c EDA
WtrM
April 2024
EP A-815-R-24-010
Maximum Contaminant Level Goals for
Perfluorooctanoic Acid (PFOA) and Perfluorooctane
Sulfonic Acid (PFOS) in Drinking Water
-------
Maximum Contaminant Level Goals for Perfluorooctanoic Acid (PFOA) and
Perfluorooctane Sulfonic Acid (PFOS) in Drinking Water
Prepared by:
U.S. Environmental Protection Agency
Office of Water (4304T)
Health and Ecological Criteria Division
Washington, DC 20460
EPA Document Number: EPA-815-R-24-010
2024
ii
-------
Disclaimer:
This document is a draft for governmental staff review purposes only. This information is
distributed solely for the purpose of review within the federal government. It has not been
formally disseminated by the U.S. Environmental Protection Agency. It does not represent and
should not be construed to represent any agency determination or policy. Mention of trade names
or commercial products does not constitute endorsement or recommendation for use.
-------
Acknowledgments
This document was prepared by the Health and Ecological Criteria Division, Office of Science
and Technology, Office of Water (OW) of the U.S. Environmental Protection Agency (EPA).
The agency gratefully acknowledges the valuable contributions of EPA scientists from the OW,
Office of Research and Development (ORD), the Office of Children's Health Protection
(OCHP), and the Office of Land and Emergency Management (OLEM). OW authors of the
document include Brittany Jacobs; Casey Lindberg; Carlye Austin; Kelly Cunningham; Barbara
Soares; Ruth Etzel; and Colleen Flaherty. ORD authors of the document include J. Michael
Wright; Elizabeth Radke; Michael Dzierlenga; Todd Zurlinden; Jacqueline Weinberger; Thomas
Bateson; Hongyu Ru; and Kelly Garcia. OCHP authors of the document include Chris
Brinkerhoff; and Greg Miller (formerly OW). EPA scientists who provided valuable
contributions to the development of the document from OW include Adrienne Keel; Joyce
Donohue (now retired); Amanda Jarvis; James R. Justice; from ORD include Timothy Buckley;
Allen Davis; Peter Egeghy; Elaine Cohen Hubal; Pamela Noyes; Kathleen Newhouse; Ingrid
Druwe; Michelle Angrish; Christopher Lau; Catherine Gibbons; and Paul Schlosser; and from
OLEM includes Stiven Foster. Additional contributions to draft document review from managers
and other scientific experts, including the ORD Toxicity Pathways Workgroup and experts from
the Office of Chemical Safety and Pollution Prevention (OSCPP), are greatly appreciated. The
agency gratefully acknowledges the valuable management oversight and review provided by
Elizabeth Behl (OW); Jamie Strong (formerly OW; currently ORD); Susan Euling (OW);
Kristina Thayer (ORD); Andrew Kraft (ORD); Viktor Morozov (ORD); Vicki Soto (ORD); and
Garland Waleko (ORD).
The systematic review work included in this assessment was prepared in collaboration with ICF
under the U.S. EPA Contracts EP-C-16-011 (Work Assignment Nos. 4-16 and 5-16) and PR-
OW-21-00612 (TO-0060). ICF authors serving as the toxicology and epidemiology technical
leads were Samantha Snow and Sorina Eftim. ICF and subcontractor authors of the assessment
include Kezia Addo; Barrett Allen; Robyn Blain; Lauren Browning; Grace Chappell; Meredith
demons; Jonathan Cohen; Grace Cooney; Ryan Cronk; Katherine Duke; Hannah Eglinton;
Zhenyu Gan; Sagi Enicole Gillera; Rebecca Gray; Joanna Greig; Samantha Goodman; Anthony
Hannani; Samantha Hall; Jessica Jimenez; Anna Kolanowski; Madison Lee; Cynthia Lin;
Alexander Lindahl; Nathan Lothrop; Melissa Miller; Rachel O'Neal; Ashley Peppriell; Mia
Peng; Lisa Prince; Johanna Rochester; Courtney Rosenthal; Amanda Ross; Karen Setty; Sheerin
Shirajan; Raquel Silva; Jenna Sprowles; Wren Tracy; Joanne Trgovcich; Janielle Vidal;
Maricruz Zarco; and Pradeep Raj an (subcontractor).
iv
-------
Contents
Acknowledgments iv
Abbreviations and Acronyms vi
Introduction 1
Background and Purpose 1
Occurrence of PFOA and PFOS in Drinking Water 1
Methods 3
Approach for Deriving an MCLG 3
Summary of the EPA's Systematic Review of the Health Effects Data for PFOA and
PFOS 5
Literature Search 5
Literature Screening 5
Study Quality Evaluation for Epidemiological Studies and Animal Toxicological
Studies 6
Data Extraction 6
Approach for Determining the Cancer Classification 7
Cancer Weight Of Evidence for Carcinogenicity and Cancer Classification 8
PFOA 8
Summary of the Weight of Evidence 8
Cancer Classification 11
PFOS 17
Summary of the Weight of Evidence 17
Cancer Classification 20
MCLG Derivation 26
PFOA 26
PFOS 26
v
-------
Tables
Table A. Comparison of the PFOA Carcinogenicity Database with the Likely Cancer
Descriptor as Outlined in the Guidelines for Carcinogen Risk Assessment
(USEPA, 2005) 13
Table B. Comparison of the PFOA Carcinogenicity Database with Cancer Descriptors as
Outlined in the Guidelines for Carcinogen Risk Assessment (USEPA, 2005) 15
Table C. Comparison of the PFOS Carcinogenicity Database with the Likely Cancer
Descriptor as Outlined in the Guidelines for Carcinogen Risk Assessment
(USEPA, 2005) 21
Table D. Comparison of the PFOS Carcinogenicity Database with Cancer Descriptors as
Outlined in the Guidelines for Carcinogen Risk Assessment (USEPA, 2005) 24
vi
-------
Abbreviations and Acronyms
CalEPA
California
OR
odds ratio
Environmental
ORD
Office of Research and
Protection Agency
Development
CAR
constitutive androstane
OSCPP
Office of Chemical
receptor
Safety and Pollution
DNA
deoxyribonucleic acid
Prevention
DWI-BW
body weight-adjusted
OW
Office of Water
drinking water intake
PACT
pancreatic acinar cell
EPA
U.S. Environmental
tumors
Protection Agency
PECO
Population, Exposure,
ER
estrogen receptor
Comparator, and
HAWC
Health Assessment
Outcome
Workplace Collaboration
PFAS
per- and polyfluoroalkyl
HERO
Health and
substances
Environmental Research
PFOA
perfluorooctanoic acid
Online
PFOS
perfluorooctane sulfonic
HESD
health effects support
acid
documents
PLCO
prostate, lung, colorectal,
IARC
International Agency for
and ovarian
Research on Cancer
POD
point of departure
IRIS
Integrated Risk
PPAR
peroxisome proliferator-
Information System
activated receptor
LCT
Leydig cell tumors
PPRTV
Provisional Peer
MCLG
Maximum Contaminant
Reviewed Toxicity
Level Goal
Value
MOA
mode of action
PWS
public water systems
NCI
National Cancer Institute
PXR
pregnane X receptor
NHANES
National Health and
QA
quality assurance
Nutrition Examination
RCC
renal cell carcinoma
Survey
RfD
reference dose
NJDWQI
New Jersey Drinking
RSC
relative source
Water Quality Institute
contribution
NPDWR
National Primary
SAB
Science Advisory Board
Drinking Water
SDWA
Safe Drinking Water Act
Regulation
SEM
systematic evidence map
NTP
National Toxicology
UCMR
Unregulated
Program
Contaminant Monitoring
OCHP
Office of Children's
Rule
Health Protection
OLEM
Office of Land and
Emergency Management
vii
-------
1 Introduction
1.1 Background and Purpose
Section 1412(a)(3) of the Safe Drinking Water Act (SDWA) requires the Administrator of the
U.S. Environmental Protection Agency (EPA) to finalize a Maximum Contaminant Level Goal
(MCLG) simultaneously with the publication of a 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." 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 for drinking water
contaminants include health effects data, toxicity values, cancer classifications, 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 relevant health effects information,
and to describe the derivation of the EPA's final individual MCLGs for perfluorooctanoic acid
(PFOA) and perfluorooctane sulfonic acid (PFOS) used in the Per- and Polyfluoroalkyl
Substances (PFAS) NPDWR (USEPA, 2024f). The individual MCLGs are based on the final
toxicity assessments for PFOA or PFOS, which were developed and finalized to support the
PFAS NPDWR (USEPA, 2024d, e). The toxicity assessments underwent both external peer
review through the EPA Science Advisory Board (USEPA OOW, 2023) and public comment
(USEPA, 2024c). This document summarizes the key elements (e.g., cancer classifications) that
the agency used as the basis for determining the individual MCLGs for PFOA and PFOS and
provides the final MCLGs for PFOA and PFOS used in the PFAS NPDWR. It is not intended to
be an exhaustive description of all health effects information or quantitative analyses provided in
the final human health toxicity assessments (USEPA, 2024d, e), nor is it a drinking water health
advisory.
1.2 Occurrence of PFOA and PFOS in Drinking Water
The EPA uses the Unregulated Contaminant Monitoring Rule (UCMR) to collect data for
contaminants that are suspected to be present in drinking water and do not have health-based
standards set under the SDWA. Under the UCMR, drinking water is monitored from public
water systems (PWSs), specifically community water systems and non-transient non-community
water systems. UCMR improves the EPA's understanding of the frequency and concentrations of
contaminants of concern occurring in the nation's drinking water systems. The first four UCMRs
collected data from a census of large water systems (serving more than 10,000 people) and from
a statistically representative sample of small water systems (serving 10,000 or fewer people).
UCMR 3 monitoring occurred between 2013 and 2015 and is currently the most comprehensive
nationally representative finished water dataset for PFOA and PFOS (USEPA, 2024f, g). Under
UCMR 3, 36,972 samples from 4,920 PWSs were analyzed. PFOA was found above the UCMR
3 minimum reporting level (20 ng/L) in 379 samples at 117 systems serving a population of
1
-------
approximately 7.6 million people located in 28 states, Tribes, or U.S. territories (USEPA, 2024f,
g). PFOS was found in 292 samples at 95 systems above the UCMR 3 minimum reporting level
(40 ng/L) (USEPA, 2024f, g). These systems serve a population of approximately 10.4 million
people located in 28 states, Tribes, or U.S. territories (USEPA, 2024f, g).
More recent state data were collected using newer EPA-approved analytical methods and some
state results reflect lower reporting limits than those in the UCMR 3. State data are available
from 32 states: Alabama, Arizona, California, Colorado, Delaware, Georgia, Idaho, Illinois,
Indiana, Iowa, Kentucky, Maine, Maryland, Massachusetts, Michigan, Minnesota, Missouri,
New Hampshire, New Jersey, New Mexico, New York, North Carolina, North Dakota, Ohio,
Oregon, Pennsylvania, South Carolina, Tennessee, Vermont, Virginia, West Virginia, and
Wisconsin (USEPA, 2024f, g). State results show continued occurrence of PFOA and PFOS in
multiple geographic locations. These data also show PFOA and PFOS occurrence at lower
concentrations and significantly greater frequencies than were measured under the UCMR 3,
likely because the more recent monitoring was able to rely on more sensitive analytical methods
(USEPA, 2024f, g). More than one-third of states that conducted non-targeted monitoring
detected PFOA and/or PFOS at more than 25% of systems (USEPA, 2024f, g). Among the
detections, PFOA concentrations ranged from 0.21 to 650 ng/L with a range of median
concentrations from 1.27 to 5.61 ng/L, and PFOS concentrations ranged from 0.24 to 650 ng/L
with a range of median concentrations from 1.21 to 12.1 ng/L (USEPA, 2024f, g). Monitoring
data for PFOA and PFOS from states that conducted targeted monitoring efforts, including 15
states, demonstrate results consistent with the non-targeted state monitoring. Within the 20 states
that conducted non-targeted monitoring there are 1,260 systems with results above 4.0 ng/L and
1,577 systems with results above 4.0 ng/L (USEPA, 2024f, g). These systems serve populations
of 12.5 and 14.4 million people, respectively. Monitoring data for PFOA and PFOS from states
that conducted targeted sampling efforts showed additional systems exceeding 4 ng/L (USEPA,
2024f, g).
Finally, the fifth UCMR (UCMR 5) was published in December 2021 and requires sample
collection and analysis for 29 PFAS, including PFOA and PFOS, between January 2023 and
December 2025 using drinking water analytical methods developed by the EPA (USEPA,
2021b). The UCMR 5 defined the minimum reporting level at 4 ng/L for PFOA and PFOS using
EPA Method 533 which is lower than the 20 and 40 ng/L, respectively, used in UCMR 3 with
EPA Method 537 (USEPA, 2021b). Therefore, UCMR 5 will be able to provide nationally
representative occurrence data for PFOA and PFOS at lower detection concentrations. While the
complete UCMR 5 dataset is not currently available, the small subset of data released (7% of the
total results that the EPA expects to receive) as of July 2023 is consistent with the results of
UCMR 3 and the state data described above (USEPA, 2024f, g).
