vvEPA
April 2024
EPA Document No. 815R24007
FINAL
Human Health Toxicity Assessment for Perfluorooctane
Sulfonic Acid (PFOS) and Related Salts
-------
APRIL 2024
FINAL
Human Health Toxicity Assessment for Perfluorooctane Sulfonic Acid (PFOS)
and Related Salts
Prepared by:
U.S. Environmental Protection Agency
Office of Water (4304T)
Health and Ecological Criteria Division
Washington, DC 20460
EPA Document Number: 815R24007
April 2024
-------
APRIL 2024
Disclaimer
This document has been reviewed in accordance with U.S. Environmental Protection Agency
(EPA) policy and approved for publication. Mention of trade names or commercial products does
not constitute endorsement or recommendation for use.
i
-------
APRIL 2024
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; and Ruth Etzel. 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 Czarina Cooper; Joyce Donohue (retired); Adrienne Keel;
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 (retired); Colleen Flaherty (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; Samantha
Hall; Anthony Hannani; 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; Kate
Weinberger; 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; Shanell Folger; Caroline Foster; Jeremy Frye; Angelina Guiducci; Tara Hamilton;
Pamela Hartman; Cara Henning; Audrey Ichida; Caroline Jasperse; Kaedra Jones; Michele
Justice; Afroditi Katsigiannakis; Gillian Laidlaw; Yi Lu; Mary Lundin; Elizabeth Martin;
Denyse Marquez Sanchez; Alicia Murphy; Emily Pak; Joei Robertson; Lucas Rocha Melogno;
Andrea Santa-Rios; Alessandria Schumacher; Swati Sriram; Nkoli Ukpabi; Harry Whately; and
Wanchen Xiong.
ii
-------
APRIL 2024
Contents
Disclaimer i
Acknowledgments ii
Contents iii
Figures vi
Tables xi
Acronyms and Abbreviations xiv
Executive Summary xx
1 Background 1-1
1.1 Purpose of This Document 1-1
1.2 Background on Per-and Polyfluoroalkyl Substances 1-2
1.3 Chemical Identity 1-3
1.4 Occurrence Summary 1-4
1.4.1 Biomonitoring 1-4
1.4.2 Ambient Water 1-5
1.4.3 Drinking Water 1-6
1.5 History of EPA's Human Health Assessment for PFOS 1-7
2 Summary of Assessment Methods 2-1
2.1 Introduction to the Systematic Review Assessment Methods 2-1
2.1.1 Literature Database 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 2-5
2.1.5 Evidence Synthesis and Integration 2-6
2.2 Dose-Response Assessment 2-7
2.2.1 Approach to POD and Candidate RfD Derivation for Noncancer Health
Outcomes 2-8
2.2.2 Cancer Assessment 2-10
2.2.3 Selecting Health Outcome-Specific and Overall Toxicity Values 2-12
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 ofPFOS by Health Outcome 3-5
3.2 Data Extraction Results 3-5
3.3 Toxicokinetic Synthesis 3-6
3.3.1 ADME 3-6
3.3.2 Pharmacokinetic Models 3-15
-------
APRIL 2024
3.4 Noncancer Health Effects Evidence Synthesis and Integration 3-22
3.4.1 Hepatic 3-22
3.4.2 Immune 3-79
3.4.3 Cardiovascular 3-137
3.4.4 Developmental 3-186
3.4.5 Evidence Synthesis and Integration for Other Noncancer Health Outcomes3-265
3.5 Cancer Evidence Study Quality Evaluation, Synthesis, Mode of Action Analysis
and Weight of Evidence 3-265
3.5.1 Human Evidence Study Quality Evaluation and Synthesis 3-265
3.5.2 Animal Evidence Study Quality Evaluation and Synthesis 3-272
3.5.3 Mechanistic Evidence 3-274
3.5.4 Weight Of Evidence for Carcinogenicity 3-289
3.5.5 Cancer Classification 3-306
4 Dose-Response Assessment 4-1
4.1 Noncancer 4-2
4.1.1 Study and Endpoint Selection 4-2
4.1.2 Estimation or Selection of Points of Departure for RfD Derivation 4-20
4.1.3 Pharmacokinetic Modeling Approaches to Convert Administered Dose to
Internal Dose in Animals and Humans 4-26
4.1.4 Application of Pharmacokinetic Modeling for Animal-Human
Extrapolation of PFOS Toxicological Endpoints and Dosimetric
Interpretation of Epidemiological Endpoints 4-37
4.1.5 Derivation of Candidate Chronic Oral Reference Doses (RfDs) 4-53
4.1.6 RfD Selection 4-62
4.2 Cancer 4-67
4.2.1 Study and Endpoint Selection 4-67
4.2.2 Candidate CSF Derivation 4-70
4.2.3 Overall CSF Selection 4-72
4.2.4 Application of Age-Dependent Adjustment Factors 4-72
5 Effects Characterization 5-1
5.1 Addressing Uncertainties in the Use of Epidemiological Studies for Quantitative
Dose-Response Analyses 5-1
5.1.1 Uncertainty Due to Potential Confounding by Co-Occurring PFAS 5-4
5.2 Comparisons Between Toxicity Values Derived from Animal Toxicological
Studies and Epidemiological Studies 5-9
5.3 Updated Approach to Animal Toxicological RfD Derivation Compared with the
2016 PFOS HESD 5-10
5.4 Reevaluation of the PFOS Carcinogenicity Database 5-12
5.5 Health Outcomes with Evidence Integration Judgments of Evidence Suggests
Bordering on Evidence Indicates 5-16
5.6 Challenges and Uncertainty in Modeling 5-18
iv
-------
APRIL 2024
5.6.1 Modeling of Animal Internal Dosimetry 5-18
5.6.2 Modeling of Human Dosimetry 5-19
5.6.3 Approach of Estimating a Benchmark Dose from a Regression Coefficient. 5-20
5.7 Human Dosimetry Models: Consideration of Alternate Modeling Approaches 5-21
5.8 Sensitive Populations 5-24
5.8.1 Fetuses, Infants, Children 5-25
5.8.2 Other Susceptible Populations 5-25
6 References 6-1
v
-------
APRIL 2024
Figures
Figure ES-1. Schematic Depicting Candidate RfDs Derived From Epidemiological and
Animal Toxicological Studies of PFOS xxiii
Figure 1-1. Distribution of PFOS Concentrations in Surface Waters by State/Waterbody
(Excluding Great Lakes) in the United States {Jarvis, 2021, 9416544} 1-6
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 3-19
Figure 3-5. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Hepatic Effects Published Before 2016 (References in the
2016 PFOS HESD) 3-23
Figure 3-6. Overall ALT Levels from 2016 PFOS HESD Epidemiology Studies Following
Exposure to PFOS 3-24
Figure 3-7. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Hepatic Effects21 3-27
Figure 3-8. Overall ALT Levels from Epidemiology Studies Following Exposure to PFOS... 3-29
Figure 3-9. Odds of Elevated ALT Levels from Epidemiology Studies Following Exposure
to PFOS 3-29
Figure 3-10. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFOS Exposure and Hepatic Effectsa'b 3-31
Figure 3-11. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFOS Exposure and Hepatic Effects (Continued) a'b 3-32
Figure 3-12. Percent Change in Serum Enzyme Levels Relative to Controls in Mice
Following Exposure to PFOSa'b 3-35
Figure 3-13. Percent Change in Serum Enzyme Levels Relative to Controls in Male Rats
Following Exposure to PFOSa'b 3-37
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 3-44
Figure 3-16. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Immune Effects Published Before 2016 (References in
2016 PFOS HESD) 3-80
vi
-------
APRIL 2024
Figure 3-17. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Immunosuppression Effects 3-84
Figure 3-18. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Immunosuppression Effects (Continued) 3-85
Figure 3-19. Overall Tetanus Antibody Levels in Children from Epidemiology Studies
Following Exposure to PFOS 3-87
Figure 3-20. Overall Diphtheria Antibody Levels in Children from Epidemiology Studies
Following Exposure to PFOS 3-88
Figure 3-21. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Immune Hypersensitivity Effects 3-95
Figure 3-22. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Autoimmune Effects 3-99
Figure 3-23. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFOS Exposure and Immune Effects21 3-101
Figure 3-24. Percent Change in Thymus Weights Relative to Controls in Rodents
Following Exposure to PFOS 3-104
Figure 3-25. Incidences of Immune Cell Histopathology in Rodents Following Exposure to
PFOS 3-106
Figure 3-26. Splenocyte Cellularity in Rodents Following Exposure to PFOS (Logarithmic
Scale)a 3-109
Figure 3-27. Thymocyte Cellularity in Rodents Following Exposure to PFOS (Logarithmic
Scale) 3-110
Figure 3-28. Summary of Mechanistic Studies of PFOS and Immune Effects 3-113
Figure 3-29. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Cardiovascular Effects Published Before 2016 (References
in the 2016 PFOS HESD) 3-138
Figure 3-30. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Cardiovascular Effects 3-141
Figure 3-31. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Cardiovascular Effects (Continued) 3-142
Figure 3-32. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Serum Lipids Published Before 2016 (References in the
2016 PFOS HESD) 3-149
Figure 3-33. Summary of Study Evaluation for Epidemiology Studies of PFOS and Serum
Lipids 3-154
Figure 3-34. Summary of Study Evaluation for Epidemiology Studies of PFOS and Serum
Lipids (Continued) 3-155
vii
-------
APRIL 2024
Figure 3-35. Summary of Study Evaluation for Epidemiology Studies of PFOS and Serum
Lipids (Continued) 3-156
Figure 3-36. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS 3-162
Figure 3-37. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS (Continued) 3-163
Figure 3-38. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS (Continued) 3-164
Figure 3-39. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS (Continued) 3-165
Figure 3-40. Odds of High Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS 3-166
Figure 3-41. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFOS Exposure and Cardiovascular Effects 3-170
Figure 3-42. Serum Lipid Levels in Animal Models Following Exposure to PFOS 3-172
Figure 3-43. Summary of Mechanistic Studies of PFOS and Cardiovascular Effects 3-173
Figure 3-44. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Developmental Effects Published before 2016 (References
from 2016 PFOS HESD) 3-189
Figure 3-45. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Birth Weight Effects a 3-198
Figure 3-46. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Weight Effects (Continued)21 3-199
Figure 3-47. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Weight Effects (Continued)21 3-200
Figure 3-48. Overall Mean Birth Weight from Epidemiology Studies Following Exposure
to PFOS 3-202
Figure 3-49. Overall Mean Birth Weight from Epidemiology Studies Following Exposure
to PFOS (Continued) 3-203
Figure 3-50. Overall Mean Birth Weight from Epidemiology Studies Following Exposure
to PFOS (Continued) 3-204
Figure 3-51. Overall Mean Birth Weight from Epidemiology Studies Following Exposure
to PFOS (Continued) 3-205
Figure 3-52. Overall Mean Birth Weight from Epidemiology Studies Following Exposure
to PFOS (Continued) 3-206
viii
-------
APRIL 2024
Figure 3-53. Odds of Small for Gestational Age in Children from High Confidence
Epidemiology Studies Following Exposure to PFOS 3-210
Figure 3-54. Odds of Small for Gestational Age in Children from Medium Confidence
Epidemiology Studies Following Exposure to PFOS 3-211
Figure 3-55. Odds of Low Birthweight in Children from Epidemiology Studies Following
Exposure to PFOS 3-212
Figure 3-56. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Small for Gestational Age and Low Birth Weight Effects.. 3-213
Figure 3-57. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Birth Length Effects a 3-215
Figure 3-58. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Birth Length Effects (Continued)a 3-216
Figure 3-59. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Birth Head Circumference Effects 3-219
Figure 3-60. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOA Exposure and Postnatal Growth 3-225
Figure 3-61. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Gestational Age 3-228
Figure 3-62. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Preterm Birth Effects 3-231
Figure 3-63. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Fetal Loss 3-234
Figure 3-64. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOA Exposure and Birth Defects 3-235
Figure 3-65. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFOA Exposure and Developmental Effects 3-236
Figure 3-66. Maternal Body Weight in Mice, Rats, and Rabbits Following Exposure to
PFOS 3-239
Figure 3-67. Mortality and Viability in Mice, Rats, and Rabbits Following Exposure to
PFOS (Logarithmic Scale) 3-241
Figure 3-68. Offspring Body Weight in Mice, Rats, and Rabbits Following Exposure to
PFOS (Logarithmic Scale, Sorted by Observation Time) 3-244
Figure 3-69. Summary of Mechanistic Studies of PFOS and Developmental Effects 3-246
Figure 3-70. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Cancer Effects Published Before 2016 (References from
2016 PFOS HESD) 3-267
ix
-------
APRIL 2024
Figure 3-71. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Cancer Effects 3-269
Figure 3-72. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFOS Exposure and Cancer Effects 3-273
Figure 3-73. Summary of Mechanistic Studies of PFOS and Cancer Effects 3-274
Figure 4-1. Model Structure for Lifestage Modeling 4-30
Figure 4-2. Gestation/Lactation Predictions of PFOS in the Rat 4-32
Figure 4-3. Comparison of Candidate RfDs Resulting from the Application of Uncertainty
Factors to PODheds Derived from Epidemiological and Animal Toxicological
Studies 4-63
Figure 4-4. Schematic Depicting Selection of the Overall RfD for PFOS 4-66
x
-------
APRIL 2024
Tables
Table ES-1. Final Toxicity Values for PFOS XXV
Table 1-1. Chemical and Physical Properties of PFOS 1-4
Table 3-1. Database Literature Search Results 3-1
Table 3-2. Associations Between Elevated Exposure to PFOS and Hepatic Outcomes From
Studies Identified in the 2016 PFOS HESD 3-25
Table 3-3. Incidences of Nonneoplastic Lesions in Male and Female Sprague-Dawley Rats,
as Reported by NTP {,2019, 5400978} 3-41
Table 3-4. Incidences of Nonneoplastic Lesions in Male and Female Sprague-Dawley Rats,
as Reported by Curran et al. {, 2008, 757871} 3-41
Table 3-5. Incidences of Nonneoplastic Lesions in Male and Female Sprague-Dawley Rats,
as Reported by Thomford {, 2002, 5029075} 3-41
Table 3-6. Evidence Profile Table for PFOS Exposure and Hepatic Effects 3-74
Table 3-7. Associations Between Elevated Exposure to PFOS and Immune Outcomes From
Studies Identified in the 2016 PFOS HESD 3-82
Table 3-8. Associations between PFOS Exposure and Vaccine Response in Faroe Island
Studies 3-89
Table 3-9. Associations Between PFOS Exposure and Natural Killer Cell Activity in Mice. 3-108
Table 3-10. Associations Between PFOS Exposure and Immune Response in Mice 3-111
Table 3-11. Effects of PFOS Exposure on Pro-Inflammatory Cytokines and Markers of
Inflammation 3-121
Table 3-12. Evidence Profile Table for PFOS Exposure and Immune Effects 3-128
Table 3-13. Associations Between Elevated Exposure to PFOS and Cardiovascular
Outcomes From Studies Identified in the 2016 PFOS HESD 3-138
Table 3-14. Associations Between Elevated Exposure to PFOS and Serum Lipids From
Studies Identified in the 2016 PFOS HESD 3-150
Table 3-15. Evidence Profile Table for PFOS Exposure and Cardiovascular Effects 3-180
Table 3-16. Associations Between Elevated Exposure to PFOS and Developmental
Outcomes in Children From Studies Identified in the 2016 PFOS HESD 3-192
Table 3-17. Evidence Profile Table for PFOS Exposure and Developmental Effects 3-257
Table 3-18. Incidences21 of Hepatocellular and Pancreatic Tumors in Male and Female
Sprague-Dawley Rats as Reported by Thomford {, 2002, 5029075} 3-273
Table 3-19. Mutagenicity Data From In Vivo Studies 3-279
xi
-------
APRIL 2024
Table 3-20. Mutagenicity Data From In Vitro Studies 3-280
Table 3-21. DNA Damage Data From In Vivo Studies 3-281
Table 3-22. DNA Damage Data From In Vitro Studies 3-282
Table 3-23. Evidence of Key Events Associated With the PPARa Mode of Action for
Hepatic Tumorsa in Male Sprague-Dawley Rats Exposed to PFOS 3-293
Table 3-24. Evidence of Key Events Associated With the PPARa Mode of Action for
Hepatic Tumorsa in Female Sprague-Dawley Rats Exposed to PFOS 3-294
Table 3-25. Evidence of Key Events Associated With the CAR Mode of Action for Hepatic
Tumorsa in Male Sprague-Dawley Rats Exposed to PFOS 3-298
Table 3-26. Evidence of Key Events Associated With the CAR Mode of Action for Hepatic
Tumorsa in Female Sprague-Dawley Rats Exposed to PFOS 3-298
Table 3-27. Evidence of Key Events Associated With the Cytotoxicity Mode of Action for
Hepatic Tumorsa in Male Sprague-Dawley Rats 3-300
Table 3-28. Evidence of Key Events Associated With the Cytotoxicity Mode of Action for
Hepatic Tumorsa in Female Sprague-Dawley Rats 3-300
Table 3-29. Incidences of Liver Tumor and Nonneoplastic Lesions in Male Sprague-
Dawley Rats at 103 Weeks, as Reported by Thomford {, 2002, 5029075} 3-301
Table 3-30. Incidences of Liver Tumor and Nonneoplastic Lesions in Female Sprague-
Dawley Rats at 103 Weeks, as Reported by Thomford {, 2002, 5029075} 3-302
Table 3-31. Comparison of the PFOS Carcinogenicity Database With the Likely Cancer
Descriptor as Outlined in the Guidelines for Carcinogen Risk Assessment {U.S.
EPA, 2005, 6324329} 3-307
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 4-16
Table 4-2. Benchmark Response Levels Selected for BMD Modeling of Health Outcomes.... 4-24
Table 4-3. PK Parameters from Wambaugh et al. {, 2013, 2850932} Meta-Analysis of
Literature Data for PFOS 4-28
Table 4-4. Model-Predicted and Literature PK Parameter Comparisons for PFOS 4-29
Table 4-5. Additional PK Parameters for Gestation/Lactation for PFOS 4-31
Table 4-6. Updated and Original Chemical-Specific Parameters for PFOS in Humans 4-35
Table 4-7. Summary of Studies Reporting the Ratio of PFOS Levels in Breastmilk and
Maternal Serum or Plasma 4-36
Table 4-8. PODheds Considered for the Derivation of Candidate RfD Values 4-39
Table 4-9. Uncertainty Factors for the Development of the Candidate Chronic RfD Values
from Epidemiological Studies {U.S. EPA, 2002, 88824} 4-56
xii
-------
APRIL 2024
Table 4-10. Uncertainty Factors for the Development of the Candidate Chronic RfD Values
From Animal Toxicological Studies {U.S. EPA, 2002, 88824} 4-58
Table 4-11. Candidate Reference Doses (RfDs) 4-60
Table 4-12. Cancer Slope Factors Derived From Results Reported by Butenhoff et al. {,
2012, 1276144}/Thomford {, 2002, 5029075}a in Sprague-Dawley Rats 4-71
Table 5-1. Correlation Coefficients Between PFOS and Other PFAS in Mutually Adjusted
Studies 5-5
Table 5-2. Impact of Co-Exposure Adjustment on Estimated Change in Mean Birth Weight
(Grams) per Unit Change (ng/mL) in PFOS Levels 5-8
Table 5-3. Comparison of Candidate RfDs Derived from Animal Toxicological Studies for
Priority Health Outcomes21 5-11
Table 5-4. Comparison of the PFOS Carcinogenicity Database with Cancer Descriptors as
Outlined in the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005,
6324329} 5-15
xiii
-------
APRIL 2024
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
Gastroenterol ogy
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
bile salt transporter
Cavg
average blood
concentration
AST
aspartate
aminotransferase
Cavg,pup,gest
area under the curve
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
xiv
-------
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
diacetate
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
ECso
half maximal effective
concentration
ECM
extracellular matrix
ESC-CM
embryonic stem cell-
derived cardiomyocyte
APRIL 2024
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
xv
-------
APRIL 2024
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
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
xvi
-------
APRIL 2024
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
xvii
-------
APRIL 2024
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
P milk
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
xviii
-------
APRIL 2024
TCR
T cell receptor
TG
triglycerides
THEMIS
thymocyte selection
associated
TLR
toll-like receptor
TLT
TREM-like transcript
cells
TNF
tumor necrosis factor
TNP
trinitrophenyl
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
subchronic 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
WHO World Health
Organization
WNT wingless-related
integration site
WoS Web of Science
WT wild type
WTCHR World Trade Center
Health Registry
ZFL zebrafish liver line
xix
-------
APRIL 2024
Executive Summary
The U.S. Environmental Protection Agency (EPA) is issuing final toxicity values for
perfluorooctane sulfonic acid (PFOS), including all isomers and nonmetal salts. The toxicity
assessment for PFOS is a scientific report that describes the evaluation of the available animal
toxicity and human epidemiology data in order to characterize the noncancer and cancer human
health hazards. This assessment also includes final toxicity values associated with noncancer
health effects (i.e., oral reference doses, or RfDs) and cancer effects (i.e., cancer slope factors, or
CSFs) following oral PFOS exposure. It is not a risk assessment, as it does not include an
exposure assessment or an overall risk characterization nor does it address the legal, policy,
social, economic, or technical considerations involved in risk management. The PFOS toxicity
assessment can be used by EPA, states, Tribes, and local communities, along with specific
exposure and other relevant information, to determine, under the appropriate regulations and
statutes, the potential risk associated with human exposures to PFOS, its isomers, and its
nonmetal salts.
This final toxicity assessment was peer reviewed by the EPA Science Advisory Board (SAB)
per- and polyfluoroalkyl substances (PFAS) Review Panel in November 2021 and underwent
public comment in March 2023. It incorporated expert scientific recommendations received from
the SAB in 2022 {U.S. EPA, 2022, 10476098} as well as feedback from the public comment
period {U.S. EPA, 2024, 11414326}. This final assessment builds upon the literature review
presented in the 2016 Health Effects Support Document for Perfluorooctane Sulfonic Acid
(PFOS) (hereafter referred to as the 2016 PFOS HESD) {U.S. EPA, 2016, 3603365} and is an
update of the SAB review draft, Proposed Approaches to the Derivation of a Draft Maximum
Contaminant Level Goal for Perfluorooctane Sulfonic Acid (PFOS) (CASRN1763-23-1) in
Drinking Water {U.S. EPA, 2022, 10668548} and the subsequent Public Comment Draft
Toxicity Assessment and Proposed Maximum Contaminant Level Goal for Perfluorooctane
Sulfonic Acid (PFOS) in Drinking Water {U.S. EPA, 2023, 10841010}.
PFOS is a member of the PFAS group. These manufactured chemicals have a history of
industrial and consumer use in the United States and are considered persistent chemicals based
on their physicochemical properties. Some of the human health concerns about exposure to
PFOS and other PFAS stem from their resistance to hydrolysis, photolysis, metabolism, and
microbial degradation in the environment and in the human body. PFAS are not naturally
occurring; they are manmade compounds that have been used widely over the past several
decades in industrial applications and consumer products since many PFAS have repellant and
surfactant properties. Frequently used as emulsifiers and as stain-, oil-, or water-repellents, PFAS
are found in a variety of environmental media and in tissues of organisms, including humans.
Most PFOS production in the United States was voluntarily phased out by its primary
manufacturer (3M) between 2000 and 2002. In 2002 and 2007, EPA took regulatory action under
the Toxic Substances Control Act (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
perfluorobutane sulfonic acid (PFBS) {3M, 4339178}. However, PFOS remains persistent in
environmental media because it is resistant to environmental degradation processes.
xx
-------
APRIL 2024
The purpose of this human health toxicity assessment is to derive toxicity values pertaining to
oral exposure for PFOS. The development of this toxicity assessment relied on a robust
systematic review process, based on the EPA peer-reviewed human health risk assessment
methodology outlined in the EPA ORD Staff Handbook for Developing IRIS Assessments {U.S.
EPA, 2022, 10367891}, to identify human epidemiological, animal toxicological, mechanistic,
and toxicokinetic data relevant to oral exposure. The PFOS systematic review protocol (see
Appendix A, {U.S. EPA, 2024, 11414344}) was developed prior to the initiation of this
assessment largely mirrors the Systematic Review Protocol for the PFBA, PFHxA, PFHxS,
PFNA, andPFDA (Anionic and Acid Forms) IRIS Assessments {U.S. EPA, 2020, 8642427}. The
protocol outlines the scoping and problem-formulation efforts and describes the systematic
review, including study quality evaluation, and the dose-response methods used to conduct this
assessment. The final assessment incorporates peer-reviewed studies captured from: EPA's 2016
PFOS HESD {U.S. EPA, 2016, 3603365}, literature searches of scientific databases and gray
literature from 2013 through February 2023, the SAB PFAS Review Panel recommendations,
and public comment. Consistent with the analysis provided in the peer-reviewed draft assessment
{U.S. EPA, 2022, 10668548} and with recommendations from external peer review (i.e., the
SAB PFAS Review Panel; {U.S. EPA, 2022, 10476098}), this final assessment focused on
qualitative and quantitative assessments of five "priority" health outcome categories based on
those with the strongest weight of evidence. These five priority health outcomes are cancer,
hepatic, developmental, cardiovascular, and immune. The results of the systematic literature
reviews and qualitative assessments for the remaining "nonpriority" health outcomes are
presented in the Appendix accompanying this final assessment {U.S. EPA, 2024, 11414344}.
Qualitative Assessment of Noncancer Effects
Overall, the available evidence indicates that PFOS exposure is likely to cause hepatic,
immunological, cardiovascular, and developmental effects in humans given sufficient exposure
conditions (e.g., at measured levels in humans as low as 0.57 to 5.0 ng/mL and at administered
doses in animals as low as 0.0017 to 0.4 mg/kg/day). These judgments are based on data from
epidemiological studies of infants, children, adolescents, pregnant individuals, and nonpregnant
adults, as well as short-term (28-day), subchronic (90-day), developmental (gestational), and
chronic (2-year) oral-exposure studies in rodents. For hepatic effects, the primary support is
evidence of increased serum liver enzyme levels (i.e., alanine transaminase (ALT)) in humans
and coherent evidence of hepatotoxicity in animals, including increased liver weights and
hepatocellular hypertrophy accompanied by necrosis, inflammation, or increased liver enzyme
levels that indicate liver injury. For immunological effects, the primary support is evidence of
developmental immunosuppression in humans, specifically decreased antibody response to
vaccination against tetanus, diphtheria, and rubella in children, and evidence of
immunosuppression and other types of immunotoxicity in studies of adult animals, including
decreased plaque forming cell response to sheep red blood cells, extramedullary hematopoiesis
in the spleen, reduced spleen and thymus weights, changes in immune cell populations, and
decreased splenic and thymic cellularity. For cardiovascular effects, the primary support is
evidence of increased serum lipids levels in humans and alterations to lipid homeostasis in
animals. For developmental effects, the primary evidence is decreased birth weight in human
infants and decreased fetal and maternal weight in animal studies. According to the protocol
described in Appendix A {U.S. EPA, 2024, 11414344} and aligned with EPA peer-reviewed
human health risk assessment methodology {U.S. EPA, 2022, 10367891}, selected quantitative
xxi
-------
APRIL 2024
data in medium and high confidence studies from these identified hazards were used to derive
toxicity values (see Table ES-1). Specific criteria for data and study selection are provided in
Appendix A {U.S. EPA, 2024, 11414344} and Section 4.1.
Quantitative Assessment of Noncancer Effects and Oral RfD
Derivation
EPA followed agency guidelines and methodologies for risk assessment in determining points of
departure (PODs) for the derivation of the RfDs for PFOS {U.S. EPA, 2022, 10367891; U.S.
EPA, 2002, 88824; U.S. EPA, 2012, 1239433; U.S. EPA, 2011, 786546; U.S. EPA, 2014,
2520260} and performed modeling following EPA's Benchmark Dose Technical Guidance
Document {U.S. EPA, 2012, 1239433}. For data from epidemiological studies, the dose-
response modeling approach was selected based on the health outcome and available data. A
hybrid modeling approach, which estimated the probability of responses at specified exposure
levels above the control, was conducted when clinically adverse outcome levels could be defined
(i.e., for developmental, hepatic, and cardiovascular effects) following EPA's Benchmark Dose
Technical Guidance Document {U.S. EPA, 2012, 1239433}. For other outcomes (i.e., immune
effects), study results from multivariate models were used to define a benchmark response
(BMR). For data from animal toxicological studies, EPA conducted benchmark dose modeling,
when possible, to empirically model the dose-response relationship in the range of observed data.
When BMDLs could not be derived, EPA used a no-observed-adverse-effect level/lowest-
observed-adverse-effect level (NOAEL/LOAEL) approach.
PODs were converted to external POD human equivalent doses (PODheds) using
pharmacokinetic modeling (see Section 4.1.3). Consistent with the recommendations presented
in EPA's A Review of the Reference Dose and Reference Concentration Processes {U.S. EPA,
2002, 88824}, EPA considered the database of information to inform the application of
uncertainty factors (UFs) to PODheds to address intraspecies variability, interspecies variability,
extrapolation from a LOAEL to NOAEL, extrapolation from a subchronic to a chronic exposure
duration, and database deficiencies. EPA derived and considered multiple candidate RfDs from
both human epidemiological and animal toxicological studies across the four priority noncancer
health outcomes that EPA determined had the strongest weight of evidence (i.e., immune,
cardiovascular, hepatic, and developmental) (see Figure ES-1 for candidate RfD values).
Additional details on candidate RfD derivation for PFOS are available in Section 4.1.
xxii
-------
APRIL 2024
Immune ,
Decreased serum
anti-tetanus antibody
concentration in children
Timmerman,
2021, 9416315;
Medium confidence
Budtz-Jorgensen,
2018, 5083631;
Medium confidence
Decreased serum
anti-diptheria antibody
concentration in children
Timmerman,
2021, 9416315;
Medium confidence
Budtz-j0rgensen,
2018, 5083631;
Medium confidence
Decreased serum
anti-rubella antibody
concentration in adolescents
Zhang, 2023, 10699594;
Medium confidence
Extramedullar
hematopoiesis in the spleen
NTP.2019, 5400978;
High confidence
Decreased PFC response
to SRBC
Zhong, 2016, 3748828;
Medium confidence
Human | Animal
•—o
RfD PODHED
*-uT°
•—o
•—o
•—o
•—o
• o
Developmental
L
'
•—o
•—o
•—o
Sagiv, 2018, 4238410;
High confidence
Wikstrom, 2020,
Decreased Birth Weight 6311677;
High confidence
Darrow, 2013, 2850966;
High confidence
Decreased Pup Body Luebker, 2005, 757857;
Weight Medium confidence
• o
t
Cardiovascular
r
•—o
•—o
I
Dong, 2019, 5080195;
Medium confidence
Increased Serum Total
Cholesterol Steenland,
2009, 1291109;
Medium confidence
Hepatic
•—o
•—o
Gallo, 2012, 1276142;
Medium confidence
Increased Serum ALT
Nian, 2019, 5080307;
Medium confidence
Butenhoff, 2012,
Individual Cell Necrosis in 1276144/ Thomford,
the Liver 2002, 5029075;
High confidence
• o
10-8 10-7 10-6 -10-5 10-4 10-3 10-2
PFOS Concentration (mg/kg-d)
Figure ES-1. Schematic Depicting Candidate RfDs Derived From Epidemiological and
Animal Toxicological Studies of PFOS
See text and Figure 4-3 in Section 4.1 for additional detail on dose-response modeling for PFOS studies.
xxiii
-------
APRIL 2024
The co-critical effects for the oral RfD of 1 x 10 7 mg/kg/day were decreased infant birth weight
{Wikstrom et al., 2020, 6311677} and increased total cholesterol in adults {Dong et al., 2019,
5080195} (see Table ES-1). These co-critical effects were selected based on the procedures
outlined in the protocol (see Appendix A, {U.S. EPA, 2024, 11414344}) and were consistent
with EPA peer-reviewed human health risk assessment methodology {U.S. EPA, 2022,
10367891}. The RfD was derived by using a total UF of 10 to account for intraspecies variability
(UFh). Notably, the RfD is protective of effects that may occur in sensitive populations (i.e.,
embryo and fetus, infants, and young children), as well as hepatic effects in adults 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.
Qualitative Carcinogenicity Assessment
Consistent with EPA's Guidelines for Carcinogen Risk Assessment {U.S. EPA, 2005, 6324329},
EPA reviewed the available data and conducted a weight of evidence evaluation across the
human epidemiological and animal toxicological studies and concluded that PFOS is Likely to Be
Carcinogenic to Humans via the oral route of exposure (see Section 3.5). Epidemiological
studies provided evidence of bladder, prostate, liver, kidney, and breast cancers in humans,
although evidence was limited or mixed for some cancer types. Animal toxicological studies
supported findings from human studies. Bioassays conducted in Sprague-Dawley rats reported
hepatocellular tumors, pancreatic islet cell tumors, and thyroid follicular cell tumors after chronic
oral exposure. Some studies observed multisite tumorigenesis (liver and pancreas) in male and
female rats. PFOS exposure is associated with multiple key characteristics of carcinogenicity
{Smith, 2016, 3160486}. Available mechanistic data suggest that multiple modes of action
(MOAs) play a role in pancreatic and hepatic tumorigenesis associated with PFOS exposure in
animal models. A full MOA analysis, including in-depth discussions on the potential MOAs for
kidney and testicular tumors, as well as discussions on the potential MOAs and human relevance
for pancreatic and liver tumors observed in rats, is presented in Section 3.5.4.2.
Quantitative Cancer Assessment and CSF Derivation
EPA followed agency guidelines for risk assessment in deriving CSFs for PFOS {U.S. EPA,
2012, 1239433; U.S. EPA, 2022, 10367891; U.S. EPA, 2005, 6324329}. EPA selected medium
and high confidence studies for derivation that met criteria outlined in the protocol (see
Appendix A, {U.S. EPA, 2024, 11414344}) and Section 4.1.1, conducted benchmark dose
modeling {U.S. EPA, 2012, 1239433}, and used the same pharmacokinetic modeling approach
as described for the derivation of noncancer RfDs above (see Section 4.2.2). Data from
epidemiological studies were not suitable for CSF derivation. From the studies that met the
criteria, EPA used multistage models to derive and consider multiple candidate CSFs from
animal toxicological studies across multiple tissue types or organ systems (i.e., liver and
pancreas). Multistage cancer models were used to predict the doses at which the selected BMR
for tumor incidence would occur. BMDLs for each tumor type served as the PODs, which were
then converted to PODheds by applying the human clearance value. Candidate CSFs were then
calculated by dividing the selected BMR by the PODheds for each tumor type.
xxiv
-------
APRIL 2024
The oral slope factor of 39.5 (mg/kg/day) 1 for hepatocellular adenomas and carcinomas in
female rats from Butenhoff et al. {, 2022, 1276144}/Thomford {, 2002, 5029075} was selected
as the basis of the overall CSF for PFOS (see Table ES-1; rationale in Section 4.2). Per EPA's
Guidelines for Carcinogen Risk Assessment and Supplemental Guidance for Assessing
Susceptibility from Early-Life Exposure to Carcinogens {U.S. EPA, 2005, 6324329; U.S. EPA
2005, 88823}, age-dependent adjustment factors were not applied during CSF derivation because
there was a lack of information to support a mutagenic MOA for PFOS, and the available
evidence was insufficient to assess susceptibility to cancer following PFOS exposure during
early life. Additional detail on candidate CSF derivation and CSF selection is provided in Table
4-12 in Section 4.2.
Final Toxicity Values for PFOS
Table ES-1. Final Toxicity Values for PFOS
Toxicity
Value Type
Critical Effect(s) Study, Confidence Strain/Species, Sex, Age
Toxicity Value3
Reference
Dose
Cancer Slope
Factor
Co-critical effects:
decreased birth weight in
infants;
increased serum total
cholesterol in adults
Combined hepatocellular
adenomas and
carcinomas
Wikstrom et al. {,
2020, 6311677},
High',
Dong et al. {, 2019,
5080195}, Medium
Butenhoff et al. {,
2012,
1276144}/Thomford
{,2002, 5029075}b,
High
Human, male and female,
PFOS concentrations in
first and second trimesters;
Human, male and female,
20-80 years
Sprague-Dawley rats,
female
1x10 7(mg/kg/d)
39.5 (mg/kg/d) 1
Notes:
a Reference doses were rounded to one significant figure.
bButenhoff et al. {, 2012, 1276144} and Thomford {, 2002, 5029075} reported data from the same experiment.
XXV
-------
APRIL 2024
1 Background
1.1 Purpose of This Document
The primary purpose of this toxicity assessment for perfluorooctane sulfonic acid (PFOS) is to
describe the best available science on the human health effects associated with PFOS exposure
and the derivation of toxicity values (i.e., noncancer reference doses (RfDs) and cancer slope
factors (CSFs)). The latest health science on PFOS was identified, evaluated using systematic
review methods, and described, and subsequently, a cancer classification was assigned and
toxicity values were developed. The final cancer classification and cancer and noncancer toxicity
values in this assessment build on the work described in the Public Comment Draft Toxicity
Assessment and Proposed Maximum Contaminant Level Goal for Perfluorooctane Sulfonic Acid
(PFOS) in Drinking Water {U.S. EPA, 2021, 10841010}, Proposed Approaches to the
Derivation of a Draft Maximum Contaminant Level Goal for Perfluorooctane Sulfonic Acid
(PFOS) (CASRN1763-23-1) in Drinking Water {U.S. EPA, 2021, 10428576}, and the Health
Effects Support Document for Perfluorooctane Sulfonate (PFOS) {U.S. EPA, 2016, 3603365}.
This final toxicity assessment for PFOS reflects expert scientific recommendations from the U.S.
Environmental Protection Agency (EPA) Science Advisory Board (SAB) {U.S. EPA, 2022,
10476098} and public comments received on the draft assessment
(https://www.regulations.gov/docketA " \ s IQ-OW-2Q2J Oil I; U.S. EPA {, 2024, 1141432}).
In addition to documenting EPA's basis for the cancer classification and toxicity values, this
document serves to:
• Describe and document transparently the literature searches conducted and systematic
review methods used to identify health effects information (epidemiological and animal
toxicological studies and physiologically based pharmacokinetic models) in the literature
(Sections 2 and 3; Appendices A andB, {U.S. EPA, 2024, 11414344}).
• Describe and document literature screening methods, including use of the Populations,
Exposures, Comparators, and Outcomes (PECO) criteria and the process for tracking
studies throughout the literature screening (Section 2; Appendix A, {U.S. EPA, 2024,
11414344}).
• Identify epidemiological and animal toxicological literature that reports health effects
after exposure to PFOS (and its related salts) as outlined in the PECO criteria (Section 3).
• Describe and document the study quality evaluations conducted on epidemiological and
animal toxicological studies considered potentially useful for point-of-departure (POD)
derivation (Section 3).
• Describe and document the data from all epidemiological studies and animal toxicological
studies that were considered for POD derivation (Section 3).
• Synthesize and document the adverse health effects evidence across studies. The
assessment focuses on synthesizing the available evidence for five priority health
outcomes that were found to have the strongest weight of evidence, as recommended by
the SAB - developmental, hepatic, immune, and cardiovascular effects, and cancer
(Section 3) - and also provides supplemental syntheses of evidence for dermal, endocrine,
gastrointestinal, hematologic, metabolic, musculoskeletal, nervous, ocular, renal, and
1-1
-------
APRIL 2024
respiratory effects, reproductive effects in males or females, and general toxicity
(Appendix C, {U.S. EPA, 2024, 11414344}).
• Evaluate and document the available mechanistic information (including toxicokinetic
understanding) associated with PFOS exposure to inform interpretation of findings related
to potential health effects in studies of humans and animals, with a focus on five priority
health outcomes (developmental, hepatic, immune, and cardiovascular effects, and cancer)
(Section 3).
• Develop and document strength of evidence judgments across studies (or subsets of
studies) separately for epidemiological, animal toxicological, and mechanistic lines of
evidence for the five priority health outcomes (Section 3).
• Develop and document integrated expert judgments across evidence streams (i.e.,
epidemiological, animal toxicological, and mechanistic streams) as to whether and to what
extent the evidence supports that exposure to PFOS has the potential to be hazardous to
humans (Section 3).
• Determine the cancer classification for PFOS using a weight-of-evidence approach
(Section 3.5.5).
• Describe and document the attributes used to evaluate and select studies for derivation of
toxicity values. These attributes are considered in addition to the study confidence
evaluation domains and enable extrapolation to relevant exposure levels (e.g., studies with
exposure levels near the range of typical environmental human exposures, broad exposure
range, or multiple exposure levels) (Section 4).
• Describe and document the dose-response analyses conducted on the studies identified for
POD derivation (Section 4).
• Derive candidate RfDs (Section 4.1) and CSFs (Section 4.2), select the final RfD (Section
4.1.6) and CSF (Section 4.2.3) for PFOS, and describe the rationale.
• Characterize hazards (e.g., uncertainties, data gaps) (Sections 3, 4, and 5).
1.2 Background on Per-and Polyfluoroalkyl Substances
Per- and polyfluoroalkyl substances (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. There is no
consensus definition of PFAS as a class of chemicals {OSTP, 2023, 11396268}. Consistent with
three related structural definitions associated with EPA's identification of PFAS included in the
fifth Contaminant Candidate List1 (CCL), the universe of environmentally relevant PFAS -
including parent chemicals, metabolites, and degradants - is approximately 15,000 compounds.2
The 2018 Organisation for Economic Co-operation and Development (OECD) New
Comprehensive Global Database of Per- and Polyfluoroalkyl Substances (PFASs) 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 1950's. PFAS have strong, stable carbon-fluorine (C-F)
1 The CCL is a list, published 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, and might require regulation under SDWA.
2 See the EPA List of PFAS Structures available at: https://comptox.epa.gov/dasliboaiil/clieniical-lists/PFASSTRIJCT.
1-2
-------
APRIL 2024
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 enable them to repel water and oil, remain chemically and thermally stable,
and exhibit surfactant properties. These properties make PFAS useful for commercial and
industrial applications and make many PFAS extremely persistent in the human body and the
environment {Calafat, 2007, 1290899; Calafat, 2019, 5381304; Kwiatkowski, 2020, 7404231}.
Because of their widespread use, physicochemical properties, persistence, and bioaccumulation
potential, many different 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.
With regard to 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. PFOS belongs to the PFAA
family of the non-polymer PFAS category and is among the most researched PFAS in terms of
human health toxicity and biomonitoring studies (for review, see Podder et al. {, 2021,
9640865}).
1.3 Chemical Identity
PFOS is a perfluoroalkyl sulfonate 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, oceangoing 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.
1-3
-------
APRIL 2024
For a more detailed discussion related to the chemical and physical properties and environmental
fate of PFOS, please see the PFAS Occurrence and Contaminant Background Support Document
for the Final PFAS National Primary Drinking Water Regulation{ U.S. EPA, 2024, 11414328},
the 2016 PFOS Health Effects Support Document {U.S. EPA, 2016, 3603365}, and the Draft
Aquatic Life Ambient Water Quality Criteria for Perfluorooctane Sulfonate (PFOS) {U.S. EPA,
2022, 10668582}.
Table 1-1. Chemical and Physical Properties of PFOS
Property
PFOS, Acidic Form;
Experimental Average
Source
Chemical Abstracts Service Registry
1763-23-1
{NLM, 2022, 10369707@@author-
Number (CASRN)3
year}
Chemical Abstracts Index Name
1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,8-
Heptadecafluoro-l-octanesulfonic acid
Synonyms
Perfluorooctane sulfonic acid;
EPA CgitidTgx Chemicals
heptadecafluoro-1-octane sulfonic
Dashboard
acid; PFOS acid
Chemical Formula
CgHFnCbS
{NLM, 2022, 10369707@@author-
year}
Molecular Weight
500.13 g/mol
{NLM, 2022, 10369707@@author-
year}
Color/Physical State
Liquid
{NLM, 2022, 10369707@@author-
year}
Boiling Point
249°C
{NLM, 2022, 10369707@@author-
year}
Melting Point
>400°C
{ATSDR, 2021, 9642134@@author-
year} (potassium salt)
Vapor Pressure
0.002 mm Hg at 25°C
{NLM, 2022, 10369707@@author-
year} (estimated)
Henry's Law Constant (KH)
4.1E-04 atm-m3/mol at 25°C
{NLM, 2022, 10369707@@author-
year} (estimated from vapor pressure
and water solubility)
Koc
1,000 ± 5.0 L/kg (mean of
{Zareitalabad, 2013,
values ± 1 standard deviation of
508056 l@@author-year} (converted
selected values)
from log Koc to Koc)
Log Kow
4.49
{NLM, 2022, 10369707@@author-
year} (estimated)
Solubility in Water
0.0032 mg/L at 25°C;
{NLM, 2022, 10369707@@author-
570 mg/L
year} (estimated)
{ATSDR, 2021, 9642134@@author-
year} (potassium salt in pure water)
Notes: CASRN = Chemical Abstracts Service Registry Number; Koc = organic carbon-water partitioning coefficient;
Kow = octanol-water partition coefficient.
aThe 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.
1.4 Occurrence Summary
1.4.1 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
1-4
-------
APRIL 2024
the general U.S. population since 1999. PFOS has 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 PFOS declined by >85% between 1999 and 2018, presumably because of restrictions on
its commercial usage in the United States {CDC, 2017, 4296146}. However, studies of residents
in locations of suspected PFAS contamination show higher serum levels of PFAS, including
PFOS, compared with the general U.S. population as reported by NHANES {Kotlarz, 2020,
6833715; Yu, 2020, 63 15796; Table 17-6 in \ITRC\ 2022, 9959768; ATS DR., 2022, 10519308}.
Most PFOS production in the United States was voluntarily phased out by its primary
manufacturer (3M) between 2000 and 2002. In 2002 and 2007, EPA took regulatory action under
the Toxic Substances Control Act (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
perfluorobutane sulfonic acid (PFBS) {3M, 2002, 4339178}. 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-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.
1.4.2 Ambient Water
Among the PFAS with established analytical methods for detection, PFOS is one of the
dominant PFAS compounds detected in ambient water both in the United States and worldwide
{Ahrens, 2011, 2657780; Benskin, 2012, 1274133; Dinglasan-Panlilio, 2014, 2545254;
Nakayama, 2007, 2901973; Remucal, 2019, 5413103; Zareitalabad, 2013, 5080561}. Although 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 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). Although 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. Figure 1-1 (excerpted from {Jarvis, 2021, 9416544@@author-year}) shows the
distribution of PFOA concentrations (ng/L) measured in surface waters for each U.S. state or
waterbody (excluding the Great Lakes) with reported data in the publicly available literature.
1-5
-------
APRIL 2024
1000000
100000
10000
C/3
o
u.
&
5
re
•J
o
in
1000
100
10
0.1
0.01
0.001
1
1
II
u
0
f—1
~ n
1
-1
AL CA CO DE FL GA LA MI MN NJ NM NY NC RI SC IN IX WA
River
Figure 1-1. Distribution of PFOS Concentrations in Surface Waters by State/Waterbody
(Excluding Great Lakes) in the United States {Jarvis, 2021, 9416544}
1.4.3 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.
EPA uses the Unregulated Contaminant Monitoring Rule (UCMR) to collect data for
contaminants that are suspected to be present in drinking water and do not have health-based
standards set under the Safe Drinking Water Act (SDWA). Under the UCMR, drinking water is
monitored from public water systems (PWSs), specifically community water systems and non-
transient, non-community water systems. The UCMR improves EPA's understanding of the
frequency and concentrations of contaminants of concern occurring in the nation's drinking
water systems. The first four UCMRs collected data from a census of large water systems
(serving more than 10,000 people) and from a statistically representative sample of small water
systems (serving 10,000 or fewer people). UCMR 3 monitoring occurred between 2013 and 2015
and is currently the most comprehensive nationally representative finished water dataset for
PFOS {USEPA, 2024, 11414345; USEPA, 2024, 11414328}. Under UCMR 3, 36,972 samples
from 4,920 PWSs were analyzed. PFOS was found in 292 samples at 95 systems above the
UCMR 3 minimum reporting level (40 ng/L). These systems serve a population of approximately
10.4 million people located in 28 states, Tribes, or U.S. territories (USEPA, 2024, 11414345;
USEPA, 2024, 11414328}.
More recent state data were collected using newer EPA-approved analytical methods and some
state results reflect lower reporting limits than those in the UCMR 3. State data are available
from 32 states: Alabama, Arizona, California, Colorado, Delaware, Georgia, Idaho, Illinois,
1-6
-------
APRIL 2024
Indiana, Iowa, Kentucky, Maine, Maryland, Massachusetts, Michigan, Minnesota, Missouri,
New Hampshire, New Jersey, New Mexico, New York, North Carolina, North Dakota, Ohio,
Oregon, Pennsylvania, South Carolina, Tennessee, Vermont, Virginia, West Virginia, and
Wisconsin {USEPA, 2024, 11414345; USEPA, 2024, 11414328}. State results show continued
occurrence of PFOS in multiple geographic locations. These data also show PFOS occurrence at
lower concentrations and significantly greater frequencies than were measured under the UCMR
3, likely because the more recent monitoring was able to rely on more sensitive analytical
methods {USEPA, 2024, 11414345; USEPA, 2024, 11414328}. More than one-third of states
that conducted nontargeted monitoring detected PFOA and/or PFOS at more than 25% of
systems {USEPA, 2024, 11414345; USEPA, 2024, 11414328}. Among the detections, PFOS
concentrations ranged from 0.24 to 650 ng/L with a range of median concentrations from 1.21 to
12.1 ng/L {USEPA, 2024, 11414345; USEPA, 2024, 11414328}. Monitoring data for PFOA and
PFOS from states that conducted targeted monitoring efforts, including 15 states, demonstrate
results consistent with the nontargeted state monitoring. Within the 20 states that conducted
nontargeted monitoring, there are 1,260 systems with results above 4.0 ng/L and 1,577 systems
with results above 4.0 ng/L {USEPA, 2024, 11414345; USEPA, 2024, 11414328}. These
systems serve populations of 12.5 and 14.4 million people, respectively. Monitoring data for
PFOS from states that conducted targeted sampling efforts showed additional systems exceeding
4 ng/L {USEPA, 2024, 11414345; USEPA, 2024, 11414328}.
Finally, the fifth UCMR (UCMR 5) was published in December 2021 and requires sample
collection and analysis for 29 PFAS, including PFOS, between January 2023 and December
2025 using drinking water analytical methods developed by EPA {USEPA, 2021, 11374428}.
The UCMR 5 defined the minimum reporting level at 4 ng/L for PFOS using EPA Method 533,
which is lower than the 40 ng/L used in the UCMR 3 with EPA Method 537 {USEPA, 2021,
11374428}. Therefore, the UCMR 5 will be able to provide nationally representative occurrence
data for PFOS at lower detection concentrations. While the complete UCMR 5 dataset is not
currently available, the small subset of data released (7% of the total results that EPA expects to
receive) as of July 2023 is consistent with the results of UCMR 3 and the state data described
above {USEPA, 2024, 11414345; USEPA, 2024, 11414328}.
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.
1.5 History of EPA's Human Health Assessment for PFOS
EPA developed an HESD for PFOS after it was listed on the third CCL (CCL 3) in 2009 {U.S.
EPA, 2009, 1508321}. An HESD is synonymous with a toxicity assessment in that they both
describe the assessment of cancer and noncancer health effects and derive toxicity values. The
2016 PFOS HESD was peer reviewed in 2014 and revised based on consideration of peer
reviewers' comments, public comments, and additional studies published through December
2015. The resulting Health Effects Support Document for Perfluorooctane Sulfonic Acid (PFOS)
{U.S. EPA, 2016, 3603365} was published in 2016 and described the assessment of cancer and
noncancer health effects and the derivation of a noncancer RfD for PFOS.
1-7
-------
APRIL 2024
EPA initiated an update to the 2016 PFOS HESD in 2021 when the agency made a determination
to regulate PFOS with a national primary drinking water regulation (NPDWR) {U.S. EPA, 2021,
9640861}. The initial update of the 2016 PFOS HESD was the Proposed Approaches to the
Derivation of a Draft Maximum Contaminant Level Goal for Perfluorooctane Sulfonic Acid
(PFOS) (CASRN 335-67-1) in Drinking Water {U.S. EPA, 2021, 10428576}. This assessment
described the systematic review of cancer and noncancer health effects, the derivation of
candidate oral cancer and noncancer toxicity values, a relative source contribution (RSC), and
cancer classification, which would subsequently be used to prepare draft and final toxicity
assessments. The agency sought peer review from the EPA SAB PFAS Review Panel on key
scientific issues, including the systematic review approach for evaluating health effects studies,
the derivation of oral toxicity values, the RSC, and the cancer classification for PFOS.
The SAB provided draft recommendations on June 3, 2022, and final recommendations on
August 23, 2022 {U.S. EPA, 2022, 10476098}. To be responsive to the SAB recommendations,
EPA developed a detailed response to comment document {USEPA OW, 2023, 11396269} and
addressed every recommendation from the SAB in the development of the Public Comment
Draft Toxicity Assessment and Proposed Maximum Contaminant Level Goal for Perfluorooctane
Sulfonic Acid (PFOS) in Drinking Water {U.S. EPA, 2023, 10841010}. Briefly, EPA:
• 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 across health outcomes;
• 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 discussion for cancer effects and rationale for the
cancer classification;
• strengthened the rationales for selection of PODs for the noncancer health outcomes; and
• clarified language related to the RSC determination, including the relevance of drinking
water exposures and the relationship between the RfD and the RSC.
EPA then released the Public Comment Draft Toxicity Assessment and Proposed Maximum
Contaminant Level Goal for Perfluorooctane Sulfonic Acid (PFOS) in Drinking Water for a 60-
day public comment period. These assessments described the systematic review of cancer and
noncancer health effects, the derivation of candidate oral cancer and noncancer toxicity values,
an RSC, and cancer classification for PFOS.
EPA incorporated feedback from public comment into the assessment and developed a detailed
response to public comment document {USEPA, 2024, 11414326}. Briefly, EPA has improved
descriptions of rationale and added clarifications related to the systematic review protocol used
for this assessment, study and endpoint selection for POD derivation, and the modeling choices
related to toxicity value derivation. Therefore, this Final Human Health Toxicity Assessment for
Perfluorooctane Sulfonic Acid (PFOS) and Related Salts incorporates feedback from external
1-8
-------
APRIL 2024
peer review and public comment and supersedes all other health effects documents produced by
the EPA Office of Water for PFOS.
1-9
-------
APRIL 2024
2 Summary of Assessment Methods
This section summarizes the methods used for the systematic review of the health effects
literature for 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). The
purposes of this systematic review were to identify the best available and most relevant health
effects literature, to evaluate studies for quality, and to subsequently identify health effects and
studies for dose-response assessment. A detailed description of these methods is provided as a
protocol in Appendix A, {U.S. EPA, 2024, 11414344}.
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 {U.S. EPA, 2022, 11181062; U.S. EPA, 2023,
11133619}. Though the IRIS Handbook was finalized concurrently with the development of this
assessment, the revisions in the final IRIS Handbook compared to the draft version do not
conflict with the methods used in this assessment. The assessment team concluded that
implementing these minor changes in study quality evaluation between the draft and final IRIS
Handbook versions 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. Additionally, the methods used to conduct the systematic review are also consistent
with and largely mirror the Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA,
andPFDA (anionic and acidforms) IRIS Assessments {U.S. EPA, 2020, 8642427}.
For this updated PFOS toxicity assessment, systematic review methods were consistent with
those in the IRIS Handbook {U.S. EPA, 2022, 10367891} and the Systematic Review Protocol
for the PFBA, PFHxA, PFHxS, PFNA, and PFDA (anionic and acid forms) IRIS Assessments
{U.S. EPA, 2020, 8642427} for the steps of literature search; screening; study quality
evaluation; data extraction; display of study evaluation results; synthesis of human and
experimental animal data; and evidence integration for all health outcomes through the 2020
literature searches, as presented in the preliminary analyses of the 2021 Proposed Approaches To
The Derivation Of A Draft Maximum Contaminant Level Goal For Perfluorooctane Sulfonic
Acid (PFOS) (CASRN1763-23-1) In Drinking Water draft document that was reviewed by the
Science Advisory Board (SAB) {U.S. EPA, 2021, 10428576; U.S. EPA, 2022, 10367891}. The
EPA then focused the remaining steps of the systematic review process (synthesis and
integration of mechanistic data; derivation of toxicity values) on health outcomes with the
strongest weight of evidence based on the conclusions presented in the 2021 draft documents,
and consistent with the recommendations of the SAB {U.S. EPA, 2022, 10476098}. These five
"priority" health outcomes are developmental, hepatic, immune, cardiovascular, and cancer. The
2-1
-------
APRIL 2024
updated systematic review focused on the priority health outcomes was published in 2023 as the
Public Comment Draft Toxicity Assessment and Proposed Maximum Contaminant Level Goal
for Perfluorooctane Sulfonic Acid (PFOS) inDrinking Water {U.S. EPA2023, 10841010}.
The following subsections provide a summary of methods used to search for and screen
identified literature, evaluate the identified studies to characterize study quality, extract data, and
select studies for dose-response analysis. Extracted data are available in interactive visual
formats (see Section 3) and can be downloaded in open access, interactive formats. The full
systematic review protocol (see Appendix A, {U.S. EPA, 2024, 11414344}) provides a detailed
description of the systematic review methods that were used. The protocol also includes the
description of the problem formulation and key science issues guiding this assessment.
2.1.1 Literature Database
The EPA assembled a database of epidemiological, animal toxicological, mechanistic, and
toxicokinetic studies for this PFOS toxicity assessment based on three main data streams: 1)
literature published from 2013 through February 6, 2023 identified via literature searches
conducted in 2019, 2020, 2022 and 2023 of a variety of publicly available scientific literature
databases, 2) literature identified via other sources (e.g., searches of the gray literature, studies
shared with EPA by the SAB, studies submitted through public comment), and 3) literature
identified in EPA's 2016 Health Effects Support Document for Perfluorooctane Sulfonic Acid
(PFOS) {U.S. EPA, 2016, 3603365}. All of these streams are described in detail below.
For the literature searches, the search strings focused on the chemical name (PFOS and its related
salts) with no limitations on lines of evidence (i.e., human/epidemiological, animal, in vitro, in
silico) or health outcomes. The EPA conducted a 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 prior to SAB review of the draft assessment (2020 literature search),
a third search covering September 2020 through February 3, 2022 prior to release of the draft
assessment for public comment (2022 literature search), and a final supplemental search
covering February 4, 2022 through February 6, 2023.
The publicly available databases listed below were searched for literature containing the
chemical search terms outlined in Appendix A {U.S. EPA, 2024, 11414344}:
• 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).
The search strings and literature sources searched are described in Appendix A {U.S. EPA, 2024,
11414344}.
For the second data stream, other review efforts and searches of publicly available sources were
used to identify relevant studies (see Appendix A, {U.S. EPA, 2024, 11414344}), as listed
below:
• studies cited in assessments published by other U.S. federal, international, and/or U.S.
state agencies (this included assessments by ATSDR {ATSDR, 2021, 9642134} and
2-2
-------
APRIL 2024
California Environmental Protection Agency {CalEPA, 2021, 9416932}),
• studies identified during mechanistic or toxicokinetic evidence 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),
• studies identified by the SAB in their final report dated August 23, 2022 {U.S. EPA,
2022, 10476098}, and
• studies submitted through public comment by May 2023
(https://www.regulations.gov/docket/EPA-HQ-OW-2022-0114).
For the third data stream, EPA relied on epidemiological and animal toxicological literature
synthesized in the 2016 PFOS HESD to identify studies relevant to the five priority health
outcomes, as recommended by SAB and consistent with preliminary conclusions from EPA's
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 PFOS HESD contained a summary of all relevant literature identified in searches
conducted through 2013. EPA's 2016 PFOS HESD relied on animal toxicological studies for
quantitative analyses whereas epidemiology studies were considered qualitatively, as a
supporting line of evidence. This updated assessment includes epidemiological studies that were
identified and presented in the 2016 PFOS HESD for the five priority health outcomes. It also
includes "key" animal toxicological studies from the 2016 PFOS HESD, which includes studies
that were selected in 2016 for dose-response modeling. The details of the studies included from
the 2016 PFOS HESD are described in Appendix A {U.S. EPA, 2024, 11414344}.
All studies identified through the data streams outlined above were uploaded into the publicly
available Health and Environmental Research Online (HERO) database
(https://hero.epa.eov/hero/index.cfm/proiect/paee/proiect id/2608).
EPA has continued to monitor the literature published since February 2023 for other potentially
relevant studies. Potentially relevant studies identified after February 2023 that were not
recommended by the SAB in their final report or via public comment 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 Appendix A, {U.S. EPA,
2024, 11414344}).
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, the EPA used
populations, exposures, comparators, and outcomes (PECO) criteria to screen the literature
identified from the literature sources outlined above in order to prioritize studies for dose-
response assessment and to identify studies containing supplemental information such as
mechanistic studies that could inform the mode of action analyses. The PECO criteria used for
screening the health effects, toxicokinetic, and mechanistic literature are provided in Appendix A
{U.S. EPA, 2024, 11414344}.
Consistent with the IRIS Handbook {U.S. EPA, 2022, 10367891} and the Systematic Review
Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA (anionic and acidforms) IRIS
Assessments {U.S. EPA, 2020, 8642427}, studies identified in the literature searches and stored
2-3
-------
APRIL 2024
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 independent reviewers using
screening tools consistent with the IRIS Handbook ({U.S. EPA, 2022, 10367891}; DistillerSR or
SWIFT ActiveScreener software), and studies that passed this screening underwent full-text
review by at least two independent reviewers. Health effects studies that met PECO inclusion
criteria following both title and abstract screening and full-text review underwent study quality
evaluation as described below (Section 2.1.3). Studies that were tagged as containing relevant
PBPK models were sent to the modeling technical experts for scientific and technical review.
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
Appendix A, {U.S. EPA, 2024, 11414344}). Supplemental studies that were identified as
mechanistic or ADME during screening did not undergo study quality evaluation.
For the supplemental literature search conducted in 2023 and literature received through public
comment, studies were screened for relevancy and considered for potential impact on the toxicity
assessments for PFOS. Consistent with the IRIS Handbook {U.S. EPA, 2022, 10367891}, the
studies identified after February 3, 2022, including studies recommended via public comment,
were "considered for inclusion only if they [were] directly relevant to the assessment PECO
criteria and [were] expected to potentially impact assessment conclusions or address key
uncertainties" {U.S. EPA, 2022, 10367891}. For the purposes of this assessment, the EPA
defined impacts on the assessment conclusions as data from a study (or studies) that, if
incorporated into the assessment, have the potential to significantly affect (i.e., by an order of
magnitude or more) the final toxicity values (i.e., RfDs and CSFs) or alter the cancer
classification for PFOS (see Appendix A, {U.S. EPA, 2024, 11414344}).
2.13 Study Quality Evaluation for Epidemiological Studies and
Animal Toxicological Studies
Study quality evaluations were performed consistent with the IRIS Handbook {U.S. EPA, 2022,
10367891} and the Systematic Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and
PFDA (anionic and acidforms) IRIS Assessments {U.S. EPA, 2020, 8642427}. For study quality
evaluation of the PECO-relevant human epidemiological and animal toxicological studies (i.e.,
studies identified in the four literature searches (all health outcomes for the 2019 and 2020
searches; the five priority health outcomes for the 2022 search; studies impacting assessment
conclusions within the five priority health outcomes for the 2023 search (see Appendix A, {U.S.
