EPA/635/R-19/049
K?c:rr\ IRIS Assessments Protocol
www.epa.gov/iris
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and
PFDA IRIS Assessments
CASRN 335-76-2 [PFDA]
CASRN 375-95-1 [PFNA]
CASRN 307-24-4 (PFHxA]
CASRN 355-46-4 [PFHxS]
CASRN 375-22-4 [PFBA]
October 2019
Integrated Risk Information System
Center for Public Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
DISCLAIMER
This document is a public comment draft. This information is distributed solely for review
purposes under applicable information quality guidelines. It has not been formally disseminated by
EPA. It does not represent and should not be construed to represent any Agency determination or
policy. It is being circulated for comments on its technical clarity and science policy implications.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
This document is a draft for review purposes only and does not constitute Agency policy.
ii DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
CONTENTS
AUTHORS | CONTRIBUTORS | REVIEWERS ii
1. INTRODUCTION 1
2. SCOPING AND PROBLEM FORMULATION SUMMARY 2
2.1. BACKGROUND 2
2.1.1. Chemical and Physical Properties 2
2.1.2. Sources, Production, and Use 5
2.1.3. Environmental Fate and Transport 6
2.1.4. Environmental Concentrations 7
2.1.5. Potential for Human Exposure 10
2.1.6. Populations and Lifestages with Potentially Greater Exposures 11
2.1.7. Other Environmental Protection Agency (EPA) Assessments of Per- and
Polyfluoroalkyl Substances (PFAS) 11
2.2.SCOPING SUMMARY 12
2.3. PROBLEM FORMULATION 15
2.3.1. Assessments and Toxicity Values from Other Sources 15
2.3.2. Preliminary Literature Inventory for the Five Per- and Polyfluoroalkyl Substances
(PFAS) Being Assessed 18
2.4. KEY SCIENCE ISSUES 21
2.4.1. Toxicokinetic Differences across Species and Sexes 22
2.4.2. Human Relevance of Effects in Animals that Involve Peroxisome
Proliferator-Activated Receptor Alpha (PPARa) Receptors 23
2.4.3. Potential Confounding by Other Per- and Polyfluoroalkyl Substances (PFAS)
Exposures in Epidemiology Studies 24
2.4.4. Toxicological Relevance of Changes in Certain Urinary and Hepatic Endpoints in
Rodents 25
2.4.5. Characterizing Uncertainty Due to Missing Chemical-Specific Information 25
3. OVERALL OBJECTIVES, SPECIFIC AIMS, AND POPULATIONS, EXPOSURES, COMPARATORS,
AND OUTCOMES (PECO) CRITERIA 26
3.1. SPECIFIC AIMS 27
3.2. POPULATIONS, EXPOSURES, COMPARATORS, AND OUTCOMES (PECO) CRITERIA 28
4. LITERATURE SEARCH AND SCREENING STRATEGIES 31
This document is a draft for review purposes only and does not constitute Agency policy.
iii DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
4.1. LITERATURE SEARCH STRATEGIES 31
4.1.1. Non-Peer-Reviewed Data 33
4.2.SCREENING PROCESS 34
4.2.1. Multiple Publications of the Same Data 37
4.2.2. Literature Flow Diagrams 37
4.3. SUMMARY-LEVEL LITERATURE INVENTORIES 42
5. REFINED EVALUATION PLAN 44
6. STUDY EVALUATION (REPORTING, RISK OF BIAS, AND SENSITIVITY) STRATEGY 55
6.1. STUDY EVALUATION OVERVIEW FOR HEALTH EFFECT STUDIES 55
6.2. EPIDEMIOLOGY STUDY EVALUATION 60
6.2.1. Epidemiology Study Evaluation Criteria Specific to These Five Per- and
Polyfluoroalkyl Substances (PFAS) 70
6.3. EXPERIMENTAL ANIMAL STUDY EVALUATION 73
6.3.1. Animal Toxicology Study Evaluation Considerations Specific to These Five
Per- and Polyfluoroalkyl Substances (PFAS) 83
6.4. PHARMACOKINETIC MODEL EVALUATION 84
6.5. MECHANISTIC STUDY EVALUATION 84
7. ORGANIZING THE HAZARD REVIEW 86
8. DATA EXTRACTION OF STUDY METHODS AND RESULTS 89
8.1. STANDARDIZING REPORTING OF EFFECT SIZES 90
8.2. STANDARDIZING ADMINISTERED DOSE LEVELS/CONCENTRATIONS 91
9. SYNTHESIS OF EVIDENCE 92
9.1. SYNTHESES OF HUMAN AND ANIMAL HEALTH EFFECTS EVIDENCE 97
9.2. MECHANISTIC INFORMATION 98
9.2.1. Toxicokinetic Information and Pharmacokinetic (PK)/Physiologically Based
Pharmacokinetic (PBPK) Models 99
9.2.2. Peroxisome Proliferator-Activated Receptor Alpha (PPARa) Dependence for
Health Effect(s) Observed in Animals 101
9.2.3. Toxicological Relevance of Select Outcomes Observed in Animals 105
9.2.4. Other Focused Mechanistic Analyses 106
10. EVIDENCE INTEGRATION 113
10.1. EVALUATING THE STRENGTH OF THE HUMAN AND ANIMAL EVIDENCE STREAMS 116
10.2. OVERALL EVIDENCE INTEGRATION JUDGMENTS 121
10.3. HAZARD CONSIDERATIONS FOR DOSE-RESPONSE 128
This document is a draft for review purposes only and does not constitute Agency policy.
iv DRAFT-DO NOT CITE OR QUOTE
This document is a draft for review purposes only and does not constitute Agency policy.
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
11. DOSE-RESPONSE ASSESSMENT: SELECTING STUDIES AND QUANTITATIVE ANALYSIS 130
11.1. SELECTING STUDIES FOR DOSE-RESPONSE ASSESSMENT 131
11.2. CONDUCTING DOSE-RESPONSE ASSESSMENTS 135
11.2.1. Dose-Response Analysis in the Range of Observation 135
11.2.2. Extrapolation: Slope Factors and Unit Risk 138
11.2.3. Extrapolation: Reference Values 139
12. PROTOCOL HISTORY 142
REFERENCES R-l
APPENDIX A. SUMMARY OF EXISTING TOXICITY VALUE INFORMATION FOR
PERFLUOROBUTANOIC ACID (PFBA), PERFLUOROHEXANOIC ACID (PFHXA),
PERFLUORONONANOIC ACID (PFNA), AND PERFLUORODECANOIC ACID (PFDA) A-l
APPENDIX B. SEARCH AND SCREENING STRATEGIES B-l
This document is a draft for review purposes only and does not constitute Agency policy.
v DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
TABLES
Table 1. Physiochemical properties of five per- and polyfluoroalkyl substances (PFAS) and their
related salts 4
Table 2. Per- and polyfluoroalkyl substances (PFAS) levels at 10 military installations 9
Table 3. Per- and polyfluoroalkyl substances (PFAS) levels in water, soil, and air at National
Priorities List sites 9
Table 4. Serum per- and polyfluoroalkyl substances (PFAS) concentrations based on National
Health and Nutrition Examination Survey (NHANES) 2013-2014 data (ng/L) 10
Table 5. Environmental Protection Agency (EPA) considerations for the selection of per- and
polyfluoroalkyl substances (PFAS) for evaluation 13
Table 6. Potential Environmental Protection Agency (EPA) needs and applications for five
per- and polyfluoroalkyl substances (PFAS) 14
Table 7. Preliminary serum half-life estimates of five per- and polyfluoroalkyl substances (PFAS)
across species and sexes 23
Table 8. Populations, exposures, comparators, and outcomes (PECO) criteria 29
Table 9. Systematic map of epidemiology outcomes 46
Table 10. Animal Endpoint Grouping Categories 50
Table 11. Questions and criteria for evaluating each domain in epidemiology studies 61
Table 12. Criteria for evaluating exposure measurement in epidemiology studies of per- and
polyfluoroalkyl substances (PFAS) and health effects 71
Table 13. Considerations to evaluate domains from animal toxicology studies 74
Table 14. Querying the evidence to organize syntheses for human and animal evidence 87
Table 15. Information most relevant to describing primary considerationsfor assessing causality
during evidence syntheses 94
Table 16. Individual and social factors that may increase susceptibility to exposure-related
health effects 97
Table 17. Examples of questions and considerations that can trigger focused analysis and
synthesis of mechanistic information 107
Table 18. Evidence profile table template 115
Table 19. Considerations that inform evaluations of the strength of the human and animal
evidence 117
Table 20. Evidence integration judgments for characterizing potential human health hazards in
the evidence integration narrative 123
Table 21. Attributes used to evaluate studies for deriving toxicity values 133
This document is a draft for review purposes only and does not constitute Agency policy.
vi DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
FIGURES
Figure 1. Chemical structures of per- and polyfluoroalkyl substances (PFAS) 3
Figure 2. Existing Oral Reference Values for (a) perfluorobutanoic acid (PFBA), (b)
perfluorohexanoic acid (PFHxA), (c) perfluorohexanesulfonate (PFHxS), (d)
perfluorononanoic acid (PFNA), and (e) perfluorodecanoic acid (PFDA) 18
Figure 3. Results of a preliminary literature inventory of five per- and polyfluoroalkyl substances
(PFAS). Data are approximated based on a cursory review of the literature
search results for studies published through 2018 (see Section 4 for details).
Health effects are based on groupings from EPA Integrated Risk Information
System (IRIS) website (https://cfpub.epa.gov/ncea/iris/search/index.cfm).a For
this summary, metabolic effects are captured under "other" and "hepatic"
includes lipid and lipoprotein measures 20
Figure 4. Literature flow diagrams for PFBA and its ammonium salt (a), PFHxA and its
ammonium and sodium salts (b), PFHxS and its potassium salt (c), PFNA and its
ammonium and sodium salts (d), and PFDA and its ammonium and sodium salts
(e) 42
Figure 5. Overview of Integrated Risk Information System (IRIS) study evaluation process, (a) An
overview of the general evaluation process [Note: see Section 5 for deviations
from independent evaluation by two reviewers for some health outcomes in
epidemiology studies], (b) The evaluation domains and definitions for ratings
(i.e., domain and overall judgments, performed on an outcome-specific basis) 56
Figure 6. Preliminary mean correlation coefficients across per- and polyfluoroalkyl substances
(PFAS) among studies in the inventory, for all media types 72
Figure 7. Preliminary proposed mechanistic pathway for per- and polyfluoroalkyl substances
(PFAS)-induced noncancer liver effects. Based on previous reviews of
perfluorooctane sulfonate (PFOS)- and perfluorooctanoic acid (PFOA)-induced
noncancer liver effects in animals (ATSDR, 2018; Li et al., 2017; Viberg and
Eriksson, 2017; U.S. EPA, 2016c, d), and proposed adverse outcome pathways
for hepatic steatosis (Mellor et al., 2016) 103
Figure 8. Process for evidence integration 114
This document is a draft for review purposes only and does not constitute Agency policy.
vii DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
ABBREVIATIONS
AD ME
AFFF
AK DEC
ALT
AOP
AST
ATSDR
BMDL
BMI
BMR
BW3/4
CAR
CAS
CASRN
CBI
CERCLA
CLA
CLH
CPAD
CPHEA
CPN
CRD
CTDPH
CTL
CWA
DNA
DTH
DWEL
ECHA
EFSA
EPA
FDA
FIFRA
FOB
FXR
GLP
GRADE
HA
HAWC
absorption, distribution, metabolism, HED
and excretion HERO
aqueous film-forming foam
Alaska Department of Environmental HFPO
Conservation hPPARa
alanine aminotransferase
adverse outcome pathway HRL
aspartate transaminase i.p.
Agency for Toxic Substances and IARC
Disease Registry
benchmark dose lower confidence limit IPCS
body mass index
benchmark response IRIS
body-weight scaling to the 3/4 power IUR
constitutive androstane receptor K
Chemical Abstracts Service LDr.o
Chemical Abstracts Service registry LOAEL
number LOD
confidential business information MAC
Comprehensive Environmental MCL
Response, Compensation, and Liability MDH
Act MF
clearance in animals MLR
clearance in humans MOA
Chemical and Pollutant Assessment MPPD
Division MRL
Center for Public Health and Na
Environmental Assessment NAFLD
chronic progressive nephropathy ND
chemical reporting data NF-kB
Connecticut Department of Health NH4+
cytotoxic T lymphocyte NHANES
Clean Water Act
deoxyribonucleic acid NH DES
delayed-type hypersensitivity
drinking water equivalent level NJ D EP
European Chemicals Agency
European Food Safely Authority NMD
Environmental Protection Agency NOAEL
Food and Drug Agency NPDWR
Federal Insecticide, Fungicide, and
Rodenticide Act NPL
functional operational battery NR
farnesoid X receptor NTP
good laboratory practices OCSPP
Grading of Recommendations
Assessment, Development and OECD
Evaluation
health advisory OLEM
Health Assessment Workspace
Collaborative OR
human equivalent dose
Health and Environmental Research
Online
hexafluoropropylene oxide
humanized peroxisome
proliferator-activated receptor alpha
health risk limit
intraperitoneal
International Agency for Research on
Cancer
International Programme on Chemical
Safety
Integrated Risk Information System
inhalation unit risk
potassium
median lethal dose
lowest-observed-adverse-effect level
limit of detection
maximum acceptable concentration
maximum contaminant level
Minnesota Department of Health
modifying factor
mixed leukocyte reaction
mode of action
multiple path particle dosimetry
minimum reporting level
sodium
nonalcoholic fatty liver disease
no data
nuclear factor kappa B pathway
ammonium
National Health and Nutrition
Examination Survey
New Hampshire Department of
Environmental Services
New Jersey Department of
Environmental Protection
normalized mean difference
no-observed-adverse-effect level
National Primary Drinking Water
Regulation
National Priorities List
nuclear receptor
National Toxicology Program
Office of Chemical Safety and Pollution
Prevention
Organisation for Economic
Co-operation and Development
Office of Land and Emergency
Management
odds ratio
This document is a draft for review purposes only and does not constitute Agency policy.
viii DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
ORD
Office of Research and Development
PWS
public water system
OSF
oral slope factor
PXR
pregnane X receptor
OW
Office of Water
RCRA
Resource Conservation and Recovery
PAC
protective action criteria
Act
PBPK
physiologically based pharmacokinetic
RfC
inhalation reference concentration
PBTK
physiologically based toxicokinetic
RfD
oral reference dose
PCL
protective concentration level
ROBINS-I
Risk of Bias in Nonrandomized Studies
PECO
populations, comparators, exposures,
of Interventions
and outcomes
ROS
reactive oxygen species
PFAS
per- and polyfluoroalkyl substances
RXR
retinoid X receptor
PFBA
perfluorobutanoic acid
SD
standard deviation
PFBS
perfluorobutane sulfonate
SDWA
Safe Drinking Water Act
PFCA
perfluoroalkyl carboxylic acid
T 0.5A
elimination half-life in animals
PFDA
perfluorodecanoic acid
T 0.5H
elimination half-life in humans
PFHxA
perfluorohexanoic acid
TCEQ
Texas Commission on Environmental
PFHxS
perfluorohexanesulfonate
Quality
PFNA
perfluorononanoic acid
TD
toxicodynamic
PFOA
perfluorooctanoic acid
TDI
tolerable daily intake
PFOS
perfluorooctane sulfonate
TEEL
temporary emergency exposure limit
PFSA
perfluoroalkane sulfonic acid
TNFa
tumor necrosis factor alpha
PI3K-Akt
phosphatidylinositol-3-kinase-
TRI
Toxics Release Inventory
serine/threonine kinase Akt
TSCA
Toxic Substances Control Act
PK
pharmacokinetic
TSCATS
Toxic Substances Control Act Test
POD
point of departure
Submissions
PPARa
peroxisome proliferator-activated
UCMR
Uncontaminated Monitoring Rule
receptor alpha
UF
uncertainty factor
PPRTV
Provisional Peer-Reviewed Toxicity
Vd
volume of distribution
Value
WHO
World Health Organization
PR
preliminary review
wt
weight
pt
point
XME
xenobiotic metabolizing enzymes
PVDF
polyvinylidene fluoride
This document is a draft for review purposes only and does not constitute Agency policy.
ii DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
AUTHORS | CONTRIBUTORS | REVIEWERS
Assessment Teams (U.S. EPA/Center for Public Health and Environmental Assessment [CPHEA])
Michelle Angrish (PFHxA co-lead)
Xabier Arzuaga (PFHxS co-lead)
Thomas Bateson
Allen Davis (PFBA and statistics co-lead)
Laura Dishaw (PFHxA co-lead)
Ingrid L. Druwe (PFHxS co-lead)
J. Phillip Kaiser (PFDA co-lead)
Andrew Kraft (PFNA co-lead)
Lucina Lizarraga (PFDA co-lead)
Pamela Noyes (PFNA co-lead)
Elizabeth Radke (epidemiology lead)
Kristen Rappazzo
Alan Sasso
Paul Schlosser (PBPK and ADME lead)
Michele Taylor (PFBA and PFNA co-lead)
Michael Wright
Jay Zhao (statistics co-lead)
Technical Experts/Contributors (U.S. EPA/CPHEA)
Audrey Galizia Brittany Schulz (ORAU contractor)
Kelly Garcia (Oak Ridge Associated Universities Andre Weaver
[ORAU] contractor, no longer with EPA)
Carolyn Gigot (ORAU contractor, no longer with Scott Wesselkamper (no longer with EPA)
EPA)
Andrew Greenhalgh (ORAU contractor, no longer Amina Wilkins
with EPA)
Belinda Hawkins George Woodall
Amanda Persad
Linda Phillips (retired)
This document is a draft for review purposes only and does not constitute Agency policy.
ii DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Technical Experts/Contributors (outside of U.S. EPA)
John Bucher
Andrew Rooney
Kyi a Taylor
National Toxicology Program/Office of Health Assessment
and Translation
National Toxicology Program/Office of Health Assessment
and Translation/Director
National Toxicology Program/Office of Health Assessment
and Translation
Executive Direction
Tina Bahadori
Samantha Jones (former PFAS assessment team
lead)
Kristina Thayer
Andrew Kraft (PFAS assessment team lead)
U.S. EPA/Office of Research and Development/Senior Science
Advisor
U.S. EPA/CPHEA/Associate Director
U.S. EPA/CPHEA/Chemical and Pollutant Assessment Division
(CPAD)/Director
U.S. EPA/CPHEA/CPAD/Senior Science Advisor
Production Team and Review (U.S. EPA/CPHEA)
Anna Chaplin
Madison Feshuk (ORAU contractor)
Catherine Gibbons
Hillary Hollinger (ORAU contractor)
Ryan Jones
Jennifer Nichols
Dahnish Shams
Vicki Soto
This document is a draft for review purposes only and does not constitute Agency policy.
iii DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
1. INTRODUCTION
Per- and polyfluoroalkyl substances (PFAS] are a large class of synthetic (man-made]
chemicals widely used in consumer products and industrial processes. The basic structure of PFAS
consists of a carbon chain surrounded by fluorine atoms, with different chemicals possessing
different end groups (see examples in Section 2.1.1]; thousands of distinct PFAS exist in commerce.
To help address this complex issue, the Environmental Protection Agency (EPA] is taking a
proactive approach. Specifically, the development of human health toxicity assessments for
exposure to individual PFAS represents only one component of the broader PFAS action plan
underway at the U.S. EPA fhttps: //www.epa.gov/pfas/epas-pfas-action-planl The five toxicity
assessments being developed according to the scope and methods outlined in this protocol build
upon several other PFAS assessments that have already been developed (see Section 2.1.7],
This protocol document presents the methods for conducting the systematic reviews and
dose-response analyses for assessments of perfluorodecanoic acid (PFDA], perfluorononanoic acid
(PFNA], perfluorohexanoic acid (PFHxA], perfluorohexanesulfonate (PFHxS], and
perfluorobutanoic acid (PFBA], and their related salts (see Figure 1], This includes a summaiy of
why these specific PFAS were prioritized for evaluation, description of the objectives and specific
aims of the assessments, draft populations, exposures, comparators, and outcomes (PECO] criteria,
and identification of key areas of scientific complexity. This assessment protocol will be posted on
the Integrated Risk Information System (IRIS] website fhttps://cfpub.epa.gov/ncea/iris2/atoz.cfml
for a 45-day comment period. The protocol will also be published in the Zenodo data repository
(https: //zen0d0.0rg/I Public input received on the protocol is considered during preparation of
the draft assessments and any adjustments made to the protocol will be reflected in an updated
version released in conjunction with the draft assessments. The literature search results for these
five PFAS will also be posted to the Health and Environmental Research Online (HERO] database1
upon public release of the protocol (the literature search results will be regularly updated during
draft development and the subsequent stages of assessment review].
1PFBA: https://hero.epa.gov/hero/index.cfm/proiect/page/proiect id/2632
PFHxA: https://hero.epa.gov/hero/index.cfm/proiect/page/proiect id/2628
PFHxS: https://hero.epa.gov/hero/index.cfm/proiect/page/proiect id/2630
PFNA: https://hero.epa.gov/hero/index.cfm/proiect/page/proiect id/2633
PFDA: https://hero.epa.gov/hero/index.cfm/proiect/page/proiect id/2614
This document is a draft for review purposes only and does not constitute Agency policy.
1 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
2. SCOPING AND PROBLEM FORMULATION
SUMMARY
2.1. BACKGROUND
2.1.1. Chemical and Physical Properties
Perfluorodecanoic acid (PFDA; CASRN 335-76-2], perfluorononanoic acid (PFNA;
CASRN 375-24-4], perfluorohexanoic acid (PFHxA, CASRN 307-24-4], perfluorohexanesulfonate
(PFHxS, CASRN 355-46-4], and perfluorobutanoic acid (PFBA, CASRN 375-22-4], and their related
salts, are members of the group PFAS. Section 2.2 ("Scoping Summary"] outlines the rationale for
why these PFAS were prioritized for assessment Buck etal. f201D define PFAS as fluorinated
substances that "contain 1 or more C atoms on which all the H substituents (present in the
nonfluorinated analogues from which they are notionally derived] have been replaced by F atoms,
in such a manner that they contain the perfluoroalkyl moiety (CnF2n+r]" More specifically, PFDA,
PFNA, PFHxA, and PFBA are classified as perfluoroalkyl carboxylic acids (PFCAs], and PFHxS is a
perfluoroalkane sulfonic acid (PFSA] fOECD. 20151. PFCAs containing seven or more
perfluorinated carbon groups and PFSAs containing six or more perfluorinated carbon units are
considered long-chain PFAS fATSDR. 2018: OECD. 2015: Buck etal.. 20111 Thus, PFDA, PFNA, and
PFHxS are considered long-chain PFAS, and PFHxA and PFBA are short-chain PFAS. The chemical
structures of PFDA, PFNA, PFHxA, PFHxS, and PFBA, and their related salts, are presented in Figure
1 (along with their CASRNs], and their physiochemical properties are provided in Table 1.
This document is a draft for review purposes only and does not constitute Agency policy.
2 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
PFDA
335-76-2
F'Y F
PFNA
375-95-1
PFHxA
307-24-4
PFHxS
355-46-4
O
OH
f-ff
PFBA
375-22-4
FFFFFFFFF
NHf F
F F F
PFDA
sodium salt
3830-45-3
PFNA
sodium salt
21049-39-8
PFHxA
sodium salt
2923-26-4
PFBA
ammonium salt
10495-86-0
PFDA
ammonium salt
3108-42-7
0*^0
PFNA
ammonium salt
4149-60-4
NHa O
PFHxA
ammonium salt
21615-47-4
F F F F F F P
F F F F F F 0
PFHxS
potassium salt
3871-99-6
Figure 1. Chemical structures of per- and polyfluoroalkyl substances (PFAS).
This document is a draft for review purposes only and does not constitute Agency policy.
3 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 1. Physiochemical properties of five per- and polyfluoroalkyl
substances (PFAS) and their related salts
Property
(unit)
PFDA + salts
PFNA + salts
PFHxA
PFHxS + salts
PFBA + salts
PFDA3
nh4+
saltb
Na
salt0
PFNAd
nh4+
salt6
Na
saltf
PFHxA®
PFHxSh
K
salt1
PFBAj
nh4+
saltk
Molecular wt.
(g/mol)
514
531
536
464
481
486
314
400
438
214
ND
Melting pt. (°C)
79.5
165*
71.4*
66.5
165*
ND
14.0
190
273
-17.5
ND
Boiling pt. (°C)
218
205*
205*
218
190*
196
157
345
236*
121
ND
Density (g/cm3)
1.79*
ND
ND
1.78*
ND
ND
1.69*
1.84*
1.84*
1.65
ND
Vapor pressure
(mm Hg)
3.89 x
10-2*
3.51 x
10-2*
3.51 x
10-2*
7.06 x
10-2*
7.38 x
10-2*
1.71 x
10"1
9.08 x
10"1
9.95 x
10-s*
9.95 x
10-s*
6.37
ND
Henry's law
constant
(atm-m3/mole)
3.55 x
10"10*
3.55 x
10"10*
3.55 x
10"10*
1.64 x
10"10*
1.64 x
10"10*
ND
2.51 x
10"10*
1.96 x
10"10*
1.96 x
10"10*
4.99 x
10-s*
ND
Water solubility
(mol/L)
8.3 x
10"6*
8.41 x
10-i*
8.41 x
10-i*
1.38 x
10-s*
8.45 x
10-i*
ND
9.39 x
10"5
6.08 x
10"4
2.96 x
10"4*
2.09 x
10"3
ND
pKa
-0.17*
ND
ND
-0.17*
ND
ND
-0.16
0.14*
ND
0.08*
ND
LogP
7.37*
4.36*
4.36*
6.59*
3.12*
ND
2.51
2.20
2.97*
1.43
ND
Soil adsorption
coefficient (L/kg)
397*
397*
397*
2,830*
2,830*
ND
1,070*
2,300*
2,300*
47.9*
ND
Bioconcentration
factor
789*
29.8*
29.8*
752*
4.95*
ND
41*
118*
5.94*
7.61*
ND
K = potassium; ND = no data; NH4+ = ammonium; Na = sodium; pt. = point; wt. = weight.
aCASRN 335-76-2. U.S. EPA (2018a) Chemistry Dashboard (https://comptox.epa.gov/dashboard/; search = PFDA)
for all values except pKa (ATSDR, 2018). Values are median experimental values (when available), or median or
average predicted values.
bCASRN 3108-42-7. Predicted average values from the U.S. EPA (2018a) Chemistry Dashboard
(search = 3108-42-7).
CCASRN 3830-45-3. Predicted average values from the U.S. EPA (2018a) Chemistry Dashboard
(search = 3830-45-3).
dCASRN 375-95-1. U.S. EPA (2018a) Chemistry Dashboard (search = PFNA) for all values except pKa (NLM, 2013).
Values are median experimental values (when available), or median or average predicted values.
eCASRN 4149-60-4. Predicted average values from the U.S. EPA (2018a) Chemistry Dashboard
(search = 4149-60-4).
fCASRN 21049-39-8. ChemNet website:
http://www.chemnet.com/cas/en/21049-39-8/sodium-heptadecafluorononanoate.html for all values.
gCASRN: 307-24-4. U.S. EPA (2018a) Chemistry Dashboard (search = 307-24-4) for all values except pKa (NLM,
2016). Values are median or average experimental values (when available), or median or average predicted
values.
hCASRN 355-46-4. U.S. EPA (2018a) Chemistry Dashboard (search = 355-46-4) for all values except pKa (NLM,
2017). Values are median or average experimental values (when available), or median or average predicted
values.
This document is a draft for review purposes only and does not constitute Agency policy.
4 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Property
(unit)
PFDA + salts
PFNA + salts
PFHxA
PFHxS + salts
PFBA + salts
PFDA3
nh4+
saltb
Na
salt0
PFNAd
nh4+
salt6
Na
saltf
PFHxA®
PFHxSh
K
salt1
PFBAj
nh4+
saltk
'CASRN 3871-99-6. U.S. EPA (2018a) Chemistry Dashboard (search = 3871-99-6) for all values. Values are median
or average experimental values (when available), or median or average predicted values.
JCASRN 375-22-4. U.S. EPA (2018a) Chemistry Dashboard (search = 375-22-4)
https://comptox.epa.gov/dashboard/dsstoxdb/results?utf8=%E2%9C%93&search=375-22-4 for all values except
pKa (ATSDR, 2018). Values are median or average experimental values (when available), or median or average
predicted values.
kCASRN 10495-86-0
* Predicted value.
1
2.1.2. Sources, Production, and Use
2 PFAS are synthetic (man-made] compounds that have been used since the 1940s in
3 consumer products and industrial applications because of their resistance to heat, oil, stains,
4 grease, and water. They have been used in stain-resistant fabrics for clothing, carpets, and
5 furniture; nonstick cookware; food packaging (e.g., popcorn bags, and fast-food containers]; and
6 personal care products (e.g., dental floss, cosmetics, and sunscreen] fATSDR. 20181 Some PFAS
7 have also been used in firefighting foam and as industrial surfactants, emulsifiers, wetting agents,
8 additives, and coatings, and in the aerospace, automotive, building and construction industries to
9 help reduce friction fATSDR. 20181 Because of their widespread use, the release of PFAS into the
10 air, water, and soil, and their persistence in the environment, most people in the United States have
11 been exposed to PFAS (see https://www.epa.gov/pfas/epas-pfas-action-plan for additional details],
12 Examples of how the five PFAS of interest have been used include:
13
14 • PFDA has been used in stain and grease-proof coatings on food packaging furniture,
15 upholstery, and carpet fHarbison etal.. 20151 and as a lubricant, wetting agent, plasticizer,
16 and corrosion inhibitor fKeml. 20151
17 • PFNA has been used as a processing aid in the production of fluoropolymers, primarily
18 polyvinylidene fluoride (PVDF], which is a plastic designed to be temperature resistant and
19 chemically nonreactive fNTDWOI. 2017: Prevedouros etal.. 20061 It has also been used in
20 aqueous film-forming foam (AFFF] for fire suppression fLaitinen etal.. 20141
21 • PFHxA is not currently a commercial product; it is a breakdown product of "stain- and
22 grease-proof coatings on food packaging and household products" (NTP. 2018bl It has
23 been proposed as a replacement for the commonly used perfluorooctanoic acid (PFOA] and
24 perfluorooctane sulfonate (PFOS] fKlaunig et al.. 20151
25 • PFHxS has been used as a surfactant to make fluoropolymers, and in water- and
26 stain-protective coatings for carpets, paper, and textiles (NTP. 2018al It may also be
27 present in certain industrial and consumer products such as "fire-fighting foams,
28 food-contact papers, water-proofing agents, cleaning and polishing products either for
This document is a draft for review purposes only and does not constitute Agency policy.
5 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
intentional uses (as surfactants or surface protection agents] or as unintentional impurities
from industrial production processes" fNorwegian Environment Agency. 20181
• PFBA is a breakdown product of other PFAS that are used in stain-resistant fabrics, paper
food packaging, and carpets; it was also used for manufacturing photographic film fMDH.
20091
The U.S. Environmental Protection Agency (EPA] has been working with companies in the
fluorochemical industry since the early 2000s to phase out the production and use of long-chain
PFAS (https://www.epa.gov/assessing-and-managing-
chemicals-under-tsca/risk-management-and-polvfluoroalkvl-substances-pfassl However, the past
production and use of these PFAS has resulted in their release to the environment through various
waste streams fNLM. 2016. 20131 Also, because products containing PFAS are still in use, they
continue to be a source of environmental PFAS contamination through their disposal and
subsequent breakdown in the environment (Kim and Kannan. 20071
Chemical reporting data (CRD] on production volumes are not available in EPA's ChemView
(U.S. EPA. 20191 for PFDA, PFNA, PFHxA, PFHxS, PFBA, or their salts. Also, because there are no
requirements to report releases to the environment from facilities manufacturing, processing, or
otherwise using PFAS, quantitative information is not available in EPA's Toxics Release Inventory
(TRI] CU.S. EPA. 2019: ATSDR. 20181
2.1.3. Environmental Fate and Transport
PFAS are very stable and persistent in the environment (ATSDR. 20181 and many are found
worldwide in the environment, wildlife, and humans (https://www.epa.gov/assessing-and-
managing-chemicals-under-tsca/risk-management-and-polvfluoroalkvl-substances-pfassl They
have been detected at a variety of sites, including private and federal facilities, and have been
associated with various sources, including AFFF, chrome-plating facilities, PFAS manufacturers, and
industries that use PFAS (e.g., textiles] (ATSDR. 20181 The environmental fate and transport of
PFAS potentially can include releases to air to soil and surficial water bodies which can then lead to
migration to subsurface soils and ground water contamination fGuelfo etal..
2 0181 (https://www.atsdr.cdc.gov/pfas/index.htmll
PFAS released to air exist in the vapor phase and resist photolysis, but particle-bound
concentrations have also been measured fNLM. 2017. 2016. 2013: Kim and Kannan. 20071 The
atmospheric half-life for degradation by reaction with photochemically produced hydroxyl radicals
is estimated to be 31 days for PFNA and PFHxA, and 115 days for PFHxS (NLM. 2017. 2016. 20131
Long-range atmospheric transport of PFAS is possible, as indicated by the detection of PFHxS in
remote arctic and marine air samples (NLM. 20171 Wet and dry deposition are potential removal
processes for particle-bound PFAS in air (e.g., to surface water or soil] (ATSDR. 20181 Standardized
analytical methods for measuring these five PFAS in ambient air is an area of ongoing research.
This document is a draft for review purposes only and does not constitute Agency policy.
6 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
In soil, the mobility of PFAS will vary depending on their soil adsorption coefficients (see
Table 1], with PFNA predicted to be the least mobile and PFBA the most mobile of the five PFAS
addressed here. Volatilization of PFNA, PFHxA, and PFHxS from moist soil is not expected to be an
important transport process fNLM. 2017. 2016. 20131 Uptake of soil PFAS to plants can occur
fATSDR. 20181 Yoo etal. f20111 estimated grass-soil accumulation factors (grass concentration
divided by soil concentration] of 3.4, 0.12, and 0.10 for PFHxA, PFNA, and PFDA, respectively, based
on samples collected from a site with bio-solids-amended soil. Zhao etal. (20161 observed that
shorter chain PFAS like PFBA were transported more readily from the roots to the shoots of wheat
plants than longer chain PFAS.
PFNA, PFHxA, and PFHxS are expected to adsorb to suspended solids and sediments in
water fNLM. 2017. 2016. 20131 and hydrolysis is not expected to be an important fate process
fATSDR. 20181 The potential for PFAS to bioaccumulate in aquatic organisms can be assessed
using their bioconcentration factors, with the predicted potential for PFDA and PFNA to
bioaccumulate being high compared to PFHxA, PFHxS, and PFBA (see Table 1], As described in
Section 2.2, standardized analytical methods for measuring these five PFAS in drinking water exist
(for 4 of the 5 PFAS to be assessed] or are under development (i.e., for PFBA], Non-drinking water
methods are currently under development.
2.1.4. Environmental Concentrations
PFDA, PFNA, PFHxA, PFHxS, and PFBA have not been evaluated under the National Air
Toxics Assessment program (https: //www.epa.gov/national-air-toxics-assessmentl However,
PFDA, PFNA, and PFHxS were measured at concentrations ranging from less than the limit of
detection (LOD] to 1.56 pg/m3 in the vapor and particle phases of air samples collected from an
urban area of Albany, NY in 2006 fKim and Kannan. 20071 PFAS have also been measured in
indoor air and dust, and they may be associated with the indoor use of consumer products such as
PFAS-treatedcarpets or other textiles fATSDR. 20181 For example, Kato etal. (20091 analyzed
dust samples collected from 39 homes in the United States, United Kingdom, Germany, and
Australia for PFAS, including PFDA, PFNA, PFHxA, and PFHxS. These PFAS were detected in 38.5,
25.6, 46.2, and 79.5% of the samples, respectively. Likewise, Strvnar and Lindstrom (20081
analyzed dust samples from 110 homes and 10 day care centers in North Carolina and Ohio, and
detected PFDA, PFNA, and PFHxA in 30.4, 42.9, and 92.9% of the samples, respectively. Indoor air
samples (N = 4] from a town in Norway had mean concentrations of 3.4 pg/m3 for PFDA, 2.7 pg/m3
for PFNA, and <4.1 pg/m3 for PFHxS (Barber etal.. 20071.
The levels of PFAS in soil and sediment surrounding perfluorochemical industrial facilities
have been measured at concentrations ranging from less than the LOD to 124 ng/g for PFBA and
less than the LOD to 3,470 ng/g for PFHxS fATSDR. 20181. PFDA, PFNA, PFHxA, PFHxS, and PFBA
were also detected at an Australian training ground where AFFFs had been used (Baduel et al„
20151. PFDA, PFNA, PFHxA, PFHxS, and PFBA were detected at 10 U.S. military sites in 67.0, 71.4,
70.3, 76.9, and 38.5% of the surface soil samples, respectively, and 48.5,12.1, 63.6, 72.7, and 24.2%
This document is a draft for review purposes only and does not constitute Agency policy.
7 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
of the sediment samples, respectively fATSDR. 20181 Table 2 shows the concentrations of these
PFAS in soil and sediment at these militaiy sites.
EPA conducted monitoring for several PFAS in drinking water as part of the third
Uncontaminated Monitoring Rule (UCMR] fU.S. EPA. 2016el Under the UCMR, all public water
systems (PWSs] serving more than 10,000 people and a representative sample of 800 PWSs serving
10,000 or fewer people were monitored for 30 unregulated contaminants between January 2013
and December 2015. PFNA and PFHxS were among the 30 contaminants monitored. PFNA was
detected above the minimum reporting level (MRL] of 0.02 |ig/L in 14 of the 4,920 PWSs tested and
in 19 of the 36,972 samples collected. PFHxS was detected above the MRL of 0.03 |ig/L in 55 of the
4,920 PWSs tested and in 207 of the 36,971 samples collected. UCMR data were not available for
PFDA, PFHxA, or PFBA. However, samples from seven municipal wells in Oakdale, MN were
analyzed for PFHxA and PFBA. The concentrations ranged from <0.025 to 0.235 |J.g/L and 0.0855 to
2.04 |ig/L, respectively (U.S. EPA. 2017bl Kim and Kannan (20071 analyzed lake water, rain water,
snow, and surface water from Albany, NY, and reported concentrations of PFDA, PFNA, and PFHxS
ranging from less than the LOD to 0.0135 |ig/L. PFAS were detected at higher concentrations in
groundwater samples from an industrial site (3M Cottage Grove] in Minnesota. PFHxS and PFBA
were detected in all 7 wells that were sampled at concentrations ranging from 6.47-40 |ig/L and
23.3-318 |ig/L, respectively [fWS. 20071 as cited in fATSDR. 20181], The concentrations ofthese
five PFAS measured at National Priorities List (NPL] sites are shown in Table 3, and the
concentrations of PFAS measured in surface water and groundwater at 10 military installations are
given in Table 2.
This document is a draft for review purposes only and does not constitute Agency policy.
8 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 2. Per- and polyfluoroalkyl substances (PFAS) levels at 10 military
installations
Media
Measure
PFDA
PFNA
PFHxA
PFHxS
PFBA
Surface soil
Frequency of detection (%)
67.03
71.43
70.33
76.92
38.46
Reporting limit (ng/kg)
0.28
0.23
0.16
0.29
0.12
Median (ng/kg)
0.980
1.30
1.75
5.70
1.00
Maximum (ng/kg)
15.0
23.0
51.0
1,300
31.0
Subsurface soil
Frequency of detection (%)
12.50
14.42
65.38
59.62
29.81
Reporting limit (ng/kg)
0.30
0.24
0.16
0.31
0.13
Median (ng/kg)
1.40
1.50
1.04
4.40
0.960
Maximum (ng/kg)
9.40
6.49
140
520
14.0
Sediment
Frequency of detection (%)
48.48
12.12
63.64
72.73
24.24
Reporting limit (ng/kg)
0.46
0.38
0.26
0.48
0.21
Median (ng/kg)
1.90
1.10
1.70
9.10
1.70
Maximum (ng/kg)
59.0
59.0
710
2,700
140
Surface water
Frequency of detection (%)
52.00
36.00
96.00
88.00
84.00
Reporting limit (ng/L)
0.008
0.017
0.003
0.007
0.010
Median (|Jg/L)
0.067
0.096
0.320
0.710
0.076
Maximum (ng/L)
3.20
10.0
292
815
110
Groundwater
Frequency of detection (%)
34.78
46.38
94.20
94.93
85.51
Reporting limit (ng/L)
0.008
0.018
0.003
0.007
0.010
Median (|Jg/L)
0.023
0.105
0.820
0.870
0.180
Maximum (ng/L)
1.80
3.00
120
290
64.0
Source: (ATSDR. 2018).
1
Table 3. Per- and polyfluoroalkyl substances (PFAS) levels in water, soil, and
air at National Priorities List sites
Media
Measure
PFDA
PFNA
PFHxA
PFHxS
PFBA
Water
Median (ppb)
ND
ND
0.25
0.26
2.15
Geometric mean (ppb)
ND
ND
0.10
1.12
1.03
Soil
Median (ppb)
ND
27.2
1,175
5,585
1,600
Geometric mean (ppb)
ND
27.2
1,175
5,585
1,600
Air
Median (ppbv)
ND
ND
ND
ND
ND
Geometric mean (ppbv)
ND
ND
ND
ND
ND
ND = no data.
Source: (ATSDR. 2018).
2
3 Schecter et al. (20121 collected 10 samples of 31 food items from five grocery stores in
4 Texas and analyzed them for persistent organic pollutants, including PFDA, PFNA, PFHxA, and
5 PFHxS. PFDA, PFNA, and PFHxA were not detected in any of the foods targeted, and PFHxS was
6 detected in cod fish at a concentration of 0.07 ng/g wet weight. PFAS have been detected in fish
This document is a draft for review purposes only and does not constitute Agency policy.
9 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
from U.S. lakes and rivers with concentrations ranging from less than the limit of quantification to
15.0 ng/g for PFDA, and <1 to 0.47 ng/g for PFHxS fATSDR. 20181 Stahl etal. (20141 characterized
PFAS in freshwater fish from 164 U.S. urban river sites and 157 Great Lakes sites. PFDA, PFNA,
PFHxA, PFHxS, and PFBA were detected in 92, 69,15, 45, and 16% of the samples, at maximum
concentrations of 13.0, 9.7, 0.8, 3.5, and 1.3 ng/g, respectively. Apart from fish, overall dietary data
for the United States are limited; however, Schaider etal. (20171 detected PFASs in food packaging
collected from U.S. fast food restaurants. Data from other countries (e.g., South Korea, Brazil, Saudi
Arabia] suggest that these PFAS can sometimes be detected in samples of food products including
shellfish, dairy products, meats, vegetables, food packaging materials, and water (both tap and
bottled] (Chen etal.. 2018: Surma etal„ 2017: Heo etal.. 2014: Moreta and Tena. 2014: Perez etal„
20141 The relevance of these detects (and the associated PFAS levels] to U.S. products is unknown.
2.1.5. Potential for Human Exposure
The general population may be exposed to PFAS through multiple routes, including
ingestion of drinking water and food, ingestion of dust, hand-to-mouth and dermal transfer in
products and materials containing these chemicals, and inhalation via indoor and outdoor air
fATSDR. 2018: NLM. 2017. 20131 The oral route of exposure has been considered the most
important route of exposure among the general population for PFAS fKlauniget al.. 20151
The presence of PFAS in human blood provides evidence of exposure among the general
population. PFAS have been monitored in the human population as part of the National Health and
Nutrition Examination Survey (NHANES], PFDA, PFNA, and PFHxS were measured in serum
samples collected in 2013-2014 from more than 2,000 survey participants. The results of these
analyses are presented in Table 4. PFDA and PFNA have also been observed in cord blood and
human milk fATSDR. 20181 Pinnevetal. f20141 and Papadopoulou et al. f20161 observed
associations between breastfeeding and elevated levels of PFHxS in the blood of children.
Table 4. Serum per- and polyfluoroalkyl substances (PFAS) concentrations
based on National Health and Nutrition Examination Survey (NHANES)
2013-2014 data (pg/L)
Population group
Measure
PFDA
PFNA
PFHxA
PFHxS
PFBA
Total population (N = 2,168)
Geometric mean
0.185
0.675
ND
1.35
ND
50th percentile
0.200
0.700
ND
1.40
ND
95th percentile
0.700
2.00
ND
5.60
ND
3 to 5 yr (N = 181)
Geometric mean
_a
0.764
ND
0.715
ND
50th percentile
0.100
0.620
ND
0.740
ND
95th percentile
0.370
3.49
ND
1.62
ND
6 to 11 yr (N = 458)
Geometric mean
_a
0.809
ND
0.913
ND
50th percentile
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
12 to 19 yr (N = 402)
Geometric mean
0.136
0.599
ND
1.27
ND
50th percentile
0.100
0.500
ND
1.10
ND
95th percentile
0.400
2.00
ND
6.30
ND
20 yr and older (N = 1,766)
Geometric mean
0.193
0.685
ND
1.36
ND
50th percentile
0.200
0.700
ND
1.40
ND
95th percentile
0.800
2.00
ND
5.50
ND
LOD = limit of detection; 0.1 (ng/L); ND = no data.
aNot calculated because the proportion of results below the limit of detection was considered too high to provide a
valid result.
Source: CDC (2018). Fourth National Report on Human Exposure to Environmental Chemicals.
2.1.6. Populations and Lifestages with Potentially Greater Exposures
Populations and lifestages that may experience exposures greater than those of the general
population include individuals in occupations that require frequent contact with PFAS-containing
products, such as firefighters or individuals who install and treat carpets fATSDR. 20181 as well as
infants and young children (due to their increased hand-to-mouth behaviors], Rotander et al.
f2 0151 analyzed serum samples from 149 Australian firefighters at an AFFF training facility. Mean
and median PFHxS concentrations were 10 to 15 times higher than those of the general population
of Australia and Canada. Laitinen etal. f 2 0141 evaluated eight firefighters' exposure to PFAS after
three training sessions in Finland in which AFFF had been used. The authors found that the
firefighters' "serum PFHxS and PFNA concentrations seemed to increase during the three training
sessions although they were not the main PFAS in used AFFF." Populations living near
fluorochemical facilities where environmental contamination has occurred may also be more highly
exposed fATSDR. 20181 Also, because PFDA can be found in ski wax, individuals who engage in
professional ski waxing may be more highly exposed because PFAS in dust may become airborne
and inhaled during this process (Harbison etal.. 20151
Populations that rely primarily on seafood for most of their diet, possibly including some
native American tribes (Byrne etal.. 20171 may also be disproportionately exposed. Christensen et
al. (20171 and Haug etal. (20101 used data on serum PFAS levels and 30-day self-reported fish and
shellfish ingestion rates from NHANES 2007-2014 to explore potential relationships between PFAS
exposures and fish consumption. PFDA, PFNA, and PFHxS were among the PFAS detected in the
serum of at least 30% of the NHANES participants, and after adjusting for demographic
characteristics, total fish consumption was associated with elevated serum PFDE and PFNA.
Shellfish consumption was associated with elevations of all the PFASs examined.
2.1.7. Other Environmental Protection Agency (EPA) Assessments of Per- and
Polyfluoroalkyl Substances (PFAS)
EPA released two PFAS assessments for peer review in 2018. Specifically, the draft
assessments of (1] 2,3,3,3-tetrafluoro-2-(l,l,2,2,3,3,3-heptafluoropropoxy]propanoic acid (also
called hexafluoropropylene oxide [HFPO] dimer acid] (CASRN 13252-13-6] and its ammonium salt
This document is a draft for review purposes only and does not constitute Agency policy.
11 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
1 2,3,3,3-tetrafluoro-2-(l,l,2,2,3,3,3-heptafluoropropoxy]propanoate (also called HFPO dimer acid]
2 (CASRN 62037-80-3] referred to as GenX chemicals and (2] perfluorobutane sulfonic acid
3 (CASRN 375-73-5] and its potassium salt potassium perfluorobutane sulfonate
4 (CASRN 29420-49-3] referred to as PFBS. These assessments summarize the available data on the
5 potential human health effects of lifetime exposure to these PFAS and included oral reference doses
6 (RfDs], which estimate (with uncertainty spanning perhaps an order of magnitude] a level of daily
7 oral exposure to the human population (including sensitive subgroups] that is likely to be without
8 an appreciable risk of deleterious noncancer health effects during a lifetime, and qualitative
9 descriptions of the carcinogenic potential of the chemicals. The PFBS assessment updates a
10 Provisional Peer-Reviewed Toxicity Value (PPRTV] assessment that was developed in support of
11 the Superfund Program and published in 2014 fPFBS PPRTV 20141 In addition, EPA released
12 Drinking Water Health Advisories for PFOA and PFOS in 2016, along with health effect support
13 documents (Drinking Water Health Advisories for PFOA and PF0S1 Health advisories are non-
14 enforceable and non-regulatory summaries of technical information on contaminants that can
15 cause human health effects and are known or anticipated to occur in drinking water.
17 was undertaken to prioritize PFAS for review. This effort was coordinated across EPA program and
18 regional offices, where staff discussed specific assessment needs as well as the timeliness of those
19 needs. While additional factors were considered during this scoping effort, Table 5 summarizes the
20 primaiy considerations for selecting the five PFAS described in this protocol, as well as two other
21 PFAS that were recently assessed by EPA: PFBS and GenX chemicals
22 (https: //www.epa.gov/pfas/genx-and-pfbs-draft-toxicitv-assessmentsI In short, these PFAS:
16
23
24
25
26
27
28
• were identified as a priority to inform decision making for EPA's Office of Water (OW],
Office of Land and Emergency Management (OLEM], Office of Chemical Safety and Pollution
Prevention (OCSPP], Office of Children's Health Protection (OCHP], EPA's regional offices,
tribes, or state departments of environmental protection. Most of these PFAS were a
priority for multiple patrons;
29
30
• had been evaluated in in vivo studies of animals and thus might be used to derive toxicity
values; and
31
32
33
• had existing (or under development] standardized analytical methods to monitor
environmental levels to allow for site-specific application of toxicity values to regulatory
decision making.
34
This document is a draft for review purposes only and does not constitute Agency policy.
12 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 5. Environmental Protection Agency (EPA) considerations for the
selection of per- and polyfluoroalkyl substances (PFAS) for evaluation
PFAS
EPA interest
Animal
dose-response
data available3
Analytical detection
methods available13
Standards
Methods
PFDA
• OLEM priority0
Yes
Yes
Yes
PFNA
• OLEM priority0
• OW(UCMR) priority
• Found in industrial effluent and AFFF
Yes
Yes
Yes
PFHxA
• OCSPP priority01
• OLEM priority0
• Region 4 (Coosa and Tennessee
Rivers)
• Found in AFFF
Yes
Yes
Yes
PFHxS
• OCSPP priority
• OLEM priority0
• OW(UCMR) priority
• Region 4 (Tennessee River)
• Found in AFFF
Yes
Yes
Yes
PFBA
• OLEM priority0
• OCSPP priority01
• Found in AFFF
Yes
Yes
Under
development
PFBS
• OLEM priority0
• OCSPP priority01
• OW(UCMR) priority
• Found in AFFF
Yes
Yes
Yes
GenX
chemicals
• OCSPP priority6
• Region 3 priority
• Region 4 priority
Yes
Yes
Yes
GenX= perfluoro(2-methyl-3-oxahexanoic) acid (CASRN 13252-13-6); Unknown = status of validated standards
and methods was unknown at scoping.
aA survey of publicly available literature on PFAS other than PFOA and PFOS (i.e., a broad PubMed search and
review of recent assessments, including ATSDR (2018) was performed to identify in vivo animal studies that
This document is a draft for review purposes only and does not constitute Agency policy.
13 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Animal
dose-response
data available3
Analytical detection
methods available13
PFAS
EPA interest
Standards
Methods
1
2
3
4
5
6
7
Table 6. Potential Environmental Protection Agency (EPA) needs and
applications for five per- and polyfluoroalkyl substances (PFAS)
EPA program
or regional
office
PFAS3
Oral
Inhalation
Dermal
Potential regulatory application and explanation
(at the time scoping was conducted)
OLEM (in
coordination
with EPA
Regions 1-10)
PFDA
PFNA
PFHxA
PFHxS
PFBA
V
V
V
Comprehensive Environmental Response, Compensation,
and Liability Act (CERCLA). CERCLA authorizes EPA to
conduct short- or long-term cleanups at Superfund sites
and later recover cleanup costs from potentially
responsible parties under Section 107. PFAS toxicological
information may be used to make risk determinations for
response actions (e.g., short-term removals, long-term
remedial response actions). An evaluation of potential
actions at Superfund sites considers all routes of exposure.
Resource Conservation and Recovery Act (RCRA). RCRA
can be drawn upon to help address waste management
and cleanup needs, including accidental releases from
potentially hazardous waste management facilities.
OW
PFNA
PFHxS
V
Safe Drinking Water Act (SDWA) and Clean Water Act
(CWA). The SDWA requires EPA to periodically review the
National Primary Drinking Water Regulation (NPDWR) for
This document is a draft for review purposes only and does not constitute Agency policy.
14 DRAFT-DO NOT CITE OR QUOTE
tested multiple PFAS exposure levels and evaluated health endpoints. The quality of the studies was not
evaluated, and while multiple PFAS are evaluated in human studies, this was not a focus of the survey.
bAs of March 2019. The methods noted are for drinking water; non-drinking water methods are being
developed.
cFound at sites, including private and federal facilities and from various sources, including AFFF, chrome-plating
facilities, PFAS manufacturers, and industries that use PFAS (e.g., textiles and electronics). These PFAS have
also been detected in environmental media (e.g., surface water; biota).
dA significant number of new chemicals submitted to EPA are based on C6 and C4 chemistry. OCSPP often
evaluates risk for these compounds based on PFHxA and PFBS, which are the terminal degradation products of
certain C6 and C4 compounds.
Replacement for PFOA (e.g., for emulsifiers) and perfluoroethers. GenX chemicals are of concern based on
occurrence in NC and because EPA has received requests to review similar types of compounds (e.g., longer
chain ethers that might break down to GenX chemicals) as new chemicals.
As described in Section 2.1.5, exposure to these five PFAS can occur via the oral, inhalation,
and dermal routes, with oral (e.g., through diet and drinking water] being the predominant one
(Klaunig et al.. 20151 Given the potential regulatory applications of these PFAS assessments (see
Table 6], these assessments will consider PFAS exposures from all exposure routes. The
assessments will consider all potential health effects of exposure, both cancer and noncancer.
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
EPA program
or regional
office
PFASa
Oral
Inhalation
Dermal
Potential regulatory application and explanation
(at the time scoping was conducted)
each contaminant and revise the regulation, if appropriate.
These potential applications focus on oral exposure.
OCSPP
PFHxA
PFHxS
PFBA
V
V
New chemical submissions to the Office of Pollution
Prevention and Toxics within OCSPP.
Region 4
PFHxA
PFHxS
V
Resource Conservation and Recovery Act (RCRA). RCRA
can be drawn upon to help address waste management
and cleanup needs, including accidental releases from
potentially hazardous waste management facilities. For
PFAS, the primary concern is potential oral exposure from
rivers in Region 4.
