vvEPA
March 2023
EPA Document No.
822P23007
PUBLIC COMMENT DRAFT
Toxicity Assessment and Proposed Maximum
Contaminant Level Goal for Perfluorooctane Sulfonic
Acid (PFOS) in Drinking Water
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
PUBLIC COMMENT DRAFT
Toxicity Assessment and Proposed Maximum Contaminant Level Goal
for Perfluorooctane Sulfonic Acid (PFOS) in Drinking Water
Prepared by:
U.S. Environmental Protection Agency
Office of Water (4304T)
Health and Ecological Criteria Division
Washington, DC 20460
EPA Document Number: EPA 822P23007
March 2023
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Disclaimer
This document is a public comment draft for review purposes only. This information is
distributed solely for the purpose of public comment. It has not been formally disseminated by
the U.S. Environmental Protection Agency. It does not represent and should not be construed to
represent any agency determination or policy. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Acknowledgments
This document was prepared by the Health and Ecological Criteria Division, Office of Science
and Technology, Office of Water (OW) of the U.S. Environmental Protection Agency (EPA).
The agency gratefully acknowledges the valuable contributions of EPA scientists from the OW,
Office of Research and Development (ORD), the Office of Children's Health Protection
(OCHP), and the Office of Land and Emergency Management (OLEM). OW authors of the
document include Brittany Jacobs; Casey Lindberg; Carlye Austin; Kelly Cunningham; Barbara
Soares; Ruth Etzel; and Colleen Flaherty. ORD authors of the document include J. Michael
Wright; Elizabeth Radke; Michael Dzierlenga; Todd Zurlinden; Jacqueline Weinberger; Thomas
Bateson; Hongyu Ru; and Kelly Garcia. OCHP authors of the document include Chris
Brinkerhoff; and Greg Miller (formerly OW). EPA scientists who provided valuable
contributions to the development of the document from OW include Adrienne Keel; Joyce
Donohue (now retired); Amanda Jarvis; James R. Justice; from ORD include Timothy Buckley;
Allen Davis; Peter Egeghy; Elaine Cohen Hubal; Pamela Noyes; Kathleen Newhouse; Ingrid
Druwe; Michelle Angrish; Christopher Lau; Catherine Gibbons; and Paul Schlosser; and from
OLEM includes Stiven Foster. Additional contributions to draft document review from managers
and other scientific experts, including the ORD Toxicity Pathways Workgroup and experts from
the Office of Chemical Safety and Pollution Prevention (OSCPP), are greatly appreciated. The
agency gratefully acknowledges the valuable management oversight and review provided by
Elizabeth Behl (OW); Jamie Strong (formerly OW; currently ORD); Susan Euling (OW);
Kristina Thayer (ORD); Andrew Kraft (ORD); Viktor Morozov (ORD); Vicki Soto (ORD); and
Garland Waleko (ORD).
The systematic review work included in this assessment was prepared in collaboration with ICF
under the U.S. EPA Contracts EP-C-16-011 (Work Assignment Nos. 4-16 and 5-16) and PR-
OW-21-00612 (TO-0060). ICF authors serving as the toxicology and epidemiology technical
leads were Samantha Snow and Sorina Eftim. ICF and subcontractor authors of the assessment
include Kezia Addo; Barrett Allen; Robyn Blain; Lauren Browning; Grace Chappell; Meredith
demons; Jonathan Cohen; Grace Cooney; Ryan Cronk; Katherine Duke; Hannah Eglinton;
Zhenyu Gan; Sagi Enicole Gillera; Rebecca Gray; Joanna Greig; Samantha Goodman; Anthony
Hannani; Samantha Hall; Jessica Jimenez; Anna Kolanowski; Madison Lee; Cynthia Lin;
Alexander Lindahl; Nathan Lothrop; Melissa Miller; Rachel O'Neal; Ashley Peppriell; Mia
Peng; Lisa Prince; Johanna Rochester; Courtney Rosenthal; Amanda Ross; Karen Setty; Sheerin
Shirajan; Raquel Silva; Jenna Sprowles; Wren Tracy; Joanne Trgovcich; Janielle Vidal;
Maricruz Zarco; and Pradeep Raj an (subcontractor).
ICF contributors to this assessment include Sarah Abosede Alii; Tonia Aminone; Caelen
Caspers; Laura Charney; Kathleen Clark; Sarah Colley; Kaylyn Dinh; Julia Finver; Lauren
Fitzharris; Caroline Foster; Jeremy Frye; Angelina Guiducci; Pamela Hartman; Cara Henning;
Audrey Ichida; Caroline Jasperse; Kaedra Jones; Michele Justice; Afroditi Katsigiannakis;
Gillian Laidlaw; Yi Lu; Denyse Marquez Sanchez; Alicia Murphy; Emily Pak; Joei Robertson;
Lucas Rocha Melogno; Andrea Santa-Rios; Alessandria Schumacher; Nkoli Ukpabi; and
Wanchen Xiong.
ii
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Contents
Disclaimer i
Acknowledgments ii
Contents iii
Figures vi
Tables xi
Acronyms and Abbreviations xiii
1 Background 1-1
1.1 National Primary Drinking Water Regulation for Per- and Polyfluoroalkyl
Substances under the Safe Drinking Water Act 1-1
1.2 Background on PFAS 1-1
1.3 Evaluation of PFOS Under SDWA 1-2
1.4 Purpose of this Document 1-3
1.5 Chemical Identity 1-5
1.6 Occurrence Summary 1-6
1.6.1 Biomonitoring 1-6
1.6.2 Ambient Water 1-7
1.6.3 Drinking Water 1-7
2 Summary of Assessment Methods 2-1
2.1 Introduction to the Systematic Review Assessment Methods 2-1
2.1.1 Literature Search 2-2
2.1.2 Literature Screening 2-3
2.1.3 Study Quality Evaluation for Epidemiological Studies and Animal
Toxicological Studies 2-4
2.1.4 Data Extraction 2-4
2.1.5 Evidence Synthesis and Integration 2-5
2.2 Dose-Response Assessment 2-6
2.2.1 Approach to POD and RfD Derivation for Non-Cancer Health Outcomes... 2-6
2.2.2 Cancer Assessment 2-8
2.3 MCLG Derivation 2-10
3 Results of the Health Effects Systematic Review and Toxicokinetics Methods 3-1
3.1 Literature Search and Screening Results 3-1
3.1.1 Results for Epidemiology Studies of PFOS by Health Outcome 3-4
3.1.2 Results for Animal Toxicological Studies of PFOS by Health Outcome 3-4
3.2 Data Extraction Results 3-5
3.3 Toxicokinetic Synthesis 3-5
3.3.1 ADME 3-5
3.3.2 Pharmacokinetic Models 3-14
3.4 Non-Cancer Health Effects Evidence Synthesis and Integration 3-21
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.1 Hepatic 3-21
3.4.2 Immune 3-78
3.4.3 Cardiovascular 3-134
3.4.4 Developmental 3-183
3.4.5 Evidence Synthesis and Integration for Other Non-Cancer Health
Outcomes 3-257
3.5 Cancer Evidence Study Quality Evaluation, Synthesis, Mode of Action Analysis
and Weight of Evidence 3-257
3.5.1 Human Evidence Study Quality Evaluation and Synthesis 3-257
3.5.2 Animal Evidence Study Quality Evaluation and Synthesis 3-264
3.5.3 Mechanistic Evidence Synthesis 3-265
3.5.4 Weight Of Evidence for Carcinogenicity 3-279
3.5.5 Cancer Classification 3-292
4 Dose-Response Assessment 4-1
4.1 Non-Cancer 4-1
4.1.1 Study andEndpoint Selection 4-1
4.1.2 Estimation or Selection of Points of Departure (PODs) for RfD
Derivation 4-13
4.1.3 Pharmacokinetic Modeling Approaches to Convert Administered Dose to
Internal Dose in Animals and Humans 4-17
4.1.4 Application of Pharmacokinetic Modeling for Animal-Human
Extrapolation of PFOS Toxicological Endpoints and Dosimetric
Interpretation of Epidemiological Endpoints 4-26
4.1.5 Derivation of Candidate Chronic Oral Reference Doses (RfDs) 4-37
4.1.6 RfD Selection 4-45
4.2 Cancer 4-50
4.2.1 Study Selection 4-50
4.2.2 CSF Development 4-50
4.2.3 CSF Selection 4-51
4.2.4 Application of Age-Dependent Adjustment Factors 4-51
5 MCLG Derivation 5-1
6 Effects Characterization 6-1
6.1 Addressing Uncertainties in the Use of Epidemiological Studies for Quantitative
Dose-Response Analyses 6-1
6.2 Comparisons Between Toxicity Values Derived from Animal Toxicological
Studies and Epidemiological studies 6-4
6.3 Updated Approach to Animal Toxicological RfD Derivation Compared to the
2016 PFOS HESD 6-5
6.4 Reevaluation of the PFOS Carcinogenicity Database 6-7
6.5 Health Outcomes with Evidence Integration Judgments of Evidence Suggests
Bordering on Evidence Indicates 6-10
6.6 Challenges and Uncertainty in Modeling 6-11
iv
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
6.6.1 Modeling of Animal Internal Dosimetry 6-11
6.6.2 Modeling of Human Dosimetry 6-12
6.6.3 Approach of Estimating a Benchmark Dose from a Regression
Coefficient 6-14
6.7 Human Dosimetry Models: Consideration of Alternate Modeling Approaches 6-14
6.8 Sensitive Populations 6-18
6.8.1 Fetuses, Infants, Children 6-18
6.8.2 Other Susceptible Populations 6-19
7 References 7-20
v
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Figures
Figure 3-1. Summary of Literature Search and Screening Process for PFOS 3-3
Figure 3-2. Summary of Epidemiology Studies of PFOS Exposure by Health System and
Study Designa 3-4
Figure 3-3. Summary of Animal Toxicological Studies of PFOS Exposure by Health
System, Study Design, and Speciesa'b 3-5
Figure 3-4. Schematic for a Physiologically Motivated Renal Resorption PK Model 3-18
Figure 3-5. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS and
Hepatic Effects 3-22
Figure 3-6. Overall ALT Levels from Pre-2016 HESD Epidemiology Studies Following
Exposure to PFOS 3-23
Figure 3-7. Summary of Study Evaluation for Epidemiology Studies of PFOS and Hepatic
I •fleets" 3-25
Figure 3-8. Overall ALT Levels from Epidemiology Studies Following Exposure to PFOS... 3-27
Figure 3-9. Odds of Elevated ALT Levels from Epidemiology Studies Following Exposure
to PFOS 3-28
Figure 3-10. Summary of Study Evaluation for Animal Toxicological Studies of PFOS and
Hepatic Effectsa'b 3-30
Figure 3-11. Summary of Study Evaluation for Animal Toxicological Studies of PFOS and
Hepatic Effects (Continued) a'b 3-31
Figure 3-12. Percent Change in Serum Enzyme Levels Relative to Controls in Mice
Following Exposure to PFOSa'b 3-34
Figure 3-13. Percent Change in Serum Enzyme Levels Relative to Controls in Male Rats
Following Exposure to PFOSa'b 3-36
Figure 3-14. Percent Change in Serum Enzyme Levels Relative to Controls in Female Rats
Following Exposure to PFOSa'b 3-39
Figure 3-15. Summary of Mechanistic Studies of PFOS and Hepatic Effects 3-43
Figure 3-16. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS
and Immune Effects 3-79
Figure 3-17. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Immunosuppression Effects 3-80
Figure 3-18. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Immunosuppression Effects (Continued) 3-81
vi
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Figure 3-19. Overall Tetanus Antibody Levels in Children from Epidemiology Studies
Following Exposure to PFOS 3-83
Figure 3-20. Overall Diphtheria Antibody Levels in Children from Epidemiology Studies
Following Exposure to PFOS 3-84
Figure 3-21. Odds of Being Below the Protective Level Against Diphtheria (Antibody
Concentrations <0.1 IU/mL) from Epidemiology Studies Following Exposure
to PFOS 3-85
Figure 3-22. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Immune Hypersensitivity Effects 3-92
Figure 3-23. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Autoimmune Effects 3-96
Figure 3-24. Summary of Study Evaluation for Toxicology Studies of PFOS and Immune
Effectsa 3-98
Figure 3-25. Percent Change in Thymus Weights Relative to Controls in Rodents
Following Exposure to PFOS 3-101
Figure 3-26. Incidences of Immune Cell Histopathology in Rodents Following Exposure to
PFOS 3-103
Figure 3-27. Splenocyte Cellularity in Rodents Following Exposure to PFOS (logarithmic
scale)a 3-106
Figure 3-28. Thymocyte Cellularity in Rodents Following Exposure to PFOS (logarithmic
scale) 3-107
Figure 3-29. Summary of Mechanistic Studies of PFOS and Immune Effects 3-110
Figure 3-30. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS
and Cardiovascular Effects 3-135
Figure 3-31. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Cardiovascular Effects 3-138
Figure 3-32. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Cardiovascular Effects (Continued) 3-139
Figure 3-33. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS
and Serum Lipids 3-146
Figure 3-34. Summary of Study Evaluation for Epidemiology Studies of PFOS and Serum
Lipids 3-150
Figure 3-35. Summary of Study Evaluation for Epidemiology Studies of PFOS and Serum
Lipids (Continued) 3-151
Figure 3-36. Summary of Study Evaluation for Epidemiology Studies of PFOS and Serum
Lipids (Continued) 3-152
vii
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Figure 3-37. Odds of High Total Cholesterol in Adults from Epidemiology Studies
Following Exposure toPFOS 3-158
Figure 3-38. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure toPFOS 3-159
Figure 3-39. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS (Continued) 3-160
Figure 3-40. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS (Continued) 3-161
Figure 3-41. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS (Continued) 3-162
Figure 3-42. Summary of Study Evaluation for Toxicology Studies of PFOS and
Cardiovascular Effects 3-166
Figure 3-43. Serum Lipid Levels in Animal Models Following Exposure to PFOS 3-168
Figure 3-44. Summary of Mechanistic Studies of PFOS and Cardiovascular Effects 3-169
Figure 3-45. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS
and Developmental Effects 3-184
Figure 3-46. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Weight Effects51 3-191
Figure 3-47. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Weight Effects (Continued)51 3-192
Figure 3-48. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Weight Effects (Continued)51 3-193
Figure 3-49. Overall Mean Birth Weight from Epidemiology Studies Following Exposure
to PFOS 3-195
Figure 3-50. Overall Mean Birth Weight from Epidemiology Studies Following Exposure
to PFOS (Continued) 3-196
Figure 3-51. Overall Mean Birth Weight from Epidemiology Studies Following Exposure
to PFOS (Continued) 3-197
Figure 3-52. Overall Mean Birth Weight from Epidemiology Studies Following Exposure
to PFOS (Continued) 3-198
Figure 3-53. Overall Mean Birth Weight from Epidemiology Studies Following Exposure
to PFOS (Continued) 3-199
Figure 3-54. Overall Mean Birth Weight from Epidemiology Studies Following Exposure
to PFOS (Continued) 3-200
Figure 3-55. Odds of Small-for-gestational-age in Children from High Confidence
Epidemiology Studies Following Exposure to PFOS 3-204
viii
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Figure 3-56. Odds of Small-for-gestational-age in Children from Medium Confidence
Epidemiology Studies Following Exposure to PFOS 3-205
Figure 3-57. Odds of Low Birthweight in Children from Epidemiology Studies Following
Exposure to PFOS 3-206
Figure 3-58. Summary of Study Evaluation for Epidemiology Studies of PFOS and Low
Birth Weight or Small for Gestational Age Effects 3-207
Figure 3-59. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Length Effectsa 3-209
Figure 3-60. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Length Effects (Continued)a 3-210
Figure 3-61. Summary of Study Evaluation for Epidemiology Studies of PFOS and Head
Circumference Effects 3-213
Figure 3-62. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Postnatal Growth Effects 3-219
Figure 3-63. Summary of Study Evaluation for Epidemiology Studies of PFOS and Preterm
Birth Effects 3-223
Figure 3-64. Summary of Study Evaluation for Epidemiology Studies of PFOS and Fetal
Loss Effects 3-226
Figure 3-65. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Defect Effects 3-227
Figure 3-66. Summary of Study Evaluation for Toxicology Studies of PFOS and
Developmental Effects 3-228
Figure 3-67. Maternal Body Weight in Mice, Rats, and Rabbits Following Exposure to
PFOS 3-230
Figure 3-68. Mortality and Viability in Mice, Rats, and Rabbits Following Exposure to
PFOS (logarithmic scale) 3-232
Figure 3-69. Mortality and Viability in Mice, Rats, and Rabbits Following Exposure to
PFOS (Continued, logarithmic scale) 3-233
Figure 3-70. Offspring Weight in Mice, Rats, and Rabbits Following Exposure to PFOS
(logarithmic scale, sorted by observation time) 3-236
Figure 3-71. Summary of Mechanistic Studies of PFOS and Developmental Effects 3-238
Figure 3-72. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS
and Cancer Effects 3-259
Figure 3-73. Summary of Study Evaluation for Epidemiology Studies of PFOS and Cancer
Effects 3-261
ix
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Figure 3-74. Summary of Study Evaluation for Toxicology Studies of PFOS and Cancer
Effects 3-264
Figure 3-75. Summary of Mechanistic Studies of PFOS and Cancer Effects 3-265
Figure 4-1. Model Structure for Life Stage Modeling 4-20
Figure 4-2. Gestation/Lactation Predictions of PFOS in the Rat 4-22
Figure 4-3. Comparison of Candidate RfDs Resulting from the Application of Uncertainty
Factors to PODheds Derived from Epidemiological and Animal Toxicological
Studies 4-46
Figure 4-4. Schematic depicting selection of the overall RfD for PFOS 4-49
x
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Tables
Table 1-1. Chemical and Physical Properties of PFOS 1-6
Table 3-1. Database Literature Search Results 3-1
Table 3-2. Incidences of Nonneoplastic Lesions in Male and Female Sprague-Dawley Rats,
as Reported by NTP (2019, 5400978) 3-40
Table 3-3. Incidences of Nonneoplastic Lesions in Male and Female Sprague-Dawley Rats,
as Reported by Curran et al. (2008, 757871) 3-40
Table 3-4. Incidences of Nonneoplastic Lesions in Male and Female Sprague-Dawley Rats,
as Reported by Thomford (2002, 5029075) 3-41
Table 3-5. Evidence Profile Table for PFOS Hepatic Effects 3-73
Table 3-6. Associations between PFOS Exposure and Vaccine Response in Faroe Island
Studies 3-86
Table 3-7. Associations Between PFOS Exposure and Natural Killer Cell Activity in Mice. 3-105
Table 3-8. Associations Between PFOS Exposure and Immune Response in Mice 3-108
Table 3-9. Effects of PFOS Exposure on Pro-Inflammatory Cytokines and Markers of
Inflammation 3-118
Table 3-10. Evidence Profile Table for PFOS Immune Effects 3-125
Table 3-11. Evidence Profile Table for PFOS Cardiovascular Effects 3-176
Table 3-12. Evidence Profile Table for PFOS Developmental Effects 3-249
Table 3-13. Incidences51 of Hepatocellular and Pancreatic Tumors in Male and Female
Sprague-Dawley Rats as Reported by Thomford (2002, 5029075) 3-264
Table 3-14. Mutagenicity Data from In Vivo Studies 3-269
Table 3-15. Mutagenicity Data from In Vitro Studies 3-270
Table 3-16. DNA Damage Data from In Vivo Studies 3-271
Table 3-17. DNA Damage Data from In Vitro Studies 3-272
Table 3-18. Evidence of Key Events Associated with the PPARa Mode of Action in Male
Sprague-Dawley Rats Exposed to PFOS 3-282
Table 3-19. Evidence of Key Events Associated with the PPARa Mode of Action in
Female Sprague-Dawley Rats Exposed to PFOS 3-283
Table 3-20. Evidence of Key Events Associated with the CAR Mode of Action in Male
Sprague-Dawley Rats Exposed to PFOS 3-286
Table 3-21. Evidence of Key Events Associated with the CAR Mode of Action in Female
Sprague-Dawley Rats Exposed to PFOS 3-286
xi
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 3-22. Evidence of Key Events Associated with the Cytotoxicity Mode of Action in
Male Sprague-Dawley Rats 3-288
Table 3-23. Evidence of Key Events Associated with the Cytotoxicity Mode of Action in
Female Sprague-Dawley Rats 3-288
Table 3-24. Incidences of Liver Tumor and Nonneoplastic Lesions in Male Sprague-
Dawley Rats at 103 weeks, as Reported by Thomford (2002, 5029075) 3-289
Table 3-25. Incidences of Liver Tumor and Nonneoplastic Lesions in Female Sprague-
Dawley Rats at 103 weeks, as Reported by Thomford (2002, 5029075) 3-289
Table 3-26. Comparison of the PFOS Carcinogenicity Database with the Likely Cancer
Descriptor as Described in the Guidelines for Carcinogen Risk Assessment
{U.S. EPA, 2005, 6324329} 3-294
Table 4-1. Summary of Endpoints and Studies Considered for Dose-Response Modeling
and Derivation of Points of Departure for All Effects in Humans and Rodents 4-10
Table 4-2. Benchmark Response Levels Selected for BMD Modeling of Health Outcomes.... 4-15
Table 4-3. PK Parameters from Wambaugh et al. (2013, 2850932) Meta-Analysis of
Literature Data for PFOS 4-18
Table 4-4. Model Predicted and Literature PK Parameter Comparisons for PFOS 4-19
Table 4-5. Additional PK Parameters for Gestation/Lactation for PFOS 4-21
Table 4-6. Updated and Original Chemical-Specific Parameters for PFOS in Humans 4-24
Table 4-7. Summary of Studies Reporting the Ratio of PFOS Levels in Breastmilk and
Maternal Serum or Plasma 4-25
Table 4-8. PODheds Considered for the Derivation of Candidate RfD Values 4-27
Table 4-9. Uncertainty Factors for the Development of the Candidate Chronic RfD Values
from Epidemiological Studies {U.S. EPA, 2002, 88824} 4-39
Table 4-10. Uncertainty Factors for the Development of the Candidate Chronic RfD Values
from Animal Toxicological Studies {U.S. EPA, 2002, 88824} 4-41
Table 4-11. Candidate Reference Doses (RfDs) 4-43
Table 4-12. Cancer Slope Factors (CSFs) derived from results reported by Butenhoff et al.
(2012, 1276144)/Thomford (2002, 5029075)a in Sprague-Dawley rats 4-51
Table 6-1. Comparison of Candidate RfDs Derived from Animal Toxicological Studies for
Priority Health Outcomes51 6-7
Table 6-2. Comparison of the PFOS Carcinogenicity Database with the Suggestive Cancer
Descriptor as Described in the Guidelines for Carcinogen Risk Assessment
{U.S. EPA, 2005, 6324329} 6-9
xii
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Acronyms and Abbreviations
AASLD
American Association for
the Study of Liver
confidence limit of a 10%
change
Diseases
BMDS
Benchmark Dose
ABC
ATP-binding cassette
Software
transporter
BMI
body mass index
ACG
American College of
BMR
benchmark response
Gastroenterology
BWT
birthweight
ADME
absorption, distribution,
BW
body weight
metabolism, and
Clast7
excretion
average concentration
over the final week of
AF:CB
amniotic fluid and cord
study
AFFF
blood ratio
CAD
coronary artery disease
aqueous film forming
foam
CalEPA
California Environmental
Protection Agency
AhR
aryl hydrocarbon receptor
CAMK
calcium/calmodulin
ALP
alkaline phosphatase
dependent protein kinase
ALSPAC
Avon Longitudinal Study
of Parents and Children
CAR
constitutive androstane
receptor
ALT
alanine aminotransferase
CASRN
Chemical Abstracts
APOB
apolipoprotein B
Service Registry Number
ApoC-III
apolipoprotein C-III
CAT
catalase
ASBT
apical sodium-dependent
Cavg
average blood
bile salt transporter
concentration
AST
aspartate
Cavg,pup,gest
area under the curve
aminotransferase
normalized per day
ATF
activating transcription
during gestation
factor
Cavg, pup, gest,lact
area under the curve
AT SDR
Agency for Toxic
Substances and Disease
normalized dose per day
during gestation/lactation
Registry
Cavg,pup,lact
area under the curve
AUC
area under the curve
normalized per day
BK
bradykinin
during lactation
BM
bone marrow
CCL
Contaminant Candidate
BMD
benchmark dose
List
BMDio
dose corresponding to a
CD
celiac disease
10% change in response
CDC
Centers for Disease
BMDL
benchmark dose lower
Control and Prevention
limits
C-F
carbon-fluorine
BMDLio
dose level corresponding
CHD
coronary heart disease
to the 95% lower
CHDS
Child Ftealth and
Development Studies
xiii
-------
DRAFT FOR PUBLIC COMMENT
CHF
congestive heart failure
CHO
Chinese hamster ovary
CI
confidence interval
CIMT
carotid artery intima-
media thickness
Cmax
maximum blood
concentration
CRP
C-reactive protein
CSF
cancer slope factor
CSM
cholestyramine
CVD
cardiovascular disease
CYP
cytochrome P450
aromatase
CYTL
cytokine like
DBP
diastolic blood pressure
DCFDA
2,7-2,7-
dichlorofluorescein
di acetate
DDIT
DNA damage inducible
transcript
DE
differentially expressed
DIPP
Diabetes Prediction and
Prevention
DMR
differentially methylated
region
DNA
deoxyribonucleic acid
DNBC
Danish National Birth
Cohort
DPP
Diabetes Prevention
Program
DPPOS
Diabetes Prevention
Program and Outcomes
Study
DTH
delayed-type
hypersensitivity response
DWI-BW
body weight-based
drinking water intake
EC
effect concentration
EC50
half maximal effective
concentration
ECM
extracellular matrix
ESC-CM
embryonic stem cell-
derived cardiomyocyte
MARCH 2023
ELISA
enzyme-linked
immunosorbent assay
EPA
U.S. Environmental
Protection Agency
ER
estrogen receptor
ERK
extracellular signal-
regulated protein kinase
Fi
first generation
f2
second generation
FGF
fibroblast growth factor
foe
soil organic carbon
fraction
FXII
Hageman factor XII
GBCA
Genetic and Biomarkers
study for Childhood
Asthma
GD
gestational day
GH
growth hormone
GF
glomerular filtration
GGT
y-glutamyltransferase
GI
gastrointestinal
gist
generalized least-squares
for trend
GSSG
glutathione disulfide
GSH
glutathione
GSH-Px
glutathione peroxidase
HAWC
Health Assessment
Workspace Collaborative
HDL
high density lipoprotein
cholesterol
HED
human equivalent dose
HERO
Health and
Environmental Research
Online
HESD
Health Effects Support
Document
HFD
high fat diet
HFMD
hand, foot, and mouth
disease
HFPO
hexafluoropropy 1 ene
oxide
Hib
Haemophilus influenzae
type b
xiv
-------
DRAFT FOR PUBLIC COMMENT
HIV
human immunodeficiency
virus
HK
high-molecular-weight
kininogen
HMOX
heme oxygenase
HMVEC
human microvascular
endothelial cells
HNF
hepatocyte nuclear factor
HOME
Health Outcomes and
Measures of the
Environment
HR
Hazard Ratio
HRL
health reference level
HSA
human serum albumin
HUVEC
human umbilical cord
endothelial cell
ICAM
intracellular adhesion
molecule
iCOS
inducible co-stimulator
iCOSL
inducible co-stimulator
ligand
IDL
intermediate density
lipoprotein
IgE
immunoglobulin E
IGF
insulin-like growth
factors
IgG
immunoglobulin G
IgM
immunoglobulin M
IHD
ischemic heart disease
IL
interleukin
IP
intraperitoneal
IPA
Ingenuity Pathway
Analysis
IPCS
International Programme
on Chemical Safety
IQR
interquartile range
IRIS
Integrated Risk
Information System
IV
intravenous
JNK
c-JUN amino-terminal
kinase
KC
Kupffer cell
MARCH 2023
KEGG
Kyoto Encyclopedia of
Genes and Genomes
KKS
kallikrein-kinin system
Kh
Henry's Law Constant
KM
Kunming mice
Kmem/w
membrane/water partition
coefficients
KO
knockout
Koc
organic carbon-water
partitioning coefficient
Kow
octanol-water partition
coefficient
LBW
low birthweight
LC
lethal concentration
LCM
liver capsular
macrophages
LC-MS
liquid chromatography-
mass spectrometry
LD
lactational day
LDL
low density lipoprotein
cholesterol
L-FABP
liver fatty acid binding
protein
LOAEL
lowest-observed-adverse-
effect level
LOEC
lowest observed effect
concentration
LOD
limit of detection
LPS
lipopolysaccharide
LSEC
liver sinusoidal
endothelial cell
LXR
liver X receptor
LYZ
lysozyme
MAIT
mucosal invariant T
MALDI
Matrix-Assisted Laser
Desorption/Ionization
MAM
mitochondria-associated
endoplasmic reticulum
membrane
MAPK
mitogen-activated protein
kinase
MCLG
Maximum Contaminant
Level Goal
xv
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
MDA
malondialdehyde
NHANES
National Health and
MDH
Minnesota Department of
Nutrition Examination
Health
Survey
MDM
monocyte-derived
NK
natural killer
macrophages
NOAEL
no-ob served-adverse-
mEB
mouse embryoid body
effect level
MEF
mouse embryonic
NOD
non-obese diabetic
fibroblast
NOS
nitric oxide synthase
MeFOSAA
2-(N-Methyl-
NPDWR
National Primary
perfluorooctane
Drinking Water
sulfonamido) acetic acid
Regulation
MEHP
mono-(2-
NFR
nuclear factor-erythroid
ethylhexyl)phthalate
factor
Me-PFOSA-AcOH 2-(N-Methyl-
NSC
neural stem cells
perfluorooctane
NT
not tested
sulfonamido) acetic acid
NTCP
sodium/taurocholate
miRNA
micro ribonucleic acid
cotransporting
MMR
measles, mumps, and
polypeptide
rubella
NTP
National Toxicology
MOA
mode of action
Program
mPLP
mouse prolactin-like
OAT
organic anion transporter
protein
OATP
organic anion
MRL
Minimum Reporting
transporting polypeptides
Level
OECD
Organisation for
mRNA
messenger ribonucleic
Economic and Co-
acid
operation and
MRP
multidrug resistance-
Development
associated protein
OR
odds ratio
MS
multiple sclerosis
OVA
ovalbumin
MTTP
microsomal triglyceride
Po
parental generation
transfer protein
PBL
peripheral blood
MWCNT
multi-walled carbon
leukocytes
nanotube
PBPK
physiologically based
NAFLD
non-alcoholic fatty liver
pharmacokinetic
disease
PcG
Polycomb group
NCBI
National Center for
PCM
peritoneal macrophages
Biotechnology
PCNA
proliferating cell nuclear
Information
antigen
NCEH
Neutral Cholesterol Ester
PDTC
pyrrolidine
Hydrolase
dithiocarbamate
NCI
National Cancer Institute
PEC AM-1
platelet endothelial cell
NF
nuclear factor
adhesion molecule
xvi
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
PECO
Population, Exposure,
Comparator, and
Outcome
PFAA
perfluoroalkyl acids
PFAS
perfluoroalkyl and
polyfluoroalkyl
substances
PFBA
perfluorobutanoic acid
PFC
plaque forming cell
PFCA
perfluorinated carboxylic
acids
PFDA
perfluorodecanoic acid
PFDoDA
perfluorododecanoic acid
PFHpA
perfluoroheptanoic acid
PFHxA
perfluorohexanoic acid
PFHxS
perfluorohexanesulfonate
PFNA
perfluorononanoic acid
PFOA
perfluorooctanoic acid
PFOS
perfluorooctane sulfonic
acid
PFSA
perfluorosulfonic acid
PHA
phytohemagglutinin
Pion
anionic permeability
PK
pharmacokinetic
Pmilk
milk:blood PFOS
partition coefficient
PND
postnatal day
PNW
postnatal week
POD
point of departure
PODhed
point of departure human
equivalent dose
POUNDS-Lost
Prevention of Obesity
Using Novel Dietary
Strategies Lost
PPAR
peroxisome proliferator
activated receptor
ppm
parts per million
PR
progesterone receptor
PRR
pattern recognition
receptor
PSA
prostate specific antigen
PTB
preterm birth
PTGS
prostaglandin-
endoperoxide synthase
PWS
public water systems
PXR
pregnane X receptor
QA
Quality Assurance
qRT-PCR
quantitative reverse
transcription polymerase
chain reaction
RAR
retinoic acid receptor
RfD
reference dose
Rfm
ratio of the concentrations
in the fetus(es) and the
mother during pregnancy
r'milk
species-specific milk
consumption rate during
lactation for the ith week
of lactation
RNS
reactive nitrogen species
ROS
reactive oxygen species
Rpm
ratio of PFOS in placenta
relative to maternal serum
RSC
relative source
contribution
RSV
respiratory syncytial virus
RXR
retinoid X receptor
SAB
Science Advisory Board
SBP
systolic blood pressure
SD
standard deviation
SDWA
Safe Drinking Water Act
SES
socioeconomic status
SGA
small for gestational age
SGP
sphingosine-1 -posphate
lyase
SHE
Syrian hamster embryo
SIRT
sirtuin
SOD
superoxide dismutase
SRBC
sheep red blood cell
T1D
type 1 diabetes
T-AOC
total antioxidant capacity
TBARS
thiobarbituric acid-
reactive substances
TC
total cholesterol
xvii
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
World Health
Organization
wingless-related
integration site
Web of Science
wild type
World Trade Center
Health Registry
zebrafish liver line
TSCATS
Toxic Substance Control
Act Test Submissions
TTE
transplacental transfer
efficiencies
TUNEL
Terminal
deoxynucleotidyl
transferase dUTP nick
end labeling
UC
ulcerative colitis
UCMR3
Third Unregulated
Contaminant Monitoring
Rule
UF
uncertainty factors
UFa
interspecies uncertainty
factor
UFd
database uncertainty
factor
UFh
intraspecies uncertainty
factor
UFl
LOAEL-to-NOAEL
extrapolation uncertainty
factor
UFs
uncertainty factor for
extrapolation from a
sub chronic to a chronic
exposure duration
UFtot
total uncertainty factors
UV-vis
ultraviolet visible
Vd
volume of distribution
Vfil
filtrate volume
VLDL
very low-density
lipoprotein cholesterol
WBC
white blood cell
TCR T cell receptor WHO
TG triglycerides
THEMIS thymocyte selection WNT
associated
TLR toll-like receptor WoS
TLT TREM-like transcript WT
cells WTCHR
TNF tumor necrosis factor
TNP trinitrophenyl ZFL
xviii
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
1 Background
1.1 National Primary Drinking Water Regulation for Per- and
Polyfluoroalkyl Substances under the Safe Drinking Water
Act
The U.S. Environmental Protection Agency (EPA) has initiated the process to develop a
Maximum Contaminant Level Goal (MCLG) and National Primary Drinking Water Regulation
(NPDWR) for per- and polyfluoroalkyl substances (PFAS), including perfluorooctane sulfonic
acid (PFOS), under the Safe Drinking Water Act (SDWA). As part of the proposed rulemaking,
EPA prepared Proposed Approaches to the Derivation of a Draft Maximum Contaminant Level
Goal for Perfluorooctane Sulfonic Acid (PFOS) (CASRN1763-23-1) in Drinking Water that
described the derivation of oral cancer and noncancer toxicity values, a relative source
contribution (RSC), and cancer classification, which could be subsequently used to derive an
MCLG for PFOS. The agency sought peer review from the EPA Science Advisory Board (SAB)
on key scientific issues related to the development of the MCLG, including the systematic
review approach, oral toxicity values, RSC, and cancer classification.
The SAB provided draft recommendations on June 3, 2022 and final recommendations on
August 23, 2022 {U.S. EPA, 2022, 10476098}, and EPA addressed those recommendations into
the development of this updated assessment, Toxicity Assessment and Proposed Maximum
Contaminant Level Goal for Perfluorooctane Sulfonic Acid (PFOS) in Drinking Water, which
derives toxicity values and an MCLG for PFOS. To be responsive to the SAB recommendations,
EPA has, for example:
• updated and expanded the scope of the studies included in the assessment;
• expanded the systematic review steps beyond study quality evaluation to include evidence
integration to ensure consistent hazard decisions;
• separated hazard identification and dose-response assessment;
• added protocols for all steps of the systematic review and more transparently described the
protocols;
• evaluated alternative pharmacokinetic models and further validated the selected model;
• conducted additional dose-response analyses using additional studies and endpoints;
• evaluated and integrated mechanistic information;
• strengthened the weight of evidence for cancer and rationale for the cancer classification;
• strengthened the rationales for selection of points of departure for the noncancer health
outcomes; and
• clarified language related to the relative source contribution determination including the
relevance of drinking water exposures and the relationship between the reference dose
(RfD) and the relative source contribution.
1.2 Background on PFAS
PFAS are a large group of anthropogenic chemicals that share a common structure of a chain of
linked carbon and fluorine atoms. The PFAS group includes PFOS, perfluorooctanoic acid
(PFOA), and thousands of other chemicals. While the number of PFAS used globally in
1-1
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
commercial products in 2021 was approximately 250 substances {Buck, 2021, 9640864}, the
universe of PFAS, including parent chemicals, metabolites, and degradants, is greater than
12,000 compounds (https://comptox.epa.eov/dashboard/chemical4ists/PFA.SMA.STER). The
Organisation for Economic Co-operation and Development (OECD) New Comprehensive Global
Database of Per- and Polyfluoroalkyl Substances (PFASs), published in 2018, includes over
4,700 PFAS {OECD, 2018, 5099062}.
PFAS have been manufactured and used in a wide variety of industries around the world,
including in the United States since the 1950s. PFAS have strong, stable carbon-fluorine (C-F)
bonds, making them resistant to hydrolysis, photolysis, microbial degradation, and metabolism
{Ahrens, 2011, 2657780; Beach, 2006, 1290843; Buck, 2011, 4771046}. The chemical
structures of PFAS make them repel water and oil, remain chemically and thermally stable, and
exhibit surfactant properties. These properties make PFAS useful for commercial and industrial
applications and purposes and are also the properties that make many PFAS extremely persistent
in the human body and the environment {Calafat, 2007, 1290899; Calafat, 2019, 5381304;
Kwiatkowski, 2020, 7404231}. Due to their widespread use, physicochemical properties,
persistence, and bioaccumulation potential, many PFAS co-occur in exposure media (e.g., air,
water, ice, sediment) as well as in tissues and blood of aquatic and terrestrial organisms,
including humans.
Based on structure, there are many families or classes of PFAS, each containing many individual
structural homologues that can exist as either branched-chain or straight-chain isomers {Buck,
2011, 4771046}. These PFAS families can be divided into two primary categories: non-polymers
and polymers. The non-polymer PFAS include perfluoroalkyl acids (PFAAs), fluorotelomer-
based substances, and per- and polyfluoroalkyl ethers. PFOA and PFOS belong to the PFAA
family of the non-polymer PFAS category and are among the most researched PFAS in terms of
human health toxicity and biomonitoring studies (for review, see Podder et al. (2021, 9640865)).
1.3 Evaluation of PFOS Under SDWA
SDWA, as amended in 1996, requires EPA to publish a list every 5 years of unregulated
contaminants that are not subject to any current proposed or promulgated NPDWRs, are known
or anticipated to occur in public water systems (PWSs), and might require regulation under
SDWA. This list is known as the Contaminant Candidate List (CCL). PFOS is included on the
third CCL (CCL 3) {U.S. EPA, 2009, 1508321} and on the fourth CCL (CCL 4) {U.S. EPA,
2016, 6115068}.
After PFOS and PFOA were listed on the CCL 3 in 2009, EPA initiated development of health
effects support documents (HESDs) for PFOA and PFOS that provided information to federal,
state, tribal, and local officials and managers of drinking water systems charged with protecting
public health when these chemicals are present in drinking water {U.S. EPA, 2016, 3603365;
U.S. EPA, 2016, 3603279}. The two HESDs were peer-reviewed in 2014 and revised based on
consideration of peer reviewers' comments, public comments, and additional studies published
through December 2015. The resulting 2016 Health Effects Support Document for
Perfluorooctane Sulfonic Acid (PFOS) {U.S. EPA, 2016, 3603365} described the assessment of
cancer and noncancer health effects and the derivation of a noncancer RfD that served as the
1-2
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
basis for the non-regulatory 2016 Drinking Water Health Advisory for Perfluorooctane Sulfonic
Acid(PFOS) {U.S. EPA, 2016, 3982043}.
SDWA requires EPA to make regulatory determinations for at least five CCL contaminants
every 5 years. EPA must begin developing an NPDWR when the agency makes a determination
to regulate based on a finding that a contaminant meets all three of the following criteria:
• The contaminant may have an adverse effect on the health of persons.
• The contaminant is known to occur or there is substantial likelihood the contaminant will
occur in PWSs with a frequency and at levels of public health concern.
• In the sole judgment of the Administrator, regulating the contaminant presents a
meaningful opportunity for health risk reductions.
To make these determinations, the agency considers a range of information, including data to
analyze occurrence of these compounds in finished drinking water and data on health effects that
represent the latest science.
In the Final Regulatory Determinations for Contaminants on the Fourth Drinking Water
Contaminant Candidate List {U.S. EPA, 2021, 9640861}, the agency made a determination to
regulate PFOA and PFOS with an NPDWR. The agency concluded that all three criteria were
met—PFOA and PFOS may have adverse health effects; they occur in PWSs with a frequency
and at levels of public health concern; and, in the sole judgment of the Administrator, regulation
of PFOA and PFOS presents a meaningful opportunity for health risk reduction for persons
served by PWSs {U.S. EPA, 2021, 7487276}. As noted above in Section 1.1, EPA prepared
Proposed Approaches to the Derivation of a Draft Maximum Contaminant Level Goal for
Perfluorooctane Sulfonic Acid (PFOS) (CASRN 335-67-1) in Drinking Water as part of this
rulemaking.
In June 2022, EPA published an interim Drinking Water Health Advisory for PFOS {U.S. EPA,
2022, 10668548} to supersede the 2016 Drinking Water Health Advisory based on analyses of
more recent data described in the Proposed Approaches to the Derivation of a Draft Maximum
Contaminant Level Goal for Perfluorooctane Sulfonic Acid (PFOS) (CASRN 1763-23-l)in
Drinking Water, which showed that PFOS can impact human health at exposure levels much
lower than reflected by the 2016 Drinking Water Health Advisory {U.S. EPA, 2016, 3982043;
U.S. EPA, 2022, 10668548}.
1.4 Purpose of this Document
Consistent with SDWA Section 1412(b)(3)(A) and (B), the primary purpose of this draft
document is to obtain public comment on EPA's toxicity assessment and proposed MCLG for
PFOS by describing the best available science on health effects in order to derive an MCLG. To
derive an MCLG, the latest science is identified, described, and evaluated, and then a cancer
classification, toxicity values (i.e., a noncancer RfD and cancer slope factor (CSF)), and RSC for
PFOS are developed (Section 2.3). The draft cancer and noncancer toxicity values, cancer
classification, and RSC values derived in this assessment build upon the work described in the
Proposed Approaches to the Derivation of a Draft Maximum Contaminant Level Goal for
Perfluorooctane Sulfonic Acid (PFOS) (CASRN 1763-23-1) in Drinking Water, the 2016 PFOS
1-3
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
HESD {U.S. EPA, 2016, 3603365}, and the previous 2016 Drinking Water Health Advisory
{U.S. EPA, 2016, 3982043}.
In addition to documenting EPA's basis for the proposed MCLG, this document also serves the
following purposes:
• Transparently describe and document the literature searches conducted and systematic
review methods used to identify health effects information (epidemiological and animal
toxicological studies and physiologically-based pharmacokinetic (PBPK) models) to
update the literature.
• Describe and document screening methods, including the Populations, Exposures,
Comparators, and Outcomes (PECO) criteria and the process for tracking studies
throughout the literature screening.
• Identify epidemiological (i.e., human) and animal toxicological literature that report health
effects after oral exposure to PFOS (and its associated salts), as outlined in the PECO
criteria.
• Evaluate and document the available mechanistic information (including toxicokinetic
understanding) associated with PFOS exposure to inform the interpretation of findings
related to potential health effects in studies of humans and animals with focus on five
main health outcomes (developmental, hepatic, immune, and cardiovascular effects, and
cancer).
• Describe and document the study quality evaluations conducted for epidemiological and
animal toxicological studies considered useful for point of departure (POD) derivation.
• Describe and document the data from high and medium confidence epidemiological and
animal toxicological studies (as determined by study quality evaluations) that were
considered for POD derivation; in cases of health effects with few available studies, data
may be extracted from low confidence studies and used in the evidence syntheses. For
dose-response assessment, only high and medium confidence studies were used to
quantify health effects.
• Synthesize and document the adverse health effects evidence across studies, assessing
health outcomes using a narrative approach. The assessment focuses on synthesizing the
available evidence for five main health outcomes—developmental, hepatic, immune, and
cardiovascular effects, and cancer—but also provides secondary syntheses for dermal,
endocrine, gastrointestinal, hematologic, metabolic, musculoskeletal, nervous, ocular,
renal, and respiratory effects; reproductive effects in males or females; and general
toxicity.
• Develop and document strength of evidence judgments across studies (or subsets of
studies) separately for epidemiological and for animal toxicological lines of evidence and
integrate mechanistic analyses into judgments for the five main health outcomes.
• Develop and document integrated expert judgments across lines of evidence (i.e.,
epidemiological or animal toxicological lines of evidence) as to whether and to what
extent the evidence supports that exposure to PFOS has the potential to be hazardous to
humans. The judgments will be directly informed by the evidence syntheses and based on
1-4
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
structured review of an adapted set of considerations for causality first introduced by
Austin Bradford Hill {Hill, 1965, 71664}.
• Describe and document the dose-response analyses conducted on the studies identified for
POD derivation.
• Derive candidate RfDs and/or CSFs and select the RfD and/or CSF for PFOS and describe
the rationale.
• Determine PFOS's cancer classification using a weight of evidence approach.
• Characterize hazards (e.g., uncertainties, data gaps).
1.5 Chemical Identity
PFOS is a PFAA that was used as an aqueous dispersion agent and emulsifier in a variety of
water-, oil-, and stain-repellent products (e.g., agricultural chemicals, alkaline cleaners, carpets,
firefighting foam, floor polish, textiles) {NLM, 2022, 10369707}. It can exist in linear- or
branched-chain isomeric form. PFOS is a strong acid that is generally present as the sulfonate
anion at typical environmental pH values. Therefore, this assessment applies to all isomers of
PFOS, as well as nonmetal salts of PFOS that would be expected to dissociate in aqueous
solutions of pH ranging from 4 to 9 (e.g., in the human body).
PFOS is stable in environmental media because it is resistant to environmental degradation
processes, such as biodegradation, photolysis, and hydrolysis. In water, no natural degradation
has been demonstrated, and it dissipates by advection, dispersion, and sorption to particulate
matter. PFOS has low volatility in its ionized form but can adsorb to particles and be deposited
on the ground and into water bodies. Because of its persistence, it can be transported long
distances in air or water, as evidenced by detections of PFOS in arctic media and biota, including
polar bears, ocean-going birds, and fish found in remote areas {Lindstrom, 2011, 1290802;
Smithwick, 2006, 1424802}.
Physical and chemical properties and other reference information for PFOS are provided in
Table 1-1. However, there is uncertainty in the estimation, measurement, and/or applicability of
certain physical/chemical properties of PFOS in drinking water, including the Koc {Li, 2018,
4238331; Nguyen, 2020, 7014622}, octanol-water partition coefficient (Kow), and Henry's Law
Constant (KH) {NCBI, 2022, 10411459; AT SDR, 2021, 9642134}. For example, for Kow, the
Agency for Toxic Substances and Disease Registry (ATSDR) (2021, 9642134) reported that a
value could not be measured because PFOS is expected to form multiple layers in octanol/water
mixtures.
For a more detailed discussion related to the chemical and physical properties and environmental
fate of PFOS, please see the PFAS Occurrence & Contaminant Background Technical Support
Document {U.S. EPA, 2023, 10692764}, the 2016 PFOS Drinking Water Health Advisory {U.S.
EPA, 2016, 3 982043 }, and the Draft Aquatic Life Ambient Water Quality Criteria for
Perfluorooctane Sulfonate (PFOS) {U.S. EPA, 2022, 10668582}.
1-5
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 1-1. Chemical and Physical Properties of PFOS
Property
PFOS, Acidic Form;
Experimental Average
Source
Chemical Abstracts Service Registry 1763-23-1
Number (CASRN)"
Chemical Abstracts Index Name
Synonyms
Chemical Formula
Molecular Weight
Color/Physical State
Boiling Point
Melting Point
Vapor Pressure
Henry's Law Constant (KH)
Ko,
Log Iv
Solubility in Water
1,000 ±5.0 L/kg (mean of
values ± 1 standard deviation of
selected values)
4.49
0.0032 mg/L at 25°C;
570 mg/L
NLM, 2022, 10369707
1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,8-
heptadecafhioro-l-octanesulfonic acid
perfluorooctane sulfonic acid;
heptadecafhioro-1 -octane sulfonic
acid; PFOS acid
C8HFi703S
500.13 g/mol
Liquid
249°C
> 400°C
0.002 mm Hg at 25°C
4.1E-04 atm-m3/mol at 25°C
EPA CompTox Chemicals
Dashboard
NLM, 2022, 10369707
NLM, 2022, 10369707
NLM, 2022, 10369707
NLM, 2022, 10369707
ATSDR, 2021, 9642134 (potassium
salt)
NLM, 2022, 10369707 (estimated)
NLM, 2022, 10369707 (estimated
from vapor pressure and water
solubility)
Zareitalabad et al., 2013, 5080561
(converted from log Koc to Koc)
NLM, 2022, 10369707 (estimated)
NLM, 2022, 10369707 (estimated)
ATSDR, 2021, 9642134 (potassium
salt in pure water)
Notes: Koc = organic carbon-water partitioning coefficient; K0w = octanol-water partition coefficient.
a The CASRN given is for linear PFOS, but the toxicity studies are based on both linear and branched; thus, this assessment
applies to all isomers of PFOS.
1.6 Occurrence Summary
1.6.1 Biomonitoring
The U.S. Centers for Disease Control and Prevention (CDC) National Health and Nutrition
Examination Survey (NHANES) has measured blood serum concentrations of several PFAS in
the general U.S. population since 1999. PFOS and PFOA have been detected in up to 98% of
serum samples taken in biomonitoring studies that are representative of the U.S. general
population. Blood levels of PFOA and PFOS dropped 60% to 80% between 1999 and 2014,
presumably due to restrictions on their commercial usage in the United States. Most PFOS
production in the United States was voluntarily phased out by its primary manufacturer (3M)
between 2000 and 2002, and in 2002 and 2007 EPA took regulatory action under TSCA to
require that EPA be notified prior to any future domestic manufacture or importation of PFOS
and 270 related PFAS {U.S. EPA, 2016, 3982043}. Manufacturers have since shifted to
alternative short-chain PFAS, such as hexafluoropropylene oxide (HFPO) dimer acid and its
ammonium salt (two "GenX chemicals"). Additionally, other PFAS were found in human blood
samples from recent (2011-2016) NHANES surveys (e.g., perfluorodecanoic acid (PFDA),
perfluorododecanoic acid (PFDoDA), perfluoroheptanoic acid (PFHpA),
perfluorohexanesulfonate (PFHxS), perfluorononanoic acid (PFNA), and 2-(N-Methyl-
1-6
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
perfluorooctane sulfonamido) acetic acid (Me-PFOSA-AcOH or MeFOSAA)). There is less
publicly available information on the occurrence and health effects of these replacement PFAS
than for PFOS, PFOA, and other members of the carboxylic acid and sulfonate PFAS categories.
1.6.2 Ambient Water
Among the PFAS with established analytical methods for detection, PFOS (along with PFOA) is
one of the dominant PFAS compounds detected in ambient water both in the U.S. and worldwide
{Ahrens, 2011, 2657780; Benskin, 2012, 1274133; Dinglasan-Panlilio, 2014, 2545254;
Nakayama, 2007, 2901973; Remucal, 2019, 5413103; Zareitalabad, 2013, 5080561}. Though it
has a history of wide usage and is highly persistent in aquatic environments, current information
on the distribution of PFOS in surface waters of the United States is somewhat limited; most
published PFOS ambient water occurrence data focuses on regions with known PFAS use or
occurrence. These regions are primarily freshwater systems in eastern states, including the
Mississippi River, Great Lakes, Cape Fear Drainage Basin, and waterbodies near Decatur,
Alabama and in northern Georgia {Jarvis, 2021, 9416544}. Additional monitoring has been
conducted in areas of known aqueous film forming foam (AFFF) use.
In a recent review, Jarvis et al. (2021, 9416544) found that concentrations of PFOS in global
surface waters ranged over eight orders of magnitude, generally in pg/L to ng/L concentrations,
but sometimes reaching |ig/L levels (range: 0.074-8,970,000 ng/L, arithmetic mean:
786.77 ng/L, geometric mean: 5.468 ng/L, median: 3.6 ng/L). Though these calculated
concentrations are not necessarily representative of all the measured PFOS concentrations in
U.S. surface waters, the majority of PFOS concentrations reported (approximately 91%) are less
than 300 ng/L.
1.6.3 Drinking Water
Ingestion of drinking water is a potentially significant source of exposure to PFOS. Serum PFOS
concentrations are known to be elevated among individuals living in communities with drinking
water contaminated from environmental discharges.
Data from the third Unregulated Contaminant Monitoring Rule (UCMR 3) are currently the best
available nationally representative finished water occurrence information for PFOS {U.S. EPA,
2017, 9419085; U.S. EPA, 2021, 7487276; U.S. EPA, 2023, 10692764}. UCMR 3 monitoring
occurred recently (between 2013 and 2015) and analyzed 36,972 samples from 4,920 PWSs for
PFOS. The minimum reporting level (MRL)1 for PFOS was 0.04 |ig/L. A total of 292 samples
from 95 PWSs (out of 36,972 total samples from 4,920 PWSs) had detections of PFOS (i.e.,
greater than or equal to the MRL). PFOS concentrations for these detections ranged from 0.04
|ig/L (the MRL) to 7 |ig/L (median concentration of 0.06 |ig/L; 90th percentile concentration of
0.25 ng/L).
Because PFOS and PFOA cause similar types of adverse health effects and their 2016 lifetime
Health Advisory values were the same, EPA recommended an additive approach when PFOA
and PFOS co-occur at the same time and location in drinking water sources {U.S. EPA, 2016,
1 The minimum reporting level is the threshold at or above which a contaminant's presence or concentration is officially
quantitated. In the case of many of EPA's nation-wide drinking water studies, the selected reporting level is known officially as
the MRL. The MRL for each contaminant in each study is set at a level that EPA believes can be achieved with specified
confidence by abroad spectrum of capable laboratories across the nation {U.S. EPA, 2021, 9640861}.
1-7
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3603365; U.S. EPA, 2016, 3603279}. This approach was used in the analysis for Regulatory
Determination for Contaminants on the Fourth Drinking Water Contaminant Candidate List
{U.S. EPA, 2021, 7487276; U.S. EPA, 2021, 9640861} and the reported maximum summed
concentration of PFOA and PFOS reported in UCMR 3 was 7.22 |ig/L2 and the median summed
value was 0.05 |ig/L. Summed PFOA and PFOS concentrations reported in UCMR 3 exceeded
one-half the health reference level (HRL)3 (0.035 |ig/L) at a minimum of 2.4% of PWSs (115
PWSs) and exceeded the HRL (0.070 |ig/L) at a minimum of 1.3% of PWSs (63 PWSs). Since
the time of UCMR 3 monitoring, some sites where elevated levels of PFOA and PFOS were
previously detected may have installed treatment for PFOA and PFOS, may have chosen to
blend water from multiple sources, or may have otherwise remediated known sources of
contamination. However, the extent of these changes is unknown. The identified 63 PWSs serve
a total population of approximately 5.6 million people and are located across 25 states, tribes, or
U.S. territories {U.S. EPA, 2017, 9419085}.
Data from more recent state monitoring efforts demonstrate occurrence in multiple geographic
locations consistent with UCMR 3 monitoring {U.S. EPA, 2021, 7487276}. IN 2021, at the time
of publication of the final regulatory determinations for PFOA and PFOS, the finished water data
available from fifteen states collected since UCMR 3 identified at least 29 PWSs where the
summed concentrations of PFOA and PFOS exceeded the EPA HRL {U.S. EPA, 2021,
7487276}. The agency notes that some of these data are from targeted sampling efforts and thus
may not be representative of levels found in all PWSs within the state or represent occurrence in
other states. The state data demonstrate occurrence in multiple geographic locations and support
EPA's finding that PFOA and PFOS occur with a frequency and at levels of public health
concern in drinking water systems across the United States.
Likewise, Glassmeyer et al. (2017, 3454569) sampled source and treated drinking water from 29
drinking water treatment plants for a suite of emerging chemical and microbial contaminants,
including 11 PFAS. PFOS was reported in source water at 88% of systems, with a median
concentration of 2.28 ng/L and maximum concentration of 48.30 ng/L. Similarly, in treated
drinking water, PFOS was detected in 80% of systems, with a median concentration of 1.62 ng/L
and maximum concentration of 36.90 ng/L.
2 Sum of PFOA + PFOS results rounded to 2 decimal places in those cases where a laboratory reported more digits.
3 An HRL is a health-based concentration against which the agency evaluates occurrence data when making decisions about
regulatory determinations. The HRL for PFOS that was used to evaluate UCMR 3 results was 0.070 (ig/L (equal to the 2016
Drinking Water Health Advisory value).
1-8
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
2 Summary of Assessment Methods
This section summarizes the methods used for the systematic review of the health literature for
PFOS, PFOA, and their related salts. The purposes of this systematic review were to identify the
best available and most relevant health effects literature, to screen studies for quality, and to
subsequently identify and consider studies that can be used for dose-response assessment. A
detailed description of these methods is provided as a protocol in the Appendix (see PFOS
Appendix).
The information that was gathered in the systematic review described in this document was used
to update EPA's 2016 HESD for PFOS {U.S. EPA, 2016, 3603365} and to derive an MCLG to
support a National Primary Drinking Water regulation under the Safe Drinking Water Act.
2.1 Introduction to the Systematic Review Assessment
Methods
The methods used to conduct the systematic review for PFOS are consistent with the methods
described in the draft and final EPA ORD Staff Handbook for Developing IRIS Assessments
{U.S. EPA, 2020, 7006986; U.S. EPA, 2022, 10367891} (hereafter referred to as the Integrated
Risk Information System (IRIS) Handbook) and a companion publication {Thayer, 2022,
10259560}. EPA's IRIS Handbook has incorporated feedback from the National Academy of
Sciences (NAS) at workshops held in 2018 and 2019 and was well regarded by the NAS review
panel for reflecting "significant improvements made by EPA to the IRIS assessment process,
including systematic review methods for identifying chemical hazards" {NAS, 2021, 9959764}.
Furthermore, EPA's IRIS program has used the IRIS Handbook to develop toxicological reviews
for numerous chemicals, including some PFAS. Though the IRIS Handbook was finalized
concurrently with this assessment, the alterations in the final IRIS Handbook compared to the
draft version did not conflict with the methods used in this assessment. In fact, many of the NAS
recommendations incorporated into the final IRIS handbook (e.g., updated methods for evidence
synthesis and integration) were similarly incorporated into this assessment protocol {NAS, 2021,
9959764}. However, some of the study evaluation refinements recommended by NAS {2021,
9959764}, including clarifications to the procedure for evaluating studies for sensitivity and
standardizing the procedure for evaluating reporting quality between human and animal studies,
were not included in this assessment protocol, consistent with a 2011 NASEM recommendation
not to delay releasing assessments until systematic review methods are finalized {NRC, 2011,
710724}. The assessment team concluded that implementing these minor changes in study
quality evaluation would not change the assessment conclusions. Therefore, EPA considers the
methods described herein to be consistent with the final IRIS Handbook and cites this version
accordingly.
For this updated toxicity assessment, systematic review methods used were comparable to those
in the IRIS Handbook for the steps of literature search, screening, study quality evaluation, data
extraction, and the display of study quality results for all health outcomes through the 2020
literature searches {U.S. EPA, 2022, 10367891}. EPA then focused the subsequent steps of the
systematic review process (synthesis of human, experimental animal, and mechanistic data;
evidence integration; derivation of toxicity values) on health effects outcomes with the strongest
weight of evidence (developmental, hepatic, immune, cardiovascular, and cancer) based on the
2-1
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
conclusions presented in EPA's preliminary analysis, Proposed Approaches to the Derivation of
a Draft Maximum Contaminant Level Goal for Perfluorooctane Sulfonic Acid (PFOS) (CASRN
1763-23-1) in Drinking Water, and consistent with the recommendations of the SAB {U.S. EPA,
2022, 10476098}.
This section provides a summary of methods used to search and screen the literature identified,
evaluate the studies and characterize study quality, extract data, and identify studies that can be
used for dose-response analysis. Extracted data are available in interactive visual formats (see
Section 3) and can be downloaded in open access formats.
The systematic review protocol (see PFOS Appendix) provides a detailed description of the
systematic review methods that were used. The particular focus of the protocol is the description
of the problem formulation and key science issues guiding this assessment.
2.1.1 Literature Search
EPA assembled an inventory of epidemiological, animal toxicological, mechanistic, and
toxicokinetic studies for this updated toxicity assessment based on three data streams: 1)
literature published from 2014 through 2019 and then updated in the course of this review (i.e.,
through February 3, 2022) identified via literature searches of a variety of publicly available
scientific literature databases, 2) literature identified via other sources (e.g., searches of the gray
literature and studies shared with EPA by the SAB), and 3) literature identified in EPA's 2016
HESDs for PFOA and PFOS {U.S. EPA, 2016, 3603279; U.S. EPA, 2016, 3603365}.
The search strings for the new searches for this updated assessment focused on the chemical
name (PFOA, PFOS, and their related salts) with no limitations on lines of evidence (i.e.,
human/epidemiological, animal, in vitro, in silico) or health outcomes. EPA conducted an
updated literature search in 2019 (covering January 2013 through April 11, 2019), which was
subsequently updated by a search covering April 2019 through September 3, 2020 (2020
literature search) and another covering September 2020 through February 3, 2022 (2022
literature search) using the same search strings used in 2019.
The publicly available databases listed below were searched for literature containing the
chemical search terms outlined in the PFOS Appendix:
• Web of Science™ (WoS) (Thomson Reuters),
• PubMed® (National Library of Medicine),
• ToxLine (incorporated into PubMed post 2019), and
• TSCATS (Toxic Substances Control Act Test Submissions).
In addition to the databases above, other review efforts and searches of publicly available
sources were used to identify relevant studies, as listed below:
• studies cited in assessments published by other U.S. federal, international, and/or state
agencies (this included assessments by ATSDR and California Environmental Protection
Agency (CalEPA)),
• studies identified during mechanistic or toxicokinetic synthesis (i.e., during manual review
of reference lists of relevant mechanistic and toxicokinetic studies deemed relevant after
screening against mechanistic- and ADME-specific PECO criteria), and
2-2
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
• studies identified by the SAB in their final report dated August 23, 2022 {U.S. EPA, 2022,
10476098}.
The details of the studies included from the 2016 HESD as well as the search strings and
literature sources searched are described in the Appendix (see PFOS Appendix).
EPA relied on epidemiological and animal toxicological literature identified in the 2016 PFOS
HESD to identify studies for this updated assessment on five major health outcomes, as
recommended by SAB and consistent with EPA's preliminary analysis in the Proposed
Approaches to the Derivation of a Draft Maximum Contaminant Level Goal for Perfluorooctane
Sulfonic Acid (PFOS) (CASRN1763-23-1) in Drinking Water. The 2016 HESD for PFOS
contained a summary of all relevant literature identified in searches conducted through 2013. The
RfD derived in EPA's 2016 HESD was based quantitatively on animal toxicological studies
whereas epidemiology studies were considered qualitatively, as a supporting line of evidence.
This updated assessment includes the study quality evaluation of epidemiological studies that
were identified and included in the 2016 HESD for the five health outcomes that had the
strongest evidence. It also includes "key" animal toxicological studies from the HESD, which
includes studies that were selected in 2016 for dose-response modeling. More details are
provided in the Appendix (See PFOS Appendix).
All studies identified in the literature searches as well as those brought forward from the 2016
PFOS HESD were uploaded into the Health and Environmental Research Online (HERO)
database (https://hero.epa.eov/hero/index.cfm/proiect/paee/proiect id/2608) and are publicly
available.
EPA has continued to monitor the literature published since February 2022 for other potentially
relevant studies published after the 2022 literature search update Potentially relevant studies
identified after February 2022 that were not recommended by the SAB in their final report are
not included as part of the evidence base for this updated assessment but are provided in a
repository detailing the results and potential impacts of new literature on the assessment (See
PFOS Appendix A.3).
2.1.2 Literature Screening
This section summarizes the methods used to screen the identified health effects, mechanistic,
and absorption, distribution, metabolism, excretion (ADME) literature. Briefly, PECO statements
were established and detail the criteria used to screen all of the literature identified from
literature searches in this assessment, prioritize the dose-response literature for dose-response
assessment, and identify supplemental studies that may inform key science questions described
in the protocol. The PECO criteria used for screening the dose-response, toxicokinetic, and
mechanistic literature are provided in the Appendix (See PFOS Appendix).
Consistent with protocols outlined in the IRIS Handbook {U.S. EPA, 2022, 10367891}, studies
identified in the literature searches and stored in HERO were imported into the Swift-Review
software platform and the software was used to identify those studies most likely to be relevant
to human health risk assessment. Studies captured then underwent title and abstract screening by
at least two reviewers using DistillerSR or SWIFT ActiveScreener software, and studies that
passed this screening underwent full-text review. Dose-response studies that met PECO inclusion
2-3
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
criteria following both title and abstract screening and full-text review underwent study quality
evaluation as described below. Studies tagged as supplemental and containing potentially
relevant mechanistic or ADME (or toxicokinetic) data following title and abstract and full-text
level screening underwent further screening using mechanistic- or ADME-specific PECO
criteria, and those deemed relevant underwent light data extraction of key study elements (e.g.,
extraction of information about the tested species or population, mechanistic or ADME
endpoints evaluated, dose levels tested; see PFOS Appendix). Supplemental studies that were
identified as mechanistic or ADME via screening did not undergo study quality evaluation.
2.1.3 Study Quality Evaluation for Epidemiological Studies
and Animal Toxicological Studies
For study quality evaluation of the PECO-relevant human epidemiological and animal
toxicological studies identified in the three literature searches (all health outcomes for the 2019
and 2020 searches; the five priority health outcomes for the 2022 search), epidemiological
studies from the 2016 HESD that reported results on one or more of the five priority health
outcomes, and key animal toxicological studies from the 2016 HESD, two or more quality
assurance (QA) reviewers, working independently, assigned ratings about the reliability of study
results {good, adequate, deficient (or "not reported"), or critically deficient) for different
evaluation domains. These study quality evaluation domains are listed below and details about
the domains, including prompting questions and suggested considerations, are described in the
PFOS Appendix.
• Epidemiological study quality evaluation domains: participant selection; exposure
measurement criteria; outcome ascertainment; potential confounding; analysis; selective
reporting; and study sensitivity.
• Animal toxicological study quality evaluation domains: reporting; allocation;
observational bias/blinding; confounding/variable control; reporting and attrition bias;
chemical administration and characterization; exposure timing, frequency, and duration;
endpoint sensitivity and specificity; and results presentation.
The independent reviewers performed study quality evaluations using a structured platform
housed within EPA's Health Assessment Workplace Collaboration (HAWC;
https://hawcproi ect.org/). Once the individual domains were rated, reviewers independently
evaluated the identified strengths and limitations of each study to reach an overall classification
on study confidence of high, medium, low, or uninformative for each PECO-relevant endpoint
evaluated in the study. A study can be given an overall mixed confidence classification if
different PECO-relevant endpoints within the study receive different confidence ratings (e.g.,
medium and low confidence classifications).
2.1.4 Data Extraction
Data extraction was conducted for all relevant human epidemiological and animal toxicological
studies determined to be of medium and high confidence after study quality evaluation. Data
were also extracted from low confidence epidemiological studies when data were limited for a
health outcome or when there was a notable effect, consistent with the IRIS Handbook {U.S.
EPA, 2020, 7006986}. Studies evaluated as being uninformative were not considered further and
therefore did not undergo data extraction. All health endpoints were considered for extraction,
2-4
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
regardless of the magnitude of effect or statistical significance of the response relative to the
control group. The level of detail in data extractions for different endpoints within a study could
differ based on how the data were presented for each outcome (i.e., ranging from a narrative to a
full extraction of dose-response effect size information).
Extractions were conducted using DistillerSR for epidemiological studies or HAWC for animal
toxicological studies. An initial reviewer conducted the extraction, followed by an independent
QA review by a second reviewer who confirmed accuracy and edited/corrected the extraction as
needed. Discrepancies in data extraction were resolved by discussion and confirmation within
the extraction team.
Data extracted from epidemiology studies included population, study design, year of data
collection, exposure measurement, and quantitative data from statistical models. Data extracted
from statistical models reported in the studies included the health effect category, endpoint
measured, sample size, description of effect estimate, covariates, and model comments. Data
extracted from animal toxicological studies included information on the experimental design and
exposure duration, species and number of animals tested, dosing regime, and endpoints
measured. Further information about data extraction can be found in the PFOS Appendix.
2.1.5 Evidence Synthesis and integration
For the purposes of this assessment, evidence synthesis and integration are considered distinct
but related processes. Evidence synthesis refers to the process of analyzing the results of the
available studies (including their strengths and weaknesses) for consistency and coherence, often
by evidence stream (e.g., human or animal) and health effect outcome. In evidence integration,
the evidence across streams is considered together and integrated to develop judgments (for each
health outcome) about whether the chemical in question poses a hazard to human health.
The evidence syntheses are summary discussions of the body of evidence for each evidence
stream (i.e., human and animal) for each health outcome analyzed. The available human and
animal health effects evidence were synthesized separately, with each synthesis resulting in a
summary discussion of the available evidence. For the animal toxicological evidence stream,
evidence synthesis included consideration of studies rated high and medium confidence. For the
epidemiological evidence stream, evidence synthesis was based primarily on studies of high and
medium confidence, including discussion of study quality considerations, according to the
recommendations of the SAB {U.S. EPA, 2022, 10476098}. Inferences drawn from studies
described in the 2016 PFOS HESD were considered when drawing health effects conclusions.
Epidemiological studies were excluded from the evidence synthesis narrative if they included
data that were reported in multiple studies (e.g., overlapping NHANES studies). Studies
reporting results from the same cohort and the same health outcome as another study were
considered overlapping evidence, and these additional studies were not discussed in the evidence
synthesis narrative to avoid duplication or overrepresentation of results from the same group of
participants. In cases of overlapping studies, the study with the largest number of participants
and/or the most accurate outcome measures was given preference. Consistent with the IRIS
Handbook {U.S. EPA, 2022, 10367891}, low confidence epidemiological studies and results
were used only in a supporting role and given less weight during evidence synthesis and
integration compared to high or medium confidence studies. Low confidence epidemiological
2-5
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
studies were included in evidence syntheses order to capture all of the available data for PFOS in
the weight of evidence analyses.
For evidence integration, integrated judgments, that took into account mechanistic considerations
for the five priority health outcomes (i.e., cancer, hepatic, immune, cardiovascular, and
developmental), were drawn for each health outcome across human and animal lines of evidence.
The evidence integration provides a summary of the causal interpretations between PFOS
exposure and health effects based on results of the available epidemiological and animal
toxicological studies, in addition to the available mechanistic evidence. Considerations when
evaluating the available studies included risk of bias, sensitivity, consistency, strength (effect
magnitude) and precision, biological gradient/dose-response, coherence, and mechanistic
evidence related to biological plausibility.
The evidence integration was conducted according to guidance outlined in the IRIS Handbook
and the Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA (anionic
and acidforms) IRIS Assessments {U.S. EPA, 2020, 8642427}. The evidence integration
included evidence stream evaluation in which the qualitative summaries on the strength of
evidence from studies in animals and humans were evaluated, and subsequent inference across
all evidence streams. Human relevance of animal models as well as mechanistic evidence to
inform mode of action were considered. Evidence integration produced an overall judgment
about whether sufficient or insufficient evidence of an association with PFOS exposure exists for
each human health outcome, as well as the rationale for each judgment. The potential evidence
integration judgments for characterizing human health effects are evidence demonstrates,
evidence indicates (likely), evidence suggests, evidence inadequate, and strong evidence
supports no effect.
Details about evidence synthesis and integration are summarized in the Appendix (See PFOS
Appendix).
2.2 Dose-Response Assessment
Evidence synthesis and integration enabled identification of the health outcomes with the
strongest weight of evidence supporting causal relationships between PFOS exposure and
adverse health effects, as well as the most sensitive cancer and noncancer endpoints. Studies
were evaluated for use in POD derivation on the basis of study design, study quality evaluation,
and data availability. For human evidence, all high or medium confidence studies were
considered; for animal evidence, only animal toxicological studies with at least two PFOS
exposure groups and of high or medium confidence were considered.
2.2.1 Approach to POD and RfD Derivation for Non-Cancer
Health Outcomes
The current, recommended EPA human health risk assessment approach described in EPA's A
Review of the Reference Dose and Reference Concentration Processes, which is a multistep
approach to dose-response assessment, includes analysis of dose and response within the range
of observation, followed by extrapolation to lower exposure levels {U.S. EPA, 2002, 88824}.
For non-cancer health outcomes, EPA performed dose-response assessments to define points of
departure (PODs) and extrapolated from the PODs to RfDs.
2-6
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
For PFOS, EPA performed benchmark dose (BMD) modeling of all animal toxicological studies
considered for dose-response to refine the POD in deriving the RfD. The BMD modeling
approach involves dose-response modeling to obtain BMDs (i.e., dose levels corresponding to
specific response levels near the low end of the observable range of the data) and identifies the
lower limits of the BMDs (BMDLs) to serve as potential PODs for deriving quantitative
estimates below the range of observation {U.S. EPA, 2012, 1239433}. EPA used the publicly
available Benchmark Dose Software (BMDS) program developed and maintained by EPA
(https://www.epa.eov/bmds). BMDS fits mathematical models to the data and determines the
dose (benchmark dose or BMD) that corresponds to a pre-determined level of response
(benchmark response or BMR). For dichotomous data, the BMR is typically set at either 5 or
10% above the background or the response of the control group. For continuous data, a BMR of
one half or one standard deviation from the control mean is typically used when there are no
outcome-specific data to indicate what level of response is biologically significant {U.S. EPA,
2012, 1239433}. For dose-response data for which BMD modeling did not produce an adequate
model fit, a no-observed-adverse-effect level (NOAEL) or lowest-observed-adverse-effect level
(LOAEL) was used as the POD.
For the epidemiological studies considered for dose-response assessment, EPA used multiple
modeling approaches to determine PODs, depending upon the health outcome and the data
provided in the studies. For the developmental, hepatic, and serum lipid dose-response studies,
EPA used a hybrid modeling approach that involves estimating the incidence of individuals
above or below a level considered to be adverse and determining the probability of responses at
specified exposure levels above the control {U.S. EPA, 2012, 1239433} for cases in which EPA
was able to define a level considered clinically adverse for these outcomes (see PFOS Appendix
for details). EPA also performed BMD modeling and provided study LOAELs/NOAELs for the
hepatic and serum lipid dose-response studies as sensitivity analyses of the hybrid approach. For
the immune studies, where a clinically defined adverse level is not well defined, EPA used
multivariate models provided in the studies and determined a BMR according to EPA guidance
to calculate BMDs and BMDLs {U.S. EPA, 2012, 1239433}.
See the PFOS Appendix for additional details on the study-specific modeling.
The general steps for deriving an RfD for PFOS are summarized below.
Step 1: Evaluate the data to identify and characterize endpoints affected by exposure to
PFOS. This step involves selecting the relevant studies and adverse effects to be considered for
BMD modeling. Once the appropriate data are collected, evaluated for study quality, and
characterized for adverse health outcomes, the risk assessor selects health endpoints/outcomes
judged to be relevant to human health and among the most sensitive, defined as effects observed
in the lower exposure range. Considerations that might influence selection of endpoints include
whether data have dose-response information, percent change from controls, adversity of effect,
and consistency across studies.
Step la (for dose-response data from a study in an animal model): Convert administered
dose to an internal dose. A pharmacokinetic model is used to predict the internal dose (in the
animals used in the toxicity studies or in humans) that would correspond to the administered
dose used in the study (see 4.1.3 for additional detail). A number of dose-metrics across life
2-7
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
stages are selected for simulation in a mouse, rat, monkey, or human. Concentrations of PFOS in
blood are considered for all the internal dose-metrics.
Step 2: Conduct dose-response modeling. See above and the PFOS Appendix for study-
specific details.
Step 3: Convert the POD to a human equivalent dose (HED) or point of departure human
equivalent dose (PODhed). The POD (a BMDL, NOAEL, or LOAEL) is converted to an HED
following the method described in Section 4.1.3. Briefly, a pharmacokinetic model for human
dosimetry is used to simulate the HED from the animal PODs from Step 2. Pharmacokinetic
modeling is also used to simulate selected epidemiological studies to obtain a chronic dose that
would result in the internal POD obtained from dose-response modeling (see Section 4). Based
on the available data, a serum PFOS concentration was identified as a suitable internal dosimetry
target for the human and animal endpoints of interest.
Step 4: Select appropriate uncertainty factors (UFs) and provide rationale for UF selection.
UFs are applied in accordance with EPA guidelines considering variations in sensitivity among
humans, differences between animals and humans (if applicable), the duration of exposure in the
critical study compared to the lifetime of the species studied, and the completeness of the
epidemiological or animal toxicological database.
Step 5: Calculate the chronic RfD. The RfD is calculated by dividing PODhed by the
composite (total) UF.
PODhed = calculated from the BMDL, NOAEL, or LOAEL using the human pharmacokinetic
(PK) model presented in Section 4.1.3.2.
UFc = Composite (total) UF calculated by multiplying the selected individual UFs for variations
in sensitivity among humans, differences between animals and humans, duration of exposure in
the critical study compared to the lifetime of the species studied, and completeness of the
toxicology database, in accordance with EPA guidelines {U.S. EPA, 2002, 88824}.
In accordance with EPA's 2005 Guidelines for Carcinogen Risk Assessment, a descriptive
weight of evidence expert judgment is made, based on all available animal, human, and
mechanistic data, as to the likelihood that a contaminant is a human carcinogen and the
conditions under which the carcinogenic effects may be expressed {U.S. EPA, 2005, 9638795}.
A narrative is developed to provide a complete description of the weight of evidence and
conditions of carcinogenicity. The potential carcinogenicity descriptors (presented in the 2005
guidelines) are:
• Carcinogenic to humans
where:
2.2.2 Cancer Assessment
2.2.2.1 Approach for Cancer Classification
2-8
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
• Likely to be carcinogenic to humans
• Suggestive evidence of carcinogenic potential
• Inadequate information to assess carcinogenic potential
• Not likely to be carcinogenic to humans
More than one carcinogenicity descriptor can be applied if a chemical's effects differ by dose,
exposure route, or mode of action (MOA)4. For example, a chemical may be carcinogenic to
humans above but not below a specific dose level if a key event in tumor formation does not
occur below that dose. MOA information informs both the qualitative and quantitative aspects of
the assessment, including the human relevance of tumors observed in animals. MOA must be
considered separately for each target organ.
2.2.2.2 Derivation of a Cancer Slope Factor
EPA's 2005 Guidelines for Carcinogen Risk Assessment recommends a two-step process for the
quantitation of cancer risk. First, a model is used to fit a dose-response curve to the data, based
on the doses and associated tumors observed. For animal toxicological studies, EPA used the
publicly available Benchmark Dose Software (BMDS) program developed and maintained by
EPA (https://www.epa.eov/bmds). For cancer data, BMDS fits multistage models and the model
is used to identify a POD for extrapolation to the low-dose region based on the BMD associated
with a significant increase in tumor incidence above the control. According to the 2005
guidelines, the POD is the lowest dose that is adequately supported by the data. The BMDio (the
dose corresponding to a 10% increase in tumors) and the BMDLio (the 95% lower confidence
limit on that dose) are also reported and are often used as the POD.
In the second step of quantitation, the POD is extrapolated to the low-dose region of interest for
environmental exposures. The approach for extrapolation depends on the MOA for
carcinogenesis (i.e., linear or nonlinear). When coverage indicates that a chemical causes cancer
through a mutagenic MOA (i.e., mutation of deoxyribonucleic acid (DNA)) or the MOA for
carcinogenicity is not known, this extrapolation is performed by drawing a line (on a graph of
dose vs. response) from the POD to the origin (zero dose, zero tumors). The slope of the line
(Aresponse/Adose) gives rise to the CSF, which can be interpreted as the risk per mg/kg/day. In
addition, according to EPA's Supplemental Guidance for Assessing Susceptibility from Early-
Life Exposure to Carcinogens {U.S. EPA, 2005, 88823}, affirmative determination of a
mutagenic MOA (as opposed to defaulting to a mutagenic MOA based on insufficient data or
limited data indicating potential mutagenicity) determines whether age-dependent adjustment
factors are applied in the quantification of risk to account for additional sensitivity of children.
In cases for which a chemical is shown to cause cancer via an MOA that is not linear at low
doses, and the chemical does not demonstrate mutagenic or other activity consistent with
linearity at low doses, a nonlinear extrapolation is conducted. EPA's 2005 Guidelines for
Carcinogen Risk Assessment state that "where tumors arise through a nonlinear MOA, an oral
RfD or inhalation reference concentration, or both, should be developed in accordance with
EPA's established practice of developing such values, taking into consideration the factors
4MOA is defined as a sequence of key events and processes, starting with interaction of an agent with a cell, proceeding through
operational and anatomical changes, and resulting in cancer formation. It is contrasted with "mechanism of action," which
implies a more detailed understanding and description of events.
2-9
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
summarized in the characterization of the POD." In these cases, an RfD-like value is calculated
based on the key event5 for carcinogenesis or the tumor response.
Once a POD is determined, a PK model is used to calculate the HED for animal oral exposures
(PODhed). The CSF is then calculated by dividing the selected BMR by the PODhed.
See the PFOS Appendix for additional details on the study-specific modeling.
2.3 MCLG Derivation
As provided in SDWA Section 1412(b)(4)(A), EPA establishes the MCLG at the level at which
no known or anticipated adverse effects on the health of persons occur and which allows an
adequate margin of safety. EPA assesses available science examining cancer and noncancer
health effects associated with oral exposure to the contaminant. Consistent with the statutory
definition of MCLG, EPA establishes MCLGs of zero for carcinogens classified as Carcinogenic
to Humans or Likely to be Carcinogenic to Humans6 for which there is insufficient information
to determine that a carcinogen has a threshold below which there are no carcinogenic effects
{U.S. EPA, 1998, 10442462; U.S. EPA, 2000, 10442463; U.S. EPA, 2001, 10442464}.
For nonlinear carcinogenic contaminants, contaminants that are suggestive carcinogens, and non-
carcinogenic contaminants, EPA establishes the MCLG based on a toxicity value, typically an
RfD, but a similar toxicity value (e.g., ATSDR Minimal Risk Level) may also be used when it
represents the best available science. A noncancer MCLG is designed to be protective of
noncancer effects over a lifetime of exposure with an adequate margin of safety, including for
sensitive populations and life stages consistent with SDWA 1412(b)(3)(C)(i)(V) and
1412(b)(4)(A). The calculation of a noncancer MCLG includes an oral toxicity reference value
5The key event is defined as an empirically observed precursor step that is itself a necessary element of the MOA or is a
biologically based marker for such an element.
6The MCLG is derived depending on the available noncancer and cancer evidence for a particular chemical. Establishing the
MCLG for a chemical has typically been accomplished in one of three ways depending upon a three-category classification
approach {U.S. EPA, 1985, 9207; U.S. EPA, 1991, 5499}. The categories are based on the available evidence of carcinogenicity
after exposure via ingestion. The starting point in categorizing a chemical is through assigning a cancer descriptor using EPA's
current Guidelines for Carcinogen Risk Assessment {U.S. EPA, 2005, 6324329}. The descriptors in the 2005 Guidelines
replaced the prior alphanumeric groupings, although the basis for the classifications is similar. In prior rulemakings, the agency
typically placed Group A, Bl, and B2 contaminants into Category I, Group C into Category II, and Group D and E into Category
III based on the agency's previous cancer classification guidelines (i.e., Guidelines for Carcinogen Risk Assessment, published in
51 FR 33992, September24, 1986 {U.S. EPA, 1986, 199530} and the 1999 interim final guidelines {U.S. EPA, 1999,41631;
U.S. EPA, 2001,10442464}):
• Category I chemicals have "strong evidence [of carcinogenicity] considering weight of evidence, pharmacokinetics,
and exposure {U.S. EPA, 1985, 9207; U.S. EPA, 1991, 5499}." EPA's 2005 cancer descriptors associated with this
category are: "Carcinogenic to Humans" or "Likely to be Carcinogenic to Humans" {U.S. EPA, 2005, 6324329}.
EPA's policy under SDWA is to set MCLGs for Category I chemicals at zero, based on the principle that any exposure
to known or likely human carcinogens might represent some finite level of risk. In cases when there is sufficient
evidence to determine a nonlinear cancer mode of action, the MCLG is based on the RfD approach described below.
• Category II chemicals have "limited evidence [of carcinogenicity] considering weight of evidence, pharmacokinetics,
and exposure {U.S. EPA, 1985, 9207; U.S. EPA, 1991, 5499}." EPA's 2005 cancer descriptor associated with this
category is: "Suggestive Evidence of Carcinogenic Potential" {U.S. EPA, 2005, 6324329}. The MCLG for Category II
contaminants is based on noncancer effects {U.S. EPA, 1985, 9207; U.S. EPA, 1991,5499}.
• Category III chemicals have "inadequate or no animal evidence [of carcinogenicity] {U.S. EPA, 1985, 9207; U.S. EPA,
1991, 5499}." EPA's 2005 cancer descriptors associated with this category are: "Inadequate Information to Assess
Carcinogenic Potential" and "Not Likely to Be Carcinogenic to Humans" {U.S. EPA, 2005, 6324329}. The MCLG for
Category III contaminants is based on noncancer effects.
2-10
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
such as an RfD, body weight-based drinking water intake (DWI-BW), and RSC as presented in
the equation below:
RfD = chronic reference dose—an estimate (with uncertainty spanning perhaps an order of
magnitude) of a daily oral exposure of the human population to a substance that is likely to be
without an appreciable risk of deleterious effects during a lifetime. The RfD is equal to a
PODhed divided by a composite uncertainty factor.
DWI-BW = An exposure factor in the form of the 90th percentile body weight-adjusted drinking
water intake value for the identified population or life stage, in units of liters of water consumed
per kilogram body weight per day (L/kg bw-day). The DWI-BW considers both direct and
indirect consumption of drinking water (indirect water consumption encompasses water added in
the preparation of foods or beverages, such as tea or coffee). Chapter 3 of EPA's Exposure
Factors Handbook {U.S. EPA, 2019, 7267482} provides DWI-BWs for various populations or
life stages within the general population for which there are publicly available, peer-reviewed
data such as NHANES data.
RSC = relative source contribution—the percentage of the total exposure attributed to drinking
water sources {U.S. EPA, 2000, 19428}, with the remainder of the exposure allocated to all
other routes or sources. The purpose of the RSC is to ensure that the level of a contaminant (e.g.,
MCLG value), when combined with other identified sources of exposure common to the
population of concern, will not result in exposures that exceed the RfD. The RSC is derived by
applying the Exposure Decision Tree approach published in EPA's Methodology for Deriving
Ambient Water Quality Criteria for the Protection of Human Health {U.S. EPA, 2000, 19428}.
Further description of the RSC for PFOS can be found in the Appendix (see PFOS Appendix).
Where:
2-11
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3 Results of the Health Effects Systematic Review
and Toxicokinetics Methods
3.1 Literature Search and Screening Results
Studies referenced in this assessment are cited as "Author Last Name, Publication Year, HERO
ID" and are available in EPA HERO: A Database of Scientific Studies and References. The
HERO ID is a unique identifier for studies available in HERO. Additional study metadata are
publicly available and can be obtained by searching for the HERO ID on the public-facing
webpage available here: https://hero.epa.eov/.
The three database searches yielded 6,007 unique records prior to running SWIFT Review. Table
3-1 shows the results from database searches conducted in April 2019, September 2020, and
February 2022.
Table 3-1. Database Literature Search Results
Database
Date Run: Results
WoS
4/10/2019: 3,081 results
9/3/2020: 1,286 results
2/2/2022: 1,021 results
PubMed
4/10/2019: 2,191 results
9/3/2020: 811 results
2/2/2022: 1,728 results
TOXLINE
4/10/2019: 60 results
TSCATS
4/11/2019: 0 results
Total number of references from all
databases for all searches3
4/2019: 3,382 results
9/2020: 1,153 results
2/2022: 1,858 results
Total number of references after
running SWIFT Review3
4/2019: 1,977 results
9/2020: 867 results
2/2022: 1,370 results
Total number of unique studies moved to
screening b
3,921
Notes:
a The number of studies includes duplicate references across search dates due to overlap between search years.
b Duplicates across search dates removed.
The additional sources of literature outlined in Section 2.1.1 (i.e., assessments published by other
agencies, studies identified during mechanistic or toxicokinetic syntheses, and studies identified
by the SAB) yielded 200 unique records.
The 3,921 studies captured with the SWIFT Review evidence streams filters and the 200 records
identified from additional sources yield a total of 4,121 unique studies. These 4,121 studies were
moved to the next stage of screening (title and abstract screening using either DistillerSR or
SWIFT ActiveScreener). Of the 4,121 unique studies, 918 moved on to full-text level review,
1,589 were excluded during title and abstract screening, and 1,614 were tagged as containing
potentially relevant supplemental material. Of the 918 screened at the full-text level, 599 were
considered to meet PECO eligibility criteria (See PFOS Appendix) and included relevant
information on PFOS. The 599 studies that were determined to meet PECO criteria after full-text
3-1
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
level screening included 423 epidemiological (human) studies, 45 animal toxicological studies, 6
PBPK studies, and 125 studies that were not extracted (e.g., low confidence studies, meta-
analyses, studies that did not evaluate effects on one of the priority health outcomes). An
additional 16 PBPK studies were identified during the toxicokinetic screening for a total of 22
PBPK studies. Details of the literature search and screening process are shown in Figure 3-1.
The 423 epidemiological studies and 45 animal toxicological studies underwent study quality
evaluation and were subsequently considered for data extraction as outlined in Sections 2.1.3 and
0 (see PFOS Appendix for more details). The results of the health outcome-specific study quality
evaluations and data extractions are described in Sections 3.4 and 3.5.
Additionally, the 22 studies tagged as containing relevant PBPK models were reviewed by PK
subject matter experts for inclusion consideration. The included studies are summarized in
Section 3.3.2 and parameters described in these studies were considered for incorporation into
the animal and human PK models, which are summarized in Section 4.1.3.
Finally, the 89 toxicokinetic and 301 mechanistic studies identified as relevant for PFOS moved
on to a limited data extraction as described in the Appendix (see PFOS Appendix). The
toxicokinetic studies pertaining to ADME are synthesized in Section 3.3.1. The mechanistic
studies relevant to the 5 prioritized health outcomes are synthesized in Sections 3.4 and 3.5 and
were considered as part of the evidence integration.
3-2
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
References identified through
database search
3,382
1.153
1,858
2013 to 2019:
2019 to 2020:
2020 to 2022:
References after
SWIFT filters applied
2013 to 2019:
2019 to 2020
2020 to 2022
1,977
867
1,370
References screened for
PFOAand PFOSa
4,121
References assessed for eligibility
for PFOA and PFOS
Relevant references for PFOS
References identified from
2016 PFOA & PFOS HESDs
References identified through
oilier sources
142
References excluded 1,589
Not PECO Relevant 1.117
Excluded by SWIFT-Aclive 472
Supplemental Tag
Other Supplemental
Mechanistic
Toxicokinetic
References excluded
Dedu plication
Not PECO Relevant
Supplemental Tag
Other Supplemental
Mechanistic
Toxicokinetic
1,614
1,309
525
232
61
14
163
132
70
19
T oxicokinetici'Meehanistic
; 829
references assessed for eligibility
Mechanistic 652
Toxicokinetic 280
Did Not Extract Human Animal
125 423 | 45
Figure 3-1. Summary of Literature Search and Screening Process for PFOS
Interactive figure and additional study details available on Tableau.
Interactive figure based on work by Magnuson et al. (2022, 10442900).
"Other sources" include assessments published by other agencies, studies identified during mechanistic or toxicokinetic
syntheses, and studies identified by the SAB.
a Includes number of unique references after deduplication of studies captured with the SWIFT Review evidence streams filters
and records identified from additional sources.
b Includes number of unique references considered to meet PECO eligibility criteria at the full text level and include relevant
information on PFOS.
c Includes number of unique references identified during title/abstract screening, full text screening, and data extraction assessed
for toxicokinetic and/or mechanistic eligibility.
d Only includes studies with relevant information on PFOS.
e Includes 6 PBPK studies determined to meet PECO criteria plus an additional 16 PBPK studies identified during the
toxicokinetic screening.
3-3
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.1.1 Results for Epidemiology Studies of PFOS by Health
Outcome
Of the 423 epidemiological studies that met the inclusion criteria, 179 studies had a cohort study
design, 167 had a cross-sectional design, 40 had a case-control design, and 37 had other study
designs (e.g., nested case-control). Epidemiological studies were categorized into 18 health
outcomes. Most studies reported on the developmental (n = 90), cardiovascular (n = 85),
metabolic (n = 74), or immune systems (n = 63). Studies that reported outcomes spanning
multiple health outcomes were not counted more than once in the grand totals shown in Figure
3-2.
Study Design
Health System
Case-control
Cohort
Cross-sectional
Other
Grand Total
Cancer
5
3
3
5
16
Cardiovascular
S
18
56
6
85
Dermal
0
1
0
0
1
Developmental
6
58
19
7
90
Endocrine
1
e
20
7
36
Gastrointestinal
1
4
0
0
5
Hematologic
0
0
8
0
8
Hepatic
1
3
18
2
24
Immune
6
31
17
9
63
Metabolic
7
32
31
4
74
Musculoskeletal
0
0
6
2
8
Nervous
3
26
5
3
37
Ocular
0
0
1
0
1
Renal
1
3
16
0
20
Reproductive, Male
0
7
15
2
24
Reproductive, Female
10
23
19
3
55
Respiratory
1
3
1
0
5
Other
0
2
3
0
5
Grand Total
40
179
167
37
423
Figure 3-2. Summary of Epidemiology Studies of PFOS Exposure by Health System and
Study Design3
Interactive figure and additional study details available on Tableau.
a A study can report on more than one health system. Column grand totals represent the number of unique studies and are not a
sum of health system tags.
3.1.2 Results for Animal Toxicological Studies of PFOS by
Health Outcome
Of the 45 animal toxicological studies that met the inclusion criteria, most studies had either
short-term (n = 19) or developmental (n = 15) study designs and most were conducted in rats
(n = 23). The rat studies had short-term (n = 12), developmental (n = 7), chronic (n = 2),
reproductive (n = 2), and subchronic (n = 1) study designs. The remaining studies reported
results for mice (n = 21) using developmental (n = 8), short-term (n = 7), subchronic (n = 5), or
3-4
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
reproductive (n = 1) study designs, monkeys (n = 1) using a chronic study design, or rabbits
(n = 1) using a developmental study design. Animal toxicological studies were categorized into
13 health outcomes. Most studies reported results for the whole body (n = 25; i.e., systemic
endpoints such as body weight), hepatic (n = 20), reproductive (n = 19), or developmental
(n = 16) systems. Studies that reported outcomes spanning multiple health outcomes, study
designs, or species were not counted more than once in the grand totals shown in Figure 3-3.
Study Design & Species
Health System
Short-term
Mouse Rat
Subchronic
Mouse Rat
Chronic
Monkey
Rat
Mouse
Developmental
Rabbit
Rat
Reproductive
Mouse
Rat
Grand Total
Cancer
0
0
0
0
0
1
0
0
0
0
0
1
Cardiovascular
1
2
2
0
1
1
0
0
2
0 1
10
Developmental
0
0
0
0
0
0
mm
1
0 2
16
Endocrine
1
4
1
0
1 1
0
3
0 1
13
Hematologic
1
3
0
0
1 0
0
0
0
0 0
5
Hepatic
2
mm
4
0
1
2
4
0
2
0 1
20
Immune
2
4
3
0
1
2
1
0
0
0 0
13
Metabolic
0
3
1
0
0
2
0
0
1
0
1
7
Nervous
2
KM
1
0
0 1 1
0
2
0
1
14
Renal
0
3
3
0
1
2
2
0
0
0
0
10
Reproductive
2
3
1
1
1
0
4
H 3 1
1 2
19
Respiratory
1
1
1
0
0
0
0
0
0
0
0
3
Whole Body
4
4
1
1
2
2
1
2
0
2 1
25
Grand Total
7
12
5
1
1
2
8
1
7
1 2
45
Figure 3-3. Summary of Animal Toxicological Studies of PFOS Exposure by Health
System, Study Design, and Speciesa'b
Interactive figure and additional study details available on Tableau.
a A study can report on more than one study design and species. Row grand totals represent the number of unique studies and are
not a sum of study design and species tags.
b A study can report on more than one health system. Column grand totals represent the number of unique studies and are not a
sum of health system tags.
3.2 Data Extraction Results
Data extracted from the 423 epidemiological studies are available via Tableau Public and data
extracted from the 45 animal toxicological studies are available in the public HAWC site,
displayed as exposure-response arrays, forest plots, and trees. See Sections 3.4 and 3.5 for health
outcome-specific data extracted for synthesis development. Additionally, the limited data
extractions from the ADME and mechanistic studies can be found via Tableau Public here and
here, respectively.
3.3 Toxicokinetic Synthesis
As described in Section 3.1, EPA identified 89 and 22 studies containing information relevant to
the toxicokinetics and PBPK modeling of PFOS, respectively. The results of these studies are
described in the subsections below and additional information related to toxicokinetic
characteristics of PFOS can be found in Appendix B.
3.3.1 ADME
PFOS is resistant to metabolic and environmental degradation due to its strong carbon-fluorine
bonds. It is not readily eliminated and can have a long half-life in humans and animals. However,
the toxicokinetic profile and the underlying mechanism for the chemical's long half-life are not
completely understood. For PFOS, membrane transporter families appear to play an important
3-5
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
role in ADME, including organic anion transporters (OATs), organic anion transporting
polypeptides (OATPs), multidrug resistance-associated proteins (MRPs), and urate transporters.
Transporters play a critical role in GI tract absorption, uptake by tissues, and excretion via bile
and the kidney. Limited data are available regarding the transporters for PFOS; however, the
toxicokinetic properties of PFOS suggest tissue uptake and renal resorption through facilitated
uptake. Some inhibition studies suggest that PFOS transport could involve the same transporters
as for PFOA, since PFOS and PFOA have similar chain lengths, renal excretion properties, and
liver accumulation.
Animal studies indicate that PFOS is well-absorbed orally and distributes to many tissues and
organs. High levels of PFOS are consistently observed in blood and liver. While PFOS can form
as a degradation product or metabolite from other per- or polyfluoroalkyl substances, PFOS itself
does not undergo further metabolism after absorption takes place. PFAS are known to activate
peroxisome proliferator activated receptor (PPAR) pathways by increasing transcription of genes
related to mitochondrial and peroxisomal lipid metabolism, as well as sterol and bile acid
biosynthesis. Based on transcriptional activation of many genes in PPARa-null mice, however,
other gene products likely modify toxicokinetics of PFOS {Andersen, 2008, 3749214}.
3.3.1.1 Absorption
Absorption data are available in laboratory animals for oral {Chang, 2012, 1289832} and
inhalation {Rusch, 1979, 7561179} exposures, and extensive data are available demonstrating
the presence of PFOS in human serum. Limited in vitro absorption data are available (see PFOS
Appendix).
Since PFOS is moderately soluble in aqueous solutions and oleophobic (i.e., minimally soluble
in body lipids), movement across interface membranes was thought to be dominated by
transporters or mechanisms other than simple diffusion across the lipid bilayer. Recent
mechanistic studies, however, support transporter-independent uptake through passive diffusion
processes. Ebert et al. (2020, 6505873) determined membrane/water partition coefficients
(Kmem/w) for PFOS and examined passible permeation into cells by measuring the passive anionic
permeability (Pion) through planar lipid bilayers. In this system, the partition coefficients were
considered high enough to explain observed cellular uptake by passive diffusion in the absence
of active uptake processes.
Uptake by cells may be influenced by interactions with lipids and serum proteins. PFOS
exhibited higher levels of binding to lipids and phospholipids relative to PFOA, which correlated
with uptake into lung epithelial cells {Sanchez Garcia, 2018, 4234856}. Phospholipophilicity
correlated to cellular accumulation better than other lipophilicity measures. The extent to which
PFOS phospholipophilicity influences absorption through the GI tract, lungs, or skin is unknown.
While there are no studies available that quantify absorption in humans, extensive data on serum
PFOS confirm uptake from the environment but do not establish an exposure route. Studies that
provide the basis for human half-life estimates rely on changes in PFOS serum levels over time.
Bioavailability of PFOS after oral exposure is very high in rats. Serum PFOS concentrations
after oral dosing were > 100% of levels measured after intravenous (IV) dosing, which may
reflect enterohepatic absorption that occurs after gavage but not IV administration {Kim, 2016,
3749289; Huang, 2019, 5387170}.
3-6
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.3.1.2 Distribution
3.3.1.2.1PFOS Binding to Blood Fractions and Serum Proteins
Detailed study descriptions of literature regarding the distribution of PFOS in humans and
animals are provided in the Appendix (see PFOS Appendix). Distribution of absorbed material
requires vascular transport from the portal of entry to receiving tissues. Distribution of PFAS to
plasma has been reported to be chain length-dependent {Jin, 2016, 3859825}. Increasing chain
length (from C6 to CI 1) correlated with an increased mass fraction in human plasma. Among
different kinds of human blood samples, PFOS accumulates to highest levels in plasma, followed
by whole blood and serum {Forsthuber, 2020, 6311640; Jin, 2016, 3859825; Poothong, 2017,
4239163}. Poothong et al. (2017, 4239163) found that median PFOS concentrations in plasma,
serum, and whole blood were 5.24, 4.77, and 2.85 ng/mL, respectively. These findings suggest
that the common practice of multiplying by a factor of 2 to convert the concentrations in whole
blood to serum {Ehresman, 2007, 1429928} will not provide accurate estimates for PFOS.
PFOS is distributed within the body by noncovalently binding to plasma proteins. Many studies
have investigated PFOS interactions with human serum albumin (HSA) {Zhang, 2009, 2919350;
Salvalaglio, 2010, 2919252; Chen, 2009, 1280480; D'Alessandro, 2013, 5084740; Liu, 2017,
3856708}. In vitro analyses found that plasma proteins can bind PFOS in plasma from humans,
cynomolgus monkeys, and rats {Kerstner-Wood, 2003, 4771364}. PFOS was highly bound
(99.8%) to albumin and showed affinity for low-density lipoproteins (95.6%) with some binding
to alpha-globulins (59.4%) and gamma-globulins (24.1%). HSA-PFOS intermolecular
interactions are mediated through van der Waals forces and hydrogen bonds {Zhang, 2009,
2919350; Chen, 2009, 1280480}. Beesoon and Martin (2015, 2850292) determined that linear
PFOS bound more strongly to calf serum albumin than the branched chain isomers in the order
of 3m < 4m < lm < 5m < 6m (iso) < linear. PFOS binding to HSA results in alterations in the
albumin secondary structure and can diminish esterase activity {Liu, 2017, 3856708}, though the
extent to which this affects the physiological functions of albumin is unknown. PFOS-mediated
conformational changes may also interfere with albumin's ability to transport its natural ligands
and pharmaceuticals, including vitamin B2 (riboflavin) and ibuprofen {D'Alessandro, 2013,
5084740}, and may interfere with PFOS uptake into cells {Sheng, 2020, 6565171}.
Binding to albumin and other serum proteins may affect transfer of PFOS from maternal blood to
the fetus {Gao, 2019, 5387135}. Since there is effectively a competition between PFOS binding
in maternal serum vs. cord blood, lower cord blood albumin levels compared to maternal blood
albumin levels are likely to reduce transfer from maternal serum across the placenta. Consistent
with this hypothesis, Pan et al. (2017, 3981900) found that a high concentration of cord serum
albumin was associated with higher PFOS transfer efficiencies, whereas high maternal serum
albumin concentration was associated with reduced transfer efficiency.
3.3.1.2.2PFOS Binding to Intracellular Proteins and Transporters
Within cells, PFOS has been shown to bind to liver fatty acid binding protein (L-FABP)
{Luebker, 2002, 1291067; Zhang, 2013, 5081488; Yang, 2020, 6356370}. L-FABP is an
intracellular lipid carrier protein that reversibly binds long-chain fatty acids, phospholipids, and
an assortment of peroxisome proliferators {Erol, 2004, 5212239} and constitutes 2—5% of the
cytosolic protein in hepatocytes.
3-7
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
PFOS entry from serum into tissues appears to be controlled by several families of membrane
transporters based on extrapolation from PFOA studies and several PFOS-specific studies. Yu et
al. (2011, 1294541) observed that PFOS exposure in rats increased hepatic OATP2 and MRP2
messenger ribonucleic acid (mRNA) expression. Transporters responsible for PFOS transport
across the placenta are not well understood, though preliminary studies examining transporter
expression identified OAT4 as a candidate receptor {Kummu, 2015, 3789332}. Thus far, no
functional studies demonstrating a role for these transporters in PFOS uptake in liver or placenta
have been identified.
3.3.1.2.3 Tissue Distribution in Humans and Animals
Evidence from human autopsy and surgical tissues demonstrates that PFOS distributes to a wide
range of tissues, organs, and matrices throughout the body. It should be noted, however, that
autopsy and surgical tissues may not accurately reflect PFAS tissue distribution in the living
body {Cao, 2021, 9959613}. Blood and liver are major sites of PFOS accumulation {Olsen,
2001, 9641811}. Two studies measured PFOS levels in cerebrospinal fluid and serum {Harada,
2007, 2919450; Wang, 2018, 5080654} and in both studies, PFOS levels in cerebrospinal fluid
were two orders of magnitude lower than in serum, suggesting that PFOS does not easily cross
the adult human blood-brain barrier.
In a study of autopsy tissues collected within 24 hours of death, Perez et al. (2013, 2325349) and
found PFOS in the liver (104 ng/g), kidney (75.6 ng/g), lung (29.1 ng/g), and brain (4.9 ng/g),
with levels below the limit of detection (LOD) in bone. PFOS also accumulates in follicular fluid
{Kang, 2020, 6356899}, raising the possibility of reproductive toxicity in humans.
Studies of tissue distribution are available for several species of animals including non-human
primates, rats, and mice. Studies of non-human primates indicate PFOS accumulates in serum in
a dose-dependent manner {Seacat, 2002, 757853; Chang, 2017, 3981378}. Limited data on liver
accumulation of PFOS in monkeys show that PFOS levels in liver were similar or slightly lower
than serum levels.
Several rodent studies identified high levels of PFOS in blood and liver across a range of dosing
regimens and study durations. Whereas monkeys had nearly a 1:1 liver to serum ratio, rodent
models were observed to accumulate far more PFOS in liver than serum {NTP, 2019, 5400978}.
Plasma PFOS concentrations were generally similar in males and females. For example, in a 28-
day toxicity study, dose-normalized plasma concentrations ([iM/mmol/kg/day) in males and
females were within 1.5-fold across the dose groups {NTP, 2019, 5400978}. Additional studies
in rats and mice documented PFOS distribution to a wide range of tissues including kidney,
heart, lungs, and spleen. Interestingly, in rodents, PFOS has been measured in moderate
quantities in the brain and testicles, indicating that PFOS does cross the blood-brain and blood-
testis barriers in rats {Qui, 2013, 2850956} and mice {Bogdanska, 2011, 2919253; Cui, 2009,
757868}.
3.3.1.2.4 Distribution During Reproduction and Development
Several studies in humans, rats, and mice quantified distribution of PFOS from pregnant females
to placenta, cord blood, and amniotic fluid, which demonstrate pathways of distribution to and
elimination from fetuses. Accumulation of PFOS in fetal tissues was found to vary by gestational
age. New studies also confirm that distribution of PFOS from nursing mothers to their infants via
3-8
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
breastmilk correlates with duration of breastfeeding. Distribution is influenced by the chemical
properties of PFAS including length, lipophilicity, and branching.
The ratio of PFOS in placenta relative to maternal serum (Rpm) ranged from 0.048 to 0.749
{Zhang, 2013, 3859792; Chen, 2017, 3859806}. Zhang etal. (2015, 2851103) observed
differential accumulation of PFOS based on branching characteristics. Specifically, Rpms of
branched PFOS isomers increased with distance of branching points away from the sulfonate
group in the order of iso-PFOS < 4m-PFOS < 3+5m-PFOS < lm-PFOS. Mamsen et al. (2019,
5080595) demonstrated that gestational age can affect PFOS concentrations in maternal serum
and placentas, estimating a placenta PFOS accumulation rate of 0.13% per day during gestation.
Several studies reported a strong positive correlation between maternal and cord serum levels of
PFOS {Kato, 2014, 2851230; Porpora, 2013, 2150057}. The ratio of PFOS in cord serum
relative to maternal serum ranged from 0.22 to 0.98 (see PFOS Appendix) and generally
increased with gestational age {Li, 2020, 6505874}. Li et al. (2020, 6505874) also showed a 6%
increase in branched PFOS accumulation compared to linear PFOS isomers. Zhao et al. (2017,
3856461) observed higher transplacental transfer efficiencies (TTEs) for lm-, 4m-, 3+5m-, and
m2-PFOS compared to n-PFOS. Together, these findings indicate that branched isomers of
PFOS transfer more efficiently from maternal blood to cord blood compared to linear isomers. In
addition to PFOS branching, maternal factors including exposure sources, parity, and other
maternal demographics are postulated to influence observed variations in cord:maternal serum
ratios {Eryasa, 2019, 5412430; Jusko, 2016, 3981718; Brochot, 2019, 5381552}.
Lower PFOS concentrations were measured in amniotic fluid compared to placenta and cord
blood {Zhang, 2013, 3859792}. The mean concentration ratio between amniotic fluid and
maternal blood (AF:MB) was lower for PFOS (0.0014) than for PFOA (0.13). The mean
concentration ratio between amniotic fluid and cord blood (AF:CB) was lower for PFOS
(0.0065) than for PFOA (0.023). Authors attributed the differences in ratios between the two
compartments to the solubilities of PFOS and PFOA and their respective protein binding
capacities in the two matrices.
PFOS also distributes widely in fetal tissues. Mamsen et al. (2017, 3858487) measured the
concentrations of five PFAS in fetuses, placentas, and maternal plasma from a cohort of 39
pregnant women in Denmark. The concentration of PFOS decreased from maternal serum to
fetal tissues as follows: maternal serum > placenta > fetal tissues. In a second study, PFAS levels
were measured in embryos and fetuses at gestational weeks 7-42 and in serum from their
matched maternal pairs {Mamsen, 2019, 5080595}. PFOS accumulated at higher levels in fetal
tissues compared to other PFAS chemicals examined in fetal tissues and across trimesters. The
concentration of PFAS in fetal tissues fluctuated across trimesters and did not follow any
particular trend. For example, PFOS concentration in the liver was higher in the second trimester
compared to the third trimester, and lowest in the lung in the second trimester compared to the
first and third trimesters.
New studies also confirm that distribution of PFOS from nursing mothers to their infants via
breastmilk correlates with duration of breastfeeding {Mondal, 2014, 2850916; Cariou, 2015,
3859840; Mogensen, 2015, 3859839; Gyllenhammar, 2018, 4778766}. Distribution is influenced
by the chemical properties of PFAS including length, lipophilicity, and branching. In the Mondal
study {Mondal, 2014, 2850916}, mean maternal serum PFOS concentrations were lower in
3-9
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
breastfeeding mothers vs. non-breastfeeding mothers. Conversely, breastfed infants had higher
mean serum PFOS than infants who were never breastfed. Maternal serum concentrations
decreased with each month of breastfeeding {Mondal, 2014, 2850916; Mogensen, 2015,
3859839}. Cariou et al. (2015, 3859840) reported that PFOS levels in breastmilk were
approximately 66-fold lower relative to maternal serum and the ratio between breastmilk and
maternal serum PFOS was 0.38 ± 0.16. The authors noted that the transfer rates of PFAS from
serum to breastmilk were lower compared to other lipophilic persistent organic pollutants such as
poly chlorinated biphenyls.
Developmental studies in rodents confirmed PFOS distribution from rat and mouse dams to
fetuses and pups, as well as variable PFOS level across many fetal tissues {Luebker, 2005,
1276160; Chang, 2009, 757876; Ishida, 2017, 3981472; Zeng, 2011, 1326732; Chen, 2012,
1276152; Borg, 2010, 2919287; Liu, 2009, 757877}.
3.3.1.2.5VoIume of Distribution in Humans and Animals
In humans, a single volume of distribution (Vd) value of 239 mL/kg has been uniformly applied
for most PFOS studies {Thompson, 2010, 2919278}. Gomis et al. (2017, 3981280) used a Vd of
235 mL/kg by averaging Vd values estimated for both humans and animals. Vd values may be
influenced by differences in distribution between males and females, between pregnant and non-
pregnant females, and across serum, plasma, and whole blood.
Vd estimates derived in monkeys, mice, and rats vary by species, age, sex, and dosing regimen.
For example, Huang et al. (2019, 5387170) calculated the apparent volume of central and
peripheral distribution in rats. In this study, a two-compartment model was the best fit for male
rats for both IV and gavage routes of administration and females dosed by the IV route, whereas
a one-compartment model was the best fit for female rats dosed by oral gavage. Vd values in
females after IV administration were lower than that observed in males in both the central and
peripheral compartments. For the oral route, striking sex differences were noted between the
central and peripheral compartments. While Vd values were quite similar in males for both
compartments, they were notably higher in the central compartment compared to the peripheral
compartment in females. Interestingly, another study found that for PFOS, a classical
compartment model was not applicable {Iwabuchi, 2017, 3859701}. Rather, the body organs
behaved as an assortment of independent one-compartments with a longer elimination half-life in
liver than serum in the elimination phase. Further discussion on the Vd for PFOS can be found in
Section 6.6.2.
3.3.1.3 Metabolism
Consistent with other reports and reviews {U.S. EPA, 2016, 3603365; ATSDR, 2018, 9642134;
Pizzuro, 2019, 5387175}, the available evidence demonstrates that PFOS is not metabolized in
humans, primates, or rodents.
3.3.1.4 Excretion
Excretion data are available for oral exposure in humans and laboratory animals. Most studies
have investigated the elimination of PFOS in humans, cynomolgus monkeys, and rats. Available
evidence supports urine as the primary route of excretion in most species, though fecal
elimination is prominent in rats. In rats, hair is another route of elimination in both males and
3-10
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
females. In females, elimination pathways include menstruation, pregnancy (cord blood,
placenta, amniotic fluid, and fetal tissues) and lactation (breast milk) (see PFOS Appendix).
3.3.1.4.1Urinary and Fecal Excretion
Urinary excretion is considered the main route of PFOS excretion in humans. Zhang et al. (2015,
2851103) estimated a daily urinary excretion rate of 16% of the estimated total daily intake for
PFOS for adults. Zhang et al. (2013, 3859849) calculated median renal clearance rates of
0.044 mL/kg/day in young women and 0.024 mL/kg/day in men and older women for total
PFOS. In a later study, Fu et al. (2016, 3859819) estimated a urinary clearance rate
0.010 mL/kg/day (geometric mean for men and women). These studies showed that PFOS daily
renal clearance values were significantly lower in males compared to females.
Several studies in rats suggest that the fecal route is as or more important than the urinary route
of excretion for PFOS. In a study by Chang et al. (2012, 1289832), excretion in urine and feces
were approximately equivalent when examined 24 and 48 hours after oral gavage administration
of 14C-PFOS. A study by Kim et al. (2016, 3749289) measured the amounts of unchanged PFOS
excreted into the urine and the feces of male and female Sprague-Dawley rats for 70 days after a
single dose of 2 mg/kg by oral or IV administration {Kim, 2016, 3749289}. PFOS levels in urine
and feces were similar in both males and females, which correlated to similar half-life estimates
for PFOS (26.44 and 28.70 days in males and 23.50 and 24.80 days in females by the oral and IV
routes, respectively).
In summary, evidence supports excretion through the fecal route in both animals and humans.
Human studies indicate excretion by the fecal route is substantially lower than that observed by
the urinary route. In rats, however, both urinary and fecal routes play prominent roles in PFOS
elimination. There are sex-specific differences in fecal excretion of PFOS. Excretion through the
fecal route appears to be more efficient in males compared to females. Also, in male rats, fecal
and urinary concentrations were similar after oral but not IV dosing. Finally, exposures to
mixtures of PFAS suggest that PFOS in the context of a mixture may be preferentially excreted
through the fecal route. The extent to which resorption by hepatic and enteric routes impacts
fecal excretion has not been established in either humans or animals.
3.3.1.4.2Renal and Enterohepatic Resorption
Early evidence of enterohepatic resorption of PFOS was revealed by Johnson et al. (1984,
5085553), who demonstrated that cholestyramine (CSM) treatment increased mean cumulative
14C elimination in feces by 9.5-fold for male CD rats administered 3.4 mg/kg 14C-PFOS. CSM is
a bile acid sequestrant, and its facilitation of PFOS gastrointestinal clearance suggests
enterohepatic circulation.
Several studies present evidence of enterohepatic excretion and potential resorption in humans
{Genuis, 2010, 2583643; Harada, 2007, 2919450}. Harada et al. (2007, 2919450) estimated a
biliary resorption rate of 0.97, which could contribute to the long half-life in humans. Genuis et
al. (2010, 2583643) described a case report of excretion analyzed after inhalation PFOS
exposure. After treatment with a bile acid sequestrant CSM for 1 week, PFOS serum levels
decreased from 23 ng/g to 14.4 ng/g. Additionally, stool PFOS concentrations increased from
undetectable before treatment (LOD = 0.5 ng/g) to 9.06 and 7.94 ng/g in the weeks after
3-11
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
treatment, suggesting that it may help with removing PFOS that gains access to the GI tract via
bile.
Zhao and colleagues (2015, 3856550; 2017, 3856461) evaluated enterohepatic transporters
identified in liver hepatocytes and intestinal enterocytes in humans and rats. Using in vitro
transfection assays, PFOS was found to be a substrate of both sodium-dependent and -
independent enterohepatic transporters involved in recirculation of bile acids. With the exception
of rat apical sodium-dependent bile salt transporter (ASBT), PFOS was demonstrated to be a
substrate for all tested transporters (sodium/taurocholate cotransporting polypeptide (NTCP),
OATP1B1, OATP1B3, OATP2B1) as well as organic solute and steroid transporter alpha/beta.
Binding efficiency to the enterohepatic transporters was chain-length dependent. NTCP
transported PFAS with decreasing affinity but increasing capacity as the chain length increased
{Zhao, 2015, 3856550}. The opposite trend was seen for OATP-mediated uptake {Zhao, 2017,
3856461}. While these in vitro studies demonstrate that PFOS is a substrate of enterohepatic
transporters found in the livers and intestines of humans and rats, it is as unknown whether and
to what extent these transporters function in vivo.
3.3.1.4.3Maternal Elimination Through Lactation and Fetal Partitioning
PFOS can readily pass from mothers to their fetuses during gestation and through breast milk
during lactation. In conjunction with elimination through menstruation discussed in Section
3.3.1.4.4, females may eliminate PFOS through routes not available to males. The total daily
elimination of PFOS in pregnant females was estimated to be 30.1 ng/day, higher than the
11.4 ng/day for PFOA {Zhang, 2014, 2850251}. The ratio of branched :total PFOS isomers in
cord blood was 0.27 and was higher in cord blood compared to maternal blood and placenta.
These findings suggest branched PFOS isomers may transfer to the fetus more readily than linear
forms. In another study in humans {Zhang, 2013, 3859792}, the mean levels in the cord blood,
placenta, and amniotic fluid were 21%, 56%, and 0.1%, respectively, of levels found in the
mother's blood, demonstrating that cord blood, placenta, and amniotic fluid are additional routes
of elimination in pregnant females. Blood loss during childbirth could be another source of
excretion. Underscoring the importance of pregnancy as a life-stage when excretion is altered,
Zhang et al., (2015, 2857764) observed that the partitioning ratio of PFOS concentrations
between urine and whole blood in pregnant women (0.0004) was lower than the ratio found in
non-pregnant women (0.0013) and may be affected by the increase in blood volume during
pregnancy {Pritchard, 1965, 9641812}.
Mamsen and colleagues (2017, 3858487) measured placental samples and fetal organs in relation
to maternal plasma levels of five PFAS in 39 Danish women {Mamsen, 2017, 3858487}. Fetal
organ levels of PFOS were lower than in maternal blood. The average concentration of PFOS
was 0.6 ng/g in fetal organs compared to 1.3 ng/g in the placenta and 8.2 ng/g in maternal
plasma. Increasing fetal PFOS levels with fetal age suggest that the rate of elimination of PFOS
from mother to fetus may increase through the gestational period.
Afterbirth, women can also eliminate PFOS via lactation {Tao, 2008, 1290895; Lee, 2017,
3983576; Thomsen, 2010, 2186079} and it was shown that PFOS levels in breastmilk are
affected by parity {Lee, 2017, 3983576; Jusko, 2016, 3981718}. In one study, mean PFOS
concentrations were 3.67, 1.38, and 0.040 ng/mL in maternal serum, cord serum, and breast milk,
respectively {Cariou, 2015, 3859840}. The observed ratio of cord serum and maternal serum for
3-12
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
PFOS was 0.38 in this study, much lower than the ratio of 0.78 for PFOA. However, the ratio
between breast milk and maternal serum was 0.038, essentially the same as PFOA. Thus, PFOS
exhibits a low transfer from maternal blood to cord blood and a 10-fold lower transfer from
maternal blood to breast milk.
3.3.1.4.40ther Routes of Elimination
Menstruation may be an important factor in the sex-specific differences observed in PFOS
elimination. Wong et al. (2014, 2851239) estimated that menstrual serum loss is 432 mL/year,
which could account for > 30% of the difference in the elimination half-life between females and
males.
Two studies supported an association between increased serum concentrations of PFOA and
PFOS and early menopause {Knox, 2011, 1402395; Taylor, 2014, 2850915}. However, are-
analysis of these data {Ruark, 2017, 3981395} suggested that this association could be explained
by reverse causality and more specifically, that pharmacokinetic bias could account for the
observed association with epidemiological data. Also challenging the assumption that this is due
to menstruation, Singer et al. (2018, 5079732) failed to find evidence of associations between
menstrual cycle length and PFAS concentrations. Furthermore, Lorber et al. (2015, 2851157)
suggested that factors other than blood loss, such as exposure to or disposition of PFOA/PFOS,
may also help explain the differences in elimination rates between males and females. Curiously,
studies providing direct measurements of PFOS in menstrual blood were not identified.
However, for PFOS to be selectively retained from the blood lost through menstruation would
require a specific mechanism for that process and no such mechanism has been demonstrated or
proposed.
Gao et al. (2015, 2850134) found that hair is potential route of PFAS elimination in rats. A dose-
dependent increase in hair PFOS concentration was observed in all exposed animals. PFOS did
not exhibit the sexual dimorphic pattern in hair noted for PFOA. While hair PFOS levels were
lower in males compared to females in the low dose group, there were no significant differences
in hair PFOS concentrations between males and females in the higher dose groups.
3.3.1.4.5Half-Life Data
There have been several studies of half-lives in humans all supporting a long residence time for
serum PFOS with estimates measured in years rather than months or weeks (see PFOS
Appendix). Because there is no evidence that PFOS is metabolized in mammals, half-life
determinations are governed by excretion. The calculated PFOS half-lives reported in the
literature vary considerably, which poses challenges in predicting both the routes and rates of
excretion. Half-life estimates vary considerably by species, being most rapid in rodents
(measured in hours to days), followed by primates (measured in days to weeks) and humans
(measured in years). Half-life estimates were shorter in human females relative to males, but sex
differences were less clear in animal studies.
Human PFOS half-life estimates range from less than 1 year in a single male child of 16 years
{Genuis, 2014, 2851045} to up to 60.9 years for males occupationally exposed in a facility in
China {Fu, 2016, 3859819} (see PFOS Appendix). With one exception {Genuis, 2014,
2851045}, half-lives estimated for males are longer than those estimated for females and show
an age-related increase {Zhang, 2013, 3859849}. Also, linear isomers exhibit longer half-lives
3-13
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
than branched isomers {Zhang, 2013, 3859849; Xu, 2020, 6781357}. While most studies were
conducted in adults and/or adolescents, at least one study estimated a PFOS half-life of 4.1 years
in newborns {Spliethoff, 2008, 2919368}.
Half-life estimates in humans rely on measured serum and/or urine concentrations. However,
relatively few studies calculated PFOS half-lives along with measured intake and serum and
urine PFOS concentrations {Xu, 2020, 6781357; Worley, 2017, 3859800; Fu, 2016, 3859819;
Zhang, 2013, 2639569} (see PFOS Appendix). PFOS half-life values among these 4 studies
varied dramatically from 1.04 years in Xu et al. (2020, 6781357) to 60.9 years in Fu et al. (2016,
3859819). These comparisons support principles suggested by the broader literature. First, sex
related differences with males exhibiting much longer half-lives compared to females which
may, at least in part, relate to menstruation as an important route of elimination in females
(especially females of reproductive age) may relate, at least in part, to menstruation as an
important route of elimination. Second, Xu et al. (2020, 6781357) suggest that linear PFOS
molecules exhibit longer half-lives than branched forms, which may reflect differential affinities
of linear vs. branched forms for resorption transporters. Third, the relationships between blood
and urine concentrations are not obvious, underscoring the role of non-urinary routes of
excretion and the difficulty in measuring renal resorption. Finally, only two studies estimated
PFOS intake in subjects {Xu, 2020, 6781357; Worley, 2017, 3859800}. Altogether, there is
insufficient data to correlate PFOS intake measurements to serum/plasma and urine
concentrations. These factors, as well as age and health status of subjects, likely contribute to the
variability in PFOS half-life estimates in humans.
In animals, half-life values are reported in days rather than in years. Values in cynomolgus
monkeys ranged from 88 to 200 days {Chang, 2012, 1289832; Seacat, 2002, 757853} and were
generally longer than those observed in rodents, but much shorter than values observed in
humans. Depending on the experimental conditions, half-lives in rats ranged from 14.5 to
43 days {Chang, 2012, 1289832; Huang, 2019, 5387170; Kim, 2016, 3749289}. In contrast to
sex-specific differences in half-lives for PFOA, PFOS half-lives showed only minor differences
between males and females.
33.2 Pharmacokinetic Models
Pharmacokinetic (PK) models are tools for quantifying the relationship between external
measures of exposure and internal measures of dose. For this assessment, PK models were
evaluated for their ability to allow for 1) cross-species PK extrapolation of animal studies of both
cancer and noncancer effects and 2) the estimation of the external dose associated with an
internal dose metric that represents the POD calculated from animal toxicological or
epidemiological studies. The following sections first describe and evaluate published PK
modeling efforts and then present conclusions from analyses that assessed the utility of the
models to predict internal doses for use in dose-response assessment.
Numerous PK models for PFOS have been developed and published over the years to
characterize the unique ADME described in Section 3.3.1. These approaches can be classified
into three categories: classical compartmental models, modified compartmental models, and
PBPK models. With classical compartmental modeling, the body is defined as either a one- or
two-compartment system with volumes and intercompartmental transfer explicitly fit to the
available PFAS PK dataset. Modified compartmental models are more physiologically based in
3-14
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
that they attempt to characterize unique aspects of in vivo ADME through protein binding,
cardiac output, and known renal elimination from the published literature. However, these
models still rely on explicit fitting of data to the non-physiological parameters. Finally, PBPK
models describe the tissues and organs of the body as discrete, physiologically-based
compartments with transport between compartments informed by available data on the
physiologically relevant quantifications of blood flow and tissue perfusion. Determining
additional, non-physiological parameters typically requires explicitly fitting the PBPK model to
time-course concentration data. However, the number of parameters estimated through data
fitting is generally fewer than for classical PK or modified compartmental models. A review of
the available PK models regarding their ability to predict PFOS ADME is provided below.
3.3.2.1 Classical Compartmental Analysis
The most common approach for the prediction of serum levels of PFOS is to apply a relatively
simple one-compartment model. This type of model describes the toxicokinetics of the substance
with a single differential equation that describes the rate of change in the amount or
concentration of the substance over time as a function of the exposure rate and the clearance rate.
This type of model describes the relationship between exposure, serum concentration, and
clearance and can be used to predict one of these values when the other two values are set.
Additionally, because the model can produce predictions of changes in exposure and serum
concentration over time, these models can be applied to fill the temporal gaps around or between
measured serum concentrations or exposures.
Some examples of one-compartment models used to predict human exposure from serum
concentrations include the work of Dassuncao et al. (2018, 4563862) who used a model to
describe historical changes in exposure in seafood and consumer products, Hu et al. (2019,
5381562) who used paired tap water and serum concentration to estimate the proportion of total
exposure that originates from drinking water, and Balk et al. (2019, 5918617) who used
measured concentration in drinking water, dust and air samples, and serum concentrations in
developing children (measured at several time points) to assess the relative proportion of
exposure that originates from dietary exposure. Zhang et al. (2019, 5080526) performed a similar
study using community tap water measurements and serum concentrations to estimate the
proportion of PFOS exposure that originates from drinking water.
Other applications are used to better understand the toxicokinetics of PFOS in humans by
combining estimated exposure values and serum values to estimate clearance and half-life in a
population of interest. One example of this type of model application was presented by Worley et
al. (2017, 3859800) who estimated the half-life of PFOS using exposure predicted from drinking
water PFAS concentration in a community with contaminated drinking water. Fu et al. (2016,
3859819) used paired serum and urine samples from an occupational cohort to estimate the half-
life separately from renal clearance (in urine) and in the whole body (in serum). One of the
largest challenges in the estimation of half-life is the problem of estimating exposure to PFOS.
One common modification of the one-compartment model is to perform a "steady-state
approximation" (i.e., to assume that the rate of change of the serum concentration is zero). This
scenario occurs when an individual experiences constant exposure, constant body habitus, and
constant clearance over a timespan of several half-lives. Due to the long half-life of PFOS,
steady state is a reasonable assumption for adults starting from the age of 25 and above.
3-15
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
However, the steady state approximation cannot be applied for ages younger than 21 years of age
(EPA defines childhood as < 21 years of age; {U.S. EPA, 2021, 9641727}) due to ongoing
development during childhood and adolescence. This growth dilutes the concentration of the
chemical in the body and results in lower levels than would be seen in its absence. Even though
pubertal development including skeletal growth typically ends several years prior to the age of
25, there is a period after growth ceases during which PFOS levels increase until the adult
steady-state level is reached. The general acceptability of the steady-state assumption in adults
has the caveat that pregnancy or breastfeeding will result in changes in serum concentration and
will not be accounted for in the steady-state approximation.
When adopting a steady-state assumption, the rate of change in serum levels over time is zero. It
follows that the ratio between exposure to the substance and clearance determines the serum
concentration. This is the approach used in the 2016 PFOS HESD to determine the constant
exposure associated with a serum concentration {U.S. EPA, 2016, 3603365}. A similar approach
was used in the recent risk assessment performed by CalEPA {CalEPA, 2021, 9416932}.
Publications reporting applications of similar models include the work of Zhang et al. (2015,
2851103) who used paired urine and serum data to estimate the total intake of PFOS and
compared it to the rate of urinary elimination, and Lorber et al. (2015, 2851157) who examined
the effects of regular blood loss due to phlebotomy on PFOS levels and extrapolated that finding
to clearance via menstruation.
In animals, two classical PK models for PFOS have been published since the 2016 HESD. In
Huang et al. (2019, 5387170), male and female Sprague-Dawley rats were dosed via oral gavage
at 2 or 20 mg/kg, through multiple administrations of PFOS at 2 mg/kg/day for five days, or
intravenously at 2 mg/kg. Following the administration of PFOS, rats were sacrificed from 5
minutes up to 140 days post-dosing to characterize the biphasic PK curve. Using plasma data
from these exposure scenarios, Huang and coworkers developed a two-compartment model to
characterize PK parameters of interest such as the alpha- and beta-phase half-life, central and
peripheral compartment volumes, and total PFOS clearance. For each dosing scenario, a single
set of PK parameters were fit, making extrapolation to other dosing scenarios difficult. However,
the authors demonstrate no significant difference between males and females in beta-phase half-
life and overall clearance which is in agreement with previous studies of PFOS PK in rats {Kim,
2016, 3749289}.
Gomis et al. (2017, 3981280) utilized the functional form of a two-compartment model with oral
gavage to predict internal dosimetry of PFOS in rats using PK data from Seacat et al. (2003,
1290852). However, because the scope of the Gomis et al. (2017, 3981280) study involved
predicting internal dose points-of-departure, PK parameters are not presented.
3.3.2.2 Modified Compartmental Models
In addition to the common one-compartment models described above, several models for
humans have been developed to extend the simple one-compartment model to describe the PK
during pregnancy and lactation. The key factors that must be introduced into the model are the
changes in body habitus that occur during pregnancy (e.g., increases in blood plasma volume and
body weight), the distribution and transfer of the substance between the maternal and fetal
tissues, the transfer from the mother to the infant during nursing, and postnatal development,
including growth of the infant during the early period of life. The mathematical formulation of
3-16
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
this type of model requires two differential equations, one describing the rate of change in
amount or concentration in the mother and one describing the rate of change in infants. One such
developmental model with a lactational component was used to predict the maternal serum
concentrations and exposure from measurements of PFOS concentrations in breast milk
{Abdallah, 2020, 6316215}. Verner et al. (2016, 3299692) presented another developmental
model to predict PFOS serum concentrations in the mother and child and predict previous
exposure using mother/child paired serum measurements at different times. This model included
all the key aspects previously mentioned for developmental PK models. Another unique
approach that extended the one-compartment framework was a publication by Shan et al. (2016,
3360127), who estimated the exposure to specific isomers of PFOS using measurements in food,
tap water, and dust to estimate the isomeric profiles of the substances in human serum.
Pharmacokinetic models that can accommodate longer half-life values than would be predicted
based on standard ADME concepts have been published as tools to estimate internal doses for
humans, monkeys, mice, and rats {Andersen, 2006, 818501; Wambaugh, 2013, 2850932;
Loccisano, 2011, 787186; Loccisano, 2012, 1289830; Loccisano, 2012, 1289833; Loccisano,
2013, 1326665; Chou, 2019, 5412429}. The underlying assumption for all the models is
saturable resorption from the kidney filtrate, which consistently returns a portion of the excreted
dose to the systemic circulation and prolongs both clearance from the body (e.g., extends half-
life) and the time needed to reach steady state.
One of the earliest PK models {Andersen, 2006, 818501} was developed for PFOS using two
dosing situations in cynomolgus monkeys. In the first, three male and three female monkeys
received a single IV dose of potassium PFOS at 2 mg/kg {Noker, 2003, 9642133}. For oral
dosing, groups of four to six male and female monkeys were administered daily oral doses of 0,
0.03, 0.15, or 0.75 mg/kg PFOS for 26 weeks {Seacat, 2002, 757853}. This model was based on
the hypothesis that saturable resorption capacity in the kidney would account for the unique half-
life properties of PFOS across species. The model structure was derived from a published model
for glucose resorption from the glomerular filtrate via transporters on the apical surface of renal
tubule epithelial cells.
The renal-resorption model includes a central compartment that receives the chemical from the
oral dose and a filtrate compartment for the glomerular filtrate from which resorption and
transfer to the central compartment can occur. Transfer from the filtrate compartment to the
central compartment decreases the rate of excretion. The resorption in the model was saturable,
meaning that there was proportionally less resorption and greater excretion at high serum PFOS
concentrations than at low concentrations. In addition to decreased renal excretion due to the
renal resorption, excretion is also reduced in the model by implementing a constant proportion of
PFOS that is bound to protein in plasma and is not available for renal filtration.
The model was parameterized using the body weight and urine output for cynomolgus monkeys
{Butenhoff, 2004, 3749227} and a cardiac output of 15 L/h-kg from the literature {Corley, 1990,
10123}. A 20% blood flow rate to the kidney was assumed based on data from humans and dogs.
Other parameters were assumed or optimized to fit the PK data for monkeys. In the IV time
course data, some time and/or dose-dependent changes occurred in distribution of PFOS between
the blood and tissue compartments, and these changes were less noticeable in the females;
therefore, only the female data were used. The simulation captured the overall time course
scenario but did not provide good correspondence with the initial rapid loss from plasma and the
3-17
-------
DRAFT FOR PUBLIC COMMENT MARCH 2023
apparent rise in plasma concentrations over the first 20 days. For oral dosing, the 0.15 mg/kg
dose simulation was uniformly lower, and the 0.75 mg/kg dose simulation was higher than the
data. When compared to PFOA, PFOS had a longer terminal half-life and more rapid approach to
steady-state with repeated oral administration.
Second
Compartment
(Vt, Ctissue)
Oral Dose ka
(Agut)
t 1
Central
Compartment
(Vc. ^central' ^plasma)
IV Dose
Qfil I : Trm Kt
" I ! Tm,
f :
Filtrate Qfi|
Compartment
(Vfi„ Cfil)
Figure 3-4. Schematic for a Physiologically Motivated Renal Resorption PK Model
Adapted from Wambaugh et al. (2013, 2850932).
Building on the work of other researchers, Wambaugh et al. (2013, 2850932) developed and
published a PK model to support the development of an EPA RfD for PFOS {U.S. EPA, 2016,
3603365}. The model was applied to data from studies conducted in monkeys, rats, or mice that
demonstrated an assortment of systemic, developmental, reproductive, and immunological
effects. A saturable renal resorption term was used. This concept has played a fundamental role
in the design of all of the published PFOS models summarized in this section. The model
structure is depicted in Figure 3-4 (adapted from Wambaugh et al. (2013, 2850392)).
Wambaugh et al. (2013, 2850932) placed bounds on the estimated values for some parameters of
the Andersen et al. (2006, 818501) model to support the assumption that serum carries a
significant portion of the total PFOS body load. The Andersen et al. (2006, 818501) model is a
modified two-compartment model in which a primary compartment describes the serum and a
secondary deep tissue compartment acts as a specified tissue reservoir. Wambaugh et al. (2013,
2850932) constrained the total Vd such that the amount in the tissue compartment was not greater
than 100 times that in the serum. As a result, the ratio of the two volumes (serum vs. total) was
estimated in place of establishing a rate of transfer from the tissue to serum, but the rate of
transfer from serum to tissue was also estimated from the data. A nonhierarchical model for
parameter values was also assumed. Under this assumption, a single numeric value represents all
individuals of the same species, sex, and strain. Body weight, the number of doses, and
magnitude of the doses were the only parameters varied for different studies. Measurement errors
were assumed to be log-normally distributed. Table 4-3. in Section 4.1.3.1.1 provides the
3-18
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
estimated and assumed PK parameters applied in the Wambaugh et al. (2013, 2850932) model
for each of the species evaluated.
The PK data that supported the Wambaugh et al. (2013, 2850932) analysis were derived from
two in vivo PFOS PK studies. The monkey PK data were derived from Seacat et al. (2002,
757853) and Chang et al. (2012, 1289832). Data for the rats (male/females) and mice were both
from Chang et al. (2012, 1289832). The data were analyzed within a Bayesian framework using
a Markov Chain Monte Carlo sampler implemented as an R package developed by EPA to allow
predictions across species, strains, and genders and identify serum levels associated with the
NOAEL and LOAEL external doses. Prior distributions for the parameters were chosen to be
vague, uniformed distributions, allowing them to be significantly informed by the data. The
values were assumed to be log-normally distributed constraining each parameter to a positive
value.
3.3.2.3 PBPK Models
An alternative approach to the use of a classical or modified compartmental model is a PBPK
model, which describes the changes in substance amount or concentration in a number of
discrete tissues. One of the main advantages of a PBPK model are the ability to define many
parameters based on physiological data, rather than having to estimate them from chemical-
specific data. Such physiological parameters include, for example, organ volumes and the blood
flow to different organs; they can be measured relatively easily and are chemical independent.
Another advantage is that amount and concentration of the substance can be predicted in specific
tissues, in addition to blood. This can be valuable for certain endpoints where it is expected that a
tissue concentration would better reflect the relevant dosimetry compared to blood concentration.
The first PBPK model developed for this chemical was reported in a series of publications by
Loccisano et al., which together describe the PK of PFOS in rats, monkeys, and humans, in both
adult and developmental (for rat and human) scenarios {Loccisano, 2011, 787186; Loccisano,
2012, 1289830; Loccisano, 2012, 1289833; Loccisano, 2013, 1326665}. These models were
developed based on an earlier "biologically motivated" model that served as a bridge between a
one-compartment model and PBPK by implementing a tissue compartment (similar to a two-
compartment model), an absorption compartment, and a renal filtrate compartment with
saturable renal resorption {Tan, 2008, 2919374}. The work of Tan et al. (2008, 2919374) was a
development of the earlier work of Andersen et al. (2006, 818501) previously discussed. The
PBPK model of Loccisano and colleagues then extended this "biologically motivated" model by
the addition of discrete tissue compartments, rather than a single compartment representing all
tissues.
A series of follow-up studies applied the Loccisano and coauthors' model structure, with
extensions, to address how PK variation in human populations could bias the result of the study.
This consisted of the work of Wu et al. (2015, 3223290) who developed a detailed model of
adolescent female development during puberty and menstrual clearance of PFOS to investigate
the interaction between chemical levels and the timing of menarche, Ruark et al. (2017,
3981395) who added a detailed description of menopause to evaluate how that affects serum
levels and the epidemiological association between early menopause and PFOS levels, Ngueta et
al. (2017, 3860773) who implemented a reduction in menstrual clearance in individuals using
oral contraceptives and the interaction between oral contraceptive use, endometriosis, and serum
3-19
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
PFOS levels, and Dzierlenga et al. (2020, 6315786; 2020, 6833691) who applied a model of
thyroid disease {Dzierlenga, 2019, 7947729} to describe changes in PFOS renal clearance due to
disease state.
In addition to this set of studies, Fabrega et al. (2014, 2850904) updated the model of Loccisano
et al. (2013, 1326665) for humans by modeling a human population using regional food and
drinking water measurements and human tissue data collected from cadavers in a region of
Spain. The use of human tissue data is relatively rare due to the challenges in sourcing human
tissue but may prove preferable to the assumption that human distribution is similar to
distribution in an animal model. However, Fabrega et al. (2014, 2850904) estimated their tissue
to blood partition coefficients from the ratio of tissue concentrations in the cadavers to the
average serum concentrations in live volunteers who lived in the same region but were sampled
several years earlier {Ericson, 2007, 3858652} and they provided no details on how their renal
resorption parameters were estimated from the human blood concentrations. This model was
further applied to a population in Norway and extended to other PFAS {Fabrega, 2015,
3223669}.
Brochot et al. (2019, 5381552) presented the application of a PBPK model for PFOS with
gestation and lactation phases to describe development and predicted maternal, infant, and
breastmilk concentrations over a variety of scenarios including the prediction of maternal levels
across multiple pregnancies.
One of the major challenges in the parameterization of PBPK models for PFOS is the estimation
of the chemical-dependent parameters such as those involved in protein binding and renal
clearance. One way to investigate this issue is to perform in vitro experiments to help inform the
parameters. Worley et al. (2015, 3981311) used in vitro measurements of renal transporter
activity to describe in detail the various steps involved in the renal filtration, resorption, and
excretion of PFOS.
Chou and Lin (2019, 5412429) developed a PFOS PBPK model for rat, mouse, monkey, and
human. Using the model structure of Worley and Fisher (2015, 3223252), parameters were
determined using a hierarchical Bayesian framework to pool datasets across studies for each
species. This model reflects saturable resorption in the proximal tubule cells of the kidney and
fecal elimination through the bile. While the Bayesian approach is ideal for handling multiple
datasets, the method for implementing the Bayesian inference raises questions about the final
posterior parameter distributions. Priors for the hierarchical model were determined using a
least-squares fitting method on the most sensitive parameters as opposed to defining priors using
information from previous studies and letting the data update those priors to determine the joint
posterior distribution of the parameter space. In a subsequent study, Chou and Lin (2021,
7542658) added a gestation/lactation element to the model and parameterized the
gestation/lactation components for rats and humans. This model structure used a three-
compartment fetal model during gestation and a physiologically motivated PK model, similar to
Wambaugh et al. (2013, 2850932) with renal resorption, for the infant. Using this model, the
authors developed HEDs using interspecies extrapolation of the average serum concentration
POD derived from the rat model. While the fits demonstrated good agreement with the
evaluation dataset, parameters for only the rat are available for developmental endpoints.
3-20
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4 Non-Cancer Health Effects Evidence Synthesis and
Integration
3.4.1 Hepatic
EPA identified 23 epidemiological studies (30 publications)7'8 and 25 animal toxicological
studies that investigated the association between PFOS and hepatic effects. Of the
epidemiological publications, 16 were classified as medium confidence, 6 as low confidence, and
7 were considered uninformative (Section 3.4.1.1). Of the animal toxicological studies, 3 were
classified as high confidence, 17 as medium confidence, and 5 were considered low confidence
(Section 3.4.1.2). Studies have mixed confidence ratings if different endpoints evaluated within
the study were assigned different confidence ratings. Though low confidence studies are
considered qualitatively in this section, they were not considered quantitatively for the dose-
response assessment (Section 4).
3.4.1.1 Human Evidence Study Quality Evaluation and Synthesis
3.4.1.1.1 Introduction and Summary of Evidence from the 2016 PFOS HESD
Serum levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) are
considered reliable markers of hepatocellular function/injury, with ALT considered more
specific and sensitive {Boone, 2005, 782862}. Bilirubin and y-glutamyltransferase (GGT) are
also routinely used to evaluate potential hepatobiliary toxicity {Boone, 2005, 782862; EMEA,
2008, 3056793; Hall, 2012, 2718645}. Elevation of liver serum biomarkers is frequently an
indication of liver injury, though not as specific as structural or functional analyses such as
histology findings and liver disease.
There are 6 epidemiological studies (7 publications)8 from the 2016 PFOS HESD {U.S. EPA,
2016, 3603365} that investigated the association between PFOS and hepatic effects. Study
quality evaluations for these 7 studies are shown in Figure 3-5.
7 Multiple publications of the same data: Jain and Ducatman (2019, 5381566); Jain and Ducatman (2019, 5080621); Jain (2019,
5381541); Jain (2020, 6833623); Omoike et al. (2020, 6988477); Liu et al. (2018, 4238514); Gleason et al. (2015,2966740) all
use NHANES data from overlapping years.
8 Olsen (2003, 1290020) is the peer-review paper of Olsen (2001,10228462).
3-21
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Alexander et al., 2003. 1291101 -
Gallo et al.. 2012, 1276142 -
Grice et al.. 2007. 4930271 -
Lin et al.. 2010. 1291111
Olsen et al . 2001. 10228462-
Olsen et al., 2003, 1290020 -
Yamaguchi et al.. 201 3, 2850970 -
¦sSS-®
~~o#>
Legend
I Good (metric) or High confidence (overall)
I Adequate (metric) or Medium confidence (overall)
I Deficient (metric) or Low confidence (overall)
I Critically deficient (metric) or Uninformative (overall)
Figure 3-5. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS and
Hepatic Effects
Interactive figure and additional study details available on HAWC.
The 2016 PFOS HESD {U.S. EPA, 2016, 3603365} describes both cross-sectional and
longitudinal studies that evaluated PFOS and liver enzymes in adults. Two available cross-
sectional studies {Lin, 2010, 1291111; Gallo, 2012, 1276142} reported positive associations
between PFOS exposure and ALT in adults of the general population (see PFOS Appendix). Lin
et al. (2010, 1291111) examined 2,216 adults in NHANES (1999-2000, and 2003-2004) and
observed that higher serum concentrations of PFOS were associated with abnormal liver
enzymes increases in the U.S. general population. With each increase in log-PFOS, serum ALT
and GGT concentrations (U/L) increased by 1.01 units (SE = 0.53) and 0.01 units (SE 0.03),
respectively {Lin, 2010, 1291111}. When PFOA, PFHxS, and PFNA were simultaneously added
in the fully adjusted regression models, one unit increase in serum log-PFOS concentration was
associated with a decrease of 0.19 units (SE = 0.63, p-value = 0.769) in serum ALT
concentration (U/L) and a 0.06 unit (SE = 0.03, p-value = 0.025) decrease in serum log-GGT
concentration (U/L). The four PFAS were moderately correlated with one another, with PFOA
and PFOS most strongly correlated (Spearman correlation coefficient of 0.68), and PFHxS and
PFNA the least correlated (Spearman correlation coefficient of 0.24). Another medium
confidence cross-sectional study (Yamaguchi, 2013, 2850970) conducted in Japan reported a
positive correlation with ALT in addition to factors influencing PFOS exposure.
Gallo et al. (2012, 1276142) reported an analysis of data from the C8 Health Project, reflective
of a highly exposed community. One of the largest studies of PFOS and ALT in adults, Gallo et
al. (2012, 1276142) evaluated 47,092 adults from the C8 Study Project living in communities in
Ohio and West Virginia impacted from a manufacturing-related PFOA-contaminated drinking
water supply. Natural log transformed serum PFOS concentrations were associated with ln-ALT
in linear regression models (regression coefficient: 0.020; 95% CI: 0.014, 0.026) and with
elevated ALT in logistic regression models across deciles of PFOS (OR = 1.13; 95% CI: 1.07,
3-22
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
1.18). There was less consistent evidence of an association between PFOS and GGT or bilirubin
in this study.
Both studies observed a slight positive association between serum PFOS levels and increased
serum ALT values. The association between PFOS and increased serum GGT was less defined.
Total or direct bilirubin showed no association with PFOS in either study. In the Gallo et al.
(2012, 1276142) study, the cross-sectional design and self-reported lifestyle characteristics are
limitations of the study, and while both Lin et al. (2010, 1291111) and Gallo et al. (2012,
1276142) showed a trend, it was not large in magnitude.
Confidence Exposure Study Exposure
Rating Reference Matrix Design Levels Sub-population Comparison EE
Effect Estimate
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Median=20.3 Regression
r I, . c ng/mL (25th - coefficient (per
Medium ~ail° Serum 75th percentile: - 1-ln ng/mL 0.02
confidence sectional 13 7.29 4 increase in
ng/mL) PFOS)
1
1
1
j •
1
1
1
Median=11.3 Regression
r.n,„n ~ ng/mL coefficient (per
Gleason et 0 Cross - 7C,. . . , . n n.
al 2015 Serum sectional (25th"75th ~ 1-in ng/mL 0.01
al--^n5> sectional percenti|e: 7 0_ increase in
18.0 ng/mL) PFOS)
I
1
1
T*~
1
1
I
Median: 23.50 Dnn„ •
/ml Regression
Unetal. Cross- (25th-75th SS'Sii"'8' na?
2010 Sewm sectional percentile: ~ 1l°9 ng/mL 0.82
15.50-33.80 E£nl!
ng/mL) '
i
1
1
•
1
1
I
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Figure 3-6. Overall ALT Levels from Pre-2016 HESD Epidemiology Studies Following
Exposure to PFOS
Interactive figure and additional study details available on Tableau.
Several cross-sectional occupational studies in PFOS production workers reported mostly null
findings with respect to biomarkers of liver disease {Olsen, 2003, 1290020; Olsen, 2001,
10228462}.
Null or inconsistent associations were reported with GGT and bilirubin. There was no evidence
of association with functional hepatic endpoints in these identified studies. No increases in
deaths from cirrhosis of the liver were found in workers at the 3M facility in Decatur, Alabama
{Alexander, 2003, 1291101}. At the same plant, nonsignificant increases in noncancerous liver
disease (including cirrhosis) were observed with cumulative exposure to PFOS {Grice, 2007,
4930271}.
3.4.1.1.2 Study Quality Evaluation Results for the Updated Literature Review
There are 17 studies (23 publications)9 from recent systematic literature search and review
efforts conducted after publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} that
investigated the association between PFOS and hepatic effects. Study quality evaluations for
these 17 studies (23 publications) are shown in Figure 3-7.
9 Multiple publications of the same data: Jain and Ducatman (2019, 5381566); Jain and Ducatman (2019, 5080621); Jain (2019,
5381541); Jain (2020, 6833623); Omoike et al. (2020, 6988477); Liu et al. (2018, 4238514); Gleason et al. (2015,2966740) all
use NHANES data from overlapping years.
3-23
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Of these, 12 were classified as medium confidence, four as low confidence, and seven were
considered uninformative. Of the informative studies, three cross-sectional {Nian, 2019,
5080307; van den Dungen, 2017, 5080340} and multiple publications of data from NHANES
{Jain, 2019, 5381541; Liu, 2018, 4238514; Omoike, 2020, 6988477; Jain, 2019, 5080621; Jain,
2019, 5381566}, one prospective cohort in elderly adults {Salihovic, 2018, 5083555}, and one
occupational cohort of fluorochemical plant workers {Olsen, 2012, 2919185} examined liver
enzymes in adults. In addition, two of the cross-sectional studies {Rantakokko, 2015, 3351439,
Liu, 2018, 4238396} examined functional liver endpoints in adults. In children and adolescents,
four studies were available including one cohort {Mora, 2018, 4239224} and three cross-
sectional studies {Khalil, 2018, 4238547; Jin, 2020, 6315720; Attanasio, 2019, 5412069}, with
one examining function liver endpoints {Jin, 2020, 6315720}. All of the studies measured PFOS
exposure using biomarkers in blood. The uninformative studies were excluded due to potential
confounding {Abraham, 2020, 6506041; Jiang, 2014, 2850910; Predieri, 3889874; Sinisalu,
2020, 7211554}, lack of information on participant selection {Sinisalu, 2021, 9959547}, or use
of PFAS as the dependent variable (in a publication with a more suitable analysis available {Jain,
2020, 6833623} or where the independent variable is a genetic variant and thus not affected by
PFAS exposure {Fan, 2014, 2967086}).
3-24
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
^V600
Abraham et al., 2020, 6506041
Attanasio, 2019, 5412069-
Fari et al., 2014, 2967086-
Gleason et al., 2015, 2966740 -
Jain and Ducatman, 2019, 5381566-
Jain et al., 2019, 5080621
Jain, 2019, 5381541
Jain, 2020, 6833623
Jiang et al., 2014, 2850910-
Jin et al., 2020, 6315720-
Khalil et al„ 2018, 4238547
Liu et al., 2018, 4238396 -I
Liu et al., 2018, 4238514-
Moraet al., 2018, 4239224-
Nian et al., 2019, 5080307-
Olsen et al., 2012, 2919185-
Omoike et al., 2020, 6988477
Predieri et al., 2015, 3889874
Rantakokko et al., 2015, 3351439
Salihovic et al., 2018, 5083555-
Sinisalu et al., 2020. 7211554 -
Sinisalu et al., 2021, 9959547
van den Dungen et al., 2017, 5080340
Legend
| Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-7. Summary of Study Evaluation for Epidemiology Studies of PFOS and Hepatic
Effects"
Interactive figure and additional study details available on HAWC.
a Multiple publications of the same data: Jain and Ducatman (2019, 5381566); Jain and Ducatman (2019, 5080621); Jain (2019,
5381541); Jain (2020, 6833623); Omoike et al (2020, 6988477); Liu (2018,4238514); Gleason et al. (2015,2966740) all use
NHANES data from overlapping years.
3-25
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.1.1.3 Synthesis of Hepatic Injury from the Updated Literature Review
Results for the eight studies that examined ALT are presented in the Appendix (see PFOS
Appendix). Of the available informative studies that measured ALT in adults, statistically
significant positive associations between ALT and PFOS (i.e., increases in ALT as a continuous
measure with higher PFOS exposure levels) were observed in two of five studies {Salihovic,
2018, 5083555; Nian, 2019, 5080307} and multiple NHANES publications, including all the
medium confidence studies. However, the positive associations in Jain et al. (2019, 5381541)
were observed only in obese participants (Figure 3-8). In non-obese participants, associations
were generally null, with an inverse association in non-obese participants with glomerular
filtration (GF) stage of 3B/4. Among low confidence studies in adults, an inverse association was
reported (p < 0.05) in Olsen et al. (2012, 2919185). However, this analysis differed from the
other studies in that the exposure measure used was change in PFOS levels during the study
period. In van den Dungen et al. (2017, 5080340), no association was observed.
In children and adolescents, positive associations were observed in girls in the fourth quartile in
Attanasio (2019, 5412069) and in the low confidence study in obese children {Khalil, 2018,
4238547}. However, inverse associations were observed in Mora et al. (2018, 4239224), which
may indicate that the associations in children are less consistent than in adults or that there are
sex differences in children. Insufficient data were available to assess the potential for effect
modification by sex.
Six studies examined AST and are presented in the Appendix (see PFOS Appendix). In adults,
statistically significant positive associations were observed in the two medium confidence studies
{Nian, 2019, 5080307} and in NHANES studies. Van den Dungen et al. (2017, 5080340)
reported a non-significant positive association. No association was observed in Olsen et al.
(2012, 2919185). In children and adolescents, the medium confidence study {Attanasio, 2019,
5412069} also observed a positive association in girls but not boys, while the low confidence
study {Khalil, 2018, 4238547} reported an inverse association, both not statistically significant.
For the other liver enzymes (bilirubin, GGT), results were generally consistent with ALT and
AST {van den Dungen, 2017, 5080340; Nian, 2019, 5080307; Attanasio, 2019, 5912069} with
the exception of inverse associations (not statistically significant) for GGT in Jain (2019,
5381541) and bilirubin in Salihovic et al. (2018, 5083555).
3-26
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Confidence Exposure Study Exposure
Rating Reference Matrix Design Levels Sub-population Comparison EE
Effect Estimate
-0.04 -0.03 -0.02 -0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Regression
mran(95%CI) coefficient (per
Medium Jainetal., Serum Cohort ' Non-obese 1-log10 ng/mL -0.02
confidence 2019 "V"9™1 increased
(5.8-6.8) pF0S)
r
i
i
i
i
•
i
i
i
i
i
Geometric Regression
mean (95% coefficient (per
Cl)= 5.5 Obese 1-log10 ng/mL 0.02
ng/mL (5.0 - increase in
6.0) PFOS)
i
i
i
i
i
i *
i
i
i
i
1
Sr2"'22 Regression
Nianetal., „ Cross- (25th-75th Excluding ijfSn'fml
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Confidence Literature Exposure
Rating Search Tag Reference Matrix Study Design Exposure Levels Sub-population Comparison EE
0.8 0.9 1.
Effect Estimate j-
0 1.1 1.2 1.3 1.4 1.5
Median=20.3 OR (per 1-ln
Medium Pre-2016 Galloetal. Serum Cross- ?gST pgtlntHe: - irXasein 113
confidence "erature 2012 sectional 13.7-29-4 ng/mL) PFOS)
1
1
1
1
1
1
OR (for decile
Median=20.3 - 2 vs. decile 1 1.01
ng/mL of PFOS)
(IQR=13.7-29.4
ng/mL)
OR (for decile
3 vs. decile 1 1.06
of PFOS)
OR (for decile
4 vs. decile 1 1.11
of PFOS)
OR (for decile
5 vs. decile 1 1.19
of PFOS)
OR (for decile
6 vs. decile 1 1.19
of PFOS)
OR (for decile
7 vs. decile 1 1.2
of PFOS)
l
1
1
1 •
1
1
OR (for decile
8 vs. decile 1 1.24
of PFOS)
1
1
1
1
1
1
OR (for decile
9 vs. decile 1 1.18
of PFOS)
1
1
1
1 •
1
1
OR (for decile
10 vs. decile 1 1.25
of PFOS)
1
1
1
1
1
1
0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
Figure 3-9. Odds of Elevated ALT Levels from Epidemiology Studies Following Exposure
to PFOS
Interactive figure and additional study details available on Tableau.
For functional measures of liver injury, two medium confidence studies (one in adults and one in
children and adolescents) examined histology endpoints. Both studies examined lobular
inflammation. Rantakokko et al. (2015, 3351439) reported higher PFOS exposure levels were
associated with reduced odds of lobular inflammation, whereas Jin et al. (2020, 6315720)
reported the opposite, with OR of 2.9 for 2-4 foci vs. none, though the results in the latter study
were non-monotonic and both were not statistically significant. Jin et al. (2020, 6315720)
additionally reported higher odds (not statistically significant) of nonalcoholic steatosis
(p < 0.05), ballooning, fibrosis, and portal inflammation. Lastly, Liu et al. (2018, 4238396)
examined hepatic fat mass and found no correlation with PFOS exposure.
In summary, across studies in the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} and the
updated systematic review, there is generally consistent evidence of a positive association
between exposure to PFOS and ALT. However, one source of uncertainty in epidemiology
studies of PFAS is confounding across the PFAS, as individuals are exposed to a mixture of
PFAS and it is difficult to disentangle the effects. This cannot be ruled out in this body of
evidence given the attenuation of the association in Lin et al. (2010, 1291111), the only general
3-28
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
population study that performed multi-pollutant modeling. In addition, associations for other
hepatic outcomes were less consistent, including for functional outcomes such as liver disease.
Thus, while there is evidence of an association between PFOS and ALT, there is residual
uncertainty.
3.4.1.2 Animal Evidence Study Quality Evaluation and Synthesis
There are 6 studies from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} and 19 studies from
recent systematic literature search and review efforts conducted after publication of the 2016
PFOS HESD that investigated the association between PFOS and hepatic effects. Study quality
evaluations for these 25 studies are shown in Figure 3-10 and Figure 3-11.
3-29
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Butenhoff et al., 2012, 1276144 -
++
++
I
NR
++
++
++
++
++
++
++
Conley et al., 2022, 10176381 -
+
+
NR
++
B
¦
B
++
B
Curran et al., 2008, 757871 -
++
NR
NR
++
a
++
++
Dong et al., 2011, 1424949-
++
+
NR
++
++
++
++
++
++
Era et al., 2009, 2919358-
+
NR
NR
++
D
D
++
B
Fuentes et al., 2006, 757859 -
+
+
NR
+
+
+
B
++
++
+
Han et al., 2018, 4238554-
+
+
NR
++
D
++
++
+
+
+
Han et al., 2018, 4355066-
++
+
NR
++
++
H
++
+
H
+
Kawamoto et al., 2011, 2919266 -
-
+
NR
¦
++
++
+
3
-
Lai et al.,2018, 5080641 -
+
+
NR
++
-
+
+
+
+
+
Lau et al., 2003, 757854-
++
+
NR
+
+
+
++
B
B
+
Lefebvre et al., 2008, 1276155-
+
NR
NR
++
+
+
++
++
++
+
Liang etal.,2019, 5412467-
+
+
NR
+
NR
+
++
B
B
-
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
NR Not reported
* Multiple judgments exist
Figure 3-10. Summary of Study Evaluation for Animal Toxicological Studies of PFOS and
Hepatic Effectsa'b
Interactive figure and additional study details available on HAWC.
aHan et al. (2018, 4238554) and Wan et al. (2016, 3981504) reported on the same hepatic data as Han et al. (2018,4355066).
bLefebvre et al. (2008, 1276155) reported on the same hepatic data as Curran et al. (2008, 757871).
3-30
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Li etal., 2021, 7643501
NTP, 2019, 5400978
Seacat et al., 2002, 757853
Seacat et al., 2003, 1290852
Thomford, 2002, 5432419
Wan etal., 2016, 3981504
Wan etal., 2020, 7174720
Xing etal., 2016, 3981506
Yan et al., 2014. 2850901
Yang etal., 2021, 7643494
Zhang etal., 2019, 5918673
Zhong etal., 2016, 3748828
B
NR
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Not reported
* Multiple judgments exist
Figure 3-11. Summary of Study Evaluation for Animal Toxicological Studies of PFOS and
Hepatic Effects (Continued) a'b
Interactive figure and additional study details available on HAWC.
aHan et al. (2018, 4238554) and Wan et al. (2016, 3981504) reported on the same hepatic data as Han et al. (2018,4355066).
bLefebvre et al. (2008,1276155) reported on the same hepatic data as Curran et al. (2008, 757871).
Hepatic effects were observed in male and female mice, rats, and monkeys after varying oral
exposure durations and PFOS doses. This includes effects such as increased absolute and relative
3-31
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
liver weight, altered clinical parameters indicating potential liver injury, and histopathological
alterations of liver tissue. Data from numerous studies provide evidence confirming the liver as a
target of PFOS toxicity.
3.4.1.2.1 Liver Weight
Significant increases in liver weight relative to body weight and absolute liver weight were
observed in several strains of male and female mice exposed to 1.25-10 mg/kg/day PFOS for
short-term, subchronic, and gestational durations {Lai, 2018, 5080641; Xing, 2016, 3981506;
Yan, 2014, 2850901; Lau, 2003, 757854; Zhong, 2016, 3748828; Yang, 2021, 7643494; Dong,
2011 1424949}. In male BALB/c mice, significant increases in both relative and absolute liver
weights were observed after 28-day exposure to PFOS doses of 1.25 and 5 mg/kg/day {Yan,
2014, 2850901}. Similarly, two short-term studies in male C57BL/6 mice reported significantly
increased relative liver weights following exposures to 2.5 {Yang et al., 2021, 7643494} or 2.5-
10 mg/kg/day PFOS {Xing, 2016, 3981506}. In a 60-day study in male C57BL/6 mice, Dong et
al. (2011, 1424949) observed a dose-related increase in relative liver weights; at doses above
0.417 mg/kg/day PFOS, the increases were statistically significant compared to control. In a 7-
week gavage study in female CD-I mice, Lai et al. (2018, 5080641) reported significant
increases in absolute and relative liver weights at 3 mg/kg/day PFOS but not 0.3 mg/kg/day.
Two developmental studies in CD-I mice observed higher liver weights in the dams following
gestational PFOS exposure {Fuentes, 2006, 757859; Wan, 2020, 7174720}. Fuentes et al. (2006,
757859) observed significantly increased absolute liver weights in dams exposed to 3 or
6 mg/kg/day PFOS and significantly increased relative liver weights in dams exposed to
6 mg/kg/day PFOS. The dams were exposed from GD 6-18 to 0, 1.5, 3, or 6 mg/kg/day PFOS.
Similarly, Wan et al. (2020, 7174720) reported significantly increased relative liver weights in
dams exposed to 3 mg/kg/day PFOS without changes in maternal body weight (absolute liver
weight not reported). Dams were exposed to 0, 1, or 3 mg/kg/day PFOS from GD 4.5-17.5.
There was a 10% increase in relative liver weight in the fetuses, but the increase was not
statistically significant and may have been related to reduced fetal weight in this group.
Two additional developmental toxicity studies in mice indicate that relative liver weights of pups
exposed to PFOS during gestation may increase and then subsequently return to control levels
after prolonged cessation of exposure during postnatal development {Zhong, 2016, 3748828;
Lau, 2003, 757854}. Zhong et al. (2016, 3748828) dosed C57BL/6J mouse dams with 0, 0.1, 1,
or 5 mg/kg/day PFOS from GD 1-17. Relative liver weights of male and female pups in the
5 mg/kg/day group were significantly increased at postnatal week 4 (PNW 4), but returned to
levels statistically indistinguishable from controls by PNW 8. Similarly, Lau et al. (2003,
757854) exposed pregnant CD-I mice to 0, 1, 5, or 10 mg/kg/day PFOS from GD 1-17 and
found significant increases in offspring liver weights in the 5 and 10 mg/kg/day dose groups at
PNDs 0 and 7 but not PND 35.
Significant increases in relative and absolute liver weights were also observed in male and
female rats exposed to 0.15-20 mg/kg/day PFOS for short-term, chronic, and gestational
durations {NTP, 2019, 5400978; Curran, 2008, 757871; Seacat, 2003, 1290852; Lau, 2003,
757854; Cui, 2009, 757868; Wan, 2012, 1332470; Wan, 2016, 3981504; Han, 2018,
4355066}(Lefebvre et al. (2008, 1276155) reported the same results as Curran et al. (2008,
757871)). An increase in relative liver weight was observed with exposure as low as
3-32
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
0.15 mg/kg/day PFOS administered to female Sprague Dawley rats for 28 days {Curran, 2008,
757871; Lefebvre, 2008, 1276155}. In males from the same study, relative liver weight was
significantly increased at 1.33 mg/kg/day. A similar study in Sprague Dawley rats found that
relative and absolute liver weights were increased in both males and females dosed with
> 0.312 mg/kg/day PFOS for 28 days {NTP, 2019, 5400978}. In a 14-week feeding study,
Seacat et al. (2003, 1290852) also observed similar responses in male and female Sprague
Dawley rats, with significant increases in relative liver weight at the highest dose tested in each
sex (1.33 and 1.56 mg/kg/day, respectively) and increased absolute liver weight (in males only)
at 1.33 mg/kg/day.
In a developmental toxicity study, Lau et al. (2003, 757854) observed inconsistent alterations in
liver weight across time points in Sprague-Dawley rat offspring exposed to 0, 1, 2, or
3 mg/kg/day PFOS from GD 2-GD 21. Significant increases in relative liver weight were
observed in the 2 and 3 mg/kg/day dose groups at PND 5, but not PND 0 or PND 35. No
significant changes in relative or absolute liver weights were observed in Sprague-Dawley rat
dams following 5-day exposure (GD 14-18) to PFOS (0, 0.1, 0.3,1, 3, 10, or 30 mg/kg/day)
{Conley, 2022, 10176381}.
In a subchronic study in cynomolgus monkeys, relative and absolute liver weights were
increased in males and females dosed with 0.75 mg/kg/day PFOS for 182 days (26 weeks)
{Seacat, 2002, 757853}.
3.4.1.2.2Clinical Chemistry Measures
Increases in serum enzymes including ALT, alkaline phosphatase (ALP), AST, and GGT
following PFOS exposure were observed across multiple species, sexes, and exposure paradigms
(Figure 3-12 (mice), Figure 3-13 (male rats), Figure 3-14 (female rats)). These enzymes are often
useful indicators of hepatic enzyme induction, hepatocellular damage, or hepatobiliary damage,
as increased serum levels are thought to be due to hepatocyte damage resulting in release into the
blood {U.S. EPA, 2002, 625713}. Alterations in serum enzyme levels are generally considered
to reach biological significance and indicate potential adversity at levels > 2-fold compared to
controls (i.e., > 100% change relative to control response) {U.S. EPA, 2002, 625713; Hall, 2012,
2718645}.
Two studies in male mice showed statistically and biologically significant increases in serum
enzymes indicative of hepatic or hepatobiliary damage after oral PFOS exposure (Figure 3-12)
{Yan, 2014, 2850901; Xing, 2016, 3981506}. Xing et al. (2016, 3981506) observed a dose-
dependent increase in ALT in male C57BL/6J mice after 30 days of PFOS exposure; ALT levels
were increased by 50% and 88% above control in the 5 and 10 mg/kg/day groups, respectively.
In comparison, in a study of 28-day exposure to 0, 1.25, or 5 mg/kg/day PFOS in male BALB/c
mice, Yan et al. (2014, 2850901) observed much larger increases in ALT in the 5 mg/kg/day
group (> 700%) change), though there was no apparent linear dose-response relationship
observed across the two tested dose levels. Both Yan et al. (2014, 2850901) and Xing et al.
(2016, 3981506) observed statistically but not biologically significant increases in AST with
increasing PFOS dose (responses did not exceed 50%> change from control at any dose level).
Xing et al. (2016, 3981506) observed a similar statistically but not biologically significant
increase in ALP level (53%> change in the 10 mg/kg/day group). Yan et al. (2014, 2850901) also
reported a large increase in ALP (321%> change relative to control) in the 5 mg/kg/day dose
3-33
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
group. Interestingly, a statistically and biologically significant dose-dependent increase in GGT
was observed by Xing et al. (2016, 3981506), with an increase of approximately 140% in the
lowest dose group (2.5 mg/kg/day) and 535% in the highest dose group (10 mg/kg/day),
indicating potential damage to the biliary system {U.S. EPA, 2002, 625713}.