EPA, 2024, 11414344})), studies recommended by the SAB, studies recommended via public
comment that reported potentially significant results on one or more of the five priority health
outcomes, epidemiological studies from the 2016 PFOS HESD that reported results on one or
more of the five priority health outcomes, and key animal toxicological studies from the 2016
PFOS HESD), two independent primary reviewers followed by a quality assurance (QA)
reviewer assigned ratings about the reliability of study results {good, adequate, deficient (or "not
reported"), or critically deficient) for different evaluation domains as described in the IRIS
Handbook {U.S. EPA, 2022, 10367891} (see Appendix A, {U.S. EPA, 2024, 11414344}). These
2-4
-------
APRIL 2024
study quality evaluation domains are listed below and details about the domains, including
prompting questions and suggested considerations, are described in Appendix A {U.S. EPA,
2024, 11414344}.
• 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 quality; 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://hawcproiect.ore/). 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 consistent with the IRIS Handbook {U.S. EPA, 2022, 10367891}. A study
can be given an overall mixed confidence rating if different PECO-relevant endpoints within the
study receive different confidence ratings (e.g., medium and low confidence ratings).
2.1.4 Do to 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. Due to
the abundance of medium and high confidence studies in this database, data were only 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, 2022,
10367891}. Studies evaluated as being uninformative for an endpoint were not considered
further when characterizing that endpoint and therefore did not undergo data extraction. All
health endpoints were considered for extraction, regardless of the magnitude of effect or
statistical significance of the response relative to the control group. The level of detail in data
extractions for different endpoints within a study could differ based on how the data were
presented for each outcome (i.e., ranging from a narrative summary to a full extraction of dose-
response effect size information).
Extractions were conducted using DistillerSR for epidemiological studies and HAWC for animal
toxicological studies. An initial reviewer conducted the extraction, followed by a second
reviewer conducting an independent QA 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
2-5
-------
APRIL 2024
measured. Further information about data extraction can be found in Appendix A {U.S. EPA,
2024, 11414344}.
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 outcome (i.e., an organ- or organ system-
level category of related health effects and endpoints). 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. Consistent
with the IRIS Handbook, groups of related outcomes within a health outcome category were
considered together as a unit of analysis during evidence synthesis and evidence integration
{U.S. EPA, 2022, 10367891}. For example, birth weight, birth length, and head circumference
were all considered under the unit of analysis of the fetal growth restriction.
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}. 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 studies were
included in evidence syntheses in order to capture all of the available data for PFOS in the
weight of evidence analyses. As described above, uninformative studies were not extracted or
included in the evidence syntheses. Results from epidemiological studies were discussed within
sections organized by population type, including children, general population adults, pregnant
women, and occupational populations. Childhood was defined as the effect of environmental
exposure during early life: from conception, infancy, early childhood and through adolescence
until 21 years of age {U.S. EPA, 2021, 9641727}. 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 on the same
health outcome as another study were considered overlapping evidence, and, to avoid duplication
or overrepresentation of results from the same group of participants, these additional studies
were not discussed in the evidence synthesis narrative. In cases of overlapping studies, the study
with the largest number of participants and/or the most accurate outcome measures was given
preference. For the five priority health outcomes, EPA also developed mechanistic syntheses.
For evidence integration, conclusions regarding the strength of evidence were drawn for each
health outcome across human and animal evidence streams. For the five priority health
outcomes, this included consideration of epidemiological studies identified in the 2016 PFOS
HESD, as well as mechanistic evidence. The evidence integration provides a summary of the
causal interpretations between PFOS exposure and health effects based on results of the available
2-6
-------
APRIL 2024
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 judgments were
directly informed by the evidence syntheses and based on structured review of an adapted set of
considerations for causality first introduced by Austin Bradford Hill {Hill, 1965, 71664}.
The evidence integration was conducted according to guidance outlined in the IRIS Handbook
{U.S. EPA, 2022, 10367891} and the Systematic Review Protocol for the PFBA, PFHxA,
PFHxS, PFNA, andPFDA (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. Considerations for each evidence
integration judgment are summarized within corresponding evidence integration sections in an
evidence profile table (EPT). EPTs were organized by evidence stream (i.e., human, animal, and
mechanistic, respectively), and, within evidence streams, units of analysis with the strongest
evidence were presented first.
Additional details about evidence synthesis and integration are summarized in Appendix A {U.S.
EPA, 2024, 11414344}.
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 within those
health outcomes. Dose-response modeling was performed for endpoints within health outcomes
with data warranting evidence integration conclusions of evidence demonstrates and evidence
indicates (likely) for noncancer endpoints and carcinogenicity descriptors of Carcinogenic to
Humans and Likely to be Carcinogenic to Humans. EPA identified specific studies for dose-
response modeling and POD derivation following attributes described in Table 7-2 of the IRIS
Handbook {U.S. EPA, 2022, 10367891}. Examples of study attributes evaluated included study
design characteristics, study confidence, and data availability, among others (see Appendix A,
{U.S. EPA, 2024, 11414344}). Human epidemiological and animal toxicological studies that
were consistent with the overall weight of evidence for a specific endpoint were considered for
dose-response. Additionally, for human evidence, all high or medium confidence studies
pertaining to a specific endpoint were considered; for animal evidence, only animal toxicological
studies with at least two PFOS exposure groups that were of high or medium confidence were
considered. Relevance of the endpoint or species reported by animal toxicological studies to
human health effects was also considered. Additional information on study selection is provided
in Appendix A {U.S. EPA, 2024, 11414344}.
2-7
-------
APRIL 2024
2.2.1 Approach to POD and Candidate RfD Derivation for
Noncancer Health Outcomes
The current recommended EPA human health risk assessment approach for noncancer POD
derivation described in EPA's A Review of the Reference Dose and Reference Concentration
Processes includes selection of a benchmark response (BMR), analysis of dose and response
within the observed dose range, followed by extrapolation to lower exposure levels {U.S. EPA,
2002, 88824}. For noncancer health outcomes, EPA performed dose-response assessments to
define PODs, including low-dose extrapolation, when feasible, and applied uncertainty factors
(UFs) to those PODs to derive candidate RfDs. An RfD is an estimate, with uncertainty spanning
perhaps an order of magnitude, of an exposure to the human population (including susceptible
subgroups) that is likely to be without an appreciable risk of deleterious health effects over a
lifetime {U.S. EPA, 2002, 88824}. For PFOS, multiple candidate RfDs were derived within a
health outcome as described in Section 4.
For PFOS animal toxicological studies, EPA attempted benchmark dose (BMD) modeling on all
studies considered for dose-response to refine the POD. BMD modeling was performed after
converting the administered dose reported by the study to an internal dose using a
pharmacokinetic model (see Section 4.1.3 for additional details). This approach resulted in dose
levels corresponding to specific response levels near the low end of the observable range of the
data and identified the lower limits of the BMDs (BMDLs) which serve as potential PODs {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 (i.e., BMD) that corresponds to a
predetermined level of response (i.e., 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. However, a POD derived
using a BMD approach typically provides a higher level of confidence in the conclusions for any
individual case, as the BMDL takes into account all the data from the dose-response curve,
incorporates the evaluation of the uncertainty in the BMD, and is related to a known and
predefined potential effect size (i.e., the BMR) {U.S. EPA, 2012, 1239433; U.S. EPA, 2022,
10367891}. For noncancer endpoints, there were several factors considered when selecting the
final model and BMD/BMDL, including the type of measured response variable (i.e.,
dichotomous or continuous), experimental design, and covariates {U.S. EPA, 2012, 1239433}.
However, as there is currently no prescriptive hierarchy, selection of model types was often
based on the goodness-of-fit and was judged based on the x2 goodness-of-fit p-value (p > 0.1),
magnitude of the scaled residuals in the vicinity of the BMR, and visual inspection of the model
fit. The Benchmark Dose Technical Guidance provides a "BMD Decision Tree" to assist in
model selection {U.S. EPA, 2012, 1239433}. See Appendix E {U.S. EPA, 2024, 11414344} for
additional details on the study-specific modeling.
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
2-8
-------
APRIL 2024
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} because the EPA was
able to define a level considered clinically adverse for these outcomes (see Appendix E, {U.S.
EPA, 2024, 11414344}). As sensitivity analyses for comparison purposes, EPA also performed
BMD modeling and provided study LOAELs/NOAELs as PODs for the epidemiological hepatic
and serum lipid dose-response studies. For the immune studies, for which a clinically defined
adverse level is not established, 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 Appendix E {U.S. EPA, 2024, 11414344} for additional details on the
study-specific modeling.
After POD derivation, EPA used a pharmacokinetic model for human dosimetry to estimate
human equivalent doses (HEDs) from both animal and epidemiological studies. A
pharmacokinetic model for human dosimetry is used to simulate the HED from the animal PODs
and is also used to simulate selected epidemiological studies to obtain a chronic dose that would
result in the internal dose POD obtained from dose-response modeling (Section 4.1.3). 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. Next, reference values are estimated by
applying relevant adjustments to the point-of-departure human equivalent doses (PODheds) to
account for five possible areas of uncertainty and variability: human variation, extrapolation
from animals to humans, extrapolation to chronic exposure duration, the type of POD being used
for reference value derivation, and extrapolation to a minimal level of risk (if not observed in the
data set). UFs used in this assessment were applied according to methods described in EPA's
Review of the Reference Dose and Reference Concentration Processes {U.S. EPA, 2002,
88824}. For additional detail on UFs, see Appendix A {U.S. EPA, 2024, 11414344}. The
PODhed for a particular candidate RfD is divided by the composite UFs.
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 confidence, 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, magnitude of response, 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) 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 stages are selected for
simulation in a mouse, rat, or monkey. Concentrations of PFOS in blood are considered for all
the internal dose-metrics.
2-9
-------
APRIL 2024
Step 2: Conduct dose-response modeling. See above and Appendix E {U.S. EPA, 2024,
11414344} 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 (e.g., BMDL, NOAEL) is converted to an HED following
the method described in Section 4.1.3.
Step 4: Select appropriate UFs and provide rationale for UF selection. UFs are applied in
accordance with EPA methodology 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 {U.S. EPA, 2002, 88824}.
Step 5: Calculate the chronic RfD. The RfD is calculated by dividing the PODhed by the
composite (total) UF specific to that PODhed.
PODhed = calculated from the internal dose POD 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 methodology {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
• Likely to Be Carcinogenic to Humans
• Suggestive Evidence of Carcinogenic Potential
• Inadequate Information to Assess Carcinogenic Potential
• Not Likely to Be Carcinogenic to Humans
where:
2.2.2 Cancer Assessment
2.2.2.1 Approach for Cancer Classification
2-10
-------
APRIL 2024
More than one carcinogenicity descriptor can be applied if a chemical's carcinogenic effects
differ by dose, exposure route, or mode of action (MOA)3. 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. The MOA analysis must be conducted separately for each target organ/tissue type {U.S.
EPA, 2005, 9638795}.
2.2.2.2 Derivation of Candidate Cancer Slope Factors
EPA's 2005 Guidelines for Carcinogen Risk Assessment recommends a two-step process for the
quantitation of cancer risk as a CSF. A CSF is a plausible upper bound lifetime cancer risk from
chronic ingestion of a chemical per unit of mass consumed per unit body weight per day (mg/kg-
day) {U.S. EPA, 2005, 9638795}. First, a model is used to fit a dose-response curve to the data,
based on the doses and associated tumors observed {U.S. EPA, 2005, 9638795}. 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 evidence 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,
the linear approach is used, and the 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.
For animal toxicological studies, EPA used the publicly available Benchmark Dose Software
(BMDS) program developed and maintained by EPA (https://www.epa.eov/bmds). First, a PK
model converted the administered dose reported by the study to an internal dose (see Section
4.1.3 for additional details). Then, BMDS fits multistage models, the preferred model type {U.S.
EPA, 2012, 1239433}, to the data 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 for that dose) are also reported and are
often used as the POD. Similar to noncancer PODs, selection of model types is often based on
the goodness-of-fit {U.S. EPA, 2012, 1239433}. For PFOS, after a POD was determined, a PK
model was used to calculate the HED for animal oral exposures (PODhed). The CSF is derived
by dividing the BMR by the PODhed. See Appendix E {U.S. EPA, 2024, 11414344} for
additional details on the study-specific modeling.
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) indicates the potential for higher cancer risks from
an early-life exposure compared to the same exposure during adulthood, and so requires that the
application of age-dependent adjustment factors (ADAFs) be considered in the quantification of
3MOA 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.
2-11
-------
APRIL 2024
risk to account for additional sensitivity of children. The ADAFs are 10- and 3-fold adjustments
that are combined with age specific exposure estimates when estimating cancer risks from early
life (<16 years of age) exposure to a mutagenic chemical.
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
summarized in the characterization of the POD" {U.S. EPA, 2005, 9638795}. In these cases, an
RfD-like value is calculated based on the key event4 for carcinogenesis or the tumor response.
2.2.3 Selecting Health Outcome-Specific and Overall Toxicity
Values
The next step is to select a health outcome-specific toxicity value for each hazard (cancer and
noncancer) identified in the assessment. This selection can be based on the study confidence
considerations, the most sensitive outcome, a clustering of values, or a combination of such
factors; the rationale for the selection is presented in the assessment. Key considerations for
candidate value selection are described in the IRIS Handbook {U.S. EPA, 2022, 10476098} and
include: 1) the weight of evidence for the specific effect or health outcome; 2) study confidence;
3) sensitivity and basis of the POD; and 4) uncertainties in modeling or extrapolations. The value
selected as the organ/system-specific toxicity value is discussed in the assessment.
The selection of final toxicity values for noncancer and cancer effects involves the study
preferences described above, consideration of overall toxicity, study confidence, and confidence
in each value, including the strength of various dose-response analyses and the possibility of
basing a more robust result on multiple data sets. The values selected as the overall RfD and CSF
are discussed in Section 4.
4The 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.
2-12
-------
APRIL 2024
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 7,160 unique records (combined for PFOA and PFOS) prior
to running SWIFT Review. Table 3-1 shows the results from database searches conducted in
April 2019, September 2020, and February 2022, and February 2023.
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
2/6/2023: 966 results
PubMed
4/10/2019:2,191 results
9/3/2020: 811 results
2/2/2022: 1,728 results
2/6/2023: 719 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
2/2023: 1,153 results
Total number of references after running SWIFT Review3
4/2019: 1,977 results
9/2020: 867 results
2/2022: 1,370 results
2/2023: 881 results
Total number of unique references moved to screeningb
4,802
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, studies identified by
the Science Advisory Board (SAB), and EPA's 2016 Health Effects Support Documents
(HESDs) for perfluorooctanoic acid (PFOA) {U.S. EPA, 2016, 3603279} and perfluorooctane
sulfonate (PFOS) {U.S. EPA, 2016, 3603365}) yielded 238 unique records (combined for PFOA
and PFOS).
The 4,802 studies captured with the SWIFT Review evidence streams filters and the 238 records
identified from additional sources yield a total of 5,011 unique studies. These 5,011 studies were
3-1
-------
APRIL 2024
moved to the next stage of screening (title and abstract screening using either DistillerSR or
SWIFT ActiveScreener). Of the 5,011 unique studies, 1,062 moved on to full-text level review,
1,697 were excluded during title and abstract screening, and 2,252 were tagged as containing
potentially relevant supplemental material. Of the 1,062 screened at the full-text level, 760 were
considered to meet PECO eligibility criteria (see Appendix A, {U.S. EPA, 2024, 11414344})
and included relevant information on PFOS. The 760 studies that were determined to meet
PECO criteria after full-text level screening included 429 epidemiological (human) studies, 45
animal toxicological studies, 11 physiologically based pharmacokinetic (PBPK) studies, and 275
studies that were not extracted (e.g., low confidence studies, meta-analyses, studies from the
2022 and 2023 searches 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 27
PBPK studies. Details of the literature search and screening process are shown in Figure 3-1.
The 429 epidemiological studies and 45 animal toxicological studies relevant to PFOS
underwent study quality evaluation and were subsequently considered for data extraction as
outlined in Sections 2.1.3 and 2.1.4 (see Appendix A, {U.S. EPA, 2024, 11414344}). The results
of the health outcome-specific study quality evaluations and data extractions are described in
Sections 3.4 and 3.5.
Additionally, the 27 studies tagged as containing relevant PBPK models for PFOS were
reviewed by pharmacokinetic (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 104 toxicokinetic and 305 mechanistic studies identified as relevant for PFOS moved
on to a limited data extraction as described in the Appendix {U.S. EPA, 2024, 11414344}. The
toxicokinetic studies pertaining to ADME are synthesized in Section 3.3.1. The mechanistic
studies relevant to the five priority health outcomes are synthesized in Sections 3.4 and 3.5 and
were considered as part of the evidence integration.
In addition to the studies identified through database searches and the other sources outlined
above, public comments submitted in response to the Public Comment Draft Toxicity Assessment
and Proposed Maximum Contaminant Level Goal for Perfluorooctane Sulfonic Acid (PFOS) in
Drinking Water {U.S. EPA 2023, 10841010} included 944 studies, relevant to PFOA and/or
PFOS, which were reviewed for relevance to the toxicity assessment. Of the 944 studies, 297
were duplicates of studies included in the toxicity assessment and 31 were duplicates of studies
included in the 2016 PFOA or PFOS HESD assessment. The 599 studies that were not identified
in the 2016 HESDs and were not included in the toxicity assessments underwent additional
review identify studies with that could impact assessment conclusions as outlined in Appendix
A.3 {U.S. EPA, 2024, 11414344}. Ultimately, none of the 599 studies were incorporated in the
toxicity assessments upon further screening. The submitted references were either deemed not
relevant after secondary review, were supplemental studies (e.g., PFOA or PFOS assessments
published by other scientific bodies, mechanistic, ADME, etc), or addressed non- priority health
outcomes. The results of this screening can be found in the docket ("Review of Public Comment
References Related to PFOA and PFOS Health Effects;"
https://www.reeiilations.eov/docket/EPA-HQ-OW-202
3-2
-------
APRIL 2024
Figure 3-1. Summary of Literature Search and Screening Process for PFOS
Interactive figure and additional study details available on HAWC.
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 References identified by SAB and through database searches were counted as identified through database search only.
3-3
-------
APRIL 2024
b Includes number of unique references after deduplication of studies captured with the SWIFT Review evidence streams filters
and records identified from additional sources.
c Includes number of unique references considered to meet PECO eligibility criteria at the full-text level and include relevant
information on PFOS.
d Includes number of unique references identified during title/abstract screening, full-text screening, and data extraction assessed
for toxicokinetic and/or mechanistic eligibility.
e Only includes references with relevant information on PFOS.
f References tagged to 'Not a priority human health system' include those identified in the 2019 search that overlap with 2016
PFOS HESD references or those identified in 2022 and 2023 searches.
g Includes 11 PBPK references determined to meet PECO criteria plus an additional 16 PBPK references identified during the
toxicokinetic screening.
3.1.1 Results for Epidemiology Studies of PFOS by Health
Outcome
Of the 429 epidemiological studies that met the inclusion criteria and underwent extraction, 181
studies had a cohort study design, 169 had a cross-sectional design, 42 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 = 86), metabolic (n = 74), or immune systems (n = 66). 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
7
3
3
5
18
Cardiovascular
5
19
56
6
86
Dermal
0
1
0
0
1
Developmental
6
58
19
7
90
Endocrine
1
8
20
7
36
Gastrointestinal
1
4
0
0
5
Hematologic
0
0
8
0
8
Hepatic
1
4
18
2
25
Immune
6
32
19
9
66
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
42
181
169
37
429
Figure 3-2. Summary of Epidemiology Studies of PFOS Exposure by Health System and
Study Design3
Interactive figure and additional study details available on HAWC.
3-4
-------
APRIL 2024
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 and underwent extraction,
most studies had either short-term (n = 19) or developmental (n = 15) study designs.
Approximately equal numbers of studies were conducted in rats (n = 23) and mice (n = 21). The
rat studies had short-term (n = 12), developmental (n = 7), chronic (n = 2), reproductive (n = 2),
and subchronic (n = 1) study designs. The mouse studies had developmental (n = 8), short-term
(n = 7), subchronic (n = 5), or reproductive (n = 1) study designs. The single monkey study used
a chronic study design and the single rabbit study used 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.
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
1
0
16
Endocrine
1
4
1
0
1 1
2
0
3
0
1
13
Hematologic
1
3
0
0
1 0
0
0
0
0 0
5
Hepatic
2
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
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
1 3
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
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 HAWC.
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
All data from this project are available in the public HAWC
(https://hawc.epa.gov/assessment/10050Q248/) site displayed as exposure-response arrays, forest
plots, and evidence maps. Data extracted from the 429 epidemiological studies are available
here. Data extracted from the 45 animal toxicological studies are available here. See Sections 3.4
and 3.5 for health outcome-specific data extracted for synthesis development. Additionally, the
3-5
-------
APRIL 2024
limited data extractions from the ADME and mechanistic studies can be found here and here.
respectively.
3.3 Toxicokinetic Synthesis
As described in Section 3.1, EPA identified 104 and 27 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 {U.S. EPA, 2024, 11414344}.
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
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. Given the 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
Appendix B, {U.S. EPA, 2024, 11414344}).
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
3-6
-------
APRIL 2024
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}.
33.1.2 Distribution
3.3.1.2.1 PFOS 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 B {U.S. EPA, 2024, 11414344}. 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
3-7
-------
APRIL 2024
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 versus cord blood, lower cord blood albumin levels compared with 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.
3.3.1.2.2 PFOS 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.
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; Maestri, 2006, 1048866}. 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} 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. Another study of post-mortem tissues found varying PFOS levels in different
tissues ranging from 1.0 ng/g in skeletal muscle to 13.6 ng/g in liver. PFOS was also detected in
brain and basal ganglia, endocrine organs (pituitary, thyroid, pancreas), liver, kidney, and
adipose tissue {Maestri, 2006, 1048866}. PFOS also accumulates in follicular fluid {Kang,
2020, 6356899} and gonads {Maestri, 2006, 1048866}, 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
3-8
-------
APRIL 2024
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}. 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}. In fact, one study in rats {Wang, 2015,
3981881} observed higher PFOS levels in the hippocampus than in serum measured on PND 1
in prenatally exposed rats. 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}. However, some sex-dependent differences in PFOS levels were observed in
rodents that varied by species, lifestage, and dose duration {Zhong, 2016, 3748828; Thomford,
2002, 5029075; Curran, 2008, 757871}.
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
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 et al. {, 2015, 2851103} observed
differential accumulation of PFOS based on branching characteristics. Specifically, RpMSof
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 placental 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 Appendix B, {U.S. EPA, 2024,
11414344}) and generally increased with gestational age {Li, 2020, 6505874}. Li et al. {, 2020,
6505874} also showed a 6% increase in branched PFOS accumulation compared with linear
PFOS isomers. Zhao et al. {, 2017, 3856461} observed higher transplacental transfer efficiencies
(TTEs) for lm-, 4m-, 3 + 5m-, and m2-PFOS compared with n-PFOS. Together, these findings
indicate that branched isomers of PFOS transfer more efficiently from maternal blood to cord
blood compared with 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}.
3-9
-------
APRIL 2024
Lower PFOS concentrations were measured in amniotic fluid compared with 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 with 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 with the third trimester, and lowest in the lung in the second trimester compared with
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
breastfeeding mothers versus 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 with other lipophilic persistent organic pollutants such
as polychlorinated 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}.
3.3.1.2.5 Volume 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
3-10
-------
APRIL 2024
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 with 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 5.6.2.
3.3.1.3 Metabolism
Consistent with other reports and reviews {U.S. EPA, 2016, 3603365; ATSDR, 2018, 9642134;
Pizzuro, 2019, 5387175}, the literature reviewed for this assessment do not provide evidence that
PFOS is 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
females. In females, elimination pathways include menstruation, pregnancy (cord blood,
placenta, amniotic fluid, and fetal tissues) and lactation (breast milk) (see Appendix B, {U.S.
EPA, 2024, 11414344}).
3.3.1.4.1 Urinary 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 with 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).
3-11
-------
APRIL 2024
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 with 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.
3.3.1.4.2 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
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 yet unknown whether
and to what extent these transporters function in vivo.
Studies describing renal resorption are discussed in Appendix B {U.S. EPA, 2024, 11414344}.
3.3.1.4.3 Maternal 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
3-12
-------
APRIL 2024
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 with 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 lifestage 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 with 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.
After birth, 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
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.
3.3.1.4.4 Other 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, a re-
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. Studies
providing direct measurements of PFOS in menstrual blood were not identified. However, for
3-13
-------
APRIL 2024
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 a 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 with females in the low dose group, there were no significant
differences in hair PFOS concentrations between males and females in the higher dose groups.
3.3.1.4.5 Half-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 Appendix B,
{U.S. EPA, 2024, 11414344}). 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 Appendix B, {U.S. EPA, 2024, 11414344}). 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 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 Appendix B, {U.S. EPA, 2024, 11414344}). PFOS half-life values
among these four 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 with 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 versus 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.
3-14
-------
APRIL 2024
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
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.
3-15
-------
APRIL 2024
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 concentrations 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. Because of the long half-life of PFOS,
steady state is a reasonable assumption for adults starting from the age of 25 and above.
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 toxicity 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.
3-16
-------
APRIL 2024
In animals, two classical PK models for PFOS have been published since the 2016 PFOS 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.
33.2.2 Modified Compartmentai 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
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 and allow for dose-dependent changes in excretion rate
compared with the classic 1- or 2- compartment approaches 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.
3-17
-------
APRIL 2024
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
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 with PFOA, PFOS had a longer terminal half-life and more rapid approach
to steady-state with repeated oral administration.
3-18
-------
APRIL 2024
Oral Dose ka
(Agut)
Second
Compartment
(V„ Ctissue)
ki2^ ^ k2i
Central
Compartment
(Vc. ^central' ^plasma)
IV Dose
Qfil I : Trm Kt
" I ! Tm,
f :
Filtrate Qfi
Compartment
(Vfl|. 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
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
3-19
-------
APRIL 2024
both from Chang et al. {, 2012, 1289832}. The data were analyzed within a Bayesian framework
using Markov Chain Monte Carlo sampler implemented as an R package developed by EPA to
allow predictions across species, strains, and sexes and to identify serum levels associated with
the no-observed-adverse-effect level (NOAEL) and lowest-observed-adverse-effect level
(LOAEL) external doses. Prior distributions for the parameters were chosen to be broad, log-
normal distributions, allowing the fitted parameters to be positive and for the posterior
distribution to be primarily informed by the data likelihood rather than by the priors.
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 with blood
concentration.
The first PBPK model developed for PFOS 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
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.
3-20
-------
APRIL 2024
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 human equivalent doses (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.
3-21
-------
APRIL 2024
3.4 Noncancer Health Effects Evidence Synthesis and
Integration
3.4.1 Hepatic
EPA identified 24 epidemiological studies (30 publications)5'6 and 25 animal toxicological
studies that investigated the association between PFOS and hepatic effects. Of the
epidemiological publications, 17 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 7 epidemiological studies (8 publications)6 from the 2016 PFOS HESD {U.S. EPA,
2016, 3603365} that investigated the association between PFOS and hepatic effects. Study
quality evaluations for these eight studies are shown in Figure 3-5. Results from studies
summarized in the 2016 PFOS HESD are described in Table 3-2 and below.
5 Multiple publications of the same data: Jain andDucatman {, 2019, 5381566}; Jain andDucatman {, 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.
6 Olsen {, 2003,1290020} is the peer-review paper of Olsen {, 2001, 10228462} and Olsen {, 2001, 10240629}; however, data
for PFOA and hepatic outcomes is reported in Olsen {,2001, 10228462}. Olsen {,2001,10240629} was considered overlapping
and not evaluated because data in the technical report was completely described in Olsen {, 2003, 1290020}.
3-22
-------
APRIL 2024
e\e&^e^e^ o, ^0
-------
APRIL 2024
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,
1.18). There was less consistent evidence of an association between PFOS and GGT or bilirubin
in this study.
Both the Gallo et al. {,2012, 1276142} and Lin et al. {,2010, 1291111} studies observed a
slight positive association between serum PFOS levels and increased serum ALT values
(Figure 3-6). 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 bating ^MamT Design Exposure 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
Gallo et al. (2012, Serum Cross - Median=20.3 - Regression
1276142), Medium sectional ng/mL (25th - 75th coefficient (per Hn
percentile: ng/mL increase in
13.7-29.4 ng/mL) PFOS) 0.02
1
1
1
j •
1
1
1
Gleason etal. (2015, Serum Cross- Median=11.3 - Regression
2966740), Medium sectional ng/mL (25th-75th coefficient (per 1-ln
percentile: 7.0 - ng/mL increase in
18.0 ng/mL) PFOS) 0.01
1
1
1
V
1
1
1
Lin etal. (2010, Serum Cross- Median: 23.50 - Regression
1291111), Medium sectional ng/mL (25th-75th coefficient (per
percentile: 1-log ng/mL
15.50-33.80 increase in PFOS) 0.82
ng/mL)
1
1
1
1 •
1
1
1
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 2016 PFOS HESD Epidemiology Studies Following
Exposure to PFOS
Interactive figure and additional study details available on HAWC.
Several cross-sectional occupational studies in PFOS production workers reported mostly null or
inconsistent findings with respect to biomarkers of liver disease. Exposure to PFOA was
generally associated with increased ALT concentrations, but findings were inconsistent for some
timepoints or in sex-stratified groups {Olsen, 2003, 1290020; Olsen, 2001, 10228462}. Null or
inconsistent associations were also 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}.
3-24
-------
APRIL 2024
Table 3-2. Associations Between Elevated Exposure to PFOS and Hepatic Outcomes From
Studies Identified in the 2016 PFOS HESD
Reference,
confidence
Study
Design
Population
ALT3
AST3
GGT3
ALP3
Liver Diseaseb
Alexander, 2003,
1291101
Cohort
Occupational
NA
NA
NA
NA
-
Low
Gallo, 2012,
1276142
Cross-
sectional
Adults
tt
NA
-
NA
NA
Medium
Grice, 2007,
4930271
Cohort
Occupational
NA
NA
NA
NA
t
Low
Lin, 2010, 1291111
Medium
Cross-
sectional
Adults
tt
NA
-
NA
NA
Olsen, 2001,
10228462
Cohort
Occupational
t
t
-
-
NA
Medium
Olsen, 2003,
1290020
Cross-
sectional
Occupational
t
-
t
t
NA
Medium
Yamaguchi, 2013,
2850970
Cross-
sectional
Adults and
adolescents
tt
tt
tt
NA
NA
Medium
Notes'. ALP = alkaline phosphatase; ALT = alanine transferase; AST = aspartate transaminase; GGT = gamma-glutamyl
transferase; NA = no analysis was for this outcome was performed; | = nonsignificant positive association; ft = significant
positive association; j = nonsignificant inverse association; jj = significant inverse association; - = no (null) association.
Jain et al., 2014,2969807 was not included in the table due to their uninformative overall study confidence ratings.
3 Arrows indicate the direction in the change of the mean response of the outcome (e.g., j indicates decreased mean birth weight).
b Arrows indicate the change in risk of the outcome (e.g., | indicates an increased risk of the outcome).
3.4.1.1.2 Study Quality Evaluation Results for the Updated Literature Review
There are 17 epidemiological studies (23 publications)7 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 3.
Of these, 12 were classified as medium confidence, four as low confidence, and seven were
considered uninformative. Of the informative studies, two cross-sectional studies {Nian, 2019,
5080307; van den Dungen, 2017, 5080340}, multiple publications of data from NHANES {Jain,
2019, 5381541; Liu, 2018, 4238514; Omoike, 2020, 6988477; Jain, 2019, 5080621; Jain, 2019,
5381566; Gleason, 2015, 2966740}, 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 cross-sectional studies {Rantakokko, 2015,
7 Multiple publications of the same data: Jain andDucatman {, 2019, 5381566}; Jain andDucatman {, 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.
3-25
-------
APRIL 2024
3351439, Liu, 2018, 4238396} examined functional liver endpoints in adults. In children and
adolescents, four studies were available including one cohort study {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}).
3-26
-------
APRIL 2024
^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 Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Hepatic Effects3
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 etal. {, 2020, 6988477}; Liu {, 2018, 4238514}; Gleason et al. {, 2015,
2966740} all use NHANES data from overlapping years.
3-27
-------
APRIL 2024
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 Appendix D {U.S. EPA, 2024,
11414344}. 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
(p < 0.05) was reported in Olsen et al. {, 2012, 2919185} (see Appendix D, {U.S. EPA, 2024,
11414344}). 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. ALT findings from low confidence studies are not
included in figures.
In children and adolescents, positive associations were observed in girls in the fourth quartile
reported by 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 Appendix D {U.S. EPA, 2024, 11414344}. In
adults, statistically significant positive associations were observed in the one medium confidence
study {Nian, 2019, 5080307} and in NHANES studies. Van den Dungen et al. {, 2017,
5080340} reported a nonsignificant 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}.
3-28
-------
APRIL 2024
Reference,
Confidence Rating
Exposure
Matrix
Study
Design
Exposure Levels
Sub-population
Comparison
EE
0.0 0.1
0.2
Effect Estimate
0.3 0.4 0.5
0.6
0.7
0.8
Jain et al. (2019,
5080621), Medium
Serum
Cohort
Geometric mean (95%
CI) = 6.3 ng/mL (5.8 -
6.8)
Non-obese
Regression
coefficient (per
1-!og10 ng/mL
increase in
PFOS)
-0.02
I
1
1
• .
1
1
1
Geometric mean (95%
Cl)= 5.5 ng/mL (5.0 -
6.0)
Obese
Regression
coefficient (per
1-log10 ng/mL
increase in
PFOS)
0.02
l
1
1
•
1
1
1
Nian et al. (2019,
5080307), Medium
Serum
Cross -
sectional
Median=24.22 ng/mL
(25th-75th percentile:
14.62-37.19 ng/mL)
Excluding
medicine takers
Regression
coefficient (per
1-in ng/mL
increase in
PFOS)
0.04
I
1
1
1
1
1
Salihovic et al. (2018.
5083555), Medium
Plasma
Cohort
Median (25th-75th
percentile): Age 70:
13.2 ng/mL (9.95-17.8);
Age 75:12.6 ng/mL
(7.97-19.2): Age 80:
0.57 ng/mL (5.36-11.5)
Regression
coefficient (per
1-ln ng/mL
increase in
PFOS)
0.03
r
i
i
~
i
i
!
00 0.1
0.2
0.3 0.4 0.5
0.6
0.7
0.8
Figure 3-8. Overall ALT Levels from Epidemiology Studies Following Exposure to PFOS
Interactive figure and additional study details available on HAWC.
Literature
Search Tag
Pre-2016
Literature
Search
Reference. E s d
Confidence ^ DMig»
Rating
Gallo et al.
(2012,
1276142),
Medium
Cross -
sectional
Exposure Levels Comparison
Median=20.3 ng/mL OR (per Hn ng/mL
(25lh - 75th percentile: increase in PFOS)
13.7-29.4 ng/mL)
Median=20.3 ng/mL OR (for decile 2 vs. c
(IQR=13.7-29.4 ng/mL) 1 of PFOS)
Effect Estimate
0.8 0.9
OR (for decile 3 vs. c
1 of PFOS)
OR (for decile 4 vs. decile
1 of PFOS)
OR (for decile 5 vs. decile
1 of PFOS)
OR (for decile 6 vs. decile
1 of PFOS)
OR (for decile 7 vs. decile
1 of PFOS)
OR (for decile 8 vs. decile
1 of PFOS)
OR (for decile 9 vs. decile
1 of PFOS)
OR (for decile 10 vs. decile
1 of PFOS)
Figure 3-9. Odds of Elevated ALT Levels from Epidemiology Studies Following Exposure
to PFOS
Interactive figure and additional study details available on HAWC.
3-29
-------
APRIL 2024
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 that higher PFOS exposure levels
were associated with reduced odds of lobular inflammation, whereas Jin et al. {, 2020, 6315720}
reported the opposite, with an OR of 2.9 for 2-4 foci versus, 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 non-alcoholic
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 of the individual contaminants. 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. 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 in epidemiological studies, there is residual uncertainty.
3.4.1.2 Animal Evidence Study Quality Evaluation and Synthesis
There are 6 animal toxicological studies from the 2016 PFOS HESD {U.S. EPA, 2016,
3603365} and 19 animal toxicological 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.
3-30
-------
APRIL 2024
Butenhoffetal., 2012, 1276144-
++
++
I
NR
++
++
++
++
++
++
++
Conley et al., 2022, 10176381 -
+
+
NR
++
B
D
B
++
B
Curran et al., 2008, 757871 -
++
NR
NR
++
B
++
++
Dong et al., 2011, 1424949-
++
+
NR
++
++
++
++
++
++
Era et al., 2009, 2919358-
+
NR
NR
++
El
B
++
H
Fuentes et al., 2006, 757859 -
+
+
NR
+
+
+
B
++
++
+
Han et al., 2018, 4238554-
+
+
NR
++
D
++
++
+
+
+
Han et al., 2018, 4355066-
++
+
NR
++
++
B
++
+
++
+
Kawamoto et al., 2011, 2919266 -
-
+
NR
¦
++
++
+
++
-
Lai et al., 2018, 5080641 -
+
+
NR
++
-
+
+
+
+
+
Lau et al., 2003, 757854-
++
+
NR
+
+
+
++
B
+
+
Lefebvre et al., 2008, 1276155 -
+
NR
NR
++
+
+
++
++
+
Liang etal., 2019, 5412467-
+
+
NR
+
NR
+
++
I
-
Legend
H
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)
F
Not reported
*
Multiple judgments exist
Figure 3-10. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFOS Exposure and Hepatic Effects1'-1'
Interactive figure and additional study details available on HAWC.
"Hanetal. {,2018,4238554} andWanetal. {,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}.
3-31
-------
APRIL 2024
Li etal., 2021, 7643501
NTP, 2019, 5400978-
Seacat et al., 2002, 757853
Seacatetal., 2003, 1290852
Thomford, 2002, 5432419
Wan etal., 2016, 3981504
Wan etal., 2020, 7174720
Xing etal., 2016, 3981506
Yan etal., 2014, 2850901 -
Yang etal., 2021, 7643494
Zhang etal., 2019, 5918673
Zhong etal., 2016, 3748828
B 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
Figure 3-11. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFGS Exposure and Hepatic Effects (Continued)
Interactive figure and additional study details available on HAWC.
aHanetal. {,2018,4238554} and Wan etal {,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
PFOS exposure durations and doses. This includes effects such as increased absolute and relative
3-32
-------
APRIL 2024
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 a 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 with 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 increased 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}.
An increase in relative liver weight was observed with exposure as low as 0.15 mg/kg/day PFOS
administered to female Sprague-Dawley rats for 28 days {Curran, 2008, 757871}. In males from
3-33
-------
APRIL 2024
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 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 2-21. 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 a relatively short 5-day exposure (GD 14-18) to PFOS concentrations of 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}.
3.4.1.2.2 Clinical 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)). Serum levels of 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 > twofold compared with controls (i.e., > 100% change relative to control response) {U.S.
EPA, 2002, 625713; Hall, 2012, 2718645}.
Two studies in male mice found 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
group. 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
3-34
-------
APRIL 2024
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 (
-------
APRIL 2024
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 twofold compared with 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 a 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 approximately 6.34 mg/kg/day
in male rats.
3-36
-------
APRIL 2024
Endpoint Study Name Study Design
Alanine Aminotransferase (ALT) Han at al,. 2018,4355066 short-term (28d) 28d
> Animal Description Dose (mg'kg/day)
Rat, Sprague-Dawley (•?, N=6) 0
PFOS Hepatic Effects - Serum Enzymes in Male Rats
| Q Statistically significant 0 Not statistically siflnificant [—I 95%
NTP. 2019, 5400978
Rat, Sprague-Dawley N=10> 0
0,312
Seacat et a!.. 2003.1290852
:rl:CD(SD)IGS BR (, C, N=10) C
Butenhoffetal.. 2012.1276144 chronic (2y) 4wk
Rat. Crl:CD(SD)IGS BR 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 ( :,
Alkaline Phosphatase (ALP) NTP. 2019, 5400978
Rat, Sprague-Dawley N=10) 0
Aspartate Aminotransferase (AST) Han etal.. 2018,4355066 short-term (28d) ;
Rat. Sprague-Dawley (;•''. N=6)
NTP. 2019. 5400978
Rat, Sprague-Dawley ( -', N=10) 0
Seacat etal. 2003.1290852 short-term (4wk) 4wk
Rat, Crt:CD(SD)IGS BR N=10) C
Butenhoffetal., 2012, 1276144 chronic (2y)
Rat, Crl:CD(SD)IGS BR N=10) 0
(-•i
'-~-H
Rat, Crl:CD(SD)IGS BR N=10) 0
Rat, Crl:CD(SD)lGS BR (. T, N=10) 0
Rat, Crl;CD(SD)IGS BR < , N=10) 0
Percent control response (%)
Figure 3-13. Percent Change in Serum Enzyme Levels Relative to Controls in Male Rats
Following Exposure to PFOS" b
Interactive figure and additional study details available on HAWC here and here.
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.
3-37
-------
APRIL 2024
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 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, 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 and
Conley et al. {, 2022, 10176381} did not observe effects on ALT levels in female Sprague-
Dawley dams treated with up to 30 mg/kg/day PFOS from GD 14-18.
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 with
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} observed no significant effect on AST in the serum of Sprague-
Dawley dams exposed to PFOS concentrations between 0.1-30 mg/kg/day from GD 14-18.
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}.
3-38
-------
APRIL 2024
PfOS Hep
atic Effect* - Serum Cnzynea In Female Rata
eimpeini SiuOy litem*
SUidy 0«s>4'i ot>u>rvauoi
Time Annual Description
Dom imjlktyO*))
Wa^itCant ^ ISo
jtatis&co»> serine
anti 95V CI |
Atemine Amnclraufcrase (ALT) Ccrtey ot al 2021.10178181
dcvetapmenlal (GDI4-18} GDIS
P0 Rat. SprapjeOawley l . N=4-6
0
•
0.t
«
H
3
h|m
•
f*-1
N«> Ml 9. 5400978
SlWt-tofrr, |2M) 290
Rat, Sfroquc-Oowlcy I , N=9-I0>
O
•4
<
* j
>
r
©
- •
•
Saacatalat, mi 1200057
chronic (I'M.)
Rat. Oi:COfSD)K3S BR ( , N-10)
0.04
«4
i
1
h
•—t
•—'
BuMnMUtal., 2QI2 1276144
ehtonic (2y) 4wk
Rat. Cit:CO(SD)*GS BR (-. N-10)
Rat. Cn-CDlSDXGS BR (:. N=10)
(tel. Cil:CD(SD»eS BR ( . N-10)
0.029
0.299
1251
-m
H
»
K
i
.. i.i...,i.LAi
, #-
~ •
Rat C" CD|SD)IG5 BR ( . N«10)
0.290
•
•
i—
•—i
Alkaline Phosenneae (AtP) NTP, 2019.SM097B
shwt-ttrai |28S) 29a
Rat Sfrt^ue-Oawley 1 , N=9 t0|
0
<
»
0.312
•
•
129
•
2.8
»
•
Aiwmna tmnxmtvau, (AST 1 Cwyotrt , 2021 10176181
dw»tofm«ntal (G014-W1 G01B
P0 Rat. Sfmtgiw-OwHoy (-, N=4-«
T
I • I
•
*
¦ m i
i ~! i
»
•
SMCAt Hi al, 2C03, 1290852
shoMwm |4w«) 4wk
Rat CifCD(SD)tGS BR (•, N-10)
0.47
'•'I
•j
1.77
•
short-tam |28d) 29d
Rat S|vaqiio-Oawtoy ( , N=9-10)
0312
125
2.5
•
•
•
B
•
dwire (If) 4i»k
ftet Ci':CD(SD)K3S BR ( , N-10)
R«tCn:CD(8D)K3SBR( . N=10)
0029
0.12
0299
1.251
0
0.029
0.12
1.251
H
-•
<
' <
M
-H
N
<
H
27»*
Rat. Crt.COlSDMGS BR (., N=10)
L
<
1 ~
>" 1
—
1 «
1
' 1
Rat. Cil:CO(SD)K3S BR (¦;, N-10)
O.Z99
1,251
1 1
—<
i
i—
•
1 1
1
1 1
00 -159 -
response (%l
ISO a
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 here and here.
ALT = alanine aminotransferase; AST = aspartate aminotransferase; ALP = alkaline phosphatase; d = day; w/wk = week;
y = year; CI = confidence interval.
aTwo publications Han et al. {, 2018, 4238554} and Wan et al. 1. 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.
3-39
-------
APRIL 2024
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 3-3). 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 concentrations; this lesion was not observed in females {Curran, 2008,
757871} (Table 3-4). 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 positive dose-response relationship was observed (Table 3-5).
3-40
-------
APRIL 2024
Table 3-3. 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 3-4. 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
Table 3-5. 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
3-41
-------
APRIL 2024
Males
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} (peer-reviewed publication of data from a report by
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 3-5); 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 quantitative data 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 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;
3-42
-------
APRIL 2024
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 to 20 mg/kg/day for approximately 4 weeks showed focal or flake-like 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 in 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 as organized by mechanistic
data category (see Appendix A, {U.S. EPA, 2024, 11414344}) and source is shown in
Figure 3-15.
3-43
-------
APRIL 2024
Mechanistic Pathway Animal Human In Vitro Grand Tot
Angiogenic, Antiangiogenic, Vascular Tissue Remodeling
Atherogenesis And Clot Formation
0
0
1
1
Big Data, Non-Targeted Analysis
g
0
6
15
Cell Growth, Differentiation, Proliferation, Or Viability
13
1
25
35
Cell Signaling Or Signal Transduction
13
1
15
25
Extracellular Matrix Or Molecules
Fatty Acid Synthesis, Metabolism, Storage, Transport, Binding, B-Oxidation
17
0
10
25
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
Not Applicable/Not Specified/Review Article
1
0
0
1
Grand Total 31 2 31 58
Figure 3-15. Summary of Mechanistic Studies of PFOS and Hepatic Effects
Interactive figure and additional study details available on HAWC.
3.4.1.3.1 Nuclear 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,
3-44
-------
APRIL 2024
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
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 [xM 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 [xM, 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 [xM {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 [xM {Behr,
2020, 6305866}. However, two additional studies reported that PFOS did not significantly
increase the activity of human PPARa up to concentrations of 100 [xM in HepG2 cells
{Rosenmai, 2018, 4220319} or 250 [xM 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 [xM and by threefold at 100 [xM 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 [xM and dissociation constant of 93.7 [xM. Mouse and rat PPARy
were also activated at 100 [xM 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 PPARS in HEK29 cells at concentrations of up to 100 [xM {Behr,
3-45
-------
APRIL 2024
2020, 6305866}, and neither human nor mouse PPARy were activated by concentrations of up to
250 [xM 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 [xM, but not at a lower concentration of 10 [xM or
at higher concentrations of 40-80 [xM. 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.
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 [xM PFOS. However, a luciferase
reporter assay in HepG2 cells showed that PFOS activates human PXR with an EC so of 7.87 [xM
{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
3.4.1.3.1.4 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
3-46
-------
APRIL 2024
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
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
3-47
-------
APRIL 2024
noted that PPARy gene expression was also upregulated 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-null 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
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-mx\\ 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 (nrlh3),
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
3-48
-------
APRIL 2024
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
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.
3.4.1.3.1.5 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 ADH1B) and increases in genes that are negatively
regulated by HNF4a targets (CCM)7, AKR1B10, and PLIN2) was observed. A study in HepaRG
cells exposed to 1-100 [xM PFOS for 24 or 48 hours corroborated these findings, as
downregulations in both HNF4a and its target gene CYP7A1 were observed {Behr, 2020,
6505973}.
3-49
-------
APRIL 2024
3.4.1.3.1.6 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.
3.4.1.3.2 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 andRar) 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.
3-50
-------
APRIL 2024
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 3 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 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 3 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-hydroxy acyl Co A 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 ninefold in females (compared with twofold 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 with 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,
3-51
-------
APRIL 2024
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.
3.4.1.3.2.2.2 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.
3-52
-------
APRIL 2024
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.
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 10-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 with WT mice. In addition, there were three-to-fourfold reductions in the genes
related to lipid homeostasis that were expressed in PFOS-exposed l'para-xwxW mice compared
with 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-null mice, there was only a twofold decrease in Hnf4a, a fourfold decrease in
Ppargcla, and a threefold 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-mx\\ mice to
10 mg/kg/day PFOS for 7 days. PFOS exposure upregulated genes related to fatty acid P-
oxidation, lipid catabolism, lipid synthesis, and lipid transport in both strains; however, the
3-53
-------
APRIL 2024
increase in expression was several-fold lower in Ppara-null 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 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 with 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 datasets
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.
3-54
-------
APRIL 2024
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
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 with 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 [xM 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
with 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 [xM 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
3-55
-------
APRIL 2024
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 10 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
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 [xM 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 and ABCG1), 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 [xM 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 [xM. 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.
3-56
-------
APRIL 2024
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
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 Appendix, {U.S. EPA, 2024, 11414344}), 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
(Abcal)) when following exposure to 5 mg/kg/day PFOS for 28 days by oral gavage. However,
ATP-binding cassette subfamily 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-null 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 [j,mol/L of PFOS for 24 hours, the authors found that eRP was
induced at 10 [j,mol/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
3-57
-------
APRIL 2024
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}.
3.4.1.3.3.1 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.
3.4.1.3.4 Xenobiotic 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-mx\\ 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
3-58
-------
APRIL 2024
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 with the control group. However, there was a nonlinear 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
(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 to 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 [xM) 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 etal. {, 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 fivefold) across all concentrations
3-59
-------
APRIL 2024
{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 pM) 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 pM) in HepG2 cells.
Franco et al. {, 2020, 6315712} also assessed gene expression of two phase II enzymes,
glutathione S-transferase mu 1 (GSTM1) andUDP glucuronosyltransferase 1A1 (UGT1A1),
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 with 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.
3.4.1.3.5 Cell 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; Wielsoe, 2015, 2533367;
Florentin, 2011, 2919235; Franco, 2020, 6315712; Ojo, 2020, 6333436; Franco, 2020, 6507465;
3-60
-------
APRIL 2024
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 [xM 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
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 [xM PFOS after for 24 or 48 hours of exposure in HepaRG cells {Franco, 2020, 6315712;
Franco, 2020, 6507465}, to 45-65 [xM after 24 or 48 hours of exposure in HepG2 cells {Wan,
2016, 3981504; Ojo, 2020, 6333436}, to 417 [xM 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 [xM for 24 hours or 400 [xM 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 (Casp8)) 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,
3-61
-------
APRIL 2024
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}.
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
3-62
-------
APRIL 2024
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,
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 [xM 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 pM PFOS for 48 or
96 hours decreased the percentage of cells at the G1/G0 (non-dividing) phases of the cell cycle
3-63
-------
APRIL 2024
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 [j,g/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
100 to 278 [xM 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.
3.4.1.3.6 Inflammation 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
3-64
-------
APRIL 2024
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.
3.4.1.3.6.1 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 [xM 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 IL-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
3-65
-------
APRIL 2024
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,
TNFa and IL-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.
3-66
-------
APRIL 2024
3.4.1.3.7 Oxidative 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
(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
3-67
-------
APRIL 2024
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 administration of 10 mg/kg/day PFOS for 3 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 with 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, etal., {, 2010, 1274165}. InWT (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-nuU 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 etal. {,
2017, 3981272} dosed rat hepatocytes with 25 [iM 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
with 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 to 250 [xM. PFOS exposure led to increases in the expression of the nitric
3-68
-------
APRIL 2024
oxide synthase 2 (Nos2 or iNos) and Hmoxl genes in mouse primary hepatocytes. In human
primary hepatocytes exposed to 100 [xM 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 [xM 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 [xM PFOS, up to
117% with 1,000 [xM, 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
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 GSSG to GSH.
In a third study using primary mouse hepatocytes, Xu et al. {, 2020, 6316207} exposed cultured
cells to 10, 100, 500, or 1,000 [xM 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 [xM 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; Wielsoe, 2015, 2533367; Shan, 2013, 2850950; Florentin, 2011,
2919235}. Two studies reported increases in ROS levels following PFOS exposure {Wan, 2016,
3981504; Wielsoe, 2015, 2533367}, while two studies did not observe statistical differences in
ROS levels following 1- or 24-hour PFOS exposures up to 400 [xM {Florentin, 2011, 2919235}
or following 3-hour PFOS exposures up to 400 [xM {Shan, 2013, 2850950}. Wanetal. {, 2016,
3981504} dosed HepG2 cells with either 0, 10, 20, 30, 40, or 50 [xM PFOS for 24 hours or with
50 [xM 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 with 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
3-69
-------
APRIL 2024
to play a key role in oxidative stress via the Nrf2 antioxidant pathway, increased nearly 12-fold
following 24-hour 50 [xM 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.
Wielsoe 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 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. 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. The positive associations with ALT are also supported by
the recent meta-analysis of 25 studies in adolescents and adults {Costello, 2022, 10285082}.
However, in several studies, the associations were not large in magnitude.
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.
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
3-70
-------
APRIL 2024
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 entirely
explain the observed associations.
Evidence for other liver enzymes and in children and adolescents is less consistent. Results for
functional measures of liver toxicity from epidemiological studies, 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. Associations for functional hepatic outcomes
such as liver disease were also less consistent than the associations between PFOS and ALT.
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 the increases in serum enzymes were
generally under twofold (100% change relative to control) as compared with controls {Seacat,
2003, 1290852; Curran, 2008, 757871; Butenhoff, 2012, 1276144; Xing, 2016, 3981506; Yan,
2014, 2850901; NTP, 2019, 5400978; Han, 2018, 4355066}. However, across the animal
toxicological database, 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 etal. {, 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 receptor PPARa,
3-71
-------
APRIL 2024
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.
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 were 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.
3-72
-------
APRIL 2024
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 3-6). 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
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.
3-73
-------
APRIL 2024
Table 3-6. Evidence Profile Table for PFOS Exposure and Hepatic Effects
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors That Increase Factors That Decrease
Certainty Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
Evidence From Studies of Exposed Humans (Section 3.4.1.1)
Serum biomarkers of
hepatic injury
12 Medium confidence
studies
3 Low confidence studies
In adults, significant
increases in ALT were
observed in medium
confidence studies (6/8).
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
occupational populations
and children were mixed.
However, significant
increases in ALT were
observed in one
occupational study in men
(1/2), and significant
increases in AST and
GGT were observed in
female children (1/3).
• Medium confidence
studies that reported
an effect
• Consistent direction of
effect for ALT
• Coherence of findings
across biomarkers
• Inconsistent direction ®©o
of effect in children. Moderate
®©o
Evidence Indicates (likely)
Evidence for hepatic
effects is based on
increases in ALT in
adults. Other supporting
evidence includes
increases in other liver
Liver disease or injury
3 Medium confidence
studies
2 Low confidence studies
Findings for markers of
liver inflammation were
mixed in medium
confidence studies (1/2).
In adults, one study
reported nonsignificant
decreased odds of lobular
inflammation (1/1). The
only study in children
reported significantly
• Medium confidence
studies
*Low confidence
studies
*Limited number of
studies examining the
outcome
»Imprecision of
findings
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
enzymes such as AST and for increased incidence of
GGT, and histological hepatic nonneoplastic
changes in children, such lesions, increased liver
as non-alcoholic steatosis, weight, and elevated serum
Minor uncertainties biomarkers of hepatic injuiy
remain regarding mixed 'n animal models. Although
associations between PFOS
exposure and other serum
biomarkers of hepatic injury
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
support for the biological
plausibility of the
phenotypic effects observed
liver enzyme findings in
children and limited
availability of high-
quality studies on liver
disease.
3-74
-------
APRIL 2024
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
increased odds of non-
alcoholic steatosis while
associations with other
histological markers of
liver injury were
generally positive but less
precise. Both low
confidence occupational
studies reported
nonsignificant increases
in liver disease (2/2), but
findings were generally
imprecise.
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
status. No studies were
conducted in children.
• Medium confidence
studies that reported
an effect
• Consistent direction of
effect for albumin
• Low confidence study
• Limited number of
studies examining the
outcome
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 Studies (Section 3.4.1.2)
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.
Liver histopathology Histopathological
2 High confidence studies alterations in the liver
5 Medium confidence were reported in rodents
studies or non-human primates
exposed to PFOS for
varying durations (6/7).
Hepatocellular
hypertrophy was most
• High and medium
confidence studies
• Consistent direction of
effects across study
design, sex, and
species
• Dose-dependent
response
• No factors noted
©0©
Robust
Evidence is based on 20
high or medium
confidence animal
toxicological studies
indicating increased
3-75
-------
APRIL 2024
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
consistently observed
across sex, species, and
duration of exposure and
in a dose-responsive
manner (5/7). 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).
• 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 groups
(i.e., hypertrophy) or
are considered severe
(i.e., cell or necrosis
and cystic
degeneration)
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
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)
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
• High and medium
confidence studies
• Consistent direction of
effects across study
design, sex, and
species
Confounding variablesWQ i§ht and hepatocellular
such as decreases in hypertrophy adverse when
body weights 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
• Limited number of
studies examining
specific endpoints
• Inconsistent direction
of effects between
sexes
3-76
-------
APRIL 2024
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
exposure increased
enzyme levels. Several
studies found increased
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
tissue was an adverse, and
not adaptive, response.
Mechanistic Evidence and Supplemental Information (Section 3.4.1.3)
Biological Events or
Pathways
Summary of Key Findings, Interpretation, and Limitations
Evidence Stream
Judgment
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 wild-type and PPARa-null mice.
Limitations:
• Conflicting results have been reported for activation of the PPARa
signaling pathway in vitro between human and rodent cells.
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
PPARa. The evidence also
suggests a role for
PPARa-independent
pathways in the MOA for
noncancer liver effects of
PFOS, particularly CAR
activation and decreased
expression of HNF4a.
3-77
-------
APRIL 2024
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
Molecular or cellular Key findings and interpretation:
initiating events — other • Activation of CAR in vivo in rodents and in vitro in both human and
pathways rodent 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.
Notes: ALP = alkaline phosphatase; ALT = alanine aminotransferase; AST = aspartate aminotransferase; CAR = constitutive androstane receptor; GGT = gamma-glutamyl
transpeptidase; HNF4a = 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.
3-78
-------
APRIL 2024
3.4.2 immune
EPA identified 47 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, 29 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, one was classified as high confidence, nine as medium confidence, one as low
confidence, and two 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 10 studies (11 publications8) from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}
that investigated the association between PFOS and immune effects. Study quality evaluations
for these 11 studies are shown in Figure 3-16. Results from studies summarized in the 2016
PFOS HESD are described in Table 3-7 and below.
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. Another study on Faroese children conducted a
pilot investigation on the association between elevated PFOS exposure and autoantibodies to
antigens indicating tissue damage, but the results were unclear {Osuna, 2014, 2851190}.
Prenatal PFOS exposure was associated with diminished vaccine response in a different birth
cohort study {Granum, 2013, 1937228}. Decreases in the anti-rubella antibody response were
significantly associated with elevated prenatal PFOS concentrations among 3-year-old children.
Stein et al. {2016, 3108691} reported significant inverse associations between PFOS exposure
8 Okada, 2012, 1332477 reports overlapping eczema results with Okada, 2014,2850407.
3-79
-------
APRIL 2024
and mumps and rubella antibody concentrations in seropositive adolescents (12-19 years old)
from multiple NHANES cycles (1999-2000, 2003-2004), but no association was observed for
measles. 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).
Evidence based on studies of infectious disease in children from the 2016 PFOS HESD was
limited. In the Danish National Birth Cohort (DNBC) study, Fei et al. {, 2010, 1290805}
reported nonsignificant increases in risk of hospitalizations for infectious diseases in children
4 years and older, but no association was observed at younger ages. In sex-stratified analyses the
risk for hospitalization for infectious disease was significantly increased in girls (IRR = 1.18,
95% CI: 1.03, 1.36), while findings for boys were null. No association was observed for
gastroenteritis or common cold in children from the Norwegian Mother and Child Cohort study
(MoBa) {Granum, 2013, 1937228}.
Dong etal., 2013, 1937230-
l
+
++
—I—
+
—I—
+
++
—I—
+
+
+
Fei et al., 2010, 1290805-
+
++
+
+
++
+
+
+
Grandjean et al., 2012, 1248827 -
B
B
++
+
+
+
+
+
Granum et al., 2013, 1937228 -
++
++
+*
+
+
-
+*
Humblet et al., 2014, 2851240 -
++
B
B
+
++
+
+
+
Looker etal., 2014, 2850913-
+
+
++
+
+
+
+
+
Okada etal., 2012, 1332477-
+
+
+
+
+
+
+
+
Okada et al., 2014, 2850407 -
+
+
+
+
+
+
+
+
Osunaet al., 2014, 2851190-
-
-
++
-
-
+
-
-
Stein etal., 2016, 3108691 -
++
++
+
+
+
+
+
+
Wang etal., 2011, 1424977-
-
+
+
+
+
+
+
+
Legend
B
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 3-16. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Immune Effects Published Before 2016 (References in 2016 PFOS
HESD)
3-80
-------
APRIL 2024
Interactive figure and additional study details available on HAWC.
3-81
-------
APRIL 2024
Table 3-7. Associations Between Elevated Exposure to PFOS and Immune Outcomes From Studies Identified in the 2016
PFOS HESD
Reference, confidence
Study
Design
Population
Tetanus
Aba
Diphtheria
Aba
Rubella
Aba
Influenza
Aba
Infectious
Diseaseb
Asthmab
Eczemab
Autoimmune
Diseaseb
Dong, 2013, 1937230
Case-control Children
NA
NA
NA
NA
NA
tt
NA
NA
Medium
Fei, 2010, 1290805
Cohort
Children
NA
NA
NA
NA
t
NA
NA
NA
Medium
Grandjean, 2012,
1248827
Cohort
Children
44
44
NA
NA
NA
NA
NA
NA
Medium
Granum, 2013,
1937228
Cohort
Children
-
NA
44
NA
-
NA
NA
NA
Mixed
Humblet, 2014,
2851240
Cross-
sectional
Adolescents
NA
NA
NA
NA
NA
-
NA
NA
Medium
Looker, 2014, 2850913
Cohort
Children
NA
NA
NA
-
NA
NA
NA
NA
Medium
Stein, 2016, 3108691
Medium
Cross-
sectional
Children
NA
NA
44
NA
NA
t
NA
NA
Okada, 2014, 2850407
Cohort
Children
NA
NA
NA
NA
NA
NA
-
NA
Medium
Wang, 2011, 1424977
Cohort
Children
NA
NA
NA
NA
NA
NA
t
NA
Medium
Notes'. Ab = antibody; NA = no analysis was for this outcome was performed; | = nonsignificant positive association; ft = significant positive association; j = nonsignificant
inverse association; jj = significant inverse association; - = no (null) association.
Osuna, 2014,2851190 analyzed autoantibody response to indicators of tissue damage and was not included in the table.
Okada, 2012,1332477 reports overlapping eczema results with Okada, 2014, 2850407, which was considered the most updated data.
a Arrows indicate the direction in the change of the mean response of the outcome (e.g., j indicates decreased mean birth weight).
b Arrows indicate the change in risk of the outcome (e.g., | indicates an increased risk of the outcome).
Granum, 2013,1937228 was rated medium confidence for antibody response, common cold, and gastroenteritis, and low confidence for all other outcomes.
3-82
-------
APRIL 2024
There are 28 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}.
3-83
-------
APRIL 2024
,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 Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Immunosuppression Effects
Interactive figure and additional study details available on HAWC.
3-84
-------
APRIL 2024
\ec^
,^Ve*6 *
,00
Kielsen etal., 2016, 4241223
Kvalem et al„ 2020, 6316210
Lopez-Espinosa et al., 2021, 7751049
Manzano-Salgado et al., 2019, 5412076
Mogensen etal., 2015, 3981889
Pilkerton etal., 2018, 5080265-
Shih etal., 2021, 9959487
Stein etal., 2016, 3860111
Timmermann etal., 2020, 6833710
Timmermann et al., 2021, 9416315
Wang etal., 2022, 10176501
Zeng etal., 2019, 5081554
Zeng etal,, 2020, 6315718
Zhang et al. 2023, 10699594
Legend
p
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 3-18. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Immunosuppression Effects (Continued)
Interactive figure and additional study details available on HAWC.
