EPA/635/R-25/012Fa
www.epa.aov/iris
IRIS Toxicological Review of Perfluorohexanesulfonic Acid
(PFHxS, CASRN 335-46-4) and Related Salts
January 2025
Integrated Risk Information System
Center for Public Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
DISCLAIMER
This document has been reviewed by the U.S. Environmental Protection Agency, Office of
Research and Development and approved for publication. Any mention of trade names, products, or
services does not imply an endorsement by the U.S. government or the U.S. Environmental
Protection Agency. EPA does not endorse any commercial products, services, or enterprises.
ii
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
CONTENTS
CONTENTS iii
AUTHORS | CONTRIBUTORS | REVIEWERS xii
EXECUTIVE SUMMARY xv
ES.l Lifetime and Subchronic Oral Reference Dose (RfD) for Noncancer Effects xviii
ES.2 Confidence in the Oral Reference Dose (RfD) and subchronic RfD xviii
ES.3 Noncancer Effects Following Inhalation Exposure xix
ES.4 Evidence for Carcinogenicity xix
1. OVERVIEW OF BACKGROUND INFORMATION AND ASSESSMENT METHODS 1-1
1.1. BACKGROUND INFORMATION ON PERFLUOROHEXANESULFONIC ACID (PFHXS) 1-1
1.1.1. Physical and Chemical Properties 1-1
1.1.2. Sources, Production, and Use 1-3
1.1.3. Environmental Fate and Transport 1-4
1.1.4. Potential for Human Exposure and Populations with Potentially Greater Exposure 1-4
1.2. SUMMARY OF ASSESSMENT METHODS 1-8
1.2.1. Literature Search and Screening 1-8
1.2.2. Evaluation of Individual Studies 1-11
1.2.3. Data Extraction 1-14
1.2.4. Evidence Synthesis and Integration 1-14
1.2.5. Dose-Response Analysis 1-16
2. LITERATURE SEARCH AND STUDY EVALUATION RESULTS 2-1
2.1. LITERATURE SEARCH AND SCREENING RESULTS 2-1
2.2. STUDY EVALUATION RESULTS 2-3
3. PHARMACOKINETICS, EVIDENCE SYNTHESIS, AND INTEGRATION 3-1
3.1. PHARMACOKINETICS 3-1
3.1.1. Absorption 3-4
3.1.2. Distribution 3-6
3.1.3. Metabolism 3-20
3.1.4. Excretion 3-20
3.1.5. Evaluation of PBPK and PK Modeling 3-40
iii
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3.1.6. Empirical Pharmacokinetic Analysis 3-44
3.1.7. Model Evaluation Conclusion and Extrapolation Approach 3-50
3.2. NONCANCER HEALTH EFFECTS 3-54
3.2.1. Thyroid Effects 3-55
3.2.2. Immune Effects 3-81
3.2.3. Developmental Effects 3-112
3.2.4. Hepatic Effects 3-204
3.2.5. Neurodevelopmental Effects 3-237
3.2.6. Cardiometabolic Effects 3-260
3.2.7. Hematopoietic Effects 3-305
3.2.8. Female Reproductive Effects 3-312
3.2.9. Male Reproductive Effects 3-337
3.2.10. Renal Effects 3-351
3.2.11. Other Noncancer Health Effects 3-361
3.3. CARCINOGENICITY 3-364
3.3.1. Cancer 3-364
4. SUMMARY OF HAZARD IDENTIFICATION CONCLUSIONS 4-1
4.1. SUMMARY OF CONCLUSIONS FOR NONCANCER HEALTH EFFECTS 4-1
4.2. SUMMARY OF CONCLUSIONS FOR CARCINOGENICITY 4-5
4.3. CONCLUSIONS REGARDING SUSCEPTIBLE POPULATIONS AND LIFESTAGES 4-5
5. DERIVATION OF TOXICITY VALUES 5-1
5.1. NONCANCER AND CANCER HEALTH EFFECT CATEGORIES CONSIDERED 5-1
5.2. NONCANCER TOXICITY VALUES 5-2
5.2.1. Oral Reference Dose (RfD) Derivation 5-2
5.2.2. Subchronic Toxicity Values for Oral Exposure (Subchronic Oral Reference Dose
[RfD]) Derivation 5-25
5.2.3. Inhalation Reference Concentration (RfC) Derivation 5-27
5.3. CANCER TOXICITY VALUES 5-27
REFERENCES 1
iv
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
TABLES
Table 1-1. Physical-chemical properties of PFHxS and related salts3 1-3
Table 1-2. Serum PFHxS concentrations based on NHANES 2013-2014 data (ng/L) 1-5
Table 1-3. PFHxS levels at 10 military installations 1-7
Table 1-4. Populations, exposures, comparators, and outcomes (PECO) criteria 1-10
Table 3-1. Estimated volume of distribution (Vd) values in rats, mice, and monkeys 3-10
Table 3-2. Measured cord serum:maternal serum ratios 3-15
Table 3-3. Summary of estimated clearance values in animals 3-23
Table 3-4. Summary of clearance values estimated for humans 3-35
Table 3-5. Summary clearance values for humans 3-40
Table 3-6. Pharmacokinetic parameters for rats, mice, monkeys, and humans 3-47
Table 3-7. Data-derived extrapolation factor (DDEF) calculations 3-53
Table 3-8. Associations between PFHxS exposure and thyroid hormone levels in medium
confidence studies of adults 3-59
Table 3-9. Associations between PFHxS exposure and thyroid hormone levels in medium
confidence studies of infants 3-63
Table 3-10. Evidence profile table for PFHxS thyroid effects 3-79
Table 3-11. Summary of PFHxS and data on antibody response to vaccines in children 3-85
Table 3-12. Summary of PFHxS and data on antibody response to vaccines in adults 3-88
Table 3-13. Summary of PFHxS and selected data on infectious disease in humans 3-91
Table 3-14. Summary of PFHxS and data on hypersensitivity in humans 3-98
Table 3-15. Animal study details 3-104
Table 3-16. Evidence profile table for PFHxS immune effects 3-109
Table 3-17. Summary of 34 epidemiologic studies of PFHxS exposure and growth restriction
measures 3-123
Table 3-18. Summary of 11 epidemiologic studies of PFHxS exposure and post-natal growth
measured 3-168
Table 3-19. Associations between PFHxS and anogenital distance in medium confidence
epidemiology studies 3-172
Table 3-20. Summary of 19 epidemiological studies of PFHxS exposure and gestational duration
measures- 3-182
Table 3-21. Evidence profile table for PFHxS-related developmental effects 3-195
Table 3-22. Associations between PFHxS and liver enzymes in medium confidence epidemiology
studies 3-210
Table 3-23. Evidence profile table for oral PFHxS exposure and liver effects 3-232
Table 3-24. Summary of results for medium confidence epidemiology studies of PFHxS exposure
and cognitive effects 3-244
Table 3-25. Summary of results for medium confidence epidemiology studies of PFHxS exposure
and attention deficit hyperactivity disorder (ADHD) 3-248
Table 3-26. Summary of results for medium confidence epidemiology studies of PFHxS exposure
and behavior 3-250
Table 3-27. Evidence profile table for PFHxS neurotoxicological effects 3-258
Table 3-28. Associations between PFHxS exposure and blood lipids in medium confidence
epidemiology studies 3-265
Table 3-29. Associations between PFHxS exposure and hypertension in medium confidence
epidemiology studies in adolescents and young adults 3-274
v
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-30. Associations between PFHxS exposure and gestational hypertension and
preeclampsia in medium confidence epidemiology studies 3-274
Table 3-31. Associations between PFHxS exposure and type 2 diabetes in epidemiology studies 3-278
Table 3-32. Associations between PFHxS exposure and gestational diabetes in epidemiology
studies 3-280
Table 3-33. Associations between PFHxS exposure and insulin resistance or blood glucose in
epidemiology studies 3-283
Table 3-34. Associations between maternal exposure to PFHxS and adiposity in children 3-292
Table 3-35. Associations between maternal exposure to PFHxS and overweight status in children
in medium confidence epidemiology studies 3-296
Table 3-36. Evidence profile table for PFHxS exposure and cardiometabolic effects 3-303
Table 3-37. Evidence profile table for PFHxS hematopoietic effects 3-311
Table 3-38. Summary of results for epidemiology studies of fecundity 3-314
Table 3-39. Associations between PFHxS and breastfeeding duration in epidemiology studies 3-324
Table 3-40. Evidence profile table for PFHxS exposure and female reproductive effects 3-333
Table 3-41. Associations between PFHxS and semen sperm parameters in medium confidence
epidemiology studies 3-339
Table 3-42. Evidence profile table for PFHxS exposure and male reproductive effects 3-348
Table 3-43. Associations between PFHxS exposure and renal function 3-354
Table 3-44. Evidence profile table for PFHxS urinary system effects 3-360
Table 3-45. Associations between PFHxS exposure and bone mineral density in medium
confidence epidemiology 3-362
Table 4-1. Hazard conclusions across published EPA PFAS human health assessments 4-3
Table 5-1. Endpoints considered for dose-response modeling and derivation of points of
departure for thyroid effects in animals 5-4
Table 5-2. Endpoints considered for dose-response modeling and derivation of points of
departure for immune (decreased serum antibody) effects in humans 5-6
Table 5-3. Mean birth weight deficit studies considered for dose-response modeling and
derivation of points of departure for developmental effects in humans 5-9
Table 5-4. Endpoints considered for dose-response modeling and derivation of points of
departure for liver effects in humans 5-11
Table 5-5. Benchmark response levels selected for BMD modeling of PFHxS outcomes 5-12
Table 5-6. Points of departure (PODs) considered for the derivation of PFHxS candidate toxicity
values 5-14
Table 5-7. Uncertainty factors for the development of the lifetime RfDfor PFHxS 5-18
Table 5-8. Lifetime candidate values for PFHxS 5-21
Table 5-9. Confidence in the organ-/system-specific RfDsfor PFHxS 5-22
Table 5-10. RfD and organ-/system-specific RfDs for PFHxS 5-24
Table 5-11. Subchronic RfD organ-/system-specific RfD values for PFHxS 5-26
vi
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
FIGURES
Figure 1-1. Chemical structure of PFHxS and related salts 1-2
Figure 2-1. Literature search for perfluorohexanesulfonic acid and related salts 2-2
Figure 3-1. Observed end-of-study of PFHxS in female and male rats in the NTP bioassay (NTP,
2019) as a function of dose 3-3
Figure 3-2. Ratio of extracellular water (% of body weight) in children versus adults 3-17
Figure 3-3. Serum concentrations of PFHxS in U.S. males versus females as a function of age 3-30
Figure 3-4. Comparison of PFHxS PBPK model predictions to IV dosimetry data (circles) of Kim et
al. (2018b) for a 10 mg/kg dose 3-41
Figure 3-5. Comparison of Female (left) and Male (right) CL values for IV and gavage exposure of
equivalent dose levels from Kim et al. (2016b), Kim et al. (2018b) and Huang et
al. (2019a) 3-46
Figure 3-6. Study evaluation results for epidemiology studies of PFHxS and thyroid effects 3-58
Figure 3-7. Study evaluation results for measures of thyroid hormone levels in PFHxS animal
toxicity studies 3-66
Figure 3-8. Summary of thyroid hormone measures in animal studies 3-68
Figure 3-9. Percent change in thyroid hormone levels following PFHxS exposure in the available
animal toxicology studies 3-69
Figure 3-10. Study evaluation results for endocrine histopathology outcomes in PFHxS animal
toxicity studies 3-71
Figure 3-11. Study evaluation results for endocrine organ weights in PFHxS animal toxicity
studies 3-72
Figure 3-12. Summary of endocrine organ weight effects in animal studies 3-73
Figure 3-13. EDSP21 results of PFHxS active assays: A: ATG_ERE_CIS_up induction assay
performed in HepG2 cells; B: NIS_RAIU_inhibition assay performed in HEK293T
cells 3-76
Figure 3-14. Summary of evaluation of epidemiology studies of PFHxS and antibody response
immunosuppression effects 3-83
Figure 3-15. Summary of evaluation of epidemiology studies of PFHxS and infectious disease
immunosuppression effects 3-90
Figure 3-16. Summary of evaluation of epidemiology studies of PFHxS and hypersensitivity
effects (e.g., asthma, allergies, and atopic dermatitis) 3-96
Figure 3-17. Study evaluation results of PFHxS animal toxicity studies with immune-related
endpoints 3-103
Figure 3-18. Summary of PFHxS immune hematology results 3-105
Figure 3-19. Study evaluation results for 39 epidemiological studies of birth weight and PFHxS 3-117
Figure 3-20. Perinatal studies of birth weight measures and subsets included in different
evaluations 3-118
Figure 3-21. Overall population birth weight results for 11 high confidence PFHxS
epidemiological studies 3-127
Figure 3-22. Overall population birth weight results for 16 medium and low confidence
epidemiological studies 3-128
Figure 3-23. Forest plot of 27 studies included for the EPA meta-analysis on changes in mean
birth weight per each In-unit PFHxS increase 3-129
Figure 3-24. Sex-specific male infants only mean birth weight results for 14 PFHxS
epidemiological studies 3-132
vii
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Figure 3-25. Sex-specific female infants only mean birth weight results for 14 PFHxS
epidemiological studies 3-133
Figure 3-26. Overall population standardized birth weight results for 12 epidemiologic studies 3-136
Figure 3-27. Sex-stratified standardized birth weight results for five epidemiologic studies (boys
above reference line, girls below) 3-137
Figure 3-28. Study evaluation results for 19 epidemiological studies of birth length and PFHxS 3-141
Figure 3-29. Overall population mean birth length results for 16 PFHxS epidemiological studies 3-142
Figure 3-30. Thumbnail schematic of Sex-stratified birth length results for 11 epidemiologic
studies (boys above reference line, girls below) 3-143
Figure 3-31. Study evaluation results for 14 epidemiological studies of head circumference and
PFHxS 3-145
Figure 3-32. Overall population head circumference results for 12 epidemiologic studies 3-147
Figure 3-33. Sex-stratified head circumference results for eight epidemiologic studies (boys
above reference line, girls below) 3-148
Figure 3-34. Study evaluation results for seven epidemiological studies of small for gestational
age and low birth weight and PFHxS 3-150
Figure 3-35. Small for gestational age and low birth weight results for seven epidemiologic
studies 3-151
Figure 3-36. Study evaluation results for 13 epidemiological studies of postnatal growth and
PFHxS 3-154
Figure 3-37. Standardized postnatal weight results for PFHxS epidemiological studies 3-156
Figure 3-38. Mean postnatal weight results for PFHxS epidemiological studies 3-157
Figure 3-39. Standardized postnatal height results for PFHxS epidemiological studies 3-159
Figure 3-40. Mean postnatal height results for PFHxS epidemiological studies 3-160
Figure 3-41. Postnatal rapid growth (weight-for-age and weight-for-length z-score) results for
PFHxS epidemiological studies 3-163
Figure 3-42. Postnatal rapid growth (length-for-age and head circumference z-score) results for
PFHxS epidemiological studies 3-164
Figure 3-43. Postnatal head circumference results for PFHxS epidemiological studies 3-165
Figure 3-44. Postnatal body mass index, adiposity, and ponderal index and weight status results
for PFHxS epidemiological studies 3-167
Figure 3-45. Summary of study evaluation for epidemiology studies of anogenital distance 3-170
Figure 3-46. Summary of study evaluation for 10 epidemiology studies of preterm birth 3-175
Figure 3-47. Preterm birth results for 10 PFHxS epidemiological studies 3-176
Figure 3-48. Study evaluation results for 19 epidemiological studies of gestational age and
PFHxS 3-178
Figure 3-49. Overall population gestational age results for 17 PFHxS epidemiological studies 3-179
Figure 3-50. Sex-stratified gestational age results for 8 PFHxS epidemiological studies 3-181
Figure 3-51. Study evaluation results for five epidemiological studies of fetal loss and PFHxS 3-184
Figure 3-52. Summary of study evaluation for two epidemiology studies of birth defects 3-186
Figure 3-53. Developmental animal study evaluation heatmap 3-187
Figure 3-54. PFHxS-induced developmental effects 3-190
Figure 3-55. Hepatic effects human study evaluation heatmap 3-206
Figure 3-56. PFHxS liver weight animal study evaluation heatmap 3-214
Figure 3-57. Liver weight responses from animal studies 3-215
Figure 3-58. Liver histopathology animal study evaluation heatmap 3-217
Figure 3-59. PFHxS liver histopathology observations from short-term animal toxicology studies 3-218
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Figure 3-60. PFHxS liver histopathology observations from developmental animal toxicity studies
(F0 generation animals) 3-219
Figure 3-61. PFHxS liver histopathology observations from developmental animal toxicity studies