OCHP
PFDA
PFNA
PFHxA
PFHxS
PFBA
V
V
V
Executive Order 13045—Protection of Children from
Environmental Health Risks and Safety Risks: Policy on
Evaluating Health Risks to Children. In accordance with
EPA's 1995 policy and EO 13045, EPA instituted and
reaffirmed an Agency-wide commitment to "consider the
risks to infants and children consistently and explicitly as
part of risk assessments generated during its
decision-making process."
aPFAS to which this protocol applies (i.e., excluding PFBS and GenX chemicals).
1
2.3. PROBLEM FORMULATION
2.3.1. Assessments and Toxicity Values from Other Sources
2 For the five PFAS addressed in this protocol, a summaiy of existing values from national,
3 international, and state agencies, as well as U.S. state action levels (current as of March 2019], is
4 provided in Figure 2. It is important to note that these values are not all directly comparable, as
5 some values take into account the potential for human exposure or other considerations. The
6 majority of current values are noncancer toxicity values based on oral exposure studies in rodents,
7 although a few inhalation toxicity values exist (see Table A-l in Appendix A for more details],
8
This document is a draft for review purposes only and does not constitute Agency policy.
15 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
(a)
PFBA Oral Reference Values
TO
"9
u>
—
E
o
o
<
00
0.1 :
0.01
0.001
Acute
$ Si
~ TCEQ RfD
8 2
~ HC Screening Value
|
~ GFS
5
c
~ MDH HRL
w
u
§
S TCEQ Tier 1 PCL
U
Intermediate I Longer-Term
Chronic / Lifetime
TCEQTIerlPCL
HC Screening Value [
EH GFS
11 MDH HRL
~
TCEQRfD
100
- 10
I— I— I— 1
10 100 1,000 10,000 100,000
Duration (Days)
.n
a.
a.
1
0)
re
g
c
c
o
c
0)
u
c
o
u
<
CO
(b)
0.001
o.ooooi
o.oooooi
PFHxA Oral Reference Values
Acute
o.oooi :
¦ Danish EPATDI*
0 TCEQ RfD
E Danish EPA QC*
EGFS
¦ HC Screening Value £
c
Q Sweden MTL*
s
~ UBAHRIV
¦ TCEQ Tier 1 PCL
H~
intermediate / Longer-Term
Chronic I Lifetime
GFS EH
UBA HRIV
0
Danish EPATDI* ¦
HC Screening Value ~
| Danish EPA QC
Sweden MTL* 0 ® SI
10
100
Duration (Days)
—H
1,000
TCEQ Tier 1 PCL
~
TCEQ RfD
1
10,000
100,000
*Applies to the sum of multiple PFAS, including PFHxA
This document is a draft for review purposes only and does not constitute Agency policy.
16 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
CO
0.0001
>-
ra
¦D
oo
E
o
~
t/t
X
0.00001 --
0.000001
PFHxS Oral Reference Values
Acute
¦ Danish EPA TDI*
Ol
El NH DES RfD
3
~ TCEQ RfD
s
~ Australia TDI*
Q
a Danish EPA QC*
~ GFS
¦ HC Screening Value
~ AK/CT AL/Mass. ORSG*
~ NH DES DWEL
_o
~ NH DES MCL
S TCEQ Tier 1 PCL
~ VT DEC DWHA*
s
~ UBA HRIV
O
~ Australia Value*
~ MDH HBV
~ Sweden MTL*
Intermediate I Longer-Term
Chronic I Lifetime
0
HC Screening Value
Danish EPATDI* ¦ E3 UBA HRIV
I
Australia TDI* ~
Ml DES DWEL ¦
Danish EPA QC*
TCEQTier 1 PCI. GFS
Sweden MTL* DES RfD
AK/CTAL/Mass. ORSG* ~ DES MCL
Australia Value*
¦II MDHHBVH |
C3- TCEQ RfD
E3- VT DEC DWHA* -
ja
Q.
a.
=L
5
C
f o.i •-
O
c
-
(11
¦o
— 0.0001 r
o
Q
<
s
0.00001
0.000001
PFNA Oral Reference Values
Acute
¦ Danish EPATDI*
¦ NH DES RfD
H TCEQ R»
S3 Danish EPA QC*
~ GFS
¦ HC Screening Value
¦ AK/CT AL/Mass. ORSG*
~ NH DES DWEL
~ NH DES Ma
0 NJ DEP MCL
@ TCEQ Tier 1PCL
~ VT DEC DWHA*
-H-
Intermediate / Longer-Term
Chronic / Lifetime
TCEQTier 1 PCL |
Danish EPA QC*
a
AK/CTAL/Mass. O'RSG*
~ I
Danish EPATDI* ¦ Q SFf
0 NH DES DWEL
TCEQ RfD ~
Q- NH DES MCL
HC Screening Value -0 ~ J
NH DES RfD IB 71 D^C DWHA*
NJ DEP MCL B
¦H-
-H-
0.1
XI
CL
a.
a
s
03
5
.5
c
o
"¦M
c
QJ
u
c
o
u
<
0.01
10 100 1,000
Duration (Days)
10,000
100,000
* Applies to the sum of multiple PFAS, including PFNA
This document is a draft for review purposes only and does not constitute Agency policy.
17 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
(e)
to
"D
2
0.01
0.001
— 0.0001 :
o
a
0.00001 :
o.oooooi
Acute
~ TCEQ RfD
~ UBAHRIV
» TCEQ Tier 1 PCL =
5
PFDA Oral Reference Values
Intermediate I Longer-Term
10 100 1,000
Duration (Days)
Chronic I Lifetime
TCEQTier 1 PCL
UBA HRIV
~
TCEQ RfD
~
—H
10,000
A
Q.
Q.
2
0)
ro
5
o.i .£
c
0
ra
L.
•M
c
01
u
c
o
u
<
o
o.oi
100,000
Figure 2. Existing Oral Reference Values for (a) perfluorobutanoic acid
(PFBA), (b) perfluorohexanoic acid (PFHxA), (c) perfluorohexanesulfonate
(PFHxS), (d) perfluorononanoic acid (PFNA), and (e) perfluorodecanoic acid
(PFDA).
2.3.2. Preliminary Literature Inventory for the Five Per- and Polyfluoroalkyl Substances
(PFAS) Being Assessed
As described in Section 2.1.1, several of these five PFAS have associated salts of potential
interest for human health assessment. Thus, the assessments will address each PFAS as follows:
• PFBA: PFBA (CASRN 375-22-4]; PFBA ammonium salt (CASRN 10495-86-0]
• PFHxA: PFHxA (CASRN 307-24-4]; PFHxA ammonium salt (CASRN 21615-47-4]; PFHxA
sodium salt (CASRN 2923-26-4]
• PFHxS: PFHxS (CASRN 355-46-4]; PFHxS potassium salt (CASRN 3871-99-6]
• PFNA: PFNA (CASRN 375-95-1]; PFNA ammonium salt (CASRN 4149-60-4]; PFNA sodium
salt (CASRN 21049-39-8]
• PFDA: PFDA (CASRN 335-76-2]; PFDA ammonia salt (CASRN 3108-42-7]; PFDA sodium salt
(CASRN 3830-45-3]
This document is a draft for review purposes only and does not constitute Agency policy.
18 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
1
2 The results of a preliminary literature inventory of health effect-related studies on these
3 five PFAS and their associated salts are presented in Figure 3. The studies summarized in this
4 preliminary literature inventory are described on the project pages for these assessments in HERO
5 fhttps: //hero.epa.gov: see Section 1 for links to the specific Health and Environmental Research
6 Online [HERO] pages].
This document is a draft for review purposes only and does not constitute Agency policy.
19 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
-h— ( 10 ¦ studies) | ++ ( ? studies)
^ ( 1 2 studies) | - (Not Studied)"
PFDAand salts
PFNA and salts
PFHxA and salts
PFHxS and salts
PFBA and salts
Oral:
Long1
Oral:
Short1
Oral:
Long1
Oral:
Short1
Oral:
Long1
Oral:
Short1
Oral:
Long1
Oral:
Short1
Oral:
Long1
Oral:
Short1
Cardiovascular
Developmental
Endocrine
(Thyroid)
Gastro-
intestinal
Hematologic
Hepatic
Musculo-
skeletal
Reproductive
Respiratory
General
Toxicity/ Other
Figure 3. Results of a preliminary literature inventory of five per- and polyfluoroalkyl substances (PFAS), Data are
approximated based on a cursory review of the literature search results for studies published through 2018 (see Section 4
for details]. Health effects are based on groupings from EPA Integrated Risk Information System (IRIS) website
fiitips://cfpub.epa.gov/ncea/iris/search/iridex.cfml.'1 For this summary, metabolic effects are captured under "other" and
"hepatic" includes lipid and lipoprotein measures.
3"Oral: iong" indicates subchronic or chronic oral exposure duration studies in animals and "Oral: short" reflects short-term and acute oral exposure studies in
animals, as well as reproductive and developmental studies.
This document is a draft for review purposes only and does not constitute Agency policy,
20 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Based on the results from the preliminary literature inventory in Figure 3, the following
health effects appear to be well studied for most PFAS of interest:
• Developmental effects
• Endocrine (primarily thyroid hormone] effects
• Hepatic effects, including lipid and lipoprotein measures
• Immune effects
• Reproductive effects in males or females
• Urinary effects
• General toxicity
As also shown in Figure 3, no studies of dermal exposure were identified. In addition, data
are sparse for assessing the potential health effects from chronic or subchronic oral exposure or for
inhalation exposure of any duration. Few studies have examined whether exposure to these PFAS
may result in carcinogenicity.
Given the potential future utility of comparing evidence across PFAS assessments (including
their respective data gaps], the five PFAS assessments will specifically address each of the potential
health effects enumerated above as "well studied." In addition, the potential for carcinogenicity will
be explicitly addressed in each assessment. Data on several other, variably studied endpoints
(i.e., cardiovascular effects, hematological effects, metabolic effects including diabetes, and nervous
system effects] will also be summarized when available. These summaries may be developed in
association with one of the health effects noted above, as a separate formal evaluation of hazard, or
as part of a qualitative summary on "other effects," depending on the assessment-specific data.
Given the paucity of available studies and the absence of exceptional evidence in the available
studies, information on other health effects (i.e., gastrointestinal effects; musculoskeletal effects;
ocular effects; and respiratory effects] may be briefly summarized but will not be formally
evaluated in any of these assessments. New literature relating to these outcomes will be monitored
during literature search updates for potential inclusion.
2.4. KEY SCIENCE ISSUES
This section describes critical areas of scientific complexity that were identified based on
the preliminary literature inventory results summarized in the previous section. These scientific
issues are essential to consider during development of these assessments, and the specific methods
for doing so within these PFAS assessments are described in subsequent sections.
This document is a draft for review purposes only and does not constitute Agency policy.
21 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
2.4.1. Toxicokinetic Differences across Species and Sexes
In humans, PFAS generally tend to remain unchanged in the body for long durations (in
general, while PFAS bind to macromolecules such as albumin and lipoproteins, they do not undergo
internal chemical reactions, and many are not metabolized]. They are typically not stored in body
fat (see Section 2.1 for PFAS-specific chemical properties, including predicted LogP], but
accumulate in locations such as the blood, liver, and kidneys (and can be transferred to offspring
through placental transfer and breast milk] (ATSDR. 2018: U.S. EPA. 2016c. d; Post etal.. 20121
However, as illustrated in Table 7, previous summaries of the existing literature suggest there are
pronounced half-life differences across species, sex, and type of PFAS. In general, PFAS with longer
chain lengths are reported to have a longer serum half-life. For the PFAS with data available, serum
half-life variation across species exhibits the following pattern: rats
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 7. Preliminary serum half-life estimates of five per- and polyfluoroalkyl
substances (PFAS) across species and sexes
PFBA (C4)
PFHxA (C6)
PFHxS(C6)
PFNA (C9)
PFDA (C10)
Female
Male
Female
Male
Female
Male
Female
Male
Female
Male
Rat
1.0-1.8 h
6-9 h
0.4-0.6 h
1.0-1.6 h
1.8 d
6.8 d
1.4 d
30.6 d
58.6 d
39.9 d
Mouse
3 h
12 h
~1.2 h
~1.6 h
24-27 d
28-30 d
26-68d
34-69 d
ND
Monkey
1.7 d
2.4 h
5.3 h
87 d
141 d
ND
ND
Human
3d
32 d
8.5 yr
4.3 yr
12 yr
"C" = carbon chain length; ND = no data.
Data are summarized in Lau (2015). Note that these values do not necessarily represent those that would be
used in qualitative or quantitative analyses for these PFAS assessments because the underlying data will be
reviewed and possibly supplemented with additional (e.g., newer) studies. Darker shading indicates longer
half-life (i.e., from hours to days to years).
2.4.2. Human Relevance of Effects in Animals that Involve Peroxisome
Proliferator-Activated Receptor Alpha (PPARa) Receptors
Activation of the peroxisome proliferator-activated receptor alpha (PPARa] by PFAS has
been reported, with in vitro evidence that the potency of human and mouse PPARa activation is
positively correlated with increasing PFCA chain length up to C9 (no human receptor activation was
noted for PFDA, although activation of the mouse receptor was only slightly less potent than PFNA]
and greater for carboxylates than sulfonates fWolfetal.. 2014: Wolfetal.. 2008: Takacs and Abbott.
2007: Shipley et al.. 2004: Malonev and Waxman. 19991 It is not known whether PFAS distribute to
the nucleus and bind directly to PPARa in vivo, or whether these substances activate the receptor
indirectly. PPARa ligand binding causes a conformational change in the protein, release of
corepressors, heterodimerization with the retinoid X receptor (RXR], and binding to cognate
peroxisome proliferator response elements in the promoters of target genes (perhaps most notably,
those related to fatty acid (3-oxidation and energy homeostasis] to modulate gene transcription.
PPARa is a ligand-activated nuclear receptor expressed in many tissues and has been at the
forefront of a longstanding debate as to whether chemical-induced PPARa modulation in rodents,
particularly in the liver, is relevant to humans (Corton etal.. 2018: Filgo etal.. 2015: Guvton et al..
20091 PPARa is active in humans and responsive to the hypolipidemic effect of fibrate drugs that
lower serum lipid levels, but the human receptor is generally considered less sensitive than PPARa
in rodents (Corton et al.. 2014: Wolfetal.. 2014: Wolfetal.. 2008: Malonev and Waxman. 19991 In
particular, research with fibrate drugs indicates that PPARa activation in rodents leads to hepatic
effects such as hepatomegaly, peroxisome proliferation, and in some instances over the long term,
cancer. These effects are not obseived in human models fCorton etal.. 20141 Evaluating the
human relevance of animal PPARa evidence is complicated by a lack of comparable model systems,
This document is a draft for review purposes only and does not constitute Agency policy.
23 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
including widely used primary cell lines that rapidly lose the capability to express nuclear receptors
such as PPARa (Soldatow et al.. 20131 and potential species-specific differences in transcriptional
coactivators and other pathway components.
With PFAS specifically, PPARa-dependent effects have been best studied in relation to
hepatic effects (and liver cancer, in particular]. However, PPARa is also known to be important to
other physiological processes in both rodents and humans, including energy homeostasis,
inflammation, reproduction, musculoskeletal function, and development (Corton etal.. 2014: Burri
etal.. 2010: Abbott. 2009: Peraza etal.. 2006: Corton etal.. 20001 Thus, although not well studied
for PFAS, the modulation of PPARa may be important to consider for developmental, metabolic,
reproductive, and immunological effects, as well as for hepatic effects.
There are additional complexities to considering the dependence on, and human relevance
of, PPARa activation by PFAS for certain health effects. The extent of PPARa activation is likely to
differ by PFAS type, making it harder to apply read-across (specifically, drawing conclusions for one
PFAS based on findings for another PFAS] or related approaches. In addition, based on conclusions
from other PFAS assessments and review articles, there is evidence to indicate that many
PFAS-mediated effects appear to include both PPARa-dependent and PPARa-independent
mechanisms, the latter of which include activation of PPARy,
phosphatidylinositol-3-kinase-serine/threonine kinase Akt (PI3K-Akt], constitutive androstane
receptor (CAR], nuclear factor kappa B pathway (NF-kB], farnesoid X receptor, liver X receptor, and
estrogen receptor a (Li et al.. 2017: Rosen etal.. 2017: FSANZ. 2016: U.S. EPA. 2014b. d; Foreman et
al.. 20091
Despite the complexities involved, it is important to evaluate the human relevance of some
PFAS exposure-mediated effects in animals (see discussion in Section 9.2],
2.4.3. Potential Confounding by Other Per- and Polyfluoroalkyl Substances (PFAS)
Exposures in Epidemiology Studies
Because different PFAS may be used in similar applications or result from similar sources,
potential confounding of associations by PFAS coexposures is an important area of uncertainty for
epidemiology studies. When associations are found for two or more moderately correlated PFAS in
a study, including those not the focus of these assessments (e.g., PFOS and PFOA], confounding is a
possible explanation. Based on a cursory review of studies identified during the preliminary
literature inventory, a complicating factor is that correlations between PFAS pairs vary
considerably across studies (see Section 6.2.1], When a study does not report the correlations in its
population, the interpretation of the risk of bias from confounding is particularly challenging. Even
when correlations are reported, there is no perfect method for eliminating confounding. Given this
variability, assessing the likelihood and impact of this source of potential confounding based on
reporting within individual studies is expected to be difficult (see discussion in Section 6.2],
This document is a draft for review purposes only and does not constitute Agency policy.
24 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
2.4.4. Toxicological Relevance of Changes in Certain Urinary and Hepatic Endpoints in
Rodents
The scientific community has identified difficulties in interpreting the toxicological
relevance of changes in certain urinary and hepatic endpoints available in rodent studies (based on
the preliminary literature inventory] for some of the five PFAS assessments. The specific rodent
endpoints in question are chronic progressive nephropathy and related urinary histopathological
changes (including alpha 2u-globulin-mediated changes], and hepatic effects that may be
considered adaptive (e.g., increased liver weight; cellular hypertrophy; single cell
necrosis/apoptosis]. For the former, some of these changes are not considered relevant to humans,
and methods exist for evaluating the dependency of observed changes on this rodent-specific
mechanism. For the latter, neither a clear scientific consensus nor specific EPA-wide guidance
defines exactly what level of change or constellation of effects is necessary to establish cause for
concern. Thus, interpretations of the toxicological relevance of changes in these specific endpoints
are expected to require additional consideration (see discussion in Section 9.2],
2.4.5. Characterizing Uncertainty Due to Missing Chemical-Specific Information
Two PFAS, PFOA and PFOS (C8], have been studied more extensively than other PFAS.
Thus, this existing knowledge base may be useful in helping to characterize existing data gaps and
uncertainties in the current five PFAS assessments. Two recently developed EPA assessments
(PFBS and GenX chemicals] could also provide information during the development of these
current assessments. For example, given knowledge regarding the health effects of PFOA and PFOS,
the potential lack of studies on immune effects for PFBA and developmental effects for PFHxA
(based on the preliminary literature inventory; see Section 2.3.2] appear to represent important
database uncertainties. In addition, given the potential for lifetime human exposure to PFAS by
multiple routes of exposure (see Section 2.1.5], the apparent scarcity of data on most of these five
PFAS other than short-term oral exposure studies in animals is expected to affect assessment
decisions and characterization of uncertainties (see discussion in Sections 10.2 and 11.2.3],
This document is a draft for review purposes only and does not constitute Agency policy.
25 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
3. OVERALL OBJECTIVES, SPECIFIC AIMS, AND
POPULATIONS, EXPOSURES, COMPARATORS,
AND OUTCOMES (PECO) CRITERIA
The overall objective of these five assessments is to identify adverse human health effects
and characterize exposure-response relationships for the effects of perfluorobutanoic acid (PFBA],
perfluorohexanoic acid (PFHxA], perfluorohexanesulfonate (PFHxS], perfluorononanoic acid
(PFNA], and perfluorodecanoic acid (PFDA] to support development of toxicity values. These
assessments will use systematic review methods to evaluate the epidemiological and toxicological
literature, including consideration of relevant mechanistic evidence (e.g., to inform key science
issues; see Section 2.4], The evaluations conducted in these assessments will be consistent with
relevant Environmental Protection Agency (EPA] guidance.2
The specific approach taken for these assessments of the potential health effects of PFBA,
PFHxA, PFHxS, PFNA, and PFDA (and their associated salts] was based on input received during
scoping as well as a preliminary literature inventory of the health effects studied for these PFAS.
As outlined in Section 2.3.2, these assessments will evaluate the potential for PFAS exposure via the
oral or inhalation route to cause health effects in humans, specifically focusing on developmental
effects; endocrine (primarily thyroid hormone] effects; hepatic effects, including lipid and
lipoprotein measures; immune effects; reproductive effects in males or females; urinary effects;
general toxicity; and carcinogenicity (see Section 5 for preliminary decisions for grouping outcomes
and endpoints within each of these predetermined health effect categories]. Data on cardiovascular
effects, hematological effects, metabolic effects including diabetes, and nervous system effects will
also be summarized when available. These summaries may be developed in association with one of
the health effects noted above either as a separate formal evaluation of hazard or as part of a
qualitative summaiy on "other effects," depending on the assessment-specific data. Given the
paucity of available studies and in the absence of exceptional evidence in any available studies,
information on other health effects (i.e., gastrointestinal effects; musculoskeletal effects; ocular
effects; and respiratory effects] will not be formally evaluated (these effects may be briefly
summarized] in any of these assessments; although new literature relating to these outcomes will
be monitored during literature search updates for potential inclusion. As outlined in the EPA PFAS
action plan,3 the characterization of the potential human health hazards from exposure to these
2EPAguidance documents: http://www.epa.gov/iris/basic-information-about-integrated-risk-information-
svstem#guidance/.
3EPA PFAS action plan: https://www.epa.gov/pfas/epas-pfas-action-plan.
This document is a draft for review purposes only and does not constitute Agency policy.
26 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
individual PFAS will be coupled with data generated from new advances in computational and
high-throughput toxicology to inform evaluations of other PFAS.
3.1. SPECIFIC AIMS
The aims of these assessments are to:
• Identify epidemiological (i.e., human] and toxicological (i.e., experimental animal] literature
reporting effects of exposure to PFBA, PFHxA, PFHxS, PFNA, and PFDA (and their associated
salts], as outlined in the PECO. These five systematic reviews will focus on identifying
studies following oral or inhalation exposure to PFAS.
• Evaluate mechanistic information (including toxicokinetic understanding] associated with
exposure to PFBA, PFHxA, PFHxS, PFNA, and PFDA, to inform the interpretation of findings
related to potential health effects in studies of humans and animals. The scope of these
analyses of mechanistic information will be determined by the complexity and confidence in
the phenotypic evidence in humans and animals, the likelihood of the analyses (e.g.,
considering the mechanistic studies available based on the literature inventory; see Section
4.2.2] to affect evidence synthesis conclusions for human health, and the directness or
relevance of the available model systems for understanding potential human health hazards
(see Section 9.2], The mechanistic evaluations will focus primarily on the key science issues
identified in Section 2.4.
• Conduct study evaluations for individual epidemiological and toxicological studies
(evaluating reporting quality, risk of bias, and sensitivity] and PBPK] (scientific and
technical review]. The evaluation of epidemiology studies will specifically consider, to the
extent possible, the likelihood and impact of potential confounding by other PFAS (see
Section 6.2.1],
• Extract data on relevant health outcomes from epidemiological and toxicological studies of
high, medium, and low confidence based on the study evaluations (full data extraction of
low confidence studies may not be performed for poorly studied health effects or for health
effects on which extensive medium and high confidence studies exist in the evidence base],
• Synthesize the evidence across studies, assessing similar health outcomes using a narrative
approach. To inform future comparisons across a range of PFAS structures and properties
(e.g., using high throughput screening, computational toxicology approaches, and chemical
informatics to fill in data gaps; see EPA PFAS action planl each of the five PFAS assessments
will synthesize the available evidence (or lack thereof] for developmental effects; endocrine
(primarily thyroid hormone] effects; hepatic effects, including lipid and lipoprotein
measures (the latter of which are also applicable to interpreting the potential for
cardiovascular toxicity]; immune effects; reproductive effects in males or females; urinary
effects; general toxicity; and carcinogenicity. Some assessments may include additional
evidence syntheses for other health effects. The toxicological relevance of changes in some
urinary and hepatic outcomes will be a point of focus in the evidence syntheses (see
Section 9.2.3],
This document is a draft for review purposes only and does not constitute Agency policy.
27 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
• For each health outcome (or grouping of outcomes], evaluate the strength-of-evidence
across studies (or subsets of studies] separately for studies of exposed humans and for
animal studies. Based on the focused mechanistic analyses specific to each PFAS
assessment (see Section 9.2], the mechanistic evidence will be used to inform evaluations of
the available health effects evidence (or lack thereof).
• For each health outcome (or grouping of outcomes], develop an integrated expert judgment
across evidence streams as to whether the evidence is sufficient (or insufficient] to indicate
that exposure to the PFAS has the potential to be hazardous to humans (in rare instances,
the evidence may be judged as sufficient to indicate that a hazard is unlikely]. The judgment
will be directly informed by the evidence syntheses and based on structured review of an
adapted set of considerations for causality first introduced by Austin Bradford Hill (Hill.
1965](see Sections 9 and 10], including consideration (e.g., based on available mechanistic
information] and discussion of biological understanding. As part of the evidence integration
narrative, characterize the strength of evidence for the available database of studies and its
uncertainties, and identify and discuss issues concerning potentially susceptible
populations and lifestages.
• Derive toxicity values (e.g., oral reference doses [RfDs], inhalation reference concentrations
[RfCs], cancer risk estimates] as supported by the available data (see Section 10.2], Apply
toxicokinetic and dosimetry modeling (possibly including PBPK modeling] to account for
interspecies differences, as appropriate. Given the apparent species and sex differences in
the toxicokinetic profile of the different PFAS (see Section 2.4], methods to address these
potential differences will be a key consideration (see Section 9.2.1],
• Characterize uncertainties and identify key data gaps and research needs across each PFAS
database, such as limitations of the available evidence, limitations of the systematic review,
and consideration of dose relevance and toxicokinetic differences when extrapolating
findings from higher dose animal studies to lower levels of human exposure.
3.2. POPULATIONS, EXPOSURES, COMPARATORS, AND OUTCOMES
(PECO) CRITERIA
The PECO criteria are used to identify the evidence that addresses the specific aims of the
assessment and to focus the literature screening, including study inclusion/exclusion, in a
systematic review (see details on literature screening in Section 4.2], Given the anticipated lack of
studies on carcinogenicity for these PFAS based on the preliminary literature inventory,
genotoxicity studies were included in the PECO criteria (see Table 8],
In addition to those studies meeting the PECO criteria, studies containing supplemental
material that are potentially relevant to the specific aims of the assessment were tracked during the
literature screening process. Although these studies did not meet PECO criteria, they were not
excluded from further consideration. The categories used to track studies as "potentially relevant
supplemental material" are also described in Section 4.2.
This document is a draft for review purposes only and does not constitute Agency policy.
28 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 8. Populations, exposures, comparators, and outcomes (PECO) criteria
PECO
element
Evidence
Populations
Human: Any population and lifestage (occupational or general population, including children and
other sensitive populations). The following study designs will be included: controlled exposure,
cohort, case-control, and cross-sectional. (Note: Case reports and case series will be tracked as
potential supplemental material.)
Animal: Nonhuman mammalian animal species (whole organism) of any lifestage (including
preconception, in utero, lactation, peripubertal, and adult stages).
Other: In vitro, in silico, or nonmammalian models of genotoxicity. (Note: Other in vitro, in silico,
or nonmammalian models will be tracked as potential supplemental material.)
Exposures
Human: Studies providing quantitative estimates of PFAS exposure based on administered dose
or concentration, biomonitoring data (e.g., urine, blood, or other specimens), environmental or
occupational-setting measures (e.g., water levels or air concentrations, residential location
and/or duration, job title, or work title). (Note: Studies that provide qualitative, but not
quantitative, estimates of exposure will be tracked as supplemental material.)
Animal: Oral or Inhalation studies including quantified exposure to a PFAS of interest based on
administered dose, dietary level, or concentration. (Note: Non-oral and non-inhalation studies
will be tracked as potential supplemental material.) PFAS mixture studies are included if they
employ an experimental arm that involves exposure to a single PFAS of interest. (Note: Other
PFAS mixture studies are tracked as potential supplemental material.)
Studies must address exposure to one or more of the following: PFDA (CASRN 335-76-2), PFDA
ammonia salt (CASRN 3108-42-7), PFDA sodium salt (CASRN 3830-45-3), PFNA (CASRN 375-95-1),
PFNA ammonium salt (CASRN 4149-60-4), PFNA sodium salt (CASRN 21049-39-8), PFHxA
(CASRN 307-24-4), PFHxA sodium salt (CASRN 2923-26-4), PFHxA ammonium salt (CASRN 21615-
47-4), PFHxS (CASRN 355-46-4), PFHxS potassium salt (CASRN 3871-99-6), PFBA
(CASRN 375-22-4), or PFBA ammonium salt (CASRN 10495-86-0). [Note: although most PFAS
tend to remain unchanged in the bodv (Nabb et al., 2007), it is possible that some PFAS mav be
bio-transformed to a PFAS of interest. Thus, studies of precursor PFAS that identify and quantify
a PFAS of interest (e.g., as a metabolite) will be tracked as potential supplemental material
(e.g., for ADME interpretations).]
Comparators
Human: A comparison or reference population exposed to lower levels (or no
exposure/exposure below detection levels) or for shorter periods of time.
Animal: Includes comparisons to historical controls or a concurrent control group that is
unexposed, exposed to vehicle-only or air-only exposures. (Note: Experiments including
exposure to PFAS across different durations or exposure levels without including one of these
control groups will be tracked as potential supplemental material [e.g., for evaluating key
science issues; Section 2.4].)
Outcomes
All cancer and noncancer health outcomes. (Note: Other than genotoxicity studies, studies
including only molecular endpoints [e.g., gene or protein changes; receptor binding or
activation] or other non-phenotypic endpoints addressing the potential biological or chemical
progression of events contributing towards toxic effects will be tracked as potential
supplemental material [e.g., for evaluating key science issues; Section 2.4].)
This document is a draft for review purposes only and does not constitute Agency policy.
29 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
PECO
element
Evidence
PBPK models
Studies describing physiologically based pharmacokinetic (PBPK) and other PK models for PFDA
(CASRN 335-76-2), PFDA ammonia salt (CASRN 3108-42-7), PFDA sodium salt (CASRN 3830-45-3),
PFNA (CASRN 375-95-1), PFNA ammonium salt (CASRN 4149-60-4), PFNA sodium salt
(CASRN 21049-39-8), PFHxA (CASRN 307-24-4), PFHxS (CASRN 355-46-4), PFHxS potassium salt
(CASRN 3871-99-6), PFBA (CASRN 375-22-4), or PFBA ammonium salt (CASRN 10495-86-0).
ADME = absorption, distribution, metabolism, and excretion; PK = pharmacokinetic.
This document is a draft for review purposes only and does not constitute Agency policy.
30 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
4. LITERATURE SEARCH AND SCREENING
STRATEGIES
The initial literature search was completed in July 2017 as part of a cross-Environmental
Protection Agency (EPA] workgroup that focused on a large set PFAS, including the five PFAS
addressed in this protocol. Subsequent literature searches were refined and are being updated
regularly. In an effort to ensure that all pertinent studies were captured, the studies identified as
relevant to PFBA, PFHxA, PFHxS, PFNA, and PFDA are being shared simultaneously with release of
this protocol.4 These search efforts reflect studies published through February 2018, several
unpublished reports, and a few recent studies not yet identified through the formal literature
updating process; the literature is currently being updated and will be updated regularly until
several months before public release of the draft assessments.5 Accordingly, the methods for
literature search and screening (as well as some of the approaches to refining the evaluation plan
based on the identified literature; see Section 5] are described in the protocol using the past tense,
whereas approaches for the other assessment methods are outlined using the future tense.
4.1. LITERATURE SEARCH STRATEGIES
The initial literature search strategy performed in July 2017 was designed to identify a
broad range of topics relevant to PFAS, including studies on physicochemical properties,
environmental fate and occurrences, human exposures, and biological effects representative of all
types of evidence (i.e., human, animal, in vitro, in silico) and health outcomes. PFAS search terms
included PFAS names (including salt, cationic, and anionic forms], all known synonyms, and CAS
registry numbers. The literature search itself encompassed a non-date-limited queiy of the
following databases:
• PubMed f National Library of Medicine!
• Web of Science (Thomson Reuters!
4PFBA: https://hero.epa.gov/hero/index.cfm/proiect/page/proiect id/2632
PFHxA: https://hero.epa.gov/hero/index.cfm/proiect/page/proiect id/2628
PFHxS: https://hero.epa.gov/hero/index.cfm/proiect/page/proiect id/2630
PFNA: https://hero.epa.gov/hero/index.cfm/proiect/page/proiect id/2633
PFDA: https://hero.epa.gov/hero/index.cfm/proiect/page/proiect id/2614
5Although not identified (yet) as part of the formal literature searches, several recent PBPK studies found through
regular monitoring of new studies are included in this protocol (see Section 6.4) so that the process for evaluating
those data can be outlined.
This document is a draft for review purposes only and does not constitute Agency policy.
31 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
• Toxline fNational Library of Medicine!
• TSCATS (Toxic Substances Control Act Test Submissions!
All literature identified in the initial search was loaded into the U.S. EPA Health and
Environmental Research Online (HERO] database. In February 2018, the literature search was
updated for the PFAS in this assessment (i.e., PFBA, PFHxA, PFHxS, PFNA, and PFDA], The updated
literature query included all PFAS nomenclature from the initial search as well as a broader
non-date-limited search of several new PFAS synonyms that had been identified since the original
search. This updated search was conducted by EPA's HERO tool to search the same databases as
were included in the initial literature query.
Because each database has its own search architecture, the resulting search strategy was
tailored to account for each database's unique search functionality. Full details of the July 2017 and
February 2018 search strategies are presented in Appendix B. No literature was restricted by
language.
Additional relevant literature not found through database searching was identified by:
• Review of studies cited in state, national (EPA, Food and Drug Administration [FDA], etc.],
and international (International Agency for Research on Cancer [IARC], World Health
Organization [WHO], European Chemicals Agency [ECHA], etc.] assessments on these five
PFAS, including parallel assessment efforts in progress (e.g., the draft Agency for Toxic
Substances and Disease Registry [ATSDR] assessment released publicly in 2018],
• Review of studies submitted to federal regulatory agencies and brought to the attention of
EPA. For example, studies submitted to EPA by the manufacturers of these five PFAS in
support of requirements under the Toxic Substances Control Act (TSCA], (Note: such
studies [or data summaries] will only be tracked in the literature flow diagrams released
with each of the five assessments when they are going to be made publicly available.]
• Identification of studies during screening for other PFAS. For example, epidemiology
studies relevant to more than one of these five PFAS were sometimes identified by searches
focused on one PFAS, but not the others.
• Other gray literature (i.e., primary studies not indexed in typical databases, such as
technical reports from government agencies or scientific research groups; unpublished
laboratory studies conducted by industry; or working reports/white papers from research
groups or committees] brought to the attention of EPA during problem formulation,
engagement with technical PFAS experts, and during future solicitation of Agency,
interagency, and public comment during the Integrated Risk Information System (IRIS]
assessment development and review process. For example, one such study was brought to
the attention of the EPA on March 29, 2018 by the National Toxicology Program (NTP]
when NTP published study tables and individual animal data from a 28-day toxicity study of
multiple PFAS (NTP. 20111 A peer-reviewed NTP Technical Report was not yet available at
the time this protocol was drafted, but these data have undergone standard NTP quality
This document is a draft for review purposes only and does not constitute Agency policy.
32 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
assurance/control processing and are publicly available, and a protocol outlining the NTP
study methods is available in HERO
fhttps://hero.epa.gov/hero/index.cfm/reference/details/reference id/43097411.
The number of studies on PFBA, PFHxA, PFHxS, PFNA, and PFDA returned from the
literature searches through February 2018 is documented in the literature flow diagrams in
Figure 4, which also reflect the literature screening decisions (see Section 4.2], Notably, the
identification and review of records submitted to the EPA, which may include confidential business
information (CBI], is ongoing. This includes exploring the possibility of making the data within any
identified records publicly available. Any records included in the assessments will be reflected in
updates to the protocol and in the draft assessments. In addition, any identified companion
documents for the included studies, such as retractions, corrections and supplemental materials,
will also be included, and the assessments will incorporate the most recent publication materials
(note: these are tracked as separate, "included" records in the literature flow diagrams [see
Section 4.2.2] and HERO; companion documents in other screening categories such as "excluded,"
which are not relevant to the target PECO, are similarly tagged as separate records within that
screening category].
The literature searches will be updated throughout the assessments' development and
review process to identify newly published literature. The last full literature search update will be
conducted prior to (several months] the planned release of the draft document for public comment.
The literature flow diagrams (see Section 4.2.2] will be updated with the results of these updates.
Although uncommon, it is possible that during assessment development and review additional
literature searches may be performed (e.g., to supplement an analysis of a specific mechanism or
biological linkage]. Any such ancillary searches will be documented in updates to the protocol.
The IRIS Program takes extra steps to ensure identification of pertinent studies by
encouraging the scientific community and the public to identify additional studies and ongoing
research; by searching for publicly available data submitted under the TSCA and the Federal
Insecticide, Fungicide, and Rodenticide Act (FIFRA]; and by considering late-breaking studies that
would affect the credibility of the conclusions, even during the review process. Studies identified
after peer review begins will only be considered for inclusion if they are directly relevant to the
PECO criteria and are expected to fundamentally alter the assessment's conclusions. Release of the
PECO-screened literature in parallel with release of the protocol for public comment provides an
opportunity for stakeholders to identify any missing studies, which if identified, will be screened as
outlined above for adherence to the PECO criteria.
4.1.1. Non-Peer-Reviewed Data
IRIS assessments rely mainly on publicly accessible, peer-reviewed studies. However, it is
possible that gray literature (i.e., studies that are not reported in the peer-reviewed literature]
directly relevant to the PECO may be identified during assessment development (e.g., good
This document is a draft for review purposes only and does not constitute Agency policy.
33 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
laboratory practices [GLP] studies submitted to EPA, dissertations, etc.]. In this case, if the data
make a substantial impact on assessment decisions or conclusions (i.e., have potential to affect the
PECO statement, hazard conclusions, or dose-response analysis], EPA can obtain external peer
review if the owners of the data are willing to have the study details and results made publicly
accessible. This independent, contractor-driven peer review would include an evaluation of the
study, as is done for peer review of a journal publication. The contractor would identify and select
two to three scientists knowledgeable in scientific disciplines relevant to the topic as potential peer
reviewers. Persons invited to serve as peer reviewers would be screened for conflict of interest
before confirming their service. In most instances, the peer review would be conducted by letter
review. The study authors would be informed of the outcome of the peer review and given an
opportunity to clarify issues or provide missing details. The study and its related information, if
used in the IRIS assessment, would become publicly available. In the assessment, EPA would
acknowledge that the document underwent external peer review managed by the EPA, and the
names of the peer reviewers would be identified. In certain cases, IRIS will conduct an assessment
for utility and data analysis based on having access to a description of study methods and raw data
that have undergone rigorous quality assurance/quality control review (e.g., ToxCast/Tox21 data;
results of National Toxicology Program [NTP] studies] but that have not yet undergone external
peer-review.
Unpublished (e.g., raw] data from personal author communication can supplement a
peer-reviewed study, as long as that information is made publicly available. If such ancillary
information is acquired, it will be documented in either the Health Assessment Workspace
Collaborative (HAWC] or HERO project page for the PFAS being assessed (depending on the nature
of the information received],
4.2. SCREENING PROCESS
As described below, PECO criteria or predefined inclusion and exclusion criteria (i.e., the
latter were used for the initial search] were used by two independent reviewers to screen and
inventory studies at the title and abstract level. For those studies considered relevant at the title
and abstract level, these criteria were then used to determine inclusion or exclusion of a reference
based on the full text. In addition to the PECO criteria, the following exclusion criteria were
applied:
• Review, commentary, other agency assessment, letter, or other record that does not contain
original data (note that these records were tracked for potential use in identifying
study-specific, original data relevant to specific scientific questions during assessment
development, including scanning of reference lists for unidentified studies; any such studies
incorporated into the assessment will be tracked under "other" as the reference source in
updates to the protocol]
This document is a draft for review purposes only and does not constitute Agency policy.
34 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
1 • Study available only as an abstract (e.g., conference abstract]
2 • Full text of the study is not available, and screening decisions could not be made at the
3 title/abstract level
4
5 In addition to including studies that meet PECO criteria, other studies containing material
6 that is potentially relevant to the assessments' objectives and specific aims were tracked during the
7 screening process as "potentially relevant supplemental material." These studies were not
8 excluded, but they may not be incorporated into the assessments unless they are deemed to be
9 relevant to addressing the key science issues, specific aims (see Sections 2.4 and 3.1], or key
10 scientific uncertainties identified at later stages of assessment development (see Section 9], Studies
11 categorized as "potentially relevant supplemental material" include the following:
12
13 • In vivo mechanistic or mode-of-action studies, including non-PECO routes of exposure and
14 populations (e.g., nonmammalian models] and studies examining potential susceptibility
15 • In vitro and in silico models
16 • ADME and toxicokinetic studies (excluding models]
17 • Exposure assessment or characterization (no health outcome] studies
18 • PFAS mixture studies (no individual PFAS comparisons]
19 • Human case reports or case-series studies
20 • Ecotoxicity studies
21 • Studies on PFAS manufacture/use
22 • Treatment/remediation studies
23 • Studies of PFAS analysis or other laboratory methods
24 • Environmental fate and transport studies
25 • Studies of other PFAS
26
27 Several of these categories of studies were further screened for consideration in addressing
28 the key science issues (described in Section 98],
29 Title and abstract screening. Following a pilot phase to calibrate screening guidance, two
30 screeners independently performed a title and abstract screen using a structured form in
31 DistillerSR (Evidence Partners; https://distillercer.com/products/distillersr-systematic-review-
32 software/I For citations with no abstract, the article was excluded if screening decisions could not
This document is a draft for review purposes only and does not constitute Agency policy.
35 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
be made based on the title and other citation information (e.g., page length] and additional attempts
to acquire the abstract or full text were unsuccessful. Screening conflicts were resolved by
discussion among the primary screeners with consultation by a third reviewer or technical advisor
(if needed] to resolve any remaining disagreements. Eligibility status of non-English studies was
assessed using the same approach with online translation tools used as needed to evaluate portions
of the study text and assess eligibility at the title and abstract level.
Studies not meeting title/abstract criteria but identified as "potentially relevant
supplemental material" were categorized (i.e., tagged] during the title and abstract screening
process (further described in Section 4.3], Conflict resolution was not required during the
screening process to identify supplemental information (i.e., tagging by a single screener was
considered adequate to identify the study as potentially relevant supplemental material for possible
inclusion during draft development].
Before beginning the Integrated Risk Information System (IRIS] PFAS assessments project,
the EPA contractor that conducted the July 2017 literature search as part of an EPA-wide
workgroup had performed a title and abstract screen to bin studies into different categories
(e.g., human, in vivo animal, excluded]. At the initiation of these PFAS assessments within IRIS, a
formalized effort was deployed with a new title and abstract screen of all studies identified in the
initial July 2017 search based on the PECO criteria in Table 8. For this initial literature screening,
specific inclusion/exclusion criteria were applied in the formalized title and abstract screen (see
Appendix B, Table B-6], Title and abstract screening of studies identified during literature search
updates will be conducted using the PECO criteria in Table 8 in DistillerSR using forms that
facilitate simultaneous initial tagging during screening (e.g., category of supplemental data;
contains data on other PFAS of interest]. An example of the questions and answers populating the
DistillerSR form for title/abstract and full-text (below] screening during literature search updates
is provided in Appendix B, Table B-7.
Full-text screening. Records that were not excluded based on the title and abstract were
advanced to full-text review. Full-text copies of these potentially relevant records were retrieved,
stored in the HERO database, and independently assessed by two screeners using a structured form
in DistillerSR to confirm eligibility. Screening conflicts were resolved by discussion among the
primary screeners with consultation by a third reviewer or technical advisor (as needed to resolve
any remaining disagreements]. As with the title and abstract screening, some studies were also
identified as "potentially relevant supplemental material" based on full-text screening. Approaches
for language translation included engagement of a native speaker from within EPA or use of
fee-based translation services.
In addition to identifying studies as included, excluded, or potential supplemental material,
the reviewers used the DistillerSR screening forms to confirm the specific PFAS (or multiple PFAS]
evaluated and to document several important experimental features of the studies (see Section 4.3],
This document is a draft for review purposes only and does not constitute Agency policy.
36 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
The results of this screening process are documented in the HERO database
fhttps: //hero.epa.gov: see Section 4.1 for links to the specific HERO pages] and literature flow
diagrams (see Figure 4], with individual studies "tagged" in HERO according to their appropriate
category descriptors (e.g., reference source; screening decisions regarding inclusion, exclusion, or
identification as supplemental; type of study],
4.2.1. Multiple Publications of the Same Data
When there are multiple publications using the same or overlapping data, all publications
on the research were included, with one selected for use as the primary study; the others were
considered as secondary publications with annotation in HAWC indicating their relationship to the
primary record during data extraction. For epidemiology studies, the primary publication was
generally the one with the longest follow-up, the largest number of cases, or the most recent
publication date. For animal studies, the primary publication was typically the one with the most
recent publication date, longest duration of exposure, or the one that assessed the outcome(s] most
informative to the PECO. For both epidemiology and animal studies, the assessments will include
relevant data from all publications of the study, although if the same data are reported in more than
one study, the data will only be extracted once (see Section 8]. For corrections, retractions, and
other companion documents to the included publications, a similar approach to annotation was
taken (see Section 4.1], and the most recently published data will be incorporated in the
assessments.
4.2.2. Literature Flow Diagrams
Figure 4 presents the literature flow diagrams for PFBA (a], PFHxA (b], PFHxS (c], PFNA (d],
and PFDA (e].6 These figures reflect literature searches through 2018. A literature search update
has been conducted and the results will be reflected in the draft assessments (and the most current
results can be viewed at any time in the HERO project pages provided in Section 4.1], Note that the
potential for updates or revisions to these figures related to CBI data and other reference decisions
is discussed in the previous sections.
6Note that the literature searches included the associated salts for each of the five PFAS, as presented in Figure 1
(Section 2.1.1). In addition, although not identified (yet) as part of the formal literature searches and not included
in these diagrams, several recent PBPK studies found through regular monitoring of new studies are included in
this protocol (see Section 6.4) so that the process for evaluating those studies can be outlined.
This document is a draft for review purposes only and does not constitute Agency policy.
37 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
(a)
PFBA
Literature Searches (through 2018)
PubMed
(n = 461)
WOS
(n = 456)
ToxLine
(n = 28)
TSCATS
(n = 0)
Other
ATSDR assessment (n = 1)
Submitted to EPA (n= J)
I
TITLE AND ABSTRACT SCREENING
This document is a draft for review purposes only and does not constitute Agency policy.
38 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
(b]
PFHxA
Literature Searches (through 2018)
PubMed
(n = 239)
WOS
(n = 245)
ToxLine
(n = 17)
TSCATS
(n = 0)
Other
ATSDR assessment (n = 17)
Submitted to EPA (n = 0)
NTP Report {n = 1)
i
TITLE AND ABSTRACT
Title & Abstract Screening
(285 records after duplicate removal)
FULL TEXT SCREENING
1
Studies Meeting PECO (n = 32)
Human health effects studies (n = 17)
Animal health effect studies (n = 10)
Genotoxicity studies (n = 4)
Susceptible population (n = 4)
Excluded (n= 205)
Not relevant to PECO (n = 205)
Excluded (n= 8)
• not relevant to PECO (n = 1), review,
commentary, or letter (n = 6), abstract-only
(n = 0), unable to obtain full text (n = 0),
other (n = 2)
Tagged as Supplemental (n= 45)
mechanistic or MOA (n = 8), ADME (n = 18),
qualitative exposure only (n = 12), mixture-
only (n = 3), non-PECO route of exposure (n
= 2), case report or case study (n = 2)
This document is a draft for review purposes only and does not constitute Agency policy.
39 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
i 00
PFHxS
Literature Searches (through 2018)
PubMed WOS
(n = 476) (n =517 )
ToxLine TSCATS
(n = 266 ) (n = 10 )
r ~\
Additional Strategies
(n=52)
^ J
I
TITLE AND ABSTRACT
Title & Abstract Screening
(507 records after duplicate removal)
Excluded (n= 254)
• Not relevant to PECO (n = 254)
FULL TEXT SCREENING
Full-Text Screening
(n =
234)
Studies Meeting PECO (n = 114)
Human health effects studies (n = 98)
Animal health effect studies (n = 14)
Genotoxicity studies (n = 1)
PBPK models (n = 1)
Excluded (n= 55)
• not relevant to PECO (n = 28), review or
regulatory document (n = 27), abstract-only
(n = 0)
Tagged as Supplemental (n= 70)
mechanistic or MOA (n = 26), ADME (n = 32),
qualitative exposure only (n = 12), mixture-only (n
= 0), non-PECO route of exposure (n = 0), case
report or case study (n = 0)
This document is a draft for review purposes only and does not constitute Agency policy.
40 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
1 [d]
PFNA
Literature Searches (through 2018)
PubMed
(n = 1,111)
WOS
(n = 707)
ToxLine
(n = 862)
TSCATS
(n = 57)
Other
ATSDR assessment (n = 36)
Submitted to EPA (n = 0)
Non-English or non-peer
reviewed (n = 37)
NTP (2018) report (n = 1)
1
TITLE AND ABSTRACT SCREENING
Studies Meeting PECO (n = 195)
• Human health effects studies (n = 166)
• Animal health effect studies (n = 20)
• Genotoxicity studies (n = 8)
• PBPK models (n = 1)
• Accessory records, such as published
corrections for included studies (n = 0)
Tagged as Supplemental (n = 124)
• mechanistic or MOA (n = 48), ADME (n =
16), exposure assessment or qualitative
exposure only (n = 64), mixture-only (n = 8),
non-PECO route of exposure (n = 10),
ecotox or zebrafish (n = 13), in silico or
modeling (n = 8), environmental occurrence
(n = ), case report or case study (n =0)
This document is a draft for review purposes only and does not constitute Agency policy.
41 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
(e)
PFDA
Literature Searches (through 2018)
PubMed
(n = 515)
WOS
(n = 613)
ToxLine
(n = 120)
TSCATS
(n = l)
Other
From draft ATSDR
assessment (n = 21)
Submitted to EPA (n = 0)
NTP report (n=l)
i
TITLE AND ABSTRACT SCREENING
Title & Abstract Screening
(730 records after duplicate removal)
FULL TEXT SCREENING
Full-Text Screening
(n = 303)
1
Studies Meeting PECO (n = 117)
• Human health effects studies (n = 95)
• Animal health effect studies (n = 13)
• Genotoxicity studies (n = 6)
• PBPK models (n = 1)
• Accessory records, such as published
corrections for included studies (n =4)
Excluded (n= 393)
Not relevant to PECO (n = 393)
Excluded (n= 16)
not relevant to PECO (n = 4), review,
commentary, or letter (n =11), abstract-only
(n = 0), unable to obtain full text (n = 1)
Tagged as Supplemental (n= 202)
• mechanistic or MOA (n = 121), ADME (n =
32), exposure assessment or qualitative
exposure only (n = 62), mixture-only (n = 4),
non-PECO route of exposure (n =74), case
report or case study (n = 0)
Figure 4. Literature flow diagrams for PFBA and its ammonium salt (a), PFHxA
and its ammonium and sodium salts (b), PFHxS and its potassium salt (c),
PFNA and its ammonium and sodium salts (d), and PFDA and its ammonium
and sodium salts (e).
4.3. SUMMARY-LEVEL LITERATURE INVENTORIES
4 As noted in Section 4.2, during title/abstract or full-text level screening, studies tagged
5 based on PECO eligibility were further categorized based on features such as evidence type (human,
6 animal, mechanistic, PBPK, etc.), health outcome(s), and/or endpointmeasure(s] included in the
7 study, and the specific PFAS (or multiple PFAS) addressed (see Appendix B, Table B-7 for
This document is a draft for review purposes only and does not constitute Agency policy.
42 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
examples]. Literature inventories for PECO-relevant studies were created to develop
summary-level, sortable lists that include some basic study design information (e.g., study
population, exposure information such as doses administered or biomarkers analyzed,
age/lifestage7 of exposure, endpoints examined, etc.]. These working literature inventories are for
internal use and facilitate subsequent review of individual studies or sets of studies by
topic-specific experts.
Inventories were also created for studies that were tagged as "potentially relevant
supplemental material" during screening, including in vitro or in silico models not addressing
genotoxicity, AD ME studies, and studies on endpoints or routes of exposure that did not meet the
specific PECO criteria, but which may still be relevant to the research question(s], Here, the
objective was to create an inventory of studies that can be tracked and further summarized as
needed—for example, by model system, key characteristic [e.g., of carcinogens, fSmithetal.. 20161].
mechanistic endpoint, or key event—to support analyses of potentially critical mechanistic
questions that arise at various stages of the systematic review (see Section 9.2 for a description of
the process for determining the specific questions and pertinent mechanistic studies to be
analyzed]. For example, AD ME data and related information are important to the next steps of
evaluating the evidence from individual PECO-specific studies, and these data will be reviewed by
subject matter experts early in the assessment process. Thus, the comprehensive identification of
studies relevant to interpreting the AD ME or toxicokinetic characteristics of these PFAS was
prioritized (see additional discussion in Section 5, and the specifics of the approach in Section 9.2],
7Age/lifestage was considered according to EPA's Guidance on Selecting Age Groups for Monitoring and Assessing
Childhood Exposures to Environmental Contaminants and EPA's A Framework for Assessing Health Risk of
Environmental Exposures to Children.
This document is a draft for review purposes only and does not constitute Agency policy.
43 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
5. REFINED EVALUATION PLAN
The primary purpose of this step is to outline any potential or expected refinements to the
set of populations, exposures, comparators, and outcomes (PECO]-relevant studies, which would
narrow the scope of studies considered for use in evidence synthesis and beyond. This optional
step is typically applied to focus an assessment with a veiy large number of PECO-relevant studies
on review of the most informative evidence (e.g., when many studies examine the same health
outcome, focusing on toxicity studies including exposures below a specified range, those studies
examining more specific or objective measures of toxicity, or those that address lifestage- or
exposure duration-specific knowledge on how the health outcome develops]. Given the relatively
small databases of animal toxicology studies for these five PFAS (see Section 2.3.2], this narrowing
is not considered applicable to these data. Thus, for these five PFAS assessments, all relevant health
outcomes in the animal toxicology studies meeting PECO criteria will be considered.