3-85
-------
APRIL 2024
3.4.2.1.1.1 Vaccine Response
Thirteen studies (14 publications9'10) 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; Zhang, 2023, 10699594}. In adults, two studies investigated antibody response to
diphtheria and tetanus {Kielsen, 2016, 4241223; Shih, 2021, 9959487}, one study investigated
hepatitis A and B vaccine response {Shih, 2021, 9959487}, one study investigated adult flu
vaccine response {Stein, 2016, 3860111}, one study measured rubella antibodies in both
adolescents (aged 12 and older) and adults {Pilkerton, 2018, 5080265}, and one study measured
rubella, measles, and mumps antibodies in adolescents {Zhang, 2023, 10699594}. 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}, six studies were
medium confidence {Grandjean, 2017, 4239492; Timmermann, 2020, 6833710; Mogensen,
2015, 3981889; Timmermann, 2021, 9416315; Shih, 2021, 9959487; Zhang, 2023, 10699594},
four were low 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 {Kielsen, 2016, 4241223}.
Of the studies that measured antibody response to vaccination in children and adolescents, four
studies were cohorts {Timmermann, 2020, 6833710; Grandjean, 2017, 3858518; Grandjean,
2017, 4239492; Mogensen, 2015, 3981889}, and four were cross-sectional {Abraham, 2020,
6506041; Timmermann, 2021, 9416315; Pilkerton, 2018, 5080265; Zhang, 2023, 10699594}
(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. In addition, two studies
measured antibody response to measles vaccination {Timmermann, 2020, 6833710; Zhang,
2023, 10699594}, two studies measured antibody response to rubella vaccination {Pilkerton,
2018, 5080265; Zhang, 2023, 10699594}, one study measured antibody response to mumps
vaccination {Zhang, 2023, 10699594}, 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 3-8 and Appendix D {U.S. EPA,
2024, 11414344}. 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-
9 Multiple publications of the same study: the study populations are the same in Grandjean et al. {, 2017, 3858518} and
Mogensen et al. {,2015,3981889}.
10 Zhang {, 2023, 10699594} analyzes NHANES cycles 2003-2004 and 2009-2010 partially overlapping with Pilkerton {, 2018,
5080265} and Stein {, 2016, 3108691} which both analyze cycles 1999-2000 and 2003-2004.
3-86
-------
APRIL 2024
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 3-8). 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. Results from all medium and high
confidence studies on tetanus and diphtheria antibody response in children are provided in
Figure 3-19 and Figure 3-20.
Reference,
Confidence Rating
Exposure Levels
Comparison
Sub-
population
Exposure
Age
Outcome
Age
Effect Estimate
EE
-50 0 50 100
Grandjean et al.
(2017a. 3858518),
High
PFOS at 7 years median (25th-75tti percent change (per doubling of
percentile^ 15.3 ng/mL (12.4-19.0 ng/mL) PFOS)
Age 7
Age 13
30 1 •
1
PFOS at 13 years median (25th-75th
percentile)=6.7 ng/mL {5.2-8.5 ng/mL)
percent change (per doubling of
PFOS)
Age 13
Age 13
22.2 1 •
1
Grandjean et al. {2012. Age 5 PFOS: Geometric mean=16.7
1248827), Medium ng/mL <25th-75th percentile=13.5-21.1
Percent difference (per doubling in
age 5 PFOS)
Age 5
Age 7
1
-23.8 1-
1
ny/itii.;
Adj for Age 5
Ab
Age 5
Age 7
1
-11.4 » 1
1
Pre-booster
Age 5
Age 5
T
-11.9 a—1—
1
Post-booster
Age 5
Age 5
T
-28.5 • 1-
1
Maternal PFOS: Geometric mean=27.3
ng/mL {25th-75th percentile=23.2-33.1
ng/mL)
Percent difference {per doubling in
maternal PFOS)
""
Prenatal
Age 7
1
35.3 -1 a
1
Adj for Age 5
Ab
Prenatal
Age 7
r
33.1 1 •
I
Age 5
1
-2.3 «
1
Pre-booster
Prenatal
Age 5
1
-10.1 •—1
1
Grandjean et al.
(2017b. 4239492),
Medium
Percent change (per doubling of
PFOS)
Cohort 3
Age 1.5
Age 5
1
-8.05 •-!
j
Age 5
Age 5
-11.86 •—1—
1 1
Cord blood
Age 5
1
-10.09 •—1
1
Cohort 3 and
5
Age 1.5
Age 5
*-|
Age 5
Age 5
l
-10.52 •—1-
1
Cord blood
Age 5
1
-10.55 •—1-
1 I 1
Cohort 5
Cord blood
Age 5
j-
-10.84 •—1—
1
Median = 4.7 ng/mL {25th - 75th
percentile: 3.5 - 6.3 ng/mL)
Percent change (per doubling of
PFOS)
Cohort 5
Age 5
Age 5
1
-9.08 •—1
1
Median = 7.1 ng/mL (25th - 75th
percentile: 4.5 -10.0 ng/mL)
Percent change (per doubling of
PFOS)
Cohort 5
Age 1.5
Age 5
1
-7.03 »-J—
1
Mogensen et al. (2015. median=15.5 ng/ml (25th-75th
3981889), Medium percentiie=12.8-19.2 ng/ml)
Percent change per doubling of
PFOS
Age 7
Age 7
Age 7
T
-9.1 •—1
1
Timmermann et al.
(2022, 9416315),
Medium
median=8.68 ng/mL (25th - 75th
percentiles: 6.52 -12.23 ng/mL)
percent difference (per unit increase Ages 7-12
in child PFOS concentration)
Age 7-12
Age 7-12
1
-3
1
median=19.16 ng/mL (25th - 75th
percentiles: 15.20 - 24.06 ng/mL)
percent difference (per unit increase
in maternal PFOS concentration)
Ages 7-12
Prenatal
Age 7-12
1'
2 ¦+¦
1
-50 0 50 100
Figure 3-19. Overall Tetanus Antibody Levels in Children from Epidemiology Studies
Following Exposure to PFOS
Interactive figure and additional study details available on HAWC.
Grandjean et al. {, 2012, 1248827} was reviewed as a part of the 2016 PFOS HESD.
3-87
-------
APRIL 2024
Reference, r . . Sub- Exposure Outcome r-r-
Confidence Rating Exposure Levels Comparison population Age Age
Effect Estimate
-60 -40 -20 0 20 40 60
Grandjean et al. PFOS at 7 years median (25th-75th percent change (per doubling of - AQ67 Age 13
(2017a. 3858518), percentile)=15.3 ng/mL (12.4-19.0 ng/mL) PFOS) ~238
Higfl PFOS at 13 years median (25th-75th percent change (per doubling of ~ Age 13 Age 13 77
percentile)=6.7 ng/mL (5.2-8.5 ng/mL) PFOS)
Grandjean eta"!. (2012, Age 5 PFOS: Geometric mean=16.7 Percent difference (per doubling in - Age 5 Age 7
1248827), Medium ng/mL (25th-75th percentile=13.5-21.1 age 5 PFOS) ~27 6
n9/mL) Adjfor Age5 Age 5 Age 7 ~~
Ab '206
Pre-booster Age 5 Age 5
Post-booster Age 5 Age 5 ^ ^ ^
Maternal PFOS: Geometric mean=27.3 Percent difference (per doubling in - Prenatal Age 7
ng/mL (25th-75th percentile=23.2-33.1 maternal PFOS) ~19 7
ng/mL) Adj for Age 5 Prenatal Age 7
Ab "10
Pre-booster Prenatal Age 5 3g g
Post-booster Prenatal Age 5
-20.6
Grandjean et al. - Percent change (per doubling of Cohort 3 Age 1.5 Age 5
(2017b. 4239492), PFOS) 21 21
Medium Age 5 Age 5
-16.02
Cord blood Age 5
-38.64
Cohort 3 and Age 1.5 Age 5
5 15.07
Age 5 Age 5 ~ 34
Cord blood Age 5
-24.47
Cohort 5 Cord blood Age 5
-14
Median = 4.7 ng/mL (25th - 75th Percent change (per doubling of Cohort5 Age 5 Age 5
percentile: 3.5 - 6.3 ng/mL) PFOS) 1717
Median = 7.1 ng/mL (25th - 75th Percent change (per doubling of Cohort 5 Age 1 5 Age 5
percentile: 4.5 -10.0 ng/mL) PFOS) 17 55
Mogensen et al. (2015, median=15.5 ng/ml (25th-75th Percent change per doubling of Age 7 Age 7 Age 7
3981889), Medium percentile=12.8-19.2 ng/ml) PFOS ~30 3
Timmermann et al. median=8.68 ng/mL (25th - 75th percent difference (per unit increase Ages 7-12 Age 7-12 Age 7-12
(2022,9416315), percentiles: 6.52 -12.23 ng/mL) in child PFOS concentration) ~9
Medium median=19.16 ng/mL (25th-75th percent difference (per unit increase Ages 7-12 Prenatal Age 7-12
percentiles: 15.20 - 24.06 ng/mL) in maternal PFOS concentration) 1
--
-60 -40 -20 0 20 40 60
Figure 3-20. Overall Diphtheria Antibody Levels in Children from Epidemiology Studies
Following Exposure to PFOS
Interactive figure and additional study details available on HAWC.
Grandjean et al. {, 2012, 1248827} was reviewed as a part of the 2016 PFOS HESD.
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 PFOS 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 does not show that confounding across PFAS is likely
to completely explain the observed effects.
3-88
-------
APRIL 2024
Table 3-8. Associations between PFOS Exposure and Vaccine Response in Faroe Island Studies
Diphtheria Antibody Associations with PFOS by Age at Tetanus Antibody Associations with PFOS by Age at
Exposure Assessment Assessment
timing levels ^ 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)
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, logPFOS)0
concentration at
cohort, logPFOS)0
concentration at age
age 5 years)b
5 years)b
Birth
IJ, (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, logPFOS)0
age 5, logPFOS)0
3-89
-------
APRIL 2024
Diphtheria Antibody Associations with PFOS by Age at Tetanus Antibody Associations with PFOS by Age at
Exposure Assessment 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)s
| (C3; sex, age at
5)g
antibody assessment,
antibody assessment,
booster type at
booster type at age
age 5)g
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)s
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.
cBudtz-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.
gGrandjean et al. {, 2017, 3858518}; medium confidence.
3-90
-------
APRIL 2024
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).
Two medium cross-sectional studies of adolescents examined associations between elevated
levels of PFOS and vaccine response {Pilkerton, 2018, 5080265; Zhang, 2023, 10699594}.
Inverse associations were observed in cross-sectional analyses in adolescents from NHANES
(2003-2004; 2009-2010) for rubella, mumps, and measles {Zhang, 2023, 10699594}, including
a significant reduction in the antibody response to rubella per 2.7-fold increase in serum PFOS.
No association was observed for rubella vaccine response in the other cross-sectional study of
adolescents {Pilkerton, 2018, 5080265}, however, an overlapping study {Stein, 2016, 3108691}
on adolescents from the same NHANES cycles (i.e., 1999-2000 and 2003-2004) reported a
significant inverse association for rubella antibody response in seropositive adolescents.
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, 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 study of the adult population in Pilkerton et al. {, 2018, 5080265} was
considered low confidence as the analysis suffered from potential exposure misclassification due
to concurrent exposure and outcome measurements, considering the amount of time since rubella
vaccination in childhood. This was less of a concern for the study of adolescent participants,
which 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,
3-91
-------
APRIL 2024
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 protective 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.
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 for 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, 10 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 1 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
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.
3-92
-------
APRIL 2024
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
Appendix D {U.S. EPA, 2024, 11414344}. The most commonly examined type of infections
was respiratory, including pneumonia/bronchitis, upper and lower respiratory tract, throat
infections, 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 {Ait Bamai, 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 {Ait Bamai, 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 Appendix D {U.S. EPA, 2024, 11414344}. 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 log2 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 to 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
3-93
-------
APRIL 2024
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.
3.4.2.1.2 Immune 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
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}.
No association was observed for eczema in a Hokkaido birth cohort study {Okada, 2014,
2850407}.
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 3-21.
3-94
-------
APRIL 2024
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 3-21. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Immune Hypersensitivity Effects
Interactive figure and additional study details available on HAWC.
3-95
-------
APRIL 2024
Thirteen studies (15 publications)12 examined asthma (or asthma symptoms) and PFOS exposure.
Ten of these studies were cohorts {Averina, 2019, 5080647; Beck, 2019, 5922599; Gaylord,
2019, 5080201; 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, {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 3-21). 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 Appendix D, {U.S. EPA, 2024, 11414344}).
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 children of the same ages who had received their MMR
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.
3-96
-------
APRIL 2024
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 with 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 nonsignificant inverse associations with asthma {Manzano-Salgado, 2019,
5412076; Smit, 2015, 2823268}, and in one study, all results were nonsignificant {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 versus 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 Appendix D {U.S.
EPA, 2024, 11414344}.
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 one study measured symptoms at 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) and self-reported food allergies and reported statistically
3-97
-------
APRIL 2024
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 Appendix D {U.S. EPA, 2024, 11414344}.
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.
3.4.2.1.3 Autoimmune 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 3-22.
Four case-control studies examined PFOS exposure and autoimmune diseases (Figure 3-22).
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 Appendix D, {U.S. EPA, 2024,
11414344}). 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,
3-98
-------
APRIL 2024
particularly control selection, was not 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 1 to 2 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.
Ammitzboll et al., 2019, 5080379 -
-
+
+
+
+
+
+
-
Gaylord et al., 2020, 6833754 -
+
+
++
+
+
+
-
+
Sinisalu etal., 2020, 7211554-
+
+
+
-
+
+
-
-
Steenland et al., 2018, 5079806 -
+
-
+
-
+
+
+
-
Figure 3-22. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure 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 with those with MS. Serum PFOS concentrations were 17% lower (95% CI:
-27%, -6%; p = 0.004) in healthy controls compared with 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 with 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
observed among adult cases of ulcerative colitis compared with healthy controls {Steenland,
2018, 5079806}.
3-99
-------
APRIL 2024
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 3-23.
3-100
-------
APRIL 2024
vC.®
Curran etal., 2008, 757871 -
Dong etal.,2011, 1424949-
Han etal., 2018, 4355066-
Lefebvre et al„ 2008, 1276155 -
Li etal., 2021, 7643501 -
Lvetal., 2015, 3981558-
NTP, 2019, 5400978
Seacat et al., 2002, 757853
Seacat et al., 2003, 1290852
Thomford, 2002, 5432419 -
Xing etal., 2016, 3981506-
Yang et al., 2021, 7643494 -
Zhong et al., 2016, 3748828 -
++
NR
NR
++
+
NR
++
+
NR
+
NR
NR
+
+
NR
++
+
NR
Legend
B
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)
F
Not reported
*
Multiple judgments exist
Figure 3-23. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFOS Exposure 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
changes in spleen and/or thymus weights, extramedullary hematopoiesis, perturbations in activity
3-101
-------
APRIL 2024
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.
3.4.2.2.1 Organ 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 4 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
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}.
3-102
-------
APRIL 2024
Similarly, relative spleen weight in male BALB/c mice was not affected at the end of a 3-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 3-24).
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 3-24).
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 1-17 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 3-24).
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).
3-103
-------
APRIL 2024
PFOS Immune Effects - Thymus Weights
Study Design Observation Time Animal Description Dose (mg/kg/day) > "Li oiatiStioal.y significant 9 Not sUifiStioally sigriificantr-^ 95'o CI j
Thomu., Wemt Absolute Ltjfebvrb et =ai <£QU8 1276155 biiort tefr-i ^JEki; 28ii
NTP, 2019,5400978 short-term 10) 1
0.312
0.625
1.25
2.5
Thymus Weight. Relative Zhong ct nl„ 2016, 3748828 developmental (G01-17) PNW4 F1 Mousa. C57BU6 { : , N=12) 0
F1 Mouse, C57BU6 ( •, N=12) 0
F1 Mouse. C57BU6 ( N=12) 0
F1 Mouse. C57BU6 {' N=12) 0
Mouse, C57BI/6 (,?. N=6) 0
Rat, Sprague-DawleyN=1S) 0
0.14
1.33
3.21
6.34
Rat. SpracfUG~Dswley{". N»15) 0
0.15
1.43
3.73
7.58
Rat, Sprague-Dawley N=10) 0
0.312
0.625
1.25
Rat. Sprngue-Dawiey ( \ N=9-10) 0
0.312
Q.625
1.25
2.5
e .
+ *-
-50 -40 -30 -20 -10 0 10
Percant control response (%)
Figure 3-24. 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; Fi = first generation
3-104
-------
APRIL 2024
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 3-25). In general, short-
term and subchronic studies have observed histopathology such as extramedullar 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 PFOS 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 nonneoplastic lesions were observed
following exposure to 2.5, 5, or 10 mg/kg/day PFOS for 3 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.
3-105
-------
APRIL 2024
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 3-25. 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}.
3.4.2.2.3 Circulating 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
3-106
-------
APRIL 2024
neutrophils in the 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}.
3.4.2.2.4 Natural 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 that NK 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 3-9).
3-107
-------
APRIL 2024
Table 3-9. 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
Keiletal. {,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.
3.4.2.2.5 Spleen Cellularity
Splenocyte sub-classes were quantified in several rodent studies (Figure 3-26). 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 nonsignificant 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
3-108
-------
APRIL 2024
percentages of T-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 1-17, and then quantified various immune cell populations in male
and female pups. Decreased splenic cell subpopulations (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
Zhong et al.. 2016. 3748628
Study Design Observatio
developmental (GD1-17) PNW4
Zhong et al., 2016, 3748828 developmental (GD1-17) PNW4
PNW8
Zhong at al., 2016, 3748828 devolopmantal (GD1-17) PNW4
PNW8
Zhong et al.. 2016. 3748828 developmental (GD1-17) PNW4
PNW8
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 significanl change A Significant increase ^ Significant decrease
Concentration (mg/kg/day)
Figure 3-26. 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, 3748828} reported data on both splenic and thymic lymphocyte populations for the same experimental
animals. Results are shown in separate figures.
3.4.2.2.6 Thymus Cellularity
Thymus cell populations were less sensitive to the effects of PFOS compared with 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 3-27). 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-
3-109
-------
APRIL 2024
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 1-17, and immune cell populations were quantified in male and female pups at 4
and 8 weeks after birth. Decreased thymic lymphocyte subpopulations (CD4+, and CD4VCD8"
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 1-17 {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 (
-------
APRIL 2024
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-
7.58 mg/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 with 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 3-10).
Table 3-10. 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. {,
28 days
0,0.14, 1.33,3.21,6.34
M
n.s.
2008, 1276155}b
(males) orO, 0.15, 1.43,
3.73, 7.58 (females)
F
n.s.
Dong et al. {, 2009,
60 days
0, 0.008, 0.083, 0.417,
M
1
142495 l}a
0.833, 2.083
0.083-2.083
Dong et al. {, 2011,
60 days
0,0.008,0.0167,0.083,
M
1
1424949}3
0.417, 0.833
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
3-111
-------
APRIL 2024
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
1332422 }a
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.
b Keyhole 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 1-17 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. Keil et al. {, 2008, 1332422} reported that the primary IgM
response to SRBC was significantly suppressed by 53% at 8-weeks in males from the
5 mg/kg/day exposure group. In females, the primary IgM response was not altered {Keil, 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
3-112
-------
APRIL 2024
were increased 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
with 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 by mechanistic data category (see Appendix
A, {U.S. EPA, 2024, 11414344}) and source is shown in Figure 3-28.
Mechanistic Pathway
In Vitro Grand Total
Big Data, Non-Targeted Analysis
1
0
0
1
Cell Growth, Differentiation, Proliferation, Or Viability
6
0
9
13
Cell Signaling Or Signal Transduction
2
0
4
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 3-28. Summary of Mechanistic Studies of PFOS and Immune Effects
Interactive figure and additional study details available on HAWC.
3.4.2.3.1 Mechanistic Evidence for PFOS-Mediated Effects on the Immune System
Since the 2016 PFOS 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.
3-113
-------
APRIL 2024
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
3-114
-------
APRIL 2024
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,
3-115
-------
APRIL 2024
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 4 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
3-116
-------
APRIL 2024
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} andPennings 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}. The N-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 [xM. Exposure to 75, 150, and 300 |iM PFOS for
3-117
-------
APRIL 2024
2, 4, 6, 8, 10, or 12 hours was associated with increased reactive oxygen species (ROS)
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 with 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 nonsignificant 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}.
3-118
-------
APRIL 2024
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-
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 upregulation 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 3-9 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 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.
3-119
-------
APRIL 2024
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
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 with 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 Appendix, {U.S. EPA, 2024, 11414344}) 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 [xM 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,
3-120
-------
APRIL 2024
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
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 3-11. 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 3-11. Effects of PFOS Exposure on Pro-Inflammatory Cytokines and Markers of
Inflammation
Study
„ . „ Cytokine or
pecies or e inflammatory Matrix and Measurement
ype Marker
Direction of Change
Following PFOS
Exposure
Mitro et al. {,
2020,6833625}
Human females IL-6
3 years
blood protein (ELISA)
None
postpartum,
Project Viva
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
serum protein
(Multispot Immunoassay)
None
C3a
serum protein (ELISA)
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,Human bronchial IL-la
5918817} epithelial cell line
culture supernatant protein
(ELISA)
T
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
Hanetal. {,2018, Sprague-Dawley IL-6
4355066} male rats
TNF-a
serum protein (ELISA)
serum protein (ELISA)
T
T
3-121
-------
APRIL 2024
„. , Species or Cell
S"'dy Type
Cytokine or
Inflammatory
Marker
Matrix and Measurement
Direction of Change
Following PFOS
Exposure
Su et al. {,2019, ICR male mice
IL-6
serum protein (ELISA)
T
5080481}
TNF-a
serum protein (ELISA)
T
Han et al. {,2018, Primary rat
IL-6
cellular mRNA (PCR) and
T
4355066} hepatocytes and
Kupffer cells
culture supernatant protein
(ELISA)
TNF-a
cellular mRNA (PCR) and
culture supernatant protein
(ELISA)
T
Zhu et al. {,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.
3-122
-------
APRIL 2024
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 |xM 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
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 [xM 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 [xM. 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 [xM PFOS {Han, 2018, 4355066}. Zhu
et al. {, 2015, 2850996} reported that treatment with 10 [xM 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 [xM 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 JNK 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.
3-123
-------
APRIL 2024
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
(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 non-alcoholic 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 upregulation 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}.
3-124
-------
APRIL 2024
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,
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 with 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, and a change in antibody
levels of approximately 11% per 2.7-fold increase of PFOS exposure was observed in
3-125
-------
APRIL 2024
adolescents from NHANES. 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 following routine vaccinations in
children. Evidence in adults does not indicate an association with immunosuppression, but high
confidence studies are not available in these populations.
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
subgroups. 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" versus "current" asthma, and no studies beyond the Dong et al. {,
2013, 1937230} described in the 2016 PFOS HESD 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 PFOS HESD, 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
3-126
-------
APRIL 2024
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
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.
3.4.2.4.1 Evidence 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 3-12). 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. Given 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 lifestages represents a major source of uncertainty in
the immunotoxicity database of PFOS.
3-127
-------
APRIL 2024
Table 3-12. Evidence Profile Table for PFOS Exposure and 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
20 Medium confidence
studies
8 Low confidence
studies
2 Mixeda confidence
studies
Studies conducted in the
Faroe Islands examined
antibody levels among
children at various
timepoints compared with
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. Similarly,
antibody levels against
rubella (2/2) were
significantly decreased in
medium confidence
studies of children.
Findings in the four
studies examining adults
were less consistent than
children. Infectious
disease was examined in
14 studies of children.
Studies examining
infections of the
respiratory system
observed some positive
associations (5/12),
although many findings
from other studies were
• High and medium
confidence studies
that reported effects
• Consistent direction
of effect
• Coherence of
findings across
antibody response
and increased
infectious disease
»Imprecision of findings
®©o
Moderate
Evidence for immune
effects is based on
decreases in childhood
antibody responses to
pathogens such as
diphtheria and tetanus,
and some effect for
rubella. 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, 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 under
relevant exposure
circumstances, the human
evidence underlying this
possibility is uncertain and
with limited support from
animal or mechanistic
studies.
3-128
-------
APRIL 2024
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase Factors that Decrease
Certainty Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
not precise. Findings for
infectious disease in
adults were mixed, with
two studies reporting
inconsistent results for
COVID-19 infections.
Immune
hypersensitivity
1 High confidence study
20 Medium confidence
studies
4 Low confidence
studies
3 Mixeda confidence
studies
Examination of immune
hypersensitivity includes
outcomes such as asthma,
allergies, and eczema.
Increased odds of asthma
were reported in multiple
medium confidence
studies (7/12), although
associations were often
inconsistent by
subgroups. Low
confidence studies
supported the findings of
increased odds of asthma
or higher exposure levels
among asthmatics,
although results were not
always consistent or
precise. Nine studies
examined allergies,
rhinitis, or
rhinoconjunctivitis. Some
positive associations (3/9)
were observed, although
this varied by outcome
timing and were at times
inconsistent. Ten studies
examined eczema, and
• High and medium
confidence studies
• Consistent direction
of effect for asthma
across medium
confidence studies
• Lnconsistent direction
of effect between
subpopulations
Human relevance and
other inferences:
Given 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 lifestages
represents a major source
of uncertainty in the
immunotoxicity database
of PFOS.
3-129
-------
APRIL 2024
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase Factors that Decrease
Certainty Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
results were generally
mixed.
Autoimmune disease
1 Medium confidence
study
3 Low confidence
studies
Lower exposure levels
were observed in healthy
controls compared with
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 with healthy
controls.
»No factors
identified
• 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 aKLH 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
hematopoiesis were
3-130
-------
APRIL 2024
Evidence Stream Summary and Interpretation
Studies and
Interpretation
Summary and Key
Findings
Factors that Increase Factors that Decrease
Certainty Certainty
Evidence Stream
Judgment
Evidence Integration
Summary Judgment
and observed no changes
in IgG 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
• High and medium
confidence studies
• Coherence of
findings across
circulating immune
cells, splenic
cellularity, and
thymic cellularity
and with
histopathological
changes
• Inconsistent direction
of effects across
studies and sex
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.
3-131
-------
APRIL 2024
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
monkeys. 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
1 High confidence study
5 Medium confidence
studies
In 1 high confidence
short-term study, a dose-
dependent increase in
both extramedullary
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),
thymus (2), or lymph
nodes (2).
• High and medium
confidence studies
• Dose-response
relationship
observed
• Coherent changes
with those observed
in circulating
immune cells,
splenic cellularity,
and thymic
cellularity
• Inconsistent direction
of effects across
studies
Organ weights
2 High confidence
studies
5 Medium confidence
studies
Mixed results were
reported for absolute and
relative spleen (7) and
thymus (5) weights. Both
studies in male and
female rats reported
»High and medium
confidence studies
• Inconsistent direction
of effects across
species and sex
• Confounding variables
such as decreases in
body weights
3-132
-------
APRIL 2024
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 absolute • Lack of dose-response
spleen (2/2) (males only) relationship
and thymus 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).
Two short-term studies • High and medium • Limited number of
found decreased globulin confidence studies studies examining
levels (2/3) in male rats specific outcomes
and no changes in female • Inconsistent direction
rats. One short-term study 0f effects across sex
found increases in subsets
of immunoglobulins (1/1)
in both male and female
Globulins and
immunoglobulins
1 High confidence
studies
4 Medium confidence
studies
3-133
-------
APRIL 2024
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
rats, and one short-term
study found no changes
in IgE (1/1) in male mice
Mechanistic Evidence and Supplemental Information (Section 3.4.2.3)
Biological events or
pathways
Summary of Key Findings, Interpretation, and Limitations
Evidence Stream
Judgment
Immune system
development and
physiology
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.
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.
Effects on adaptive
immune responses
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.
Limitations:
• Association between gene expression changes and apical endpoints need
further confirmation.
Autoimmune diseases
Key findings and interpretation:
• PFOS-mediated effects on pro-inflammatory T-helper cells, specifically
increased IL-17 and IL-22 production, in mice.
Limitations:
3-134
-------
APRIL 2024
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
• Only a single study; no studies directly evaluated the mechanism by which
PFOS could promote autoimmunity.
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.
Notes: HFMD = hand, foot, and mouth disease; 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.
3-135
-------
APRIL 2024
a Studies 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).
3-136
-------
APRIL 2024
3.4.3 Cardiovascular
EPA identified 106 epidemiological and 13 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, 2 and were
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
3.4.3.1.1 Cardiovascular 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
buildup 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. Cardiovascular outcomes were synthesized
separately by population (i.e., adults, children, occupational populations), and definitions of
certain conditions may vary by age. For example, high blood pressure and/or hypertension is
generally defined as SBP >140 mmHg and DBP > 90 mmHg in adults and SBP >130 mmHg
and DBP > 80 mmHg in children and adolescents, although consistent blood pressure
measurements in youth can be challenging {Falkner, 2023, 11279612}.
The 2016 PFOS HESD {U.S. EPA, 2016, 3603365} 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 3-29. Results from
studies summarized in the 2016 PFOS HESD are described in Table 3-13 and below.
The developmental section in the 2016 PFOS HESD describes results from Geiger et al. {, 2014,
2851286} which reported no association with hypertension in 1,655 children aged 12-18 years
from theNHANES (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
3-137
-------
APRIL 2024
was considered low confidence due to concerns about healthy work effect and potential residual
confounding by smoking status and race/ethnicity.
i i i i ¦ ¦
<0^
Alexander et al., 2003, 1291101 -
-
+
+
-
+
+
-
-
Geiger et al., 2014, 2851286 -
+
++
++
+
+
+
+
+
Figure 3-29. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Cardiovascular Effects Published Before 2016 (References in the 2016
PFOS HESD)
Interactive figure and additional study details available on HAWC.
Table 3-13. Associations Between Elevated Exposure to PFOS and Cardiovascular
Outcomes From Studies Identified in the 2016 PFOS HESD
Reference,
confidence
Study
Design
Population
. , Heart Disease
Hypertension13 MortaHtyb
Cerebrovascular
Disease
Mortality13
Alexander, 2003,
1291101
Low
Cohort
Occupational
NA |
4
Geiger, 2014,
2851286
Medium
Cross-
sectional
Children
NA
NA
Notes: NA = no analysis was for this outcome was performed; \ = nonsignificant positive association; ft = significant positive
association; J, = nonsignificant inverse association; J. J. = significant inverse association; - = no (null) association.
a Arrows indicate the direction in the change of the mean response of the outcome (e.g., J. indicates decreased mean birth weight).
b Arrows indicate the change in risk of the outcome (e.g., \ indicates an increased risk of the outcome).
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
3-138
-------
APRIL 2024
hypertension is discussed in the synthesis on female reproductive effects (see Appendix D, {U.S.
EPA, 2024, 11414344}). 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, 6988476; Pitter, 2020, 6988479; Yu, 2021,
8453076; Ye, 2021, 6988486}. The three controlled trial studies {Cardenas, 2019, 5381549; Liu,
2018, 4238396; Osorio-Yanez, 2021, 7542684} 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 available 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
3-139
-------
APRIL 2024
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 3-30 and Figure 3-31.
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,
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.
3-140
-------
APRIL 2024
Averina et al., 2021, 7410155-
+
+
*
+
+
+
+
+
Bao et al., 2017, 3860099-
+
+
+
+
++
+
+
+
Canova etal.,2021, 10176518-
+
+
+
++
+
+
+
+
Cardenas et al., 2019, 5381549-
+
++
+
+
+
+
+
+
Chen et al., 2019, 5387400-
-
+
++
+
+
+
+
+*
Christensen et al.; 2016, 3858533 -
-
+
-
~
+
-
Christerisen et al., 2019, 5080398-
+
+
+
+
+
+
+
+
Donat-Vargas et al., 2019, 5080588 -
+
++
+
++
+
+
+
Fry et al., 2017,4181820-
++ ++
+
+
+
+
+
+
Graber et al., 2019, 5080653-
-
+
-
+
+
-
Honda-Kohmo et al., 2019, 5080551 -
rr
+
+ +
+
Huang et al., 2018, 5024212-
++ ++
+
+
+
+
+
+
Hutcheson et al., 2020, 6320195-
~
-
+
+
+
Jain et al., 2020, 6988488-
+
+
+
+
+
+
+
+
Jain, 2020, 6311650-
4
+
+
+
+
+
+
~
Jain, 2020, 6833623-
~
+
+
+
+
+
Khalilet al., 2018, 4238547-
-
+
+
+
+
Khalilet al., 2020, 7021479-
-
+
+
+
+
-
Koshy et al., 2017, 4238478 -
+
+
+
+
Koskela et al., 2022, 10176386-
-
+
++
+
+
+
-
Leary et al., 2020, 7240043 -
L_
+
+
+
+
~
Li et al., 2021, 7404102-
*
++
+
++
+
Liao et al., 2020, 6356903-
++
++
++
+
++
+
+
++
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 3-30. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Cardiovascular Effects
Interactive figure and additional study details available on IiAWC.
3-141
-------
APRIL 2024
>°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 3-31. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Cardiovascular Effects (Continued)
Interactive figure and additional study details available on IiAWC.
3-142
-------
APRIL 2024
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 Appendix D, {U.S. EPA, 2024,
11414344}). Among 2,251 NHANES (2003-2012) adolescents (mean age 15.5 years) Ma et al.
{, 2019, 5413104} observed a positive association with DBP, which was 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 with 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 postnatal 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 in children or adolescents {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
3-143
-------
APRIL 2024
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 Appendix D, {U.S. EPA, 2024, 11414344}). 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 nonlinear (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 postpartum 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 nonsignificant 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
3-144
-------
APRIL 2024
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
toward 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 with
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 with 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 nonsignificant among males and females in stratified analysis. No
association was observed for elevated blood pressure in two medium confidence studies
{Christensen, 2019, 5080398; Liu, 2018, 4238514} and for hypertension in one 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}).
3-145
-------
APRIL 2024
Yu 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 et al. {, 2020,
6988486} reported a nonsignificant 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 et al. {, 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 nonsignificant 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 nonsignificant 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 with 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 et al. {, 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 et al. {, 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 not associated with markers of carotid artery atherosclerosis, including
3-146
-------
APRIL 2024
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 with a mild score (< 11) was
increased with respect to PFOS exposure, and the odds of a severe score (> 400) compared with
a mild score were significantly increased. Koskela et al. {, 2022, 10176376}, a low confidence
study, examined abdominal aortic calcification among participants aged 40 years and older in
NHANES (2013-2014) and 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 postpartum 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.
3.4.3.1.2 Serum 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. Evidence for changes in serum lipids was synthesized
by population (i.e., children, pregnant women, adults, occupational populations), and there may
be differences in the interpretation of an effect depending on age. For example, while elevated
levels of TC, LDL, and triglycerides are associated with increased cardiovascular risks in adults,
serum lipid changes in children are age-dependent and fluctuate during puberty {Daniels, 2008,
6815477}.
There are 15 studies (17 publications)13 from the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}
that investigated the association between PFOS and serum lipid effects. Study quality
13 Olsen {, 2003, 1290020} is the peer-review paper of Olsen {, 2001, 10228462} and Olsen {, 2001, 10240629}.
3-147
-------
APRIL 2024
evaluations for these 15 studies are shown in Figure 3-32. Results from studies summarized in
the 2016 PFOS HESD are described in Table 3-14 and below.
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,
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 separate cohorts (C8 Health Project
and Canadian Health Measures Survey) {Steenland, 2009, 1291109; Fisher, 2013, 2919156}.
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.
3-148
-------
APRIL 2024
<\®
,o®
Chateau-Degat etal., 2010, 2919285-
+
I
+
I
+
'
+
I
I
+
I
+
+
Eriksen etal., 2013, 2919150-
+
+
+
+
++
+
+
+
Fisher etal., 2013, 2919156-
0
+
+
+
+ +
+
+
+
Fitz-Simon et al., 2013, 2850962-
+
++
+ *
+
++
+
+
+ *
Frisbee et al., 2010, 1430763 -
+
+
+ *
+
+
+
+
+ *
Fu etal,, 2014, 3749193-
+
+ *
+
+
+
Geiger et al., 2014, 2850925 -
+
++ ++
+
+
+
+
+
Jain, 2014, 2969807-
+
+
+
"
+
+
~
Lin etal., 2009, 1290820-
+
+
+
+
+
+
+
+
Maisonet et al., 2015, 3981585-
+
+
+ *
+
+
+
+
+*
Nelson etal., 2010, 1291110 -
++
+
+
+
++
+
+
+
Olsen etal., 2001, 10228462-
+
-*
+
+
+
+
-*
Olsen etal., 2003, 1290020-
+
"
+
+
+
+
-
Starling etal., 2014, 2850928-
+
+
+ *
+
++
+
+
+*
Steeriland etal.. 2009, 1291109-
+
+
+ *
+
+
+
+
+*
Timmermann etal., 2014, 2850370-
+
+
+
+
+
+
+
+
Legend
B
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 Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Serum Lipids Published Before 2016 (References in the 2016 PFOS
HESD)
Interactive figure and additional study details available on HAWC.
3-149
-------
APRIL 2024
Table 3-14. Associations Between Elevated Exposure to PFOS and Serum Lipids From Studies Identified in the 2016 PFOS
HESD
Reference, confidence
Study Design
Population
TCa
HDLa
LDLa
TG
Chateau-Degat, 2010,
2919285
Medium
Cross-sectional
Adults
t
tt
1
1
Eriksen, 2013,2919150
Medium
Cross-sectional
Adults
tt
NA
NA
NA
Fisher, 2013,2919156
Medium
Cross-sectional
Adults
-
-
-
-
Fitz-Simon, 2013, 2850962
Mixed b
Cohort
Adults
t
4
t
-
Frisbee, 2010, 1430763
Mixed b
Cross-sectional
Children
tt
-
tt
t
Fu, 2014, 3749193
Low
Cross-sectional
Adults and children
t
4
t
t
Geiger, 2014, 2850925
Medium
Cross-sectional
Adolescents
tt
-
tt
1
Lin, 2009, 1290820
Medium
Cross-sectional
Adults
NA
tt
NA
-
Maisonet, 2015, 3981585
Mixed b
Cohort
Children
-
-
-
1
Nelson, 2010, 1291110
Medium
Cross-sectional
Adults
tt
t
t
NA
Olsen, 2001, 10228462
Mixed b
Cohort
Adults
t
1
NA
t
Olsen, 2003, 1290020
Low
Cohort
Occupational
-
NA
NA
-
3-150
-------
APRIL 2024
Reference, confidence
Study Design
Population
TCa
HDLa
LDLa
TG
Starling, 2014, 2850928
Mixedh
Cohort
Children
tt
tt
t
-
Steenland, 2009, 1291109
Mixedh
Cross-sectional
Occupational
t
t
t
t
Timmerman, 2014, 2850370
Medium
Cohort
Children
NA
NA
NA
t
Notes'. HDL = high-density lipoprotein cholesterol; LDL = low-density lipoprotein; NA = no analysis was for this outcome was performed; TC = total cholesterol;
TG = triglycerides; | = nonsignificant positive association; ft = significant positive association; j = nonsignificant inverse association; jj = significant inverse association;
- = no (null) association.
Jain et al., 2014,2969807 was not included in the table due to their uninformative overall study confidence ratings.
a Arrows indicate the direction in the change of the mean response of the outcome (e.g., j indicates decreased mean birth weight).
b Mixed confidence studies were rated medium confidence for TC and HDL and low confidence for LDL and TG due to non-fasted blood samples.
3-151
-------
APRIL 2024
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 10 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
methods in Appendix A {U.S. EPA, 2024, 11414344} and Section 2.1.3. 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 = 3.4 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 3-33, Figure 3-34, and Figure 3-35.
Consistent with 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 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,
14 Dong et al. {, 2019, 5080195} counted as two studies, one in adolescents and one in adults.
3-152
-------
APRIL 2024
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 den Dungen, 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 only).
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 den Dungen, 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
confidence studies were the focus, although low confidence studies were still considered for
consistency in the direction of association.
3-153
-------
APRIL 2024
% 0^CO
-------
APRIL 2024
,<^eXe,^e#
^!>46*°
Jensen et al., 2020, 6833719 -
+
+ I
..
i
+
'
+
i
+
-
+*
Kang et al., 2018, 4937567 -
+
+
+
+
+
+
+
+
Khalil et al., 2018, 4238547-
-
+
+
-
+
+
-
Khalil et al., 2020, 7021479-
-
+
+
-
+
+
-
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 -
-
+
+
-
+
+
:s
Li et al., 2020, 6315681 -
¦
+
+*
-
+
+
Li et al., 2021, 7404102-
+
+
+
+
Linet al., 2013, 2850967-
-
+
+
-
+
+
+
Linet al., 2019, 5187597-
++
+
++
+
++
+
-
+
Linet al., 2020, 6315756-
-
+
+
+
+
+
+
+
Linet al., 2020, 6988476-
-
+
-
+
+
+
•f
+
Liuet al., 2018, 4238396-
-
+
++
+
+
+
+
Liuet al., 2018, 4238514-
+
+
+
+
+
+
+
+
Liuet al., 2020, 6318644-
+
+
+
+
++
+
+
+
Liu et al., 2021, 10176563-
-
-*
++
+
-
+
Manzano-Salgado et al., 2017, 4238509 -
-
++
+*
+
+
++
+*
Matilla-Santander et al., 2017, 4238432 -
+
+
+*
+
+
+*
Mora et al., 2018, 4239224-
+
+
+
+
+
+
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 3-34. Summary of Study Evaluation for Epidemiology Studies of PFOS and Serum
Lipids (Continued)
Interactive figure and additional study details available on HAWC.
3-155
-------
APRIL 2024
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
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 3-35. Summary of Study Evaluation for Epidemiology Studies of PFOS and Serum
Lipids (Continued)
Interactive figure and additional study details available on IiAWC.
3-156
-------
APRIL 2024
3.4.3.1.2.3 Findings From Children
Results for the studies that examined TC in children are presented in Appendix D {U.S. EPA,
2024, 11414344}. 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 10
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}. 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 with adolescents. Significant positive associations were observed in 9-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}. 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 Appendix D, {U.S. EPA, 2024, 11414344}) {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;
3-157
-------
APRIL 2024
Zeng, 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, 11 medium confidence, and 3 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}. 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 et al. {, 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
3-158
-------
APRIL 2024
{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 upward. 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
One high confidence study {Gardener, 2021, 7021199} and four medium confidence studies
examined the association between PFOS and TC in pregnant women and four reported positive
associations between PFOS and TC (see Appendix D, {U.S. EPA, 2024, 11414344}) {Matilla-
Santander, 2017, 4238432; Skuladottir, 2015, 3749113; DallaZuanna, 2021, 7277682}. A
significant positive trend across quartiles of PFOS exposure was observed for TC in a cohort
study of pregnant women from the United States {Gardener, 2021, 7021199}. Skuladottir et al.
{, 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 {DallaZuanna, 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 nonsignificant, association between PFOS and HDL (see Appendix D, {U.S.
EPA, 2024, 11414344}) {Starling, 2017, 3858473}. No association between PFOS and HDL
was observed in a Chinese study of pregnant women {Yang, 2020, 7021246}.
One high confidence, one medium confidence and three low confidence studies examined the
association between PFOS and triglycerides in pregnant women. A significant positive trend
across quartiles of PFOS exposure was observed for triglycerides in a cohort study of pregnant
3-159
-------
APRIL 2024
women from the United States {Gardener, 2021, 7021199}. The medium confidence study
reported no association between PFOS and triglycerides (see Appendix D, {U.S. EPA, 2024,
11414344}) {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
nonsignificant inverse association. All low 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. The inverse associations observed in the low confidence studies
provides support for an inverse association between PFOS and triglycerides. These inverse
associations are inconsistent with the finding in the only medium confidence study. In summary,
the available evidence suggests an inverse association between PFOS and triglycerides in
pregnant women. However, additional high or medium confidence evidence is needed to confirm
this association.
Kishi et al. {, 2015, 2850268} additionally examined the association between PFOS and select
fatty acids in serum. Except for stearic acid and eicosapentaenoic acid, 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 12 low confidence studies examined PFOS and TC or
hypercholesterolemia in adults. All studies examined the cross-sectional association {Dong,
2019, 5080195; 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 10 medium confidence studies, nine reported positive associations (Figure 3-36,
Figure 3-37, Figure 3-38, and Figure 3-39). 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 TC. Canova et al. {, 2020, 7021512}
also reported a concentration-response curve for risk of high TC when PFOS was categorized in
quartiles or deciles, with a higher slope at higher PFOS concentrations (Figure 3-40). 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
3-160
-------
APRIL 2024
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; Jain, 2019, 5080642; Liu,
2018, 4238514; Fan, 2020, 7102734} (see Appendix D, {U.S. EPA, 2024, 11414344}). The
association was statistically significant when data from all cycles were pooled in analyses
{Dong, 2019, 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 et al. {, 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 et al. {, 2019, 5080588} were small and few were
statistically significant. It is noteworthy that all participants in Lin et al. {, 2019, 5187597} were
prediabetic, approximately half of all participants in Han et al. {, 2021, 7762348} were diabetic,
all participants in Liu et al. {, 2020, 6318644} were obese and enrolled in a weight loss trial, and
all participants in Donat-Vargas et al. {, 2019, 5080588} were free of diabetes for at least
10 years of follow-up. It is unclear whether 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 10 of 12 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
other studies {Bjorke-Monsen, 2020, 7643487; Cong, 2021, 8442223; Liu, 2021, 10176563}.
Further, He et al. {, 2018, 4238388} used similar data as the four medium NHANES studies and
thus added little unique information. Considering medium and low confidence studies together,
small increases in TC with increased PFOS were observed, though less consistently.
3-161
-------
APRIL 2024
Reference, FxDosure
Confidence , - Study Design Exposure Levels Sub-population Comparison EE
r»_»: Matrix
Rating
Effect Estimate
-5 0 5 10 15 20 25 30
Canova et al. serum (Soss-sectional median=3.7 ng/mL — regression
(2020, (25th-75th percentile: coefficient [per
7021512), 2.5-5.6 ng/mL) Hn(PFOS) mg/mL 4.99
Medium increase PFOS]
-
females regression
coefficient [per
l-ln(PFOS) mg/mL 4.07
increase PFOS]
—
regression
coefficient [for 02
vs. 01 PFOS] 2.86
—
regression
coefficient [for 03
vs. 01 PFOS] 4.18
regression
coefficient [for 04
vs. 01 PFOS] 6.32
males regression
coefficient [per
Hn(PFOS) mg/mL 5.48
increase PFOS]
—
regression
coefficient [for 02
vs. 01 PFOS] 2.16
regression
coefficient [for 03
vs. 01 PFOS] 4.8
regression
coefficient [for 04
vs. 01 PFOS] 8.09
Chateau- plasma Cross-sectional Geometric mean: 18.6 - regression
Degat et al. ( ug/L (95% confidence coefficient (per
2010, interval: 17.8-19.5) 1-ug/L increase in 0
2919285), PFOS)
Medium
-5 0 5 10 15 20 25 30
Figure 3-36. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS
Interactive figure and additional study details available on HAWC.
3-162
-------
APRIL 2024
Reference, FxDosure
Confidence , - Study Design Exposure Levels Sub-population Comparison EE
r»_»: Matrix
Rating
Effect Estimate
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6
Donat-Vargas plasma Cohort Baseline median: 20 - prospective regression
etal. (2019, (ng/ml) (25th-75th coefficient (per nni-
5080588), percentile: 15-26 ng7ml) 1-SD7.42 ng/mL U UD
Medium Follow-up median: 15 PFOS)
1
i •
i
i
(ng/ml) (25th-75th prospective regression
percentile: 9.7-21 ng/ml) coefficient (mean Q Q1
change) for tertile 2 vs
fertile 1 PFOS
i
i
i
prospective regression
coefficient (mean _ _
change) for tertile 3 vs
tertile 1 PFOS
i
1 •
i
i
regression coefficient
(per 1-SD 8.62 ng/mL n nQ
PFOS) "uus
i
•—h
i
i
regression coefficient
(mean change) for n ,.
tertile 2 vs tertile 1
PFOS
i
• i
i
regression coefficient
(mean change) for n „
tertile 3 vs tertile 1
PFOS
i
• i
i
median: 20 (ng/ml) baseline regression coefficient
(25th-75th percentile: (per 1-SD 8.42 ng/mL n _.
15-26 ng/ml) PFOS) ~u^
i
• i
i
regression coefficient
(mean change) for n „
tertile 2 vs tertile 1
PFOS
i
• i
i
regression coefficient
(mean change) for n _
tertile 3 vs tertile 1
PFOS
i
• i
i
i
median: 15 (ng/ml) follow-up regression coefficient
(25th-75th percentile: (per 1-SD 7.94 ng/mL n ni
9.7-21 ng/ml) PFOS) 1
i
i
i
regression coefficient
(mean change) for n _.
tertile 2 vs tertile 1
PFOS
i
• i
i
i
regression coefficient
(mean change) for n ni
tertile 3 vs tertile 1 u u 1
PFOS
i
i
Dong etal. serum Cross-sectional Median: 10.9 ng/ml; Adults (20-80 Regression Coefficient
(2019, Mean ± SD: 15.6 ± 17.8 years) (change in TC per unit q*
5080195), ng/ml increase in serum
Medium PFOS)
i
i »
i
Figure 3-37. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS (Continued)
Interactive figure and additional study details available on HAWC.
3-163
-------
APRIL 2024
Referent*
Confidence
Rating
En •> n i-t ?! p rmc
50),
F=m f>t
(2020,
- P-14I
Exposure
Matnx
Study Design Exposure Levels
CrDjh-jfttio^al mesn 26 i no-'ml
Gross-sectional median=5.14 ug/L
<25tft-75th percentile:
2.80-3.31 ug/L)
Si b pi^pulatin Com an oi
ii-fsrfcn e | f
l""erQt Hr |h rd It- if
PFOS
Cross-sectional geometric mean (SDi =
8.40 ug/L (2.04)
Irs
- FOS)
Jain et ai.
(2019,
5080642).
Case-control
^ |i v^nti p 44"
-10.55 ng/mL): Controls'
1101^^=0 ^ n-1 it 2^th
"rc * cen ei iIp 5 -*i> 1 u*
no/mLs
Cross-sectional Gi-oti»" t n-dp-" 4 rnl f cn ui t-se
25m - 75tfi percentiles: 4.4 - females
13 1 SD 2.5
cient (per
•mge
re"re:sfun
coeffment >ppr
-en SO untrhange
ir PFOS.
Geometric mean-11.5
r th
ppr^ntiP i: SD
Non-obese males aJju tp rsgre ni
coefficients (beta)
between ioglO
FF 'b ma L =iiid
iK « te' i mq dLi
regression
1 U rfS ant pp
' o'liun ihsnp
r PF> ->
Figure 3-38. Overall Levels of Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS (Continued)
Interactive figure and additional study details available on HAWC.
3-164
-------
APRIL 2024
Reference, FxDosure
Confidence , - Study Design Exposure Levels Sub-population Comparison EE
r»_»: Matrix
Rating
Effect Estimate
-5 0 5 10 15 20 25 30
Linetal. plasma Cohort and median=27.2 ng/ml participants who Regression coefficient
(2019, cross-sectional (25th-75th percentile: were not on (mean difference) (per
5187597), 18.0-40.4) lipid-lowering doubling of baseline 2.53
Medium medication plasma PFOS)
1
1
1 .
1
1
1
Regression coefficient
(mean difference) (for
Q2vsQ1) 1.13
1
1
1 .
1
1
Regression coefficient
(mean difference) (for
Q3vsQ1) 5.05
1
1
1 _
1
1
1
Regression coefficient
(mean difference) (for
Q4vsQ1) 5.13
1
1
1 ,
1
1
1
Linetal. serum Cross-sectional median=16.2 ng/mL Participants not Regression
(2020, (25th-75th percentile: taking lipid Coefficient (for Q2 vs
6988476), 10.1-24.1) lowering Q1 ) 15.06
Medium medication
i
1
1 a
1
1
Regression
Coefficient (for Q3 vs
Q1 ) 11.47
1
1
1 -
1
1
1
Regression
Coefficient (for Q4 vs
Q1 ) 10.18
1
1
1 _
1
1
Nelson etal. serum Cross-sectional median: 21.0 ug/L 20-to regression coefficient
(2010, (range: 1.4-392.0 ug/L) 80-year-olds (per ug/L increase in
1291110), PFOS) 0.27
Medium
1
1
j»
1
1
regression coefficient
[for Q4 (28.2-392.0
ug/L) vs. Q1 (1.4-13.6 13.4
ug/L) PFOS]
1
1
1 -
1
1
i
Olsen et al. serum Cohort Antwerp Mean (SD) = Males Regression coefficient
(2003, 0.96 ppm (0.97); (per unit increase in
1290020), Decatur=1.40 ppm (1.15) PFOS) 0.01
Medium
1
1
1
Steenland et serum Cross-sectional Median: 19.6 ng/mL - regression coefficient
al. (2009, (min-max: 0.25-759.2 (per 1-ln ng/mL
1291109), ng/mL) increase in PFOS) 0.03
Medium
|
1
1
-5 0 5 10 15 20 25 30
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 HAWC.
3-165
-------
APRIL 2024
Reference, pXD0Sure
Confidence ,. Study Design Exposure Levels Sub-population Comparison EE
n„»: Matrix
Rating
Effect Estimate
0.9 1 1.1 1.2 1.35 1.5 1.65 1.8 2
CanovaetaT serum Cross-sectional median=3.7 ng/mL - OR[forQ2 vs. Q1)
(2020, (25th-75th percentile:
7021512), 2.5-5.6 ng/mL) 1 1Q
Medium
OR [for Q3 vs. Q1]
1.37
OR [for Q4 vs. Q1)
1.58
females OR [for 02 vs. Q1]
1.21
OR [for Q3 vs. Q1]
1.31
OR [for Q4 vs. Q1]
1.44
males OR [for Q2 vs. Q1]
1.12
OR [for Q3 vs. Q1]
1.4
OR [for Q4 vs. Q1]
1.62
0.9 1 1.1 1.2 1.35 1.5 1.65 1.8 2
Figure 3-40. Odds of High Total Cholesterol in Adults from Epidemiology Studies
Following Exposure to PFOS
Interactive figure and additional study details available on HAWC.
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; Jain, 2019, 5080642; Liu,
2018, 4238514}, but the association was statistically significant in obese women only {Jain,
2019, 5080642} (see Appendix D, {U.S. EPA, 2024, 11414344}). 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 55-75-year-olds. This analysis combined cases of type 2 diabetes and
healthy controls, and it is unclear whether 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 et al. {, 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-III-containing IDL and LDL are strongly associated with increased cardiovascular risks.
3-166
-------
APRIL 2024
Thus, the positive associations with cholesterol in ApoC-III-containing fractions of IDL and
LDL were coherent with the 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, 9 of the 10 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
exposure and health problems {Li, 2020, 6315681} were major concerns in these studies. In
addition, He et al. {, 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 13 low confidence studies examined PFOS and HDL or clinically
defined low HDL in adults. All studies examined the cross-sectional association {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} including Dong et al. {, 2019, 5080195} in the adult portion
of the study. 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 Appendix D, {U.S. EPA, 2024, 11414344}). 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 5 of 13 studies {Lin, 2020, 6315756; Li, 2020, 6315681; He,
2018, 4238388; Yang, 2018, 4238462; Liu, 2018, 4238396} (association with concurrent HDL).
Of note, in Lin et al. {, 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.
3-167
-------
APRIL 2024
Nine medium confidence and 13 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 Appendix D, {U.S. EPA,
2024, 11414344}). 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 et al. {, 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 qualitatively 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 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,
3-168
-------
APRIL 2024
3859842}; one investigated the association between baseline PFOS and changes in TC over the
course of a fluorochemical plant demolition project {Olsen, 2012, 2919185}. PFOS was
positively associated with TC in Rotander et al. {, 2015, 3859842}, but the association was not
statistically significant. The other cross-sectional study simply reported no significant association
{Wang, 2012, 2919184}. Olsen et al. {, 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. On the basis of 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 et al. {, 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 et al. {, 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 et al. {, 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 et al. {, 2015, 3859842}, but this
association was not statistically significant. Wang et al. {, 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, the available studies examining associations between PFOS serum concentrations
and serum lipids among workers was limited to 3 low confidence studies. 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 9 studies from
recent systematic literature search and review efforts conducted after publication of the 2016
3-169
-------
APRIL 2024
PFOS HESD that investigated the association between PFOS and cardiovascular effects. Study
quality evaluations for these 13 studies are shown in Figure 3-41.
^eV°
_l I I
Conley et al„ 2022, 10176381 -
i
+
i
+
I
NR
++
1
+
i
+
DD
++
+
Curran et al.. 2008, 757871 -
++
NR
NR
++
+
+
++
++
++
+
Dangudubiyyam et al., 2022, 10429383 -
+
++
++
+
-
+
+*
Lai et al., 2018, 5080641 -
+
+
NR
++
-
+
+
+
+
+
Li et al., 2021, 7643501 -
+
+
NR
++
-
+
+
-
+
Luebker et al., 2005, 757857 -
+
+
NR
+
-
-
+
+
+
NTP, 2019, 5400978-
++
++
++
++
Seacat et al. 2002, 757853 -
++
++
++
++
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
++
D
++
++
++
++
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 3-41. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFOS Exposure and Cardiovascular Effects
Interactive figure and additional study details available on HAWC.
3-170
-------
APRIL 2024
Cardiovascular effects, including blood pressure, heart weight, heart histopathology, and/or
serum lipid levels, following exposure to PFOS were minimal {Curran, 2008, 757871;
Dangudubiyyam, 2022, 10429383; NTP, 2019, 5400978; Rogers, 2014, 2149155; Li, 2021,
7643501; Xia et al., 2011, 2919267}. In male and female mice (sexes combined), relative heart
weight was increased at PND 21 after gestational exposure (GD 2-21) 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 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 {NTP, 2019, 5400978}. 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. 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}. In female
Sprague-Dawley rats exposed to 50 [j,g/mL PFOS in drinking water from GD 4-20, H&E and
Trichrome-Masson staining of the heart revealed a significant increase in ventricular wall
thickness as well as a slight increase in the percentage of fibrotic area (approximately 1% in the
control animals and 2% in the exposed animals) {Dangudubiyyam, 2022, 10429383}.
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. However, in
female Sprague-Dawley rats exposed to 0.5-50 |ig/mL PFOS in drinking water from GD 4-20,
blood pressure was significantly increased at GD 20 {Dangudubiyyam, 2022, 10429383}. Adult
Sprague-Dawley offspring of dams treated with PFOS (18.75 mg/kg/day) via oral gavage from
GD 2-6 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 with 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, 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
3-171
-------
APRIL 2024
primates and rodent models in subchronic, chronic, and developmental studies of oral exposure
to PFOS (Figure 3-42). Decreased serum TC, triglycerides, and/or HDL 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 twofold increase in serum LDL levels in male
BALB/c mice following exposure to 5 mg/kg/day PFOS by gavage for 28 days {Yan, 2014,
2850901}.
Endpolnt Study Name Study Design Observation Time Animal Description
PFOS Cardiovascular Effects - Serum Lipids
A No signilicanl change A Significant increase ~ Si gnifi can I decrease |
High Density Lipoprotein (HDL) Seacat et al., 2Q02,757853 chronic (26wk) 182d Monkey. Cynomolgus (r.C, IM=4-6)
Monkey. Cynomolgus (T, N=4-6)
Yan el al., 2014,2850901 short-term (28d) 28d Mouse, BALB/c (-*. N=6)
¦ i
1
Luebker el al., 2005, 757857 reproductive <76d (42d pre-cohabitalion, 14d maling. GDO-20J) GD21 P0 Rat. Crl:Cd(Sd)lgs Vaf/Plus (£. N=8)
F1 Rat, Cr1:Cd(Sd)lgs Vaf/Plus ( Jl. N=8)
reproductive <42d prior mating-LD4) LD5 P0 Rat. Crl:Cd{Sd)lgs Vaf/Plus N=17)
F1 Rat, Cr1:Cd(Sd)lgs Vaf/Plus . N=17)
Low Density Lipoprotein (LDL) Yan etal., 2014,2850901 short-term (28d) 28d Mouse, BALB/c (_', N=6)
Luebker et al., 2005, 757B57 reproductive (76d (42d pre-cohabltatlon, 14d mating, GDo-20)) GD21 P0 Rat, Crl:Cd(Sd)lgs Vaf/Plus (¦"-, N=8)
F1 Rat, Cil:Cd(Sd)lgs Vaf/Plus (;?•-, N=8)
A
^
reproductive <42d prior mating-LD4) LD5 P0 Rat. Crl:Cd'day)
0
Figure 3-42. 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,
3-172
-------
APRIL 2024
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 with 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.
3.4.3.3 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 nine 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 organized by mechanistic data category (see Appendix A,
{U.S. EPA, 2024, 11414344}) and source is shown in Figure 3-43.
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 3-43. Summary of Mechanistic Studies of PFOS and Cardiovascular Effects
Interactive figure and additional study details available on HAWC.
3.4.3.3.1 Fatty 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
3-173
-------
APRIL 2024
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.
3.4.3.3.2 Serum 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
study summarized in EPA's 2016 Health Effects Support Document for Perfluorooctane
Sulfonate (PFOS){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.
3.4.3.3.3 Oxidative 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 buildup 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}.
3-174
-------
APRIL 2024
3.4.3.3.4 Endothelial Dysfunction
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
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 with 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 with 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 (PPARyj 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
3-175
-------
APRIL 2024
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 (xM) 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.
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 with typical human exposure levels in the general
population. PFAS with longer carbon chain lengths activated the KKS at a much lower
concentration compared with PFOS (e.g., PFHxDA activated the KKS at 30 [xM). 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 [xM) 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
3-176
-------
APRIL 2024
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}.
3.4.3.3.5 Coagulation and Fibrinolysis
The coagulation and fibrinolytic pathways can contribute to the progression of atherosclerosis.
Two studies from the literature published after the 2016 PFOS 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, 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 (J,g/mL, including measures of
prothrombin time, activated partial thromboplastin time, and fibrinogen.
3.4.3.4 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
TC and LDL. 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.
On the basis of this systematic review of 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 to
be 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 PFOS HESD. The positive associations with TC are
also supported by the recent meta-analysis restricted to general population studies in adults
{EPA, 2024, 11414059 J. 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 human epidemiological studies identified since the 2016 PFOS HESDs provided additional
clarity regarding the association between PFOS and CVD outcomes. Most of the CVD-related
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
3-177
-------
APRIL 2024
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 postpartum. 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 PFOS HESD also indicated association with blood pressure in children.
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.
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 of 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 (NCEH1 and PPARa) in samples from women, while
there were no associations in men. The results for PFOS-induced changes in serum lipid levels
contrast between rodents (generally decreased) and humans (generally increased). PFOS
exposure led to upregulation 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 evidenced 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
3-178
-------
APRIL 2024
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.
3.4.3.4.1 Evidence 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 3-15).
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. The consistent findings for serum lipids are also supported by evidence of
associations with blood pressure in adult populations in high and medium confidence studies.
3-179
-------
APRIL 2024
Table 3-15. Evidence Profile Table for PFOS Exposure and 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 Examination of serum
2 High confidence studies lipids included measures
28 Medium confidence of TC, LDL, HDL, TG,
studies and VLDL. In studies of
21 Low confidence studies serum lipids in adults from
12 Mixed* studies the general population
(33), there is evidence of
positive associations with
TC (13/15) in the medium
confidence studies.