F.ndpoint Study Name Study Design Observation Time Animal Description Dose (ing/kg/day) | Q Statistically significant 9 Not statistically significant I 195% CI |
Alanine Aminotransferase (ALT) Yan et al., 2014,2850901 short-term (28d) 28d Mouse, BALB/c (cf, N=6) 0
1.25
5
1—<
•
H 1
Xing et al., 2016.3981506 subchronic <30dJ 3ld Mouse, C57BL/6J (tf, N=10) 0
2.5
5
10
*
•
©
_o
Alkaline Phosphatase (ALP) Yan ct al.. 2014,2850901 short-term (28d) 28d Mouse. BALB/c Kf. N=6) 0
1.25
5
1
n i
» i
i
| ^
Xing ctal., 2016,3981506 subchronic (30d) 31d Mouse. C57BL/6J (
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
levels in male rats exposed to PFOS via the diet for 4, 14, 27, or 53 weeks, though this study
tested relatively low doses (approximately 0.02 to 1 mg/kg/day).
As with ALT levels, AST levels in male Sprague Dawley rats exposed to PFOS for varying
durations were increased, but the increases did not exceed two-fold compared to controls. Han et
al. (2018, 4355066) reported a statistically significant increase in AST in male rats dosed with
10 mg/kg/day PFOS for 28 days, but the increase was less than a 20% change from the control.
Three other 28-day studies assessing AST levels in male rats either reported changes in AST that
were not dose-dependent {NTP, 2019, 5400978} or not statistically significant between treated
and control groups {Seacat, 2003, 1290852; Curran, 2008, 757871}. Butenhoff et al. (2012,
1276144) also did not observe statistically significant changes in AST levels in male rats
exposed to PFOS via the diet for 4, 14, 27, or 53 weeks at doses up to 0.984 mg/kg/day.
NTP (2019, 5400978) reported statistically significant increases in ALP in male rats after 28-day
PFOS exposure at dose levels as low as 0.625 mg/kg/day. However, these increases only ranged
from approximately 15%—35% change across all doses with statistically significant responses.
Similarly, Curran et al. (2008, 757871) did not observe consistent effects of 28-day dietary
consumption of PFOS on ALP levels at dose levels up to -6.34 mg/kg/day in male rats.
3-35
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Endpoint Study Name Study Design
Alanine Aminotransferase (ALT) Han at al,. 2018,4355066 short-term (28d) 28d
NTP. 2019, 5400978
Seacat et al„ 2003.1290852 chronic (14wk)
Alkaline Phosphatase (ALP) NTP. 2019, 5400978
Aspartate Aminotransferase (AST) Han etal.. 2018.4355066 short-term (28d)
NTP. 2019. 5400978
Butenhoffetal., 2012, 1276144 chronic (2y)
> Animal Description Dose (mg'kg/day)
Rat, Sprague-Dawley (•?, N=6) 0
Rat, Sprague-Dawley (• '. N=10) 0
0,312
PFOS Hepatic Effects - Serum Enzymes in Male Rats
| Q Statistically significant % Not statistically significant [—I 95%
:rl:CD(SD)IGS BR (, C, N=10) C
Butenhoffetal..2012.1276144 chronic (2y) 4wk Rat.Crl:CD(SD)K3SBR(,-T. N=10) 0
Rat. Crl:CD(SD)lGS BR N=10) 0
Rat. Crl:CD(SD)IGS BR N=10) 0
it, Crl:CD(SD)IGS BR ( :,
Rat, Sprague-Dawley N=10) 0
Rat. Sprague-Dawley (;•''. N=6)
n (28d) 29d Rat, Sprague-Dawley ( , N=10) 0
Seacat etal. 2003. 1290852 short-term (4wk) 4wk Rat. Cr1:CD(SD)IGS BR N=10) C
Rat, Crl:CD(SD)IGS BR N=10) 0
Rat, Crl:CD(SD}IGS BR N=10) 0
27wk Rat, Crl:CD(SD)lGS BR N=10) 0
53wk Rat, Crl;CD(SD)IGS BR ( ', N-10) 0
: •
(-•i
'-~-H
-150 -100 -50
Percent control response (%)
Figure 3-13. Percent Change in Serum Enzyme Levels Relative to Controls in Male Rats
Following Exposure to PFOSa,b
Interactive figure and additional study details available on HAWC and Tableau.
ALT = alanine aminotransferase; AST = aspartate aminotransferase; ALP = alkaline phosphatase; d = day; w/wk = week;
y = year; CI = confidence interval.
'Two publications Han et al. (2018, 4238554) and Wan et al. (2016, 3981504) reported on the same data as Han et al. (2018,
4355066) and are not shown in the figure.
b The red dashed lines indicate a 100% increase and decrease from the control response.
3-36
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
As generally observed in male Sprague Dawley rats, there were also statistically but not
biologically significant alterations in serum enzyme levels observed in female Sprague Dawley
rats exposed to PFOS for 4-53 weeks {NTP, 2019, 5400978; Seacat, 2003, 1290852; Butenhoff,
2012, 1276144; Curran, 2008, 757871}. In a 28-day study in female rats, NTP (2019, 5400978)
reported dose-dependent increases in ALT, though these increases reached only approximately
62% change with the highest dose tested (10 mg/kg/day). A second dietary 28-day study in
female rats reported no statistically significant difference between the control group and groups
treated with up to -7.58 mg/kg/day PFOS (Curran et al., 2008, 757871). Similarly, Seacat et al.
(2003, 1290852) observed no significant differences in ALT levels of female rats exposed to
dietary concentrations of PFOS up to -1.56 mg/kg/day for 14 weeks. Butenhoff et al. (2012,
1276144) also did not observe significant changes in ALT levels in female rats exposed to
dietary concentrations of PFOS for 4, 14, 27, or 53 weeks with doses up to -1.25 mg/kg/day.
Both Curran et al. (2008, 757871) and Butenhoff et al. (2012, 1276144) observed statistically
significant decreases in AST levels of female rats exposed to PFOS for 28 days at the highest
dose tested in each study (7.58 and 1.251 mg/kg/day, respectively). These alterations were
approximately 25-26% decreases from control levels in both studies. In contrast, two other 28-
day studies in female rats did not observe significant changes in AST levels compared to controls
{NTP, 2019, 5400978; Seacat, 2003, 1290852} and the statistically significant decrease observed
by Butenhoff et al. (2012, 1276144) at the high dose at the 4-week time point were not observed
at the 14-, 27-, or 53-week time points.
In a developmental exposure paradigm, Conley et al. (2022, 10176381) exposed Sprague-
Dawley dams to PFOS (0, 0.1, 0.3,1, 3, 10, or 30 mg/kg/day) from GD 14-18, and no significant
effects were observed on levels of ALT or AST in serum.
NTP (2019, 5400978) reported statistically but not biologically significant increases in ALP at
dose levels of 2.5 and 5 mg/kg/day in female rats exposed to PFOS for 28 days (increases did not
exceed 35% change with either dose). In another 28-day study, ALP levels in female rats
administered up to 7.58 mg/kg/day PFOS were not significantly different from control levels
{Curran, 2008, 757871}.
3-37
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Endpoint
Alanine Aminotransferase (ALT)
Study Name
Conley etal.. 2021,10176381
NTP. 2019, 5400978
Ssacat at »l„ 2003,1290852 chronic (14wk)
Butenhoff at aL 2012. 1276144 chronic (2y)
Study Design Observation Time Animal Description
developmental (GD14-18} GD18 PO Rat. Sprague-Oawley (,. N
Dose (mg/kgftlay)
PFOS Hepatic Effects - Serum Enzymes in Female Rats
| ® Statistically significant ^ Not statistically significant!—I 95% CI |
Rat. Spraguc-Dawley (';, N=9-10)
0.312
0.625
Rat. Ctl:CD(SD)IGS BR (v, N=10) 0
0.04
0.15
Rat, CrtCD(SD)IGS BR (¦-, N=10) 0
0.029
0.12
0.299
1.251
Rat. Crt:CD(SD)lGS BR (£, N=10) 0
0,020
0.12
0.299
Rat. Cil:CD(SD)IGS BR ('?, N=10)
0.029
0.12
0.299
1.251
Rat, Cri:CD(5D)IGS BR <; . N=10) 0
Alkaline Phosphatase (ALP) NTP. 2019, 5400978
Aspartate Aminotransferase (AST) Conley etal.. 2021,10176381 developmental (G014-16) GD18
Rat. Sprague-Dawley (•.,
0.029
0.12
0.296
1.251
0
0.312
0.625
Seacat at al.: 2003,1290852
NTP, 2019, 5400978
P0 Rat. Sprague-Dawley (,, N=4-6) 0
Rat, Cri:CD(SD)IGS BR r-, N=10) 0
0.05
0.22
0.47
Rat, Spraguo-Dawloy ((,, N=9-10)
0.312
0.625
I.. 2012, 1276144 chronic <2y)
Rat. Cil:CD(SD)IGS BR ( ¦, N=10) 0
0.029
0.12
0.299
1.251
Rat, Cri:CD(SD)IGS BR (!,, N=10) 0
0.029
0.12
0.299
Rat. Cri:CD(SD)lGS BR (i, N=10) 0
0.12
0.299
1.251
Ral. Crf;CD(SD)IGS BR (5, N=10) 0
0.029
0.12
0.299
hS®h
i©
~
3-38
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Figure 3-14. Percent Change in Serum Enzyme Levels Relative to Controls in Female Rats
Following Exposure to PFOSa'b
Interactive figure and additional study details available on HAWC and Tableau.
ALT = alanine aminotransferase; AST = aspartate aminotransferase; ALP = alkaline phosphatase; d = day; w/wk = week;
y = year; CI = confidence interval.
a Two publications Han et al. (2018, 4238554) and Wan et al. (2016, 3981504) reported on the same data as Han et al. (2018,
4355066) and are not shown in the figure.
b The red dashed lines indicate a 100% increase or 100% decrease from the control response.
Neither ALT nor ALP were significantly altered in male or female cynomolgus monkeys dosed
with up to 0.75 mg/kg/day PFOS for 26 weeks {Seacat, 2002, 757853}.
Levels of bilirubin, albumin, and bile salt/acids were also observed to be altered in several
studies in mice, rats, and monkeys. However, these clinical chemistry measurements were
generally altered at higher concentrations of PFOS than were serum enzymes, and changes were
inconsistent across studies. Bilirubin (direct, indirect, or total) was either unchanged or increased
in male rats exposed to > 5 mg/kg/day PFOS and in female rats exposed to > 2.5 mg/kg/day
PFOS {NTP, 2019, 5400978; Curran, 2008, 757871; Seacat, 2003, 1290852}. Total bilirubin
was decreased in male monkeys exposed to 0.75 mg/kg/day for 91-182 days, but there was no
statistically significant response in female monkeys {Seacat, 2002, 757853}. Six studies
examined albumin levels, but only two studies found significant alterations due to PFOS
treatment {Yan, 2014, 2850901; NTP, 2019, 5400978; Seacat, 2003, 1290852; Butenhoff, 2012,
1276144; Curran, 2008, 757871; Conley, 2022, 10176381}. In male mice dosed with 1.25 or
5 mg/kg/day of PFOS for 28 days, albumin was significantly increased above control levels at
both doses {Yan, 2014, 2850901}. In rats dosed with PFOS for 28 days, albumin was
significantly increased in females dosed with 1.25-5 mg/kg/day and in males dosed with
5 mg/kg/day {NTP, 2019, 5400978}. Bile salt/acids were significantly increased in male rats
exposed to 5 mg/kg/day PFOS and in female rats exposed to 2.5 and 5 mg/kg/day PFOS {NTP,
2019, 5400978}. In monkeys, serum bile acids were significantly increased in males, but not in
females, dosed with 0.75 mg/kg/day PFOS {Seacat, 2002, 757853}.
3.4.1.2.3 Histopathology
Liver lesions were confirmed microscopically in male mice and male and female rats in several
short-term and subchronic studies {Wan, 2012, 1332470; Xing, 2016, 3981506; Curran, 2008,
757871; Cui, 2009, 757868; Han, 2018, 4238554; Han, 2018, 4355066; Wan, 2016, 3981504;
NTP, 2019, 5400978; Li, 2021, 7643501} and in two chronic studies of male and female rats and
monkeys {Seacat, 2002, 757853; Butenhoff, 2012, 1276144}. Only three of these studies
provided quantitative incidence data {NTP, 2019, 5400978; Butenhoff, 2012, 1276144; Curran,
2008, 757871).
Hepatocellular hypertrophy was shown to be significantly increased in male Sprague Dawley rats
dosed with 2.5 and 5 mg/kg/day PFOS and in females dosed with 5 mg/kg/day PFOS for 28 days
{NTP, 2019, 5400978} (Table 3-2). Cytoplasmic vacuolation and alterations were significantly
increased in a dose-dependent manner in male and female rats, respectively, in the 2.5 (females
only) and 5 mg/kg/day (males and females) exposure groups {NTP, 2019, 5400978}. Another
28-day study in Sprague Dawley rats observed higher incidence of hepatocellular hypertrophy in
zone 3 of the liver in males exposed to 3.21 and 6.24 mg/kg/day PFOS, the two highest
3-39
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
concentrations; no incidence was seen in females {Curran, 2008, 757871} (Table 3-3). A higher
incidence of cytoplasmic homogeneity in zone 3 of the liver was also observed in both males and
females exposed to 3.21 and 6.24 mg/kg/day PFOS {Curran, 2008, 757871}. In the chronic study
in Sprague Dawley rats {Butenhoff, 2012, 1276144; Thomford, 2002, 5029075}, hepatocellular
hypertrophy was significantly increased in males exposed to 0.098-0.984 mg/kg/day of PFOS
and in females exposed to 0.299-1.251 mg/kg/day for 103 weeks; a dose-response relationship
was observed (Table 3-4).
Table 3-2. Incidences of Nonneoplastic Lesions in Male and Female Sprague-Dawley Rats,
as Reported by NTP (2019, 5400978)
0 mg/kg/day 0.312 mg/kg/day 0.625 mg/kg/day 1.25 mg/kg/day 2.5 mg/kg/day 5 mg/kg/day
Males
Hepatocyte, 0/10 0/10 0/10 3/10 8/10** 10/10**
Hypertrophy
Hepatocyte, 0/10 0/10 0/10 0/10 2/10 4/10*
Vacuolization,
Cytoplasmic
Females
Hepatocyte, 0/10 0/10 0/10 2/10 3/10 10/10**
Hypertrophy
Hepatocyte, 0/10 0/10 0/10 3/10 5/10* 10/10**
Cytoplasmic
Alteration
Notes:
* Statistically significant at p < 0.05; ** p < 0.01.
Table 3-3. Incidences of Nonneoplastic Lesions in Male and Female Sprague-Dawley Rats,
as Reported by Curran et al. (2008, 757871)
Males
0 mg/kg/day
0.14 mg/kg/day
1.33 mg/kg/day
3.21 mg/kg/day
6.34 mg/kg/day
Hepatocyte,
Hypertrophy in
Zone 3
0/4
0/4
0/4
1/4
3/4
Cytoplasmic
Homogeneity in
Zone 3
0/4
0/4
0/4
1/4
3/4
Females
0 mg/kg/day
0.15 mg/kg/day
1.43 mg/kg/day
3.73 mg/kg/day
7.58 mg/kg/day
Hepatocyte,
Hypertrophy in
Zone 3
0/4
0/4
0/4
0/4
0/4
Cytoplasmic
Homogeneity in
Zone 3
0/4
0/4
0/4
1/4
3/4
3-40
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 3-4. Incidences of Nonneoplastic Lesions in Male and Female Sprague-Dawley Rats,
as Reported by Thomford (2002, 5029075)
Males
0 mg/kg/day 0.024 mg/kg/day 0.098 mg/kg/day 0.242 mg/kg/day 0.984 mg/kg/day
Hypertrophy,
0/50
2/50
4/50
17/50
29/50
Hepatocellular,
Centrilobular
Vacuolation,
2/50
3/50
6/50
10/50
10/50
Hepatocellular
Midzonal/Centrilobular
Hyperplasia, Bile Duct
19/50
20/50
25/50
24/50
25/50
Necrosis, Individual
3/50
2/50
6/50
4/50
10/50
Hepatocyte
Altered Hepatocellular,
13/50
21/50
23/50
24/50
24/50
Clear/Eosinophilic Cell
Degeneration, Cystic
5/50
15/50
19/50
17/50
22/50
Females
0 mg/kg/day
0.029 mg/kg/day 0.120 mg/kg/day 0.299 mg/kg/day
1.251 mg/kg/day
Hypertrophy,
2/50
1/50
4/50
15/50
39/50
Hepatocellular,
Centrilobular
Hyperplasia, Bile Duct
21/50
25/50
19/50
17/50
27/50
Necrosis, Individual
3/50
4/50
4/50
5/50
9/50
Hepatocyte
Infiltrate,
33/50
37/50
33/50
36/50
42/50
Lymphohistiocytic
Infiltrate, Macrophage,
2/50
3/50
5/50
6/50
20/50
Pigmented
Degeneration, Cystic
0/50
1/50
1/50
2/50
4/50
Butenhoff et al. (2012, 1276144) and Thomford (2002, 5029075) also observed a dose-
dependent increase in cystic degeneration in male rats exposed to 0.024-0.984 mg/kg/day of
PFOS (Table 3-4); this effect was observed at lower incidences in female rats, but also appeared
to follow a dose-dependent positive trend. Lymphohistiocytic and macrophage infiltrate were
increased in a dose-dependent manner in females exposed to 1.251 mg/kg/day. A dose-response
relationship was also observed with hepatocellular single cell necrosis, which was increased in
males and females exposed to 0.984 and 1.251 mg/kg/day PFOS, respectively {Butenhoff, 2012,
1276144; Thomford, 2002, 5029075}.
The most consistently observed liver lesions following short-term, subchronic, and chronic
exposure to PFOS were hepatocellular hypertrophy and vacuolization. Other liver lesions
commonly observed include single-cell and/or focal necrosis, hepatocytic or cystic degeneration,
and inflammatory cell infiltration. However, in many instances these are qualitatively described
as being observed by the study authors without incidence provided. A single study in male mice
dosed with PFOS for 30 days observed hepatocellular hypertrophy and cytoplasmic vacuolation
in all treatment groups (2.5, 5, and 10 mg/kg/day), but did not provide incidence data to evaluate
3-41
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
a dose response {Xing, 2016, 3981506}. Cytoplasmic vacuolation was also observed in one
study of female mice exposed to 0.1 mg/kg/day PFOS for 60 days {Li, 2021, 7643501}. Male
rats were used in multiple studies and this effect was observed at a range of exposures. Three
studies from the same lab observed hepatocellular hypertrophy in male Sprague Dawley rats
dosed with 1 mg/kg/day of PFOS for 28 days {Han, 2018, 4238554; Han, 2018, 4355066; Wan,
2016, 3981504}; however, none of the studies provided incidence data. Hepatocellular
hypertrophy and centrilobular vacuolation were also observed in another 28-day rat study that
was conducted with higher concentrations of PFOS (5 and 20 mg/kg/day) {Cui, 2009, 757868}.
Hepatocellular hypertrophy was also observed in male and female cynomolgus monkeys exposed
to 0.75 mg/kg/day PFOS for 182 days (incidence data not provided) {Seacat, 2002, 757853}.
Hepatocytic or cystic degeneration, inflammatory cell infiltration, and/or necrosis, were observed
in several short-term and subchronic studies (28-30 days) in male mice and rats {Xing, 2016,
3981506; Cui, 2009, 757868; Han, 2018, 4238554; Han, 2018, 4355066; Wan, 2016, 3981504}.
Livers of male C57BL/6J mice and Sprague Dawley rats dosed with PFOS concentrations
ranging from 2.5-20 mg/kg/day for approximately 4 weeks showed focal or flakelike necrosis,
hepatocytic degeneration, and/or inflammatory cell infiltration {Xing, 2016, 3981506; Cui, 2009,
757868}. Three publications from the same lab described hepatocyte degeneration and
inflammatory infiltration in male Sprague Dawley rats dosed with lower concentrations of
1 mg/kg/day PFOS for 28 days {Han., 2018, 4238554; Han, 2018, 4355066; Wan, 2016,
3981504}. Hepatocytic degeneration and inflammatory cell infiltration were noted in a single
study of female mice, with hepatocyte degeneration being observed in mice exposed to
0.1 mg/kg/day for 60 days and focal infiltration of inflammatory cells being observed in mice
exposed to 1 mg/kg/day {Li, 2021, 7643501}. However, no quantification or statistical analyses
were performed on these studies.
3.4.1.3 Mechanistic Evidence
Mechanistic evidence linking PFOS exposure to adverse hepatic outcomes is discussed in
Sections 3.2.2, 3.2.3, 3.2.5, 3.3.4, 3.3.5, and 3.4.1.1 of the 2016 PFOS HESD {U.S. EPA, 2016,
3603365}. There are 56 studies from recent systematic literature search and review efforts
conducted after publication of the 2016 PFOS HESD that investigated the mechanisms of action
of PFOS that lead to hepatic effects. A summary of these studies is shown in Figure 3-15.
3-42
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Mechanistic Pathway
Animal
Human
In Vitro
Grand Total
Atherogenesis And Clot Formation
0
0
1
1
Big Data, Non-Targeted Analysis
9
0
5
14
Cell Growth, Differentiation, Proliferation, Or Viability
13
1
25
35
Cell Signaling Or Signal Transduction
13
1
15
25
Fatty Acid Synthesis, Metabolism, Storage. Transport, Binding, B-Oxidation
16
0
10
24
Hormone Function
3
1
0
4
Inflammation And Immune Response
5
1
2
7
Oxidative Stress
6
0
7
12
Renal Dysfunction
1
0
0
1
Xenobiotic Metabolism
3
1
6
10
Other
3
0
0
3
Grand Total
30
2
30
56
Figure 3-15. Summary of Mechanistic Studies of PFOS and Hepatic Effects
Interactive figure and additional study details available on Tableau.
3.4.1.3.1Nuclear Receptor Activation
3.4.1.3.1.1 Introduction
The ability of PFOS to mediate hepatotoxicity via receptor activation has been investigated for
several receptor-signaling pathways, including that of the peroxisome proliferator-activated
receptor (PPAR), pregnane X receptor (PXR), constitutive androstane receptor (CAR), liver X
receptor (LXR), and retinoic acid receptor (RAR). Activation of PPARa has been cited as a
mechanism of action for PFAS, including PFOS, because of the association between increased
liver weight and peroxisome proliferation downstream of PPARa activation in rats. However,
increased hepatic lipid content in the absence of a strong PPARa response (i.e., activation of
downstream target genes) is a characteristic of exposure to PFOS, and many of the genes
activated by PFOS are associated with nuclear receptors other than PPARa, namely CAR and
LXR {U.S. EPA, 2016, 3603365}. PPAR, PXR, CAR, LXR, and RAR are nuclear receptors that
can form heterodimers with one another to induce transcription of linked genes, and therefore,
the effects of PFOS on one or multiple receptors may contribute to mechanisms underlying
hepatotoxicity {U.S. EPA, 2016, 3603365}. Additionally, hepatic effects observed with PFAS
exposure including inflammation and necrosis cannot be fully explained by PPARa activation
(Section 3.4.1.2.3). This updated assessment includes studies that have examined activation of
PPARs (including PPARa, p/S, and y), CAR, PXR, LXR, and/or retinoid X receptor (RXR)
activation, as well as the downregulation of hepatocyte nuclear factor 4-alpha (HNF4a) as
potential mechanisms underlying the hepatic health effects induced by PFOS.
3.4.1.3.1.2 Receptor Binding and Activation
Receptor binding and activation assays have been conducted in vitro with the goal of examining
the potential association between activation of PPARs, CAR, PXR, and LXR and PFOS-
mediated hepatotoxicity. PPARs modulate gene expression in response to exogenous or
3-43
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
endogenous ligands and play essential roles in lipid metabolism, energy homeostasis,
development, and cell differentiation {U.S. EPA, 2016, 3603365}.
Several studies used luciferase reporter assays to examine the activation of PPARa by PFOS in
vitro with human and animal cell lines transfected with human or mouse PPARa with varying
results {Wolf, 2014, 2850908; Rosenmai, 2018, 4220319; Takacs, 2007, 783393; Wolf, 2008,
716635; Behr, 2020, 6305866}. In COS-1 cells transfected with mouse PPARa, PPARa was
activated in a concentration-dependent manner, with an approximate half maximal effective
concentration (EC50) of 65 |iM in one study {Wolf, 2014, 2850908} and a lowest observed
effect concentration (LOEC) of 90 |iM for PPARa activation in another study {Wolf, 2008,
716635}. However, a third study in transfected COS-1 cells found that PFOS activated mouse
PPARa, with a significant increase in activity only at a concentration of 120 |iM, but not at lower
concentrations of 1-90 |iM or at higher concentrations of 150 or 250 |iM {Takacs, 2007,
783393}. In cell lines transfected with human PPARa, one study showed that PPARa was
activated in COS-1 cells in a dose-dependent manner, with a LOEC of 30 |iM {Wolf, 2008,
716635}. A second study in HEK293T cells showed that human PPARa was only activated (i.e.,
upregulated by approximately 1.5-fold) at the highest concentration of 100 |iM {Behr, 2020,
6305866}. However, two additional studies reported that PFOS did not significantly increase the
activity of human PPARa up to concentrations of 100 |iM in HepG2 cells {Rosenmai, 2018,
4220319} or 250 |iM in COS-1 cells {Takacs, 2007, 783393}. In every study that compared the
ability of PFOS to activate PPARa with that of PFOA, PFOS was a weaker PPARa activator
{Wolf, 2014, 2850908; Rosenmai, 2018, 4220319; Takacs, 2007, 783393; Wolf, 2008, 716635;
Behr, 2020, 6305866}.
In vitro luciferase reporter assays have also been used to examine the ability of PFOS to activate
other PPAR receptors, namely PPARy and PPARp/S {Bagley, 2017, 4238503; Takacs, 2007,
783393; Zhang, 2014, 5081455; Behr, 2020, 6305866}. One study showed that PFOS
significantly activates human PPARy by 1.5-fold at 10 |iM and by 3-fold at 100 |iM in a
luciferase assay in HepG2 cells {Zhang, 2014, 5081455}. The authors also performed a cell-free
binding assay to show that PFOS binds to human PPARy with a half maximal inhibitory
concentration (IC50) of 13.5 |iM and dissociation constant of 93.7 |iM. Mouse and rat PPARy
were also activated at 100 |iM with a luciferase reporter assay conducted in Chinese hamster
ovary (CHO) cells {Bagley, 2017, 4238503}. However, two other studies did not observe
activation of PPARy by PFOS {Behr, 2020, 6305866; Takacs, 2007, 783393}: PFOS did not
activate human PPARy or PPAR8 in HEK29 cells at concentrations of up to 100 |iM {Behr,
2020, 6305866}, and neither human nor mouse PPARy were activated by concentrations of up to
250 |iM PFOS in COS-1 cells {Takacs, 2007, 783393}. This study conducted in COS-1 cells
also examined activation of human and mouse PPARp/S and observed activation of mouse
PPARp/S only at concentrations of 20 and 30 |iM, but not at a lower concentration of 10 |iM or
at higher concentrations of 40-80 |iM. Human PPARp/S was not shown to be activated by PFOS
in this study. Furthermore, this study demonstrated that the activities of mouse PPARa, y, and
p/S were more responsive than their human counterparts to positive control agonists and
antagonists, demonstrating species-specific differences in receptor-activation {Takacs, 2007,
783393}. Given the discrepancies in the ability and magnitude of PFOS to activate either mouse
or human PPAR receptors, the role of PPAR activation in mediating hepatotoxicity of PFOS is
not fully understood.
3-44
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Two studies examined the activation of CAR/PXR and/or LXR/RXR in vitro with luciferase
reporter assays using HEK293 cells or CHO cells {Bagley, 2017, 4238503; Behr, 2020,
6305866}. No activation of human CAR, human PXR, rat PXR, rat LXRP, human LXRa, or
human RXRa was observed with concentrations of up to 100 |iM PFOS. However, a luciferase
reporter assay in HepG2 cells showed that PFOS activates human PXR with an EC so of 7.87 |iM
{Zhang, 2017, 3604013}. Notably, these studies did not examine endogenous receptor
activation, though other lines of evidence are available that evaluate endogenous receptor
signaling in vivo and in vitro.
3.4.1.3.1.3 Receptor Signaling
3.4.1.3.1.3.1 In Vivo Models
PFOS can activate PPARa in rodents and humans. However, the extent to which activation of
PPARa mediates hepatoxicity may be species-specific, and activation of other receptors may
also contribute to toxicity {U.S. EPA, 2016, 3603365}. Indeed, several studies in Sprague
Dawley rats have found evidence that PFOS may activate both PPARa and CAR/PXR in the
liver {Dong, 2016, 3981515; NTP, 2019, 5400978; Martin, 2007, 758419; Elcombe, 2012,
1401466; Chang, 2009, 757876; Elcombe, 2012, 1332473}. In an acute/short term study, male
rats were exposed to 10 mg/kg/day PFOS for 1, 3, or 5 days, and gene expression changes were
assessed in their livers with an expression microarray {Martin, 2007, 758419}. Although PFOS
exposure induced PPARa-regulated genes and pathway analysis revealed that PFOS clustered
with PPARa agonists (e.g., bezafibrate, clofibric acid, and fenofibrate), the correlation between
the gene response to PFOS and that of known peroxisome proliferators was weak (with a
correlation coefficient of 0.26 for PFOS, in comparison to 0.76 for PFOA). Changes in
cytochrome P450 3 A (Cyp3a) genes were also observed, consistent with the activation of
CAR/PXR.
Another transcriptomics study of the liver of rats exposed to 50 mg PFOS/kg diet for 28 days had
similar results using an expression microarray {Dong, 2016, 3981515}. Upstream regulator
analysis using Ingenuity Pathway Analysis (IPA, Qiagen) revealed that PFOS likely activated
both PPARa and CAR/PXR, with alterations in 48 genes that have evidence of being regulated
by PPARa in the IPA reference database (approximately 10% of all known genes in this
pathway), and 29 genes from the reference database for the CAR/PXR pathway (approximately
14% of all known genes in this pathway). Two other studies support these results, reporting that
genes regulated by either PPARa or CAR/PXR are altered by PFOS, according to qPCR analysis
{NTP, 2019, 5400978; Chang, 2009, 757876}. In a developmental rat study, dams were dosed
with 1 mg/kg/day PFOS from GD 0-19, and the expression of both PPARa- and CAR/PXR-
regulated genes was found to be increased in liver samples from the dams on GD 20 and male
offspring on PND 21; female offspring were not tested {Chang, 2009, 757876}. A 28-day study
in male and female rats found increases in the expression of both PPARa-regulated genes
(Cyp4al, Acoxl) and CAR-regulated genes (Cyp2bl, Cyp2b2) at all exposure concentrations
tested (0.312-10 mg/kg/day) {NTP, 2019, 5400978}. However, there were apparent sex
differences in this study; PPARa-regulated genes were increased by 2- to 31-fold in males and
by 1.3- to 3-fold in females, while CAR-regulated genes were increased by 6- to 400-fold in
males and 32- to 1,227-fold in females. Although Acoxl was the least responsive gene in males,
with increased expression in males exposed to 5 and 10 mg/kg/day and in females exposed to
3-45
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
0.312-10 mg/kg/day, the corresponding enzyme activity (acyl-CoA oxidase) was increased in
males exposed to 5 and 10 mg/kg/day, but not in females.
Two studies in male rats provided additional evidence of PFOS activation of PPARa, CAR, and
PXR through the use of enzymatic biomarkers {Elcombe, 2012, 1332473; Elcombe, 2012,
1401466}. In one study, rats were fed diets containing either 20 or 100 ppm (approximately 2
and 10 mg/kg/day, respectively) PFOS for 7 days, and livers were collected on days 1, 28, 56,
and 84 post-exposure {Elcombe, 2012, 1332473}. In the second study, rats were fed the same
dietary PFOS concentrations for up to 28 days, with livers collected on days 1, 7, and 28 of the
exposure {Elcombe, 2012, 1401466}. PPARa, CAR, and PXR activities [as measured by lauric
acid 12-hydroxylation (CYP4A activity), pentoxyresorufin- O-depentylation (PROD; CYP2B
activity), and testosterone 6B-hydroxylation (CYP3 A activity), respectively] were found to be
increased in the liver microsomes of rats exposed to PFOS at most time points and in both
exposure concentrations tested. Liver palmitoyl CoA oxidase (ACOX activity), another marker
of PPARa activity, was not changed after 7 days of exposure to PFOS {Elcombe, 2012,
1332473}, but was shown to be significantly increased at both concentrations after 28 days of
exposure {Elcombe, 2012, 1401466}. However, in another study in male rats exposed to 0.643-
2.205 mg/kg/day PFOS for 28 days or 14 weeks, ACOX activity was unchanged {Seacat, 2003,
1290852}.
Studies in various strains of wild-type (WT) mice also examined PPARa activation as a
mechanism of PFOS-induced liver toxicity {Huck, 2018, 5079648; Wang, 2014, 2851252; Wan,
2012, 1332470; Bijland, 2011, 1578502; Rosen, 2009, 2919338; Lai, 2017, 3981375}. Through
genetic studies and pathway analysis, changes in PPARa signaling or expression of PPARa
and/or downstream target genes were found to be associated with PFOS exposure in several
studies {Wang, 2014, 2851252; Wan, 2012, 1332470; Bijland, 2011, 1578502; Rosen, 2009,
2919338; Lai, 2017, 3981375}. However, these studies also found evidence of upregulation of
other receptors such as PPARy, CAR/PXR, or LXR/RXR. In one study, the authors concluded
that the main mechanism of action of PFOS for observed changes in liver endpoints (increased
absolute liver weight and histopathological changes including cytoplasmic vacuolization and
steatosis) may be mitochondrial P-oxidation, which leads to the accumulation of free fatty acids
and subsequent activation of PPARa {Wan, 2012, 1332470}. In another study, the authors did
not report any changes in the expression of PPARa or a subset of the downstream target genes
examined by qPCR (Acoxl, Pdk4, Cptl) in mice exposed to PFOS with or without high fat diet-
induced hepatic steatosis {Huck, 2018, 5079648}. The authors suggested that alterations in
PPARy may be a mechanism of PFOS-induced liver hepatotoxicity, based on the fact that
PPARy gene expression was induced by PFOS in mice fed a normal diet. However, it should be
noted that PPARy gene expression was also up-regulated in the livers of mice fed a high fat diet
in the absence of PFOS, and PPARy was unchanged in mice exposed to PFOS and fed a high fat
diet.
Two additional studies comparing 129Sl/SvlmJ WT mice to Ppara-mx\\ mice support PPARa
activation as a mechanism of PFOS-toxicity, but also support the hypothesis that other
mechanisms, including the activation of CAR/PXR, may play a role {Rosen, 2010, 1274165;
Rosen, 2017, 3859803}. The first study found that PPARa-regulated genes were altered in WT
mice dosed with 10 mg/kg/day PFOS for 7 days {Rosen, 2010, 1274165}. However, other genes
and pathways were affected in both WT and Ppara-null mice, including changes related to lipid
3-46
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
metabolism, inflammation, xenobiotic metabolism, and CAR activation (as indicated by
upregulation of Cyp2bl0) {Rosen, 2010, 1274165}. In a connected study, the authors reanalyzed
their data using different expression analysis software than the initial analysis {Rosen, 2017,
3859803}. They found that only approximately 15% of the PFOS-responsive gene changes in the
liver were PPARa-independent, including CAR activation. In both WT and Ppara-null mice,
there were significant similarities in gene expression changes induced by PFOS in comparison to
the CAR biomarker gene set and the CAR agonist phenobarbital {Rosen, 2017, 3859803}. Two
gene expression compendium studies further analyzed these data using gene expression
biomarker signatures built using microarray profiles from livers of WT, Car-null mice {Oshida,
2015, 2850125}, and Ppara-mA\ mice {Oshida, 2015, 5386121}. These analyses found that both
CAR and PPAR were activated by PFOS, and that CAR activation was generally more
significant in Ppara-null mice. The authors concluded that CAR likely plays a subordinate role
to PPARa in mediating the adverse hepatic effects of PFOS {Oshida, 2015, 2850125}.
Comparisons of 129Sl/SvlmJ WT and Ppara-null mice also suggest that increases in liver
weights may not be solely due to activation of PPARa. In the Rosen et al. {2010, 1274165}
study, absolute and relative liver weights were significantly increased in both WT and Ppara-
null mice exposed to 10 mg/kg/day PFOS for 7 days. The absolute liver weights were increased
by 63% in WT mice and by 42% in Ppara-null mice, while relative liver weights were increased
by 44% in both strains. Similarly, in a study of male C57BL/6 (H-2b) mice and Ppara-null
129/Sv mice exposed to 0.005% and 0.02% PFOS in diet for 10 days, absolute liver weight in
WT mice was increased by 95% and 122% in the 0.005%> and 0.02% groups, respectively {Qazi,
2009, 1937260}. In Ppara-null mice, absolute liver weights were increased by 49% and 95% in
the 0.005%) and 0.02% groups, respectively. In a study by Abbott et al. (2009, 2919376), WT
mice were dosed with 4.5-10.5 mg/kg/day PFOS and Ppara-null mice were dosed with 8.5 or
10.5 mg/kg/day from GD 15-18. The authors reported that gestational exposure to 10.5
mg/kg/day resulted in increased relative liver weights in both WT (14%) and Ppara-null (29%)
mouse pups. WT and Ppara-null mouse dams showed 11% and 14% increases, respectively, in
relative liver weights, though these increases were not statistically significant.
A zebrafish study supports the involvement of CAR/PXR and LXR/RXR in PFOS-mediated
hepatic steatosis {Cheng, 2016, 3981479}. Gene expression of liver X receptor alpha (nrlhS),
retinoic acid receptor alpha (rara), retinoid X receptor gamma b (rxrgb), and pregnane X
receptor (nrll2) was elevated in WT male zebrafish livers after exposure to 0.5 |iM PFOS for 5
months, which was accompanied by increased relative liver weight and lipid droplet
accumulation. In female zebrafish, only a slight increase in nrll2 and mild lipid droplet
accumulation was observed; there was no change in relative liver weight.
In comparison to the nuclear receptors mentioned above, the involvement of the nuclear receptor
HNF4a, a regulator of hepatic differentiation and quiescence, has been less frequently studied in
PFOS-induced liver toxicity. Only one in vivo study examined compared gene expression
changes in male WT mice exposed to 10 mg/kg/day PFOS for 7 days with genes regulated by
HNF4a {Beggs, 2016, 3981474}. This study reported that 90 out of 681 genes (13%) altered by
PFOS exposure were regulated by HNF4a. PFOS exposure was shown to decrease the protein
expression of HNF4a in male WT mice. Increased relative liver weight in WT mice was also
3-47
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
observed in this study, and the authors concluded that hepatomegaly, along with other liver
effects such as steatosis and hepatocellular carcinoma (which were not observed in this short-
term study) may be mediated by PFOS-induced dysregulation of HNF4a.
3.4.1.3.1.3.2 In Vitro Models
In vitro genetic studies corroborate the in vivo findings in rodents that suggest PPARa
contributes to the mechanism of PFOS hepatotoxicity but is likely not the only contributor
{Rosen, 2013, 2919147; Bjork, 2009, 2325339; Louisse, 2020, 6833626; Song, 2016, 9959776}.
Two studies conducted in primary rodent and human hepatocytes had conflicting results, with
one study finding no clear pattern of the differential expression of genes associated with PPARa
activation in either mouse or human hepatocytes {Rosen, 2013, 2919147}, and the other study
finding evidence of PPARa activation by altered expression of PPARa signaling pathway genes
in rat hepatocytes, but not in human hepatocytes, neither primary nor HepG2 cells {Bjork, 2009,
2325339}. In a third study in primary human hepatocytes, pathway analysis of gene expression
changes induced by PFOS exposure were not significantly similar to those induced by known
PPARa agonists, which is in contrast to changes following PFOA exposure {Beggs, 2016,
3981474}. However, transcripts associated with CAR/PXR activation were upregulated in
human hepatocytes {Rosen, 2013, 2919147}. In contrast to the results from primary human
hepatocytes, PFOS upregulated PPARa target genes in two human cell lines derived from the
liver, HepaRG and HepG2 cells {Louisse, 2020, 6833626; Song, 2016, 9959776}. Gene
expression patterns in PFOS-exposed HepG2 cells were also consistent with activation of LXR
{Louisse, 2020, 6833626}. Another study in HepG2 cells, however, reported reduced gene
expression of PXR and LXR following treatment with 10-100 [xM PFOS for 24 hours, with the
reduction in PXR being attenuated by 48 hours {Behr, 2020, 6505973}.
The involvement of HNF4a in PFOS-induced hepatotoxicity was examined in two in vitro
studies, and the results support the findings of the in vivo study described above {Beggs, 2016,
3981474; Behr, 2020, 6505973}. In one study, protein levels of HNF4a were decreased in
primary human hepatocytes after 48 and 98 hours of exposure to 10 |iM PFOS {Beggs, 2016,
3981474}. A corresponding decrease in the expression of genes that are positively regulated by
HNF4a (CLDN1, CYP7A1, TAT, and AI)HIB) and increases in genes that are negatively
regulated by HNF4a targets (CCNI)l, AKR1B10, and PLIN2) was observed. A study in HepaRG
cells exposed to 1-100 |iM PFOS for 24 or 48 hours corroborated these findings, as
downregulations in both HNF4a and its target gene CYP7A1 were observed {Behr, 2020,
6505973}.
3.4.1.3.1.4 Conclusions
Although activation of PPARa is a widely cited mechanism of liver toxicity induced by PFAS
exposure, PFOS has been shown to activate a number of other nuclear receptors, including
PPARy, PPARp/S, CAR/PXR, and LXR/RXR. Many of these nuclear receptors, including CAR
and PPARy, are also known to play important roles in liver homeostasis and have been
implicated in liver dysfunction, including steatosis {Armstrong, 2019, 6956799}. Therefore,
PFOS exposure may lead to liver toxicity through the activation of multiple nuclear receptors in
both rodents and humans.
3-48
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.1.3.2 Lipid Metabolism, Transport, and Storage
3.4.1.3.2.1 Introduction
The liver is the primary driver of lipid metabolism, transport, and storage. It is responsible for
the absorption, packaging, and secretion of lipids and lipoproteins. Lipids are absorbed from
digestion through biliary synthesis and secretion, where they are converted to fatty acids {Trefts,
2017, 10284972}. These fatty acids are then transported into hepatocytes, cells that make up
roughly 80% of the liver mass, via a variety of transport proteins such as CD36, FATP2, and
FATP5 {Lehner, 2016, 10284974}. Fatty acids can be converted to triglycerides, which can be
packaged with high or very-low-density lipoproteins (HDL or VLDL, respectively) for secretion.
Lipid handling for the liver is important for energy metabolism (e.g., fatty acid P-oxidation) in
other organs and for the absorption of lipid-soluble vitamins. De novo cholesterol synthesis is
another vital function of the liver {Huang, 2011, 10284973}. Cholesterol is important for the
assembly and maintenance of plasma membranes. Dysregulation of any of these functions of the
liver can have implications for metabolic and homeostatic processes within the liver itself and
other organs and contribute to the development of diseases such as non-alcoholic fatty liver
disease, steatosis, hepatomegaly, and obesity.
The liver is a major site of PFOS deposition and as such, not only influences hepatic lipid levels
but can also alter gene expression for a variety of pathways involved in biological processes
{U.S. EPA, 2016, 3603365}. PFAS have been shown to induce steatosis and increase hepatic
triglyceride levels in rodents via inducing changes in genes directly involved with fatty acid and
triglyceride synthesis. These include genes such as fatty acid binding protein 1 (Fabpl), sterol
regulatory element binding protein 1 (Srebpl), VLDL receptor (Vldlr), and lipoprotein lipase
(Lpll) {Armstrong, 2019, 6956799}. These genes can be altered through PPARa and PPARy
induction pathways due to regulation of HNF4a. PFOS upregulates hepatic nuclear receptor
genes directly involved in lipid metabolism (e.g., Pxr andRaf) and the P-oxidation of fatty acids
(e.g., acyl-CoA oxidase 1 (Acoxl) and carnitine palmitoyltransferase 1A (Cptla)) {Lee, 2020,
6323794}. The responses of lipids, bile acids, and associated genes and processes to PFOS
exposure are dose-, model-, and, for some responses, sex-dependent.
3.4.1.3.2.2 In Vivo Models
While the sections below focus on hepatic-specific measurements of lipids from the available
literature, measurements of lipids in the serum are also important indicators of lipid homeostasis
and alterations in lipid metabolism, transport, and storage due to PFOS exposure. Serum lipid
metrics from both animal and epidemiological studies are reported in Section 3.4.3.2 and Section
3.4.3.1, respectively.
3.4.1.3.2.2.1 Rats
Two studies conducted in both male and female Sprague Dawley rats reported marked effects on
lipid metabolism including sex-dependent effects of PFOS on hepatic outcomes {Bagley, 2017,
4238503; NTP, 2019, 5400978}.
In a study by Bagley et al. (2017, 4238503), male and female rats were exposed to 0 or 100 ppm
of PFOS in their diet for three weeks. In males, the authors observed increased liver choline, an
organic cation critical for the assembly/secretion of lipoproteins and the solubilization of
cholesterol in bile; females fed PFOS diets had no change in liver choline levels. An increase in
3-49
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
hepatic free fatty acids, triglycerides, and liver lipid area percent was also observed in males fed
PFOS, while a decrease was observed in females. This is indicative of hepatic steatosis occurring
in males but not in females. Serum was collected from animals on days 2, 9, 16, and 23 during
the three weeks of dietary PFOS exposure and subsequently analyzed for serum clinical
chemistry. There were transient effects on the serum levels of enzymes related to lipid
metabolism (e.g., lipase, lactate dehydrogenase) in the PFOS-fed groups. In comparison to
controls, there was a reduction in lipase and lactate dehydrogenase in PFOS-fed males at all four
of the timepoints tested. PFOS-fed females had similar reductions in lipase and lactate
dehydrogenase concentrations at every timepoint except day 23. For days 2, 9, and 16, animals
were not fasted prior to serum collection; on day 23, animals were instead fasted overnight, and
serum was collected via exsanguination at necropsy. The gene expression of enoyl-CoA
hydratase and 3-hydroxyacyl CoA dehydrogenase (Ehhadh), one of the enzymes involved in
peroxisomal P-oxidation, was upregulated to a larger degree in females than in males (4.1-fold
vs. 3.7-fold). Similarly, stearoyl-CoA desaturase-1 (Scdl), involved in the conversion of oleic
acid to stearate, was upregulated 9-fold in females (compared to 2-fold in males, a change that
was not significantly different from the control males). While nuclear receptors (such as CAR,
PXR, LXR-a, LXR-P, and PPAR-y) are involved in lipid accumulation, and an upregulation of
the mRNA for enzymes involved in this process (such as Scdl) would indicate their activation,
there was no lipid accumulation in females. Ehhadh was increased in both sexes compared to
controls. Together, this may indicate that steatosis in rats is not induced by activation of these
nuclear receptors or transcription levels of protein involved in key steatosis pathways. The
authors also investigated the effect of choline supplementation along with PFOS administration
and found that the steatosis phenotype persisted in males. The authors hypothesize that increased
efficiency of female hepatic cytosolic fatty acid binding protein results in greater mobilization
from lipid to VLDL causing faster excretion into serum and thus adipose tissue. However, the
authors note that this apparent sex difference in lipid accumulation warrants further study
{Bagley, 2017, 4238503}.
NTP (2019, 5400978) used an oral dosing paradigm of 0, 0.312, 0.625, 1.25, 2.5, or 5 mg/kg/day
for 28 days and measured serum cholesterol and triglyceride concentrations (Section 3.4.3.2).
Notably however, both males and females exhibited an increase in lipid metabolism/oxidation
related genes (Acoxl, Cyp4al, Cyp2bl, and Cyp2b2). An increase in these genes indicates
increases in PPARa and CAR activity.
In addition to the sex differences in liver lipid levels described Bagley et al. (2017, 4238503),
Luebker (2005, 757857) reported that there may also be differences depending on the
developmental stage. Female rats were exposed to 0, 0.4, 0.8, 1.0, 1.2, 1.6, or 2.0 mg/kg/day
PFOS for 42 days (6 weeks) prior to mating through either GD 20 or LD 4. In the GD 20 group,
dams were sacrificed and fetuses collected at GD 21, and liver cholesterol and triglycerides were
measured in dams and fetuses exposed to 0, 1.6, or 2.0 mg/kg/day. In dams, liver cholesterol was
significantly reduced at both doses of PFOS, whereas triglycerides were unchanged. No changes
were observed in fetuses at this timepoint. In the LD 5 groups, dams and pups were sacrificed to
measure liver cholesterol and triglycerides. In dams, liver cholesterol was unchanged at this time
point, and liver triglycerides were significantly increased at 1.6 and 2.0 mg/kg/day. In pups, liver
cholesterol was also unchanged; however, liver triglycerides were significantly decreased in pups
exposed to 1.0-2.0 mg/kg/day in both sexes.
3-50
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.1.3.2.2.2 Mice
Several studies in a variety of mouse models were conducted to investigate the effects of PFOS
on the transcription and translation of lipid metabolism and biliary pathways. The focus of these
studies was to identify key regulators affected by PFOS exposure and the extent to which
pathways were affected. To this end, the studies employed expression microarray, quantitative
reverse transcription polymerase chain reaction (qRT-PCR), Kyoto Encyclopedia of Genes and
Genomes (KEGG) and Ingenuity Pathway Analysis (IPA), and other biochemical measures such
as Western Blot and enzyme-linked immunosorbent assay (ELISA).
3.4.1.3.2.2.2.1 Biochemical and Related Histological Changes
Many biochemical changes occurred with lipids and bile within the liver as well as lipid
transport out of the liver (serum/plasma values). In several mouse studies, triglycerides, total
cholesterol, and/or LDL levels were altered in liver {Lai, 2018, 5080641; Liang, 2019, 5412467;
Huck, 2018, 5079648; Xu, 2017, 3981352}. These changes often had potentially associated
histopathological consequences, with steatosis and other lesions being observed in affected livers
{Liang, 2019, 5412467; Huck, 2018, 5079648; Su, 2019, 5080481}.
In a 4-week study, decreased liver cholesterol was observed in male C57BL/6 mice dosed with 5
mg/kg/day PFOS {Xu, 2017, 3981352}; the mechanism of action was attributed to estrogen
receptor 13 (ER13) and is further described in Section 3.4.1.3.3. In a 7-week study, increased liver
triglycerides were observed in female CD-I mice exposed to 0.3 or 3 mg/kg/day PFOS {Lai,
2018, 5080641}. A yellowish appearance was also noted in the livers of the 3 mg/kg/day group,
which the authors associated with lipid accumulation. The authors hypothesized that the
increased hepatic triglycerides may be due to an impairment in lipid catabolism and/or lipid
export.
A study in Kunming mice investigated lipid metabolism markers within pregnant mice and the
offspring exposed prenatally {Liang, 2019, 5412467}. Lipid dysregulation was present in both
mother and offspring. Specifically, the authors observed increased liver weight and triglyceride
content at the 5 mg/kg/day dose of PFOS in both the mother and offspring. In maternal livers,
hepatomegaly along with hepatic steatosis was observed. Further, the authors also found
increased protein expression of CYP4A14 in offspring. This cytochrome P450 catalyzes the
omega(co)-hydroxylation of medium-chain fatty acids and arachidonic acid in mice and is a
common indicator of PPARa activation. Authors also observed increases in CD36 protein levels,
which has a direct effect on fatty acid uptake by hepatocytes, and decreased levels of the proteins
apolipoprotein B (APOB), a cholesterol transporter, and FGF21 in the PND 1 mouse liver.
Together, this evidence indicates that PFOS undergoes gestational transfer, impairing lipid
homeostasis in the offspring.
In ICR mice exposed to 10 mg/kg/day PFOS for 21 days, lipid-based vacuolization was observed
in the liver, which was accompanied by decreased fibroblast growth factor 21 (FGF21) protein
concentration {Su et al. 2019, 5080481}. This hormone is produced by hepatocytes and regulates
the metabolism of sugar and lipids through receptors in the hypothalamus. Interestingly, vitamin
C showed a protective effect in the study, lowering the effect size of some of the increased
parameters and reducing liver lesions. This indicates that nutritional status can mediate the
hepatotoxicity of PFOS.
3-51
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Beggs et al. (2016, 3981474) observed a decrease in hepatocyte nuclear factor alpha (HNF4a)
protein, a master regulator of hepatic differentiation, in the livers of ten-week-old CD-I mice
exposed to 3 or 10 mg/kg/day PFOS by oral gavage for 7 days. HNF4a regulates liver
development (hepatocyte quiescence and differentiation), transcription of specific liver genes,
and lipid metabolism. This decrease in HNF4a protein occurred without a subsequent reduction
in messenger ribonucleic acid (mRNA) levels but appeared to cause a subsequent upregulation of
genes that are negative targets of HNF4a. For example, downstream proteins such as CYP7al
and perilipin 2 (PLIN2) were reduced. HNF4a is considered an orphan receptor with various
fatty acids as its endogenous ligands. These fatty acids maintain the structure of the receptor
homodimer. PFOA and PFOS are analogous in structure to fatty acids and may also provide
stabilization of the homodimer. The authors investigated the role of PFOS interaction with this
protein via in silico docking models, which showed a displacement of fatty acids by PFOS and
PFOA, possibly tagging HNF4a for degradation. Although the authors, do not directly look at
liver pathology, they hypothesize that steatosis, hepatomegaly, and carcinoma in rodents may be
a consequence of the loss of this protein and also presents a potential mechanism for PFOS
induced hepatic effects in humans {Beggs, 2016, 3981474}.
3.4.1.3.2.2.2.2 Microarray Analyses and RT-PCR
Several studies observed perturbations in lipid transport, fatty acid synthesis, triglyceride
synthesis, and cholesterol synthesis in PFOS-exposed mice {Das, 2017, 3859817; Rosen, 2017,
3859803; Su, 2019, 5080481; Liang, 2019, 5412467; Huck, 2018, 5079648}. Two of these
studies, Das et al. (2017, 3859817) and Rosen et al. (2017, 3859803), investigated the effects of
PFOS on lipid metabolism and homeostasis without the influence of PPARa using nullizygous
models. After exposure to 3 or 10 mg/kg/day PFOS for 7 days, Das et al. (2017, 3859817)
observed that a smaller subset of genes related to lipid homeostasis was activated in l'para-xwxW
mice compared to WT mice. In addition, there were 3-to-4-fold reductions in the genes related to
lipid homeostasis that were expressed in PFOS-exposed Ppara-null mice compared to WT mice,
including carbohydrate response element binding protein (Chrebp), Hnf4a, Ppary coactivator la
(Ppargcla), and sterol regulatory element binding transcription factor 2 (Srebf2). In Ppara-mx\\
mice, there was only a 2-fold decrease in Hnf4a, a 4-fold decrease in Ppargcla, and a 3-fold
increase in Srebfl. Srebf genes encode transcription factors that bind to the sterol regulatory
element-1 motif that is found in the promoter of genes involved in sterol biosynthesis. This
indicates that some of the effects on lipid metabolism are independent of, or only partially
dependent on, PPARa as an upstream regulator.
The results from Das et al. (2017, 3859817) are concurrent with the findings in another study by
the same authors {Rosen et al. 2017, 3859803}, which exposes WT and Ppara-null mice to 10
mg/kg/day PFOS for 7 days. PFOS exposure up-regulated genes related to fatty acid P-oxidation,
lipid catabolism, lipid synthesis, and lipid transport in both strains; however, the increase in
expression was several-fold lower in l'para-xwxW mice than in WT mice. In fact, the authors
suggest that the transcriptome of the mice resembled that of mice treated with PPARy agonists,
thus suggesting a role for other PPAR receptors in the dysregulation of lipid synthesis that occurs
with PFOS exposure. Xu et al. (2017, 3981352), in their investigations using Erfi-null mice
(Section 3.4.1.3.3), found a difference in lipid metabolism and bile acid synthesis between Erfi-
null and WT mice exposed to PFOS. In mice exposed to PFOS, mRNA levels of cholesterol-7a-
hydroxylase (Cyp7al), the rate limiting enzyme in the conversion of cholesterol to bile acid, was
3-52
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
downregulated in WT but not in Erfi-null mice, supporting a role for pathways independent of
PPARa in hepatic lipid responses to PFOS exposure.
Genes involved in lipid homeostasis and regulation were found to be differentially expressed in
mice exposed to PFOS {Su, 2019, 5080481; Liang, 2019, 5412467; Huck, 2018, 5079648}. Key
regulators of fatty acid oxidation including Cyp4al4 and Cd36 were upregulated in the livers of
PND 1 mice exposed during gestation to PFOS {Liang, 2019, 5412467}. Interestingly, genes
related to hepatic export of lipids, such as Apob and Fgf21, were downregulated. Downregulation
of these genes may play a role in the hepatic steatosis, hepatomegaly, and hepatocyte
hypertrophy observed across multiple studies. A study using C57BL/6 mice dosed at 1
mg/kg/day PFOS in the diet for 6 weeks, found that a high fat diet (HFD) protected against
PFOS-induced steatosis and hepatomegaly by inducing Apoal, Apoa2, Apob, and the
microsomal triglyceride transfer protein (Mttp) gene expression {Huck, 2018, 5079648}. Srebfl,
a regulator of hepatic lipogenesis, was significantly induced in PFOS-exposed mice in the HFD
group compared to those fed normal diets. Similarly, gene expression of Cd36, a major lipid
importer, was induced by PFOS in mice fed normal diet but was suppressed in HFD groups,
suggesting that co-administration of PFOS and HFD mitigates steatosis and hepatomegaly.
Together, these results suggest that diet could be a mediating factor in PFOS toxicity and
warrants consideration for evaluation of human hepatic effects.
3.4.1.3.2.2.2.3 Kyoto Encyclopedia of Genes and Genomes (KEGG) and Ingenuity Pathway
Analyses (IPA)
KEGG and IPA tools (Qiagen) are useful for analysis and interpretation of large data sets
generated from transcriptomic profiling. Two studies extensively utilized these tools to
characterize the changes to liver lipid homeostasis. Much like in the studies described in the
previous two subsections, many genes related to the synthesis of fatty acids, including lipid, fatty
acid, triglyceride, linoleic acid and arachidonic acid metabolism, lipid transport, fatty acid
biosynthesis, and triglyceride homeostasis were differentially expressed in mice administered
PFOS {Beggs, 2016, 3981474; Lai, 2017, 3981375}.
Beggs et al. (2016, 3981474) exposed CD-I mice to 0 or 10 mg/kg/day PFOS for 7 days. The
pathway for hydroxylation of lipids was significantly dysregulated in the PFOS-exposed group.
Lai et al. (2017, 3981375) exposed pregnant CD-I mice to 0 or 0.3 mg/kg/day PFOS before
mating through to embryonic day 18.5. Pathway enrichment analysis using KEGG and IPA to
understand the signaling pathways and biological processes that were affected, as evidenced by
differentially expressed genes, highlighted changes in fatty acid metabolism including the
deregulation of the PPAR signaling pathway (not specific to any isoform), fat digestion and
absorption, the biosynthesis of unsaturated fatty acids, and bile secretion in both the maternal and
offspring livers.
3.4.1.3.2.2.3 Zebrafish
Zebrafish have been increasingly used as a model to investigate the toxicity of PFAS. Several
studies have evaluated the toxicity of PFOS in zebrafish, specifically in regard to effects on lipid
metabolism. Similar to the results in rodent models, fatty acid oxidation enzymes and related
gene expression, as well as lipidosis, was increased in PFOS-treated animals {Cheng, 2016,
3981479; Khazaee, 2019, 5918850; Cui, 2017, 3981467; Du, 2014, 2851143}. The authors of
these studies also reported increases in triglycerides, total cholesterol, and free fatty acid
3-53
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
receptors in liver samples from PFOS-exposed zebrafish. Interestingly, as seen in rodent models,
there can be a temporal shift in the levels of proteins or genes involved in lipid metabolism, with
PFOS exposure. Khazaee et al. (2019, 5918850) found that expression levels of the fatty acid
binding protein 1-A gene fabpla, which binds free fatty acids and their coenzyme A derivatives
and is involved in their intracellular transport into the liver, varied over a 30-day period of
exposure to 0.1 or 1 mg/L PFOS. Expression in the liver peaked at day 14 of exposure but being
below control levels at day 30 of exposure. This suggests that lipid metabolism is dynamic, and
the authors concluded that more research is needed to understand if a key time point exists for
evaluating such gene expression changes versus examining such changes over time.
Sex-dependent differences were also observed in a few studies in PFOS-treated zebrafish
{Cheng, 2016, 3981479; Cui, 2017, 3981467}. In one study in which zebrafish were exposed to
0.5 |iM for 5 months beginning at 8 hours post-fertilization (hpf), males tended to have increased
fatty accumulation and reduced hepatic glycogen storage compared to females {Cheng, 2016,
3981479}. In a 2-generation study, Cui et al. (2017, 3981467) observed that the offspring of
zebrafish exposed to PFOS from 8 hpf until 180 days post-fertilization (dpf) tended to have
increased expression of the leptin a (lepa) and insulin receptor a (insr) genes. Diacylglycerol O-
acyltransferase 1 (dgatlb), a metabolic enzyme in triglyceride biosynthesis, and apoal, which
regulates cholesterol transport, were downregulated by PFOS exposure. The authors also noted
that along with indicators of lipid dysregulation, there were morphologically different
mitochondria, potentially exacerbating lipid homeostasis.
3.4.1.3.2.3 In Vitro Models
Two studies reported genetic profiles and pathway analyses in mouse and human hepatocytes to
determine the effect of PFOS treatment on lipid homeostasis and bile synthesis. Rosen et al.
(2013, 2919147) exposed mouse and human primary hepatocytes to 0-250 |iM PFOS for 48
hours. Gene expression was evaluated using microarrays, IP A, and qRT-PCR. For PFOS-
exposed murine hepatocytes, a much smaller group of genes was found to be altered compared to
the whole liver (described in Section 3.4.1.3.4). These included genes associated with P-
oxidation and fatty acid synthesis such as Ehhadh and Fabpl, which were both upregulated with
PFOS exposure. In contrast to the transcriptome of primary mouse hepatocytes, in primary
human hepatocytes, a relatively large group of genes related to lipid metabolism including
PLIN2 and CYPT1A were differentially expressed with PFOS exposure. The authors attribute
some of these differences between mouse and human hepatocytes to a less robust activation of
PPARa in humans. Further, many of the genes investigated were chosen to explore effects of
PFOS exposure that are independent of PPARa activation but may include other nuclear
receptors such as CAR, LXR, PXR and the aryl hydrocarbon receptor (AhR) (Section 3.4.1.3.1).
Beggs et al. (2016, 3981474) exposed human primary hepatocytes to 0.01-100 |iM PFOS for 48
or 96 hours, to determine pathways affected by PFOS exposure. PFOS treatment altered genes
primarily associated with liver necrosis and carcinogenesis. However, pathways associated with
lipid metabolism and bile synthesis (hydroxylation of lipids), including several CYP450 enzymes
associated with lipid homeostasis such as CYP2B6, CYP2C8, CYP3A4, CYP3A5, CYP4A11,
CYP4A22, and CYP7A1 were also altered. Notably, CYP7A1 was among the top ten most
downregulated genes with a fold change of -7.13 indicating potential limitations in the
conversion of cholesterol to bile acid. Importantly, HNF4a, a master regulator of liver function,
regulates many differentially expressed genes related to lipid metabolism which includes all the
aforementioned CYP450s. Together these studies indicate PFOS-induced activation of CYP450
3-54
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
through a variety of PPARa-dependent and independent pathways. Interestingly, there may be
crosstalk between some of these receptors. Beggs et al. (2016, 3981474) notes that HNF4a can
regulate PPARa in mice.
There are several studies that investigated the effect of PFOS on lipid homeostasis using human
cells such as HepG2, HepaRG, and HL-7702 cells. Various endpoints were also investigated in
these cell lines such as mRNA expression through microarray and qRT-PCR assays; lipid,
triglyceride, cholesterol, and choline content; and protein levels via ELISA or Western Blot.
In human hepatic cell lines such as HepaRG or HepG2, PFOS treatment correlated with
suppression of gene expression for genes regulating cholesterol homeostasis. Louisse et al.
(2020, 6833626) noted a concentration-dependent increase in triglycerides, a decrease of
cholesterol, and downregulation of cholesterogenic genes, predominantly with the highest dose
tested, in HepaRG cells exposed to 0-100 |iM PFOS for 24 hours. Cellular cholesterol
biosynthesis genes are regulated by SREBPs, which were also downregulated with PFOS
exposure. In contrast, PPARa-responsive genes were upregulated with PFOS exposure,
particularly at higher doses. Behr et al. (2020, 6505973) also exposed HepaRG cells to 0-100 |iM
PFOS for 24 or 48 hours. Similar to the results from Louisse et al. (2020, 6833626), at 24 hours,
genes related to cholesterol synthesis and transport were downregulated at the highest dose
except for several genes that were upregulated, including bile and cholesterol efflux transporters
(UGT1A1 &n&ABCGl\ and genes involved in bile acid detoxification (CYP3A4). The gene
profiles after 48 hours of exposure were similar, except at the high dose, which saw some
attenuation of the response in cholesterol synthesis and transport. Cholesterol content was
significantly higher in the supernatant at the highest dose of 100 |iM but there was no significant
difference after 48 hours between treated cells and controls, in line with the genetic data of some
response attenuation.
Franco et al. (2020, 6507465) exposed HepaRG cells to 0.0001-1 |iM. Interestingly, lipid levels
were elevated with the lower PFOS concentrations and reduced with the higher PFOS
concentrations. PFOS increased diglyceride levels in a dose-dependent manner except for a
decrease that was observed at the highest concentration. In contrast, triglyceride levels were not
significantly different from controls. This study provides evidence of potential non-monotonic
dose-responses that could result from low-dose PFOS exposures, a potential area that may
require further consideration.
While alterations in lipid metabolism have been reported, Das et al. (2017, 3859817) found that
PFOS did not inhibit palmitate-supported respiration (i.e., mitochondrial metabolism) in
HepaRG cells. There was no effect on oxidation or translocation of palmitoylcarnitine, an ester
involved in the metabolism of fatty acids which plays a role in the tricarboxylic acid cycle.
3.4.1.3.2.4 Conclusions
As described in Section 3.4.3.2, serum lipid concentrations generally decrease with increasing
PFOS doses in rodent bioassays. It is thought that the activation of PPARa, which is less robust
in humans, mediates the effect seen in rodents. In the mechanistic evidence synthesized above, it
appears that PFOS exposure in mammalian and non-mammalian species is associated with
increased lipid accumulation within the liver. Interestingly, studies that measure both serum and
liver lipid content generally follow this trend and report a decrease in serum lipids and an
increase in liver lipid content; this effect may be contributing to the observed PFOS-induced
3-55
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
hepatomegaly and steatosis. Additional data on human liver lipid accumulation would clarify
whether the effects on liver lipid contents in animal bioassays are mechanistically relevant to
humans.
Effects on hepatic lipid metabolism can be observed through the influence of PFOS on not only
PPARa, but other key regulators of hepatic lipid homeostasis such as HNF4a. Gene ontology
using receptor null mice has shown that lipid homeostasis is complex and PFOS is likely acting
on more than one key regulator. Other PPAR isoforms and hormone receptors such as ERP play
a role in regulating lipid and bile metabolism/catabolism, transport, and storage. While minor
conflicts exist between some cell line studies, the evidence supports that PFOS causes lipid
dyshomeostasis and contributes to liver dysfunction and disease, likely through the modulation
of multiple nuclear receptors.
3.4.1.3.3 Hormone Function and Response
While much of the literature relevant to hormone function and response is focused on
reproductive outcomes (See PFOS Appendix), recent literature has also shown a relationship
between hepatic hormonal effects and PFOS exposure. For example, PFOS has been found to
have estrogenic effects. Xu et al. (2017, 3981352) reported an induction of ERP, but not estrogen
receptor alpha (ERa), when wild-type (C57BL/6) male mice were dosed with 5 mg/kg/day PFOS
via oral gavage for 4 weeks. To further explore this relationship, the authors investigated PFOS
administration in male wild-type (WT) and ErP-null mice. They observed no significant changes
in either WT or Erfi- null mice in genes related to lipid metabolism and bile synthesis (3-
hydroxy-3-methylglutaryl-CoA reductase [Hmgcr], scavenger receptor class B type I [Srbi\ low-
density lipoprotein [Ldl], ATP-binding cassette transporter [A heal]) when following exposure to
5 mg/kg/day PFOS for 28 days by oral gavage. However, ATP-binding cassette sub-family G
member 5 (Abcg5), a gene involved in sterol excretion, was increased due to PFOS exposure in
WT mice but not in Erfi-null mice, while cholesterol 7a hydroxylase (Cypla711), the initiator of
cholesterol catabolism, was reduced due to PFOS exposure in WT mice but not in Erfi-null mice.
Further, liver cholesterol levels were significantly decreased in WT PFOS-treated animals but
not in Erfi-mxW mice. This suggests that ERP mediates PFOS hepatotoxicity via altered
cholesterol and bile synthesis. To confirm induction of ERP, the authors also investigated the
response to PFOS exposure in HEPG2 cells. After exposing the cells to 0, 10, or 100 |imol/L of
PFOS for 24 hours, the authors found that ERP was induced at 10 |imol/L, but not at the highest
dose, potentially indicating a non-monotonic dose response.
There is also in vitro evidence that in the liver, genes responsible for a response to hormone
stimulus and hormone metabolism are altered with PFOS exposure {Popovic, 2014, 2713517;
Song, 2016, 9959776}. Differentially expressed genes due to PFOS treatment in these studies
encode proteins such as serine peptidase inhibitor, clade A, proprotein convertase
subtilisin/kexin type 9, activin A receptor type IC, and insulin-like growth factor binding protein
7, all of which are associated with hormone stimulus and/or metabolism. However, it should be
noted that these genes were more significantly altered with PFOA exposure; the authors
indicated that while PFOS was more cytotoxic, PFOA exposure induced more gene alterations,
suggesting that PFOS may be a relatively weak agonist or activator for the transcription factors
or nuclear response elements involved in regulating their transcription {Song, 2016, 9959776}.
3-56
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.1.3.3.1 Conclusions
While there is a small number of studies regarding hormone function and response specifically
within the liver, there is evidence that PFOS has the potential to perturb hormonal balance and
hormonal metabolism in hepatic cells. There is also some evidence from one in vivo study in
mice that PFOS hepatotoxicity may be partially modulated by ER|3, This could have implications
for hormone function and responses in other organ systems and may also be important for mode
of action considerations for hepatotoxicity.
3.4.1.3.4Xenobiotic Metabolism
3.4.1.3.4.1 Introduction
Xenobiotic metabolism is the transformation and elimination of endogenous and exogenous
chemicals via enzymes (i.e., cytochrome P450 [CYP] enzymes) and transporters (i.e., organic
anion transporting peptides [OATPs]) {Lee, 2011, 3114850}. As described in Section 3.3.1.3,
the available evidence demonstrates that PFOS is not metabolized in humans or other species.
However, several studies have investigated how PFOS could alter activation of PXR/CAR as
described in Section 3.4.1.3.1; subsequently, xenobiotic metabolism is altered via manipulation
of the expression of key genes. For instance, the genes for OATP expression (i.e., slcoldl and
slco2bl) in zebrafish or phase I and II biotransformation enzymes in human hepatocytes (i.e.,
CYP3A4), responsible for the transport or metabolism of xenobiotics, may be upregulated or
downregulated following PFOS exposure.
Overall, results from both in vivo and in vitro model systems suggest that genes responsible for
xenobiotic metabolism are upregulated as a result of PFOS exposure.
3.4.1.3.4.2 In Vivo Models
Four studies investigated xenobiotic metabolism endpoints with three studies using Sprague
Dawley rats {Elcombe, 2012, 1401466; Curran, 2008, 757871; Chang, 2009, 757876} and the
remaining study using Ppara-null and WT mice {Rosen, 2010, 1274165}. In a gestational and
lactational exposure study, Chang et al. (2009, 757876) reported increased Cyp2b2 expression in
dams and male pups (2.8-fold and 1.8-fold, respectively). Elcombe et al. (2012, 1401466) also
reported the induction of CYP2B1/2, in addition to CYP2E1 and CYP3A1 proteins, following
test diets of 20 ppm or 100 ppm PFOS. Additionally, Curran et al. (2008, 757871) and Rosen et
al. (2010, 1274165) reported upregulation of Cyp4a22 and Cyp2bl0 expression.
Two studies examined xenobiotic metabolism endpoints, including CYP450 expression and
CYP2B enzyme activity via the PROD biomarker response, in rats {Elcombe, 2012, 1332473;
NTP, 2019, 5400978}. Sprague Dawley rats were exposed to 0, 20, or 100 ppm PFOS for a 7-
day dietary treatment and then were assessed for CYP450 protein expression in the liver at
recovery days 28, 56, and 84 {Elcombe, 2012, 1332473}. Total CYP450 concentration in liver
microsomes was measured via carbon monoxide difference spectrum of ferrocytochrome P450.
Across each dose group and recovery day, mean CYP450 concentrations were increased 123—
189% compared to the control group. However, there was a non-linear PROD dose-response
relationship; the 20 ppm group had decreased mean PROD activity across all recovery days, but
the 100 ppm group had increased activity on recovery days 1 and 28, followed by similar activity
on recovery day 56, then statistically significant decreased PROD activity by recovery day 84.
NTP (2019, 5400978) also assessed Sprague Dawley rats following 28-day treatment of PFOS
3-57
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
(0, 1.25, 2.5, or 5 mg/kg/day) by gavage. Across all treatments of PFOS, females and males both
had increased hepatic expression of Cyp2bl, Cyp2b2, and Cyp4al.
One study examined the expression of genes related to xenobiotic metabolism in zebrafish
{Jantzen, 2016, 3860109}. AB strain zebrafish embryos were exposed to PFOS from 3 to 120
hpf and evaluated at 180 dpf. Female zebrafish had significant reductions in slcoldl expression,
while males had significant reductions in both slcoldl and slco2bl expression {Jantzen, 2016,
3860109}, which are the genes responsible for OATPs and significant in the transport of
xenobiotics {Popovic, 2014, 2713517}. Jantzen et al. (2016, 3860109) noted that in their
previous study, PFOS exposure from 5-14 dpf resulted in significantly reduced slco2bl
expression in zebrafish at 5 dpf but significantly increased expression at 14 dpf {Jantzen, 2016,
3860114}. While their current study reported alterations in gene expression long-term, further
studies with additional time points are needed to elucidate the effect of PFOS exposure on OATP
expression.
3.4.1.3.4.3 In Vitro Models
Gene expression of CYP enzymes responsible for xenobiotic metabolism were assessed in one
study using primary human (e.g., CYP2B6 and CYP3A4 genes) and mouse (e.g., Cyplal and
Cyp3all genes) hepatocytes {Rosen, 2013, 2919147}. Results varied between human and mouse
hepatocytes, with CYP2B6 and CYP3A4 expression upregulated in human hepatocytes, but not in
mouse hepatocytes. The authors noted that the reasons for the differences in gene expression in
the human and mouse hepatocytes were unclear; however, cell density, collection methods, and
time in culture were possible factors, as these were not consistent between models.
Xenobiotic metabolism endpoints were assessed in five studies using hepatic cell lines, including
HepG2 {Shan, 2013, 2850950; Song, 2016, 9959776} andHepaRG {Behr, 2020, 6505973;
Franco, 2020, 6315712; Louisse, 2020, 6833626}. Franco et al. (2020, 6315712) assessed
several phase I biotransformation enzymes following exposure to PFOS concentrations (0.0001,
0.001, 0.01, 0.1, or 1.0 |iM) for 24 or 48 hours. Gene expression of phase I enzymes varied
across concentrations and between the 24- and 48-hour exposures. For CYP1A2, after 24 hours,
the two lowest concentrations resulted in significant increases in expression; however, after 48
hours, the two highest concentrations resulted in significant decreases (~ 10-fold) in expression.
For CYP2C19, after 24 hours, there were no clear trends; however, after 48 hours, expression
was significantly reduced across all concentrations {Franco, 2020, 6315712}.
Evidence varied for CYP3 A4 induction, depending on the model and duration of exposure, as
well as whether gene expression or enzyme activity was assessed {Franco, 2020, 6315712; Behr,
2020, 6505973; Louisse, 2020, 6833626; Shan, 2013, 2850950}. Franco et al. (2020, 6315712)
reported that after 24 hours, there were no clear trends in CYP3A4 expression. However, after 48
hours, CYP3A4 expression was significantly reduced (up to five-fold) across all concentrations
{Franco, 2020, 6315712}. Conversely, Behr et al. (2020, 6505973) and Louisse et al. (2020,
6833626) reported upregulation of CYP3 A4 enzyme activity following 24- or 48-hour PFOS
exposure (1, 10, 25, 50, and 100 |iM) in HepaRG cells, while Shan et al. (2013, 2850950)
reported no significant changes in CYP3A4 enzyme activity following PFOS exposure (0, 100,
200, 300, and 400 |iM) in HepG2 cells.
Franco et al. (2020, 6315712) also assessed gene expression of two phase II enzymes,
glutathione S-transferase mu 1 (GSTM1) and UDP-glucuronosyltransferase 1 Al (UGT1A1),
3-58
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
which were not significantly affected in differentiated HepaRG cells by exposure to PFOS after
24 or 48 hours. The authors noted that it was unclear how PFOS alters gene expression of phase I
enzymes but not phase II enzymes. Further research is needed to determine whether altered gene
expression occurs by interference with cytoplasm receptors, inhibition of nuclear translocation,
or inhibition of the interaction of nuclear translocator complexes with DNA sequences {Franco,
2020, 6315712}.
Song et al. (2016, 9959776) analyzed expression of over 1,000 genes via microarray and gene
ontology analysis in HepG2 cells exposed to PFOS. HepG2 cells were first exposed to 0-1,000
|iM PFOS for 48 h to determine cell viability and cytotoxicity; an IC20 dose of 278 |iM PFOS
was determined from these results. HepG2 cells were then treated with 278 |iM PFOS for 48
hours and used in microarray analysis. As a result of 278 |iM PFOS treatment, 279 genes had
>1.5-fold change in compared to the control group, including genes related to xenobiotic
metabolism by cytochrome P450s such as flavin containing dimethylaniline monoxygenase 5
(FM05), UDP glucuronosyltransferase family 1 member A6 (UGT1A6), glutathione S-
transferase alpha 5 (GSTA5), alcohol dehydrogenase 6 (class V) (ADH6), and glutathione S-
transferase alpha 2 (GSTA2).
3.4.1.3.4.4 Conclusions
Several studies are available that assessed xenobiotic metabolism endpoints as a response to
PFOS exposure, including studies in rats {Elcombe, 2012, 1332473; NTP, 2019, 5400978},
zebrafish {Jantzen, 2016, 3860109}, primary hepatocytes {Rosen, 2013, 2919147}, or hepatic
cell lines {Shan, 2013, 2850950; Song, 2016, 9959776; Behr, 2020, 6505973; Franco, 2020,
6315712; Louisse, 2020, 6833626}. Jantzen et al. (2016, 3860109) reported significant
reductions in the expression of OATPs (slcoldl and slco2bl). While the majority of studies
reported upregulation of gene expression of CYP enzymes {Elcombe, 2012, 1332473; NTP,
2019, 5400978; Franco, 2020, 6315712; Rosen, 2013, 2919147; Behr, 2020, 6505973; Louisse,
2020, 6833626; Song, 2016, 9959776}, direction and magnitude of change varied across doses
and exposure times. Jantzen et al. (2016, 3860109) and Franco et al. (2020, 6315712) both noted
the need for further studies to elucidate any potential relationships between PFOS exposure and
xenobiotic metabolism.
3.4.1.3.5CeII Viability, Growth and Fate
3.4.1.3.5.1 Cytotoxicity
Many in vitro studies have examined the potential for PFOS to cause cytotoxicity with various
cell viability assays in both primary hepatic cell cultures {Khansari, 2017, 3981272; Xu, 2019,
5381556} and in hepatic cell lines {Louisse, 2020, 6833626; Rosenmai, 2018, 4220319; Shan,
2013, 2850950; Sheng, 2018, 4199441; Bagley, 2017, 4238503; Wiels0e, 2015, 2533367;
Florentin, 2011, 2919235; Franco, 2020, 6315712; Ojo, 2020, 6333436; Franco, 2020, 6507465;
Huang, 2014, 2851292; Oh, 2017, 3981364; Wan, 2016, 3981504; Cui, 2015, 3981517; Behr,
2020, 6505973; Song, 2016, 9959776}, with varying results depending on the exposure time and
culturing methods. In mouse primary hepatocytes, cell viability was reduced by approximately
10% as determined by the CCK-8 assay after 24 hours of exposure to 10 |iM PFOS {Xu, 2019,
5381556} and by 64%, as determined by a trypan blue exclusion assay in rat primary
hepatocytes exposed to 25 |iM PFOS for 3 hours {Khansari, 2017, 3981272}. However, another
study in mouse and human primary hepatocytes reported that 100 |iM PFOS did not induce
3-59
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
cytotoxicity after 48 hours, determined by a lack of treatment effect in genes related to cell
damage such as heme oxygenase 1 (HMOX1), DNA damage inducible transcript 3 (1)1)113), and
activating transcription factor 3 (ATF3) {Rosen, 2013, 2919147}.
Median lethal concentration (LC50) values in hepatic cell lines ranged from approximately 13
|iM PFOS after for 24 or 48 hours of exposure in HepaRG cells {Franco, 2020, 6315712;
Franco, 2020, 6507465}, to 45-65 |iM after 24 or 48 hours of exposure in HepG2 cells {Wan,
2016, 3981504; Ojo, 2020, 6333436}, to 417 |iM after 24 hours of exposure in HL-7702 cells
{Sheng, 2018, 4199441}. However, two studies in HepG2 cells {Rosenmai, 2018, 4220319} and
HepaRG cells {Louisse, 2020, 6833626} showed no effect on cell viability up to concentrations
of 100 |iM for 24 hours or 400 |iM for 72 hours, respectively. A subset of these studies looked
further into the mechanisms of cytotoxicity, including the induction of apoptotic pathways
(Section 3.4.1.3.5.2.2).
3.4.1.3.5.2 Apoptosis
3.4.1.3.5.2.1 In Vivo Models
Apoptosis induced by PFOS exposure was assessed in five studies in male rats {Elcombe, 2012,
1332473; Elcombe, 2012, 1401466; Eke, 2017, 3981318; Wan, 2016, 3981504; Han, 2018,
4238554} and two studies in male mice {Xing, 2016, 3981506; Lv, 2018, 5080395}, with
varying results. Two short-term dietary studies exposed rats to 20 or 100 ppm PFOS (equivalent
to approximately 2 and 10 mg/kg/day, respectively), and apoptosis was assessed through the
TUNEL assay {Elcombe, 2012, 1332473; Elcombe 2012, 1401466}. In one of these studies, rats
were exposed for 7 days and allowed to recover for 1, 28, 56, or 84 days {Elcombe, 2012,
1332473}, while the other study exposed rats for 1, 7, or 28 days and collected liver directly after
exposure {Elcombe, 2012, 1401466}. In the recovery study, at both PFOS exposure
concentrations, a decreased apoptotic index was observed at all timepoints tested. In the 28-day
study, the apoptotic index was decreased with 100 ppm PFOS at days 7 and 28, and increased at
20 ppm on day 7; no changes were observed at other timepoints. It should be noted that cell
proliferation was markedly increased, particularly with the higher dose (100 ppm), in both
studies (Section 3.4.1.3.5.3); increases in the total number of cells due to cell proliferation may
confound certain metrics of apoptosis that do not report comparisons of the absolute number of
apoptotic cells along with cell percentages.
Contrary to the dietary studies, three short-term gavage studies in rats showed an increase in
expression of apoptotic genes (caspase 3 [Casp3] and caspase 8 [CaspH]) and proteins (e.g.,
cleaved poly-ADP-ribose polymerases [PARP], CASP3, and BCL2 associated X, apoptosis
regulator [Bax]) in livers collected after administrations of up to 10 mg/kg/day PFOS for 28 days
{Eke, 2017, 3981318; Wan, 2016, 3981504; Han, 2018, 4238554}. Similarly, two short-term
gavage studies in male mice showed an increase in liver apoptosis {Xing, 2016, 3981506; Lv,
2018, 5080395}. Increased apoptosis in the liver, as determined via the TUNEL assay, was
observed in male mice administered 2.5-10 mg/kg/day PFOS for 30 days {Xing, 2016,
3981506}. Increased apoptosis was also observed in liver tissue of male mice dosed with 10
mg/kg/day PFOS for 21 days, as measured by an increased expression of apoptotic-related
proteins (tumor suppressor p53 [p53] and BAX) and a corresponding decrease in B cell
leukemia/lymphoma 2 (BCL2) and by an increase in CASP3 enzyme activity {Lv, 2018,
5080395}.
3-60
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Several studies further examined the mechanisms by which PFOS exposure may lead to
apoptosis in the liver {Han, 2018, 4238554; Lv, 2018, 5080395; Xing, 2016, 3981506; Xu, 2020,
6316207; Oh, 2017, 3981364; Huang, 2014, 2851292; Yao, 2014, 2850398}. One rat study
suggested that hepatic apoptosis was induced through mitochondrial damage, as shown by an
increased level of cytoplasmic cytochrome c and decreased level of mitochondrial cytochrome c
{Han, 2018, 4238554}. Two mouse studies concluded that hepatic apoptosis was induced by
increases in oxidative stress, as evidenced by a decrease in antioxidant enzymes and a
corresponding increase in lipid peroxidation {Lv, 2018, 5080395; Xing, 2016, 3981506}. In a
third mouse study that examined microRNA (miRNA) expression in the liver, an increase in the
expression of miR-34a-5p, which has been shown to recapitulate p53-mediated apoptosis, was
observed {Yan, 2014, 2850901}.
3.4.1.3.5.2.2 In Vitro Models
In vitro, apoptosis has been examined in primary mouse hepatocytes and mouse and human cell
lines after exposure to various concentrations of PFOS {Xu, 2019, 5381556; Xu, 2020, 6316207;
Song, 2016, 9959776; Huang, 2014, 2851292; Oh, 2017, 3981364; Wan, 2016, 3981504; Cui,
2015, 3981517; Yao, 2016, 3981442}. PFOS was shown to increase the percentage of apoptotic
cells {Xu, 2019, 5381556; Huang, 2014, 2851292; Oh, 2017, 3981364; Cui, 2015, 3981517;
Yao, 2016, 3981442}, to increase the expression of proteins and genes in apoptotic pathways
{Song, 2016, 9959776; Wan, 2016, 3981504}, or to increase CASP3 enzyme activity {Yao,
2016, 3981442}. Only one study in HL-7702 cells showed no change in the percentage of
apoptotic cells {Cui, 2015, 3981568}.
In mouse primary hepatocytes, PFOS induced apoptosis through activation of Caspase 3, which
was mediated by PFOS-induced mitochondrial membrane damage and increased intracellular
calcium levels {Xu, 2020, 6316207}. One study in the Chang liver cell line suggested that
apoptosis following exposure to PFOS may be caused by endoplasmic reticulum stress, mediated
by the phosphorylation of extracellular signal-regulated protein kinases 1 and 2 (ERK1/2) {Oh,
2017, 3981364}. A study in human L-02 cells suggested that PFOS exposure may lead to
apoptosis through the activation of p53 and myc proto-oncogene (myc) pathways {Huang, 2014,
2851292}. In two studies in HepG2 cells, PFOS exposure led to increases in apoptosis and
alterations in autophagy, leading the authors to conclude that hepatotoxicity induced by PFOS
exposure may be at least partially attributed to autophagy-dependent apoptosis {Yao, 2014,
2850398; Yao, 2016, 3981442}.
No in vitro study directly evaluated cellular necrosis, although one RNA-sequencing study in
primary human hepatocytes found that PFOS exposure was associated with changes in gene
expression that aligned with cell death and hepatic system disease, including necrosis,
cholestasis, liver failure, and cancer {Beggs, 2016, 3981474}. Another RNA-sequencing study
showed that PFOS induced genetic changes in WT zebrafish that were comparable to those seen
in a zebrafish model of fatty liver disease; pathways involved in apoptosis of hepatocytes and
focal necrosis of liver were upregulated {Fai Tse, 2016, 3981456}.
3.4.1.3.5.3 Cell Cycle and Proliferation
3.4.1.3.5.3.1 In Vivo Models
Alterations in cell proliferation and the cell cycle were also seen in many in vivo and in vitro
studies {Thomford, 2002, 5029075; Elcombe, 2012, 1332473; Elcombe, 2012, 1401466; Han,
3-61
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
2018, 4355066; Huck, 2018, 5079648; Lai, 2017, 3981375; Beggs, 2016, 3981474; Cui, 2015,
3981568; Song, 2016, 9959776; Louisse, 2020, 6833626; Cui, 2015, 3981517}. Two short-term
studies in male rats with PFOS doses of 20 or 100 ppm (approximately 2 and 10 mg/kg/day,
respectively) found increased proliferation in the liver, as seen through increased BrdU staining,
which was accompanied by increased liver weights {Elcombe, 2012, 1332473; Elcombe, 2012,
1401466}. In a third study in male rats dosed with 1 or 10 mg/kg/day PFOS for 28 days,
proliferation in the liver was also observed, via an increase in the percentage of cells staining for
proliferating cell nuclear antigen (PCNA) and expression of proliferation-related proteins
(PCNA, c-JUN, c-MYC, and CCND1) {Han, 2018, 4355066}. Increased liver weight at 10
mg/kg/day was also observed. These results in short-term studies are in contrast to one chronic
dietary study in male and female rats which did not identify significant increases in cell
proliferation (as determined with PCNA or BrdU immunohistochemistry) after 4, 14, or 52
weeks of dietary PFOS administration {Thomford, 2002, 5029075}. However, the study authors
noted that a biologically significant and test-compound related mild increase in proliferation was
observed at week 4 in two out of five females in both of the highest dose groups. The biological
significance was defined as having twice the mean of the controls and being greater than that of
the highest control. Notably, this study did not use concentrations of PFOS greater than
approximately 1 mg/kg/day.
Similarly, in mice exposed to 10 mg/kg/day PFOS for 7 days, proliferation in the liver, as seen
through PCNA staining, was increased {Beggs, 2016, 3981474}; increased relative liver weights
were also observed. However, no changes in PCNA positive cells or PCNA protein expression
was observed in a second study in mice exposed to 1 mg/kg PFOS in their diet for 6 weeks
{Huck, 2018, 5079648}. Using RNAseq, one study examined the fetal livers of mice exposed
gestationally to 0.3 mg/kg/day PFOS and showed a positive association between PFOS exposure
and pathways involved in the alteration of liver cell and hepatocyte proliferation {Lai, 2017,
3981375}.
3.4.1.3.5.3.2 In Vitro Models
In one study in primary rat hepatocytes, increased proliferation, as seen by an increased
percentage of EdU-positive cells, was observed with PFOS exposures of 50 |ig/mL for 24 hours
{Han, 2018, 4355066}. A study in human HL-7702 cells found increased proliferation with 50-
200 |iM PFOS exposures for 48 or 96 hours using the MTT assay; they also reported an
association between PFOS exposure and proteomic changes that correlated with increased
proliferation {Cui, 2015, 3981568}. This same study found that approximately half of the
proteins changed with PFOS exposure were involved in the cell cycle. Using flow cytometry,
Cui et al. (2015, 3981568) further found that in HL-7702 cells, 50-200 |iM PFOS for 48 or 96
hours decreased the percentage of cells at the G1/G0 (non-dividing) phases of the cell cycle
while increasing the percentage of cells at the S phase (DNA synthesis); the percentage of cells
at G2/M phase (interphase growth/mitosis) was increased at the 100 |iM exposure after 48 hours
of exposure but was decreased at the 200 |iM exposure after 48 and 96 hours. Another study in a
zebrafish liver cell line (ZFL) also used flow cytometry to examine changes in the cell cycle after
PFOS exposure {Cui, 2015, 3981517}. In corroboration with the study in HL-7702 cells, PFOS
concentrations of 27.9 and 56.8 |ig/mL for 48 hours were shown to decrease the percentage of
cells at the G1/G0 phases while increasing the percentage of cells at G2/M and S phases. In
addition, two microarray studies in hepatic cell lines found that PFOS exposures ranging from
3-62
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
100-278 |iM for 24 or 48 hours were associated with pathways involved in the regulation of
cellular proliferation or the cell cycle {Song, 2016, 9959776; Louisse, 2020, 6833626}.
Several in vitro and in vivo studies mention pathways through which PFOS may be inducing
proliferation. The RNAseq study of fetal livers of mice exposed gestationally to 0.3 mg/kg/day
PFOS described above suggested that proliferation may be induced by PFOS activating RAC and
Wnt/p-catenin signaling pathways {Lai, 2017, 3981375}. Additionally, in two studies, PFOS has
been shown to decrease the expression of HNF4a {Behr, 2020, 6505973; Beggs, 2016,
3981474}, a regulator of hepatic differentiation and quiescence that has been suggested as a
mediator of steatosis following PFOS exposure {Armstrong, 2019, 6956799}. In one study by
Beggs et al. (2016, 3981474) (as described in Section 3.4.1.3.1.3), the authors concluded that
PFOS may be causing cellular proliferation by down-regulating positive targets of HNF4a,
including differentiation genes, and by inducing the expression of negative targets of HNF4a,
including pro-mitogenic genes such as CCND1 and protein levels of stem cell markers such as
NANOG, leading to hepatocyte de-differentiation.
3.4.1.3.5.4 Conclusions
Although some results were conflicting, there is generally strong evidence that PFOS exposure
can disrupt the balance between cell proliferation and cell death/apoptosis. Out of the multitude
of studies examining cell proliferation both in vivo and in vitro, only a single in vivo study
showed that PFOS did not alter hepatic cellular proliferation, with increased cell proliferation
observed in all other studies. Although most in vitro studies suggested that PFOS could induce
apoptosis, several in vivo studies showed that PFOS either did not alter or decreased apoptosis.
Disruption in cell cycle and the reduction of HNF4a were the most frequently cited mechanisms
of proliferation induced by PFOS. This increase in proliferation in the liver could be linked to
increased liver weights, steatosis, and cancer. Similarly, many pathways were implicated in
PFOS-mediated apoptosis, including mitochondrial dysfunction, endoplasmic reticulum stress,
and alterations in autophagy.
3.4.1.3.6Inflammation and Immune Response
The liver is an important buffer between the digestive system and systemic circulation and is
thus exposed to compounds that are potentially immunogenic that result in protective immune
and inflammatory responses. Kupffer cells constitute the majority of the liver-resident
macrophages and make up one third of the non-parenchymal cells in the liver. Kupffer cells
phagocytose particles, dead erythrocytes, and other cells from the liver sinusoids and play a key
role in preventing immunoreactive substances from portal circulation from entering systemic
circulation {Dixon, 2013, 10365841}. While Kupffer cells can be protective in drug- and toxin-
induced liver toxicity, dysregulation of Kupffer cell-mediated inflammatory responses is
associated with a range of liver diseases, including steatosis. Other liver-resident immune cells
include natural killer (NK) cells, invariant NKT cells, mucosal associated invariant T (MAIT)
cells, yST cells, and memory CD8+T cells {Wang et al., 2019, 10365737}. The non-immune
cells of the liver, liver sinusoidal endothelial cells (LSECs), hepatocytes, and stellate cells, also
participate in immunity. They can express pattern recognition receptors and present antigens to T
cells {Robinson, 2016, 10284350}. However, the impact of PFOS on the immune function of
these cell types has not been thoroughly investigated.
3-63
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.1.3.6.1 In Vivo and In Vitro Models
Investigations into the liver immune response has been reported in an epidemiological study in
the C8 Health Project cohort {Bassler, 2019, 5080624}, rat models {Han, 2018, 4355066; Han,
2018, 4238554}, mouse models {Lai, 2017, 3981375; Su, 2019, 5080481}, and in vitro models
{Han, 2018, 4355066; Song, 2016, 9959776}. Bassler et al. (2019, 5080624) collected 200
serum samples from participants of the C8 Health Project to analyze mechanistic biomarkers of
non-alcoholic fatty liver disease (NAFLD) and test the hypothesis that PFAS exposures are
associated with increased hepatocyte apoptosis and decreased pro-inflammatory cytokines. PFOS
levels were significantly correlated with decreases in serum levels of two pro-inflammatory
cytokines, tumor necrosis factor a (TNFa) and IL-8. The authors state that these results are
consistent with other findings that PFAS are immunotoxic and downregulate some aspects of the
immune responses, but paradoxically result in increased apoptosis, which may subsequently
result in progression of liver diseases including NAFLD.
In 6-week-old male Sprague Dawley rats gavaged with 0, 1, or 10 mg/kg/day PFOS for 28 days,
changes in immune related end points in the liver were measured through western blot, qRT-
PCR, histopathology, and ELISA {Han, 2018, 4355066; Han, 2018, 4238554}. In contrast to the
C8 Panel study in humans {Bassler, 2019, 5080624}, the authors reported dose-dependent
increases in both serum TNFa and hepatic Tnfa mRNA levels, indicating an increased pro-
inflammatory response to PFOS exposure. Likewise, in a histopathological analysis of the liver
of these PFOS-exposed animals, the authors noted intense inflammatory infiltrates in the
periportal area and an increase in inflammatory foci. Han et al. (2018, 4355066) also reported
increased TNFa in the free supernatant and Tnfa mRNA in primary Kupffer cells treated with
100 |iM PFOS for up to 48 hours. These increases were not linear over time; supernatant levels
and hepatic mRNA levels appeared to peak at 24 hours and 1 hour, respectively. Altered
supernatant TNFa concentrations were not observed in similarly treated primary hepatocytes.
Similar effects were also reported by Han et al. (2018, 4355066) for interleukin-6 (IL-6), which
is a contributor to inflammatory responses in cells. Dose-dependent increases in IL-6 levels were
observed in rat serum and increases in 11-6 mRNA were observed in rat liver tissue after the 28-
day in vivo exposure. The authors also reported increased IL-6 free supernatant concentrations
and mRNA levels in primary Kupffer cells treated with 100 |iM PFOS for up to 48 hours. In the
primary Kupffer cells, supernatant IL-6 levels and mRNA levels peaked at 1 and 6 hours of
treatment, respectively. No changes in IL-6 concentrations were observed in supernatant from
primary hepatocytes treated with 100 |iM PFOS for up to 48 hours. In activation/inhibition
assays targeting the c-JUN amino-terminal kinase (JNK), IkB, and nuclear factor-KB (NF-kB)
signaling pathways in Kupffer cells (all of which are associated with cellular stress and/or
immune/inflammatory responses) PFOS exposure induced JNK and IkB phosphorylation and
NF-kB activity. Han et al. (2018, 4355066) further reported partial mediation of the TNF-a and
IL-6 response in Kupffer cells co-treated with PFOS and either a NF-kB or JNK inhibitor,
indicating that these two pathways are at least partially responsible for hepatic inflammatory
responses to PFOS. In addition to cytokine levels, Han et al. (2018, 4355066) used the F4/80
antibody as a macrophage marker and found dose-dependent increases in F4/80+ cells of the
livers of rats treated with either 1 or 10 mg/kg/day PFOS for 28 days. The authors suggest that
the increase in hepatic macrophages may be a result of Kupffer cell activation.
In mice, the observed changes were similar to the rat data in that inflammatory markers and
pathways were upregulated with PFOS exposure. In one study conducted in male ICR mice,
3-64
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
TNFa and JL-6 were significantly increased in serum of mice treated with 10 mg/kg/day PFOS
for 21 days {Su, 2019, 5080481}. The authors also observed increased TNFa positive liver cells.
In prenatally exposed CD-I mouse offspring whose dams were treated with 0 or 0.3 mg/kg/day
PFOS the day after mating until embryonic day 18.5, there was an upregulation of inflammatory
pathways in the PFOS exposed fetuses {Lai, 2017, 3981375}. Using IP A, the authors identified
numerous inflammatory genes that were upregulated in the fetal liver tissue. KEGG pathway
analysis highlighted the deregulation of adipocytokines, pro-inflammatory cytokines produced
by adipocytes, and TGFP signaling. Interestingly, activation of TGFP is associated with anti-
inflammatory responses, immunosuppression, and tumor promoting pathways.
In another study investigating the hepatic effects of PFOS in vitro, Song et al. (2016, 9959776)
saw much of the same effects using human liver hepatocellular carcinoma line, HepG2. After
exposing these cells to 278 |iM PFOS (the IC20 dose) for 48 hours, through KEGG pathway
analyses, the authors reported that genes related to immune response were the fifth most
differentially expressed biological process out of the 189 processes with altered genetic profiles.
Within the immune response, 17 genes were differentially expressed, including those related to
the TNF signaling pathway, as well as genes involved in the KEGG pathways of nucleotide-
binding and oligomerization domain (NOD)-like receptor signaling, cytokine-cytokine receptor
interactions, and the complement and coagulation cascade system.
3.4.1.3.6.2 Conclusions
While there are not many studies investigating the immunotoxicity of PFOS specifically related
to the liver, evidence presented from various methods and biomarkers strongly indicate that
PFOS can disrupt normal hepatic immunological function. However, the immune response to
PFOS exposure in humans does not appear to be consistent with rodent and in vitro models.
While a single study in the C8 Health Project cohort suggests that immunosuppression may be
involved in the progression of NAFLD and potentially other types of liver disease, studies in
rats, mice, primary hepatic (Kupffer) cells, and immortalized cell lines suggest that pro-
inflammatory immune responses generally result from PFOS exposure. Specifically, there is
evidence that activation through the JNK/NF-kB pathways may stimulate the production of pro-
inflammatory cytokines such as TNFa and IL-6. Although further assessment of human
populations and in human cell lines may be needed to understand the differences in responses
between humans and laboratory models, both lines of evidence suggest PFOS exposure can alter
the hepatic immune and inflammatory responses.
3.4.1.3.70xidative Stress and Antioxidant Activity
3.4.1.3.7.1 Introduction
Oxidative stress, caused by an imbalance of reactive oxygen species (ROS) production and
detoxification processes, is a key part of several pathways, including inflammation, apoptosis,
mitochondrial function, and other cellular functions and responses. In the liver, oxidative stress
contributes to the progression and damage associated with chronic diseases, such as alcoholic
liver disease, non-alcoholic fatty liver disease, hepatic encephalopathy, and Hepatitis C viral
infection {Cichoz-Lach, 2014, 2996796}. Indicators of oxidative stress include but are not
limited to increased oxidative damage (e.g., malondialdehyde (MDA) formation); increased
reactive oxygen species (ROS) production (e.g., hydrogen peroxide and superoxide anion);
altered antioxidant enzyme levels or activity (e.g., superoxide dismutase (SOD) and catalase
3-65
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
(CAT) activity); changes in total antioxidant capacity (T-AOC); changes in antioxidant levels
(e.g., glutathione [GSH] and glutathione disulfide [GSSG] ratios); and changes in gene or protein
expression (e.g., nuclear factor-erythroid factor 2-related factor 2 [Nrf2] protein levels). PFOS
has been demonstrated to induce these indicators of oxidative stress, inflammation, and cell
damage.
3.4.1.3.7.2 In Vivo Models
Several studies in rats and mice assessed hepatic oxidative stress in response to PFOS exposure.
In male Sprague-Dawley rats, a positive association between markers of oxidative stress,
potentially due to decreased antioxidant capacity, and oral PFOS exposure (1 or 10 mg/kg/day of
for 28 days) was reported {Wan, 2016, 3981504; Han, 2018, 4238554}. In hepatocytes extracted
from dosed rats, Wan et al. (2016, 3981504) found decreased Nrf2 total protein levels and
decreased activated Nrf2 in the nuclei at 10 mg/kg/day PFOS. Nrf2 is known for its role as a
regulator of antioxidant response elements and is generally activated upon oxidant exposure.
Additionally, liver lysates from rats at the highest PFOS dose showed decreases in expression of
both heme oxygenase-1 (Hmoxl) and NAD(P)H quinone dehydrogenase 1 (Nqol) genes, both of
which are associated with antioxidant, anti-inflammatory, and/or stress responses, revealing an
inhibition of the Nrf2 signaling pathway following PFOS exposure. Results from Han et al.
(2018, 4238554) also provide evidence of increased hepatic oxidative stress following PFOS
exposure. PFOS-exposed rats had significant dose-dependent increases in ROS, as measured by
the 2,7-dichlorofluorescein diacetate (DCFDA) fluorescent probe, and significant increases in
hepatic inducible nitric oxide synthase (iNos) and Cyp2el mRNA expression, key producers of
oxidants in the cell. MDA levels, an indicator of lipid peroxidation, were also significantly
increased at both 1 and 10 mg/kg/day. Simultaneously, significant decreases were observed in
CAT and SOD activities in liver tissues. Antioxidants typically responsible for returning cells to
their homeostatic state were altered in the liver following PFOS exposure, including decreases in
GSH levels, increases in GSSG levels, and a decrease in the GSH/GSSG ratio. A decrease in this
ratio generally indicates an imbalance of the oxidation-reduction (redox) state of the cell.
Four additional studies examined indicators of oxidative stress in male mice {Rosen, 2010,
1274165; Liu, 2009, 757877; Xing, 2016, 3981506; Lv, 2018, 5080395}. Rosen et al. (2010,
1274165) found exposure to PFOS in mice downregulated genes associated with oxidative
phosphorylation. In their assessment of Kunming (KM) mice that were administered PFOS via
subcutaneous injection, Liu et al. (2009, 757877) found evidence of oxidative damage that
included decreased SOD activity in the male brain and female liver and decreased T-AOC in
male and female livers. Overall, oxidative damage was observed in younger offspring and was
slightly more evident among males. In a subchronic exposure study, evidence of increased
oxidative stress was observed among male C57BL/6 mice dosed once with 0, 2.5, 5, or 10
mg/kg/day PFOS via oral gavage for 30 days {Xing, 2016, 3981506}. Dose-dependent
reductions were observed for levels of the antioxidant enzymes SOD, CAT, and glutathione
peroxidase (GSH-Px) in the liver; the T-AOC (i.e., free radical scavenging capacity) was also
reduced in hepatic tissues, with the lowest capacity observed at the highest dose. Lipid
peroxidation reported as MDA levels were significantly increased in hepatic tissues of rats
exposed to PFOS. The highest MDA levels were observed in the highest dose group. Results
from the Lv et al. (2018, 5080395) subchronic exposure study also showed evidence of increased
oxidative stress and decreased mechanisms of defense against oxidative stress following PFOS
exposure {Lv, 2018, 5080395}. In an unspecified species of male mice, intragastric
3-66
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
administration of 10 mg/kg/day PFOS for three weeks resulted in significant increases in MDA
and hydrogen peroxide production and significant decreases in SOD activity and GSH levels in
the liver. Nrf2 protein expression was significantly decreased following PFOS exposure
compared to unexposed controls. Additionally, transcriptional levels of Sod, Cat, and Ho-1
mRNA were significantly decreased in the liver.
One gene expression compendium study aimed to examine the relationship between activation of
xenobiotic receptors, Nrf2, and oxidative stress by comparing the microarray profiles in mouse
livers (strain and species not specified) {Rooney, 2019, 6988236}. The study authors compiled
gene expression data from 163 chemical exposures found within Illumina's BaseSpace
Correlation Engine. Gene expression data for PFOS exposure was obtained from a previously
published paper by Rosen, et al., (2010, 1274165). In WT (129Sl/SvlmJ) male mice, Nrf2
activation was observed (as seen by increases in gene expression biomarkers) after a 7-day
exposure to 10 mg/kg/day PFOS via gavage. In Pppara-null mice, this activation was observed
at both the 3 and 10 mg/kg/day doses. CAR was similarly activated in these two strains of mice.
The authors proposed that CAR activation by chemical exposure (PFOS or otherwise) leads to
Nrf2 activation and that oxidative stress may be a mediator.
3.4.1.3.7.3 In Vitro Models
Several studies examined oxidative stress endpoints in hepatic primary cells {Khansari, 2017,
3981272; Rosen, 2013, 2919147; Xu, 2019, 5381556; Xu, 2020, 6316207}. Khansari et al.
(2017, 3981272) dosed rat hepatocytes with 25 [j,M PFOS for three hours and demonstrated
significantly increased production of ROS, measured with the DCFDA probe, and lipid
peroxidation, measured as thiobarbituric acid-reactive substances (TBARS) content, compared to
controls. Additionally, PFOS treatment resulted in increased damage of lysosomal membranes,
likely caused by lipid peroxidation and increased levels of ROS. The authors also noted that
PFOS treatment resulted in mitochondrial membrane potential collapse; disruptions in
mitochondrial membrane potential in itself may result in increased ROS production, which could
then create a positive feedback loop of further mitochondrial dysfunction and increased ROS.
The authors suggest that these results demonstrate a potential oxidative stress-related mechanism
underlying PFOS hepatoxicity.
Rosen et al. (2013, 2919147) assessed oxidative stress-related gene expression changes using
TaqMan low density arrays (TLDA) in both mouse and human primary hepatocytes exposed to
PFOS ranging from 0-250 |iM. PFOS exposure led to increases in the expression of the nitric
oxide synthase 2 (Nos2 or /Was) and Hmoxl genes in mouse primary hepatocytes. In human
primary hepatocytes exposed to 100 |iM PFOS, NOS2 expression decreased while HMOX1
expression increased.
Xu et al. (2019, 5381556) exposed primary hepatocytes from C57B1/6J male mice to 10, 100,
500, or 1,000 |iM PFOS for 24 hours. ROS levels, measured by a CM-H2DCFA fluorescent
probe, were significantly increased in cells exposed to the highest level of PFOS. Interestingly,
SOD activity was significantly increased in cells exposed to 500 and 1,000 |iM PFOS, up to
117% with 1,000 |iM, while CAT activity was reduced by 59% in cells at the highest dose level.
PFOS exposure also led to alterations in the structure of SOD, with PFOS exposure resulting in
an increased percentage of a-helix structures (26.9%) and a decreased percentage of P-sheet
structures (21.9%), providing evidence of polypeptide chain shortening. These structural changes
3-67
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
suggest that PFOS interacts directly with SOD. Alterations in the resonance light scattering
(RLS) measures further revealed the impact of PFOS exposure on SOD protein structures in that
protein aggregations were observed at low doses of PFOS, but the aggregations were destroyed
at higher doses of PFOS, leading to increased SOD activity. The authors suggest that this may
result from agglomerate dispersion following the destruction of the solvent shell on the surface
of SOD at high doses of PFOS or from protein collapse following PFOS binding. Additionally,
GSH content was increased by 199% in cells exposed to the highest dose level; the authors
suggest that increases in GSH may reflect cellular adaptations to oxidative stress and can lead to
detoxification of oxidized GSSGto GSH.
In a third study using primary mouse hepatocytes, Xu et al. (2020, 6316207) exposed cultured
cells to 10, 100, 500, or 1,000 |iM of PFOS for 24 hours to examine oxidative stress related cell
apoptosis. The authors examined the impact of PFOS exposure on endogenous levels of
lysozyme (LYZ), an enzyme that inhibits oxidative stress-induced damage, and demonstrated
that PFOS exposure impacted LYZ molecular structure, subsequently decreasing activity levels,
leading to oxidative stress-induced apoptosis. Decreases in peak intensity at 206 nm during
ultraviolet-visible (UV-vis) absorption spectrometry represented an unfolding of the LYZ
molecule following exposure to PFOS, which inhibited enzyme activity. At exposure levels of
100 |iM and above, LYZ enzyme activity decreased to 761% of control levels. Such an impact
on LYZ activity was deemed to be related to the high affinity of PFOS for key central binding
sites on the LYZ molecule.
Four additional studies examined oxidative stress endpoints following PFOS exposure in HepG2
cell lines {Wan, 2016, 3981504; Wiels0e, 2015, 2533367; Shan, 2013, 2850950; Florentin, 2011,
2919235}. Two studies reported increases in ROS levels following PFOS exposure {Wan, 2016,
3981504; Wiels0e, 2015, 2533367}, while two studies did not observe statistical differences in
ROS levels following 1- or 24-hour PFOS exposures up to 400 |iM {Florentin, 2011, 2919235}
or following 3-hour PFOS exposures up to 400 |iM {Shan, 2013, 2850950}. Wan et al. (2016,
3981504) dosed HepG2 cells with either 0, 10, 20, 30, 40, or 50 [xM PFOS for 24 hours or with
50 |iM PFOS for 1, 3, 6, 12, or 24 hours. ROS generation, analyzed using DCFH-DA, was
increased in a dose-dependent manner in cells dosed with 50 [xM across multiple time points,
with a peak in levels observed at 12 hours of exposure and a decrease in levels at 24 hours of
exposure; ROS production was significantly increased compared to control levels at 24 hours.
Significant decreases were observed in GSH and protein expression of total-Nrf2, HO-1, and
NQO-1 in a dose- and time-dependent manner. Expression of miR-155, a microRNA suspected
to play a key role in oxidative stress via the Nrf2 antioxidant pathway, increased nearly 12-fold
following 24-hour 50 |iM PFOS exposure. When cells were pre-treated with CAT prior to PFOS
exposure, ROS production was decreased along with miR-155 expression. SOD pre-treatment
did not lead to significant effects. Wan et al. (2016, 3981504) concluded that miR-155 plays a
key role in the inhibition of the Nrf2 signaling pathway and can be upregulated with PFOS
exposure.
Wiels0e et al. (2015, 2533367) incubated HepG2 cells with up to 200|xM PFOS to detect
changes in ROS, T-AOC, and DNA damage. PFOS exposure significantly increased ROS
production, as measured with the carboxy-H2DCFDA probe, as well as DNA damage, as
indicated by increased mean percent tail intensity in a comet assay, which is an indicator of DNA
strand breaks. Shan et al., 2013 exposed HepG2 cells to 100, 200, 300, or 400 [xM PFOS for 3
3-68
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
hours and found an increase in ROS generation with only 100 |iM PFOS, though the effect was
not statistically significant. Additionally, no changes were observed in the GSH/GSSG ratio.
3.4.1.3.7.4 Conclusions
Results from new studies published since the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}
further support the conclusions that implicate PFOS in inducing oxidative stress leading to
hepatocytic damage. Evidence of increased oxidative stress in the liver, including increased ROS
levels, changes in GSH and GSSG levels, and decreases in T-AOC, were observed following
both in vivo and in vitro exposures to PFOS. PFOS exposure was also associated with increased
levels of markers of oxidative damage and decreased activity or levels of protective antioxidants
that play a role in the reduction of oxidative damage. Interestingly, PFOS exposure appeared to
result in inhibition of the Nrf2 signaling pathway, with evidence of decreased Nrf2 protein levels
and reductions of the expression and activity of genes and proteins downstream of this
transcription factor. There was also evidence that PFOS can disrupt the structure and subsequent
function of crucial enzymes that mitigate ROS production and oxidative damage, SOD and LYZ.
While further research is needed to fully understand the mechanisms by which PFOS disrupts
oxidative stress responses, it is clear that PFOS induces oxidative stress in hepatic tissues.
3.4.1.4 Evidence Integration
There is moderate evidence for an association between PFOS exposure and hepatic effects in
humans based on associations with liver biomarkers, especially ALT, in several medium
confidence studies. Consistent with the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}, the
epidemiological data provide consistent evidence of a positive association between PFOS
exposure and ALT in adults. However, the associations were not large in magnitude, and it is
unclear whether the observed changes are clinically adverse. Evidence for other liver enzymes
and in children and adolescents is less consistent. Results for functional measures of liver
toxicity, specifically histology results, are mixed. There is some indication of higher risk of liver
disease with higher exposure, coherent with the liver enzyme findings, but there is inconsistency
for lobular inflammation among the two available studies, which decreases certainty. Among the
studies of ALT in adults, two presented correlations across PFAS {Nian, 2019, 5080307;
Salihovic, 2018, 5083555}; PFOA and PFOS were moderately correlated in both studies
(r = 0.4-0.5). Jin et al. (2020, 6315720), which reported positive associations with histology,
reported fairly low correlations between PFOS/PFOA (r = 0.14), which reduces the concern for
confounding in that population. It is not possible to rule out potential confounding across PFAS
with this evidence, but there is also no evidence that confounding can explain the observed
associations.
In summary, across studies in the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} and this
updated systematic review, there is generally consistent evidence of a positive association
between exposure to PFOS and ALT. However, one source of uncertainty in epidemiology
studies of PFAS is confounding across the PFAS as individuals are exposed to a mixture of
PFAS and it is difficult to disentangle the effects. This cannot be ruled out in this body of
evidence given the attenuation of the association in Lin et al. (2010, 1291111), the only general
population study that performed multi-pollutant modeling. The positive associations with ALT
are also supported by the recent meta-analysis of 25 studies in adolescents and adults {Costello,
2022, 10285082}. Associations for other hepatic outcomes were less consistent, including for
3-69
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
functional outcomes such as liver disease. Thus, while there is evidence of an association
between PFOS and ALT, there is residual uncertainty.
The animal evidence for an association between PFOS exposure and hepatic toxicity is robust
based on 20 high or medium confidence studies that show hepatic alterations. However, it is
important to distinguish between alterations that may be non-adverse (e.g., hepatocellular
hypertrophy alone) and those that indicate functional impairment or lesions {U.S. EPA, 2002,
625713; FDA, 2009, 6987952; EMEA, 2010, 3056796; Hall, 2012, 2718645}. EPA considers
responses such as increased relative liver weight and hepatocellular hypertrophy adverse when
accompanied by hepatotoxic effects such as necrosis, inflammation, or biologically significant
increases in enzymes indicative of liver toxicity {U.S. EPA, 2002, 625713}.
Multiple studies in mice and rats report increases in relative liver weights accompanied by
statistically significant increases in serum enzymes, though these increases were generally under
two-fold (100% change relative to control) as compared to control {Seacat, 2003, 1290852;
Curran, 2008, 757871; Butenhoff, 2012, 1276144; Xing, 2016, 3981506; Yan, 2014, 2850901;
NTP, 2019, 5400978; Han, 2018, 4355066}. However, these changes in serum enzyme levels
were accompanied by histopathological evidence of damage.
Of the four available animal toxicological studies with quantitative histopathological data, a
chronic study in rats {Butenhoff, 2012, 1271644} was the only study that identified dose-
dependent increases in hepatocellular hypertrophy, hepatocellular vacuolation, hepatocytic
necrosis, and inflammatory cell infiltration, though these effects were qualitatively reported in
other studies {Xing, 2016, 3981506; Han, 2018, 4355066; Cui, 2009, 757868}. A 28-day study
in male and female rats also reported dose-dependent increases in hepatocellular hypertrophy and
cytoplasmic alterations {NTP, 2019, 5400978}. A second short-term study in rats {Curran, 2008,
757871} only had a limited simple size of 4 rats/sex/treatment group, though there were apparent
dose-dependent increases in hypertrophy and cytoplasmic alterations in PFOS-exposed rats.
These two studies are supportive of the results observed by Butenhoff et al. (2012, 1271644).
Mechanistic data can contribute to the understanding toxicity in the context of relevance of data
collected from laboratory models in relation to observed human effects and the application of
such data in human hazard. There are several studies that have proposed potential underlying
mechanisms of the hepatotoxicity observed in rodents exposed to PFOS, some of which have
also been tested in human cells in vitro. Mechanistic evidence supports a role of nuclear
receptors, including the activation of PPARa and CAR and a decrease in HNF4a, in PFOS-
induced hepatotoxicity based on data collected in vivo in rodents and in vitro in both human and
rodent models. Findings support a role of these nuclear receptors in steatosis and hepatomegaly
observed in rodents in laboratory studies. However, it should be noted that although substantial
evidence exists demonstrating expression changes in gene targets of the nuclear receptors
PPARa, conflicting results have been reported for activation of the PPARa signaling pathway in
vitro between human and rodent cells, as well as across studies in different cells/cells lines from
the same species. Nonetheless, cells transfected with human PPARa demonstrated that PFOS can
increase PPAR activation. Gene expression signatures for CAR and PPAR activation has been
observed in mice exposed to PFOS, with CAR activation generally more significant in PPARa-
null mice, leading authors to conclude that CAR likely plays a subsequent role to PPARa in
mediating the adverse hepatic effects of PFOS. PPARa and CAR are known to play important
roles in liver homeostasis and have been implicated in liver dysfunction, including steatosis.
3-70
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Therefore, PFOS exposure may lead to liver toxicity through the activation of multiple nuclear
receptors in both rodents and humans.
HNF4a appears to play an important role in hepatotoxic effects related to PFOS exposure. PFOS
exposure led to a decrease in the protein expression of HNF4a in mice, which was associated
with an increase in relative liver weight. The in vivo alterations to HNF4a have been confirmed
by in vitro studies conducted in primary human hepatocytes and HepaRG cells, in which HNF4a
protein and gene expression was decreased. Importantly, increased cell proliferation in the liver
is related to reduction in HNF4a, both of which are reported effects of PFOS.
Regarding the cytotoxic potential of PFOS, results from in vitro exposure of both human and
rodent cells are variable and inconsistent in the concentrations at which PFOS causes
cytotoxicity, as well as whether or not PFOS is cytotoxic at any concentration tested in vitro.
Some studies evaluated mechanisms of the cell death, such as induction of apoptotic pathways,
with inconsistent results. In vivo, increases and decreases in apoptosis was observed in the livers
of mice, with variations related to duration of exposure, type of exposure (dietary or gavage), and
whether or not a recovery period was included in the study design. Oxidative stress, alterations to
p53 signaling, and mitochondrial damage have been reported in vivo in rodent studies as well as
in vitro in rodent cells; however, additional research is necessary to fully characterize the
involvement of such events in alterations to apoptotic signaling. While necrosis was not directly
evaluated, two transcriptomic analyses (one in primary human hepatocytes and one in zebrafish)
reported that PFOS induced changes in the expression of genes involved in liver necrosis and
damage. Increased hepatic cell proliferation has been more consistently reported in in vivo and in
vitro models, and is associated with increased liver weights and steatosis, which have also been
observed in rodents exposed to PFOS.
Inflammation and immunomodulation have also been reported in relation to PFOS, and
molecular-level alterations in inflammatory and immune response pathways can be linked to
inflammation observed in the livers of rodents exposed to PFOS. In rats, PFOS resulted in
increased serum TNFa and hepatic Tnfa gene expression, indicating an increased pro-
inflammatory response, which was accompanied by intense inflammatory infiltrates in the
periportal area and an increase in inflammatory foci. Decreased serum TNFa has been observed
in humans in relation to PFOS exposure, indicating that alterations to TNFa may have species
differences and/or be dependent upon exposure duration and dose. Alterations to inflammatory
response pathway genes have been reported in human cells in vitro (HepG2 cells), supporting the
observation in rodents that PFOS exposure leads to inflammatory response. Although further
assessment of human populations and human cell lines is needed to clarify the ability of PFOS to
induce inflammatory and immune responses in humans, the currently available evidence suggest
PFOS exposure can alter the hepatic immune and inflammatory responses.
3.4.1.4.1 Evidence Integration Judgment
Overall, considering the available evidence from human, animal, and mechanistic studies,
evidence indicates that PFOS exposure is likely to cause hepatotoxicity in humans under
relevant exposure circumstances (Table 3-5). This conclusion is based primarily on coherent
liver effects in animal models following exposure to doses as low as 0.02 mg/kg/day PFOS. The
available mechanistic information overall provide support for the biological plausibility of the
phenotypic effects observed in exposed animals as well as the activation of relevant molecular
3-71
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
and cellular pathways across human and animal models in support of the human relevance of the
animal findings. In human studies, there is generally consistent evidence of a positive association
with ALT, at median plasma PFOS levels as low as 0.57 ng/mL. Although a few associations
between other liver serum biomarkers and PFOS exposure were identified in medium confidence
epidemiological studies, there is considerable uncertainty in the results due to inconsistency
across studies.
3-72
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 3-5. Evidence Profile Table for PFOS Hepatic Effects
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
Evidence from Studies of Exposed Humans (Section 3.4.1.1)
Serum biomarkers of
hepatic injury
8 Medium confidence
studies
3 Low confidence
studies
In adults, significant
increases in ALT were
observed in medium
confidence studies (3/5).
Findings for AST and
GGT were similar to ALT,
indicating increased levels
of these enzymes,
however, some analyses
stratified by sex or weight
status (i.e., obesity) were
less consistent. Findings
for liver enzymes in
children were mixed.
Medium
confidence
studies
Consistent
direction of
effect for ALT
Coherence of
findings between
liver enzyme
increases
• Low confidence
studies
0©O
Moderate
Inconsistent
direction of effect
in children.
Liver disease or injury
3 Medium confidence
studies
Findings for markers of
liver inflammation were
mixed. In adults, one
study (1/2) reported
decreased odds of lobular
inflammation while
another study (1/2)
reported increased odds of
lobular inflammation and
non-alcoholic steatosis.
Results from the only
study in children were
imprecise.
Medium
confidence
studies
Evidence for hepatic
effects is based on
increases in ALT in adults.
Other supporting evidence
includes increases in other
liver enzymes such as
AST and GGT and
increased incidence of
liver disease mortality in
occupational settings.
Minor uncertainties
Limited number r,/ rcmam regarding mixed
studies examining liver cn/> mc flndln8s in
the outcome
Imprecision of
findings
children and coherence of
liver enzyme and albumin
findings.
Serum protein
2 Medium confidence
studies
1 Low confidence study
Three studies in adults
reported significantly
increased albumin (3/3).
For one study,
significance varied by
glomerular filtration rate
Medium
confidence
studies
Consistent
direction of
• Low confidence
study
• Limited number of
studies examining
the outcome
0©O
"Evidence Indicates (likely)
Primary basis and cross-
stream coherence:
Human data indicated
consistent evidence of
hepatoxicity as noted by
increased serum
biomarkers of hepatic
injury (primarily ALT)
with coherent results for
increased incidence of
hepatic nonneoplastic
lesions, increased liver
weight, and elevated serum
biomarkers of hepatic
injury in animal models.
Although a few
associations between other
serum biomarkers of
hepatic injury and PFOS
exposure were identified in
medium confidence
epidemiological studies,
there is considerable
uncertainty in the results
due to inconsistency across
studies.
Human relevance and
other inferences:
The available mechanistic
information overall provide
3-73
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
status. No studies were
conducted in children.
effect for
albumin
Serum iron
1 Medium confidence
study
Only one large cross-
sectional study examined
serum iron concentrations
and reported a significant
positive association.
• Medium
confidence
study
• Limited number of
studies examining
the outcome
Evidence from In Vivo Animal Toxicological Studies (Section 3.4.1.2)
Liver histopathology
2 High confidence
studies
5 Medium confidence
studies
Histopathologic^
alterations in liver were
reported in rodents or non-
human primates exposed
to PFOS for varying
durations (6/7).
Hepatocellular
hypertrophy was most
consistently (5/7)
observed across sex,
species, and duration of
exposure and in a dose-
responsive manner. Other
observed lesions included:
cystic or hepatocyte
degeneration (2/7), focal
or flake-like necrosis
(2/7), steatosis (1/7),
centrilobular or
cytoplasmic vacuolation
(6/7) and inflammatory
cellular infiltration into
liver tissue (4/7).
• High and
medium
confidence
studies
• Consistent
direction of
effects across
study design,
sex, and species
• Dose-dependent
response
• Coherence of
findings in other
endpoints
indicating liver
damage (i.e.,
increased serum
biomarkers and
liver weight)
• Large magnitude
of effect, with
some responses
reaching 100%
incidence in
some dose
No factors noted
©0©
Robust
Evidence is based on 20
high or medium
confidence animal
toxicological studies
indicating increased
incidence of hepatic
nonneoplastic lesions,
increased liver weight, and
elevated serum biomarkers
of hepatic injury.
However, it is important to
distinguish between
alterations that may be
non-adverse (e.g.,
hepatocellular hypertrophy
alone) and those that
indicate functional
impairment or lesions.
EPA considers responses
such as increased relative
liver weight and
hepatocellular hypertrophy
adverse when
support for the biological
plausibility of the
phenotypic effects
observed in exposed
animals as well as the
activation of relevant
molecular and cellular
pathways across human
.and animal models in
support of the human
relevance of the animal
findings.
3-74
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
groups (i.e.,
hypertrophy) or
are considered
severe (i.e., cell
or necrosis and
cystic
degeneration)
Liver weight
2 High confidence
studies
14 Medium confidence
studies
Liver weights were
increased in male and
female mice, rats, and
non-human primates at
higher doses across a
variety of study designs
including developmental,
short-term, subchronic,
and chronic (11/14). Liver
weight increases in pups
exposed in utero were also
observed (2/5).
High and
medium
confidence
studies
Consistent
direction of
effects across
study design,
sex, and species
Coherence of
effects with
other responses
indicating
increased liver
size (e.g.,
hepatocellular
hypertrophy)
Confounding
variables such as
decreases in body
weights
Serum biomarkers of
hepatic injury
3 High confidence
studies
7 Medium confidence
studies
ALT (7/7), AST (4/7),
ALP (3/4), and GGT (1/1)
levels were increased in
male adult rodents.
Measurements of ALT
(1/5), AST (0/5), and ALP
(1/2) in females found
little evidence that PFOS
exposure increased
enzyme levels. Several
studies found increased
• High and
medium
confidence
studies
• Consistent
direction of
effects across
study design,
sex, and species
Limited number of
studies examining
outcome
Inconsistent
direction of effects
between sex
Evidence Integration
Summary Judgment
accompanied by
hepatotoxic effects such as
necrosis and inflammation.
Many of the studies
discussed in this section
reported dose-dependent
increases in liver weight
"and hepatocellular
hypertrophy in rodents of
both sexes. However, a
limited number of these
studies additionally
examined functional or
histopathological hepatic
impairment to provide
evidence that the
enlargement of hepatic
tissue was an adverse, and
not adaptive, response.
3-75
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
bilirubin (3/3), albumin
(2/2), and
albumin/globulin ratio
(2/2) in male and female
animals, with an increase
in total protein in females
only (1/2), occurring
predominantly in high
dose groups only.
Increased concentrations
of bile salts/acids were
found in males (2/3) and
females (1/2).
• Dose-dependent
response
• Coherence of
findings with
other responses
indicating
hepatobiliary
damage (i.e.,
histopathological
lesions)
• Large magnitude
of effect, with
evidence of
biologically
significant
increases (i.e.,
>100% control
responses) in
serum liver
enzymes
indicating
adversity
Mechanistic Evidence and Supplemental Information (Section 3.4.1.3)
Biological Events or
Pathways
Summary of Key Findings, Interpretation, and Limitations
Evidence Stream
Judgement
Molecular initiating
events - PPARa
Key findings and interpretation:
Activation of PPARa in vivo in rodents and in vitro in human and rodent cells.
Increased expression of PPARa-target genes in vitro in rat and human
hepatocytes, and cells transfected with human PPARa.
Altered expression of genes involved in lipid metabolism and lipid
homeostasis.
Gene expression changes related to lipid metabolism were observed in both wildPPARa. The evidence also
type and PPARa-null mice. suggests a role for
Limitations: PPARq-independent
Overall, studies in rodent
and human in vitro models
and in vivo in rodent
studies suggest that PFOS
induces hepatic effects, at
least in part, through
3-76
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Conflicting results have been reported for activation of the PPARa signaling
pathway in vitro between human and rodent cells.
Molecular or cellular Key findings and interpretation:
initiating events - other Activation of CAR in vivo in rodents and in vitro in both human and rodent
pathways models.
Gene expression signatures for CAR activation observed in mice; more
significant in Ppara-null mice than in wild type mice.
Decrease in HNF4a protein expression, and changes in the expression of genes
regulated by HNF4a in vivo in mice.
Decrease in HNF4a gene and protein expression in vitro in human hepatocytes.
Reduction in HNF4a is associated with increased cell proliferation, which was
observed separately inPFOS-exposed animals.
Upregulation of PPARy, CAR/PXR, or LXR/RXR in mice.
Limitations:
Evidence is limited for some receptors, such as PPARy and LXR/RXR.
pathways in the MOA for
_noncancer liver effects of
PFOS, particularly CAR
activation and decreased
expression of HNF4a.
Evidence Integration
Summary Judgment
Notes: ALP = alkaline phosphatase; ALT = alanine aminotransferase; AST = aspartate aminotransferase; CAR = constitutive androstane receptor; GGT = gamma-glutamyl
transpeptidase; HNT4a = hepatocyte nuclear factor 4-alpha; LXR = liver X receptor; PPARa = peroxisome proliferator-activated receptor alpha; MOA = mode of action;
PPARy = peroxisome proliferator-activated receptor gamma; PXR = pregnane X receptor; RXR = retinoid X receptor.
3-77
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.2 immune
EPA identified 46 epidemiological and 13 animal toxicological studies that investigated the
association between PFOS and immune effects. Of the epidemiological studies, 2 were classified
as high confidence, 28 as medium confidence, 10 as low confidence, 5 as mixed (5 medium/low)
confidence, and 1 was considered uninformative (Section 3.4.2.1). Of the animal toxicological
studies, 1 was classified as high confidence, 9 as medium confidence, 1 as low confidence, and 2
were considered mixed {high/low and medium/low) (Section 3.4.2.2). Studies have mixed
confidence ratings if different endpoints evaluated within the study were assigned different
confidence ratings. Though low confidence studies are considered qualitatively in this section,
they were not considered quantitatively for the dose-response assessment (Section 4).
3.4.2.1 Human Evidence Study Quality Evaluation and Synthesis
3.4.2.1.1 Immunosuppression
Immune function—specifically immune system suppression—can affect numerous health
outcomes, including risk of common infectious diseases (e.g., colds, influenza, otitis media) and
some types of cancer. The WHO guidelines for immunotoxicity risk assessment recommend
measures of vaccine response as a measure of immune effects, with potentially important public
health implications {WHO, 2012, 9522548}.
There are 8 studies from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} that investigated the
association between PFOS and immune effects. Study quality evaluations for these 8 studies are
shown in Figure 3-16.
In the 2016 PFOS HESD, there was consistent evidence of an association between PFOS
exposure and immunosuppression in children. Two studies reported decreases in response to one
or more vaccines in relation to higher exposure to PFOS in children {Grandjean, 2012, 1248827;
Granum, 2013, 1937228}. In one study of adults, no association was observed {Looker, 2014,
2850913}. Antibody responses for diphtheria and tetanus in children (n = 587) were examined at
multiple timepoints in a study on a Faroese birth cohort {Grandjean, 2012, 1248827}. Prenatal
and age five serum PFOS concentrations were inversely associated with childhood diphtheria
antibody response at all measured timepoints, and the association was significant for anti-
diphtheria antibody concentrations pre-booster at age five and at age seven, modeled using
prenatal and age five serum PFOS concentrations, respectively. The antibody response for
tetanus was inversely associated with prenatal and age five serum PFOS concentrations but was
only significant for the association between age five serum PFOS concentrations and post-
booster anti-tetanus antibody concentrations. Prenatal PFOS exposure was associated with
diminished vaccine response in a different birth cohort study {Granum, 2013, 1937228, MoBa}.
Decreases in the anti-rubella antibody response were significantly associated with elevated
prenatal PFOS concentrations among 3-year-old children. No association was observed for the
only study {Looker, 2014, 2850913} in adults, examining influenza vaccine responses in a high-
exposure community (C8 Health Project).
3-78
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
,0^
&
I I
Dong et al„ 2013, 1937230-
+
++
+
+
++
+
+
+
Fei etal., 2010, 1290805-
+
++
+
+
++
+
+
+
Grandjean et al„ 2012, 1248827-
B
B
++
+
+
+
+
+
Granum etal., 2013, 1937228-
++
++
+*
+
+
-
+*
Humblet et al„ 2014, 2851240 -
++
B
B
+
++
+
+
+
Looker et al„ 2014, 2850913-
+
+
++
+
+
+
+
+
Okada et al„ 2012, 1332477-
+
+
+
+
+
+
+
+
Wang etal., 2011, 1424977-
-
+
+
+
+
+
+
+
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-16. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS
and Immune Effects
Interactive figure and additional study details available on HAWC.
There are 27 new studies from recent systematic literature search and review efforts conducted
after publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} that investigated the
association between PFOS and immunosuppression effects. Study quality evaluations for these
27 studies are shown in Figure 3-17 and Figure 3-18. One study from the 2016 PFOS HESD
{Grandjean, 2012, 1248827} was updated during this period, and the update was included in the
systematic review {Grandjean, 2017, 3858518}.
3-79
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
,C.®
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Abraham et al., 2020, 6506041 -
+
+
%
-
-
+
+
-
Ait Bamai et al., 2020, 6833636 -
+
+
+
+
++
+
+
+
Bulka et al., 2021, 7410156-
++
+
+
+
+
+
+
+
Dalsager et al., 2016, 3858505 -
-
++
-
+
+
+
-
-
Dalsager et al., 2021, 7405343 -
+
+
+
+
+
+
+
+
Goudarzi et al., 2017, 3859808 -
++
+
+
+
+
+
+
+
Grandjean et al., 2017, 3858518 -
+
++
++
+
+
+
+
++
Grandjean et al., 2017, 4239492-
+
++
++
-
+
+
+
+
Grandjean et al., 2020, 7403067 -
-
++
B
+
+
+
+
+
Huang et al., 2020, 6988475 -
+
+
+
+
+
+
+
+
Impinen et al., 2018, 4238440-
+
+
+
+
+
-
+*
Impinen et al., 2019, 5080609 -
++
++
-
+
++
+
-
-
Ji et al., 2021,7491706-
-
+
+
+
+
+
-
+
Figure 3-17. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Immunosuppression Effects
Interactive figure and additional study details available on HAWC.
3-80
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Kielsen etal., 2016, 4241223
Kvalem et al., 2020, 6316210
Lopez-Espinosa etal., 2021, 7751049
Manzano-Salgado et al., 2019, 5412076
Mogensen et al., 2015, 3981889
Pilkerton et al., 2018, 5080265
Shih etal., 2021, 9959487
Smitetal., 2015, 2823268
Stein et al., 2016, 3860111
Timmermann et al., 2020, 6833710
Timmermann et al., 2021, 9416315
Wang etal., 2022, 10176501
Zeng etal., 2019, 5081554
Zeng etal., 2020, 6315718
,e^
,ce
far
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-18. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Immunosuppression Effects (Continued)
Interactive figure and additional study details available on HAWC.
3-81
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.2.1.1.1 Vaccine Response
Twelve studies (thirteen publications10) studied the relationship between antibody response to
vaccination and PFOS exposure. Six of these studies investigated antibody response to
vaccination in children {Timmermann, 2020, 6833710; Abraham, 2020, 6506041; Grandjean,
2017, 3858518; Mogensen, 2015, 3981889; Grandjean, 2017, 4239492; Timmermann, 2021,
9416315}. In adults, two studies investigated antibody response to diphtheria and tetanus
{Kielsen, 2016, 4241223; Shih, 2021, 9959487}, one study investigated hepatitis vaccine
response {Shih, 2021, 9959487}, one study investigated adult flu vaccine response {Stein, 2016,
3860111}, and one study measured rubella antibodies in both adolescents (aged 12 and older)
and adults {Pilkerton, 2018, 5080265}. In addition, one study {Zeng, 2019, 5081554} measured
natural antibody exposure to hand, foot, and mouth disease (HFMD), and one study {Zeng,
2020, 6315718} measured hepatitis b antibodies in adults. Overall, one study was high
confidence { Grandjean, 2017, 3858518}, five studies were medium confidence {Grandjean,
2017, 4239492; Timmermann, 2020, 6833710; Mogensen, 2015, 3981889; Timmermann, 2021,
9416315; Shih, 2021, 9959487}, four were tow confidence {Stein, 2016, 3860111; Zeng, 2019,
5081554; Zeng, 2020, 6315718; Abraham, 2020, 6506041}, one was mixed (medium/low
confidence) {Pilkerton, 2018, 5080265}, and one was uninformative.
Of the studies that measured antibody response to vaccination in children, four studies were
cohorts {Timmermann, 2020, 6833710; Grandjean, 2017, 3858518; Grandjean, 2017, 4239492;
Mogensen, 2015, 3981889}, and two were cross-sectional {Abraham, 2020, 6506041;
Timmermann, 2021, 9416315} (maternal serum was available for a subset of participants in
Timmermann et al. (2021, 9416315)). These included multiple prospective birth cohorts in the
Faroe Islands, one with enrollment in 1997-2000 and subsequent follow-up to age 13
{Grandjean, 2017, 3858518} and one with enrollment in 2007-2009 and follow-up to age 5
{Grandjean, 2017, 4239492} (one additional cohort in the Faroe Islands examined outcomes in
adults with enrollment in 1986-1987 and follow-up to age 28 {Shih, 2021, 9959487}). Five of
these studies measured antibody response to tetanus vaccination {Abraham, 2020, 6506041;
Grandjean, 2017, 3858518; Grandjean, 2017, 4239492; Mogensen, 2015, 3981889;
Timmermann, 2021, 9416315}; the same studies also measured antibody response to diphtheria
vaccination; one study measured antibody response to measles vaccination {Timmermann, 2020,
6833710}, and one study to Haemophilus influenzae type b (Hib) vaccination {Abraham, 2020,
6506041}.
The results for this set of studies in children are shown in Table 3-6 and the Appendix (see PFOS
Appendix). The Faroe Islands studies {Grandjean, 2017, 3858518; Grandjean, 2017, 4239492;
Mogensen, 2015, 3981889} observed associations between higher levels of PFOS and lower
antibody levels against tetanus and diphtheria in children at 18 months, age 5 years (pre-and
post-booster), and at age 7 years, with some being statistically significant. These studies
measured exposure levels in maternal blood during the perinatal period and at later time periods
from children at age 5, 7, and 13 years (Table 3-6). No biological rationale has been identified as
to whether one particular time period or duration of exposure or outcome measurement is more
sensitive to an overall immune response to PFOS exposure.
10 Multiple publications of the same study: the study populations are the same in Grandjean et al. (2017, 3858518) and Mogensen
etal. (2015, 3981889).
3-82
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Effect Estimate
Percent charge (per dautrtr>g ol ^
PFOS 317 years median
Grand,can e< (25lh-75th perceotie)-15.3
al„ 2017a n^VnL (12.4 19.0 nptaiL)
PFOS ai 13 years median
(251h-75tti paroerrSope.7
nptaiL (5.2-8.5 09*1*.) 1
Grandiean et Age 5 PFOS Geometric Percent deference (per :'ojbirrj f|u|
at. 2012 rrearr=10.7 ng'mi. in age S PFOS)
<25th-75th
pereentfle=13.5-21.1
nglmL)
Exposure Outcome
Aff» Age EE
Age 7 Ago 13 30
Age 13 Age 13 22.2
far Age d Ab Age 5
Age 5 20.5
Pre boaster Age 5 Age 5
Matercai PFOS: Geometric Percent difference (per d<
rroan-27.3 ng'mL in maternal PFOS)
<25lh-75tn
percenlile^23J-33.1
npSmLJ
Adj for Age 5 Ab Pienala
Pte-boosler Pienalaf
Age7
Age 7
Age 5
Age 5
Ago 5
Median - 4.7 ngfrnL <2Slh _ . ,
75
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Effect Estimate
PFOS at 7 years median (2Slh-75th ~
percentile r 15.3 ngrtnL (12.4-19.0 a(PrOS^
PrOS at 13 years med-an.:25th 75#i
percentile 1^6.7 ng'mL (5.2 3.S
n^mL)
Age 5 PFOS- Geamatrc mean-16-7 Percent cSflerence (par
rtgMiL (2S£i 75th doubting in age 5 PFOS)
percentile^ 3.5-21.1 ngtaL)
Age >3 Age 13 8.6
Aii[ far Age d Ab Age 5
«t booster Age 5
Age 5 Age 5
smal PFOS.: Geometric
1^27.3 ngftnL (25th 75th
entile-232-33.1 agfenl.)
Percent difference "[per Kid Ptenafitf Age 7 -19.7
doubting in matornji PFOSl
Adj for Age S Ab Prenatal Age 7 10
Part-booster Prenalai Age 5 20.6
PrelKMStw Prenatal Age 5 38.6
Age 1JS Age 5 >2121
Age 5 Age 5 -16.02|
Cord blood Age 5 38.64
Cohort 3and5 Age 1J Age 5 15.07
Age 5 Age 5 -1.34
Age 5 Age 5 17.17
nt change {per daubing
PFOS Ages 7-12
«{par unit
mi PFOS Ages 7-12
Age 1.5 Age 5 17j56
Age 7 Age 7 -30.3
Age 7-12 Age 7-12 -9
Prenatal1 Age 7-12 1
Figure 3-20. Overall Diphtheria Antibody Levels in Children from Epidemiology Studies
Following Exposure to PFOS
Interactive figure and additional study details available on Tableau.
Grandjean et al., 2012 was reviewed as a part of the 2016 HESD
3-84
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
npVnL (2Slh - OR (per in it
75tfi percenllea: increase in
6.S2 - 12.23 PFOS)
Figure 3-21. Odds of Being Below the Protective Level Against Diphtheria (Antibody
Concentrations < 0.1 IU/mL) from Epidemiology Studies Following Exposure to PFOS
Interactive figure and additional study details available on Tableau.
It is plausible that the observed associations with PFOS exposure could be explained by
confounding across the PFAS, however, exposure levels to PFOS were higher than PFOA (PFOS
17 ng/mL, PFOA 4 ng/mL) in the Faroe Island studies. Though there was a moderately high
correlation between PFOS and PFOA, PFHxS, and PFNA (0.50, 0.57, 0.48, respectively), the
study authors assessed the possibility of confounding in a follow-up paper {Budtz-Jorgensen,
2018, 5083631} where estimates were adjusted for PFOA and there was no notable attenuation
of the observed effects. The other available studies did not perform multipollutant modeling.
Overall, the available evidence suggests that confounding across PFAS is unlikely to completely
explain the observed effects.
3-85
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 3-6. Associations between PFOS Exposure and Vaccine Response in Faroe Island Studies
Exposure
measurement
timing, levels
(ng/mL)a
Diphtheria Antibody Associations with PFOS by Age at
Assessment
Tetanus Antibody Associations with PFOS by Age at
Assessment
5 years
(Pre-Booster)
(C3 and/or C5)
7 years
(C3 only)
13 years
(C3 only)
5 years
(Pre-Booster)
(C3 and/or C5)
7 years
(C3 only)
13 years
(C3 only)
Maternal
| (C3; age, sex)b
| (C3; age, sex,
| (C3; age, sex)b
tt (C3; age, sex,
C3: GM: 27.3
booster type, and the
booster type, and the
(23.2-33.1)
BMD/BMDL
child's specific
BMD/BMDL
child's specific
(C3&5; sex, birth
antibody
(C3&5; sex, birth
antibody
cohort, log-PFOS)°
concentration at
cohort, log-PFOS)°
concentration at age
age 5 years)b
5 years)b
Birth
J, J, (C3; age, sex)d
-
| (C3; age, sex)d
-
(modeled)
(C3&5; age, sex)d
| (C3&5; age, sex)d
| (C5; age, sex)d
| (C5; age, sex)d
18 months
| (C3; age, sex)d
- -
| (C3; age, sex)d
- -
C3:NR
C5: 7.1 (4.5-
| (C3&5; age, sex)d
| (C3&5; age, sex)d
10.0)
| (C5; age, sex)d
| (C5; age, sex)d
5 years
|| (C3; age, sex)b
| (C3; age, sex,
| (C3; age, sex)b
| (C3; age, sex,
C3: GM: 16.7
booster type, and the
booster type, and the
(13.5-21.1)
| (C3; age, sex)d
child's specific
| (C3; age, sex)d
child's specific
C5: 4.7 (3.5-
antibody
antibody
6.3)
| (C3&5; age, sex)d
concentration at
| (C3&5; age, sex)d
concentration at age
age 5 years)b
5 years)b
| (C5; age, sex)d
| (C5; age, sex)d
BMD/BMDL (C3;
BMD/BMDL (C3;
sex, age, and booster
sex, age, and booster
type at age 5)e
type at age 5)e
BMD/BMDL (C3;
BMD/BMDL (C3;
sex, booster type at
sex, booster type at
age 5, log-PFOS)°
age 5, log-PFOS)°
3-86
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Exposure
Diphtheria Antibody Associations with PFOS by Age at
Assessment
Tetanus Antibody Associations with PFOS by Age at
Assessment
measurement
timing, levels
5 years
7 years
13 years
5 years
7 years
13 years
(ng/mL)a
(Pre-Booster)
(C3 only)
(C3 only)
(Pre-Booster)
(C3 only)
(C3 only)
(C3 and/or C5)
(C3 and/or C5)
7 years
J, J, (C3; age, sex,
(C3; sex, age at
| (C3; age, sex,
| (C3; sex, age at
C3: 15.3 (12.4-
booster type)f
antibody assessment,
booster type)f
antibody assessment,
19.0)
booster type at
booster type at age
| (C3; sex, age at
age 5)g
| (C3; sex, age at
5)g
antibody assessment,
antibody assessment,
booster type at
booster type at age
age 5)s
5)g
13 years
-
| (C3; sex, age at
-
| (C3; sex, age at
C3: 6.7 (5.2-
antibody assessment,
antibody assessment,
8.5)
booster type at
booster type at
age 5)g
age 5)g
Notes: C3 = cohort 3, born 1997-2000; C5 = cohort 5, born 2007-2009; GM = geometric mean; NR = not reported.
Arrows indicate direction of association with PFOS levels; double arrows indicate statistical significance (p < 0.05) where reported. Arrows are followed by parenthetical
information denoting the cohort(s) studied and confounders (factors the models presented adjusted for).
a Exposure levels reported from serum as median (25th-75th percentile) unless otherwise noted.
bGrandjean et al. (2012,1248827); medium confidence.
c Budtz-Jergensen and Grandjean (2018, 5083631); medium confidence.
dGrandjean et al. (2017,4239492); medium confidence.
e Grandjean and Budtz-Jergensen (2013,1937222); medium confidence.
fMogensen et al. (2015, 3981889); medium confidence.
g Grandjean et al. (2017, 3858518); medium confidence.
3-87
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
The cross-sectional study of these antibodies in Greenlandic children {Timmermann, 2021,
9416315} reported results that differed in direction of association based on the covariate set
selected. The exposure measurement in these analyses may not have represented an etiologically
relevant window; cross-sectional analyses in the Faroe Islands studies at similar ages also found
weaker associations than analyses for some other exposure windows. However, a subset of the
study population did have maternal samples available, and those results were null. On the other
hand, this study was the only one to examine the odds ratio for not being protected against
diphtheria (antibody concentrations, which has clear clinical significance, and they reported
elevated odds of not being protected (based on antibody concentrations <0.1 IU/mL, OR (95%
CI) per unit increase in exposure: 1.14 (1.04, 1.26)). Looking at other vaccines, Timmermann et
al. (2020, 6833710) also observed inverse associations between elevated levels of PFOS and
lower adjusted antibody levels against measles (statistically significant only in group with fewer
measles vaccinations). Lastly, the low confidence cross-sectional study at age one, Abraham et
al. (2020, 6506041), did not observe associations between adjusted tetanus, Hib, and diphtheria
antibody levels and PFOS concentrations.
Of the three studies that measured vaccine response in adults or adolescents, two were cohorts
{Stein, 2016, 3860111; Shih, 2021, 9959487}, and one was a cross-sectional analysis {Pilkerton,
2018, 5080265}. Shih et al. (2021, 9559487) measured exposure in cord blood and at multiple
points through childhood to early adulthood, with outcome measurement at age 28 years; this
study was medium confidence. Stein et al. (2016, 3860111) utilized a convenience sampling to
recruit participants, had low seroconversion rates, and was at high risk of residual confounding,
so was low confidence. The adult population in Pilkerton et al. (2018, 5080265) suffered from
potential exposure misclassification due to concurrent exposure and outcome measurements and
was also low confidence, but this was less of a concern for the adolescent participants so this
sub-population was rated as medium confidence for adolescence antibody response to
vaccinations. Shih et al. (2021, 9959487) reported inconsistent direction of associations across
exposure windows and vaccines (diphtheria, tetanus, Hepatitis A, Hepatitis B). Results also
differed by sex, but without a consistent direction (i.e., stronger associations were sometimes
observed in women and sometimes men). Similar to the results in 13-year-olds in the other Faroe
Island cohorts, this may indicate that by age 28, the effect of developmental exposure is less
relevant. Neither of the other studies reported associations with immunosuppression.
In addition to these studies of antibody response to vaccination, there are two studies that
examined antibody response to HFMD {Zeng, 2019, 5081554} and hepatitis B infection {Zeng,
2020, 6315718}. This birth cohort in China {Zeng, 2019, 5081554} measured antibody levels in
infants at birth and age 3 months, which represent passive immunity from maternal antibodies.
This study {Zeng, 2019, 5081554} was rated low confidence because the clinical significance of
the outcome is difficult to interpret in infants and there are concerns for confounding by timing
of HFMD infection as well as other limitations. Statistically significant increased odds of HFMD
antibody concentration below clinically protected levels per doubling of PFOS were observed.
This is coherent with the vaccine antibody results, but there is uncertainty due to study
deficiencies. Zeng et al. (2020, 6315718) observed negative associations between serum n-PFOS
concentration and hepatitis B surface antibody; however, there are study limitations due to
concurrent measurement of exposure and outcome and potential for reverse causality.
3-88
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
In a C8 Health project study, Lopez- Espinoza et al. (2021, 7751049) measured serum PFAS and
white blood cell types in 42,782 (2005-2006) and 526 (2010) adults from an area with PFOA
drinking water contamination in the Mid-Ohio Valley (USA). Generally positive monotonic
associations between total lymphocytes and PFOS were found in both surveys (difference range:
1.95-3.39% for count and 0.61-0.77 for percentage, per PFOS IQR increment). Significant
decreasing associations were observed neutrophils across the surveys and total white blood cell
count percent difference in the 2005-2006 survey. Findings were inconsistent for lymphocyte
subtypes.
3.4.2.1.1.2 Infectious Disease
Overall, ten studies (11 publications11) measured associations between PFOS exposure and
infectious diseases (or disease symptoms) in children with follow-ups between one and 16 years.
Infectious diseases measured included: common cold, lower respiratory tract infections,
respiratory syncytial virus (RSV), otitis media, pneumonia, chickenpox, varicella, bronchitis,
bronchiolitis, ear infections, gastric flu, urinary tract infections, and streptococcus. Of the studies
measuring associations between infectious disease and PFOS exposure, eight (nine publications)
were cohorts {AitBamai, 2020, 6833636; Dalsager, 2016, 3858505; Dalsager, 2021, 7405343;
Kvalem, 2020, 6316210; Manzano-Salgado, 2019, 5412076; Goudarzi, 2017, 3859808; Impinen,
2019, 5080609; Wang, 2022, 10176501; Huang, 2020, 6988475}, one was a case control study
nested in a cohort {Impinen, 2018, 4238440}, and one was a cross-sectional study {Abraham,
2020, 6506041}. Five studies measured PFOS concentrations from mothers during pregnancy
{AitBamai, 2020, 6833636; Dalsager, 2016, 3858505; Manzano-Salgado, 2019, 5412076;
Goudarzi, 2017, 3859808; Impinen, 2019, 5080609}. Impinen et al. (2018, 4238440) measured
PFOS concentrations from cord blood at delivery. Two studies measured PFOS concentrations in
children's serum at age one year {Abraham, 2020, 6506041} and at age 10 years {Kvalem, 2020,
6316210}.
Several of the studies measured infectious disease incidences as parental self-report, which may
have led to outcome misclassification {Kvalem, 2020, 6316210; Abraham, 2020, 6506041;
Impinen, 2018, 4238440; Impinen, 2019, 5080609}. Four studies measured infections as the
doctor-diagnosed incidence of disease over a particular period {Goudarzi, 2017, 3859808;
Manzano-Salgado, 2019, 5412076; AitBamai, 2020, 6833636; Huang, 2020, 6988475}, and
Wang et al. (2022, 10176501) used a combination of parental report and medical records. One
study used hospitalizations as an outcome, with events identified based on medical records
{Dalsager, 2021, 7405343}. Overall, seven studies were medium confidence {Abraham, 2020,
6506041; AitBamai, 2020, 6833636; Goudarzi, 2017, 3859808; Manzano-Salgado, 2019,
5412076; Dalsager, 2021, 7405343; Wang, 2022, 10176501; Huang, 2020; 6988475} and four
were low confidence {Dalsager, 2016, 3858505; Impinen, 2018, 4238440; Impinen, 2019,
5080609; Kvalem, 2020, 6316210}.
Increased incidence of some infectious diseases in relation to PFOS exposure was observed,
although results were not consistent across studies. Results from these studies are available in the
Appendix (see PFOS Appendix). The most commonly examined type of infections was
respiratory, including pneumonia/bronchitis, upper and lower respiratory tract, throat infections,
11 Multiple publications of the same study: both Dalsager et al. (2016, 3858505) and Dalsager et al. (2021, 7405343) use data
from the Odense cohort in Denmark and thus have overlapping, though not identical populations. They received different ratings
due to outcome ascertainment methods.
3-89
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
and common colds. Dalsager et al. (2021, 7405343), a medium confidence study, reported higher
rates of hospitalization for upper and lower respiratory tract infections with higher PFOS
exposure (statistically significant for lower respiratory tract). Among studies that examined
incidence, two studies (one medium and one low confidence) examining pneumonia/bronchitis
observed statistically significant associations between elevated PFOS concentration and
increased risk of developing pneumonia in 0- to 3-year-old children {Impinen, 2019, 5080609}
and 7-year-old children {AitBamai, 2020, 6833636}; however, two other medium confidence
studies did not report an increase in infections {Abraham, 2020, 6506041; Wang, 2022,
10176501}. Huang et al. (2020, 6988475) examined recurrent respiratory infections and found a
positive association with recurrent respiratory infections but not total infections. Two low and
one medium confidence studies found positive associations with lower respiratory infection
{Kvalem, 2020, 6316210; Impinen, 2018, 4238440; Dalsager, 2021, 7405343}, while another
medium confidence study reported no association {Manzano-Salgado, 2019, 5412076}. There
were also non statistically significant positive associations seen for PFOS in relation to
chickenpox {AitBamai, 2020, 6833636}, common cold {Wang, 2022, 10176501}, and cough
{Dalsager, 2016, 3858505}, but statistically significant inverse associations were observed for
RSV {AitBamai, 2020, 6833636} and common cold {Impinen, 2018, 4238440}. Outside of
respiratory infections, two medium confidence studies examined total infectious diseases.
Dalsager et al. (2021, 7405343) reported higher rates of hospitalization for any infections with
higher PFOS exposure (not statistically significant), while {Goudarzi, 2017, 3859808} reported
higher odds of total infectious diseases. Results for other infection types, including
gastrointestinal, generally did not indicate a positive association.
In addition to the studies in children, three studies examined infectious disease in adults, {Ji,
2021, 7491706; Grandjean, 2020, 7403067; Bulka, 2021, 7410156}. Results from these studies
are available in the Appendix (see PFOS Appendix). All three studies were medium confidence.
Ji et al. (2021, 7491706) was a case-control study of COVID-19 infection. They reported higher
odds of infection with higher exposure (OR (95% CI) per log-2 SD increase in PFOS: 1.94 (1.39,
2.96)). In contrast, a cross-sectional study examining severity of COVID-19 illness in Denmark
using biobank samples and national registry data Grandjean et al. (2020, 7403067) reported no
association between PFOS exposure and increased COVID-19 severity. Bulka et al. (2021,
7410156) used NHANES data from 1999-2016 in adolescents and adults and examined
immunoglobulin G (IgG) antibody levels to several persistent infections, including
cytomegalovirus, Epstein Barr virus, hepatitis C and E, herpes simplex 1 and 2, human
immunodeficiency virus (HIV), Toxoplasma gondii and Toxocara species. High levels of these
antibodies were interpreted as presence of a persistent infection. They found higher prevalence of
Herpes simplex viruses 1 and 2, Toxoplasma gondii and Toxocara species and total pathogen
burden with higher PFOS exposure in adults (not statistically significant for HSV-2 and
Toxoplasma gondii) but no association with other individual pathogens.
3.4.2.1.2Immune Hypersensitivity
Another major category of immune response is the evaluation of sensitization-related or allergic
responses resulting from exaggerated immune reactions (e.g., allergies or allergic asthma) to
foreign agents {IPCS, 2012, 1249755}. A chemical may be either a direct sensitizer (i.e.,
promote a specific immunoglobulin E (IgE)-mediated immune response to the chemical itself) or
may promote or exacerbate a hypersensitivity-related outcome without evoking a direct response.
For example, chemical exposure could promote a physiological response resulting in a
3-90
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
propensity for sensitization to other allergens (pet fur, dust, pollen etc.). Hypersensitivity
responses occur in two phases. The first phase, sensitization, is without symptoms, and it is
during this step that a specific interaction is developed with the sensitizing agent so that the
immune system is prepared to react to the next exposure. Once an individual or animal has been
sensitized, contact with that same (or, in some cases, a similar) agent leads to the second phase,
elicitation, and symptoms of allergic disease. Although these responses are mediated by
circulating factors such as T cells, IgE, and inflammatory cytokines, there are many health
effects associated with hypersensitivity and allergic response. Functional measures of sensitivity
and allergic response consist of health effects such as allergies or asthma and skin prick tests.
In the 2016 PFOS HESD, one of two studies reported higher odds of asthma with higher PFOS
exposure in children. A case-control study {Dong, 2013, 1937230} of children in Taiwan
reported an increased odds of asthma with increasing childhood PFOS exposure. The magnitude
of association was particularly large comparing each of the highest quartiles of exposure to the
lowest. In cross-sectional analyses of asthmatic children, the study authors reported monotonic
increases by quartile of exposure for IgE in serum, absolute eosinophil counts, eosinophilic
cationic protein, and asthma severity score. No association for current or ever asthma was
observed among NHANES (1999-2000, 2003-2008) adolescents {Humblet, 2014, 2851240}.
There are 23 studies from recent systematic literature search and review efforts conducted after
publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} that investigated the
association between PFOS and immune hypersensitivity (i.e., asthma, allergy, and eczema)
effects. Study quality evaluations for these 23 studies are shown in Figure 3-22.
3-91
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Ait Bamai et al., 2020, 6833636
Averina et al., 2019, 5080647 ¦
Impinen et al., 2019, 5080609
Jackson-Browne et al., 2020, 6833598
Manzano-Salgado et al., 2019, 5412076 ¦
Shen etal.,2022, 10176753
Workman et al., 2019, 5387046 ¦
+
+
+
+
++
+
+
+
+
3
+
+
+
-
+
+
+
-
+
+
+
-
+
~
+
++
++
B
+
++
+
+
+
+
+
+
+
+
+
+
+
+
-
+
+
+
+
+
+
++
+
+
+
+
++
+
+
+
+
+
•
+
a
+
4+
+
-
+
D
+
++
+
+
+
+
+
+
+*
+
+
+
+*
+
/A
+
+
+
+
+
+
+
-
+
+
-
+
+
++
++
++
+
++
+
+
+
+
+
+
+
+
+*
+
+
+
+
B
++
++
+
+
+
+
+
+
-
+
-
+
++
+
-
+
-
-
+
++
+
++
+
-
+
-
+
++
+
+
+
+
+
-
+
-
+
+
+
+
-
+
+
+
+
+
+
-
+
-
+
-
+
+
+
Legend
| Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
^ Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-22. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Immune Hypersensitivity Effects
Interactive figure and additional study details available on HAWC.
3-92
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Thirteen studies (fifteen publications)12 examined asthma (or asthma symptoms) and PFOS
exposure. Nine of these studies were cohorts {Averina, 2019, 5080647; Beck, 2019, 5922599;
Kvalem, 2020, 6316210; Manzano-Salgado, 2019, 5412076; Zeng, 2019, 5412431; Impinen,
2019, 5080609; Smit, 2015, 2823268; Timmermann, 2017, 3858497; Workman, 2019,
5387046}; three studies (five publications) were case-control investigations {Zhou, 2016,
3981296; Zhou, 2017, 3858488; Zhu, 2016, 3360105}, including one nested case-control,
{Gaylord, 2019, 5080201; Impinen, 2018, 4238440}; and one was a cross-sectional analysis
{Jackson-Browne, 2020, 6833598}. Seven studies measured the prevalence of "current" asthma
for at least one time point {Averina, 2019, 5080647; Beck, 2019, 5922599; Manzano-Salgado,
2019, 5412076; Kvalem, 2020, 6316210; Impinen, 2018, 4238440; Impinen, 2019, 5080609;
Zeng, 2019, 5412431}. Eight studies measured "ever" asthma for at least one time point
{Averina, 2019, 5080647; Manzano-Salgado, 2019, 5412076; Jackson-Browne, 2020, 6833598;
Gaylord, 2019, 5080201; Impinen, 2018, 4238440; Impinen, 2019, 5080609; Smit, 2015,
2823268; Timmermann, 2017, 3858497}. Incident or recurrent wheeze was examined in one
study {Workman, 2019, 5387046}. Overall, nine studies were rated medium confidence, and six
studies were low confidence for asthma (Figure 3-22). Timmermann et al. (2017, 3858497) was
low confidence for asthma because the questionnaire used to ascertain status was not validated.
Averina et al. (2019, 5080647) was considered low confidence because results were not provided
quantitatively. Studies from the Genetic and Biomarkers study for Childhood Asthma (GBCA)
{Zhou, 2016, 3981296; Zhou, 2017, 3858488; Zhu, 2016, 3360105} were considered low
confidence based on participant selection. Cases and controls were recruited from different
catchment areas, and the resulting differences between cases and controls indicated potential for
residual confounding by age. Additionally, the timing of exposure assessment in relation to
outcome assessment was unclear, and it was not reported whether outcome status was confirmed
in controls.
Results across these studies were inconsistent (see PFOS Appendix). Several studies observed
positive associations with ORs greater than 1.2 between PFOS concentration levels and
increased "current" or "ever" asthma {Beck, 2019, 5922599; Timmermann, 2017, 3858497;
Jackson-Browne, 2020, 6833598; Zeng, 2019, 5412431; Impinen, 2018, 4238440; Averina,
2019, 5080647}, but often only within population subgroups. Averina et al. (2019, 5080647)
observed statistically significant increased odds of self-reported doctor diagnosed asthma among
adolescents in their first year of high school. Jackson-Browne et al. (2020, 6833598) reported
statistically significant increased odds of "ever" asthma from increased PFOS concentrations in
children aged 3 to 5 years. No association was observed at ages 6-11 years, and the overall
association was small (OR: 1.1). Beck et al. (2019, 5922599) observed increased odds of self-
reported asthma per PFOS increase in boys (p > 0.05), but this was not observed in girls. For
doctor diagnosed asthma in the same study, an inverse association (p > 0.05) was observed in
boys and a positive association (p > 0.05) was observed in girls. Zeng et al. (2019, 5412431)
observed a positive association in boys and an inverse association in girls (both p > 0.05).
Impinen et al. (2018, 4238440) reported higher odds of ever asthma. The low confidence study,
Timmermann et al. (2017, 3858497), observed positive associations (p > 0.05) between increased
asthma odds and elevated PFOS concentrations in small subset of children aged 5 and 13 who
did not receive their measles, mumps, and rubella (MMR) vaccination before age 5. However, in
12 Three publications {Zhou, 2016, 3981296; Zhou, 2017, 3858488; Zhu, 2016, 3360105} reported on the same cohort (Genetic
and Biomarker study for Childhood Asthma) and outcome and are considered one study.
3-93
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
children of the same ages who had received their MMR vaccination before age 5, no association
was observed. Low confidence studies from the GBCA study {Zhou, 2016, 3981296; Zhou,
2017, 3858488; Zhu, 2016, 3360105} observed elevated PFOS levels (p = 0.002) in children
with asthma compared to those without {Zhou, 2016, 3981296}, and the odds of current asthma
was also found to be elevated among boys and girls with increasing PFOS exposure {Zhu, 2016,
3360105}. One other study {Impinen, 2019, 5080609} observed a small positive association
(OR: 1.1) with current asthma in boys only. Two studies reported non-significant inverse
associations with asthma {Manzano-Salgado, 2019, 5412076; Smit, 2015, 2823268}, and in one
study, all results were non-significant {Gaylord, 2019, 5080201}. One low confidence study did
not observe a significant effect for recurrent wheeze {Workman, 2019, 5387046}.
In addition to the studies of asthma in children, one medium confidence study {Xu, 2020,
6988472} using data from NHANES examined fractional exhaled nitric oxide (FeNO), a
measure of airway inflammation, in adults. Among participants without current asthma, this
study found higher FeNO levels with higher PFOS exposure, indicating greater inflammation
(percent change (95% CI) for tertiles vs. Tl, T2: 1.80 (-1.53, 5.25); T3: 5.02 (1.40, 8.77)).
Seven studies observed associations between PFOS exposure and allergies, specifically allergic
rhinitis or rhinoconjunctivitis, skin prick test, and food or inhaled allergies. Five of these studies
were cohorts {Goudarzi, 2016, 3859523; AitBamai, 2020, 6833636; Kvalem, 2020, 6316210;
Impinen, 2019, 5080609; Timmermann, 2017, 3858497}, one study was a case-control analysis
{Impinen, 2018, 4238440}, and one study was a cross-sectional study using data from NHANES
2005-2006 and 2007-2010 {Buser, 2016, 3859834}. All studies were considered medium
confidence for allergy outcomes. Results for these outcomes are presented in the Appendix (see
PFOS Appendix).
Three studies conducted skin prick tests on participants to determine allergy sensitization at age
10 years {Kvalem, 2020, 6316210; Impinen, 2018, 4238440}, at age 13 years {Timmermann,
2017, 3858497}, and at age 16 years {Kvalem, 2020, 6316210}. Skin prick tests were conducted
to test sensitization to dust mites, pets, grass, trees and mugwort pollens and molds, cow's milk,
wheat, peanuts, and cod. Results were inconsistent across studies. Kvalem et al. (2020, 6316210)
reported a statistically significant but small association (OR: 1.09) with a positive skin prick test
at age 16 years (results were similar at age 10 years but p > 0.05). Timmermann et al. (2017,
3858497) also reported a positive association (p > 0.05) in children who had received an MMR
before age 5 years, but an inverse association in those who had not received an MMR, and
Impinen et al. (2018, 4238440) reported an inverse association (p > 0.05). Five studies measured
symptoms of "current" or "ever" allergic rhinitis or rhinoconjunctivitis {Goudarzi, 2016,
3859523; AitBamai, 2020, 6833636; Impinen, 2018, 4238440; Kvalem, 2020, 6316210;
Timmermann, 2017, 3858497}, and 16 years old {Kvalem, 2020, 6316210}. Rhinitis was
defined as at least one symptom of runny or blocked nose or sneezing. Rhinoconjunctivitis was
defined as having symptoms of rhinitis, in addition to itchy and watery eyes. Results were null
for these outcomes in all five studies. Impinen et al. (2019, 5080609) measured parent-reported,
doctor-diagnosed "current" or "ever" allergy symptoms at 7 years old, in addition to known food
and inhaled allergies and reported higher odds of "ever" inhaled allergies (p > 0.05) but no
associations with food allergies or "current" inhaled allergies. Buser et al. (2016, 3859834)
measured food sensitization (defined as having at least 1 food-specific serum IgE > 0.35 kU/L)
3-94
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
and self-reported food allergies and reported statistically significant positive associations with
self-reported food allergies in NHANES 2007-2010 but not in in NHANES 2005-2006.
Seven studies measured the association between PFOS concentration and eczema (described by
some authors as atopic dermatitis). Six of these studies were cohorts {Goudarzi, 2016, 3859523;
Wen, 2019, 5387152; Wen, 2019, 5081172; Manzano-Salgado, 2019, 5412076; Chen, 2018,
4238372; Timmermann, 2017, 3858497}, and one was a case-control analysis {Impinen, 2018,
4238440}. Four studies measured PFOS concentrations in cord blood at delivery {Wen, 2019,
5387152; Wen, 2019, 5081172; Chen, 2018, 4238372; Impinen, 2018, 4238440}, three studies
measured PFOS concentrations in pregnancy {Goudarzi, 2016, 3859523; Manzano-Salgado,
2019, 5412076; Timmermann, 2017, 3858497}, and one study measured child blood at age 5 and
13 years {Timmermann, 2017, 3858497}. All the studies were considered medium confidence
for eczema. Results are presented in the Appendix (see PFOS Appendix).
