the early 1990s until 2001 when it was phased
out and replaced by fluorotelomer containing AFFF
until 2011. Concentrations reported in 1:100 diluted
AFFF. Number of samples not reported but assumed
to be 1. Sampling year assumed to be 2016 based on
when samples were collected for groundwater and
soil.
1:100 diluted foam:
n= 1, point = 3,100 ng/L
Relative contribution to total PFAS = <0.1%
(LOQ = 0.3 ng/L)
Dauchy et al. (2017)
France (unspecified)
Nine firefighting foam concentrates were provided by
a professional user and included alcohol-resistant film-
forming fluoroprotein foams (n = 5), alcohol-resistant
aqueous film-forming foams (n = 2), film-forming
fluoroprotein foams (n = 1), and fluorine-free foams (n
= 1). These concentrates were manufactured after
2002 by four different manufacturers. Concentrate
sampling year was not reported, though water
sampling in the same study was conducted in
November 2014.
N = 9, DF (frequency of quantification) 0%
(LOQ = 5,000 ng/L)
Laitinen et al. (2014)
Finland (Oulu)
Sthamex 3% AFFF liquid, manufactured in Germany
and available commercially in Finland, used by
firefighters during training in the simulation of aircraft
accidents. Samples collected in 2010.
N = NR., DF (frequency of quantification) 0%
*Low concentrations of PFNA were detected,
but were below LOQ
(LOQ = 20 ng/mL)
Orijjin Unspecified
Becanova et al. (2016)
Not specified
One hundred twenty-six samples of (1) household
equipment (textiles, floor coverings, electrical and
electronic equipment (EEE), and plastics; includes
Household equipment: n = 55, DFa 2%, range
=
-------
FINAL
2024
Study
Location
Site Details
Results
children-related items such as teddy bear filling, teddy
bear cover, and plush); (2) building materials (oriented
strand board, other composite wood and wood,
insulation materials, mounting and sealant foam, I
materials, polystyrene, air conditioner components);
(3) car interior materials; and (4) wastes of electrical
and electronic equipment (WEEE) purchased (for new
materials) or collected from various sources (for older
and used materials). Production year ranged from
1981 to 2010. Origin of production and location and
year of purchase/collection not reported.
Building materials: n = 54, DFa 0%
Car interior materials: n = 10, DFa 0%
WEEE: n = 7, DFa 14%, range =
-------
FINAL
2024
C.3.4. Indoor Dust
In a Wisconsin Department of Health Services study, Knobeloch et al. (2012) examined levels of
16 perfluoroalkyl chemicals in vacuum cleaner dust from 39 Wisconsin homes across 16
counties in March and April 2008 (Table C-8). Samples from these homes built between 1890
and 2005 were collected during a pilot study to assess residential exposure to persistent
contaminants found in the Great Lakes Basin. PFNA was found in all samples at a median
concentration of 12 ng/g. The number of rooms with synthetic, wall-to-wall carpeting and the
square footage of the homes were both significantly positively correlated with dust
concentrations of PFNA. Based on the results of this study, the authors suggested that
perfluoroalkyl chemicals may be ubiquitous contaminants in U.S. homes. In an EPA study of
112 indoor dust samples collected from vacuum cleaner bags from homes and daycare centers in
North Carolina and Ohio in 2000-2001 (EPA's Children's Total Exposure to Persistent
Pesticides and Other Persistent Organic Pollutants (CTEPP) study), samples were collected from
102 homes and 10 daycare centers in North Carolina (49 homes, 5 daycare centers) and Ohio (53
homes, 5 daycare centers) (Strynar and Lindstrom, 2008). Results were not reported separately
for homes and daycares. Overall, PFNA was detected in 42.9% of all samples (n = 112) with
mean and median concentrations of 22.1 ng/g and 7.99 ng/g, respectively. The authors concluded
that the study measured perfluorinated compounds in house dust at levels that may represent an
important pathway for human exposure.
Additional peer-reviewed studies based in the U.S. were identified that evaluated the occurrence
of PFNA and other PFAS in dust of indoor environments, primarily in homes, as well as in
schools, childcare facilities, offices, and vehicles (Zheng et al., 2020; Scher et al., 2019; Byrne et
al., 2017; Karaskova et al., 2016; Wu et al., 2014; Fraser et al., 2013; Kato et al., 2009) (Table C-
8). For those studies with results stratified for U.S. homes, PFNA levels and detection
frequencies were lowest in a study of remote Alaska Native villages (35% detection, median
below 0.2 ng/g), while in other U.S. locations, PFNA was detected in at least 65% of samples
(some studies reporting 100% detection) at widely varying mean and median levels across the
studies (from approximately 4 ng/g to 70 ng/g). Few studies sampled childcare centers, vehicles,
and offices, and none of the reviewed studies reported measurements in other microenvironments
(e.g., public libraries, universities).
Several studies reported results from dust samples collected only from homes (Scher et al., 2019;
Byrne et al., 2017; Wu et al., 2014), with one study sampling from locations near a PFAS
production facility. Scher et al. (2019) evaluated indoor dust in 19 homes in Minnesota within a
groundwater contamination area (GCA) in the vicinity of a former 3M PFAS production facility.
Homes within the GCA had previous or ongoing PFAS contamination in drinking water and
were served by the Oakdale, Minnesota PWS or a private well previously tested and shown to
have detectable levels of PFOA or PFOS. In the house dust samples, collected from July to
September 2010, the detection frequencies for PFNA were 68% and 95% for entryways to the
yard and interior living spaces such as the family or living rooms, respectively (n = 19 each),
with median concentrations of 9.7 ng/g and 26 ng/g, respectively. PFAS concentrations in both
sampling locations were higher than corresponding soil concentrations, suggesting that interior
sources were the main contributors to PFAS in house dust.
C-85
-------
FINAL
2024
Byrne et al. (2017) assessed exposure to PFNA and other PFAS among residents of two remote
Alaska Native villages on St. Lawrence Island. PFAS concentrations were measured in dust
collected from the surfaces of floors and furniture of 49 homes on St. Lawrence Island during
February-April of 2013 and 2014. Residents were asked not to sweep or dust for one week prior
to sampling. The authors described the overall PFAS levels in dust samples as "on the lower end
of those reported worldwide in other studies." PFNA was detected in 35% of all samples (n = 49)
with a median value below the LOD (0.1 ng/g-0.2 ng/g). Wu et al. (2014) measured
concentrations of five PFCs in residential dust in California in 2008-2009. Dust samples were
collected from the carpet or area rug in the main living area of the home. Homes of parents with
young children and homes with older adults were differentiated to characterize the relationship
between serum concentrations of PFCs and several other factors, including PFC concentrations
in residential dust. PFNA was detected in 65% of samples from households with young children
in Northern California (n = 82), with mean and median concentrations of 67.4 ng/g and
9.70 ng/g, respectively. PFNA was detected in 72% of samples from households of older adults
in central California (n = 42), with mean and median concentrations of 58.5 ng/g and 11.85 ng/g,
respectively.
Apart from the information reported by Strynar and Lindstrom (2008), one other study included
childcare centers in the locations sampled (Zheng et al., 2020). Zheng et al. (2020) collected dust
samples from seven childcare centers in Seattle, Washington (n = 14) and one childcare facility
in West Lafayette, Indiana (n = 6 across six rooms); the sampling year was not reported. The
included childcare facilities consisted of several building types, including multiple classrooms, a
former church, and a former home. Because centers were vacuumed and mopped daily, dust
samples were obtained from elevated surfaces (shelving, tops of bookcases/storage cubbies)
along with floor dust. PFNA was detected in all samples at mean and median concentrations of
3.2 ng/g and 1.7 ng/g, respectively.
One study evaluated PFNA levels in vehicles and offices, in addition to homes. Fraser et al.
(2013) collected dust samples between January and March 2009 from three microenvironments
of 31 individuals in Boston, Massachusetts (offices (n = 31), homes (n = 30), and vehicles with
sufficient dust for analysis (n = 13)). Study participants worked in separate offices located across
seven buildings, which were categorized as Building A (n = 6), Building B (n = 17), or Other
(n = 8). Building A was a newly constructed (approximately one year prior to study initiation)
building with new carpeting and new upholstered furniture in each office; Building B was a
partially renovated (approximately one year prior to study initiation) building with new carpeting
throughout hallways and in about 10% of offices. The other buildings had no known recent
renovation occurred. Study offices were not vacuumed during the sampling week and
participants were asked not to dust or vacuum their homes and vehicles for at least one week
prior to home sampling. PFNA was detected in 94%, 67%, and 85% of office, home, and vehicle
dust samples, respectively, with geometric mean concentrations of 63.0 ng/g, 10.9 ng/g, and
14.7 ng/g, respectively. Geometric mean PFNA concentrations were statistically significantly
higher in offices compared to homes and vehicles. The study also observed that PFNA
concentration in house dust was significantly predictive of PFNA serum concentration.
Two studies evaluated dust samples collected across multiple continents (Karaskova et al., 2016;
Kato et al., 2009). Karaskova et al. (2016) examined PFAS levels in house dust collected
between April and August 2013 from the living rooms and bedrooms of 14 homes in the United
C-86
-------
FINAL
2024
States, 15 homes in Canada, and 12 homes in the Czech Republic (locations unspecified). PFNA
was detected in all U.S. samples (n = 20) at mean and median concentrations of 10.9 ng/g and
3.9 ng/g, respectively. The authors reported PFNA concentrations were significantly higher in
North America compared to the Czech Republic, which they indicated may suggest a faster shift
from long-chain PFAS to their shorter-chain homologues in Europe than in North America.
Overall, no significant differences in total PFAS concentrations were found between the
bedroom and living room in the same household although significant relationships were found
based on type of floors, number of residents, and age of the house. A second multicontinental
study (Kato et al., 2009) measured PFC concentrations in 39 household dust samples collected in
2004 from homes in the United States (Atlanta, GA) (n = 10), United Kingdom (n = 9), Germany
(n = 10), and Australia (n = 10). Across all 39 homes, PFNA was detected in 25.6% of samples
with a median concentration below the LOQ (2.6 ng/g). The authors did not report stratified
PFNA data by country.
Results for all of the identified non-U.S. studies are presented in detail in Table C-8 and are
summarized here. One study conducted in Canada evaluated dust samples in homes by using the
vacuum cleaners of participants and found PFNA detected in 69% of samples (n = 48) with a
mean concentration of 0.71 ng/g (Makey et al., 2017). Studies conducted in Europe evaluated
indoor dust concentrations from homes, schools, workplaces, and/or cars. One of these studies,
Huber et al. (2011) from Norway, also sampled a storage room which contained chemicals and
highly contaminated samples and found an elevated concentration at 43.4 ng/g when compared
to other locations within the study. Other samples within the study include living rooms, carpets,
sleeping rooms, sofas, and offices and concentrations ranged between 0.2-26.7 ng/g. Four
studies sampled floor dust exclusively in homes with results ranging from not detected to 37 ng/g
(de la Torre et al., 2019; Winkens et al., 2018; Padilla-Sanchez and Haug, 2016; Jogsten et al.,
2012). Another study in Norway evaluated dust from elevated surfaces such as bookshelves and
windowsills (n = 41) and reported a range of 3.9-92 ng/g, notably higher than other studies
(Haug et al., 2011). One study from Sweden assessed 20 dust samples from elevated surfaces in
preschools and reported a median and 95th percentile of 1.09 ng/g and 56.0 ng/g, respectively
(Giovanoulis et al., 2019). Harrad et al. (2019) evaluated dust concentrations from living rooms,
cars, and classrooms in Ireland ranging from <0.05 ng/g to 14 ng/g and, notably, from offices
which had concentrations ranging from <0.05 ng/g to 120 ng/g. A study in Belgium evaluated
randomly selected homes (n = 43) and offices (n = 10) and found dust samples containing
median (95th percentile) values of 0.1 (2.1) ng/g dw and 0.4 (62) ng/g dw, respectively
(D'Hollander et al., 2010).
C-87
-------
FINAL
2024
Table C-8. Summary of PFNA Indoor Dust Data
Study
Location
Site Details
Results
United States
Scheretal. (2019)
United States (Twin Cities
metropolitan region, Minnesota)
Nineteen homes in three cities within a GCA near
former 3M PFAS production facility as well as from
three homes in the Twin Cities Metro outside the
GCA. Dust samples collected from an entryway to the
yard and from an interior living space (e.g., family
room, living room) in each home in July-September
2010. Homes within the GCA had previous or ongoing
PFAS contamination in drinking water and were
served by the Oakdale, Minnesota public water system
or a private well previously tested and shown to have
detectable levels ofPFOA orPFOS. Results were not
reported for homes outside the GCA.
Entryway: n = 19, DF 68%, median (range) =
9.7 (
-------
FINAL
2024
Study
Location
Site Details
Results
(shelving, tops of bookcases/storage cubbies) were
collected along with floor dust in the same sample.
Strynar and Lindstrom (2008)
United States (North Carolina;
Ohio)
Dust samples from vacuum cleaner bags were
obtained in 2000-2001 during the EPA's Children's
Total Exposure to Persistent Pesticides and Other
Persistent Organic Pollutants (CTEPP) study from
North Carolina (49 homes, 5 daycare centers) and
Ohio (53 homes, 5 daycare centers). Vacuum cleaner
bags were only collected if available at each site.
n = 112; DF 42.9%, mean, median (maximum)
= 22.1, 7.99 (263) ng/g
(LOQ =11.3 ng/g)
*Values below the LOQ assigned a value of
LOQ/V2
Fraseretal. (2013)
United States (Boston,
Massachusetts)
Dust samples were collected in January-March 2009
from offices (n = 31), homes (n = 30), and vehicles (n
= 13) of 31 individuals. Study participants worked in
separate offices located across seven buildings, which
were categorized into Building A, Building B, and
Other. Six samples were collected from Building A, a
newly constructed (approximately one year prior to
study initiation) building with new carpeting and new
upholstered furniture in each office. Seventeen
samples were collected from Building B, a partially
renovated (approximately one year prior to study
initiation) building with new carpeting throughout
hallways and in about 10% of offices. Eight samples
were collected from the other five remaining buildings
where no known recent renovation occurred. Study
offices were not vacuumed during the sampling week
and homes and vehicles were not vacuumed for at
least one week prior to sampling. Entire accessible
floor surface areas and tops of immovable furniture
were vacuumed in offices and the main living area of
homes. Entire surface areas of the front and back seats
of vehicles were vacuumed.
Number of home dust samples was reduced to 30
because 1 participant lived in a boarding house with
no main living area. Sufficient mass of dust for
analysis was available from only 13 vehicles.
Homes: n = 30, DF 67%, GM (range) = 10.9
(6.21-1,420 ng/g)
Offices: n = 31, DF 94%, GM (range) = 63.0
(10.9-639) ng/g
Vehicles: n = 13, DF 85%, GM (range) = 14.7
(4.95-101 ng/g)
(LOQ =5 ng/g)
*GM calculated by replacing values
-------
FINAL
2024
Study
Location
Site Details
Results
cleaner bags or subsampling bagless or central
vacuums.
Europe
de la Torre et al. (2019)
Spain (unspecified), Belgium
(unspecified), Italy (unspecified)
Sixty-five homes belonging to the partners of Test-
Achats (Belgium), Altroconsumo (Italy), and OCU
Ediciones SA (Spain). Home occupants vacuumed the
entire floor of their home from September 2016 to
January 2017 and vacuum bags were collected
Total: n = 65, DF 46%, median (range) = 0.04
(ND-9.04) ng/g
Spain: n = 21, DF 48%, median (range) = 0.04
(ND-5.70) ng/g
Belgium: n = 22, DF 36%, median (range) =
0.04 (ND-9.04) ng/g
Italy: n = 22, DF 55%, median (range) = 0.10
(ND-6.54) ng/g
(LOQ = 0.06 ng/g)
*Values below LOQ replaced with
LOQ/(square root of 2)
Winkens et al. (2018)
Finland (Kuopio)
Sixty-three private households from the birth cohort
study, LUKAS2. Floor dust samples collected in
2014/2015 from the children's bedroom (entire floor).
Participants were instructed not to vacuum clean the
room at least a week before sampling. For 55 rooms,
dust samples were collected at the end of a 3-week air
sampling period (indoor air results reported in a
different study).
n = 63, DF 52.4%), mean, median (range) =
1.76,1.05 (BDL-14.8)ng/g
(MDL = 0.73 ng/g)
*Values
-------
FINAL
2024
Study
Location
Site Details
Results
Giovanoulis et al. (2019)
Sweden (Stockholm)
Twenty preschools that had been previously sampled
in 2015 and then participated in the "chemical smart
preschool" initiative to reduce the presence of
hazardous chemicals in the indoor environment; 2015
results are reported elsewhere. Samples for this study
were collected during January to February 2018. One
settled dust sample was collected from elevated
surfaces (50-250 cm above the floor) from different
areas of a play room at each preschool.
n = 20, DF 55%, medi™ (95th percentile) =
1.09 (56.0) ng/g
(LOD = 0.5 ng/g)
*Values
-------
FINAL
2024
Study
Location
Site Details
Results
¦"Concentrations MQL
*Median calculated by replacing values
-------
FINAL
2024
C.3.S. Air
Perfluoroalkyl chemicals have been released to air from wastewater treatment plants, waste
incinerators, and landfills (Ahrens et al., 2011), though there is limited information on the
detection levels or frequencies of PFNA in either indoor or ambient air. ATSDR (2021) notes
perfluoroalkyl chemicals have been detected in air and they can be transported long distances via
the atmosphere. For example, in a study performed from April 2007 to January 2009, PFNA was
detected at an average concentration of 0.3 pg/m3 in 8% of 141 atmospheric samples from
Atlantic and Southern Oceans and coastal areas of the Baltic Sea (NCBI, 2022b; Dreyer et al.,
2009). PFNA is not expected to be broken down directly by photolysis (NCBI, 2022b). PFNA
can undergo hydroxylation in the atmosphere, with a (predicted average) atmospheric
hydroxylation rate of 8.41 x 10 13 cmVmolecule - second to a (derived) rate of
5.2 x 10 " cm3/molecule - second (with corresponding estimated half-life of 31 days for this
reaction in air) (USEPA, 2022b; NCBI, 2022b). With a vapor pressure of 4.83 x 10~3 mm Hg at
20°C (extrapolated), 8.3 x 10~2 mm Hg at 25°C (estimated), 8.4 mm Hg at 99.63°C (measured),
and a (measured) range of 4.80 x 10 3 mm Hg to 9.77 x 10 3 mm Hg, volatilization is not
expected to be an important fate process for this chemical (USEPA, 2022b; NCBI, 2022b;
ATSDR, 2021). The EPA's Toxics Release Inventory reported release data for PFNA in 2020,
with a total onsite disposal, offsite disposal, and other releases concentration of 0 pounds from
one facility (USEPA, 2022c). PFNA is not listed as a hazardous air pollutant (USEPA, 2022d).
C.3.5.1. Indoor Air
No studies from the U.S. reporting levels of PFNA in indoor air were identified from the primary
or gray literature. However, the EPA identified studies from Canada and Europe that are
summarized below and in Table C-9 (Harrad et al., 2019; Makey et al., 2017; Jogsten et al.,
2012; Barber et al., 2007). All of these studies sampled from homes, while only one study also
sampled from offices, cars, and classrooms and one study also sampled from a laboratory.
Two studies reported results from indoor air samples collected only from homes (Makey et al.,
2017; Jogsten et al., 2012). In the Canadian study, Makey et al. (2017) collected indoor air,
vacuum cleaner dust, and serum samples in 2007-2008 from the homes of women in the second
trimester of pregnancy and analyzed the samples for levels of PFAAs. Participants were part of
the Chemicals, Health, and Pregnancy (CHirP) Study. PFNA was detected in 42% of indoor air
samples (n = 39), with a geometric mean of 1.5 pg/m3. In Spain, Jogsten et al. (2012) sampled
indoor air (n = 10) from selected homes in Catalonia in December 2009 and evaluated levels of
27 PFCs. PFNA was not detected (LOD = 3.1-280 pg/m3 for all ionic PFAS).
The remaining two studies evaluated PFNA levels in offices, vehicles, and/or schools, in addition
to homes. In Ireland, Harrad et al. (2019) collected air samples in homes (living rooms, n = 34),
offices (n = 34), cars (n = 31), and school classrooms (n = 28) between August 2016 and January
2017. PFNA was detected in all four indoor microenvironments in 18%, 91%, 90%, and 93% of
samples for homes, offices, cars, and classrooms, respectively. The mean (median)
concentrations were 2.1 (1.7) pg/m3 in homes, 3.7 (2.5) pg/m3 in offices, 5.2 (2.1) pg/m3 in cars,
and 3.5 (2.5) pg/m3 in classrooms. In Norway, neutral and ionic PFAS were analyzed in indoor
air samples collected from three homes and one laboratory in Troms0 between April and June
2005 (Barber et al., 2007). The study detected PFNA in all four samples, with a mean
concentration of 2.7 pg/m3.
C-93
-------
FINAL
2024
Table C-9. Summary of Studies Reporting the Occurrence of PFNA in Indoor Air
Study
Location
Site Details
Results
Canada
Makey et al. (2017)
Canada (Vancouver)
Samples were collected from a subset of Chemicals,
Health, and Pregnancy (CHirP) Study participants'
homes in 2007-2008; dust and serum samples were
also collected. Indoor air was sampled using passive
samplers deployed in participants' bedrooms for four
weeks.
n = 39, DF 42%, GM = 1.5 pg/m3
(DL = 0.02 pg/m3)
Europe
Jogsten et al. (2012)
Spain (Catalonia)
Indoor air sampling was performed in December 2009
from ten households at approximately 1 m above the
floor. Samples were collected out of convenience and
may not be representative of the entire Catalan
population. Both particulate and gas phases collected.
n = 10, DF 0%
(LOD = 3.1-280 pg/m3 for all ionic PFAS)
Harrad et al. (2019)
Ireland (Dublin, Gal way,
Limerick)
Air samples collected from homes (living rooms),
offices, cars, and school classrooms; dust samples also
collected. Samples collected between August 2016
and January 2017. Sample numbers were split
approximately equally from each of the three counties.
Gas or particulate phase not specified.
Homes: n = 34, DF 18%, mean, median
(range) = 2.1, 1.7 (<0.3-13) pg/m3
Offices: n = 34, DF 91%, mean, median
(range) = 3.7, 2.5 (<0.3-18) pg/m3
Cars: n = 31, DF 90%, mean, median (range)
= 5.2,2.1 (<0.3-24) pg/m3
Classrooms: n = 28, DF 93%), mean, median
(range) = 3.5, 2.5 (<0.3—15) pg/m3
(LOD = 0.3 pg/m3)
*When analyte peaks are
-------
FINAL
2024
C.3.5.2. Ambient Air
A single U.S. study measured levels of PFNA in ambient air (Kim and Kannan, 2007). Kim and
Kannan (2007) analyzed particle phase (n = 8) and gas phase (n = 8) concentrations of
perfluorinated acids in ambient air samples collected in and around Albany, New York in May
and July 2006 to examine the relative importance of certain media pathways to the contamination
of urban lakes. PFNA was detected in all gas phase samples with mean and median
concentrations of 0.21 pg/m3 and 0.20 pg/m3, respectively. PFNA was also detected in the
particulate phase, but the detection frequency was not reported. Authors reported particulate
phase mean and median concentrations of 0.13 pg/m3 and below the LOQ (0.12 pg/m3),
respectively.
One Canadian PFNA study by Ahrens et al. (2011) analyzed the temporal ambient PFNA
concentrations at two municipal solid waste (MSW) landfills and at one wastewater treatment
plant. At the MSW landfills, they recorded a mean (range) concentration of 2.11 (0.97-3.24)
pg/m3 upwind of the site and 10.3 (4.82-15.8) pg/m3 at the site. At the wastewater treatment
plant, they measured the PFNA concentration within the facility in the primary clarifier, aeration
tank, and secondary clarifier, finding mean concentrations between 2.97 and 3.62 pg/m3. They
also evaluated reference sites near to (<200m) and distant from (~600m) the facility, finding
higher ambient PFNA concentrations at the near sites (mean of 1.64 pg/m3) compared to the
distant sites (single point estimate of 0.88 pg/m3).
Among the European studies—conducted in Spain (Jogsten et al., 2012; Beser et al., 2011),
Ireland (Harrad et al., 2019; Barber et al., 2007), Norway (Barber et al., 2007), and the United
Kingdom (Barber et al., 2007)— each collected ambient (outdoor) air for PFNA measurements.
Reported mean PFNA concentrations among these studies range from 0.08-3.8 pg/m3. Beser et
al. (2011) measured PFNA (PM2.5-bound) in both residential and industrial areas (five sampling
stations in total) of the Alicante Province in Spain in 2010, with mean concentrations ranging
from 1.65-3.8 pg/m3. Jogsten et al. (2012) did not detect PFNA in ten sites across Catalonia,
Spain the year prior in 2009, however, they identified their limit of detection as 3.1 pg/m3, which
was higher than the mean PFNA concentration detected in four of the five sites in Spain
evaluated by Beser et al. (2011). Harrad et al. (2020) compared PFNA concentrations both
upwind and downwind of ten Irish MSW landfills in late 2018 and early 2019, finding
comparable concentration ranges upwind (<0.08-0.31 pg/m3) and downwind (0.08-0.52 pg/m3).
Barber et al. (2007) also detected ambient PFNA at detectable but not quantifiable levels (<3.3
pg/m3 on average) in rural Mace Head, Ireland. United Kingdom samples (n = 15) from
Hazelrigg and Manchester were reported in ranges of <0.06-0.9 pg/m3 and two samples from
Kjeller, Norway were reported at a range of 0.10-0.13 pg/m3 (Barber et al., 2007).
C-95
-------
FINAL
2024
Table C-10. Summary of Peer-Reviewed Studies Reporting the Occurrence of PFNA in Ambient Air
Study
Location
Site Details
Results
United States
Kim and Kannan (2007)
United States (Albany, New
York)
Roof of a lakehouse building located at Washington
Park Lake in May and July 2006. Both particulate and
gas phases collected.
Gas: n = 8, DFa 100%, mean, median (range)
= 0.21,0.20 (0.16-0.31)pg/m3
Particle: n = 8, DF NR, mean, median (range)
= 0.13,
-------
FINAL
2024
Study
Location
Site Details
Results
biomedical waste. Gas or particulate phase not
specified.
Jogsten et al. (2012)
Spain (Catalonia)
Outdoor air sampling conducted in December 2009 for
the purposes of comparison to indoor air and dust
samples. Number of sites not specified but assumed to
be ten because indoor air was sampled from ten
homes. Samples were collected out of convenience
and may not be representative of the entire Catalan
population. Both particulate and gas phases collected.
n = 10; DF 0%
(LOD = 3.1-280 pg/m3 for all ionic PFAS)
Beser et al. (2011)
Spain (Alicante province)
Samples collected from April to July 2010 from five
stations. Two stations were placed in Elche (one in a
residential area and the other in an industrial area).
The third station was placed in a residential area of
Alicante City. The fourth station was in a rural area of
Pinoso and the last station was in a residential area of
Alcoy. Concentrations reported for PM2.5-bound
PFNA.
Elche (residential): n = 11, DFa 55%, mean =
2.7 pg/m3
Elche (industrial): n = 13, DFa 69%, mean =
3.8 pg/m3
Alicante City: n = 11, DFa 36%, mean =1.7
pg/m3
Pinoso: n = 3, DFa 67%, mean = 1.65 pg/m3
Alcoy: n = 3, DFa 100%), mean = 2.2 pg/m3
(MQL = 1.4 pg/m3)
*Mean calculated from values >MQL
Barber et al. (2007)
United Kingdom (Hazelrigg,
Manchester); Ireland (Mace
Head); Norway (Kjeller)
Samples collected from four field sites in Europe:
Hazelrigg (semirural) and Manchester (urban) were
sampled in two sampling events in February-March
2005 and November 2005-January 2006; Mace Head
(rural) was sampled in March 2006; and Kjeller (rural)
was sampled in November-December 2005. PFNA was
measured in the particulate phase.
Hazelrigg first sampling event:
n = 2, DF NR, mean = <13.8 pg/m3
(MQL =13.9 pg/m3)
*The glass-fibre filters were analyzed in a
batch of samples that showed contamination
problems, so the high associated blank value
used to calculate the MQL put most analytes
-------
FINAL
2024
Study
Location
Site Details
Results
(MQL = 3.32 pg/m3)
Kjeller:
n = 2, DFa 100%, mean (range) = 0.12
(0.10-0.13) pg/m3
(MQL = 0.10 pg/m3)
*Means calculated from values >MQL
C-98
-------
FINAL
2024
C.3.6. Soil
The use and production of PFNA could result in its release to soils through various waste streams
(NCBI, 2022b). When released to soil, based on its physico-chemical properties, PFNA is
expected to have no mobility (NCBI, 2022b). PFNA has been measured in grass samples grown
in soil containing PFNA and other PFAS near Decatur, Alabama (ATSDR, 2021; Yoo et al.,
2011). In addition, PFNA has been found to accumulate in the roots of maize plants grown in
soil containing PFNA and other PFAS (ATSDR, 2021; Krippner et al., 2014).
Seven U.S. studies were identified that evaluated the occurrence of PFNA and other PFAS in soil
(Galloway et al., 2020; Nickerson et al., 2020; Zhu and Kannan, 2019; Eberle et al., 2017;
Anderson et al., 2016; Venkatesan and Halden, 2014; Blaine et al., 2013) (Table C-l 1). Among
these studies, three analyzed soils potentially impacted by past AFFF use. The PFNA detection
frequencies varied widely (from less than 20% to over 90%) but mean concentrations tended to
be below 5 ng/g. Few studies analyzed soils in the vicinity of fluoropolymer manufacturing
facilities or by contaminated soil amendments. Other than control soils in two greenhouse and
field studies and one reference site, the U.S. studies did not evaluate soils without amendments
or without a nearby current or historical PFAS source.
Two studies analyzed soils in the vicinity of fluoropolymer manufacturing facilities (Galloway et
al., 2020; Zhu and Kannan, 2019). Galloway et al. (2020) collected soil samples in December
2016 and March 2018 near a fluoropolymer production facility outside Parkersburg, West
Virginia. The 2016 sampling included sites 4.0 km-48.1 km downwind to the north and
northeast of the facility and the 2018 sampling included sites 1.3 km-45.4 km north of the
facility. PFNA was detected in six of eight of the 2016 samples, however only one was above the
LOQ with a concentration of 1.63 ng/g. PFNA was also detected in six of seven of the 2018
samples, however only one was above the LOQ with a concentration of 1.92 ng/g at a distance of
1.3 km. Both the 2016 and 2018 samples that were above the LOQ were reported at the site
closest to the facility. In Zhu and Kannan (2019), authors studied PFAS concentrations in soil
contaminated by a nearby fluoropolymer manufacturing facility in Little Hocking, Ohio, which
had been manufacturing fluorochemicals for over five decades. The 45-acre well field located in
a floodplain meadowland was divided into quadrants and surface soil samples were collected
from multiple locations within each quadrant in October 2009. PFNA was detected in all 19
samples with mean and median concentrations of 2.7 ng/g and 2.5 ng/g, respectively.
Three studies analyzed soils potentially impacted by AFFF use (Nickerson et al., 2020; Eberle et
al., 2017; Anderson et al., 2016). Anderson et al. (2016) assessed 40 sites across 10 active Air
Force installations throughout the continental United States and Alaska between March and
September 2014. Installations were included if there was known historic AFFF release in the
period 1970-1990. It is assumed that the measured PFAS profiles at these sites reflect the net
effect of several decades of all applicable environmental processes. The selected sites were not
related to former fire training areas and were characterized according to volume of AFFF release
- low, medium, and high. Across all sites, the PFNA detection frequency was 71.43% in 100
surface soil samples (median concentration of detects was 1.3 ng/g) and 14.42% in 112
subsurface soil samples (median concentration of detects was 1.5 ng/g). PFNA was detected
more frequently at high-volume release sites (50.8% in 32 surface soil samples with mean
concentration of 2.5 ng/g; 84.4% in 31 subsurface soil samples with mean concentration of
C-99
-------
FINAL
2024
2.4 ng/g) than at low-volume sites (50.0% in 12 surface soil samples with mean concentration of
2.7 ng/g; 17.6% in 17 subsurface soil samples with mean concentration of 1.0 ng/g) and medium-
volume sites (38.3%) in 56 surface soil samples with mean concentration of 2.2 ng/g; 67.9% in 64
subsurface soil samples with mean concentration of 2.1 ng/g). Authors noted that given PFNA is
not present in 3M AFFF formulations, there may be some degree of telomer-based AFFF
contamination. Nickerson et al. (2020) developed a method to quantify anionic, cationic, and
zwitterionic PFAS from AFFF-impacted soils. The method was applied to two soil cores
collected from two different AFFF-impacted former fire training areas; the sampling year and
geographic location were not provided. Eleven soil samples, corresponding to 11 depths ranging
from 0.46 m to 15.1 m, were evaluated from Core E, and 12 soil samples, at depths ranging from
0.30 to 14.2 m, were evaluated from Core F. In Core E, PFNA was detected in 5 of 11 samples at
depths both at the surface and further below ground with PFNA concentrations ranging from
below the LOQ to 1.96 ng/g dw. In Core F, PFNA was detected in 5 of 12 samples at the five
depths closest to the surface, with concentrations ranging from below the LOQ to 4.17 ng/g dw
(LOQ not reported). Eberle et al. (2017) investigated the effects of an in situ chemical oxidation
treatment for remediation of chlorinated volatile organic compounds and PFAAs co-
contaminants. Soil samples were collected in 2012-2013 before and after a pilot scale field test
at a former fire training site at Joint Base Langley-Eustis, Virginia. Monthly fire training
activities were conducted at the site from 1968 to 1980 and irregular fire training activities
continued until 1990. Impacted soil was excavated in 1982 but details were not provided. PFNA
was detected in 1 of 5 pre-treatment samples and in 13 of 14 post-treatment samples. Of the
available three paired pre- and post-treatment soil samples, PFNA was not detected pre-treatment
in two pairings but detected post-treatment at 0.07 ng/g and 0.05 ng/g post-treatment. For the
third pairing, PFNA was detected at 1.1 ng/g pre-treatment and below the LOQ (0.06 ng/g) post-
treatment.
Of the remaining two studies conducted in the United States, Venkatesan and Halden (2014)
conducted outdoor mesocosm studies to examine the fate of PFAS in biosolids-amended soil
collected during 2005-2008. Biosolids were obtained from a wastewater treatment plant
(WWTP) in Baltimore that primarily treated wastewater from domestic sources with only minor
contribution (1.9%) from industry. The number of samples was not provided but PFNA was
detected in the control (nonamended) soil at levels below 0.5 ng/g dw and in the biosolids-
amended soil at a level not reported by the authors. In a field and greenhouse study, Blaine et al.
(2013) studied the uptake of PFAS into edible crops grown in control and biosolids-amended
soil. In the field study, urban biosolids were obtained from a WWTP receiving both domestic
and industrial waste while rural solids were obtained from a WWTP receiving domestic waste
only. PFNA was detected in soils from urban (mean = 0.20 ng/g, 0.28 n/g, and 0.40 ng/g in
control, lx9 and 2x amended fields, respectively) and rural fields (mean = 0.06 ng/g and
0.75 ng/g in control and 0.5 x amended fields, respectively). In the greenhouse study, three soils
(nonamended control, industrially impacted, and municipal) were investigated. Industrially
impacted soils contained composted biosolids from a small municipal WWTP that was impacted
by PFAA manufacturing while municipal soils were obtained from a reclamation site in Illinois
where municipal biosolids were applied for 20 years. PFNA was detected in all three soils at an
average concentration of 0.30 ng/g, 20.15 ng/g, and 6.11 ng/g in control, industrially impacted,
9 0.5 x, 1 /. or 2/ is defined as 'A I, or 2 times the agronomic rate of biosolids application to meet nitrogen
requirements of the crop.
C-100
-------
FINAL
2024
and municipal soil, respectively. Authors noted that the trace levels of PFAS detected in the
control soil may be due to minor cross-contamination from plowing, planting, or atmospheric
deposition from the surrounding area where biosolids have been applied.
Results for all of the identified non-U.S. studies are presented in detail in Table C-l 1 and are
summarized here. The EPA identified three Canadian studies, two were conducted at locations
with prior AFFF use (Cabrerizo et al., 2018; Mejia-Avendano et al., 2017). Cabrerizo et al.
(2018) evaluated soil in two locations, one of which was relatively remote and largely did not
have direct human contact and the other of which was previously used as a military training
facility. The remote location (n = 19) had concentrations ranging from 0.0262 ng/g dw to 0.8749
ng/g dw and the historic military training location (n = 8) similarly had concentrations from
0.0836 ng/g dw to 0.7794 ng/g dw. Mejia-Avendano et al. (2017) investigated soil samples at the
site of the 2013 Lac-Megantic train accident, where approximately 33,000 L of AFFF
concentrates were used to put out fires. In 2013, 12 sample concentrations ranged from 0.138-19
ng/g dw and in 2015, two years after the incident, 11 sample concentrations ranged from 0.031-
0.777 ng/g dw. In the third Canadian study, Dreyer et al. (2012) sampled bog peat cores in an
undisturbed and well investigated location to determine historic atmospheric contamination.
Estimated core segment dated back to 1912, with PFNA concentrations ranging between not
detected and 0.412 ng/g.
Of the European studies, three were conducted at locations near firefighting facilities (Dauchy et
al., 2019; Skaar et al., 2019; Hale et al., 2017). Dauchy et al. (2019) and Skaar et al. (2019)
found varying results from not detected to 15 ng/g dw while Hale et al. (2017) reported a range
of 2.8-41.3 ng/g for the 40% of samples that were detected. One study in Belgium (Groffen et
al., 2019) evaluated soil samples at a perfluorochemical plant and at four sites increasing in
distance from the plant. PFNA levels were detected at the plant ranging from not detected to 2.53
ng/g dw and ranging from not detected to 0.53 ng/g dw away from the plant in no discernable
pattern. Harrad et al. (2020) investigated soil samples upwind and downwind from ten municipal
solid waste landfills in Ireland. PFNA was found in all samples upwind of the landfills, ranging
in concentrations from 0.0029-0.033 ng/g dw, and in 89% of samples downwind, ranging in
concentrations from not detected to 0.0077 ng/g dw. A study in Norway (Grannestad et al.,
2019) found mean PFNA concentrations from a popular skiing location in Granasen to be lower
than that of the study's selected reference site in Jonsvatnet. Sammut et al. (2019) sampled soil
from six random small urban fields in Malta and found concentrations ranging from 0.66-0.87
ng/g.
C-101
-------
FINAL
2024
Table C-ll. Summary of PFNA Data in Soil
Study
Location
Site Details
Results
United States
Galloway et al. (2020)
United States (Parkersburg,
West Virginia)
Soil samples collected near a fluoropolymer facility in
two sampling trips in December 2016 and March
2018. The 2016 sampling trip included a collection
radius 4.0^18.1 km downwind to the north and
northeast of the facility. The 2018 sampling trip
focused on samples collected to the north of the
facility with a radius of 1.3^15.4 km.