Positive associations were
also observed for LDL
(9/11) medium confidence
studies. Results for HDL
and TG were mixed, with
some positive associations
for HDL (8/14) and some
inverse associations for
TG (8/13) in medium
confidence studies.
Evidence from studies of
children (21), reported
significant increases in TC
(7/16) and LDL (7/16),
though others observed no
association. While some
studies observed
significantly increased
HDL (3/17), others
reported significant
decreases or no
associations. Studies
examining pregnant
• High and medium
confidence studies
• Consistent findings of
positive associations
for LDL and TC across
study populations
• Coherence of findings
across serum lipids
• Low confidence
occupational studies
0©O
Moderate
Evidence for
cardiovascular effects is
based on numerous
medium confidence
studies reporting positive
associations with serum
lipids (LDL and TC) 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.
However, interpretations
of changes in serum lipids
for children are less clear.
High and medium
confidence studies
reported positive
associations with blood
pressure and increased
risk of hypertension. Low
confidence studies
reported nonsignificant
associations, while most
mixed confidence studies
reported significant
associations. Observed
0©O
Evidence Indicates
(likely)
Primary basis and cross-
stream coherence:
Human evidence indicated
consistent evidence of
serum lipids response and
animal evidence showed
coherent results for
perturbations in lipid
homeostasis in non-human
primates and rodent
models in developmental,
subchronic, and chronic
studies following exposure
to PFOS. The consistent
findings for serum lipids
are also supported by
evidence of associations
with blood pressure in
adult populations in high
and medium confidence
studies.
Human relevance and
other inferences:
No specific factors are
noted.
3-180
-------
APRIL 2024
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. Most
occupational studies (5)
were considered low
confidence (4/5), and no
association was observed
for TC or HDL-C in the
single medium confidence
occupational study.
effects were inconsistent
for CVD and imprecise for
atherosclerotic changes
across all study
populations.
Blood pressure and
hypertension
Studies examining
changes in blood pressure,
• High and medium
confidence studies
2 High confidence studies including DBP and SBP,
17 Medium confidence
studies
7 Low confidence studies
and risk of hypertension in
general population adults
showed consistent positive
associations with
increased risk of
hypertension (4/7),
positive associations for
SBP (7/9) and DBP (7/8),
including four medium or
high confidence studies
reporting significant
increases (4/6). Studies in
children (10) reported
mostly nonsignificant
associations with blood
pressure and/or
• Inconsistent findings in
children, likely due to
variation in measured
exposure windows
3-181
-------
APRIL 2024
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
hypertension, though one
study in adolescents
reported significantly
increased DBP (1/10) and
another reported decreased
(1/10) SBP. No studies
examined blood pressure
or hypertension in
occupational populations.
Cardiovascular disease
1 High confidence study
4 Medium confidence
studies
5 Low confidence studies
In adults from the general
population (8),
significantly decreased
odds of stroke (1/2) and
significantly increased
odds of MVD (1/1), heart
attack (1/1), and CVD in
the third exposure group
(1/4), were observed.
Other studies of stroke,
CHD, and CVD reported
nonsignificant
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 reported a
significant inverse
relationship between
employees in high-
• High and medium
confidence studies
• Limited number of
studies examining
specific outcomes
• Lnconsistent findings
for CVD-related
outcomes
• Lmprecision of
findings, particularly
for two studies with
self-reported outcome
measures
3-182
-------
APRIL 2024
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
exposure jobs and all heart
disease mortality (1/2).
Atherosclerotic changes
1 High confidence study
4 Medium confidence
studies
3 Low confidence studies
In studies of children and
young adults (3), two
studies observed
significant associations
with CIMT across
exposure groups (2/3),
among females, and
among those ages 12-19.
In studies of adults from
the general population (5),
two focused on adults
older than 70 years of age.
One study reported a
significant increase in left
ventricular end-diastolic
diameter and a significant
decrease in relative wall
thickness (1/2). One
medium confidence study
in prediabetic adults aged
over 25 also reported
significantly increased
odds of an Agatatson
Scores over 400, a
measure of arterial
calcification (1/1).
1 High and medium
confidence studies
• Imprecision of
findings across
children and adult
study populations
• Limited number of
studies examining
specific outcomes
Evidence from In Vivo Animal Toxicological Studies (Section 3.4.3.2)
Serum lipids
2 High confidence studies
6 Medium confidence
studies
Significant decreases in
serum TG were observed
in 5/7 studies that
examined this endpoint,
regardless of species, sex,
or study design. No
1 High and medium
confidence studies
' Consistency of
findings across
species, sex, or study
design
• Incoherence of
findings in other
cardiovascular
outcomes
0©O
Moderate
Evidence based on eight
high or medium
confidence studies
3-183
-------
APRIL 2024
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
changes were observed in
one monkey study and 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 was observed in
pups. Fewer studies
examined HDL and LDL,
with decreases in HDL
(2/3) and increases in LDL
(2/2) being observed.
• Dose-response
relationship observed
within multiple
studies
> Biological significance observed that PFOS
of the magnitude of
effect is unclear
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
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 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
Organ weight
1 High confidence study,
2 Medium confidence
studies
Mixed results were
reported for absolute and
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
• High and medium
confidence studies
• Limited number of
studies examining
outcome
the studies identified may
not be adequate in
exposure duration to
Confounding variables assess potential toxicity to
such as decreases in cardiovascular system,
body weights may
limit ability to interpret
these responses
3-184
-------
APRIL 2024
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
absolute heart weight and
an increase in relative
heart weight which was
confounded by decreases
in body weights.
Blood pressure and
heart rate
3 Medium confidence
studies
A developmental study
found increased blood
pressure in dams. A short-
term study found no effect
on blood pressure in male
and female rats. One
developmental study
found no effect on heart
rate.
• Medium confidence
studies
• Limited number of
studies examining
outcome
Mechanistic Evidence and Supplemental Information (Section 3.4.3.3)
Summary of Key Findings, Interpretation, and Limitations
Evidence Stream
Judgment
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.
• PFOS can bind to human FXII in vitro, which is the initial zymogen of plasma KKS activation, a
regulator 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.
Findings support the
plausibility that PFOS
exposure can lead to
changes in the expression
of genes involved in
cholesterol regulation, as
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.
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.
3-185
-------
APRIL 2024
3.4.4 Developmental
EPA identified 96 epidemiological and 20 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, 4 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.
3.4.4.1.2 Study 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. Further discussion of considerations for potential
confounding by co-occurring PFAS can be found in Section 5.1.1.
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 catchup periods during childhood {Perng, 2016,
3-186
-------
APRIL 2024
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).
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 toward 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).
3-187
-------
APRIL 2024
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. While
they could impact precision and study sensitivity, they were not be considered a major concern
for risk of bias.
3.4.4.1.3 Summary of Evidence From the 2016 PFOS HESD
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 3-44.
Studies included those conducted both in the general population as well as in communities
known to have experienced relatively high PFAS exposure (e.g., the C8 population in West
Virginia and Ohio). Results from studies summarized in the 2016 PFOS HESD are described in
Table 3-16 and below.
3-188
-------
APRIL 2024
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
1 1 1 I I I 1 1 L
Andersen etal., 2010, 1429893-
+
+
+
+
+
+
+
+
Apelberg et al., 2007, 1290833 -
+
+
+
++
+
+
+
Apelberg etal., 2007, 1290900-
+
+
+
+
+
+
n
Chen etal., 2012, 1332466-
+
+
+
+
+
+
Darrow et al., 2013, 2850966 -
++
+
+
++
+
+
++
Darrow et al., 2014, 2850274 -
+
+
-
+
+
+
+
+
Fei et al., 2007, 1005775-
+
++
++
+
++
+
+
+
Fei et al., 2008, 1290822-
+
++
H
+
+
+
+
+
Fei et al., 2008, 2349574-
+
+
+
+
+
+
+
+
Fei etal., 2010, 1430760-
+
++
-
+
+
+
+
-
Grice etal., 2007, 4930271 -
+
+
-
-
+
+
+
-
Hamm et al., 2010, 1290814 -
+
+
+
+
+
+
+
Maisonet et al., 2012, 1332465 -
+
•
-
+
+
+
+
Monroy et al., 2008, 2349575 -
-
¦
-
+
Olsen etal., 2004, 5081321 -
-
"
+
-
+
+
-
-
Stein etal., 2009, 1290816-
+
+
-
+
+
+
+
+
Washinoetal., 2009, 1291133-
+
+
+
+
+
+
+
+
Whitworth et al., 2012, 2349577 -
++
++
id
+
++
-
++
++
>oe
Figure 3-44. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Developmental Effects Published before 2016 (References from 2016
PFOS HESD)
Interactive figure and additional study details available on HAWC.
3-189
-------
APRIL 2024
As noted in the 2016 PFOS 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 (j,g/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 (j,g/mL) was associated with reduced 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, 1291133} 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 PFOS 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, 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 |ig/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, 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 2016 PFOS 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).
3-190
-------
APRIL 2024
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.
3-191
-------
APRIL 2024
Table 3-16. Associations Between Elevated Exposure to PFOS and Developmental Outcomes in Children From Studies
Identified in the 2016 PFOS HESD
Reference, confidence
Study
Design
Birth
Weight3
LBWb
SGAb
Gestational
Duration3
Preterm
Birthb
Birth
Defectsb
Pregnancy Lossb
PNG3
Andersen, 2010, 1429893°
Medium
Cohort
4
NA
NA
NA
NA
NA
NA
II
Apelberg, 2007, 1290833
Medium
Cross-
sectional
4
NA
NA
t
NA
NA
NA
NA
Chen, 2012, 1332466°
Medium
Cohort
II
t
tt
II
tt
NA
NA
NA
Darrow, 2014, 2850274
Medium
Cohort
NA
NA
NA
NA
NA
NA
t
NA
Darrow, 2013,2850966
High
Cohort
1
t
NA
NA
—
NA
NA
NA
Fei, 2007, 1005775d
Medium
Cohort
1
t
—
NA
t
NA
NA
NA
Grice, 2007, 4930271°
Low
Cohort
—
NA
NA
NA
NA
NA
—
NA
Hamm, 2010, 1290814
Medium
Cohort
—
NA
—
—
t
NA
NA
NA
Maisonet, 2012, 1332465
Medium
Cohort
II
NA
NA
—
NA
NA
NA
t
Olsen, 2004, 5081321
Low
Cross-
sectional
NA
NA
NA
NA
t
—
NA
NA
Stein, 2009, 1290816
Medium
Cohort
NA
tt
NA
NA
t
t
—
NA
Washino, 2009, 1291133f
Medium
Cohort
II
NA
NA
NA
NA
NA
NA
NA
3-192
-------
APRIL 2024
Reference, confidence
Study
Design
Birth
Weight3
LBWb
SGAb
Gestational
Duration3
Preterm
Birthb
Birth
Defectsb
Pregnancy Lossb
PNG3
Whitworth, 2012, 2349577
High
Cohort
4
NA
t
NA
4
NA
NA
NA
Notes'. LBW = low birth weight; NA = no analysis was for this outcome was performed; PNG = postnatal growth; SGA = small-for-gestational age;| = nonsignificant positive
association; ft = significant positive association; j = nonsignificant inverse association; jj = significant inverse association; - = no (null) association.
Apelberg et al. {, 2007, 1290900} and Monroy et al. {, 2008, 2349575} were not included in the table due to their uninformative overall study confidence ratings. Fei et al. {, 2008,
1290822}, Fei et al. {, 2008, 2349574}, and Fei et al. {, 2010, 1430760} were not included in the table because the studies only analyzed other developmental outcomes that were
more prone to measurement error (see Study Evaluation Considerations in Section 3.4.4.1.2) or were not as heavily studied (i.e., other measures of fetal growth restriction such as
birth length and head circumference and breastfeeding duration or developmental milestones, respectively).
a Arrows indicate the direction in the change of the mean response of the outcome (e.g., j indicates decreased mean birth weight).
b Arrows indicate the change in risk of the outcome (e.g., | indicates an increased risk of the outcome).
cChen, 2012, 1332466 reports results from a population overlapping with Chen et al. {, 2017 3981292}, which was considered the most updated data.
dFei, 2007, 1005775 reports results from a population overlapping with Meng et al. {, 2018, 4829851}, which was considered the most updated data.
e Grice, 2007, 4930271 reported results from children born to women in an occupational cohort.
fWashino et al. {, 2009, 1291133} reports results from a population overlapping with Kashino et al. {, 2020, 6311632}, which was considered the most updated data.
3-193
-------
APRIL 2024
3.4.4.1.4 Study Inclusion For Updated Literature Searches
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). More specifically, 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 therefore not further considered here.
In general, to best gauge consistency and magnitude of reported associations, EPA largely
focused on the most accurate and most prevalent measures within each fetal growth endpoint.
Studies with overlapping cohorts were included in the synthesis, as each study provided some
unique data for different endpoints. Specifically, 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 (received via communication with
study 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.
Following exclusion of the nine studies noted above, 69 developmental epidemiological studies
were included in the synthesis that were not included in the 2016 PFOS HESD. Six additional
3-194
-------
APRIL 2024
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, 12 examined postnatal growth, 5 each examined fetal loss,
and birth defects.
3.4.4.1.5 Growth Restriction: Fetal Growth
3.4.4.1.5.1 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 3-45,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 postpartum {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
one biomarker measure was used here (e.g., preferably maternal samples when collected in
3-195
-------
APRIL 2024
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
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 12 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 27
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,
3-196
-------
APRIL 2024
3858531; Wang, 2019, 5080598} as shown in
Alkhalawi etal., 2016, 3859818
Ashley-Martin et al„ 2017, 3981371
Bach etal., 2016, 3981534
Bell etal., 2018, 5041287
Bjerregaard-Olesen et al., 2019, 5083648
Callan et al., 2016, 3858524
Caoetal., 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 etal, 2020, 7102797
Espindola Santos etal, 2021, 8442216-
Gao etal, 2019, 5387135-
Gennings et al, 2020, 7643497 -
Govarts et al, 2016, 3230364
Gross etal, 2020, 7014743
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 3-45, Figure 3-46, and Figure 3-47. The median PFOS 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}.
3-197
-------
APRIL 2024
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 3-45. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Birth Weight Effects "
Interactive figure and additional study details available on IiAWC.
aIncludes six overlapping studies {Bjerregaard-Olesen, 2019, 5083648; Kishi, 2015, 2850268; Kobayashi, 2017, 3981430; Li,
2017, 3981358; Minatoya, 2017, 3981691; Rokoff, 2018, 4238310}.
3-198
-------
APRIL 2024
A\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
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 3-46. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Weight Effects (Continued)3
Interactive figure and additional study details available on IiAWC.
aInclndes six overlapping studies {Bjerregaard-Olesen, 2019, 5083648; Kishi, 2015, 2850268; Kobayashi, 2017, 3981430; Li,
2017, 3981358; Minatoya, 2017, 3981691; Rokoff, 2018, 4238310}.
3-199
-------
APRIL 2024
&
pe
_i I
Marks etal., 2019, 5081319-
-
++
++
-
+
+
+
-
Meng et al„ 2018, 4829851 -
+
B
++
+
++
+
++
+
Minatoya et al., 2017, 3981691 -
++
++
++
+
++
+
+
++
Robledo et al,, 2015, 2851197-
B
++
+
++
+
++
•
Rokoff et al., 2018, 4238310 -
++
++
++
+
++
+
t
++
Sagivetal., 2018, 4238410-
++
+ +
+
++
+
++
++
Shi etal., 2017, 3827535-
¦
++
++
B
B
+
+
Shoaff etal., 2018, 4619944-
++
++
++
++
++
+
++
++
Starling et al., 2017, 3858473-
++
++
++
++
+
++
++
Valvi et al., 2017, 3983872-
++
++
++
++
+
++
++
Wang et al., 2019, 5080598 -
B
++
++
++
+
-
~
Wikstrom et al., 2020, 6311677 -
++
++
++
++
+
++
++
Woods et al., 2017, 4183148 -
XI
B
++
++
+
++
B
Workman et al., 2019, 5387046 -
X
++
++
++
-
•
~
Xiao etal., 2020, 5918609-
++
++
++
++
+
++
++
Xu et al., 2019, 5381338-
¦
++
++
B
+
+
Yao etal., 2021,9960202-
++
++
++
+
+
++
de Cock etal., 2016, 3045435-
+
+
++
+
+
+
-
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 3-47. Summary of Study Evaluation for Epidemiology Studies of PFOS and Birth
Weight Effects (Continued)3
Interactive figure and additional study details available on HAWC.
3-200
-------
APRIL 2024
3.4.4.1.5.1.1 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 3-48, Figure 3-49, Figure 3-50, Figure 3-51, and Figure 3-52).
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 10 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 from -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
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. However, five {Bach, 2016, 3981534; Hjermitslev, 2020, 5880849; Meng, 2018,
4829851; Sagiv, 2018, 4238410; Wikstrom, 2020, 6311677} of eight medium and high
confidence studies still reported evidence of mean BWT deficits based on early pregnancy
biomarker samples.
3-201
-------
APRIL 2024
3.4.4.1.5.1.2 Mean BWT-Overall Population Summary
Eighteen of the 19 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 10 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.
Reference, P
Period° Confidence Study design Matrix^ Sub-population Exposure levels Comparison EE
Rating
Effect Estimate
-150 -100 -50 0 50 100
Earty Bach et al. (2016, Cohort maternal Term births (GA >= median=8.3 ng/mL (25th-75th Regression
pregnancy 3981534), High serum 37 weeks) percentile: 6.0-10.8 ng/mL) coefficient per IQR
(4.8 ng/mL) increase ^
1
1
1
1
1
1
Regression
coefficient for Q2
(6.03-8.29 ng/mL)
vs. Q1 (<6.03
ng/mL)
1
1
1
• 1
1
1
1
Regression
coefficient for Q3
(8.30-10.80 ng/mL) „
vs. Q1 (<6.03
ng/mL)
1
1
1
• 1—
1
1
1
Regression
coefficient for Q4
(10.81-36.10 ng/mL)
vs. Q1 (<6.03
ng/mL)
1
1
1
• I
1
1
1
Later Bell et al. (2018, Cross-sectional blood Singleton median=1.72 ng/mL Regression
pregnancy 5041287), High (25th-75th percentile: coefficient (per
1.14-2.44 ng/mL) log(PFOS+1) unit ,18 3
increase)
i
i
i
i
i
i
Chu et al. (2020, Cohort maternal - median=7.153 ng/mL (25th Regression
6315711), High serum percentiles.361 ng/mL, 75th coefficient (per 11n
percentile=11.928 ng/mL) change in PFOS) 833
i
i
i
• i
i
i
i
Eick at al. (2020, Cohort serum full-term births median= 1.93 ng/mL Regression
7102797), High (25th-75th percentile= 1.18 - Coefficient [for T2
3.13 ng/mL) (1.40-2.56 ng/ml) vs. . 6
T1 (<1.40 ng/ml)]
I
i
i
•
i
i
I
Regression
Coefficient [for T3
(>2.56 ng/ml) vs. T1 14
(<1.40 ng/ml)]
i
i
i
1 »
i
i
i
-150 -100 -50 0 50 100
Figure 3-48. Overall Mean Birth Weight from Epidemiology Studies Following Exposure to
PFOS
Interactive figure and additional study details available on HAWC
3-202
-------
APRIL 2024
Sampling
Period
Early
pregnancy
Later
pregnancy
Study design
Manzano-Saigado e* Cohort
ai. (2017, 423S465i
High
Govarts et at. {2016 Cohort
3230364). High
Launc^r a
i2.P -i9S14P
hgii
L^cetal 102!
Q^SyclUi Hish
Exposure
Matrix
p:asma
maternal
blood
Sub-population
Exposure levels
Comparison
cord blood -
riatPHJ
serum
maternal
blood, cord
blood
Mpan ""D ^ ny nL 2 T4 P^nn- in
ngj/nU ujeffi-wut ,uhona«
ip . irfh HQhtpter
rhuP nj o PFjs
Ptjgre jr
l efi mo -sr "irth
R-jr* n
niJ*fi ent'crhrtli
weight (03 vs 01 j
Rearession
Mi-iiupnt f-rtirth
wetgm i,u4 « ay
geometric mean = 2.63 ug/L R~qi>
(25ttv75th percentile = m'-'frem per IQR
170-3.80 ug/L) '
i hue* .n FFOS
Fw c i ed etn=c "*4 ni nL Reirn i.n
r
-------
APRIL 2024
Sampling
Period
Early
pregnancy
Later
pregnancy
Reference,
Confidence Study di
Rating
Manzano-Salgado et Cohort
al. (2017, 4238465)
High
Exposure
Matrix
plasma,
maternal
blood
Sub-population Exposure levels
Comparison
Effect Estimate
Sagivetal. (2018,
4238410), High
maternal
blood
Wikstrom et al.
(2020, 6311677),
High
maternal
serum
Starling et al. (2017, Cohort
3858473), High
Valvi et al. (2017,
3983872), High
maternal
serum
maternal
serum
Yao et al. (2021,
9960202), High
Cross-sectional other
maternal
serum
- Mean (SD): 6.05 Regression coefficient
ng/mL (2.74 ng/mL) (change in birth weight per 0.4
doubling of PFOS)
Regression coefficient for
birth weight (Q2vsQ1) 23.6
Regression coefficient for
birth weight (Q3vsQ1) 38.7
Regression coefficient for
birth weight (Q4vsQ1) 8.2
- median=25.7 ng/mL Regression coefficient per
(IQR: 16.0 ng/mL) IQR increase -17.9
Regression coefficient (for
Q2 [18.9 - 25.6 ng/mL] vs -27 8
Q1 [0.1 -18.8 ng/mL))
Regression coefficient (for
Q3VSQ1) -36.3
Regression coefficient (for
Q4vsQ1) -57.6
- Median=5.38 ng/mL Regression coefficient (per
(25th-75th 1-ln ng/mL change in -46
percentiles: PFOS)
3.97-7.60 ng/mL) Regression coefficient (for
Q2vsQ1) -27
Regression coefficient (for
Q3vsQ1) -22
Regression coefficient (for
Q4vsQ1) -80
- median=2.4 ng/mL Regression coefficient (per
(25th percentile=1.5. 11n increase in PFOS) -13.8
75th percentile=3.7)
Regression coefficient for
tertiie 2 (1.8-3.2 ng/mL) vs. -33.8
fertile 1 (
-------
APRIL 2024
«. .. Reference, F
Period9 Confidence Study design Matrix Sub-population Exposure levels Comparison EE
Rating
Effect Estimate
0 200 400 600 800
Early Chang et al. (2022, Cohort maternal term births median: 2.19 ng/mL Regression
pregnancy 9959688), Medium serum (25th-75th percentile: coefficient (per
1.45-3.24 ng/mL) doubling in PFOS) j
1
1
1
1
1
1
Regression
coefficient [for 02
(1.44-2.19 ng/mL) 78
vs. Q1 (1899 .™4
ng/L) vs. Tertile 1
(<1200 ng/L)
i
i
i
i •
i
i
i
Gyllenhammar et al. Cohort and maternal - - Regression
(2018,4238300), cross-sectional serum coefficient per
Medium unit-log increase in 1Q r
PFOS "Jy:>
i
i
i
i
i
i
0 200 400 600 800
Figure 3-51. Overall Mean Birth Weight from Epidemiology Studies Following Exposure to
PFOS (Continued)
Interactive figure and additional study details available on HAWC.
3-205
-------
APRIL 2024
^Pcrtod^ Confidence Study design Sub-population Exposure levels Comparison EE
Rating
Effect Estimate
-150 -100 -50 0 50
Early Hjermitslev et al. Cohort maternal - median=8.99 Regression coefficient
pregnancy (2020,5880849), serum ng/mL (min-max: (per 11n-ng/mL
Medium 1.50-61.3) increase in PFOS) ~5-5
1
1
1
1
male median=8.99 Regression coefficient
ng/mL (min-max: (per 1In-ng/mL
1.50-61.3) increase in PFOS) "3 ®
1
—
1
1
female median=8.99 Regression coefficient
ng/mL (min-max: (per 11n-ng/mL
1.50-61.3) increase in PFOS) "48
1
1
1
1
Meng et al. (2018. Cohort maternal - Median (25th-75tfi Regression coefficient
4829851), Medium serum percentiles): 30.1 (per doubling of
ng/mL (22.9-39.0 PFOS) "45 2
ng/mL)
1
1
• 1
1
1
Regression coefficient
for Q2 vs. Q1
24.7
1
1 a
1
1
Regression coefficient
forQ3vs. Q1
-50.1
1
» 1
I
1
Regression coefficient
for 04 vs. 01
-48 2
1
1
» I
I
1
Later Kashino et al. (2020, Cohort plasma - Median=3.4 ng/mL Regression coefficient
pregnancy 6311632), Medium (25th-75th (per Iog10 change in
percentile: 2.6-4.7 PFOS) "35
ng/mL)
I
• 1
I
1
Kwon et al. (2016, Cross-sectional cord blood - median=0.64 Regression coefficient
3858531), Medium ng/mL (25th-75th (per 1 log-unit change
percentile = in PFOS) "49 4
0.29-1.09 ng/mL)
1
1
« 1
1
1
Lenters et al. Cohort maternal - geometric Regression coefficient
(2016,5617416), serum mean=9.357 per 2-SD (1.600
Medium ng/mL (2-SD ng/mL) increase in ~688
In-PFOS: 1.600) In-ng/mL PFOS
1
1
• >
1
1
Wang et al. (2019, Cross-sectional cord blood - Median (25th-75th Regression coefficient
5080598), Medium percentiles)= (per 1-log10 change
0.65ng/mL in PFOS) "54 5
(0.40-1.19ng/mL)
1
m 1
1
I
Woods et al. (2017, Cohort maternal - median=14.4 ug/L Regression coefficient
4183148), Medium serum (25tti-75th (perlog10-ug/L
percentile: 10-17.9 increase maternal "8 7
ug/L) PFOS)
1
I
+ 1
1
1
-150 -100 -50 0 50
Figure 3-52. Overall Mean Birth Weight from Epidemiology Studies Following Exposure to
PFOS (Continued)
Interactive figure and additional study details available on HAWC.
3.4.4.1.5.1.3 Mean Birth Weight Study Results: Sex-Specific Studies
Ten of 16 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 15 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 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 (P per each ln-unit PFOS increase: -8.5 g; 95% CI: -15.9, -1.1) and showed an
3-206
-------
APRIL 2024
exposure-response with reported large deficits in PFOS tertile 2 (P: -26.6 g; 95% CI: -147.3,
94.2) and tertile 3 (P: -83.9 g; 95% CI: -201.4, 33.7) compared with 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 inverse 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 with early sampled studies (38%). Study sensitivity was not an
explanatory factor among the null studies in either sex.
3.4.4.1.5.1.4 Standardized Birth Weight Measures
Fifteen studies examined standardized BWT measures including 14 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 3-45, Figure 3-46, Figure 3-47).
Nine of the 15 studies showed some evidence of inverse 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 ratio (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 14 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 quartile 2 (P: -0.15; 95% CI: -0.29, -0.02) and quartile 4 (P: -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} (P: -0.15; 95% CI: -0.41, 0.11) and Shoaff et al.
{,2018, 4619944} reported associations similar in magnitude for their overall population (P:
-0.12; 95% CI: -0.36, 0.13). The medium confidence study by Chen et al. {, 2017, 3981292}
3-207
-------
APRIL 2024
reported inverse associations in the overall population (P: -0.14; 95% CI: -0.26, -0.01) with
comparable 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 (P: -0.13; 95% CI: 0.26, 0.00); the largest association in this study
was found for male neonates (P: -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 inverse
associations (P per each ln-unit increase: -0.10; 95% CI: -0.20; -0.004) as well as in quartile 4
in the overall population (P: -0.17; 95% CI: -0.37, -0.03); these results appeared to be driven by
associations detected in female neonates (P per each ln-unit increase: -0.17; 95% CI: -0.30, -
0.03; P for quartile 4: -0.30; 95% CI: -0.49, -0.10). The high confidence study by Xiao et al. {,
2019, 5918609} reported z-scores fairly similar in magnitude for the overall population (P:
-0.47; 95% CI: -0.85, -0.09), male neonates (P: -0.40; 95% CI: -0.89, 0.08), and female
neonates (P: -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 (P: -0.62; 95% CI: -0.96 to -0.29). The authors also reported large
deficits for both males (P: -0.81; SE = 0.24; p-value = 0.001) and females (P: -0.46; SE = 0.29;
p-value = 0.11) for PFOS levels greater than the mean level.
3.4.4.1.5.1.5 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 studies in general. Twice as many studies
showing inverse 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, 9 different studies out of 15 showed some
suggestion of inverse associations in the overall population or either or both sexes.
3.4.4.1.5.2 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, 11 studies examined either or both LBW or SGA in relation to PFOS
exposure with 4 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
3-208
-------
APRIL 2024
confidence, {Chang, 2022, 9959688; Souza, 2020, 6833697; Xu, 2019, 5381338} and one as
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 increased risk, 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 increased
risk 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 3-53
and Figure 3-54) 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 per each ln-unit change = 1.19; 95% CI: 0.87, 1.64) 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 (Figure 3-53 and Figure 3-54). SGA findings from low confidence
studies are not included in figures. 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. Few patterns were discernible across study characteristics
or study confidence for these SGA findings, although the number of studies was small.
3-209
-------
APRIL 2024
Sampling
Period
Reference,
Confidence
Rating
Exposure Study
Matrix Design
Exposure Levels
Sub-population
Comparison
EE
Effect Estimate
0 1 2 3 4 5 <
5 7
Early
pregnancy
Manzano-
Salgado et al.
(2017,
plasma, Cohort
maternal
blood
Mean (SD): 6.05 ng/mL (2.74 ng/mL)
Boys
OR (per doubling in
maternal plasma PF..
1.01
i
-f-
4238465)
High
Girls
OR (per doubling in
maternal plasma PF..
0.84
1
1
OR (per doubling in
maternal plasma PF..
0.92
J-
1
Wikstrom et
al. (2020,
6311677),
maternal Cohort
serum
Median=5.38 ng/mL (25th-75th percentiles: 3.97-7.60
ng/mL)
Boys
OR (per 1-ln ng/mL
change in PFOS)
1.08
1*
High
OR (for Q2vsQ1)
1.26
! •
i
OR (for Q3 vs Q1)
0.86
i
i
OR (for 04 vs Q1)
1.3
!
i
Giris
OR (per Hn ng/mL
change in PFOS)
1.4
i
¦ «
1
OR (for Q2 vs Q1)
0.89
1
-i—
OR (for 03 vs 01)
0.82
1
i
OR (for 04 vs 01)
2.05
1
¦ a
i
OR (per 1-ln ng/mL
change in PFOS)
1.19
1
i
OR (for 02 vs 01)
0.69
i
i
OR (for 03 vs 01)
0.79
1
-i
OR (for 04 vs 01)
1.56
|
i *
Later
pregnancy
Lauritzen et
al. (2017,
3981410)
maternal Cohort
serum
median=9.74 ng/mL (range: 0.95-59.6 ng/mL)
Norway
OR (per In unit
increase in PFOS)
0.71
*!
High
median=16.4 ng/mL (range: 2.28-55.2 ng/mL)
Sweden
OR (per In unit
increase in PFOS)
2.51
i
¦ s
i
Norway: median=9.74 ng/mL (range: 0.95-59.6 ng/mL);
Sweden: median=16.4 ng/mL (range: 2.28-55.2 ng/mL)
OR (per In unit
increase in PFOS)
0.95
i
0 12 3
4 5 (
> 7
Figure 3-53. 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 HAWC.
Small for gestational age defined as birthweight below the 10th percentile for the reference population.
3-210
-------
APRIL 2024
Sampling
Period
Reference,
Confidence
Rating
Exposure
Matrix
Study Design
Exposure Levels
Sub-population
Comparison
EE
Effect Estimate
0.0 0.5 1.0 1.5 2.0 2.5
Early
pregnancy
Chang et al.
(2022,
9959688),
Medium
maternal
serum
Cohort
median: 2.19 ng/mL
(25th-75th percentile:
1.45-3.24 ng/mL)
term births
OR (per doubling in
PFOS)
1.12
1
1
1
—1—•
1
1
1
OR [for Q2
(1.44-2.19 ng/mL)
vs. Q1 (
-------
APRIL 2024
„ Reference, Measured „ c. .
pePod» —ce ^ MPamx fig, Sub,„pulat,o„ Comparison EE
Effect Estimate
0 5 10 15 20 25
Early Manzano-Salgado Low birth plasma, Cohort - OR (per doubling in
pregnancy et al. (2017, weight maternal maternal plasma PFOS) 06
4238465) High blood
1
1
-f"
l
Boys OR (per doubling in
maternal plasma PFOS) 19
1
1
1 *
1
Girls OR (per doubling in
maternal plasma PFOS) q.73
I
1
1
Low birth plasma, Cohort - OR (per doubling in
weight at maternal maternal plasma PFOS) o.91
term blood
I
1
¦#*
l
Boys OR (per doubling in
maternal plasma PFOS) 1 68
1
1
1 •
1
Girls OR (per doubling in
maternal plasma PFOS) o.73
1
l
T
1
Hjermitslev et al. Low birth maternal Cohort - OR (per 11n-ng/mL
(2020,5880849), weight serum change in PFOS) 103
Medium
1
1
1
Meng et al. (2018. Low birth maternal Cohort - OR (per doubling of
4829851), weight serum PFOS) 13
Medium
1
1
T*-
1
OR (for 02 vs. Q1)
1.4
1
_l^
1
OR (for 03 vs. Q1)
1.8
I
1
1 *
1
OR (for 04 vs. 01)
1.2
1
1
1
Later Chu et al. (2020. low birth maternal Cohort - OR (per 1 In ng/mL
pregnancy 6315711), High weight serum increase in PFOS) 2.43
1
1
*
1
OR for 02 (> 4.36 to 7.15
ng/mL PFOS) vs. 01 0 83
(<=4.36 ng/mL PFOS)
f
1
H
1
OR for 03 (>7.15 to
11.93 ng/mL PFOS) vs. 1 41
Q1 (<=4.36 ng/mL PFOS)
1
1
i*
1
OR for 04 (>11.93
ng/mL PFOS) vs. 01 37
(<=4.36 ng/mL PFOS)
1
1
1 •
1
0 5 10 15 20 25
Figure 3-55. Odds of Low Birthweight in Children from Epidemiology Studies Following
Exposure to PFOS
Interactive figure and additional study details available on HAWC.
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 = 1) 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.
3-212
-------
APRIL 2024
<\6^
vC.0
Arbuckle et al., 2013, 2152344 ¦
Chang et al., 2022, 9959688
Chu et al., 2020, 6315711
Govarts et al., 2018, 4567442 •
Gundacker et al., 2021, 10176483'
Hjermitslev et al., 2020, 5880849
Lauritzen etal., 2017, 3981410'
Manzano-Salgado et al., 2017, 4238465 ¦
Meng etal., 2018, 4829851
Souza et al., 2020, 6833697
Wikstrom et al., 2020, 6311677 ¦
Xu etal., 2019, 5381338-
—1—
+
I
+
I
+
I
I
+
I
+
I
+
~
+
B
&
+
++
+
+
+* 1
+
++
++
+
++
+
++
++
+
B
++
+
++
+
+
Ui
+
H
1
B
-
+
-
n
+
M
+
+
-
++
B
„
++
B
+
++
+
++
++
++
++
+
++
+
++
++
B
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)
* Multiple judgments exist
Figure 3-56. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Small for Gestational Age and Low Birth Weight Effects
Interactive figure and additional study details available on HAWC.
3.4.4.1.5.3 Birth Length
Thirty-one birth length studies were considered as part of the study evaluation as shown in
Figure 3-57. and Figure 3-58. Four studies were considered uninformative {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;
Kobayashi, 2022, 10176408} were not further considered for multiple publications from the
same cohort studies. Twenty-three non-overlapping and informative studies examined birth
3-213
-------
APRIL 2024
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.
3-214
-------
APRIL 2024
^drisvs_ a0c^!^6s s®rt^x &o^
,o®
Aikhalawi et al., 2016, 3859818 -
-
+
+
-
-
+
+
Bach et al., 2016, 3981534-
+
+
+
+
+
Bell et al., 2018, 5041287-
+
+
+
-
Bjerregaard-Olesen et al., 2019, 5083648 -
+
+
+
Callan et al., 2016, 3858524 -
+
+
+
-
+
+
+
-
Cao et al., 2018, 5080197-
-
+
+
-
+
+
+
-
Chen et al., 2017, 3981292-
+
+
++
+
++
+
++
+
Chen et al., 2021, 7263985-
+
++
+
+
++
+
++
+
Espindola Santos et al., 2021, 8442216 -
+
++
+
+
-
+
-
-
Gao et al., 2019, 5387135-
+
++
+
-
-
+
+
-
Gundackeretal., 2021, 10176483-
+
+
+
B
-
+
-
~
Gyllenhammar et al., 2018, 4238300 -
+
+
+
+
++
+
++
+
Hjermitslev et al., 2020, 5880849 -
+
+
+
+
+
-
++
+
Jin et al., 2020, 6316202-
-
+
+
H
-
+
-
n
Kashino et al., 2020, 6311632 -
+
++
+
+
+
+
++
+
Kishi et al.,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 3-57. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Birth Length Effects a
Interactive figure and additional study details available on IiAWC.
aIncludes three overlapping studies: Bjerregaard-Olsen et al. {, 2019, 5083648}; Kishi et al. {, 2015, 2850268}; Kobayashi et al.
{,2017,3981430}.
3-215
-------
APRIL 2024
.fiCk
c,oSe<
-------
APRIL 2024
Of the 23 studies examining either standardized birth length or mean birth length measures,
seven studies showed some inverse 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 (P per each log2 increase in PFOS: -0.33; 95%
CI: -0.69, 0.03) in the overall population, as well as for both male (P: -0.41; 95% CI: -0.87,
0.05) and female neonates (P: -0.23; 95% CI: -0.75, 0.30). Although smaller in magnitude, the
medium confidence study by Chen et al. {, 2017, 3981292} also reported a birth length deficit of
-0.16 per each ln-unit PFOS increase (95% CI: -0.31, -0.02) in the overall population as well as
male (|3: -0.15; 95% CI: -0.33, 0.03) and female neonates (P: -0.20; 95% CI: -0.44, 0.05). The
other high confidence study by Shoaff et al. {, 2018, 4619944} of standardized birth length
measures showed a deficit only for tertile 3 (P: -0.24; 95% CI: -0.64, 0.15) compared with
tertile 1.
Four {Callan, 2017, 3858524; Chen, 2021, 7263985: Lauritzen, 2017, 3981410; Workman,
2019, 5387046} of the 16 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 (P: -0.3 cm; 95% CI:
-0.7, 0.1), but detected the strongest association when restricted to the Swedish population (P:
-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 (P per each PFOS ln-unit increase: -
0.27 cm; 95% CI: -0.51, -0.02), males (P: -0.14 cm; 95% CI: -0.55, 0.26), and females (P: -
0.40 cm; 95% CI: -0.74, -0.06). The low confidence study by Workman et al. {, 2019, 5387046}
reported a non-statistically significant birth length reduction (P per each ln-unit PFOS increase: -
0.16 cm; 95% CI: -0.92, 0.60). 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 inverse
associations (P for tertile 3 vs. tertile 1: -0.52 cm; 95% CI: -1.05, 0.01). The high confidence
study by Valvi et al. {,2017, 3983872} reported no associations in the overall population but did
detect a nonsignificant birth length deficit in male neonates (P per each PFOS log2 exposure
increase: -0.18 cm; 95% CI: -0.60, 0.23). The low confidence study Wang et al. {, 2019,
5080598} study also reported a nonsignificant birth length deficit in males that was similar in
magnitude (P: -0.17 cm; 95% 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 (P per each IQR PFOS increase: -0.3 cm; 95% CI: -0.7,
0.0). 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 inverse associations
either in the overall population, or in either or both sexes. Two of 10 studies in females and four
3-217
-------
APRIL 2024
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 with the overall population. None of the five studies examining categorical data in
either 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.
3.4.4.1.5.4 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 confidence
{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 11 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}.
3-218
-------
APRIL 2024
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 3-59. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Birth Head Circumference Effects
Interactive figure and additional study details available on IiAWC.
3-219
-------
APRIL 2024
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 (P: -0.26; 95% CI: -0.68, 0.16), as well as male (P:
-0.15; 95% CI: -0.68, 0.39) and female neonates (P: -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 (P per each ln-unit
PFOS change: -0.4 cm; 95% CI: -0.9, 0.04).
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,
nonsignificant head circumference differences (P per each PFOS IQR increase: -0.1 cm; 95% CI:
-0.2, 0.1). In their low confidence study, Cao et al. {, 2018, 5080197} reported a nonsignificant
inverse association in the overall population (P per each ln-unit PFOS: -0.23 cm; 95% CI: -1.19,
0.73) as did the low confidence study by Callan et al. {, 2016, 3858524} (P per each ln-unit
PFOS: -0.39 cm; 95% CI: -0.98, 0.20).
Two of 10 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 (P per each doubling of serum PFOS: -0.28 cm; 95% CI: -0.65, 0.09),
while head circumference increases were found for female neonates (P: 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 (P: -0.13 cm; 95% CI: -0.45, 0.19) and 3 (P:
-0.31 cm; 95% CI: -0.62, 0.01) compared with tertile 1. The medium confidence study by
Kashino et al. {, 2020, 6311632} reported smaller deficits only in male neonates (P per each
loglO PFOS: -0.14 cm; 95% CI: -0.61, 0.32). 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 (P per each IQR PFOS increase:
-0.1 cm; 95% CI: -0.3, 0.1). The low confidence study by Cao et al. {, 2018, 5080197} found a
3-220
-------
APRIL 2024
large head circumference difference (P for tertile 3 vs. 1: -1.22 cm; 95% CI: -2.70, 0.25) 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 subgroups precluded a detailed
assessment, few patterns were evident across the 10 different studies that showed some inverse
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 inverse 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.
3.4.4.1.5.5 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. However, there were also
inverse associations observed in multiple studies based on early pregnancy biomarker samples.
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 inverse 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.
3-221
-------
APRIL 2024
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 sizable 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 (studies published since the 2016 PFOS HESD) is fairly strong 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 the 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 postpartum maternal samples) that may be more prone to pregnancy
hemodynamic impacts. This observation is in agreement with the results of Dzierlenga et al. {,
2020, 7643488}, though there was also evidence of associations in studies less likely to be
biased by pregnancy hemodynamics (i.e., preconception or trimester 1 sampling). Therefore,
despite consistency in evidence across some of 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.
3.4.4.1.6 Postnatal growth
Eleven studies examined PFOS exposure in relation to postnatal growth measures (Figure 3-60).
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 3-60, seven postnatal growth studies had
good study sensitivity {Chen, 2017, 3981292; Gyllenhammar, 2018, 4238300; Jensen, 2020,
3-222
-------
APRIL 2024
6833719; Lee, 2018, 4238394; Manzano-Salgado, 2017, 4238509; Shoaff, 2018, 4619944;
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 for height at age 2 years (P
per each PFOS ln-unit increase: -0.77 cm; 95% CI: -1.27, -0.15) as well as height change from
birth to 2 years (P: -0.71 cm; 95% CI: -1.27, -0.15). Small differences were seen for mean
weight differences at age 2 years (P: -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 with 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 (P per each ln-unit PFOS increase: -0.25; 95% CI: -0.47, -0.04)
or 12 to 24 months (P: -0.25; 95% CI: -0.41, -0.06). Females seemed to drive the deficit
detected in the overall population (P per each ln-unit PFOS increase: -0.13; 95% CI: -0.32,
0.07) 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 with 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) (P: 1.37 cm; 95% CI: -0.5, 3.28), while large
postnatal weight deficits were reported for PFOS tertiles 2 (P: -138 g; 95% CI: -574, 298) and 3
(P: -78 g; 95% CI: -532, 375).
Associations at five months of age in the overall population (P: -0.28; 95% CI: -0.51, -0.05) and
females (P: -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
(P: 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 (P: -0.33; 95% CI: -0.65, -0.01) and weight-for-length z-score (P:
-0.34; 95% CI: -0.59, -0.08) in PFOS tertile 3 compared with tertile 1. Small deficits that were
not statistically significant were observed in tertile 3 for length for age z-score (P: -0.22; 95%
CI: -0.49, 0.04). In their high confidence study, Yeung et al. {, 2019, 5080619} reported
3-223
-------
APRIL 2024
statistically significant negative growth trajectories weight-for-length z-scores in relation to each
log SD increase in PFOS exposures among singletons followed for 3 years. No associations were
detected for 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 (P: -0.10; 95% CI:
-0.16, -0.05) and weight z-score (P: -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 (P: -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 inverse associations with PFOS.
Study quality ratings, including study sensitivity and overall confidence, did not appear to be
explanatory factors for heterogeneous results across studies.
3-224
-------
APRIL 2024
Cao et al., 2018, 5080197
Chen etal., 2017, 3981292-
Gross et al„ 2020, 7014743
Gyllenhammar et al,, 2018, 4238300
Jensen et al., 2020, 6833719 -
Lee etal., 2018, 4238394
Manzano-Salgado et al., 2017, 4238509
Shoaff et al., 2018, 4619944
Starling et al., 2019, 5412449
Yeung et al,, 2019, 5080619
de Cock et al., 2014, 2713590
y W© , A\ O0^
-
+
++
-
+
+
+
-
+
+
++
+
++
+
++
+
+
+
+
+
-
+
-
+
+
++
+
++
+
++
B
+
+
+
+
+
+
++
++
D
++
B
+
++
B
¦
++
++
B
++
+
++
B
++
++
++
++
++
+
++
++
++
++
&
++
++
+
++
++
++
B
++
B
++
+
B
++
D
H
++
+
-
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 3-60. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOA Exposure and Postnatal Growth
Interactive figure and additional study details available on HAWC.
3.4.4.1.6.1 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 (P:
-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
driven by results in females especially for the 6 to 12 months (P: -0.33; 95% CI: 0.59, -0.08)
3-225
-------
APRIL 2024
and 12 to 24 months (P: -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 3 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 (P: -0.11; 95% CI: -0.17, -0.05) and BMI (P: -0.16 kg/m2;
95%) CI: -0.24, -0.08). In the high confidence study by Starling et al. {, 2019, 5412449},
decreased adiposity ((3: -2.08; 95%> CI: -3.81, -0.35) among girls was detected in PFOS tertile 3
compared with 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.
3.4.4.1.6.2 Rapid Weight Gain
Four high confidence studies {Manzano-Salgado, 2017, 4238509; Shoaff, 2018, 4619944;
Starling, 2019, 5412449; Yeung, 2019, 5080619} examined rapid infant growth. Limited
evidence of associations was reported, as only one {Starling, 2019, 5412449} of four studies
{Manzano-Salgado, 2017, 4238509; Shoaff, 2018, 4619944; Starling, 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.
3.4.4.1.6.3 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 inverse 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 postnatal growth measures up
to 2 years of age.
3.4.4.1.7 Gestational 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
3-226
-------
APRIL 2024
gestational age measures, while two studies only examined PTB {Liu, 2020, 6833609; Gardener,
2021, 7021199}.
3.4.4.1.7.1 Gestational Age
Seventeen of the 20 studies reporting gestational age estimates in relation to PFOS exposures
were considered (Figure 3-61). Two studies were deemed uninformative {Gundacker, 2021,
10176483; Lee, 2013, 3859850} and were excluded and one study was excluded 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 and one study reported
sex-specific results only {Lind, 2017, 3858512}.
Among the 17 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,
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, 2018, 5041287}.
3-227
-------
APRIL 2024
#*•
,0©
Bach etal.,2016, 3981534-
Bangma et al., 2020, 6833725-
Bell et al., 2018, 5041287-^
Chu et al., 2020, 6315711 -
Eick et al.. 2020, 7102797
Gao et al., 2019, 5387135
Gundacker et al., 2021, 10176483
Gyllenhammar et al., 2018, 4238300
Hjermitslev et al., 2020, 5880849
Huo et al., 2020, 6835452
Lauritzen et al., 2017, 3981410
Lee et al., 2013, 3859850-
Li et al., 2017, 3981358-
Lind et al., 2017, 3858512^
Manzano-Salgado et al., 2017, 4238465 -j
Meng et al., 2018, 4829851 -\
Sagiv et al., 2018, 4238410^
Workman et al., 2019, 5387046 -
Xu et al., 2019, 5381338-
Yang et al., 2022, 10176806-
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 3-61. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Gestational Age
Interactive figure and additional study details available on IiAWC.
3-228
-------
APRIL 2024
Nine of the 16 studies examining mean gestational age change in the overall population reported
some deficits. Among these nine studies, 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 IQRPFOS 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 (P: -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 (P
per each IQR increase: -0.19 weeks; 95% CI: -0.33, -0.05). The high confidence study by Chu
et al. {, 2020, 6315711} reported similar deficits in the overall population (P: -0.32 weeks; 95%
CI: -0.53, -0.11) which was driven by female neonates (P: -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 (P: -0.4 weeks; 95% CI: -0.9, 0.2). Compared with tertile 1, the low
confidence study by Gao et al. {, 2019, 5387135} reported deficits in tertile 2 (P: -0.40 weeks;
95% CI: -0.92, 0.12) and tertile 3 (P: -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 (P: -0.31 weeks; 95% CI: -0.55, -0.06) compared with quartile 1. Despite relatively
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 (P:
-1.26 weeks; 95% CI: - 2.46, -0.05) and linear PFOS (P per each IQR increase: -1.80 weeks;
95% CI: -3.24, -0.37), with results larger results in female (P: -1.06 weeks; 95% CI: -2.87,
0.74) than male neonates (P: -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 nonsignificant
decrease (P per each ln-unit PFOS change: -0.17 weeks; 95% CI: -0.52, 0.18).
Lind et al. {, 2017, 3858512} reported sex-specific changes in mean gestational age only.
Inverse associations were observed for both boys (P per ln-unit increase: -0.5 days, 95% CI: -
3.4, 2.3) and girls (P: -1.0, 95% CI: -4.2, 2.1), but neither was significant.
Overall, nine of the 16 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 to -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 inverse 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.
3-229
-------
APRIL 2024
3.4.4.1.7.2 Preterm Birth
As shown in Figure 3-62, 11 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 11 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, 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).
3-230
-------
APRIL 2024
,©xe ^
5^
5\VNx A\i>&
c,©^9 v\ o°^
Bach et al., 2016, 3981534-
1
+
i
+
++
i
+
++
—I—
+
++
++
Chu et al., 2020, 6315711 -
+
++
+
+
++
+
++
++
Eick et al., 2020, 7102797-
++
++
+
++
+
B
++
Gardener et al.. 2021, 7021199 -
++
++
+
++
+
B
++
Hjermitslev et al., 2020, 5880849 -
+
+
+
+
+
-
++
B
Huo et al., 2020, 6835452-
++
++
++
+
++
+
++
++
Liu et al., 2020, 6833609-
++
B
IB
-
++
+
+
•
Manzano-Salgado et al., 2017, 4238465-
++
++
++
+
++
+
++
++
Meng et al., 2018, 4829851 -
B
m
++
+
++
+
++
B
Sagiv et al,, 2018, 4238410 -
++
+
+
+
++
+
++
++
Yang et al„ 2022, 10176806 -
+
+
+
+
+
+
+
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)
Figure 3-62. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Preterm Birth Effects
Interactive figure and additional study details available on HAWC.
An increased risk was 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 at. {, 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, 3.32) as well as an exposure-response relationship across upper three quartiles (OR
3-231
-------
APRIL 2024
range: 2.22-4.99) exposures when compared with 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; tertile 3 OR= 1.87; 95% CI: 0.72, 4.88, compared with 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 logiounit 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 with 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 IQRPFOS 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
sizable 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 strong with respect to an increased risk 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 increased risk had limited exposure contrasts.
3.4.4.1.7.3 Gestational 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 studies showed some
increased risk of gestational duration deficits. 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 increased risk), 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 the qualitative analysis.
3-232
-------
APRIL 2024
3.4.4.1.8 Fetal Loss
As shown in Figure 3-63, five (two high, two medium, and one low confidence) studies examined
PFOS exposure and fetal loss. All of these studies had good study sensitivity owing largely to
very large 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 nonsignificant 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
relationship for miscarriages across PFOS quartiles (OR range: 1.1-1.4). Similarly, the low
confidence study by Jensen et al. {, 2015, 2850253} reported increased nonsignificant 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.
3-233
-------
APRIL 2024
,6c^ ^
,e^
&
Buck Louis et al., 2016, 3858527 -
+
+
+
+
+
+
++
B
Jensen et al., 2015, 2850253 -
+
++
+
+
+
+
++
Liew et al., 2020, 6387285-
B
B
++
+
++
+
++
B
Wang et al., 2021, 10176703-
++
++
++
+
++
+
++
++
Wikstrom et al., 2021, 7413606 -
++
++
++
+
++
+
++
++
Figure 3-63. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Fetal Loss
Interactive figure and additional study details available on HAWC.
3.4.4.1.9 Birth Defects
As shown in Figure 3-64, five (three medium and two 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 the 75th PFOS percentile level (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 per each ln-unit increase in PFOS = 0.51; 95% CI: 0.21-1.20). This risk
seemed to be largely driven by boys from Finland. The medium confidence study by 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,
3-234
-------
APRIL 2024
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
with 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.
Anand-Ivell et al., 2018, 4728675-
+
+
+
-
-
+
-
-
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 3-64. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOA Exposure and Birth Defects
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 16 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 20 studies are shown in Figure 3-65.
3-235
-------
APRIL 2024
&
Argus, 2000, 5080012
Butenhoffetal., 2009, 757873
Chen etal., 2018, 5080460
Conley etal., 2022, 10176381
Dangudubiyyam etal., 2022, 10429383
Era etal., 2009, 2919358
Fuentes et al., 2006, 757859
Lai etal., 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 etal., 2020, 7174720-
Xia et al., 2011, 2919267-
Zhang et al, 2021, 6988534 -
Zhang etal., 2020, 6315674
Zhong etal., 2016, 3748828-
•* ^
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 3-65. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFOA Exposure and Developmental Effects
Interactive figure and additional study details available on HAWC.
3-236
-------
APRIL 2024
Evidence indicates 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
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 Appendix C {U.S.
EPA, 2024, 11414344}).
3.4.4.2.1 Maternal Effects
Multiple developmental studies evaluated maternal weight outcomes in rats, mice, and rabbits
(Figure 3-66). 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 with controls. Dams in
the 2 and 8 mg/kg/day dose groups had significantly lower mean body weights on GD 14-17. 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 1-20). 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 with 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 pre-
cohabitation 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
3-237
-------
APRIL 2024
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
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 4-20) 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.
3-238
-------
APRIL 2024
Endpoint
Study Name
Study Design
Observation Timi
e Animal Description
Maternal Body Weight
Leeet al., 2015,2851075
developmental (GD11-16)
GD17
P0 Mouse. CD-1 { -,N=10)
Wan et al., 2020. 7174720
developmental (GD4.5-17.5)
GD17.5
P0 Mouse. CD-1 ( .-,N=8)
Fuentes et al,. 2006, 757859
developmental (GD6-18)
GD18
Mouse, CD-1 (i, N=10-11)
Lai et al., 2017, 3981773
developmental (GD1-17)
GD1-20
P0 Mouse. CD-1 (", N=18)
Mshaty et al.. 2020, 6833692
developmental (LD1-14)
PND21
P0 Mouse. C57BL/6J ( N=0-15)
Li etal., 2016, 3981495
developmental (GD12-18)
GD18
P0 Rat, Sprague-Dawley (i. N=10)
Butenhoff et al., 2009, 757873
developmental (GD0-PND20)
GD20
PO Rat, Crl:CD(SD) (" , N=23-25)
PND1
P0 Rat, Crl:CD(SD) ( :, N=23-25)
PND21
P0 Rat, Crl:CD(SD) (_:, N=23-25)
Luebkeretal., 2005, 1276160
reproductive (42d prior mating-LD20)
LD1
PO Rat, Crl:Cd (Sd)lgs Br Vaf fi'. N=24-25)
LD21
P0 Ral, Crl:Cd (Sd)lgs Br Vaf
-------
APRIL 2024
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
statistically 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 per doe, number of dead fetuses per 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) with 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
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 19-20, 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 1-14) 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).
3-240
-------
APRIL 2024
PFOS Dovclopmental Effects - Mortality
Endpoirt
Study Name
Study Design
Observation Tim
Animal Description
£ No significant change^^ Significant increase v Significant decrease
Abortions
Argus, 2000, 5080012
developmental (007-20)
GD29
P0 Rabbit. New Zealand (S, N=17-21)
Dams with Stillborn Pups
Fuentes et al.. 2006. 757859
developmental (GD6-18)
GD18
Mouse. CD-1 (*,N=10-11)
Luebkeret al.. 2005. 757857
reproductive (42d prior mating-LD4)
PND0
P0 Rat. Crl:Cd(Scl)lgs Vaf/Plus (u. N=17)
A 'WW
Luebkeret al.. 2005,1276160
reproductive (42d prior mating-LD20)
PND1
P0 Rat. Crl:Cd (Sd)lgs Br Vaf N=20-25)
. . —A
Fetuses. Dead
Argus. 2000. 5080012
developmental (GD7-20)
GD29
P0 Raboit. New Zealand (L1, N-12-20)
Lee eta!.. 2015,2851075
developmental (GD11-16)
GD17
P0 Mouse. CD-1 ('J. N=10)
A A A
Fuentes etal.. 2006. 757859
developmental (GD6-18)
GD18
Mouse. CD-1 {+, N=10-11}
Luebkeret al.. 2005. 757857
reproductive (76d (42d pre-cohabitation, 14d mating, GD0-20))
GD21
P0 Rat. Crl:Cd(Sd)lgs Vaf/Plus (L. N=8)
Fetuses. Dead per Litter
Fuentes etal.. 2006. 757859
developmental (GD6-18)
GD16
Mouse, CD-1 [+. N=10-11)
Fetuses. Live
Argus, 2000. 5080012
developmental (GD7-20)
GD29
P0 Rabbit. New Zealand (y, N=12-20)
Fuentes et al.. 2006. 757859
developmental (GD6-18)
GD18
Mouse. CD K.-,N=10-11>
Fetuses. Live (No. per Live Litter)
Conley et al.. 2022,10176381
developmental (GD14-18)
GD18
P0 Rat. Sprague-Dawley (J. N=4-6)
Implantation
Argus. 2000. 5080012
developmental (GD7-20)
GD29
P0 Raboit. New Zealand (L1, N=12-20)
Luebkeret al.. 2005. 757857
reproductive (42d prior mating-LD4>
LD5
P0 Rat. Cr1:Cd(Sd)lgs Vaf/Plus <•' . N=17)
Implantation Sites. Per Delivered Lister
Fuentes etal.. 2006. 757859
developmental (GOS-18)
GD18
Mouse, CD-1 N=10-11)
Live Pups Bom
Luebkeretal..2005. 757857
reproductive (42d prior mating-LD4)
PND0
P0 Rat. GtCdlSdilgs Vbf/Plus N=17)
Luebker etal.. 2005.1276160
reproductive (42d prior mating-LD20)
PNDl
F1 Rat. CrtCd (Sd)lfis Br Vaf < ,. N=20-25j
¦ y
Liveborn Puss. Mean/Litter
Zhang etal. 2021.6980534
developmental (GD12-18)
PND1
P0 Rat. Sprague-Dawley (' . N=8)
No. Dams with All Pups Dying. PND 1-4
Luebker etal.. 2005.1276160
reproductive (42d prior mating-LD20)
LD1-4
P0 Rat. Crl:Cd (Sd)lgs Br Vaf ("\ N=20-25)
¦ A
No. Dams with All Pups Dying. PND 1-5
Luebker etal.. 2005. 757857
reproductive (42d prior mating-LD4)
LD1-5
P0 Rat. Crl:Cd(Sd)lgs Vaf/Plus <- . N=17)
Mortality
Xiaotal., 2011.2919267
developmental (GD2-21)
PND3
F1 Rat, Spraguo-Dawloy N=10)
¦ . A
Offspring Survival
Lau Ct al.. 2003, 757854
developmental (GD1-171
PND0
F1 Mouse. CD-1 (? , N=7)
PND6
F1 Mouse, CD-1 < ? N=7)
¦ ¥W
PND24
F1 Mouse, CD-1 (?>,N=7>
« ¦ ¥W
Zhang etal, 2021.6988534
developmental (GD12-18)
PND14
F1 Rat, Sprague-Dawlay (. N-93-98)
Butonhoff ot al., 2009. 757873
developmental (GD0-PND20)
PND0-4
F1 Rat, Crl:CD(SD)< ' , N-23-25)
PND4-21
F1 Rat, Cil:CD(SD) , N=23-25>
Lau et al.. 2003, 757654
developmental (GD2-21)
PND0
F1 Rat, Sprague-Dawley (. . N-9)
PND5
F1 Rat, Sprague-Dawley (. N-9)
, w V V
PND22
F1 Rat, Sprague-Dawley (.. ?.N=9)
. wv v
Post-Implantation Loss
Lee etal., 2015,2851075
develop mental (GD11 -16}
GD17
P0 Mouse, CD-1 (•-, N=10)
A A A
Fuentes etal.. 2006, 757859
developmental (GD6-18)
GD18
Mouse, CD-1 {-r. N=10-11)
Resorptions, Any
Argus, 2OD0, 5080012
developmental (6D7-2Q)
GD2P
P0 Rabbit, New Zealand (N=12-20)
Resorptions, Early
Argus, 2000, 5080012
developmental (GD7-20)
GD29
PO Rabbit, New Zealand (N=12-20)
Fuentes etal.. 2006, 757859
developmental (GD&-1U)
GD18
Mouse, CD-1 (y , N=10-11>
Resorptions, Late
Argus, 2000, 5080012
developmental (GD7-20)
GD29
PO Rabbit. New Zealand ( \ N=12-20)
Fuentes et al., 2006, 757859
developmental (GD6-18)
GD18
Mouse, CD-I (y . N=10-11)
Resorptions, Mean.titter
Luebker etal., 2005, 757857
reproductive (7Sd (42d pra-cohabitalion, 14d mating, GDQ-20))
GD21
P0 Rat. Crl:Cd(Sd)lgs Vaf/Plus N=8)
•W
Resorptions, PercBnULittsr
Argus, 2000, 5080012
developmental (GD7-20)
GD29
P0 Rabbil. New Zealand N=12-20)
RBsorptions, Total
Conley et al., 2022,1017B381
developmental (G014-1B)
GDIS
PO Rat, Sprague-Dawley (" , N=4-6)
Stillborn Pups
LuBbkerat al., 2005,1276160
reproductive (42d prior mallng-LD20)
PND1
F1 Rat, CrS:Cd (Sd)lgs Br Vaf (£• T, N=2D-25)
Total Litter Resorbed
Argus, 2000, 5080012
developmental (GD7-20)
GD29
PO Rabiiit. New Zealand i;-J, N=12-20)
Viability Index
l uBbkerst al.. 2005,1276100
reproductive (42d prior mating-I.D20)
PND1-4
F1 Rat, Cri:Cd (Sd)lgs Br Vaf | , N=156-346)
V V
0.01
0.1 1 10 100
Concentration (mg/kgrday)
Figure 3-67. Mortality and Viability in Mice, Rats, and Rabbits Following Exposure to
PFOS (Logarithmic Scale)
PFQS concentration is presented in logarithmic scale to optimize the spatial presentation of data.
Interactive figure and additional study details available on IiAWC.
GD = gestation day; PND = postnatal day; LD = lactational day; Po = parental generation; Fi = first generation; d = day.
3.4.4.2.3 Skeletal, Soft Tissue, and Gross Effects
Skeletal defects in offspring, including bone ossification, have been observed in mice, rats, and
rabbits gestationally exposed to PFOS. 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
3-241
-------
APRIL 2024
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, asymmetric 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 versus control
(4.82 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 1 mg/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 versus
control (38.9%), 15.8%, 6.2%, 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) versus 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 versus 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
versus 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 versus
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 66.1% 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).