(F1 generation animals) 3-220
Figure 3-62. PFHxS liver serum biomarkers animal study evaluation heatmap 3-221
Figure 3-63. PFHxS liver/hepatobiliary serum biomarkers 3-222
Figure 3-64. PFHxS liver hepatic lipid content study evaluation heatmap 3-224
Figure 3-65. Mode of action for PFHxS-induced liver effects 3-227
Figure 3-66. Summary of study evaluation for epidemiology studies of neurodevelopment 3-239
Figure 3-67. Confidence scores of neurodevelopmental system effects from repeated PFHxS
dose animal toxicity studies 3-254
Figure 3-68. Study evaluation results for epidemiology studies of PFHxS and blood lipids 3-262
Figure 3-69. Study evaluation results for epidemiology studies of PFHxS and cardiovascular
disease risk factors 3-273
Figure 3-70. Study evaluation results for epidemiology studies of PFHxS and cardiovascular
disease 3-276
Figure 3-71. Summary of study evaluation for PFHxS and type 2 diabetes in epidemiology
studies 3-277
Figure 3-72. Heatmap of study evaluations for PFHxS and gestational diabetes 3-279
Figure 3-73. Heatmap of study evaluations for insulin resistance and blood glucose3-282
Figure 3-74. Summary of study evaluations for epidemiology studies of PFHxS and metabolic
syndrome 3-289
Figure 3-75. Summary of study evaluations for epidemiology studies of adiposity 3-290
Figure 3-76. Cardiometabolic effects, heart weight/histopathology - animal study evaluation
heatmap 3-298
Figure 3-77. Cardiometabolic effects, serum lipids - animal study evaluation heatmap 3-300
Figure 3-78. Serum cholesterol responses from animal studies 3-300
Figure 3-79. Hematological animal study confidence scores from repeated PFHxS dose animal
toxicity studies 3-306
Figure 3-80. Hematopoietic effects of PFHxS exposure in animals 3-309
Figure 3-81. Summary of study evaluation for epidemiology studies of fecundity 3-313
Figure 3-82. Summary of study evaluations for epidemiology studies of female reproductive
hormones 3-317
Figure 3-83. Summary of study evaluation for epidemiology studies of other female reproductive
effects (menstrual cycle characteristics, gynecological conditions, ovarian
reserve, and pubertal development) 3-320
Figure 3-84. PFHxS mating and fertility animal study evaluation heatmap 3-326
Figure 3-85. PFHxS estrous cycle animal study evaluation heatmap 3-327
Figure 3-86. PFHxS hormone levels animal study evaluation heatmap 3-328
Figure 3-87. PFHxS female reproductive histopathology animal study evaluation heatmap 3-329
Figure 3-88. PFHxS female reproductive organ weight animal study evaluation heatmap 3-330
Figure 3-89. PFHxS female reproductive sexual differentiation and maturation animal study
evaluation heatmap 3-331
Figure 3-90. Semen parameters epidemiology study evaluation heatmap 3-338
Figure 3-91. Summary of study evaluation for epidemiology studies of male reproductive
hormones 3-340
Figure 3-92. Male reproductive animal study evaluation heatmap - sperm measures 3-343
Figure 3-93. Male reproductive histopathology animal study evaluation heatmap 3-344
ix
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Figure 3-94. Male reproductive animal study evaluation heatmap - reproductive hormones 3-345
Figure 3-95. Male reproductive animal study evaluation heatmap - reproductive organ weights 3-346
Figure 3-96. Male reproductive animal study evaluation heatmap - developmental effects and
functional measures 3-347
Figure 3-97. Renal effects human study evaluation heatmap 3-352
Figure 3-98. Renal effects - animal study evaluation heatmap 3-357
Figure 3-99. Musculoskeletal effects human study evaluation heatmap 3-361
Figure 3-100. Study evaluation results for epidemiology studies of PFHxS and cancer 3-365
x
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
ABBREVIATIONS AND ACRONYMS
ADHD
attention deficit hyperactivity disorder
MNPCE
micronucleated polychromatic
AIC
Akaike's information criterion
erythrocyte
ALT
alanine aminotransferase
MOA
mode of action
AST
aspartate aminotransferase
MTD
maximum tolerated dose
atm
atmosphere
ATSDR
Agency for Toxic Substances and
NCI
National Cancer Institute
Disease Registry
NOAEL
no-observed-adverse-effect level
BMD
benchmark dose
NTP
National Toxicology Program
BMDL
benchmark dose lower confidence limit
NZW
New Zealand White (rabbit breed)
BMDS
Benchmark Dose Software
ORD
Office of Research and Development
BMR
benchmark response
osRfD
organ-specific reference dose
BUN
blood urea nitrogen
PBPK
physiologically based pharmacokinetic
BW
body weight
PFHxS
perfluorohexanesulfonic acid
CA
chromosomal aberration
PND
postnatal day
CASRN
Chemical Abstracts Service registry
POD
point of departure
number
POD [AD J]
duration-adjusted POD
CHO
Chinese hamster ovary (cell line cells)
QSAR
quantitative structure-activity
CPHEA
Center for Public Health and
relationship
Environmental Assessment
RD
relative deviation
CL
confidence limit
RfC
inhalation reference concentration
CNS
central nervous system
RfD
oral reference dose
CYP450
cytochrome P450
RGDR
regional gas dose ratio
DAF
dosimetric adjustment factor
RNA
ribonucleic acid
DDEF
data-derived extrapolation factor
SAR
structure activity relationship
DMSO
dimethylsulfoxide
SCE
sister chromatid exchange
DNA
deoxyribonucleic acid
SD
standard deviation
EPA
Environmental Protection Agency
SDH
sorbitol dehydrogenase
ER
extra risk
SE
standard error
FDA
Food and Drug Administration
SEM
Systematic Evidence Map
FEVi
forced expiratory volume of 1 second
SGOT
glutamic oxaloacetic transaminase, also
GD
gestation day
known as AST
GDH
glutamate dehydrogenase
SGPT
glutamic pyruvic transaminase, also
GGT
y-glutamyl transferase
known as ALT
GLP
good laboratory practices
TSCATS
Toxic Substances Control Act Test
GSH
glutathione
Submissions
GST
glutathione-S-transferase
TWA
time-weighted average
HBCD
hexabromocyclododecane
UF
uncertainty factor
Hb/g-A
animal blood:gas partition coefficient
UFa
animal-to-human uncertainty factor
Hb/g-H
human blood:gas partition coefficient
UFd
database deficiencies uncertainty factor
HEC
human equivalent concentration
UFh
human variation uncertainty factor
HED
human equivalent dose
UFl
LOAEL-to-NOAEL uncertainty factor
HERO
Health and Environmental Research
UFs
subchronic-to-chronic uncertainty
Online
factor
i.p.
intraperitoneal
WOS
Web of Science
IRIS
Integrated Risk Information System
i.v.
intravenous
LCso
median lethal concentration
LD50
median lethal dose
LOAEL
lowest-observed-adverse-effect level
MN
micronuclei
xi
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
AUTHORS | CONTRIBUTORS | REVIEWERS
Assessment Managers (Lead Authors)
Xabier Arzuaga. Ph.D. U.S. EPA/ORD/CPHEA
Ingrid L. Druwe. Ph.D.
Authors
Thomas F. Bateson, Sc.D., M.P.H. U.S. EPA/ORD/CPHEA
I. Allen Davis. M.S.P.H.
Michael Dzierlenga, Ph.D.
Andrew Kraft, Ph.D.
Alexandra L. Lee, Ph.D.
Elizabeth Radke, Ph.D.
Hongyu Ru, Ph.D.
Paul Schlosser, Ph.D.
Shana White, Ph.D.
|ohn Michael Wright, Sc.D.
lay Zhao, Ph.D.
lason C. Lambert, Ph.D. U.S. EPA/ORD/CCTE
Contributors
Michelle M. Angrish, Ph.D. U.S. EPA/ORD/CPHEA
Laura V. Dishaw, Ph.D.
Mary Gilbert, Ph.D.
Barbara Glenn, Ph.D. (retired)
Christopher Lau, Ph.D.
Geniece M. Lehmann, Ph.D.
Andrew Hotchkiss, Ph.D.
Anuradha Mudipalli, Ph.D.
Kathleen Newhouse, Ph.D.
Pamela D. Noyes, Ph.D.
Katherine L. O'Shaughnessy, Ph.D.
Kristen Rappazzo, Ph.D.
Susan Makris, Ph.D. (retired)
Tammy Stoker, Ph.D.
Andre Weaver, Ph.D.
Erin Yost, Ph.D.
Chris Corton, Ph.D.
Stephanie Kim, Ph.D. (former)
Andrew Rooney, Ph.D.
Kyla Taylor, Ph.D.
Dori Germolec, Ph.D.
Alexis Agbai
Timothy Decoff
Angela Scafidi (former)
Robyn B. Blain, Ph.D.
Alexandra E. Goldstone, M.P.H.
Alexander |. Lindahl, M.P.H.
Christopher A. Sibrizzi, M.P.H.
U.S. EPA/ORD/CCTE
U.S. EPA/Region 2
DTT/NIEHS
Oak Ridge Associated Universities (ORAU) Contractor
ICF
xii
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Project Management and Production Team
JackRehmann U.S. EPA/ORD/CPHEA
Ryan Jones
Dahnish Shams
Avanti Shirke
Jessica Soto Hernandez
Vicki Soto
Samuel Thacker
Garland Waleko
Grace Kaupas (former) Oak Ridge Associated Universities (ORAU) Contractor
Rebecca Schaefer
Jacqueline Weinberger (former)
Executive Direction
Wayne Cascio, M.D., FACC
V. Kay Holt, M.S.
Samantha Jones, Ph.D.
Kristina Thayer, Ph.D.
Andrew Kraft, Ph.D.
Paul White, Ph.D.
Ravi Subramaniam, Ph.D. (retired)
Shannon Hanna, Ph.D.
Janice Lee, Ph.D.
Glenn Rice, Sc.D.
Viktor Morozov, Ph.D.
Vicki Soto, B.S.
Review
CPAD Executive Review Committee
Kristina Thayer CPAD Division Director
Paul White CPHEA/CPAD/Senior Science Advisor
Janice Lee CPHEA/CPAD/Toxic Effects Assessment (RTP) Branch Chief
Glenn Rice CPHEA/CPAD/Science Assessment Methods Branch Chief
Ravi Subramaniam (retired) CPHEA/CPAD/Toxic Effects Assessment (DC) Branch Chief
Karen Hogan CPHEA/CPAD/Emeritus
Alan Stern NJDEP (retired), Contractor
Agency Review
This assessment was provided for review to scientists in EPA's program and regional offices.
Comments were submitted by:
Office of the Administrator/Office of Children's Health Protection
Office of Air and Radiation/Office of Air Quality and Standards
Office of Land and Emergency Management
Office of Water
Region 1, Boston, MA
Region 3, Philadelphia, PA
Region 4, Atlanta, GA
Region 8, Denver, CO
CPHEA Center Director
CPHEA Deputy Center Director
CPHEA Associate Director
CPAD Division Director
CPAD Associate Division Director, IRIS PFAS Team Lead
CPAD Senior Science Advisor
CPAD Senior Science Advisor
CPHEA/CPAD/Toxic Effects Assessment (DC) Branch Chief
CPHEA/CPAD/Toxic Effects Assessment (RTP) Branch Chief
CPHEA/CPAD/Science Assessment Methods Branch Chief
CPHEA/CPAD/Quantitative Assessment Branch Chief
CPHEA/CPAD/Assessment Management Branch Chief
xiii
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Interagency Review
This assessment was provided for review to other federal agencies and the Executive Office of the
President (EOP). Comments were submitted by:
Department of Defense (DoD)
Department of Health and Human Services
• Agency for Toxic Substance and Disease Registry
• National Institute for Occupational Safety and Health
• National Institute of Environmental Health Sciences
Executive Office of the President
• Council on Environmental Quality
• Office of Information and Regulatory Affairs
• Office of Science and Technology Policy
United States Department of Agriculture
xiv
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
EXECUTIVE SUMMARY
Perfluorohexanesulfonic acid (PFHxS, CASRN 355-46-4),1 and its related salts (such as
potassium perfluorohexanesulfonate [PFHxS-K, CASRN 3871-99-6], ammonium
perfluorohexanesulfonate [PFHxS-Nl-U, CASRN 68259-08-5], and sodium perfluorohexanesulfonate
[PFHxS-Na, CASRN 82382-12-5]), are members of the group per- and polyfluoroalkyl substances
(PFAS). This assessment applies to PFHxS as well as nonmetal and alkali metal salts of PFHxS that
would be expected to fully dissociate in aqueous solutions of pH ranging from 4 to 9 (e.g., in the
human body) and not release other moieties that would cause toxicity independent of PFHxS. The
synthesis of evidence and toxicity value derivation presented in this assessment focuses on the free
acid of PFHxS and its potassium, sodium, and ammonium salts given the currently available
toxicity data.
Concerns about PFHxS and other PFAS stem from the resistance of these compounds to
hydrolysis, photolysis, and biodegradation, which leads to their persistence in the environment
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 as many PFAS are
resistant to heat and are used to confer resistance of products (e.g., textiles) to stains by repelling
oil, grease, and water. PFAS are also used in a wide range of other applications, including electrical
insulation and to confer frictionless coatings onto surfaces. PFAS in the environment are found at
industrial sites, military fire training areas, wastewater treatment plants, and in commercial
products (see Appendix A, Section 2.1.2).
The Integrated Risk Information System (IRIS) Program is developing a series of five PFAS
assessments (i.e., perfluorohexane sulfonate [PFHxS], perfluorobutanoic acid [PFBA],
perfluorohexanoic acid [PFHxA], perfluorononanoic acid [PFNA], perfluorodecanoic acid [PFDA],
and their associated salts) (see December 2018 IRIS Program Outlook) at the request of EPA
national programs and regions. Specifically, the development of human health toxicity assessments
for exposure to these individual PFAS represents only one component of the broader PFAS strategic
roadmap at the EPA f https://www.epa.gov/pfas/pfas-strategic-roadmap-epas-commitments-
action-2021-20241. The systematic review protocol (see Appendix A) for these five PFAS
assessments outlines the related scoping and problem-formulation efforts, including a summary of
other federal and state assessments of PFHxS. The protocol also describes the systematic review
1 The CASRN given here is for linear PFHxS; the source of PFHxS used in toxicity studies was reported to be
98% pure and reagent grade, generally giving this CASRN. None of the studies referenced in this assessment
explicitly state that only the linear form was used. Therefore, there is the possibility that a minor proportion
of the PFHxS used in the studies were branched isomers and thus observed health effects may apply to the
total linear and branched isomers in a given exposure source.
xv
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
and dose-response methods used to conduct this review (see also Section 1.2). In addition to these
ongoing IRIS PFAS toxicity assessments, EPA's Office of Research and Development is carrying out
several other activities related to PFAS, including the creation of PFAS systematic evidence maps
(SEMs) fRadke etal.. 2022: Carlson etal.. 20221 and consolidating and updating PFAS data on
chemical and physical properties, human health toxicity, and pharmacokinetics, as well as
ecotoxicity.
Human epidemiological studies have examined possible associations between PFHxS
exposure and health outcomes, including immune responses, birth weight, hematopoietic effects,
thyroid hormone effects, liver enzyme effects, serum lipids effects, cardiovascular disease,
hematological effects, reproductive effects, neurodevelopmental effects, and cancer. The ability to
draw conclusions from the epidemiological evidence for the assessed health outcomes is limited
(apart from immune effects) by the overall quality and lack of consistency in the available studies.
Animal studies of PFHxS exposure exclusively examined the oral exposure route; therefore,
no inhalation assessment was conducted nor was an inhalation reference concentration (RfC)
derived (see Section 5.2.3). The available animal studies of oral PFHxS exposure examined a variety
of noncancer endpoints, including those relevant to the thyroid, immune system, developmental
effects, hematopoietic system, hepatic effects, cardiometabolic effects, reproductive (male and
female) system, nervous system, and renal effects. Some limitations in the animal database include
the types of studies identified (e.g., few subchronic and single chronic exposure studies were
available), and few studies per health outcome.
Overall, the available evidence indicates that PFHxS exposure is likely to cause thyroid and
developmental immune effects in humans, given sufficient exposure conditions. For thyroid effects,
the primary supporting evidence for this hazard conclusion included evidence of decreased thyroid
hormone levels, abnormal histopathology results, and changes in organ weight in experimental
animals. For immune effects, the primary supporting evidence included decreased antibody
responses to vaccination against tetanus or diphtheria in children. Selected quantitative data from
these identified hazards were used to derive toxicity values (see Table ES-1; see Sections 3.2.1 and
3.2.2 for evidence synthesis and integration analyses).