2
-------
2 Methods
2.1 Approach for Deriving an MCLG
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." Consistent with SDWA Section 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." To establish the MCLG, the EPA assesses the available data examining cancer and
noncancer health effects associated with oral exposure to the contaminant. For known or likely
linear carcinogenic contaminants, where there is a proportional relationship between dose and
carcinogenicity at low concentrations or where 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 long-standing practice of establishing the MCLG at zero (see USEPA, (2001, 2000b,
1998); see S. Rep. No. 169, 104th Cong., 1st Sess. (1995) at 3). This is called the linear default
extrapolation approach and ensures that the MCLG is set at a level where there are no anticipated
adverse health effects, allowing for an adequate margin of safety. For nonlinear carcinogenic
contaminants, contaminants that are designated as Suggestive Human Carcinogens (USEPA,
2005), and non-carcinogenic contaminants, the EPA typically establishes the MCLG based on a
noncancer RfD. 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. A nonlinear carcinogen 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.
A noncancer MCLG is designed to be protective of noncancer effects over a lifetime of exposure
with an adequate margin of safety, including for sensitive populations and lifestages, 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 value, body weight-adjusted drinking water intake (DWI-BW), and a
relative source contribution (RSC), as presented in the equation below:
RfD = chronic reference dose - an estimate (with uncertainty spanning perhaps an order
of magnitude) of daily oral exposure of the human population to a substance that is likely
to be without an appreciable risk of deleterious effects during a lifetime. The RfD is equal
to a point of departure (POD) human equivalence dose (HED) (PODhed) divided by a
composite uncertainty factor.
DWI-BW = An exposure factor for the 90th percentile body weight-adjusted drinking
water intake value for the identified population or lifestage, in units of liters of water
Where:
3
-------
consumed per kilogram body weight per day (L/kg bw-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, 2019) provides the most
up-to-date DWI-BWs for various populations or lifestages within the U.S. general
population for which there are publicly available, peer-reviewed data such as from the
National Health and Nutrition Examination Survey (NHANES).
RSC = relative source contribution - the percentage of total exposure attributed to
drinking water sources (USEPA, 2000a), 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 (e.g., MCLG value), 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 published in EPA's Methodology for Deriving Ambient Water Quality
Criteria for the Protection of Human Health (USEPA, 2000a).
Because the cancer classification of the chemical determines which approach that the EPA used
to derive the MCLGs, the EPA summarizes the carcinogenic data evaluated for cancer
classification selection below. The EPA followed a transparent systematic review process to
evaluate the best available science and to determine the weight of evidence for carcinogenicity
and the cancer classifications for PFOA and PFOS, individually, according to agency guidance
(USEPA, 2005). Following this guidance, and as detailed below, the EPA determined that PFOA
and PFOS are each classified as Likely to Be Carcinogenic to Humans based on sufficient
evidence of carcinogenicity in the available human and animal studies. The EPA also determined
that a linear default extrapolation approach is appropriate for PFOA and PFOS as there is no
available evidence demonstrating a threshold level of exposure below which there is no
appreciable cancer risk for either compound (USEPA, 2016c, 2005). Therefore, the EPA
concluded that there is no known threshold for carcinogenicity. Because of these cancer
conclusions the noncancer health effects that the EPA identified as hazards in the draft toxicity
assessments (e.g., decreased immune response in children, increased serum alanine
aminotransferase (ALT), decreased birth weight, increased serum total cholesterol) are not the
basis for the final MCLGs and are not, therefore, described in this document. Details related to
the noncancer effects associated with PFOA and PFOS can be found in the final human health
toxicity assessments for PFOA and PFOS (USEPA, 2024d, e).
4
-------
2.2 Summary of the EPA's Systematic Review of the Health
Effects Data for PFOA and PFOS
The EPA conducted the systematic review of the cancer health effects data for PFOA and PFOS
consistent with the methods described in the EPA ORD Staff Handbook for Developing IRIS
Assessments (USEPA, 2022a) (hereafter referred to as the Integrated Risk Information System
(IRIS) Handbook) and a companion publication (Thayer et al., 2022). The agency's systematic
review incorporated and considered studies that are consistent with the SDWA mandate to "use
(i) the best available, peer-reviewed science and supporting studies conducted in accordance with
sound and objective scientific practices; and (ii) data collected by accepted methods or best
available methods (if the reliability of the method and the nature of the decision justifies use of
the data)" (SDWA(b)(3)(A)). Full details of the systematic review methodology can be found in
Appendix A of the toxicity assessments for PFOA and PFOS (USEPA, 2024a, b).
2.2.1 Literature Search
The EPA assembled an inventory of epidemiological, animal toxicological, and mechanistic
studies based on three data sources: 1) literature published from 2014 through 2019 and then
updated throughout the course of this review (i.e., through February 3, 2022) identified via
literature searches of a variety of publicly available scientific literature databases; 2) literature
identified via other sources (e.g., searches of the gray literature and studies shared with the EPA
by the Science Advisory Board (SAB)); and 3) literature identified in the EPA's 2016 health
effects support documents (HESDs) for PFOA and PFOS (USEPA, 2016a, b). Additionally, the
EPA identified studies from a supplemental literature search conducted in February 2023 as well
as studies received through public comments and included those studies that were deemed to
have the potential to quantitatively affect the final toxicity values (i.e., RfDs and cancer slope
factors) or MCLGs for PFOA or PFOS in a significant way (i.e., by an order of magnitude or
more). For additional details related to the literature included, please refer to Sections 2.1 and 3.1
in the final human health toxicity assessments USEPA (2024e); and USEPA (2024b) as well as
Section A. 1.5 in USEPA (2024d) and USEPA (2024a).
2.2.2 Literature Screening
The EPA used populations, exposures, comparators, and outcomes (PECO) criteria to screen all
of the literature identified from the literature sources outlined above in order to prioritize the
dose-response studies for dose-response assessment and to identify studies containing
supplemental information such as mechanistic studies that could inform the mode of action
analysis.
Consistent with protocols outlined in the IRIS Handbook (USEPA, 2022b), studies identified in
the literature searches and stored in HERO were imported into the Swift-Review software
platform and the software was subsequently used to identify those studies most likely to be
relevant to human health risk assessment. Studies captured then underwent title and abstract
screening by at least two reviewers using screening tools consistent with the IRIS Handbook
(USEPA (2022b); DistillerSR or SWIFT ActiveScreener software), and studies that passed this
screening underwent full-text review. Dose-response studies that met PECO inclusion criteria
following both title and abstract screening and full-text review underwent study quality
evaluation as described below. Studies tagged as supplemental and containing potentially
5
-------
relevant mechanistic data following title and abstract and full-text level screening underwent
further screening using mechanistic-specific PECO criteria, and those deemed relevant
underwent light data extraction of key study elements (e.g., mechanistic endpoints evaluated,
dose levels tested). Supplemental studies that were identified as mechanistic via screening did
not undergo study quality evaluation. For additional details related to literature screening, please
refer to Section A. 1.8 in the Final Human Health Toxicity Assessment for PFOA (USEPA,
2024e) and PFOS (USEPA, 2024d).
2.2.3 Study Quality Evaluation for Epidemiological Studies
and Animal Toxicological Studies
For study quality evaluation of the PECO-relevant human epidemiological and animal
toxicological studies identified for cancer, two or more quality assurance (QA) reviewers,
working independently, assigned ratings about the reliability of study results {good, adequate,
deficient (or "not reported"), or critically deficient) for different evaluation domains consistent
with the IRIS Handbook (USEPA, 2022b). These study quality evaluation domains are listed
below and details about the domains, including prompting questions and suggested
considerations, are described in Section A. 1.9 in the Final Human Health Toxicity Assessment
for PFOA (USEPA, 2024e) and PFOS (USEPA, 2024d).
Epidemiological study quality evaluation domains: participant selection; exposure
measurement criteria; outcome ascertainment; potential confounding; analysis; selective
reporting; and study sensitivity.
Animal toxicological study quality evaluation domains: reporting; allocation;
observational bias/blinding; confounding/variable control; reporting and attrition bias;
chemical administration and characterization; exposure timing, frequency, and duration;
endpoint sensitivity and specificity; and results presentation.
The independent reviewers performed study evaluations using a structured platform housed
within the EPA's Health Assessment Workplace Collaboration (HAWC;
https://hawcproi ect.org/Y Once the individual domains were rated, reviewers independently
evaluated the identified strengths and limitations of each study to reach an overall classification
on study confidence of high, medium, low, or uninformative for each relevant endpoint evaluated
in the study. A study can be given an overall mixed confidence classification if different PECO-
relevant endpoints within the study receive different confidence ratings (e.g., medium and low
confidence classifications). All study evaluations are publicly available in HAWC at
https://hawc.epa.eov/stiidy/assessment/100500248/. For additional details related to study
evaluation, please refer to Section A. 1.9 in the Final Human Health Toxicity Assessment for
PFOA (USEPA, 2024e) and PFOS (USEPA, 2024d).
2.2.4 Data Extraction
Data extraction was conducted for all relevant human epidemiological and animal toxicological
studies determined to be of medium and high confidence based on study quality evaluation. Data
were also extracted from low confidence epidemiological studies when data were limited for a
health outcome or when there was a notable effect, consistent with the IRIS Handbook (USEPA,
2022b). Data extracted from low confidence studies was considered qualitatively only. Studies
evaluated as being uninformative were not considered further and therefore did not undergo data
6
-------
extraction. All health endpoints were considered for extraction, regardless of the magnitude of
effect or statistical significance of the response relative to the control group. The level of detail
in data extractions for different endpoints within a study could differ based on how the data were
presented for each outcome (i.e., ranging from a narrative to a full extraction of dose-response
effect size information).
Extractions were conducted using DistillerSR for epidemiological studies and HAWC for animal
toxicological studies. An initial reviewer conducted the extraction, followed by an independent
QA review by a second reviewer who confirmed accuracy and edited/corrected the extracted data
as needed. Discrepancies in data extraction were resolved by discussion and confirmation within
the extraction team.
Data extracted from epidemiology studies included population, study design, year of data
collection, exposure measurement, and quantitative analyses of the data from statistical models.
Results extracted from statistical models performed in the studies included the health effect
category, endpoint measured, sample size, description of effect estimate, covariates, and model
comments. Data extracted from animal toxicological studies included information on the
experimental design and exposure duration, species and number of animals tested, dosing
regime, and endpoints measured. For additional details related to data extraction, please refer to
Sections A. 1.10 and A. 1.11 in the Final Human Health Toxicity Assessment for PFOA (USEPA,
2024e) and PFOS (USEPA, 2024d).
2.3 Approach for Determining the Cancer Classification
In accordance with the EPA's 2005 Guidelines for Carcinogen Risk Assessment, a descriptive
weight of evidence expert judgment is made, based on all available animal, human, and
mechanistic data, as to the likelihood that a contaminant is a human carcinogen and the
conditions under which the carcinogenic effects may be expressed (USEPA, 2005). A narrative
is developed to provide a complete description of the weight of evidence evaluation and
conditions of carcinogenicity. The potential carcinogenicity descriptors presented in the EPA's
2005 guidelines are:
Carcinogenic to Humans
Likely to Be Carcinogenic to Humans
Suggestive Evidence of Carcinogenic Potential
Inadequate Information to Assess Carcinogenic Potential
Not Likely to Be Carcinogenic to Humans
More than one carcinogenicity descriptor can be applied in cases when a chemical's carcinogenic
effects differ by dose, exposure route, or mode of action (MOA)1. MO A information informs
both the qualitative and quantitative aspects of the assessment, including the human relevance of
tumors observed in animals. The MOA analysis must be conducted separately for each target
organ/tissue type according to EPA guidance (USEPA, 2005).
'MOA is defined as a sequence of key events and processes, starting with interaction of an agent with a cell, proceeding through
operational and anatomical changes, and resulting in cancer formation. It is contrasted with "mechanism of action," which
implies a more detailed understanding and description of events.
7
-------
3 Cancer Weight of Evidence for Carcinogenicity
and Cancer Classification
3.1 PFOA
3.1.1 Summary of the Weight of Evidence
The carcinogenicity of PFOA has been documented in both epidemiological and animal
toxicological studies. The evidence from medium quality epidemiological studies is primarily
based on the incidence of kidney and testicular cancer, as well as some evidence of increased
breast cancer incidence in susceptible subpopulations. Other cancer types have been observed in
humans, although the evidence for these is generally limited to low confidence studies. The
evidence of carcinogenicity in animal models is provided in three high or medium confidence
chronic oral animal bioassays in Sprague-Dawley rats which together identified neoplastic
lesions of the liver, pancreas, and testes. The available mechanistic data suggest that multiple
MO As could play a role in the renal, testicular, pancreatic, and hepatic tumorigenesis associated
with PFOA exposure in human populations as well as animal models.