2016 sampling:
Drag Strip Road (4.0 km) = 1.63 ng/g
Veto Lake (8.0 km) =
-------
FINAL
2024
Study
Location
Site Details
Results
process; active remedies had not been applied at any
of the sites selected. Approximately ten samples were
collected between March and September 2014 at each
site for surface and subsurface soil; sites were grouped
according to volume of AFFF release—low-volume
typically had a single AFFF release, medium-volume
had one to five releases, and high-volume had multiple
releases.
Testing and Maintenance (high-volume
release):
n = 32, DF 50.8%, mean (range) = 2.5
(0.24-23) ng/g
(RL = 0.23 ng/g)
Subsurface soil:
Overall: n = 112, DF 14.42%, median
(maximum) =1.5 (6.49) ng/g
Breakdown by site:
Emergency Response (low-volume release):
n = 17, DF 17.6%), mean (range) = 1.0
(0.5-1.5) ng/g
Flangars and Buildings (medium-volume
release):
n = 64, DF 67.9%o, mean (range) = 2.1
(0.21-12) ng/g
Testing and Maintenance (high-volume
release):
n = 31, DF 84.4%o, mean (range) = 2.4
(0.24-23) ng/g
(RL = 0.24 ng/g)
*Median calculated using quantified
detections
*Non-detects were substituted with Vi the
reporting limit
Nickerson et al. (2020)
United States (unspecified)
Soil cores E and F from two different AFFF-impacted
fire training areas; sampling year and geographic
location not provided. Soil core E contained 11- 0.3 m
increment samples from 0.3-15.2 m below ground
surface and was collected in an area where the
surficial soils were likely disturbed due to regrading
and other soil redistribution activities. Soil core F
contained 12- 0.61 m increment samples from 0-14.2
m below ground surface and was collected in an area
where the surficial soils were highly permeable only
within the upper 0.5 to 1 m, and the underlying
impermeable clay layer exhibited a relatively high
cation exchange capacity and organic carbon content.
Core E:
0.46 m = 1.96 ng/g dw
2.9 m =
-------
FINAL
2024
Study
Location
Site Details
Results
The water table was relatively shallow (depth <3 m) at
both sites.
1.22 m = 4.17 ng/g dw
1.83 m = 3.23 ng/g dw
2.44 m = 1.04 ng/g dw
3.05 m = 0.64 ng/g dw
4.11 m =
-------
FINAL
2024
Study
Location
Site Details
Results
Center; number of sampling sites and number of
samples not provided.
Biosolids-amended soil obtained by mixing biosolids
and soil at a volumetric ratio of 1:2. Biosolids were
from Back River WWTP in Baltimore, a full-scale
activated sludge treatment plant. Raw wastewater was
primarily from domestic sources with only minor
contribution (1.9%) from industry.
PFNA, PFDA, and PFUnA in the control
soil accounted for 0.3-3% of their initial
levels in the amended soil mix
(MDL = 0.08 ng/g)
Blaine et al. (2013)
United States (Midwest)
Urban and rural full-scale field study with control
(nonamended) and biosolids-amended plots. Three
agricultural fields were amended (0.5x, 1 x, or 2x)
with municipal biosolids. Urban biosolids (1 x and 2x)
were from a WWTP receiving both domestic and
industrial waste. Rural biosolids (0.5x) were from a
WWTP receiving domestic waste only. Control plots
were proximal to the rural and urban amended corn
grain and corn stover field sites; sampling year not
provided.
Greenhouse study with control (nonamended) and
biosolids-amended soil. Nonamended soil obtained
from a field that received commercial fertilizers and
had a similar cropping system as the nearby municipal
soil site. Municipal soil was obtained from a
reclamation site in Illinois where municipal biosolids
were applied at reclamation rates for 20 years,
reaching the cumulative biosolids application rate of
1,654 Mg/ha. Industrially impacted soil was created
by mixing composted biosolids from a small
municipal (but impacted by PFAA manufacturing)
WWTP with control soil on a 10% mass basis.
Sampling year not provided.
Field study:
Urban non-amended: n = 3-7, DF NR, mean
= 0.20 ng/g
Urban 1 x: n = 3-7, DF NR, mean = 0.28
ng/g
Urban 2x: n = 3-7, DF NR, mean = 0.40
ng/g
Rural non-amended: n = 3-7, DF NR, mean
= 0.06 ng/g
Rural 0.5x: n = 3-7, DF NR, mean = 0.75
ng/g
(LOQ not reported)
Greenhouse study:
Nonamended: n = 3-5, DF NR, mean = 0.30
ng/g
Industrially impacted: n = 3-5, DF NR,
mean = 20.15 ng/g
Municipal: n = 3-5, DF NR, mean = 6.11
ng/g
(LOQ not reported)
Canada
Cabrerizo et al. (2018)
Canada (Melville and
Cornwallis Islands)
Catchment areas of lakes in the Cape Bounty Arctic
Watershed Observatory on southern Melville Island
(West, East, and Headwater lakes) during summer
(late July-early August) 2015 and 2016, representing
an environment largely unimpacted by direct human
activity; data for 19 sampling sites available (S6, SI 1—
S28).
Melville Island lakes:
n = 19, DFa 100%), meana (range) = 0.3248
(0.0262- 0.8749) ng/g dw
Cornwallis Island lakes:
n = 8, DFa 100%o, meana (range) = 0.3333
(0.0836-0.7794) ng/g dw
(LOD = 0.0001-0.018 ng/g for all PFAS)
C-105
-------
FINAL
2024
Study
Location
Site Details
Results
Catchment areas of lakes on Cornwallis Island
(Resolute, North, Small, Meretta, 9 Mile, and Amituk
lakes) near the community of Resolute Bay during
summer (late July-early August) 2015 and 2016.
Resolute Bay has a military and civilian airport which
discharged its wastewaters into the upper area of the
catchment until 1997, three old solid waste landfills
1.5-2 km west of the airport used until the mid-1990s,
and Arctic research and military training facilities
close to the airport that support activities such as
vehicle use, firefighting, and construction/demolition;
eight sampling sites (S29-S36).
Mejia-Avendano et al. (2017)
Canada (Lac-Megantic, Quebec)
Site of July 2013 Lac-Megantic train accident where
63 out of 72 train cars carrying 8 million liters of
crude oil derailed and a major oil fire ignited. Seven
types of AFFFs and approximately 33,000 L of AFFF
concentrates were used. Samples were collected in
July 2013 weeks after the accident from the western
shores of Chaudiere River, at the point where the oil
and AFFF runoff reached the river, approximately 500
m from the edge of the derailment site; in July 2015
from the fire burn site and adjacent area in downtown
Lac-Megantic where the soil was continuously
excavated for remediation (the site was the closest to
the accident site among the areas open to sampling);
and from a background, nonimpacted area next to
Chaudiere River, about 5 km from the accident site, on
the east shore of the river and on the opposite side of
the accident.
Background:
n = 3, DF NR, mean = 0.212 ng/g dw
2013:
n = 12 (from 12 sites), DFa 100%, meana
(range) = 4.41 (0.138-19) ng/g dw
2015:
n = 11 (from 9 sites), DFa 100%, meana
(range) = 0.274 (0.031-0.777) ng/g dw
(LOD = 0.02 ng/mL; LOQ = 0.05 ng/mL)
Dreyer et al. (2012)
Canada (Ottawa, Ontario)
Two ombrotrophic Mer Bleue bog peat core samples
collected in October 2009 and cut into 5-cm segments
(nine segments for the first core, eight segments for
the second core); Mer Bleue selected because it is
undisturbed and well investigated and is located in a
meltwater channel of the postglacial Ottawa River.
Peat cores sampled to determine their suitability for
determining historic atmospheric contamination;
contaminants present due to atmospheric deposition
only. The year for each segment was estimated
through dating of Mer Bleue peat cores collected in
the same year for a different study.
First core (first parallel; second parallel):
2009: 0.041; 0.033 ng/g
2006: 0.143; 0.192 ng/g
2001: 0.229; 0.223 ng/g
1992: 0.259; 0.271 ng/g
1983: 0.241; 0.234 ng/g
1973: 0.203; 0.322 ng/g
1962: 0.263; 0.297 ng/g
1945: 0.148; 0.156 ng/g
1927: 0.069; 0.089 ng/g
1912: 0.044; 0.044 ng/g
C-106
-------
FINAL
2024
Study
Location
Site Details
Results
Second core (first parallel; second parallel):
2009: 0.052; 0.062 ng/g
2006: 0.193; 0.166 ng/g
2001: 0.162; 0.234 ng/g
1992: 0.320; 0.319 ng/g
1983: 0.396; 0.412 ng/g
1973: 0.206; 0.206 ng/g
1962: 0.082; 0.081 ng/g
1945: 0.160; 0.149 ng/g
1927:
-------
FINAL
2024
Study
Location
Site Details
Results
every 25 cm in the topmost meter and then every 50
cm). Second sampling campaign collected 14 soil
cores between 4 m and 15 m from areas 1-6 (thickness
of composite soil samples ranged from 25 cm to 100
cm) in October 2016.
Area 1 stored raw oil products when the oil refinery
was operating; a preliminary survey showed
hydrocarbon traces in the area, suggesting that an
incident had occurred and that fluorinated surfactants
could have been used. Area 2 is one of the main areas
used for firefighting activities since 1987; training
sessions held directly on the ground before 10-cm
thick concrete slab was built in the 1990s. Area 3 was
used for firefighting activities since 1987 and is
situated on a 1-meter thick concrete slab on the
foundations of the former oil refinery. Area 4
corresponds to the site's WWTP where sludge and
sediment from a lagoon were stored directly on the
ground; influents of the WWTP are highly
contaminated by PFAS. Area 5 was used for
firefighting training exercises by the former oil
refinery. Area 6 is used for firefighting exercises with
tank trucks.
SC-60 = <20, <20, <4, <4, <2, <2, <2 ng/g
dw (0-0.25, 0.25-0.5, 0.5-0.75, 0.75-1,
1-1.5,2-2.5,2.5-3 m)
SC-61 = <20, <20, <4, <4, <20, <4, <4 ng/g
dw (0-0.25, 0.25-0.5, 0.5-0.75, 0.75-1,
1-1.5,1.5-2,2-2.5 m)
SC-62 = <20, <20, <4, <20, <4 ng/g dw (0-
0.25, 0.5-0.75,1-1.5, 2-2.5, 3.5^1 m)
SC-63 = <20, <20, <4, <2, <2, <2, <2, <2,
<2 ng/g dw (0-0.25, 0.25-0.5, 0.5-0.75,
0.75-1, 1-1.5, 1.5-2,2-2.5,2.5-3,3-3.5
m)
SC-64 = <2, <2, <4, <4, <2, <4 ng/g dw (0-
0.25, 0.25-0.5, 0.5-0.75, 0.75-1,1-1.5,
1.5-2,2-2.5 m)
SC-65 = <20, <20, <20, <4, <4, <2, <2 ng/g
dw (0-0.25, 0.25-0.5, 0.5-0.75, 0.75-1,
1-1.5,3-3.5,3.5^ m)
SC-66 = <4, <4, <4, <2, <2, <2, <2, <2, <2,
<2 ng/g dw (0-0.25, 0.25-0.5, 0.5-0.75,
0.75-1, 1-1.5, 1.5-2,2-2.5,2.5-3,3-3.5,
3.5^1 m)
SC-58b = <10 ng/g dw (4-5, 5-6, 9-10,14-
15 m)
SC-59b = <10, <10,2, <10, <10 ng/g dw (3-
4,4-5,6-7, 9-10,14-15 m)
SC-65b = <10 ng/g dw (4-5, 7-9, 9-11,14-
15 m)
SC-67 = <20, <20, <10, <10 ng/g dw (0-1,
1.3-2,2-3,4-5 m)
Area 3:
SC-40 = <4 ng/g dw (0-1 m)
SC-43 = <2 ng/g dw (1-2 m)
SC-45 = <2 ng/g dw (0-1 m)
SC-47 = <20 ng/g dw (0-1 m)
SC-48 = <4 ng/g dw (0-1 m)
SC-41 = <10 ng/g dw (0-0.25,1-2 m)
SC-42 = <10 ng/g dw (0-0.25,1-2, 3^1 m)
Area 4:
SC-33 = <20 ng/g dw (0-1 m)
SC-34 = <4 ng/g dw (1-2 m)
C-108
-------
FINAL
2024
Study
Location
Site Details
Results
SC-35 = <2 ng/g dw (0-1 m)
SC-36 = <4 ng/g dw (0.3-1 m)
SC-37 = <20 ng/g dw (0.1-1.1 m)
SC-37b = <10 ng/g dw (0-0.25,1-1.5, 3^1
m)
SC-38 = <10 ng/g dw (0.25-1,2-3 m)
Area 5:
SC-10 = <2 ng/g dw (0-0.25, 0.25-0.5, 0.5-
0.75, 0.75-1,1-1.5,1.5-2, 2-2.5, 2.5-3,
3-3.5 m)
SC-11 = <2 ng/g dw (0-0.25, 0.25-0.5, 0.5-
0.75,0.75-1,1-1.5,1.5-2, 2-2.5 m)
SC-12 = <2 ng/g dw (0-0.25, 0.25-0.5, 0.5-
0.75 m)
Area 6:
SC-21 = 5, <10, <10, <10, <10 ng/g dw (0-
0.25, 0.25-1,2-3, 8-9,13-15 m)
SC-22 = 3,2, <10, <20 ng/g dw (0-0.25,
0.25-1, 1-2, 3^1 m)
SC-23 = 15, 3, <10, <10 ng/g dw (0-0.25,
0.25-1, 1-2, 3^1 m)
SC-24 = 10, <10, 3 ng/g dw (0-0.25,1-2,
3^1 m)
SC-25 = 3, <10 ng/g dw (0-0.25, 1.5-2 m)
SC-26 = <10 ng/g dw (2-3, 4-5 m)
(LOQ = 2 ng/g dw)
Skaaretal. (2019)
Norway (Ny-Alesund)
Samples collected in June 2016 in and around the
international research facilities (Ny-Alesund) near
local firefighting training site. Background soil
samples were collected at representative locations.
Background: n = 8, DF 0%
Contaminated: n = 2, DFa 50%, range =
<0.005-0.73 ng/g dw
(IDL = 0.026 ng; LOD = 0.005 ng/g dw; LOQ
= 0.01 ng/g dw)
*Table 1 and Table S2 reported a total of nine
samples across background and contaminated
sites; however, Tables Sll and SI3 report a
total of ten samples (two contaminated sites
from Table Sll and eight background sites
from Table SI3
Hale et al. (2017)
Norway (Gardermoen)
Samples collected in June 2015 from six locations
around a firefighting training facility west of the Oslo
airport site. Samples were taken at 0-1 m, 1-2 m, 2-3
n = 22, DF 40%, range = 2.8^11.3 ng/g
(LOD =1 ng/g)
C-109
-------
FINAL
2024
Study
Location
Site Details
Results
m, and 3 to groundwater table level (which was in all
cases above 4 m). Facility was established in 1989 and
AFFF was used extensively. AFFF containing PFOS
was banned at the facility in 2007 and a complete ban
on organofluorine AFFF was enforced in 2011. The
soil is known to be contaminated with a range of
perfluorinated compounds.
*Range reported for detects
*The DF and range extracted are reported in
the results (Section 3.1); however, Table S2 of
the individual sample data show all
concentrations ranging from <1.8 to <2.5 ng/g
Harrad et al. (2020)
Ireland (multiple cities)
Samples collected from ten municipal solid waste
landfills upwind and downwind at each site between
November 2018 and January 2019. At each
upwind/downwind location, nine sub-samples of soil
were taken in a "W" formation. Samples were
collected from the same areas as air samples and were
taken within the boundaries of the landfill operational
facility. Soil used as capping on landfill cells was not
sampled to ensure soil samples were not collected
from soil placed after landfill operations ceased and
that farming activities would not influence
concentrations found. Waste accepted by the landfills
included: municipal solid waste, industrial (non-
hazardous) waste, construction & demolition, and
biomedical waste.
Downwind:
n = 9, DFa 89%, mean, median (range) =
0.0045, 0.0043 (<0.001-0.0077) ng/g dw
Upwind:
n = 7, DFa 100%, mean, median (range) =
0.014, 0.006 (0.0029-0.033) ng/g dw
(LOD = <0.001 ng/g dw)
*Non-detects replaced by Vi LOD
*Soil samples from three upwind locations
and one downwind location destroyed in
transit from field to laboratory
Grannestad et al. (2019)
Norway (Granasen, Jonsvatnet)
Upper layer soil samples (3-10 cm in depth) collected
in June 2017 and 2018 from Granasen (skiing area)
and Jonsvatnet (reference site). Five samples per year
were analyzed for each site. Located 10 km from
Trondheim city center, Granasen is the main arena for
winter sports in Trondheim and hosts an annual ski
jumping World Cup event and regional, national, and
international competitions in cross-country skiing.
Located 15 km away from Trondheim city center,
Jonsvatnet is a natural forest area not used for ski-
sports and is in the vicinity of an ecological farm next
to Lake Jonsvatnet. The two study areas have similar
vegetation.
Reference area: n = 10, DF 70%, mean (range)
= 0.198 (
-------
FINAL 2024
Notes: AFFF = aqueous film-forming foam; DF = detection frequency; dw = dry weight; LOQ = limit of quantitation; LHWA = Little Hocking Water Association; LOD = limit
detection; MDL = method detection limit; 0.5x, 1 x; or 2* = Vi, 1, or 2 times the agronomic rate of biosolids application to meet nitrogen requirements of the crop; ND = not
detected; NR = not reported; PFAA = perfluoroalkyl acids ; RL = reporting limit; WWTP = wastewater treatment plant.
C-lll
-------
FINAL
2024
C.3.7. Sediment
When released into water, based on its physico-chemical properties, PFNA is expected to adsorb
to suspended solids and sediments (NCBI, 2022b). The EPA did not identify studies conducted
in the U.S. that reported the occurrence of PFNA in sediment. Concentrations of PFNA in
sediment samples collected from the Hudson Bay region of northeast Canada ranged
from <0.06 ng/g to 0.14 ng/g (dry weight) (NCBI, 2022b; Kelly et al., 2009).
C.4. Recommended RSC
The EPA followed the Exposure Decision Tree approach to determine the RSC for PFNA
(USEPA, 2000b). The EPA first identified three potential populations of concern (Box 1):
pregnant women and their developing fetuses, lactating women, and women of childbearing age
(see Section 2.3.2). However, limited information was available regarding specific exposure of
these populations to PFNA in different environmental media. The EPA considered exposures in
the general U.S. population as likely being applicable to these two populations. Second, the EPA
identified several relevant PFNA exposures and pathways (Box 2), including dietary
consumption, incidental oral consumption via exposure to dust, consumer products, and soil,
dermal exposure via soil, consumer products, and dust, and respiration via ambient air. Several
of these may be potentially significant exposure sources. Third, the EPA determined that there
was not adequate quantitative data to describe the central tendencies and high-end estimates for
all of the potentially significant sources (Box 3). For example, studies from Canada and Europe
indicate that indoor air may be a significant source of exposure to PFNA. At the time of the
literature search, the EPA was unable to identify studies assessing PFNA concentrations in
indoor air samples from the U.S. and therefore, the agency does not have adequate quantitative
data to describe the central tendency and high-end estimate of exposure for this potentially
significant source in the U.S. population. However, the agency determined there were sufficient
data, physical/chemical property information, fate and transport information, and/or generalized
information available to characterize the likelihood of exposure to relevant sources (Box 4).
Notably, based on the studies summarized in the sections above, there are significant known or
potential uses/sources of PFNA other than drinking water (Box 6), though there is not
information available on each source to make a characterization of exposure (Box 8A). For
example, there are several studies from the U.S. indicating that PFNA may occur in dust sampled
from various microenvironments (e.g., homes, offices, daycare centers, vehicles). However, the
majority of studies sampled in only one location and few studies examined dust samples outside
of the home (e.g., one study assessed PFNA occurrence in dust sampled from vehicles).
Additionally, though several studies from around the U.S. measured PFNA concentrations in
dust from houses, the detection frequencies in these studies varied widely (from 35% to 100%)
and may be a result of uncertainties including home characteristics, behaviors of the residents,
and the presence or absence of PFNA-containing materials or products (Haug et al., 2011).
Therefore, it is not possible to determine whether dust can be considered a major or minor
contributor to total PFNA exposure. Similarly, it is not possible to determine whether the other
potentially significant exposure sources such as seafood and consumer products should be
considered major or minor contributors to total PFNA exposure. Given these considerations,
following recommendations of the Exposure Decision Tree (USEPA, 2000b), the EPA
recommends an RSC of 20% (0.20) for PFNA.
C-112
-------
FINAL
2024
Appendix D. PFHxS: Summary of Occurrence in
Water and Detailed Relative Source Contribution
D.l. Occurrence in Water
The production of PFHxS and its use as a raw material or precursor for manufacturing PFAS-
based products, as well as its previous use in firefighting foam and carpet treatment solutions,
could result in its release to the aquatic environment through various waste streams (NCBI,
2022a). PFHxS has an estimated water solubility of 6,200 |ig/L (6.2 mg/L) at 25°C and when
released to surface water, it is not expected to adsorb to suspended solids and sediment (NCBI,
2022a). Volatilization from water surfaces is not expected to be an important fate process for
PFHxS (NCBI, 2022a).
D.l.l. Groundwater
Several studies have evaluated the occurrence of PFHxS in groundwater in both the United
States and Europe. PFHxS was detected in at least one groundwater sample site in each study in
the U.S (Table D-l). Lindstrom et al. (2011) analyzed well water samples in Decatur, Alabama.
The samples were collected in February 2009 from farms that had applied PFC-contaminated
biosolids to local agricultural fields as a soil amendment for at least 12 years. PFHxS was
detected in two wells at concentrations of 56.5 and 87.5 ng/L. In another study, median and
maximum groundwater of 870 ng/L and 290,000 ng/L (0.870 |ig/L and 290 |ig/L), respectively,
were detected at 10 U.S. military installations (Anderson et al., 2016). Three other studies of
groundwater known to be impacted by nearby AFFF use similarly had PFHxS concentrations
ranging from 36-120,000 ng/L and detection frequencies of 100% (Steele et al., 2018; Eberle et
al., 2017; Moody et al., 2003).
Post et al. (2013) assessed raw water from public drinking water system intakes that were chosen
to represent New Jersey geographically but were not necessarily associated with any known
PFAS release. PFHxS was found in 2 of 18 systems at levels <10 ng/L. Appleman et al. (2014)
evaluated groundwater contaminated by wastewater effluent discharge. At this site, detection
frequency was 100%, but PFHxS levels did not exceed 11 ng/L. Procopio et al. (2017) collected
groundwater from areas downstream of a manufacturer of PF AS-containing products but found
minimal PFHxS in only 5% of samples, all of which ranged from non-detects to 5.5 ng/L. Boone
et al. (2019) evaluated 17 PFAS in source and treated waters collected in 2010-2012. Of the
three groundwater sources evaluated, PFHxS was detected in two out of three samples at levels
of 1.88 and 44.8 ng/L. In the final U.S.-based study, Quinones and Snyder (2009) examined
levels of eight PFAS at two sites in Las Vegas Wash, Nevada that were highly impacted from
treated wastewater. Samples were collected in 2008 as part of a study to assess both raw and
treated water from utilities producing at least 75 megaliters of finished water per day. Mean
PFHxS levels at the two sites were 6.8 and 5.6 ng/L at sites 1 (n = 7) and 2 (n = 8), respectively.
Of the studies conducted in Europe, 4 studies (Barreca et al., 2020; Sammut et al., 2019) were
conducted in areas not associated with any known PFAS release. At these sites, the detection
frequency of PFHxS was 0-40% with a maximum level of 32 ng/L. The remaining nine
D-l
-------
FINAL
2024
European studies evaluated groundwater samples from sites with known or suspected PFAS
releases associated with fluorochemical manufacturing (Boiteux et al., 2012; Loos et al., 2010)
or AFFF use (Boiteux et al., 2017; Dauchy et al., 2012). Source categories for Gobelius et al.
(2018) included fire training sites, but also included landfill/waste disposal sites, skiing areas,
urban areas, and areas of unspecific industries. Of the sites with known sources of
contamination, higher detection frequencies (up to 100%) and greater PFHxS levels (up to 3,470
ng/L) were reported at sites with AFFF use. Two related studies (Boiteux et al., 2017; Dauchy et
al., 2012) sampled alluvial wells downstream of a fluorochemical manufacturing facility in
France. Preliminary results showed low levels of PFHxS (up to 11 ng/L; (Dauchy et al., 2012)),
but PFHxS was not detected in samples from the latter study (Boiteux et al., 2017).
D-2
-------
FINAL
2024
Table D-l. Studies Reporting Occurrence of PFHxS in Groundwater
Study
Location
Site Details
Results
United States
Procopio etal. (2017)
United States (New Jersey)
Groundwater from an industrial/business park located
within the South Branch Metedeconk River watershed,
where there was suspected illicit discharge to soil and
groundwater from a manufacturer of industrial fabrics,
composites, and elastomers that use or produce
products containing PFAAs. Samples were collected
following the installation of 16 temporary monitoring
wells by the NJ Geological and Water Survey or a
contract driller during August 2013 (sampling event
#7) and June-July 2014 (sampling event #8). Samples
were taken from the upper 1.5 m (5 ft) of the water
table from each well, except for one "profile well" in
which samples were collected at three different depths
(3.7-4.6, 6.7-7.6, and 10.7-11.6 m below grade; 12-
15,22-25, and 35-38 ft below grade, respectively).
n = 19, DFa 5%, range = <5-5.5 ng/L
(Minimum RL = 5 ng/L)
Post et al. (2013)
United States (New Jersey)
Raw water collected from public drinking water
system intakes in two sampling campaigns. In the first
sampling campaign, samples from 18 drinking water
systems were collected between August 2009 and
February 2010 from 1 confined well (sunk into an
aquifer located between two impermeable strata) and
17 unconfined wells in the upper unconfmed aquifer;
sites were chosen to represent NJ geographically and
included 1 site with a nearby industrial facility that
previously used large quantities of PFNA (site 5). In
the second sampling campaign, samples from two
drinking water systems (PWS-A and PWS-B) were
collected in 2010-2013 from five unconfined wells.
Groundwater at these two systems were known to be
contaminated by PFOA.
1st sampling campaign:
n = 18, DF 11%, range = ND-10 ng/L
2nd sampling campaign:
PWS-A, WF1A: n = 5, DF 0%
PWS-A, WF1B: n = 4, DF 0%
PWS-A, WF2A: n = 9, DFa14%, range =
ND-6 ng/L
PWS-A, WF2B: n = 9, DF 0%
PWS-B: n = 8, DF 0%
(Minimum RL = 5 ng/L)
Lindstrom et al. (2011)
United States (Decatur,
Alabama)
Thirteen samples collected in February 2009 from 13
wells located on farms with historical land application
of PFC-contaminated biosolids to local agricultural
fields between 1995 and 2008. Biosolids obtained
from local municipal WWTP where sources
discharging to the WWTP included facilities involved
in the production and use of fluoropolymers,
n = 13, DF (frequency of quantification)315%,
range =
-------
FINAL
2024
Study
Location
Site Details
Results
fluorocarbon fibers, polymers, polymer films, and
resins.
Boone et al. (2019)
United States (unspecified)
Three groundwater sites used as source waters for
three DWTPs, collected in 2010-2012; some locations
with known or suspected sources of wastewater in the
source water, but study did not differentiate which
locations had known or suspected sources.
n = 3, DFa 67%, range = NCM4.8 ng/L
(LCMRL = 0.034 ng/L)
Appleman et al. (2014)
United States (New Jersey)
Groundwater source water for five DWTPs, sampled
November 2011 to September 2012. Majority of the
utilities were selected because they were either known
from previous monitoring or expected based on their
source waters to contain detectable PFAS (i.e.,
impacted by upstream wastewater effluent discharge).
Two sites were sampled twice and three sites were
sampled only Once.
n = 7, DFa 100%, meana (range) = 5.5 (0.48-
11) ng/L
(Method RL = 0.25 ng/L)
Quinones and Snyder (2009)
United States (Nevada)
Samples collected in 2008 from two groundwater sites
in Las Vegas Wash, Nevada that were highly impacted
from treated wastewater.
Site 1: n = 7, DF NR, mean (maximum) = 6.8
(24) ng/L
Site 2: n = 8, DF NR, mean (maximum) = 5.6
(13) ng/L
(Method RL= 1.0 ng/L)
Anderson et al. (2016)
United States (national)
Forty AFFF-impacted sites from ten active U.S. Air
Force installations with historic AFFF release between
1970 and 1990 that were not related to former fire
training areas. It is assumed that the measured PFAS
profiles at these sites reflect the net effect of several
decades of all applicable environmental processes.
AFFF-impacted sites included emergency response
locations, hangers and buildings, and testing and
maintenance related to regular maintenance and
equipment performance testing of emergency vehicles
and performance testing of AFFF solution. Previous
remedial activities for co-occurring contaminants were
not specifically controlled for in the site selection
process; active remedies had not been applied at any
of the sites selected. Approximately ten samples were
collected between March and September 2014 at each
site; sites were grouped according to volume of AFFF
release—low-volume typically had a single AFFF
release, medium-volume had one to five releases, and
high-volume had multiple releases. Groundwater
Overall: n = 149, DF 94.93%, median
(maximum) = 870 (290,000) ng/L
Breakdown by site group:
Emergency Response (low-volume release):
n = 24, DF 87.5%), mean (range) = 20,100
(10-270,000) ng/L
Hangars and Buildings (medium-volume
release):
n = 100, DF 90.6%o, mean (range) = 71,400
(23-910,000) ng/L
Testing and Maintenance (high-volume
release):
n = 25, DF 100.0%o, mean (range) = 400
(390^120) ng/L
(Median RL = 7 ng/L)
*Minimum of detected values reported
D-4
-------
FINAL
2024
Study
Location
Site Details
Results
samples were collected from existing monitoring wells
and temporary monitoring wells installed with direct
push technology.
*Median calculated using quantified
detections
*Non-detects were substituted with Vi the
reporting limit
Steele et al. (2018)
United States (Alaska)
Monthly samples collected from a military installation
during July 2016-March 2017; six wells from around
the installation were sampled each month, along with
a seventh well that was only sampled in July 2016.
PFAS contamination predominately from prior legacy
AFFF use. Wells selected based on historical sample
data indicating PFAS contamination.
Data for July, August, September, October,
November, December, January, and February,
respectively:
Well A: 170,140,130,180,150, 96,120,100
ng/L
Well B: 220, 360, 360, 370, 410, 300, 400,
370 ng/L
Well D: 120 ng/L (for July only; no values
provided for other months)
Well E: 36, 39,48, 240, 210, 94, 85, 77 ng/L
Well F: 82,110,110,240,150,110,100,110
ng/L
DK: 460, 590, 590, 700, 700, 530, 690, 740
ng/L
FG: 60, 69, 75, 61, 93, 64, 67, 59 ng/L
(Minimum DL not reported)
Eberle et al. (2017)
United States (Joint Base
Langley-Eustis, Virginia)
Pilot testing area in former fire training area (Training
Site 15) at Joint Base Langley-Eustis where monthly
fire training activities were conducted from 1968 to
1980 in a zigzag pattern burn pit. Facility was
abandoned in 1980 but irregular fire training activities
using an above-ground germed burn pit continued
until 1990. Groundwater samples collected for
screening/site characterization (April and December
2012), and for pre- (April 2013) and post- (October
2013 and February 2014) in situ chemical oxidation
treatment using a peroxone activated persulfate
(OxyZone) technology. Treatment was conducted in
Test Cell 1 over 113 days (April through August
2013). Pre-treatment samples were collected from 14
wells screened in the deep zone, and 3 wells screened
in the shallow zone. Post-treatment samples were
collected from the same wells as the pre-treatment
samples with an additional three wells (two shallow,
one deep) sampled. Wells EC-1, EC-2, EC-3, EC-4,1-
Screening/site characterization:
EC-1 (deep, sentry): 19,000 ng/L
EC-2 (deep, sentry): 39,000 ng/L
1-1 (deep): 32,000 ng/L
1-2 (deep, sentry): 57,000 ng/L
1-4 (deep): 80,000 ng/L
1-5 (shallow): 13,000 ng/L
1-6 (shallow): 13,000 ng/L
MW-2904 (deep): 9,400 ng/L
U-16D (deep): 44,000 ng/L
U-16S (shallow): 19,000 ng/L
D-5
-------
FINAL
2024
Study
Location
Site Details
Results
2, and 1-3 were sentry wells to monitor the possible
migration of oxidants and contaminants outside Test
Cell 1. Limited data reported for post-treatment
samples.
Pre-treatment: (values reported for two
different laboratories: TA; CSM)
EC-2 (deep, sentry): 40,000; 48,000 ng/L
EC-3 (deep, sentry): 46,000; 59,400 ng/L
1-1 (deep): 25,000; 24,900 ng/L
1-2 (deep, sentry): 23,000; 20,400 ng/L
1-4 (deep): 64,000; 66,400 ng/L
Post-treatment:
1-4 (deep), U-16D (deep), U-17D (deep),
and U-20D (deep) showed a 56% reduction
in PFHxS compared to pre-treatment values
(LOQ = 1 ng/L)
Moody et al. (2003)
United States (Oscoda,
Michigan)
Groundwater samples collected from ten wells during
November 1998 and June 1999 from plume at Fire
Training Area Two (FTA-02) at the former Wurtsmith
Air Force Base; FTA-02 was used from 1952 to 1993
to train U.S. military personnel in firefighting
procedures and included flooding a concrete pad with
flammable liquids, igniting the fluids, and
extinguishing the fire with firefighting agents
including AFFF. Minimum of five years since active
firefighting activity.
Well ID (distance from fire training pad):
FT2 (17 m): n = 1, point = 120,000 ng/L
FT3 (18 m): n = 1, point = 104,000 ng/L
ML3 (114 m): n = 1, point = 70,000 ng/L
ML8 (121 m): n = 1, point = 39,000 ng/L
FT8 (183 m): n = 1, point = 30,000 ng/L
FT9 (183 m): n = 1, point = 46,000 ng/L
FT12 (305 m): n = 1, point = 23,000 ng/L
FT14 (305 m): n = 1, point = 27,000 ng/L
FT18 (518 m): n = 1, point = 33,000 ng/L
FT17 (540 m): n = 1, point = 9,000 ng/L
(LOD = 3,000 ng/L; LOQ = 13,000 ng/L)
*Point value reported is from five replicate
analyses of one sample
Europe
Bach et al. (2017)
France (southern)
Samples were collected from alluvial wells that
provide source water for two DWTPs. The two
DWTPs are located on both sides of a river, ~15 km
downstream from an industrial site where two
facilities produce fluoropolymers; the industrial site
Alluvial wells for DWTP A:
April 2013: n = 7, DFa 29%, range = <4-7
ng/L
D-6
-------
FINAL
2024
Study
Location
Site Details
Results
discharges its effluents at three points along a river.
The alluvial wells are located along the river, with
wells for the first DWTP (DWTP A) located on the
left shore and alluvial wells for the second DWTP
(DWTP B) located on the right shore, on an island
formed by a backwater. Sample collection occurred in
April, July, October, and December 2013.
July 2013: n = 7, DFa 43%, range = <4-8
ng/L
October 2013: n = 7, DFa 43%, range = <4-
8 ng/L
December 2013: n = 7, DFa 51%, range =
<4-6 ng/L
Alluvial wells for DWTP B:
April 2013: n = 8, DFa 50%, range = <4-7
ng/L
July 2013: n = 8, DFa 63%, range = <4-8
ng/L
October 2013: n = 7, DFa 51%, range = <4-
8 ng/L
December 2013: n = 8, DFa 63%), range =
<4-6 ng/L
(LOQ = 4 ng/L)
*DF represents frequency of quantification
Boiteux et al. (2017)
France (northern)
Samples were collected in four sampling campaigns
(May, July, October, and December 2013) from
alluvial wells that provide source water for two
DWTPs. The two DWTPs (A and B) are located
downstream of an industrial WWTP that processes
raw sewage from a facility that manufactures
fluorotelomer-based products and side-chain-
fluorinated polymers used in firefighting foams and
stain repellents.
DWTP A is located 15 km downstream from the
WWTP and is supplied by five alluvial wells. DWTP
B is located 20 km downstream of the WWTP and is
supplied by four alluvial wells.
DWTP A:
May 2013: n = 5, DF 0%
July 2013: n = 5, DF 0%
October 2013: n = 5, DF 0%
December 2013: n = 5, DF 0%
DWTP B:
May 2013: n = 4, DF 0%
July 2013: n = 4, DF 0%
October 2013: n = 4, DF 0%
December 2013: n = 4, DF 0%
(LOQ = 4 ng/L)
*DF represents frequency of quantification
D-7
-------
FINAL
2024
Study
Location
Site Details
Results
Dauchy et al. (2012)
France (unspecified)
Raw water sampled in June 2010 from four
monitoring wells at a fluoropolymer manufacturing
plant (PI3, PI4, PI 5, P01). Groundwater flowed from
well P14 to P01 and well PI 5 is nearest to the
polyvinylidene fluoride production area.
Raw water resources also collected from two DWTPs
(five sampling sites - DWA-1, DWA-2, DWA-3,
DWA-4, DWB-1); the first DWTP (DWA) is supplied
by four alluvial wells, and the second DWTP (DWB)
is supplied by one alluvial well. The two DWTPs are
located on both sides of a river, 15 km downstream of
fluorochemical manufacturing facility. The river
receives wastewater from many domestic and
industrial activities.
Fluoropolymer manufacturing plant:
P13: n = NR, DF NR, 68 ng/L
P14: n = NR, DF NR, 8 ng/L
PI 5 and P01 not quantifiable due to dilution
or matrix effects
DWA:
DWA-1: n = NR, DF NR, 11 ng/L
DWA-2: n = NR, DF NR, 8 ng/L
DWA-3: n = NR, DF NR, 4 ng/L
DWA-4: n = NR, DF NR, <4 ng/L
DWB:
DWB-1: n = NR, DF NR, 5 ng/L
(LOQ = 4 ng/L)
* Study did not indicate whether
concentrations reported were point values or
means
Harrad et al. (2020)
Ireland (multiple cities)
Groundwater samples collected between November
2018 and January 2019 from ten municipal solid waste
landfills at two sampling points down-gradient from
the main body of each landfill. Each sampling point
consisted of a borehole leading down to water
reservoirs at a minimum depth of 5 m below ground
level. Waste accepted by the landfills included:
municipal solid waste, industrial (non-hazardous)
waste, construction and demolition, and biomedical
waste.
n = 10, DF 20%, mean, median (range) = <0.1,
<0.1 (<0.1-0.28) ng/L
(LOD = <0.1 ng/L)
*Non-detects replaced by Vi LOD
Gobelius et al. (2018)
Sweden (national)
Sampling conducted between May and August 2015,
with the majority in July and a few samples in
September and November 2015. Samples were
collected in 21 regional counties by the County
Administration Boards. Sampling locations selected
based on potential vicinity of PFAS hot spots (i.e., fire
training sites, unspecific industry, sewage treatment
plant effluent, landfill/waste disposal, skiing, and
urban areas) and/or importance as a drinking water
source. Sample numbers varied for each county and
sampling sites were spread unevenly across Sweden.
n = 161, DFa 37%, range = <0.15-80 ng/L
(MDL = 0.15 ng/L)
D-8
-------
FINAL
2024
Study
Location
Site Details
Results
Boiteux et al. (2012)
France (national)
Raw water from DWTPs distributed across 100
French departments to represent -20% of the national
water supply flow; samples collected during two
sampling campaigns in July-September 2009 (first
campaign) and June 2010 (second cam-aign - focused
on sites from first sampling campaign that had PFC
levels >LOQ). Some sites possibly affected by
commercial/industrial releases.