3.4.4.2.4 Fetal or Pup Body Weight
Several studies in different species reported data on fetal body weight (Figure 3-68). 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
3-242
-------
APRIL 2024
the 3 mg/kg/day group compared with 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 statistical
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. 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 3-68). 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 in
the one-generation study {Luebker, 2005, 757857} where the LOAEL was the lowest dose tested
of 0.4 mg/kg/day for decreased pup body weight, decreased maternal body weight, and decreased
gestation length. In both the one- and two-generation studies, the decreased pup body weight was
observed across multiple time points (PND 0 and LD 5 and PND 1, 4, 7, 14, and 21,
respectively) in the first generation. In the second generation, decreased pup weight was only
observed in the highest dose group tested (0.4 mg/kg/day) on PND 7 and 14 {Luebker, 2005,
1276160}. 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. In contrast, 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.
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
3-243
-------
APRIL 2024
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.
End point
Fetal Body Weight
Study Name Study
Argus, 2000.5080012 developmental (GD7-20)
Log at al.. 2015, 2851075 developmental (GD11-16)
Wan et al., 2020. 7174720 developmental (GD4.5-17.5)
Li al al., 2021, 9959491 developmental (GD1-17)
Fuenlas al al., 2006, 757B59 developmental (GDB-10)
Conley el al., 2022,10176381 developmental (GDH-18)
Li al al„ 2Q1I5,3981495 developmental (GD12-18)
Observation Time
GD29
G017
GD17.5
GD1B
GD1B
GD18
GD18
Pup Body Weight
Luebker et al., 2005. 757857 reproductive (76d (42d pre-coftabitation. 14d mating. GD0-20))
Zhong et al.. 2016,3748828 developmental (GD1-17)
GD21
PNW4
Lau at al.. 2003. 757854 developmental (GD1-17)
developmental (GD2-21)
Xia et al.. 2011,2919267 developmental (GD2-21)
Zhang et al. 2021. 6988534 developmental (GD12-18)
Butenhoff et al.. 2009.757873 developmental (GD0-PND2O)
Pup Body Weighl Relative lo Lilter LuebKer el al., 2005, 757857 reproduclive (42d prior maling-LD4)
Luebker et al.. 2005. 1276160 reproductive (42d prior mating-LD20)
iBpiaduclive (GD0-PND21)
PND0
PND21
PND35
PNDO
PND21
PND35
PNDO
PND1
PND3
PND7
PND14
PND1
LD5
Animal Description
F1 Rabbit New Zealand (N-12-20)
F1 Mouse, CD-1 {,<¦-. N=10)
F1 Mouse, CD-1 {<.?¦-, N=8)
F1 Mouse, Nol Specified N=3>
Mouse, CD-I (J, N=10-11)
P0 Rat. Sprague-Dawley (i, N=4-6)
F1 Ral, Sprague-Dawley (o2, N=10)
F1 Rat, Sprague-Dawley (+. N=10)
F1 Rat, Sprague-Dawley ( J. N=10)
F1 Rat. Crl:Cd(Sd)lgs Vaf.'Plus < N=8>
F1 Mouse. C57BL/6 ( ?. N=12)
F1 Mouse, C57BL/6 (y, N-12)
F1 Mouse. CD-1 N-20)
F1 Mouse. CD-1 fy-', N=20)
F1 Mouse, CD-1 N=20)
F1 Ral, Spiague-Dawlay N=5-8)
F1 Ral, Sprague-Dawley (;-¦ N=B)
F1 Rat. Sprague-Dawley ( N=8)
F1 Rat, Sprague-Dawley (:'J. N=10)
F1 Rat, Sprague-Dawley (N=8)
F1 Rat, Sprague-Dawley ( > N=8)
F1 Rat, Sprague-Dawley (? •. N=8)
F1 Rat, Sprague-Dawley ( N=6)
PND1
PND4 (preculling)
PND4 (postculling)
PND7
PND14
PND21
PND1
PND4 (preculling)
PND4 (postwilling)
PND7
PND14
PND21
F1 Rat, Crl
F1 Rat, Crl
F1 Ral, Crl
F1 Rat, Crl
F1 Ral, Crl
F1 Rat. Crl
F1 Rat, Crl
F1 Rat. Crl
F1 Rat, Crl
F1 Rat, Crl
F1 Rat. Crl
F1 Rat, Crl
F2 Rf
F2Ri
F2 Rat, Crl
F2 Rat, Crl
F2 Rat, Crl
F2 Rat, Crl
CD(SD) N=20)
CD(SD) (-'. N=20)
CD(SD) ( - , N=20)
CD(SD) (i. N=20)
Cd (Sri)lgs VbRPIus (;
Cd(Sd)lgs Vaf.'Plus (.
Cd (Sd)lgsBrVaf ( ';
Cd (Sd)lgsBrVaf ( '.
Cd (Sd)lgs BrVaf
Cd (Sd)lgsBrVaf ("
Cd (Sd)lgsBrVaf
Cd (Sd)lgs Br Vaf (.;
Cd (Sd)lgsBrVal (.;
Cd (Sd)lgs BrVal (o
Cd (Sd)lgs Br Val
Cd (Sd)lgs Br Vat :
Cd (Sd)lgs Br Vaf (
Cd (Sd)lgs Br Vaf (t*=
N=17)
N=17)
N=20-2£
N=20-25
N=20-25
N=20-25
N=20-25
N=20-25
N=22-25
N=22
N=22
N=22
PFOS Developmental Effects - Offspring Weight
> No significant change A, Significant increase~^T"slgnificant decrease |
-¥¦ W
VV V
V VW.V
-w
-JF
Figure 3-68. Offspring Body 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
in controls). No significant difference in placental weights was detected in dams exposed to
5 mg/kg/day PFOS relative to control. 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
3-244
-------
APRIL 2024
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 with controls, though the biological
significance of that effect is unclear. Wan et al. {, 2020, 7174720} found no effects 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. 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. 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.
Developmental PFOS exposure also had adverse effects on lung development, further described
in the Respiratory Section of Appendix C {U.S. EPA, 2024, 11414344}.
3.4.43 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 34 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 by mechanistic data category (see Appendix A, {U.S. EPA, 2024,
11414344}) and source is shown in Figure 3-69.
3-245
-------
APRIL 2024
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
8
0
15
20
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 SpecifiedyReview Article
1
0
0
1
Grand Total
14
7
16
34
Figure 3-69. Summary of Mechanistic Studies of PFOS and Developmental Effects
Interactive figure and additional study details available on HAWC.
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
3-246
-------
APRIL 2024
Dang et al. {, 2018, 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 [xM 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
3-247
-------
APRIL 2024
several epigenetic factors (including DNA methylation, histone deacetylation, and histone
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 programming 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
3-248
-------
APRIL 2024
larvae. This appeared to be mediated through the growth hormone/insulin-like growth factor
(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 corticosterone 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 [xM 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
3-249
-------
APRIL 2024
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-£elik, 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-£elik, 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 postnatal 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,
3-250
-------
APRIL 2024
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 (IDso), defined as the concentration at which PFOS inhibited
the development of contracting cardiomyocytes by 50%, ranged from 40 [xM {Zhang, 2016,
3981565} to 73 [xM {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 [xM 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 [xM) 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 [xg/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 with 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 [xM 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,
3-251
-------
APRIL 2024
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
3-252
-------
APRIL 2024
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 Leydig 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 Leydig cells in vitro but did not affect cell proliferation. In vivo, PFOS exposure
reduced serum testosterone levels, and reduced sperm production. LHCGR, CYP11A1, and
CYP17A1 levels in Leydig 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 upregulated 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,
3-253
-------
APRIL 2024
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 epidemiological literature reviewed in the 2016 PFOS HESD and the updated
literature searches. As noted in the epidemiological fetal growth restriction summary, there is
robust 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} (see Appendix A, {U.S. EPA, 2024,
11414344}). Comparing the postnatal growth results in infants with birth-related measures is
challenging due to complex growth dynamics including rapid growth catchup periods for those
with fetal restriction. Nonetheless, the evidence for postnatal weight deficits was comparable to
that seen for BWT.
The consistent and strong evidence for decreases in birth weight in PFOS-exposed population is
further supported by coherent evidence for other developmental effects. There is evidence of an
impact of PFOS exposure on gestational duration measures (i.e., either preterm birth or
gestational age measures) with most of the studies showing increased risk of gestational duration
deficits. 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 increased risk), 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.
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 for
studies of associations between PFOS exposure and 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. Several meta-analyses also show
associations between PFOS and preterm birth {Deji, 2021, 7564388; Gao, 2021, 99596011;
Yang, 2022, 10176603} (see Appendix A, {U.S. EPA, 2024, 11414344}).
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, and Section 5.1.1
3-254
-------
APRIL 2024
provides a further discussion on considerations for potential confounding by co-occurring PFAS.
Very few of the existing studies performed multipollutant modeling in comparison with single
pollutant estimates of PFOS associations. For studies that provided this comparison, the results
were often mixed, 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.
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 in
rodents 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 (e.g., ossification), and developmental delays
(e.g., 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 all metrics of fetal or pup viability).
Reductions in litter size or fetal 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 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. It is unclear at
this time whether one effect should be considered a precursor for the other.
The mechanistic data are primarily focused on gene expression changes and epigenetic
alterations related to exposure to PFOS during developmental stages. The PFOS-induced
alterations to the expression of genes related to growth and development support the
3-255
-------
APRIL 2024
observations of developmental effects in animals and humans (e.g., fetal growth restriction).
Molecular alterations (primarily epigenetic alterations) measured in human cord blood were
related to PFOS levels in the same biological samples. Specifically, global DNA
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). The authors of these studies of epigenetic
alterations did not measure gene expression changes to confirm that the epigenetic alterations
indeed affected gene expression, nor were adverse postnatal outcomes measured in the same
study. 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 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.
3.4.4.4.1 Evidence 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 3-17). 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 measures
of decreased gestational duration and other 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.
3-256
-------
APRIL 2024
Table 3-17. Evidence Profile Table for PFOS Exposure and Developmental 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 3.4.4.1)
Fetal growth
restriction
21 High confidence
studies
26 Medium confidence
studies
11 Low confidence
studies
2 Mixed confidence
studies
Deficits in mean birth
weight were observed in
most studies (27/39) in
the overall population.
Studies on changes in
standardized birth weight
measures reported some
inverse associations
(12/18) in the overall
population or among
boys or girls. Ten of 17
studies observed
increased risk of low
birth weight or SGA.
Deficits in birth weight-
related measures were
supported by decreases
in related FGR outcomes
such as birth length
(15/28) and head
circumference (12/23).
• High and medium
confidence studies
• Coherence across
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
Gestational duration
10 High confidence
studies
11 Medium confidence
studies
7 Low confidence
studies
Some inverse
associations with
gestational age measures
were observed in high or
medium confidence
studies in the overall
population (10/18).
Increased risk of preterm
birth was also observed
in high or medium
confidence studies
High and medium
confidence studies
Consistency in the
magnitude of
gestational age
deficits
• Limited evidence of
exposure-response
relationships in studies
examining preterm
birth
®©o
Moderate
Evidence for
developmental effects is
based on consistent
adverse effects for FGR
including birthweight
measures which are the
most accurate endpoint.
Inverse associations
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 arises
due to the potential
impact of hemodynamics
in later pregnancy due to
use of biomonitoring
samples from the second
and third trimester or
postpartum. However,
©0O
' 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). Further support is
provided by consistent
inverse associations with
gestational age measures in
high or medium confidence
epidemiological studies in
the overall population and
consistent findings of dose-
dependent decreases in fetal
weight in animal models
gestationally exposed to
PFOS.
Human relevance and other
inferences:
The available mechanistic
information provides
support for the biological
plausibility of the
3-257
-------
APRIL 2024
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
(12/17).
Fetal loss
Increased risk of fetal
• High and medium • No factors noted
3 High confidence
loss was observed (4/7)
confidence studies
studies
although results were
• Good sensitivity
3 Medium confidence
mostly nonsignificant.
across all studies
studies
One high confidence
• Consistent
1 Low confidence study
study observed a
magnitude of effect
significant increase in
• Exposure-response
risk for miscarriage for
relationship
some quintiles of
exposure in subgroup
analyses. One medium
confidence study
reported an inverse
association.
several studies present
associations for samples
collected pre-pregnancy
or in the first trimester.
phenotypic effects observed
in exposed animals in
support of the human
relevance of the animal
findings.
Postnatal growth
4 High confidence
studies
7 Medium confidence
studies
3 Low confidence
studies
Most studies (8/10)
reported an inverse
association for infant
weight or BMI changes.
There was some
evidence of an exposure-
response relationship in
two studies (2/4)
reporting categorical
exposures. Decreases in
infant height were mixed
(2/4). Inverse
associations were
observed for infant
weight in most medium
and high confidence
studies (6/10), while two
studies observed positive
associations (2/10). In
• High and medium
confidence studies
• Exposure-
response relationship
• Good or adequate
sensitivity for most
studies
• Inconsistent timing of
follow-up evaluation
3-258
-------
APRIL 2024
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
medium and high
confidence studies,
inverse associations with
infant BMI or adiposity
were observed in some
studies (4/9), but other
studies reported positive
associations (1/8) or
mixed associations by
sex and timepoint (2/8).
Birth defects
One low confidence • Medium confidence
• Low confidence
4 Medium confidence
study (1/2) observed a studies
studies
studies
small increased risk for
• Lmprecision of some
3 Low confidence
total or combined birth
positive associations
studies
defects. One medium
may suggest statistical
confidence study
power was limited
reported increased risk
• Limited number of
for septal defects,
studies examining
conotruncal defects, and
individual defects
total congenital heart
defects, but results were
imprecise.
Cryptorchidism was
examined in three
studies. Of the two
medium confidence
studies, one reported a
nonsignificant inverse
association and the other
reported a null
association.
3-259
-------
APRIL 2024
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 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
• Exposure-response
relationship
• Inconsistent direction
of effects across
species
Offspring body weight
15 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.
• Medium confidence
studies
• Dose-dependent
response
• Inconsistent direction
of effects across
species for postnatal
body weight
Offspring mortality
11 Medium confidence
studies
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
to mating through
lactation, and the study
with higher doses
• Medium confidence
studies
• Consistent direction
of effects
• Dose-dependent
response
• No factors noted
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
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.
3-260
-------
APRIL 2024
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
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.
Placental effects Decreased placental
6 Medium confidence weight (2/3), decreased
studies placental diameter (1/1)
and decreased placental
capacity (1/1) were
• Medium confidence
studies
• Dose-response
relationship
• Inconsistent direction
of effects
• Limited number of
studies examining
outcomes
3-261
-------
APRIL 2024
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
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.
• Coherence of
findings
Structural
abnormalities
2 Medium confidence
studies
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).
• Medium confidence
studies
• Limited number of
studies examining
outcomes
Developmental timing
and organ maturation
4 Medium confidence
studies
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
• Medium confidence
studies
• Coherence of effects
with other
developmental
delays
• Limited number of
studies examining
outcomes
3-262
-------
APRIL 2024
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
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/lX
Mechanistic Evidence and Supplemental Information (Section 3.4.4.3)
Summary of Key Findings, Interpretation, and Limitations ^'judgment ^^
Key findings and interpretation: The evidence
• Evidence from zebrafish embryo assays demonstrate that PFOS exposure can lead to embryonic demonstrates that PFOS
and/or larval malformation and delays/reduction in hatching. exposure during
• Alterations to the expression of genes related to growth and development in vivo in zebrafish and development can alter
rodents, and in human embryonic cell lines. the epigenome and the
• Alterations to DNA methylation (global hypomethylation and gene-specific hypermethylation) in expression of genes that
human cord blood and in placenta from rodent studies. control regular growth
Limitations: and development; it is
• The role of epigenetic mechanisms in changes at the mRNA level is not clear, nor is the relationship possible that such
between molecular changes and apical developmental outcomes. changes are related,
3-263
-------
APRIL 2024
Evidence Stream Summary and Interpretation
Evidence Integration
Summary Judgment
Studies and
Interpretation
Summary and Key Factors that Increase Factors that Decrease
Findings Certainty Certainty
Evidence Stream
Judgment
although the relationship
has not been directly
measured.
Notes: SGA = small-for-gestational age; FGR = fetal growth restriction; PND = postnatal day; GD = gestational day; BMI = body mass index; DNA = deoxyribonucleic acid;
mRNA = messenger ribonucleic acid.
3-264
-------
APRIL 2024
3.4.5 Evidence Synthesis and Integration for Other Noncancer
Health Outcomes
Consistent with the SAB's recommendation {U.S. EPA, 2022, 10476098}, EPA concluded that
the noncancer 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 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 {U.S. EPA, 2021, 10428576}. Based on this
analysis, these outcomes were not prioritized for the subsequent literature search update efforts;
the evidence synthesis and integration for these outcomes are presented in Appendix C {U.S.
EPA, 2024, 11414344}. In addition, Section 5.5 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 17 epidemiological and 1 animal toxicological study (2 overlapping publications)
that investigated the association between PFOS and cancer. Of the epidemiological studies, eight
were classified as medium confidence, seven as low confidence, and two were 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 and Summary of Evidence from the 2016 PFOS
HESD
There are eight epidemiological studies (nine publications15) from the 2016 PFOS HESD {U.S.
EPA, 2016, 3603365} that investigated the association between PFOS and cancer effects. Study
quality evaluations for these seven studies are shown in Figure 3-70.
The 2016 PFOS HESD {U.S. EPA, 2016, 3603365} concluded that there was no evidence of
carcinogenic effects for PFOS from human studies, 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 nonsignificant incidence ratios that were 1.7- to twofold higher
15 Ghisari, 2014,2920449 analyzes interactions between gene polymorphisms and PFOS exposure on breast cancer risk in the
same population analyzed in Bonefeld-fcrgensen, 2011, 2150988.
3-265
-------
APRIL 2024
among exposed workers {Alexander, 2007, 4727072}. Mean PFOS serum levels were
94.1 ng/mL. In the same 3M cohort, Grice et al. {, 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
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 risk was observed for bladder, liver, or pancreatic cancer 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}. No elevated risk of colorectal cancer was observed in community
participants of the C8 Health Project {Innes, 2014, 2850898}. Elevated nonsignificant 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.
3-266
-------
APRIL 2024
I I 1 I 1 1 I 1 L
Alexander and Olsen, 2007, 4727072 -
+
+
"
+
+
-
"
Alexander et al., 2003, 1291101 -
Bonefeld-Jorgensen etal., 2011, 2150988-
Bonefeld-Jorgensen et al., 2014, 2851186 -
Eriksen et al., 2009, 2919344 -
Ghisari et al., 2014, 2920449 -
Grice et al., 2007, 4930271 -
Hardell etal., 2014, 2968084-
"
+
+
"
+
+
"
+
+
+
+
+
+
+
+
+
+
-
+
-
+
-
++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
"
+ +
+
+
+
"
Innes et al., 2014, 2850898 -
+
"
+
+
+
+
+
"
~
+
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)
Figure 3-70. Summary of Study Quality Evaluation Results for Epidemiology Studies of
PFOS Exposure and Cancer Effects Published Before 2016 (References from 2016 PFOS
HESD)
Interactive figure and additional study details available on HAWC.
Since publication of the 2016 PFOS HESD {U.S. EPA, 2016, 3603365}, 17 studies have been
published that investigated the association between PFOS and cancer (see Appendix D, {U.S.
EPA, 2024, 11414344}). 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}, six case-control studies {Wielsoe,
2017, 3858479; Tsai, 2020, 6833693; Lin, 2020, 6835434; Itoh, 2021, 9959632; Liu, 2021,
10176563; Cao, 2022, 10412870}, six nested case-control studies {Ghisari, 2017, 3860243;
Hurley, 2018, 5080646; Cohn, 2020, 5412451; Mancini, 2020, 5381529; Shearer, 2021,
7161466; Goodrich, 2022, 10369722}, 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; Cao, 2022, 10412870}, 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;
3-267
-------
APRIL 2024
Christensen, 2016, 3858533; Ducatman, 2015, 3859843; Shearer, 2021, 7161466; Hurley, 2018,
5080646; Cohn, 2020, 5412451; Omoike, 2021, 7021502; Goodrich, 2022, 10369722}. All the
studies measured PFOS in 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}, liver {Cao, 2022, 10412870; Goodrich, 2022, 10369722}, 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
Study quality evaluations for the 17 studies identified since the 2016 PFOS HESD are shown in
Figure 3-71. Of these 17 studies, eight were considered medium confidence and seven were low
confidence {Christensen, 2016, 3858533; Itoh, 2021, 9959632; Lin, 2020, 6835434; Liu, 2021,
10176563; Omoike, 2021, 7021502; Tsai, 2020, 6833693; Cao, 2022, 10412870}. 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 are not further considered in this review. Concerns in 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 and other lifestyle
factors, which could be associated with both exposure and cancer outcome. Although PFOS has
a relatively 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
was noted to be a concern 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 small sample sizes. Lack of details or reporting issues were also
a concern for some low confidence studies which resulted in difficulty in quantitatively
interpreting analysis results {Cao, 2022, 10412870}. Cao et al. {, 2022, 10412870} was
determined to have mixed confidence (low and uninformative). The uninformative metric was the
liver cancer biomarker analysis included in this study which did not provide sufficient
information on biomarker measurement methods {Cao, 2022, 10412870}. The biomarker
analysis portion of this study is not further considered in this review.
3-268
-------
APRIL 2024
><$¦
,®x ^
iw
0^e
-------
APRIL 2024
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 1 week after diagnosis.
However, the study did not observe an increased risk of germ cell tumors association with a per
ng/mL increase in blood PFOS. One low confidence study examined liver cancers in children
and adults {Cao, 2022, 10412870}, but since results are not presented separately by age group,
this study will be reviewed in the following section.
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
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 with a
per doubling change in PFOS after adjusting for other PFAS.
Seven general population studies published since the 2016 PFOS HESD, 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, lifestyle factors and
exposure to 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
decrease in the daughters' breast cancer 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
3-270
-------
APRIL 2024
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
with 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
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 with 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 Appendix D, {U.S. EPA,
2024, 11414344}). 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).
Two general population studies published since the 2016 PFOS HESD examined liver cancer
{Cao, 2022, 10412870; Goodrich, 2022, 10369722}. One study was considered medium
confidence {Goodrich, 2022, 10369722} and one study was considered low confidence {Cao,
2022, 10412870}. The medium confidence nested case-control study of U.S. adults observed a
significant increase in risk of liver cancer when comparing participants with PFOS exposures
above the 85th percentile (54.9 ng/mL) compared with those at or below (OR = 4.50, 95% CI:
1.20, 16.00) {Goodrich, 2022, 10369722}. The association remained elevated but not statistically
significant in analyses of continuous PFOS exposure. The study was nested in the large
Multiethnic Cohort study of California and Hawaii; however, the sample size was small (n = 50
cases and controls each) which likely limited study sensitivity. A significantly elevated risk of
liver cancer was also observed in a low confidence case-control study of Chinese children and
adults (OR per log-ng/mL increase in PFOS exposure = 2.609; 95% CI: 1.179, 4.029) {Cao,
2022, 10412870}. However, confidence in the study results was considered low due to limited or
lacking information regarding selection of controls, diagnosis method for liver cancer,
adjustment for potential confounding, and details on the statistical analysis.
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
3-271
-------
APRIL 2024
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 Appendix D, {U.S. EPA, 2024, 11414344}). 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 excrete PFAS could change when the thyroid
becomes cancerous by causing abnormal thyroid hormone levels which can affect the glomerular
filtration rate {Dzierlenga, 2020, 6833691}, thereby changing the PFAS concentrations.
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 [j,g/L {Christensen, 2016, 3858533}. Christensen
et al. {, 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.
3.5.2 Animal Evidence Study Quality Evaluation and Synthesis
There is one study (2 overlapping publications) 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 3-72.
3-272
-------
APRIL 2024
,Ge
fSPP*"^"'c^g^ e^°s ^VQO^"^' &#
Butenhoff et al., 2012, 1276144-
++ ++
NR
++
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)
NR
Not reported
Figure 3-72. Summary of Study Quality Evaluation Results for Animal Toxicological
Studies of PFOS Exposure and Cancer Effects
Interactive figure and additional study details available on HAWC.
A single chronic cancer bioassay in animals was previously identified for PFOS {Thomford,
2002, 5432392; Butenhoff, 2012, 1276144}. In this study, conducted by Thomford {, 2002,
5432392} and published in part by Butenhoff et al. {, 2012, 1276144}, male and female
Crl:CD®(SD)IGS BR rats were administered diets containing 0, 0.5, 2, 5, or 20 ppm PFOS for
103-104 weeks. Increased incidence of hepatocellular adenomas in the high-dose groups for
male (7/43; 16%) and female rats (5/31; 16%) and combined adenomas/carcinomas in high-dose
group females (6/32; 19%) were observed (Table 3-18). There was also a statistically significant
positive trend of each of these responses in both male and female rats (all p < 0.01). At
105 weeks there was an accompanying increase in eosinophilic clear cell foci, and cystic
hepatocellular degeneration in males given 2, 5, and 20 ppm PFOS. Low levels of single cell
necrosis in all dose groups for both males and females were identified, though the increase
compared with controls was significant only at the highest dose in each sex.
Table 3-18. 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
0/28 (0%)**
1/29 (3%)
1/16 (6%)
1/31 (3%)
6/32 (19%)*
Hepatocellular
Adenomas and
Carcinomas
3-273
-------
APRIL 2024
Treatment group
Sex Tumor Type
0 ppm 0.5 ppm 2 ppm 5 ppm 20 ppm
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'3
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%)
8/46 (17%)
9/44 (20%)
Notes: *Statistically significant compared with the control group at p < 0.05. **Statistically significant compared with 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.
b Statistical significance determined by EPA using the Cochran-Armitage test.
In addition to hepatocellular tumors, Thomford {, 2002, 5029075} reported increased incidences
of pancreatic islet cell carcinomas in male rats (Table 3-18). Though the 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 (p < 0.05; Cochran-Armitage test).
Thyroid and mammary gland tumors were also observed but did not exhibit linear dose-response
relationships {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 27 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 by data source is shown in Figure 3-73.
Evidence Stream
Animal Human In Vitro Grand Total
Figure 3-73. Summary of Mechanistic Studies of PFOS and Cancer Effects
Interactive figure and additional study details available on HAWC..
In 2016, 10 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
3-274
-------
APRIL 2024
"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
10 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 7 of the 10 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
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-characterized 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 Gene Mutation
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 3-19).
3.5.3.1.1.2 In Vitro Evidence
Several studies have demonstrated that PFOS is not mutagenic in vitro (Table 3-20). 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 redBA/gam loci, a region
often used to determine point mutations and deletions.
3-275
-------
APRIL 2024
3.5.3.1.2 DNA Damage
3.5.3.1.2.1 In Vivo Evidence
3.5.3.1.2.1.1 Human Studies
One study reported on the genotoxic potential of PFOS exposure in humans (Table 3-21).
Governini et al. {, 2015, 3981589} collected semen samples from healthy nonsmoking men and
evaluated aneuploidy, diploidy, and DNA fragmentation. The occurrence of aneuploidy and
diploidy in sperm cells, which are normally haploid, was significantly higher in the PFAS-
positive samples (PFOS was detected in 25% of the samples) when compared with PFAS-
negative samples. This suggests that PFAS exposure is related to errors in cell division leading to
aneugenicity. Additionally, fragmented chromatin levels were also significantly increased for the
PFAS-positive group compared with the PFAS-negative group.
3.5.3.1.2.1.2 Animal Toxicological Studies
Evaluations of PFOS exposure in rat, mouse, and zebrafish models were identified, which
predominantly demonstrated evidence of genotoxicity (Table 3-21). 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. Quantifying micronuclei
formation in rats via optimal and reliable methods has been previously described {Witt, 2000,
783839; WHO, 2020, 11347555; WHO, 2009, 10455024}.
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 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 with 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 {Qelik, 2013, 2919161; Eke, 2016,
2850124; Eke, 2017, 3981318}. 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, similar to the results from NTP {, 2019, 5400978}, the study also
demonstrated that PFOS exposure decreased the ratio of PCEs to normochromic erythrocytes
(NCEs), indicating 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.
3-276
-------
APRIL 2024
Notably, these two studies did not report the ratio of PCEs to NCEs which limits the ability to
interpret these data further. Given 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.
Micronucleus frequency was slightly elevated 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}. Similarly, EPA's 2016 PFOS HESD {U.S.
EPA, 2016, 3603365} reported mouse bone marrow micronucleus assays to be negative after
high-dose acute exposures (237.5, 475, and 950 mg/kg; measured after approximately 24, 48, and
72 hours) to PFOS {Murli, 1996, 10228098}. 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}.
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.
3.5.3.1.2.2 In Vitro Evidence
3.5.3.1.2.2.1 Chromosomal aberrations
EPA's 2016 PFOS HESD {U.S. EPA, 2016, 3603365} reports that PFOS exposure did not
induce chromosomal aberrations in human lymphocytes (Table 3-22) {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 [j,g/mL to 4,000 [j,g/mL 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 [j,g/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 HAWC visual 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
3-22). 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 [xM PFOS (no statistical analysis was reported). Direct exposure of suspended calf thymus
3-277
-------
APRIL 2024
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 [j,g/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 [xM for 24 hours
{Kawamoto, 2010, 1274162}.
Florentin et al. {, 2011, 2919235} exposed HepG2 cells to PFOS (5-300 [xM) 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 [xM. In contrast, a study conducted by Wielsoe et al. {, 2015,
2533367} exposed HepG2 cells to PFOS (2 x icr7 to 2 x icr5 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.
3-278
-------
APRIL 2024
Table 3-19. Mutagenicity Data From In Vivo Studies
Reference
Species, Strain
(Sex)
Tissue
Results
PFOS Concentration
(Dosing Regimen)
Wang et al. {, 2015,
Mouse, Gpt delta
Liver
Negative
1-10 mg/kg/day
2850220}
transgenic
(daily via gavage for 28 days)
(Male)
3-279
-------
APRIL 2024
Table 3-20. 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 (ig/plate
Litton Bionetics, Inc. {, 1979,
10228135}
Saccharomyces cerevisiae (D4)
Not Reported
Negative
0.1-1,000 (ig/plate
Mecchi {, 1999, 10228133}
Salmonella typhimurium (TA98, TA100, TA1535,
TA1537)
Negative
Negative
0.333-5,000 (ig/plate
Mecchi {, 1999, 10228133}
Escherichia coli (WP2wrA)
Negative
Negative
33.3-5,000 (ig/plate
NTP {, 2019, 5400978}
Salmonella typhimurium (TA98, TA100)
Negative
Negative
100-5,000 (ig/plate
NTP {, 2019, 5400978}
Escherichia coli (WP2MvrA/pkM101)
Negative
Negative
100-10,000 (ig/plate
Simmon {, 1978, 10228131}
Salmonella typhimurium (TA1535, TA1537,
TA1538, TA98, TA100)
Negative
Negative
10-5,000 (ig/plate
Simmon {, 1978, 10228131}
Salmonella cerevisiae (D3)
Negative
Negative
0.1-5 (ig/plate
Wang et al. {, 2015, 2850220}
gpt Delta transgenic mouse embryonic fibroblasts
Not reported
Positive3
1-20 (iM
(24 hours)
Notes:
a Mutagens were present in cells exposed >10 jxM.
3-280
-------
APRIL 2024
Table 3-21. DNA Damage Data From In Vivo Studies
Reference
Species, Strain
(Sex)
Tissue
Results
PFOS Concentration
(Dosing Regimen)
DNA Strand Breakage
Governini et al. {, 2015,
3981589}
Human
(Male)
Semen
Positive
Average Seminal Plasma Concentration of
5.37 ng/gf.w.
DNA Damage via Comet Assay
Cclik et al. {, 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 et al. {,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 et al. {,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
Cclik et al. {, 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 et al. {,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 et al. {,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)
Notes: f.w. = formula weight.
" Findings based on the 2016 EPA's Health Effects Support Document for Perfluorooctane Sulfonate (PFOS) {U.S. EPA, 2016, 3603365}, concentrations)
unknown.
3-281
-------
APRIL 2024
Table 3-22. 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
10-170 (ig/mL
(3 hours)
Unscheduled DNA Synthesis
Cifone {, 1999, 10228136}
Fisher 344 male rat hepatocytes
Negative
0.25-4,000 (ig/mL
DNA Damage
Wang et al. {, 2015, 2850220}
gpt Delta transgenic mouse embryonic
fibroblasts
(Y-H2AX foci)
Positive
0-30 pM
(24 hours)
Jacquet et al. {, 2012, 2919219}
Syrian hamster embryo cells
(comet assay)
Negative
2 x 10~4-50 (ig/mL
(7 days)
Kawamoto et al. {, 2010, 1274162}
Paramecium caudatum
(comet assay)
Negative
10-100 (iM
(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 l0-7-2 x 10-5 M
(24 hours)
Florentin et al. {,2011,2919235}
HepG2
(comet assay)
Negative
5-300 pM
(1 or 24 hours)
3-282
-------
APRIL 2024
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}.
3.5.3.2.1 In Vivo Evidence
3.5.3.2.1.1 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 HAWC visual for additional supporting information and study details).
3.5.3.2.1.2 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 HAWC visual 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.
3-283
-------
APRIL 2024
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 HAWC visual for additional supporting
information and study details).
3.5.3.2.2 In Vitro Evidence
Pierozan et al. {, 2020, 6833637} evaluated PFOS (10 [xM) 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 HAWC visual 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, 2016,
3160486 J.
3.5.3.3.1 In 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 HAWC visual 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
3-284
-------
APRIL 2024
malondialdehyde (MDA) did not differ between controls and exposed animals, and that
superoxide dismutase activity was lower in treated versus 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 H.A.WC visual for additional supporting
information and study details).
3.5.3.3.2 In 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 |iinol/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
glutathione-.S'-transferase and glutathione peroxidase levels in cells. In contrast, Florentin et al. {,
2011, 2919235} exposed HepG2 cells to PFOS (5-300 pM) 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 [xM 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 with 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 HA.WC visual 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
damage. PFOS. For additional information, please see the hepatic mechanistic section (Section
3-285
-------
APRIL 2024
3.4.1.3; refer to the interactive H.A.WC visual 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
H.A.WC visual 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 H.A.WC visual 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, 3160486J.
3.5.3.6.1 In 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
3-286
-------
APRIL 2024
receptors in both rodents and humans. For additional information, please see the hepatic
mechanistic section (Section 3.4.1.3; refer to the interactive H.A.WC visual for additional
supporting information and study details).
3.5.3.6.2 In 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 (j,g/mL) for 5 and 24 hours. Evaluation of PPAR gene expression by qPCR indicated a
threefold increase of ppar-b/d mRNA level at a PFOS concentration of 0.2 [j,g/mL after 24 hours.
Subsequent exposure of SHE cells to PFOS (0.02-20 |ig/mL) for 1 week found overexpression
of PPAR-target genes and a significant increase of ppar-b/d mRNA at 0.2 [j,g/mL (twofold
increase) and 2 [j,g/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 (j,g/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 (J,g/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 #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}.
3.5.3.7.1 In 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 HA.WC visual 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, 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,
3-287
-------
APRIL 2024
2003, 1290852}. However, the same authors continued this same dietary PFOS exposure in
Sprague-Dawley rats for up to 2 years and found liver effects consistent with PPAR activation
{Thomford 2002, 5029075; Butenhoff, 2012, 1276144}. This 2-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 In 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 [xM PFOS exposure for 72 hours stimulated MCF-10A cell migration and invasion. A follow-
up study evaluating PFOS (10 [xM; 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 [xg/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 HA.WC visual for additional supporting information and study details).
3-288
-------
APRIL 2024
3.5.4 Weigh t Of E vide nee for Carcinogenicity
3.5.4.1 Summary of Evidence
The carcinogenicity of PFOS has been documented in both epidemiological and animal
toxicological studies. The available epidemiology studies report elevated risk of liver, bladder,
kidney, prostate, and breast cancers after chronic PFOS exposure in some studies, though limited
evidence for some tumor types (i.e., liver and renal) and mixed results for other tumor types
(i.e., bladder, prostate, breast) provide plausible but not definitively causal evidence of a
relationship between PFOS exposure and cancer outcomes from the epidemiological evidence
alone. The animal chronic cancer bioassay provides additional support for carcinogenicity with
the identification of multi-site tumorigenesis (liver and pancreas) in both male and female rats.
The available mechanistic data suggest that multiple MO As could play role in the hepatic and
pancreatic tumorigenesis associated with PFOS exposure based on animal model study findings.
3.5.4.1.1 Evidence From Epidemiological Studies
Results for liver cancer from one low confidence occupational {Alexander, 2003, 1291101} and
one medium confidence general population-based {Eriksen, 2009, 2919344} study of PFOS
exposure published approximately 15-20 years ago were generally imprecise (i.e., null results
with wide confidence intervals), but more recent studies have reported statistically significant
increased risk of liver cancer associated with increased PFOS exposure {Cao, 2022, 10412870;
Goodrich, 2022, 10369722}. Kmedium confidence nested case-control study of adults from the
Multiethnic Cohort (MEC) study reported a significant increased risk of liver cancer when
comparing those in the 85th percentile of PFOS exposure to those at or below the 85th percentile
{Goodrich, 2022, 10369722}. Positive, but not statistically significant, associations were
observed in analyses of continuous PFOS exposure which supported the study's overall
conclusion of an increased risk of liver cancer with increasing PFOS exposure. The study's
sensitivity was limited by the small number of cases and controls (n = 50 each). Consistent with
this finding, a Chinese general population case-control study of children and adults reported a
significant increase in risk of liver cancer in analyses of continuous PFOS exposure; however,
the study was considered low confidence due to lack of information on control selection,
outcome ascertainment, and statistical analysis {Cao, 2022, 10412870}.
Studies of the association between PFOS serum concentrations and bladder cancer have mixed
(positive and null) findings. An elevated risk of bladder cancer mortality was associated with
PFOS exposure in an occupational study {Alexander, 2003, 1291101} but a subsequent study to
ascertain cancer incidence in this cohort with four additional years of observation observed
elevated but not statistically significant incidence ratios that were 1.7- to twofold higher among
workers with higher cumulative exposure to PFOS {Alexander, 2007, 4727072}. Some of the
limitations of these studies include the lack of precision of the risk estimates due to the small
number of cases, and the lack of control for the potential confounding of smoking. A nested
case-control study in a general population Danish cohort did not observe elevated bladder cancer
risk with increasing PFOS serum levels {Eriksen, 2009, 2919344}. Overall, there is plausible
evidence of a relationship between PFOS exposure and bladder cancer, particularly for high-
exposure communities.
One study in the general population reported a statistically significant increase in risk of RCC in
the highest PFOS exposure quartile and in continuous analyses of PFOS exposure (i.e., per
3-289
-------
APRIL 2024
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
reported association when evaluated on a per doubling of PFOS after adjusting for other PFAS.
Elevated nonsignificant 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 cancers. When cumulative PFOS exposure
measures were analyzed, elevated ORs were reported for prostate cancer, however, they did not
reach statistical significance. Length of follow-up may not have been adequate to detect cancer
incidence in this cohort as approximately one-third of the participants had worked <5 years in
their jobs, and only 41.7% were employed >20 years {Grice, 2007, 4930271}. 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 anNHANES population, Omoike et al. {, 2021,
7021502} observed a significantly inverse association between PFOS exposure and prostate
cancer.
The majority of studies examining associations between PFOS exposure and cancer outcomes
were on breast cancer. One study of Inuit females in Greenland observed positive associations
between PFOS levels and risk for breast cancer {Bonefeld-J0rgensen, 2011, 2150988}, although
the association was of a low magnitude and could not be separated from the effects of other
perfluorosulfonated compound exposures (i.e., PFHxS and PFOSA). 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}
reported that increased breast cancer risk was associated with increased PFOS serum
concentrations in Danish individuals with a specific polymorphism in the CYP19 gene (for
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 increased risk of breast cancer in all women 50 years old or younger
(including ER+ and ER- participants), and in ER+ participants aged 50 years or younger.
Statistically significant increases in breast cancer risk were also observed in an NHANES
population in the two highest quartiles of exposure, but the association was inverse in the second
quartile {Omoike, 2021, 7021502}. No association was identified between PFOS and breast
cancer in either case-control or nested case-control studies of Danish and California cancer
registry populations, respectively {Bonefeld-J0rgensen, 2014, 2851186; Hurley, 2018,
5080646}. Another general population study in the United States suggested that maternal PFOS
exposure combined with high maternal cholesterol may decrease the daughters' risk of breast
cancer but did not examine breast cancer subtypes or individuals with genetic variants that may
have increased susceptibility {Cohn, 2020, 5412451}. A recent study in a Japanese population
observed an inverse association across serum PFOS quartiles with a significant dose-response
3-290
-------
APRIL 2024
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 exposure and subsequent cancer assessment is not available. Overall,
study design limitations with specific studies, lack of replication of the results, and a lack of
mechanistic understanding of specific breast cancer subtypes or susceptibilities of specific
populations limit firm conclusions regarding PFOS and breast cancer. However, there is
suggestive evidence that PFOS exposure may be associated with an increased breast cancer risk
based on studies in susceptible populations, such as those with specific polymorphisms and for
specific types of breast tumors.
3.5.4.1.2 Evidence From Animal Bioassays
One available chronic toxicity/carcinogenicity bioassay for PFOS, a 104-week dietary study in
rats, provides evidence of multi-sex and multi-site tumorigenesis resulting from PFOS exposure
{Thomford, 2002, 5029075; Butenhoff, 2012, 1276144}. This study was originally published as
a 3M-sponsored report by Thomford {, 2002, 5029075} and some of the data were later
published in a peer-reviewed study by Butenhoff et al. {, 2012, 1276144}. Statistically
significant increases in the incidence of hepatocellular adenomas in the high-dose (20 ppm) male
(7/43; 16%) and female (5/31; 16%) rat groups and combined adenomas/carcinomas in the
females (6/32; 19%; five adenomas, one carcinoma) were observed. The observation of one
carcinoma in the female rats is a relatively rare occurrence according to NTP's historical controls
for female Sprague-Dawley rats (1/639 historical control incidence) {NTP, 2020, 10368689}.
Historical control incidence rates for these tumor types were not provided by Thomford {, 2002,
5029075}. Additionally, there were statistically significant dose-related trends in the hepatic
tumor responses of both males and females. A statistically significant trend of increased
incidence of pancreatic islet cell carcinomas with increased PFOS dose was also observed in the
male rats, though the individual dose groups were not statistically different from the control
group. The percentages of animals with islet cell carcinomas in the highest dose group (12.5%)
exceeds NTP's historical controls for male Sprague-Dawley rats by over an order of magnitude
(12/638; 1.9%) {NTP, 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 {Thomford, 2002, 5029075; Butenhoff, 2012, 1276144}. 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 few follicular cell adenomas/carcinomas in
the females (4 total, excluding the recovery group) with a nonlinear dose response. Mammary
gland tumors, primarily combined fibroma adenoma and adenoma, were also observed in
females, though there was a high background incidence of mammary gland tumors in the control
animals, and the incidence lacked dose response for all tumor classifications.
3-291
-------
APRIL 2024
3.5.4.2 Mode of Action Analysis
As PFOS has been associated with multi-site tumorigenesis in both epidemiological studies and
animal toxicological studies, not always with site concordance, it is reasonable to assume that it
may act through multiple carcinogenic MOAs. In the 2016 PFOS HESD {U.S. EPA, 2016,
3603365}, EPA suggested that the induction of tumors may be related to nuclear receptor
activation, mitochondrial effects, and gap junction intercellular communication. As described in
the following subsections, the available mechanistic data continue to suggest that multiple
MOAs could play a role in the tumorigenesis associated with PFOS exposure in animal models
and human populations.
3.5.4.2.1 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}. These findings in rats are supported by recent
epidemiological studies that have reported associations between PFOS and hepatocellular
carcinoma in humans {Cao, 2022, 10412870; Goodrich, 2022, 10369722}.
The EPA previously concluded that, "the data are inadequate to support a PPARa-linked MOA
for the liver and thyroid adenomas observed by Thomford (2002)/Butenhoff et al. (2012)" {U.S.
EPA, 2016, 3603365}. As described in the subsections below, the available mechanistic data
continue to suggest that multiple MOAs may underlie the hepatocellular tumors observed after
PFOS exposure. Specifically, the available studies provide varying levels of support for the role
of several plausible MOAs: PPARa activation, CAR activation, HNF4a suppression,
cytotoxicity, genotoxicity, oxidative stress, and immunosuppression.
3.5.4.2.1.1 PPARa 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} (Table 3-23, Table 3-24). This AOP
is associated with but not necessarily causally related to nonneoplastic effects including
3-292
-------
APRIL 2024
peroxisome proliferation, 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 3-23. Evidence of Key Events Associated With the PPARa Mode of Action for
Hepatic Tumors3 in Male Sprague-Dawley Rats Exposed to PFOS
Canonical
MOA
Key Event 1:
PPARa
Activation
Key Event
2: Altered
Cell Growth
Signaling
Key Event 3a:
Increased
Hepatic Cell
Proliferation
Key Event 3b:
Inhibition of
Apoptosis
Key Event 4:
Preneoplastic
Clonal
Expansion
Outcome:
Hepatic
Tumors
Dose
(mg/kg/day)b
PPARa
Activation0
Altered Cell
Growth
Signaling
Hepatic Cell
Proliferation
Apoptosis
Preneoplastic
Clonal
Expansion
Hepatic
Tumors
0.024
- (4, 14w)
- (4w)
- (4, 14w)
-(14, 103w)
NR
- (103w)
0.098
- (4, 14w)
- (4w)
- (4, 14w)
-(14, 103w)
NR
- (103w)
0.242
- (4, 14w)
- (4w)
- (4, 14w)
-(14, 103w)
NR
- (103w)
0.312
T(4w)
NR
NR
- (4w)
NR
NR
0.625
T(4w)
NR
NR
- (4w)
NR
NR
0.984
T(4w)
-(14w)
T (4w)
T(4w)
- (14, 53w)
i (103w)
- (14, 53w)
NR
T (103w)
1
| (.Ft PND 21)
NR
NR
NR
NR
NR
1.25
T(4w)
NR
NR
- (4w)
NR
NR
1.33/1.51
- (4,14w)
NR
- (4w)
NR
NR
NR
1.66
T (28(1)
- (1, 7(1)
NR
T(7d)
-(1,28(1)
T(7d)
-(1. 28d)
NR
NR
1.93
- (7d)
NR
T(7d)
1 (7d)
NR
NR
Notes: | = statistically significant increase in response compared with controls; - = no significant response; j = statistically
significant decrease in response compared with controls; MOA = mode of action; PPARa = peroxisome proliferator-activated
receptor a; NR = not reported; d = day(s); w = week(s); Fi = first generation of offspring; PND = postnatal day.
Cells in bolded text with blue shading indicate that the response direction is concordant with the key event in the published
MOA. Cells with NR (not reported) indicate that no data were measured for that particular key event at that dose in the studies
reviewed.
Data represented in table extracted from NTP {, 2019, 5400978}; Chang et al. {, 2009, 757876}; Elcombe et al. {, 2012,
1332473}; Elcombeetal. {,2012, 1401466}; Seacat et al. {,2003, 1290852}; and Butenhoff et al. {,2012, 1276144}/Thomford
{,2002, 5029075}.
aReviewed in Klaunig et al. {, 2003, 5772415}; Corton et al. {, 2014, 2215399}; and Corton et al. {, 2018, 4862049}.
bDoses for 0.024, 0.098, 0.242, and 0.984 mg/kg/day correspond to 0.5, 2, 5, and 20 ppm in feed, respectively, in Butenhoff et al.
{, 2012, 1276144}. Dose for 1.33/1.51 mg/kg corresponds to 20 ppm in feed for animals exposed for 14 and 4 weeks,
respectively, in Seacat et al., {, 2003, 1290852}. Dose for 1.66 mg/kg corresponds to 20 ppm in feed in Elcombe et al. {, 2012,
1401466}. Dose for 1.93 mg/kg corresponds to 20 ppm in feed in Elcombe et al. {, 2012, 1332473}.
c Indirect measurement of PPARa induction provided as Cyp4al, Cyp2b2, or ACoA mRNA expression in Chang et al. {, 2009,
757876}; as hepatic palmitoyl-CoA oxidase activity in Butenhoff et al. {, 2012, 1276144}/Thomford {, 2002, 5029075}, Seacat
et al. {, 2003, 1290852}, Elcombe et al. {, 2012, 1332473}, and Elcombe et al. {, 2012, 1401466}; and as Cyp4al, Cyp2bl,
Cyp2b2, and Acoxl gene expression or hepatic acyl-CoA oxidase activity in NTP {, 2019, 5400978}.
3-293
-------
APRIL 2024
Table 3-24. Evidence of Key Events Associated With the PPARa Mode of Action for
Hepatic Tumors3 in Female Sprague-Dawley Rats Exposed to PFOS
Canonical
MOA
Key Event 1:
PPARa
Activation
Key Event 2:
Altered Cell
Growth
Signaling
Key Event 3a:
Increased
Hepatic Cell
Proliferation
Key Event 3b:
Inhibition of
Apoptosis
Key Event 4:
Preneoplastic
Clonal
Expansion
Outcome:
Hepatic
Tumors
Dose
PPARa
Altered Cell
Hepatic Cell
Apoptosis
Preneoplastic
Hepatic
(mg/kg/day)b
Activation0
Growth
Signaling
Proliferation
Clonal
Expansion
Tumors
0.029
- (4, 14w)
NR
- (4, 14w)
- (14, 103w)
NR
- (103w)
0.120
4 (4w)
- (14w)
NR
- (4, 14w)
- (14, 103w)
NR
- (103w)
0.299
- (4, 14w)
NR
- (4, 14w)
- (14, 103w)
NR
- (103w)
0.312
t (4w)
NR
NR
- (4w)
NR
NR
0.47
4 (4w)
NR
- (4w)
NR
NR
NR
0.625
T(4w)
NR
NR
- (4w)
NR
NR
1
t (Po GD 1—20)
NR
NR
NR
NR
NR
1.25
t (4w)
NR
NR
- (4w)
NR
NR
1.251
- (4, 14w)
NR
- (4, 14, 53w)
4 (103w)
- (14, 53w)
NR
t (103w)
1.56/1.77
- (4, 14w)
NR
- (4w)
NR
NR
NR
Notes: | = statistically significant increase in response compared with controls; - = no significant response; j = statistically
significant decrease in response compared with controls; MOA = mode of action; PPARa = peroxisome proliferator-activated
receptor a; NR = not reported; w = week(s); Po = parental generation; GD = gestational day.
Cells in bolded text with blue shading indicate that the response direction is concordant with the key event in the published
MOA. Cells with NR (not reported) indicate that no data were measured for that particular key event at that dose in the studies
reviewed.
Data represented in table extracted from NTP {, 2019, 5400978}; Chang et al. {, 2009, 757876}; Seacat et al., {, 2003,
1290852}; andButenhoff et al. {,2012, 1276144}/Thomford {,2002, 5029075}.
aReviewed in Klaunig et al. {, 2003, 5772415}; Corton et al. {, 2014, 2215399}; and Corton et al. {, 2018, 4862049}.
bDoses for 0.029, 0.120, 0.299, and 1.251 mg/kg/day correspond to 0.5, 2, 5, and 20 ppm in feed, respectively, in Butenhoff et al.
{, 2012, 1276144}. Dose for 0.47 corresponds to 5 ppm in feed in Seacat et al. {, 2003, 1290852}. Dose for 1.56/1.77 mg/kg
corresponds to 20 ppm in feed for animals exposed for 14 and 4 weeks, respectively, in Seacat et al. {, 2003, 1290852}.
c Indirect measurement of PPARa induction provided as Cyp4al, Cyp2b2, or ACoA mRNA expression in Chang et al. {, 2009,
757876}, as hepatic palmitoyl-CoA oxidase activity at 4 and 14 weeks in Butenhoff et al. {, 2012, 1276144}/Thomford {, 2002,
5029075}, and as Cyp4al, Cyp2bl, Cyp2b2, and Acoxl gene expression in NTP {, 2019, 5400978}.
The published in vivo and in vitro literature suggests that PFOS is a relatively weak PPARa
agonist compared with 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
3-294
-------
APRIL 2024
further details, see Section 3.4.1.3. Consistent with these findings, studies of WT 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 differentially expressed genes in WT mice) compared with responses in mice treated with
3 mg/kg/day Wyeth-14,643 (902 genes WT, 10 genes PPARa-null) or PFOA (879 genes WT,
176 genes PPARa-null) {Rosen, 2010, 1274165}, consistent with multiple MOAs for PFOS
hepatic effects.
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
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
3-295
-------
APRIL 2024
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 with 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 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 necessarily expected to induce cell proliferation or suppress
apoptosis of hepatocytes in humans {Corton, 2018, 4862049}. Specifically, some have argued
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,
3-296
-------
APRIL 2024
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}.
3.5.4.2.1.2 Other 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} (Table 3-25, Table 3-26). 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 and associative events {Dong, 2016, 3981515; NTP, 2019, 5400978;
Martin, 2007, 758419; Elcombe, 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 with PPARa, depending on the sex and/or model of interest. As with PPARa-mediated
tumorigenesis, 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
3-297
-------
APRIL 2024
understand the mechanistic underpinnings of PFOS-induced hepatic cell proliferation and
whether it is related to CAR activation.
Table 3-25. Evidence of Key Events Associated With the CAR Mode of Action for Hepatic
Tumors3 in Male Sprague-Dawley Rats Exposed to PFOS
Canonical
MOA
Key Event 1:
CAR
Activation
Key Event 2:
Altered Gene
Expression
Key Event 3:
Increased
Hepatic Cell
Proliferation
Key Event 4:
Preneoplastic
Clonal
Expansion
Outcome:
Hepatic Tumors
Dose
(mg/kg/day)b
CAR
Activation
Altered Gene
Expression
Hepatic Cell
Proliferation
Preneoplastic
Clonal
Expansion
Hepatic Tumors
0.024
NR
NR
- (4, 14w)
NR
- (103w)
0.098
NR
NR
- (4, 14w)
NR
- (103w)
0.242
NR
NR
- (4, 14w)
NR
- (103w)
0.312
NR
T(4w)
NR
NR
NR
0.625
NR
t (4 w)
NR
NR
NR
0.984
NR
NR
T(4w)
- (14, 53w)
NR
t (103w)
1
NR
| (Fi PND 21)
NR
NR
NR
Notes: | = statistically significant increase in response compared with controls; - = no significant response; MOA = mode of
action; CAR = constitutive androstane receptor; NR = not reported; w = week(s); GD = gestational day; Fi = first generation of
offspring; PND = postnatal day.
Cells in bolded text with blue shading indicate that the response direction is concordant with the key event in the published
MOA. Cells with NR (not reported) indicate that no data were measured for that particular key event at that dose in the studies
reviewed.
Data represented in table extracted from NTP {, 2019, 5400978}; Chang et al. {, 2009, 757876}; and Butenhoff et al. {, 2012,
1276144}/Thomford {, 2002, 5029075}.
a Re viewed in Felteretal. {,2018, 9642149}.
bDoses for 0.024, 0.098, 0.242, and 0.984 mg/kg/day correspond to 0.5, 2, 5, and 20 ppm in feed, respectively, in Butenhoff et al.
{,2012, 1276144}.
Table 3-26. Evidence of Key Events Associated With the CAR Mode of Action for Hepatic
Tumors3 in Female Sprague-Dawley Rats Exposed to PFOS
Canonical
MOA
Key Event 1:
CAR
Activation
Key Event 2:
Altered Gene
Expression
Key Event 3:
Increased
Hepatic Cell
Proliferation
Key Event 4:
Preneoplastic
Clonal
Expansion
Outcome:
Hepatic
Tumors
Dose
(mg/kg/day)b
CAR
Activation
Altered Gene
Expression
Hepatic Cell
Proliferation
Preneoplastic
Clonal
Expansion
Hepatic
Tumors
0.029
NR
NR
- (4, 14w)
NR
- (103w)
0.120
NR
NR
- (4, 14w)
NR
- (103w)
0.299
NR
NR
- (4, 14w)
NR
- (103w)
0.312
NR
T(4w)
NR
NR
NR
0.625
NR
t (4w)
NR
NR
NR
1.251
NR
t (Po GD 1-20)
- (4, 14, 53w)
NR
t (103w)
Notes: | = statistically significant increase in response compared with controls; - = no significant response; MOA = mode of
action; CAR = constitutive androstane receptor; NR = not reported; w = week(s); Po = parental generation; GD = gestational
day.
3-298
-------
APRIL 2024
Cells in bolded text with blue shading indicate that the response direction is concordant with the key event in the published
MOA. Cells with NR (not reported) indicate that no data were measured for that particular key event at that dose in the studies
reviewed.
Data represented in table extracted from NTP {, 2019, 5400978}; Chang et al. {, 2009, 757876}; and Butenhoff et al. {, 2012,
1276144}/Thomford {, 2002, 5029075}.
a Reviewed in Felter et al. {, 2018, 9642149}
bDoses for 0.029, 0.120, 0.299, and 1.251 mg/kg/day correspond to 0.5, 2, 5, and 20 ppm in feed, respectively, in Butenhoff et al.
{,2012, 1276144}.
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
noncancer hepatic effects including hepatomegaly, steatosis, altered lipid metabolism, and fatty
liver disease. Beggs et al. {, 2016, 3981474} exposed human primary hepatocytes to 0.01-10 [xM
PFOS and determined after 48 and 96 hours of 10 |iM PFOS, HNF4a protein expression was
significantly decreased. Beggs et al. {, 2016, 3981474} also observed a decrease in HNF4a
protein in the livers of 10-week-old CD-I mice exposed to 10 mg/kg/day PFOS once daily by
oral gavage for 7 days. 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}.
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), as well as 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.
3.5.4.2.1.3 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 corresponding dose response for cytotoxicity and formation of tumors; and
6) reversibility upon cessation of exposure (Table 3-27,Table 3-28). As discussed above in the
genotoxicity section (Section 3.5.4.2.1.4), there is no experimental support that PFOS can induce
3-299
-------
APRIL 2024
DNA damage and/or micronuclei formation in liver tissue, which supports the first key event in
the cytotoxicity MOA. 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 non-alcoholic steatohepatitis (p < 0.05), ballooning, fibrosis, and portal
inflammation.
Table 3-27. Evidence of Key Events Associated With the Cytotoxicity Mode of Action for
Hepatic Tumors3 in Male Sprague-Dawley Rats
Canonical
MOA
Key Event 1:
Cytotoxicity
Key Event 2:
Increased
Serum
Enzymes
Key Event 3:
Regenerative
Proliferation
Key Event 4:
Hyperplasia and/or
Preneoplastic
Lesions
Outcome:
Hepatic
Tumors
Dose
(mg/kg/day)b
Cytotoxicity
Serum
Enzymes
Regenerative
Proliferation
Hyperplasia and/or
Preneoplastic
Lesions
Hepatic
Tumors
0.024
0.098
0.242
0.312
-(14, 103w)
-(14, 103w)
-(14, 103w)
- (4w)
- (4, 14, 27,
53w)
- (4, 14, 27,
53w)
- (4, 14, 27,
53w)
- (4w)
- (4, 14w)
- (4, 14w)
- (4, 14w)
NR
-(14, 103w)
-(14, 103w)
-(14, 103w)
- (4w)
- (103w)
- (103w)
- (103w)
NR
0.625
- (4w)
T(4w)
NR
- (4w)
NR
0.984
t (103w)
- (4,14, 53w)
T (4, 14, 53w)
- (27w)
T(4w)
- (14, 53w)
t (1.03w)
- (14, 53w)
t (1.03w)
Notes: | = statistically significant increase in response compared with controls; - = no significant response; MOA = mode of
action; w = week(s); NR = not reported.
Cells in bolded text with blue shading indicate that the response direction is concordant with the key event in the published
MOA. Cells with NR (not reported) indicate that no data were measured for that particular key event at that dose in the studies
reviewed.
Data represented in table extracted from: NTP {, 2019, 5400978} and Butenhoff et al. {, 2012, 1276144}/Thomford {, 2002,
5029075}.
a Re viewed in Felteretal. {,2018, 9642149}.
bDoses for 0.024, 0.098, 0.242, and 0.984 mg/kg/day correspond to 0.5, 2, 5, and 20 ppm in feed, respectively, in Butenhoff et al.
{,2012, 1276144}.
Table 3-28. Evidence of Key Events Associated With the Cytotoxicity Mode of Action for
Hepatic Tumors3 in Female Sprague-Dawley Rats
Canonical
MOA
Key Event 1:
Cytotoxicity
Key Event 2:
Increased
Serum
Enzymes
Key Event 3:
Regenerative
Proliferation
Key Event 4:
Hyperplasia and/or
Preneoplastic
Lesions
Outcome:
Hepatic
Tumors
Dose
(mg/kg/day)b
Cytotoxicity
Serum
Enzymes
Regenerative
Proliferation
Hyperplasia and/or
Preneoplastic
Lesions
Hepatic
Tumors
0.029
0.120
-(14, 103w)
-(14, 103w)
- (4, 14, 27,
53w)
- (4, 14, 27,
53w)
- (4, 14w)
- (4, 14w)
-(14, 103w)
-(14, 103w)
- (103w)
- (103w)
3-300
-------
APRIL 2024
Canonical
MOA
Key Event 1:
Cytotoxicity
Key Event 2:
Increased
Serum
Enzymes
Key Event 3:
Regenerative
Proliferation
Key Event 4:
Hyperplasia and/or
Preneoplastic
Lesions
Outcome:
Hepatic
Tumors
0.299
-(14, 103w)
- (4, 14, 27,
- (4, 14w)
-(14, 103w)
- (103w)
53w)
0.312
- (4w)
- (4w)
NR
- (4w)
NR
0.625
- (4w)
- (4w)
NR
- (4w)
NR
1.251
t (103w)
- (4, 14, 27,
- (4, 14, 53w)
t (103w)
t (103w)
- (4, 14,53w)
53w)
- (14, 53w)
Notes: | = statistically significant increase in response compared with controls; - = no significant response; MOA = mode of
action; w = week(s); NR = not reported.
Cells in bolded text with blue shading indicate that the response direction is concordant with the key event in the published
MOA. Cells with NR (not reported) indicate that no data were measured for that particular key event at that dose in the studies
reviewed.
Data represented in table extracted from: NTP {, 2019, 5400978} and Butenhoff et al. {, 2012, 1276144}/Thomford {, 2002,
5029075}.
a Re viewed in Felteretal. {,2018, 9642149}.
b Doses for 0.029, 0.120, 0.299, and 1.251 mg/kg/day correspond to 0.5, 2, 5, and 20 ppm in feed, respectively, in Butenhoff et al.
{,2012, 1276144}.
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
twofold (100% change relative to control) compared with 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.
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 corresponding dose response for cytotoxicity and the formation of liver tumors as
evidence in Table 3-29 and Table 3-30, 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 3-29. 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
Adenomas
0/41**
3/42 3/47 1/44 7/43**
3-301
-------
APRIL 2024
0 mg/kg/day 0.024 mg/kg/day 0.098 mg/kg/day 0.242 mg/kg/day 0.984 mg/kg/day
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 with 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 3-30. 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
0/28**
1/29
1/16
1/31
6/32*
Hepatocellular
Adenomas &
Carcinomas
Necrosis, Individual
3/50
4/50
4/50
5/50
9/50
Hepatocyte
Infiltrate,
2/50
3/50
5/50
6/50
20/50
Macrophage,
Pigmented
Infiltrate,
33/50
37/50
33/50
36/50
42/50
Lymphohistiocytic
Hyperplasia, Bile
21/50
25/50
19/50
17/50
27/50
Duct
Notes: Statistical significance for an exposed group indicates a significant pairwise test compared with the vehicle control group.