Positive associations (p > 0.05) with eczema were observed in two studies (three publications)
{Wen, 2019, 5387152; Wen, 2019, 5081172; Chen, 2018, 4238372}, as well as a small positive
association at age 0-2 years in Impinen et al. (2018, 4238440). However, inverse associations
(p > 0.05) were reported in Manzano-Salgado et al. (2019, 5412076), Timmermann et al. (2017,
3858497), Goudarzi et al. (2016, 3859523), and at age 10 years in Impinen et al. (2018,
4238440).
One medium confidence nested case-control study examined chronic spontaneous urticaria
{Shen, 2022, 10176753}. They found no association between PFOS exposure and case status.
3.4.2.1.3Autoimmune Disease
Autoimmunity and autoimmune disease arise from immune responses against endogenously
produced molecules. The mechanisms of autoimmune response rely on the same innate and
adaptive immune functions responding to foreign antigens: inflammatory mediators, activation
of T lymphocytes, or the production of antibodies for self-antigens {IPCS, 2012, 1249755}.
Chemical exposures that induce immune response or immunosuppression may initiate or
exacerbate autoimmune conditions through the same functions. Autoimmune conditions can
affect specific systems in the body, such as the nervous system (e.g., multiple sclerosis (MS)), or
the effects can be diffuse, resulting in inflammatory responses throughout the body (e.g., lupus).
The 2016 PFOS HESD did not identify epidemiological evidence examining the association
between PFOS exposure and autoimmune conditions. There are 4 studies from recent systematic
literature search and review efforts conducted after publication of the 2016 PFOS HESD {U.S.
EPA, 2016, 3603365} that investigated the association between PFOS and autoimmune disease
effects. Study quality evaluations for these 4 studies are shown in Figure 3-23.
Four case-control studies examined PFOS exposure and autoimmune diseases (Figure 3-23).
Two studies examined MS {Ammitzb0ll, 2019, 5080379} and ulcerative colitis {Steenland,
2018, 5079806} in adults, and two studies examined celiac disease in children {Sinisalu, 2020,
7211554} and young adults {Gaylord, 2020, 6833754}. PFOS was measured in blood
components (i.e., blood, plasma, or serum) for all studies (see PFOS Appendix). One study was
medium confidence {Gaylord, 2020, 6833754} with minimal deficiencies, and three studies were
considered low confidence {Ammitzb0ll, 2019, 5080379; Steenland, 2018, 5079806; Sinisalu,
2020, 7211554}. Information on participant selection, particularly control selection, was not
3-95
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
reported in Ammitzboll et al. (2019, 5080379). Additionally, PFOS was evaluated as a
dependent rather than independent variable, making no informative determinations about
associations between PFOS exposure and risk of MS, and contributed to a low confidence rating.
Steenland et al. (2018, 5079806) examined exposure concentrations one to two years after
diagnosis of celiac disease, resulting in some concern for reverse causation. Additionally, there
was potential for residual confounding by SES which was not considered in the analysis. These
factors together contributed to a low confidence rating.
//////>''/
,o®
Ammitzb0ll et al., 2019, 5080379 -
-
+
+
+
+
+
+
-
Gaylord et al., 2020, 6833754 -
+
+
++
+
+
+
-
+
Sinisalu etal., 2020, 7211554-
+
+
+
-
+
+
-
-
Steenland et al., 2018, 5079806 -
+
-
+
-
+
+
+
-
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-23. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Autoimmune Effects
Interactive figure and additional study details available on HAWC.
Ammitzboll et al. (2019, 5080379) observed lower PFOS concentrations among healthy controls
compared to those with MS. Serum PFOS concentrations were 17% lower (95% CI: -27%, -6%;
p = 0.004) in healthy controls compared to cases of relapsing remitting MS and clinically
isolated MS. Restricting the analysis to men, serum PFOS levels were 28% lower (95% CI:
-32%, -3%; p = 0.023) in healthy controls compared to cases. The result was similar among
women but did not reach significance (p = 0.093).
In children and young adults, the odds of celiac disease were elevated but not significantly
{Gaylord, 2020, 6833754}. However, the effect was much stronger in females only (OR: 12.8;
95% CI: 1.17, 141; p < 0.05). A marginally significant (p = 0.06) decrease in serum PFOS was
3-96
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
observed among adult cases of ulcerative colitis compared to healthy controls {Steenland, 2018,
5079806}.
In the prospective observational Finnish Diabetes Prediction and Prevention (DIPP) study in
which children genetically at risk to develop type 1 diabetes (T1D) and celiac disease (CD) were
followed from birth, with blood samples taken at birth and 3 months of age {Sinisalu, 2020,
7211554}, there was no significant difference in the levels of PFOS exposure in those children
that later developed CD, which may be due to the small sample size, but age at diagnosis of CD
was strongly associated with the PFOS exposure.
Overall, the associations between PFOS exposure and autoimmune disease were very limited and
mostly null, with one study with evidence of elevated odds of celiac disease. Two studies
observed that PFOS levels in healthy controls were either higher than UC cases {Steenland,
2018, 5079806} or lower than in MS cases {Ammitzb0ll, 2019, 5080379}.
3.4.2.2 Animal Evidence Study Quality Evaluation and Synthesis
There are 3 studies from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} and 10 studies from
recent systematic literature search and review efforts conducted after publication of the 2016
PFOS HESD that investigated the association between PFOS and hepatic effects. Study quality
evaluations for these 13 studies are shown in Figure 3-24.
3-97
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Curran etal., 2008, 757871 -
++
NR
NR
Dong etal., 2011, 1424949-
++
+
NR
Han etal., 2018, 4355066-
++
+
NR
Lefebvre et al., 2008, 1276155 -
+
NR
NR
Li etal., 2021, 7643501 -
+
+
NR
Lv et al., 2015, 3981558-
+
NR
NTP, 2019, 5400978-
++
++
Seacat et al„ 2002, 757853
Seacat et al„ 2003, 1290852
tr
NR
NR
Thomford, 2002, 5432419-
~
-
NR
Xing etal., 2016, 3981506-
++
+
NR
Yang et al., 2021, 7643494-
++
+
NR
Zhong et al., 2016, 3748828-
-
NR
NR
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
9 Critically deficient (metric) or Uninformative (overall)
NR Not reported
~ Multiple judgments exist
++
++
++
+ +
++
Figure 3-24. Summary of Study Evaluation for Toxicology Studies of PFOS and Immune
Effects"
Interactive figure and additional study details available on HAWC.
a Lefebvre et al. (2008, 1276155) reported on the same animals as Curran et al. (2008, 757871).
The immune system could be a target of PFOS toxicity as effects have been observed across
animal toxicological studies of varying durations of oral exposure to PFOS. Effects include
3-98
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
changes in spleen and/or thymus weights, extramedullary hematopoiesis, perturbations in activity
level or composition of various immune cell populations, and diminished ability to generate an
immune response. Studies indicate that PFOS exposure may result in dose- and sex-specific
immunomodulatory effects.
3.4.2.2.lOrgan Weight
Several rodent studies have reported changes in thymus and/or spleen weights following oral
exposure to PFOS.
3.4.2.2.1.1 Spleen
Two separate 28-day studies reported absolute and relative spleen weights in male and female
rats exposed to PFOS. Lefebvre et al. (2008, 1276155) observed reduced absolute spleen weights
in male rats of the highest exposure group in Sprague-Dawley rats given PFOS in diet (0.14-
6.34 mg/kg/day in males and 0.15-7.58 mg/kg/day in females). When expressed as percent body
weight, these changes were not significant and were within 5% of control for any given exposed
group. In contrast, absolute spleen weights were not affected by PFOS exposure in females, but
relative spleen weights were significantly higher (18% higher than controls) in the highest
exposure group. The increased relative spleen weights in females may be explained by lower
body weights of the two highest exposure groups. Another 28-day study by NTP (2019,
5400978) administered PFOS (0.312, 0.625, 1.25, 2.5, or 5 mg/kg/day) to Sprague-Dawley rats
for 28 days and observed dose dependent reductions in absolute spleen weights at
1.25 mg/kg/day and higher in males only; no effects were observed in females. Spleen weights
relative to body weight were not significantly reduced in either sex. While body weights were
not significantly different throughout treatment, the high-dose group tended to have lower body
weight with a significant, but < 10%, difference from the control. Therefore, differences in body
weight cannot explain the decreased absolute weight.
In four separate studies, male C57BL/6 mice were administered 5, 20, or 40 mg/kg/day PFOS for
7 days {Zheng, 2009, 1429960}, fed chow with 0.001, 0.005, or 0.02% PFOS (equivalent to
-40 mg/kg/day) for 10 days {Qazi, 2009, 1937260}, 0.008-2.083 mg/kg/day PFOS for 60 days
{Dong, 2009, 1424951}, or administered 0.008-0.833 mg/kg/day PFOS for 60 days via
gavage{Dong, 2011, 1424949}. Decreased absolute and relative splenic weights tended to be
observed only at the highest doses for each study. Female mice were not assessed. These
findings are complimented by Xing et al. (2016, 3981506), where a reduction in relative spleen
weight was observed in male C57BL/6J mice following exposure to 10 mg/kg/day PFOS for
30 days via gavage. No effects were observed at other doses (2.5 and 5 mg/kg/day) {Xing, 2016,
3981506}.
In a developmental study, spleens were weighed in 4- and 8-week-old offspring of pregnant
C57BL/6 mice given 0, 0.1, 1, or 5 mg/kg/day PFOS from GD 1-17 via gavage. Relative spleen
weights were reduced in male pups from the 5 mg/kg/day exposure group at four-weeks. No
significant effects were observed in lower dose groups, at the 8-week time point, or in females
{Zhong, 2016, 3748828}.
In three separate mouse studies, spleen weights were not significantly altered following short-
term exposure to PFOS, including a study of male and female B6C3F1 mice administered
0.00017-0.166 mg/kg/day PFOS for 28 days {Peden-Adams, 2008, 1424797}, male C57BL/6
3-99
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
mice exposed to 0.25 or 2.5 mg/kg/day PFOS for 28 days {Yang, 2021, 7643494}, and male
C57BL/6 (H-2b) mice administered 0.005% PFOS in the diet for 10 days {Qazi, 2010, 1276154}.
Similarly, relative spleen weight in male BALB/c mice was not affected at the end of a three-
week exposure to 2.5-5 mg/kg/day PFOS {Lv, 2015, 3981558}. Although Qazi et al. (2010,
1276154), observed that relative spleen weight was slightly reduced in C57BL/6 mice following
10-day exposure to 0.005% PFOS, the effects did not reach significance.
3.4.2.2.1.2 Thymus
Reductions in thymus weight have been reported across studies of varying durations (7-60 days)
and species (mice or rats). It is unclear whether sex has an influence on toxicity, as a number of
studies did not include females in their investigations.
The aforementioned 28-day studies by NTP (2019, 5400978) and Lefebvre et al. (2008,
1276155) reported reductions in absolute and/or relative thymus weights in male Sprague-
Dawley rats administered oral PFOS, at the highest doses of 5-7.58 mg/kg/day (Figure 3-25).
Reductions in absolute thymus weight were also observed in females of the highest dose in
Lefebvre et al. (2008, 1276155). In contrast, females in the NTP study exhibited reduced
absolute thymus weights at doses as low as 1.25 mg/kg/day, suggesting a higher sensitivity in
females {NTP, 2019, 5400978} (Figure 3-25).
Similarly, reduced thymic weights were observed in male C57BL/6 mice administered 20 or
40 mg/kg/day PFOS via gavage for 7 days {Zheng, 2009, 1429960}, 0.02% PFOS for 10 days in
diet {Qazi, 2009, 1937260}, or 0.417-2.083 mg/kg/day PFOS for 60 days {Dong, 2009,
1424951}. A follow up from the latter study {Dong, 2009, 1424951} by Dong et al. (2011
1424949) also exposed adult male C57BL/6 to 0.008-0.833 mg/kg/day PFOS for 60 days via
gavage, but reductions in relative thymus weight were only observed in the highest dose. Female
mice were not assessed in these studies. Yang et al. (2021, 7643494) exposed male C57BL/6
mice to 0.25 or 2.5 mg/kg/day PFOS for 28 days and observed an 18% and 24%, respectively,
reduction in relative thymus weight although these changes were not statistically significant.
In a developmental exposure study, the thymus was weighed in 4- and 8-week-old offspring of
pregnant C57BL/6 mice given 0, 0.1, 1, or 5 mg/kg/day PFOS from GD 1-GD 17 via gavage. In
male pups from the 5 mg/kg/day exposure group, relative thymus weights were reduced at 4 and
8 weeks of age. However, no effects were observed in lower dose groups or in females {Zhong,
2016, 3748828} (Figure 3-25).
In contrast to the several studies that reported reductions in thymus weight, Qazi et al. (2010,
1276154) and Peden-Adams et al. (2008, 1424797) did not observe any changes in thymus
weight. Qazi et al. (2010, 1276154) exposed male C57BL/6 (H-2b) mice to 0.005% PFOS in the
diet for 10 days, while Peden-Adams et al. (2008, 1424797) exposed male and female B6C3F1
mice to 0.00017-0.166 mg/kg/day PFOS for 28 days. The contrasting results of the 28-day study
by Peden-Adams et al. (2008, 1424797) and NTP (2019, 5400978) may underscore species
differences, however, the dose levels used in the mouse study were generally below the LOEL of
the NTP study (5 mg/kg/day).
3-100
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
PFOS Immune Effects - Thymus Weights
Endpoint Study Name Study Design Observation Time Animal Description Dose (mg/kg/day) | (£> Statistically significant 0 Not statistically significant!—I 95% CI [
Thymus Weight, Absolute Lefebvre etal., 2008, 1276155 short-term {28d) 28d
NTP. 2019,5400978 short-term (28d) 29d
Thymus Weight, Relative Zhong ot al., 2016,3748828 developmental (GD1-17) PNW4
Rat. Sprague-Dawley (•*. N=15) 0
0.14
1.33
3.21
6.34
Rat, Sprague Dawley {', N=15) 0
0.15
1.43
3.73
7.58
Rat, Sprague-Dawley (;•*, N=10) 0
0.312
0.625
1.25
2.5
5
Rat, Sprague-Dawley N=9-10) 0
0.312
0.625
1.25
2.5
5
F1 Mouse, C57BL/6 N=12) 0
F1 Mouse, C57BL/6 (?, N=12) 0
PNW8
F1 Mouse. C57BLf6 { :, N=12)
0
0.1
5
F1 Mouse, C57BU6 { \ N=12)
0
0.1
5
Yang et al., 2021. 7643494 short-term <28d)
29d
Mouse. C57BI/6 (,?, N=6)
0
0.25
2.5
Lefebvre etal.. 2008.1276155 short-term <28d)
28d
Rat, Sprague-Dawley (-', N=15)
0
0.14
1.33
3.21
6.34
NTP, 2019,5400978 short-term (28d) 29d
Rat, Sprague-Dawley ( -, N=15) 0
0.15
1.43
3.73
7.58
Rat, Sprague-Dawley N=10) 0
0.312
0.625
1.25
2.5
5
Rat, Sprague-Dawley (' . N=9-10) 0
0.312
0.625
1.25
2.5
Percent control response (%)
Figure 3-25. Percent Change in Thymus Weights Relative to Controls in Rodents Following
Exposure to PFOS
Interactive figure and additional study details available on HAWC.
GD = gestation day; PNW = postnatal week; Ei = first generation
3-101
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.2.2.2 Histopathology
Histopathology of the spleen, thymus, and/or lymph nodes has been evaluated following oral
exposure to PFOS across studies of varying durations in rodents (Figure 3-26). In general, short-
term and subchronic studies have observed histopathology such as extramedullary hematopoiesis
{NTP 2019, 5400978}, bone marrow hypocellularity {NTP, 2019, 5400978}, and other
aberrations in the immune organs {Qazi, 2009, 1937260; Lv, 2015, 3981558}.
One study included in the 2016 HESD {U.S. EPA, 2016, 3603365} by Qazi et al. (2009,
1937260) described perturbations in the thymus of male C57BL/6 (H-2b) mice exposed to 0.02%
(equivalent to -40 mg/kg/day) PFOS in feed for 10 days; the thymic cortex was smaller and
devoid of cells and the cortical/medullary junction was indistinguishable. These observations
may coincide with the reduction in thymus weight described above {Qazi, 2009, 1937260; NTP,
2019, 5400978}. However, the 28-day study in rats by NTP did not observe histopathologic
effects in the thymus of males or females following exposure to 0.312-5 mg/kg/day PFOS
{NTP, 2019, 5400978}, and this finding was complemented by a chronic non-human primate
study by Seacat et al. (2002 757853), which also found no effects in the thymus of males or
females following PFOS exposure (0, 0.03, or 0.15 mg/kg/day).
In spleens of male BALB/c mice, no significant increases in non-neoplastic lesions were
observed following exposure to 2.5, 5, or 10 mg/kg/day PFOS for three weeks, though
quantitative results were not reported {Lv, 2015, 3981558}. However, the authors {Lv, 2015,
3981558} state that alterations in spleen architecture were observed at the end of the exposure in
the 5 and 10 mg/kg/day groups. Moreover, splenic sinusoids, which drain into pulp veins, were
dilated and hyperemic. Peripheral splenic pulp structure and splenic cords (also known as red
pulp cords or cords of Billroth) were destroyed, the marginal zone disappeared, and
megakaryocytes (myeloid cell precursors) were abundant.
3-102
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
F.ndpoint
Bone Marrow. Hypocellularity
Study Name Study Design Observation Time Animal Description IKise (mg/kg/day)
NTP, 2019.5400978 short-term <28d) 29d Rat. Sprague-Dawley (d\ N=10) 0
0.312
0.625
5
Spleen, Intramedullary Heinaiopoiesis NTP. 2019.5400978 shori-ierm (28d) 29d
Rai. Sprague-Dawley (9. N=10) 0
0.312
0.(525
1.25
Rai. Sprague-Dawley ( 0
0.312
0.625
PFOS Immune Effects - Histopathology
^J^^^lalislicall^iignificaii^^^^^Jo^iignican^
0 10 20 30 40 50 60 70 80 90 100
Incidence (%)
Figure 3-26. Incidences of Immune Cell Histopathology in Rodents Following Exposure to
PFOS
Interactive figure and additional study details available on HAWC.
Xing et al. (2016, 3981506) examined spleens of male C57BL/6J mice for histopathology; no
distinguishable morphological differences were observed between any exposure group (2.5, 5, or
10 mg/kg/day for 30 days) and control. Similarly, Li et al (2021, 7643501) reported that there
were no significant lesions observed in the spleen among female BALB/c mice exposed via
gavage to 0.1 or 1 mg/kg/day PFOS for 60 days.
One study reported histology for the lymphatic system, but no histopathology was observed in
the lymph nodes (mandibular and mesenteric) following PFOS exposure {NTP, 2019, 5400978}.
3.4.2.2.3Circulating Immune Cells
Effects of PFOS exposure on circulating immune cells have been reported in rodents and non-
human primates. Alterations in neutrophil and white blood cell (WBC) populations in the
circulation have been observed in rodents, but the directionality of the effect is often
inconsistent, possibly reflecting differences in the timing of exposure.
Qazi et al. (2009, 1937259) performed a study to see if exposure to PFOS influenced circulating
immune cells. Male C57BL/6 mice were fed chow containing 0.02% PFOS for 10 consecutive
days, after which levels of WBCs were evaluated in blood collected from retroorbital puncture.
The absolute WBC count was significantly reduced and was mainly a reflection of decreased
lymphocytes, as no change in neutrophils was seen. A significant reduction of the relative
proportion and absolute number of macrophages in the bone marrow was also reported {Qazi,
2009, 1937259}. In a study by Seacat et al. (2003, 1290852), male and female Sprague-Dawley
rats were exposed to 0, 0.5, 2, 5, or 20 ppm PFOS for 14 weeks and WBC counts were
determined. The only statistically significant change was an increase in neutrophils in the
3-103
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
20 ppm exposure group (1.33 mg/kg/day dose equivalent) in the males only. No effects were
observed at lower exposure groups (0.5, 2.0, 5.0 ppm) nor in females {Seacat, 2003, 1290852}.
A shorter (28-day) study in male and female Sprague-Dawley rats exposed to 0.14-
7.58 mg/kg/day PFOS did not observe any statistically significant effects on circulating white
blood cell populations {Lefebvre, 2008 1276155}. The authors examined a myriad of circulating
immune cell endpoints, including WBC, total lymphocytes, as well as the number and
percentages of CD3+ (all T cells), CD3+/CD8+ (Cytotoxic T cells), CD3+/CD4+ (Helper T
cells), CD45RA+ (B-cells). Although not significant, Helper T cell counts in males and females
were elevated from control by 35% or 42%, respectively, which coincided with a 29% or 41%
increase in total T cell counts, suggesting that there may be a specific effect of PFOS on helper T
cell populations. Similarly, Yang et al. (2021, 7643494) found that exposure of male C57BL/6
mice to 2.5 mg/kg/day PFOS for 28 days did not significantly alter WBC counts, nor percent or
number of neutrophils, total lymphocytes, eosinophils, monocytes, and basophils in the serum.
Evidence from one paper {Seacat, 2002, 757853} suggests that the effects of PFOS on WBCs
that have been noted in some rodent studies do not extend to non-human primates. Male and
female cynomolgus monkeys, orally administered 0.3-0.75 mg/kg/day PFOS for 26 weeks,
exhibited no significant change in WBC counts, including neutrophils and total lymphocytes
{Seacat, 2003, 757853}. In contrast, reduced numbers of neutrophils were observed in male rats,
but not females, in an NTP (2019, 5400978) study. In that report, NTP also reported that male
rats, and not females, exhibited significantly reduced WBC counts {NTP, 2019, 5400978}.
3.4.2.2.4Natural Killer Cell Activity
The available data on the effect of PFOS exposure on natural killer (NK) cell activity indicate
that there may be different effects in NK cell activity based on dose, but there are too few studies
to make any determination and no single study assesses the continuum of doses to see if there is
an opposing effect at different areas of the dose response curve. Oral administration of 0.00017-
0.166 mg/kg/day PFOS to male and female B6C3F1 mice for 28 days resulted in increased NK
cell activity in males only exposed to 0.017, 0.033, and 0.166 mg/kg/day {Peden-Adams, 2008,
1424797}. Male C57BL/6 mice exposed to 0.083 mg/kg/day PFOS daily for 60 days displayed
significantly increased NK cell activity by 38%, but treatment with 0.833 and 2.083 mg/kg/day
resulted in decreased NK cell activity {Dong, 2009, 1424951}. Female mice were not assessed in
this study. In another assessment of male C57BL/6 mice administered 0-40 mg/kg/day for
7 days, NK cell activity was reduced following exposure to 20 and 40 mg/kg/day {Zheng, 2009,
1429960}. Similarly, Zhong et al. (2016, 3748828) reported thatNK cell activity was decreased
in 4-week-old male offspring from the 5 mg/kg/day group and also reduced in 8-week-old
offspring from the 1 or 5 mg/kg/day group. The latter result was recapitulated in the study by
Keil et al. (2008, 1332422) where the female C57BL/6 mice were mated with C3H to derive
B6C3F1 offspring. Female offspring from both studies were less sensitive to the PFOS-induced
reduction in NK cell activity {Keil, 2008, 1332422; Zhong, 2016, 3748828} as indicated by the
lack of statistically significant changes in females exposed to 1 mg/kg/day in each study.
Moreover, at 8 weeks, NK cell activity was suppressed by 42.5% and 32.1% in males at the 1
and 5 mg/kg/day treatments, respectively, and was suppressed by 35.1% in females at the
5 mg/kg/day treatment {Keil, 2008, 1332422}. These studies indicate that male mice may be
more susceptible to PFOS-induced altered NK cell activity, and that NK cell activity can be
increased or decreased following low or high PFOS exposure, respectively (Table 3-7).
3-104
-------
DRAFT FOR PUBLIC COMMENT MARCH 2023
Table 3-7. Associations Between PFOS Exposure and Natural Killer Cell Activity in Mice
Reference
Exposure Length
Dose
(mg/kg/day)
Sex
Change
Peden-Adams et al.
(2008, 1424797)
28 days
0, 0.00017, 0.0017, 0.0033,
0.017,0.033,0.166
M
4
0.017-0.166 mg/kg/day
F
n.s.
Dong et al. (2009,
1424951)
60 days
0,0.008,0.083,0.417,0.833, M
2.083
t
(0.083 mg/kg/day)
4
(0.833-2.083 mg/kg/day)
Zheng et al. (2009,
1429960)
7 days
0, 5, 20, 40
M
1
(20-40 mg/kg/day)
Zhong et al. (2016,
3748828)
GD 1-17
4-week assessment
0,0.1, 1,5
M
1
5 mg/kg/day
F
n.s.
GD 1-17
8-week assessment
0,0.1, 1,5
M
1
1-5 mg/kg/day
F
1
5 mg/kg/day
Keil et al. (2008,
GD 1-17
0,0.1, 1,5
M
n.s.
1332422)
4-week assessment
F
n.s
GD 1-17
8-week assessment
0,0.1, 1,5
M
1
1-5 mg/kg/day
0,0.1, 1,5
F
1
5 mg/kg/day
Notes: F = female; M = male; n.s. = nonsignificant.
3.4.2.2.5Spleen Cellularity
Splenocyte sub-classes were quantified in several rodent studies (Figure 3-27). Splenic T cell
immunophenotypes were slightly affected in male and female B6C3F1 mice exposed to oral
administration of 0.00017-0.166 mg/kg/day PFOS for 28 days {Peden-Adams, 2008, 1424797}.
In males, CD4"/CD8+ and CD4VCD8" cells were increased, whereas numbers of CD4+/CD8" and
CD4+/CD8+ cells were decreased beginning at 0.0033 mg/kg/day. In females, splenic CD4"/CD8+
and CD4+/CD8- cells were decreased beginning at 0.0033 mg/kg/day. Significantly decreased
splenocyte populations were also observed in male C57BL/6 mice exposed to 0.02% PFOS for
10 days {Qazi, 2009, 1937260}, 20 or 40 mg/kg/day PFOS for 7 days {Zheng, 2009, 1429960},
and 0.417-2.083 mg/kg/day for 60 days {Dong, 2009, 1424951}. Female mice were not
evaluated in these studies.
Altered splenic cellular composition was observed in a study by Lv et al. (2015, 3981558) where
male BALB/c mice were exposed to 0, 2.5, 5, or 10 mg/kg/day PFOS for 3 weeks {Lv, 2015,
3981558}, and spleens harvested for lymphocyte counting and phenotyping. Fluctuations in
lymphocyte counts and T cell proliferation were apparent at the 3-week timepoint. A dose-
dependent increase in the number of splenic T cells (CD3+) relative to controls was observed at
the end of 3 weeks, reaching significance in the 2.5 and 10 mg/kg/day exposure groups. This
coincided with a non-significant increase in T-helper (CD3+CD4+) and T-cytotoxic (CD3+CD8+)
lymphocytes in the 5 and 10 mg/kg/day groups, all relative to controls. The percentages of T-
3-105
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
helper (CD3+CD4+) and T-cytotoxic (CD3+CD8+) lymphocytes were increased in the
10 mg/kg/day groups {Lv, 2015, 3981558}.
Further effects of PFOS on immune cell composition in the spleen have also been reported
following developmental exposure by Keil et al. (2008, 1332422) and Zhong et al. (2016,
3748828). Zhong et al. (2016, 3748828) exposed pregnant female C57BL/6 mice to 0.1-
5 mg/kg/day PFOS from GD 1-GD 17, and then quantified various immune cell populations in
male and female pups. Decreased splenic cell sub-populations (CD4+ and CD8+ cell counts) were
observed in the 4-week-old male pups from the 5 mg/kg/day exposure group. At 8-weeks,
reductions in CD8+ cells in the spleen were observed in the 5 mg/kg/day exposure group {Zhong,
2016, 3748828}.
B220+ Cell Count
CD4+ Cell Count
CD8+ Coll Count
CD4+/CD8+- Cell Count
Study Name Study Design Observatio
Zhong el al.. 2016. 3748828 developmental (GD1-17) PNW4
Zhong et al., 2016, 3748828 developmental (GD1-17) PNW4
PNW8
Zhong at al., 2016, 3748828 developmental (GD1-17) PNW4
PNW8
Zhong et al.. 2016. 3748828 developmental (GD1-17) PNW4
i Time Animal Description
F1 Mouse. C57BL/6 (:
F1 Mouse,
F1 Mouse
F1 Mouse
F1 Mouse
F1 Mouse
F1 Mouse
F1 Mouse
F1 Mouse
F1 Mouse
F1 Mouse
F1 Mouse
F1 Mouse
F1 Mouse
F1 Mouse
N=12)
C57BL/6 (-. N=12)
C57BL/6 ( N=12)
C57BL'6 (2, N=12)
C57BL/6 ( ; , N=12)
C57BL/6 (2, N=12)
C57BL/6 (/'¦, N=12)
C57BL/6 ( N=12)
C57BL/6 (c , N=12)
C57BL/6 ( N=12)
C57BL/6 . N=12)
C57BL/6 N=12)
C57BL.'6 N=12)
CS7BL/6 N=12)
C57BL/6 N=12)
C57BL/6 ( N=12)
Splenic Cellularity, Lymphocytes, CD3* L
el al.. 2015
3981558
short-term
(21 d)
3wk
Mouse.
BALB/c (•-
N=4)
Splenic Cellularity, Lymphocytes, CD3+ (Normalized to Control) L
etal.,2015
3981558
short-term
(21 d)
3wk
Mouse,
BALB/c ( -
N=4)
Splenic Cellularity, Lymphocytes, CD3+CD4+ L
et al.. 2015
3981558
short-term
(21 d)
3wk
Mouse,
BALB/c (;'
N=4)
Splenic Cellularity, Lymphocytes, CD3+CD4+ (Normalized to Control) L
etal.,2015
3981558
short-term
(21 d)
3wk
Mouse,
BALB/c (.;
N=4)
Splenic Cellularity, Lymphocytes, CD3+CD8+ L
etal.,2015
3981558
short-term
(21 d)
3wk
Mouse.
BALB/c (
N=4)
Splenic Cellularity, Lymphocytes, CD3+CD8+ (Normalized to Control) L
et al.. 2015
3981558
short-term
(21 d)
3wk
Mouse.
BALB/c (r
N=4)
PFOS Immune Effects - Splenic Immune Cellularity
0 No significant change^ Significant increase ^ Significant decrease
Concentration (mg/kg/day)
Figure 3-27. Splenocyte Cellularity in Rodents Following Exposure to PFOS (logarithmic
scale)3
PFOS concentration is presented in logarithmic scale to optimize the spatial presentation of data. Interactive figure and additional
study details available on HAWC.
GD = gestation day; PNW = postnatal week; Fi = first generation.
a Zhong et al., 2016 reported data on both splenic and thymic lymphocyte populations for the same experimental animals. Results
are shown in separate figures.
3.4.2.2.6Thymus Cellularity
Thymus cell populations were less sensitive to the effects of PFOS compared to the effects
observed in the spleen, as determined by the dose where the change occurred and the number of
endpoints that changed following PFOS exposure (Figure 3-28). Indeed, while all splenic T cell
CD4/CD8 subpopulations were altered in one study of male B6C3F1 mice beginning at 0.1
mg/kg/day exposures, none of the thymic T cell subpopulations were affected. Furthermore, the
effects appeared to also have a female-bias; although thymic CD4"/CD8+ cells were increased in
female B6C3F1 mice exposed to 0.033 or 0.166 mg/kg/day, no effects were observed in males
{Peden-Adams, 2008, 1424797}. In contrast, significantly decreased thymocyte populations
were observed in male C57BL/6 mice exposed to 0.02% PFOS for 10 days {Qazi, 2009,
1937260}, 20 or 40 mg/kg/day PFOS for 7 days {Zheng, 2009, 1429960}, and 0.417-
3-106
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
2.083 mg/kg/day for 60 days {Dong, 2009, 1424951}. Female mice were not evaluated in these
studies.
Effects of PFOS on immune cell composition in the thymus have also been reported following
developmental exposure. Pregnant female C57BL/6 mice were dosed with 0.1-5 mg/kg/day
PFOS from GD 1-GD 17, and immune cell populations were quantified in male and female pups
at 4 and 8 weeks after birth. Decreased thymic lymphocyte sub-populations (CD4+, and CD4"
/CD8" cell counts) and decreased thymic cellularity were observed in the 4-week-old male pups
from the 5 mg/kg/day exposure group, and no effects were observed in females {Zhong, 2016,
3748828}. At 8-weeks, no effects were observed in females and reductions in thymic CD4+ cells
were observed in males from the 5 mg/kg/day exposure group. These findings were
complimented by Keil et al. (2008, 1332422), who observed a reduction in CD3+ and CD4+
thymocytes in 8-week C57BL/6N male mice following exposure to 0.1-5 mg/kg/day from GD
1-GD 17 {Keil, 2008, 1332422}.
PFOS Immune Effects - Thymic Immune Cellularity
Endpoint
CD4+ Cell Count
CD8+ Cell Count
Study Name Study Design Observation Time
Zhong et al., 2016, 3748828 developmental (GD1-17) PNW4
Zhong et al.. 2016, 3748828 developmental (GD1-17) PNW4
CD4+/CD8+ Cell Count Zhong et al., 2016, 3748828 developmental (GD1-17) PNW4
CD4-/CD8-Cell Count Zhong et al.. 2016, 3748828 developmental (GD1-17) PNW4
Animal Description
F1 Mouse, C57BL/6 ( 5, N=
F1 Mouse, C57BL/6 (<>, N=
F1 Mouse, C57BL/6 (
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
blood cell (SRBC) plaque-forming cell (PFC) response, which measures IgM-producing cells,
was reduced in male and female B6C3F1 mice administered 0.0017-0.166 mg/kg/day PFOS for
28 days. The response was suppressed at lower PFOS doses in male mice (effect first observed at
0.0017 mg/kg/day) than female mice (effect first observed at 0.017 mg/kg). Because IgM
suppression can result from effects on both T and B cells, antibody production was also
measured in response to a bacteria-like challenge, trinitrophenyl (TNP)-lipopolysaccharide
(LPS), which would induce a T-independent response. Following the TNP-LPS challenge, a
decrease in IgM titers was observed in female B6C3F1 mice that had been exposed to
0.334 mg/kg/day PFOS for 21 days. Male animals were not assessed in this study {Peden-
Adams, 2008, 1424797}. Similarly, Dong et al. (2009, 1424951) observed a dose-dependent
reduction in the SRBC-specific IgM PFC response in male C57BL/6 mice exposed to PFOS
daily for 60 days. These results are consistent with a similar study by the same authors in 2011,
including a dose-dependent reduction in IgM levels in serum {Dong, 2011 1424949}. The
authors also examined the delayed-type hypersensitivity response (DTH) to SRBC. Although
IgM levels were reduced in groups exposed to 0.0833 mg/kg/day PFOS or higher, IgG, IgGl,
and IgE levels were elevated only in the highest exposure group (0.833 mg/kg/day), and no
change was observed in IgG2a levels {Dong, 2011 1424949}. To further assess the DTH
response, footpad thickness was measured using digital calipers on the foot used to sensitize the
mice to SRBC relative to the non-sensitized foot; no significant increase in footpad swelling was
observed. Female mice were not assessed in either of these studies. The DTH response was also
assessed by Lefebvre et al. (2008, 1276155) in male and female rats sensitized with the T-
dependent antigen, keyhole limpet hemocyanin (KLH), during a 28-day exposure to 0.14-
7.58mg/kg/day PFOS (on days 14 and 21) and challenged at the end of study with KLH. There
were no significant changes in anti-KLH IgG titers in males or females compared to control, and
there were no changes in footpad swelling. Zheng et al. (2009, 1429960) also found that the PFC
response to a SRBC challenge was suppressed in male C57BL/6 mice given 5, 20, or
40 mg/kg/day PFOS for 7 days. These rodent studies provide evidence of a PFOS-induced
suppression of the immune response to a SRBC challenge that may be more sensitive in male
mice (Table 3-8).
Table 3-8. Associations Between PFOS Exposure and Immune Response in Mice
Reference
Exposure Length
Dose
(mg/kg/day)
Sex
Change
Peden-Adams et al.
28 days
0, 0.00017, 0.0017,
M
4
(2008, 1424797)3
0.0033,0.017,0.033,
0.0017-0.166 mg/kg/day
0.166
F
4
0.017-0.166 mg/kg/day
Lefebvre et al. (2008,
28 days
0,0.14, 1.33,3.21,6.34
M
n.s.
1276155)b
(males) orO, 0.15, 1.43,
3.73, 7.58 (females)
F
n.s.
Dong et al. (2009,
142495 If
60 days
0,0.008,0.083,0.417,
0.833,2.083
M
1
0.083-2.083
Dong et al. (2011,
1424949)3
60 days
0, 0.008, 0.0167, 0.083,
0.417,0.833
M
1
0.083-0.833
Zheng et al. (2009,
1429960)3
7 days
0, 5, 20, 40
M
1
5-40 mg/kg/day
Zhong et al. (2016,
GD 1-17
0,0.1, 1,5
M
1
3748828)a
4-week assessment
1-5 mg/kg/day
3-108
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Reference
Exposure Length
Dose
(mg/kg/day)
Sex
Change
F
4
5 mg/kg/day
GD 1-17
0,0.1, 1,5
M
n.s.
8-week assessment
F
n.s.
Keil et al. (2008,
GD 1-17
0,0.1, 1,5
M
4
1332422f
8-week assessment
F
5 mg/kg/day
n.s.
Notes: F = female; M = male; n.s = nonsignificant.
a Sheep red blood cell-specific IgM production.
bKeyhole limpet hemocyanin-specific IgG production.
Similar observations were reported in two developmental PFOS exposure studies. Keil et al.
(2008, 1332422) and Zhong et al. (2016, 3748828), each exposed pregnant female C57BL/6
mice to 0.1-5 mg/kg/day PFOS from GD 1-GD 17 and then tested the immune responses in
offspring at 4 and 8 weeks of age. Four days before sacrifice, mice were injected with SRBC to
induce an immune response. In males from the 5 mg/kg/day exposure group, the primary IgM
response to SRBC was significantly suppressed by 53% at 8-weeks. In females, the primary IgM
response was not altered Keil et al. (2008, 1332422). Similarly, Zhong et al. (2016, 3748828)
observed that SRBC-specific IgM production by B-lymphocytes in the spleens of 4-week old
mouse pups exposed to 1 or 5 mg/kg/day PFOS in utero was reduced by 15% or 28%,
respectively. In females, the SRBC-specific IgM response was significantly suppressed by 24%
in the 5 mg/kg/day group only. However, no significant changes were observed at 8 weeks.
Alterations in the serum levels of globulin can be associated with decreases in antibody
production {FDA, 2002, 88170}. Two 28-day studies {NTP, 2019, 5400978; Curran, 2008,
757871} in male and female Sprague-Dawley rats reported effects on serum globulin levels. In
the first study, rats were orally administered 0.312-5 mg/kg/day PFOS. Male rats exhibited
significantly decreased globulin while globulin in females did not significantly differ from
control values {NTP, 2019, 5400978}. These findings are complemented by a study by Curran et
al. {2008, 757871}, in which male and female rats fed diets containing 2-100 mg/kg PFOS
(equivalent to 0.14-6.34 mg/kg/day in males and 0.15-7.58 mg/kg/day in females) for 28 days.
In male rats, serum albumin/globulin ratios were elevated in the highest exposure group in
conjunction with a significant dose-related negative trend in globulin levels. In female rats, no
changes were observed in albumin/globulin ratio or globulin levels. In a separate study
{Lefebvre, 2008, 1276155} the same authors also reported total levels of IgM, IgG, IgGl,
IgG2a, IgG2b, and IgG2c in serum of male and female rats exposed to 0, 2, 20, 50, or 100
mg/kg/day PFOS for 28 days. In males, significant reductions in IgGl levels were observed at
the two lowest doses and a significant positive trend was observed for trend for IgG, IgG2a, and
IgG2c. In females, both IgM and IgG2c levels were significantly elevated in the highest dose
group.
Two studies by Lee et al. (2018, 5085013) and Yang et al. (2021, 7643494) found evidence that
PFOS exposure can exacerbate an allergic immune response in mice. Lee et al. sensitized male
ICR mice with ovalbumin (OVA) on day 0 and day 7 and exposed them to 50-150 mg/kg/day
PFOS on study day 9, 11, and 13. Serum histamine, TNF-a, IgE, and IgG levels were increased
3-109
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
following exposure, suggesting that PFOS exacerbates mast cell-mediated allergic inflammation.
These findings are complemented by studies in male C57BL/6 mice by Yang et al. (2021,
7643494). In that study, mice were exposed to PFOS for 28 days via gavage, sensitized to OVA
and adjuvant via subcutaneous injection on days 4 and 11, and challenged with an aerosol of 1%
OVA on days 26 to 28. In the serum, exposure to OVA alone or to OVA + PFOS did not lead to
elevations in WBC counts, nor percent or number of neutrophils, total lymphocytes, eosinophils,
monocytes, and basophils. Serum IgE levels and anti-OVA IgE antibodies were elevated in
groups exposed to 0.25 or 2.5 mg/kg/day PFOS + OVA compared to OVA alone or untreated
controls. Mice exposed to 0.25 or 2.5 mg/kg/day PFOS alone showed a low level of serum IgE,
similar to the control group.
3.4.2.3 Mechanistic Evidence
Mechanistic evidence linking PFOS exposure to adverse immune outcomes is discussed in
Sections 3.1.1.6, 3.3.2, 3.3.4, and 3.3.6 of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}.
There are 24 studies from recent systematic literature search and review efforts conducted after
publication of the 2016 PFOS HESD that investigated the mechanisms of action of PFOS that
lead to immune effects. A summary of these studies is shown in Figure 3-29.
Mechanistic Pathway Animal Human In Vitro Grand Total
Big Data, Non-Targeted Analysis
1
0
0
1
Cell Growth, Differentiation, Proliferation, Or Viability
6
0
mm
13
Cell Signaling Or Signal Transduction
2
0
'
6
Extracellular Matrix Or Molecules
0
0
2
2
Fatty Acid Synthesis, Metabolism, Storage. Transport, Binding, B-Oxidation
1
0
1
2
Hormone Function
1
0
0
1
Inflammation And Immune Response
6
5
12
19
Oxidative Stress
1
0
3
4
Other
1
0
0
1
Grand Total
8
5
15
24
Figure 3-29. Summary of Mechanistic Studies of PFOS and Immune Effects
Interactive figure and additional study details available on Tableau.
3.4.2.3.1Mechanistic Evidence for PFOS-mediated Effects on the Immune System
Since the 2016 HESD advisory was released, 26 studies were identified that inform the
mechanism by which PFOS may alter or perturb immune system function or immune system
development and physiology. Recent studies provide mechanistic insights into PFOS effects on
immune system development and physiology (5 studies), adaptive immune responses (6 studies),
innate immune responses (4 studies), intrinsic cellular defense (1 study), and disruption of
inflammatory responses (9 studies). Mechanistic pathways associated with the immune system
identified in the recent PFOS literature included inflammation, immune responses, cell viability,
cell signaling, oxidative stress, and hormone function.
3-110
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.2.3.1.1 Mechanistic Evidence for PFOS-mediated Effects on Immune
System Development and Physiology
Alterations in immune and allergic responses in exposed children may suggest PFOS-mediated
effects in immune system development. In addition, changes in white blood cell count {Oulhote,
2017, 3748921} and alterations in gene expression related to immune and inflammation
responses in human cord blood {Pennings, 2016, 3352001} present potential mechanisms of
immunotoxicity in children. In animals, PFOS-related health effects related to immune system
development and physiology are described in Sections 3.4.2.2.1 to 3.4.2.2.7. Briefly, effects in
mice and rats included reduced spleen and thymus weights, alterations in spleen and thymus
morphology, and changes in the cellularity and immunophenotypes of lymphocytes. Effects
varied by sex and strain.
Three mechanistic studies in mice suggest that changes in immune physiology and development
following exposure to PFOS can be sex-dependent. Zhong et al. (2016, 3748828) demonstrated
sex-specific impacts of PFOS on immune organ development and physiology in C57BL/6 mice
exposed during development. Pups were evaluated after maternal oral exposure to PFOS (0.1,
1.0, or 5.0 mg PFOS/kg/day) from gestational day (GD) 1-17. Sex-dependent alterations in
spleen and thymus organ weights, cellularity, and cellular immunophenotypes are discussed in
Section 3.4.2.2. These may be linked to sex hormones during development as there was a
significant interaction between sex and PFOS concentrations for serum testosterone at 4 and 8
weeks of age, and estradiol at 4 weeks of age. The authors suggest that sex-dependent
differences in PFOS excretion, the endocrine-disrupting properties of PFOS, or male or female
sex hormone-differences may influence the sex-specific impact on spleen and thymus
physiology.
Lv et al. (2015, 3981558) reported disrupted splenic architecture and reduced absolute numbers
(albeit increased percentages) of T helper (CD3+CD4+) and cytotoxic T (CD3+CD8+) cells in
the spleen of male BALB/c mice administered 10 mg/kg/day PFOS via gastric gavage for 3
weeks followed by a 1-week recovery. Gene expression profiling identified differential
regulation of genes involved in mitogen-activated protein kinase (MAPK) signal transduction
pathways and in cellular responses to oxidative stress. The effects on gene expression paralleled
a dose-dependent increase in intracellular free calcium ([Ca2+], which plays an important role in
immune cell proliferation in response to foreign antigens) concentration in splenocytes of
exposed animals, suggesting that activation of MAPK signaling pathway and/or oxidative stress
genes in response to PFOS may alter splenic architecture via induction of apoptosis in
lymphocytes.
Qazi et al. (2012, 1937236) also observed decreased spleen and thymus weights and cellularity
as well as reduced numbers of myeloid, pro/pre-B, and immature B cells in bone marrow (BM).
In male C57BL/6 (H-2b) mice fed diets containing PFOS compounds (0.001-0.02%, w/w) for 10
days, atrophy of the thymus and spleen as well as hypocellularity of BM was observed at the
higher dose of 0.02%. PFOS exposure caused reduced feed consumption and atrophy of the
thymus and spleen and hypocellularity of bone marrow cells. Histopathological and flow
cytometric analysis of BM showed significant reductions in the total numbers of bone marrow
cells as well as the numbers of pro/pre-B (CD 19 + CD138 + IgM+) and immature B (CD 19+
CD138+ IgM+) cells. Myeloid (Grl+ CD1 lb+) cells and B-lymphoid (CD19+) cells were also
reduced in mice administered the high dose of PFOS. After 10 days of withdrawal of PFOS from
3-111
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
feed, the effects in bone marrow partially or completely reversed. Interestingly, food restriction
alone in the absence of PFOS exposure also led to reduced cell numbers in the thymus and
spleen and resulted in reductions of the total numbers of B-lymphoid cells, pro/pre-B, and
immature B cells. These findings indicate that immunotoxicity of PFOS may, at least in part, be
a consequence of reduced food consumption. Additionally, perturbation of the bone marrow may
contribute to reduced numbers of splenic B cells, atrophy of the spleen, and impaired humoral
immune responses caused by exposure to PFOS.
3.4.2.3.2 Mechanistic Evidence for PFOS-mediated Effects on Adaptive Immune
Responses
3.4.2.3.2.1 Mechanistic data informing suppression of immune responses to
vaccines and infectious diseases
The effects of prenatal, childhood, or adult PFOS exposure on responses to vaccines and
infectious diseases are described in Section 3.4.2.1. Briefly, studies observed an inverse
association between PFOS exposure and vaccine-induced antibody levels to tetanus and to
pathogens including human foot and mouth disease (HFMD) and hepatitis B infection. Other
studies identified associations between PFOS exposure and increased incidence of infections
including those caused by pneumonia and chickenpox, though PFOS was associated with a
decrease in the incidence of respiratory syncytial virus (RSV), common cold, ear infection, and
urinary tract infection. Six new mechanistic studies were identified that inform PFOS-mediated
effects on adaptive immunity (3 in humans and 3 in mice). One mechanistic study directly
evaluated PFOS-mediated effects on adaptive immune responses specific to vaccines and
infectious disease {Pennings, 2016, 3352001}, and 5 mechanistic studies evaluated non-allergic
adaptive immune responses.
As described in Section 3.4.2.1.1, in children exposed to PFOS in utero, Granum et al. (2013,
1937228) previously reported an inverse association between maternal serum concentrations of
PFOS and anti-rubella antibody levels in serum of 3-year-old children, as well as an increased
incidence of the common cold, using samples and data from the Norwegian BraMat cohort. In a
follow-up study of early life immunosuppression again using Norwegian BraMat cohort data,
Pennings et al. (2016, 3352001) conducted a whole genome transcriptomic microarray analysis
of neonatal cord blood samples and compared the results to maternal levels of PFOS (as well as
PFOA, perfluorononanoic acid (PFNA), and perfluorohexane sulfonate (PFHxS)) in the blood.
Dose-response relationships between PFOS and expression of individual genes, rubella antibody
levels, and episodes of the common cold were analyzed. Expression of 636 genes was positively
associated with PFOS exposure, and 671 were negatively correlated. A set of 27 genes were
correlated between all four of the PFAS evaluated and the number of common cold episodes. Of
these, three genes were related to immunological and/or hematopoietic functions, including
peroxisome proliferator activated receptor delta (PPARD), SHC adaptor protein 4 (SHC4), and
cytokine like 1 (CYTL1), expressed in CD34+ in bone marrow and cord blood mononuclear
cells. Of the six genes related to development and/or morphogenesis, two overlapped with
immune and hematopoietic functions (PPARD and CYTL1). Interestingly, another gene
associated with development and morphogenesis, sphingosine-1 -phosphate lyase 1 (SGPL1), has
been recently associated with immune responses to viral infections including inhibition of
influenza virus replication by promoting antiviral type I interferon innate immune responses
{Wolf, 2019, 10259528}. A set of 26 genes overlapped between PFAS and rubella titers,
3-112
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
including two genes also identified in pathway analysis as relevant to regulation of T cell
activation (interleukin 27 [IL27] and the adenosine A2a receptor [ADORA2A]). Only one gene
(CYTL1) was in common between the sets of genes that overlapped with PFAS exposure and
common cold episodes, and PFAS exposure and rubella titers. However, a clear understanding of
the function of CYTL1 in hematopoiesis and immune function is lacking. While the correlation
between gene expression changes and changes in protein expression or function in cord blood
was not investigated in this study, these represent potential candidate genes that mediate the
mechanism(s) of early childhood immunotoxicity associated with prenatal exposure to PFOS and
other PFAS chemicals.
Lv et al. (2015, 3981558) examined T cells in male BALB/c mice administered 10 mg/kg/day
PFOS via gavage for 3 weeks followed by 1-week recovery. Gene expression profiling in spleens
was performed using GeneChip® Mouse Genome 430 2.0 Array (Affymetrix Inc., Santa Clara,
CA, USA) and quantitative real time PCR (qRT-PCR). The authors identified 1,327
differentially expressed genes (4% of all analyzed genes) in response to PFOS exposure.
Biological processes associated with differentially expressed genes included cell cycle, DNA
metabolism, mitosis, and DNA replication. Pathway analysis identified significantly upregulated
pathways related to the T cell receptor (TCR) and to immune signaling (primary
immunodeficiency signaling, inducible co-stimulator [iCOS] - iCOS ligand [iCOSL] signaling
in T helper cells, OX40 signaling pathway, and calcium-induced T lymphocyte apoptosis).
However, the transducer of ErbB-2.1 (TOB) T cell signaling pathway was significantly
downregulated, as were genes associated with nuclear factor erythroid derived 2 like 2 (Nrf2)-
mediated oxidative stress response (such as GSTM3 and MGST3). During the recovery period
following four weeks of PFOS exposure, immunoblotting confirmed a dose-dependent
upregulation of protein levels in spleens for several genes involved in TCR signaling and
calcium signaling, including thymocyte selection associated (THEMIS), the CD3 gamma subunit
of T-cell receptor complex (CD3G), and calcium/calmodulin dependent protein kinase IV
(CAMK4). Additionally, in splenocytes of exposed animals, [Ca2+]i increased in a
concentration-dependent manner, and T-cell proliferation in response to Concanavalin A (Con
A) stimulation was inhibited by PFOS. The authors suggest that activation of MAPK signaling
pathway and/or oxidative stress genes in response to PFOS may alter splenic architecture via
induction of apoptosis in lymphocytes. These findings also suggest that altered expression of cell
cycle genes, upregulation of genes involved in TCR signaling, and altered calcium homeostasis
impact T cell function through inhibition of T cell proliferation and induction of T cell anergy
(intrinsic functional inactivation of lymphocytes following an antigen encounter).
Li et al. (2020, 6833655) used an integrative 'omics approach to evaluate perturbations in the
transcriptome and lipidome in human lymphocytes that may impact adaptive immune responses
to vaccines or infectious diseases. Lymphocytes were isolated from human donors and cultured
before treatment with 50 mM PFOS for 72 hours. PFOS treatment led to a significant induction
of the cytokines IL-1, IL-4, IL-6, and IL-8 cytokines relative to controls, as measured by ELISA.
Subsequent deep sequencing of RNA for PFOS-treated lymphocytes revealed that numerous
differentially expressed genes were related to lymphocyte function and biological processes
related to immunity, including immune responses, innate immune responses, and inflammatory
responses. Enrichment analysis using the Kyoto Encyclopedia of Genes and Genomes (KEGG)
database linked PFOS treatment to stimulation of cytokine-cytokine receptor interactions,
extracellular matrix (ECM)-receptor interactions, the PI3K-Akt signaling pathway, the
3-113
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
peroxisome proliferator-activated receptor (PPAR) signaling pathway, cholesterol metabolism,
and phagosome and lysosome regulation at the gene expression level. The analysis identified
differentially expressed genes associated with cytokines, growth factors, and differentiation and
migration of antigen-presenting cells. Additionally, the authors conducted a lipidomic analysis of
treated cells using liquid chromatography-mass spectrometry (LC-MS). Lipid metabolites (40
upregulated and 56 downregulated) were identified in PFOS-exposed lymphocytes relative to
control lymphocytes. Clusters of lipids associated with immune function were dysregulated,
including lipids involved in glycerophospholipid metabolism, sphingolipid metabolism,
glycerolipid metabolism, adipocytokine signaling, regulation of autophagy, and arachidonic acid
metabolism. Taken together with the transcriptomic and functional analyses reported by Lv et al.
(2015, 3981558) and Pennings et al. (2016, 3352001), these findings suggest that PFOS exposure
may disrupt adaptive immunity through dysregulation of genes and lipids involved in
lymphocyte survival, proliferation, and anergy.
The potential for PFOS to suppress immune responses to vaccines and infection are also
informed by studies investigating PFOS-mediated effects on THl/TH2-type cytokines in mice
{Zhong, 2016, 3748828}, glycosylation of immunoglobulins in humans {Liu, 2020, 6833599},
and lymphocyte toxicity in vitro {Zarei, 2018, 5079848}. Zhong et al. (2016, 3748828) exposed
pregnant female C57BL/6 mice to PFOS (0.1, 1.0, or 5.0 mg/kg/day) from GD 1-17 and cultured
splenocytes of male pups at 4 and 8 weeks of age. Spontaneous IL-4 formation was increased
and spontaneous production of TH1 cytokines (i.e., IL-2) was decreased in the 5 mg/kg/day
group at 8 weeks. Functionally, lymphocyte proliferation was significantly decreased in
splenocytes from both males and females exposed to the highest dose at 4 weeks, and natural
killer (NK) cell activity exhibited a decreasing trend with dose (males only at 4 weeks, males and
females at 8 weeks). Given the reductions in serum testosterone at 4 and 8 weeks of age, and
increased estradiol levels in male pups at 4 weeks of age (discussed in Section 3.4.2.2), these
findings suggest that in utero exposure may elicit sex-specific alterations in TH1 and TH2
cytokine profiles in immune cells as well as diminished lymphocyte and NK functions.
A recent study suggests that PFOS may also alter antibody glycosylation patterns {Liu, 2020,
6833599}. Altered IgG glycosylation patterns are associated with disease states and immune
functions including cancer immunosurveillance and anti-inflammatory reactions {Cobb, 2020,
10284268}. TheN-glycome profiles of immunoglobulins from serum samples of adults and
children were analyzed by subjecting the IgG fraction to glycan release, derivatization, and
matrix-assisted laser desorption/ionization-MS (MALDI-MS) analysis. Specifically, increasing
PFOS exposure was associated with decreased galactosylation, increased fucosylation and
sialylation in adults, and increased agalactosylation, bisecting GlcNAcylation, sialylation and
decreased galactosylation in children. The authors suggested several mechanisms by which
altered IgG glycosylation impacts immunity including antibody-dependent cellular cytotoxicity
(ADCC). While no functional studies were conducted, these preliminary findings provide a
potential mechanism for altered antibody-dependent immune responses in PFOS-exposed
persons.
Zarei et al. (2018, 5079848) isolated lymphocytes from the blood of healthy humans and
analyzed cytotoxicity in vitro in response to exposure to 100-500 |iM PFOS for 12 hours. The
IC50 for cytotoxicity was calculated to be 163.5 |iM. Exposure to 75, 150, and 300 |iM PFOS
for 2, 4, 6, 8, 10, or 12 hours was associated with increased reactive oxygen species (ROS)
3-114
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
formation, lipid peroxidation, and glutathione depletion. PFOS also damaged mitochondrial and
lysosomal membranes and was associated with significantly increased levels of cellular
proteolysis and caspase 3 activity. These findings suggest that PFOS could mediate
immunosuppressive effects through direct cytotoxicity of lymphocytes.
3.4.2.3.2.2 Mechanistic data informing autoimmune diseases
As described in Section 3.4.2.1, two studies reported that PFOS levels in healthy controls were
either higher than in ulcerative colitis (UC) cases {Steenland, 2018, 5079806} or lower than in
multiple sclerosis (MS) cases {Ammitzb0ll, 2019, 5080379}. While no mechanistic studies
directly investigated the mechanism by which PFOS could promote the development of
autoimmunity, one study evaluated PFOS effects on TH17 cells, implicated in the
pathophysiology of both MS and UC {Chen, 2020, 10284264; Fu, 2020, 10284269}. Suo et al.
(2017, 3981310) examined the effects of 2 mg/kg PFOS in a mouse model of Citrobacter
rodentium infection. PFOS was administered for 7 days by oral gavage before mice were
infected with C. rodentium and throughout the early and late phases of infection. Large intestinal
lamina proprial lymphocytes were isolated 5 days after infection and analyzed by flow cytometry
after treatment with immune stimulators. Levels of IL-17 and IL-22 produced by Thl7 cells were
significantly elevated in PFOS-treated mice compared to the control group. These findings
support that PFOS-mediated effects on pathogenic TH17 cells may impact development of
autoimmune diseases as well as bacterial infections of the gut.
3.4.2.3.2.3 Mechanistic data informing allergic responses
Several studies were identified that evaluated associations between PFOS exposure and immune
hypersensitivity, including asthma, allergy, and eczema as described in Section 3.4.2.1.2. Five
new mechanistic studies informed allergy and asthma. Oulhote et al. (2017, 3748921) observed a
significant association between PFAS exposures and increased basophil counts between birth
and age 5 in human children. Although PFAS exposure was analyzed collectively (included
PFOA, PFOS, PFHxS, PFNA, and perfluorodecanoic acid [PFDA]), PFOS showed the highest
serum concentrations at all ages. The authors suggested that enhanced basophil levels could be
associated with dysregulated allergic and asthma-related responses, possibly by promoting TH2-
type responses.
Zhu et al. (2016, 3360105) evaluated 231 asthmatic children and 225 non-asthmatic control
children from Northern Taiwan. A significant positive association was identified for PFOS blood
levels and TH2 cytokines while a non-significant inverse association was found for TH1
cytokines among asthmatic children. Male asthmatics exhibited elevated IgE levels with
increasing PFOS levels. Also, in males only, significant positive associations between PFOS
levels in blood and TH2:TH1 cytokine ratios were observed for both the IL-4/IFN-y ratio and IL-
5/IFN-y ratio. This finding suggests that PFOS may exacerbate asthma by altering availability of
key TH1 and TH2 cytokines. However, the effects of PFOS on TH1- and TH2-type cytokine
profiles may be dependent on disease context or the cell types under study. For example, in
earlier studies of human peripheral blood leukocytes (PBLs) treated with phytohemagglutinin
(PHA), PFOS exposure led to diminished IL-4, IL10, and IFN-y {NTP, 2016, 5080063; Corsini,
2011, 1937246; Corsini, 2012, 1937239}.
Lee et al. (2018, 5085013) used an albumin-induced active systemic anaphylaxis model to
evaluate type I hypersensitivity in mice. After sensitization with ovalbumin (OVA), PFOS (50-
3-115
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
150 mg/kg) was orally administered on days 9, 11, and 13. On day 14, OVA was administered
by intraperitoneal (IP) injection, and mice were evaluated for signs of allergy. PFOS
significantly aggravated allergic symptoms such as hypothermia and significantly increased
serum histamine, TNF-a, IgE, and IgGl relative to controls. Further findings suggest the
mechanism of aggravated allergic responses mediated by PFOS is through release of histamine
and P hexosaminidase associated with up-regulation of intracellular calcium in IgE-stimulated
mast cells. Elevated levels of inflammatory cytokines (TNF-a, IL-ip, IL-6, and IL-8) were also
observed in PFOS-exposed non-sensitized rat basophilic leukemia cells, which were linked to
NF-kB activation. Together, these findings provide a plausible pathway for PFOS-mediated
exacerbation of allergic responses.
3.4.2.3.2.4 Mechanistic Evidence for PFOS-mediated Effects on Innate
Immune Responses
As described in Sections 3.4.2.2.3 and 3.4.2.2.4, several studies in animals suggest PFOS may
negatively impact NK cells and macrophage function, indicating innate immune effector cells are
susceptible to perturbations by PFOS. Very few studies were identified that evaluated the
mechanisms by which PFOS may alter innate immunity and no studies evaluated the
mechanisms by which PFOS alters NK cell activity. Among the studies reporting NK activity in
Table 3-7 in section 3.4.2.2.4, most studies observed decreased NK activity, though at least one
study observed enhanced NK responses at low doses of exposure {Dong, 2009, 1424951}. In all
of these studies, NK cells were obtained from animals exposed in vivo and analyzed in vitro
using target cells that were not exposed to PFOS, suggesting PFOS directly alters NK maturation
or activity. Whether PFOS alters the spectrum of activating and inhibiting receptors on NK cells
or some other aspect of NK activity is not known. At least one study treated NK and target YAC-
1 cells in vitro, though neither NK receptor nor ligand expression were evaluated {Wirth, 2014,
1937219}. Thus, an important outstanding mechanistic question that may directly impact
observations of dose- and sex-dependent effects is whether PFOS alters expression of NK cell
receptors or target cell ligands for NK receptors.
Two studies were identified that evaluated mechanisms of PFOS activity on innate immune
responses mediated by macrophages, and one evaluated PFOS effects on gut immunity and
innate lymphoid cells (ILC3). Rainieri et al. (2017, 3860104) measured PFOS effects in TREM-
like transcript (TLT) cells, a human macrophage-derived cell line. Treatment of cells with 15.6-
500 mg/L PFOS for 24 hours increased cell viability relative to controls, which was associated
with a significant decrease in the number of apoptotic cells. Using non-confluent cell cultures,
500 mg/L PFOS treatment significantly decreased the number of cells in the G2/M phase. PFOS
treatment significantly increased ROS production. However, Berntsen et al. (2018, 4167035)
found no PFOS-specific effects on macrophage phagocytosis in primary cells including
peritoneal macrophages (PCM) from adult Wistar rats and C57B1/6 mice, non-obese diabetic
(NOD) mice, IL-1 knockout (KO) mice, and newly born rats. In addition, PFOS did not alter
phagocytosis in human or rat monocyte-derived macrophages (MDM). Taken together, these
limited findings suggest that while PFOS does not alter macrophage function, it may affect
viability and induce ROS and lipid peroxidation in macrophage cell lines.
Suo et al. (2017, 3981310) examined effects of PFOS in a mouse model of C. rodentium
infection. PFOS at 2 mg/kg or vehicle control was administered for 7 days before infecting mice
with C. rodentium and throughout the observation period of infection. Part of this study
3-116
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
evaluated effects on ILC3s, which have been suggested to be important in controlling C.
rodentium at the early phase of infection prior to induction of adaptive immune responses. ILC3s
secrete IL-17 and IL-22 that act to stimulate epithelial cells to secrete anti-microbial peptides or
through recruitment of neutrophils {Takatori, 2009, 9811595; Zheng, 2008, 10284265; Ishigame,
2009, 10284267}. PFOS inhibited the expansion of C. rodentium by promoting IL-22
production in ILC3 cells in an aryl hydrocarbon receptor (AhR)-dependent manner. However,
PFOS also led to decreased mucin production from goblet cells, which may contribute to the
observation that PFOS altered the gut microbiome. Specifically, PFOS-exposed mice at late
stages of infection exhibited decreased levels of Lactobacillus casei and Lactobacillus johnsonii,
and increased levels of E. coli. The authors crossed Ahrf/f mice (in which the Ahr gene is
flanked by loxP sites) to mice in which the ere recombinase gene is driven by the RAR-related
orphan receptor gamma promoter (RORc-cre) to delete Ahr in ILC3 and T cells (Ahrf/f RORc-
cre). Cells isolated from either Ahrf/f RORc-cre or Ahrf/f mice were exposed to PFOS, and
cytokines were analyzed using flow cytometry. PFOS-exposed mice exhibited increased IFN-y
production from CD3- non-T cells compared to control mice, indicating a pro-inflammatory role
of PFOS. Taken together, PFOS-associated dysbiosis and persistent inflammation in the intestine
ultimately led to a failure to clear C. rodentium at the late phase of infection. These findings
suggest PFOS may impact gastrointestinal health in animals (See PFOS Appendix) and raises the
possibility that immune mechanisms associated with AhR activation are disrupted by PFOS.
3.4.2.3.2.5 Mechanistic Evidence for PFOS-mediated Effects on Intrinsic
Cellular Defense Pathways
There is limited evidence of PFOS exposure related to the disruption of intrinsic cellular defense
pathways. S0rli et al. (2020, 5918817) used HBEC3-KT human bronchial epithelial cells to study
inflammatory changes in response to PFOS, including modulation of the inflammatory response
induced by polyinosinic:polycytidylic acid (Poly I:C), a toll-like receptor 3 (TLR3) ligand. In
cells exposed to 30 or 60 |iM PFOS for 48 hours, IL-la/p release was elevated, indicative of a
pro-inflammatory response. In cells treated with 5 [j,g/mL poly I:C for 3 hours followed by
exposure to 10 |iM PFOS for 48 hours, release of the chemokines CXCL8 and CXCL10 was
suppressed, but IL-1 a/p release was enhanced. The authors hypothesized that IL- p release may
be related to the fact that it requires only proteolytic cleavage of preformed IL-1 in the cytosol,
and thus may not be dependent on TLR3-dependent gene expression. The authors also
hypothesized that PFOS may inhibit NF-kB activation in a cell type-dependent manner in the
lung. TLR3 stability and/or function, other double-stranded RNA sensors in these cells, or
associated signal transduction pathways were not evaluated. These results indicate that PFOS can
exert divergent effects on chemokine and cytokine release in a dose-dependent manner in human
bronchial epithelial cells and modulates the activity of intrinsic cellular defense responses
mediated by toll receptors and/or other double-stranded RNA sensors.
3.4.2.3.2.6 Mechanistic Evidence for PFOS-mediated Effects on Inflammation
PFOS-mediated effects on inflammation may impact a wide range of diseases given that chronic
inflammation can be a key driver of many diseases such as cancer, cardiovascular, metabolic,
and neurological diseases {Hunter, 2012, 10284266}. Earlier studies suggest that PFOS
differentially impacts pro-inflammatory cytokine release in a cell type and tissue-specific
manner. For example, as described in 2016 PFOS HESD {U.S. EPA, 2016, 3603365}, cells
isolated from the peritoneal cavity and bone marrow, but not spleen, of mice exposed to high
3-117
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
levels of PFOS had enhanced levels of the pro-inflammatory cytokines, TNF-a and IL-6, in
response to stimulation by lipopolysaccharide (LPS). The levels of these cytokines in the serum
were not elevated {Qazi, 2009, 1937259}. Since the 2016 document, 9 additional mechanistic
studies reported correlations between PFOS exposure and modulation of pro-inflammatory
cytokines or serum markers of inflammation. Consequences of PFOS exposure are not consistent
across species and are summarized in Table 3-9. Pro-inflammatory cytokines were elevated in
PFOS-exposed rodents and in human and animal cells in culture. In both studies evaluating
human subjects {Bassler, 2019, 5080624; Mitro, 2020, 6833625}, either no significant changes
were observed in serum cytokine or marker levels (IL-6, IFN-y, C-reactive protein [CRP], or
C3a) or levels were reduced (TNF-a, IL-8) relative to subjects with lower PFOS exposures.
Table 3-9. Effects of PFOS Exposure on Pro-Inflammatory Cytokines and Markers of
Inflammation
Study
„ . „ „ Cytokine or
Species or Cell T x.
T Inflammatory
ype Marker
Matrix and Measurement
Direction of Change
Following PFOS
Exposure
Mitro et al. (2020, Human females 3 IL-6
blood protein (ELISA)
None
6833625)
years postpartum,
Project Viva CRp
blood protein
(immunoturbidimetric high-
sensitivity assay)
None
Bassler et al.
Human males andIL-6
serum protein
None
(2019, 5080624)
females, C8
Health Project
TNF-a
(Multispot Immunoassay)
serum protein
(Multispot Immunoassay)
1
IL-8
serum protein
(Multispot Immunoassay)
1
IFN-y
C3a
serum protein
(Multispot Immunoassay)
serum protein (ELISA)
None
1
Li et al. (2020,
6833655)
Human IL-1
lymphocytes
culture supernatant protein
(ELISA)
T
IL-6
culture supernatant protein
(ELISA)
T
Sorli et al. (2020,
5918817)
Human bronchial IL-1 a
epithelial cell
culture supernatant protein
(ELISA)
T
line IL-1 (3
culture supernatant protein
(ELISA)
T
Liao et al. (2013,
Human umbilical IL-6
cellular mRNA (qRT-PCR)
T
1937227)
vein
endothelial cells
(HUVECs) IL-1 (3
cellular mRNA (qRT-PCR)
T
Han et al. (2018,
4355066)
Sprague-Dawley IL-6
male rats
TNF-a
serum protein (ELISA)
serum protein (ELISA)
T
T
Su et al. (2019,
5080481)
ICR male mice IL-6
TNF-a
serum protein (ELISA)
serum protein (ELISA)
T
T
3-118
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Study
Species or Cell
Type
Cytokine or
Inflammatory
Marker
Matrix and Measurement
Direction of Change
Following PFOS
Exposure
Han et al. (2018,
4355066)
Primary rat
hepatocytes and
Kupffer cells
IL-6
cellular mRNA (PCR) and
culture supernatant protein
(ELISA)
T
TNF-a
cellular mRNA (PCR) and
culture supernatant protein
(ELISA)
T
Zhuetal. (2015,
Murine
IL-6
cellular mRNA (PCR) and
T
2850996)
microglial cell
line
culture supernatant protein
(ELISA)
TNF-a
cellular mRNA (PCR) and
culture supernatant protein
(ELISA)
T
Notes: C3a = cohort 3a; CRP = C-reactive protein; ELISA = enzyme-linked immunosorbent assay; IL-la = interleukin 1 alpha;
IL-ip = interleukin 1 beta; IL-6 = interleukin 6; IL-8 = interleukin 8; PCR = polymerase chain reaction; TNF-a = tumor necrosis
factor alpha; qRT-PCR = quantitative reverse transcription polymerase chain reaction.
3.4.2.3.2.6.1 Animal Toxicological Studies
Han et al. (2018, 4355066) investigated PFOS-effects on hepatic inflammation in male Sprague-
Dawley (SD) rats exposed to 1 or 10 mg/kg body weight PFOS by gavage and in isolated
primary rat Kupffer cells cultured in vitro. In vivo, PFOS induced Kupffer cell activation and
elevated serum TNF-a and IL-6 and stimulated release of these cytokines from cultured primary
Kupffer cells in vitro. Studies with a Kupffer cell-blocking and depleting agent, gandolinium
chloride (GdCL3), demonstrated that PFOS exposure stimulated Kupffer cell release of TNF-a
and IL-6 in vivo (measured by ELISA) and in vitro (increased mRNA expression measured by
PCR and protein expression measured by ELISA). Furthermore, Kupffer cell activation was
mitigated by treatment with anti-TNF-a or anti-IL-6 antibodies. In vivo, PFOS exposure
upregulated the protein expression of proliferating cell nuclear antigen (PCNA), c-Jun, c-MYC,
and Cyclin D1 (CyDl) in liver, a finding mirrored in Kupffer cells cultured in vitro. Treatment
with a drug inhibitor of NF-kB (pyrrolidine dithiocarbamate [PDTC]) and a c-Jun N-terminal
kinase (INK) inhibitor (SP600125) significantly inhibited production of PFOS-induced TNF-a
and IL-6. Together, these findings suggest that PFOS induces Kupffer cell activation, leading to
NF-kB/TNF- a/IL-6-dependent hepatocyte proliferation.
Su et al. (2019, 5080481) also examined liver-specific immunotoxicity. Male ICR mice were
dosed with 10 mg/kg/day for 21 days. TNF-a and IL-6 were significantly elevated, whereas
fibroblast growth factor 21 (FGF21) was significantly reduced in sera from these mice. Co-
treatment with 200 mg/kg per day of vitamin C led to a significant reversal in PFOS-induced
changes in serum TNF-a, IL-6, and FGF21, consistent with results of immunostaining for TNF-a
and FGF21 in liver cells. The mechanism by which vitamin C exerts protection from
inflammatory responses in this model was not elucidated.
3.4.2.3.2.6.2 In Vitro Studies
Four studies demonstrated increased inflammatory cytokine expression in human cells cultured
in vitro. PFOS exposure at concentrations of >30 |iM led to increased IL-la/p release in HBEC3-
KT human bronchial epithelial cells {S0rli, 2020, 5918817}. Li et al. (2020, 6833655)
demonstrated induction of IL-1 and IL-6 in human lymphocytes that were isolated from human
3-119
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
donors and exposed in culture to 50 mM PFOS for 72 hours. Gimenez-Bastida and Surma (2015,
3981569) investigated inflammatory cytokine responses in human CCD-18 Co myofibroblasts as
a model of colonic subepithelial myofibroblasts in the intestinal lamina propria. Cells were
exposed to PFOS at concentrations ranging from 0.6 to 100 |iM in combination with IL-ip (1
ng/mL). Exposure to PFOS reduced IL-ip-induced IL-6 production at all doses except 100 [xM,
but this reduction only reached significance at 6 |iM, Liao et al. (2013, 1937227) pretreated
human umbilical cord endothelial cells (HUVECs) with 100 mg/L PFOS for 5 hours and then co-
treated with polyphenols (Flos Lonicerae extract and chlorogenic acid) for 24 or 48 hours. PFOS
exposure resulted in increased levels of mRNA transcripts for inflammatory cytokines (IL-ip,
IL-6) as well as COX-2 (cyclooxygenase 2) and NOS3 (nitric oxide synthase 3), the protein
products of which function in cellular defense and prostaglandin synthesis. PFOS exposure also
led to upregulation of transcripts for adhesion molecules P-Selectin (SELP) and ICAM1
(intercellular adhesion molecule 1). Functionally, PFOS treatment for 48 h increased adhesion of
THP-1 monocytes to HUVECs. These PFOS-mediated changes in HUVECs were mitigated by
co-treatment of cells with polyphenols.
In immortalized murine BV2 microglial cells, which are brain resident macrophage-like cells
that are considered central to inflammatory responses in the brain, PFOS exposure increased
inflammatory cytokine expression {Zhu, 2015, 2850996} via similar pathways observed in
primary rat hepatocytes and Kupffer cells exposed to 100 |iM PFOS {Han, 2018, 4355066}. Zhu
et al. (2015, 2850996) reported that treatment with 10 |iM PFOS for 6 hours resulted in increased
levels of Tnfa and 116 gene expression. Time course studies were performed using 1 [xM PFOS
and indicated that elevated Tnf-a and IL-6 mRNA expression occurs within 1 hour, peaks at 3
hours, and begins to diminish by 6 hours of PFOS exposure. Protein levels of these cytokines in
culture supernatant continually increased with 6, 12, and 24 hours of 1 |iM PFOS treatment.
Transcriptional activation of TNF-a and IL-6 correlated with activation of NF-kB (measured by
immunoblot of the phosphorylated form) and was mitigated by targeting INK and the
extracellular regulate kinase (ERK1/2) with a drug inhibitor (SP600125) or blocker (PD98059).
Together, the data support a role for MAPK signaling pathways and NF-kB activation in PFOS-
mediated inflammatory gene expression in cultured microglial cells and primary Kupffer cells.
In addition to activation of MAPK signal transduction pathways, epigenetic mechanisms may
impact inflammatory gene expression mediated by PFOS. Park et al. (2019, 5412425) found
increased gene expression of sirtuin (SIRT) genes in RAW 264.7 macrophage cells (cell line
derived from BALB/c mice). The SIRT family of proteins act to deacetylate the lysine residues
of histone proteins, but they also can deacetylate nonhistone substrates, such as inflammation-
related transcription factors including NF-kB (Frescas, 2005, 10284417; Yeung, 2004,
10284418}. PFOS exposure increased expression of Sirt2, Sirt3, Sirt5, and Sirt6. The authors did
not investigate the effect of increased expression of Sirt genes observed after PFOS on the
acetylation status or expression of inflammatory proteins.
3.4.2.3.2.6.3 Human Studies
Bassler et al. (2019, 5080624) examined 200 adult participants of the C8 Health Project to test
the hypothesis that environmental perfluoroalkyl acids (PFAAs) are associated with increased
hepatocyte apoptosis and decreased pro-inflammatory cytokines in serum. In support of this
hypothesis, PFOS levels were associated with significantly reduced serum TNF-a and IL-8
serum levels. However, there was no correlation between PFOS serum levels and other cytokines
3-120
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
(IL-6, IFN-y), inflammatory markers (cleaved complement C3a) or markers of hepatocyte cell
death (caspase 3 cleaved cytokeratin 18). The authors hypothesized that under certain
circumstances such as with nonalcoholic fatty liver disease (NAFLD), PFAAs are associated
with immunotoxic suppressive effects on innate immunity and inflammation.
Mitro et al. (2020, 6833625) set out to evaluate PFAS exposures and cardiometabolic health in
pregnant women and in the years postpartum as part of Project Viva. The study obtained 3-year
postpartum anthropometry measurements and blood biomarker measurements of inflammation
including IL-6 and CRP. While exposure to some PFAS was associated with elevated IL-6 levels
3 years postpartum, no significant associations were observed for PFOS. None of the PFAS
chemicals examined other than 2-(N-methyl-perfluorooctane sulfonamido) acetic acid
(MeFOSAA) showed a strong association with CRP levels in this study.
3.4.2.3.2.7 Summary
Since publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}, new mechanistic
information has emerged informing immune system physiology, innate and adaptive immune
functions, intrinsic cellular defense, and inflammation. Earlier studies summarized in the 2016
PFOS HESD {U.S. EPA, 2016, 3603365} linked PFOS-mediated PPARy activation to decreased
spleen and thymus weight and reduced spleen and thymus cellularity {Yang, 2002, 1332453;
NTP, 2016, 4613766}. Recent studies such as Zhong et al. (2016, 3748828) suggest a role for
PFOS in disrupting spleen and thymic weights and cellularity through sex hormones, activation
of MAPK signaling pathway and/or oxidative stress genes associated with apoptosis in
lymphocytes {Lv, 2015, 3981558}, and reduced numbers of myeloid, pro/pre-B, immature B,
and early mature B cells in bone marrow {Qazi, 2012, 1937236}.
New mechanistic insights into PFOS-mediated suppression of adaptive immune responses
include PFOS-mediated effects on THl/TH2-type cytokines and IgE titers in response to
allergens in mice and humans {Zhong, 2016, 3748828; Zhu, 2016, 3360105}, glycosylation of
immunoglobulins in humans {Liu, 2020, 6833599}, and lymphocyte toxicity in vitro {Zarei,
2018, 5079848}. Effects of PFOS exposure on allergy {Lee, 2018, 5085013} included release of
histamine and P hexosaminidase associated with up-regulation of intracellular calcium in IgE-
stimulated mast cells and release of inflammatory cytokines linked to NF-kB activation. PFOS
was also found to stimulate release of IL-17 and IL-22 from TH17 cells in an animal model of
intestinal infection {Suo, 2017, 3981310}. Additional insights were provided by transcriptomic
and lipidomic studies {Lv, 2015, 3981558; Li, 2020, 6833655; Pennings, 2016, 3352001}.
Transcriptomic studies identified candidate genes that may mediate immunotoxicity in children
exposed in utero to PFOS including SHC4, PPARD, CYTL1, IL-27, and ADORA2A {Pennings,
2016, 3352001}. In mice, PFOS exposure upregulated THEMIS and CD3G and altered calcium
homeostasis, cell cycle genes that may impact T cell immunophenotypes observed in spleen, and
T cell function through inhibition of T cell proliferation and induction of T cell anergy {Lv,
2015, 3981558}.
With respect to innate immune responses, PFOS is associated with a depression of NK cell
activity. An important outstanding mechanistic question that may directly impact observations of
dose- and sex-dependent effects is whether PFOS alters NK cells directly or influences NK cell
receptor ligand expression on potential target cells. Two new studies evaluated mechanisms of
PFOS activity on innate immune responses mediated by macrophages and ILC3 {Rainieri, 2017,
3-121
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3860104; Berntsen, 2018, 4167035}. Together, these findings suggest that while PFOS does not
alter macrophage function, it may induce ROS and lipid peroxidation in macrophage cell lines.
Also, Suo et al. (2017, 3981310) examined effects of PFOS in a mouse model of C. rodentium
infection. PFOS inhibited the expansion of C. rodentium by promoting IL-22 production in ILC3
cells in an AhR-dependent manner and increased IFN-y production from CD3- non-T cells
compared to control mice.
Very little information is available regarding whether PFOS impacts intrinsic cellular defenses.
One recent study, S0rli et al. (2020, 5918817), demonstrated that PFOS exerts divergent effects
on chemokine and cytokine release in a dose-dependent manner in human bronchial epithelial
cells. This study also proposed that PFOS can modulate the activity of intrinsic cellular defense
responses mediated by toll receptors and/or other double-stranded RNA sensors.
Nine recent studies reported correlations between PFOS exposure and modulation of pro-
inflammatory cytokines or serum markers of inflammation; however, the inflammatory
responses to PFOS exposure are not consistent across species. Pro-inflammatory cytokines were
elevated in PFOS-exposed rodents and in human and animal cells in culture through activation of
MAPK signaling pathways and activation of NF-kB {Han, 2018, 4355066; Zhu, 2015,
2850996}. In contrast, the available studies evaluating human subjects observed either no
changes in serum cytokine or marker levels (IL-6, IFN-y, or CRP) or reduced levels (TNF-a, IL-
8, or C3a) relative to subjects with lower PFOS exposures.
Despite recent research informing a range of immunotoxicity endpoints, a comprehensive
understanding of the mechanisms by which PFOS alters immune system development,
physiology, and function is lacking. Data from transcriptomic studies have advanced the
understanding regarding the potential of PFOS to disrupt lymphocyte signaling and function. A
particularly promising area of research relates to the observation that PFOS exposure in human
lymphocytes is associated with dysregulated lipid profiles that encompass glycerophospholipid
metabolism, sphingolipid metabolism, glycerolipid metabolism, adipocytokine signaling,
regulation of autophagy, and arachidonic acid metabolism {Li, 2020, 6833655}. However,
further studies are needed to determine if these gene expression changes result in altered protein
accumulation and if gene expression and lipid profile changes mediate functional changes in
immunity.
3.4.2.4 Evidence Integration
There is moderate evidence for an association between PFOS exposure and immunosuppressive
effects in human studies based on largely consistent decrease in antibody response following
vaccinations (against three different infectious agents) in multiple medium confidence studies in
children. Reduced antibody response is an indication of immunosuppression and may result in
increased susceptibility to infectious disease. Changes in antibody levels of 10-20% per
doubling of PFOS exposure were observed in the Faroe Islands cohorts. The variability in the
results, including null and positive associations, could be related to differences in sample sizes,
individual variation, vaccine type, and differences in timing of the boosters, as well as
differences in timing of antibody measurements in relation to the last booster. However, these
factors cannot be explored further with currently available data. Overall, the evidence indicates
an association between increased serum PFOS levels and decreased antibody production
3-122
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
following routine vaccinations in children. Evidence in adults does not indicate an association
with immunosuppression, but high confidence studies are not available.
There is slight evidence for sensitization and allergic responses from studies in humans, but
notable limitations and uncertainties in the evidence base remain. Associations in
epidemiological studies measuring PFOS exposure and hypersensitivity outcomes were mixed.
There is some evidence from epidemiological studies of an association between PFOS exposure
and asthma, but there is considerable uncertainty due to inconsistency across studies and sub-
groups. Sex-specific differences were reported in multiple studies, but there was inconsistency in
the direction of association within each sex. There is not an obvious pattern of results by analysis
of "ever" vs. "current" asthma, and no studies beyond the Dong et al. (2013, 1937230) described
in the 2016 Health Assessment examined asthma incidence. For allergy and eczema outcomes,
results were inconsistent across studies.
There is limited evidence of an association between PFOS exposure and infectious diseases.
While one medium confidence study reported higher odds of total infectious diseases, results
from studies examining individual diseases including respiratory infections, chickenpox, cough,
RSV, common cold, ear infections, and urinary tract infections were inconsistent.
Epidemiological evidence on autoimmune effects was limited to three studies reporting on
different autoimmune conditions. Similar to the findings from the 2016 Health Assessment, there
was insufficient information to draw conclusions on the effect of PFOS exposure on autoimmune
disease.
The animal evidence for an association between PFOS exposure and immunosuppressive
responses is moderate based on decreased PFC responses and NK cell activities observed in 12
high or medium confidence rodent studies. Additionally, fluctuations in splenic and thymic cell
populations and increased bone marrow hypocellularity in conjunction with extramedullary
hematopoiesis were observed. Extramedullary hematopoiesis, blood cell production outside of
the bone marrow, occurs when normal cell production is impaired. Bone marrow hypocellularity
in parallel with extramedullary hematopoiesis suggest that PFOS impedes hematopoiesis in the
bone marrow. As such, EPA concluded that elevated extramedullary hematopoiesis and bone
marrow hypocellularity, as well as reduced ability to generate an immune response to a bacteria-
like challenge and reduced PFC response indicate toxicity of relevance to humans exposed to
PFOS.
It is clear that PFOS can alter immune cells and signaling in experimental systems. However, the
connection between various alterations to immune and inflammation signaling and immunologic
effects reported in humans is not clear. Transcriptomics data represent some of the most
informative findings in regard to potential underlying mechanisms of immunotoxicity of PFOS.
Together, the findings from transcriptomic and functional analyses reported in human
lymphocytes exposed to PFOS, in human cord blood samples from gestational exposure to
PFOS, and in mice treated with PFOS suggest that PFOS exposure may disrupt adaptive
immunity through the dysregulation of genes and lipids involved in lymphocyte survival,
proliferation, and inactivation. PFOS effects on gene expression paralleled a dose-dependent
increase in intracellular free calcium (which plays an important role in immune cell proliferation
in response to foreign antigens) concentration in splenocytes of mice treated with PFOS,
suggesting that activation of MAPK signaling pathway and/or oxidative stress genes in response
3-123
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
to PFOS may alter splenic architecture via induction of apoptosis in lymphocytes. Relatedly,
additional in vitro transcriptomic data collected from mouse microglial cells and rat hepatocytes
and Kuppfer cells demonstrate activation of TNF-a and IL-6, correlated with activation of NF-
kB. These data support a role for MAPK signaling pathways and NF-kB activation in PFOS-
mediated inflammatory gene expression in vitro. TNF-a, IL-6, and NF-kB are all related to
inflammation, allergy, and other immune responses.
Despite recent research informing a range of immunotoxicity endpoints, a comprehensive
understanding of the mechanisms by which PFOS alters immune system development,
physiology, and function is lacking. A particularly promising area of research relates to the
observation that PFOS exposure in human lymphocytes is associated with dysregulated lipid
profiles that encompass glycerophospholipid metabolism, sphingolipid metabolism, glycerolipid
metabolism, adipocytokine signaling, regulation of autophagy, and arachidonic acid metabolism.
Additional research is needed to determine if these gene expression changes result in altered
protein accumulation and if gene expression and lipid profile changes mediate functional
changes in immunity; specifically, alterations to antibody response and susceptibility to
infection, as reported in humans.
3.4.2.4.1Evidence Integration Judgment
Overall, considering the available evidence from human, animal, and mechanistic studies, the
evidence indicates that PFOS exposure is likely to cause adverse immune effects, specifically
immunosuppression, in humans under relevant exposure circumstances (Table 3-10). The hazard
judgment is driven primarily by consistent evidence of reduced antibody response from
epidemiological studies at levels of 0.8 ng/mL PFOS (median exposure in studies observing an
adverse effect). The evidence in animals showed coherent immunomodulatory responses at doses
as low as 0.0017 mg/kg/day that are consistent with potential immunosuppression and supportive
of the human studies, although issues with overt organ/systemic toxicity raise concerns about the
biological significance of some of these effects. While there is some evidence that PFOS
exposure might also have the potential to affect sensitization and allergic responses in humans
given relevant exposure circumstances, the human evidence underlying this possibility is
uncertain and with limited support from animal or mechanistic studies. Based on the antibody
response data in humans, children and young individuals exposed during critical developmental
windows may represent a potential susceptible population for the immunosuppressive effects of
PFOS. The absence of additional epidemiological studies or any long-term/chronic exposure
studies in animals examining alterations in immune function or immune-related disease
outcomes during different developmental life stages represents a major source of uncertainty in
the Immunotoxicity database of PFOS.
3-124
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 3-10. Evidence Profile Table for PFOS Immune Effects
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
Evidence from Studies of Exposed Humans (Section 3.4.2.1)
Immunosuppression
1 High confidence study
16 Medium confidence
studies
6 Low confidence studies
2 Mixed1 confidence
studies
Studies conducted in the
Faroe Islands examined
antibody levels among
children at various
timepoints compared to
exposure measured
prenatally and throughout
childhood. Lower
antibody levels against
tetanus and diphtheria
were observed in children
at birth, 18 months, age 5
years (pre-and post-
booster), and at age 7
years, with some being
statistically significant.
Findings in the three
studies examining adults
and adolescents were less
consistent than children.
Infectious disease was
examined in 11 studies of
children. Studies
examining infections of
the respiratory system
observed some positive
associations (5/11),
although many findings
from other studies were
not precise. Findings for
infectious disease in adults
were mixed, with two
studies reporting
• High and
medium
confidence
studies
• Consistent
direction of
effect
• Coherence of
findings
between
antibody
response and
increased
infectious
disease
• Low confidence
studies
• Lmprecision of
findings
0©O
Moderate
Evidence for immune
effects is based on
decreases in childhood
antibody responses to
pathogens such as
diphtheria and tetanus.
Reductions in antibody
response were observed at
multiple timepoints in
childhood, using both
prenatal and childhood
exposure levels. An
increased risk of upper
and lower respiratory tract
infections was observed
among children, coherent
with findings of reduced
antibody response. There
was also supporting
evidence of increased risk
of asthma, eczema, and
autoimmune disease,
however, the number of
studies examining the
same type of autoimmune
disease was limited.
®©o
' Evidence Indicates (likely)
Primary basis and cross-
stream coherence:
Human data indicated
consistent evidence of
reduced antibody response.
Evidence in animals
showed coherent
immunomodulatory
responses that are
consistent with potential
immunosuppression and
supportive of the human
studies, although issues
with overt organ/systemic
toxicity raise concerns
about the biological
significance of some of
these effects. While there is
some evidence that PFOS
exposure might also have
the potential to affect
sensitization and allergic
responses in humans given
relevant exposure
circumstances, the human
evidence underlying this
possibility is uncertain and
with limited support from
animal or mechanistic
studies.
3-125
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Evidence Integration
Studies and Summary and Key Factors that Increase Factors that Decrease Evidence Stream Summary Judgment
Interpretation Findings Certainty Certainty Judgment
inconsistent results for
COVID-19 infections.
Human relevance and other
inferences:
Immune
Examination of immune
• High and
• Low confidence
Based on the antibody
hypersensitivity
hypersensitivity includes
medium
studies
response data in humans,
1 High confidence study
outcomes such as asthma,
confidence
• Lnconsistent
children and young
17 Medium confidence
allergies, and eczema.
studies
direction of effect
individuals exposed during
studies
Increased odds of asthma
• Consistent
between
critical developmental
3 Low confidence studies
were reported in most
direction of
subpopulations
windows may represent a
2 Mixed1 confidence
medium confidence
effect for
potential susceptible
studies
studies (6/9), although
asthma across
population for the
associations were often
medium
immunosuppressive effects
inconsistent by subgroups.
confidence
of PFOS. The absence of
Low confidence studies
studies
additional epidemiological
supported the findings of
studies or any long-
increased odds of asthma
term/chronic exposure
or higher exposure levels
studies in animals
among asthmatics,
examining alterations in
although results were not
immune function or
always consistent or
immune-related disease
precise. Seven studies
outcomes during different
examined allergies,
developmental life stages
rhinitis, or
represents a major source
rhinoconjunctivitis. Some
of uncertainty in the
positive associations (3/7)
immunotoxicity database of
were observed, although
PFOS.
this varied by outcome
timing and were at times
inconsistent. Significantly
increased odds of eczema
were observed in three
(3/7) studies for those in
the highest exposure
group, however, inverse
associations were also
observed (3/7).
3-126
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
Autoimmune disease
1 Medium confidence
study
3 Low confidence studies
Lower exposure levels
were observed in healthy
controls compared to
multiple sclerosis cases in
one study of adults. An
increased risk of celiac
disease was also observed
in a study of children and
young adults. Another
study observed lower
exposure levels among
ulcerative colitis cases
compared to health
controls. There was no
significant difference in
exposure levels based on
type 1 diabetes status.
• Medium
confidence
study
• Low confidence
studies
• Limited number of
studies examining
outcome
Evidence from In Vivo Animal Toxicological Studies (Section 3.4.2.2)
Immune response
4 Medium confidence
studies
In response to a SRBC
challenge, decreased IgM
response in the PFC assay
was reported (2/2) in a
subchronic and
developmental study in
mice, and was dose-
dependent in males. In the
developmental study, NK
cell activity was reduced
up to 8 weeks after a
gestational exposure (1/1).
One short-term study in
rats examined the effect of
PFOS on a delayed-type
hypersensitivity response
to a KLH challenge (1/1)
• Medium
confidence
studies
• Dose-response
relationship
seen within
multiple studies
Limited number of
studies examining
specific outcomes
0©O
Moderate
Evidence is based on
decreased immune
responses and NK cell
activities observed in
several high or medium
confidence rodent studies.
Additionally, fluctuations
in splenic and thymic cell
populations and increased
bone marrow
hypocellularity in
conjunction with
extramedullary
3-127
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
and observed no changes
inlgG levels (1/1) or
footpad swelling (1/1).
Another short-term study
observed no changes in
circulating white blood
cells but an increase in IgE
after an OVA challenge
(1/1).
Immune cellularity
2 High confidence
studies
6 Medium confidence
studies
Of the studies that
measured circulating
WBCs and differentials
(5/8), one short-term rat
study found decreases in
WBCs and segmented
neutrophils in males only,
while a chronic rat study
found increases in
segmented neutrophils in
males only. In another
short-term study in rats, a
negative trend for subsets
of T-cells and a positive
trend for B-cells were
observed in males. In
females a positive trend
was observed for WBCs,
lymphocytes, and subsets
of T-cells; a negative trend
was observed for B-cells.
No effects on WBCs or
differentials were seen in a
short-term study of male
mice and in a chronic
study in monkeys.
High and
medium
confidence
studies
Coherence of
findings across
circulating
immune cells,
splenic
cellularity, and
thymic
cellularity and
with
histopathologica
1 changes
Inconsistent
direction of effects
across studies and
sex
hematopoiesis were
observed. Extramedullary
hematopoiesis, blood cell
production outside of the
bone marrow, occurs
when normal cell
production is impaired.
Bone marrow
hypocellularity in parallel
with extramedullary
hematopoiesis suggest that
PFOS impedes
hematopoiesis in the bone
marrow.
3-128
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Evidence Integration
Studies and Summary and Key Factors that Increase Factors that Decrease Evidence Stream Summary Judgment
Interpretation Findings Certainty Certainty Judgment
Decreases in total spleen
cellularity and/or subsets
of splenic cells were
observed in 2 short-term
studies in male and female
rats and mice. Similar
decreases were seen in the
thymus in these studies;
however, no changes were
observed in females.
Histopathology
In 1 high confidence
• High and
• Inconsistent
1 High confidence
short-term study, a dose-
medium
direction of effects
studies
dependent increase in both
confidence
across studies and
5 Medium confidence
extramedullary
studies
sex
studies
hematopoiesis in the
spleen and hypocellularity
in the bone marrow was
observed in male and
female rats. No changes
were observed in the
thymus or lymph nodes.
None of the medium
confidence studies (5)
reported histopathologic
changes in the spleen (4),
• Dose-response
relationship
observed
• Coherent
changes with
those observed
in circulating
immune cells,
splenic
cellularity, and
thymic
thymus (2), or lymph
cellularity
nodes (2).
Organ weights
Mixed results were
• High and
• Inconsistent
2 High confidence
reported for absolute and
medium
direction of effects
studies
relative spleen (7) and
confidence
across species and
5 Medium confidence
thymus (5) weights. Both
studies
sex
studies
studies in male and female
rats reported decreases in
absolute spleen (2/2)
(males only) and thymus
• Confounding
variables such as
decreases in body
weights
3-129
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
weights (2/2) (males and
females), which generally
coincided with decreases
in body weights. Relative
spleen weights were
unchanged (2/2) or
increased (1/2) in rats,
while relative thymus
weights were unchanged
(1/2) or decreased (1/2). In
mouse studies, absolute
spleen and thymus weights
were not reported.
Decreased relative spleen
weights were observed in
mice (4/5); however, this
result was not always
consistent between sex
and timepoint. Relative
thymus weights were
decreased in male mice
(2/2) and unchanged in
female mice (1/1).
• Lack of dose-
response
relationship
Globulins and
immunoglobulins
1 High confidence
studies
4 Medium confidence
studies
Two short-term studies
found decreased globulin
levels (2/3) in male rats
and no changes in female
rats. One short-term study
found increases in subsets
of immunoglobulins (1/1)
in both male and female
rats, and one short-term
study found no changes in
IgE (1/1) in male mice.
• High and
medium
confidence
studies
Limited number of
studies examining
specific outcomes
Inconsistent
direction of effects
across sex
Mechanistic Evidence and Supplemental Information (Section 3.4.2.3)
3-130
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key Factors that Increase Factors that Decrease
Findings Certainty Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
Biological events or
pathways
Summary of Key Findings, Interpretation, and Limitations
Evidence Stream
Judgement
Immune system
development and
physiology
Effects on adaptive
immune responses
Key findings and interpretation:
• Changes in WBC and alterations in expression of immune and
inflammation-related genes in human cord blood have been reported.
• Reduction in immune organ weight, cellularity, and morphology (spleen
and thymus) in mice and rats.
• Disrupted splenic architecture and reduction in T helper and cytotoxic T
cells in the spleen in mice.
Limitations:
• No direct effects related to immune system development or physiology in
humans to anchor mechanistic findings.
Key findings and interpretation:
• Inverse association between PFOS exposure and vaccine-induced
antibody levels in human studies (in utero exposure to PFOS).
• Dysregulation of genes and lipids involved in lymphocyte survival,
proliferation, and anergy in vitro in human lymphocytes.
• Alterations to the expression of genes involved in adaptive immune
responses (i.e., immunological and/or hematopoietic functions) in cord
blood of samples from cases of maternal exposure to PFOS, as well as in
spleens of PFOS-exposed mice, and in human lymphocytes exposed to
PFOS in vitro.
PFOS can alter immune
cells and signaling in
experimental systems.
However, the connection
between various
alterations to immune and
inflammation signaling
and immunologic effects
reported in humans is not
clear.
Autoimmune diseases
Limitations:
• Association between gene expression changes and apical endpoints need
further confirmation.
Key findings and interpretation:
• PFOS-mediated effects on pro-inflammatory T helper cells, specifically
increased IL-17 and IL-22 production, in mice.
Limitations:
• Only a single study; no studies directly evaluated the mechanism by
which PFOS could promote autoimmunity.
3-131
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Evidence Integration
Studies and Summary and Key Factors that Increase Factors that Decrease Evidence Stream Summary Judgment
Interpretation Findings Certainty Certainty Judgment
Allergic responses Key findings and interpretation:
• Serum levels PFAS, including PFOS, was positively associated with
basophil counts in children between birth and age 5.
• PFOS blood levels were associated with alterations in cytokines in
asthmatic and non-asthmatic children, with some effects being specific to
asthmatic children.
Limitations:
• Human data include exposure to other PFAS in addition to PFOS.
Innate Immunity Key findings and interpretation:
• Conflicting results for NK cell activity across studies of cells from
animals exposed to PFOS in vivo.
• Alterations to apoptosis and cell cycle stage in a human macrophage-
derived cell line.
Limitations:
• Limited database, no human studies of innate immunity endpoints.
Effects on Intrinsic Key findings and interpretation:
Cellular Defense • PFOS-induced elevation of IL-la/p release, indicative of a pro-
Pathways inflammatory response, in human bronchial epithelial cells exposed in
vitro.
Limitations:
• Only a single study.
Effects on Key findings and interpretation:
Inflammation • Altered levels of pro-inflammatory cytokines or serum markers of
inflammation have been reported in humans, mice, and rats both in vivo
as well as in vitro.
• No association between PFOS exposure and increased acute or chronic
inflammatory responses in humans in vivo.
Limitations:
• Limited database.
3-132
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Notes: COVID-19 = coronavirus disease 2019; SRBC = sheep red blood cells; IgM = immunoglobulin M; PFC = plaque forming cell; NK = natural killer; KLH = keyhole limpet
hemocyanin; IgG = immunoglobulin G; IgE = immunoglobulin E; OVA = ovalbumin; WBC = white blood cells; IL-17 = interleukin 17; IL-22 = interleukin 22; IL-la/p =
interleukin 1 alpha/beta.
aStudies may be of mixed confidence due to differences in how individual outcomes within the same study were assessed (e.g., clinical test vs self-reported data).
3-133
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.3 Cardiovascular
EPA identified 106 epidemiological and 12 animal toxicological studies that investigated the
association between PFOS and cardiovascular effects. Of the 46 epidemiological studies
addressing cardiovascular endpoints, 4 were classified as high confidence, 24 as medium
confidence, 11 as low confidence, 3 as mixed (1 high/medium and 2 medium/low) confidence,
and 4 were considered uninformative (Section 3.4.3.1). Of the 80 epidemiological studies
addressing serum lipid endpoints, 2 were classified as high confidence, 29 as medium
confidence, 26 as low confidence, 16 as mixed (1 high/medium and 15 medium/low) confidence,
and 7 were considered uninformative (Section 3.4.3.1). Of the animal toxicological studies, 2
were classified as high confidence, 7 as medium confidence, 2 as low confidence, 1 and was
considered mixed {medium/low) (Section 3.4.3.2). Studies have mixed confidence ratings if
different endpoints evaluated within the study were assigned different confidence ratings.
Though low confidence studies are considered qualitatively in this section, they were not
considered quantitatively for the dose-response assessment (Section 4).
3.4.3.1 Human Evidence Study Quality Evaluation and Synthesis
3.4.3.l.lCardiovascular Endpoints
3.4.3.1.1.1 Introduction
Cardiovascular disease (CVD) is the primary cause of death in the United States with
approximately 12% of adults reporting a diagnosis of heart disease {Schiller, 2012, 1798736}.
Studied health effects include ischemic heart diseases (IHD), coronary artery disease (CAD),
coronary heart disease (CHD), hypertension, cerebrovascular disease, atherosclerosis (plaque
build-up inside arteries and hardening and narrowing of their walls), microvascular disease,
markers of inflammation (e.g., C-reactive protein), and mortality. These health outcomes are
interrelated—IHD is caused by decreased blood flow through coronary arteries due to
atherosclerosis resulting in myocardial ischemia.
The 2016 Health Advisory {U.S. EPA, 2016, 3982043} andHESD {U.S. EPA, 2016, 3603365}
assessments did not assess evidence for associations between CVD diseases and PFOS, besides
the review of its effects on serum lipids which are further described in subsequent sections.
There are 2 studies from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} that investigated the
association between PFOS and cardiovascular effects. Study quality evaluations for these 2
studies are shown in Figure 3-30.
The developmental section in the 2016 Health Advisory describes results from Geiger et al.
(2014, 2851286) which reported no association with hypertension in 1,655 children aged 12-18
years from the NHANES (1999-2000 and 2003-2008 cycles). An occupational study
{Alexander, 2003, 1291101} reported an inverse association for mortality from heart disease
among all cohort members. The decreased SMR was consistent in sensitivity analyses of cohort
members ever employed in a high exposure job and those only working in non-exposed jobs. The
study was considered low confidence due to concerns about healthy work effect and potential
residual confounding by smoking status and race/ethnicity.
3-134
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
,0^
kS®
VGO*°VS V*1 V
&
Alexander et al., 2003, 1291101 -
-
+
+
-
+
+
-
-
Geiger et al., 2014, 2851286 -
+
++
++
+
+
+
+
+
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-30. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS
and Cardiovascular Effects
Interactive figure and additional study details available on HAWC.
Since publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}, 44 new
epidemiological studies report on the association between PFOS and CVD, including outcomes
such as hypertension, CAD, congestive heart failure (CHF), microvascular diseases, and
mortality. Of these, 19 examined blood pressure or hypertension in adults. Pregnancy-related
hypertension is discussed in Section 3.4.3.1.1. All studies were conducted on the general
population with six {Honda-Kohmo, 2019, 5080551; Hutcheson, 2020, 6320195; Bao, 2017,
3860099; Mi, 2020, 6833736; Yu, 2021, 8453076; Ye, 2021, 6988486} conducted in a high-
exposure community in China (i.e., C8 Health Project and "Isomers of C8 Health Project"
populations), and three studies {Canova, 2021, 10176518; Pitter, 2020, 6988479;Zare Jeddi,
2021, 7404065} were conducted in a high-exposure community in Italy (i.e., Vento Region).
Different study designs were also used including three controlled trial studies {Cardenas, 2019,
5381549; Liu, 2018, 4238396; Osorio-Yanez, 2021, 7542684}, 11 cohort studies {Fry, 2017,
4181820; Donat-Vargas, 2019, 5080588; Lin, 2020, 6311641; Manzano-Salgado, 2017,
4238509; Matilla-Santander, 2017, 4238432; Mitro, 2020, 6833625; Warembourg, 2019,
5881345; Li, 2021, 7404102; Papadopoulou, 2021, 9960593}, one case-control study {Mattsson,
2015, 3859607}, and 33 cross-sectional studies {Bao, 2017, 3860099; Chen, 2019, 5387400;
Christensen, 2016, 3858533; Christensen, 2019, 5080398; Graber, 2019, 5080653; Honda-
Kohmo, 2019, 5080551; Huang, 2018, 5024212; Hutcheson, 2020, 6320195; Jain, 2020,
6311650; Jain, 2020, 6833623; Khalil, 2018, 4238547; Koshy, 2017, 4238478; Liao, 2020,
6356903; Lin, 2013, 2850967; Lin, 2016, 3981457; Lind, 2017, 3858504; Liu, 2018, 4238514;
Ma, 2019, 5413104; Mi, 2020, 6833736; Mobacke, 2018, 4354163; Yang, 2018, 4238462;
Averina, 2021, 7410155; Canova, 2021, 10176518; Jain, 2020, 6988488; Zare Jeddi, 2021,
7404065; Khalil, 2020, 7021479; Koskela, 2022, 10176386; Leary, 2020, 7240043; Lin, 2020,
3-135
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
6988476; Pitter, 2020, 6988479; Yu, 2021, 8453076; Ye, 2021, 6988486}. The two controlled
trial studies {Cardenas, 2019, 5381549; Liu, 2018, 4238396} were not controlled trials of PFAS
exposures, but rather health interventions: prevention of type 2 diabetes in Diabetes Prevention
Program and Outcomes Study (DPPOS) {Cardenas, 2019, 5381549; Osorio-Yanez, 2021,
7542684} and weight loss in the Prevention of Obesity Using Novel Dietary Strategies Lost
(POUNDS-Lost) Study {Liu, 2018, 4238396}. Thus, these studies could be interpreted as cohort
studies for evaluating cardiovascular risk purposes.
The studies were conducted in different study populations with the majority of studies conducted
in the United States {Cardenas, 2019, 5381549; Christensen, 2016, 3858533; Christensen, 2019,
5080398; Fry, 2017, 4181820; Graber, 2019, 5080653; Honda-Kohmo, 2019, 5080551; Huang,
2018, 5024212; Hutcheson, 2020, 6320195; Jain, 2020, 6311650; Jain, 2020, 6833623; Khalil,
2018, 4238547; Koshy, 2017, 4238478; Liao, 2020, 6356903; Lin, 2020, 6311641; Liu, 2018,
4238396; Liu, 2018, 4238514; Ma, 2019, 5413104; Mi, 2020, 6833736; Mitro, 2020, 6833625;
Jain, 2020, 6988488; Khalil, 2020, 7021479; Koskela, 2022, 10176386; Leary, 2020, 7240043;
Li, 2021, 7404102; Jain, 2020, 6988488; Osorio-Yanez, 2021, 7542684}. The remaining studies
were conducted in China {Bao, 2017, 3860099; Yang, 2018, 4238462; Yu, 2021, 8453076; Ye,
2021, 6988486}, Taiwan {Lin, 2013, 2850967; Lin, 2016, 3981457}, Spain {Manzano-Salgado,
2017, 4238509; Matilla-Santander, 2017, 4238432}, Croatia {Chen, 2019, 5387400}, Sweden
{Donat-Vargas, 2019, 5080588; Lind, 2017, 3858504; Mattsson, 2015, 3859607; Mobacke,
2018, 4354163}, Denmark {Jensen, 2020, 6833719}, Italy {Canova, 2021, 10176518; Ye, 2021,
6988486; Zare Jeddi, 2021, 7404065; Pitter, 2020, 6988479}, Norway {Averina, 2021,
7410155}, and two studies conducted in several European countries {Papadopoulou, 2021,
9960593; Warembourg, 2019, 5881345}. All the studies measured PFOS in blood components
(i.e., serum or plasma) with three studies measuring levels in maternal serum {Papadopoulou,
2021, 9960593; Li, 2021, 7404102;Warembourg, 2019, 5881345}, and four studies measuring
levels in maternal plasma {Papadopoulou, 2021, 9960593; Warembourg, 2019, 5881345;
Manzano-Salgado, 2017, 4238509; Mitro, 2020, 6833625}.
3.4.3.1.1.2 Study Quality
There are 45 studies from recent systematic literature search and review efforts conducted after
publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} that investigated the
association between PFOS and cardiovascular effects. Study quality evaluations for these 45
studies are shown in Figure 3-31 and Figure 3-32.
Of the 45 studies identified since the 2016 assessment, 4 studies were high confidence, 23 were
medium confidence, 10 were low confidence, 4 studies were mixed (1 high/medium due to
difference exposure estimates and 3 medium/low for different cardiovascular endpoints)
confidence, and 4 studies included an outcome considered uninformative {Jain, 2020, 6833623;
Jain, 2020, 6311650; Seo, 2018, 4238334; Leary, 2020, 7240043}. The main concerns with the
low confidence studies included the possibility of outcome misclassification (e.g., reliance on
self-reporting) in addition to the potential for residual confounding or selection bias (e.g.,
unequal recruitment and participation among subjects with outcome of interest, lack of
consideration and potential exclusion due to medication usage). Residual confounding was
possible due to socioeconomic status (SES), which can be associated with both exposure and the
cardiovascular outcome. Although PFOS has a long half-life in the blood, concurrent
measurements may not be appropriate for cardiovascular effects with long latencies. Further,
3-136
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
temporality of PFOS exposure could not be established for several low confidence studies due to
their cross-sectional design. Several of the low confidence studies also had sensitivity issues due
to limited sample sizes {Christensen, 2016, 3858533; Girardi, 2019, 6315730; Graber, 2019,
5080653;Khalil, 2018, 4238547}. Two studies were rated adequate for all domains, indicating
lower risk-of-bias; however, both studies treated PFOS as the dependent variable, resulting in
both studies being considered uninformative {Jain, 2020, 6833623; Jain, 2020, 6311650}.
Analyses treating PFOS as the dependent variable support inferences for characteristics (e.g.,
kidney function, disease status, race/ethnicity, etc.) that affect PFOS levels in the body, but it
does not inform the association between exposure to PFOS and incidence of cardiovascular
disease. Small sample size (n = 45) and missing details on exposure measurements were the
primary concerns of the remaining uninformative study {Leary, 2020, 7240043}. Studies
considered uninformative were not considered further.
3-137
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
1 1 1 1
Averina et al., 2021, 7410155-
+
+
+
+
+
+
+
+
Bao et al., 2017, 3860099-
+
+
+
+
++
+
+
+
Canovaetal., 2021, 10176518 -
+
+
+
++
+
+
+
+
Cardenas etal., 2019, 5381549-
+
++
+
+
+
+
+
+
Chen etal., 2019, 5387400-
+
++
+
+
+
+
+ *
Christensen et al.. 2016, 3858533 -
+
-
-
+
+
-
Christensen et al., 2019, 5080398-
+
+
+
+
+
+
+
Donat-Vargas et al., 2019, 5080588 -
+
+
++
+
++
+
+
+
Fry et al., 2017, 4181820-
++ ++
+
+
+
+
+
+
Graberetal., 2019, 5080653-
+
-
-
+
+
-
Honda-Kohmo et al., 2019, 5080551 -
+
+
-
-
+
+
+
Huang etal., 2018, 5024212-
++
++
+
+
+
+
+
+
Hutcheson et al., 2020, 6320195-
+
-
-
+
+
+
+
+
Jain et al., 2020, 6988488 -
+
+
+
+
+
4-
+
+
Jain, 2020, 6311650-
+
+
+
+
+
+
~
Jain, 2020, 6833623-
+
+
+
+
+
+
~
Khalil et al., 2018, 4238547-
+
+
-
+
+
-
Khalilet al., 2020, 7021479-
+
+
-
+
+
-
Koshy etal., 2017, 4238478-
+
+
+*
-
+
+
+
Koskela et al., 2022, 10176386-
++
+
+
4-
+
Leary etal., 2020, 7240043-
+
-
+
+•
~
Li et al., 2021, 7404102-
+
++
+
++
+
+
Liao et al., 2020, 6356903 -
++
++
++
+
++
+
+
++
B Legend
Good (metric) or High confidence (overall)
+ Adequate (metric) or Medium confidence (overall)
- Deficient (metric) or Low confidence (overall)
^ Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-31. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Cardiovascular Effects
Interactive figure and additional study details available on HAWC.
3-138
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
>°v,e
#*¦
,C.®
Lin et al., 2013, 2850967-
-*
+
-
+
+
+
-*
Lin et al., 2016, 3981457-
+
+
+
+
+
+
+
Lin et al., 2020, 6311641-
+
+
+
+
+
+
Lin et al., 2020, 6988476-
-
+
+
+
+
+
+
-*
Lind et al., 2017, 3858504-
+
++ ++
+
++
+
+
Liu et al., 2018, 4238396-
-
+
+
+
+
+
+
+
Liu et al., 2018, 4238514-
+
+
+
+
+
+
+
+
Manzario-Salgado et al., 2017, 4238509-
-
WA
+
+
++
+
Matilla-Santander et al., 2017, 4238432 -
+
+
+*
+
+
+
Mattsson et al., 2015, 3859607 -
++
+
++
+
+
+
+
++
Mitro et al., 2020, 6833625 -
++
++
++
++
+
+
+
++
Mobacke et al., 2018, 4354163 -
+
+
++
+
+
+
+
++
Osorio-Yanez et al., 2021, 7542684 -
-
+
++
+
4
+
+
+
Papadopoulou et al., 2021, 9960593 -
+
+
+
+
+
+
+
+
Pitter et al., 2020, 6988479 -
+
+
+
+
++
+
+
+
Seo et al., 2018, 4238334-
-
+
- B
-
-
B
Varshavsky et al., 2021, 7410195-
+
+
+
+
+
+
+
Warembourg et al., 2019, 5881345-
+
•
+
+
++
+
+
+
Yang et al., 2018, 4238462 -
-
+
-
-
+
+
+
-
Ye et al., 2020, 6988486 -
-
+
+
+
++
+
+
-
Yu et al., 2021, 8453076-
-
-»•
+
+
++
+
+
-
Zare Jeddi et al., 2021, 7404065 -
+
+
-*
+
~
+
+
+
Legend
D
Good (metric) or High confidence (overall)
+
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
B
Critically deficient (metric) or Uninformative (overall)
*
Multiple judgments exist
Figure 3-32. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Cardiovascular Effects (Continued)
Interactive figure and additional study details available on HAWC.
3-139
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.3.1.1.3 Findings from Children
The single high confidence study examined the association between PFOS at several ages
(prenatal, cord blood, 3 years, 8 years, and 12 years) and blood pressure at age 12 and all
observed associations were essentially null. Of the six medium confidence studies that examined
blood pressure in children and adolescents, one reported positive association with diastolic blood
pressure (DBP) only {Ma, 2019, 5413104}, one reported an inverse association with systolic
blood pressure (SBP) and DBP in adolescents, and one reported an increased risk of
hypertension among first-level high school students {Averina, 2021, 7410155}. Results from the
remaining medium confidence studies were essentially null} (see PFOS Appendix). Among
2,251 NHANES (2003-2012) adolescents (mean age 15.5 years) Ma et al. (2019, 5413104)
observed a positive association with DBP, but significant only in boys (0.025; 95% CI: 0.001,
0.049). The study also reported that male adolescents with PFOS levels in the highest quintile
(>18 ng/mL) had mean DBP values that were 2.70% greater (95% CI: 0.32%, 5.02%) than the
lowest quartile (< 6.2 ng/mL). Blood pressure also was examined in children (n = 2,693) and
adolescents (n = 6,669) participating in a health surveillance program in a high-exposure
community (Italy, Veneto Region). Inverse associations were observed for both SBP and DBP in
adolescents which were significant for DBP in continuous analyses. Inverse associations for
DBP were observed in quartile analyses of children, but none reached significance. No
association was observed for SBP in children. In contrast, an increased risk of hypertension was
observed among first-level high school students (n = 940) participating in the Fit Futures Study
{Averina, 2021, 7410155}. In quartile analyses, the association was positive for the second to
fourth quartiles compared to the first but was only significant for the fourth quartile comparison.
No association was observed for DBP among female adolescents, or for SBP among all
adolescents. Manzano-Salgado et al. (2017, 4238509) reported that maternal PFOS was not
associated with blood pressure in combined or in gender-stratified analyses at age 4 and 7 years.
In a cohort of 1,277 children (age 6-11 years), Warembourg et al. (2019, 5881345) observed that
PFOS measured in maternal blood during the pre-natal period, and in plasma during the post-
natal period were not associated with blood pressure in single-pollutant models. Results from an
overlapping study {Papadopoulou, 2021, 9960593} on the same cohort were consistent with
Warembourg et al. (2019, 5881345)
Two low confidence studies did not observe associations between serum PFOS and blood
pressure {Khalil, 2018, 4238547; Lin, 2013, 2850967}.
Other cardiovascular conditions reported in the recent literature include carotid artery intima-
media thickness (CIMT) and brachial artery distensibility. Two medium confidence studies
examined CIMT among 664 {Lin, 2013, 2850967} and 848 {Lin, 2016, 3981457} adolescents
and young adults from the Young Taiwanese Cohort Study. Both studies observed a statistically
significant increase in the mean CIMT with higher serum PFOS levels (p < 0.001 in test for
trend). A low confidence study of children and adolescents from the World Trade Center Health
Registry (WTCHR) reported that the association between PFOS and brachial artery distensibility
was borderline significant (p = 0.06), with no association reported for pulse wave velocity
{Koshy, 2017, 4238478}. However, concerns for residual confounding by age and SES
contributed to the low confidence.
Overall, the limited evidence available among children and adolescents was inconsistent and
indicates PFOS is not associated with blood pressure in these age groups. The evidence for an
3-140
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
association between PFOS and other CVD-related endpoints assessed in this study population
was limited and inconsistent.
3.4.3.1.1.4 Findings from the General Adult Population
Most of the studies identified since the last assessment were conducted among general
population adults (see PFOS Appendix). A total of 16 studies examined PFOS in association
with SBP, DBP, hypertension, and elevated blood pressure {Bao, 2017, 3860099; Chen, 2019,
5387400; Christensen, 2016, 3858533; Christensen, 2019, 5080398; Donat-Vargas, 2019,
5080588; Mitro, 2020, 6833625; Liao, 2020, 6356903; Lin, 2020, 6311641; Liu, 2018, 4238514;
Liu, 2018, 4238396; Mi, 2020, 6833736; Yang, 2018, 4238462; Pitter, 2020, 6988479; Zare
Jeddi, 2021, 7404065; Ye, 2021, 6988486; Yu, 2021, 8453076}.
Of the eight studies that examined blood pressure as a continuous measure, five observed
statistically significant positive associations {Liao, 2020, 6356903; Mitro, 2020, 6833625; Bao,
2017, 3860099; Mi, 2020, 6833736; Liu, 2018, 4238396}. However, the results were not always
consistent between SBP and DBP. A high confidence study in 6,967 participants 20 years and
older in NHANES (2003-2012) reported a statistically significant positive association with SBP
(per 10-fold change in PFOS: 1.35; 95% CI: 0.18, 2.53) {Liao, 2020, 6356903}. Using a
generalized additive model and restricted cubic splines, a non-linear (J-shaped) relationship
between PFOS and DBP was observed, with the inflection point of PFOS at 8.20 ng/mL. Each
10-fold increase in PFOS was inversely associated with DBP (OR: -2.62; 95% CI: -4.73, -0.51)
on the left side of the inflection point and positively associated on the right side of the inflection
point (OR: 1.23; 95% CI: -0.42, 2.88). A high confidence study {Mitro, 2020, 6833625}
conducted in 761 women that examined associations between PFOS concentrations measured
during pregnancy and blood pressure assessed at 3 years post-partum reported significantly
higher SBP levels among all women (beta per doubling of PFOS: 1.2; 95% CI: 0.3, 2.2) and
among women 35 years or older (percent difference per doubling of PFOS: 2.3; 95% CI: 0.9,
3.6). No association was observed with DBP.
Two medium confidence cross-sectional studies with overlapping data from the "Isomers of C8
Health Project", a high-exposed population of Shenyang, China {Mi, 2020, 6833736; Bao, 2017,
3860099} also reported positive associations for blood pressure. In adults with very high PFOS
levels (median 24.22 ng/mL), Bao et al. (2017, 3860099) observed statistically significant
increases in DBP (2.70; 95% CI: 1.98, 3.42) and SBP (4.84; 95% CI: 3.55, 6.12). A positive
trend for the association between PFOS, linear (n-PFOS), and branched isomers, and blood
pressure was highly significant (p < 0.001). In adults with high PFOS levels (median
10.33 ng/mL) Mi et al. (2020, 6833736) reported statistically significant increases in SBP (2.23;
95%) CI: 0.58, 3.89). After stratification by sex, significant positive associations were observed in
women only for SBP, the estimate was 3.08 (95% CI: 1.53, 4.62; p-value for interaction by
sex = 0.03). For DBP, the associations were positive but non-significant overall or among
women. Another high-exposure community study {Pitter, 2020, 6988479} examined risk of
hypertension in a large population (n = 15,786) of young adults (20-39 years old) living in a
PF AS-contaminated region of Italy (Veneto Region) and observed an increased risk of
hypertension. The risk of hypertension was significantly increased in continuous analyses (OR
per ln-ng/mL PFOS: 1.12; 95% CI: 1.02, 1.22), but quartile analyses indicated the association
may have been driven by males in the highest two quartiles of exposure. An overlapping study
{Zare Jeddi, 2021, 7404065} on the same population examined blood pressure as a criterion for
3-141
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
metabolic syndrome and results were consistent with an increased risk of hypertension among
the whole population.
Lin et al. (2020, 6311641) using data from the Diabetes Prevention Program, a randomized
controlled health intervention trial, reported that higher baseline PFOS concentrations were
significantly associated with a decrease in SBP over time (year 2: -2.13 mmHg; 95% CI: -3.54,
-0.71) among participants assigned to the lifestyle intervention arm, but no association was
observed in participants in the placebo-medication arm. However, the study authors attribute the
negative findings for BP trajectories (decreases over time) in the lifestyle group to regression
towards the mean, a statistical phenomenon in which a more extreme value from the population
mean can experience a greater change toward the mean; however, it is unclear why this
phenomenon would apply only to the lifestyle arm.
In a weight loss-controlled trial population (POUNDS-Lost study) Liu et al. (2018, 4238396)
observed that baseline PFOS was positively correlated with DBP (p < 0.001) but at 6- and 24-
month follow-up assessments no associations were observed for SBP or DBP.
No association was observed for blood pressure in two low confidence studies {Chen, 2019,
5387400; Yang, 2018, 4238462}.
Of the eight studies that examined risk of elevated blood pressure (hypertension), two reported
statistically significant associations {Bao, 2017, 3860099; Mi, 2020, 6833736}. Hypertension
was defined as average SBP >140 mmHg and average DBP > 90 mmHg, or self-reported use of
prescribed anti-hypertensive medication. Mi et al. (2020, 6833736) and Bao et al. (2017,
3860099), which had overlapping data on high exposed Isomers of C8 Health Project
participants, reported significant associations. Bao et al. (2017, 3860099) reported significantly
higher odds of hypertension (OR: 1.24; 95% CI: 1.08, 1.44) for PFOS, and for several PFOS
isomers. The associations remained significant in women for PFOS (OR: 1.63; 95% CI: 1.24,
2.13; p-value for interaction by sex = 0.016), and some isomers. These results suggest branched
PFOS isomers have a stronger association with increased risk of hypertension compared to linear
isomers (n-PFOS). Mi et al. (2020, 6833736) reported a significant positive association for
hypertension (OR: 2.52; 95% CI: 1.91, 3.33) overall, and in women (OR 2.32; 95% CI: 1.38,
3.91; p-value for interaction by sex < 0.01).
The high confidence study {Liao, 2020, 6356903} reported in a fully adjusted analysis that the
OR among adults exposed to PFOS levels in the highest tertile compared to the lowest tertile and
the test of trend, respectively, were not significant. Additionally, a significant interaction was
observed between gender and hypertension (p = 0.016), although the association between PFOS
and hypertension was non-significant among males and females in stratified analysis. No
association was observed for elevated blood pressure in two medium studies {Christensen, 2019,
5080398; Liu, 2018, 4238514} and for hypertension in on ^medium {Lin, 2020, 6311641} and
one low confidence study {Christensen, 2016, 3858533}. One medium confidence study {Donat-
Vargas, 2019, 5080588} reported a significant protective effect for hypertension (OR: 0.71; 95%
CI: 0.56, 0.89).
Increased risk of elevated blood pressure was also observed in both low confidence studies {Ye,
2021, 6988486; Yu, 2021, 8453076}, both of which examined participants of the Isomers of C8
Health Project (overlapping with Mi et al. (2020, 6833736) and Bao et al. (2017, 3860099)). Yu
3-142
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
et al. (2021, 8453076) examined components of metabolic syndrome and reported significantly
increased risk of elevated blood pressure. The association was significant in continuous analyses
and the trend was significant in quartile analyses. When stratified by sex, the association was
more pronounced in women and was not significant in men. Ye, 2020, 6988486 reported a non-
significant increased risk in elevated blood pressure. The magnitude of association for total
PFOS was similar to individual PFOS isomers.
Nine studies examined other CVD-related outcomes in adults, including CHD, stroke, carotid
artery atherosclerosis, angina pectoris, C-reactive protein, CHF, microvascular disease, and
mortality.
Graber, 2019, 5080653 reported a positive, borderline significant association with self-reported
cardiovascular conditions (i.e., high blood pressure, CAD, stroke) (1.08; 95% CI: 0.98, 1.21).
However, potential selection bias is a major concern for this study owing to the recruitment of
volunteers who already knew their PFAS exposure levels and were motivated to participate in a
lawsuit.
Among the four studies that examined CHD, the findings were mixed, with three studies
reporting positive non-significant associations, and one study reporting negative associations. A
high confidence study {Mattsson, 2015, 3859607}, a medium confidence NHANES study
{Huang, 2018, 5024212}, and a low confidence study {Christensen, 2016, 3858533} reported
positive non-significant associations with CHD. A low confidence study from the C8 Health
Project {Honda-Kohmo, 2019, 5080551} reported a significant inverse association between
PFOS and CHD among adults with and without diabetes. However, study limitations that may
have influenced these findings include the reliance on self-reporting of a clinician-based
diagnosis for CHD outcome classification and residual confounding by SES.
A medium confidence study of 10,850 NHANES participants (1999-2014) {Huang, 2018,
5024212} reported significantly higher odds of heart attack for the third quartile (OR: 1.56; 95%
CI: 1.01, 2.43) compared to the first quartile, and a very similar but not significant effect in the
fourth quartile. No associations were observed with stroke, CHF, and angina pectoris. A medium
confidence study {Hutcheson, 2020, 6320195} of 3,921 adults with and 44,285 without diabetes
participating in the C8 Health Project found a significant inverse association with history of
stroke (OR: 0.90; 95% CI: 0.82, 0.98; p = 0.02). A significant inverse association with history of
stroke (OR: 0.81; 0.70-0.90) was observed among people with diabetes. No association with
stroke was observed among those without diabetes.
Cardenas, 2019, 5381549 reported significant increases in risk of any microvascular disease, that
were significant only in the lifestyle arm of a health interventions-controlled trial (OR: 1.37;
95%) CI: 1.04, 1.84). No associations were observed for nephropathy, retinopathy, or neuropathy.
Two studies assessed potential PFOS-associated changes in heart structure {Mobacke, 2018,
4354163} and carotid atherosclerosis {Lind, 2017, 3858504} in participants 70 years and older,
with mixed results. Mobacke, 2018, 4354163 evaluated alterations of left ventricular geometry, a
risk factor for CVD and reported that serum PFOS (linear isomer) was significantly associated
with higher left ventricular end-diastolic diameter (0.47; 95%> CI: 0.08, 0.87; p = 0.02) and lower
relative wall thickness (-0.01; 95%> CI: -0.01, -0.001; p = 0.03). PFOS was not significantly
associated with left ventricular mass. Lind et al. (2017, 3858504) reported that plasma PFOS was
3-143
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
not associated with markers of carotid artery atherosclerosis, including atherosclerotic plaque,
the intima-media complex, and the CIMT, a measure used to diagnose the extent of carotid
atherosclerotic vascular disease. Aortic and coronary artery calcification was examined in a
medium confidence study {Osorio-Yanez, 2021, 7542684} on prediabetic participants from the
DPPOS. A significantly increased risk of ascending aortic calcification was reported along with
increased risk of coronary artery calcification. Coronary artery calcification was represented as a
score of severity (Agatston score) indicating mild, moderate, or severe calcification. The odds of
a moderate score (11-400) compared to a mild score (< 11) was increased with respect to PFOS
exposure, and the odds of a severe score (> 400) compared to a mild score were significantly
increased. Koskela, 2022, 10176376, a low confidence study, examined abdominal aortic
calcification among participants aged 40 years and older in NHANES (2013-2014) did not
observe an association.
No association between PFOS and C-reactive protein levels, a risk factor for CVD, was observed
in two studies of pregnant and post-partum women {Mitro, 2020, 6833625; Matilla-Santander,
2017, 4238432}.
Mortality due to heart/cerebrovascular diseases was examined in one medium confidence study
{Fry, 2017, 4181820}. Among a cohort of 1,043 NHANES participants 60 years and older,
PFOS was not associated with mortality due to heart/cerebrovascular diseases.
Overall, the findings from a single high confidence study and several medium confidence studies
conducted among the general population provided consistent evidence for an association
between PFOS and blood pressure. The directionality of this association was mostly positive,
although a single medium confidence study {Lin, 2020, 6311641} reported an inverse
association. The limited evidence for an association between PFOS and increased risk of
hypertension was inconsistent. There was evidence suggesting an increased risk of hypertension
among women {Liao, 2020, 6356903; Bao, 2017, 3860099} in the general adult population, but
additional studies are needed to confirm this finding. Evidence for other CVD-related endpoints
also was limited and inconsistent. No occupational studies examining PFOS exposure and CVD
were identified.
3.4.3.1.2Serum Lipids
3.4.3.1.2.1 Introduction
Serum cholesterol and triglycerides are well-established risk factors for CVDs. Major cholesterol
species in serum include LDL and HDL cholesterol. Elevated levels of total cholesterol (TC),
LDL, and triglycerides are associated with increased cardiovascular risks, whereas higher levels
of HDL are associated with reduced risks.
There are 14 studies (15 publications)13 from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}
that investigated the association between PFOS and serum lipid effects. Study quality
evaluations for these 15 studies are shown in Figure 3-33.
In the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}, the epidemiologic evidence overall
supported an association between PFOS and increased TC. An association between PFOS and
small increases in TC in the general population was observed in several studies {Steenland,
13 Olsen (2003,1290020) is the peer-review paper of Olsen (2001, 10228462).
3-144
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
2009, 1291109; Geiger, 2014, 2850925; Eriksen, 2013, 2919150; Frisbee, 2010, 1430763;
Nelson, 2010, 1291110}. Steenland {2009, 1291109} examined serum PFOS levels among over
46,000 C8 Health Project participants and reported significant positive associations for all serum
lipids except HDL. A cross-sectional study {Frisbee, 2010, 1430763} of children enrolled in the
C8 Health Project also reported significantly increased TC and LDL, with increasing serum
PFOS. Positive associations were seen in another general population study {Eriksen, 2013,
2919150} conducted among Danish adults (50-65 years old). A positive association between
PFOS and hypercholesterolemia also was observed in two cohorts {Steenland, 2009, 1291109,
C8 Health Project; Fisher, 2013, 2919156, Canadian Health Measures Survey}. Cross-sectional
occupational studies {Olsen, 2001, 10228462; Olsen, 2003, 1290020} reported positive
associations between PFOS and increased TC and triglycerides (TG), however, the association
was not observed in longitudinal analyses. Evidence for associations between other serum lipids
and PFOS was mixed including HDL, LDL, VLDL, non-HDL cholesterol, and triglycerides.
3-145
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
,e^
&
Chateau-Degat et al., 2010, 2919285 -
+
+
+
+
-
+
+
+
Eriksen etal., 2013, 2919150-
+
+
+
+
++
+
+
+
Fisher etal., 2013, 2919156-
&
+
+
+
++
+
+
+
Fitz-Simon et al., 2013, 2850962 -
+
++
+ *
+
++
+
+
+ *
Frisbee et al., 2010, 1430763 -
+
+
+ *
+
+
+
+
+ *
Fu etal., 2014, 3749193-
-
+
+ *
-
+
+
+
Geiger et al., 2014, 2850925 -
+
++
++
+
+
+
+
+
Lin etal., 2009, 1290820-
+
+
+
+
+
+
+
+
Maisonet et al., 2015, 3981585 -
+
+
+ *
+
+
+
+
+ *
Nelson etal., 2010, 1291110 -
++
+
+
+
++
+
+
+
Olsen et al., 2001, 10228462-
+
-*
+
+
+
+
-*
Olsen etal., 2003, 1290020-
-
+
-
+
+
+
+
Starling et al., 2014, 2850928 -
+
+
+ *
+
++
+
+
+ *
Steenland et al., 2009, 1291109 -
+
+
+ *
+
+
+
+
+ *
Timmermann et al., 2014, 2850370 -
+
+
+
+
+
+
+
+
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
U Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-33. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS
and Serum Lipids
3-146
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Interactive figure and additional study details available on HAWC.
Since publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}, 66 new epidemiologic
studies (65 publications)14 were identified. These studies examined the associations between
PFOS and serum lipids in children (n = 24), in pregnant women (n = 7), in the general adult
population (n = 32), and in workers (n = 3). Except for ten studies {Olsen, 2012, 2919185;
Domazet, 2016, 3981435; Lin, 2019, 5187597; Liu, 2020, 6318644; Donat-Vargas, 2019,
5080588; Liu, 2018, 4238396; Blomberg, 2021, 8442228; Sinisalu, 2020, 7211554; Li, 2021,
7404102; Tian, 2020, 7026251}, all studies were cross-sectional. Some cohort studies provided
additional cross-sectional analyses {Blomberg, 2021, 8442228; Sinisalu, 2020, 7211554; Li,
2021, 7404102}. Most studies assessed exposure to PFOS using biomarkers in blood, and
measured serum lipids with standard clinical biochemistry methods. Serum lipids were
frequently analyzed as continuous outcomes, but some studies examined the prevalence or
incidence of hypercholesterolemia, hypertriglyceridemia, and low HDL based on the clinical cut-
points, medication use, doctor's diagnosis, or criteria for metabolic syndrome.
3.4.3.1.2.2 Study Quality
All studies were evaluated for risk of bias, selective reporting, and sensitivity following the EPA
IRIS protocol. Three considerations were specific to evaluating the quality of studies on serum
lipids. First, because lipid-lowering medications strongly affect serum lipid levels, unless the
prevalence of medication use is expected to be low in the study population (e.g., children),
studies that did not account for the use of lipid-lowering medications by restriction, stratification,
or adjustment were rated as deficient in the participant selection domain. Second, because
triglycerides levels are sensitive to recent food intake {Mora, 2016, 9564968}, outcome
measurement error is likely substantial when TG is measured without fasting. Thus, studies that
did not measure triglycerides in fasting blood samples were rated deficient in the outcome
measures domain for triglycerides. The outcome measures domain for LDL was also rated
deficient if LDL was calculated based on triglycerides. Fasting status did not affect the outcome
measures rating for TC, directly measured LDL, and HDL because the serum levels of these
lipids change minimally after a meal {Mora, 2016, 9564968}. Third, measuring PFOS and serum
lipids concurrently was considered adequate in terms of exposure assessment timing. Given the
long half-life of PFOS (median half-life = 3.5 years) {Li, 2018, 4238434}, current blood
concentrations are expected to correlate well with past exposures. Furthermore, although reverse
causation due to hypothyroidism {Dzierlenga, 2020, 6833691} or enterohepatic cycling of bile
acids {Fragki, 2021, 8442211} has been suggested, there is yet clear evidence to support these
reverse causal pathways.
There are 65 studies from recent systematic literature search and review efforts conducted after
publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} that investigated the
association between PFOS and serum lipid effects. Study quality evaluations for these 65 studies
are shown in Figure 3-34, Figure 3-35, and Figure 3-36.
Based on the considerations mentioned, 2 studies were considered high confidence, 1 study was
rated high for one exposure measurement and medium for the other, 22 studies were rated
medium confidence for all lipid outcomes, 9 studies were rated medium confidence for TC or
HDL, but low confidence for triglycerides or LDL, 24 studies were rated low confidence for all
14 Dong 2019, 5080195 counted as two studies, one in adolescents and one in adults.
3-147
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
lipid outcomes, and 7 studies were rated uninformative for all lipid outcomes {Seo, 2018,
4238334; Abraham, 2020, 6506041; Predieri, 2015, 3889874; Huang, 2018, 5024212; Leary,
2020, 7240043; Sinisalu, 2021, 9959547}. Notably, nine studies {Zeng, 2015, 2851005;
Manzano-Salgado, 2017, 4238509; Canova, 2020, 7021512; Matilla-Santander, 2017, 4238432;
Blomberg, 2021, 8442228; Canova, 2021, 10176518; DallaZuanna, 2021, 7277682; Tian, 2020,
7026251; Yang, 2020, 7021246} were rated low confidence specifically for triglycerides and/or
LDL because these studies measured triglycerides in non-fasting blood samples. The low
confidence studies had deficiencies in participant selection {Wang, 2012, 2919184; Khalil, 2018,
4238547; Lin, 2013, 2850967; Lin, 2020, 6315756;van denDungen, 2017, 5080340; Chen,
2019, 5387400; Li, 2020, 6315681; He, 2018, 4238388; Yang, 2018, 4238462; Christensen,
2016, 3858533; Graber, 2019, 5080653; Sun, 2018, 4241053; Rotander, 2015, 3859842; Liu,
2018, 4238396; Cong, 2021, 8442223; Khalil, 2020, 7021479; Kobayashi, 2021, 8442188; Liu,
2021, 10176563; Ye, 2021, 6988486; Yu, 2021, 8453076}, outcome measures {Koshy, 2017,
4238478; Yang, 2018, 4238462; Christensen, 2016, 3858533; Kishi, 2015, 2850268; Graber,
2019, 5080653; Rotander, 2015, 3859842; Kobayashi, 2021, 8442188}, confounding {Wang,
2012, 2919184; Khalil, 2018, 4238547; Koshy, 2017, 4238478; Olsen, 2012, 2919185; Lin,
2013, 2850967; Lin, 2020, 6315756; van den Dungen, 2017, 5080340; Li, 2020, 6315681; Yang,
2018, 4238462; Christensen, 2016, 3858533; Graber, 2019, 5080653; Khalil, 2020, 7021479;
Liu, 2021, 10176563; Sinisalu, 2020, 7211554}, analysis {He, 2018, 4238388; Sun, 2018,
4241053; Liu, 2018, 4238396}, sensitivity {Wang, 2012, 2919184; Khalil, 2018, 4238547;
Olsen, 2012, 2919185; Christensen, 2016, 3858533; Graber, 2019, 5080653; Rotander, 2015,
3859842; van den Dungen, 2017, 5080340; Khalil, 2020, 7021479; Sinisalu, 2020, 7211554}, or
selective reporting {Dong, 2019, 5080195} (adolescent portion).
The most common reason for a low confidence rating was concerns for participant selection.
These concerns include a lack of exclusion based on use of lipid-lowering medications {Wang,
2012, 2919184; Lin, 2020, 6315756; Chen, 2019, 5387400; Li, 2020, 6315681; He, 2018,
4238388; Yang, 2018, 4238462; Sun, 2018, 4241053; van denDungen, 2017, 5080340; Liu,
2018, 4238396; Cong, 2021, 8442223; Liu, 2021, 10176563; Ye, 2021, 6988486; Yu, 2021,
8453076}, potential for self-selection {Li, 2020, 6315681; Christensen, 2016, 3858533; Graber,
2019, 5080653; Rotander, 2015, 3859842; van den Dungen, 2017, 5080340}, highly unequal
recruitment efforts in sampling frames with potentially different joint distributions of PFOS and
lipids {Lin, 2013, 2850967}, and missing key information on the recruitment process {Khalil,
2018, 4238547; Yang, 2018, 4238462; Khalil, 2020, 7021479}. Another common reason for low
confidence was a serious risk for residual confounding by SES {Wang, 2012, 2919184; Khalil,
2018, 4238547; Koshy, 2017, 4238478; Olsen, 2012, 2919185; Lin, 2013, 2850967; Lin, 2020,
6315756; van den Dungen, 2017, 5080340; Li, 2020, 6315681; Yang, 2018, 4238462;
Christensen, 2016, 3858533; Graber, 2019, 5080653; Sinisalu, 2020, 7211554}. Frequently,
deficiencies in multiple domains contributed to an overall low confidence rating. The
uninformative studies had critical deficiencies in at least one domain or were deficient in several
domains. These critical deficiencies include a lack of control for confounding {Seo, 2018,
4238334; Huang, 2018, 5024212; Abraham, 2020, 6506041}, convenience sampling {Sinisalu,
2021, 9959547}, and treating PFOS as an outcome of all lipids instead of an exposure, which
limits the ability to make causal inference for the purpose of hazard determination {Predieri,
2015, 3889874}. Small sample size (n = 45) and missing details on exposure measurements were
the primary concerns of the remaining uninformative study {Leary, 2020, 7240043}. Studies
considered uninformative were not considered further. In the evidence synthesis below, medium
3-148
-------
DRAFT FOR PUBLIC COMMENT MARCH 2023
confidence studies were the focus, although low confidence studies were still considered for
consistency in the direction of association.
3-149
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Abraham et al., 2020, 6506041 -
I
+
I
+
L
+ *
¦
I
I
+
•n
Averina et al., 2021, 7410155-
+
+
+ *
+
+
+
+
+
Bjorke-Monsen et al., 2020, 7643487 -
+
+ *
-
+
-
+
-
Blomberg et al., 2021, 8442228-
+
+
+ *
-
+
+*
Canova et al., 2020, 7021512 -
+
+
+ *
+
+
+*
Canova etal., 2021, 10176518-
+
+
+ *
+
+
+
+*
Chen et al., 2019, 5387400-
+
++
+
+
+
+
Christensen et al., 2016, 3858533-
+
-
+
+
-
Christensen et al., 2019, 5080398-
+
+
+
+
+
+
+
+
Cong et al., 2021, 8442223 -
-
+
+
+
+
+
+
Dalla Zuanna et al., 2021, 7277682-
-
+
+ *
+
+
+
+
+*
Domazet et al., 2016, 3981435-
+
+
-
+
+
+
+
+
Donat-Vargas et al., 2019, 5080588 -
+
+
++
+
++
+
+
+
Dong et al., 2019, 5080195-
+
+
+
+
+
-
+
+
Fan et al., 2020, 7102734-
-
+
++
++
+
+
+
Gardener et al., 2021, 7021199-
++
++
++
+
++
+
+
++
Graber et al., 2019, 5080653 -
+
-
+
-
-
Han et al., 2021, 7762348-
+
+
+
¦
+
+
+
+
He et al., 2018, 4238388-
+
++
+
-
+
+
Huang et al., 2018, 5024212-
+
-
+
+
Jain et al., 2018, 5079656-
+
+
+
+
+
+
~
Jain et al., 2019, 5080642-
+
+
+
+
+
+
+
~
Legend
I Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-34. Summary of Study Evaluation for Epidemiology Studies of PFOS and Serum
Lipids
Interactive figure and additional study details available on HAWC.
3-150
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Jensen et al., 2020, 6833719 -
I
+
'
+
, i ,
+*
"
+
'
+
'
+
-
+*
Kang et al.,2018, 4937567-
+
+
+
+
+
+
+
+
Khalil et al.,2018, 4238547-
-
+
+
-
+
+
-
-
Khalil et al., 2020, 7021479-
-
+
+
-
+
4
-
Kishi et al., 2015, 2850268-
+
+
-
+
+
+
-
Kobayashi et al., 2021, 8442188 -
-
+
*'
+
+
+
+
Kobayashi et al., 2022, 10176408 -
+
++
-
+
+
+
++
+
Koshy et al., 2017, 4238478 -
-
+
-
-
+
+
+
Leary et al.. 2020, 7240043 -
+
+
-
+ |
4
-
s
Li et al., 2020, 6315681 -
-
+
+*
-
++
+
+
Li et al., 2021, 7404102-
+
8
++
+
++
+
+
yA
Linet al., 2013, 2850967-
¦
+
+
"
+
+
+
Linet al., 2019, 5187597-
++
+
++
+
++
+
-
+
Linet al., 2020, 6315756-
-
+
+
+
+
+
+
+
Lin et al., 2020, 6988476 -
-
+
-
+
+
+
+
+
Liu et al.,2018, 4238396-
-
+
++
+
+
+
+
Liuet al.,2018, 4238514-
-
+
+
+
+
~
+
+
Liuet al., 2020, 6318644 J
+
+
+
+
++
+
+
+
Liu et al., 2021, 10176563 J
-
-*
++
¦
+
-
+
Manzano-Salgado et al., 2017, 4238509 H
+*
+
-
++
+*
Matilla-Santander et al., 2017, 4238432 -\
+
+
+*
+
+
+*
Mora et al., 2018, 4239224 -J
_
-
+
+
+
+
+
+
Legend
| Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
^ Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-35. Summary of Study Evaluation for Epidemiology Studies of PFOS and Serum
Lipids (Continued)
Interactive figure and additional study details available on HAWC.
3-151
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Cfsfil
Olsenetal., 2012, 2919185-
l
+
i
+
i
+
i
i
+
'
+
i
Papadopoulou et al., 2021, 9960593 -
+
+
+
+
+
+
+
+
Predieri et al., 2015, 3889874 -
+
+
-
-
+
+
¦
~
Rotander et al., 2015, 3859842 -
-
+
-
+
+
+
-
~
Seo et al., 2018, 4238334-
¦
+
B
-
-
-
~
Sinisalu et al., 2020, 7211554-
+
+
* B
-
+
-
~
Sinisalu et al., 2021, 9959547 -
B
+
+
-
-
+
-
~
Skuladottir et al., 2015, 3749113 -
+
+
+
+
+
+
+
+
Spratlen et al., 2020, 5915332 -
+
+
+*
+
++
+
+
+
Starling et al., 2017, 3858473 -
+
+
++
+
++
++
Sunet al., 2018, 4241053-
-
+
+
-
+
+
-
Tianet al., 2020, 7026251 -
+
+
+*
+
++
+
+
+
Varshavsky et al., 2021, 7410195 -
-
+
-
+
+
+
+
-
Wang etal., 2012, 2919184-
-
+
+
-
+
+
-
-
Yang etal., 2018, 4238462-
+
-
-
+
+
+
-
Yang etal., 2020, 7021246-
+
+
+*
+
+
+
+
+*
Ye etal., 2020, 6988486-
-
+
+
+
++
+
+
-
Yu etal., 2021, 8453076-
-
+
+
+
++
+
+
-
Zare Jeddi et al., 2021, 7404065 -
+
+
-*
+
+
+
+
+
Zeng etal... 2015, 2851005-
+
+
+*
+
+
+
+
+*
van den Dungen et al., 2017, 5080340 -
-
+
+
-
+
+
-
-
Legend
| Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
^ Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-36. Summary of Study Evaluation for Epidemiology Studies of PFOS and Serum
Lipids (Continued)
Interactive figure and additional study details available on HAWC.
3-152
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.3.1.2.3 Findings from Children
Results for the studies that examined TC in children are presented in the Appendix (see PFOS
Appendix). Eleven medium confidence and three low confidence studies examined the
association between PFOS and TC in children. Of these, four studies examined the association
between prenatal PFOS exposure and TC in childhood {Spratlen, 2020, 5915332; Jensen, 2020,
6833719; Manzano-Salgado, 2017, 4238509; Mora, 2018, 4239224}, one examined exposure
and TC at multiple timepoints throughout childhood {Blomberg, 2021, 8442228}, and ten
examined the association between childhood PFOS exposure and concurrent TC {Mora, 2018,
4239224; Jain, 2018, 5079656; Zeng, 2015, 2851005; Kang, 2018, 4937567; Khalil, 2018,
4238547; Koshy, 2017, 4238478; Averina, 2021, 7410155; Canova, 2021, 10176518; Tian,
2020, 7026251; Dong, 2019, 5080195} (adolescent portion). Higher PFOS was significantly
associated with higher TC in all children in five medium confidence studies {Jain, 2018,
5079656; Zeng, 2015, 2851005; Canova, 2021, 10176518; Averina, 2021, 7410155; Blomberg,
2021, 8442228}. Notably, significant positive associations were observed among children
{n = 2,693} and adolescents (n = 6,669) of a high-exposure community in Italy (Veneto Region).
The associations were significant in continuous and all quartile analyses and were more
prominent in children compared to adolescents. Significant positive associations were observed
in nine-year old cross-sectional analyses and one prospective comparison (PFOS measured at 5
years, TC measured at 9 years of age) of children belonging to a Faroese cohort {Blomberg,
2021, 8442228}. Comparisons of PFOS and TC measured at other timepoints were less
consistent. Positive associations were also found in four other medium confidence studies
{Spratlen, 2020, 5915332; Jensen, 2020, 6833719; Manzano-Salgado, 2017, 4238509; Mora,
2018, 4239224}, but the associations were small and statistically not significant except for girls
in mid-childhood {Mora, 2018, 4239224}. In contrast, one medium confidence study {Tian,
2020, 7026251} reported inverse associations, however, this analysis was only conducted
concurrently in cord blood. In two out of three low confidence studies, positive associations were
reported, including a statistically significant finding in Koshy 2017, 4238478 {Khalil, 2018,
4238547; Koshy, 2017, 4238478}. However, residual confounding by SES may have positively
biased the results of both studies. Taken together, these studies support a positive association
between PFOS and TC in children, particularly for childhood exposure.
Five medium confidence and seven low confidence studies examined the association between
PFOS and LDL in children. Of these, three examined prenatal exposure {Jensen, 2020, 6833719;
Manzano-Salgado, 2017, 4238509; Mora, 2018, 4239224}, one examined prenatal and childhood
exposure {Papadopoulou, 2021, 9960593} and nine examined childhood exposure {Mora, 2018,
4239224; Zeng, 2015, 2851005; Kang, 2018, 4937567; Khalil, 2018, 4238547; Koshy, 2017,
4238478; Averina, 2021, 7410155; Canova, 2021, 10176518; Tian, 2020, 7026251; Dong, 2019,
5080195} (adolescent portion). The medium studies generally found small, positive associations
between PFOS and LDL, but only one study in first-level high school students reported a
significant association {Averina, 2021, 7410155}. None of the associations were statistically
significant in the remaining medium confidence studies (see PFOS Appendix) {Jensen, 2020,
6833719; Mora, 2018, 4239224; Kang, 2018, 4937567}. Most low confidence studies found a
positive association between PFOS and LDL {Khalil, 2018, 4238547; Koshy, 2017, 4238478;
Manzano-Salgado, 2017, 4238509; Zeng, 2015, 2851005; Canova, 2021, 10176518}, including
statistically significant findings in three studies {Khalil, 2018, 4238547; Koshy, 2017, 4238478;
Canova, 2021, 10176518}. However, residual confounding by SES {Khalil, 2018, 4238547;
Koshy, 2017, 4238478} and the use of non-fasting samples {Canova, 2021, 10176518; Zeng,
3-153
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
2015, 2851005; Manzano-Salgado, 2017, 4238509} were concerns in these studies. Overall,
increases in LDL with increasing PFOS were observed in children, but the magnitudes were
small.
One high confidence, eleven medium confidence, and three low confidence studies examined the
association between PFOS and HDL in children. Of these, three examined prenatal exposure
{Jensen, 2020, 6833719; Manzano-Salgado, 2017, 4238509; Mora, 2018, 4239224}, one
examined prenatal and postnatal exposure {Papadopoulou, 2021, 9960593}, two examined
exposure and HDL at multiple timepoints throughout childhood {Blomberg, 2021, 8442228;Li,
2021, 7404102}, and six examined childhood exposure {Mora, 2018, 4239224; Jain, 2018,
5079656; Zeng, 2015, 2851005; Khalil, 2018, 4238547; Koshy, 2017, 4238478; Dong, 2019,
5080195; Averina, 2021, 7410155; Canova, 2021, 10176518; Tian, 2020, 7026251} (adolescent
portion). The only high confidence study {Li, 2021, 7404102} reported significant positive
associations for HDL at 12 years of age among child participants of the HOME study. PFOS
measured at 8 years of age and concurrently at 12 years of age was significantly associated with
increased HDL. The associations for PFOS measured prenatally, at birth, and at 3 years of age
were all non-significantly positive. Higher PFOS was significantly associated with higher HDL
in children in mid-childhood in two medium confidence studies {Mora, 2018, 4239224; Canova,
2021, 10176518}. The positive association observed in Canova, 2021, 10176518 was consistent
when examining adolescent participants. In Faroese children {Blomberg, 2021, 8442228},
higher PFOS was significantly associated with higher HDL when measured concurrently at
9 years of age. Comparisons of other timepoints (18-month concurrent measurements, 18-month
PFOS and 9-year HDL, and 5-year PFOS and 9-year HDL) were all positively associated with
HDL with increasing PFOS concentrations. Other medium confidence studies found positive
{Jain 2018, 5079656}, inverse (HDL at 18 months in Jensen et al. (2020, 6833719);
Papadopoulou et al. (2021, 9960593), prenatal PFOS; Manzano-Salgado et al. (2017, 4238509);
Zeng et al. (2015, 2851005); Tian et al. (2020, 7026251)), or close to zero (HDL at 3 months in
Jensen et al. (2020, 6833719); Papadopoulou et al. (2021, 9960593), postnatal PFOS)
associations; none of these associations were statistically significant. Two of the three low
confidence studies found positive associations between PFOS and HDL {Khalil, 2018, 4238547;
Koshy, 2017, 4238478}. In summary, mixed associations were found between PFOS and HDL
in children.
Five medium confidence studies and four low confidence studies examined the association
between PFOS and triglycerides in children. Of these, four examined prenatal exposure
{Spratlen, 2020, 5915332; Jensen, 2020, 6833719; Manzano-Salgado, 2017, 4238509; Mora,
2018, 4239224} and six examined childhood exposure {Domazet, 2016, 3981435;Mora, 2018,
4239224; Zeng, 2015, 2851005; Kang, 2018, 4937567; Khalil, 2018, 4238547; Koshy, 2017,
4238478}. Higher mid-childhood PFOS exposure was significantly associated with lower
triglycerides in one medium confidence study {Mora, 2018, 4239224}. The other medium
confidence studies reported positive {Spratlen, 2020, 5915332; Kang, 2018, 4937567}, inverse
(triglycerides at 3 months in Jensen et al. (2020, 6833719); PFOS exposure at age 9 years in
Domazet et al. (2016, 3981435)), or close to zero associations (triglycerides at 18 months in
Jensen et al. (2020, 6833719); PFOS exposure at age 15 years in Domazet et al. (2016,
3981435)); none of these associations were statistically significant. Of note, in Jensen et al.
(2020, 6833719) and Domazet et al. (2016, 3981435), the direction of association changed
depending on the timing of outcome or exposure assessment. One medium confidence study
3-154
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
{Kobayashi, 2022, 10176408} and one low confidence study {Kobayashi, 2021, 8442188}
conducted on mother-child pairs from the Hokkaido Study on Environment and Children's
Health examined the association between prenatal PFOS exposure, maternal polymorphisms of
nuclear receptor genes, and triglyceride levels in infants. Inverse associations for PFOS and TG
were observed, but both studies reported no significant interaction between maternal nuclear
gene polymorphisms and PFOS exposure on triglyceride levels. All other low confidence studies
reported positive associations between PFOS and triglycerides, but all associations were small
and not statistically significant {Manzano-Salgado, 2017, 4238509; Zeng, 2015, 2851005;
Khalil, 2018, 4238547; Koshy, 2017, 4238478; Sinisalu, 2020, 7211554}. The use of non-fasting
samples and residual confounding by SES may have biased these results upwards. Overall,
mixed associations were found between PFOS and triglycerides in children.
In summary, the available evidence supports positive associations between PFOS and TC and
LDL in children. The associations with HDL and triglycerides were mixed.
3.4.3.1.2.4 Findings from Pregnant Women
Four medium confidence studies examined the association between PFOS and TC in pregnant
women and three reported positive associations between PFOS and TC (see PFOS Appendix)
{Matilla-Santander, 2017, 4238432; Skuladottir, 2015, 3749113; DallaZuanna, 2021, 7277682}.
Skuladottir 2015, 3749113, reported a statistically significant linear trend of increasing TC with
increasing PFOS. Positive associations also were observed in an Italian high-exposure
community study {Dalla Zuanna, 2021, 7277682} on pregnant women. The association from
continuous analyses indicated non-significantly increased TC, which was supported by positive
associations when analyzing the second and fourth quartile of exposure but not the second. No
association between PFOS and TC was observed in a Chinese study of pregnant women {Yang,
2020, 7021246}. No association was found in the single low confidence study {Varshavsky,
2021, 7410195} on total serum lipids after adjustment for race/ethnicity, insurance type, and
parity. These findings suggest a consistently positive association between PFOS and TC in
pregnant women.
Two studies {DallaZuanna, 2021, 7277682; Yang, 2020, 7021246} considered low confidence
for LDL due to lack of fasting did not observe an association between PFOS exposure and LDL
in pregnant women. Three medium confidence studies examined the association between PFOS
and HDL, and two reported positive associations. In a high-exposure community study {Dalla
Zuanna, 2021, 7277682}, serum HDL was significantly increased among pregnant Italian
women (beta per ln-ng/mL PFOS: 4.84; 95% CI: 2.15, 7.54), and the association was consistent
in quartile analyses. A study on pregnant women in the Healthy Start Study reported a positive,
though statistically non-significant, association between PFOS and HDL (see PFOS Appendix)
{Starling, 2017, 3858473}. No association between PFOS and HDL was observed in a Chinese
study of pregnant women {Yang, 2020, 7021246}.
One medium confidence and three low confidence studies examined the association between
PFOS and triglycerides in pregnant women. The medium confidence study reported no
association between PFOS and triglycerides (see PFOS Appendix) {Starling, 2017, 3858473}.
Two low confidence studies reported statistically significant, inverse associations between PFOS
and triglycerides {Matilla-Santander, 2017, 4238432; Kishi, 2015, 2850268} while the
remaining study {Yang, 2020, 7021246} reported a non-significant inverse association. All low
3-155
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
confidence studies were limited by their use of non-fasting blood samples. Given that recent food
intake is associated with increased triglycerides and may be a source of PFOS, using non-fasting
blood samples is expected to positively bias the PFOS- triglycerides association. That inverse
associations were still observed in the low confidence studies provides support for an inverse
association between PFOS and triglycerides. This inverse association is inconsistent with the
finding in the only medium confidence study. In sum, the available evidence suggests an inverse
association between PFOS and triglycerides in pregnant women. However, high-quality evidence
is lacking to confirm this association.
Kishi 2015, 2850268 additionally examined the association between PFOS and select fatty acids
in serum. Except for stearic acid and EPA, PFOS was inversely associated with serum fatty
acids; most of these associations were statistically significant {Kishi, 2015, 2850268}. This
study suggests PFOS may disrupt fatty acid metabolism in pregnant women, but additional
studies are needed to confirm this finding.
In summary, the available evidence supports a positive association between PFOS and TC in
pregnancy. The available evidence does not support a consistent, positive association between
PFOS and triglycerides and HDL. Finally, the available evidence is too limited or non-existent to
determine the association between PFOS and LDL in pregnant women.
3.4.3.1.2.5 Findings from the General Adult Population
Ten medium confidence and twelve low confidence studies examined PFOS and TC or
hypercholesterolemia in adults. All studies examined the cross-sectional association {Dong,
2019, 5080195, adult portion; Jain, 2019, 5080642; Liu, 2018, 4238514; Liu, 2020, 6318644;
Lin, 2019, 5187597; Donat-Vargas, 2019, 5080588; Wang, 2012, 2919184; Chen, 2019,
5387400; Li, 2020, 6315681; He, 2018, 4238388; Christensen, 2016, 3858533; Graber, 2019,
5080653; Sun, 2018, 4241053; Liu, 2018, 4238396; Canova, 2020, 7021512; Fan, 2020,
7102734; Lin, 2020, 6988476; Han, 2021, 7762348; Bjorke-Monsen, 2020, 7643487; Cong,
2021, 8442223; Liu, 2021, 10176563; Khalil, 2020, 7021479}; two studies additionally
examined the association between baseline PFOS and changes in TC or incident
hypercholesterolemia {Liu, 2020, 6318644; Lin, 2019, 5187597}.
Of the ten medium confidence studies, nine reported positive associations. In a population of
young adults aged 20 to 39 years in Veneto region, Italy, an area with water contamination by
PFAS, Canova 2020, 7021512 reported statistically positive associations with TC. Canova 2020,
7021512 also reported a concentration-response curve when PFOS was categorized in quartiles
or deciles, with a higher slope at higher PFOS concentrations. Another high-exposure
community study {Lin, 2020, 6988476} conducted in Taiwan provided a sensitivity analysis of
older adults (age 55-75 years), restricting to those participants not taking lipid lowering or anti-
hypertensive medications. In quartile analyses of TC, the association was significantly positive
for the second (beta for Q2 vs. Q1: 15.06; 95% CI: 4.66, 25.46) and third quartile (beta for Q3
vs. Ql: 11.47; 95% CI: 1.03, 21.91) of exposure. The magnitude of association was similar for
the fourth quartile of exposure but did not reach significance.
Four medium studies using overlapping data from NHANES 2003-2014 reported positive
associations between PFOS and TC in adults {Dong, 2019, 5080195, adult portion; Jain, 2019,
5080642; Liu, 2018, 4238514; Fan, 2020, 7102734} (see PFOS Appendix). The association was
statistically significant when data from all cycles were pooled in analyses {Dong, 2019,
3-156
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
5080195}. A cross-sectional analysis {Han, 2021, 7762348} of type 2 diabetes cases and healthy
controls in China reported a positive association for TC, but it did not reach significance. PFOS
also was associated with slightly higher TC at baseline in the POUNDS-Lost cohort {Liu, 2020,
6318644} and the DPPOS {Lin, 2019, 5187597}, but neither association was statistically
significant. The DPPOS also reported that PFOS was associated with a slightly higher prevalence
of hypercholesterolemia at baseline (OR = 1.02, 95% CI: 0.85, 1.21)) and a slightly higher
incidence of hypercholesterolemia prospectively (HR = 1.01, 95% CI: 0.91, 1.12). In contrast to
these findings, Donat-Vargas 2019, 5080588 reported inverse associations between PFOS and
concurrently measured TC. Further, it reported positive associations between PFOS averaged
between baseline and follow-up and TC at follow-up {Donat-Vargas, 2019, 5080588}. All
associations in Donat-Vargas 2019, 5080588 were small and few were statistically significant. It
is noteworthy that all participants in Lin {2019, 5187597} were prediabetic, approximately half
of all participants in Han {2021, 7762348} were diabetic, all participants in Liu {2020,
6318644} were obese and enrolled in a weight loss trial, and all participants in Donat-Vargas
{2019, 5080588} were free of diabetes for at least 10 years of follow-up. It is unclear if
differences in participants' health status explained the studies' conflicting findings.
In low confidence studies, positive associations between PFOS and TC or hypercholesterolemia
were reported in ten of twelve studies {Chen, 2019, 5387400; Li, 2020, 6315681; He, 2018,
4238388; Christensen, 2016, 3858533; Graber, 2019, 5080653; Sun, 2018, 4241053; Liu, 2018,
4238396; Bjorke-Monsen, 2020, 7643487; Cong, 2021, 8442223; Liu, 2021, 10176563}.
However, oversampling of persons with potentially high PFOS exposure and health problems
was a concern in three of these studies {Li, 2020, 6315681; Christensen, 2016, 3858533; Graber,
2019, 5080653}. Medication status and potential residual confounding by SES was a concern in
three studies {Bjorke-Monsen, 2020, 7643487; Cong, 2021, 8442223; Liu, 2021, 10176563}.
Further, He 2018, 4238388 used similar data as the four medium NHANES studies and thus
added little information. Considering medium and low confidence studies together, small
increases in TC with increased PFOS were observed, though less consistently.
3-157
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Confidence Exposure Study Exposure
Rating Reference Matrix Design Levels Sub-population Comparison EE
Effect Estrnate
05 0.9 15 1.1 1.2 U M 15 15 1.7 1.8 1.9 25
Medium CanovaelaL, Serum Cross- Median-3.7 OR |for Q2 re.Q1] 1.19
confidence 2020 sectional npVnL
<25th75tti
OR ffor Q3 vs. Q1] 1.37
OR |for 04 vs. Q1] 1.58
Females OR|fixQ2 VS.Q1] 1.21
OR (for 03 vs. Q1] 1.31
OR (fix 04 vs. Q1] 1.44
Males OR (fir 02 vs. 01] 1.12
OR par 03 vs. Q1] 1.4
OR (for 04 vs. Q1] 1.62
Median=5.66
Low Graber et aL, Cross upl (2Sth-75lh OR (per 1-ug/L increase
confidence 2019 serum sectional percentile^. in senxn PFOS)
09-928 ug'Ll
0.8 0-9 1.0 1.1 1.2 U 1.4 15 15 1.7 15 1.9 25
Figure 3-37. Odds of High Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS
Interactive figure and additional study details available on Tableau.
3-158
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Mcdian-3.7 np'inL
(25lh-75tti percentie:
2.5 5.6 ns'ni)
Regression Coefficient [for Q2 vs. 01 PFOS] 2M
5 6
10 11
Regression Coefficient [for Q3 vs. Q1 PFOS] 4.13
Regression Coefficient [for Q4 vs. 01 PFOS] 6.32
Regression Coefficient [per l-ln(PFOS) rng'mL c
incraase PFOS] a'"
Regression Coefficient [for 02 w. Q1 PFOS] 2.16
Regression Coefficient [far Q3 vs. 01 PFOS] 4.8
Regression Coefficient [far 04 vs. Q1 PFOS] 8.0!
Chateau Degat _ Cross- Geo mean: 16.6 ugil.
el al., 2010 ftosma sectoral (95* Ck 17.&-19.S)
2 3
5 6 7 6
10 11
Figure 3-38. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS
Interactive figure and additional study details available on Tableau.
3-159
-------
DRAFT FOR PUBLIC COMMENT
Pbsina Cohort
Basc&ra median. 20
(25th ?5th
percentile. 15-28 ng/ml)
FofcnvLf) median. 15
(ngrtnl) (25Ih-?5th
percentile: 9.7-21 ngM)
MARCH 2023
Effort Estinate
Median. 20 (ng'rril
(25th-75tti pefcentie:
15-28 ngM)
Median: 15 (n^irtf)
(25th-75th percertle.
9.7-21 rigfoil)
RegressiOT Coefficient for tnrlite 2 _
v, iiwiiu 1 crrvQ 'a-*<1
Regression Coefficient for tertita 3
Dong ct si.. 2019 Scram
Figure 3-39. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS (Continued)
Interactive figure and additional study details available on Tableau.
3-160
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Confidence Exposure Study
Rating Reference Matrix Ocsgn Exposure Lirrcfe Sub papulatoon Comparison EE
Effect Estimate
012 345678
. i Erikscnd Cress .. . .. Difference per pitErquarale
r
i
i
i m
i
i
i
x
Fan et SS., r. Cross Modian=5.14 ugit (25fh 7S9i percentile. Regression Cceflicierrl [per , a,
2020 :*:nirn sectional 2.80-9-31 u»l» t-tog(IO) increase in PFOS) J la
1
i
i
i .
i
i
i
1
Fisher etal. Crass Gncrnotnc moan |SD> = 8-40 tnil. Regression CwrfficrerR (per In - -,
2013 Pta5,na sectional (204) " unit increase PfOS> 0 01
1
i
i
f
i
i
j
Cases: med=7.80 no'rnL (2St-75th n
Han et at., Case perrcnUle: 4.47 10.56 ngimL); Consols: ETTTl
2021 Swunl control mod=a.4S ngtaL (2575th peramble toglO nymL m-rease«. Q.CB
5.4011.95 ng"mLJ
1
i
i
h
1
1
L
Jain el aL, Serum Cross Geomelnc rraonr7 4 n»
-------
DRAFT FOR PUBLIC COMMENT
Exposure Lenb
MARCH 2023
Effect Estimate
confidence 2019
Regression Coefficient (for
Median-16.2 ngi'mL
(25th 75th perccmie:
iai-24.i)
Regression Coefficient (lor , c Ae
ra«.nn 1506
Regression Coefficient (for n j-t
',-4'
Coefficient (for m jg
, Serum Cross - Median: 21.0 ugfl- (range: 20- to 80-year-olils
Regression Coefficient (per
upl increase in PFOS)
Antwerp Mean (SO) = 0.96
a" Serum Cohort ppm (0.97J; Decalur^ 1 40
ppm(1-15>
Regression Coefficient [for
04 (28.2 392.0 upl) vs. 01
(1.4-13.6 upt) PTOSJ
SteenliTidet _ Cross Median: 19.6 ng'rnL
al., 2009 i*!n,nl sectional (mh-nax: 0.2S-75S.2 ng
Regression Coefficient (per „
1-ln ngfmL increase r PFOS) 0 03
Figure 3-41. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS (Continued)
Interactive figure and additional study details available on Tableau.