The strongest evidence of an association between PFOA exposure and cancer in human
populations is from studies of kidney cancer. Two medium confidence studies of the C8 Health
Project population reported positive associations between PFOA levels (mean at enrollment
0.024 |ig/mL) and kidney cancer among the residents living near the DuPont plant in
Parkersburg, West Virginia (Barry et al., 2013; Vieira et al., 2013). Vieira et al. (2013) reported
elevated risk of kidney cancer in residents of the Little Hocking water district of Ohio (OR: 1.7,
95% CI: 0.4, 3.3; n = 10) and the Tuppers Plains water district of Ohio (OR: 2.0, 95% CI: 1.3,
3.1; n = 23). Barry et al. (2013) extended this work, and found increased risk of kidney cancer
(HR: 1.10, 95% CI: 0.98, 1.24; n = 105), though the levels did not reach statistical significance.
The high-exposure occupational study by Steenland and Woskie (2012) evaluated kidney cancer
mortality in workers from West Virginia and observed significant elevated risk of kidney cancer
death in the highest exposure quartile. As part of the C8 Health Project, the C8 Science Panel
(2012) concluded a probable link between PFOA exposure and kidney cancer (Steenland et al.,
2020).
The findings of another recently published medium quality study add support to the previous
evidence of an association between PFOA and kidney cancer (Shearer et al., 2021). Shearer et al.
(2021) is a multi-center case-control study nested within the National Cancer Institute (NCI)
Prostate, Lung, Colorectal and Ovarian (PLCO) cancer screening trial (n = 326). The authors
reported a statistically significant increase in risk of renal cell carcinoma (RCC) with pre-
diagnostic serum levels of PFOA (OR = 2.63; 95% CI: 1.33, 5.20 for the highest vs. lowest
quartiles; p-trend = 0.007, or per doubling of PFOA: OR: 1.71; 95% CI: 1.23, 2.37). The
association remained significant in analyses on a per doubling increase in PFOA after adjusting
for other PFAS. The increase in the highest exposure quartile remained and the magnitude was
similar (i.e., OR = 2.63 without adjusting for other PFAS vs. 2.19 after adjusting for other
PFAS), but it was no longer statistically significant. Statistically significant increased odds of
RCC were observed in a subgroup of participants ages 55-59 years, and in men and in women,
analyzed separately. A recent critical review and meta-analysis of the epidemiological literature
concluded that there was an increased risk for kidney tumors (16%) for every 10 ng/mL increase
8
-------
in serum PFOA (Bartell and Vieira, 2021). Although the authors concluded that the associations
were likely causal, they noted the limited number of studies and therefore, additional studies with
larger cohorts would strengthen the conclusion. Taken together, the recent pooled analysis of the
NCI nested case-control study (Shearer et al., 2021) of 324 cases and controls and the C8
Science Panel Study (Barry et al., 2013) of 103 cases and 511 controls provide evidence of
concordance in kidney cancer findings from studies of the general population and studies of
high-exposure communities (Steenland et al., 2022). CalEPA (2021) similarly concluded,
"[t]here is evidence from epidemiologic studies that exposure to PFOA increases the risk of
kidney cancer."
There is also evidence of associations between PFOA serum concentrations and testicular cancer
in humans, though no new epidemiological studies reporting these associations have been
published since the studies described in the EPA's 2016 PFOA HESD (USEPA, 2016b). Similar
to their results for kidney cancer, Vieira et al. (2013) reported an increased adjusted OR for
testicular cancer (OR: 5.1, 95% CI: 1.6, 15.6; n = 8) in residents of the Little Hocking water
district of Ohio. Barry et al. (2013) also found significantly increased testicular cancer risk with
an increase in estimated cumulative PFOA serum levels (HR: 1.34, 95% CI: 1.00, 1.79; n = 17).
The C8 Science Panel (2012) concluded that a probable link also exists between PFOA exposure
and testicular cancer (Steenland et al., 2020). A recent critical review and meta-analysis of the
epidemiological literature concluded that there was an increased risk for testicular tumors (3%)
for every 10 ng/mL increase in serum PFOA (Bartell and Vieira, 2021) (see Appendix A, Table
A-42, USEPA (2024e)). In their review of the available epidemiological data, IARC (2016)
concluded that the evidence for testicular cancer was "considered credible and unlikely to be
explained by bias and confounding, however, the estimate was based on small numbers."
Similarly, CalEPA (2021) concluded, "[ojverall, the epidemiologic literature to date suggests
that PFOA is associated with testicular cancer."
The majority of epidemiological studies examining the carcinogenicity after PFOA exposure
reported on breast cancer risk. Two nested case-control studies found associations between
PFOA exposure and breast cancer, but only in participants with known genetic susceptibility
(e.g., specific genotype or tumor estrogen receptor (ER) type) (Mancini et al., 2020; Ghisari et
al., 2017). In Taiwan, Tsai et al. (2020) observed an increased risk of breast cancer only in all
women 50 years old or younger (including ER+ and ER- participants), and in ER+ participants
aged 50 years or younger, along with a decrease in risk for ER- breast cancers in participants
aged 50 years or younger. Significantly increased odds of breast cancer were also observed in an
NHANES population across serum PFOA quartiles with a significant dose-response trend
(Omoike et al., 2021). Two nested case-control studies did not report an association between
breast cancer and PFOA concentrations measured in maternal serum throughout pregnancy and
1-3 days after delivery (Cohn et al., 2020) or in serum after case diagnosis and breast cancer
(Hurley et al., 2018). One nested case-cohort study did not report an association between breast
cancer and PFOA concentrations measured in a group of predominantly premenopausal women
(Bonefeld-J0rgensen et al., 2014). In the C8 Health Project cohort, Barry et al. (2013) observed a
significant inverse association with breast cancer for both untagged (i.e., concurrent) and 10-year
lagged (i.e., cumulative exposures occurring 10 years in the past) estimated cumulative PFOA
serum concentrations. Similarly, a recent study in a Japanese population reported an inverse
association across serum PFOA quartiles with a significant dose-response trend (Itoh et al.,
2021). Overall, study design differences, lack of replication of the results, and a lack of
9
-------
mechanistic understanding of specific breast cancer subtypes or susceptibilities of specific
populations limit firm conclusions regarding PFOA and breast cancer. However, there is
suggestive evidence that PFOA exposure may be associated with an increased breast cancer risk
based on studies in populations with specific genetic polymorphisms conferring increased
susceptibility and for specific types of breast tumors.
In addition to the available epidemiological data, two multi-dose bioassays and one single-dose
chronic cancer bioassay are available that investigate the relationship between dietary PFOA
exposure and carcinogenicity in male and female rats (NTP, 2020b; Butenhoff et al., 2012b;
Biegel et al., 2001). Increased incidences of neoplastic lesions were primarily observed in male
rats, though results in females are supportive of potential carcinogenicity of PFOA. Testicular
Ley dig cell tumors (LCTs) were identified in both the Butenhoff et al. (2012b) and Biegel et al.
(2001) studies. LCT incidence at similar dose levels was comparable between the two studies
(11% and 14%). Pancreatic acinar cell tumors (PACTs) were observed in both the NTP (2020b)
and Biegel et al. (2001) studies. NTP (2020b) reported increased incidences of pancreatic acinar
cell adenomas and adenocarcinomas in males in all treatment groups compared to their
respective controls. These pancreatic tumor types were also observed in female rats in the
highest dose group, a rare occurrence compared to historical controls (0/340), though these
increases did not reach statistical significance. Biegel et al. (2001) similarly reported increases in
the incidence of PACTs in male rats treated with PFOA, with zero incidences observed in control
animals. In addition, NTP (2020b) reported dose-dependent increases in the incidence of liver
adenomas and carcinomas in male rats and Biegel et al. (2001) also observed increased incidence
of adenomas in male rats. Overall, NTP concluded that in their 2-year feeding studies, there was
clear evidence of carcinogenic activity of PFOA in male Sprague-Dawley rats and some
evidence of carcinogenic activity of PFOA in female Sprague-Dawley rats based on the observed
tumor types (NTP, 2020b).
The report from NTP (2020b) provides evidence that chronic oral exposure accompanied by
perinatal exposure (i.e., exposure beginning at gestation day 5 through lactation) to PFOA does
not increase cancer risk when compared to chronic exposure scenarios beginning during the
postnatal (i.e., exposure initiated after weaning) stage. The incidences of all tumor types
examined did not differ significantly between the treatment groups administered PFOA during
both perinatal and postweaning periods compared with the postweaning-only treatment groups
(see further study design details in Section 3.4.4.2.1.2 of the Final Human Health Toxicity
Assessment for PFOA (USEPA, 2024e). Lifestage-dependent sensitivity to the carcinogenic
effects of PFOA exposure was previously assessed in the study by Filgo et al. (2015) which
exposed two mouse strains during gestation only (i.e., prenatal exposure with no comparisons to
mice exposed through adulthood). Filgo et al. (2015) observed a non-monotonic increase in
hepatocellular adenomas in the female offspring of one strain (CD-I) and hepatocellular
adenoma incidence in approximately 13% of all PFOA-exposed peroxisome proliferator-
activated receptor (PPAR) a-knockout mice. However, these results are not conclusive due to the
study's limited sample size and study design.
In the 2016 PFOA HESD (USEPA, 2016b), the EPA concluded that the induction of tumors was
likely due to multiple MO As, specifically noting interactions with nuclear receptors,
perturbations in the endocrine system, interruption of intercellular communication, mitochondrial
effects, and/or perturbations in the DNA replication and cell division processes. Since that time,
10
-------
the available mechanistic data continue to suggest that multiple MO As could play role in the
renal, testicular, pancreatic, and hepatic tumorigenesis associated with PFOA exposure in human
populations as well as animal models. The few available mechanistic studies focusing on PFOA-
induced renal toxicity highlight several potential underlying mechanisms of PFOA exposure-
induced renal tumorigenesis, including altered cell proliferation and apoptosis, epigenetic
alterations, and oxidative stress. However, due to data limitations, it is difficult to distinguish
which mechanism(s) are operative for PFOA-induced kidney cancer. Similarly for testicular
cancer, the available literature highlights several potential MO As by which PFOA exposure may
result in increased incidence of LCTs in animals, though it is unclear whether these MO As are
relevant to testicular cancers associated with PFOA exposure in humans. Combined, the
epidemiological and animal toxicological literature indicate that the testes are a common site of
PFOA-induced tumorigenesis. A full MOA analysis, including in-depth discussions on the
potential MO As for kidney and testicular tumors, as well as discussions on the potential MO As
and human relevance for pancreatic and liver tumors observed in rats, is presented in Section
3.5.4.2 of th e Final Human Health Toxicity Assessment for PFOA (USEPA, 2024e). Overall, the
EPA concluded that the available mechanistic data suggest that multiple MO As could play role
in the renal, testicular, pancreatic, and hepatic tumorigenesis associated with PFOA exposure in
studies of human populations and animal models. IARC (Zahm et al., 2023; IARC, 2016),
CalEPA (CalEPA, 2021) and NJDWQI (Gleason et al., 2017) similarly concluded that there is
evidence for many potential mechanisms for PFOA-induced carcinogenicity. For example, IARC
concluded there is strong mechanistic evidence of carcinogenicity in exposed humans and that
PFOA is immunosuppressive, induces epigenetic alterations, induces oxidative stress, modulates
receptor-mediated effects (via (PPAR) a, constitutive androstane receptor/pregnane X receptor
[CAR/PXR], and PPARy), and alters cell proliferation, cell death, and nutrient and energy supply
(Zahm et al., 2023).
3.1.2 Cancer Classification
3.1.2.1 PFOA Is Determined to be Likely to Be Carcinogenic to
Humans
Under the Guidelines for Carcinogen Risk Assessment (USEPA, 2005), the EPA reviewed the
weight of the evidence and determined that PFOA is Likely to Be Carcinogenic to Humans, as
"the evidence is adequate to demonstrate carcinogenic potential to humans but does not reach the
weight of evidence for the descriptor Carcinogenic to HumansThis determination is based on
the evidence of kidney and testicular cancer in humans and LCTs, PACTs, and hepatocellular
adenomas in rats.