Evidence primarily from epidemiological studies suggests but is insufficient to infer that
PFHxS exposure might affect fetal development, specifically resulting in decreased birth weight (see
Section 3.2.3). However, because of limitations and uncertainties in the currently available studies,
a hazard could not be clearly identified, and these data were not considered for use in deriving
toxicity values. While no reference dose (RfD) was derived for developmental effects, a point of
departure (POD) was derived and presented for comparison purposes (see Section 5.2.1).
Evidence from epidemiological and animal studies suggests but is insufficient to infer that
PFHxS exposure might cause hepatic effects, specifically increases in serum biomarkers (see section
3.2.4). However, because of limitations and uncertainties in the currently available studies, a hazard
could not be clearly identified, and these data were not considered for use in deriving toxicity
xvi
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
values. While no reference dose (RfD) was derived for hepatic effects, a POD was derived and
presented for comparison purposes (see Section 5.2.1).
In addition, evidence from human and animal studies suggests but is insufficient to infer
that PFHxS exposure may cause neurodevelopmental and cardiometabolic effects in humans.
Lastly, although evidence from humans and or animals was also identified for
hematopoietic, reproductive, renal, and carcinogenic effects, the currently available evidence is
inadequate to assess whether PFHxS exposure may be capable of causing these health effects in
humans, and these outcomes were not considered for use in deriving toxicity values.
Table ES-1. Health effects with evidence available to synthesize and draw
summary judgments and derived toxicity values3
Organ/
system
Evidence
integration
judgment
Toxicity
value
Value
(mg/kg-d)
Confidence
UFa
UFh
UFs
UFl
UFd
UFC
Basis
Immune (i.e.,
developmental
immune)
Evidence
indicates
(likely)
Lifetime
osRfD
4 x 10"10
(RfD)
Medium
1
10
1
1
3
30
Decreased serum
anti-tetanus
antibody
concentration in
children at age 7 yr
(Grandiean et al..
2012: Budtz-
J0rgensen and
Grandjean, 2018)
Subchronic
osRfD
4 x 10"10
Medium
1
10
1
1
3
30
Decreased serum
anti-tetanus
antibody
concentration in
children at age 7 yr
(Grandiean et al..
2012: Budtz-
J0rgensen and
Grandiean. 2018)
Thyroid
Evidence
indicates
(likely)
Lifetime
osRfD
2 x 10"7
Medium
3
10
1
1
3
100
Decreased serum-
total T4 levels in F1
Wistar rats pups at
PND 16/17
(Ramh0i et al..
2018)
Subchronic
osRfD
2 x 10"7
Medium
3
10
1
1
3
100
Decreased serum-
total T4 levels in F1
Wistar rats pups at
PND 16/17
(Ramh0i et al..
2018)
RfD = reference dose (in mg/kg-d) for lifetime exposure; subchronic RfD = reference dose (in mg/kg-d) for less-than-
lifetime exposure; osRfD = organ-/system-specific reference dose (in mg/kg-d); UFA = animal to human uncertainty
factor; UFC = composite uncertainty factor; UFD = evidence base deficiencies uncertainty factor; UFH = human
variation uncertainty factor; UFL = LOAEL to NOAEL uncertainty factor; UFS = subchronic to chronic uncertainty
factor.
aA summary of pharmacokinetic parameters used for this evaluation is provided in Table 3-6 in Section 3.1.6,
Empirical Pharmacokinetic Analysis.
xvii
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
ES.l LIFETIME AND SUBCHRONIC ORAL REFERENCE DOSE (RfD) FOR NONCANCER EFFECTS
From the identified hazards with sufficient qualitative and quantitative information to
support the derivation of candidate lifetime values (i.e., immune and thyroid), decreased serum
anti-tetanus antibody concentrations in children (male and female) fGrandiean etal.. 2012: Budtz-
Targensen and Grandjean. 2018) was selected as the basis for the oral RfD of 4 x 10"10 mg/kg-day. A
BMDL^sd of 2.82 x 10"4 mg/L in serum was identified for this endpoint and was used as the
PODintemai- The human equivalent dose POD (PODhed) of 1.16 x 10"8 mg/kg-day was derived by
multiplying the PODintemai by the human clearance of 4.1 x 10"5 L/kg-day to estimate human
equivalent doses from an internal dose. The overall RfD for PFHxS was calculated by dividing the
PODhed by a composite uncertainty factor of 30 to account for interindividual differences in human
susceptibility (UFh = 10) and deficiencies in the toxicity evidence base (UFd = 3). The immune
organ-/system-specific osRfD is based on the lowest overall PODhed and UFc; therefore, the selected
RfD based on decreased serum anti-tetanus antibody concentration in children (a susceptible
lifestage for this effect) is considered protective of the observed health effects associated with
lifetime PFHxS exposure. The selection considered both available osRfDs as well as the overall
confidence and composite uncertainty for those osRfDs. The thyroid osRfD was based on
application of a composite uncertainty threefold greater than that applied in deriving the immune
osRfD (UFc = 100 for thyroid versus UFC= 30 for developmental immune effects). Further, when
comparing the sensitivity of thyroid and immune osRfDs, the thyroid value is 500-fold higher than
the developmental immune endpoint. Selection of the RfD on the basis of developmental immune
effects is presumed to be protective of possible thyroid and other potential adverse health effects
(including potential effects on birth weight and adverse hepatic effects) in humans. Finally, because
the developmental immune osRfD is based on effects observed in males and females, the overall
RfD would be protective for both sexes. The same study fGrandiean etal.. 2012: Budtz-largensen
and Grandjean. 2018) endpoint (decreased serum anti-tetanus antibody concentration in children)
and value were selected as the basis for the subchronic RfD of 4 x 10"10 mg/kg-day.
ES.2 CONFIDENCE IN THE ORAL REFERENCE DOSE (RfD) AND SUBCHRONIC RfD
The overall confidence in the RfD and subchronic RfD is medium and is driven by medium
confidence in the overall evidence base for immune effects, medium confidence in the fGrandiean et
al.. 2012: Budtz-large risen and Grandiean. 20181 study (HAWCJink), and medium confidence in
quantitation of the POD (see Section 5.2. and Table 5-8).
xviii
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
ES.3 NONCANCER EFFECTS FOLLOWING INHALATION EXPOSURE
No studies that examine toxicity in humans or experimental animals following inhalation
exposure are available and no acceptable physiologically based pharmacokinetic (PBPK) models
are available to support route-to-route extrapolation; therefore, no RfC was derived.
ES.4 EVIDENCE FOR CARCINOGENICITY
Under EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA. 20051. EPA concluded
there is inadequate information to assess carcinogenic potential for PFHxS by either the oral or
inhalation routes of exposure. This conclusion is based on the lack of adequate data to inform the
potential carcinogenicity of PFHxS in the database. This precludes the derivation of quantitative
estimates for either oral (oral slope factor [OSF]) or inhalation (inhalation unit risk [IUR])
exposure.
xix
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
1. OVERVIEW OF BACKGROUND INFORMATION
AND ASSESSMENT METHODS
A series of five PFAS assessments (Perfluorohexanesulfonic acid [PFHxS],
perfluorohexanoic acid [PFHxA], perfluorobutanoic acid [PFBA], perfluorononanoic acid [PFNA],
perfluorodecanoic acid [PFDA], and their associated salts; see December 2018 IRIS Outlook) is
being developed by the Integrated Risk Information System (IRIS) Program at the request of the
U.S. Environmental Protection Agency (EPA) national programs and regions. Appendix A is the
systematic review protocol for these five PFAS assessments. The protocol outlines the scoping and
problem-formulation efforts relating to these assessments, including a summary of other federal
and state reference values for PFHxS. The protocol also lays out the systematic review and dose-
response methods used to conduct this review (see also Section 1.2). This systematic review
protocol was released for public comment in November 2019 and was subsequently updated based
on those public comments. Appendix A includes a link to the updated protocol, including a
summary of the updates in the protocol history section (see Section 12). In addition to these
ongoing IRIS PFAS toxicity assessments, EPA's Office of Research and Development is carrying out
several other activities related to PFAS, including creation of PFAS systematic evidence maps
(SEMs) and consolidating and updating PFAS data on chemical and physical properties, human
health toxicity, and pharmacokinetics, as well as ecotoxicity.
1.1. BACKGROUND INFORMATION ON PERFLUOROHEXANESULFONIC
ACID (PFHxS)
Section 1.1 provides a brief overview of aspects of the physicochemical properties, human
exposure, and environmental fate characteristics of perfluorohexanesulfonic acid (PFHxS; CASRN
335-46-4), and its related salts that mightprovide useful context for this assessment This overview
is not intended to provide a comprehensive description of the available information on these topics.
The reader is encouraged to refer to the source materials cited below, more recent publications on
these topics, and authoritative reviews or assessments focused on these topics.
1.1.1. Physical and Chemical Properties
PFHxS and its related salts such as potassium, sodium, and ammonium PFHxS salts covered
in this assessment are members of the group per- and polyfluoroalkyl substances (PFAS). Buck et
al. (2011) defines PFAS as fluorinated substances that "contain 1 or more C atoms on which all the
H substituents (present in the nonfluorinated analogs from which they are notionally derived) have
been replaced by F atoms, in such a manner that they contain the perfluoroalkyl moiety CnF2n+i-)-"
1-1
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
More specifically, PFHxS is classified as a perfluoroalkane sulfonic acid [PFSA; (OECD. 20151].
PFSAs containing six or more perfluorinated carbons are considered long-chain PFASs fOECD.
2015: Buck etal.. 2011: ATSDR. 20211. Thus, PFHxS is a long-chain PFAS. The chemical structures of
PFHxS and its related salts are presented in Figure 1-1.2 The physical-chemical properties of PFHxS
and related salts are provided in Table 1-1.
0=S OH
II
O
PFHxS
355-46-4
X+ 0—s-
: 1
: 1
:
:
PFHxS
related salts
Figure 1-1. Chemical structure of PFHxS and related salts (see
https://comptox.epa.goV/dashboard/l. X represents the cations for potassium
(CASRN 3871-99-6), sodium (CASRN 82382-12-5), and ammonium (CASRN 68259-
08-5).
2While this figure shows the linear chemical structures, the assessment may also apply to other nonlinear
isomers of PFHxS, and related salts as described in the Executive Summary.
1-2
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 1-1. Physical-chemical properties of PFHxS and related salts3
Property (unit)
Value
PFHxS
355-46-4b
PFHxS
Potassium salt
3871-99-6c
PFHxS
Ammonium salt
68259-08-5c
PFHxS
Sodium salt
82382-12-5c
Molecular weight (g/mol)
400
438
417c*
422*
Melting point (°C)
190
273
111*
217*
Boiling point (°C)
246
303*
228*
238*
Density (g/cm3)
1.84*
1.84*
1.84*
1.84*
Vapor pressure (mm Hg)
8.10 x 10"9
8.19 x 10"9*
8.19 x 10"9*
8.19 x 10"9*
Henry's law constant
(atm-m3/mol)
1.94 x 10 10*
1.94 x 10 10*
1.94 x 10 10*
1.94 x 10 10*
Water solubility (mol/L)
6.08 x 10"4d
3.52 x 10"2*
6.10 x 10"4*
7.03 x 10"2*
pKa
0.14*
ND
ND
ND
LogP
2.20d
2.71*
3.48*
2.91*
Soil adsorption coefficient
(L/kg)
2,300*
2,300*
2,300*
2,300*
Bioconcentration factor (BCF)
175*
271*
271*
5.94*
*Average predicted value. These values are more uncertain and, in general, less reliable than experimental values.
ND = no data.
aThis information is provided as part of a general overview providing background context only and should not be
used for decision purposes. Up-to-date primary references should be consulted. A summary of pharmacokinetic
parameters used for this evaluation is provided in Table 3-6 and the method for calculating the human equivalent
dose values (prior to application of UFs) is described in Approach for Animal-Human Extrapolation of PFHxS
Dosimetry Section 3.1.7.
bCompTox Chemicals Dashboard (U.S. EPA, 2018a) for all values except pKa. The value of pKa was obtained from
ECHA: https://echa.europa.eu/documents/10162/lf48372e-97dd-db9f-4335-8cec7ae55eee. Questions and
corrections to the CompTox Chemicals Dashboard can be submitted at: https://comptox.epa.aov/dashboard/.
c (U.S. EPA, 2018a). Questions and corrections to the CompTox Chemicals Dashboard can be submitted at:
https://comptox.epa.gov/dashboard/.
dAs of April 2023 these values are indicated as 'experimental' in the CompTox Chemicals Dashboard (U.S. EPA,
2018a); however, they appear to be predicted values based on the citations provided, and therefore may be more
uncertain. Note that these values are not used for dosimetric extrapolation in this assessment, which was based
on available empirical pharmacokinetic data (see Section 3.1.7).
1.1.2. Sources, Production, and Use
PFAS are not naturally occurring in the environment (ATSDR. 2024). They are manmade
compounds that have been used widely over the past several decades in consumer products and
industrial applications because of their resistance to heat, oil, stains, grease, and water. PFHxS has
been used as a surfactant to make fluoropolymers, and in water and stain-protective coatings for
carpets, paper, packaging and textiles fNTP. 2018c: Norwegian Environment Agency. 20181. It may
also be present in certain industrial and consumer products, such as electronics, industrial fluids,
1-3
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
"food-contact papers, water-proofing agents, cleaning and polishing products either for intentional
uses (as surfactants or surface protection agents) or as unintentional impurities from industrial
production processes" fNorwegian Environment Agency. 20181. It has also been used in aqueous
film-forming foam (AFFF) for fire suppression fLaitinen etal.. 20141.
EPA has been working with companies in the fluorochemical industry since the early 2000s
to phase out the production and use of long-chain PFAS such as PFHxS
(https://www.epa.gov/assessing-and-managing-chemicals-under-tsca/risk-management-and-
polyfluoroalkyl-substances-pfassl. However, in addition to the environmental persistence of PFHxS
(see below), products containing PFHxS are still in use and may be imported into the United States;
thus, there may continue to be a source of environmental contamination due to disposal or
breakdown in the environment fKim and Kannan. 20071.
No chemical reporting data on production volume are available in EPA's ChemView fU.S.
EPA. 2019a) for PFHxS or its salts. As part of the National Defense Authorization Act for Fiscal Year
2020 (see Section 7321), 172 per- and polyfluoroalkyl substances including PFHxS were added to
the EPA's Toxic Release Inventory (TRI) list (https: //www.epa.gov/toxics-release-inventory-tri-
program /tri-listed-chemicalsl. The reporting requirements apply to a de minimus limit of 1% and a
manufacture process, or otherwise use a threshold of 100 pounds. Currently, incomplete
quantitative information is available in EPA's Toxic Release Inventory or other informational
repositories regarding PFHxS releases to the environment from facilities that manufacture, process,
or use imported/previously manufactured products that contain or dispose of imported/previously
manufactured products containing PFHxS.
1.1.3. Environmental Fate and Transport
PFAS, including PFHxS, are very stable and persistent in the environment (Harbison etal..
2015: ATSDR. 20241. and many are found worldwide in the environment, wildlife, and humans
(https://www.epa.gov/assessing-and-managing-chemicals-under-tsca/risk-management-and-
polyfluoroalkyl-substances-pfass). Long-chain PFAS have been found at sites, including private and
federal facilities, and have been associated with various sources, including AFFF for fire
suppression, and PFAS manufacturers and industries that use PFAS (e.g., textiles) (ATSDR. 20241.
PFAS that are released to air exist in the vapor phase in the atmosphere and resist
photolysis, but particle-bound concentrations have also been measured (Kim and Kannan. 20071.
In soil, the mobility of PFHxS depends on the soil adsorption coefficients (see Table 1-1).
Volatilization of PFHxS from moist soil is not expected to be an important transport process (NLM.
2013. 2016. 2017). Furthermore, PFHxS is expected to adsorb to suspended solids and sediments in
water fNLM. 2013. 2016. 20171.
1.1.4. Potential for Human Exposure and Populations with Potentially Greater Exposure
The general population may be exposed to PFAS via inhalation of indoor or outdoor air,
ingestion of drinking water and food, and dermal contact with PFAS-containing products fNLM.
1-4
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
2013. 2017: ATSDR. 20241. Exposure may also occur via hand-to-mouth transfer of materials
containing these compounds fATSDR. 20241. However, the oral route of exposure has been
considered the most important route of exposure among the general population. This conclusion is
based on several studies that have investigated the various routes of PFAS exposure f Sunderland et
al.. 20191.