Overall: n = 196, DF (frequency of
quantification) 18%, maximum = 32 ng/L
1st Sampling Campaign
n = 163, DF 31%), mean, median (maximum)
= 1, <1 (32) ng/L
2nd Sampling Campaign
n = 33, results not reported
(LOD =1.3 ng/L, LOQ = 4 ng/L)
Loos et al. (2010)
23 European countries
Groundwater collected from 164 groundwater
monitoring stations of participating European Union
Member State laboratories during an 8-week window
in Fall 2008. There were no strict selection criteria for
the sampling sites such as "representative" or
"contaminated". Most monitoring stations were
"official" monitoring stations also used for drinking
water abstraction.
n = 164, DF 34.8%o, mean, median (maximum)
= 1,0 (19) ng/L
(LOD = 0.4 ng/L)
Dauchy et al. (2019)
France (unspecified)
Samples collected in two sampling campaigns in and
around site where fluoro surfactant-based foams have
been used extensively. From 1969 to 1984, the site
was an oil refinery, with the exact location of the
firefighting training area, frequency of training
sessions, and history of firefighting training activities
unknown. From 1987 to date, it has been a large
training area for firefighters. First sampling campaign
collected 13 samples from 9 monitoring wells and 4
springs in June 2015. Second sampling campaign
collected from four monitoring wells in October 2016.
Monitoring wells MW-1 to MW-5 were located
upgradient from the firefighter training site around a
landfill site. Monitoring well MW-11 and springs SW-
A, SW-B, and SW-D located downgradient from the
landfill or firefighter training site but not in the
direction of groundwater flow. Monitoring wells MW-
6 to MW-13 and spring SW-C were located
downgradient from the firefighter training site in the
direction of groundwater flow.
Upgradient:
Monitoring wells: n = 5, DFa 40%o, range =
<4^12 ng/L
Downgradient but not in the direction of
groundwater flow:
Monitoring wells: n = 1, point = 42 ng/L
Spring water: n = 3, DFa 33%o, range = <4-7
ng/L
Downgradient in the direction of groundwater
flow:
Monitoring wells: n = 7, DFa 100%o, meana
(range) = 660 (26-2,860) ng/L
Spring water: n = 1, point =122 ng/L
(LOQ = 4 ng/L)
Lfoisaster et al. (2019)
Norway (unspecified)
Firefighting training site with an airport that
extensively used AFFF containing PFOS since the
early 1990s until 2001 when it was replaced by
fluorotelomer containing AFFF. All PFAS containing
n = 19, DF NR, mean* = 2,900 ng/L
(LOD/LOQ not reported)
*Mean estimated from Figure 4b in the paper
D-9
-------
FINAL
2024
Study
Location
Site Details
Results
firefighting foams was banned at the airport in 2011.
Groundwater samples collected in 2016 at five
pumping wells installed down gradient of the site to
intercept and pump and treat the plume spreading
frlOirefightrefighting training site. A total of 19
sampling campaigns were performed.
Dauchy et al. (2017)
France (unspecified)
Samples collected in the vicinity of three sites (A, C,
D) where fluoro surfactant-based foams are or were
being heavily used. Site A is an oil storage depot
located in a river port. In June 1987, a large explosion
occurred in the depot and the fire was extinguished by
applying a large amount of fluoro surfactant-based
foams. Two groundwater samples were collected in
October 2014 and March 2015 from a monitoring well
located in the center of the depot. The water table lies
2.5 - 3.5 m below the ground.
Site C is a military airport, with the exact location of
the training area, frequency of the training sessions,
and history of the firefighting training activities
unknown. The well supplying the DWTP was sampled
in March 2015.
Site D is a training center for firefighters. From 1969
to 1984, the site was an oil refinery. Starting in 1987,
the site became a training area for firefighters, with
exercises carried out directly on the soil. From the
1990s, some exercise areas were covered with
concrete. In November 2014, groundwater samples
were collected from five springs.
Site A:
October 2014: n = 1, point =139 ng/L
March 2015: n= 1, point = 136 ng/L
Site C: n = 1, point = 25 ng/L
Site D: n = 5, DFa 20%, range = <4-132 ng/L
(LOQ = 4 ng/L)
Filipovic et al. (2015)
Sweden (Stockholm)
Groundwater samples collected at the airfield and in
the vicinity of the airport of the closed air force base
F18 in Tullinge Riksten, 19 km south of Stockholm
city center, where AFFFs were used. Samples
collected in two sampling campaigns in December
2011 and May 2012. The air force base was formally
demobilized in 1986 but continued to be used as an air
force school for combat command and air surveillance
until 1994. Of note, the air force base encountered
numerous accidents and incidents during the transfer
from propeller era to the jet engine era, including
planes crashing upon takeoff and landing, fire
incidents, accidental dispersion of iet engine starting
n = 16, DF 69%, range = <0.5-3,470 ng/L
*Highest concentrations of 2,960 and 3,470
ng/L detected at sites G5 and G6
(MDL = <0.5 ng/L)
D-10
-------
FINAL
2024
Study
Location
Site Details
Results
fuel. The area was sold to a land developer in 1996
and is in the process of being transformed into a
municipal area. Groundwater flow is directed from the
military airfield towards Lake Tullingesjon. Sampling
sites included locations under the main firefighting
training facility (sites G5 and G6). Other groundwater
sampling sites were not mapped to specific locations
(note that soil samples were collected at the main
firefighting training facility, intermediate soil depot,
J34 Hawker Hunter site, old fire station, and soil
depot).
Gyllenhammar et al. (2015)
Sweden (Uppsala)
Three observation well sites (Tuna backar: n = 3
wells; Svartbacken: n = 1 well, Libroback: n = 2
wells;) were sampled from September 2012 to January
2013.
Four DWTP production well sites (Storvad: n = 9
wells; Galgbacken: n = 1 well; Stadstradgarden and
Kronasen: n = 6 wells; Sunnersta: n = 5 wells) were
sampled from July 2012 to February 2014.
One private well (Klastorp) was sampled in September
2012.
All wells located downstream of a military airport
with firefighting training activities up to the year
2003. It is not known when the usage of AFFF started.
Observation wells:
Tuna backar: n = 3, DF 100%, median =
690 ng/L
Libroback: n = 4, DF 0%
Svartbacken: n = 3, DF 100%, median =
250 ng/L
Production wells:
Storvad: n = 12, DF 0%
Galgbacken: n = 7, DF 0%
Stadstradgarden and Kronasen: n = 103, DF
100%), median = 83 ng/L
Sunnersta: n = 50, DFa 52%o, median = 8
ng/L
Private well:
Klastorp: n = 1, point = 16 ng/L
(MDL = 10 ng/L)
Wagner et al. (2013)
Germany (unspecified)
Groundwater samples collected downstream from a
site contaminated by PFC-based AFFFs from
firefighting activities. Sampling year not provided.
Samples used to test out a new analytical protocol to
determine trace levels of adsorbable organic fluorine.
n = 3, DFa 100%o, mean (range) = 368 (230-
510) ng/L
(LOD = 50 ng/L F"; LOQ = 150 ng/L F")
*PFHxS concentrations were calculated using
the fluorine concentrations reported in Table 4
D-ll
-------
FINAL
2024
Study
Location
Site Details
Results
Barreca et al. (2020)
Italy (Lombardia region)
Fifty-seven groundwater sampling stations throughout
the region. Samples collected in 2018.
n = 130, DF 0%
(LOQ = 1 ng/L)
Sammut et al. (2019)
Malta
Groundwater collected from ten boreholes at different
areas on the island during November and December
2015 and January 2016. Collection sites were the most
commonly used extraction sites by the Malta Water
Services Corporation for water extraction as well as
for sampling for water quality analysis.
n = 10, DF 70%, range = ND-2.22 ng/L
(LOD = 0.02 ng/L; LOQ = 0.04 ng/L)
*Of the ten samples, three were ND and three
were
-------
FINAL
2024
D.1.2. Surface Water
Overall, almost all U.S. studies reported PFHxS detected in at least one surface water sample site
in each study. Three studies investigated surface water upstream and downstream of
fluoropolymer facilities, with some sites also downstream of other potential PFAS sources (e.g.,
landfills, WWTP) (Galloway et al., 2020; Newsted et al., 2017; Newton et al., 2017). Galloway
et al. (2020) assessed several rivers and tributaries along the Ohio River in three sampling trips
in 2016. The sampling sites ranged from upstream, downstream, and north/northeast of a
fluoropolymer facility and known PFAS containing landfills. In June 2016, samples were
collected on a 188 km stretch of the Ohio River, from 130 km downstream to 58 km upstream of
the facility, and tributaries that pass near known PFAS-containing landfills. In July 2016,
samples were collected from lakes, rivers, and creeks to the north and northeast of the facility as
far as 16 km downwind. The December 2016 trip expanded the collection radius to more than 48
km downwind to the north and northeast of the facility. PFHxS was detected in 96% of samples
(n = 26) in June 2016, however all detects were below the LOQ (10 ng/L). Of the second
sampling trip, PFHxS was detected at levels above the LOQ in three samples in July 2016
(n = 25), ranging from 64.5-79.0 ng/L. Finally, in December 2016, PFHxS was detected at levels
above the LOQ in three samples ranging from 10.1-14.4 ng/L, detected but below the LOQ in 16
samples, and not detected in 21 samples. In Newsted et al. (2017), surface water samples were
collected in August 2011 from a 3-mile section of the Upper Mississippi River: ten sampling
reaches (three samples each) in an area between Ford Dam (between Minneapolis and St. Paul)
and Hastings Dam (near Hastings) and which had been subject to 10-15 years of actions to
reduce PFAS contamination from 3M Cottage Grove plant and other commercial/industrial
entities. PFHxS was detected in one sample from reach 10, immediately downstream the 3M
Cottage Grove facility outfall, at a concentration of 60.5 ng/L. PFHxS in all other samples was
below the LOQ of 2.0 ng/L. Authors were not able to observe a clear and consistent time trend in
water concentrations. Newton et al. (2017) investigated surface water upstream and downstream
of facilities that manufactured or used fluorinated materials along the Tennessee River near
Decatur, Alabama. Six sampling sites were located upstream of the manufacturing facilities and
three sites were downstream. Among the upstream sites, three were also upstream of a WWTP.
All samples were collected in October 2015. PFHxS was detected at one downstream site at a
concentration of 39 ng/L; authors suggested elevated PFAS concentrations at downstream sites
resulted from infiltration from groundwater or runoff from soil.
In five studies, sampling locations included surface waters potentially impacted by current and/or
historic use of AFFFs (Genualdi et al., 2017; Anderson et al., 2016; Post et al., 2013; Nakayama
et al., 2010; Nakayama et al., 2007). Genualdi et al. (2017) investigated a cranberry bog in
Massachusetts approximately 10 miles from a military base with a history of AFFF usage. Bog
water samples were collected in November 2016 and PFHxS was detected in all four samples
(six total samples collected, two samples were lost due to evaporation) with a mean
concentration of 10.98 ng/L. Authors concluded that given the presence and ratio of PFHxS and
PFOS in the bog water samples, it was likely the surface water contamination was related to the
previous AFFF usage. Anderson et al. (2016) assessed 40 sites across 10 active Air Force
installations throughout the continental United States and Alaska between March and September
2014. Installations were included if there was known historic AFFF release in the period 1970-
1990. The selected sites were not related to former fire training areas and were characterized
according to volume of AFFF release—low (n = 2), medium (n = 32), and high (n = 2). PFHxS
D-13
-------
FINAL
2024
was detected at both sites characterized to have high-volume AFFF releases, with mean
concentration of 5,600 ng/L. Detection frequencies for sites with low- and medium-volume
AFFF releases were 50.0% and 74.3%, respectively, and mean concentrations were 7.1 and
196,800 ng/L, respectively. Across all sites, the median concentration of detects was 710 ng/L.
Post et al. (2013) evaluated raw surface water samples from 12 public drinking water system
intakes collected between August 2009 and February 2010. Six rivers and six reservoirs,
including two reservoirs in Atlantic County near a civil-military airport with possible AFFF use,
were selected to represent New Jersey geographically. PFHxS was below the minimum RL (5
ng/L) in all six river samples. In reservoir samples, PFHxS was detected at two of six sites (n =
16) at concentrations of 44 and 46 ng/L. These two sites corresponded to the sites near the civil-
military airport; given the presence of PFHxS and other PFCs, authors reported the
contamination to be indicative of AFFF usage. Two studies from Nakayama et al. (2010; 2007)
assessed surface water samples from the Upper Mississippi River, Missouri River, and Cape Fear
River Basins. In Nakayama et al. (2010), a large-scale evaluation of the Upper Mississippi River
Basin and portion of the Missouri River Basin was conducted to provide preliminary PFC data
given the importance of the two basins in supplying drinking water. Between the two basins, 173
samples were collected across 88 sampling sites in March-August 2008 by several different
agencies—Minnesota Pollution Agency, Wisconsin Department of Natural Resources, Illinois
Environmental Protection Agency, and U.S. EPA Region 7 Water Quality Monitoring Team.
Overall, the detection frequency of PFHxS was 89% with a median concentration of 0.71 ng/L.
Authors reported higher PFC concentrations adjacent to chemical manufacturers, downstream of
WWTPs receiving waste from those types of manufacturers, and near an airport with historic use
of firefighting foams. In Nakayama et al. (2007), authors evaluated the performance of a new
method for the collection and analysis of PFCs using samples collected from the Cape Fear River
Basin, North Carolina during spring 2006. Authors noted possible sources of PFCs to the basin
included firefighting foam from nearby air force bases and commercial/industrial facilities. One
hundred surface water samples were taken from 80 sites selected to reflect water quality
throughout the basin. PFHxS was detected in 98.7% of samples with mean and median
concentrations of 7.29 and 5.66 ng/L, respectively. The highest concentrations were found in the
middle reaches of the Cape Fear River and its two major tributaries.
Two studies examined surface water near or downstream of land application sites where PFC-
contaminated WWTP effluent or biosolids were applied (Lasier et al., 2011; Lindstrom et al.,
2011). In the first study, Lasier et al. (2011) sampled the Coosa River, Georgia during summer
2008; samples included two sites upstream (sites 1 and 2) and six sites downstream (sites 3-8) of
a land application site, where treated effluent from a WWTP was sprayed. The WWTP processed
effluents from multiple carpet manufacturers who were reported to use significant quantities of
PFCs. Additionally, site 2 was downstream of a local airport, and site 4 was downstream of a
manufacturing facility of latex and polyurethane back material—inputs for the carpet
manufacturers. PFHxS was below the MDL (0.014 ng/g) at both of the upstream sites and at the
most downstream sites (sites 7 and 8). Mean concentrations for sites 3-6 were 30, 31, 17, and 13
ng/L, respectively. Authors reported highest concentrations downstream of the land application
and backing-material sites and then decreased concentrations increasingly downstream as a result
of dilution. In the second study, Lindstrom et al. (2011) analyzed surface water samples from
ponds and streams in Decatur, Alabama. The samples were collected in February 2009 from
farms that had applied PFC-contaminated biosolids to local agricultural fields as a soil
amendment for at least 12 years. The biosolids were obtained from a local municipal WWTP
D-14
-------
FINAL
2024
where authors noted that sources discharging to the WWTP included facilities involved in the
production and use of fluoropolymers, fluorocarbon fibers, polymers, polymer films, and resins,
although specific sources could not be characterized. PFHxS was detected in 22% of samples (n
= 32), with levels ranging from below the LOQ (10 ng/L) to 218 ng/L.
Three studies evaluated surface water potentially impacted by wastewater (Boone et al., 2019;
Subedi et al., 2015; Appleman et al., 2014). Boone et al. (2019) evaluated 17 PFAS in source and
treated waters collected in 2010-2012. Authors attempted to select locations with known or
suspected sources of wastewater in the source water, but ultimately the site selection was
dependent upon the willingness of DWTPs to participate. The study did not differentiate which
locations had known or suspected sources. Of the 22 surface water sources evaluated (16 river
and 6 lake/reservoir), PFHxS was detected in 95% of samples (n = 22), ranging from not
detected (LCMRL = 0.034 ng/L) to 19.7 ng/L. Subedi et al. (2015) collected 28 lake water
samples from 3 sampling events in August-September 2012 and four sampling events in May-
September 2013 from Skaneateles Lake. Sites were selected to be along the shoreline of homes
that use an enhanced treatment unit for onsite wastewater treatment. Wastewater effluents were
identified as a source of contamination to the lake. PFHxS was detected in 79% of samples with
mean and median concentrations of 0.56 and 0.28 ng/L, respectively. Appleman et al. (2014)
assessed source water from 11 utilities in Alaska, Alabama, Colorado, Nevada, New Jersey,
Ohio, Oklahoma, and Wisconsin from August 2011 to May 2012, the majority of which were
selected because they were either known from previous monitoring or expected to contain
detectable PFAS because they were impacted by upstream wastewater effluent discharge.
Authors evaluated the utilities and their effectiveness for removing PFAS. The study did not
report an average concentration for PFHxS, but PFHxS was detected in 18 of 25 samples (from 8
of 11 utilities) with a maximum concentration of 13 ng/L.
In three studies, surface water samples were collected from locations with potential sources of
PFAS that were not related to AFFF use (Procopio et al., 2017; Zhang et al., 2016; Sinclair et al.,
2006). Procopio et al. (2017) evaluated samples collected between September 2011 and July
2014 from the Metedeconk River. Eight sampling events were conducted as part of a source
trackdown study to identify potential sources of PFAS contamination after elevated PFOA levels
were discovered at a raw surface water intake of the Brick Township Municipal Utilities
Authority. In all 56 samples, PFHxS was below the minimum laboratory RL of 5 ng/L. Zhang et
al. (2016) conducted analyses to determine major sources of surface water PFAS contamination.
Freshwater sample collection sites included 22 sites in the state of Rhode Island (sampled June
2014) and 6 sites in the New York Metropolitan Area (sampled October 2014). Surface water
sites were creeks and rivers in urban and rural locations. PFHxS was detected in 89% of samples
(n = 28) ranging from below the limit of detection (LOD) to 35.022 ng/L. Authors identified
potential PFAS sources at these sites to be metal coating plating; paint, coating, adhesive
manufacturing; paper manufacturing; petroleum coal products manufacturing; printing activity;
printing ink manufacturing; semiconductor manufacturing; sewage treatment; textile mills; waste
management including landfills, and airports. PFHxS levels were below the LOD (0.06 ng/L) at
three rural sites corresponding to a background site with no recorded upstream industrial
facilities; a Pawcatuck River site 1 km upstream of a military, tactical, and performance synthetic
and synthetic blend textiles manufacturer; and a Secret Lake-Oak Hill Brook site 2 km east of a
legacy landfill site. Authors reported significantly higher concentrations in urban regions, with
the highest being possibly attributed mainly to T.F. Green Airport near Mill Cove, Rhode Island.
D-15
-------
FINAL
2024
In July 2004, Sinclair et al. (2006) collected 51 samples from nine major water bodies of New
York to assess the distribution of PFAS. Water bodies included Lake Ontario, Niagara River,
Lake Erie, Finger Lakes, Lake Onondaga, Lake Oneida, Erie Canal, Hudson River, and Lake
Champlain. PFHxS was detected in 50 samples with median concentrations across all lakes
ranged from 0.9 to 7.4 ng/L. The highest concentrations were detected at Lake Onondaga and
Erie Canal with median concentrations of 7.4 and 2.6 ng/L, respectively. Authors noted that Lake
Onondaga is a Superfund site, is influenced by several industrial sources located along the lake,
and also receives effluent from the Metropolitan Syracuse sewage treatment plant. Based on the
results of other PFAS detected, the authors also suggested that there may be greater industrial use
of fluoropolymers and telomer-alcohol in the region, including the Erie Canal.
Of the remaining studies (Bradley et al., 2020; Pan et al., 2018; Boone et al., 2014; Quinones and
Snyder, 2009; Kim and Kannan, 2007), Bradley et al. (2020) analyzed samples of Lake Michigan
untreated intake water as part of a study that also analyzed home tap water samples. Samples
were collected in 2017 at four intake sites. PFHxS was detected in all seven samples with a mean
concentration of 1.0 ng/L. Boone et al. (2014) developed and tested the accuracy and precision of
an analytical method to determine PFCs in environmental and drinking waters. The authors
analyzed PFCs in water samples collected from the surface of the Mississippi River at a low flow
level (2.95 ft) in September 2010 and a high flow level (8.32 ft) in June 2009. Results were
presented as means based on an average of primary and duplicate samples or an average of four
replicates. PFHxS levels were 1.315 and 1.07 ng/L at low and high flow levels, respectively. In
Quinones and Snyder (2009), surface water samples from the Boulder Basin, Hoover Dam, and
the lower Colorado River were collected in 2008, as part of a study to assess both raw and
treated water from utilities producing at least 75 megaliters of finished water per day. PFC
occurrence had not been previously determined or reported for these sites. Mean PFHxS levels at
all sites were below the method RL (1.0 ng/L). Kim and Kannan (2007) sampled two urban lakes
in Albany, New York during five sampling trips from February-November 2006. The lakes,
Washington Park Lake are Rensselaer Lake, are located in downtown Albany and receive surface
runoff from nearby roadways and residential areas during stormwater runoff. PFHxS was
detected in Washington Park Lake (n = 6) at a mean and median concentration of 0.33 ng/L.
PFHxS was detected in Rensselaer Lake (n = 5) at mean and median concentrations of 3.09 and
3.25 ng/L, respectively. Overall, PFHxS was detected in 81.8% of the 11 total samples. Finally,
in a multicontinental study, Pan et al. (2018) assessed surface water samples from several
countries including the United States (Delaware River), United Kingdom (Thames River),
Germany and the Netherlands (Rhine River), and Sweden (Malaren Lake). Twelve samples were
collected in September-December 2016 along the Delaware River that spanned seven cities—
Trenton, Bristol, Philadelphia, Chester, Delaware, Smyrna, and Frederica. Authors noted that all
sampling sites were along the main stream of the studied rivers and not proximate to known
point sources of any fluorochemical facilities. PFHxS was detected in all samples from the
Delaware River with a mean concentration of 1.68 ng/L and were similar to levels found in the
Rhine River and Malaren Lake.
Detailed results of the occurrence of PFHxS in European surface waters are presented in
Table D-2. Nine studies conducted in Europe evaluated sites with no known point
fluorochemical source (Barreca et al., 2020; Munoz et al., 2018; Pan et al., 2018; Shafique et al.,
2017; Eriksson et al., 2013; Wagner et al., 2013; Ahrens et al., 2009b; Ahrens et al., 2009a;
Ericson et al., 2008b). Pan et al. (2018) performed a study that included surface water sampling
D-16
-------
FINAL
2024
sites in the United Kingdom (Thames River), Germany and the Netherlands (Rhine River), and
Sweden (Malaren Lake). None of the sites sampled were proximate to known sources of PFAS,
but for all three water bodies, detection frequency for PFHxS was 100%. The highest PFHxS
levels were detected in the Thames River (maximum = 11.3 ng/L), which was about 3 to 4 times
greater than the maximum levels found in the other water bodies. For the remaining nine studies,
most reported PFHxS levels were relatively low (<7.8 ng/L) or were not detected at all. Eight
studies in Europe evaluated urban areas possibly affected by industrial activities (Lorenzo et al.,
2015; Zhao et al., 2015; Boiteux et al., 2012; Eschauzier et al., 2012; Kovarova et al., 2012;
Rostkowski et al., 2009) or wastewater effluent discharges (Lorenzo et al., 2015; Labadie and
Chevreuil, 2011; Moller et al., 2010). PFHxS occurrence in these studies varied, with some
studies reported 0% detections and some reporting detectable levels in all samples. The
remaining studies conducted in Europe evaluated surface water samples from sites with known
or suspected PFAS releases associated with fluorochemical manufacturing (Bach et al., 2017;
Boiteux et al., 2017; Gebbink et al., 2017; Valsecchi et al., 2015) or AFFF use (Mussabek et al.,
2019; Gobelius et al., 2018; Dauchy et al., 2017; Filipovic et al., 2015). Of the four studies
potentially impacted by nearby fluorochemical manufacturing sites, two conducted in France
found no PFHxS (Bach et al., 2017; Boiteux et al., 2017). PFHxS levels were not detected or
relatively low at most sampling locations of the remaining two studies (<3 ng/L) except for River
Brenta in Italy with an approximately 10-fold higher maximum level (35.6 ng/L). This site is
also impacted by nearby textile and tannery manufacturers (Valsecchi et al., 2015). Consistent
with U.S.-based studies, the highest PFHxS levels were found at the AFFF-impacted sites (up to
7,550 ng/L).
D-17
-------
FINAL
2024
Table D-2. Studies Reporting PFHxS Occurrence in Surface Water
Study
Location
Site Details
Results
United States
Galloway et al. (2020)
United States (Ohio; West
Virginia)
Rivers and tributaries near a fluoropolymer facility
sampled throughout three trips on June, July, and
December 2016. In June 2016, samples were collected
on a 188 km stretch of the Ohio River, from 130 km
downstream to 58 km upstream of the facility, and
tributaries that pass near known PFAS-containing
landfills. In July 2016, samples were collected from
lakes, rivers, and creeks to the north and northeast of
the facility as far as 16 km downwind. The December
2016 trip expanded the collection radius to more than
48 km downwind to the north and northeast of the
facility.
June 2016:
n = 26, DFa = 96%*
*PFHxS was detected but below the LOQ in
25 samples, and ND in 1 sample
July 2016:
n = 25, DFa = 12%*, range = ND-79.0 ng/L
December 2016:
n = 40; DFa = 48%)*, range3 =
-------
FINAL
2024
Study
Location
Site Details
Results
1970 and 1990 that were not related to former fire
training areas. It is assumed that the measured PFAS
profiles at these sites reflect the net effect of several
decades of all applicable environmental processes.
AFFF-impacted sites included emergency response
locations, hangers and buildings, and testing and
maintenance related to regular maintenance and
equipment performance testing of emergency vehicles
and performance testing of AFFF solution. Previous
remedial activities for co-occurring contaminants were
not specifically controlled for in the site selection
process; active remedies had not been applied at any
of the sites selected. Approximately ten samples were
collected between March and September 2014 at each
site; sites were grouped according to volume of AFFF
release—low-volume typically had a single AFFF
release, medium-volume had one to five releases, and
high-volume had multiple releases. Surface water
sample locations included engineered storm water
channels, engineered AFFF ponds, and natural
streams.
Breakdown by site:
Emergency Response (low-volume release):
n = 2, DF 50.0%, mean (range) = 7.1 (7.1-
7.1) ng/L
Hangars and Buildings (medium-volume
release):
n = 32, DF 74.3%, 196,800 (360-2,700,000)
ng/L
Testing and Maintenance (high-volume
release):
n = 2, DF 100.0%), mean (range) = 5,600
(4,400-6,700) ng/L
(Median RL = 7 ng/L)
*Median calculated using quantified
detections
*Non-detects were substituted with Vi the
reporting limit
Post et al. (2013)
United States (New Jersey)
Raw water collected from 12 public drinking water
system intakes between August 2009 and February
2010 from 6 rivers and 6 reservoirs. Sites were chosen
to represent NJ geographically and included two
reservoir sites near a civil-military airport with
possible AFFF use.
Overall n = 12, DF 17%, range = ND^6 ng/L
Rivers: n = 6, DF 0%
Reservoirs: n = 6, DF 33%, range = <5^16
ng/L
(RL = 5 ng/L)
Nakayama et al. (2010)
United States (Illinois; Iowa;
Minnesota; Missouri;
Wisconsin)
Eighty-eight sampling sites collected between March
and August 2008 from tributaries and streams in the
Upper Mississippi River Basin and a portion of the
Missouri River Basin. Samples were collected by the
Minnesota Pollution Agency, Wisconsin Department
of Natural Resources, Illinois Environmental
Protection Agency, and U.S. EPA Region 7 Water
Quality Monitoring Team. Each agency selected
sampling sites with the intention of providing
preliminary PFC data to the individual regions.
Sampling sites included locations adjacent to chemical
manufacturers, downstream of WWTPs receiving
waste from those types of manufacturers, and near an
airport with historic use of firefighting foams.
n = 173, DF 89%o, median (range) = 0.71
(ND-169) ng/L
(LOD = 0.02 ng/L)
*ND data points were substituted with
LOD/sqrt(2) = 0.014 ng/L
*The maximum concentration of 169 ng/L
was collected from a waterway that potentially
receives a run-off from a historical fire
training site at the Duluth International
Airport. Excluding this location, the maximum
PFHxS level in surface water is 14.5 ng/L
D-19
-------
FINAL
2024
Study
Location
Site Details
Results
Nakayama et al. (2007)
United States (North Carolina)
Eighty sampling sites in river basin during spring
2006. The sites were selected to reflect water quality
throughout the basin. Possible sources of PFCs include
use of firefighting foam from Fort Bragg and Pope Air
Force, metal-plating facilities, textile, and paper
production, and other industries.
n = 100, DF 98.7%, mean, GM, median
(range) = 7.29, 5.73, 5.66 (
-------
FINAL
2024
Study
Location
Site Details
Results
Subedi et al. (2015)
United States (New York)
Lake water along the shoreline of residences that use
an enhanced treatment unit for onsite wastewater
treatment; samples were collected -40 ft from the
lakeshore about 2 ft below surface. Sampling occurred
August-September 2012 (three sampling events) and
May-September 2013 (four sampling events).
Wastewater effluents identified as source of
contamination.
n = 28, DFa 79%, mean, median, (maximum)
= 0.56, 0.28 (2.57) ng/L
(LOQ = 0.2 ng/L)
*Data points
-------
FINAL
2024
Study
Location
Site Details
Results
Sinclair et al. (2006)
United States (New York)
Nine major water bodies—Lake Ontario, Niagara
River, Lake Erie, Finger Lakes, Lake Onondaga, Lake
Oneida, Erie Canal, Hudson River, and Lake
Champlain—were sampled in July 2004 to represent
the major water bodies of New York State. Lake
Onondaga is a Superfund site, is influenced by
industrial sources along the lake, and receives WWTP
effluent.
Lake Ontario: n = 13, DF 100%, median
(range) = 1.4 (1.2-2.8) ng/L
Niagara River: n = 3, DF 100%, median
(range) = 1.2 (1.2-1.4) ng/L
Lake Erie: n = 3, DF 100%, median (range) =
1.2 (1.2-1.6) ng/L
Finger Lakes: n = 13, DF 100%, median
(range) = 0.9 (0.7-1.3) ng/L
Lake Onondaga: n = 3, DF 100%), median
(range) = 7.4 (4.2-8.5) ng/L
Lake Oneida: n = 1, point =0.9 ng/L
Erie Canal: n = 3, DF 100%, median (range) =
2.6 (2.5-5.6) ng/L
Hudson River: n = 8, DF 100%), median
(range) = 0.9 (0.7-1.6) ng/L
Lake Champlain: n = 4, DFa 75%o, median
(range) =1.3 (ND-1.6) ng/L
(DL = 0.5 ng/L)
Bradley et al. (2020)
United States (Chicago, Illinois;
East Chicago, Indiana)
Lake Michigan untreated intake water at four intake
sites. Samples collected in July and November 2017 at
the intakes of the Chicago North and Chicago South
WTPs and in July and November 2017 at the intakes
of the two East Chicago DwTPs.
n = 7, DFa 100%o, meana (range) = 1.0 (0.8-
1.3) ng/L
(LOQ = 0.120-0.580 ng/L)
*Quantitative (>LOQ) and semiquantitative
(between LOQ and MDL) results treated as
detections
Boone et al. (2014)
United States (New Orleans,
Louisiana)
Surface samples from the Mississippi River collected
in June 2009 when the river was at a high flow level
(8.32 ft) and in September 2010 when the river was at
a low flow level (2.95 ft).
Low flow (2.95 ft): mean based on a primary
and duplicate sample = 1.315 ng/L
High flow (8.32 ft): mean based on four
replicates =1.07 ng/L
(DL = 0.016 ng/L; LCMRL = 0.034 ng/L)
Quinones and Snyder (2009)
United States (Arizona; Nevada)
Samples collected in 2008 from three sites in Boulder
Basin, one site in Hoover Dam, and two sites from the
lower Colorado River. PFC occurrence had not been
previously determined or reported for these sites.
n = 40, DF* 0%
(Method RL= 1.0 ng/L)
*Mean values at all sites were
-------
FINAL
2024
Study
Location
Site Details
Results
trips from February-November 2006. Both lakes are
located in downtown Albany and receive surface
runoff from nearby roadways and residential areas
during stormwater runoff.
Washington Park Lake: n = 6, DF NR, mean,
median (range) = 0.33, 0.33 (
-------
FINAL
2024
Study
Location
Site Details
Results
Europe
Bachetal. (2017)
France (southern)
Grab water samples were collected from six locations
along the shore of a river in April, July, October, and
December 2013. The river selected for the study
receives effluent at three points along the river from
an industrial site where two facilities produce
fluoropolymers. The first facility has been active since
the 1960s, with production including PTFE synthesis
from the beginning of the 1960s to 1985 with PFOA
as a processing aid; more recently, PVDF has been
synthesized since the early 1970s with fluorotelomer
sulfonic acid (6:2 FTSA) or PFNA as a processing aid.
The second facility, established in 2002, produced
fluoropolymers with PFOA as a processing aid until
2008 when it was replaced with PFHxA. Samples
were collected starting ~1.3 km upstream from the
industrial site and covered ~15 km of the river.
Samples (point #1 to #6) were collected from
upstream to downstream.
Upstream:
Sampling point #1: n = 1, DF 0% for April,
July, October, and December 2013
Downstream:
Sampling points #2-6: n = 1, DF 0% for
April, July, October, and December 2013
(LOQ = 4 ng/L)
*DF represents frequency of quantification
Boiteux et al. (2017)
France (northern)
Grab water samples were collected from seven
locations along a river in May, July, October, and
December 2013. The river selected for the study
receives wastewater from an industrial WWTP that
treats raw sewage coming from a facility that
manufactures fluorotelomer-based products and side-
chain-fluorinated polymers used in firefighting foams
and stain repellents. Samples were collected starting
~1.2 km upstream of the WWTP discharge and
encompassed ~65 km of the river. Samples were
collected from upstream to downstream.
Upstream:
Sampling point #1: n = 1, DF 0% for May,
July, October, and December 2013
Downstream:
Sampling points #3,4, 5,7, 9,11: n= 1,DF
0% for May, July, October, and
December 2013
(LOQ = 4 ng/L)
*DF represents frequency of quantification
Gebbink et al. (2017)
The Netherlands (Dordecht)
River water samples collected in October 2016 at sites
downstream (R1-R13) and upstream (R14-R16) of
the Dordrecht fluorochemical production plant.
Samples (R17-R18) were also collected from
different waterbodies at control sites.
Control sites: n = 2, DFa (frequency of
quantification) 100%, meana (range) = 1.85
(1.7-2.0) ng/L
Upstream: n = 3, DFa (frequency of
quantification) 100%, meana (range) = 2.1
(2.0-2.2) ng/L
Downstream: n = 13, DFa (frequency of
quantification) 100%, meana (range) = 2.0
(1.5-2.2) ng/L
(MQL = 0.02 ng/L)
D-24
-------
FINAL
2024
Study
Location
Site Details
Results
Valsecchi et al. (2015)
Italy (River Basins Po, Brenta,
Adige, Tevere, and Arno)
Five river basins were sampled between 2008 and
2013. Two river basins (Po and Brenta) receive
discharges from two chemical plants that produce
fluorinated polymers and intermediates; two river
basins (Tevere and Adige) are not impacted by
relevant industrial activities; and one river basin
(Arno) has textile and tannery districts located along
parts of the river. In total, 20 rivers were sampled at
their basin closure stations. Rivers Arno, Tevere, and
Po were also sampled along the course of the river.
Po: n = 105, DF 0%
Brenta: n = 5, DFa 20%, range =
-------
FINAL
2024
Study
Location
Site Details
Results
source. Sample numbers varied for each county and
sampling sites were spread unevenly across Sweden.
Surface water samples collected approximately 10 cm
below the water surface.
Dauchy et al. (2017)
France (unspecified)
Samples collected in the vicinity of three sites (B, C,
D) where fluorosurfactant-based foams are or were
being heavily used. Site B is an international civilian
airport built in 1974. The exact location of the training
area, frequency of training sessions, and history of
firefighting training activities are unknown. In
November 2014, surface water samples were collected
in the only river running alongside the airport.
Downstream from the airport, this river joins two
other rivers, which were also sampled.
Site C is a military airport, with the exact location of
the training area, frequency of the training sessions,
and history of the firefighting training activities
unknown. In April 2014, surface water samples were
collected in several rivers surrounding the military
base.
Site D is a training center for firefighters. From 1969
to 1984, the site was an oil refinery. Starting in 1987,
the site became a training area for firefighters, with
exercises carried out directly on the soil. From the
1990s, some exercise areas were covered with
concrete. In November 2014, two surface water
samples were collected from the river receiving
effluent from a WWTP at the site, one upstream and
one downstream of the discharge pipe.
Site B: n = 5, DFa 100%, meana (range) = 11.2
(5-18) ng/L
Site C: n = 9, DF 0%
Site D: n = 2, DFa 50%, range = <4-251 ng/L
(LOQ = 4 ng/L)
Filipovic et al. (2015)
Sweden (Stockholm)
Fourteen lakes and ponds surrounding the closed air
force base F18 in Tullinge Riksten, 19 km south of
Stockholm city center, where AFFFs were used.
Samples collected in two sampling campaigns in
December 2011 and April 2012. The air force base
was formally demobilized in 1986 but continued to be
used as an air force school for combat command and
air surveillance until 1994. Of note, the air force base
encountered numerous accidents and incidents during
the transfer from propeller era to the jet engine era,
including planes crashing upon takeoff and landing,
fire incidents, accidental dispersion of jet engine
n = 14, DF 64%), range = <0.5-25.1 ng/L
(MDL = <0.5 ng/L)
D-26
-------
FINAL
2024
Study
Location
Site Details
Results
starting fuel. The area was sold to a land developer in
1996 and is in the process of being transformed into a
municipal Area.
Ciofietal. (2018)
Italy (Tuscany)
Surface water samples were collected at 13 locations.
Sampling year was not reported.