The Guidelines (USEPA, 2005) provide examples of data that may support the Likely to Be
Carcinogenic to Humans descriptor; the available PFOA data are consistent with the following
factors:
"an agent demonstrating a plausible (but not definitively causal) association between
human exposure and cancer, in most cases with some supporting biological, experimental
evidence, though not necessarily carcinogenicity data from animal experiments";
"an agent that has tested positive in animal experiments in more than one species, sex,
strain, site, or exposure route, with or without evidence of carcinogenicity in humans";
11
-------
"a rare animal tumor response in a single experiment that is assumed to be relevant to
humans";
"a positive tumor study that is strengthened by other lines of evidence, for example, either
plausible (but not definitively causal) association between human exposure and cancer or
evidence that the agent or an important metabolite causes events generally known to be
associated with tumor formation (such as DNA reactivity or effects on cell growth
control) likely to be related to the tumor response in this case" (USEPA, 2005).
The available evidence indicates that PFOA has carcinogenic potential in humans and at least
one animal model. A plausible, though not definitively causal, association exists between human
exposure to PFOA and kidney and testicular cancers in the general population and highly
exposed populations. As stated in the Guidelines for Carcinogen Risk Assessment, "an inference
of causality is strengthened when a pattern of elevated risks is observed across several
independent studies." Two medium confidence independent studies provide evidence of an
association between kidney cancer and elevated PFOA serum concentrations (Shearer et al.,
2021; Vieira et al., 2013), while two studies in the same cohort provide evidence of an
association between testicular cancer and elevated PFOA serum concentrations (Barry et al.,
2013; Vieira et al., 2013). The PFOA cancer database would benefit from additional large high
confidence cohort studies in independent populations.
The evidence of carcinogenicity in animals is based on three studies that used the same strain of
rat. Taken together, these results provide evidence of increased incidence of three different tumor
types (LCTs, PACTs, and hepatocellular tumors) in males administered diets contaminated with
PFOA. Additionally, pancreatic acinar cell adenocarcinomas are a rare tumor type (NTP, 2020b),
and their occurrence in PFOA-treated animals in this study increases the confidence that this
incidence is treatment-related since these tumors are unlikely to be observed in the absence of a
carcinogenic agent (USEPA, 2005). The historical control incidence for pancreatic acinar cell
adenocarcinomas in the female rats is 0/340 and in the male rats is 2/340, highlighting the rarity
of this particular tumor type (NTP, 2020b). Importantly, site concordance is not always assumed
between humans and animal models; agents observed to produce tumors may do so at the same
or different sites in humans and animals (USEPA, 2005). While site concordance was present
between human studies of testicular cancer and animal studies reporting increased incidence of
LCTs, evidence of carcinogenicity of PFOA from other cancer sites where concordance between
humans and animals is not present is still relevant to the carcinogenicity determination for
PFOA. See Table A below for specific rationale on how PFOA aligns with examples supporting
the Likely to Be Carcinogenic to Humans cancer descriptor in the Guidelines for Carcinogen
Risk Assessment (USEPA, 2005).
12
-------
Table A. Comparison of the PFOA Carcinogenicity Database with the Likely Cancer
Descriptor as Outlined in the Guidelines for Carcinogen Risk Assessment (USEPA, 2005)
Likely to Be Carcinogenic to Humans
"An agent demonstrating a plausible (but not
definitively causal) association between human
exposure and cancer, in most cases with some
supporting biological, experimental evidence, though
not necessarily carcinogenicity data from animal
experiments." (USEPA, 2005)
"An agent that has tested positive in animal
experiments in more than one species, sex, strain, site,
or exposure route, with or without evidence of
carcinogenicity in humans." (USEPA, 2005)
"A positive tumor study that raises additional
biological concerns beyond that of a statistically
significant result, for example, a high degree of
malignancy, or an early age at onset." (USEPA, 2005)
"A rare animal tumor response in a single experiment
that is assumed to be relevant to humans." (USEPA,
2005)
"A positive tumor study that is strengthened by other
lines of evidence, for example, either plausible (but
not definitively causal) association between human
exposure and cancer or evidence that the agent or an
important metabolite causes events generally known to
be associated with tumor formation (such as DNA
reactivity or effects on cell growth control) likely to be
related to the tumor response in this case." (USEPA,
2005)
Notes: DNA = deoxyribonucleic acid; MOA = mode of action.
PFOA data are consistent with this description.
Epidemiological evidence supports a plausible
association between PFOA exposure and kidney and
testicular cancer, though there are uncertainties
regarding the MO As for tumor types observed in
humans. There is supporting experimental evidence,
including carcinogenicity data from animal experiments.
PFOA data are consistent with this description.
PFOA has tested positive in one species (rat), both sexes,
and multiple sites (liver, pancreas, testes, uterus). There
is also evidence of carcinogenicity in humans.
This description is not applicable to PFOA. The report
by NTP (2020b) does not indicate that perinatal exposure
exacerbates the carcinogenic potential of PFOA.
PFOA data are consistent with this description. The
pancreatic adenocarcinomas observed in multiple male
dose groups are a rare tumor type in this strain (NTP,
2020b).
PFOA data are consistent with this description.
Multiple positive tumor studies in the same strain of rat
are supported by plausible associations between human
exposure and kidney and testicular cancer.
The EPA recognizes that other state and international health agencies have recently classified
PFOA as carcinogenic to humans (IARC as reported in Zahm et al., 2023; CalEPA, 2021). As
the SAB PFAS Panel (USEPA, 2022c) noted, "the criteria used by California EPA, for
determination that a chemical is a carcinogen, are not identical to the criteria in the U.S. EPA's
Guidelines for Carcinogen Risk Assessment (USEPA, 2005)" and, similarly, IARC's
classification criteria are not identical to the EPA's guidelines (IARC, 2019). Rationale for why
PFOA does not meet the Carcinogenic to Humans descriptor according to the EPA's Guidelines
for Carcinogen Risk Assessment (USEPA, 2005) is detailed in the following section.
3.1.2.2 PFOA Surpasses the Suggestive but Does Not Meet the
Carcinogenic to Humans Classification
While reviewing the weight of evidence for PFOA, the EPA also evaluated consistencies of the
carcinogenicity database with other cancer descriptors according to the Guidelines for
Carcinogen Risk Assessment (USEPA, 2005). In the 2016 PFOA HESD, the EPA determined
that the available carcinogenicity database for PFOA at that time was consistent with the
descriptions for Suggestive Evidence of Carcinogenic Potential (USEPA, 2016b). Upon
reevaluation for this assessment, the agency identified several new studies reporting on cancer
13
-------
outcomes that strengthened the evidence. As a result of conducting a weight of evidence
evaluation of the available carcinogenicity database, the EPA determined that PFOA is
consistent with the descriptions for Likely to Be Carcinogenic to Humans according to the
Guidelines for Carcinogen Risk Assessment (USEPA, 2005), as described above. More
specifically, the available data for PFOA surpass many of the descriptions for Suggestive
Evidence of Carcinogenic Potential provided in the Guidelines for Carcinogen Risk Assessment
(USEPA, 2005). The examples for which the PFOA database exceeds the Suggestive descriptions
(outlined below) include:
"a small, and possibly not statistically significant, increase in tumor incidence observed
in a single animal or human study that does not reach the weight of evidence for the
descriptor "Likely to Be Carcinogenic to Humans." The study generally would not be
contradicted by other studies of equal quality in the same population group or
experimental system (see discussions of conflicting evidence and differing results,
below);
a small increase in a tumor with a high background rate in that sex and strain, when there
is some but insufficient evidence that the observed tumors may be due to intrinsic factors
that cause background tumors and not due to the agent being assessed;
a statistically significant increase at one dose only, but no significant response at the
other doses and no overall trend" (USEPA, 2005).
There are multiple medium or high confidence human and animal toxicological studies that
provide evidence of multiple tumor types resulting from exposure to PFOA. The observed tumor
types are generally consistent across human subpopulations (i.e., kidney (Shearer et al., 2021;
Vieira et al., 2013) and testicular (Barry et al., 2013; Vieira et al., 2013)) and studies of equal
quality did not provide conflicting evidence for these cancer types. Studies within the same
species of rat are consistently demonstrating multi-site tumorigenesis (i.e., testicular, pancreatic,
and hepatic (NTP, 2020b; Butenhoff et al., 2012b; Biegel et al., 2001)) and there is no indication
that a high background incidence or other intrinsic factors related to these tumor types are
driving the observed responses. The SAB PFAS Review Panel agreed that: "a) the evidence for
potential carcinogenicity of PFOA has been strengthened since the 2016 HESD; b) the results of
human and animal studies of PFOA are consistent with the examples provided above and support
a designation of 'likely to be carcinogenic to humans'; and c) the data exceed the descriptors for
the three designations lower than 'likely to be carcinogenic'" (USEPA, 2022b). See Table B
below for specific details on how PFOA exceeds the examples supporting the Suggestive
Evidence of Carcinogenic Potential cancer descriptor in the Guidelines for Carcinogen Risk
Assessment (USEPA, 2005).
While the SAB panel agreed that the data for PFOA exceed a Suggestive cancer descriptor, the
final report also recommends "explicit description of how the available data for PFOA do not
meet the criteria for the higher designation as "carcinogenic" (USEPA, 2022b). After reviewing
the descriptions of the descriptor Carcinogenic to Humans, the EPA has determined that at this
time, the evidence supporting the carcinogenicity of PFOA does not warrant a descriptor
exceeding Likely to Be Carcinogenic to Humans. The Guidelines indicate that a chemical agent
can be deemed Carcinogenic to Humans if it meets all of the following conditions:
14
-------
"there is strong evidence of an association between human exposure and either cancer or
the key precursor events of the agent's mode of action but not enough for a causal
association, and
there is extensive evidence of carcinogenicity in animals, and
the mode(s) of carcinogenic action and associated key precursor events have been
identified in animals, and
there is strong evidence that the key precursor events that precede the cancer response in
animals are anticipated to occur in humans and progress to tumors, based on available
biological information" (USEPA, 2005).
As discussed in the subsection above, convincing epidemiological evidence supporting a causal
association between human exposure to PFOA and cancer are currently lacking. The SAB
similarly concluded that "the available epidemiologic data do not provide convincing evidence of
a causal association but rather provide evidence of a plausible association, and thus do not
support a higher designation of 'carcinogenic to humans'" (USEPA, 2022b).
Additionally, though the available evidence indicates that there are positive associations between
PFOA and multiple cancer types, there is uncertainty regarding the identification of carcinogenic
MOA(s) for PFOA, particularly for renal cell carcinomas and testicular cancer in humans. The
evidence of carcinogenicity in animals is limited to a single strain of rat, although PFOA tested
positive for multi-site tumorigenesis. The animal database does not provide clarity to discern the
MOA(s) of PFOA in humans, though there is some animal toxicological study evidence
supporting hormone-mediated MO As for testicular tumors and oxidative stress-mediated MO As
for pancreatic tumors. The full mode of action analysis, including in-depth discussions on the
potential MO As for kidney and testicular tumors, as well as discussions on the potential MO As
and human relevance for pancreatic and liver tumors observed in rats, is presented in Section
3.5.4.2 of th e Final Human Health Toxicity Assessment for PFOA(XJSEP A, 2024e). See TableB
below for specific details on how PFOA does not align with the examples supporting the
Carcinogenic to Humans cancer descriptor in the Guidelines for Carcinogen Risk Assessment
(USEPA, 2005).
Table B. Comparison of the PFOA Carcinogenicity Database with Cancer Descriptors as
Outlined in the Guidelines for Carcinogen Risk Assessment (USEPA, 2005)
Comparison of Evidence for Suggestive and Carcinogenic Cancer Descriptors
Suggestive Evidence of Carcinogenic Potential
"A small, and possibly not statistically significant, PFOA data exceed this description. Statistically
increase in tumor incidence observed in a single significant increases in tumor incidence of multiple tumor
animal or human study that does not reach the weight types were observed across several human and animal
of evidence for the descriptor "Likely to Be toxicological studies.
Carcinogenic to Humans." The study generally would
not be contradicted by other studies of equal quality
in the same population group or experimental
system." (USEPA, 2005)
"A small increase in a tumor with a high background This description is not applicable to the tumor types
rate in that sex and strain, when there is some but observed after PFOA exposure,
insufficient evidence that the observed tumors may
be due to intrinsic factors that cause background
15
-------
Comparison of Evidence for Suggestive and Carcinogenic Cancer Descriptors
tumors and not due to the agent being assessed.'
(USEPA, 2005)
"Evidence of a positive response in a study whose
power, design, or conduct limits the ability to draw a
confident conclusion (but does not make the study
fatally flawed), but where the carcinogenic potential
is strengthened by other lines of evidence (such as
structure-activity relationships)." (USEPA, 2005)
PFOA data exceed this description. The studies from
which carcinogenicity data are available were determined to
be high or medium confidence during study quality
evaluation.