Six medium confidence studies examined PFOS and LDL in adults and all reported positive
associations. The four studies using overlapping data from NHANES 2003-2014 reported
positive associations between PFOS and LDL {Dong, 2019, 5080195, adult portion; Jain, 2019,
5080642; Liu, 2018, 4238514}, but the association was statistically significant in obese women
only {Jain, 2019, 5080642} (see PFOS Appendix). The association was inverse, but not
statistically significant, in non-obese persons {Jain, 2019, 5080642}. A cross-sectional analysis
{Han, 2021, 7762348} of a case-control study conducted in China reported a significant positive
association among 5 5-75-year-olds. This analysis combined cases of type 2 diabetes and healthy
controls, and it is unclear if the health status of cases explained some of the association. Positive
association between PFOS and LDL also was reported at baseline in the DPPOS, but this
association was not statistically significant {Lin, 2019, 5187597}. This study additionally
reported that PFOS was significantly associated with higher VLDL and non-HDL {Lin, 2019,
5187597}, which are cholesterol species related to LDL and known to increase cardiovascular
risks. Liu 2020, 6318644 reported that PFOS was associated with slightly higher cholesterol in
combined fractions of intermediate-density (IDL) and LDL that contained apolipoprotein C-III
(ApoC-III), but this association was not statistically significant. ApoC-Ill-containing IDL and
LDL are strongly associated with increased cardiovascular risks. Thus, the positive associations
with cholesterol in ApoC-III-containing fractions of IDL and LDL were coherent with the
3-162
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
positive associations found for LDL in the other medium confidence studies. APOB was also
examined in a single medium confidence NHANES study {Jain, 2020, 6988488} that reported a
significantly positive association among non-diabetic, non-lipid-lowering medication users.
Consistent with these findings, nine of the ten low confidence studies reported positive
associations between PFOS and LDL {Lin, 2020, 6315756; Lin, 2013, 2850967; Li, 2020,
6315681; He, 2018, 4238388; Canova, 2020, 7021512; Liu, 2018, 4238396; Bjorke-Monsen,
2020, 7643487; Cong, 2021, 8442223; Khalil, 2020, 7021479}. However, residual confounding
by SES {Lin, 2020, 6315756; Lin, 2013, 2850967; Bjorke-Monsen, 2020, 7643487; Cong, 2021,
8442223} and oversampling of persons with potentially high PFOS and health problems {Li,
2020, 6315681} were major concerns in these studies. In addition, He 2018, 4238388 provided
little new information because it used similar data as the four medium confidence NHANES
studies. Altogether, the available evidence supports a positive association between PFOS and
LDL. Few available findings were statistically significant, however, suggesting that the
association between PFOS and LDL may be relatively small.
Eleven medium confidence and thirteen low confidence studies examined PFOS and HDL or
clinically defined low HDL in adults. All studies examined the cross-sectional association
{Dong, 2019, 5080195, adult portion; Jain, 2019, 5080642; Christensen, 2019, 5080398; Liu,
2018, 4238514; Liu, 2020, 6318644; Lin, 2019, 5187597; Wang, 2012, 2919184; van den
Dungen, 2017, 5080340; Lin, 2020, 6315756; Chen, 2019, 5387400; Li, 2020, 6315681; He,
2018, 4238388; Yang, 2018, 4238462; Fan, 2020, 7102734; Canova, 2020, 7021512; Liu, 2018,
4238396; Bjorke-Monsen, 2020, 7643487; Cong, 2021, 8442223; Khalil, 2020, 7021479; Lin,
2020, 6988476; Han, 2021, 7762348; Zare Jeddi, 2021, 7404065; Ye, 2021, 6988486}. Two
studies additionally examined the association between baseline PFOS and changes in HDL {Liu,
2020, 6318644; Liu 2018, 4238396}. In a population of young adults aged 20 to 39 years in
Veneto region, Italy, an area with water contamination by PFAS, Canova et al. (2020, 7021512)
reported statistically positive associations with HDL. Canova et al. (2020, 7021512) also
reported a concentration-response curve when PFOS was categorized in deciles. An overlapping
study {Zare Jeddi, 2021, 7404065} in the same community was consistent with Canova et al.
(2020, 7021512), reporting significantly decreased odds of reduced HDL (< 40 mg/L, male;
< 50 mg/L, female) in young adults (aged 20 to 39 years). PFOS was associated with lower HDL
at baseline in the DPPOS, but this association was not statistically significant {Lin, 2019,
5187597} (see PFOS Appendix). The POUNDS-Lost study {Liu, 2020, 6318644}, most cycles
of NHANES 2003-2014 {Dong, 2019, 5080195}, a study conducted in a Taiwanese high-
exposure community {Lin, 2020, 6988476}, and a cross-sectional analysis {Han, 2021,
7762348} of type 2 diabetes cases and healthy controls reported no association between PFOS
and HDL. In low confidence studies, PFOS was positively associated with HDL in five of
thirteen studies {Lin, 2020, 6315756; Li, 2020, 6315681; He, 2018, 4238388; Yang, 2018,
4238462; Liu, 2018, 4238396} (association with concurrent HDL). Of note, in Lin 2020
6315756, the positive association was limited to linear PFOS only; the association between
branched PFOS and HDL was inverse and statistically significant {Lin, 2020, 6315756}. The
low confidence studies had limitations in participant selection, residual confounding by SES, and
analysis. It is unclear to what extent these limitations explained the inconsistent findings between
medium and low confidence studies. Overall, the available evidence does not support a
consistently inverse association between PFOS and HDL in adults.
3-163
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Nine medium confidence and thirteen low confidence studies examined the association between
PFOS and TG or hypertriglyceridemia. All studies examined the cross-sectional association
{Jain, 2019, 5080642; Christensen, 2019, 5080398; Liu, 2018, 4238514; Liu, 2020, 6318644;
Lin, 2019, 5187597; Donat-Vargas, 2019, 5080588; Wang, 2012, 2919184; Lin, 2013, 2850967;
Lin, 2020, 6315756; Chen, 2019, 5387400; Li, 2020, 6315681; He, 2018, 4238388; Yang, 2018,
4238462; Sun, 2018, 4241053; Canova, 2020, 7021512; Fan, 2020, 7102734; Liu, 2018,
4238396; Cong, 2021, 8442223; Khalil, 2020, 7021479; Ye, 2021, 6988486; Han, 2021,
7762348; Zare Jeddi, 2021, 7404065}; three studies additionally examined the association
between baseline PFOS and changes in TG or incident hypertriglyceridemia {Liu, 2020,
6318644; Lin, 2019, 5187597; Liu, 2018, 4238396}. Higher PFOS was significantly associated
with higher levels of TG in the DPPOS {Lin, 2019, 5187597} (see PFOS Appendix). This study
also reported that PFOS was associated with higher odds of hypertriglyceridemia at baseline and
higher incidence of hypertriglyceridemia prospectively; the prospective association was
particularly strong in participants enrolled in the placebo arm of the DPPOS {Lin, 2019,
5187597}. In contrast, PFOS was not associated with triglycerides or changes in triglycerides in
the POUNDS-Lost study {Liu, 2020, 6318644}, a cross-sectional analysis {Han, 2021,
7762348} of type 2 diabetes cases and healthy controls, and a high-exposure community study in
Italian young adults (aged 20-39 years) {Zare Jeddi, 2021, 7404065}. Furthermore, PFOS was
inversely associated with TG in the three studies using overlapping NHANES data {Jain, 2019,
5080642; Christensen, 2019, 5080398; Liu, 2018, 4238514} and in Donat-Vargas 2019,
5080588. In this latter study, there was a statistically significant, linear trend of lower TG with
increasing PFOS, regardless of whether PFOS was measured concurrently with TG or averaged
between baseline and follow-up {Donat-Vargas, 2019, 5080588}. In low confidence studies, five
reported inverse associations {Lin, 2013, 2850967; Lin, 2020, 6315756; Li, 2020, 6315681; He,
2018, 4238388; Liu, 2018, 4238396}, six reported essentially null associations {Chen, 2019,
5387400; Sun, 2018, 4241053; Canova, 2020, 7021512; Cong, 2021, 8442223; Khalil, 2020,
7021479; Ye, 2021, 6988486}, one reported a positive association {Yang, 2018, 4238462}, and
one stated the association was not statistically significant {Wang, 2012, 2919184}. Altogether,
the association between PFOS and TG was inconsistent.
In summary, in the general adult population, the available evidence generally supports positive
associations between PFOS and TC and LDL, although some inconsistency exists. The available
evidence does not support a consistent association between PFOS and reduced HDL and elevated
TG.
3.4.3.1.2.6 Findings from Occupational Studies
Workers are usually exposed to higher levels of PFOS, in a more regular manner, and potentially
for a longer duration than adults in the general population. At the same time, according to the
"healthy worker effect," workers tend to be healthier than non-workers, which may lead to
reduced susceptibility to toxic agents {Shah, 2009, 9570930}. Because of these potential
differences in exposure characteristics and host susceptibility, occupational studies are
summarized separately from studies among adults in the general population.
Three low confidence studies examined the association between PFOS and TC in workers. Of
these, two examined the cross-sectional association between PFOS and TC in fluorochemical
plant workers or firefighters exposed to AFFF {Wang, 2012, 2919184; Rotander, 2015,
3859842}; one investigated the association between baseline PFOS and changes in TC over the
3-164
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
course of a fluorochemical plant demolition project {Olsen, 2012, 2919185}. PFOS was
positively associated with TC in Rotander (2015, 3859842), but the association was not
statistically significant. The other cross-sectional study simply reported no significant association
{Wang, 2012, 2919184}. Olsen (2012, 2919185) reported an inverse or positive association
between changes in PFOS and changes in TC, depending on whether the outcome was log-
transformed {Olsen, 2012, 2919185}. This pattern is unusual and suggests different data subsets
may have been used for analyses with and without log-transformed outcome. Taken together, the
occupational studies are limited in both quantity and quality. Based on these studies, it is difficult
to discern the pattern of association between PFOS and TC in workers.
Two studies examined PFOS and LDL in workers. One study examined PFOS and non-HDL, of
which LDL is a major component. All studies were considered low confidence. PFOS was
positively associated with LDL in Rotander (2015, 3859842), but this association was not
statistically significant. The other cross-sectional study simply stated that no significant
association was found {Wang, 2012, 2919184}. The study examining non-HDL found that
changes in PFOS during the fluorochemical plant demolition project were inversely associated
with changes in non-HDL, but the association was not statistically significant {Olsen, 2012,
2919185}. Overall, these studies suggest no consistent association between PFOS and elevated
LDL in workers.
The studies that examined LDL or non-HDL also examined the association between PFOS and
HDL {Wang, 2012, 2919184; Rotander, 2015, 3859842; Olsen, 2012, 2919185}. PFOS was
positively associated with HDL in Rotander (2015, 3859842), but this association was not
statistically significant. The other cross-sectional study simply stated that no significant
association was found {Wang, 2012, 2919184}. In Olsen (2012, 2919185), changes in PFOS
over the demolition project was positively associated with changes in HDL {Olsen, 2012,
2919185}. Together, the occupational studies suggest a positive association between PFOS and
HDL in workers, although these findings were limited by potentially unmeasured confounding
{Rotander, 2015, 3859842; Olsen, 2012, 2919185} and self-selection of subjects {Rotander,
2015, 3859842}.
Two low confidence cross-sectional studies examined PFOS and TG in workers and found that
PFOS was inversely associated with TG in Rotander (2015, 3859842), but this association was
not statistically significant. Wang (2012, 2919184) only reported that no significant association
was found. Given these limited data, it is not possible to determine the pattern of association
between PFOS and TG in workers.
In summary, among workers, a positive association between PFOS and HDL was observed in
some studies. There was not a consistent positive association between PFOS and elevated LDL.
The evidence is too limited to determine the association between PFOS and TC and TG in
workers.
3.4.3.2 Animal Evidence Study Quality Evaluation and Synthesis
There are 4 studies from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} and 8 studies from
recent systematic literature search and review efforts conducted after publication of the 2016
PFOS HESD that investigated the association between PFOS and cardiovascular effects. Study
quality evaluations for these 12 studies are shown in Figure 3-42.
3-165
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Conley et al., 2022, 10176381 -
+
+
NR
++
+
+
Curran et al., 2008, 757871 -
++
NR
NR
++
+
+
Lai et al., 2018, 5080641 -
+
+
NR
++
-
+
Li etal., 2021, 7643501 -
+
+
NR
++
-
+
Luebker et al., 2005, 757857 -
+
+
NR
+
-
-
_l l_^ I . I I l_ 1 L
NTP, 2019, 5400978 -4
&
++ ++ ++ ++
Seacat et al., 2002, 757853 -
++
D
D
B
B
++
++
++
Seacat et al., 2003, 1290852 -
++ ++
NR
+
+
+
+
+
+
Thomford, 2002, 5432419-
+
-
NR
++
++
++
++
++
-
Xia et al., 2011, 2919267-
+
NR
NR
++
B
++
+*
+*
Yan et al., 2014, 2850901 -
++
+
NR
++
++
++
++
++
Zhang et al., 2019, 5918673 -
-
+
NR
+
+
-
+
-
~
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
B Critically deficient (metric) or Uninformative (overall)
Not reported
* Multiple judgments exist
Figure 3-42. Summary of Study Evaluation for Toxicology Studies of PFOS and
Cardiovascular Effects
Interactive figure and additional study details available on HAWC.
Cardiovascular effects, including blood pressure, heart weight, heart histopathology, and/or
serum lipid levels, following exposure to PFOS were minimal {Curran, 2008, 757871; NTP,
2019, 5400978; Rogers, 2014, 2149155; Li, 2021, 7643501; Xia et al„ 201X, 2919267}. In male
and female mice (sexes combined), relative heart weight was increased at PHD 21 after
gestational exposure (GD 2-GD 21) to 2 mg/kg/day PFOS; however, this was confounded by
decreased body weights; absolute heart weights were unchanged {Xia et al., 2011, 2919267}. In
3-166
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
10-11-week-old Sprague-Dawley rats exposed daily by gavage for 28 days, a decrease in
absolute (14% relative to control animals) and relative (9% relative to control animals) heart
weight were reported in females exposed to 5 mg/kg/day while a decrease in absolute (9%
relative to control animals) heart weight was reported in male rats exposed to 5 mg/kg/day. The
authors note that the biological significance of this is not clear. No alterations were observed in
the heart following histopathological analysis in either sex {NTP, 2019, 5400978}. It should be
noted that this study design (e.g., 28-day duration) is not sufficient to address whether PFOS
exposure leads to injuries in the cardiovascular system like plaque formation in atherosclerosis as
this often requires 10-12 weeks for development to accurately be evaluated in a rodent model
{Daugherty, 2017, 5932343}. H&E staining of tissues extracted from PFOS-exposed female
BALB/c mice revealed that exposure (0.1 or 1 mg/kg/day for 2 months) accumulated in the
epicardial area of the heart that correlated regionally with inflammatory cell infiltration (results
reported qualitatively) {Li, 2021, 7643501}.
Curran et al. (2008, 757871) measured blood pressure in 35-37-day old Sprague-Dawley rats
exposed to PFOS in the diet (doses up to approximately 6.34 mg/kg/day for males and
7.58 mg/kg/day for females) for 28 days; no significant change in blood pressure measurements
were observed across the groups, though results were not quantitatively reported. Adult Sprague-
Dawley offspring of dams treated with PFOS (18.75 mg/kg/day) via oral gavage from GD 2-GD
6 had increased blood pressure measurements {Rogers, 2014, 2149155}. Male offspring
exhibited an 18% and 12% increase in systolic blood pressure at 7 and 52 weeks of age,
respectively. Female offspring exhibited a 24% and 19% increase in systolic blood pressure at 37
and 65 weeks of age, respectively; no change in blood pressure was noted at the 7-week
timepoint. In male offspring, increased systolic blood pressure was associated with a
significantly decreased number of nephrons in the kidney (measurements were taken at PND 22;
body weights and kidney weights were not significantly different compared to control animals).
Rogers et al. (2014, 2149155) discussed that the association is a consequence of a higher load on
the available nephrons. The higher load results in a cycle of sclerosis and pressure natriuresis that
can increase blood pressure. However, the exact mechanisms have yet to be elucidated. In
contrast to the results of Rogers et al. (2014, 2149155), no changes in blood pressure were
observed at PND 21 in male and female mice gestationally exposed to 0.2-2 mg/kg/day PFOS
{Xia et al., 2011, 2919267}. Heart rate was also unchanged in this study.
PFOS has been observed to cause perturbations in lipid homeostasis, which may have effects on
the cardiovascular system. Alterations in serum lipid levels have been observed in non-human
primates and rodent models in subchronic, chronic, and developmental studies of oral exposure
to PFOS (Figure 3-43). Decreased serum TC, triglycerides, HDL, LDL, and/or VLDL levels
occurred in rhesus monkeys {Goldenthal, 1979, 9573133}, cynomolgus monkeys {Seacat, 2002,
757853}, rats {Seacat, 2003, 1290852; Thibodeaux, 2003, 5082311; Luebker, 2005, 757857;
Curran, 2008, 757871; NTP, 2019, 5400978; Conley, 2022, 10176381}, and mice {Bijland,
2011, 1578502; Wan, 2012, 1332470; Wang, 2014, 2851252; Yan, 2014, 2850901; Lai, 2018,
5080641} following PFOS exposure. In Sprague-Dawley rats exposed daily by gavage for
28 days, significant decreases in serum TC (males) and triglyceride (females) levels were
reported following PFOS exposure as low as 0.312 and 2.5 mg/kg/day, respectively {NTP, 2019,
5400978}. Serum triglyceride levels were significantly decreased in female CD-I mice exposed
daily by gavage to 3 mg/kg/day PFOS for 7 weeks {Lai, 2018, 5080641}. One study reported
decreased serum HDL levels but an approximate 2-fold increase in serum LDL levels in male
3-167
-------
DRAFT FOR PUBLIC COMMENT MARCH 2023
BALB/c mice following exposure to 5 mg/kg/day PFOS by gavage for 28 days {Yan, 2014,
2850901}.
Endpolnt Study Name
High Density Lipoprotein (HDL) Seacat et al., 2002, 757853
Van el al., 2014,2850901
Luebker el al., 2005, 757857
Observatlo
182d
shorl-lerm (28d) :
reproductive (76d (42d pre-cohabilalion, 14d mating. GDO-20)) i
reproductive (42d prior mating-LD4)
Low Density Lipoprotein (LDL) Yan et al., 2014,2850901
Luebker et al., 2005, 757857
Total Cholesterol
short-term (28d) :
reproductive (76d (42d pie-cohabitation. 1dd mating, GDO-20)) i
reproductive (42d prior mating-LD4)
Seacat et al.. 2002, 757853
Yanetal., 2014, 2850901
Conleyet al., 2022,10176381
Luebker et al., 2005, 757857
short-term (28d)
developmental (GD14-18)
reproductive (76d (42d pre-
GD18
d mating, GDO-20)) GD21
Curran et al.. 2008. 757871
NTP.2019, 5400978
Seacat et al., 2003, 1290852
Seacat etal.. 2002, 757853
Yan et al., 2014.2850901
Lai etal.. 2018, 5080641
Conleyet al., 2022,10176381
Luebker et al., 2005, 757857
Curran et al.. 2008. 757871
NTP, 2019, 5400978
reproductive (42d prior maling-LD4)
short-term (28d)
short-torm (28d)
chronic (I4wk)
chronic (26wk)
short-term (28d)
subchronic (49d)
developmental (GD14-18)
reproductive (76d (42d ore-cohabitatio
reproductive (42d prior malirig-LD4)
short-term (28d)
short-to rm (28d)
GD18
d mating, GDO-20)) GD21
Time Animal Description
Monkey. Cynomolgus IM=4-6)
Monkey, Cynomolgus (r, N=4-6)
Mouse, BALB/c , N=(5)
P0 Ral, Crl;Cd(Sd)lg5 Vafi'Plus (£. N=8)
F1 Rat. Crl:Cd(Sd)lgs Vaf/Plus ( N=8>
P0 Rat. Crl:Cd{Sd)lgs Vaf/Plus (l_. N=17)
F1 Rat, Cr1:Cd(Sd)lgs Vaf/Plus N=17)
Mouse. BALB/c (J. N=6)
P0 Rat. Crl:Cd(Sd)lgs Vaf/Plus (¦:, N=8)
F1 Rat, Cil:Cd(Sd)lgs Vaf/Plus N=8>
P0 Rat. Crl:Cd(Sd)lgs Vaf/Plus (£. N=17)
F1 Rat, Crl:Cd(Sd)lgs Vaf/Plus (rJS. N=17)
Monkey, Cynomolgus ( N=4-C)
Monkey. Cynomolgus (V, N=4-6)
Mouse, BALB/c ( J. N=6)
P0 Rat, Sprague-Dawley ( r. N=4-6)
P0 Rat. Crl:Cd(3d)lgs Vaf/Plus (¦-•, N=8)
F1 Ral, Cri;Cd(Sd)lgs Vaf/Plus (fTS. N=8>
P0 Ral, Crl;Cd{Sd)lgs Vaf/Plus (2. N=17)
F1 Rat, Cri:Cd(Sd)lgs Vaf/Plus( ;L. N=17)
Rat, Sprague-Dawley (:-\ N=15)
Rat. Sprague-Dawley ('J. N=15)
Rat, Sprague-Dawley N=10)
Rat. Sprague-Dawley N=9-10)
Rat, Crl:CD(SD)IGS BR (N=10)
Rat, Crl:CD(SD)IGS BR (•-. N=10)
Monkey, Cynomolgus ( -", N=4-C)
Monkey, Cynomolgus (i, N=4-6)
Mouse. BALB/c (•?, N=6)
Mouse, CD-1 (y, N=4)
P0 Rat, Sprague-Dawley (T, N=4-6)
P0 Rat, Crl:Cd(Sd)lgs Vaf.'Plus N=8)
F1 Rat, Crl:Cd(Sd)igs Vaf/Plus N=8)
P0 Ral. Crl:Cd(Sd)lgs Vaf/Plus (£. N=17)
F1 Rat, Cri:Cd(Sd)lgs Vaf/Plus (-Ti. N=17)
Rat. Sprague-Dawley N=15)
Rat. Sprague-Dawley N=15)
Rat, Spraguc-Dawlcy (c', N=10)
Rat, Sprague-Dawley N=9-10)
PFOS Cardiovascular Effects - Serum Lipids
# No significarl change A Significant increase V Significant
-AL
-W
-W
-AL
—W
V V
V V
v v V V V
V V
V V
Concentration (rng/kg.'day)
Figure 3-43. Serum Lipid Levels in Animal Models Following Exposure to PFOS
PFOS concentration is presented in logarithmic scale to optimize the spatial presentation of data. Interactive figure and additional
study details available on HAWC.
GD = gestation day; Po = parental generation; PND = postnatal day; PNW = postnatal week; Fi = first generation.
Conclusions from these studies are limited by differences in serum lipid composition between
humans and commonly used rodent models, which may impact the relevance of the results to
human exposures {Getz, 2012, 1065480; Oppi, 2019, 5926372}. Some rodent studies {Yan,
2014, 2850901} exhibit a biphasic dose response where low exposure concentrations lead to
increased serum lipid levels while high exposure concentrations lead to decreased serum lipid
levels. This has called in the validity of using rodent models to predict human lipid outcomes.
Additionally, food consumption and food type may confound these results {Cope, 2021,
10176465; Schlezinger, 2020, 6833593; Fragki, 2021, 8442211}, as diet is a major source of
lipids, yet studies do not consistently report a fasting period before serum collection and
laboratory diets contain a lower fat content compared to typical Westernized human diets. More
research is needed to understand the influence of diet on the response of serum cholesterol levels
in rodents treated with PFOS.
3-168
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.3.3 Mechanistic Evidence
Mechanistic evidence linking PFOS exposure to adverse cardiovascular outcomes is discussed in
Section 3.2.6 of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}. There are 9 studies from
recent systematic literature search and review efforts conducted after publication of the 2016
PFOS HESD that investigated the mechanisms of action of PFOS that lead to cardiovascular
effects. A summary of these studies is shown in Figure 3-44.
Mechanistic Pathway Animal Human In Vitro Grand Total
Angiogenic, Antiangiogenic, Vascular Tissue Remodeling
0
1
1
2
Atherogenesis And Clot Formation
1
1
2
4
Cell Growth, Differentiation, Proliferation, Or Viability
0
1
1
2
Cell Signaling Or Signal Transduction
0
0
2
2
Fatty Acid Synthesis, Metabolism, Storage, Transport, Binding, B-Oxidation
1
0
1
2
Inflammation And Immune Response
0
0
2
2
Oxidative Stress
0
2
2
4
Grand Total
2
3
4
9
Figure 3-44. Summary of Mechanistic Studies of PFOS and Cardiovascular Effects
Interactive figure and additional study details available on Tableau.
3.4.3.3.1Fatty acid synthesis, metabolism, storage, transport, and binding
One study published in 2019 found that in vivo exposure to PFOS significantly upregulated the
expression of genes associated with fatty acid metabolism in zebrafish heart tissue {Khazaee,
2019, 5918850}. Fatty acid binding proteins are highly expressed in tissues involved in active
lipid metabolism, such as the heart and liver, and they act as intracellular lipid chaperones
{Nguyen, 2020, 727986}. In this study, adult male and female zebrafish were exposed to 0.1 or
1 mg/L PFOS for 30 days, and the expression of genes that encode fatty acid binding proteins
fabpla.fabplOa, and fabp2 was measured in several tissues (liver, heart, intestine, and ovary) at
four timepoints. PFOS upregulated the expression of fatty acid binding proteins fabplOa and
fabp2 in the heart tissue of males and females at all timepoints, while fabpla expression was not
detected in heart tissue. The authors found that the heart had the most consistent results out of all
tissues examined {Khazaee, 2019, 5918850}. For additional information on the disruption of
fatty acid synthesis, metabolism, transport, and storage in the liver following PFOS exposure,
please see Section 3.4.1.3.2.
3.4.3.3.2Serum Lipid Homeostasis
Epidemiological studies (Section 3.4.3.1) provide consistent evidence that PFOS alters serum
lipid levels, demonstrated by significant positive associations between PFOS and TC and LDL
cholesterol. The mechanisms underlying these associations have not yet been determined. One
3-169
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
study summarized in EPA's 2016 Health Effects Support Document {U.S. EPA. 2016, 3603365}
provides mechanistic evidence related to these outcomes {Fletcher, 2013, 2850968}. The authors
of this study evaluated a subset of 290 adults in the C8 Health Project for evidence that PFOS
can influence the expression of genes involved in cholesterol metabolism, mobilization, or
transport measured in whole blood. When both sexes were analyzed together, a positive
association was found between PFOS and a gene involved in cholesterol mobilization (Neutral
Cholesterol Ester Hydrolase 1 (NCEH1)), and a negative relationship was found between PFOS
and a transcript involved in cholesterol transport (Nuclear Receptor Subfamily 1, Group H,
Member 3 (NR1H3)). When males and females were analyzed separately, serum PFOS was
positively associated with expression of genes involved in cholesterol mobilization and transport
in females (NCEH1 and PPARa), but no associations were found in males. For additional
information on the disruption of lipid metabolism, transport, and storage in the liver following
PFOS exposure, please see Section 3.4.1.3.2.
3.4.3.3.30xidative stress, apoptosis, inflammation, and vascular permeability
leading to atherogenesis
Epidemiological studies (Section 3.4.3.1) provide consistent evidence for an association between
PFOS and blood pressure in some human populations, and limited evidence for an association
between PFOS and increased risk of hypertension. The biological mechanisms underlying the
association between PFOS and elevated blood pressure are still largely unknown, but pathways
that have been proposed include PFOS-induced oxidative stress leading to endothelial
dysfunction and impaired vasodilation, intra-uterine exposure leading to reduced number of
nephrons at birth, interference with signaling pathways of thyroid hormones that regulate blood
pressure, and transcriptional induction of aldosterone {Pitter, 2020, 6988479}.
Oxidative damage, inflammation, and increased vascular permeability are all pathways
associated with the early stages of atherosclerosis. Atherosclerosis is an inflammatory disease of
vessel walls characterized by plaque build-up inside arteries caused by high blood lipid levels
and endothelial dysfunction. Atherosclerosis is an established risk factor for cardiovascular
diseases including myocardial infarction and stroke {Nguyen, 2020, 7279862}. One
epidemiological study found no significant associations between PFOS and carotid artery
atherosclerotic plaque or CIMT {Lind, 2017, 3858504}, but two other studies found significant
associations between PFOS and CIMT {Lin, 2013, 2850967; Lin, 2016, 3981457}.
3.4.3.3.4Endothelial disfunction
3.4.3.3.4.1 In Vivo Evidence
A cross-sectional study in adolescents and young adults in Taiwan (1992-2000) studied the
associations between serum PFOS, CIMT, circulating endothelial and platelet microparticles,
and urinary 8-hydroxydeoxyguanosine (8-OHdG) {Lin, 2016, 3981457}. CIMT is a measure
used to diagnose the extent of carotid atherosclerotic vascular disease. Cluster of differentiation
31 (CD31), also known as platelet endothelial cell adhesion molecule (PECAM-1), is a protein
involved in cell-to-cell adhesion. CD42 is a protein expressed on the surface of platelets that is
involved in platelet adhesion and plug formation at sites of vascular injury. This study evaluated
serum CD31+/CD42a- as a marker of endothelial apoptosis and serum CD31+/CD42a+ as a
marker of platelet apoptosis. The results showed that both markers of apoptosis increased
significantly across quartiles of PFOS exposure. No significant associations were found between
3-170
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
PFOS and CD62E, a marker of endothelial activation, or between PFOS and CD62P, a marker of
platelet activation. In addition, no significant associations were found between serum PFOS and
urinary 8-OhdG, a marker of DNA oxidative stress. The authors observed a positive association
between PFOS and CIMT that was stronger when serum markers of endothelial and platelet
apoptosis were higher. The adjusted odds ratio (OR) for CIMT with PFOS was 2.86 (95% CI:
1.69, 4.84), P < 0.001) when the levels of CD31+/CD42a- and CD31+/CD42a+ were both above
50%, compared to the OR of 1.72 (95% CI: 0.84, 3.53, P = 0.138) when both apoptosis markers
were below 50%. The authors postulated that PFOS may play a role in atherosclerosis by
inducing apoptosis of endothelial and platelet cells {Lin, 2016, 3981457}.
Another cross-sectional study in Taiwanese adults (2009-2011) evaluated the associations
between serum PFOS and urinary 8-OhdG and 8-nitroguanine (8-N02Gua) as biomarkers of
DNA oxidative and nitrative stress {Lin, 2020, 6315756}; however, unlike Lin et al. (2016,
3981457), this study found significant associations between PFOS and biomarkers of oxidative
DNA damage. Linear PFOS levels were positively associated with adjusted levels of 8-OhdG
and 8-N02Gua, while no association was found for branched PFOS levels. The authors also
evaluated the associations between PFOS and serum lipid profiles (LDL, small dense LDL,
HDL, triglycerides), and found that the adjusted OR for elevated LDL (> 75th percentile) with
linear PFOS was higher when each DNA stress marker was above 50% compared to below 50%
(OR 3.15, 95% CI: 1.45, 6.64, P = 0.003 for both stress markers above 50% vs. OR 1.33, 95%
CI: 0.78, 2.27, P = 0.302 for both stress markers below 50%). Linear PFOS levels were also
positively correlated with HDL, but the relationship with stress markers was not studied.
3.4.3.3.4.2 In Vitro Evidence
Liao et al. (2013, 1937227) found that expression of peroxisome proliferator-activated receptor-
gamma (PPARy) and estrogen receptor-alpha {Era) were significantly upregulated in human
umbilical vein endothelial cells (HUVECs) exposed to PFOS (100 mg/L) for 48 hours. PFOS
exposure also significantly upregulated expression of six inflammatory response-related genes
(interleukin-l-beta (IL-lfi), interkeukin-6 (IL-6), prostaglandin-endoperoxide synthase 2
(PTGS2) also known as COX2, nitric oxide synthase 3 (NOS3), P-Selectin, and intracellular
adhesion molecule 1 (ICAM1)) and increased the generation of intracellular reactive oxygen
species (ROS) in HUVECs. In addition, adhesion of monocytes onto HUVECs was increased
2.1-fold over the control when the cells were treated with PFOS (100 mg/L) for 48 hours. The
authors postulated that the PFOS-induced inflammatory response in this in vitro system was
mediated by PPARy, Era, and ROS, and that PFOS upregulation of ICAM1 and P-Selectin may
play an important role in adhesion of monocytes to vascular epithelium leading to vascular
inflammation.
Similarly, Qian et al. (2010, 2919301) found that PFOS induced ROS production in human
microvascular endothelial cells (HMVECs) even at low concentrations (2-5 |iM) within one
hour. These authors also studied permeability changes in HMVEC monolayers following PFOS
exposure by measuring transendothelial electrical resistance. The results showed that PFOS
induced endothelial permeability in a concentration-dependent manner. Confocal microscopy
imaging analysis revealed many gaps in the PFOS-treated HMVEC monolayers that increased in
a concentration-dependent manner. PFOS also induced actin filament remodeling. Pretreating
HMVEC monolayers with catalase, a ROS scavenger, prior to PFOS exposure substantially
blocked the PFOS-induced gap formation and actin filament remodeling.
3-171
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Two studies evaluated the potential for PFOS and other PFAS to activate the plasma kallikrein-
kinin system (KKS) using in vitro and ex vivo activation assays and in silico molecular docking
analysis {Liu, 2017, 4238579; Liu, 2018, 4238499}. The plasma KKS plays important roles in
regulating inflammation, blood pressure, coagulation, and vascular permeability. Activation of
the plasma KKS can release the inflammatory peptide, bradykinin (BK), which can lead to
dysfunction of vascular permeability {Liu, 2018, 4238499}. The cascade activation of KKS
involves autoactivation of Hageman factor XII (FXII), cleavage of plasma prekallikrein (PPK),
and activation of high-molecular-weight kininogen (HK) {Liu, 2018, 4238499}. These studies
examined the potential for PFOS and other PFAS chemicals to act as FXII activators due to their
structural similarities to natural long-chain fatty acids {Liu, 2017, 4238579}. The addition of
PFOS (1-5 mM) to mouse plasma ex vivo resulted in dose-dependent PPK activation measured
by analysis of PPK and plasma kallikrein expression levels after 2 hours of incubation, and the
approximate lowest-observed-effect concentration (LOEC) for PFOS was 3 mM {Liu, 2017,
4238579}. This demonstrated the potential for PFOS to activate the plasma KKS, but at a
relatively high concentration compared to typical human exposure levels in the general
population. PFAS with longer carbon chain lengths activated the KKS at a much lower
concentration compared to PFOS (e.g., PFHxDA activated the KKS at 30 |iM). Time course
experiments showed that PPK activation occurred within 5 min after addition of PFOS or other
PFAS to mouse plasma {Liu, 2017, 4238579}.
The potential effects of PFOS on KKS activation in mouse plasma ex vivo were also evaluated
using protease activity assays. Plasma samples were incubated with PFOS (100-5,000 |iM) for
15 minutes and then analyzed for FXIIa activity and kallikrein-like activity. PFOS significantly
increased FXIIa activity only at the highest concentration tested (5 mM) Liu et al. (2018,
4238499), and kallikrein-like activity was significantly increased only at 3 and 5 mM PFOS
{Liu, 2017, 4238579; Liu, 2018, 4238499}. Western blot analyses demonstrated that 5 mM
PFOS could induce the KKS waterfall cascade activation both in vitro, utilizing human plasma
zymogens FXII, PPK, and HK, and ex vivo utilizing plasma from human volunteers {Liu, 2017,
4238579}.
Binding of PFOS with purified human FXII was further evaluated by Liu et al. (2017, 4238579)
using native PAGE separation and FXII Western blot assay. Two hours of incubation of FXII
with PFOS (1 or 3 mM) reduced the amount of free FXII in a concentration-related manner. The
results from ex vivo, in vitro, and in silico experiments were compared for different PFAS, and
the authors concluded that the degree of KKS activation was related to structural properties such
as carbon chain length, terminal groups, and fluorine atom substitution. For example, PFAS
terminated with sulfonic acid, including PFOS, demonstrated a stronger binding affinity for FXII
and higher capability of inducing KKS activation than PFAS terminated with carboxylic acid or
other terminal groups. {Liu, 2017, 4238579}.
3.4.3.3.5Coagulation and fibrinolysis
The coagulation and fibrinolytic pathways can contribute to the progression of atherosclerosis.
Two studies from the literature published after the 2016 HESD evaluated the potential of PFOS
to affect these pathways. Bassler et al. (2019, 5080624) evaluated a subset of 200 individuals
from the C8 Health Project for a variety of disease biomarkers including plasminogen activator
inhibitor (PAI-1), a glycoprotein that inhibits the formation of plasmin from plasminogen and
thus prevents clot lysis in vessel walls. Elevated PAI-1 levels are associated with thrombotic risk,
3-172
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
but this study found no significant association between PFOS and PAI-1 levels. Likewise, Chang
et al. (2017, 3981378) saw no significant changes in coagulation parameters measured in male
and female cynomolgus monkeys following acute oral exposure to PFOS with serum
concentrations up to 165 |ig/mL, including measures of prothrombin time, activated partial
thromboplastin time, and fibrinogen.
3.43.4 Evidence Integration
There is moderate evidence for an association between PFOS exposure and cardiovascular
effects in humans based on consistent positive associations with serum lipid levels, specifically
LDL and TC. Additional evidence of positive associations with blood pressure and hypertension
in adults supported this classification. The available data for CVD and atherosclerotic changes
was limited and addressed a wider range of outcomes, resulting in some residual uncertainty for
the association between PFOS exposure and these outcomes.
The human epidemiological studies identified since the 2016 health assessments provided
additional clarity regarding the association between PFOS and CVD. Most of the CVD evidence
identified focused on blood pressure in general adult populations (12 studies). The findings from
one high confidence study and five medium confidence studies provide evidence for a positive
association between PFOS and blood pressure, although the results were not always consistent
between SBP and DBP, and one study reported an inverse association. The limited evidence for
an association between PFOS and increased risk of hypertension was inconsistent. There was,
evidence suggesting an increased risk of hypertension among women, but additional studies are
needed to confirm this finding. One high confidence study in women with PFOS measured
during pregnancy reported a positive association with blood pressure assessed at 3 years post-
partum. Evidence in children and adolescents is also less consistent. The six studies available
among children and adolescents suggest PFOS was not associated with elevated blood pressure.
Evidence for other CVD-related outcomes across all study populations was more limited and
inconsistent. The limited evidence for CVD outcomes discussed in the 2016 assessment also
indicated association with blood pressure in children.
Based on this systematic review of 44 epidemiologic studies, the available evidence supports a
positive association between PFOS and TC in the general population, including children and
pregnant women. The available evidence also generally supports a positive association between
PFOS and LDL in children and adults in the general population. Although PFOS appeared not
associated with elevated TC and LDL in workers, this conclusion is uncertain as the occupational
studies included in this review are limited in both quantity and quality. Finally, for all
populations, the association between PFOS and HDL and TG were mixed, suggesting no
consistent associations between PFOS and reduced HDL and elevated TG. Overall, these
findings are largely consistent with the 2016 Health Assessment. The positive associations with
TC are also supported by the recent meta-analysis restricted to general population studies in
adults {EPA, 2022, 10369698}. Similarly, a recent meta-analysis including data from I I
studies reported consistent associations between serum PFOS or a combination of several
PFCs including PFOA and PFOS, and increased serum TC, LDL, triglyceride levels in
children and adults ! Abdullah Soheimi, 2021, 9959584}.
The animal evidence for an association between PFOS exposure and cardiovascular toxicity is
moderate based on serum lipids effects observed in eight high or medium confidence studies.
3-173
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
The most consistent results are for total cholesterol and triglycerides, although direction of effect
can vary by dose. In animal toxicological studies, no effects or minimal alterations were noted
for blood pressure, heart weight, and histopathology in the heart. However, many of the studies
identified may not be adequate in exposure duration to assess potential toxicity to the
cardiovascular system. The biological significance of the decrease in various serum lipid levels
observed in these animal models regardless of species, sex, or exposure paradigm is unclear;
however, these effects do indicate a disruption in lipid metabolism.
The mechanisms underlying the positive associations between PFOS and serum TC, LDL, and
blood pressure in humans have yet to be determined. Data from the C8 Health Project
demonstrated that serum PFOS was positively associated with expression of genes involved in
cholesterol mobilization and transport in samples from women (NCEH1 and PPARa), while
there were no associations in men. The results for PFOS-induced changes to serum lipid levels
are in contrast in rodents (generally decreased) compared to humans (generally increased). PFOS
exposure led to up-regulation of genes that encode fatty acid binding proteins in zebrafish, which
play a role in lipid binding, particularly in the heart. Evidence is ultimately limited in regard to
clear demonstration of mechanisms of alterations to serum lipid homeostasis caused by PFOS
exposure.
Regarding the potential for PFOS to lead to atherosclerosis as evaluated by related mechanisms
or mechanistic indicators, one epidemiologic study found no association between PFOS and
carotid artery atherosclerotic plaque or CIMT, while two other epidemiologic studies found
significant associations between PFOS and CIMT. The two studies that reported PFOS-
associated CIMT demonstrated endothelial dysfunction via increases in markers of endothelial
and platelet apoptosis in the serum: increased serum CD31+/CD42a-, which is a marker of
endothelial apoptosis, and increased serum CD31+/CD42a+, which is a marker of platelet
apoptosis. Markers of serum and platelet activation were not changed, nor was there evidence of
DNA oxidative damage (no change in urinary 8-OhdG). The authors of the study postulated that
PFOS-induced apoptosis of endothelial and platelet cells may play a role in the development of
atherosclerosis. In contrast, another human study reported increased urinary 8-OhdG and 8-
nitroguanine (8-N02Gua) resulting in limited and inconsistent results for oxidative damaging
potential of PFOS. In vitro, PFOS was shown to induce oxidative stress and upregulate
inflammatory response genes in human umbilical vein endothelial cells. The authors concluded
that oxidative stress and changes in the expression of genes involved in adhesion of monocytes
to vascular epithelium may lead to vascular inflammation. Binding of PFOS to human FXII was
demonstrated, which is the initial zymogen of plasma kallikrein-kinin system (KKS) activation,
an important regulator of inflammation, blood pressure, coagulation, and vascular permeability.
The authors attributed the degree of KKS activation to structural properties of PFOS (among
other PFAS). There was no association between PFOS and disease biomarkers related to clotting
and coagulation in both human and non-human primate data. While there is mechanistic
evidence that PFOS exposure can lead to molecular and cellular changes that are related to
atherosclerosis, human studies identified herein reported a lack of an association between PFOS
exposure and markers of atherosclerosis. Thus, the relevance of these mechanistic data is
unclear.
3-174
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.3.4.1Evidence Integration Judgment
Overall, considering the available evidence from human, animal, and mechanistic studies, the
evidence indicates that PFOS exposure is likely to cause adverse cardiovascular effects,
specifically serum lipids effects, in humans under relevant exposure circumstances (Table 3-11).
The hazard judgment is driven primarily by consistent evidence of serum lipids response from
epidemiological studies at median PFOS levels between 3.7-36.1 ng/mL (range of median
exposure in studies observing an adverse effect). The evidence in animals showed coherent
results for perturbations in lipid homeostasis in non-human primates and rodent models in
developmental, subchronic, and chronic studies following exposure to doses as low as
0.03 mg/kg/day PFOS. While there is some evidence that PFOS exposure might also have the
potential to affect blood pressure and other cardiovascular responses in humans given relevant
exposure circumstances, the human evidence underlying this possibility is uncertain and without
support from animal or mechanistic studies.
3-175
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 3-11. Evidence Profile Table for PFOS Cardiovascular Effects
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
Evidence from Studies of Exposed Humans (Section 3.4.3.1)
Serum lipids Among studies of High and medium
2 High confidence studies children (20), several confidence studies
20 Medium confidence studies reported evidence Consistent findings of
Low confidence studies
studies
25 Low confidence
studies
11 Mixed' studies
of significant increases in positive associations for
TC (6/20) and LDL
(5/20), though others
observed no association.
While some studies
observed significantly
increased HDL (6/20),
LDL and TC across
study populations
Coherence of observed
associations in adults
from the general
population with
others reported significantprevious evidence from
decreases or no serum lipid effects
associations. Medium and
mixed confidence studies
in adults (19) observed
significant positive
associations in HDL
(7/19), LDL (7/19), and
TC (8/19). Results for TG
were mixed, with two
studies reporting
increased levels and three
studies finding decreased
levels, with one study in
obese females. Non-
significant inverse results
were observed for HDL
(4/19), LDL (3/19), TC
(6/19), and TG (7/19).
Low confidence studies
followed the similar trend
of mixed results. Studies
examining pregnant
0©O
Moderate
0©O
'Evidence Indicates (likely)
Evidence for
cardiovascular effects is
based on numerous
medium confidence studiesserum hPlds resPonse 311(1
animal evidence showed
Primary basis and cross-
stream coherence:
Human evidence indicated
consistent evidence of
reporting positive
associations with serum
lipids (HDL, LDL, TC,
TG) in adults from the
general population.
Studies of children
reported mixed findings in
most serum lipids, but
results were largely
consistent for LDL and
TC, with some reaching
significance. High and
coherent results for
perturbations in lipid
homeostasis in non-human
primates and rodent models
in developmental,
subchronic, and chronic
studies following exposure
to PFOS. While there is
some evidence that PFOS
exposure might also have
the potential to affect blood
/H«///w/ cmVndcrKC St"udicsPrcssurc 311(1 other
reported positive cardiovascular responses in
associations with blood humans 8ivcn relevant
pressure and increased riskexPosure circumstances,
the human evidence
underlying this possibility
is uncertain and without
support from animal or
mechanistic studies.
of hypertension. Low
confidence studies
reported non-significant
associations, while most
mixed confidence studies
reported significant
associations. Observed
effects were inconsistent
for CVD and imprecise forNo sPeciflc factors are
noted.
Human relevance and other
inferences:
3-176
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
women were of medium
and mixed confidence and
reported mixed results
(6). While three studies
reported evidence of
increased HDL and TC
levels, the others failed to
reach significance or
reported inverse
associations.
atherosclerotic changes
across all study
populations.
Blood pressure and Results from studies of High and medium
hypertension varying confidence confidence studies
2 High confidence studies reported mixed results for
16 Medium confidence
studies
7 Low confidence studies
changes in blood
pressure, including DBP
and SBP, and risk of
hypertension for all study
populations. Studies in
children (10) reported
mostly non-significant
associations with blood
pressure and/or
hypertension, though two
studies in adolescents
reported significantly
increased (1/10) and
decreased (1/10) DBP in
males. In adults (13), one
study reported a
significantly increased
risk of hypertension
(1/13), but associations
from other studies did not
reach significance (3/13).
When stratified by sex,
there were mixed results.
Low confidence studies
Inconsistent findings of
effects observed in children,
likely due to variation in
measured exposure
windows
3-177
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
One study reported a
higher risk of
hypertension for males
(1/13), while another
reported higher risk for
females (1/13). One study
reported an inverse
association for DBP
(1/13), while others
reported positive
associations for DBP
(6/13), but only three
studies reached
significance. SBP was
significantly increased for
all adults (4/13), in
females only (2/13), and
in males only (1/13). No
studies examined blood
pressure or hypertension
in occupational
populations.
Cardiovascular disease
1 High confidence study
4 Medium confidence
studies
2 Low confidence studies
In adults from the general High and medium
population (6), confidence studies
significantly decreased
odds of stroke (1/6) and
significantly increased
odds of MVD (1/6), heart
attack and CVD in the
third exposure group
(1/6), CVD in males
(1/6), and self-reported
cardiovascular conditions
(1/6) were observed.
Other studies of stroke,
CHD, and CVD reported
Low confidence studies
Inconsistent findings for
CVD-related outcomes
Imprecision of findings,
particularly for two studies
with self-reported outcome
measures
3-178
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
non-significant
associations, including
one high confidence study
that reported no
associations with CHD
among Swedish men and
a medium confidence
study that reported no
association with mortality
from CVD or other heart
diseases. One low
confidence occupational
study (1) examined male
anglers over age 50 and
did not observe an
association with CHD or
any cardiovascular
conditions.
Atherosclerotic changes
In studies of children (3), High and medium
Low confidence study
1 High confidence study
one study observed confidence studies
Imprecision of findings
4 Medium confidence
significant associations
across children and adult
studies
with CIMT across
study populations
1 Low confidence study
exposure groups, among
Limited number of studies
females, and among those
examining specific
ages 12-19 (1/3). A
outcomes
cohort study of children
and young adults reported
significant increases in
CIMT for all exposure
groups and significant
increases in some
endothelial microparticle
levels for atherosclerosis
(1/3). Findings were
mixed among adults older
than 70 years of age. One
3-179
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
study did not observe
significant effects while a
separate study reported a
significant increase in left
ventricular end-diastolic
diameter and a significant
decrease in relative wall
thickness (1/3). One
medium confidence study
also reported significantly
increased odds in
Agatatson Scores of over
400, a measure of arterial
calcification, in
prediabetic adults aged
over 25.
Evidence from In Vivo Animal Toxicological Studies (Section 3.4.3.2)
Serum lipids Significant decreases in High and medium
2 High confidence studies serum TG were observed confidence studies
6 Medium confidence in 5/7 studies that Consistency of findings
studies
examined this endpoint, across species, sex, or
regardless of species, sex, study design
or study design. No Dose-response
changes were observed in relationship observed
one monkey study and within multiple studies
one short-term study in
male mice. Similar
decreases were observed
in serum TC (6/7), with
no changes being
observed in one short-
term study in male mice.
In a developmental study,
decreases were observed
in dams, but no change
Incoherence of findings in
other cardiovascular
outcomes
Biological significance of
the magnitude of effect is
unclear
0©O
Moderate
Evidence based on eight
high or medium
confidence studies
observed that PFOS
affects serum lipids in
animal models. The most
consistent results are for
total cholesterol and
triglycerides, although
direction of effect can vary
by dose. The biological
significance of the
decrease in various serum
lipid levels observed in
3-180
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
was observed in pups.
Fewer studies examined
HDL and LDL, with
decreases in HDL (2/3)
and increases in LDL
(2/2) being observed.
Histopathology
1 High confidence study
2 Medium confidence
studies
No changes in heart
histopathology were
reported in 2 rat studies.
One study in female mice
qualitatively reported an
increase in inflammatory
cell infiltration.
High and medium
confidence studies
Limited number of studies
examining outcome
Organ weight
1 High confidence study,
2 Medium confidence
studies
Mixed results were High and medium
reported for absolute and confidence studies
relative heart weight.
Two short-term studies
reported decreases in
absolute heart weights in
male and female rats, but
mixed results (no change
or decreases) were
reported for relative heart
weights. A developmental
study reported no change
in absolute heart weight
and an increase in relative
heart weight which was
confounded by decreases
Limited number of studies
examining outcome
Confounding variables such
as decreases in body
weights may limit ability to
interpret these responses
these animal models
regardless of species, sex,
or exposure paradigm is
unclear; however, these
effects do indicate a
disruption in lipid
metabolism. No effects or
minimal alterations were
noted for blood pressure,
heart weight, and
histopathology in the
heart. However, many of
the studies identified may
not be adequate in
exposure duration to
assess potential toxicity to
the cardiovascular
system.
Blood pressure and
A short-term and a Medium confidence
Limited number of studies
heart rate
developmental study studies
examining outcome
3 Medium confidence
found no effect on blood
studies
pressure in male and
female rats. One
developmental study
3-181
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Studies and Summary and Key Factors that Increase Factors that Decrease
Interpretation Findings Certainty Certainty
Evidence Stream
Judgment
found no effect on heart
rate.
Mechanistic Evidence and Supplemental Information (Section 3.4.3.3)
Summary of Key Findings, Interpretation, and Limitations
Evidence Stream
Judgement
Key findings and interpretation:
PFOS exposure was associated with changes in the expression of genes involved in cholesterol
metabolism, mobilization, or transport in whole blood of adult humans.
PFOS induced oxidative stress and upregulated inflammatory response genes in human umbilical vein
endothelial cells exposed in vitro, which can lead to vascular inflammation.
Findings support the
plausibility that PFOS
exposure can lead to
changes in the expression
of genes involved in
PFOS can bind to human FXII in vitro, which is the initial zymogen of plasma KKS activation, a regulatorcholesterol regulation, as
of inflammation, blood pressure, coagulation, and vascular permeability.
Limitations:
Small database; the only in vivo evidence is reported in two human studies with conflicting results for
markers of platelet activation.
Results regarding the association between PFOS exposure and carotid artery atherosclerotic plaques or
CIMT, which are mechanisms of atherosclerosis, are inconsistent in human epidemiological studies.
well as molecular and
cellular changes that are
related to atherosclerosis,
although no association
was observed between
PFOS exposure and
atherosclerosis in human
epidemiological studies.
Evidence Integration
Summary Judgment
Notes: CHD = coronary heart disease; CIMT = carotid intima-media thickness; CVD = cardiovascular disease; DBP = diastolic blood pressure; FXII = Factor XII; HDL = high
density lipoprotein; KKS = kallikrein-kinin system; LDL = low density lipoprotein; density lipoprotein; SBP = systolic blood pressure; MVD = microvascular disease; TC = total
cholesterol; TG = triglycerides.
¦Mixed confidence studies had split confidence determinations for different serum lipid measures with some measures rated medium confidence and others rated low confidence.
3-182
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.4 Developmental
EPA identified 96 epidemiological and 19 animal toxicological studies that investigated the
association between PFOS and developmental effects. Of the epidemiological studies, 28 were
classified as high confidence, 37 as medium confidence, 20 as low confidence, 3 as mixed (2
high/medium and 1 medium/low) confidence, and 8 were considered uninformative (Section
3.4.4.1). Of the animal toxicological studies, 15 were classified as medium confidence, 3 as low
confidence, and 1 was considered mixed(medium/uninformative) (Section 3.4.4.2). Studies have
mixed confidence ratings if different endpoints evaluated within the study were assigned
different confidence ratings. Though low confidence studies are considered qualitatively in this
section, they were not considered quantitatively for the dose-response assessment (Section 4).
3.4.4.1 Human Evidence Study Quality Evaluation and Synthesis
3.4.4.1.1 Introduction
This section describes studies of PFOS exposure and potential in utero and perinatal effects or
developmental delays, as well as effects attributable to developmental exposure. Developmental
endpoints include gestational age, measures of fetal growth (e.g., birth weight), and miscarriage,
as well as infant/child development.
The 2016 PFOS HESD {U.S. EPA, 2016, 3603365} summarized epidemiological studies of
developmental effects in relation to PFOS exposure. There are 18 studies from the 2016 PFOS
HESD {U.S. EPA, 2016, 3603365} that investigated the association between PFOS and
developmental effects. Study quality evaluations for these 18 studies are shown in Figure 3-45.
Studies included ones conducted both in the general population as well as in communities known
to have experienced high PFOS exposure (e.g., the C8 population in West Virginia and Ohio).
Results of eleven high or medium confidence epidemiological studies (see Section 2.1.3 for
information about study quality evaluations) discussed in the 2016 PFOS HESD are summarized
below.
3-183
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
&
Andersen et al., 2010, 1429893 -
Apelberg et al., 2007, 1290833-
Apelberg etal., 2007, 1290900-
Chen etal., 2012, 1332466-
Darrow et al., 2013, 2850966
Darrow et al„ 2014, 2850274
Fei etal., 2007, 1005775-
Fei etal., 2008, 1290822-
Fei etal., 2008, 2349574
Fei etal., 2010, 1430760'
Grice et al., 2007, 4930271
Hamm etal., 2010, 1290814'
Maisonet et al., 2012, 1332465
Monroy et al., 2008, 2349575 •
Olsen etal., 2004, 5081321
Stein etal., 2009, 1290816
Washino et al., 2009, 1291133
Whitworth et al., 2012, 2349577
+
+
+
+
+
+
+
+
+
+
8
+
++
+
+
+
+
+
+
+
+
+
-
1
+
+
+
+
+
+
++
+
+
++
+
+
++
+
+
-
+
+
+
+
+
+
++
++
+
++
+
+
+
+
++
D
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
-
+
+
+
+
-
+
+
-
-
+
+
+
-
+
+
+
+
+
+
+
+
+
-
+
+
+
~
-
++
++
-
-
+
+
-
-
+
-
+
+
-
+
+
-
+
+
+
+
+
+
+
+
+
+
+
+
+
++
++
+
++
-
++
++
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
^ Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-45. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS
and Developmental Effects
Interactive figure and additional study details available on HAWC.
3-184
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
As noted in the 2016 HESD, several available studies measured fetal growth outcomes. Apelberg
et al. (2007, 1290833) found that birth weight, head circumference, and ponderal index were
inversely associated with umbilical cord PFOS concentration in 293 infants born in Maryland in
2004-2005. In particular, large deficits in mean birth weight per one ln-unit increase in PFOS
concentration were found (P = -69; 95% CI: -149, 10; PFOS was detected in > 99% of samples
at a mean concentration of 0.005 |ig/mL). Maisonet et al. (2012, 1332465) evaluated fetal growth
outcomes in 395 singleton female births of participants in the Avon Longitudinal Study of
Parents and Children (ALSPAC) and found that increased maternal PFOS concentration (median
concentration of 0.0196 |ig/mL) was associated with lower birth weights, but not with lower 20-
month body weights. A study of 252 pregnant women in Alberta, Canada found no statistically
significant association between birth weight or gestation length and PFOS concentration
measured in maternal blood during the second trimester (mean concentration of 0.009 (j,g/mL)
(Hamm, 2010, 1290814), although mean birth weight increased slightly by increasing PFOS
tertiles (3,278 g for < 0.006 [j,g/mL; 3,380 g for 0.006-0.010 [j,g/mL; 3,387 g for > 0.010-
0.035 (j,g/mL). In a prospective cohort study in Japan (2002-2005), Washino et al. (2009) found
an inverse association between PFOS concentration in maternal blood during pregnancy (mean
PFOS concentration of 0.006 (j,g/mL) and birth weight. As noted in the 2016 HESD, these
researchers reported large reductions in mean birth weight (P = -149; 95% CI: -297.0, -0.5 g)
for each log-10 change in maternal PFOS concentration, especially among female infants
(P = -269.4; 95% CI: -465.7, -73.0 g). Chen et al. (2012, 1332466) examined 429 mother-infant
pairs from the Taiwan Birth Panel Study and found that umbilical cord blood PFOS
concentration (geometric mean of 5.94 ng/mL) was inversely associated with gestational age
(P = -0.37, 95% CI -0.60, -0.13, weeks), birth weight (P = -110.2, 95% CI -176.0, -44.5, g),
and head circumference (P = -0.25, 95% CI -0.46, -0.05, cm). Additionally, ORs for preterm
birth, low birth weight, and small for gestational age increased with PFOS exposure (adjusted
OR (95% CI) = 2.45 (1.47, 4.08), 2.61 (0.85, 8.03) and 2.27 (1.25, 4.15), respectively).
Some studies evaluated fetal growth parameters in the prospective Danish National Birth Cohort
(DNBC; 1996-2002) {Andersen, 2010, 1429893; Fei et al., 2007, 1005775; Fei, 2008,
2349574}. Maternal blood samples were taken in the first and second trimester. The median
maternal plasma PFOS concentration was 0.0334 [j,g/mL (range of 0.0064-0.1067 (j,g/mL). Fei et
al. (2007, 1005775) found no associations between maternal PFOS concentration (blood samples
taken in the first and second trimester) and birth weight. Also, as noted in the 2016 HESD, these
researchers found that ORs for preterm birth (OR range: 1.43-2.94) were consistent in
magnitude across the upper three PFOS quartiles, and that ORs for low birth weight (OR range:
3.39-6.00) were consistently elevated across the upper three quartiles. The HESD notes,
however, that analyses in this study were limited by small cell sizes due to low incidence of these
outcomes. Fei et al. (2008, 2349574) found an inverse association between maternal PFOS levels
and birth length and ponderal index in the DNBC in a stratified analysis, but the associations
were not statistically significant. Andersen et al. (2010, 1429893) examined the association
between maternal PFOS concentrations and birth weight, birth length, and infant body mass
index (BMI) and body weight at 5 and 12 months of age in DNBC participants. They found an
inverse association between PFOS concentration and birth weight in girls (P = -3.2; 95% CI:
-6.0, -0.3), 12-month body weight in boys (P = -9; 95% CI: -15.9, -2.2), and 12-month BMI in
boys (P = -0.017; 95% CI: -0.028, -0.005).
3-185
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Some studies described in the 2016 PFOS HESD evaluated developmental outcomes in the C8
Health Project study population, which comprises a community known to have been subjected to
high PFAS exposure. The C8 Health Project included pregnancies within 5 years prior to
exposure measurement, and many of the women may not have been pregnant at the time of
exposure measurement. Stein et al. (2009, 1290816) found an association between maternal
PFOS concentration and increased risk of low birth weight (adjusted OR =1.5; 95% CI: 1.1,1.9;
dose-related relationship for the 50th-75th, 75th-90th and > 90th percentile PFOS exposure
concentrations), but not pre-term birth. Mean PFOS serum concentration was 0.014 [j,g/mL.
Darrow et al. (2013, 2850966) evaluated birth outcomes in 1,630 singleton live births from 1,330
women in this study population and found an inverse association between maternal PFOS
concentration and birth weight (-29 g per log unit increase; 95% CI: -66, -7); they found no
association with preterm birth or low birth weight. Darrow et al. (2014, 2850274) and Stein et al.
(2009, 1290816) found no association between maternal serum PFOS and increased risk for
miscarriage in this population.
3.4.4.1.2Study Evaluation Considerations
There were multiple outcome-specific considerations that informed domain-specific ratings and
overall study confidence. For the Confounding domain, downgrading of studies occurred when
key confounders of the fetal growth and PFAS relationship, such as parity, were not considered.
Some hemodynamic factors related to physiological changes during pregnancy were also
considered in this domain as potential confounders (e.g., glomerular filtration rate and blood
volume changes over the course of pregnancy), because these factors may be related to both
PFOS levels and the developmental effects examined here. More confidence was placed in the
epidemiologic studies that adjusted for glomerular filtration rate in their regression models or if
they limited this potential source of confounding by sampling PFAS levels earlier in pregnancy.
An additional source of uncertainty was the potential for confounding by other PFAS (and other
co-occurring contaminants). Although scientific consensus on how best to address PFAS co-
exposures remains elusive, this was considered in the study quality evaluations and as part of the
overall weight of evidence determination.
For the Exposure domain, all the available studies analyzed PFAS in serum or plasma using
standard methods. Given the estimated long half-life of PFOS in humans as described in Section
3.3, samples collected during all three trimesters, before birth or and shortly after birth) were
considered adequately representative of the most critical in utero exposures for fetal growth and
gestational duration measures. The postnatal anthropometric studies were evaluated with
consideration of fetal programming mechanisms (i.e., Barker hypothesis) where in utero
perturbations, such as poor nutrition, can lead to developmental effects such as fetal growth
restriction and ultimately adult-onset metabolic-related disorders and related complications (see
more on this topic in {De Boo, 2009, 6937194} and {Perng, 2016, 6814341}). There is some
evidence that birth weight deficits can be followed by increased weight gain that may occur
especially among those with rapid growth catch-up periods during childhood {Perng, 2016,
6814341}. Therefore, the primary critical exposure window for measures of postnatal (and early
childhood) weight and height change is assumed to be in utero for study evaluation purposes,
and studies of this outcome were downgraded in the exposure domain if exposure data were
collected later during childhood or concurrently with outcome assessment (i.e., cross-sectional
analyses).
3-186
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Studies were also downgraded for study sensitivity, for example, if they had limited exposure
contrasts and/or small sample sizes, since this can impact the ability of studies to detect
statistically significant associations that may be present (e.g., for sex-stratified results). In the
Outcome domain, specific considerations address validation and accuracy of specific endpoints
and adequacy of case ascertainment for some dichotomous (i.e., binary) outcomes. For example,
birthweight measures have been shown to be quite accurate and precise, while other fetal and
early childhood anthropometric measures may result in more uncertainty. Mismeasurement and
incomplete case ascertainment can affect the accuracy of effect estimates by impacting both
precision and validity. For example, the spontaneous abortion studies were downgraded for
incomplete case ascertainment in the outcome domain given that some pregnancy losses go
unrecognized early in pregnancy (e.g., before implantation). This incomplete ascertainment,
referred to as left truncation, can result in decreased study sensitivity and loss of precision.
Often, this type of error can result in bias towards the null if ascertainment of fetal loss is not
associated with PFOS exposures (i.e., non-differential). In some situations, differential loss is
possible and bias away from the null and can manifest as an apparent protective effect. Fetal and
childhood growth restriction were examined using several endpoints including low birth weight,
small for gestational age (SGA), ponderal index (i.e., birth weight grams/birth length (cm3) x
100), abdominal and head circumference, as well as upper arm/thigh length, mean height/length,
and mean weight either at birth or later during childhood. The developmental effects synthesis is
largely focused on the higher quality endpoints (i.e., classified as good in the Outcome domain)
that were available in multiple studies to allow for an evaluation of consistency and other
considerations across studies. However, even when databases were more limited, such as for
spontaneous abortions, the evidence was evaluated for its ability to inform developmental
toxicity more broadly, even if available in only one study.
Overall, mean birth weight and birth weight-related measures are considered very accurate and
were collected predominately from medical records; therefore, more confidence was placed in
these endpoints in the Outcome domain judgments. Some of the adverse endpoints of interest
examined here included fetal growth restriction endpoints based on birth weight such as mean
birth weight (or variations of this endpoint such as standardized birthweight z-scores), as well as
binary measures such as SGA (e.g., lowest decile of birthweight stratified by gestational age and
other covariates) and low birth weight (i.e., typically < 2500 grams; 5 pounds, 8 ounces) births.
Sufficient details on the SGA percentile definitions and stratification factors as well as sources of
standardization for z-scores were necessary to be classified as good for these endpoints in this
domain. In contrast, other measures of fetal growth that are subject to more measurement error
(e.g., head circumference and body length measures such as ponderal index) were given a rating
of adequate {Shinwell, 2003, 6937192}. These sources of measurement error are expected to be
non-differential with respect to PFOS exposure status and, therefore, would not typically be a
major concern for risk of bias but could impact study sensitivity.
Gestational duration measures were presented as either continuous (i.e., per each gestational
week) or binary endpoints such as preterm birth (typically defined as gestational age <37
weeks). Although changes in mean gestational age may lack some sensitivity, especially given
the potential for measurement error, many of the studies were based on ultrasound measures
early in pregnancy, which should increase the accuracy of estimated gestational age and the
ability to detect associations that may be present. Any sources of error in the classification of
these endpoints would also be anticipated to be non-differential with respect to PFOS exposure.
3-187
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
While they could impact precision and study sensitivity, they were not be considered a major
concern for risk of bias.
3.4.4.1.3Study Inclusion
There are 78 studies from recent systematic literature search and review efforts conducted after
publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} that investigated the
association between PFOS and developmental effects. Although every study is included in the
study evaluation heat maps for comprehensiveness, eight developmental epidemiological studies
identified in the literature search were excluded for consideration in this synthesis because other
studies report results for the same health outcomes and from the same study cohorts (i.e., were
considered duplicative). For example, the Rokoff et al. (2018, 4238310) study overlapped with
the Project Viva study by Sagiv et al. (2018, 4238410). The Gennings et al. (2020, 7643497)
study is also not further considered here as it is a smaller subset of the Aarhus Birth Cohort
described in Wikstrom et al. (2020, 6311677). Similarly, the Li et al. (2017, 3981358)
Guangzhou Birth Cohort Study overlapped with a more recent study by Chu et al. (2020,
6315711). Four studies {Kishi, 2015, 2850268; Kobayashi, 2017, 3981430; Minatoya, 2017,
3981691; Kobayashi, 2022, 10176408} were also not considered in this synthesis, because they
provided overlapping data from the same Hokkaido Study on Environment and Children's Health
birth cohort population as Kashino et al. (2020, 6311632). For those Japanese studies with the
same endpoints such as mean birthweight (BWT), the analysis with the largest sample size was
used in forest plots and tables (e.g., Kashino et al., (2020, 6311632)). Although the Kobayashi et
al. (2017, 3981430) study included a unique endpoint called ponderal index, this measure is
more prone to measurement error and was not considered in any study given the wealth of other
fetal growth restriction data. Similarly, the Costa et al., (2019, 5388081) study that examined a
less accurate in utero growth estimate was not considered in lieu of their more accurate birth
outcomes measures reported in the same cohort {Manzano-Salgado, 2017, 4238465}. One
additional study by Bae et al. (2015, 2850239) was the only study to examine sex ratio and was
not further considered here.
In general, to best gauge consistency and magnitude of reported associations U.S. EPA largely
focused on the most accurate and most prevalent measures within each fetal growth endpoint.
Two other studies with overlapping cohorts were included in the synthesis, as each study
provided some unique data for different endpoints. For example, the Woods et al. (2017,
4183148) publication on the Health Outcomes and Measures of the Environment (HOME) cohort
overlaps with Shoaff et al. (2018, 4619944) but has additional mean BWT data (communication
with author). The mean BWT results for singleton and twin births from Bell et al. (2018,
5041287) are included in forest plots here as are the postnatal growth trajectory data in the same
UPSTATE KIDS cohort by Yeung et al. (2019, 5080619) as they target different developmental
windows. The Bjerregaard-Olesen et al. (2019, 5083648) study from the Aarhus birth cohort also
overlaps with Bach et al. (2016, 3981534). The main effect results are comparable for head
circumference and birth length in both studies despite a smaller sample size in the Aarhus birth
cohort subset examined in Bjerregaard-Olesen et al. (2019, 5083648). Given that additional sex-
specific data are available in the Bjerregaard-Olesen et al. (2019, 5083648) study, the synthesis
for head circumference and birth length are based on this subset alone. Chen et al., (2021,
7263985) reported an implausibly large effect estimate for head circumference. After
correspondence with study authors, an error was identified, and the study was not considered for
head circumference.
3-188
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Following exclusion of the nine studies noted above, 69 developmental epidemiological studies
were included in the synthesis that were not included in the 2016 HESD report. Six additional
studies {Alkhalawi, 2016, 3859818; Gundacker, 2021, 10176483; Jin, 2020, 6315720; Lee,
2013, 3859850; Lee, 2016, 3981528; Maekawa, 2017, 4238291} were considered uninformative
due to critical study deficiencies in some risk of bias domains (e.g., confounding) or multiple
domain deficiencies and are not further examined here. Thus, 63 studies were included across
various developmental endpoints for further examination and synthesis.
Forty-three of the 63 different studies examined PFOS in relation to fetal growth restriction
measured by the following endpoints: small for gestational age (SGA), low BWT, head
circumference, as well as mean and standardized BWT and birth length measures. Twenty-two
studies examined gestation duration, twelve examined post-natal growth, five each examined
fetal loss, and birth defects.
3.4.4.1.4Growth Restriction: Fetal Growth
3.4.4.1.4.1 Birth Weight
Of the 40 informative and non-overlapping studies that examined BWT measures in relation to
PFOS exposures, 34 studies examined mean BWT differences. Fifteen studies examined
standardized BWT measures (e.g., z-scores) with nine of these reporting results for mean and
standardized BWT {Ashley-Martin, 2017, 3981371; Bach, 2016, 3981534; Eick, 2020, 7102797;
Gyllenhammar, 2018, 4238300; Meng, 2018, 4829851; Sagiv, 2018, 4238410; Wang, 2019,
5080598; Wikstrom, 2020, 6311677; Workman, 2019, 5387046}. Twenty-five of the 34 mean
BWT studies shown in Figure 3-46 and Figure 3-47 provided results based on a prospective birth
cohort study design, and the remaining nine were cross-sectional analyses defined here as if
biomarker samples were collected at birth or post-partum {Bell, 2018, 5041287; Callan, 2016,
3858524; de Cock, 2016, 3045435; Gao, 2019, 5387135; Gyllenhammar, 2018, 4238300; Kwon,
2016, 3858531; Shi, 2017, 3827535; Wang, 2019, 5080598; Xu, 2019, 5381338}.
Overall, eight of the PFOS studies relied on umbilical cord measures {Cao, 2018, 5080197; de
Cock, 2016, 3045435; Govarts, 2016, 3230364; Kwon, 2016, 3858531; Shi, 2017, 3827535;
Wang, 2019, 5080598; Workman, 2019, 5387046; Xu, 2019, 5381338}, and one collected blood
samples in infants 3 weeks following delivery {Gyllenhammar, 2018, 4238300}. Results from
the Bell et al. (2018, 5041287) study were based on infant whole blood taken from a heel stick
and captured onto filter paper cards at 24 hours or more following delivery, and one study used
both maternal serum samples collected 1-2 days before delivery and cord blood samples
collected immediately after delivery {Gao, 2019, 5387135}. One study examined pre-conception
maternal serum samples {Robledo, 2015, 2851197}. Twenty-one studies had maternal serum or
plasma PFOS measures that were sampled during trimesters one {Ashley-Martin, 2017,
3981371; Bach, 2016, 3981534; Lind, 2017, 3858512; Manzano-Salgado, 2017, 4238465; Sagiv,
2018, 4238410}, two {Lauritzen, 2017, 3981410}, or three {Callan, 2016, 3858524; Chu, 2020,
6315711; Kashino, 2020, 6311632; Luo, 2021, 9959610; Valvi, 2017, 3983872; Yao, 2021,
9960202}, or across multiple trimesters {Chang, 2022, 9959688; Chen, 2021, 7263985; Eick,
2020, 7102797; Hjermitslev, 2020, 5880849; Lenters, 2016, 5617416; Marks, 2019, 5081319;
Starling, 2017, 3858473; Wikstrom, 2020, 6311677; Woods, 2017, 4183148}. The study by
Meng et al. (2018, 4829851) pooled exposure data from two study populations, one which
measured PFOS in umbilical cord blood and one which measured PFOS in maternal blood
samples collected in trimesters 1 and 2. For comparability with other studies of mean BWT, only
3-189
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
one biomarker measure was used here (e.g., preferably maternal samples when collected in
conjunction with umbilical cord samples or maternal only when more than parent provided
samples). In addition, other related publications (e.g., Gyllenhammar et al., 2017, 7323676)) or
additional information or data (e.g., Woods et al., 2017, 4183148)) provided by study authors
(communication with author) were used.
Fifteen of the 34 mean BWT studies included in the synthesis were rated high in overall study
confidence {Ashley-Martin, 2017, 3981371; Bach, 2016, 3981534; Bell, 2018, 5041287; Chu,
2020, 6315711; Eick, 2020, 7102797; Govarts, 2016, 3230364; Lauritzen, 2017, 3981410; Lind,
2017, 3858512; Luo, 2021, 9959610; Manzano-Salgado, 2017, 4238465; Sagiv, 2018, 4238410;
Starling, 2017, 3858473; Valvi, 2017, 3983872; Wikstrom, 2020, 6311677; Yao, 2021,
9960202}, while twelve were rated medium {Chang, 2022, 9959688; Chen, 2021, 7263985; de
Cock, 2016, 3045435; Gyllenhammar, 2018, 4238300; Hjermitslev, 2020, 5880849; Kashino,
2020, 6311632; Kwon, 2016, 3858531; Lenters, 2016, 5617416; Meng, 2018, 4829851;
Robledo, 2015, 2851197; Wang, 2019, 5080598; Woods, 2017, 4183148}, and seven were
classified as low {Callan, 2016, 3858524; Cao, 2018, 5080197; Gao, 2019, 5387135; Marks,
2019, 5081319; Shi, 2017, 3827535; Workman, 2019, 5387046; Xu, 2019, 5381338}. Twenty-
three of the twenty-seven high or medium confidence studies detailed in this synthesis were
classified as having good study sensitivity {Ashley-Martin, 2017, 3981371; Bach, 2016,
3981534; Chen, 2021, 7263985; Gyllenhammar, 2018, 4238300; Hjermitslev, 2020, 5880849;
Kashino, 2020, 6311632; Lauritzen, 2017, 3981410; Lenters, 2016, 5617416; Lind, 2017,
385812; Manzano-Salgado, 2017, 4238465; Meng, 2018, 4829851; Robledo, 2015, 2851197;
Sagiv, 2018, 4238410; Starling, 2017, 3858473; Wikstrom, 2020, 6311677; Valvi, 2017,
3983872; Woods, 2017, 4183148} or adequate study sensitivity {Chang, 2022, 9959688; Chu,
2020, 6315711; Eick, 2020, 7102797; Govarts, 2016, 3230364; Luo, 2021, 9959610; Yao, 2021,
9960202}, while four had deficient study sensitivity {Bell, 2018, 5041287; de Cock, 2016,
3045435; Kwon, 2016, 3858531; Wang, 2019, 5080598} as shown in Figure 3-46, Figure 3-47,
and Figure 3-48. The median exposure values across all of the studies were quite variable and
ranged from 0.38 ng/mL {Kwon, 2016, 3858531} to 30.1 ng/mL {Meng, 2018, 4829851}.
3-190
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Alkhalawi et al., 2016, 3859818
Ashley-Martin et al„ 2017, 3981371
Bach etal., 2016, 3981534
Bell et al., 2018, 5041287
Bjerregaard-Olesen et al., 2019, 5083648
Callan etal., 2016, 3858524
Cao etal., 2018, 5080197
Chang et al., 2022, 9959688
Chen etal., 2017, 3981292
Chen etal., 2021, 7263985
Chu etal., 2020, 6315711
Costa et al., 2019, 5388081
Eick et al., 2020, 7102797
Espindola Santos et al., 2021, 8442216
Gao etal., 2019, 5387135
Gennings et al., 2020, 7643497
Govarts et al., 2016, 3230364
Gross etal., 2020, 7014743
Figure 3-46. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Weight Effects3
Interactive figure and additional study details available on HAWC.
aIncludes six overlapping studies (Bjerregaard-Olesen, 2019, 5083648; Kishi, 2015, 2850268; Kobayashi, 2017, 3981430; Li,
2017, 3981358; Minatoya, 2017, 3981691; Rokoff, 2018,4238310).
3-191
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
A\0^
0*^
,G®
Gundacker etal., 2021, 10176483
Gyllenhammar et al., 2018, 4238300-
Hjermitslev et al., 2020, 5880849
Jin et al., 2020, 6316202-
Kashino et al., 2020, 6311632
Kishi et al., 2015, 2850268
Kobayashi et al., 2017, 3981430
Kobayashi et al., 2022, 10176408
Kwon et al„ 2016, 3858531 H
Lauritzen et al„ 2017, 3981410
Lee etal., 2013, 3859850
Lee et al., 2016, 3981528
Lenters et al., 2016, 5617416
Li et al., 2017, 3981358
Lindetal., 2017, 3858512
Luo etal., 2021, 9959610
Maekawa et al., 2017, 4238291
Manzano-Salgado et al., 2017, 4238465
Legend
| Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-47. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Weight Effects (Continued)3
Interactive figure and additional study details available on HAWC.
aIncludes six overlapping studies (Bjerregaard-Olesen, 2019, 5083648; Kishi, 2015, 2850268; Kobayashi, 2017, 3981430; Li,
2017, 3981358; Minatoya, 2017, 3981691; Rokoff, 2018,4238310).
3-192
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
,o®
Marks etal., 2019, 5081319-
Meng etal., 2018, 4829851 -
Minatoya et al., 2017, 3981691 -
Robledo et al., 2015, 2851197-
Rokoff et al., 2018, 4238310 -
Sagivetal., 2018, 4238410-
Shi etal., 2017, 3827535-
Shoaff etal., 2018, 4619944
Starling et al., 2017, 3858473
Valvi etal., 2017, 3983872
Wang et al., 2019, 5080598 - +
Wikstrom et al., 2020, 6311677
Woods et al., 2017, 4183148 -
Workman et al., 2019, 5387046
Xiao etal., 2020, 5918609
Xu etal., 2019, 5381338-
Yao etal., 2021,9960202
de Cook et al., 2016, 3045435 -
II 1 1 1 1 1 1
-
++
++
-
+
+
+
+
D
++
+
++
+
++
+
++
++
++
+
++
+
+
++
++
B
++
+
++
+
++
J
++
++
++
+
++
+
+
++
~1
+
+
++
+
1 ++ ++
-
+
+
| +
n
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-48. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Weight Effects (Continued)3
Interactive figure and additional study details available on HAWC.
3-193
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.4.1.4.1.1 Mean Birth Weight Study Results: Overall Population Studies
Thirty of the 34 included studies that examined mean BWT data in the overall population {Bach,
2016, 3981534; Bell, 2018, 5041287; Callan, 2016, 3858524; Cao, 2018, 5080197; Chang, 2022,
9959688; Chen, 2021, 7263985; Chu, 2020, 6315711; de Cock, 2016, 3045435; Eick, 2020,
7102797; Gao, 2019, 5387135; Govarts, 2016, 3230364; Gyllenhammar, 2018, 4238300;
Hjermitslev, 2020, 5880849; Kashino, 2020, 6311632; Kwon, 2016, 3858531; Lauritzen, 2017,
3981410; Lenters, 2016, 5617416; Luo, 2021, 9959610; Manzano-Salgado, 2017, 4238465;
Marks, 2019, 5081319; Meng, 2018, 4829851; Robledo, 2015, 2851197; Shi, 2017, 3827535;
Starling, 2017, 3858473; Valvi, 2017, 3983872; Wikstrom, 2020, 6311677; Woods, 2017,
4183148; Wu, 2012, 2919186; Xu, 2019, 5381338; Yao, 2021, 9960202}, while four only
reported sex-specific data only {Ashley-Martin, 3981371; Lind, 2017, 3858512; Marks, 2019,
5081319; Robledo, 2015, 2851197}. Nineteen of the 30 PFOS studies with analyses based on an
overall population reported some mean BWT deficits, albeit some of these were not statistically
significant (Figure 3-49, Figure 3-50, Interactive figure and additional study details available on
Tableau.
Effect Estimate
Confidence Rating Sampling Period
Sludy Deaiiyi Exposure Malrix Exposure levels
Cohort maternal plasma Mean (SD* 6.05 ngtatL <2.74 ngAnL)
M al.. 2020 Cotrat
Starling et al., 2017 Cohort
Valvi etal.. 2017
Cohort maternal sc
iian-25.7 ngAnL (IOR 16.0 ngVnL)
Regression coefficient for bit
Regression coefficient per IOR inc
npfnL])
Regression coefficient (ft* Q3 re Q1)
Regression coefficient (Tor 04 re Q1)
Regression coefficient (for Q3 re Q1)
Regression coefficient (fcr 04 re 01 *|
nt (for 02(18.9 25.6 ngimL) re Q110.1 -'
nt (per 1 -fci ngi'mL change in PFOS)
iian: 4.55 ng'mL (range: Q.55 29.65 ng/mL)
Regression coefficient [per doubling of serum PFOS]
Regression coefficient (per 1 tn ngi'mL increase in male
Figure 3-51, Interactive figure and additional study details available on Tableau.
3-194
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Confidence Rainq
Sampling Period
Reference
Sludy Deaiyi E]
pgsure Matrix Exposure levels
-200
Comparison EE
Effect Eslriiale
-100 0 100 200 300 400
Medium confidence
Early pregnancy
Chang etaL, 2022
Uo.
alernal serum median: 2.19 ngM.
(25th-75lh percerrtie:
1.45-3.24 npVnL>
Regression coefficient (per doubling in PFOSl -7.0
1
1
1
H-
I
I
i
Regression coefficienl [for 02 (1.44-2.19 n^YnL) vs. 01 _n n
(1899 n^LI vs. T1 (<1200 .
nffl) a3a'4
i
i
i
i
i
i
i
Gylenhammar et al., 2013
alernal serum Nul
Regression coefficient per unit log increase in PFOS -39.5
i
i
i
• ¦
i
i
i
-200 -100 0 100 200 300 400
Figure 3-52, Figure 3-53, and Interactive figure and additional study details available on
Tableau.
3-195
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Effect Estrnale
Reference
Study Degigi
Exposure Matrix Exposure levels
Comparison EE
400 200 0 200 400
Calanetd-, 2016
median-1.99 ngi'mL (range:
Regression coefficient (per one ln-i*irt change in „ „
1
1
1
maternal blood 0.45-8.1 ng/mL)
maternal PFOS) 69 0
1
1
i
Cao el al., 2018
Cohort
card blood rnedian^l.01 ng/mL (25lh-75th
percentile: 0.60 1.76 npYnL)
Regression coefficient for T2(0.74-1-52 ng'mL) vs T1 ( c
<0.74 ng'mL) 103'5
I
l
I
H ~
1
1
1
Regression coefficient far T3(>1.52 ng/mL) vs. T1 (<0.74 _
npVnL) 1/6
1
1
1
—H—
i
i
i
Gao Hal., 2019
maternal blood, cord Median (Min-Max) = 4.07 n^'inL
blaod (LOD 22.6 ng/mL)
Regression coefficient (maternal serum PFOS 3.09-5.16 ,
*s <3.09 ngltnL) w"'
i
i
i
1 •
i
i
i
Regression coefficienl (maternal serum PFOS >5.16 vs n _
<3.09 ng/mL) °-5
i
i
4
l
l
l
Shi etsL. 2017
Cross-sectional
. Median-0.974 ngVnL (25*i-75th
cora ouoo percentile: 0.626 1.584 n^rnt)
Regression coefficienl (per 10-fald change in PFOS) 160.4
l
l
l
J •
l
l
l
Workman Mai.. 2019
Cohort
median= 2.2 ng>mL (range:
Regression coefficient (per 1 -fci ng/mL increase in „
I
I
maternal nicoa Q.18-21 nglmL)
maternal prenatal PFOS) 50 0
I
I
I
XuetaL. 2019
Median [2&tfi-7Slh percentiles)^
Regression coefficienl (per loglO-unit change in PFOS) -417.3
1
I
I
i
card blood 4.07 ngftnL (2.88-8.05 ngihiL)
l
l
l
400 -200 0 200 400
Figure 3-54). Nine mean BWT studies in the overall population reported null associations {Cao,
2018, 5080197; Chang, 2022, 9959688; Chen, 2021, 7263985; Eick, 2020, 7102797; Gao, 2019,
5387135; Govarts, 2016, 3230364; Hjermitslev, 2020, 5880849; Manzano-Salgado, 2017,
4238465; Woods, 2017, 4183148}, while two reported increased mean BWT deficits {de Cock,
2016, 3045435; Shi, 2017, 3827535}. Only two studies {Starling, 2017, 3858473; Sagiv, 2018,
4238410} out of ten studies which examined categorical data {Bach, 2016, 3981534, Cao, 2018,
5080197; Chang, 2022, 9959688; Eick, 2020, 7102797; Gao, 2019, 5387135; Govarts, 2016,
3230364; Manzano-Salgado, 2017, 4238465; Meng, 2018, 4829851; Sagiv, 2018, 4238410;
Starling, 2017, 3858473; Wikstrom, 2020, 6311677} showed inverse monotonic exposure-
response relationships. Although two studies {Bach, 2016, 3981534; Meng, 2018, 4829851} also
showed large BWT deficits consistent in magnitude in the upper two quartiles (-50 to -62 g and
-50 to -48 g relative to their quartile 1 referents, respectively).
Although there was a wide distribution of BWT deficits (range: -14 to -417 grams) in the overall
population (i.e., both sexes combined) across both categorical and continuous exposure
estimates, 18 of these ranged from -14 to -93 grams per each PFOS unit increase. This included
all 10 high confidence studies with five of these reporting deficits ranging from 14 to 18 grams
per each unit PFOS increase. The six medium confidence studies reporting deficits showed larger
associations with an even narrower distribution ranging -35 to -69 grams per each unit PFOS
increase. The three low confidence studies reporting deficits showed the largest associations
ranging -0 to -417 grams per each unit PFOS increase including three studies ranging from -50
to -69 grams. Thus, there was some suggestion of larger and more variable BWT deficits in low
confidence studies which have a higher potential for bias. There was also a preponderance of
3-196
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
inverse associations based on studies with later biomarker sampling timing (i.e., trimester two
onward) including 15 of the overall 19 studies and 7 of the 10 high confidence studies only; this
may be related to pregnancy hemodynamic influences on the PFOS biomarkers during
pregnancy.
3.4.4.1.4.1.2 Mean BWT-Overall Population Summary
Eighteen of the nineteen studies that reported deficits based on either categorical or continuous
expression ranged from -14 to -93 grams. A pattern of larger and more variable results was
detected across study confidence with smaller and less variable BWT deficits among the higher
confidence studies. Overall, there was evidence of an adverse monotonic exposure-response in
two of ten studies, but an additional two studies showed large and consistent results in the upper
two quartiles. Most of the evidence of mean birth weight difference was detected among the
medium (6 of 12) or high (10 of 15) confidence studies. Study sensitivity was not an explanatory
factor of the null BWT studies. There was some suggestion of a relationship between PFOS
sample timing and magnitude of associations with the six of the largest deficits detected among
studies that used maternal serum with some or all samples collected during trimester 3 or were
based on umbilical cord samples. There was also a preponderance of inverse associations based
on studies with later biomarker sampling timing (i.e., trimester two onward) that may be related
to pregnancy hemodynamic influences on the PFOS biomarkers during pregnancy.
3-197
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
fidenoeRainn Sampling Period Reference Sludy Design Malm Expo
i confidence Early pregnancy Bach rtal.. 2016 Cohort maternal median-a.3 np'rri (25lb-75lh percentile: 6.0-10.8 ngfmLl Regression coefficient per IOR (4.8 n^knL) increase -14.
st pregnancy Sell el al.. 2018 Cross sc
Chu el al., 2020 Cohort
Eickatal.,2020 Cohtrt
edian-1.72 ngfmL (25lh 75lh perce rrtte: 1.14-2.44 ng/mL)
1=7.153noVnL |2Sth percentile-4.361 na'rnL, 75tfi _ _ . . _
percentile-11.928 n^VnL) Regress.cn coefBoent (p« 11n chaise in PFOS| -83.3
lian= 1.93 ng/mL (2Stti-7Stti percentile- 1.18-3.13 ngiWiL)
Coefficient [for T2 <1.40-2.56 nglirt) vs. T1
Regression Coefficient [for T3 (>2.56 n^rnl) vs. T1
(<1.40 n»W)] 143
Effect Estimate
-100 0 100
0 100
Figure 3-49. Overall Mean Birth Weight from Epidemiology Studies Following Exposure to
PFOS
Interactive figure and additional study details available on Tableau.
3-198
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Effect Estimate
x Rating Sampling P*
Study Dcajyn M*i* Expo
onficencc Earty pregnancy Ma-uano SalgadD e!t al., Cohort
=r pregnancy Govarts et al„ 2016 Cohort
at aL, 2017 Cohort
I.. 2021 Cohcrt
Moan (SD* 6.05 ng.hiL (2.74 ng/ntL|
doubling at PTOS;
wbiith weight |Q2wQ1) 23.6
Ih weight |Q3 vs 011 36.7
oefficienl for birth weight |Q4 vsQ1) 8.2
nt (per IQR change in PFOS
d, oord median |25»i-75th percentile): S.01 ngTmL (3.32-7.62)
in InPFOS -15.1
nt (per h n^VnL increase PFOS) -93.3
Figure 3-50. Overall Mean Birth Weight from Epidemiology Studies Following Exposure to
PFOS (Continued)
Interactive figure and additional study details available on Tableau.
3-199
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Confidenoe Rating Sampling Penod
Study Design Exposure
Saijiv el al., 2G1B Cohort
Wikslram a al.. 2020 Cohort
Starting el al.. 2017 Cohort
Valvi Dt al.. 2017 Cohort
seal. 2021 Cross se
Exposure levels
Moan (SD>. 6.05 ngiYitL (2.74 ngfmL)
coefficient for birth waighs (Q3 re Ot J
coefficient liar birth wegh! |Q4 re Ql |
!5.7 ngftnL (IOR: ?8.©ti(j!mL) Regression coefficient per IQR irxzeasc
Regression coefficient (for 02 [1B.9 25.B ng'mLj re Q1|0.1
nginL])
coefficient (for Q3 re 01
200 100
Regression coefficient (fcr 04 re Q
coefficient (for Q3 re O
Regression coefficient |fcr Q4 re O
coefficient (for 02 re O
coefficient [per deuiifeig of serum Pros]
edian: d.56 nglinL (range. 0.55 29.65 ngrmLl SjrgsJ"
200 100
Figure 3-51. Overall Mean Birth Weight from Epidemiology Studies Following Exposure to
PFOS (Continued)
Interactive figure and additional study details available on Tableau.
3-200
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Confidence Rainq
Sampling Period
Reference
Sludy Deaiyi E]
pgsure Matrix Exposure levels
-200
Comparison EE
Effect Eslriiale
-100 0 100 200 300 400
Medium confidence
Early pregnancy
Chang etaL, 2022
Uo.
alernal serum median: 2.19 ngM.
(25th-75lh percerrtie:
1.45-3.24 npVnL)
Regression coefficient (per doubling in PFOS) -7.0
1
1
1
H-
I
I
i
Regression coefficient [for 02 (1.44-2.19npVnL)vs. 01 7n n
<inL)] 'BU
I
I
I
—1
I
I
I
Regression coeffinenl [for 03 (2.19-3.24 n^VnL) vs. 01
(1899 n^L) vs. T1 (<1200 .
nffl) a3a'4
i
i
i
i
i
i
i
Gylenhammar et al., 2013
alernal serum Nul
Regression coefficient per unit log increase in PFOS -39.5
i
i
i
• ¦
i
i
i
-200 -100 0 100 200 3O0 400
Figure 3-52. Overall Mean Birth Weight from Epidemiology Studies Following Exposure to
PFOS (Continued)
Interactive figure and additional study details available on Tableau.
3-201
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
edian-8.99 ng/mL (min-max: 1.50-61.3) Regression coefficient (per 11n-ngihtL increase in PFOS) 5.6
Effect Estimate
-150 -100 -50
eng et al., 2018 Cohort
.1 ng'mL Regression coefficient (per doubling of PFOS) -45.2
Regression coefficient for Q2 vs. Q1
Regression coefficient for 03 vs. 01
~efficient far Q4 vs. 01
~efficient (per bglfl change in PFOS)
Kwon et al., 2016 Cross se
Regression coefficient (per 1 log-unit change in PFOS) -49.4
Wang el al., 2019 CtnsMeESaiul card b.
~efficient (per 1 -bg10 change in PFOS) -54.5
Woods etal., 2017 Cohort
regression coefficient (per log10-ug'L in<
-150 -100 -SO
Figure 3-53. Overall Mean Birth Weight from Epidemiology Studies Following Exposure to
PFOS (Continued)
Interactive figure and additional study details available on Tableau.
3-202
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Effect Estrnale
Confidenoe Raflng Sampling Period Reference Study Design Exposure Matrix Expo!
er pregnancy Calan et s(_, 2016 Cross se
Caoetal.,2018 Cohort
Shi etaL. 2017 Cross se
1.. 2019 Cohort
ri.01 ngjftiL <25lh-75th Ro^on coefficient far T2(0.74-1^2 n^W.)« T1 (
te: 0.60-1.76 rnVnL) <0 74 n9"nL>
coefficient far T3(>1.52 ngAnL) vs. T1 (<0.7^
maternal serum PFOS 3.09-5.16
^efficient (maternal serum PFOS >5.16 v=
coefficient (per 10-fald change in PFOS) 160.4
nt (per taglO-unit change in PFOS) -417.3
| 400 -200
200 400
| 400 -200
200 400
Figure 3-54. Overall Mean Birth Weight from Epidemiology Studies Following Exposure to
PFOS (Continued)
Interactive figure and additional study details available on Tableau.
3.4.4.1.4.1.3 Mean Birth Weight Study Results: Sex Specific Studies
Ten of sixteen epidemiological studies examining sex-specific results in male neonates showed
some BWT deficits. The remaining six studies {Ashley-Martin, 2017, 3981371; Cao, 2018,
5080197; de Cock, 2016, 3045435; Hjermitslev, 2020, 5880849; Robledo, 2015, 2851197; Shi,
2017, 3827535} in male neonates were either null or showed larger birth weights with increasing
PFOS exposures. Six of fifteen epidemiological studies examining sex-specific results in female
neonates showed some BWT deficits. The magnitude of associations was much more variable in
boys (range: -9 to -150 grams) than in girls (range: -20 to -85 grams) per each unit PFOS
increase. There was also little evidence of exposure-response relationships in either sex as only
1 out of 5 studies with categorical data showed monotonicity.
Six of the 15 studies examining mean BWT associations in both boys and girls detected some
deficits in both sexes. Two of these six studies showed sex-specific deficits comparable in
magnitude among boys and girls {Chu, 2010, 6315711; Wang, 2019, 5080598}. Three of these
studies {Bach, 2016, 3981534; Meng, 2018, 4829851; Wikstrom, 2020, 6311677} showed larger
deficits among girls and one showed larger deficits among boys {Kashino, 2020, 6311632}. The
low confidence study by Marks et al. (2019, 5081319) of males only detected a small statistically
significant association (-8.5 g; 95% CI: -15.9, -1.1) per each ln-unit PFOS increase and showed
an exposure-response with reported large deficits in PFOS tertile 2 (-26.6 g; 95% CI: -147.3,
3-203
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
94.2) and tertile 3 (-83.9 g; 95% CI: -201.4, 33.7) compared to the tertile 1 referent. Four other
studies reported mean BWT deficits only in boys {Lind, 2017, 3858512; Manzano-Salgado,
2017, 4238465; Valvi, 2017, 3983872}; no studies reported deficits in girls only.
Overall, there was more evidence of adverse associations detected in boys, but the magnitude of
associations detected was more consistent in girls. There was an exposure-response relationship
detected in only one of five studies with categorical data in both sexes. Study confidence and
most other study characteristics did not seem to be explanatory patterns for the results, as, for
example, nearly all (9 of 10 in boys) or all (6 of 6 girls) were either high or medium confidence.
Definitive patterns by sample timing were also not evident in the male neonates across all study
confidence levels but a larger proportion of the later sampled studies (60%) showed inverse
associations in females compared to early sampled studies (38%). Study sensitivity was not an
explanatory factor among the null studies in either sex.
3.4.4.1.4.1.4 Standardized Birth Weight Measures
Fifteen studies examined standardized BWT measures including fourteen studies reporting a
change in BWT z-scores on a continuous scale per each PFOS comparison. Eight of the 15
studies were high confidence studies {Ashley-Martin, 2017, 3981371; Bach, 2016, 3981534;
Eick, 2020, 7102797; Gardener, 2021, 7021199; Sagiv, 2018, 4238410; Shoaff, 2018, 4619944;
Wikstrom, 2020, 6311677; Xiao, 2019, 5918609}, four were medium {Chen, 2017, 3981292;
Gyllenhammar, 2018, 438300; Meng, 2018, 4829851; Wang, 2019, 5080598} and three were
low confidence {Espindola-Santos, 2021, 8442216; Gross, 2020, 7014743; Workman, 2019,
5387046} (Figure 3-46, Figure 3-47, Figure 3-48).
Nine of the fifteen studies showed some evidence of adverse associations between PFOS
exposures and BWT z-scores. Six of these were high confidence {Bach, 2016, 3981534;
Gardener, 2021, 7021199; Sagiv, 2018, 4238410; Shoaff, 2018, 4619944; Wikstrom, 2020,
6311677; Xiao, 2019, 5918609}, two were medium confidence {Chen, 2017, 3981292; Wang,
2019, 5080598} and one was low confidence {Gross, 2020, 7014743}. None of the four studies
reporting categorical data showed evidence of monotonicity across tertiles or quartiles. The high
confidence study by Gardener et al. (2021, 7021199) reported that participants in the highest
PFOS exposure quartile (relative to the lowest quartile) had a higher odds (OR = 1.41; 95% CI:
0.66, 2.03) of being in the lowest standardized birthweight category (vs. the top 3 BWT z-score
quartiles). Four studies reporting associations in the overall population also reported standardized
birth weight deficits in either or both male and female neonates. Two studies {Gardener, 2021,
7021199; Gyllenhammar, 2018, 4238300} also reported that there were no statistically
significant interactions for their BWT-z measures by sex.
Among the fourteen studies examining continuous BWT z-score measures in the overall
population, eight reported associations for different PFOS exposures. The high confidence study
by Bach et al. (2016, 3981534) reported a statistically significant association between mean
BWT z-score and PFOS quartiles 2 (-0.15; 95% CI: -0.29, -0.02) and quartile 4 (-0.11; 95%
CI: -0.25, 0.02) only, with no exposure-response relationship detected. Although not statistically
significant, both Wang et al. (2019, 5080598) (-0.15; 95% CI: -0.41, 0.11) and Shoaff et al.
(2018, 4619944) reported associations similar in magnitude for their overall population (-0.12;
95% CI: -0.36, 0.13). The medium confidence study by Chen et al. {2017, 3981292} reported
adverse associations in the overall population (-0.14; 95% CI: -0.26, -0.01) with comparable
3-204
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
results in both male and female neonates (BWT z-score range: -0.13 to -0.15). The high
confidence study by Sagiv et al. (2018, 4238410) reported associations for PFOS quartile 4 in the
overall population (-0.13; 95% CI: 0.26, 0.00); the largest association in this study was found for
male neonates (-0.19; 95% CI: -0.33, -0.05) per each interquartile range (IQR) increase. The
high confidence study by Wikstrom et al. {2020, 6311677} reported adverse associations per
each ln-unit increase (-0.10; 95%CI: -0.20; -0.004) as well as in quartile 4 in the overall
population (-0.17; 95% CI: -0.37, -0.03); these results appeared to be driven by associations
detected in female neonates (-0.17; 95% CI: -0.30, -0.03 per each ln-unit increase; -0.30; 95%
CI: -0.49, -0.10 for quartile 4). The high confidence study by Xiao et al. (2019, 5918609)
reported z-scores fairly similar in magnitude for the overall population (-0.47; 95% CI: -0.85,
-0.09), male neonates (-0.40; 95% CI: -0.89, 0.08), and female neonates (-0.56; 95% CI:
-1.12, 0). Among the eight studies showing some deficits, the largest association was detected in
the low confidence study by Gross et al. (2020, 7014743) for the overall population (-0.62; 95%
CI: -0.96 to -0.29). The authors also reported large deficits for both males (-0.81; SE=0.24; p-
value=0.001) and females (-0.46; SE=0.29; p-value=0.11) for PFOS levels greater than the mean
level.
3.4.4.1.4.1.5 BWT z-score Summary
Nine out of 15 studies showed some associations between standardized BWT scores and PFOS
exposures including eight medium or high confidence studies. None of the five studies with
categorical data reported strong evidence of exposure-response relationships. No patterns by
sample timing were evident as three of these studies had trimester one maternal samples;
however, the strongest associations were seen in studies with later biomarker sampling. Study
sensitivity did not seem to be an explanatory factor in the six null studies of standardized BWT
most of these studies had moderate or large exposure contrasts and sufficient sample sizes.
Although some studies may have been underpowered to detect associations small in magnitude
relative to PFOS exposure, there was consistent lower BWT z-scores reported in these studies.
There was no apparent pattern related to magnitude of deficits across study confidence, but more
associations were evident across high confidence levels in general. Twice as many studies
showing adverse associations were based on later (6 of 9) versus early (i.e., at least some
trimester one maternal samples) pregnancy sampling (3 of 9); this might be reflective of some
impact of pregnancy hemodynamics on biomarker concentrations over time. Few differences
were seen across sexes including magnitude of associations as the majority of studies in both
male (3 of 5 studies; 2 were medium or high confidence) and female (4 of 5 studies; 3 of 4 were
medium or high confidence) neonates showed some associations between decreased standardized
birth weights and increasing PFOS exposures. Overall, nine different studies out of fifteen
showed some suggestion of inverse associations in the overall population or either or both sexes.
3.4.4.1.4.2 Small for Gestational Age/Low Birth Weight
Ten informative and non-overlapping epidemiological studies examined associations between
PFOS exposure and different dichotomous fetal growth restriction endpoints, such as SGA (or
related intrauterine growth retardation endpoints), LBW, or both (i.e., Manzano-Salgado et al.
(2017, 4238465)). Overall, eleven studies examined either or both LBW or SGA in relation to
PFOS exposure with four classified as high confidence {Chu, 2020, 6315711; Lauritzen, 2017,
3981410; Manzano-Salgado, 2017, 4238465; Wikstrom, 2020, 6311677}, three as medium
confidence {Govarts, 2018, 4567442; Hjermitslev, 2020, 5880849; Meng, 2018, 4829851}, three
as low {Chang, 2022, 9959688; Souza, 2020, 6833697; Xu, 2019, 5381338} and one as
3-205
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
uninformative {Arbuckle, 2013, 2152344}. Six of these studies had good sensitivity {Chu, 2020,
6315711; Hjermitslev, 2020, 5880849; Lauritzen, 2017, 3981410; Manzano-Salgado, 2017,
4238465; Meng, 2018, 4829851; Wikstrom, 2020, 6311677}, while five were considered
adequate {Arbuckle, 2013, 2152344; Chang, 2022, 9959688; Govarts, 2018, 4567442; Souza,
2020, 6833697; Xu, 2019, 5381338).
Four {Lauritzen, 2017, 3981410; Wikstrom, 2020, 6311677; Souza, 2020, 6833697; Xu, 2019,
5381338} of the seven SGA studies reporting main effects showed some adverse associations,
while three studies were null {Chang, 2022, 9959688; Govarts, 2018, 4567442; Manzano-
Salgado, 2017, 4238465}. The magnitude of odds ratios (ORs) across the four studies showing
adverse associations in the overall population (OR range: 1.19 to 4.14) was variable whether the
effect estimates were based on either categorical or continuous exposures (per each unit increase)
(Figure 3-58) with the two low confidence studies showing the largest risks. For example, Xu et
al. (2019, 5381338) reported an OR of 4.14 (95% CI: 1.07, 16.0) for each loglO unit increase in
PFOS. Souza et al (2020, 6833697) reported an OR of 3.67 (1.38-9.74) in quartile 4 relative to
quartile 1. The high confidence Lauritzen et al. (2017, 3981410) study did not show an increased
risk in the overall population per each ln-unit PFOS increase, but they did show a larger
association among participants from Sweden (OR = 2.51; 95% CI: 0.93, 6.77). The high
confidence study by Wikstrom et al. (2020, 6311677) reported an OR of 1.56 (95%CI: 1.09; 2.22
per each ln-unit increase) with a larger OR in girls (OR = 2.05; 95% CI: 1.00, 4.21) than boys
(OR = 1.30; 95%) CI: 0.70, 2.40). Similarly, a slight increased risk in their overall population
(OR= 1.19; 95%) CI: 0.87, 1.64) expressed per each ln-unit change was largely driven by results
in girls (OR = 1.40; 95% CI: 0.83, 2.35).
Overall, four (2 high and 2 low confidence studies) reported increased risks for SGA with
increasing PFOS exposures. The magnitude in risk across many of these studies were relatively
large, but neither of two studies examining categorical exposures showed any evidence of an
exposure-response relationship. Although the number of studies was small, few patterns were
discernible across study characteristics or overall confidence for these SGA findings.
3-206
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Sampling Exposure Study
Period Reference Matrix Design Exposure Levels Sub-population Comparison EE
Effect Estimate
01234567
OR (per doubl ing in
Early Manzand- Plasma. Cohort Mean (SD): 6.05 ng/mL (2.74 Boys : , 1.01
' „ , ..... ,. \ matemel plasma PFOS)
pregnancy Salgado et Maternal ng/mL)
I
—J
al., 2017 Blood OR (per doubling in
maternal plasma PFOS)
OR {per doubling in ^
maternal plasma PFOS)
OR (per 1-ln ng/mL change
Wikstrom Maternal Cohort Median=5.33 ng/mL Boys . 1.08
in PFOS)
etal.,2020 Serum (25th-75th percentiles:
3.97-7.60 „9/mL) OR (for Q2 vs Ql} 1.26
OR (for Q3 vs Ql} 0.86
OR(forQ4vs Ql) 1.3
OR (per 1-ln ng/mL change
Ir 5 in PFOS)
OR (for Q2 vs Ql) 0.89
OR (for Q3 vs Ql) 0.82
OR (for Q4 vs Ql) 2.05
OR (per Hn ng/mL change
in PFOS) 1,19
OR(fbrQ2vsQl) 0.69
OR (for Q3 vs Ql) 0.79
OR (for Q4 vs Ql) 1.56
Median=9.74 ng/mL (range: .. OR (per In unit increase in _
Uttr Lairitzol M.ttrml CotiDrt 0 95_59 6 „^mL) Norway 0.71
pregnancy etal., 2017 Serum
Median=i64 ng/mL (range: OR (per in unit increase in
2.28-55.2 ng/mL) Sweden pp0S) 251
Norway: median=9.74 ng/mL OR (per In unit increase in
(range: 0.95-59.6 ng/mL); S.. PFOS)
01 23456 7
Figure 3-55. Odds of Small-for-gestational-age in Children from High Confidence
Epidemiology Studies Following Exposure to PFOS
Interactive figure and additional study details available on Tableau.
Small-for-gestational-age defined as birthweight below the 10th percentile for the reference population.
3-207
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Sampling Exposure Study
Period Reference Matrix Design Exposure Levels Sub-population Comparison EE
Effect Estimate
0.0 0.5 1.0 1.5 2.0 2.5
Early Chang et Maternal Cohort Median: 2.19 ng/mL Term Births OR (per doubling in 1
pregnancy al., 2022 Serum (25th-75th percentile:
1.45-3.24 ng/mL)
1
1
1
1
1
1 •
1
1
1
1
1
OR [for Q2
(1.44-2.19 ng/mL)
vs. Q1 (
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
overall population with imprecise increased risks shown for PFOS exposure quartile 3 (OR =
1.41; 95% CI: 0.23, 8.82) and quartile 4 (OR = 3.70; 95% CI: 0.61, 22.6) compared to the
quartile one referent.
Confidence Sampling Measured Exposure Study
Rating Period Reference Effect/Endpoints Matrix Design Sub-population Comparison EE
Effect Estimate it
0 2 4 6 8 10
^ .. . ., , _. . ^ OR (per doubling in maternal ,
High Early Manzano- Low Birth Weight Plasma, Cohort - . 1.06
.. . plasma PFOS)
confidence pregnancy Salgadoetal. Maternal
1
1
, 2017 Blood 777
OR (per doubling in maternal
Boys plasma PFOS) "
1
1
1
1
, OR (per doubling in maternal _
Girls , ficncl 0.73
plasma PFOS)
1
1
~*T
l
.. _ , OR (per doubling in maternal
Term Low Birth P asma, Cohort - . v nc«c\ 0-91
,«. • *.4. .... . plasma PFOS)
Weight Maternal '
1
I
OR (per doubling in maternal
Boys . v M 1.68
plasma PFOS)
1
I
1 •
I
. OR (per doubling in maternal
plasma PFOS)
1
Later Chuetal., Low Birth Weight Maternal Cohort - O R( per llnng/mL increase in ^
pregnancy 2020 Serum
1
I
1 •
I
OR for Q2 (> 4.36 to 7.15 ng/mL
PFOS) vs. 01 (<=4.36 ng/mL PFOS)
1
.1
•l
1
OR for Q3 (> 7.15 to 11.93 ng/mL
PFOS) vs. 01 (<=4.36 ng/mL PFOS)
1
1 _
1 *
I
OR for Q4 (> 11.93 ng/mL PFOS) vs.
01 (<=4.36 ng/mL PFOS)
1
1
1
1
Medium Early et a^^O]^' Low Birth Weight Cohort — OR (per lln-ng/mL change in PFOS) 1.03
confidence pregnancy
1
1
Mengetal., Low Birth Weight Maternal Cohort - OR (per doubling of PFOS) 1.3
2018 Serum
1
OR (for Q2 vs. Ql) 1.4
1
I
I •
1
OR (for Q3 vs. 01) 1.8
l
1
l
1
OR (for Q4 vs. Ql) 1.2
l
1 _
l
I
0 2 4 6 8 10
Figure 3-57. Odds of Low Birthweight in Children from Epidemiology Studies Following
Exposure to PFOS
Interactive figure and additional study details available on Tableau.
Low birthweight defined as birthweight < 2,500 g.
Collectively, the majority (7 of 10) of SGA and LBW studies were supportive of an increased
risk with increasing PFOS exposures. The increased odds ranged from 1.19 to 4.14 although
evidence of exposure-response relationships was lacking. There was no evidence of differences
by study confidence as five of these seven were either high (n=4) or medium (n=l) confidence.
There was also no evidence of sample timing differences as the majority of studies with
associations were reported in studies based on early sampling periods.
3-209
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Arbuckle et al„ 2013, 2152344
Chang et al., 2022, 9959688
Chu et a!., 2020, 6315711
Govarts et al., 2018, 4567442
Gundacker et al., 2021, 10176483
Hjermitslev et al,, 2020, 5880849
Lauritzen et al., 2017, 3981410
Manzano-Salgado et al,, 2017, 4238465
Meng et al., 2018, 4829851
Souza et al., 2020, 6833697
Wikstrom et al., 2020, 6311677
Xu et al., 2019, 5381338
+
+
+
-
+
+
+
~
+
+
+
+
+* 1
+
++
+
++
++
+
+
++
B
+
+
J
+
+
--
-
+
-
~
+
+
++
B
+
-
++
B
++ ++
+
+
++
+
++
++
++ ++
+
++
+
++
++
D
D
++
+
++
+
++
B
+
+
+
-
-
+
+
++ ++
++
+
++
+
++
++
++
++
+
+
+
+
~
Legend
Q
Good (metric) or High confidence (overall)
+
Adequate (metric) or Medium confidence (overall)
-
Deficient (metric) or Low confidence (overall)
b
Critically deficient (metric) or Uninformative (overall)
*
Multiple judgments exist
Figure 3-58. Summary of Study Evaluation for Epidemiology Studies of PFOS and Low
Birth Weight or Small for Gestational Age Effects
Interactive figure and additional study details available on HAWC.
3.4.4.1.4.3 Birth Length
Thirty-one birth length studies were considered as part of the study evaluation as shown in
Figure 3-59 and Figure 3-60. Four studies were considered iminformative {Alkhalawi, 2016,
3859818; Gundacker, 2021, 10176483; Jin, 2020, 6315720; Lee, 2013, 3859850} and four more
studies noted above {Bach, 2016, 3981534; Kishi, 2015, 2850268, Kobayashi, 2017, 3981430;
3-210
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Kobayashi, 2022, 10176408} were not further considered for multiple publications from the
same cohort studies. Twenty-three non-overlapping and informative studies examined birth
length in relation to PFOS with five of these examining standardized birth length measures only
{Chen, 2017, 3981292; Espindola-Santos, 2021, 8442216; Gyllenhammar, 2018, 4238300;
Shoaff, 2018, 4619944; Xiao, 2019, 5918609}, and one evaluating both measures {Workman,
2019, 5387046}. Twelve studies examined sex-specific data with two studies {Marks, 2019,
5081319; Robledo, 2015, 2851197} reporting only sex-specific results. Eighteen studies
examined mean birth length differences in the overall study population.
Seven of these 23 included studies were high confidence {Bell, 2018, 5041287; Bjerregaard-
Olesen, 2019, 5083648; Lauritzen, 2017, 3981410; Manzano-Salgado, 2017, 4238465; Shoaff,
2018, 4619944; Valvi, 2017, 3983872; Xiao, 2019, 5918609}, eight were medium confidence
{Chen, 2017, 3981292; Chen, 2021, 7263985; Gyllenhammar, 2018, 4238300; Hjermitslev,
2020, 5880849; Kashino, 2020, 6311632; Luo, 2021, 9959610; Robledo, 2015, 2851197; Wang,
2019, 5080598} and eight were low confidence studies {Callan, 2016, 3858524; Cao, 2018,
5080197; Espindola-Santos, 2021, 8442216; Gao, 2018, 5387135; Marks, 2019, 5081319; Shi,
2017, 3827535; Workman, 2019, 5387046; Xu, 2019, 5381338}. Twelve PFOS studies had good
study sensitivity {Bjerregaard-Olesen, 2019, 5083648; Chen, 2017, 3981292; Chen, 2021,
7263985; Gyllenhammar, 2018, 4238300; Hjermitslev, 2020, 5880849; Kashino, 2020, 6311632;
Lauritzen, 2017, 3981410; Manzano-Salgado, 2017, 4238465; Robledo, 2015, 2851197; Shoaff,
2018, 4619944; Valvi, 2017, 3983872; Xiao, 2019, 5918609}, while eight had adequate
sensitivity {Callan, 2016, 3858524; Cao, 2018, 5080197; Gao, 2018, 5387135; Luo, 2021,
9959610; Marks, 2019, 5081319; Shi, 2017, 3827535; Workman, 2019, 5387046; Xu, 2019,
5381338} and three {Bell, 2018, 5041287; Espindola-Santos, 2021, 8442216; Wang, 2019,
5080598} were considered deficient.
3-211
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
,ce
Alkhalawi et al„ 2016, 3859818
Bach etal., 2016, 3981534-
Bell etal., 2018, 5041287
Bjerregaard-Olesen et al., 2019, 5083648
Callan et al., 2016, 3858524 - + +
Cao et al., 2018, 5080197
Chen etal., 2017, 3981292-
Chen etal., 2021, 7263985
Espindola Santos et al., 2021, 8442216 -
Gao et al., 2019, 5387135
Gundackeretal., 2021, 10176483
Gyllenhammar et al., 2018, 4238300
Hjermitslev et al., 2020, 5880849
Jin etal., 2020, 6316202
Kashino et al., 2020, 6311632
Klshi etal., 2015, 2850268-
B Legend
Good (metric) or High confidence (overall)
+ Adequate (metric) or Medium confidence (overall)
- Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-59. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Length Effects3
Interactive figure and additional study details available on HAWC.
•* Includes three overlapping studies: Bjerregaard-Olsen et al. (2019, 5083648); Kishi et al. (2015, 2850268); Kobayashi et al.
(2017,3981430).
3-212
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
.fiCk
I I I I I I I II
9^^
Kobayashi et al„ 2017, 3981430-
Kobayashi etal., 2022, 10176408-
Lauritzen et al., 2017, 3981410-
Leeetal., 2013, 3859850-
Luoetal., 2021, 9959610-
Manzano-Salgado et al., 2017, 4238465-
Marks et al., 2019, 5081319-
Robledo et al., 2015, 2851197 -
Shi et al., 2017, 3827535-
Shoaff etal., 2018, 4619944
Valvi et al., 2017, 3983872
Wang etal., 2019, 5080598
Workman et al., 2019, 5387046
Xiao etal., 2020, 5918609
Xu et al., 2019, 5381338
++
a
B
++
++
++
++
a
-
+
B
++
++
++
B
++
++
+
-
++
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-60. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Length Effects (Continued)a
Interactive figure and additional study details available on HAWC.
aIncludes three overlapping studies: Bjerregaard-Olsen et al. (2019, 5083648); Kishi et al. (2015,2850268); Kobayashi et al.
(2017,3981430).
3-213
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Of the 23 studies examining either standardized birth length or mean birth length measures,
seven studies showed some adverse associations based on the overall population. This included
three of the six {Chen, 2017, 3981292; Espindola-Santos, 2021, 8442216; Gyllenhammar, 2018,
4238300; Shoaff, 2018, 461994; Workman, 2019, 5387046; Xiao, 2019, 5918609} studies that
reported standardized birth length data. The high confidence study by Xiao et al. (2019,
5918609) reported reduced birth length z-scores (-0.33; 95% CI: -0.69, 0.03) in the overall
population, as well as for both male (-0.41; 95% CI: -0.87, 0.05) and female neonates (-0.23;
95%) CI: -0.75, 0.30) per each log2 increase in PFOS. Although smaller in magnitude, the
medium confidence study by Chen et al. (2017, 3981292) also reported a birth length deficit of
-0.16 (95%o CI: -0.31, -0.02) in the overall population as well as male (-0.15; 95%> CI: -0.33,
0.03) and female neonates (-0.20; 95%> CI: -0.44, 0.05 per each In unit PFOS increase). The
other high confidence study by Shoaff et al. (2018, 4619944) of standardized birth length
measures showed a deficit only for tertile 3 (-0.24; 95%> CI: -0.64, 0.15) compared to tertile 1.
Four {Callan, 2017, 3858524; Chen, 2021, 7263985: Lauritzen, 2017, 3981410; Workman et al.,
2019, 5387046} of the sixteen studies examining mean birth length in the overall population in
relation to PFOS showed some evidence of reductions. The high confidence study by Lauritzen
et al. (2017, 3981410) showed a small deficit in the overall population (-0.3 cm; 95%> CI: -0.7,
0.1), but detected the strongest association when restricted to the Swedish population (-1.2 cm;
95% CI: -2.1, -0.3). The medium confidence study by Chen et al. (2021, 7263985) reported
birth length deficits in the overall population (-0.27 cm; 95%> CI: -0.51, -0.02), males (-0.14
cm; 95% CI: -0.55, 0.26), and females (-0.40 cm; 95%> CI: -0.74, -0.06) per each PFOS ln-unit
increase. The low confidence study by Workman et al. (2019, 5387046) reported a non-
statistically significant birth length reduction of-0.16 cm (95%> CI: -0.92, 0.60) per each ln-unit
PFOS increase. The low confidence study by Callan et al. (2017, 3858524) reported a slightly
larger birth length reduction of-0.22 cm (95%> CI: -1.0, 0.57) per each ln-unit PFOS increase.
Five different sex-specific studies reported some birth length deficits in either or both male (4 of
11) and female (2 of 10) neonates including the Chen et al. (2021, 7263985) results noted above.
Among the two sex-specific only studies {Robledo, 2015, 2851197; Marks, 2019, 5081319}, the
Marks et al. {2019, 5081319} low confidence study of boys only showed adverse associations
(-0.52 cm; 95%> CI: -1.05, 0.01 for tertile 3 vs. tertile 1). The high confidence study by Valvi et
al. (2017, 3983872) reported no associations in the overall population but did detect a non-
significant birth length deficit in male neonates (-0.18 cm; 95%> CI: -0.60, 0.23 per each PFOS
log2 exposure increase). The low confidence study Wang et al. (2019, 5080598) study also
reported a non-significant birth length deficit in males that was similar in magnitude (-0.17 cm;
95%o CI: -0.71, 0.37). Although it was not statistically significant, the high confidence study by
Bjerregaard-Olesen et al. (2019, 5083648) detected a difference in mean birth length among girls
only (-0.3 cm; 95%>CI: -0.7, 0.0 per each IQR PFOS increase). One study not reporting sex-
specific differences did report that there were no statistically significant interactions by sex for
their birth length and PFOS measures {Gyllenhammar, 2018, 4238300}.
In summary, of the 23 birth length studies, 11 different ones showed some adverse associations
either in the overall population, or in either or both sexes. Two of 10 studies in females and four
of 11 studies in males reported some birth length deficits. Although there were more studies in
males that reported decreased birth length, there was little consistency across sex or even
compared to the overall population. None of the five studies examining categorical data in either
3-214
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
sex or the overall population showed any evidence of an adverse exposure-response relationship.
Few patterns were evident across study characteristics or confidence levels, although the
database may be prone to bias due to pregnancy hemodynamics as eight of the studies that
showed associations relied on later biomarker samples.
3.4.4.1.4.4 Head Circumference at Birth
Nineteen informative studies that examined head circumference were considered in the synthesis.
Seven studies were rated as medium {Chen, 2021, 7263985; Gyllenhammar, 2018, 4238300;
Hjermitslev, 2020, 5880849; Kashino, 2020, 6311632; Lind, 2017, 3858512; Robledo, 2015,
2851197; Wang, 2019, 5080598} confidence, while six were high confidence {Bell, 2018,
5041287; Bjerregaard-Olesen, 2019, 5083648; Lauritzen, 2017, 3981410; Manzano-Salgado,
2017, 4238465; Valvi, 2017, 3983872; Xiao, 2019, 5918609} and six were low {Callan, 2016,
3858524; Cao, 2018, 5080197; Espindola-Santos, 2021, 8442216; Marks, 2019, 5081319;
Workman, 2019, 5387046; Xu, 2019, 5381338}. Three studies were deficient in study sensitivity
{Bell, 2018, 5041287; Espindola-Santos, 2021, 8442216; Wang, 2019, 5080598}, while eleven
had good {Bjerregaard-Olesen, 2019, 5083648; Chen, 2021, 7263985; Gyllenhammar, 2018,
4238300; Hjermitslev, 2020, 5880849; Kashino, 2020, 6311632; Lauritzen, 2017, 3981410;
Lind, 2017, 3858512; Manzano-Salgado, 2017, 4238465; Robledo, 2015, 2851197; Valvi, 2017,
3983872; Xiao, 2019, 5918609} and five had adequate study sensitivity {Callan, 2016, 3858524;
Cao, 2018, 5080197; Marks, 2019, 5081319; Workman, 2019, 5387046; Xu, 2019, 5381338}.
3-215
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
i©
Bell etal., 2018, 5041287
Bjerregaard-Olesen et al., 2019, 5083648-
Calian etal., 2016, 3858524-
Cao etal., 2018, 5080197-
Chen et al., 2021, 7263985 -
Espindola Santos et al., 2021, 8442216 -
Gundacker et al., 2021, 10176483-
Gyllenhammar et al., 2018, 4238300 -
Hjermitslev et al., 2020, 5880849 -
Kashino et al., 2020, 6311632 -
Lauritzen etal., 2017, 3981410-j
Lind etal., 2017, 3858512 J
Manzano-Salgado et al., 2017, 4238465 -
Marks etal., 2019, 5081319-
Robledo et al., 2015, 2851197 -
Valvi etal., 2017, 3983872^
Wang etal., 2019, 5080598-
Workman et al., 2019, 5387046 -
Xiao et al., 2020, 5918609 -j
Xu etal., 2019, 5381338-
b : i: :rn
&
Legend
| Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-61. Summary of Study Evaluation for Epidemiology Studies of PFOS and Head
Circumference Effects
Interactive figure and additional study details available on HAWC.
3-216
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Sixteen of the 19 included studies examined PFOS in relation to mean head circumference
differences including 13 studies with results in the overall population and 11 different studies
with sex-specific data. Three of the mean head circumference studies {Lind, 2017, 3858512;
Marks, 2019, 5081319; Robledo, 2015, 2851197} only reported sex-specific data, including the
low confidence study by Marks et al. (2019, 5081319) which only examined male neonates. The
three remaining studies {Espindola-Santos, 2021, 8442216; Gyllenhammar, 2018, 4238300;
Xiao, 2019, 5918609} examined unitless standardized measures.
Five of the 16 studies with data based on the overall population reported some associations
between PFOS and different head circumference measures. This included one study based on
standardized head circumference and four studies examining mean head circumference. The high
confidence study by Xiao et al. (2019, 5918609) showed consistent head circumference z-score
deficits across their overall population (-0.26; 95% CI: -0.68, 0.16), as well as male (-0.15;
95% CI: -0.68, 0.39) and female neonates (-0.42; 95% CI: -1.05, 0.21) per each log2 increase
in PFOS. Although the high confidence study by Lauritzen et al. (2017, 3981410) reported a null
association in the combined Norwegian and Swedish population, they did detect a large head
circumference reduction amongst their Swedish population only (-0.4 cm; 95% CI: -0.9, 0.04)
per each ln-unit PFOS change.
Only three of the 14 studies examining mean head circumference differences in the overall
population reported any evidence of associations with none of these reaching statistical
significance. The high confidence study by Bach et al. (2016, 3981534) showed a small, non-
significant head circumference differences (-0.1 cm; 95% CI: -0.2, 0.1 per each PFOS IQR
increase). In their low confidence study, Cao et al. (2018, 5080197) reported a non-significant
inverse association in the overall population (-0.23 cm; 95% CI: -1.19, 0.73 per each ln-unit
PFOS) as did the low confidence study by Callan et al. (2016, 3858524) (-0.39 cm; 95% CI:
-0.98, 0.20 per each In unit PFOS).
Two of ten studies examining female neonates and four of 11 examining male neonates reported
some inverse associations between increasing PFOS and mean head circumference. One study
not reporting sex-specific differences did report that there were no statistically significant
interactions by sex for their head circumference and PFOS measures {Gyllenhammar, 2018,
4238300}. The head circumference reductions were consistently around -0.3 cm in males in
three (one each low, medium, and high confidence) of four studies. The medium confidence study
by Lind et al. (2017, 3858512) reported deficits across all quartiles (range: -0.3 to -0.4 cm) but
only in males. The high confidence study by Valvi et al. (2017, 3983872) also reported deficits
only in male neonates (-0.28 cm; 95%CI: -0.65, 0.09 per each doubling of serum PFOS
exposures), while head circumference increases were found for female neonates (0.48 cm;
95%CI: 0.05, 0.90). The low confidence study of boys only by Marks et al. (2019, 5081319)
reported monotonic deficits across PFOS tertiles 2 (-0.13 cm; 95% CI: -0.45, 0.19) and 3 (-0.31
cm; 95%) CI: -0.62, 0.01) compared to tertile 1. The medium confidence study by Kashino et al.
(2020, 6311632) reported smaller deficits only in male neonates (-0.14 cm; 95% CI: -0.61, 0.32
per each loglO PFOS). Although it was not statistically significant, the high confidence study by
Bjerregaard-Olesen et al. (2019, 5083648) detected a small difference in mean head
circumference among girls only (-0.1 cm; 95% CI: -0.3, 0.1 per each IQR PFOS increase). The
low confidence study by Cao et al. (2018, 5080197) found a large head circumference difference
3-217
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
(-1.22 cm; 95% CI: -2.70, 0.25 for tertile 3 vs. 1) among females with some evidence of an
exposure-response relationship.
Although there were nine different studies that showed some evidence of associations between
PFOS and head circumference in the overall population or different subsets by countries or sex,
there was limited epidemiological evidence of associations among the overall population with
only four of 13 studies showing any inverse associations. Mean sex-specific head circumference
deficits were detected in six different studies including four in male neonates and two others in
females only. An additional study with standardized head circumference measures showed
deficits in both sexes, but larger deficits were noted among females. One of two studies in each
sex showed some evidence of an exposure-response relationship. A very large association was
seen in one low confidence study among females, but more consistent results were seen across
four studies in males (two high, one medium and one low confidence). Although limited numbers
across different study characteristic or overall confidence level sub-groups precluded a detailed
assessment, few patterns were evident across the ten different studies that showed some adverse
associations with head circumference. Only two {Bjerregaard-Olesen, 2019, 5083648; Lind,
2017, 3858512} of these nine studies had any early pregnancy (i.e., trimester 1) samples, with
seven studies {Callan, 2016, 3858524; Cao, 2018, 5080197; Kashino, 2020, 6311632; Lauritzen,
2017, 3981410; Marks, 2019, 5081319; Valvi, 2017, 3983872; Xiao, 2019, 5918609} based on
either second and/or third trimester maternal samples or later. Overall, nine of 19 studies
showing some evidence of adverse associations with some uncertainty as to what degree these
results may be influenced by pregnancy hemodynamics due to later sample timing. There was
considerable heterogeneity of results within and across both sexes and different studies.
3.4.4.1.4.5 Fetal Growth Restriction Summary
The majority of studies examining fetal growth restriction showed some evidence of associations
with PFOS exposures especially those that included BWT data (i.e., SGA, low BWT, as well as
mean and standardized BWT measures). The evidence for two fetal growth measures such as
head circumference and birth length were less consistent. For many of these endpoints, there was
a preponderance of associations amongst studies with later biomarker samples that may be more
prone to potential biases from pregnancy hemodynamic impacts. There was limited evidence of
exposure-response relationships in either analyses specific to the overall population or different
sexes, although the categorical data generally supported the linearly expressed associations that
were detected.
Among the most accurate fetal growth restriction endpoints examined here, there was generally
consistent evidence for BWT deficits across different measures and types of PFOS exposure
metrics considered. BWT deficits were detected in the roughly two-thirds of included studies
whether measured as mean BWT or standardized z-scores. This included 19 out of 30 mean
BWT studies in the overall population and 16 of 27 medium or high confidence studies. Most of
the sex-specific mean BWT studies showed some adverse associations in either male or female
neonates, and although it was not consistent across studies, more deficits were found in male
neonates. As noted above, many of the individual study results lacked precision and were not
statistically significant especially the sex-stratified results which may have been largely
underpowered to detect sex-specific differences.
3-218
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
The magnitude of some fetal growth measures were at times considered large especially when
considering the per unit PFOS increases across the exposure distributions. Although some of the
other endpoints were fairly small in magnitude, the birth weight deficits and odds ratios for
birthweight-related measures were more sizeable especially when considering most were
expressed on a per-unit increase basis. For example, for all but one of the 19 studies showing
mean BWT deficits in the overall population, reported deficits ranging from -14 to -93 grams
per each PFOS unit increase. Associations were also seen for the majority of studies examining
small for gestational age and low birth weight measures.
The current database (since the 2016 HESD) is fairly robust given the wealth of studies included
here, with most studies considered high or medium confidence (e.g., 23 out of 30 mean BWT)
and most having adequate or good study sensitivity. As noted earlier, one source of uncertainty is
that previous meta-analyses of PFOS by Dzierlenga et al. (2020, 7643488) and PFOA by
Steenland et al. (2018, 5079861) have shown that some measures like mean BWT may be prone
to bias from pregnancy hemodynamics especially in studies with sampling later in pregnancy.
Although a limited number of studies across some strata does not fully lend itself to
differentiating patterns across different study characteristics, like study confidence and sample
timing, some patterns emerged across the study results. For many of these endpoints, there was a
preponderance of associations, such as birth weight measures, amongst studies with later
biomarker samples (i.e., either exclusive trimester 2 maternal sample or later, such as umbilical
cord or post-partum maternal samples) that may be more prone to pregnancy hemodynamic
impacts. This would seem to comport with the PFOS meta-analysis by Dzierlenga et al. (2020
7643488) that suggested that results for mean BWT may be impacted by some bias due to
pregnancy hemodynamics. Therefore, despite some consistency in evidence across these fetal
growth endpoints, some important uncertainties remain mainly around the degree that some of
the results examined here may be influenced by sample timing.
3.4.4.1.5Postnatal growth
Eleven studies examined PFOS exposure in relation to postnatal growth measures (Figure 3-62).
The synthesis here is focused on postnatal growth measures including mean and standardized
weight {Cao, 2018, 5080197; Chen, 2017, 3981292; de Cock, 2014, 2713590; Gyllenhammar,
2018, 4238300; Lee, 2018, 4238394; Manzano-Salgado, 2017, 4238509; Shoaff, 2018, 4619944;
Starling, 2019, 5412449; Yeung, 2019, 5080619} and height {Cao, 2018, 5080197; Chen, 2017,
3981292; de Cock, 2014, 2713590; Gyllenhammar, 2018, 4238300; Lee, 2018, 4238394; Shoaff,
2018, 4619944; Yeung, 2019, 5080619}, as well as body mass index (BMI)/adiposity measures
{Chen, 2017, 3981292; de Cock, 2014, 2713590; Gross, 2020, 7014743; Jensen, 2020, 6833719;
Shoaff, 2018, 4619944; Starling, 2019, 5412449; Yeung, 2019, 5080619} and estimates of rapid
growth during infancy {Manzano-Salgado, 2017, 4238509; Shoaff, 2018, 4619944; Starling,
2019, 5412449; Yeung, 2019, 5080619}.
Four postnatal growth studies were high confidence {Jensen, 2020, 6833719; Shoaff, 2018,
4619944; Starling, 2019, 5412449; Yeung, 2019, 5080619}, four were medium confidence
{Chen, 2017, 3981292; de Cock, 2014, 2713590; Gyllenhammar, 2018, 4238300; Manzano-
Salgado, 2017, 4238509}, and three were low confidence {Cao, 2018, 5080197; Gross, 2020,
7014743; Lee, 2018, 4238394}. As shown in Figure 3-62 seven postnatal growth studies had
good study sensitivity {Chen, 2017, 3981292; Gyllenhammar, 2018, 4238300; Jensen, 2020,
6833719; Lee, 2018, 4238394; Manzano-Salgado, 2017, 4238509; Shoaff, 2018, 4619944;
3-219
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Starling, 2019, 5412449}, two each were adequate {Cao, 2018, 5080197; Yeung, 2019,
5080619} or deficient {de Cock, 2014, 2713590; Gross, 2020, 7014743}. The medium
confidence study by de Cock et al. (2014, 2713590) did not report effect estimates but indicated
that there were no statistically significant associations between PFOS quartiles and infant BMI
(p-value=0.59), infant weight (p-value=0.80), and infant height (p-value=0.98) measures up to
11 months of age. But their lack of reporting of effect estimates precluded consideration of
magnitude and direction of any associations and are not further examined below in the
summaries.
The medium confidence study by Manzano-Salgado et al. (2017, 4238509) reported null
associations for their overall population, female, and male neonates for weight gain z-score
measured at 6 months per each log2 PFOS increase. The low confidence study by Lee et al.
(2018, 4238394) reported statistically significant inverse associations per each PFOS In unit
increase for height at age 2 years (-0.77 cm; 95% CI: -1.27, -0.15) as well as height change
from birth to 2 years (-0.71 cm; 95% CI: -1.27, -0.15). Small differences were seen for mean
weight differences at age 2 years (-0.17 cm; 95% CI: -0.38, 0.04) but not for weight change
from birth to 2 years. Although no exposure-response relationships were detected when
examined across PFOS categories those with the highest exposure saw smaller statistically
significant height increases at age 2 compared to lower exposures. Although a statistically
significant birth length association was detected, the medium confidence study by Chen et al.
(2017, 3981292) reported no association with infant height z-score up to 24 months. They did
report statistically significant lower infant weight z-scores among female neonates comparable in
magnitude for 6 to 12 months (-0.25; 95% CI: -0.47, -0.04) or 12 to 24 months (-0.25; 95% CI:
-0.41, -0.06) per each In unit PFOS increase. Females seemed to drive the deficit detected in the
overall population (-0.13; 95% CI: -0.32, 0.07 per each In unit PFOS increase) for the 6-to-12-
month window. The medium confidence study by Gyllenhammar et al. (2018, 4238300) did not
detect standardized BWT deficits per each IQR PFOS change, but they showed slight weight
deficits (—0.2) at 3 months that persisted throughout 60 months of age. In contrast, standardized
birth length measures were null for increasing PFOS exposures regardless of the time windows
examined. Compared to the tertile 1 referent, the low confidence study of infants followed up to
a median age of 19.7 months by Cao et al. (2018, 5080197) reported slight increases in postnatal
length (i.e., height) (1.37 cm; 95% CI: -0.5, 3.28), while large postnatal weight deficits were
reported for PFOS tertiles 2 (-138 g; 95% CI: -574, 298) and 3 (-78 g; 95%CI: -532, 375).
Associations at five months of age in the overall population (-0.28; 95% CI: -0.51, -0.05) and
females (-0.56; 95% CI: -0.87, -0.26) from the high confidence study by Starling et al. (2019,
5412449) were detected for weight-for-age z-scores, as well as weight-for-length z-scores
(overall: -0.26; 95% CI: -0.53, 0.00; females; -0.52; 95% CI: -0.88, -0.17). Exposure-response
relationships were observed across tertiles for both of these measures. In their high confidence
study of repeated measures at 4 weeks, 1 year and 2 years of age, Shoaff et al. (2018, 4619944)
detected statistically significant deficits and exposure-response relationships for infant weight-
for-age z-score (-0.33; 95% CI: -0.65, -0.01) and weight-for-length z-score (-0.34; 95% CI:
-0.59, -0.08) in PFOS tertile 3 compared to tertile 1. Small deficits that were not statistically
significant were observed in tertile 3 for length for age z-score (-0.22; 95% CI: -0.49, 0.04). In
their high confidence study, Yeung et al. (2019, 5080619) reported statistically significant
negative growth trajectories weight-for-length z-scores in relation to each log SD increase in
PFOS exposures among singletons followed for three years. No associations were detected for
3-220
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
infant length (i.e., height) measures. Some sex-specific results were detected with larger
associations seen in singleton females for weight for length z-score (-0.10; 95%CI: -0.16, -0.05)
and weight z-score (-0.07; 95%CI: -0.13, -0.01). An infant weight deficit of-22.0 g (95% CI:
-59.5, 15.6 per each 1 log SD PFOS increase) was also observed that was driven by results in
females (-51.6 g; 95% CI: -102.3, -0.8).
Overall, seven of 8 studies with quantitative estimates (including 5 high and medium confidence
studies) showed some associations between PFOS exposures and different measures of infant
weight. Two of four studies with categorical data showed some evidence of inverse monotonic
exposure-response relationships. Two of six studies with quantitative estimates examining
different infant height measures showed some evidence of adverse associations with PFOS.
Study quality ratings, including study sensitivity and overall confidence, did not appear to be
explanatory factors for heterogeneous results across studies.
3-221
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
5\e°
,Ge
,„G^® .a-j^S-^g0T>
Cao et al., 2018, 5080197-
Chen et al., 2017, 3981292-
Gross et al., 2020, 7014743-
Gyllenhammar et al., 2018, 4238300 -
Jensen et al., 2020, 6833719 -
Lee et al., 2018, 4238394
Manzano-Salgado et al., 2017, 4238509
Shoaffet al., 2018, 4619944
Starling et al., 2019, 5412449
Yeung et al., 2019, 5080619
de Cock et al., 2014, 2713590
II 1 1 1 1 1 1
-
+
++
-
+
+
+
+
+
++
+
++
+
++
+
+
+
+
+
-
+
-
-
+
+
++
+
++
+
++
B
+
+
+
+
+
+
++
++
B
++
BB
+
++
B
¦
++
++
B
++
+
++
B
++
++
++
++
++
+
++
++
++
++
++
++
+
++
++
++
B
++
B
++
+
B
++
a
B
++
+
-
J
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
| Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-62. Summary of Study Evaluation for Epidemiology Studies of PFOS and
Postnatal Growth Effects
Interactive figure and additional study details available on HAWC.
3.4.4.1.5.1 Adiposity/BMI
In their high confidence study of repeated measures at 4 weeks, 1 year and 2 years of age, Shoaff
et al. (2018, 4619944) detected statistically significant decreases in infant BMI z-score (~0.36;
95% CI: -0.60, -0.12). Although they were not statistically significant, the medium confidence
Chen et al. (2017, 3981292) reported consistently small BMI z-scores across infant
developmental windows (range: -0.08 to -0.10) per each In unit PFOS. These results seem to be
3-222
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
driven by results in females especially for the 6 to 12 months (-0.33; 95% CI: -0.59, -0.08) and
12 to 24 months (-0.25; 95% CI: -0.45, -0.05) developmental periods. In their high confidence
study, Yeung et al. (2019, 5080619) reported statistically significant negative growth trajectories
for BMI and BMI z-score in relation to each log SD increase in PFOS exposures among
singletons followed for three years. No exposure-response relationship was detected for BMI z-
scores. Some sex-specific results were detected with larger associations seen in singleton females
BMI z-score (-0.11; 95% CI: -0.17, -0.05) and BMI (-0.16 kg/m2; 95% CI: -0.24, -0.08). In
the high confidence study by Starling et al. (2019, 5412449), decreased adiposity (-2.08; 95%
CI: -3.81, -0.35) among girls were detected in PFOS tertile 3 compared to the tertile 1 referent.
The high confidence study by Jensen et al. (2020, 6833719) reported null associations between
adiposity and per each 1-unit increase in PFOS measured at 3 and 18 months. The low
confidence study by Gross et al. (2020, 7014743) reported an inverse association (OR = 0.43;
95%) CI: 0.17 to 1.09) of being overweight at 18 months for PFOS levels greater than the mean
level. They also reported a lower odds ratio of being overweight at 18 months in males (OR =
0.19; p-value=0.04) than females (OR = 0.85; p-value=0.85). Mixed results were seen for
measures of adiposity and increased BMI with increasing PFOS exposures.
3.4.4.1.5.2 Rapid Weight Gain
Four high confidence studies {Manzano-Salgado, 2017, 4238509; Shoaff, 2018, 4619944;
Starling et al. 2019, 5412449; Yeung, 2019, 5080619} examined rapid infant growth. Limited
evidence of associations was reported, as only one {Starling et al., 2019, 5412449} of four
studies {Manzano-Salgado, 2017, 4238509; Shoaff, 2018, 4619944; Starling et al. 2019,
5412449; Yeung, 2019, 5080619} showed increased odds or rapid weight gain with increasing
PFOS. For example, Starling et al. (2019, 5412449) reported a small OR of 1.36 for rapid growth
in the overall population based on either weight for length-based z-scores. Study sensitivity was
not an explanatory factor for the null studies.
3.4.4.1.5.3 Postnatal Growth Summary
Seven (3 high, 2 medium, and 2 low confidence) of the 8 studies with quantitative estimates
examining different infant weight measures showed some evidence of adverse associations with
PFOS exposures either in the overall population or either/or both male or female neonates. There
was some evidence of exposure-response relationships as two of the four studies on infant weight
showed adverse monotonic relationships across PFOS categories. No patterns by study
characteristics or study confidence were evident. Only two (one low and one high confidence) of
the seven studies with quantitative estimates examining different infant height measures showed
some evidence of adverse associations with PFOS exposures. Two of the six postnatal growth
studies with quantitative estimates showed increased infant BMI or adiposity with increasing
PFOS exposures, while three showed decreased risk of higher BMI or adiposity. Only one out of
four high confidence studies showed any evidence of rapid growth among infants following
PFOS exposures. Although the data for some endpoints was less consistent, the majority of
infant weight studies indicated that PFOS may be associated with post-natal growth measures up
to two years of age.
3.4.4.1.6Gestational Duration
Twenty-two different studies examined gestational duration measures (i.e., PTB or gestational
age measures) in relation to PFOS exposures. Nine of these studies examined both PTB and
3-223
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
gestational age measures, while two studies only examined PTB {Liu, 2020, 6833609; Gardener,
2021, 7021199}.
3.4.4.1.6.1 Gestational Age
Seventeen of the 20 studies reporting gestational age estimates in relation to PFOS exposures
were considered informative and included here including two that were uninformative
{Gundacker, 2021, 10176483; Lee, 2013, 3859850} and one excluded study based on an
overlapping cohort {Li, 2017, 3981358}. Sixteen non-overlapping and informative studies
examined mean gestational age (in weeks) in relation to PFOS exposures including one study
reporting sex-specific results only {Lind, 2017, 3858512}.
Among the 16 different studies included here, nine were high confidence {Bach, 2016, 3981534;
Bell, 2018, 5041287; Chu, 2020, 6315711; Eick, 2020, 7102797; Huo, 2020, 6835452;
Lauritzen, 2017, 3981410; Lind, 2017, 3858512; Manzano-Salgado, 2017, 4238465; Sagiv et al.
2018, 4238410}, four were medium {Gyllenhammar, 2018, 4238300; Hjermitslev, 2020,
5880849; Meng, 2018, 4829851; Yang, 2022, 10176806} and four were low confidence
{Bangma, 2020, 6833725; Gao, 2019, 5387135; Workman, 2019, 5387046; Xu, 2019,
5381338}. Ten of these studies had good study sensitivity, six were adequate {Bangma, 2020,
6833725; Eick, 2020, 7102797; Gao, 2019, 5387135; Workman, 2019, 5387046; Xu, 2019,
5381338; Yang, 2022, 10176806} and one was deficient {Bell et al., 2018, 5041287}.
Nine of the 15 studies examining mean gestational age change in the overall population reported
some deficits. Among these, four were high confidence, and three were medium and two were
low confidence. The medium confidence study by Gyllenhammar et al. (2018, 4238300) reported
a deficit of-0.29 weeks (95% CI: -0.59, 0.01) per each IQR PFOS change; they also reported
that there were no statistically significant interactions by sex for their PFOS measures. The high
confidence study by Sagiv et al. (2018, 4238410) reported a similar gestational age reduction in
the overall population (-0.36 weeks; 95% CI: -0.64, -0.09) for PFOS quartile 4 versus quartile
1; this seemed to be driven by associations among boys only (z-score: -0.19; 95% CI: -0.33,
-0.05) per each IQR increase). The high confidence study by Chu et al. (2020, 6315711)
reported similar deficits in the overall population (-0.32 weeks; 95% CI: -0.53, -0.11) which
was driven by female neonates (-0.61 weeks; 95% CI: -0.90, -0.32). The high confidence study
by Lauritzen et al. (2017, 3981410) only showed deficits among their Swedish population (-0.4
weeks; 95%CI: -0.9, 0.2). Compared to tertile 1, the low confidence study by Gao et al. (2019,
5387135) reported deficits in tertile 2 (-0.40 weeks; 95% CI: -0.92, 0.12) and tertile 3 (-0.20;
95%CI: -0.61, 0.20). The high confidence study by Manzano-Salgado et al. (2017, 4238465)
reported deficits in quartile 4 among the overall population (-0.31 weeks; 95% CI: -0.55, -0.06)
compared to quartile 1. Despite low overall PFOS concentrations, the medium confidence study
by Yang et al. {2022, 10176806} showed reduced gestational age only among pre-term births for
both total PFOS (-1.26 weeks; 95%CI: - 2.46, -0.05) and linear PFOS (-1.80 weeks; 95%CI:
-3.24, -0.37) per each IQR increase, with results larger results in female (-1.06 weeks; 95%CI:
-2.87, 0.74) than male neonates (-0.41 weeks; 95%CI: -2.20, 1.37). The medium confidence
study by Meng et al. (2018, 4829851) reported statistically significant gestational age deficits
(range: -0.16 to -0.29 weeks) across all quartiles but no evidence of an exposure-response
relationship. The low confidence study by Workman et al. (2019, 5387046) reported a non-
significant decrease (-0.17 weeks; 95% CI: -0.52, 0.18) per each ln-unitPFOS change.
3-224
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Overall, nine of the 15 studies based on the overall population showed some evidence of inverse
associations between PFOS and gestational age. This included seven medium or high confidence
studies. The four high confidence studies showed deficits in the overall population consistent in
magnitude (range: -0.30 to -0.40 weeks). Apart from one study with very large deficits, the
remaining two medium and two low confidence studies all ranged from -0.17-0.30 weeks for
different PFOS contrasts). No exposure-response relationships were detected in any study, and
no definitive patterns were seen based on other study characteristics or in the other few studies
with sex-specific data. For example, 3 of 7 studies showed decreased gestational ages in relation
to PFOS exposures among both male or female neonates. Study sensitivity did not seem to be an
explanatory factor as five of six studies that did not show adverse associations had good or
adequate study sensitivity. Lastly, sample timing did not seem to be an explanatory factor of the
results as an equal proportion (60%) of studies showing inverse associations between PFOS and
gestational age deficits were based on earlier and later biomarker sampling.
3.4.4.1.6.2 Preterm Birth
As shown in Figure 3-63, eleven studies examined the relationship between PFOS and preterm
birth (PTB); all of the studies were either medium {Hjermitslev, 2020, 5880849; Liu, 2020,
6833609; Meng, 2018, 4829851; Yang 2022, 10176806} or high confidence {Bach, 2016,
3981534; Chu, 2020, 6315711; Eick, 2020, 7102797; Gardener, 2021, 7021199; Huo, 2020,
6835452; Manzano-Salgado, 2017, 4238465; Sagiv, 2018, 4238410}. Nine of the eleven studies
were prospective birth cohort studies, while the two studies by Liu et al. (2020, 6833609) and
Yang et al. (2022, 10176806) were case-control studies nested with prospective birth cohorts.
Four studies had maternal exposure measures that were sampled during trimester one {Bach,
2016, 3981534; Manzano-Salgado, 2017, 4238465; Sagiv, 2018, 4238410}, or trimester three
{Gardener, 2021, 7021199}. The high confidence study by Chu et al. {2020, 6315711} sampled
during the late third trimester or within three days of delivery. Four studies collected samples
across multiple trimesters {Eick, 2020, 7102797; Hjermitslev, 2020, 5880849; Huo, 2020,
6835452; Liu, 2020, 6833609}. One study used umbilical cord serum samples {Yang 2022,
10176806}. The medium confidence study by Meng et al. (2018, 4829851) pooled umbilical cord
blood and maternal serum (trimester 1 and 2) exposure data from two study populations. Seven
studies had good study sensitivity, while four others were considered adequate {Eick, 2020,
7102797; Liu, 2020, 6833609; Gardener, 2021, 7021199; Yang 2022, 10176806} with the
median exposure values in the overall population ranging from 1.79 ng/mL {Liu et al. 2020,
6833609} to 30.1 ng/mL {Meng, 2018, 4829851}. Lower levels were also seen for a total PFOS
measure in Yang et al. {2022, 10176806} for both cases (median (IQR) = 0.27 (0.30) ng/mL)
and controls (0.21 (0.37) ng/mL).
3-225
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
vcs^ «
se^ ^ cP^
* r**'
Bach etal., 2016, 3981534-
I
+
i
+
++
—1—
+
++
—1—
+
++
++
Chu etal., 2020, 6315711 -
B
++
B
+
++
+
++
++
Eick et al., 2020, 7102797-
++
++
B
+
++
+
B
++
Gardener et al., 2021, 7021199 -
++
++
++
+
++
+
B
++
Hjermitslev et al., 2020, 5880849 -
+
+
+
+
+
-
++
B
Huo et al., 2020, 6835452 -
++
++
++
+
++
+
++
++
Liu et al., 2020, 6833609 -
++
B
11
-
++
+
+
*
Manzano-Salgado et al., 2017, 4238465 -
++
++
++
+
++
+
++
++
Meng etal., 2018, 4829851 -
B
B
++
+
++
+
++
B
Sagiv et al., 2018, 4238410-
++
+
+
+
++
+
++
++
Yang et al., 2022, 10176806 -
+
+
+
+
+
+
+
~
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-63. Summary of Study Evaluation for Epidemiology Studies of PFOS and Preterm
Birth Effects
Interactive figure and additional study details available on HAWC.
Adverse associations were reported in seven of the 11 PTB studies with ORs from 1.5- to 5-fold
higher for elevated PFOS exposures. The medium confidence study by Meng et al. (2018,
4829851) study reported statistically significant non-monotonic increased ORs for PTB in the
upper three PFOS quartiles (OR range: 1.9-3.3), as well as per each doubling of PFOS exposures
(OR = 1.5; 95% CI: 1.1, 2.2). The high confidence study by Chu et al. (2020, 6315711) reported
some statistically significant increased ORs per each In unit increase (OR = 2.03; 95% CI: 1.24,
3-226
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.32) as well as an exposure-response relationship across upper three quartiles (OR range: 2.22-
4.99) exposures when compared to the referent. The high confidence study by Eick et al. (2020,
7102797) reported an exposure-response relationship as well (tertile 2 OR = 1.21; 95% CI: 0.50,
2.91; tertile3 OR = 1.87; 95% CI: 0.72, 4.88, compared to tertile 1). Although they were not
statistically significant, the medium confidence study by Liu et al. (2020, 6833609) reported
increased ORs of similar magnitude per each logi0unit increase (OR = 1.30; 95% CI: 0.76, 2.21)
or when quartile 3 (OR= 1.51; 95% CI: 0.85, 2.69) and quartile 4 (OR= 1.35; 95% CI: 0.74,
2.45) exposures were compared to the referent. The high confidence study by Sagiv et al. (2018,
4238410) study reported consistently elevated non-monotonic ORs for PTB in the upper three
PFOS quartiles (OR range: 2.0-2.4), but smaller ORs when examined per each IQR PFOS
increase (OR =1.1; 95% CI: 1.0, 1.3). The high confidence study by Gardener et al. (2021,
7021199) reported that participants in the PFOS exposure quartiles 2 (OR = 1.94; 95% CI: 0.66,
5.68) and 4 (OR = 1.41; 95% CI: 0.46, 4.33) had higher odds of preterm birth (relative to the
lowest quartile). Despite low overall PFOS concentrations, the medium confidence study by
Yang et al. (2022, 10176806) showed statistically significant increased odds of preterm birth per
each IQR increase in total PFOS (OR = 1.44; 95% CI: 1.18, 1.79), linear PFOS (OR = 1.41; 95%
CI: 1.19, 1.73), and branched PFOS (OR= 1.11; 95% CI: 1.01, 1.29). No differences were
observed for male or female stratified results (OR range: 1.40-1.45). Null or inverse associations
were reported by Bach et al. (2016, 3981534), Huo et al. (2020, 6835452), Manzano-Salgado et
al. (2017, 4238465) and Hjermitslev et al. (2019, 5880849). Overall, only two {Chu, 2020,
6315711; Eick, 2020, 7102797} out of eight studies showed evidence of exposure-response
relationships.
Overall, 7 of 11 studies reported increased odds of preterm birth in relation to PFOS with some
sizeable relative risks reported. There was some limited evidence of exposure-response
relationships as well. Although small numbers limited the confidence in many of the sub-strata
comparisons, few patterns in the PTB results emerged based on study confidence (all 11 studies
were medium or high confidence), sample timing or other study characteristics. For example,
three of the four null studies were considered to have good sensitivity to detect associations that
may be present. The results for preterm birth are robust with respect to adverse associations
detected with increasing PFOS exposures.
Few patterns in the PTB results emerged based on study confidence or other study
characteristics. Since nearly all studies had good study sensitivity, study sensitivity did not
largely appear to be a concern in this database. In addition, only one out of the four studies that
did not show adverse associations had limited exposure contrasts.
3.4.4.1.6.3Gestational Duration Summary
Overall, there is robust evidence of an impact of PFOS exposure on gestational duration
measures (i.e., either preterm birth or gestational age measures) as most of studies showed some
adverse associations. This was strengthened by consistency in the reported magnitude of
gestational age deficits despite different exposure levels and metrics examined. Although they
were not as consistent in magnitude (60% of the PTB studies showed some adverse
associations), some of the effect estimates were large for preterm birth in relation to PFOS
exposures with limited evidence of exposure-response relationships. Few patterns were evident
as explanatory factors for heterogeneous results based on our qualitative analysis.
3-227
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.4.1.7 Fetal Loss
As shown in Figure 3-64, five (2 high, 2 medium and 1 low confidence) studies examined PFOS
exposure and fetal loss. All of these studies had good study sensitivity owing largely to very
large sample size and sufficient sample sizes {Buck Louis, 2016, 3858527; Jensen, 2015,
2850253; Liew, 2020, 6387285; Wang, 2021, 10176703; Wikstrom, 2021, 7413606}.
The high confidence study by Wikstrom et al. (2021, 7413606) showed little evidence of
association between PFOS and miscarriages (OR = 1.13; 95% CI: 0.82, 1.52 per doubling of
PFOS exposures). The authors did not report an exposure-response relationship across PFOS
quartiles but did show elevated non-significant ORs of approximately 1.2 and 1.3 for the upper
two quartiles. Although the ORs were not statistically significant in the medium confidence study
by Liew et al. (2020, 6387285), there was some suggestion of an exposure-response for
miscarriages across PFOS quartiles (OR range: 1.1-1.4). Similarly, the low confidence study by
Jensen et al. (2015, 2850253) reported increased non-significant risks across tertiles 2 and 3 (OR
range: 1.15-1.33). No association was detected in the high confidence study by Wang et al.
(2021, 10176703) (OR = 0.95; 95%CI: 0.87, 1.04) or the medium confidence study by Buck
Louis et al. (2016, 3858527) (hazard ratio (HR) = 0.81; 95% CI: 0.65, 1.00 per each SD PFOS
increase).
Overall, there was positive evidence for fetal loss with increased relative risk estimates in three
out of five studies. In those three studies, the magnitude of associations detected were low but
consistently reported in the range of 1.1 of 1.4 with an exposure-response relationship detected in
one study. No patterns in the results were detected by study confidence ratings including
sensitivity.
3-228
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
i aq awn ¦
**» ^ ^
&
I I I
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Buck Louis et al., 2016, 3858527 -
+
+
+
+
+
+
++
B
Jensen et al., 2015, 2850253 -
+
+
+
+
+
++
Liew et al., 2020, 6387285-
B
++
+
++
+
++
B
Wang et al., 2021, 10176703-
++ ++
++
+
++
+
++
++
Wikstrom et al., 2021, 7413606-
++ ++
++
+
++
+
++
++
Figure 3-64. Summary of Study Evaluation for Epidemiology Studies of PFOS and Fetal
Loss Effects
Interactive figure and additional study details available on HAWC.
3.4.4.1.8 Birth Defects
As shown in Figure 3-65, five (3 medium and 2 low confidence) studies examined PFOS
exposure in relation to birth defects. Four of the five studies had adequate sensitivity. This
included a medium confidence study by Ou et al. (2021, 7493134) that reported increased risks
for septal defects (OR = 1.92; 95% CI: 0.80, 4.60), conotruncal defects (OR = 1.65; 95% CI:
0.59, 4.63) and total congenital heart defects (OR=1.61; 95%CI: 0.91, 2.84) among participants
with maternal serum levels over >75th PFOS percentile (relative to those <75th percentile). A
low confidence study of a non-specific grouping of all birth defects {Cao, 2018, 5080197}
reported a small but imprecise increased risk (OR = 1.27; 95% CI: 0.59, 2.73). Interpretation of
all birth defect groupings is challenging given that etiological heterogeneity may occur across
individual defects.
Three studies examined PFOS exposures in relation to cryptorchidism. The medium confidence
study by Vesterholm Jensen et al. (2014, 2850926) detected an inverse association for
cryptorchidism (OR = 0.51; 95% CI: 0.21-1.20) per each ln-unit increase in PFOS exposures.
This risk seemed to be largely driven by boys from Finland. The medium confidence study by
3-229
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Toft et al. (2016, 3102984) reported null associations per each ln-unit increase in PFOS
exposures and both cryptorchidism (OR = 0.99; 95% CI: 0.75, 1.30) and hypospadias (OR =
0.87; 95% CI: 0.57, 1.34). The low confidence study by Anand-Ivell et al. (2018, 4728675) did
not find statistically significant PFOS exposure differences among cryptorchidism or hypospadia
cases compared to controls, but they did not examine this in a multivariate fashion adjusting for
confounders.
Overall, there was very limited evidence of associations between PFOS and birth defects based
on the available epidemiological studies. This was based on cryptorchidism, hypospadias or all
birth defect groupings. As noted previously, there is considerable uncertainty in interpreting
results for broad any defect groupings which are anticipated to have decreased sensitivity to
detect associations.
,0®
Anand-Ivell et al., 2018, 4728675-
I
+
I
+
I
+
I
I
I
+
i
Cao et al., 2018, 5080197-
-
+
-
-
+
+
+
-
Ou et al., 2021, 7493134-
++
+
++
+
+
+
+
+
Toftet al., 2016, 3102984-
+
+
++
+
++
+
+
+
Vesterholm Jensen et al., 2014, 2850926 -
+
+
++
+
++
+
+
+
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-65. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Defect Effects
Interactive figure and additional study details available on HAWC.
3.4.4.2 Animal Evidence Study Quality Evaluation and Synthesis
There are 4 studies from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} and 15 studies from
recent systematic literature search and review efforts conducted after publication of the 2016
PFOS HESD that investigated the association between PFOS and developmental effects. Study
quality evaluations for these 19 studies are shown in Figure 3-66.
3-230
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Argus, 2000, 5080012
Butenhoff et al., 2009, 757873
Chen et al., 2018, 5080460-
Conley et al., 2022, 10176381
Era et al., 2009, 2919358
Fuentes et al., 2006, 757859
Lai et al., 2017, 3981773
Lau et al., 2003, 757854
Lee etal., 2015, 2851075
Li etal., 2016, 3981495
Li etal., 2021, 9959491
Luebker et al., 2005, 1276160
Luebker et al., 2005, 757857 -
Mshaty et al., 2020, 6833692 -
Wan et al., 2020, 7174720-
Xia et al., 2011, 2919267
Zhang et al, 2021, 6988534
Zhang et al., 2020, 6315674 -
Zhong et al., 2016, 3748828 -
Legend
| Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
S Critically deficient (metric) or Uninformative (overall)
Not reported
~ Multiple judgments exist
~ Bias away from null
Figure 3-66. Summary of Study Evaluation for Toxicology Studies of PFOS and
Developmental Effects
Interactive figure and additional study details available on HAWC.
Evidence suggests that PFOS exposure can adversely affect development. Oral studies in mice,
rats, and rabbits report effects in offspring including decreased survival, decreased body weights,
structural abnormalities (e.g., reduced skeletal ossification), histopathological changes in the
lung, and delayed eye opening, among others. Effects in offspring primarily occurred at similar
doses as those seen in the maternal animals. Adverse effects observed in dams include alterations
in gestational weight and gestational weight gain, as well as evidence of altered placental
3-231
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
histology. In some cases, adverse developmental effects of PFOS exposure that relate to other
health outcomes may be discussed in the corresponding health outcome section (e.g., fetal and
neonatal pulmonary effects are discussed in the respiratory section found in the PFOS
Appendix).
3.4.4.2.1 Maternal Effects
Multiple developmental studies evaluated maternal weight outcomes in rats, mice, and rabbits
(Figure 3-67). Yahia et al. (2008, 2919381) observed a decrease in body weight in ICR mouse
dams administered 20 mg/kg/day PFOS from gestational day 1 to 17 (GD 1 to GD 17) or GD 18.
The dams exhibited no clinical signs of toxicity. Thibodeaux et al. (2003, 757855) observed
significantly decreased maternal body weight gain in CD-I mice at exposed to 20 mg/kg/day
PFOS (highest dose tested in the study); food and water consumption were not affected by
treatment. Lee et al. (2015, 2851075) also reported reduced maternal body weight gain in CD-I
mice treated with 2 or 8 mg/kg/day PFOS (not 0.5 mg/kg/day) compared to controls. Dams in the
2 and 8 mg/kg/day dose groups had significantly lower mean body weights on GD 14-GD 17. In
contrast, Lai et al. (2017, 3981773) did not observe a significant difference in maternal body
weight in CD-I mouse dams orally exposed to 0, 0.3, or 3 mg/kg/day throughout gestation (GD
1-GD 20). The authors determined that there were no observable maternal effects related to
PFOS exposure at the relatively low doses evaluated. Wan et al. (2020, 7174720) found no effect
of PFOS on maternal body weight in CD-I mouse dams orally dosed with 0, 1, or 3 mg/kg/day
from GD 4.5 to GD 17.5. Likewise, Fuentes et al. (2006, 757859) found no treatment-related
effects on maternal body weight, maternal body weight gain, or maternal food consumption in
CD-I mouse dams orally exposed to 0, 1.5, 3, or 6 mg/kg/day PFOS from GD 6 to GD 18.
Mshaty et al. (2020, 6833692) orally administered PFOS to C57BL/6J mice from postnatal day 1
(PND 1) to PND 14, resulting in lactational exposure to pups. Mean maternal body weights were
evaluated at PND 21 and determined to be comparable between the control and the 1 mg/kg/day
dose groups.
Thibodeaux et al. (2003, 757855) observed significant, dose-dependent decreases in maternal
body weight, food consumption, and water consumption in Sprague Dawley rats dosed with
> 2 mg/kg/day PFOS from GD 2 to GD 20. Xia et al. (2011, 2919267) also observed reduced
body weight on GD 21 in Sprague Dawley rats dosed with 2 mg/kg/day from GD 2 to GD 21. In
a 2-generation reproductive toxicity study in rats, Luebker et al. (2005, 1276160) similarly
observed dose-dependent decreases in maternal body weight in the 3.2 mg/kg/day dose group of
the parental generation (Po) from day 15 of the premating exposure through lactation day 1 (LD
1), the last recorded weight; this dose group also had significantly decreased maternal weight
gain from GD 0 to GD 20. The 1.6 mg/kg/day dams experienced transient decreases in maternal
weight compared to controls in the window between GD 3 and GD 11. There were no reported
differences in the maternal weight of adult first generation (Fi) females during precohabitation
until the end of lactation, though the highest dose tested in these females was only
0.4 mg/kg/day. Following the 2-generation study, Luebker et al. (2005, 757857) conducted a
follow up 1-generation study that examined additional PFOS doses during development.
Crl:Cd(Sd)Igs Vaf/Plus rat dams were gavaged with 0, 0.4, 0.8, 1, 1.2, 1.6, or 2 mg/kg/day
PFOS. Dosing started 6 weeks prior to mating and continued through mating and gestation with
the final dose on LD 4. The authors observed no treatment-related effects on body weight change
during gestation, but body weight gain was reduced in the 0.8, 1, 1.6, and 2 mg/kg/day groups
relative to controls during lactation. They also reported a general trend for reduced food
3-232
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
consumption with increasing dose during gestation and lactation {Luebker, 2005, 757857}. In
another study with Sprague Dawley rats dosed with 0, 5, or 20 mg/kg/day PFOS from GD 12 to
GD 18, Li et al. (2016, 3981495) also reported reduced mean maternal body weights in the
20 mg/kg/day dose group. In another study, Conley et al. (2022, 10176381) reported a significant
43% weight gain reduction relative to controls in Sprague-Dawley (Crl:CD(SD)) rat dams dosed
with 30 mg/kg/day PFOS from GD 14 to GD 18; no significant effects were observed for the 0.1,
0.3, 1, 3, or 10 mg/kg/day PFOS groups. Zhang et al. (2021, 6988534) also reported no
significant treatment-related effects on maternal body weight in Sprague-Dawley rat dams dosed
with 0, 1, or 5 mg/kg/day PFOS from GD 12 to GD 18. Butenhoff et al. (2009, 757873) observed
comparable maternal body weight and body weight gain during gestation in Sprague Dawley rat
dams dosed with 0, 0.1, 0.3, or 1 mg/kg/day PFOS from GD 0 to LD 20 but observed
significantly lower absolute body weights during lactation (PND 4-PND 20) in dams treated
with 1 mg/kg/day PFOS. Transient decreases in food consumption were observed in the 0.3 and
1.0 mg/kg/day groups throughout the study, though these findings were not considered
treatment-related or adverse.
In a single rabbit study, Argus Research Laboratories (2000, 5080012) reported significantly
decreased maternal body weight gain from GD 7 to GD 21 at PFOS doses > 1 mg/kg/day (mean
body weight change of 0.38, 0.3, 0.2, and -0.01 kg with 0, 1, 2.5, and 3.75 mg/kg/day PFOS,
respectively); no significant effect was observed from GD 21 to GD 29. There were observations
of scant or no feces for some does in the 1.0, 2.5, and 3.75 mg/kg/day groups. Observations of
scant feces were significant relative to control at 3.75 mg/kg/day. Significant reductions in
absolute (g/day) and relative (g/kg/day) feed consumption was also observed in the 2.5 and
3.75 mg/kg/day dose groups.