The presence of PFHxS in human blood provides evidence of exposure among the general
population. PFHxS has been monitored in the human population as part of the National Health and
Nutrition Examination Survey (NHANES). PFHxS was measured in serum samples collected in
2013-2014 from more than 2,000 survey participants fCDC. 20221. The results of these analyses
are presented in Table 1-2.
Table 1-2. Serum PFHxS concentrations based on NHANES 2013-2014
data (ng/L)
Population group3
Value
Total population (N = 2,168)
Geometric mean
1.35
50th percentile
1.40
95th percentile
5.60
3 to 5 yr (N = 181)
Geometric mean
0.715
50th percentile
0.740
95th percentile
1.62
6 to 11 yr (N = 458)
Geometric mean
0.913
50th percentile
0.850
95th percentile
4.14
12 to 19 yr (N = 402)
Geometric mean
1.27
50th percentile
1.10
95th percentile
6.30
20 yr and older (N = 1,766)
Geometric mean
1.36
50th percentile
1.40
95th percentile
5.50
Source: CDC (2022). Fourth National Report on Human Exposure to Environmental Chemicals.
aThis table provides only general context on serum PFHxS levels from a single study and within a narrow time
period (environmental PFHxS levels are changing over time). Note that PFHxS is expected to bioaccumulate over a
lifetime (see Sections 1.1.3 and 3.1). Up-to-date information from authoritative bodies should be used in any
decisional context.
Air and Dust
PFHxS has not been evaluated under the Air Toxics Screening Assessment
fhttps://www.epa.gov/AirToxScreenl. However, PFHxS was measured at concentrations ranging
1-5
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
from less than the limit of detection to 1.56 pg/m3 in the vapor and particle phases of air samples
collected from an urban area of Albany, New York, in 2006 fKim and Kannan. 20071.
PFAS, including PFHxS, have also been measured in indoor air and dust and may be
associated with the indoor use of consumer products such as PFAS-treated carpets or other textiles
fATSDR. 20241. For example, Kato etal. f20091 analyzed dust samples collected from 39 homes in
the United States, United Kingdom, Germany, and Australia for PFAS, including PFHxS, which was
detected in 79.5% of the samples. Furthermore, indoor air samples (N = 4) from a town in Norway
had PFHxS mean concentrations of <4.1 pg/m3 for PFHxS (Barber etal.. 20071.
Water
EPA conducted monitoring for several PFAS, including PFHxS, in drinking water as part of
the third Unregulated Contaminant Monitoring Rule (UCMR) fU.S. EPA. 2016cl. Under the UCMR3,
all public water systems (PWSs) serving more than 10,000 people and a representative sample of
800 PWSs serving 10,000 or fewer people were monitored for 30 unregulated contaminants
between January 2013 and December 2015. PFHxS was among the 30 contaminants monitored and
was detected above the minimum reporting level (MRL) of 0.03 ng/L in 55 of the 4,920 PWSs tested
and in 207 of the 36,971 samples collected. Kim and Kannan (20071 analyzed lake water, rainwater,
snow, and surface water from Albany, New York, and reported concentrations of PFHxS ranging
from less than the LOD to 0.0135 ng/L. PFAS were detected at higher concentrations in
groundwater samples from an industrial site (3M Cottage Grove) in Minnesota. PFHxS was detected
in all seven wells that were sampled at concentrations ranging from 6.47 to 40 ng/L (WS. 20071 as
cited in ATSDR (20211. In additon UMCR5 is currently underway and will include monitoring
requirements for 29 PFAS compounds, including PFHxS, in both small and large public water
systems (over 10,000 systems combined), over the 2023 - 2025 timeframe. The minimum
reporting limit (MRL) for PFHxS under UCMR5 is 0.003 ng/L, which is 10-times lower than that
required under UCMR3, and sufficient to determine potential impacts below the MCL of 10 ng/L
(https://www.epa.gov/system/files/documents/2022-02/ucmr5-factsheet.pdf).
Aqueous Film-Forming Foam (AFFF) Training and Military Sites
The levels of PFHxS in soil and sediment surrounding perfluorochemical industrial facilities
have been measured at concentrations ranging from less than the LOD to 3,470 ng/g fATSDR.
20211. PFHxS was also detected at an Australian training ground where AFFFs had been used
(Baduel etal.. 2015). PFHxS was detected at 10 U.S. military sites in 76.9% of the surface soil
samples and 72.7% of sediment samples (ATSDR. 2021). Table 1-3 shows the concentration of
PFHxS in soil and sediment at these military sites.
1-6
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 1-3. PFHxS levels at 10 military installations
Media
Value
Surface Soil
Frequency of detection (%)
76.92
Median (ng/kg)
5.70
Maximum (ng/kg)
1,300
Subsurface Soil
Frequency of detection (%)
59.62
Median (ng/kg)
4.40
Maximum (ng/kg)
520
Sediment
Frequency of detection (%)
72.73
Median (ng/kg)
9.10
Maximum (ng/kg)
2,700
Surface Water
Frequency of detection (%)
88.00
Median (ng/L)
0.710
Maximum (ng/L)
815
Groundwater
Frequency of detection (%)
94.93
Median (ng/L)
0.870
Maximum (ng/L)
290
Source: Anderson et al. (2016); ATSDR (2024).
Other Exposures
Schecter etal. f20121 collected 10 samples of 31 food items from five grocery stores in
Texas and analyzed them for persistent organic pollutants, including PFHxS, which was detected in
cod fish at a concentration of 0.07 ng/g wet weight Stahl etal. (2014) characterized PFAS in
freshwater fish from 164 U.S. urban river sites and 157 Great Lakes sites. PFHxS was detected in
45% of the samples at maximum concentrations of 3.5 ng/g and method detection limit of
0.12 ng/g (Stahl etal.. 20141. PFHxS was not detected in U.S. grocery store finfish and shellfish
samples fRuffle etal.. 20201. Apart from fish, overall dietary data for the United States are limited.
Data from other countries (e.g., South Korea, Brazil, Saudi Arabia) suggest that long-chain PFAS
such as PFHxS can sometimes be detected in samples of food products including shellfish, dairy
products, meats, vegetables, food packaging materials, infant formula, and water (both tap and
bottled) (Surmaetal.. 2017: Perez etal.. 2014: MoretaandTena. 2014: Lakindetal.. 2023: Heo et
al.. 2014: Chen etal.. 2018b). The relevance of these detects (and the associated PFHxS levels) to
U.S. products is unknown.
Populations with Potentially Greater Exposures
Populations that may experience exposures greater than those of the general population
may include individuals in occupations that require frequent contact with PFHxS-containing
1-7
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
products, such as individuals who install and treat carpets or firefighters f ATSDR. 2024). Rotander
etal. f2015al analyzed serum samples from 149 Australian firefighters at an AFFF training facility.
Mean and median PFHxS concentrations were 10 to 15 times higher than those of the general
population of Australia and Canada. Laitinen et al. f20141 evaluated eight firefighters' exposure to
PFHxS after three training sessions in Finland in which AFFF had been used. The authors found that
the firefighters "serum PFHxS concentrations seemed to increase during the three training sessions
although it was not the main PFAS used in AFFF." Populations living near fluorochemical facilities
where environmental contamination has occurred may also be more highly exposed (ATSDR.
20211.
Populations that rely primarily on seafood for most of their diet, possibly including some
Native American tribes f Byrne etal.. 20171. may also be disproportionately exposed to PFHxS.
Christensen et al. f20171 and Haugetal. f20101 used data on serum PFAS levels and 30-day self-
reported fish and shellfish ingestion rates from NHANES 2007-2014 to explore potential
relationships between PFAS exposures and fish consumption. PFHxS was detected in the serum of
99% of the NHANES participants, and after adjusting for demographic characteristics shellfish
consumption was associated with elevated levels of PFHxS f Christensen etal.. 20171.
1.2. SUMMARY OF ASSESSMENT METHODS
The methods used to conduct this systematic review and dose-response analysis are
summarized in the remainder of this section. A more detailed description of the methods for each
step of the assessment development process is provided in the systematic review protocol released
in 2019 (see Appendix A); the literature inventory for PFHxS in the protocol was not updated after
its release (see Section 2.1). The protocol includes additional problem-formulation details,
including the specific aims and key science issues identified for this assessment
1.2.1. Literature Search and Screening
The detailed search approach, including the query strings and populations, exposures,
comparators, and outcomes (PECO) criteria (see Table 1-4), are provided in Appendix B. The results
of the literature search and screening efforts are documented in Section 2.1. Briefly, a literature
search was first conducted in 2017, and regular yearly updates are performed. The most recent
literature search update that was fully incorporated into the assessment is from April 2022. A
literature search was also performed in April of 2023. However, only studies through April 2022
are fully incorporated into the assessment. Studies available after April 2022 were only fully
incorporated if they would have a material impact on the assessment conclusions (for additional
details please see Appendix B and Table B-5). The results of this literature update and any
additional unscreened studies identified during public comment were screened against the PECO
criteria and presented in a table that was included as an Appendix to the assessment (Appendix B,
Table B-5). The table provides the identified studies that met PECO criteria or certain supplemental
1-8
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
evidence categories (i.e., in vivo mechanistic or MOA studies, including non-PECO routes of
exposure and populations; in vitro and in silico models; and ADME and pharmacokinetic studies)
and EP As judgment on whether the studies have a material impact on the assessment conclusions
(i.e., identified hazards or toxicity values) presented in the public comment draft The external peer
reviewers were asked to consider EPA's disposition of these newly identified studies and make
recommendations, as appropriate (see Charge Question 1).
The literature search queried the following databases (no date or language restrictions
were applied):
• PubMed (National Library of Medicine)
• Web of Science fThomson Reuters!
• Toxline fNational Library of Medicine 1
• TSCATS (Toxic Substances Control Act Test Submissions)
In addition, relevant literature not found through database searching was identified by:
• Review of citations in studies meeting the PFHxS PECO criteria or published reviews of
PFHxS; finalized or publicly available U.S. federal and international assessments (e.g., the
2021 Agency for Toxic Substances and Disease Registry [ATSDR] Toxicological Profile for
Perfluoroalkyls).
• Searches of published PFAS Systematic Evidence Maps (SEMs) fPelch etal.. 2022: Carlson et
al.. 20221 starting in 2021.
• Review of studies submitted to federal regulatory agencies and brought to the attention of
EPA. For example, studies submitted to EPA by the manufacturers in support of
requirements under the Toxic Substances Control Act (TSCA).
• Identification of studies during literature screening for other EPA PFAS assessments. For
example, epidemiology studies relevant to PFHxS were sometimes identified by searches
focused on one of the other four PFAS currently being assessed by the Integrated Risk
Information System (IRIS) Program.
• Other gray literature (e.g., primary studies not indexed in typical databases, such as
technical reports from government agencies or scientific research groups; unpublished
laboratory studies conducted by industry; or working reports/white papers from research
groups or committees) brought to the attention of EPA.
All literature is tracked in the U.S. EPA Health and Environmental Research Online (HERO)
database (https://heronet.epa.gov/heronet/index.cfm/proiect/page/proiect id/2630). The PECO
criteria (see Table 1-4) identify the evidence that addresses the specific aims of the assessment and
to focus the literature screening including study inclusion/exclusion.
1-9
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 1-4. Populations, exposures, comparators, and outcomes (PECO) criteria
PECO
element
Evidence
Populations
Human: Any population and lifestage (occupational or general population, including children and other
sensitive populations). The following study designs will be included: controlled exposure, cohort, case
control, and cross sectional. (Note: Case reports and case series will be tracked as potential supplemental
material.)
Animal: Nonhuman mammalian animal species (whole organism) of any lifestage (including preconception,
in utero, lactation, peripubertal, and adult stages).
Other: In vitro, in silico, or nonmammalian models of genotoxicity. (Note: Other in vitro, in silico, or
nonmammalian models will be tracked as potential supplemental material.)
Exposures
Human: Studies providing quantitative estimates of PFHxS exposure based on administered dose or
concentration, biomonitoring data (e.g., urine, blood, or other specimens), environmental or
occupational-setting measures (e.g., water levels or air concentrations, residential location and/or duration,
job title, or work title). (Note: Studies that provide qualitative, but not quantitative, estimates of exposure
will be tracked as supplemental material.)
Animal: Oral or inhalation studies including quantified exposure to PFHxS based on administered dose,
dietary level, or concentration. (Note: Nonoral and noninhalation studies will be tracked as potential
supplemental material). PFHxS mixture studies are included if they employ an experimental arm that
involves exposure to a single PFHxS. (Note: Other PFHxS mixture studies are tracked as potential
supplemental material.)
Studies must address exposure to following: PFHxS (CASRN 355-46-4), PFHxS potassium salt (CASRN 3871-
99-6) or PFHxS ammonium salt (CASRN 68259-08-5).
Comparators
Human: A comparison or reference population exposed to lower levels (or no exposure/exposure below
detection levels) or for shorter periods of time.
Animal: Includes comparisons to historical controls3 or a concurrent control group that is unexposed,
exposed to vehicle-only or air-only exposures. (Note: Experiments including exposure to PFHxS across
different durations or exposure levels without including one of these control groups will be tracked as
potential supplemental material [e.g., for evaluating key science issues; Section 2.4 of the protocol].)
Outcomes
All cancer and noncancer health outcomes. (Note: Other than genotoxicity studies, studies including only
molecular endpoints [e.g., gene or protein changes; receptor binding or activation] or other nonphenotypic
endpoints addressing the potential biological or chemical progression of events contributing toward toxic
effects will be tracked as potential supplemental material [e.g., for evaluating key science issues; Section
2.4 of the protocol].)
aWhile concurrent controls are strongly preferred, historical controls can be useful, for example in the evaluation
of rare tumors or when considering the similarity between current experimental animals and laboratory
conditions to historical controls. However, use of historical controls only is noted as a limitation during study
evaluation, so concurrent and historical controls are not considered equal. It is noted that no studies using only
historical controls were identified in the literature searches for PFHxS.
In addition to those studies meeting the PECO criteria and studies excluded as not relevant
to the assessment, studies containing supplemental material potentially relevant to the specific
aims of the assessment were inventoried during the literature screening process. Although these
studies did not meet PECO criteria, they were not excluded. Rather, they were considered for use in
addressing the identified key science issues (see Appendix A, Section 2.4) and other potential
1-10
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
scientific uncertainties identified during assessment development but unanticipated at the time of
protocol posting. Studies categorized as "potentially relevant supplemental material" included the
following:
• In vivo mechanistic or mode of action studies, including non-PECO routes of exposure
(e.g., intraperitoneal injection) and populations (e.g., nonmammalian models)
• In vitro and in silico models
• Absorption, distribution, metabolism, and excretion (ADME) and pharmacokinetic studies
(excluding models)3
• Exposure assessment or characterization (no health outcome) studies
• Human case reports or case series studies
The literature was screened by two independent reviewers with a process for conflict
resolution, first at the title and abstract level and subsequently the full-text level, using structured
forms in DistillerSR (Evidence Partners; https://www.distillersr.com/products/distillersr-
systematic-review-software ). Literature inventories for PECO-relevant studies and studies tagged
as "potentially relevant supplemental material" during screening were created to facilitate
subsequent review of individual studies or sets of studies by topic-specific experts.
1.2.2. Evaluation of Individual Studies
The detailed approaches used for the evaluation of epidemiologic and animal toxicological
studies used in the PFHxS assessment are provided in the systematic review protocol (see
Appendix A, Section 6). The general approach for evaluating PECO-relevant health effect studies is
the same for epidemiology and animal toxicological studies, although the specifics of applying the
approach differ; thus, they are described in detail in Appendix A (see Sections 6.2 and 6.3,
respectively). Approaches for study evaluation for mechanistic studies are described in detail in
Appendix A (see Section 6.5).
The key concerns for the review of epidemiology and animal toxicological studies are
potential bias (systematic errors or deviations from the truth related to internal validity that affect
the magnitude or direction of an effect in either direction) and insensitivity (factors that limit the
ability of a study to detect a true effect and can lead to a false negative). For example, any type of
random measurement error that may lead to attenuation of study results (i.e., bias toward the null).
In evaluating individual studies, two or more reviewers independently arrived at judgments
regarding the reliability of the study results (reflected as study confidence determinations; see
below) with regard to each outcome or outcome grouping of interest; thus, different judgments
3Given the known importance of ADME data, this supplemental tagging was used as the starting point for a
separate screening and review of pharmacokinetics data (see Appendix A, Section 9.2 for details).