In contrast to the animal studies, there are many epidemiology studies. To make the
systematic review of the epidemiology literature more pragmatic and efficient and focus the set of
studies undergoing study evaluation a systematic map of the available evidence was developed
after literature screening. The PECO criteria described in Section 3.2 were intentionally broad and
inclusive and are well suited for application to systematic mapping. Using the literature inventoiy,
one epidemiologist per outcome reviewed the available evidence and summarized at a high level
the direction and consistency of observed associations. In the systematic map (Table 9], the
summary of available evidence for each outcome is shaded based on the consistency of direction of
association in the studies. These summaries do not account for study risk of bias, with the exception
of easily identified critical deficiencies that would make a study or set of studies uninformative (e.g.,
considerable concern for exposure measurement, confounding, or reverse causation].
Based on the systematic map, outcomes were classified into one of three different tiers of
further review based on the likely impact of the outcome on hazard identification and
dose-response analyses: (1] systematic review with formal study evaluation (see Section 6.2] with
at least two reviewers and evidence synthesis; (2] systematic review with formal study evaluation
with one reviewer and evidence synthesis; or (3] systematic map (SM] only - no study evaluation or
synthesis of the evidence, although the available database might be mentioned in the assessments
to inform data gaps]. The depth of review for the first two tiers is equivalent other than the number
of reviewers for study evaluation, though the syntheses for the second tier will typically be more
succinct due to weaker available evidence.
The determination for review tier was based on the following aspects: (1] consistency of
direction of available evidence; (2] consideration of null results in the context of study sensitivity
(e.g., if all studies for an outcome reported no association, but also had poor sensitivity due to, for
This document is a draft for review purposes only and does not constitute Agency policy.
44 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
example, PFAS exposure levels being below or near the limit of detection [LOD], SM only would be
used due to an assumed inability to draw any conclusion with confidence]; (3] identification of
critical deficiencies that would make a study or set of studies uninformative as described above;
and (4] consideration of the amount of available animal evidence addressing the human health
outcome and the likelihood that a review of the available epidemiology evidence would inform
hazard or dose-response conclusions.
In general, an outcome with only uninformative (i.e., having critical deficiencies] or null
studies with low sensitivity (shaded light blue in the systematic map] would receive no further
review. The systematic review with one reviewer category was used for outcomes in which the
available evidence was sparse or the reported results were inconsistent (shaded yellow in the
systematic map], but for which there was still a need to conduct a review to provide an informed
summary of the available data to support other parts of the assessment (e.g., outcomes where a
detailed animal evidence synthesis was expected based on the literature survey in Section 2.3.2;
identification of potential research needs or important uncertainties]. Generally, outcomes with
some or more consistency (shaded light or dark pink in the systematic map] were classified as
systematic review with two reviewers, but in a few cases, outcomes with some consistency (shaded
light pink] were classified as requiring one reviewer if the complexity appeared to be low (e.g., a
small number of studies]. In addition, there were logistical reasons that some outcomes were
upgraded to two reviewers (e.g., training staff, NTP participation, the same studies used in another
outcome receiving two reviews]. Outcomes classified as one reviewer can be upgraded if additional
complexity is identified during the review.
This approach of tiered reviews is consistent with recommendations from the National
Academies of Science encouraging the U.S. Environmental Protection Agency (EPA] to explore ways
to make systematic review more feasible, including conducting a "rapid review in which
components of the systematic review process are simplified or omitted (e.g., the need for two
independent reviewers]" (NAS. 20141
This document is a draft for review purposes only and does not constitute Agency policy.
45 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 9. Systematic map of epidemiology outcomes
Health effect/
Outcome
Number of publications3
Summary of available evidence
Study
evaluation
approach
(# reviewers)
PFNA
PFHxS
PFDA
PFHxA
PFBA
MALE REPRODUCTIVE EFFECTS
Semen parameters/
sperm DNA damage
5
5
2
1
1
Lower semen parameters in one study for PFNA, PFDA, and PFHxS (ns]
1
Repro hormones
7
7
4
1
0
Mix of higher and lower levels of estradiol and testosterone for PFNA
AND PFHxS. No association for PFDA
2b
Anogenital distance
1
1
1
0
0
Smaller AGD in boys for PFHxS (significant], but not PFNA or PFDA.
1
Penile width
1
1
1
0
0
No association in single study with low sensitivity
SM
FEMALE REPRODUCTIVE EFFECTS
Repro hormones
6
6
4
1
0
Mix of higher and lower levels of estradiol and testosterone for PFNA
AND PFHxS. No association for PFDA
2b
Fecundity
6
6
3
0
0
Three studies observed some positive association with PFNA (one
significant], one with PFHxS, but not consistently in subpopulations.
One study with inverse association with PFNA, PFHxS, and PFDA.
1
Pubertal timing
1
1
0
0
0
No association in single study with low sensitivity
SM
Menstrual cycle
characteristics
2
1
2
0
0
One study reported significant association with irregular cycle; concern
for reverse causality. No association reported with PFDA
1
Endometriosis
3
3
2
0
0
Two studies reported association with endometriosis for PFNA
(significant] and PFHxS (ns], one study with association for PFDA (ns];
concern for reverse causality
1
Menopause
1
1
0
0
0
Earlier age at menopause in one study. High potential for reverse
causality
SM
Anogenital distance
1
1
1
0
0
Smaller AGD in girls in single study (significant]
1
DEVELOPMENTAL EFFECTS
Birth size/fetal
growth restriction
17
20
14
0
1
Majority of PFAS studies showed some evidence of birth weight deficits
either in the overall population, or among male or female neonates, with
statistical significance in some.
2
Preterm birth/
gestational duration
7
8
4
0
1
Shorter gestational duration in 2 studies for PFNA, PFHxS, and PFDA.
2C
Postnatal growth
4
5
3
0
0
Inverse association with height and/or weight in 2 studies for PFNA,
PFHxS, and PFDA (some significant], but not consistent for all measures.
Positive association in 2 studies for PFHxS and 1 study for PFNA (ns].
1
Spontaneous abortion
2
2
2
0
0
Positive association observed in one study, inverse (PFNA and PFDA] or
no (PFHxS] association in one study
1
Sex ratio
1
0
1
0
0
No association in single study with low sensitivity
SM
This document is a draft for review purposes only and does not constitute Agency policy.
46 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
ENDOCRINE EFFECTS
Thyroid hormones
and disease
22
20
10
2
1
Majority of studies reported no association, mix of positive and inverse
associations in remaining studies
1
IMMUNE EFFECTS
Asthma
11
11
8
3
0
Higher asthma in two studies for PFNA and one study for PFHxS
(significant]
2d
Allergy
6
6
3
0
0
No association in several studies with low sensitivity
2d
Antibody response
6
6
4
0
0
All studies of diphtheria and tetanus vaccination reported lower
antibody response (some significant] for at least some exposure and
outcome measure timing combinations.
2
Infections
4
4
2
0
0
Higher rate of infections observed in two studies for PFNA and PFHxS
(significant]. No association for PFDA.
2
Atopic dermatitis
6
5
6
0
0
Higher atopic dermatitis in one study for PFNA and two studies for
PFHxS (significant]
2d
HEPATIC EFFECTS
Liver enzymes
6
5
1
1
0
Higher liver enzyme levels in two studies for PFNA and PFHxS
(significant in one study], lower levels in one study for PFNA
(significant]. No association for PFDA.
1
Albumin
5
5
1
1
0
No association in studies with low sensitivity; concerns for reverse
causality
SM
Liver disease
1
1
1
0
0
Single study evaluated as uninformative
SM
URINARY/RENAL EFFECTS
Renal function tests
7
8
2
1
0
Associations with impaired renal function with PFNA (3 studies, 2
significant], PFHxS (3 studies, 3 significant], and PFDA (1 study, ns].
High potential for reverse causality.
1
CARDIOVASCULAR EFFECTS
Blood pressure
7
6
4
1
1
Higher odds of hypertension in two studies for PFNA (significant]. No
association with PFHxS or PFDA.
1
Serum lipids
20
18
9
3
1
Majority of studies report no association; higher serum lipid levels in
small number of studies
1
Atherosclerosis
3
2
2
0
0
Higher atherosclerosis in one study for PFNA and PFDA (ns]. No
association for PFHxS.
1
Coronary heart
disease
2
2
2
0
0
No association in two studies with low sensitivity
SM
Ventricular geometiy
1
1
1
0
0
Changes in ventricular geometiy in single study for PFNA (significant]
and PFDA (ns]. No association for PFHxS.
1
NERVOUS SYSTEM EFFECTS
Neurodevelopment
17
15
10
0
0
Small number of studies per specific outcome (e.g., cognition, motor,
attention, autism spectrum disorder, cerebral palsy]. Mix of positive,
inverse, and no associations observed.
1
Adult neurologic
effects
3
3
2
0
0
Single study per effect (memoiy, depression, sleep], with latter two
evaluated as uninformative
SM
This document is a draft for review purposes only and does not constitute Agency policy.
47 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
METABOLIC EFFECTS
Diabetes
7
6
2
0
0
Higher odds of diabetes in two studies for PFNA and PFHxS (ns] and one
study for PFDA (significant]. Lower odds in one study for PFHxS
(significant] and PFDA (ns].
1
Gestational diabetes
3
3
2
0
0
Two studies report higher odds of gestational diabetes (ns] for PFHxS.
No association for PFNA and PFDA.
1
Insulin resistance
12
12
4
1
0
Higher insulin resistance in two studies for PFNA (ns] and PFHxS (one
significant]. No association for PFDA.
1
Adiposity
10
9
3
0
0
For PFNA and PFDA, most studies report higher adiposity in one
outcome measure (ns], but not consistently across measures. No
association with adiposity for PFHxS. One study for PFNA, PFDA, and
PFHxS reported higher weight gain (significant].
2
Metabolic syndrome
3
3
1
0
0
One study evaluated as uninformative; no association in other studies
with low sensitivity
SM
OTHER EFFECTS
Cancer
6
7
4
0
0
Two studies were evaluated as uninformative. One study available per
cancer type with the exception of breast cancer. Higher odds of breast
cancer in one study and lower odds in one study (both significant for
PFHxS and ns for PFNA].
1
Hematologic effects
1
1
1
1
0
No association in single study with low sensitivity
SM
Mortality
1
1
0
0
0
No association in single study with low sensitivity
SM
a Number of publications does not account for multiple publications of the same study
b Number of reviewers for study evaluation increased to 2 due to staff training on evaluation of hormonal measures with this approach (see Section 6.2)
c Number of reviewers for study evaluation increased to 2 due to the same studies being used for birth size
d Number of reviewers for study evaluation increased to 2 because systematic review for immune outcomes was performed by NTP.
SM = systematic map only; ns = not statistically significant; BMI = body mass index; DNA = deoxyribonucleic acid; PFBA = perfluorobutanoic acid;
PFDA = perfluorodecanoic acid; PFHxA = perfluorohexanoic acid; PFHxS = perfluorohexanesulfonate; PFNA = perfluorodecanoic acid.
1
2 Legend for table shading
Number of publications
1-4 10+
Summary of available evidence
No association in set of studies or the direction of the association(s) observed is not considered to be toxicologically relevant
Mix of positive, inverse, and no association results in set of studies, with less than 1/2 of studies in uniform direction.
OR 1/2 of studies report association in uniform direction but none are statistically significant
OR only one study is available, and results are in the direction of a detrimental effect but are not statistically significant
OR >1/2 of studies report association in uniform direction but there is considerable uncertainty due to potential for reverse causality
>1/2 of studies report association in uniform direction but none are statistically significant
OR 1/2-2/3 of studies report association in uniform direction and at least one is statistically significant
OR only one study is available and reports statistically significant association
>2/3 of studies report an association in uniform direction and at least one is statistically significant
3
This document is a draft for review purposes only and does not constitute Agency policy.
48 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
To promote consistency in evaluation and presentation across assessments, preliminary
decisions were made regarding the grouping of related endpoints for outcome-specific study
evaluations and discussion in the evidence synthesis. This helps implement the study evaluation
criteria (see Section 6] because those evaluations are outcome and analysis specific. Preliminary
decisions for grouping of endpoints from animal toxicology studies for discussion within each
assessed human health effect category are described in Table 10. Parallel groupings for outcomes
assessed in the available epidemiology studies are captured in Table 9. These groupings are meant
to serve as a starting place for consistency in presentation and analysis across studies and
assessments, although assessment-specific deviations are possible (e.g., depending on the
assessment-specific database of endpoints in the available studies or PFAS-specific understanding
of mechanistic relationships across outcomes].
This document is a draft for review purposes only and does not constitute Agency policy.
49 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 10. Animal Endpoint Grouping Categories
Relevant human health
effect category3
Examples of animal endpoints included
Notes
General toxicity
• Body weight (not maternal or pup weights, or
weights after developmental-only exposure)
• Mortality, survival, or LD50s
• Growth curve
• Clinical observations (non-behavioral)
• Clinical chemistry
endpoints are under
Hepatic or Hematologic
• Maternal or pup
body-weight endpoints
are under Developmental
• Pathology (including
gross lesions) is organ
specific
Hepatic (toxicity)15
• Liver weight and histopathology
• Serum or tissue liver enzymes (e.g., ALT and
AST from clinical chemistry)
• Other liver tissue enzyme activity
(e.g., catalase) or protein/DNA content
• Other liver tissue biochemical markers (e.g.,
albumin; glycogen; glucose)
• Liver-specific serum biochemistry
(e.g., albumin; albumin/globulin)
• Liver tissue lipids: triglycerides, cholesterol
• Serum lipids (Note: also informative to
cardiovasculard)
• Biochemical markers
such as albumin or
glucose are under
Hematological
• Liver tissue cytokines are
under Immune
• Serum glucose is under
Metabolic
Cardiovascular (toxicity)bc
• Heart weight and histopathology
• Serum lipids (note: also informative to
Hepatic)
• Blood pressure
• Blood measures of cardiovascular risk (e.g.,
C-reactive protein)
• Other blood measures
are under Hepatic,
Immune, or Hematologic
Hematologic (effects)bc
• Red blood cells
• Blood hematocrit or hemoglobin
• Corpuscular volume
• Blood platelets or reticulocytes
• Blood biochemical measures (e.g., sodium,
calcium, phosphorus)
• White blood cell count
and globulin are under
Immune
• Serum lipids are under
Cardiovascular
• Serum liver markers are
under Hepatic
Immune (effects)15
• Host resistance
• Red blood cells are under
Hematological
This document is a draft for review purposes only and does not constitute Agency policy.
50 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Relevant human health
effect category3
Examples of animal endpoints included
Notes
• Allergic, autoimmune or infectious disease
• Hypersensitivity
• General immune assays (e.g., white blood cell
counts, immunological factors or cytokines in
blood, lymphocyte phenotyping or
proliferation)
• Any measure in lymphoid tissues (weight;
histopathology; cell counts; etc.)
• Immune cell counts or immune-specific
cytokines in non-lymphoid tissues
• Other immune functional assays
(e.g., antibody production, natural killer cell
function, DTH, MLR, CTL, phagocytosis or
bacterial killing by monocytes)
• Immune responses in the respiratory system
• Stress-related factors in blood
(e.g., glucocorticoids or other adrenal
markers)
• Non-immune measures
of pulmonary function
are under Respiratory
Urinary (toxicity)15
• Kidney weight and histopathology
• Urinary measures (e.g., protein, volume, pH,
specific gravity)
Nervous system (effects)bc
• Brain weight
• Behavioral measures (including FOB and
cage-side observations)
• Nervous system histopathology
Endocrine (effects)15
• Thyroid weight and histopathology
• Hormonal measures in any tissue or blood
(non-reproductive)
• Reproductive hormones
are under Reproductive
This document is a draft for review purposes only and does not constitute Agency policy.
51 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Relevant human health
effect category3
Examples of animal endpoints included
Notes
Metabolic (effects)bc
• Free fatty acids
• Serum glucose or insulin, or other measures
related to diabetes
• Pancreatic effects relevant to diabetes
• Induced-obesity or BMI
• Any of the above endpoints after
developmental exposure will be primarily
discussed in this health effect category, and
then referenced under developmental effects
Reproductive (toxicity)15
Note: Evidence synthesis
and evidence integration
conclusions in assessments
are developed separately
for male and female
reproductive effects
(toxicity)
• Reproductive organ weight and
histopathology
• Markers of sexual differentiation or
maturation (e.g., preputial separation in
males; vaginal opening or estrous cycling in
females)
• Mating parameters (e.g., success, mount
latency)
• Reproductive hormones
• Birth parameters
(e.g., litter size;
resorptions,
implantations, viability)
are under Developmental
• If data indicate altered
birth parameters are
likely attributable to
female fertility, these
data may be discussed
under Female
Reproductive
Developmental (effects)15
Note: Evidence synthesis of
these endpoints in the
assessments is termed
"offspring growth and early
development," but
evidence integration
conclusions will be drawn
on the broader category of
"developmental effects"
(which also considers
organ/system-specific
effects after exposure
during development)
• Dam health (e.g., weight gain, food
consumption)
• Pup viability/survival or other birth
parameters (e.g., number of pups per litter)
• Pup weight or growth (includes measures into
adulthood after developmental-only
exposure)
• Developmental landmarks (eye opening, etc.,
but not including markers for other
organ/system-specific toxicities)
• Histopathology and
markers of development
specific to other systems
are organ/system-
specific (e.g., vaginal
opening is under Female
Reproductive; tests of
sensory maturation are
under Nervous System)
This document is a draft for review purposes only and does not constitute Agency policy.
52 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Relevant human health
effect category3
Examples of animal endpoints included
Notes
Carcinogenicity15
• Tumors
• Precancerous lesions (e.g., dysplasia)
ALT = alanine aminotransferase; AST = aspartate transaminase; BMI = body mass index; CTL = cytotoxicT
lymphocyte; DNA = deoxyribonucleic acid; DTH = delayed-type hypersensitivity; FOB = functional operational
battery; LD50 = median lethal dose; MLR = mixed leukocyte reaction.
aGiven the paucity of available studies and the absence of exceptional new evidence, information on
gastrointestinal effects, musculoskeletal effects, ocular effects, and respiratory effects will not be formally
evaluated in these assessments, although short summaries of the evidence may be included for context, and
new literature relating to these outcomes will be monitored during literature search updates for potential
inclusion.
bAny of the health effect-relevant endpoints observed after developmental exposure will be discussed primarily
in the health effect category indicated, and then referenced under developmental effects.
cThe primary focus of these assessments will be on developmental effects; endocrine effects; hepatic effects,
including lipid and lipoprotein measures; immune effects; reproductive effects in males or females; urinary
effects; general toxicity; and carcinogenicity. Data on cardiovascular effects, hematological effects, metabolic
effects including diabetes, and nervous system effects will be summarized when available. These summaries
may be developed in association with one of the health effects noted above, as a separate formal evaluation of
hazard, or as part of a qualitative summary on "other effects," depending on the assessment-specific data.
dSome outcomes are relevant to multiple health effects. These outcomes may be categorized under only a single
health effect in Table 10 for clarity. However, in the assessments, such outcome data would be discussed in the
first relevant health effect synthesis (syntheses will generally follow the pattern of most-to-least available
evidence) and then this synthesis will be cited in the syntheses of other relevant health effects. The evidence
(for or against an effect) will contribute to evidence integration decisions for all relevant health effects.
Assessment-specific refinements to the evaluation plan described in later sections of this
protocol may be justified after review of the key areas of scientific complexity outlined in
Section 2.4. Although not expected based on the relatively small database of studies for these PFAS,
one such refinement includes the potential prioritization of studies testing specific (lower]
exposure levels, exposure lifestages, routes of exposure, or toxic metabolites as identified based on
conclusions made regarding the AD ME properties of these PFAS. As noted in Section 2.4,
consideration of the available AD ME data for these five PFAS will be prioritized (see additional
discussion in Section 9.2], This will serve multiple purposes, including updating the data in Table 7
on serum half-lives across species and sexes. Notably, it is not expected that there will be enough
data to examine lifestage-specific differences in AD ME (including metabolic pathways for
toxification or detoxification] that might inform evidence evaluation and synthesis decisions. (This
is distinct from lifestage-specific differences in exposure, e.g., due to the higher intake of food per kg
body weight [B W] of young children or ingestion of dust.] However, a few anticipatory refinements
will be applied to study evaluations based on the preliminary data presented in Table 7.
Specifically, given the apparent sex-specific differences in PFAS half-life in rats and mice (note:
toxicology studies in nonhuman primates were not identified for these PFAS], examining and
reporting data for both sexes will be reviewed as a potential source of study insensitivity during
This document is a draft for review purposes only and does not constitute Agency policy.
53 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
study evaluation (see Section 6.3], particularly for PFAS that seem to vary largely for this parameter
(e.g., PFHxS; PFNA], These half-life data will also be considered when evaluating the
experiment-specific sensitivity of the frequency of exposures and the timing of endpoint testing
after exposure in experimental animals (see Section 6.3], as well as the potential for using exposure
biomarkers in exposed animals and humans (see Sections 6.2 and 6.3], The apparent AD ME
differences across species will be a critical consideration in these assessments. This consideration
will be applied to evidence synthesis and integration decisions (e.g., exploring ADME differences
between rats and mice as a potential explanation if there are differences in sensitivity in
outcome-specific responses; see Sections 9 and 10], as well as in extrapolating dosimetry
(i.e., exposure levels and duration] from experiments in animal models to quantitative estimates
relevant to humans, possibly including application within existing pharmacokinetic (PK] and PBPK
models (see additional discussion in Sections 6.4 and 11.2].
Lastly, based on the key areas of scientific complexity outlined in Section 2.4, some of the
analyses performed in support of these assessments may need to consider a broader array of
studies than those available for these five PFAS. One example includes the need to consider the
human relevance of certain outcomes observed in animals, including the role of receptors such as
PPARa (see additional discussion in Section 9.2], In addition, it is possible that additional literature
identification and evaluation strategies will be developed to address the other key areas of
scientific complexity outlined in Section 2.4, or other assessment-specific issues that arise during
review. Any such approaches will be documented in updates to the PFAS-specific assessment
protocols] (i.e., as assessment-specific updates to this document included as appendix materials
for each of the five PFAS assessments].
This document is a draft for review purposes only and does not constitute Agency policy.
54 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
6. STUDY EVALUATION (REPORTING, RISK OF BIAS,
AND SENSITIVITY) STRATEGY
The general approach for evaluating PECO-relevant primary health effect studies is
described in Section 6.1 and is the same for epidemiology and animal toxicology experiments, but
the specifics of applying the approach differ; thus, they are described separately for epidemiology
and animal toxicology studies in Sections 6.2 and 6.3, respectively. No controlled human exposure
studies for these PFAS were identified (see Section 4], Although they have not yet been formally
identified by the systematic literature searches (this is expected during the next literature search
update], PBPK modeling studies were recently identified for PFHxS fKim etal.. 20181 and for PFDA
and PFNA (Kim et al„ 20191 In addition, a two-compartment PK model for gestational and
lactational transfer of PFHxS in humans has been described by Verner etal. (20161 The specific
approach for reviewing these studies is described in Section 6.4. Different approaches are used to
evaluate mechanistic studies (see Sections 6.5 and 9.2],
6.1. STUDY EVALUATION OVERVIEW FOR HEALTH EFFECT STUDIES
Key concerns for the review of epidemiology and animal toxicology studies are potential
bias (factors that affect the magnitude or direction of an effect in either direction] and insensitivity
(factors that limit the ability of a study to detect a true effect; low sensitivity is a bias towards the
null when an effect exists]. The study evaluations are aimed at discerning the expected magnitude
of any identified limitations (focusing on limitations that could substantively change a result],
considering also the expected direction of the bias. The study evaluation approach is designed to
address a range of study designs, health effects, and chemicals. The general approach for reaching
an overall judgment for the study (or a specific analysis within a study] regarding confidence in the
reliability of the results is illustrated in Figure 5.
This document is a draft for review purposes only and does not constitute Agency policy.
55 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Study evaluation process (b)
Individual evaluation domains
Develop assessment-
specific considerations
Pilot testing
(and possible refinement)
Independent evaluation by
two reviewers
Finalize domain judgments
and overall study rating
Animal
Epidemiology
Selection and performance
¦ Allocation
Participant selection
¦ Observational bias/blinding
Confounding/variable control
Confounding
Selective reporting and attrition
Selective reporting
Exposure methods sensitivity
¦ Chemical administration and
Exposure measurement
characterization
¦ Exposure timing, frequency, and duration
Outcome measures and results display
Outcome ascertainment
• Endpoint sensitivity and specificity
Analysis
• Results presentation
Reporting quality
Other sensitivity
Domain judgments
Judgment
0 Good
Adequate
Deficient
Critically
Deficient
Interpretation
Appropriate study conduct relating to the domain and
minor deficiencies not expected to influence results
A study that may have some limitations relating to the
domain, but they are not likely to be severe or to
have a notable impact on results.
Identified biases or deficiencies interpreted as likely
to have had a notable impact on the results or
prevent reliable interpretation of study findings.
A serious flaw identified that makes the observed
effect(s) uninterpretable. Studies with a critical
deficiency will almost always be considered
*uninformative" overall.
Overall study rating for an outcome
Rating
Interpretation
High
No notable deficiencies or concerns identified; potential
for bias unlikely or minimal; sensitive methodology.
Medium
Possible deficiencies or concerns noted, but resulting
bias or lack of sensitivity is unlikely to be of a notable
degree.
Low
Deficiencies or concerns were noted, and the potential
for substantive bias or inadequate sensitivity could have
a significant impact on the study results or their
interpretation.
Uninformative
Serious flaw(s) makes study results unusable for hazard
identification or dose response.
Figure 5. Overview of Integrated Risk Information System (IRIS) study
evaluation process, (a) An overview of the general evaluation process [Note: see
Section 5 for deviations from independent evaluation by two reviewers for some
health outcomes in epidemiology studies], (b) The evaluation domains and
definitions for ratings (i.e., domain and overall judgments, performed on an
outcome-specific basis).
This document is a draft for review purposes only and does not constitute Agency policy.
56 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
With the exceptions noted in the refined evaluation plan for select outcomes reported in
epidemiology studies (see Section 5], at least two reviewers will independently evaluate health
effect studies to identify characteristics that bear on the informativeness (i.e., validity and
sensitivity] of the results. The independent reviewers will use the structured platform for study
evaluation housed within the Environmental Protection Agency's (EPA's] version of the Health
Assessment Workplace Collaboration (HAWC]8 to record separate judgements for each domain and
the overall study for each outcome, to reach consensus between reviewers, and when necessary,
resolve differences by discussion between the reviewers or consultation with additional
independent reviewers. For some domains, additional chemical- or outcome-specific knowledge
will be applied to evaluating the experimental design and methodology, as described below.
In general, considerations for reviewing a study with regard to its conduct for specific
health outcomes is based on the use of existing guidance documents when available, including EPA
guidance for carcinogenicity, neurotoxicity, reproductive toxicity, and developmental toxicity (U.S.
EPA. 2005a. 1998.1996a. 1991a], For some aspects of the study evaluations (e.g., review of
exposure assessment in epidemiology studies], additional considerations are developed in
consultation with topic-specific technical experts. To calibrate the assessment-specific
considerations, the study evaluations will include a pilot phase to assess and refine the evaluation
process. Additionally, as reviewers examine a group of studies, additional chemical-specific
knowledge or methodologic concerns may emerge and a second pass of all pertinent studies may
become necessary. Refinements to the study evaluation process made during the pilot phase and
subsequent implementation across all relevant studies will be acknowledged as updates to the
protocol.
Authors may be queried to obtain critical information, particularly that involving missing
reporting quality information or other data (e.g., information on variability] or additional analyses
that could address potential study limitations. The decision on whether to seek missing
information includes consideration of what additional information would be useful, specifically
with respect to any information that could result in a reevaluation of the overall study confidence
for an outcome. Outreach to study authors will be documented and considered unsuccessful if
researchers do not respond to an email or phone request within one month of the attempt to
contact. Only information or data that can be made publicly available (e.g., within HAWC or Health
and Environmental Research Online [HERO]] will be considered.
8HAWC is a free and open source software application that provides a modular, web-based interface to help
develop human health assessments of chemicals: https://hawcproiect.org/portal/. Standard operating
procedures provided to the reviewers to facilitate consistent and relevant documentation of their judgments
using the HAWC software can be found as attachments embedded within the online tool
(https://hawcprd.epa.gov/assessment/100000Q39/).
This document is a draft for review purposes only and does not constitute Agency policy.
57 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
1 When evaluating studies9 that examine more than one outcome, the evaluation process will
2 be performed separately for each outcome, because the utility of a study can vary for different
3 outcomes. If a study examines multiple endpoints for the same outcome,10 evaluations may be
4 performed at a more granular level if appropriate, but these measures may still be grouped for
5 evidence synthesis.
6 During review, for each evaluation domain reviewers will reach a consensus judgment of
7 good, adequate, deficient, not reported, or critically deficient. If a consensus is not reached, a third
8 reviewer will perform conflict resolution. It is important to stress that these evaluations are
9 performed in the context of the study's utility for identifying individual hazards. While limitations
10 specific to the usability of the study for dose-response analysis are useful to note to inform those
11 later decisions, they do not contribute to the study confidence classifications.
12 These four categories are applied to each evaluation domain for each study as follows:
13
14 • Good represents a judgment that the study was conducted appropriately in relation to the
15 evaluation domain, and any minor deficiencies that are noted would not be expected to
16 influence the study results.
17 • Adequate indicates a judgment that there may be methodological limitations relating to the
18 evaluation domain, but that those limitations are not likely to be severe or to have a notable
19 impact on the results.
20 • Deficient denotes identified biases or deficiencies that are interpreted as likely to have had a
21 notable impact on the results or that prevent interpretation of the study findings.
22 • Not reported indicates that the information necessary to evaluate the domain question was
23 not available in the study. Generally, this term carries the same functional interpretation as
24 deficient for the purposes of the study confidence classification (described below],
25 Depending on the number of unreported items and severity of other limitations identified in
26 the study, it may or may not be worth reaching out to the study authors for this information
27 (see discussion above],
28 • Critically deficient reflects a judgment that the study conduct relating to the evaluation
29 domain question introduced a serious flaw that is interpreted to be the primary driver of
30 any observed effect(s] or makes the study uninterpretable. Studies with a determination of
31 critically deficient in an evaluation domain will not be used for hazard identification or
32 dose-response analysis, but they may be used to highlight possible research gaps.
9Note: "study" is used instead of a more accurate term (e.g., "experiment") throughout these sections owing to an
established familiarity within the field for discussing a study's risk of bias or sensitivity, etc. However, all
evaluations discussed herein are explicitly conducted at the level of an individual outcome within a population or
cohort of humans or animals exposed in a similar manner (e.g., unexposed or exposed at comparable lifestages
and for the same duration of exposure), or to a sample of the study population within a study.
10Note: "outcome" will be used throughout these methods; the same methods also apply to an endpoint assessed
separately within a larger outcome. "Endpoint" refers to a more granular measurement (e.g., for the outcome of
liver histopathology, different endpoints might include necrosis and cellular hypertrophy).
This document is a draft for review purposes only and does not constitute Agency policy.
58 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Once the evaluation domains have been rated, the identified strengths and limitations will
be considered as a whole to reach a study confidence classification of high, medium, or low
confidence, or uninformative for a specific health outcome. This classification is based on the
reviewer judgments across the evaluation domains and considers the likely impact that inadequate
reporting or the noted deficiencies in bias and sensitivity have on the outcome-specific results. The
classifications, which reflect a consensus judgment between reviewers, are defined as follows:
• High confidence: No notable deficiencies or concerns were identified; the potential for bias
is unlikely or minimal, and the study used sensitive methodology, ///^-confidence studies
generally reflect judgments of good across all or most evaluation domains.
• Medium confidence: Possible deficiencies or concerns were noted, but the limitations are
unlikely to be of a notable degree. Generally, medium-confidence studies include adequate
or good judgments across most domains, with the impact of any identified limitation not
being judged as severe.
• Low confidence: Deficiencies or concerns are noted, and the potential for bias or inadequate
sensitivity could have a significant impact on the study results or their interpretation.
Typically, /ow-confidence studies have a deficient evaluation for one or more domains,
although some mec/Zum-con fide nee studies may have a deficient rating in domain(s]
considered to have less influence on the magnitude or direction of the outcome-specific
results]. Low-confidence results are given less weight compared to high- or
medium-confidence results during evidence synthesis and integration (see Section 10.1,
Table 20and Table 21], and are generally not used on their own for hazard identification or
dose-response analyses unless they are the only studies available or they inform data gaps
unexamined in the high- or medium-confidence studies. Studies rated as medium- or
low-confidence only because of sensitivity concerns about bias towards the null will be
asterisked or otherwise noted because they may require additional consideration during
evidence synthesis. Effects observed in studies biased toward the null may actually
increase confidence in the results, assuming the study is otherwise well conducted (see
Section 9],
• Uninformative: Serious flaw(s] make the study results unusable for hazard identification.
Studies with critically deficient judgements in any evaluation domain are almost always
classified as uninformative (see explanation above]. Studies with multiple deficient
judgments across domains may also be considered uninformative. Uninformative studies
will not be considered further in the synthesis and integration of evidence, except perhaps
to highlight possible research gaps.
As previously noted, study evaluation determinations reached by each reviewer and the
consensus judgment between reviewers will be recorded in HAWC. Final study evaluations housed
in HAWC will be made available when the draft is publicly released. The study confidence
classifications and their rationales will be carried forward and considered as part of evidence
synthesis (see Section 9] to help interpret the results across studies.
This document is a draft for review purposes only and does not constitute Agency policy.
59 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
6.2. EPIDEMIOLOGY STUDY EVALUATION
Evaluation of epidemiology studies of health effects to assess risk of bias and study
sensitivity will be conducted for the following domains: exposure measurement, outcome
ascertainment, participant selection, potential confounding, analysis, study sensitivity, and selective
reporting. Bias can result in false positives and negatives (i.e., Types I and II errors], while study
sensitivity is typically concerned with identifying the latter.
The principles and framework used for evaluating epidemiology studies are based on the
Cochrane Risk of Bias in Nonrandomized Studies of Interventions [ROBINS-I; (Sterne etal.. 20161]
but modified to address environmental and occupational exposures. The underlying philosophy of
ROBINS-I is to describe attributes of an "ideal" study with respect to each of the evaluation domains
(e.g., exposure measurement, outcome classification, etc.]. Core and prompting questions are used
to collect information to guide evaluation of each domain.
Core and prompting questions for each domain are presented in Table 11. Core questions
represent key concepts, while the prompting questions help the reviewer focus on relevant details
under each key domain. Table 11 also includes criteria that apply to all exposures and outcomes.
PFAS-specific criteria are described in Section 6.2.1. As mentioned in Section 6.1, any additions to
or refinements of the criteria will be documented as updates to the protocol.
This document is a draft for review purposes only and does not constitute Agency policy.
60 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 11. Questions and criteria for evaluating each domain in epidemiology studies
Domain and
core question
Prompting questions
Follow-up
questions
Criteria that apply to most exposures and outcomes
Exposure
measurement
Does the
exposure
measure reliably
distinguish
between levels
of exposure in a
time window
considered most
relevant for a
causal effect
with respect to
the development
of the outcome?
For all:
• Does the exposure measure capture the
variability in exposure among the
participants, considering intensity,
frequency, and duration of exposure?
• Does the exposure measure reflect a
relevant time window? If not, can the
relationship between measures in this
time and the relevant time window be
estimated reliably?
• Was the exposure measurement likely to
be affected by a knowledge of the
outcome?
• Was the exposure measurement likely to
be affected by the presence of the
outcome (i.e., reverse causality)?
For case-control studies of occupational exposures:
• Is exposure based on a comprehensive job
history describing tasks, setting, time
period, and use of specific materials?
For biomarkers of exposure, general population:
• Is a standard assay used? What are the
intra- and inter-assay coefficients of
variation? Is the assay likely to be affected
by contamination? Are values less than
Is the degree of
exposure
misclassification
likely to vary by
exposure level?
If the
correlation
between
exposure
measurements
is moderate, is
there an
adequate
statistical
approach to
ameliorate
variability in
measurements?
If there is a
concern about
the potential
for bias, what is
the predicted
direction or
distortion of the
bias on the
effect estimate
(if there is
enough
information)?
Good
Valid exposure assessment methods used, which represent
the etiologically relevant time period of interest.
Exposure misclassification is expected to be minimal.
Adequate
• Valid exposure assessment methods used, which represent
the etiologically relevant time period of interest.
• Exposure misclassification may exist but is not expected to
greatly change the effect estimate.
Deficient
• Valid exposure assessment methods used, which represent
the etiologically relevant time period of interest. Specific
knowledge about the exposure and outcome raise
concerns about reverse causality, but there is uncertainty
whether it is influencing the effect estimate.
• Exposed groups are expected to contain a notable
proportion of unexposed or minimally exposed individuals,
the method did not capture important temporal or spatial
variation, or there is other evidence of exposure
misclassification that would be expected to notably change
the effect estimate.
This document is a draft for review purposes only and does not constitute Agency policy.
61 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Domain and
core question
Prompting questions
Follow-up
questions
Criteria that apply to most exposures and outcomes
the limit of detection dealt with
adequately?
• What exposure time period is reflected by
the biomarker? If the half-life is short,
what is the correlation between serial
measurements of exposure?
Critically deficient
• Exposure measurement does not characterize the
etiologically relevant time period of exposure or is not
valid.
• There is evidence that reverse causality is very likely to
account for the observed association.
• Exposure measurement was not independent of outcome
status.
Outcome
ascertainment
Does the
outcome
measure reliably
distinguish the
presence or
absence (or
degree of
severity) of the
outcome?
For all:
• Is outcome ascertainment likely to be
affected by knowledge of, or presence of,
exposure (e.g., consider access to health
care, if based on self-reported history of
diagnosis)?
For case-control studies:
• Is the comparison group without the
outcome (e.g., controls in a case-control
study) based on objective criteria with
little or no likelihood of inclusion of people
with the disease?
For mortality measures:
• How well does cause-of-death data reflect
occurrence of the disease in an individual?
How well do mortality data reflect
incidence of the disease?
For diagnosis of disease measures:
Is there a
concern that
any outcome
misclassification
is non-
differential,
differential, or
both?
What is the
predicted
direction or
distortion of the
bias on the
effect estimate
(if there is
enough
information)?
Good
• High certainty in the outcome definition (i.e., specificity
and sensitivity), minimal concerns with respect to
misclassification.
• Assessment instrument was validated in a population
comparable to the one from which the study group was
selected.
Adequate
• Moderate confidence that outcome definition was specific
and sensitive, some uncertainty with respect to
misclassification but not expected to greatly change the
effect estimate.
• Assessment instrument was validated but not necessarily
in a population comparable to the study group.
Deficient
• Outcome definition was not specific or sensitive.
• Uncertainty regarding validity of assessment instrument.
This document is a draft for review purposes only and does not constitute Agency policy.
62 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Domain and
core question
Prompting questions
Follow-up
questions
Criteria that apply to most exposures and outcomes
• Is the diagnosis based on standard clinical
criteria? If it is based on self-report of the
diagnosis, what is the validity of this
measure?
For laboratory-based measures (e.g., hormone
levels):
• Is a standard assay used? Does the assay
have an acceptable level of inter-assay
variability? Is the sensitivity of the assay
appropriate for the outcome measure in
this study population?
Critically deficient
• Invalid/insensitive marker of outcome.
• Outcome ascertainment is very likely to be affected by
knowledge of, or presence of, exposure.
Note: Lack of blinding should not be automatically construed to be
critically deficient.
Participant
selection
Is there evidence
that selection
into or out of the
study (or analysis
sample) was
jointly related to
exposure and to
outcome?
For longitudinal cohort:
• Did participants volunteer for the cohort
based on knowledge of exposure and/or
preclinical disease symptoms? Was entry
into the cohort or continuation in the
cohort related to exposure and outcome?
For occupational cohort:
• Did entry into the cohort begin with the
start of the exposure?
• Was follow-up or outcome assessment
incomplete, and if so, was follow-up
related to both exposure and outcome
status?
• Could exposure produce symptoms that
would result in a change in work
assignment/work status ("healthy worker
survivor effect")?
Were
differences in
participant
enrollment and
follow-up
evaluated to
assess bias?
If there is a
concern about
the potential
for bias, what is
the predicted
direction or
distortion of the
bias on the
effect estimate
(if there is
enough
information)?
Good
• Minimal concern for selection bias based on description of
recruitment process (e.g., selection of comparison
population, population-based random sample selection,
recruitment from sampling frame including current and
previous employees).
• Exclusion and inclusion criteria specified and would not
induce bias.
• Participation rate is reported at all steps of study
(e.g., initial enrollment, follow-up, selection into analysis
sample). If rate is not high, there is appropriate rationale
for why it is unlikely to be related to exposure
(e.g., comparison between participants and
nonparticipants or other available information indicates
differential selection is not likely).
Adequate
• Enough of a description of the recruitment process to be
comfortable that there is no serious risk of bias.
This document is a draft for review purposes only and does not constitute Agency policy.
63 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Domain and
core question
Prompting questions
Follow-up
questions
Criteria that apply to most exposures and outcomes
For case-control study:
• Were controls representative of
population and time periods from which
cases were drawn?
• Are hospital controls selected from a
group whose reason for admission is
independent of exposure?
• Could recruitment strategies, eligibility
criteria, or participation rates result in
differential participation relating to both
disease and exposure?
For population-based survey:
• Was recruitment based on advertisement
to people with knowledge of exposure,
outcome, and hypothesis?
Were
appropriate
analyses
performed to
address
changing
exposures over
time in relation
to symptoms?
Is there a
comparison of
participants and
nonparticipants
to address
whether
differential
selection is
likely?
• Inclusion and exclusion criteria specified and would not
induce bias.
• Participation rate is incompletely reported but available
information indicates participation is unlikely to be related
to exposure.
Deficient
• Little information on recruitment process, selection
strategy, sampling framework, and/or participation OR
aspects of these processes raises the potential for bias
(e.g., healthy worker effect, survivor bias).
Critically deficient
• Aspects of the processes for recruitment, selection
strategy, sampling framework, or participation result in
concern that selection bias is likely to have had a large
impact on effect estimates (e.g., convenience sample with
no information about recruitment and selection, cases and
controls are recruited from different sources with different
likelihood of exposure, recruitment materials stated
outcome of interest and potential participants are aware
of or are concerned about specific exposures).
Confounding
Is confounding of
the effect of the
exposure likely?
Is confounding adequately addressed by
considerations in:
• Participant selection (matching or
restriction)?
• Accurate information on potential
confounders and statistical adjustment
procedures?
If there is a
concern about
the potential
for bias, what is
the predicted
direction or
distortion of the
bias on the
effect estimate
(if there is
Good
• Conveys strategy for identifying key confounders. This
may include a priori biological considerations, published
literature, causal diagrams, or statistical analyses, with the
recognition that not all "risk factors" are confounders.
• Inclusion of potential confounders in statistical models not
based solely on statistical significance criteria (e.g., p < 0.05
from stepwise regression).
This document is a draft for review purposes only and does not constitute Agency policy.
64 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Domain and
core question
Prompting questions
Follow-up
questions
Criteria that apply to most exposures and outcomes
• Lack of association between confounder
and outcome, or confounder and
exposure in the study?
• Information from other sources?
Is the assessment of confounders based on a
thoughtful review of published literature, potential
relationships (e.g., as can be gained through
directed acyclic graphing), and minimizing potential
overcontrol (e.g., inclusion of a variable on the
pathway between exposure and outcome)?
enough
information)?
• Does not include variables in the models that are likely to
be influential colliders or intermediates on the causal
pathway.
• Key confounders are evaluated appropriately and
considered to be unlikely sources of substantial
confounding. This often will include:
o Presenting the distribution of potential
confounders by levels of the exposure of interest
and/or the outcomes of interest (with amount of
missing data noted);
o Consideration that potential confounders were
rare among the study population, or were
expected to be poorly correlated with exposure of
interest;
o Consideration of the most relevant functional
forms of potential confounders;
o Examination of the potential impact of
measurement error or missing data on
confounder adjustment; or
o Presenting a progression of model results with
adjustments for different potential confounders, if
warranted.
Adequate
• Similar to good but may not have included all key
confounders, or less detail may be available on the
evaluation of confounders (e.g., sub-bullets in good). It is
possible that residual confounding could explain part of
the observed effect, but concern is minimal.
This document is a draft for review purposes only and does not constitute Agency policy.
65 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Domain and
core question
Prompting questions
Follow-up
questions
Criteria that apply to most exposures and outcomes
Deficient
• Does not include variables in the models that have been
shown to be influential colliders or intermediates on the
causal pathway.
• And any of the following:
o The potential for bias to explain some of the
results is high based on an inability to rule out
residual confounding, such as a lack of
demonstration that key confounders of the
exposure-outcome relationships were considered;
o Descriptive information on key confounders
(e.g., their relationship relative to the outcomes
and exposure levels) are not presented; or
o Strategy of evaluating confounding is unclear or is
not recommended (e.g., only based on statistical
significance criteria or stepwise regression
[forward or backward elimination]).
Critically deficient
• Includes variables in the models that are colliders and/or
intermediates in the causal pathway, indicating that
substantial bias is likely from this adjustment; or
• Confounding is likely present and not accounted for,
indicating that all the results were most likely due to bias.
Analysis
Does the analysis
strategy and
presentation
convey the
necessary
• Are missing outcome, exposure, and
covariate data recognized, and if
necessary, accounted for in the analysis?
If there is a
concern about
the potential
for bias, what is
the predicted
direction or
Good
• Use of an optimal characterization of the outcome
variable, including presentation of subgroup- or lifestage-
specific comparisons (as appropriate for the outcome).
This document is a draft for review purposes only and does not constitute Agency policy.
66 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Domain and
core question
Prompting questions
Follow-up
questions
Criteria that apply to most exposures and outcomes
familiarity with
the data and
assumptions?
• Does the analysis appropriately consider
variable distributions and modeling
assumptions?
• Does the analysis appropriately consider
subgroups or lifestages of interest
(e.g., based on variability in exposure level
or duration or susceptibility)?
• Is an appropriate analysis used for the
study design?
• Is effect modification considered, based
on considerations developed a priori?
• Does the study include additional analyses
addressing potential biases or limitations
(i.e., sensitivity analyses)?
distortion of the
bias on the
effect estimate
(if there is
enough
information)?
• Quantitative results presented (effect estimates and
confidence limits or variability in estimates) (i.e., not
presented only as a p-value or "significant"/"not
significant").
• Descriptive information about outcome and exposure
provided (where applicable).
• Amount of missing data noted and addressed
appropriately (discussion of selection issues—missing at
random vs. differential).
• Where applicable, for exposure, includes LOD (and
percentage below the LOD), and decision to use log
transformation.
• Includes analyses that address robustness of findings,
e.g., examination of exposure-response (explicit
consideration of nonlinear possibilities, quadratic, spline,
or threshold/ceiling effects included, when feasible);
relevant sensitivity analyses; effect modification examined
based only on a priori rationale with sufficient numbers.
• No deficiencies in analysis evident. Discussion of some
details may be absent (e.g., examination of outliers).
Adequate
• Same as 'Good', except:
• Descriptive information about exposure provided (where
applicable) but may be incomplete; might not have
discussed missing data, cut-points, or shape of
distribution(s).
• Includes analyses that address robustness of findings
(examples in 'Good'), but some important analyses are not
performed.
This document is a draft for review purposes only and does not constitute Agency policy.
67 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Domain and
core question
Prompting questions
Follow-up
questions
Criteria that apply to most exposures and outcomes
Deficient
• Does not conduct analysis using optimal characterization of
the outcome variable.
• Descriptive information about exposure levels not
provided (where applicable).
• Effect estimate and p-value presented, without standard
error or confidence interval.
• Results presented as statistically "significant"/"not
significant."
Critically deficient
• Results of analyses of effect modification examined
without clear a priori rationale and without providing
main/principal effects (e.g., presentation only of
statistically significant interactions that were not
hypothesis driven).
• Analysis methods are not appropriate for design or data of
the study.
Selective
reporting
Is there reason
to be concerned
about selective
reporting?
• Were results provided for all the primary
analyses described in the methods
section?
• Is there appropriate justification for
restricting the amount and type of results
that are shown?
• Are only statistically significant results
presented?
If there is a
concern about
the potential
for bias, what is
the predicted
direction or
distortion of the
bias on the
effect estimate
(if there is
enough
information)?
Good
• The results reported by study authors are consistent with
the primary and secondary analyses described in a
registered protocol or methods paper.
Adequate
• The authors described their primary (and secondary)
analyses in the methods section and results were reported
for all primary analyses.
This document is a draft for review purposes only and does not constitute Agency policy.
68 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Domain and
core question
Prompting questions
Follow-up
questions
Criteria that apply to most exposures and outcomes
Deficient
• Concerns were raised based on previous publications, a
methods paper, or a registered protocol indicating that
analyses were planned or conducted that were not
reported, or that hypotheses originally considered to be
secondary were represented as primary in the reviewed
paper.
• Only subgroup analyses were reported, suggesting that
results for the entire group were omitted.
• Only statistically significant results were reported.
Sensitivity
Is there a
concern that
sensitivity of the
study is not
adequate to
detect an effect?
• Is the exposure range adequate to detect
associations and exposure-response
relationships?
• Was the appropriate population or
lifestage included?
• Was the length of follow-up adequate? Is
the time/age of outcome ascertainment
optimal given the interval of exposure and
the health outcome?
• Are there other aspects related to risk of
bias or otherwise that raise concerns
about sensitivity?
Adequate
• The range of exposure levels provides adequate variability
to evaluate the associations of relevance.
• The population was exposed to levels expected to have an
impact on response.
• The study population was sensitive to the development of
the outcomes of interest (e.g., ages, lifestage, sex).
• The timing of outcome ascertainment was appropriate
given expected latency for outcome development
(i.e., adequate follow-up interval).
• The study was adequately powered to observe an effect.
• No other concerns raised regarding study sensitivity.
Deficient
• Concerns were raised about the issues described for
adequate that are expected to notably decrease the
sensitivity of the study to detect associations for the
outcome.
1
This document is a draft for review purposes only and does not constitute Agency policy.
69 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
6.2.1. Epidemiology Study Evaluation Criteria Specific to These Five Per- and
Polyfluoroalkyl Substances (PFAS)
The exposure criteria described in Table 12 below are modified from the criteria developed
by NTP OHAT11 for their assessment of the association between PFOA and immune effects.
The estimated serum half-lives of PFAS in humans were presented in Table 7 (see
Section 2.4.1], In considering temporality concerns, some PFAS (PFHxS, PFDA, and PFNA] are
persistent compounds with longer (multiple year] half-lives in humans, so current exposure levels
may be indicative of critical exposure windows that were narrow or past exposures that extended
beyond the anticipated half-lives. In contrast, other PFAS appear to have half-lives of 1 month or
less (PFBA, PFHxA], and current exposure levels may not be indicative of past exposures that
extend beyond the anticipated half-lives. Some evidence suggests that the half-lives vary based on
sex, parity, interval between pregnancy, reproductive hormones, and gynecological disorders (Lau
et al.. 20071: therefore, these factors will be considered depending on the population^], critical
windows, and outcomes being examined.
Standard analytical methods of individual PFAS in serum or whole-blood using quantitative
techniques such as liquid chromatography-triple quadrupole mass spectrometry are considered to
be well-established methods fCDC. 2019a. b; ATSDR.2018: CDC. 2015: U.S. EPA. 2014a. b; CDC.
20091
"National Toxicology Program (NTP) Report:
https://ntp.niehs.nih.gov/ntp/ohat/pfoa pfos/pfoa pfosmonograph 508.pdf
NTP protocol: https://ntp.niehs.nih.gov/ntp/ohat/pfoa pfos/protocol 201506 508.pdf.
This document is a draft for review purposes only and does not constitute Agency policy.
70 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 12. Criteria for evaluating exposure measurementin epidemiology
studies of per- and polyfluoroalkyl substances (PFAS) and health effects
Rating
Criteria
Good
• Evidence that exposure was consistently assessed using well-established analytical
methods that directly measure exposure (e.g., measurement of PFAS in blood, serum, or
plasma).
OR
• Exposure was assessed using less established methods (e.g., measurement of PFAS in
breast milk) or methods that indirectly measure exposure (e.g., drinking water
concentrations and residential location/history, questionnaire or occupational exposure
assessment by a certified industrial hygienist) that are validated against well-established
direct methods (i.e., inter-methods validation: one method vs. another) in the target
population of interest.
And all the following:
• Exposure was assessed in a relevant time-window (i.e., temporality is established, and
sufficient latency occurred prior to disease onset) for development of the outcome based
on current biological understanding.
• There is evidence that sufficient exposure data measurements are above the limit of
quantification for the assay.
• The laboratory analysis included data on standard quality control measures with
demonstrated precision and accuracy.
Adequate
• Exposure was assessed using less established methods or indirect measures that are
validated but not in the target population of interest.
OR
• Evidence that exposure was consistently assessed using methods described in 'Good', but
there were some concerns about quality control measures or other potential for non-
differential misclassification.
And all the following:
• Exposure was assessed in a relevant time-window for development of the outcome
• There is evidence that sufficient exposure data measurements are above the limit of
quantification for the assay.
• The laboratory analysis included some data on standard quality control measures with
demonstrated precision and accuracy.
Deficient
Any of the following:
• Some concern, but no direct evidence, that the exposure was assessed using methods
that have not been validated or empirically shown to be consistent with methods that
directly measure exposure.
• Exposure was assessed in a relevant time window(s) for development of the outcome,
but there could be some concern about the potential for bias due to reverse causality3
between exposure and outcome, yet no direct evidence that it is present.
This document is a draft for review purposes only and does not constitute Agency policy.
71 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Rating
Criteria
Critically
deficient
Any of the following:
• Exposure was assessed in a time window that is unknown or not relevant for
development of the outcome. This could be due to clear evidence of bias from reverse
causality between exposure and outcome, or other concerns such as the lack of temporal
ordering of exposure and disease onset, insufficient latency, or having exposure
measurements that are not reliable measures of exposure during the etiologic
window(s).
• Direct evidence that bias was likely because the exposure was assessed using methods
with poor validity.
• Evidence of differential exposure misclassification (e.g., differential recall of self-reported
exposure).
• There is evidence that an insufficient number of the exposure data measurements were
above the limit of quantification for the assay.
aReverse causality refers to a situation where an observed association between exposure and outcome is not due
to causality from exposure to outcome, but rather due to the outcome of interest causing a change in the
measured exposure.
1
2 In addition, there are PFAS-specific considerations for the evaluation of confounding. As
3 discussed in Section 2.4.3, confounding across PFAS is an important area of uncertainty when
4 interpreting the results of epidemiology studies for individual PFAS (i.e., quantifying the effected of
5 an individual PFAS can potentially be confounded by other PFAS], Based on preliminary analyses,
6 correlations differ across the PFAS (see Figure 6], While some pairs have correlation coefficients
7 consistently above 0.6 (e.g., PFNA and PFDA], the correlations for most vary from 0.1 to 0.6
8 depending on the study, and little data is available on correlations with less commonly occurring or
9 detected PFAS like PFBA and PFHxA.