SW-1: Arno river before the "Canale Maestro della
Chiana" (Arezzo), receiving agricultural runoff and
untreated urban wastewater
SW-2: Arno river after the "Canale Maestro della
Chiana" (Arezzo)
SW-3: Arno river before entering in Florence
SW-4: Arno river after the discharge of the Florence
WWTP
SW-5: Arno river after the confluence of the Bisenzio
river
SW-6: Arno river after the city of Empoli (Florence)
SW-7: Arno river after receiving the WWTP effluent
from the leather industrial district of Santa Croce
(Pisa)
SW-8: Arno river in the proximity of the mouth (Pisa)
SW-9: Bisenzio river before the confluence with Arno
river (Florence)
SW-10: Serchio river in the proximity of the mouth
(Lucca)
SW-11: East area of the coastal lake "Massaciuccoli"
(Lucca)
SW-12: West area of the coastal lake "Massaciuccoli"
(Lucca)
SW-13: Central area of the artificial lake "Bilancino"
(Florence)
SW-1: n = 1, point = <0.010 ng/L
SW-2: n = 1, point = <0.010 ng/L
SW-3: n = 1, point = <0.011 ng/L
SW-4: n = 1, point = <0.010 ng/L
SW-5: n = 1, point = <0.010 ng/L
SW-6: n = 1, point = <0.026 ng/L
SW-7: n = 1, point = <0.010 ng/L
SW-8: n = 1, point = <0.010 ng/L
SW-9: n = 1, point = <0.010 ng/L
SW-10: n = 1, point = <0.010 ng/L
SW-11: n = 1, point = <0.010 ng/L
SW-12: n = 1, point = <0.026 ng/L
SW-13: n= 1, point = <0.013 ng/L
(MDL = 0.010-0.013 ng/L; MQL = 0.026
ng/L)
*MDL/MQL varied by sample. MDL
provided for 11 of 13 samples; MQL provided
for 2 of 13 samples
Munoz et al. (2018)
France (Marnay-sur-Sein;
Bougival; Triel-sur-Seine)
Surface water along the Seine River was collected
during four sampling campaigns between September
2011 and December 2012, each conducted in a
different season. For each campaign, two to four
samples were collected over a one-month period.
Three sampling sites were investigated: Marnay-sur-
Sein, located 200 km upstream from Paris, was
selected as a reference site, non-affected by the
Greater Paris region; Bougival, situated 40 km
downstream from Paris, was chosen to investigate the
impact of Greater Paris on PFAS levels; Triel-sur-
n = 36, DF 100%, range = 0.28-7.8 ng/L
(LOD = 0.02-0.3 ng/L for all PFAS)
*PFF[xS concentrations were not reported in
text or table by site but relative abundance by
site is available in Figure 1
D-27
-------
FINAL
2024
Study
Location
Site Details
Results
Seine, another 40 km further downstream, was
selected to assess the global influence of the Paris
urban area, including other inputs such as WWTps.
Lorenzo et al. (2015)
Spain (Guadalquivir River
Basin; Ebro River Basin)
Surface water was collected from the Guadalquivir
River and its main tributaries and from the Ebro River
and its main tributaries in October 2010. Guadalquivir
sampling locations included downstream of WWTPs,
near industrial areas, near a military camp, or through
major cities; Ebro sampling locations included nearby
ski resorts and downstream of WWTP and industrial
areas.
Guadalquivir: n = 24, DF 13%, mean (range)
= 4.1 (1.5-88.5) ng/L
Ebro: n = 24, DF 13%), mean (range) = 0.5
(1.1-5.8) ng/L
*Minimum reported is the lowest amount
quantified
*Mean was calculated with not detected
concentrations as zeros
(MQL = 0.004 ng/L)
Zhaoetal. (2015)
Germany (Elbe River)
Four sampling campaigns conducted in February,
April, August, and October 2011 to represent the four
seasons. Freshwater samples (sites E619 to E689 with
salinity <1 PSU) were collected at nine locations in
the river Elbe. Some sampling sites were near
Hamburg city and experienced occasional discharge of
wastewater from industrial plants.
February: n = 8, DFa 87.5%o, range = <0.08-
0.96 ng/L
April: n = 9, DFa 100%o, meana (range) = 0.46
(0.08-0.77) ng/L
August: n = 9, DFa 100%o, meana (range) =
0.60 (0.42-1.0) ng/L
October: n = 8, DFa 100%, meana (range) =
0.38 (0.18-0.52) ng/L
(MDL = 0.03 ng/L)
Boiteux et al. (2012)
France (national)
Raw water from rivers used as source water for
DWTPs. Sites distributed across 100 French
department to represent -20% of the national water
supply flow; samples collected during two sampling
campaigns in July-September 2009 (first campaign)
and June 2010 (second campaign - focused on sites
from first sampling campaign that had PFC levels
>LOQ). Some sites possibly affected by
commercial/industrial releases.
Overall: n = 135, DF (frequency of
quantification) 7%, maximum = 8 ng/L
1st Sampling Campaign
n = 99, DF 48%o, mean, median (maximum)
= 1, <1 (8) ng/L
2nd Sampling Campaign
n = 36, results not reported
(LOD =1.3 ng/L, LOQ = 4 ng/L)
Eschauzier et al. (2012)
The Netherlands (Amsterdam)
Intake water from the Lek canal (n = 2) was collected
in January and September 2010 to determine the
behavior of PFAAs during the drinking water
treatment processes. The Lek canal, a tributary of the
river Rhine, is the source of drinking water for the city
of Amsterdam and is downstream of an industrial
point source in the German part of the Lower Rhine.
n = 2, DFa 100%o, mean (range) = 2.0(1.9-2.2)
ng/L
(LOQ = 0.55 ng/L)
D-28
-------
FINAL
2024
Study
Location
Site Details
Results
Kovarova et al. (2012)
Czech Republic (Brno)
Seven locations in the Svitava and Svratka Rivers
upstream and downstream of Brno, a city with highly
developed chemical, engineering, textile, and food-
processing industries. A sampler was installed at each
site for 30 days twice a year (May and September
2008). Due to technical problems, samples were
produced from only four of seven sites in May and
from five of seven sites in September.
n = 9, DF 0%
(LOD not reported)
Llorca et al. (2012)
Germany (Hesse), Spain
(national)
Forty-eight surface river waters were sampled in
2010-2011 (24 from Spain and 24 from Germany).
Samples from Germany were collected from
agriculturally or industrially influenced streams.
Samples from Spain were collected from the Xuquer
River Basin, Llobregat River Basin, and Ebro River
Basin.
Germany: n = 24, DF 21%, mean, median
(range) = 1.9, 0.5 (0.06-5.6) ng/L
Spain: n = 24, DF 21%, mean, median (range)
= 16, 5.8 (0.06-37) ng/L
(MLOQ = 0.06 ng/L)
*MLOQ reported above is from Table 3;
Table 2 reports MLOD = 0.27 ng/L and
MLOQ = 0.90 ng/L
Labadie and Chevreuil (2011)
France (Paris)
Samples collected weekly in January-May 2010 in an
urban stretch of the River Seine at the Austerlitz Quay,
downtown Paris during a flood cycle. The sampling
station is under the influence of two major WWTPs
and two major combined sewer overflow outfalls.
n = 16, DF 100%), mean, median (range) = 7.1,
6.8 (3.9-12.0) ng/L
(LOQ = 0.15 ng/L)
Moller et al. (2010)
Germany (Rhine River
watershed)
Raw freshwater samples collected in September-
October 2008 along the River Rhine (stations 1-36)
and major tributaries of the River Rhine (e.g., Rivers
Neckar, Main, Rhur, stations 37^18). Along the River
Rhine, samples were taken upstream and downstream
of Leverkusen, where effluent of a WWTP treating
industrial wastewater was discharged. All samples
taken at a water depth <1 m.
Rhine upstream Leverkusen: n = 27, DF NR,
mean (range) = 3.04 (<0.51-14.5) ng/L
Rhine downstream Leverkusen: n = 9, DF
100%o, mean (range) = 1.93 (1.66-2.44)
ng/L
River Ruhr: n = 3, DF NR, mean (range) =
0.18 (<0.51-0.53)
River Moehne: n = 1, point = 1.03 ng/L
Other tributaries: n = 8, DF NR, mean (range)
= 1.41 (<0.51-2.93) ng/L
(MDL = 0.51 ng/L)
Rostkowski et al. (2009)
Poland (national)
Inland surface water samples were collected at 12
locations in the southern part of Poland and 14
locations in the northern part of Poland in October and
December 2004. Inland surface waters included rivers,
lakes, and streams. The northern locations flowed
through forested, agricultural, and rural areas; these
North: n = 14, DFa 79%o, range = <0.02-
2.8 ng/L
South: n = 11, DFa 91%o, range = <0.67-
113 ng/L
(LOQ = 0.1 ng/L)
D-29
-------
FINAL
2024
Study
Location
Site Details
Results
areas are considered unpolluted with industrial
chemicals. Some southern locations were near
chemical industrial activities.
Barreca et al. (2020)
Italy (Lombardia Region)
Fifty-two surface water sampling stations (rivers and
streams) throughout the region. Samples collected in
2018.
n = 286, DFa 4%
(LOQ = 1 ng/L)
Shafique et al. (2017)
Germany (River Elster; River
PleiBe; River Saale; and River
Elbe)
Surface water samples were collected from the River
Elster, River PleiBe, River Saale, and River Elbe at
the start of 2015.
Elster: n = 4, DF NR, mean = 0.42 ng/L
PleiBe: n = 2, DF NR, mean = 0.37 ng/L
Saale (Site A): n = 10, DF NR, mean =0.13
ng/L
Saale (Site B): n = 10, DF NR, mean = 0.60
ng/L
Saale (Site C): n = 10, DF NR, mean = 0.07
ng/L
Elbe: n = 2, DF NR, mean = 0.61 ng/L
(MDL = 0.11 ng/L)
*Values extracted from SI, which provides a
more detailed breakdown of sites compared to
that reported in the main text (where Elster
and PleiBe sites were combined and Saale
sites were combined)
Eriksson et al. (2013)
Denmark (Faroe Islands)
Grab samples collected in April-May 2012 from
Lakes Leitisvatn, Havnardal, Kornvatn, and A
Myranar.
Leitisvatn: n = 1, point = <0.058 ng/L
Havnardal Lake: n = 1, point = <0.024 ng/L
Kornvatn Lake: n = 1, point = <0.027 ng/L
A Myranar: n = 1, point = <0.024 ng/L
(LOD = 0.035 ng/L)
Wagner et al. (2013)
Germany (Rhine River)
Surface water samples were collected from the Rhine
River. Sampling year not provided. Samples used to
test out a new analytical protocol to determine trace
levels of adsorbable organic fluorine.
n = 2, DFa 100%, mean (range) = 2.4 (1.6-3.2)
ng/L
(LOD = 50 ng/L F"; LOQ = 150 ng/L F")
*PFHxS concentrations were calculated using
the fluorine concentrations reported in Table 4
Ahrens et al. (2009a)
Germany (Hamburg;
Laurenburg)
Nine samples collected from the river Elbe in
Hamburg city (sites 16-18) and from Laurenburg to
Hamburg (sites 19-24) in August 2006. Samples were
collected at a water depth of 1 m. Dissolved and
Hamburg:
Dissolved: n = 3, DFa 100%, mean (range) =
0.60 (0.56-0.67) ng/L
Particulate: n = 3, DF 0%
D-30
-------
FINAL
2024
Study
Location
Site Details
Results
particulate phases were analyzed for each of the water
samples.
Laurenburg to Elamburg:
Dissolved: n = 6, DFa 100%, mean (range) =
0.36 (0.24-0.49) ng/L
Particulate: n = 6, DFa 33%, mean (range) =
0.029 (ND-0.098) ng/L
(MDL = 0.140 ng/L for dissolved phase; 0.045
ng/L for particulate phase)
Ahrens et al. (2009b)
Germany (Elbe River)
Samples collected at 53 to 122 km (sites 1 to 9)
upstream of estuary mouth of Elbe River in June 2007.
*Only locations with conductivity <1.5 mS/cm were
assumed to be freshwater and extracted
Site 1 (122 km): n = NR, DF NR, mean =
0.85 ng/L
Site 2(118 km): n = NR, DF NR, mean =
0.9 ng/L
Site 3(115 km): n = NR, DF NR, mean =
1.0 ng/L
Site 4 (103 km): n = NR, DF NR, mean =
1.2 ng/L
Site 5 (90 km): n = NR, DF NR, mean =
0.8 ng/L
Site 6 (80 km): n = NR, DF NR, mean =
1.3 ng/L
Site 7 (74 km): n = NR, DF NR, mean =
0.9 ng/L
Site 8 (64 km): n = NR, DF NR, mean =
1.1 ng/L
Site 9 (53 km): n = NR, DF NR, mean =
0.9 ng/L
(MDL = 0.17 ng/L; MQL = 0.57 ng/g)
Ericson et al. (2008b)
Spain (Tarragona Province)
River water samples collected from the Ebro (at two
points, Garcia and Mora), Francoli, and Cortiella
Rivers in February 2007.
Ebro site 1: n = 1, point = 0.40 ng/L
Ebro site 2: n = 1, point = 0.43 ng/L
Francoli: n = 1, point = 0.78 ng/L
Cortiella: n = 1, point = <0.18 ng/L
(LOD = 0.18 ng/L)
Multiple Continents
Pan et al. (2018)
United States (Delaware River)
Samples were collected from the Delaware River
between September and December 2016. Sampling
sites were not proximate to known point sources of
any fluorochemical facilities. Cities included Trenton,
n = 12, DFa 100%), mean, median (range) =
1.68,1.72 (0.65-2.63) ng/L
(MDL = 0.05 ng/L)
D-31
-------
FINAL
2024
Study
Location
Site Details
Results
Bristol, Philadelphia, Chester, Delaware, Smyrna, and
Frederica.
United Kingdom (Thames
River)
Samples were collected from the Thames River in
October 2016. Sampling sites were not proximate to
known point sources of any fluorochemical facilities.
Cities included Oxford and London.
n = 6, DFa 100%, mean, median (range) =
7.14, 6.42 (4.96-11.3) ng/L
(MDL = 0.05 ng/L)
Germany and The Netherlands
(Rhine River)
Samples were collected from the Rhine River in
December 2016. Sampling sites were not proximate to
known point sources of any fluorochemical facilities.
Cities in Germany included Offenbach, Frankfurt,
Goarshausen, Rheinbrohl, Bonn, Cologne,
Leverkusen, Dormagen, Dusseldorf, Duisburg, Wesel,
and Emmerich. Cities in The Netherlands included
Arnhem, Lienden, Duurstede, Nijmegen, Wamel, and
Zaltbommel.
n = 20, DFa 100%, mean, median (range) =
1.98,2.03 (0.12-3.90) ng/L
(MDL = 0.05 ng/L)
Sweden (Malaren Lake)
Samples were collected from Malaren Lake in
September 2016. Sampling sites were not proximate to
known point sources of any fluorochemical facilities.
Cities included Orebro and Stockholm.
n = 10, DFa 100%), mean, median (range) =
1.30,0.97 (0.56-2.79) ng/L
(MDL = 0.05 ng/L)
D-32
-------
FINAL
2024
D.2. RSC for PFHxS, Literature Search and Screening
Methodology
The EPA applies an RSC to the RfD when calculating an MCLG based on noncancer effects or
for carcinogens that are known to act through a nonlinear mode of action to account for the
fraction of an individual's total exposure allocated to drinking water (USEPA, 2000b). The EPA
emphasizes that the purpose of the RSC is to ensure that the level of a chemical allowed by a
criterion (e.g., the MCLG for drinking water) or multiple criteria, when combined with other
identified sources of exposure (e.g., diet, ambient and indoor air) common to the population of
concern, will not result in exposures that exceed the RfD. In other words, the RSC is the portion
of total daily exposure equal to the RfD that is attributed to drinking water ingestion (directly or
indirectly in beverages like coffee tea or soup, as well as from transfer to dietary items prepared
with drinking water) relative to other exposure sources; the remainder of the exposure equal to
the RfD is allocated to other potential exposure sources. For example, if for a particular
chemical, drinking water were to represent half of total exposure and diet were to represent the
other half, then the drinking water contribution (or RSC) would be 50%. The EPA considers any
potentially significant exposure source when deriving the RSC.
The RSC is derived by applying the Exposure Decision Tree approach published in the EPA's
Methodology for Deriving Ambient Water Quality Criteria for the Protection of Human Health
(USEPA, 2000b). The Exposure Decision Tree approach allows flexibility in the RfD
apportionment among sources of exposure and considers several characteristics of the
contaminant of interest, including the adequacy of available exposure data, levels of the
contaminant in relevant sources or media of exposure, and regulatory agendas (i.e., whether there
are multiple health-based criteria or regulatory standards for the contaminant). The RSC is
developed to reflect the exposure to the U.S. general population or a sensitive population within
the U.S. general population and may be derived qualitatively or quantitatively, depending on the
available data.
A quantitative RSC determination first requires "data for the chemical in question...
representative of each source/medium of exposure and... relevant to the identified population(s)"
(USEPA, 2000b). The term "data" in this context is defined as ambient sampling measurements
in the media of exposure, not internal human biomonitoring metrics. More specifically, the data
must adequately characterize exposure distributions including the central tendency and high-end
exposure levels for each source and 95% confidence intervals for these terms (USEPA, 2000b).
Frequently, an adequate level of detail is not available to support a quantitative RSC derivation.
When adequate quantitative data are not available, the agency relies on the qualitative
alternatives of the Exposure Decision Tree approach. A qualitatively-derived RSC is an estimate
that incorporates data and policy considerations and thus, is sometimes referred to as a "default"
RSC (USEPA, 2000b). Both the quantitative and qualitative approaches recommend a "ceiling"
RSC of 80%) and a "floor" RSC of 20% to account for uncertainties including unknown sources
of exposure, changes to exposure characteristics over time, and data inadequacies (USEPA,
2000b).
In cases in which there is a lack of sufficient data describing environmental monitoring results
and/or exposure intake, the Exposure Decision Tree approach results in a recommended RSC of
D-33
-------
FINAL
2024
20%. In the case of MCLG development, this means that 20% of the exposure equal to the RfD
is allocated to drinking water and the remaining 80% is reserved for other potential sources, such
as diet, air, consumer products, etc. This 20% RSC value can be replaced if sufficient data are
available to develop a scientifically defensible alternative value. If scientific data demonstrating
that sources and routes of exposure other than drinking water are not anticipated for a specific
pollutant, the RSC can be raised as high as 80% based on the available data, allowing the
remaining 20% for other potential sources (USEPA, 2000b). Applying a lower RSC (e.g., 20%)
is a more conservative approach to public health and results in a lower MCLG.
D.2.1. Literature Search and Screening
In 2020, the EPA conducted a literature search to evaluate evidence for pathways of human
exposure to eight PFAS chemicals (PFOA, PFOS, PFBA, PFBS, PFDA, PFHxA, PFHxS, and
PFNA) (Holder et al., 2023). This search was not date limited and spanned the information
collected across the WOS, PubMed, and ToxNet/ToxLine (now ProQuest) databases. The results
of the PFHxS literature search of publicly available sources are available through the EPA's
Health & Environmental Resource Online website at
https://hero.epa.eov/hero/index.cfm/proiect/paee/proiect id/2630.
The 950 literature search results for PFHxS were imported into SWIFT-Review (Sciome, LLC,
Research Triangle Park, NC) and filtered through the Evidence Stream tags to identify human
studies and nonhuman (i.e., those not identified as human) studies (Holder et al., 2023). Studies
identified as human studies were further categorized into seven major PFAS pathways (Cleaning
Products, Clothing, Environmental Media, Food Packaging, Home Products/Articles/Materials,
Personal Care Products, and Specialty Products) as well as an additional category for Human
Exposure Measures. Nonhuman studies were grouped into the same seven major PFAS pathway
categories, except that the Environmental Media category did not include soil, wastewater, or
landfill. Only studies published between 2003 and 2020 were considered. Application of the
SWIFT-Review tags identified 654 peer-reviewed papers matching these criteria for PFHxS.
Holder et al. (2023) screened the 654 papers to identify studies reporting measured occurrence of
PFHxS in human matrices and media commonly related to human exposure (human
blood/serum/urine, drinking water, food, food contact materials, consumer products, indoor dust,
indoor and ambient air, and soil). For this synthesis, additional screening was conducted to
identify studies relevant to surface water (freshwater only) and groundwater using a keyword10
search for water terms.
Following the PECO criteria outlined in Table D-3, the title and abstract of each study were
independently screened for relevance by two screeners using litstream™. A study was included
as relevant if it was unclear from the title and abstract whether it met the inclusion criteria. When
two screeners did not agree whether a study should be included or excluded, a third reviewer was
consulted to make a final decision. The title and abstract screening of Holder et al. (2023) and of
this synthesis resulted in 494 unique studies being tagged as relevant (i.e., having data on
occurrence of PFHxS in exposure media of interest) that were further screened with full-text
10 Keyword list: water, aquifer, direct water, freshwater, fresh water, groundwater, ground water, indirect water,
lake, meltwater, melt water, natural water, overland flow, recreation water, recreational water, river, riverine water,
riverwater, river water, springwater, spring water, stream, surface water, total water, water supply
D-34
-------
FINAL
2024
review using the same inclusion criteria. After additional review of the evidence collected by
Holder et al. (2023), 109 studies originally identified for other PFAS also contained information
relevant to PFHxS. Based on full-text review, 172 studies were identified as having relevant,
extractable data for PFHxS from the United States, Canada, or Europe for environmental media,
not including studies with only human biomonitoring data. Of these 172 studies, 161 were
identified from Holder et al. (2023), where primary data were extracted into a comprehensive
evidence database. Parameters of interest included: sampling dates and locations, numbers of
collection sites and participants, analytical methods, limits of detection and detection
frequencies, and occurrence statistics. Eleven of the 172 studies were identified in this synthesis
as containing primary data on only surface water and/or groundwater.
Table D-3. Populations, Exposures, Comparators, and Outcomes (PECO) Criteria
PECO Element Inclusion Criteria
Population Adults and/or children in the general population and populations in the
vicinity of PFAS point sources from the United States, Canada, or Europe
Exposure Primary data from peer-reviewed studies collected in any of the following
media: ambient air, consumer products, drinking water, dust, food, food
packaging, groundwater51, human blood/serum/urine, indoor air, landfill,
sediment, soil, surface watera (freshwater), wastewater/biosolids/sludge
Comparator Not applicable
Outcome Measured concentrations of PFHxS (or measured emissions from food
packaging and consumer products only)
iVofes.-PFHxS = perfluorohexanesulfonic acid.
a Surface water and groundwater were not included as relevant media in Holder et al. (2023). Studies were re-screened for these
two media in this synthesis.
The evidence database of Holder et al. (2023) additionally identified 18 studies for which the
main article was not available for review. As part of this synthesis, 17 of the 18 studies could be
retrieved. An additional three peer-reviewed references were identified through gray literature
sources, described below, that were included to supplement the search results. The combined 20
studies underwent full-text screening using the inclusion criteria in Table D-3. Based on full-text
review, five studies were identified as relevant.
Using the screening results from the evidence database and this synthesis, a total of 177 peer-
reviewed studies were identified as relevant. Fifty of these contained information relevant to the
United States.
D.2.2. Additional Screening
The EPA also searched the following publicly available gray literature sources for information
related to relative exposure of PFHxS for all potentially relevant routes of exposure (oral,
inhalation, dermal) and exposure pathways relevant to humans:
• AT SDR's Toxicological Profiles.,
• CDC's national reports on human exposures to environmental chemicals;
• EPA's CompTox Chemicals Dashboard;
D-3 5
-------
FINAL
2024
• EPA's fish tissue studies;
• EPA's Toxics Release Inventory;
• Relevant documents submitted under the Toxic Substances Control Act and relevant
reports from EPA's Office of Chemical Safety and Pollution Prevention;
• U.S. Food and Drug Administration's (FDA's) Total Diet Studies and other similar
publications from FDA, U.S. Department of Agriculture, and Health Canada;
• NOAA's National Centers for Coastal Ocean Science data collections;
• National Science Foundation direct and indirect food and/or certified drinking water
additives;
• Throwaway Packaging, Forever Chemicals: European wide survey o/PFAS in
disposable food packaging and tableware (Strakova et al., 2021);
• PubChem compound summaries;
• Relevant sources identified in the relative source contribution discussions (Section 5) of
EPA's Proposed Approaches to the Derivation of a Draft Maximum Contaminant Level
Goal for Perfluorooctanoic Acid (PFOA)/Perfluorooctane Sulfonic Acid (PFOS) in
Drinking Water; and
• Additional sources, as needed.
The EPA has included available information from these gray literature sources for PFHxS
relevant to its uses, chemical and physical properties, and for occurrence in ambient or indoor
air, foods (including fish and shellfish), soil, dust, and consumer products. The EPA has included
available information specific to PFHxS below on any regulations that may restrict PFHxS levels
in media (e.g., water quality standards, air quality standards, food tolerance levels).
D.3. Summary of Potential Exposure Sources of PFHxS Other
than Water
D.3.1. Dietary Sources
D.3.1.1. Seafood
PFHxS was detected in 71 of 157 fish tissue composite samples collected during the EPA's
National Lake Fish Tissue Study, with a maximum concentration of 3.50 ng/g and a 50th
percentile concentration of <0.12 ng/g (Stahl et al., 2014). It was not detected in the 162 fish
tissue composite samples collected during the EPA's 2008-2009 National Rivers and Streams
Assessment (NRSA) (Stahl et al., 2014). More recently, PFHxS was detected in 32 of 349 fish
tissue composite samples at concentrations ranging from 0.121 ng/g to 0.980 ng/g in the EPA's
2013-2014 NRSA (USEPA, 2020a). PFHxS was also detected in 1 of 152 fish tissue composite
samples at a concentration of 0.96 ng/g in the EPA's 2015 Great Lakes Human Health Fish Fillet
Tissue Study (USEPA, 2021g). PFHxS has been detected in a mixture of fish fillet samples
collected from Mississippi River sites in Minnesota at a concentration of 0.47 ng/g (ATSDR,
2021; Delinsky et al., 2010). PFHxS has been detected in Irish pompano (Diapterus auratus),
silver porgy (Diplodus argenteus), and gray snapper (Lutjanus griseus) from the St. Lucie
Estuary in in NOAA's National Centers for Coastal Ocean Science, National Status and Trends
Data (NOAA, 2022). Burkhard (2021) identified 47 studies reporting BAFs for PFHxS and
D-36
-------
FINAL
2024
calculated a median (standard deviation) bioaccumulation factor (BAF) in muscle tissue/fillet of
19.95 ± 7.94 L/kg wet weight (reported as a logBAF of 1.30 ± 0.90 L/kg).
Five additional U.S. studies were identified that evaluated PFHxS levels in seafood (Young et
al., 2022; Chiesa et al., 2019; Byrne et al., 2017; Young et al., 2013; Schecter et al., 2010)
(Table D-4). One study evaluated fish samples collected directly from rivers and lakes (Byrne et
al., 2017). As part of a study to assess exposure to PFHxS and other PFAS among residents of
two remote Alaska Native villages on St. Lawrence Island, Byrne et al. (2017) measured PFAS
concentrations in stickleback and Alaska blackfish, resident fish used as sentinel species to detect
accumulation of PFAS in the local environment. Stickleback were collected from three locations
- Suqitughneq (Suqi) River watershed (n = 9 composite samples), Tapisaggak (Tapi) River
(n = 2 composite samples), and Troutman Lake (n = 3 composite samples). Blackfish were
collected from the Suqi River (n = 29) but were not found in the other water bodies. Authors
reported that the Suqi River watershed was upstream and downstream of a formerly used defense
site and Tapi River was approximately 5 km east of a military site, however at the start of the
study none of the sites were known to be contaminated with PFAS. The sample dates were not
reported. PFHxS was not detected in any of the stickleback and blackfish samples, despite the
authors noting that stickleback from Troutman Lake had "exceptionally high" total PFAS
concentrations.
The remaining four studies purchased seafood from stores and fish markets (Young et al., 2022;
Chiesa et al., 2019; Young et al., 2013; Schecter et al., 2010). Young et al. (2013) assessed fish
and shellfish collected in 2010-2012 from retail markets across the continental United States.
Retail markets in California, Florida, Illinois, Mississippi, New Jersey, New York, Tennessee,
Texas, and Washington, D.C., were represented. Authors selected the 10 most consumed fish and
shellfish in the United States that were farm raised, wild caught, or had unknown origin. Among
the crab meat, shrimp, striped bass, catfish, clams, cod, flounder, pangasius, pollock, tuna,
salmon, scallops, and tilapia, PFHxS was only detected in 1 of 10 samples of striped bass at a
concentration of 0.66 ng/g. Young et al. (2022) evaluated fish and shellfish purchased from retail
markets in the Washington. D.C. metropolitan area and online markets (clams only) from March
2021 through May 2022. Seafood samples represented 8 of the top 10 consumed fish and
shellfish in the United States. Seafood samples were farm raised, wild caught, or of unknown
origin, and location of harvest was provided when known. PFHxS was only detected in two
seafood types, crab and clam meat. All samples of clam meat (n = 10) had detectable
concentrations of PFHxS, ranging from 51 to 605 ng/kg. Only two samples of crabs (n = 11) had
detectable levels of 112 and 242 ng/kg. Authors also analyzed food packaging for PFAS analytes
and did not identify any packaging samples with detectable levels of PFAS. Schecter et al.
(2010) evaluated PFHxS and other PFAS in seafood collected from five Dallas, Texas grocery
stores in 2009. The origin or source of seafood was not described. Seafood included canned
sardines in water, canned tuna, fresh catfish fillet, cod, frozen fish sticks, salmon, and tilapia
(n = 1 composite sample for each seafood type). PFHxS was only detected in cod at a
concentration of 0.07 ng/g ww. Finally, in a multicontinental study, Chiesa et al. (2019) collected
salmon from a wholesale fish market in Milan, Italy; the sampling year was not reported. Wild-
caught salmon samples originated from the United States (n = 7), Canada (n = 15), and Scotland
(n = 2), while farmed salmon samples originated from Norway (n = 25) and Scotland (n = 17).
Among the salmon that originated from the United States Pacific Ocean (FAO 67 and 77), two
species - Oncorhynchus kisutch and Oncorhynchus keta - were analyzed, with PFHxS not
D-37
-------
FINAL
2024
detected in either species (LOQ = 0.015 ng/g). PFHxS was also not detected in wild-caught
salmon from Canada and Scotland.
In studies outside of the U.S., PFHxS was detected in multiple fish and shellfish species. Results
for all of the identified non-U.S. studies are presented in detail in Table D-4and are summarized
here. Approximately half of the studies including samples from outside the U.S. reported at least
one sample with detectable levels of PFHxS. Bhavsar et al. (2014) reported detections of PFHxS
in both cooked and raw samples of carp, lake trout, and walleye, but did not find detectable
levels in raw or cooked Chinook salmon, all of which were caught recreationally from rivers in
Ontario, Canada. PFHxS was also detected in several European studies that examined fresh-
caught or farmed seafood, including market-bought samples (Hansen et al., 2016; Carlsson et al.,
2014; Johansson et al., 2014; Yamada et al., 2014; Falandysz et al., 2006). The maximum PFHxS
level detected in seafood was 2.22 ng/g ww in brown trout collected from Norway (Hansen et al.,
2016). Though several studies reported on a variety of different species, many presented results
for relatively small sample sizes (n<5).
D-38
-------
FINAL
2024
Table D-4. Summary of PFHxS Data in Seafood
Study
Location and Source
Seafood Type
Results
United States
Byrne et al. (2017)
United States (Alaska)
Stickleback collected from three locations on St.
Lawrence Island: Suqitughneq (Suqi) River watershed
(upstream and downstream of a formerly used defense
site), Tapisaggak (Tapi) River (located approximately 5
km east of military site), and Troutman Lake, a coastal
lake situated adjacent to the village of Gambell.
Alaska blackfish collected from the Suqi River but were
absent from the other water bodies.
Sampling year not reported. No sites were known to be
contaminated with PFASs at the initiation of the study.
Stickleback and Alaska
blackfish
Stickleback:
Troutman Lake: n = 3*; DF 0%
Suqi River: n = 9*; DF 0%
Tapi River: n = 2*; DF 0%
Blackfish: n = 29; DF 0%
(LOQ = 0.5-1 ng/g ww for all PFAS)
*Number of composite samples, each composed of
~10 stickleback fish
Young et al. (2013)
United States (California; Illinois; Mississippi;
Tennessee; Florida; New Jersey; New York; Texas;
Washington, D.C.)
Fish and shellfish collected from retail markets in 11
areas across the continental United States from 2010-
2012. The fish and shellfish included farm raised, wild
caught, and unknown origin, as well as freshwater fish,
saltwater fish, and euryhaline fish.
Crab meat, clams, cod, flounder, pangasius, salmon,
scallops, and tilapia purchased from Washington, D.C.
Shrimp purchased from Orlando, Florida; Memphis,
Tennessee; and Nashville, Tennessee. Striped bass
purchased from New York, New York and Cherry Hill,
New Jersey. Catfish purchased from Indianola,
Mississippi; Dallas, Texas; Tampa, Florida; and Orlando,
Florida. Pollock purchased from Huntington Beach,
California. Tuna purchased from Chicago, Illinois.
Crab, shrimp, striped bass,
catfish, clams, cod, flounder,
pangasius, pollock, tuna (can
and pouch), salmon, scallops
(bay and sea), tilapia
Striped bass: n = 10, DFa 10%, range = ND-0.66*
ng/g
Crab meat: n = 1, DF 0%
Shrimp: n = 9, DF 0%
Catfish: n= 13, DF 0%
Clams: n = 1, DF 0%
Cod: n= 1,DF0%
Flounder: n = 1, DF 0%
Pangasius: n = 1, DF 0%
Pollock: n = 1, DF 0%
Tuna: n = 3, DF 0%
Salmon: n = 2, DF 0%
Scallops: n = 2, DF 0%
Tilapia: n = 1, DF 0%
(MDL = 0.55 ng/g for all seafood)
*This value was above the MDL but below the LOQ;
LOQ is estimated as 3x the MDL
Schecter et al. (2010)
United States (Texas)
Seafood samples from five different grocery stores in
Dallas, Texas were collected in 2009. Ten individual
samples were collected for each food type and combined
to form composite samples. The origin/source of the food
samples were not reported.
Salmon, canned tuna, fresh
catfish fillet, tilapia, cod,
canned sardines, frozen fish
sticks
Cod: n = 1, point = 0.07 ng/g ww, LOD = NR
Salmon: n = 1, DF 0%, LOD = 0.07 ng/g ww
Canned tuna: n = 1, DF 0%, LOD = 0.05 ng/g ww
Fresh catfish fillet: n = 1, DF 0%, LOD = 0.06 ng/g
WW
Tilapia: n = 1, DF 0%, LOD = 0.04 ng/g ww
Canned sardines: n = 1, DF 0%, LOD = 0.06 ng/g ww
Frozen fish sticks: n = 1, DF 0%, LOD = 0.09 ng/g
WW
*Number of composite samples, each composed of
~10 individual samples
D-39
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
Canada
Bhavsar et al. (2014)
Canada (Ontario)
Recreationally caught fish from four rivers - Credit
River, Thames River, Niagara River, Welland River - in
summer and fall of 2010 and 2011. Chinook salmon
were caught from Credit River, common carp from
Thames River, lake trout from Niagara River, and
walleye from Welland River. Elevated PFASs
concentrations were expected in the fish based on nearby
industrial activities or previous monitoring work
conducted by the Ontario Ministry of Environment. Raw
fish were analyzed, as well as cooked fish using three
different cooking methods (baking, broiling, and frying).
Raw and cooked fish
(chinook salmon, common
carp, lake trout, walleye)
Chinook salmon:
Raw: n = 5, DF NR, mean = <0.006 ng/g ww
Baked: n = 5, DF NR, mean = <0.006 ng/g ww
Broiled: n = 5, DF NR, mean = <0.006 ng/g ww
Fried: n = 5, DF NR, mean = <0.006 ng/g ww
Common carp:
Raw: n = 5, DF NR, mean = 0.292 ng/g ww
Baked: n = 5, DF NR, mean = 0.341 ng/g ww
Broiled: n = 5, DF NR, mean = 0.291 ng/g ww
Fried: n = 5, DF NR, mean = 0.359 ng/g ww
Lake trout:
Raw: n = 4, DF NR, mean = 0.258 ng/g ww
Baked: n = 4, DF NR, mean = 0.248 ng/g ww
Broiled: n = 4, DF NR, mean = 0.263 ng/g ww
Broiled; n = 4, DF NR, mean = 0.245 ng/g ww
Walleye:
Raw: n = 5, DF NR, mean = 0.080 ng/g ww
Baked: n = 5, DF NR, mean = 0.098 ng/g ww
Broiled; n = 5, DF NR, mean = 0.088 ng/g ww
Fried; n = 5, DF NR, mean = 0.083 ng/g ww
(LOQ not reported)
Europe
Hansen et al. (2016)
Norway (Evenes; Skanland)
Fish were sampled from Lake Langvatnet, Lake
Lavangsvatnet, River Tarstadelva, and the reference
Lake Strandvatnet. A civilian airport (location also
shared with the Air Station of the Royal Norwegian Air
Force) is situated on a ridge between Lake Langvatnet
and Lake Lavangsvatnet. These waters are affected by
PFAS due to AFFF emissions from the airport. Lake
Lavangsvatnet drains into the river Tarstadelva and Lake
Strandvatnet is ~15 km away from the airport with no
connection to the airport runoff. Samples of the
dorsolateral muscle were taken from 10 salmon, 10
anadromous brown trout, 12 stationary brown trout, and
3 European flounder by local fishermen and by personnel
from Sweco, an environmental consulting company. The
samples were collected in August and September 2014.
Brown trout, European
flounder, salmon
Brown trout (stationary):
Lake Langvatnet: n = 6, DFa 83%, range =
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
River Tarstadelva: n = 5, DFa 20%, range =
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
Yamada et al. (2014)
France
Marine fish sampled were selected based on the fish
consumption habits of the population of four areas - La
Rochelle in Gironde-Charente Maritime Sud, Le Havre
in Normandy-Baie de Seine, Lorient in South Brittany,
and Toulon in Mediterranean-Var. Five primary samples
of fish were bought from the fish market and/or
supermarket in each region for each species in January-
April 2005.
Freshwater fish sampled were selected based on the
individual dietary consumption analysis of anglers or
their family members of the ICAR-PCB study.
Freshwater fish were collected in six major French rivers
with each river divided into three or four section in
2008-2009. Half of the samples were composite
samples.