"A statistically significant increase at one dose only,
but no significant response at the other doses and no
overall trend." (USEPA, 2005)
PFOA data exceed this description. Increases in kidney
cancer in humans were statistically significant in two
exposure groups in one study (Vieira et al., 2013) and there
was a statistically significant increasing trend across
exposure quartiles in a second study (Shearer et al., 2021).
Increases in hepatic and pancreatic tumors in male rats were
observed in multiple dose groups with a statistically
significant trend overall (NTP, 2020b).
Carcinogenic to Humans
"This descriptor is appropriate when there is
convincing epidemiologic evidence of a causal
association between human exposure and cancer."
(USEPA, 2005)
PFOA data are not consistent with this description.
There is evidence of a plausible association between PFOA
exposure and cancer in humans, however, the database is
limited to only two independent populations, there is
uncertainty regarding the potential confounding of other
PFAS, and there is limited mechanistic information that
could contribute to the determination of a causal
relationship.
Or, this descriptor may be equally appropriate with a
lesser weight of epidemiologic evidence that is
strengthened by other lines of evidence. It can be
used when all of the following conditions are met:
"There is strong evidence of an association
between human exposure and either cancer or the
key precursor events of the agent's MO A but not
enough for a causal association." (USEPA, 2005)
PFOA data are not consistent with this description.
There is evidence of an association between human
exposure and cancer, however, there is limited mechanistic
information that could contribute to the determination of a
causal relationship.
"There is extensive evidence of carcinogenicity
in animals." (USEPA, 2005)
PFOA data are not consistent with this description.
While there are three chronic cancer bioassays available,
each testing positive in at least one tumor type, they were all
conducted in the same strain of rat. The database would
benefit from high confidence chronic studies in other
species and/or strains.
"The mode(s) of carcinogenic action and
associated key precursor events have been
identified in animals." (USEPA, 2005)
PFOA data are not consistent with this description. A
definitive MOA has not been identified for each of the
PFOA-induced tumor types identified in rats.
"There is strong evidence that the key precursor
events that precede the cancer response in
animals are anticipated to occur in humans and
progress to tumors, based on available biological
information." (USEPA, 2005)
PFOA data are not consistent with this description. The
animal database does not provide significant clarity on the
MOA(s) of PFOA in humans, though there is some evidence
supporting hormone-mediated MO As for testicular tumors
and oxidative stress-mediated MOAs for pancreatic tumors.
Notes: MOA = mode of action.
16
-------
3.2 PFOS
3.2.1 Summary of the Weight of Evidence
The carcinogenicity of PFOS has been documented in both epidemiological and animal
toxicological studies. The available epidemiology studies report elevated risk of liver, bladder,
kidney, prostate, and breast cancers after chronic PFOS exposure in some studies, though limited
evidence for some tumor types (i.e., liver and renal) and mixed results for other tumor types (i.e.,
bladder, prostate, breast) provide plausible but not definitively causal evidence of a relationship
between PFOS exposure and cancer outcomes from the epidemiological evidence alone. The
animal chronic cancer bioassay provides additional support for carcinogenicity with the
identification of multi-site tumorigenesis (liver and pancreas) in both male and female rats. The
available mechanistic data suggest that multiple MO As could play role in the hepatic and
pancreatic tumorigenesis associated with PFOS exposure based on animal model study findings.
Results for liver cancer from occupational (Alexander et al., 2003) and general population-based
(Eriksen et al., 2009) studies of PFOS exposure published -15-20 years ago were generally
imprecise (i.e., null results with wide confidence intervals), but more recent studies have
reported statistically significant increased risk of liver cancer associated with increased PFOS
exposure (Cao et al., 2022; Goodrich et al., 2022). A nested case-control study of adults from the
Multiethnic Cohort study reported a significant increased risk of liver cancer when comparing
those in the 85th percentile of PFOS exposure to those at or below the 85th percentile (Goodrich
et al., 2022). Positive, but not statistically significant, associations were observed in analyses of
continuous PFOS exposure which supported the study's overall conclusion of an increased risk
of liver cancer with increasing PFOS exposure. The study's sensitivity was limited by the small
number of cases and controls (n = 50 each). Consistent with this finding, a Chinese general
population case-control study of children and adults reported a significant increase in risk of liver
cancer in analyses of continuous PFOS exposure; however, the study was considered low
confidence due to lack of information on control selection, outcome ascertainment, and statistical
analysis (Cao et al., 2022).
Studies of the association between PFOS serum concentrations and bladder cancer have mixed
(positive and null) findings. An elevated risk of bladder cancer mortality was associated with
PFOS exposure in an occupational study (Alexander et al., 2003) but a subsequent study to
ascertain cancer incidence in this cohort with four additional years of observation observed
elevated but not statistically significant incidence ratios that were 1.7- to 2-fold higher among
workers with higher cumulative exposure to PFOS (Alexander and Olsen, 2007). Some of the
limitations of these studies include the lack of precision of the risk estimates due to the small
number of cases, and the lack of control for the potential confounding of smoking. A nested
case-control study in a general population Danish cohort did not observe elevated bladder cancer
risk with increasing PFOS serum levels (Eriksen et al., 2009). Overall, there is suggestive
evidence of a relationship between PFOS exposure and bladder cancer, particularly for high-
exposure communities.
One study in the general population reported a statistically significant increase in risk of RCC in
the highest PFOS exposure quartile and in continuous analyses of PFOS exposure (i.e., per
doubling of PFOS concentration) (Shearer et al., 2021). Although the trend was significant
across quartiles, the effect in the third quartile was null. Additionally, the association with PFOS
17
-------
was attenuated after adjusting for other PFAS, and it was lower in the third quartile than in the
second quartile, indicating potential confounding by correlated PFAS exposures. There was no
reported association when evaluated on a per doubling of PFOS after adjusting for other PFAS.
Elevated non-significant ORs for prostate cancer were reported for the occupationally exposed
cohort examined by Alexander and Olsen (2007) and the Danish population-based cohort
examined by Eriksen et al. (2009). In the same occupational cohort studied by Alexander and
Olsen (2007), Grice et al. (2007) observed that prostate cancers were among the most frequently
reported cancers. When cumulative PFOS exposure measures were analyzed, elevated ORs were
reported for prostate cancer, however, they did not reach statistical significance. Length of
follow-up may not have been adequate to detect cancer incidence in this cohort as approximately
one-third of the participants had worked < 5 years in their jobs, and only 41.7% were employed
> 20 years (Grice et al., 2007). No association between PFOS exposure and prostate cancer was
reported in either a second case-control study in Denmark (Hardell et al., 2014) or in a study of
the association between PFOS serum concentrations and prostate specific antigen (a biomarker
of prostate cancer) from the C8 Health Project (Ducatman et al., 2015). In an NHANES
population, Omoike et al. (2021) observed a significantly inverse association between PFOS
exposure and prostate cancer.
The majority of studies examining associations between PFOS exposure and cancer outcomes
were on breast cancer. One study of Inuit females in Greenland observed positive associations
between PFOS levels and risk for breast cancer (Bonefeld-Jorgensen et al., 2011), although the
association was of a low magnitude and could not be separated from the effects of other
perfluorosulfonated compound exposures (i.e., perfluorohexanesulfonate and
perfluorooctanesulfonamide). Three studies indicated potential associations between PFOS
exposure and increased breast cancer risk in specific subgroups or increased risk for specific
breast cancer subtypes. Ghisari et al. (2017) reported that increased breast cancer risk was
associated with increased PFOS serum concentrations in Danish individuals with a specific
polymorphism in the cytochrome P450 aromatase gene (for aromatase, associated with estrogen
biosynthesis and metabolism). Mancini et al. (2020) reported that increased PFOS serum
concentrations were associated specifically with increased risk of ER+ and PR+ tumors, whereas
risk of ER- and PR- tumors did not follow a dose-dependent response. In a Taiwanese
population, Tsai et al. (2020) observed a statistically significant increased risk of breast cancer in
all women 50 years old or younger (including ER+ and ER- participants), and in ER+
participants aged 50 years or younger. Statistically significant increases in breast cancer risk
were also observed in an NHANES population in the two highest quartiles of exposure, but the
association was inverse in the second quartile (Omoike et al., 2021). No association was
identified between PFOS and breast cancer in either case-control or nested case-control studies
of Danish and California cancer registry populations, respectively (Hurley et al., 2018; Bonefeld-
J0rgensen et al., 2014). Another general population study in the U.S. suggested that maternal
PFOS exposure combined with high maternal cholesterol may decrease the daughters' risk of
breast cancer but did not examine breast cancer subtypes or individuals with genetic variants that
may have increased susceptibility (Cohn et al., 2020). A recent study in a Japanese population
observed an inverse association across serum PFOS quartiles with a significant dose-response
trend (Itoh et al., 2021). The association remained significantly inverse in both pre- and
postmenopausal women in the highest tertile of exposure, with a significant dose-response trend.
However, in some of the studies PFOS levels were measured after or near the time of cancer
18
-------
diagnosis (Omoike et al., 2021; Tsai et al., 2020). Given the long half-life of PFOS in human
blood, the exposure levels measured in these studies could represent exposures that occurred
prior to cancer development. However, this is currently difficult to evaluate since data on the
latency of PFOS exposure and subsequent cancer assessment is not available. Overall, study
design limitations with specific studies, lack of replication of the results, and a lack of
mechanistic understanding of specific breast cancer subtypes or susceptibilities of specific
populations limit firm conclusions regarding PFOS and breast cancer. However, there is
suggestive evidence that PFOS exposure may be associated with an increased breast cancer risk
based on studies in susceptible populations, such as those with specific polymorphisms and for
specific types of breast tumors.
One available chronic toxicity/carcinogenicity bioassay for PFOS, a 104-week dietary study in
rats, provides evidence of multi-sex and multi-site tumorigenesis resulting from PFOS exposure
(Butenhoff et al., 2012a; Thomford, 2002). This study was originally published as a 3M-
sponsored report by Thomford (2002) and some of the data were later published in a peer-
reviewed article by Butenhoff et al. (2012a). Statistically significant increases in the incidence of
hepatocellular adenomas in the high-dose (20 ppm) male (7/43; 16%) and female rat groups
(5/31; 16%) and combined adenomas/carcinomas in the females (6/32; 19%; five adenomas, one
carcinoma) were observed. The observation of one carcinoma in the female rats is a relatively
rare occurrence according to NTP's historical controls for female Sprague-Dawley rats (1/639
historical control incidence) (NTP, 2020a). Historical control incidence rates for these tumor
types were not provided by Thomford (2002). Additionally, there were statistically significant
dose-related trends in the hepatic tumor responses of both males and females. A statistically
significant trend of increased incidence of pancreatic islet cell carcinomas with increased PFOS
dose was also observed in the male rats, though the individual dose groups were not statistically
different from the control group. The percentages of animals with islet cell carcinomas in the
highest dose group (12.5%) exceeds NTP's historical controls for male Sprague-Dawley rats by
over an order of magnitude (12/638; 1.9%) (NTP, 2020a).
Thyroid tumors (follicular cell adenomas and carcinomas) were observed in males and females,
though these responses were not statistically significant in any dose group, nor was there a linear
dose-response trend (Butenhoff et al., 2012a; Thomford, 2002). In males, the incidence of
thyroid tumors was significantly elevated only in the high-dose, recovery group males exposed
for 52 weeks (10/39) but not in the animals receiving the same dose for 105 weeks. However,
Thomford (2002) indicated that the number of thyroid tumors observed in the recovery group
males were outside the range of historical control values at that time, similar to what NTP
(2020a) has reported for its laboratories (3/637 combined follicular cell adenoma or carcinoma).
There were few follicular cell adenomas/carcinomas in the females (4 total, excluding the
recovery group) with a nonlinear dose response. Mammary gland tumors, primarily combined
fibroma adenoma and adenoma, were also observed in females, though there was a high
background incidence of mammary gland tumors in the control animals, and the incidence lacked
dose response for all tumor classifications.
Based on the weight of evidence evaluation of the available peer-reviewed scientific evidence,
PFOS has the potential to induce hepatic tumors in humans and rodents via multiple MOAs,
most notably via the modulation of nuclear receptors (i.e., PPARa and CAR) and cytotoxicity.
There is also limited evidence supporting additional potential MOAs of genotoxicity,
19
-------
immunosuppression, and oxidative stress. The conclusions from the weight of evidence analysis
of the available data for PFOS are consistent with literature reviews recently published by two
state health agencies which concluded that the hepatotoxic effects of PFOS are not entirely
dependent on PPARa activation (CalEPA, 2021; NJDWQI, 2018). Similarly, IARC (Zahm et al.,
2023) found strong mechanistic evidence of carcinogenicity in exposed humans and that PFOS is
immunosuppressive and induces epigenetic alterations in humans, induces oxidative stress in
human primary cells and experimental systems and modulates multiple receptors. No established
MOA was identified for pancreatic islet cell carcinogenicity in animals. A full mode of action
analysis, including in-depth discussions on the potential MO As for hepatic and pancreatic tumors
is presented in Section 3.5.4 of th q Final Human Health Toxicity Assessment for PFOS (USEPA,
2024d). Overall, the EPA concluded that there is an absence of definitive mechanistic data
supporting a single MOA for PFOS and therefore, both tumor types may be relevant to humans
in accordance with the Guidelines for Carcinogen Risk Assessment (USEPA, 2005).