10
PFBA
PFDA
PFHxA
PFHxS
PFNA
PFOA
PFOS
PFBA*
1.00
0.01
0.45
0.14
0.15
0.03
0.07
PFDA
1.00
-0.03
0.28
0.73
0.42
0.48
PFHxA*
1.00
0.08
-0.07
0.19
-0.04
PFHxS
1.00
0.35
0.43
0.50
PFNA
1.00
0.54
0.51
Figure 6. Preliminary mean correlation coefficients across per- and
polyfluoroalkyl substances (PFAS) among studies in the inventory, for all
media types.
*PFBA and PFHxA correlations were based on three studies for PFOS, PFOA, and PFHxS, two studies for each other,
and one study for PFNA and PFDA, so these estimates are less stable than the other PFAS, which were all based
on >10 studies.
PFOS = perfluorooctane sulfonate.
This document is a draft for review purposes only and does not constitute Agency policy.
72 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Rather than rating each study with lower confidence because of this issue, potential
confounding by other PFAS will be considered during the evidence synthesis phase. This may
include looking across studies in populations with different exposure profiles (e.g., observing an
association in a population with much higher exposure to one PFAS due to proximity to an
industrial plant would increase confidence for that PFAS], In situations where there is considerable
uncertainty regarding the impact of residual confounding across PFAS, this will be captured as a
factor that decreases evidence strength (see Section 10).
6.3. EXPERIMENTAL ANIMAL STUDY EVALUATION
Using the principles described in Section 6.1, the evaluation of animal studies of health
effects to assess risk of bias and sensitivity will be conducted for the following domains: reporting
quality, risk of bias (selection or performance bias, confounding/variable control, and reporting or
attrition bias), study sensitivity (exposure methods sensitivity, and outcome measures and results
display) (see Table 13). Several additional considerations specific to assessing these five PFAS are
outlined in Section 6.3.1.
The rationale for judgments will be documented clearly and consistently at the outcome
level. In addition, similar to the evaluation of epidemiology studies, for domains other than
reporting quality the evaluation documentation in HAWC will include the identified limitations and
consider their impact on the overall confidence level. This will, to the extent possible, reflect an
interpretation of the potential influence on the outcome-specific results (including the direction
and/or magnitude of influence).
This document is a draft for review purposes only and does not constitute Agency policy.
73 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 13. Considerations to evaluate domains from animal toxicology studies
Evaluation
concern
Domain—core question
Prompting questions
General considerations
Reporting quality
Reporting quality
Does the study report
information for evaluating
the design and conduct of
the study for the
endpoints/outcomes of
interest?
Note:
This domain is limited to
reporting. Other aspects of
the exposure methods,
experimental design, and
endpoint evaluation
methods are evaluated
using the domains related
to risk of bias and study
sensitivity.
Does the study report the
following?
Critical information necessarv
to perform study evaluation:
• Species, test article name,
levels and duration of
exposure, route (e.g., oral;
inhalation), qualitative or
quantitative results for at
least one endpoint of
interest
Important information for
evaluating the study methods:
• Test animal: strain, sex,
source, and general
husbandry procedures
• Exposure methods: source,
purity, method of
administration
• Experimental design:
frequency of exposure,
animal age and lifestage
during exposure and at
endpoint/outcome
evaluation
• Endpoint evaluation
methods: assays or
procedures used to measure
the endpoints/outcomes of
interest
A judgment and rationale for this domain
will generally be given for the study. In the
rationale, reviewers will also indicate when
a study adhered to GLP, or to OECD (or
similar) testing guidelines.
• Good: All critical and important
information is reported or inferable for
the endpoints/outcomes of interest.
• Adequate: All critical information is
reported, but some important
information is missing. However, the
missing information is not expected to
significantly impact the study
evaluation.
• Deficient: All critical information is
reported, but important information is
missing that is expected to significantly
reduce the ability to evaluate the study.
• Critically deficient: Study report is
missing anv pieces of critical
information. Studies that are Critically
Deficient for reporting are
Uninformative for the overall rating and
not considered further.
This document is a draft for review purposes only and does not constitute Agency policy.
74 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Evaluation
concern
Domain—core question
Prompting questions
General considerations
Risk of bias
Selection and performance bias
Allocation
Were animals assigned to
experimental groups using a
method that minimizes
selection bias?
For each study:
• Did each animal or litter
have an equal chance of
being assigned to any
experimental group
(i.e., random allocation3)?
• Is the allocation method
described?
• Aside from randomization,
were any steps taken to
balance variables across
experimental groups during
allocation?
A judgment and rationale for this domain
will be given for each cohort or experiment
in the study.
• Good: Experimental groups were
randomized, and any specific
randomization procedure was
described or inferable
(e.g., computer-generated scheme).
(Note that normalization is not the
same as randomization [see response
for 'Adequate'].)
• Adequate: Authors report that groups
were randomized but do not describe
the specific procedure used
(e.g., "animals were randomized").
Alternatively, authors used a
nonrandom method to control for
important modifying factors (i.e., with
respect to the outcome of interest)
across experimental groups
(e.g., body-weight normalization).
• Not reported (interpreted as Deficient):
No indication of randomization of
groups or other methods
(e.g., normalization) to control for
important modifying factors across
experimental groups.
• Critically deficient: Bias in the animal
allocations was reported or inferable.
This document is a draft for review purposes only and does not constitute Agency policy.
75 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Evaluation
concern
Domain—core question
Prompting questions
General considerations
Observational bias/blinding
For each endpoint/outcome or
A judgment and rationale for this domain
Did the study implement
grouping of outcomes in a study:
will be given for each endpoint/outcome or
measures to reduce
observational bias?
• Does the study report
blinding or other
methods/procedures for
reducing observational bias,
as appropriate for the assays
of interest?
group of outcomes investigated in the study.
• Good: Measures to reduce
observational bias were described
(e.g., blinding to conceal treatment
groups during endpoint evaluation;
consensus-based evaluations of
• If not, did the study use a
histopathology lesions3).
0)
3
design or approach for
• Adequate: Methods for reducing
C
-l->
which such procedures can
observational bias (e.g., blinding) can
C
o
be inferred?
be inferred or were reported but
"D
0)
3
o
V)
CO
• What is the expected impact
described incompletely.
C
'+¦»
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Evaluation
concern
Domain—core question
Prompting questions
General considerations
Confounding
For each study:
A judgment and rationale for this domain
Are variables with the
• Are there differences across
the treatment groups
(e.g., co-exposures, vehicle,
diet, palatability, husbandry,
health status, surgery) that
could bias the results?
will be given for each cohort or experiment
potential to confound or
modify results controlled
for and consistent across all
experimental groups?
in the study, noting when the potential for
confounding is restricted to specific
end points/outcomes.
• Good: Outside of the exposure of
interest, variables that are likely to
• If differences are identified,
to what extent are they
expected to impact the
confound or modify results appear to
"D
0)
o
*_
¦*->
c
o
be controlled for and consistent across
experimental groups.
3
C
DO
across groups, but these are expected
4—
c
to have a minimal impact on the
o
yi
c
3
o
results.
cc
M-
c
o
u
• Deficient: Notable concern that
potentially confounding variables were
uncontrolled or inconsistent across
groups and that they are expected to
substantially impact the results.
• Critically deficient: Confounding
variables were presumed to be
uncontrolled or inconsistent across
groups, and they are expected to be a
primary driver of the results.
This document is a draft for review purposes only and does not constitute Agency policy.
77 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Evaluation
concern
Domain—core question
Prompting questions
General considerations
Risk of bias (continued)
Selective reporting and attrition bias
Selective reporting and
attrition
Did the study report results
for all prespecified
outcomes and tested
animals?
Note:
This domain does not
consider the
appropriateness of the
com par is ons/res ults
presentation. This aspect of
study quality is evaluated in
another domain.
For each study:
Selective reporting bias:
• Are all results presented for
end points/outcomes
described in the methods
(see note)?
Attrition bias:
• Do the results account for all
animals?
• If there are discrepancies, do
the authors provide an
explanation (e.g., death or
unscheduled sacrifice during
the study)?
• If unexplained results
omissions and/or attrition
are identified, what is the
expected impact on the
interpretation of the results?
A judgment and rationale for this domain
will be given for each cohort or experiment
in the study.
• Good: Quantitative or qualitative
results were reported for all
prespecified outcomes (explicitly stated
or inferred), exposure groups and
evaluation time points. Data not
reported in the primary article are
available from supplemental material.
If results omissions or animal attrition
are identified, the authors provide an
appropriate explanation, and the
omissions or attrition are not expected
to impact the interpretation of the
results.
• Adequate: Quantitative or qualitative
results are reported for most
prespecified outcomes (explicitly stated
or inferred), exposure groups, and
evaluation time points. Omissions
and/or attrition are not explained, but
they are not expected to significantly
impact the interpretation of the results.
• Deficient: Quantitative or qualitative
results are missing for many
prespecified outcomes (explicitly stated
or inferred), exposure groups and
evaluation time points, or there is high
animal attrition; omissions and/or
attrition are not explained and are
expected to significantly impact the
interpretation of the results.
• Critically deficient: Extensive results
omission and/or animal attrition are
identified and prevent comparisons of
results across treatment groups.
This document is a draft for review purposes only and does not constitute Agency policy.
78 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Evaluation
concern
Domain—core question
Prompting questions
General considerations
>
">
a>
to
>
>
¦a
o
-C
*
E
ai
o
a.
x
Chemical administration
and characterization
Did the study adequately
characterize exposure to
the chemical of interest and
the exposure administration
methods?
Note:
These considerations are
limited to oral exposure, as
only a single inhalation
study focusing on acute
toxicity (i.e., after PFNA
exposure) was identified
(see Section 2.3.2).
For each study:
• Does the study report the
source and purity and/or
composition (e.g., identity
and percent distribution of
different isomers) of the
chemical? If not, can the
purity and/or composition
be obtained from the
supplier (e.g., as reported on
the website)?
• Was independent analytical
verification of the test article
purity and composition
performed?
• Are there concerns about
the methods used to
administer the chemical
(e.g., gavage volume)?
• If necessary, based on
consideration of
chemical-specific knowledge
(e.g., instability in solution;
volatility) and/or exposure
design (e.g., the frequency
and duration of exposure),
were the chemical
concentrations in the dosing
solutions or diet analytically
confirmed?
A judgment and rationale for this domain
will be given for each cohort or experiment
in the study.
• Good: Chemical administration and
characterization is complete
(i.e., source, purity, and analytical
verification of the test article are
provided). There are no concerns
about the composition, stability, or
purity of the administered chemical, or
the specific methods of administration.
• Adequate: Some uncertainties in the
chemical administration and
characterization are identified, but
these are expected to have minimal
impact on interpreting the results
(e.g., source and vendor-reported
purity are presented, but not
independently verified; purity of the
test article is suboptimal but not
concerning).
• Deficient: Uncertainties in the exposure
characterization are identified and
expected to substantially impact the
results (e.g., source of the test article is
not reported; levels of impurities are
substantial or concerning; deficient
administration methods, such as use of
a gavage volume considered too large
for the species and/or lifestage at
exposure).
• Critically deficient: Uncertainties in the
exposure characterization are
identified, and there is reasonable
certainty that the results are largely
attributable to factors other than
exposure to the chemical of interest
(e.g., identified impurities are expected
to be a primary driver of the results).
This document is a draft for review purposes only and does not constitute Agency policy.
79 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Evaluation
concern
Domain—core question
Prompting questions
General considerations
¦a
a>
o
u
>
">
a>
1/1
¦a
ai
o
¦M
>
-a
o
.c
tu
E
a>
o
a.
x
Exposu retiming,
frequency, and duration
Was the timing, frequency,
and duration of exposure
sensitive for the
endpoint(s)/outcome(s) of
interest?
For each endpoint/outcome or
grouping of outcomes in a study:
• Does the exposure period
include the full critical
window of sensitivity, based
on current biological
understanding?
• Was the duration and
frequency of exposure
sensitive for detecting the
endpoint of interest?
A judgment and rationale for this domain
will be given for each endpoint/outcome or
group of outcomes investigated in the study.
• Good: The duration and frequency of
the exposure was sensitive, and the
exposure included the critical window
of sensitivity (if known).
• Adequate: The duration and frequency
of the exposure was sensitive, and the
exposure covered most of the critical
window of sensitivity (if known).
• Deficient: The duration and/or
frequency of the exposure is not
sensitive and did not include most of
the critical window of sensitivity (if
known). These limitations are expected
to bias the results towards the null.
• Critically deficient: The exposure
design was not sensitive and is
expected to strongly bias the results
towards the null. The rationale should
indicate the specific concern(s).
This document is a draft for review purposes only and does not constitute Agency policy.
80 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Evaluation
concern
Domain—core question
Prompting questions
General considerations
-a
ai
>
'>
ai
to
>
a.
ro
ai
E
ai
E
o
Endpoint sensitivity and
specificity
Are the procedures
sensitive and specific for
evaluating the
endpoint(s)/outcome(s) of
interest?
Note:
• Sample size alone is not
a reason to conclude
an individual study is
critically deficient.
• Considerations related
to adjustments/
corrections to endpoint
measurements
(e.g., organ weight
corrected for body
weight) are addressed
under results
presentation.
For each endpoint/outcome or
grouping of outcomes in a study:
• Are there concerns
regarding the sensitivity,
specificity, and/or validity of
the outcome measurement
protocols?
• Are there serious concerns
regarding the sample size?
• Are there concerns
regarding the timing of the
endpoint assessment?
A judgment and rationale for this domain
will be given for each endpoint/outcome or
group of outcomes investigated in the study.
Examples of potential concerns include:
• Selection of protocols that are
insensitive or nonspecific for the
endpoint of interest.
• Evaluations did not include all
treatment groups (e.g., only control and
high dose).
• Use of unreliable or invalid methods to
assess the outcome.
• Assessment of endpoints at
inappropriate or insensitive ages, or
without addressing known endpoint
variation (e.g., due to circadian
rhythms, estrous cyclicity).
• Decreased specificity or sensitivity of
the response due to the timing of
endpoint evaluation, as compared with
exposure (e.g., immediate endpoint
assessment after exposure to chemicals
with short-acting depressant or irritant
effects; insensitivity due to prolonged
period of non-exposure before testing).
This document is a draft for review purposes only and does not constitute Agency policy.
81 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Evaluation
concern
Domain—core question
Prompting questions
General considerations
¦a
ai
o
u
>
'>
a>
to
¦a
ai
o
a.
ro
ai
E
ai
E
o
Results presentation
Are the results presented in
a way that makes the data
usable and transparent?
Note:
Potential issues associated
with statistical analyses will
be flagged for review by
EPA statisticians and
possible reanalysis (if
information is available to
do so, any reanalysis will be
transparently presented).
Any remaining limitations
will be discussed during
evidence synthesis or
dose-response analyses
(depending on the identified
issue).
For each endpoint/outcome or
grouping of outcomes in a study:
• Does the level of detail allow
for an informed
interpretation of the results?
• Are the data analyzed,
compared, or presented in a
way that is inappropriate or
misleading?
A judgment and rationale for this domain
will be given for each endpoint/outcome or
group of outcomes investigated in the study.
Examples of potential concerns include:
• Non-preferred presentation
(e.g., developmental toxicity data
averaged across pups in a treatment
group, when litter responses are more
appropriate; presentation of absolute
organ-weight data when relative
weights are more appropriate).
• Failing to present quantitative results
either in tables or figures.
• Pooling data when responses are
known or expected to differ
substantially (e.g., across sexes or
lifestages).
• Failing to report on or address overt
toxicity when exposure levels are
known or expected to be highly toxic.
• Lack of full presentation of the data
(e.g., presentation of mean without
variance data; concurrent control data
are not presented).
a>
¦c
a>
>
O
Overall confidence
Considering the identified
strengths and limitations,
what is the overall
confidence rating for the
endpoint(s)/outcome(s) of
interest?
Note:
Reviewers will mark studies
that are rated lower than
high confidence due only to
low sensitivity (i.e., bias
towards the null) for
additional consideration
during evidence synthesis.
If the study is otherwise well
conducted and an effect is
observed, the confidence
may be increased.
For each endpoint/outcome or
grouping of outcomes in a study:
• Were concerns
(i.e., limitations or
uncertainties) related to the
reporting quality, risk of
bias, or sensitivity
identified?
• If yes, what is their expected
impact on the overall
interpretation of the
reliability and validity of the
study results, including
(when possible)
interpretations of impacts
on the magnitude or
direction of the reported
effects?
The overall confidence rating considers the
likely impact of the noted concerns
(i.e., limitations or uncertainties) in
reporting, bias and sensitivity on the results.
A confidence rating and rationale will be
given for each endpoint/outcome or group
of outcomes investigated in the study.
Confidence rating definitions are described
above (see Section 6.1).
This document is a draft for review purposes only and does not constitute Agency policy.
82 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Evaluation
concern
Domain—core question
Prompting questions
General considerations
aSeveral studies have characterized the relevance of randomization, allocation concealment, and blind outcome
assessment in experimental studies (Hirst et al., 2014; Krauth et al., 2013; Macleod, 2013; Higgins and Green, 2011).
GLP = good laboratory practice; OECD = Organisation for Economic Co-operation and Development.
6.3.1. Animal Toxicology Study Evaluation Considerations Specific to These Five Per- and
Polyfluoroalkyl Substances (PFAS)
One of the key uncertainties in these assessments has to do with the toxicokinetics of these
PFAS. The apparent differences in toxicokinetics across animal species will not be addressed at the
individual study level but will be considered during evidence integration (see Section 10] and is
expected to be most influential when developing toxicity values for potential human health hazards
(see Section 11], However, based on Table 8 (see Section 2.4.1], the clearance of some of these
PFAS, and the sex-specific differences in serum half-lives, represent important considerations for
potential sources of insensitivity during study evaluation. Specifically, studies that fail to account
for the short serum half-lives of PFBA in female rats and mice (half-lives of ~ 1-3 hours; half-lives in
males are close to half a day and of less concern] and PFHxA in rats and mice of both sexes
(half-lives of ~0.5-1.5 hours] by including, for example, multiple daily exposures may be judged as
insensitive. Half-lives in rodents for PFDA, PFNA, and PFHxS are on the order of days or longer, and
so insensitivity due to short half-life in rodents does not represent a concern (note: no nonhuman
primate health effect studies were identified]. Similarly, given the profound apparent differences in
clearance between male and female rats for PFNA (i.e., females appear to clear PFNA 25 x faster],
studies that examined both sexes are preferred, and any study that tested female rats only may be
judged as insensitive. This consideration may also be applied, but to a lesser extent, for studies of
PFHxS in rats and PFBA in mice or rats (females appear to clear these PFAS ~4-6x faster than
males].
These five PFAS are considered stable and nonreactive, and the presence of potentially toxic
impurities within these readily available chemicals has not been identified as an issue in the
literature. Thus, failure to describe preparation and storage of dosing solutions will not be
considered an issue of concern, and a lack of information on chemical purity will not be considered
a significant limitation. However, given the more relevant possibility of contamination of these five
PFAS with other PFAS, a lack of analytical verification of the test article will be flagged as a
limitation, although this alone will not significantly affect overall study confidence ratings.
A wide variety of outcomes have been assessed in the available animal studies for these five
PFAS. Considerations specific to each outcome are not included in this protocol (outcome-specific
concerns will be available in HAWC when the assessments are released]. As examples, a few
specific considerations that will be applied include better domain ratings for studies that address
potential differences in time of day] for evaluations of hormone levels (due to fluctuations
This document is a draft for review purposes only and does not constitute Agency policy.
83 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
throughout the day] and for studies that address fasting status for metabolic-related
measurements.
6.4. PHARMACOKINETIC MODEL EVALUATION
A similar approach for evaluation will be applied to the full PBPK models for PFHxS (Kimet
al.. 20181 and for PFDA and PFNA (Kim etal.. 20191 as well as to the two-compartment PK model
for gestational and lactational transfer of PFHxS in humans described by Verner etal. (20161
Models will be preferred for use in these assessments when an applicable one exists and no equal
or better alternative for dosimetric extrapolation is available. Given these preferences, sound
justification will be provided for not using a PBPK (or classical PK] model when an applicable one
exists and no equal or better alternative for dosimetric extrapolation is available. Note, however,
that these preferences only apply to models that faithfully represent current scientific knowledge
and accurately translate the science into computational code in a reproducible, transparent
manner. In practice, it has been found that many models have errors that affect their predictions to
varying degrees; hence, an evaluation of a model is required before it can be used in an assessment.
Thus, the models identified above and any other models identified at later stages of developing
these assessments will be evaluated as described below.
Considerations for judging the suitability of a model are separated into two categories:
scientific and technical. In summary, the scientific criteria focus on whether the biology, chemistry,
and other information available for chemical mode(s] of action (M0A[s]] are appropriately
represented by the model structure and equations. Significant to the overall efficiency of this
process, the scientific criteria can be judged by reading the publication or report that describes the
model, without requiring an evaluation of the computer code. Preliminary technical criteria include
the availability of the computer code and apparent completeness of parameter listing and
documentation. The in-depth technical and scientific criteria focus on the accurate implementation
of the conceptual model in the computational code, use of correct or biologically consistent
parameters in the model, and reproducibility of model results reported in journal publications and
other documents. Specific details for this evaluation are provided in the Quality Assurance Project
Plan for PBPK models CU.S. EPA. 2018bl
6.5. MECHANISTIC STUDY EVALUATION
Sections 9 and 10 outline an approach for focused consideration of information from
mechanistic studies (including in vitro, in vivo, ex vivo, and in silico studies] where the specific
analytical approach is targeted to the assessment needs, depending in part on the extent and nature
of the phenotypic human and animal evidence. In this way, the mechanistic synthesis for a given
health effect might range from a high-level summary (or detailed analysis] of potential mechanisms
of action to specific, focused questions needed to address important and impactful assessment
uncertainties unaddressed by the available phenotypic studies (e.g., expected shape of the
This document is a draft for review purposes only and does not constitute Agency policy.
84 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
dose-response curve in the low-dose region, applicability of the animal evidence to humans,
addressing susceptible populations]. Individual study-level evaluation of mechanistic endpoints
will not typically be pursued. However, it may be necessary to identify assay-specific
considerations for study endpoint evaluations on a case-by-case basis to provide a more detailed
summary and evaluation for the most relevant individual mechanistic studies addressing a key
assessment uncertainty. This may be done, for example, when the scientific understanding of a
critical mechanistic event or MOA lacks scientific consensus, when the reported findings on a
critical mechanistic endpoint are conflicting, when the available mechanistic evidence addresses a
complex and influential aspect of the assessment, or when in vitro or in silico data make up the bulk
of the evidence base and there is little or no evidence from epidemiological studies or animal
bioassays. As noted in Section 3 and Section 4, genotoxicity studies were identified as meeting
PECO criteria; these data will be summarized in each PFAS assessment to describe evidence
relevant to carcinogenicity even in the absence of more phenotypic data. Based on the
considerations above, if these studies are interpreted as adequate to draw a hazard conclusion
(i.e., other than "insufficient"], individual study-level evaluations of some or all the genotoxicity
studies will be informative to this decision. If necessary, based on the assessment-specific issues
identified during study evaluation and evidence synthesis (see Section 9.2], the specific approach to
evaluating individual studies other than those addressed in Sections 6.2-6.4 will be outlined in the
assessments and included as an update to the protocol.
This document is a draft for review purposes only and does not constitute Agency policy.
85 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
7. ORGANIZING THE HAZARD REVIEW
The organization and scope of the hazard evaluation is determined by the available
evidence for each PFAS regarding routes of exposure, metabolism and distribution, outcomes
evaluated, and number of studies pertaining to each outcome, as well as the results of the
evaluation of sources of bias and sensitivity. The hazard evaluations will be organized around
organ systems (e.g., nervous system] informed by one or multiple related outcomes, as described in
Section 5, and a decision will be made as to what level (e.g., organ system or subsets of outcomes
within an organ system] to organize the synthesis.
Table 14 lists some questions that may be asked of the evidence to assist with this decision.
These questions extend from considerations and decisions made during development of the refined
evaluation plan to include review of the concerns raised during individual study evaluations as well
as the direction and magnitude of the study-specific results. Resolution of these questions will then
inform critical decisions about the organization of the hazard evaluation and help determine what
studies may be useful in dose-response analyses.
This document is a draft for review purposes only and does not constitute Agency policy.
86 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 14. Querying the evidence to organize syntheses for human and animal
evidence
Evidence
Questions
Follow-up questions
ADME
Given the known ADME issues for these PFAS, do the
data appear to differ by route of exposure studied,
lifestage when exposure occurred, sex, species, or
dosing regimens used?
Will separate analyses be needed by factors
such as sex, route of exposure, or by methods
of dosing within a route of exposure (e.g., are
large differences expected between gavage
and dietary exposures)?
Are data available to inform which lifestages
and what dosing regimens are more relevant
to human exposure scenarios?
Is there toxicity information for metabolites that also
should be evaluated for hazard?
What exposures will be included in the
evaluation?
Outcomes
What outcomes are reported in studies? Are the
data reported in a comparable manner across studies
(similar output metrics at similar levels of specificity,
such as adenomas and carcinomas quantified
separately)?
At what level (hazard, grouped outcomes, or
individual outcomes) will the synthesis be
conducted?
What commonalities will the outcomes be
grouped by:
• health effect,
• exposure levels,
• functional or population-level
consequences (e.g., endpoints all
ultimately leading to decreased
fertility or impaired cognitive
function),
• involvement of related biological
pathways
How well do the assessed human and animal
outcomes relate within a level of grouping?
Are there inter-related outcomes? If so, consider
whether some outcomes are more useful and/or of
greater concern than others.
Does the evidence indicate greater sensitivity to
effects (at lower exposure levels or severity) in
certain subgroups (by age, sex, ethnicity, lifestage)?
Should the hazard evaluation include a subgroup
analysis?
Does incidence or severity of an outcome increase
with duration of exposure or a particular window of
exposure. What exposure time frames are relevant
to development or progression of the outcome?
Is there mechanistic evidence that informs how
outcomes might be grouped together?
How robust is the evidence for specific outcomes?
• What outcomes are reported by both human
and animal studies and by one or the other?
Were different animal species and sexes (or
other important population-level
differences) tested?
• In general, what are the study confidence
conclusions of the studies (high, medium,
low, uninformative) for the different
outcomes? Is there enough evidence from
high- and medium-confidence studies to
draw conclusions about causality?
What outcomes should be highlighted?
Should the others be synthesized at all?
Would comparisons by specific limitations be
informative?
This document is a draft for review purposes only and does not constitute Agency policy.
87 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Evidence
Questions
Follow-up questions
Dose-
response
Did some outcomes include better coverage of
exposure ranges that may be most relevant to human
exposure than others?
What outcomes and studies are informative
for developing toxicity values?
For which outcomes are there sufficient data
available to draw conclusions about dose-response?
Are there outcomes with study results of sufficient
similarity (e.g., an established linkage in a biological
pathway) to allow examination or calculation of
common measures of effect across studies? Do the
mechanistic data identify surrogate or precursor
outcomes that are adequate for dose-response
analysis?
Are there subgroups that exhibit responses at lower
exposure levels than others?
Are there findings from ADME studies that could
inform data-derived extrapolation factors, or link
toxicity observed via different routes of exposure, or
link effects between humans and experimental
animals?
What studies might be used to develop non-
default UFs? Is there a common internal dose
metric that can be used to compare species
or routes of exposure?
ADME = absorption, distribution, metabolism, and excretion; UF = uncertainty factor.
1
This document is a draft for review purposes only and does not constitute Agency policy.
88 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
8. DATA EXTRACTION OF STUDY METHODS AND
RESULTS
Data extraction and content management will be carried out using the Health Assessment
Workplace Collaborative (HAWC; web links will be shared in the individual assessments], A
consistent approach to data extraction will be applied across these PFAS assessments to facilitate
their anticipated future use in addressing poorly studied PFAS (e.g., through coupling with
computational toxicology data generated as described in the Environmental Protection Agency
[EPA] PFAS action plan]. Data extraction elements that may be collected from epidemiological,
controlled human exposure, animal toxicological, and in vitro studies are described in HAWC
f https: //hawcprd.epa.gOv/about/l Not all studies that meet the PECO criteria go through data
extraction. Studies evaluated as being Uninformative are not considered further and therefore
would not undergo data extraction. In addition, outcomes determined to be less relevant during
PECO refinement (see Section 5] may not go through data extraction or may have only minimal data
extraction. The same may be true for /ow-confidence studies if enough medium- and
/?zg/?-confidence studies (e.g., on an outcome] are available. All findings are considered for
extraction, regardless of statistical significance. The level of extraction for specific outcomes within
a study may differ (i.e., ranging from a narrative to full extraction of dose-response effect size
information]. In part, this is determined based on the level of detail to be discussed in the evidence
synthesis for that health effect (e.g., a detailed extraction will not be necessary for health effects
with veiy few available studies; these will only be briefly summarized in a short narrative].
Similarly, decisions about data extraction for /ow-confidence studies are typically made while
implementing the protocol and are based on the quality and extent of the available evidence. If
necessary, the version of the protocol released with the draft assessment will outline how
low-confidence studies were treated for extraction and evidence synthesis.
The data extraction results for included studies will be presented in the assessment (and
made available for download from EPA HAWC in Excel format] when the draft is publicly released.
(Note: The following browsers are supported for accessing HAWC: Google Chrome [preferred],
Mozilla Firefox, and Apple Safari. There are errors in functionality when viewed with Internet
Explorer.] For quality control, data extraction will be performed by one member of the evaluation
team and independently verified by at least one other member. Discrepancies in data extraction
will be resolved by discussion or consultation with a third member of the evaluation team. Digital
rulers, such as WebPlotDigitizer fhttps://automeris.io/WebPlotDigitizer/I. will be used to extract
numerical information from figures, and their use will be documented during extraction.
This document is a draft for review purposes only and does not constitute Agency policy.
89 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
As previously described, routine attempts will be made to obtain missing information from
human and animal health effect studies, if it is considered influential during study evaluations (see
Section 6) or when it can provide information important for dose-response analysis or
interpretations of significance (e.g., missing group size or variance descriptors such as standard
deviation or confidence interval]. Missing data from individual mechanistic (e.g., in vitro] studies
generally will not be sought. Outreach to study authors or designated contact persons will be
documented and considered unsuccessful if researchers do not respond to email or phone requests
within 1 month of initial attempt(s) to contact.
8.1. STANDARDIZING REPORTING OF EFFECT SIZES
In addition to providing quantitative outcomes in their original units for all study groups,
results from outcome measures will be transformed, when possible, to a common metric to help
compare distinct but related outcomes that are measured with different scales. These standardized
effect size estimates facilitate systematic evaluation and evidence integration for hazard
identification (see Section 9.1). The following summary of effect size metrics by data type outlines
issues in selecting the most appropriate common metric for a collection of related endpoints
(Vesterinen et al„ 20141
Common metrics for continuous outcomes include:
• Absolute difference in means. This metric is the difference between the means in the control
and treatment groups, expressed in the units in which the outcome is measured. When the
outcome measure and its scale are the same across all studies, this approach is the simplest
to implement.
• Percent control response (or normalized mean difference [NMD]). Percent control group
calculations are based on means. Standard deviation (or standard error) values presented
in the studies for these normalized effect sizes can also be estimated if sufficient
information has been provided. Note that some outcomes reported as percentages, such as
mean percentage of affected offspring per litter, can lead to distorted effect sizes when
further characterized as percentage change from control. Such measures are better
expressed as absolute difference in means, or rather preferably transformed to incidences
using approaches for event or incidence data (see below).
• Standardized mean difference. The NMD approach above is relevant to ratio scales, but
sometimes it is not possible to infer what a "normal" animal would score, such as when data
for animals without lesions are not available. In these circumstances, standardized mean
differences can be used. The difference in group means is divided by a measure of the
pooled variance to convert all outcome measures to a standardized scale with units of
standard deviations. This approach can also be applied to data for which different
measurement scales are reported for the same outcome measure (e.g., different measures of
lesion size such as infarct volume and infarct area).
Common metrics for event or incidence data include:
This document is a draft for review purposes only and does not constitute Agency policy.
90 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
• Percent change from control. This metric is analogous to the NMD approach described for
continuous data above.
• For binary outcomes such as the number of individuals that developed a disease or died,
and with only one treatment evaluated, data can be represented in a 2 x 2 table. Note that
when the value in any cell is 0, 0.5 is added to each cell to avoid problems with the
computation of the standard error. For each comparison, the odds ratio (OR] and its
standard error can be calculated. Odds ratios are normally combined on a logarithmic scale.
An additional approach for epidemiology studies is to extract adjusted statistical estimates
when possible rather than unadjusted or raw estimates.
It is important to consider the variability associated with effect size estimates, with better
powered studies generally showing more precise estimates. Effect size estimation can be affected,
however, by such factors as variances that differ substantially across treatment groups, or by lack of
information to characterize variance, especially for animal studies in biomedical research
fVesterinen et al.. 20141 The assessments will consider the nature of any variance issues and
ensure that the associated uncertainties are clarified and accounted for during the evidence
synthesis process (see Section 9],
8.2. STANDARDIZING ADMINISTERED DOSE LEVELS/CONCENTRATIONS
Exposures will be standardized to common units. Exposure levels in oral studies will be
expressed in units of mg PFAS/kg-day. Where study authors provide exposure levels in
concentrations in the diet or drinking water, dose conversions will be made using study-specific
food or water consumption rates and body weights when available. Otherwise, EPA defaults will be
used (U.S. EPA. 19881 addressing age and study duration as relevant for the species/strain and sex
of the animal of interest Exposure levels in inhalation studies will be expressed in units of mg/m3.
Assumptions used in performing dose conversions will be documented in HAWC or the specific
assessments.
This document is a draft for review purposes only and does not constitute Agency policy.
91 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
9. SYNTHESIS OF EVIDENCE
For the purposes of this assessment, evidence synthesis and integration are considered
distinct, but related processes. The syntheses of separate bodies of evidence (i.e., human, animal,
and mechanistic evidence] described in this section will directly inform the integration across all
evidence to draw an overall judgment for each of the assessed human health effects (as described in
Section 10). The phrase "evidence integration" used here is analogous to the phrase "weight of
evidence" used in some other assessment processes (EFSA. 2017: U.S. EPA. 2017a: NRC. 2014: U.S.
EPA. 2005al12
For each potential human health effect (or smaller subset of related outcomes), the U.S.
Environmental Protection Agency (EPA) will separately synthesize the available phenotypic human
and animal evidence pertaining to that potential health effect Mechanistic evidence will also be
considered in targeted analyses conducted prior to, during and after developing syntheses of the
phenotypic human and animal evidence. The results of the analyses of mechanistic evidence will be
used to inform key uncertainties; as a result, the scope of the mechanistic analyses will generally
depend on the extent and nature of the human and animal evidence (see Sections 9.2 and 10).
Thus, apart from the pre-defined mechanistic analyses (see Sections 9.2.1-9.2.3), the human and
animal evidence syntheses (or the lack of phenotypic data in humans and animals) help determine
the approach to be taken in synthesizing the available mechanistic evidence (see Section 9.2.4). In
this way, a mechanistic evidence synthesis might range from a high-level summaiy of potential
toxicity mechanisms discussed in the published literature to a detailed analysis of multiple
potential modes-of-action, or it might evaluate specific, focused questions that inform key
uncertainties unaddressed by the phenotypic human and animal evidence (e.g., shape of the
dose-response curve at low doses, applicability of the animal evidence to humans, addressing
susceptible populations). Each synthesis will provide a summary discussion of the available
evidence that addresses considerations regarding causation (see Table 15). These considerations
are adapted from considerations for causality introduced by Austin Bradford Hill ("Hill.
19651: consistency, dose-response relationship, strength of the association, temporal relationship,
biological plausibility, coherence, and "natural experiments" in humans [see additional discussion
12This revision has been adopted primarily based on the 2014 NAS review of IRIS (NRC, 2014): "The present
committee found that the phrase weight of evidence has become far too vague as used in practice today and thus
is of little scientific use. In some accounts, it is characterized as an oversimplified balance scale on which evidence
supporting hazard is placed on one side and evidence refuting hazard on the other... The present committee found
the phrase evidence integration to be more useful and more descriptive of what is done at this point in an IRIS
assessment—that is, IRIS assessments must come to a judgment about whether a chemical is hazardous to human
health and must do so by integrating a variety of evidence."
This document is a draft for review purposes only and does not constitute Agency policy.
92 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
in U.S. EPA f2005al and U.S. EPA f19941]. Importantly, the evidence synthesis process explicitly
considers and incorporates the conclusions from the individual study evaluations (see Section 6],
This document is a draft for review purposes only and does not constitute Agency policy.
93 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 15. Information most relevant to describing primary considerations for
assessing causality during evidence syntheses
Consideration
Description of the consideration and its application in IRIS syntheses
Study confidence
Description: Incorporates decisions about studv confidence within each of the
considerations.
Application: In evaluating the evidence for each of the causalitv considerations described
in the following rows, the syntheses will consider study confidence decisions.
H/g/?-confidence studies carry the most weight. The syntheses will consider the specific
limitations and strengths identified during study evaluation and describe how these
informed each consideration.
Consistency
Description: Examines the similarity of results (e.g., direction; magnitude) across studies.
Application: Syntheses will evaluate the homogeneity of findings on a given outcome or
endpoint across studies. When inconsistencies exist, the syntheses consider whether
results were "conflicting" (i.e., unexplained positive and negative results in similarly
exposed human populations or in similar animal models) or "differing" (i.e., mixed results
explainable by, for example, differences between human populations, animal models,
exposure conditions, or studv methods) (U.S. EPA, 2005a). These considerations are
based on analyses of potentially important explanatory factors such as:
• Confidence in studies' results, including study sensitivity (e.g., some study
results that appear to be inconsistent may be explained by potential biases or
other attributes that affect sensitivity).
• Exposure, including route (if applicable) and administration methods, levels,
duration, timing with respect to outcome development (e.g., critical windows),
and exposure assessment methods (i.e., in epidemiology studies), including
analytical units and specific groups being compared.
• Specificity and sensitivity of the endpoint for evaluating the health effect in
question (e.g., functional measures can be more sensitive than organ weights).
• Populations or species, including consideration of potential susceptible groups
or differences across lifestage at exposure or endpoint assessment.
• Toxicokinetic information explaining observed differences in responses across
route of exposure, other aspects of exposure, species, sexes, or lifestages.
The interpretation of consistency will emphasize biological significance, to the extent
that it is understood, over statistical significance. Statistical significance from suitably
applied tests (this may involve consultation with an EPA statistician) adds weight when
biological significance is not well understood. Consistency in the direction of results
increases confidence in that association even in the absence of statistical significance. In
some cases, it may be helpful to consider the potential for publication bias to provide
context to interpretations of consistency.3
This document is a draft for review purposes only and does not constitute Agency policy.
94 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Consideration
Description of the consideration and its application in IRIS syntheses
Strength (effect
magnitude) and
precision
Description: Examines the effect magnitude or relative risk, based on what is known
about the assessed endpoint(s), and considers the precision of the reported results
based on analyses of variability (e.g., confidence intervals; standard error). This may
include consideration of the rarity or severity of the outcomes.
Application: Syntheses will analvze results both within and across studies and mav
consider the utility of combined analyses (e.g., meta-analysis). While larger effect
magnitudes and precision (e.g., p < 0.05) help reduce concerns about chance, bias, or
other factors as explanatory, syntheses should also consider the biological or
population-level significance of small effect sizes.
Biological gradient/
dose-response
Description: Examines whether the results (e.g., response magnitude; incidence;
severity) change in a manner consistent with changes in exposure (e.g., level; duration),
including consideration of changes in response after cessation of exposure.
Application: Syntheses will consider relationships both within and across studies,
acknowledging that the dose-response relationship (e.g., shape) can vary depending on
other aspects of the experiment, including the biology underlying the outcome and the
toxicokinetics of the chemical. Thus, when dose-dependence is lacking or unclear, the
synthesis will also consider the potential influence of such factors on the response
pattern.
Coherence
Description: Examines the extent to which findings are cohesive across different
endpoints that are related to, or dependent on, one another (e.g., based on known
biology of the organ system or disease, or mechanistic understanding such as
toxicokinetic/dynamic understanding of the chemical or related chemicals). In some
instances, additional analyses of mechanistic evidence from research on the chemical
under review or related chemicals that evaluate linkages between endpoints or
organ-specific effects may be needed to interpret the evidence. These analyses may
require additional literature search strategies.
Application: Syntheses will consider potentially related findings, both within and across
studies, particularly when relationships are observed within a cohort or within a
narrowly defined category (e.g., occupation; strain or sex; lifestage of exposure).
Syntheses will emphasize evidence indicative of a progression of effects, such as
temporal- or dose-dependent increases in the severity of the type of endpoint observed.
If an expected coherence between findings is not observed, possible explanations should
be explored including those related to the biology of the effects as well as the sensitivity
and specificity of the measures used.
This document is a draft for review purposes only and does not constitute Agency policy.
95 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Consideration
Description of the consideration and its application in IRIS syntheses
Mechanistic evidence
related to biological
plausibility
Description: There are multiple uses for mechanistic information, and this consideration
overlaps with "coherence." This consideration examines the biological support (or lack
thereof) for findings from the human and animal health effect studies and becomes
more influential on the hazard conclusions when notable uncertainties in the strength of
those sets of studies exist. These analyses can also improve understanding of dose- or
duration-related development of the health effect. In the absence of human or animal
evidence of apical health endpoints, the synthesis of mechanistic information may drive
evidence integration conclusions (when such information is available).
Application: Syntheses can evaluate evidence on precursors, biomarkers, or other
molecular or cellular changes related to the health effect(s) of interest to describe the
likelihood that the observed effects result from exposure. This evaluation will entail an
analysis of existing evidence, and not simply whether a theoretical pathway can be
postulated. This analysis may not be limited to evidence relevant to the PECO but may
also include evaluations of biological pathways (e.g., for the health effect; established for
other, possibly related, chemicals). Any such synthesis of mechanistic evidence will
consider the sensitivity of the mechanistic changes and the potential contribution of
alternative or previously unidentified mechanisms of toxicity.
Natural experiments
Description: Specific to epidemiology studies and rarelv available, this consideration
examines effects in populations that have experienced well-described, pronounced
changes in chemical exposure (e.g., lead exposures before and after banning lead in
gasoline).
Application: Compared to other observational designs, natural experiments have the
benefit of dividing people into exposed and unexposed groups without them influencing
their own exposure status. During synthesis, associations in medium- and
/?/g/?-confidence natural experiments can substantially reduce concerns about residual
confounding.
Publication bias involves the influence of the direction, magnitude, or statistical significance of the results on the
likelihood of a paper being published; it can result from decisions made, consciously or unconsciously, by study
authors, journal reviewers, and journal editors (Dickersin, 1990). When evidence of publication bias is present
for a set of studies, less weight may be placed on the consistency of the findings for or against an effect during
evidence synthesis and integration.
PECO = populations, exposures, comparators, and outcomes.
1
2 Data permitting, the syntheses will also discuss analyses relating to potential susceptible
3 populations.13 These analyses will be based on knowledge about the health outcome or organ
4 system affected, demographics, genetic variability, lifestage, health status, behaviors or practices,
5 and social determinants (see Table 16). This information will be used to draw conclusions
13Various terms have been used to characterize populations that may be at increased risk of developing health
effects from exposure to environmental chemicals, including "susceptible," "vulnerable," and "sensitive."
Furthermore, these terms have been inconsistently defined across the scientific literature. The term susceptibility
is used in this protocol to describe populations or lifestages at increased risk, focusing on intrinsic biological factors
that can modify the effect of a specific exposure, but also considering social determinants or behaviors that may
increase susceptibility. However, factors resulting in higher exposures to specific groups (e.g., proximity, housing,
occupation) will typically not be analyzed to describe increased risk among specific populations or subgroups.
This document is a draft for review purposes only and does not constitute Agency policy.
96 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
regarding potential susceptibility among specific populations or subgroups in a separate section
(see Section 10.3], This summary will describe concerns across the available evidence for all
potential human health effects and will inform both hazard identification and dose-response
analyses.
Table 16. Individual and social factors that may increase susceptibility to
exposure-related health effects
Factor
Examples
Demographic
Gender, age, race/ethnicity, education, income, occupation, geography
Genetic variability
Polymorphisms in genes regulating cell cycle, DNA repair, cell division,
cell signaling, cell structure, gene expression, apoptosis, and metabolism
Lifestage
In utero, childhood, puberty, pregnancy, women of childbearing age, old
age
Health status
Pre-existing conditions or disease such as psychosocial stress, elevated
body mass index, frailty, nutritional status, chronic disease
Behaviors or practices
Diet, mouthing, smoking, alcohol consumption, pica, subsistence or
recreational hunting and fishing
Social determinants
Income, socioeconomic status, neighborhood factors, health care
access, and social, economic, and political inequality
EPA ExpoBox Exposure Assessment Tools, based on EPA's Guidelines for Exposure Assessment (U.S. EPA, 1992).
DNA = deoxyribonucleic acid.
9.1. SYNTHESES OF HUMAN AND ANIMAL HEALTH EFFECTS EVIDENCE
The syntheses of the human and animal health effects evidence will focus on describing
aspects of the evidence that best inform causal interpretations, including the exposure context
examined in the sets of studies. Each evidence synthesis will be based primarily on studies of high
and medium confidence. Low-confidence studies may be used if few or no studies with higher
confidence are available to help evaluate consistency, or if the study designs of the /ow-confidence
studies address notable uncertainties in the set of high- or medium-confidence studies on a given
health effect. If low-confidence studies are used, then a careful examination of risk bias and
sensitivity with potential impacts on the evidence synthesis conclusions will be included in the
narrative.
As previously described, these syntheses will articulate the strengths and the weaknesses of
the available evidence organized around the considerations described in Table 15, as well as issues
that stem from the evaluation of individual studies (e.g., concerns about bias or sensitivity]. If
possible, results across studies will be compared using graphs and charts or other data
visualization strategies. The analysis will typically include examination of results stratified by any
This document is a draft for review purposes only and does not constitute Agency policy.
97 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
or all of the following: study confidence classification (or specific issues within confidence
evaluation domains], population or species, exposures (e.g., level, patterns [intermittent or
continuous], duration, intensity], sensitivity (e.g., low vs. high], and other factors that may have
been identified during study evaluation or analyses of key science issues (see Section 2.4], The
number of studies and the differences encompassed by the studies will determine the extent to
which specific factors can be examined for use in stratifying study results. Additionally, for both the
human and animal evidence syntheses, if supported by the available data, additional analyses
across studies (such as meta-analysis] may also be conducted.
9.2. MECHANISTIC INFORMATION
The synthesis of mechanistic information informs the integration of health effects evidence
for both hazard identification (i.e., biological plausibility or coherence of the available human or
animal evidence; inferences regarding human relevance, or the identification of susceptible
populations and lifestages across the human and animal evidence] and dose-response evaluation.
As introduced in prior sections, several key science issues that are essential to consider in these five
assessments will involve a focused analysis and synthesis of mechanistic information (see
Sections 9.2.1-9.2.3], Other potential assessment-specific uncertainties for which mechanistic
analyses might be conducted, and the considerations for including those analyses in an assessment,
are outlined in Section 9.2.4. Deviations from the approaches described in Sections 9.2.1-9.2.3, as
well as the specific methods for any analyses conducted based on the considerations described in
Section 9.2.4, will be tracked as updates to the protocol.
Mechanistic evidence includes any experimental measurement related to a health outcome
that provides information about the biological or chemical events associated with phenotypic
effects; these measurements can improve understanding of the mechanisms involved in the toxic
effects following exposure to a chemical but are not generally considered adverse outcomes.
Mechanistic data are reported in a diverse array of observational and experimental studies across
species, model systems, and exposure paradigms, including in vitro, in vivo (by various routes of
exposure], ex vivo, and in silico studies, and across a wide spectrum of diverse endpoints.
Evaluations of mechanistic information typically differ from evaluations of phenotypic
evidence (e.g., from routine toxicology studies]. This is primarily because mechanistic data
evaluations consider the support for and involvement of specific events or sets of events within the
context of a broader research question (e.g., support for a hypothesized mechanism; consistency
with known biological processes], rather than evaluations of individual apical endpoints considered
in relative isolation. Such analyses are complicated because a chemical may operate through
multiple mechanistic pathways, even if one hypothesis dominates the literature (U.S. EPA. 2005al
Similarly, multiple mechanistic pathways might interact to cause an adverse effect. Thus, pragmatic
and stepwise approaches to considering and reviewing this evidence for these PFAS assessments
are outlined below. The format of these syntheses is expected to vary from a short narrative
This document is a draft for review purposes only and does not constitute Agency policy.
98 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
summary of existing knowledge to an in-depth analysis and weighing of the evidence underlying
multiple mechanistic events, depending on data availability and the criticality of the
assessment-specific uncertainty (ies].
9.2.1. Toxicokinetic Information and Pharmacokinetic (PK)/Physiologically Based
Pharmacokinetic (PBPK) Models
One key mechanistic issue has to do with the toxicokinetics of these chemicals, particularly
their serum half-life values because these values are useful for extrapolating doses from exposed
animals to humans. Toxicokinetic studies were extracted for consideration (from the broad PFAS
literature searches] by subject matter experts using two different methods: (1] tagging of studies
during literature screening (see Sections 4.2-4.3], noting that this tagging was not conducted by
ADME subject matter experts, and (2] use of SWIFT Review software
[https: //www.sciome.com/swift-review/: (Howard etal.. 20161] to categorize the literature via
health outcome tags for ADME from the title and abstract. For identification of AD ME-related
studies to be reviewed using SWIFT Active Screener
fhttps: //www.sciome.com/swift-activescreener /l the results using the health outcome tags for
ADME embedded within SWIFT Review were confirmed using a search string developed by experts
in toxicokinetics within the IRIS program.14 This resulted in 813 potentially relevant studies that
were imported into SWIFT Active Screener for review by two independent reviewers with
demonstrated expertise in ADME (conflicts were resolved through discussion], A basic set of PECO
criteria were used for this review:
• Population: in vivo studies in humans, nonhuman primates, rats, or mice. (Note: in vitro
studies in these species were tagged as potentially supportive; see explanation below.]
• Exposure: any route of administration of a single chemical compound that is expected to
occur for human exposure for PFBA, PFHxS, PFHxA, PFDA, or PFNA. Exposure to metabolic
precursors of these chemicals was also included. (Note: intraperitoneal [i.p.] injection
studies and in vitro studies were tagged as potentially supportive; see explanation below.]
14tiab: (adme OR admet OR bile OR biliary OR bioavail* OR biodistribut* OR biologic-avail* OR biological-avail* OR
biologically-avail* OR biotrans* OR clearance OR detox* OR distribut* OR dosim* OR eliminat* OR endocytosis
OR enterohepatic OR "entero hepatic" OR excret* OR exhalation OR hepatobiliary OR inhalation OR metaboli* OR
"partition coefficient" OR permeability OR persistence OR phagocytosis OR pharmacokinetic* OR
physiologic-avail* OR physiological-avail* OR physiologically-avail* OR pinocytosis OR protein-bind* OR
reabsorption OR retention OR secretion* OR toxicokinetic* OR transport OR uptake OR urination OR ((absorb OR
absorbs OR absorbed OR absorption* OR deposition) NOT (atomic OR optical OR spectra* OR spectros* OR
spectrum* OR infrared)) OR title : ("gas exchange" AND (alveolar OR lung OR lungs OR pulmonary OR respirat*))
OR mesh_mh: ("biological transport" OR "enterohepatic circulation" OR pharmacokinetics) OR mesh_sh:
(pharmacokinetics) OR mesh_mh: (toxicokinetics)).
This string identified two fewer potentially relevant studies than the SWIFT review (including all studies identified
using the non-SWIFT string). So, the studies identified by SWIFT Review were screened in SWIFT Active Screener.
This document is a draft for review purposes only and does not constitute Agency policy.
99 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
• Comparator: vehicle control or reference population.
• Outcome: data to quantify AD ME processes, steady state analysis, empirical
pharmacokinetic (PK], full PBPK.
This screening (i.e., to 96% predicted completion based on the machine-learning software]
resulted in the identification of 99 studies relevant to toxicokinetics across the five PFAS
assessments. These data will be considered for use in the assessments as described below.
All PK and PBPK models will be formally evaluated for use in the assessments, as described
in Section 6.4. The specific approaches for determining the most appropriate method for
dosimetric extrapolation, if necessary for these assessments (note: this is likely to be necessary,
based on the preliminary literature inventory], as well as other potential quantitative approaches
for using the PK/PBPK models and ADME data, are outlined in Section 11.2.
To draw conclusions regarding the most appropriate serum half-life measures, the ADME
studies identified by the screening methods described above will be considered as outlined in U.S.
EPA (2018bl Briefly, the studies relevant to updating the data presented in Table 7 in
Section 2.4.1, including the studies underlying the current data in the table, will be reviewed, and
data that are highly unreliable will be excluded (e.g., data points below the limit of detection [LOD];
values based on uncertain exposure estimates, or other unvalidated assumptions]. Study
characteristics that will be reviewed by subject matter experts to determine whether studies are
informative to the PFAS-specific half-life values include appropriateness of the analytic method, the
number of exposure levels tested, the human relevance of the exposure range, and the number of
time points and tissues sampled. Although ADME data from in vitro studies and i.p. injection
studies were tracked as potentially relevant during screening, additional considerations will apply
to the potential incorporation of these data into the assessments, given their inherent uncertainties
(e.g., difficulties interpreting the relevance of bioavailability or peak concentration data from i.p.
injection studies]. Specifically, regarding in vitro studies, it is expected that there maybe no in vivo
toxicokinetic data on the rate of conversion of precursor compounds to the PFAS of interest, in
which case conversion rates measured in vitro can be extrapolated to in vivo as the next best means
of predicting this mechanism of exposure. Even if such extrapolation is determined to be
quantitatively uncertain, these data might still provide useful qualitative information.
In general, when there are multiple studies informative to a given ADME parameter, if the
values from two or more studies of the same species, strain, and sex are similar, a numerical
average among those values will be used. Because significant differences in the half-life between
males and females of a given species have been observed for some PFAS, these sex differences will
be assumed to be real in general across species. Specifically, when feasible, the data for males and
females of each species, for each PFAS, will be analyzed separately, even if the difference is not
statistically significant. If the values differ significantly across studies for the same
This document is a draft for review purposes only and does not constitute Agency policy.
100 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
species/strain/sex, a more detailed review of the study methods indicated above will be conducted
to determine whether one study is more likely to provide accurate information than another.
Given that PFAS inhalation exposure is expected to be via adsorption to particulates, if
sufficient data are available for any of the assessed PFAS, inhalation exposure rates for
PFAS-containing particles will be predicted using the multiple-path particle dosimetry (MPPD)
model fhttps://www.ara.com/products/multiple-path-particle-dosimetry-model-mppd-v-3041
This model predicts inhaled particle deposition in laboratory animals and humans as a function of
particle size. Particle sizes used in controlled animal studies or measured in ambient
environmental or workplace exposure studies in humans will be used as inputs. If PK data are
identified that allow the bioavailability of inhaled particulate PFAS to be estimated, the mass
deposition predicted by the MPPD model will be adjusted accordingly. Otherwise 100% absorption
of PFAS from inhalation deposition will be assumed. Note that while some inhaled particles are
later moved by the mucociliary apparatus to the larynx and swallowed, the PK bioavailability for
oral ingestion can then be applied to that fraction. Any predictions will be considered for use in
comparing findings across oral and inhalation routes of exposure during evidence integration (see
Section 10). In addition, see Section 11.2 for the application of these predictions to developing
quantitative estimates. If necessary, inhalation of PFAS in free ionic form can be estimated based
on the inhalation uptake of other chemicals with high liquid: air partition coefficients (i.e., assuming
nearly complete absorption of any free ions that contact the airway lining).