Freshwater fish, fresh or
frozen marine fish
Results presented for lower bound and upper bound if
LB value different from UB value
Fresh and frozen marine fish:
Total LB: n = 95, DF NR, mean (range) = 0.00 (0-
0.03) ng/g ww
Total UB: n = 95, DF NR, mean (range) = 0.03
(0.02-0.04) ng/g ww
Anchovy: n = 1, LB-UB = 0-0.03 ng/g ww
Monkfish: n = 4, LB-UB = 0-0.02 ng/g ww
Catshark: n = 4, LB-UB = 0.02-0.04 ng/g ww
Cod: n = 4, LB-UB = 0-0.02 ng/g ww
Common dab: n = 4, LB-UB = 0-0.02 ng/g ww
Orange roughy: n = 3, LB-UB = 0-0.03 ng/g ww
Plaice/witch: n = 2, LB-UB = 0-0.02 ng/g ww
Goatfish: n = 3, LB-UB = 0-0.03 ng/g ww
Grenadier: n = 4, LB-UB = 0-0.02 ng/g ww
Gurnard: n = 1, LB-UB = 0.03
Haddock: n = 2, LB-UB = 0-0.02 ng/g ww
Hake: n = 4, LB-UB = 0-0.02 ng/g ww
Halibut: n = 4, LB-UB = 0-0.03 ng/g ww
John dory: n = 2, LB-UB = 0.02-0.04 ng/g ww
Ling: n = 4, LB-UB = 0-0.02 ng/g ww
Mackerel: n = 4, LB-UB = 0-0.03 ng/g ww
Pollack: n = 3, LB-UB = 0-0.02 ng/g ww
Pout: n = 1, LB-UB = 0-0.03 ng/g ww
Ray: n = 4, LB-UB = 0-0.02 ng/g ww
Saithe: n = 4, LB-UB = 0-0.02 ng/g ww
Salmon: n = 4, LB-UB = 0-0.03 ng/g ww
Sardine: n = 4, LB-UB = 0-0.03 ng/g ww
Scorpionfish: n = 1, LB-UB = 0-0.03 ng/g ww
Seabass: n = 4, LB-UB = 0-0.03 ng/g ww
Sea bream: n = 4, LB-UB = 0-0.03 ng/g ww
Sole: n = 4, LB-UB = 0-0.02 ng/g ww
Swordfish: n = 4, LB-UB = 0-0.03 ng/g ww
Tuna: n = 4, LB-UB = 0-0.03 ng/g ww
Whiting: n = 4, LB-UB = 0-0.02 ng/g ww
Freshwater fish:
Barbel: n = 5, LB-UB = 0.19-0.22 ng/g ww
Bleak: n = 9, LB-UB = 0.02-0.41 ng/g ww
Brown trout: n = 31, LB-UB = 0.06-0.17 ng/g ww
Chub: n = 9, LB-UB = 0.05-0.16 ng/g ww
Common carp: n = 7, LB-UB = 0.67-0.76 ng/g ww
D-42
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
Common roach: n = 67, LB-UB = 0.06-0.18 ng/g
WW
Minnow: n = 1, LB-UB = 0.4 ng/g ww
European eel: n = 137, LB-UB = 0.67-0.77 ng/g
WW
European perch: n = 9, LB-UB = 0.11-0.2 ng/g ww
Freshwater bream: n = 34, LB-UB = 0.1-0.24 ng/g
WW
Gudgeon: n = 5, LB-UB = 0.36-0.39 ng/g ww
Northern pike: n = 8, LB-UB = 0.02-0.18 ng/g ww
White bream: n = 22, LB-UB = 0.21-0.3 ng/g ww
Thicklip grey mullet: n = 6, LB-UB = 0.01-0.16
ng/g ww
Wels catfish: n = 14, LB-UB = 0.01-0.12 ng/g ww
Western vairone: n = 1, LB-UB = 0-0.17 ng/g ww
Pike-perch: n = 22, LB-UB = 0.06-0.16 ng/g ww
(LOD = 0.007-0.95 ng/g for PFAAs other than PFOA
andPFOS)
*Lower bound (LB) scenario defined as values
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
were formed of 12 subsamples of the same food and of
equal weight. The fish were cooked according to the
practices reported in the survey of practices.
Sadia et al. (2020)
Sweden (Orebro)
Three fish samples from different brands were purchased
from a local supermarket in February 2019.
Fish
n = 3, DFa 33%, range = ND-0.0055 ng/g ww
(LOD = 0.0011; LOQ = 0.0049 ng/g)
Barbosaetal. (2018)
Belgium, France, The Netherlands, Portugal
Fish were collected from different markets based on the
assumption that the fish species were frequently
consumed in European Union countries and the fish
species contained high levels of contaminants of
emerging concern. Sampling year not reported. The
following fish species (origin, market country) were
included:
P. platessa: Channel, Belgium
M. australis: South America, Portugal
M. capenis: South Africa, Portugal
K. pelamis: Azores, Portugal
M. edulis: North Sea, The Netherlands; France, France
Raw and steamed fish (P.
platessa, M. australis, M.
capenis, K. pelamis, and M.
edulis)
P. platessa: n = 25, DF 0%
M. australis: n = 25, DF 0%
M. capenis: n = 25, DF 0%
K. pelamis: n = 25, DF 0%
M. edulis:
The Netherlands: n = 50, DF 0%
France: n = 50, DF 0%
(LOD = <0.01 ng/g ww for all PFCs)
Gebbink et al. (2015)
Sweden
Food items were purchased at two major grocery store
chains in four major Swedish cities in 1999 and 2005. In
2010, sampling was limited to Uppsala city since no
systematic geographical differences in food
contamination was observed in the two earlier market
basket studies. The food items were selected based on
Swedish food and production statistics and were not
cooked before analysis. Homogenates of fish products
(fresh and frozen fillets of fish, canned fish products,
shellfish) were prepared for each collection year by
mixing food items proportionally according to food
consumption statistics. Results were not reported for the
2005 and 2010 fish product composite samples (only
reported pooled with other food types).
Fish
1999: n = 1, point = 0.0069 ng/g
(MLOQ = 0.0001 ng/g)
*n represents number of composite samples
Vassiliadou et al. (2015)
Greece
Samples were obtained during the winter and early
spring of 2011. Finfish, squids, and shrimps were
purchased from the local fish market in Kallithea,
Athens, while mussels were obtained from a mariculture
farm within the Saronikos Gulf, Attika. Samples were
analyzed raw as well as cooked in the ways favored in
Greek cuisine (pan-fried in olive oil and/or grilled).
Anchovy, bogue, hake,
picarel, sand smelt, sardine,
striped mullet, mussel,
shrimp, squid
Anchovy (raw, fried, grilled), bogue (raw, fried,
grilled), hake (raw, fried, grilled), picarel (raw, fried),
sand smelt (raw, fried), sardine (raw, fried, grilled),
striped mullet (raw, fried, grilled), mussel (raw, fried),
shrimp (raw, fried), and squid (raw, fried, grilled): n =
4 for each, DF 0%
*n represents number of composite samples
(LOD = 0.18 ng/g ww; LOQ = 0.54 ng/g ww)
D-44
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
Quadruplicate composite samples were created for each
food type, each consisting of four to six items of raw or
cooked fish or shellfish.
Carlsson et al. (2014)
Greenland (Nuuk)
Seafood was purchased at the local fish market and
grocery shops in June 2010. All items were originally
caught in the vicinity of the Nuuk area and/or along the
West coast of Greenland and represented the common
food items consumed by the local Inuit population.
Salmon, halibut
Raw salmon fillet: n = 6, DF 0%
Smoked salmon fillet: n = 6, DF 0%
Smoked halibut fillet: n = 6, DF 0%
(LOD = 0.014-0.224 ng/g for all PFAS)
Perez et al. (2014)
Serbia (Belgrade and Novi Sad), Spain (Barcelona,
Girona, and Madrid)
Between September 2011 and February 2013, samples
were purchased from different supermarkets and retail
stores in representative cites around the world, including
cities in Serbia and Spain in Europe.
Bivalves, whiting, cod, hake,
salmon, herring, pangasius,
trout, tuna
Spain:
Bivalves (n = 28), whiting (n = 7), cod (n = 12),
hake (n = 3), salmon (n = 9), herring (n = 22),
pangasius (n = 9), trout (n = 19), tuna (n = 9): DF
0%
Serbia:
Canned tuna (n = 1), pangasius (n = 1), cod (n = 2),
herring (n = 5), trout (n = 2): DF 0%
(MLOD = 0.107 ng/g; MLOQ = 0.356 ng/g)
Domingo et al. (2012)
Spain (Catalonia)
Foods purchased from 4 shops/stores of each of the 12
representative cities of Catalonia (Barcelona, l'Hospitalet
de Llobregat, Vilanova I la Geltru, Matarr, Sabadell,
Terrassa, Girona, Tarragona, Reus, Tortosa, Lleida and
Manresa) in September 2011. Shops/stores included local
markets, small stores, supermarkets, and big grocery
stores. For each food item, two composite samples were
prepared for analysis, where each composite sample
consisted of 24 individual units. Only edible parts of
each food item were included in the composites.
Fish and seafood (sardine,
tuna, anchovy, swordfish,
salmon, hake, red mullet,
sole, cuttlefish, clam, mussel,
and shrimp)
n = 2, DF NR, mean = 0.045 ng/g fw
(LOD not reported)
*n represents number of composite samples
Vestergren et al. (2012)
Sweden (Malmoe, Gothenburg, Uppsala, Sundsvall)
Purchasing locations of the two largest retail chains in
Sweden were selected in each of four major Swedish
cities. All purchases were made in spring/summer of
1999,2005, and 2010. In 2010, the study was limited to
the largest retail chains in Uppsala located in close
vicinity to Stockholm. An equal amount of each food
group from each of the four cities was combined into one
sample pool to provide a representative sample for the
Swedish urban population.
Fish products (fresh and
frozen fillets of fish, canned
fish products, shellfish)
1999: n = 1, point = 0.0217 ng/g
2005: n = 1, point = 0.0088 ng/g
2010: n = 1, point = 0.0092 ng/g
(MDL = 0.0020 ng/g; MQL = 0.0059 ng/g)
*n represents number of composite samples
D-45
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
Noorlander et al. (2011)
The Netherlands
Fish randomly purchased from several Dutch retail stores
with nationwide coverage in November 2009. Fish
samples were ground, homogenized, and pooled for
analysis.
Fatty fish (herring, eel,
mackerel, salmon), lean fish
(cod, plaice, pollack, tuna),
crustaceans (mussels, shrimp,
crab)
Fatty fish: n = 1, point = 0.009 ng/g
Lean fish: n = 1, point = 0.023 ng/g
Crustaceans: n = 1, point = 0.044 ng/g
(LOD not reported)
*n represents number of composite samples
Jogsten et al. (2009)
Spain (Catalonia)
Fish samples purchased from local markets, large
supermarkets, and grocery stores from two different
areas of Tarragona Province, Catalonia in January and
February 2008. The cities of Tarragona and Reus were
sampled in the northern area and L'Ametlla de Mar and
Tortosa in the southern area. For each food item, two
composite samples were analyzed (one composite for the
northern area and one for the southern area). Each
composite was formed of a minimum of six individual
sub-samples of the same product.
Marinated salmon
(homemade and packaged)
Homemade: n = 2, DF NR, mean = 0.014 ng/g
Packaged: n = 2, DF NR, mean = <0.003 ng/g
(LOD = 0.001 ng/g)
*n represents number of composite samples
*Values of ND were replaced with i^xLOD
Ericson et al. (2008a)
Spain
Food samples purchased from local markets, large
supermarkets, and grocery stores within Tarragona
County in July 2006. Food samples were randomly
purchased with origin source not specified. Each of the
food samples were duplicated and combined to analyze a
composite sample. Only the edible part of each food was
included in the composite samples. Composite samples
included the following:
White fish: hake, whiting blue, sea bass, monkfish
Seafood: mussel, shrimp
Tinned fish: tuna, sardine, mussel
Blue fish: salmon, sardine, and tuna
White fish, seafood, tinned
fish, blue fish
White fish: n = 2, DF 0%
Seafood: n = 2, DF 0%
Tinned fish: n = 2, DF 0%
Blue fish: n = 2, DF 0%
(LOD = 0.001-0.65 ng/g fw)
*n represents number of composite samples
Johansson et al. (2014)
Sweden
Farmed rainbow trout (whole fish) were collected from
fish farms along the Swedish Baltic Sea coast (brackish
water). Only fish older than 12 months were sampled.
Samples were collected annually from 1999 to 2010
within the Swedish National Food Agency's official food
control program.
Rainbow trout
Total: n = 36, DFa 58%, range = <0.011-0.04 ng/g
fw
1999: n = 10, DFa 90%, range = <0.011-0.034 ng/g
fw
2000: n = 3, DFa 100%, mean3 (range) = 0.014
(0.012-0.017 ng/g fw
2001: n = 4, DFa 75%, range = <0.011-0.040 ng/g
fw
2002: n = 1, point = 0.017 ng/g fw
2003: n = 2, DF 0%
2004: n = 1, point = 0.020 ng/g fw
D-46
-------
FINAL
2024
Study
Location and Source
Seafood Type
Results
2005: n = 4, DFa 25%, range = <0.011-0.020 ng/g
t\v
2006: n = 3, DFa 33%, range = <0.011-0.014 ng/g
t\v
2007: n = 2, DFa 50%, range = <0.011-0.029 ng/g
t\v
2008: n = 1, DF 0%
2009: n = 4, DFa 25%, range = <0.011-0.015 ng/g
t\v
2010: n =1, DF 0%
(MDL = 0.011 ng/g fw; MQL = 0.025 ng/g fw)
Multiple Continents
Chiesaetal. (2019)
United States (Pacific Ocean)
Wild-caught fish were collected at a wholesale fish
market in Milan, Italy. Sampling year was not reported.
The wild-caught salmon were from USA-Pacific Ocean
(Food and Agriculture Organization Area 67 and 77).
Wild-caught salmon
(Oncorhynchus kisutch and
Oncorhynchus keta)
Oncorhynchus kisutch. n = 5, DF 0%
Oncorhynchus keta: n = 2, DF 0%
(LOQ = 0.015 ng/g)
Canada
Wild-caught fish were collected at a wholesale fish
market in Milan, Italy. Sampling year was not reported.
The wild-caught salmon were from Canada (Food and
Agriculture Organization Area 67).
Wild-caught salmon
(Oncorhynchus nerka)
n = 15, DF 0%
(LOQ = 0.015 ng/g)
Norway
Farmed fish were collected at a wholesale fish market in
Milan, Italy. Sampling year was not reported. The wild-
caught salmon were from Norway (Food and Agriculture
Organization Area 27).
Farmed salmon (Salmo salar)
n = 25, DF 0%
(LOQ = 0.015 ng/g)
Scotland
Wild-caught and farmed fish were collected at a
wholesale fish market in Milan, Italy. Sampling year was
not reported. The wild-caught salmon were from
Scotland (Food and Agriculture Organization Area 27).
Wild-caught and farmed
salmon (Salmo salar)
Wild-caught: n = 2, DF 0%
Farmed: n = 17, DF 0%
(LOQ = 0.015 ng/g)
Notes: DF = detection frequency; LOD = limit of detection; LOQ = limit of quantitation; MDL = method detection limit; ND = not detected; NR = not reported; ww = wet weight.
Bold indicates detected levels of PFHxS in food.
a The DF and/or mean was not reported in the study and was calculated in this synthesis. Means were calculated only when DF = 100%.
D-47
-------
FINAL
2024
4.1.1.1 Other Food Sources
PFHxS was included in a suite of PFAS evaluated in FDA's 2019, 2021, and 2022 Total Diet
Study Sampling (US FDA, 2022b, 2022a, 2021b, 2021a, 2020b, 2020a); however, it was not
detected in any of the food samples tested. It should be noted that FDA indicated that the sample
sizes used in the PFAS 2019, 2021, and 2022 Total Diet Study Sampling were limited and that
the results should not be used to draw definitive conclusions about PFAS levels in the general
food supply (US FDA, 2022c). PFHxS was detected in milk samples collected from a farm with
groundwater known to be contaminated with PFAS; however, it was not detected in produce
collected from an area near a PFAS production plant, in FDA studies of the potential exposure to
the U.S. population to PFAS (US FDA, 2021c, 2018). PFHxS is not a registered pesticide under
the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), and the EPA does not set a 40
CFR Part 180 pesticide tolerance in food and feed commodities for PFHxS (US GPO, 2022).
Maximum residue levels for PFHxS were not found in the Global Maximum Residue Level
Database (Bryant Christie Inc, 2022).
Several peer-reviewed studies were identified that examined PFHxS in food sources other than
seafood, including cereals, dairy, fats/other (e.g., eggs, oils, and spices), fruits and vegetables,
meat, and breastmilk (Table D-5). Few U.S. studies analyzed foods from any one origin - four
sampled crops grown in areas with known or suspected PFAS contamination, including
biosolids-amended soils, two sampled from crops as part of greenhouse and field studies, one
studied wild-caught alligator meat. Only two studies sampled from store- or market-bought
meats, eggs, produce, and dairy.
Scher et al. (2018) evaluated garden produce samples from homes in Minnesota within and
outside of a GCA in the vicinity of a former 3M PFAS production facility. Twenty homes within
the GCA had previous or ongoing PFAS contamination in drinking water and were served by the
Oakdale, Minnesota public water system or a private well previously tested and shown to have
detectable levels of PFOA or PFOS. A total of 279 produce samples (232 inside GCA, 47 outside
GCA) were collected between May and October 2010. PFHxS was detected in 1% of the 232
produce samples from inside the GCA (one floret sample and one leaf sample). The authors
suggested that the two detections were associated with PFAS present in irrigation water that had
accumulated in produce. They also noted that accumulation of PFAS was particularly high in
florets. Three homes that were outside the GCA served as a reference. No PFHxS was detected
in produce samples from home gardens outside the GCA. Genualdi et al. (2017) analyzed PFAS
contamination in a Massachusetts cranberry bog approximately 10 miles from a military base
with a history of AFFF usage. Ten cranberry samples were taken directly from trucks
transporting cranberries and 32 cranberry samples were collected directly from the bog water in
November 2016. PFHxS was not detected in any samples (MDL = 0.79 ng/g).
Two studies purchased food items from stores and markets for evaluation (Young et al., 2012;
Schecter et al., 2010). Schecter et al. (2010) assessed PFHxS and other PFAS in food samples
collected from five Dallas, Texas grocery stores in 2009. The origin or source of each food was
not described. Food items included meat products (bacon, canned chili, chicken breast, ground
beef, roast beef, ham, sausage, and turkey), dairy (butter, cheeses, frozen yogurt, ice cream, milk,
and yogurt), eggs, grains (cereal), fruits and vegetables (apples, potatoes), and fats/other (canola
oil, margarine, olive oil, peanut butter). PFHxS was not detected in any of the food samples. In
Young et al. (2012), cow milk was purchased from retail markets across the continental United
D-48
-------
FINAL
2024
States representing 17 states; the sampling year was not reported. Cow milk samples included
organic milk, vitamin D added milk, and ultra-pasteurized milk. PFHxS was not detected in any
of the 49 retail milk samples (MDL = 0.15 ng/g).
One study investigated PFAS levels in wild meat (Tipton et al., 2017). Tipton et al. (2017)
assessed alligator tail meat that was collected during the South Carolina recreational hunting
season between September to October 2015. Tail meat samples were collected from four
different public hunt units - Southern Coastal, Middle Coastal, Midlands, and Pee Dee. PFHxS
was detected in all samples from all hunt units. Median concentrations from Southern Coastal
(n = 19), Middle Coastal (n = 17), Midlands (n = 5), and Pee Dee (n = 2) were 0.087 ng/g,
0.099 ng/g, 0.0816 ng/g, and 0.093 ng/g wet mass, respectively.
Two studies by Blaine et al. (2014; 2013) evaluated PFHxS in crops grown in greenhouse and
field studies. In Blaine et al. (2014), PFAS levels were measured in celery root, pea fruit, and
radish root grown in a greenhouse with control (unamended) soil, industrially impacted soil, and
municipal soil (n = 3-5). PFHxS was detected in radish root from all three soils, celery shoot
from the industrially impacted and municipal soil, and pea fruit from only industrially impacted
soil. Mean concentrations of PFHxS in radish root for the control, industrially impacted, and
municipal soil were 3.81 ng/g, 2.84 ng/g, and 4.33 ng/g, respectively. Mean concentrations of
PFHxS in celery shoot for the industrially impacted and municipal soil were 3.19 ng/g and
0.38 ng/g, respectively. The mean concentration of PFHxS in pea fruit in the industrially
impacted soil was 0.24 ng/g. Authors noted minor cross-contamination of the control soil due to
the proximity of the unamended soil to biosolids-amended soil. In Blaine et al. (2013), authors
studied the uptake of PFAS into edible crops in both field and greenhouse studies. In the field
study, PFAS levels were measured in corn grain and corn stover grown with control
(unamended), urban biosolids-amended, and rural biosolids-amended soil (n = 3-7). Mean
PFHxS concentrations were below the LOQ in both corn grain and corn stover grown in any
field study plots (<0.04 ng/g for corn grain; <0.29 ng/g for corn stover). In the greenhouse study,
lettuce and tomato plants were grown in control soil, industrially impacted soil, or municipal soil
(n = 3-5). Mean PFHxS concentrations were below the LOQ for lettuce and tomato grown in the
control soil and for tomato grown in municipal soil; however, mean PFHxS levels were
10.44 ng/g and 5.54 ng/g for lettuce grown in industrially impacted and municipal soils,
respectively, and 0.76 ng/g for tomato grown in industrially impacted soil. Sampling year was
not reported.
The remaining two U.S. studies evaluated the occurrence of PFHxS in breastmilk (von
Ehrenstein et al., 2009; Kuklenyik et al., 2004). von Ehrenstein et al. (2009) collected breastmilk
samples between December 2004 and July 2005 from women between the ages of 18 and 38 at
the time of recruitment as part of the pilot study Methods Advancement for Milk Analysis
(MAMA). Women provided milk samples at two visits - the first visit was 2-7 weeks
postpartum, and the second visit was 3-4 months postpartum. PFHxS was not detected in any of
the samples from the first visit (n = 18) or second visit (n = 20). Similarly, PFHxS was below the
LOD (0.3 ng/mL) in the samples reported by Kuklenyik et al. (2004). Kuklenyik et al. (2004) did
not report information on the breastmilk donors or the sampling procedure as it was unavailable;
PFHxS was not detected in either of the two samples.
Of the studies conducted in Europe examining non-breastmilk and non-seafood food items, 15
found PFHxS in at least one food type. Results for all of the identified non-U. S. studies are
D-49
-------
FINAL
2024
presented in detail in Table D-5 and are summarized here. Across the European studies, PFHxS
was found in animal products such as meat, dairy, and eggs, as well as fruits and vegetables.
Similarly to studies examining seafood, most studies reported on a variety of different food
types, but the majority presented results for relatively small sample sizes (n<5). Lastly, eight of
the twelve studies conducted outside the U.S. examining PFHxS in breastmilk had detectable
levels.
D-50
-------
FINAL
2024
Table D-5. Summary of PFHxS Data in Other Food
Study
Location and Source
Food Types
Results
United States
Scheretal. (2018)
United States (Minnesota)
Home garden produce samples were collected
between May and October 2010 from 20 homes in 3
cities within a GCA as well as 3 homes in the Twin
Cities Metro outside the GCA. Homes within the
GCA were near a former 3M PFAS production
facility, had previous or ongoing PFAS
contamination in drinking water, and were served by
the Oakdale, Minnesota public water system or were
formerly or currently using a private well previously
tested and shown to have detectable levels of PFOA
orPFOS.
279 produce samples (232 within GCA and 47
outside GCA) consisting of mature, edible portions
of plants were analyzed. Plant part categories
included floret, fruit, leaf, root, seed, and stem.
Fruits and vegetables
Within GCA:
All: n = 232, DF 1%, median (range) = ND (ND-
0.066) ng/g
Floret: n = 5, DF 20%, median (range) = ND
(ND-0.066) ng/g
Leaf: n = 35, DF 3%, median (range) = ND
(ND-0.046) ng/g
Garden fruit (n = 98), yard fruit (n = 13), root (n =
29), seed (n = 29), and stem (n = 23): DF 0%
Outside GCA:
All: n = 47, DF 0%
Floret (n = 1), garden fruit (n = 15), yard fruit (n =
4), leaf (n = 12), root (n = 5), seed (n = 5), and
stem (n = 5): DF 0%
(MDL = 0.003 to 0.029 ng/g depending on the
analyte and type of produce)
Genualdi et al. (2017)
United States (Massachusetts)
Samples from cranberry bog with surface water
contaminated with PFAS—likely due to proximity
to a military base with a history of AFFF usage. The
bog was located approximately 10 miles from the
military base. Ten cranberry samples taken directly
from trucks transporting cranberries (five samples
each from two trucks) and 32 cranberry samples
taken directly from 12 sections of the bog water.
Samples collected in November 2016.
Fruits
n = 42, DF 0%
(MDL = 0.79 ng/g)
Schecter et al. (2010)
United States (Texas)
Food samples from five different grocery stores in
Dallas, Texas were collected in 2009. Ten individual
samples were collected for each food type and
combined to form composite samples. The
origin/source of the food samples were not reported.
Dairy; fruits and vegetables;
grains; meat; fats/other
Meat
Hamburger: n = 1, DF 0%, LOD = 0.04 ng/g ww
Bacon: n = 1, DF 0%, LOD = 0.05 ng/g ww
D-51
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Sliced turkey: n = 1, DF 0%, LOD = 0.02 ng/g ww
Sausages: n = 1, DF 0%, LOD = 0.04 ng/g ww
Ham: n = 1, DF 0%, LOD = 0.02 ng/g ww
Sliced chicken breast: n = 1, DF 0%, LOD = 0.02
ng/g ww
Roast beef: n = 1, DF 0%, LOD = 0.02 ng/g ww
Canned chili: n = 1, DF 0%, LOD = 0.01 ng/g ww
Dairy and Eggs
Butter: n = 1, DF 0%, LOD = 0.09 ng/g ww
American cheese: n = 1, DF 0%, LOD = 0.04 ng/g
WW
Other cheese: n = 1, DF 0%, LOD = 0.04 ng/g ww
Whole milk: n = 1, DF 0%, LOD = 0.02 ng/g ww
Ice cream: n = 1, DF 0%, LOD = 0.03 ng/g ww
Frozen yogurt: n = 1, DF 0%, LOD = 0.02 ng/g
WW
Whole milk yogurt: n = 1, DF 0%, LOD = 0.03
ng/g ww
Cream cheese: n = 1, DF 0%, LOD = 0.02 ng/g
WW
Eggs: n = 1, DF 0%, LOD = 0.04 ng/g ww
Grains
Cereals: n = 1, DF 0%, LOD = 0.04 ng/g ww
Fruits and Vegetables
Apples: n = 1, DF 0%, LOD = 0.02 ng/g ww
Potatoes: n = 1, DF 0%, LOD = 0.04 ng/g ww
Fats/Other
D-52
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Olive oil: n = 1, DF 0%, LOD = 0.3 ng/g ww
Canola oil: n = 1, DF 0%, LOD = 0.5 ng/g ww
Margarine: n = 1, DF 0%, LOD = 0.03 ng/g ww
Peanut butter: n = 1, DF 0%, LOD = 0.03 ng/g ww
*n reflects number of composite samples, each
composed of ~10 individual samples
Young et al. (2012)
United States (17 states)
Retail cow's milk samples were all pasteurized
whole milk, commercially available, and purchased
at retail markets across the continental United States
representing 17 states. Samples were organic milk,
vitamin D added milk, and ultra-pasteurized milk.
Sampling year not reported.
Dairy
n = 49, DF 0%,
(MDL = 0.15 ng/g)
Tipton et al. (2017)
United States (South Carolina)
Alligator tail meat samples were collected from a
local wild game meat processer during the South
Carolina recreational hunt season between
September to October 2015. Samples were from
four different public hunt units—Southern Coastal,
Middle Coast, Midlands, and Pee Dee.
Meat
Alligator tail:
Southern coastal: n = 19, DFa 100%, median
(range) = 0.087 (0.051-0.252) ng/g wet mass
Middle coastal: n = 17, DFa 100%, median
(range) = 0.099 (0.063-0.272) ng/g wet mass
Midlands: n = 5, DFa 100%, median (range) =
0.0816 (0.054-0.158) ng/g wet mass
Pee Dee: n = 2, DFa 100%, median (range) =
0.093 (0.071-0.115) ng/g wet mass
(RL not reported)
Blaine et al. (2014)
United States (Midwest)
Crops grown in in greenhouse study with control
(unamended), industrially impacted soil, or
municipal soil. Control soil had minor cross-
contamination due to proximity to biosolids-
amended fields. Industrially impacted soil was
amended with industrially impacted biosolids, and
municipal soil was amended with municipal
biosolids for over 20 years.
Crops grown in the greenhouse study were grown
from seed in pots, which were randomly arranged
within the greenhouse. Sampling year not reported.
Fruits and vegetables
Radish root:
Control: n = 3-5, DF NR, mean = 3.81 ng/g
Industrially impacted; n = 3-5, DF NR, mean =
2.84 ng/g
Municipal: n = 3-5, DF NR, mean = 4.33 ng/g
Celery shoot:
Control: n = 3-5, DF 0%
Industrially impacted: n = 3-5, DF NR, mean =
3.19 ng/g
D-53
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Municipal: n = 3-5, DF NR, mean = 0.38 ng/g
Pea fruit:
Control: n = 3-5, DF 0%
Industrially impacted: n = 3-5, DF NR, mean =
0.24 ng/g
Municipal: n = 3-5, DF 0%
(LOQ = 0.03 ng/g)
Blaine et al. (2013)
United States (Midwest)
Crops grown in urban and rural full-scale field study
with control (unamended) and biosolids-amended
soil. Three agricultural fields were amended (0.5 x,
1 x, or 2x) with municipal biosolids. Urban biosolids
(1 x and 2*) were from a WWTP receiving both
domestic and industrial waste. Rural biosolids (0.5x)
were from a WWTP receiving domestic waste only.
Control plots were proximal to the rural and urban
amended corn grain and corn stover field sites;
sampling year not provided.
Crops grown in greenhouse study with control
(nonamended) and biosolids-amended soil.
Nonamended soil obtained from a field that received
commercial fertilizers and had a similar cropping
system as the nearby municipal soil site. Municipal
soil was obtained from a reclamation site in Illinois
where municipal biosolids were applied at
reclamation rates for 20 years, reaching the
cumulative biosolids application rate of 1,654
Mg/ha. Industrially impacted soil was created by
mixing composted biosolids from a small municipal
(but impacted by PFAA manufacturing) WWTP
with control soil on a 10% mass basis. Sampling
year not provided.
Fruits and vegetables; grains
Field study:
Corn grain:
Urban nonamended: n = 3-7, DF NR, mean =
<0.04 ng/g
Urban 1 x: n = 3-7, DF NR, mean = <0.04 ng/g
Urban 2x: n = 3-7, DF NR, mean = <0.04 ng/g
Rural nonamended: n = 3-7, DF NR, mean =
<0.04 ng/g
Rural 0.5x: n = 3-7, DF NR, mean = <0.04 ng/g
Corn stover:
Urban nonamended: n = 3-7, DF NR, mean =
<0.29 ng/g
Urban 1 x: n = 3-7, DF NR, mean = <0.29 ng/g
Urban 2x: n = 3-7, DF NR, mean = <0.29 ng/g
Rural nonamended: n = 3-7, DF NR, mean =
<0.29 ng/g
Rural 0.5x: n = 3-7, DF NR, mean = <0.29 ng/g
(LOQ = 0.04 ng/g for corn grain; LOQ = 0.29 ng/g
for corn stover)
Greenhouse study:
Lettuce:
D-54
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Nonamended: n = 3-5, DF NR, mean = <0.01
ng/g
Industrially impacted: n = 3-5, DF NR, mean
= 10.44 ng/g
Municipal: n = 3-5, DF NR, mean = 5.54 ng/g
Tomato:
Nonamended: n = 3-5, DF NR, mean = <0.03
ng/g
Industrially impacted: n = 3-5, DF NR, mean
= 0.76 ng/g
Municipal: n = 3-5, DF NR, mean = <0.03 ng/g
(LOQ = 0.01 ng/g for lettuce; LOQ = 0.03 ng/g for
tomato)
von Ehrenstein et al. (2009)
United States (North Carolina)
As part of the Methods Advancement for Milk
Analysis (MAMA) pilot study, 34 breastfeeding
women aged 18 to 38 years at recruitment provided
breastmilk samples at two visits. The first visit
occurred 2-7 weeks postpartum, and the second visit
occurred 3^1 months postpartum. Both visits were
between December 2004 and July 2005.
Breastmilk
Visit #1: n= 18,DF0%
Visit #2: n = 20, DF 0%
(LOQ = 0.30 ng/mL)
Kuklenyik et al. (2004)
United States (Georgia)
Authors reported that no information was provided
on the human milk donors or the sampling
procedure.
Breastmilk
n = 2, DF 0%
(LOD = 0.3 ng/mL)
Canada
Kubwabo et al. (2013)
Canada (Ontario)
Breastmilk samples were collected in the Kingston
region of Ontario in 2003-2004.
Breastmilk
n = 13, DF 0%
(MDL = 0.125 ng/mL; LOQ = 0.416 ng/mL)
Europe
Scordo et al. (2020)
Italy
Commercially available strawberry and olive fruits
were purchased in two Italian supermarkets in 2018.
Fruits
Strawberries: n = 2, DFa 100%, mean3 (range) =
0.469 (0.148-0.790) ng/g dw
(MDL = 0.010 ng/g; MQL = 0.037 ng/g)
D-55
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Olives: n = 2, DFa 50%, range = <0.0024-0.022
ng/g dw
(MDL = 0.0024 ng/g; MQL = 0.0086 ng/g)
Sadia et al. (2020)
Sweden (Orebro)
Three samples of each food type (cow milk, butter,
chicken meat, beef) of different brands were
purchased from a local supermarket in February
2019.
Dairy; fats/other
Milk: n = 3, DFa 100%, mean3 (range) = 0.0021
(0.0014-0.0030) ng/g ww
Butter: n = 3, DFa 100%, mean3 (range) = 0.0082
(0.0026-0.018) ng/g ww
Beef: n = 3, DFa 67%, range = ND-0.0044 ng/g
WW
Chicken: n = 3, DFa 100%, mean3 (range) =
0.0035 (0.0033-0.0036) ng/g ww
(LOD = 0.0011; LOQ = 0.0049 ng/g)
Sznajder-Katarzynska et al.
(2019)
Poland
Milk and milk products were purchased in Polish
markets in 2017. Commercially available samples of
each product were obtained from five different
suppliers.
Dairy
All daily: n = 35, DF 48.6%, sum PFHxS = 0.47
ng/g
Milk: n = 5, DFa 60%, range = 0.01-0.01 ng/g
Cottage cheese: n = 5, DFa 100%, range = 0.04-
0.05 ng/g
Natural yogurt: n = 5, DFa 60%, range = 0.01-
0.02 ng/g
Kefir/Bonny clabber: n = 5, DF3 60%, range =
0.01-0.02 ng/g
Butter: n = 5, DFa 60%, range = 0.02-0.03 ng/g
Sour cream (n = 5), camembert-type cheese
(n = 5): DF 0%
(LOD = 0.003 ng/g; LOQ = 0.010 ng/g)
*Range reported for detected values
Riviere et al. (2019)
France
Samples collected between July 2011 and July 2012
in the center region of France. Food items were
selected based on the results of a national
consumption survey to obtain a representative and
general view of children's (0-3 years old) food
consumption. All analyzed samples were formed of
12 subsamples of the same food and of equal
weight. The products purchased were prepared in a
way that reflected as closely as possible what is
done in the home (preparation and cooking).
Meat; dairy; fruits and
vegetables; fats/other
Infant-specific foods:
Milk-based beverage (n = 8), cereals (n = 5), milk-
based desserts (n = 6), growing-up milk (n = 9),
soups and puree (n = 11), fruit puree (n = 4),
vegetable-based ready-to-eat meal (n = 20),
meat/fish-based ready-to-eat meal (n = 45), infant
formula (n = 28), follow-on formula (n = 33): DF
0%
Common foods:
D-56
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Non-alcoholic beverages (n = 1), dairy-based
desserts (n = 1), milk (n = 1), mixed dishes (n = 1),
ultra-fresh dairy products (n = 1), meat (n = 1),
poultry and game (n = 1): DF 0%
(LOD = 0.0002-3.7 ng/g for all PFAS)
*n represents number of composite samples
Sznajder-Katarzynska et al.
(2018)
Poland
Samples were purchased in Polish markets in 2017.
Individual food items were selected among the most
frequently consumed in Poland. Vegetables
(potatoes, beetroots, carrots, white cabbage,
tomatoes) and fruits (apples, cherries, strawberries)
of Polish origin were bought in season when
naturally ripe. Bananas, lemons, and oranges were
bought after being imported to Poland. Five
different samples of each fruit or vegetable were
collected.
Fruits and vegetables
Apples, bananas, cherries, lemons, oranges,
strawberries, beetroots, carrots, tomatoes, potatoes,
and white cabbage: n = 5 for each, DF 0%
(LOD = 0.006 ng/g; LOQ = 0.017 ng/g)
Sunna et al. (2017)
Spain, Slovakia
Spice samples were collected in powder form from
Spain and Slovakia. Sampling year not reported.
Fats/other
Spain:
Anise (n = 1), star anise (n = 1), white pepper (n =
1), fennel (n = 1), cardamom (n = 1), clove (n = 1),
coriander (n = 1), nutmeg (n = 1), allspice (n = 1),
cinnamon (n = 2), vanilla (n = 1), ginger (n = 1),
peppermint (n = 1), parsley (n = 1), thyme (n = 1),
laurel (n = 1), garlic (n = 1), cumin (n = 1), black
pepper (n = 1), mild hot pepper (n = 1), hot hot
pepper (n = 1), oregano (n = 2): DF 0%
Slovakia:
Anise (n = 1), star anise (n = 1), white pepper (n =
1), fennel (n = 1), cardamom (n = 1), clove (n = 1),
coriander (n = 1), nutmeg (n = 1), allspice (n = 1),
cinnamon (n = 1), vanilla (n = 1), ginger (n = 1):
DF 0%
(LOD = 0.013 ng/g; LOQ = 0.039 ng/g)
Zafeiraki et al. (2016a)
Greece, The Netherlands
Home and commercially-produced eggs were
collected from different regions in the Netherlands
and Greece in August 2013-August 2014. Home-
produced eggs were voluntarily provided, and
Fats/other
Domestic eggs:
The Netherlands: n = 73, DF 7%, median
(range) = 1.1 (<0.05-5.2) ng/g
D-57
-------
FINAL
2024
Study
Location and Source
Food Types
Results
commercial eggs were purchased from
supermarkets. The yolks of the same sample of eggs
were pooled, homogenized, and then analyzed.
Greece: n = 45, DF 0%, median (range) = <0.5
(<0.5)ng/g
Commercial eggs:
The Netherlands: n = 22, DF 0%
Greece: n = 31, DF 0%
(LOD = 0.15 ng/g; LOQ = 0.5 ng/g)
*Median calculated only on the concentrations above
LOQ
Zafeiraki et al. (2016b)
The Netherlands
Samples purchased from local markets and
slaughterhouses in the Netherlands in 2014. Samples
included liver samples of horse, sheep, bovine, pig,
and chicken.