Endpoint
Maternal Body Weight
Study Name Study Design
Lee et al., 2015, 2851075 developmental (GD11-16)
Wan et al., 2020, 7174720 developmental (GD4.5-17.5)
Fuentes et al„ 2006, 757859 developmental (GD6-18)
Lai etal., 2017, 3981773 developmental (GD1-17)
Mshaty et al.. 2020, 6833692 developmental (LD1-14)
Li et al., 2016, 3981495 developmental (GD12-18)
Butenhoff et al., 2009, 757873 developmental (GD0-PND20)
Luebker et al., 2005, 1276160 reproductive (42d prior mating-LD20) LD1
reproductive (GD0-PND112)
Maternal Body Weight Change Conley et al.. 2022.10176381 developmental (GD14-18)
Argus, 2000, 5080012 developmental (GD7-20)
Butenhoff ot al., 2009, 757873 developmental (GD0-PND20)
Luebker et al.. 2005,757857 reproductive (42d prior mating-LD4)
Luebker et al.. 2005,1276160 reproductive (42d prior mating-LD20)
Observation Time Animal Description
GD17 P0 Mouse, CD-1 ( -, N=10)
GD17.5 P0 Mouse. CD-1 N=8)
GD18 Mouse, CD-1 (i, N=10-11)
GD1-20 P0 Mouse, CD-1 (". N=18)
PND21 P0 Mouse. C57BL/6J (v. N=0-15)
GD18 P0 Rat, Sprague-Dawley (*. N=10)
GD20 PO Rat, Crl:CD(SD) (' , N=23-25)
PND1 P0 Rat, Crl:CD(SD) (,;, N=23-25)
PND21 P0 Rat, Crl:CD
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.4.2.2 Viability
Decreases in both fetal and pup survival and viability with perinatal PFOS exposure were
observed in multiple studies (Figure 3-68). Lee et al. (2015, 2851075) reported a significantly
higher incidence of resorptions, post-implantation loss, and dead fetuses at GD 17 after dosing
pregnant CD-I mice by gavage with 0.5, 2, or 8 mg/kg/day from GD 11 to GD 16; however,
there was no significant difference in the mean number of implantations. A significant decrease
in mean number of live fetuses was also observed in the 2.0 and 8.0 mg/kg/day dose groups vs.
controls. A decrease in the mean number of live fetuses was also reported in the 0.5 mg/kg/day
dose group but this difference was not significant relative to control. Administration of 0, 1, 5,
10, 15, or 20 mg/kg/day PFOS to CD-I mice from GD 1 to GD 17 did not affect the number of
implantation sites but resulted in a significant increase in post-implantation loss, as measured by
decrease in mean percentage of live fetuses, in dams administered 20 mg/kg/day {Thibodeaux,
2003, 757855}. In another study, CD-I mouse dams were dosed with 0, 3, or 6 mg/kg/day PFOS
from GD 6 to GD 18. The authors found no treatment-related effects on the number of litters
with dead fetuses, the total number of dead fetuses, dead fetuses per litter, or live fetuses per
litter, and there were no effects of PFOS on the number of implantation sites, the percentage of
post-implantation loss, the number of early or late resorptions, or fetal sex ratio {Fuentes, 2006,
757859}.
Mice appear to be more sensitive to alterations in fetal viability than rats. Thibodeaux et al.
(2003, 757855) dosed pregnant Sprague-Dawley rats with 0, 1, 2, 3, 5, or 10 mg/kg PFOS daily
by gavage from GD 2 to GD 20. The number of implantations was not affected by treatment and
there were no treatment-related effects observed on the live rat fetuses at term. Likewise, Zhang
et al. {2021, 6988534} dosed Sprague-Dawley rat dams with 0, 1, or 5 mg/kg/day PFOS from
GD 12 to GD 18 and found no treatment-related effects on liveborn pups per litter, pup survival,
or pup sex ratio. Butenhoff et al. (2009, 757873) also observed no treatment-related effects on
the number of implantation sites or resorptions in pregnant Sprague-Dawley rats exposed to 0.1,
0.3, or 1.0 mg/kg/day by gavage from GD 0 to PND 20. Similarly, Conley et al. (2022,
10176381) found no effects of PFOS on the number of live fetuses per litter or total resorptions
in a study wherein Sprague-Dawley (Crl:CD(SD)) rat dams were dosed with 0, 0.1, 0.3, 1, 3, 10,
or 30 mg/kg/day PFOS from GD 14 to GD 18.
In pregnant New Zealand white rabbits cesarean sectioned on GD 29 after gestational exposure
to PFOS, Argus Research Laboratories (2000, 5080012) reported no significant effects on
implantations or resorptions. However, Argus Research Laboratories (2000, 5080012) did report
abortions among New Zealand white rabbits orally dosed with 2.5 mg/kg/day (1/17 does, 5.9%)
or 3.75 mg/kg/day (9/21 does, 42.8%) from GD 7 to GD 20. The abortion rate was significantly
greater relative to control for the 3.75 mg/kg/day dose group. Argus Research Laboratories
(2000, 5080012) reported no significant effects on the mean number of live fetuses/doe, number
of dead fetuses/doe, mean litter size, and offspring viability.
Altered pup viability was observed in studies of both rats and mice. In one- and two-generation
reproductive toxicity studies in Sprague Dawley rats, Luebker et al. (2005, 757857; 2005,
1276160) observed reduced pup viability index (ratio of the number of pups alive at PND 5 to
the number of live pups born) at higher maternal PFOS doses. A significant decrease in pup
viability for the one-generation study was associated with a dose of 1.6 mg/kg/day {Luebker,
2005, 757857}; the number of dams with all pups dying between PND 1 and PND 5 was also
3-234
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
significantly increased in the 2 mg/kg/day dose group. The dose associated with a decreased
viability index in Fi pups was also 1.6 mg/kg/day in the two-generation study {Luebker, 2005,
1276160}; between PND 1 and PND 4, 100% of dams had all pups dying in the 3.2 mg/kg/day
dose group. Following gestational exposure to PFOS on GD 19 GD 20, Grasty et al. (2003,
5085464) observed survival of 98%, 66%, and 3% of rat pups in the control, 25, and
50 mg/kg/day groups, respectively, on PND 5. Similarly, Xia et al. (2011, 2919267) found
decreased number of delivered pups per litter and increased pup mortality between birth and
PND 3 for rats treated with 2 mg/kg/day on GD 2 to GD 21. Chen et al. (2012, 1276152) also
observed decreased pup survival through PND 3 in rat pups exposed to 2 mg/kg/day PFOS from
GD 1 to GD 21. Thibodeaux et al. (2003, 757855) and Lau et al. (2003, 757854) similarly
observed decreased pup survival in rats exposed to > 2.0 mg/kg/day PFOS from GD 2 to GD 21.
Lau et al. (2003, 757854) also reported PFOS-related effects on survival in mice following
gestational exposure to PFOS. Briefly, most mouse pups from dams administered 15 or
20 mg/kg/day did not survive for 24 hours after birth. Fifty percent mortality was observed at
10 mg/kg/day. Survival of pups in the 1 and 5 mg/kg/day treated dams was similar to controls.
Yahia et al. (2008, 2919381) also observed significant effects on pup survival. In this study,
pregnant ICR mice/group were administered 0, 1, 10, or 20 mg/kg of PFOS daily by gavage from
GD 1 to GD 17 or GD 18. All neonates in the 20 mg/kg/day dose group were born pale, weak,
and inactive, and all died within a few hours of birth. At 10 mg/kg/day, 45% of those born died
within 24 hours. Survival of the 1 mg/kg/day group was similar to that of controls. Of the
developmental studies identified in the most recent literature search, only Mshaty et al. (2020,
6833692) evaluated the impact of lactational (PND 1-PND 14) PFOS exposure on pup survival.
Mshaty et al. (2020, 6833692) observed no difference in C57BL/6J mouse pup survival through
PND 21 between control group pups and pups exposed to 1 mg/kg/day PFOS (quantitative data
not provided).
PFOS Dcvolopmontal Effects - Mortality
Eridpolnt
Study Name
Study Design
Observation Time
Animal Description
0 No significant changed Significan
increase ~ Significant decrease
Abortions
Argus. 2000. 5080012
developmental (GD7-20)
P0 Rabbit. New Zealand (i.. N=17-21)
Dams with Still burr. Pups
Fue nles el al., 2006, 757859
Luobkor et al.. 2005.757857
developmental (GD6-18)
reproductive (42d prior mating-LD4)
Mouse, CD-1 , N=10-11)
P0 Rat, Crl:Cd(Sd)lgs Vaf.'Plus (", N=17)
A
w
Luebker et al.. 2005,1276160
reproductive (42C prior mating-LD20)
PND1
P0 Rut. CrhCd (SdJIgs Br Va« N=20-25)
Fetuses, Dead
Argus. 2000. 5080012
developmental (GD7-20)
P0 Rabbit. Now Zealand ( ,. N=12-20)
Lee etal.. 2015.2851075
developmental (GD11-16)
GD17
P0 Mouse. CD-1 ('-. N=10)
A
A A
Fuentes et al., 2006, 757859
developmental (GD6-18)
GD18
Mouse. CD-1!,, N=10-11)
Luebker etal. 2005.757857
reproductive (76c (42d pre-cohabitation. I4d mating. GD0-20))
GD21
PO Rat. Crl:Cd(Sd)lgs Vaf.'Plus ('_'. N=8)
FbIusbs, Dead pat Liltar
Fuantes el al„ 2006, 757859
developmental (GD6-18)
GDIS
Mousa,CD-I (,, N*10-11)
Fetuses, Live
Argus. 2000. 5080012
developmental (GD7-20)
GD29
P0 Rabbit, Now Zealand (,, N=l2-20>
Fuentes etal., 2006,757859
developmental (GD8-18)
GDIS
Mouse, CD-1 C.N=10-11)
Fetuses, Live (No. per Live Litter)
Conley et al.. 2022, 10176381
developmental (GD14-18)
GD18
P0 Rat, Sptague-Dawley (•••, N=4-6)
Implantation
Argus. 2000,5080012
developmental (GD7-20)
GD29
P0 Rabbit, New Zealand (+, N=12-20)
Luebkei et al., 2005,757857
lepicxiuctive (42d prior malinq-LD4)
LD5
P0 Rat, Crl:Cd(Sd)lgs Vaf.'Plus (- N=17)
Implantation Sites. Per Delivered Litter
Fuentes et al.. 2006.757859
developmental (GD6-18)
GD18
Mouse. CD-1 (+. N=10-11)
Liva Pups Born
Luebker el si., 2005, 757857
reproductive (42ri prior maling-LD4)
PND0
PO Ral. Crl:Cd(Sd)lgs Vaf,'Plus ('"', N=17)
Luobkor ot al., 2005,1276160
reproductive (42d prior mating-LD20)
PND1
F1 Rat, Crt:Cd (Sd)lgs Br Vaf (• . N-20-2S)
• W
Livebom Pups, Mean/Litter
Zhang eta I. 2021, 6988534
developmental (GD12-18)
PND1
P0 Rat. Sprague-Dawley (_. N=8)
No. Dams with All Pups Dying, PND 1-
Luebkei et al., 2005,1276160
lepicductive (42d prioi mating-LD20)
P0 Rat CrlrCd (Sdjlgs Bi Vaf (',, N=20-25)
No. Dams with All Pups Dying. PND 1-
Morlalily
Luebker et al,. 2005,757857
Xia el al., 2011,29197(57
reproductive (420 prior mating-LD4)
developmental (GD2-21)
LD1-5
PND3
P0 Rat. Crl:Cd(Sd)lgs Vaf.'Plus ('J. N=17)
.A
0.01
0.1 1
10 10O
Concentration (mg.'kg/day)
Figure 3-68. Mortality and Viability in Mice, Rats, and Rabbits Following Exposure to
PFOS (logarithmic scale)
PFOS concentration is presented in logarithmic scale to optimize the spatial presentation of data.
Interactive figure and additional study details available on HAWC.
GD = gestation day; PND = postnatal day; LD = lactational day; Po = parental generation; Fi = first generation; d = day.
3-235
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Endpoint
Study Name
Study Design
Observation Time
PFOS Developmental Effects • Mortality
Animal Description
% No significant change A Significant increase ~ Significant decrease
Offspring Survival
Lau el al.. 2003, 737864
developmental (GD1-17)
PND0
F1 Mouse. CD-1 ( N=7)
PND6
F1 Mouse. CD-1 (¦>-. N=7)
« . vw
PND24
F1 Mouse, CD-1 (-;i. N=7)
4 VW
Zhang et al, 2021. 6988534
developmental (GD12-18)
PND14
F1 Rat. Sprague-Daivley . N=93-98)
ButenhofT et al.. 2009,757873
developmental (GD0-PND20)
PND0-4
F1 Rat. CrtCD(SD) >, N=23-2S)
PND4-21
F1 Rat, Crl:CD(SD) \ N=23-25)
Lau et al.. 2003, 757854
developmental (GD2-21)
PND0
F1 Rat. Sprague-Dawley N=9)
PND5
F1 Rat, Sprague-Dawley {/•', N=9)
~ . VVV V
PND22
F1 Rat, Sprague-Dawley , , N=9)
VVV V
Post-Implantation Loss
Leeetal.. 2015,2851075
developmental (GD11-16)
GD17
P0 Mouse. CD-1 (' . N=10)
« A A A
Fuentes etal.; 2006, 757859
developmental (GD6-18)
GDIS
Mouse, CD-1 { N=10-11)
Resorptions. Any
Argus, 2000, 5080012
developmental (GD7-20)
GD29
P0 Rabbit, New Zealand (¦. N=12-20)
Resorptions, Early
Argus, 2000.5080012
developmental (GD7-20)
GD29
P0 Rabbit, Now Zealand ( , N=12-20)
Fuentes et al., 2006, 757859
developmental (GD6-18)
GD18
Mouse, CD-1 { '. N=10-11)
Resorptions. Late
Argus. 2000, 5080012
developmental (GD7-2Q)
GD29
P0 Rabbit. New Zealand ( N=12-20)
Fuentes et al., 2006, 757859
developmental (GD6-18)
GD18
Mouse, CD-1 ('-\ N=10-11)
Resorptions, Mean.'Litter
Luebkeretal., 2005, 757857
reproductive (76d (42d pre-cohabitaiion, 14d mating, GDO-20))
GD21
P0 Rat, Crt:Cd(Sd)lgs Vaf,-Plus (V, N=8)
Resorptions. Parcent/Litte
Argus, 2000, 5080012
developmental (GD7-20)
GD20
P0 Rabbit. New Zealand (>¦, N=12-20)
Resorptions, Total
Conley et al.. 2022. 10176381
developmental (GD14-18)
GD18
P0 Rat, Sprague-Dawley (T, N=4-6)
Stillborn Pups
Luebker et aL 2005, 1276160
reproductive (42d prior mating-LD20)
PND1
F1 Rat. Cri:Cd (Sd)lgs Br Vaf (.; y, N=20-25)
« . . A
Total Litter Rasorbed
Argus, 2000,5080012
developmental (GD7-20)
GD2S
P0 Rabbit. New Zealand (-, , N=12-20)
Viability Index
Luabker at al.. 2005, 1276160
reproductive (42d prior mating-LD20)
PND1-4
F1 Rat. CrlrCd (Sd)lgs Br Vaf (.; V, N-156-346)
.01 0.1 1 10 100
Concentration (mg/kg/day)
Figure 3-69. Mortality and Viability in Mice, Rats, and Rabbits Following Exposure to
PFOS (Continued, logarithmic scale)
PFOS concentration is presented in logarithmic scale to optimize the spatial presentation of data.
Interactive figure and additional study details available on HAWC.
GD = gestation day; PND = postnatal day; LD = lactational day; Po = parental generation; Pi = first generation; d = day.
3.4.4.2.3 Skeletal, Soft Tissue, and Gross Effects
Skeletal defects in offspring, including bone ossification, are a known effect of gestati onal PFOS
exposure. In one study, 0, 1, 10, or 20 mg/kg of PFOS was administered daily by gavage to
pregnant ICR mice from GD 1 to GD 17 or GD 18 {Yahia, 2008, 2919381}. Five dams/group
were sacrificed on GD 18 for fetal external and skeletal effects. In the fetuses from dams treated
with 20 mg/kg/day, there were significant increases in the numbers of fetuses with cleft palates
(98.56%), sternal defects (100%), delayed ossification of phalanges (57.23%), wavy ribs
(84.09%), spina bifida occulta (100%), and curved fetus (68.47%). In mice, Thibodeaux et al.
(2003, 757855) observed significantly increased incidences of cleft palate at 15 and
20 mg/kg/day PFOS, sternal defects at 5, 10, 15, and 20 mg/kg/day PFOS, and ventricular septal
defects at 20 mg/kg/day PFOS. Thibodeaux et al. (2003, 757855) also observed significantly
increased incidences of these deformities in rats. The authors reported incidences of cleft palate
at 10 mg/kg/day PFOS and sternal defects at 2 and 10 mg/kg/day PFOS. In another study, CD-I
mouse dams were exposed to 0, 1.5, 3, or 6 mg/kg/day PFOS from GD 6 to GD 18 {Fuentes,
2006, 757859}. The authors reported a lower incidence of incomplete calcaneus ossification in
the 3 mg/kg/day group (6% fetal incidence, 20% litter incidence) relative to controls (46% fetal
incidence, 80% litter incidence). The same study observed no treatment-related effects on fetal or
litter incidence of the following skeletal development outcomes: supernumerary ribs, asymmetri c
sternebra, incomplete ossification of vertebra, or total skeletal malformations {Fuentes, 2006,
757859}.
Skeletal malformations in fetal and neonatal rabbits were reported in Argus Research
Laboratories (2000, 5080012) at comparatively lower PFOS doses than those described in rat and
mouse studies. A significant decrease in the mean number of isolated ossification sites of the
metacarpal per fetus per litter was observed in the 3.75 mg/kg/day dose group vs. control (4.82
3-236
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
vs. 4.98, respectively); no significant change in mean number of ossification sites per fetus per
litter was reported in the 0.1 (4.97), 1 (4.99), or 2.5 mg/kg/day (4.97) dose groups. A significant
decrease in the mean number of sternal center ossification sites per fetus per litter was observed
in the 2.5 and 3.75 mg/kg/day dose groups relative to control (3.81 and 3.82, respectively,
relative to 3.98 for the control group); no significant change in the mean number of sternal center
ossification sites per fetus per litter was detected in the 0.1 (3.92) and lmg/kg/day (3.95) dose
groups. A significant difference in fetal incidence of irregular ossification of the skull was
reported in both the 2.5 and 3.75 mg/kg/day dose groups relative to control (0.8% and 9.2%
incidence respectively, relative to 4% in the control); no significant difference was observed in
the 0.1 (5.6%) and 1 mg/kg/day (2%) dose groups. There were no significant differences in litter
incidence of irregular ossification of the skull in the 0.1, 1, 2.5, and 3.75 dose groups vs. control
(38.9%), 15.8%o, 6.2%o, and 25%, respectively, vs. 30%>). A significant decrease in mean number
of ossification sites in the hyoid body per fetus per litter was reported in the 3.75 mg/kg/day dose
group (0.92) vs. Control (1); no change in mean number of hyoid ossification sites was reported
in other dose groups (mean of 1 for the 0.1, 1, and 2.5 mg/kg/day dose groups). A significant
increase in fetal incidence of a hole in the parietal bone was observed in the 3.75 mg/kg/day dose
group vs. Control (6.5%> vs. 0%); no holes were detected in the 0.1, 1, and 2.5 mg/kg/day dose
groups. Litter incidence of a hole in the parietal was 1 (8.3%>) in the 3.75 mg/kg/day dose group
and 0 (0%>) in the 0, 0.1, 1, and 2.5 mg/kg/day dose groups. Fetal incidence of unossified pubis
was also significantly increased in the 3.75 mg/kg/day group vs. Control (3.7% vs. 0%>). No other
dose groups exhibited unossified pubis. A significant increase in litter incidence of unossified
pubis was observed in the 3.75 mg/kg/day group vs. Control (16.7%> vs. 0%>). The rest of the dose
groups exhibited 0% litter incidence of unossified pubis. However, fetal alterations were
observed in a similar percentage of litters across all dose groups (70%, 61.1%, 47.4%, 25%, and
66.7%) in the 0, 0.1, 1, 2.5, and 3.75 mg/kg/day dose groups, respectively). No significant
difference was seen in the mean percentage of fetuses per litter with any alteration (14.1%>, 17%,
9.5%), 3.6%), and 17.4% in the 0, 0.1, 1, 2.5, and 3.75 mg/kg/day dose groups, respectively).
3.4.4.2.4Fetal or Pup Body Weight
Several studies in different species reported data on fetal body weight (Figure 3-70). In a study in
CD-I mice with gestational PFOS exposure from GD 11 to GD 16, Lee et al. (2015, 2851075)
reported mean fetal body weights on GD 17 of 1.72, 1.54, 1.3, and 1.12 g in the 0, 0.5, 2, and
8 mg/kg/day dose groups, respectively. The mean fetal weights reported for the 2 and
8 mg/kg/day groups were significantly lower than those reported for the control dose group. In
another study with CD-I mice that were exposed to 0, 1, or 3 mg/kg/day PFOS from GD 4.5 to
GD 17.5, Wan et al. (2020, 7174720) reported a significant reduction in fetal body weight in the
3 mg/kg/day group compared to controls. In contrast, Fuentes et al. (2006, 757859) found no
treatment-related effects on mean fetal weight per litter on GD 18 in CD-I mice exposed to 0,
1.5, 3, or 6 mg/kg/day PFOS from GD 6 to GD 18. Li et al. (2021, 9959491) observed a dose-
dependent decrease in fetal body weight in mice (strain not specified) exposed to 0, 0.5, 2.5, or
12.5 mg/kg/day PFOS from GD 1 to GD 17, whereby the mean fetal weights in the 2.5 and
12.5 mg/kg/day groups were decreased by approximately 17% and 24%, respectively, relative to
controls. However, the reduction in weight did not reach significance, though it should be noted
that the sample size was small (n = 3 litters/group). Li et al. (2016, 3981495) reported mean GD
18.5 fetal body weights of 2.73, 2.68, and 2.48 g in the 0, 5, and 20 mg/kg/day dose groups
(sexes combined) following exposure of Sprague-Dawley rat to PFOS from GD 12 to GD 18.
3-237
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Mean fetal body weight for the 20 mg/kg/day dose group was significantly different from that of
the control group. Mean fetal body weight in males alone was also significantly decreased at
20 mg/kg/day (2.79, 2.74, and 2.43 g for the 0, 5, and 20 mg/kg/day dose groups, respectively).
Thibodeaux et al. (2003, 757855) similarly observed a decrease in rat fetal weight following
gestational exposure to 10 mg/kg/day PFOS. In a one-generation reproductive study in Sprague
Dawley rats, Luebker et al. (2005, 757857) reported no effect on pooled fetal body weights with
PFOS doses up to 2 mg/kg/day. Similarly, Conley et al. (2022, 10176381) found no effects of
PFOS on fetal body weight on GD 18 in Sprague-Dawley rats (Crl:CD(SD)) exposed to 0, 0.1,
0.3, 1, 3, 10, or 30 mg/kg/day from GD 14 to GD 18. In a study in New Zealand white rabbits,
Argus Research Laboratories (2000, 5080012) reported mean live fetal body weights of 44.15,
41.67, 42.37, 39.89, and 33.41 g/litter in 0, 0.1, 1, 2.5, and 3.75 mg/kg/day dose groups,
respectively. Fetal body weights for the 2.5 and 3.75 mg/kg/day dose groups were significantly
lower than fetal body weight reported in the control group.
Several other studies measured body weights of pups after birth (Figure 3-70). Zhang et al.
(2021, 6988534) found no PFOS-related effects on pup body weight on PND 1, 3, 7, and 14 in
Sprague-Dawley rat pups exposed to 0, 1, or 5 mg/kg/day from GD 12 to GD 18. The most
sensitive endpoint in the one- and two-generation reproductive studies in Sprague Dawley rats
(dams treated with PFOS pre-conception through gestation for 63 or 84 days, respectively) was
decreased pup body weight {Luebker, 2005, 757857; Luebker, 2005, 1276160}. The NOAEL
and LOAEL for pup body weight effects was 0.1 and 0.4 mg/kg/day, respectively, in the two-
generation study {Luebker, 2005, 1276160}; the lowest dose of 0.1 mg/kg/day was not tested
(NT) in the one-generation study {Luebker, 2005, 757857} where the LOAEL was
0.4 mg/kg/day for decreased pup body weight, decreased maternal body weight, and decreased
gestation length. Lau et al. (2003, 757854) also reported significant weight deficits in Sprague
Dawley rat pups on PND 0 after gestational PFOS exposures of 2, 3, or 5 mg/kg/day, but not
1 mg/kg/day. Similarly, Xia et al. (2011, 2919267) observed significantly reduced pup body
weights in Sprague Dawley rats on PND 0 and PND 21 following gestational exposure to 2
mg/kg/day PFOS.
For this endpoint, rats appear to be more sensitive than mice. Yahia et al. (2008, 2919381)
reported significant decreases in ICR mouse neonatal weight at relatively high doses of 10 and
20 mg/kg/day. Lau et al. (2003, 757854) did not report statistically significant reductions in pup
body weights of CD-I mice gestationally exposed to PFOS doses up to 20 mg/kg/day. Zhong et
al. (2016, 3748828) measured body weights of C57BL/6 mouse pups that had been exposed to 0,
0.1, 1, or 5 mg/kg/day PFOS in utero from GD 1 to GD 17. They did not see significant
differences in body weight measurements of male or female mice at 4 and 8 weeks of age.
Mshaty et al. (2020, 6833692) also reported no effects on C57BL/6J mouse pup body weight at
PND 21 following lactational exposure to 1 mg/kg/day PFOS from PND 1 to PND 14.
3-238
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Endpoint
Fetal Body Weight
Pup Body Weight
Study Name Study
Argus, 2000, 5080012 developmental (GD7-20)
Lee et al„ 2015, 2851075 developmental (GD11-16)
Wan at al., 2020, 7174720 developmental (GD4.5-17.S)
Li al al., 2021, 9<559491 developmental (GD1-17)
Fuerilas al al., 2006, 757B59 developmental (GDfi-18)
Conley et al., 2022,10176381 developmental (GD14-18)
Li el al„ 2016, 3981495 developmental (GD12-18)
Luebker et al., 2005. 757857 reproductive (76d (42d pre-cohabitation. 1
Zhong et al.. 2016. 3748828 developmental (GD1-17)
Observation Tims
GD29
GD17
GD17.5
GD1B
GD1B
GD18
GD18
d mating. GDO-20)) GD21
PNW4
Animal Description
Rabbit, New Zealand (-C'-', N=12-20)
Mouse, CD-1 {«•--, N=10)
Mouse, CD-1 {y'2, N=B)
Mouse, No I Sped lied N=3)
msa, CD-1 (;,N=10-11)
P0 Rat. Sprague-Dawley (i, N=4-6)
Ral, Sprague-Dawley (c?i, N=10)
Rat, Sprague-Dawley {'+. N=10)
Rat, Sprague-Dawley (t. N=10)
Rat, Crl:Cd[Sd)lgs Vaf,'Plus ( Vj. N=8)
Mouse. C57BL'6 ('*, N=12)
Mouse, C57BL/8 ( N=12)
Lau et al.. 2003, 757854
developmental (GD1-17)
PND0
F1 Mouse, CD-1 N=20)
PND21
F1 Mouse. CD-1 {n'--, N=20)
PND35
F1 Mouse,
CD-1 N=20)
developmental (GD2-21)
PNDO
F1 Ral, Spiayue-DnwIay (o"\ N
5-8)
PND21
F1 Rat, Sp
ague-Dawley (c?2, N=8)
PND35
F1 Rat, Sprague-Dawley ( "J, N=8)
Xiaetal.. 2011, 2919267
developmental (GD2-21)
PNDO
F1 Rat, Sprague-Dawley ( N
10)
Zhang et al. 2021. 6988534
developmental (GD12-18)
PND1
F1 Rat. Sprague-Dawley ( N=8)
PND3
F1 Rat. Sprague-Dawley (J'\ N=8)
PND7
F1 Rat. Sprague-Dawley (N=8)
PND14
F1 Rat, Sprague-Dawley N
6)
Butanhoff et al., 2009, 757873
developmental (GD0-PND20)
PND1
F1 Rat, Crl
CD(SD) (o, N=20)
F1 Rat, Cr
CD(SD) N=20)
PND21
F1 Ral, Crl
CD(SD) N=20)
F1 Rat, Cr
CD(SD) (i, N=20)
up Body Weight Relfllivs lo Lit
er LuebKer el al., 2005, 767857
reproductive (42d prior mating-LD4)
PNDO
F1 Ral, Cr
Cd(Sd)lgs Var.'Plus
i, N=17)
LD5
F1 Rat, Cr
Cd(Sd)lgs Vaf.'Plus (,
ii. N=17)
Luebkeretal.,2005. 1276160
reproductive (42d prior mating-LD20)
PND1
F1 Rat, Cr
Cd (Sd)lgs Br Vaf (;*
. N=20-25)
PND4 (preculling)
F1 Rat, Cr
Cd (Sd)lgs Br Vaf (:-
. N=20-25)
PND4 (postculling)
F1 Rat. Cr
Cd (Sd)lgsBrVaf (r'
. N=20-25)
PND7
F1 Rat, Cr
Cd (Sd)lgsBrVaf (c
. N=20-25)
PND14
F1 Rat, Crl
Cd (Sd)lgsBrVaf (<;
. N=20-25)
PND21
F1 Rat, Cr
Cd (Sd)lgs Br Vaf | ;
. N=20-25)
reproductive (GD0-PND21)
PND1
F2 Ral, Cr
Cd (Sd)lgs Br Val (o*
, N=22-25)
PND4 (preeulhg)
F2 Ral, Cr
Cd (Sd)lgsBrVaf (.2
. N=22-25)
PND4 (postculling)
F2 Ral, Cr
Cd (Sd)lgsBrVal(.J
, N=22-25)
PND7
F2 Rat, Cr
Cd (Sd)lgs Br Vaf fc-
, N=22-25)
PND14
F2 Rat, Cr
Cd (Sd)lgs Br Vaf fc-
. N=22-25)
PND21
F2 Rat, Cr
Cd (Sd)lgs Br Vaf (r-
. N=22-25)
PFOS Developmental Effects - Offspring Weight
| 0 No significant change A Significant increase ^ Significant decrease]
-w-v
V
V
V V
-w
-w
-w
-w
-w
Figure 3-70. Offspring Weight in Mice, Rats, and Rabbits Following Exposure to PFOS
(logarithmic scale, sorted by observation time)
PFOS concentration is presented in logarithmic scale to optimize the spatial presentation of data.
Interactive figure and additional study details available on HAWC.
GD = gestation day; PND = postnatal day; LD = lactational day; Fi = first generation; F2 = second generation; d = day.
3.4.4.2.5 Placenta
Placental endpoints were reported in six studies with rats, mice, or rabbits. Li et al. (2016,
3981495) reported a significant decrease in mean placental weight in Sprague-Dawley rat dams
exposed to 20 mg/kg/day PFOS from GD 12 to GD 18 relative to control (442.8 mg vs.
480.4 mg). No significant difference in placental weights was detected in dams exposed to
5 mg/kg/day PFOS relative to control (455.1 mg vs. 480.4 mg). At > 0.5 mg/kg/day, Lee et al.
(2015, 2851075) observed significant decreases in mean absolute placental weight (185.63,
177.32, 163.22, and 151.54 mg at 0, 0.5, 2, and 8 mg/kg/day, respectively) and placental capacity
(ratio of fetal weight/placental weight; 9.3, 8.68, 7.96, and 7.39 at 0, 0.5, 2, and 8 mg/kg/day,
respectively) in mice exposed to PFOS from GD 11 to GD 16 and sacrificed at GD 17. In the
same study, microscopic evaluation revealed necrotic changes and dose-dependent decreases in
the frequency of glycogen trophoblast cells and sinusoidal trophoblast cells at dose levels > 2.0
and > 0.5 mg/kg/day, respectively {Lee, 2015, 2851075}. Li et al. (2021, 9959491) dosed mouse
dams (strain not specified) with 0, 0.5, 2.5, or 12.5 mg/kg/day PFOS from GD 1 to GD 17 and
observed smaller placental diameter in the 12.5 mg/kg/day group compared to controls, though
the statistical significance of that effects is unclear. Wan et al. (2020, 7174720) found no effects
3-239
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
on absolute or relative placenta weight, junctional zone area, labyrinth zone area, or the ratio of
labyrinth to junctional zone area in CD-I mice exposed to 0, 1, or 3 mg/kg/day PFOS from GD
4.5 to GD 17.5. Argus Research Laboratories (2000, 5080012) did not observe any placental
effects in exposed rabbits and Luebker et al. (2005, 757857) observed no changes in placental
size, color, or shape in exposed rats.
3.4.4.2.6 Postnatal Development
Gestational PFOS exposure is associated with effects on postnatal development. Lau et al. (2003,
757854) observed delayed eye opening in rats and mice following developmental exposure to
PFOS. A significant, treatment-related delay in eye opening was reported in mice following
gestational exposure to PFOS (eye opening at PND 14.8 in control vs. Eye opening at PND 15.1,
PND 15.5, and PND 15.6 at 1, 5, and 10 mg/kg/day, respectively). TheNOAEL for delays in eye
opening in rats was 1 mg/kg/day PFOS. Mshaty et al. (2020, 6833692) evaluated age at eye
opening in mice exposed to 1 mg/kg/day from PND 1 through PND 14 and found no significant
effects. A two-generation reproduction study in rats {Luebker, 2005, 1276160} evaluated
various developmental landmarks in the Fi offspring and observed significant delays in pups
attaining pinna unfolding, eye opening, surface righting, and air righting in the 1.6 mg/kg/day
dose group. Eye opening was also slightly, but significantly, delayed in pups exposed to 0.4
mg/kg/day.
Developmental PFOS exposure also had adverse effects on lung development, further described
in the Respiratory Section of the PFOS Appendix (Section C.7).
3.4.4.3 Mechanistic Evidence
Mechanistic evidence linking PFOS exposure to adverse developmental outcomes is discussed in
Section 3.3.4 of the 2016 PFOS HESD (EPA, 2016, 3603365). There are 33 studies from recent
systematic literature search and review efforts conducted after publication of the 2016 PFOS
HESD that investigated the mechanisms of action of PFOS that lead to developmental effects. A
summary of these studies is shown in Figure 3-71.
3-240
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Mechanistic Pathway
Animal
Human
In Vitro
Grand Total
Angiogenic, Antiangiogenic, Vascular Tissue Remodeling
1
0
0
1
Big Data, Non-Targeted Analysis
5
6
4
14
Cell Growth, Differentiation, Proliferation, Or Viability
7
0
15
19
Cell Signaling Or Signal Transduction
5
1
5
10
Extracellular Matrix Or Molecules
0
0
1
1
Fatty Acid Synthesis, Metabolism, Storage, Transport, Binding, B-Oxidation
3
1
2
6
Hormone Function
2
0
1
2
Inflammation And Immune Response
0
1
1
2
Oxidative Stress
1
1
3
5
Xenobiotic Metabolism
1
0
2
3
Not Applicable/Not Specified/Review Article
1
0
0
1
Grand Total
13
7
16
33
Figure 3-71. Summary of Mechanistic Studies of PFOS and Developmental Effects
Interactive figure and additional study details available on Tableau.
Mechanistic data available from in vitro, in vivo, and epidemiological studies were evaluated to
inform the mode of action of developmental effects of PFOS. Outcomes included early survival,
general development, and gross morphology; fetal growth and placental effects; metabolism;
lung development; hepatic development; testes development; cardiac development; and
neurological development.
3.4.4.3.1 Early Survival, General Development, Gross Morphology
Mechanisms through which PFOS exposure may alter survival and development were studied in
several zebrafish embryo bioassay studies. Several of these studies identified in the current
assessment were included in a recent review of developmental effects of PFOS in zebrafish
models {Lee, 2020, 6323794}. In general, PFOS can lead to embryo and/or larva malformation,
delays in hatching, and decreases in body length. Wang et al. (2017, 3981383) exposed embryos
to 0.2, 0.4, 0.8, or 1.6 mg/L PFOS and observed significant and dose-dependent reductions in
hatching rate and heart rate as well as significant increases in mortality and malformations in the
spine and swim bladder. The overt effects in general development and gross morphology
coincided with dose-dependent increases in reactive oxygen species (ROS), lipid peroxidation,
and antioxidant enzyme activity (including catalase (CAT), superoxide dismutase (SOD), and
glutathione peroxidase (GSH-Px)). Interestingly, co-exposure of the embryos with PFOS and
multi-walled carbon nanotubes (MWCNTs) reduced toxicity in several of these endpoints and
attenuated the increase in oxidative stress biomarkers caused by PFOS, suggesting that oxidative
stress is a key event that mediates alterations in development and gross morphology following
exposure to PFOS. Another zebrafish embryo bioassay conducted by Dang et al. (2018,
3-241
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
4651759) reported that exposure to 0.1, 1, or 10 |iM PFOS did not affect hatching and survival
rates, but did increase malformation rates by 7%, possibly due to downregulation of the growth
hormone/insulin-like growth factors (GH/IGFs) axis. Blanc et al. (2019, 5413062) determined
the lethal/effect concentrations (LC/ECs) for zebrafish embryos at 96-hours post-fertilization
(hpf). The 50% lethal effect concentration (LCso) was 88 [xM, which is lower than the previously
determined value of 109 |iM by Hagenaars et al. (2011, 1279113). The 10% lethal effect
concentration (LCio) was 35 |iM and was used in subsequent experiments to explore mechanisms
that may contribute to the developmental toxicity at the transcriptional and epigenetic level,
which are described in the Section below {Blanc, 2019, 5413062}. Lastly, Chen et al. (2014,
2540874) found that PFOS exposure of zebrafish embryos led to several malformations,
including uninflated swim bladder, underdeveloped gut, and curved spine, which paralleled
histological alterations in the swim bladder and gut. To complement the functional data, the
authors examined differential gene expression by microarray analysis, which revealed
upregulated genes involved in nucleic and macromolecule metabolism, cell differentiation and
proliferation, neuron differentiation and development, and voltage-gated channels. Genes that
were downregulated were associated with cellular protein metabolic processes, macromolecular
complex assembly, protein-DNA complex assembly, and positive regulation of translation and
multicellular organism growth. The authors also used the genomic data to identify the top
predicted developmental toxicity pathways initiated by PFOS exposure, including Peroxisome
Proliferator-Activated Receptor alpha (PPARa)-mediated pathways, decreases of transmembrane
potential of mitochondria and mitochondrial membrane, and cardiac necrosis/cell death.
Two in vitro studies by Xu et al. (2013, 2968325; 2015, 2850066) examined the effects of PFOS
on changes in mouse embryonic stem cell (mESC) pluripotency markers, which control normal
cell differentiation and development. Xu et al. (2013, 2968325) found that PFOS exposure did
not affect cell viability. However, PFOS exposure decreased mRNA and protein levels of the
pluripotency markers Sox2 and Nanog, but not Oct4. They also measured several miRNAs,
including miR-145 and miR-490-3p, which can regulate Sox2 and Nanog, and found them to be
increased, supporting the epigenetic mechanisms of control of these markers. In Xu et al. (2015,
2850066), cell differentiation effects on mouse embryoid bodies (mEBs) were examined. eBs are
formed when embryonic stem cells spontaneously differentiate into the three germ cell layers,
mimicking early gastrulation. The authors found that mEB formation was unaffected by PFOS,
but that PFOS exposure increased the mRNA and protein levels of the previously studied
pluripotency markers (Oct4, Sox2, and Nanog); this is notably a reversal of the findings from
their previous study in mESCs {xu, 2013, 2968325}. Xu et al. (2015, 2850066) found that PFOS
exposure in mEBs decreased differentiation markers (Soxl7, FOXA2, SMA, Brachyury, Nestin,
Fgf5), as well as Polycomb group (PcG) proteins and several miRNAs also involved in
differentiation. These alterations could disturb the dynamic equilibrium of embryonic
differentiation and induce developmental toxicity. Altogether, the results suggest that PFOS
exposure can disturb the expression of pluripotency factors that are essential during early
embryonic development, potentially via miRNA dysregulation, which may reflect mechanisms
of toxicity that are relevant during a critical window of embryonic development.
Global epigenetic changes in response to PFOS exposure were measured in several studies,
including in one zebrafish study and two epidemiological studies. Blanc et al. (2019, 5413062)
found that PFOS induced global DNA hypermethylation, minor alterations in gene expression of
several epigenetic factors (including DNA methylation, histone deacetylation, and histone
3-242
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
demethylation factors) following PFOS exposure. Moreover, the genes encoding the DNA
methyltransferase dnmt3ab and the H3K4 histone demethylase kdm5ba were significantly
downregulated. H3K4 methylation is associated with open, transcriptionally active regions and
depleted of DNA methylation. The authors did not measure methylation patterns on H3K4 or
other histones; to confirm alterations to H3K4 methylation status, additional studies are required.
In cord blood samples from a Japanese birth cohort study, Miura et al. (2018, 5080353)
measured PFOS levels in tandem with epigenetic modifications during fetal development. The
authors found significant associations between global hypermethylation and PFOS exposure. The
top differentially methylated regions (DMRs) of the genome that were associated with PFOS
exposure included hypermethylation of CpG sites of CYP2E1, SMAD, and SLC17A9; however,
the authors did not measure the expression level of these genes to confirm the effect of the
epigenetic alterations. In contrast, another study of human cord blood samples conducted by Liu
et al. (2018, 4926233) found that PFOS exposure was associated with low methylation of Alu
retrotransposon family in cord blood DNA samples, indicating global hypomethylation.
Demethylation of Alu elements has been proposed to induce insertion and/or homologous
recombination and cause alterations to genomic stability and, subsequently, gene transcription. In
another study of human cord blood samples, PFOS exposure was associated with DNA
methylation changes at key CpG sites associated with genes in pathways important for several
physiological functions and diseases, including nervous system development, tissue morphology,
digestive system development, embryonic development, endocrine system development, cancer,
eye disease, organ abnormalities, cardiovascular disease, and connective tissue disorders {Leung,
2018, 4633577}.
Lastly, in a study of human cord blood in a prospective cohort in China, PFOS exposure was
associated with significantly shorter leukocyte telomere lengths and increased ROS in female
newborns. Interestingly, the effects were not observed in male newborns, suggesting sex-specific
effects in early-life sensitivity to PFOS exposure at the molecular level. The authors determined
that the effect of PFOS on shortened leukocyte telomere length was partially mediated through
ROS in females, indicating a programing role of PFOS on telomere length during gestation {Liu,
2018, 4239494}.
3.4.4.3.2 Fetal Growth and Placental Development
Growth was measured in developing zebrafish larvae in three studies. Wang et al. (2017,
3981383), reported a dose-dependent reduction in body length that coincided with dose-
dependent increases in ROS generation, lipid peroxidation, and the activities of antioxidant
enzymes in larvae exposed to 0.2, 0.4, 0.8, or 1.6 mg/L PFOS. Reduction in body length was
likely due to PFOS-related increased oxidative stress and lipid peroxidation. In Jantzen et al.
(2016, 3860114), the morphometric endpoints of interocular distance, total body length, and yolk
sac area were measured in zebrafish embryos. PFOS exposure significantly decreased all three
parameters relative to controls, indicating slowed embryonic development, at values 5- to 25-fold
below previously calculated LCso values. The authors found alterations in the expression of
several genes involved in development, including calcium ion binding (calm3a), cell cycle
regulation (cdknla), aromatic compound metabolism (cypla), and angiogenesis (flkl), as well as
increased tfc3a (muscle development) expression and decreased apis (protein transport). Lastly,
Dang et al. (2018, 4651759) found that PFOS significantly inhibited body length and growth of
larvae. This appeared to be mediated through the growth hormone/insulin-like growth factor
3-243
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
(GH/IGF) axis, as several GH/IGF axis genes had decreased expression, including the genes
growth hormone releasing hormone (ghrh), growth hormone receptors a and b (ghra and ghrb),
insulin-like growth factor 1 receptor a and b (igflra and igflrb), insulin-like growth factor 2
receptor (igf2r), insulin-like growth factor 2a (igf2a), and insulin-like growth factor binding
protein 2a and 2b (;lgfbp2a and igfbp2b).
In three in vivo rodent studies, fetal growth and placental disruption in response to maternal
PFOS exposure were measured. In a mouse study, Lee et al. (2015, 2851075) reported a
relationship between gene expression of prolactin-family hormones and placental and fetal
outcomes following maternal exposure to 0, 0.5, 2.0, or 8.0 mg/kg/day PFOS from GD 11-16 via
gavage. Dose-dependent increases in placental histopathological lesions and reductions in
placental weights, fetal weights, and number of live fetuses were significantly correlated with
reductions in gene expression of mouse placental lactogen (mPL-Il), prolactin-like protein Ca
(,mPLP-Ca), and prolactin-like protein K (mPLP-K). Given the alterations in prolactin-family
gene expression, the authors propose that this placental disruption is related to endocrine (i.e.,
prolactin) dysfunction. Li et al. (2016, 3981495) also found that maternal PFOS exposure
reduced fetal and placental weight, which coincided with increased corti coster one in fetal serum.
In the placenta, activity of 1 lb-hydroxysteroid dehydrogenase 2, and expression of several genes
involved in development (i.e., extracellular matrix, growth factors and hormones, ion
transporters, signal transducers, and structural constituents) were downregulated, suggesting
intrauterine growth restriction was related to altered placental development and functionality. Li
et al. (2020, 6833703) also found that PFOS exposure was associated with reduced placental size
in mice and proposed that the disruption was mediated by the dysregulation of a long non-coding
RNA, H19 which plays a role in regulation of embryonic growth {Monnier, 2013, 10439067},
which was altered in placental tissues (i.e., hypomethylation of the H19 promoter and increased
expression of the gene). In vitro experiments in human placental trophoblast cells (HTR-
8/sVneo) provided further support for a mechanism involving H19; cell growth that was
inhibited by PFOS was partially alleviated following suppression of H19 via transfection with si-
H19 {Li, 2020, 6833703}.
Sonkar et al. (2019, 5918797) also used HTR-8/sVneo cells to evaluate the epigenetic
mechanisms through which PFOS exposure adversely effects the placenta. The authors reported
increased ROS production, possibly due to alterations of several DNA methyltransferases and
sirtuins, which consequently led to a reduction in global DNA methylation and increased protein
lysine acetylation. The authors propose that ROS production could lead to pregnancy
complications, such as preeclampsia and intrauterine growth restrictions.
In a human placental choriocarcinoma cell line (JEG-3), PFOS exposure was found to induce
placental cell cytotoxicity and inhibition of aromatase activity {Gorrochategui, 2014, 2324895}.
In Yang et al. (2016, 3981458), 0.1 |iM PFOS inhibited decidualization of the first trimester
human decidual stromal cells (collected from the uterine lining). PFOS also downregulated 11-
hydroxysteroid dehydrogenase 1 (11P-HSD1), an enzyme that converts the inactive form of
Cortisol to the active form of Cortisol, and inhibited the glucocorticoid-driven reduction of the
proinflammatory cytokines IL-6 and IL1-P, which could result in a reduced immune-tolerance
environment in early pregnancy. In human amnion and fetal lung cells exposed to PFOS in vitro,
PFOS exposure upregulated the gene expression of Caspase3 and apoptotic peptidase activating
3-244
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
factor 1 (APAF1), genes that initiate apoptosis. This effect was concentration (between 10 4 and
1CT6 M PFOS) and time-dependent (between 24 and 48 hours) {Karakas-Celik, 2014, 2850400}.
Lastly, in humans, Ouidir et al. (2020, 6833759) recruited pregnant women and measured plasma
PFOS levels during the first trimester of the pregnancy and examined global methylation in the
placenta at birth. The authors found significant associations between PFOS exposure and DNA
methylation changes in the placenta, and the associated downregulation of certain genes,
particularly the reduced gene expression of several genes associated with anthropometry
parameters such as shorter birth length, reduced birth weight, and reduced head circumference
that were previously associated with PFAS exposure {Buck, 2018, 5016992}. These data suggest
that the prenatal toxicity of PFOS might be driven by epigenetic changes in the placenta {Ouidir,
2020, 6833759}.
3.4.4.3.3 Metabolism
Metabolomic profiles in relation to PFOS exposure were analyzed in humans in two studies. In a
cross-sectional study in 8-year-old children in Cincinnati, OH, the authors conducted untargeted,
high-resolution metabolomic profiling in relation to serum PFOS concentrations. They found that
PFOS exposure was associated with several lipid and dietary factors, including arginine, proline,
aspartate, asparagine, and butanoate metabolism {Kingsley, 2019, 5405904}. In a study of
mothers that were part of the Child Health and Development Studies (CHDS) cohort, maternal
serum was analyzed for PFOS as well as underwent metabolomics profiling to determine if
metabolic alterations reflected in measurements from maternal serum could possibly contribute
to later health outcomes in their children {Hu, 2019, 5412445}. PFOS exposure was associated
with a distinct metabolic profile, including a positive association with urea cycle metabolites and
a positive association with carnitine shuttle metabolites. This profile indicates disruption of fatty
acid metabolism, which could possibly cause developmental alterations in offspring {Hu, 2019,
5412445}.
3.4.4.3.4 Lung Development
In a human fetal lung fibroblast cell line (Hel299), PFOS exposure upregulated the expression of
Caspase3 and Apafl, genes that initiate apoptosis. This effect was dose and time-dependent
{Karakas-Celik, 2014, 2850400}. These results indicate that PFOS can cause in vitro toxicity
(via apoptotic mechanisms) in embryonic cells, possibly affecting the development.
3.4.4.3.5 Hepatic Development
Liang et al. (2019, 5412467) studied the effects of developmental exposure to PFOS on
metabolic liver function in Kunming mice, in post-natal day 1 offspring. They found that PFOS
exposure during gestation increased liver triglycerides, total cholesterol, and low density
lipoprotein (LDL), and decreased high density lipoprotein (HDL) in the offspring. The mRNA of
several factors involved in fatty acid oxidation, update, and hepatic export of livers were altered,
indicating developmental perturbation of lipid metabolic function. These in vivo results show
that PFOS may disrupt hepatic lipid metabolism through negative effects on hepatocellular lipid
trafficking in mice developmentally exposed to PFOS.
3.4.4.3.6 Cardiac Development
Several in vitro studies examined developmental toxicity of PFOS using embryonic stem cell-
derived cardiomyocytes (ESC-CMs) as a model of the early stages of heart development {Cheng,
3-245
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
2013, 2850971; Zhou, 2017, 3981356; Zhang, 2016, 3981565; Tang, 2017, 3981359; Liu, 2020,
6833698; Yang, 2020, 6315676}. Most of the studies utilized mouse ESC-CMs but one study,
Yang et al. (2020, 6315676), used a human ESC-CM model of cardiac differentiation. Cardiac
differentiation was inhibited in PFOS-treated mouse ESC-CMs, shown by a concentration-
dependent decrease in the contract positive rate (i.e., percentage of beating embryoid bodies) on
differentiation days 8-10 {Cheng, 2013, 2850971; Zhou, 2017, 3981356; Zhang, 2016, 3981565;
Tang, 2017, 3981359} and a decreased proportion of a-actinin-positive cells (a marker of
cardiomyocytes) on differentiation day 10 {Zhang, 2016, 3981565; Tang, 2017, 3981359}. The
median inhibition of differentiation (ID50), defined as the concentration at which PFOS inhibited
the development of contracting cardiomyocytes by 50%, ranged from 40 |iM (Zhang et al. (2016,
3981565) to 73 |iM {Zhou, 2017, 3981356}. Collectively, these results provide in vitro evidence
of potential developmental cardiotoxicity following PFOS exposure.
Several in vitro studies have demonstrated that PFOS can significantly alter gene and protein
expression at multiple time points during differentiation of cardiomyocytes from mouse or
human ESCs, specifically for genes in the myosin heavy chain, myosin light chain, and cardiac
troponin T families. In human ESC-CMs, 0.1-60 |iM PFOS significantly inhibited the
expression of cardiac-specific homeobox gene Nk2 homeobox 5 (NKX2.5), myosin heavy chain
6 (MYH6), and myosin light chain 7 (MYL7'), and significantly reduced protein levels of NKX2.5
and cardiac troponin T2 (TNNT2) on day 8 and/or day 12 of differentiation {Yang, 2020,
6315676}. In mouse ESC-CMs, on differentiation day 5, PFOS (20-40 |iM) reduced gene and
protein levels of Brachyury (mesodermal marker), cardiac transcription factors GATA binding
protein 4 (GATA4), and myocyte enhancer factor 2C (MEF2C) {Zhang, 2016, 3981565}. On
differentiation days 9-10, PFOS reduced the expression ofMyh6 and Tnnt2 (i.e., cTnT) in a
dose-dependent manner from 2.5 to 160 |ig/mL PFOS {Cheng, 2013, 2850971; Zhou, 2017,
3981356}. Cheng et al. (2013, 2850971) found that PFOS significantly altered the chronological
order of gene expression during in vitro cardiogenesis. Expression of important cardiac genes
were significantly lower in PFOS-treated cells compared to controls on day 9, but expression of
Nkx2.5 and Mlcla were significantly higher in PFOS-treated cells by day 14 of differentiation
{Cheng, 2013,2850971}.
Proteomic analysis during cardiac differentiation of mouse ESCs revealed 176 differentially
expressed proteins (67 upregulated and 109 downregulated) {Zhang, 2016, 3981565}. The
differentially expressed proteins were mainly associated with catalytical activity, protein binding,
nucleotide binding, and nucleic acid binding. PFOS significantly affected 32 signaling pathways,
with metabolic pathways the most affected. The PPAR signaling pathway and mitogen-activated
protein kinase (MAPK) signaling pathways were also significantly affected by PFOS.
Yang et al. (2020, 6315676) studied global gene expression during cardiac differentiation of
human ESCs exposed to 60 |iM PFOS. Their analysis revealed 584 differentially expressed
genes (247 upregulated and 337 downregulated) on differentiation day 8, and 707 differentially
expressed genes (389 upregulated and 318 downregulated) on differentiation day 12. In total,
199 genes were affected on both days 8 and 12. The majority of affected genes are related to
extracellular matrix and cell membrane. Seven Kyoto Encyclopedia of Genes and Genomes
(KEGG) pathways were affected by PFOS on both days (mostly neural-related pathways and a
few general pathways), but cardiac pathways were not greatly affected. PFOS downregulated
cardiac markers such as natriuretic peptide A (NPPA), natriuretic peptide B (NI'I'Bj, NKX2.5,
3-246
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
MYH6, MYL2, and MYH7, but upregulated epicardial markers WT1 transcription factor (WT1)
and T-box transcription factor 18 (TBX18). Wingless-related integration site (WNT) signaling
pathway-related genes (secreted frizzled related protein 2 (SFRP2) and frizzled-related protein
(FRZB)) and IGF signaling pathway genes (IGF2 and IGF binding protein 5 (IGFBP5')) were
significantly upregulated in PFOS-treated cells. The authors postulated that PFOS stimulated
differentiation to epicardial cells more than to cardiomyocytes by stimulating the WNT signaling
pathway.
Mouse ESC cardiac differentiation assays have demonstrated that exposure to PFOS can cause
mitochondrial toxicity in these cells. In contrast, one study in human ESCs-derived
cardiomyocytes {Yang, 2020, 6315676} found that PFOS did not affect mitochondrial integrity
on day 12 of differentiation.
Cheng et al. (2013, 2850971) found that PFOS reduced ATP production, increased accumulation
of ROS, and stimulated apoptosis in mouse ESC-CMs. However, Tang et al. (2017, 3981359)
demonstrated that PFOS decreased intracellular ATP and lowered mitochondrial membrane
potential in mouse ESC-CMs without inducing apoptosis. Exposure to PFOS during cardiac
differentiation also caused structural damage to mitochondria (e.g., swelling, vacuolar structure,
loss of cristae) and the mitochondria-associated endoplasmic reticulum membrane (MAM).
Furthermore, PFOS increased intracellular lactate production, fatty acid content, and disrupted
calcium fluxes. Analysis of protein expression demonstrated that destruction of the MAM
structure occurred along with activation of Rictor/mTORC2 signaling pathway via
phosphorylation of epidermal growth factor receptor, which led to accumulation of intracellular
fatty acid and resulted in blocking of the [Ca2+]mito transient.
The mechanisms behind PFOS mitochondrial toxicity were further explored by Liu et al. (2020,
6833698) who found that PFOS-treated ESC-CMs displayed autophagosome accumulation
accompanied by increased levels of p62 and ubiquitinated proteins, increased lysosomal pH, and
decreased the levels of lysosome-associated membrane protein (Lamp2a) and the mature form of
Cathepsin D (lysosomal protease), suggesting an impairment of autophagy-lysosome
degradation. PFOS also blocked mitophagy, the removal of damaged mitochondria through
autophagy, thereby disrupting the balance between mitophagy and biogenesis {Liu, 2020,
6833698}. The authors postulated that the mechanism of PFOS-induced toxicity to ESC-CMs
involves reduced lysosomal acidification, inhibited maturation of cathepsin D, blocked fusion
between lysosomes and autophagosomes, accumulation of autophagosomes, and dysfunctional
mitochondria.
One study included in the prior 2016 PFOS HESD {U.S. EPA, 2016 3603365} investigated
cardiac mediated apoptosis in weaned rats exposed to PFOS (0, 0.1, 0.6, or 2 mg/kg/day) on GD
2-21 {Zeng, 2014, 2851284}. The pups were sacrificed at the end of the lactation period, and
trunk blood and the heart were recovered. Apoptotic cells in the heart tissue from six animals per
dose group were measured using a Terminal deoxynucleotidyl transferase dUTP nick end labeling
(TUNEL) staining assay. PFOS exposure was associated with a dose dependent increase in the
percentage of TUNEL positive nuclei. The 0.6 mg/kg/day dose was the LOAEL and the
0.1 mg/kg/day dose the NOAEL. The researchers found that biomarkers for apoptosis were
supportive of the TUNEL results. The expression of BCL2-associated X protein and cytochrome
c were upregulated and bcl-2 downregulated. The concentration of caspase 9 was significantly
3-247
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
increased above the control levels at all doses and caspase 3 levels were significantly increased
for all but the lowest dose level.
3.4.4.3.7 Testicular Development
Two rat studies examined PFOS effects on testicular development. Zhang et al. (2013, 1598626)
isolated primary Sertoli cells and gonocytes from 5-day-old rat pups and created a Sertoli
cell/gonocyte coculture system to mimic in vivo interactions. PFOS exposure reduced cell
viability and induced ROS production in a concentration-dependent manner, although PFOS did
not appear to increase apoptosis. PFOS exposure altered and inhibited the cytoskeletal proteins
vimentin and F-actin in Sertoli cells, indicating PFOS could adversely affect developing testes
via ROS and cytoskeleton disruption. Li et al. (2018, 4241058) examined the effects of PFOS on
pubertal Ley dig cell development, both in vitro and in vivo. In vitro, PFOS inhibited androgen
secretion via the downregulation of 17b-hydroxysteroid dehydrogenase 3 (HSD17B3, gene
Hsdl8b3), as measured by Hsdl8b3 mRNA expression. PFOS also promoted apoptosis of
immature Ley dig cells in vitro but did not affect cell proliferation. In vivo, PFOS exposure
reduced serum testosterone levels, and reduced sperm production. LHCGR, CYP11 Al, and
CYP17A1 levels in Ley dig cells were reduced, suggesting that PFOS exposure downregulates
critical Leydig cell gene expression, indicating delayed maturation of these cells.
3.4.4.3.8 Neurological Development
PFOS effects on neurodevelopment and behavior in zebrafish were examined in two studies. In
the zebrafish embryo assay by Jantzen et al. (2016, 3860114), embryonic exposure to PFOS
resulted in hyperactive locomotor activity in larvae, possibly mediated through altered
expression of development-associated genes (calm3a, cdknla, cypla,flkl, tfc3a, and apis).
Stengel et al. (2018, 4238489) developed a neurodevelopmental toxicity test battery in zebrafish
embryos and evaluated the effect of PFOS exposure. Although PFOS exposure had significant
adverse effects on neuromast cells, including degeneration, no changes were observed in the
olfactory or retinal toxicity assays.
Rat embryonic neural stem cells (NSCs) were used to examine the effects of PFOS on neuronal
and oligodendrocytic differentiation. PFOS exposure at 25 or 50 nM reduced cell proliferation
but showed increased protein levels in markers associated with differentiation (TuJl, CNPase).
Exposure also reduced the number of cells with spontaneous calcium activity. These effects
appeared to be mediated through PPAR pathways, as indicated by increases in PPARy and the
downstream target UCP2. Results were confirmed using a PPARy agonist that showed similar
effects in the cells. This study also evaluated effects of PFOS exposure on the PPAR system in
vivo. In PFOS-treated neonatal mice, PPARy and UCP3 were up-regulated in brain cortical tissue
{Wan Ibrahim, 2013, 2919149}.
Lastly, Leung et al. (2018, 4633577) conducted a genome-wide methylation study on mothers
and infants from the Faroese birth cohort study, which has been extensively studied for
associations between neurodevelopmental deficits in children exposed to various chemicals,
including PFAS. In cord blood samples from males, PFOS exposure was significantly associated
with 10,598 methylation changes in CpG sites, 15% of which were enriched in cytobands of the
X chromosome associated with neurological disorders. Other CpG sites were associated with
genes in pathways of key physiological functions and diseases, including nervous system
development, tissue morphology, digestive system development, embryonic development,
3-248
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
endocrine system development, cancer, eye disease, organ abnormalities, cardiovascular disease,
and connective tissue disorders. The same effects were not observed in cord blood from females.
3.4.4.3.9 Conclusion
The available mechanistic studies suggest that the developing liver, developing heart, and
placenta may be affected by PFOS at the molecular level (i.e., differential methylation of genes,
gene expression changes, mitochondrial dysregulation), which may be related to developmental
health effects described in Sections 3.4.4.1 and 3.4.4.2. Some effects tend to vary by sex or by
developmental timepoint of outcome evaluation (e.g., early gastrulation, late gestation, lactation).
Oxidative stress in parallel with epigenetic alterations in the placenta were consistently reported.
3.4.4.4 Evidence Integration
The evidence of an association between PFOS and developmental effects in humans is moderate
based on the recent epidemiological literature. As noted in the epidemiological fetal growth
restriction summary, there is evidence that PFOS may impact fetal growth restriction in humans.
Several meta-analyses also support evidence of associations between maternal or cord blood
serum PFOS and BWT or BWT-related measures {Verner, 2015, 3150627; Negri, 2017,
3981320; Dzierlenga, 2020, 7643488; Cao, 2021, 9959525; Yang, 2022, 10176603} (Table A-
41, PFOS Appendix A). Comparing the postnatal growth results in infants with birth-related
measures is challenging due to complex growth dynamics including rapid growth catch-up
periods for those with fetal restriction. Nonetheless, the evidence for postnatal weight deficits
was comparable to that seen for BWT. Overall, there was inconsistent evidence of PFOS impacts
on rapid growth measures, postnatal height and postnatal adiposity measures up to age 2. There
was less evidence available in recent studies of PFOS exposure for other endpoints such as fetal
loss and birth defects. The evidence for an association between PFOS exposure and
cryptorchidism or hypospadias were primarily negative but overall inconsistent. In contrast, there
was fairly consistent evidence of an impact of PFOS exposure on gestational duration measures
(i.e., either preterm birth or gestational age measures) as the majority of studies showed some
adverse associations. Several meta-analyses also show associations between PFOS and preterm
birth {Deji, 2021, 7564388; Gao, 2021, 99596011; Yang, 2022, 10176603} (Table A-41, PFOS
Appendix A).
As noted previously there is some uncertainty as to what degree the available evidence may be
impacted by pregnancy hemodynamic and sample timing differences across studies, as this may
result in either confounding or reverse causality {Steenland, 2018, 5079861}. Additional
uncertainty exists due to the potential for confounding by other PFAS. Very few of the existing
studies performed multipollutant modeling in comparison with single pollutant estimates of
PFOS associations. The results were often mixed from those that did this with some estimates
increasing and some decreasing although PFOS was rarely chosen amongst dimension-reducing
statistical approaches from models with various PFAS and or other environmental contaminants.
There is some concern that controlling for other highly correlated co-exposures in the same
model may amplify the potential confounding bias of another co-exposure rather than removing
it {Weisskopf, 2018, 7325521}. Given these interpretation difficulties and potential for this co-
exposure amplification bias, it remains unclear whether certain mutually adjusted models give a
more accurate representation of the independent effect of specific pollutants for complex PFAS
mixture scenarios.
3-249
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
The animal evidence for an association between PFOS exposure and developmental toxicity is
moderate based on 16 medium confidence animal toxicological studies. Dose-dependent
maternal and offspring effects were reported in mice, rats, and rabbits; however, a few studies
did not observe effects. The studies evaluated demonstrate that PFOS exposure is associated with
various developmental toxicity endpoints including increased mortality (pup mortality, fetal
death, stillbirth, abortion), decreased body weight or body weight change (fetal, pup, and
maternal), skeletal and soft tissue effects, and delayed eye opening. The most consistent effects
observed across studies were decreased maternal body weight (encompassing decreases in
maternal body weight and maternal body weight change), decreased offspring weight during the
perinatal developmental period (encompassing fetal weight and pup weight prior to weaning),
and increased mortality (encompassing abortion, stillbirth, fetal death, and pup mortality).
Reductions in litter size or fetal/pup weight may be the driver of reductions seen in maternal
weight. For all but one study, decreased maternal weight was observed at the same doses as the
potential confounding effects of reduced fetal weight, increased incidence of abortion, increased
pup mortality/stillbirth, and others. However, Argus Research Laboratories (2000, 5080012)
reported reduced maternal body weight change in the absence of statistically significant effects
on pups that could influence maternal weight. In this case, maternal body weight may be an
influential precursor to or sensitive indicator of potential offspring mortality.
Similarly, Luebker et al. (2005, 757857; 2005, 1276160) observed decreased pup weights as an
average per litter at lower dose levels than effects on viability endpoints including decreases in
implantations, increased number of dams with all pups dying, and decreased number of live pups
per litter. These results are supported by Lau et al. (2003, 757854) who found significant
decreases in rat pup body weight at birth and increases in pup mortality in the first 24-48 hours
after birth. Significant reductions in both endpoints occurred at the same dose of 2 mg/kg/day. A
final study {Lee, 2015, 2851075} also observed increased fetal death and decreased fetal weight.
However, in this study, increased incidence of fetal death was statistically significant at all dose
levels whereas fetal weight was not affected at the lowest dose of 0.5 mg/kg/day.
The mechanistic data are primarily focused on gene expression changes and epigenetic
alterations related to exposure to PFOS during developmental stages. PFOS-induced alterations
to the expression of genes related to growth and development supports the observations in
animals and humans (e.g., fetal growth restriction). Molecular alterations (primarily epigenetic
alterations) were also measured in human cord blood and were related to PFOS levels in the
same biological samples. Global hypomethylation, a marker of genomic instability, was
associated with PFOS exposure, as was hypermethylation of genes related to xenobiotic
metabolism. Another study in human cord blood reported changes in DNA methylation at
genomic sites associated with genes related to normal development of several tissue and organ
systems (e.g., nervous system development and endocrine system development, among others).
However, the authors of these studies did not measure gene expression changes to confirm the
epigenetic alterations affected the transcriptome, nor did the authors report any adverse post-
natal effects to which to anchor the epigenetic alterations. In addition to human data, mechanistic
data related to developmental effects and PFOS have been collected in vivo in zebrafish and
rodent studies, as well as in human and rodent in vitro models. In zebrafish embryos exposed to
PFOS, changes in genes that are related to growth and development (e.g., growth factors, among
3-250
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
others) were observed along with growth inhibition, decreased hatch rate, embryonic
malformations, and other metrics of development, indicating that PFOS-induced effects on
growth and development are related to alterations to the transcriptome of developing zebrafish.
Alterations to individual genes or pathways that are also seen in tissues from adult animals in
laboratory studies (e.g., PPAR and markers of apoptosis in the liver, or cardiac-specific
pathways) were observed in developing animals and/or embryonic cell lines. Alterations to the
epigenome were observed in several animal toxicological studies, including in the placenta of
pregnant rodents exposed to PFOS. Such alterations occurred at the global and gene-specific
levels, indicating that epigenetic regulation of normal development can be altered by PFOS
exposure. Overall, there is robust evidence of an impact of PFOS exposure on gestational
duration measures (i.e., either preterm birth or gestational age measures) as most of the studies
showed some adverse associations. This was strengthened by consistency in the reported
magnitude of gestational age deficits despite different exposure levels and metrics examined.
Although they were not as consistent in magnitude (60% of the PTB studies showed some
adverse associations), some of the effect estimates were large for preterm birth in relation to
PFOS exposures with limited evidence of exposure-response relationships. Few patterns were
evident as explanatory factors for heterogeneous results based on our qualitative analysis.
3.4.4.4.1Evidence Integration Judgment
Overall, considering the available evidence from human, animal, and mechanistic studies, the
available human and animal evidence indicates that PFOS exposure is likely to cause
developmental toxicity in humans under relevant exposure circumstances (Table 3-12). This
conclusion is based primarily on evidence of decreased birth weight from epidemiologic studies
in which PFOS was measured during pregnancy, primarily with median PFOS ranging from 5.0
to 30.1 ng/mL. The conclusion is supported by coherent epidemiological evidence for
biologically related effects (e.g., decreased postnatal growth and birth length) and consistent
findings of dose-dependent decreases in fetal and maternal weight, with the effects observed in
animal models gestationally exposed to PFOS at doses as low as 0.4 mg/kg/day. The available
mechanistic information provides support for the biological plausibility of the phenotypic effects
observed in exposed animals and humans.
3-251
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 3-12. Evidence Profile Table for PFOS Developmental Effects
Evidence Stream Summary and Interpretation
Evidence Integration
Summary Judgment
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase
Certainty
Factors that Decrease
Certainty
Evidence Stream
Judgment
Evidence from Studies of Exposed Humans (Section 3.4.4.1)
Fetal growth
restriction
20 High confidence
studies
13 Medium confidence
studies
10 Low confidence
studies
Some deficits in mean
birth weight were
observed in most studies
(19/30) in the overall
population, but evidence
for the exposure-
response relationship
was limited. Most
evidence for deficits in
mean birth weight were
reported from high or
medium confidence
studies (16/27). Studies
on changes in
standardized birth weight
measures reported some
inverse associations
(9/15) in the overall
population or either/both
sexes. Seven of 11
studies observed
increased risk of low
birth weight or SGA.
Deficits in birth weight-
related measures were
supported by in related
FGR outcomes such as
birth length (11/23) and
head circumference
(10/19).
High and medium
confidence studies
Coherence between
different measures of
FGR
Good or adequate
sensitivity in most
studies
Limited evidence of
exposure-response
relationships based on
categorical data
Potential bias due to
hemodynamic
differences noted in
studies using samples
from later pregnancy
0©O
Moderate
Evidence for
developmental effects is
based on consistent
adverse effects for FGR
including birthweight
measures which are the
most accurate endpoint.
Some deficits were
consistently reported for
birth weight and
standardized birth weight
in many high and
medium confidence
cohort studies. Effects on
birth weight were
supported by findings for
other measures of FGR,
including birth length
and head circumference,
and impacts on
gestational duration.
Some uncertainty due to
the potential impact of
hemodynamics in later
pregnancy due to use of
biomonitoring samples
from the second and
®©o
" Evidence Indicates (likely)
Primary basis and cross-
stream coherence:
Evidence consisted of
decreased birth weight from
epidemiologic studies in
which PFOS was measured
during pregnancy. This is
supported by coherent
epidemiological evidence
for biologically related
effects (e.g., decreased
postnatal growth and birth
length) and consistent
findings of dose-dependent
decreases in fetal weight,
with the effects observed in
animal models gestationally
exposed to PFOS.
Human relevance and other
inferences:
The available mechanistic
information provides
support for the biological
plausibility of the
phenotypic effects observed
in exposed animals in
support of the human
relevance of the animal
findings.
3-252
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Evidence Integration
Summary Judgment
Gestational duration
10 High confidence
studies
5 Medium confidence
studies
5 Low confidence
studies
Fetal loss
2 High confidence
studies
2 Medium confidence
studies
1 Low confidence
study
Post-natal growth
4 High confidence
studies
4 Medium confidence
studies
3 Low confidence
studies
Some associations with
gestational age measures
in the overall population
were observed (9/15),
with most (7/9)
considered high or
medium confidence.
Increased risk of preterm
birth was also observed
in most studies (7/11).
Increased risk of fetal
loss was observed (3/5).
One medium confidence
study reported an inverse
association.
Most studies (7/8)
reported an adverse
association for infant
weight changes. There
was some evidence of a
dose-response
relationship in two
studies (2/4) reporting
categorical exposures.
Decreases in infant
height were only
observed in two studies
(2/4). Results for BMI
and adiposity were
mixed with studies
reporting both increased
(2/6) and decreased (3/6)
High and medium
confidence studies
Consistency in the
magnitude of
gestational age
deficits
High and medium
confidence studies
Good sensitivity
across all studies
Consistent magnitude
of effect
Dose-dependent
response
High and medium
confidence studies
Dose-dependent
response
Good or adequate
sensitivity for most
studies
third trimester or post-
partum.
Limited number of
studies examining
preterm birth
No factors noted
Lnconsistent timing of
follow-up evaluation
3-253
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Evidence Integration
Summary Judgment
Birth defects
3 Medium confidence
studies
2 Low confidence
studies
risk for adverse adiposity
outcomes.
One low confidence
study observed a small
increased risk for total or
combined birth defects.
One medium confidence
study reported increased
risk for septal defects,
conotruncal defects, and
total congenital heart
defects, but results were
imprecise.
Cryptorchidism was
examined in three
studies. Of the two
medium confidence
studies, one reported a
non-significant inverse
association and the other
reported a null
association.
Medium confidence
studies
Low confidence studies
Lmprecision of some
positive associations
may suggest statistical
power was limited
Limited number of
studies examining
individual defects
Evidence from In Vivo Animal Studies (Section 3.4.4.2)
Maternal body
weight
12 Medium confidence
studies
Maternal body weight
and/or body weight gain
during gestation and
lactation were dose-
dependently reduced in
several studies in rats,
mice, and rabbits (8/12).
Remaining studies (4/12)
in mice found no effects
on maternal body
weight
Medium confidence
studies
Dose-response
relationship
Lnconsistent direction of
effects
0©O
Moderate
Evidence based on 16
high or medium
confidence animal
studies indicates that the
developing fetus is a
target of PFOS toxicity.
Dose-dependent maternal
and offspring effects
were reported in mice,
rats, and rabbits;
however, a few studies
3-254
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Evidence Integration
Summary Judgment
Offspring body
weight
15 Medium confidence
studies
Offspring mortality
11 Medium confidence
studies
Fetal body weights were
dose-dependently
reduced (4/8) in studies
in rats, mice, and rabbits.
Pup birth weights and/or
body weights during
lactation were dose-
dependently reduced
(4/9), with significant
effects observed in rats
but not mice.
Increased fetal mortality
(2/7) was reported in
rats, mice, and rabbits
that evaluated endpoints
such as abortion,
implantation, resorption,
and dead/live fetus
counts prior to
parturition. Two studies
exposed female rats prior
Medium confidence
studies
Dose-dependent
response
Medium confidence
studies
Consistent direction of
effects
Dose-dependent
response
No factors noted
did not observe effects.
The studies evaluated
demonstrate that PFOS
exposure is associated
with various
developmental toxicity
endpoints including
increased mortality (pup
mortality, fetal death,
stillbirth, abortion),
decreased body weight or
body weight change
(fetal, pup, and
maternal), skeletal and
soft tissue effects, and
delayed eye opening.
Inconsistent direction of
effects across species for
postnatal body weight
3-255
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Placental effects
6 Medium confidence
studies
to mating through
lactation, and the study
with higher doses
observed decreased
number of implantation
sites per delivered litter
and liveborn litter size,
and increased number of
stillborn pups per litter
(1/2). Four studies began
exposure during
gestation and allowed
natural delivery of litters,
and only one (1/4)
observed decreased
liveborn litter size. No
studies reported an effect
on sex ratio (percentage
of male pups delivered
per litter) (0/6). Postnatal
survival was dose-
dependently decreased in
several studies in mice
and rats (5/8). For the
two studies with
exposure prior to mating
through lactation, both
reported decreased pup
viability index and
increased numbers of
dams with all pups dying
in the first 4-5 days
postpartum.
Decreased placental
weight (2/3), decreased
placental diameter (1/1),
and decreased placental
Medium confidence
studies
Dose-response
relationship
Inconsistent direction of
effects
Limited number of studies
examining outcomes
Evidence Integration
Summary Judgment
3-256
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
Evidence Integration
Summary Judgment
Structural
abnormalities
2 Medium confidence
studies
Developmental timing
and organ
maturation
4 Medium confidence
studies
capacity (1/1) were
observed in rat and
mouse studies, but two
other studies in rats and
rabbits reported normal
placental size and
appearance.
Histopathology was
evaluated in two mouse
studies; one study
observed no changes in
the placenta while the
other study observed
necrotic changes and
dose-dependent
decreases in
trophoblasts.
No external or visceral
abnormalities were
detected in mouse or
rabbit fetuses (2/2).
Lower incidence of
diminished calcaneus
ossification was
observed in mice (1/1)
and delayed skeletal
ossification was
observed in rabbits
(1/1).
Delayed eye-opening
(2/3) was reported in rats
and mice following
gestational PFOS
exposure. In a two-
generation study in rats,
delayed pinna unfolding,
air righting, and surface
Coherence of findings
Medium confidence
studies
Limited number of studies
examining outcomes
Medium confidence
studies
Coherence of effects
with other
developmental delays
Limited number of studies
examining outcomes
3-257
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Evidence Stream Summary and Interpretation
righting was also
observed (1/1). In
contrast, eye opening in
mice exposed from PND
1-14 was unaffected (pup
body weight was also
unaffected in that study).
In general, the studies
that observed
developmental delays
also reported growth
deficits and decreased
viability during the
lactation period.
PFOS exposure from GD
12-18 affected lung
development and
maturation in rats when
observed on PND 1-14
(1/1).
Mechanistic Evidence and Supplemental Information (Section 3.4.4.3)
Summary of Key Findings, Interpretation, and Limitations
Evidence Stream
Judgement
Key findings and interpretation:
Evidence from zebrafish embryo assays demonstrate that PFOS exposure can lead to embryo and/or
larva malformation and delays/reduction in hatching.
Alterations to the expression of genes related to growth and development in vivo in zebrafish and
rodents, and in human embryonic cell lines.
Alterations to DNA methylation (global hypomethylation and gene-specific hypermethylation) in
human cord blood and in placenta from rodent studies.
Limitations:
The role of epigenetic mechanisms in changes at the mRNA level is not clear, nor is the relationship
between molecular changes and apical developmental outcomes.
The evidence
demonstrates that PFOS
exposure during
development can alter
the epigenome and the
expression of genes that
control regular growth
and development; it is
possible that such
changes are related,
although the relationship
has not been directly
measured.
Evidence Integration
Summary Judgment
3-258
-------
DRAFT FOR PUBLIC COMMENT MARCH 2023
Notes: SGA = small-for-gestational age; FGR = fetal growth restriction; PND = postnatal day; GD = gestational day; BMI = body mass index; DNA = deoxynucleic acid; mRNA =
messenger ribonucleic acid.
3-259
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.4.5 Evidence Synthesis and Integration for Other Non-
Cancer Health Outcomes
Consistent with the SAB's recommendation, EPA concluded that the non-cancer health
outcomes with the strongest evidence are hepatic, immune, cardiovascular and developmental.
For all other health outcomes (e.g., reproductive and endocrine), EPA concluded that the
epidemiological and animal toxicological evidence available at this time from the preliminary
scoping considered in the Proposed Approaches to the Derivation of a Draft Maximum
Contaminant Level Goal for Perfluorooctane Sulfonate (PFOS) (CASRN1763-23-1) in Drinking
Water is either suggestive of associations or inadequate to determine associations between PFOS
and the health effects described. Based on this analysis, these outcomes were not prioritized for
the MCLG assessment and the evidence synthesis and integration for these other outcomes are
presented in the PFOS Appendix. In addition, Section 6.5 further describes rationale for evidence
integration judgments for health outcomes which EPA determined had evidence suggestive of
associations between PFOS and related adverse health effects, though the databases for those
health outcomes shared some characteristics with the evidence indicates judgment.
3.5 Cancer Evidence Study Quality Evaluation, Synthesis, Mode
of Action Analysis and Weight of Evidence
EPA identified 15 epidemiological and 1 animal toxicological study that investigated the
association between PFOS and cancer. Of the epidemiological studies, 8 were classified as
medium confidence, 6 as low confidence, and 1 was considered uninformative (Section 3.5.1).
The single animal toxicological study was considered a high confidence study (Section 3.5.2).
Studies have mixed confidence ratings if different endpoints evaluated within the study were
assigned different confidence ratings. Though low confidence studies are considered
qualitatively in this section, they were not considered quantitatively for the dose-response
assessment (Section 4).
3.5.1 Human Evidence Study Quality Evaluation and
Synthesis
3.5.1.1 Introduction
There are 7 studies from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} that investigated the
association between PFOS and cancer effects. Study quality evaluations for these 7 studies are
shown in Figure 3-72.
The 2016 Health Effects Support Document for PFOS {U.S. EPA, 2016, 3603365} concluded
that there was no evidence of carcinogenic effects for PFOS, but that the small number, breadth,
and scope of the studies were not adequate to make definitive conclusions. Although an elevated
risk of bladder cancer mortality was observed in an occupational study of workers at the 3M
Decatur, Alabama plant {Alexander, 2003, 1291101}, a subsequent study to ascertain cancer
incidence in the cohort observed elevated but non-significant incidence ratios that were 1.7- to 2-
fold higher among exposed workers {Alexander, 2007, 4727072}. Mean PFOS serum levels
were 94.1 ng/mL. In the same 3M cohort, {Grice, 2007, 4930271} observed that prostate,
melanoma, and colon cancer were the most frequently reported malignancies. When cumulative
exposure measures were analyzed, elevated odds ratios were reported for melanoma, colon, and
3-260
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
prostate cancer, however, they did not reach statistical significance. Length of follow-up may not
have been adequate to detect cancer incidence in this cohort as approximately one-third of the
participants had worked < 5 years in their jobs, and only 41.7% were employed > 20 years.
No elevated bladder cancer risk was observed in a nested case-control study in a Danish cohort
with plasma PFOS concentrations at enrollment ranging 1-130.5 ng/mL {Eriksen, 2009,
2919344}. Elevated non-significant ORs for prostate cancer were reported for the occupational
cohort examined by Alexander and Olsen (2007, 4727072) and the Danish population-based
cohort examined by Eriksen et al. (2009, 2919344), and no association was reported by another
case-control study in Denmark {Hardell, 2014, 2968084}. A case-control study of breast cancer
among Inuit females in Greenland with similar serum PFOS levels to those of the Danish
population (1.5-172 ng/mL) reported an association of low magnitude that could not be
separated from other perfluorosulfonated acids, and the association was not confirmed in a
Danish population {Bonefeld-J0rgensen, 2011, 2150988; Bonefeld-J0rgensen, 2014, 2851186}.
Some studies evaluated associations with serum PFOS concentration at the time of cancer
diagnosis and the impact of this potential exposure misclassification on the estimated risks is
unknown {Bonefeld-J0rgensen, 2011, 2150988; Hardell, 2014, 2968084}. No associations were
adjusted for other perfluorinated chemicals in serum in any of the occupational and population-
based studies.
3-261
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
<\
,0®
Alexander and Olsen, 2007, 4727072 -
Alexander et al., 2003, 1291101
Bonefeld-Jorgensen et al., 2011, 2150988
Bonefeld-Jorgensen etal., 2014, 2851186-
Eriksen et al„ 2009, 2919344
Grice et al., 2007, 4930271 -
Hardell etal., 2014, 2968084
¦
++
B
++
++
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-72. Summary of Study Evaluation for Pre-2016 Epidemiology Studies of PFOS
and Cancer Effects
Interactive figure and additional study details available on HAWC.
Since publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}, 15 studies have been
published that investigated the association between PFOS and cancer (see PFOS Appendix). All
studies were conducted on the general population with one in a high-exposure community (i.e.,
C8 population). Different study designs were also used including two cohort studies {Fry, 2017,
4181820; Li, 2022, 9961926}, five case-control studies {Wielsoe, 2017, 3858479; Tsai, 2020,
6833693; Lin, 2020, 6835434; Itoh, 2021, 9959632; Liu, 2021, 10176563}, five nested case-
control studies {Ghisari, 2017, 3860243; Hurley, 2018, 5080646; Cohn, 2020, 5412451;
Mancini, 2020, 5381529; Shearer, 2021, 7161466}, and three cross-sectional studies
{Christensen, 2016, 3858533; Ducatman, 2015, 3859843; Omoike, 2021, 7021502}. The studies
were conducted in different study populations including populations from China {Lin, 2020,
6835434; Liu, 2021, 10176563}, Denmark {Ghisari, 2017, 3860243}, France {Mancini, 2020,
5381529}, Greenland {Wielsoe, 2017, 3858479}, Japan {Itoh, 2021, 9959632}, Sweden {Li,
2022, 9961926}, Taiwan {Tsai, 2020, 6833693}, and the United States {Fry, 2017, 4181820;
Christensen, 2016, 3858533; Ducatman, 2015, 3859843; Shearer, 2021, 7161466; Hurley, 2018,
5080646; Cohn, 2020, 5412451; Omoike, 2021, 7021502}. All the studies measured PFOS in
3-262
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
study subject's blood components (i.e., serum or plasma) with one study measuring the levels in
the maternal serum {Cohn, 2020, 5412451}. Cancers evaluated included breast {Cohn, 2020,
5412451; Ghisari, 2017, 3860243; Hurley, 2018, 5080646; Itoh, 2021, 9959632; Li, 2022,
9961926; Mancini, 2020, 5381529; Omoike, 2021, 7021502; Tsai, 2020, 6833693; Wielsoe,
2017, 3858479}, germ cell tumors {Lin, 2020, 6835434}, kidney {Shearer, 2021, 7161466},
melanoma {Li, 2022, 9961926}, ovarian {Omoike, 2021, 7021502}, prostate {Ducatman, 2015,
3859843; Omoike, 2021, 7021502}, thyroid {Liu, 2021, 10176563} uterine {Omoike, 2021,
7021502}, and any cancer {Christensen, 2016, 3858533; Fry, 2017, 4181820; Li, 2022,
9961926}.
3.5.1.2 Study Quality
There are 15 studies from recent systematic literature search and review efforts conducted after
publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} that investigated the
association between PFOS and cancer effects. Study quality evaluations for these 15 studies are
shown in Figure 3-73.
Of the 15 studies identified since the 2016 assessment (Figure 3-73), seven were considered
medium confidence and six were low confidence {Christensen, 2016, 3858533; Itoh, 2021,
9959632; Lin, 2020, 6835434; Liu, 2021, 10176563; Omoike, 2021, 7021502; Tsai, 2020,
6833693}. One study conducted in the high exposure to PFAS Ronneby Register Cohort in
Sweden was uninformative {Li, 2022, 9961926} because of concerns about exposure assessment
and lack of data on important covariates. One study conducted in Greenland was considered
uninformative { Wielsoe, 2017, 3858479} because of concerns about exposure assessment and
participant selection, As a result, these two studies will not be further considered in this review.
Concerns with the low confidence studies included the possibility of outcome misclassification,
confounding or potential selection bias. Residual confounding was also a concern, including lack
of considering co-exposures by other PFAS, and lack of appropriately addressing SES (SES) and
other life-style factors, which could be associated with both exposure and cancer outcome.
Although PFOS has a long half-life in the blood, concurrent measurements may not be
appropriate for cancers with long latencies. Temporality of exposure measure in terms of cancer
development wase noted to be an issue in several low confidence studies {Tsai, 2020, 6833693;
Itoh, 2021, 9959632; Liu, 2021, 10176563; Omoike, 2021, 7021502}. Many of the low
confidence studies also had sensitivity issues due to limited sample size.
3-263
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
\e^
Vs4V&^> Va'
&
Christenseri et al., 2016, 3858533 -
Cohnet al., 2020, 5412451 -
Ducatman et al., 2015, 3859843
Fry et al., 2017, 4181820
Ghisari et al., 2017, 3860243 -
Hurley et al., 2018, 5080646 -
Itoh et al., 2021, 9959632-
Li et al., 2022, 9961926^
Lin et al., 2020, 6835434-
Liu et al., 2021, 10176563
Mancini et al., 2019, 5381529
Omoike et al., 2020, 7021502
Shearer etal., 2021, 7161466
Tsai et al., 2020, 6833693-
Wielsoe et al., 2017, 3858479 -J
I Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-73. Summary of Study Evaluation for Epidemiology Studies of PFOS and Cancer
Effects
Interactive figure and additional study details available on HAWC.
3.5.1.3 Findings from Children
One low confidence study examined cancers in children {Lin, 2020, 6835434} and reported a
statistically significant higher median PFOS concentration in 42 pediatric germ cell tumor cases
compared with 42 controls in blood samples collected from the children one week after
diagnosis. However, the study did not observe an increased risk of germ cell tumors when
evaluated on a per ng/mL increase in blood PFOS.
3.5.1.4 Findings from the General Adult Population
PFOS was associated with an increased risk of kidney cancer (i.e., renal cell carcinoma) in a
medium confidence study {Shearer, 2021, 7161466}. A case-control study nested within the
3-264
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
National Cancer Institute's (NCI) Prostate, Lung, Colorectal, and Ovarian Screening Trial,
reported a statistically significant positive trend in risk of renal cell carcinoma with pre-
diagnostic serum levels of PFOS (OR = 2.51; 95% CI: 1.28, 4.92 for the highest vs. lowest
quartiles; p-trend = 0.009, or per doubling of PFOS: OR: 1.39; 95% CI: 1.04, 1.86) {Shearer,
2021, 7161466}. Although the trend was significant across quartiles, the effect in the third
quartile was null (OR = 0.92; 95% CI: 0.45, 1.88). Additionally, the association with PFOS was
attenuated after adjusting for other PFAS (OR = 1.14; 95% CI: 0.45, 2.88 for the highest us.
lowest quartiles; p-trend = 0.64), and it was lower in the third quartile than in the second quartile,
indicating potential confounding by correlated PFAS exposures. There was no association when
evaluated on a per doubling of PFOS after adjusting for other PFAS.
Seven general population studies published since the 2016 assessment, evaluated PFOS and risk
for breast cancer {Cohn, 2020, 5412451; Ghisari, 2017, 3860243; Hurley, 2018, 5080646; Itoh,
2021, 9959632; Mancini, 2020, 5381529; Omoike, 2021, 7021502; Tsai, 2020, 6833693} with
mixed results. All studies were case-control studies (with some nested case-controls), except for
one cross-sectional NHANES-based study {Omoike, 2021, 7021502}. Three studies were
considered low confidence {Itoh, 2021, 9959632; Omoike, 2021, 7021502; Tsai, 2020, 6833693}
because of concerns about temporality of exposure measurements and breast cancer
development, the control status was not confirmed via examination or medical records {Tsai,
2020, 6833693}, and potential for residual confounding due to SES, life-style factors and other
PFAS. The remaining studies were all medium confidence. A nested case-control study did not
observe an association between breast cancer identified through California cancer registry and
PFOS concentrations in serum after case diagnosis, max PFOS concentration of 99.8 ng/mL
{Hurley, 2018, 5080646}. A nested case-control study in a prospective (pregnancy) cohort study,
the CHDS, suggested that maternal PFOS was associated with a decreased daughters' breast
cancers risk in the first or fourth quartile of TC {Cohn, 2020, 5412451}, but the study did not
examine breast cancer subtypes or genetic variants. Two nested case-control studies and one low
confidence case-control study found associations between PFOS and breast cancer, but only in
specific groups of participants {Ghisari, 2017, 3860243; Mancini, 2020, 5381529; Tsai, 2020,
6833693}. Ghisari et al. (2017, 3860243) reported an increased risk for breast cancer identified
from the cancer registry with increasing PFOS concentrations only in participants with a CC
genotype (n = 36 cases and 47 controls) in the CYP19 gene (cytochrome P450 aromatase). A
nested case-control study (194 pairs of breast cancer cases and controls) within the French E3N
cohort found an 86% higher risk of breast cancer in the 2nd and 3rd quartiles of PFOS (13.6-
17.3 ng/mL, and 17.3-22.5 ng/mL) compared to the 1st quartile (5.8-13.6 ng/mL) (OR = 1.94;
95% CI: 1.00, 3.78, and OR = 2.03; 95% CI: 1.02, 4.04) in the full adjusted model {Mancini,
2020, 5381529}. Mancini et al. (2020, 5381529) reported that the risk for breast cancer (93%
verified pathologically confirmed from medical records after self-reported cancer diagnosis)
varied by type of cancer with a statistically significant increasing trend in estrogen receptor
positive (ER+) and progesterone receptor positive (PR+) breast cancers. The study also observed
a significant increase in estrogen receptor- (ER-) and progesterone receptor- (PR-) breast cancers
in the second quartile with elevated risks also observed in the other quartiles, but with no trend.
The sample size was small with 26 participants having ER- breast cancers and 57 having PR-
breast cancers.
One low confidence study {Tsai, 2020, 6833693} conducted in Taiwan observed a statistically
significant increase in risk of breast cancer with increasing log transformed PFOS, but only in
3-265
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
participants aged 50 years or younger and in ER+ breast cancer in participants aged 50 years or
younger. Statistically significant increased odds of breast cancer were also observed in a low
confidence NHANES study (2005-2012) {Omoike, 2021, 7021502} both per ng/mL increase in
PFOS (OR = 1.011; 95% CI: 1.011, 1.011) and in the two highest quartiles of exposure. The
association was significantly inverse in the second quartile compared to the lowest (OR = 0.87;
95% CI: 0.86, 0.89). One low confidence case-control study conducted in Japanese women {Itoh,
2021, 9959632} observed a significant inverse association across serum PFOS quartiles with a
significant dose-response trend (p-value < 0.0001) (see PFOS Appendix). Median PFOS levels
ranged from 7.6 ng/mL in the lowest quartile to 24.67 ng/mL in the highest quartile. The
association remained significantly inverse in both pre- and postmenopausal women in the highest
tertile of exposure, with a significant dose-response trend (p-values for trend = 0.007 and 0.001,
respectively).
One medium confidence study based on the C8 Health Project {Ducatman, 2015, 3859843}
examined prostate-specific antigen (PSA) as a biomarker for prostate cancer in adult males over
age 20 years who lived, worked, or went to school in one of the six water districts contaminated
by the DuPont Washington Works facility. No association was observed between PSA levels in
either younger (i.e., aged 20-49 years) or older (i.e., aged 50-69 years) men and concurrent
mean serum PFOS concentrations up to 25 ng/mL. In an NHANES population, Omoike et al.
{2021, 7021502} observed a significantly inverse association with prostate cancer (OR = 0.994;
95% CI: 0.994, 0.994).
Omoike et al. (2021, 7021502) also observed statistically significant increased odds of ovarian
cancer both per ng/mL increase in PFOS (OR = 1.012; 95% CI: 1.012, 1.013) and in the two
highest quartiles of exposure, although the association was significantly inverse for the second
quartile of PFOS exposure (see PFOS Appendix). A significant inverse association also was
observed for uterine cancer (OR = 0.945; 95% CI: 0.944, 0.945 per ng/mL increase in PFOS)
{Omoike, 2021, 7021502}.
One low confidence study conducted in Shandong Province, in eastern China {Liu, 2021,
10176563} observed a statistically significant inverse association with thyroid cancer across
quartiles of serum PFOS (p-value for trend = 0.001). The median serum PFOS levels were higher
in controls than in cases (7.5 vs. 5.5 ng/mL, p-value < 0.001). However, there is some concern
about possible reverse causality. The ability to metabolize PFAS could change when the thyroid
becomes cancerous, thereby changing the PFAS concentrations. The abnormality of thyroid
hormones may also disturb the PFAS levels.
Two studies examined all cancers together, but collected different information on cancer (i.e.,
incidence verses mortality) and obtained the information using different methods. Cancer
mortality based on Public-use Linked Mortality Files was not associated with PFOS exposure in
a medium confidence study of participants over 60 years of age from NHANES, with median
PFOS concentration 4.3 ng/g lipid {Fry, 2017, 4181820}; PFOS also was not found to be
associated with self-reported cancer incidence in a low confidence study among male anglers
over 50 years, median PFOS concentration 19 |ig/L ({Christensen, 2016, 3858533}. Christensen,
2016, 3858533 was considered low confidence due to the potential of self-selection because
participants were recruited from flyers and other methods and filled out an online survey
including self-reported outcomes.
3-266
-------
DRAFT FOR PUBLIC COMMENT MARCH 2023
3.5.2 Animal Evidence Study Quality Evaluation and
Synthesis
There is one study from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} that investigated the
association between PFOS and cancer effects. Study quality evaluation for this one study is
shown in Figure 3-74.
cP°
Butenhoffetal., 2012, 1276144-
++ ++
NR
++
Legend
B
Good (metric) or High confidence (overall)
+
Adequate (metric) or Medium confidence (overall)
-
Deficient (metric) or Low confidence (overall)
E
Critically deficient (metric) or Uninformative (overall)
NR
Not reported
Figure 3-74. Summary of Study Evaluation for Toxicology Studies of PFOS and Cancer
Effects
Interactive figure and additional study details available on HAWC.
Table 3-13. Incidences3 of Hepatocellular and Pancreatic Tumors in Male and Female
Sprague-Dawley Rats as Reported by Thomford (2002, 5029075)
Treatment group
Sex
Tumor Type
0 ppm
0.5 ppm
2 ppm
5 ppm
20 ppm
Male
Hepatocellular
Adenomas
0/41 (0%)**
3/42 (7%)
3/47 (6%)
1/44 (2%)
7/43 (16%)**
Female
Hepatocellular
Adenomas
0/28 (0%)**
1/26 (4%)
1/15 (7%)
1/28 (4%)
5/31 (16%)*
Female
Hepatocellular
Carcinomas
0/28 (0%)
0/29 (0%)
0/16 (0%)
0/31 (0%)
1/32 (3%)
Female
Combined
Hepatocellular
Adenomas and
Carcinomas
0/28 (0%)**
1/29 (3%)
1/16 (6%)
1/31 (3%)
6/32 (19%)*
Male
Pancreatic Islet Cell
Adenomas
4/44 (9%)
3/45 (7%)
4/48 (8%)
4/46 (9%)
4/44 (9%)
Male
Pancreatic Islet Cell
Carcinomas
1/38 (3%)*
2/41 (5%)
2/44 (5%)
5/44(11%)
5/40 (13%)
Male
Combined
Pancreatic Islet Cell
Adenomas and
Carcinomas
5/44(11%)
5/45 (11%)
6/48 (13%)
8/46a (17%)
9/44 (20%)
3-267
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Notes:
* Statistically significant compared to the control group at p < 0.05. ** Statistically significant compared to the control group at
p < 0.01. Denoted significance for the control groups indicate statistically significant trends.
a Tumor incidence is expressed as the number of animals with tumors over the number of animals alive at the time of first
occurrence of the tumor.
In addition to hepatocellular tumors, Thomford et al. (2002, 5029075) reported increased
incidences of pancreatic islet cell carcinomas in males (Table 3-13). Though the slight increases
in the number of animals with carcinomas in the 5 and 20 ppm dose groups were not statistically
different from the control group, there was a statistically significant trend of increased incidence
with increased dose.
Thyroid and mammary gland tumors were also observed but did not exhibit linear dose-response
{Thomford, 2002, 5029075; Butenhoff, 2012, 1276144}. The most frequent thyroid tumor type
in females was C-cell adenomas, but the highest incidence was that for the controls and there
was a lack of dose-response among the exposed groups. There was also a high background
incidence in mammary gland tumors in the female rats, primarily combined fibroma adenoma
and adenoma, but the incidence lacked dose-response for all tumor classifications.
3.5.3 Mechanistic Evidence
Mechanistic evidence linking PFOS exposure to adverse cancer outcomes is discussed in Section
3.4.3 of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}. There are 26 studies from recent
systematic literature search and review efforts conducted after publication of the 2016 PFOS
HESD that investigated the mechanisms of action of PFOS that lead to cancer effects. A
summary of these studies is shown in Figure 3-75.
Evidence Stream
Animal Human In Vitro Grand Total
Figure 3-75. Summary of Mechanistic Studies of PFOS and Cancer Effects
Interactive figure and additional study details available on Tableau.
In 2016, ten key characteristics of carcinogens were selected by a multi-disciplinary working
group of the International Agency for Research on Cancer (IARC), based upon common
empirical observations of chemical and biological properties associated with human carcinogens
(i.e., Group 1 carcinogens as determined by IARC) {Smith, 2016, 3160486}. In contrast to the
"Hallmarks of cancer" as presented by Hanahan and Weinberg {Hanahan, 2022, 10164687;
Hanahan, 2011, 758924; Hanahan, 2000, 188413}, the key characteristics focus on the properties
of human carcinogens that induce cancer, not the phenotypic or genotypic traits of cancers. The
ten key characteristics provide a framework to systematically identify, organize, and summarize
mechanistic information for cancer hazard evaluations {Smith, 2016, 3160486}.
To aid in the evaluation of the carcinogenic potential of PFOS, the studies containing
mechanistic data were organized by the proposed key characteristics of carcinogens for the
following section. Evidence related to seven of the ten key characteristics of carcinogens was
identified in the literature included in this assessment: 'Is Genotoxic', 'Induces Epigenetic
Effects', 'Induces Oxidative Stress', 'Modulates Receptor Mediated Effects', 'Alters Cell
3-268
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Proliferation, Cell Death, and Nutrient Supply', 'Is Immunosuppressive', and 'Induced Chronic
Inflammation'. No studies from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365} and recent
systematic literature search and review efforts were identified for the following key
characteristics: 'Is Electrophilic or Can Be Metabolically Activated to Electrophiles', 'Alters
DNA Repair and Causes Genomic Instability', and 'Causes Immortalization'.
3.5.3.1 Key Characteristic #2: Is Genotoxic
Genotoxicity is a well-studied mode of action for carcinogens, defined as alterations to DNA
through single or double strand breaks, alterations to DNA synthesis, and DNA adducts, all of
which can result in chromosomal aberrations, formation of micronuclei, and mutagenesis if not
effectively repaired.
3.5.3.1.1 Mutagenicity
3.5.3.1.1.1 In Vivo Evidence
Male gpt delta transgenic mice, a strain that was designed to facilitate the quantification of point
mutations and deletions, were exposed to PFOS (4 and 10 mg/kg/day) for 28 days {Wang, 2015,
2850220}. The mutation frequencies at the targeted redBA and gam loci in the liver of exposed
male mice were increased at concentrations of 4 and 10 mg/kg/day relative to controls, but the
increase was not significant, and the variance of the high dose group was relatively large. The
evidence for mutagenicity of PFOS in vivo is negative based on this single study (Table 3-14).
3.5.3.1.1.2 In Vitro Evidence
Several studies have demonstrated that PFOS is not mutagenic in vitro (Table 3-15). Of the four
publications that tested PFOS for mutagenicity in Salmonella typhimurium, Saccharomyces
cerevisiae, and Escherichia coli {Litton Bionetics, Inc., 1979, 10228135; Mecchi, 1999,
10228133; Simmon, 1978, 10228131; NTP, 2019, 5400978}, no evidence of DNA mutagenesis
has been described in the presence or absence of metabolic activation. In contrast, Wang et al.
(2015, 2850220) exposed gpt delta transgenic mouse embryonic fibroblast cells to PFOS and
found concentration-dependent increases in mutation frequencies at the / lJBA/gam loci, a region
often used to determine point mutations and deletions.
3.5.3.1.2DNA Damage
3.5.3.1.2.1 In Vivo Evidence
Evaluations of PFOS exposure in rat, mouse, and zebrafish models were identified, which
predominantly demonstrated evidence of genotoxicity (Table 3-16). The majority of studies
presented data on potential micronuclei formation in bone marrow, peripheral blood, and/or the
liver, though some also reported different metrics of DNA damage. It is important to note that rat
models could be ineffective for determining micronucleus formation if study authors do not use
appropriate methodologies because the spleen will remove micronucleated cells {Schlegel, 1984,
10368697}. However, this would generally bias studies towards the null, not result in false
positives.
Particular methods such as flow cytometry can be used to effectively identify micronucleated
cells prior to splenic removal {Dertinger, 2004, 10328871}. For instance, NTP (2019, 5400978)
reported using flow cytometry to analyze micronuclei formation in immature polychromatic
erythrocytes from the peripheral blood of male and female Sprague Dawley rats treated with
3-269
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
0.312-5 mg/kg/day PFOS by gavage for 28 days. No effects on the number of micronucleated
polychromatic erythrocytes (PCEs) were observed in males, though there was a significant
increase in the number of PCEs in the 5 mg/kg/day females. Importantly, NTP (2019, 5400978)
noted that while there was a statistically significant trend for increasing micronucleated PCEs,
and that the response in the 5 mg/kg/day group was statistically significant compared to controls
indicating a positive test, the response was nonetheless within the range of historical control
levels. NTP (2019, 5400978) also reported that there were significant dose-dependent decreases
in the percentage of PCEs in the peripheral blood of both males and females, suggesting that
PFOS exposure may induce bone marrow toxicity.
Three other studies published by the same primary authors also reported the induction of
micronuclei formation in male or female Swiss Albino rats {(^elik, 2013, 2919161; Eke, 2016,
2850124; Eke, 2017, 3981318}. These studies used a valid fluorescence microscopy-based
technique, though they did not use the OECD recommended number of cells according to OECD
Guideline Test No. 474: Mammalian Erythrocyte Micronucleus Test (at least 500 erythrocytes
for bone marrow and at least 2,000 erythrocytes for peripheral blood) {Dertinger, 2004,
10328871}. Qelik et al. (2013, 2919161) found that oral treatment with PFOS (<2.5 mg/kg/day)
administered every other day for 30 days induced genetic damage as measured with the comet
assay, as well as the formation of micronuclei in female rat bone marrow samples. However, the
study also demonstrated that PFOS exposure decreased the ratio of PCEs to normochromic
erythrocytes (NCEs), similar to the results from NTP (2019, 5400978). These findings indicate
that the genetic damage may be a result of bone marrow toxicity rather than direct genotoxicity
of PFOS. Two subsequent studies in male rats using the same exposure paradigm (30-day
exposure administered every other day) found similar results. Eke and Qelik (2016, 2850124)
reported increased micronuclei formation and genetic damage indices (calculated using results of
a comet assay) in peripheral blood while Eke et al. (2017, 3981318) reported increased
micronuclei formation and genetic damage indices in liver tissue. Notably, these two studies did
not report the ratio of PCEs to NCEs which limits the ability to interpret these data further. Based
on the results from Qelik et al. (2013, 2919161) and considering the similarities in study design,
it is reasonable to assume that the genetic damage observed may be due to bone marrow or
hepatic toxicity.
Micronuclei frequency was higher in the bone marrow male gpt delta transgenic mice exposed to
PFOS (4 and 10 mg/kg/day) for 28 days than in controls; however, these results were not
statistically significant {Wang, 2015, 2850220}. This is a potential contradiction to previous
findings reported in the EPA's 2016 HESD {U.S. EPA, 2016, 3603365} that found mouse bone
marrow micronuclei assays to be negative after high dose acute exposures (approximately 24, 48,
and 72 hours) to PFOS {Murli, 1996, 10228098}.
In another study, male and female zebrafish embryos were exposed to PFOS concentrations of
0.4, 0.8, or 1.6 mg/L for 30 days {Du, 2014, 2851143}. Following exposure, Du et al. (2014,
2851143) found significant dose-dependent increases in micronucleus formation. Du et al. (2014,
2851143) also reported increases in the number of DNA single-strand breaks, though none of the
PFOS doses tested resulted in significant effects. Notably, the high dose exposure resulted in
increased rates of developmental malformations, which could potentially confound these results.
3-270
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Subchronic 28-day exposure of Sprague Dawley rats to PFOS did not alter micronuclei
formation in reticulocytes in exposed males, while data derived from exposed female rats was
equivocal {NTP, 2019, 5400978).
3.5.3.1.2.2 In Vitro Evidence
3.5.3.1.2.2.1 Chromosomal aberrations
EPA's 2016 HESD {U.S. EPA, 2016, 3603365} reports that PFOS exposure did not induce
chromosomal aberrations in human lymphocytes (Table 3-17) {Murli, 1999, 10228132}. No new
studies were identified that measure chromosomal aberrations after PFOS exposure in the
updated literature search.
3.5.3.1.2.2.2 DNA Synthesis
A study by Cifone (1999, 10228136) evaluated the effects of 15 different PFOS concentrations
ranging from 0.25 |ag/m L to 4,000 |ag/m L in Fisher 344 male rat hepatocytes. No evidence of
increased DNA synthesis was observed, denoted by the lack of elevated mean net nuclear grains.
Cytotoxicity significantly increased at approximately 50 |ig/mL.
An additional study, detailed elsewhere, noted increased DNA synthesis (increased cells in S
phase) following exposure in rodent hepatocytes. For additional information, please see the
hepatic mechanistic section (Section 3.4.1.3; refer to the interactive Tableau for additional
supporting information and study details).
3.5.3.1.2.2.3 DNA Damage
Several assays of DNA damage have been performed on a variety of in vitro models (Table
3-17). Wang et al. (2015, 2850220) exposed gpt delta transgenic mouse embryonic fibroblasts to
PFOS and found evidence of concentration-dependent increase in phosphorylated histone H2AX
(y-H2AX), a biomarker of DNA double strand breaks (DSBs), after exposure to 1 or 20 |iM
PFOS (no statistical analysis was reported). Direct exposure of suspended calf thymus DNA to
10 |iM PFOS for 30 minutes modified DNA structure, attenuated DNA charge transport, and led
to PFOS-DNA adduct formation {Lu, 2012, 2919198}.
In contrast, several studies found no evidence of DNA damage after exposure. Jacquet et al.
(2012, 2919219) exposed Syrian hamster embryos to PFOS (<50 |ig/mL) and found no evidence
of DNA damage by a comet assay. Similarly, there was no evidence of DNA damage via a comet
assay in the protist species Paramecium caudatum exposed to 10-100 |iM for 24 hours
{Kawamoto, 2010, 1274162}.
Florentin et al. (2011, 2919235) exposed HepG2 cells to PFOS (5-300 |iM) for 1 or 24 hours.
There was no evidence of DNA damage in a comet assay nor change in micronucleus frequency
at any concentration or time point. However, within the 24-hour exposure assay, significant
cytotoxic effects were noted at 300 |iM. In contrast, a study conducted by Wielsoe et al. (2015,
2533367) exposed HepG2 cells to PFOS (2 x 10"7 to 2 x 10"5 M) for 24 hours and used a comet
assay to measure DNA damage. Following exposure, the cells demonstrated a dose-dependent
increase in DNA damage at all tested concentrations.
3-271
-------
DRAFT FOR PUBLIC COMMENT MARCH 2023
Table 3-14. Mutagenicity Data from In Vivo Studies
Reference
Species, Strain
(Sex)
Tissue
Results
PFOS Concentration
(Dosing Regimen)
Wangetal. (2015,
Mouse, Gpt delta
Liver
Negative
1-10 mg/kg/day
2850220)
transgenic
(daily via gavage for 28 days)
(Male)
3-272
-------
DRAFT FOR PUBLIC COMMENT MARCH 2023
Table 3-15. Mutagenicity Data from In Vitro Studies
Reference
Cell Line or Bacterial Strain
Results
Concentration
(Duration of exposure)
S9-Activated
Non-Activated
Litton Bionetics, Inc. (1979,
10228135)
Salmonella typhimurium (TA1535, TA1537,
TA1538, TA98, TA100)
Negative
Negative
0.1 - 1,000 ng/plate
Litton Bionetics, Inc. (1979,
10228135)
Saccharomyces cerevisiae (D4)
Not Reported
Negative
0.1 - 1,000 ng/plate
Mecchi (1999, 10228133)
Salmonella typhimurium (TA98, TA100, TA1535,
TA1537)
Negative
Negative
0.333 - 5,000 ng/plate
Mecchi (1999, 10228133)
Escherichia coli (WPluvrA)
Negative
Negative
33.3 - 5,000 ng/plate
NTP (2019, 5400978)
Salmonella typhimurium (TA98, TA100)
Negative
Negative
100 - 5,000 ng/plate
NTP (2019, 5400978)
Escherichia coli (WP2MvrA/pkM101)
Negative
Negative
100 - 10,000 ng/plate
Simmon (1978, 10228131)
Salmonella typhimurium (TA1535, TA1537,
TA1538, TA98, TA100)
Negative
Negative
10 - 5,000 ng/plate
Simmon (1978, 10228131)
Salmonella cerevisiae (D3)
Negative
Negative
0.1-5 ng/plate
Wang et al. (2015, 2850220)
gpt Delta transgenic mouse embryonic fibroblasts
Not reported
Positive3
1 - 20 nM
(24 hours)
a Mutagens were Present in cells exposed >10 (iM.
3-273
-------
DRAFT FOR PUBLIC COMMENT MARCH 2023
Table 3-16. DNA Damage Data from In Vivo Studies
Reference
Species, Strain
(Sex)
Tissue
Results
PFOS Concentration
(Dosing Regimen)
DNA Damage via Comet Assay
Celiketal. (2013,2919161)
Rat, Swiss Albino
(Female)
Bone marrow
Positive
0.6 - 2.5 mg/kg/day
(every other day via gavage for 30 days)
Du etal. (2014, 2851143)
Zebrafish, AB
(Male and female)
Peripheral blood cells
Negative
0.4 - 1.6 mg/L
(single dose to rearing water)
Eke and Qelik (2016, 2850124)Rat, Swiss Albino
(Male)
Peripheral blood cells
Positive
0.6 - 2.5 mg/kg/day
(every other day via gavage for 30 days)
Eke etal. (2017, 3981318)
Rat, Swiss Albino
(Male)
Liver
Positive
0.6 - 2.5 mg/kg/day
(every other day via gavage for 30 days)
Micronuclei Formation
Celiketal. (2013,2919161)
Rat, Swiss Albino
(Female)
Bone marrow
Positive
0.6 - 2.5 mg/kg/day
(every other day via gavage for 30 days)
Du etal. (2014, 2851143)
Zebrafish, AB
(Male and female)
Peripheral blood cells
Positive
0.4 - 1.6 mg/L
(single dose to rearing water for 30 days)
Eke and Qelik (2016, 2850124)Rat, Swiss Albino
(Male)
Peripheral blood cells
Positive
0.6 - 2.5 mg/kg/day
(every other day via gavage for 30 days)
Eke etal. (2017, 3981318)
Rat, Swiss Albino
(Male)
Liver
Positive
0.6 - 2.5 mg/kg/day
(every other day via gavage for 30 days)
Murli (1996, 10228098)
Mouse, Crl:CD-l
(Male and female)
Bone marrow
Negative
a
NTP (2019, 5400978)
Rat, Sprague Dawley
(Male)
Peripheral blood cells
Negative
0.312-5 mg/kg/day
(daily via gavage for 28 days)
NTP (2019, 5400978)
Rat, Sprague Dawley
(Female)
Peripheral blood cells
Equivocal
0.312-5 mg/kg/day
(daily via gavage for 28 days)
Wang et al. (2015, 2850220)
Mouse, Gpt delta
transgenic
(Male)
Bone marrow
Negative
1-10 mg/kg/day
(daily via gavage for 28 days)
a Findings based on the 2016 EPA's Health Effects Support Document for Perlluorooctane Sulfonic Acid {U.S. EPA, 2016, 3603365}, concentrations) unknown.
3-274
-------
DRAFT FOR PUBLIC COMMENT MARCH 2023
Table 3-17. DNA Damage Data from In Vitro Studies
Reference
In Vitro Model
(Assay)
Results
Concentration
(Duration of exposure)
Chromosomal Aberrations
Murli (1999, 10228132)
Human lymphocytes
Negative
10 - 470 ng/mL
(3 hours)
Unscheduled DNA Synthesis
Cifone (1999, 10228136)
Fisher 344 male rat hepatocytes
Negative
0.25 - 4,000 ng/mL
DNA Damage
Wang et al. (2015, 2850220)
gpt Delta transgenic mouse embryonic
fibroblasts
(Y-H2AX foci)
Positive
0 - 30 nM
(24 hours)
Jacquet et al. (2012, 2919219)
Syrian hamster embryo cells
(comet assay)
Negative
2 x 10"4 - 50 ng/mL
(7 days)
Kawamoto etal. (2010, 1274162)
Paramecium caudatum
(comet assay)
Negative
10 - 100 nM
(1-24 hours)
Lu et al. (2012, 2919198)
Calf thymus DNA
(X-ray photoelectron spectroscopic and
electrochemical impedance spectroscopy)
Positive
10 |imol/L
(30 minutes)
Wielsoe et al. (2015, 2533367)
HepG2
(comet assay)
Positive
2 x 10"7 - 2 x 10"5 M
(24 hours)
Florentin et al. (2011, 2919235)
HepG2
(comet assay)
Negative
5 - 300 nM
(1 or 24 hours)
3-275
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.5.3.2 Key Characteristic #4: Induces Epigenetic Alterations
Epigenetic alterations are modifications to the genome that do not change genetic sequence.
Epigenetic alterations include DNA methylation, histone modifications, changes in chromatin
structure, and dysregulated microRNA expression, all of which can affect the transcription of
individual genes and/or genomic stability {Smith, 2016, 3160486}.
3.5.3.2.lln Vivo Evidence
3.5.3.2.1.11.2.1.1. Humans
A cohort of singleton term births were recruited from Faroese hospitals over an eighteen-month
period from 1986 to 1987 {Leung, 2018, 4633577}. At delivery, samples of umbilical cord
whole blood and scalp hair from the mothers were collected and used to measure toxicant levels
as well as evaluation of DNA methylation. PFOS levels were significantly correlated with the
number of methylated CpG sites (10,598 sites) in male newborn umbilical cord whole blood
samples. Data from the male samples were then used to evaluated potential gene networks or
pathways enriched based on the genes related to the methylated CpG sites; specifically, to
evaluate potential relationships between physiological functions/diseases and the PFOS-induced
aberrant methylation patterns. The top physiological function related to the methylation changes
was "nervous system development and function." Additionally, CpG sites for which PFOS
exposure altered the methylation status were associated with individual genes related to cancer.
A subset of adults enrolled in the C8 Health Project between August 1, 2005 and August 31,
2006 were evaluated for exposure to perfluoroalkyl acids (PFAAs) via drinking water {Watkins,
2014, 2850906}. The cross-sectional survey consisted only of residents within the mid-Ohio
River Valley. A second, short-term follow-up study including another sample collection was
conducted in 2010 to evaluate epigenetic alterations in relation to serum PFOS concentrations.
Serum concentrations of PFOS decreased slightly between enrollment (2005-2006) and follow-
up (2010). Methylation of long interspersed nuclear elements (LINE-1) transposable DNA
elements in peripheral blood leukocytes at the follow-up timepoint in 2010 was significantly
associated with PFOS exposure, with an unadjusted 0.265% increase in LINE-1 methylation (per
12 ng/mL increase in mean serum PFOS). This association between LINE-1 methylation and
PFOS exposure remained significant after adjusting for covariates; a 0.20% increase was
observed when the data were adjusted for age, gender, BMI, smoking status, and drinking status.
Additional epidemiological studies of prenatal or birth cohorts have identified epigenetic
alterations associated with PFOS, indicating exposure can induce global DNA methylation
changes and alterations to methylation of CpG sites that are associated with genes involved in
several physiological functions and diseases related to development. For additional information,
please see the developmental mechanistic section (Section 3.4.4.3; refer to the
interactive Tableau for additional supporting information and study details).
3.5.3.2.1.21.2.1.2. Animals
Dysregulation of long non-coding RNAs in rodent in vivo studies following PFOS exposure has
been demonstrated, leading to reduced placental size. For additional information, please see the
developmental mechanistic section (Section 3.4.4.3; refer to the interactive Tableau for
additional supporting information and study details). It should be noted that such effects were not
seen in other tissues or in relation to other effects that may be more relevant to cancer outcomes.
3-276
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Additional rodent evidence examined liver microRNA (miRNA) expression and found an
increase in the expression of miR-34a-5p, which is involved in p53-mediated apoptosis,
following exposure to PFOS. For additional information, please see the hepatic mechanistic
section (Section 3.4.1.3; refer to the interactive Tableau for additional supporting information
and study details).
3.5.3.2.2/n Vitro Evidence
Pierozan et al. (2020, 6833637) evaluated PFOS (10 |iM) in the MCF-10A breast cell line. After
72 hours of exposure, PFOS-treated cells exhibited decreased acetylation of histone H3K9
(H3K9ac). In contrast, no alterations were found in the levels of H3K9 methylation and H3K26
acetylation.
Several additional studies have evaluated the potential of PFOS to alter the epigenome within
various in vitro systems designed to test developmental effects. The available mechanistic
studies suggest that the developing liver, developing heart, and placenta may be affected by
PFOS at the molecular level (i.e., differential methylation of genes, gene expression changes,
mitochondrial dysregulation). For additional information, please see the developmental
mechanistic section (Section 3.4.4.3; refer to the interactive Tableau for additional supporting
information and study details).
3.5.3.3 Key Characteristic #5: Induce Oxidative Stress
Reactive oxygen and nitrogen species (ROS and RNS, respectively) are byproducts of energy
production that occur under normal physiological conditions. An imbalance in the detoxification
of reactive such species can result in oxidative (or nitrosative) stress, which can play a role in a
variety of diseases and pathological conditions, including cancer. The primary mechanism by
which oxidative stress leads to the carcinogenic transformation of normal cells is by inducing
oxidative DNA damage that leads to genomic instability and/or mutations 1 Smith et al., 2016,
3160486 J.
3.5.3.3.lln Vivo Evidence
3.5.3.3.1.1 Humans
Several human epidemiological studies have reported that PFOS exposure induces oxidative
stress, leading to cardiological dysregulation (e.g., endothelial dysfunction, impaired
vasodilation, increased 8-OHdG and 8-N02Gua). For additional information, please see the
cardiovascular mechanistic section (Section 3.4.3.3; refer to the interactive Tableau for
additional supporting information and study details).
3.5.3.3.1.2 Animals
Male Sprague Dawley rats were administered 1 or 10 mg/kg/day PFOS orally for 28 days {Han,
2018, 4238554}. Following exposure, significant increases in ROS production and nitric oxide
synthase mRNA expression were noted in the liver. Elevation of oxidative stress was associated
with decreased intracellular antioxidant defense by aberrant catalase and superoxide dismutase
activities.
Liu et al. (2009, 757877) studied markers of oxidative stress in the liver and brain in KM mice
exposed to PFOS and found that there was no treatment effect. The authors found that levels of
3-277
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
malondialdehyde (MDA) did not differ between controls and exposed animals, and that
superoxide dismutase activity was lower in treated vs. control mice, indicating that oxidative
stress was not induced.
Evidence of increased oxidative stress in the liver, including increased ROS levels, changes in
GSH and GSSG levels, and decreases in antioxidant enzymes, was observed in rodents in vivo
following oral exposure to PFOS. For additional information, please see the hepatic mechanistic
section (Section 3.4.1.3; refer to the interactive Tableau for additional supporting information
and study details).
3.5.3.3.2/n Vitro Evidence
Several studies have evaluated ROS production in HepG2 cells exposed to PFOS, reporting
varied results. A study by Hu and Hu (2009, 2919334) demonstrated PFOS exposure (50-200
|imol/L; 24-72 hours) induced a significant increase in ROS. This effect correlated with
decreased mitochondrial membrane potential and apoptosis. Furthermore, PFOS exposure caused
increased superoxide dismutase, catalase, and glutathione reductase levels but decreased
glutathiones-transferase and glutathione peroxidase levels in cells. In contrast, Florentin et al.
(2011, 2919235) exposed HepG2 cells to PFOS (5-300 |iM) for 24 hours and found a decrease in
ROS generation by approximately 23%.
A study by Wang et al. (2015, 2850220) used mouse embryonic fibroblast (MEF) cells to
identify intercellular ROS induced by PFOS exposure (1 or 20 |iM). Using a fluorescent free
radical probe CM-FhDCFDA kit to evaluate ROS levels, cells exposed to 20 |iM PFOS had a
significantly higher level of florescence than controls, indicating PFOS induced intercellular
oxidative stress. To better understand the role of H2O2 in this PFOS-induced cytotoxicity
(Section 3.5.3.7) and genotoxicity (Section 3.5.3.1), Wang et al. treated cells concurrently with a
cell membrane-permeating catalase to initiate the breakdown of H2O2 and protect cells from
oxidative damage. In the presence of catalase, cytotoxicity and DNA double strand break
frequency were decreased in PFOS-exposed cells. Mutation frequencies were also significantly
suppressed in cells exposed to both PFOS and catalase when compared to cells exposed to PFOS
alone. These results in Wang et al. (2015, 2850220) suggest that PFOS-induced genotoxicity is
mediated by the induction of ROS.
Wielsoe et al. (2014, 2533367) exposed HepG2 cells to PFOS (2 x 10"7 to 2 x 10"5 M) for 24
hours. Following exposure, the cells demonstrated significant increase in intercellular ROS at all
tested PFOS concentrations.
Several studies have identified the potential of PFOS to induce oxidative stress within various in
vitro testing systems that are designed to understand effects during developmental stages. The
available mechanistic studies demonstrated that oxidative stress mediates alterations in
development and gross morphology following PFOS exposure. PFOS. For additional
information, please see the developmental mechanistic section (Section 3.4.4.3; refer to the
interactive Tableau for additional supporting information and study details).
Further evidence of the ability of PFOS to induce oxidative stress is described elsewhere. PFOS
exposure has been shown to be associated with increased markers of oxidative damage and
decreased activity of protective antioxidants that play a role in the reduction of oxidative
3-278
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
damage. PFOS. For additional information, please see the hepatic mechanistic section (Section
3.4.1.3; refer to the interactive Tableau for additional supporting information and study details).
3.5.3.4 Key Characteristic #6: Induces Chronic Inflammation
The induction of chronic inflammation includes increased white blood cells, altered chemokine
and/or cytokine production, and myeloperoxidase activity {Smith, 2016, 3160486}. Chronic
inflammation has been associated with several forms of cancer, and a role of chronic
inflammation in the development of cancer has been hypothesized. However, there are biological
links between inflammation and oxidative stress and genomic instability, such that the
contribution of each in carcinogenic progression is not always clear.
Several studies have identified the potential of PFOS to increase inflammation within various in
vivo and in vitro models. It is important to note that in vitro models may be used for the
evaluation of changes in inflammatory markers and response, they are generally not effective in
modeling the events that are associated with chronic inflammation. For additional information,
please see the immune (Section 3.4.2.3), hepatic (Section 3.4.1.3), developmental (Section
3.4.4.3), and cardiovascular (Section 3.4.3.3) mechanistic sections (refer to the interactive
Tableau for additional supporting information and study details).
3.5.3.5 Key Characteristic #7: Is Immunosuppressive
Immunosuppression refers to the reduction in the response of the immune system to antigen,
which is important in cases of tumor antigens 1 Smith, 2016, 3 1604861. It is important to note
that immunosuppressive agents do not directly transform cells, but rather can facilitate immune
surveillance escape of cells transformed through other mechanisms (e.g., genotoxicity).
Studies have identified the immunosuppressive potential of PFOS in in vivo and in vitro testing
systems. Specifically, PFOS has been associated with depression of natural killer cell activity,
reduced macrophage function, and changes in the cellularity and immunophenotypes of
lymphocytes. For additional information, please see the immune mechanistic section (Section
3.4.2.3; refer to the interactive Tableau for additional supporting information and study details).
3.5.3.6 Key Characteristic #8: Modulates Receptor-Mediated Effects
Modulation of receptor-mediated effects involves the activation or inactivation of receptors (e.g.,
PPAR, AhR) or the modification of endogenous ligands (including hormones) 1 Smith, 2016,
3160486 J.
3.5.3.6.lln Vivo Evidence
Several studies have reported the potential of PFOS to modulate nuclear receptor- and hormone-
mediated effects within various in vivo and in vitro testing systems, specifically models relevant
to the hepatic system.
PFOS has been shown to activate several nuclear receptors, including PPARa, PPARy,
PPARp/S, CAR/PXR, and LXR/RXR. Many of these nuclear receptors, including PPARa and
CAR, are known to play an important role in liver homeostasis and have been implicated in liver
dysfunction. PFOS exposure may lead to liver toxicity through the activation of multiple nuclear
receptors in both rodents and humans. For additional information, please see the hepatic
3-279
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
mechanistic section (Section 3.4.1.3; refer to the interactive Tableau for additional supporting
information and study details).
3.5.3.6.2/n Vitro Evidence
3.5.3.6.2.1 PPAR Mediated Effects
Liver-expressed peroxisome PPARa regulates transcription of genes involved in peroxisome
proliferation, cell cycle control, apoptosis, and lipid metabolism. Data for PFOS illustrates the
ability of PFOS to activate PPARa {Shipley, 2004, 2990378; Martin, 2007, 758419; Wolf, 2008,
716635; Wolf, 2014, 2850908}.
Jacquet et al. (2012, 2919219) exposed Syrian hamster embryo (SHE) cells to PFOS (<50
|ig/mL) for 5 and 24 hours. Evaluation of PPAR gene expression by qPCR indicated a 3.0-fold
increase ofppar-b/d mRNA level at a PFOS concentration of 0.2 |ig/mL after 24 hours.
Subsequent exposure of SHE cells to PFOS (0.02-20 (ag/m L) for 1 week found overexpression of
PPAR-target genes and a significant increase of ppar-b/d mRNA at 0.2 |ag/m L (2-fold increase)
and 2 |ag/mL (2.5-fold increase). mRNA levels ofppar-y were significant increased after 7 days
at all PFOS exposure concentrations. Interestingly, upregulation of the ppar-a gene was found at
the lowest concentration tested (0.2 |ig/mL). A study using MCF-7 human breast cancer cells
demonstrated that PFOS increased proliferation in a dose-dependent manner at concentrations of
0.01 and 30 |ig/mL, a response that was observed in tandem with the maximal estrogen (E2)
response, suggesting that PFOS may be an estrogen receptor agonist at these concentrations
{Henry, 2013, 1805116}.
3.5.3.7 Key Characteristic ft 10: Alters Cell Proliferation, Cell Death,
or Nutrient Supply
Aberrant cellular proliferation, cell death, and/or nutrient supply is a common mechanism among
carcinogens. This mechanism includes aberrant proliferation, decreased apoptosis or other
evasion of terminal programming, changes in growth factors, angiogenesis, and modulation of
energetics and signaling pathways related to cellular replication or cell cycle control {Smith,
2016, 3160486}.
3.5.3.7.lln Vivo Evidence
3.5.3.7.1.1 Humans
Epidemiological studies found an association between PFOS exposure and increased markers of
endothelial and platelet apoptosis. For additional information, please see the cardiovascular
mechanistic section (Section 3.4.3.3; refer to the interactive Tableau for additional supporting
information and study details).
3.5.3.7.1.2 Animals
Proliferation of peroxisomes has been suggested as a mechanism of action for several non-
genotoxic carcinogens that induce liver tumors upon chronic administration to rats and mice
{Ashby, 1994, 630327; Rao and Reddy, 1996, 1334694}, and PFOS has been shown to activate
PPARs. In a study of male and female Sprague Dawley rats administered PFOS in the diet at 0,
0.5, 2, 5, or 20 ppm for 4 or 14 weeks, there was no evidence of increased hepatic cell
proliferation {Seacat, 2003, 1290852}. However, the same authors continued this same dietary
3-280
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
PFOS exposure in Sprague Dawley rats for up to two years and found liver effects consistent
with PPAR activation {Thomford 2002, 5029075; Butenhoff, 2012, 1276144}. This two-year
cancer bioassay found that the only neoplastic response that was attributable to PFOS exposure
was an increased incidence of hepatocellular adenoma in both male and female rats in the
20 ppm PFOS group.
3.5.3.7.2/n Vitro Evidence
Two human giant cell tumor (GCT)-derived cell lines (COV434 and KGN) were exposed to
PFOS (0.08-8,000 ng/mL) for 72 hours {Gogola, 2018, 5016947}. PFOS significantly increased
proliferation in both cell lines in a dose-dependent manner. Specifically, PFOS treatment at
0.08 ng/mL increased COV434 and KGN proliferation by 1.4-fold and 1.9-fold, respectively.
Follow up studies by the same authors did not observe any change in caspase 3 or 7 activities in
cells exposed to concentrations of PFOS (0.8, 8, or 80 ng/ml; 72 hours), both of which play a
role in apoptosis {Gogola. 2020, 6316203; Gogola, 2020, 6316206}.
The potential of PFOS to induce tumorigenic activity (proliferation, cell-cycle progression, and
malignant phenotype) was evaluated in MCF-10A breast epithelial cells {Pierozan, 2018,
4238459}. Exposure to 10 |iM promoted proliferation by accelerating GO/Gl-to-S phase
transition of the cell cycle after 24, 48, and 72 hours of exposure. PFOS exposure increased
CDK4 while simultaneously decreased p27, p21, and p53 levels in MCF-10A cells. Furthermore,
10 |iM PFOS exposure for 72 hours stimulated MCF-10A cell migration and invasion. A follow
up study evaluating PFOS (10 |iM; 72 hours) in MCF-10A cells induced proliferation and
alteration of regulatory cell-cycle proteins (cyclin Dl, CDK6, p21, p53, p27, ERK1, ERK2, and
p38) {Pierozan, 2020, 6833637}. Additionally, PFOS exposure increased cell migration and
invasion in unexposed daughter cells of exposed cells, as evidenced by a reduction in the levels
of E-cadherin, occludin, and P-integrin. A study in MCF-7 human breast cancer cells
demonstrated that PFOS increased proliferation in a dose-dependent manner at concentrations of
0.01 and 30 |ig/mL, a response that may be the result of estrogen receptor activation {Henry,
2013, 1805116}. These results elucidate PFOS's potential carcinogenic effects through alteration
of cell proliferation.
In contrast to these results, no changes in cellular proliferation were observed in MCF-7 breast
adenocarcinoma cells exposed to PFOS (0.1-100 [xM) for 24 hours {Maras, 2006, 2952988}.
However, a small but significant downregulation of estrogen-responsive genes (TFFI and ESR1)
was noted following PFOS exposure.
In a study designed to determine the effect of PFOS effect on the tumor suppressor protein SHP-
2, HepG2 cells were exposed to sub-cytotoxic concentrations of PFOS for 24 hours before SHP-
2 was immunoprecipitated from the cell lysates {Yang, 2017, 3981427}. While PFOS exposure
increased SHP-2 gene expression in a concentration-dependent manner, it was also found to have
an inverse proportional decrease in SHP-2 enzyme activity. Interestingly, a 1.4-fold increase in
SHP-2 protein levels was observed in exposed cells, indicating that PFOS inhibits SHP-2 by
blocking enzymatic activity post-translationally.
For additional information, please see the developmental mechanistic section (Section 3.4.4.3;
refer to the interactive Tableau for additional supporting information and study details).
3-281
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
3.5.4 Weight Of Evidence for Carcinogenicity
3.5.4.1 Summary of Evidence
Several epidemiological studies and a single chronic cancer bioassay comprise the evidence
database for the carcinogenicity of PFOS. The available epidemiology studies report elevated
risk of bladder, prostate, kidney, and breast cancers after chronic PFOS exposure. However, the
study designs, analyses, and mixed results do not allow for a definitive conclusion on the
relationship between PFOS exposure and cancer outcomes in humans. The sole animal chronic
cancer bioassay study provide support for multi-site tumorigenesis in male and female rats.
3.5.4.1.1Evidence from Epidemiological Studies
Studies of the association between PFOS serum concentrations and bladder cancer have mixed
findings. An elevated risk of bladder cancer mortality was associated with PFOS exposure in an
occupational study {Alexander, 2003, 1291101} but a subsequent study to ascertain cancer
incidence in the cohort observed elevated but not statistically significant incidence ratios that
were 1.7- to 2-fold higher among workers with higher cumulative exposure {Alexander, 2007,
4727072}. The risk estimates lacked precision because the number of cases was small, and the
study did not control for the potential confounding of smoking. A nested case-control study in a
general population Danish cohort did not observe elevated bladder cancer risk with increasing
PFOS serum levels {Eriksen, 2009, 2919344}.
Elevated non-significant ORs for prostate cancer were reported for the occupationally exposed
cohort examined by Alexander and Olsen (2007, 4727072) and the Danish population-based
cohort examined by Eriksen et al. (2009, 2919344). In the same occupational cohort studied by
Alexander and Olsen (2007, 4727072), Grice et al. (2007, 4930271) observed that prostate
cancers were among the most frequently reported malignancies. When cumulative exposure
measures were analyzed, elevated ORs were reported for prostate cancer, however, they did not
reach statistical significance. Length of follow-up may not have been adequate to detect cancer
incidence in this cohort as approximately one-third of the participants had worked < 5 years in
their jobs, and only 41.7% were employed > 20 years. No association between PFOS exposure
and prostate cancer was reported in either a second case-control study in Denmark {Hardell,
2014, 2968084} or in a study of the association between PFOS serum concentrations and
prostate specific antigen (a biomarker of prostate cancer) from the C8 Health Project {Ducatman,
2015, 3859843}. In an NHANES population, Omoike et al. (2021, 7021502) observed a
significantly inverse association with prostate cancer.
One study in the general population reported a statistically significant increase in risk of renal
cell carcinoma in the highest PFOS exposure quartile and per doubling of PFOS concentration
{Shearer, 2021, 7161466}. Although the trend was significant across quartiles, the effect in the
third quartile was null. Additionally, the association with PFOS was attenuated after adjusting
for other PFAS, and it was lower in the third quartile than in the second quartile, indicating
potential confounding by correlated PFAS exposures. There was no association when evaluated
on a per doubling of PFOS after adjusting for other PFAS.
The majority of studies examining associations between PFOS and cancer outcomes were on
breast cancer. No association was identified between PFOS and breast cancer in either a case-
control or a nested case-control studies of Danish and California cancer registry populations,
3-282
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
respectively {Bonefeld-J0rgensen, 2014, 2851186; Hurley, 2018, 5080646}. One study of Inuit
females in Greenland observed positive associations between PFOS levels and risk for breast
cancer {Bonefeld-fergensen, 2011, 2150988}, although the association was of a low magnitude
and could not be separated from the effects of other perfluorosulfonated compound exposures.
Three studies indicated potential associations between PFOS exposure and increased breast
cancer risk in specific subgroups or increased risk for specific breast cancer subtypes. Ghisari et
al. (2017, 3860243) found that increased breast cancer risk was associated with increased PFOS
serum concentrations in Danish individuals with a specific polymorphism in the CYP19 gene
(aromatase; associated with estrogen biosynthesis and metabolism). Mancini et al. (2019,
5381529) reported that increased PFOS serum concentrations were associated specifically with
increased risk of ER+ and PR+ tumors, whereas risk of ER- and PR- tumors did not follow a
dose-dependent response. In a Taiwanese population Tsai et al. (2020, 6833693) observed a
statistically significant increase in risk of breast cancer, but only in participants aged 50 years or
younger, and in ER+ breast cancer in participants aged 50 years or younger. Another general
population study in the U.S. suggested that maternal PFOS exposure combined with high
maternal cholesterol may decrease the daughters' risk of breast cancer but did not examine breast
cancer subtypes or genetic variants {Cohn, 2020, 5412451}. Significantly increased breast
cancer risk was also observed in an NHANES population in the two highest quartiles of
exposure, but the association was inverse in the second quartile {Omoike, 2021, 7021502}. A
recent study in a Japanese population observed inverse association across serum PFOS quartiles
with a significant dose-response trend {Itoh, 2021, 9959632}. The association remained
significantly inverse in both pre- and postmenopausal women in the highest tertile of exposure,
with a significant dose-response trend. However, in some of the studies PFOS levels were
measured after or near the time of cancer diagnosis {Tsai, 2020, 6833693; Omoike, 2021,
7021502}. Given the long half-life of PFOS in human blood, the exposure levels measured in
these studies could represent exposures that occurred prior to cancer development. However, this
is currently difficult to evaluate since data on the latency of PFOS-related cancer is not available.
Overall, study design issues, lack of replication of the results, and a lack of mechanistic
understanding of PFOS on specific breast cancer subtypes or in subpopulations limit firm
conclusions regarding PFOS and breast cancer. These findings are supported by other recent
assessments and reviews {ATSDR2021, 9642134; Steenland, 2021, 7491705; CalEPA, 2021,
9416932}.
3.5.4.1.2 Evidence from Animal Bioassays
The single available chronic toxicity/carcinogenicity bioassay for PFOS in animals is a 104-week
dietary study in rats {Thomford, 2002, 5029075; Butenhoff, 2012, 1276144}. Statistically
significant increases in the incidence of hepatocellular adenomas in the high dose (20 ppm) male
(7/43; 16%) and female rat groups (5/31; 16%) and combined adenomas/carcinomas in the
females (6/32; 19%; 5 adenomas, 1 carcinoma) were observed. The observation of a carcinoma
in the female rats is a relatively rare occurrence according to NTP's historical controls for female
Sprague-Dawley rats (1/639 historical control incidence) {NTP, 2020, 10368689}. Historical
control incidence rates for these tumor types were not provided by Thomford (2002, 5029075).
Additionally, there were statistically significant trends in the hepatic tumor responses of both
males and females. A statistically significant trend of increased incidence of pancreatic islet cell
carcinomas with increased PFOS dose was also observed in the male rats, though the individual
dose groups were not statistically different from the control group. The percentages of animals
3-283
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
with islet cell carcinomas in the highest dose group (12.5%) exceeds NTP's historical controls
for male Sprague-Dawley rats by over an order of magnitude (12/638; 1.9%) {NTP, 2020,
10368689}.
Thyroid tumors (follicular cell adenomas and carcinomas) were observed in males and females,
though these responses were not statistically significant in any dose group, nor was there a linear
dose-response trend. In males, the incidence of thyroid tumors was significantly elevated only in
the high-dose, recovery group males exposed for 52 weeks (10/39) but not in the animals
receiving the same dose for 105 weeks. However, Thomford (2002, 5029075) indicated that the
number of thyroid tumors observed in the recovery group males were outside the range of
historical control values at that time, similar to what NTP (2020, 10368689) has reported for its
laboratories (3/637 combined follicular cell adenoma or carcinoma). There were very few
follicular cell adenomas/carcinomas in the females (4 total, excluding the recovery group) with a
non-linear dose-response. There was also a high background incidence of mammary gland
tumors in the female rats, primarily combined fibroma adenoma and adenoma, but the incidence
lacked dose-response for all tumor classifications.
3.5.4.2 Mode of Action for Hepatic Tumors
The strongest evidence of the carcinogenicity of PFOS comes from a high confidence chronic
rodent study identifying hepatocellular tumors in both male and female rats {Butenhoff, 2012,
1276144; Thomford, 2002, 5029075}. As described in the subsections below, the available
mechanistic data suggest that multiple MO As may underlie the hepatocellular tumors observed
after PFOS exposure. Specifically, the available studies provide varying levels of support for the
role of several plausible MO As: PPARa activation, CAR activation, HNF4a suppression,
cytotoxicity, genotoxicity, oxidative stress, and immunosuppression.
3.5.4.2.1PPARa activation
There is considerable debate over the relevance of PFAS-induced hepatic tumors to human
health. Exposure to some PFAS have been shown to activate PPARa, which is characterized by
downstream cellular or tissue alterations in peroxisome proliferation, cell cycle control (e.g.,
apoptosis and cell proliferation), and lipid metabolism {U.S. EPA, 2016, 3603365}. Notably,
human expression of PPARa mRNA and protein is only a fraction of what is expressed in rodent
models, though there are functional variant forms of PPARa that are expressed in human liver to
a greater extent than rodent models {Klaunig, 2003, 5772415; Corton, 2018; 4862049}.
Therefore, for PPARa activators that act solely or primarily through PPARa-dependent
mechanisms (e.g., Wyeth-14,643, di-2-ethyl hexyl phthalate), the hepatic tumorigenesis observed
in rodents may be expected to be reduced in frequency or severity or not observed in humans
{Klaunig, 2003, 5772415; Corton, 2014, 2215399; Corton, 2018, 4862049}.
The adverse outcome pathway (AOP) for the PPARa MOA for hepatic tumors has been
characterized to include the following set of key events: 1) PPARa activation in hepatic cells; 2)
alterations in cell growth signaling pathways (e.g., increases in Kupffer cell activation leading to
increases in TNFa); 3) perturbations of hepatocyte growth and survival (i.e., increased cell
proliferation and inhibition of apoptosis); and 4) selective clonal expansion of preneoplastic foci
cells leading to 5) increases in hepatocellular adenomas and carcinomas {Klaunig, 2003,
5772415; Corton, 2014, 2215399; Corton, 2018, 4862049}. This AOP is associated with but not
necessarily causally related to non-neoplastic effects including peroxisome proliferation,
3-284
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
hepatocellular hypertrophy, Kupffer cell-mediated events, and increased liver weight. There is
also some overlap between signaling pathways and adverse outcomes, including tumorigenesis,
associated with PPARa activation and the activation or degradation of other nuclear receptors,
such as CAR, PXR, HNF4a, and PPARy {Rosen, 2017, 3859803; Huck, 2018, 5079648; Beggs,
2016, 3981474; Corton, 2018, 4862049}.
Table 3-18. Evidence of Key Events Associated with the PPARa Mode of Action in Male
Sprague-Dawley Rats Exposed to PFOS
Dose
(mg/kg/day)
Key Event 1
(PPARa
activation)
Key Event 2
(altered cell
growth
signaling)
Key Event 3a
(altered
hepatocyte
growth)
Key Event 3b
(altered
hepatocyte
survival)
Key Event 4
(preneoplastic
clonal
expansion)
Key Event 5
(hepatic
tumors)
0.024
-
-
- (4 & 14w)
- (14 & 103w)
-
- (103w)
0.098
- (4w)
-
- (4 & 14w)
- (14 & 103w)
-
- (103w)
0.242
-
-
- (4 & 14w)
- (14 & 103w)
-
- (103w)
0.312
T(4w)
-
-
- (4w)
-
-
0.625
T(4w)
-
-
- (4w)
-
-
0.984-1
t (4w, GD 1-
PND 20 Fi
PND21)
T(4w)
V- (4w)
- (14 & 53w)
- (14 & 53w)
4 (103w)
t (103w)
1.5-1.93
-(Id)
t (7d & 4w)
-(Id)
T(7d)
V- (4w)
-(Id)
!(7d)
-/T (4w)
Notes: | = statistically significant increase in response compared to controls; - = no significant response; j = statistically
significant decrease in response compared to controls; a "/" separating symbols for direction of effect indicates that multiple
studies assessed the key event at the same dose and time point but reported conflicting results; d = day(s); w = week(s).
Data represented in table extracted from NTP (2019, 5400978); Chang et al. (2009, 757876); Elcombe et al. (2012,1332473);
Elcombe et al. (2012,1401466); Curran et al. (2008, 757871); Han et al. (2018,4355066); Butenhoff et al. (2012,1276144)/
Thomford (2002, 5029075).
The published in vivo and in vitro literature suggests that PFOS is a relatively weak PPARa
agonist compared to other known PPARa agonists such as PFOA {Martin, 2007, 758419; Wolf,
2012, 1289836; Behr, 2020, 6305866; Rosen, 2013, 2919147}. While in vitro PPARa activation
assay results indicate overall effective activation of PPARa by PFOS, the magnitude of that
activation has been found to be relatively lower than chemicals that induce toxicity primarily
through PPARa activation (e.g., di-2-ethyl hexyl phthalate). There is in vivo rodent assay
evidence of PFOS-induced PPARa-associated transcriptional and enzymatic responses (e.g.,
upregulation of Acoxl and acyl-CoA activity) as well. However, consistent with the in vitro
activation assays, these in vivo responses were relatively weaker than PFOA and/or other PPARa
activators and were often reported to be accompanied by transcriptional responses associated
with other nuclear receptor signaling pathways (e.g., CAR and PPARy), consistent with multiple
modes of action {Martin, 2007, 758419; Dong, 2016, 3981515; NTP, 2019, 5400978; Chang,
2009, 757876; Elcombe, 2012, 1332473; Elcombe, 2012, 1401466}. For further details, see
Section 3.4.1.3. Consistent with these findings, studies ofWT and PPARa-null mice reported
that 808 differentially expressed genes responsive to a 7-day 10 mg/kg/day PFOS exposure were
expressed in PPARa-null mouse livers while 906 genes were differentially expressed in WT
mice, corroborating the likelihood of an active PPARa-independent MOA(s) {Rosen, 2010,
1274165}. Robust PPARa-independent effects in null mice were observed even at the lowest
dose of PFOS (3 mg/kg/day; 630 differentially expressed genes in PPARa-null mice vs. 81
3-285
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
differentially expressed genes in WT mice) compared to responses in mice treated with
3 mg/kg/day Wyeth-14,643 or PFOA (902 genes WT, 10 genes PPARa-null and 879 genes WT,
176 genes PPARa-null, respectively) {Rosen, 2010, 1274165}, consistent with multiple MOAs
for PFOS hepatic effects.
Table 3-19. Evidence of Key Events Associated with the PPARa Mode of Action in Female
Sprague-Dawley Rats Exposed to PFOS
Dose
(mg/kg/day)
Key Event 1
(PPARa
activation)
Key Event 2
(altered cell
growth
signaling)
Key Event 3a
(altered
hepatocyte
growth)
Key Event 3b
(altered
hepatocyte
survival)
Key Event 4
(preneoplastic
clonal
expansion)
Key Event 5
(hepatic
tumors)
0.029
-
-
- (4 & 14w)
- (14 & 103w)
-
- (103w)
0.12
4 (4w)
-
- (4 & 14w)
- (14 & 103w)
-
- (103w)
0.299
-
-
- (4 & 14w)
- (14 & 103w)
-
- (103w)
0.312
T(4w)
-
-
- (4w)
-
-
0.625
T(4w)
-
-
- (4w)
-
-
1.251
- (4w & GD 1-
GD 20 dam)
-
- (4, 14 & 53w)
- (14 & 53w)
4 (103w)
-
t (103w)
Notes: | = statistically significant increase in response compared to controls; - = no significant response; j = statistically
significant decrease in response compared to controls; d = day(s); w = week(s).
Data represented in table extracted from NTP (2019, 5400978); Chang et al. (2009, 757876); Curran et al. (2008, 757871);
Butenhoff et al. (2012, 1276144)/ Thomford (2002, 5029075).
There is evidence from in vivo animal bioassays and in vitro studies of Kupffer cell activation, an
indicator of alterations in cell growth, in response to PFOS treatment. Though this mechanism is
itself PPARa-independent, factors secreted upon Kupffer cell activation may be required for
increased cell proliferation by PPARa activators {Corton, 2018, 4862049}. Two short-term
exposure in vivo rodent studies reported increased serum TNFa levels after 3-4 weeks of PFOS
administration {Han, 2018, 4355066; Su, 2019, 5080481}; TNFa is a pro-inflammatory cytokine
that can be released upon activation of Kupffer cells {Corton, 2018, 4862049}. In addition to
serum TNFa levels, Han et al. (2018, 4355066) reported increased TNFa mRNA in hepatic
tissues of PFOS-exposed rats. The authors also extracted primary Kupffer cells from untreated
rats and cultured them with PFOS in vitro for 48 hours and reported increased supernatant TNFa
levels and cellular TNFa mRNA levels. These results indicate that rodent hepatic tissues may be
primed for perturbations of PPARa-dependent cell growth upon PFOS exposure. However,
further study is needed to understand the potential role of other mediators of Kupffer cell
activation since unlike PPARa, PPARy is expressed in Kupffer cells and can also be activated by
PFOS.
While there is some evidence of alterations in cell growth signaling pathways due to PFOS
exposure, there is conflicting evidence related to the ability of PFOS to induce hepatic cell
proliferation and inhibit apoptosis. The available rodent in vivo study results indicate that
increases in proliferation may be dose- and exposure duration-dependent whereas changes in
apoptosis may be species- or dose-dependent. In the only available chronic rodent bioassay for
PFOS {Thomford, 2002, 5029075; Butenhoff, 2012, 1276144}, significant increases in the
number of hepatic tumors were observed at the highest dose levels in each sex (20 ppm in diet or
approximately 1 mg/kg/day) without corresponding increases in the incidence or severity of cell
proliferation at 52 weeks in the livers of male or female rats. Additionally, there were transient
3-286
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
effects on hepatic peroxisomal proliferation in males or females at weeks 4 and 14 as indicated
by the palmitoyl-CoA assay {Thomford, 2002, 5029075; Seacat, 2003, 1290852}. In contrast,
there is evidence of hepatic cell and/or peroxisome proliferation from short-term studies that
administered higher PFOS dose levels than the Thomford report (2002, 5029075) (i.e., 2-
10 mg/kg/day) {Elcombe, 2012, 1401466; Elcombe, 2012, 1332473; NTP, 2019, 5400978; Han,
2018, 4355066}. Results were not always consistent across time points or sexes and were
accompanied by evidence of increased activation of other nuclear receptors (i.e., CAR and PXR),
which could also influence cell proliferation. The characteristics of typical PPARa-induced cell
proliferation includes an early burst that recovers to a level that is slightly higher than
background, the latter of which is difficult to detect for compounds that are weak PPARa
activators {Corton, 2014, 4862049}. This likely explains, at least in part, the inconsistencies in
cell proliferation patterns across timepoints and lends support to the evidence of relatively weak
PPARa activation by PFOS. Additionally, Elcombe et al. (2012, 1401466) reported substantially
greater palmitoyl-CoA oxidation after 50 ppm Wyeth-14,643 administration in male Sprague-
Dawley rats compared to 20 or 100 ppm (approximately 1.7 and 7.9 mg/kg/day, respectively)
PFOS administration for up to 28 days, lending further support for PFOS as a relatively weak
PPARa activator.
In addition to the observation of increased hepatic cell proliferation on day 1 of recovery in male
rats administered 20 or 100 ppm PFOS (approximately 1.93 and 9.65 mg/kg/day, respectively)
for 7 days, Elcombe et al. (2012, 1332473) also reported decreased hepatic apoptotic indices
(i.e., the percent of apoptotic nuclei out of the total number cell nuclei in a unit of area) in both
dose groups, which is an indication of PPARa-dependent hepatotoxicity. However, these results
were inconsistent with the results of the second Elcombe et al. (2012, 1401466) study, which
reported an increased apoptotic index after 7 days of 20 ppm dietary PFOS administration. The
authors observed no other statistically significant changes in the apoptotic indices of rats from
the 20 ppm group in the two additional timepoints tested (1 day and 28 days), though they did
report decreases in the apoptotic indices of rats in the 100 ppm group at all three time points,
similar to the results of Elcombe et al. (2012, 1332473; 2012, 1401466). The underlying reason
for the inconsistent apoptosis findings in the 20 ppm dose groups between the two studies is
unclear. Increased hepatic apoptosis was observed in mice administered 2.5-10 mg/kg/day PFOS
for 30 days {Xing, 2016, 3981506}, and short-term PFOS studies in both rats and mice reported
increases in apoptosis-related hepatic gene expression and/or protein activity/expression {Eke,
2017, 3981318; Wan, 2016, 3981504; Han, 2018, 4238554; Lv, 2018, 5080395}. Further
descriptions of these in vivo studies, as well as in vitro studies examining hepatic cell
proliferation and apoptosis can be found in Section 3.4.1.3.
There are several studies of the hepatic effects resulting from PFOS exposure observed in
PPARa-null mice with either short-term or gestational exposure durations but therefore, lack an
ability to assess tumor incidence or chronic histopathological effects. The studies of Qazi et al.
(2009, 1937260), Abbott et al. (2009, 2919376), and Rosen et al. (2010, 1274165) all observed
increased absolute and/or relative liver weight in PPARa-null adults orally administered PFOS
or pups exposed to PFOS in utero. Along with the PPARa-independent cell signaling effects in
PPARa-null mice reported by Rosen et al. (2010, 1274165; 2017, 3859803), these studies
corroborate that the hepatomegaly observed in WT rodents administered PFOS is not entirely
PPARa-dependent. Several other signaling pathways may contribute to the observed
hepatomegaly due to PFOS exposure, though the relationship of these liver effects with tumor
3-287
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
formation is unclear. Further descriptions of studies utilizing PPARa-null mice can be found in
Section 3.4.1.3.
In general, PPARa activators are not expected to induce cell proliferation or suppress apoptosis
of hepatocytes in humans {Corton, 2018, 4862049}. Specifically, there is strong consensus that
the MOA for liver tumor induction by PPARa activators in rodents has limited-to-no relevance
to humans, due to differences in cellular expression patterns of PPARa and related proteins (e.g.,
cofactors and chromatin remodelers), as well as differences in binding site affinity and
availability {Corton, 2018, 4862049; Klaunig, 2003, 5772415}. Nonetheless, several studies
have reported increased cell proliferation or markers of cell proliferation in vitro in human liver
cell lines exposed to PFOS {Cui, 2015, 3981568; Song, 2016, 9959776; Louisse, 2020,
6833626} (see Section 3.4.1.3). For example, Cui et al. (2015, 3981568) found increased
proliferation using the MTT assay in the non-tumor fetal human liver cell line HL-7702. These
increases in cell proliferation were accompanied by corresponding proteomic changes indicative
of increased proliferation. Using flow cytometry, Cui et al. (2015, 3981568) also found that
increased percentages of cells were in cell phases associated with DNA synthesis and/or
interphase growth and mitosis (S and G2/M phases), depending on the length of exposure and
dose of PFOS. Corroborative transcriptional results were observed in two additional human cell
lines (HepG2 and HepaRG) {Song, 2016, 9959776; Louisse, 2020, 6833626}. There was no
mention of changes in apoptosis accompanying increased cell proliferation in two of the studies
of human hepatocytes {Cui, 2015, 3981568; Louisse, 2020, 6833626}, while Song et al. (2016,
9959776) reported that genes related to "regulation of apoptosis" were significantly altered,
although the direction of the change is not specified. Beggs et al. (2016, 3981474) reported that a
human primary cell line exposed to PFOS predominantly showed changes in the expression of
genes involved in carcinogenesis and cell death signaling, among other biological
pathways/functions related to hepatotoxicity and hepatic diseases. The authors linked these
transcriptional changes to the loss of HNF4a functionality which is known to promote the
development of hepatocellular carcinoma, providing evidence of a PPARa-independent
mechanism of hepatotoxicity and carcinogenicity. In addition to HNF4a-mediated
hepatocarcinogenicity, Benninghoff et al. (2012, 1274145) proposed that promotion of
hepatocarcinogenesis by PFOS in an initiation-promotion model in rainbow trout, which are
similarly insensitive to PPARa as humans, is potentially the result of activation of the trout liver
estrogen receptor. Specifically, dietary PFOS treatment promoted hepatocarcinogenesis (i.e.,
increased the incidence of hepatocellular carcinomas and adenomas) and increased tumor
promotion and cell proliferation in rainbow trout exposed to aflatoxin Bi as a cancer initiator
{Benninghoff, 2012, 1274145}.
3.5.4.2.20ther Nuclear Receptors
In addition to PPARa, there is some evidence that other nuclear receptors may play a role in the
MOA for hepatic tumors resulting from PFOS exposure. For example, CAR, which has an
established adverse outcome pathway of key events similar to PPARa, has been implicated in
hepatic tumorigenesis in rodents. The key events of CAR-mediated hepatic tumors are: 1)
activation of CAR; 2) altered gene expression specific to CAR activation; 3) increased cell
proliferation; 4) clonal expansion leading to altered hepatic foci; and 5) liver tumors {Felter,
2018, 9642149}. Associative events include hypertrophy, induction of CAR-specific CYP
enzymes (e.g., CYP2B) and inhibition of apoptosis. As described in Section 3.4.1.3, there is both
in vivo and in vitro evidence that PFOS can activate CAR and initiate altered gene expression
3-288
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
and associative events {Dong, 2016, 3981515; NTP, 2019, 5400978; Martin, 2007, 758419;
El combe, 2012, 1401466; Chang, 2009, 757876; Elcombe, 2012, 1332473; Rosen, 2010,
1274165; Rosen, 2013, 2919147; Rosen, 2017, 3859803}. Some studies, such as NTP (2019,
5400978), report greater activation of CAR with PFOS treatment compared to PPARa,
depending on the sex and/or model of interest. As with PPARa-mediated turn oogenesis, there
are claims that CAR-mediated tumorigenesis is not relevant to humans because CAR activators
such as phenobarbital have been shown to induce cell proliferation and subsequent tumorigenesis
in rodents but do not induce cell proliferation in human cell lines {Elcombe, 2014, 2343661}.
However, as outlined above, several studies have reported increased cell proliferation or markers
of cell proliferation due to PFOS treatment in human cell lines {Cui, 2015, 3981568; Song,
2016, 9959776; Louisse, 2020, 6833626}. Further study is needed to understand the mechanistic
underpinnings of PFOS-induced hepatic cell proliferation and whether it is related to CAR
activation.
Table 3-20. Evidence of Key Events Associated with the CAR Mode of Action in Male
Sprague-Dawley Rats Exposed to PFOS
Dose (mg/kg/day)
Key Event 1
(CAR
activation)
Key Event 2
(Altered Gene
Expression)
Key Event 3
(Cell
Proliferation)
Key Event 4
(clonal
expansion)
Key Event 5
(Hepatic Tumors)
0.024
-
-
- (4 & 14 w)
-
- (103 w)
0.098
-
-
- (4 & 14 w)
-
- (103 w)
0.242
-
-
- (4 & 14 w)
-
- (103 w)
0.312
-
t (4 w)
-
-
-
0.625
-
t (4 w)
-
-
-
0.984-1
-
t (GD 1-PND
20, PND 21 Fi)
V- (4 w)
- (14 & 53 w)
-
t (103 w)
Notes: | = statistically significant increase in response compared to controls; - = no significant response; j = statistically
significant decrease in response compared to controls; a "/" separating symbols for direction of effect indicates that multiple
studies assessed the key event at the same dose and time point but reported conflicting results; d = day(s); w = week(s).
Data represented in table extracted from NTP (2019, 5400978); Chang et al. (2009, 757876); Elcombe et al. (2012,1332473);
Elcombe et al. (2012,1401466); Han et al. (2018, 4355066); Butenhoff et al. (2012, 1276144)/ Thomford (2002, 5029075).
Table 3-21. Evidence of Key Events Associated with the CAR Mode of Action in Female
Sprague-Dawley Rats Exposed to PFOS
Key Event 5
(Hepatic Tumors)
Key Event 1 Key Event 2 Key Event 3 Key Event 4
Dose (ppm) (CAR (Altered Gene (Cell (clonal
activation) Expression) Proliferation) expansion)
0.024/ - - - (4 & 14 w) - - (103 w)
0.029
0.098/0.12 - - - (4 & 14 w) - - (103 w)
0.242/0.299 - - - (4 & 14 w) - - (103 w)
0.312 - t(4w)
0.625 - t (4 w)
0.984/ 1.251 - | (GD 1-GD 20, - (4, 14 & 53 w) - | (103 w)
dam)
Notes: | = statistically significant increase in response compared to controls; - = no significant response; j = statistically
significant decrease in response compared to controls; d = day(s); w = week(s).
3-289
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Data represented in table extracted from NTP (2019, 5400978); Chang et al. (2009, 757876); Butenhoff et al. (2012, 1276144)/
Thomford (2002, 5029075).
HNF4a is known as a master regulator of hepatic differentiation and plays a role in tumor
suppression as well as general liver maintenance and function {Beggs, 2016, 3981474}.
Interestingly, PFOS exposure appears to downregulate HNF4a and its target genes. Studies
utilizing primary human hepatocytes, HepG2 cells, and in vivo mouse models have reported
decreased HNF4a protein expression as well as corresponding changes in downstream HNF4a
target genes with PFOS treatment {Beggs, 2016, 3981474; Behr, 2020, 6505973}. Beggs et al.
(2016, 3981474) reported that PFOS induced changes in genes involved in carcinogenesis and
cell death signaling and linked the loss of HNF4a functionality to potential hepatocellular tumor
promotion. The authors also suggested that loss of HNF4a functionality may play a role in non-
cancer hepatic effects including hepatomegaly, steatosis, altered lipid metabolism, and fatty liver
disease.
There is additional evidence from in vivo and in vitro studies that PFOS has the ability to activate
and modulate the targets of other nuclear receptors. As described in Section 3.4.1.3, PFOS has
been reported to modulate the activity of PPARs other than PPARa (i.e., PPARp/S, and PPARy),
PXR, LXR, RXR, RAR, and ErP, though the evidence of activation is sometimes conflicting
across different cell lines, assays, and species. Several of these nuclear receptors, such as
PPARy, are known to play a role in liver homeostasis and disease and may be driving factors in
the hepatotoxicity observed after PFOS exposure, though their role in tumorigenesis is less clear.
As described in Section 3.5.3, there is also evidence that PFOS modulates endogenous ligands
for nuclear receptors, most notably thyroid and reproductive hormones. However, it is also
unclear what role, if any, these receptors and ligands may be playing in PFOS-induced hepatic
tumorigenesis.
3.5.4.2.3 Cytotoxicity
There is suggestive evidence that PFOS may act through a cytotoxic MOA. Felter et al. (2018,
9642149) identified the following key events for establishing a cytotoxicity MOA: 1) the
chemical is not DNA reactive; 2) clear evidence of cytotoxicity by histopathology such as the
presence of necrosis and/or increased apoptosis; 3) evidence of toxicity by increased serum
enzymes indicative of cellular damage that are relevant to humans; 4) presence of increased cell
proliferation as evidenced by increased labeling index and/or increased number of hepatocytes;
5) demonstration of a parallel dose response for cytotoxicity and formation of tumors; and 6)
reversibility upon cessation of exposure. As discussed above in the genotoxicity section, there is
some evidence that PFOS can induce DNA damage and/or micronuclei formation in liver tissue
{Eke, 2017, 3981318; Wang, 2015, 2850220}. These data indicate that PFOS may be DNA
reactive (either directly or indirectly), but it is unclear if this DNA reactivity is the source of the
tumor findings {Holsapple, 2006, 194740}. Quantitative liver histopathology is limited to three
studies, however the one available chronic study {Butenhoff, 2012, 1276144} reported
significant trends in increased individual hepatocyte necrosis in male and female Sprague-
Dawley rats which was also statistically significant in the highest dose groups. Liver
histopathology in humans is also limited, however, Jin et al. (2020, 6315720) reported higher
odds (not necessarily statistically significant) of nonalcoholic steatohepatitis (p < 0.05),
ballooning, fibrosis, and portal inflammation.
3-290
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 3-22. Evidence of Key Events Associated with the Cytotoxicity Mode of Action in
Male Sprague-Dawley Rats
Key Event 4
Dose
(mg/kg/day)
Key Event 1
(Cytotoxicity)
Key Event 2
Key Event 3
(Hyperplasia
Key Event 5
(Serum
Enzymes)
(Regenerative
Proliferation)
and/or
Preneoplastic
Lesions)
(Hepatic
Tumors)
0.024
- (14 & 103 w)
- (4, 14, 27 &
53 w)
- (4 & 14 w)
- (14 & 103 w)
- (103 w)
0.098
- (14 & 103 w)
- (4, 14, 27 &
53 w)
- (4 & 14 w)
- (14 & 103 w)
- (103 w)
0.242
- (14 & 103 w)
- (4, 14, 27 &
53 w)
- (4 & 14 w)
- (14 & 103 w)
- (103 w)
0.312
- (4 w)
- (4 w)
-
- (4 w)
-
0.625
- (4 w)
t (4 w)
-
- (4 w)
-
0.984
- (4, 14 & 53 w)
t (4, 14 &
V- (4 w)
- (14 & 53 w)
t (103 w)
t (103 w)
53 w)
- (27 w)
- (14 & 53 w)
t (103 w)
Notes: | = statistically significant increase in response compared to controls; - = no significant response; j = statistically
significant decrease in response compared to controls; a "/" separating symbols for direction of effect indicates that multiple
studies assessed the key event at the same dose and time point but reported conflicting results; d = day(s); w = week(s).
NTP (2019, 5400978); Elcombe et al. (2012, 1332473); Elcombe et al. (2012,1401466); Han et al. (2018,4355066); Butenhoff
et al. (2012, 1276144)/ Thomford (2002, 5029075).
There is evidence in both humans and animals that exposure to PFOS increases serum liver
enzymes. Specifically, statistically significant positive associations between ALT and PFOS (i.e.,
increased ALT as a continuous measure with higher PFOS exposure levels) were observed in
several studies {Salihovic, 2018, 5083555; Nian, 2019, 5080307; Jain, 2019, 5381541; Costa,
2009, 1429922; Gallo, 2012, 1276142; Olsen, 2003, 1290020}. These individual findings are
supported by a meta-analysis of epidemiological studies reporting biomarkers of liver injury
reporting a statistically significant (p < 0.001) weighted z-score suggesting a positive association
between PFOS and increased ALT in adults and children {Costello, 2022, 10285082}.
Statistically significant increases in serum enzymes (i.e., ALT, AST, ALP, and GGT) were also
observed in several animal toxicological studies, though these increases were generally less than
two-fold (100% change relative to control) compared to control {Seacat, 2003, 1290852; Curran,
2008, 757871; Butenhoff, 2012, 1276144; Xing, 2016, 3981506; Yan, 2014, 2850901; NTP,
2019, 5400978; Han, 2018, 4355066}. However, these changes in serum enzyme levels were
accompanied by histopathological evidence of damage, as outlined above, and coherence is
observed in humans.
Table 3-23. Evidence of Key Events Associated with the Cytotoxicity Mode of Action in
Female Sprague-Dawley Rats
Dose
(mg/kg/day)
Key Event 1
(Cytotoxicity)
Key Event 2
(Serum
Enzymes)
Key Event 3
(Regenerative
Proliferation)
Key Event 4
(Hyperplasia
and/or
Preneoplastic
Lesions)
Key Event 5
(Hepatic
Tumors)
0.029
- (14 & 103 w)
- (4, 14, 27 &
53 w)
- (4 & 14 w)
- (14 & 103 w)
- (103 w)
3-291
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Dose
(mg/kg/day)
Key Event 1
(Cytotoxicity)
Key Event 2
(Serum
Enzymes)
Key Event 3
(Regenerative
Proliferation)
Key Event 4
(Hyperplasia
and/or
Preneoplastic
Lesions)
Key Event 5
(Hepatic
Tumors)
0.12
- (14 & 103 w)
- (4, 14, 27 &
- (4 & 14 w)
- (14 & 103 w)
- (103 w)
53 w)
0.299
- (14 & 103 w)
- (4, 14, 27 &
- (4 & 14 w)
- (14 & 103 w)
- (103 w)
53 w)
0.312
- (4 w)
- (4 w)
-
- (4 w)
-
0.625
- (4 w)
- (4 w)
-
- (4 w)
-
1.251
- (4, 14 & 53 w)
- (4, 14, 27 &
- (4, 14 & 53 w)
- (14 & 53 w)
t (103 w)
t (103 w)
53 w)
t (103 w)
Notes: | = statistically significant increase in response compared to controls; - = no significant response; j = statistically
significant decrease in response compared to controls; d = day(s); w = week(s).
NTP (2019, 5400978); Butenhoff et al. (2012, 1276144)/ Thomford (2002, 5029075).
As highlighted in the PPARa activation section, several studies have reported increased cell
proliferation or markers of cell proliferation in human cell lines {Cui, 2015, 3981568; Song,
2016, 9959776; Louisse, 2020, 6833626}, though there is limited quantitative histopathological
data to determine the ability of PFOS to induce hepatic hyperplasia. Finally, the available data
indicate a parallel dose response for cytotoxicity and the formation of liver tumors as evidence in
Table 3-24 and Table 3-25, though dose spacing (i.e., the gap in dosing between the mid-high
and high doses administered) may limit the precision of a dose response curve.
Table 3-24. Incidences of Liver Tumor and Nonneoplastic Lesions in Male Sprague-Dawley
Rats at 103 weeks, as Reported by Thomford (2002, 5029075)
0 mg/kg/day 0.024 mg/kg/day 0.098 mg/kg/day 0.242 mg/kg/day 0.984 mg/kg/day
Hepatocellular
0/41**
3/42
3/47
1/44
7/43**
Adenomas
Necrosis, Individual
3/50
2/50
6/50
4/50
10/50
Hepatocyte
Altered
13/50
21/50
23/50
24/50
24/50
Hepatocellular,
Clear/Eosinophilic
Cell
Cystic Degeneration
5/50
15/50
19/50
17/50
22/50
Hyperplasia, Bile
19/50
20/50
25/50
24/50
25/50
Duct
Notes:
Statistical significance for an exposed group indicates a significant pairwise test compared to the vehicle control group. Statistical
significance for the vehicle control indicates a significant trend test.
* Statistically significant at p < 0.05; ** p < 0.01.
Table 3-25. Incidences of Liver Tumor and Nonneoplastic Lesions in Female Sprague-
Dawley Rats at 103 weeks, as Reported by Thomford (2002, 5029075)
0 mg/kg/day
0.029 mg/kg/day 0.120 mg/kg/day 0.299 mg/kg/day 1.251 mg/kg/day
Combined
Hepatocellular
0/28**
1/29 1/16 1/31 6/32*
3-292
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
0 mg/kg/day 0.029 mg/kg/day 0.120 mg/kg/day 0.299 mg/kg/day 1.251 mg/kg/day
Adenomas &
Carcinomas
Necrosis, Individual
Hepatocyte
Infiltrate,
Macrophage,
Pigmented
Infiltrate,
Lymphohistiocytic
Hyperplasia, Bile
Duct
Notes:
Statistical significance for an exposed group indicates a significant pairwise test compared to the vehicle control group. Statistical
significance for the vehicle control indicates a significant trend test.