9.2.2. Peroxisome Proliferator-Activated Receptor Alpha (PPARa) Dependence for Health
Effect(s) Observed in Animals
A second area of focused mechanistic analysis is evaluating the human relevance of effects
in animals that appear to involve (at least in part) a PPARa-mediated MOA. The approach outlined
below focuses on hepatic effects, which are expected to be the primary health effect area in these
assessments for which this analysis is useful, and for which data are likely to exist to conduct an
analysis. The specifics of applying this approach may vary across the five PFAS assessments,
depending on the availability of data to address this question and the strength of the evidence
indicating PPARa involvement. During assessment development, for other health effects with
evidence that a PPARa-mediated MOA might be operant, the mechanistic syntheses will include
consideration of this issue. These analyses will depend on the amount of information available and
the strength of the evidence indicating PPARa involvement. Thus, the analyses might range from a
short summary of the available evidence when data are sparse to an evaluation approximating the
one described below when extensive data are available.
To identify the literature most relevant to addressing the question of the
PPARa-dependence of hepatic effects observed in experimental animals, a PFAS assessment with
This document is a draft for review purposes only and does not constitute Agency policy.
101 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
extensive evidence of liver effects and potential PPARa involvement will screen15 the "potentially
relevant supplemental material" studies on a given PFAS at the full-text level as follows:
• Population: in vivo animal studies in mammalian models; in vitro and human experiments
using primary or immortalized liver cell lines
• Exposure: PFAS of interest (parent compound only]
• Comparator: vehicle control
• Outcome: mechanistic outcomes relevant to the hepatobiliary system (e.g., in liver tissues or
cells]
Any additional assessment-specific strategies for identifying other information of potential
relevance on molecular mechanistic data for these five PFAS, or from the more extensive literature
on perfluorooctanoic acid (PFOA] and PFAS (e.g., as points of comparison], will be described as
updates to this protocol.
The pool of studies identified based on the strategies outlined above will be inventoried into
a database to allow for the organization and evaluation of these data. Specifically, the following
information will be extracted for each reference: a reference identifier; test compound; exposure
route and duration; the sex, species, and strain of the organism; age at exposure; and endpoint
evaluation of the test organism or test system. Additionally, the inventory(ies] will capture a
succinct description of the assessed endpoints and the potential mechanistic event(s] informed by
those endpoints in each study. The mechanistic events in the proposed mechanisms pathway for
which there are data will then be organized according to the following levels of biological
organization: molecular target(s], cellular response(s], tissue/organ response(s], and organism
response(s], in accordance with the levels of biological organization used to develop adverse
outcome pathways (AOPs].16
Although refinements based on the assessment-specific evidence are anticipated, these
assessments will first consider the use of the preliminary pathway outlined in Figure 7 as an
organizing AOP for these data. The preliminary, proposed AOP displayed in Figure 7 is based on
molecular initiating events, key events, and adverse outcomes identified in previous evaluations on
PFOS and PFOA and proposed AOPs for chemical-induced noncancer liver toxicity [see Li et al.
15Although the specifics of this screening process may vary across PFAS, this protocol describes that screening will
occur by at least two reviewers and use of DistillerSR to track decisions and resolve differences. Any deviations
from this will be tracked on an assessment-specific basis as updates to the protocol.
16Although the World Health Organization (WHO)-lnternational Programme on Chemical Safety (IPCS)-MOA and
the Organisation for Economic Co-operation and Development (OECD)-AOP frameworks are similar in the
identification and analysis of key events following modified Bradford Hill criteria (Meek et al., 2014), AOPs are
chemical agnostic whereas MOA analyses are intended to inform health assessments of individual (or groups of)
chemical(s) (Edwards et al., 2016).
This document is a draft for review purposes only and does not constitute Agency policy.
102 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
C20171 Mellor et al. C20161 Wang et al. C20141 U.S. EPA C2016dl U.S. EPA C2016dl ATSDR C20181
and NTDWOI (20171], Prior evaluations of PFOS and PFOAhave discussed studies using wild type,
PPARa knockout and humanized PPARa (hPPARa] mice showing that exposure leads to fatty acid
and triglyceride accumulation in the liver and steatosis via both PPARa-dependent
and -independent pathways (ATSDR. 2018: Li etal.. 2017: Viberg and Eriksson. 20171 In addition
to PPARa, these reviews have implicated other nuclear receptor (NR] and cell signaling pathways
with PFOA- and PFOS-induced noncancer liver effects, including PPARp/5, PPARy, constitutive
androstane receptor (CAR] and pregnane X receptor (PXR], the farnesoid X receptor (FXR], the
phosphatidylinositol 3-kinase-serine/threonine kinase Akt (PI3K-Akt] signal transduction pathway,
and the nuclear factor kappa B pathway (NF-kB] (Li etal.. 2017: Viberg and Eriksson. 20171
Activation of these pathways can be associated with alterations in lipid and glucose metabolism,
increased cellular stress, and inflammation (Mackowiak etal.. 2018: Li etal.. 2017: Mellor et al..
2016: Wang etal.. 20141 Thus, the potential involvement and contribution of these different
signaling responses to hepatic effects after exposure to the five PFAS will also be considered.
Figure 7. Preliminary proposed mechanistic pathway for per- and
polyfluoroalkyl substances (PFAS)-induced noncancer liver effects. Based on
previous reviews of perfluorooctane sulfonate (PFOS]- and perfluorooctanoic acid
(PFOA]-induced noncancer liver effects in animals fATSDR. 2018: Li etal.. 2017:
Viberg and Eriksson. 2017: U.S. EPA. 2016c. d], and proposed adverse outcome
pathways for hepatic steatosis fMellor etal.. 20161
NAFLD = nonalcoholic fatty liver disease; ROS = reactive oxygen species; TNFa = tumor necrosis factor alpha;
XME = xenobiotic metabolizing enzymes.
This document is a draft for review purposes only and does not constitute Agency policy.
103 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
The analysis of the involvement of PPARa and these other signaling cascades in hepatic
toxicity after exposure to these five PFAS will focus on the concordance of changes in the specific
mechanistic events or separate pathways to effects (i.e., in Figure 7, and as otherwise identified
during assessment-specific evaluations] across species to ascertain the relevance of animal studies
to human health. The analyses of evidence for each mechanistic event and potential pathway will
be qualitatively analyzed for various aspects of the Hill considerations outlined in the EPA Cancer
Guidelines framework for MOA analysis (U.S. EPA. 2005al Given the focus of these analyses, the
review will stress the aspects of consistency, coherence, and biological plausibility to ascertain the
level of support (or lack thereof), depending on the availability of data. To facilitate this analysis,
the following prompting questions and clarifying considerations will be used, depending on the
assessment-specific data:
• What is the level of evidentiary support (or lack thereof) for the mechanistic events or
signaling pathways, based on the assessment-specific PFAS data? In parallel, are
assessment-specific data available to inform the strength of the linkages between events in
the pathway or across pathways? Based on the assessment-specific feasibility of doing so
(i.e., an adequate database size], answering these questions will incorporate the use of
general categories of evidence strength so that the different evaluations can be succinctly
summarized. In general, the categories will be based on the following descriptions, each of
which will be clarified as support either for or against involvement of an event or pathway:
"strong"—independent studies using different experimental models provide consistent
and/or coherent evidence; "marginal"—some consistent and/or coherent evidence,
although some results may be equivocal or vary from one model to another; and
"unclear"—largely inconsistent evidence or insufficient evidence to evaluate.
• Are sufficient assessment-specific data available to inform exposure duration- or
level-dependencies for any of the evaluated mechanistic events or pathways?
• Is the assessment-specific evidence (on specific events or pathways in general] consistent
with the general biology of the human liver or mechanisms known to be associated with
noncancer liver effects in humans? To consider this question, assessments will compare the
endpoint-level results across studies on a particular PFAS against the mechanistic
understanding/underlying biology for similar effects in the human liver. (Note: this analysis
might be informed by studies or reviews on the more robust PFOA/perfluorooctane
sulfonate [PFOS] evidence bases.]
• Are responses across studies for these five PFAS assessments indicative of activation of
specific mechanisms or signaling pathways conserved across experimental models and
designs? To consider this question, assessments will include an evaluation of consistency
and coherence across different species and strains of animals, human and animal cell
culture models, and in vivo humanized animal models, depending on data availability.
• Does the assessment-specific mechanistic information indicate there are likely to be
populations or lifestages that may be more susceptible to PFAS-induced liver effects?
This document is a draft for review purposes only and does not constitute Agency policy.
104 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
The assessment-specific conclusions (and attendant uncertainties] regarding these
questions will be used to draw judgments regarding the human relevance of these animal effects,
and the rationale for these judgments will be documented transparently within each assessment.
As described in EPA guidance fU.S. EPA. 2005al human relevance is the default and mechanistic
evidence will need to be compelling and strong to reach a conclusion otherwise.
9.2.3. Toxicological Relevance of Select Outcomes Observed in Animals
Lastly, the preliminary literature inventory identified the potential for PFAS
exposure-mediated effects on several health outcomes for which it is expected to be difficult to
identify whether any observed changes (or a lack of changes] are toxicologically relevant,
specifically some changes in the kidney and liver (see below]. It is expected that in some instances,
the synthesis will need to address this issue to inform whether the effects in animals are relevant to
interpreting the potential for PFAS exposure to cause a human health effect, and in other instances
addressing this issue might be necessary for identifying a level of change for use in determining the
potential for adversity or for use in dose-response analysis. It is possible that additional outcomes
with similar questions of health relevance might be identified during the development of these
assessments. If so, the specifics of the approach selected to address those outcomes will be
documented in the assessments] and as an update to this protocol. For the aforementioned
outcomes, different approaches will be taken, specifically:
1) Kidney changes in rats, including chronic progressive nephropathy (CPN) and effects that
appear to be mediated by an alpha 2u-globulin MOA. Because the rodent (i.e., male
rat]-specific alpha 2u-globulin MOA is not considered relevant to humans, assessments with
evidence indicating its involvement will include an evaluation and judgment of the evidence
supporting (or not supporting] dependence on this MOA. Specifically, these data will be
evaluated against the predefined criteria established by the U.S. EPA (1991a) and/or more
recently established criteria, such as those published by Swenberg and Lehman-McKeeman
f19991 Relatedly (and possibly overlapping the evaluation of alpha 2u-globulin, because
this MOA may exacerbate CPN], there is no human disease analog to the constellation of
changes observed in rodent CPN. CPN represents a complex disease process in rats, and its
etiology is unknown. Thus, these evaluations will include judgments as to whether all or a
subset of the observed changes have adequate evidence to identify dependence on
rodent-specific processes, including whether it can be concluded (i.e., based on biological
understanding] that the observed kidney endpoints are associated with CPN and the
potential for exacerbation of human-relevant disease processes can be ruled out. Data
permitting, the assessments will consider whether these conclusions vary by exposure level.
2) Hepatic changes. Some individual liver endpoints (and even some constellations of
endpoints] might be considered adaptive in nature, possibly leading to the interpretation
that some statistically significant changes are not indicative of adverse effects. These
endpoints include increased liver weight, cellular hypertrophy, and single cell
necrosis/apoptosis. To draw inferences regarding the adversity of these types of liver
effects, these assessments will consider the panel recommendations outlined by Hall et al.
This document is a draft for review purposes only and does not constitute Agency policy.
105 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
f20121 to draw assessment-specific judgments regarding adversity. Briefly, these include
evaluation of the available histological data and results suggesting structural degeneration
or cellular demise (e.g., apoptosis, oncosis, and/or necrosis], and clinical evidence of
hepatocyte damage. As the Hall etal. (20121 recommendations were developed in the
context liver tumor formation, consultation of additional reference materials will be
considered on an assessment-specific basis. Each assessment will include an explanatory
rationale documenting the application of the Hall etal. (20121 recommendations (and any
other considerations] to the available evidence.
9.2.4. Other Focused Mechanistic Analyses
Other analyses within the syntheses of mechanistic information will focus on the evidence
most useful for informing key uncertainties in the human or animal health effect evidence.
This means that, for example, if extensive and consistent /7/^-confidence human or animal
evidence is available, the need to synthesize all relevant mechanistic evidence will likely be
diminished. In these cases, the analyses will focus on the review and interpretation of smaller sets
of mechanistic studies that specifically address controversial or outstanding issues that are
anticipated to have a substantial impact on the assessment conclusions. Generally, key
uncertainties will be addressed in the mechanistic evidence syntheses by considering the biological
understanding of how the effect(s] in question develop or are related. In this way, the analyses can
provide information on, for example, (1] potential precursor events when the apical data are
uncertain (or unusable for dose-response analyses], (2] the human relevance of animal results
when their relevance is unclear or controversial and the human evidence is weak, (3] the shape of
the dose-response curve at low exposure levels when this understanding is highly uncertain and
data informing this uncertainty are known to exist, or (4] the identification of likely susceptible
populations and lifestages. Thus, consideration of biological understanding represents an important
component of the evidence analysis. However, mechanistic understanding is not a prerequisite for
drawing a conclusion that a chemical causes a given health effect fNTP. 2015: NRC. 20141.
To identify the focused set(s] of studies to use in analyzing critical mechanistic questions
other than those outlined in Sections 9.2.1-9.2.3, a stepwise approach will be applied to
progressively define the scope of the mechanistic information to be considered throughout
assessment development. This stepwise scoping begins during the literature search and screening
steps and depends primarily on the potential health hazard signals that arise from the individual
human and/or animal health effect studies, or from mechanistic studies that signal potential health
hazards that have not been examined in studies of phenotypic, potentially adverse effects.
Examples of the focused questions or scenarios triggering these mechanistic evaluations, as well as
when during the systematic review they are likely to apply, are listed in Table 17. While the specific
methods for evaluating the evidence most relevant to each question will vary, some general
considerations for judging the evidence strength in these syntheses are provided below, and if
necessary, assessment-specific refinements will be included as updates to the protocol.
This document is a draft for review purposes only and does not constitute Agency policy.
106 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 17. Examples of questions and considerations that can trigger focused
analysis and synthesis of mechanistic information
Key
assessment-
specific
uncertainties
Examples of questions and PFAS-specific considerations for identifying the
uncertainties and key evidence to analyze
Addressing
database
completeness
based on
literature
inventories of
human,
animal, and
mechanistic
information
• Are there mechanistic studies on an organ system or potential health hazard that were
not examined by human or animal studies meeting the PECO criteria?
o Depending on the extent of the available data, consider the utility of
developing a separate synthesis of evidence versus the utility of a concise,
narrative summary (or evidence mapping) to describe these knowledge gaps.
Consider whether the mechanistic evidence might be sufficient to substantiate
a conclusion on its own (if so, a separate synthesis will be developed).
Addressing
questions of
inconsistency
within the
human and
animal
evidence
• For the health effects of potential concern, is a mechanistic evaluation(s) warranted to
inform questions regarding the consistency of the available human or animal studies?
Typically, this consideration would focus on health effects that show some indication of
an association in epidemiological studies or causality in experimental studies during
evidence synthesis. Based on the literature inventory, consider whether mechanistic
data are available to inform the specific, key uncertainties in question. Examples of
specific scenarios for evaluation include:
o If cancer has been observed and tumor types appear to differ across
populations (e.g., species or sex), review the literature inventory for
mechanistic data that might be relevant to interpreting such differences, and
conduct analyses as warranted based on that review. Approaches outlined in
the EPA cancer guidelines (U.S. EPA, 2005a) that mav be relevant to these
analyses will be applied, as appropriate.
o If pronounced and unexplained differences in health effect(s)-specific
responses are observed across lifestages or populations (e.g., animal strain;
human demographic), first consider toxicokinetic differences for the specific
PFAS, and then the mechanistic evidence relevant to assessing the potential
for health effect-specific biological differences in response (toxicodynamics).
Further, inconsistent evidence (i.e., heterogeneous results) across different
animal species or human populations might be informed by a review of the
evidence relevant to whether different mechanisms may be operant in the
different populations (e.g., evidence demonstrating that certain species are
more or less sensitive to a certain biological perturbation; evidence that gene
polymorphisms are related to variability in response).
This document is a draft for review purposes only and does not constitute Agency policy.
107 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Key
assessment-
specific
uncertainties
Examples of questions and PFAS-specific considerations for identifying the
uncertainties and key evidence to analyze
Addressing
questions of
biological
plausibility3
and coherence
within the
human and
animal
evidence, and
coherence
across bodies
of evidence
For the health effects of potential concern, would a mechanistic evaluation(s) of
biological plausibility (usually for an individual outcome) or coherence (usually across
outcomes) provide meaningful information for interpreting the evidence strength?
Typically, this consideration would focus on effects for which the evidence strength for
an individual outcome (either for or against an effect) is questionable (e.g., primarily
studies of low confidence), when a substantial outstanding methodological concern(s)
across the relevant studies exists, or when evidence exists for multiple, potentially
related (e.g., biologically) outcomes. Based on the literature inventory, consider
whether there are mechanistic data available to inform the specific, key uncertainties in
question. Examples of specific scenarios for evaluation may include:
o If the evidence for a given outcome is weak or uncertain, or when unaddressed
methodological concerns identify critical uncertainties in the human or animal
findings for a health effect, identify data on mechanistic changes in exposed
humans or animals that are likely to be linked to the development or
occurrence of the health outcome in question. If enough suitable studies are
available, analyze data on changes expected to be related to the phenotypic
finding(s) of interest, which can either increase or decrease the evidence
strength that the finding(s) is real. It is important to note that the absence of a
mechanistic explanation for an association (e.g., the MOA is not understood)
will not be used to reduce confidence in observations from human or animal
studies. However, the plausibility of an association observed in human or
animal studies may be diminished if expected mechanistic findings (e.g., based
a known biological dependence) are tested and not apparent. The mechanistic
evidence on possible precursors or effects that are known to co-occur with the
health outcome of interest are particularly impactful when the changes are
observed in the same exposed population presenting the outcome of interest.
An understanding of mechanistic pathways (e.g., by identifying and analyzing
mechanistic precursor events linked qualitatively or quantitatively to apical
health effect[s]; see Section 9.2.2 for additional context) can inform the
strength of the evidence integration judgments (see Section 10).
o If evidence on multiple health outcomes within an organ system, or possibly
across organ systems (e.g., thyroid and nervous system), are available and the
strength of the evidence for any single outcome is uncertain, identify biological
data that can inform understanding of the relatedness of outcomes within and
across systems. Biological understanding or strong mechanistic support (e.g., a
shared mechanistic event) for linkages across outcomes can increase the
strength of the evidence when changes are related. However, evidence
strength may be diminished if an expected pattern among biologically linked
outcomes is not observed. Interpretation of the pattern of changes across the
outcomes will consider the underlying biological understanding (e.g., one
outcome may be expected to precede the other, or be more sensitive). These
same considerations inform analyses of the coherence of observed effects
across evidence streams during evidence integration (see Section 10.2).
This document is a draft for review purposes only and does not constitute Agency policy.
108 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Key
assessment-
specific
uncertainties
Examples of questions and PFAS-specific considerations for identifying the
uncertainties and key evidence to analyze
Addressing
questions on
the human
relevance of
findings in
animals
• For the health effects of potential concern, does the available evidence raise questions
of human relevance? Typically, this consideration applies when human evidence is
lacking or has results that differ from animal studies, noting that some differences in
responses across humans and animals are acknowledged [e.g., for cancer, site
concordance is not a requirement for determining the relevance of animal data for
humans (U.S. EPA, 2005a); for noncancer nervous svstem effects, behavioral changes
can manifest differently between animals and humans]. The identification of potential
differences will also consider ADME information across species, primarily relating to
distribution (e.g., to the likely target tissue) and PFAS half-life. Examples of information
to identify from the literature inventory, as well as specific scenarios and considerations
for these analyses may include:
o If there is no evidence indicating that the animal results are irrelevant to
humans, summarize existing knowledge on the development of the health
effect in each species, including potential differences in PFAS toxicokinetics,
and assess the relatedness across species. Note that in the absence of
sufficient evidence to the contrary, effects in animal models are assumed to be
relevant to humans (ATSDR, 2018; NIEHS, 2015; U.S. EPA, 2005a).b
o If there is evidence indicating that the mechanisms underlying the effects in
animals may not operate in humans, or that the available animal model(s) may
not be suitable for the human health outcome(s) of interest, present and
analyze the strength of the evidence for and against the human relevance of
the observed findings. In addition to considerations specific to the outcome of
interest, the analysis will evaluate observations of mechanistic changes in
exposed humans for similarities or biological coherence with mechanistic or
toxicological changes in experimental animals interpreted to be associated
with the health outcome under evaluation. It may also include an evaluation
of findings across species known or presumed to be more or less relevant for
interpreting potential human toxicity for the health effect(s) in question. In
rare instances or for controversial decisions that are likely to drive key
assessment conclusions, the analysis may extend to a detailed analysis of a
plausible mechanistic pathway(s) or MOA(s) within which each key event and
key event relationship is evaluated regarding the likelihood of similarities
(e.g., in presence or function) across species. These analyses, regardless of
their rigor, will lead to a definitive judgment about whether the animal
response is relevant to humans during evidence integration (see Section 10).
This document is a draft for review purposes only and does not constitute Agency policy.
109 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Key
assessment-
specific
uncertainties
Examples of questions and PFAS-specific considerations for identifying the
uncertainties and key evidence to analyze
Addressing
questions on
potential
susceptibility
for hazard
identification
and
dose-response
analysis
• For the health effects of potential concern, do the results from the human and animal
health effect studies appear to differ by categories that indicate the apparent presence
of susceptible populations (e.g., across demographics, species, strains, sexes, or
lifestages)? Separately, are there human or animal study data that could identify or
clarify population differences in response (e.g., experiments testing sensitivity of
responses across lifestages or across genetic variations; observed differences
attributable to genetic polymorphisms)? Are there mechanistic data (i.e., based on the
literature inventory) that address potential susceptibility factors?0 If evidence exists for
any of these scenarios, information on susceptibility will be reviewed and, if impactful
to assessment conclusions, analyzed in detail. Examples of when these analyses are
important include:
o If the analysis of evidence indicates the likely presence of a sensitive
population or lifestage in humans, the groups likely to be at greatest risk will
be captured in the evidence integration narrative (see Section 10). In addition,
this narrative will discuss whether the appropriate analogous exposures and
populations or lifestages were adequately represented or tested in the
available human or animal studies, and if not, will identify studies on the most
susceptible populations or lifestages as key research needs (see Section 10).
o If the analysis of evidence indicates the likely presence of a sensitive
population or lifestage in humans, this information will be used to select
studies for quantitative analysis (e.g., prioritizing those studies that include
such populations [see Section 11]). If specific studies addressing these
susceptibilities are unusable for quantitative analysis, susceptibility data may
be used to support refined human variability uncertainty factors or
probabilistic uncertainty analyses (see Section 11).
This document is a draft for review purposes only and does not constitute Agency policy.
110 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Key
assessment-
specific
uncertainties
Examples of questions and PFAS-specific considerations for identifying the
uncertainties and key evidence to analyze
Addressing
questions on
biological
understanding
to optimize
dose-response
analysis
• If the human and/or animal health effect data amenable to dose-response analysis are
weakd or only at high exposure levels, or if the selection of critical parameters for
modeling is uncertain, the following analyses will be considered:
o When the apical health effect data are highly uncertain or cannot be used with
confidence for the purpose of deriving quantitative estimates, mechanistic
precursor events linked qualitatively or quantitatively to the phenotypic effect
can be evaluated for use as surrogate markers (e.g., based on the strength and
completeness of the linkage between mechanistic and phenotypic effects) for
deriving quantitative estimates.
o When there is a notable lack of understanding of the appropriate exposure
metric, biomarker, or modeling parameter for developing quantitative
estimates, toxicokinetic and mechanistic understanding of the development of
the health effect can inform the most biologically appropriate measure.
o When there are dose-response modeling decisions or uncertainties that would
be substantially improved by biological or toxicokinetic understanding,
mechanistic analyses can improve selection of particular models (e.g., a linear,
nonlinear, or threshold model) and help evaluate the appropriateness of
integrating/combining data across related outcomes (e.g., based on biological
coherence or a conserved MOA). For cancer toxicity values, existing guidance
will be consulted (U.S. EPA, 2005a).
aAs applied herein, biological plausibility describes mechanistic information that either strengthens or weakens
an interpretation of the likelihood of an association between exposure and the health effect. The interpretation
of biological plausibility considers the existing biological understanding of how the health effect develops and
can involve analyses of information at different levels of biological complexity (e.g., molecular, cellular, tissue).
bAs described in the EPA RfD/RfC Technical Report (2002), "one of the major default assumptions in EPA's risk
assessment guidelines is that animal data are relevant for humans [e.g., U.S. EPA (1998), U.S. EPA (1991a), and
U.S. EPA (1996b)1. Such defaults are intended to be used in the absence of experimental data that can provide
direct information on the relevance of animal data" (U.S. EPA, 2002).
Susceptibility factors include lifestage, demographics and social determinants, behavioral factors, health status,
and genetic variability. Although not considered in these analyses, factors that can increase vulnerability, such
as other pollutant exposures or differential proximity to exposure sources, are typically considered during a full
risk assessment.
dNote that "weak" here refers to the study's usability for dose-response analysis specifically. Such studies may be
judged to be of medium or high confidence for the purposes of identifying potential hazards but possess
limitations preventing their use for deriving reliable quantitative estimates.
PECO = populations, exposures, comparators, and outcomes; RfC = inhalation reference concentration; RfD = oral
reference dose.
1
2 If focused areas for additional mechanistic evaluations are identified to inform key
3 assessment-specific uncertainties (e.g., by applying Table 13], the assessments will identify the
4 most impactful studies for evaluation. This could represent only a subset of the potentially relevant
5 studies, particularly if there are many mechanistic studies relevant to the specific questionfs].
This document is a draft for review purposes only and does not constitute Agency policy.
Ill DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Because the potential influence of the information provided by the available studies can vary
depending on the question(s] or the associated mechanistic events or pathways, the rigor of the
analyses will likewise vary from cursory insights drawn from sets of unanalyzed results to detailed
evaluations of a subset of the most relevant, individual mechanistic studies. Although the specifics
that might be applicable across potential mechanistic topic areas cannot be predefined, the analyses
will first consider the studies based on their toxicological relevance to answering the specific
question (e.g., model system; specificity of the assay for the effect of interest], potentially refining
the focus to a subset of the most relevant studies. This will be particularly important when the
set(s] of studies are inconsistent and potentially conflicting, include potentially more informative
studies that challenge the necessity of proposed mechanistic relationships between exposure and
an apical effect (e.g., altering a receptor-mediated pathway through chemical intervention or using
knockout animals], or when it is apparent that particular study design aspects in some or all of the
relevant studies are likely to have significant flaws or important uncertainties (e.g., for certain
questions, a preliminary review of the exposure methods across the relevant mechanistic studies
can flag serious deficiencies]. In general, across these assessments, relevant mechanistic
information from in vivo studies will be prioritized, with preference given to PFAS- and
endpoint-relevant exposure routes and exposure designs. Analysis of ex vivo and in vitro studies
will then be considered, prioritizing those most informative to evaluating the mechanistic events
indicated by the in vivo data, including studies conducted under conditions most relevant to human
exposures and in model systems best replicating in vivo human biology.
In some instances, additional literature searches may be warranted, targeting mechanistic
events or biological pathways that are not specific to a particular PFAS or group of PFAS. When
more rigorous mechanistic analyses are deemed necessary, the review will be aided by the use of
pathway-based organizational methods and, if available, established evidence evaluation
frameworks. These approaches provide transparency and objectivity for integrating and
interpreting the mechanistic events and pathways anchored to the specific questions that have
been identified (e.g., anchored to a specific health effect] across diverse sets of relevant data
(e.g., human, animal, and in vitro studies]. The approaches may be facilitated by using
organizational tools or frameworks, such as AOPs (see example in Section 9.2.2], As noted above,
any additional assessment-specific literature searches and evaluation methods will be described in
updates to the protocol.
Based on the analyses and considerations outlined in Sections 9.2.1-9.2.4, the results of the
health effect- and assessment-specific mechanistic evidence syntheses will inform both evidence
integration and dose-response analyses (see Sections 10 and 11],
This document is a draft for review purposes only and does not constitute Agency policy.
112 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
10. EVIDENCE INTEGRATION
For the analysis of human health outcomes that might result from chemical exposure, these
PFAS assessments will draw integrated judgments across human, animal, and mechanistic evidence
for each assessed health effect (see Section 9], As previously discussed in Section 9.2, the approach
to evaluating the mechanistic evidence relevant to each assessed health effect will follow a step-
wise approach and is expected to vaiy depending on the nature and impact of the uncertainties
identified within each evidence base, as well as the specific mechanistic information available to
address those uncertainties. This includes evaluations of mechanistic evidence relevant to the
identified key science issues (Section 2.4] prior to or in parallel with evaluations of the phenotypic
data in human and animal studies, as well as other focused mechanistic analyses identified during
draft development to address key assessment uncertainties (see Section 9.2.4 for a discussion of
these scenarios]. During evidence integration, a structured and documented process will be used,
as follows (and depicted in Figure 8]:
• Building from the separate syntheses of the human and animal evidence (see Section 9.1],
the strength of the evidence from the available human and animal health effect studies will
be summarized in parallel, but separately, using a structured evaluation of an adapted set of
considerations first introduced by Sir Bradford Hill fHill. 19651 Table 19 describes these
structured evaluations and the explicit consideration of study confidence within each
evaluation domain. Based on the approaches and considerations described in Section 9.2,
these summaries will incorporate the relevant mechanistic evidence (or MOA
understanding] that informs the biological plausibility and coherence within the available
human or animal health effect studies.
• The strength of the animal and human evidence will be considered together in light of
inferences across evidence streams. Specifically, the inferences considered during this
integration include the human relevance of the animal and mechanistic evidence, coherence
across the separate bodies of evidence, and other important information (e.g., judgments
regarding susceptibility]. Note that without evidence to the contrary, the human relevance
of animal findings is assumed.
• A summary judgment is drawn as to whether the available evidence base for each potential
human health effect as a whole is sufficient (or insufficient] to indicate that PFAS exposure
has the potential to be hazardous to humans17.
17 Due to the expected rarity of scenarios where there is "sufficient evidence to judge that a hazard is unlikely" (see
description in Table 20 and Section 10.2) and to improve readability, this judgment is not specified in some instances.
This document is a draft for review purposes only and does not constitute Agency policy.
113 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Evidence Stream Evaluation
Based on the structured review of adapted Hill
considerations (including biological understanding),
as part of the evidence integration narrative:
• Qualitatively summarize the strength of the
evidence from studies in humans.
• Qualitatively summarize the strength of the
evidence from animal studies.
+
Inference Across Evidence Streams
•Information on the human relevance of
the animal and mechanistic evidence
•Coherence across bodies of evidence or
with related effects
•Other (e.g., read-across; susceptibility)
Evidence Integration
Summary Judgment
Overall judgments across
evidence for each
potential human health
effect, including evidence
basis rationale
Figure 8. Process for evidence integration. Note that "sufficient evidence" could
indicate a judgment of "sufficient evidence for hazard" or "sufficient evidence to
judge that a hazard is unlikely", depending on the nature and extent of the available
evidence (see Table 20).
2
3 The decision points within the structured evidence integration process will be summarized
4 in an evidence profile table for each health effect category (see Table 18 for a preliminary template
5 version) in support of the evidence integration narrative. The specific decision frameworks for the
6 structured evaluation of the strength of the human and animal evidence streams and for drawing
7 the overall evidence integration judgment are described in Section 10.1. This process is similar to
8 that used by the Grading of Recommendations Assessment, Development, and Evaluation [GRADE;
9 (Morgan et al„ 2016: Guvattet al„ 2011: Schunemann et al„ 20111]. which arrives at an overall
10 integration conclusion based on consideration of the body of evidence. As described in Section 9,
11 the human, animal, and mechanistic evidence syntheses serve as inputs providing a foundation for
12 the evidence integration decisions; thus, the major conclusions from these syntheses will be
13 summarized in the evidence profile table (see Table 18 for a preliminary template version)
14 supporting the evidence integration narrative. The evidence profile tables on each potential human
15 health effect evaluated will summarize the judgments and their evidence basis for each step of the
16 structured evidence integration process. Separate sections are included for summarizing the
17 human and animal evidence, for the inference drawn across evidence streams, and for the overall
18 evidence integration judgment. The table presents the key information from the different bodies of
19 evidence that informs each decision.
This document is a draft for review purposes only and does not constitute Agency policy.
114 ' DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 18. Evidence profile table template
Studies and
interpretation
Factors that increase
strength
Factors that decrease
strength
Summary of evidence streams
Inferences across evidence
streams
Overall Evidence Integration
Judgment
[Health Effect or Outcome Grouping]
a
Evidence in Studies of Humans [may be separated by exposure route ]
• Human relevance of
findings in animals
• Cross-stream coherence
• Other inferences:
o Information on
susceptibility
o MOA analysis
inferences
o Relevant
information from
other sources (e.g.,
read across)
Describe judgment regarding
whether there is sufficient (or
insufficient) evidence to identify
a potential human health
hazard, integrating evidence
across streams and including a
summary of the models and
range of dose levels upon which
the judgment is primarily
reliant.
• References
• Study confidence
• Study design
description
• Consistency
• Dose-response
gradient
• Coherence of
observed effects
• Effect size
• Mechanistic evidence
providing plausibility
• Medium or high
confidence studies'3
• Unexplained
inconsistency
• Imprecision
• Low confidence
studies'3
• Evidence
demonstrating
implausibility
Qualitative summary of the strength of
the evidence from human studies
based on the factors at left, including
the primary evidence basis and
considering:
• Results across human
epidemiological and controlled
exposure studies
• Human mechanistic evidence
informing biological plausibility
(e.g., precursor events linked to
adverse outcomes)
a
Evidence from Animal Studies [may be separated by exposure route ]
• References
• Study confidence
• Study design
description
• Consistency and/or
Replication
• Dose-response
gradient
• Coherence of
observed effects
• Effect size
• Mechanistic evidence
providing plausibility
• Medium or high
confidence studies'3
• Unexplained
inconsistency
• Imprecision
• Low confidence
studies'3
• Evidence
demonstrating
implausibility
Qualitative summary of the strength of
the evidence for an effect in animals
based on the factors at left, including
the primary evidence basis and
considering:
• Results across animal toxicological
studies
• Animal mechanistic evidence
informing biological plausibility
(e.g., precursor events linked to
adverse outcomes)
aln addition to exposure route, the summaries of the strength of each evidence stream may include multiple rows- e.g., by study confidence, population, or species, if this
informed the analysis of results heterogeneity.
1 bStudy confidence, based on evaluation of risk of bias and study sensitivity (see Section 6), and information on susceptibility will be considered when evaluating each of the
2 other factors that increase or decrease strength (e.g., consistency). Notably, lack of findings in studies deemed insensitive neither increases or decreases strength.
3
This document is a draft for review purposes only and does not constitute Agency policy.
115 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
10.1. EVALUATING THE STRENGTH OF THE HUMAN AND ANIMAL
EVIDENCE STREAMS
As summarized above, prior to drawing overall evidence integration judgments about
whether exposure to one of these five PFAS has the potential to cause certain health effect(s] in
humans given relevant exposure circumstances, the strength of evidence for the available human
and animal evidence will be evaluated and summarized. For each assessed health effect or health
effect grouping (see Section 5 for examples of the endpoints that will be considered within each
health effect category], the relevant mechanistic evidence in exposed humans and animals (or in
their cells, relevant NAMs or in silico models], which will be synthesized based on the approaches
and considerations in Section 9.2, will be integrated with the evidence from the available studies of
phenotypic effects in humans and animals. The considerations previously outlined in Table 15 (the
different features of the evidence considered and summarized during evidence synthesis; see
Section 9] will be evaluated by the specific PFAS assessment teams within the context of how they
affect judgments of the strength of evidence (see Table 19], which will directly inform the overall
evidence integration judgment (see Section 10.2], The evaluation of the strength of the human or
animal health effects evidence (i.e., based on the considerations in Table 19] will preferably occur at
the most specific health outcome level possible (e.g., an analysis at the level of decreased
pulmonary function is generally preferable to an analysis of respiratory system effects], if there is
an adequate set of studies for analyses at this level and considering the interrelatedness of the
available outcomes. If studies on a target system are sparse or varied, or if the interpretation of
evidence strength relies largely on the consideration of coherence across related outcomes, then
the analyses may need to be conducted at a broader health effect level. The factors judged to
increase or decrease the strength of the evidence will be summarized in tabular format using the
evidence profile table template in Table 18 to transparently convey expert judgments made
throughout the evidence synthesis and integration processes. The evidence profile table allows for
consistent documentation of the supporting rationale for each decision.
This document is a draft for review purposes only and does not constitute Agency policy.
116 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 19. Considerations that inform evaluations of the strength of the human and animal evidence
Consideration
Increased evidence strength
(of the human or animal evidence)
Decreased evidence strength
(of the human or animal evidence)
The structured categories and criteria inTable 20 (Section 10.2) will guide the application of strength-of-evidence judgments for an outcome or health effect.
Evidence synthesis scenarios that do not warrant an increase or decrease in evidence strength for a given consideration will be considered "neutral" and are
not described in this table (and, in general, will not be captured in the assessment-specific evidence profile tables).
Risk of bias; sensitivity (across studies)
• An evidence base of high- or
medium-confidence studies increases
strength.
• An evidence base of mostly /ow-confidence studies
decreases strength. An exception to this is an
evidence base of studies in which the primary issues
resulting in low confidence are related to insensitivity.
This may increase evidence strength in cases where an
association is identified because the expected impact
of study insensitivity is towards the null.
• Decisions to increase strength for other
considerations in this table should generally not be
made if there are serious concerns for risk of bias.
Consistency
• Similarity of findings for a given outcome
(e.g., of a similar magnitude, direction) across
independent studies or experiments increases
strength, particularly when consistency is
observed across populations (e.g.,
geographical location) or exposure scenarios
in human studies, and across laboratories,
populations (e.g., species), or exposure
scenarios (e.g., duration; route; timing) in
animal studies.
• Unexplained inconsistency (i.e., conflicting evidence;
see U.S. EPA (2005a)) decreases strength. Generally,
strength should not be decreased if discrepant
findings can be reasonably explained by study
confidence conclusions; variation in population or
species, sex, or lifestage; exposure patterns
(e.g., intermittent or continuous); exposure levels (low
or high); or exposure duration.
This document is a draft for review purposes only and does not constitute Agency policy.
117 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Consideration
Increased evidence strength
(of the human or animal evidence)
Decreased evidence strength
(of the human or animal evidence)
Strength (effect magnitude) and precision
• Evidence of a large magnitude effect
(considered either within or across studies)
can increase strength. Effects of a concerning
rarity or severity can also increase strength,
even if they are of a small magnitude.
• Precise results from individual studies or
across the set of studies increases strength,
noting that biological significance is prioritized
over statistical significance.
• Strength may be decreased if effect sizes that are
small in magnitude are concluded not to be
biologically significant, or if there are only a few
studies with imprecise results.
This document is a draft for review purposes only and does not constitute Agency policy.
118 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Consideration
Increased evidence strength
(of the human or animal evidence)
Decreased evidence strength
(of the human or animal evidence)
Biological gradient/dose-response
• Evidence of dose-response increases strength.
Dose-response may be demonstrated across
studies or within studies and it can be
dose- or duration-dependent. It also may not
be a monotonic dose-response (monotonicity
should not necessarily be expected,
e.g., different outcomes may be expected at
low vs. high doses due to activation of
different mechanistic pathways or induction
of systemic toxicity at very high doses).
• Decreases in a response after cessation of
exposure (e.g., symptoms of current asthma)
also may increase strength by increasing
certainty in a relationship between exposure
and outcome (this is most applicable to
epidemiology studies because of their
observational nature).
• A lack of dose-response when expected based on
biological understanding and having a wide range of
doses/exposures evaluated in the evidence base can
decrease strength.
• In experimental studies, strength may be decreased
when effects resolve under certain experimental
conditions (e.g., rapid reversibility after removal of
exposure). However, many reversible effects are of
high concern. Deciding between these situations is
informed by factors such as the toxicokinetics of the
chemical and the conditions of exposure fsee U.S. EPA
(1998)1, endpoint severity, judgments regarding the
potential for delayed or secondary effects, as well as
the exposure context focus of the assessment
(e.g., addressing intermittent or short-term
exposures).
• In rare cases, and typically only in toxicology studies,
the magnitude of effects at a given exposure level
might decrease with longer exposures (e.g., due to
tolerance or acclimation). Like the discussion of
reversibility above, a decision about whether this
decreases evidence strength depends on the exposure
context focus of the assessment and other factors.
• If the data are not adequate to evaluate a
dose-response pattern, then strength is neither
increased or decreased.
This document is a draft for review purposes only and does not constitute Agency policy.
119 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Consideration
Increased evidence strength
(of the human or animal evidence)
Decreased evidence strength
(of the human or animal evidence)
Coherence
• Biologically related findings within an organ
system, or across populations (e.g., sex)
increase strength, particularly when a
temporal- or dose-dependent progression of
related effects is observed within or across
studies, or when related findings of increasing
severity are observed with increasing
exposure.
• An observed lack of expected coherent changes
(e.g., well-established biological relationships) will
typically decrease evidence strength. However, the
biological relationships between the endpoints being
compared and the sensitivity and specificity of the
measures used need to be carefully examined. The
decision to decrease evidence strength depends on
the availability of evidence across multiple related
endpoints for which changes would be anticipated,
and it considers factors (e.g., dose and duration of
exposure, strength of expected relationship) across
the studies of related changes.
Mechanistic evidence related to biological
plausibility
• Mechanistic evidence of precursors or health
effect biomarkers in well-conducted studies
of exposed humans or animals, in
appropriately exposed human or animal cells,
or other relevant human, animal, or in silico
models (including new approach methods,
NAMs) increases strength, particularly when
this evidence is observed in the same
cohort/population exhibiting the phenotypic
health outcome.
• Evidence of changes in biological pathways or
support for a proposed MOA in appropriate
models also increases strength, particularly
when support is provided for rate-limiting or
key events, or is conserved across multiple
components of the pathway or MOA.
• Mechanistic understanding is not a prerequisite for
drawing a conclusion that a chemical causes a given
health effect; thus, an absence of knowledge should
not be used a basis for decreasing strength (NTP,
2015; NRC, 2014).
• Mechanistic evidence in well-conducted studies (see
examples of evidence types at left) that demonstrates
that the health effect(s) are unlikely to occur, or only
likely to occur under certain scenarios (e.g., above
certain exposure levels), can decrease evidence
strength. A decision to decrease strength depends on
an evaluation of the strength of the mechanistic
evidence supporting vs. opposing biological
plausibility, as well as the strength of the health
effect-specific findings (e.g., stronger health effect
data require more certainty in mechanistic evidence
opposing plausibility).
This document is a draft for review purposes only and does not constitute Agency policy.
120 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
1 For human and animal evidence, the analyses of each consideration in Table 19 will be used
2 to qualitatively summarize the strength-of-evidence for the separate evidence streams in the
3 evidence integration narrative. Table 20 provides the criteria that will guide how to draw the
4 judgment for each health effect, and the terms that will be used to summarize those evidence
5 integration judgments.
10.2. OVERALL EVIDENCE INTEGRATION JUDGMENTS
6 Evidence integration combines decisions regarding the strength of the animal and human
7 evidence with considerations regarding mechanistic information on the human relevance of the
8 animal evidence, relevance of the mechanistic evidence to humans (especially in cases where
9 animal evidence is lacking], coherence across bodies of evidence, and information on susceptible
10 populations and lifestages, based on the considerations and analyses outlined in Section 9.2. This
11 evidence integration decision process will culminate in an evidence integration narrative that
12 summarizes the judgments regarding the evidence for each potential health effect (i.e., each
13 noncancer health effect and specific type of cancer, or broader grouping of related outcomes]. For
14 each health effect, this narrative will include:
15
16 • A descriptive summary of the primary judgments about the evidence informing the
17 potential for health effects in exposed humans, based on the following analyses:
18 ° evaluations of the strength of the available human and animal evidence (see
19 Section 10.1];
20 ° consideration of the coherence of findings (i.e., the extent to which the evidence for
21 health effects and relevant mechanistic changes are similar] across human and animal
22 studies;
23 ° other information on the human relevance of findings in animals (see Section 9.2]; and
24 ° conclusions drawn based on the predefined mechanistic analyses (see
25 Sections 9.2.1-9.2.3], as well as those based on analyses identified during stepwise
26 consideration of the health effect-specific evidence during draft development (see
27 Section 9.2.4],
28 • A summary of key evidence supporting these judgments, highlighting the evidence that was
29 the primary driver of these judgments and any notable issues (e.g., data quality; coherence
30 of the results], and a narrative expression of confidence (a summary of strengths and
31 remaining uncertainties] for these judgments.
32 • Information on the general conditions of expression of these health effects (e.g., exposure
33 routes and levels in the studies that were the primary drivers of these judgments], noting
34 that these conditions will be clarified during dose-response analysis (see Section 11],
This document is a draft for review purposes only and does not constitute Agency policy.
121 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
• Indications of potentially susceptible populations or lifestages (i.e., an integrated summary
of the available evidence on potential susceptible populations and lifestages drawn across
the syntheses of the human, animal, and mechanistic evidence]18.
• A summary of key assumptions used in the analysis, which are generally based on EPA
guidelines and which are largely captured in this protocol.
• Strengths and limitations of the evidence integration judgments, including key uncertainties
and data gaps, as well as the limitations of the systematic review. As noted in Section 4.2.2,
for one or more of these five PFAS assessments, characterization of the uncertainties in the
animal evidence is expected to include a discussion of the reliance on short-term oral
exposure studies in rats. Similarly, the characterization of uncertainty in the human
evidence is expected to include a discussion of potential confounding by PFAS other than
the PFAS of interest.
In short, the evidence integration narrative will present a qualitative summary of the
strength of each evidence stream and an overall judgment across all relevant evidence, with
exposure context provided. For each health effect or specific cancer type of potential concern, the
first sentence of the evidence integration narrative will include the summary judgment [see
description below for how these judgments help inform selection of a descriptor for carcinogenicity
(U.S. EPA. 2005a1], Assessments will also include an evidence profile table (see Table 18] to
support the evidence integration narrative by providing the major decisions and supporting
rationale. Table 20 describes the categories of evidence integration judgments that will be used in
these PFAS assessments and provides examples of database scenarios that fit each category of
evidence. These summary judgments provide a succinct and clear representation of the decisions
from the more detailed analyses of whether (or not] the evidence strength indicates that PFAS
exposure has the potential to cause the human health effect(s] under the necessary conditions of
exposure. Consistent with EPA non-cancer and cancer guidelines, a judgment that the evidence
supports an apparent lack of an effect of PFAS exposure on the health effect(s] will only be used
when the available data are considered robust for deciding that there is no basis for human hazard
concern; lesser levels of evidence suggesting a lack of an effect will be characterized as
"insufficient."
18One or more of these five PFAS assessments may include consideration of information outside of their
PFAS-specific database to address this aspect of the evidence integration narrative. These PFAS-specific data gaps
and uncertainties appear to extend beyond poorly studied health effects, and the discussion of missing
information on potential populations, sexes, or lifestages that are likely to be more susceptible to developing a
specific health effect may consider information from reviews of other PFAS.
This document is a draft for review purposes only and does not constitute Agency policy.
122 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 20. Evidence integration judgments for characterizing potential human health hazards in the evidence
integration narrative
Evidence
Integration
Judgment1
Evidence in Studies of Humans
Evidence in Animal Studies
Inferences Across Evidence Streams
Sufficient
evidence for
hazard
A judgment of sufficient evidence for hazard
evidence in animal studies, incorporating the
this judgment span a broad range of overall e
• Strong mechanistic evidence in well-
conducted studies of exposed humans
(medium or high confidence) or human cells
(including NAMs), in the absence of other
substantive data, where an informed
evaluation has determined that the data are
reliable for assessing toxicity relevant to
humans and the mechanistic events have
been causally linked to the development of
the health effect of interest3.
• A single high or medium confidence study
without supporting coherent evidence or
concern for unexplained inconsistency.
Specifically, there are no comparable
studies of similar confidence and sensitivity
providing conflicting evidence, or the
differences can be reasonably explained by,
e.g., the populations or exposure levels
studied (U.S. EPA, 2005a).
• Multiple studies showing generally
consistent findings, including at least one
high or medium confidence study and
supporting evidence, but with some residual
uncertainty due to potential chance, bias, or
confounding (e.g., effect estimates of low
magnitude or small effect sizes given what is
known about the endpoint; uninterpretable
requires th
considerat
vidence str
10
o
i_
LU
i_
(D
DO
C
o
1_
+-»
LO
V
at a scenario below is met for either the evider
ons outlined under inferences across evidence
ength and examples are provided below, starti
• Strong mechanistic evidence in well-
conducted studies of animals or animal
cells (including NAMs), in the absence of
other substantive data, where an
informed evaluation has determined the
assays are reliable for assessing toxicity
relevant to humans and the mechanistic
events have been causally linked to the
development of the health effect3.
• A single high or medium confidence
experiment in the absence of comparable
experiment(s) of similar confidence and
sensitivity providing conflicting evidence4
(evidence that cannot be reasonably
explained, e.g., by respective study
designs or differences in animal model;
(U.S. EPA, 2005a).
• At least one high or medium confidence
study with supporting information
increasing the strength of the evidence.
Although the results are largely
consistent, notable uncertainties remain.
However, in scenarios when inconsistent
evidence or evidence indicating
nonspecific effects exist, it is not judged to
reduce or discount the level of concern
regarding the positive findings, or it is not
ice in studies of humans OR
streams. The scenarios justifying
ng with the weakest evidence 2.
• Supplemental evidence (e.g.,
structure-activity data; chemical
class information; other NAMs) is
judged to increase the strength of
limited or near-equivocal, chemical-
specific human or animal evidence
to sufficient evidence for hazard.
• Coherent or biologically consistent
findings across evidence streams
increases the strength of limited or
near-equivocal human or animal
evidence (e.g., single or few high or
medium confidence studies with
some conflicting evidence) to
sufficient evidence for hazard.
• The strength of the evidence is
decreased because mechanistic
information (even if it does not
provide MOA understanding) raises
uncertainties regarding the human
and/or animal evidence, but overall
the evidence is still considered
strong enough to result in a
judgment of sufficient evidence for
hazard.
• The strength of the evidence is
decreased because findings across
This document is a draft for review purposes only and does not constitute Agency policy.
123 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
patterns with respect to exposure levels).
Alternatively, a single high or medium
confidence study with a large magnitude or
severity of the effect, a dose-response
gradient, or other factors that increase the
evidence strength, without serious residual
uncertainties. In both scenarios,
associations with related endpoints,
including mechanistic evidence from
exposed humans, can address uncertainties
relating to exposure response, temporality,
coherence, and biological plausibility, and
any conflicting evidence is not from a
comparable body of higher confidence,
sensitive studies4.
A set of high or medium confidence
independent studies reporting an
association between the exposure and the
health outcome, with reasonable
confidence that alternative explanations,
including chance, bias, and confounding, can
be ruled out across studies. The set of
studies is primarily consistent, with
reasonable explanations when results differ;
and an exposure response gradient is
demonstrated. Supporting evidence, such
as associations with biologically related
endpoints in human studies (coherence) or
large estimates of risk or severity of the
response, may help to rule out alternative
explanations. Similarly, mechanistic
evidence from exposed humans may serve
to address uncertainties relating to
exposure-response, temporality, coherence,
and biological plausibility (i.e., providing
evidence consistent with an explanation for
how exposure could cause the health effect
l/>
o
i_
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
based on current biological knowledge).
mechanistic evidence (e.g., precursor
events linked to adverse outcomes) in
animal models may exist to address
uncertainties in the evidence base.
Insufficient
evidence
A iudgment of insufficient evidence reauires that a scenario below is met for both the evidence in studies of humans AND evidence in
animal studies, incorporating the considerations outlined under inferences across evidence streams.
• A body of evidence, including scenarios with
one or more high or medium confidence
studies reporting an association between
exposure and the health outcome, where
either (1) conflicting evidence exists in
studies of similar confidence and sensitivity
4,5 OR (2) considerable methodological
uncertainties remain across the body of
evidence (typically related to exposure or
outcome ascertainment, including
temporality), AND there is no supporting
coherent evidence that increases the overall
evidence strength.
• A set of only low confidence studies.
• No studies of exposed humans or well-
conducted studies of human cells.
• A set of largely null studies that does not
meet a scenario for sufficient evidence to
judge that a hazard is unlikely.
• A body of evidence, including scenarios with one or
more high or medium confidence experiments
reporting effects but without supporting coherent
evidence that increases the overall evidence
strength, where conflicting evidence exists from a set
of sensitive experiments of similar or higher
confidence (can include mechanistic evidence)4-5.
• A set of only low confidence experiments.
• No animal studies or well-conducted studies of
animal cells.
• The available endpoints are not informative to the
hazard question under evaluation.
• A set of largely null studies that does not meet the
criteria for sufficient evidence to judge that a hazard
is unlikely.
• The evidence in animal studies
meets a scenario for sufficient
evidence for hazard, but strong
experimental evidence (e.g., a
MOA interpreted with reasonable
certainty) indicates the findings in
animals are unlikely to be relevant
to humans.
• The evidence meets a scenario for
sufficient evidence to judge that a
hazard is unlikely, but there is
inadequate testing of susceptible
populations and lifestages, or data
conflict across evidence streams.
• The evidence in animal studies
meets a scenario for sufficient
evidence to judge that a hazard is
unlikely, but the database lacks
experimental support that the
models are relevant to humans for
the effect of interest.
Sufficient
evidence to
judge that a
hazard is
unlikely 6
A iudgment of sufficient evidence to iudae that a hazard is unlikelv reauires that a scenario below is met for either the evidence in studies
of humans OR evidence in animal studies, incorporating the considerations outlined under inferences across evidence streams.
• Several high confidence studies showing null
results (for example, an odds ratio of 1.0),
ruling out alternative explanations including
chance, bias, and confounding with
reasonable confidence. Each of the studies
should have used an optimal outcome and
• A set of high confidence experiments examining a
reasonable spectrum of endpoints relevant to a type
of toxicity that demonstrate a lack of biologically
significant effects across multiple species, both sexes
(if applicable), and a broad range of exposure levels.
The data are compelling in that the experiments have
• There is adequate testing of
susceptible populations and
lifestages.
• When the evidence in animal
studies meets a scenario for this
This document is a draft for review purposes only and does not constitute Agency policy.
125 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
judgment, there is experimental
support that the models are
relevant to humans for the effect of
interest and no conflicting human
evidence exists.
• When the evidence in studies of
humans meets a scenario for this
judgment and conflicting animal
data exist, mechanistic information
indicates that the animal data are
unlikely to be relevant to humans.