Meat
Horse: n = 19, DF 0%
Sheep: n = 18, DF 0%
Bovine: n = 22, DF 0%
Pig: n = 20, DF 0%
Chicken: n = 20, DF 0%
(LOQ = 0.5 ng/g ww)
D'Hollander et al. (2015)
Czech Republic, Belgium, Norway, Italy
The Czech Republic, Belgium, Norway, and Italy
were selected to represent eastern, western, northern,
and southern Europe. Sampling took place between
spring and summer 2010 as part of the PERFOOD
study. Individual items were randomly selected in
three national retail stores covering different brands
or countries of origin. Of each item, three to ten
single samples were combined to create a pooled
sample. The food items examined were:
Cereals: rice, wheat (white), oats, rye
Sweets: sugar (beet), sugar (cane), honey
Fruits - berries: strawberries
Fruits - citrus fruit: oranges, tangerines, lemons,
grapefruits
Fruits - pipe and stone fruit: apples, pears,
peaches, plums
Fruits - others and exotic fruit: melons, grape,
bananas
Grain; fruits; fats/other
Czech Republic:
Cereals: wheat (white), oats, rye: n = 1 each, point
= <0.010 ng/g
Sweets: sugar (beet), honey: n = 1 each, point =
<0.004 ng/g
Fruits - berries: strawberries: n = 1, point = <0.004
ng/g
Fruits - citrus fruit: oranges, tangerines: n = 1,
point = <0.004 ng/g
Fruits - pipe and stone fruit: apples, pears,
peaches: n = 1, point = <0.004 ng/g
Fruits - others and exotic fruit: melons: n = 1,
point = <0.004 ng/g
Miscellaneous: rock salt: n = 1, point = <0.004
ng/g
Italy:
Cereals:
D-58
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Miscellaneous: "rock" salt, "marine" salt
Rice, maize: n = 1 each, point = <0.004 ng/g
Wheat (white): n = 1, point = <0.010 ng/g
Sweets: sugar (beet), honey: n = 1 each, point =
<0.004 ng/g
Fruits - citrus fruit: lemons: n = 1, point = <0.004
ng/g
Fruits - pipe and stone fruit:
Apples: n = 1, point = 0.015 ng/g
Pears: n = 1, point = <0.004 ng/g
Peaches: n = 1, point = 0.016 ng/g
Plums: n = 1, point = <0.004 ng/g
Fruits - others and exotic fruit:
Grapes: n = 1, point = <0.004 ng/g
Bananas: n = 1, point = 0.008 ng/g
Miscellaneous: marine salt: n = 1, point = <0.004
ng/g
Norway:
Cereals: wheat (white): n = 1, point = <0.010 ng/g
Sweets: sugar (cane), honey: n = 1 each, point =
<0.004 ng/g
Fruits - berries: strawberries: n = 1, point = <0.004
ng/g
Fruits - citrus fruit:
Oranges: n = 1, point = <0.004 ng/g
Grapefruits: n = 1, point = 0.0189 ng/g
Fruits - pipe and stone fruit: apples, pears: n = 1
each, point = <0.004 ng/g
D-59
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Fruits - others and exotic fruit: melons: n = 1,
point = 0.0038 ng/g
Miscellaneous: rock salt: n = 1, point = <0.004
ng/g
Belgium:
Cereals: rice, wheat (white), wheat (dark), oats: n
= 1 each, point = <0.004 ng/g
Sweets: sugar (beet), honey: n = 1 each, point =
<0.004 ng/g
Fruits - berries: strawberries: n = 1, point =
0.123 ng/g
Fruits - citrus fruit: oranges, lemons: n = 1 each,
point = <0.004 ng/g
Fruits - pipe and stone fruit:
Apples: n = 1, point = 0.197 ng/g
Pears: n = 1, point = <0.004 ng/g
Plums: n = 1, point = <0.004 ng/g
Fruits - others and exotic fruit: grapes: n = 1, point
= <0.004 ng/g
(LOD = 0.004 or 0.010 ng/g)
*n represents number of composite samples
Gebbink et al. (2015)
Sweden
Food items were purchased at two major grocery
store chains in four major Swedish cities in 1999
and 2005. In 2010, sampling was limited to Uppsala
city since no systematic geographical differences in
food contamination was observed in the two earlier
market basket studies. The food items were selected
based on Swedish food and production statistics and
were not cooked before analysis. The food items
were divided into 12 groups and homogenates for
each food group were prepared by mixing food
items proportionally according to food consumption
Fruits and vegetables; meat;
grains; fats/other
1999:
Dairy products: n = 1, point = 0.0007 ng/g
Meat products: n = 1, point = 0.0059 ng/g
Egg: n = 1, point = 0.034 ng/g
Potatoes: n = 1, point = 0.0002 ng/g
Soft drinks: n = 1, mean = 0.0002 ng/g
Fats: n = 1, point = <0.0001 ng/g
D-60
-------
FINAL
2024
Study
Location and Source
Food Types
Results
statistics. Results by food group were not reported
for the 2005 and 2010 years. For all three sampling
years, a homogenate was prepared by mixing
proportional amounts of each food group according
to consumption data for the respective year (includes
fish samples).
Pastries: n = 1, point = <0.0001 ng/g
Cereal products: n = 1, point = <0.0001 ng/g
Vegetables: n = 1, point = <0.0001 ng/g
Fruit: n = 1, point = <0.0001 ng/g
Sugar and sweets: n = 1, point = <0.0001 ng/g
Year pool: n = 12, point = 0.0022 ng/g
2005:
Year pool: n = 12, point = 0.0010 ng/g
2010:
Year pool: n = 12, point = 0.0007 ng/g
(MLOQ = 0.0001 ng/g)
*n represents number of composite samples
Carlsson et al. (2014)
Greenland (Nuuk)
Meat was purchased at the local fish market and
grocery shops in June 2010. All items were
originally caught in the vicinity of the Nuuk area
and/or along the West coast of Greenland and
represented the common food items consumed by
the local Inuit population.
Meat
Seal beef: n = 2, DFa 100%, range = 0.2-0.6 ng/g
WW
Narwhal: n = 6, DF NR, median =
-------
FINAL
2024
Study
Location and Source
Food Types
Results
(Cereals: MLOD, MLOQ = 0.100, 0.333 ng/g;
Juice MLOD, MLOQ = 0.157, 0.522 ng/g;
Milk: MLOD, MLOQ = 0.118, 0.393 ng/g;
Olive oil: MLOD, MLOQ = 0.051, 0.161 ng/g;
Meat: MLOD, MLOQ = 0.121, 0.365 ng/g)
*Artichoke was reported as one of the five cereal
samples in Spain - unclear if this was a typo/error
Eschauzier et al. (2013)
The Netherlands (Amsterdam)
Brewed coffee samples (n = 12) from different
coffee machines were collected from all over the
city. Coffee beans from four of these locations were
collected to manually brew coffee. Post-mixed cola
was collected (n = 4) together with corresponding
tap water and an additional three cola samples from
different parts of town. Sampling was conducted
between February and April 2011 at various
locations (cafes, universities, and supermarkets).
Fats/other
Post-mixed cola: n = 6, DF NR, mean = <0.63 ng/L
Brewed coffee from coffee machines: n = 12, analyte
peak areas could not be quantified due to strong
matrix effects
Manually brewed coffee: n = 4, analyte peak areas
could not be quantified due to strong matrix effects
(LOQ = 0.63 ng/L for cola; 0.95 ng/L for coffee)
Herzke et al. (2013)
Belgium, Czech Republic, Italy, Norway
The Czech Republic, Belgium, Norway, and Italy
were selected to represent eastern, western, northern,
and southern Europe. Sampling took place between
spring 2010 and 2011 as part of the PERFOOD
study. Individual items were randomly selected in
three national retail stores covering different brands
or countries of origin. Of each item, three to ten
single samples were combined to create one pooled
sample per country. The following items were
sampled:
Root vegetables: carrots
Bulb vegetables: onions
Fruiting vegetables: tomatoes, courgettes,
cucumbers, aubergine, peppers
Brassica vegetables: cauliflower, cabbage,
broccoli
Vegetables
Belgium: n = 21, DF NR, mean = 0.00032 ng/g t\v
Czech Republic: n = 16, DF 0%
Italy: n = 15, DF 0%
Norway: n = 17, DF 0%
(MQL = 0.002-0.050 ng/g)
*n represents number of composite samples
* Values below the MQL were substituted with the
MQL value
D-62
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Leaf vegetables: lettuce and other salads,
spinaches, chicory, pre-packed lettuce mix, pre-
packed and minced frozen spinach
Stem vegetables: asparagus, celery, fennel,
cultivated mushrooms
Starchy root tubers: potatoes, prepacked ready-to-
cook pommes frites
Legumes, beans, dried: peas, beans
Hlouskova et al. (2013)
Belgium, Czech Republic, Italy, Norway
Food products were randomly purchased in several
nationwide supermarkets in four European regions
during summer 2010. Within the sampling
campaign, the collection of at least one food item
per subcategory (meat, fish, hen eggs, milk and
dairy products, and butter) in all four countries was
acquired. Food items within each subcategory
included the following:
Meat: beef, canned pork meat, poultry, pork,
pig/bovine liver, rabbit, and/or sheep/lamb
Fish: farmed freshwater fish, farmed marine fish,
and/or seafood)
Hen eggs
Milk and dairy products: ultra-high temperature
whole cow milk, ultra-high temperature
skimmed cow milk, cheese (yellow,
Gouda/Edamer, etc.), and butter
Samples were pooled within a respective food
category but not across food groups.
Pooled milk/dairy products,
meat, fish, hen eggs
n = 50, DF 7%, mean, median (range) = 0.0335,
0.0264 (0.00485-0.0763) ng/g
(MQL = 0.002 ng/g for fish and seafood, meat, hen
eggs, and cheese; 0.001 ng/mL for milk, and 0.006
ng/g for butter)
*n represents number of pooled samples
*Results not reported for individual food groups
Domingo et al. (2012)
Spain (Catalonia)
Foods purchased from 4 shops/stores of each of the
12 representative cities of Catalonia (Barcelona,
l'Hospitalet de Llobregat, Vilanova I la Geltru,
Matarr, Sabadell, Terrassa, Girona, Tarragona, Reus,
Tortosa, Lleida and Manresa) in September 2011.
Shops/stores included local markets, small stores,
Meat; fruits and vegetables;
dairy; grains; fats/other
Meat and meat products: n = 2, DF NR, 0.0032
ng/g t\v
Vegetables: n = 2, DF NR, 0.0045 ng/g fw
Tubers: n = 2, DF NR, mean = <0.0019 ng/g fw
Fruits: n = 2, DF NR, mean = <0.0019 ng/g fw
Eggs: n = 2, DF NR, mean = <0.002 ng/g fw
Milk: n = 2, DF NR, mean = <0.0026 ng/g fw
D-63
-------
FINAL
2024
Study
Location and Source
Food Types
Results
supermarkets, and big grocery stores. Analyzed
samples included 40 items:
Meat and meat products: veal steak, loin of pork,
chicken breast, steak of lamb, boiled ham,
"Frankfurt"-type sausage, cured ham
Vegetables and tubers: lettuce, tomato, potato,
carrot
Fresh fruits: apple, orange, banana
Milk and dairy products: whole and semi-
skimmed milk, yogurt, cheese I - low fat, cheese
II - medium fat, cheese III - extra fat
Cereals: French bread, pasta
Pulses: lentils
Industrial bakery: cookies
Eggs: hen eggs
Oils and fats: olive oil
For each food item, two composite samples were
prepared for analysis, where each composite sample
consisted of 24 individual units. Only edible parts of
each food item were included in the composites.
Dairy products: n = 2, DF NR, mean = <0.0011 ng/g
fw
Cereals: n = 2, DF NR, mean = <0.0006 ng/g fw
Pulses: n = 2, DF NR, mean = <0.0013 ng/g fw
Oils: n = 2, DF NR, mean = <0.0007 ng/g fw
Industrial bakery: n = 2, DF NR, mean = <0.00056
ng/g fw
(LOD not reported)
*n represents number of composite samples
Vestergren et al. (2012)
Sweden (Malmoe, Gothenburg, Uppsala, Sundsvall)
Purchasing locations of the two largest retail chains
in Sweden were selected in each of four major
Swedish cities. All purchases were made in
spring/summer of 1999, 2005, and 2010. In 2010,
the study was limited to the largest retail chains in
Uppsala located in close vicinity to Stockholm. An
equal amount of each food group from each of the
four cities was combined into one sample pool to
provide a representative sample for the Swedish
urban population.
Dairy; meat; grains; fruits and
vegetables; fats/other
Meat products:
1999: n = 1, point = 0.0085 ng/g
2005: n = 1, point = 0.0051 ng/g
2010: n = 1, point = 0.0045 ng/g
(MDL = 0.0019 ng/g; MQL = 0.0058 ng/g)
Egg:
1999: n = 1, point = 0.039 ng/g
2005: n = 1, point =
-------
FINAL
2024
Study
Location and Source
Food Types
Results
(MDL = 0.0019 ng/g; MQL = 0.0058 ng/g)
Dairy products:
1999: n = 1, point =
-------
FINAL
2024
Study
Location and Source
Food Types
Results
(MDL = 0.0010 ng/g; MQL = 0.0030 ng/g)
Soft drinks, lemonade:
1999: n = 1, point = 0.0007 ng/g (estimated)
2005: n = 1, point =
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Noorlander et al. (2011)
The Netherlands
Food products randomly purchased from several
Dutch retail stores with nationwide coverage in
November 2009. Food samples were ground,
homogenized, and pooled for analysis. Food items
within each subcategory included the following:
Flour: whole wheat flour, flour
Pork: sausage, slice of bacon, pork chop, bacon,
minced meat rolled in bacon
Eggs: chicken eggs
Bakery products: cake, almond paste cake,
biscuits, brown spiced biscuit, pie
Vegetables/fruit: apple, orange, grape, banana,
onion, carrot, beet, chicory or leek, tomato,
cucumber, paprika, mushroom, cauliflower,
broccoli, white cabbage, red cabbage, brussel
sprout, spinach, endive, lettuce, French beans
Cheese: gouda cheese, edammer cheese, cheese
(>48% fat, less salt), cheese (>30% fat), brie
cheese
Beef: ground beef, beefburger, stewing steak,
braising steak, minced steak
Chicken/poultry: chicken leg, quarter chicken,
chicken filet, chicken burger, collared chicken
Butter: butter salt-free, salted, low-fat
Milk: half cream milk
Vegetable oil: margarine (solid/fluid), low-fat
margarine, frying fat (vegetable), frying oil
(vegetable), sunflower oil
Industrial oil: low-fat margarine, frying fat
(industrial), frying oil (industrial)
Meat; dairy; fruits and
vegetables; grains; fats/other
Butter: n = 1, point = 0.016 ng/g
Chicken/poultry: n = 1, point = 0.003 ng/g
Bakery products: n = 1, point = 0.006 ng/g
Flour: point = n = 1, 0.018 ng/g
Industrial oil: n = 1, point = 0.007 ng/g
Cheese: n = 1, point = <0.025 ng/g
Milk: n = 1, point = <0.002 ng/g
Eggs: n = 1, point = <0.006 ng/g
Pork: n = 1, point = <0.005 ng/g
Beef: n = 1, point = <0.004 ng/g
Vegetables/fruit: n = 1, point = <0.012 ng/g
Vegetable oil: n = 1, point = <0.002 ng/g
(LOD not reported)
*n represents number of composite samples
D-67
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Jogsten et al. (2009)
Spain (Catalonia)
Food samples purchased from local markets, large
supermarkets, and grocery stores from two different
areas of Tarragona Province, Catalonia in January
and February 2008. The cities of Tarragona and
Reus were sampled in the northern area and
L'Ametlla de Mar and Tortosa in the southern area.
For each food item, two composite samples were
analyzed (one composite for the northern area and
one for the southern area). Each composite was
formed of a minimum of six individual sub-samples
of the same product.
Fruits and vegetables; meat;
fats/other
Raw veal: n = 2, DF NR, mean = <0.003 ng/g
Grilled veal: n = 2, DF NR, mean = <0.001 ng/g
Fried veal: n = 2, DF NR, mean = <0.003 ng/g
Raw pork: n = 2, DF NR, mean = <0.001 ng/g
Grilled pork: n = 2, DF NR, mean = <0.001 ng/g
Fried pork: n = 2, DF NR, mean = <0.002 ng/g
Raw chicken: n = 2, DF NR, mean = <0.001 ng/g
Grilled chicken: n = 2, DF NR, mean = <0.001 ng/g
Fried chicken: n = 2, DF NR, mean = <0.001 ng/g
Fried chicken nuggets (packaged): n = 2, DF NR,
mean = <0.005 ng/g
Black pudding: n = 2, DF NR, mean = <0.008 ng/g
Lamb liver: n = 2, DF NR, mean = <0.250 ng/g
Pate of pork liver (packaged): n = 2, DF NR, mean =
<0.088 ng/g
Foie gras of duck (packaged): n = 2, DF NR, mean =
<0.043 ng/g
"Frankfurt" sausages (packaged): n = 2, DF NR,
mean = <0.006 ng/g
Lettuce: n = 2, DF NR, mean = <0.001 ng/g
Lettuce (packaged): n = 2, DF NR, mean = <0.003
ng/g
Common salt (packaged): n = 2, DF NR, mean =
<0.010 ng/g
(LOD = 0.001 ng/g)
*n represents number of composite samples
*Values of ND were replaced with i^xLOD
Ericson et al. (2008a)
Spain
Food samples purchased from local markets, large
supermarkets, and grocery stores within Tarragona
County in July 2006. Food samples were randomly
purchased with origin source not specified. Each of
the food samples were duplicated and combined to
analyze a composite sample. Composite samples
included the following:
Meat; dairy; fruits and
vegetables; grains; fats/other
Vegetables (n = 2), pulses (n = 2), cereals (n = 2),
pork (n = 2), chicken (n = 2), veal (n = 2), lamb (n =
2), eggs (n = 2), dairy products (n = 2), whole milk
(n = 2), semi-skimmed milk (n = 2), fruits (n = 2),
margarine (n = 2), oil (n = 2): DF 0%
(LOD = 0.001-0.65 ng/g fw)
*n represents number of composite samples
D-68
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Vegetables: lettuce, tomato, green bean, spinach
Pulses: lentils, beans, chickpeas
Cereals: rice, spaghetti, bread
Pork: sausage, hot dogs, steak, hamburger, ham
Chicken: breast, thighs, sausage
Veal: steak, hamburger
Lamb: steak
Dairy products: three different kinds of cheese,
yogurt, "petit-Swiss" creamy yogurt, cream
caramel, custard
Fruits: apple, orange, pear, banana
Oil: olive oil, sunflower oil, corn oil
Fats: margarine
Eggs
Papadopoulou et al. (2017)
Norway
Participants of the A-TEAM project collected a
duplicate portion of all consumed foods and drinks,
prepared as for consumption, over two consecutive
weekdays. Only the samples collected in the first
day were analyzed. Sampling year not reported.
Solid foods: cereals and cereal
products, dairy products (not
milk), fish and seafood, meat
and meat products, sugar and
sugar products, fats and oils,
vegetables and nuts, fruits,
salty snacks, eggs, potatoes;
liquid foods: coffee, tea and
cocoa, milk, water, alcoholic
beverages, soft drinks
Solid foods:
n = 61, DF 66%, median (range) = 0.00088 (0-
0.1) ng/g
(LOQ = 0.0001 lng/g)
Liquid foods:
n = 61, DF 8%, median (range) = 0 (0-0.002)
ng/g
(LOQ = 0.00002 ng/g)
¦"Concentrations
-------
FINAL
2024
Study
Location and Source
Food Types
Results
forbade the use of anti-stick cookware. For each
canteen, lunch meals related to five school days
(from Monday to Friday) were weighed, pooled, and
homogenized. Beverages were not included in the
composites.
Fromme et al. (2007)
Germany (Bavaria)
Thirty-one participants provided 24-hour duplicate
diet samples over seven consecutive days, including
one weekend, in April-October 2005. One
participant only provided samples over four
consecutive days. All diet samples were a normal
mixed diet; no participants were on a special diet or
vegetarian.
Duplicate diet samples
n = 214, DFa 3%, range = 0.05-3.03 ng/g fw
(LOD = 0.1 ng/g fw)
*Values
-------
FINAL
2024
Study
Location and Source
Food Types
Results
sample consisted of a pool of 10-12 eggs from one
Total: n = 36, DFa 61%, range = <0.010-0.128
producer. The pooled samples comprised eggs from
ng/g fw
both conventional and organic production.
Information on the number of organic eggs sampled
1999: n = 3, DFa 100%, mean3 (range) = 0.038
was not available.
(0.019-0.072) ng/g fw
Fresh milk was sampled from the tanks of milk
2000: n = 3, DFa 100%, mean3 (range) = 0.038
transport vehicles between 1999 and 2009 as part of
(0.016-0.069) ng/g fw
the food control program. The tanks generally
contained milk from ten dairy farms. In 2010, milk
2001: n = 3, DFa 100%, mean3 (range) = 0.033
samples were taken from the milk storage tanks on
(0.015-0.051) ng/g fw
individual dairy farms. Between 10-25 milk samples
were collected each year. The milk samples were
2002: n = 3, DFa 67%, range = <0.010-0.128
extracted in two different batches.
ng/g fw
2003: n = 3, DFa 67%, range = <0.010-0.054
ng/g fw
2004: n = 3, DFa 33%, range = <0.010-0.011
ng/g fw
2005: n = 3, DFa 100%, mean3 (range) = 0.015
(0.011-0.018) ng/g fw
2006: n = 3, DF 0%
2007: n = 3, DFa 33%, range = <0.010-0.020
ng/g fw
2008: n = 3, DFa 33%, range = <0.010-0.011
ng/g fw
2009: n = 3, DFa 33%, range = <0.010-0.012
ng/g fw
2010: n = 3, DFa 67%, range = <0.010-0.020
ng/g fw
(MDL = 0.010 ng/g fw; MQL = 0.033 ng/g fw)
Cow's milk (1st batch):
Total: n = 18, DFa22%, range = <0.0010-0.0011
ng/g fw
1999: n= 1,DF0%
D-71
-------
FINAL
2024
Study
Location and Source
Food Types
Results
2000: n = 2, DFa 50%, range = <0.0010-0.0010
ng/g t\v
2001: n = 2,DF0%
2002: n = 2, DFa 50%, range = <0.0010-0.0011
ng/g t\v
2003: n= 1,DF0%
2004: n = 2, DF 0%
2005: n= 1,DF0%
2006: n = 1, point = 0.0010 ng/g fw
2007: n= 1,DF0%
2008: n = 2, DFa 50%, range = <0.0010-0.0011
ng/g fw
2009: n= 1,DF0%
2010: n = 2DF0%
(MDL = 0.0010 ng/g fw; MQL = 0.0033 ng/g fw)
Cow's milk (2nd batch):
Total: n = 18, DF 0%
(MDL = 0.0023 ng/g fw; MQL = 0.0063 ng/g fw)
Eriksson et al. (2013)
Denmark (Faroe Islands)
Locally produced food items sampled in 2011-2012.
Packaged dairy products were supplied by Faroe
Islands, Meginfelag BunaSarmanna - dairy products
included samples of milk, low fat (0.5%), semi-
skimmed (1.5%), yoghurt with banana and pear
(3.4% fat), low fat (0.9%) plain yoghurt, and creme
fraiche (18% fat). Yoghurt with banana and pear
was sampled from two production batches, and the
low fat plain yoghurt and creme fraiche was
sampled from one production batch. Potatoes were
sampled from two different farms.
Dairy; fruits and vegetables
Milk (n = 6), yogurt (n = 3), creme fraiche (n = 1),
potatoes (n = 2): DF 0%
(LOD = 0.0058 ng/L for milk; LOD = 0.0017 ng/g
for dairy; LOD = 0.0016 ng/g for potato)
Vestergren et al. (2013)
Sweden (Karsta)
Meat; dairy
Dairy farm cow's milk: n = 6, DF 0%
D-72
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Study was conducted at a dairy cattle farm that was
selected to represent a background contaminated
agricultural area with no known point sources of
PFAS in the proximity. The farm had no history of
sewage sludge application to the pasture land. Milk
samples were collected between November 2010
and April 2011 from a milk tank, where milk from
the entire farm is stored after milking. Muscle, liver,
and whole blood samples were obtained from five
individual cows from the slaughterhouse on two
different occasions (April and June 2011).
Cow liver: n = 5, DF 0%
Cow blood: n = 5, DF 0%
Cow muscle: n = 5, DF 0%
(MDL not reported)
Falandysz et al. (2006)
Poland
Eider duck samples were collected from the Gulf of
Gdansk in the Baltic Sea (south coast of Poland) in
February 2003.
Meat
n = 16, DF NR, mean, median (range) = 1.1,1.1
(0.4-2.9) ng/mL
(LOD not reported)
*Values reported for animal whole blood samples
Lankova et al. (2013)
Czech Republic
Breastmilk samples were obtained from 50 women
living in the Olomouc region from April to August
2010. The age of participating mothers ranged from
20 to 43 years.
Six types of infant formula from the Czech retail
market were also examined: one powdered formula
for infants, two formulas for toddlers, one formula
for babies with lactose intolerance, one formula for
premature babies, and one soya-based formula for
babies with non-milk diets. Sampling year not
provided.
Fats/other; breastmilk
Breastmilk:
n = 50, DF (frequency of quantification) 8%,
range = <0.006-0.022 ng/mL
(LOQ = 0.006 ng/mL)
Infant formula:
n = 6, DF (frequency of quantification) 0%
(LOQ = 0.005 ng/g)
Abdallah et al. (2020)
Ireland (Dublin)
Breastmilk samples obtained from mothers recruited
from breastfeeding clinics at two Irish maternity
hospitals. Mothers provided samples between 3 and
8 weeks postpartum. Mothers were up to 41 years of
age, primiparas, in good health, and exclusively
feeding one infant. Sampling year not reported.
Breastmilk
n = 16, DF 31%, mean, median (range) = <0.04,
<0.04 (<0.04-0.087 ng/mL)
(LOQ = 0.04 ng/mL)
*Values
-------
FINAL
2024
Study
Location and Source
Food Types
Results
native Swedish mothers, who were predominately
non-smokers and primiparous. There were a total of
20 pooled samples analyzed from Stockholm (1972-
2016), containing 9-116 individual samples per
pool, and 11 pooled samples from Gothenburg
(2007-2015), containing 5-11 individuals per pool.
In addition, samples collected in 2012 (16 from
Gothenburg and 20 from Stockholm) and in 2016
(10 from Stockholm) were analyzed individually.
Stockholm (pooled): n = 20, DFa 90%, range =
<0.0008-0.021 ng/mL
Gothenburg (pooled): n = 11, DFa 91%, range =
<0.0008-0.012 ng/mL
Stockholm (2012, individual): n = 20, DFa 95%,
range = <0.0008-0.025 ng/mL
Gothenburg (2012, individual): n = 16, DFa
94%, range = <0.0008-0.014 ng/mL
Stockholm (2016, individual): n = 10, DFa 80%,
range = <0.0008-0.013 ng/mL
Br-PFHxS:
Stockholm (pooled): n = 20, DFa 25%, range =
<0.0008-0.0029 ng/mL
Gothenburg (pooled): n = 11, DFa 18%, range =
<0.0008-0.004 ng/mL
Stockholm (2012, individual): n = 20, DF 0%
Gothenburg (2012, individual): n = 16, DFa 6%,
range = <0.0008-0.0012 ng/mL
Stockholm (2016, individual): n = 10, DFa 30%,
range = <0.0008-0.004 ng/mL
(MDL = 0.0008 ng/mL)
Cariou et al. (2015)
France (Toulouse)
Breastmilk samples obtained from female volunteers
hospitalized between June 2010 and January 2013
for planned caesarean delivery. Samples were
collected between the fourth and fifth day after
delivery.
Breastmilk
n = 61, DF 15%, mean, median (range) = 0.026,
-------
FINAL
2024
Study
Location and Source
Food Types
Results
with instructions on breastmilk collection. Milk
samples could be collected during several lactation
sessions. On average, 15 aliquot samples of 10 mL
were collected for each participant and pooled into
one sample for analysis.
Croes et al. (2012)
Belgium (Flanders)
Breastfeeding mothers were recruited from 9
maternities in 24 rural communities in East and
West Flanders and Flemish Brabant in May 2009 -
June 2010. Breastmilk samples were collected
between two and eight weeks after delivery and a
subset was analyzed for perfluorinated compounds.
Breastmilk
n = 40, DF 20%, median (10th-90th percentile) =
-------
FINAL
2024
Study
Location and Source
Food Types
Results
*n represents number of pooled samples
Karrman et al. (2010)
Spain (Catalonia)
Breastmilk samples were collected from healthy
primiparae mothers aged 30-39 years who lived in
Tarragona County for at least the last five years.
Babies were aged 41-60 days when milk samples
were collected in 2007.
Breastmilk
n = 10, DF 100%, mean, median (range) = 0.04,
0.04 (0.02-0.11) ng/mL
(LOQ = 0.01 ng/mL)
Karrman et al. (2007)
Sweden (Uppsala, Goteborg, Lund, Lycksele)
Individual breastmilk samples from 12 women in
Uppsala, Sweden were collected in 2004.
Composite samples were created from breastmilk
samples collected from 25-90 women each year
between 1996 and 2004 and pooled into an annual
composite sample. Donors originated from four
regions in Sweden (Uppsala 1996 -2000, 2002;
Goteborg 2001; Lund 2003; Lycksele 2003-2004).
All samples were collected from primiparous
women (19^11 years old) during the third week after
delivery.
Breastmilk
Individual samples:
2004: n = 12, DFa 100%, mean, median (range)
= 0.085, 0.070 (0.031-0.172) ng/mL
Composite samples:
1996: n = 1, point = 0.037 ng/mL
1997: n = 1, point = 0.030 ng/mL
1998: n = 1, point = 0.040 ng/mL
1999: n = 1, point = 0.044 ng/mL
2000: n = 1, point = 0.028 ng/mL
2001: n = 1, point = 0.028 ng/mL
2002: n = 1, point = 0.051 ng/mL
2003: n = 1, point = 0.025 ng/mL
2003-2004: n = 1, point = 0.016 ng/mL
*n represents number of composite samples
(DL = 0.01 ng/mL)
Beseretal. (2019)
Spain (Valencian region)
Breastmilk samples were collected from 14 Spanish
women (aged 30-39 years) living in the Valencian
region and recruited by the perinatology group of
the Health Research Institute La Fe in Valencia.
Milk samples were collected at different stages after
birth during 2015.
Breastmilk
n = 20, DFa 30%
*Six samples were >MDL but
-------
FINAL
2024
Study
Location and Source
Food Types
Results
Pooled breastmilk samples were collected from 109
first-time mothers at four centers across Ireland.
Sampling year not reported.
*n represents number of pooled samples
Notes: AFFF = aqueous film-forming foam; DF = detection frequency; GCA = groundwater contamination area; LOD = limit of detection; LOQ = limit of quantitation;
MAMA = Methods Advancement for Milk Analysis; MDL = method detection limit; ND = not detected; NR = not reported; RL = reporting limit; ww = wet weight;
WWTP = wastewater treatment plant.
Bold indicates detected levels of PFHxS in food.
a The DF and/or mean was not reported in the study and was calculated in this synthesis. Means were calculated only when DF = 100%.
D-77
-------
FINAL
2024
D.3.2. Food Contact Materials
No studies were identified that evaluated the occurrence of PFHxS in food packaging or food
contact materials (FCMs) purchased in the United States. In an analysis performed at the
Department of Food Analysis and Nutrition of the University of Chemistry and Technology in
Prague, Czech Republic, PFHxS was not detected in 42 samples of disposable food packaging
and tableware purchased from six different European countries between May and December
2020 (LOQ = 1.7 mg/kg) (Strakova et al., 2021). The five additional peer-reviewed European
studies identified are summarized below and in Table D-6 (Vavrous et al., 2016; Kotthoff et al.,
2015; Surma et al., 2015; Vestergren et al., 2015; Zafeiraki et al., 2014). Three of these studies
reported no detection of PFHxS in FCMs while two reported detectable levels of PFHxS in
FCMs. Of these two studies, PFHxS was detected in 6% of paper-based FCM samples purchased
recently in Germany (at the time of the study: 2010) and also was detected in samples purchased
before 2010, but in both cases the median concentration was below the LOQ of 0.5 ng/g. For
FCMs purchased in Poland, PFHxS was detected in one brand of cellulose wrapping paper (0.29
pg/cm2) but was below the LOD (0.01 pg/cm2) or below the LOQ (0.03 pg/cm2) in other
cellulose and polyether sulfone FCMs. Additional research is needed to evaluate PFHxS in
FCMs purchased in U.S. and Canada and for FCMs with different countries of origin.
Two studies reported PFHxS in food contact materials (Kotthoff et al., 2015; Surma et al., 2015).
Kotthoff et al. (2015) analyzed for PFSA and PFCA compounds in random samples of recent
(purchased in 2010) individual paper-based FCMs (n = 33) from local retailers in Germany.
Samples were purchased from local retailers or collected by co-workers of the institute in the
first until third quarter of 2010. Baking paper purchased before 2010 (sample age ranging from a
few years to decades) was collected from staff of the institutes and referred to as archived
samples. PFHxS was detected in 6% of recent paper-based FCM samples with a median
concentration below the LOQ (0.5 ng/g). PFHxS was also detected in archived FCM samples
with a median concentration below the LOQ. In Surma et al. (2015), the authors measured levels
of PFHxS in three different brands of FCMs that included wrapping papers (n = 3), breakfast
bags (n = 3), baking papers (n = 3), and roasting bags (n = 3); sampling year was not reported.
Items were obtained from typical, commercially available food contact products in Poland.
Roasting bags were made of polyether sulfone; the remaining items were made of cellulose.
PFHxS was detected in one brand of wrapping paper (brand B) at 0.29 pg/cm2, but PFHxS was
below the LOD (0.01 pg/cm2) or below the LOQ (0.03 pg/cm2) in all other FCMs. The authors
reported that the highest content of perfluorinated compounds were reported for B brand FCM.
They also reported that FCMs based on cellulose contained more PFCAs than PFASs; on the
other hand, FCMs based on polyether sulfone contained more PFASs than PFCAs.
The remaining three studies did not detect PFHxS in FCMs (Vavrous et al., 2016; Vestergren et
al., 2015; Zafeiraki et al., 2014). Vavrous et al. (2016) analyzed 15 samples of paper FCM (11
with direct food contact and 4 with indirect food contact) acquired from a market in the Czech
Republic, including paper packages of wheat flour (n = 2), paper bags for bakery products (n =
2), sheets of paper for food packaging in food stores (n = 2), cardboard boxes for packaging of
various foodstuffs (n = 3), coated bakery release papers for oven baking at temperatures up to
220°C (n = 3), and paper filters for coffee preparation (n = 3). PFHxS was below the LOQ
(0.0030 mg/kg) in all samples. In Vestergren et al. (2015), the authors analyzed a random sample
of FCMs collected in November 2012, including a baking mold (n = 1), baking cover (n = 1),
D-78
-------
FINAL
2024
paper cup (n = 1), and paper plates (n = 2) that were purchased from major retail stores in
Norway but were imported from China. PFHxS was below the MDL (0.01 |ig/m2) in all samples.
Finally, Zafeiraki et al. (2014) analyzed 42 samples of FCMs made of paper, paperboard, or
aluminum foil randomly obtained from retailers. All products except for microwave popcorn and
rice bags were manufactured in Greece. Sampled packaging materials included unused items and
used items (i.e., contained food products). Beverage and ice cream cups, wrappers, and paper
boxes were collected in Athens from October to December 2012 from popular Greek fast food
chain restaurants, coffee shops, and multiplex cinemas. Other FCMs (muffin cups, baking
papers, and microwave popcorn and rice bags) were purchased from large supermarkets. PFHxS
was below the LOD (0.18 ng/g) in all samples.
D-79
-------
FINAL
2024
Table D-6. Studies Reporting PFHxS Occurrence in Food Contact Materials
Study
Location
Site Details
Results
Europe
Kotthoff et al. (2015)
Germany (Schmallenberg)
Thirty-three random samples of recent individual
paper-based FCMs collected in the first until the third
quarter of 2010 in Germany. Individual samples were
bought from local retailers or collected by coworkers
of the involved institutes. Sampled products spanned
all quality levels from entry level to cutting edge
products. The age of the samples ranged from a few
years to decades. Country of origin not reported.
"Archived" older samples of FCMs (baking paper
purchased before 2010) were collected from the staff
of the institutes. The age of these samples ranged from
a few years to decade. Country of origin not reported.
Recent samples: n = 33, DF 6%, median
(range) =
-------
FINAL
2024
Study
Location
Site Details
Results
= 3), and paper filters for coffee preparation (n = 3).
Sampling year and country of origin for products not
reported.
Vestergren et al. (2015)
Norway (Tromse, Trondheim)
Five samples of FCMs (one baking mold, two paper
plates, one baking cover, and one paper cup) were
purchased from major retail stores in November 2012.
Sampling campaign designed to evaluate consumer
products in product categories that were previously
found to contain PFAS residuals and that were
representative of products imported from China in
large quantities. Individual products randomly selected
without prior knowledge of surface treatment with
PFAS. Year of manufacture not reported.
n = 5, DFa 0%
(MDL = 0.010 (ig/m2)
Zafeiraki et al. (2014)
Greece
Forty-two samples of FCMs made of paper,
paperboard, or aluminum foil were obtained randomly
from retailers. Their exact composition was not stated
and there was no information about
perfluorochemicals used in their manufacturing
process or not. Beverage and ice cream cups,
wrappers, and paper boxes were collected in Athens
from October to December 2012 from popular Greek
fast food chain restaurants, coffee shops, and
multiplex cinemas. Other FCMs (muffin cups, baking
papers, and microwave popcorn and rice bags) were
purchased from large supermarkets. All products
except for microwave popcorn and rice bags were
manufactured in Greece. Sampled packaging materials
included unused items and used items (i.e., contained
food products).
A microwave popcorn bag was also analyzed before
and after cooking.
Beverage cups: n = 8, DF 0%
Ice cream cup: n = 1, DF 0%
Fast food paper boxes: n = 8, DF 0%
Fast food wrappers: n = 6, DF 0%
Paper materials for baking: n = 2, DF 0%
Microwave bags: n = 3, DF 0%
Before cooking: n = 1, point =
-------
FINAL
2024
D.3.3. Consumer Products
PFHxS has been used in laboratory applications and as a raw material or a precursor for the
manufacture of PFAS/perfluoroalkyl sulfonate-based products, though production of PFHxS in
the United States was phased out by its major manufacturer in 2002 (NCBI, 2022a; Sigma-
Aldrich, 2014; Backe et al., 2013; Buck et al., 2011; OECD, 2011). PFHxS has also been used in
firefighting foam and carpet treatment solutions, and it has been used as a stain and water
repellant (NCBI, 2022a; Garcia and Harbison, 2015). PFHxS has been detected in aqueous film-
forming foam, aftermarket carpet protection products, chipboards, leather, membranes for
apparel, treated apparel, and photoprint ink and laser ink (NCBI, 2022a; Gliige et al., 2020;
Norwegian Environment Agency, 2018; Becanova et al., 2016; Kotthoff et al., 2015; Liu et al.,
2014; Backe et al., 2013; Buck et al., 2011; Herzke et al., 2009).
Two U.S.-based studies were identified that analyzed PFHxS concentrations in a range of
consumer products, including children's nap mats, household carpet/fabric-care liquids, and
textiles (Zheng et al., 2020; Liu et al., 2014) (Table D-7). Few U.S. studies analyzed children's
products, fabric treatments, treated fabrics, sealants, and similar products, and none of the U.S.
studies reviewed sampled for PFAS in other household products and articles, such as cosmetics,
cleaners, paints, upholstered furniture, etc. Of the U.S. studies, the majority of the consumer
products evaluated are likely used by adults (e.g., floor waxes), can come into contact with both
adults and children (e.g., treated upholstery), or the user was not specified (e.g., clothing).
Zheng et al. (2020) determined the occurrence of ionic and neutral PFAS in the childcare
environment (dust and nap mats). Samples of children's nap mats were collected from seven
Seattle childcare centers (n = 26; 20 polyurethane foam, 6 vinyl cover samples). PFHxS was
detected in 73% of nap mat samples with a mean concentration of 0.32 ng/g. Half of the
analyzed mats were purchased as new products and the other half were used. The authors
reported that total PFAS levels in the new versus used mats were not significantly different.
Total PFAS levels in mat foam versus mat covers were also similar. Based on these results, the
authors suggested that indoor air was not the major source of PFAS in mats and that PFAS in
mats could be associated with the manufacturing process.