3.2.2 Cancer Classification
3.2.2.1 PFOS Is Likely to Be Carcinogenic to Humans
Under the Guidelines for Carcinogen Risk Assessment (USEPA, 2005), the EPA reviewed the
weight of the evidence and determined that PFOS is Likely to Be Carcinogenic to Humans, as
"the evidence is adequate to demonstrate carcinogenic potential to humans but does not reach the
weight of evidence for the descriptor Carcinogenic to HumansThe Guidelines provide
descriptions of data that may support the Likely to Be Carcinogenic to Humans descriptor; the
available PFOS data are consistent with the following factors:
"an agent that has tested positive in animal experiments in more than one species, sex,
strain, site, or exposure route, with or without evidence of carcinogenicity in humans;
a rare animal tumor response in a single experiment that is assumed to be relevant to
humans; or
a positive tumor study that is strengthened by other lines of evidence, for example, either
plausible (but not definitively causal) association between human exposure and cancer or
evidence that the agent or an important metabolite causes events generally known to be
associated with tumor formation (such as DNA reactivity or effects on cell growth
control) likely to be related to the tumor response in this case" (USEPA, 2005).
The available evidence indicates that PFOS has carcinogenic potential in one animal model for
multiple sites and both sexes, as well as supporting evidence from human studies, consistent with
the examples described in the Guidelines for Carcinogen Risk Assessment for the Likely
descriptor. The epidemiological evidence of associations between PFOS and cancer found mixed
results across tumor types. However, the available study findings support a plausible correlation
between PFOS exposure and carcinogenicity in humans. The single chronic cancer bioassay
performed in rats is positive for multi-site and -sex tumorigenesis (Butenhoff et al., 2012a;
Thomford, 2002). In this study, statistically significant increases in the incidences of
hepatocellular adenomas or combined adenomas and carcinomas were observed in both male and
female rats. There was also a statistically significant trend of this response in both sexes
indicating a relationship between the magnitude/direction of response and PFOS dose. As
described in Section 3.5.4.2 of the Final Human Health Toxicity Assessment for PFOS (USEPA,
2024d), the available mechanistic evidence is consistent with multiple potential MO As for this
20
-------
tumor type; therefore, the hepatocellular tumors observed by Butenhoff et al./Thomford (2012a;
2002) may be relevant to humans. These findings in rats and their potential human relevance are
supported by recent epidemiological studies that have reported associations between PFOS and
hepatocellular carcinoma in humans (Cao et al., 2022; Goodrich et al., 2022).
In addition to hepatocellular tumors, Thomford (2002) reported increased incidences of
pancreatic islet cell carcinomas with a statistically significant dose-dependent positive trend, as
well as modest increases in the incidence of thyroid follicular cell tumors. The findings of
multiple tumor types provide additional support for potential multi-site tumorigenesis resulting
from PFOS exposure. Importantly, site concordance is not always assumed between humans and
animal models; agents observed to produce tumors may do so at the same or different sites in
humans and animals (USEPA, 2005). While site concordance was present between human
studies of liver cancer and animal studies reporting increased incidence of hepatocellular tumors,
evidence of carcinogenicity of PFOS from other cancer sites where concordance between
humans and animals is not present is still relevant to the carcinogenicity determination for PFOS.
See Table C below for specific details on how PFOS aligns with the examples supporting the
Likely to Be Carcinogenic to Humans cancer descriptor in the Guidelines for Carcinogen Risk
Assessment (USEPA, 2005).
Table C. Comparison of the PFOS Carcinogenicity Database with the Likely Cancer
Descriptor as Outlined in the Guidelines for Carcinogen Risk Assessment (USEPA, 2005)
Likely to Be Carcinogenic to Humans
"An agent demonstrating a plausible (but not definitively
causal) association between human exposure and cancer,
in most cases with some supporting biological,
experimental evidence, though not necessarily
carcinogenicity data from animal experiments."
(USEPA, 2005)
"An agent that has tested positive in animal experiments
in more than one species, sex, strain, site, or exposure
route, with or without evidence of carcinogenicity in
humans." (USEPA, 2005)
"A positive tumor study that raises additional biological
concerns beyond that of a statistically significant result,
for example, a high degree of malignancy, or an early
age at onset." (USEPA, 2005)
"A rare animal tumor response in a single experiment
that is assumed to be relevant to humans." (USEPA,
2005)
"A positive tumor study that is strengthened by other
lines of evidence, for example, either plausible (but not
definitively causal) association between human exposure
and cancer or evidence that the agent or an important
metabolite causes events generally known to be
PFOS data are consistent with this description.
Epidemiological evidence supports a plausible
association between PFOS exposure and liver cancer
which is consistent with evidence of liver cancer in
animals. Epidemiological studies evaluating the
association between human exposure to PFOS and other
cancers are mixed. Supporting carcinogenicity data are
available from animal experiments.
PFOS data are consistent with this description. PFOS
has tested positive in animal experiments in more than
one sex and site. Hepatic tumors were observed in male
and female rats (statistically significant at high dose and
statistically significant trend tests for each) and islet cell
carcinomas show a statistically significant positive trend
in male rats.
This description is not applicable to PFOS.
PFOS data are consistent with this description. The
hepatocellular carcinoma observed in the high-dose
female rats is a rare tumor type in this strain (NTP,
2020b).
PFOS data are consistent with this description. The
positive multi-site, multi-sex chronic cancer bioassay is
supported by mechanistic data indicating that PFOS is
associated with events generally known to be associated
with tumor formation such as inducing nuclear receptor
21
-------
Likely to Be Carcinogenic to Humans
associated with tumor formation (such as DNA reactivity activation, cytotoxicity, genotoxicity, oxidative stress,
or effects on cell growth control) likely to be related to and immunosuppression.
the tumor response in this case." (USEPA, 2005)
Notes: MOA = mode of action.
The EPA recognizes that other state and international health agencies have recently classified
PFOS as either "possibly carcinogenic to humans" (IARC as reported in Zahm et al., 2023) or
carcinogenic to humans (CalEPA, 2021). As the SAB PFAS Review Panel (USEPA, 2022c)
noted, "the criteria used by California EPA, for determination that a chemical is a carcinogen, are
not identical to the criteria in the U.S. EPA's Guidelines for Carcinogen Risk Assessment
(USEPA, 2005)" and, similarly, IARC's classification criteria are not identical to the EPA's
guidelines (IARC, 2019). Rationale for why PFOS exceeds the Suggestive Evidence of
Carcinogenic Potential descriptor and does not meet the Carcinogenic to Humans descriptor
according to the EPA's Guidelines for Carcinogen Risk Assessment (USEPA, 2005) is detailed in
the following section.
3.2.2.2 PFOS Surpasses the Suggestive but Does Not Meet the
Carcinogenic to Humans Classification
To provide further support for that PFOS is Likely to Be Carcinogenic to Humans, the EPA also
evaluated consistencies of the carcinogenicity database with other cancer descriptors according
to the Guidelines for Carcinogen Risk Assessment (USEPA, 2005). Upon reassessment of the
PFOS carcinogenicity database, including the epidemiological, animal toxicological, and
mechanistic databases, the agency has determined the available data for PFOS surpass many of
the descriptions for Suggestive Evidence of Carcinogenic Potential according to the Guidelines
for Carcinogen Risk Assessment (USEPA, 2005). The examples for which the PFOS database
exceeds the Suggestive Evidence descriptions outlined in the Guidelines for Carcinogen Risk
Assessment include:
"a small, and possibly not statistically significant, increase in tumor incidence observed in
a single animal or human study that does not reach the weight of evidence for the
descriptor Likely to Be Carcinogenic to Humans;
a small increase in a tumor with a high background rate in that sex and strain, when there
is some but insufficient evidence that the observed tumors may be due to intrinsic factors
that cause background tumors and not due to the agent being assessed;
evidence of a positive response in a study whose power, design, or conduct limits the
ability to draw a confident conclusion; and
a statistically significant increase at one dose only, but no significant response at the other
doses and no overall trend" (USEPA, 2005).
The strongest evidence for the carcinogenicity of PFOS is from one chronic animal bioassay
which presents findings surpassing several of these criteria (Butenhoff et al., 2012a; Thomford,
2002). The Thomford/Butenhoff et al. (2012a; 2002) study is a high confidence study that
observed statistically significant increases at individual dose levels and/or statistically significant
trends in two tumor types and in one or more sexes, even with the relatively low dose levels
used. The background incidence of these tumor types was low or negligible.
22
-------
In the initial draft of this toxicity assessment (e.g., the Proposed Approaches document)
(USEPA, 2021a) published for SAB review and the 2016 HESD (USEPA, 2016a), the EPA
relied upon the tumor incidences provided in Butenhoff et al. (2012a), which is the peer-
reviewed manuscript of an unpublished industry report - Thomford (2002). Upon further review
of the results presented in the Thomford (2002) report prior to rule proposal (USEPA, 2023), the
agency identified two factors that limited previous qualitative and quantitative interpretations of
the data: 1) the Butenhoff et al. (2012a) study reported combined incidences of neoplastic lesions
in the control and high-dose groups from the interim time point (52 weeks of dietary exposure;
n = 10) and terminal time point (104 weeks of dietary exposure; n = 50); and 2) the Butenhoff et
al. (2012a) study did not report incidences for pancreatic islet cell neoplasms. The first factor
resulted in statistical dilution of tumor incidence in the high-dose group as many of the tumor
types observed in the study, including hepatocellular neoplasms, were not reported until
approximately 70 weeks of treatment or later. Therefore, the EPA conducted a re-analysis that
excluded animals sacrificed at the interim time point from statistical analyses as it was
biologically implausible for the 10 animals from the interim time point to have presented with
neoplasms. The second factor impacted the EPA from recognizing the statistically significant
trend in a second tumor site/type (pancreatic islet cell carcinomas) observed in the chronic cancer
bioassay. As a result of identifying the second tumor site/type, PFOS does meet an additional
characteristic for the designation of Likely to Be Carcinogenic to Humans: "an agent that has
tested positive in animal experiments in more than one species, sex, strain, site, or exposure
route, with or without evidence of carcinogenicity in humans" (emphasis added) (USEPA, 2005).
Overall, the Thomford/Butenhoff et al. (2012a; 2002) report, along with plausible associations
between PFOS exposure and carcinogenicity reported in epidemiological studies, provides
substantive evidence that PFOS exceeds the designation of Suggestive Evidence of Carcinogenic
Potential and is consistent with Likely Evidence of Carcinogenic Potential in Humans (see
Section 3.5.5 of the Final Human Health Toxicity Assessment for PFOS (USEPA, 2024d) for
more information on the Likely determination). See Table D below for specific details on how
PFOS exceeds the examples supporting the Suggestive Evidence of Carcinogenic Potential
cancer descriptor in the Guidelines for Carcinogen Risk Assessment (USEPA, 2005).
After reviewing the examples of the descriptor Carcinogenic to Humans, the EPA has
determined that at this time, the evidence supporting the carcinogenicity of PFOS does not
warrant a descriptor exceeding Likely to Be Carcinogenic to Humans. The Guidelines indicate
that a chemical agent can be deemed Carcinogenic to Humans if it meets all of the following
conditions:
"there is strong evidence of an association between human exposure and either cancer or
the key precursor events of the agent's mode of action but not enough for a causal
association, and
there is extensive evidence of carcinogenicity in animals, and
the mode(s) of carcinogenic action and associated key precursor events have been
identified in animals, and
there is strong evidence that the key precursor events that precede the cancer response in
animals are anticipated to occur in humans and progress to tumors, based on available
biological information" (USEPA, 2005).
23
-------
As discussed in the subsection above, convincing epidemiological evidence supporting a causal
association between human exposure to PFOS and cancer are currently lacking. Additionally,
though the available evidence indicates that there are positive associations between PFOS and
multiple cancer types, there is uncertainty regarding the identification of carcinogenic MOA(s)
and associated key precursor events for PFOS in animals. See Table D below for specific details
on how PFOS does not align with the examples supporting the Carcinogenic to Humans cancer
descriptor in the Guidelines for Carcinogen Risk Assessment (USEPA, 2005).