• When multiple high confidence
animal experiments and studies in
humans indicate lack of an effect,
but the evidence does not meet a
scenario for sufficient evidence to
judge that a hazard is unlikely,
strong mechanistic evidence in
models relevant to humans
supports lack of an effect such that
the totality of evidence supports
this judgment.
incancer health effects (i.e., (U.S. EPA,
pplied herein, the process of 'evidence
integration' is synonymous with 'weight of evidence'). The evidence integration judgment for each assessed health effect will be included as part of an evidence
integration narrative, the specific documentation of the various expert decisions and evidence-based (or default) rationales are summarized in an evidence
profile table, and the judgement will be contextualized based on the primary supporting evidence (experimental model or observed population, and exposure
levels tested or estimated). Importantly, as discussed in Section 10.1, these judgments may be based on analyses of grouped outcomes at different levels of
granularity (e.g., motor activity versus neurobehavioral effects versus nervous system effects) depending on the specifics of the health effect evidence base.
Health effects characterized as having sufficient evidence for hazard will be evaluated for use in dose-response assessment.
2 Qualitative descriptions of differences in the strength of the evidence across different health effects judged as having sufficient evidence for hazard are useful
for other assessment decisions, including prioritizing outcomes in quantitative analyses and characterizing assessment uncertainties. Thus, for all evidence
scenarios, but particularly for those in the lower end of this range, it is important to characterize the uncertainties in the evidence base within the evidence
integration narrative and convey the evidence strength to subsequent steps, including toxicity values developed based on those effects. Existing guidance defines
the minimum evidence necessary to judge that a health hazard could exist as one adverse endpoint from one well-conducted study U.S. EPA (1998); this has
been expanded in this table to better incorporate mechanistic evidence, including new approach methods (NAMs) data.
exposure assessment and adequate sample
size (specifically for higher exposure groups
and for susceptible populations). The
overall set should include the full range of
levels of exposures that human beings are
known to encounter, and an evaluation of
an exposure-response gradient.
examined the range of scenarios across which health
effects in animals could be observed, and an
alternative explanation (e.g., inadequately controlled
features of the studies' experimental designs;
inadequate sample sizes) for the observed lack of
effects is not available. The experiments were
designed to specifically test for the effects of
interest, including suitable exposure timing and
duration, post- exposure latency, and endpoint
evaluation procedures.
1 These categories are based on those indicated for use in hazard characterization from the existing EPA guidelines for nc
1996a, 1991b, 1998)) and, as described in those guidance documents, they depend heavily on expert judgment (note: as a
This document is a draft for review purposes only and does not constitute Agency policy.
126 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
3 Scientific understanding of toxicity mechanisms and of the human implications of new toxicity testing methods (e.g., from high-throughput screening, from
short-term in vivo testing of alternative species, or from new in vitro and in silico testing and other NAMs) will continue to increase. Thus, the sufficiency of
mechanistic evidence alone for identifying potential human health hazards is expected to increase as the science evolves. The decision to identify a potential
human hazard based on these data is an expert judgment dependent on the state-of-the-science at the time of review.
4 Scenarios with unexplained heterogeneity across sets of studies with similar confidence and sensitivity can be considered either sufficient evidence for hazard
or insufficient evidence, depending on the expert judgment of the overall weight of evidence. Specifically, this judgment considers the level of support (or lack
thereof) provided by evaluations of the magnitude or severity of the effects, coherence of related findings (including mechanistic evidence), dose-response,
and biological plausibility, as well as the comparability of the supporting and conflicting evidence (e.g., the specific endpoints tested, or the methods used to
test them; the specific sources of bias or insensitivity in the respective sets of studies). The evidence-specific factors supporting either evidence integration
judgment will be clearly articulated in the evidence integration narrative.
5 When the database includes at least one well-conducted study and a hazard characterization judgment of insufficient evidence is drawn, quantitative analyses
may still be useful for some purposes (e.g., providing a sense of the magnitude and uncertainty of estimates for health effects of potential concern, ranking
potential hazards, or setting research priorities), but not for others (see related discussions in (U.S. EPA, 2005a)). It is critical to transparently convey the
extreme uncertainty in any such estimates.
6 The criteria for this category are intentionally more stringent than those justifying a conclusion of sufficient evidence for hazard, consistent with the "difficulty
of proving a negative" (as discussed in (U.S. EPA, 1996a, 1991b, 1998)).
This document is a draft for review purposes only and does not constitute Agency policy.
127 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
For evaluations of carcinogenicity, consistent with EPA's Cancer Guidelines fU.S. EPA.
2005a], one of EPA's standardized cancer descriptors will be used as a shorthand characterization
of the evidence integration narrative, describing the overall potential for human carcinogenicity
across all potential cancer types. These are: (1] carcinogenic to humans, (2] likely to be carcinogenic
to humans, (3] suggestive evidence of carcinogenic potential, (4] inadequate information to assess
carcinogenic potential, or (5] not likely to be carcinogenic to humans. More than one descriptor can
be used when a chemical's effects differ by exposure level or route (U.S. EPA. 2005a1: if the
database supports such an analysis, these decisions will be clarified based on a more thorough
review of the mechanistic evidence or more detailed dose-response analysis (see Section 11). In
some cases, mutagenicity will also be evaluated (e.g., when there is evidence of carcinogenicity),
because it influences the approach to dose-response assessment and subsequent application of
adjustment factors for exposures early in life (U.S. EPA. 2005a. b).
An appropriate cancer descriptor will be selected as described in EPA Cancer Guidelines
(U.S. EPA. 2005al For each cancer subtype, an evidence integration narrative and summary
judgment will be provided, as described above. The cancer descriptor will consider the
interrelatedness of cancer types potentially related to PFAS exposure, consistency across the
human and animal evidence for any cancer type [noting that site concordance is not required (U.S.
EPA. 2005a)]. and the uncertainties associated with each assessment-specific conclusion. In
general, however, if a systematic review of more than one cancer type was conducted, then the
overall judgment and discussion of evidence strength in the evidence integration narrative for the
cancer type(s) with the strongest evidence for hazard will be used to inform selection of the cancer
descriptor, with each assessment providing a transparent description of the decision rationale. The
cancer descriptor and evidence integration narrative for potential carcinogenicity, including
application of the MOA framework, will consider the conditions of carcinogenicity, including
exposure (e.g., route; level) and susceptibility (e.g., genetics; lifestage), as the data allow fFarland.
2005: U.S. EPA. 2005a. bl
10.3. HAZARD CONSIDERATIONS FOR DOSE-RESPONSE
This section outlines how these assessments will consider and describe the transition from
hazard identification to dose-response analysis, highlighting (1) information that will inform the
selection of outcomes or broader health effect categories for which toxicity values will be derived,
(2) whether toxicity values can be derived to protect specific populations or lifestages, (3) how
dose-response modeling will be informed by toxicokinetic information, and (4) information aiding
the identification of biologically based benchmark response (BMR) levels. The pool of outcomes
and study-specific endpoints will be discussed to identify which categories of effects and study
designs are considered the strongest and most appropriate for quantitative assessment of a given
health effect. Health effects that were analyzed in human studies in relation to exposure levels
within or closer to the range of exposures encountered in the environment will be considered
This document is a draft for review purposes only and does not constitute Agency policy.
128 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
particularly informative, as are animal studies testing a broad range of exposure levels and
including levels in the lower dose region. When there are multiple endpoints for an organ/system,
considerations for characterizing the overall impact on this organ/system will be discussed,
including the severity and longevity of the effects. For example, if there are multiple
histopathological alterations relevant to liver function changes, liver necrosis maybe selected as
the most representative endpoint to consider for dose-response analysis. This section may review
or clarify which endpoints or combination of endpoints in each organ/system characterize the
overall effect for dose-response analysis. For cancer types, consideration will be given to the
overall risk of multiple types of tumors. Multiple tumor types (if applicable] will be discussed, and
a rationale given for any grouping.
Biological considerations that are important for dose-response analysis (e.g., that could help
with selection of a BMR) will be discussed. The impact of route of exposure on toxicity to different
organs/systems will be examined, if appropriate. The existence and validity of PBPK models or
toxicokinetic information that may allow the estimation of internal dose for route-to-route
extrapolation will be presented (see additional discussion and decision points in Section 11.2). In
addition, mechanistic evidence analyses that will influence the dose-response analyses will be
highlighted (see Section 9.2 for specific considerations), for example, evidence related to
susceptibility or potential shape of the dose-response curve.
This section will also describe the evidence regarding populations and lifestages that
appear to be susceptible to the health hazards identified and factors that are likely to increase the
risk of developing (or exacerbating) these health effects, depending on the available evidence. This
section will include this discussion even if there are no specific data on the effects of exposure to
the PFAS of interest in the potentially susceptible population. Table 16 in Section 9 outlines some
of the specific factors that will be considered for discussion and summaries of the evidence with
respect to patterns across studies pertinent to consistency, coherence, and the magnitude and
direction of effect measures. At a minimum, consideration will be given to discussion of
information relevant to infants and children, pregnant women, and women of childbearing age.
The section will consider options for using susceptible population data in the dose-response
analysis. In particular, an attempt will be made to highlight where it might be possible to develop
separate risk estimates for a specific population or lifestage or to determine whether evidence is
available to select a data-derived uncertainty factor.
This document is a draft for review purposes only and does not constitute Agency policy.
129 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
11. DOSE-RESPONSE ASSESSMENT: SELECTING
STUDIES AND QUANTITATIVE ANALYSIS
The previous sections of this protocol describe how systematic review principles will be
applied to evaluate studies (for potential bias and sensitivity] and reach evidence integration
conclusions on potential human health effects associated with exposure to the PFAS of interest.
Selection of specific data sets for dose-response assessment and performance of the dose-response
assessment will be conducted after hazard identification is complete and involves database- and
chemical-specific biological judgments that build from decisions made at earlier stages of
assessment development. A number of Environmental Protection Agency (EPA] guidance and
support documents detail data requirements and other considerations for dose-response modeling,
especially EPA's Benchmark Dose Technical Guidance (U.S. EPA. 20121 EPA's Review of the Reference
Dose and Reference Concentration Processes fU.S. EPA. 20021 Guidelines for Carcinogen Risk
Assessment fU.S. EPA. 2005al and Supplemental Guidance for Assessing Susceptibility from Early-Life
Exposure to Carcinogens (U.S. EPA. 2005bl This section of the protocol provides an overview of
considerations for conducting the dose-response assessment, particularly statistical considerations
specific to dose-response analysis that support quantitative risk assessment. Importantly, these
considerations do not supersede existing EPA guidance.
Dose-response assessments will be performed for both noncancer and cancer health
hazards, and for both oral and inhalation routes of exposure following exposure19 to the chemical of
interest, if supported by existing data. For noncancer hazards, an oral reference dose (RfD] and/or
an inhalation reference concentration (RfC] will be derived when possible. An RfD or an RfC is an
estimate, with uncertainty spanning perhaps an order of magnitude, of an exposure to the human
population (including susceptible subgroups] that is likely to be without an appreciable risk of
deleterious health effects over a lifetime (U.S. EPA. 20021 These health effects may also include
cancer effects [e.g., in a case where a nonlinear MOA is concluded that indicates a key precursor
event necessary for carcinogenicity does not occur below a specific exposure level fU.S. EPA.
2005a]; see Section 11.2.3], Reference values are not predictive risk values; that is, they provide no
information about risks at higher or lower exposure levels.
When low-dose linear extrapolation for cancer effects is supported, particularly for
chemicals with direct mutagenic activity or those for which the data indicate a linear component
19For most health outcomes (e.g., this would not apply to outcomes related to developmental toxicity),
dose-response assessments will be preferably based on studies of chronic exposure. However, analyses will also
be conducted for shorter durations, particularly when the evidence base for a PFAS indicates potential risks
associated with shorter exposures to the chemical (U.S. EPA, 2002).
This document is a draft for review purposes only and does not constitute Agency policy.
130 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
below the point of departure (POD], an oral slope factor (OSF] and/or an inhalation unit risk (IUR]
will be used to estimate human cancer risks. In general, this will also be the case when no data are
available to inform the evaluation of linearity. An OSF is a plausible upper bound lifetime cancer
risk from chronic ingestion of a chemical per unit of mass consumed per unit body weight per day
(mg/kg-day], An IUR is a plausible upper bound lifetime cancer risk from chronic inhalation of a
chemical per unit of air concentration (expressed as ppm or [ig/m3]. In contrast with reference
values (RfVs], an OSF or IUR can be used in conjunction with exposure information to predict
cancer risk at a given dose.
As discussed in Section 2 "Scoping and Problem Formulation Summary" for these PFAS
assessments, Integrated Risk Information System (IRIS] will conduct the assessments with a goal of
developing any toxicity values that are reasonably supported by the available data, based on
judgments of the evidence drawn during hazard identification and the suitability of studies for
dose-response analysis.
The derivation of reference values and cancer risk estimates will depend on the nature of
the health hazard conclusions drawn during evidence integration (see Section 10.2], Specifically,
EPA generally conducts dose-response assessments and derives cancer values for chemicals that
are classified as carcinogenic or likely to be carcinogenic to humans. When there is suggestive
evidence of carcinogenicity to humans, EPA generally would not conduct a dose-response
assessment or derive a cancer value except when the evidence includes a well-conducted study and
quantitative analyses may be useful for some purposes, for example, providing a sense of the
magnitude and uncertainty of potential risks, ranking potential hazards, or setting research
priorities fU.S. EPA. 2005al A parallel approach will be taken for potential noncancer health effects
in these assessments. Specifically, for noncancer outcomes these assessments will generally
include dose-response assessments when the evidence integration judgments indicate there is
"sufficient evidence for hazard," with preference given to health effects with stronger evidence
scenarios within that category (Section 10.2], and quantitative analyses generally will not be
attempted for "insufficient evidence."
11.1. SELECTING STUDIES FOR DOSE-RESPONSE ASSESSMENT
The dose-response assessment will begin with a review of the important health effects
highlighted during hazard identification, particularly among the studies of highest quality and that
exemplify the study attributes summarized in Table 23. This review will also consider whether
there are opportunities for quantitative evidence integration, although it is considered unlikely that
the data available to do so will be available for these assessments based on the preliminary
literature inventory. Examples of quantitative integration, from simplest to more complex, include
(1] the combination of results for an outcome across sex (within a study]; (2] characterizing overall
toxicity, as in combining effects that comprise a syndrome, or occur on a continuum
(e.g., precursors and overt toxicity, benign tumors that progress to malignant tumors]; and
This document is a draft for review purposes only and does not constitute Agency policy.
131 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
(3] meta-analysis or meta-regression of all studies addressing a category of important health
effects.
Some studies that were used qualitatively for hazard identification may or may not be
considered useful quantitatively for dose-response analysis in these five assessments because of
factors like the lack of quantitative measures of exposure or of variability measures for response
data. If the needed information cannot be located (e.g., by contacting study authors and making any
information publicly available], a semiquantitative analysis (e.g., via no-observed-adverse-effect
level [NOAEL]/lowest-observed-adverse-effect level [LOAEL]] will be considered. Studies of low
sensitivity may be considered less useful if they failed to detect an effect or reported points of
departure with wide confidence limits, but such studies will still be considered for inclusion in a
meta-analysis.
Among the studies that support the evidence integration conclusions, those that are most
useful for dose-response analysis will generally have at least one exposure level in the region of the
dose-response curve near the benchmark response (the response level to be used for deriving
toxicity values] to minimize low-dose extrapolation. Such studies will also have more exposure
levels and larger sample sizes overall (U.S. EPA. 20121 These attributes support a more complete
characterization of the shape of the exposure-response curve and decrease the uncertainty in the
associated exposure-response metric (e.g., IUR or RfC] by reducing statistical uncertainty in the
POD and minimizing the need for low-dose extrapolation. In addition to these more general
considerations, specific issues that may be considered for their potential to impact the feasibility of
dose-response modeling for individual data sets are described in more detail in the Benchmark Dose
Technical Guidance fU.S. EPA. 20121
This document is a draft for review purposes only and does not constitute Agency policy.
132 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table 21. Attributes used to evaluate studies for deriving toxicity values
Study attributes
Considerations
Human studies
Animal studies
Rationale for choice of
species
Human data are preferred over animal data to
eliminate interspecies extrapolation uncertainties
(e.g., in toxicodynamics, relevance of specific health
outcomes to humans).
Animal studies provide supporting evidence when adequate human
studies are available and are considered principal studies when
adequate human studies are not available. For some hazards, studies
of animal species known to respond similarly to humans would be
preferred over studies of other species.
Relevance of
exposure
paradigm
Exposure
route
Studies involving human environmental exposures
(oral, inhalation).
Studies by a route of administration relevant to human
environmental exposure are preferred. A validated toxicokinetic
model can also be used to extrapolate across exposure routes.
Exposure
durations
When developing a chronic toxicity value, chronic- or subchronic-duration studies are preferred over studies of acute exposure.
Exceptions exist, such as when a susceptible population or lifestage is more sensitive in a certain time window
(e.g., developmental exposure).
Exposure
levels
Exposures near the range of typical environmental human exposures are preferred. Studies with a broad exposure range and
multiple exposure levels are preferred to the extent that they can provide information about the shape of the
exoosure-resDonse relationship [see the EPA Benchmark Dose Technical Guidance; (U.S. EPA, 2012)1 and facilitate extrapolation
to more relevant (generally lower) exposures.
Subject selection
Studies that provide risk estimates in the most susceptible groups are preferred.
Controls for possible
confounding3
Studies with a design (e.g., matching procedures, blocking) or analysis (e.g., covariates or other procedures for statistical
adjustment) that adequately address the relevant sources of potential critical confounding for a given outcome are preferred.
Measurement of exposure
Studies that can reliably distinguish between levels of
exposure in a time window considered most relevant
for a causal effect with respect to the development of
the outcome are preferred. Exposure assessment
methods that reduce measurement error and methods
that provide measurement of exposure at the level of
the individual are preferred. Measurements of
exposure should not be influenced by knowledge of
health outcome status.
Studies providing actual measurements of exposure (e.g., analytical
inhalation concentrations versus target concentrations) are
preferred. Relevant internal dose measures may facilitate
extrapolation to humans, as would availability of a suitable animal
PBPK model in conjunction with an animal study reported in terms of
administered exposure.
This document is a draft for review purposes only and does not constitute Agency policy.
133 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Study attributes
Considerations
Human studies
Animal studies
Measurement of health
outcome(s)
Studies that can reliably distinguish the presence or absence (or degree of severity) of the outcome are preferred. Outcome
ascertainment methods using generally accepted, standardized approaches) are preferred.
Studies with individual data are preferred in general. Examples include characterizing experimental variability more realistically
and characterizing overall incidence of individuals affected by related outcomes (e.g., phthalate syndrome).
Study size and design
Preference is given to studies using designs reasonably expected to have power to detect responses of suitable magnitude.15
This does not mean that studies with substantial responses but low power would be ignored, but that they should be
interpreted in the context of a confidence interval or variance for the response. Studies that address changes in the number at
risk (through decreased survival, loss to follow-up) are preferred.
aAn exposure or other variable that is associated with both exposure and outcome but is not an intermediary between the two.
bPower is an attribute of the design and population parameters, based on a concept of repeatedly sampling a population; it cannot be inferred post hoc using
data from one experiment (Hoenig and Heisev, 2001).
This document is a draft for review purposes only and does not constitute Agency policy.
134 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
11.2. CONDUCTING DOSE-RESPONSE ASSESSMENTS
Consistent with EPA practice, these PFAS assessments will apply a two-step approach for
dose-response assessment that distinguishes analysis of the dose-response data in the range of
observation from any inferences about responses at lower environmentally relevant exposure
levels flJ.S. EPA. 2012. 2005a]:
1) Within the observed dose range, the preferred approach will be to use dose-response
modeling to incorporate as much of the data set as possible into the analysis. This modeling
to derive a POD should include an exposure level ideally near the lower end of the range of
observation, without significant extrapolation to lower exposure levels (see Section 11.2.1
for more details],
2) As derivation of cancer risk estimates and reference values nearly always involves
extrapolation to exposures lower than the POD; the approaches to be applied in these
assessments are described in more detail in Section 11.2.2 and Section 11.2.3, respectively.
When sufficient and appropriate human and laboratory animal data are available for the
same outcome, human data will be generally preferred for the dose-response assessment because
use of human data eliminates the need to perform interspecies extrapolations.
For reference values, these assessments will typically derive a candidate value from each
suitable data set, whether in humans or animals (see Section 11.1]. Evaluation of these candidate
values grouped within a given organ/system will yield a single organ/system-specific value for each
organ/system under consideration. Next, evaluation of these organ/system-specific values will
result in the selection of a single overall reference value to cover all health outcomes across all
organs/systems. While this overall reference value represents the focus of these dose-response
assessments, the organ/system-specific values can be useful for subsequent cumulative risk
assessments that consider the combined effect of multiple PFAS (or other agents] acting at a
common organ/system.
For cancer, if there are multiple tumor sites that can be quantified individually, the final
cancer risk estimate(s] will typically address overall cancer risk, to the extent the data allow.
For both cancer and noncancer toxicity values, uncertainties in these estimates will be
transparently characterized and discussed.
11.2.1. Dose-Response Analysis in the Range of Observation
Toxicodynamic ("biologically based"] modeling is generally preferred when there are
sufficient, reliable data to ascertain the MOA and quantitatively support model parameters that
represent rates and other quantities associated with the key precursor events of the MOA. Such
data, however, do not appear to be available for these five PFAS.
This document is a draft for review purposes only and does not constitute Agency policy.
135 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Because a toxicodynamic model will not be available for dose-response assessment,
empirical modeling will be used to fit the data (on the apical outcome or a key precursor event] in
the range of observation. For this purpose, EPA has developed a standard set of models
fhttp://www.epa.gov/ncea/bmdsl that can be applied to typical data sets, including those that are
nonlinear. In situations where there are alternative models with significant biological support, the
decision maker will be informed by the presentation of these alternatives in the assessments]
along with the models' strengths and uncertainties. EPA has developed guidance on modeling
dose-response data, assessing model fit, selecting suitable models, and reporting modeling results
[see the EPA Benchmark Dose Technical Guidance; (U.S. EPA. 20121], Additional judgment or
alternative analyses will be used if the procedure fails to yield reliable results; for example, if the fit
is poor, modeling may be restricted to the lower doses, especially if there is competing toxicity at
higher doses.
For each modeled response, a POD from the observed data will be estimated to mark the
beginning of extrapolation to lower doses. The POD is an estimated dose (expressed in
human-equivalent terms] near the lower end of the observed range without significant
extrapolation to lower doses. The POD will be used as the starting point for subsequent
extrapolations and analyses. For linear extrapolation of cancer risk, the POD will be used to
calculate an OSF or IUR, and for nonlinear extrapolation the POD will be used in the calculation of
an RfD or RfC.
The response level at which the POD is calculated will be guided by the severity of the
endpoint. If linear extrapolation is used, standard values near the low end of the observable range
will generally be used (for example, 10% extra risk for cancer bioassay data, 1% for epidemiologic
data, lower for rare cancers]. For nonlinear approaches, both statistical and biological significance
will be considered. For dichotomous data, a response level of 10% extra risk will generally be used
for minimally adverse effects, 5% or lower for more severe effects. For continuous data, a response
level ideally will be based on an established definition of biologic significance. In the absence of
such definition, one control standard deviation from the control mean will generally be used for
minimally adverse effects, and one-half standard deviation for more severe effects. The point of
departure will be the 95% lower bound on the dose associated with the selected response level.
EPA has developed standard approaches to determine the relevant dose for use in
dose-response modeling in the absence of appropriate toxicokinetic modeling. These standard
approaches can also aide comparison across exposure patterns and species in the absence of a
validated pharmacokinetic (PK] model (see below]. The general approaches and considerations to
be used to extrapolate PFAS dosimetry from (1] shorter to longer durations within studies, (2] from
animals to humans, and (3] across routes of exposure are outlined below:
• Intermittent study exposures will be standardized to a daily average over the duration of
exposure. For chronic effects, daily exposures will be averaged over the life span.
This document is a draft for review purposes only and does not constitute Agency policy.
136 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Exposures during a critical period, however, will not be averaged over a longer duration
fU.S. EPA. 2005a. 1991al Note that this will typically be done after modeling because the
conversion is linear.
• The preferred approach for dosimetry extrapolation from animals to humans will be
through PBPK or PK modeling. This approach will be considered first for PFAS and
lifestages with existing PBPK models or where an existing model structure can be readily
adapted (see Section 6.4 on PBPK modeling],
• Because PK data for the PFAS being evaluated exist in at least one relevant animal species
(rats or monkeys] and in humans (see Section 2.4.1], a data-informed extrapolation
approach will also be considered for any PFAS that either lacks a PK model or has a model
determined to be of inadequate quality. Briefly, the ratio of the elimination half-life in
animal to that in the human, T0.5,a:T0.5,h, or the ratio of the clearance in the human to the
animal, CLH:CLA, will be considered for use in converting an oral dose-rate in animals
(mg/kg/day] to a human equivalent dose rate (i.e., the human exposure that should yield
the same blood concentration as the animal exposure from which it is being extrapolated].
Note that clearance and half-life are inversely related. The assessments will consider these
metrics as follows:
° Of these two metrics, T0.5 and CL, the half-life is a less complete measure of elimination
but one that can be evaluated from more minimal PK data. A half-life can be estimated
by observing the decline in an individual's blood concentration of a compound after an
exposure has ended. In this way, the total exposure or body burden of the chemical
does not have to be known. However, PFAS elimination may go through several phases
during which distinct half-lives apply, and the blood concentration that occurs during
ongoing exposure may effectively reflect an average among these. The specific
approaches and considerations for estimating PFAS half-life are outlined in
Section 9.2.1.
0 The clearance, on the other hand, is a measure of average elimination but requires more
data to estimate. One must also quantify a companion variable, the volume of
distribution (Vd], which in turn requires a measure of total exposure or dose in
well-conducted studies. Although more rigorous assessment-specific evaluations will
be performed, based on a preliminary review of studies in the literature inventory, the
data necessary for the reliable quantification of Vd are expected to be lacking.
Specifically, it does not appear that accurate estimates of dose are available in human
exposure studies, and the identified animal studies demonstrate considerable
interstudy variability in Vd estimates.
• Based on the selection of half-life as the preferred metric and a POD identified from a
health-effects study in animals, the human equivalent dose (HED] will be calculated as:
HED = (To.5,a[s]/To.5,h[s]] x POD
0 Here, the [s] in the subscript indicates that the value may be sex specific. When there
are sex-specific values (significant differences between males and females] in both
animals and humans, the T0.5 values for females would be used to extrapolate health
effects in female animals to women, the T0.5 values for males used to extrapolate male
animal health effects to men. If human data are available to estimate separate half-lives
This document is a draft for review purposes only and does not constitute Agency policy.
137 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
1 for women and men, the T0.5 for women will likewise be used to estimate HED values in
2 women and the T0.5 in men used to estimate HEDs in men. If human data are not
3 sufficient to provide distinct values for men and women, a common T0.5 for humans will
4 be used.
5 • In the absence of PK data/half-lives, oral doses will be scaled allometrically using
6 mg/kg3/4 day as the equivalent dose metric across species. Allometric scaling pertains to
7 equivalence across species, not across lifestages, and will not be used to scale doses from
8 adult humans or mature animals to infants or children fU.S. EPA. 2011a. 2005a. 19941
9 Using this approach, the HED will be calculated as:
10 HED = (BWh/BWa]0 25 x POD (mg/kg-day]
11 • Inhalation exposures will be scaled using dosimetry models that apply species-specific
12 physiologic and anatomic factors and consider whether the effect occurs at the site of first
13 contact or after systemic circulation (U.S. EPA. 2012.19941
14 • It can be informative to convert doses across exposure routes. If this is done, the
15 assessment will describe the underlying data, algorithms, and assumptions fU.S. EPA.
16 2005al Depending on the availability of sufficient data (see Section 9.2] and/or suitable
17 models (see Section 6.4], route-to-route extrapolations in these assessments will be
18 accomplished by using the inhalation exposure rates for PFAS-containing particles
19 predicted using the MPPD model (see Section 9.2] as an ingestion rate in the PK analysis
20 (PBPK/PK model or AD ME adjustment], under the assumption that once absorbed into
21 general circulation, the toxic effect is only a function of the body burden or blood
22 concentration.
23 • In the absence of study-specific data on, for example, intake rates or body weight, the EPA
24 has developed recommended values for use in dose-response analysis (U.S. EPA. 19881
25
11.2.2. Extrapolation: Slope Factors and Unit Risk
26 An OSF or IUR will be used to estimate human cancer risks when low-dose linear
27 extrapolation for cancer effects is supported by the PFAS-specific evidence, particularly for PFAS
28 with direct mutagenic activity or those for which the data indicate a linear component below the
29 POD. Low-dose linear extrapolation will also be used as a default when the data are insufficient to
30 establish the MOA (U.S. EPA. 2005al If the PF AS-specific data are sufficient to ascertain that one or
31 more modes of action are consistent with low-dose nonlinearity, or to support their biological
32 plausibility, low-dose extrapolation will use the reference-value approach when suitable data are
33 available fU.S. EPA. 2005al see Section 11.2.3 below.
34 Differences in susceptibility will be considered for use in deriving multiple slope factors or
35 unit risks, with separate estimates for susceptible populations and lifestages (U.S. EPA. 2005al If
36 appropriate chemical-specific data on susceptibility from early life exposures are available, then
37 these data will be used to develop cancer slope factors or unit risks that specifically address any
38 potential for differential potency in early lifestages (Farland. 2005: U.S. EPA. 2005al If such data
This document is a draft for review purposes only and does not constitute Agency policy.
138 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
are not available, the evidence integration analyses supports a mutagenic MOA for carcinogenicity,
and the extrapolation approach is linear, the dose-response assessment will indicate to decision
makers thatin the development of risk estimates, the default age-dependent adjustment factors
should be used with the cancer slope factor or unit risk and age-specific estimates of exposure fU.S.
EPA. 2005a. b], In this scenario, the final cancer risk value presented in the assessments] will
reflect this adjustment, with the requisite calculations provided.
The derivation of an OSF and IUR for any of these five PFAS conducted as part of the current
assessments will be performed in a manner consistent with EPA guidance.
11.2.3. Extrapolation: Reference Values
Reference value derivation is EPA's most frequently used type of nonlinear extrapolation
method, and it will be used in these PFAS assessments for noncancer effects. This approach will
also be used for cancer effects if the available data are sufficient to ascertain the MOA and conclude
that it is not linear at low doses (see Section 11.2.2], In this case, reference values for each relevant
route of exposure will be developed following EPA's established practices (U.S. EPA. 2005al in
general, the reference value will be based not on tumor incidence, but on a key precursor event in
the MOA that is necessary for tumor formation. The derivation of an RfD or RfC (if feasible]
conducted as part of the assessments for perfluorobutanoic acid (PFBA], PFHxA,
perfluorohexanesulfonate (PFHxS], perfluorononanoic acid (PFNA], and perfluorodecanoic acid
(PFDA] will be performed in a manner consistent with EPA guidance.
For each data set selected, reference values will be estimated by applying relevant
adjustments (i.e., uncertainty factors [UFs]] to the PODs to account for the conditions of the
reference value definition. These factors account for human variation, extrapolation from animals
to humans, extrapolation to chronic exposure duration, extrapolation to a minimal level of risk (if
not observed in the data set], and database deficiencies, as outlined below. Increasingly, data-based
adjustments (U.S. EPA. 2014cl probabilistic approaches (Chiu etal.. 2018: Chiu and Slob. 20151
and Bayesian methods for characterizing population variability (NAS. 20141 are becoming feasible
and may be distinguished from the UF considerations outlined below, if such data exist for these
five PFAS. These assessments will discuss the scientific bases (or lack thereof) for each selected UF,
including any data-based adjustments based on the following considerations.
• Animal-to-human extrapolation: If animal results are used to make inferences about
humans, the reference value derivation will incorporate the potential for cross-species
differences, which may arise from differences in toxicokinetics or toxicodynamics. The POD
will be standardized to equivalent human terms or be based on toxicokinetic or dosimetry
modeling, that may range from detailed chemical-specific to default approaches (U.S. EPA.
2014c. 2011a], and a factor of 10°5 (rounded to 3] will be applied to account for the
remaining uncertainty involving toxicokinetic and toxicodynamic differences. Data-derived
adjustments for toxicodynamic differences across species may include qualitative decisions
This document is a draft for review purposes only and does not constitute Agency policy.
139 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
regarding key science issues (e.g., if, during evaluation of PPARa-dependency, it is
concluded that humans are not more sensitive than rodents],
• Human variation: The assessments will account for variation in susceptibility across the
human population and the possibility that the available data may not represent individuals
who are most susceptible to the effect. If appropriate data or models for the effect or for
characterizing the internal dose are available, the potential for data-based adjustments for
toxicodynamics or toxicokinetics will also be considered fU.S. EPA. 2014c. 2002120'21 When
sufficient data are available, an intraspecies UF either less than or greater than 10-fold may
be justified (U.S. EPA. 20021 A reduction in this UF will be considered if the POD is derived
from or adjusted specifically for susceptible individuals, but not for a general population
that includes both susceptible and non-susceptible individuals (U.S. EPA. 2002. 1998.
1996a. 1994.1991a], In general, when the use of such data or modeling is not supported, a
UF with a default value of 10 will be used.
• LOAEL to NOAEL: When a POD is based on a LOAEL, the assessment will include an
adjustment to an exposure level where such effects are not expected. This can be a matter
of great uncertainty if no evidence is available at lower exposures. A factor of 3 or 10 will
generally be applied to extrapolate to a lower exposure expected to be without appreciable
effects. A factor other than 10 may also be considered, depending on the magnitude and
nature of the response and the shape of the dose-response curve (U.S. EPA. 2002.1998.
1996a, 1994,1991a],
• Subchronic-to-chronic exposure: When using studies of less-than-chronic exposure to make
inferences about chronic/lifetime exposure, the assessment will consider whether lifetime
exposure could reasonably be interpreted to result in effects at lower levels of exposure,
including consideration of the specific health outcome(s] in question. A factor of up to 10
will be considered, depending on the duration of the studies and the nature of the response
flJ.S. EPA. 2002. 1998. 19941
• Database deficiencies: In addition to the adjustments above, if database deficiencies raise
concern that further studies might identify a more sensitive effect, organ system, or
lifestage, the assessment will apply a database UF (U.S. EPA. 2002.1998.1996a. 1994.
1991a], The size of the factor will depend on the nature of the database deficiency. For
example, the EPA typically follows the recommendation that a factor of 10 be applied if both
a prenatal toxicity study and a two-generation reproduction study are missing and a factor
of 10°5 (i.e., 3] if either one or the other is missing fU.S. EPA. 20021 As noted in
Section 2.4.5, the evaluation of database completeness for these five PFAS will also consider
existing knowledge gained through reviewing other, potentially similar, PFAS to identify
data gaps. For example, given the potential for exposure to PFAS to cause developmental
effects (based on reviews of perfluorooctanoic acid [PFOA] and perfluorooctane sulfonate
20Examples of adjusting the toxicokinetic portion of interhuman variability include the Integrated Risk Information
System (IRIS) boron assessment's use of nonchemical-specific kinetic data [e.g., glomerular filtration rate in
pregnant humans as a surrogate for boron clearance (U.S. EPA, 2004)1 and the IRIS trichloroethylene assessment's
use of population variability in trichloroethylene metabolism, via a PBPK model, to estimate the lower 1st
percentile of the dose metric distribution for each POD (U.S. EPA, 2011b).
21Note that when a PBPK model is available for relating human internal dose to environmental exposure, relevant
portions of this UF may be more usefully applied prior to animal-to-human extrapolation, depending on the
correspondence of any nonlinearities (e.g., saturation levels) between species.
This document is a draft for review purposes only and does not constitute Agency policy.
140 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
1 [PFOS]] and the lack of such studies for PFHxA, consideration of the potential for PFHxA
2 exposure to cause developmental effects might review knowledge gained through the
3 assessment of the other C6 PFAS, PFHxS, or the other short-chain perfluoroalkyl carboxylic
4 acid, perfluorobutanoic acid (PFBA], In such cases, an interpretation of the relatedness
5 between the PFAS of interest and the PFAS used for comparison will inform selection of the
6 factor.
7 The POD for a particular RfV will be divided by the product of these factors. Based on the
8 RfD/RfC Review fU.S. EPA. 20021 recommendation that any composite factor that exceeds 3,000
9 represents excessive uncertainty, values with >3,000 UFC will not be used to derive RfVs. An
10 RfD/RfC may be based on the POD for a single endpoint within a study, or on a collection of related
11 PODs within or across studies, if such biological relationships are substantiated by the evidence.
This document is a draft for review purposes only and does not constitute Agency policy.
141 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
12. PROTOCOL HISTORY
This section is a placeholder for tracking information on the original protocol release and
any potential protocol updates.
This document is a draft for review purposes only and does not constitute Agency policy.
142 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
REFERENCES
Abbott. BP. (2009], Review of the expression of peroxisome proliferator-activated receptors alpha
(PPAR alpha], beta (PPAR beta], and gamma (PPAR gamma] in rodent and human
development [Review], Reprod Toxicol 27: 246-257.
http://dx.doi.0rg/lO.lOl6/i.reprotox.2OO8.lO.OOl
AKDEC (Alaska Department of Environmental Conservation]. (2018], Technical memorandum:
action levels for PFAs in water and guidance on sampling groundwater and drinking water.
https://dec.alaska.gov/spar/csp/sites/north-pole-refinerv/documents/pfas-fact-sheet-
l/#axn
ATSDR (Agency for Toxic Substances and Disease Registry], (2018], Toxicological profile for
perfluoroalkyls. Draft for public comment. Atlanta, GA: U.S. Department of Health and
Human Services, Centers for Disease Control and Prevention.
https://www.atsdr.cdc.gov/toxprofiles/tp200.pdf
Baduel. C: Paxman. CI: Mueller. IF. (2015], Perfluoroalkyl substances in a firefighting training
ground (FTG], distribution and potential future release. J Hazard Mater 296: 46-53.
http://dx.doi.Org/10.1016/i.ihazmat.2015.03.007
Barber. TL: Berger. U: Chaemfa. C: Huber. S: Tahnke. A: Temme. C: Tones. KC. (2007], Analysis of per-
and polyfluorinated alkyl substances in air samples from Northwest Europe. J Environ
Monit9: 530-541. http://dx.d0i.0rg/l0.1039/h701417a
Buck. RC: Franklin. 1: Berger. U: Conder. TM: Cousins. IT: de Voogt. P: Tensen. AA: Kannan. K: Mabuiy.
SA: van Leeuwen. SP. (2011], Perfluoroalkyl and polyfluoroalkyl substances in the
environment: terminology, classification, and origins [Review], Integr Environ Assess
Manag 7: 513-541. http://dx.doi.org/10.1002/ieam.258
Burri. L: Thoresen. GH: Berge. RK. (2010], The Role of PPARa Activation in Liver and Muscle. PPAR
Research 2010. http://dx.doi.org/10.1155/2010/542359
Butenhoff. TL: Biork. TA: Chang. SC: Ehresman. DT: Parker. GA: Das. K: Lau. C: Lieder. PH: van
Otterdiik. FM: Wallace. KB. (2012a], Toxicological evaluation of ammonium
perfluorobutyrate in rats: twenty-eight-day and ninety-day oral gavage studies. Reprod
Toxicol 33: 513-530. http://dx.doi.Org/10.1016/i.reprotox.2011.08.004
Butenhoff. TL: Chang. SC: Ehresman. DT: York. RG. (2009], Evaluation of potential reproductive and
developmental toxicity of potassium perfluorohexanesulfonate in Sprague Dawley rats.
Reprod Toxicol 27: 331-341. http://dx.doi.Org/10.1016/i.reprotox.2009.01.004
Butenhoff. TL: Chang. SC: Olsen. GW: Thomford. PI. (2012b], Chronic dietary toxicity and
carcinogenicity study with potassium perfluorooctanesulfonate in Sprague Dawley rats.
Toxicology 293: 1-15. http://dx.doi.Org/10.1016/i.tox.2012.01.003
Bvrne. S: Seguinot-Medina. S: Miller. P: Waghivi. V: von Hippel. FA: Buck. CL: Carpenter. DO. (2017],
Exposure to polybrominated diphenyl ethers and perfluoroalkyl substances in a remote
population of Alaska Natives. Environ Pollut 231: 387-395.
http://dx.doi.0rg/lO.lOl6/i.envpol.2Oi7.O8.O2O
CDC (Centers for Disease Control and Prevention], (2009], Fourth national report on human
exposure to environmental chemicals. Atlanta, GA: U.S. Department of Health and Human
Services, Centers for Disease Control and Prevention, http: //www.cdc.gov/exposurereport/
CDC (Centers for Disease Control and Prevention]. (2015], Fourth national report on human
exposure to environmental chemicals, updated tables, February 2015. Atlanta, GA.
https://www.cdc.gov/biomonitoring/pdf/FourthReport UpdatedTables Feb2015.pdf
This document is a draft for review purposes only and does not constitute Agency policy.
R-l DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
CDC (Centers for Disease Control and Prevention]. (2018], Fourth national report on human
exposure to environmental chemicals, updated tables, March 2018, volume one. Atlanta, GA:
U.S. Department of Health and Human Services.
https://www.cdc.gov/exposurereport/pdf/FourthReport UpdatedTables Volumel Mar20
18.pdf
CDC (Centers for Disease Control and Prevention], (2019a], Fourth National Report on Human
Exposure to Environmental Chemicals, Updated Tables, January 2019, Volume 1.
https://www.cdc.gov/exposurereport/pdf/FourthReport UpdatedTables Volumel Tan201
9-508.pdf
CDC (Centers for Disease Control and Prevention], (2019b], Fourth National Report on Human
Exposure to Environmental Chemicals, Updated Tables, January 2019, Volume 2. Atlanta,
GA: U.S. Department of Health and Human Services.
https://www.cdc.gov/exposurereport/pdf/FourthReport UpdatedTables Volume2 Tan201
9-508.pdf
Chang. S: Butenhoff. IL: Parker. GA: Coder. PS: Zitzow. ID: Krisko. RM: Biork. I A: Wallace. KB: Seed.
JG. (2018], Reproductive and developmental toxicity of potassium
perfluorohexanesulfonate in CD-I mice. Reprod Toxicol 78: 150-168.
http://dx.doi.Org/10.1016/i.reprotox.2018.04.007
Chen. WL: Bai. FY: Chang. YC: Chen. PC: Chen. CY. (2018], Concentrations of perfluoroalkyl
substances in foods and the dietary exposure among Taiwan general population and
pregnant women. J Food Drug Anal 26: 994-1004.
http://dx.doi.Org/10.1016/i.ifda.2017.12.011
Chiu. WA: Guvton. KZ: Martin. MT: Reif. DM: Rusvn. I. (2018], Use of high-throughput in vitro
toxicity screening data in cancer hazard evaluations by IARC Monograph Working Groups.
ALTEX 35: 51-64. http://dx.doi.org/10.14573/altex.1703231
Chiu. WA: Slob. W. (2015], A unified probabilistic framework for dose-response assessment of
human health effects [Review], Environ Health Perspect 123: 1241-1254.
http://dx.doi.org/10.1289/ehp.1409385
Christensen. KY: Raymond. M: Blackowicz. M: Liu. Y: Thompson. BA: Anderson. HA: Turvk. M.
(2017], Perfluoroalkyl substances and fish consumption. Environ Res 154: 145-151.
http://dx.doi.Org/10.1016/i.envres.2016.12.032
Connecticut DPH (Connecticut Department of Public Health], (2016], Drinking water action level for
perfluorinated alkyl substances (PFAS] (pp. 1-9], https://portal.ctgov/-
/media /D epartments-and-
Agencies/DPH/dph/environmental health/eoha/Toxicology Risk Assessment/DrinkingWa
terActionLevelPerfluorinatedAlkvlSubstances-PFAS.pdf
Corton. TC: Anderson. SP: Stauber. A. (2000], Central role of peroxisome proliferator-activated
receptors in the actions of peroxisome proliferators [Review], Annu Rev Pharmacol Toxicol
40: 491-518. http://dx.doi.Org/10.1146/annurev.pharmtox.40.l.491
Corton. TC: Cunningham. ML: Hummer. BT: Lau. C: Meek. B: Peters. TM: Popp. TA: Rhomberg. L: Seed.
1: Klaunig. IE. (2014], Mode of action framework analysis for receptor-mediated toxicity:
The peroxisome proliferator-activated receptor alpha (PPARa] as a case study [Review],
Crit Rev Toxicol 44: 1-49. http://dx.doi.org/10.3109/10408444.2Q13.835784
Corton. TC: Peters. TM: Klaunig. IE. (2018], The PPARa-dependent rodent liver tumor response is not
relevant to humans: addressing misconceptions [Review], Arch Toxicol 92: 83-119.
http://dx.doi.org/10.1007/s00204-017-2Q94-7
Danish EPA (Danish Environmental Protection Agency], (2015], Perfluoroalkylated substances:
PF0A, PF0S and PF0SA. Evaluation of health hazards and proposal of a health based quality
criterion for drinking water, soil and ground water. (Environmental project No. 1665],
https://www2.mst.dk/Udgiv/publications/2015/04/978-87-93283-01-5.pdf
This document is a draft for review purposes only and does not constitute Agency policy.
R-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Das. KP: Grey. BE: Rosen. MB: Wood. CR: Tatum-Gibbs. KR: Zehr. RD: Strvnar. MT: Lindstrom. AB:
Lau. C. (2015], Developmental toxicity of perfluorononanoic acid in mice. Reprod Toxicol
51: 133-144. http://dx.doi.Org/10.1016/i.reprotox.2014.12.012
Dickersin. K. (1990], The existence of publication bias and risk factors for its occurrence. JAMA 263:
1385-1389.
DOE (U.S. Department of Energy], (2016], Temporary emergency exposure limits for chemicals:
methods and practice. (DOE-HDBK-1046-2016],
https://sp.eota.energy.gOv/EM/Shared%20Documents/D0E-HDBK-1046-2016.pdf
DOE (U.S. Department of Energy], (2018], Table 2: Protective Action Criteria (PAC] Rev. 29a based
on applicable 60-minute AEGLs, ERPGs, or TEELs.
https://sp.eota.energv.gov/pac/docs/Revision 29A Table2.pdf
Edwards. SW: Tan. YM: Villeneuve. PL: Meek. ME: Mcqueen. CA. (2016], Adverse outcome
pathways-organizing toxicological information to improve decision making [Review], J
Pharmacol Exp Ther 356: 170-181. http://dx.doi.org/10.1124/ipet.115.228239
EFSA (European Food Safety Authority], (2008], Perfluorooctane sulfonate (PFOS],
perfluorooctanoic acid (PFOA] and their salts. Scientific opinion of the panel on
contaminants in the food chain. EFSAJ653: 1-131.
http://dx.doi.Org/10.2903/i.efsa.2008.653
EFSA (European Food Safety Authority], (2017], Guidance on the use of the weight of evidence
approach in scientific assessments. EFSA J 15: 1-69.
http://dx.doi.Org/10.2903/i.efsa.2017.4971
EIC (ENVIRON International Corporation], (2014], Assessment of POP criteria for specific short-
chain perfluorinated alkyl substances. (Project Number: 0134304A],
https: / /fluorocouncil.com /wp-content/uploads /2 017 /0 3 /ENVIRON-Assessment-of-P OP -
Criteria-Resources-l.pdf
Fabrega. F: Kumar. V: Benfenati. E: Schuhmacher. M: Domingo. TL: Nadal. M. (2015], Physiologically
based pharmacokinetic modeling of perfluoroalkyl substances in the human body. Toxicol
Environ Chem 97: 814-827. http://dx.doi.org/10.1080/02772248.2015.1060976
Far land. WH. (2005], [Memo to Science Policy council regarding implementation of the cancer
guidelines and accompanying supplemental guidance - Science Policy Council Cancer
Guidelines. Implementation Workgroup communication I: Application of the mode of action
framework in mutagenicity determinations for carcinogenicity]. Available online at
https://www.epa.gov/sites/production/files/2015-
01 /documents/cgiwgcommuniation i.pdf
Filgo. AT: Ouist. EM: Hoenerhoff. MT: Brix. AE: Kissling. GE: Fenton. SE. (2015], Perfluorooctanoic
Acid (PFOA]-induced Liver Lesions in Two Strains of Mice Following Developmental
Exposures: PPARa Is Not Required. Toxicol Pathol 43: 558-568.
http://dx.doi.org/10.1177/0192623314558463
Foreman. IE: Chang. SC: Ehresman. DT: Butenhoff. TL: Anderson. CR: Palkar. PS: Kang. BH: Gonzalez.
FT: Peters. TM. (2009], Differential Hepatic Effects of Perfluorobutyrate Mediated by Mouse
and Human PPAR-alpha. Toxicol Sci 110: 204-211.
http://dx.doi.org/10.1093/toxsci/kfpQ77
FSANZ (Food Standards Australia New Zealand], (2016], Hazard assessment report:
Perfluorooctane sulfonate (PFOS], perfluorooctanoic acid (PFOA], perfluorohexane
sulfonate (PFHxS],
http://www.health.gov.au/internet/main/publishing.nsf/Content/2200FE086D480353CA
2580C900817CDC/$File/6.sdl-Hazard-assessment-report.pdf
Goeden. HM: Greene. CW: Tacobus. TA. (2019], A transgenerational toxicokinetic model and its use in
derivation of Minnesota PFOA water guidance. J Expo Sci Environ Epidemiol 29: 183-195.
http://dx.doi.org/10.1038 /s41370-018-0110-5
This document is a draft for review purposes only and does not constitute Agency policy.
R-3 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Guelfo. TL: Marlow. T: Klein. DM: Savitz. DA: Frickel. S: Crimi. M: Suuberg. EM. (2018], Evaluation
and Management Strategies for Per- and Polyfluoroalkyl Substances (PFASs] in Drinking
Water Aquifers: Perspectives from Impacted U.S. Northeast Communities. Environ Health
Perspect 126: 065001. http://dx.doi.org/10.1289/EHP2727
Guvatt. G: Oxman. AD: Akl. EA: Kunz. R: Vist. G: Brozek. 1: Norris. S: Falck-Ytter. Y: Glasziou. P:
DeBeer. H: laeschke. R: Rind. D: Meerpohl. 1: Dahm. P: Schunemann. HI. (2011], GRADE
guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin
Epidemiol 64: 383-394. http://dx.doi.Org/10.1016/i.iclinepi.2010.04.026
Guvton. KZ: Chiu. WA: Bateson. TF: Tinot. 1: Scott. CS: Brown. RC: Caldwell. TC. (2009], A
reexamination of the PPAR-alpha activation mode of action as a basis for assessing human
cancer risks of environmental contaminants [Review], Environ Health Perspect 117: 1664-
1672. http://dx.doi.org/10.1289/ehp.0900758
Hall. AP: Elcombe. CR: Foster. TR: Harada. T: Kaufmann. W: Knippel. A: Kuttler. K: Malarkev. DE:
Maronpot. RR: Nishikawa. A: Nolte. T: Schulte. A: Strauss. V: York. Ml. (2012], Liver
hypertrophy: a review of adaptive (adverse and non-adverse] changes-conclusions from
the 3rd International ESTP Expert Workshop [Review], Toxicol Pathol 40: 971-994.
http://dx.doi.org/10.n77/0192623312448935
Harbison. RD: Bourgeois. MM: Tohnson. GT. (2015], Hamilton & Hardys industrial toxicology (6th
ed.]. Hoboken, NJ: John Wiley & Sons, Inc.
http://www.wilev.eom/WilevCDA/WilevTitle/productCd-0470929731.subiectCd-
CH50.html
Haug. LS: Thomsen. C: Brantsaeter. AL: Kvalem. HE: Haugen. M: Becher. G: Alexander. I: Meltzer.
HM: Knutsen. HK. (2010], Diet and particularly seafood are major sources of perfluorinated
compounds in humans. Environ Int 36: 772-778.
http://dx.doi.Org/10.1016/i.envint.2010.05.016
Health Canada. (2018a], Drinkingwater screening values: Perfluoroalkylated substances. In Water
Talk (Cat. No.: H144-47/2017E-PDF ed.].
https://www.canada.ca/en/services/health/publications/healthy-living/water-talk-
drinking-water-screening-values-perfluoroalkvlated-substances.html
Health Canada. (2018b], Guidelines for Canadian drinkingwater quality: Guideline technical
document - Perfluorooctane sulfonate (PFOS], In Guidelines for Canadian Drinking Water
Quality. Ottawa, Ontario: Minister of Health, https: //www.canada.ca/en/health-
canada/services/publications/healthy-living/guidelines-canadian-drinking-water-qualitv-
guideline-technical-document-perfluorooctane-sulfonate/documenthtml
Health Canada. (2018c], Guidelines for Canadian drinkingwater quality: Guideline technical
document - Perfluorooctanoic acid (PFOA], In Guidelines for Canadian Drinking Water
Quality. Ottawa, Ontario: Minister of Health, https: //www.canada.ca/en/health-
canada/services/publications/healthy-living/guidelines-canadian-drinking-water-qualitv-
technical-document-perfluorooctanoic-acid/document.html
Heo. II: Lee. TW: Kim. SK: Oh. IE. (2014], Foodstuff analyses show that seafood and water are major
perfluoroalkyl acids (PFAAs] sources to humans in Korea. J Hazard Mater 279: 402-409.
http://dx.doi.Org/10.1016/i.ihazmat.2014.07.004
Higgins. TPT: Green. S. (2011], Cochrane handbook for systematic reviews of interventions. Version
5.1.0 (Updated March 2011]: The Cochrane Collaboration. Retrieved from
http://handbook.cochrane.org/
Hill. AB. (1965], The environment and disease: Association or causation? Proc R Soc Med 58: 295-
300.
Hirst. TA: Howick. I: Aronson. IK: Roberts. N: Perera. R: Koshiaris. C: Heneghan. C. (2014], The need
for randomization in animal trials: an overview of systematic reviews [Review], PLoS ONE
9: e98856. http://dx.doi.org/10.1371/iournal.pone.0098856
This document is a draft for review purposes only and does not constitute Agency policy.
R-4 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Hoenig. TM: Heisev. DM. (2001], The abuse of power: The pervasive fallacy of power calculations for
data analysis. Am Stat 55: 19-24.