Liu et al. (2014) analyzed the occurrence of PFAS in consumer products (including commercial
carpet/fabric-care liquids, household carpet/fabric-care liquids, treated apparel, treated home
textile and upholstery, treated floor waxes and stone-wood sealants, membranes for apparel, and
thread-sealant tapes and pastes) purchased between March 2007 and September 2011 from local
retailers and online stores in the United States. PFHxS was analyzed in a subset of these
consumer products, originating from the United States, England, Dominican Republic, Vietnam,
and China, and was detected in one out of two commercial carpet/fabric-care liquids samples at
194 ng/g, in two out of four household carpet/fabric-care liquids and foams samples at 88.8 ng/g
and 155 ng/g, in one out of two treated children's apparel samples at 1.70 ng/g (in boy's uniform
pants), in one out of two treated home textile and upholstery samples at 12.1 ng/g, in one apparel
membrane sample at 7.10 ng/g, and in one out of two thread-sealant tapes and pastes samples at
60.3 ng/g. PFHxS was not detected in one treated floor wax and stone/wood sealant sample.
Detection limits were not reported in the study.
D-82
-------
FINAL
2024
Beesoon et al. (2012) detected PFHxS in all cleaning formulation-treated carpet samples (n = 9)
from various rooms of a Canadian household whose members had previously been identified as
having disproportionately high blood serum levels of PFHxS. The Scotchgard formulation-
treated carpet samples had a mean (range) of 512 (12-2,880) ng/g of PFHxS, with a level of 17
ng/g detected in one untreated, uninstalled carpet sample taken from the same house.
Of the European studies, van der Veen et al. (2020) examined the effects of weathering on PFAS
content in durable water repellent clothing collected from six suppliers in Sweden. Two pieces of
each of the 13 fabrics were cut. One piece was exposed to elevated ultraviolet radiation,
humidity, and temperature in an aging device for 300 hours (assumed lifespan of outdoor
clothing); the other was not aged. Pieces of textile, aged and not aged, were analyzed for ionic
PFAS (including PFHxS) and volatile PFAS. For 10 of 13 fabrics, PFHxS was not detected
before or after aging. For one fabric, PFHxS was detected before and after aging, increasing from
0.11 to 0.68 |ig/m2. For one fabric, PFHxS was detected prior to aging at 0.89 |ig/m2but was not
detected afterward. For the remaining fabric, PFHxS was not detected prior to aging but was not
analyzed after aging.
Kotthoff et al. (2015) analyzed 82 samples of consumer products collected in Germany,
including outdoor textiles, carpets, cleaning agents, impregnating agents, leather samples, and ski
waxes. Individual samples were bought from local retailers or collected by coworkers of the
involved institutes or local clubs in Germany. The age of the samples ranged from a few years to
decades. PFHxS was detected in 35% of ski wax samples (n = 13) up to 9.3 ng/g and in 96% of
leather samples (n = 13) up to 10.1 |ig/m2. PFHxS was not detected in cleaning agents (n = 6),
wood glue (n = 1), impregnating sprays (n = 3), outdoor textiles (n = 3), carpet (n = 6), gloves (n
= 3), or awning cloth (n = 1). Favreau et al. (2016) analyzed the occurrence of 41 PFASs in a
wide variety of liquid products (n = 194), including impregnating agents, lubricants, cleansers,
polishes, AFFFs, and other industrial products purchased from stores and supermarkets in
Switzerland. PFHxS was not detected in impregnation products (n = 60), cleansers (n = 24), or
polishes (n = 18). PFHxS was detected in 4% of a miscellaneous category of products (n = 23)
that included foam-suppressing agents for the chromium industry, paints, ski wax, inks, and
tanning substances, with a maximum concentration of 1,700 ng/g.
The remaining two European studies did not detect PFHxS in the consumer products analyzed.
Vestergren et al. (2015) analyzed furniture textile, carpet, and clothing samples (n = 40)
purchased from major retail stores in Troms0 and Trondheim, Norway. All samples represented
materials that had been imported from China. PFHxS was not detected in any of the samples.
Schultes et al. (2018) determined levels of 39 PFAS in 31 cosmetic products collected in
Sweden. The study found that 16 out of 31 samples contained measurable concentrations of at
least one PFAS; however, PFHxS was not detected amongst the samples.
Of the two studies where the location of purchase was not specified, Gremmel et al. (2016)
determined levels of 23 PFAS in 16 new outdoor jackets. PFHxS was detected in one jacket at a
concentration of 0.01 ng/g. Becanova et al. (2016) analyzed 126 samples of (1) household
equipment (textiles, floor coverings, electrical and electronic equipment (EEE), and plastics); (2)
building materials (oriented strand board, other composite wood and wood, insulation materials,
mounting and sealant foam, facade materials, polystyrene, air conditioner components); (3) car
interior materials; and (4) wastes of electrical and electronic equipment (WEEE) for 15 target
PFAS, including PFHxS. The condition (new versus used) and production year of the samples
D-83
-------
FINAL 2024
varied; the production year ranged from 1981 to 2010. PFHxS was detected in 42%, 22%, 30%,
and 14% of household equipment, building materials, car interior, and WEEE samples,
respectively. The highest level was 24.5 ng/g found in a drywall sample.
D-84
-------
FINAL
2024
Table D-7. Summary of PFHxS Consumer Product Data
Study
Location
Site Details
Results
United States
Zheng et al. (2020)
United States (Seattle,
Washington)
Children's nap mat samples (n = 26, finely cut) from
seven Seattle childcare centers, including
polyurethane foam (n = 20) and vinyl cover (n = 6)
samples. Sampling year not reported.
n = 26, DF 73%, mean, median (range) =
0.32, 0.30 (
-------
FINAL
2024
Study
Location
Site Details
Results
room. The last application was in 2007, with six
intermittent applications. The house did not have a
fan-forced air circulation system and there was also
no fresh-air intake to the house from the outdoors.
Two additional pieces of the same carpet (untreated,
uninstalled) stored in the basement for many years
were also evaluated.
(LOD/LOQ not reported)
Europe
van der Veen et al. (2020)
Sweden (unspecified)
Samples of durable water repellent outdoor clothing
collected from six suppliers from the outdoor textile
industry in Sweden (one pair of outdoor trousers, six
jackets, and six fabrics for outdoor clothes*). Each
sample was cut into two pieces - one exposed to
elevated UV radiation, humidity, and temperature for
300 hours (assumed lifespan of outdoor clothing) and
one untreated (not aged). Sampling year not reported.
Year of manufacturing not reported for nine of the 13
samples; the remaining four samples (samples 4-7)
reported a manufacturing year of 2012/2013. Country
of origin not reported.
*The breakdown of the 13 items of outdoor clothing
is reported differently in Section 2.2and Table 1 of
the article. Section 2.2 reports one pair of outdoor
trousers, seven jackets, four fabrics for outdoor
clothes, and one outdoor overall. Table 1 shows one
pair of outdoor trousers, six jackets, and six fabrics
for outdoor clothes.
Values presented as not aged, aged for L-
PFHxS.
Samples 1, 3,4, 6-12 (1 outdoor trousers, 4
jackets, and 5 fabrics for outdoor clothes):
n = 1 (for each sample), ND, ND
Sample 2 (fabric for jacket): n = 1, 0.89
(ig/m2, ND
Sample 5 (men's jacket): n = 1, 0.11, 0.68
Hg/m2
Sample 13 (fabric for outdoor clothes): n =
1, ND, NA
(LOD = 0.02-0.1 (ig/m2 for ionic PFAS)
Schultes et al. (2018)
Sweden (unspecified)
Thirty-one cosmetic products from five product
categories (moisturizing cream, foundation, eye
pencil, powder and eye shadow, shaving foam)
purchased from the Swedish market in 2016-2017.
Cosmetic products were selected based on (i) the
2015 KEMI survey which reported the most
frequently reported PFAS in cosmetic products and
(ii) a database of ingredient lists compiled by the
Swedish Society for Nature Conservation. Twenty-
four products listing nine different PFAS as active
ingredients were purchased. In addition, seven
products which did not list PFAS in their ingredients
were also purchased from the same stores as control
Control: n = 7, DF 0%
PFAS-containing: n = 24, DF 0%
(LOD = not reported)
D-86
-------
FINAL
2024
Study
Location
Site Details
Results
samples. Year of manufacture and country of origin
not reported.
Favreau et al. (2016)
Switzerland (national)
Liquid consumer products, including impregnation
agents, cleansers, polishes, lubricants, miscellaneous
items, and commercial AFFFs purchased in 2012 and
2013 from stores and supermarkets throughout
Switzerland. Products were purchased from 82
different producers and were selected based on their
susceptibility to contain PFAS according to previous
screenings. Miscellaneous "other" products included
foam-suppressing agents for the chromium industry,
paints, ski wax, inks, and tanning substances. AFFFs
were divided into two sets based on the sampling
source. AFFF set 1 was derived from stock solution
in fire installation of industrial sites storing chemicals
and petroleum products and samples may be the
result of multiple AFFF fillings over the years (1990—
2010 was the last documented filling date). AFFF set
2 came from commercially available AFFFs between
2012 and 2013 from six producers. Results reported
for L-PFHxS.
Impregnation products: n = 60, DF 0%
Cleansers: n = 24, DF 0%
Polishes: n = 18, DF 0%
Others: n = 23, DF 4%, mean (range) = 1,700
(1,700-1,700) ng/g
AFFF set 1: n = 27, DF 81%, mean, median
(range) = 293,000, 89,500 (100-1,025,000)
ng/g
AFFF set 2: n = 35, DF 0%
(LOQ = 0.5 ng/mL)
*Mean and range values only include samples
where L-PFHxS was detected
*ND treated as 0 for median calculations
Kotthoff et al. (2015)
Germany (Schmallenberg)
Forty-nine random samples of consumer products
collected in the first until the third quarter of 2010 in
Germany, including outdoor textiles, carpets,
cleaning agents, impregnating agents, leather
samples, and ski waxes. Individual samples were
bought from local retailers or collected by coworkers
of the involved institutes or local clubs (e.g., ski
waxes from local skiing club). Sampled products
spanned all quality levels from entry level to cutting
edge products. The age of the samples ranged from a
few years to decades. Country of origin not reported.
Cleaner: n = 6, DF 0%
Wood glue: n = 1, DF 0%
Nanosprays and impregnation sprays: n = 3,
DF 0%
Outdoor textiles: n = 3, DF 0%
Carpet: n = 6, DF 0%
Gloves: n = 3, DF 0%
Ski wax: n = 13, DF 35%, median
(maximum) =
-------
FINAL
2024
Study
Location
Site Details
Results
purchased from three major retail stores during
November 2012-February 2013. Sampling campaign
designed to evaluate consumer products in product
categories that were previously found to contain
PFAS residuals and that were representative of
products imported from China in large quantities.
Individual products randomly selected without prior
knowledge of surface treatment with PFAS. Outdoor
clothing was excluded. Year of manufacture not
reported.
Cotton/Leather clothing: n = 4, DFa 0%
(MDL = 0.010 (ig/m2)
Orijjin Unspecified
Becanova et al. (2016)
Not specified
One hundred twenty-six samples of (1) household
equipment (textiles, floor coverings, electrical and
electronic equipment (EEE), and plastics; includes
children-related items such as teddy bear filling,
teddy bear cover, and plush); (2) building materials
(oriented strand board, other composite wood and
wood, insulation materials, mounting and sealant
foam, facade materials, polystyrene, air conditioner
components); (3) car interior materials; and (4)
wastes of electrical and electronic equipment
(WEEE) purchased (for new materials) or collected
from various sources (for older and used materials).
Production year ranged from 1981 to 2010. Origin of
production and location and year of
purchase/collection not reported.
Household equipment: n = 55, DFa 42%,
range =
-------
FINAL
2024
D.3.4. Indoor Dust
Several studies from the U.S. and abroad were identified that evaluated the occurrence of PFHxS
and other PFAS in dust of indoor environments, primarily in homes, as well as in schools,
childcare facilities, offices, and vehicles. In a Wisconsin Department of Health Services study,
Knobeloch et al. (2012) examined levels of 16 perfluoroalkyl chemicals in vacuum cleaner dust
from 39 Wisconsin homes across 16 counties in March and April 2008 (Table D-8). Samples
from these homes built between 1890 and 2005 were collected during a pilot study to assess
residential exposure to persistent contaminants found in the Great Lakes Basin. PFHxS was
found in all samples at a median concentration of 16 ng/g. Mean levels of PFHxS in dust were
significantly higher in homes built between 1968 and 1995 (219 ng/g vs. 57 ng/g in homes
constructed in other years). Based on the results of this study, the authors suggested that
perfluoroalkyl chemicals may be ubiquitous contaminants in U.S. homes. In an EPA study of
112 indoor dust samples collected from vacuum cleaner bags from homes and daycare centers in
North Carolina and Ohio in 2000-2001 (EPA's CTEPP study), samples were collected from 102
homes and 10 daycare centers in North Carolina (49 homes, 5 daycare centers) and Ohio (53
homes, 5 daycare centers) (Strynar and Lindstrom, 2008). Results were not reported separately
for homes and daycares. Overall, PFHxS was detected in 77.7% of all samples (n = 112) at mean
and median concentrations of 874 and 45.5 ng/g, respectively. The authors concluded that the
study measured perfluorinated compounds in house dust at levels that may represent an
important pathway for human exposure.
Additional peer-reviewed studies were identified that evaluated the occurrence of PFHxS and
other PFAS in dust of indoor environments, primarily in homes, as well as in schools, childcare
facilities, offices, and vehicles (Zheng et al., 2020; Scher et al., 2019; Byrne et al., 2017;
Karaskova et al., 2016; Wu et al., 2014; Fraser et al., 2013; Knobeloch et al., 2012; Kato et al.,
2009) (Table D-8). For those studies with results stratified for U.S. homes, PFHxS levels and
detection frequencies were lowest in a study of remote Alaska Native villages (27% detection,
median below 0.2 ng/g), while in other U.S. locations, PFHxS was detected in at least 40% of
samples (some studies reporting 100% detection) at widely varying mean and median levels
across the studies (from on the order of 10 ng/g to on the order of 200 ng/g) with one study
reporting the highest mean value (219 ng/g) from homes built between 1968 and 1995. The two
studies also reporting home measurements from other countries differed in how PFHxS levels in
the United States ranked relative to other countries, with one study ranking the U.S. highest and
the other second lowest. Few studies sampled childcare centers, vehicles, and offices, and none
of the reviewed studies reported measurements in other microenvironments (e.g., public libraries,
universities).
Several studies reported results from dust samples collected only from homes (Scher et al., 2019;
Byrne et al., 2017); Wu et al. (2014), with one study sampling from locations near a PFAS
production facility. Scher et al. (2019) evaluated indoor dust in 19 homes in Minnesota within a
GCA in the vicinity of a former 3M PFAS production facility. Homes within the GCA had
previous or ongoing PFAS contamination in drinking water and were served by the Oakdale,
Minnesota public water system or a private well previously tested and shown to have detectable
levels of PFOA or PFOS. In the house dust samples, collected from July to September 2010, the
detection frequencies for PFHxS were 68% and 84% for entryways to the yard and interior living
spaces such as the family or living rooms, respectively (n = 19 each), with median concentrations
D-89
-------
FINAL
2024
of 8.2 ng/g and 18 ng/g, respectively. PFAS concentrations in both sampling locations were
higher than corresponding soil concentrations, suggesting that interior sources were the main
contributors to PFAS in house dust.
Byrne et al. (2017) assessed exposure to PFHxS and other PFAS among residents of two remote
Alaska Native villages on St. Lawrence Island. PFAS concentrations were measured in dust
collected from the surfaces of floors and furniture of 49 homes on St. Lawrence Island during
February-April of 2013 and 2014. Residents were asked not to sweep or dust for one week prior
to sampling. The authors described the overall PFAS levels in dust samples as "on the lower end
of those reported worldwide in other studies." PFHxS was found in 27% of all samples (n = 49)
with a median value below the LOD (0.1 ng/g-0.2 ng/g). Wu et al. (2014) measured
concentrations of five PFAS in residential dust in California in 2008-2009. Dust samples were
collected from the carpet or area rug in the main living area of the home. Homes of parents with
young children and homes with older adults were differentiated to characterize the relationship
between serum concentrations of PFAS and several other factors, including PFAS concentrations
in residential dust. PFHxS was detected in 51% of samples from households with young children
in Northern California (n = 82), with mean and median concentrations of 142 ng/g and 5.30 ng/g,
respectively. PFHxS was detected in 52% of samples from households of older adults in central
California (n = 42), with mean and median concentrations of 55 ng/g and 5.55 ng/g, respectively.
Apart from the information reported by Strynar and Lindstrom (2008), one other study included
childcare centers in the locations sampled (Zheng et al., 2020). Zheng et al. (2020) collected dust
samples from seven childcare centers in Seattle, Washington (n = 14) and one childcare facility
in West Lafayette, Indiana (n = 6 across six rooms); the sampling year was not reported. The
included childcare facilities consisted of several building types, including multiple classrooms, a
former church, and a former home. Because centers were vacuumed and mopped daily, dust
samples were obtained from elevated surfaces (shelving, tops of bookcases/storage cubbies)
along with floor dust. PFHxS was detected in 95% of samples at mean and median
concentrations of 0.34 ng/g and 0.25 ng/g, respectively.
One study evaluated PFHxS levels in vehicles and offices, in addition to homes. Fraser et al.
(2013) collected dust samples between January and March 2009 from three microenvironments
of 31 individuals in Boston, Massachusetts (offices (n = 31), homes (n = 30), and vehicles with
sufficient dust for analysis (n = 13)). Study participants worked in separate offices located across
seven buildings, which were categorized as Building A (n = 6), Building B (n = 17), or Other
(n = 8). Building A was a newly constructed (approximately one year prior to study initiation)
building with new carpeting and new upholstered furniture in each office. Building B was a
partially renovated (approximately one year prior to study initiation) building with new carpeting
throughout hallways and in about 10% of offices. The Other buildings had no known recent
renovation occurred. Study offices were not vacuumed during the sampling week and
participants were asked not to dust or vacuum their homes and vehicles for at least one week
prior to home sampling. Because PFHxS was detected in less than 50% of samples in all three
microenvironments, geometric means were not reported. The detection frequencies for PFHxS
were 23%, 40%, and 46% for offices, homes, and vehicles, respectively, with the range of
detected values reported as 5.24 ng/g-18.5 ng/g, 6.05 ng/g-430 ng/g, and 5.22 ng/g-108 ng/g,
respectively.
D-90
-------
FINAL
2024
Two studies evaluated dust samples collected across multiple continents (Karaskova et al., 2016;
Kato et al., 2009). Karaskova et al. (2016) examined PFAS levels in house dust collected
between April and August 2013 from the living rooms and bedrooms of 14 homes in the United
States, 15 homes in Canada, and 12 homes in the Czech Republic (locations unspecified). PFHxS
was detected in all U.S. samples (n = 20) at mean and median concentrations of 13.8 ng/g and
8.7 ng/g, respectively. The authors reported significant differences between countries for PFHxS
concentrations, with a trend of U.S. > Canada ~ Czech Republic and suggested that the
differences may be explained by differences in the market, import history, and usage of these
substances. Overall, no significant differences in total PFAS concentrations were found between
the bedroom and living room in the same household although significant relationships were
found based on type of floors, number of residents, and age of the house. A second
multicontinental study (Kato et al., 2009) measured PFC concentrations in 39 household dust
samples collected in 2004 from homes in the United States (Atlanta, GA) (n = 10), United
Kingdom (n = 9), Germany (n = 10), and Australia (n = 10). Across all 39 homes, PFHxS was
detected in 79.5% of samples with a median concentration of 185.5 ng/g. Authors presented the
median and maximum PFHxS concentrations by country in a bar chart, which showed that
PFHxS was detected in all countries. The median and maximum PFHxS concentrations for the
10 United States (Atlanta, GA) house dust samples were approximately 96.4 ng/g and
231.3 ng/g, respectively. The highest median was found in Australia, followed by the United
Kingdom, the United States, and Germany in decreasing order; statistical significance on the
comparison of median PFHxS concentrations by country was not reported.
In general, PFHxS concentrations in dust were higher in North America than Europe, with five
studies in the United States or Canada reporting maximum concentrations >1,000 ng/g in homes
and daycares (Wu et al., 2014; Beesoon et al., 2012; Knobeloch et al., 2012; Strynar and
Lindstrom, 2008; Kubwabo et al., 2005) compared to one study from Europe for an office
storage room (Huber et al., 2011). An additional study (Kato et al., 2009) measured PFHxS
concentrations in both continents and reported higher maximum concentrations in the United
States compared to Germany and the United Kingdom. One European study, conducted by Haug
et al. (2011), indicated that house dust concentrations are likely influenced by a number of
factors related to the building (e.g., size, age, floor space, flooring type, ventilation); the
residents or occupants (e.g., number of people, housekeeping practices, consumer habits such as
buying new or used products); and the presence and use of certain products (e.g., carpeting,
carpet or furniture stain-protective coatings, waterproofing sprays, cleaning agents, kitchen
utensils, clothing, shoes, cosmetics, insecticides, electronic devices). In addition, the extent and
use of the items affects the distribution patterns of PFAS compounds in dust of these buildings.
Results from the remaining studies conducted in Europe are presented in Table D-8.
D-91
-------
FINAL
2024
Table D-8. Summary of PFHxS Indoor Dust Data
Study
Location
Site Details
Results
United States
Scheretal. (2019)
United States (Twin Cities
metropolitan region, Minnesota)
Nineteen homes in three cities within a GCA near
former 3M PFAS production facility as well as from
three homes in the Twin Cities Metro outside the
GCA. Dust samples collected from an entryway to the
yard and from an interior living space (e.g., family
room, living room) in each home in July-September
2010. Homes within the GCA had previous or ongoing
PFAS contamination in drinking water and were
served by the Oakdale, Minnesota public water system
or a private well previously tested and shown to have
detectable levels ofPFOA orPFOS. Results were not
reported for homes outside the GCA.
Entryway: n = 19, DF 68%, median (range) =
8.2 (
-------
FINAL
2024
Study
Location
Site Details
Results
Strynar and Lindstrom (2008)
United States (North Carolina;
Ohio)
Dust samples from vacuum cleaner bags were
obtained in 2000-2001 during EPA's Children's Total
Exposure to Persistent Pesticides and Other Persistent
Organic Pollutants (CTEPP) study from North
Carolina (49 homes, 5 daycare centers) and Ohio (53
homes, 5 daycare centers). Vacuum cleaner bags were
only collected if available at each site.
n = 112, DF 77.7%, mean, median (maximum)
= 874, 45.5 (35,700) ng/g
(LOQ = 12.9 ng/g)
*Values below the LOQ assigned a value of
LOQ/V2
Fraseretal. (2013)
United States (Boston,
Massachusetts)
Dust samples were collected between January-March
2009 from offices (n = 31), homes (n = 30), and
vehicles (n = 13) of 31 individuals. Study participants
worked in separate offices located across seven
buildings, which were categorized into Building A,
Building B, and Other. Six samples were collected
from Building A, a newly constructed (approximately
one year prior to study initiation) building with new
carpeting and new upholstered furniture in each office.
Seventeen samples were collected from Building B, a
partially renovated (approximately one year prior to
study initiation) building with new carpeting
throughout hallways and in about 10% of offices.
Eight samples were collected from the other five
remaining buildings where no known recent
renovation occurred. Study offices were not vacuumed
during the sampling week and homes and vehicles
were not vacuumed for at least one week prior to
sampling. Entire accessible floor surface areas and
tops of immovable furniture were vacuumed in offices
and the main living area of homes. Entire surface areas
of the front and back seats of vehicles were vacuumed.
Number of home dust samples was reduced to 30
because 1 participant lived in a boarding house with
no main living area. Sufficient mass of dust for
analysis was available from only 13 vehicles.
Homes: n = 30, DF 40%, range = 6.05^130
ng/g
Offices: n = 31, DF 23%, range = 5.24-18.5
ng/g
Vehicles: n = 13, DF 46%, range = 5.22-108
ng/g
(LOQ =5 ng/g)
*Range of detected values reported
Canada
Beesoon et al. (2012)
Canada (Alberta)
Dust samples collected from the vacuum bag of the
central vacuum system on the same day that carpets
were sampled in a house built in 1989 in the
Edmonton area. Samples were collected in September
2008 prior to a major renovation where all carpets
were removed. The house was identified after
abnormally high serum levels of PFHxS were
2008 sampling: n = 1, point = 2,780 ng/g
2012 sampling: n = 1, point = 253 ng/g
(LOD/LOQ not reported)
*The 2008 value is reported as 2,780 ng/g in
the text but 2,900 ng/g in Figure 2
D-93
-------
FINAL
2024
Study
Location
Site Details
Results
identified in a husband and wife that were enrolled as
volunteer control participants in a clinical research
project. Many rooms in the house had wall-to-wall
carpeting installed on top of the heated floors. Since
1991, Scotchgard has been applied to the carpets in the
main floor family room and occasionally to the dining
room. The last application was in 2007, with six
intermittent applications. The house did not have a
fan-forced air circulation system and there was also no
fresh-air intake to the house from the outdoors.
Vacuum dust sample also collected in January 2012.
Kubwabo et al. (2005)
Canada (Ottawa)
Sixty-seven randomly selected homes with home
selection described in a previous study. Samples
collected in the winter of 2002-2003 from houses with
varying age and percentage of surface area covered by
carpet.
n = 67, DF 85%, mean, median (range) =
391.96,23.1 (2.28^1,305) ng/g
(MDL = 4.56 ng/g)
*Non-detects and values
-------
FINAL
2024
Study
Location
Site Details
Results
Padilla-Sanchez and Haug
(2016)
Norway (Oslo)
Homes of staff from the Norwegian Institute of Public
Health. Dust samples collected from vacuum cleaner
bags provided by staff. Sampling year not provided.
n = 7, DFa 71%, range = ND-1 ng/g
(MDL = 0.012 ng/g; MQL = 0.040 ng/g)
Jogsten et al. (2012)
Spain (Catalonia)
Dust sampling was performed in December 2009 from
ten households using household vacuum cleaner dust
bags. Samples were collected out of convenience and
may not be representative of the entire Catalan
population.
n = 10, DFa 100%, meana (range) = 1.07
(0.17-5.3) ng/g
(LOD = 0.003 ng/g)
Haug etal. (2011)
Norway (Oslo)
Forty-one homes of breastfeeding mothers recruited
for a study on exposure pathways. House dust samples
collected between February and May 2008 on two
consecutive days while the residence was in regular
use. Samples taken from elevated surfaces such as
bookshelves and window sills (deposited dust) and not
from the floor.
n = 41, DF (frequency of quantification)3 54%,
mean, median (range) = 8.4, 0.60 (0.21-142)
ng/g
(LOQ = 0.23-1.1 ng/g)
¦"Concentrations that were not detected or
-------
FINAL
2024
Study
Location
Site Details
Results
Huberetal. (2011)
Norway (Tromse)
Homes and workplaces sampled in winter 2007-2008.
Home samples included seven different living rooms
(L-l to L-7), one sleeping room (S, related to L-3),
one sofa (a stain repellant fabric, related to L-7), and
one carpet (related to L-4). Workplace samples
included an office and a storage room at Fram Center;
old documents and chemicals and highly contaminated
samples were stored in the storage room. Samples
were taken from bookshelves, commodes, TVs,
electrical heaters, picture frames, window sills and sun
blinds. Dust from the floor was not sampled.
All homes: n = 7, DF NR, median = 1.4 ng/g
Living room: n = 7, DFa 86%, mean, median
(range) = 1.7, 1.4 (0.8-3.1) ng/g
Carpet: n = 1, point = 0.5 ng/g
Sleeping room: n = 1, point =2.1 ng/g
Sofa: n = 1, point =1.7 ng/g
Office: n = 1, point = 27.8 ng/g
Storage room: n = 1, point = 1,814 ng/g
(LOD on column = 0.006 ng; MDL = 0.1-1.85
ng/g)
D'Hollander et al. (2010)
Belgium (Flanders)
Forty-three randomly selected homes and ten
randomly selected offices throughout Flanders.
Samples collected using a vacuum from bare floor,
possibly covered with carpet, in 2008. In homes, the
living room, bedroom, kitchen, and working area were
sampled.
Homes: n = 43, DF 40%, median (95th
percentile) = 0.1 (9) ng/g dw
Offices: n = 10, DF NR, median (95th
percentile) = 0.2 (5.1) ng/g dw
(LOQ = 0.1 ng/g)
¦"Concentration MQL
*Median calculated by replacing values
-------
FINAL
2024
Study
Location
Site Details
Results
Kato et al. (2009)
United States (Atlanta, Georgia),
Germany (unspecified), United
Kingdom (unspecified),
Australia (unspecified)
Thirty-nine household dust samples from the United
States (n = 10), Germany (n = 10), United Kingdom (n
= 9), and Australia (n = 10) collected in 2004 for
method validation. Dust sampling procedures not
described.
n = 39, DF 79.5%, median (range) = 185.5
(
-------
FINAL
2024
D.3.S. Air
Perfluoroalkyl chemicals have been released to air from wastewater treatment plants, waste
incinerators, and landfills (Ahrens et al., 2011), though there is limited information on the
detection levels or frequencies of PFHxS in either indoor or ambient air. NCBI (2022a) notes
that PFHxS has been detected in particulates and in the vapor phase in air and it can be
transported long distances via the atmosphere; it has been detected at low concentrations in areas
as remote as the Arctic and ocean waters. For example, PFHxS was detected in particle-phase air
samples collected in 2007 and 2008 from the Atlantic Ocean, Southern Ocean, and Baltic Sea
(NCBI, 2022a; Dreyer et al., 2009). PFHxS is not expected to be broken down directly by
photolysis (NCBI, 2022a). PFHxS in the particle-phase can be removed from the atmosphere
through wet and dry deposition (NCBI, 2022a). PFHxS in the vapor phase can undergo
hydroxylation in the atmosphere, with a (predicted average) atmospheric hydroxylation rate of
2.16 x 10-15 cm3/molecule - second to a (derived) rate of 1.4 x io~13 cm3/molecule - second at
25°C (with corresponding estimated half-life of 115 days for this reaction in air) (USEPA,
2022a; NCBI, 2022a). With a vapor pressure of 0.0046 mm Hg at 25°C (estimated),
8.10 x io~9 mm Hg (measured average), and 8.19 x io~9 mm Hg (predicted average),
volatilization is not expected to be an important fate process for this chemical (USEPA, 2022a;
NCBI, 2022a). EPA's Toxics Release Inventory reported release data for PFHxS in 2020, with
total onsite disposal, offsite disposal, and other releases concentrations of 1 pound at an
individual facility and 122 pounds at a second facility (USEPA, 2022b). PFHxS is not listed as a
hazardous air pollutant (USEPA, 2022c).
D. 3.5.1. Indoor Air
No studies from the U.S. reporting levels of PFHxS in indoor air were identified from the
primary or gray literature. However, several studies from Canada and Europe were identified and
are discussed briefly below and presented in Table D-9 (Harrad et al., 2019; Beesoon et al.,
2012; Goosey and Harrad, 2012; Jogsten et al., 2012; Barber et al., 2007). All of these studies
sampled from homes, while two studies also sampled from offices, one study also sampled a
laboratory, and only one study also sampled from cars and classrooms. In studies exclusively in
homes, two studies did not detect PFHxS in indoor air while the remaining three studies had
PFHxS detection frequencies ranging from 21 to 100%. In one of these studies, a Canadian
household with wall-to-wall Scotchgard-treated carpets on top of heated floors (and whose
residents had previously been found to have disproportionately high blood serum levels of
PFHxS) found 27.4 pg/m3 PFHxS in the family room and 426 pg/m3 PFHxS in the basement
(with no clear reason for the higher levels in the basement). Additional research is needed to
evaluate PFHxS in indoor air from U.S. locations and from a variety of non-home
microenvironments (offices, cars, classrooms, and laboratories) in Canada and Europe.
Jogsten et al. (2012) sampled indoor air (n = 10) from selected homes in Catalonia, Spain in
December 2009 and evaluated levels of 27 PFCs, though PFHxS was not detected. The
remaining studies evaluated PFHxS levels in offices, vehicles, and/or schools, in addition to
homes (Harrad et al., 2019; Goosey and Harrad, 2012; Barber et al., 2007). In Ireland, Harrad et
al. (2019) collected air samples in homes (living rooms, n = 34), offices (n = 34), cars (n = 31),
and school classrooms (n = 28) between August 2016 and January 2017. PFHxS was detected in
all four indoor environments in 21%, 44%, 23%, and 25% of samples for homes, offices, cars,
and classrooms, respectively. However, the median concentrations were below the LOD for all
D-98
-------
FINAL
2024
environments. Goosey and Harrad (2012) detected PFHxS with mean (median) concentrations of
36 (23) pg/m3 in homes (n = 20) and 94 (84) pg/m3 in offices (n = 12) sampled in the United
Kingdom between September 2008 and March 2009. In addition, Goosey and Harrad (2012) also
examined seasonal variation in these same five indoor microenvironments by sampling monthly
between September 2008 and August 2009. The relative standard deviation for PFHxS was
between 52-106%. The observed variation could not be attributed to changes in room contents or
to seasonality. The study also measured ambient air concentrations in the same location and
concluded that indoor air concentrations significantly exceeded ambient air concentrations; the
authors suggested that indoor emissions may influence both indoor and outdoor PFAS levels. In
Norway, neutral and ionic PFAS were analyzed in indoor air samples collected from three homes
and one laboratory in Troms0 between in April and June 2005 (Barber et al., 2007). Results for
15 neutral PFAS and 16 ionic PFAS (including PFHxS) were presented but PFHxS was not
detected (method quantification limit [MQL] = 4.09 pg/m3; authors considered this a high MQL
and likely due to much lower sampling volume).
D-99
-------
FINAL
2024
Table D-9. Summary of PFHxS in Indoor Air
Study
Location
Site Details
Results
Canada
Beesoon et al. (2012)
Canada (Alberta)
Indoor air samples collected from one home in the
Edmonton area, built in 1989. The house was
identified after abnormally high serum levels of
PFHxS were identified in a husband and wife that
were enrolled as volunteer control participants in a
clinical research project. Many rooms in the house had
wall-to-wall carpeting installed on top of the heated
floors. Since 1991, Scotchgard has been applied to the
carpets in the main floor family room and occasionally
to the dining room. The last application was in 2007,
with six intermittent applications. The house did not
have a fan-forced air circulation system and there was
also no fresh-air intake to the house from the outdoors.
At the time of indoor air sampling in September 2008,
renovations had begun in the basement. Analysis for
PFHxS conducted on Fraction 1 (particulate phase)
from the main floor family room and finished
basement.
Main floor family room: n = 1, point = 27.4
pg/m3
Basement: n = 1, point = 426 pg/m3
(LOD/LOQ not reported)
Europe
Jogsten et al. (2012)
Spain (Catalonia)
Indoor air sampling was performed in December 2009
from ten households at approximately 1 m above the
floor. Samples were collected out of convenience and
may not be representative of the entire Catalan
population. Both particulate and gas phases collected.
n = 10, DF 0%
(LOD = 3.1-280 pg/m3 for all ionic PFAS)
Harrad et al. (2019)
Ireland (Dublin, Gal way,
Limerick)
Air samples collected from homes (living rooms),
offices, cars, and school classrooms; dust samples also
collected. Samples collected between August 2016
and January 2017. Sample numbers were split
approximately equally from each of the three counties.
Gas or particulate phase not specified.
Homes: n = 34, DF 21%, mean, median
(range) = <0.4, <0.4 (<0.4-0.46) pg/m3
Offices: n = 34, DF 44%, mean, median
(range) = 0.40, <0.4 (<0.4-1.4) pg/m3
Cars: n = 31, DF 23%, mean, median (range)
= 0.15, <0.4 (<0.4-0.55) pg/m3
Classrooms: n = 28, DF 25%, mean, median
(range) = <0.4, <0.4 (<0.4-2.3) pg/m3
(LOD = 0.4 pg/m3)
*When analyte peaks are
-------
FINAL
2024
Study
Location
Site Details
Results
Goosey and Harrad (2012)
United Kingdom (Birmingham)
Samples collected from homes and offices between
September 2008 and March 2009. In addition, air
samples were collected monthly from one office and
four living rooms between September 2008 and
August 2009. Gas or particulate phase not specified.
Homes: n = 20, DFa 90%, mean, median
(range) = 36,23 (<1.1-220) pg/m3
Offices: n = 12, DFa 83%, mean, median
(range) = 94, 84 (<1.1—330) pg/m3
Seasonal Variations: concentrations reported
for each month from Sept 2008 to Aug 2009
Home 1: 12, 100,24,21,22, 17,49,35,4,
<1.1 pg/m3
Home 2: 110,49,30, 7, 30, 9,44, 50, 110,
44,22 pg/m3
Home 3: <1.1,24, 11, 12, 49,100, 9, 16, 84,
17,19, <1.1 pg/m3
Home 4: <1.1, 37, 44, 5,27, 23, 47, 25, 3,
27,22, 37 pg/m3
Office 1: <1.1, 8, 30,16, 50, 33, 12, 12,18,
9, <1.1 pg/m3
(DL =1.1 pg/m3)
*Where concentration
-------
FINAL
2024
D.3.5.2. Ambient Air
The EPA identified a single study reporting ambient air concentrations of PFHxS from the U.S.
Kim and Kannan (2007) analyzed particle phase (n = 8) and gas phase (n = 8) concentrations of
perfluorinated acids in ambient air samples collected in and around Albany, New York, in May
and July 2006 to examine the relative importance of certain media pathways to the contamination
of urban lakes. PFHxS was detected in all gas phase samples with mean, and median
concentrations of 0.31 pg/m3 and 0.34 pg/m3, respectively, but was not detected in the particle
phase (LOQ = 0.12 pg/m3).
Additional studies were identified that examined PFHxS in ambient air from sampling efforts
conducted in Canada and Europe. These studies are summarized below and presented in
Table D-10. In Canada, Ahrens et al. (2011) did not detect PFHxS in air around a WWTP and
two landfill sites. In Europe, Barber et al. (2007) collected air samples from four field sites in the
United Kingdom, Ireland, and Norway for analysis of neutral and ionic PFAS. PFHxS was
detected in the ambient air particle phase at two sampling events in Manchester, UK at 0.1 and
1.0 pg/m3, at one of two sampling events in Hazelrigg, UK at 0.04 pg/m3, at one sampling event
in Norway at 0.05 pg/m3, and at one sampling event in Ireland at 0.07 pg/m3. Goosey and Harrad
(2012) collected ambient air samples from ten different locations within a 1.5 km radius of the
University of Birmingham campus in the United Kingdom. PFHxS was detected with a mean
(maximum) concentration of 7.0 (30) pg/m3. Beser et al. (2011) detected PFHxS in 5% of
ambient air particulate samples from five locations in Alicante province, Spain (3 residential, 1
rural, 1 industrial), with a mean (maximum) concentration of 8.7 (15.9) pg/m3. The highest
concentration observed was at the industrial site. Harrad et al. (2020) analyzed air samples near
ten Irish municipal solid waste landfills in nonindustrial areas. PFHxS was detected in more than
20% of the samples, with mean (maximum) concentrations at downwind and upwind locations of
0.34 (0.79) pg/m3 and 0.23 (0.81) pg/m3, respectively. One European study (Jogsten et al., 2012)
did not detect PFHxS in ambient air samples collected outside homes in Catalonia, Spain.