* Statistically significant at p < 0.05; ** p < 0.01.
3.5.4.2.4Genotoxicity
Several relatively recent studies, primarily published by the same laboratory, have shown the
potential for PFOS to act as a genotoxicant (see Section 3.5.3); previously, EPA had not
identified evidence supporting genotoxicity as a potential MOA for PFOS {U.S. EPA, 2016,
3603365}. Two in vivo studies, the first a 30-day study in male Swiss Albino rats and the second
a 28-day study in male gpt delta transgenic mice, provided evidence of DNA damage and/or
micronuclei formation in liver tissue of animals administered up to 2.5 or 10 mg/kg/day PFOS,
respectively {Eke, 2017, 3981318; Wang, 2015, 2850220}. However, there are concerns about
the Interpretation of these studies regarding the genotoxicity and mutagenicity of PFOS because
results reported as not statistically significant, concerns about the study design, or unclear
relationship of the observed effects to genotoxicity of PFOS vs. Secondary effects from
hepatoxicity (e.g., oxidative stress).
Several other 28-30-day studies in male and female rats and mice also observed DNA damage
and/or micronuclei formation in bone marrow or peripheral blood cells {(^elik, 2013, 2919161;
Eke, 2016, 2850124; NTP, 2019, 5400978}, though there are similar concerns about whether
these responses are attributable to direct genotoxicity of PFOS. For example, NTP (2019,
5400978) reported increased numbers of micronucleated polychromatic erythrocytes in the blood
of female rats administered 5 mg/kg/day PFOS (highest dose group) for 28 days, but also
reported concomitant decreases in the percentage of polychromatic erythrocytes in the peripheral
blood, indicative of bone marrow toxicity. This potential bone marrow toxicity may be driving
micronuclei formation rather than the direct mutagenicity of PFOS. NTP (2019, 5400978) also
noted that the observed responses of the high dose females were within historical control ranges
and considered these results to be equivocal. From this very limited database, it does not appear
that genotoxicity in male and female Sprague-Dawley rats occurs at doses at or below those that
result in tumorigenesis.
In addition to rodent studies, Du et al. (2014, 2851143) reported increased DNA strand breaks
and micronuclei formation in peripheral blood cells of male and female zebrafish exposed to
PFOS for 30 days and several other studies reported increased DNA damage in vitro {Wang,
2015, 2850220; Lu, 2012, 2919198; Wielsoe, 2014, 2533367}. However, the majority of in vitro
3/50 4/50 4/50 5/50 9/50
2/50 3/50 5/50 6/50 20/50
33/50 37/50 33/50 36/50 42/50
21/50 25/50 19/50 17/50 27/50
3-293
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
studies (described in Section 3.5.3) report negative results for genotoxic endpoints including
chromosomal aberrations, unscheduled DNA synthesis, mutagenicity, and various types of DNA
damage.
The available in vivo evidence suggests that exposure to PFOS may result in genotoxicity and is
particularly compelling for the potential indirect genotoxicity that could stem from
hepatotoxicity and/or bone marrow toxicity. At this time, there are no generally accepted
mechanistic explanations for PFOS directly interacting with genetic material. Additionally, while
there is some in vivo evidence of PFOS-induced mutagenicity as primarily evidenced by
micronuclei formation in rats, mice, and zebrafish, there are several uncertainties that limit the
interpretation of these results. There is currently no robust evidence to support a mutagenic
MOA for PFOS, though overall, genotoxicity cannot be ruled out as a potential MOA for PFOS.
3.5.4.2.5Consideration of Other Plausible MOAs
In addition to the evidence supporting modulation of receptor-mediated effects, and potential
genotoxicity, PFOS also exhibits several other key characteristics (KCs) of carcinogens (see
Section 3.5.3), some of which are similarly directly evident in hepatic tissues.
For example, PFOS appears to induce oxidative stress, another KC of carcinogens, particularly in
hepatic tissues (see Section 3.4.1.3). Several studies in rats and mice showed evidence of
increased oxidative stress and reduced capacity for defense against oxidants and oxidative
damage in hepatic tissues. Interestingly, two studies, one 28-day study in rats and one 30-day
study in mice, reported reduced Nrf2 protein levels or expression in hepatic tissues after PFOS
exposure {Wan, 2016, 3981504; Lv, 2018, 5080395}. Nrf2 is an important regulator of
antioxidant response elements and is generally activated in response to pro-oxidant exposure and
oxidative stress. Accordingly, these studies and others noted a reduction in the hepatic
expression of genes that are implicated in antioxidant, anti-inflammatory, and/or stress response
functions (e.g., hmoxl, nqol) as well as reduced antioxidant enzyme levels and activities (e.g.,
CAT, SOD) {Wan, 2016, 3981504; Lv, 2018, 5080395; Han, 2018, 4238554; Liu, 2009, 757877;
Xing, 2016, 3981506}. Several in vivo exposure studies also noted increases in hepatic ROS and
markers of oxidative damage (e.g., MDA) {Han, 2018, 4238554; Liu, 2009, 757877; Xing, 2016,
3981506; Wan, 2016, 3981504; Lv, 2018, 5080395}. Notably, Han et al. (2018, 4238554)
reported several indicators of oxidative stress in male Sprague-Dawley rats gavaged for 28 days
with 1 mg/kg/day PFOS (lowest dose tested in the study), a comparable dose to that which
caused tumorigenesis in the chronic study in male rats. Taken together, these results provide
some support for disruption of the oxidative stress response in hepatic tissues leading to
accumulation of ROS and subsequent oxidative damage.
Immunosuppression is the reduction of an individual's immune system to respond to foreign
cells or antigens, including tumor cells {Smith, 2020, 6956443}. The immune system plays an
important role in the identification and eventual destruction of cancer cells; immunosuppression
may allow for the evasion of this process by cancer cells and subsequently lead to tumorigenesis.
As discussed in Section 3.4.2.1.1, PFOS serum levels are associated with markers of
immunosuppression, particularly in children. Several studies reported inverse associations
between PFOS serum concentrations and antibody production following vaccinations in children
{Grandjean, 2017, 3858518; Grandjean, 2017, 4239492; Mogensen, 2015, 3981889;
Timmermann, 2020, 6833710}. Additionally, one medium confidence study reported higher odds
3-294
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
of total infectious diseases with increasing PFOS serum concentrations {Goudarzi, 2017,
3859808}, though it should be noted that studies reporting odds ratios for individual infectious
diseases had mixed results. Animal toxicological studies also report markers of
immunosuppression, including reductions in natural killer cell activity. As described in Section
3.4.2.2, there are several reports of decreased natural killer cell activity in male and female, adult
and Fi generation mice from short-term, subchronic, and gestational studies {Dong, 2009,
1424951; Peden-Adams, 2008, 1424797; Keil, 2008, 1332422; Zhong, 2016, 3748828; Zheng,
2009, 1429960}. While one short-term study in male mice reported increases in splenic T-helper
(CD3+CD4+) and T-cytotoxic (CD3+CD8+) lymphocytes {Lv, 2015, 3981558}, two gestational
studies reported reductions in thymic CD4+ cells in male offspring {Zhong, 2016, 3748828; Keil,
2008, 1332422}. There is also limited evidence of immunosuppression in the form of reduced
white blood cell counts (primarily lymphocytes) from two short-term rodent studies in male mice
and rats, respectively {Qazi, 2009, 1937259; NTP, 2019, 5400978}. This short-term report is the
only available study in Sprague-Dawley rats and does not indicate that immunosuppressive
effects are occurring at or below doses that result in tumorigenesis {NTP, 2019, 5400978}.
However, it is difficult to discount immunosuppression as a potential MOA for PFOS, given the
limited database for rats and stronger databases indicating immunosuppression in mice and
humans.
3.5.4.2.6Conclusions
Based on the weight of evidence evaluation of the available peer-reviewed scientific evidence,
PFOS has the potential to induce hepatic tumors via multiple MOAs in rodents, most notably via
the modulation of nuclear receptors (i.e., PPARa and CAR) and cytotoxicity. There is also
limited evidence supporting potential MOAs of genotoxicity, immunosuppression, and oxidative
stress. The conclusions from the weight of evidence analysis of the available data for PFOS are
consistent with literature reviews recently published by two state health agencies which
concluded that the hepatotoxic effects of PFOS are not entirely dependent on PPARa activation
{CalEPA, 2021, 9416932; NJDWQI, 2018, 5026035}.
As described in the Guidelines for Carcinogen Risk Assessment {U.S. EPA, 2005, 6324329},
"[i]n the absence of sufficiently, scientifically justifiable mode of action information, EPA
generally takes public health-protective, default positions regarding the interpretation of
toxicologic and epidemiologic data; animal tumor findings are judged to be relevant to humans,
and cancer risks are assumed to conform with low dose linearity." For the available data
regarding the MOA of PFOS-induced hepatic carcinogenesis, there is an absence of definitive
information supporting a single, scientifically justified MOA; in fact, there is evidence
supporting the potential for multiple plausible MOAs. Therefore, EPA concludes that the hepatic
tumors observed by Thomford (2002, 5029075) and Butenhoff et al. (2012, 1276144) can be
relevant to human health and support the positive, albeit, limited, tumor findings from
epidemiological studies.
3.5.5 Cancer Classification
Under the Guidelines for Carcinogen Risk Assessment {U.S. EPA, 2005, 6324329}, EPA
reviewed the weight of the evidence and determined that PFOS is Likely to Be Carcinogenic to
Humans, as "the evidence is adequate to demonstrate carcinogenic potential to humans but does
not reach the weight of evidence for the descriptor Carcinogenic to HumansThe Guidelines
3-295
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
provide descriptions of data that may support the Likely to Be Carcinogenic to Humans
descriptor; the available PFOS data are consistent with the following factors:
• "an agent that has tested positive in animal experiments in more than one species, sex,
strain, site, or exposure route, with or without evidence of carcinogenicity in humans;
• a rare animal tumor response in a single experiment that is assumed to be relevant to
humans; or
• a positive tumor study that is strengthened by other lines of evidence, for example, either
plausible (but not definitively causal) association between human exposure and cancer or
evidence that the agent or an important metabolite causes events generally known to be
associated with tumor formation (such as DNA reactivity or effects on cell growth
control) likely to be related to the tumor response in this case" {U.S. EPA, 2005,
6324329}.
The epidemiological evidence of associations between PFOS and cancer found mixed results
across tumor types. However, the available study findings support a plausible correlation
between PFOS exposure and carcinogenicity in humans. The single chronic cancer bioassay
performed in rats is positive for multi-site and -sex tumorigenesis {Thomford, 2002, 5029075;
Butenhoff, 2012, 1276144}. In this study, statistically significant increases in the incidences of
hepatocellular adenomas or combined adenomas and carcinomas were observed in both male and
female rats. There was also a statistically significant trend of this response in both sexes
indicating a relationship between the magnitude/direction of response and PFOS dose. As
described in Section 3.5.4.2, the available mechanistic evidence is consistent with multiple
potential MO As for this tumor type; therefore, the hepatocellular tumors observed by
Thomford/Butenhoff et al. (2002, 5029075; 2012; 1276144) may be relevant to humans. In
addition to hepatocellular tumors, Thomford/Butenhoff etal. (2002, 5029075; 2012; 1276144)
reported increased incidences of pancreatic islet cell tumors with a statistically significant dose-
dependent positive trend, as well as modest increases in the incidence of thyroid follicular cell
tumors. The findings of multiple tumor types provide additional support for potential multi-site
tumorigenesis resulting from PFOS exposure.
The PFOA carcinogenicity database includes both epidemiological studies and animal bioassays
that support its designation as Likely to be Carcinogenic to Humans. Structural similarities
between PFOS and PFOA add to the weight of evidence for carcinogenicity of PFOS. Notably, a
similar set of non-cancer effects have been observed after exposure to either PFOA or PFOS in
humans and animal toxicological studies including similarities in hepatic, developmental,
immunological, cardiovascular, and endocrine effects, among others.
3-296
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 3-26. Comparison of the PFOS Carcinogenicity Database with the Likely Cancer
Descriptor as Described in the Guidelines for Carcinogen Risk Assessment {U.S. EPA,
2005, 6324329}
Likely to be Carcinogenic to Humans
An agent demonstrating a plausible (but not
definitively causal) association between
human exposure and cancer, in most cases
with some supporting biological,
experimental evidence, though not
necessarily carcinogenicity data from animal
experiments
Epidemiological studies evaluating the association between human
exposure to PFOS and cancer are mixed. Supporting carcinogenicity
data are available from animal experiments.
An agent that has tested positive in animal
experiments in more than one species, sex,
strain, site, or exposure route, with or without
evidence of carcinogenicity in humans
PFOS data are consistent with this description. PFOS has tested
positive in animal experiments in more than one sex and site.
Hepatic tumors were observed in male and female rats (statistically
significant at high dose and statistically significant trend tests for
each) and islet cell carcinomas show a statistically significant
positive trend in male rats.
A positive tumor study that raises additional
biological concerns beyond that of a
statistically significant result, for example, a
high degree of malignancy, or an early age at
onset
This description is not applicable to PFOS.
A rare animal tumor response in a single
experiment that is assumed to be relevant to
humans
PFOS data are consistent with this description. The
hepatocellular carcinoma observed in the high-dose female rats is a
rare tumor type in this strain {NTP, 2020, 7330145}.
A positive tumor study that is strengthened
by other lines of evidence, for example, either
plausible (but not definitively causal)
association between human exposure and
cancer or evidence that the agent or an
important metabolite causes events generally
known to be associated with tumor formation
(such as DNA reactivity or effects on cell
growth control) likely to be related to the
tumor response in this case
PFOS data are consistent with this description. The positive
multi-site, multi-sex chronic cancer bioassay is supported by
mechanistic data indicating that PFOS is associated with events
generally known to be associated with tumor formation such as
inducing nuclear receptor activation, cytotoxicity, genotoxicity,
oxidative stress, and immunosuppression.
While reviewing the weight of evidence for PFOS, EPA evaluated consistencies of the
carcinogenicity database with other cancer descriptors according to the Guidelines for
Carcinogen Risk Assessment {U.S. EPA, 2005, 6324329}. A discussion on these findings is
presented in Section 6.4.
3-297
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
4 Dose-Response Assessment
4.1 Non-Cancer
4.1.1 Study and End point Selection
There is evidence from both epidemiological and animal toxicological studies that oral PFOS
exposure may result in adverse health effects across many health outcomes (Section 3.4). Per
recommendations made by the SAB and the conclusions presented in EPA's preliminary
analysis, Proposed Approaches to the Derivation of a Draft Maximum Contaminant Level Goal
for Perfluorooctane Sulfonic Acid (PFOS) (CASRN1763-23-1) in Drinking Water, EPA has
focused its toxicity value derivation efforts "on those health outcomes that have been concluded
to have the strongest evidence" {U.S. EPA, 2022, 10476098}. EPA prioritized health outcomes
and endpoints with the strongest overall weight of evidence (evidence demonstrates or evidence
indicates) based on human and animal evidence (Section 3.4 and 3.5) for POD derivation using
the systematic review methods described in Section 2 and the Appendix (see PFOS Appendix).
For PFOS these health outcomes are immunological, developmental, cardiovascular (serum
lipids), and hepatic effects. EPA considered both epidemiological and animal toxicological
studies for POD derivation.
In the previous section, for hazard judgment decisions (Section 3.4 and 3.5), EPA qualitatively
considered high, medium, and, at times, low confidence studies to characterize the weight of
evidence for each health outcome. However, given the robust database for PFOS, only well-
conducted high or medium confidence human and animal toxicological studies were considered
for POD derivation, as recommended in the IRIS Handbook {U.S. EPA, 2022, 10367891}. Such
human epidemiological studies were available for immunotoxicity, developmental, serum lipid,
and hepatic effects. Preferred animal toxicological studies consisted of medium and high
confidence studies of longer exposure duration (e.g., chronic or subchronic studies vs. 28-day
studies) or with exposure during sensitive windows of development (i.e., perinatal periods) with
exposure levels near the lower dose range of doses tested across the evidence base, along with
medium or high confidence animal toxicological studies evaluating exposure periods relevant to
developmental outcomes. These types of animal toxicological studies increase the confidence in
the RfD relative to other animal toxicological studies because they are based on data with
relatively low risk of bias and are associated with less uncertainty related to low-dose and
exposure duration extrapolations. See Section 6.3 for a discussion of animal toxicological studies
and endpoints selected for POD derivation for this updated assessment compared to those
selected for the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}.
For all other health outcomes (e.g., reproductive, endocrine, nervous, hematological,
musculoskeletal), the evidence integration summary judgment for the human and animal
evidence was suggestive or inadequate and these outcomes were not assessed quantitatively.
Uncertainties related to health outcomes for which the results were suggestive are discussed in
the evidence profile tables provided in the Appendix (See PFOS Appendix), as well as Section
6.5.
4.1.1.1 Hepatic effects
As reviewed in Section 3.4.1.4, evidence indicates that elevated exposures to PFOS are
associated with hepatic effects in humans. As described in Table 3-5, the majority of
4-1
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
epidemiological studies assessed endpoints related to serum biomarkers of hepatic injury (8
medium confidence studies), while several studies also reported on liver disease or injury (3
medium confidence studies) and other serum markers of liver function (2 medium confidence
studies). EPA prioritized endpoints related to serum biomarkers of injury for quantitative
analyses as the reported effects on these endpoints, particularly ALT, were well-represented
within the database and were generally consistent across the available medium confidence
studies. Specifically, three medium confidence studies (out of five total) reported statistically
significant positive associations between PFOS serum concentrations and ALT in adults.
Findings for AST and GGT in adults were generally positive and are supportive of the selection
of ALT as an endpoint for dose-response modeling, though results from stratified analyses of
these two endpoints were less consistent.
Serum ALT measures are considered a reliable indicator of impaired liver function because
increased serum ALT is indicative of leakage of ALT from damaged hepatocytes {Boone, 2005,
782862; Liu, 2014, 10473988; U.S. EPA, 2002, 625713}. Additionally, evidence from both
human epidemiological and animal toxicological studies indicates that increased serum ALT is
associated with liver disease {Ioannou, 2006, 10473853; Ioannou, 2006, 10473854; Kwo, 2017,
10328876; Roth, 2021, 9960592}. Human epidemiological studies have demonstrated that even
low magnitude increases in serum ALT can be clinically significant. For example, a
Scandinavian study in people with no symptoms of liver disease observed that relatively small
increases in serum ALT were associated with liver diseases such as steatosis and chronic
hepatitis C {Mathiesen, 1999, 10293242}. Additionally, a study in Korea found that the use of
lowered thresholds for "normal" serum ALT values showed good prediction power for liver-
related adverse outcomes such as mortality and hepatocellular carcinoma {Park, 2019,
10293238}. Others have questioned the biological significance of relatively small increases in
serum ALT (i.e., less than 2-fold) reported in animal toxicological studies {Hall, 2012,
2718645}, though measures of ALT in these studies can be supported by histopathological
evidence of liver damage.
Additionally, numerous studies have demonstrated an association between elevated ALT and
liver-related mortality (reviewed by Kwo et al. (2017, 10328876)). Furthermore, the American
Association for the Study of Liver Diseases (AASLD) recognizes serum ALT as an indicator of
overall human health and mortality {Kim, 2008, 7757318}. For example, as reported by Kwo et
al. (2017, 10328876), Kim et al. (2004, 10473876) observed that higher serum ALT
concentrations corresponded to an increased risk of liver-related death in Korean men and
women; similarly, Ruhl and Everhart (2009, 3405056; 2013, 2331047) analyzed NHANES data
and observed an association between elevated serum ALT and increased mortality, liver-related
mortality, coronary heart disease in Americans, and Lee et al. (2008, 10293233) found that
higher serum ALT was associated with higher mortality in men and women in Olmstead County,
Minnesota. Furthermore, the American College of Gastroenterology (ACG) recommends that
people with ALT levels greater than 33 (men) or 25 IU/L (women) undergo screenings and
assessments for liver diseases, alcohol use, and hepatotoxic medication use {Kwo, 2017,
10328876}. Results of human and animal toxicological studies as well as the positions of the
AASLD and the ACG demonstrate the clinical significance of increased serum ALT. It is also
important to note that while evaluation of direct liver damage is possible in animal studies, it is
difficult to obtain biopsy-confirmed histological data in humans. Therefore, liver injury is
typically assessed using serum biomarkers of hepatotoxicity {Costello et al, 2022, 10285082}.
4-2
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Results reported in animal toxicological studies are consistent with the observed elevated ALT
indicative of hepatic damage in epidemiological studies. Specifically, studies in rodents found
that oral PFOS treatment resulted in increased liver weight (11/14 high and medium confidence
studies in adult rodents), alterations in levels of serum biomarkers of liver injury, particularly in
male rodents (i.e., ALT (7/7 studies), AST (4/7 studies), ALP (3/4 studies), and GGT (1/1
study)), and evidence of histopathological alterations including hepatocellular damage (5/7 high
and medium confidence studies). These hepatic effects, particularly the increases in serum
enzymes and histopathological evidence of liver damage are supportive of elevated ALT
observed in human populations. Mechanistic studies in rodents and limited evidence from in
vitro studies and animal models provide additional support for the biological plausibility and
human relevance of the apical effects observed in animals and suggest possible PPARa-
dependent and -independent MOA for PFOS induced liver toxicity. EPA prioritized studies that
quantitatively reported histopathological evidence of hepatic damage for dose-response modeling
because these endpoints are more direct measures of liver injury than serum biomarkers.
However, the observed increases in liver enzymes in rodents are supportive of the hepatic
damage confirmed during histopathological examinations in several studies.
Three medium confidence epidemiological studies {Gallo, 2012, 1276142; Lin, 2010, 1291111;
Nian, 2019, 5080307} and one high confidence animal toxicological study {Butenhoff, 2012,
1276144} were considered for POD derivation (Table 4-1). The largest study of PFOS and ALT
in adults is by Gallo et al. (2012, 1276142), which was conducted in over 30,000 adults from the
C8 Study. Two additional studies {Lin, 2010, 1291111; Nian, 2019, 5080307} were considered
by EPA for POD derivation because they reported significant associations in general populations
in the U.S and a highly exposed population in China, respectively. Nian et al. (2019, 5080307)
examined a large population of adults in Shenyang (one of the largest fluoropolymer
manufacturing centers in China) as part of the Isomers of C8 Health Project. In an NHANES
adult population, Lin et al. (2010, 1291111) observed elevated ALT levels per log-unit increase
in PFOS in the models adjusted for age, gender, and race/ethnicity, but not in the fully adjusted
models, or in the models additionally adjusted for PFOA, PFHxS and PFNA. While this is a
large nationally representative population, several methodological limitations, including lack of
clarity about base of logarithmic transformation applied to PFOS concentrations in regression
models and the choice to model ALT as an untransformed variable ultimately preclude its use for
POD derivation.
EPA identified one study in male rats by Butenhoff et al. (2012, 1276144), a chronic dietary
study, for POD derivation. Butenhoff et al. (2012, 1276144) conducted histopathological
examinations of liver tissue in male and female rats and reported dose-dependent increases in the
incidence of individual hepatocellular necrosis. As this is the only available chronic PFOS
toxicity studies with a large sample size, numerous and relatively low dose levels, and a
comprehensive suite of endpoints, individual cell necrosis in the liver in females was considered
for derivation of PODs. This effect was also observed in males but was accompanied by
inflammatory cell responses in the livers of female animals.
4.1.1.2 Immunological Effects
As reviewed in Section 3.4.2.4, evidence indicates that elevated exposures to PFOS are
associated with immunological effects in humans. As described in Table 3-10, the majority of
epidemiological studies assessed endpoints related to immunosuppression (1 high and 16
4-3
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
medium confidence studies) and immune hypersensitivity (1 high and 17 medium confidence
studies), while one study {medium confidence) also reported on endpoints related to autoimmune
disease. Endpoints related to autoimmune diseases were not further considered for quantitative
assessments as there were a limited number of medium and high confidence studies and a limited
number of total studies that assessed the same specific diseases (e.g., rheumatoid arthritis, celiac
disease). Endpoints related to immune hypersensitivity were also not considered for dose-
response analyses. Although the majority (6/9) of the available medium confidence studies
reported consistent increases in the odds of asthma, there were inconsistencies in effects reported
in subgroups across studies. These inconsistencies limited the confidence needed to select
particular studies and populations for dose-response modeling. Other immune hypersensitivity
endpoints, such as odds of allergies and rhinoconjunctivitis, had less consistent results reported
across medium and high confidence studies and were therefore excluded from further
consideration, though they are supportive of an association between PFOS and altered immune
function.
Evidence of immunosuppression in children reported by epidemiological studies were consistent
across studies and endpoints. Specifically, epidemiological studies reported reduced humoral
immune response to routine childhood immunizations, including lower levels of tetanus and anti-
diphtheria antibodies {Timmerman, 2021, 9416315; Grandjean, 2012, 1248827; Budtz-
J0rgensen, 2018, 5083631} and rubella {Granum, 2013, 1937228; Pilkerton, 2018, 5080265;
Stein, 2016, 3108691} antibody titers. Reductions in antibody response were observed at
multiple timepoints throughout childhood, using both prenatal and childhood exposure levels,
and were consistent across study populations from medium confidence studies.
Measurement of antigen-specific antibodies following vaccinations is an overall measure of the
ability of the immune system to respond to a challenge. The antigen-specific antibody response
is extremely useful for evaluating the entire cycle of adaptive immunity and is a sweeping
approach to detect immunosuppression across a range of cells and signals {Myers, 2018,
10473136}. The SAB's PFAS review panel noted that reduction in the level of antibodies
produced in response to a vaccine represents a failure of the immune system to respond to a
challenge and is considered an adverse immunological health outcome {U.S. EPA, 2022,
10476098}. This is in line with a review by Selgrade (2007, 736210) who suggested that specific
immunotoxic effects observed in children may be broadly indicative of developmental
immunosuppression impacting these children's ability to protect against a range of immune
hazards—which has the potential to be a more adverse effect that just a single immunotoxic
effect. Thus, decrements in the ability to maintain effective levels of antitoxins following
immunization may be indicative of wider immunosuppression in these children exposed to
PFOS.
As noted by Dewitt et al. (2017, 5926400; 2019, 5080663) as well as subject matter experts on
the SAB's PFAS review panel {U.S. EPA, 2022, 10476098}, the clinical manifestation of a
disease is not a prerequisite for a chemical to be classified as an immunotoxic agent and the
ability to measure clinical outcomes as a result of mild to moderate immunosuppression from
exposure to chemicals in traditional epidemiological studies can be challenging. Specifically, the
SAB noted that "[djecreased antibody responses to vaccines is relevant to clinical health
outcomes and likely to be predictive of risk of disease" {U.S. EPA, 2022, 10476098}. The WHO
Guidance for immunotoxicity risk assessment for chemicals similarly recommends measures of
4-4
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
vaccine response as a measure of immune effects as "childhood vaccine failures represent a
significant public health concern" {WHO, 2012, 10633091}. This response is also translatable
across multiple species, including rodents and humans, and extensive historical data indicate that
suppression of antigen-specific antibody responses by exogenous agents is predictive of
immunotoxicity.
When immunosuppression occurs in the developing immune system, the risks of developing
infectious diseases and other immunosuppression-linked diseases may increase {Dietert, 2010,
644213}. Immunosuppression linked with chemical stressors is not the same as an
immunodeficiency associated with, for example, genetic-based diseases, but still is an endpoint
associated with potential health risks. Studies of individuals exposed at the extremes of age,
those with existing immunodeficiencies, and those exposed to chronic stress, show that what
may be considered mild to moderate immunosuppression in the general population could result
in increased risk of infections in these more susceptible populations {Selgrade, 2007, 736210}.
Finally, the immune system continues developing after birth; because of this continued
development, exposures to PFAS may have serious and long-lasting consequences {DeWitt,
2019, 5080663; MacGillivray, 2014, 6749084; Selgrade, 2007, 736210}. Hessel et al. (2015,
5750707) reviewed the effect of exposure to nine toxicants on the developing immune system
and found that the developing immune system was at least as sensitive or more sensitive than the
general (developmental) toxicity parameters. Immunotoxicity that occurs in the developing
organism generally occurs at doses lower than required to affect the adult immune system, thus
providing a more sensitive endpoint for assessing risk {vonderEmbse, 2018, 6741321}. Luster et
al. (2005, 2174509) similarly noted that responses to childhood vaccines may be sensitive
enough to detect changes in populations with moderate degrees of immunosuppression, such as
those exposed to an immunotoxic agent.
Results reported in animal toxicological studies are consistent with the observed
immunosuppression in epidemiological studies. Specifically, studies in rodents found that oral
PFOS treatment resulted in reduced immune responses (e.g., reduced plaque-forming cell (PFC)
responses, reduced natural killer (NK) cell activity) (4 medium confidence studies) and altered
immune cell populations (e.g., bone marrow hypocellularity, altered splenic and thymic
cellularity, white blood cell counts) (2 high and 3 medium confidence studies). EPA prioritized
endpoints from both categories for quantitative analyses for several reasons. First,
immunosuppression evidenced by functional assessments of the immune responses, such as
analyses of PFC and NK responses, are coherent decreased antibody responses seen in human
populations. EPA prioritized PFC responses over NK cell activity for POD derivation because
several studies {Dong, 2009, 1424951; Peden-Adams, 2008, 1424797; Zhong, 2016, 3748828}
reported non-monotonic dose-response curves for NK cell activity. Second, altered immune cell
populations were reported in two high confidence studies and supported by several medium
confidence studies, strengthening the weight of evidence for these endpoints. EPA prioritized
results from NTP (2019, 5400978) as this was a high confidence study reporting consistent
effects of PFOS treatment on multiple endpoints related to immune cellularity (i.e., increased
bone marrow hypocellularity, increased splenic extramedullary hematopoiesis, and reduced
leukocytes, neutrophils, and white blood cell counts) in male rats.
Two medium confidence epidemiologic studies {Budtz-j0rgensen, 2018, 5083631; Timmerman,
2021, 9416315} and one high and one medium confidence animal toxicological studies {Zhong,
4-5
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
2016, 3748828; NTP, 2019, 5400978} were considered for POD derivation (Table 4-1). The
candidate epidemiological studies offer data characterizing antibody responses to vaccinations in
children using a variety of PFOS exposure measures across various populations and
vaccinations. Budtz-j0rgensen and Grandjean (2018, 5083631) investigated anti-tetanus and
anti-diphtheria responses in Faroese children aged 5-7 and Timmerman et al. (2021, 9416315)
investigated anti-tetanus and anti-diphtheria responses in Greenlandic children aged 7-12. In
addition to the results from epidemiological studies, altered PFC response in male PNW 4 mice
gestationally exposed to PFOS from GD 1-17 reported in Zhong et al. (2016, 3748828) and
extramedullary hematopoiesis in male and female rats gavaged with PFOS for 28 days reported
in NTP (2019, 5400978), supported the evidence of immunotoxicity in humans and were also
considered for POD derivation.
4.1.1.3 Cardiovascular effects
As reviewed in Section 3.4.3.4, evidence indicates that elevated exposures to PFOS are
associated with cardiovascular effects in humans. As described in Table 3-11, the majority of
epidemiological studies assessed endpoints related to serum lipids (2 high and 20 medium
confidence studies) and blood pressure and hypertension (2 high and 16 medium confidence
studies), while several studies also reported on cardiovascular disease (1 high and 4 medium
confidence studies) and atherosclerosis (1 high and 4 medium confidence studies). Endpoints
related to cardiovascular disease and atherosclerosis were not prioritized for dose-response as
they reported mixed or primarily null results. Endpoints related to blood pressure and
hypertension were also not prioritized for quantitative analyses because studies reported no
effects or generally mixed associations, though there was evidence of associations between
PFOS exposure and at least one measure of continuous blood pressure in adults (5 medium or
high confidence studies reported positive associations). There is some uncertainty associated
with the blood pressure endpoints as there was not often concordance between SBP and DBP
within study populations. However, these results are supportive of an association between PFOS
and cardiovascular effects in humans.
Studies in adults from the general population, including high-exposure communities, reported
positive associations between PFOS serum concentrations and serum lipids. Specifically,
medium confidence epidemiological studies in the general population reported positive
associations between PFOS exposure and total cholesterol (TC) (9/10 studies) and low-density
lipoprotein (LDL) (6/6 studies). Associations between PFOS and high-density lipoprotein (HDL)
or triglycerides in the general population were inconsistent. EPA prioritized TC for quantitative
assessments because the association was consistently positive in adults, with some studies
reporting statistically significant ORs, the response was more consistently positive with a greater
magnitude of change in other populations (i.e., children and pregnant women) compared to LDL,
and elevations in TC were reported in a marginally larger number of studies. Additionally, the
positive associations with TC were supported by a recent meta-analysis restricted to 14 general
population studies in adults {U.S. EPA, 2022, 10369698}.
Increased serum cholesterol is associated with changes in incidence of cardiovascular disease
events such as myocardial infarction (MI, i.e., heart attack), ischemic stroke (IS), and
cardiovascular mortality occurring in populations without prior CVD events {D'Agostino, 2008,
10694408; Goff, 2014, 3121148; Lloyd-Jones, 2017, 10694407}. Additionally, disturbances in
cholesterol homeostasis contribute to the pathology of non-alcoholic fatty liver disease (NAFLD)
4-6
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
and to accumulation of lipids in hepatocytes {Malhotra, 2020, 10442471}. Cholesterol is made
and metabolized in the liver, and thus the evidence indicating that PFOS exposure disrupts lipid
metabolism, suggests that toxic disruptions of lipid metabolism by PFOS are indications of
hepatoxicity. Associations between PFOS and other serum lipids (i.e., TG and HDL) were less
certain, though there was some evidence of positive associations with blood pressure and
hypertension in adults.
Though results reported in animal toxicological studies support the alterations in lipid
metabolism observed in epidemiological studies, variations in the direction of effect with dose
increases the uncertainty of the biological relevance of these responses in rodents to humans.
Additionally, the available mechanistic data does not help to explain the non-monotonicity of
serum lipid levels and decreased serum lipid levels at higher PFOS dose levels in rodents
(Section 3.4.3.3). EPA did not derive PODs for animal toxicological studies reporting
cardiovascular effects, such as altered serum lipid levels, due to uncertainties about the human
relevance of these responses.
Three medium confidence epidemiologic studies were considered for POD derivation (Table 4-1)
{Dong, 2019, 5080195; Lin, 2019, 5187597; Steenland, 2009, 1291109}. These candidate
studies offer a variety of PFOS exposure measures across various populations. Dong et al. (2019,
5080195) investigated the NHANES population (2003-2014), while Steenland et al. (2009,
1291109) investigated effects in a high-exposure community (the C8 Health Project study
population). Lin et al. (2019, 5187597) collected data from prediabetic adults from the Diabetes
Prevention Program (DPP) and DPP Outcomes Study at baseline (1996-1999). Dong et al.
(2019, 5080195) and Steenland et al. (2009, 1291109) excluded individuals prescribed
cholesterol medication from their analyses, a potential confounder for the total cholesterol
endpoint, whereas Lin et al. (2019, 5187597) did not.
4.1.1.4 Developmental effects
As reviewed in Section 3.4.4.4, evidence indicates that elevated exposures to PFOS are
associated with developmental effects in humans. As described in Table 3-12, the majority of
epidemiological studies assessed endpoints related to fetal growth restriction (20 high and 13
medium confidence studies) and gestational duration (10 high and 5 medium confidence studies),
while several studies also reported on endpoints related to fetal loss (2 high and 2 medium
confidence studies) and birth defects (3 medium confidence studies). Findings from the small
number of studies reporting on birth defects were mixed and generally limited in terms of the
number of studies reporting specific effects and therefore were not prioritized for quantitative
assessments. Although half of the available high and medium confidence studies reported
increased incidence of fetal loss (2/4), EPA did not prioritize this endpoint for dose-response
analyses as there were a relatively limited number of studies compared to endpoints related to
gestational duration and fetal growth restriction and the evidence from high confidence studies
was mixed. The impacts observed on fetal loss are supportive of an association between PFOS
exposure and adverse developmental effects.
Approximately half of the available studies reporting metrics of gestational duration observed
increased risk associated with PFOS exposure. Seven of the thirteen medium or high confidence
studies reported adverse effects on gestational age at birth and seven of the eleven medium or
high confidence studies reported an association with preterm birth. There were generally
4-7
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
consistent associations with adverse effects on preterm birth, particularly from the high
confidence studies, with several studies reporting statistically significant results. However,
several studies did not report exposure-response relationships and no definitive patterns or
explanations were seen based on study characteristics that would help to explain why some
studies reported associations while others did not. Results for gestational age were also relatively
consistent but there was similar uncertainty due to a lack of rationale explaining inconsistent
responses between studies. While overall there appears to be associations between PFOS
exposure and gestational duration, the inconsistencies in the database reduce the level of
confidence in the responses preferred for endpoints prioritized for dose-response modeling.
The adverse effects on gestational duration were consistent with effects on fetal growth
restriction. The majority of high and medium confidence epidemiological studies (16/27)
reported associations between PFOS and decreased mean birth weight in infants. Studies on
changes in standardized birth weight measures (i.e., z-scores) also generally reported inverse
associations (8/12 studies; 6 high and 2 medium confidence). Low birth weight is clinically
defined as birth weight less than 2,500 g (approximately 5.8 lbs.) and can include babies born
small for gestational age (SGA; birth weight below the 10th percentile for gestational age, sex,
and parity) {JAMA, 2002, 10473200; Mclntire, 1999, 15310; U.S. EPA, 2013, 4158459}. Low
birth weight is widely considered a useful measure of public health {Cutland, 2017, 10473225;
Lira, 1996, 10473966; Vilanova, 2019, 10474271; WHO, 2004, 10473140} and is on the World
Health Organization's (WHO's) global reference list of core health indicators {WHO, 2014,
10473141; WHO, 2018, 10473143}.
Substantial evidence links low birth weight to a variety of adverse health outcomes at various
stages of life. It has been shown to predict prenatal mortality and morbidity {Cutland, 2017,
10473225; U.S. EPA, 2013, 4158459; WHO, 2014, 10473141} and is a leading cause of infant
mortality in the United States {CDC, 2020, 10473144}. Low-birth-weight infants are also more
likely to have underdeveloped and/or improperly functioning organ systems (e.g., respiratory,
hepatic, cardiovascular), clinical manifestations of which can include breathing problems, red
blood cell disorders (e.g., anemia), and heart failure {Guyatt, 2004, 10473298; JAMA, 2002,
10473200; U.S. EPA, 2013, 4158459; WHO, 2004, 10473140; Zeleke, 2012, 10474317}.
Additionally, low-birth-weight infants evaluated at 18 to 22 months of age demonstrated
impaired mental development {Laptook, 2005, 3116555}.
Low birth weight is also associated with increased risk for diseases in adulthood, including
obesity, diabetes, and cardiovascular disease {Gluckman, 2008, 10473269; Osmond, 2000,
3421656; Risnes, 2011, 2738398; Smith, 2016, 10474151; Ong, 2002, 10474127, as reported in
Yang et al. (2022, 10176603). Poor academic performance, cognitive difficulties {Hack, 2002,
3116212; Larroque, 2001, 10473940}, and depression {Loret de Mola, 2014, 10473992} in
adulthood have also been linked to low birth weight. These associations between low birth
weight and infant mortality, childhood disease, and adult disease establish low birth weight as an
adverse effect. Given the known consequences of this effect, as well as the consistency of the
database and large number of high confidence studies reporting statistically significant odds of
this effect, the endpoint of low birth weight in humans was considered for dose-response
modeling.
Results reported in animal toxicological studies are consistent with the observed developmental
toxicity in epidemiological studies. Specifically, studies in rodents found that gestational PFOS
4-8
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
treatment resulted in reduced offspring weight (8/14 medium confidence studies), decreased
offspring survival (5/9 medium confidence studies), and altered maternal weight (6/12 medium
confidence studies). Though limited in number, several other studies also reported consistent
effects on placental endpoints, reduced ossification, and developmental delays.
Given the large number of adverse effects identified in the animal toxicological database for the
developmental health outcome, EPA considered only the most sensitive effects in pups supported
by multiple studies for derivation of PODs. EPA focused on the animal studies with effects in the
offspring, as opposed to maternal effects, because these effects provide concordance with the
approximate timing of low birth weight observed in human infants. The one study reporting
altered maternal weight without confounding effects on the offspring {Argus, 2000, 5080012}
could not be considered for derivation of a POD because the study was in rabbits and the
pharmacokinetic model EPA used to predict internal dose in the animal models is parameterized
for mice, rats and monkeys but rabbits. EPA also focused on endpoints for which multiple
studies corroborated the observed effect, thereby increasing the confidence in that effect.
Multiple animal toxicological studies observed effects at low dose levels and demonstrated a
dose-related response in pups for decreased fetal and pup body weight and decreased offspring
survival. EPA also focused on studies with exposure durations lasting through the majority of
gestation and/or lactation (i.e., from GD 1 until postnatal development) rather than those that
targeted a specific period of gestation as they were more likely to be sensitive for detection of
developmental effects. Overall, the developmental effects seen in the offspring of rodents treated
with PFOS are supportive of low birth weight and potential consequences of low birth weight
observed in human populations.
Six high confidence epidemiologic studies {Chu, 2020, 6315711; Darrow, 2013, 2850966;
Sagiv, 2018, 4238410; Starling, 2017, 3858473; Wikstrom, 2020, 6311677; Yao, 2021,
9960202} and 2 medium confidence animal toxicological studies {Lee, 2015, 2851075; Luebker,
2005, 757857} were considered for POD derivation (Table 4-1). The candidate epidemiological
studies offer a variety of PFOS exposure measures across the fetal and neonatal window. All
studies reported their exposure metric in units of ng/mL and reported the P
coefficients per ng/mL or ln(ng/mL), along with 95% confidence intervals, estimated from linear
regression models. Given the consistency of effects on offspring weight and survival across
studies and species, decreased pup body weight at LD 5 from a reproductive study (exposure to
dams from 42 days prior to mating until LD 5) as reported by Luebker et al. (2005, 757857), and
increased fetal death and decreased fetal weight in offspring exposed to PFOS from GD 11-16
reported by Lee et al. (2015, 2851075) were considered for the derivation of PODs.
Table 4-1 summarizes the studies and endpoints considered for POD derivation.
4-9
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 4-1. Summary of Endpoints and Studies Considered for Dose-Response Modeling and Derivation of Points of Departure
for All Effects in Humans and Rodents
Endpoint
Reference,
Confidence
Strain/
Species/Sex
POD Derived?
Notes
Immune Effects
Reduced Antibody
Concentrations for
Diphtheria and
Tetanus
Budtz-Jorgensen and Human, male
Grandj ean (2018, and female
508363 l)a children
Medium
Timmerman et al.
(2021, 9416315)
Medium
Yes
. Decreases in childhood antibody responses to pathogens such as diphtheria
and tetanus were observed at multiple timepoints in childhood, using both
prenatal and childhood exposure levels. Effect was large in magnitude and
generally coherent with epidemiological evidence for other antibody effects
Reduced Antibody
Concentrations for
Rubella
Granum et al. (2013,
1937228)
Medium
Human (male
and female
children)
No
Effect was large in magnitude and generally coherent with epidemiological
evidence for other antibody effects, however, the data were not suitable for
application of a BMR of 1 SD and Vi SD to provide a reasonably good estimate
of 10% and 5% extra risk. The Benchmark Dose Technical Guidance {U.S.
EPA, 2012, 1239433} explains that in a control population where 1.4%are
considered to be at risk of having an adverse effect, a downward shift in the
control mean of one SD results in about 10% extra risk of being at risk of
having an adverse effect. The cut off value of 0.1 IU/mL resulted in 0.003% of
the control population at risk of having an adverse effect, a value much smaller
than 1.4% which in turn did not result in 10% extra risk (see PFOS Appendix).
Decreased Plaque
Forming Cell (PFC)
Response to SRBC
Zhong et al. (2016,
3748828)
Medium
C57BL/6 Mice,
Fi males
Yes
Indicative of immunosuppression. Effect was consistently observed across
multiple studies: Peden-Adams et al. (2008, 1424797), Dong et al. (2009,
1424951), Zheng et al. (2009, 1429960), and Keil et al. (2008, 1332422). Zhong
et al. (2016, 3748828) was selected because the study tested a relatively low
dose range and the effect was measured in a sensitive lifestage and time point
(pups at PNW 4).
Extramedullary NTP(2019,
Hematopoiesis in the 5400978)
Spleen High
Sprague-
Dawley Rats,
male and female
Yes
Blood cell production outside of the bone marrow which occurs when normal
cell production is impaired. Selected for POD derivation because the results
were from a high confidence study, histopathologically confirmed, consistent
across both sexes, accompanied by evidence of bone marrow hypocellularity,
and consistent with other studies that reported alterations in circulating immune
cells, splenic cellularity, and thymic cellularity.
Developmental Effects
4-10
-------
DRAFT FOR PUBLIC COMMENT MARCH 2023
„ , . , Reference, Strain/
Endpoint „ POD Derived?
Confidence Species/Sex
Decreased Birth Chu et al. (2020,
Weight 6315711)
High
Darrow et al. (2013.
2850966)
High
Sagiv et al. (2018,
4238410)
High
Starling et al. (2017
3858473)
High
Wikstrom et al.
(2020, 6311677)
High
Yao et al. (2021,
9960202)
High
Decreased Fetal Lee et al. (2015, CD-I Mice, Fi Yes
Body Weight 2851075) males and
Medium females
Decreased Pup Body Luebker et al. (2005, Sprague- Yes
Weight 757857) Dawley Rats, Fi
Medium male and female
Increased Number of Lee etal. (2015, CD-I Mice, Yes
Dead Fetuses 2851075) females
Medium
Notes
Evidence for developmental effects is based on consistent adverse effects for
FGR including birthweight measures which are the most accurate endpoint.
Some deficits were consistently reported for birth weight and standardized birth
weight in many high and medium confidence cohort studies. Effect was
generally large in magnitude and coherent with epidemiological evidence for
other biologically related effects.
Effect was consistently observed across multiple studies and species {Argus,
2000, 5080012; Li, 2016, 3981495} and is coherent with epidemiological
evidence of low birth weight. Lee et al. (2015, 2851075) was selected because
there is a pharmacokinetic model available to extrapolate from exposures in
mice to exposures in humans, the study tested a relatively low dose range, and
mice appear to be a more sensitive model for this endpoint than rats.
Effect was consistently observed across multiple studies and species {Luebker,
2005, 1276160; Lau, 2003, 757854} and is coherent with epidemiological
evidence of low birth weight. Luebker et al. (2005, 757857) was selected
because the study tested a relatively large number of dose groups and a low
dose range. This study was previously selected as the overall RfD for PFOS in
the 2016 HESD {U.S. EPA, 2016, 3603365}.
Decreased offspring survival was consistently observed across multiple studies
and species and is also consistent with other developmental effects related to
survival observed in rodents {Luebker, 2005, 1276160; Argus, 2000, 5080012;
Lau, 2003, 757854; Luebker, 2005, 757857}. Lee et al. (2015, 2851075) was
selected because there is a pharmacokinetic model available to extrapolate from
exposures in mice to exposures in humans, the study tested a relatively low dose
range, and mice appear to be a more sensitive model for this endpoint than rats.
Human, male Yes
and female
infants
4-11
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Endpoint
Reference,
Confidence
Strain/
Species/Sex
POD Derived?
Notes
Serum Lipid Effects
Increased Total
Cholesterol
Dong et al. (2019,
5080195)
Medium
Lin et al. (2019,
5187597)
Medium
Steenland et al.
(2009, 1291109)b
Medium
Human, male
and female
Yes
Effect supported by an association in PFOS and blood pressure from the
epidemiological studies. Effect observed in studies designed to exclude
individuals prescribed cholesterol medication, minimizing concerns of bias due
to medical intervention {Dong, 2019, 5080195; Steenland, 2009, 1291109}.
Hepatic Effects
Increased ALT
Galloetal. (2012,
1276142)
Medium
Nian et al. (2019,
5080307)
Medium
Human (male
and female
adults)
Yes
Effect was consistent and observed across multiple populations including
general population adults {Lin, 2010, 1291111} (NHANES) and high-exposure
communities {Gallo, 2012, 1276142} (C8 Health Project); Nian, 2019,
5080307} (Isomers of C8 Health Project in China)
Increased ALT
Linetal. (2010,
1291111)
Medium
Human (male
and female
adults)
No
While this is a large nationally representative population, several
methodological limitations preclude its use for POD derivation. Limitations
include lack of clarity about base of logarithmic transformation applied to
PFOS concentrations in regression models, and the choice to model ALT as an
untransformed variable, a departure from the typically lognormality assumed in
most of the ALT literature.
Individual Cell Butenhoff et al.
Necrosis in the Liver (2012,1276144)
High
Sprague-
Dawley rats,
females
Yes
Effect was supported by a similar response in males from the same study
{Butenhoff, 2012, 1276144}. Effect was accompanied by a hepatic
inflammatory cell response in females. Effect was qualitatively observed in
Xing et al. (2016, 3981506) and Cui et al. (2009, 757868), and further
supported by increases in serum enzyme levels associated with hepatic damage
in both animals and humans.
Notes: PNW = postnatal week; ALT = alanine transaminase; Fi =first generation.
a Supported by Grandjean et al. (2012, 1248827); Grandjean et al. (2017, 3858518); Grandjean et al. (2017, 4239492).
b See Section 6.6.3 for discussion on the approach to estimating BMDs from regression coefficients.
4-12
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
4.1.2 Estimation or Selection of Points of Departure (PODs)
for RfD Derivation
Consistent with EPA's Benchmark Dose Technical Guidance {U.S. EPA, 2012, 1239433}, the
BMD and 95% lower confidence limit on the BMD (BMDL) were estimated using a BMR
intended to represent a minimal, biologically significant level of change. The Benchmark Dose
Technical Guidance {U.S. EPA, 2012, 1239433} describes a hierarchy by which BMRs are
selected, with the first and preferred approach being the use of a biological or toxicological basis
to define what minimal level of response or change is biologically significant. If that biological
or toxicological information is lacking, the guidance document recommends BMRs that could be
used in the absence of information about a minimal clinical or biological level of change
considered to be adverse—specifically, a BMR of one standard deviation (SD) change from the
control mean for continuous data or a BMR of 10% extra risk for dichotomous data. When
severe or frank effects are modeled, a lower BMR can be adopted. For example, developmental
effects are frequently serious effects, and the Benchmark Dose Technical Guidance suggests that
studies of developmental effects can support lower BMRs. BMDs for these effects may employ a
BMR of 0.5 SD change from the control mean for continuous data or a BMR of 5% for
dichotomous data {U.S. EPA, 2012, 1239433}. A lower BMR can also be used if it can be
justified on a biological and/or statistical basis. The Benchmark Dose Technical Guidance (page
23; {U.S. EPA, 2012, 1239433}) shows that in a control population where 1.4% are considered
to be at risk of having an adverse effect, a downward shift in the control mean of one SD results
in a -10%) extra risk of being at risk of having an adverse effect. A BMR smaller than 0.5 SD
change from the control mean is generally used for severe effects (e.g., 1% extra risk of cancer
mortality).
Based on rationales described in EPA's Benchmark Dose Technical Guidance {U.S. EPA, 2012,
1239433}, the IRIS Handbook {U.S. EPA, 2022, 10367891} and past IRIS assessment
precedent, BMRs were selected for dose-response modeling of PFOS-induced health effects for
individual study endpoints as described below and summarized in Table 4-2 along with the
rationales for their selection. For this assessment, EPA took statistical and biological
considerations into account to select the BMR. For dichotomous responses, the general approach
was to use 10% extra risk as the BMR for borderline or minimally adverse effects and either 5%
or 1% extra risk for adverse effects, with 1% reserved for the most severe effects. For continuous
responses, the preferred approach for defining the BMR was to use a preestablished cutoff for the
minimal level of change in the endpoint at which the effect is generally considered to become
biologically significant (e.g., greater than or equal to 42 IU/L serum ALT in human males
{Valenti, 2021, 10369689}) In the absence of an established cutoff, a BMR of 1 SD change from
the control mean, or 0.5 SD for effects considered to be severe, was generally selected. Specific
considerations for BMR selection for endpoints under each of the priority non-cancer health
outcomes are described in the subsections below. Considerations for BMR selection for cancer
endpoints are described in Section 4.2.
4.1.2.1 Hepatic Effects
Modeling elevated human ALT used cutoff levels of 42 IU/L for males and 30 IU/L for females,
based on the most recent sex-specific upper reference limits {Valenti, 2021, 10369689}. The
baseline prevalence of elevated ALT is estimated as 14% and 13% in U.S. male and female
4-13
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
adults (aged 20 and older), respectively (see PFOS Appendix). Therefore, the BMR was defined
as a 5% increase in the number of people with ALT values above the cutoffs. Although the
Benchmark Dose Technical Guidance {U.S. EPA, 2012, 1239433} recommends a BMR of 10%
extra risk for dichotomous data when biological information is not sufficient to identify the
BMR, in this situation, such a BMR would result in a doubling of risk.
For the adverse effect of individual cell necrosis observed in livers of rats following PFOS
exposure, there is currently inadequate available biological or toxicological information to permit
determination of a minimal biologically significant response level. Therefore, in accordance with
EPA's Benchmark Dose Technical Guidance {U.S. EPA, 2012, 1239433}, a BMR of 10% extra
risk was used (dichotomous data; see Table 4-2).
4.1.2.2 Immune Effects
For the developmental immune endpoint of decreased diphtheria and tetanus antibody response
in children found to be associated with PFOS exposure, the BMD and the BMDL were estimated
using a BMR of 0.5 SD change from the control mean (see Table 4-2). Consistent with EPA's
Benchmark Dose Technical Guidance {U.S. EPA, 2012, 1239433}, EPA typically selects a 5%
or 0.5 standard deviation (SD) benchmark response (BMR) when performing dose response
modeling of data from an endpoint resulting from developmental exposure. Because Budtz-
J0rgensen and Grandjean (2018, 5083631) and Timmerman et al. (2021, 9416315) assessed
antibody response after PFAS exposure during gestation and childhood, these are considered
developmental studies {U.S. EPA, 1991, 732120} based on EPA's Guidelines for Developmental
Toxicity Risk Assessment, which includes the following definition:
"Developmental toxicology - The study of adverse effects on the developing
organism that may result from exposure prior to conception (either parent), during
prenatal development, or postnatally to the time of sexual maturation. Adverse
developmental effects may be detected at any point in the lifespan of the
organism."
EPA guidance recommends the use of a 1 or 0.5 SD change in cases where there is no accepted
definition of an adverse level of change or clinical cut-off for the health outcome {U.S. EPA,
2012, 1239433}. A 0.5 SD was selected since the health outcome is developmental and there is
no accepted definition of an adverse level of change or clinical cut-off for reduced antibody
concentrations in response to vaccination. Therefore, EPA performed the BMDL modeling using
a BMR equivalent to a 0.5 SD change in log2-transformed antibody concentrations, as opposed
to a fixed change in the antibody concentration distributions {U.S. EPA, 2012, 1239433}.
For the adverse effects of decreased plaque forming cell (PFC) response to SRBC observed in
mice and splenic extramedullary hematopoiesis in rats following PFOS exposure, there is
currently inadequate available biological or toxicological information to permit determination of
minimal biologically significant response levels. Therefore, in accordance with EPA's
Benchmark Dose Technical Guidance {U.S. EPA, 2012, 1239433}, a BMR of 1 SD change from
the control mean was employed for the effect on PFC response (continuous data) and a BMR of
10%) extra risk was used for the increased incidence of extramedullary hematopoiesis
(dichotomous data) (see Table 4-2).
4-14
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
4.1.2.3 Cardiovascular Effects
Modeling human cholesterol used a cutoff level of 240 mg/dL for elevated serum total
cholesterol, consistent with the American Heart Association's definition of hypercholesterolemia
{NCHS, 2019, 10369680}. Recent data (for years 2015-2018) show that the percentage of U.S.
adults aged 20 and older with total cholesterol >240 mg/dL is 11.5% {NCHS, 2019, 10369680}.
Therefore, the BMR was defined as a 5% increase in the number of people with total cholesterol
values above 240 mg/dL. Although the Benchmark Dose Technical Guidance {U.S. EPA, 2012,
1239433} recommends a BMR of 10% extra risk for dichotomous data when biological
information is not sufficient to identify the BMR, in this situation, such a BMR would result in a
doubling of risk.
4.1.2.4 Developmental Effects
For the developmental endpoint of decreased birth weight in infants associated with PFOS
exposure, the BMD and the BMDL were estimated using a BMR of 5% extra risk, given that this
level of response is typically used when modeling developmental responses from animal
toxicology studies, and that low birthweight confers increased risk for adverse health effects
throughout life {Hack, 1995, 8632216; Reyes, 2005, 1065677; Tian, 2019, 8632212}. Low birth
weight is clinically defined as birth weight less than 2,500 g (approximately 5.8 lbs) and can
include babies born SGA (birth weight below the 10th percentile for gestational age, sex, and
parity) {JAMA, 2002, 10473200; Mclntire, 1999, 15310; U.S. EPA, 2013, 4158459}.
For decreased fetal and pup weights and decreased pup survival observed in animal studies, a
BMR of 5% relative deviation and 0.5 SD from the control was employed, respectively (see
Table 4-2). This is consistent with EPA's Benchmark Dose Technical Guidance {U.S. EPA,
2012, 1239433} and the IRIS Handbook {U.S. EPA, 2022, 10367891}, which note that studies
of adverse developmental effects represent a susceptible lifestage and can support BMRs that are
lower than 10% extra risk (dichotomous data) and 1 SD change from the control mean
(continuous data).
A 5% relative deviation in markers of growth in gestational exposure studies (e.g., fetal weight)
that do not lead to death has generally been considered an appropriate biologically significant
response level and has been used as the BMR in final IRIS assessments (e.g., U.S. EPA (2003,
1290574), U.S. EPA (2004, 198783), and U.S. EPA (2012, 3114808)). Additionally, the 5%
BMR selection is statistically supported by data which compared a BMR of 5% relative
deviation for decreased fetal weight to NOAELs and other BMR measurements, including 0.5
standard deviation, and found they were statistically similar {Kavlock, 1995, 75837}.
Table 4-2. Benchmark Response Levels Selected for BMD Modeling of Health Outcomes
Endpoint BMR Rationale
Immune Effects
Reduced antibody concentrations for 0.5 SD Consistent with EPA guidance. EPA typically selects a 5% or
diphtheria and tetanus in children 0.5 standard deviation (SD) benchmark response (BMR) when
(developmental immune endpoint) performing dose response modeling of data from an endpoint
resulting from developmental exposure and selects a 1 or 0.5 SD
change in cases where there is no accepted definition of an
4-15
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Endpoint
BMR
Rationale
Decreased Plaque Forming Cell
(PFC) Response to SRBC
adverse level of change or clinical cut-off for the health outcome
{U.S. EPA, 2012, 1239433}
1 SD Insufficient information available to determine minimal
biologically significant response level. The available biological
or toxicological information does not allow for determination of
a minimal biologically significant response level for this adverse
effect, and so a BMR of one SD was used as per EPA guidance
{U.S. EPA, 2012, 1239433}
Extramedullary Hematopoiesis in the
Spleen
10% Insufficient information available to determine minimal
biologically significant response level. The available biological
or toxicological information does not allow for determination of
a minimal biologically significant response level for this adverse
effect, and so a BMR of 10% was used as per EPA guidance
{U.S. EPA, 2012, 1239433}
Developmental Effects
Decreased Birth Weight in Infants or 5%
Decreased Pup Body Weight in
Rodent Offspring
Increased Number of Dead Fetuses 0.5 SD
Consistent with EPA guidance. EPA typically selects a 5% or
0.5 standard deviation (SD) benchmark response (BMR) when
performing dose response modeling of data from an endpoint
resulting from developmental exposure {U.S. EPA, 2012,
1239433}
Consistent with EPA guidance. EPA typically selects a 5% or
0.5 standard deviation (SD) benchmark response (BMR) when
performing dose response modeling of data from an endpoint
resulting from developmental exposure {U.S. EPA, 2012,
1239433}
Serum Lipids
Increased Cholesterol
5% Response rate of 5% extra risk is reasonable, whereas a 10%
BMR would result in a doubling of risk. Although EPA's
Benchmark Dose Technical Guidance {U.S. EPA, 2012,
1239433 } recommends a BMR based on a 10% extra risk for
dichotomous endpoints when biological information is not
sufficient to identify the BMR, in this situation such a BMR
would result in a highly improbable doubling of risk.
Hepatic Effects
Increased ALT
Individual Cell Necrosis
5% Response rate of 5% extra risk is reasonable, whereas
a 10% BMR would result in a doubling of risk. Although
EPA's Benchmark Dose Technical Guidance {U.S. EPA, 2012,
1239433 } recommends a BMR based on a 10% extra risk for
dichotomous endpoints when biological information is not
sufficient to identify the BMR, in this situation such a BMR
would result in a highly improbable doubling of risk
10% Insufficient information available to determine minimal
biologically significant response level. The available biological
or toxicological information does not allow for determination of
a minimal biologically significant response level for this adverse
effect, and so a BMR of 10% was used as per EPA guidance
{U.S. EPA, 2012, 1239433}
Notes: ALT = alanine transaminase; BMD = benchmark dose; BMR = benchmark response; CDC = Centers for Disease Control;
SD = standard deviation.
4-16
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
4.1.3 Pharmacokinetic Modeling Approaches to Convert
Administered Dose to internal Dose in Animals and
Humans
4.1.3.1 Pharmacokinetic Model for Animal Internal Dosimetry
Following review of the available models in the literature, EPA chose the Wambaugh et al.
(2013, 2850932) model to describe PFOS dosimetry in experimental animals based on the
following criteria:
• availability of model parameters across the species of interest,
• agreement with out-of-sample datasets (see PFOS Appendix), and
• flexibility to implement life course modeling.
These criteria originated from the goal of accurately predicting internal dose metrics for
toxicology studies that were selected for dose-response analysis. These studies involved rats,
mice, and non-human primates, and these were the species of interest necessary to have available
model parameters. Good agreement with out-of-sample datasets shows that the model
performance is good compared to both the data used to identify model parameters and to external
data. This increases confidence that the model can be used to make accurate predictions of
internal dose metrics for the toxicology studies, which can also be seen as external. The ability to
implement life-course modeling was necessary to properly predict internal dose metrics for
developmental studies and endpoints as the animal transitioned through numerous life-stages.
In this case, an oral dosing version of the original model structure introduced by Andersen et al.
(2006, 818501) and summarized in Section 3.3.2 was selected for having the fewest number of
parameters that would need estimation. In addition, the Wambaugh et al. (2013, 2850932)
approach allowed for a single model structure to be used for all species in the toxicological
studies allowing for model consistency for the predicted dose metrics associated with LOAELs
and NOAELs from 13 animal toxicological studies of PFOS.
The Wambaugh et al. (2013, 2850932) model was selected for pharmacokinetic modeling for
animal internal dosimetry for several important reasons: 1) it allowed for sex-dependent
concentration-time predictions for PFOS across all three species of interest, 2) it adequately
predicted dosimetry of newer datasets published after model development, and 3) it was
amendable to addition of a life stage component for predicting developmental study designs.
These analyses are further described below. Uncertainties and limitations of the selected
modeling approach are described in Section 6.6.1.
4-17
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
4.1.3.1.1AnimaI Model Parameters
Table 4-3. PK Parameters from Wambaugh et al. (2013, 2850932) Meta-Analysis of Literature Data for PFOS
Parameter
Units
CD1 Mouse
(FT
CD1 Mouse
(M)a
Sprague- Dawley Rat
(FT
Sprague- Dawley Rat CynomolgusMonkey
(M)a (M/F)a
Body weightb (BW) kg
Cardiac Output0 (Qcc) L/h/kg°74
0.02
5.68
0.02
5.68
0.203
12.39
0.222
12.39
3.42
19.£
Absorption rate (ka)
1/h
1.16
433.4
4.65
0.836
132
(0.617-42,400)
(0.51-803.8)
(3.02-1,980)
(0.522-1.51)
(0.225-72,100)
Central Compartment L/kg
0.264
0.292
0.535
0.637
0.303
Volume (Vcc)
(0.24-0.286)
(0.268-0.317)
(0.49-0.581)
(0.593-0.68)
(0.289-0.314)
Intercompartment
1/h
0.0093
2,976
0.0124
0.00524
0.00292
transfer rate (ki2)
(2.63 x e~10-38,900)
(2.8 x e~10-
4.2 x e4)
(3.1 x e~10-46,800)
(2.86 x e~10-43,200)
(2.59 x e~10-34,500)
Intercompartment
Unitless
1.01
1.29
0.957
1.04
1.03
ratio (Rv2:V2i)
(0.251-4.06)
(0.24-4.09)
(0.238-3.62)
(0.256-4.01)
(0.256-4.05)
Maximum resorption
|imol/h
57.9
1.1 xe4
1,930
1.34 x e~6
15.5
rate (TmaXc)
(0.671-32,000)
(2.1-7.9 xe4)
(4.11-83,400)
(1.65 x e~10-44)
(0.764-4,680)
Renal resorption
|imol
0.0109
381
9.49
2.45
0.00594
affinity (KT)
(1.44 x e 5-1.45)
(2.6 x e~5-2.9 x e3)
(0.00626-11,100)
(4.88 x e~lcl-60,300)
(2.34 x e~5-0.0941)
Free fraction
Unitless
0.00963
0.012
0.00807
0.00193
0.0101
(0.00238-0.0372)
(0.0024-0.038)
(0.00203-0.0291)
(0.000954-0.00249)
(0.00265-0.04)
Filtrate flow rate
Unitless
0.439
27.59
0.0666
0.0122
0.198
(Qfiic)
(0.0125-307)
(0.012-283)
(0.0107-8.95)
(0.0101-0.025)
(0.012-50.5)
Filtrate volume (Vr,ic)
L/kg
0.00142
0.51
0.0185
0.000194
0.0534
(4.4 x e~10-6.2)
(3.5 x e~10-6.09)
(8.2 x e~7-7.34)
(1.48 x e 9-5.51)
(1.1 x e~7-8.52)
Notes: F = female; M = male.
Means and 95% credible intervals (in parentheses) from Bayesian analysis are reported. For some parameters the distributions are quite wide, indicating uncertainty in that
parameter (i.e., the predictions match the data equally well for a wide range of values).
aData sets modeled for the mouse and rat were from Chang et al. (2012, 1289832) and for the monkey from Seacat et al. (2002, 757853) and Chang et al. (2012,1289832).
b Average bodyweight for species:individual-specific bodyweights.
c Cardiac outputs obtained from Davies and Morris (1993,192570).
4-18
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
4.1.3.1.20ut-of-Sample Comparisons
To evaluate the model's ability to predict PFOS concentration-time data in the species of
interest, EPA compared model fits to in vivo datasets published following the 2016 HESD (Table
4-4). For rats, the data of Kim et al. (2016, 3749289) and Huang et al. (2019, 7410147) were
used. Model simulations demonstrated good agreement with available data for adult time-course
PFOS PK predictions in the rat. However, there was no comparable PK dataset for PFOS in
mice. Therefore, only the original study used for parameter determination {Chang, 2012,
1289832} was compared to model simulations. This comparison approach demonstrated
agreement with the in vivo data.
Using the Wambaugh et al. (2013, 2850932) model, EPA predicted the half-life, Vd, and
clearance and compared these species-specific predictions to values obtained from in vivo studies
when data were available.
Following out-of-sample dataset evaluation of the female rat PK parameters (Table 4-4) and
visual inspection of the resulting concentration-time fits, EPA determined that only male PK
model parameters would be used for all rat-specific modeling. This assumption agrees with Kim
et al. (2016, 3749289) where they report no PK differences between the sexes for PFOS ADME.
Table 4-4. Model Predicted and Literature PK Parameter Comparisons for PFOS
Male
Female
tl/2,P
Vd,p
CL
tl/2,P
Vd,p
CL
(days)
(L/kg)
(L/d/kg)
(days)
(L/kg)
(L/d/kg)
Rat
Model
44.13
0.638
0.01
282.05
0.538
0.0013
Literature
28.7a, 39.7b
0.3823, 0.681b
0.00923, 0.013b
24.8a- 32.8b
0.2883, 0.421b
0.008a, 0.009b
Mouse
Model
134.83
0.472
0.0024
38.4
1.41
0.0255
Literature
-
-
-
-
-
-
Notes: CL = clearance; PK = pharmacokinetic; ti/2,p = terminal-phase elimination half-life; Vd, P = volume of distribution during
the terminal phase.
a Information obtained from Kim et al. (2016, 3749289).
bInformation obtained from Huang et al. (2019, 5387170).
4.1.3.1.3Life Course Modeling
The Wambaugh et al. (2013, 2850932) model was modified to allow for a gestation, lactation,
and post-weaning phase (Figure 4-1). Using the original model structure and published
parameters, simulations assumed that dams were dosed prior to conceptions and up to the date of
parturition. Following parturition, a lactational phase involved PFOS transfer from the
breastmilk to the suckling pup where the pup was modeled with a simple one-compartment PK
model. Finally, a post-weaning phase utilized the body-weight scaled Wambaugh model to
simulate dosing to the growing pup and accounted for filtrate rate as a constant fraction of
cardiac output.
4-19
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Gestation Lactation Post-weaning
Figure 4-1. Model Structure for Life Stage Modeling
Model parameters for three-compartment model are the same as those described earlier. Pup-specific parameters include milk
consumption in kgmiik/day (Rmiik), infant-specific volume of distribution (Vd), and infant-specific half-life (ti/2).
This methodology was adapted from Kapraun et al. (2022, 9641977) and relies on the following
assumptions for gestation/lactation modeling:
• During gestation and up through the instant birth occurs, the ratio of the fetal
concentration (mg of substance per mL of tissue) to the maternal concentration is
constant.
• Infant animal growth during the lactational period is governed by the infant growth curves
outlined in Kapraun et al. (2022, 9641977).
• Rapid equilibrium between maternal serum PFOS and milk PFOS is assumed and
modeled using a serum:milk partition coefficient.
• All (100%) of the substance in the breast milk ingested by the offspring is absorbed by the
offspring.
• The elimination rate of the substance in offspring is proportional to the amount of
substance in the body and is characterized by an infant-specific half-life that is a fixed
constant for any given animal species as described in Table 4-5 below.
• Following the lactation period, infant time course concentrations are tracked using the
more physiologically-based Wambaugh model to model post-weaning exposure and infant
growth.
A simple one-compartment model for infant lactational exposure was chosen because of
differences between PFOS Vd reported in the literature and Wambaugh et al. (2013, 2850932)
model-predicted Vd following extrapolation to a relatively low infant body weights. Because Vd
is assumed to be extracellular water in humans, Goeden et al. (2019, 5080506) adjusts for life
stage-specific changes in extracellular water using an adjustment factor where infants have 2.1
times more extracellular water than adults resulting in a larger Vd. However, this large difference
in extracellular water is not observed in rats {Johanson, 1979, 9641334}. Johanson (1979,
9641334) demonstrated a 5% decrease in blood water content from early postnatal life (-0.5
weeks) to adulthood (> 7 weeks) in the rat. Therefore, EPA used the literature reported Vd {Kim,
2016, 3749289; Chang, 2012, 1289832} for the one compartment model to describe infant
4-20
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
toxicokinetics (Table 4-5). Finally, the Wambaugh et al. (2013, 2850932) model was not
parameterized for a post-partum infant, and it was not possible to evaluate the mechanistic
assumptions for renal elimination with postnatal toxicokinetic data. Therefore, the parameters
listed in Table 4-5 in a one-compartment gestation/lactation model were used in conjunction with
the parameters published in Wambaugh et al. (2013, 2850932) to predict developmental dose
metrics for PFOS.
Table 4-5. Additional PK Parameters for Gestation/Lactation for PFOS
Parameter
Units
Rat
Mouse
Maternal Milk:Blood Partition Coefficient (Pmiik) Unitless
0.13a
0.32e
Fetus:Mother Concentration Ratio (Rfm)
Unitless
0.83b
0.41f
Elimination Half-Life (ti/2)
Days
0
O
"t
36.878
Volume of Distribution (Vd)
L/kg
0.28d
0.268
Starting Milk Consumption Rate (r!,min:)
kgmiik/dav
0.001h
0.00011
Week 1 Milk Consumption Rate (rVik)
kgmiik/dav
0.003h
0.00031
Week 2 Milk Consumption Rate (r2miik)
kgmiik/dav
0.0054h
0.000541
Week 3 Milk Consumption Rate (r3miik)
kgmiik/dav
0.0059h
0.000591
Notes: PK = pharmacokinetic.
information obtained from Loccisano et al. (2013, 1326665) (derived from Kuklenyik et al. (2004, 1598132)).
bInformation obtained from Lau et al. (2003, 757854).
c Average of male/female half-lives reported in Huang et al. (2019, 5387170), Kim et al. (2016, 3749289), and Chang et al.
(2012,1289832).
information obtained from Kim et al. (2016, 3749289).
e Assume same Pmiik as PFOA (lack of mouse data).
fInformation obtained from Wan et al. (2020, 7174720).
information obtained from Chang et al. (2012,1289832).
information obtained from Kapraun et al. (2022, 9641977) (adapted from Lehmann et al. (2014, 2447276)).
1 Information obtained from Kapraun et al. (2022, 9641977) (mouse value is 10% of rat based on assumption that milk ingestion
rate is proportional to body mass).
These developmental-specific parameters include the maternal milk:blood PFOS partition
coefficient (Pmiik), the ratio of the concentrations in the fetus(es) and the mother during
pregnancy (Rfm), the species-specific in vivo determined half-life (ti/2) and Vd for PFOS, and the
species-specific milk consumption rate during lactation (r'miik) for the ith week of lactation. Milk
rate consumptions are defined as:
• r°miik, the starting milk consumption rate in kg milk per day (kg/d);
• ^miik, the (average) milk consumption rate (kg/d) during the first week of lactation (and
nursing);
• r2miik, the (average) milk consumption rate (kg/d) during the second week of lactation; and
• r3miik, the (average) milk consumption rate (kg/d) during the third week of lactation.
where Rmiik used in the model is a piecewise linear function comprising each r'miik depending on
the week of lactation.
Using this gestation/lactation model, EPA fit one study for PFOS exposure in rats to ensure the
model predicted the time-course concentration curves for both the dam and the pup. For all
gestation/lactation studies, time zero represents conception followed by a gestational window (21
days for the rat). Dosing prior to day zero represents pre-mating exposure to PFOS.
4-21
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Figure 4-2 demonstrates the model's ability to predict gestation/lactation study design in the rat
for dams exposed to 1.6 mg/kg/day PFOS giving birth to pups who are exposed through lactation
{Luebker, 2005, 1276160}. For developmental PK simulations, the original Wambaugh et al.
(2013, 2850932) model with increasing maternal weight predicts dam concentrations in female
rats while the one-compartmental lactational transfer model predicts infant concentrations for
pups exposed both in utero and through lactation only.
PFOS: 1.6 mg/kg/day
5 103
"t5i
£ 102
£ u
m c
Q o
un Iff3
O
£ 10"1
3 103
&
£
3c 10s
£l. o
u
lti
O
u_
a- 101
-40
-20
1
0
20
40
birth
-40
-20
0
20
40
Time [days]
Figure 4-2. Gestation/Lactation Predictions of PFOS in the Rat
Top panel represents predicted dam concentrations with open diamonds (0) representing the dam concentrations reported in
Luebker et al. (2005, 1276160). Bottom panel represents predicted pup concentrations with open diamonds (0) representing the
reported pup concentrations in Luebker et al. (2005, 1276160) where the source of PFOS exposure is from the breast milk.
Vertical dashed line represents birth.
The purpose of the animal PBPK model is to make predictions of internal dose in lab animals
used in toxicity studies and extrapolate these internal dose points-of-departure to humans.
Therefore, to evaluate its predictive utility for risk assessment, a number of dose-metrics across
life stages were selected for simulation in a mouse, rat, monkey, or human. Concentrations of
PFOS in blood were considered for all the dose-metrics. For studies in adult animals the dose-
metric options were generally a maximum blood concentration (Cmax, mg/L) and a time averaged
blood concentration (i.e., the area under the curve over the duration of the study (AUC,
mg * day/L)) or the blood concentration over the last 7 days of the study (Ciast7, mg/L). In
developmental studies, dose-metrics were developed for the dam, the fetus (during gestation),
and the pup (during lactation) for both time maximum blood concentrations (Cmax) and average
blood concentrations (Cavg). In the dam, the Cmax and Cavg were calculated over a range of life
stages: during gestation (Cavg dam gest), during lactation (Cavg damjact), or combined gestation and
lactation (Cavg dam gestjact). In pups for Cmax, two different life stages were calculated either during
gestation or lactation (Cmax pup gest, Cmax_puPjact). In pups for time averaged metrics, a Cavg was
calculated for during gestation, lactation or combined gestation and lactation (Cavg PuP gest,
Cavg pup Jact and Cavg pup gest lact).
4-22
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
4.1.3.2 Pharmacokinetic Model for Human Dosimetry
The key factors considered in model determination were to implement a human model from the
literature that was able to model gestational and lactational exposure to infants, that was able to
describe time course changes in serum concentration due to changes in bodyweight during
growth, and that required minimal new development. Previous modeling efforts suggested that
limiting model complexity helps to prevent errors and facilitates rapid implementation
{Bernstein, 2021, 9639956}. For the human and animal endpoints of interests, serum
concentration was identified as a suitable internal dosimetry target which provides support for
using a simpler model that did not have individual tissue dosimetry. For these reasons, EPA
selected the one compartment human developmental model published by Verner et al. (2016,
3299692). Several alternative models to EPA's updated version of the Verner et al. (2016,
3299692) model for the calculation of PODhed from an internal POD were considered. This
included consideration of full PBPK models (i.e., the Loccisano family of models {Loccisano,
2011, 787186; Loccisano, 2012, 1289830; Loccisano, 2012, 1289833; Loccisano, 2013,
1326665} and a developmental PBPK model in rats {Chou, 2021, 7542658}), as well as other
one-compartment PK models (e.g., Goeden et al. (2019, 5080506)). Discussion on the
justification for selection of the Verner et al. (2016, 3299629) model as the basis for the
pharmacokinetic modeling approach used for PFOS is available in Sections 6.6.2 and 6.7.
Several adjustments were undertaken to facilitate the application of the model to our use. First,
the model was converted from acslX language to an R/MCSim framework. This allows for the
code to be more accessible to others by updating it to a contemporary modeling language, as
acslX software is no longer available or supported. The starting point for the conversion to
R/MCSim was another model with a similar structure that was in development by EPA at that
time {Kapraun, 2022, 9641977}. Second, body weight curves for non-pregnant adults were
revised based on U.S. Centers for Disease Control and Prevention (CDC) growth data for
juveniles and values from EPA's Exposure Factors Handbook in adults {Kuczmarski, 2002,
3490881; U.S. EPA, 2011, 786546}. Linear interpolation was used to connect individual
timepoints from these two sources to produce a continuous function over time. Bodyweight
during pregnancy was defined based on selected studies of maternal body weight changes during
pregnancy {Portier, 2007, 192981; Carmichael, 1997, 1060457; Thorsdottir, 1998, 4940407;
Dewey, 1993, 1335605; U.S. EPA, 2011, 786546}. Age-dependent breastmilk intake rates were
based on the 95th percentile estimates from EPA's Exposure Factors Handbook and was defined
relative to the infant's bodyweight {U.S. EPA, 2011, 786546}.
A third modification was the update of parameters: the half-life, Vd, the ratio of PFOS
concentration in cord blood to maternal serum, and the ratio of PFOS concentration in breastmilk
and maternal serum. Details for how these parameters were updated are given in the following
paragraphs. In the model, half-life and Vd are used to calculate the clearance, which is used in the
model directly and is also used for calculation of steady-state concentrations in adults. Other than
half-life and, because of that, clearance, the updated parameters were similar to the original
parameters (Table 4-6). The results of the new R model and updated acslX model with the
original parameters were essentially identical (see PFOS Appendix). With the updated
parameters, the predicted PFOS serum concentrations are approximately 60% of the original
values during pregnancy, and the child's serum concentration is approximately 80% of the
original values during the first year of life.
4-23
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
The use of the Verner model in humans presents a substantial advancement in approach for
endpoints in children compared to the previous EPA assessment of PFOS {U.S. EPA, 2016,
3603365}. The previous assessment did not explicitly model children, but instead applied an
uncertainty factor to an RfD based on long-term adult exposure to account for the potential for
increased susceptibility. The current approach explicitly models PFOS exposure to infants during
nursing and the rapid growth of children, who do not reach steady state until near adulthood.
This allows for a more accurate estimation of exposures associated with either serum levels in
children or dose metric from developmental animal toxicological studies. The Verner model also
explicitly models the mother from her birth through the end of breastfeeding which allows for
the description of accumulation in the mother prior to pregnancy followed by decreasing
maternal levels during pregnancy. Detailed modeling of this period is important for dose metrics
based on maternal levels during pregnancy, especially near term, and on cord blood levels.
Application of the updated Verner model to three cohorts with paired maternal measurements
and subsequent samples in children between ages of 6 months and 6 years showed good
agreement between reported and predicted serum levels in the children (See PFOS Appendix).
This suggests that the assumptions made governing lactational transfer and the selected half-life
value are reasonable. A local sensitivity analysis was also performed to better understand the
influence of each parameter on model output (See PFOS Appendix).
Table 4-6. Updated and Original Chemical-Specific Parameters for PFOS in Humans
Parameter
Updated Value
Original Value"
Volume of Distribution (mL/kg)
230b
230
Half-life (yr)
3.4°
5.5
Clearance (mL/kg/d)
0.128d
0.079
Cord Serum:Maternal Serum Ratio
0.40e
0.42
Milk: Serum Partition Coefficient
0.016f
0.014
a Verner et al. (2016, 3299692).
b Thompson et al. (2010, 2919278).
c Li et al. (2018,4238434).
d Calculated from half-life (ti/2) and volume of distribution (Vd). Clearance (CI) = Vd * ln(2)/ti/2.
e Average values for total PFOA Cord Serum:Maternal Serum ratios (see PFOS Appendix). This is a similar approach to that used
by Verner et al. (2016, 3299692), but also includes studies made available after the publication of that model.
f Average value of studies as reported in Table 4-7. This is a similar approach to that used by Verner et al. (2016, 3299692), but
also includes studies made available after the publication of that model.
EPA's approach for selection of half-life for this effort was to select a reported value from an
exposure to the general population, with a clear decrease in exposure, a high number of
individuals, and a long follow-up time. With these criteria, a half-life of 3.4 years for PFOS was
selected {Li, 2018, 4238434}. This value for PFOS comes from a community with contaminated
drinking water with serial samples of 106 individuals for a relatively short follow-up time of
2 years. A summary of PFOS half-life values is presented in the Appendix (See PFOS
Appendix). Uncertainties related to EPA's selected half-life are discussed in Section 6.6.2.
The updated value for human Vd, 230 mL/kg, was sourced from Thompson et al. (2010,
2919278). To estimate the Vd for PFOS, Thompson et al. (2010, 2919278) scaled the value they
obtained for PFOA by the ratio of VdS obtained by Andersen et al. (2006, 818501) in the
parameterization of that PK model using PK data in monkey. That is, VdPFOA, human) = Vd
4-24
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
(PFOA, human*Vd (PFOS, monkey)/Vd (PFOA, monkey). Vd is a parameter that is relatively
easily obtained from an analysis of PK data from a controlled experimental study, as it is related
to the peak concentration observed after dosing and is expected to be similar between human and
non-human primates {Mordenti, 1991, 9571900}. For comparison, the optimized Vd value from
oral dosing in monkeys was 220 mL/kg for PFOS {Andersen, 2006, 818501}.
A summary of PFOS Vd values is presented in the Appendix (see PFOS Appendix).
Uncertainties related to EPA's selected Vd are discussed in Section 6.6.2.
In the original model, the ratio of PFOS concentration in cord blood to maternal serum, and the
ratio of PFOS concentration in breastmilk and maternal serum were based on an average of
values available in the literature; here, EPA identified literature made available since the original
model was published and updated those parameters with the averages of all identified values
(Table 4-7). The values for cord blood to maternal serum ratio are presented in the Appendix
(see PFOS Appendix). One restriction implemented on the measurements of the cord blood to
maternal serum ratio was to only include reports where the ratio was reported, and not to
calculate the ratio from reported mean cord and maternal serum values. This was due to potential
bias that could be introduced if a greater proportion of cord blood measurements are below the
limit of detection compared to maternal serum.
Table 4-7. Summary of Studies Reporting the Ratio of PFOS Levels in Breastmilk and
Maternal Serum or Plasma
Source
HERO ID
Milk: Maternal
Plasma Ratio
Included in Verner et al.
(2016,3299692) Analysis
Haug et al. (2011,2577501)
2577501
0.014
No
Seung-Kyu Kim etal. (2011, 2919258)
2919258
0.011
No
Liu et al. (2011,2919240)
2919240
0.020
No
Karrman et al. (2007, 1290903)
1290903
0.010
No
Cariou et al. (2015, 3859840)3
3859840
0.011
Yes
Sunmi Kim et al. (2011, 1424975)b
1424975
0.030
Yes
Verner et al. (2016, 3299692)
3299692
0.014°
-
Additional Studies
-
0.016d
-
Whether studies were included in the analysis of Verner et al. (2016, 3299692) is noted. The reported values were based on the
mean of ratios in the study populations except when noted otherwise.
a Median result based on the report of Pizzurro et al. (2019, 5387175).
b Median result as reported by the authors.
c Average value of milk:maternal plasma ratio used by Verner et al. (2016, 3299692).
d Average value of milk:maternal plasma ratio with the inclusion of additional studies not in the original analysis. This value was
used in the human PK model.
This updated model was used to simulate the human equivalent doses (FLED) from the animal
PODs that were obtained from BMD modeling of the animal toxicological studies (see PFOS
Appendix). It was also used to simulate selected epidemiological studies (Section 4.1.4) to obtain
a chronic dose that would result in the internal POD obtained from dose-response modeling (see
PFOS Appendix). For PODs resulting from chronic exposure, such as a long-term animal
toxicological study or an epidemiological study on an adult cohort, the steady state
approximation was used to calculate a PODhed that would result in the same dose metric after
chronic exposure. For PODs from exposure to animals in developmental scenarios, the updated
4-25
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Verner model was used to calculate a PODhed that results in the same dose metric during the
developmental window selected. The updated Verner model was also used to calculate a PODhed
for PODs based on epidemiological observations of maternal serum concentration during
pregnancy, cord blood concentration, and serum concentrations in children.
The pharmacokinetic modeling code for both the updated Wambaugh et al. (2013, 850932) and
Verner et al. (2013, 299692) models that was used to calculate human equivalence doses is
available in an online repository (https://github.com/LJSEPA/OW-PFOS-1 VtCLG-support-
PK-modelsY The model code was thoroughly QA'd through the established EPA Quality
Assurance Project Plan (QAPP) for PBPK models {U.S. EPA, 2018, 4326432}.
4.1.4 Application of Pharmacokinetic Modeling for Animal-
Human Extrapolation of PF OS Toxicological End points and
Dosimetric Interpretation of Epidemiological End points
Table 4-8 displays the POD and estimated internal and PODheds for immune, developmental,
cardiovascular (serum lipids), and hepatic endpoints from animal and/or human studies selected
for the derivation of candidate RfDs. The PODs from epidemiological studies (immune,
developmental, hepatic, and serum lipid endpoints) were derived using benchmark dose
modeling (see PFOS Appendix) which provided an internal serum concentration in mg/L. The
internal dose PODs were converted to a PODhed using the modified Verner model described in
Section 4.1.3.1.3 to calculate the dose that results in the same serum concentrations. Specifically,
reverse dosimetry was performed by multiplying an internal dose POD by a model predicted
ratio of a standard exposure and the internal dose for that standard exposure. This expedited
procedure can be performed because the human model is linear, that is, the ratio of external and
internal dose is constant with dose. Additional details are provided below and in Table 4-8.
The PODs from the animal toxicological studies were derived by first converting the
administered dose to an internal dose as described in Section 4.1.3.1.1. The rationale for the
internal dosimetric selected for each endpoint is described in the Appendix (see PFOS
Appendix). Because a toxicological endpoint of interest results from the presence of chemical at
the organ-specific site of action, dose response modeling is preferentially performed on internal
doses rather than administered doses and assumes the internal dose metric is proportional to the
target tissue dose. In addition, the non-linear elimination described in Wambaugh et al. (2013,
2850932) requires conversion to an internal dose as the relationship between internal and
external dose will not scale linearly. The internal doses were then modeled using the Benchmark
Dose Software (BMDS) (see PFOS Appendix for additional modeling details). The internal dose
animal PODs were converted to a PODhed using the model described in Section 4.1.3.1.3.
Reverse dosimetry for the animal PODs used the ratio of standard exposure and internal dose as
was applied to PODs from epidemiological data. For animal toxicological studies using the
average concentration over the final week of the study (Ciast7), the PODhed is the human dose
that would result in the same steady-state concentration in adults. When a concentration internal
dose metric in the pup during lactation and/or gestation was selected, the PODhed is the dose to
the mother that results in the same average concentration in the fetus/infant over that period.
4-26
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 4-8. PODheds Considered for the Derivation of Candidate RfD Values
Endpoint
Reference,
Confidence
Strain/ POD Type,
Species/Sex Model
POD
POD Internal
Dose/Internal
Dose Metric3
PODhed
(mg/kg/day)
Immunological Effects
Decreased serum anti-
Budtz-Jorgensen (2018,
Human, male and female;
BMDLo.5sd,
18.5 ng/mL
2.71>
<10-6
Grandjean (2018,
508363 l)b
PFOS concentrations in
the mother0 and anti-
Linear
Medium
tetanus antibody serum
concentrations at age
5 years
Timmerman et al. (2021,
Human, male and female;
BMDLo.5sd,
9.66 ng/mL
1.78>
<10-6
9416315)
PFOS concentrations and
Linear
Medium
anti-tetanus antibody
concentrations at ages 7-
10 years
Decreased serum anti-
Budtz-Jorgensen (2018,
Human, male and female;
BMDLo.5sd,
12.5 ng/mL
1.83 >
<10-6
diphtheria antibody
508363 l)b
PFOS concentrations at
Linear
concentration in children
Medium
age five years and anti-
diphtheria antibody serum
concentrations at
age seven years
Budtz-Jorgensen and
Human, male and female;
BMDLo.5sd,
20.0 ng/mL
3.48>
<10-6
Grandjean (2018,
508363 l)b
PFOS concentrations in
the mother0 and anti-
Linear
Medium
tetanus antibody serum
concentrations at age
5 years
Timmerman et al. (2021,
Human, male and female;
BMDLo.5sd,
5.61 ng/mL
1.03 >
<10-6
9416315)
PFOS concentrations and
Linear
Medium
anti-diphtheria antibody
4-27
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Endpoint
Reference,
Confidence
Strain/
Species/Sex
POD Type,
Model
POD
POD Internal
Dose/Internal
Dose Metric"
PODhed
(mg/kg/day)
concentrations at ages 7-
10 years
Decreased PFC response
to SRBC
Zhong et al. (2016,
3748828)
Medium
C57BL/6 Mice, PNW 4 Fi
males
BMDLisd,
Hill
3.3 mg/L
C(iyg pup lact
5.32xl0-4
Extramedullar
Hematopoiesis in the
Spleen
NTP (2019, 5400978)
High
Sprague-Dawley Rats,
female
BMDLiord,
Multistage
Degree 1
2.27 mg/L
Clast7
2.91 x10~4
Extramedullar
Hematopoiesis in the
Spleen
NTP (2019, 5400978)
High
Sprague-Dawley Rats,
male
BMDLiord,
Logistic
9.59 mg/L
Clast7
1.23xl0-3
Developmental Effects
Low Birth Weight
Chu et al. (2020, 6315711) Human, male and female;
High PFOS serum
concentrations in third
trimester
BMDL5RD,
Hybrid
7.3 ng/mL
1.27xl0-6
Sagiv et al. (2018,
4238410)
High
Human, male and female;
PFOS serum
concentrations in first
trimester
BMDL5RD,
Hybrid
41.0 ng/mL
6.00x10-®
Starling et al. (2017,
3858473)
High
Human, male and female;
PFOS serum
concentrations in second
and third trimesters
BMDL5RD,
Hybrid
5.7 ng/mL
9.26 x 10-7
Wikstrom et al. (2020,
6311677)
High
Human, male and female;
PFOS serum
concentrations in first and
second trimesters
BMDL5RD,
Hybrid
7.7 ng/mL
1.13xl0-6
Darrow et al. (2013,
2850966)
High
Human, male and female:
maternal PFOS serum
concentrations taken at
time of enrollment in C8
projectd
BMDL5RD,
Hybrid
17.4 ng/mL
2.51xl0-6
4-28
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Endpoint
Reference,
Confidence
Strain/
Species/Sex
POD Type,
Model
POD
POD Internal
Dose/Internal
Dose Metric"
PODhed
(mg/kg/day)
Yao et al. (2021, 9960202) Human, male and female;
High PFOS serum
concentrations in third
trimester
BMDL5RD,
Hybrid
5.0 ng/L
8.70 x 10 -7
Decreased Fetal Body
Weight
Lee et al. (2015,2851075)
Medium
CD-I Mice, Fi males and
females
NOAELe
0.5 mg/kg/day
8.75 x 10-1 mg/L
Cava pup gest
3.40xl0-4
Decreased Pup Body
Weight
Luebker et al. (2005,
757857)
Medium
Sprague-Dawley Rats, Fi
male and female
BMDL5RD,
Polynomial
Degree 6
10.2 mg/L
Cavg_pup gest
3.96xl0-3
Increased Number of
Dead Fetuses
Lee et al. (2015,2851075)
Medium
CD-I Mice, females
LOAELe
0.5 mg/kg/day
2.13 mg/L
Cavg dam gest
3.32xl0-4
Cardiovascular Effects (Serum Lipids)
Increased Total
Cholesterol
Dong et al. (2019,
5080195)
Medium
Human, male and female;
excluding individuals
prescribed cholesterol
medication
BMDL5RD,
Hybrid
9.34 ng/mL
1.20X10"6
Steenland et al. (2009,
1291109)
Medium
Human, male and female;
excluding individuals
prescribed cholesterol
medication
BMDL5RD,
Hybrid
9.52 ng/mL
1.22xl0-6
Lin et al. (2019, 5187597)
Medium
Human, male and female
BMDL5RD,
Hybrid
66.5 ng/mL
8.51x10-®
Hepatic Effects
Elevated ALT
Galloetal. (2012,
1276142)
Medium
Human, female
BMDL5RD,
Hybrid
56.8 ng/mL
7.27x10-®
Nian et al. (2019,
5080307)
Medium
Human, female
BMDL5RD,
Hybrid
15.1 ng/mL
1.94xl0-6
Increased individual Cell
Necrosis in the Liver
Butenhoff et al. (2012,
1276144)/ Thomford
(2002, 5029075)f
High
Sprague-Dawley rats,
females
BMDLiord,
Log-logisitic
27.0 mg/L
Clast7
3.45X10-3
4-29
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Notes: ALT = alanine aminotransferase; AUC = area under the curve; BMDLo.5sd = lower bound on the dose level corresponding to the 95% lower confidence limit for a change in
the mean response equal to 0.5 standard deviation from the control mean; BMDLisd = lower bound on the dose level corresponding to the 95% lower confidence limit for a
change in the mean response equal to 1 standard deviation from the control mean; BMDLsrd = lower bound on the dose level corresponding to the 95% lower confidence limit for
a 5% change in response; BMDLiord = lower bound on the dose level corresponding to the 95% lower confidence limit of a 10%) change in response; Cavg_pup_gest = average
blood concentration during gestation; Ciast7 = blood concentration over the last 7 days; Fi = first generation; LOAEL = lowest-observed-adverse-effect level; NOAEL = no-
observed-adverse-effect level; PFC = plaque forming cell; PNW = postnatal week; POD = point of departure; PODhed = point of departure human equivalent dose;
RfD = reference dose; SRBC = sheep red blood cell.
a See PFOS Appendix for additional details on BMD modeling.
b Supported by Grandjean et al. (2012,1248827); Grandjean et al. (2017, 3858518); Grandjean et al. (2017,4239492).
c Maternal serum concentrations were taken either in the third trimester (32 weeks) or about two weeks after the expected term date.
d 99% of the pregnancies of participants in Darrow et al. (2013, 2850966) were within 3 years of the serum PFOS measurement.
e No models provided adequate fit; therefore, a NOAEL/LOAEL approach was selected.
f Butenhoff et al. (2012, 1276144) and Thomford et al. (2002, 5029075) reported the same data.
4-30
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
4.1.4.1 Hepatic Effects
Increased ALT in individuals aged 18 and older {Gallo, 2012,1276142; Nian, 2019,
5080307}
The POD for increased ALT in adults was derived by quantifying a benchmark dose using a
hybrid modeling approach (see PFOS Appendix) on the measured PFOS serum concentrations
collected from adults aged 18 years and older {Gallo, 2012, 1276142; Nian, 2019, 5080307},
which provided an internal serum concentration POD in mg/L. The internal serum POD was
converted to an external dose (PODhed), in mg/kg/day. Specifically, the PODhed was calculated
as the external dose that would result in a steady-state serum concentration equal to the internal
serum POD. This calculation is simply the POD multiplied by the selected clearance value
(0.128 mL/kg/day; calculated from half-life and volume of distribution; CI = Vd * ln(2)/ti/2)).
Individual Cell Necrosis in the Liver, Sprague-Dawley rats, females, Ciast7 {Butenhoff, 2012,
1276144}
Increased incidence of individual cell necrosis in the liver was observed in female Sprague-
Dawley Crl:CD(SD)IGS BR rats. Dichotomous models were used to fit dose-response data. A
BMR of 10% extra risk was chosen. The Ciast7 was selected for this model rather than alternate
metrics such as Cmax because the average blood concentration is expected to better correlate with
an accumulation of individual cell necrosis in the liver. The BMDS produced a BMDL in mg/L.
A PODhed was calculated as the external dose that would result in a steady-state serum
concentration in humans equal to the POD from the animal analysis. This calculation is simply
the POD multiplied by the selected clearance value (0.128 mL/kg/day; calculated from half-life
and volume of distribution; CI = Vd * ln(2)/ti/2)).
4.1.4.2 Immune Effects
Decreased Diphtheria and Tetanus antibody response in vaccinated children at age 7
{Budtz-Jorgensen, 2018, 5083631}
The POD for decreased antibody production at age 7 was derived by quantifying a benchmark
dose (see PFOS Appendix) on the measured PFOS serum concentrations at age 5, which
provided an internal serum concentration POD in mg/L. The internal serum POD was converted
to an external dose (PODhed), in mg/kg/day, using the updated Verner model (described in
4.1.3.1.3). For this, the model was run starting at the birth of the mother, with constant exposure
relative to bodyweight. Pregnancy began at 24.25 years maternal age and birth occurred at
25 years maternal age. The initial concentration in the child is governed by the observed ratio
between maternal serum and cord blood at delivery. Then the model is run through the 1 year
breastfeeding period, where the exposure to the child is only through lactation, which is much
greater than the exposure to the mother. After 1 year, the exposure to the child, relative to
bodyweight, is set to the same value as the mother. The model provides predictions up to a child
age of 5 years, when the serum concentrations used to determine the POD were collected, and
reverse dosimetry was used to determine the PODhed that results in the POD serum
concentration. Because of different growth curves used for male and female children used in the
model, the model predicted slightly different (less than 5%) serum concentrations for them. The
lower HED was then selected as it was the most health protective.
4-31
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Decreased Diphtheria and Tetanus antibody response in vaccinated children at age 5
{Budtz-Jorgensen, 2018, 5083631}
The POD for decreased antibody production at age 5 was derived by quantifying a benchmark
dose (see PFOS Appendix) on the measured PFOS serum concentrations collected from the
mother either in the third trimester (32 weeks) or about two weeks after the expected term date,
which provided an internal serum concentration POD in mg/L. The internal serum POD was
converted to an external dose (PODhed), in mg/kg/day, using the updated Verner model
(described in Section 4.1.3.1.3). For this, the model was run similarly to the endpoint based on
antibodies at age 7, except that the model was only run until the maternal age of 25 years, when
delivery occurs in the model. As the POD was based on maternal serum concentrations taken
before and after birth, the time of delivery was chosen as an average of the two. Reverse
dosimetry was performed on model predicted maternal serum concentration at that time to
calculate the PODhed. This metric is independent of the sex of the child in the model.
Decreased Diphtheria and Tetanus antibody response in vaccinated children at ages 7-12
{Timmerman, 2021, 9416315}
The POD for decreased antibody production in children aged 7-12 was derived by quantifying a
benchmark dose (see PFOS Appendix) on the measured PFOS serum concentrations at ages 7-
12, which provided an internal serum concentration POD in mg/L. The internal serum POD was
converted to an external dose (PODhed), in mg/kg/day, using the updated Verner model
(described in Section 4.1.3.1.3). For this, the model was run similarly to the endpoint based on
antibodies at age 7 {Budtz-Jorgensen, 2018, 5083631}, but the model was run until the median
age of this cohort at blood collection, 9.9 years. Reverse dosimetry was used to calculate the
PODhed that resulted in a serum level equal to the POD at that age. Because of different growth
curves used for male and female children, the model predicted slightly different serum
concentrations for them. The lower HED was then selected as it was the most health protective.
Decreased plaque forming cell (PFC) response to SRBC, C57BL/6 Mice, PNW 4 Fi males,
Cavg pup gest lact {Zhong, 2016, 3748828}
Decreased mean level of PFC response of splenic cells was observed in Fi male C57BL/6 mice.
Continuous models were used to fit dose-response data. Using the Wambaugh et al. (2013,
2850932) model, daily exposure to PFOS through oral gavage was simulated from GD1-GD17
days using female CD1 mice parameters (C57BL/6 mice parameters are not available for PFOS).
An average concentration in the pup during gestation and lactation (Cavg PuP gest lact) was
calculated as the internal dose metric for each dose group. A benchmark response (BMR) of a
change in the mean equal to 1 SD from the control mean was chosen per EPA's Benchmark Dose
Technical Guidance {U.S. EPA, 2012, 1239433}. The Cavg_pup gest lact was selected for this model
rather than alternate metrics such as Cmax because the average blood concentration is expected to
better correlate with an accumulation of decreased plaque forming cell response of splenic cells
from across the gestation and lactation lifestages. The BMDS produced a BMDL in mg/L. The
internal serum POD, based on the predicted average serum concentration in the pup during
gestation and lactation, was converted to an external dose (PODhed), in mg/kg/day, using the
updated Verner model (described in Section 4.1.3.1.3). For this, the model was run starting at the
birth of the mother, with constant exposure relative to bodyweight. Pregnancy began at
24.25 years maternal age and birth occurred at 25 years maternal age. The initial concentration in
4-32
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
the child is governed by the observed ratio between maternal serum and cord blood at delivery.
Then the model is run through the 1 year breastfeeding period. The average serum concentration
in the infant through gestation and lactation is determined for this scenario and reverse dosimetry
is used to calculate the exposure that results in the same value as the POD. A male infant was
used for this calculation to match the sex of the animals.
Extramedullar hematopoiesis in the spleen, Sprague-Dawley Rats, female and male, Ciast7
{NTP, 2019, 5400978}
Increased incidence of extramedullar hematopoiesis in the spleen was observed in male
Sprague-Dawley rats. Using the Wambaugh et al. (2013, 2850932) model, daily exposure to
PFOS through oral gavage was simulated for 28 days using Sprague-Dawley rat parameters. An
average concentration over the last seven days of exposure (Ciast7) was calculated as the internal
dose metric for each dose group. Dichotomous models were used to fit dose-response data. A
BMR of 10% extra risk was chosen per EPA's Benchmark Dose Technical Guidance {U.S. EPA,
2012, 1239433}. The Ciast7 was selected for this model rather than alternate metrics such as Cmax
because the average blood concentration is expected to better correlate with an accumulation of
extramedullary hematopoiesis in the spleen. The BMDS produced a BMDL in mg/L. A PODhed
was calculated as the external dose that would result in a steady-state serum concentration in
humans equal to the POD from the animal analysis. This calculation is simply the POD
multiplied by the selected human clearance value (0.128 mL/kg/day; calculated from half-life
and volume of distribution; CI = Vd * ln(2)/ti/2)).
4.1.4.3 Cardiovascular Effects
Increased total cholesterol in individuals aged 20-80, excluding individuals prescribed
cholesterol medication {Dong, 2019, 5080195}
The POD for increased TC in adults was derived by quantifying a benchmark dose using a
hybrid modeling approach (see PFOS Appendix) on the measured PFOS serum concentrations
collected from adults aged 20-80 years not prescribed cholesterol medication through the
NHANES, which provided an internal serum concentration POD in mg/L. The internal serum
POD was converted to an external dose (PODhed), in mg/kg/day. Specifically, the PODhed was
calculated as the external dose that would result in a steady-state serum concentration equal to
the internal serum POD. This calculation is simply the POD multiplied by the selected human
clearance value (0.128 mL/kg/day; calculated from half-life and volume of distribution; CI = Vd
* ln(2)/ti/2)).
Increased total cholesterol in individuals aged 18 and older, excluding individuals
prescribed cholesterol medication {Steenland, 2009,1291109}
The POD for increased TC in adults was derived by quantifying a benchmark dose using a
hybrid modeling approach (see PFOS Appendix) on the measured PFOS serum concentrations
collected from adults aged 18 years and older not prescribed cholesterol medication from the C8
study population, which provided an internal serum concentration POD in mg/L. The internal
serum POD was converted to an external dose (PODhed), in mg/kg/day. Specifically, the
PODhed was calculated as the external dose that would result in a steady-state serum
concentration equal to the internal serum POD. This calculation is simply the POD multiplied by
4-33
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
the selected human clearance value (0.128 mL/kg/day; calculated from half-life and volume of
distribution; CI = Vd * ln(2)/ti/2)).
Increased total cholesterol in individuals aged 25 and older {Lin, 2019, 5187597}
The POD for increased TC in adults was derived by quantifying a benchmark dose using a
hybrid modeling approach (see PFOS Appendix) on the measured PFOS serum concentrations
collected in adults 25 years and older who were at high risk of developing type 2 diabetes and
hyperlipidemia from the Diabetes Prevention Program (DPP) and Outcomes Study (DPPOS),
which provided an internal serum concentration POD in mg/L. The internal serum POD was
converted to an external dose (PODhed), in mg/kg/day. Specifically, the PODhed was calculated
as the external dose that would result in a steady-state serum concentration equal to the internal
serum POD. This calculation is simply the POD multiplied by the selected human clearance
value (0.128 mL/kg/day; calculated from half-life and volume of distribution; CI = Vd *
ln(2)/ti/2)).
4.1.4.4 Developmental Effects
Decreased birthweight using the mother's serum PFOS concentration collected in third
trimester {Chu, 2020, 6315711}
The POD for decreased birth weight in infants was derived by quantifying a benchmark dose
using a hybrid modeling approach (see PFOS Appendix) on the measured PFOS serum
concentrations collected from the mother in the third trimester (blood was collected within
3 days after delivery), which provided an internal serum concentration POD in mg/L. The
internal serum POD was converted to an external dose (PODhed), in mg/kg/day, using the
updated Verner model (described in Section 4.1.3.1.3). This calculation was performed similarly
for each of the birthweight endpoints. The model was run starting at the birth of the mother, with
constant exposure relative to bodyweight. Pregnancy began at 24.25 years maternal age. The
model was stopped at a time to match the median gestational age of the cohort at sample time for
samples taken during pregnancy, or at delivery (25 years maternal age) in the case of maternal
samples at delivery or samples of cord blood. Reverse dosimetry was performed to calculate the
PODhed resulting in serum levels matching the POD at the model end time. For this study,
maternal blood was drawn within a few days of the birth of the child, so delivery was chosen as
the model end time. This metric is independent of the sex of the child in the model.
Decreased birthweight using the mother's serum PFOS concentration collected in in first
trimester {Sagiv, 2018, 4238410}
The POD for decreased birth weight in infants was derived by quantifying a benchmark dose
using a hybrid modeling approach (see PFOS Appendix) on the measured PFOS serum
concentrations collected from the mother in the first trimester (median gestational age of 9
weeks), which provided an internal serum concentration POD in mg/L. The internal serum POD
was converted to an external dose (PODhed), in mg/kg/day, using the updated Verner model
(described in Section 4.1.3.1.3). This was performed as described for the Chu et al. (2020,
6315711) study. The model was stopped at the median gestational age of this cohort, 9 weeks.
The time after conception was calculated as the fraction of pregnancy competed after 9 weeks
(9/39 weeks), times the pregnancy duration of 0.75 year. Reverse dosimetry was performed to
4-34
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
calculate the PODhed that resulted in the POD in maternal serum at that time. This metric is
independent of the sex of the child in the model.
Decreased birthweight using the mother's serum PFOS concentration collected in second
and third trimesters {Starling, 2017, 3858473}
The POD for decreased birth weight in infants was derived by quantifying a benchmark dose
using a hybrid modeling approach (see PFOS Appendix) on the measured PFOS serum
concentrations collected from the mother in the trimesters 2 and 3 (median gestational age of 27
weeks), which provided an internal serum concentration POD in mg/L. The internal serum POD
was converted to an external dose (PODhed), in mg/kg/day, using the updated Verner model
(described in Section 4.1.3.1.3). This was performed as described for the Chu et al. (2020,
6315711) study. The model was stopped at the median gestational age of this cohort, 27 weeks.
The time after conception was calculated as the fraction of pregnancy completed after 27 weeks
(27/39 weeks), times the pregnancy duration of 0.75 year. Reverse dosimetry was performed to
calculate the PODhed that resulted in the POD in maternal serum at that time. This metric is
independent of the sex of the child in the model.
Decreased birthweight using the mother's serum PFOS concentration collected in first and
second trimesters {Wikstrom, 2020, 6311677}
The POD for decreased birth weight in infants was derived by quantifying a benchmark dose
using a hybrid modeling approach (see PFOS Appendix) on the measured PFOS serum
concentrations collected from the mother in the trimesters 1 and 2 (median gestational age of 10
weeks), which provided an internal serum concentration POD in mg/L. The internal serum POD
was converted to an external dose (PODhed), in mg/kg/day, using the updated Verner model
(described in Section 4.1.3.1.3). This was performed as described for the Chu et al. (2020,
6315711) study. The model was stopped at the median gestational age of this cohort, 10 weeks.
The time after conception was calculated as the fraction of pregnancy completed at 10 weeks
(10/39 weeks), times the pregnancy duration of 0.75 year. Reverse dosimetry was performed to
calculate the PODhed that resulted in the POD in maternal serum at that time. This metric is
independent of the sex of the child in the model.
Decreased birthweight using the mother's serum PFOS concentration collected in third
trimester {Yao, 2021, 9960202}
The POD for decreased birth weight in infants was derived by quantifying a benchmark dose
using a hybrid modeling approach (see PFOS Appendix) on the measured PFOS serum
concentrations collected from the mother in the third trimester (blood was collected within 3
days of delivery, at hospital admittance), which provided an internal serum concentration POD in
mg/L. The internal serum POD was converted to an external dose (PODhed), in mg/kg/day,
using the updated Verner model (described in Section 4.1.3.1.3). This calculation was performed
similarly for each of the birthweight endpoints. The model was run starting at the birth of the
mother, with constant exposure relative to bodyweight. Pregnancy began at 24.25 years maternal
age and birth occurred at 25 years maternal age. The model was stopped at a time to match the
median gestational age of the cohort at sample time for samples taken during pregnancy, or at
delivery in the case of maternal samples at delivery or samples of cord blood. Reverse dosimetry
was performed to calculate the PODhed resulting in serum levels matching the POD at the model
4-35
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
end time. For these studies, maternal blood was drawn withing a few days of the birth of the
child, so delivery was chosen as the model end time. This metric is independent of the sex of the
child in the model.
Decreased birthweight using the mother's serum PFOS concentration collected at
enrollment into the C8 study {Darrow, 2013, 2850966}
The POD for decreased birth weight in infants was derived by quantifying a benchmark dose
using a hybrid modeling approach (see PFOS Appendix) on the measured PFOS serum
concentrations collected from the mother at the time of enrollment in the C8 project, which
provided an internal serum concentration POD in mg/L. The internal serum POD was converted
to an external dose (PODhed), in mg/kg/day, using the updated Verner model (described in
4.1.3.1.3). This was performed as described for the Chuetal. (2020, 6315711) study. In this
cohort, blood samples were taken from women before (52%), during (22%), and after (26%)
pregnancy. Because most samples were drawn prior to pregnancy, the PODhed was calculated
based on a maternal age of 24.25 years, prior to any pharmacokinetic effects related to
pregnancy. Reverse dosimetry was performed to calculate the PODhed that resulted in the POD
in maternal serum at that time.
Decreased Fetal Body Weight, CD-I Mice, Fi males and females, Cavgjmp_gest {Lee, 2015,
2851075}
Decreased mean response of fetal body weight was observed in Fi male and female CD-I mice.
Continuous models were used to fit dose-response data. A BMR of 5% extra risk was selected as
described in Section 4.1.2, and a change in the mean equal to 0.5 standard deviations from the
control mean was provided for comparison purposes (See PFOS Appendix). The average blood
concentration of the pup during gestation (Cavg_pup_gest) was selected for this model rather than
alternate metrics such as Cmax because the average blood concentration during gestation is
expected to better correlate with an accumulation of effect resulting in decreased fetal body
weight. The BMDS did not produce a model with adequate fit, so a NOAEL approach was taken.
The internal serum POD, based on the predicted average serum concentration in the pup during
gestation, was converted to an external dose (PODhed), in mg/kg/day, using the updated Verner
model (described in Section 4.1.3.1.3). For this endpoint, the model was run starting at the birth
of the mother, with constant exposure relative to bodyweight. Pregnancy began at 24.25 years
maternal age and birth occurred at 25 years maternal age. The model was run up to the birth of
the child. The average serum concentration in the infant during gestation was determined for this
scenario and reverse dosimetry was used to calculate the exposure that results in the same value
as the POD. Before birth, model predictions for male and female children are equivalent.
Increased Number of Dead Fetuses, CD-I Mice, females, Cavg^dam_gest {Lee, 2015, 2851075}
Increased number of dead fetuses was observed in P0 female CD-I mice. Continuous models
were used to fit dose-response data. A BMR of a change in the mean equal to 0.5 standard
deviations from the control mean was chosen. The average blood concentration of the dam
during gestation (Cavg dam gest) and maximum maternal concentration during gestation (C max dam)
were both considered (see PFOS Appendix) because fetal death could be a result of exposure
during a sensitive window of development. The average blood concentration of the dam during
gestation (Cavg dam gest) was ultimately selected because this metric is expected to better correlate
4-36
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
with an accumulation of effect resulting in decreased fetal survival. The BMDS did not produce
a model with adequate fit, so a LOAEL approach was taken. The internal serum POD, based on
the predicted average serum concentration in the pup during gestation, was converted to an
external dose (PODhed), in mg/kg/day, using the updated Verner model (described in Section
4.1.3.1.3). For this, the model was run starting at the birth of the mother, with constant exposure
relative to bodyweight. Pregnancy began at 24.25 years maternal age and birth occurred at 25
years maternal age. The model was run up to the birth of the child. The average serum
concentration in the infant during gestation was determined for this scenario and reverse
dosimetry was used to calculate the exposure that results in the same value as the POD. Before
birth, model predictions for male and female children are equivalent.
Decreased Pup Body Weight, Sprague-Dawley Rats, Fi male and female,
C avg pup gest {Luebker, 2005, 757857}
Decreased mean pup body weight relative to the litter at LD 5 was observed in Fi male and
female Sprague-Dawley rats. Continuous models were used to fit dose-response data. A BMR of
5% extra risk was selected as described in Section 4.1.2, and a change in the mean equal to 0.5
standard deviations from the control mean was provided for comparison purposes (See PFOS
Appendix). The Cavg pup gest was selected for this model rather than alternate metrics such as Cmax
because the average blood concentration of the pup during gestation is expected to better
correlate with an accumulation of effect resulting in decreased pup body weight. The BMDS
produced a BMDL in mg/L. The internal serum POD, based on the predicted average serum
concentration in the pup during gestation, was converted to an external dose (PODhed), in
mg/kg/day, using the updated Verner model (described in Section 4.1.3.1.3). For this, the model
was run starting at the birth of the mother, with constant exposure relative to bodyweight.
Pregnancy began at 24.25 years maternal age and birth occurred at 25 years maternal age. The
model was run up to the birth of the child. The average serum concentration in the infant during
gestation was determined for this scenario and reverse dosimetry was used to calculate the
exposure that results in the same value as the POD. Before birth, model predictions for male and
female children are equivalent.
4.1.5 Derivation of Candidate Chronic Oral Reference Doses
(RfDs)
Though multiple PODheds were derived for multiple health systems from both epidemiological
and animal toxicological studies, EPA selected the PODheds with the greatest strength of
evidence and the lowest risk of bias represented by high or medium confidence studies for
candidate RfD derivation, as described below. As presented in Table 4-1, epidemiological data
representing the four prioritized health outcomes represented the most sensitive effects after
PFOS exposure in the lower dose range. Four endpoints from epidemiological studies
representing the four health outcomes were considered for candidate RfD derivation. These
endpoints are decreased antibody response, low birth weight, increased total cholesterol, and
elevated ALT. As described in the subsections below, EPA further evaluated studies within each
endpoint to determine those most suitable for candidate RfD derivation.
EPA also further evaluated animal toxicological studies to determine which were the most
suitable for candidate RfD derivation. Factors considered included study confidence (i.e., high
confidence studies were prioritized over medium confidence studies), amenability to benchmark
4-37
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
dose modeling, and health effects observed after exposure in the lower dose range among the
animal toxicological studies. As described in the subsections below, this examination led to the
exclusion a number of studies considered for POD derivation, including both epidemiological
and animal toxicological studies, from further consideration.
4.1.5.1 Hepatic Effects
Two medium confidence epidemiological studies were carried forward for candidate RfD
determination {Gallo, 2012, 1276142; Nian, 2019, 5080307}. EPA considered both studies as
they represented the low-dose range of effects across hepatic endpoints and provided data from
relatively large populations, including U.S. populations.
One high confidence animal toxicological study was carried forward for candidate RfD
determination {Butenhoff, 2012, 1276144}. This study was prioritized for candidate RfD
development because it was determined to be a high confidence study and it was the only study
with a chronic exposure duration that histopathologically examined animals treated with PFOS.
4.1.5.2 Immune Effects
Two medium confidence epidemiological studies were carried forward for candidate RfD
determination {Budtz-j0rgensen, 2018, 5083631; Timmerman, 2021, 9416315}. EPA considered
both studies as they both represented the low-dose range of effects across immunological
endpoints and provided data regarding sensitive populations (i.e., children). Although EPA
derived PODheds for two time points reported by Budtz-j0rgensen and Grandjean (2018,
5083631) (i.e., PFOS serum concentrations at age 5 and antibody concentrations at age 7; PFOS
serum concentrations in the mother during the third trimester or approximately 2 weeks after the
expected term date and antibody concentrations at age 5), EPA did not carry forward PODheds
based on serum PFOS concentrations measured in the mother for candidate RfD derivation
because of concerns surrounding bias due to pregnancy-related hemodynamic effects.
One high and one medium confidence animal toxicological studies were carried forward for
candidate RfD determination {NTP, 2019, 5400978; Zhong, 2016, 3748828}. NTP (2019,
5400978) is a high confidence study reporting the effect of extramedullary hematopoiesis of the
spleen in both male and female rats, female rats being marginally more sensitive than males.
This effect was accompanied by increased bone marrow hypocellularity, suggesting that PFOS
disrupts hematopoiesis in the bone marrow. As extramedullary hematopoiesis was observed in a
high confidence study, in in both sexes, and was amenable to BMD modeling, this endpoint was
carried forward for candidate RfD derivation. The endpoint of reduced PFC response as reported
by Zhong et al. (2016, 3748828) was also selected for candidate RfD derivation because the
effect was reported by multiple studies and represented effects in the low-dose range for immune
effects reported by animal toxicological studies. In addition, Zhong et al. (2016, 3748828)
reported this effect in pups exposed to PFOS during gestation and therefore encompasses a
sensitive population that is coherent with the developmental immunotoxicity observed in
humans.
4.1.5.3 Cardiovascular Effects
Two medium confidence epidemiological studies were carried forward for candidate RfD
determination {Dong, 2019, 5080195; Steenland, 2009, 1291109}. Of the three studies for which
4-38
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
PODheds were derived, Dong et al. (2019, 5080195) and Steenland et al. (2009, 1291109)
exclude individuals who were prescribed cholesterol medication, minimizing concerns
surrounding confounding due to the medical intervention altering serum total cholesterol levels.
Therefore, these two studies were considered further for candidate RfD derivation.
4.1.5.4 Developmental Effects
Two high confidence epidemiological studies were carried forward for candidate RfD
determination {Sagiv, 2018, 4238410; Wikstrom, 2019, 6311677}. Of the six epidemiological
studies for which PODheds were derived, Sagiv et al. (2018, 4238410) and Wikstrom et al.
(2019, 6311677) assessed maternal PFOS serum concentrations primarily or exclusively in the
first trimester, minimizing concerns surrounding bias due to pregnancy-related hemodynamic
effects. Therefore, these two studies were considered further for candidate RfD derivation.
One medium confidence animal toxicological study was carried forward for candidate RfD
determination {Luebker, 2005, 757857}. The endpoint of reduced pup weight from this study
was amenable to benchmark dose modeling (i.e., BMD modeling produced viable model fits),
unlike the endpoints of fetal death and fetal weight reported by Lee et al. (2015, 2851075), which
had a LOAEL and NOAEL as the basis of the PODheds, respectively. As the endpoint of
decreased pup weight reported Luebker et al. (2005, 757857) encompasses sensitive populations
and is coherent with the observed effect of low birth weight in humans, this study was
considered further for candidate RfD derivation. The selection of this study is consistent with
critical study selection in the 2016 HESD {U.S. EPA, 2016, 3603365}.
4.1.5.5 Application of Uncertainty Factors (UFs)
To calculate the candidate RfD values, EPA applied UFs to the PODheds derived from selected
epidemiological and animal toxicological studies (Table 4-9 and Table 4-10). UFs were applied
according to methods described in EPA's Review of the Reference Dose and Reference
Concentration Processes {U.S. EPA, 2002, 88824}.
Table 4-9. Uncertainty Factors for the Development of the Candidate Chronic RfD Values
from Epidemiological Studies {U.S. EPA, 2002, 88824}
UF
Value
Justification
UFa
1
A UFa of 1 is applied to effects observed in epidemiological studies as the study
population is humans.
UFh
10
A UFh of 10 is applied when information is not available relative to variability in
the human population.
UFs
1
A UFS of 1 is applied when effects are observed in adult human populations that
are assumed to have been exposed to a contaminant over the course of many years.
A UFS of 1 is applied for developmental effects because the developmental period
is recognized as a susceptible life stage when exposure during a time window of
development is more relevant to the induction of developmental effects than
lifetime exposure {U.S. EPA, 1991, 732120}.
UFl
1
A UFl of 1 is applied for LOAEL to NOAEL extrapolation when the POD is a
BMDL or a NOAEL.
UFd
1
A UFd of 1 is applied when the database for a contaminant contains a multitude of
studies of adequate quality that encompass a comprehensive array of endpoints in
various life stages and populations and allow for a complete characterization of the
contaminant's toxicity.
4-39
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
UF
Value
Justification
UFC
10
Composite UFC = UFA x UFH x UFS x UFL x UFD
Notes: LTFa = interspecies uncertainty factor; UFd = database uncertainty factor; UFh = intraspecies uncertainty factor;
LTFl = LOAEL-to-NOAEL extrapolation uncertainty factor; UFs = uncertainty factor for extrapolation from a subchronic to a
chronic exposure duration; UFc = composite uncertainty factors.
An interspecies UF (UFa) of 1 was applied to PODheds derived from epidemiological studies
because the dose response information from these studies is directly relevant to humans. There is
no need to account for uncertainty in extrapolating from laboratory animals to humans.
An intraspecies UF (UFh) of 10 was applied to PODheds derived from epidemiological studies to
account for variability in the responses within the human populations because of both intrinsic
(toxicokinetic, toxicodynamic, genetic, life stage, and health status) and extrinsic (lifestyle)
factors that can influence the response to dose. No information to support a UFh other than 10
was available to quantitatively characterize interindividual and age-related variability in the
toxicokinetics or toxicodynamics.
A LOAEL-to-NOAEL extrapolation UF (UFl) of 1 was applied to PODheds derived from
epidemiological studies because a BMDL is used as the basis for the PODhed derivation. When
the POD type is a BMDL, the current approach is to address this factor as one of the
considerations in selecting a BMR for BMD modeling.
A UF for extrapolation from a subchronic to a chronic exposure duration (UFs) of 1 was applied
to PODheds derived from epidemiological studies. A UFS of 1 was applied to the hepatic and
cardiovascular endpoints because the effects were observed in adult populations that were
assumed to have been exposed to PFOS over the course of many years. A UFs of 1 was applied
to the developmental endpoints because the developmental period is recognized as a susceptible
life stage when exposure during a time window of development is more relevant to the induction
of developmental effects than lifetime exposure {U.S. EPA, 1991, 732120}. A UFs of 1 was also
applied to the immune endpoints in children because the developing immune system is
recognized as a susceptible lifestage; therefore, exposure during this time window can be
considered more relevant than lifetime exposure {U.S. EPA, 1991, 732120}. According to the
WHO/ International Programme on Chemical Safety (IPCS) Immunotoxicity Guidance for Risk
Assessment, developmental immunotoxicity encompasses the prenatal, neonatal, juvenile and
adolescent life stages and should be viewed differently from the immune system of adults from a
risk assessment perspective {IPCS, 2012, 1249755}.
A database UF (UFd) of 1 was applied to account for deficiencies in the database for PFOS. In
animals, comprehensive oral short term, subchronic, and chronic studies in three species and
several strains of laboratory animals have been conducted and published in the peer reviewed
literature. Additionally, there are several neurotoxicity studies (including developmental
neurotoxicity) and several reproductive (including one- and two-generation reproductive toxicity
studies) and developmental toxicity studies including assessment of immune effects following
developmental exposure. Moreover, there is a robust epidemiological database which was used
quantitatively in this assessment. Typically, the specific study types lacking in a chemical's
database that influence the value of the UFd to the greatest degree are developmental toxicity and
multigenerational reproductive toxicity studies. Effects identified in developmental and
4-40
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
multigenerational reproductive toxicity studies have been quantitatively considered in this
assessment.
The composite UFs applied to all epidemiological studies considered for candidate RfD
derivation were the same value (UFc = 10) (Table 4-9).
Increased uncertainty is associated with the use of animal toxicological studies as the basis of
candidate RfDs. The composite UFs applied to animal toxicological studies considered for
candidate RfD derivation were either one of two values, depending on the duration of exposure
(i.e., chronic vs. subchronic) or exposure window (e.g., gestational) (Table 4-10).
Table 4-10. Uncertainty Factors for the Development of the Candidate Chronic RfD Values
from Animal Toxicological Studies {U.S. EPA, 2002, 88824}
UF
Value
Justification
UFa
3
A UFa of 3 is applied for the extrapolation from animal models to humans due to
the implementation of a PK model for animal PODhed derivation.
UFh
10
A UFh of 10 is applied when information is not available relative to variability in
the human population.
UFs
1 or 10
A UFS of 10 is applied for the extrapolation of subchronic to chronic exposure
durations. A UFS of 1 is applied to studies with chronic exposure durations or that
encompass a developmental period (i.e., gestation). The developmental period is
recognized as a susceptible life stage when exposure during a time window of
development is more relevant to the induction of developmental effects than
lifetime exposure {U.S. EPA, 1991, 732120}.
UFl
1
A UFl of 1 is applied for LOAEL to NOAEL extrapolation when the POD is a
BMDL or a NOAEL.
UFd
1
A UFd of 1 is applied when the database for a contaminant contains a multitude of
studies of adequate quality that encompass a comprehensive array of endpoints in
various life stages and populations and allow for a complete characterization of the
contaminant's toxicity.
UFC
30 or 300
Composite UFC = UFA x UFH x UFS x UFL x UFD
Notes: UFa = interspecies uncertainty factor; UFd = database uncertainty factor; UFh = intraspecies uncertainty factor;
UFl = LOAEL-to-NOAEL extrapolation uncertainty factor; UFs = uncertainty factor for extrapolation from a subchronic to a
chronic exposure duration; UFc = composite uncertainty factors.
A UFa of 3 was applied to PODheds derived from animal toxicological studies to account for
uncertainty in extrapolating from laboratory animals to humans (i.e., interspecies variability).
The 3-fold factor is applied to account for toxicodynamic differences between the animals and
humans. The HEDs were derived using a model that accounted for PK differences between
animals and humans.
A UFh of 10 was applied to PODheds derived from animal toxicological studies to account for
variability in the responses within human populations because of both intrinsic (toxicokinetic,
toxicodynamic, genetic, life stage, and health status) and extrinsic (lifestyle) factors can
influence the response to dose. No information to support a UFh other than 10 was available to
characterize interindividual and age-related variability in the toxicokinetics or toxicodynamics.
A UFl of 1 was applied to PODheds derived from animal toxicological studies because a BMDL
is used as the basis for the PODhed derivation. When the POD type is a BMDL, the current
4-41
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
approach is to address this factor as one of the considerations in selecting a BMR for BMD
modeling.
A UFs of 1 was applied to PODheds derived from chronic animal toxicological studies as well as
animal toxicological studies that encompass a developmental period (i.e., gestation). A UFS of 1
was applied to developmental endpoints because the developmental period is recognized as a
susceptible life stage when exposure during a time window of development is more relevant to
the induction of developmental effects than lifetime exposure {U.S. EPA, 1991, 732120}. A UFS
of 10 was applied to PODheds derived from studies that implemented a less-than-chronic
exposure duration because extrapolation is required to translate from a subchronic PODhed to a
chronic RfD.
A UFd of 1 was applied to account for deficiencies in the database for PFOS. In animals,
comprehensive oral short term, subchronic, and chronic studies in three species and several
strains of laboratory animals have been conducted and published in the peer reviewed literature.
Additionally, there are several neurotoxicity studies (including developmental neurotoxicity) and
several reproductive (including one- and two-generation reproductive toxicity studies) and
developmental toxicity studies including assessment of immune effects following developmental
exposure. Moreover, there is a robust epidemiological database which was used quantitatively in
this assessment. Typically, the specific study types lacking in a chemical's database that
influence the value of the UFd to the greatest degree are developmental toxicity and
multigenerational reproductive toxicity studies. Effects identified in developmental and
multigenerational reproductive toxicity studies have been quantitatively considered in this
assessment.
4.1.5.6 Candidate RfDs
Table 4-11 shows the UFs applied to each candidate study to subsequently derive the candidate
RfDs.
4-42
-------
DRAFT FOR PUBLIC COMMENT
Table 4-11. Candidate Reference Doses (RfDs)
MARCH 2023
Endpoint
Reference,
Confidence
Strain/ Species/Sex
PODhed
(mg/kg/day)
UFa
UFh
UFs
UFl
UFd
UFtot
Candidate RfDa
(mg/kg/day)
Immune Effects
Decreased Serum Anti-
Budtz-Jorgensen and
Human, male and
2.71 xKT6
1
10
1
1
1
10
2.71xl0-7 = 3xl0-7
Tetanus Antibody
Concentration in
Children
Grandjean (2018,
5083631)
Medium
female
Timmerman et al.
Human, male and
1.78X10"6
1
10
1
1
1
10
1.78xl0-7 = 2xl0-7
(2021, 9416315)
Medium
female
Decreased Serum Anti-
Budtz-Jorgensen and
Human, male and
1.83 xKT6
1
10
1
1
1
10
1.83x10-'= 2xl0"7
Diphtheria Antibody
Concentration in
Children
Grandjean (2018,
5083631)
Medium
female
Timmerman et al.
Human, male and
1.03 xKT6
1
10
1
1
1
10
1.03x10-'= 1x10-'
(2021, 9416315)
Medium
female
Decreased Plaque
Zhong et al. (2016,
C57BL/6 Mice, PNW 4
5.32xKT4
3
10
1
1
1
30
1.77xl0-5 = 2xl0-5
Forming Cell (PFC)
Response to SRBC
3748828)
Medium
Fi males
Extramedullary
Hematopoiesis in the
Spleen
NTP (2019, 5400978)
High
Sprague-Dawley rats,
female
2.91xl0-4
3
10
10
1
1
300
9.70x10-'= lxlQ-6
Developmental Effects
Low Birth Weight
Sagiv et al. (2018,
4238410)
High
Human, male and
female
6.00x10-®
1
10
1
1
1
10
6.00x10-'= 6x10-'
Wikstrom et al. (2019,
Human, male and
1.13xl0-6
1
10
1
1
1
10
1.13x10-'= 1x10-'
6311677)
High
female
Decreased Pup Body
Weight
Luebker et al. (2005,
757857)
Medium
Sprague-Dawley Rats,
Fi male and female
3.96xKT3
3
10
1
1
1
30
1.32xl0-4 = lxlQ-4
Cardiovascular Effects
4-43
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Reference, PODhfd Candidate RfDa
Endpoint Strain/ Species/Sex , " " , UFa UFh UFs UFl UFd UFtot
Confidence (mg/kg/day) (mg/kg/day)
Increased Serum Total Dong et al. (2019, Human, male and 1.20 x 10 6 1 10 1 1 1 10 1.20xl0~7 = lxl0~7
Cholesterol 5080195) female, excluding
Medium individuals prescribed
cholesterol medication
Steenland et al. (2009, Human, male and 1.22 x 10"6 I 10 I I I 10 1.22x10 " = 1/10
1291109) female, excluding
Medium individuals prescribed
cholesterol medication
Hepatic Effects
Increased Serum ALT Gallo et al. (2012, Human, female 7.27 x 10 6 1 10 1 1 1 10 7.27xl0~7 = 7xl0~7
1276142)
Medium
Nian et al. (2019, Human, female 1.94 x 10"' I 10 I I I 10 1.94x10 ~ = 2x10 ~
5080307)
Medium
Individual Cell Butenhoff et al. (2012, Sprague-Dawley rats, 3.45 x 10 3 3 10 1 1 1 30 1.15x10 4 = 1 x io 1
Necrosis in the Liver 1276144)/Thomford females
(2002, 5029075)b
High
Notes: ALT = alanine transaminase; UFa = interspecies uncertainty factor; UFd = database uncertainty factor; UFh = intraspecies uncertainty factor; UFs = subchronic-to-chronic
extrapolation uncertainty factor; UFl = extrapolation from a LOAEL to a NOAEL uncertainty factor; UFtot = composite uncertainty factor.
aRfDs were rounded to one significant figure.
b Butenhoff et al. (2012,1276144) and Thomford et al. (2002, 5029075) reported data from the same experiment.
4-44
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
4.1.6 RfD Selection
As presented in Section 4.1.5 (Table 4-11), EPA derived and considered multiple candidate RfDs
across the four non-cancer health outcomes that EPA determined had the strongest weight of
evidence (i.e., immune, cardiovascular, hepatic, and developmental). EPA derived candidate
RfDs based on both epidemiological and animal toxicological studies. As depicted in Figure 4-3
the candidate RfDs derived from epidemiological studies were all within 1 order of magnitude of
each other (10"6 to 10"7 mg/kg/day), regardless of endpoint, health outcome, or study population.
Candidate RfDs derived from animal toxicological studies were generally 2-3 orders of
magnitude higher than candidate RfDs derived from epidemiological studies. However, EPA
does not necessarily expect concordance between animal and epidemiological studies in terms of
the adverse effect(s) observed, as well as the dose level that elicits the adverse effect(s). For
example, EPA's Guidelines for Developmental Toxicity Risk Assessment states that "the fact that
every species may not react in the same way could be due to species-specific differences in
critical periods, differences in timing of exposure, metabolism, developmental patterns,
placentation, or mechanisms of action" {U.S. EPA, 1991, 732120}. Additionally, for
developmental effects, the guidance says that "the experimental animal data were generally
predictive of adverse developmental effects in humans, but in some cases, the administered dose
or exposure level required to achieve these adverse effects was much higher than the effective
dose in humans" {U.S. EPA, 1991, 732120}.
As shown in Table 4-11 and Figure 4-3, there is greater uncertainty associated with the use of
animal toxicological studies as the basis of RfDs than human epidemiological studies. Though
there are some uncertainties in the use of epidemiological studies for quantitative dose-response
analyses (see Section 6.1), human data eliminate the uncertainties associated with interspecies
extrapolation and the toxicokinetic differences between species which are major uncertainties
associated with the PFOS animal toxicological studies due to the half-life differences between
humans and other species. These uncertainties may explain why the candidate RfDs derived from
animal toxicological studies were several orders of magnitude higher in value than the candidate
RfDs derived from epidemiological studies. Moreover, the human epidemiological studies also
have greater relevance of exposure to human exposure because they directly measure
environmental or serum concentrations of PFOS. In accordance with EPA's current best
practices for systematic review, "animal studies provide supporting evidence when adequate
human studies are available, and they are considered to be the studies of primary interest when
adequate human studies are not available" {U.S. EPA, 2022, 10367891}. For these reasons, EPA
determined that candidate RfDs based on animal toxicological studies would not be further
considered for health outcome-specific RfD selection or overall RfD selection. See Section 6.2
for further comparisons between toxicity values derived from epidemiological and animal
toxicological studies.
4-45
-------
DRAFT FOR PUBLIC COMMENT
Decreased
serum
anti-tetanus
antibody
concentration
in children
Decreased
serum
anti-diptheria
antibody
concentration
in children
Extramedullar
hematopoiesis
in the spleen
Decreased
PFC response
to SRBC
Timmerman et al. (2021, 9416315);
Medium confidence
Budtz-Jorgensen and Grandjean
(2018, 5083631);
Medium confidence
Timmerman et al. (2021, 9416315);
Medium confidence
Budtz-J0rgensen and Grandjean
(2018, 5083631);
Medium confidence
NTP (2019, 5400978);
High confidence
Zhong et al. (2016, 3748828);
Medium confidence
MARCH 2023
Human Animal
¦o
RfD
UF
PODHED
O
-o
¦o
-o
¦o
•-
-o
Decreased
Birth Weight
Sagiv et al. (2018, 4238410);
High confidence
Wikstrom et al. (2019, 6311677);
High confidence
Decreased
Pup Body
Weight
Luebker et al. (2005, 757857);
Medium confidence
o
¦o
#-
-o
Increased
Serum Total
Cholesterol
Dong et al. (2019, 5080195);
Medium confidence
Steenland et al. (2009, 1291109);
Medium confidence
-o
-o
Increased
Serum ALT
Individual Cell
Necrosis in
the Liver
Gallo et al. (2012, 1276142);
Medium confidence
Nian et al. (2019, 5080307);
Medium confidence
Butenhoff et al. (2012, 1276144)/
Thomford (2002, 5029075);
High confidence
-o
-o
¦o
10-8 -I0-7 10-6 10-5 10
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
As described in the subsections below, EPA selected amongst the candidate RfDs to identify an
RfD representative of each of the four prioritized health outcomes (i.e., health outcome-specific
RfDs), as well as an overall RfD that is protective of the effects of PFOS on all health outcomes
and endpoints (Figure 4-4).
4.1.6.1 Health outcome-Specific RfDs
4.1.6.1.1Hepatic Effects
Two medium confidence epidemiological studies were selected as candidates for RfD derivation
for the endpoint of elevated ALT {Gallo, 2012, 1276142; Nian, 2019, 5080307}. The larger
study of PFOS and ALT in adults {Gallo, 2012, 1276142} was conducted in over 30,000 adults
from the C8 Study. The other study {Nian, 2019, 5080307} examined a large population of
adults in Shenyang (one of the largest fluoropolymer manufacturing centers in China) as part of
the Isomers of C8 Health Project and observed significant increases in lognormal ALT per each
ln-unit increase in PFOS, as well significant increases in odds ratios of elevated ALT. The RfD
for increased ALT from Nian et al. (2019, 5080307) was ultimately selected as the health
outcome-specific RfD for hepatic effects because PFOS was the predominating PFAS in this
study which reduces concern about potential confounding by other PFAS. The resulting health
outcome-specific RfD is 2 x 10"7mg/kg/day (Figure 4-4). Note that both candidate RfDs based
on epidemiological studies for the hepatic outcome were within one order of magnitude of the
selected health outcome-specific RfD.
4.1.6.1.2Immune Effects
Two medium confidence epidemiological studies were considered for RfD derivation for the
endpoint of decreased antibody production in response to various vaccinations in children
{Budtz-j0rgensen, 2018, 5083631; Timmerman, 2021, 9416315}. These candidate studies offer
a variety of PFOS exposure measures across various populations and various vaccinations.
Budtz-j0rgensen and Grandjean (2018, 5083631) investigated anti-tetanus and anti-diphtheria
responses in Faroese children aged 5-7 and Timmerman et al. (2021, 9416315) investigated anti-
tetanus and anti-diphtheria responses in Greenlandic children aged 7-12. Though both are
medium confidence studies, the study by Budtz-j0rgensen and Grandjean (2018, 5083631) has
two features that strengthen the results: 1) the response reported by this study reached statistical
significance, and 2) the analysis considered co-exposures of other PFAS. The RfDs for anti-
diphtheria responses in 7-year-old Faroese children from Budtz-j0rgensen and Grandjean (2018,
5083631) was ultimately selected as the basis for the health outcome-specific RfD for immune
effects because the response reported by this study reached statistical significance, this analysis
considered co-exposures of other PFAS, and it was the more health-protective of the two
vaccine-specific responses reported by Budtz-j0rgensen and Grandjean (2018, 5083631). The
resulting health outcome-specific RfD is 2 x 10"7 mg/kg/day (Figure 4-4). Note that all candidate
RfDs based on epidemiological studies for the immune outcome were within one order of
magnitude of the selected health outcome-specific RfD.
4.1.6.1.3Cardiovascular Effects
Two medium confidence epidemiological studies were considered for RfD derivation for the
endpoint of increased total cholesterol {Dong, 2019, 5080195; Steenland, 2009, 1291109}.
These candidate studies offer a variety of PFOS exposure measures across various populations.
Dong et al. (2019, 5080195) investigated the NHANES population (2003-2014), while
4-47
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Steenland et al. (2009, 1291109) investigated effects in a high-exposure community (the C8
Health Project study population). Both of these studies excluded individuals prescribed
cholesterol medication which minimizes concerns of confounding due to medical intervention.
The RfD for increased TC from Dong et al. (2019, 5080195) was ultimately selected for the
health outcome-specific RfD for cardiovascular effects as there is marginally increased
confidence in the modeling from this study. Steenland et al. (2009, 1291109) presented analyses
using both PFOS and TC as categorical and continuous variables. The results using the natural
log transformed TC and the natural log transformed PFOS were stated to fit the data slightly
better than the ones using untransformed PFOS. However, the dramatically different changes in
regression slopes between the two analyses by Steenland et al. (2009, 1291109) resulting in
extremely different PODs raise concerns about the appropriateness of using the data for RfD
derivation. Therefore, the resulting health outcome-specific RfD based on results from Dong et
al. (2019, 5080195) is 1 x 10"7 mg/kg/day (Figure 4-4). Note that the candidate RfDs for the
cardiovascular outcome were nearly identical.
4.1.6.1.4Developmental Effects
Two high confidence epidemiological studies were considered for RfD derivation for the
endpoint of low birth weight {Sagiv, 2018, 4238410; Wikstrom, 2019, 6311677}. These
candidate studies assessed maternal PFOS serum concentrations primarily or exclusively in the
first trimester, minimizing concerns surrounding bias due to pregnancy-related hemodynamic
effects. Both were high confidence prospective cohort studies with many study strengths
including sufficient study sensitivity and sound methodological approaches, analysis, and design,
as well as no evidence of bias. The RfD for low birth weight from Wikstrom et al. (2020,
6311677) was selected as the basis for the organ-specific RfD for developmental effects as it was
the lowest and therefore most health protective candidate RfD from these two studies. The
resulting health outcome-specific RfD is 1 x 10"7 mg/kg/day (Figure 4-4). Note that both
candidate RfDs based on epidemiological studies for the developmental outcome were within
one order of magnitude of the selected health outcome-specific RfD.
4-48
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Immune
Developmental
Anti-tetanus
antibody
response
Anti-diphtheria
antibody
response
Low birth weight 1
Cardiovascular
Increased total
cholesterol
Hepatic
Health Outcome
Elevated ALT
Budtz-J0rgensen
and Grandjean
(2018, 5083631)
3x 10-7
Timmerman et al.
2 x 10~7
'(2021, 9416315)
Budtz-j0rgensen
and Grandjean
(2018, 5083631)
2x107
Timmerman et al.
1 x 10 7
'(2021, 9416315)
Sagiv et al.
6 x 10 7
' (2018,4238410)
Wikstrom et al.
1 x 10'7
(2019,6311677)
Dong et al.
1 x 10-7
(2019, 5080195)
Steenland et al.
1 x 10"7
' (2009, 1291109)
Gallo et al.
7 x 10-7
' (2012, 1276142)
Nian et al.
2 x 107
' (2019, 5080307)
2 x 107
1 x 1&7
1 x 10"7
2x 10
I-7
Candidate RfD
(mg/kg/day)
Endpoint i Study
Figure 4-4. Schematic depicting selection of the overall RID for PFOS
Health Outcome
Specific RfD
(mg/kg/day)
1 x 10~7
Overall RfD
(mg/kg/day)
4.1.6.2 Overall RfD
The available evidence indicates there are effects across immune, developmental, cardiovascular,
and hepatic organ systems at the same or approximately the same level of PFOS exposure. In
fact, candidate RfDs within the developmental and cardiovascular outcomes are the same value
(i.e., 1 x 10~7 mg/kg/day). Therefore, EPA has selected an overall RfD for PFOS of 1 x 10"7
mg/kg/day (Figure 4-4). The developmental and cardiovascular RfDs based on endpoints of low
birth weight and increased total cholesterol, respectively, serve as co-critical effects for this RfD.
Notably, the RfD is protective of effects that may occur in sensitive populations (i.e., infants and
children; see Section 6.8), as well as immune and hepatic effects that may result from PFOS
exposure. As one of the co-critical effects identified for PFOS is a developmental endpoint and
can potentially result from a short-term exposure during critical periods of development, EPA
concludes that the overall RfD for PFOS is applicable to both short-term and chronic risk
assessment scenarios.
4-49
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
The critical studies that serve as the basis of the RfD are all medium or high confidence
epidemiological studies. The critical studies are supported by multiple other medium or high
confidence studies in both humans and animal models and have health outcome databases for
which EPA determined that either evidence indicates or evidence demonstrates that oral PFOS
exposure is associated with adverse effects. Additionally, the selected critical effects can lead to
clinical outcomes in a sensitive lifestage (children) and/or yield the lowest PODHEDand
candidate RfDs and therefore, is expected to be protective of all other health effects in humans.
4.2 Cancer
4.2.1 Study Selection
Several medium and high confidence epidemiological studies and a single high confidence
animal chronic cancer bioassay comprise the evidence database for the carcinogenicity of PFOS.
The available epidemiology studies report elevated risks of bladder, prostate, kidney, and breast
cancers after chronic PFOS exposure. While there are reports of cancer incidence from
epidemiological studies, the study designs, analyses, and mixed results preclude definitive
conclusions about the relationship between PFOS exposure and cancer outcomes in humans and
also limit the potential for quantitative assessment of these data (i.e., dose-response modeling for
CSF derivation).
The sole animal chronic cancer bioassay study provides evidence of multi-site tumorigenesis in
male and female rats. The Thomford (2002, 5029075)/Butenhoff et al. (2012, 1276144) chronic
cancer study was determined to be high confidence and provides several multi-dose tumor
incidence findings in male and female rats that are suitable for dose-response modeling and
subsequent CSF derivation, further described below.
4.2.2 CSF Development
EPA derived PODs and candidate CSFs for four endpoints reported by Thomford (2002,
5029075)/Butenhoff et al. (2012, 1276144): hepatocellular adenomas in male rats; hepatocellular
adenomas in female rats; combined hepatocellular adenomas and carcinomas in female rats; and
pancreatic islet cell carcinomas in male rats (Table 4-12). As noted in Table 3-13, EPA
expressed tumor incidence as the number of animals with reported tumors over the number of
animals alive at the time of first occurrence of the tumor. Expressing incidence in this way
quantitatively eliminates animals that died prior to the PFOS treatment duration plausibly
required to result in tumor formation in the critical study. BMDLs were derived using the BMDS
3.2 program. Multistage models were used consistent with the long-standing practice of EPA to
prefer multistage models to fit tumor dose-response data and a BMR of 10% extra risk was
chosen per EPA's Benchmark Dose Technical Guidance {U.S. EPA, 2012, 1239433}. AUC
averaged over study duration (AUCavg), equivalent to mean serum concentration during the
duration of the study, was selected for this model because the AUC accounts for the
accumulation of effects expected to precede the increased incidence of adenomas and/or
carcinomas. The BMDS produced a BMDL in mg/L. The animal POD was converted to a
PODhed by multiplying the POD by the human clearance value (Table 4-6). This PODhed is
equivalent to the constant exposure, per bodyweight, that would result in serum concentration
equal to the POD at steady state. The CSF is then calculated by dividing the BMR of 10% by the
PODhed.
4-50
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 4-12. Cancer Slope Factors (CSFs) derived from results reported by Butenhoff et al.
(2012,1276144)/Thomford (2002, 5029075)a in Sprague-Dawley rats
Tumor Type
Sex
POD Type, Model
POD Internal Dose
PODhed
Candidate CSF
/Internal Dose Metricb
(BMR/PODhed)
Hepatocellular
Male
BMDLio
25.6 mg/L (AUC
3.28xl0-3
mg/kg/day
Adenomas
Multistage Degree 4
Model
normalized per day
(AUCavg))
30.5 (mg/kg/day)-1
Hepatocellular
Female
BMDLio
21.8 mg/L (AUC
2.79xl(T3
mg/kg/day
Adenomas
Multistage Degree 1
Model
normalized per day
(AUCavg))
35.8 (mg/kg/day)"1
Combined
Female
BMDLio
19.8 mg/L (AUC
Hepatocellular
Multistage Degree 1
normalized per day
2.53X10"3
39.5 (mg/kg/day)1
Adenomas and
Model
(AUCavg))
mg/kg/day
Carcinomas0
Pancreatic Islet Cell
Male
BMDLio
26.1 mg/L (AUC
3.34xl(T3
mg/kg/day
Carcinomas
Multistage Degree 1
Model
normalized per day
(AUCavg))
29.9 (mg/kg/day)"1
Notes: BMDLio = benchmark dose level corresponding to the 95% lower confidence limit of a 10% change.
a Butenhoff et al. (2012,1276144) and Thomford (2002, 5029075) reported data from the same experiment.
b See PFOS Appendix for additional details on benchmark dose modeling.
c Endpoint is bolded to note that it was selected as the basis for the selected cancer slope factor.
4.2.3 CSF Selection
EPA selected the hepatocellular adenomas and carcinomas in female rats reported by Butenhoff
et al. (2012, 1276144)/Thomford (2002, 5029075) as the basis of the CSF for PFOS. This
endpoint was selected because: 1) there was a statistically significant increase in tumor incidence
in the highest dose group; 2) a statistically significant trend of increased incidence with
increasing PFOS concentrations across dose groups; and 3) it is representative of both
hepatocellular tumor types observed in male and female rats. The resulting CSF is
39.5 (mg/kg/day)"1.
Selection of hepatocellular adenomas and carcinomas in female rats is supported by statistically
significant increases in hepatocellular tumor incidence in the high dose group as well as a
statistically significant trend of this response observed in the male rats. The critical effect of
pancreatic islet cell carcinomas was not selected as the basis of the CSF because the response of
the high dose group was not statistically different from the control group, though the trend of
response across dose groups was statistically significant. Regardless, the resulting CSF from this
endpoint is relatively close in value to the CSFs derived from hepatocellular tumors and bolsters
the confidence in the selected CSF.
4.2.4 Application of Age-Dependent Adjustment Factors
EPA's Guidelines for Carcinogen Risk Assessment and Supplemental Guidance for Assessing
Susceptibility from Early-Life Exposure to Carcinogens require the consideration of applying
age-dependent adjustment factors (ADAFs) to CSFs to address potential increased risk for cancer
due to early life stage susceptibility to chemical exposure {U.S. EPA, 2005, 6324329; U.S. EPA,
2005, 88823}. ADAFs are only to be used for carcinogenic chemicals with a mutagenic MOA
when chemical-specific data about early-life susceptibility are lacking. For carcinogens with any
4-51
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
MO A, including mutagens and non-mutagens, but with available chemical specific data for
early-life exposure, those data should be used.
As described in Section 3.5.3.1.1, the limited number of in vivo and in vitro studies assessing
mutagenicity following PFOS exposure were primarily negative. Therefore, EPA has determined
that PFOS is unlikely to cause tumorigenesis via a mutagenic MOA. Given the lack of evidence
of a mutagenic MOA, EPA does not recommend applying ADAFs when quantitatively
determining the cancer risk for PFOS {U.S. EPA, 2011, 783747}.
Additionally, there is insufficient information available from epidemiological and animal
toxicological studies to adequately determine whether PFOS exposure during early-life periods,
per EPA's above-referenced supplemental guidance, may increase incidence or reduce latency
for cancer compared with adult-only exposure. No current studies allow for comparisons of
cancer incidence after early-life vs. adult-only PFOS exposure.
4-52
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
5 MCLG Derivation
Consistent with the Guidelines for Carcinogen Risk Assessment {U.S. EPA, 2005, 6324329},
EPA reviewed the weight of the evidence and determined that PFOS is Likely to Be
Carcinogenic to Humans, as "the evidence is adequate to demonstrate carcinogenic potential to
humans but does not reach the weight of evidence for the descriptor Carcinogenic to Humans
This determination is based on the evidence of hepatocellular tumors in male and female rats,
pancreatic islet cell carcinomas in male rats, and mixed but plausible evidence of bladder,
prostate, kidney, and breast cancers in humans as outlined in Section 3.5.4. As previously noted,
the results reported by one chronic cancer bioassay in rats exceeds the descriptor of Suggestive
Evidence of Carcinogenic Potential as it provides evidence of multi-site and multi-sex
tumorigenesis {Thomford, 2002, 5029075; Butenhoff, 2012, 1276144}.
Unless a non-linear mode of action is determined, EPA establishes MCLGs of zero for
carcinogens classified as Carcinogenic to Humans or Likely to be Carcinogenic to Humans
consistent with the statutory definition of MCLG, which requires EPA to establish MCLGs at a
level where there are "no known or anticipated adverse effects" on public health and with "an
adequate margin of safety." Under SDWA, where there is insufficient information to determine
that a carcinogen has a threshold below which there are no carcinogenic effects, EPA takes the
health-protective approach of assuming that there is no such threshold and that carcinogenic
effects should therefore be extrapolated linearly to zero {U.S. EPA, 1985, 9207; U.S. EPA, 1991,
5499; U.S. EPA, 2016, 6557097}. This approach, known as the linear default extrapolation
approach, ensures that the MCLG is set at a level where there are no adverse health effects with a
margin of safety. EPA has determined that PFOS is Likely to be Carcinogenic to Humans based
on sufficient evidence of carcinogenicity in humans and animals, that there is not sufficient
evidence of a threshold for PFOS, and that therefore a linear default extrapolation approach is
appropriate {U.S. EPA, 2005, 6324329}. Based upon a consideration of the best available peer
reviewed science and a consideration of an adequate margin of safety, EPA proposes a MCLG of
zero for PFOS in drinking water.
5-1
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
6 Effects Characterization
6.1 Addressing Uncertainties in the Use of Epidemiological
Studies for Quantitative Dose-Response Analyses
In the 2016 PFOS HESD and Drinking Water Health Advisory {U.S. EPA, 2016, 3982043; U.S.
EPA, 2016, 3603365}, EPA qualitatively considered epidemiological studies as a supporting line
of evidence but did not quantitatively consider them for POD derivation, citing the following as
reasons to exclude the epidemiological data that were available at that time from quantitative
analyses:
• inconsistencies in the epidemiological database,
• the use of mean serum PFOS concentrations rather than estimates of exposure,
• declining serum PFOS values in the U.S. general population over time {CDC, 2017,
4296146},
• uncertainties related to potential exposure to additional PFAS, telomer alcohols that
metabolically break down into PFOS, and other bio-persistent contaminants, and
• uncertainties related to the clinical significance of effects observed in epidemiological
studies.
Since 2016, EPA has identified many additional epidemiology studies that have increased the
database of information for PFOS (see Sections 3.1.1, 3.4, and 3.5). Further, new tools that have
facilitated the use of study quality evaluation as part of systematic review have enabled EPA to
systematically assess study quality in a way that includes consideration of confounding. As a
result, EPA is now in a position to be able to quantitatively consider epidemiological studies for
POD derivation in this assessment.
In this assessment EPA has assessed the strength of epidemiological and animal evidence
systematically, a process that was not followed in 2016. By performing an updated assessment
using systematic review methods, EPA determined that four health outcomes and four
epidemiological endpoints within these outcomes (i.e., decreased antibody response to
vaccination in children, decreased birthweight, elevated total cholesterol and elevated ALT) have
sufficient weight of evidence to consider quantitatively. Each endpoint quantified in this
assessment has consistent evidence from multiple medium and/or high confidence
epidemiological and animal toxicological studies supporting an association between PFOS
exposure and the adverse effect. Each of the endpoints were also specifically supported by
multiple epidemiological studies in different populations, including general and highly exposed
populations. Several of these supporting studies have been published since 2016 and have
strengthened the weight of evidence for this assessment.
As described in Section 4.1.1.3, EPA has improved upon the pharmacokinetic modeling
technique used in 2016. Though there are challenges in estimations of human dosimetry from
measured or modeled serum concentrations (see Section 6.6.2), EPA has evaluated the available
literature and developed a pharmacokinetic model that estimates PFOS exposure concentrations
6-1
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
from the serum PFOS concentrations provided in epidemiological studies, which reduces
uncertainties related to exposure estimations in humans. This new approach is supplemented
with the UF accounting for intraspecies variation of lOx applied to each PODhed, which
accounts for the sensitivities of specific populations, including those that may have increased
susceptibility to PFOS toxicity due to differential toxicokinetics.
An additional source of uncertainty in using epidemiological data for POD derivation is the
documented declined in human serum PFOS levels over time, which raises concerns about
whether one-time serum PFOS measurements are a good representation of lifetime peak
exposure. Because of PFOS's long half-life in serum, however, one-time measurements likely
reflect several years of exposure. Importantly, EPA considered multiple time periods when
estimating PFOS exposure, ranging from the longest period with available data on PFOS serum
levels within the U.S. population (1999-2018) to the shortest and most recent period (2017-2018)
(see PFOS Appendix E), when performing dose-response modeling of the ALT and TC
endpoints in the epidemiological data. EPA selected PODs for these two endpoints using PFOS
exposure estimates based on the serum PFOS data for 1999-2018, which is likely to capture the
peak PFOS exposures in the U.S. which occurred in the 1990s {Dong, 2019, 5080195}. The
modeling results show that the BMDL estimates for increased TC derived using these exposure
data are consistently lower than those based on the 2017-2018 PFOS exposure data whereas for
ALT, the BMDL estimates using data from the longest exposure period are consistently higher
than those based on the 2017-2018 PFOS exposure data. Based on these analyses, it appears that
selection of one exposure time-period over another does not predictably impact the modeling
results. Therefore, for this assessment, EPA decided to consistently select the time periods more
likely to capture peak PFOS exposures (e.g., 1999-2018) as the basis of BMDL estimates for all
endpoints of interest (see PFOS Appendix E).
It is plausible that observed associations between adverse health effects and PFOS exposure
could be explained in part by confounding from other PFAS exposures, including the metabolism
of precursor compounds to PFOS in the human body. However, for four of the five priority
health outcomes, at least one available study performed multi-pollutant modeling. For example,
for the decreased antibody production endpoint, Budtz-Jorgensen and Grandjean (2018,
5083631) performed a follow-up analysis of the study by Grandjean et al. (2012, 1248827) in
which results were additionally adjusted for PFOA, and there was no notable attenuation of the
observed association between PFOS exposure and decreased antibody response. For an extended
review of the uncertainties associated with PFAS co-exposures, see Systematic Review Protocol
for the PFBA, PFHxA, PFHxS, PFNA, and PFDA (anionic and acid forms) IRIS Assessments
{U.S. EPA, 2020, 8642427}.
Additionally, there is uncertainty about the magnitude of the contribution of PFAS precursors to
PFOS serum concentrations, especially as biotransformation efficiency appears to vary
depending on the precursor of interest {Mcdonough, 2022, 10412593; Vestergren, 2008,
2558842; D'eon, 2011, 2903650}. The contributions of PFAS precursors to serum
concentrations also varies between populations with differing PFAS exposure histories (i.e.,
individuals living at or near sites with AFFF use may have different precursor PFOS
contributions than the general population).
In addition, some populations may be disproportionately exposed to other contaminants, such as
polychlorobiphenyls and methylmercury. To address this, EPA quantified associations between
6-2
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
PFOS serum concentrations and endpoints of interest in populations with varying exposure
histories, including the general population and high-exposure communities. EPA observed
associations for endpoints in populations known to have been predominantly exposed to PFOS
(e.g., Isomers of C8 Health Project participants), reducing the uncertainty related to potential
confounding of other contaminants, including PFAS precursor compounds. These sensitivity
analyses are supportive of EPA's conclusions regarding the effects of PFOS reported across
many epidemiological studies.
In this assessment, studies were not excluded from consideration based primarily on lack of or
incomplete adjustments for potential confounders including socioeconomic status (SES) or
race/ethnicity. A small number of studies examining PFAS serum levels across SES and
racial/ethnic groups were identified. These studies (most with sampling from the early-mid
2000s) reported conflicting results regarding the relationship between race/ethnicity and serum
PFOS concentrations, with studies differing depending on locations sampled, further
stratification of results by age, cohort characteristics, etc. {Kato, 2014, 2851230; Nelson, 2012,
4904674; Calafat, 2007, 1290899; Park, 2019, 5381560}. EPA acknowledges that in
observational epidemiological studies, potential residual confounding may result from SES and
racial/ethnic disparities. Additional racially and ethnically diverse studies in multiple U.S.
communities are needed to fill this important data gap. The PFOS Appendix provides detailed
information on the available epidemiological studies and identifies the study-specific
confounding variables that were considered, such as SES.
Lastly, the potential uncertainty related to the clinical significance of effects observed in the
PFOS epidemiological studies is sometimes cited for dismissing the epidemiological data
quantitatively. However, as described in section 4.1.1, increased ALT levels, decreased antibody
responses in children, increased serum cholesterol levels, and decreased birthweight are
clinically meaningful effects, and EPA's A Review of the Reference Dose and Reference
Concentration Processes, states that a RfD should be based on an adverse effect or a precursor to
an adverse effect (e.g., increased risk of an adverse effect occuring) {U.S. EPA, 2002, 88824}.
Briefly, evidence from both human epidemiological and animal toxicological studies indicates
that increased serum ALT is associated with increased risk for liver disease {Ioannou, 2006,
10473853; Ioannou, 2006, 10473854; Kwo, 2017, 10328876; Roth, 2021, 9960592}. Human
epidemiological studies have also demonstrated that even low magnitude increases in serum
ALT can be clinically significant (See section 4.1.1.1). It is also important to note that while
evaluation of direct liver damage is possible in animal studies, it is difficult to obtain biopsy-
confirmed histological data in humans. Therefore, liver injury is typically assessed using serum
biomarkers of hepatotoxicity {Costello et al, 2022, 10285082}. The SAB's PFAS review panel
noted that reduction in the level of antibodies produced in response to a vaccine represents a
failure of the immune system to respond to a challenge and is considered an adverse
immunological health outcome {U.S. EPA, 2022, 10476098}. Further, a review by Selgrade
(2007, 736210) suggests that specific immunotoxic effects, such as antibody response, observed
in children may be broadly indicative of developmental immunosuppression impacting these
children's ability to protect against a range of immune hazards.
Additionally, increased serum cholesterol is associated with changes in incidence of
cardiovascular disease events such as myocardial infarction (MI, i.e., heart attack), ischemic
stroke (IS), and cardiovascular mortality occurring in populations without prior CVD events
6-3
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
{D'Agostino, 2008, 10694408; Goff, 2014, 3121148; Lloyd-Jones, 2017, 10694407}. Moreover,
disturbances in cholesterol homeostasis contribute to the pathology of non-alcoholic fatty liver
disease (NAFLD) and to accumulation of lipids in hepatocytes {Malhotra, 2020, 10442471},
providing further evidence of effects in the liver. Finally, substantial evidence links low birth
weight to a variety of adverse health outcomes at various stages of life. It has been shown to
predict prenatal mortality and morbidity {Cutland, 2017, 10473225; U.S. EPA, 2013, 4158459;
WHO, 2014, 10473141} and is a leading cause of infant mortality in the United States {CDC,
2020, 10473144}.Low birth weight is also associated with increased risk for diseases in
adulthood, including obesity, diabetes, and cardiovascular disease {Gluckman, 2008, 10473269;
Osmond, 2000, 3421656; Risnes, 2011, 2738398; Smith, 2016, 10474151; Ong, 2002,
10474127, as reported in Yang et al. (2022, 10176603).
There are challenges associated with quantitative use of epidemiological data for risk assessment
{Deener, 2018, 6793519} as described above; however, improvements such as methodological
advancements that minimize bias and confounding, strengthened methods to estimate and
measure exposure, and updated systematic review practices facilitate the use of epidemiological
studies to quantitatively inform risk.
6.2 Comparisons Between Toxicity Values Derived from Animal
Toxicological Studies and Epidemiological studies
As recommended by the SAB {U.S. EPA, 2022, 10476098}, EPA derived candidate RfDs and
CSFs for multiple health outcomes using data from both epidemiological and animal
toxicological studies. Candidate RfDs from epidemiological and animal toxicological studies
within a health outcome differed by approximately two to three orders of magnitude (see Figure
4-3), with epidemiological studies producing lower values. EPA does not necessarily expect
concordance between animal and epidemiological studies in terms of the adverse effect(s)
observed, as well as the dose level that elicits the adverse effect(s). For example, EPA's
Guidelines for Developmental Toxicity Risk Assessment states that "the fact that every species
may not react in the same way could be due to species-specific differences in critical periods,
differences in timing of exposure, metabolism, developmental patterns, placentation, or
mechanisms of action" {U.S. EPA, 1991, 732120}. EPA further describes these factors in
relation to this assessment below.
First, there are well-established differences in the toxicokinetics between humans and animal
models such as rats and mice. As described in Section 3.3.1.4.5, PFOS half-life estimates vary
considerably by species, being lowest in rodents (hours to days) and several orders of magnitude
higher in humans (years). All candidate toxicity values based on animal toxicological studies
were derived from studies conducted in rats or mice, adding a potential source of uncertainty
related to toxicokinetic differences in these species compared to humans. To address this
potential source of uncertainty, EPA utilized a PK model to estimate the internal dosimetry of
each animal model and convert the values into predicted levels of human exposure that would
result in the corresponding observed health effects. However, the outputs of these models are
estimates and may not fully account for species-specific toxicokinetic differences, particularly
differences in excretion. The application of uncertainty factors (i.e., UFa) also may not precisely
reflect animal-human toxicokinetic differences.
6-4
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Second, candidate toxicity values derived from epidemiological studies are based on responses
associated with actual environmental exposure levels, whereas animal toxicological studies are
limited to the tested dose levels which are often several orders of magnitude higher than the
ranges of exposure levels in humans. Extrapolation from relatively high experimental doses to
environmental exposure levels introduces a potential source of uncertainty for toxicity values
derived from animal toxicological studies; exposures at higher dose levels could result in
different responses, perhaps due to differences in mechanisms activated, compared to responses
to lower dose levels. One example of this is the difference between epidemiological and animal
toxicological studies in the effect of PFOS exposure on serum lipid levels (i.e., potential non-
monotonic dose-response relationships that are not easily assessed in animal studies due to low
dose levels needed to elicit the same response observed in humans).
Third, there may be differences in mechanistic responses between humans and animal models.
Two examples of this is the PPARa and CAR responses. It is unclear to what extent PPARa and
CAR influence the responses to PFOS exposure observed in humans, though the rodent PPARa
and CAR responses may differ from those observed in humans (see Section 3.4.1.3.1).
Mechanistic differences could influence dose-response relationships and subsequently result in
differences between toxicity values derived from epidemiological and animal toxicological
studies. There may be additional mechanisms that differ between humans and animal models that
could contribute to the magnitude of responses and doses required to elicit responses across
species.
The factors described above represent some but not all potential contributors that may explain
the differences between toxicity values derived from epidemiological and animal toxicological
studies. In this assessment, EPA prioritized epidemiological studies of medium or high
confidence for the selection of health outcome-specific and overall RfDs and CSFs (see Section
4.1.6). The use of human data to derive toxicity values removes uncertainties and assumptions
about human relevance inherent in extrapolating from and interpreting animal toxicological data
in quantitative risk assessment.
6.3 Updated Approach to Animal Toxicological RfD Derivation
Compared to the 2016 PFOS HESD
For POD derivation in this assessment, EPA considered the studies identified in the recent
literature searches and also re-examined the candidate RfDs derived in the 2016 PFOS HESD
{U.S. EPA, 2016, 3603365} and the animal toxicological studies and endpoints on which they
were based. The updated approach used for hazard identification and dose response in the current
assessment as compared to the 2016 HESD led to some differences between animal toxicological
studies and endpoints used as the basis of candidate RfDs for each assessment. These updates
and the resulting differences are further described below.
For the 2016 PFOS HESD, EPA did not use BMD modeling to derive PODs, and instead relied
on the NOAEL/LOAEL approach for all candidate studies and endpoints {U.S. EPA, 2016,
3603365}. The NOAEL/LOAEL approach allows for the incorporation of multiple endpoints
from a single study to derive a single POD, if the endpoints have the same NOAEL and/or
LOAEL. For example, in the 2016 PFOS HESD, EPA derived a candidate RfD based on the
endpoints of increased ALT and increased blood urea nitrogen (BUN) reported by Seacat et al.
(2003, 1290852), both of which shared a common POD (NOAEL). For the current assessment,
6-5
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
EPA preferentially used BMD modeling to derive PODs because it allows for greater precision
than the NOAEL/LOAEL approach and considers the entirety of the dose-response curve. This
approach requires the consideration of endpoints on an individual basis and further examination
of the weight of evidence for particular endpoints, as well as the dose-response trend reported for
each endpoint, in order to derive a BMDL. When considering an effect on a standalone basis
rather than grouped with other effects occurring at the same exposure level, EPA sometimes
determined the weight of evidence was not sufficient to consider an individual endpoint for POD
derivation. For the current assessment, EPA used a systematic review approach consistent with
the IRIS Handbook {U.S. EPA, 2022, 10367891} to consider the weight of evidence for both the
health outcomes as well as for individual endpoints of interest when selecting endpoints and
studies for dose-response modeling. In the case of the endpoints selected in the 2016 PFOS
HESD from the Seacat et al. (2003, 1290852) study, renal effects such as increased BUN were
reevaluated and determined to have evidence suggestive of an association with PFOS exposure.
As described in Section 4.1.1, in this assessment, EPA only derived PODs for endpoints from
health outcomes with evidence indicating or evidence demonstrating an association with PFOS
exposure.