1-11
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
were possible for different outcomes within the same study. The results of these reviews were
tracked within EPA's version of the Health Assessment Workplace Collaboration (HAWC). To
develop these judgments, each reviewer assigned a category of good, adequate, deficient (or not
reported, which generally carried the same functional interpretation as deficient), or critically
deficient (listed from best to worst methodological conduct; see Appendix A, Section 6 for
definitions) related to each evaluation domain representing the different characteristics of the
study methods that were evaluated based on the criteria outlined in HAWC.
Once all evaluation domains were evaluated, the reviewers collectively considered the
identified strengths and limitations to reach a final study confidence classification:
• High confidence: No notable deficiencies or concerns were identified; the potential for bias
is unlikely or minimal, and the study used sensitive methodology.
• Medium confidence: Possible deficiencies or concerns were noted, but the limitations are
unlikely to be of a notable degree or to have a notable impact on the results.
• Low confidence: Deficiencies or concerns were noted, and the potential for bias or
inadequate sensitivity could have a significant impact on the study results or their
interpretation. Low confidence results were given less weight than high or medium
confidence results during evidence synthesis and integration (see Sections 1.2.4 and 1.2.5).
• Uninformative: Serious flaw(s) were identified that make the study results unusable.
Uninformative studies were not considered further, except to highlight possible research
gaps.
Using the HAWC platform (and conflict resolution by an additional reviewer, as needed), the
reviewers reached a consensus judgment regarding each evaluation domain and overall
(confidence) determination. The specific limitations identified during study evaluation were carried
forward to inform the synthesis (see Section 1.2.4) within each body of evidence for a given health
effect (i.e., study confidence determinations were not used to inform judgments in isolation).
Additional Epidemiology Considerations
While the detailed methods for epidemiology study evaluation are described in the IRIS
PFAS systematic review protocol ((U.S. EPA. 2021c) see Appendix A, Section 6.2.1), a few
considerations have been developed further; these are described here.
As noted above, study sensitivity is an important consideration given that it could lead to
false negative (i.e., null) results (Type II error) if a study is underpowered or not designed with
adequate sensitivity to detect an association that may exist A key element for study sensitivity,
along with others described in the systematic review protocol, is whether exposure
contrasts/gradients are sufficient across populations to detect differences in risk. For example,
measurement error resulting in inaccurate exposure estimates can lead to exposure
1-12
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
misclassification and influence the ability to detect an association as well as an exposure-response
relationship that may be evident of a biologic gradient
Confounding across PFAS is a potential source of uncertainty when interpreting the results
of epidemiology studies of individual PFAS (e.g., quantifying the effect of an individual PFAS can
potentially be confounded by other PFAS). For confounding to occur, co-pollutants would have to
be associated with PFAS of interest, associated with the endpoint, and not act as an intermediate in
the causal pathway. One way to begin to assess whether co-exposure is occurring is through
examination of correlations. While some PFAS pairs have correlation coefficients consistently above
0.6 (e.g., PFNA and PFDA), the correlations for most PFAS, including PFHxS, vary from 0.1 to 0.6
depending on the study (see Appendix A, Section 6). For this reason, it was not considered
appropriate to assume that co-exposure to other PFAS was necessarily an important confounder in
all studies. The potential for confounding across PFAS is incorporated in individual study
evaluations and assessed across studies in evidence synthesis. In most studies, it is difficult to
determine the likelihood of confounding without considering additional information not typically
included in individual study evaluation (e.g., associations of other PFAS with the outcome of
interest and correlation profiles of PFAS within and across studies). In addition, even when this
information is considered or the study authors perform analyses to adjust for other PFAS, it is often
not possible to fully disentangle the associations due to high correlations. This challenge stems
from the potential for amplification bias in which bias can occur following adjustment of highly
correlated PFAS (Weisskopf et al.. 2018). Thus, in most studies, there may be some residual
uncertainty about the risk of confounding by other PFAS. A "Good" rating for the confounding
domain is reserved for situations in which there is minimal concern for substantial confounding
across PFAS as well as for other sources of confounding. Examples that would obtain this rating
include results for a PFAS that predominates in a population (such as a contamination event) or
studies that demonstrate robust results following multi-PFAS adjustment (i.e., similar results to
single-PFAS models), which would also indicate minimal concern for amplification bias. Because of
the challenge in evaluating individual studies for confounding across PFAS, this issue is also
assessed across studies during the evidence synthesis phase, as described in the systematic review
protocol (see link in Appendix A, Section 6.2), primarily when there is support for an association
with adverse health effects in the epidemiology evidence (i.e., moderate, or robust evidence in
humans, as described below). Analyses used include comparing results across studies in
populations with different PFAS exposure mixture profiles, considering results of multipollutant
models when available, and examining strength of associations for other correlated PFAS. In
situations for which there is considerable uncertainty regarding the impact of residual confounding
across PFAS, a factor is captured that decreases the overall strength of evidence (see link in
Appendix A, Section 10).
1-13
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
1.2.3. Data Extraction
The detailed data extraction approach is provided in Appendix A, Section 8. Briefly, data
extraction and content management were carried out using HAWC for all health effects for animal
studies and some health effects for epidemiological studies. Data extraction elements collected from
epidemiological, controlled human exposure, animal toxicological, and in vitro studies are
described in HAWC (https: //hawcprd.epa.gov/about/). For epidemiological studies not extracted
in HAWC, extraction was performed into Word tables and the extraction elements depended on
information needed for presentation. Not all studies that meet the PECO criteria went through data
extraction: studies evaluated as being uninformative were not considered further and therefore did
not undergo data extraction, and outcomes determined to be less relevant during PECO refinement
did not go through data extraction. The same was true for low confidence studies when medium and
high confidence studies (e.g., on an outcome) were available. All findings are considered for
extraction, regardless of the statistical significance of their findings. The level of extraction for
specific outcomes within a study may differ (i.e., ranging from a narrative to full extraction of
dose-response effect size information). For quality control, data extraction was performed by one
member of the evaluation team and independently verified by at least one other member.
Discrepancies in data extraction were resolved by discussion or consultation within the
evaluation team.
1.2.4. Evidence Synthesis and Integration
For the purposes of this assessment, evidence synthesis and integration are considered
distinct but related processes (see Appendix A, Sections 9 and 10 for full details). For each assessed
health effect, the evidence syntheses provide a summary discussion of each body of evidence
considered in the review that directly informs the integration across evidence to draw an overall
judgment for each health effect The available human and animal evidence pertaining to the
potential health effects are synthesized separately, with each synthesis providing a summary
discussion of the available evidence that addresses considerations regarding causation that are
adapted from Hill (1965). Mechanistic evidence is also synthesized as necessary to help inform key
decisions regarding the human and animal evidence; processes for synthesizing mechanistic
information are covered in detail in Appendix A, Section 9.2.
The syntheses of the human and animal health effects evidence focus on describing aspects
of the evidence that best inform causal interpretations, including the exposure context examined in
the sets of studies. The evidence synthesis is based primarily on studies of high and medium
confidence. Low confidence studies could be used if few or no studies with higher confidence are
available to help evaluate consistency, or if the study designs of the low confidence studies address
notable uncertainties in the set of high or medium confidence studies on a given health effect. If low
confidence studies are used, a careful examination of the study evaluation and sensitivity with
potential effects on the evidence synthesis conclusions will be included in the narrative. When
1-14
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
possible, results across studies are compared using graphs and charts or other data visualization
strategies. The synthesis of mechanistic information informs the integration of health effects
evidence for both hazard identification (e.g., biological plausibility or coherence of the available
human or animal evidence; inferences regarding human relevance, or the identification of
susceptible populations and lifestages across the human and animal evidence) and dose-response
evaluation (e.g., selection of benchmark response levels, selection of uncertainty factors).
Evaluations of mechanistic information typically differ from evaluations of phenotypic evidence
(e.g., from routine toxicological studies) primarily because mechanistic data evaluations consider
the support for and involvement of specific events or sets of events within the context of a broader
research question (e.g., support for a hypothesized mode of action; consistency with known
biological processes), rather than evaluations of individual apical endpoints considered in relative
isolation.
Following the synthesis of human and animal health effects data and mechanistic data,
integrated judgments are drawn across all lines of evidence for each assessed health effect. During
evidence integration, a structured and documented two-step process is used, as follows:
Building from the separate syntheses of human and animal evidence, the strength of the
evidence from the available human and animal health effect studies are summarized in parallel, but
separately, using a structured evaluation of an adapted set of considerations first introduced by Sir
Bradford Hill (Hill. 1965). This process is similar to that used by the Grading of Recommendations
Assessment, Development, and Evaluation (GRADE) (Schiinemann etal.. 2011: Morgan etal.. 2016:
Guvattetal.. 2011). which arrives at an overall integration conclusion based on consideration of the
body of evidence. These summaries incorporate the relevant mechanistic evidence (or mode of
action [MOA] understanding) that informs the biological plausibility and coherence within the
available human or animal health effect studies. The terms associated with the different strength-
of-evidence judgments within evidence streams are robust, moderate, slight, indeterminate, and
compelling evidence of no effect.
The animal, human, and mechanistic evidence judgments are then combined to draw an
overall judgment that incorporates inferences across evidence streams. Specifically, the inferences
considered during this integration include the human relevance of the animal and mechanistic
evidence, coherence across the separate bodies of evidence, and other important information
(e.g., judgments regarding susceptibility). Note that without evidence to the contrary, the human
relevance of animal findings is assumed. The final output is a summary judgment of the evidence
base for each potential human health effect across evidence streams. The terms associated with
these summary judgments are evidence demonstrates, evidence indicates (likely), evidence suggests,
evidence inadequate, and strong evidence of no effect. The decision points within the structured
evidence integration process are summarized in an evidence profile table for each considered
health effect
1-15
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
As discussed in the protocol (see Appendix A), the methods for evaluating the potential
carcinogenicity of PFAS follow processes laid out in the EPA cancer guidelines fU.S. EPA. 20051:
however, for PFHxS, data relevant to cancer were sparse and did not allow for such an evaluation
(see Appendix A, Section 3.3).
1.2.5. Dose-Response Analysis
The details for the dose-response employed in this assessment can be found in Appendix A,
Section 11. Briefly a dose response assessment was performed for noncancer health hazards,
following exposure to PFHxS via the oral route, as supported by existing data. For oral noncancer
hazards, oral reference doses (RfDs) are derived when possible. 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). The derivation of a reference value like the RfD
depends on the nature of the health hazard conclusions drawn during evidence integration. For
noncancer outcomes, a dose response assessment was conducted for evidence integration
conclusions of evidence demonstrates or evidence indicates (likely). In general, toxicity values
are not developed for noncancer hazards with evidence suggests conclusions (see Appendix A,
Section 10.2 for exceptions). Consistent with EPA practice, the PFHxS assessment applied a two-
step approach for dose response assessment that distinguishes analysis of the dose response data
in the range of observation from any inferences about responses at lower environmentally relevant
exposure levels (U.S. EPA. 2005. 2012):
• Within the observed dose range, the preferred approach was to use dose-response
modeling to incorporate as much of the dataset as possible into the analysis. This modeling
to derive a point of departure (POD) ideally includes an exposure level near the lower end
of the range of observation, without significant extrapolation to lower exposure levels.
• As derivation of cancer risk estimates and reference values nearly always involves
extrapolation to exposures lower than the POD; the approaches to be applied in these
assessments are described in more detail in Appendix A, Section 11.2.
When sufficient and appropriate human and laboratory animal data are available for the
same outcome, human data are generally preferred for the dose-response assessment because use
of human data eliminates the need to perform interspecies extrapolations. For reference values,
this assessment will derive a candidate value from each suitable dataset. Evaluation of these
candidate values will yield a single organ/system-specific value for each organ/system under
consideration from which a single overall reference value will be selected to cover all health
outcomes across all organs/systems. While this overall reference value represents the focus of
these dose-response assessments, the organ/system-specific values can be useful for subsequent
cumulative risk assessments that consider the combined effect of multiple PFAS (or other agents)
1-16
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
acting at a common organ/system. For noncancer toxicity values, uncertainties in these estimates
are characterized and discussed.
For dose-response purposes, EPA has developed a standard set of models
fhttp: //www.epa.gov/bmds] that can be applied to typical datasets, including those that are
nonlinear. In situations for which there are alternative models with significant biological support
(e.g., pharmacodynamic models), those models are included as alternatives in the assessment(s)
along with a discussion of the models' strengths and uncertainties. EPA has developed guidance on
modeling dose-response data, assessing model fit, selecting suitable models, and reporting
modeling results [see the EPA Benchmark Dose Technical Guidance fU.S. EPA. 20121], For each
modeled response, a POD from the observed data was estimated to mark the beginning of
extrapolation to lower doses. The POD is an estimated dose (expressed in human-equivalent terms)
near the lower end of the observed range without significant extrapolation to lower doses. The POD
is used as the starting point for subsequent extrapolations and analyses. For noncancer effects, the
POD is used in calculating the RfD.
1-17
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
2. LITERATURE SEARCH AND STUDY EVALUATION
RESULTS
2.1. LITERATURE SEARCH AND SCREENING RESULTS
The database searches yielded 4,432 records; of these records 162 were identified from
additional sources, such as posted National Toxicology Program (NTP) study tables and during
review of reference lists from other authoritative sources (ATSDR. 2021) (see Figure 2-1). No
studies were submitted to EPA. After deduplication, 1,935 unique records were identified, 862 were
excluded during title and abstract screening, and 806 were reviewed at the full text level. Of the 806
screened at the full text level, 445 were considered to meet the populations, exposures,
comparators, and outcomes (PECO) eligibility criteria (see Table 1-4). The studies meeting PECO at
the full text level included 415 epidemiologic studies and 20 animal studies. High-throughput
screening data on perfluorohexane sulfonate (PFHxS) are currently available from the EPA's
Chemicals Dashboard (U.S. EPA. 2019b) and relevant information is presented and analyzed in
Appendix d (see Section 3).
2-1
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
PFHxS
Literature Searches (through April 2022)/"'
PubMed
(n = 1002)
WOS
(n =1064 )
ToxLine
(n= 566)
TSCATS
(n = 10)
Other
Pelch database (n=582)
SCOPUS tn = 1208)
Additional search
strategies tn=162)
i
TITLE AND ABSTRACT
Title & Abstract Screening
(1935 records after duplicate removal)
FULL TEXT SCREENING
Full-Text Screening
(n = 806 )
Studies Considered Further (n =445)
Human health effects studies (n = 415)
Animal health effect studies (n = 20)
Genotoxicity studies (n = 0)
PBPK models (n = 10)
Excluded (n= 862)
Not relevant to PECO (n = 862)
Excluded (n=39)
* not relevant to PECO (n = 13), review/reg
docs (n = 26), abstract-only (n = 0)
Tagged as Supplemental (n= 613)
mechanistic or MOA (n = 56), ADME (n = 45),
qualitative exposure only (n = 182), mixture-only
(n = 6), non-PECO route of exposure (n = 15), case
report or case study (n = 0); in vitro or in silico
(non-genotox) studies (n= 68); Environmental fate
or occurrence (n =61); manufacture, engineering,
use treatment, remediation, or laboratory
methods (n =60); Other assessments or records
with no original data (n = 51); Ecotoxicity studies
(n =80); Susceptible populations (rt =5)
Figure 2-1. Literature search for perfluorohexanesulfonic acid and related
salts.
2-2
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
2.2. STUDY EVALUATION RESULTS
Four hundred fifteen epidemiologic studies were identified that met the PECO criteria and
report on the potential association between PFHxS and human health effects. The database of
animal toxicity studies for PFHxS consists of two short-term oral exposure studies using rats (NTP.
2018a: 3M. 2000al. two short-term study in mice (Vibergetal.. 2013: Das et al.. 20171. three
subchronic studies using mice (Yin etal.. 2021: He etal.. 2022: Biiland etal.. 20111. and five
multigenerational studies using rats or mice f T etzlaff et al.. 2 0 21: Ramhai etal.. 2018: Ramhai etal..
2020: Marques etal.. 2021: Chang etal.. 2018: Bute nhoff etal.. 2009: 3M. 20031 and one chronic
study in C57BL/6j mice fPfohl etal.. 20201.
Graphical representations of outcome-specific study evaluations are presented and
discussed within the hazard sections (see Sections 3.2.1-3.3.1). In cases for which a study was rated
medium or low confidence for one or more of the evaluated outcomes, the specific limitations are
explained in the synthesis section(s). Detailed rationales for each domain and overall confidence
rating are available in Health Assessment Workspace Collaborative fHAWCl.