Statistical significance for the vehicle control indicates a significant trend test.
* Statistically significant at p < 0.05; ** p < 0.01.
3.5.4.2.1.4 Genotoxicity
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 versus 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;
3-302
-------
APRIL 2024
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
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 at levels resulting in cytotoxicity
(e.g., hepatotoxicity, bone marrow toxicity) can lead to secondary genotoxicity in target tissues.
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 or key event in PFOS tumor formation.
3.5.4.2.1.5 Consideration 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 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. 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,
3-303
-------
APRIL 2024
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, 2012, 1248827; Grandjean, 2017, 3858518; Grandjean, 2017, 4239492; Mogensen,
2015, 3981889; Budtz-j0rgensen, 2018, 5083631; Timmermann, 2020, 6833710; Granum, 2013,
1937228; Stein, 2016, 3108691; Zhang, 2023, 10699594}. Additionally, one medium confidence
study reported higher odds of total infectious diseases with increasing PFOS serum
concentrations {Goudarzi, 2017, 3859808}, though it should be noted that studies reporting odds
ratios for specific 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.
3.5.4.2.2 Mode of Action for Pancreatic Tumors
Additional evidence of the carcinogenicity of PFOS comes from a high confidence chronic
rodent study identifying pancreatic islet cell carcinomas in male rats {Thomford, 2002,
5029075}. From a review of the literature, no established MOA was identified for pancreatic
islet cell carcinogenicity in animals. Considerable uncertainty remains in the underlying
mechanisms of PFOS-induced pancreatic islet tumors.
A recent review of the molecular mechanisms of pancreatic islet cell (i.e., neuroendocrine)
tumors indicates pancreatic neuroendocrine tumors primarily originate from aberrant cell
proliferation in the endocrine pancreas {Maharjan, 2021, 11248923}. However, these tumors can
also develop from pluripotent cells of the exocrine pancreas {Maharjan, 2021, 11248923}. The
3-304
-------
APRIL 2024
human islet is similar to the rodent islet, with similarities in P-cell numbers, islet cell patterns,
and blood vessel-islet structure and interactions {Bonner-Weir, 2015, 11321029}. Some
evidence suggests a role for PPARa and PPARy in rat and human pancreatic islet cell function
{Sugden, 2001, 11321031; Dubois, 2000, 11321030; Roduit, 2000, 5213235; Eibl, 2001,
11321032}, though PPARa activation has been argued to be related to pancreatic acinar cell
tumors rather than to islet cell tumors {Klaunig, 2003, 5772415}. Other studies have shown that
PFOS exposure can reduce pancreatic islet cell size and viability and can induce ROS {Qin,
2022, 10176395}.
Although an established MOA is currently unknown for this tumor type, the observation of
pancreatic islet cell tumors in rodents provides additional evidence for the carcinogenic potential
of PFOS.
3.5.4.2.3 Conclusions
Based on the weight of evidence evaluation of the available literature, PFOS has the potential to
induce hepatic tumors in humans and rodents via multiple MO As, most notably via the
modulation of nuclear receptors (i.e., PPARa and CAR) and cytotoxicity. There is also limited
evidence supporting additional potential MO As 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}. No established MOA was
identified for pancreatic islet cell carcinogenicity in rats.
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 and pancreatic 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 MO As. Therefore, EPA concludes that
the hepatic and pancreatic 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, particularly findings of increased risk of hepatocellular carcinoma, from
epidemiological studies.
Several health agencies have reviewed the available mechanistic literature and have come to
similar conclusions regarding the multiple potential MO As for PFOS-induced tumorigenesis. For
example, CalEPA's Office of Environmental Health Hazard Assessment recently concluded that
PFOS "possesses] several of the key characteristics of carcinogens, including the ability to
induce oxidative stress, inflammation, and modulate receptor-mediated effects. Additionally,
there is suggestive evidence that... PFOS [is] genotoxic, thus a genotoxic MOA for cancer
remains plausible" {CalEPA, 2021, 9416932}. Zahm et al. (2023, 3982387) also concluded that
there is moderate evidence for many potential mechanisms for PFOS-induced toxicity and
specifically noted that PFOS can induce epigenetic alterations, immunosuppression, and
oxidative stress and cause endocrine- and receptor-mediated effects.
3-305
-------
APRIL 2024
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
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 available evidence indicates that PFOS has carcinogenic potential in one animal model for
multiple sites and both sexes, as well as supporting evidence from human studies, consistent with
the examples described in the Guidelines for Carcinogen Risk Assessment for the Likely
descriptor. The epidemiological evidence of associations between PFOS and cancer found mixed
results across tumor types. However, the available study findings support a plausible correlation
between PFOS exposure and carcinogenicity in humans. The single chronic cancer bioassay
performed in rats is positive for multi-site and -sex tumorigenesis {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 male and
female rats, respectively. 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
{, 2002, 5029075}/Butenhoff et al. {, 2012, 1276144} may be relevant to humans. These
findings in rats and their potential human relevance are supported by recent epidemiological
studies that have reported associations between PFOS and hepatocellular carcinoma in humans
{Cao, 2022, 10412870; Goodrich, 2022, 10369722}.
In addition to hepatocellular tumors, Thomford {, 2002, 5029075} reported increased incidences
of pancreatic islet cell carcinomas with a statistically significant dose-dependent positive trend,
as well as modest increases in the incidence of thyroid follicular cell tumors. The findings of
multiple tumor types provide additional support for potential multi-site tumorigenesis resulting
from PFOS exposure. Importantly, site concordance is not always assumed between humans and
animal models; agents observed to produce tumors may do so at the same or different sites in
humans and animals {U.S. EPA, 2005, 6324329}. While site concordance was present between
human studies of liver cancer and animal studies reporting increased incidence of hepatocellular
tumors, evidence of carcinogenicity of PFOS from other cancer sites where concordance
3-306
-------
APRIL 2024
between humans and animals is not present is still relevant to the carcinogenicity determination
for PFOS. See Table 3-31 below for specific details on how PFOS aligns with the examples
supporting the Likely to Be Carcinogenic to Humans cancer descriptor in the Guidelines for
Carcinogen Risk Assessment {U.S. EPA, 2005, 6324329}.
Table 3-31. Comparison of the PFOS Carcinogenicity Database With the Likely Cancer
Descriptor as Outlined 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." {U.S.
EPA, 2005, 6324329}
"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}
"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." {U.S. EPA, 2005, 6324329}
"A rare animal tumor response in a single experiment
that is assumed to be relevant to humans." {U.S. EPA,
2005,6324329}
"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}
Notes: MOA = mode of action.
PFOS data are consistent with this description.
Epidemiological evidence supports a plausible
association between PFOS exposure and liver cancer
which is consistent with evidence of liver cancer in
animals. Epidemiological studies evaluating the
association between human exposure to PFOS and other
cancers are mixed. Supporting carcinogenicity data are
available from animal experiments.
PFOS data are consistent with this description. PFOS
has tested positive in animal experiments in more than
one sex and site. Hepatic tumors were observed in male
and female rats (statistically significant at high dose and
statistically significant trend tests for each) and islet cell
carcinomas show a statistically significant positive trend
in male rats.
This description is not applicable to PFOS.
PFOS data are consistent with this description. The
hepatocellular carcinoma observed in the high-dose
female rats is a rare tumor type in this strain {NTP,
2020, 7330145}.
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.
EPA recognizes that other state and international health agencies have recently classified PFOS
as either "possibly carcinogenic to humans" (IARC as reported in Zahm et al. {, 2023,
3982387}) or carcinogenic to humans {CalEPA, 2021, 9416932}. As the SAB PFAS Review
Panel {U.S. EPA, 2022} noted, "the criteria used by California EPA, for determination that a
chemical is a carcinogen, are not identical to the criteria in the U.S. EPA (2005) Guidelines for
Carcinogen Risk Assessment' and, similarly, IARC's classification criteria are not identical to
EPA's guidelines {IARC, 2019, 11320796}. Rationale for why PFOS exceeds the Suggestive
Evidence of Carcinogenic Potential descriptor and does not meet the Carcinogenic to Humans
3-307
-------
APRIL 2024
descriptor according to EPA's Guidelines for Carcinogen Risk Assessment {U.S. EPA, 2005,
6324329} is detailed in Section 5.4.
3-308
-------
APRIL 2024
4 Dose-Response Assessment
Considerations in Selecting Studies and Endpoints for Dose-Response Analysis
There is evidence from both human epidemiological and animal toxicological studies that oral
perfluorooctane sulfonic acid (PFOS) exposure can result in adverse health effects across a range
of health outcomes. In response to recommendations made by the EPA's Science Advisory
Board (SAB) and the conclusions presented in the U.S. Environmental Protection Agency's
(EPA's) preliminary analysis, the 2021 SAB review draft Proposed Approaches to the
Derivation of a Draft Maximum Contaminant Level Goal for Perfluorooctane Sulfonic Acid
(PFOS) (CASRN1763-23-1) in Drinking Water {U.S. EPA, 2021, 10428576}, EPA focused its
final toxicity value derivation efforts herein "on those health outcomes that have been concluded
to have the strongest evidence" {U.S. EPA, 2022, 10476098}. Therefore, EPA prioritized health
outcomes and endpoints with the strongest overall weight of evidence which were the health
outcomes with evidence demonstrates or evidence indicates integration judgments based on
human, animal, and mechanistic evidence (Sections 3.4 and 3.5) for points of departure (POD)
derivation using the systematic review methods described in Section 2 and Appendix A {U.S.
EPA, 2024, 11414344}. For PFOS, the health outcomes with the strongest weight of evidence
are cancer (described in Section 4.2) and the noncancer health outcomes of immunological,
developmental, cardiovascular (serum lipids), and hepatic effects (described in Section 4.1). For
all other health outcomes (e.g., reproductive, endocrine, nervous, hematological,
musculoskeletal), the evidence integration summary judgment for the human epidemiological
and animal toxicological evidence was suggestive or inadequate and these outcomes were not
assessed quantitatively. For transparency, health outcomes for which the results were suggestive
are discussed in the evidence profile tables provided in Appendix C {U.S. EPA, 2024,
11414344}.
In the previous sections describing the hazard judgment decisions (Sections 3.4 and 3.5), EPA
qualitatively considered high, medium, and sometimes low confidence studies of PFOS exposure
to characterize the weight of evidence for each health outcome. For the quantitative analyses
described in the following subsections, EPA focused exclusively on high or medium confidence
human epidemiological and animal toxicological studies for POD derivation, as recommended in
Chapter 7.2 of the IRIS Handbook {U.S. EPA, 2022, 10367891}. While the IRIS Handbook also
includes consideration of low confidence studies for dose-response analysis under certain
circumstances, this EPA assessment did not consider low confidence studies because of the
relatively large and robust database for PFOS. At this stage, EPA considered additional study
attributes to enable extrapolation to relevant exposure levels in humans. These attributes are
described in Table 7-2 of the IRIS Handbook and include relevance of the test species, relevance
of the studied exposure to human environmental exposures, quality of measurements of exposure
and outcomes, and other aspects of study design including specific reconsideration of the
potential for bias in the reported association between exposure and outcomes {U.S. EPA, 2022,
10367891}.
Consideration of these attributes facilitates the transparent selection of studies and data for dose-
response modeling and potential RfD or CSF derivation. Studies exhibiting these attributes are
expected to provide more accurate human equivalent toxicity values and are therefore preferred
4-1
-------
APRIL 2024
in the selection process. Consideration of these attributes in the study selection process are
described below for noncancer and cancer endpoints.
4.1 Noncancer
4.1.1 Study and End point Selection
For study and endpoint selection for noncancer health outcomes, the human studies that provided
all necessary analytical information (e.g., exposure distribution or variance, dose-response data)
for POD derivation, analyzed the outcome of interest in the general population or susceptible
population, and demonstrated the dose-response attributes outlined above were preferred. If
available, high and medium confidence studies with exposures levels near the range of typical
environmental human exposures, especially exposure levels comparable to human exposure in
the United States, were preferred over studies reporting considerably higher exposure levels
(e.g., occupational exposure levels). Exposure levels near the typical range of environmental
human exposure can facilitate extrapolation to the lower dose range of exposure levels that are
relevant to the overall population. When available for a given health outcome, studies with
analyses that addressed potential confounding factors affecting exposure concentrations (e.g.,
addressing temporal variations of PFOS concentrations during pregnancy due to hemodynamics)
were also preferred. Additionally, when studies presented overlapping data on the same cohort or
study population, various factors were considered to facilitate selection of one study for POD
derivation. These factors included the duration of exposure, the length of observation of the
study cohort, and the comprehensiveness of the analysis of the cohort in order to capture the
most relevant results for dose-response analysis.
The preferred animal toxicological studies consisted of medium and high confidence studies with
exposure durations appropriate for the endpoint of interest (e.g., chronic or subchronic studies vs.
short-term studies for chronic effects) or with exposure during sensitive windows of
development and with exposure levels near the lower dose range of doses tested across the
evidence base. 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 5.3 for a discussion of animal toxicological studies and endpoints
selected for POD derivation for this updated assessment compared with those selected for the
2016 PFOS HESD {U.S. EPA, 2016, 3603365}.
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 3-6, the majority of
epidemiological studies assessed endpoints related to serum biomarkers of hepatic injury (12
medium confidence studies), while fewer studies reported on liver disease or injury (3 medium
confidence studies) and other serum markers of liver function (2 medium confidence studies).
EPA prioritized studies that evaluated endpoints related to serum biomarkers of injury for
quantitative analyses because the reported effects on these endpoints were well-represented
within the database and were generally consistent across the available medium confidence
studies. Additionally, 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}. Specifically, all five
4-2
-------
APRIL 2024
medium confidence studies in general population adults from the updated literature searches
reported positive associations between PFOS serum concentrations and ALT, three of which
reported statistically significant responses {Nian, 2019, 5080307; Jain, 2019, 5080621;
Salihovic, 2018, 5083555; Gleason, 2015, 2966740; Jain, 2019, 5381541}. These more recently
published studies provided additional evidence for increased ALT in adults beyond the three
medium confidence reporting positive associations for ALT from the 2016 PFOS HESD {Lin,
2010, 1291111; Yamaguchi, 2013, 2850970; Gallo, 2012, 1276142}. Findings from studies of
other liver enzymes, AST and GGT, in adults generally reported a positive association, though
less consistently than studies of ALT; therefore, studies of AST and GGT are supportive of the
selection of ALT as an endpoint for POD derivation because these results demonstrate coherence
across the different liver serum enzyme endpoints.
As mentioned above, 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 when extrapolated to the overall population {Gilbert, 2006, 174259}. For
example, a Scandinavian study in people without any symptoms of liver disease but with
relatively small increased serum ALT levels were later diagnosed 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}.
Numerous studies have also 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}. Taken together, results of human epidemiological 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 in humans is typically assessed using serum
biomarkers of hepatotoxicity {Costello, 2022, 10285082}.
4-3
-------
APRIL 2024
Among the available medium confidence epidemiological studies reporting alterations in serum
ALT in humans, studies of adults in the general population were prioritized over studies in other
populations (e.g., occupational) or life stages (e.g. children), as the adult studies provided the
most consistent evidence of increases in ALT (see Section 3.4.1.1). Several of these medium
confidence studies reporting increases in ALT in adults were excluded from POD derivation for
reasons such as combined adolescent and adult populations {Gleason, 2015, 2966740},
populations consisting of only elderly adults {Salihovic, 2018, 5083555}, use of correlation
analyses only {Yamaguchi, 2013, 2850970}, and reporting analyses stratified by glomerular
filtration without stratifying by exposure level, which were not amenable to modeling {Jain,
2019, 5381541; Jain, 2019, 5381541}.
Exclusions of these studies resulted in the consideration of three medium confidence studies for
POD derivation {Gallo, 2012, 1276142; Lin, 2010, 1291111; Nian, 2019, 5080307} (Table 4-1).
These studies exhibited many of the study attributes outlined in Section 4 above and in Appendix
A {U.S. EPA, 2024, 11414344}. For example, Gallo et al. {, 2012, 1276142}, is the largest study
assessing PFOS and ALT in adults which was conducted in over 30,000 individuals from the
general population, aged 18-years and older, as part of the C8 Health Project in the United
States. Further, Gallo et al. {, 2012, 1276142} demonstrated a statistically significant trend in
increased ALT across deciles. Two additional studies {Lin, 2010, 1291111; Nian, 2019,
5080307} were considered for POD derivation because they reported associations in general
populations in the United States and a Chinese population located near a PFAS manufacturing
facility, 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 and reported significantly increased level of ALT associated with
PFOS. Lin et al. {, 2010, 1291111} was also considered for POD derivation since it is a large
(2,216 men and 1,063 women) nationally representative study in an NHANES adult population
and observed increased ALT levels per log-unit increase in PFOS in the models adjusted for age,
gender, and race/ethnicity. However, the association no longer remained in the fully adjusted
models, or in the models additionally adjusted for PFOA, PFHxS, and PFNA. Additionally,
several methodological limitations precluded its use for POD derivation. Limitations include lack
of clarity about the base of logarithmic transformation applied to PFOS concentrations in
regression models, and the choice to model ALT as an untransformed variable, which is a
departure from the lognormality assumed in most of the ALT literature. Therefore, two medium
confidence epidemiological studies were prioritized for POD derivation {Gallo, 2012, 1276142;
Nian, 2019, 5080307} (Table 4-1).
Liver toxicity results reported in animal toxicological studies after PFOS exposure are
concordant with the observed increased ALT indicative of hepatic damage observed in
epidemiological studies. Specifically, studies in rodents found that oral PFOS treatment resulted
in increased liver weight (11/14 high and medium confidence studies), increased 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 increased ALT observed in human populations. Mechanistic studies in
mammals and evidence from in vitro studies and nonmammalian animal models provide
additional support for the biological plausibility and human relevance of the PFOA-induced
4-4
-------
APRIL 2024
hepatic effects observed in animals. These studies suggest multiple potential MO As for the
observed liver toxicity, including PPARa-dependent and -independent mechanisms of action
(MOAs). The observed increases in liver enzymes (e.g., ALT) in rodents are supportive of the
hepatic damage confirmed during histopathological examinations in several studies. Taken
together, the study results suggest that at least some mechanisms for PFOS-induced hepatic
effects are relevant to humans.
For animal toxicological hepatic endpoints, EPA preferred studies reporting quantitative
biologically or statistically significant specific measures of severe toxicity (i.e., histopathological
lesions related to cell or tissue death or necrosis) with study designs best suited for quantitative
analysis (e.g., large sample size, reported effects in the lower dose range). Of the three studies
that quantitatively reported incidences of hepatic histopathological alterations, two were
excluded because they had relatively small sample sizes (i.e., n < 10) and used short-term
exposure durations (i.e., 28 days) {NTP, 2019, 5400978; Curran, 2008, 757871} as compared to
Butenhoff et al., {, 2012, 1276144}. Butenhoff et al. {, 2012, 1276144} was a chronic dietary
study which 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 study with a large sample size (i.e., n = 50) ,
numerous and relatively low-dose levels, and data examining a suite of endpoints, individual cell
necrosis in the liver in females was considered for derivation of PODs (Table 4-1). This endpoint
was supported by the observation of non-monotonic increases in single cell necrosis in males
from the same study.
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 3-12, the majority of
epidemiological studies assessed endpoints related to immunosuppression (1 high and 21
medium confidence studies) and immune hypersensitivity (1 high and 20 medium confidence
studies), while one study {medium confidence) also reported on endpoints related to autoimmune
disease. Studies that reported on specific autoimmune diseases were excluded from POD
derivation because there were a limited number of studies that assessed the same diseases (e.g.,
rheumatoid arthritis, celiac disease). Studies that evaluated endpoints related to immune
hypersensitivity (e.g., asthma) were also not considered for POD derivation because there were
inconsistencies in the direction and precision of effects across gender or age subgroups in the
available 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, reported differing results across medium and high
confidence studies and were therefore excluded from further consideration, though they provide
qualitative support of an association between PFOS exposure and altered immune function.
Evidence of immunosuppression in children associated with exposure to PFOS reported by
epidemiological studies were consistent across studies and endpoints. Specifically,
epidemiological studies reported associations between PFOS exposure and reduced humoral
immune response to routine childhood immunizations, including lower levels of tetanus and
diphtheria {Timmerman, 2021, 9416315; Grandjean, 2012, 1248827; Budtz-j0rgensen, 2018,
5083631} and rubella {Granum, 2013, 1937228; Stein, 2016, 3108691; Zhang, 2023, 10699594}
antibody titers. Reductions in antibody response were observed at multiple timepoints during
4-5
-------
APRIL 2024
childhood (specifically ages between 3-19 years in these studies), for either prenatal or postnatal
childhood PFOS exposure levels, and were consistent across studies in children populations from
medium confidence studies. Therefore, reduced antibody response in children was selected as an
endpoint for POD derivation.
Measurement of antigen-specific antibodies following vaccination(s) is a measure of the overall
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, which is a type of
immunity that develops when a person's immune system responds to a foreign substance or
microorganism, and it has been used as a comprehensive 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 specific challenge and is considered an adverse
immunological health outcome" {U.S. EPA, 2022, 10476098}. This is consistent with a review
article 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. {, 2016, 10293267;, 2017, 5926400;, 2019, 5080663} and in comments
from other subject matter experts on the SAB's PFAS review panel {U.S. EPA, 2022,
10476098}, the clinical manifestation of a disease after chemical exposure is not required 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 in response to chemical exposure 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 vaccine
response as a measure of immune effects as "childhood vaccine failures represent a significant
public health concern" {WHO, 2012, 10633091}. Decreases in antibody response, even at
smaller magnitudes in individuals, are clinically relevant when extrapolated to the overall
population {Gilbert, 2006, 174259}. This response also translates across multiple species,
including rodents, and extensive historical data indicate that suppression of antigen-specific
antibody responses by exogenous agents is predictive of immunotoxicity.
Studies of developmental exposure to environmental toxicants demonstrate an association with
immune suppression {Selgrade, 2007, 736210}. When immunosuppression occurs during
immune system development, the risks of developing infectious diseases and other
immunosuppression-linked diseases may increase {Dietert, 2010, 644213}. The immune system
continues developing postnatally; because of this, exposures to PFAS and other immunotoxic
agents during development may have serious, long-lasting, and irreversible health consequences
{DeWitt, 2019, 5080663; MacGillivray, 2014, 6749084; Selgrade, 2007, 736210}. Indeed,
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 that were
4-6
-------
APRIL 2024
assessed. Developmental immunotoxicity as a result of chemical exposure is generally observed
at doses lower than required to elicit immunotoxicity in adults {vonderEmbse, 2018, 6741321}.
Therefore, developmental immunotoxicity is generally a highly sensitive health outcome, both
when considering other types of developmental toxicity and when comparing it to
immunotoxicity observed in exposed adults. Luster et al. {, 2005, 2174509} similarly noted that
the specific immunotoxic endpoint of 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.
One high and 10 medium confidence studies {Grandjean, 2012, 1248827; Granum, 2013,
1937228; Mogensen, 2015, 3981889; Grandjean, 2017, 3858518; Grandjean, 2017, 4239492;
Budtz-j0rgensen, 2018, 5083631; Timmerman, 2021, 9416315; Pilkerton, 2018, 5080265; Shih,
2021, 9959487; Stein, 2016, 3108691; Zhang, 2023, 10699594} reported findings on antibody
response to tetanus, diphtheria, or rubella in children or adolescents. At least two medium
confidence studies representing two different populations of children or adolescents reported
inverse associations or increased risks of falling below seroprotective levels between each
vaccine type and PFOS concentrations. For diphtheria and tetanus, this included five medium and
one high confidence studies on the Faroe Island cohort {Grandjean, 2012, 1248827; Mogensen,
2015, 3981889; Grandjean, 2017, 3858518; Grandjean, 2017, 4239492; Shih, 2021, 9959487;
Budtz-j0rgensen, 2018, 5083631} and one medium confidence study in Greenlandic children
{Timmerman, 2021, 9416315}. For rubella, this included one medium confidence study in
Norwegian children {Granum. 2013, 1937228} and two medium confidence studies on partially
overlapping sets of children from the United States {Stein, 2016, 3108691; Zhang, 2023,
10699594}. Given the consistency of this response across multiple vaccine types and
populations, including children from the United States, EPA considered studies reporting on all
three vaccines for POD derivation. Specifically, EPA selected one medium confidence study
representing each population (i.e., children or adolescents from the United States, Faroe Islands,
Norway, and Greenland) for POD derivation.
Five separate studies {Grandjean, 2012, 1248827; Mogensen, 2015, 3981889; Grandjean, 2017,
3858518; Grandjean, 2017, 4239492; Shih, 2021, 9959487} reported on diphtheria and tetanus
antibody responses in the same population (i.e., the same individuals) of Faroese children. One
study reported on the same Faroese children cohort in a more recent medium confidence
publication {Budtz-j0rgensen, 2018, 5083631}. Because this most recent medium confidence
study is the only one of the five studies that provided dose-response data with untransformed
PFOA concentrations which are more amenable to BMD modeling, only results from Budtz-
J0rgensen and Grandjean {, 2018, 5083631} were prioritized for POD derivation and the four
other studies conducted in the Faroe Island population were excluded. For rubella, the NHANES
populations examined in Zhang et al. {, 2023, 10699594}, Stein et al. {, 2016, 3108691}, and
Pilkerton et al. {, 2018, 5080265} partially overlapped, and Zhang et al. {, 2023, 10699594} was
selected for POD derivation as it reported more recent data and a significant inverse response.
In total, four medium confidence epidemiologic studies {Budtz-j0rgensen, 2018, 5083631;
Timmerman, 2021, 9416315; Granum, 2013, 5083631; Zhang, 2023, 10699594} exhibited many
of the study attributes outlined in Section 4 above and in Appendix A {U.S. EPA, 2024,
11414344} and were considered for POD derivation (Table 4-1). Budtz-j0rgensen and
Grandjean {, 2018, 5083631} investigated anti-tetanus and anti-diphtheria responses in Faroese
4-7
-------
APRIL 2024
children aged 5-7 and Timmerman et al. {, 2021, 9416315} investigated anti-tetanus and anti-
diphtheria responses in Greenlandic children aged 7-12. Granum et al. {, 2013, 1937228}
investigated rubella responses in Norwegian children aged 3 and Zhang et al. {, 2023,
10699594} investigated rubella responses in U.S. adolescents.
Immunotoxicity results reported in animal toxicological studies are concordant 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) (two high and three 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 concordant with 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, increasing the
uncertainty in the dose-response relationship for that endpoint. Of the six studies reporting
reductions in PFC response in rodents {Peden-Adams, 2008, 1424797; Dong, 2009, 1424951;
Dong, 2011, 1424949; Zheng, 2009, 1429960; Zhong, 2016, 3748828; Keil, 2008, 1332422},
one medium confidence study {Zhong, 2016, 3748828}was selected for POD derivation because
the study tested a relatively low-dose range compared with the other five studies, the response
was observed in both males and females, and the effect was measured in pups treated with PFOS
in utero, consistent with the sensitive lifestage (i.e., children) identified from human studies
(Table 4-1). 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 immunological endpoints. EPA prioritized results from NTP {,2019,
5400978} for POD derivation over the other high confidence study {Lv, 2015, 3981558}
because it reported consistent effects of PFOS treatment on a suite of endpoints related to
immune cellularity which were confirmed by histopathological evidence (if applicable),
including increased bone marrow hypocellularity, increased splenic extramedullary
hematopoiesis, and reduced leukocytes, neutrophils, and white blood cell counts in male and
female rats. The endpoint of splenic extramedullary hematopoiesis was observed in both sexes
and was consistent with other high and medium confidence studies that reported alterations in
circulating immune cells, splenic cellularity, and thymic cellularity, both of which increase the
confidence in this endpoint (Table 4-1).
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 3-15, the majority of
epidemiological studies assessed endpoints related to serum lipids (2 high, 28 medium, and 12
mixed16 confidence studies) and blood pressure and hypertension (2 high and 17 medium
confidence studies), while some studies also reported on cardiovascular disease (1 high and 4
medium confidence studies) and atherosclerosis (1 high and 4 medium confidence studies).
16 Mixed confidence studies on serum lipids were primarily of medium confidence for total cholesterol and HDL cholesterol, and
low confidence for LDL cholesterol and triglycerides.
4-8
-------
APRIL 2024
Endpoints related to cardiovascular disease and atherosclerosis were excluded from
consideration for POD derivation because there were limited high and medium confidence
studies and they reported mixed (i.e., positive and inverse associations) or mostly null results.
Endpoints related to blood pressure and hypertension were also excluded from quantitative
analyses because there was higher confidence in analytically determined serum lipid levels
compared with blood pressure measurements and there was a larger evidence base for serum
lipids as compared to blood pressure. However, there was evidence of associations between
PFOS exposure and at least one measure of continuous blood pressure in adults and increased
risk of hypertension. These results are qualitatively supportive of an association between PFOS
and cardiovascular effects in humans.
The majority of studies in adults from the general population, including high-exposure
communities, reported positive associations between PFOS serum concentrations and serum
lipids. Studies in adults were prioritized due to the current understanding that serum lipid
changes in children are age-dependent and fluctuate during puberty {Daniels, 2008, 6815477},
which may impact the consistency of results from studies in children. Specifically, medium
confidence epidemiological studies in the general population reported positive associations
between PFOS exposure and total cholesterol (TC) (18/23 studies) and low-density lipoprotein
(LDL) (13/18 studies). Associations between PFOS and high-density lipoprotein (HDL) or
triglycerides in the general population were inconsistent and were therefore excluded from POD
derivation. EPA selected TC for quantitative assessments because the association was the most
consistently observed in adults, and studies for TC were of higher confidence for outcome
measurements compared with LDL. Additionally, the positive associations with TC in these
studies were further supported by a recent meta-analysis that included 14 general population
studies in adults {U.S. EPA, 2024, 11414059} and reported an association between increased
cholesterol and increased PFOS exposure.
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 nonalcoholic fatty liver disease (NAFLD)
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. Increases in serum cholesterol, even at smaller magnitudes at the individual level,
are clinically relevant when extrapolated to the overall population {Gilbert, 2006, 174259}. This
is because, at the population level, even small magnitude increases in serum cholesterol could
shift the distribution of serum cholesterol in the overall population relative to the clinical cut-off,
leading to an increased number of individuals at risk for cardiovascular disease. The SAB PFAS
Panel agreed with this interpretation, stating that "an increase in the number of subjects with a
clinically abnormal value is also expected from the overall change (shift in the distribution
curve) in the abnormal direction. While the clinical relevance of exposure to.. PFAS cannot be
predicted on an individual basis, the increased number of individuals within a population with
clinically defined abnormal values is of public health concern" {U.S. EPA, 2022, 10476098}.
4-9
-------
APRIL 2024
A total of 13 medium confidence studies {Chateau-Degat, 2010, 2919285; Nelson, 2010,
1291110; Liu, 2018, 4238514; Dong, 2019, 5080195; Jain, 2019, 5080642; Fan, 2020, 7102734;
Steenland, 2009, 1291109; Eriksen, 2013, 2919150; Fitz-Simon, 2013, 2850962; Lin, 2019,
5187597; Canova, 2020, 7021512; Lin, 2020, 6988476; Olsen, 2003, 1290020} reported on
positive associations between exposure to PFOS and total cholesterol in adults from the general
population. One study evaluated occupational adult populations only {Olsen, 2003, 1290020}
was not considered as exposure pathways and concentrations in this population did not represent
typical exposure scenarios for human environmental exposure. Three studies {Eriksen, 2013,
2919150; Lin, 2020, 6988476; Canova, 2020, 7021512} were excluded from POD derivation due
to narrow age ranges (i.e., 50-65 years of age, 55-75 years of age, and 20-39 years of age,
respectively) of the study populations that were less comprehensive than the age groups included
by other studies and therefore, may not apply across the general adult population. One study
{Jain, 2019, 5080642} was excluded form POD derivation because the study reported findings
stratified by BMI status but was not stratified by exposure.
Although the positive associations between PFOS and TC were supported by a recent meta-
analysis that included 14 general population studies of adults {U.S. EPA, 2024, 11414059}, EPA
did not use the pooled effect from this meta-analysis for POD derivation. This meta-analysis was
not comprehensive of the entire database of studies on PFOS and TC because it was performed
specifically with the purpose of informing aspects of the Pooled Cohort Atherosclerotic
Cardiovascular Disease (ASCVD) model which relies on CVD risk reduction analysis for those
ages 40-89 {U.S. EPA, 2024, 11414059}. The results of another recent meta-analysis on PFOS
and serum lipids {Abdullah Soheimi, 2021, 9959584} was excluded from POD derivation
because the pooled effects reported combined 11 studies with TC, triglycerides and LDL in
multiple populations (i.e., children, adolescents, pregnant women, and adults). As previously
noted, serum lipids rise in early childhood and fluctuate in puberty {Daniels, 2008, 6815477},
and combining study populations at different lifestages would likely result in unaddressed
confounding by age.
Four studies presented overlapping data from NHANES {Nelson, 2010, 1291110; Liu, 2018,
4238514; Fan, 2020, 7102734; Dong, 2019, 5080195}. Of these four, Dong et al. {, 2019,
5080195} was selected for POD derivation because this larger study included data from all
NHANES cycles between 2003 and 2014, while the other three studies reported results for only
one or two cycles within the 2003-2014 range and were therefore not further considered.
Similarly, two studies {Fitz-Simon, 2013, 2850962; Steenland, 2009, 1291109} presented data
on the C8 Health Project population. Fitz-Simon et al. {, 2013, 2850962} was not selected for
POD derivation because it was a part of a short-term follow-up and was not as comprehensive as
the population examined by Steenland et al. {, 2009, 1291109}. Likewise, another higher
exposure community study {Chateau-Degat, 2010, 2919285} reported TC changes in
approximately 700 Nunavik Inuit adults which was not as comprehensive as Steenland et al. {,
2009, 1291109} which investigated over 46,000 adults. Therefore, Steenland et al. {, 2009,
1291109} was also selected for POD derivation. Finally, Lin et al. {, 2019, 5187597} was also
selected for POD derivation because it provided data for a large number of adults (n = 940) in the
general U.S. population from the Diabetes Prevention Program (DPP) population, with PFOS
levels at baseline comparable to those from NHANES 1999-2000.
4-10
-------
APRIL 2024
In summary, 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
and exhibited many of the study attributes outlined in Section 4 above and in Appendix A {U.S.
EPA, 2024, 11414344}. 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 DPP and DPPOS at baseline (1996-1999).
Though results reported in animal toxicological studies support the alterations in lipid
metabolism observed in epidemiological studies, there are species differences direction of effect
with dose. As a result of these differences, there is some uncertainty about the human relevance
of these observed responses in rodents. Additionally, the available mechanistic data do not
increase the understanding about the non-monotonicity of serum lipid levels and decreased
serum lipid levels at higher dose levels in rodents (Section 3.4.3.3). Due to the uncertainties
regarding human relevance of the animal toxicology studies, EPA did not derive PODs for
animal toxicological studies reporting cardiovascular effects, such as altered serum lipid levels.
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 3-17, the majority of
epidemiological studies assessed endpoints related to fetal growth restriction (21 high and 26
medium confidence studies) and gestational duration (10 high and 11 medium confidence
studies), while fewer studies reported on endpoints related to fetal loss (3 high and 3 medium
confidence studies) and birth defects (4 medium confidence studies). Evidence for birth defects
was limited in that there are only 4 medium confidence studies and those studies provided mixed
findings. Therefore, birth defects not prioritized for POD derivation. Although half of the
available high and medium confidence studies reported increased incidence of fetal loss (3/6),
EPA did not prioritize this endpoint for dose-response analyses as there were a relatively limited
number of studies compared with 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.
The majority of the available studies reporting metrics of gestational duration observed increased
risk associated with PFOS exposure, including among high confidence studies. Seven of the 13
medium or high confidence studies reported inverse associations for gestational age at birth and 7
of the 11 medium or high confidence studies reported an association with preterm birth. These
findings are supportive of an association between PFOS exposure and adverse developmental
effects. There were generally consistent associations with increased risk of preterm birth,
particularly from the high confidence studies, with several studies reporting statistically
significant results. While overall there appears to be consistent associations between PFOS
exposure and gestational duration, the database for fetal growth restriction demonstrated
consistent associations between PFOS and fetal growth restriction and was also both larger and
consisted of more medium and high confidence studies than gestational duration. Therefore,
studies demonstrating fetal growth restriction were prioritized for POD derivation.
4-11
-------
APRIL 2024
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 reported inverse associations (8/12
studies; 6 high and 2 medium confidence). Endpoints characterizing fetal growth restriction were
included for POD derivation multiple studies reported effects on these endpoints, particularly
decreased birth weight, and reported generally consistent findings across high and medium
confidence studies. As noted in the Developmental Human Evidence Study Evaluation
Considerations (Section 3.4.4.1.2), measures of birth weight were considered higher confidence
outcomes compared with other measures of fetal growth restriction such as birth length, head
circumference, or ponderal index because birth weight measures are less prone to measurement
error {Shinwell, 2003, 6937192}. Studies reporting changes in mean birth weight were more
amenable to modeling compared with those reporting changes in standardized (e.g., z-score)
birth weight measurements. Standardized measurements depend on sources of standardization
and are harder to interpret and compare across studies. As a result, measures of mean changes in
birth weight were considered for quantitative analysis.
Low birth weight (LBW) 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}. LBW is widely considered a useful population level public health measure {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}. Decreases in birthweight, even at smaller
magnitudes at the individual level, are clinically relevant when extrapolated to the overall
population {Gilbert, 2006, 174259}. This is because, at the population level, even small
magnitude decreases in birthweight could shift the distribution of birthweight in the overall
population relative to the clinical cut-off, leading to an increased number of individuals at risk
for decreased birthweight and subsequent effects related to decreased birthweight. The SAB
PFAS Panel agreed with this interpretation, stating that "an increase in the number of subjects
with a clinically abnormal value is also expected from the overall change (shift in the distribution
curve) in the abnormal direction. While the clinical relevance of exposure to PFOA.. .cannot be
predicted on an individual basis, the increased number of individuals within a population with
clinically defined abnormal values is of public health concern" {U.S. EPA, 2022, 10476098}.
Substantial evidence links LBW 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}.
LBW is also associated with increased risk for diseases in adulthood, including obesity, diabetes,
and cardiovascular disease {Gluckman, 2008, 10473269; Osmond, 2000, 3421656; Risnes, 2011,
4-12
-------
APRIL 2024
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 LBW. These associations between LBW and infant mortality, childhood disease, and
adult disease establish LBW as an adverse effect. Considering the established consequences of
LBW, as well as the consistency of the database and large number of medium and high
confidence studies reporting mean birth weight and other binary birth weight-related measures,
the endpoint of decreased birth weight in infants was selected for POD derivation.
Given the abundance of high confidence epidemiological studies evaluating decreases in birth
weight, low and medium confidence studies were excluded from POD derivation. Thus, 15 high
confidence studies reporting inverse associations between exposure to PFOS and mean birth
weight {Ashley-Martin, 2017, 3981371; Bell, 2018, 5041287; Chu, 2020, 6315711; Darrow,
2013, 2850966; Gardener, 2021, 7021199; Lauritzen, 2017, 3981410; Lind, 2017, 3858512; Luo,
2021, 9959610; Sagiv, 2018, 4238410; Starling, 2017, 3858473; Valvi, 2017, 3983872;
Wikstrom, 2020, 6311677; Whitworth, 2012, 2349577; Xiao, 2020, 5918609; Yao, 2021,
9960202} were considered for POD derivation. Four studies {Ashley-Martin, 2017, 3981371;
Gardener, 2021, 7021199; Whitworth, 2012, 2349577; Xiao, 2020, 5918609} were excluded
because they reported sex-stratified results rather than results in both sexes or results for the
overall population in terms of standardized measurements (e.g., z-score) only. Analyses utilizing
standardized measurements as the dependent variable are internally valid, but this type of
analysis estimates a change in birthweight relative to the study population, which would not be
generalizable to other populations. Two studies {Bell, 2018, 5041287; Luo, 2021, 9959610}
were not considered due to the use of non-preferred exposure characterizations such as infant
whole blood samples from a heel stick and postpartum maternal exposure samples, which are
prone to exposure misclassification. Three studies {Lauritzen, 2017, 3981410; Lind, 2017,
3858512; Valvi, 2017, 3983872} were not considered further due to inconsistencies by sex or
location with no clear biological explanation for the inconsistency.
As a result of these exclusions, six remaining high confidence epidemiologic studies {Chu, 2020,
6315711; Darrow, 2013, 2850966; Sagiv, 2018, 4238410; Starling, 2017, 3858473; Wikstrom,
2020, 6311677; Yao, 2021, 9960202} met the preferred criteria outlined in Section 4.1.1 and
were considered for POD derivation (Table 4-1). The candidate epidemiological studies offer a
variety of PFOS exposure measures across different developmental windows (i.e., preconception,
fetal, neonatal). All six 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. Two of the six studies examined PFOS primarily during trimester one
{Sagiv, 2018, 4238410, Wikstrom, 2020, 6311677}, one during trimesters two and three
{Starling, 2017, 3858473} and one examined PFOS during trimester three {Yao, 2021,
9960202}. One study examined PFOS collected within days ofbirth {Chu, 2020, 6315711} and
another study {Darrow, 2013, 2850966} examined PFOS collected at the time of enrollment in
the C8 Health Project. In the latter cohort, two sets of analyses were conducted: one analysis
including all births identified from women enrolling in the study and one analysis of only the
mother's first prospective birth following enrollment (i.e., only births following blood collected
during enrollment). EPA identified the first prospective birth analysis as the analysis to consider
for POD derivation due to increased confidence in the temporal relationship between exposure
measurement and outcome assessment (i.e., not including mothers with samples collected after
4-13
-------
APRIL 2024
pregnancy). The Wikstrom et al. {, 2020, 6311677} study reported on the large Swedish
Environmental Longitudinal, Mother and child, Asthma and allergy (SELMA) study cohort with
samples collected between 2007 and 2010. Sagiv et al. {, 2018, 4238410} reported associations
between first trimester PFOS samples collected between 1999-2002 in a Project Viva birth
cohort in the United States. Darrow et al. {, 2013, 2850966} reported large inverse associations
between PFOS collected during C8 Health Project enrollment (2005-2006) in the Mid-Ohio
Valley. Chu et al. {, 2020, 6315711} reported on associations between maternal PFOS collected
within three days of delivery and birth weight in the Chinese Guangzhou Birth Cohort Study
(2013). Starling et al. {, 2017, 3858473} reported on associations between PFOS collected in
later pregnancy (range: 20 to 34 weeks gestational age) in the Healthy Start prospective cohort in
Colorado (2009-2014). Yao et al. {, 2021, 9960202} reported associations between PFOS
measured in maternal blood collected three days prior to delivery and decreased birth weight in
the Chinese Laizhou Wan Birth Cohort (2010-2013).
Developmental toxicity results reported in animal toxicological studies are concordant with the
observed developmental effects in epidemiological studies. Specifically, studies in rodents found
that gestational PFOS exposure resulted in reduced offspring weight (8/14 medium confidence
studies) and decreased offspring survival (5/9 medium confidence studies). Though limited in
number, several other studies also reported consistent effects on placental endpoints, reduced
ossification, and developmental delays. Some of these developmental effects seen in the
offspring of rodents treated with PFOA (e.g., reduced offspring weight) are consistent with the
decreases in birth weight and adverse effects associated with LBW observed in human
populations.
Given the large number of adverse effects identified in the animal toxicological database for the
developmental health outcome, EPA prioritized only the most sensitive effects (i.e., those
observed at lower dose levels and/or higher magnitude) in offspring that were 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 decreased 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 POD derivation 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 not rabbits (see Section 4.1.3). EPA also focused on endpoints
for which results from multiple animal toxicological studies corroborated the observed effect,
thereby increasing the confidence in that effect. EPA additionally focused on studies with
exposure durations lasting through the majority of gestation and/or lactation (i.e., from GD 1
through early postnatal development) rather than those that targeted a specific period of gestation
or postnatal development as they were more likely to be sensitive for detection of developmental
effects. Multiple animal toxicological studies observed effects at low dose levels and
demonstrated a dose-related response in decreased fetal weight, offspring body weight and
decreased offspring survival. Therefore, these endpoints were prioritized for dose-response
analysis.
Five studies in rats and mice reported decreased pup body weight with PFOS exposure {Luebker,
2005, 757857; Luebker, 2005, 1276160; Lau, 2003, 757854; Xia, 2011, 2919267; Yahia, 2008,
2919381}. For this endpoint, EPA selected studies in rats as the effect was observed more
4-14
-------
APRIL 2024
consistently in this species and rats appeared to be more sensitive to pup weight changes than
mice. Of the four studies reporting this effect in rats, EPA selected the data presented in the 1-
and 2-generation studies by Luebker et al. {, 2005, 757857} and Luebker et al. {, 2005,
1276160} (Fi generation only) because the exposure duration spanned prior to mating through
lactation, there were more dose groups tested than any of the other available studies, the dosing
paradigm encompassed relatively low-dose levels, the authors reported pup weight relative to
litter weight, and the effect was observed at multiple time points. Specifically, EPA selected the
time points of LD 0 and LD 5 from Luebker et al. {, 2005, 757857} and PND 1 (Fl only) from
Luebker et al. {, 2005, 1276160}.
Six studies in mice, rats, and rabbits reported decreased fetal body weight with gestational PFOS
exposure {Lee, 2015, 2851075; Wan, 2020, 7174720; Li, 2021, 9959491; Li, 2016, 3981495;
Thibodeaux, 2003, 757855; Argus, 2000, 5080012}. While the majority of studies reporting this
endpoint did not use an exposure paradigm that encompassed the earliest period of gestation (i.e.,
GD 1-4), thus increasing the uncertainty about the sensitivity of the data selected for dose-
response modeling, EPA modeled this endpoint due to the consistency of the response across
species and for comparison to PODs derived for pup weight. EPA selected studies in mice as this
species appeared to be more sensitive to fetal weight changes than rats at lower dose levels and
as described above, the PK model used in this assessment is not parameterized for rabbits.
Ultimately, Lee et al. {, 2015, 2851075} was selected for POD derivation as it reported fetal
weight for a relatively greater number of dose groups, incorporated a lower dose level than other
studies reporting this effect, and reported more than one dose group with a statistically
significant response.
Reduced offspring survival or viability was also observed with developmental PFOS exposure in
both rats and mice. Various metrics were used to assess prenatal mortality, including measures of
post-implantation loss, stillbirths, abortions, resorptions, and fetal death. Though the response
was not entirely consistent between studies, potentially due to study design and differences in the
endpoint measurement, reduced prenatal viability was observed in mice, rats, and rabbits and
qualitatively supports the observation of reduced pup survival in rats and mice. Given these
considerations, reduced fetal survival was not selected for dose-response modeling. Eight studies
reporting reduced pup survival; seven in rats and two in mice (Lau et al. {, 2003, 757854}
reported on both species). Therefore, EPA considered studies in rats for POD derivation. EPA
then selected the metric of pup survival {Gratsy, 2003, 5085464; Xia, 2011, 2919267; Chen,
2012, 1276152; Thibodeaux, 2003, 757855; Lau, 2003, 757854} over pup viability {Luebker,
2005, 757857; Luebker, 2005, 1276160} since more studies reported on the former (5 vs. 2).
Ultimately, EPA selected pup survival at PND 5 and PND 22 as reported by Lau et al. {, 2003,
757854} for POD derivation because this was a medium confidence study that presented data for
a greater number of dose groups as compared to the other studies, provided data at multiple time
points, incorporated relatively low-dose levels as compared to the other studies, used an
exposure duration that encompassed the majority of gestation (GD 2-21), and reported more than
one dose group with a statistically significant.
Table 4-1 summarizes the studies and endpoints considered for POD derivation.
4-15
-------
APRIL 2024
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/Se
X
POD
Derived?
Justification
Immune Effects
Reduced
Budtz-
Human,
Antibody
Jorgensen and
male and
Concentrations
Grandjean {,
female
for Diphtheria,
2018,
children or
Tetanus, and
508363 l}a
adolescents
Rubella
Medium
Timmerman et
al. {, 2021,
9416315}
Medium
Granum et al. {,
2013, 1937228}
Medium
Zhang et al. {,
2023,
10699594}
Medium
Yes Decreases in antibody responses to pathogens such as diphtheria, tetanus, and rubella were
observed at multiple timepoints in childhood and during adolescence, using both prenatal and
childhood PFOS exposure levels. Effect was large in magnitude and generally coherent with
epidemiological and animal toxicological evidence for other immunosuppressive effects. Effects
were observed in multiple populations, including adolescents from the United States.
Decreased PFC
Response to
SRBC
Zhongetal. {, C57BL/6
2016,3748828} Mice, Fi
Medium males
Extramedullar NTP {, 2019,
Hematopoiesis 5400978}
in the Spleen High
Yes Functional assessment indicative of immunosuppression 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).
Sprague- Yes Blood cell production outside of the bone marrow which occurs when normal cell production is
Dawley impaired. Selected for POD derivation because the results were from a high confidence study,
Rats, adult histopathologically confirmed, consistent across both sexes, accompanied by evidence of bone
male and marrow hypocellularity, and consistent with other studies that reported alterations in circulating
female immune cells, splenic cellularity, and thymic cellularity.
Developmental Effects
Decreased Birth
Weight
Chuetal. {, Human,
2020,6315711} male and
Yes Evidence for developmental effects is based on consistent inverse effects for FGR including
birthweight measures which are the most accurate endpoint examined. Some deficits were
4-16
-------
APRIL 2024
„ , . Reference, „ . POD
Endpoint „ Species/Se „ .
1 Confidence Derived:
x
Justification
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
female
infants
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.
Decreased Fetal Leeetal. {, CD-I Mice,
Body Weight 2015, 2851075} Fi males
Medium and females
Decreased Pup Luebkeretal. {, Sprague-
Body Weight
(relative to
litter)
2005,757857}
Medium
Luebker et al. {,
2005, 1276160}
Dawley
Rats, Fi
male and
female (LD
0 and LD 5
{Luebker,
2005,
757857};
PND 1
Yes Effect was consistently observed across six studies and three species {Argus, 2000, 5080012; Li,
2016, 3981495; Lee, 2015, 2851075; Wan, 2020, 7174720; Li, 2021, 9959491; Thibodeaux,
2003, 757855} and is coherent with epidemiological evidence of decreased birth weight and
evidence of reduced pup weight in rodents. 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, incorporates a relatively greater number of
dose groups, and reported more than one dose group with a statistically significant response
compared with other studies reporting this effect, and because mice appear to be a more sensitive
model for this endpoint than rats.
Yes Effect was consistently observed across five studies and two species and is coherent with
epidemiological evidence of decreased birth weight and evidence of decreased fetal weight in
rodents. Luebker et al. {, 2005, 757857} and Luebker et al. {, 2005, 1276160} were selected
because rats appear to be more sensitive than mice to this endpoint, the studies are designed to
be sensitive to this effect (i.e., multigeneration studies testing relatively large numbers of dose
groups and low-dose ranges), the studies reported effects as relative to litter and the studies
observed effects in multiple generations or multiple time points and in multiple dose groups.
4-17
-------
APRIL 2024
„ , . Reference, „ . POD
Endpoint „ Species/Se „ .
1 Confidence Derived:
x
Justification
{Luebker,
2005,
1276160})
Decreased Pup Lauetal. {, Sprague- Yes Decreased offspring survival was consistently observed across eight studies and two species and
Survival 2003, 757854} Dawley is also supported by reduced fetal survival observed in rodents. Lau et al. {, 2003, 757854} was
Medium Rats, Fi selected because rats appeared to be more sensitive to this effect than mice and because the study
male and presented data for a greater number of dose groups and at multiple time points compared with
female the other four studies in rats, incorporated relatively low-dose levels, used an exposure duration
(PND 5 and that encompassed the majority of gestation (GD 2-21), and reported more than one dose group
PND 22) with a statistically significant response.
Serum Lipid Effects
Increased Total
Cholesterol
Dong et al. {,
2019, 5080195}
Medium
Linetal. {,
2019, 5187597}
Medium
Steenland et al.
{, 2009,
1291109}b
Medium
Human, Yes Effect was consistent and observed across multiple adult populations including general
male and population adults in NHANES {Dong, 2019, 5080195}; from prediabetic adults from the DPP
female and DPPOS cohort {Lin, 2019, 5187597} and the C8 Health project high-exposure community
adults {Steenland, 2009, 1291109}, as well as when study designs excluded individuals prescribed
cholesterol medication, minimizing concerns of bias due to medical intervention {Dong, 2019,
5080195; Steenland, 2009, 1291109}. Endpoint is supported by associations between PFOS and
blood pressure.
Hepatic Effects
Increased ALT Gallo et al. {,
Human
Yes
Effect was consistent and observed across multiple populations including general population
2012, 1276142}
(male and
adults {Lin, 2010, 1291111} (NHANES) and high-exposure communities {Gallo, 2012,
Medium
female
1276142} (C8 Health Project); {Nian, 2019, 5080307} (Isomers of C8 Health Project in China).
Nian et al. {,
adults)
2019, 5080307}
Medium
Increased ALT
Linetal. {, Human
2010,1291111} (male and
Medium 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.
4-18
-------
APRIL 2024
Reference,
Confidence
Strain/
POD
Derived?
Endpoint
Species/Se
X
Justification
Individual Cell
Necrosis in the
Liver
Butenhoff et al.
{,2012,
1276144}
High
Sprague-
Dawley
rats,
females
Yes
Effect was supported by a non-monotonic response in males from the same study {Butenhoff,
2012, 1276144}. 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 etal. {,2012, 1248827}; Grandjean et al. {,2017, 3858518}; Grandjean et al. {,2017,4239492}.
b See Section 5.6.3 for discussion on the approach to estimating BMDs from regression coefficients.
4-19
-------
APRIL 2024
4.1.2 Estimation or Selection of Points of Departure 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 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 1 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 serious effects that can result in irreversible structural or functional changes to the
organism, and thq 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. Thq 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 1 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 noncancer health
outcomes are described in the subsections below and alongside the modeling methods and results
provided in Appendix E {U.S. EPA, 2024, 11414344}. Considerations for BMR selection for
cancer endpoints are described in Section 4.24.2 and Appendix E {U.S. EPA, 2024, 11414344}.
4.1.2.1 Hepatic Effects
For the hepatic endpoint of increased serum ALT in adults associated with PFOS exposure, the
BMD and the BMDL were estimated using a BMR of 5% extra risk from the biologically
4-20
-------
APRIL 2024
significant adverse serum ALT level (see Table 4-2). As described in detail in Appendix E {U.S.
EPA, 2024, 11414344}, EPA reviewed the available information regarding potential clinical
definitions of adversity for the endpoint of elevated ALT. Specifically EPA modeled elevated
human ALT using cutoff levels of 42 IU/L for males and 30 IU/L for females {Valenti, 2021,
10369689}. These are the most updated clinical consensus cutoffs which post-date the American
Association for the Study of Liver Diseases (AASLD) journal of Clinical Liver Disease
recommended values of 30 IU/L for males, and 19 IU/L for females {Kasarala, 2016, 11350060;
Ducatman, 2023, 11412135}. Valenti et al. (2021, 1036989) determined the values using the
same approach at the same center, but using an updated standardized method, a large cohort of
apparently healthy blood donors (ages 18-65 years) and showed that the updated cutoffs were
able to better predict liver disease.
Because EPA identified a biological basis for BMR selection, EPA used the hybrid approach
(see Section 2.3.3.1 of USEPA {, 2012, 1239433}) to estimate the probability of an individual
with an adverse serum ALT level as a function of PFOS exposure. This approach effectively
dichotomizes the data; therefore, EPA considered BMRs of 1%, 5%, and 10% extra risk for this
endpoint. As described in the Benchmark Dose Technical Guidance {U.S. EPA, 2012,
1239433}, a 10% BMR is often used to describe quantal data, however, "for epidemiological
data, response rates of 10% extra risk would often involve upward extrapolation, in which case it
is desirable to use lower levels, and 1% extra risk is often used as a BMR." EPA considered
BMRs of 5%> and 10% extra risk. EPA did not select a 1% BMR because it is often used for
frank effects and cancer incidence {U.S. EPA, 2012, 1239433}, neither of which apply to the
endpoint of elevated serum ALT.
EPA selected a BMR of 5% extra risk because studies have demonstrated that ALT levels at or
slightly above the selected cutoff levels can be associated with more severe liver diseases
{Mathiesen, 1999, 10293242; Wedmeyer, 2010, 11374673}, increased risk of liver-related
mortality {Park, 2019, 10293238; Ruhl, 2009, 3405056; Kim, 2004, 10473876}, and mortality
{Lee, 2008, 10293233}. Based on the severity of the health effects associated with increased
ALT, EPA determined that a BMR of 5% extra risk is warranted {U.S. EPA, 2012, 1239433}; a
10%) extra risk would result in a greater number of individuals, especially those in sensitive
subpopulations, experiencing more severe liver diseases such as advanced fibrosis, chronic liver
disease, and even liver-related death. Since there is currently a relatively high prevalence of
elevated ALT in the general population (14% and 13% of U.S. male and female adults,
respectively, aged 20 and older {Valenti, 2021, 10369689}), a small increase in the prevalence
of elevated ALT associated with PFOA exposure would likely increase the number of
individuals with severe liver-related health effects. This also supports using a more health
protective BMR of 5% extra risk (rather than 10%) for POD derivation. EPA presents PODs with
a 10%) BMR for comparison purposes in Appendix E {U.S. EPA, 2024, 11414344}, as
recommended by agency guidance {U.S. EPA, 2012, 1239433}.
For the adverse effect of individual cell necrosis observed in livers of adult rats following PFOS
exposure, there is currently inadequate available biological or toxicological information to permit
determination of an effect-specific 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 because it is considered the standard reporting level for quantal
(dichotomous) data and a minimally biologically significant response level (see Table 4-2).
4-21
-------
APRIL 2024
4.1.2.2 Immune Effects
For the developmental immune endpoint of decreased diphtheria, rubella, and tetanus antibody
response in children 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 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}, Timmermanetal. {, 2021, 9416315}, Granum et al. {,2013, 1937228}, and
Zhang et al. {, 2023, 10699594} measured antibody concentrations in childhood and PFOS
exposure during gestation or childhood, these are considered developmental studies based on
EPA's Guidelines for Developmental Toxicity Risk Assessment {U.S. EPA, 1991, 732120},
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 cutoff for the health outcome {U.S. EPA,
2012, 1239433}. As described in detail in Appendix E {U.S. EPA, 2024, 11414344}, EPA
reviewed the available information regarding potential clinical definitions of adversity for this
effect. A blood concentration for tetanus and diphtheria antibodies of 0.1 IU/mL has been cited
in the literature as a "protective level" {Grandjean, 2017, 4239492; Galazka, 1989, 9642152}.
However, in the Immunological Basis for Immunization Series of modules {WHO, 2018,
10406857}, the WHO argued that assay-specific "protective levels" of tetanus antitoxin may not
actually guarantee immunity. Galazka et al. {, 1993, 10228565} similarly argued that several
factors give rise to variability in the vulnerability of individuals to diphtheria and there is no
consensus on what level offers full protection. For rubella, 10 IU/mL has been cited in the
literature as a protective level {Skenzdel, 1996, 11374672}, however, the geographical
variability, lack of consensus, and relatively dated assessment of this cutoff precludes its use as
the basis of the BMR {Charlton, 2016, 11374674}. As such, EPA determined that there is no
clear definition of an adverse effect threshold for the endpoints of reduced tetanus, rubella, or
diphtheria antibody concentrations in children or adolescents.
With these two factors in mind, a 0.5 SD was selected as the BMR because: 1) the health
outcome is developmental, and 2) there is no accepted definition of an adverse level of change or
clinical cutoff 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. EPA also presented BMDL modeling using a BMR equivalent to a 1 SD change, as
recommended by agency guidance {U.S. EPA, 2012, 1239433}.
For the adverse effects of decreased PFC response to SRBC observed in PNW 4 mice and
splenic extramedullary hematopoiesis in adult rats following PFOS exposure, there is currently
inadequate available biological or toxicological information to permit determination of minimal
biologically significant response levels. In accordance with EPA's Benchmark Dose Technical
4-22
-------
APRIL 2024
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).
4.1.2.3 Cardiovascular Effects
For the cardiovascular endpoint of increased serum TC in adults associated with PFOS exposure,
the BMD and the BMDL were estimated using a BMR of 5% extra risk from the biologically
significant adverse serum TC concentration {Dong, 2019, 5080195; Steenland, 2009, 1291109}
or a BMR of 0.5 SD {Lin, 2019, 5187597}, depending on the data provided by the study (see
Table 4-2). As described in detail in Appendix E {U.S. EPA, 2024, 11414344}, EPA reviewed
the available information regarding potential clinical definitions of adversity for this effect and
identified the definition of hypercholesterolemia from the American Heart Association {NCHS,
2019, 10369680} as providing the most recent upper reference limit for clinically adverse serum
TC. Specifically, when possible, EPA modeled human cholesterol using a cutoff level of
240 mg/dL for elevated serum total cholesterol {NCHS, 2019, 10369680}.
Because EPA identified a biological basis for BMR selection, EPA used the hybrid approach
(see Section 2.3.3.1 of USEPA {, 2012, 1239433}) to estimate the probability of an individual
with an adverse TC level as a function of PFOS exposure. This approach effectively
dichotomizes the data; therefore, EPA considered BMRs of 1%, 5%, and 10% extra risk for this
endpoint. As described in the Benchmark Dose Technical Guidance {U.S. EPA, 2012,
1239433}, a 10% BMR is often used to describe quantal data, however, "for epidemiological
data, response rates of 10% extra risk would often involve upward extrapolation, in which case it
is desirable to use lower levels, and 1% extra risk is often used as a BMR." EPA considered
BMRs of 5% and 10% extra risk. EPA did not select a 1% BMR because it is often used for
frank effects and cancer incidence {U.S. EPA, 2012, 1239433}, neither of which apply to the
effect of elevated serum TC. For Lin {, 2019, 5187597), EPA relied on the mean serum TC
concentrations reported across PFOS quartiles (i.e., continuous data) provided by the study, and
therefore considered a BMR of a change in the mean equal to 0.5 SD or 1 SD from the control
mean.
Increased serum cholesterol is associated with changes in incidence of cardiovascular disease
events such as myocardial infarction (MI, i.e., heart attack), IS, and cardiovascular mortality
occurring in populations without prior CVD events {D'Agostino, 2008, 10694408; Goff, 2014,
3121148; Lloyd-Jones, 2017, 10694407}. Based on the severity of the cardiovascular-related
health effects associated with increased cholesterol, EPA determined that selection of a BMR of
5% extra risk or 0.5 SD is warranted {U.S. EPA, 2012, 1239433}; a 10% extra risk or 1SD
would result in a greater number of individuals, especially those in sensitive subpopulations,
experiencing increased incidence of cardiovascular disease events. Since there is currently a
relatively high prevalence of elevated TC in the general population (11.5% of U.S. adults aged
20 and older {NCHS, 2019, 10369680}), a small increase in the prevalence of elevated TC
associated with PFOA exposure would likely increase risk of severe health outcomes, such as
cardiovascular-related events. Thus, this supports using a more conservative BMR of 5% extra
risk or 0.5 SD for POD derivation. EPA presents PODs with a BMR of 10% extra risk {Dong,
2019, 5080195; Steenland, 2009, 1291109} or 1 SD {Lin, 2019, 5187597} for comparison
4-23
-------
APRIL 2024
purposes in Appendix E {U.S. EPA, 2024, 11414344}, as recommended by agency guidance
{U.S. EPA, 2012, 1239433}.