Howard. BE: Phillips. 1: Miller. K: Tandon. A: Mav. D: Shah. MR: Holmgren. S: Pelch. KE: Walker. V:
Roonev. AA: Macleod. M: Shah. RR: Thayer. K. (2016], SWIFT-Review: a text-mining
workbench for systematic review. Syst Rev 5: 87. http://dx.doi.org/10.1186/sl3643-016-
0263-z
Kato. K: Calafat. AM: Needham. LL. (2009], Polyfluoroalkyl chemicals in house dust. Environ Res
109: 518-523. http://dx.doi.Org/10.1016/i.envres.2009.01.005
KemI (Swedish Chemicals Agency], (2015], International chemical identification:
Nonadecaflourodecanoic acid (PFDA] and its ammonium and sodium salts.
https://echa.europa.eu/documents/10162/0ddl63d4-56e5-3f3b-8a91-745ec9af2a9f
Kim. ST: Choi. El: Choi. GW: Lee. YB: Cho. HY. (2019], Exploring sex differences in human health risk
assessment for PFNA and PFDA using a PBPK model. Arch Toxicol 93: 311-330.
http://dx.doi.org/10.1007/sQ0204-018-2365-v
Kim. ST: Shin. H: Lee. YB: Cho. HY. (2018], Sex-specific risk assessment of PFHxS using a
physiologically based pharmacokinetic model. Arch Toxicol 92: 1113-1131.
http://dx.doi.org/10.1007/s00204-017-2116-5
Kim. SK: Kannan. K. (2007], Perfluorinated acids in air, rain, snow, surface runoff, and lakes:
relative importance of pathways to contamination of urban lakes. Environ Sci Technol 41:
8328-8334. http://dx.doi.org/10.1021/es072107t
Klaunig. IE: Shinohara. M: Iwai. H: Chengelis. CP: Kirkpatrick. IB: Wang. Z: Bruner. RH. (2015],
Evaluation of the chronic toxicity and carcinogenicity of perfluorohexanoic acid (PFHxA] in
Sprague-Dawley rats. Toxicol Pathol 43: 209-220.
http://dx.doi.org/10.1177/0192623314530532
Krauth. D: Woodruff. TT: Bero. L. (2013], Instruments for assessing risk of bias and other
methodological criteria of published animal studies: a systematic review [Review], Environ
Health Perspect 121: 985-992. http://dx.doi.org/10.1289/ehp.1206389
Laitinen. TA: Koponen. I: Koikkalainen. I: Kiviranta. H. (2014], Firefighters' exposure to
perfluoroalkyl acids and 2-butoxyethanol present in firefighting foams. Toxicol Lett 231:
227-232. http: / /dx.doi.org/10.1016/i.toxlet.2014.09.007
Lau. C. (2015], Perfluorinated compounds: An overview. InJC DeWitt (Ed.], Toxicological effects of
perfluoroalkyl and polyfluoroalkyl substances (pp. 1-21]. New York: Springer.
http://dx.doi.org/10.1007/978-3-319-15518-0 1
Lau. C: Anitole. K: Hodes. C: Lai. D: Pfahles-Hutchens. A: Seed. 1. (2007], Perfluoroalkyl acids: a
review of monitoring and toxicological findings [Review], Toxicol Sci 99: 366-394.
http: / /dx.doi.or g/10.109 3 /toxsci /kfm 128
Lau. C: Thibodeaux. TR: Hanson. RG: Narotskv. MG: Rogers. TM: Lindstrom. AB: Strvnar. Ml. (2006],
Effects of perfluorooctanoic acid exposure during pregnancy in the mouse. Toxicol Sci 90:
510-518. http://dx.doi.org/10.1093/toxsci/kfil05
Li. K: Gao. P: Xiang. P: Zhang. X: Cui. X: Ma. LP. (2017], Molecular mechanisms of PFOA-induced
toxicity in animals and humans: Implications for health risks [Review], Environ Int 99: 43-
54. http: //dx.doi.Org/10.1016/i.envint.2016.ll.014
Luebker. DT: Case. MT: York. RG: Moore. TA: Hansen. KT: Butenhoff. TL. (2005], Two-generation
reproduction and cross-foster studies of perfluorooctanesulfonate (PFOS] in rats.
Toxicology 215: 126-148. http://dx.doi.Org/10.1016/i.tox.2005.07.018
Mackowiak. B: Hodge. 1: Stern. S: Wang. H. (2018], The roles ofxenobiotic receptors: Beyond
chemical disposition [Review], Drug Metab Dispos 46: 1361-1371.
http://dx.doi.org/10.1124/dmd.118.081Q42
This document is a draft for review purposes only and does not constitute Agency policy.
R-5 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Macleod. MR. (2013], Systematic reviews of experimental animal studies. Presentation presented at
Workshop on weight of evidence; US National Research Council Committee to review the
Integrated Risk Information System (IRIS] process, March 27-28, 2013, Washington, DC.
Malonev. EK: Waxman. DT. (1999], trans-Activation of PPARalpha and PPARgamma by structurally
diverse environmental chemicals. Toxicol Appl Pharmacol 161: 209-218.
http://dx.doi.org/10.1006/taap.1999.8809
MassDEP (Massachusetts Department of Environmental Protection], (2018], MassDEP ORS
Guideline for PFAS: current regulatory limit: per- and polyfluoroalkyl substances (PFAS],
including the US EPAUCMR3 analytes.
https://www.mass.gOv/files/documents/2018/06/ll/orsg-pfas-20180608.pdf
MDH (Minnesota Department of Health], (2009], East Metro Perfluorochemical Biomonitoring Pilot
Project. (8EHQ-80-373], St. Paul, MN: 3M Company.
MDH (Minnesota Department of Health], (2018], Toxicological summary for: Perfluorobutanoate.
Health Based Guidance for Water Health Risk Assessment Unit, Environmental Health
Division.
https://www.health.state.mn.us/communities/environment/risk/docs/guidance/gw/pfba
2summ.pdf
MDH (Minnesota Department of Health], (2019], Health Based Guidance for Water: Toxicological
Summary for: Perfluorohexane sulfonate (PFHxS],
https://www.health.state.mn.us/communities/environment/risk/docs/guidance/gw/pfhx
s.pdf
Meek. ME: Boobis. A: Cote. I: Dellarco. V: Fotakis. G: Munn. S: Seed. I: Vickers. C. (2014], New
developments in the evolution and application of the WH0/IPCS framework on mode of
action/species concordance analysis [Review], J Appl Toxicol 34: 1-18.
http://dx.doi.org/10.1002/iat.2949
Mellor. CL: Steinmetz. FP: Cronin. MT. (2016], The identification of nuclear receptors associated
with hepatic steatosis to develop and extend adverse outcome pathways [Review]. Crit Rev
Toxicol 46: 138-152. http: //dx.doi.org/10.3109/10408444.2015.1089471
Moreta. C: Tena. MT. (2014], Determination of perfluorinated alkyl acids in corn, popcorn and
popcorn bags before and after cooking by focused ultrasound solid-liquid extraction, liquid
chromatography and quadrupole-time of flight mass spectrometry. J Chromatogr A 1355:
211-218. http://dx.doi.Org/10.1016/i.chroma.2014.06.018
Morgan. RL: Thaver. KA: Bero. L: Bruce. N: Falck-Ytter. Y: Ghersi. D: Guvatt. G: Hooiimans. C:
Langendam. M: Mandrioli. D: Mustafa. RA: Rehfuess. EA: Roonev. AA: Shea. B: Silbergeld. EK:
Sutton. P: Wolfe. MS: Woodruff. TI: Verbeek. TH: Hollowav. AC: Santesso. N: Schunemann. HI.
(2016], GRADE: Assessing the quality of evidence in environmental and occupational health.
Environ Int 92-93: 611-616. http://dx.doi.Org/10.1016/i.envint.2016.01.004
Nabb. PL: Szostek. B: Himmelstein. MW: Mawn. MP: Gargas. ML: Sweeney. LM: Stadler. TC: Buck. RC:
Fasano. WT. (2007], In vitro metabolism of 8-2 fluorotelomer alcohol: interspecies
comparisons and metabolic pathway refinement. Toxicol Sci 100: 333-344.
http://dx.doi.org/10.1093/toxsci/kfm230
NAS (National Academy of Sciences], (2014], Review of EPA's Integrated Risk Information System
(IRIS] process. Washington, DC: National Academies Press (US],
http://dx.doi.org/10.17226/18764
New Hampshire PES (New Hampshire Department of Environmental Services], (2019], Summary
report on the New Hampshire Pepartment of Environmental Services development of the
Maximum Contaminant Levels and Ambient Groundwater Quality Standards for
perfluorooctanesulfonic acid (PF0S], perfluorooctanoic acid (PF0A], perfluorononoic acid
(PFNA] and perfluorohexanesulfonic acid (PFHxS], (R-WP-19-01],
https://www4.des.state.nh.us/nh-pfas-investigation/?p=923
This document is a draft for review purposes only and does not constitute Agency policy.
R-6 PRAFT-P0 NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
NHMRC (Australian Government National Health and Medical Research Council], (2019], Physical
and chemical characteristics: Fact sheets: Per-fluoroalkyl and poly-fluoroalkyl substances
(PFAS], In Australian Drinking Water Guidelines.
NICNAS (National Industrial Chemicals Notification and Assessment Scheme], (2014], Public
Report: Polyfluorinated polymer in Capstone FS-81 and Capstone TR. (File No. LTD/1406],
NICNAS (National Industrial Chemicals Notification and Assessment Scheme], (2015], Public
Report: Efka SL 3239. (File No. LTD/1572],
NIEHS (National Institute of Environmental Health Sciences], (2015], Handbook for conducting a
literature-based health assessment using OHAT approach for systematic review and
evidence integration. Office of Health Assessment and Translation (OHAT], National
Toxicology Program, U.S Department of Health and Human Services.
https://ntp.niehs.nih.gov/ntp/ohat/pubs/handbookian2015 508.pdf
NTDWOI (New Jersey Drinking Water Quality Institute], (2015], Health-based maximum
contaminant level support document: perfluorononanoic acid (PFNA], NJDWQI Health
Effects Subcommittee, https://www.state.ni.us/dep/watersupplv/pdf/pfha-health-
effects.pdf
NTDWOI (New Jersey Drinking Water Quality Institute], (2017], Public review draft: Health-based
maximum contaminant level support document: Perfluorooctane sulfonate (PF0S] (CAS #:
1763-23-1; Chemical Formula: C8HF1703S], NJDWQI Health Effects Subcommittee.
https://cswab.org/wp-content/uploads/2018/09/New-Tersev-Scientific-Documentation-
Supporting-PF0S-13-ppt.pdf
NLM (National Institutes of Health, National Library of Medicine], (2013], HSDB: Perfluoro-n-
nonanoic acid. Available online at https: //toxnet.nlm.nih.gov/cgi-
bin/sis/search/a?dbs+hsdb:(5)term+(5)DOCNO+8040
NLM (National Institutes of Health, National Library of Medicine], (2016], HSDB: Perfluorohexanoic
acid. Available online at https://toxnet.nlm.nih.gov/cgi-
bin/sis/search/a?dbs+hsdb:(5)term+@DOCNO+8299
NLM (National Institutes of Health, National Library of Medicine], (2017], HSDB:
Perfluorohexanesulfonic acid. Available online at https://toxnet.nlm.nih.gov/cgi-
bin/sis/search/a?dbs+hsdb:(5)term+@DOCNO+8274
Norwegian Environment Agency. (2018], Investigation of sources to PFHxS in the environment.
Munich, Germany: BiPRO GmbH.
https://www.miliodirektoratet.no/globalassets/publikasioner/M961/M961.pdf
NRC (National Research Council], (2014], Review of EPA's Integrated Risk Information System
(IRIS] process. Washington, DC: The National Academies Press.
http://www.nap.edu/catalog.php7record id=18764
NTP (National Toxicology Program], (2011], NTP Protocol Outline of the Class Study of
Perfluorinated Chemicals [perfluorobutane sulfonate (PFBS], perfluorohexane sulfonate
potassium salt (PFHSKslt], perfluorooctane sulfonate (PFOS], perfluorohexanoic acid
(PFHxA], perfluorooctanoic acid (PFOA], perfluorononanoic acid (PFNA], perfluorodecanoic
acid (PFDA] andWY 14643 (WY]J in Harlan SpragueDawley rats administered daily by
gavage for 28-Days [NTP],
NTP (National Toxicology Program], (2015], Handbook for conducting a literature-based health
assessment using OHAT approach for systematic review and evidence integration. U.S. Dept.
of Health and Human Services, National Toxicology Program.
https://ntp.niehs.nih.gov/ntp/ohat/pubs/handbookian2015 508.pdf
NTP (National Toxicology Program], (2018a], TOX-96: 1-Perfluorobutanesulfonic acid (375-73-5],
potassium perfluorohexanesulfonate (3871-99-6], perfluorooctane sulfonate (1763-23-1],
WY-14643 (50892-23-4], Chemical Effects in Biological Systems (CEBS], Available online at
This document is a draft for review purposes only and does not constitute Agency policy.
R-7 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
https://manticore.niehs.nih.gov/cebssearch/publication/TOX-96 (accessed August 6,
2018],
NTP (National Toxicology Program], (2018b], TOX-97: Perfluorohexanoic acid (307-24-4],
perfluorooctanoic acid (335-67-1], perfluorononanoic acid (375-95-1], perfluorodecanoic
acid (335-76-2], WY-14643 (50892-23-4], Chemical Effects in Biological Systems (CEBS],
Available online at https: //manticore.niehs.nih.gov/cebssearch/publication/TOX-97
(accessed April 24, 2018],
NTP (National Toxicology Program], (2019], TOX-96: Toxicity Report Tables and Curves for Short-
term Studies: Perfluorinated Compounds: Sulfonates. Available online at
https://tools.niehs.nih.gov/cebs3/views/?action=main.dataReview&bin id=3874
OECD (Organisation for Economic Co-operation and Development], (2015], Working towards a
global emission inventory of PFASS: focus on PFCAS - status quo and the way forward. Paris,
France.
http://www.oecd.org/chemicalsafetv/Working%20Towards%20a%20Global%20Emission
% 2 01nventorv% 2 0of%2 OPFASS.pdf
Papadopoulou. E: Sabaredzovic. A: Namork. E: Nvgaard. UC: Granum. B: Haug. LS. (2016], Exposure
of Norwegian toddlers to perfluoroalkyl substances (PFAS]: The association with
breastfeeding and maternal PFAS concentrations. Environ Int 94: 687-694.
http://dx.doi.Org/10.1016/i.envint.2016.07.006
Peraza. MA: Burdick. AD: Marin. HE: Gonzalez. FT: Peters. TM. (2006], The toxicology of ligands for
peroxisome proliferator-activated receptors (PPAR] [Review], Toxicol Sci 90: 269-295.
http://dx.doi.org/10.1093/toxsci/kfi062
Perez. F: Llorca. M: Kock-Schulmever. M: Skrbic. B: Oliveira. LS: da Boit Martinello. K: Al-Dhabi. NA:
Antic. I: Farre. M: Barcelo. D. (2014], Assessment of perfluoroalkyl substances in food items
at global scale. Environ Res 135: 181-189. http: //dx.doi.Org/10.1016/i.envres.2014.08.004
Perkins. RG: Butenhoff. TL: Kennedy. GL: Palazzolo. Ml. (2004], 13-week dietary toxicity study of
ammonium perfluorooctanoate (APF0] in male rats. Drug Chem Toxicol 27: 361-378.
http://dx.doi.org/10.1081 /DCT-200039773
Pinnev. SM: Biro. FM: Windham. GC: Herrick. RL: Yaghivan. L: Calafat. AM: Succop. P: Sucharew. H:
Ball. KM: Kato. K: Kushi. LH: Bornschein. R. (2014], Serum biomarkers of polyfluoroalkyl
compound exposure in young girls in Greater Cincinnati and the San Francisco Bay Area,
USA. Environ Pollut 184: 327-334. http://dx.doi.Org/10.1016/i.envpol.2013.09.008
Post. GB: Cohn. PD: Cooper. KR. (2012], Perfluorooctanoic acid (PF0A], an emerging drinking water
contaminant: a critical review of recent literature [Review], Environ Res 116: 93-117.
http://dx.doi.Org/10.1016/i.envres.2012.03.007
Prevedouros. K: Cousins. IT: Buck. RC: Korzeniowski. SH. (2006], Sources, fate and transport of
perfluorocarboxylates [Review], Environ Sci Technol 40: 32-44.
http://dx.doi.org/10.1021/es0512475
Rosen. MB: Das. KP: Roonev. 1: Abbott. B: Lau. C: Corton. TC. (2017], PPARa-independent
transcriptional targets of perfluoroalkyl acids revealed by transcript profiling. Toxicology
387: 95-107. http://dx.doi.org/10.1016/i.tox.2017.05.013
Rotander. A: Karrman. A: Toms. LM: Kay. M: Mueller. IF: Gomez Ramos. Ml. (2015], Novel
fluorinated surfactants tentatively identified in firefighters using liquid chromatography
quadrupole time-of-flight tandem mass spectrometry and a case-control approach. Environ
Sci Technol 49: 2434-2442. http://dx.doi.org/10.1021/es503653n
Schaider. LA: Balan. SA: Blum. A: Andrews. DO: Strvnar. Ml: Dickinson. ME: Lunderberg. DM: Lang.
TR: Peaslee. GF. (2017], Fluorinated compounds in US fast food packaging. Environ Sci
Technol Lett 4: 105-111. http: //dx.doi.org/10.1021/acs.estlett.6b00435
Schecter. A: Malik-Bass. N: Calafat. AM: Kato. K: Colacino. TA: Gent. TL: Hvnan. LS: Harris. TR: Malla.
S: Birnbaum. L. (2012], Polyfluoroalkyl compounds in Texas children from birth through 12
This document is a draft for review purposes only and does not constitute Agency policy.
R-8 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
years of age. Environ Health Perspect 120: 590-594.
http://dx.doi.org/10.1289/ehp.1104325
Schunemann. H: Hill. S: Guvatt. G: Akl. EA: Ahmed. F. (2011], The GRADE approach and Bradford
Hill's criteria for causation. J Epidemiol Community Health 65: 392-395.
http://dx.doi.org/10.1136/iech.2010.119933
Seacat. AM: Thomford. PI: Hansen. KT: Olsen. GW: Case. MT: Butenhoff. TL. (2002], Subchronic
toxicity studies on perfluorooctanesulfonate potassium salt in cynomolgus monkeys.
Toxicol Sci 68: 249-264. http://dx.doi.Org/10.1093/toxsci/68.l.249
Shipley. TM: Hurst. CH: Tanaka. SS: DeRoos. FL: Butenhoff. TL: Seacat. AM: Waxman. DT. (2004],
Trans-activation of PPARalpha and induction of PPARalpha target genes by
perfluorooctane-based chemicals. Toxicol Sci 80: 151-160.
http://dx.doi.org/10.1093/toxsci/kfhl30
Smith. MT: Guvton. KZ: Gibbons. CF: Fritz. TM: Portier. CI: Rusvn. I: DeMarini. DM: Caldwell. TC:
Kavlock. RT: Lambert. PF: Hecht. SS: Bucher. TR: Stewart. BW: Baan. RA: Cogliano. VI: Straif.
K. (2016], Key characteristics of carcinogens as a basis for organizing data on mechanisms
of carcinogenesis [Review], Environ Health Perspect 124: 713-721.
http://dx.doi.org/10.1289/ehp.1509912
SNFA (Swedish National Food Agency], (2018], [Risk management - PFAS in drinking water and
fish]. Available online at https: //www.livsmedelsverket.se/livsmedel-och-
innehall/oonskade-amnen/miliogifter/pfas-polv-och-perfluorerade-
alkvlsubstanser/riskhantering-pfaa-i-dricksvatten
Soldatow. VY: Lecluvse. EL: Griffith. LG: Rusvn. I. (2013], In vitro models for liver toxicity testing.
Toxicology Research 2: 23-39. http://dx.doi.org/10.1039/c2tx20051a
Stahl. LL: Snvder. BP: Olsen. AR: Kincaid. TM: Wathen. IB: McCartv. HB. (2014], Perfluorinated
compounds in fish from U.S. urban rivers and the Great Lakes. Sci Total Environ 499: 185-
195. http://dx.doi.Org/10.1016/i.scitotenv.2014.07.126
Sterne. TAC: Hernan. MA: Reeves. BC: Savovic. 1: Berkman. ND: Viswanathan. M: Henry. D: Altman.
DG: Ansari. MT: Boutron. I: Carpenter. TR: Chan. AW: Churchill. R: Peeks. 11: Hrobiartsson. A:
Kirkham. 1: Tuni. P: Loke. YK: Pigott. TP: Ramsay. CR: Regidor. D: Rothstein. HR: Sandhu. L:
Santaguida. PL: Schunemann. HT: Shea. B: Shrier. I: Tugwell. P: Turner. L: Valentine. TC:
Waddington. H: Waters. E: Wells. GA: Whiting. PF: Higgins. TPT. (2016], R0BINS-I: A tool for
assessing risk of bias in non-randomised studies of interventions. Br Med J 355: i4919.
Strvnar. MT: Lindstrom. AB. (2008], Perfluorinated compounds in house dust from Ohio and North
Carolina, USA. Environ Sci Technol 42: 3751-3756. http://dx.doi.org/10.1021/es7032058
Stump. DG: Holson. IF: Murphy. SR: Farr. CH: Schmit. B: Shinohara. M. (2008], An oral two-
generation reproductive toxicity study of S-111-S-WB in rats. Reprod Toxicol 25: 7-20.
http://dx.doi.Org/10.1016/i.reprotox.2007.10.002
Surma. M: Piskula. M: Wiczkowski. W: Zielinski. H. (2017], The perfluoroalkyl carboxylic acids
(PFCAs] and perfluoroalkane sulfonates (PFSAs] contamination level in spices. European
Food Research and Technology 243: 297-307. http://dx.doi.org/10.1007/s00217-016-
2744-7
Swenberg. TA: Lehman-McKeeman. LP. (1999], Alpha 2-urinary globulin-associated nephropathy as
a mechanism of renal tubule cell carcinogenesis in male rats. In CC Capen; E Dybing; JM
Rice; JD Wilbourn (Eds.], Species Differences in Thyroid, Kidney, and Urinary Bladder
Carcinogenesis (pp. 95-118], Lyon, France: International Agency for Research on Cancer.
http://apps.who.int/bookorders/anglais/detartl.isp?sesslan=l&codlan=l&codcol=73&cod
cch=147
Takacs. ML: Abbott. BP. (2007], Activation of mouse and human peroxisome proliferator-activated
receptors (alpha, beta/delta, gamma] by perfluorooctanoic acid and perfluorooctane
sulfonate. Toxicol Sci 95: 108-117. http://dx.doi.org/10.1093/toxsci/kfll35
This document is a draft for review purposes only and does not constitute Agency policy.
R-9 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
TCEQ (Texas Commission on Environmental Quality], (2015], TCEQ guidelines to develop toxicity
factors. (Revised RG-442], Austin, TX. https://www.tceq.texas.gov/assets/public/
comm exec/pubs/rg/rg-442.pdf
TCEQ (Texas Commission on Environmental Quality], (2013], Tiered Development of
Human Health PCLs. https: //www.tceq.texas.gov/assets/public/comm exec/pubs/rg/rg-366-trrp-
22.pdf
TCEQ (Texas Commission on Environmental Quality], (2018], TRRP protective concentration levels:
April 2018 PCL and supporting tables. Retrieved from
https://www.tceq.texas.gov/remediation/trrp/trrppcls.html
Thomford. PI. (2002], 104-week dietary chronic toxicity and carcinogenicity study with
perfluorooctane sulfonic acid potassium salt (PFOS; T-6295] in rats. Study No. 6329-183.
Madison, WI: Convance Laboratories.
U.S. EPA (U.S. Environmental Protection Agency], (1988], Recommendations for and documentation
of biological values for use in risk assessment [EPA Report] (pp. 1-395], (EPA/600/6-
87/008], Cincinnati, OH: U.S. Environmental Protection Agency, Office of Research and
Development, Office of Health and Environmental Assessment.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=34855
U.S. EPA (U.S. Environmental Protection Agency], (1991a], Guidelines for developmental toxicity
risk assessment (pp. 1-71]. (EPA/600/FR-91/001], Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum.
http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=23162
U.S. EPA (U.S. Environmental Protection Agency], (1991b], Guidelines for developmental toxicity
risk assessment. Fed Reg 56: 63798-63826.
U.S. EPA (U.S. Environmental Protection Agency], (1992], Guidelines for Exposure Assessment [EPA
Report], In Risk Assessment Forum. (EPA/600/Z-92/001], Washington, DC: U. S.
Environmental Protection Agency.
https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=15263
U.S. EPA (U.S. Environmental Protection Agency], (1994], Methods for derivation of inhalation
reference concentrations and application of inhalation dosimetry [EPA Report].
(EPA/600/8-90/066F], Research Triangle Park, NC: U.S. Environmental Protection Agency,
Office of Research and Development, Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office.
https://cfpub.epa.gov/ncea/risk/recordisplav.cfm?deid=71993&CFID=51174829&CFTOKE
N=25006317
U.S. EPA (U.S. Environmental Protection Agency], (1996a], Guidelines for reproductive toxicity risk
assessment (pp. 1-143], (EPA/630/R-96/009], Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum.
https://www.epa.gov/sites/production/files/2Q14-
11/documents/guidelines repro toxicitv.pdf
U.S. EPA (U.S. Environmental Protection Agency], (1996b], Guidelines for reproductive toxicity risk
assessment. Fed Reg 61: 56274-56322.
U.S. EPA (U.S. Environmental Protection Agency], (1998], Guidelines for neurotoxicity risk
assessment [EPA Report] (pp. 1-89], (EPA/630/R-95/001F], Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment Forum.
http://www.epa.gov/risk/guidelines-neurotoxicitv-risk-assessment
U.S. EPA (U.S. Environmental Protection Agency], (2002], A review of the reference dose and
reference concentration processes (pp. 1-192], (EPA/630/P-02/002F], Washington, DC:
U.S. Environmental Protection Agency, Risk Assessment Forum.
http://www.epa.gov/osa/review-reference-dose-and-reference-concentration-processes
This document is a draft for review purposes only and does not constitute Agency policy.
R-10 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
U.S. EPA (U.S. Environmental Protection Agency], (2004], Toxicological review of boron and
compounds. In support of summary information on the Integrated Risk Information System
(IRIS] [EPAReport], (EPA/635/04/052]. Washington, DC: U.S. Environmental Protection
Agency, IRIS. http://nepis.epa.gov/exe/ZvPURL.cgi?Dockev=P 1006CK9.txt
U.S. EPA (U.S. Environmental Protection Agency], (2005a], Guidelines for carcinogen risk
assessment [EPA Report] (pp. 1-166], (EPA/630/P-03/001F], Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment Forum.
https://www.epa.gov/sites/production/files/2Q13-
09/documents/cancer guidelines final 3-25-05.pdf
U.S. EPA (U.S. Environmental Protection Agency], (2005b], Supplemental guidance for assessing
susceptibility from early-life exposure to carcinogens [EPA Report], (EPA/630/R-03/003F],
Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
https://www3.epa.gov/airtoxics/childrens supplement final.pdf
U.S. EPA (U.S. Environmental Protection Agency], (2011a], Recommended use of body weight 3/4
as the default method in derivation of the oral reference dose (pp. 1-50],
(EPA/lOO/Rll/OOOl], Washington, DC: U.S. Environmental Protection Agency, Risk
Assessment Forum, Office of the Science Advisor.
https://www.epa.gov/risk/recommended-use-bodv-weight-34-default-method-derivation-
oral-reference-dose
U.S. EPA (U.S. Environmental Protection Agency], (2011b], Toxicological review of
trichloroethylene (CASRN 79-01-6] in support of summary information on the Integrated
Risk Information System (IRIS] [EPA Report], (EPA/635/R-09/011F], Washington, DC.
https://cfpub.epa.gov/ncea/iris/iris documents/documents/toxreviews/0106tr.pdf
U.S. EPA (U.S. Environmental Protection Agency], (2012], Benchmark dose technical guidance.
(EPA/100/R-12/001], Washington, DC: U.S. Environmental Protection Agency, Risk
Assessment Forum, https: //www.epa.gov/risk/benchmark-dose-technical-guidance
U.S. EPA (U.S. Environmental Protection Agency], (2014a], Draft: Health effects document for
perfluorooctane sulfonate (PFOS] [EPAReport], (EPA/822/R-14/002], Washington, DC: U.S.
Environmental Protection Agency, Office of Water, Health and Ecological Criteria Division.
https://peerreview.versar.com/epa/pfoa/pdf/Health-Effects-Document-for-
Perfluorooctane-Sulfonate-fPFOSlpdf
U.S. EPA (U.S. Environmental Protection Agency], (2014b], Draft: Health effects document for
Perfluorooctanoic Acid (PFOA] [EPAReport], (EPA Document Number: 822R14001], Office
of Water, Health and Ecological Criteria Division.
https://peerreview.versar.com/epa/pfoa/pdf/Health-Effects-Document-for-
Perfluorooctanoic-Acid-fPFOAlpdf
U.S. EPA (U.S. Environmental Protection Agency], (2014c], Guidance for applying quantitative data
to develop data-derived extrapolation factors for interspecies and intraspecies
extrapolation [EPAReport], (EPA/100/R-14/002F], Washington, DC: Risk Assessment
Forum, Office of the Science Advisor, https://www.epa.gov/sites/production/files/2015-
01 /documents/ddef-final.pdf
U.S. EPA (U.S. Environmental Protection Agency], (2014d], Health effects document for
perfluorooctane sulfonate (PFOS]: Draft [EPA Report], (EPA/822/R-14/002], Washington,
DC: U.S. Environmental Protection Agency, Office of Water, Health and Ecological Criteria
Division. https://peerreview.versar.com/epa/pfoa/pdf/Health-Effects-Document-for-
Perfluorooctane-Sulfonate-fPFOSlpdf
U.S. EPA (U.S. Environmental Protection Agency], (2016a], Drinking water health advisory for
perfluorooctane sulfonate (PFOS] [EPA Report], (EPA 822-R-16-004], Washington, DC: U.S.
Environmental Protection Agency, Office of Water.
https://www.regulations.gov/document?D=EPA-H0-0W-2 014-0138-0038
This document is a draft for review purposes only and does not constitute Agency policy.
R-ll DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
U.S. EPA (U.S. Environmental Protection Agency], (2016b], Drinking water health advisory for
perfluorooctanoic acid (PFOA], (EPA 822-R-16-005], Washington, DC: U.S. Environmental
Protection Agency, Office of Water. https://www.regulations.gov/document?D=EPA-HO-
QW-2014-0138-0041
U.S. EPA (U.S. Environmental Protection Agency], (2016c], 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/2016-05/documents/pfos hesd final 508.pdf
U.S. EPA (U.S. Environmental Protection Agency], (2016d], 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, Health and Ecological Criteria Division.
https://www.epa.gOv/sites/production/files/2016-05/documents/pfoa hesd final -
plain.pdf
U.S. EPA (U.S. Environmental Protection Agency], (2016e], The Third Unregulated Contaminant
Monitoring Rule. Available online at https: //www.epa.gov/dwucmr/third-unregulated-
contaminant-monitoring-rule
U.S. EPA (U.S. Environmental Protection Agency], (2017a], Guidance to assist interested persons in
developing and submitting draft risk evaluations under the Toxic Substances Control Act.
(EPA/740/R17/001], Washington, DC: U.S Environmental Protection Agency, Office of
Chemical Safety and Pollution Prevention.
https://www.epa.gov/sites/production/files/2017-
06/documents/tsca ra guidance final.pdf
U.S. EPA (U.S. Environmental Protection Agency], (2017b], The Third Unregulated Contaminant
Monitoring Rule (UCMR 3]: Data summary, January 2017. (EPA 815-S-17-001], Washington,
DC: U.S Environmental Protection Agency, Office ofWater.
https://www.epa.gov/sites/production/files/2017-02/documents/ucmr3-data-summarv-
ianuarv-2017.pdf
U.S. EPA (U.S. Environmental Protection Agency], (2018a], Chemistry Dashboard. Washington, DC.
Retrieved from https: //comptox.epa.gov/dashboard
U.S. EPA (U.S. Environmental Protection Agency], (2018b], An umbrella Quality Assurance Project
Plan (QAPP] for PBPK models [EPA Report], (ORD QAPP ID No: B-0030740-QP-1-1],
Research Triangle Park, NC.
U.S. EPA (U.S. Environmental Protection Agency], (2019], ChemView [Database], Retrieved from
https://chemview.epa.gov/chemview
van Otterdiik. FM. (2007], Repeated dose 28-day oral toxicity study with MTDID-8391 by daily
gavage in the rat, followed by a 21-day recovery period. (Study Number 06-226],
Maplewood, MN: 3M.
Verner. MA: Ngueta. G: Tensen. ET: Fromme. H: Voelkel. W: Nvgaard. UC: Granum. B: Longnecker.
MP. (2016], A simple pharmacokinetic model of prenatal and postnataleExposure to
Perfluoroalkyl Substances (PFASs], Environ Sci Technol 50: 978-986.
http://dx.doi.org/10.1021/acs.est.5b04399
Vesterinen. HM: Sena. ES: Egan. KT: Hirst. TC: Churolov. L: Currie. GL: Antonic. A: Howells. DW:
Macleod. MR. (2014], Meta-analysis of data from animal studies: a practical guide. J
Neurosci Methods 221: 92-102. http://dx.doi.Org/10.1016/i.ineumeth.2013.09.010
Viberg. H: Eriksson. P. (2017], Chapter 43 - Perfluorooctane sulfonate and perfluorooctanoic acid.
In RC Gupta (Ed.], Reproductive and Developmental Toxicology (2nd ed., pp. 811-827], San
Diego, CA: Academic Press. http://dx.doi.org/10.1016/B978-0-12-804239-7.00043-3
von der Trenck. KT: Konietzka. R: Biegel-Engler. A: Brodskv. I: Hadicke. A: Ouadflieg. A: Stockerl. R:
Stahl. T. (2018], Significance thresholds for the assessment of contaminated groundwater:
This document is a draft for review purposes only and does not constitute Agency policy.
R-12 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
perfluorinated and polyfluorinated chemicals [Review], Environ Sci Eur 30: 19.
http://dx.doi.org/10.1186/sl2302-018-0142-4
VT DOH [Vermont Department of Health], (2018], [Memorandum to Emily Boedecker from Mark A.
Levine regarding drinking water health advisory for five PFAs (per- and polyfluorinated
alkyl substances]]. Available online at
http://www.healthvermont.gov/sites/default/files/documents/pdf/ENV DW PFAS Health
Advisorv.pdf
Wambaugh. IF: Setzer. RW: Pitruzzello. AM: Liu. 1: Reif. DM: Kleinstreuer. NC: Wang. NC: Sipes. N:
Martin. M: Das. K: Dewitt. TC: Strvnar. M: Tudson. R: Houck. KA: Lau. C. (2013], Dosimetric
anchoring of in vivo and in vitro studies for perfluorooctanoate and
perfluorooctanesulfonate. Toxicol Sci 136: 308-327.
http://dx.doi.org/10.1093/toxsci/kft204
Wang. Y: Wang. L: Li. 1: Liang. Y: Ti. H: Zhang. 1: Zhou. 0: Tiang. G. (2014], The mechanism of
immunosuppression by perfluorooctanoic acid in BALB/c mice. Toxicology Research 3: 205-
213. http://dx.doi.org/10.1039/c3tx50096a
Wilhelm. M: Bergmann. S: Dieter. HH. (2010], Occurrence of perfluorinated compounds (PFCs] in
drinking water of North Rhine-Westphalia, Germany and new approach to assess drinking
water contamination by shorter-chained C4-C7 PFCs. IntJ HygEnviron Health 213: 224-
232. http://dx.doi.Org/10.1016/i.iiheh.2010.05.004
Wolf. CI: Rider. CV: Lau. C: Abbott. BP. (2014], Evaluating the additivity of perfluoroalkyl acids in
binary combinations on peroxisome proliferator-activated receptor-a activation. Toxicology
316: 43-54. http://dx.doi.org/10.1016/i.tox.2013.12.002
Wolf. CI: Takacs. ML: Schmid. IE: Lau. C: Abbott. BP. (2008], Activation of mouse and human
peroxisome proliferator-activated receptor alpha by perfluoroalkyl acids of different
functional groups and chain lengths. Toxicol Sci 106: 162-171.
http://dx.doi.org/10.1093/toxsci/kfhl66
WS (Weston Solutions Inc], (2007], Remedial investigation report. Phase 2. Fluorochemical (FC]
data assessment report for the Cottage Grove, MN site. St. Paul, MN: 3M Corporate
Toxicology, https://www.pca.state.mn.us/sites/default/files/pfc-cottagegrove-
remedialinvestigationreportpdf
Yoo. H: Washington. TW: Tenkins. TM: Ellington. 11. (2011], Quantitative determination of
perfluorochemicals and fluorotelomer alcohols in plants from biosolid-amended fields using
LC/MS/MS and GC/MS. Environ Sci Technol 45: 7985-7990.
http://dx.doi.org/10.1021/esl02972m
Zhao. P: Xia. X: Pong. 1: Xia. N: Tiang. X: Li. Y: Zhu. Y. (2016], Short- and long-chain perfluoroalkyl
substances in the water, suspended particulate matter, and surface sediment of a turbid
river. Sci Total Environ 568: 57-65. http://dx.doi.Org/10.1016/i.scitotenv.2016.05.221
This document is a draft for review purposes only and does not constitute Agency policy.
R-13 PRAFT-P0 NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
APPENDIX A. SUMMARY OF EXISTING TOXICITY VALUE INFORMATION
FOR PERFLUOROBUTANOIC ACID (PFBA), PERFLUOROHEXANOIC ACID
(PFHXA), PERFLUORONONANOIC ACID (PFNA), AND
PERFLUORODECANOIC ACID (PFDA)
Table A-l. Details on derivation of the available health effect reference values for inhalation exposure to selected
per- and polyfluoroalkyl substances (PFAS)
Reference
value name
Duration
PFAS
Reference value
Health
effect
Point of
departure
Qualifier
Source
Uncertainty
factors
Notes on
derivation
Review
status
(mg/m3)
(ppm)
Emergency response
PAC-3
lh
PFBA
3.3 x 101
3.6 x 10°
NR
NR
NR
NR
PAC values
derived via an
approach
developed by
the Department
of Enerev (DOE,
2016)
Final
(DOE, 2018)
PAC-2
lh
PFBA
5.5 x 10°
6.0 x 10 1
Based on
PAC-3
-
-
-
-
Based on PAC-3a
PAC-1
lh
PFBA
5.0 x 10 1
5.5 x 10"2
Based on
PAC-2
-
-
-
-
Based on PAC-2b
General
TCEQ RfC
Chronic
PFBA
1.0 x 10"2
1.1 x 10"3
NR
NR
NR
NR
RfCs developed
with TCEQ's
protocol
(TCEQ, 2012)
Final
(TCEQ,
2018)
PFDA
5.3 x 10"5
2.5 x 10"6
NR
NR
NR
NR
PFNA
2.8 x 10"5
1.4 x 10"6
NR
NR
NR
NR
This document is a draft for review purposes only and does not constitute Agency policy.
A-l DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Reference
value name
Duration
PFAS
Reference value
Health
effect
Point of
departure
Qualifier
Source
Uncertainty
factors
Notes on
derivation
Review
status
(mg/m3)
(ppm)
PFHxS
1.3 x 10"5
7.8 x 10"7
NR
NR
NR
NR
aPAC-2 = PAC-3 6 = 33 mg/m3^ 6 = 5.5 mg/m3.
bPAC-l = PAC-1 t11= 5.5 mg/m3-Ml = 0.5 mg/m3.
NR = not reported; PAC = Protective Action Criteria; PFAS = per- and polyfluoroalkyl substances; PFBA = perfluorobutanoic acid; PFDA = perfluorodecanoic
acid; PFHxS = perfluorohexanesulfonate; PFNA = perfluorononanoic acid; RfC = inhalation reference concentration; TCEQ = Texas Commission on
Environmental Quality; UFA= animal to human variability; UFD = database uncertainty; UFH = interhuman variability; UFL = LOAEL-to-NOAEL adjustment;
UFs = subchronic-to-chronic adjustment.
This document is a draft for review purposes only and does not constitute Agency policy.
A-2 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table A-2. Details on derivation of the available health effect reference values for oral exposure to selected
per- and polyfluoroalkyl substances (PFAS)
Reference
value
name
Duration
PFAS
Reference
value
(mg/kg-d)
Health effect
Point of
departure
Qualifier
Source
Uncertainty
factors
Notes on
derivation
Review
status
EPA RfD
(OW)
Chronic
PFOA
2 x 10"5
Decreased
ossification
and
accelerated
male puberty
in F1 mice
1 mg/kg-d,
38 mg/L serum
0.0053 mg/kg-d
LOAEL
LOAELhed
Lau et al.
(2006)
Total UF = 300
UFa = 3
UFh = 10
UFL= 10
UFs = 1
UFd=1
Average serum
concentration
derived using a
PBPK model
developed by
Wambaugh et
al. (2013)
HED adjusted3
Final
(U.S. EPA,
2016b)
PFOS
2 x 10"5
Reduced body
wt. in F2 rats
0.1 mg/kg-d,
6.26 Hg/mL
0.00051 mg/kg-d
NOAEL
NOAELhed
Luebker
et al.
(2005)
Total UF = 30
UFA =3
UFh = 10
UFL= 1
UFs = 1
UFD = 1
Average serum
concentration
derived using a
PBPK model
developed by
Wambaugh et
al. (2013)
HED adjusted15
Final
(U.S. EPA,
2016a)
EFSA TDI
Chronic
PFOS
1.5 x 10"4
Changes in
lipids and
thyroid
hormones in
monkeys
0.03 mg/kg-d
NOAEL
Seacat et
al. (2002)
Total UF = 200
UFA = 10
UFh = 10
UFD = 2
Final
(EFSA,
2008)
This document is a draft for review purposes only and does not constitute Agency policy.
A-3 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Reference
value
name
Duration
PFAS
Reference
value
(mg/kg-d)
Health effect
Point of
departure
Qualifier
Source
Uncertainty
factors
Notes on
derivation
Review
status
NH DES
RfD
Chronic
PFHxS
9.3 x 10"6
Reduced litter
size in mice
exposed for
14 d
0.3 mg/kg-d,
27,200 ng/L
serum
90.7 ng/mL
NOAEL
Target
human
serum
level
Chang et
al. (2018)
Total UF = 300
UFA = 3
UFh = 10
MFC = 10
Target human
serum level =
BMDL-MJF
Calculatedd
Final
(New
Hampshire
DES, 2019)
PFNA
2.5 x 10"6
Increased
relative liver
wts. in mice
4,900 ng/L serum
16.3 ng/mL serum
BMDL
Target
human
serum
level
Das et al.
(2015)
Total UF = 300
UFA = 3
UFh = 10
MFe= 10
Target human
serum level =
BMDL-MJF
Calculated'
TCEQ RfD
Chronic
PFBA
2.9 x 10"3
NR
NR
NR
NR
RfDs
developed
with
TCEQ's
protocol
(TCEQ, 2012)
Final
(TCEQ,
2018)
PFDA
1.5 x 10"5
NR
NR
NR
NR
PFHxA
3.8 x 10"6
NR
NR
NR
NR
PFHxS
3.8 x 10"6
NR
NR
NR
NR
PFNA
1.2 x 10"5
NR
NR
NR
NR
Australia
Dept. of
Health TDI
Chronic
Combined
PFOSand
PFHxS
2 x 10"5
Decreased
body wt. gain
in FO female
rats
0.1 mg/kg-d,
7.14 Hg/mL
0.0006 mg/kg-d
NOAEL
NOAELhed
Luebker
et al.
(2005)
Total UF = 30
UFA = 3
UFh = 10
HED Adjusted8
Final
(FSANZ,
2016)
This document is a draft for review purposes only and does not constitute Agency policy.
A-4 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Reference
value
name
Duration
PFAS
Reference
value
(mg/kg-d)
Health effect
Point of
departure
Qualifier
Source
Uncertainty
factors
Notes on
derivation
Review
status
Danish EPA
TDI
Chronic
Combined
PFOS,
PFBS,
PFDS,
PFHpA,
PFHxA,
PFHxS,
and PFNA
3 x 10"5
Liver lesions in
male rats
0.033 mg/kg-d
0.0008 mg/kg-d
BMDLio
BMDLhed
Thomford
(2002)
Total UF = 30
UFA = 3
UFh = 10
HED Adjustedh
Pharmaco-
kinetic
adjustments
based on
those in U.S.
EPA (2014a)
Final
(Danish
EPA, 2015)
aLOAELHED = LOAEL x volume of distribution x (In2 -f ti/2) = 38 mg/L x 0.17 L/kg x (0.693 4 839.5 days) = 0.0053 mg/kg-day.
bNOAELHED = NOAEL x volume of distribution x (In2 4 ti/2) = 6.26 ng/mL x 0.23 L/kg x (0.693 4 1,971 days) = 0.00051 mg/kg-day.
cThe modifying factor of 10 is to account for "the limited number of studies on PFHxS, both animal and epidemiological, as well as uncertainty for associated
effects on other physiological processes including the thyroid system."
dRfD = THSL x volume of distribution x (In2 4 ti/2) = 90.7 ng/mL x 0.287 L/kg x (0.693 4 [5.3 years x 365 days/year]) x 1,000 mL/L = 9.3 ng/kg-day.
eThe modifying factor of 10 is to account for "the limited number of studies on PFNA, specifically the lack of information for a serum half-life in humans, as
well as uncertainty for associated effects on other physiological processes including the immune system."
fRfD = THSL x volume of distribution x (In2 4 ti/2) = 16.3 ng/mL x 0.2 L/kg x (0.693 4 [2.5 years x 365 days/year]) x 1,000 mL/L = 2.5 ng/kg-day.
STDI = NOAEL x volume of distribution x (In2 4 ti/2) = 7.14 ng/mL x 0.23 L/kg x (0.693 4 1,971 days) = 0.0006 mg/kg-day.
hBMDLHED = BMDLio^- ([volume of distribution x (In2 4 ti/2Rat)] 4 [volume of distribution x (In2 4 ti/2HUman)]) = 0.033 mg/kg-day -f
([0.23 L/kg x (0.693 4 48 days)] 4 [0.23 L/kg x (0.693 4 1,971 days)]) = 0.0008 mg/kg-day.
BMDL = benchmark dose lower confidence limit; EFSA = European Food Safety Authority; EPA = Environmental Protection Agency; HED = human equivalent
dose; LOAEL = lowest-observed-adverse-effect level; MDH = Minnesota Department of Health; MF = modifying factor; NH DES = New Hampshire Department
of Environmental Services; NOAEL = no-observed-adverse-effect level; NR = not reported; OW = Office of Water; PBPK = physiologically based
pharmacokinetic; PFAS = per- and polyfluoroalkyl substances; PFBA = perfluorobutanoic acid; PFBS = perfluorobutane sulfonate; PFDA = perfluorodecanoic
acid; PFHxA = perfluorohexanoic acid; PFHxS = perfluorohexanesulfonate; PFNA = Perfluorononanoic acid; PFOS = perfluorooctane sulfonate;
PK = pharmacokinetic; RfD = oral reference dose; SD = standard deviation; TCEQ = Texas Commission on Environmental Quality; TDI = tolerable daily intake;
THSL = target human serum level; UFA= animal to human variability; UFD = database uncertainty; UFH = interhuman variability; UFL = LOAEL-to-NOAEL
adjustment; UFS = subchronic-to-chronic adjustment.
This document is a draft for review purposes only and does not constitute Agency policy.
A-5 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table A-3. Details on derivation of the available drinking water standards for selected per- and polyfluoroalkyl
substances (PFAS)
Reference
value
name
Duration
PFAS
Reference
value
(|og/L)
Health effect
Point of
departure
Qualifier
Source
Uncertainty
factors
Notes on
derivation
Review
status
EPA DWEL
(OW)
Chronic
PFOA
0.37
Based on RfD
Based on
RfDa
Final
(U.S. EPA,
2016b)
PFOS
0.37
Final
(U.S. EPA,
2016a)
EPA HA
(OW)
Chronic
Combined
PFOA and
PFOS
0.07
Based on PFOA
and PFOS
DWELs
Based on
PFOA and
PFOS
DWELsb
Final
(U.S. EPA,
2016a, b)
MDH HRLC
1-30 d
Subchronic
Chronic
PFBA
7
7
7
Decreased
cholesterol,
serum total
thyroxine, and
dialysis free
thyroxine and
increased
relative
thyroid wt. in
rats
3.01 mg/kg-day
0.38 mg/kg-day
0.0038 mg/kg-
day
BMDLlsd
BMDLhed
RfD
(van
Otterdiik,
2007)
Total
UF= 100
UFA = 3
UFh = 10
UFD = 3
HED
Adjustedd
Calculated6
Final
(MDH,
2018)
This document is a draft for review purposes only and does not constitute Agency policy.
A-6 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Reference
value
name
Duration
PFAS
Reference
value
(|og/L)
Health effect
Point of
departure
Qualifier
Source
Uncertainty
factors
Notes on
derivation
Review
status
MDH HBV
1-30 d
Subchronic
Chronic
PFHxS
0.047
0.047
0.047
Decreased free
and total T4
and
triiodothyroni
ne (T3),
changes in
cholesterol
levels, and
increased
hepatic focal
necrosis in rats
32.4 mg/L
serum
0.00292 mg/kg-
day
0.0000097
mg/kg-day
BMDL
BMDLhed
RfD
(NTP,
2019)
Total
U F = 300
UFA = 3
UFh = 10
UFD= 10
HED
Adjusted'
Developed
from RfD
using
toxicokinetic
model
(Goeden et
al., 2019)
Final
(MDH,
2019)
Mass.
ORSG
Chronic
Combined
PFOA, PFOS,
PFNA, PFHpA,
and PFHxS
0.07
Adopted EPA
HA for PFOA
and PFOS
Adopted
EPA HA for
PFOA and
PFOS
Final
(MassDEP,
2018)
CTDPH
Drinking
Water
Action
Level
Chronic
Combined
PFOA, PFOS,
PFNA, PFHpA,
and PFHxS
0.07
Adopted EPA
HA for PFOA
and PFOS
Adopted
EPA HA for
PFOA and
PFOS
Final
(Connectic
utDPH,
2016)
AK DEC
Action
Level
Chronic
Combined
PFOA, PFOS,
PFNA, PFHpA,
and PFHxS
0.07
Adopted EPA
HA for PFOA
and PFOS
Adopted
EPA HA for
PFOA and
PFOS
Final
(AK DEC,
2018)
NH DES
DWEL
Chronic
PFHxS
0.1691
Based on RfD
--
--
--
--
Based on
RfDg
Final
(New
Hampshire
DES, 2019)
PFNA
0.0455
Based on RfD
--
--
--
--
Based on
RfDh
NH DES
MCL
Chronic
PFHxS
0.085
Based on
DWEL
--
--
--
--
Based on
DWEL1
This document is a draft for review purposes only and does not constitute Agency policy.
A-7 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Reference
value
name
Duration
PFAS
Reference
value
(|og/L)
Health effect
Point of
departure
Qualifier
Source
Uncertainty
factors
Notes on
derivation
Review
status
PFNA
0.023
Based on
DWEL
--
--
--
--
Based on
DWELJ
NJ DEP
MCL
Chronic
PFNA
0.013
Increased liver
wts. in
pregnant mice
4,900 ng/mL
serum
4,900 ng/L
serum
BMDL
Target
human
serum
level
Das et al.
(2015)
Total
UF= 1,000
UFa = 3
UFh = 10
UFS= 10
UFD = 3
Target
human
serum level
= BMDL-MJF
Calculatedk
Final
(NJDWQI,
2015)
VT DEC
Drinking
Water HA
Chronic
Combined
PFOA, PFOS,
PFHxS,
PFHpA, PFNA
0.02
Based on EPA
OW RfD
Based on
EPA OW RfD1
Final
(VT DOH,
2018)
TCEQTier 1
PCL
Chronic
PFBA
71
NR
NR
NR
NR
PCL derived
in
accordance
with TCEQ
protocol
(TCEQ.
2013)
Final
(TCEQ,
2018)
PFDA
0.37
NR
NR
NR
NR
PFHxA
0.093
NR
NR
NR
NR
PFHxS
0.093
NR
NR
NR
NR
PFNA
0.29
NR
NR
NR
NR
Australia
Dept. of
Health
Drinking
Water
Guideline
Chronic
Combined
PFOS and
PFHxS
0.07
Based on TDI
Based on
TDIm
Final
(NHMRC,
2019)
This document is a draft for review purposes only and does not constitute Agency policy.
AS DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Reference
value
name
Duration
PFAS
Reference
value
(|og/L)
Health effect
Point of
departure
Qualifier
Source
Uncertainty
factors
Notes on
derivation
Review
status
Danish EPA
QC
Chronic
Combined
PFOA, PFOS,
PFBS, PFDS,
PFHpA,
PFHxA,
PFHxS, and
PFNA
0.1
Based on TDI
Based on
TDI"
Final
(Danish
EPA, 2015)
GFS
Chronic
PFBA
10
Hepatocellular
hypertrophy
and thyroid
hyperplasia
and
hypertrophy in
rats exposed
for 90 d
6 mg/kg-d
3 ng/kg-d
NOAEL
NOAELhed
Butenhoff
et al.
(2012a)
and van
Otterdiik
(2007)
Total
UF = 250
UFa = 2.5
UFh = 10
UFS= 10
HED
adjusted0
Calculated*5
Final
(von der
Trenck et
al., 2018)
PFNA
0.06
Liver lesions
and
hepatocellular
hypertrophy in
rats
0.025 mg/kg-d
0.0167 ng/kg-d
LOAEL
LOAELhed
Stump et
al. (2008)
Total UF = 30
UFl = 3
UFD= 10
HED
adjustedq
Calculatedr
PFHxA
6
Renal toxicity
and lower
urine pH in
rats exposed
for 104 wk
15 mg/kg-d
1.84 ng/kg-d
NOAEL
NOAELhed
Klaunig et
al. (2015),
NICNAS
(2015),
EIC
(2014),
and
NICNAS
(2014)
Total UF = 25
UFA = 2.5
UFh = 10
HED
adjusted5
Calculated'
This document is a draft for review purposes only and does not constitute Agency policy.
A-9 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Reference
value
name
Duration
PFAS
Reference
value
(|og/L)
Health effect
Point of
departure
Qualifier
Source
Uncertainty
factors
Notes on
derivation
Review
status
PFHxS
0.1
Increased liver
wt.,
hepatocellular
hypertrophy,
thyroid
hyperplasia,
and decreased
hematocrit in
rats exposed
for 42 d
1 mg/kg-d
0.03 ng/kg-d
NOAEL
NOAELhed
Butenhoff
et al.
(2009)
Total
UF = 375
UFa = 2.5
UFh = 10
UFS= 15
HED
adjusted"
Calculated7
UBA HRIVW
Chronic
PFDA
0.1
NR
NR
NR
NR
Based on
estimated
half-lives
Provisional
(von der
Trenck et
al., 2018)
PFHxA
1
NR
NR
NR
NR
Provisional
(Wilhelm et
al., 2010)
PFHxS
0.3
NR
NR
NR
NR
Health
Canada
MAC
Chronic
PFOA
0.2
Hepatocellular
hypertrophy
and increased
liver wt. in rats
0.05 mg/kg-day
0.000521
mg/kg-day
0.000021
mg/kg-day
BMDLio
BMDLhed
TDI
(Perkins et
al., 2004)
Total UF = 25
UFa = 2.5
UFh = 10
HED
Adjusted"
Calculated*
Final
(Health
Canada,
2018c)
This document is a draft for review purposes only and does not constitute Agency policy.