D-102
-------
FINAL
2024
Table D-10. Summary of the Occurrence of PFHxS in Ambient Air
Study
Location
Site Details
Results
United States
Kim and Kannan (2007)
United States (Albany, New
York)
Roof of a lakehouse building located at Washington
Park Lake in May and July 2006. Both particulate and
gas phases collected.
Gas: n = 8, DF NR, mean, median (range) =
0.31,0.34 (0.13-0.44) pg/m3
Particle: n = 8, DF 0%
(LOQ = 0.12 pg/m3)
*Non-detects were set to zero; values below
the LOQ were set to 'A LOQ
Canada
Ahrens et al. (2011)
Canada (Ontario)
Samples collected on and around one municipal
WWTP for 63 days between July and September
2009. Samplers were placed at the primary clarifier,
aeration tank, secondary clarifier, and at four reference
sites (three near [within 200 m of the treatment tanks]
and one distant [-600 m from the perimeter of the
WWTP]).
Samples also collected at two municipal solid waste
landfills between June and August 2009 for 55 days.
The two landfills were 60 km apart. Samplers were
located upwind and onsite of the active zone of each
landfill site and one field blank was collected at each
site. Both sites collected landfill gas and the active
area of the landfill was kept to a minimum by covering
the waste with soil and a plastic film.
The passive sampling configuration used resulted in
the collection of mainly PFAS in the gas phase.
WWTP:
Reference sites (near): n = 3, DF 0%
Primary clarifier: n = 2, DF 0%
Aeration tank: n = 3, DF 0%
Secondary clarifier: n = 2, DF 0%
Reference site (distant): n = 1, DF 0%
Landfills:
Upwind: n = 2, DF 0%
On site: n = 2, DF 0%
(MDL = 0.04-0.87 pg/m3 for PFCAs, PFSAs,
andPFOSA)
Europe
Harrad et al. (2020)
Ireland (multiple cities)
Samples collected from ten municipal solid waste
landfills upwind and downwind at each site between
November 2018 and January 2019. Location of
sampling sites based on wind direction data taken
from the Irish Meteorological Service, with slight
modification where necessary based on local
information from site operators and ease of access.
Sample sites were between 150 and 500 m of the
center of the landfill. Waste accepted by the landfills
included: municipal solid waste, industrial (non-
hazardous) waste, construction and demolition, and
Downwind:
n = 10, DFa 60%, mean, median (range) =
0.34,0.23 (<0.15-0.79) pg/m3
Upwind:
n = 10, DFa 40%, mean, median (range) =
0.23,0.08 (<0.15-0.81) pg/m3
(LOD = 0.15 pg/m3)
*Non-detects replaced by Vi LOD
D-103
-------
FINAL
2024
Study
Location
Site Details
Results
biomedical waste. Gas or particulate phase not
specified.
Goosey and Harrad (2012)
United Kingdom (Birmingham)
Ten different locations within a 1.5 km radius of the
University of Birmingham campus. Samples collected
in March 2009. In addition, air samples were collected
monthly from three outdoor locations between
September 2008 and August 2009 where two of the
locations were among the ten sampled in March 2009
and the third location was Harwell, Oxfordshire in
Southeast England, a semi-rural site. Gas or
particulate phase not specified.
n = 10, DFa 50%, mean, median (range) = 7.0,
2.7 (<1.1-30) pg/m3
(DL assumed to be 1.1 pg/m3)
Seasonal Variations: concentrations reported
for each month from Sept 2008 to Aug 2009
Birmingham site 1: 12,11, 3.8, 2.0, 420,
<0.8, <0.8, <0.8,1.3, 1.4 pg/m3
Birmingham site 7: 1.5, 8.3, <0.8, 3.8, 1.9,
5.3, 5.6, <0.8, 12, <0.8, 20 pg/m3
Harwell site: <0.8, 1.6,4.9, <0.8, 1.3, <0.8,
1.0, <0.8,4.9, 1.5,20, 8.8 pg/m3
(DL assumed to be 0.8 pg/m3)
*For concentration < DL, VixDL was used for
calculations
Jogsten et al. (2012)
Spain (Catalonia)
Outdoor air sampling conducted in December 2009 for
the purposes of comparison to indoor air and dust
samples. Number of sites not specified but assumed to
be ten because indoor air was sampled from ten
homes. Samples were collected out of convenience
and may not be representative of the entire Catalan
population. Both particulate and gas phases collected.
n = 10, DF 0%
(LOD = 3.1-280 pg/m3 for all ionic PFAS)
Beser et al. (2011)
Spain (Alicante province)
Samples collected from April to July 2010 from five
stations. Two stations were placed in Elche (one in a
residential area and the other in an industrial area).
The third station was placed in a residential area of
Alicante City. The fourth station was in a rural area of
Pinoso and the last station was in a residential area of
Alcoy. Concentrations reported for PM2.5-bound
PFHxS.
Elche (residential): n = 11, DFa 0%
Elche (industrial): n = 13, DFa 7.7%, mean =
15.9 pg/m3
Alicante City: n = 11, DFa 0%
Pinoso: n = 3, DFa 0%
Alcoy: n = 3, DFa 33%, mean =1.5 pg/m3
(MQL = 1.4 pg/m3)
*Mean calculated from values >MQL
Barber et al. (2007)
United Kingdom (Hazelrigg,
Manchester); Ireland (Mace
Head); Norway (Kjeller)
Samples collected from four field sites in Europe:
Hazelrigg (semirural) and Manchester (urban) were
sampled in two sampling events in February-March
2005 and November 2005-January 2006; Mace Head
(rural) was sampled in March 2006; and Kjeller (rural)
was sampled in November-December 2005. PFHxS
was measured in the particulate phase.
Hazelrigg first sampling event:
n = 2, DF NR, mean = <5.9 pg/m3
(MQL = 5.93 pg/m3) Feb/Mar 2005
*The glass-fibre filters were analyzed in a
batch of samples that showed contamination
problems, so the high associated blank value
D-104
-------
FINAL
2024
Study
Location
Site Details
Results
used to calculate the MQL put most analytes
MQL
D-105
-------
FINAL
2024
D.3.6. Soil
The production of PFHxS and its use as a raw material or precursor for manufacturing PFAS-
based products, as well as its previous use in firefighting foam and carpet treatment solutions,
could result in its release to soils through various waste streams (NCBI, 2022a). When released
to soil, PFHxS is expected to have very high mobility (NCBI, 2022a). The maximum
concentration of PFHxS in soil samples collected from a PFAS production facility in Minnesota
was 3,470 ng/g, with PFHxS detected in 90% of the samples collected (NCBI, 2022a; ATSDR,
2021; 3M, 2007). The maximum concentration of PFHxS in soil samples collected at a fire
training area at a PFAS production facility in Minnesota was 62.2 ng/g, with PFHxS detected in
100% of samples (ATSDR, 2021; 3M, 2007). PFHxS was also detected in soil samples collected
from a former sludge incorporation area at a PFAS production facility in Decatur, Alabama, with
average levels ranging from 3.56 ng/g to 270 ng/g, with PFHxS detected in 86% of the samples
collected (ATSDR, 2021; 3M, 2012). PFHxS has been found to accumulate in the roots of maize
plants grown in soil containing PFHxS and other PFAS (ATSDR, 2021; Krippner et al., 2015).
Several additional peer-reviewed studies were identified that evaluated the occurrence of PFHxS
and other PFAS in soil collected in the U.S. Two studies by Scher et al. (2019; 2018) evaluated
soils collected in 2010 from the garden planting area of 20 homes in Minnesota within a GCA
impacted by the former 3M PFAS production facility. Homes within the GCA had previous or
ongoing PFAS contamination in drinking water and were served by the Oakdale, Minnesota
public water system or a private well previously tested and shown to have detectable levels of
PFOA or PFOS. Both studies reported similar median PFHxS levels of 0.08 ng/g and 0.057 ng/g
(n = 20-34) from the 2019 and 2018 publications, respectively. Scher et al. (2018) also reported
a median PFHxS concentration of 0.078 ng/g from six samples collected outside the GCA.
Three studies analyzed soils potentially impacted by AFFF use (Nickerson et al., 2020; Eberle et
al., 2017; Anderson et al., 2016). Anderson et al. (2016) examined surface and subsurface soil
from 40 sites across 10 active Air Force installations throughout the continental United States
and Alaska between March and September 2014. Installations were included if there was known
historic AFFF release in the period 1970-1990. It is assumed that the measured PFAS profiles at
these sites reflect the net effect of several decades of all applicable environmental processes. The
selected sites were not related to former fire training areas and were characterized according to
volume of AFFF release - low, medium, and high. Across all sites, the PFHxS detection
frequency was 76.92%> in 100 surface soil samples (median concentration of detects was
5.7 ng/g) and 59.62%> in 112 subsurface soil samples (median concentration of detects was
4.4 ng/g). PFHxS was detected more frequently at high-volume release sites (82.5%> in 32
surface soil samples with mean concentration of 19.7 ng/g; 87.5%> in 31 subsurface soil samples
with mean concentration of 9.3 ng/g) than at low-volume sites (75.0% in 12 surface soil samples
with mean concentration of 13.9 ng/g; 58.8%> in 17 subsurface soil samples with mean
concentration of 57.9 ng/g) and medium-volume sites (59.2%> in 56 surface soil samples with
mean concentration of 39.4 ng/g; 71.4% in 64 subsurface soil samples with mean concentration
of 55.4 ng/g). Nickerson et al. (2020) developed a method to quantify anionic, cationic, and
zwitterionic PFAS from AFFF-impacted soils. The method was applied to two soil cores
collected from two different AFFF-impacted former fire training areas; the sampling year and
geographic location were not provided. Eleven soil samples, corresponding to 11 depths ranging
from 0.46 m to 15.1 m, were evaluated from Core E, and 12 soil samples, at depths ranging from
D-106
-------
FINAL
2024
0.30 m to 14.2 m, were evaluated from Core F. PFHxS was detected at all depths analyzed for
both cores, with concentrations ranging from 1.17 ng/g dw to 160.6 ng/g dw for Core E and
0.66 ng/g dw to 296.4 ng/g dw for Core F. Eberle et al. (2017) investigated the effects of an in
situ chemical oxidation treatment for remediation of chlorinated volatile organic compounds and
PFAAs co-contaminants. Soil samples were collected in 2012-2013 before and after a pilot scale
field test at a former fire training site at Joint Base Langley-Eustis, Virginia. Monthly fire
training activities were conducted at the site from 1968 to 1980 and irregular fire training
activities continued until 1990. Impacted soil was excavated in 1982 but details were not
provided. PFHxS was detected in 4 of 5 pre-treatment samples and in all 14 post-treatment
samples. In the available three paired pre- and post-treatment soil samples, two pairings showed
a decrease in PFHxS concentration after treatment, from 6.7 ng/g to 1.40 ng/g in one pairing and
from 12 ng/g to 1.44 ng/g in the other pairing. In the third pairing, PFHxS was not detected pre-
treatment and was found at 0.31 ng/g post-treatment.
Of the remaining two studies conducted in the United States, Venkatesan and Halden (2014)
conducted outdoor mesocosm studies to examine the fate of PFAS in biosolids-amended soil
collected during 2005-2008. Biosolids were obtained from a WWTP in Baltimore that primarily
treated wastewater from domestic sources with only minor contribution (1.9%) from industry.
PFHxS was not detected in the control (nonamended) soil and not consistently detected in the
biosolids-amended soil (MDL = 0.03-0.14 ng/g dw). In a field and greenhouse study, Blaine et
al. (2013) studied the uptake of PFAS into edible crops grown in control and biosolids-amended
soil. In the field study, urban biosolids were obtained from a WWTP receiving both domestic
and industrial waste while rural solids were obtained from a WWTP receiving domestic waste
only. Mean PFHxS concentrations were below the LOQ (0.01 ng/g) in the urban control and
biosolids-amended soils and in the rural control soil. In the rural biosolids-amended soil, the
mean PFHxS concentration was 0.016 ng/g. In the greenhouse study, three soils (nonamended
control, industrially impacted, and municipal) were investigated. Industrially impacted soils
contained composted biosolids from a small municipal WWTP that was impacted by PFAA
manufacturing while municipal soils were obtained from a reclamation site in Illinois where
municipal biosolids were applied for 20 years. PFHxS was detected in all three soils at an
average concentration of 0.44 ng/g, 1.38 ng/g, and 5.11 ng/g in control, industrially impacted,
and municipal soil, respectively. Authors noted that the trace levels of PFAS detected in the
control soil may be due to minor cross-contamination from plowing, planting, or atmospheric
deposition from the surrounding area where biosolids have been applied.
Studies conducted in Europe and Canada were also identified and are summarized below and
presented in Table D-l 1. Of the European studies, a study in Ireland (Harrad et al., 2020)
examined soil samples collected upwind and downwind from ten municipal solid waste landfills
and found PFHxS levels to be lower at downwind locations. Based on the overall study findings,
however, the authors concluded there was no discernible impact of the landfills on
concentrations of PFAS in soil surrounding these facilities. In the Maltese islands, Sammut et al.
(2019) sampled soil from small urban fields and found PFHxS concentrations to range from non-
detectable levels to 0.12 ng/g. Grannestad et al. (2019) investigated soils from a skiing area in
Norway to elucidate exposure routes of PFAS into the environment from ski products, such as
ski waxes. PFHxS was below the limit of quantification in soil samples from both the Granasen
skiing area and the Jonsvatnet reference area. One study performed in Belgium (Groffen et al.,
2019) evaluated soils collected at a 3M fluorochemical plant in Antwerp and at four sites located
D-l 07
-------
FINAL
2024
at increasing distances from the plant. PFHxS levels were elevated at the plant site but not
detected at other locations. Four studies were conducted at locations near firefighting training
sites, which found varying results from non-detected levels to 2,344 ng/g dw (Dauchy et al.,
2019; Skaar et al., 2019; Hale et al., 2017; Filipovic et al., 2015).
D-108
-------
FINAL
2024
Table D-ll. Summary of PFHxS Data in Soil
Study
Location
Site Details
Results
United States
Scheretal. (2019)
United States (Twin Cities
metropolitan region, Minnesota)
Area near former 3M PFAS production facility. Soil
composite samples (200-500 g) collected in 2010
from the garden planting area of 20 homes in 3 cities
within a GCA as well as from 3 homes in the Twin
Cities Metro outside the GCA. Homes within the GCA
had previous or ongoing PFAS contamination in
drinking water and were served by the Oakdale,
Minnesota public water system or a private well
previously tested and shown to have detectable levels
of PFOA or PFOS. Results were not reported for
homes outside the GCA.
n = 20, DF 85%, median (90th percentile) =
0.08 (0.16) ng/g
(RL = 0.0008-0.033 ng/g for all PFAS)
Scheretal. (2018)
United States (Twin Cities
metropolitan region, Minnesota)
Area near former 3M PFAS production facility. Soil
samples collected in 2010 at a sample depth of 0-6
inches from the garden planting area of 20 homes in 3
cities within a GCA as well as from 3 homes in the
Twin Cities Metro outside the GCA. Homes within the
GCA had previous or ongoing PFAS contamination in
drinking water and were served by the Oakdale,
Minnesota public water system or was formerly or
currently using a private well previously tested and
shown to have detectable levels of PFOA or PFOS. At
14 homes, soil samples were collected on two separate
days.
Within GCA: n = 34, DF 71%, median (range)
= 0.057 (ND-0.24)ng/g
Outside GCA: n = 6, DF 100%), median
(range) = 0.078 (0.028-0.11) ng/g
(MDL = 0.008-0.033 ng/g for all PFAS)
*Values below the method reporting limit but
above the lowest calibration limit (estimated
values) were included in all analyses as
quantitative results
*Values below the lowest calibration limit
were replaced with Vi of the limit
Anderson et al. (2016)
United States (national)
Forty AFFF-impacted sites from ten active U.S. Air
Force installations with historic AFFF release between
1970 and 1990 that were not related to former fire
training areas. It is assumed that the measured PFAS
profiles at these sites reflect the net effect of several
decades of all applicable environmental processes.
AFFF-impacted sites included emergency response
locations, hangers and buildings, and testing and
maintenance related to regular maintenance and
equipment performance testing of emergency vehicles
and performance testing of AFFF solution. Previous
remedial activities for co-occurring contaminants were
not specifically controlled for in the site selection
process; active remedies had not been applied at any
of the sites selected. Approximately ten samples were
Surface soil:
Overall: n = 100, DF 76.92%o, median
(maximum) = 5.7 (1,300) ng/g
Breakdown by site:
Emergency Response (low-volume release):
n = 12, DF 75.0%), mean (range) = 13.9
(0.38-64) ng/g
Hangars and Buildings (medium-volume
release):
n = 56, DF 59.2%o, mean (range) = 39.4
(0.34-1,300) ng/g
Testing and Maintenance (high-volume
release):
D-109
-------
FINAL
2024
Study
Location
Site Details
Results
collected between March and September 2014 at each
site for surface and subsurface soil; sites were grouped
according to volume of AFFF release—low-volume
typically had a single AFFF release, medium-volume
had one to five releases, and high-volume had multiple
releases.
n = 32, DF 82.5%, mean (range) = 19.7
(0.29-180) ng/g
(RL = 0.29 ng/g)
Subsurface soil:
Overall: n = 112, DF 59.62%, median
(maximum) = 4.4 (520) ng/g
Breakdown by site:
Emergency Response (low-volume release):
n = 17, DF 58.8%), mean (range) = 57.9
(0.37-520) ng/g
Flangars and Buildings (medium-volume
release):
n = 64, DF 71.4%, mean (range) = 55.4
(0.35-1,300) ng/g
Testing and Maintenance (high-volume
release):
n = 31, DF 87.5%o, mean (range) = 9.3
(1-1^0) ng/g
(RL = 0.31 ng/g)
*Median calculated using quantified
detections
*Non-detects were substituted with Vi the
reporting limit
Nickerson et al. (2020)
United States (unspecified)
Soil cores E and F from two different AFFF-impacted
fire training areas; sampling year and geographic
location not provided. Soil core E contained 11- 0.3 m
increment samples from 0.3-15.2 m below ground
surface and was collected in an area where the
surficial soils were likely disturbed due to regrading
and other soil redistribution activities. Soil core F
contained 12- 0.61 m increment samples from 0-14.2
m below ground surface and was collected in an area
where the surficial soils were highly permeable only
within the upper 0.5 to 1 m, and the underlying
impermeable clay layer exhibited a relatively high
cation exchange capacity and organic carbon content.
The water table was relatively shallow (depth <3 m) at
both sites.
Core E:
0.46 m = 1.44 ng/g dw
2.9 m = 2.12 ng/g dw
3.66 m = 4.17 ng/g dw
3.96 m = 15.21 ng/g dw
4.27 m = 28.68 ng/g dw
4.57 m = 4.13 ng/g dw
4.88 m = 5.73 ng/g dw
7.01 m = 13.86 ng/g dw
8.38 m = 160.6 ng/g dw
10.5 m = 139.0 ng/g dw
15.1 m = 1.17 ng/g dw
Core F:
0.30 m = 11.07 ng/g dw
1.22 m = 296.4 ng/g dw
1.83 m = 276.2 ng/g dw
D-110
-------
FINAL
2024
Study
Location
Site Details
Results
2.44 m = 106.2 ng/g dw
3.05 m = 42.69 ng/g dw
4.11 m = 28.78 ng/g dw
7.62 m = 14.19 ng/g dw
8.84 m = 6.26 ng/g dw
9.45 m = 3.25 ng/g dw
10.5 m = 0.66 ng/g dw
11.9 m = 3.06 ng/g dw
14.2 m = 7.96 ng/g dw
(LOD/LOQ not reported)
Eberle et al. (2017)
United States (Joint Base
Pilot testing area in former fire training area (Training
Pre-treatment:
Langley-Eustis, Virginia)
Site 15) at Joint Base Langley-Eustis where monthly
fire training activities were conducted from 1968 to
1980 in a zigzag pattern burn pit. Facility was
abandoned in 1980 but irregular fire training activities
using an above-ground germed burn pit continued
until 1990. Impacted soil was removed in 1982 but
additional details of the excavation are not well
known. Soil samples collected for pre- (April and
September 2012) and post- (December 2013) in situ
chemical oxidation treatment using a peroxone
activated persulfate (OxyZone) technology. Treatment
was conducted in Test Cell 1 over 113 days (April-
August 2013). Soil samples were collected adjacent to
wells; wells outside Test Cell 1 were used as sentry
wells. Well IDs for pre- and post-sampling were not
provided but the following three pairings were
assumed based on Table 2 in the paper: U-20 with SB-
106; U-16 with SB-112; and 1-1 with SB-109.
1-1 (1.2-4.3 m) = 12 ng/g
1-2 (1.2-4.3 m) = 83 ng/g
U-12 (2.1m) =1.2 ng/g
U-16 (3.0 m) = 6.7 ng/g
U-20 (1.8 m) = ND
(LOQ = 0.68-0.72 ng/g)
Post-treatment:
SB-101 (4.3 m)= 8.08 ng/g
SB-105 (1.8 m) = 0.83 ng/g
SB-106/U-20 (1.8 m) = 0.31 ng/g
SB-106 (4.3 m) = 5.08 ng/g
SB-107 (1.8 m) = 2.11 ng/g
SB-107 (4.3 m) = 3.99 ng/g
SB-108 (1.8 m)= 1.48 ng/g
SB-108 (4.3 m) = 4.83 ng/g
SB-109/I-1 (3 m)= 1.44 ng/g
SB-Ill (4.3 m)= 11.85 ng/g
SB-112 (1.8 m) = 2.57 ng/g
SB-112/U-16 (3 m) = 1.4 ng/g
SB-114 (1.8 m) = 3.63 ng/g
SB-114 (4.3 m)= 16.17 ng/g
(LOQ = 0.12 ng/g)
Venkatesan and Halden (2014)
United States (Baltimore,
Archived agricultural soil (nonamended) collected
Nonamended: n = NR, DF 0%
Maryland)
during 2005-2008 at a depth of 0-20 cm from the
United States Department of Agriculture-Agricultural
Research Service Beltsville Agricultural Research
Amended: n = NR, authors reported that
PFHxS was not consistently detected in
bio solids-amended mesocosms
Center; number of sampling sites and number of
samples not provided.
(MDL = 0.03-0.14 ng/g dw for all PFAS)
D-lll
-------
FINAL
2024
Study
Location
Site Details
Results
Biosolids-amended soil obtained by mixing biosolids
and soil at a volumetric ratio of 1:2. Biosolids were
from Back River WWTP in Baltimore, a full-scale
activated sludge treatment plant. Raw wastewater was
primarily from domestic sources with only minor
contribution (1.9%) from industry.
Blaine et al. (2013)
United States (Midwest)
Urban and rural full-scale field study with control
(nonamended) and biosolids-amended plots. Three
agricultural fields were amended(0.5 x, 1 x, or 2x)
with municipal biosolids. Urban biosolids (1 x and 2*)
were from a WWTP receiving both domestic and
industrial waste. Rural biosolids (0.5x) were from a
WWTP receiving domestic waste only. Control plots
were proximal to the rural and urban amended corn
grain and corn stover field sites; sampling year not
provided.
Greenhouse study with control (nonamended) and
biosolids-amended soil. Nonamended soil obtained
from a field that received commercial fertilizers and
had a similar cropping system as the nearby municipal
soil site. Municipal soil was obtained from a
reclamation site in Illinois where municipal biosolids
were applied at reclamation rates for 20 years,
reaching the cumulative biosolids application rate of
1,654 Mg/ha. Industrially impacted soil was created
by mixing composted biosolids from a small
municipal (but impacted by PFAA manufacturing)
WWTP with control soil on a 10% mass basis.
Sampling year not provided.
Field study:
Urban non-amended: n = 3-7, DF NR, mean
< 0.01 ng/g
Urban 1 x: n = 3-7, DF NR, mean < 0.01
ng/g
Urban 2x: n = 3-7, DF NR, mean < 0.01
ng/g
Rural non-amended: n = 3-7, DF NR, mean
< 0.01 ng/g
Rural 0.5x: n = 3-7, DF NR, mean = 0.16
ng/g
(LOQ = 0.01 ng/g)
Greenhouse study:
Nonamended: n = 3-5, DF NR, mean = 0.44
ng/g
Industrially impacted: n = 3-5, DF NR,
mean = 1.38 ng/g
Municipal: n = 3-5, DF NR, mean =5.11
ng/g
(LOQ not reported)
Canada
Cabrerizo et al. (2018)
Canada (Melville and
Cornwallis Islands)
Catchment areas of lakes in the Cape Bounty Arctic
Watershed Observatory on southern Melville Island
(West, East, and Headwater lakes) during summer
(late July-early August) 2015 and 2016, representing
an environment largely unimpacted by direct human
activity; data for 19 sampling sites available (S6, SI 1—
S28).
Catchment areas of lakes on Cornwallis Island
(Resolute, North, Small, Meretta, 9 Mile, and Amituk
lakes) near the community of Resolute Bay during
summer (late July-early August) 2015 and 2016.
Melville Island lakes:
n = 19, DFa 95%, range =
-------
FINAL
2024
Study
Location
Site Details
Results
Resolute Bay has a military and civilian airport which
discharged its wastewaters into the upper area of the
catchment until 1997, three old solid waste landfills
1.5-2 km west of the airport used until the mid-1990s,
and Arctic research and military training facilities
close to the airport that support activities such as
vehicle use, firefighting, and construction/demolition;
eight sampling sites (S29-S36).
Mejia-Avendano et al. (2017)
Canada (Lac-Megantic, Quebec)
Site of July 2013 Lac-Megantic train accident where
63 out of 72 train cars carrying 8 million liters of
crude oil derailed and a major oil fire ignited. Seven
types of AFFFs and approximately 33,000 L of AFFF
concentrates were used. Samples were collected in
July 2013 weeks after the accident from the western
shores of Chaudiere River, at the point where the oil
and AFFF runoff reached the river, approximately 500
m from the edge of the derailment site; in July 2015
from the fire burn site and adjacent area in downtown
Lac-Megantic where the soil was continuously
excavated for remediation (the site was the closest to
the accident site among the areas open to sampling);
and from a background, nonimpacted area next to
Chaudiere River, about 5 km from the accident site, on
the east shore of the river and on the opposite side of
the accident.
Background:
n = 3, DF NR, mean = 0.015 ng/g dw
2013:
n = 12 (from 12 sites), DFa 92%, range =
-------
FINAL
2024
Study
Location
Site Details
Results
1983: 0.010; 0.018 ng/g
1973: 0.011; 0.007 ng/g
1962: 0.017; 0.015 ng/g
1945: 0.012; 0.008 ng/g
1927: ND;ND
(IDL = 0.0022 ng/g)
*Authors estimated the year for each core
segment using cores for a different study that
underwent dating
Europe
Groffen et al. (2019)
Belgium (Antwerp)
3M perfluorochemical plant and four sites with
increasing distance from plant were selected based on
prior biomonitoring studies in the vicinity of the plant.
The four sites are: Vlietbos (1 km SE from 3M), Rot-
Middenvijver (2.3 km ESE from 3M), Burchtse Weel
(3 km SE from 3M), and Fort 4(11 km SE from 3M).
Samples collected in June 2016.
Plant: n = 13, DF 31%, mean, median (range)
= 6.88,
-------
FINAL
2024
Study
Location
Site Details
Results
thick concrete slab was built in the 1990s. Area 3 was
SC-64 = 6,15, 29, 49, 14, 6, <4 ng/g dw (0-
used for firefighting activities since 1987 and is
0.25, 0.25-0.5, 0.5-0.75, 0.75-1, 1-1.5,
situated on a 1-meter thick concrete slab on the
1.5-2,2-2.5 m)
foundations of the former oil refinery. Area 4
SC-65 = 12, 30,25, <4,17,18, 21 ng/g dw
corresponds to the site's WWTP where sludge and
(0-0.25, 0.25-0.5, 0.5-0.75, 0.75-1, 1-
sediment from a lagoon were stored directly on the
1.5, 3-3.5, 3.5—4- m)
ground; influents of the WWTP are highly
SC-66 = 8, <4, <4, 8, 3, 2, <2, <2, <2, 8 ng/g
contaminated by PFAS. Area 5 was used for
dw (0-0.25, 0.25-0.5, 0.5-0.75, 0.75-1,
firefighting training exercises by the former oil
1-1.5, 1.5-2, 2-2.5, 2.5-3, 3-3.5, 3.5-4
refinery. Area 6 is used for firefighting exercises with
m)
tank trucks.
SC-58b = 17, 2, <10, 4 ng/g dw (4-5, 5-6,
9-10, 14-15m)
SC-59b = 14,39, 34,11, 63 ng/g dw (3^,
4-5, 6-7, 9-10, 14-15 m)
SC-65b = 3, 15, 5, 5 ng/g dw (4-5, 7-9, 9-
11,14-15 m)
SC-67 = 4, 5, 5,12 ng/g dw (0-1, 1.3-2, 2-
3, 4-5 m)
Area 3:
SC-40 = <4 ng/g dw (0-1 m)
SC-43 = <2 ng/g dw (1-2 m)
SC-45 = <2 ng/g dw (0-1 m)
SC-47 = 18 ng/g dw (0-1 m)
SC-48 = <4 ng/g dw (0-1 m)
SC-41 = 5, <10 ng/g dw (0-0.25, 1-2 m)
SC-42 = <10 ng/g dw (0-0.25,1-2, 3^1 m)
Area 4:
SC-33 = <20 ng/g dw (0-1 m)
SC-34 =13 ng/g dw (1-2 m)
SC-35 = <2 ng/g dw (0-1 m)
SC-36 = <4 ng/g dw (0.3-1 m)
SC-37 = 43 ng/g dw (0.1-1.1 m)
SC-37b = 11, 6, <10 ng/g dw (0-0.25, 1-
1.5, 3—4 m)
SC-38 = 35, 2 ng/g dw (0.25-1, 2-3 m)
Area 5:
SC-10 = <2 ng/g dw (0-0.25, 0.25-0.5, 0.5-
0.75, 0.75-1,1-1.5,1.5-2, 2-2.5, 2.5-3,
3-3.5 m)
SC-11 = <2 ng/g dw (0-0.25, 0.25-0.5, 0.5-
0.75,0.75-1,1-1.5,1.5-2, 2-2.5 m)
D-115
-------
FINAL
2024
Study
Location
Site Details
Results
SC-12 = <2 ng/g dw (0-0.25, 0.25-0.5, 0.5-
0.75 m)
Area 6:
SC-21 = 148, 290, 273, 278, 98 ng/g dw (0-
0.25, 0.25-1, 2-3, 8-9, 13-15 m)
SC-22 = 8, 11, 6, 7 ng/g dw (0-0.25, 0.25-1,
1-2, 3-4 m)
SC-23 = 59, 29, 29,16 ng/g dw (0-0.25,
0.25-1, 1-2, 3^1 m)
SC-24 = 85, 31, 43 ng/g dw (0-0.25, 1-2, 3-
4 m)
SC-25 = 7, 26 ng/g dw (0-0.25,1.5-2 m)
SC-26 = 4, 3 ng/g dw (2-3,4-5 m)
(LOQ = 2 ng/g dw)
Skaaretal. (2019)
Norway (Ny-Alesund)
Samples collected in June 2016 in and around the
international research facilities (Ny-Alesund) near
local firefighting training site. Background soil
samples were collected at representative locations.
Background: n = 8, DF 0%
Contaminated: n = 2, DFa 100%, meana
(range) = 29.42 (13.82^15.02) ng/g dw
(IDL = 0.003 ng; LOD = 0.001 ng/g dw; LOQ
= 0.002 ng/g dw)
*Values reported for sum of branched and
linear PFHxS isomers
*Table 1 and Table S2 reported a total of nine
samples across background and contaminated
sites; however, Tables Sll and SI3 report a
total of ten samples (two contaminated sites
from Table Sll and eight background sites
from Table SI3
Hale et al. (2017)
Norway (Gardermoen)
Samples collected in June 2015 from six locations
around a firefighting training facility west of the Oslo
airport site. Samples were taken at 0-1 m, 1-2 m, 2-3
m, and 3 to groundwater table level (which was in all
cases above 4 m). Facility was established in 1989 and
AFFF was used extensively. AFFF containing PFOS
was banned at the facility in 2007 and a complete ban
on organofluorine AFFF was enforced in 2011. The
soil is known to be contaminated with a range of
perfluorinated compounds.
n = 22 (across all depths), DF 36%, range =
3.0-25.3 ng/g
(LOD =1 ng/g)
*Range reported for detects
*The DF and range extracted are reported in
the results (Section 3.1); however, Table S2 of
the individual sample data show all
concentrations ranging from <1.8 to <2.5 ng/g
Filipovic et al. (2015)
Sweden (Stockholm)
Five locations at a closed air force base F18 in
Tullinge Riksten, 19 km south of Stockholm city
Intermediate soil depot:
n = 10, DF 80%), range = <0.02-3.1 ng/g dw
D-116
-------
FINAL
2024
Study
Location
Site Details
Results
center, where AFFFs were used. Samples collected in
two sampling campaigns in December 2011 and May
2012. The air force base was formally demobilized in
1986 but continued to be used as an air force school
for combat command and air surveillance until 1994.
Of note, the air force base encountered numerous
accidents and incidents during the transfer from
propeller era to the jet engine era, including planes
crashing upon takeoff and landing, fire incidents,
accidental dispersion of jet engine starting fuel. The
area was sold to a land developer in 1996 and is in the
process of being transformed into a municipal area. Of
the five sites samples, the intermediate soil depot site
(sampling depth 0-3 m) and the soil depot site
(sampling depth 0-3 m) were assembled by the land
developing corporation; the J34 Hawker Hunter site
was the collision site of two J34 Hawker Hunter
aircrafts in the 1960s, resulting in huge amounts of
firefighting foam being dispersed (sampling depth
0.5-1.0 m); the main firefighting training facility was
intensively used between 1970-1985 and sparsely
used between 1985-1994 (sampling depth 0.5-1.0 m);
and the old fire station was used to store AFFF stock
solution (sampling depth 0.5-1.0 m).
J34 Hawker Hunter site:
n = 5, DF 0%
Main firefighting training facility:
n = 15, DF 100%, meana (range) = 3.9 (0.1-
21.3)ng/g dw
Old fire station:
n = 5, DF 100%, meana (range) = 3.7(1.6-
6) ng/g dw
Soil depot:
n = 10, DF 40%, range = <0.02-0.33 ng/g
dw
(MDL = <0.02-<0.1 ng/g dw)
Harrad et al. (2020)
Ireland (multiple cities)
Samples collected from ten municipal solid waste
landfills upwind and downwind at each site between
November 2018 and January 2019. At each
upwind/downwind location, nine sub-samples of soil
were taken in a "W" formation. Samples were
collected from the same areas as air samples and were
taken within the boundaries of the landfill operational
facility. Soil used as capping on landfill cells was not
sampled to ensure soil samples were not collected
from soil placed after landfill operations ceased and
that farming activities would not influence
concentrations found. Waste accepted by the landfills
included: municipal solid waste, industrial (non-
hazardous) waste, construction & demolition, and
biomedical waste.
Downwind:
n = 9, DFa 11%), mean, median (range) =
0.00077, <0.001 (<0.001-0.0029) ng/g dw
Upwind:
n = 7, DFa 57%o, mean, median (range) =
0.0018, 0.0023 (<0.001-0.0037) ng/g dw
(LOD = <0.001 ng/g dw)
*Non-detects replaced by Vi LOD
*Soil samples from three upwind locations
and one downwind location destroyed in
transit from field to laboratory
Grannestad et al. (2019)
Norway (Granasen, Jonsvatnet)
Upper layer soil samples (3-10 cm in depth) collected
in June 2017 and 2018 from Granasen (skiing area)
Reference area: n = 10, DF 0%
Skiing area: n = 10, DF 0%
D-117
-------
FINAL
2024
Study
Location
Site Details
Results
and Jonsvatnet (reference site). Five samples per year
were analyzed for each site. Located 10 km from
Trondheim city center, Granasen is the main arena for
winter sports in Trondheim and hosts an annual ski
jumping World Cup event and regional, national, and
international competitions in cross-country skiing.
Located 15 km away from Trondheim city center,
Jonsvatnet is a natural forest area not used for ski-
sports and is in the vicinity of an ecological farm next
to Lake Jonsvatnet. The two study areas have similar
vegetation.
(LOQ = 0.032 ng/g dw)
*For the reference area, Table S2 reported a
DF = 10% but a range, mean, and standard
deviation of
-------
FINAL
2024
D.3.7. Sediment
When released into water, PFHxS is not expected to adsorb to suspended solids and sediments
based on its physico-chemical properties (NCBI, 2022a). However, PFHxS was detected in 28%
of sediment samples collected along the Mississippi River shoreline in the vicinity of a PFAS
production facility in Minnesota, at a maximum concentration of 11.5 ng/g (NCBI, 2022a;
ATSDR, 2021; 3M, 2007). PFHxS was also detected in 96% of sediment samples collected from
two coves of the Mississippi River (East and West coves) located at the eastern and western ends
of the PFAS production facility, at a maximum concentration of 126 ng/g (ATSDR, 2021; 3M,
2007). PFHxS was not detected in any Mississippi River transect sediment samples (collected at
points crossing the river near the PFAS facility) (ATSDR, 2021; 3M, 2007). PFHxS has been
detected in sediment core samples collected from three Canadian Arctic lakes in 2003 and 2005
at concentrations ranging from approximately 1 ng/g to 10 ng/g (NCBI, 2022a; Stock et al.,
2007).
D.4. Recommended RSC
The EPA followed the Exposure Decision Tree approach to determine the RSC for PFHxS
(USEPA, 2000b). The EPA first identified the U.S. general population as the population of
concern (Box 1; see Section 2.4.2). Second, the EPA identified several relevant PFHxS
exposures and pathways (Box 2), including dietary consumption, incidental oral consumption via
exposure to dust, consumer products, and soil, dermal exposure via soil, consumer products, and
dust, and respiration via ambient air. Several of these may be potentially significant exposure
sources. Third, the EPA determined that there was not adequate quantitative data to describe the
central tendencies and high-end estimates for all of the potentially significant sources (Box 3).
For example, studies from Canada and Europe indicate that indoor air may be a significant
source of exposure to PFHxS. At the time of the literature search, the EPA was unable to identify
studies assessing PFHxS concentrations in indoor air samples from the U.S. and therefore, the
agency does not have adequate quantitative data to describe the central tendency and high-end
estimate of exposure for this potentially significant source in the U.S. population. However, the
agency determined there were sufficient data, physical/chemical property information, fate and
transport information, and/or generalized information available to characterize the likelihood of
exposure to relevant sources (Box 4). Notably, based on the studies summarized in the sections
above, there are significant known or potential uses/sources of PFHxS other than drinking water
(Box 6), though there is not information available on each source to make a characterization of
exposure (Box 8A). For example, there are several studies from the U.S. indicating that PFHxS
may occur in multiple food products (e.g., eggs, seafood, meats, vegetables, fruit) and consumer
products (e.g., building materials, clothing, furniture). However, the majority of studies
examined very few samples (i.e., n=l-5) of each type of media. Therefore, it is not possible to
determine which source, if any, can be considered major or minor contributors to total PFHxS
exposure. Given these considerations, following recommendations of the Exposure Decision
Tree (USEPA, 2000b), the EPA recommends an RSC of 20% (0.20) for PFHxS.
D-119
------- |