Additionally, for the current assessment, EPA preferentially selected endpoints that were
amenable to BMD modeling, had dose-dependent trends in responses, were supported by at least
one other study in the available literature, and were direct/specific measures of toxicity for POD
derivation. For some studies considered in the 2016 PFOS HESD and reevaluated during the
current assessment, EPA attempted BMD modeling for specific endpoints but the efforts did not
result in viable model fits. For the current assessment, EPA elected to derive a candidate RfD for
hepatic effects based on histopathological lesions observed in the liver as reported by Butenhoff
et al. (2012, 1276144)/Thomford (2002, 5029075) rather than serum ALT reported by Seacat et
al. (2003, 1290852), as the Butenhoff et al. (2012, 1276144)/Thomford (2002, 5029075) studies
were rated as high confidence (vs. the medium confidence Seacat et al. (2003, 1290852)), used a
chronic study design (vs. the 14-week exposure used by Seacat et al. (2003, 1290852)), and
histopathological lesions reflect direct damage to the liver whereas ALT is an indicator of liver
damage. In animal studies, evaluation of direct liver damage is possible, however in humans, it is
difficult to obtain biopsy-confirmed histological data. Therefore, liver injury is typically assessed
using serum biomarkers of hepatotoxicity {Costello et al, 2022, 10285082}.
For some health outcomes, new studies have been published since 2016 that improve upon the
weight of evidence determined in the 2016 PFOS HESD. For example, in 2016, EPA did not
derive a candidate RfD based on immune effects. Since that time, several high and medium
confidence studies (both animal toxicological and epidemiological) have been published that
increased the strength of evidence for this health outcome. As described in Section 3.4.2.4,
evidence indicates that PFOS exposure is associated with immune effects and therefore, in this
assessment, EPA derived candidate RfDs for the immune health outcome.
For transparency, EPA has provided a comparison of studies and endpoints used to derive
candidate RfDs for both the 2016 PFOS HESD and the present assessment in Table 6-1.
6-6
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Table 6-1. Comparison of Candidate RfDs Derived from Animal Toxicological Studies for
Priority Health Outcomes"
Studies and Effects Used in 2016 for Studies and Effects Used in 2023 for
Candidate RfD Derivationb Candidate RfD Derivation
Immune
Zhong et al. (2016, 3748828), medium
confidence - decreased PFC response to
NA SRBC
NTP (2019, 5400978), high confidence -
extramedullar hematopoiesis
Developmental
Luebker et al. (2005, 757857) medium Luebker et al. (2005, 757857), medium
confidence - decreased pup body weight confidence - decreased pup body weight
Luebker etal. (2005, 1276160), medium
confidence - decreased pup survival
Lau et al. (2003, 757854), medium confidence
- decreased pup survival
Hepatic
Seacat et al. (2003, 1290852), medium Butenhoff et al. (2012, 1276144)/Thomford
confidence - increased ALT (and increased (2002, 5029075), high confidence - individual
BUN) cell necrosis in the liver
Notes: RfD = reference dose; NA = not applicable; PFC = plaque forming cell; SRBC = sheep red blood cell; NTP = National
Toxicology Program; ALT = alanine aminotransferase; BUN = blood urea nitrogen.
a Note that candidate RfDs for the fourth priority health outcome (i.e., cardiovascular) are not presented in this table because
candidate RfDs based on animal toxicological studies representing this health outcome were not derived in the 2016 HESD or
the current assessment.
b Candidate RfDs from the 2016 HESD that correspond to non-prioritized health outcomes (e.g., nervous) are not presented here.
6.4 Reevaluation of the PFOS Carcinogenicity Database
In November 2021, EPA published the draft Proposed Approaches to the Derivation of a
Maximum Contaminant Level Goal for Perjluorooctane Sulfonic Acid (PFOS) (CASRN1763-23-
1) in Drinking Water for review by the SAB PFAS Review Panel {U.S. EPA, 2021, 10428576}.
As part of the review process, EPA charged the SAB panel with providing comment on the
rationale and conclusion for the PFOS cancer classification. Prior to SAB review, EPA had
concluded that the weight of evidence supported the determination of PFOS as having Suggestive
Evidence of Carcinogenicity, similar to the conclusions of the 2016 PFOS HESD {U.S. EPA,
2016, 3603365}. This was, in part, because no new animal toxicological studies have been
published since publication of the HESD and the new epidemiological literature continue to
provide mixed results.
As part of the final report, the SAB noted, "[sjeveral new studies have been published that
warrant further evaluation to determine whether the "likely" designation is appropriate" {U.S.
EPA, 2022, 10476098}. The SAB recommended EPA reevaluate several aspects of the
carcinogenicity database for PFOS to confirm or update the draft Proposed Approaches
conclusion that PFOS has Suggestive Evidence of Carcinogenic Potential, including
epidemiological studies reporting kidney cancer (i.e., Shearer et al. (2021, 7161466) and Li et al.
(2022, 9961926)), mechanistic data (e.g., Benninghoff et al. (2012, 1274145)), and "overly
6-7
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
conservative" conclusions about animal toxicological data in rats (i.e., Butenhoff et al. (2012,
1276144)). EPA has reevaluated these aspects of the database and relevant discussions of the
recommended studies are provided in Section 3.5.
Upon reassessment of the PFOS carcinogenicity database, including the epidemiological, animal
toxicological, and mechanistic databases, EPA has now determined the available data for PFOS
surpass many of the descriptions for Suggestive Evidence of Carcinogenic Potential according to
the Guidelines for Carcinogen Risk Assessment {U.S. EPA, 2005, 6324329}. The examples for
which the PFOS database exceeds the descriptions outlined in the Guidelines for Carcinogen
Risk Assessment include:
• "a small, and possibly not statistically significant, increase in tumor incidence observed in
a single animal or human study that does not reach the weight of evidence for the
descriptor 'Likely to Be Carcinogenic to Humans;'
• a small increase in a tumor with a high background rate in that sex and strain, when there
is some but insufficient evidence that the observed tumors may be due to intrinsic factors
that cause background tumors and not due to the agent being assessed;
• evidence of a positive response in a study whose power, design, or conduct limits the
ability to draw a confident conclusion; and
• a statistically significant increase at one dose only, but no significant response at the other
doses and no overall trend" {U.S. EPA, 2005, 6324329}.
The strongest evidence for the carcinogenicity of PFOS is primarily from one chronic animal
bioassay which presents findings surpassing several of these criteria. {Thomford, 2002,
5029075/Butenhoff, 2012, 1276144}. The Thomford/Butenhoff et al. (2002, 5029075; 2012,
1276144) study is a high confidence study that observed statistically significant increases at
individual dose levels and/or statistically significant trends in two tumor types and in one or
more sexes, even with the relatively low dose levels used. The background incidence of these
tumor types was low or negligible.
In the draft Proposed Approaches document, EPA relied upon the tumor incidences provided in
Butenhoff et al. (2012, 1276144), which is the peer-reviewed manuscript of an unpublished
industry report - Thomford (2002, 5029075). Upon further review of the Thomford (2002,
5029075) report, EPA recognized two factors that influenced previous qualitative and
quantitative interpretations of the data: 1) the Butenhoff et al. (2012, 1276144) study reported
combined incidences of neoplastic lesions in the control and high dose groups from the interim
time point (52 weeks of dietary exposure; n = 10) and terminal time point (104 weeks of dietary
exposure; n = 50); and 2) the Butenhoff et al. (2012, 1276144) study did not report incidences for
pancreatic islet cell neoplasms. The first factor resulted in statistical dilutions of tumor incidence
in the high dose group as many of the tumor types observed in the study, including
hepatocellular neoplasms, were not reported until approximately 70 weeks of treatment or later.
Therefore, EPA excluded animals sacrificed at the interim time point from statistical analyses as
it was biologically implausible for the 10 animals from the interim time point to have presented
with neoplasms. The second factor resulted in a previous lack of recognition by EPA that a
statistically significant trend in a second tumor site/type (pancreatic islet cell carcinomas) was
6-8
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
observed in the chronic cancer bioassay. This factor importantly results in PFOS meeting an
additional characteristic for the designation of Likely to be Carcinogenic to Humans: "an agent
that has tested positive in animal experiments in more than one species, sex, strain, site, or
exposure route, with or without evidence of carcinogenicity in humans" {U.S. EPA, 2005,
6324329}.
Although the study design (i.e., low dose levels and relatively large gap between the highest and
next highest dose group levels (5 and 20 ppm)) may limit the ability to interpret the dose-
response relationship presented by Thomford/Butenhoff et al. (2002, 5029075; 2012, 1276144),
these results are quantifiable and have been used to derive CSFs within this assessment. Overall,
the Thomford/Butenhoff et al. (2002, 5029075; 2012, 1276144) report, along with plausible
associations between PFOS exposure and carcinogenicity reported by epidemiological studies,
provides substantive evidence that PFOS exceeds the designation of Suggestive Evidence of
Carcinogenic Potential and is consistent with Likely Evidence of Carcinogenic Potential in
Humans (see Section 3.5.5 for more information on the Likely determination).
Table 6-2. Comparison of the PFOS Carcinogenicity Database with the Suggestive Cancer
Descriptor as Described in the Guidelines for Carcinogen Risk Assessment {U.S. EPA,
2005, 6324329}
Suggestive Evidence of Carcinogenic Potential
A small, and possibly not statistically PFOS data exceed this description. Observed statistically
significant, increase in tumor incidence significant increases in hepatic tumors (adenomas in males and
observed in a single animal or human study adenomas and carcinomas in females) at the high dose and a
that does not reach the weight of evidence for statistically significant trend overall in both sexes,
the descriptor "Likely to Be Carcinogenic to
Humans." The study generally would not be
contradicted by other studies of equal quality
in the same population group or experimental
system
A small increase in a tumor with a high This description is not applicable to the tumor types observed
background rate in that sex and strain, when after PFOS exposure,
there is some but insufficient evidence that
the observed tumors may be due to intrinsic
factors that cause background tumors and not
due to the agent being assessed.
Evidence of a positive response in a study PFOS data exceed this description. The study from which
whose power, design, or conduct limits the carcinogenicity data are available was determined to be high
ability to draw a confident conclusion (but confidence during study quality evaluation.
does not make the study fatally flawed), but
where the carcinogenic potential is
strengthened by other lines of evidence (such
as structure-activity relationships)
A statistically significant increase at one dose PFOS data exceed this description. Observed statistically
only, but no significant response at the other significant increases in hepatic tumors (adenomas in males and
doses and no overall trend adenomas and carcinomas in females) at the high dose and a
statistically significant trend overall. Also observed statistically
significant trend of increased pancreatic islet cell tumors with
increasing dose.
6-9
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
6.5 Health Outcomes with Evidence Integration Judgments of
Evidence Suggests Bordering on Evidence indicates
EPA evaluated sixteen non-cancer health outcomes as part of this assessment. In accordance with
recommendations from the SAB {U.S. EPA, 2022, 10476098} and the IRIS Handbook {U.S.
EPA, 2022, 10367891}, for both quantitative and qualitative analyses in the current assessment,
EPA prioritized health outcomes with either evidence demonstrating or evidence indicating
associations between PFOS exposure and adverse health effects. Health outcomes reaching these
tiers of judgment were the hepatic, immune, developmental, cardiovascular, and cancer
outcomes. Some other health outcomes were determined to have evidence suggestive of
associations between PFOS and adverse health effects as well as some characteristics associated
with the evidence indicates tier, and EPA made judgments on these health outcomes as described
below.
For PFOS, two health outcomes that had characteristics of both evidence suggests and evidence
indicates were the endocrine and nervous system outcomes. Endpoints relevant to these two
health outcomes had been previously considered for POD derivation in the Proposed Approaches
to the Derivation of a Draft Maximum Contaminant Level Goal for Perfluorooctane Sulfonic
Acid (PFOS) (CASRN1763-23-1) in Drinking Water. However, upon further examination using
the protocols for evidence integration outlined in the PFOS Appendix and Section 2.1.5, EPA
concluded that the available epidemiological and animal toxicological evidence did not meet the
criteria necessary for subsequent quantitative dose-response analyses. Although these health
outcomes were not prioritized in the current assessment, based on the available data, EPA
concluded that PFOS exposure may cause adverse endocrine or nervous system effects.
Epidemiological studies considered for evidence integration for adverse endocrine effects
include many high and medium confidence studies. There was slight evidence to suggest human
endocrine toxicity, including associations between PFOS exposure and thyroid disease.
However, this evidence was limited to one high confidence study {Kim, 2020, 6833758}. In
addition, the available evidence supports the relationship between PFOS exposure and thyroid
stimulating hormone (TSH) in children and, to a lesser extent, adults. Similar to what was
concluded in the 2016 PFOS HESD, evidence supporting adverse endocrine effects was
inconsistent among epidemiological studies. Animal toxicological studies considered for
evidence integration consisted of 13 high or medium confidence studies. The animal evidence for
an association between PFOS exposure and effects on the endocrine system was considered
moderate, based on observed disruptions of normal thyroid function (i.e., decreased free
thyroxine (T4), total T4 and total triiodothyronine (T3)). In addition, reductions in hormones
associated with the hypothalamic-pituitary-adrenal axis were observed, although the
corresponding histopathological data was inconsistent. Overall, the available human and animal
evidence was suggestive but not indicative of, adverse endocrine effects due to PFOS exposure.
Therefore, EPA did not prioritize this outcome for dose-response modeling. See Appendix C for
a detailed description of endocrine evidence synthesis and integration.
Similar endocrine effects are observed among the family of PFAS chemicals. For example, the
thyroid was identified as a target for oral exposure to PFBS {U.S. EPA, 2021, 7310530}.
Additionally, the draft IRIS Toxicological Review ofPFBA concluded that the available evidence
indicates the observed thyroid effects were likely due to PFBA exposure {U.S. EPA, 2021,
6-10
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
10064222}. Given the similarities across PFAS, these findings support potential associations
between PFOS and adverse endocrine effects.
There was also slight evidence from epidemiological studies that supported a relationship
between PFOS exposure and adverse nervous system effects, but study results were mostly
mixed or limited. For example, studies evaluating neurodevelopmental, neuropsychological, and
cognitive outcomes were limited with only one study supporting an adverse effect of PFOS
exposure on hearing {Li, 2020, 6833686}. Although multiple studies examining associations
between PFOS and ADHD were available, only one study reported a significant relationship
between PFOS and ADHD {Lenters, 2019, 5080366}. There was an indication of a potential
relationship between PFOS and autistic behaviors or ASD diagnosis in some studies {Braun,
2014, 2345999; Oulhote, 2016, 3789517; Shin, 2020, 6507470}, however there were
methodology concerns associated with these studies. Animal studies considered for evidence
integration suggest a relationship between PFOS exposure and nervous system effects,
specifically in relation to learning and memory and neurotransmitter concentrations. Although
there is moderate evidence to support adverse effects on the nervous system following exposure
to PFOS from animal toxicological studies, EPA concluded there is considerable uncertainty in
the results due to inconsistency across studies and limited number of studies. Overall, the
available human and animal evidence was suggestive but not indicative of adverse nervous
system effects due to PFOS exposure. Therefore, EPA did not prioritize this outcome for dose-
response modeling. See Appendix C for a detailed description of endocrine evidence synthesis
and integration.
As the databases for endocrine and nervous system outcomes were suggestive of human health
effects resulting from PFOS exposure, they were not prioritized during the updated literature
review conducted in February 2022. However, EPA acknowledges that future studies of these
currently "borderline" associations could impact the strength of the association and the weight of
evidence for these health outcomes. The currently available studies indicate the potential for
endocrine and nervous system effects after PFOS exposure. Studies on endocrine and nervous
system health outcomes represent two important research needs.
6.6 Challenges and Uncertainty in Modeling
6.6.1 Modeling of Animal Internal Dosimetry
There are several limitations and uncertainties associated with using pharmacokinetic models in
general and estimating animal internal dosimetry. In this assessment, EPA utilized the
Wambaugh et al. (2013, 2850932) animal internal dosimetry model because it had availability of
model parameters across almost all species of interest, agreement with out-of-sample datasets
(see PFOS Appendix), and flexibility to implement life-course modeling (see Section 4.1.3.1).
However, there were some limitations to this approach.
First, posterior parameter distributions summarized in Table 4-3 for each sex/species
combination were determined using a single study. Therefore, uncertainty in these parameters
represents only uncertainty in fitting that single study; any variability between studies or
differences in study design were not accounted for in the uncertainty of these parameters.
Second, issues with parameter identifiability for some sex/species combinations resulted in
6-11
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
substantial uncertainty for some parameters. For example, filtrate volume (Vfil) represents a
parameter with poor identifiability when determined using only serum data due to lack of
sensitivity to serum concentrations (See PFOS Appendix). Measurements in additional matrices,
such as urine, would help inform this parameter and reduce the uncertainty reflected in the wide
credible intervals of the posterior distribution. These parameters with wide posterior CIs
represent parameters that are not sensitive to the concentration-time datasets on which the model
was trained (See PFOS Appendix). However, these uncertain model parameters will not impact
the median prediction used for BMD modeling and simply demonstrate that the available data
are unable to identify all parameters across every species over the range of doses used for model
calibration. Finally, the model is only parameterized using adult, single dose, PFOS study
designs. Gestational and lactational PK modeling parameters were later identified from
numerous sources (Table 4-5) to allow for the modeling of these life stages with a more detailed
description of the life-course modeling in Section 4.1.3.1.3.
The Wambaugh et al. (2013, 2850932) model fit the selected PFOS developmental study data
well, though there are several limitations to using this method to model developmental life
stages. First, perinatal fetal concentrations assume instantaneous equilibration across the placenta
and do not account for the possibility of active transporters mediating distribution to the fetus.
Second, clearance in the pup during lactation is assumed to be a first-order process governed by
a single half-life. At low doses, this assumption is in line with adult clearance, but it is unclear
how physiological changes during development impact the infant half-life. Finally, PFOS
concentrations in breast milk are assumed to partition passively from the maternal blood. This
assumption does not account for the presence of active transport in the mammary gland or time-
course changes for PFOS uptake to the milk. Despite these limitations, the incorporation of
model parameters related to developmental life stages is a significant improvement over the
model used in the 2016 HESD which did not implement life course modeling {U.S. EPA, 2016,
3603365}.
6.6.2 Modeling of Human Dosimetry
Uncertainties may stem from efforts to model human dosimetry. One limitation is that the
clearance parameter, which is a function of the measured half-life and Vd values, is difficult to
estimate in the human general population. Specifically for PFOS, the measurement of half-life is
hindered by slow excretion and ongoing exposure. Additionally, it is unclear whether some of
the variability in measured half-life values reflects actual variability in the population, as
opposed to uncertainty in the measurement of the value. There is also a lack of reported Vd
values in humans because this parameter requires knowledge of the total dose or exposure. Vd
values are difficult to determine from environmental exposures, and only one reported value is
available {Thompson, 2010, 5082271}.
In the Verner et al. (2016, 3299692) model, half-life, Vd, and hence clearance values are assumed
to be constant across ages and sexes. The excretion of PFOS in children and infants is not well
understood. The ontogeny of renal transporters, age-dependent changes in overall renal function,
and the amount of protein binding (especially in serum) could all play a role in PFOS excretion
and could vary between children and adults. It is even difficult to predict the overall direction of
change in excretion in children (higher or lower than in adults) without a clear understanding of
6-12
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
these age-dependent differences. Vd is also expected to be different in children. Children have a
higher body water content, which results in a greater distribution of hydrophilic chemicals to
tissues compared to blood in neonates and infants compared to adults {Fernandez, 2011,
9641878}. This is well known for pharmaceuticals, but PFOS is unlike most pharmaceuticals in
that it undergoes extensive protein interaction, such that its distribution in the body is driven
primarily by protein binding and active transport. Hence, it is difficult to infer the degree to
which increased body water content will impact the distribution of PFOS.
The updated half-life value was developed based upon a review of recent literature (see Section
3.3.1.4.5). Many half-life values have been reported for the clearance of PFOS in humans (see
PFOS Appendix). The slow excretion of PFOS requires measurement of a small change in serum
concentration over a long time; the difficulties associated with making these measurements may
represent one reason for the variance in reported values. Another challenge is the ubiquity of
PFOS exposure. Ongoing exposure will result in a positive bias in observed half-life values if not
considered {Russell, 2015, 2851185}. In studies that calculate the half-life in a population with
greatly decreased PFOS exposures, typically due to the end of occupational exposure or the
introduction of drinking water filtration, the amount of bias due to continuing exposure will be
related to the ratio of the prior and ongoing exposure. That is, for a given ongoing exposure, a
higher prior exposure may be less likely to overestimate half-life compared to a lower prior
exposure. However, a half-life value determined from a population with very high exposure may
not be informative of the half-life in typical exposure scenarios because of non-linearities in PK
that may occur due to the saturation of PFAS-protein interactions. This will likely take the form
of an under-estimation of the half-life that is relevant to lower levels, which are more
representative of the general population, due to saturation of renal resorption and increased
urinary clearance in the study population.
Because the derivation of the Vd for PFOS relied on the value for PFOA, it is important to
consider alternate values for Vd for PFOA. For PFOA, the Vd calculation depended on the half-
life. Thompson et al. (2010, 2919278) used 2.3 years, which was estimated within their
population. If EPA chosen half-life of 2.7 years was used instead, the Vd for PFOA would be
200 mL/kg, which results in a PFOS value of 271 mL/kg. EPA did not update the Vd values
based on the updated half-life because the value of 2.3 years was calculated based on the same
data as the Vd and this half-life may be more representative of that population at that specific
time. Gomis et al. (2017, 3981280) also calculated Vd by taking the average of reported animal
and human values and estimated values of 235 mL/kg for PFOS. This calculation included the
value from Thompson et al. (2010, 2919278) and did not include additional values derived from
human data. This average value shows that the value from Thompson et al. (2010, 2919278),
which was selected based on the fact that it was derived only from human and non-human
primate data, is reasonable.
Lastly, the description of breastfeeding in the updated Verner et al. (2016, 3299692) model relied
on a number of assumptions: that infants were exclusively breastfed for one year, that there was
a constant relationship between maternal serum and breastmilk PFOS concentrations, and that
weaning was an immediate process with the infant transitioning from a fully breastmilk diet to
the background exposure at one year. This is a relatively long duration of breastfeeding, only
27% of children in the U.S. are being breastfed at one year of age {CDC, 2013, 1936457}. Along
with using the 95th percentile of breastmilk consumption, this provides a scenario of high but
6-13
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
realistic lactational exposure. Lactational exposure to the infant is much greater than background
exposure so the scenario of long breastfeeding is a conservative approach and will result in a
lower PODhed than a scenario with earlier weaning. Children in the U.S. are very unlikely to be
exclusively breastfed for up to one year, and this approach does not account for potential PFOS
exposure via the introduction of solid foods. However, since lactational exposure is much greater
than exposure after weaning, a breastfeeding scenario that does not account for potential PFOS
exposure from introduction of infants to solid foods is not expected to introduce substantial error.
6.63 Approach of Estimating a Benchmark Dose from a
Regression Coefficient
EPA identified epidemiological studies (e.g., Steenland et al. (2009, 1291109)) that reported
associations between PFOS exposure and diseases or clinical outcomes as regression
coefficients. BMD modeling of regression coefficients results in a non-traditional BMD, where
the BMR is associated with a change in the regression coefficient of the response variable rather
than the measured biological response variable. As a result, there is some uncertainty about the
biological relevance of this non-traditional BMD associated with a regression coefficient.
However, as this regression coefficient is associated with a change in the biological response
variable, it is biologically meaningful and EPA concluded that it can therefore be used for POD
derivation. EPA modeled these regression coefficients using the same approach that EPA used to
model for studies that reported measured response variables which is similar to the approach
followed by CalEPA in their draft Public Health Goal for PFOS {CalEPA, 2021, 9416932}.
To evaluate this potential uncertainty, EPA obtained the measured dose response data across
exposure deciles from Steenland et al. (2009, 1291109) (kindly provided to EPA on June 30,
2022 via email communication with the corresponding study author) and conducted sensitivity
analyses to compare BMDs produced by the reported regression coefficients with the measured
response variable (i.e., mean total cholesterol and odds ratios of elevated total cholesterol). These
analyses are presented in detail in the PFOS Appendix.
For PFOS, BMDLs values estimated using the regression coefficient and using the measured
response variable were 9.52 ng/L and 26.39 ng/L, respectively. The two BMDL estimates from
the two approaches are within an order of magnitude, less than a 3-fold difference, and the RfD
allows for an order of magnitude (10-fold or 1,000%) uncertainty in the estimate. Therefore,
EPA is confident in its use regression coefficients as the basis of PODheds.
6.7 Human Dosimetry Models: Consideration of Alternate
Modeling Approaches
PBPK models are typically preferred over a one-compartment approach because they can
provide individual tissue information and have a one-to-one correspondence with the biological
system that can be used to incorporate additional features of pharmacokinetics, including tissue-
specific internal dosimetry and local metabolism. In addition, though PBPK models are more
complex than one-compartment models, many of the additional parameters are chemical-
independent and have widely accepted values. Even some of the chemical-dependent values can
be extrapolated from animal toxicological studies when parameterizing a model for humans,
where data are typically scarcer.
6-14
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
The decision to select a non-physiologically based model as opposed to one of the PBPK models
was influenced in part by past issues identified during evaluation of the application of PBPK
models to other PFAS for the purpose of risk assessment. During the process of adapting a
published PBPK model for EPA needs, models are subjected to an extensive EPA internal QA
review. During initial review of the Loccisano family of models {Loccisano, 2011, 787186;
Loccisano, 2012, 1289830; Loccisano, 2012, 1289833; Loccisano, 2013, 1326665}, an unusual
implementation of PFOS plasma binding appeared to introduce a mass balance error. Due to the
stated goal of minimizing new model development (see Section 4.1.3.2), EPA did not pursue
resolution of the discrepancies, which would have required modifications to one of these models
for application in this assessment.
A new publication describing a developmental PBPK model in rats and humans was also
evaluated for this effort {Chou, 2021, 7542658}. This model used the in vitro extrapolation that
was previously developed by Worley et al. (2015, 3981311) for PFOA as an initial point for
parameter optimization for PFOS. The complex nature of this renal model, with processes for
resorption, secretion, and passive diffusion presented multiple competing options for
parameterization based on the available human data. Specifically, the set of available model
parameters can take numerous values that fit the human observations equally well. However,
when the model is applied within similar conditions to the human observations, predicting the
exact values of the parameters may not impact the model's ability to predict the targeted
biomarkers (i.e., human milk, fetal serum, and maternal serum). For our purposes, it was not
clear, whether the exposure and internal doses that needed modeling would be within the bounds
of the doses used to parameterize the Chou et al. (2021, 7542658) model.
Due to the previous issues that EPA encountered for other PFAS when implementing PBPK
models, the known issue with the Loccisano model and the models based upon it, and the
concerns about application of the Chou et al. (2021, 7542658) model outside its original
parameterization space, EPA concluded that a one-compartment model was the strongest
approach to predict blood (or serum/plasma) concentrations. Serum/plasma is a good biomarker
for exposure, because a major proportion of the PFOS in the body is found in serum/plasma due
to albumin binding {Forsthuber, 2020, 6311640}. There were no other specific tissues that were
considered essential to describe the dosimetry of PFOS. A full PBPK model can predict serum
concentrations equally well, but with many more parameters, many of which are difficult to
predict for PFOS due to parameter identifiability issues. PFOS presents an unusually high barrier
in this regard because much of its PK is dependent on the interaction between PFOS and proteins
in the form of binding {Frosthuber, 2020, 6311640} and active transport {Zhao, 2017,
3856461}. These protein interactions are more difficult to extrapolate from animal toxicological
studies to humans than PK that is dependent on blood flow and passive diffusion.
The only one-compartment approach identified in the literature for PFOS was the model of
Verner et al. (2016, 3299692). EPA also considered the model developed by the Minnesota
Department of Health (MDH model), which was published as a PFOA model, but has been
applied to other PFAS, including PFOS {Goeden, 2019, 5080506}. These two models are
structurally very similar, with a single compartment each for mother and child, first-order
excretion from those compartments, and a similar methodology for describing lactational transfer
from mother to child. The following paragraphs describe the slight differences in model
implementations, but it is first worth emphasizing the similarity in the two approaches. The
6-15
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
overall agreement in approach supports its validity for the task of human health risk assessment
for PFOS.
One advantage of the Verner model is that it explicitly models the mother from birth through the
end of breastfeeding. The MDH model, however, is limited to predictions for the time period
after the birth of the child with maternal levels set to an initial steady-state level. An explicit
description of maternal blood levels allows for the description of accumulation in the mother
prior to pregnancy followed by decreasing maternal levels during pregnancy, as has been
observed for serum PFOS in serial samples from pregnant women {Glynn, 2012, 1578498}. This
decrease occurs due to the relatively rapid increase in body weight during pregnancy (compared
to the years preceding pregnancy) and the increase in blood volume that occurs to support fetal
growth { Sibai, 1995, 1101373}. Detailed modeling of this period is important for dose metrics
based on maternal levels during pregnancy, especially near term, and on cord blood levels.
Another distinction of the Verner model is that it is written in terms of rates of change in mass
rather than concentrations, as in the MDH model. This approach includes the effect of dilution of
PFOS during childhood growth, without the need for an explicit term in the equations. Not
accounting for growth will result in the overprediction of serum concentration in individuals
exposed during growth. Despite this, PFOS concentration in infants at any specific time is driven
more by recent lactational exposure than by earlier exposure (either during pregnancy or early
breastfeeding), which tends to minimize the impact of growth dilution. Additionally, this
structural consideration best matches the approach taken in our animal model, presenting a
harmonized approach. These structural considerations favor the application of the updated
Verner model over the MDH model.
EPA evaluated two other factors that were present in the MDH model: the application of a
scaling factor to increase the Vd in children and the treatment of exposure as a drinking water
intake rather than a constant exposure relative to bodyweight. After testing these features within
the updated Verner model structure, EPA determined that neither of these features were
appropriate for this assessment, primarily because they did not meaningfully improve the
comparison of model predictions to validation data.
In the MDH model, Vd in children starts at 2.4 times the adult Vd and decreases relatively
quickly to 1.5 times the adults Vd between 6 and 12 months, reaching the adult level at 10 years
of age. These scaling values originated from measurements of body water content relative to
weight compared to the adult value. There is no chemical-specific information to suggest that Vd
is larger in children compared to adults for PFOS. However, it is generally accepted in
pharmaceutical research that hydrophilic chemicals have greater Vd in children {Batchelor, 2015,
3223516}, which is attributed to increased body water. Still, PFOS is amphiphilic, not simply
hydrophilic, and its distribution is driven by interactions with binding proteins and transporters,
not by passive diffusion with body water. While it is plausible that Vd is larger in children, it is
unknown to what degree.
Since increased Vd in children is plausible, but neither supported nor contradicted by direct
evidence, EPA evaluated the effect of variable Vd by implementing this change in the updated
Verner model and comparing the results with constant and variable Vd (see PFOS Appendix).
This resulted in reduced predictions of serum concentrations, primarily during their peak in early
childhood. The model with variable Vd did not decrease the average relative error or the average
6-16
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
absolute value of relative error compared to the model with constant Vd (with PFOA and PFOS
results combined). Since the model with constant Vd had marginally better performance and was
an overall simpler solution, EPA did not implement variable Vd in the application of the model
for PODhed calculation.
The other key difference between the MDH model and the updated Verner model is that instead
of constant exposure relative to body weight, exposure in the MDH model was based on drinking
water consumption, which is greater relative to bodyweight in young children compared to
adults. Drinking water consumption is also greater in lactating women. To evaluate the potential
impact of calculating a drinking water concentration directly, bypassing the RfD step, EPA
implemented drinking water consumption in the modified Verner model (see PFOS Appendix).
EPA evaluated this decision for PFOA and PFOS together because the choice of units used for
human exposure represents a substantial difference in risk assessment methodology. For reasons
explained below, EPA ultimately decided to continue to calculate an RfD in terms of constant
exposure, with an MCLG calculated thereafter using life-stage specific drinking water
consumption values.
When comparing exposure based on drinking water consumption to the traditional RfD
approach, the impact on the serum concentrations predicted by the updated Verner model
differed between PFOA and PFOS. For PFOA, the predicted serum concentration in the child
was qualitatively similar, with the main effect seen in overprediction of timepoints that occur
later in childhood. These timepoints are more susceptible to changes in exposure as early
childhood exposure is dominated by lactational exposure. Lactational exposure is slightly
increased in this scenario, because of increased drinking water consumption during lactation.
However, the main source of PFOA or PFOS in breastmilk in the model with exposure based on
drinking water consumption is that which accumulated over the mother's life prior to childbirth,
not that which was consumed during lactation. For PFOS, the increased exposure predicted
based on children's water intake results in much greater levels in later childhood compared to the
model with constant exposure relative to bodyweight. Use of water ingestion rates to adjust the
dose in the Verner model fails to match the decrease in PFOS concentration present in the
reported data with multiple timepoints and overestimates the value for the Norwegian Mother,
Father, and Child Cohort Study (MoBa) cohort with a single timepoint. There is a much greater
effect on PFOS model results relative to PFOA. This comparison suggests that incorporating
variations in drinking water exposure in this way is not appropriate for the updated Verner
model.
In addition to the comparison with reported data, EPA's decision to use the Verner model was
also considered in the context of the effect on the derivation of MCLGs. The epidemiological
endpoints can be placed into three categories based on the age of the individuals: adults,
children, and pregnant women. Because increased drinking water exposure is only applied to
children and lactating women, the group of endpoints in children are the only ones that would be
affected. While the RfD estimated using the updated Verner model assumed constant exposure,
the MCLG is an algebraic calculation that incorporates the RfD, RSC, and drinking water intake.
The drinking water intake used for the MCLG calculation is chosen based on the target
population relevant to the critical effect that serves as the basis of the RfD. Therefore, even if the
RfD does not incorporate increased drinking water intake in certain lifestages, the subsequent
6-17
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
MCLG calculation does take this into account. Furthermore, derivation of an RfD is useful for
general assessment of risk and not limited to drinking water exposure.
For these reasons and based on EPA's analyses, EPA determined that the updated Verner model
was the most appropriate available model structure for PODhed calculation for PFOS. Including
the determination that assuming Vd in children equal to the adult values was appropriate, and that
calculating a RfD assuming a constant dose (mg/kg/day) was appropriate for this assessment.
6.8 Sensitive Populations
Some populations may be more susceptible to the potential adverse health effects of toxic
substances such as PFOS. These potentially susceptible populations include populations
exhibiting a greater response than others despite similar PFOS exposure due to increased
biological sensitivity, as well as populations exhibiting a greater response due to higher PFOS
exposure and/or exposure to other chemicals or non-chemical stressors. Populations with greater
biological sensitivity may include pregnant women and their developing fetuses, lactating
women, the elderly, and people with certain underlying medical conditions (see Section 6.8.1).
Populations that could exhibit a greater response to PFOS exposure due to higher exposures to
PFOS or other chemicals include communities overburdened by chemical exposures or non-
chemical stressors such as communities with environmental justice concerns (see Section 6.8.2).
The potential health effects after PFOS exposure have been evaluated in some sensitive
populations (e.g., pregnant women, children) and a small number of studies have assessed
differences in exposure to PFOS across populations to assess whether racial/ethnic or
socioeconomic differences are associated with greater PFOS exposure. However, the available
research on PFOS's potential impacts on sensitive populations is limited and more research is
needed. Health effects differences in sensitivity to PFOS exposure have not allowed for the
identification or characterization of all potentially sensitive subpopulations. This lack of
knowledge about susceptibility to PFOS represents a potential source of uncertainty in the
assessment of PFOS.
6.8.1 Fetuses, Infants, Children
One of the more well-studied sensitive populations to PFOS exposure is developing fetuses,
infants, and children. Both animal toxicological and epidemiological data suggest that the
developing fetus is particularly sensitive to PFOS-induced toxicity. As described in Section
3.4.4.1, results of some epidemiological studies indicate an association between PFOS exposure
during pregnancy and adverse birth outcomes such as low birth weight, and studies of PFOS
exposure during early childhood, which may also reflect in utero exposure, suggest an
association between PFOS exposure and effects on development, including immune system
development (Section 3.4.2.1). The available animal toxicological data lend support to these
findings; as described in Section 3.4.4.2, numerous studies in rodents report effects similar to
those seen in humans (e.g., decreased body weights in offspring exposed to PFOS during
gestation). Additionally, PFOS exposure during certain life stages or exposure windows (e.g.,
prenatal or early postnatal exposure windows) may be more consequential than others. For
example, as described in Section C.7.2 of the PFOS Appendix, Grasty et al. (2003, 1332670;
2005, 2951495) identified GD 19-21 as a critical exposure window for neonatal lung
development and subsequent neonatal mortality in rats. These potentially different effects in
6-18
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
different populations and/or exposure windows have not been fully characterized. More research
is needed to fully understand the specific critical windows of exposure during development.
With respect to the decreased antibody production endpoint, children who have autoimmune
diseases (e.g., juvenile arthritis) or are taking medications that weaken the immune system would
be expected to be more likely to mount a low antibody response and would therefore represent
potentially susceptible populations for PFOS exposure. There are also concerns about declines in
vaccination status {Smith, 2011, 9642143; Bramer, 2020, 9642145} for children overall, and the
possibility that diseases which are considered eradicated (such as diphtheria or tetanus) could
return to the United States {Hotez, 2019, 9642144}. As noted by Dietert et al. (2010, 644213),
the risks of developing infectious diseases may increase if immunosuppression occurs in the
developing immune system.
6.8.2 Other Susceptible Populations
As noted in the SAB PFAS review panel's final report {U.S. EPA, 2022, 10476098}, there is
uncertainty about whether there are susceptible populations, such as certain racial/ethnic groups,
that might be more sensitive to the health effects of PFOS exposure because of either greater
biological sensitivity or higher exposure to PFOS and/or other environmental chemicals.
Although some studies have evaluated differences in PFAS exposure levels across SES and
racial/ethnic groups (see Section 6.1), studies of differential health effects incidence and PFOS
exposure are limited. To fully address equity and environmental justice concerns about PFOS,
these data gaps regarding differential exposure and health effects after PFOS exposure need to be
addressed. In the development of the proposed PFAS NPDWR, EPA conducted an analysis to
evaluate potential environmental justice impacts of the proposed regulation (See Chapter 8 of the
Economic Analysis for the Proposed PFAS National Primary Drinking Water Regulation {U.S.
EPA, 2023, 10692765}). EPA acknowledges that exposure to PFOS, and PFAS in general, may
have a disproportionate impact on certain communities (e.g., low SES communities; tribal
communities; minority communities; communities in the vicinity of areas of historical PFOS
manufacturing and/or contamination) and that studies of these communities are high priority
research needs.
6-19
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
7 References
3M. (2000). Determination of serum half-lives of several fluorochemicals, Interim report #1,
June 8, 2000 [TSCA Submission], In TSCA 8(e) Supplemental Notice for Sulfonate-
based and Carboxylic-based Fluorochemicals-DocketNumbers 8EHQ-1180-373; 8EHQ-
1180-374; 8EHQ-0381- 0394; 8EHQ-0598-373. (8EHQ-80-373. 8EHQ-0302-00373.
89(811844Q). AR226-0611).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/8568548
3M. (2002). Determination of serum half-lives of several fluorochemicals, Interim report #2,
January 11, 2002 [TSCA Submission], In TSCA 8(e) Supplemental Notice for Sulfonate-
based and Carboxylic-based Fluorochemicals-DocketNumbers 8EHQ-1180-373; 8EHQ-
1180-374; 8EHQ-0381- 0394; 8EHQ-0598-373. (8EHQ-80-373. 8EHQ-0302-00373.
89(811844Q). AR226-1086).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6574114
Abbott, BD; Wolf, CJ; Das, KP; Zehr, RD; Schmid, JE; Lindstrom, AB; Strynar, MJ; Lau, C.
(2009). Developmental toxicity of perfluorooctane sulfonate (PFOS) is not dependent on
expression of peroxisome proliferator activated receptor-alpha (PPAR alpha) in the
mouse. Reprod Toxicol 27: 258-265.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919376
Abdallah, MA; Wemken, N; Drage, DS; Tlustos, C; Cellarius, C; Cleere, K; Morrison, JJ; Daly,
S; Coggins, MA; Harrad, S. (2020). Concentrations of perfluoroalkyl substances in
human milk from Ireland: Implications for adult and nursing infant exposure.
Chemosphere 246: 125724.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6316215
Abdullah Soheimi, SS; Abdul Rahman, A; Abd Latip, N; Ibrahim, E; Sheikh Abdul Kadir, SH.
(2021). Understanding the impact of perfluorinated compounds on cardiovascular
diseases and their risk factors: A meta-analysis study [Review], Int J Environ Res Public
Health 18: 8345.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9959584
Abraham, K; Mielke, H; Fromme, H; Volkel, W; Menzel, J; Peiser, M; Zepp, F; Willich, SN;
Weikert, C. (2020). Internal exposure to perfluoroalkyl substances (PFASs) and
biological markers in 101 healthy 1-year-old children: associations between levels of
perfluorooctanoic acid (PFOA) and vaccine response. Arch Toxicol 94: 2131-2147.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6506041
Agier, L; Basagana, X; Maitre, L; Granum, B; Bird, PK; Casas, M; Oftedal, B; Wright, J;
Andrusaityte, S; de Castro, M; Cequier, E; Chatzi, L; Donaire-Gonzalez, D;
Grazuleviciene, R; Haug, LS; Sakhi, AK; Leventakou, V; Mceachan, R; Nieuwenhuijsen,
M; Petraviciene, I; Robinson, O; Roumeliotaki, T; Sunyer, J; Tamayo-Uria, I; Thomsen,
C; Urquiza, J; Valentin, A; Slama, R; Vrijheid, M; Siroux, V. (2019). Early-life
exposome and lung function in children in Europe: an analysis of data from the
longitudinal, population-based HELIX cohort. The Lancet Planetary Health 3: e81-e92.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5043613
Ahrens, L. (2011). Polyfluoroalkyl compounds in the aquatic environment: a review of their
occurrence and fate [Review], J Environ Monit 13: 20-31.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2657780
Ahrens, L; Shoeib, M; Harner, T, om; Lee, S; Guo, R, ui; Reiner, EJ. (2011). Wastewater
Treatment Plant and Landfills as Sources of Polyfluoroalkyl Compounds to the
7-20
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Atmosphere. Environ Sci Technol 45: 8098-8105.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2325317
Aimuzi, R; Luo, K; Chen, Q; Wang, H; Feng, L; Ouyang, F; Zhang, J. (2019). Perfluoroalkyl
and polyfluoroalkyl substances and fetal thyroid hormone levels in umbilical cord blood
among newborns by prelabor caesarean delivery. Environ Int 130: 104929.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5387078
Aimuzi, R; Luo, K; Huang, R; Huo, X; Nian, M; Ouyang, F; Du, Y; Feng, L; Wang, W; Zhang,
J. (2020). Perfluoroalkyl and polyfluroalkyl substances and maternal thyroid hormones in
early pregnancy. Environ Pollut 264: 114557.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6512125
Ait Bamai, Y; Goudarzi, H; Araki, A; Okada, E; Kashino, I; Miyashita, C; Kishi, R. (2020).
Effect of prenatal exposure to per- and polyfluoroalkyl substances on childhood allergies
and common infectious diseases in children up to age 7 years: The Hokkaido study on
environment and children's health. Environ Int 143: 105979.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833636
Alam, MN; Han, X; Nan, B; Liu, L; Tian, M; Shen, H; Huang, Q. (2021). Chronic low-level
perfluorooctane sulfonate (PFOS) exposure promotes testicular steroidogenesis through
enhanced histone acetylation. Environ Pollut 284: 117518.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9959508
Alderete, TL; Jin, R; Walker, DI; Valvi, D; Chen, Z; Jones, DP; Peng, C; Gilliland, FD; Berhane,
K; Conti, DV; Goran, MI; Chatzi, L. (2019). Perfluoroalkyl substances, metabolomic
profiling, and alterations in glucose homeostasis among overweight and obese Hispanic
children: A proof-of-concept analysis. Environ Int 126: 445-453.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080614
Alexander, BH; Olsen, GW. (2007). Bladder cancer in perfluorooctanesulfonyl fluoride
manufacturing workers. Ann Epidemiol 17: 471-478.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4727072
Alexander, BH; Olsen, GW; Burris, JM; Mandel, JH; Mandel, JS. (2003). Mortality of
employees of a perfluorooctanesulphonyl fluoride manufacturing facility. Occup Environ
Med 60: 722-729.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291101
Ammitzb0ll, C; Bornsen, L; Petersen, ER; Oturai, AB; S0ndergaard, HB; Grandjean, P;
Sellebjerg, F. (2019). Perfluorinated substances, risk factors for multiple sclerosis and
cellular immune activation. J Neuroimmunol 330: 90-95.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080379
Anand-Ivell, R; Cohen, A; N0rgaard-Pedersen, B; Jonsson, BAG; Bonde, JP; Hougaard, DM;
Lindh, CH; Toft, G; Lindhard, MS; Ivell, R. (2018). Amniotic fluid INSL3 measured
during the critical time window in human pregnancy relates to cryptorchidism,
hypospadias, and phthalate load: a large case-control study. Front Physiol 9: 406.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4728675
Andersen, ME; Butenhoff, JL; Chang, SC; Farrar, DG; Kennedy, GL; Lau, C; Olsen, GW; Seed,
J; Wallace, KB. (2008). Perfluoroalkyl acids and related chemistries—toxicokinetics and
modes of action. Toxicol Sci 102: 3-14.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3749214
Andersen, ME; Clewell, HJ; Tan, YM; Butenhoff, JL; Olsen, GW. (2006). Pharmacokinetic
modeling of saturable, renal resorption of perfluoroalkylacids in monkeys—probing the
7-21
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
determinants of long plasma half-lives. Toxicology 227: 156-164.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/818501
Apelberg, BJ; Goldman, LR; Calafat, AM; Herbstman, JB; Kuklenyik, Z; Heidler, J; Needham,
LL; Halden, RU; Witter, FR. (2007). Determinants of fetal exposure to polyfluoroalkyl
compounds in Baltimore, Maryland. Environ Sci Technol 41: 3891-3897.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290900
Arbuckle, T; Kubwabo, C; Walker, M; Davis, K; Lalonde, K; Kosarac, I; Wen, S; Arnold, DL.
(2013). Umbilical cord blood levels of perfluoroalkyl acids and polybrominated flame
retardants. Int J Hyg Environ Health 216: 184-194.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2152344
Arbuckle, TE; Macpherson, S; Foster, WG; Sathyanarayana, S; Fisher, M; Monnier, P;
Lanphear, B; Muckle, G; Fraser, WD. (2020). Prenatal Perfluoroalkyl Substances and
Newborn Anogenital Distance in a Canadian Cohort. Reprod Toxicol 94: 31-39.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6356900
Argus Research Laboratories. (2000). Support: Oral (stomach tube) developmental toxicity study
of PFOS in rabbits with attachments and cover letter dated 0425000. (EPA/OTS; Doc
#8EHQ-0800-0373).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080012
Armstrong, LE; Guo, GL. (2019). Understanding environmental contaminants' direct effects on
non-alcoholic fatty liver disease progression. Curr Environ Health Rep 6: 95-104.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6956799
Arrebola, JP; Ramos, JJ; Bartolome, M; Esteban, M; Huetos, O; Canas, AI; Lopez-Herranz, A;
Calvo, E; Perez-Gomez, B; Castano, A; BIOAMBIENT.ES. (2019). Associations of
multiple exposures to persistent toxic substances with the risk of hyperuricemia and
subclinical uric acid levels in BIOAMBIENT.ES study. Environ Int 123: 512-521.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080503
Ashley-Martin, J; Dodds, L; Arbuckle, TE; Bouchard, MF; Fisher, M; Morriset, AS; Monnier, P;
Shapiro, GD; Ettinger, AS; Dallaire, R; Taback, S; Fraser, W; Piatt, RW. (2017).
Maternal concentrations of perfluoroalkyl substances and fetal markers of metabolic
function and birth weight. Am J Epidemiol 185: 185-193.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981371
Ashley-Martin, J; Dodds, L; Arbuckle, TE; Morisset, AS; Fisher, M; Bouchard, MF; Shapiro,
GD; Ettinger, AS; Monnier, P; Dallaire, R; Taback, S; Fraser, W. (2016). Maternal and
Neonatal Levels of Perfluoroalkyl Substances in Relation to Gestational Weight Gain. Int
J Environ Res Public Health 13: 146.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859831
ATSDR. (2021). Toxicological profile for perfluoroalkyls [ATSDR Tox Profile], Atlanta, GA:
U.S. Department of Health and Human Services, Public Health Service.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9642134
Attanasio, R. (2019). Sex differences in the association between perfluoroalkyl acids and liver
function in US adolescents: Analyses of NHANES 2013-2016. Environ Pollut 254:
113061. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5412069
Audet-Delage, Y; Ouellet, N; Dallaire, R; Dewailly, E; Ayotte, P. (2013). Persistent organic
pollutants and transthyretin-bound thyroxin in plasma of inuit women of childbearing
age. Environ Sci Technol 47: 13086-13092.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2149477
7-22
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Averina, M; Brox, J; Huber, S; Furberg, AS. (2021). Exposure to perfluoroalkyl substances
(PFAS) and dyslipidemia, hypertension and obesity in adolescents. The Fit Futures study.
Environ Res 195: 110740.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7410155
Averina, M; Brox, J; Huber, S; Furberg, AS; S0rensen, M. (2019). Serum perfluoroalkyl
substances (PFAS) and risk of asthma and various allergies in adolescents. The Troms0
study Fit Futures in Northern Norway. Environ Res 169: 114-121.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080647
Aylward, LL; Hays, SM; Kirman, CR; Marchitti, SA; Kenneke, JF; English, C; Mattison, DR;
Becker, RA. (2014). Relationships of chemical concentrations in maternal and cord
blood: a review of available data [Review], J Toxicol Environ Health B Crit Rev 17: 175-
203. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2920555
Bach, C; Matthiesen, B; Olsen; Henriksen, B. (2018). Conditioning on parity in studies of
perfluoroalkyl acids and time to pregnancy: an example from the danish national birth
cohort. Environ Health Perspect 126: 117003.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080557
Bach, CC; Bech, BH; Nohr, EA; Olsen, J; Matthiesen, NB; Bossi, R; Uldbjerg, N; Bonefeld-
J0rgensen, EC; Henriksen, TB. (2015). Serum perfluoroalkyl acids and time to pregnancy
in nulliparous women. Environ Res 142: 535-541.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981559
Bach, CC; Liew, Z; Bech, BH; Nohr, EA; Fei, C; Bonefeld-Jorgensen, EC; Henriksen, TB;
Olsen, J. (2015). Perfluoroalkyl acids and time to pregnancy revisited: An update from
the Danish National Birth Cohort. Environ Health 14: 59.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981738
Bagley, BD; Chang, SC; Ehresman, DJ; Eveland, A; Zitzow, JD; Parker, GA; Peters, JM;
Wallace, KB; Butenhoff, JL. (2017). Perfluorooctane Sulfonate-Induced Hepatic
Steatosis in Male Sprague Dawley Rats Is Not Attenuated by Dietary Choline
Supplementation. Toxicol Sci 160: 284-298.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238503
Balk, FGP; Winkens Piitz, K; Ribbenstedt, A; Gomis, MI; Filipovic, M; Cousins, IT. (2019).
Children's exposure to perfluoroalkyl acids - a modelling approach. Environ Sci Process
Impacts 21: 1875-1886.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5918617
Bangma, J; Eaves, LA; Oldenburg, K; Reiner, JL; Manuck, T; Fry, RC. (2020). Identifying Risk
Factors for Levels of Per- and Polyfluoroalkyl Substances (PFAS) in the Placenta in a
High-Risk Pregnancy Cohort in North Carolina. Environ Sci Technol 54: 8158-8166.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833725
Bao, WW; Qian, ZM; Geiger, SD; Liu, E; Liu, Y; Wang, SQ; Lawrence, WR; Yang, BY; Hu,
LW; Zeng, XW; Dong, GH. (2017). Gender-specific associations between serum isomers
of perfluoroalkyl substances and blood pressure among Chinese: Isomers of C8 Health
Project in China. Sci Total Environ 607-608: 1304-1312.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860099
Barber, JL; Berger, U; Chaemfa, C; Huber, S; Jahnke, A; Temme, C; Jones, KC. (2007).
Analysis of per- and polyfluorinated alkyl substances in air samples from Northwest
Europe. J Environ Monit 9: 530-541.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1049488
7-23
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Bartell, SM; Calafat, AM; Lyu, C; Kato, K; Ryan, PB; Steenland, K. (2010). Rate of Decline in
Serum PFOA Concentrations after Granular Activated Carbon Filtration at Two Public
Water Systems in Ohio and West Virginia. Environ Health Perspect 118: 222-228.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/379025
Bassler, J; Ducatman, A; Elliott, M; Wen, S; Wahlang, B; Barnett, J; Cave, MC. (2019).
Environmental perfluoroalkyl acid exposures are associated with liver disease
characterized by apoptosis and altered serum adipocytokines. Environ Pollut 247: 1055-
1063. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080624
Batchelor, HK; Marriott, JF. (2015). Paediatric pharmacokinetics: key considerations. Br J Clin
Pharmacol 79: 395-404.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3223516
Beach, SA; Newsted, JL; Coady, K; Giesy, JP. (2006). Ecotoxicological evaluation of
perfluorooctanesulfonate (PFOS) [Review], Rev Environ Contam Toxicol 186: 133-174.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290843
Beck, IH; Timmermann, CAG; Nielsen, F; Schoeters, G; J0hnk, C; Kyhl, HB; Ftest, A; Jensen,
TK. (2019). Association between prenatal exposure to perfluoroalkyl substances and
asthma in 5-year-old children in the Odense Child Cohort. Environ Health 18: 97.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5922599
Beesoon, S; Martin, JW. (2015). Isomer-Specific Binding Affinity of Perfluorooctanesulfonate
(PFOS) and Perfluorooctanoate (PFOA) to Serum Proteins. Environ Sci Technol 49:
5722-5731. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850292
Beesoon, S; Webster, GM; Shoeib, M; Harner, T, om; Benskin, JP; Martin, JW. (2011). Isomer
Profiles of Perfluorochemicals in Matched Maternal, Cord, and House Dust Samples:
Manufacturing Sources and Transplacental Transfer. Environ Health Perspect 119: 1659-
1664. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2050293
Beggs, KM; Mcgreal, S. R.; McCarthy, A; Gunewardena, S; Lampe, JN; Lau, C; Apte, U. (2016).
The role of hepatocyte nuclear factor 4-alpha in perfluorooctanoic acid- and
perfluorooctanesulfonic acid-induced hepatocellular dysfunction. Toxicol Appl
Pharmacol 304: 18-29.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981474
Behr, AC; Kwiatkowski, A; Stahlman, M; Schmidt, FF; Luckert, C; Braeuning, A; Buhrke, T.
(2020). Impairment of bile acid metabolism by perfluorooctanoic acid (PFOA) and
perfluorooctanesulfonic acid (PFOS) in human HepaRG hepatoma cells. Arch Toxicol
94: 1673-1686.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6505973
Behr, AC; Plinsch, C; Braeuning, A; Buhrke, T. (2020). Activation of human nuclear receptors
by perfluoroalkylated substances (PFAS). Toxicol In Vitro 62: 104700.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6305866
Benninghoff, AD; Orner, GA; Buchner, CH; Hendricks, JD; Duffy, AM; Williams, DE. (2012).
Promotion of Hepatocarcinogenesis by Perfluoroalkyl Acids in Rainbow Trout. Toxicol
Sci 125: 69-78.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1274145
Benskin, JP; Muir, DCG; Scott, BF; Spencer, C; De Silva, AO; Kylin, H; Martin, JW; Morris, A;
Lohmann, R; Tomy, G; Rosenberg, B; Taniyasu, S; Yamashita, N. (2012). Perfluoroalkyl
Acids in the Atlantic and Canadian Arctic Oceans. Environ Sci Technol 46: 5815-5823.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1274133
7-24
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Berg, V; N0st, TH; Hansen, S; Elverland, A; Veyhe, AS; Jorde, R; Odland, JO; Sandanger, TM.
(2015). Assessing the relationship between perfluoroalkyl substances, thyroid hormones
and binding proteins in pregnant women; a longitudinal mixed effects approach. Environ
Int 77: 63-69. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851002
Berg, V; N0st, TH; Pettersen, RD; Hansen, S; Veyhe, AS; Jorde, R; Odland, JO; Sandanger, TM.
(2017). Persistent organic pollutants and the association with maternal and infant thyroid
homeostasis: a multipollutant assessment. Environ Health Perspect 125: 127-133.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3350759
Berk, M; Williams, LJ; Andreazza, A; Pasco, JA; Dodd, S; Jacka, FN; Moylan, S; Reiner, EJ;
Magalhaes, PVS. (2014). Pop, heavy metal and the blues: secondary analysis of persistent
organic pollutants (POP), heavy metals and depressive symptoms in the NHANES
National Epidemiological Survey. BMJ Open 4: e005142.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2713574
Bernstein, AS; Kapraun, DF; Schlosser, PM. (2021). A Model Template Approach for Rapid
Evaluation and Application of Physiologically Based Pharmacokinetic Models for Use in
Human Health Risk Assessments: A Case Study on Per- and Polyfluoroalkyl Substances.
Toxicol Sci 182: 215-228.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9639956
Berntsen, HF; B0lling, AK; Bj0rklund, CG; Zimmer, K; Ropstad, E; Zienolddiny, S; Becher, R;
Holme, JA; Dirven, H; Nygaard, UC; Bodin, J. (2018). Decreased macrophage
phagocytic function due to xenobiotic exposures in vitro, difference in sensitivity
between various macrophage models. Food Chem Toxicol 112: 86-96.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4167035
Biesemeier, J; Harris, D. (1974). Eye and skin irritation report on sample T-l 117. (Project No.
4102871). Madison, WI: WARF Institute Inc.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4467668
Bijland, S; Rensen, PCN; Pieterman, EJ; Maas, ACE; van Der Hoorn, JW; van Erk, MJ;
Havekes, LM; Willems van Dijk, K; Chang, SC; Ehresman, DJ; Butenhoff, JL; Princen,
HMG. (2011). Perfluoroalkyl sulfonates cause alkyl chain length-dependent hepatic
steatosis and hypolipidemia mainly by impairing lipoprotein production in APOE*3-
Leiden CETP mice. Toxicol Sci 123: 290-303.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1578502
Bjork, JA; Wallace, KB. (2009). Structure-Activity Relationships and Human Relevance for
Perfluoroalkyl Acid-Induced Transcriptional Activation of Peroxisome Proliferation in
Liver Cell Cultures. Toxicol Sci 111: 89-99.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2325339
Bjorke-Monsen, AL; Varsi, K; Averina, M; Brox, J; Huber, S. (2020). Perfluoroalkyl substances
(PFASs) and mercury in never-pregnant women of fertile age: association with fish
consumption and unfavorable lipid profile. 3: 277-284.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7643487
Blake, BE; Pinney, SM; Hines, EP; Fenton, SE; Ferguson, KK. (2018). Associations between
longitudinal serum perfluoroalkyl substance (PFAS) levels and measures of thyroid
hormone, kidney function, and body mass index in the Fernald Community Cohort.
Environ Pollut 242: 894-904.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080657
Blanc, M; Riiegg, J; Scherbak, N; Keiter, SH. (2019). Environmental chemicals differentially
7-25
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
affect epigenetic-related mechanisms in the zebrafish liver (ZF-L) cell line and in
zebrafish embryos. Aquat Toxicol 215: 105272.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5413062
Blomberg, AJ; Shih, YH; Messerlian, C; J0rgensen, LH; Weihe, P; Grandjean, P. (2021). Early-
life associations between per- and polyfluoroalkyl substances and serum lipids in a
longitudinal birth cohort. Environ Res 200: 111400.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/8442228
Bogdanska, J; Borg, D; Sundstrom, M; Bergstrom, U; Halldin, K; Abedi-Valugerdi, M;
Bergman, A; Nelson, B; Depierre, J; Nobel, S. (2011). Tissue distribution of 35S-labelled
perfluorooctane sulfonate in adult mice after oral exposure to a low environmentally
relevant dose or a high experimental dose. Toxicology 284: 54-62.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919253
Bonefeld-Jorgensen, EC; Long, M; Bossi, R; Ayotte, P; Asmund, G; Kriiger, T; Ghisari, M;
Mulvad, G; Kern, P; Nzulumiki, P; Dewailly, E. (2011). Perfluorinated compounds are
related to breast cancer risk in Greenlandic Inuit: a case control study. Environ Health 10:
88. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2150988
Bonefeld-J0rgensen, EC; Long, M; Fredslund, SO; Bossi, R; Olsen, J. (2014). Breast cancer risk
after exposure to perfluorinated compounds in Danish women: a case-control study
nested in the Danish National Birth Cohort. Cancer Causes Control 25: 1439-1448.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851186
Boone, L; Meyer, D; Cusick, P; Ennulat, D; Bolliger, AP; Everds, N; Meador, V; Elliott, G;
Honor, D; Bounous, D; Jordan, H. (2005). Selection and interpretation of clinical
pathology indicators of hepatic injury in preclinical studies [Review], Vet Clin Pathol 34:
182-188. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/782862
Borg, D; Bogdanska, J; Sundstrom, M; Nobel, S; Hakansson, H; Bergman, A; Depierre, JW;
Halldin, K; Bergstrom, U. (2010). Tissue distribution of (35)S-labelled perfluorooctane
sulfonate (PFOS) in C57B1/6 mice following late gestational exposure. Reprod Toxicol
30: 558-565. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919287
Borg, D; Ivarsson, J. (2017). Analysis of PFASs and TOF in Products. (TemaNord 2017:543).
Nordic Council of Ministers.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9416541
Borghese, MM; Walker, M; Helewa, ME; Fraser, WD; Arbuckle, TE. (2020). Association of
perfluoroalkyl substances with gestational hypertension and preeclampsia in the MIREC
study. Environ Int 141: 105789.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833656
Boulanger, B; Peck, AM; Schnoor, JL; Hornbuckle, KC. (2005). Mass budget of perfluorooctane
surfactants in Lake Ontario. Environ Sci Technol 39: 74-79.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1289802
Braun, JM; Chen, A; Romano, ME; Calafat, AM; Webster, GM; Yolton, K; Lanphear, BP.
(2016). Prenatal perfluoroalkyl substance exposure and child adiposity at 8 years of age:
The HOME study. Obesity (Silver Spring) 24: 231-237.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859836
Braun, JM; Kalkbrenner, AE; Just, AC; Yolton, K; Calafat, AM; Sjodin, A; Hauser, R; Webster,
GM; Chen, A; Lanphear, BP. (2014). Gestational exposure to endocrine-disrupting
chemicals and reciprocal social, repetitive, and stereotypic behaviors in 4- and 5-year-old
children: the HOME study. Environ Health Perspect 122: 513-520.
7-26
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2345999
Brochot, C; Casas, M; Manzano-Salgado, C; Zeman, FA; Schettgen, T; Vrijheid, M; Bois, FY.
(2019). Prediction of maternal and foetal exposures to perfluoroalkyl compounds in a
Spanish birth cohort using toxicokinetic modelling. Toxicol Appl Pharmacol 379:
114640. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5381552
Buck, CO; Eliot, MN; Kelsey, KT; Calafat, AM; Chen, A; Ehrlich, S; Lanphear, BP; Braun, JM.
(2018). Prenatal exposure to perfluoroalkyl substances and adipocytokines: the HOME
Study. PediatrRes 84: 854-860.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080288
Buck Louis, GM; Chen, Z; Schisterman, EF; Kim, S; Sweeney, AM; Sundaram, R; Lynch, CD;
Gore-Langton, RE; Barr, DB. (2015). Perfluorochemicals and human semen quality: The
LIFE Study. Environ Health Perspect 123: 57-63.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851189
Buck Louis, GM; Zhai, S; Smarr, MM; Grewal, J; Zhang, C; Grantz, KL; Hinkle, SN; Sundaram,
R; Lee, S; Honda, M; Oh, J; Kannan, K. (2018). Endocrine disruptors and neonatal
anthropometry, NICHD Fetal Growth Studies - Singletons. Environ Int 119: 515-526.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5016992
Buck, RC; Franklin, J; Berger, U; Conder, JM; Cousins, IT; de Voogt, P; Jensen, AA; Kannan,
K; Mabury, SA; van Leeuwen, SP. (2011). Perfluoroalkyl and polyfluoroalkyl substances
in the environment: terminology, classification, and origins [Review], Integr Environ
Assess Manag 7: 513-541.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4771046
Buck, RC; Korzeniowski, SH; Laganis, E; Adamsky, F. (2021). Identification and classification
of commercially relevant per- and poly-fluoroalkyl substances (PFAS). Integr Environ
Assess Manag 17: 1045-1055.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9640864
Budtz-j0rgensen, E; Grandjean, P. (2018). Application of benchmark analysis for mixed
contaminant exposures: Mutual adjustment of perfluoroalkylate substances associated
with immunotoxicity. PLoS ONE 13: e0205388.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5083631
Bulka, CM; Avula, V; Fry, RC. (2021). Associations of exposure to perfluoroalkyl substances
individually and in mixtures with persistent infections: Recent findings from NHANES
1999-2016. Environ Pollut 275: 116619.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7410156
Buser, MC; Scinicariello, F. (2016). Perfluoroalkyl substances and food allergies in adolescents.
Environ Int 88: 74-79.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859834
Butenhoff, JL; Chang, SC; Olsen, GW; Thomford, PJ. (2012). Chronic dietary toxicity and
carcinogenicity study with potassium perfluorooctanesulfonate in Sprague Dawley rats.
Toxicology 293: 1-15.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1276144
Butenhoff, JL; Ehresman, DJ; Chang, SC; Parker, GA; Stump, DG. (2009). Gestational and
lactational exposure to potassium perfluorooctanesulfonate (K+PFOS) in rats:
developmental neurotoxicity. Reprod Toxicol 27: 319-330.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757873
Butenhoff, JL; Kennedy, GL; Hinderliter, PM; Lieder, PH; Jung, R; Hansen, KJ; Gorman, GS;
7-27
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Noker, PE; Thomford, PJ. (2004). Pharmacokinetics of perfluorooctanoate in
cynomolgus monkeys. Toxicol Sci 82: 394-406.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3749227
Butt, CM; Berger, U; Bossi, R; Tomy, GT. (2010). Levels and trends of poly- and perfluorinated
compounds in the arctic environment [Review], Sci Total Environ 408: 2936-2965.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291056
Byrne, S; Seguinot-Medina, S; Miller, P; Waghiyi, V; von Hippel, FA; Buck, CL; Carpenter,
DO. (2017). Exposure to polybrominated diphenyl ethers and perfluoroalkyl substances
in a remote population of Alaska Natives. Environ Pollut 231: 387-395.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4165183
Byrne, SC; Miller, P; Seguinot-Medina, S; Waghiyi, V; Buck, CL; von Hippel, FA; Carpenter,
DO. (2018). Exposure to perfluoroalkyl substances and associations with serum thyroid
hormones in a remote population of Alaska Natives. Environ Res 166: 537-543.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079678
Cai, D; Li, QQ; Chu, C; Wang, SZ; Tang, YT; Appleton, AA; Qiu, RL; Yang, BY; Hu, LW;
Dong, GH; Zeng, XW. (2020). High trans-placental transfer of perfluoroalkyl substances
alternatives in the matched maternal-cord blood serum: Evidence from a birth cohort
study. Sci Total Environ 705: 135885.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6318671
Calafat, AM; Kato, K; Hubbard, K; Jia, T; Botelho, JC; Wong, LY. (2019). Legacy and
alternative per- and polyfluoroalkyl substances in the U.S. general population: Paired
serum-urine data from the 2013-2014 National Health and Nutrition Examination Survey.
Environ Int 131: 105048.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5381304
Calafat, AM; Wong, LY; Kuklenyik, Z; Reidy, JA; Needham, LL. (2007). Polyfluoroalkyl
chemicals in the US population: Data from the National Health and Nutrition
Examination Survey (NHANES) 2003-2004 and comparisons with NHANES 1999-2000.
Environ Health Perspect 115: 1596-1602.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290899
CalEPA. (2021). Public Health Goals: Perfluorooctanoic Acid and Perfluorooctane Sulfonic
Acid in Drinking Water (First Public Review Draft ed.). California Environmental
Protection Agency, Office of Environmental Health Hazard Assessment, Pesticide and
Environmental Toxicology Branch.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9416932
Campbell, S; Raza, M; Pollack, AZ. (2016). Perfluoroalkyl substances and endometriosis in US
women in NHANES 2003-2006. Reprod Toxicol 65: 230-235.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860110
Canova, C; Barbieri, G; Zare Jeddi, M; Gion, M; Fabricio, A; Dapra, F; Russo, F; Fletcher, T;
Pitter, G. (2020). Associations between perfluoroalkyl substances and lipid profile in a
highly exposed young adult population in the Veneto Region. Environ Int 145: 106117.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7021512
Canova, C; Di Nisio, A; Barbieri, G; Russo, F; Fletcher, T; Batzella, E; Dalla Zuanna, T; Pitter,
G. (2021). PFAS Concentrations and Cardiometabolic Traits in Highly Exposed Children
and Adolescents. Int J Environ Res Public Health 18.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10176518
Cao, T; Qu, A; Li, Z; Wang, W; Liu, R; Wang, X; Nie, Y; Sun, S; Zhang, X; Liu, X. (2021). The
7-28
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
relationship between maternal perfluoroalkylated substances exposure and low birth
weight of offspring: a systematic review and meta-analysis. Environ Sci Pollut Res Int
28: 67053-67065.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9959525
Cardenas, A; Gold, DR; Hauser, R; Kleinman, KP; Hivert, MF; Calafat, AM; Ye, X; Webster,
TF; Horton, ES; Oken, E. (2017). Plasma concentrations of per- and polyfluoroalkyl
substances at baseline and associations with glycemic indicators and diabetes incidence
among high-risk adults in the Diabetes Prevention Program trial. Environ Health Perspect
125: 107001. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4167229
Cardenas, A; Hivert, MF; Gold, DR; Hauser, R; Kleinman, KP; Lin, PD; Fleisch, AF; Calafat,
AM; Ye, X; Webster, TF; Horton, ES; Oken, E. (2019). Associations of perfluoroalkyl
and polyfluoroalkyl substances with incident diabetes and microvascular disease.
Diabetes Care 42: 1824-1832.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5381549
Cariou, R; Veyrand, B; Yamada, A; Berrebi, A; Zalko, D; Durand, S; Pollono, C; Marchand, P;
Leblanc, JC; Antignac, JP; Le Bizec, B. (2015). Perfluoroalkyl acid (PFAA) levels and
profiles in breast milk, maternal and cord serum of French women and their newborns.
Environ Int 84: 71-81.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859840
Carmichael, S; Abrams, B; Selvin, S. (1997). The pattern of maternal weight gain in women with
good pregnancy outcomes. Am J Public Health 87: 1984-1988.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1060457
Caron-Beaudoin, E; Ayotte, P; Laouan Sidi, EA; Simon, CoL; Nation, CoWLPF; Nutashkuan,
CTKo; Shipu, CoU; Gros-Louis McHugh, N; Lemire, M. (2019). Exposure to
perfluoroalkyl substances (PFAS) and associations with thyroid parameters in First
Nation children and youth from Quebec. Environ Int 128: 13-23.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5097914
Caserta, D; Bordi, G; Ciardo, F; Marci, R; LaRocca, C; Tait, S; Bergamasco, B; Stecca, L;
Mantovani, A; Guerranti, C; Fanello, EL; Perra, G; Borghini, F; Focardi, SE; Moscarini,
M. (2013). The influence of endocrine disruptors in a selected population of infertile
women. Gynecol Endocrinol 29: 444-447.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2001177
Caserta, D; Ciardo, F; Bordi, G; Guerranti, C; Fanello, E; Perra, G; Borghini, F; La Rocca, C;
Tait, S; Bergamasco, B; Stecca, L; Marci, R; Lo Monte, G; Soave, I; Focardi, S;
Mantovani, A; Moscarini, M. (2013). Correlation of endocrine disrupting chemicals
serum levels and white blood cells gene expression of nuclear receptors in a population of
infertile women. International Journal of Endocrinology 2013: 510703.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2000966
Caserta, D; Pegoraro, S; Mallozzi, M; Di Benedetto, L; Colicino, E; Lionetto, L; Simmaco, M.
(2018). Maternal exposure to endocrine disruptors and placental transmission: a pilot
study. Gynecol Endocrinol 34: 1-4.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4728855
CDC. (2011). Tetanus surveillance — United States, 2001-2008. MMWR Morb Mortal Wkly
Rep 60: 365-369.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9998272
CDC. (2013). Breastfeeding report card : United States 2013. Atlanta, GA.
7-29
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1936457
CDC. (2017). Fourth national report on human exposure to environmental chemicals, updated
tables, January 2017, volume one. Atlanta, GA.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4296146
Qelik, A; Eke, D; Ekinci, SY; Yildirim, S. (2013). The protective role of curcumin on
perfluorooctane sulfonate-induced genotoxicity: single cell gel electrophoresis and
micronucleus test. Food Chem Toxicol 53: 249-255.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919161
Cellesi, C; Michelangeli, C; Rossolini, GM; Giovannoni, F; Rossolini, A. (1989). Immunity to
diphtheria, six to 15 years after a basic three-dose immunization schedule. Journal of
Biological Standardization 17: 29-34.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9642154
Chan, E; Burstyn, I; Cherry, N; Bamforth, F; Martin, JW. (2011). Perfluorinated acids and
hypothyroxinemia in pregnant women. Environ Res 111: 559-564.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1402500
Chang, S; Allen, BC; Andres, KL; Ehresman, DJ; Falvo, R; Provencher, A; Olsen, GW;
Butenhoff, JL. (2017). Evaluation of serum lipid, thyroid, and hepatic clinical chemistries
in association with serum perfluorooctanesulfonate (PFOS) in cynomolgus monkeys after
oral dosing with potassium PFOS. Toxicol Sci 156: 387-401.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981378
Chang, SC; Ehresman, DJ; Bjork, JA; Wallace, KB; Parker, GA; Stump, DG; Butenhoff, JL.
(2009). Gestational and lactational exposure to potassium perfluorooctanesulfonate
(K+PFOS) in rats: toxicokinetics, thyroid hormone status, and related gene expression.
Reprod Toxicol 27: 387-399.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757876
Chang, SC; Noker, PE; Gorman, GS; Gibson, SJ; Hart, JA; Ehresman, DJ; Butenhoff, JL.
(2012). Comparative pharmacokinetics of perfluorooctanesulfonate (PFOS) in rats, mice,
and monkeys. Reprod Toxicol 33: 428-440.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1289832
Chen, A; Jandarov, R; Zhou, L; Calafat, AM; Zhang, G; Urbina, EM; Sarac, J; Augustin, DH;
Caric, T; Bockor, L; Petranovic, MZ; Novokmet, N; Missoni, S; Rudan, P; Deka, R.
(2019). Association of perfluoroalkyl substances exposure with cardiometabolic traits in
an island population of the eastern Adriatic coast of Croatia. Sci Total Environ 683: 29-
36. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5387400
Chen, F; Yin, S; Kelly, BC; Liu, W. (2017). Chlorinated polyfluoroalkyl ether sulfonic acids in
matched maternal, cord, and placenta samples: A study of transplacental transfer. Environ
Sci Technol 51: 6387-6394.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981340
Chen, F; Yin, S; Kelly, BC; Liu, W. (2017). Isomer-specific transplacental transfer of
perfluoroalkyl acids: Results from a survey of paired maternal, cord sera, and placentas.
Environ Sci Technol 51: 5756-5763.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859806
Chen, J; Liu, X; Zhong, Y. (2020). Interleukin-17A: The Key Cytokine in Neurodegenerative
Diseases [Review], Front Aging Neurosci 12: 566922.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284264
Chen, J; Tanguay, RL; Tal, TL; Gai, Z; Ma, X; Bai, C; Tilton, SC; Jin, D; Yang, D; Huang, C;
7-30
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Dong, Q. (2014). Early life perfluorooctanesulphonic acid (PFOS) exposure impairs
zebrafish organogenesis. Aquat Toxicol 150: 124-132.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2540874
Chen, MH; Ha, EH; Liao, HF; Jeng, SF; Su, YN; Wen, TW; Lien, GW; Chen, CY; Hsieh, WS;
Chen, PC. (2013). Perfluorinated compound levels in cord blood and neurodevelopment
at 2 years of age. Epidemiology 24: 800-808.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850933
Chen, Q; Huang, R; Hua, L; Guo, Y; Huang, L; Zhao, Y; Wang, X; Zhang, J. (2018). Prenatal
exposure to perfluoroalkyl and polyfluoroalkyl substances and childhood atopic
dermatitis: A prospective birth cohort study. Environ Health 17: 1-12.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238372
Chen, Q; Zhang, X; Zhao, Y; Lu, W; Wu, J; Zhao, S; Zhang, J; Huang, L. (2019). Prenatal
exposure to perfluorobutanesulfonic acid and childhood adiposity: A prospective birth
cohort study in Shanghai, China. Chemosphere 226: 17-23.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080578
Chen, T; Zhang, L; Yue, JQ; Lv, ZQ; Xia, W; Wan, YJ; Li, YY; Xu, SQ. (2012). Prenatal PFOS
exposure induces oxidative stress and apoptosis in the lung of rat off-spring. Reprod
Toxicol 33: 538-545.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1276152
Chen, YM; Guo, LH. (2009). Fluorescence study on site-specific binding of perfluoroalkyl acids
to human serum albumin. Arch Toxicol 83: 255-261.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1280480
Cheng, J; Lv, S; Nie, S; Liu, J; Tong, S; Kang, N; Xiao, Y; Dong, Q; Huang, C; Yang, D. (2016).
Chronic perfluorooctane sulfonate (PFOS) exposure induces hepatic steatosis in
zebrafish. Aquat Toxicol 176: 45-52.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981479
Cheng, W; Ng, CA. (2018). Predicting relative protein affinity of novel per- and polyfluoroalkyl
substances (PFASs) by an efficient molecular dynamics approach. Environ Sci Technol
52: 7972-7980.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5024207
Cheng, W; Yu, Z; Feng, L; Wang, Y. (2013). Perfluorooctane sulfonate (PFOS) induced
embryotoxicity and disruption of cardiogenesis. Toxicol In Vitro 27: 1503-1512.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850971
Chou, HC; Wen, LL; Chang, CC; Lin, CY; Jin, L; Juan, SH. (2017). From the cover: 1-carnitine
via ppary- and sirtl-dependent mechanisms attenuates epithelial-mesenchymal transition
and renal fibrosis caused by perfluorooctanesulfonate. Toxicol Sci 160: 217-229.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4239391
Chou, WC; Lin, Z. (2019). Bayesian evaluation of a physiologically based pharmacokinetic
(PBPK) model for perfluorooctane sulfonate (PFOS) to characterize the interspecies
uncertainty between mice, rats, monkeys, and humans: Development and performance
verification. Environ Int 129: 408-422.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5412429
Chou, WC; Lin, Z. (2021). Development of a Gestational and Lactational Physiologically Based
Pharmacokinetic (PBPK) Model for Perfluorooctane Sulfonate (PFOS) in Rats and
Humans and Its Implications in the Derivation of Health-Based Toxicity Values. Environ
Health Perspect 129: 37004.
7-31
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7542658
Christensen, KY; Maisonet, M; Rubin, C; Holmes, A; Calafat, AM; Kato, K; Flanders, WD;
Heron, J; Mcgeehin, MA; Marcus, M. (2011). Exposure to polyfluoroalkyl chemicals
during pregnancy is not associated with offspring age at menarche in a contemporary
British cohort. Environ Int 37: 129-135.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290803
Christensen, KY; Raymond, M; Meiman, J. (2019). Perfluoroalkyl substances and metabolic
syndrome. Int J Hyg Environ Health 222: 147-153.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080398
Christensen, KY; Raymond, M; Thompson, BA; Anderson, HA. (2016). Perfluoroalkyl
substances in older male anglers in Wisconsin. Environ Int 91: 312-318.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858533
Christensen, KY; Raymond, MR; Thompson, BA; Anderson, HA. (2016). Fish consumption,
levels of nutrients and contaminants, and endocrine-related health outcomes among older
male anglers in Wisconsin. J Occup Environ Med 58: 668-675.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3350721
Christenson, B; Bottiger, M. (1986). Serological immunity to diphtheria in Sweden in 1978 and
1984. Scand J Infect Dis 18: 227-233.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9978484
Chu, C; Zhou, Y; Li, QQ; Bloom, MS; Lin, S; Yu, YJ; Chen, D; Yu, HY; Hu, LW; Yang, BY;
Zeng, XW; Dong, GH. (2020). Are perfluorooctane sulfonate alternatives safer? New
insights from a birth cohort study. Environ Int 135: 105365.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315711
Cichoz-Lach, H; Michalak, A. (2014). Oxidative stress as a crucial factor in liver diseases
[Review], World J Gastroenterol 20: 8082-8091.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2996796
Cifone, MA. (1999). Unscheduled DNA synthesis in rat liver primary cell cultures with PFOS.
Final report. (Covance Study No. 207840447). Vienna, VA: Covance Laboratories.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10228136
Cluett, R; Seshasayee, SM; Rokoff, LB; Rifas-Shiman, SL; Ye, X; Calafat, AM; Gold, DR;
Coull, B; Gordon, CM; Rosen, CJ; Oken, E; Sagiv, SK; Fleisch, AF. (2019). Per- and
Polyfluoroalkyl Substance Plasma Concentrations and Bone Mineral Density in
Midchildhood: A Cross-Sectional Study (Project Viva, United States). Environ Health
Perspect 127: 87006.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5412438
Cobb, BA. (2020). The history of IgG glycosylation and where we are now [Review],
Glycobiology 30: 202-213.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284268
Cohn, BA; La Merrill, MA; Krigbaum, NY; Wang, M; Park, JS; Petreas, M; Yeh, G; Hovey, RC;
Zimmermann, L; Cirillo, PM. (2020). In utero exposure to poly- and perfluoroalkyl
substances (PFASs) and subsequent breast cancer. Reprod Toxicol 92: 112-119.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5412451
Cong, J; Chu, C; Li, QQ; Zhou, Y; Qian, ZM; Geiger, SD; Vaughn, MG; Zeng, XW; Liu, RQ;
Hu, LW; Yang, BY; Chen, G; Zeeshan, M; Sun, X; Xiang, M; Dong, GH. (2021).
Associations of perfluorooctane sulfonate alternatives and serum lipids in Chinese adults.
Environ Int 155: 106596.
7-32
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/8442223
Conley, JM; Lambright, CS; Evans, N; Medlock-Kakaley, E; Hill, D; McCord, J; Strynar, MJ;
Wehmas, LC; Hester, S; MacMillan, DK; Gray, LE, Jr. (2022). Developmental toxicity
of Nafion byproduct 2 (NBP2) in the Sprague-Dawley rat with comparisons to
hexafluoropropylene oxide-dimer acid (HFPO-DA or GenX) and perfluorooctane
sulfonate (PFOS). Environ Int 160: 107056.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10176381
Convertino, M; Church, TR; Olsen, GW; Liu, Y; Doyle, E; Elcombe, CR; Barnett, AL; Samuel,
LM; Macpherson, IR; Evans, TRJ. (2018). Stochastic pharmacokinetic-
pharmacodynamic modeling for assessing the systemic health risk of perfluorooctanoate
(pfoa). Toxicol Sci 163: 293-306.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080342
Conway, B; Innes, KE; Long, D. (2016). Perfluoroalkyl substances and beta cell deficient
diabetes. J Diabetes Complications 30: 993-998.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859824
Conway, BN; Badders, AN; Costacou, T; Arthur, JM; Innes, KE. (2018). Perfluoroalkyl
substances and kidney function in chronic kidney disease, anemia, and diabetes.
Diabetes, Metabolic Syndrome and Obesity 11: 707-716.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080465
Corley, RA; Mendrala, AL; Smith, FA; Staats, DA; Gargas, ML; Conolly, RB; Andersen, ME;
Reitz, RH. (1990). Development of a physiologically based pharmacokinetic model for
chloroform. Toxicol Appl Pharmacol 103: 512-527.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10123
Cornelis, C; D'Hollander, W; Roosens, L; Covaci, A; Smolders, R; Van Den Heuvel, R; Govarts,
E; Van Campenhout, K; Reynders, H; Bervoets, L. (2012). First assessment of population
exposure to perfluorinated compounds in Flanders, Belgium. Chemosphere 86: 308-314.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2569108
Corsini, E; Avogadro, A; Galbiati, V; Dell'Agli, M; Marinovich, M; Galli, CL; Germolec, DR.
(2011). In vitro evaluation of the immunotoxic potential of perfluorinated compounds
(PFCs). Toxicol Appl Pharmacol 250: 108-116.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1937246
Corsini, E; Sangiovanni, E; Avogadro, A; Galbiati, V; Viviani, B; Marinovich, M; Galli, CL;
Dell'Agli, M; Germolec, DR. (2012). In vitro characterization of the immunotoxic
potential of several perfluorinated compounds (PFCs). Toxicol Appl Pharmacol 258:
248-255. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1937239
Corton, JC; Cunningham, ML; Hummer, BT; Lau, C; Meek, B; Peters, JM; Popp, JA;
Rhomberg, L; Seed, J; Klaunig, JE. (2014). Mode of action framework analysis for
receptor-mediated toxicity: The peroxisome proliferator-activated receptor alpha
(PPARa) as a case study [Review], CritRev Toxicol 44: 1-49.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2215399
Corton, JC; Peters, JM; Klaunig, JE. (2018). The PPARa-dependent rodent liver tumor response
is not relevant to humans: addressing misconceptions [Review], Arch Toxicol 92: 83-
119. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4862049
Costa, G; Sartori, S; Consonni, D. (2009). Thirty years of medical surveillance in
perfluooctanoic acid production workers. J Occup Environ Med 51: 364-372.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1429922
7-33
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Costello, E; Rock, S; Stratakis, N; Eckel, SP; Walker, DI; Valvi, D; Cserbik, D; Jenkins, T;
Xanthakos, SA; Kohli, R; Sisley, S; Vasiliou, V; La Merrill, MA; Rosen, H; Conti, DV;
Mcconnell, R; Chatzi, L. (2022). Exposure to per- and Polyfluoroalkyl Substances and
Markers of Liver Injury: A Systematic Review and Meta-Analysis [Review], Environ
Health Perspect 130: 46001.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10285082
Crawford, NM; Fenton, SE; Strynar, M; Hines, EP; Pritchard, DA; Steiner, AZ. (2017). Effects
of perfluorinated chemicals on thyroid function, markers of ovarian reserve, and natural
fertility. Reprod Toxicol 69: 53-59.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859813
Crissman, JW; Goodman, DG; Hildebrandt, PK; Maronpot, RR; Prater, DA; Riley, JH; Seaman,
WJ; Thake, DC. (2004). Best practices guideline: Toxicologic histopathology. Toxicol
Pathol 32: 126-131.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/51763
Crump, KS. (1995). Calculation of benchmark doses from continuous data. Risk Anal 15: 79-89.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2258
Cui, L; Liao, CY; Zhou, QF; Xia, TM; Yun, ZJ; Jiang, GB. (2010). Excretion of PFOA and
PFOS in male rats during a subchronic exposure. Arch Environ Contam Toxicol 58: 205-
213. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919335
Cui, L; Zhou, QF; Liao, CY; Fu, JJ; Jiang, GB. (2009). Studies on the toxicological effects of
PFOA and PFOS on rats using histological observation and chemical analysis. Arch
Environ Contam Toxicol 56: 338-349.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757868
Cui, Q; Pan, Y; Wang, J; Liu, H; Yao, B; Dai, J. (2020). Exposure to per- and polyfluoroalkyl
substances (PFASs) in serum versus semen and their association with male reproductive
hormones. Environ Pollut 266 Pt. 2: 115330.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833614
Cui, R; Zhang, H; Guo, X; Cui, Q; Wang, J; Dai, J. (2015). Proteomic analysis of cell
proliferation in a human hepatic cell line (HL-7702) induced by perfluorooctane
sulfonate using iTRAQ. J Hazard Mater 299: 361-370.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981568
Cui, Y; Liu, W; Xie, W; Yu, W; Wang, C; Chen, H. (2015). Investigation of the Effects of
Perfluorooctanoic Acid (PFOA) and Perfluorooctane Sulfonate (PFOS) on Apoptosis and
Cell Cycle in a Zebrafish (Danio rerio) Liver Cell Line. Int J Environ Res Public Health
12: 15673-15682.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981517
Cui, Y; Lv, S; Liu, J; Nie, S; Chen, J; Dong, Q; Huang, C; Yang, D. (2017). Chronic
perfluorooctanesulfonic acid exposure disrupts lipid metabolism in zebrafish. Hum Exp
Toxicol 36: 207-217.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981467
Curran, I; Hierlihy, SL; Liston, V; Pantazopoulos, P; Nunnikhoven, A; Tittlemier, S; Barker, M;
Trick, K; Bondy, G. (2008). Altered fatty acid homeostasis and related toxicologic
sequelae in rats exposed to dietary potassium perfluorooctanesulfonate (PFOS). J Toxicol
Environ Health A 71: 1526-1541.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757871
D'Agostino, RB; Vasan, RS; Pencina, MJ; Wolf, PA; Cobain, M; Massaro, JM; Kannel, WB.
7-34
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
(2008). General cardiovascular risk profile for use in primary care: the Framingham
Heart Study. Circulation 117: 743-753.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10694408
D'Alessandro, ML; Ellis, DA; Carter, JA; Stock, NL; March, RE. (2013). Competitive binding of
aqueous perfluoroctanesulfonic acid and ibuprofen with bovine serum albumin studied by
electrospray ionization mass spectrometry. Int J Mass Spectrom 345: 28-36.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5084740
Dalla Zuanna, T; Savitz, DA; Barbieri, G; Pitter, G; Zare Jeddi, M; Dapra, F; Fabricio, ASC;
Russo, F; Fletcher, T; Canova, C. (2021). The association between perfluoroalkyl
substances and lipid profile in exposed pregnant women in the Veneto region, Italy.
Ecotoxicol Environ Saf209: 111805.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7277682
Dalsager, L; Christensen, N; Halekoh, U; Timmermann, CAG; Nielsen, F; Kyhl, HB; Husby, S;
Grandjean, P; Jensen, TK; Andersen, HR. (2021). Exposure to perfluoroalkyl substances
during fetal life and hospitalization for infectious disease in childhood: A study among
1,503 children from the Odense Child Cohort. Environ Int 149: 106395.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7405343
Dalsager, L; Christensen, N; Husby, S; Kyhl, H; Nielsen, F; H0st, A; Grandjean, P; Jensen, TK.
(2016). Association between prenatal exposure to perfluorinated compounds and
symptoms of infections at age l-4years among 359 children in the Odense Child Cohort.
Environ Int 96: 58-64.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858505
Dang, Y; Wang, F; Liu, C. (2018). Real-time PCR array to study the effects of chemicals on the
growth hormone/insulin-like growth factors (GH/IGFs) axis of zebrafish embryos/larvae.
Chemosphere 207: 365-376.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4651759
Darrow, LA; Stein, CR; Steenland, K. (2013). Serum perfluorooctanoic acid and perfluorooctane
sulfonate concentrations in relation to birth outcomes in the Mid-Ohio Valley, 2005-
2010. Environ Health Perspect 121: 1207-1213.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850966
Das, KP; Wood, CR; Lin, MT; Starkov, AA; Lau, C; Wallace, KB; Corton, JC; Abbott, BD.
(2017). Perfluoroalkyl acids-induced liver steatosis: Effects on genes controlling lipid
homeostasis. Toxicology 378: 37-52.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859817
Dassuncao, C; Hu, XC; Nielsen, F; Weihe, P; Grandjean, P; Sunderland, EM. (2018). Shifting
Global Exposures to Poly- and Perfluoroalkyl Substances (PFASs) Evident in
Longitudinal Birth Cohorts from a Seafood-Consuming Population. Environ Sci Technol
52: 3738-3747.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4563862
Davies, B; Morris, T. (1993). Physiological parameters in laboratory animals and humans
[Review], PharmRes 10: 1093-1095.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/192570
De Felip, E; Abballe, A; Albano, FL; Battista, T; Carraro, V; Conversano, M; Franchini, S;
Giambanco, L; Iacovella, N; Ingelido, AM; Maiorana, A; Maneschi, F; Marra, V;
Mercurio, A; Nale, R; Nucci, B; Panella, V; Pirola, F; Porpora, MG; Procopio, E; Suma,
N; Valentini, S; Valsenti, L; Vecchie, V. (2015). Current exposure of Italian women of
7-35
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
reproductive age to PFOS and PFOA: A human biomonitoring study. Chemosphere 137:
1-8. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850114
de Vos, MG; Huijbregts, MA; van den Heuvel-Greve, MJ; Vethaak, AD; Van de Vijver, KI;
Leonards, PE; van Leeuwen, SP; de Voogt, P; Hendriks, AJ. (2008). Accumulation of
perfluorooctane sulfonate (PFOS) in the food chain of the Western Scheldt estuary:
Comparing field measurements with kinetic modeling. Chemosphere 70: 1766-1773.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919394
Dean, WP; Jessup, DC; Thompson, G; Romig, G; Powell, D. (1978). Fluorad Fluorochemical
Surfactant FC-95 Acute Oral Toxicity (LD50) Study in Rats. (Study No. 137-083).
International Research and Development Corporation.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9579905
Deji, Z; Liu, P; Wang, X; Zhang, X; Luo, Y; Huang, Z. (2021). Association between maternal
exposure to perfluoroalkyl and polyfluoroalkyl substances and risks of adverse pregnancy
outcomes: A systematic review and meta-analysis [Review], Sci Total Environ 783:
146984. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7564388
Delinsky, AD; Strynar, MJ; Mccann, PJ; Varns, JL; Mcmillan, L; Nakayama, SF; Lindstrom,
AB. (2010). Geographical distribution of perfluorinated compounds in fish from
Minnesota lakes and rivers. Environ Sci Technol 44: 2549-2554.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2587663
Deluca, NM; Angrish, M; Wilkins, A; Thayer, K; Cohen Hubal, EA. (2022). Human exposure
pathways to poly- and perfluoroalkyl substances (PFAS) from indoor media: A
systematic review protocol. Environ Int 146: 106308.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7277659
Dertinger, SD; Camphausen, K; Macgregor, JT; Bishop, ME; Torous, DK; Avlasevich, S;
Cairns, S; Tometsko, CR; Menard, C; Muanza, T; Chen, Y; Miller, RK; Cederbrant, K;
Sandelin, K; Ponten, I; Bolcsfoldi, G. (2004). Three-color labeling method for flow
cytometric measurement of cytogenetic damage in rodent and human blood. Environ Mol
Mutagen 44: 427-435.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10328871
Dewey, KG; Heinig, MJ; Nommsen, LA. (1993). Maternal weight-loss patterns during prolonged
lactation. Am J ClinNutr 58: 162-166.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1335605
Dewitt, J; Cox, K; Savitz, D. (2019). Health based drinking water value recommendations for
PFAS in Michigan. Lansing, MI: Michigan Science Advisory Work Group.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6982827
Dewitt, JC; Keil, DE. (2017). Current issues in developmental immunotoxicity. In G Parker
(Ed.), Immunopathology in Toxicology and Drug Development: Volume 1,
Immunobiology, Investigative Techniques, and Special Studies (pp. 601-618). Totowa,
NJ: Humana Press.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5926400
Dhingra, R; Winquist, A; Darrow, LA; Klein, M; Steenland, K. (2017). A study of reverse
causation: Examining the associations of perfluorooctanoic acid serum levels with two
outcomes. Environ Health Perspect 125: 416-421.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981432
Di Nisio, A; Sabovic, I; Valente, U; Tescari, S; Rocca, MS; Guidolin, D; Dall'Acqua, S;
Acquasaliente, L; Pozzi, N; Plebani, M; Garolla, A; Foresta, C. (2019). Endocrine
7-36
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
disruption of androgenic activity by perfluoroalkyl substances: clinical and experimental
evidence. J Clin Endocrinol Metab 104: 1259-1271.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080655
Ding, N; Harlow, SD; Randolph, JF; Calafat, AM; Mukheijee, B; Batterman, S; Gold, EB; Park,
SK. (2020). Associations of perfluoroalkyl substances with incident natural menopause:
The study of women's health across the nation. J Clin Endocrinol Metab 105: E3169-
E3182. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833612
Ding, N; Park, SK. (2020). Perfluoroalkyl substances exposure and hearing impairment in US
adults. Environ Res 187: 109686.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6711603
Dinglasan-Panlilio, MJ; Prakash, SS; Baker, JE. (2014). Perfluorinated compounds in the surface
waters of Puget Sound, Washington and Clayoquot and Barkley Sounds, British
Columbia. Mar Pollut Bull 78: 173-180.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2545254
Dixon, LJ; Barnes, M; Tang, H; Pritchard, MT; Nagy, LE. (2013). Kupffer cells in the liver
[Review], Compr Physiol 3: 785-797.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10365841
Domazet, SL; Grantved, A; Timmermann, AG; Nielsen, F; Jensen, TK. (2016). Longitudinal
associations of exposure to perfluoroalkylated substances in childhood and adolescence
and indicators of adiposity and glucose metabolism 6 and 12 years later: The European
Youth Heart Study. Diabetes Care 39: 1745-1751.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981435
Domazet, SL; Jensen, TK; Wedderkopp, N; Nielsen, F; Andersen, LB; Grantved, A. (2020).
Exposure to perfluoroalkylated substances (PFAS) in relation to fitness, physical activity,
and adipokine levels in childhood: The European youth heart study. Environ Res 191:
110110. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833700
Domingo, JL; Nadal, M. (2017). Per- and Polyfluoroalkyl Substances (PFASs) in Food and
Human Dietary Intake: A Review of the Recent Scientific Literature [Review], J Agric
Food Chem 65: 533-543.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981385
Donat-Vargas, C; Bergdahl, IA; Tornevi, A; Wennberg, M; Sommar, J; Kiviranta, H; Koponen,
J; Rolandsson, O; Akesson, A. (2019). Perfluoroalkyl substances and risk of type II
diabetes: A prospective nested case-control study. Environ Int 123: 390-398.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5083542
Donat-Vargas, C; Bergdahl, IA; Tornevi, A; Wennberg, M; Sommar, J; Koponen, J; Kiviranta,
H; Akesson, A. (2019). Associations between repeated measure of plasma perfluoroalkyl
substances and cardiometabolic risk factors. Environ Int 124: 58-65.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080588
Dong, GH; Liu, MM; Wang, D; Zheng, L; Liang, ZF; Jin, YH. (2011). Sub-chronic effect of
perfluorooctanesulfonate (PFOS) on the balance of type 1 and type 2 cytokine in adult
C57BL6 mice. Arch Toxicol 85: 1235-1244.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1424949
Dong, GH; Tung, KY; Tsai, CH; Liu, MM; Wang, D; Liu, W; Jin, YH; Hsieh, WS; Lee, YL;
Chen, PC. (2013). Serum polyfluoroalkyl concentrations, asthma outcomes, and
immunological markers in a case-control study of Taiwanese children. Environ Health
Perspect 121: 507-513, 513e501-508.
7-37
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1937230
Dong, GH; Zhang, YH; Zheng, L; Liu, W; Jin, YH; He, QC. (2009). Chronic effects of
perfluorooctanesulfonate exposure on immunotoxicity in adult male C57BL/6 mice. Arch
Toxicol 83: 805-815.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1424951
Dong, H; Curran, I; Williams, A; Bondy, G; Yauk, CL; Wade, MG. (2016). Hepatic miRNA
profiles and thyroid hormone homeostasis in rats exposed to dietary potassium
perfluorooctanesulfonate (PFOS). Environ Toxicol Pharmacol 41: 201-210.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981515
Dong, Z; Wang, H; Yu, YY; Li, YB; Naidu, R; Liu, Y. (2019). Using 2003-2014 U.S. NHANES
data to determine the associations between per- and polyfluoroalkyl substances and
cholesterol: Trend and implications. Ecotoxicol Environ Saf 173: 461-468.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080195
Donley, GM; Taylor, E; Jeddy, Z; Namulanda, G; Hartman, TJ. (2019). Association between in
utero perfluoroalkyl substance exposure and anti-Mullerian hormone levels in adolescent
females in a British cohort. Environ Res 177: 108585.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5381537
Dreyer, AF; Jensen, RC; Glintborg, D; Schmedes, AV; Brandslund, I; Nielsen, F; Kyhl, HB;
Jensen, TK; Andersen, MS. (2020). Perfluoroalkyl substance exposure early in pregnancy
was negatively associated with late pregnancy cortisone levels. J Clin Endocrinol Metab
105: E2834-E2844.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833676
Du, J; Wang, S; You, H; Jiang, R; Zhuang, C; Zhang, X. (2014). Developmental toxicity and
DNA damage to zebrafish induced by perfluorooctane sulfonate in the presence of ZnO
nanoparticles. Environ Toxicol 31: 360-371.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851143
Duan, Y; Sun, H; Yao, Y; Meng, Y; Li, Y. (2020). Distribution of novel and legacy per-
/polyfluoroalkyl substances in serum and its associations with two glycemic biomarkers
among Chinese adult men and women with normal blood glucose levels. Environ Int 134:
105295. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5918597
Ducatman, A; Zhang, J; Fan, H. (2015). Prostate-specific antigen and perfluoroalkyl acids in the
C8 health study population. J Occup Environ Med 57: 111-114.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859843
Dufour, P; Pirard, C; Seghaye, MC; Charlier, C. (2018). Association between organohalogenated
pollutants in cord blood and thyroid function in newborns and mothers from Belgian
population. Environ Pollut 238: 389-396.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4354164
Dzierlenga, M, .W.; Crawford, L, .; Longnecker, M, .P. (2020). Birth weight and perfluorooctane
sulfonic acid: a random-effects meta-regression analysis. Environmental Epidemiology 4:
e095. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7643488
Dzierlenga, MW; Allen, BC; Clewell, HJ; Longnecker, MP. (2020). Pharmacokinetic bias
analysis of an association between clinical thyroid disease and two perfluoroalkyl
substances. Environ Int 141: 105784.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833691
Dzierlenga, MW; Allen, BC; Ward, PL; Clewell, HJ; Longnecker, MP. (2019). A model of
functional thyroid disease status over the lifetime. PLoS ONE 14: e0219769.
7-38
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7947729
Dzierlenga, MW; Moreau, M; Song, G; Mallick, P; Ward, PL; Campbell, JL; Housand, C; Yoon,
M; Allen, BC; Clewell, HJ; Longnecker, MP. (2020). Quantitative bias analysis of the
association between subclinical thyroid disease and two perfluoroalkyl substances in a
single study. Environ Res 182: 109017.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315786
East, A; Egeghy, PP; Hubal, E; Slover, R; Vallero, DA. (2021). Computational estimates of daily
aggregate exposure to PFOA/PFOS from 2011 to 2017 using a basic intake model. J
Expo Sci Environ Epidemiol Online ahead of print.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9416543
Ebert, A; Allendorf, F; Berger, U; Goss, KU; Ulrich, N. (2020). Membrane/water partitioning
and permeabilities of perfluoroalkyl acids and four of their alternatives and the effects on
toxicokinetic behavior. Environ Sci Technol 54: 5051-5061.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6505873
ECCC. (2018). Canadian Environmental Protection Act, 1999. Federal Environmental Quality
Guidelines. Perfluorooctane Sulfonate (PFOS).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9638786
EFSA. (2020). Risk to human health related to the presence ofperfluoroalkyl substances in food.
EFSA J 18. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6984182
Egeghy, PP; Lorber, M. (2011). An assessment of the exposure of Americans to perfluorooctane
sulfonate: A comparison of estimated intake with values inferred from NHANES data. J
Expo Sci Environ Epidemiol 21: 150-168.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/723765
Ehresman, DJ; Froehlich, JW; Olsen, GW; Chang, SC; Butenhoff, JL. (2007). Comparison of
human whole blood, plasma, and serum matrices for the determination of
perfluorooctanesulfonate (PFOS), perfluorooctanoate (PFOA), and other
fluorochemicals. Environ Res 103: 176-184.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1429928
Eke, D; Qelik, A. (2016). Curcumin prevents perfluorooctane sulfonate-induced genotoxicity and
oxidative DNA damage in rat peripheral blood. Drug Chem Toxicol 39: 97-103.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850124
Eke, D; Qelik, A; Yilmaz, MB; Aras, N; Kocatiirk Sel, S; Alptekin, D. (2017). Apoptotic gene
expression profiles and DNA damage levels in rat liver treated with perfluorooctane
sulfonate and protective role of curcumin. Int J Biol Macromol 104: 515-520.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981318
Elcombe, CR; Elcombe, BM; Foster, JR; Chang, SC; Ehresman, DJ; Butenhoff, JL. (2012).
Hepatocellular hypertrophy and cell proliferation in Sprague-Dawley rats from dietary
exposure to potassium perfluorooctanesulfonate results from increased expression of
xenosensor nuclear receptors PPARa and CAR/PXR. Toxicology 293: 16-29.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1401466
Elcombe, CR; Elcombe, BM; Foster, JR; Chang, SC; Ehresman, DJ; Noker, PE; Butenhoff, JL.
(2012). Evaluation of hepatic and thyroid responses in male Sprague Dawley rats for up
to eighty-four days following seven days of dietary exposure to potassium
perfluorooctanesulfonate. Toxicology 293: 30-40.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332473
Elcombe, CR; Peffer, RC; Wolf, DC; Bailey, J; Bars, R; Bell, D; Cattley, RC; Ferguson, SS;
7-39
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Geter, D; Goetz, A; Goodman, JI; Hester, S; Jacobs, A; Omiecinski, CJ; Schoeny, R; Xie,
W; Lake, BG. (2014). Mode of action and human relevance analysis for nuclear receptor-
mediated liver toxicity: A case study with phenobarbital as a model constitutive
androstane receptor (CAR) activator [Review], Crit Rev Toxicol 44: 64-82.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2343661
EMEA. (2008). Non-clinical guideline on drug induced hepatotoxicity. (Doc. Ref.
EMEA/CHMP/SWP/150115/2006). London, UK.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3056793
EMEA. (2010). Reflection paper on non-clinical evaluation of drug-induced liver injury (DILI).
(EMEA/CHMP/SWP/150115/2006). London, UK.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3056796
Ericson, I; Gomez, M; Nadal, M; van Bavel, B; Lindstrom, G; Domingo, JL. (2007).
Perfluorinated chemicals in blood of residents in Catalonia (Spain) in relation to age and
gender: a pilot study. Environ Int 33: 616-623.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858652
Eriksen, KT; Raaschou-Nielsen, O; Mclaughlin, JK; Lipworth, L; Tj0nneland, A; Overvad, K;
S0rensen, M. (2013). Association between plasma PFOA and PFOS levels and total
cholesterol in a middle-aged Danish population. PLoS ONE 8: e56969.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919150
Eriksen, KT; S0rensen, M; Mclaughlin, JK; Lipworth, L; Tj0nneland, A; Overvad, K; Raaschou-
Nielsen, O. (2009). Perfluorooctanoate and perfluorooctanesulfonate plasma levels and
risk of cancer in the general Danish population. J Natl Cancer Inst 101: 605-609.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919344
Ernst, A; Brix, N; Lauridsen, LLB; Olsen, J; Parner, ET; Liew, Z; Olsen, LH; Ramlau-Hansen,
CH. (2019). Exposure to perfluoroalkyl substances during fetal life and pubertal
development in boys and girls from the danish national birth cohort. Environ Health
Perspect 127: 17004.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080529
Erol, E; Kumar, LS; Cline, GW; Shulman, GI; Kelly, DP; Binas, B. (2004). Liver fatty acid
binding protein is required for high rates of hepatic fatty acid oxidation but not for the
action of PPARalpha in fasting mice. FASEB J 18: 347-349.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5212239
Eryasa, B; Grandjean, P; Nielsen, F; Valvi, D; Zmirou-Navier, D; Sunderland, E; Weihe, P;
Oulhote, Y. (2019). Physico-chemical properties and gestational diabetes predict
transplacental transfer and partitioning of perfluoroalkyl substances. Environ Int 130:
104874. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5412430
Etzel, TM; Braun, JM; Buckley, JP. (2019). Associations of serum perfluoroalkyl substance and
vitamin D biomarker concentrations in NHANES, 2003-2010. Int J Hyg Environ Health
222: 262-269.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5043582
Fabrega, F; Kumar, V; Benfenati, E; Schuhmacher, M; Domingo, JL; Nadal, M. (2015).
Physiologically based pharmacokinetic modeling of perfluoroalkyl substances in the
human body. Toxicol Environ Chem 97: 814-827.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3223669
Fabrega, F; Kumar, V; Schuhmacher, M; Domingo, JL; Nadal, M. (2014). PBPK modeling for
PFOS and PFOA: validation with human experimental data. Toxicol Lett 230: 244-251.
7-40
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850904
Fai Tse, WK; Li, JW; Kwan Tse, AC; Chan, TF; Hin Ho, JC; Sun Wu, RS; Chu Wong, CK; Lai,
KP. (2016). Fatty liver disease induced by perfluorooctane sulfonate: Novel insight from
transcriptome analysis. Chemosphere 159: 166-177.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981456
Fan, H; Ducatman, A; Zhang, J. (2014). Perfluorocarbons and Gilbert syndrome (phenotype) in
the C8 Health Study Population. Environ Res 135: 70-75.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2967086
Fan, Y; Lu, C; Li, X; Xu, Q; Zhang, Y; Yang, X; Han, X; Du, G; Xia, Y; Wang, X. (2020).
Serum albumin mediates the effect of multiple per- and polyfluoroalkyl substances on
serum lipid levels. Environ Pollut 266 Pt 2: 115138.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7102734
Fang, S; Chen, X; Zhao, S; Zhang, Y; Jiang, W; Yang, L; Zhu, L. (2014). Trophic magnification
and isomer fractionation of perfluoroalkyl substances in the food web of Taihu Lake,
China. Environ Sci Technol 48: 2173-2182.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850900
Fassler, CS; Pinney, SE; Xie, C; Biro, FM; Pinney, SM. (2019). Complex relationships between
perfluorooctanoate, body mass index, insulin resistance and serum lipids in young girls.
Environ Res 176: 108558.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315820
FDA. (2002). Guidance for industry: immunotoxicology evaluation of investigational new drugs.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/88170
FDA. (2009). Drug-induced liver injury: Premarketing clinical evaluation. (Docket no. FDA-
2008-D-0128). Rockville, MD: U.S. Department of Health and Human Services, Food
and Drug Administration.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6987952
FDA. (2016). Analytical Results for PFAS in 2016 Carbonated Water and Non-Carbonated
Bottled Water Sampling (parts per trillion). Retrieved from
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9419013
FDA. (2018). Analytical Results for PFAS in 2018 Produce Sampling (Parts Per Trillion).
Retrieved from
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9419064
FDA. (2019). Analytical Results for PFAS in 2019 Total Diet Study Sampling (Parts Per
Trillion)—Dataset 1: U. S. Food and Drug Administration (FDA). Retrieved from
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9638790
FDA. (2019). Analytical Results for PFAS in 2019 Total Diet Study Sampling (Parts Per
Trillion)—Dataset 2: U.S. Food and Drug Administration (FDA). Retrieved from
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9638792
FDA. (2020). Authorized Uses of PFAS in Food contact Applications.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9419078
FDA. (2020). FDA Announces Voluntary Agreement with Manufacturers to Phase Out Certain
Short-Chain PFAS Used in Food Packaging.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9419079
FDA. (2021). Analytical Results of Testing Food for PFAS from Environmental Contamination.
Retrieved from
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9419076
7-41
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Fei, C; Mclaughlin, JK; Lipworth, L; Olsen, J. (2008). Prenatal exposure to perfluorooctanoate
(PFOA) and perfluorooctanesulfonate (PFOS) and maternally reported developmental
milestones in infancy. Environ Health Perspect 116: 1391-1395.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290822
Fei, C; Mclaughlin, JK; Lipworth, L; Olsen, J. (2010). Prenatal exposure to PFOA and PFOS and
risk of hospitalization for infectious diseases in early childhood. Environ Res 110: 773-
777. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290805
Fei, C; Mclaughlin, JK; Tarone, RE; Olsen, J. (2007). Perfluorinated chemicals and fetal growth:
A study within the Danish National Birth Cohort. Environ Health Perspect 115: 1677-
1682. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1005775
Fei, C; Olsen, J. (2011). Prenatal exposure to perfluorinated chemicals and behavioral or
coordination problems at age 7 years. Environ Health Perspect 119: 573-578.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/758428
Fei, CY; Mclaughlin, JK; Lipworth, L; Olsen, J. (2009). Maternal levels of perfluorinated
chemicals and subfecundity. Hum Reprod 24: 1200-1205.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291107
Felter, SP; Foreman, JE; Boobis, A; Corton, JC; Doi, AM; Flowers, L; Goodman, J; Haber, LT;
Jacobs, A; Klaunig, JE; Lynch, AM; Moggs, J; Pandiri, A. (2018). Human relevance of
rodent liver tumors: key insights from a toxicology forum workshop on nongenotoxic
modes of action. Regul Toxicol Pharmacol 92: 1-7.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9642149
Fenton, SE; Ducatman, A; Boobis, A; DeWitt, JC; Lau, C; Ng, C; Smith, JS; Roberts, SM.
(2021). Per- and polyfluoroalkyl substance toxicity and human health review: Current
state of knowledge and strategies for informing future research [Review], Environ
Toxicol Chem 40: 606-630.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6988520
Fisher, M; Arbuckle, TE; Wade, M; Haines, DA. (2013). Do perfluoroalkyl substances affect
metabolic function and plasma lipids?—Analysis of the 2007-2009, Canadian Health
Measures Survey (CHMS) Cycle 1. Environ Res 121: 95-103.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919156
Fleisch, AF; Rifas-Shiman, SL; Mora, AM; Calafat, AM; Ye, X; Luttmann-Gibson, H; Gillman,
MW; Oken, E; Sagiv, SK. (2017). Early-life exposure to perfluoroalkyl substances and
childhood metabolic function. Environ Health Perspect 125: 481-487.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858513
Fletcher, T; Galloway, TS; Melzer, D; Holcroft, P; Cipelli, R; Pilling, LC; Mondal, D; Luster,
M; Harries, LW. (2013). Associations between PFOA, PFOS and changes in the
expression of genes involved in cholesterol metabolism in humans. Environ Int 57-58: 2-
10. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850968
Florentin, A; Deblonde, T; Diguio, N; Hautemaniere, A; Hartemann, P. (2011). Impacts of two
perfluorinated compounds (PFOS and PFOA) on human hepatoma cells: cytotoxicity but
no genotoxicity? Int J Hyg Environ Health 214: 493-499.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919235
Forns, J; Iszatt, N; White, RA; Mandal, S; Sabaredzovic, A; Lamoree, M; Thomsen, C; Haug,
LS; Stigum, H; Eggesb0, M. (2015). Perfluoroalkyl substances measured in breast milk
and child neuropsychological development in a Norwegian birth cohort study. Environ
Int83:176-182.
7-42
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3228833
Forsthuber, M; Kaiser, AM; Granitzer, S; Hassl, I; Hengstschlager, M; Stangl, H; Gundacker, C.
(2020). Albumin is the major carrier protein for PFOS, PFOA, PFHxS, PFNA and PFDA
in human plasma. Environ Int 137: 105324.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6311640
Fragki, S; Dirven, H; Fletcher, T; Grasl-Kraupp, B; Bjerve Giitzkow, K; Hoogenboom, R;
Kersten, S; Lindeman, B; Louisse, J; Peijnenburg, A; Piersma, AH; Princen, HMG; Uhl,
M; Westerhout, J; Zeilmaker, MJ; Luijten, M. (2021). Systemic PFOS and PFOA
exposure and disturbed lipid homeostasis in humans: what do we know and what not?
Crit Rev Toxicoll41-164.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/8442211
Franco, ME; Fernandez-Luna, MT; Ramirez, AJ; Lavado, R. (2020). Metabolomic-based
assessment reveals dysregulation of lipid profiles in human liver cells exposed to
environmental obesogens. Toxicol Appl Pharmacol 398: 115009.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6507465
Franco, ME; Sutherland, GE; Fernandez-Luna, MT; Lavado, R. (2020). Altered expression and
activity of phase I and II biotransformation enzymes in human liver cells by
perfluorooctanoate (PFOA) and perfluorooctane sulfonate (PFOS). Toxicology 430:
152339. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315712
Fraser, AJ; Webster, TF; Watkins, DJ; Strynar, MJ; Kato, K; Calafat, AM; Vieira, VM; Mcclean,
MD. (2013). Polyfluorinated compounds in dust from homes, offices, and vehicles as
predictors of concentrations in office workers' serum. Environ Int 60: 128-136.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2325338
Frescas, D; Valenti, L; Accili, D. (2005). Nuclear trapping of the forkhead transcription factor
FoxOl via Sirt-dependent deacetylation promotes expression of glucogenetic genes. J
Biol Chem 280: 20589-20595.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284417
Frisbee, SJ; Shankar, A; Knox, SS; Steenland, K; Savitz, DA; Fletcher, T; Ducatman, AM.
(2010). Perfluorooctanoic acid, perfluorooctanesulfonate, and serum lipids in children
and adolescents: results from the C8 Health Project. Arch Pediatr Adolesc Med 164: 860-
869. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1430763
Fromme, H; Mosch, C; Morovitz, M; Alba-Alejandre, I; Boehmer, S; Kiranoglu, M; Faber, F;
Hannibal, I; Genzel-Boroviczeny, O; Koletzko, B; Volkel, W. (2010). Pre- and postnatal
exposure to perfluorinated compounds (PFCs). Environ Sci Technol 44: 7123-7129.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290877
Fromme, H; Tittlemier, SA; Volkel, W; Wilhelm, M; Twardella, D. (2009). Perfluorinated
compounds - Exposure assessment for the general population in western countries
[Review], Int J Hyg Environ Health 212: 239-270.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291085
Fry, K; Power, MC. (2017). Persistent organic pollutants and mortality in the United States,
NHANES 1999-2011. Environ Health 16: 105.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4181820
Fu, J; Gao, Y; Cui, L; Wang, T; Liang, Y; Qu, G; Yuan, B; Wang, Y; Zhang, A; Jiang, G.
(2016). Occurrence, temporal trends, and half-lives of perfluoroalkyl acids (PFAAs) in
occupational workers in China. Sci Rep 6: 38039.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859819
7-43
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Fu, SH; Chien, MW; Hsu, CY; Liu, YW; Sytwu, HK. (2020). Interplay between Cytokine
Circuitry and Transcriptional Regulation Shaping Helper T Cell Pathogenicity and
Plasticity in Inflammatory Bowel Disease [Review], International Journal of Molecular
Sciences 21. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284269
Fuentes, S; Colomina, MT; Rodriguez, J; Vicens, P; Domingo, JL. (2006). Interactions in
developmental toxicology: concurrent exposure to perfluorooctane sulfonate (PFOS) and
stress in pregnant mice. Toxicol Lett 164: 81-89.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757859
Fuentes, S; Colomina, MT; Vicens, P; Franco-Pons, N; Domingo, JL. (2007). Concurrent
exposure to perfluorooctane sulfonate and restraint stress during pregnancy in mice:
effects on postnatal development and behavior of the offspring. Toxicol Sci 98: 589-598.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757863
Fuentes, S; Vicens, P; Colomina, MT; Domingo, JL. (2007). Behavioral effects in adult mice
exposed to perfluorooctane sulfonate (PFOS). Toxicology 242: 123-129.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757865
Gabrielsson, J; Weiner, D. (2000). Pharmacokinetic and pharmacodynamic data analysis:
concepts and applications (3rd ed.). Stockholm: Swedish Pharmaceutical Press.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9642135
Galazka, A; Kardymowicz, B. (1989). Immunity against diphtheria in adults in Poland.
Epidemiol Infect 103: 587-593.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9642152
Galazka, AM; Milstien, JB; Robertson, SE; Cutts, FT. (1993). The immunological basis for
immunization module 2 : Diphtheria. (WHO/EPI/Gen/93.11-18). Galazka, AM; Milstien,
JB; Robertson, SE; Cutts, FT.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10228565
Gallo, V; Leonardi, G; Brayne, C; Armstrong, B, en; Fletcher, T. (2013). Serum perfluoroalkyl
acids concentrations and memory impairment in a large cross-sectional study. BMJ Open
3. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2272847
Gallo, V; Leonardi, G; Genser, B; Lopez-Espinosa, MJ; Frisbee, SJ; Karlsson, L; Ducatman,
AM; Fletcher, T. (2012). Serum perfluorooctanoate (PFOA) and perfluorooctane
sulfonate (PFOS) concentrations and liver function biomarkers in a population with
elevated PFOA exposure. Environ Health Perspect 120: 655-660.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1276142
Gao, B; He, X; Liu, W; Zhang, H; Saito, N; Tsuda, S. (2015). Distribution of perfluoroalkyl
compounds in rats: Indication for using hair as bioindicator of exposure. J Expo Sci
Environ Epidemiol 25: 632-638.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851191
Gao, K, e; Zhuang, T; Liu, X; Fu, J; Zhang, J; Fu, J, ie; Wang, L; Zhang, A; Liang, Y; Song, M;
Jiang, G. (2019). Prenatal Exposure to Per- and Polyfluoroalkyl Substances (PFASs) and
Association between the Placental Transfer Efficiencies and Dissociation Constant of
Serum Proteins-PFAS Complexes. Environ Sci Technol 53: 6529-6538.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5387135
Gao, X; Ni, W; Zhu, S; Wu, Y; Cui, Y; Ma, J; Liu, Y; Qiao, J; Ye, Y; Yang, P; Liu, C; Zeng, F.
(2021). Per- and polyfluoroalkyl substances exposure during pregnancy and adverse
pregnancy and birth outcomes: A systematic review and meta-analysis. Environ Res 201:
111632. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9959601
7-44
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Gao, Y; Fu, J; Cao, H; Wang, Y; Zhang, A; Liang, Y; Wang, T; Zhao, C; Jiang, G. (2015).
Differential Accumulation and Elimination Behavior of Perfluoroalkyl Acid Isomers in
Occupational Workers in a Manufactory in China. Environ Sci Technol 49: 6953-6962.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850134
Gaylord, A; Berger, KI; Naidu, M; Attina, TM; Gilbert, J; Koshy, TT; Han, X; Marmor, M;
Shao, Y; Giusti, R; Goldring, RM; Kannan, K; Trasande, L. (2019). Serum perfluoroalkyl
substances and lung function in adolescents exposed to the World Trade Center disaster.
Environ Res 172: 266-272.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080201
Gaylord, A; Trasande, L; Kannan, K; Thomas, KM; Lee, S; Liu, M; Levine, J. (2020). Persistent
organic pollutant exposure and celiac disease: A pilot study. Environ Res 186: 109439.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833754
Gebbink, WA; Berger, U; Cousins, IT. (2015). Estimating human exposure to PFOS isomers and
PFCA homologues: the relative importance of direct and indirect (precursor) exposure.
Environ Int 74: 160-169.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850068
Geiger, SD; Xiao, J; Ducatman, A; Frisbee, S; Innes, K; Shankar, A. (2014). The association
between PFOA, PFOS and serum lipid levels in adolescents. Chemosphere 98: 78-83.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850925
Geiger, SD; Xiao, J; Shankar, A. (2013). Positive association between perfluoroalkyl chemicals
and hyperuricemia in children. Am J Epidemiol 177: 1255-1262.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919148
Geiger, SD; Xiao, J; Shankar, A. (2014). No association between perfluoroalkyl chemicals and
hypertension in children. Integrated Blood Pressure Control 7: 1-7.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851286
Genuis, SJ; Birkholz, D; Ralitsch, M; Thibault, N. (2010). Human detoxification of
perfluorinated compounds. Public Health 124: 367-375.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2583643
Genuis, SJ; Liu, Y; Genuis, QI; Martin, JW. (2014). Phlebotomy treatment for elimination of
perfluoroalkyl acids in a highly exposed family: a retrospective case-series. PLoS ONE 9:
el 14295. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851045
Getz, GS; Reardon, CA. (2012). Animal models of atherosclerosis [Review], Arterioscler
Thromb Vase Biol 32: 1104-1115.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1065480
Ghassabian, A; Bell, EM; Ma, WL; Sundaram, R; Kannan, K; Buck Louis, GM; Yeung, E.
(2018). Concentrations of perfluoroalkyl substances and bisphenol A in newborn dried
blood spots and the association with child behavior. Environ Pollut 243: 1629-1636.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080189
Ghisari, M; Long, M; R0ge, DM; Olsen, J; Bonefeld-J0rgensen, EC. (2017). Polymorphism in
xenobiotic and estrogen metabolizing genes, exposure to perfluorinated compounds and
subsequent breast cancer risk: A nested case-control study in the Danish National Birth
Cohort. Environ Res 154: 325-333.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860243
Giesy, JP; Kannan, K. (2001). Global distribution of perfluorooctane sulfonate in wildlife.
Environ Sci Technol 35: 1339-1342.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290854
7-45
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Gimenez-Bastida, JA; Surma, M; Zielinski, H. (2015). In vitro evaluation of the cytotoxicity and
modulation of mechanisms associated with inflammation induced by
perfluorooctanesulfonate and perfluorooctanoic acid in human colon myofibroblasts
CCD-I8C0. Toxicol In Vitro 29: 1683-1691.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981569
Girardi, P; Merler, E. (2019). A mortality study on male subjects exposed to polyfluoroalkyl
acids with high internal dose of perfluorooctanoic acid. Environ Res 179: 108743.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315730
Glassmeyer, ST; Furlong, ET; Kolpin, DW; Batt, AL; Benson, R; Boone, JS; Conerly, O;
Donohue, MJ; King, DN; Kostich, MS; Mash, HE; Pfaller, SL; Schenck, KM; Simmons,
JE; Varughese, EA; Vesper, SJ; Villegas, EN; Wilson, VS. (2017). Nationwide
reconnaissance of contaminants of emerging concern in source and treated drinking
waters of the United States. Sci Total Environ 581-582: 909-922.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3454569
Gleason, JA; Post, GB; Fagliano, JA. (2015). Associations of perfluorinated chemical serum
concentrations and biomarkers of liver function and uric acid in the US population
(NHANES), 2007-2010. Environ Res 136: 8-14.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2966740
Glynn, A; Berger, U; Bignert, A; Ullah, S; Aune, M; Lignell, S; Darnerud, PO. (2012).
Perfluorinated alkyl acids in blood serum from primiparous women in Sweden: serial
sampling during pregnancy and nursing, and temporal trends 1996-2010. Environ Sci
Technol 46: 9071-9079.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1578498
Goeden, HM; Greene, CW; Jacobus, JA. (2019). A transgenerational toxicokinetic model and its
use in derivation of Minnesota PFOA water guidance. J Expo Sci Environ Epidemiol 29:
183-195. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080506
Goff, DC; Lloyd-Jones, DM; Bennett, G; Coady, S; D'Agostino, RB; Gibbons, R; Greenland, P;
Lackland, DT; Levy, D; O'Donnell, CJ; Robinson, JG; Schwartz, JS; Shero, ST; Smith,
SC; Sorlie, P; Stone, NJ; Wilson, PW. (2014). 2013 ACC/AHA guideline on the
assessment of cardiovascular risk: a report of the American College of
Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation
129: S49-S73.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3121148
Gogola, J; Hoffmann, M; Nimpsz, S; Ptak, A. (2020). Disruption of 17P-estradiol secretion by
persistent organic pollutants present in human follicular fluid is dependent on the
potential of ovarian granulosa tumor cell lines to metabolize estrogen. Mol Cell
Endocrinol 503: 110698.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6316203
Gogola, J; Hoffmann, M; Ptak, A. (2019). Persistent endocrine-disrupting chemicals found in
human follicular fluid stimulate the proliferation of granulosa tumor spheroids via
GPR30 and IGF1R but not via the classic estrogen receptors. Chemosphere 217: 100-110.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5016947
Gogola, J; Hoffmann, M; Ptak, A. (2020). Persistent endocrine-disrupting chemicals found in
human follicular fluid stimulate IGF1 secretion by adult ovarian granulosa cell tumor
spheroids and thereby increase proliferation of non-cancer ovarian granulosa cells.
Toxicol In Vitro 65: 104769.
7-46
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6316206
Goldenthal, E; Jessup, DC; Geil, RG; Mehring, JS. (1978). Ninety-day subacute rhesus monkey
toxicity study: Fluorad " Fluorochemical FC-143. (Study No. 137-090). St. Paul, MN:
Report prepared for 3M by Institutional Research and Devlopment Corporation
(Mattawan, MN).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291068
Goldenthal, EI; Jessup, DC; Geil, RB; Mehring, JS. (1979). Ninety-Day Subacute Rhesus
Monkey Toxicity Study. (Study No. 137-087). Mattawan, MI: International Research and
Development Corporation.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9573133
Gomis, MI; Vestergren, R; Macleod, M; Mueller, JF; Cousins, IT. (2017). Historical human
exposure to perfluoroalkyl acids in the United States and Australia reconstructed from
biomonitoring data using population-based pharmacokinetic modelling. Environ Int 108:
92-102. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981280
Gorrochategui, E; Perez-Albaladejo, E; Casas, J; Lacorte, S; Porte, C. (2014). Perfluorinated
chemicals: Differential toxicity, inhibition of aromatase activity and alteration of cellular
lipids in human placental cells. Toxicol Appl Pharmacol 277: 124-130.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2324895
Goudarzi, H; Araki, A; Itoh, S; Sasaki, S; Miyashita, C; Mitsui, T; Nakazawa, H; Nonomura, K;
Kishi, R. (2017). The association of prenatal exposure to perfluorinated chemicals with
glucocorticoid and androgenic hormones in cord blood samples: The Hokkaido study.
Environ Health Perspect 125: 111-118.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981462
Goudarzi, H; Miyashita, C; Okada, E; Kashino, I; Chen, CJ; Ito, S; Araki, A; Kobayashi, S;
Matsuura, H; Kishi, R. (2017). Prenatal exposure to perfluoroalkyl acids and prevalence
of infectious diseases up to 4 years of age. Environ Int 104: 132-138.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859808
Goudarzi, H; Miyashita, C; Okada, E; Kashino, I; Kobayashi, S; Chen, CJ; Ito, S; Araki, A;
Matsuura, H; Ito, YM; Kishi, R. (2016). Effects of prenatal exposure to perfluoroalkyl
acids on prevalence of allergic diseases among 4-year-old children. Environ Int 94: 124-
132. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859523
Goudarzi, H; Nakajima, S; Ikeno, T; Sasaki, S; Kobayashi, S; Miyashita, C; Ito, S; Araki, A;
Nakazawa, H; Kishi, R. (2016). Prenatal exposure to perfluorinated chemicals and
neurodevelopment in early infancy: The Hokkaido Study. Sci Total Environ 541: 1002-
1010. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981536
Graber, JM; Alexander, C; Laumbach, RJ; Black, K; Strickland, PO; Georgopoulos, PG;
Marshall, EG; Shendell, DG; Alderson, D; Mi, Z; Mascari, M; Weisel, CP. (2019). Per
and polyfluoroalkyl substances (PFAS) blood levels after contamination of a community
water supply and comparison with 2013-2014 NHANES. J Expo Sci Environ Epidemiol
29: 172-182. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080653
Grandjean, P; Andersen, EW; Budtz-j0rgensen, E; Nielsen, F; M0lbak, K; Weihe, P; Heilmann,
C. (2012). Serum vaccine antibody concentrations in children exposed to perfluorinated
compounds. JAMA 307: 391-397.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1248827
Grandjean, P; Bateson, T. (2021). RE: Benchmark analysis for PFAS immunotoxicity. Available
online 9959716
7-47
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Grandjean, P; Heilmann, C; Weihe, P; Nielsen, F; Mogensen, UB; Budtz-j0rgensen, E. (2017).
Serum Vaccine Antibody Concentrations in Adolescents Exposed to Perfluorinated
Compounds. Environ Health Perspect 125: 077018.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858518
Grandjean, P; Heilmann, C; Weihe, P; Nielsen, F; Mogensen, UB; Timmermann, A; Budtz-
J0rgensen, E. (2017). Estimated exposures to perfluorinated compounds in infancy
predict attenuated vaccine antibody concentrations at age 5-years. J Immunotoxicol 14:
188-195. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4239492
Grandjean, P; Timmermann, CAG; Kruse, M; Nielsen, F; Vinholt, PJ; Boding, L; Heilmann, C;
M0lbak, K. (2020). Severity of COVID-19 at elevated exposure to perfluorinated
alkylates. PLoS ONE 15: e0244815.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7403067
Granum, B; Haug, LS; Namork, E; St0levik, SB; Thomsen, C; Aaberge, IS; van Loveren, H;
L0vik, M; Nygaard, UC. (2013). Pre-natal exposure to perfluoroalkyl substances may be
associated with altered vaccine antibody levels and immune-related health outcomes in
early childhood. J Immunotoxicol 10: 373-379.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1937228
Grasty, RC; Bjork, JA; Wallace, KB; Lau, CS; Rogers, JM. (2005). Effects of prenatal
perfluorooctane sulfonate (PFOS) exposure on lung maturation in the perinatal rat. Birth
Defects Res B Dev Reprod Toxicol 74: 405-416.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2951495
Grasty, RC; Wolf, DC; Grey, BE; Lau, CS; Rogers, JM. (2003). Prenatal window of
susceptibility to perfluorooctane sulfonate-induced neonatal mortality in the Sprague-
Dawley rat. Birth Defects Res B Dev Reprod Toxicol 68: 465-471.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332670
Grasty, RC; Wolf, DC; Grey, BE; Lau, CS; Rogers, JM. (2006). Prenatal window of
susceptibility to perfluorooctane sulfonate-induced neonatal mortality in the Sprague-
Dawley rat. Birth defects research (Part B) (vol 68, pg 465, 2003) [Erratum], Birth
Defects Res B Dev Reprod Toxicol 77: 86.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5085464
Gray, LE; Ostby, J; Sigmon, R; Ferrell, J; Rehnberg, G; Linder, R; Cooper, R; Goldman, J;
Laskey, J. (1988). The development of a protocol to assess reproductive effects of
toxicants in the rat [Review], Reprod Toxicol 2: 281-287.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332904
Greenland, S; Longnecker, MP. (1992). Methods for trend estimation from summarized dose-
response data, with applications to meta-analysis. Am J Epidemiol 135: 1301-1309.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5069
Gremmel, C; Fromel, T; Knepper, TP. (2016). Systematic determination of perfluoroalkyl and
polyfluoroalkyl substances (PFASs) in outdoor jackets. Chemosphere 160: 173-180.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858525
Grice, MM; Alexander, BH; Hoffbeck, R; Kampa, DM. (2007). Self-reported medical conditions
in perfluorooctanesulfonyl fluoride manufacturing workers. J Occup Environ Med 49:
722-729. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4930271
Guo, R; Reiner, EJ; Bhavsar, SP; Helm, PA; Mabury, SA; Braekevelt, E; Tittlemier, SA. (2012).
Determination of polyfluoroalkyl phosphoric acid diesters, perfluoroalkyl phosphonic
acids, perfluoroalkyl phosphinic acids, perfluoroalkyl carboxylic acids, and
7-48
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
perfluoroalkane sulfonic acids in lake trout from the Great Lakes region. Anal Bioanal
Chem 404: 2699-2709.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919419
Gyllenhammar, I; Benskin, JP; Sandblom, O; Berger, U; Ahrens, L; Lignell, S; Wiberg, K;
Glynn, A. (2018). Perfluoroalkyl acids (PFAAs) in serum from 2-4-month-old infants:
Influence of maternal serum concentration, gestational age, breast-feeding, and
contaminated drinking water. Environ Sci Technol 52: 7101-7110.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4778766
Gyllenhammar, I; Diderholm, B; Gustafsson, J; Berger, U; Ridefelt, P; Benskin, JP; Lignell, S;
Lampa, E; Glynn, A. (2018). Perfluoroalkyl acid levels in first-time mothers in relation to
offspring weight gain and growth. Environ Int 111: 191-199.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238300
Hack, M; Klein, NK; Taylor, HG. (1995). Long-term developmental outcomes of low birth
weight infants. Future Child 5: 176-196.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/8632216
Hagenaars, A; Vergauwen, L; De Coen, W; Knapen, D. (2011). Structure-activity relationship
assessment of four perfluorinated chemicals using a prolonged zebrafish early life stage
test. Chemosphere 82: 764-772.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1279113
Hall, AP; Elcombe, CR; Foster, JR; Harada, T; Kaufmann, W; Knippel, A; Kiittler, K; Malarkey,
DE; Maronpot, RR; Nishikawa, A; Nolte, T; Schulte, A; Strauss, V; York, MJ. (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. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2718645
Hammer, T; Lophaven, SN; Nielsen, KR; Petersen, MS; Munkholm, P; Weihe, P; Burisch, J;
Lynge, E. (2019). Dietary risk factors for inflammatory bowel diseases in a high-risk
population: Results from the Faroese IBD study. 7: 924-932.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/8776815
Han, R; Hu, M; Zhong, Q; Wan, C; Liu, L; Li, F; Zhang, F; Ding, W. (2018). Perfluorooctane
sulphonate induces oxidative hepatic damage via mitochondria-dependent and NF-
KB/TNF-a-mediated pathway. Chemosphere 191: 1056-1064.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238554
Han, R; Zhang, F; Wan, C; Liu, L; Zhong, Q; Ding, W. (2018). Effect of perfluorooctane
sulphonate-induced Kupffer cell activation on hepatocyte proliferation through the NF-
KB/TNF-a/IL-6-dependent pathway. Chemosphere 200: 283-294.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4355066
Han, W; Gao, Y; Yao, Q; Yuan, T; Wang, Y; Zhao, S; Shi, R; Bonefeld-Jorgensen, EC; Shen, X;
Tian, Y. (2018). Perfluoroalkyl and polyfluoroalkyl substances in matched parental and
cord serum in Shandong, China. Environ Int 116: 206-213.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080230
Han, X; Meng, L; Zhang, G; Li, Y; Shi, Y; Zhang, Q; Jiang, G. (2021). Exposure to novel and
legacy per- and polyfluoroalkyl substances (PFASs) and associations with type 2
diabetes: A case-control study in East China. Environ Int 156: 106637.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7762348
Hanahan, D. (2022). Hallmarks of cancer: New dimensions [Review], Cancer Discovery 12: 31-
46. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10164687
7-49
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Hanahan, D; Weinberg, RA. (2000). The hallmarks of cancer [Review], Cell 100: 57-70.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/188413
Hanahan, D; Weinberg, RA. (2011). Hallmarks of cancer: The next generation [Review], Cell
144: 646-674. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/758924
Hanssen, L; Dudarev, AA; Huber, S; Odland, J0; Nieboer, E; Sandanger, TM. (2013). Partition
of perfluoroalkyl substances (PFASs) in whole blood and plasma, assessed in maternal
and umbilical cord samples from inhabitants of arctic Russia and Uzbekistan. Sci Total
Environ 447: 430-437.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859848
Hanssen, L; Rollin, H; Odland, J0; Moe, MK; Sandanger, TM. (2010). Perfluorinated
compounds in maternal serum and cord blood from selected areas of South Africa: results
of a pilot study. J Environ Monit 12: 1355-1361.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919297
Harada, KH; Hashida, S; Kaneko, T; Takenaka, K; Minata, M; Inoue, K; Saito, N; Koizumi, A.
(2007). Biliary excretion and cerebrospinal fluid partition of perfluorooctanoate and
perfluorooctane sulfonate in humans. Environ Toxicol Pharmacol 24: 134-139.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919450
Hardell, E; Karrman, A; van Bavel, B; Bao, J; Carlberg, M; Hardell, L. (2014). Case-control
study on perfluorinated alkyl acids (PFAAs) and the risk of prostate cancer. Environ Int
63: 35-39. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2968084
Harris, MH; Oken, E; Rifas-Shiman, SL; Calafat, AM; Ye, X; Bellinger, DC; Webster, TF;
White, RF; Sagiv, SK. (2018). Prenatal and childhood exposure to per- and
polyfluoroalkyl substances (PFASs) and child cognition. Environ Int 115: 358-369.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4442261
Hartman, TJ; Calafat, AM; Holmes, AK; Marcus, M; Northstone, K; Flanders, WD; Kato, K;
Taylor, EV. (2017). Prenatal exposure to perfluoroalkyl substances and body fatness in
girls. Child Obes 13: 222-230.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859812
Haug, LS; Huber, S; Becher, G; Thomsen, C. (2011). Characterisation of human exposure
pathways to perfluorinated compounds—comparing exposure estimates with biomarkers
of exposure. Environ Int 37: 687-693.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2577501
Haukas, M; Berger, U; Hop, H; Gulliksen, B; Gabrielsen, GW. (2007). Bioaccumulation of per-
and polyfluorinated alkyl substances (PFAS) in selected species from the Barents Sea
food web. Environ Pollut 148: 360-371.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2158020
He, X; Liu, Y; Xu, B; Gu, L; Tang, W. (2018). PFOA is associated with diabetes and metabolic
alteration in US men: National Health and Nutrition Examination Survey 2003-2012. Sci
Total Environ 625: 566-574.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238388
Heffernan, AL; Cunningham, TK; Drage, DS; Aylward, LL; Thompson, K; Vijayasarathy, S;
Mueller, JF; Atkin, SL; Sathyapalan, T. (2018). Perfluorinated alkyl acids in the serum
and follicular fluid of UK women with and without polycystic ovarian syndrome
undergoing fertility treatment and associations with hormonal and metabolic parameters.
Int J Hyg Environ Health 221: 1068-1075.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079713
7-50
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Henry, ND; Fair, PA. (2013). Comparison of in vitro cytotoxicity, estrogenicity and anti-
estrogenicity of triclosan, perfluorooctane sulfonate and perfluorooctanoic acid. J Appl
Toxicol 33: 265-272.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1805116
Hill, AB. (1965). The environment and disease: Association or causation? Proc R Soc Med 58:
295-300. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/71664
Hoffman, K; Webster, TF; Weisskopf, MG; Weinberg, J; Vieira, VM. (2010). Exposure to
polyfluoroalkyl chemicals and attention deficit/hyperactivity disorder in U.S. children 12-
15 years of age. Environ Health Perspect 118: 1762-1767.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291112
Holder, C; Deluca, N; Luh, J; Soleymani, P; Vallero, D; Cohen Hubal, E. (2021 in prep.). (In
Press) Systematic evidence mapping of potential exposure pathways for per- and poly-
fluoroalkyl (PFAS) chemicals based on measured occurrence in multiple media.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9419128
Holsapple, MP; Pitot, HC; Cohen, SM; Boobis, AR; Klaunig, JE; Pastoor, T; Dellarco, VL;
Dragan, YP. (2006). Mode of action in relevance of rodent liver tumors to human cancer
risk. Toxicol Sci 89: 51-56.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/194740
Honda-Kohmo, K; Hutcheson, R; Innes, KE; Conway, BN. (2019). Perfluoroalkyl substances are
inversely associated with coronary heart disease in adults with diabetes. J Diabetes
Complications 33: 407-412.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080551
Houde, M; Martin, JW; Letcher, RJ; Solomon, KR; Muir, DCG. (2006). Biological Monitoring
of Polyfluoroalkyl Substances: A Review [Review], Environ Sci Technol 40: 3463-3473.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290875
H0yer, BB; Ramlau-Hansen, CH; Obel, C; Pedersen, HS; Hernik, A; Ogniev, V; Jonsson, BA;
Lindh, CH; Rylander, L; Rignell-Hydbom, A; Bonde, JP; Toft, G. (2015). Pregnancy
serum concentrations of perfluorinated alkyl substances and offspring behaviour and
motor development at age 5-9 years—a prospective study. Environ Health 14: 2.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851038
Hu, X; Li, S; Cirillo, PM; Krigbaum, NY; Tran, V; Jones, DP; Cohn, BA. (2019). Metabolome
wide association study of serum poly and perfluoroalkyl substances (PFASs) in
pregnancy and early postpartum. Reprod Toxicol 87: 70-78.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5412445
Hu, XDC; Tokranov, AK; Liddie, J; Zhang, XM; Grandjean, P; Hart, JE; Laden, F; Sun, Q;
Yeung, LWY; Sunderland, EM. (2019). Tap Water Contributions to Plasma
Concentrations of Poly- and Perfluoroalkyl Substances (PFAS) in a Nationwide
Prospective Cohort of U.S. Women. Environ Health Perspect 127: 67006.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5381562
Hu, XZ; Hu, DC. (2009). Effects of perfluorooctanoate and perfluorooctane sulfonate exposure
on hepatoma Hep G2 cells. Arch Toxicol 83: 851-861.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919334
Hu, Y; Liu, G; Rood, J; Liang, L; Bray, GA; de Jonge, L; Coull, B; Furtado, JD; Qi, L;
Grandjean, P; Sun, Q. (2019). Perfluoroalkyl substances and changes in bone mineral
density: A prospective analysis in the POUNDS-LOST study. Environ Res 179: 108775.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315798
7-51
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Huang, H; Yu, K; Zeng, X; Chen, Q; Liu, Q; Zhao, Y; Zhang, J; Zhang, X; Huang, L. (2020).
Association between prenatal exposure to perfluoroalkyl substances and respiratory tract
infections in preschool children. Environ Res 191: 110156.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6988475
Huang, JS; Borensztajn, J; Reddy, JK. (2011). Hepatic lipid metabolism. In SPS Monga (Ed.),
Molecular pathology of liver diseases (pp. 133-146). Boston, MA: Springer.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284973
Huang, M; Jiao, J; Zhuang, P; Chen, X; Wang, J; Zhang, Y. (2018). Serum polyfluoroalkyl
chemicals are associated with risk of cardiovascular diseases in national US population.
Environ Int 119: 37-46.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5024212
Huang, MC; Dzierlenga, AL; Robinson, VG; Waidyanatha, S; Devito, MJ; Eifrid, MA;
Granville, CA; Gibbs, ST; Blystone, CR. (2019). Toxicokinetics of perfluorobutane
sulfonate (PFBS), perfluorohexane-l-sulphonic acid (PFHxS), and perfluorooctane
sulfonic acid (PFOS) in male and female Hsd:Sprague Dawley SD rats after intravenous
and gavage administration. Toxicology Reports 6: 645-655.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5387170
Huang, MC; Dzierlenga, AL; Robinson, VG; Waidyanatha, S; Devito, MJ; Eifrid, MA;
Granville, CA; Gibbs, ST; Blystone, CR. (2021). Corrigendum to "Toxicokinetics of
perfluorobutane sulfonate (PFBS), perfluorohexane-l-sulphonic acid (PFHxS), and
perfluorooctane sulfonic acid (PFOS) in male and female Hsd:Sprague Dawley SD rats
after intravenous and gavage administration" [Toxicol. Rep. 6 (2019) 645-655]
[Erratum], Toxicology Reports 8: 365.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7410147
Huang, Q; Zhang, J; Peng, S; Du, M; Ow, S; Pu, H; Pan, C; Shen, H. (2014). Proteomic analysis
of perfluorooctane sulfonate-induced apoptosis in human hepatic cells using the iTRAQ
technique. J Appl Toxicol 34: 1342-1351.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851292
Huang, R; Chen, Q; Zhang, L; Luo, K; Chen, L; Zhao, S; Feng, L; Zhang, J. (2019). Prenatal
exposure to perfluoroalkyl and polyfluoroalkyl substances and the risk of hypertensive
disorders of pregnancy. Environ Health 18:5.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5083564
Huck, I; Beggs, K; Apte, U. (2018). Paradoxical protective effect of perfluorooctanesulfonic acid
against high-fat diet-induced hepatic steatosis in mice. Int J Toxicol 37: 383-392.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079648
Humblet, O; Diaz-Ramirez, LG; Balmes, JR; Pinney, SM; Hiatt, RA. (2014). Perfluoroalkyl
chemicals and asthma among children 12-19 years of age: NHANES (1999-2008).
Environ Health Perspect 122: 1129-1133.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851240
Hunter, P. (2012). The inflammation theory of disease. The growing realization that chronic
inflammation is crucial in many diseases opens new avenues for treatment. EMBO Rep
13: 968-970.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284266
Huo, X; Huang, R; Gan, Y; Luo, K; Aimuzi, R; Nian, M; Ao, J; Feng, L; Tian, Y; Wang, W; Ye,
W; Zhang, J. (2020). Perfluoroalkyl substances in early pregnancy and risk of
hypertensive disorders of pregnancy: A prospective cohort study. Environ Int 138:
7-52
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
105656. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6505752
Hurley, S; Goldberg, D; Wang, M; Park, JS; Petreas, M; Bernstein, L; Anton-Culver, H; Nelson,
DO; Reynolds, P. (2018). Breast cancer risk and serum levels of per- and poly-
fluoroalkyl substances: a case-control study nested in the California Teachers Study.
Environ Health 17: 83.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080646
Hutcheson, R; Innes, K; Conway, B. (2020). Perfluoroalkyl substances and likelihood of stroke
in persons with and without diabetes. Diab Vase Dis Res 17: 1-8.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6320195
ILEPA. Title 35: Environmental Protection. Subtitle: Public Water Supplies. Chapter I: Pollution
Control Board. Part 620 Groundwater Quality, (2019).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9417528
Impinen, A; Longnecker, MP; Nygaard, UC; London, SJ; Ferguson, KK; Haug, LS; Granum, B.
(2019). Maternal levels of perfluoroalkyl substances (PFASs) during pregnancy and
childhood allergy and asthma related outcomes and infections in the Norwegian Mother
and Child (MoBa) cohort. Environ Int 124: 462-472.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080609
Impinen, A; Nygaard, UC; L0drup Carlsen, KC; Mowinckel, P; Carlsen, KH; Haug, LS;
Granum, B. (2018). Prenatal exposure to perfluoralkyl substances (PFASs) associated
with respiratory tract infections but not allergy- and asthma-related health outcomes in
childhood. Environ Res 160: 518-523.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238440
Inoue, K; Okada, F; Ito, R; Kato, S; Sasaki, S; Nakajima, S; Uno, A; Saijo, Y; Sata, F;
Yoshimura, Y; Kishi, R; Nakazawa, H. (2004). Perfluorooctane sulfonate (PFOS) and
related perfluorinated compounds in human maternal and cord blood samples: assessment
of PFOS exposure in a susceptible population during pregnancy. Environ Health Perspect
112: 1204-1207.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2994839
Inoue, K; Ritz, B; Andersen, SL; Ramlau-Hansen, CH; H0yer, BB; Bech, BH; Henriksen, TB;
Bonefeld-J0rgensen, EC; Olsen, J; Liew, Z. (2019). Perfluoroalkyl Substances and
Maternal Thyroid Hormones in Early Pregnancy; Findings in the Danish National Birth
Cohort. Environ Health Perspect 127: 117002.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5918599
IPCS. (2012). Harmonization project document no. 10: Guidance for immunotoxicity risk
assessment for chemicals. Geneva, Switzerland: World Health Organization.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1249755
Ipsen, J. (1946). Circulating antitoxin at the onset of diphtheria in 425 patients. J Immunol 54:
325-347. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10228563
Ishida, K; Tsuyama, Y; Sanoh, S; Ohta, S; Kotake, Y. (2017). Perfluorooctane sulfonate induces
neuronal vulnerability by decreasing GluR2 expression. Arch Toxicol 91: 885-895.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981472
Ishigame, H; Kakuta, S; Nagai, T; Kadoki, M; Nambu, A; Komiyama, Y; Fujikado, N;
Tanahashi, Y; Akitsu, A; Kotaki, H; Sudo, K; Nakae, S; Sasakawa, C; Iwakura, Y.
(2009). Differential roles of interleukin-17A and -17F in host defense against
mucoepithelial bacterial infection and allergic responses. Immunity 30: 108-119.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284267
7-53
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Itoh, H; Harada, KH; Kasuga, Y; Yokoyama, S; Onuma, H; Nishimura, H; Kusama, R;
Yokoyama, K; Zhu, J; Harada Sassa, M; Tsugane, S; Iwasaki, M. (2021). Serum
perfluoroalkyl substances and breast cancer risk in Japanese women: A case-control
study. Sci Total Environ 800: 149316.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9959632
Itoh, S; Araki, A; Mitsui, T; Miyashita, C; Goudarzi, H; Sasaki, S; Cho, K; Nakazawa, H;
Iwasaki, Y; Shinohara, N; Nonomura, K; Kishi, R. (2016). Association of perfluoroalkyl
substances exposure in utero with reproductive hormone levels in cord blood in the
Hokkaido Study on Environment and Children's Health. Environ Int 94: 51-59.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981465
Itoh, S; Araki, A; Miyashita, C; Yamazaki, K; Goudarzi, H; Minatoya, M; AitBamai, Y;
Kobayashi, S; Okada, E; Kashino, I; Yuasa, M; Baba, T; Kishi, R. (2019). Association
between perfluoroalkyl substance exposure and thyroid hormone/thyroid antibody levels
in maternal and cord blood: The Hokkaido Study. Environ Int 133: 105139.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5915990
Iwabuchi, K; Senzaki, N; Mazawa, D; Sato, I; Hara, M; Ueda, F; Liu, W; Tsuda, S. (2017).
Tissue toxicokinetics of perfluoro compounds with single and chronic low doses in male
rats. J Toxicol Sci 42: 301-317.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859701
Jaacks, LM; Boyd Barr, D; Sundaram, R; Grewal, J; Zhang, C; Buck Louis, GM. (2016). Pre-
pregnancy Maternal Exposure to Persistent Organic Pollutants and Gestational Weight
Gain: A Prospective Cohort Study. Int J Environ Res Public Health 13.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981711
Jackson-Browne, MS; Eliot, M; Patti, M; Spanier, AJ; Braun, JM. (2020). PFAS (per- and
polyfluoroalkyl substances) and asthma in young children: NHANES 2013-2014. Int J
Hyg Environ Health 229: 113565.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833598
Jacquet, N; Maire, MA; Landkocz, Y; Vasseur, P. (2012). Carcinogenic potency of
perfluorooctane sulfonate (PFOS) on Syrian hamster embryo (SHE) cells. Arch Toxicol
86: 305-314. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919219
Jain, R. (2013). Association between thyroid profile and perfluoroalkyl acids: Data from
NHNAES 2007-2008. Environ Res 126: 51-59.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2168068
Jain, RB. (2019). Concentration of selected liver enzymes across the stages of glomerular
function: The associations with PFOA and PFOS. Heliyon 5: e02168.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5381541
Jain, RB. (2020). Associations between selected perfluoroalkyl acids in serum and hemoglobin
in whole blood, a biomarker of anemia: Impact of deteriorating kidney function. Environ
Pollut 263: 114458.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6333438
Jain, RB. (2020). Impact of the co-occurrence of obesity with diabetes, anemia, hypertension,
and albuminuria on concentrations of selected perfluoroalkyl acids. Environ Pollut 266
Pt. 2: 115207.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833623
Jain, RB. (2020). Variabilities in concentrations of selected perfluoroalkyl acids among
normotensives and hypertensives across various stages of glomerular function. Arch
7-54
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Environ Occup Health 76: 1-11.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6311650
Jain, RB; Ducatman, A. (2018). Associations between lipid/lipoprotein levels and perfluoroalkyl
substances among US children aged 6-11 years. Environ Pollut 243: 1-8.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079656
Jain, RB; Ducatman, A. (2019). Dynamics of associations between perfluoroalkyl substances and
uric acid across the various stages of glomerular function. Environ Sci Pollut Res Int 26:
12425-12434.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080378
Jain, RB; Ducatman, A. (2019). Perfluoroalkyl acids and thyroid hormones across stages of
kidney function. Sci Total Environ 696: 133994.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315816
Jain, RB; Ducatman, A. (2019). Perfluoroalkyl acids serum concentrations and their relationship
to biomarkers of renal failure: Serum and urine albumin, creatinine, and albumin
creatinine ratios across the spectrum of glomerular function among US adults. Environ
Res 174: 143-151.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5381566
Jain, RB; Ducatman, A. (2019). Roles of gender and obesity in defining correlations between
perfluoroalkyl substances and lipid/lipoproteins. Sci Total Environ 653: 74-81.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080642
Jain, RB; Ducatman, A. (2019). Selective associations of recent low concentrations of
perfluoroalkyl substances with liver function biomarkers: nhanes 2011 to 2014 data on us
adults aged >20 years. J Occup Environ Med 61: 293-302.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080621
Jain, RB; Ducatman, A. (2020). Associations between apolipoprotein B and selected
perfluoroalkyl substances among diabetics and nondiabetics. Environ Sci Pollut Res Int
2020: 13819-13828.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6988488
Jantzen, CE; Annunziato, KA; Bugel, SM; Cooper, KR. (2016). PFOS, PFNA, and PFOA sub-
lethal exposure to embryonic zebrafish have different toxicity profiles in terms of
morphometries, behavior and gene expression. Aquat Toxicol 175: 160-170.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860114
Jantzen, CE; Annunziato, KM; Cooper, KR. (2016). Behavioral, morphometric, and gene
expression effects in adult zebrafish (Danio rerio) embryonically exposed to PFOA,
PFOS, and PFNA. Aquat Toxicol 180: 123-130.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860109
Jarvis, AL; Justice, JR; Elias, MC; Schnitker, B; Gallagher, K. (2021). Perfluorooctane Sulfonate
in US Ambient Surface Waters: A Review of Occurrence in Aquatic Environments and
Comparison to Global Concentrations. Environ Toxicol Chem 40: 2425-2442.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9416544
Jeddy, Z; Hartman, TJ; Taylor, EV; Poteete, C; Kordas, K. (2017). Prenatal concentrations of
perfluoroalkyl substances and early communication development in British girls. Early
HumDev 109: 15-20.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859807
Jeddy, Z; Tobias, JH; Taylor, EV; Northstone, K; Flanders, WD; Hartman, TJ. (2018). Prenatal
concentrations of perfluoroalkyl substances and bone health in British girls at age 17.
7-55
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Archives of Osteoporosis 13: 84.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079850
Jensen, RC; Andersen, MS; Larsen, PV; Glintborg, D; Dalgard, C; Timmermann, CAG; Nielsen,
F; Sandberg, MB; Andersen, HR; Christesen, HT; Grandjean, P; Jensen, TK. (2020).
Prenatal Exposures to Perfluoroalkyl Acids and Associations with Markers of Adiposity
and Plasma Lipids in Infancy: An Odense Child Cohort Study. Environ Health Perspect
128: 77001. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833719
Jensen, RC; Glintborg, D; Gade Timmermann, CA; Nielsen, F; Kyhl, HB; Frederiksen, H;
Andersson, AM; Juul, A; Sidelmann, JJ; Andersen, HR; Grandjean, P; Andersen, MS;
Jensen, TK. (2020). Prenatal exposure to perfluorodecanoic acid is associated with lower
circulating concentration of adrenal steroid metabolites during mini puberty in human
female infants. The odense child cohort. Environ Res 182: 109101.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6311643
Jensen, RC; Glintborg, D; Timmermann, CAG; Nielsen, F; Kyhl, HB; Andersen, HR; Grandjean,
P; Jensen, TK; Andersen, M. (2018). Perfluoroalkyl substances and glycemic status in
pregnant Danish women: The Odense Child Cohort. Environ Int 116: 101-107.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4354143
Jensen, TK; Andersen, LB; Kyhl, HB; Nielsen, F; Christesen, HT; Grandjean, P. (2015).
Association between Perfluorinated Compound Exposure and Miscarriage in Danish
Pregnant Women. PLoS ONE 10: e0123496.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850253
Ji, J; Song, L; Wang, J; Yang, Z; Yan, H; Li, T; Yu, L; Jian, L; Jiang, F; Li, J; Zheng, J; Li, K.
(2021). Association between urinary per- and poly-fluoroalkyl substances and COVID-19
susceptibility. Environ Int 153: 106524.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7491706
Jiang, W; Zhang, Y; Zhu, L; Deng, J. (2014). Serum levels of perfluoroalkyl acids (PFAAs) with
isomer analysis and their associations with medical parameters in Chinese pregnant
women. Environ Int 64: 40-47.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850910
Jin, H; Zhang, Y; Jiang, W; Zhu, L; Martin, JW. (2016). Isomer-Specific Distribution of
Perfluoroalkyl Substances in Blood. Environ Sci Technol 50: 7808-7815.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859825
Jin, R; Mcconnell, R; Catherine, C; Xu, S; Walker, DI; Stratakis, N; Jones, DP; Miller, GW;
Peng, C; Conti, DV; Vos, MB; Chatzi, L. (2020). Perfluoroalkyl substances and severity
of nonalcoholic fatty liver in Children: An untargeted metabolomics approach. Environ
Int 134: 105220.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315720
Joensen, UN; Bossi, R; Leffers, H; Jensen, AA; Skakkebsek, NE; J0rgensen, N. (2009). Do
perfluoroalkyl compounds impair human semen quality? Environ Health Perspect 117:
923-927. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1405085
Joensen, UN; Veyrand, B; Antignac, JP; Jensen, MB; Petersen, JH; Marchand, P; Skakkebaek,
NE; Andersson, AM; Le Bizec, B; J0rgensen, N. (2014). PFOS
(perfluorooctanesulfonate) in serum is negatively associated with testosterone levels, but
not with semen quality, in healthy men. Hum Reprod 29: 1600-1600.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851244
Johanson, CE. (1979). Distribution of fluid between extracellular and intracellular compartments
7-56
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
in the heart, lungs, liver and spleen of neonatal rats. 36: 282-289.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9641334
Johansson, N; Fredriksson, A; Eriksson, P. (2008). Neonatal exposure to perfluorooctane
sulfonate (PFOS) and perfluorooctanoic acid (PFOA) causes neurobehavioural defects in
adult mice. Neurotoxicology 29: 160-169.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1276156
Johnson, JD; Gibson, SJ; Ober, RE. (1984). Cholestyramine-enhanced fecal elimination of c-14
in rats after administration of ammonium [c-14] perfluorooctanoate or potassium [c-
14]perfluorooctanesulfonate. Fundam Appl Toxicol 4: 972-976.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5085553
J0rgensen, KT; Specht, 10; Lenters, V; Bach, CC; Rylander, L; Jonsson, BA; Lindh, CH;
Giwercman, A; Heederik, D; Toft, G; Bonde, JP. (2014). Perfluoroalkyl substances and
time to pregnancy in couples from Greenland, Poland and Ukraine. Environ Health 13:
116. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851025
Jusko, TA; Oktapodas, M; Palkovicova Murinova, L; Babinska, K; Babjakova, J; Verner, MA;
Dewitt, JC; Thevenet-Morrison, K; Conka, K; Drobna, B; Chovancova, J; Thurston, SW;
Lawrence, BP; Dozier, AM; Jarvinen, KM; Patayova, H; Trnovec, T; Legler, J; Hertz -
Picciotto, I; Lamoree, MH. (2016). Demographic, reproductive, and dietary determinants
of perfluorooctane sulfonic (PFOS) and perfluorooctanoic acid (PFOA) concentrations in
human colostrum. Environ Sci Technol 50: 7152-7162.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981718
Kang, H; Lee, HK; Moon, HB; Kim, S; Lee, J; Ha, M; Hong, S; Kim, S; Choi, K. (2018).
Perfluoroalkyl acids in serum of Korean children: Occurrences, related sources, and
associated health outcomes. Sci Total Environ 645: 958-965.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4937567
Kang, Q; Gao, F; Zhang, X; Wang, L; Liu, J; Fu, M; Zhang, S; Wan, Y; Shen, H; Hu, J. (2020).
Nontargeted identification of per- and polyfluoroalkyl substances in human follicular
fluid and their blood-follicle transfer. Environ Int 139: 105686.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6356899
Kannan, K; Tao, L; Sinclair, E; Pastva, SD; Jude, DJ; Giesy, JP. (2005). Perfluorinated
compounds in aquatic organisms at various trophic levels in a Great Lakes food chain.
Arch Environ Contam Toxicol 48: 559-566.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290874
Kapraun, DF; Zurlinden, TJ; Verner, M-A; Chiang, C; Dzierlenga, MW; Carlson, LM;
Schlosser, PM; Lehmann, GM. (2022). A Generic Pharmacokinetic Model for
Quantifying Mother-to-Offspring Transfer of Lipophilic Persistent Environmental
Chemicals, https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9641977
Karakas-Celik, S; Aras, N. (2014). An in-vitro investigation of the effect of perfluorooctane
sulphonate on cell lines of embryonic origin. Mol Biol Rep 41: 3755-3759.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850400
Karlsen, M; Grandjean, P; Weihe, P; Steuerwald, U; Oulhote, Y; Valvi, D. (2017). Early-life
exposures to persistent organic pollutants in relation to overweight in preschool children.
Reprod Toxicol 68: 145-153.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858520
Karrman, A; Domingo, JL; Llebaria, X; Nadal, M; Bigas, E; van Bavel, B; Lindstrom, G. (2010).
Biomonitoring perfluorinated compounds in Catalonia, Spain: concentrations and trends
7-57
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
in human liver and milk samples. Environ Sci Pollut Res Int 17: 750-758.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2732071
Karrman, A; Ericson, I; van Bavel, B; Darnerud, PO; Aune, M; Glynn, A; Lignell, S; Lindstrom,
G. (2007). Exposure of perfluorinated chemicals through lactation: Levels of matched
human milk and serum and a temporal trend, 1996-2004, in Sweden. Environ Health
Perspect 115: 226-230.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290903
Karrman, A; van Bavel, B; Jarnberg, U; Hardell, L; Lindstrom, G. (2006). Perfluorinated
chemicals in relation to other persistent organic pollutants in human blood. Chemosphere
64: 1582-1591.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2159543
Kataria, A; Trachtman, H; Malaga-Dieguez, L; Trasande, L. (2015). Association between
perfluoroalkyl acids and kidney function in a cross-sectional study of adolescents.
Environ Health 14: 89.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859835
Kato, K; Wong, LY; Chen, A; Dunbar, C; Webster, GM; Lanphear, BP; Calafat, AM. (2014).
Changes in serum concentrations of maternal poly- and perfluoroalkyl substances over
the course of pregnancy and predictors of exposure in a multiethnic cohort of Cincinnati,
Ohio pregnant women during 2003-2006. Environ Sci Technol 48: 9600-9608.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851230
Kato, S; Itoh, S; Yuasa, M; Baba, T; Miyashita, C; Sasaki, S; Nakajima, S; Uno, A; Nakazawa,
H; Iwasaki, Y; Okada, E; Kishi, R. (2016). Association of perfluorinated chemical
exposure in utero with maternal and infant thyroid hormone levels in the Sapporo cohort
of Hokkaido Study on the Environment and Children's Health. Environ Health Prev Med
21: 334-344. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981723
Kavlock, RJ; Allen, BC; Faustman, EM; Kimmel, CA. (1995). Dose-response assessments for
developmental toxicity. IV. Benchmark doses for fetal weight changes. Toxicol Sci 26:
211-222. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/75837
Kawamoto, K; Oashi, T; Oami, K; Liu, W; Jin, YH; Saito, N; Sato, I; Tsuda, S. (2010).
Perfluorooctanoic acid (PFOA) but not perfluorooctane sulfonate (PFOS) showed DNA
damage in comet assay on Paramecium caudatum. J Toxicol Sci 35: 835-841.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1274162
Kawamoto, K; Sato, I; Tsuda, S; Yoshida, M; Yaegashi, K; Saito, N; Liu, W; Jin, Y. (2011).
Ultrasonic-induced tonic convulsion in rats after subchronic exposure to perfluorooctane
sulfonate (PFOS). J Toxicol Sci 36: 55-62.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919266
Keil, DE; Mehlmann, T; Butterworth, L; Peden-Adams, MM. (2008). Gestational exposure to
perfluorooctane sulfonate suppresses immune function in B6C3F1 mice. Toxicol Sci 103:
77-85. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332422
Keller & Heckman LLP. (2021). Attack on PFAS extends to food packaging. National Law
Review X. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9419081
Kelly, BC; Ikonomou, MG; Blair, JD; Surridge, B; Hoover, D; Grace, R; Gobas, FA. (2009).
Perfluoroalkyl contaminants in an Arctic marine food web: trophic magnification and
wildlife exposure. Environ Sci Technol 43: 4037-4043.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1276129
Kerstner-Wood, C; Coward, L; Gorman, G; Southern Research Institute. (2003). Protein binding
7-58
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
of perfluorobutane sulfonate, perfluorohexane sulfonate, perfluorooctane sulfonate and
perfluorooctanoate to plasma (human, rat, and monkey), and various human-derived
plasma protein fractions. Study ID 9921.7 [TSCA Submission], (8EHQ-04-15845A;
88040000364). St. Paul, MN: 3M Company.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4771364
Khalil, N; Chen, A; Lee, M; Czerwinski, SA; Ebert, JR; Dewitt, JC; Kannan, K. (2016).
Association of Perfluoroalkyl Substances, Bone Mineral Density, and Osteoporosis in the
US Population in NHANES 2009-2010. Environ Health Perspect 124: 81-87.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3229485
Khalil, N; Ducatman, AM; Sinari, S; Billheimer, D; Hu, C; Littau, S; Burgess, JL. (2020). Per-
and polyfluoroalkyl substance and cardio metabolic markers in firefighters. J Occup
Environ Med 62: 1076-1081.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7021479
Khalil, N; Ebert, JR; Honda, M; Lee, M; Nahhas, RW; Koskela, A; Hangartner, T; Kannan, K.
(2018). Perfluoroalkyl substances, bone density, and cardio-metabolic risk factors in
obese 8-12 year old children: A pilot study. Environ Res 160: 314-321.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238547
Khansari, MR; Yousefsani, BS; Kobarfard, F; Faizi, M; Pourahmad, J. (2017). In vitro toxicity
of perfluorooctane sulfonate on rat liver hepatocytes: probability of distinctive binding to
CYP 2E1 and involvement of cellular proteolysis. Environ Sci Pollut Res Int 24: 23382-
23388. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981272
Khazaee, M; Guardian, MGE; Aga, DS; Ng, CA. (2019). Impacts of sex and exposure duration
on gene expression in zebrafish following perfluorooctane sulfonate (PFOS) exposure.
Environ Toxicol Chem 39: 437-449.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5918850
Kielsen, K; Shamim, Z; Ryder, LP; Nielsen, F; Grandjean, P; Budtz-j0rgensen, E; Heilmann, C.
(2016). Antibody response to booster vaccination with tetanus and diphtheria in adults
exposed to perfluorinated alkylates. J Immunotoxicol 13: 270-273.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4241223
Kim, DH; Kim, UJ; Kim, HY; Choi, SD; Oh, JE. (2016). Perfluoroalkyl substances in serum
from South Korean infants with congenital hypothyroidism and healthy infants - Its
relationship with thyroid hormones. Environ Res 147: 399-404.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3351917
Kim, DH; Lee, JH; Oh, JE. (2019). Assessment of individual-based perfluoroalkly substances
exposure by multiple human exposure sources. J Hazard Mater 365: 26-33.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080673
Kim, HY; Kim, KN; Shin, CH; Lim, YH; Kim, JI; Kim, BN; Hong, YC; Lee, YA. (2020). The
relationship between perfluoroalkyl substances concentrations and thyroid function in
early childhood: A prospective cohort study. Thyroid 30: 1556-1565.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833758
Kim, MJ; Moon, S; Oh, BC; Jung, D; Ji, K; Choi, K; Park, YJ. (2018). Association between
perfluoroalkyl substances exposure and thyroid function in adults: A meta-analysis. PLoS
ONE 13: eO 197244.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079795
Kim, S; Choi, K; Ji, K; Seo, J; Kho, Y; Park, J; Kim, S; Park, S; Hwang, I; Jeon, J; Yang, H;
Giesy, JP. (2011). Trans-placental transfer of thirteen perfluorinated compounds and
7-59
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
relations with fetal thyroid hormones. Environ Sci Technol 45: 7465-7472.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1424975
Kim, SJ; Heo, SH; Lee, DS; Hwang, IG; Lee, YB; Cho, HY. (2016). Gender differences in
pharmacokinetics and tissue distribution of 3 perfluoroalkyl and polyfluoroalkyl
substances in rats. Food Chem Toxicol 97: 243-255.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3749289
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.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1289790
Kim, SK; Lee, KT; Kang, CS; Tao, L; Kannan, K; Kim, KR; Kim, CK; Lee, JS; Park, PS; Yoo,
YW; Ha, JY; Shin, YS; Lee, JH. (2011). Distribution of perfluorochemicals between sera
and milk from the same mothers and implications for prenatal and postnatal exposures.