A-10 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Reference
value
name
Duration
PFAS
Reference
value
(|og/L)
Health effect
Point of
departure
Qualifier
Source
Uncertainty
factors
Notes on
derivation
Review
status
PFOS
0.6
Hepatocellular
hypertrophy in
rats
0.021 mg/kg-
day
0.0015 mg/kg-
day
0.00006 mg/kg-
day
NOAEL
NOAELhed
TDI
(Butenhof
fetal.,
2012b)
Total UF = 25
UFa = 2.5
UFh = 10
HED
Adjusted2
Calculated33
Final
(Health
Canada,
2018b)
Health
Canada
Drinking
Water
Screening
Value
Chronic
PFBA
30
NR
NR
NR
NR
NR
Final
(Health
Canada,
2018a)
PFNA
0.02
NR
NR
NR
NR
NR
PFHxA
0.2
Adopted MAC
for PFOA
Adopted
MAC for
PFOA
PFHxS
0.6
Adopted MAC
for PFOS
—
—
—
—
Adopted
MAC for
PFOS
Sweden
MTL
Chronic
Combined
PFOA, PFBS,
PFHpA,
PFHxA,
PFHxS,
PFHpA, and
PFPeA
0.09
Based on EFSA
TDI for PFOS
Based on
EFSA TDI for
PFOSbb
Final
(SNFA,
2018)cc
This document is a draft for review purposes only and does not constitute Agency policy.
A-ll DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
aDWEL = RfD 4 water intake rate = 0.00002 mg/kg-day 4 0.054 L/kg-day = 0.00037 mg/L.
bHA = DWEL x RSC = 0.00037 mg/L x 0.2 = 0.00007 mg/L
cThe HRL was adopted as a drinking water quality standard in Italy.
dBMDLHED = BMDLisd 4 (t1/2 Human ^ 11/2Male Rat) = 3.01 mg/kg-day 4 72 h 4 9.22 h = 0.38 mg/kg-day.
eHRL = RfD x RSC 4 short-term intake rate = 0.0038 mg/kg-day x 0.5 4 0.285 L/kg-day = 0.0067 mg/L x 1,000 |Jg/mg = 7 ng/L-
fBMDLHED = BMDL x volume of distribution x (In2 4 ti/2) = 32.4 mg/L x 0.25 L/kg x 0.693 4 1935 days = 0.00292 mg/kg-day.
gDWEL = RfD 4 water intake rate = 9.3 ng/kg-day 4 0.055 L/kg-day = 169.1 ng/L.
hDWEL = RfD 4 water intake rate = 2.5 ng/kg-day 4 0.055 L/kg-day = 45.5 ng/L.
'MCL = DWEL x RSC = 169.1 ng/L x 0.5 = 85 ng/L.
JMCL = DWEL x RSC = 45.5 ng/L x 0.5 = 23 ng/L.
kMCL = THSL x RSC 4 serunrdrinking water ratio = 4,900 ng/L serum x 0.5 4 (200 ng/L serum 4 1 ng/L drinking water) = 13 ng/L.
'DWHA = RFD x RSC x(U BWAIR) = 0.00002 mg/kg-day x0.2xU 0.175 L/kg-day x 1,000 ng/mL = 0.07 |jg/L.
mDW guideline =TDI x BW x RSC 4 water intake rate = 0.00002 mg/kg-day x 70 kg x 0.14 2 L/day = 0.00007 mg/L x 1,000 ng/mL = 0.07 ng/L.
"QC = TDI x = RSC 4 water intake rate = 0.03 ng/kg-day x 0.14 0.03 L/kg-day = 0.1 ng/L.
°NOAELhed = NOAEL4 (t1/2Human 4 ti/2Rat) 4 UF = 6 mg/kg-day 4 8 4 250 = 0.003 mg/kg-day.
PGFS = NOAELhed x BW x RSC 4 water intake rate = 3 ng/kg-day x 70 kg x 0.14 2 L/day = 10 ng/L-
qNOAELHED = NOAEL4 (ti/2Human 4 ti/2Rat) 4 UF = 0.025 mg/kg-day 4 50 4 30 = 0.0000167 mg/kg-day.
rGFS = NOAELhed x BW x RSC 4 water intake rate = 0.0167 ng/kg-day x 70 kg x 0.14 2 L/day = 0.058 ng/L-
sNOAELHEd = NOAEL 4 (tl/2 Human ¦ tl/2 Rat) 4 UF = 15 mg/kg-day 4 327 4 25 = 0.00184 mg/kg-day.
'GFS = NOAELhed x BW x RSC 4 water intake rate = 1.84 ng/kg-day x 70 kg x 0.14 2 L/day = 6.4 ng/L-
"NOAELhed = NOAEL4 (ti/2Human 4 ti/2Rat) 4 UF = 1 mg/kg-day 4 90 4 375 = 0.00003 mg/kg-day.
VGFS = NOAELhed x BW x RSC 4 water intake rate = 0.03 ng/kg-day x 70 kg x 0.14 2 L/day = 0.1037 ng/L-
"The HRIV for PFHxA was adopted as a drinking water quality standard in Italy.
"BMDLhed = BMDLio 4 species- and dose-specific adjustment factor = 0.05 mg/kg-day 4 96 = 0.000521 mg/kg-day.
yMAC = TDI x BW x RSC 4 water intake rate = 0.000021 mg/kg-day x 70 kg x 0.2 4 1.5 L/day = 0.0002 mg/L.
zNOAELHed = NOAEL 4 species- and dose-specific adjustment factor = 0.021 mg/kg-day 4 14 = 0.0015 mg/kg-day.
aaMAC = TDI x BW x RSC 4 water intake rate = 0.00006 mg/kg-day x 70 kg x 0.2 4 1.5 L/day = 0.0006 mg/L.
bbThe Danish EPA's summary of other existing regulations states that Sweden's MTL was derived "considering an exposure scenario where 10% of [the PFOS
TDI] was allocated to the consumption of infant formula based on drinking water."
ccSource document not available in English.
AK DEC = Alaska Department of Environmental Conservation; BMDL = benchmark dose lower confidence limit; CT DPH = Connecticut Department of Health;
DWEL = drinking water equivalent level; EFSA = European Food Safety Authority; EPA = Environmental Protection Agency; GFS = Geringfugigkeitsschwelle, a
significance threshold for the assessment of contaminated groundwater; HA = health advisory; HBV = health-based value; HED = human equivalent dose;
HRIV = health risk indication value; HRL= health risk limit; LOAEL= lowest-observed-adverse effect- level; MAC = maximum acceptable concentration; Mass.
ORSG = MassDEP Office of Research and Standards Guidelines; MCL = maximum contaminant level; MDH = Minnesota Department of Health; MTL = maximal
tolerable level; NH DES = New Hampshire Department of Environmental Services; NJ DEP = New Jersey Department of Environmental Protection;
NOAEL = no-observed-adverse effect- level; NR = not reported; OW = Office of Water; PBPK = physiologically based pharmacokinetic; PCL = protective
This document is a draft for review purposes only and does not constitute Agency policy.
A-12 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Reference
Reference
value
value
Point of
Uncertainty
Notes on
Review
name
Duration
PFAS
(|og/L)
Health effect
departure
Qualifier
Source
factors
derivation
status
concentration level; PFAS = per and polyfluoroalkyl substances; PFBA = perfluorobutanoic acid; PFBS = perfluorobutane sulfonate; PFDA = perfluorodecanoic
acid; PFHxA = perfluorohexanoic acid; PFHxS = perfluorohexanesulfonate; PFNA = Perfluorononanoic acid; PFOS = perfluorooctane- sulfonate;
PK = pharmacokinetic; QC = quality criterion RfD = oral reference dose; RSC = relative source contribution; TCEQ = Texas Commission on Environmental
Quality; TDI = tolerable daily intake; THSL = target human serum level; UBA = Umweltbundesamt, the German Federal Environment Agency; UFA = animal to
human variability; UFD = database uncertainty; UFH = interhuman variability; UFL = LOAEL-to-NOAEL adjustment; UFS = subchronic-to-chronic adjustment; VT
DEC = Vermont Department of Environmental Conservation.
This document is a draft for review purposes only and does not constitute Agency policy.
A-13 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
APPENDIX B. SEARCH AND SCREENING STRATEGIES
Table B-l. Perfluorobutanoic acid (PFBA) database search strategy
Search
Search strategy
Dates of search
PubMed
Search
terms
375-22-4[rn] OR "Heptafluoro-l-butanoic acid"[tw] OR "Heptafluorobutanoic
acid"[tw] OR "Heptafluorobutyric acid"[tw] OR "Kyselina
heptafluormaselna"[tw] OR "Perfluorobutanoic acid"[tw] OR
"Perfluorobutyric acid"[tw] OR "Perfluoropropanecarboxylic acid"[tw] OR
"2,2,3,3,4,4,4-heptafluoro-Butanoic acid"[tw] OR "Butanoic acid,
2,2,3,3,4,4,4-heptafluoro-"[tw] OR "Butanoic acid, heptafluoro-"[tw] OR
"Perfluoro-n-butanoic acid"[tw] OR "Perfluorobutanoate"[tw] OR
"2,2,3,3,4,4,4-Heptafluorobutanoic acid"[tw] OR "Butyric acid,
heptafluoro-"[tw] OR "Fluorad FC 23"[tw] OR "H 0024"[tw] OR "NSC 820"[tw]
OR «PFBA[tw] OR "FC 23"[tw] OR HFBA[tw]) AND (fluorocarbon*[tw] OR
fluorotelomer*[tw] OR polyfluoro*[tw] OR perfluoro-*[tw] OR
perfluoroa*[tw] OR perfluorob*[tw] OR perfluoroc*[tw] OR perfluorod*[tw]
OR perfluoroe*[tw] OR perfluoroh*[tw] OR perfluoron*[tw] OR
perfluoroo*[tw] OR perfluorop*[tw] OR perfluoros*[tw] OR perfluorou*[tw]
OR perfluorinated[tw] OR fluorinated[tw] OR PFAS[tw] OR PFOS[tw] OR
PFOA[tw]))
No date
limit—7/19/2017
Literature
update
search
terms
(((375-22-4[rn] OR "Heptafluoro-l-butanoic acid"[tw] OR
"Heptafluorobutanoic acid"[tw] OR "Heptafluorobutyric acid"[tw] OR
"Kyselina heptafluormaselna"[tw] OR "Perfluorobutanoic acid"[tw] OR
"Perfluorobutyric acid"[tw] OR "Perfluoropropanecarboxylic acid"[tw] OR
"2,2,3,3,4,4,4-heptafluoro-Butanoic acid"[tw] OR "Butanoic acid,
2,2,3,3,4,4,4-heptafluoro-"[tw] OR "Butanoic acid, heptafluoro-"[tw] OR
"Perfluoro-n-butanoic acid"[tw] OR "Perfluorobutanoate"[tw] OR
"2,2,3,3,4,4,4-Heptafluorobutanoic acid"[tw] OR "Butyric acid,
heptafluoro-"[tw] OR "Fluorad FC 23"[tw] OR "H 0024"[tw] OR "NSC 820"[tw]
OR «PFBA[tw] OR "FC 23"[tw] OR HFBA[tw]) AND (fluorocarbon*[tw] OR
fluorotelomer*[tw] OR polyfluoro*[tw] OR perfluoro-*[tw] OR
perfluoroa*[tw] OR perfluorob*[tw] OR perfluoroc*[tw] OR perfluorod*[tw]
OR perfluoroe*[tw] OR perfluoroh*[tw] OR perfluoron*[tw] OR
perfluoroo*[tw] OR perfluorop*[tw] OR perfluoros*[tw] OR perfluorou*[tw]
OR perfluorinated[tw] OR fluorinated[tw] OR PFAS[tw] OR PFOS[tw] OR
PFOA[tw])) AND ("2017/08/01"[PDAT] : "2018/02/14"[PDAT])
8/1/2017-2/14/2018
This document is a draft for review purposes only and does not constitute Agency policy.
B-l DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Search
Search strategy
Dates of search
Web of Science
Search
terms
TS="Heptafluoro-l-butanoic acid" OR TS="Heptafluorobutanoic acid" OR
TS="Heptafluorobutyric acid" ORTS="Kyselina heptafluormaselna" OR
TS="Perfluorobutanoic acid" ORTS="Perfluorobutyric acid" OR
TS="Perfluoropropanecarboxylic acid" OR
TS="2,2,3,3,4,4,4-heptafluoro-Butanoic acid" OR TS="Butanoic acid,
2,2,3,3,4,4,4-heptafluoro-" ORTS="Butanoic acid, heptafluoro-" OR
TS="Perfluoro-n-butanoic acid" ORTS="Perfluorobutanoate" OR
TS="2,2,3,3,4,4,4-Heptafluorobutanoic acid" OR TS="Butyric acid,
heptafluoro-" OR TS="Fluorad FC 23" ORTS="H 0024" OR TS="NSC 820" OR
(TS=(PFBA OR "FC 23" OR HFBA) AND TS=(fluorocarbon* OR fluorotelomer*
OR polyfluoro* OR perfluoro-* OR perfluoroa* OR perfluorob* OR
perfluoroc* OR perfluorod* OR perfluoroe* OR perfluoroh* OR perfluoron*
OR perfluoroo* OR perfluorop* OR perfluoros* OR perfluorou* OR
perfluorinated ORfluorinated OR PFAS OR PFOS OR PFOA))
No date
limit-7/20/2017
Literature
update
search
terms
((TS="Heptafluoro-l-butanoic acid" ORTS="Heptafluorobutanoic acid" OR
TS="Heptafluorobutyric acid" ORTS="Kyselina heptafluormaselna" OR
TS="Perfluorobutanoic acid" ORTS="Perfluorobutyric acid" OR
TS="Perfluoropropanecarboxylic acid" OR
TS="2,2,3,3,4,4,4-heptafluoro-Butanoic acid" OR TS="Butanoic acid,
2,2,3,3,4,4,4-heptafluoro-" ORTS="Butanoic acid, heptafluoro-" OR
TS="Perfluoro-n-butanoic acid" ORTS="Perfluorobutanoate" OR
TS="2,2,3,3,4,4,4-Heptafluorobutanoic acid" OR TS="Butyric acid,
heptafluoro-" OR TS="Fluorad FC 23" ORTS="H 0024" OR TS="NSC 820") OR
TS=(PFBA OR "FC 23" OR HFBA) AND TS=(fluorocarbon* OR fluorotelomer* OR
polyfluoro* OR perfluoro-* OR perfluoroa* OR perfluorob* OR perfluoroc*
OR perfluorod* OR perfluoroe* OR perfluoroh* OR perfluoron* OR
perfluoroo* OR perfluorop* OR perfluoros* OR perfluorou* OR
perfluorinated ORfluorinated OR PFAS OR PFOS OR PFOA)) AND
PY=2017-2018
2017-2018
Toxline
Search
terms
(375-22-4 [rn] OR "heptafluoro-l-butanoic acid" OR "heptafluorobutanoic
acid" OR "heptafluorobutyric acid" OR "kyselina heptafluormaselna" OR
"perfluorobutanoic acid" OR "perfluorobutyric acid" OR
"perfluoropropanecarboxylic acid" OR "2,2,3,3,4,4,4-heptafluoro-butanoic
acid" OR "butanoic acid 2 2 3 3 4 4 4-heptafluoro-" OR "butanoic acid
heptafluoro-" OR "perfluoro-n-butanoic acid" OR "perfluorobutanoate" OR
"2,2,3,3,4,4,4-heptafluorobutanoic acid" OR "butyric acid heptafluoro-" OR
"fluorad fc 23" OR "h 0024" OR "nsc 820" OR (( pfba OR "fc 23" OR hfba ) AND
(fluorocarbon* OR fluorotelomer* OR polyfluoro* OR perfluoro* OR
perfluorinated ORfluorinated OR pfas OR pfos OR pfoa ))) AND (ANEUPL
[org] OR BIOSIS [org] OR CIS [org] OR DART [org] OR EMIC [org] OR EPIDEM
[org] OR HEEP [org] OR HMTC [org] OR IPA [org] OR RISKLINE [org] OR
MTGABS [org] OR NIOSH [org] OR NTIS [org] OR PESTAB [org] OR PPBIB [org] )
AND NOT PubMed [org] AND NOT pubdart [org]
No date
limit-7/20/2017
This document is a draft for review purposes only and does not constitute Agency policy.
B-2 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Search
Search strategy
Dates of search
Literature
update
search
terms
@AND+@OR+("heptafluoro-l-butanoic
acid"+"heptafluorobutanoic+acid"+"heptafluorobutyric+acid"+"kyselina+hept
afluormaselna"+"perfluorobutanoic+acid"+"perfluorobutyric+acid"+"perfluor
opropanecarboxylic +acid"+"2 2 3 3 4 4
4-heptafluoro-butanoic+acid"+"butanoic+acid+2 2 3 3 4 4
4-heptafluoro-"+"butanoic+acid+heptafluoro-"+"perfluoro-n-butanoic
acid"+"perfluorobutanoate"+"2 2 3 3 4 4
4-heptafluorobutanoic+acid"+"butyric+acid+heptafluoro-"+"fluorad+fc+23"+"
h0024"+"nsc+820"+@TERM+@rn+375-22-4("pfba"+"fc+23"+"hfba"))+(
fluorocarbon*+
fluorotelomer*+polyfluoro*+perfluoro*+perfluorinated+fluorinated+pfas+pfo
s+pfoa)+@RANGE+yr+2017+2018
2017-2018
TSCATS
Search
terms
375-22-4[rn] AND tscats[org]
No date
limit-7/20/2017
This document is a draft for review purposes only and does not constitute Agency policy.
B-3 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table B-2. Perfluorodecanoic acid (PFDA) database search strategy
Search
Search strategy
Dates of search
PubMed
Search
terms
335-76-2[rn] OR "Ndfda"[tw] OR "Nonadecafluoro-n-decanoic acid"[tw] OR
"Nonadecafluorodecanoic acid"[tw] OR "Perfluoro-n-decanoic acid"[tw] OR
"Perfluorodecanoic acid"[tw] OR
"2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-Decanoic acid"[tw]
OR "Decanoic acid,
2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-"[tw] OR "Decanoic
acid, nonadecafluoro-"[tw] OR "Perfluorodecanoate"[tw] OR "PFDeA"[tw] OR
"PFDcA"[tw] OR ("PFDA"[tw] AND (fluorocarbon*[tw]
ORfluorotelomer*[tw] OR polyfluoro*[tw] OR perfluoro-*[tw] OR
perfluoroa*[tw] OR perfluorob*[tw] OR perfluoroc*[tw] OR perfluorod*[tw]
OR perfluoroe*[tw] OR perfluoroh*[tw] OR perfluoron*[tw] OR
perfluoroo*[tw] OR perfluorop*[tw] OR perfluoros*[tw] OR perfluorou*[tw]
OR perfluorinated[tw] OR fluorinated[tw] OR PFAS[tw] OR PFOS[tw] OR
PFOA[tw]))
No date
limit—7/26/2017
Literature
update
search
terms
((335-76-2[rn] OR "Ndfda"[tw] OR "Nonadecafluoro-n-decanoic acid"[tw] OR
"Nonadecafluorodecanoic acid"[tw] OR "Perfluoro-n-decanoic acid"[tw] OR
"Perfluorodecanoic acid"[tw] OR
"2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-Decanoic acid"[tw]
OR "Decanoic acid,
2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-"[tw] OR "Decanoic
acid, nonadecafluoro-"[tw] OR "Perfluorodecanoate"[tw] OR "PFDeA"[tw] OR
"PFDcA"[tw] OR ("PFDA"[tw] AND (fluorocarbon*[tw] OR fluorotelomer*[tw]
OR polyfluoro*[tw] OR perfluoro-*[tw] OR perfluoroa*[tw] OR
perfluorob*[tw] OR perfluoroc*[tw] OR perfluorod*[tw] OR perfluoroe*[tw]
OR perfluoroh*[tw] OR perfluoron*[tw] OR perfluoroo*[tw] OR
perfluorop*[tw] OR perfluoros*[tw] OR perfluorou*[tw] OR
perfluorinated[tw] ORfluorinated[tw] OR PFAS[tw] OR PFOS[tw] OR
PFOA[tw])) AND ("2017/08/01"[Date - Publication] :
"2018/03/01"[Date - Publication])
8/1/2017-2/14/2018
This document is a draft for review purposes only and does not constitute Agency policy.
B-4 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Search
Search strategy
Dates of search
Web of Science
Search
terms
TS="PFDeA" ORTS="PFDcA" ORTS="Ndfda" OR
TS="Nonadecafluoro-n-decanoic acid" OR TS="Nonadecafluorodecanoic acid"
OR TS="Perfluoro-n-decanoic acid" OR TS="Perfluorodecanoic acid" OR
TS="2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-Decanoic acid"
OR TS="Decanoic acid,
2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-" ORTS="Decanoic
acid, nonadecafluoro-" ORTS="Perfluorodecanoate" OR (TS=PFDA AND
TS=(fluorocarbon* OR fluorotelomer* OR polyfluoro* OR perfluoro-* OR
perfluoroa* OR perfluorob* OR perfluoroc* OR perfluorod* OR perfluoroe*
OR perfluoroh* OR perfluoron* OR perfluoroo* OR perfluorop* OR
perfluoros* OR perfluorou* OR perfluorinated OR fluorinated)) OR (TS=PFDA
AND TS=(fluorocarbon* OR fluorotelomer* OR polyfluoro* OR perfluoro-* OR
perfluoroa* OR perfluorob* OR perfluoroc* OR perfluorod* OR perfluoroe*
OR perfluoroh* OR perfluoron* OR perfluoroo* OR perfluorop* OR
perfluoros* OR perfluorou* OR perfluorinated OR fluorinated OR PFAS OR
PFOS OR PFOA
No date
limit—7/26/2017
Literature
update
search
terms
TS="PFDeA" ORTS="PFDcA" ORTS="Ndfda" OR
TS="Nonadecafluoro-n-decanoic acid" OR TS="Nonadecafluorodecanoic acid"
OR TS="Perfluoro-n-decanoic acid" OR TS="Perfluorodecanoic acid" OR
TS="2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-Decanoic acid"
OR TS="Decanoic acid,
2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-" ORTS="Decanoic
acid, nonadecafluoro-" ORTS="Perfluorodecanoate" OR (TS=PFDA AND
TS=(fluorocarbon* OR fluorotelomer* OR polyfluoro* OR perfluoro-* OR
perfluoroa* OR perfluorob* OR perfluoroc* OR perfluorod* OR perfluoroe*
OR perfluoroh* OR perfluoron* OR perfluoroo* OR perfluorop* OR
perfluoros* OR perfluorou* OR perfluorinated OR fluorinated)) OR (TS=PFDA
AND TS=(fluorocarbon* OR fluorotelomer* OR polyfluoro* OR perfluoro-* OR
perfluoroa* OR perfluorob* OR perfluoroc* OR perfluorod* OR perfluoroe*
OR perfluoroh* OR perfluoron* OR perfluoroo* OR perfluorop* OR
perfluoros* OR perfluorou* OR perfluorinated OR fluorinated OR PFAS OR
PFOS OR PFOA)) AND PY=2017-2018
2017-2018
Toxline
Search
terms
(335-76-2 [rn] OR
"2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluorodecanoic acid" OR
"2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-decanoic acid" OR
"decanoic acid 2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-nonadecafluoro-" OR
"decanoic acid nonadecafluoro-" OR "nonadecafluoro-n-decanoic acid" OR
"nonadecafluorodecanoic acid" OR "perfluoro-1-nonanecarboxylic acid" OR
"perfluoro-n-decanoic acid" OR "perfluorocapric acid" OR
"perfluorodecanoate" OR "perfluorodecanoic acid" OR "ndfda" OR "PFDeA"
OR "PFDcA" OR ( pfda AND (fluorocarbon* OR fluorotelomer* OR polyfluoro*
OR perfluoro* OR perfluorinated OR fluorinated OR pfas OR pfos OR pfoa )))
AND (ANEUPL [org] OR BIOS IS [org] OR CIS [org] OR DART [org] OR EMIC [org]
OR EPIDEM [org] OR HEEP [org] OR HMTC [org] OR IPA [org] OR RISKUNE [org]
OR MTGABS [org] OR NIOSH [org] OR NTIS [org] OR PESTAB [org] OR PPBIB
[org]) AND NOT PubMed [org] AND NOT pubdart [org]
No date
limit—7/21/2017
This document is a draft for review purposes only and does not constitute Agency policy.
B-5 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Search
Search strategy
Dates of search
Literature
update
search
terms
2017-2018
TSCATS
Search
terms
335-76-2[rn] AND TSCATS[org]
No date
limit—7/21/2017
This document is a draft for review purposes only and does not constitute Agency policy.
B-6 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table B-3. Perfluorononanoic acid (PFNA) database search strategy
Search
Search strategy
Dates of search
PubMed
Search
terms
"375-95-l"[rn] OR "2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluorononanoic
acid"[tw] OR "Nonanoic acid,
2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluoro-"[tw] OR "Nonanoic acid,
heptadecafluoro-"[tw] OR "Perfluoro-n-nonanoic acid"[tw] OR
"Perfluorononan-l-oic acid"[tw] OR "Perfluorononanoate"[tw] OR
"Perfluorononanoic acid"[tw] OR "Perfluorononanonic acid"[tw] OR
"Perfluoropelargonic acid"[tw] OR "heptadecafluorononanoic acid"[tw] OR
(("PFNA"[tw] OR "C 1800"[tw]) AND (fluorocarbon*[tw] OR fluorotelomer*[tw]
OR polyfluoro*[tw] OR perfluoro-*[tw] OR perfluoroa*[tw] OR perfluorob*[tw]
OR
perfluoroc*[tw] OR perfluorod*[tw] OR perfluoroe*[tw] OR perfluoroh*[tw]
OR perfluoron*[tw] OR perfluoroo*[tw] OR perfluorop*[tw] OR
perfluoros*[tw] OR perfluorou*[tw] OR perfluorinated[tw] ORfluorinated[tw]
OR PFAS[tw] OR PFOS[tw] OR PFOA[tw]))
No date
limit—7/26/2017
Literature
update
and
additional
PFNA
synonyms
search
terms
((("2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluorononanoic acid" [tw] OR
"Nonanoic acid, 2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluoro-" [tw] OR
"Nonanoic acid, heptadecafluoro-" [tw] OR "Perfluoro-n-nonanoic acid" [tw]
OR "Perfluorononan-l-oic acid" [tw] OR "Perfluorononanoate" [tw] OR
"Perfluorononanoic acid" [tw] OR "Perfluorononanonic acid" [tw] OR
"Perfluoropelargonic acid" [tw] OR "heptadecafluorononanoic acid" [tw] OR
"PFNA" [tw] OR "C 1800" [tw] OR "Methyl-nl-Perfluorononanoic acid" [tw] OR
"PFNA-nlCH3" [tw] OR "EINECS 206-801-3" [tw] OR
"Heptadecafluornonansaeure" [tw] OR "Heptadekafluornonansaeure" [tw] OR
"Perfluornonansaeure" [tw] OR "Perfluorononanoic acid (PFNA)" [tw] OR
"UNII-5830Z6S63M" [tw] OR "perfluoro-n-nonanoic acid" [tw] OR
"perfluorononan-l-oic acid" [tw] OR "perfluorononanoic acid" [tw] OR
"Ammonium Perfluorononanoate" [tw] OR "Ammonium perfluorononanoate"
[tw] OR "PFNA-H3N" [tw]))) AND ("2017/01/01"[Date - Publication] :
"3000"[Date- Publication])
1/2017-4/2018
Web of Science
Search
terms
((TS=PFNA ORTS="C 1800") AND TS=(fluorocarbon* OR fluorotelomer* OR
polyfluoro* OR perfluoro-* OR perfluoroa* OR perfluorob* OR perfluoroc* OR
perfluorod* OR perfluoroe* OR perfluoroh* OR perfluoron* OR perfluoroo*
OR perfluorop* OR perfluoros* OR perfluorou* OR perfluorinated OR
fluorinated OR PFAS OR PFOS OR PFOA)) OR
TS="2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluorononanoic acid" OR
TS="Nonanoic acid, 2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluoro-" OR
TS="Nonanoic acid, heptadecafluoro-" ORTS="Perfluoro-n-nonanoic acid" OR
TS="Perfluorononan-l-oic acid" ORTS="Perfluorononanoate" OR
TS="Perfluorononanoic acid" OR TS="Perfluorononanonic acid" OR
TS="Perfluoropelargonic acid" OR TS="heptadecafluorononanoic acid"
No date
limit—7/26/2017
This document is a draft for review purposes only and does not constitute Agency policy.
B-7 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Search
Search strategy
Dates of search
Literature
update
and
additional
PFNA
synonyms
search
terms
(TS="PFNA" OR TS="C 1800" OR
TS="2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluorononanoic acid" OR
TS="Nonanoic acid, 2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluoro-" OR
TS="Nonanoic acid, heptadecafluoro-" OR TS="Methyl-nl-Perfluorononanoic
acid" ORTS="PFNA-nlCH3" ORTS="EINECS 206-801-3" OR
TS="Heptadecafluornonansaeure" OR TS="Heptadekafluornonansaeure" OR
TS="Perfluornonansaeure" OR TS="Perfluorononanoic acid (PFNA)" OR
TS="UNII-5830Z6S63M" ORTS="perfluoro-n-nonanoic acid" OR
TS="perfluorononan-l-oic acid" ORTS="perfluorononanoic acid" OR
TS="Ammonium Perfluorononanoate" ORTS="Ammonium
perfluorononanoate" ORTS="PFNA-H3N") AND PY=2017-2018
2017-2018
Toxline
Search
terms
(( pfna OR "c 1800") AND (fluorocarbon* OR fluorotelomer* OR polyfluoro*
OR perfluoro* OR perfluorinated ORfluorinated OR pfas OR pfos OR pfoa ) OR
"375-95-1" [rn] OR "2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluorononanoic
acid" OR "nonanoic acid 2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluoro-" OR
"nonanoic acid heptadecafluoro-" OR "perfluoro-n-nonanoic acid" OR
"perfluorononan-l-oic acid" OR "perfluorononanoate" OR "perfluorononanoic
acid" OR "perfluorononanonic acid" OR "perfluoropelargonic acid" OR
"heptadecafluorononanoic acid")) AND (ANEUPL [org] OR BIOS IS [org] OR CIS
[org] OR DART [org] OR EMIC [org] OR EPIDEM [org] OR HEEP [org] OR HMTC
[org] OR IPA [org] OR RISKLINE [org] OR MTGABS [org] OR NIOSH [org] OR NTIS
[org] OR PESTAB [org] OR PPBIB [org] ) AND NOT PubMed [org] AND NOT
pubdart [org]
No date
limit—7/26/2017
Literature
update
and
additional
PFNA
synonyms
search
terms
@AND+@OR+(pfna+"c
1800"+fluorocarbon*+"2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluorononan
oic+acid"+"nonanoic+acid+2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,9-heptadecafluoro-"+
"nonanoic+acid+heptadecafluoro-"+"perfluoro-n-nonanoic+acid"+"perfluorono
nan-l-oic+acid"+perfluorononanoate+"perfluorononanoic+acid"+"perfluoropel
argon ic+acid"+"heptadecafluorononanoic+acid"+"Methyl-nl-Perfluorononanoi
c+acid"+"PFNA-nlCH3"+"EINECS
206-801-3"+"Heptadecafluornonansaeure"+"Heptadekafluornonansaeure"+"P
erfluornonansaeure"+"Perfluorononanoic+acid
(PFNA)"+"UNII-5830Z6S63M"+"perfluoro-n-nonanoic+acid"+"perfluorononan-l
-oic+acid"+"perfluorononanoic+acid"+"Ammonium+Perfluorononanoate"+"Am
monium+perfluorononanoate"+" PFNA-H3N"+@TERM+@rn+375-95-l)+@ RAN
GE+yr+2017+2018
2017-2018
TSCATS
Search
terms
"375-95-1" [rn] ANDTSCATS[org]
No date
limit-7/20/2017
This document is a draft for review purposes only and does not constitute Agency policy.
B-8 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Search
Search strategy
Dates of search
Literature
@TERM+@rn+375-95-l+@RANGE+yr+2017+2018
2017-2018
update
and
additional
PFNA
synonyms
search
terms
This document is a draft for review purposes only and does not constitute Agency policy.
B-9 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table B-4. Perfluorohexanoic acid (PFHxA) database search strategy
Search
Search strategy
Dates of search
PubMed
Search
terms
((307-24-4[rn] OR "2,2,3,3,4,4,5,5,6,6,6-undecafluorohexanoic acid"[tw] OR
"2,2,3,3,4,4,5,5,6,6,6-undecafluoro-hexanoic acid"[tw] OR "hexanoic acid,
2,2,3,3,4,4,5,5,6,6,6-undecafluoro-"[tw] OR "hexanoic acid,
undecafluoro-"[tw] OR "perfluorohexanoic acid"[tw] OR
"perfluoro-l-pentanecarboxylic acid"[tw] OR "perfluorocaproic acid"[tw] OR
"perfluorohexanoate"[tw] OR "perfluorohexanoic acid"[tw] OR
"undecafluoro-l-hexanoic acid"[tw] OR "undecafluorocaproic acid"[tw] OR
"undecafluorohexanoic acid"[tw] OR "PFHxA"[tw])) AND
("2017/08/01"[PDAT] : "2018/02/28"[PDAT])
No date
limit—7/21/2017
Literature
update
search
terms
((92612-52-7[EC/RN Number]) OR 355-38-4[EC/RN Number]) OR
2062-98-8[EC/RN Number]) OR "PFHxA_ion"[tw]) OR
"Perfluorohexanoate"[tw]) OR "Hexanoyl fluoride,
2,2,3,3,4,4,5,5,6,6,6-undecafluoro-"[tw]) OR "Hexanoyl fluoride,
undecafluoro-"[tw]) OR "Perfluorohexanoyl fluoride"[tw]) OR
"Undecafluorohexanoyl fluoride"[tw]) OR
"Perfluoro(2-methyl-3-oxahexanoyl) fluoride"[tw]) OR "Propanoyl fluoride,
2,3,3,3-tetrafluoro-2-(l,l,2,2,3,3,3-heptafluoropropoxy)-" [tw]) OR
"Propanoyl fluoride, 2,3,3,3-tetrafluoro-2-(heptafluoropropoxy)-" [tw]) OR
"Propionyl fluoride, tetrafluoro-2-(heptafluoropropoxy)-" [tw]) OR
"2,2,3,3,4,4,5,5,6,6,6-Undecafluorohexanoic acid"[tw]) OR "EINECS
206-196-6"[tw]) OR "NSC 5213"[tw]) OR "Perfluoro-l-pentanecarboxylic
acid"[tw]) OR "Perfluoro-n-hexanoic acid"[tw]) OR "UNII-ZP34Q2220R"[tw])
OR "Undecafluorocaproic acid"[tw]) OR "Ammonium
Perfluorohexanoate"[tw]) OR "PFHxA-H3N"[tw]) OR "PFHxA-Na"[tw]) OR
"Sodium Perfluorohexanoate"[tw]))
8/1/2017-2/14/2018
Web of Science
Search
terms
((TS="2,2,3,3,4,4,5,5,6,6,6-undecafluorohexanoic acid" OR
TS="2,2,3,3,4,4,5,5,6,6,6-undecafluoro-hexanoic acid" OR TS="hexanoic acid,
2,2,3,3,4,4,5,5,6,6,6-undecafluoro-" OR TS="hexanoic acid, undecafluoro-"
ORTS="perfluorohexanoic acid" ORTS="perfluoro-l-pentanecarboxylic acid"
ORTS="perfluorocaproic acid" ORTS="perfluorohexanoate" OR
TS="perfluorohexanoic acid" ORTS="undecafluoro-l-hexanoic acid" OR
TS="undecafluorocaproic acid" OR TS="undecafluorohexanoic acid" OR
TS="PFHxA")) AND PY=2017-2018
No date
limit-7/24/2017
This document is a draft for review purposes only and does not constitute Agency policy.
B-10 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Search
Search strategy
Dates of search
Literature
update
search
terms
TS="PFHxA_ion" ORTS="Perfluorohexanoate" ORTS="Hexanoyl fluoride,
2,2,3,3,4,4,5,5,6,6,6-undecafluoro-" OR TS="Hexanoyl fluoride,
undecafluoro-" ORTS="Perfluorohexanoyl fluoride" OR
TS="Undecafluorohexanoyl fluoride" OR
TS="Perfluoro(2-methyl-3-oxahexanoyl) fluoride" ORTS="Propanoyl fluoride,
2,3,3,3-tetrafluoro-2-(l,l,2,2,3,3,3-heptafluoropropoxy)-" ORTS="Propanoyl
fluoride, 2,3,3,3-tetrafluoro-2-(heptafluoropropoxy)-" OR TS="Propionyl
fluoride, tetrafluoro-2-(heptafluoropropoxy)-" OR
TS="2,2,3,3,4,4,5,5,6,6,6-Undecafluorohexanoic acid" ORTS="EINECS
206-196-6" ORTS="NSC 5213" ORTS="Perfluoro-l-pentanecarboxylic acid"
OR TS="Perfluoro-n-hexanoic acid" OR TS="UNII-ZP34Q2220R" OR
TS="Undecafluorocaproic acid" OR TS="Undecafluorohexanoic acid" OR
TS="Ammonium Perfluorohexanoate" ORTS="PFHxA-H3N" OR
TS="PFHxA-Na" ORTS="Sodium Perfluorohexanoate"
2017-2018
Toxline
Search
terms
@AND+@OR+("2,2,3,3,4,4,5,5,6,6,6-undecafluorohexanoic+acid"+"2,2,3,3,4,
4,5,5,6,6,6-undecafluoro-hexanoic+acid"+"hexanoic+acid+2,2,3,3,4,4,5,5,6,6,
6-undecafluoro-"+"hexanoic+acid+undecafluoro-"+"perfluorohexanoic+acid"
+"perfluoro-l-pentanecarboxylic+acid"+"perfluorocaproic+acid"+"perfluoroh
exanoate"+"perfluorohexanoic
acid"+"undecafluoro-l-hexanoic+acid"+"undecafluorocaproic+acid"+"undec
afluorohexanoic+acid"+"pfhxa"+@TERM+@rn+(307+24+4)+@RANGE+yr+20
17+2018+@NOT+@org+"nih+reporter"
No date
limit—7/21/2017
Literature
update
search
terms
@AND+@OR+("PFHxA_ion"+"Perfluorohexanoate"+"Hexanoyl+fluoride,+2,2,
3,3,4,4,5,5,6,6,6-undecafluoro-"+"Hexanoyl+fluoride,+undecafluoro-"+"Perfl
uorohexanoyl+fluoride"+"Undecafluorohexanoyl+fluoride"+"Perfluoro(2-me
thyl-3-oxahexanoyl)+fluoride"+"Propanoyl+fluoride,+2,3,3,3-tetrafluoro-2-(l,
l,2,2,3,3,3-heptafluoropropoxy)-"+"Propanoyl+fluoride,+2,3,3,3-tetrafluoro-
2-(heptafluoropropoxy)-"+"Propionyl+fluoride,+tetrafluoro-2-(heptafluoropr
opoxy)-"+"2,2,3,3,4,4,5,5,6,6,6-Undecafluorohexanoic+acid"+"EINECS+206+l
96+6"+"NSC+5213"+"Perfluoro-l-pentanecarboxylic+acid"+"Perfluoro-n-hex
anoic+acid"+"UNII-ZP34Q2220R"+"Undecafluorocaproic+acid"+"Undecafluor
ohexanoic+acid"+"Ammonium+Perfluorohexanoate"+"PFHxA-H3N"+"PFHxA-
Na"+"Sodium+Perfluorohexanoate")+@NOT+@org+"nih+reporter"
@AND+@OR+(@TERM+@rn+92612+52+7+@TERM+@rn+355+38+4+@TER
M+@rn+2062+98+8)+@NOT+@org+"nih+reporter"
2017-2018
TSCATS
Search
terms
307-24-4[rn] AND tscats[org]
No date
limit-7/20/2017
Literature
update
search
terms
@AND+@OR+(@TERM+@rn+92612+52+7+@TERM+@rn+355+38+4+@TER
M+@rn+2062+98+8)+@org+tscat
This document is a draft for review purposes only and does not constitute Agency policy.
B-ll DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table B-5. Perfluorohexanesulfonate (PFHxS) database search strategy
Search
Search strategy
Dates of search
PubMed
Search
terms
108427-53-8[rn] OR 355-46-4[rn] OR
"1,1,2,2,3,3,4,4,5,5,6,6,6-Tridecafluorohexane-l-sulfonic acid"[tw] OR
"1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-l-Hexanesulfonic acid"[tw] OR
"1-Hexanesulfonic acid, l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-"[tw] OR
"1-Hexanesulfonic acid, tridecafluoro-"[tw] OR "1-Perfluorohexanesulfonic
acid"[tw] OR "Perfluoro-l-hexanesulfonate"[tw] OR "Perfluorohexane
sulfonic
acid"[tw] OR "Perfluorohexane-1-sulphonic acid"[tw] OR
"Perfluorohexanesulfonate"[tw] OR "Perfluorohexanesulfonic acid"[tw] OR
"Perfluorohexylsulfonate"[tw] OR "Tridecafluorohexanesulfonic acid"[tw] OR
"tridecafluoro-l-Hexanesulfonic acid"[tw] OR "PFHxS"[tw]
No date
limit—7/21/2017
Literature
update
search
terms
((108427-53-8[EC/RN Number]) OR 423-50-7[EC/RN Number]) OR
"1-Hexanesulfonic acid, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-, ion(l-)"[tw])
OR "PFHxS ion(l-)"[tw]) OR "PFHxS_ion"[tw]) OR
"Perfluorohexanesulfonate"[tw]) OR "Tridecafluorohexane-l-sulfonate"[tw])
OR "perfluorohexyl sulfonate"[tw]) OR
"1,1,2,2,3,3,4,4,5,5,6,6,6-Tridecafluoro-l-hexanesulfonyl fluoride"[tw]) OR
"1-Hexanesulfonyl fluoride, l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-"[tw]) OR
"1,1,2,2,3,3,4,4,5,5,6,6,6-Tridecafluoro-l-hexanesulfonic acid"[tw]) OR "EC
206-587-l"[tw]) OR "EINECS 206-587-l"[tw]) OR "PFHS"[tw]) OR
"Perfluorhexan-l-sulfonsaure"[tw]) OR "Perfluorohexane sulfonic acid
(PFHxS)"[tw]) OR "Perfluorohexane-1-sulphonic acid"[tw]) OR "acide
perfluorohexane-l-sulfonique"[tw]) OR "acido
perfluorohexano-l-sulfonico"[tw]) OR "perfluorohexane-1-sulphonic
acid"[tw]) OR "perfluorohexanesulfonic acid"[tw]) OR "Ammonium
Perfluorohexanesulfonate"[tw]) OR "Ammonium
perfluorohexanesulfonate"[tw]) OR "PFHxS-H3N"[tw]) OR "PFHxS-K"[tw]) OR
"Potassium Perfluorohexanesulfonate"[tw]) OR "Potassium
perfluorohexanesulfonate"[tw]) OR "Lithium Perfluorohexanesulfonate"[tw])
OR "Lithium perfluorohexanesulfonate"[tw]) OR "PFHxS-Li"[tw]))
8/1/2017-2/14/2018
Web of Science
Search
terms
TS="l,l,2,2,3,3,4,4,5,5,6,6,6-Tridecafluorohexane-l-sulfonic acid" OR
TS="l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-1-Hexanesulfonic acid" OR
TS="l-Hexanesulfonic acid, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-" OR
TS="l-Hexanesulfonic acid, tridecafluoro-" OR
TS="l-Perfluorohexanesulfonic acid" OR TS="Perfluoro-l-hexanesulfonate"
ORTS="Perfluorohexane sulfonic acid" ORTS="Perfluorohexane-1-sulphonic
acid" ORTS="Perfluorohexanesulfonate" ORTS="Perfluorohexanesulfonic
acid" ORTS="Perfluorohexylsulfonate" ORTS="Tridecafluorohexanesulfonic
acid" OR TS="tridecafluoro-1-Hexanesulfonic acid" ORTS="PFHxS"
No date
limit-7/24/2017
This document is a draft for review purposes only and does not constitute Agency policy.
B-12 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Search
Search strategy
Dates of search
Literature
update
search
terms
TS="l-Hexanesulfonic acid, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-, ion(l-)"
ORTS="PFHxS ion(l-)" ORTS="PFHxS_ion" OR
TS="Perfluorohexanesulfonate" ORTS="Tridecafluorohexane-l-sulfonate"
ORTS="perfluorohexyl sulfonate" OR
TS="l,l,2,2,3,3,4,4,5,5,6,6,6-Tridecafluoro-l-hexanesulfonyl fluoride" OR
TS="l-Hexanesulfonyl fluoride, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-" OR
TS="l,l,2,2,3,3,4,4,5,5,6,6,6-Tridecafluoro-l-hexanesulfonic acid" ORTS="EC
206-587-1" ORTS="EINECS 206-587-1" ORTS="PFHS" OR
TS="Perfluorhexan-l-sulfonsaure" OR TS="Perfluorohexane sulfonic acid
(PFHxS)" OR TS="Perfluorohexane-l-sulphonic acid" OR TS="acide
perfluorohexane-l-sulfonique" OR TS="acido perfluorohexano-l-sulfonico"
ORTS="perfluorohexane-l-sulphonic acid" ORTS="perfluorohexanesulfonic
acid" ORTS="Ammonium Perfluorohexanesulfonate" ORTS="Ammonium
perfluorohexanesulfonate" OR TS="PFHxS-H3N" OR TS="PFHxS-K" OR
TS="Potassium Perfluorohexanesulfonate" ORTS="Potassium
perfluorohexanesulfonate" ORTS="Lithium Perfluorohexanesulfonate" OR
TS="Lithium perfluorohexanesulfonate" ORTS="PFHxS-Li"
2017-2018
Toxline
Search
terms
(108427-53-8[rn] OR 355-46-4[rn] OR
"1,1,2,2,3,3,4,4,5,5,6,6,6-Tridecafluorohexane-l-sulfonic acid" OR
"1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-l-Hexanesulfonic acid" OR
"1-Hexanesulfonic acid, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-" OR
"1-Hexanesulfonic acid, tridecafluoro-" OR "1-Perfluorohexanesulfonic acid"
OR "Perfluoro-l-hexanesulfonate" OR "Perfluorohexane sulfonic acid" OR
"Perfluorohexane-l-sulphonic acid" OR "Perfluorohexanesulfonate" OR
"Perfluorohexanesulfonic acid" OR "Perfluorohexylsulfonate" OR
"Tridecafluorohexanesulfonic acid" OR "tridecafluoro-1-Hexanesulfonic acid"
OR "PFHxS") AND (ANEUPL [org] OR BIOSIS [org] OR CIS [org] OR DART [org]
OR EMIC [org] OR EPIDEM [org] OR HEEP [org] OR HMTC [org] OR IPA [org]
OR RISKLINE [org] OR MTGABS [org] OR NIOSH [org] OR NTIS [org] OR
PESTAB [org] OR PPBIB [org] ) [not] PubMed [org] [not] pubdart [org]
No date
limit—7/21/2017
Literature
update
search
terms
@AND+@OR+("l-Hexanesulfonic+acid,+l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluor
o-,+ion(l-)"+"PFHxS+ion(l-)"+"PFHxS_ion"+"Perfluorohexanesulfonate"+"Tri
decafluorohexane-l-sulfonate"+"perfluorohexyl+sulfonate"+"l, 1,2,2,3,3,4,4,
5,5,6,6,6-Tridecafluoro-l-hexanesulfonyl+fluoride"+"l-Hexanesulfonyl+fluori
de,+l, l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-"+"l, 1,2,2,3,3,4,4,5,5,6,6,6-Tridec
afluoro-l-hexanesulfonic+acid"+"EC+206-587-l"+"EINECS+206-587-l"+"PFH
S"+"Perfluorhexan-l-sulfonsaure"+"Perfluorohexane+sulfonic+acid+(PFHxS)"
+" Perfluorohexane-l-sulphonic+acid"+"acide+perfluorohexane-l-sulfonique
"+"acido+perfluorohexano-l-sulfonico"+"perfluorohexane-l-sulphonic+acid"
+"perfluorohexanesulfonic+acid"+"Ammonium+Perfluorohexanesulfonate"+
"Ammonium+perfluorohexanesulfonate"+"PFHxS-H3N"+"PFHxS-K"+"Potassi
um+Perfluorohexanesulfonate"+"Potassium+perfluorohexanesulfonate"+"Lit
hium+Perfluorohexanesulfonate"+"Lithium+perfluorohexanesulfonate"+"PF
HxS-Li")+@NOT+@org+"nih+reporter"
@OR+(@TERM+@rn+108427+53+8+@TERM+@rn+423+50+7)+@NOT+@org
+"nih+reporter"
2017-2018
This document is a draft for review purposes only and does not constitute Agency policy.
B-13 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Search
Search strategy
Dates of search
TSCATS
Search
terms
@OR+(@term+@rn+355-46-4+@term+@rn+108427-53-8)
+@AND+@org+tscats
No date
limit—7/21/2017
Literature
update
search
terms
@OR+(@TERM+@rn+"108427+53+8"+@TERM+@rn+"423+50+7")+@org+tsc
ats
2017-2018
This document is a draft for review purposes only and does not constitute Agency policy.
B-14 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table B-6. Title/abstract-level screening criteria for the initial literature
searches
Inclusion criteria
Exclusion criteria
Populations
• Humans
• Standard mammalian animal models, including
rat, mouse, rabbit, guinea pig, hamster, monkey,
dog
• Alternative animal models in standard laboratory
conditions (e.g., Xenopus, zebrafish, minipig)
• Human or animal cells, tissues, or organs (not
whole animals); bacteria, nonmammalian
eukaryotes; other nonmammalian laboratory
species
• Ecological species
Exposures
• Exposure is to a PFAS compound
• Exposure via oral, inhalation, dermal,
intraperitoneal, or intravenous injection routes
• Exposure is measured in air, dust, drinking
water, diet, gavage, injection or via a biomarker
of exposure (PFAS levels in whole blood, serum,
plasma, or breastmilk)
• Study population is not exposed to a
PFAS compound
• Exposure is to a mixture only
Outcomes
• Studies that include a measure of one or more
health effect endpoints, including but not limited
to, effects on reproduction, development,
developmental neurotoxicity, liver, thyroid,
immune system, nervous system, genotoxicity,
and cancer
• In vivo and/or in vitro studies related to toxicity
mechanisms, physiological effects/adverse
outcomes, and studies useful for elucidating
toxic modes of action (MOAs)
• Qualitative or quantitative description of
absorption, distribution, metabolism, excretion,
toxicokinetic and/or toxicodynamic models
(e.g., PBPK, PBTK, PBTK/TD)
• Studies addressing risks to infants, children,
pregnant women, occupational workers, the
elderly, and any other susceptible or
differentially exposed populations
This document is a draft for review purposes only and does not constitute Agency policy.
B-15 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Inclusion criteria
Exclusion criteria
Other
• Structure and physiochemical properties
• Not on topic, including:
• Reviews and regulatory documents
• Abstract only, inadequately reported
abstract, or no abstract and not
considered further because study was
not potentially relevant
• Bioremediation, biodegradation, or
chemical or physical treatment of
PFAS compounds, including
evaluation of wastewater treatment
technologies and methods for
remediation or contaminated water
and soil
• Ecosystem effects
• Studies of environmental fate and
transport of PFAS compounds in
environmental media
• Analytical methods for
detecting/measuring PFAS
compounds in environmental media
and use in sample preparations and
assays
• Studies describing the manufacture
and use of PFAS compounds
• Not chemical specific (studies that do
not involve testing of PFAS
compounds)
• Studies that describe measures of
exposure to PFAS compounds
without data on associated health
effects
MOA = mode of action; PBPK = physiologically based pharmacokinetic; PBTK = physiologically based toxicokinetic;
TD = toxicodynamic; PFAS = per- and polyfluoroalkyl substances.
1
This document is a draft for review purposes only and does not constitute Agency policy.
B-16 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
Table B-7. Example DistillerSR form questions to be used for title/abstract- and full text-level screening for
literature search updates from 2019
Used in title/abstract and full-text screening
Used in full text only
Source of study
Which
If meets PECO and
if not identified
Does the article
If meets PECO,
PFAS did
endocrine outcome,
from database
meet PECO
what type of
If supplemental, what
the study
If meets PECO, which health
which endocrine
Question
search?
criteria?
evidence?
type of information?
report?
outcome(s) apply?
tags apply?
Answer
options
(can
select
multiple
options)
• Source other
than HERO
database
search
• Yes
• No
• Unclear
• Tag as
potentially
relevant
supplemental
information
• Human
• Animal
(mammalia
n models)
• In vitro or in
silico
genotoxicity
• PBPKorPK
model
• In vivo mechanistic or
MOA studies, including
non-PECO routes of
exposure
(e.g., injection) and
populations
(e.g., nonmammalian)
• In vitro or in silico
studies (non-
genotoxicity)
• ADME/
toxicokinetic (excluding
models)
• Exposure assessment
or characterization (no
health outcome)
• PFAS Mixture Study (no
individual PFAS
comparisons)
• Human case reports or
case series
• Ecotoxicity studies
• PFBA
• PFHxA
• PFHxS
• PFNA
• PFDA
• General toxicity, including
body weight, mortality, and
survival
• Cancer
• Cardiovascular, including
serum lipids
• Endocrine (hormone)
• Gastrointestinal
• Genotoxicity
• Growth (early life) and
development
• Hematological, including
non-immune/hepatic/
renal clinical chemistry
measures
• Hepatic, including liver
measures and serum
markers (e.g., ALT; AST)
• Immune/Inflammation
• Musculoskeletal
• Adrenal
• Sex hormones
(e.g., androgen;
estrogen;
progesterone)
• Neuroendocrine
• Pituitary
• Steroidogenesis
• Thyroid
This document is a draft for review purposes only and does not constitute Agency policy.
B-17 DRAFT-DO NOT CITE OR QUOTE
-------
Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA IRIS Assessments
• Environmental fate or
occurrence (including
food)
• Manufacture,
engineering, use,
treatment,
remediation, or
laboratory methods
• Other assessments or
records with no
original data
(e.g., reviews,
editorials,
commentaries)
• Nervous system, including
behavior and sensory
function
• Nutrition and metabolic
• Ocular
• PBPK or PK model
• Renal, including urinary
measures (e.g., protein)
• Reproductive
• Respiratory
• Skin and connective tissue
effects
ADME = absorption, distribution, metabolism, and excretion; ALT = alanine aminotransferase; AST = aspartate transaminase; HERO = Health and
Environmental Research Online; MOA = mode of action; PBPK = physiologically based pharmacokinetic; PECO = populations, exposures, comparators, and
outcomes; PFAS = per- and polyfluoroalkyl substances; PFBA = perfluorobutanoic acid; PFDA = perfluorodecanoic acid; PFHxA = perfluorohexanoic acid;
PFHxS = perfluorohexanesulfonate; PFNA = Perfluorononanoic acid; PK = pharmacokinetic.
This document is a draft for review purposes only and does not constitute Agency policy.
B-18 DRAFT-DO NOT CITE OR QUOTE
------- |