2-3
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3. PHARMACOKINETICS, EVIDENCE SYNTHESIS,
AND INTEGRATION
3.1. PHARMACOKINETICS
The following sections review the scientific evidence for the absorption, distribution,
metabolism, and excretion (ADME) of perfluorohexanesulfonic acid (PFHxS). In general, the
evidence described below demonstrates that PFHxS has ADME characteristics comparable with
other perfluoroalkyl acids (PFAA) (of which PFSAs are a subclass) that are readily absorbed in the
gastrointestinal tract following oral exposure irrespective of sex or species.
Multiple PFHxS isomers have been identified. Benskin etal. (2009) found evidence of three
PFHxS isomers as minor fractions in a PFOS standard generated using electrochemical fluorination.
They identified the most prevalent of these as the linear isomer (n-PFHxS) and the two others as
branched isomers. The branched isomers were present as a small fraction relative to the linear
isomer 4 but were a majority of the PFHxS found in urine 3 days after dosing, as branched isomers
are eliminated more quickly than n-PFHxS. By day 38, the branched isomers, not including n-PFHxS,
were essentially absent in blood (Benskin et al.. 2009). Some pharmacokinetic studies specifically
identified the isomer used (e.g., Sundstrom etal. (2012) used the linear isomer), but others did not.
Results from other studies based on measured PFHxS concentrations in blood were therefore
assumed to represent n-PFHxS unless otherwise specified. The current evidence is too sparse to
draw separate judgments for branched and linear isomers, although this review of PFHxS ADME is
interpreted as primarily focused on evidence for n-PFHxS. While branched PFHxS isomers are likely
to have many similar pharmacokinetic (and pharmacodynamic) properties as n-PFHxS, their
contribution to the summary information below (and the toxicity data in Section 3.2) cannot
currently be specified.
Both animal and human data suggest that PFHxS has a high affinity for protein binding.
Bischel etal. f20111 measured 99% bound in a solution of bovine serum albumin and Kim et al.
f2018bl estimated less than 0.08% free in rat plasma and 0.03% free in human plasma. Significant
sex differences in urinary excretion have been reported, suggesting hormonal regulation of
transporters involved in renal reuptake (Yang etal.. 2009). The PFHxS serum concentrations
reported at the end of the 28-day NTP bioassay (NTP. 2019) were in fact strongly suggestive both of
sex differences and of saturable resorption in the elimination of PFHxS by rats (see Figure 3-1).
While the dose range was greater for female rats (0-50 mg/kg-day) than for male rats (0-
4Based on peak height in a representative chromatogram shown in Figure 1 of Benskin et al. (2009).
quantified by digitization of the published plot, the two branched isomers had concentrations of about 8%
and 15% of the linear isomer in the dosing solution.
3-1
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
10 mg/kg-day), it is still clear thatplasma levels in the males at 10 mg/kg-day (198 mg/L) were
three times higher than the plasma concentration in females given 12 mg/kg-day (64 mg/L) at the
end of the 28-day study. This sex difference was clearly reflected by the differences in clearance and
half-life for male and female rats seen in multiple studies, discussed subsequently. The NTP f20191
data also clearly indicated strong pharmacokinetic nonlinearity (see Figure 3-1). If absorption and
clearance were independent of concentration, the plasma concentrations in Figure 3-1 would be
approximately linear with dose. The PK data discussed below also indicated nonlinearity in either
or both the absorption and clearance. In particular, Huang etal. f2019al estimated clearance levels
1.5 to 2 times higher after a 32 mg/kg dose than after 4 and 16 mg/kg and a decrease in
bioavailability of about 50% between 4 and 32 mg/kg in both male and female rats. However,
because those PK experiments only used a single dose, they may not have achieved plasma
concentrations high enough to demonstrate the extent of the difference in clearance that might be
needed to explain the NTP data.
3-2
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
i r
4 6
Dose (mg/kg/d)
10
Figure 3-1. Observed end-of-study of PFHxS in female and male rats in the NTP
bioassay (NTP. 20191 as a function of dose. The plasma concentrations were
measured 1 day after the final dose, i.e., day 29. While the two data sets look similar
as shown with their respective dose scales, note that significant saturation occurs in
male rats by a dose of 5 mg/kg-day, where the plasma concentration is 80% of that
observed at the highest administered dose, while a dose of about 20 mg/kg-day is
needed to achieve the same degree of saturation in females, while the highest
concentration in males is twice that in females. The similarity in shape may occur
because of binding of PFHxS to the same transporter determines the nonlinearity in
both sexes.
3-3
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Serum binding also appears to limit distribution of PFHxS into other tissues, with the tissue:
blood or plasma ratio reported as less than 0.2 for liver and much lower for all other tissues (Kimet
al.. 2016b: Benskin et al.. 20091. After the liver, the next highest tissue levels were observed in
kidney, lung heart, and spleen. Similar to other PFAAs, PFHxS has been presumed to be
metabolically inert, but Sundstrom et al. f20121 only recovered 45%-55% of material between
serum, liver, urine, and feces 96 hours after dosing to Sprague-Dawley rats. The majority (~90%) of
PFHxS was excreted in the urine rather than the feces (Kim etal.. 2018b).
A pharmacokinetic (PK) approach was used to extrapolate toxicity points of departure from
animal PFHxS doses and human blood PFHxS levels to a human equivalent (external oral) dose. A
review of the ADME information for rats and humans directly informed the PK approach. Although
no endpoints in mice or monkeys were advanced for dose-response modeling, evaluation of ADME
in those species provided a broader context for interpreting the results in rats and humans. For
example, to what extent might significant differences between PK in male and female rats be
predictive of possible sex differences in humans? Differences or similarities between rats and
monkeys can likewise be indicative of the comparison between rats and humans.
Two key parameters determined were clearance (CL; L/kg-day) and volume of distribution
(Vd; L/kg). For convenience, the following analysis of published data used units of mL/kg-day.
Options for physiologically based pharmacokinetic (PBPK) and PK modeling were evaluated (see
Section 3.1.5). That evaluation informed the specific choice for dose extrapolation described in
Approach for Animal-Human Extrapolation of PFHxS Dosimetry in Section 3.1.7), while the
literature used to support the selection of the PK parameters and rationale for the approach used
are discussed in the relevant pharmacokinetics sections below.
3.1.1. Absorption
For the most part, PFHxS data showed near complete absorption after oral dosing. Kim et al.
(2016b) estimated total AUC in blood (AUCo-oo) that was greater after oral compared with IV doses
(4 mg/kg PFHxS) in both male and female rats. This result is counter to general pharmacokinetic
understanding, which assumes that the oral AUC will be lower than the IV AUC because of
incomplete absorption in the gastrointestinal tract These results may have been an artifact of
experimental variability and of the PK analysis used but they indicated complete absorption. Kim et
al. (2018b) then estimated ~90% absorption in female SD rats (92% and 88% absorption at 1 and
4 mg/kg doses, respectively) and 96% in male SD rats (10 mg/kg dose) based on observations to
14 days postexposure. While Sundstrom et al. (2012) showed results indicating only 50% oral
uptake in SD rats, these results were based on only two animals for the oral PK and observations
only to 24-hour post-dose, and so were more uncertain. Huang etal. (2019a) estimated a decline in
the fraction of PFHxS absorbed with increasing dose in rats: 98%, 82%, and 52% absorbed in males
and apparent values of 142%, 112%, and 71% in females at respective doses of 4,16, and
32 mg/kg. As noted above, reduced absorption at higher doses would explain in part the observed
dose-dependence seen in Figure 3-1.
3-4
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
While the results discussed above indicate a decrease in bioavailability at higher doses,
pharmacokinetic extrapolation from animals to humans is focused on low doses for which most of
the available data indicated complete absorption, if not greater bioavailability after oral exposure
than IV dosing. A more comprehensive computational analysis of the PK data was conducted (see
Section 3.1.6), including consideration of less than 100% bioavailability; however, that analysis was
unable to resolve the uncertainty in bioavailability. Therefore, 100% bioavailability was assumed
for the purpose of low-dose extrapolation from rats to humans.
The rate of absorption appeared to be more rapid in female rats than in males. Kim et al.
f2016bl reported a Tmax- of 1.4-1.5 hours (0.06 days) in female rats and 3 days in male rats and Kim
etal. f2018bl likewise reported 1.4 hours in females and 3.1 days in males. However, this difference
in timing may also be confounded by the much slower clearance in male versus female rats (see
below). Huang etal. f2019al obtained a Tmax- of 2-3 hours in female rats and 5-7 hours in male rats,
with a decreasing trend as dose increased. Transporter-mediated processes and protein binding
may have caused dose-dependence of Tmaxfor PFHxS, but the differences in Tmax between dose
groups was not reported as statistically significant by Huang etal. (2019a) and the range of values
for each sex was not large enough to be of consequence for dose extrapolation.
While these results indicated somewhat slower absorption in male rats than in female rats,
it is only by a factor of 2 or 3 fHuang etal.. 2019al. Sundstrom etal. f20121 observed a Tmax of only
0.5 hours in female SD rats and could not estimate a value for male rats because of the short 24-
hour window of observation. The cause for the discrepancy from other studies discussed just above
was unclear. Plotted data indicated very rapid initial absorption in both males and females (Kim et
al.. 2016b: Kim etal.. 2018b) and by definition peak concentration occurs when the rate of
clearance equals the rate of absorption (which decreases as the remaining dose in the
gastrointestinal tract declines). So, it may simply be that it took longer for the absorption rate to fall
below the slow clearance rate of PFHxS in male rats than female rats.
In male CD-I mice Sundstrom etal. (2012) the observed Tmax was 8 hours at a dose of
1 mg/kg and 4 hours at a dose of 20 mg/kg while Tmax was 2 days in females at 1 mg/kg but only
4 hours in female mice at 20 mg/kg. Thus, the predominant results indicated that the majority of
absorption occurs in less than 8 hours in mice, consistent with uptake being in the range of 90% or
higher. It was unclear why Tmax was lower at the higher doses in both males and females. No specific
methodological flaws were identified, but the exact value of Tmax from an experiment depends on
the timing of blood samples (experimental design) and can be affected by experimental variability.
Serum concentrations were measured starting at 2 hours and it is possible that the value of "2" for
female mice dosed with 1 mg/kg PFHxS was actually 2 hours, rather than 2 days. While
bioavailability was not measured in primates, it is reasonable to assume that uptake in monkeys
and humans is likewise fairly efficient
A study on the toxicological response upon dermal exposure to a technical mixture
containing PFHxS showed the presence of PFHxS in serum during the 28-day dosing period and
3-5
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
after a 14-day recovery period (3M. 20041. Male and female rats were exposed to the product as a
liquid on cotton gauze or as a solid dried onto cotton gauze. PFHxS from both the liquid and dried
product entered systemic circulation through the skin as determined by measurements of serum
PFHxS levels. Male rats showed higher PFHxS serum levels compared with female rats, which was
likely an effect of differential excretion, rather than differential absorption. Male rats showed a
clear accumulation of PFHxS in serum over the duration of the 2 8-day dosing period and levels
appeared to decrease during the recovery period in the group exposed to the dried formulation.
Male rats exposed to the liquid formulation had peak levels observed after the recovery period. In
female rats, peak concentrations were seen after 14 days of exposure and lower levels were seen
after 28 days of exposure. Levels were lower still after the recovery period. These data suggested a
concern for dermal exposure to PFHxS in both liquid and dried formulations, but further research is
needed to quantify rates of absorption, the resulting relationship between external and internal
dose, and the extrapolation of this information to human exposure.
No data on absorption of PFHxS through the respiratory tract has been found.
There is no direct quantification of oral absorption of PFHxS in humans. However, an
epidemiological study by Stubleski etal. f20161 identified a qualitative association between PFHxS
concentrations in human serum and concentrations in drinking water. Specifically, a 54% increase
in serum levels was observed during the observation period after a large contamination event, but
serum levels only declined 20% after an intervention that decreased drinking water levels by 60%.
The lack of exact correlation may have been due to the timing of sampling versus the contamination
event, as well as to the long half-life of PFHxS in humans.
Given the generally high absorption reported in rats (e.g., 90% for female rats and 96% for
male rats) by Kim etal. f2018bl. humans will be assumed to absorb 100% of ingested PFHxS, which
is slightly more health protective compared with assuming 90%-96%.
3.1.2. Distribution
While PFHxS was found at some level in all tissues evaluated, the largest amounts have been
in the liver, followed by the kidneys and lung, with much lower levels in other tissues. For example,
Benskin et al. (2009) reported tissue:blood ratios in male rats on day 3 of dosing at 0.03 mg/kg as
being 17% for liver, 10% for lungs, 5% for heart and kidney, with other tissues being 4% or lower.
Kim etal. (2016b) measured ratios after 72 days in male and 14 days in female rats from 4 mg/kg
doses and obtained ratios of 17% and 11% for male and female liver, respectively; 13% and 8% for
kidney; 5% and 4% for heart; 4% and 3% for lung (each for males and females, respectively); and
2% for spleen in both sexes. This distribution appears to be fairly rapid compared with the overall
time-course in blood: Huang etal. (2019a) showed essentially constant tissue:plasma ratios in
female rat liver and kidney from day 0 to day 8 and in the male rat kidney from 0 to 50 days after a
16 mg/kg dose. Interestingly, the ratio in the male rat liver quickly rose to 50%-60% but then
gradually increased to over 80% on day 50 fHuangetal.. 2019al. This time-dependence may have
been due to slower clearance from the male rat liver than the blood and other tissues which may
3-6
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
confound interpretation of PK data. If the percent distribution to the liver (relative to plasma)
increased over time, then the observed decline in plasma concentrations was not proportional to
whole-body elimination.
The order of tissue concentrations was observed to be the same in mice as in rats, but with
the mouse liver having 2 5%-40% of serum levels and the kidney ~10% fSundstrom etal.. 20121.
However, measurements of PFAS levels in human cadavers indicated a different ordering of
concentration, with the highest levels in kidney (median 18 ng/g), followed by lung (median
5.7 ng/g), then brain, liver, and bone (2.3,1.8, and 1.2 ng/g respectively) (Perez etal.. 20131. These
human results should be interpreted with some caution since they do not provide ratios from
matched samples and the specific method of collecting tissues likely differed to some extent (details
on the human tissue collection are not available). However, the difference between kidney and liver
may be large enough to suggest a difference between human and rodent PFHxS distribution for
these tissues.
Karrman et al. (20101 also examined postmortem liver concentrations in 12 human samples
and compared those to serum concentrations previously observed in the region. This comparison is
severely limited as the serum and liver samples were sourced from different individuals.
Yeung etal. f20131 evaluated PFHxS concentrations in liver versus serum of humans with
hepatocellular carcinoma (HCC) or cirrhosis due to chronic hepatitis C virus (HCV). In these
patients, the liver concentration was 15% of the serum in HCC patients (n = 11) and 9% of the
serum in HCV patients (n = 32). These results need to be interpreted with caution because of the
disease status, but they indicated somewhat lower distribution into the human liver than observed
in rodents. The authors did not have paired liver and serum from healthy individuals for
comparison. In addition to the evidence of distribution to the brain in cadavers, PFHxS has been
observed in the cerebrospinal fluid of neonates, with a median cerebrospinal fluid: blood serum
ratio of 0.0290 from two paired samples fLiu etal.. 2022bl. Given evidence from other PFAS in
humans and rats that the authors reviewed, this ratio is expected to be higher in neonates
compared with adults due to ongoing development of the blood-cerebrospinal fluid barrier.
Intracellular concentrations of PFHxS in the brain are expected to be much higher than the
concentration in the cerebrospinal fluid due to interactions between PFHxS and cytoplasmic
proteins.
A recent study evaluated levels of several PFAS, including PFHxS, in human serum as a
function of various measures of body composition as well as localized measurements of adipose
content throughout the body generated by dual-energy X-ray absorptiometry (DXA) and whole-
body magnetic resonance imaging (WB-MRI) (Lind etal.. 2022). There was not an association with
traditional measures of body composition, such as body mass index (BMI). PFHxS was however
inversely related to total lean mass, leg lean mass, subcutaneous adipose tissue in the arms, trunk
and thigh, and skeletal muscle volume in the arms and legs in men but not in women. Given the
minimal distribution of PFHxS to adipose and muscle tissues described above, one might expect
3-7
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
essentially no effect of the volume of these tissues on serum levels. However, one would predict a
negative correlation between Vd and body fat, the results in men may be consistent with that
prediction if glomerular filtration increases with body mass or surface area. It is also possible that
the correlation was due to variation in exposure related to body fat or muscle volume that occur
particularly in males. Matched estimates of exposure from dietary surveys or samples or matched
measures of urinary clearance (PFAS concentrations in urine) are ultimately needed to determine
whether or not the correlations actually reflect PK variation.