4.1.2.4 Developmental Effects
For the developmental endpoint of decreased birth weight associated with PFOS exposure, the
BMD and the BMDL were estimated using a BMR of 5% extra risk from the biologically
significant birth weight deficit (see Table 4-2). As described in Appendix E {U.S. EPA, 2024,
11414344}, LBW is clinically defined as birth weight less than 2,500 g (approximately 5.8 lbs)
and can include but is not exclusive to 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}.
Consistent with EPA's Benchmark Dose Technical Guidance {U.S. EPA, 2012, 1239433}, EPA
typically selects a 5% or 0.5 SD benchmark response (BMR) when performing dose-response
modeling of data from an endpoint resulting from developmental exposure. Low birthweight is
associated with increased risk for adverse health effects throughout life {Hack, 1995, 8632216;
Reyes, 2005, 1065677; Tian, 2019, 8632212} and therefore, supports a more stringent BMR
below 10% (for dichotomous data) or 1 SD (for continuous data). Because EPA identified a
biological basis for BMR selection, EPA used the hybrid approach (see Section 2.3.3.1 of U.S.
EPA {, 2012, 1239433}) to estimate the probability of an individual with a birth weight deficit as
a function of PFOS exposure. This approach effectively dichotomized the data, resulting in a
BMR defined as a 5% increase in the number of infants with birth weights below 2,500 g.
For decreased fetal and pup weights and decreased pup survival observed in animal studies,
BMRs of 5% relative deviation and 0.5 SD from the control were employed, respectively (see
Table 4-2). As with human data, these BMRs are 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) 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 SD, and found they were statistically similar {Kavlock, 1995,
75837}.. EPA presented modeling results using a BMR of 0.5 SD for decreased fetal and pup
body weight and a BMR of 0.1 SD for the frank effect of decreased pup survival for comparison
purposes, as recommended by EPA guidance {U.S. EPA, 2012, 1239433} (see Appendix, {U.S.
EPA, 2024, 11414344}).
Table 4-2. Benchmark Response Levels Selected for BMD Modeling of Health Outcomes
Endpoint
BMR
Rationale
Immune Effects
Reduced antibody
concentrations for diphtheria,
0.5 SD
Consistent with EPA guidance. EPA typically selects a 5% or
0.5 SD benchmark response (BMR) when performing dose-
response modeling of data from an endpoint resulting from
4-24
-------
APRIL 2024
Endpoint
BMR
Rationale
rubella, and tetanus in
children or adolescents
Decreased PFC Response to
SRBC (PNW 4)
Extramedullar
Hematopoiesis in the Spleen
developmental exposure in consideration of the severity of the
effect and selects a 1 or 0.5 SD change in cases where there is no
accepted definition of an adverse level of change or clinical
cutoff 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 1 SD was used as per EPA guidance
{U.S. EPA, 2012, 1239433}
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
5% extra risk of
exceeding
adversity cutoff
Decreased Fetal or Pup
Weight
Decreased Pup Survival
Increased Cholesterol
Consistent with EPA guidance. EPA typically selects a 5% or
0.5 SD benchmark response (BMR) when performing dose-
response modeling of data from an endpoint resulting from
(hybrid approach) developmental exposure in consideration of the severity of the
effect {U.S. EPA, 2012, 1239433}. The use of the hybrid
approach results in dichotomization of the data and therefore a
5% BMR was selected {U.S. EPA, 2012, 1239433}
5% Consistent with EPA guidance. EPA typically selects a 5% or
0.5 SD benchmark response (BMR) when performing dose-
response modeling of data from an endpoint resulting from
developmental exposure in consideration of the severity of the
effect {U.S. EPA, 2012, 1239433}
0.5 SD Consistent with EPA guidance. EPA typically selects a 5% or
0.5 SD benchmark response (BMR) when performing dose-
response modeling of data from an endpoint resulting from
developmental exposure in consideration of the severity of the
effect {U.S. EPA, 2012, 1239433}
Cardiovascular Effects
5% extra risk of Although EPA's Benchmark Dose Technical Guidance {U.S.
exceeding EPA, 2012, 1239433 } recommends a BMR based on a 10%
adversity cutoff extra risk for dichotomous endpoints when biological
(hybrid approach) information is not sufficient to identify the BMR, "for
epidemiological data, response rates of 10% extra risk would
often involve upward extrapolation, in which case it is desirable
to use lower levels" {U.S. EPA, 2012, 1239433}. Because
increased TC is not a frank effect but is associated with
increased incidence of severe cardiovascular-related effects a 5%
extra risk was used as the BMR. The response rate of 5% extra
risk limits further increases in the prevalence of this effect.
0.5 SD Because increased TC is not a frank effect but is associated with
increased incidence of severe cardiovascular-related effects, a
0.5 SD was used as the BMR. A change from the mean of 0.5
SD limits further increases in the prevalence of this effect
4-25
-------
APRIL 2024
Endpoint
BMR
Rationale
Hepatic Effects
Increased ALT 5% extra risk of Although EPA's Benchmark Dose Technical Guidance {U.S.
exceeding EPA, 2012, 1239433 } recommends a BMR based on a 10%
adversity cutoff extra risk for dichotomous endpoints when biological
(hybrid approach) information is not sufficient to identify the BMR, "for
epidemiological data, response rates of 10% extra risk would
often involve upward extrapolation, in which case it is desirable
to use lower levels" {U.S. EPA, 2012, 1239433}. Because
increased ALT is not a frank effect but is associated with
increased incidence of severe liver-related effects a 5% extra risk
was used as the BMR. The response rate of 5% extra risk limits
further increases in the prevalence of this effect
Individual Cell Necrosis 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.
4.1.3 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 (see Section 3.3.2), 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 Appendix, {U.S. EPA, 2024, 11414344}), 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. The species used in the
toxicological studies (i.e., species of interest) were rats, mice, and nonhuman primates; model
parameters for these species of interest were available. Good agreement with training and test
(out-of-sample) datasets shows that the model performance is good compared with both the data
used to identify model parameters and to external data. This was assessed using the mean square
log error (MSLE) to compare model predicted concentration values to observed PFOS serum
concentrations following single dose exposure to animals. Training set data demonstrated an
MSLE of 0.17 for PFOS, respectively. For test set data, the MSLE was 0.38 for PFOS. The
general agreement between test and training datasets increases confidence that the model can be
used to make accurate predictions of internal dose metrics for the dose magnitudes used in the
available toxicology studies. 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 lifestages.
4-26
-------
APRIL 2024
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 lifestage component for predicting developmental study designs.
These analyses are further described below. Uncertainties and limitations of the selected
modeling approach are described in Section 5.6.1.
4-27
-------
APRIL 2024
4.1.3.1.1 Animal Model Parameters
Pharmacokinetic parameters for different species and strains represented in the Wambaugh et al. {, 2013, 2850932} model are
presented in Table 4-3.
Table 4-3. PK Parameters from Wambaugh et al. {, 2013, 2850932} Meta-Analysis of Literature Data for PFOS
Parameter
Units
CD1 Mouse
(F)a
CD1 Mouse
(M)a
Sprague-Dawley Rat
(FT
Sprague-Dawley Rat
(M)a
CynomolgusMonkey
(M/F)a
Body weightb (BW)
kg
0.02
0.02
0.203
0.222
3.42
Cardiac Output0 (Qcc)
L/h/kg074
8.68
8.68
12.39
12.39
19.8
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)
(0.251-1.06)
(0.24-4.09)
(0.238-3.62)
(0.256-4.01)
(0.256-1.05)
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 x e4)
(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 (Vmc)
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).
aDatasets 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 body weight for species: individual-specific bodyweights.
cCardiac outputs obtained from Davies and Morris {, 1993, 192570}.
4-28
-------
APRIL 2024
4.1.3.1.2 Out-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 PFOS 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 with 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.
information obtained from Kim et al. {, 2016, 3749289}.
b Information obtained from Huang et al. {, 2019, 5387170}.
4.1.3.1.3 Life 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.
4-29
-------
APRIL 2024
Gestation Lactation Post-weaning
Figure 4-1. Model Structure for Lifestage 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
lifestage-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
4-30
-------
APRIL 2024
infant toxicokinetics (Table 4-5). Finally, the Wambaugh et al. {, 2013, 2850932} model was not
parameterized for a postpartum 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
40°
36.87 g
Volume of Distribution (Vd)
L/kg
0.28d
0.26 g
Starting Milk Consumption Rate (r°miik)
kgmiik/day
0.001h
o.ooor
Week 1 Milk Consumption Rate (rVik)
kgmiik/dav
0.003h
0.00031
Week 2 Milk Consumption Rate (i2™^)
kgmiik/dav
0.0054h
0.000541
Week 3 Milk Consumption Rate (r\nin:)
kgmiik/dav
0.0059h
0.000591
Notes: PK = pharmacokinetic.
aInformation obtained from Loccisano etal. {,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}.
gInformation obtained from Chang et al. {, 2012,1289832}.
information obtained from Kapraun et al. {, 2022, 9641977} (adapted from Lehmann et al. {, 2014, 2447276}).
'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.
4-31
-------
APRIL 2024
Figure 4-2 demonstrates the model's ability to predict gestation and lactation study designs in rat
dams exposed to 1.6 mg/kg/day PFOS that gave birth to pups who are exposed through gestation
and lactation until weaning {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.
3 1031
E 1Q2 i
I ^ 10l 1
0 O
£ 10° i
o
£ 10"1 ¦
=r 103 i
oi
£
lti
O
u_
^ 101
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 laboratory
animal species used in toxicity studies and extrapolate these internal dose POD to humans.
Therefore, to evaluate its predictive utility for risk assessment, a number of dose metrics across
lifestages 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
lifestages: during gestation (Cavg dam gest), during lactation (Cavg damjact), or combined gestation
and lactation (Cavg dam gest jact). In pups for Cmax, two different lifestages 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_lact and Cavg_pup_gest_lact).
PFOS: 1.6 mg/kg/day
-40
-20
1
0
20
40
birth
-40 -20 0 20 40
Time [days]
4-32
-------
APRIL 2024
EPA selected the metric of Ciast7 for studies examining noncancer effects using non-
developmental exposure paradigms. This metric provides a consistent internal dose for use
across disparate chronic and subchronic study designs where steady state may or may not have
been reached in the animal following continuous dosing. When the animal has reached steady
state, Ciast7 is equal to the steady-state concentration and for non-steady-state study designs, this
metric averages the concentration variability over a week's worth of dosing rather than using a
single, maximal concentration. This allows for extrapolation to the human model where a steady-
state assumption is implemented for adult dose-metric calculations.
For developmental endpoints, the metric of Cmax is typically used when there is a known MOA
related to a threshold effect during a specific window of susceptibility. From the Guidance for
applying quantitative data to develop data-derived extrapolation factors for interspecies and
intraspecies extrapolation {U.S. EPA, 2014, 2520260}, the choice of this metric "depends on
whether toxicity is best ascribed to a transient tissue exposure or a cumulative dose to the target
tissue." Furthermore, the guidance clarifies that "for chronic effects, in the absence of MOA
information to the contrary, it is generally assumed that some integrated cumulative measure of
tissue exposure to the active toxicant is the most appropriate dose metric (e.g., AUC)" {U.S.
EPA, 2014, 2520260}. Repeat dosing coupled with a lack of a defined MOA for the apical
endpoints used for dose-response modeling resulted in EPA excluding Cmax as an internal dose
metric for animal toxicological endpoints. However, EPA provides modeling results using Cmax
for comparison purposes in Appendix E {U.S. EPA, 2024, 11414344}.
EPA selected the metric of Cavg for studies with reproductive or developmental exposure designs
encompassing gestation and/or lactation. One factor considered for this selection pertains to the
long half-life of PFOA and the degree of accumulation throughout pregnancy and lactation.
Because PFOA is not cleared within 24 hours, daily dosing throughout pregnancy/lactation will
result in a Cmax that falls on the final day of pregnancy or lactation or a Ciast7 only representative
of the final days of gestation or lactation, even if dosing ceases after birth, due to ongoing
lactational exposure. The endpoints in this assessment (decreased fetal or pup weight, decreased
pup survival, delayed time to eye opening) do not have established MO As or known windows of
susceptibility and instead are expected to result from sustained internal dose from repeated
exposures. If, as anticipated, this window of susceptibly for a given endpoint is not on the final
day or the last week of exposure, the Cmax or Ciast7 will not capture the exposure at the time
associated with the adverse effect. A Cavg metric is more representative of the exposure
throughout the potential window of susceptibility. This selection is also supported by the
Guidelines for Developmental Toxicity Risk Assessment {U.S. EPA, 1991, 11346204}, which
state that when pharmacokinetic data are available, as is the case for PFOA, "adjustments may be
made to provide an estimate of equal average concentration at the site of action for the human
exposure scenario of concern." The selection of Cavg for developmental animal studies is
therefore consistent with the guidance for humans.
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 body weight during
growth, and that required minimal new development. Previous modeling efforts suggested that
4-33
-------
APRIL 2024
limiting model complexity helps to prevent errors and facilitates rapid implementation
{Bernstein, 2021, 9639956}. For the human epidemiological and animal toxicological 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 5.6.2 and 5.7.
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. Body weight
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 body weight {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 Appendix, {U.S. EPA, 2024, 11414344}).
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.
The use of the Verner model in humans presents a substantial advancement in approach for
endpoints in children compared with the previous EPA assessment of PFOS {U.S. EPA, 2016,
3603365}. The previous 2016 HESD 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 in children. The current approach explicitly models PFOS exposure to
4-34
-------
APRIL 2024
infants during nursing who are undergoing rapid development, including growth, through
childhood, and 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 Appendix, {U.S.
EPA, 2024, 11414344}). 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 Appendix, {U.S. EPA,
2024, 11414344}).
Table 4-6. Updated and Original Chemical-Specific Parameters for PFOS in Humans
Parameter
Updated Value
Original Value3
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
Notes:
a Verner et al. {, 2016, 3299692}.
bThompsonetal. {,2010,2919278}.
cLi etal. {,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 Appendix, {U.S. EPA, 2024,11414344}). 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 selected a reported half-life value from an exposure to a study population that is
demographically representative of the general population, with a clear decrease in exposure at a
known time, with a high number of participants and a long follow-up time. Based on 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 blood 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 {U.S. EPA, 2024, 11414344}. Uncertainties related to EPA's
selected half-life are discussed in Section 5.6.2.
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
(PFOA, human*Vd(PFOS, monkey)/Vd (PFOA, monkey). Vd is a parameter that is relatively
4-35
-------
APRIL 2024
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
nonhuman primates {Mordenti, 1991, 9571900}. For comparison, the optimized Vd value from
oral dosing in monkeys was 220 mL/kgforPFOS {Andersen, 2006, 818501}.
A summary of PFOS Vd values is presented in the Appendix {U.S. EPA, 2024, 11414344}.
Uncertainties related to EPA's selected Vd are discussed in Section 5.6.2.
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
{U.S. EPA, 2024, 11414344}. 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 with 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 et al. {,2011,
2919258
0.011
No
2919258}
Liu et al. {,2011,2919240}
2919240
0.020
No
Karrmanetal. {, 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
-
Notes:
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
Appendix, {U.S. EPA, 2024, 11414344}). 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 Appendix, {U.S. EPA, 2024, 11414344}). 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
4-36
-------
APRIL 2024
would result in the same dose metric after chronic exposure. For PODs from exposure to animals
in developmental scenarios, the updated 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 (http s://eithub. com/LI SEP A/O W -PF O S -PF O A-MCLG- support-
PK-models). 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 PFOS Toxicological End points and
Dosimetric Interpretation of Epidemiological End points
Different approaches were taken to estimate PODheds depending on the species (i.e., human vs.
animal model) and lifestage (e.g., developmental, adult). The PODs from epidemiological studies
(immune, developmental, hepatic, and serum lipid endpoints) were derived using hybrid or
benchmark dose modeling (see Appendix E.l, {U.S. EPA, 2024, 11414344}) which provided an
internal serum concentration in ng/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 Appendix E.2 {U.S. EPA, 2024,
11414344}. 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 Appendix E, {U.S. EPA, 2024, 11414344}). If BMD modeling did
not produce a viable model, a NOAEL or LOAEL approach was used consistent with EPA
guidance {U.S. EPA, 2012, 1239433}. 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
4-37
-------
APRIL 2024
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.
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.
4-38
-------
APRIL 2024
Table 4-8. PODheds Considered for the Derivation of Candidate RfD Values
Endpoint
Reference,
Confidence
Strain/Species/Sex/Age
POD Type,
Model
POD Internal
Dose/Internal
Dose Metric3
PODhed
(mg/kg/day)
Notes on Modeling
Immunological Effects
Decreased serum
anti-tetanus
antibody
concentration in
children
Budtz-Jorgensen
and Grandjean {,
2018, 508363l}b
Medium
Human, male and female;
PFOS concentrations at age
5 and anti-tetanus antibody
serum concentrations at age
7
BMDLo.5 sd
18.5 ng/mL
2.71 x 10-6
Single- and multi-PFAS
models resulted in
comparable BMDLs though
there was a 55% change in
the effect size when
controlling for PFOA;
selected BMDL was based on
a non-significant regression
parameter
Budtz-Jorgensen
and Grandjean {,
2018, 508363l}b
Medium
Human, male and female;
PFOS concentrations in the
mother0 and anti-tetanus
antibody serum
concentrations at age 5
BMDLo.5 sd
29.9 ng/mL
5.21 x 10-6
PFOS concentrations may be
influenced by pregnancy
hemodynamics; single- and
multi-PFAS models resulted
in poor quality of model fits;
selected BMDL was based on
a non-significant regression
parameter
Timmerman et al. {,
2021,9416315}
Medium
Human, male and female;
PFOS concentrations and
anti-tetanus antibody
concentrations at ages 7-12
BMDLo.5 sd
9.66 ng/mL
1.78 x 10-6
BMDL based on non-
significant regression
parameter and resulted in a
poor quality of model fit;
BMR of 0.5 SD may not be a
reasonably good estimate of
5% extra risk
Decreased serum
anti-diphtheria
antibody
concentration in
children
Budtz-Jorgensen
and Grandjean {,
2018, 508363l}b
Medium
Human, male and female;
PFOS concentrations at age
5 and anti-diphtheria
antibody serum
concentrations at age 7
BMDLo.5 sd
12.5 ng/mL
1.83 x 10-6
Single- and multi-PFAS
models resulted in
comparable BMDLs though
there was a 36% change in
the effect size when
controlling for PFOA;
selected BMDL was based on
a significant regression
parameter
4-39
-------
APRIL 2024
Endpoint
Reference,
Confidence
Strain/Species/Sex/Age
POD Type,
Model
POD Internal
Dose/Internal
Dose Metric3
PODhed
(mg/kg/day)
Notes on Modeling
Budtz-Jorgcnscn
and Grandjean {,
2018, 508363l}b
Medium
Human, male and female;
PFOS concentrations in the
mother0 and anti-tetanus
antibody serum
concentrations at age 5
BMDLo
20.0 ng/mL
3.48 x 10 6 PFOS concentrations may be
influenced by pregnancy
hemodynamics; single- and
multi-PFAS models resulted
in comparable BMDLs
though there was a 22%
change in the effect size when
controlling for PFOA;
selected BMDL was based on
a significant regression
parameter
Timmerman et al. {,
2021,9416315}
Medium
Human, male and female;
PFOS concentrations and
anti-diphtheria antibody
concentrations at ages 7-12
BMDLo
5.61 ng/mL
1.03 x 10~6 BMDL based on model with
poor quality of fit; BMDL
based on significant
regression parameter; BMR
of 0.5 SD may not be a
reasonably good estimate of
5% extra risk
Decreased serum
anti-rubella
antibody
concentration in
children or
adolescents
Granumetal. {,
2013,1937228}
Medium
Human, male and female;
PFOS concentrations in the
mother at delivery and anti-
rubella antibody
concentrations at age 3
BMDLo
1.6 ng/mL
2.79 x 10 PFOS concentrations may be
influenced by pregnancy
hemodynamics; BMRs ofZi
or 1 SD provide reasonably
good estimates of 5% and
10% extra risk; selected
BMDL was based on a
significant regression
parameter
Zhang et al. {,
2023,10699594}
Medium
Human, male and female;
PFOS concentrations and
anti-rubella antibody
concentrations at ages 12-19
BMDLo
24.3 ng/mL
4.31 x 10 6 Selected BMDL was based on
a significant regression
parameter; BMRs of Vi or
1 SD may not be reasonably
good estimates of 5% and
10% extra risk
Decreased PFC
response to SRBC
Zhong et al. {,
2016,3748828}
Medium
C57BL/6 Mice, PNW 4 Fl BMDL, SD,
males Hill
1.8 mg/L 2.88 x 10 4 Selected model showed
Cavg_pup_gest_iact adequate fit (p > 0.1) and
presented most protective
4-40
-------
APRIL 2024
Endpoint
Reference,
Confidence
Strain/Species/Sex/Age
POD Type,
Model
POD Internal
Dose/Internal
Dose Metric3
PODhed
(mg/kg/day)
Notes on Modeling
BMDL associated with the
effect in a sensitive lifestage;
AICs from all models were
comparable
Extramedullar
Hematopoiesis in
the Spleen
NTP {,2019,
5400978}
High
Sprague-Dawley Rats,
female, adults
BMDLiord,
Multistage
Degree 1
2.27 mg/L
Clast7,avg
2.91 >
< 10~4
Selected model showed
adequate fit (p > 0.1) and
presented most protective
BMDL; all BMDLs from
adequate fitting models were
comparable
NTP {,2019,
5400978}
High
Sprague-Dawley Rats, male,
adults
BMDLiord,
Logistic
9.59 mg/L
Clast7,avg
1.23 >
< 10~3
Selected model showed
adequate fit (p > 0.1) and
lowest AIC
Developmental Effects
Decreased Birth
Weight
Chuetal. {,2020,
6315711}
High
Human, male and female;
PFOS serum concentrations
in third trimester
BMDL5RD,
Hybrid
7.3 ng/mL
1.27 >
< 10~6
PFOS concentrations may be
influenced by pregnancy
hemodynamics; selected
BMDL based on significant
regression parameter
Sagiv et al. {, 2018,
4238410}
High
Human, male and female;
PFOS serum concentrations
in first and second trimesters
BMDL5RD,
Hybrid
41.0 ng/mL
6.00 >
< 10~6
Selected BMDL based on
non-significant regression
parameter
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 >
< 10~7
PFOS concentrations may be
influenced by pregnancy
hemodynamics; selected
BMDL based on non-
significant regression
parameter
Wikstrom et al. {,
2020,6311677}
High
Human, male and female;
PFOS serum concentrations
in first and second trimesters
BMDL5RD,
Hybrid
7.7 ng/mL
1.13 >
< 10~6
Selected BMDL based on
significant regression
parameter
Darrow et al. {,
2013,2850966}
High
Human, male and female,
maternal PFOS serum
BMDL5RD,
Hybrid
17.4 ng/mL
2.51 >
< 10~6
Modeled based on first
prospective birth analysis
(i.e., PFOS concentrations
4-41
-------
APRIL 2024
Endpoint
Reference,
Confidence
Strain/Species/Sex/Age
POD Type,
Model
POD Internal
Dose/Internal
Dose Metric3
PODhed
(mg/kg/day)
Notes on Modeling
concentrations taken at time
of enrollment in C8 projectd
measured prior to pregnancy);
selected BMDL based on
significant regression
parameter
Yaoetal. {,2021,
9960202}
High
Human, male and female;
PFOS serum concentrations
in third trimester
BMDL5RD,
Hybrid
5.0 ng/L
8.70 >
< 10~7
PFOS concentrations may be
influenced by pregnancy
hemodynamics; selected
BMDL based on non-
significant regression
parameter
Decreased Fetal
Body Weight
Lee et al. {, 2015,
2851075}
Medium
CD-I Mice, Fi males and
females (GD 17)
NOAELe
(0.5 mg/kg/day)
8.75 x 10 1 mg/L
Cavg_pup gest
3.40 >
< 10~4
No models had adequate fit
(residuals at BMD or control
were greater than 2, or the
BMDL was 3x lower than the
lowest tested dose); NOAEL
approach taken
Decreased Pup
Body Weight
Luebker et al. {,
2005,757857}
Medium
Sprague-Dawley Rats, Fi
male and female (LD 1)
BMDL5RD,
Exponential 3
14.7 mg/L
Cavg_pup gest
5.71 >
< 10~3
Selected model showed
adequate fit (p > 0.1) and
lowest AIC
Luebker et al. {,
2005,757857}
Medium
Sprague-Dawley Rats, Fi
male and female (LD 5)
BMDL5RD,
Polynomial
Degree 6
2.30 mg/L
Cavg_pup gest lact
3.65>
<10~4
Selected model showed
adequate fit (p > 0.1) and
lowest AIC
Luebker et al. {,
2005,1276160}
Medium
Sprague-Dawley Rats, Fi
male and female (LD 1)
BMDL5RD,
Exponential 4
11.3 mg/L
Cavg_pup gest
4.39 >
< 10-3
Selected model showed
adequate fit (p > 0.1) and
lowest AIC
Decreased Pup
Survival
Lau et al. {, 2003,
757854}
Medium
Sprague-Dawley Rats, Fi
male and female (PND 5)
NOAELe
(1 mg/kg/day)
13.0 mg/L
Cavg_pup gest lact
2.06 >
< 10~3
No models had adequate fit
(for all models, all model
control response SD was 1.5x
greater than actual response
SD, and for most models, the
calculated BMD was 3x
lower than the lowest
administered dose); NOAEL
approach taken
4-42
-------
APRIL 2024
Endpoint
Reference,
Confidence
Strain/Species/Sex/Age
POD Type,
Model
POD Internal
Dose/Internal
Dose Metric3
PODhed
(mg/kg/day)
Notes on Modeling
Lau et al. {, 2003,
757854}
Medium
Sprague-Dawley Rats, Fi
male and female (PND 22)
NOAELe 17.3 mg/L 2.75 x 10 3 No models had adequate fit
(1 mg/kg/day) C,n2 pll|, gest_lact (for all models, all model
control response SD was 1.5x
greater than actual response
SD, and for most models, the
calculated BMD was 3x
lower than the lowest
administered dose); NOAEL
approach taken
Cardiovascular Effects (Serum Lipids)
Increased Total
Cholesterol
Dong et al. {,
5080195}
Medium
2019, Human, male and female, BMDL5rd,
age 20-80 Hybrid
9.34 ng/mL 1.20 x 10~6 BMDL based on analyses
excluding individuals
prescribed cholesterol
medication and significant
regression parameter
Steenland et al. {,
2009,1291109}
Medium
Human, male and female, BMDL5rd,
age 18 and older Hybrid
9.52 ng/mL 1.22 x 10~6 BMDL based on analyses
excluding individuals
prescribed cholesterol
medication and significant
regression parameter
Lin et al. {, 2019,
5187597}
Medium
Human, male and female, BMDL5rd,
age 25 and older Linear
66.5 ng/mL 8.51 x 10~6 BMDL based on analyses
including individuals
prescribed cholesterol
medication and non-
significant regression
parameter
Hepatic Effects
Elevated ALT
Gallo et al. {,
2012, 1276142}
Medium
Human, female, age 18 and BMDL5rd,
older Hybrid
56.8 ng/mL 7.27 x 10 6 BMDL based on significant
regression parameter
Nianetal. {,2019,
5080307}
Medium
Human, female, age 22 and BMDL5rd,
older Hybrid
15.1 ng/mL 1.94 x 10 6 BMDL based on significant
regression parameter
4-43
-------
APRIL 2024
Endpoint
Reference,
Confidence
Strain/Species/Sex/Age
PODhed
(mg/kg/day)
Notes on Modeling
Increased
Individual Cell
Necrosis in the
Liver
Butenhoff et al. {, Sprague-Dawley Rats,
2012, 1276144}/ females, adults
Thomford {, 2002,
BMDLiord, 27.0 mg/L
Log-Logistic Clast7,avg
3.45 x io 3 Selected model showed
5029075}
High
f
adequate fit (p > 0.1) and
lowest AIC among models
with BMD/BMDL ratio < 3
Notes: ALT = alanine aminotransferase; AUC = area under the curve; BMDL0.5 sd = lower bound on the dose level corresponding to the 95% lower confidence limit for a change
in the mean response equal to 0.5 SD from the control mean; BMDLi sd = lower bound on the dose level corresponding to the 95% lower confidence limit for a change in the
mean response equal to 1 SD 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,avg = average 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 Appendix {U.S. EPA, 2024, 11414344} 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.
eNo models provided adequate fit; therefore, a NOAEL/LOAEL approach was selected.
f Butenhoff et al. {, 2012, 1276144} and Thomford {, 2002, 5029075} reported the same data.
4-44
-------
APRIL 2024
4.1.4.1 Hepatic Effects
Increased ALT in individuals aged 18 and older {Gallo, 2012,1276142} or 22 and older
{Nian,2019, 5080307}
The POD for increased ALT in adults was derived by quantifying a benchmark dose using a
hybrid modeling approach (see Appendix E. 1, {U.S. EPA, 2024, 11414344}) 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.
A BMR of 5% extra risk was chosen per EPA's Benchmark Dose Technical Guidance {U.S.
EPA, 2012, 1239433} (Section 4.1.2) The internal serum POD was converted to an external dose
(PODhed), in mg/kg/day (Section 4.1.3.2). 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 was 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,avg {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 per EPA's Benchmark Dose Technical Guidance {U.S. EPA,
2012, 1239433} (Section 4.1.2). The Ciast7,avg was selected for all non-developmental studies
rather than alternate metrics such as Cmax to provide a consistent internal dose for use across
chronic and subchronic study designs where steady state may or may not have been reached and
to allow extrapolation to the human PK model (Section 4.1.3.1.3). 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 (Section
4.1.3.2). This calculation was 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 Appendix E.l, {U.S. EPA, 2024, 11414344}) on the measured PFOS serum
concentrations at age 5, which provided an internal serum concentration POD in mg/L. A BMR
of 0.5 SD was chosen per EPA's Benchmark Dose Technical Guidance {U.S. EPA, 2012,
1239433} (Section 4.1.2). The internal serum POD was converted to an external dose (PODhed),
in mg/kg/day, using the updated Verner model (described in 4.1.3.2). For this, the model was run
starting at the birth of the mother, with constant exposure relative to body weight. 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 body weight, is set to the same value
as the mother. The model provides predictions up to a child age of 5 years, when the serum
4-45
-------
APRIL 2024
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 each. The slightly lower HED in males was
then selected as it was the most health protective.
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 Appendix E, {U.S. EPA, 2024, 11414344}) 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. A BMR of 0.5 SD was chosen per EPA's Benchmark Dose Technical Guidance (U.S.
EPA, 2012, 1239433} (Section 4.1.2). The internal serum POD was converted to an external
dose (PODhed), in mg/kg/day, using the updated Verner model (described in Section 4.1.3.2).
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 Appendix E, (U.S. EPA, 2024, 11414344}) on the measured PFOS serum
concentrations at ages 7-12, which provided an internal serum concentration POD in mg/L. A
BMR of 0.5 SD was chosen per EP A's Benchmark Dose Technical Guidance (U.S. EPA, 2012,
1239433} (Section 4.1.2). The internal serum POD was converted to an external dose (PODhed),
in mg/kg/day, using the updated Verner model (described in Section 4.1.3.2). For this, the model
was run similarly to the endpoint based on antibodies at age 7 {Budtz-j0rgensen, 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 different growth curves specific to male and female
children were used in the model, the model predicted slightly different (less than 5%) serum
concentrations for each sex. The lower HED was then selected as it was the most health
protective.
Decreased Rubella antibody response in vaccinated adolescents at ages 12-19 {Zhang,
2023,10699594}
The POD for decreased antibody production in adolescents aged 12-19 was derived by
quantifying a benchmark dose (see Appendix E, {U.S. EPA, 2024, 11414344}) on the measured
PFOS serum concentrations at ages 12-19, which provided an internal serum concentration POD
in mg/L. A BMR of 0.5 SD was chosen per EPA's Benchmark Dose Technical Guidance {U.S.
EPA, 2012, 1239433} (Section 4.1.2). The internal serum POD was converted to an external
4-46
-------
APRIL 2024
dose (PODhed), in mg/kg/day, using the updated Verner model (described in Section 4.1.3.2).
For this, the model was ran similarly to the endpoint based on antibodies at age 7 (Budtz-
J0rgensen, 2018, 5083631}, but the model was run until the median age of this cohort at blood
collection, 15.5 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 Rubella antibody response in vaccinated children at age 3 {Granum, 2013,
1937228}
The POD for decreased antibody production at age 3 was derived by quantifying a benchmark
dose (see Appendix E, {U.S. EPA, 2024, 11414344}) on the measured PFOS serum
concentrations collected from the mother at delivery, which provided an internal serum
concentration POD in mg/L. A BMR of 0.5 SD was chosen per EPA's Benchmark Dose
Technical Guidance (U.S. EPA, 2012, 1239433} (Section 4.1.2). The internal serum POD was
converted to an external dose (PODhed), in mg/kg/day, using the updated Verner model
(described in Section 4.1.3.2). 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 at
the time of delivery. Reverse dosimetry was performed on model-predicted maternal serum
concentration at that time to calculate the PODhed. This metric was independent of the sex of the
child in the model.
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.
Using the Wambaugh et al. {, 2013, 2850932} model, daily exposure to PFOS through oral
gavage was simulated from GD 1-GD 17 using female CD1 mice parameters (C57BL/6 mice
parameters are not available for PFOS; Section 4.1.3.1). The Cavg,PuP,gest_iact internal dose metric
was selected for this model since an average concentration metric is expected to better correlate
with this developmental effect that may have resulted from exposure during gestation or lactation
(Section 4.1.3.1.3). Continuous models were used to fit dose-response data. 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} (Section 4.1.2). 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 4.1.3.2). For
this, the model was run starting at the birth of the mother, with constant exposure relative to
body weight. 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 was 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 was 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.
4-47
-------
APRIL 2024
Extramedullar hematopoiesis in the spleen, Sprague-Dawley Rats, female and male,
Ciast7,avg {NTP, 2019, 5400978}
Increased incidence of extramedullary hematopoiesis in the spleen was observed in male and
female 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 (Section 4.1.3.1). The Ciast7,avg was selected for all non-developmental studies rather
than alternate metrics such as Cmax to provide a consistent internal dose for use across chronic
and subchronic study designs where steady state may or may not have been reached and to allow
extrapolation to the human PK model (Section 4.1.3.1.3). 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} (Section 4.1.2). 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 (Section 4.1.3.2). This
calculation was 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 Appendix E, (U.S. EPA, 2024, 11414344}) 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. A BMR of 5% extra risk was chosen per EPA's Benchmark Dose Technical Guidance
(U.S. EPA, 2012, 1239433} (Section 4.1.2). The internal serum POD was converted to an
external dose (PODhed), in mg/kg/day (Section 4.1.3.2). 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 was 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 Appendix E, (U.S. EPA, 2024, 11414344}) 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. A BMR of 5% extra risk was chosen per EPA's Benchmark Dose
Technical Guidance (U.S. EPA, 2012, 1239433} (Section 4.1.2). The internal serum POD was
converted to an external dose (PODhed), in mg/kg/day. Specifically, the PODhed was calculated
as the external dose (in mg/kg/day) that would result in a steady-state serum concentration equal
to the internal serum POD (Section 4.1.3.2). This calculation was the POD multiplied by the
selected human clearance value (0.128 mL/kg/day; calculated from half-life and volume of
distribution; CI = Yd * ln(2)/ti/2)).
4-48
-------
APRIL 2024
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 BMDS
(see Appendix E, {U.S. EPA, 2024, 11414344}) from 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 DPP and Outcomes Study (DPPOS), which provided an internal serum
concentration POD in mg/L. A BMR of 0.5 SD extra risk was chosen per EPA's Benchmark
Dose Technical Guidance {U.S. EPA, 2012, 1239433} (Section 4.1.2). The internal serum POD
was converted to an external dose (PODhed), in mg/kg/day (Section 4.1.3.2). 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 was 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 Appendix E, {U.S. EPA, 2024, 11414344}) 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. A BMR of 5% extra risk was chosen per EPA's Benchmark Dose Technical Guidance
{U.S. EPA, 2012, 1239433} (Section 4.1.2). The internal serum POD was converted to an
external dose (PODhed), in mg/kg/day, using the updated Verner model (described in Section
4.1.3.2). 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 body weight.
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 the first
and second trimesters {Sagiv, 2018, 4238410}
The POD for decreased birth weight in infants was derived by quantifying a benchmark dose
using a hybrid modeling approach (see Appendix E, {U.S. EPA, 2024, 11414344}) on the
measured PFOS serum concentrations collected from the mother primarily in the first trimester
(median gestational age of 9 weeks), which provided an internal serum concentration POD in
mg/L. A BMR of 5% extra risk was chosen per EPA's Benchmark Dose Technical Guidance
{U.S. EPA, 2012, 1239433} (Section 4.1.2). The internal serum POD was converted to an
external dose (PODhed), in mg/kg/day, using the updated Verner model (described in Section
4.1.3.2). 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
4-49
-------
APRIL 2024
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 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 Appendix E, {U.S. EPA, 2024, 11414344}) on the
measured PFOS serum concentrations collected from the mother in trimesters 2 and 3 (median
gestational age of 27 weeks), which provided an internal serum concentration POD in mg/L. A
BMR of 5% extra risk was chosen per EPA's Benchmark Dose Technical Guidance {U.S. EPA,
2012, 1239433} (Section 4.1.2). The internal serum POD was converted to an external dose
(PODhed), in mg/kg/day, using the updated Verner model (described in Section 4.1.3.2). 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 Appendix E, {U.S. EPA, 2024, 11414344}) 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. A BMR of 5% extra risk was chosen per EPA's Benchmark Dose Technical Guidance
{U.S. EPA, 2012, 1239433} (Section 4.1.2). The internal serum POD was converted to an
external dose (PODhed), in mg/kg/day, using the updated Verner model (described in Section
4.1.3.2). 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 Appendix E, {U.S. EPA, 2024, 11414344}) 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. A BMR of 5% extra risk was chosen per EPA's Benchmark Dose
Technical Guidance {U.S. EPA, 2012, 1239433} (Section 4.1.2). The internal serum POD was
4-50
-------
APRIL 2024
converted to an external dose (PODhed), in mg/kg/day, using the updated Verner model
(described in Section 4.1.3.2). 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 body weight. 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 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 Appendix E, {U.S. EPA, 2024, 11414344}) on the
measured PFOS serum concentrations collected from the mother prior to conception, which
provided an internal serum concentration POD in mg/L. A BMR of 5% extra risk was chosen per
EPA's Benchmark Dose Technical Guidance (U.S. EPA, 2012, 1239433} (Section 4.1.2). The
internal serum POD was converted to an external dose (PODhed), in mg/kg/day, using the
updated Verner model (described in 4.1.3.2). This was performed as described for the Chu et al.
{, 2020, 6315711} study. In the selected cohort, blood samples were taken from women before
conception. Therefore, 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, Cavg_pup_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 a 5% change from the control
mean was chosen per EPA's Benchmark Dose Technical Guidance {U.S. EPA, 2012, 1239433}
(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 Appendix E, {U.S. EPA, 2024, 11414344}). The
Cavg,pup,gest internal dose metric was selected for this model since an average concentration metric
is expected to better correlate with this developmental effect that may have resulted from
exposure any time during gestation (Section 4.1.3.1.3). 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 4.1.3.2). For this
endpoint, the model was run starting at the birth of the mother, with constant exposure relative to
body weight. 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.
4-51
-------
APRIL 2024
Decreased Pup Body Weight, Sprague-Dawley Rats, Fi male and female (LD 5),
Cavg pup gest lact {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
a 5% change from the control mean was chosen per EP A's Benchmark Dose Technical Guidance
{U.S. EPA, 2012, 1239433} (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 Appendix E, (U.S.
EPA, 2024, 11414344}). The Cavg,PuP,gest_iact internal dose metric was selected for this model since
an average concentration metric is expected to better correlate with this developmental effect that
may have resulted from exposure during gestation or lactation (Section 4.1.3.1.3). 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 4.1.3.2). For this, the model
was run starting at the birth of the mother, with constant exposure relative to body weight.
Pregnancy began at 24.25 years maternal age and birth occurred at 25 years maternal age. The
initial concentration in the child was governed by the observed ratio between maternal serum and
cord blood at delivery. Then the model was run through the entire 1-year breastfeeding period
Then the model was run through the entire 1-year breastfeeding period because the lactational
duration in humans that equates to lactational day 5 in rodents is unknown. Additionally, there is
currently no mechanistic information to identify a specific window of susceptibility in lactation
for this endpoint. The average serum concentration in the infant through gestation and lactation
was determined for this scenario and reverse dosimetry was used to calculate the exposure that
results in the same value as the POD. Because of different growth curves used for male and
female children, the model predicted slightly different serum concentrations for males and
females. The lower HED was selected to be more health protective.
Decreased Pup Body Weight, Sprague-Dawley Rats, Fi male and female (LD 1),
Cavg pup gest {Luebker, 2005,1276160; Luebker, 2005, 757857}
Decreased mean pup body weight relative to the litter at LD 1 (the day of birth) was observed in
Fi male and female Sprague-Dawley rats in 1-generation and 2-generation reproductive studies.
Continuous models were used to fit dose-response data. A BMR of a 5% change from the control
mean was chosen per EPA's Benchmark Dose Technical Guidance {U.S. EPA, 2012, 1239433}
(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 Appendix E, {U.S. EPA, 2024, 11414344}). The
Cavg,pup,gest internal dose metric was selected for this model since an average concentration metric
is expected to better correlate with this developmental effect that may have resulted from
exposure any time during gestation (Section 4.1.3.1.3). 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 4.1.3.2). For this, the model was run starting at the birth of the
mother, with constant exposure relative to body weight. 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.
4-52
-------
APRIL 2024
Decreased Pup Survival, Sprague-Dawley Rats, Fi male and female (PND 5 and 22),
Cavg pup gest lact {Lau, 2003, 757854}
Decreased pup survival at PND 5 and PND 22 was observed in Fi male and female Sprague-
Dawley rats. Continuous models were used to fit dose-response data. A BMR of 0.5 SD was
chosen per EPA's Benchmark Dose Technical Guidance {U.S. EPA, 2012, 1239433} (Section
4.1.2) and a BMR of a change in the mean equal to 0.1 standard deviations from the control
mean was provided for comparison purposes because decreased pup survival is a severe, frank
effect (U.S. EPA, 2012, 1239433} (see Appendix E, (U.S. EPA, 2024, 11414344}). The
Cavg,pup,gestjact internal dose metric was selected for this model since an average concentration
metric is expected to better correlate with this developmental effect that may have resulted from
exposure during gestation or lactation (Section 4.1.3.1.3). 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 4.1.3.2).
For this, the model was run starting at the birth of the mother, with constant exposure relative to
body weight. Pregnancy began at 24.25 years maternal age and birth occurred at 25 years
maternal age. The initial concentration in the child was governed by the observed ratio between
maternal serum and cord blood at delivery. Then the model was run through the entire 1-year
breastfeeding period for both timepoints because the lactational duration in humans that equates
to lactational day 5 in rodents is unknown. Additionally, there is currently no mechanistic
information to identify a specific window of susceptibility in lactation for this endpoint. The
average serum concentration in the infant through gestation and lactation was determined for this
scenario and reverse dosimetry was used to calculate the exposure that results in the same value
as the POD. Because of different growth curves used for male and female children, the model
predicted slightly different serum concentrations for males and females. The lower HED was
selected to be more health protective.
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. For epidemiological studies, similar to the
discussion of study selection factors in Section 4 and Section 4.1.1, EPA critically considered
attributes for each PODhed including timing of endpoint collection or measurement,
uncertainties associated with modeling (see Appendix E {U.S. EPA, 2024, 11414344} and Table
4-8), and consideration of confounding. For animal toxicological studies, attributes considered
included study confidence (i.e., high confidence studies were prioritized over medium confidence
studies), amenability to benchmark dose modeling, study design, sensitive lifestages, and health
effects observed after exposure in the lower dose range among the animal toxicological studies.
As described in the subsections below, this examination of epidemiological and toxicological
studies led to the exclusion of a number of studies from consideration for candidate RfD
derivation. Health outcome- and study-specific considerations are discussed in Sections 4.1.5.1
(Hepatic) 4.1.5.2 (Immune) 4.1.5.3 (Cardiovascular), and 4.1.5.4 (Developmental).
4-53
-------
APRIL 2024
Once studies and their corresponding PODheds were prioritized for candidate RfD derivation,
EPA applied uncertainty factors (UFs) according to methods described in EPA's Review of the
Reference Dose and Reference Concentration Processes {U.S. EPA, 2002, 88824}.
Considerations for individual UFs differed between epidemiological and animal toxicological
studies and are further described in Section 4.1.5.5. Presentation of the candidate RfDs for each
health outcome is provided in Section 4.1.5.6.
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 the U.S. population. Additionally, these studies had many
study strengths including sufficient study sensitivity and sound methodological approaches,
analysis, and design, as well as no evidence of bias. The two studies reported analyses examining
different forms of confounding factors, sensitivity analyses excluding participants with lifestyle
characteristics (e.g., excluding smokers, drinkers, medicine takers) impacting outcome
assessment {Nian, 2019, 5080307}, and non-linear exposure-response relationships {Gallo,
2012, 1276142}. Both studies provided the necessary data for modeling.
One high confidence animal toxicological study was carried forward for candidate RfD
determination {Butenhoff, 2012, 1276144/Thomford, 2002, 5029075}. This study was
prioritized for candidate RfD development because it was determined to be a high confidence
study, was amenable to BMD modeling, and was the only animal toxicological study with a
chronic exposure duration that histopathologically examined the liver of animals treated with
PFOS.
4.1.5.2 Immune Effects
Three medium confidence epidemiological studies were carried forward for candidate RfD
determination {Zhang, 2023, 10699594; Budtz-j0rgensen, 2018, 5083631; Timmerman, 2021,
9416315}. EPA considered all three studies as they represented the low-dose range of effects
across immunological endpoints and provided data regarding sensitive populations (i.e.,
children) across three vaccine types. 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 potentially increased risk bias due to pregnancy-related hemodynamic effects.
Similarly, EPA did not carry forward PODheds derived from Granum et al. {,2013, 1937228}
because PFOS serum concentrations were measured in the mother at the time of delivery and
therefore, this study also had potential for increased risk of bias due to pregnancy-related
hemodynamic effects. EPA also derived candidate RfDs for both tetanus and diphtheria vaccine
responses from Timmerman et al. {, 2021, 9416315} for comparison to a second population of
children. Zhang et al. {, 2023, 10699594} was also selected for candidate RfD derivation
because it provided results in adolescents from the U.S. population for a third vaccine type (i.e.,
rubella). Additionally, the BMDL derived from this study was based on a significant regression
4-54
-------
APRIL 2024
parameter. In total, five immunological PODheds from three epidemiological studies were
carried forward for candidate RfD derivation.
Two animal toxicological studies, one high and one medium confidence, 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 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 encompassed a
sensitive population that is coherent with the developmental immunotoxicity observed in
humans. For these reasons, EPA determined that both of these effects warranted candidate RfD
derivation.
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
PODheds were derived, Dong et al. {, 2019, 5080195} and Steenland et al. {, 2009, 1291109}
excluded individuals who were prescribed cholesterol medication, minimizing concerns
surrounding confounding due to the medical intervention altering serum total cholesterol levels.
This is in contrast to Lin et al. {, 2019, 5187597} which did not control for individuals
prescribed cholesterol medication and was therefore excluded from further consideration.
Modeling of both Dong et al. {, 2019, 5080195} and Steenland et al. {, 2009, 1291109} resulted
in PODheds with minimal risk of bias, representing both the general population and a high-
exposure community, respectively and thus were both considered further for candidate RfD
derivation.
4.1.5.4 Developmental Effects
Three high confidence epidemiological studies were carried forward for candidate RfD
determination for the endpoint of decreased birth weight {Sagiv, 2018, 4238410; Wikstrom,
2019, 6311677; Darrow, 2013, 2850966}. Of the six epidemiological studies for which PODheds
were derived, Darrow et al. {, 2013, 2850966}, Sagiv et al. {, 2018, 4238410}, and Wikstrom et
al. {, 2019, 6311677} assessed maternal PFOS serum concentrations either prior to conception or
primarily in the first trimester, minimizing concerns surrounding bias due to pregnancy-related
hemodynamic effects. Although Wikstrom et al. {, 2020, 6311677} collected approximately 4%
of samples during early weeks of the second trimester, sensitivity analyses showed no
differences when trimester two samples were excluded. Additionally, these studies had many
study strengths including sufficient study sensitivity and sound methodological approaches,
analysis, and design, as well as no evidence of bias and reflected two different study populations.
Therefore, all three studies were considered further for candidate RfD derivation. The three
excluded studies assessed PFOS concentrations in either umbilical cord blood or primarily
4-55
-------
APRIL 2024
during the second or third trimesters, increasing the uncertainty associated with the derived
PODheds due to potential pregnancy-related hemodynamic effects, and as a result, were
excluded from consideration for candidate RfD derivation {Chu, 2020, 6315711; Starling, 2017,
3858473; Yao, 2021, 9960202}.
One medium confidence animal toxicological study was carried forward for candidate RfD
determination {Luebker, 2005, 757857}. The endpoint of reduced pup weight at LD 5 from this
study was amenable to benchmark dose modeling (i.e., BMD modeling produced viable model
fits), unlike the endpoints of decreased fetal weight reported by Lee et al. {, 2015, 2851075} and
decreased pup survival reported by Lau et al. {, 2003, 757854}, which had NOAELs as the basis
of the PODheds. Decreased pup weight at LD 5 was selected over the other time point reported
by Luebker et al. {, 2005, 757857} (i.e., LD 1) and decreased pup weight reported by Luebker et
al. {, 2005, 1276160} (also LD 1) because it was the most protective of the three PODheds, all of
which were derived from BMDLs. The endpoint of decreased pup weight reported by Luebker et
al. {, 2005, 757857} encompassed a sensitive population and was coherent with the observed
effect of decreased birth weight in humans and was therefore selected for candidate RfD
derivation.
4.1.5.5 Application of Uncertainty Factors
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 lifestage 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 lifestages and populations and allow for a complete characterization of the
contaminant's toxicity.
UFC
10
CompositeUFC = UFA x UFH XUFSX 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.
4-56
-------
APRIL 2024
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, lifestage, 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
lifestage 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 observed in children and adolescents because exposure is
assumed to occur from gestation through childhood, when the response variable was measured.
There is uncertainty regarding the critical window of exposure that results in these immune
effects in children and adolescents. Therefore, EPA expects that any exposure during this period
of development has the potential to impact this response {U.S. EPA, 1991, 732120}. According
to the WHO/International Programme on Chemical Safety (IPCS) Immunotoxicity Guidance for
Risk Assessment, developmental immunotoxicity is assessed during the prenatal, neonatal,
juvenile and adolescent life stages because immune system development occurs throughout these
life stages and should be viewed differently in part due to increased susceptibility compared with
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 large number of medium and high confidence
epidemiological studies 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.
4-57
-------
APRIL 2024
The composite UF 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 UF 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
lor 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 lifestage 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 lifestages 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 threefold 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, lifestage, 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
was used as the basis for the PODhed derivation. BMDLs were available for all animal
toxicological endpoints and studies advanced for candidate RfD derivation.
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
4-58
-------
APRIL 2024
was applied to developmental endpoints because the developmental period is recognized as a
susceptible lifestage 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 large number of medium and high confidence epidemiological
studies 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.
In summary, the composite UF that was applied to candidate RfDs derived from all of the
epidemiological studies were the same value (UFc = 10) (Table 4-9). The composite UF that was
applied to candidate RfDs derived from animal toxicological studies was either UFc = 30 or 300
(Table 4-10). In all of these cases, the total uncertainty is well below the maximum
recommended UFc = 3,000 {U.S. EPA, 2002, 88824}.
4.1.5.6 Candidate RfDs
Table 4-11 shows the UFs applied to each candidate study to subsequently derive the candidate
RfDs.
4-59
-------
APRIL 2024
Table 4-11. Candidate Reference Doses (RfDs)
Endpoint
Reference,
Confidence
Strain/Species/
Sex/Age
PODheo IJFa UFh UFs UFl UFd UFtot
(mg/kg/day)
Candidate RfDa
(mg/kg/day)
Immune Effects
Human, male and 2.71 x 10 6 1 10 1 1 1 10 2.71 x 10 7 = 3 x 10 7
female, PFOS
concentrations at age 5
and antibody
concentrations at age 7
Human, male and 1.78 x 10 6 1 10 1 1 1 10 1.78 x 10"7 = 2 x 10"7
female, PFOS and
antibody concentrations
at age 7-12
Human, male and 1.83 x 10 6 I 10 I I I 10 1.83 x 10 7 = 2 x 10~7
female, PFOS
concentrations at age 5
and antibody
concentrations at age 7
Human, male and 1.03 x 10 6 1 10 1 1 1 10 1.03 x 10"7 = 1 x 10"7
female, PFOS and
antibody concentrations
at age 7-12
Decreased Serum Anti-
Rubella Antibody
Concentration in
Adolescents
Zhang et al. {, 2023,
10699594}
Medium
Human, male and
female, PFOS and
antibody concentrations
at age 12-19
4.31 >
< 10~6 1
10
1
1
1
10
4.31 >
< 10~7 = 4 >
< 10~7
Decreased Plaque
Forming Cell (PFC)
Response to SRBC
Zhong et al. {, 2016,
3748828}
Medium
C57BL/6 Mice, PNW 4
Fi males
2.88 >
< 10~4 3
10
1
1
1
30
9.60 >
< 10~6 = 1 >
< 10~5
Extramedullary
Hematopoiesis in the
Spleen
NTP {, 2019, 5400978} Sprague-Dawley rats,
High female, adults
2.91 >
< 10~4 3
10
10
1
1
300
9.70 >
< 10~7 = 1 >
< 10~6
Developmental Effects
Decreased Birth
Weight
Sagiv et al. {, 2018,
4238410}
High
Human, male and
female, PFOS
concentrations in first
and second trimesters
6.00 >
< 10~6 1
10
1
1
1
10
6.00 >
< 10~7 = 6 >
< 10~7
Decreased Serum Anti-
Tetanus Antibody
Concentration in
Children
Budtz-Jorgensen and
Grandjean {, 2018,
5083631}
Medium
Timmerman et al. {,
2021,9416315}
Medium
Decreased Serum Anti- Budtz-Jorgensen and
Diphtheria Antibody Grandjean {, 2018,
Concentration in 5083631}
Children Medium
Timmerman et al. {,
2021,9416315}
Medium
4-60
-------
APRIL 2024
Endpoint
Reference,
Confidence
Strain/Species/
Sex/Age
PODhed
(mg/kg/day)
UFa
UFh
UFs
UFl
UFd
UFtot
Candidate RfDa
(mg/kg/day)
Wikstrom et al. {,
Human, male and
1.13
X 10-6
1
10
1
1
1
10
1.13 x 10-7= 1 x l(T7
2019,6311677}
female, PFOS
High
concentrations in first
and second trimesters
Darrow et al. {, 2013,
Human, male and
2.51 x
10~6
1
10
1
1
1
10
2.51 x 10 7 = 3 x 10~7
2850966}
female, PFOS
High
concentrations at time
of enrollment13
Decreased Pup Body
Luebker et al. {, 2005,
Sprague-Dawley Rats,
3.65
X 10~4
3
10
1
1
1
30
1.22 x 10-5= 1 x 10-5
Weight
757857}
Medium
Fi male and female
(LD 5)
Cardiovascular Effects
Increased Serum Total
Dong et al. {,2019,
Human, male and
1.20
X 10~6
1
10
1
1
1
10
1.20 x 10-7 = 1 x l(T7
Cholesterol
5080195}
Medium
female, ages 20-80
Steenland et al. {, 2009, Human, male and
1.22
X 10~6
1
10
1
1
1
10
1.22 x 10-7 = 1 x l(T7
1291109}
female, age 18 and
Medium
older
Hepatic Effects
Increased Serum ALT
Galloetal. {,2012,
1276142}
Medium
Human, female, age 18
and older
7.27
X 10~6
1
10
1
1
1
10
7.27 x 10-7 = 7 x 10-7
Nianetal. {,2019,
Human, female, at age
1.94
X 10~6
1
10
1
1
1
10
1.94 x 10~7 = 2 x 10-7
5080307}
22 and older
Medium
Individual Cell
Butenhoff et al. {,
Sprague-Dawley rats,
3.45
X 103
3
10
1
1
1
30
1.15 x 10-4= 1 x l(T4
Necrosis in the Liver
2012,
1276144}/Thomford {,
2002, 5029075}°
High
females, adults
Notes: ALT = alanine transaminase; LTFa = interspecies uncertainty factor; LTFd = database uncertainty factor; LTFh = intraspecies uncertainty factor; LTFs = subchronic-to-chronic
extrapolation uncertainty factor; LTFl = extrapolation from a LOAEL to a NOAEL uncertainty factor; UFtot = composite uncertainty factor.
aRfDs were rounded to one significant figure.
b 99% of the pregnancies of participants in Darrow et al. {, 2013, 2850966} were within 3 years of the serum PFOS measurement.
c Butenhoff et al. {, 2012, 1276144} and Thomford {, 2002, 5029075} reported data from the same experiment.
4-61
-------
APRIL 2024
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 noncancer 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 7mg/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 or 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 Sections 5.1, 5.6, and 5.7), 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 and sex-specific toxicokinetic differences in rodent species These uncertainties may
explain, in part, the higher magnitude of candidate RfDs derived from animal toxicological
studies compared to the candidate RfDs derived from epidemiological studies. Moreover, the
human epidemiological studies also have greater relevance to human exposure than animal
toxicological studies 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 5.2 for further comparisons between toxicity values derived
from epidemiological and animal toxicological studies.
4-62
-------
APRIL 2024
Decreased serum
anti-tetanus antibody
concentration in children
Decreased serum
anti-diptheria antibody
concentration in children
Decreased serum
anti-rubella antibody
concentration in adolescents
Timmerman,
2021, 9416315;
Medium confidence
Budtz-Jargensen,
2018, 5083631;
Medium confidence
Timmerman,
2021, 9416315;
Medium confidence
Budtz-Jorgensen,
2018, 5083631;
Medium confidence
Zhang, 2023, 10699594;
Medium confidence
Extramedullar
hematopoiesis in the spleen
NTP.2019, 5400978;
High confidence
Decreased PFC response
to SRBC
Zhong, 2016, 3748828;
Medium confidence
-o
-o
Human Animal
RfD PODHED
-O
-O
-o
-o
-o
Sagiv, 2018, 4238410;
High confidence
Wikstrdm, 2020,
Decreased Birth Weight
6311677;
High confidence
Darrow, 2013, 2850966;
High confidence
Decreased Pup Body
Luebker, 2005, 757857;
Weight
Medium confidence
•—o
•—o
•—o
• o
Increased Serum Total
Cholesterol
Dong, 2019, 5080195;
Medium confidence
Steenland,
2009, 1291109;
Medium confidence
•—o
•—o
Increased Serum ALT
Gallo, 2012, 1276142;
Medium confidence
Nian, 2019, 5080307;
Medium confidence
•—o
•—o
Butenhoff, 2012,
Individual Cell Necrosis in 1276144/ Thomford, W O
the Liver 2002, 5029075;
High confidence
108 10-7 10-6 10-5 10-4 10-3 10-2
PFOS Concentration (mg/kg-d)
Figure 4-3. Comparison of Candidate RfDs Resulting from the Application of Uncertainty
Factors to PODheds Derived from Epidemiological and Animal Toxicological Studies
4-63
-------
APRIL 2024
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
At least two candidate RfDs were derived from epidemiological studies for each of the four
prioritized noncancer health outcomes. EPA considered several factors when selecting health
outcome-specific RfDs, including relevance of exposure or population characteristics to the
general population, potential confounding factors, and characteristics of the modeled data. Health
outcome- and study-specific considerations are discussed in Sections 4.1.6.1.1 (Hepatic),
4.1.6.1.2 (Immune), 4.1.6.1.3 (Cardiovascular), and 4.1.6.1.4 (Developmental), below.
4.1.6.1.1 Hepatic Effects
Two medium confidence epidemiological studies were selected for candidate RfD derivation for
the endpoint of increased 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
candidate 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 in the
population of interest. The resulting health outcome-specific RfD is 2 x 10 7 mg/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.
4.1.6.1.2 Immune Effects
Candidate RfDs were derived from three medium confidence epidemiological studies for the
endpoint of decreased antibody production in response to various vaccinations in children
{Budtz-j0rgensen, 2018, 5083631; Timmerman, 2021, 9416315; Zhang, 2023, 10699594}.
Candidate RfDs derived from Timmerman et al. {, 2021, 9416315} were considered lower
confidence candidate RfDs than those derived from Budtz-j0rgensen and Grandjean {,2018,
5083631}. PODheds derived from Timmerman et al. {, 2021, 9416315} were considered to have
increased uncertainty compared with Budtz-j0rgensen and Grandjean {, 2018, 5083631} due to
two features of the latter study that strengthen the confidence in the PODheds: 1) the response
reported by this study was more precise in that it reached statistical significance, and 2) the
analysis considered co-exposures of other PFAS. Therefore, the candidate RfDs from
Timmerman et al. {, 2021, 9416315} were not considered for selection as the health outcome-
specific RfD. Similarly, the candidate RfD derived from Zhang {, 2023, 10699594} was also not
considered since the analysis did not consider co-occurring PFAS and the resulting health
outcome-specific RfD would be less protective.
The RfD 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
4-64
-------
APRIL 2024
RfD for immune effects because the PODhed were based on models with adequate quality of fit
and significant regression parameters, the analysis considered co-exposures of other PFAS and
indicated minimal potential for confounding in the value of the PODhed due to PFOA, and the
response was more consistently observed across the two time points reported in the study
between 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.
4.1.6.1.3 Cardiovascular Effects
Two medium confidence epidemiological studies were selected for candidate 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 theNHANES population (2003-2014), while
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 candidate 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 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 the same.
4.1.6.1.4 Developmental Effects
Three high confidence epidemiological studies were considered for candidate RfD derivation for
the endpoint of decreased birth weight {Sagiv, 2018, 4238410; Wikstrom, 2019, 6311677;
Darrow, 2013, 2850966}. These candidate studies assessed maternal PFOS serum concentrations
before birth {Darrow, 2013, 2850966} or primarily in the first trimester {Sagiv, 2018, 4238410;
Wikstrom, 2019, 6311677} minimizing concerns for bias due to pregnancy-related
hemodynamic effects. All three studies were high confidence prospective cohort studies with
many strengths including sufficient study sensitivity and sound methodological approaches,
analysis, and design, as well as no evidence of bias. Between these three studies, PFOS exposure
concentrations observed in Wikstrom et al. {, 2020, 6311677} are more comparable to current
exposure levels in the United States and therefore may be more relevant to the general
population than the candidate RfD derived from Sagiv et al. {, 2018, 4238410} or Darrow et al.,
{, 2013, 2850966}. Additionally, theBMDL derived from Wikstrom et al. {, 2020, 6311677}
was based on a statistically significant regression parameter. For these reasons, the RfD for
decreased birth weight from Wikstrom et al. {, 2020, 6311677} was selected as the basis for the
organ-specific RfD for developmental effects. The resulting health outcome-specific RfD is
1 x 10 7 mg/kg/day (Figure 4-4). Note that all three candidate RfDs based on epidemiological
4-65
-------
APRIL 2024
studies for the developmental outcome were within one order of magnitude of the selected health
outcome-specific RfD.
Immune
Developmental
Cardiovascular
Hepatic
Anti-tetanus
antibody
response
Anti-diphtheria
antibody
response
Anti-rubella
antibody
response
Decreased
birth weight
Increased total
cholesterol
Elevated ALT
Health Outcome Endpoint
Budtz-J0rgensen
and Grandjean
(2018, 5083631)
Budtz-Jergensen
and Grandjean —*-
(2018, 5083631)
^Timmerman et al.