Environ Pollut 159: 169-174.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919258
Kim, YR; White, N; Braunig, J; Vijayasarathy, S; Mueller, JF; Knox, CL; Harden, FA; Pacella,
R; Toms, LL. (2020). Per- and poly-fluoroalkyl substances (PFASs) in follicular fluid
from women experiencing infertility in Australia. Environ Res 190: 109963.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833596
Kingsley, SL; Walker, DI; Calafat, AM; Chen, A; Papandonatos, GD; Xu, Y; Jones, DP;
Lanphear, BP; Pennell, KD; Braun, JM. (2019). Metabolomics of childhood exposure to
perfluoroalkyl substances: across-sectional study. Metabolomics 15: 95.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5405904
Kishi, R; Nakajima, T; Goudarzi, H; Kobayashi, S; Sasaki, S; Okada, E; Miyashita, C; Itoh, S;
Araki, A; Ikeno, T; Iwasaki, Y; Nakazawa, H. (2015). The association of prenatal
exposure to perfluorinated chemicals with maternal essential and long-chain
polyunsaturated fatty acids during pregnancy and the birth weight of their offspring: the
hokkaido study. Environ Health Perspect 123: 1038-1045.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850268
Klamt, A; Huniar, U; Spycher, S; Keldenich, J. (2008). COSMOmic: a mechanistic approach to
the calculation of membrane-water partition coefficients and internal distributions within
membranes and micelles. JPhys Chem B 112: 12148-12157.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9641966
Klaunig, JE; Babich, MA; Baetcke, KP; Cook, JC; Corton, JC; David, RM; Deluca, JG; Lai, DY;
Mckee, RH; Peters, JM; Roberts, RA; Fenner-Crisp, PA. (2003). PPARalpha agonist-
induced rodent tumors: modes of action and human relevance [Review], Crit Rev Toxicol
33: 655-780. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5772415
Knox, SS; Jackson, T; Javins, B; Frisbee, SJ; Shankar, A; Ducatman, AM. (2011). Implications
of early menopause in women exposed to perfluorocarbons. J Clin Endocrinol Metab 96:
1747-1753. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1402395
Kobayashi, S; Azumi, K; Goudarzi, H; Araki, A; Miyashita, C; Kobayashi, S; Itoh, S; Sasaki, S;
Ishizuka, M; Nakazawa, H; Ikeno, T; Kishi, R. (2017). Effects of prenatal perfluoroalkyl
acid exposure on cord blood IGF2/H19 methylation and ponderal index: The Hokkaido
Study. J Expo Sci Environ Epidemiol 27: 251-259.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981430
Kobayashi, S; Sata, F; Goudarzi, H; Araki, A; Miyashita, C; Sasaki, S; Okada, E; Iwasaki, Y;
7-60
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Nakajima, T; Kishi, R. (2021). Associations among
perfluorooctanesulfonic/perfluorooctanoic acid levels, nuclear receptor gene
polymorphisms, and lipid levels in pregnant women in the Hokkaido study. Sci Rep 11:
9994. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/8442188
Kobayashi, S; Sata, F; Ikeda-Araki, A; Miyashita, C; Goudarzi, H; Iwasaki, Y; Nakajima, T;
Kishi, R. (2022). Relationships between maternal perfluoroalkyl substance levels,
polymorphisms of receptor genes, and adverse birth outcomes in the Hokkaido birth
cohort study, Japan. Reprod Toxicol 107: 112-122.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10176408
Koshy, TT; Attina, TM; Ghassabian, A; Gilbert, J; Burdine, LK; Marmor, M; Honda, M; Chu,
DB; Han, X; Shao, Y; Kannan, K; Urbina, EM; Trasande, L. (2017). Serum
perfluoroalkyl substances and cardiometabolic consequences in adolescents exposed to
the World Trade Center disaster and a matched comparison group. Environ Int 109: 128-
135. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238478
Koskela, A; Ducatman, A; Schousboe, JT; Nahhas, RW; Khalil, N. (2022). Perfluoroalkyl
Substances and Abdominal Aortic Calcification. J Occup Environ Med.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10176386
Kotlarz, N; Mccord, J; Collier, D; Lea, CS; Strynar, M; Lindstrom, AB; Wilkie, AA; Islam, JY;
Matney, K; Tarte, P; Polera, ME; Burdette, K; Dewitt, J; May, K; Smart, RC; Knappe,
DRU; Hoppin, JA. (2020). Measurement of Novel, Drinking Water-Associated PFAS in
Blood from Adults and Children in Wilmington, North Carolina. Environ Health Perspect
128: 77005. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833715
Kotthoff, M; Miiller, J; Jiirling, H; Schlummer, M; Fiedler, D. (2015). Perfluoroalkyl and
polyfluoroalkyl substances in consumer products. Environ Sci Pollut Res Int 22: 14546-
14559. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850246
Kuczmarski, RJ; Ogden, CL; Guo, SS; Grummer-Strawn, LM; Flegal, KM; Mei, Z; Wei, R;
Curtin, LR; Roche, AF; Johnson, CL. (2002). 2000 CDC growth charts for the United
States: Methods and development. (Vital and Health Statistics: Series 11, No. 246).
Hyattsville, MD: National Center for Health Statistics.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3490881
Kuklenyik, Z; Reich, JA; Tully, JS; Needham, LL; Calafat, AM. (2004). Automated solid-phase
extraction and measurement of perfluorinated organic acids and amides in human serum
and milk. Environ Sci Technol 38: 3698-3704.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1598132
Kummu, M; Sieppi, E; Koponen, J; Laatio, L; VaoHaoKangas, K; Kiviranta, H; Rautio, A;
Myllynen, P. (2015). Organic anion transporter 4 (OAT 4) modifies placental transfer of
perfluorinated alkyl acids PFOS and PFOA in human placental ex vivo perfusion system.
Placenta 36: 1185-1191.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3789332
Kvalem, HE; Nygaard, UC; L0drup Carlsen, KC; Carlsen, KH; Haug, LS; Granum, B. (2020).
Perfluoroalkyl substances, airways infections, allergy and asthma related health outcomes
- Implications of gender, exposure period and study design. Environ Int 134: 105259.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6316210
Kvist, L; Giwercman, YL; Jonsson, BA; Lindh, CH; Bonde, JP; Toft, G; Strucinski, P; Pedersen,
HS; Zvyezday, V; Giwercman, A. (2012). Serum levels of perfluorinated compounds and
sperm Y:X chromosome ratio in two European populations and in Inuit from Greenland.
7-61
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Reprod Toxicol 34: 644-650.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919170
Lai, KP; Lee, JC; Wan, HT; Li, JW; Wong, AY; Chan, TF; Oger, C; Galano, JM; Durand, T;
Leung, KS; Leung, CC; Li, R; Wong, CK. (2017). Effects of in Utero PFOS Exposure on
Transcriptome, Lipidome, and Function of Mouse Testis. Environ Sci Technol 51: 8782-
8794. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981773
Lai, KP; Li, JW; Cheung, A; Li, R; Billah, MB; Chan, TF; Wong, CK. (2017). Transcriptome
sequencing reveals prenatal PFOS exposure on liver disorders. Environ Pollut 223: 416-
425. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981375
Lai, KP; Ng, AH; Wan, HT; Wong, AY; Leung, CC; Li, R; Wong, CK. (2018). Dietary exposure
to the environmental chemical, pfos on the diversity of gut microbiota, associated with
the development of metabolic syndrome. FMICB 9: 2552.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080641
Lange, F; Schmidt, C; Brauch, H-J. (2006). Perfluoroalkylcarboxylates and -sulfonates:
Emerging Contaminants for Drinking Water Supplies? Nieuwegein, The Netherlands:
Association of River Waterworks - RIWA.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10411376
Lau, C; Thibodeaux, JR; Hanson, RG; Rogers, JM; Grey, BE; Stanton, ME; Butenhoff, JL;
Stevenson, LA. (2003). Exposure to perfluorooctane sulfonate during pregnancy in rat
and mouse. II: postnatal evaluation. Toxicol Sci 74: 382-392.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757854
Lauritzen, HB; Larose, TL; 0ien, T; Sandanger, TM; Odland, J0; van de Bor, M; Jacobsen, GW.
(2018). Prenatal exposure to persistent organic pollutants and child overweight/obesity at
5-year follow-up: a prospective cohort study. Environ Health 17: 9.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4217244
Leary, DB; Takazawa, M; Kannan, K; Khalil, N. (2020). Perfluoroalkyl substances and
metabolic syndrome in firefighters a pilot study. J Occup Environ Med 62: 52-57.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7240043
Lebeaux, RM; Doherty, BT; Gallagher, LG; Zoeller, RT; Hoofnagle, AN; Calafat, AM; Karagas,
MR; Yolton, K; Chen, A; Lanphear, BP; Braun, JM; Romano, ME. (2020). Maternal
serum perfluoroalkyl substance mixtures and thyroid hormone concentrations in maternal
and cord sera: The HOME Study. Environ Res 185: 109395.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6356361
Lee, CK; Kang, SG; Lee, JT; Lee, SW; Kim, JH; Kim, DH; Son, BC; Kim, KH; Suh, CH; Kim,
SY; Park, YB. (2015). Effects of perfluorooctane sulfuric acid on placental PRL-family
hormone production and fetal growth retardation in mice. Mol Cell Endocrinol 401: 165-
172. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851075
Lee, J; Lee, S; Choi, YA, e; Jin, M; Kim, YY; Kang, BC; Kim, M; Dhakal, H; Lee, S; Kim, S;
Khang, D; Kim, SH. (2018). Perfluorooctane sulfonate exacerbates mast cell-mediated
allergic inflammation by the release of histamine. Mol Cell Toxicol 14: 173-181.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5085013
Lee, JS; Ward, WO; Liu, J; Ren, H; Vallanat, B; Delker, D; Corton, JC. (2011). Hepatic
xenobiotic metabolizing enzyme and transporter gene expression through the life stages
of the mouse. PLoS ONE 6: e24381.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3114850
Lee, JW; Choi, K; Park, K; Seong, C; Yu, SD; Kim, P. (2020). Adverse effects of perfluoroalkyl
7-62
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
acids on fish and other aquatic organisms: A review [Review], Sci Total Environ 707:
135334. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6323794
Lee, S; Kim, S; Park, J; Kim, HJ; Choi, G; Choi, S; Kim, S; Kim, SY; Kim, S; Choi, K; Moon,
HB. (2017). Perfluoroalkyl substances (PFASs) in breast milk from Korea: Time-course
trends, influencing factors, and infant exposure. Sci Total Environ 612: 286-292.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3983576
Lee, YJ; Kim, MK; Bae, J; Yang, JH. (2013). Concentrations of perfluoroalkyl compounds in
maternal and umbilical cord sera and birth outcomes in Korea. Chemosphere 90: 1603-
1609. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859850
Lefebvre, DE; Curran, I; Armstrong, C; Coady, L; Parenteau, M; Liston, V; Barker, M; Aziz, S;
Rutherford, K; Bellon-Gagnon, P; Shenton, J; Mehta, R; Bondy, G. (2008).
Immunomodulatory effects of dietary potassium perfluorooctane sulfonate (PFOS)
exposure in adult Sprague-Dawley rats. J Toxicol Environ Health A 71: 1516-1525.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1276155
Lehmann, GM; Verner, MA; Luukinen, B; Henning, C; Assimon, SA; Lakind, JS; Mclanahan,
ED; Phillips, LJ; Davis, MH; Powers, CM; Hines, EP; Haddad, S; Longnecker, MP;
Poulsen, MT; Farrer, DG; Marchitti, SA; Tan, YM; Swartout, JC; Sagiv, SK; Welsh, C;
Campbell, JL; Foster, WG; Yang, RS; Fenton, SE; Tornero-Velez, R; Francis, BM;
Barnett, JB; El-Masri, HA; Simmons, JE. (2014). Improving the risk assessment of
lipophilic persistent environmental chemicals in breast milk [Review], Crit Rev Toxicol
44: 600-617. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2447276
Lehner, R; Quiroga, AD. (2016). Chapter 5: Fatty acid handling in mammalian cells. In ND
Ridgway; RS McLeod (Eds.), Biochemistry of lipids, lipoproteins and membranes (6th
ed., pp. 149-184). Amsterdam, Netherlands: Elsevier.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284974
Lenters, V; Iszatt, N; Forns, J; Cechova, E; Kocan, A; Legler, J; Leonards, P; Stigum, H;
Eggesb0, M. (2019). Early-life exposure to persistent organic pollutants (OCPs, PBDEs,
PCBs, PFASs) and attention-deficit/hyperactivity disorder: A multi-pollutant analysis of
a Norwegian birth cohort. Environ Int 125: 33-42.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080366
Leter, G; Consales, C; Eleuteri, P; Uccelli, R; Specht, 10; Toft, G; Moccia, T; Budillon, A;
Jonsson, BA; Lindh, CH; Giwercman, A; Pedersen, HS; Ludwicki, JK; Zviezdai, V;
Heederik, D; Bonde, JP; Spano, M. (2014). Exposure to perfluoroalkyl substances and
sperm DNA global methylation in Arctic and European populations. Environ Mol
Mutagen 55: 591-600.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2967406
Leung, YK; Ouyang, B; Niu, L; Xie, C; Ying, J; Medvedovic, M; Chen, A; Weihe, P; Valvi, D;
Grandjean, P; Ho, SM. (2018). Identification of sex-specific DNA methylation changes
driven by specific chemicals in cord blood in a Faroese birth cohort. Epigenetics 13: 290-
300. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4633577
Lewis, RC; Johns, LE; Meeker, JD. (2015). Serum Biomarkers of Exposure to Perfluoroalkyl
Substances in Relation to Serum Testosterone and Measures of Thyroid Function among
Adults and Adolescents from NHANES 2011-2012. Int J Environ Res Public Health 12:
6098-6114. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3749030
Li, H; Hammarstrand, S; Midberg, B; Xu, Y; Li, Y; Olsson, DS; Fletcher, T; Jakobsson, K;
Andersson, EM. (2022). Cancer incidence in a Swedish cohort with high exposure to
7-63
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
perfluoroalkyl substances in drinking water. Environ Res 204: 112217.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9961926
Li, J; Cai, D; Chu, C; Li, QQ; Zhou, Y; Hu, LW; Yang, BY; Dong, GH; Zeng, XW; Chen, D.
(2020). Transplacental Transfer of Per- and Polyfluoroalkyl Substances (PFASs):
Differences between Preterm and Full-Term Deliveries and Associations with Placental
Transporter mRNA Expression. Environ Sci Technol 54: 5062-5070.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6505874
Li, J; Quan, X; Lei, S; Huang, Z; Wang, Q; Xu, P. (2021). PFOS inhibited normal functional
development of placenta cells via PPARy signaling. Biomedicines 9: 677.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9959491
Li, J; Quan, XJ; Chen, G; Hong, JW; Wang, Q; Xu, LL; Wang, BH; Yu, ZH; Yu, HM. (2020).
PFOS-induced placental cell growth inhibition is partially mediated by IncRNA H19
through interacting with miR-19a and miR-19b. Chemosphere 261: 127640.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833703
Li, K; Sun, J; Yang, J; Roberts, SM; Zhang, X; Cui, X; Wei, S; Ma, LQ. (2017). Molecular
Mechanisms of Perfluorooctanoate-Induced Hepatocyte Apoptosis in Mice Using
Proteomic Techniques. Environ Sci Technol 51: 11380-11389.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238518
Li, L; Li, X; Chen, X; Chen, Y; Liu, J; Chen, F; Ge, F; Ye, L; Lian, Q; Ge, RS. (2018).
Perfluorooctane sulfonate impairs rat Ley dig cell development during puberty.
Chemosphere 190: 43-53.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4241058
Li, MC. (2020). Serum Per- and Polyfluoroalkyl Substances Are Associated with Increased
Hearing Impairment: A Re-Analysis of the National Health and Nutrition Examination
Survey Data. Int J Environ Res Public Health 17.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833686
Li, N; Liu, Y; Papandonatos, GD; Calafat, AM; Eaton, CB; Kelsey, KT; Cecil, KM; Kalkwarf,
HJ; Yolton, K; Lanphear, BP; Chen, A; Braun, JM. (2021). Gestational and childhood
exposure to per- and polyfluoroalkyl substances and cardiometabolic risk at age 12 years.
Environ Int 147: 106344.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7404102
Li, R; Guo, C; Lin, X; Chan, TF; Lai, KP; Chen, J. (2020). Integrative omics analyses uncover
the mechanism underlying the immunotoxicity of perfluorooctanesulfonate in human
lymphocytes. Chemosphere 256: 127062.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833655
Li, X; Li, T; Wang, Z; Wei, J; Liu, J; Zhang, Y; Zhao, Z. (2021). Distribution of perfluorooctane
sulfonate in mice and its effect on liver lipidomic. Talanta 226: 122150.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7643501
Li, X; Ye, L; Ge, Y; Yuan, K; Zhang, Y; Liang, Y; Wei, J; Zhao, C; Lian, QQ; Zhu, X; Ge, RS.
(2016). In utero perfluorooctane sulfonate exposure causes low body weights of fetal rats:
A mechanism study. Placenta 39: 125-133.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981495
Li, Y; Barregard, L; Xu, Y; Scott, K; Pineda, D; Lindh, CH; Jakobsson, K; Fletcher, T. (2020).
Associations between perfluoroalkyl substances and serum lipids in a Swedish adult
population with contaminated drinking water. Environ Health 19: 33.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315681
7-64
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Li, Y; Cheng, Y; Xie, Z; Zeng, F. (2017). Perfluorinated alkyl substances in serum of the
southern Chinese general population and potential impact on thyroid hormones. Sci Rep
7: 43380. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3856460
Li, Y; Fletcher, T; Mucs, D; Scott, K; Lindh, CH; Tallving, P; Jakobsson, K. (2018). Half-lives
of PFOS, PFHxS and PFOA after end of exposure to contaminated drinking water. Occup
Environ Med 75: 46-51.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238434
Li, Y; Oliver, DP; Kookana, RS. (2018). A critical analysis of published data to discern the role
of soil and sediment properties in determining sorption of per and polyfluoroalkyl
substances (PFASs) [Review], Sci Total Environ 628-629: 110-120.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238331
Li, Y; Yu, N; Du, L; Shi, W; Yu, H; Song, M; Wei, S. (2020). Transplacental Transfer of Per-
and Polyfluoroalkyl Substances Identified in Paired Maternal and Cord Sera Using
Suspect and Nontarget Screening. Environ Sci Technol 54: 3407-3416.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6506038
Li, ZR; Hromchak, R; Mudipalli, A; Bloch, A. (1998). Tumor suppressor proteins as regulators
of cell differentiation. Cancer Res 58: 4282-4287.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/598342
Liang, X; Xie, G; Wu, X; Su, M; Yang, B. (2019). Effect of prenatal PFOS exposure on liver cell
function in neonatal mice. Environ Sci Pollut Res Int 26: 18240-18246.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5412467
Liao, S; Yao, W; Cheang, I; Tang, X; Yin, T; Lu, X; Zhou, Y; Zhang, H; Li, X. (2020).
Association between perfluoroalkyl acids and the prevalence of hypertension among US
adults. Ecotoxicol Environ Saf 196: 110589.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6356903
Liao, Y; Dong, S; Kiyama, R; Cai, P; Liu, L; Shen, H. (2013). Flos lonicerae extracts and
chlorogenic acid protect human umbilical vein endothelial cells from the toxic damage of
perfluorooctane sulphonate. Inflammation 36: 767-779.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1937227
Lien, GW; Huang, CC; Shiu, JS; Chen, MH; Hsieh, WS; Guo, YL; Chen, PC. (2016).
Perfluoroalkyl substances in cord blood and attention deficit/hyperactivity disorder
symptoms in seven-year-old children. Chemosphere 156: 118-127.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860112
Liew, Z; Luo, J; Nohr, EA; Bech, BH; Bossi, R; Arah, OA; Olsen, J. (2020). Maternal Plasma
Perfluoroalkyl Substances and Miscarriage: A Nested Case-Control Study in the Danish
National Birth Cohort. Environ Health Perspect 128: 47007.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6387285
Liew, Z; Ritz, B; Bach, CC; Asarnow, RF; Bech, BH; Nohr, EA; Bossi, R; Henriksen, TB;
Bonefeld-J0rgensen, EC; Olsen, J. (2018). Prenatal exposure to perfluoroalkyl substances
and iq scores at age 5; a study in the danish national birth cohort. Environ Health
Perspect 126: 067004.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079744
Liew, Z; Ritz, B; Bonefeld-J0rgensen, EC; Henriksen, TB; Nohr, EA; Bech, BH; Fei, C; Bossi,
R; von Ehrenstein, OS; Streja, E; Uldall, P; Olsen, J. (2014). Prenatal exposure to
perfluoroalkyl substances and the risk of congenital cerebral palsy in children. Am J
Epidemiol 180: 574-581.
7-65
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2852208
Liew, Z; Ritz, B; von Ehrenstein, OS; Bech, BH; Nohr, EA; Fei, C; Bossi, R; Henriksen, TB;
Bonefeld-J0rgensen, EC; Olsen, J. (2015). Attention deficit/hyperactivity disorder and
childhood autism in association with prenatal exposure to perfluoroalkyl substances: A
nested case-control study in the Danish National Birth Cohort. Environ Health Perspect
123: 367-373.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851010
Lin, CY; Chen, PC; Lin, YC; Lin, LY. (2009). Association among serum perfluoroalkyl
chemicals, glucose homeostasis, and metabolic syndrome in adolescents and adults.
Diabetes Care 32: 702-707.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290820
Lin, CY; Chen, PC; Lo, SC; Torng, PL; Sung, FC; Su, TC. (2016). The association of carotid
intima-media thickness with serum Level of perfluorinated chemicals and endothelium-
platelet microparticles in adolescents and young adults. Environ Int 94: 292-299.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981457
Lin, CY; Lee, HL; Hwang, YT; Su, TC. (2020). The association between total serum isomers of
per- and polyfluoroalkyl substances, lipid profiles, and the DNA oxidative/nitrative stress
biomarkers in middle-aged Taiwanese adults. Environ Res 182: 109064.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315756
Lin, CY; Lin, LY; Chiang, CK; Wang, WJ; Su, YN; Hung, KY; Chen, PC. (2010). Investigation
of the Associations Between Low-Dose Serum Perfluorinated Chemicals and Liver
Enzymes in US Adults. Am J Gastroenterol 105: 1354-1363.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291111
Lin, CY; Lin, LY; Wen, TW; Lien, GW; Chien, KL; Hsu, SH; Liao, CC; Sung, FC; Chen, PC;
Su, TC. (2013). Association between levels of serum perfluorooctane sulfate and carotid
artery intima-media thickness in adolescents and young adults. Int J Cardiol 168: 3309-
3316. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850967
Lin, HW; Feng, HX; Chen, L; Yuan, XJ; Tan, Z. (2020). Maternal exposure to environmental
endocrine disruptors during pregnancy is associated with pediatric germ cell tumors.
Nagoya J Med Sci 82: 323-333.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6835434
Lin, LY; Wen, LL; Su, TC; Chen, PC; Lin, CY. (2014). Negative association between serum
perfluorooctane sulfate concentration and bone mineral density in US premenopausal
women: NHANES, 2005-2008. J Clin Endocrinol Metab 99: 2173-2180.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079772
Lin, P; Cardenas, A; Hauser, R; Gold, DR; Kleinman, K; Hivert, MF; Fleisch, AF; Calafat, AM;
Webster, TF; Horton, ES; Oken, E. (2019). Per- and polyfluoroalkyl substances and
blood lipid levels in pre-diabetic adults-longitudinal analysis of the diabetes prevention
program outcomes study. Environ Int 129: 343-353.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5187597
Lin, PD; Cardenas, A; Hauser, R; Gold, DR; Kleinman, KP; Hivert, MF; Calafat, AM; Webster,
TF; Horton, ES; Oken, E. (2020). Per- and polyfluoroalkyl substances and blood pressure
in pre-diabetic adults-cross-sectional and longitudinal analyses of the diabetes prevention
program outcomes study. Environ Int 137: 105573.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6311641
Lin, TW; Chen, MK; Lin, CC; Chen, MH; Tsai, MS; Chan, DC; Hung, KY; Chen, PC. (2020).
7-66
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Association between exposure to perfluoroalkyl substances and metabolic syndrome and
related outcomes among older residents living near a Science Park in Taiwan. Int J Hyg
Environ Health 230: 113607.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6988476
Lind, DV; Priskorn, L; Lassen, TH; Nielsen, F; Kyhl, HB; Kristensen, DM; Christesen, HT;
J0rgensen, JS; Grandjean, P; Jensen, TK. (2017). Prenatal exposure to perfluoroalkyl
substances and anogenital distance at 3 months of age in a Danish mother-child cohort.
Reprod Toxicol 68: 200-206.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858512
Lind, L; Zethelius, B; Salihovic, S; van Bavel, B; Lind, PM. (2014). Circulating levels of
perfluoroalkyl substances and prevalent diabetes in the elderly. Diabetologia 57: 473-479.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2215376
Lind, PM; Salihovic, S; van Bavel, B; Lind, L. (2017). Circulating levels of perfluoroalkyl
substances (PFASs) and carotid artery atherosclerosis. Environ Res 152: 157-164.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858504
Litton Bionetics. (1979). Mutagenicity evaluation of T-2014 CoC in the ames
salmonella/microsome plate test. Final report. Beverly Hills, CA: Litton Bionetics, Inc.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10228135
Liu, CY; Chen, PC; Lien, PC; Liao, YP. (2018). Prenatal perfluorooctyl sulfonate exposure and
Alu DNA hypomethylation in cord blood. Int J Environ Res Public Health 15: 1066.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4926233
Liu, D; Liu, NY; Chen, LT; Shao, Y; Shi, XM; Zhu, DY. (2020). Perfluorooctane sulfonate
induced toxicity in embryonic stem cell-derived cardiomyocytes via inhibiting
autophagy-lysosome pathway. Toxicol In Vitro 69: 104988.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833698
Liu, G; Dhana, K; Furtado, JD; Rood, J; Zong, G; Liang, L; Qi, L; Bray, GA; Dejonge, L; Coull,
B; Grandjean, P; Sun, Q. (2018). Perfluoroalkyl substances and changes in body weight
and resting metabolic rate in response to weight-loss diets: A prospective study. PLoS
Med 15: el002502.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238396
Liu, G; Zhang, B; Hu, Y; Rood, J; Liang, L; Qi, L; Bray, GA; Dejonge, L; Coull, B; Grandjean,
P; Furtado, JD; Sun, Q. (2020). Associations of Perfluoroalkyl substances with blood
lipids and Apolipoproteins in lipoprotein subspecies: the POUNDS-lost study. Environ
Health 19: 5. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6318644
Liu, H; Chen, Q; Lei, L; Zhou, W; Huang, L; Zhang, J; Chen, D. (2018). Prenatal exposure to
perfluoroalkyl and polyfluoroalkyl substances affects leukocyte telomere length in female
newborns. Environ Pollut 235: 446-452.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4239494
Liu, H; Pan, Y; Jin, S; Li, Y; Zhao, L; Sun, X; Cui, Q; Zhang, B; Zheng, T; Xia, W; Zhou, A;
Campana, AM; Dai, J; Xu, S. (2020). Associations of per-/polyfluoroalkyl substances
with glucocorticoids and progestogens in newborns. Environ Int 140: 105636.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6569227
Liu, HS; Wen, LL; Chu, PL; Lin, CY. (2018). Association among total serum isomers of
perfluorinated chemicals, glucose homeostasis, lipid profiles, serum protein and
metabolic syndrome in adults: NHANES, 2013-2014. Environ Pollut 232: 73-79.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238514
7-67
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Liu, J; Li, J; Liu, Y; Chan, HM; Zhao, Y; Cai, Z; Wu, Y. (2011). Comparison on gestation and
lactation exposure of perfluorinated compounds for newborns. Environ Int 37: 1206-
1212. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919240
Liu, J; Liu, S; Huang, Z; Fu, Y; Fei, J; Liu, X; He, Z. (2020). Associations between the serum
levels of PFOS/PFOA and IgGN-glycosylation in adult or children. Environ Pollut 265:
114285. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833599
Liu, L; Liu, W; Song, J; Yu, H; Jin, Y; Oami, K; Sato, I; Saito, N; Tsuda, S. (2009). A
comparative study on oxidative damage and distributions of perfluorooctane sulfonate
(PFOS) in mice at different postnatal developmental stages. J Toxicol Sci 34: 245-254.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757877
Liu, M; Zhang, G; Meng, L; Han, X; Li, Y; Shi, Y; Li, A; Turyk, ME; Zhang, Q; Jiang, G.
(2021). Associations between novel and legacy per- and polyfluoroalkyl substances in
human serum and thyroid cancer: A case and healthy population in Shandong Province,
East China. Environ Sci Technol.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10176563
Liu, QS; Hao, F; Sun, Z; Long, Y; Zhou, Q; Jiang, G. (2018). Perfluorohexadecanoic acid
increases paracellular permeability in endothelial cells through the activation of plasma
kallikrein-kinin system. Chemosphere 190: 191-200.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238499
Liu, QS; Sun, Y; Qu, G; Long, Y; Zhao, X; Zhang, A; Zhou, Q; Hu, L; Jiang, G. (2017).
Structure-Dependent Hematological Effects of Per- and Polyfluoroalkyl Substances on
Activation of Plasma Kallikrein-Kinin System Cascade. Environ Sci Technol 51: 10173-
10183. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238579
Liu, X; Guo, Z; Krebs, KA; Pope, RH; Roache, NF. (2014). Concentrations and trends of
perfluorinated chemicals in potential indoor sources from 2007 through 2011 in the US.
Chemosphere 98: 51-57.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2324799
Liu, X; Zhang, L; Chen, L; Li, J; Wang, Y; Wang, J; Meng, G; Chi, M; Zhao, Y; Chen, H; Wu,
Y. (2019). Structure-based investigation on the association between perfluoroalkyl acids
exposure and both gestational diabetes mellitus and glucose homeostasis in pregnant
women. Environ Int 127: 85-93.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5881135
Liu, Y; Cao, Z; Zong, W; Liu, R. (2017). Interaction rule and mechanism of perfluoroalkyl
sulfonates containing different carbon chains with human serum albumin. RSC Adv 7:
24781-24788.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3856708
Lloyd-Jones, DM; Huffman, MD; Karmali, KN; Sanghavi, DM; Wright, JS; Pelser, C; Gulati,
M; Masoudi, FA; Goff, DC. (2017). Estimating Longitudinal Risks and Benefits From
Cardiovascular Preventive Therapies Among Medicare Patients: The Million Hearts
Longitudinal ASCVD Risk Assessment Tool: A Special Report From the American Heart
Association and American College of Cardiology [Review], Circulation 135: e793-e813.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10694407
Loccisano, AE; Campbell, JL, Jr; Andersen, ME; Clewell, HJ, III. (2011). Evaluation and
prediction of pharmacokinetics of PFOA and PFOS in the monkey and human using a
PBPK model. Regul Toxicol Pharmacol 59: 157-175.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/787186
7-68
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Loccisano, AE; Campbell, JL, Jr; Butenhoff, JL; Andersen, ME; Clewell, HJ, III. (2012).
Comparison and evaluation of pharmacokinetics of PFOA and PFOS in the adult rat
using a physiologically based pharmacokinetic model. Reprod Toxicol 33: 452-467.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1289830
Loccisano, AE; Campbell, JL; Butenhoff, JL; Andersen, ME; Clewell, HJ. (2012). Evaluation of
placental and lactational pharmacokinetics of PFOA and PFOS in the pregnant, lactating,
fetal and neonatal rat using a physiologically based pharmacokinetic model. Reprod
Toxicol 33: 468-490.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1289833
Loccisano, AE; Longnecker, MP; Campbell, JL, Jr; Andersen, ME; Clewell, HJ, III. (2013).
Development of pbpk models for pfoa and pfos for human pregnancy and lactation life
stages. J Toxicol Environ Health A 76: 25-57.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1326665
Loi, EIH; Yeung, LWY; Taniyasu, S; Lam, PKS; Kannan, K; Yamashita, N. (2011). Trophic
Magnification of Poly- and Perfluorinated Compounds in a Subtropical Food Web.
Environ Sci Technol 45: 5506-5513.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1274155
Long, M; Ghisari, M; Kjeldsen, L; Wiels0e, M; N0rgaard-Pedersen, B; Mortensen, EL;
Abdallah, MW; Bonefeld-J0rgensen, EC. (2019). Autism spectrum disorders, endocrine
disrupting compounds, and heavy metals in amniotic fluid: a case-control study.
Molecular autism 10: 1.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080602
Long, Y; Wang, Y; Ji, G; Yan, L; Hu, F; Gu, A. (2013). Neurotoxicity of perfluorooctane
sulfonate to hippocampal cells in adult mice. PLoS ONE 8: e54176.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850984
Looker, C; Luster, MI; Calafat, AM; Johnson, VJ; Burleson, GR; Burleson, FG; Fletcher, T.
(2014). Influenza vaccine response in adults exposed to perfluorooctanoate and
perfluorooctanesulfonate. Toxicol Sci 138: 76-88.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850913
Lopez-Doval, S; Salgado, R; Fernandez-Perez, B; Lafuente, A. (2015). Possible role of serotonin
and neuropeptide Y on the disruption of the reproductive axis activity by perfluorooctane
sulfonate. Toxicol Lett 233: 138-147.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2848266
Lopez-Doval, S; Salgado, R; Pereiro, N; Moyano, R; Lafuente, A. (2014). Perfluorooctane
sulfonate effects on the reproductive axis in adult male rats. Environ Res 134: 158-168.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850091
Lopez-Espinosa, MJ; Carrizosa, C; Luster, MI; Margolick, JB; Costa, O; Leonardi, GS; Fletcher,
T. (2021). Perfluoroalkyl substances and immune cell counts in adults from the Mid-Ohio
Valley (USA). Environ Int 156: 106599.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7751049
Lopez-Espinosa, MJ; Fletcher, T; Armstrong, B, en; Genser, B; Dhatariya, K; Mondal, D;
Ducatman, A; Leonardi, G. (2011). Association of Perfluorooctanoic Acid (PFOA) and
Perfluorooctane Sulfonate (PFOS) with Age of Puberty among Children Living near a
Chemical Plant. Environ Sci Technol 45: 8160-8166.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1424973
Lopez-Espinosa, MJ; Mondal, D; Armstrong, B; Bloom, MS; Fletcher, T. (2012). Thyroid
7-69
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
function and perfluoroalkyl acids in children living near a chemical plant. Environ Health
Perspect 120: 1036-1041.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291122
Lopez-Espinosa, MJ; Mondal, D; Armstrong, BG; Eskenazi, B; Fletcher, T. (2016).
Perfluoroalkyl Substances, Sex Hormones, and Insulin-like Growth Factor-1 at 6-9 Years
of Age: A Cross-Sectional Analysis within the C8 Health Project. Environ Health
Perspect 124: 1269-1275.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859832
Lorber, M; Eaglesham, GE; Hobson, P; Toms, LM; Mueller, JF; Thompson, JS. (2015). The
effect of ongoing blood loss on human serum concentrations of perfluorinated acids.
Chemosphere 118: 170-177.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851157
Louis, GM; Sapra, KJ; Barr, DB; Lu, Z; Sundaram, R. (2016). Preconception perfluoroalkyl and
polyfluoroalkyl substances and incident pregnancy loss, LIFE Study. Reprod Toxicol 65:
11-17. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858527
Louis, GMB; Peterson, CM; Chen, Z; Hediger, ML; Croughan, MS; Sundaram, R; Stanford, JB;
Fujimoto, VY; Varner, MW; Giudice, LC; Kennedy, A; Sun, L; Wu, Q; Kannan, K.
(2012). Perfluorochemicals and endometriosis: The ENDO study. Epidemiology 23: 799-
805. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1597490
Louisse, J; Rijkers, D; Stoopen, G; Janssen, A; Staats, M; Hoogenboom, R; Kersten, S;
Peijnenburg, A. (2020). Perfluorooctanoic acid (PFOA), perfluorooctane sulfonic acid
(PFOS), and perfluorononanoic acid (PFNA) increase triglyceride levels and decrease
cholesterogenic gene expression in human HepaRG liver cells. Arch Toxicol 94: 3137-
3155. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833626
Lu, L; Xu, L; Kang, T; Cheng, S. (2012). Investigation of DNA damage treated with
perfluorooctane sulfonate (PFOS) on Zr02/DDAB active nano-order film. Biosens
Bioelectron 35: 180-185.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919198
Luebker, DJ; Case, MT; York, RG; Moore, JA; Hansen, KJ; Butenhoff, JL. (2005). Two-
generation reproduction and cross-foster studies of perfluorooctanesulfonate (PFOS) in
rats. Toxicology 215: 126-148.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1276160
Luebker, DJ; Hansen, KJ; Bass, NM; Butenhoff, JL; Seacat, AM. (2002). Interactions of
fluorochemicals with rat liver fatty acid-binding protein. Toxicology 176: 175-185.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291067
Luebker, DJ; York, RG; Hansen, KJ; Moore, JA; Butenhoff, JL. (2005). Neonatal mortality from
in utero exposure to perfluorooctanesulfonate (PFOS) in Sprague-Dawley rats: dose-
response, and biochemical and pharamacokinetic parameters. Toxicology 215: 149-169.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757857
Lum, KJ; Sundaram, R; Barr, DB; Louis, TA; Buck Louis, GM. (2017). Perfluoroalkyl
Chemicals, Menstrual Cycle Length, and Fecundity: Findings from a Prospective
Pregnancy Study. Epidemiology 28: 90-98.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858516
Lv, QY; Wan, B; Guo, LH; Yang, Y; Ren, XM; Zhang, H. (2015). In vivo immunotoxicity of
perfluorooctane sulfonate in BALB/c mice: Identification of T-cell receptor and calcium-
mediated signaling pathway disruption through gene expression profiling of the spleen.
7-70
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Chem Biol Interact 240: 84-93.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981558
Lv, Z; Li, G; Li, Y; Ying, C; Chen, J; Chen, T; Wei, J; Lin, Y; Jiang, Y; Wang, Y; Shu, B; Xu,
B; Xu, S. (2013). Glucose and lipid homeostasis in adult rat is impaired by early-life
exposure to perfluorooctane sulfonate. Environ Toxicol 28: 532-542.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850947
Lv, Z; Wu, W; Ge, S; Jia, R; Lin, T; Yuan, Y; Kuang, H; Yang, B; Wu, L; Wei, J; Zhang, D.
(2018). Naringin protects against perfluorooctane sulfonate-induced liver injury by
modulating NRF2 andNF-KB in mice. Int Immunopharmacol 65: 140-147.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080395
Lyall, K; Yau, VM; Hansen, R; Kharrazi, M; Yoshida, CK; Calafat, AM; Windham, G; Croen,
LA. (2018). Prenatal maternal serum concentrations of per- and polyfluoroalkyl
substances in association with autism spectrum disorder and intellectual disability.
Environ Health Perspect 126: 017001.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4239287
Lyngs0, J; Ramlau-Hansen, CH; H0yer, BB; St0vring, H; Bonde, JP; Jonsson, BA; Lindh, CH;
Pedersen, HS; Ludwicki, JK; Zviezdai, V; Toft, G. (2014). Menstrual cycle
characteristics in fertile women from Greenland, Poland and Ukraine exposed to
perfluorinated chemicals: a cross-sectional study. Hum Reprod 29: 359-367.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850920
Ma, S; Xu, C; Ma, J; Wang, Z; Zhang, Y; Shu, Y; Mo, X. (2019). Association between
perfluoroalkyl substance concentrations and blood pressure in adolescents. Environ Pollut
254: 112971. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5413104
Maekawa, R; Ito, R; Iwasaki, Y; Saito, K; Akutsu, K; Takatori, S; Ishii, R; Kondo, F; Arai, Y;
Ohgane, J; Shiota, K; Makino, T; Sugino, N. (2017). Evidence of exposure to chemicals
and heavy metals during pregnancy in Japanese women. Reproductive Medicine and
Biology 16: 337-348.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238291
Magnuson, K; Lemeris, C; McGill, R; Sibrizzi, C; Rooney, ASK; Taylor, K; Walker, V. (2022).
Using Interactive Literature Flow Diagrams to Increase Transparency in Systematic
Reviews (unpublished work).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10442900
Maisonet, M; Calafat, AM; Marcus, M; Jaakkola, JJ; Lashen, H. (2015). Prenatal exposure to
perfluoroalkyl acids and serum testosterone concentrations at 15 years of age in female
ALSPAC study participants. Environ Health Perspect 123: 1325-1330.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859841
Makey, CM; Webster, TF; Martin, JW; Shoeib, M; Harner, T; Dix-Cooper, L; Webster, GM.
(2017). Airborne precursors predict maternal serum perfluoroalkyl acid concentrations.
Environ Sci Technol 51: 7667-7675.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860102
Malhotra, P; Gill, RK; Saksena, S; Alrefai, WA. (2020). Disturbances in Cholesterol
Homeostasis and Non-alcoholic Fatty Liver Diseases [Review], Front Med 7: 467.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10442471
Mamsen, LS; Bjorvang, RD; Mucs, D; Vinnars, MT; Papadogiannakis, N; Lindh, CH; Andersen,
CY; Damdimopoulou, P. (2019). Concentrations of perfluoroalkyl substances (PFASs) in
human embryonic and fetal organs from first, second, and third trimester pregnancies.
7-71
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Environ Int 124: 482-492.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080595
Mamsen, LS; Jonsson, BAG; Lindh, CH; Olesen, RH; Larsen, A; Ernst, E; Kelsey, TW;
Andersen, CY. (2017). Concentration of perfluorinated compounds and cotinine in
human foetal organs, placenta, and maternal plasma. Sci Total Environ 596-597: 97-105.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858487
Mancini, FR; Cano-Sancho, G; Gambaretti, J; Marchand, P; Boutron-Ruault, MC; Severi, G;
Arveux, P; Antignac, JP; Kvaskoff, M. (2020). Perfluorinated alkylated substances serum
concentration and breast cancer risk: Evidence from a nested case-control study in the
French E3N cohort. Int J Cancer 146: 917-928.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5381529
Mancini, FR; Rajaobelina, K; Praud, D; Dow, C; Antignac, JP; Kvaskoff, M; Severi, G; Bonnet,
F; Boutron-Ruault, MC; Fagherazzi, G. (2018). Nonlinear associations between dietary
exposures to perfluorooctanoic acid (PFOA) or perfluorooctane sulfonate (PFOS) and
type 2 diabetes risk in women: Findings from the E3N cohort study. Int J Hyg Environ
Health 221: 1054-1060.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079710
Manzano-Salgado, CB; Casas, M; Lopez-Espinosa, MJ; Ballester, F; Basterrechea, M; Grimalt,
JO; Jimenez, AM; Kraus, T; Schettgen, T; Sunyer, J; Vrijheid, M. (2015). Transfer of
perfluoroalkyl substances from mother to fetus in a Spanish birth cohort. Environ Res
142: 471-478.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3448674
Manzano-Salgado, CB; Casas, M; Lopez-Espinosa, MJ; Ballester, F; Iniguez, C; Martinez, D;
Romaguera, D; Fernandez-Barres, S; Santa-Marina, L; Basterretxea, M; Schettgen, T;
Valvi, D; Vioque, J; Sunyer, J; Vrijheid, M. (2017). Prenatal exposure to perfluoroalkyl
substances and cardiometabolic risk in children from the Spanish INMA birth cohort
study. Environ Health Perspect 125: 097018.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238509
Manzano-Salgado, CB; Granum, B; Lopez-Espinosa, MJ; Ballester, F; Iniguez, C; Gascon, M;
Martinez, D; Guxens, M; Basterretxea, M; Zabaleta, C; Schettgen, T; Sunyer, J; Vrijheid,
M; Casas, M. (2019). Prenatal exposure to perfluoroalkyl substances, immune-related
outcomes, and lung function in children from a Spanish birth cohort study. Int J Hyg
Environ Health 222: 945-954.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5412076
Maras, M; Vanparys, C; Muylle, F; Robbens, J; Berger, U; Barber, JL; Blust, R; De Coen, W.
(2006). Estrogen-like properties of fluorotelomer alcohols as revealed by mcf-7 breast
cancer cell proliferation. Environ Health Perspect 114: 100-105.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2952988
Marks, KJ; Jeddy, Z; Flanders, WD; Northstone, K; Fraser, A; Calafat, AM; Kato, K; Hartman,
TJ. (2019). Maternal serum concentrations of perfluoroalkyl substances during pregnancy
and gestational weight gain: The Avon Longitudinal Study of Parents and Children.
Reprod Toxicol 90: 8-14.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5381534
Martin, JW; Smithwick, MM; Braune, BM; Hoekstra, PF; Muir, DC; Mabury, SA. (2004).
Identification of long-chain perfluorinated acids in biota from the Canadian Arctic.
Environ Sci Technol 38: 373-380.
7-72
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291044
Martin, MT; Brennan, RJ; Hu, W; Ayanoglu, E; Lau, C; Ren, H; Wood, CR; Corton, JC;
Kavlock, RJ; Dix, DJ. (2007). Toxicogenomic study of triazole fungicides and
perfluoroalkyl acids in rat livers predicts toxicity and categorizes chemicals based on
mechanisms of toxicity. Toxicol Sci 97: 595-613.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/758419
Martinsson, M; Nielsen, C; Bjork, J; Rylander, L; Malmqvist, E; Lindh, C; Rignell-Hydbom, A.
(2020). Intrauterine exposure to perfluorinated compounds and overweight at age 4: A
case-control study. PLoS ONE 15: e0230137.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6311645
MassDEP. (2019). Per- and polyfluoroalkyl substances (pfas):an updated subgroup approach to
groundwater and drinking water values. Boston, MA.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6983120
Matilla-Santander, N; Valvi, D; Lopez-Espinosa, MJ; Manzano-Salgado, CB; Ballester, F;
Ibarluzea, J; Santa-Marina, L; Schettgen, T; Guxens, M; Sunyer, J; Vrijheid, M. (2017).
Exposure to Perfluoroalkyl Substances and Metabolic Outcomes in Pregnant Women:
Evidence from the Spanish INMA Birth Cohorts. Environ Health Perspect 125: 117004.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238432
Mattsson, K; Rignell-Hydbom, A; Holmberg, S; Thelin, A; Jonsson, BA; Lindh, CH; Sehlstedt,
A; Rylander, L. (2015). Levels of perfluoroalkyl substances and risk of coronary heart
disease: Findings from a population-based longitudinal study. Environ Res 142: 148-154.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859607
Mccoy, JA; Bangma, JT; Reiner, JL; Bowden, JA; Schnorr, J; Slowey, M; O'Leary, T; Guillette,
LJ; Parrott, BB. (2017). Associations between perfluorinated alkyl acids in blood and
ovarian follicular fluid and ovarian function in women undergoing assisted reproductive
treatment. Sci Total Environ 605-606: 9-17.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858475
MDH. (2020). Health Based Guidance for Water: Toxicological Summary for: Perfluorooctane
Sulfonate (PFOS).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5030007
Mecchi, MS. (1999). Salmonella-Escherichia coli/mammalian-microsome reverse mutation assay
with PFOS. Final report. (Covance Study No. 20784-0-409). Vienna, VA: Covance
Laboratories.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10228133
Melzer, D; Rice, N; Depledge, MH; Henley, WE; Galloway, TS. (2010). Association between
serum perfluorooctanoic acid (PFOA) and thyroid disease in the U.S. National Health and
Nutrition Examination Survey. Environ Health Perspect 118: 686-692.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290811
Mi, X; Yang, YQ; Zeeshan, M; Wang, ZB; Zeng, XY; Zhou, Y; Yang, BY; Hu, LW; Yu, HY;
Zeng, XW; Liu, RQ; Dong, GH. (2020). Serum levels of per- and polyfluoroalkyl
substances alternatives and blood pressure by sex status: Isomers of C8 health project in
China. Chemosphere 261: 127691.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833736
Midasch, O; Drexler, H; Hart, N; Beckmann, MW; Angerer, J. (2007). Transplacental exposure
of neonates to perfluorooctanesulfonate and perfluorooctanoate: a pilot study. Int Arch
Occup Environ Health 80: 643-648.
7-73
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290901
Minatoya, M; Itoh, S; Miyashita, C; Araki, A; Sasaki, S; Miura, R; Goudarzi, H; Iwasaki, Y;
Kishi, R. (2017). Association of prenatal exposure to perfluoroalkyl substances with cord
blood adipokines and birth size: The Hokkaido Study on environment and children's
health. Environ Res 156: 175-182.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981691
Mitro, SD; Sagiv, SK; Fleisch, AF; Jaacks, LM; Williams, PL; Rifas-Shiman, SL; Calafat, AM;
Hivert, MF; Oken, E; James-Todd, TM. (2020). Pregnancy per- and polyfluoroalkyl
substance concentrations and postpartum health in project viva: A prospective cohort. J
Clin Endocrinol Metab 105: e3415-e3426.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833625
Miura, R; Araki, A; Miyashita, C; Kobayashi, S; Kobayashi, S; Wang, SL; Chen, CH; Miyake,
K; Ishizuka, M; Iwasaki, Y; Ito, YM; Kubota, T; Kishi, R. (2018). An epigenome-wide
study of cord blood DNA methylations in relation to prenatal perfluoroalkyl substance
exposure: The Hokkaido study. Environ Int 115: 21-28.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080353
Mobacke, I; Lind, L; Dunder, L; Salihovic, S; Lind, PM. (2018). Circulating levels of
perfluoroalkyl substances and left ventricular geometry of the heart in the elderly.
Environ Int 115: 295-300.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4354163
Mogensen, UB; Grandjean, P; Heilmann, C; Nielsen, F; Weihe, P; Budtz-j0rgensen, E. (2015).
Structural equation modeling of immunotoxicity associated with exposure to
perfluorinated alkylates. Environ Health 14: 47.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981889
Mogensen, UB; Grandjean, P; Nielsen, F; Weihe, P; Budtz-j0rgensen, E. (2015). Breastfeeding
as an Exposure Pathway for Perfluorinated Alkylates. Environ Sci Technol 49: 10466-
10473. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859839
Mondal, D; Weldon, RH; Armstrong, BG; Gibson, LJ; Lopez-Espinosa, MJ; Shin, HM; Fletcher,
T. (2014). Breastfeeding: a potential excretion route for mothers and implications for
infant exposure to perfluoroalkyl acids. Environ Health Perspect 122: 187-192.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850916
Monge Brenes, AL; Curtzwiler, G; Dixon, P; Harrata, K; Talbert, J; Vorst, K. (2019). PFOA and
PFOS levels in microwave paper packaging between 2005 and 2018. Food Addit Contam
Part B Surveill 12: 191-198.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080553
Monnier, P; Martinet, C; Pontis, J; Stancheva, I; Ait-Si-Ali, S; Dandolo, L. (2013). HI9 IncRNA
controls gene expression of the Imprinted Gene Network by recruiting MBD1. Proc Natl
Acad Sci USA 110: 20693-20698.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10439067
Monroy, R; Morrison, K; Teo, K; Atkinson, S; Kubwabo, C; Stewart, B; Foster, WG. (2008).
Serum levels of perfluoroalkyl compounds in human maternal and umbilical cord blood
samples. Environ Res 108: 56-62.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2349575
Mora, AM; Fleisch, AF; Rifas-Shiman, SL; Woo Baidal, JA; Pardo, L; Webster, TF; Calafat,
AM; Ye, X; Oken, E; Sagiv, SK. (2018). Early life exposure to per- and polyfluoroalkyl
substances and mid-childhood lipid and alanine aminotransferase levels. Environ Int 111:
7-74
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
1-13. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4239224
Mora, AM; Oken, E; Rifas-Shiman, SL; Webster, TF; Gillman, MW; Calafat, AM; Ye, X; Sagiv,
SK. (2017). Prenatal exposure to perfluoroalkyl substances and adiposity in early and
mid-childhood. Environ Health Perspect 125: 467-473.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859823
Mora, S. (2016). Nonfasting for Routine Lipid Testing: From Evidence to Action. JAMA Intern
Med 176: 1005-1006.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9564968
Mordenti, J; Chen, SA; Moore, JA; Ferraiolo, BL; Green, JD. (1991). Interspecies scaling of
clearance and volume of distribution data for five therapeutic proteins. Pharm Res 8:
1351-1359. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9571900
MPCA. (2008). PFCs in Minnesota's Ambient Environment: 2008 Progress Report.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9419086
Mshaty, A; Haijima, A; Takatsuru, Y; Ninomiya, A; Yajima, H; Kokubo, M; Khairinisa, MA;
Miyazaki, W; Amano, I; Koibuchi, N. (2020). Neurotoxic effects of lactational exposure
to perfluorooctane sulfonate on learning and memory in adult male mouse. Food Chem
Toxicol 145: 111710.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833692
Murli, H. (1996). Mutagenicity test on T-6295 in an in vivo mouse micronucleus assay. Final
report. (CHV Study No. 17403-0-455). Vienna, VA: Corning Hazleton.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10228098
Murli, H. (1999). Chromosomal aberrations in human whole blood lymphocytes with PFOS.
Final report. (Covance Study No. 2784-0-499; EPA-AR-226-0131). Vienna, VA:
Covance Laboratories.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10228132
Nakayama, S; Strynar, MJ; Helfant, L; Egeghy, P; Ye, X; Lindstrom, AB. (2007). Perfluorinated
compounds in the Cape Fear Drainage Basin in North Carolina. Environ Sci Technol 41:
5271-5276. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2901973
NCBI. (2022). PubChem: Compound summary: Perfluorooctanesulfonic acid. Available online
at https://pubchem.ncbi.nlm.nih.gov/compound/Perfluorooctanesulfonic-acid (accessed
April 21,
2022).https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10411449
NCHS. (2019). Health, United States - Data Finder. 2019: Table 23. Hyattsville, MD.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10369680
Needham, LL; Grandjean, P; Heinzow, B; torgensen, PJ; Nielsen, F; Patterson, DG; Sjodin, A;
Turner, WE; Weihe, P. (2011). Partition of environmental chemicals between maternal
and fetal blood and tissues. Environ Sci Technol 45: 1121-1126.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1312781
Negri, E; Metruccio, F; Guercio, V; Tosti, L; Benfenati, E; Bonzi, R; La Vecchia, C; Moretto, A.
(2017). Exposure to PFOA and PFOS and fetal growth: a critical merging of
toxicological and epidemiological data [Review], Crit Rev Toxicol 47: 482-508.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981320
Nelson, JW; Hatch, EE; Webster, TF. (2010). Exposure to Polyfluoroalkyl Chemicals and
Cholesterol, Body Weight, and Insulin Resistance in the General US Population. Environ
Health Perspect 118: 197-202.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291110
7-75
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
New Hampshire DES. (2019). Technical background report for the june 2019 proposed
maximum contaminant levels (MCLs) and ambient groundwater quality standards
(AGQSs) for perfluorooctane sulfonic acid (PFOS), perfluorooctanoic acid (PFOA),
perfluorononanoic acid (PFNA), and perfluorohexane sulfonic acid (PFHXs).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5949029
Ngo, HT; Hetland, RB; Sabaredzovic, A; Haug, LS; Steffensen, IL. (2014). In utero exposure to
perfluorooctanoate (PFOA) or perfluorooctane sulfonate (PFOS) did not increase body
weight or intestinal tumorigenesis in multiple intestinal neoplasia (Min/+) mice. Environ
Res 132: 251-263.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850267
Ngueta, G; Longnecker, MP; Yoon, M; Ruark, CD; Clewell, HJ; Andersen, ME; Verner, MA.
(2017). Quantitative bias analysis of a reported association between perfluoroalkyl
substances (PFAS) and endometriosis: The influence of oral contraceptive use. Environ
Int 104: 118-121.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860773
Nguyen, HC; Qadura, M; Singh, KK. (2020). Role of the Fatty Acid Binding Proteins in
Cardiovascular Diseases: A Systematic Review [Review], 9.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7279862
Nguyen, TMH; Braunig, J; Thompson, K; Thompson, J; Kabiri, S; Navarro, DA; Kookana, RS;
Grimison, C; Barnes, CM; Higgins, CP; Mclaughlin, MJ; Mueller, JF. (2020). Influences
of Chemical Properties, Soil Properties, and Solution pH on Soil-Water Partitioning
Coefficients of Per- and Polyfluoroalkyl Substances (PFASs). Environ Sci Technol 54:
15883-15892.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7014622
Nian, M; Li, QQ; Bloom, M; Qian, ZM; Syberg, KM; Vaughn, MG; Wang, SQ; Wei, Q;
Zeeshan, M; Gurram, N; Chu, C; Wang, J; Tian, YP; Hu, LW; Liu, KK; Yang, BY; Liu,
RQ; Feng, D; Zeng, XW; Dong, GH. (2019). Liver function biomarkers disorder is
associated with exposure to perfluoroalkyl acids in adults: Isomers of C8 Health Project
in China. Environ Res 172: 81-88.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080307
Nielsen, C; Andersson Hall, U; Lindh, C; Ekstrom, U; Xu, Y; Li, Y; Holmang, A; Jakobsson, K.
(2020). Pregnancy-induced changes in serum concentrations of perfluoroalkyl substances
and the influence of kidney function. Environ Health 19: 80.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833687
Niu, J; Liang, H; Tian, Y; Yuan, W; Xiao, H; Hu, H; Sun, X; Song, X; Wen, S; Yang, L; Ren, Y;
Miao, M. (2019). Prenatal plasma concentrations of Perfluoroalkyl and polyfluoroalkyl
substances and neuropsychological development in children at four years of age. Environ
Health 18: 53.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5381527
NJDWQI. (2018). Health-based maximum contaminant level support document: Perfluorooctane
sulfonate (PFOS): Appendix A.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5026035
NLM. (2022). ChemlDplus: Perfluorooctane sulfonic acid. Available online at
https://chem.nlm.nih.gov/chemidplus/name/pfos 10369707
NOAA. (2017). NOAA's National Status and Trends Data. Silver Spring, MD: U.S. Department
of Commerce, National Oceanic and Atmospheric Administration, National Centers for
7-76
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Coastal Ocean Science. Retrieved from
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9638787
Noker, PE; Gorman, G. (2003). A Pharmacokinetic Study of Potassium Perfluorooctanesulfonate
in the Cynomolgus Monkey. (U.S. EPA docket AR-226-1356). Washington, D C.: U.S.
Environmental Protection Agency.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9642133
Noorlander, CW; van Leeuwen, SP; Te Biesebeek, JD; Mengelers, MJ; Zeilmaker, MJ. (2011).
Levels of perfluorinated compounds in food and dietary intake of PFOS and PFOA in the
Netherlands. J Agric Food Chem 59: 7496-7505.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919242
NRC. (2011). Review of the Environmental Protection Agency's draft IRIS assessment of
formaldehyde (pp. 1-194). Washington, DC: The National Academies Press.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/710724
NTP. (2016). NTP Monograph on Immunotoxicity Associated with Exposure to
Perfluorooctanoic Acid (PFOA) or Perfluorooctane Sulfonate (PFOS).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080063 (use
4613766)
NTP. (2016). NTP Monograph: Immunotoxicity associated with exposure to perfluorooctanoic
acid (PFOA) or perfluorooctane sulfonate (PFOS). Research Triangle Park, NC: U.S.
Department of Health and Human Services, Office of Health Assessment and
Translation, https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4613766
NTP. (2019). NTP technical report on the toxicity studies of perfluoroalkyl sulfonates
(perfluorobutane sulfonic acid, perfluorohexane sulfonate potassium salt, and
perfluorooctane sulfonic acid) administered by gavage to Sprague Dawley (Hsd:Sprague
Dawley SD) rats. (Toxicity Report 96). Research Triangle Park, NC.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5400978
NTP. (2020). NTP historical controls report all routes and vehicles Harlan Sprague-Dawley rats
[NTP], Research Triangle Park, NC.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10368689
NYSDOH. (2018). NYSDOH Drinking Water Quality Council (DWQC) - October 17, 2018.
Available online at https://www.youtube.com/watch?v=2JIXCla6cHM&feature=youtu.be
6984171
Ode, A; Kallen, K; Gustafsson, P; Rylander, L; Jonsson, BA; Olofsson, P; Ivarsson, SA; Lindh,
CH; Rignell-Hydbom, A. (2014). Fetal exposure to perfluorinated compounds and
attention deficit hyperactivity disorder in childhood. PLoS ONE 9: e95891.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851245
OECD. (2001). Test no. 416: Two-generation reproduction toxicity. In OECD guidelines for the
testing of chemicals, Section 4: Health effects. Paris, France.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3421602
OECD. (2018). Toward a new comprehensive global database of per- and polyfluoroalkyl
substances (PFASs): Summary report on updating the OECD 2007 list of per- and
polyfluoroalkyl substances (PFASs). (ENV/JM/MONO(2018)7).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5099062
Oh, JH; Kim, EY; Choi, YH; Nam, TJ. (2017). Negative regulation of ERK1/2 by PI3K is
required for the protective effects of Pyropia yezoensis peptide against perfluorooctane
sulfonate-induced endoplasmic reticulum stress. Mol Med Rep 15: 2583-2587.
7-77
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981364
Ojo, AF; Peng, C; Ng, JC. (2020). Combined effects and toxicological interactions of
perfluoroalkyl and polyfluoroalkyl substances mixtures in human liver cells (HepG2).
Environ Pollut 263: 114182.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6333436
Okada, E; Sasaki, S; Saijo, Y; Washino, N; Miyashita, C; Kobayashi, S; Konishi, K; Ito, YM;
Ito, R; Nakata, A; Iwasaki, Y; Saito, K; Nakazawa, H; Kishi, R. (2012). Prenatal
exposure to perfluorinated chemicals and relationship with allergies and infectious
diseases in infants. Environ Res 112: 118-125.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332477
Olsen, G, .W., B.,urlew, M.,.M.; Burris, J, .M.; Mandel, J, .H. (2001). A Longitudinal Analysis
of Serum Perfluorooctane Sulfonate (PFOS) and Perfluorooctanoate (PFOA) Levels in
Relation to Lipid and Hepatic Clinical Chemistry Test Results from Male Employee
Participants of the 1994/95, 1997 and 2000 Fluorochemical Medical Surveillance
Program. Final Report. (Epidemiology, 220-3W-05). St. Paul, MN: 3M Company.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10228462
Olsen, GW; Burris, JM; Burlew, MM; Mandel, JH. (2003). Epidemiologic assessment of worker
serum perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA) concentrations
and medical surveillance examinations. J Occup Environ Med 45: 260-270.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290020
Olsen, GW; Burris, JM; Ehresman, DJ; Froehlich, JW; Seacat, AM; Butenhoff, JL; Zobel, LR.
(2007). Half-life of serum elimination of
perfluorooctanesulfonate,perfluorohexanesulfonate, and perfluorooctanoate in retired
fluorochemical production workers. Environ Health Perspect 115: 1298-1305.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1429952
Olsen, GW; Ehresman, DJ; Buehrer, BD; Gibson, BA; Butenhoff, JL; Zobel, LR. (2012).
Longitudinal assessment of lipid and hepatic clinical parameters in workers involved with
the demolition of perfluoroalkyl manufacturing facilities. J Occup Environ Med 54: 974-
983. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919185
Olsen, GW; Hansen, KJ; Clemen, LA; Burris, JM; Mandel, JH. (2001). Identification of
Fluorochemicals in Human Tissue. (U.S. Environmental Protection Agency
Administrative Record 226-1030a022). 3M.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9641811
Olsen, GW; Hansen, KJ; Stevenson, LA; Burris, JM; Mandel, JH. (2003). Human donor liver
and serum concentrations of perfluorooctanesulfonate and other perfluorochemicals.
Environ Sci Technol 37: 888-891.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3005572
Olsen, GW; Lange, CC; Ellefson, ME; Mair, DC; Church, TR; Goldberg, CL; Herron, RM;
Medhdizadehkashi, Z; Nobiletti, JB; Rios, JA; Reagen, WK; Zobel, LR. (2012).
Temporal trends of perfluoroalkyl concentrations in American Red Cross adult blood
donors, 2000-2010. Environ Sci Technol 46: 6330-6338.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1578499
Omoike, OE; Pack, RP; Mamudu, HM; Liu, Y; Strasser, S; Zheng, S; Okoro, J; Wang, L. (2020).
Association between per and polyfluoroalkyl substances and markers of inflammation
and oxidative stress. Environ Res 196: 110361.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6988477
7-78
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Omoike, OE; Pack, RP; Mamudu, HM; Liu, Y; Wang, L. (2021). A cross-sectional study of the
association between perfluorinated chemical exposure and cancers related to deregulation
of estrogen receptors. Environ Res 196: 110329.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7021502
Onishchenko, N; Fischer, C; Wan Ibrahim, WN; Negri, S; Spulber, S; Cottica, D; Ceccatelli, S.
(2011). Prenatal exposure to PFOS or PFOA alters motor function in mice in a sex-
related manner. Neurotox Res 19: 452-461.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/758427
Oppi, S; Liischer, TF; Stein, S. (2019). Mouse models for atherosclerosis research- Which is my
line? [Review], Front Cardiovasc Med 6: 46.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5926372
Oshida, K; Vasani, N; Jones, C; Moore, T; Hester, S; Nesnow, S; Auerbach, S; Geter, DR;
Aleksunes, LM; Thomas, RS; Applegate, D; Klaassen, CD; Corton, JC. (2015).
Identification of chemical modulators of the constitutive activated receptor (CAR) in a
gene expression compendium. Nuclear Receptor Signaling 13: e002.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850125
Oshida, K; Vasani, N; Jones, C; Moore, T; Hester, S; Nesnow, S; Auerbach, S; Geter, DR;
Aleksunes, LM; Thomas, RS; Applegate, D; Klaassen, CD; Corton, JC. (2015).
Identification of chemical modulators of the constitutive activated receptor (CAR) in a
gene expression compendium. Nuclear Receptor Signaling 13: e002.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850125
Oshida, K; Vasani, N; Thomas, RS; Applegate, D; Rosen, M; Abbott, B; Lau, C; Guo, G;
Aleksunes, LM; Klaassen, C; Corton, JC. (2015). Identification of modulators of the
nuclear receptor peroxisome proliferator-activated receptor a (PPARa) in a mouse liver
gene expression compendium. PLoS ONE 10: eOl 12655.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5386121
Osorio-Yanez, C; Sanchez-Guerra, M; Cardenas, A; Lin, PID; Hauser, R; Gold, DR; Kleinman,
KP; Hivert, MF; Fleisch, AF; Calafat, AM; Webster, TF; Horton, ES; Oken, E. (2021).
Per- and polyfluoroalkyl substances and calcifications of the coronary and aortic arteries
in adults with prediabetes: Results from the diabetes prevention program outcomes study.
Environ Int 151: 106446.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7542684
Ouidir, M; Mendola, P; Louis, GMB; Kannan, K; Zhang, C; Tekola-Ayele, F. (2020).
Concentrations of persistent organic pollutants in maternal plasma and epigenome-wide
placental DNA methylation. Clinical Epigenetics 12: 103.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833759
Oulhote, Y; Coull, B; Bind, MA; Debes, F; Nielsen, F; Tamayo, I; Weihe, P; Grandjean, P.
(2019). Joint and independent neurotoxic effects of early life exposures to a chemical
mixture: A multi-pollutant approach combining ensemble learning and g-computation.
Environmental Epidemiology 3: e063.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6316905
Oulhote, Y; Shamim, Z; Kielsen, K; Weihe, P; Grandjean, P; Ryder, LP; Heilmann, C. (2017).
Children's white blood cell counts in relation to developmental exposures to
methylmercury and persistent organic pollutants. Reprod Toxicol 68: 207-214.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3748921
Oulhote, Y; Steuerwald, U; Debes, F; Weihe, P; Grandjean, P. (2016). Behavioral difficulties in
7-79
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
7-year old children in relation to developmental exposure to perfluorinated alkyl
substances [Review], Environ Int 97: 237-245.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3789517
Pan, Y; Cui, Q; Wang, J; Sheng, N; Jing, J; Yao, B; Dai, J. (2019). Profiles of Emerging and
Legacy Per-/Polyfluoroalkyl Substances in Matched Serum and Semen Samples: New
Implications for Human Semen Quality. Environ Health Perspect 127: 127005.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315783
Pan, Y; Zhu, Y; Zheng, T; Cui, Q; Buka, SL; Zhang, B; Guo, Y; Xia, W; Yeung, LW; Li, Y;
Zhou, A; Qiu, L; Liu, H; Jiang, M; Wu, C; Xu, S; Dai, J. (2017). Novel Chlorinated
Polyfluorinated Ether Sulfonates and Legacy Per-/Polyfluoroalkyl Substances: Placental
Transfer and Relationship with Serum Albumin and Glomerular Filtration Rate. Environ
Sci Technol 51: 634-644.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981900
Papadopoulou, E; Poothong, S; Koekkoek, J; Lucattini, L; Padilla-Sanchez, JA; Haugen, M;
Herzke, D; Valdersnes, S; Maage, A; Cousins, IT; Leonards, PEG; Smastuen Haug, L.
(2017). Estimating human exposure to perfluoroalkyl acids via solid food and drinks:
Implementation and comparison of different dietary assessment methods. Environ Res
158: 269-276.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859798
Papadopoulou, E; Stratakis, N; Basagana, X; Brantsseter, AL; Casas, M; Fossati, S;
Grazuleviciene, R; Smastuen Haug, L; Heude, B; Maitre, L; Mceachan, RRC; Robinson,
O; Roumeliotaki, T; Sabido, E; Borras, E; Urquiza, J; Vafeiadi, M; Zhao, Y; Slama, R;
Wright, J; Conti, DV; Vrijheid, M; Chatzi, L. (2021). Prenatal and postnatal exposure to
PFAS and cardiometabolic factors and inflammation status in children from six European
cohorts. Environ Int 157: 106853.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9960593
Park, MH; Gutierrez-Garcia, AK; Choudhury, M. (2019). Mono-(2-ethylhexyl) phthalate
aggravates inflammatory response via sirtuin regulation and inflammasome activation in
RAW 264.7 cells. Chem Res Toxicol 32: 935-942.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5412425
Patel, JC; Mehta, BC. (1999). Tetanus: Study of 8,697 cases. Indian J Med Sci 53: 393-401.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10176842
Peden-Adams, MM; Keller, JM; Eudaly, JG; Berger, J; Gilkeson, GS; Keil, DE. (2008).
Suppression of humoral immunity in mice following exposure to perfluorooctane
sulfonate. Toxicol Sci 104: 144-154.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1424797
Penland, TN; Cope, WG; Kwak, TJ; Strynar, MJ; Grieshaber, CA; Heise, RJ; Sessions, FW.
(2020). Trophodynamics of Per- and Polyfluoroalkyl Substances in the Food Web of a
Large Atlantic Slope River. Environ Sci Technol 54: 6800-6811.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6512132
Pennati, G; Corno, C; Costantino, ML; Bellotti, M. (2003). Umbilical flow distribution to the
liver and the ductus venosus in human fetuses during gestation: an anatomy-based
mathematical modeling. Med Eng Phys 25: 229-238.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9642023
Pennings, JLA; Jennen, DGJ; Nygaard, UC; Namork, E; Haug, LS; van Loveren, H; Granum, B.
(2016). Cord blood gene expression supports that prenatal exposure to perfluoroalkyl
7-80
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
substances causes depressed immune functionality in early childhood. J Immunotoxicol
13: 173-180. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3352001
Pereiro, N; Moyano, R; Blanco, A; Lafuente, A. (2014). Regulation of corticosterone secretion is
modified by PFOS exposure at different levels of the hypothalamic-pituitary-adrenal axis
in adult male rats. Toxicol Lett 230: 252-262.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2230732
Perez, F; Nadal, M; Navarro-Ortega, A; Fabrega, F; Domingo, JL; Barcelo, D; Farre, M. (2013).
Accumulation of perfluoroalkyl substances in human tissues. Environ Int 59: 354-362.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2325349
Petersen, MS; Hailing, J; torgensen, N; Nielsen, F; Grandjean, P; Jensen, TK; Weihe, P. (2018).
Reproductive function in a population of young Faroese men with elevated exposure to
polychlorinated biphenyls (pcbs) and perfluorinated alkylate substances (pfas). Int J
Environ Res Public Health 15: n/a.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080277
Petro, EM; D'Hollander, W; Covaci, A; Bervoets, L; Fransen, E; De Neubourg, D; De Pauw, I;
Leroy, JL; Jorssen, EP; Bols, PE. (2014). Perfluoroalkyl acid contamination of follicular
fluid and its consequence for in vitro oocyte developmental competence. Sci Total
Environ 496: 282-288.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850178
Pierozan, P; Cattani, D; Karlsson, O. (2020). Perfluorooctane sulfonate (PFOS) and
perfluorooctanoic acid (PFOA) induce epigenetic alterations and promote human breast
cell carcinogenesis in vitro. Arch Toxicol 94: 3893-3906.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833637
Pierozan, P; Karlsson, O. (2018). PFOS induces proliferation, cell-cycle progression, and
malignant phenotype in human breast epithelial cells. Arch Toxicol 92: 705-716.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238459
Pilkerton, CS; Hobbs, GR; Lilly, C; Knox, SS. (2018). Rubella immunity and serum
perfluoroalkyl substances: Sex and analytic strategy. PLoS ONE 13: e0203330.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080265
Pitter, G; Zare Jeddi, M; Barbieri, G; Gion, M; Fabricio, ASC; Dapra, F; Russo, F; Fletcher, T;
Canova, C. (2020). Perfluoroalkyl substances are associated with elevated blood pressure
and hypertension in highly exposed young adults. Environ Health 19: 102.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6988479
Pizzurro, DM; Seeley, M; Kerper, LE; Beck, BD. (2019). Interspecies differences in
perfluoroalkyl substances (PFAS) toxicokinetics and application to health-based criteria
[Review], Regul Toxicol Pharmacol 106: 239-250.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5387175
Podder, A, .; Sadmani, A, .; Reinhart, D, .; Chang, N, . B.; Goel, R, . (2021). Per and poly-
fluoroalkyl substances (PFAS) as a contaminant of emerging concern in surface water: A
transboundary review of their occurrences and toxicity effects. J Hazard Mater 419:
126361. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9640865
Poothong, S; Padilla-Sanchez, JA; Papadopoulou, E; Giovanoulis, G; Thomsen, C; Haug, LS.
(2019). Hand Wipes: A Useful Tool for Assessing Human Exposure to Poly- and
Perfluoroalkyl Substances (PFASs) through Hand-to-Mouth and Dermal Contacts.
Environ Sci Technol 53: 1985-1993.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080584
7-81
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Poothong, S; Papadopoulou, E; Padilla-Sanchez, JA; Thomsen, C; Haug, LS. (2020). Multiple
pathways of human exposure to poly- and perfluoroalkyl substances (PFASs): From
external exposure to human blood. Environ Int 134: 105244.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6311690
Poothong, S; Thomsen, C; Padilla-Sanchez, JA; Papadopoulou, E; Haug, LS. (2017).
Distribution of novel and well-known poly- and perfluoroalkyl substances (PFASs) in
human serum, plasma, and whole blood. Environ Sci Technol 51: 13388-13396.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4239163
Popovic, M; Zaja, R; Fent, K; Smital, T. (2014). Interaction of environmental contaminants with
zebrafish organic anion transporting polypeptide, Oatpldl (Slcoldl). Toxicol Appl
Pharmacol 280: 149-158.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2713517
Porpora, MG; Lucchini, R; Abballe, A; Ingelido, AM; Valentini, S; Fuggetta, E; Cardi, V;
Ticino, A; Marra, V; Fulgenzi, AR; Felip, ED. (2013). Placental transfer of persistent
organic pollutants: a preliminary study on mother-newborn pairs. Int J Environ Res
Public Health 10: 699-711.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2150057
Portier, K; Tolson, JK; Roberts, SM. (2007). Body weight distributions for risk assessment. Risk
Anal 27: 11-26.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/192981
Powley, CR; George, SW; Russell, MH; Hoke, RA; Buck, RC. (2008). Polyfluorinated
chemicals in a spatially and temporally integrated food web in the Western Arctic.
Chemosphere 70: 664-672.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332751
Predieri, B; Iughetti, L; Guerranti, C; Bruzzi, P; Perra, G; Focardi, SE. (2015). High Levels of
Perfluorooctane Sulfonate in Children at the Onset of Diabetes. International Journal of
Endocrinology 2015: 234358.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3889874
Preston, EV; Rifas-Shiman, SL; Hivert, MF; Zota, AR; Sagiv, SK; Calafat, AM; Oken, E;
James-Todd, T. (2020). Associations of per- and polyfluoroalkyl substances (PFAS) with
glucose tolerance during pregnancy in project viva. J Clin Endocrinol Metab 105: E2864-
E2876. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833657
Preston, EV; Webster, TF; Oken, E; Claus Henn, B; Mcclean, MD; Rifas-Shiman, SL; Pearce,
EN; Braverman, LE; Calafat, AM; Ye, X; Sagiv, SK. (2018). Maternal plasma per- and
polyfluoroalkyl substance concentrations in early pregnancy and maternal and neonatal
thyroid function in a prospective birth cohort: Project Viva (USA). Environ Health
Perspect 126: 027013.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4241056
Pritchard, JA. (1965). Changes in the blood volume during pregnancy and delivery [Review],
Anesthesiology 26: 393-399.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9641812
Puttige Ramesh, N; Arora, M; Braun, JM. (2019). Cross-sectional study of the association
between serum perfluorinated alkyl acid concentrations and dental caries among US
adolescents (NHANES 1999-2012). BMJ Open 9: e024189.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080517
Qazi, MR; Abedi, MR; Nelson, BD; Depierre, JW; Abedi-Valugerdi, M. (2010). Dietary
7-82
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
exposure to perfluorooctanoate or perfluorooctane sulfonate induces hypertrophy in
centrilobular hepatocytes and alters the hepatic immune status in mice. Int
Immunopharmacol 10: 1420-1427.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1276154
Qazi, MR; Bogdanska, J; Butenhoff, JL; Nelson, BD; Depierre, JW; Abedi-Valugerdi, M.
(2009). High-dose, short-term exposure of mice to perfluorooctanesulfonate (PFOS) or
perfluorooctanoate (PFOA) affects the number of circulating neutrophils differently, but
enhances the inflammatory responses of macrophages to lipopolysaccharide (LPS) in a
similar fashion. Toxicology 262: 207-214.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1937259
Qazi, MR; Nelson, BD; Depierre, JW; Abedi-Valugerdi, M. (2012). High-dose dietary exposure
of mice to perfluorooctanoate or perfluorooctane sulfonate exerts toxic effects on
myeloid and B-lymphoid cells in the bone marrow and these effects are partially
dependent on reduced food consumption. Food Chem Toxicol 50: 2955-2963.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1937236
Qazi, MR; Xia, Z; Bogdanska, J; Chang, SC; Ehresman, DJ; Butenhoff, JL; Nelson, BD;
Depierre, JW; Abedi-Valugerdi, M. (2009). The atrophy and changes in the cellular
compositions of the thymus and spleen observed in mice subjected to short-term exposure
to perfluorooctanesulfonate are high-dose phenomena mediated in part by peroxisome
proliferator-activated receptor-alpha (PPARalpha). Toxicology 260: 68-76.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1937260
Qian, Y; Ducatman, A; Ward, R; Leonard, S; Bukowski, V; Lan Guo, N; Shi, X; Vallyathan, V;
Castranova, V. (2010). Perfluorooctane sulfonate (PFOS) induces reactive oxygen
species (ROS) production in human microvascular endothelial cells: role in endothelial
permeability. J Toxicol Environ Health A 73: 819-836.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919301
Qin, XD; Qian, Z; Vaughn, MG; Huang, J; Ward, P; Zeng, XW; Zhou, Y; Zhu, Y; Yuan, P; Li,
M; Bai, Z; Paul, G; Hao, YT; Chen, W; Chen, PC; Dong, GH; Lee, YL. (2016). Positive
associations of serum perfluoroalkyl substances with uric acid and hyperuricemia in
children from Taiwan. Environ Pollut 212: 519-524.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981721
Qin, XD; Qian, ZM; Dharmage, SC; Perret, J; Geiger, SD; Rigdon, SE; Howard, S; Zeng, XW;
Hu, LW; Yang, BY; Zhou, Y; Li, M; Xu, SL; Bao, WW; Zhang, YZ; Yuan, P; Wang, J;
Zhang, C; Tian, YP; Nian, M; Xiao, X; Chen, W; Lee, YL; Dong, GH. (2017).
Association of perfluoroalkyl substances exposure with impaired lung function in
children. Environ Res 155: 15-21.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3869265
Qiu, L; Qian, Y; Liu, Z; Wang, C; Qu, J; Wang, X; Wang, S. (2016). Perfluorooctane sulfonate
(PFOS) disrupts blood-testis barrier by down-regulating junction proteins via p38
MAPK/ATF2/MMP9 signaling pathway. Toxicology 373: 1-12.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981408
Qiu, L; Wang, H; Dong, T; Huang, J; Li, T; Ren, H; Wang, X; Qu, J; Wang, S. (2020).
Perfluorooctane sulfonate (PFOS) disrupts testosterone biosynthesis via
CREB/CRTC2/StAR signaling pathway in Leydig cells. Toxicology 449: 152663.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7276729
Qiu, L; Zhang, X; Zhang, X; Zhang, Y; Gu, J; Chen, M; Zhang, Z; Wang, X; Wang, SL. (2013).
7-83
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Sertoli cell is a potential target for perfluorooctane sulfonate-induced reproductive
dysfunction in male mice. Toxicol Sci 135: 229-240.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850956
Qu, A; Cao, T; Li, Z; Wang, W; Liu, R; Wang, X; Nie, Y; Sun, S; Liu, X; Zhang, X. (2021). The
association between maternal perfluoroalkyl substances exposure and early attention
deficit hyperactivity disorder in children: a systematic review and meta-analysis. Environ
Sci Pollut Res Int 28: 67066-67081.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9959569
Qu, JH; Lu, CC; Xu, C; Chen, G; Qiu, LL; Jiang, JK; Ben, S; Wang, YB; Gu, AH; Wang, XR.
(2016). Perfluorooctane sulfonate-induced testicular toxicity and differential testicular
expression of estrogen receptor in male mice. Environ Toxicol Pharmacol 45: 150-157.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981454
Quaak, I; de Cock, M; de Boer, M; Lamoree, M; Leonards, P; van de Bor, M. (2016). Prenatal
Exposure to Perfluoroalkyl Substances and Behavioral Development in Children. Int J
Environ Res Public Health 13.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981464
Rahman, ML; Zhang, C; Smarr, MM; Lee, S; Honda, M; Kannan, K; Tekola-Ayele, F; Buck
Louis, GM. (2019). Persistent organic pollutants and gestational diabetes: A multi-center
prospective cohort study of healthy US women. Environ Int 124: 249-258.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5024206
Rainieri, S; Conlledo, N; Langerholc, T; Madorran, E; Sala, M; Barranco, A. (2017). Toxic
effects of perfluorinated compounds at human cellular level and on a model vertebrate.
Food Chem Toxicol 104: 14-25.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860104
Rantakokko, P; Mannisto, V; Airaksinen, R; Koponen, J; Viluksela, M; Kiviranta, H;
Pihlajamaki, J. (2015). Persistent organic pollutants and non-alcoholic fatty liver disease
in morbidly obese patients: A cohort study. Environ Health 14: 79.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3351439
Reardon, AJF; Khodayari Moez, E; Dinu, I; Goruk, S; Field, CJ; Kinniburgh, DW; Macdonald,
AM; Martin, JW; Study, A. (2019). Longitudinal analysis reveals early-pregnancy
associations between perfluoroalkyl sulfonates and thyroid hormone status in a Canadian
prospective birth cohort. Environ Int 129: 389-399.
https://hero.epa.gov/hero/index.cfim/reference/details/reference_id/5412435
Remucal, CK. (2019). Spatial and temporal variability of perfluoroalkyl substances in the
Laurentian Great Lakes [Review], Environ Sci Process Impacts 21: 1816-1834.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5413103
Ren, Y; Jin, L; Yang, F; Liang, H; Zhang, Z; Du, J; Song, X; Miao, M; Yuan, W. (2020).
Concentrations of perfluoroalkyl and polyfluoroalkyl substances and blood glucose in
pregnant women. Environ Health 19: 88.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833646
Reyes, L; Manalich, R. (2005). Long-term consequences of low birth weight [Review], Kidney
Int Suppl 68: S107-S111.
https://hero.epa.gov/hero/index.cfim/reference/details/reference_id/1065677
Robinson, MW; Harmon, C; O'Farrelly, C. (2016). Liver immunology and its role in
inflammation and homeostasis [Review], Cell Mol Immunol 13: 267-276.
https://hero.epa.gov/hero/index.cfim/reference/details/reference_id/10284350
7-84
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Rogers, JM; Ellis-Hutchings, RG; Grey, BE; Zucker, RM; Norwood, J, Jr; Grace, CE; Gordon,
CJ; Lau, C. (2014). Elevated blood pressure in offspring of rats exposed to diverse
chemicals during pregnancy. Toxicol Sci 137: 436-446.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2149155
Romano, ME; Xu, Y; Calafat, AM; Yolton, K; Chen, A; Webster, GM; Eliot, MN; Howard, CR;
Lanphear, BP; Braun, JM. (2016). Maternal serum perfluoroalkyl substances during
pregnancy and duration of breastfeeding. Environ Res 149: 239-246.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981728
Rooney, JP; Oshida, K; Kumar, R; Baldwin, WS; Corton, JC. (2019). Chemical Activation of the
Constitutive Androstane Receptor Leads to Activation of Oxidant-Induced Nrf2. Toxicol
Sci 167: 172-189.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6988236
Rosen, MB; Das, KP; Rooney, J; Abbott, B; Lau, C; Corton, JC. (2017). PPARa-independent
transcriptional targets of perfluoroalkyl acids revealed by transcript profiling. Toxicology
387: 95-107. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859803
Rosen, MB; Das, KP; Wood, CR; Wolf, CJ; Abbott, BD; Lau, C. (2013). Evaluation of
perfluoroalkyl acid activity using primary mouse and human hepatocytes. Toxicology
308: 129-137.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919147
Rosen, MB; Schmid, JE; Das, KP; Wood, CR; Zehr, RD; Lau, C. (2009). Gene expression
profiling in the liver and lung of perfluorooctane sulfonate-exposed mouse fetuses:
comparison to changes induced by exposure to perfluorooctanoic acid. Reprod Toxicol
27: 278-288. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919338
Rosen, MB; Schmid, JR; Corton, JC; Zehr, RD; Das, KP; Abbott, BD; Lau, C. (2010). Gene
expression profiling in wild-type and PPAR alpha-null mice exposed to perfluorooctane
sulfonate reveals PPAR alpha-independent effects. PPAR Research 2010: 794739.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1274165
Rosenmai, AK; Ahrens, L; le Godec, T; Lundqvist, J; Oskarsson, A. (2018). Relationship
between peroxisome proliferator-activated receptor alpha activity and cellular
concentration of 14 perfluoroalkyl substances in HepG2 cells. J Appl Toxicol 38: 219-
226. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4220319
Rotander, A; Toms, LM; Aylward, L; Kay, M; Mueller, JF. (2015). Elevated levels of PFOS and
PFHxS in firefighters exposed to aqueous film forming foam (AFFF). Environ Int 82: 28-
34. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859842
Ruark, CD; Song, G; Yoon, M; Verner, MA; Andersen, ME; Clewell, HJ; Longnecker, MP.
(2017). Quantitative bias analysis for epidemiological associations of perfluoroalkyl
substance serum concentrations and early onset of menopause. Environ Int 99: 245-254.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981395
Ruffle, B; Vedagiri, U; Bogdan, D; Maier, M; Schwach, C; Murphy-Hagan, C. (2020).
Perfluoroalkyl Substances in U.S. market basket fish and shellfish. Environ Res 190:
9932-9932. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833737
Rusch, GM. (1979). An Acute Inhalation Toxicity Study of T-2307 CoC in the Rat. (3M-EPA-
0029829). Rusch, GM.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7561179
Russell, MH; Waterland, RL; Wong, F. (2015). Calculation of chemical elimination half-life
from blood with an ongoing exposure source: The example of perfluorooctanoic acid
7-85
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
(PFOA). Chemosphere 129: 210-216.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851185
Rylander, L; Lindh, CH; Hansson, S. R.; Broberg, K; Kallen, K. (2020). Per- and polyfluoroalkyl
substances in early pregnancy and risk for preeclampsia: a case-control study in Southern
Sweden. Toxics 8: 43.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833607
Sagiv, SK; Rifas-Shiman, SL; Fleisch, AF; Webster, TF; Calafat, AM; Ye, X; Gillman, MW;
Oken, E. (2018). Early Pregnancy Perfluoroalkyl Substance Plasma Concentrations and
Birth Outcomes in Project Viva: Confounded by Pregnancy Hemodynamics? Am J
Epidemiol 187: 793-802.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238410
Salgado, R; Lopez-Doval, S; Pereiro, N; Lafuente, A. (2016). Perfluorooctane sulfonate (PFOS)
exposure could modify the dopaminergic system in several limbic brain regions. Toxicol
Lett 240: 226-235.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3179088
Salgado, R; Pereiro, N; Lopez-Doval, S; Lafuente, A. (2015). Initial study on the possible
mechanisms involved in the effects of high doses of perfluorooctane sulfonate (PFOS) on
prolactin secretion. Food Chem Toxicol 83: 10-16.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981583
Salgado-Freiria, R; Lopez-Doval, S; Lafuente, A. (2018). Perfluorooctane sulfonate (PFOS) can
alter the hypothalamic-pituitary-adrenal (HPA) axis activity by modifying CRF1 and
glucocorticoid receptors. Toxicol Lett 295: 1-9.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079767
Salihovic, S; Stubleski, J; Karrman, A; Larsson, A; Fall, T; Lind, L; Lind, PM. (2018). Changes
in markers of liver function in relation to changes in perfluoroalkyl substances - A
longitudinal study. Environ Int 117: 196-203.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5083555
Salvalaglio, M; Muscionico, I; Cavallotti, C. (2010). Determination of energies and sites of
binding of PFOA and PFOS to human serum albumin. J Phys Chem B 114: 14860-
14874. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919252
Sanchez Garcia, D; Sjodin, M; Hellstrandh, M; Norinder, U; Nikiforova, V; Lindberg, J;
Wincent, E; Bergman, A; Cotgreave, I; Munic Kos, V. (2018). Cellular accumulation and
lipid binding of perfluorinated alkylated substances (PFASs) - A comparison with
lysosomotropic drugs. Chem Biol Interact 281: 1-10.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4234856
Sasaki, K; Harada, K; Saito, N; Tsutsui, T; Nakanishi, S; Tsuzuki, H; Koizumi, A. (2003).
Impact of airborne perfluorooctane sulfonate on the human body burden and the
ecological system. Bull Environ Contam Toxicol 71: 408-413.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5081390
Schaider, LA; Balan, SA; Blum, A; Andrews, DQ; Strynar, MJ; Dickinson, ME; Lunderberg,
DM; Lang, JR; Peaslee, GF. (2017). Fluorinated compounds in US fast food packaging.
Environ Sci Technol Lett 4: 105-111.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981864
Schecter, A; Colacino, J; Haffner, D; Patel, K; Opel, M; Papke, O; Birnbaum, L. (2010).
Perfluorinated compounds, polychlorinated biphenyls, and organochlorine pesticide
contamination in composite food samples from Dallas, Texas, USA. Environ Health
7-86
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Perspect 118: 796-802.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/729962
Schiller, JS; Lucas, JW; Ward, BW; Peregoy, JA. (2012). Summary health statistics for U.S.
adults: National Health Interview Survey, 2010. (DHHS Publication No. (PHS) 2012-
1580). Hyattsville, MD: National Center for Health Statistics.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1798736
Schlegel, R, MacGregor, J.,.T. (1984). The persistence of micronucleated erythrocytes in the
peripheral circulation of normal and splenectomized Fischer 344 rats: Implications for
cytogenetic screening. MutatRes 127: 169-174.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10368697
Schreder, E; Dickman, J. (2018). Take Out Toxics: PFAS Chemicals in Food Packaging.
Schreder, Erika Schreder; Dickman, Jennifer.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9419077
Schumann, G; Bonora, R; Ceriotti, F; Ferard, G; Ferrero, CA; Franck, PFH; Gella, FJ; Hoelzel,
W; J0rgensen, PJ; Kanno, T; Kessner, A; Klauke, R; Kristiansen, N; Lessinger, JM;
Linsinger, TPJ; Misaki, H; Panteghini, M; Pauwels, J; Schiele, F; Schimmel, HG. (2002).
IFCC Primary Reference Procedures for the Measurement of Catalytic Activity
Concentrations of Enzymes at 37°C. Part 4. Reference Procedure for the Measurement of
Catalytic Concentration of Alanine Aminotransferase. 40: 718-724.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10369681
Scinicariello, F; Buser, MC; Balluz, L; Gehle, K; Murray, HE; Abadin, HG; Attanasio, R.
(2020). Perfluoroalkyl acids, hyperuricemia and gout in adults: Analyses of NHANES
2009-2014. Chemosphere 259: 127446.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833670
Seacat, AM; Thomford, PJ; Hansen, KJ; Clemen, LA; Eldridge, SR; Elcombe, CR; Butenhoff,
JL. (2003). Sub-chronic dietary toxicity of potassium perfluorooctanesulfonate in rats
[Erratum], Toxicology 183: 117-131.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290852
Seacat, AM; Thomford, PJ; Hansen, KJ; Olsen, GW; Case, MT; Butenhoff, JL. (2002).
Subchronic toxicity studies on perfluorooctanesulfonate potassium salt in cynomolgus
monkeys. Toxicol Sci 68: 249-264.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757853
Selgrade, MK. (2007). Immunotoxicity: The risk is real [Review], Toxicol Sci 100: 328-332.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/736210
Seo, SH; Son, MH; Choi, SD; Lee, DH; Chang, YS. (2018). Influence of exposure to
perfluoroalkyl substances (PFASs) on the Korean general population: 10-year trend and
health effects. Environ Int 113: 149-161.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238334
Shah, D. (2009). Healthy worker effect phenomenon. Indian J Occup Environ Med 13: 77-79.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9570930
Shah-Kulkarni, S; Kim, BM; Hong, YC; Kim, HS; Kwon, EJ; Park, H; Kim, YJ; Ha, EH. (2016).
Prenatal exposure to perfluorinated compounds affects thyroid hormone levels in
newborn girls. Environ Int 94: 607-613.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859821
Shan, G; Wang, Z, hi; Zhou, L; Du, P, in; Luo, X; Wu, Q; Zhu, L. (2016). Impacts of daily
intakes on the isomeric profiles of perfluoroalkyl substances (PFASs) in human serum.
7-87
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Environ Int 89-90: 62-70.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3360127
Shan, G; Ye, M; Zhu, B; Zhu, L. (2013). Enhanced cytotoxicity of pentachlorophenol by
perfluorooctane sulfonate or perfluorooctanoic acid in HepG2 cells. Chemosphere 93:
2101-2107. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850950
Shankar, A; Xiao, J; Ducatman, A. (2011). Perfluoroalkyl chemicals and chronic kidney disease
in US adults. Am J Epidemiol 174: 893-900.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919232
Shapiro, GD; Dodds, L; Arbuckle, TE; Ashley-Martin, J; Ettinger, AS; Fisher, M; Taback, S;
Bouchard, MF; Monnier, P; Dallaire, R; Morisset, AS; Fraser, W. (2016). Exposure to
organophosphorus and organochlorine pesticides, perfluoroalkyl substances, and
poly chlorinated biphenyls in pregnancy and the association with impaired glucose
tolerance and gestational diabetes mellitus: The MIREC Study. Environ Res 147: 71-81.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3201206
Shearer, JJ; Callahan, CL; Calafat, AM; Huang, WY; Jones, RR; Sabbisetti, VS; Freedman, ND;
Sampson, JN; Silverman, DT; Purdue, MP; Hofmann, JN, (2021). Serum concentrations
of per- and polyfluoroalkyl substances and risk of renal cell carcinoma. J Natl Cancer Inst
113: 580-587.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7161466
Shen, M; Xiao, Y; Huang, Y; Jing, D; Su, J; Luo, D; Duan, Y; Xiao, S; Li, J; Chen, X. (2022).
Perfluoroalkyl substances are linked to incident chronic spontaneous urticaria: A nested
case-control study. Chemosphere 287: 132358.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10176753
Sheng, N; Cui, R; Wang, J; Guo, Y; Wang, J; Dai, J. (2018). Cytotoxicity of novel fluorinated
alternatives to long-chain perfluoroalkyl substances to human liver cell line and their
binding capacity to human liver fatty acid binding protein. Arch Toxicol 92: 359-369.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4199441
Sheng, N; Wang, J; Guo, Y; Wang, J; Dai, J. (2020). Interactions of perfluorooctanesulfonate
and 6:2 chlorinated polyfluorinated ether sulfonate with human serum albumin: A
comparative study. Chem Res Toxicol 33: 1478-1486.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6565171
Shih, YH; Blomberg, AJ; Bind, MA; Holm, D; Nielsen, F; Heilmann, C; Weihe, P; Grandjean, P.
(2021). Serum vaccine antibody concentrations in adults exposed to per- and
polyfluoroalkyl substances: A birth cohort in the Faroe Islands. J Immunotoxicol 18: 85-
92. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9959487
Shin, HM; Bennett, DH; Calafat, AM; Tancredi, D; Hertz-Picciotto, I. (2020). Modeled prenatal
exposure to per- and polyfluoroalkyl substances in association with child autism
spectrum disorder: A case-control study. Environ Res 186: 109514.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6507470
Shipley, JM; Hurst, CH; Tanaka, SS; Deroos, FL; Butenhoff, JL; Seacat, AM; Waxman, DJ.
(2004). Trans-activation of PPARalpha and induction of PPARalpha target genes by
perfluorooctane-based chemicals. Toxicol Sci 80: 151-160.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2990378
Shoeib, M; Harner, T; M Webster, G; Lee, SC. (2011). Indoor sources of poly- and
perfluorinated compounds (PFCS) in Vancouver, Canada: implications for human
exposure. Environ Sci Technol 45: 7999-8005.
7-88
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1082300
Shrestha, S; Bloom, MS; Yucel, R; Seegal, RF; Rej, R; Mccaffrey, RJ; Wu, Q; Kannan, K;
Fitzgerald, EF. (2017). Perfluoroalkyl substances, thyroid hormones, and
neuropsychological status in older adults. Int J Hyg Environ Health 220: 679-685.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981382
Sibai, BM; Frangieh, A. (1995). Maternal adaptation to pregnancy. Curr Opin Obstet Gynecol 7:
420-426. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1101373
Simmon, VF. (1978). In-vitro microbiological mutagenicity assays of 3M Company compounds
T-2247 CoC and T-2248 CoC. Final report. (SRI Project LSC-442-016). Menlo Park,
CA: SRI International.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10228131
Singer, AB; Whitworth, KW; Haug, LS; Sabaredzovic, A; Impinen, A; Papadopoulou, E;
Longnecker, MP. (2018). Menstrual cycle characteristics as determinants of plasma
concentrations of perfluoroalkyl substances (PFASs) in the Norwegian Mother and Child
Cohort (MoBa study). Environ Res 166: 78-85.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079732
Sinisalu, L; Sen, P; Salihovic, S; Virtanen, SM; Hyoty, H; Ilonen, J; Toppari, J; Veijola, R;
Oresic, M; Knip, M; Hyotylainen, T. (2020). Early-life exposure to perfluorinated alkyl
substances modulates lipid metabolism in progression to celiac disease. Environ Res 188:
109864. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7211554
Sinisalu, L; Yeung, LWY; Wang, J; Pan, Y; Dai, J; Hyotylainen, T. (2021). Prenatal exposure to
poly-/per-fluoroalkyl substances is associated with alteration of lipid profiles in cord-
blood. Metabolomics 17: 103.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9959547
Skogheim, TS; Villanger, GD; Weyde, KVF; Engel, SM; Suren, P; 0ie, MG; Skogan, AH; Biele,
G; Zeiner, P; 0vergaard, KR; Haug, LS; Sabaredzovic, A; Aase, H. (2019). Prenatal
exposure to perfluoroalkyl substances and associations with symptoms of attention-
deficit/hyperactivity disorder and cognitive functions in preschool children. Int J Hyg
Environ Health 223: 80-92.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5918847
Skuladottir, M; Ramel, A; Rytter, D; Haug, LS; Sabaredzovic, A; Bech, BH; Henriksen, TB;
Olsen, SF; Halldorsson, TI. (2015). Examining confounding by diet in the association
between perfluoroalkyl acids and serum cholesterol in pregnancy. Environ Res 143: 33-
38. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3749113
Smit, LA; Lenters, V; H0yer, BB; Lindh, CH; Pedersen, HS; Liermontova, I; Jonsson, BA;
Piersma, AH; Bonde, JP; Toft, G; Vermeulen, R; Heederik, D. (2015). Prenatal exposure
to environmental chemical contaminants and asthma and eczema in school-age children.
Allergy 70: 653-660.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2823268
Smith, MT; Guyton, KZ; Gibbons, CF; Fritz, JM; Portier, CJ; Rusyn, I; DeMarini, DM;
Caldwell, JC; Kavlock, RJ; Lambert, PF; Hecht, SS; Bucher, JR; Stewart, BW; Baan,
RA; Cogliano, VJ; Straif, K. (2016). Key characteristics of carcinogens as a basis for
organizing data on mechanisms of carcinogenesis [Review], Environ Health Perspect
124: 713-721.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3160486
Smith, MT; Guyton, KZ; Kleinstreuer, N; Borrel, A; Cardenas, A; Chiu, WA; Felsher, DW;
7-89
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Gibbons, CF; Goodson, WH; Houck, KA; Kane, AB; La Merrill, MA; Lebrec, H; Lowe,
L; Mchale, CM; Minocherhomji, S; Rieswijk, L; Sandy, MS; Sone, H; Wang, A; Zhang,
L; Zeise, L; Fielden, M. (2020). The Key Characteristics of Carcinogens: Relationship to
the Hallmarks of Cancer, Relevant Biomarkers, and Assays to Measure Them [Editorial],
Cancer Epidemiol Biomarkers Prev.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6956443
Song, M, i-Kyung; Cho, Y, oon; Jeong, S, eung-Chan; Ryu, J, ae-Chun. (2016). Analysis of gene
expression changes in relation to hepatotoxicity induced by perfluorinated chemicals in a
human hepatoma cell line. Toxicol Environ Health Sci 8: 114-127.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9959776
Song, X; Tang, S; Zhu, H; Chen, Z; Zang, Z; Zhang, Y; Niu, X; Wang, X; Yin, H; Zeng, F; He,
C. (2018). Biomonitoring PFAAs in blood and semen samples: Investigation of a
potential link between PFAAs exposure and semen mobility in China. Environ Int 113:
50-54. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4220306
Sonkar, R; Kay, MK; Choudhury, M. (2019). PFOS modulates interactive epigenetic regulation
in first-trimester human trophoblast cell line HTR-8/SVneo. Chem Res Toxicol 32: 2016-
2027. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5918797
S0rli, JB; Lag, M; Ekeren, L; Perez-Gil, J; Haug, LS; Da Silva, E; Matrod, MN; Giitzkow, KB;
Lindeman, B. (2020). Per- and polyfluoroalkyl substances (PFASs) modify lung
surfactant function and pro-inflammatory responses in human bronchial epithelial cells.
Toxicol In Vitro 62: 104656.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5918817
Spliethoff, HM; Tao, L; Shaver, SM; Aldous, KM; Pass, KA; Kannan, K; Eadon, GA. (2008).
Use of newborn screening program blood spots for exposure assessment: declining levels
of perluorinated compounds in New York State infants. Environ Sci Technol 42: 5361-
5367. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919368
Spratlen, MJ; Perera, FP; Lederman, SA; Rauh, VA; Robinson, M; Kannan, K; Trasande, L;
Herbstman, J. (2020). The association between prenatal exposure to perfluoroalkyl
substances and childhood neurodevelopment. Environ Pollut 263: 114444.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6364693
Spratlen, MJ; Perera, FP; Lederman, SA; Robinson, M; Kannan, K; Herbstman, J; Trasande, L.
(2020). The association between perfluoroalkyl substances and lipids in cord blood. J
Clin Endocrinol Metab 105: 43-54.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5915332
Starling, AP; Adgate, JL; Hamman, RF; Kechris, K; Calafat, AM; Ye, X; Dabelea, D. (2017).
Perfluoroalkyl substances during pregnancy and offspring weight and adiposity at birth:
Examining mediation by maternal fasting glucose in the healthy start study. Environ
Health Perspect 125: 067016.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858473
Starling, AP; Engel, SM; Richardson, DB; Baird, DD; Haug, LS; Stuebe, AM; Klungsoyr, K;
Harmon, Q; Becher, G; Thomsen, C; Sabaredzovic, A; Eggesbo, M; Hoppin, JA; Travlos,
GS; Wilson, RE; Trogstad, LI; Magnus, P, er; Longnecker, MP. (2014). Perfluoroalkyl
Substances During Pregnancy and Validated Preeclampsia Among Nulliparous Women
in the Norwegian Mother and Child Cohort Study. Am J Epidemiol 179: 824-833.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2446669
StataCorp. (2021). Stata Statistical Software: Release 17 [Computer Program], College Station,
7-90
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
TX: StataCorp LLC. Retrieved from
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10406419
Steenland, K; Barry, V; Savitz, D. (2018). Serum perfluorooctanoic acid and birthweight: an
updated meta-analysis with bias analysis. Epidemiology 29: 765-776.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079861
Steenland, K; Kugathasan, S; Barr, DB. (2018). PFOA and ulcerative colitis. Environ Res 165:
317-321. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079806
Steenland, K; Tinker, S; Frisbee, S; Ducatman, A; Vaccarino, V. (2009). Association of
Perfluorooctanoic Acid and Perfluorooctane Sulfonate With Serum Lipids Among Adults
Living Near a Chemical Plant. Am J Epidemiol 170: 1268-1278.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291109
Steenland, K; Tinker, S; Shankar, A; Ducatman, A. (2010). Association of perfluorooctanoic
acid (PFOA) and perfluorooctane sulfonate (PFOS) with uric acid among adults with
elevated community exposure to PFOA. Environ Health Perspect 118: 229-233.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290810
Steenland, K; Winquist, A. (2021). PFAS and cancer, a scoping review of the epidemiologic
evidence [Review], Environ Res 194: 110690.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7491705
Stein, CR; Ge, Y; Wolff, MS; Ye, X; Calafat, AM; Kraus, T; Moran, TM. (2016). Perfluoroalkyl
substance serum concentrations and immune response to FluMist vaccination among
healthy adults. Environ Res 149: 171-178.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860111
Stein, CR; Savitz, DA; Dougan, M. (2009). Serum levels of perfluorooctanoic acid and
perfluorooctane sulfonate and pregnancy outcome. Am J Epidemiol 170: 837-846.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290816
Stein, CR; Wolff, MS; Calafat, AM; Kato, K; Engel, SM. (2012). Comparison of polyfluoroalkyl
compound concentrations in maternal serum and amniotic fluid: a pilot study. Reprod
Toxicol 34: 312-316.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332468
Stengel, D; Wahby, S; Braunbeck, T. (2018). In search of a comprehensible set of endpoints for
the routine monitoring of neurotoxicity in vertebrates: sensory perception and nerve
transmission in zebrafish (Danio rerio) embryos. Environ Sci Pollut Res Int 25: 4066-
4084. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238489
Stock, NL; Furdui, VI; Muir, DC; Mabury, SA. (2007). Perfluoroalkyl contaminants in the
Canadian Arctic: evidence of atmospheric transport and local contamination. Environ Sci
Technol 41: 3529-3536.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1289794
Stratakis, N; Rock, S; La Merrill, MA; Saez, M; Robinson, O; Fecht, D; Vrijheid, M; Valvi, D;
Conti, DV; McConnell, R; Chatzi, VL. (2022). Prenatal exposure to persistent organic
pollutants and childhood obesity: A systematic review and meta-analysis of human
studies. Obes Rev23(Sl): el3383.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10176437
Stram, M; Hansen, S; Olsen, SF; Haug, LS; Rantakokko, P; Kiviranta, H; Halldorsson, TI.
(2014). Persistent organic pollutants measured in maternal serum and offspring
neurodevelopmental outcomes—a prospective study with long-term follow-up. Environ
Int 68: 41-48. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2922190
7-91
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Su, M; Liang, X; Xu, X; Wu, X; Yang, B. (2019). Hepatoprotective benefits of vitamin C against
perfluorooctane sulfonate-induced liver damage in mice through suppressing
inflammatory reaction and ER stress. Environ Toxicol Pharmacol 65: 60-65.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080481
Su, TC; Kuo, CC; Hwang, JJ; Lien, GW; Chen, MF; Chen, PC. (2016). Serum perfluorinated
chemicals, glucose homeostasis and the risk of diabetes in working-aged Taiwanese
adults. Environ Int 88: 15-22.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860116
Sun, Q; Zong, G; Valvi, D; Nielsen, F; Coull, B; Grandjean, P. (2018). Plasma concentrations of
perfluoroalkyl substances and risk of Type 2 diabetes: A prospective investigation among
U.S. women. Environ Health Perspect 126: 037001.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4241053
Suo, C; Fan, Z; Zhou, L; Qiu, J. (2017). Perfluorooctane sulfonate affects intestinal immunity
against bacterial infection. Sci Rep 7: 5166.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981310
Takacs, ML; Abbott, BD. (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.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/783393
Takatori, H; Kanno, Y; Watford, WT; Tato, CM; Weiss, G; Ivanov, II; Littman, DR; O'Shea, JJ.
(2009). Lymphoid tissue inducer-like cells are an innate source of IL-17 and IL-22. J Exp
Med 206: 35-41.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9811595
Tan, YM; Clewell, HJ; Andersen, ME. (2008). Time dependencies in perfluorooctylacids
disposition in rat and monkeys: a kinetic analysis. Toxicol Lett 177: 38-47.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919374
Tang, LL; Wang, JD; Xu, TT; Zhao, Z; Zheng, JJ; Ge, RS; Zhu, DY. (2017). Mitochondrial
toxicity of perfluorooctane sulfonate in mouse embryonic stem cell-derived
cardiomyocytes. Toxicology 382: 108-116.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981359
Tao, L; Kannan, K; Wong, CM; Arcaro, KF; Butenhoff, JL. (2008). Perfluorinated compounds in
human milk from Massachusetts, USA. Environ Sci Technol 42: 3096-3101.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290895
Taylor, KW; Hoffman, K; Thayer, KA; Daniels, JL. (2014). Polyfluoroalkyl chemicals and
menopause among women 20-65 years of age (NHANES). Environ Health Perspect 122:
145-150. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850915
TCEQ. (2016). Perfluoro compounds (PFCs).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5975349
Thibodeaux, JR; Hanson, RG; Rogers, JM; Grey, BE; Barbee, BD; Richards, JH; Butenhoff, JL;
Stevenson, LA; Lau, C. (2003). Exposure to perfluorooctane sulfonate during pregnancy
in rat and mouse. I: maternal and prenatal evaluations. Toxicol Sci 74: 369-381.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757855
Thibodeaux, JR; Hanson, RG; Rogers, JM; Grey, BE; Barbee, BD; Richards, JH; Butenhoff, JL;
Stevenson, LA; Lau, C. (2004). Exposure to perfluorooctane sulfonate during pregnancy
in rat and mouse: Maternal and prenatal evaluations (vol 74, pg 369, 2003). Toxicol Sci
82: 359-359. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5082311
7-92
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Thomford, PJ. (2002). 4-week capsule toxicity study with perfluorooctane sulfonic acid
potassium salt (PFOS; T-6295) in Cynomolgus monkeys. (Covance 6329-222; 3M Study
No. T-6295.6; US EPA OPPT AR226-0144). St. Paul, MN: 3M.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5432419
Thomford, PJ. (2002). 104-week dietary chronic toxicity and carcinogenicity study with
perfluorooctane sulfonic acid potassium salt (PFOS; T-6295) in rats (pp. 002148-
002363). (Study No. 6329-183). Madison, WI: Covance Laboratories.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5029075
Thompson, J; Lorber, M; Toms, LM; Kato, K; Calafat, AM; Mueller, JF. (2010). Use of simple
pharmacokinetic modeling to characterize exposure of Australians to perfluorooctanoic
acid and perfluorooctane sulfonic acid. Environ Int 36: 390-397.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919278
Thomsen, C; Haug, LS; Stigum, H; Frashaug, M; Broadwell, SL; Becher, G. (2010). Changes in
concentrations of perfluorinated compounds, polybrominated diphenyl ethers, and
polychlorinated biphenyls in Norwegian breast-milk during twelve months of lactation.
Environ Sci Technol 44: 9550-9556.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/759807
Thomsen, C; Haug, LS; Stigum, H; Froshaug, Ma; Broadwell, SL; Becher, G. (2011). Changes
in Concentrations of Perfluorinated Compounds, Polybrominated Diphenyl Ethers and
Polychlorinated Biphenyls in Norwegian Breast-Milk during Twelve Months of Lactation
(vol 44, pg 9550, 2010). Environ Sci Technol 45: 3192-3192.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2186079
Thorsdottir, I; Birgisdottir, BE. (1998). Different weight gain in women of normal weight before
pregnancy: postpartum weight and birth weight. Obstet Gynecol 92: 377-383.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4940407
Tian, M; Reichetzeder, C; Li, J; Hocher, B. (2019). Low birth weight, a risk factor for diseases in
later life, is a surrogate of insulin resistance at birth. J Hypertens 37: 2123-2134.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/8632212
Tian, Y; Liang, H; Miao, M; Yang, F; Ji, H; Cao, W; Liu, X; Zhang, X; Chen, A; Xiao, H; Hu,
H; Yuan, W. (2019). Maternal plasma concentrations of perfluoroalkyl and
polyfluoroalkyl substances during pregnancy and anogenital distance in male infants.
Hum Reprod 34: 1356-1368.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5390052
Tian, Y; Miao, M; Ji, H; Zhang, X; Chen, A; Wang, Z; Yuan, W; Liang, H. (2020). Prenatal
exposure to perfluoroalkyl substances and cord plasma lipid concentrations. Environ
Pollut 268: 115426.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7026251
Tian, YP; Zeng, XW; Bloom, MS; Lin, S; Wang, SQ; Yim, SHL; Yang, M; Chu, C; Gurram, N;
Hu, LW; Liu, KK; Yang, BY; Feng, D; Liu, RQ; Nian, M; Dong, GH. (2019). Isomers of
perfluoroalkyl substances and overweight status among Chinese by sex status: Isomers of
C8 Health Project in China. Environ Int 124: 130-138.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080586
Timmermann, CA; Budtz-j0rgensen, E; Petersen, MS; Weihe, P; Steuerwald, U; Nielsen, F;
Jensen, TK; Grandjean, P. (2017). Shorter duration of breastfeeding at elevated exposures
to perfluoroalkyl substances. Reprod Toxicol 68: 164-170.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981439
7-93
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Timmermann, CAG; Jensen, KJ; Nielsen, F; Budtz-j0rgensen, E; van Der Klis, F; Benn, CS;
Grandjean, P; Fisker, AB. (2020). Serum Perfluoroalkyl Substances, Vaccine Responses,
and Morbidity in a Cohort of Guinea-Bissau Children. Environ Health Perspect 128:
87002. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833710
Timmermann, CAG; Pedersen, HS; Weihe, P; Bjerregaard, P; Nielsen, F; Heilmann, C;
Grandjean, P. (2021). Concentrations of tetanus and diphtheria antibodies in vaccinated
Greenlandic children aged 7-12 years exposed to marine pollutants, a cross sectional
study. Environ Res 203: 111712.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9416315
Toft, G; Jonsson, BA; Bonde, JP; N0rgaard-Pedersen, B; Hougaard, DM; Cohen, A; Lindh, CH;
Ivell, R; Anand-Ivell, R; Lindhard, MS. (2016). Perfluorooctane Sulfonate
Concentrations in Amniotic Fluid, Biomarkers of Fetal Ley dig Cell Function, and
Cryptorchidism and Hypospadias in Danish Boys (1980-1996). Environ Health Perspect
124:151-156.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3102984
Toft, G; Jonsson, BAG; Lindh, CH; Giwercman, A; Spano, M; Heederik, D; Lenters, V;
Vermeulen, R; Rylander, L; Pedersen, HS; Ludwicki, JK; Zviezdai, V; Bonde, JP.
(2012). Exposure to perfluorinated compounds and human semen quality in arctic and
European populations. Hum Reprod 27: 2532-2540.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332467
Tomy, GT; Budakowski, W; Halldorson, T; Helm, PA; Stern, GA; Friesen, K; Pepper, K;
Tittlemier, SA; Fisk, AT. (2004). Fluorinated organic compounds in an eastern Arctic
marine food web. Environ Sci Technol 38: 6475-6481.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332758
Trefts, E; Gannon, M; Wasserman, DH. (2017). The liver. Curr Biol 27: R1147-R1151.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284972
Trudel, D; Horowitz, L; Wormuth, M; Scheringer, M; Cousins, IT; Hungerbuhler, K. (2008).
Estimating consumer exposure to PFOS and PFOA.[erratum appears in Risk Anal. 2008
Jun;28(3):807], Risk Anal 28: 251-269.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/214241
Tsai, MS; Chang, SH; Kuo, WH; Kuo, CH; Li, SY; Wang, MY; Chang, DY; Lu, YS; Huang,
CS; Cheng, AL; Lin, CH; Chen, PC. (2020). A case-control study of perfluoroalkyl
substances and the risk of breast cancer in Taiwanese women. Environ Int 142: 105850.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833693
Tsai, MS; Lin, CC; Chen, MH; Hsieh, WS; Chen, PC. (2017). Perfluoroalkyl substances and
thyroid hormones in cord blood. Environ Pollut 222: 543-548.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860107
Tsai, MS; Lin, CY; Lin, CC; Chen, MH; Hsu, SH; Chien, KL; Sung, FC; Chen, PC; Su, TC.
(2015). Association between perfluoroalkyl substances and reproductive hormones in
adolescents and young adults. Int J Hyg Environ Health 218: 437-443.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850160
U.S. EPA. (1991). Guidelines for developmental toxicity risk assessment. Fed Reg 56: 63798-
63826. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/732120
U.S. EPA. (1996). 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://hero.epa.gov/hero/index.cfm/reference/details/reference_id/30019
7-94
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
U.S. EPA. (1998). Health effects test guidelines OPPTS 870.3800 reproduction and fertility
effects [EPA Report], (EPA 712-C-98-208). Washington D.C.: U.S. Environmental
Protection Agency, Office of Prevention, Pesticides and Toxic Substances.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2229410
U.S. EPA. (2000). Methodology for deriving ambient water quality criteria for the protection of
human health (2000). (EPA/822/B-00/004). Washington, DC: U.S. Environmental
Protection Agency, Office of Water.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/19428
U.S. EPA. (2002). Hepatocellular hypertrophy. HED guidance document #G2002.01 [EPA
Report], Washington, DC.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/625713
U.S. EPA. (2002). A review of the reference dose and reference concentration processes.
(EPA630P02002F). Washington, DC.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/88824
U.S. EPA. (2005). Guidelines for carcinogen risk assessment [EPA Report], (EPA630P03001B).
Washington, DC.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9638795 (use
6324329)
U.S. EPA. (2005). Guidelines for carcinogen risk assessment [EPA Report], (EPA630P03001F).
Washington, DC.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6324329
U.S. EPA (U.S. Environmental Protection Agency). (2005). 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://hero.epa.gov/hero/index.cfm/reference/details/reference_id/88823
U.S. EPA. (2009). Final Contaminant Candidate List 3 Chemicals: Screening to a PCCL.
(815R09007).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1508321
U.S. EPA. (2010). 2008-2009 National Rivers and Streams Assessment Fish Tissue Study.
Washington, DC.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10369692
U.S. EPA. (2011). 2010 Great Lakes Human Health Fish Tissue Study. Washington, DC: U.S.
Environmental Protection Agency, National Coastal Condition Assessment.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10369695
U.S. EPA. (2011). Exposure factors handbook: 2011 edition [EPA Report], (EPA/600/R-
090/052F). Washington, DC: U.S. Environmental Protection Agency, Office of Research
and Development, National Center for Environmental Assessment.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/786546
U.S. EPA. (2012). Benchmark dose technical guidance. (EPA/100/R-12/001). Washington, DC:
U.S. Environmental Protection Agency, Risk Assessment Forum.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1239433
U.S. EPA. (2015). 2013-2014 National Rivers and Streams Assessment Fish Tissue Study.
Washington, DC.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10369694
U.S. EPA. (2016). Drinking water health advisory for perfluorooctane sulfonate (PFOS) [EPA
Report], (EPA 822-R-16-004). Washington, DC: U.S. Environmental Protection Agency,
7-95
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Office of Water.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3982043
U.S. EPA. (2016). Drinking water health advisory for perfluorooctanoic acid (PFOA) [EPA
Report], (EPA 822-R-16-005). Washington, DC: U.S. Environmental Protection Agency,
Office of Water.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3982042
U.S. EPA. (2016). 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://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3603365
U.S. EPA. (2016). 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://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3603279
U.S. EPA. (2016). National Coastal Condition Assessment: 2015 Results. Washington, DC.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10369696
U.S. EPA. (2016). Safe Drinking Water Act (SDWA) - Section 1412(b) Fourth Contaminant
Candidate List (81 FR 81099). Washington, DC.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6115068
U.S. EPA. (2017). Occurrence Data for the Unregulated Contaminant Monitoring Rule: UCMR 3
(2013-2015). Washington, D.C.: U.S. Environmental Protection Agency, Office of
Water. Retrieved from
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9419085
U.S. EPA. (2018). 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.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4326432
U.S. EPA. (2019). Systematic review protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA
IRIS assessments [EPA Report], (EPA635R19050). Integrated Risk Information System.
Center for Public Health and Environmental Assessment. Office of Research and
Development. U.S. Environmental Protection Agency.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6572089
U.S. EPA. (2020). ORD staff handbook for developing IRIS assessments (public comment draft)
[EPA Report], (EPA/600/R-20/137). Washington, DC: U.S. Environmental Protection
Agency, Office of Research and Development, Center for Public Health and
Environmental Assessment.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7006986
U.S. EPA. Announcement of final regulatory determinations for contaminants on the Fourth
Drinking Water Contaminant Candidate List, 86 FR 12272 (2021).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7487276
U.S. EPA. (2021). Final Regulatory Determination 4 Support Document [EPA Report], (EPA
815R21001). U.S. Environmental Protection Agency (EPA).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9640861
U.S. EPA. (2022). Draft Aquatic Life Ambient Water Quality Criteria for Perfluorooctane
Sulfonate (PFOS). (EPA-842-D-22-002). Washington, DC: U.S. Environmental
Protection Agency, Office of Water.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10668582
7-96
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
U.S. EPA. (2022). Draft Economic Analysis for the Proposed PFAS Rule. Washington, DC.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10369698
U.S. EPA. (2022). INTERIM Drinking Water Health Advisory: Perfluorooctane Sulfonic Acid
(PFOS) CASRN 1763-23-1. (EPA/822/R-22/004). Washington, DC: U.S. Environmental
Protection Agency, Office of Water, Office of Science and Technology.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10668548
U.S. EPA. (2022). ORD staff handbook for developing IRIS assessments [EPA Report], (EPA
600/R-22/268). Washington, DC: U.S. Environmental Protection Agency, Office of
Research and Development, Center for Public Health and Environmental Assessment.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10367891
U.S. EPA. (2022). Review of EPA's Analysis to Support EPA's National Primary Drinking
Water Rulemaking for PFAS. (EPA-SAB-22-008). U.S. Environmental Protection
Agency, Science Advisory Board.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10476098
U.S. EPA. (2023). Economic analysis for the proposed PFAS national primary drinking water
regulation [EPA Report], (EPA-822-P-22-001).
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10692765
Uhl, SA; James-Todd, T; Bell, ML. (2013). Association of Osteoarthritis with
Perfluorooctanoate and Perfluorooctane Sulfonate in NHANES 2003-2008. Environ
Health Perspect 121: 447-452, 452e441-443.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1937226
Vagi, SJ; Azziz-Baumgartner, E; Sjodin, A; Calafat, AM; Dumesic, D; Gonzalez, L; Kato, K;
Silva, MJ; Ye, X; Azziz, R. (2014). Exploring the potential association between
brominated diphenyl ethers, polychlorinated biphenyls, organochlorine pesticides,
perfluorinated compounds, phthalates, and bisphenol a in polycystic ovary syndrome: a
case-control study. BMC Endocrine Disorders 14: 86.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2718073
Valvi, D; Oulhote, Y; Weihe, P; Dalgard, C; Bjerve, KS; Steuerwald, U; Grandjean, P. (2017).
Gestational diabetes and offspring birth size at elevated environmental pollutant
exposures. Environ Int 107: 205-215.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3983872
van den Dungen, MW; Murk, AJ; Kampman, E; Steegenga, WT; Kok, DE. (2017). Association
between DNA methylation profiles in leukocytes and serum levels of persistent organic
pollutants in Dutch men. Environ Epigenet 3: dvxOOl.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080340
van der Veen, I; Hanning, AC; Stare, A; Leonards, PEG; de Boer, J; Weiss, JM. (2020). The
effect of weathering on per- and polyfluoroalkyl substances (PFASs) from durable water
repellent (DWR) clothing. Chemosphere 249: 126100.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6316195
Varshavsky, JR; Robinson, JF; Zhou, Y; Puckett, KA; Kwan, E; Buarpung, S; Aburajab, R;
Gaw, SL; Sen, S; Gao, S; Smith, SC; Park, JS; Zakharevich, I; Gerona, RR; Fisher, SJ;
Woodruff, TJ. (2021). Organophosphate Flame Retardants, Highly Fluorinated
Chemicals, and Biomarkers of Placental Development and Disease during Mid-Gestation.
Toxicol Sci 181: 215-228.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7410195
Velez, MP; Arbuckle, TE; Fraser, WD. (2015). Maternal exposure to perfluorinated chemicals
7-97
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
and reduced fecundity: the MIREC study. Hum Reprod 30: 701-709.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851037
Verner, MA; Loccisano, AE; Morken, NH; Yoon, M; Wu, H; Mcdougall, R; Maisonet, M;
Marcus, M; Kishi, R; Miyashita, C; Chen, MH; Hsieh, WS; Andersen, ME; Clewell, HJ;
Longnecker, MP. (2015). Associations of Perfluoroalkyl Substances (PFAS) with Lower
Birth Weight: An Evaluation of Potential Confounding by Glomerular Filtration Rate
Using a Physiologically Based Pharmacokinetic Model (PBPK). Environ Health Perspect
123:1317-1324.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3150627
Verner, MA; Longnecker, MP. (2015). Comment on "enhanced elimination of
perfluorooctanesulfonic acid by menstruating women: evidence from population-based
pharmacokinetic modeling" [Letter], Environ Sci Technol 49: 5836-5837.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850226
Verner, MA; Ngueta, G; Jensen, ET; Fromme, H; Voelkel, W; Nygaard, UC; Granum, B;
Longnecker, MP. (2016). A simple pharmacokinetic model of prenatal and postnatal
exposure to perfluoroalkyl substances (PFASs). Environ Sci Technol 50: 978-986.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3299692
Vestergren, R; Cousins, IT. (2009). Tracking the pathways of human exposure to
perfluorocarboxylates [Review], Environ Sci Technol 43: 5565-5575.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290815
Volkel, W; Genzel-Boroviczeny, O; Demmelmair, H; Gebauer, C; Koletzko, B; Twardella, D;
Raab, U; Fromme, H. (2008). Perfluorooctane sulphonate (PFOS) and perfluorooctanoic
acid (PFOA) in human breast milk: results of a pilot study. Int J Hyg Environ Health 211:
440-446. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3103448
von Ehrenstein, OS; Fenton, SE; Kato, K; Kuklenyik, Z; Calafat, AM; Hines, EP. (2009).
Polyfluoroalkyl chemicals in the serum and milk of breastfeeding women. Reprod
Toxicol 27: 239-245.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/194805
Vuong, A; Yolton, K; Webster, GM; Sjodin, A; Calafat, AM; Braun, JM; Dietrich, K; Lanphear,
BP; Chen, A. (2016). Prenatal polybrominated diphenyl ether and perfluoroalkyl
substance exposures and executive function in school-age children. Environ Res 147:
556-564. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3352166
Vuong, AM; Braun, JM; Yolton, K; Wang, Z; Xie, C; Webster, GM; Ye, X; Calafat, AM;
Dietrich, KN; Lanphear, BP; Chen, A. (2018). Prenatal and childhood exposure to
perfluoroalkyl substances (PFAS) and measures of attention, impulse control, and visual
spatial abilities. Environ Int 119: 413-420.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079693
Vuong, AM; Xie, C; Jandarov, R; Dietrich, KN; Zhang, H; Sjodin, A; Calafat, AM; Lanphear,
BP; Mccandless, L; Braun, JM; Yolton, K; Chen, A. (2020). Prenatal exposure to a
mixture of persistent organic pollutants (POPs) and child reading skills at school age. Int
J Hyg Environ Health 228: 113527.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833684
Vuong, AM; Yolton, K; Braun, JM; Sjodin, A; Calafat, AM; Xu, Y; Dietrich, KN; Lanphear,
BP; Chen, A. (2020). Polybrominated diphenyl ether (PBDE) and poly- and
perfluoroalkyl substance (PFAS) exposures during pregnancy and maternal depression.
Environ Int 139: 105694.
7-98
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6356876
Vuong, AM; Yolton, K; Wang, Z; Xie, C; Webster, GM; Ye, X; Calafat, AM; Braun, JM;
Dietrich, KN; Lanphear, BP; Chen, A. (2018). Childhood perfluoroalkyl substance
exposure and executive function in children at 8 years. Environ Int 119: 212-219.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079675
Vuong, AM; Yolton, K; Xie, C; Dietrich, KN; Braun, JM; Webster, GM; Calafat, AM;
Lanphear, BP; Chen, A. (2019). Prenatal and childhood exposure to poly- and
perfluoroalkyl substances (PFAS) and cognitive development in children at age 8 years.
Environ Res 172: 242-248.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080218
WA DOH. (2020). DOH Approach to Developing PFAS State Action Levels.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9418278
Wan, C; Han, R; Liu, L; Zhang, F; Li, F; Xiang, M; Ding, W. (2016). Role of miR-155 in
fluorooctane sulfonate-induced oxidative hepatic damage via the Nrf2-dependent
pathway. Toxicol Appl Pharmacol 295: 85-93.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981504
Wan, HT; Wong, AY; Feng, S; Wong, CK. (2020). Effects of In Utero Exposure to
Perfluorooctane Sulfonate on Placental Functions. Environ Sci Technol 54: 16050-16061.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7174720
Wan, HT; Zhao, YG; Leung, PY; Wong, CK. (2014). Perinatal exposure to perfluorooctane
sulfonate affects glucose metabolism in adult offspring. PLoS ONE 9: e87137.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850405
Wan, HT; Zhao, YG; Wei, X; Hui, KY; Giesy, JP; Wong, CK. (2012). PFOS-induced hepatic
steatosis, the mechanistic actions on P-oxidation and lipid transport. Biochim Biophys
Acta Gen Subj 1820: 1092-1101.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332470
Wan Ibrahim, WN; Tofighi, R; Onishchenko, N; Rebellato, P; Bose, R; Uhlen, P; Ceccatelli, S.
(2013). Perfluorooctane sulfonate induces neuronal and oligodendrocytic differentiation
in neural stem cells and alters the expression of PPARy in vitro and in vivo. Toxicol Appl
Pharmacol 269: 51-60.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919149
Wang, B; Zhang, R; Jin, F; Lou, H; Mao, Y; Zhu, W; Zhou, W; Zhang, P; Zhang, J. (2017).
Perfluoroalkyl substances and endometriosis-related infertility in Chinese women.
Environ Int 102: 207-212.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3856459
Wang, H; Du, H; Yang, J; Jiang, H; O, K; Xu, L; Liu, S; Yi, J; Qian, X; Chen, Y; Jiang, Q; He,
G. (2019). PFOS, PFOA, estrogen homeostasis, and birth size in Chinese infants.
Chemosphere 221: 349-355.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080598
Wang, H; Yang, J; Du, H; Xu, L; Liu, S; Yi, J; Qian, X; Chen, Y; Jiang, Q; He, G. (2018).
Perfluoroalkyl substances, glucose homeostasis, and gestational diabetes mellitus in
Chinese pregnant women: A repeat measurement-based prospective study. Environ Int
114: 12-20. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080352
Wang, IJ; Hsieh, WS; Chen, CY; Fletcher, T; Lien, GW; Chiang, HL; Chiang, CF; Wu, TN;
Chen, PC. (2011). The effect of prenatal perfluorinated chemicals exposures on pediatric
atopy. Environ Res 111: 785-791.
7-99
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1424977
Wang, J; Pan, Y; Cui, Q; Yao, B; Wang, J; Dai, J. (2018). Penetration of PFASs across the blood
cerebrospinal fluid barrier and its determinants in humans. Environ Sci Technol 52:
13553-13561.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080654
Wang, J; Zeng, XW; Bloom, MS; Qian, Z; Hinyard, LJ; Belue, R; Lin, S; Wang, SQ; Tian, YP;
Yang, M; Chu, C; Gurram, N; Hu, LW; Liu, KK; Yang, BY; Feng, D; Liu, RQ; Dong,
GH. (2019). Renal function and isomers of perfluorooctanoate (PFOA) and
perfluorooctanesulfonate (PFOS): Isomers of C8 Health Project in China. Chemosphere
218: 1042-1049.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080583
Wang, J; Zhang, Y; Zhang, W; Jin, Y; Dai, J. (2012). Association of perfluorooctanoic acid with
HDL cholesterol and circulating miR-26b and miR-199-3p in workers of a
fluorochemical plant and nearby residents. Environ Sci Technol 46: 9274-9281.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919184
Wang, L; Wang, Y; Liang, Y; Li, J; Liu, Y; Zhang, J; Zhang, A; Fu, J; Jiang, G. (2014). PFOS
induced lipid metabolism disturbances in BALB/c mice through inhibition of low density
lipoproteins excretion. Sci Rep 4: 4582.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851252
Wang, S; Zhuang, C; Du, J; Wu, C; You, H. (2017). The presence of MWCNTs reduces
developmental toxicity of PFOS in early life stage of zebrafish. Environ Pollut 222: 201-
209. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981383
Wang, W; Zhou, W; Wu, S; Liang, F; Li, Y; Zhang, J; Cui, L; Feng, Y; Wang, Y. (2019).
Perfluoroalkyl substances exposure and risk of polycystic ovarian syndrome related
infertility in Chinese women. Environ Pollut 247: 824-831.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080500
Wang, Y; Han, W; Wang, C; Zhou, Y; Shi, R; Bonefeld-J0rgensen, EC; Yao, Q; Yuan, T; Gao,
Y; Zhang, J; Tian, Y. (2019). Efficiency of maternal-fetal transfer of perfluoroalkyl and
polyfluoroalkyl substances. Environ Sci Pollut Res Int 26: 2691-2698.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5083694
Wang, Y; Liu, W; Zhang, Q; Zhao, H; Quan, X. (2015). Effects of developmental
perfluorooctane sulfonate exposure on spatial learning and memory ability of rats and
mechanism associated with synaptic plasticity. Food Chem Toxicol 76: 70-76.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851030
Wang, Y; Rogan, WJ; Chen, HY; Chen, PC; Su, PH; Chen, HY; Wang, SL. (2015). Prenatal
exposure to perfluroalkyl substances and children's IQ: The Taiwan maternal and infant
cohort study. Int J Hyg Environ Health 218: 639-644.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3860120
Wang, Y; Rogan, WJ; Chen, PC; Lien, GW; Chen, HY; Tseng, YC; Longnecker, MP; Wang, SL.
(2014). Association between maternal serum perfluoroalkyl substances during pregnancy
and maternal and cord thyroid hormones: Taiwan maternal and infant cohort study.
Environ Health Perspect 122: 529-534.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850394
Wang, Y; Starling, AP; Haug, LS; Eggesbo, M; Becher, G; Thomsen, C; Travlos, G; King, D;
Hoppin, JA; Rogan, WJ; Longnecker, MP. (2013). Association between perfluoroalkyl
substances and thyroid stimulating hormone among pregnant women: a cross-sectional
7-100
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
study. Environ Health 12: 76.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4241230
Wang, Y; Zhang, C. (2019). The Roles of Liver-Resident Lymphocytes in Liver Diseases
[Review], Front Immunol 10: 1582.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10365737
Wang, Y; Zhang, L; Teng, Y; Zhang, J; Yang, L; Li, J; Lai, J; Zhao, Y; Wu, Y. (2018).
Association of serum levels of perfluoroalkyl substances with gestational diabetes
mellitus and postpartum blood glucose. J Environ Sci 69: 5-11.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079666
Wang, Y; Zhang, X; Wang, M; Cao, Y; Wang, X; Liu, Y; Wang, J; Wang, J; Wu, L; Hei, TK;
Luan, Y; Xu, A. (2015). Mutagenic Effects of Perfluorooctanesulfonic Acid in gpt Delta
Transgenic System Are Mediated by Hydrogen Peroxide. Environ Sci Technol 49: 6294-
6303. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850220
Wang, Z; Shi, R; Ding, G; Yao, Q; Pan, C; Gao, Y; Tian, Y. (2022). Association between
maternal serum concentration of perfluoroalkyl substances (PFASs) at delivery and acute
infectious diseases in infancy. Chemosphere 289: 133235.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10176501
Warembourg, C; Maitre, L, ea; Tamayo-Uria, I; Fossati, S; Roumeliotaki, T; Aasvang, GM;
Andrusaityte, S; Casas, M; Cequier, E; Chatzi, L; Dedele, A; Gonzalez, J. R.;
Grazuleviciene, R; Haug, LS; Hernandez-Ferrer, C; Heude, B; Karachaliou, M; Krog,
NH; Mceachan, R; Nieuwenhuijsen, M; Petraviciene, I; Quentin, J; Robinson, O; Sakhi,
AK; Slama, R; Thomsen, C; Urquiza, J; Vafeiadi, M; West, J; Wright, J; Vrijheid, M;
Basagana, X. (2019). Early-life environmental exposures and blood pressure in children.
J Am Coll Cardiol 74: 1317-1328.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5881345
Watkins, DJ; Josson, J; Elston, B; Bartell, SM; Shin, HM; Vieira, VM; Savitz, DA; Fletcher, T;
Wellenius, GA. (2013). Exposure to perfluoroalkyl acids and markers of kidney function
among children and adolescents living near a chemical plant. Environ Health Perspect
121: 625-630.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850974
Webster, GM; Venners, SA; Mattman, A; Martin, JW. (2014). Associations between
perfluoroalkyl acids (PFASs) and maternal thyroid hormones in early pregnancy: a
population-based cohort study. Environ Res 133: 338-347.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850208
Weisskopf, MG; Seals, RM; Webster, TF. (2018). Bias amplification in epidemiologic analysis
of exposure to mixtures. Environ Health Perspect 126.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7325521
Wen, HJ; Wang, SL; Chen, PC; Guo, YL. (2019). Prenatal perfluorooctanoic acid exposure and
glutathione s-transferase Tl/Ml genotypes and their association with atopic dermatitis at
2 years of age. PLoS ONE 14: e0210708.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5081172
Wen, HJ; Wang, SL; Chuang, YC; Chen, PC; Guo, YL. (2019). Prenatal perfluorooctanoic acid
exposure is associated with early onset atopic dermatitis in 5-year-old children.
Chemosphere 231: 25-31.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5387152
Wen, LL; Lin, CY; Chou, HC; Chang, CC; Lo, HY; Juan, SH. (2016). Perfluorooctanesulfonate
7-101
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Mediates Renal Tubular Cell Apoptosis through PPARgamma Inactivation. PLoS ONE
11: e0155190.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981471
Wen, LL; Lin, LY; Su, TC; Chen, PC; Lin, CY. (2013). Association between serum
perfluorinated chemicals and thyroid function in U.S. adults: the National Health and
Nutrition Examination Survey 2007-2010. J Clin Endocrinol Metab 98: E1456-E1464.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850943
Weng, J; Hong, C; Tasi, J; Shen, CY, u; Su, P; Wang, S. (2020). The association between
prenatal endocrine-disrupting chemical exposure and altered resting-state brain fMRI in
teenagers. Brain Struct Funct 225: 1669-1684.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6718530
Whitworth, KW; Haug, LS; Baird, DD; Becher, G; Hoppin, JA; Skjaerven, R; Thomsen, C;
Eggesbo, M; Travlos, G; Wilson, R; Longnecker, MP. (2012). Perfluorinated compounds
and subfecundity in pregnant women. Epidemiology 23: 257-263.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332476
WHO. (2012). Guidance for immunotoxicity risk assessment for chemicals. (Harmonization
Project Document No. 10). Geneva, Switzerland.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9522548
WHO. (2018). The immunological basis for immunization series. Module 3: Tetanus. Geneva.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10406857
Wiels0e, M; Kern, P; Bonefeld-J0rgensen, EC. (2017). Serum levels of environmental pollutants
is a risk factor for breast cancer in Inuit: a case control study. Environ Health 16: 56.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858479
Wiels0e, M; Long, M; Ghisari, M; Bonefeld-J0rgensen, EC. (2015). Perfluoroalkylated
substances (PFAS) affect oxidative stress biomarkers in vitro. Chemosphere 129: 239-
245. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2533367
Wiener, RC; Waters, C. (2019). Perfluoroalkyls/polyfluoroalkyl substances and dental caries
experience in children, ages 3-11 years, National Health and Nutrition Examination
Survey, 2013-2014. J Public Health Dent 79: 307-319.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5386081
Wikstrom, S; Lin, PI; Lindh, CH; Shu, H; Bornehag, CG. (2019). Maternal serum levels of
perfluoroalkyl substances in early pregnancy and offspring birth weight. Pediatr Res 87:
1093-1099. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6311677
Wikstrom, S; Lindh, CH; Shu, H; Bornehag, CG. (2019). Early pregnancy serum levels of
perfluoroalkyl substances and risk of preeclampsia in Swedish women. Sci Rep 9: 9179.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5387145
Wimsatt, J; Villers, M; Thomas, L; Kamarec, S; Montgomery, C; Yeung, LW; Hu, Y; Innes, K.
(2016). Oral perfluorooctane sulfonate (PFOS) lessens tumor development in the
APC(min) mouse model of spontaneous familial adenomatous polyposis. BMC Cancer
16: 942. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981396
Wolf, CJ; Rider, CV; Lau, C; Abbott, BD. (2014). Evaluating the additivity of perfluoroalkyl
acids in binary combinations on peroxisome proliferator-activated receptor-a activation.
Toxicology 316: 43-54.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850908
Wolf, CJ; Schmid, JE; Lau, C; Abbott, BD. (2012). Activation of mouse and human peroxisome
proliferator-activated receptor-alpha (PPARa) by perfluoroalkyl acids (PFAAs): further
7-102
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
investigation of C4-C12 compounds. Reprod Toxicol 33: 546-551.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1289836
Wolf, CJ; Takacs, ML; Schmid, JE; Lau, C; Abbott, BD. (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.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/716635
Wolf, DC; Moore, T; Abbott, BD; Rosen, MB; Das, KP; Zehr, RD; Lindstrom, AB; Strynar, MJ;
Lau, C. (2008). Comparative hepatic effects of perfluorooctanoic acid and WY 14,643 in
PPAR-alpha knockout and wild-type mice. Toxicol Pathol 36: 632-639.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290827
Wolf, JJ; Studstill, CJ; Hahm, B, . (2019). Emerging Connections of SIP-Metabolizing Enzymes
with Host Defense and Immunity During Virus Infections [Review], Viruses 11: 1097.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10259528
Wong, F; Macleod, M; Mueller, JF; Cousins, IT. (2014). Enhanced elimination of
perfluorooctane sulfonic acid by menstruating women: evidence from population-based
pharmacokinetic modeling. Environ Sci Technol 48: 8807-8814.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851239
Woodcroft, MW; Ellis, DA; Rafferty, SP; Burns, DC; March, RE; Stock, NL; Trumpour, KS;
Yee, J; Munro, K. (2010). Experimental characterization of the mechanism of
perfluorocarboxylic acids' liver protein bioaccumulation: the key role of the neutral
species. Environ Toxicol Chem 29: 1669-1677.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919284
Workman, CE; Becker, AB; Azad, MB; Moraes, TJ; Mandhane, PJ; Turvey, SE; Subbarao, P;
Brook, JR; Sears, MR; Wong, CS. (2019). Associations between concentrations of
perfluoroalkyl substances in human plasma and maternal, infant, and home characteristics
in Winnipeg, Canada. Environ Pollut 249: 758-766.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5387046
Worley, RR; Fisher, J. (2015). Application of physiologically-based pharmacokinetic modeling
to explore the role of kidney transporters in renal reabsorption of perfluorooctanoic acid
in the rat. Toxicol Appl Pharmacol 289: 428-441.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3223252
Worley, RR; Moore, SM; Tierney, BC; Ye, X; Calafat, AM; Campbell, S; Woudneh, MB;
Fisher, J. (2017). Per- and polyfluoroalkyl substances in human serum and urine samples
from a residentially exposed community. Environ Int 106: 135-143.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859800
Worley, RR; Yang, X; Fisher, J. (2017). Physiologically based pharmacokinetic modeling of
human exposure to perfluorooctanoic acid suggests historical non drinking-water
exposures are important for predicting current serum concentrations. Toxicol Appl
Pharmacol 330: 9-21.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981311
Wu, H; Yoon, M; Verner, MA; Xue, J; Luo, M, an; Andersen, ME; Longnecker, MP; Clewell,
HJ, III. (2015). Can the observed association between serum perfluoroalkyl substances
and delayed menarche be explained on the basis of puberty-related changes in physiology
and pharmacokinetics? Environ Int 82: 61-68.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3223290
Wu, XM; Bennett, DH; Calafat, AM; Kato, K; Strynar, M; Andersen, E; Moran, RE; Tancredi,
7-103
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
DJ; Tulve, NS; Hertz-Picciotto, I. (2014). Serum concentrations of perfluorinated
compounds (PFC) among selected populations of children and Adults in California.
Environ Res 136C: 264-273.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2533322
Xiao, C; Grandjean, P; Valvi, D; Nielsen, F; Jensen, TK; Weihe, P; Oulhote, Y. (2019).
Associations of exposure to perfluoroalkyl substances with thyroid hormone
concentrations and birth size. J Clin Endocrinol Metab 105: 735-745.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5918609
Xing, J; Wang, G; Zhao, J; Wang, E; Yin, B; Fang, D; Zhao, J; Zhang, H; Chen, YQ; Chen, W.
(2016). Toxicity assessment of perfluorooctane sulfonate using acute and subchronic
male C57BL/6J mouse models. Environ Pollut 210: 388-396.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981506
Xu, B; Chen, X; Mao, Z; Chen, M; Han, X; Du, G; Ji, X; Chang, C; Rehan, VK; Wang, X; Xia,
Y. (2013). Perfluorooctane sulfonate disturbs Nanog expression through miR-490-3p in
mouse embryonic stem cells. PLoS ONE 8: e74968.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2968325
Xu, B; Chen, X; Mao, Z; Chen, M; Han, X; Du, G; Ji, X; Chang, C; Rehan, VK; Wang, X; Xia,
Y. (2019). Correction: Perfluorooctane Sulfonate Disturbs Nanog Expression through
miR-490-3p in Mouse Embryonic Stem Cells [Erratum], PLoS ONE 14: e0221486.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5413068
Xu, B; Ji, X; Chen, X; Yao, M; Han, X; Chen, M; Tang, W; Xia, Y. (2015). Effect of
perfluorooctane sulfonate on pluripotency and differentiation factors in mouse embryoid
bodies. Toxicology 328: 160-167.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850066
Xu, C; Jiang, ZY; Liu, Q; Liu, H; Gu, A. (2017). Estrogen receptor beta mediates hepatotoxicity
induced by perfluorooctane sulfonate in mouse. Environ Sci Pollut Res Int 24: 13414-
13423. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981352
Xu, H; Zhou, Q; Zhang, J; Chen, X; Zhao, H; Lu, H; Ma, B; Wang, Z; Wu, C; Ying, C; Xiong,
Y; Zhou, Z; Li, X. (2020). Exposure to elevated per- and polyfluoroalkyl substances in
early pregnancy is related to increased risk of gestational diabetes mellitus: A nested
case-control study in Shanghai, China. Environ Int 143: 105952.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6833677
Xu, J; Guo, CS; Zhang, Y; Meng, W. (2014). Bioaccumulation and trophic transfer of
perfluorinated compounds in a eutrophic freshwater food web. Environ Pollut 184: 254-
261. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079760
Xu, M; Liu, G; Li, M; Huo, M; Zong, W; Liu, R. (2020). Probing the Cell Apoptosis Pathway
Induced by Perfluorooctanoic Acid and Perfluorooctane Sulfonate at the Subcellular and
Molecular Levels. J AgricFood Chem 68: 633-641.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6316207
Xu, M; Wan, J; Niu, Q; Liu, R. (2019). PFOA and PFOS interact with superoxide dismutase and
induce cytotoxicity in mouse primary hepatocytes: A combined cellular and molecular
methods. Environ Res 175: 63-70.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5381556
Xu, Y; Fletcher, T; Pineda, D; Lindh, CH; Nilsson, C; Glynn, A; Vogs, C; Norstrom, K; Lilja, K;
Jakobsson, K; Li, Y. (2020). Serum Half-Lives for Short- and Long-Chain Perfluoroalkyl
Acids after Ceasing Exposure from Drinking Water Contaminated by Firefighting Foam.
7-104
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Environ Health Perspect 128: 77004.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6781357
Xu, Y; Li, Y; Scott, K; Lindh, CH; Jakobsson, K; Fletcher, T; Ohlsson, B; Andersson, EM.
(2020). Inflammatory bowel disease and biomarkers of gut inflammation and
permeability in a community with high exposure to perfluoroalkyl substances through
drinking water. Environ Res 181: 108923.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315709
Yahia, D; Tsukuba, C; Yoshida, M; Sato, I; Tsuda, S. (2008). Neonatal death of mice treated
with perfluorooctane sulfonate. J Toxicol Sci 33: 219-226.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919381
Yan, S; Wang, J; Zhang, W; Dai, J. (2014). Circulating microRNA profiles altered in mice after
28 d exposure to perfluorooctanoic acid. Toxicol Lett 224: 24-31.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850901
Yang, C; Lee, HK; Zhang, Y; Jiang, LL; Chen, ZF; Chung, ACK; Cai, Z. (2019). In Situ
Detection and Imaging of PFOS in Mouse Kidney by Matrix-Assisted Laser
Desorption/Ionization Imaging Mass Spectrometry. Anal Chem 91: 8783-8788.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5387049
Yang, D; Han, J; Hall, DR; Sun, J; Fu, J; Kutarna, S; Houck, KA; Lalone, CA; Doering, JA; Ng,
CA; Peng, H. (2020). Nontarget Screening of Per- and Polyfluoroalkyl Substances
Binding to Human Liver Fatty Acid Binding Protein. Environ Sci Technol 54: 5676-
5686. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6356370
Yang, J; Wang, H; Du, H; Fang, H; Han, M; Xu, L; Liu, S; Yi, J; Chen, Y; Jiang, Q; He, G.
(2020). Serum perfluoroalkyl substances in relation to lipid metabolism in Chinese
pregnant women. Chemosphere 273: 128566.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7021246
Yang, L; Li, J; Lai, J; Luan, H; Cai, Z; Wang, Y; Zhao, Y; Wu, Y. (2016). Placental transfer of
perfluoroalkyl substances and associations with thyroid hormones: Beijing prenatal
exposure study. Sci Rep 6: 21699.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858535
Yang, M; Li, LY; Qin, XD; Ye, XY; Yu, S; Bao, Q; Sun, L; Wang, ZB; Bloom, MS; Jalava, P;
Hu, LW; Yu, HY; Zeng, XW; Yang, BY; Dong, GH; Li, CW. (2021).
Perfluorooctanesulfonate and perfluorooctanoate exacerbate airway inflammation in
asthmatic mice and in vitro. Sci Total Environ 766: 142365.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7643494
Yang, Q; Guo, X; Sun, P; Chen, Y; Zhang, W; Gao, A. (2018). Association of serum levels of
perfluoroalkyl substances (PFASs) with the metabolic syndrome (MetS) in Chinese male
adults: A cross-sectional study. Sci Total Environ 621: 1542-1549.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238462
Yang, Q; Wang, W; Liu, C; Wang, Y; Sun, K. (2016). Effect of PFOS on glucocorticoid-induced
changes in human decidual stromal cells in the first trimester of pregnancy. Reprod
Toxicol 63: 142-150.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981458
Yang, Q; Xie, Y; Alexson, SE; Nelson, BD; Depierre, JW. (2002). Involvement of the
peroxisome proliferator-activated receptor alpha in the immunomodulation caused by
peroxisome proliferators in mice. Biochem Pharmacol 63: 1893-1900.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332453
7-105
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Yang, R; Liu, S; Liang, X; Yin, N; Ruan, T; Jiang, L; Faiola, F. (2020). F-53B and PFOS
treatments skew human embryonic stem cell in vitro cardiac differentiation towards
epicardial cells by partly disrupting the WNT signaling pathway. Environ Pollut 261:
114153. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315676
Yang, Y; Lv, QY; Guo, LH; Wan, B; Ren, XM; Shi, YL; Cai, YQ. (2017). Identification of
protein tyrosine phosphatase SHP-2 as a new target of perfluoroalkyl acids in HepG2
cells. Arch Toxicol 91: 1697-1707.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981427
Yang, Z; Liu, HY; Yang, QY; Chen, X; Li, W; Leng, J; Tang, NJ. (2022). Associations between
exposure to perfluoroalkyl substances and birth outcomes: A meta-analysis.
Chemosphere 291: 132909.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10176603
Yao, Q; Gao, Y; Zhang, Y; Qin, K; Liew, Z; Tian, Y. (2021). Associations of paternal and
maternal per- and polyfluoroalkyl substances exposure with cord serum reproductive
hormones, placental steroidogenic enzyme and birth weight. Chemosphere 285: 131521.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9960202
Yao, Q; Shi, R; Wang, C; Han, W; Gao, Y; Zhang, Y; Zhou, Y; Ding, G; Tian, Y. (2019). Cord
blood per- and polyfluoroalkyl substances, placental steroidogenic enzyme, and cord
blood reproductive hormone. Environ Int 129: 573-582.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5187556
Yao, X; Sha, S; Wang, Y; Sun, X; Cao, J; Kang, J; Jiang, L; Chen, M; Ma, Y. (2016).
Perfluorooctane Sulfonate Induces Autophagy-Dependent Apoptosis through Spinster 1-
Mediated lysosomal-Mitochondrial Axis and Impaired Mitophagy. Toxicol Sci 153: 198-
211. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981442
Yao, XF; Cao, J; Xu, LM; Sun, XC; Kang, J; Yang, G; Jiang, LP; Geng, CY; Gao, CZ; Zhong,
LF; Ma, YF. (2014). Perfluorooctane sulfonate blocked autophagy flux and induced
lysosome membrane permeabilization in HepG2 cells. Food Chem Toxicol 67: 96-104.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850398
Ye, L; Zhao, B; Yuan, K; Chu, Y; Li, C; Zhao, C; Lian, QQ; Ge, RS. (2012). Gene expression
profiling in fetal rat lung during gestational perfluorooctane sulfonate exposure. Toxicol
Lett 209: 270-276.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919212
Ye, WL; Chen, ZX; Xie, YQ; Kong, ML; Li, QQ; Yu, S; Chu, C; Dong, GH; Zeng, XW. (2021).
Associations between serum isomers of perfluoroalkyl acids and metabolic syndrome in
adults: Isomers of C8 Health Project in China. Environ Res 196: 110430.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6988486
Yeung, F; Hoberg, JE; Ramsey, CS; Keller, MD; Jones, DR; Frye, RA; Mayo, MW. (2004).
Modulation of NF-kappaB-dependent transcription and cell survival by the SIRT1
deacetylase. EMBO J 23: 2369-2380.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284418
Yeung, LWY; Robinson, SJ; Koschorreck, J; Mabury, SA. (2013). Part II. A Temporal Study of
PFOS and Its Precursors in Human Plasma from Two German Cities in 1982-2009.
Environ Sci Technol 47: 3875-3882.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850973
Young, W; Wiggins, S; Limm, W; Fisher, CM; Dejager, L; Genualdi, S. (2022). Analysis of Per-
and Poly(fluoroalkyl) Substances (PFASs) in Highly Consumed Seafood Products from
7-106
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
U.S. Markets. J Agric Food Chem 70: 13545-13553.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10601281
Yu, S; Feng, WR; Liang, ZM; Zeng, XY; Bloom, MS; Hu, GC; Zhou, Y; Ou, YQ; Chu, C; Li,
QQ; Yu, Y; Zeng, XW; Dong, GH. (2021). Perfluorooctane sulfonate alternatives and
metabolic syndrome in adults: New evidence from the Isomers of C8 Health Project in
China. Environ Pollut 283: 117078.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/8453076
Yu, WG; Liu, W; Jin, YH. (2009). Effects of perfluorooctane sulfonate on rat thyroid hormone
biosynthesis and metabolism. Environ Toxicol Chem 28: 990-996.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/757872
Yu, WG; Liu, W; Liu, L; Jin, YH. (2011). Perfluorooctane sulfonate increased hepatic
expression of OAPT2 and MRP2 in rats. Arch Toxicol 85: 613-621.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1294541
Yu, Y; Wang, C; Zhang, X; Zhu, J; Wang, L; Ji, M; Zhang, Z; Ji, XM; Wang, SL. (2019).
Perfluorooctane sulfonate disrupts the blood brain barrier through the crosstalk between
endothelial cells and astrocytes in mice. Environ Pollut 256: 113429.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5918598
Zabaleta, I; Blanco-Zubiaguirre, L; Baharli, EN; Olivares, M; Prieto, A; Zuloaga, O; Elizalde,
MP. (2020). Occurrence of per- and polyfluorinated compounds in paper and board
packaging materials and migration to food simulants and foodstuffs. Food Chem 321:
126746. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6505866
Zabaleta, I; Negreira, N; Bizkarguenaga, E; Prieto, A; Covaci, A; Zuloaga, O. (2017). Screening
and identification of per- and polyfluoroalkyl substances in microwave popcorn bags.
Food Chem 230: 497-506.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981827
Zafeiraki, E; Gebbink, WA; Hoogenboom, R; Kotterman, M; Kwadijk, C; Dassenakis, E; van
Leeuwen, SPJ. (2019). Occurrence of perfluoroalkyl substances (PFASs) in a large
number of wild and farmed aquatic animals collected in the Netherlands. Chemosphere
232: 415-423.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5387058
Zare Jeddi, M; Dalla Zuanna, T; Barbieri, G; Fabricio, ASC; Dapra, F; Fletcher, T; Russo, F;
Pitter, G; Canova, C. (2021). Associations of Perfluoroalkyl Substances with Prevalence
of Metabolic Syndrome in Highly Exposed Young Adult Community Residents-A Cross-
Sectional Study in Veneto Region, Italy. Int J Environ Res Public Health 18: 1194.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7404065
Zare Jeddi, M; Soltanmohammadi, R; Barbieri, G; Fabricio, ASC; Pitter, G; Dalla Zuanna, T;
Canova, C. (2021). To which extent are per-and poly-fluorinated substances associated to
metabolic syndrome? [Review], Rev Environ Health.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/8347183
Zarei, MH; Hosseini Shirazi, SF; Aghvami, M; Pourahmad, J. (2018). Perfluorooctanesulfonate
(PFOS) Induces Apoptosis Signaling and Proteolysis in Human Lymphocytes through
ROS Mediated Mitochondrial Dysfunction and Lysosomal Membrane Labialization. Iran
J Pharm Res 17: 995-1007.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079848
Zareitalabad, P; Siemens, J; Hamer, M; Amelung, W. (2013). Perfluorooctanoic acid (PFOA)
and perfluorooctanesulfonic acid (PFOS) in surface waters, sediments, soils and
7-107
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
wastewater - A review on concentrations and distribution coefficients [Review],
Chemosphere 91: 725-732.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080561
Zeeshan, M; Yang, Y; Zhou, Y; Huang, W; Wang, Z; Zeng, XY; Liu, RQ; Yang, BY; Hu, LW;
Zeng, XW; Sun, X; Yu, Y; Dong, GH. (2020). Incidence of ocular conditions associated
with perfluoroalkyl substances exposure: Isomers of C8 Health Project in China. Environ
Int 137: 105555.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315698
Zeng, HC; Zhang, L; Li, YY; Wang, YJ; Xia, W; Lin, Y; Wei, J; Xu, SQ. (2011). Inflammation-
like glial response in rat brain induced by prenatal PFOS exposure. Neurotoxicology 32:
130-139. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1326732
Zeng, X; Chen, Q; Zhang, X; Li, H; Liu, Q; Li, C; Ma, M; Zhang, J; Zhang, W; Zhang, J; Huang,
L. (2019). Association between prenatal exposure to perfluoroalkyl substances and
asthma-related diseases in preschool children. Environ Sci Pollut Res Int 26: 29639-
29648. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5412431
Zeng, XW; Bloom, MS; Dharmage, SC; Lodge, CJ; Chen, D; Li, S; Guo, Y; Roponen, M;
Jalava, P; Hirvonen, MR; Ma, H; Hao, YT; Chen, W; Yang, M; Chu, C; Li, QQ; Hu, LW;
Liu, KK; Yang, BY; Liu, S; Fu, C; Dong, GH. (2019). Prenatal exposure to
perfluoroalkyl substances is associated with lower hand, foot and mouth disease viruses
antibody response in infancy: Findings from the Guangzhou Birth Cohort Study. Sci
Total Environ 663: 60-67.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5081554
Zeng, XW; Li, QQ; Chu, C; Ye, WL; Yu, S; Ma, H; Zeng, XY; Zhou, Y; Yu, HY; Hu, LW;
Yang, BY; Dong, GH. (2020). Alternatives of perfluoroalkyl acids and hepatitis B virus
surface antibody in adults: Isomers of C8 Health Project in China. Environ Pollut 259:
113857. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315718
Zeng, XW; Lodge, CJ; Dharmage, SC; Bloom, MS; Yu, Y; Yang, M; Chu, C; Li, QQ; Hu, LW;
Liu, KK; Yang, BY; Dong, GH. (2019). Isomers of per- and polyfluoroalkyl substances
and uric acid in adults: Isomers of C8 Health Project in China. Environ Int 133: 105160.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5918630
Zeng, XW; Qian, Z; Emo, B; Vaughn, M; Bao, J; Qin, XD; Zhu, Y; Li, J; Lee, YL; Dong, GH.
(2015). Association of polyfluoroalkyl chemical exposure with serum lipids in children.
Sci Total Environ 512-513: 364-370.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851005
Zhang, C; Sundaram, R; Maisog, J; Calafat, AM; Barr, DB; Buck Louis, GM. (2015). A
prospective study of prepregnancy serum concentrations of perfluorochemicals and the
risk of gestational diabetes. Fertil Steril 103: 184-189.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2857764
Zhang, D, eY; Xu, XL, u; Ruan, Q, in; Shen, X, iuY; Lu, Y, in. (2019). Subchronic Effects of
Perfluorooctane Sulphonate on the Testicular Morphology and Spermatogenesis in Mice.
Pakistan J Zool 51: 2217-2223.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5918673
Zhang, H; Lu, H; Yu, L; Yuan, J; Qin, S; Li, C; Ge, RS; Chen, H; Ye, L. (2021). Effects of
gestational exposure to perfluorooctane sulfonate on the lung development of offspring
rats. Environ Pollut 272: 115535.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6988534
7-108
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Zhang, H; Yolton, K; Webster, GM; Ye, X; Calafat, AM; Dietrich, KN; Xu, Y; Xie, C; Braun,
JM; Lanphear, BP; Chen, A. (2018). Prenatal and childhood perfluoroalkyl substances
exposures and children's reading skills at ages 5 and 8 years. Environ Int 111: 224-231.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238294
Zhang, J, ie; Liang, J; Zhu, H; Li, C; Wu, Q. (2013). PFOS and PCB 153 have direct adverse
effects on neonatal testis modeled using a coculture of primary gonocyte and Sertoli cells.
Environ Toxicol 28: 322-331.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1598626
Zhang, L; Ren, XM; Guo, LH. (2013). Structure-based investigation on the interaction of
perfluorinated compounds with human liver fatty acid binding protein. Environ Sci
Technol 47: 11293-11301.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5081488
Zhang, LY; Ren, XM; Wan, B; Guo, LH. (2014). Structure-dependent binding and activation of
perfluorinated compounds on human peroxisome proliferator-activated receptor y.
Toxicol Appl Pharmacol 279: 275-283.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5081455
Zhang, Q; Liu, W; Zhao, H; Zhang, Z; Qin, H; Luo, F; Niu, Q. (2019). Developmental
perfluorooctane sulfonate exposure inhibits long-term potentiation by affecting AMPA
receptor trafficking. Toxicology 412: 55-62.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080461
Zhang, S; Kang, Q; Peng, H; Ding, M; Zhao, F; Zhou, Y; Dong, Z; Zhang, H; Yang, M; Tao, S;
Hu, J. (2019). Relationship between perfluorooctanoate and perfluorooctane sulfonate
blood concentrations in the general population and routine drinking water exposure.
Environ Int 126: 54-60.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080526
Zhang, S; Tan, R; Pan, R; Xiong, J; Tian, Y; Wu, J; Chen, L. (2018). Association of
perfluoroalkyl and polyfluoroalkyl substances with premature ovarian insufficiency in
Chinese women. J Clin Endocrinol Metab 103: 2543-2551.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079665
Zhang, T; Qin, X. (2014). Assessment of fetal exposure and maternal elimination of
perfluoroalkyl substances. Environ Sci Process Impacts 16: 1878-1881.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850251
Zhang, T; Sun, H; Lin, Y; Qin, X; Zhang, Y; Geng, X; Kannan, K. (2013). Distribution of poly-
and perfluoroalkyl substances in matched samples from pregnant women and carbon
chain length related maternal transfer. Environ Sci Technol 47: 7974-7981.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859792
Zhang, T; Sun, H; Qin, X; Gan, Z; Kannan, K. (2015). PFOS and PFOA in paired urine and
blood from general adults and pregnant women: assessment of urinary elimination.
Environ Sci Pollut Res Int 22: 5572-5579.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2851103
Zhang, X; Chen, L; Fei, XC; Ma, YS; Gao, HW. (2009). Binding of PFOS to serum albumin and
DNA: Insight into the molecular toxicity of perfluorochemicals. BMC Mol Biol 10: 16.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2919350
Zhang, Y; Beesoon, S; Zhu, L; Martin, JW. (2013). Biomonitoring of perfluoroalkyl acids in
human urine and estimates of biological half-life. Environ Sci Technol 47: 10619-10627.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859849
7-109
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
Zhang, Y; Beesoon, S; Zhu, L; Martin, JW. (2013). Isomers of perfluorooctanesulfonate and
perfluorooctanoate and total perfluoroalkyl acids in human serum from two cities in
North China. Environ Int 53: 9-17.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2639569
Zhang, YM; Dong, XY; Fan, LJ; Zhang, ZL; Wang, Q; Jiang, N; Yang, XS. (2017). Poly- and
perfluorinated compounds activate human pregnane X receptor. Toxicology 380: 23-29.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3604013
Zhang, YY; Tang, LL; Zheng, B; Ge, RS; Zhu, DY. (2016). Protein profiles of cardiomyocyte
differentiation in murine embryonic stem cells exposed to perfluorooctane sulfonate. J
Appl Toxicol 36: 726-740.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981565
Zhao, L; Zhang, Y; Zhu, L; Ma, X; Wang, Y; Sun, H; Luo, Y. (2017). Isomer-Specific
Transplacental Efficiencies of Perfluoroalkyl Substances in Human Whole Blood.
Environ Sci Technol Lett 4: 391-398.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5085130
Zhao, W; Zitzow, JD; Ehresman, DJ; Chang, SC; Butenhoff, JL; Forster, J; Hagenbuch, B.
(2015). Na+/taurocholate cotransporting polypeptide and apical sodium-dependent bile
acid transporter are involved in the disposition of perfluoroalkyl sulfonates in humans
and rats. Toxicol Sci 146: 363-373.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3856550
Zhao, W; Zitzow, JD; Weaver, Y; Ehresman, DJ; Chang, SC; Butenhoff, JL; Hagenbuch, B.
(2017). Organic anion transporting polypeptides contribute to the disposition of
perfluoroalkyl acids in humans and rats. Toxicol Sci 156: 84-95.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3856461
Zheng, L; Dong, GH; Jin, YH; He, QC. (2009). Immunotoxic changes associated with a 7-day
oral exposure to perfluorooctanesulfonate (PFOS) in adult male C57BL/6 mice. Arch
Toxicol 83: 679-689.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1429960
Zheng, Y; Valdez, PA; Danilenko, DM; Hu, Y; Sa, SM; Gong, Q; Abbas, AR; Modrusan, Z;
Ghilardi, N; de Sauvage, FJ; Ouyang, W. (2008). Interleukin-22 mediates early host
defense against attaching and effacing bacterial pathogens. Nat Med 14: 282-289.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284265
Zhong, SQ; Chen, ZX; Kong, ML; Xie, YQ; Zhou, Y; Qin, XD; Paul, G; Zeng, XW; Dong, GH.
(2016). Testosterone-Mediated Endocrine Function and TH1/TH2 Cytokine Balance after
Prenatal Exposure to Perfluorooctane Sulfonate: By Sex Status. International Journal of
Molecular Sciences 17.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3748828
Zhou, R; Cheng, W; Feng, Y; Wei, H; Liang, F; Wang, Y. (2017). Interactions between three
typical endocrine-disrupting chemicals (EDCs) in binary mixtures exposure on
myocardial differentiation of mouse embryonic stem cell. Chemosphere 178: 378-383.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981356
Zhou, W; Zhang, L; Tong, C; Fang, F; Zhao, S; Tian, Y; Tao, Y; Zhang, J. (2017). Plasma
perfluoroalkyl and polyfluoroalkyl substances concentration and menstrual cycle
characteristics in preconception women. Environ Health Perspect 125: 067012.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859799
Zhou, Y; Bao, WW; Qian, ZM; Dee Geiger, S; Parrish, KL; Yang, BY; Lee, YL; Dong, GH.
7-110
-------
DRAFT FOR PUBLIC COMMENT
MARCH 2023
(2017). Perfluoroalkyl substance exposure and urine CC16 levels among asthmatics: A
case-control study of children. Environ Res 159: 158-163.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981296
Zhou, Y; Hu, LW; Qian, ZM; Chang, JJ; King, C; Paul, G; Lin, S; Chen, PC; Lee, YL; Dong,
GH. (2016). Association of perfluoroalkyl substances exposure with reproductive
hormone levels in adolescents: By sex status. Environ Int 94: 189-195.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3856472
Zhou, Y; Hu, LW; Qian, ZM; Geiger, SD; Parrish, KL; Dharmage, SC; Campbell, B; Roponen,
M; Jalava, P; Hirvonen, MR; Heinrich, J; Zeng, XW; Yang, BY; Qin, XD; Lee, YL;
Dong, GH. (2017). Interaction effects of polyfluoroalkyl substances and sex steroid
hormones on asthma among children. Sci Rep 7: 899.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3858488
Zhu, J; Qian, W; Wang, Y; Gao, R; Wang, J; Xiao, H. (2015). Involvement of mitogen-activated
protein kinase and NF-kB signaling pathways in perfluorooctane sulfonic acid-induced
inflammatory reaction in BV2 microglial cells. J Appl Toxicol 35: 1539-1549.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2850996
Zhu, Y; Qin, XD; Zeng, XW; Paul, G; Morawska, L; Su, MW; Tsai, CH; Wang, SQ; Lee, YL;
Dong, GH. (2016). Associations of serum perfluoroalkyl acid levels with T-helper cell-
specific cytokines in children: By gender and asthma status. Sci Total Environ 559: 166-
173. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3360105
Zong, G; Grandjean, P; Wang, X; Sun, Q. (2016). Lactation history, serum concentrations of
persistent organic pollutants, and maternal risk of diabetes. Environ Res 150: 282-288.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3350666
7-111
------- |