Kangetal. (2020) measured the levels of PFAS in the follicular fluid of women undergoing
oocyte retrieval for in vitro fertilization in relation to their serum levels and observed a median
ratio of 0.84, which is much higher than seen for other various tissues described above. This result
suggested that PFHxS can pass readily through the follicular walls (theca and granulosa cells), and
that binding to proteins in the follicular fluid is similar to that in serum.
Zhao etal. (2015) and Zhao etal. (2017) investigated the role of renal transporters known
to be involved in enterohepatic recirculation of bile acids. Zhao etal. (2015) showed that PFHxS is a
substrate for the human and ratNa+/taurocholate co-transporting polypeptide (NTCP) expressed in
vitro and Zhao etal. f20171 showed that multiple human and rat organic anion transporting
polypeptides (OATPs) likewise transported PFHxS. These active transport processes may
contribute to the relatively high distribution of PFHxS observed in the liver and its long half-life in
rats and humans by limiting biliary excretion. Excretion is also limited by protein binding in the
liver; for example, observed in interactions with human liver fatty acid-binding protein (hL-FABP)
(Yang etal.. 2020a: Shengetal.. 2016). and in serum, discussed subsequently in the Distribution in
Blood/Proteins section. The impact of serum protein binding on renal clearance is also discussed in
the Excretion section (see Section 3.1.4) under the Clearance Versus Glomerular Filtration Rate and
Free Fraction in Serum subsection.
Volume of Distribution
Vd is a pharmacokinetic parameter that quantifies the extent to which a chemical
distributes between the blood and the body as a whole and is effectively an average of tissue-
specific distribution ratios. Vd is key in evaluating internal dose because it quantifies the blood
concentration for a given total amount in the body. See Section 3.1.6, Empirical Pharmacokinetic
Analysis, for details of EPA's computational analysis. In rats, mean Vd ranged from 123 to
327 mL/kg among studies, doses, and routes of administration, without a clear sex difference
(Sundstrom etal.. 2012: Kim etal.. 2016b: Kim etal.. 2018b: Huang etal.. 2019a). Only Sundstrom
etal. (2012) evaluated the Vd in mice at two oral doses, and while the values were approximately
25% lower in females than in males at a given dose, the value for female mice given 20 mg/kg was
between the values for male mice given 1 versus 20 mg/kg. The overall range of Vd in mice (96-
195 mL/kg) strongly overlapped the observed range in rats. The Vd in monkeys was also evaluated
by Sundstrom et al. f20121. though only at a single IV dose (10 mg/kg) and was likewise in the
range reported for rats: 213 mL/kg in female monkeys and 287 mL/kg in male monkeys.
3-8
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
The fact that reported values ofVd were generally below 300 mL/kg and that most tissue-
specific levels were low compared with blood (see previous section) indicated that PFHxS primarily
distributes with extracellular fluid, with the exception of the liver.
Reported values of Vd are listed in Table 3-1, grouped by species and sex. No data to
determine Vd in humans were found.
The biochemical and physiological factors that determine tissue distribution have been
generally presumed to be evolutionarily conserved among mammalian species, an assumption
which was supported by the overall similarity of values across species seen in Table 3-1. However,
species differences in Vd can occur, especially given that the tissue fraction in the body varies
among species, and as shown by Kim etal. f2018bl the distribution to different tissues varies
several-fold. Since nonhuman primates were expected to be closer to humans in body composition
than rats or mice, the Vd values in human males and females was assumed equal to the values
estimated by Sundstrom et al. (2012) for male and female monkeys, respectively. There is
uncertainty in this assumption, which would be reduced by measurements of the PFHxS Vd in
humans.
A Bayesian PK analysis was conducted that combines data from across studies and doses
listed in Table 3-1 for male and female rats and mice (summary in Section 3.1.6, details provided in
Appendix E). This analysis provided both an overall mean and a credible interval for the Vd for each
of these species and sexes. The analysis for rats was restricted to oral dosimetry data because the
reported PK parameters indicated some discrepancy between the results for IV and oral dosimetry
that were unlikely to be resolved by the empirical modeling approach used here, and the bioassay
results that will be extrapolated using the PK parameters are from oral exposures. Because only IV
route data were available for monkeys, those data were used for that species.
3-9
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-1. Estimated volume of distribution (Vd) values in rats, mice, and
monkeys
Study
Vd (miykg)
Notes
Male rats
Sundstrom et al. (2012)
275 ± 5a
N/A
10 mg/kg IV, n = 4,10 wk time-course
Kim et al. (2016b)
269 ± 52b
N/A
4 mg/kg IV, n = 5, 72 d
278 ± 4b
279.0 (234.7-323.7)
4 mg/kg oral, n = 5, 72 d
Kim et al. (2018b)
315 ± 23b
N/A
10 mg/kg IV, n = 5,14 d
327 ±10b
313.7 (298.9-327.9)
10 mg/kg oral, n = 5,14 d
Huang et al. (2019a)
224 ± 32°
N/A
4 mg/kg IV, n = 3/time point, 50 d
123 ± lld
136.0 (115.0 -155.9^
4 mg/kg oral, n = 3/time point, 50 d
137 ± 9d
142.7 (119.6-164.2)
16 mg/kg oral, n = 3/time point, 50 d
192 ± 17d
206.2 (173.6-237.4)
32 mg/kg oral, n = 3/time point, 50 d
Population mean
208.2 (136.3-278.1)
Female rats
Sundstrom et al. (2012)
278 ± 66a
N/A
10 mg/kg IV, n = 3, 24 h
126 ± 14a
N/A
10 mg/kg IV, n = 4. 10 wk
Kim et al. (2016b)
289 ± 24b
N/A
4 mg/kg IV, n = 5,14 d
256 ±18b
299.2 (271.8-326.8)
4 mg/kg oral, n = 5,14 d
Kim et al. (2018b)
176 ± llb
N/A
0.5 mg/kg IV, n = 5,14 d
191 ± 7.5b
N/A
1 mg/kg IV, n = 5,14 d
130 ± 5.5b
N/A
4 mg/kg IV, n = 5,14 d
3-10
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Vd (miykg)
Notes
154 ± 20b
N/A
10 mg/kg IV, n = 5,14 d
187 ± 3.5b
198.3 (182.9-214.5)
1 mg/kg oral, n = 5,14 d
159 ± 7.8b
240.6(221-259.1)
4 mg/kg oral, n = 5,14 d
Huang et al. (2019a)
144 ± 18°
N/A
4 mg/kg IV, n = 3/time point, 22 d
155 ± 9d
170.7 (149.8-190.7)
4 mg/kg oral, n = 3/time point, 22 d
186 ± 14d
190.8 (169.9-212.7)
16 mg/kg oral, n = 3/time point, 22 d
264 ± 20d
254.8 (224 - 284.3)
32 mg/kg/ oral, n = 3/time point, 22 d
Population mean
222.6 (177.8-263.9)
Male mice
Sundstrom et al. (2012)
129b
1 mg/kg oral, n = 4/time point, 23 wk
195b
20 mg/kg oral, n = 4/time point, 23 wk
Population mean
150.6 (136.1-164.7)
Female mice
Sundstrom et al. (2012)
96b
1 mg/kg oral, n = 4/time point, 23 wk
147b
20 mg/kg oral, n = 4/time point, 23 wk
Population mean
120.7(112.8-128.9)
Male monkeys
Sundstrom et al. (2012)
287 ± 52b
272.0(239.3-303.2)
10 mg/kg IV, n = 3,171 d
Female monkeys
Sundstrom et al. (2012)
213 ± 28b
222.9 (197.9-249)
10 mg/kg IV, n = 3,171 d
Values in italics are the mean (90% credible interval) from the Bayesian analysis described in Appendix E (oral
exposure data).
aVdSS from two-compartment PK model.
bVd from noncompartmental PK analysis.
dVd from one-compartment PK model.
3-11
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
While Vd in rodents for several PFAS have generally been found to be less than 1,000 mL/kg
(1 L/kg), reported values do vary considerably. For example, Huang etal. f2019al reported
respective male and female rat values for total Vd of:
• 170-340 and 170-420 mL/kg for PFBS;
• 300-680 and 220-420 mL/kg for PFOS given doses of 2 mg/kg; but
• 79 and 56 mL/kg for PFOS given a dose of 20 mg/kg
(Dzierlenga et al.. 2019) reported respective male and female rat values for total Vd of:
• 300-620 and 223-560 mL/kg for PFHxA;
• 150-200 and 79-340 mL/kg for PFOA; and
• 410-630 and 270-410 mL/kg for PFDA.
In part, these ranges, and differences in reported Vd values between laboratories reflected
both experimental variability and differences in the pharmacokinetic analyses used, which may
have been more or less sensitive to variability in the data. Experimental design, such as the
timepoints selected for measurement and duration of a PK study, also impact Vd estimates.
However, some of the variability demonstrated here between different PFAS almost certainly
represents true differences in their chemical properties. A comprehensive review of such factors is
beyond the scope of this assessment, but these data indicated that the reported Vd values for PFHxS
were well within the overall range observed for several other PFAS.
The only study to evaluate Vd in humans directly from human data for PFHxS (versus using
a value obtained for other PFAS or in other species) was that of Chiu etal. f20221. who applied a
one-compartment PK model in a Bayesian analysis of human serum concentrations matched with
drinking water (DW) concentrations of several PFAS, including PFHxS, from multiple community
studies. The analysis only included adults who were determined unlikely to have occupational
exposure (i.e., for whom DW was likely to be the primary exposure) with corresponding DW
concentrations measured prior to measurement of their serum concentration. The overall approach
and parameter estimation method were considered sound. The value of Vd obtained for PFHxS
(95% CI) was 0.25 (0.15, 0.42) L/kg, which is almost identical to the average of the Vd values
estimated for male and female monkeys (see Table 3-1).
Distribution in Blood/Proteins
The low estimated volume of distribution of PFHxS reflects the relatively high amount of the
chemical found in plasma. A major factor in this distribution was attributed to the interaction
between PFHxS and proteins in plasma, including albumin and transthyretin fWeiss etal.. 2009:
Forsthuber etal.. 2020: Bischel etal.. 2011: Alesio etal.. 2022). An investigation of protein binding
3-12
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
showed that in human plasma PFHxS was 99.98% bound to protein with no sex-specific difference
fKim etal.. 2018bl. The same study reported 99.92% binding to protein in male rat plasma and
99.93% binding to protein in female rat plasma fKim etal.. 2018bl. More recently, Smeltz et al.
£2023} measured 99.91 ± 0.03% bound in a mixed pool of human plasma. Binding to plasma
proteins may also drive the partitioning of PFHxS within blood components for which greater levels
of PFHxS were measured in serum and plasma compared with whole blood. Poothong etal. (2017)
found median ratios of 1.06 between serum and plasma, 1.88 between serum and whole blood, and
1.75 between plasma and whole blood in adult men and women. Hanssenetal. (20131 found a
median ratio of 1.58 between plasma and whole blood in women just after the delivery of a child.
Tin etal. f20161 determined a mass fraction in plasma of 0.87 in adult men and women. Liu et al.
f20231 obtained a similar mean fraction in plasma of 0.84 specifically for n-PFHxS, but higher
fractions of 0.9 and 0.93 for two branched isomers.
Fetal Blood and Placenta
Studies of the associations between maternal serum levels and umbilical cord blood levels
of PFHxS demonstrated transfer through the placenta fZhang et al.. 2 013a: Monrov et al.. 2008: Li et
al.. 2020a: Lee etal.. 2013: Kang etal.. 2021: Hanssen etal.. 2013: Fromme etal.. 2010: Chen etal..
20171. Lee etal. f20131. Chen etal. f 20171. Kang etal. f20211. Li etal. f2020al and Zhang et al.
f2013al showed greater concentrations of PFHxS in maternal serum relative to cord serum, a
phenomenon that also has been observed for other PFAS such as PFOA and PFOS (e.g., Li et al.
(2020all. Lee etal. (20131 analyzed pairwise data to determine a cord serum:maternal serum ratio
of 0.57 ± 0.29 (mean ± SD). Chen etal. (20171 similarly found a geometric mean cord
serum: maternal serum ratio of 0.54. Kang etal. f20211 calculated an arithmetic mean cord
serum:maternal serum ratio of 0.365. Hanssen et al. (20131 observed a median cord:maternal ratio
of 0.53 in plasma and a median cord:maternal ratio of 0.43 in whole blood from pairwise data.
Zhang etal. (2013a) also examined the ratio in whole blood and found a cord:maternal blood ratio
of 0.294. Li etal. (2020a) compared cord:maternal serum ratios from preterm versus full-term
deliveries and reported a median ratio of 0.40 for preterm versus 0.72 for full-term, with the
difference being statistically significant. The authors suggest that this increase in distribution may
be due to placental aging resulting in a reduced capacity to limit transfer of xenobiotics, though
they also consider simple accumulation with time as a mechanism (Li etal.. 2020a). Li etal. (2020a)
also evaluated the role of nine placental transporters, testing for correlation between their
expression and the cord:maternal serum ratio. However, the only significant correlation was with
folate receptor alpha (FRa) in preterm deliveries (i.e., not full term), with a positive correlation
coefficient, indicating that FRa facilitates transfer to the fetus.
In contrast, Monrov etal. (2008) observed cord serum concentrations that were
significantly higher than maternal serum concentrations based on a paired t-test and linear
regression analysis. However, these data were highly censored, with the prevalence of samples
above the level of detection in umbilical cord serum (20%) lower than in maternal serum (45.5%).
3-13
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
The observed relationship between maternal serum and umbilical cord serum could be an artifact
due to the higher prevalence of umbilical cord samples below the level of detection.
To quantitatively compare the distribution between tissues and maternal blood matrices
among different studies, adjustments were made to correct for the distribution among blood
components. As described above, Poothong etal. f20171 measured a median ratio of 1.88 for
serum:whole blood, 1.75 for plasma: whole blood, and 1.06 for serum:plasma concentrations of
PFHxS. These values were used to adjust subsequent tissue:blood matrix ratios to tissue:serum,
when reported for whole blood or plasma.
Serum and plasma are components of whole blood, with the main other component (by
volume) being red blood cells. Assuming that PFHxS partitions completely into the plasma and not
the red blood cells, a theoretical maximum ratio between the plasma and whole blood was
calculated, that is, as if whole blood is a dilution of plasma with red blood cells. The small additional
volume contribution from other components of whole blood is not present in plasma or serum were
assumed to not substantially affect this theoretical ratio. The most common metric for the
composition of whole blood is the hematocrit (Hct), which is the ratio of the volumes of red blood
cells and whole blood. In terms of Hct, the theoretical maximum ratio of plasma:whole blood was
calculated as 1/(1-Hct). The normal range of hematocrit for men is 42%-52% and for women is
37%-48% flordan etal.. 19921. Inputting a typical human male Hct of 45% gave a plasma: whole
blood ratio of 1.82. In females, Hct is typically lower, which resulted in a lower estimated maximum
plasma: whole blood ratio. Using the reported plasma: whole blood ratio of 1.7 and a Hct of 45% the
fraction of PFHxS in plasma (Fp) was calculated to be 1.7 x (1-Hct) = 93.5%, which is very high but
consistent with the high level of plasma protein binding described above. The median ratio of 1.88
serum:whole blood reported by Poothong et al. f20171 is greater than the theoretical maximum and
implies a Hct of >46.8%, which is in the normal range for men, but slightly higher than the normal
range for women. The population of Poothong etal. f20171 was approximately 75% women, which
may indicate a deviation from the ideal behavior assumed for the calculation, variation in Hct, or an
experimental error in the measurement of concentrations or in the separation. Partitioning of
PFHxS and other PFAAs between human plasma and blood cells was also investigated by Tin etal.
(20161. who obtained a mean Fp = 91% and report a mean serum:whole blood ratio of 1.6. The
average of serum:blood ratio of 1.6 from Tin etal. f20161 and 1.88 from Poothong etal. f20171 is
1.7. Given Hct = 0.45, this value implies 95.7% of PFHxS is in serum, which is still reasonable.
Therefore, a serum:blood ratio of 1.7 was used to convert tissue partitioning data relative to whole-
blood concentrations to serum-based concentrations below.
The empirical data ofHanssen etal. (2013). although limited by a modest number of
subjects with data over the limit of detection, indicated generally higher serum:whole blood ratios
in cord serum and blood than maternal serum and blood, with ratios for multiple samples
(subjects) reported as 2.2 or higher. This difference can be explained in part by a higher hematocrit
in later gestation and newborns than in adults (mean hematocrit ~51% for gestation week 42 and
3-14
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
full-term newborns) (Toplingetal.. 2009). One study included in Table 3-2 below (Zhang etal..