(2021, 9416315)'
Zhang et al.
"(2023, 10699594)"
Sagivetal.
(2018, 4238410)
Wikstrom et al.
(2020, 6311677) '
Darrow et al.
(2013, 2850966)
Dong et al.
(2019, 5080195) '
Steenland et al.
(2009, 1291109)
Galio et al.
(2012, 1276142) '
Nian et al.
(2019, 5080307)
Study
3 x 10"7
Timmerman et al.
' (2021, 9416315)
I 1 1
2 x 10 7
1
2 x 10"7
1 x 10"7
4 x 10"7
6 x 10-7
1 x 10"7
3 x 10"7
1 x 10"7
1 x 10-7
7 x 10"7
2 x 107
Candidate RfD
(mg/kg/day)
2 x 10'7
1 x 10"7
1 x 10"7
2x 10-7
Health Outcome
Specific RfD
(mg/kg/day)
1 x 10-7
Overall RfD
(mg/kg/day)
Figure 4-4. Schematic Depicting Selection of the Overall RfD for PFOS
4.1.6.2 Overall Noncancer 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 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 decreased 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
4-66
-------
APRIL 2024
populations (i.e., infants and children; see Section 5.8), 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.
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 evidence indicates 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 therefore, the overall RfD is expected to be protective of all other
noncancer health effects in humans.
4.2 Cancer
As described in the introduction of Section 3, there is evidence from both epidemiological and
animal toxicological studies that oral PFOS exposure may result in adverse health effects across
many health outcomes, including cancer (Section 3.5). In Section 3.5.5, EPA concluded that
PFOS is Likely to Be Carcinogenic to Humans in accordance with the Guidelines for Carcinogen
Risk Assessment {U.S. EPA, 2005, 6324329}. Therefore, the quantification of cancer effects was
prioritized along with the four noncancer health outcomes that are described in Section 4.1. EPA
considered only high or medium confidence human and animal toxicological studies for CSF
derivation.
4.2.1 Study and Endpoint Selection
Human studies selected for CSF derivation reported all necessary analytical information (e.g.,
exposure distribution or variance) for the outcome of interest (any cancer). If available, high and
medium confidence studies with exposures levels near the range of typical environmental human
exposures, especially exposure levels comparable to human exposure in the general population,
were preferred over studies reporting considerably higher exposure levels. Exposure levels near
the typical range of environmental human exposure can facilitate extrapolation to exposure levels
that may be more relevant to the U.S. general population. Additionally, the most recent and
comprehensive publication on a single study population was preferred over prior publications on
the same or portions of the same population.
Preferred animal toxicological studies consisted of medium and high confidence studies with
chronic exposure durations to capture potential latency of cancer effects. Studies with exposure
durations during development (e.g., gestation) were also considered informative for assessing
potential early lifestage susceptibility to cancer. Studies encompassing lower dose ranges were
also preferred. These types of animal toxicological studies increase the confidence in the CSF
relative to other animal toxicological studies because they are based on data with relatively low
risk of bias, have sufficient study designs to observe the critical effects, and are associated with
less uncertainty related to low-dose and exposure duration extrapolations.
4-67
-------
APRIL 2024
4.2.1.1 Epidemiological Studies
The available epidemiology studies report elevated risk of liver, bladder, kidney, prostate, and
breast cancers after chronic PFOS exposure in some studies, though limited evidence for some
tumor types (i.e., liver and renal) and mixed results for other tumor types (i.e., bladder, prostate,
breast) provide plausible but not definitively causal evidence of a relationship between PFOS
exposure and cancer outcomes from the epidemiological evidence alone. The animal chronic
cancer bioassay provides additional support for carcinogenicity with the identification of multi-
site tumorigenesis (liver and pancreas) in both male and female rats.
The limited renal or mixed results (breast, bladder, prostate) preclude definitive conclusions
about the relationship between PFOS exposure and these cancer outcomes in humans and
therefore limits the potential for quantitative assessment of these data. For example, Shearer et
al., {, 2021, 7161466} is a medium confidence study which suggests an association between
PFOS and increased kidney cancer. However, it is the only study indicating an association for
kidney cancer. Furthermore, the magnitude of the association between PFOS and kidney cancer
was lower than that for PFOA and after adjustment for other PFAS, the adjusted OR for the
highest quartile was relatively low in magnitude and not statistically significant. For these
reasons, Shearer et al., 2021 was not considered for CSF derivation. Additionally, the breast
cancer studies provide mixed evidence, with associations between PFOS and breast cancer
observed in some studies, but only in specific groups of participants or for certain sub-types of
breast cancer. Without plausible evidence for MO As that inform these responses in specific
populations, there is not strong support for quantitative analyses of these studies.
Recently published studies have provided additional evidence of an increased risk of liver cancer
with PFOS exposure. Importantly, these data are concordant with the liver tumors observed in
the published rodent studies {Thomford, 2002, 5432392; Butenhoff, 2012, 1276144}, providing
cross-stream concordance for liver cancer which strengthens the weight of evidence for this
endpoint. Results from publications considered in the 2016 PFOS HESD {U.S. EPA, 2016,
3603365}, a low confidence occupational study {Alexander, 2003, 1291101} and & medium
confidence general population-based study {Eriksen, 2009, 2919344}, investigating associations
between liver cancer and PFOS exposure reported non-significant associations, though these
studies were considered imprecise (i.e., null results with wide confidence intervals). Recently,
statistically significant increased risk of liver cancer has been reported in two additional studies,
a medium confidence nested case-control study in the U.S. {Goodrich, 2022, 10369722} and a
low confidence general population study in China {Cao, 2022, 10412870}. Given the
concordance of tumor site between these studies in humans and the available animal
toxicological study, discussed further in Section 4.2.1.2, EPA considered liver cancer reported by
Goodrich et al. {, 2022, 10369722} for CSF derivation. EPA did not consider Cao et al. {, 2022,
10412870} as there were several concerns with this study, including: the potential for selection
bias due to lack of information on case recruitment and on source of healthy controls;
uncertainties related to outcome assessment due to lack of liver cancer diagnosis detail; and
potential for residual confounding because the list of confounders included in PFAS and liver
cancer analyses was not provided. These concerns resulted in low confidence rating.
Goodrich et al. {, 2022, 10369722}, is a medium confidence study which reported on a small,
nested case-control study of adults from the large Multiethnic Cohort (MEC) in California and
Hawaii. The study examined incident non-viral hepatocellular carcinoma cases and individually
4-68
-------
APRIL 2024
matched controls {Goodrich, 2022, 10369722}. EPA identified several factors that also
precluded use of Goodrich et al. {, 2022, 10369722} from dose-response analyses. First, there
was a lack of association observed in continuous analyses of PFOS exposure indicating a lack of
dose-response. Thus, the study lacks a precise estimate of the slope needed for POD derivation.
Second, the elevated risk in this study was observed only in analyses comparing participants with
PFOS concentrations at or above the 85th percentile of PFOS (i.e., 54.9 (J,g/L). This indicates that
only the highest exposure group demonstrated a response, making 54.9 [j,g/L PFOS the LOAEL.
With only a LOAEL from this dataset, EPA is unable to conduct a low-dose linear extrapolation
or derive a CSF. Lastly, the elevated exposure level at which the response was observed in this
study is outside the reported PFOS environmental human exposures ranges typical for U.S. and
international populations. For example, the mean 90th percentile PFOS serum concentration
from the 2017-2018 NHANES cycle was 11.5 (J,g/L. The small sample size for the study (50
cases and 50 controls) may have limited the study's sensitivity. For these reasons, Goodrich et al.
{, 2022, 10369722} was not selected for CSF derivation.
4.2.1.2 Animal Toxicological Studies
A single high confidence animal chronic cancer bioassay comprises the animal toxicological
evidence database for the carcinogenicity of PFOS. This high confidence chronic cancer
bioassay study, first published as an industry-sponsored report {Thomford, 2002, 5029075} and
later published as a peer-reviewed journal article {Butenhoff, 2012, 1276144} provides evidence
of multisite tumorigenesis in male and female rats.
Hepatocellular tumors were observed in both male and female rats {Butenhoff, 2012, 1276144}.
In males, there was a statistically significant increase in the incidence of hepatocellular
adenomas in the highest dose group tested (20 ppm or approximately 1 mg/kg/day) and a
significant trend of increased incidence with increasing PFOS dose. A similar response was
observed in females, with the addition of one incidence of hepatocellular carcinoma in a rat from
the highest dose group tested (20 ppm or approximately 1.25 mg/kg/day). As these tumors were
observed in both sexes with similar sensitivity and since this effect is concordant with the
associations between PFOS and liver cancer observed in humans, the endpoints of hepatocellular
adenomas in male rats and hepatocellular adenomas or carcinomas in female rats were both
selected for candidate CSF derivation.
Increased incidence of pancreatic islet cell tumors were also observed in male rats {Butenhoff,
2012, 1276144}. Though there were similar incidences of islet cell adenomas in control and
PFOS-treated rats, there was a statistically significant trend of increased incidence of islet cell
carcinomas with increasing PFOS dose. EPA additionally selected the incidence of pancreatic
islet cell carcinomas in male rats for candidate CSF derivation as this is a malignant tumor and
appears to be similar in sensitivity as the hepatocellular tumors observed in male and female rats.
EPA also considered incidences of combined islet cell adenomas and carcinomas for quantitative
analyses, the modeling for which is presented in Appendix E {U.S. EPA, 2024, 11414344} but
was not selected for candidate CSF derivation because there was no dose-response relationship
observed with the adenomas alone and combining the two tumor types resulted in a slight
attenuation of the effect, evidenced by a loss of the statistically significant trend of response.
4-69
-------
APRIL 2024
4.2.2 Candidate CSF Derivation
As described above, EPA did not identify epidemiological studies suitable for CSF derivation.
However, 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
3-18, 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. For comparison purposes, EPA
presents BMDLs derived using the number of animals in each dose group at the start of the study
in Appendix E {U.S. EPA, 2024, 11414344}. All BMDLs were derived using the BMDS 3.2
program.
Multistage models were used consistent with the longstanding practice of EPA to prefer
multistage models to fit tumor dose-response data {U.S. EPA, 2005, 6324329} and a BMR of
10% extra risk was chosen per EPA's Benchmark Dose Technical Guidance {U.S. EPA, 2012,
1239433}. EPA selected the AUC averaged over the study duration (AUCavg), equivalent to the
mean serum concentration over the duration of the study, as the dose metric for modeling cancer
endpoints. This is consistent with the Guidelines for Carcinogen Risk Assessment {U.S. EPA,
2005, 6324329} and the IRIS Handbook {U.S. EPA, 2022, 10367891}, which recommend the
cumulative dose received over a lifetime as the measure of exposure to a carcinogen when
modeling chronic cancer effects. 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 body weight, 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.
4-70
-------
APRIL 2024
Table 4-12. Cancer Slope Factors Derived From Results Reported by Butenhoff et al. {, 2012,1276144}/Thomford {, 2002,
5029075}3 in Sprague-Dawley Rats
Sex
POD Type, Model
POD Internal Dose
PODhed
Candidate CSF
Tumor Type
/Internal Dose Metricb
(BMR/PODhed) Notes on Modeling
Hepatocellular
Male
BMDLio
25.6 mg/L (AUC
3.28 x 10~3mg
/kg/day
Model with the lowest AIC was
Adenomas
Multistage Degree 4
normalized per day
30.5 (mg/kg/day) 1 selected as all models had adequate fit
Model
(AUCavg))
and BMDLs were sufficiently close.
Hepatocellular
Female
BMDLio
21.8 mg/L (AUC
2.79 x l(T3mg
/kg/day
Model with the lowest AIC was
Adenomas
Multistage Degree 1
normalized per day
35.8 (mg/kg/day) 1 selected as all models had adequate fit
Model
(AUCavg))
and BMDLs were sufficiently close.
Combined
Female
BMDLio
19.8 mg/L (AUC
Model with the lowest AIC was
Hepatocellular
Multistage Degree 1
normalized per day
2.53 x l(T3mg
, „ ,, , i selected as all models had adequate fit
39.5 (mg/kg/day) and BMDLs were sufficiently close.
Adenomas and
Model
(AUCavg))
/kg/day
Carcinomas
Pancreatic Islet Cell
Male
BMDLio
26.1 mg/L (AUC
3.34 x 10~3 mg
/kg/day
Model with the lowest AIC was
Carcinomas
Multistage Degree 1
normalized per day
29.9 (mg/kg/day)-1 selected as all models had adequate fit
Model
(AUCavg))
and BMDLs were sufficiently close.
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 Appendix {U.S. EPA, 2024, 11414344} for additional details on benchmark dose modeling.
4-71
-------
APRIL 2024
4.2.3 Overall 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 overall CSF for PFOS.
This endpoint was selected because: 1) there is concordance between the observed hepatocellular
tumors in rats with the liver cancer observed in human epidemiological studies; 2) the derived
candidate CSF is representative of both malignant and benign tumors; 3) the endpoint is
supported by the observation of hepatocellular adenomas in male rats; 4) there was a statistically
significant increase in tumor incidence in the highest dose group; and 5) a statistically significant
trend of increased incidence with increasing PFOS concentrations across dose groups. The
resulting CSF is 39.5 (mg/kg/day)
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 lifestage 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
MOA, 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 versus adult-only PFOS exposure.
4-72
-------
APRIL 2024
5 Effects Characterization
5.1 Addressing Uncertainties in the Use of Epidemiological
Studies for Quantitative Dose-Response Analyses
In the 2016 Health Effects Support Document for Perfluorooctane Sulfonate (PFOS) and
Drinking Water Health Advisory {U.S. EPA, 2016, 3982043; U.S. EPA, 2016, 3603365}, the
U.S. Environmental Protection Agency (EPA) qualitatively considered epidemiological studies
as a supporting line of evidence but did not quantitatively consider them for point-of-departure
(POD) derivation, citing the following as reasons to exclude the epidemiological data that were
available at that time from quantitative analyses:
• Unexplained 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 Sections3.1.1, 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 studies 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 of PFOS
for POD derivation in this assessment.
In this assessment EPA has assessed the strength of epidemiological and animal evidence
following current agency best practices for systematic review {U.S. EPA, 2022, 10367891}, a
process that was not followed in 2016. By performing an updated assessment using systematic
review methods, EPA determined that four noncancer health outcomes and four epidemiological
endpoints within these outcomes (i.e., decreased antibody response to vaccination in children,
decreased birthweight, increased total cholesterol, and increased alanine aminotransferase
(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 high and/or medium confidence epidemiological studies with low risk of bias in
different populations, including general and highly exposed populations. Many of these
supporting studies have been published since 2016 and have strengthened the weight of evidence
for this assessment.
5-1
-------
APRIL 2024
As described in Section 4.1.1.34.1.3, EPA has improved upon the pharmacokinetic modeling
approach used in 2016. Though there are challenges in estimations of human dosimetry from
measured or modeled serum concentrations (see Section 5.6.2), EPA has evaluated the available
literature and developed a pharmacokinetic model that estimates PFOS exposure concentrations
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 uncertainty factor (UF) accounting for intraspecies variation of 10x 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 of chronic,
nondevelopmental effects, is the documented decline 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 Appendix E, {U.S. EPA, 2024, 11414344}), 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 United
States that occurred in the 1990's {Gallo, 2012, 1276142; Nian, 2019, 5080307; Dong, 2019,
5080195; Steenland, 2009, 1291109}. The modeling results show that the benchmark dose lower
confidence limit (BMDL) estimates for increased TC derived using the longest range of exposure
data (1999-2018) 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. Given 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 consistently selected 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 Appendix E, {U.S. EPA, 2024, 11414344}).
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, mixture analysis remains an area
of emerging research {Taylor, 2016, 11320539}, and there is no scientific consensus yet for the
best approach to account for exposure by co-occurring PFAS. Additionally, multipollutant
analyses from studies included in this assessment did not provide direct evidence that
associations between exposure to PFOS and health effects are confounded by or are fully
attributable to confounding by co-occurring PFAS. A detailed discussion of statical approaches
for accounting for co-occurring PFAS and results from studies performing multipollutant
analysis is provided in Section 5.1.1. 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 acidforms) 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
5-2
-------
APRIL 2024
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
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
complexities related to SES and racial/ethnic disparities. Additional racially and ethnically
diverse studies in multiple U.S. communities are needed to fill this important data gap. Appendix
D {U.S. EPA, 2024, 11414344} 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, the four selected critical effects (i.e.,
decreased antibody response to vaccination, increased serum ALT, increased TC, and decreased
birthweight) are biologically significant effects and/or precursors to disease (e.g., CVD), which,
according to agency guidance and methods, both warrant consideration as the basis of RfDs for
PFOA {U.S. EPA, 2002,88824 ; U.S. EPA, 2005, 6324329; U.S. EPA, 2022, 10367891}. EPA's
A Review of the Reference Dose and Reference Concentration Processes, states that a reference
dose (RfD) should be based on an adverse effect or a precursor to an adverse effect
(e.g., increased risk of an adverse effect occurring) {U.S. EPA, 2002, 88824}. Also, at the
individual level, the interpretation and impact of small magnitude changes in endpoints such as
increased TC, increased ALT, decreased birth weight, and decreased antibody response to
vaccination may be less clear. However, at the population level, even small magnitude changes
in these effects will shift the distribution in the overall population and increase the number of
individuals at risk for diseases, such as cardiovascular disease and liver disease{Gilbert, 2006,
174259}.
5-3
-------
APRIL 2024
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.
5.1.1 Uncertainty Due to Potential Confounding by Co-Occurring
PFAS
5.1.1.1 PFAS Co-Exposure Statistical Approaches and Confounding
Analysis
A potential source of uncertainty in epidemiologic studies examining associations between a
particular PFAS and health outcomes is confounding by other co-occurring PFAS. In studies of
PFOS, such confounding may occur if there are other PFAS that are moderately or highly
correlated with PFOS, associated with the outcome of interest, and not on the causal pathway
between PFOS and the outcome. If the association between co-occurring PFAS and the outcome
is in the same direction as the association between PFOS and that outcome, the anticipated
direction of bias resulting from not accounting for other PFAS would be away from the null. For
an extended review of the uncertainties associated with PFAS co-exposures, see the Systematic
Review Protocol for the PFBA, PFHxA, PFHxS, PFNA, and PFDA (anionic and acidforms)
IRIS Assessments {U.S. EPA, 2020, 8642427}.
Several statistical methods are currently used to estimate associations while accounting for
potential confounding by co-occurring PFAS and other pollutants. One common approach is to
include co-occurring PFAS as covariates in regression models. This approach allows for an
estimation of the association between PFOS and the outcome of interest, adjusted for other
covariates and the copollutants. Another approach is to screen large groups of exposures to
identify which ones are most strongly related to the outcome, using methods such as principal
components analysis, elastic net regression, and Bayesian kernel machine regression (BKMR).
Each of these approaches has strengths and limitations. For example, when PFOS and the
copollutants are highly correlated, then multipollutant models could be affected by
multicollinearity or result in amplification bias, rather than reduce confounding bias compared
with single-pollutant models {Weisskopf, 2018, 7325521}. Additionally, accounting for a
variable in a multivariable regression model that is not a significant predictor of the response
variable reduces the degrees of freedom and effectively dilutes the significance of the other
exposure variables that are predictors of the response. The use of screening approaches, while
effective at accounting for copollutants, can result in estimates that are not easily interpretable
and make it difficult to differentiate the impact and contribution of individual PFAS {Meng,
2018, 4829851}. Mixture analysis is an emerging research area {Taylor, 2016, 11320539; Liu,
2022, 10606356}, and there is no scientific consensus yet on the best approach for estimating
independent effects of PFOS within complex PFAS mixtures.
In this assessment, the risk of bias due to confounding by co-occurring PFAS was considered as
part of the study quality evaluation process. To further support the assessment, Section 5.1.1.2
below summarizes evidence from high and medium confidence studies that included at least one
of the approaches described above (hereafter referred to collectively as "multipollutant models")
5-4
-------
APRIL 2024
to account for copollutants, in order to assesses the extent to which there may be confounding by
other PFAS in studies reporting the associations between PFOS and birth weight.
5.1.1.2 Multipollutant Models of PFOS and Birth Weight
When assessing the associations between PFOS and a health effect of interest (e.g., decreased
birth weight), there is concern for potential confounding by other PFAS when there is a strong
correlation between the occurrence of PFOS and another PFAS and when the magnitude of the
association between the co-exposure and the health effect is large.
To identify co-occurring PFAS with potential for confounding, Table 5-1 shows correlations
between PFOS and other PFAS exposures in nine studies evaluating the association between
exposure to PFOS and birth weight, each of which included mutually adjusted models. Four of
these studies are medium confidence {Lenters, 2016, 5617416; Meng, 2018, 4829851; Robledo,
2015, 2851197; Woods, 2017, 4183148} and five are high confidence studies {Ashley-Martin,
2017, 3981371; Luo, 2021, 9959610; Manzano-Salgado, 2017, 4238465; Shoaff, 2018, 4619944;
Starling, 2017, 3858473}. Moderately positive correlations (-0.6) between PFOS and PFOA
were consistently observed in six of the seven studies that reported such information.
Correlations between PFOS and other commonly examined PFAS, including PFNA (four
studies), PFDA (four studies), and PFHxS (five studies), were less consistent than correlations
with PFOA, ranging from weak (i.e., 0.0-0.3) to strong (i.e., 0.7-1.0). These results suggest that
other PFAS may not consistently co-occur with PFOS.
Table 5-1. Correlation Coefficients Between PFOS and Other PFAS in Mutually Adjusted
Studies
Correlations with PFOS
Reference Study Setting
PFOA
PFNA
PFDA
PFHxS
Ashley-Martin et al. {, 2017, 3981371 }a
Canada (10 cities)
0.59
0.55
High
Luo et al. {, 2021, 9959610}3
High
Guangzhou, China
0.11
0.63
0.68
0.01
Manzano-Salgado et al. {, 2017, 4238465}° Gipuzkoa, Sabadell, and
\k
\k
\k
\k
High
Valencia, Spain
Shoaff etal. {, 2018, 4619944}d
Cincinnati, Ohio, USA
High
Starling et al. {, 2017, 3858473}e
High
Colorado, USA
0.68
0.62
0.49
0.65
Lenters et al. {, 2016, 5617416}e
Greenland; Kharkiv,
0.61
0.42
0.78
0.34
Medium
Ukraine; Warsaw, Poland
Meng etal. {, 2018, 4829851}d
Medium
Denmark
0.66
0.48
0.48
0.30
Robledo etal. {, 2015, 2851197}c
Michigan and Texas,
\k
\k
\k
\k
Medium
USA
Woods etal. {, 2017, 4183148}f
Cincinnati, Ohio, USA
+g
+
+
+
Medium
Notes'. NR = not reported.
Shaded cells indicate analytes for which a correlation with PFOA was not measured or reported.
a Pearson correlation of loglO-transformed {Ashley-Martin, 2017, 3981371} and ln-transformed {Luo, 2021, 9959610} PFAS
values.
b Analyte not measured.
5-5
-------
APRIL 2024
c Correlation coefficients not reported.
dPearson correlation of PFAS values, unclear if transformed prior to correlation analysis.
e Spearman rank correlation of PFAS values.
f Correlation coefficient type not specified
g Correlations not reported numerically. Heat map of correlation coefficients (Figure S2, in Woods et al. {, 2017, 4183148})
shows positive correlations between PFOS and PFOA, PFNA, PFHxS, and PFDA, ranging from about 0.6 to about 0.1,
respectively.
Results from mutually adjusted models are summarized and compared in Table 5-2. The
statistical approaches for accounting for PFAS co-exposures varied across the studies. Six
studies included at least one additional PFAS as a predictor in ordinary least squares (OLS)
regression models {Ashley-Martin, 2017, 3981371; Manzano-Salgado, 2017, 4238465; Meng,
2018, 4829851; Robledo, 2015, 2851197; Shoaff, 2018, 4619944; Starling, 2017, 3858473}.
Woods et al. {2017, 4183148} included multiple PFAS as predictors in aBayesian hierarchical
linear model. Three studies {Lenters, 2016, 5617416; Starling, 2017, 3858473; Woods, 2017,
4183148} used elastic net regression, and one study used BKMR {Luo, 2021, 9959610}. The
impact of other PFAS adjustment on the association between PFOS and birth weight is evaluated
by comparing the magnitude and direction of the effects from the single-PFOS model (when
available) to those from mutually adjusted models.
Six studies provided results from both single and multipollutant models {Lenters, 2016,
5617416; Luo, 2021, 9959610; Manzano-Salgado, 2017, 4238465; Meng, 2018, 4829851;
Shoaff, 2018, 4619944; Starling, 2017, 3858473}. Multipollutant models in these six studies
included PFOA but varied with respect to other PFAS considered (Table 5-2). Lenters et al. {,
2016, 5617416} also adjusted for other types of chemicals (such as phthalates and
organochlorides) in addition to several PFAS. Generally, the direction of effect estimates
remained the same following adjustment for other PFAS, but precision was reduced. None of the
studies that showed birth weight deficits in single-pollutant models reported greater magnitude or
more precision of the association following statistical adjustment for other PFAS.
Three studies reported large inverse associations (range: -45 to -83 g) between PFOS and mean
birth weight in single-pollutant (i.e., PFOS only) models {Lenters, 2016, 5617416; Luo, 2021,
9959610; Meng, 2018, 4829851}. In Luo et al. {, 2021, 9959610}, the association remained
statistically significant in a BKMR model that included 11 other PFAS. In Meng et al. {, 2018,
4829851}, the association was slightly attenuated (from -45 to -38 g) and no longer statistically
significant following adjustment for PFOA. Lenters et al. {, 2016, 5617416} observed a
nonsignificant inverse association between PFOS and reduced birth weight in single-pollutant
models, but PFOS was not selected for inclusion in an elastic net regression model that included
other pollutants. Manzano-Salgado et al. {, 2017, 4238465}, Shoaff et al. {, 2018, 4619944}, and
Starling {, 2017, 3858473} reported null results in single and in multi-PFAS regression models.
Additionally, Starling {, 2017, 3858473} reported that PFOS was not selected for inclusion in an
elastic net regression model. Although found in the minority of studies, the large inverse
associations (range: -38 to -109 g) from two multipollutant OLS studies were comparable in
magnitude to the single-pollutant models.
Three studies provided results only from multipollutant models {Ashley-Martin, 2017, 3981371;
Robledo, 2015, 2851197; Woods, 2017, 4183148}, thus making assessment of the impact of
copollutants difficult. None of these studies reported statistically significant associations between
PFOS and birth weight, and PFOS was not selected for the elastic net regression model in Woods
5-6
-------
APRIL 2024
et al. {, 2017, 4183148}, which reported on the same cohort as Shoaff et al. {,2018, 4619944},
that included other endocrine-disrupting chemicals in addition to PFAS.
In summary, in the six studies that included both single and multipollutant models, associations
were attenuated to various degrees while others were strengthened following adjustment for
other PFAS {Lenters, 2016, 5617416; Luo, 2021, 9959610; Manzano-Salgado, 2017, 4238465;
Meng, 2018, 4829851; Shoaff, 2018, 4619944; Starling, 2017, 3858473}.Three additional
studies presented results from multipollutant models only, making it difficult to determine the
extent to which confounding by other PFAS may have impacted the PFOS-birth weight
associations {Ashley-Martin, 2017, 3981371; Robledo, 2015, 2851197; Woods, 2017, 4183148}.
Considering all nine studies (8 different cohorts) together, it is challenging to draw conclusions
about the extent of confounding by co-occurring PFAS, particularly given differences in
modeling approaches, PFAS considered in the adjustment, and exposure contrasts used across
studies. Additionally, these studies represented only a small fraction of the total number of
studies examining associations between PFOS and birth weight and it is unclear whether their
results are generalizable to the broader evidence base. Although it is an important source of
uncertainty, there is no evidence in the entirety of the large evidence base that the observed
associations between PFOS and birth weight deficits are fully attributable to confounding by co-
occurring PFAS.
Similar conclusions can be drawn for other health outcomes. Budtz-Jorgensen {, 2018, 5083631}
evaluated the possibility of confounding across PFAS in analyses of decreased antibody
response. The study reported significant decreases in the antibody response with elevated PFOS
exposure, and there was no notable attenuation of the observed effects after estimates were
adjusted for PFOA (see Section 3.4.2.1.1.1) {Budtz-Jorgensen, 2018, 5083631}. Alimited
number of studies performed co-exposure analyses for increased ALT and increased TC in
adults. Lin et al. {,2010, 1291111} performed multipollutant modeling for the effects on serum
ALT, but multipollutant modeling results for the association between PFOS exposure and ALT
was not reported. Fan et al. {, 2020, 7102734} examined cross-sectional associations between
exposure to PFOS and increased TC in single- and multipollutant models in a sample of adults
from NHANES (2012-2014). Exposure to PFOS was associated with significantly elevated TC
in the single-pollutant model, but the association was no longer significant in multipollutant
analyses. A significantly positive association was also observed for PFAS mixture and TC in
WQS regression analyses {Fan, 2020, 7102734}.
Overall, there is no evidence that the consistently observed associations between exposures to
PFOS and the four priority noncancer health outcomes are confounded or are fully attributable to
confounding by co-occurring PFAS.
5-7
-------
APRIL 2024
Table 5-2. Impact of Co-Exposure Adjustment on Estimated Change in Mean Birth Weight (Grams) per Unit Change (ng/mL)
in PFOS Levels.
Reference
Single PFAS Model
Result (95% CI)ab
Multi-PFAS Model
Result (95% CI)ab
Elastic Net
Regression
Resultb
Exposure Comparison
Effect of PFAS
Adjustment on PFOA
Birth Weight Results
PFAS Adjustments
Ashley-Martin et al.
{,2017,3981371}
High
NR
Girls: 94.31 (-76.30,
264.92)
Bovs: -11.15 (-174.26,
151.95)
_C
logio-unit (ng/mL)
increase
PFOA, PFHxS
Luo et al. {, 2021,
9959610}
High
-93.34 (-157.92,
-28.75)
-109 (-215, -4)d
Sinsle PFAS model: ln-
unit (ng/mL) increase
Multi-PFAS model:
75th vs. 25th percentile
Results not directly PFOA, PFBA, PFNA,
comparable due to different PFDA, PFUnDA,
exposure comparisons, but PFDoDA, PFTrDA,
both models showed large PFBS, PFHxS, 6:2 Cl-
inverse associations PFESA, 8:2 Cl-PFESA
Manzano-Salgado et
al. {, 2017,
4238465}
High
0.44 (-32.48, 33.36)
18.64 (-26.08, 63.36)
log2-unit (ng/mL)
increase
Strengthened (increased
birth weight)
PFOA, PFNA, PFHxS
Shoaff et al. {,2018,
4619944}
High
-0.06 (-0.16, 0.04)e
-0.06 (-0.26, 0.15)e
log2-unit (ng/mL)
increase
No change
PFOA, PFNA, PFHxS
Starling et al. {,
2017,3858473}
High
-13.8 (-53.8, 26.3)
29.09 (-32.56, 90.75)
N/S
ln-unit (ng/mL) increase Attenuated/changed
direction
PFOA, PFNA, PFDA,
PFHxS
Lenters et al. {,
2016,5617416}
Medium
-68.84 (-152.90,
15.22)
N/S
2 SD ln-unit (ng/mL)
increase
Attenuated
PFOA, PFNA, PFDA,
PFHxS, PFHpA,
PFUnDA, PFDoDA
Meng et al. {, 2018,
482985 l}f
Medium
-45.2 (-76.8, -13.6)
-38.11 (-82.09, 5.88)
log2-unit (ng/mL)
increase
Slightly Attenuated
PFOA
Robledo et al. {,
2015, 2851197}®
Medium
NR
Girls: 14.16 (-81.83,
110.15)
Bovs: 37.51 (-73.45,
148.46)
1 SD ln-unit (ng/mL)
increase
PFOA, PFDA, PFNA,
PFOSA, Et-PFOSA-
AcOH, Me-PFOSA-
AcOH
Woods et al. {,
2017,4183148}
Medium
NR
-9 (-53, 35)h
N/S
logio-unit (ng/mL)
increase
PFOA, PFNA, PFDA,
PFHxS
Notes: NR = not reported; N/S = not sufficient.
5-8
-------
APRIL 2024
5.2 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
4-4), 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 with humans. To address this
potential source of uncertainty, EPA utilized a pharmacokinetic (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.
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 with
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 nonmonotonic 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.
One example of this is the PPARa response. It is unclear to what extent PPARa influences the
responses to PFOS exposure observed in humans, though the rodent PPARa response 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
5-9
-------
APRIL 2024
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.
5.3 Updated Approach to Animal Toxicological RfD Derivation
Compared with 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 Health
Effects Support Document (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 with the 2016 PFOS
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 no-observed-adverse-effect level/lowest-observed-adverse-effect level (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, 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 relationship 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, in this assessment, EPA only derived PODs for endpoints from health
outcomes with evidence indicating or evidence demonstrating an association with PFOS exposure.
5-10
-------
APRIL 2024
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 a less specific
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, 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 5-3.
Table 5-3. Comparison of Candidate RfDs Derived from Animal Toxicological Studies for
Priority Health Outcomes3
Studies and Effects Used in 2016 for Candidate Studies and Effects Used in 2024 for Candidate
RfD Derivationb
RfD Derivation
Immune
NA
Zhong et al. {, 2016, 3748828}, medium confidence - decreased
pup PFC response to SRBC
NTP {, 2019, 5400978}, high confidence - extramedullar
hematopoiesis in the spleen
Developmental
Luebker et al. {, 2005, 757857} medium Luebker et al. {, 2005, 757857}, medium confidence - decreased
confidence - decreased pup body weight 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 {, 2002, 5029075},
confidence - increased ALT (and increased BUN) high confidence - individual 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.
5-11
-------
APRIL 2024
a Note that candidate RfDs for the fourth priority noncancer 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
PFOS HESD or the current assessment.
b Candidate RfDs from the 2016 PFOS HESD that correspond to nonprioritized health outcomes (e.g., nervous) are not presented
here.
5.4 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}, which was, in part, because no new animal toxicological studies had been
published since publication of the 2016 PFOS HESD and the new epidemiological literature
published up until 2021 continued 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" for
PFOS and requested that the agency provide an "explicit description of why the available data
for PFOS do not meet the EPA Guidelines for Carcinogen Risk Assessment (USEPA, 2005)
criterion for the higher designation as 'likely carcinogenic'" {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 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, the agency has determined the available data for PFOS
surpass many of the descriptions for Suggestive Evidence of Carcinogenic Potential according to
the Guidelines for Carcinogen Risk Assessment {U.S. EPA, 2005, 6324329} and meet the
descriptions for Likely to Be Carcinogenic to Humans, as described in Section 3.5.5. This
conclusion was based on four independent factors. First, EPA considered the SAB's request that
EPA "reevaluate the 2012 Butenhoff study" {U.S. EPA, 2022, 10476098}. After reviewing the
available data, as described in Sections 3.5.2, 3.5.5, and below in this subsection, EPA
subsequently agreed with the SAB that the agency's prior "interpretation of the hepatocellular
carcinoma data from the Butenhoff (2012b) study in the 2016 PFOS HESD is overly
conservative in dismissing the appearance of a dose-response relationship for this endpoint,
particularly in females" {U.S. EPA, 2022, 10476098}. Second, as requested by the SAB, and
following agency methodology {U.S. EPA, 2022, 10367891}, EPA incorporated syntheses of
mechanistic literature, which served as the basis of EPA's conclusions that multiple, potentially
human-relevant MO As may contribute to the hepatocellular tumors reported in PFOS
toxicological studies of rats (see Section 3.5.4.2). This conclusion aligned with the SAB's
comments that "multiple MO As may be operative" in the reported hepatocellular tumorigenesis
5-12
-------
APRIL 2024
and that "the rodent liver tumors caused by PFOS do not appear to be PPAR-a dependent,"
{U.S. EPA, 2022, 10476098}. Third, EPA considered the SAB's comment that there were
inconsistencies between EPA's draft conclusions and "the California EPA conclusions based on
the same human, animal, and mechanistic evidence presented in the EPA PFOS document,"
leading the EPA to re-review the CalEPA's draft Public Health Goals for PFOA and PFOS
technical document {CalEPA, 2021, 9416932} and identify data indicating the occurrence of
tumorigenesis in a second tumor site in male rats (i.e., pancreatic islet cell tumors) {U.S. EPA,
2022, 10476098}. Fourth, EPA identified new supporting epidemiological literature resulting
from the SAB's recommendation that EPA update the literature search prior to finalization of the
toxicity assessments for PFOA and PFOS {U.S. EPA, 2022, 10476098}. This new
epidemiological literature included two studies reporting increased risk of hepatocellular
carcinoma associated with increased PFOS exposure in humans {Goodrich, 2022, 10369722;
Cao, 2022, 10412870}, which provided concordant evidence between one of the tumor types and
sites observed in the available animal toxicological study. This concordance further supports the
potential human relevance of the hepatocellular tumors observed in animal toxicological studies
of PFOS.
More specifically, the examples for which the PFOS database exceeds the Suggestive Evidence
descriptions outlined in the Guidelines for Carcinogen Risk Assessment include:
• "a small, and possibly not statistically significant, increase in tumor incidence observed in
a single animal or human study that does not reach the weight of evidence for the
descriptor 'Likely to Be Carcinogenic to Humans;'
• a small increase in a tumor with a high background rate in that sex and strain, when there
is some but insufficient evidence that the observed tumors may be due to intrinsic factors
that cause background tumors and not due to the agent being assessed;
• evidence of a positive response in a study whose power, design, or conduct limits the
ability to draw a confident conclusion; and
• a statistically significant increase at one dose only, but no significant response at the other
doses and no overall trend." {U.S. EPA, 2005, 6324329}.
The strongest evidence for the carcinogenicity of PFOS is 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. As described in Section 3.5.4.2, EPA determined that these tumor types
are potentially relevant to humans.
In the initial draft of this toxicity assessment published for SAB review (i.e., the Proposed
Approaches document) {U.S. EPA, 2021, 10428576}, as well as the 2016 PFOS HESD {U.S.
EPA, 2016, 3603365}, EPA relied upon the tumor incidences provided in Butenhoff et al. {,
2012, 1276144}, which is the peer-reviewed manuscript of an industry report - Thomford {,
2002, 5029075}. Upon further review of the results presented in the Thomford {, 2002,
5029075} report prior to finalization of this assessment, the agency identified two factors that
limited previous qualitative and quantitative interpretations of the data: 1) the Butenhoff et al. {,
5-13
-------
APRIL 2024
2012, 1276144} study reported combined incidences of neoplastic lesions in the control and
high-dose groups (males and females) 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 dilution of tumor incidence in the high-dose group as many of
the tumor types observed in the study, including hepatocellular neoplasms, were not reported
until approximately 70 weeks of treatment or later. Therefore, EPA conducted a re-analysis that
excluded animals sacrificed at the interim time point from statistical analyses as it was
biologically implausible for the 10 animals from the interim time point to have presented with
neoplasms. As a result of this reanalysis, EPA agreed with the SAB that the original analysis was
"overly conservative in dismissing the appearance of a dose-response relationship for this
endpoint, particularly in females" {U.S. EPA, 2022, 10476098}.
The second factor prevented EPA from previously identifying the statistically significant trend in
a second tumor site/type (pancreatic islet cell carcinomas) observed in the chronic cancer
bioassay. As a result of identifying the second tumor site/type and updating the conclusions
regarding hepatocellular tumors in females, the EPA concluded that PFOS met an additional
characteristic for the designation of Likely to Be Carcinogenic to Humans: "an agent that has
tested positive in animal experiments in more than one species, sex, strain, site, or exposure
route, with or without evidence of carcinogenicity in humans" (emphasis added) {U.S. EPA,
2005, 6324329}.
Overall, the Thomford/Butenhoff et al. {, 2002, 5029075;, 2012, 1276144} report, along with
plausible associations between PFOS exposure and carcinogenicity reported in epidemiological
studies, particularly for hepatocellular carcinoma, provide 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). See Table 5-4 below for specific details on how PFOS exceeds the
examples supporting the Suggestive Evidence of Carcinogenic Potential cancer descriptor in the
Guidelines for Carcinogen Risk Assessment {U.S. EPA, 2005, 6324329}.
After reviewing the examples of the descriptor Carcinogenic to Humans, EPA has determined
that at this time, the evidence supporting the carcinogenicity of PFOS does not warrant a
descriptor exceeding Likely to Be Carcinogenic to Humans. The Guidelines indicate that a
chemical agent can be deemed Carcinogenic to Humans if it meets all the following conditions:
• "there is strong evidence of an association between human exposure and either cancer or
the key precursor events of the agent's mode of action but not enough for a causal
association, and
• there is extensive evidence of carcinogenicity in animals, and
• the mode(s) of carcinogenic action and associated key precursor events have been
identified in animals, and
• there is strong evidence that the key precursor events that precede the cancer response in
animals are anticipated to occur in humans and progress to tumors, based on available
biological information" {U.S. EPA, 2005, 6324329}.
5-14
-------
APRIL 2024
As discussed in Section 3.5.5, convincing epidemiological evidence supporting a causal
association between human exposure to PFOS and cancer is currently lacking. Additionally,
though the available evidence indicates that there are positive associations between PFOS and
multiple cancer types, there is uncertainty regarding the identification of carcinogenic modes of
action (MOAs) and associated key precursor events for PFOS in animals. See Table 5-4 below
for specific details on how PFOS does not align with the examples supporting the Carcinogenic
to Humans cancer descriptor in the Guidelines for Carcinogen Risk Assessment {U.S. EPA,
2005, 6324329}.
Table 5-4. Comparison of the PFOS Carcinogenicity Database with Cancer Descriptors as
Outlined in the Guidelines for Carcinogen Risk Assessment {U.S. EPA, 2005, 6324329}
Comparison of Evidence for Suggestive and Carcinogenic Cancer Descriptors
Suggestive Evidence of Carcinogenic Potential
"A small, and possibly not statistically significant,
increase in tumor incidence observed in a single
animal or human study that does not reach the weight
of evidence for the descriptor "Likely to Be
Carcinogenic to Humans." The study generally would
not be contradicted by other studies of equal quality in
the same population group or experimental system"
{U.S. EPA, 2005, 6324329}
PFOS data exceed this description. Observed statistically
significant increases in hepatic tumors in rats (adenomas in
males and adenomas and carcinomas in females) at the
high dose and a statistically significant trend overall in both
sexes. Concordant evidence of increased risk of
hepatocellular carcinoma from two human epidemiological
studies. Observed statistically significant trend of increased
incidence of pancreatic islet cell tumors in male rats.
"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." {U.S. EPA,
2005,6324329}
This description is not applicable to the tumor types
observed after PFOS exposure.
"Evidence of a positive response in a study whose PFOS data exceed this description. The animal study
power, design, or conduct limits the ability to draw a from which carcinogenicity data are available was
confident conclusion (but does not make the study determined to be high confidence during study quality
fatally flawed), but where the carcinogenic potential is evaluation,
strengthened by other lines of evidence (such as
structure-activity relationships)." {U.S. EPA, 2005,
6324329}
"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}
PFOS data exceed this description. Observed statistically
significant increases in hepatic tumors (adenomas in males
and adenomas and carcinomas in females) at the high dose
and a statistically significant trend overall. Also observed
statistically significant trend of increased pancreatic islet
cell carcinomas with increasing dose.
Carcinogenic to Humans
This descriptor is appropriate when there is
convincing epidemiologic evidence of a causal
association between human exposure and cancer.
PFOS data are not consistent with this description.
There is evidence of a plausible association between PFOS
exposure and cancer in humans, however, the database is
limited, there is uncertainty regarding the potential
confounding of other PFAS, and there is limited
mechanistic information that could contribute to the
determination of a causal relationship. The database would
benefit from large high confidence cohort studies in
independent populations.
5-15
-------
APRIL 2024
Comparison of Evidence for Suggestive and Carcinogenic Cancer Descriptors
Or, this descriptor may be equally appropriate with a
lesser weight of epidemiologic evidence that is
strengthened by other lines of evidence. It can be used
when all of the following conditions are met:
There is strong evidence of an association between
human exposure and either cancer or the key precursor
events of the agent's MOA but not enough for a causal
association.
There is extensive evidence of carcinogenicity in
animals.
The mode(s) of carcinogenic action and associated key
precursor events have been identified in animals.
There is strong evidence that the key precursor events
that precede the cancer response in animals are
anticipated to occur in humans and progress to tumors,
based on available biological information.
Notes: MOA = mode of action.
PFOS data are not consistent with this description.
There is evidence of an association between human
exposure and cancer, however, there is limited mechanistic
information that could contribute to the determination of a
causal relationship.
PFOS data are not consistent with this description. Only
one chronic cancer bioassay is available for PFOS. The
database would benefit from high confidence chronic
studies in other species and/or strains.
PFOS data are not consistent with this description. A
definitive MOA has not been identified for each of the
PFOS-induced tumor types identified in rats.
PFOS data are not consistent with this description. The
animal database does not provide significant clarity on the
MOA(s) of PFOS in animals.
5.5 Health Outcomes with Evidence Integration Judgments of
Evidence Suggests Bordering on Evidence indicates
EPA evaluated 16 noncancer 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 final 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 {U.S. EPA, 2021, 10428576}. However,
upon further examination using the protocols for evidence integration outlined in Appendix A
{U.S. EPA, 2024, 11414344} and Section 2.1.5, EPA concluded that the available
epidemiological and animal toxicological evidence did not meet the criteria recommended 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.
5-16
-------
APRIL 2024
Epidemiological studies published since the 2016 PFOS HESD considered for evidence
integration for adverse endocrine effects include high and medium confidence studies, though
EPA determined that there was slight evidence to suggest human endocrine toxicity, including
associations between PFOS exposure and thyroid disease. The available evidence supports the
relationship between PFOS exposure and thyroid stimulating hormone (TSH) in children and, to
a lesser extent, adults. However, 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 {U.S. EPA, 2024, 11414344} 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 final IRIS Toxicological Reviews for both PFBA {U.S. EPA, 2022, 11181062}
and PFHxA {U.S. EPA, 2023, 11181061} concluded that the available evidence indicates that
the observed thyroid effects were likely due to PFBA and PFHxA exposure, respectively. Given
the similarities across PFAS, these findings support potential associations between PFOS and
adverse endocrine effects.
There was also slight evidence from epidemiological studies published since the 2016 PFOS
HESD 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 {U.S. EPA, 2024, 11414344} for a
detailed description of endocrine evidence synthesis and integration.
5-17
-------
APRIL 2024
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
reviews conducted in February 2022 and 2023. 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.
5.6 Challenges and Uncertainty in Modeling
5.6.1 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 Appendix F, {U.S. EPA, 2024, 11414344}), 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
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 Appendix F, {U.S. EPA, 2024, 11414344}).
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 Appendix, {U.S. EPA, 2024,
11414344}). 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 lifestages 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 lifestages.
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-
5-18
-------
APRIL 2024
course changes for PFOS uptake to the milk. Despite these limitations, the incorporation of
model parameters related to developmental lifestages is a significant improvement over the
model used in the 2016 PFOS HESD which did not implement life-course modeling {U.S. EPA,
2016, 3603365}.
5.6.2 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 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 with blood in neonates and infants compared with adults
{Fernandez, 2011, 9641878}. This behavior 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
Appendix B, {U.S. EPA, 2024, 11414344}). 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 with 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
5-19
-------
APRIL 2024
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
nonhuman 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 1 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 1 year. This is a relatively long duration of breastfeeding, only
27% of children in the United States are being breastfed at 1 year of age {CDC, 2013, 1936457}.
Along with using the 95th percentile of breastmilk consumption, this provides a scenario of high
but 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 United States are
very unlikely to be exclusively breastfed for up to 1 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.
5.6.3 Approach of Estimating a Benchmark Dose from a
Regression Coefficient
EPA identified epidemiological studies that reported associations between PFOS exposure and
response variables as regression coefficients. Since such a regression coefficient is associated
with a change in the biological response variable, it is biologically meaningful and can therefore
be used for POD derivation. EPA modeled these regression coefficients using the same approach
used to model studies that reported measured response variables. The SAB PFAS Review Panel
agreed with this approach, stating, "it would seem straightforward to apply the same
methodology to derive the beta-coefficients ("re-expressed," if necessary, in units of per ng/mL)
for antibody responses to vaccines and other health-effect-specific endpoints. Such a coefficient
could then be used for deriving PODs" {U.S. EPA, 2022, 10476098}. When modeling regression
coefficients that were reported per log-transformed units of exposure, EPA used the SAB's
recommended approach and re-expressed the reported P coefficients in units of per ng/mL (see
5-20
-------
APRIL 2024
Appendix E, {U.S. EPA, 2024, 11414344}). Sensitivity analyses to evaluate the potential impact
of re-expression in a hybrid approach when modeling hepatic and serum lipid studies for PFOS
showed little impact on BMDLs (see Appendix E, {U.S. EPA, 2024, 11414344}).
To evaluate this potential uncertainty in BMDLs derived based on regression coefficients, 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
Appendix E {U.S. EPA, 2024, 11414344}.
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 threefold difference. The RfD
allows for an order of magnitude (10-fold or 1,000%) uncertainty in the estimate. Therefore,
EPA is confident in its use of regression coefficients, re-expressed or not, as the basis of
PODheds.
5.7 Human Dosimetry Models: Consideration of Alternate
Modeling Approaches
Physiologically based pharmacokinetic (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, for which data are typically scarcer.
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. Because of
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
5-21
-------
APRIL 2024
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.
Because of 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 two one-compartment approaches identified in the literature for PFOA was the model of
Verner et al. (2016, 3299692) and the model developed by the Minnesota Department of Health
(MDH model) {Goeden, 2019, 5080506}, 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
overall agreement in approach between the two models 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
with 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
5-22
-------
APRIL 2024
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 body weight. 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 with the adult value. There is no chemical-specific information to suggest that
Vd is larger in children compared with 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 it is 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 Appendix F, {U.S.
EPA, 2024, 11414344}). This resulted in reduced predictions of serum concentrations, primarily
during their peak in early childhood. The model with variable Vd did not decrease the root mean
squared error compared with the model with constant Vd. Because the model with constant Vd
had 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 body weight in young children compared with
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 Appendix F, {U.S.
EPA, 2024, 11414344}). 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 a maximum contaminant level goal (MCLG) calculated
thereafter using lifestage 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
5-23
-------
APRIL 2024
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 with
the model with constant exposure relative to body weight. 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 was a much
greater effect on PFOS model results relative to PFOA, but in both cases model performance, as
quantified by root mean squared error, was superior with constant exposure compared with
exposure based on drinking water consumption. 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 under SDWA. The
epidemiological endpoints can be placed into three categories based on the age of the individuals
at the time of exposure measurement: 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 based on noncancer effects or for
nonlinear carcinogens is an algebraic calculation that incorporates the RfD, RSC, and drinking
water intake. The drinking water intake used for this type of MCLG calculation would be chosen
based on the target population relevant to the exposure interval used in the critical study and/or
timing of exposure measurement and the response variable 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 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 presented in Appendix F {U.S. EPA, 2024,
11414344}, EPA determined that the updated Verner model was the most appropriate available
model structure for PODhed calculation for PFOS. Specifically, the EPA concluded that the
determination that assuming Vd in children equal to the adult values and calculating a RfD
assuming a constant dose (mg/kg/day) were appropriate for this assessment.
5.8 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 more severe response than others despite similar PFOS exposure due to increased
biological sensitivity, as well as populations exhibiting a more severe response due to higher
PFOS exposure and/or exposure to other chemicals or nonchemical stressors. Populations with
greater biological sensitivity may include pregnant women and their developing fetuses,
5-24
-------
APRIL 2024
children, adolescents, lactating women, the elderly, and people with certain underlying medical
conditions (see Section 5.8.1). 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 nonchemical stressors such as communities with
environmental justice concerns (see Section 5.8.2).
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.
5.8.1 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 Sections 0
and 3.4.2.1, results of some epidemiological studies indicate an association between PFOS
exposure during pregnancy and/or early childhood and adverse outcomes such as decreased birth
weight and decreased antibody response to vaccinations. 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 lifestages or exposure
windows (e.g., prenatal or early postnatal exposure windows) may be more consequential than
others. For example, as described in Appendix C {U.S. EPA, 2024, 11414344}, 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
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 mount a relatively low antibody response compared to other children 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.
5.8.2 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,
5-25
-------
APRIL 2024
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 5.1), 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 Final Per- and Polyfluoroalkyl Substances National Primary
Drinking Water Regulation {U.S. EPA, 2024, 11414059}). 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.
5-26
-------
APRIL 2024
6 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
6-1
-------
APRIL 2024
Treatment Plant and Landfills as Sources of Polyfluoroalkyl Compounds to the
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
AitBamai, 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
6-2
-------
APRIL 2024
modeling of saturable, renal resorption of perfluoroalkylacids in monkeys—probing the
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.
6-3
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2149477
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.
6-4
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1049488
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.
6-5
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1274133
Berg, V; N0st, TH; Hansen, S; Elverland, A; Veyhe, AS; Jorde, R; Odland, J0; 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
6-6
-------
APRIL 2024
Blanc, M; Riiegg, J; Scherbak, N; Keiter, SH. (2019). Environmental chemicals differentially
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
6-7
-------
APRIL 2024
children: the HOME study. Environ Health Perspect 122: 513-520.
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
6-8
-------
APRIL 2024
Butenhoff, JL; Kennedy, GL; Hinderliter, PM; Lieder, PH; Jung, R; Hansen, KJ; Gorman, GS;
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 withNHANES 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
6-9
-------
APRIL 2024
Cao, T; Qu, A; Li, Z; Wang, W; Liu, R; Wang, X; Nie, Y; Sun, S; Zhang, X; Liu, X. (2021). The
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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/reference/details/reference_id/2001177
Caserta, D; Ciardo, F; Bordi, G; Guerranti, C; Fanello, E; Perra, G; Borghini, F; LaRocca, 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.cfim/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.cfim/reference/details/reference_id/4728855
CDC. (2011). Tetanus surveillance — United States, 2001-2008. MMWRMorb Mortal Wkly
Rep 60: 365-369.
https://hero.epa.gov/hero/index.cfim/reference/details/reference_id/9998272
6-10
-------
APRIL 2024
CDC. (2013). Breastfeeding report card : United States 2013. Atlanta, GA.
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
6-11
-------
APRIL 2024
Chen, J; Tanguay, RL; Tal, TL; Gai, Z; Ma, X; Bai, C; Tilton, SC; Jin, D; Yang, D; Huang, C;
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
6-12
-------
APRIL 2024
Health Perspect 129: 37004.
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.
6-13
-------
APRIL 2024
Environ Int 155: 106596.
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; B adders, 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], Crit Rev 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.
6-14
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1429922
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
6-15
-------
APRIL 2024
D'Agostino, RB; Vasan, RS; Pencina, MJ; Wolf, PA; Cobain, M; Massaro, JM; Kannel, WB.
(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 Saf 209: 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,
6-16
-------
APRIL 2024
N; Valentini, S; Valsenti, L; Vecchie, V. (2015). Current exposure of Italian women of
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;
6-17
-------
APRIL 2024
Acquasaliente, L; Pozzi, N; Plebani, M; Garolla, A; Foresta, C. (2019). Endocrine
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; Mukherjee, 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
6-18
-------
APRIL 2024
Perspect 121: 507-513, 513e501-508.
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
6-19
-------
APRIL 2024
functional thyroid disease status over the lifetime. PLoS ONE 14: e0219769.
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
6-20
-------
APRIL 2024
Elcombe, CR; Peffer, RC; Wolf, DC; Bailey, J; Bars, R; Bell, D; Cattley, RC; Ferguson, SS;
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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/reference/details/reference_id/3223669
Fabrega, F; Kumar, V; Schuhmacher, M; Domingo, JL; Nadal, M. (2014). PBPK modeling for
6-21
-------
APRIL 2024
PFOS and PFOA: validation with human experimental data. Toxicol Lett 230: 244-251.
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
6-22
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9419076
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
6-23
-------
APRIL 2024
Int 83: 176-182.
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.
6-24
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859819
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:
6-25
-------
APRIL 2024
111632. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/9959601
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.
6-26
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1290854
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 noncancer ovarian granulosa cells.
6-27
-------
APRIL 2024
Toxicol In Vitro 65: 104769.
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
6-28
-------
APRIL 2024
online 9959716
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
6-29
-------
APRIL 2024
acids, perfluoroalkyl phosphinic acids, perfluoroalkyl carboxylic acids, and
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-
6-30
-------
APRIL 2024
46. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10164687
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.
6-31
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5079713
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.
6-32
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315798
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
6-33
-------
APRIL 2024
hypertensive disorders of pregnancy: A prospective cohort study. Environ Int 138:
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.
6-34
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284267
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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/reference/details/reference_id/6833623
Jain, RB. (2020). Variabilities in concentrations of selected perfluoroalkyl acids among
6-35
-------
APRIL 2024
normotensives and hypertensives across various stages of glomerular function. Arch
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
6-36
-------
APRIL 2024
concentrations of perfluoroalkyl substances and bone health in British girls at age 17.
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
6-37
-------
APRIL 2024
Johanson, CE. (1979). Distribution of fluid between extracellular and intracellular compartments
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).
6-38
-------
APRIL 2024
Biomonitoring perfluorinated compounds in Catalonia, Spain: concentrations and trends
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
6-39
-------
APRIL 2024
Kerstner-Wood, C; Coward, L; Gorman, G; Southern Research Institute. (2003). Protein binding
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;
6-40
-------
APRIL 2024
Giesy, JP. (2011). Trans-placental transfer of thirteen perfluorinated compounds and
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.cfim/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.cfim/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.cfim/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: a cross-sectional study. Metabolomics 15: 95.
https://hero.epa.gov/hero/index.cfim/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.cfim/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.cfim/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.cfim/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.cfim/reference/details/reference_id/3981430
6-41
-------
APRIL 2024
Kobayashi, S; Sata, F; Goudarzi, H; Araki, A; Miyashita, C; Sasaki, S; Okada, E; Iwasaki, Y;
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; Jurling, 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
6-42
-------
APRIL 2024
sperm Y:X chromosome ratio in two European populations and in Inuit from Greenland.
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
6-43
-------
APRIL 2024
Lee, JW; Choi, K; Park, K; Seong, C; Yu, SD; Kim, P. (2020). Adverse effects of perfluoroalkyl
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;
6-44
-------
APRIL 2024
Andersson, EM. (2022). Cancer incidence in a Swedish cohort with high exposure to
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.
6-45
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6315681
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
6-46
-------
APRIL 2024
Epidemiol 180: 574-581.
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
6-47
-------
APRIL 2024
Lin, TW; Chen, MK; Lin, CC; Chen, MH; Tsai, MS; Chan, DC; Hung, KY; Chen, PC. (2020).
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 Pollut235: 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.
6-48
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/4238514
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 IgG N-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.
6-49
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/787186
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
6-50
-------
APRIL 2024
Lopez-Espinosa, MJ; Mondal, D; Armstrong, B; Bloom, MS; Fletcher, T. (2012). Thyroid
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-
6-51
-------
APRIL 2024
mediated signaling pathway disruption through gene expression profiling of the spleen.
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
6-52
-------
APRIL 2024
human embryonic and fetal organs from first, second, and third trimester pregnancies.
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.
6-53
-------
APRIL 2024
Environ Sci Technol 38: 373-380.
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
6-54
-------
APRIL 2024
Occup Environ Health 80: 643-648.
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
6-55
-------
APRIL 2024
substances and mid-childhood lipid and alanine aminotransferase levels. Environ Int 111:
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; J0rgensen, 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.
6-56
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1291110
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
6-57
-------
APRIL 2024
of Commerce, National Oceanic and Atmospheric Administration, National Centers for
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
6-58
-------
APRIL 2024
sulfonate-induced endoplasmic reticulum stress. Mol Med Rep 15: 2583-2587.
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.
6-59
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6988477
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
6-60
-------
APRIL 2024
Oulhote, Y; Steuerwald, U; Debes, F; Weihe, P; Grandjean, P. (2016). Behavioral difficulties in
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; Brants aster, 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.
6-61
-------
APRIL 2024
(2016). Cord blood gene expression supports that prenatal exposure to perfluoroalkyl
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.
6-62
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5080584
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
6-63
-------
APRIL 2024
Qazi, MR; Abedi, MR; Nelson, BD; Depierre, JW; Abedi-Valugerdi, M. (2010). Dietary
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 Ley dig cells. Toxicology 449: 152663.
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/7276729
6-64
-------
APRIL 2024
Qiu, L; Zhang, X; Zhang, X; Zhang, Y; Gu, J; Chen, M; Zhang, Z; Wang, X; Wang, SL. (2013).
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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.
6-65
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10284350
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
6-66
-------
APRIL 2024
from blood with an ongoing exposure source: The example of perfluorooctanoic acid
(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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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
6-67
-------
APRIL 2024
contamination in composite food samples from Dallas, Texas, USA. Environ Health
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
6-68
-------
APRIL 2024
intakes on the isomeric profiles of perfluoroalkyl substances (PFASs) in human serum.
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
polychlorinated 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
6-69
-------
APRIL 2024
exposure. Environ Sci Technol 45: 7999-8005.
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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/reference/details/reference_id/3160486
6-70
-------
APRIL 2024
Smith, MT; Guyton, KZ; Kleinstreuer, N; Borrel, A; Cardenas, A; Chiu, WA; Felsher, DW;
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
HealthPerspect 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
6-71
-------
APRIL 2024
StataCorp. (2021). Stata Statistical Software: Release 17 [Computer Program], College Station,
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
Strain, 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
6-72
-------
APRIL 2024
Int 68: 41-48. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/2922190
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
6-73
-------
APRIL 2024
82: 359-359. https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/5082311
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
poly chlorinated 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
Poly chlorinated 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.
6-74
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3981439
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
6-75
-------
APRIL 2024
Forum, https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/30019
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
6-76
-------
APRIL 2024
Report], (EPA 822-R-16-004). Washington, DC: U.S. Environmental Protection Agency,
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.
6-77
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/10668582
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
6-78
-------
APRIL 2024
Velez, MP; Arbuckle, TE; Fraser, WD. (2015). Maternal exposure to perfluorinated chemicals
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.
6-79
-------
APRIL 2024
Environ Int 139: 105694.
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
6-80
-------
APRIL 2024
atopy. Environ Res 111: 785-791.
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
6-81
-------
APRIL 2024
substances and thyroid stimulating hormone among pregnant women: a cross-sectional
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
6-82
-------
APRIL 2024
Wen, LL; Lin, CY; Chou, HC; Chang, CC; Lo, HY; Juan, SH. (2016). Perfluorooctanesulfonate
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
6-83
-------
APRIL 2024
proliferator-activated receptor-alpha (PPARa) by perfluoroalkyl acids (PFAAs): further
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 Pollut249: 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
6-84
-------
APRIL 2024
Wu, XM; Bennett, DH; Calafat, AM; Kato, K; Strynar, M; Andersen, E; Moran, RE; Tancredi,
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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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 Agric Food Chem 68: 633-641.
https://hero.epa.gov/hero/index.cfim/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.cfim/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
6-85
-------
APRIL 2024
Acids after Ceasing Exposure from Drinking Water Contaminated by Firefighting Foam.
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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.cfim/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.
6-86
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/1332453
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-
6-87
-------
APRIL 2024
and Poly(fluoroalkyl) Substances (PFASs) in Highly Consumed Seafood Products from
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; DallaZuanna, 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)
6-88
-------
APRIL 2024
and perfluorooctanesulfonic acid (PFOS) in surface waters, sediments, soils and
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.
6-89
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/6988534
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.
6-90
-------
APRIL 2024
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/3859849
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
6-91
-------
APRIL 2024
Zhou, Y; Bao, WW; Qian, ZM; Dee Geiger, S; Parrish, KL; Yang, BY; Lee, YL; Dong, GH.
(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
6-92
------- |