Table D. Comparison of the PFOS Carcinogenicity Database with Cancer Descriptors as
Outlined in the Guidelines for Carcinogen Risk Assessment (USEPA, 2005)
Comparison of Evidence for Suggestive and Carcinogenic Cancer Descriptors
Suggestive Evidence of Carcinogenic Potential
"A small, and possibly not statistically significant,
increase in tumor incidence observed in a single
animal or human study that does not reach the weight
of evidence for the descriptor "Likely to Be
Carcinogenic to Humans." The study generally would
not be contradicted by other studies of equal quality in
the same population group or experimental system"
(USEPA, 2005)
PFOS data exceed this description. Observed statistically
significant increases in hepatic tumors (adenomas in males
and adenomas and carcinomas in females) at the high dose
and a statistically significant trend overall in both sexes.
"A small increase in a tumor with a high background
rate in that sex and strain, when there is some but
insufficient evidence that the observed tumors may be
due to intrinsic factors that cause background tumors
and not due to the agent being assessed." (USEPA,
2005)
This description is not applicable to the tumor types
observed after PFOS exposure.
"Evidence of a positive response in a study whose
power, design, or conduct limits the ability to draw a
confident conclusion (but does not make the study
fatally flawed), but where the carcinogenic potential is
strengthened by other lines of evidence (such as
structure-activity relationships)." (USEPA, 2005)
PFOS data exceed this description. The study from
which carcinogenicity data are available was determined to
be high confidence during study quality evaluation.
"A statistically significant increase at one dose only,
but no significant response at the other doses and no
overall trend." (USEPA, 2005)
PFOS data exceed this description. Observed statistically
significant increases in hepatic tumors (adenomas in males
and adenomas and carcinomas in females) at the high dose
and a statistically significant trend overall. Also observed
statistically significant trend of increased pancreatic islet
cell carcinomas with increasing dose.
Carcinogenic to Humans
"This descriptor is appropriate when there is
convincing epidemiologic evidence of a causal
association between human exposure and cancer."
(USEPA, 2005)
PFOS data are not consistent with this description.
There is evidence of a plausible association between PFOS
exposure and cancer in humans, however, the database is
limited, there is uncertainty regarding the potential
confounding of other PFAS, and there is limited
mechanistic information that could contribute to the
determination of a causal relationship.
Or, this descriptor may be equally appropriate with a
lesser weight of epidemiologic evidence that is
strengthened by other lines of evidence. It can be used
when all of the following conditions are met:
24
-------
Comparison of Evidence for Suggestive and Carcinogenic Cancer Descriptors
"There is strong evidence of an association
between human exposure and either cancer or the
key precursor events of the agent's MO A but not
enough for a causal association." (USEPA, 2005)
"There is extensive evidence of carcinogenicity in
animals." (USEPA, 2005)
"The mode(s) of carcinogenic action and associated
key precursor events have been identified in animals."
(USEPA, 2005)
"There is strong evidence that the key precursor events
that precede the cancer response in animals are
anticipated to occur in humans and progress to tumors,
based on available biological information." (USEPA,
2005)
Notes: MOA = mode of action.
PFOS data are not consistent with this description.
There is evidence of an association between human
exposure and cancer, however, there is limited mechanistic
information that could contribute to the determination of a
causal relationship.
PFOS data are not consistent with this description. Only
one chronic cancer bioassay is available for PFOS. The
database would benefit from high confidence chronic
studies in other species and/or strains.
PFOS data are not consistent with this description. A
definitive MOA has not been identified for each of the
PFOS-induced tumor types identified in rats.
PFOS data are not consistent with this description. The
animal database does not provide significant clarity on the
MOA(s) of PFOS in animals.
25
-------
4 MCLG Derivation
4.1 PFOA
Consistent with the Guidelines for Carcinogen Risk Assessment (USEPA, 2005), the EPA
reviewed the weight of evidence and determined that PFOA is Likely to Be Carcinogenic to
Humans because "the evidence is adequate to demonstrate carcinogenic potential to humans but
does not reach the weight of evidence for the descriptor Carcinogenic to HumansThis
determination is based on the evidence of kidney and testicular cancer in humans and LCTs,
PACTs, and hepatocellular tumors in rats as described in Section 3.1 above, and Section 3.5 of
the Final Human Health Toxicity Assessment for PFOA (USEPA, 2024e).
Consistent with the statutory definition of MCLG, the EPA establishes MCLGs of zero for
carcinogens classified as either Carcinogenic to Humans or Likely to Be Carcinogenic to
Humans where there is a proportional relationship between dose and carcinogenicity at low
concentrations or where there is insufficient information to determine that a carcinogen has a
threshold dose below which no carcinogenic effects have been observed. In these situations, the
EPA takes the health protective approach of assuming that carcinogenic effects should therefore
be extrapolated linearly to zero (USEPA, 2005). This is called the linear default extrapolation
approach and ensures that the MCLG is set at a level where there are no anticipated adverse
health effects, allowing for an adequate margin of safety (USEPA, 2016c, 1991, 1985). Here, the
EPA has determined that PFOA is Likely to Be Carcinogenic to Humans based on sufficient
evidence of carcinogenicity in humans and animals. The EPA has also determined that a linear
default extrapolation approach is appropriate as there is no evidence demonstrating a threshold
level of exposure below which there is no appreciable cancer risk (USEPA, 2005) and therefore,
there is no known threshold for carcinogenicity. Based upon a consideration of the best available
peer-reviewed science and data collected by accepted or best available methods, as well as the
statutory directive to set the MCLG "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 EPA
has finalized an MCLG of zero for PFOA in drinking water.
4.2 PFOS
Consistent with the Guidelines for Carcinogen Risk Assessment (USEPA, 2005), the EPA
reviewed the weight of evidence and determined that PFOS is Likely to Be Carcinogenic to
Humans because "the evidence is adequate to demonstrate carcinogenic potential to humans but
does not reach the weight of evidence for the descriptor Carcinogenic to HumansThis
determination is based on the evidence of hepatocellular tumors in male and female rats, which is
further supported by recent evidence of hepatocellular carcinoma in humans (Cao et al., 2022;
Goodrich et al., 2022), pancreatic islet cell carcinomas in male rats, and mixed but plausible
evidence of bladder, prostate, kidney, and breast cancers in humans as described in Section 3.2
above, and Section 3.5 ofth q Final Human Health Toxicity Assessment for PFOS (USEPA,
2024d).
Consistent with the statutory definition of MCLG, the EPA establishes MCLGs of zero for
carcinogens classified as either Carcinogenic to Humans or Likely to Be Carcinogenic to
Humans where there is a proportional relationship between dose and carcinogenicity at low
concentrations or where there is insufficient information to determine that a carcinogen has a
26
-------
threshold dose below which no carcinogenic effects have been observed. In these situations, the
EPA takes the health protective approach of assuming that carcinogenic effects should therefore
be extrapolated linearly to zero (USEPA, 2005). This is called the linear default extrapolation
approach and ensures that the MCLG is set at a level where there are no anticipated adverse
health effects, allowing for an adequate margin of safety (USEPA (1985); USEPA (1991);
USEPA (2016c); See S. Rep. No. 169, 104th Cong., 1st Sess. (1995) at 3). Here, the EPA has
determined that PFOS is Likely to Be Carcinogenic to Humans based on sufficient evidence of
carcinogenicity in humans and animals. The EPA has also determined that a linear default
extrapolation approach is appropriate as there is no evidence demonstrating a threshold level of
exposure below which there is no appreciable cancer risk (USEPA, 2005) and therefore, there is
no known threshold for carcinogenicity. Based upon a consideration of the best available peer-
reviewed science and data collected by accepted or best available methods, as well as the
statutory directive to set the MCLG "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 EPA
has finalized an MCLG of zero for PFOS in drinking water.
27
-------
5 References
Alexander, BH; Olsen, GW. (2007). Bladder cancer in perfluorooctanesulfonyl fluoride manufacturing
workers. Annals of Epidemiology 17: 471-478.
http://dx.doi.iH ' (0 hMu/i.annepidem.200 ' 01 036
Alexander, BH; Olsen, GW; Burris, JM; Mandel, JH; Mandel, JS. (2003). Mortality of employees of a
perfluorooctanesulphonyl fluoride manufacturing facility. Occupational and Environmental
Medicine 60: 722-729. http://dx.doij /oem.60..1.0.722
Barry, V; Winquist, A; Steenland, K. (2013). Perfluorooctanoic acid (PFOA) exposures and incident
cancers among adults living near a chemical plant. Environmental Health Perspectives 121: 1313-
1318. http://dx.doi.oi 89/ehp..1.3066.1.5
Bartell, SM; Vieira, VM. (2021). Critical Review on PFOA, Kidney Cancer, and Testicular Cancer.
Journal of the Air and Waste Management Association.
http://dx.doi.org/10.1080/10962247.2021. .1.909668
Biegel, LB; Hurtt, ME; Frame, S. R.; O'Connor, JC; Cook, JC. (2001). Mechanisms of extrahepatic tumor
induction by peroxisome proliferators in male CD rats. Toxicological Sciences 60: 44-55.
http://dx.doi.org/ .1.0. .1.093/toxsci/'
Bonefeld-Jorgensen, EC; Long, M; Bossi, R; Ayotte, P; Asmund, G; Kriiger, T, et al. (2011).
Perfluorinated compounds are related to breast cancer risk in Greenlandic Inuit: a case control
study. Environmental Health: A Global Access Science Source 10: 88.
http://dx.dob ^ 10 1 I ^ V1476-069X-10-88
Bonefeld-Jorgensen, 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 and Control 25: 1439-1448.
http://dx.doi.org/ .1.0. .1.007/s .1. C 46-7
Butenhoff, JL; Chang, SC; Olsen, GW; Thomford, PJ. (2012a). Chronic dietary toxicity and
carcinogenicity study with potassium perfluorooctanesulfonate in Sprague Dawley rats.
Toxicology 293: 1-15. http://dx.doit /j.tox.2012.01.003
Butenhoff, JL; Kennedy, GL; Chang, SC; Olsen, GW. (2012b). Chronic dietary toxicity and
carcinogenicity study with ammonium perfluorooctanoate in Sprague-Dawley rats. Toxicology
298: 1-13.
C8 Science Panel. (2012). C8 portable link reports, http://www.c8sciencepanel.org/index.html
CalEPA. (2021). Public Health Goals: Perfluorooctanoic Acid and Perfluorooctane Sulfonic Acid in
Drinking Water (First Public Review Draft ed.). California Environmental Protection Agency,
Office of Environmental Health Hazard Assessment, Pesticide and Environmental Toxicology
Branch.
https://oehha.ca.gov/sites/default/files/media/downloads/cmr/pfoapfosphgdraft061021.pdf
Cao, L; Guo, Y; Chen, Y; Hong, J; Wu, J; Hangbiao, J. (2022). Per-/polyfluoroalkyl substance
concentrations in human serum and their associations with liver cancer. Chemosphere 296:
134083. http://dx.doi.org ;hemosphere.2022..1.34083
Cohn, BA; La Merrill, MA; Krigbaum, NY; Wang, M; Park, JS; Petreas, M, et al. (2020). In utero
exposure to poly- and perfluoroalkyl substances (PFASs) and subsequent breast cancer.
Reproductive Toxicology 92: 112-119.
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.
Eriksen, KT; Sorensen. M; Mclaughlin, JK; Lipworth, L; Tjonncland. A; Overvad, K; Raaschou-Nielsen,
O. (2009). Perfluorooctanoate and perfluorooctanesulfonate plasma levels and risk of cancer in
the general Danish population. Journal of the National Cancer Institute 101: 605-609.
http ://dx.doi .org/ .1.0. .1.093/i nci/di p04.1.
28
-------
Filgo, AJ; Quist, EM; Hoenerhoff, MJ; Brix, AE; Kissling, GE; Fenton, SE. (2015). Perfluorooctanoic
acid (PFOA)-induced liver lesions in two strains of mice following developmental exposures:
PPARalpha is not required. Toxicologic Pathology 43: 558-568.
http://dx.doi.org/10. .1.177/01926233.1.4558463
Ghisari, M; Long, M; Roge. DM; Olsen, J; Bonefcld-lorgenscn. EC. (2017). Polymorphism in xenobiotic
and estrogen metabolizing genes, exposure to perfluorinated compounds and subsequent breast
cancer risk: A nested case-control study in the Danish National Birth Cohort. Environmental
Research 154: 325-333.
Gleason, JA; Cooper, KR; Klotz, JB; Post, GB; Van Orden, G; New Jersey Drinking Water Quality
Institute (NJDWQI). (2017). Health-based maximum contaminant level support document:
Perfluorooctanoic acid (PFOA): Appendix A. https://www.state.ni.us/dep/watersupply/pdf/pfoa-
appendixa.pdf
Goodrich, JA; Walker, D; Lin, X; Wang, H; Lim, T; Mcconnell, R, et al. (2022). Exposure to
perfluoroalkyl substances and risk of hepatocellular carcinoma in a multiethnic cohort. 4:
100550. http://dx.doi.org hepr.2022..1.00550
Grice, MM; Alexander, BH; Hoffbeck, R; Kampa, DM. (2007). Self-reported medical conditions in
perfluorooctanesulfonyl fluoride manufacturing workers. Journal of Occupational and
Environmental Medicine 49: 722-729. http://dx.doi.org/10.1097/JOM.0b013e3180582043
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. http://dx.doi.Ct' (0 10Ir'i.envint AM.- 10 005.
Hurley, S; Goldberg, D; Wang, M; Park, JS; Petreas, M; Bernstein, L, et al. (2018). Breast cancer risk and
serum levels of per- and poly-fluoroalkyl substances: a case-control study nested in the California
Teachers Study. Environmental Health: A Global Access Science Source 17: 83.
IARC (International Agency for Research on Cancer). (2016). Some chemicals used as solvents and in
polymer manufacture Perfluorooctanoic acid. In IARC Monographs on the Evaluation of the
Carcinogenic Risk of Chemicals to Humans. Lyons, France: World Health Organization.
http://monographsiarc.fr/ENG/Monographs/voll.10/mot
IARC. (2019). Preamble to the IARC Monographs. In IARC Monographs on the Identification of
Carcinogenic Hazards to Humans. Lyon, France, https://rnonographs.iarc.who.int/wp~
content/uploads/20.1.9/07/.Preamble-2019.pdf
Itoh, H; Harada, KH; Kasuga, Y; Yokoyama, S; Onuma, H; Nishimura, H, et al. (2021). Serum
perfluoroalkyl substances and breast cancer risk in Japanese women: A case-control study.
Science of the Total Environment 800: 149316. http://dx.doi.org scitotenv.202.1.. .1.493.1.6
Mancini, FR; Cano-Sancho, G; Gambaretti, J; Marchand, P; Boutron-Ruault, MC; Severi, G, et al. (2020).
Perfluorinated alkylated substances serum concentration and breast cancer risk: Evidence from a
nested case-control study in the French E3N cohort. International Journal of Cancer 146: 917-
928.
NJDWQI. (2018). Health-based maximum contaminant level support document: Perfluorooctane
sulfonate (PFOS): Appendix A. https://www.state.ni.us/dep/watersupplv/pdf/pfos-
recommendation-appendix-a.pdf
NTP. (2020a). NTP historical controls report all routes and vehicles Harlan Sprague-Dawley rats [NTP],
Research Triangle Park, NC.
https://ntp.niehs.nih.gov/ntp/liistorical controls/ntp2000 2020/r herpt allrte20201200.pdf
NTP. (2020b). NTP technical report on the toxicology and carcinogenesis studies of perfluorooctanoic
acid (CASRN 335-67-1) administered in feed to Sprague Dawley (Hsd:Sprague Dawley SD) rats
[NTP], (Technical Report 598). Research Triangle Park, NC.
Omoike, OE; Pack, RP; Mamudu, HM; Liu, Y; Wang, L. (2021). A cross-sectional study of the
association between perfluorinated chemical exposure and cancers related to deregulation of
estrogen receptors. Environmental Research 196: 110329.
http://dx.doij >/i.envres.2020.1.1.0329
29
-------
Shearer, JJ; Callahan, CL; Calafat, AM; Huang, WY; Jones, RR; Sabbisetti, VS, et al. (2021). Serum
concentrations of per- and polyfluoroalkyl substances and risk of renal cell carcinoma. Journal of
the National Cancer Institute 113: 580-587. http://dx.doi.org 3/inci/diaal43
Steenland, K; Fletcher, T; Stein, CR; Bartell, SM; Darrow, L; Lopez-Espinosa, MJ, et al. (2020). Review:
Evolution of evidence on PFOA and health following the assessments of the C8 Science Panel
[Review]. Environment International 145: 106125.
http://dx.dou ¦/i.envint.2020.106.1.25
Steenland, K; Hofmann, JN; Silverman, DT; Bartell, SM. (2022). Risk assessment for PFOA and kidney
cancer based on a pooled analysis of two studies. Environment International 167: 107425.
http://dx.doii ¦/i.envint.2022. .1.07425
Steenland, K; Woskie, S. (2012). Cohort mortality study of workers exposed to perfluorooctanoic acid.
American Journal of Epidemiology 176: 909-917. http://dx.doi.org/10.1093/aie/kws 171
Thayer, KA; Angrish, M; Arzuaga, X; Carlson, LM; Davis, A; Dishaw, L, et al. (2022). Systematic
evidence map (SEM) template: report format and methods used for the US EPA integrated risk
information system (iris) program, provisional peer reviewed toxicity value (PPRTV) program,
and other "fit for purpose" literature-based human health analyses (manuscript-in-progress) (pp.
1-69). Thayer, KA; Angrish, M; Arzuaga, X; Carlson, LM; Davis, A; Dishaw, L; Druwe, I;
Gibbons, C; Glenn, B; Jones, R; Kaiser, JP; Keshava, C; Keshava, N; Kraft, A; Lizarraga, L;
Persad, A; Radke, EG; Rice, G; Schulz, B; Shaffer, R; Shannon. T; Shapiro, A; Thacker, S;
Vulimiri, S; Williams, AJ; Woodall, G; Yost, E; Blain, R; Duke, K; Goldstone, AE; Hartman, P;
Hobbie, K; Ingle, B; Lemeris, C; Lin, C; Lindahl, A; McKinley, K; Soleymani, P; Vetter, N.
Thomford, PJ. (2002). 104-week dietary chronic toxicity and carcinogenicity study with perfluorooctane
sulfonic acid potassium salt (PFOS; T-6295) in rats (pp. 002148-002363). (Study No. 6329-183).
Madison, WI: Covance Laboratories.
Tsai, MS; Chang, SH; Kuo, WH; Kuo, CH; Li, SY; Wang, MY, et al. (2020). A case-control study of
perfluoroalkyl substances and the risk of breast cancer in Taiwanese women. Environment
International 142: 105850.
USEPA. (1985). National primary drinking water regulations; synthetic organic chemicals, inorganic
chemicals and microorganisms. Federal Register 50: 46936-47025.
USEPA. (1991). National primary drinking water regulationssynthetic organic chemicals and inorganic
chemicals; monitoring for unregulated contaminants; national primary drinking water regulations
implementation; national secondary drinking water regulations. Federal Register 56: 3526-3597.
USEPA. (1998). National Primary Drinking Water Regulations: Disinfectants and Disinfection
Byproducts, Federal Register: 63 FR 69390. 63: 69390-69476.
USEPA. (2000a). Methodology for deriving ambient water quality criteria for the protection of human
health (2000). (EPA/822/B-00/004). Washington, DC: U.S. Environmental Protection Agency,
Office of Water, http://www.epa.gov/waterscience/criteria/liumanhealth/method/complete.pdf
USEPA. (2000b). National Primary Drinking Water Regulations; Radionuclides; Final Rule. Federal
Register: 65 FR 76708. 65: 76708-76753.
USEPA. (2001). National Primary Drinking Water Regulations; Arsenic and Clarifications to Compliance
and New Source Contaminants Monitoring: Delay of Effective Date. Federal Register: 66 FR
28342. 66:28342-28350.
USEPA. (2005). Guidelines for carcinogen risk assessment [EPA Report]. (EPA630P03001F).
Washington, DC. https://www.epa.gov/sites/prodi.ietion/files/2013~
09/documents/cancer guidelines final 3-25-05.pdf
USEPA. (2016a). Health effects support document for perfluorooctane sulfonate (PFOS) [EPA Report].
(EPA 822-R-16-002). Washington, DC: U.S. Environmental Protection Agency, Office of Water,
Health and Ecological Criteria Division, https://www.epa.gov/sites/production/files/
05/documents/pfos hesd final 508.pdf
USEPA. (2016b). Health effects support document for perfluorooctanoic acid (PFOA) [EPA Report].
(EPA 822-R-16-003). Washington, DC: U.S. Environmental Protection Agency, Office of Water,
30
-------
Health and Ecological Criteria Division, https://www.epa.gov/sites/production/files/
05/documents/pfoa hesd final-plain.txlf
USEPA. (2016c). Six-year review 3 - Health effects assessment for existing chemical and radionuclide
national primary drinking water regulations summary report [EPA Report]. (EPA 822-R-16-008).
Washington, DC. https://nepis.epa.gov/Exe/ZvPURL.cgi?Dockev=P 10000WI.txt
USEPA. (2019). Exposure factors handbook chapter 3 (update): Ingestion of water and other select
liquids [EPA Report], (EPA/600/R-18/259F). Washington, DC.
https://cfbub.epa.gov/ncea/efb/recordisplav.cfm ?deid=343661
USEPA. (2021a). EXTERNAL PEER REVIEW DRAFT: Proposed Approaches to the Derivation of a
Draft Maximum Contaminant Level Goal for Perfluorooctane Sulfonic Acid (PFOS) (CASRN
1763-23-1) in Drinking Water [EPA Report]. Washington, DC: U.S. Environmental Protection
Agency (EPA). https://sab.epa.gov/ords/sab/f?p= .1.00: .1.8:164909479° ' M >:260.1.
USEPA. (2021b). Revisions to the Unregulated Contaminant Monitoring Rule (UCMR 5) for Public
Water Systems and Announcement of Public Meetings. Federal Register 86: 73131-73157.
USEPA. (2022a). ORD staff handbook for developing IRIS assessments [EPA Report]. (EPA 600/R-
22/268). Washington, DC: U.S. Environmental Protection Agency, Office of Research and
Development, Center for Public Health and Environmental Assessment.
https://cfbub.epa.gov/ncea/iris drafts/recordisplav.cfm?deid=356370
USEPA. (2022b). Review of EPA's Analysis to Support EPA's National Primary Drinking Water
Rulemaking for PFAS. (EPA-SAB-22-008). U.S. Environmental Protection Agency, Science
Advisory Board.
USEPA. (2022c). Transmittal of the Science Advisory Board Report titled, "Review of EPA's Analyses
to Support EPA's National Primary Drinking Water Rulemaking for PFAS.". (EPA-22-008).
https://sab.epa.gOv/ords/sab/i ^ II! L t'.'' *'V '77846
USEPA. (2023). Public Comment Draft - Toxicity Assessment and Proposed Maximum Contaminant
Level Goal (MCLG) for Perfluorooctane Sulfonic Acid (PFOS) (CASRN 1763-23-1) in Drinking
Water. (EPA-822-P-23-007).
USEPA. (2024a). Appendix - Office of Water Final Human Health Toxicity Assessment for
Perfluorooctane Sulfonic Acid (PFOS). (815R24009).
USEPA. (2024b). Appendix - Office of Water Final Human Health Toxicity Assessment for
Perfluorooctanoic Acid (PFOA). (815R24008).
USEPA. (2024c). EPA's responses to public comments on the proposed PFAS National Primary Drinking
Water Regulation [EPA Report]. (EPA-815-R-24-005).
USEPA. (2024d). Human health toxicity assessment for perfluorooctane sulfonic acid (PFOS) [EPA
Report]. Washington, DC: U.S. Environmental Protection Agency, Office of Water.
USEPA. (2024e). Human health toxicity assessment for perfluorooctanoic acid (PFOA) [EPA Report].
Washington, DC: U.S. Environmental Protection Agency, Office of Water.
USEPA. (2024f). PFAS National Primary Drinking Water Regulation rulemaking. Final rule. Federal
Register [pending publication].
USEPA. (2024g). PFAS occurrence and contaminant background support document for the final PFAS
NPDWR [EPA Report], (EPA-815-R-24-013).
USEPA OOW. (2023). EPA response to final Science Advisory Board recommendations (August 2022)
on four draft support documents for the EPA's proposed PFAS National Primary Drinking Water
Regulation [EPA Report]. (EPA Document No. 815D23001). Washington, DC: U1S.
Environmental Protection Agency, Office of Water.
https://www.epa.gov/system/files/documents/2Q23-
03/Response%20to%20SAB%20recommendations 3.9.23 For%20Proposal l.pdf
Vieira, VM; Hoffman, K; Shin, HM; Weinberg, JM; Webster, TF; Fletcher, T. (2013). Perfluorooctanoic
acid exposure and cancer outcomes in a contaminated community: a geographic analysis.
Environmental Health Perspectives 121: 318-323. http://dx.doi.org/10.1289/ehp. .1.205829
31
-------
Zahm, S; Bonde, JP; Chiu, WA; Hoppin, J; Kanno, J; Abdallah, M, et al. (2023). Carcinogenicity of
perfluorooctanoic acid and perfluorooctanesulfonic acid. The Lancet Oncology.
http://dx.doM ^ 10 lOlo N '0-2045(23)00622-8
32
------- |