2013a) reported concentrations of PFHxS for whole maternal and cord blood, rather than serum
levels. Therefore, the resulting ratios for matched samples (obtained from the supplemental data of
Zhang etal. f2013all were adjusted by the ratio 0.55;0.49, that is, (1-Hctaduit)/(1-Hctfetus) to account
for the expectation that serum:whole blood concentrations will be higher in the fetal cord blood
than in the adult.
With the adjustment noted above, median (mean) values of cord serum: maternal serum
ratios in humans at childbirth were 0.53 on average (see Table 3-2). That the value is roughly 50%
indicated that the placenta may limit transfer of PFHxS from the mother to the fetus, but if
distribution to fetal tissues is increased in proportion to water content of tissue, as discussed
below, then an overall higher concentration in the fetus versus maternal tissue is predicted. There
was not an apparent trend in the ratio related to the maternal sample timing relative to childbirth
(i.e., whether taken before, at, or after childbirth) or the fraction of cord or maternal serum
measurement below the limit of detection, although as described above Li etal. (2020a) reported a
significant increase in the ratio from preterm to full-term deliveries. Examination of the standard
deviation of the mean of medians and mean of means shows that the two values are, on average,
similar, suggesting that the distribution of cord serum:maternal serum ratio is symmetric. However,
it is notable that the reported median value is lower than the mean value in almost every study.
Table 3-2. Measured cord serurmmaternal serum ratios
Cord serum:maternal
serum ratio
% > LOD
Study
Median
Mean
Cord
Maternal
Maternal sample timing
Chen et al. (2017)
0.55
0.6
97%
97%
Within 3 d prior to delivery
Hanssen et al. (2013)
0.54
0.63
100%
100%
3-5 d after delivery
Kane et al. (2021)
0.315
0.365
97%
100%
At delivery, exact timing not
clear
Kim et al. (2011b)
0.65
0.64
100%
100%
20th-41st wk of pregnancy,
mostly in 3rd trimester
Lee et al. (2013)
0.5
0.57
100%
100%
At delivery, exact timing not
clear
Liu etal. (2011)
0.73
0.95
96%
98%
Within 1 wk after delivery
Yang et al. (2016b)
0.35
0.43
100%
100%
1-2 d before delivery
Yang et al. (2016c)a
0.52
0.63
96%
100%
Within 1 wk after delivery
Zhang et al. (2013a)
0.332
0.387
100%
100%
Within 1 hr prior to delivery
Li et al. (2020a) preterm
0.40
NR
81%
81%
Within 1 wk before delivery
3-15
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Cord serum:maternal
serum ratio
% > LOD
Study
Median
Mean
Cord
Maternal
Maternal sample timing
Li et al. (2020a) full-term
0.72
NR
94%
94%
Within 1 wk before delivery
Overall meanb
0.50±0.14
0.58±0.17
NR = not reported.
aCord:maternal serum ratios for this study are the ratio of the reported median (mean) values for cord and
maternal serum.
bMean and standard deviation of the set of medians or means.
After correction for the serum:whole blood ratio as described above, comparisons between
maternal serum and placenta were reasonably consistent: Chen etal. f20171 observed median
(mean) placenta:maternal serum = 0.421 (0.429) and applying the serum:whole blood factor of 1.7
to the results of Zhang etal. (2013a) the EPA obtained median (mean) = 0.266 (0.289). Chen etal.
(2017) suggested that the difference between their results and those of Zhang etal. (2013a) was
due to variation in isomeric composition between the two study populations or the greater range in
concentration in the placentas in the study of Zhang etal. (2013a). but with the correction applied,
it appears to be modest. The volume of distribution estimated for PFHxS in female monkeys was
Vd = 0.213 L/kg (Sundstrom etal.. 2012). which represents the average of distribution into all
tissues. While the placenta distribution measurements in humans of Chen et al. (2017) and Zhang et
al. (2013a) were 1.5 to 2 times higher than this value for female monkeys, Kim etal. (2018b)
showed greater variability in PFHxS concentrations between specific tissues of rats. Hence, the
reported placenta: serum levels of Chen etal. (20171 and Zhang etal. f2013al were not outside the
range one would expect for a specific tissue given an overall Vd of 0.213 L/kg, i.e., if distribution to
adipose and muscle was substantially less than internal organs, as was observed for rats by Kim et
al. (2018bl.
As umbilical cord blood followed the same trend as in adult blood, the results from Chen et
al. (2017) and Zhang etal. (2013a) were consistent with a concentration trend of cord
serum > placenta > cord whole blood.
One study that distinguished between isomers of PFHxS found the greatest prevalence of
the linear relative to the branched isomer in cord serum (97% linear), followed by maternal serum
(86% linear) and placenta (77% linear) fChen etal.. 20171.
Distribution in Fetal Tissues and Children
One study provides a relatively unique dataset of PFHxS concentrations in human fetal
tissues obtained from voluntary abortion (gestation week < 12) or after intrauterine fetal death in
the second and third trimester, and in maternal serum collected at these times (Mamsen et al..
20191. However, PFHxS was detected in only 6% of fetal tissues, making it difficult to interpret
these data quantitatively.
3-16
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Pharmacokinetic modeling of PFOA dosimetry in humans by Goeden etal. (2019) suggested
a reason why observed tissue levels of PFAS in the fetus and young children may have been greater
than in adults: the greater amount of extracellular water in the tissues of fetuses and children
fFriis-Hansen. 19611 led to a greater distribution of PFAS into these tissues. As noted above, the Vd
values estimated for adult rats, mice, and monkeys are consistent with the assumption of
distribution in body water. The amount of extracellular water in newborns was estimated to be 2.4
times higher than adults (Friis-Hansen. 19611 (see Figure 3-2).
2.6
2.4 4
2.2 1
2
.9 1.8
cc 1.6
1.4
1.2
1
0.8
C
Ratio of ExtracellularTissue Water in Children vs. Adults
) 1
l ;
t 3 '
I 5 f
Age (y)
7 8 9 10
Figure 3-2. Ratio of extracellular water (% of body weight) in children versus
adults. Values (points) were calculated from results in Friis-Hansen (1961) and
plotted at the midpoint for the corresponding age ranges evaluated.
Mamsenetal. (2019) (described briefly above) only detected PFHxS in 6% of fetal tissue
samples and did not report ratios of fetal tissue to maternal serum for PFHxS. So, while their data
may indicate that average fetal levels are much lower than maternal levels, they cannot be used to
quantify the fetal-maternal relationship. Since PFHxS is amphiphilic, with Vd < 1 in adults, it is not
expected to distribute with or in proportion to body fat, and therefore fetal body fat content is not
considered an appropriate predictor of fetal PFHxS distribution. Given the overall lack of data on
fetal distribution of PFHxS, EPA considers any estimate of such distribution to be uncertain. In the
face of this uncertainty, EPA chose the simplest assumption for prediction of fetal body burdens:
that distribution between fetal serum and fetal tissues is the same as the distribution between
serum and tissues in the newborn. The alternative, which would be to assume that there is a
discontinuity (sudden increase or decrease) in the body burden of the offspring at the moment of
birth, would require a more specific assumption about the magnitude and direction of that
discontinuity. Likewise, assuming any other change in Vd over the time of fetal development and
birth would also have no supporting data and therefore involve equal or greater uncertainty. There
3-17
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
are no clear developmental PK data for PFHxS that could be used to guide a choice among these
alternatives. Hence, EPA simply assumed that the ratio of body water in the newborn versus adults
(2.4) also applies to the fetus.
Since the Vd in a human woman (mother) is assumed to be the same as in monkeys, given
the assumption that Vd in a fetus is 2.4 times higher than in an adult, the estimated Vd in a female
fetus relative to fetal serum is 2.4 x 0.213 L/kg = 0.511 L/kg and in a male fetus
2.4 x 0.287 L/kg = 0.689 L/kg. However, as described above, the average ratio of PFHxS in cord
serum, which is assumed to be fetal serum, compared with maternal serum was rf:m = 0.52.
Together, these values and assumptions led to the prediction that relative to maternal serum, the
Vd for the fetus as a whole is 0.52 x 0.511 L/kg = 0.266 L/kg for females and likewise 0.358 L/kg for
males, indicating average fetal tissue concentrations is 25% higher than average maternal tissues
for girls and 68% higher for boys. Hence, the body burden in the newborn can be estimated using
the following equation:
amount of PFHxS in newborn = rf:m x Cmother x Vdnewborn x BWnewborn, (3-1)
where rf:m = 0.52 and Vdnewborn is 0.511 L/kg for girls and 0.689 L/kg for boys.
The average weight of a newborn is only 5% of maternal body weight (3.4 versus 68 kg), so
while distribution into the male fetus was estimated to be 68% higher than maternal tissues, the
effect on Vd of the mother and fetus together (i.e., total amount in the mother and fetus compared
with maternal serum concentration) was thereby estimated to be less than 3.4% (5% x 68%).
Therefore, the Vd for mother and fetus together during pregnancy was simply assumed equal to the
value for the adult woman (0.213 L/kg), although the amount in the newborn child was calculated
as described above. Because the maternal weight just after childbirth is reduced by more than the
weight of the newborn, reflecting the loss of amniotic fluid, placenta, etc., this choice effectively
assumed slightly less PFHxS mass is lost with those fluids than would be calculated if total maternal
and fetal Vd were increased. The interpolation function shown in Figure 3-2 can be multiplied by
the adult Vd (L/kg) to obtain the corresponding value for children under 10 years of age, as was
done by Goeden et al. (2019). However, an opposing factor is the approximately 20% larger blood
volume as a fraction of BW in young children compared with older children and adults (Darrow et
al.. 1928). given that a high fraction of PFHxS is bound to blood proteins. More specifically, the mass
of PFHxS bound to blood proteins would increase in proportion to the total mass of those proteins,
which one might expect to increase in proportion to blood volume. Hence, a 20% larger blood
volume could be expected to reduce the PFHxS available for distribution to tissues by 20%. So,
instead of an increase of 2.4-fold in Vd in newborns one might predict an increase of 1.9-fold (i.e.,
80% x 2.4).
3-18
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Trend in Pregnancy
Four studies investigated how PFHxS levels tend to change during pregnancy and nursing.
Monrovetal. f20081 found that mean maternal serum PFHxS concentration did not change
between sampling at 24-28 weeks and sampling at delivery. Likewise, Oh etal. f20221 observed
only a slight average decrease in maternal PFHxS over the course of pregnancy not statistically
significant. (Varsi etal.. 2022) observed PFHxS serum concentrations in pregnant women at 18, 28,
and 3 6 weeks. Total PFHxS concentrations were relatively constant during this time, but there were
differences observed between PFHxS isomers. Linear PFHxS decreased during pregnancy and was
lower than concentrations observed in women who had never been pregnant at all timepoints.
Branched PFHxS, however, was highest at the 3 6-week timepoint, compared with concentrations at
18 and 28 weeks and compared with the nonpregnant women. Glynn etal. f20121 presented data
for other PFAS on the relative serum concentrations during pregnancy and nursing but did not
present that information for PFHxS, although PFHxS was included in other analyses in that study.
Breast Milk
PFHxS has been observed in human breastmilk, indicating that nursing acts as a route of
excretion for the mother and a route of exposure for her infant fKim etal.. 2011b: Karrman et al..
2007: Karrman etal.. 20101. Blomberg etal. f20231 evaluated longitudinal changes in breast milk
concentrations of PFHxS between delivery and up to 8 months postpartum; while milk
concentrations declined among the women with the highest levels at 0-2 months postpartum (i.e.,
over 500 pg/mL), they were more constant among those with early concentrations of 300 pg/mL or
lower. This decrease can be viewed as supporting this hypothesis, but some caution is needed in
interpreting these data as the drinking water source for the most highly exposed part of the cohort
was switched to a less contaminated source as soon as the contamination was identified, i.e.,
decreased exposure through drinking water could also drive decreased breast milk concentrations,
independent of excretion through breast milk. However, Oh etal. (2022) observed a significant
decline in maternal serum levels (average decline of 5.6%) during the first 6 months postpartum in
a population with typical PFHxS exposure (with no intervention to reduce exposure). This
observation provides some additional potential evidence of increased excretion of PFHxS after
giving birth, without an artificial change in PFHxS exposure.
In paired milk and maternal serum samples, the concentrations were highly correlated
(Pearson r2 = 0.8) (Karrman et al.. 2007). The concentration of PFHxS in breastmilk was reported to
be lower than the concentration in paired maternal serum, with ratios between milk and maternal
serum of 0.02 (Karrman et al.. 2007) and 0.008 (Kim etal.. 2011b). Karrman et al. (2010) reported
PFHxS concentrations in breast milk samples but did not have paired maternal blood levels, which
limits the ability to specify the distribution into breast milk compared with other body
compartments. Another study found that PFHxS was below the limit of detection in all breast milk
samples collected fLiu etal.. 20111. Mondaletal. f20141 investigated the association between
3-19
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
PFHxS concentration in maternal and infant serum and the length of breastfeeding and found that,
although there were associations consistent with breastfeeding acting as a route of excretion for the
mother and a route of exposure for the infant, none of the associations rose to the level of
significance. Significant associations were found for other PFAS studied and negative associations
for maternal serum and length of breastfeeding and positive associations for infant serum and
length of breastfeeding were consistent across PFAS. Varsi etal. (2022) observed paired maternal
and infant serum concentrations, with one infant timepoint at 6 months of age, and six maternal
timepoints, three during pregnancy and four postpartum. At 6 months after delivery, the relative
concentrations of PFHxS in the infant and mother differed by isomer, with the infants having a
higher median linear PFHxS concentration and a lower median branched PFHxS compared with the
mothers. Similarly, the branched: linear isomeric ratio was lower in the infant compared with the
mother. This could indicate a preferential transfer of the linear isomer to the infant, either during
gestation or lactation. Potential evidence for gestational transfer is the increase in maternal
branched: linear isomeric ratio that the authors observed between the 28th and 36th week of
pregnancy. Evidence for lactational transfer is the association the authors observed between infant
linear PFHxS concentration and months of exclusive breastfeeding a relationship that was not
present for the branched isomer.
3.1.3. Metabolism
Because of the high stability of the perfluoroalkyl bonds, PFHxS is thought to not be
metabolized in mammals, as was seen for similar PFAS (Lau etal.. 20071. Studies have examined
similar PFAS, including perfluorooctanoic acid (PFOA) and perfluorodecanoic acid (PFDA) and
identified only the parent compound in excreta fVanden Heuvel et al.. 1991a. b). The sulfonate
analog of PFOA, perflurosulfonic acid (PFOS), is also not metabolized (Lau etal.. 20071.
3.1.4. Excretion
Animals
Several studies examined the excretion of PFHxS from animals, particularly rats, after a
controlled exposure (Sundstrom etal.. 2012: Kim etal.. 2016b: Kim etal.. 2018b: Huang etal..
2019a: Benskin et al.. 20091. Excretion has been observed in urine and feces, with renal excretion
being the most prominent route. Other studies have only indirect observation of excretion through
the decreasing amounts of PFHxS in the serum over time. As PFHxS is not metabolized, decreases in
serum concentration after the distribution phase were attributed to excretion, assuming a constant
serum:tissue ratio. As noted above, the distribution phase may not be complete after a relatively
short time given the shifts in liver:serum ratio observed over 50 days (Huang etal.. 2019a). To
quantify the impact of such a shift on estimated excretion would require a PBPK model for PFHxS
that accounts for the time-dependence in specific tissue volumes and distribution, which is not in
the realm of available science. Since the extended time-dependent distribution appears to be
3-20
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
confined to the liver, the analysis based on empirical evaluation of excretion was still assumed to
provide a sufficient approximation for dosimetric extrapolation.
In animal studies, urinary excretion was greater than fecal excretion. There was a strong
sex-dependence in rats and mice in renal excretion with female rats excreting more of the total
dose in urine. Specifically, Kim etal. f2018bl reported 15.9% of the initial IV dose was excreted in
urine and 1.3% of the dose was excreted in feces in male rats and 39.1% of the dose was excreted in
urine and 3.1% of the dose was excreted in feces in female rats after 14 days. Similarly, after an oral
dose, in male rats 18.5% of the dose was excreted in urine and 2.8% in feces, while in female rats
36.8% of the dose was excreted in urine and 3.3% in feces. In another study Kim etal. f2016bl.
reported that female rats excreted 28.02% of an IV dose in urine after 14 days while male rats
excreted 8.26% of the dose in urine after 72 days. Sundstrom et al. f20121 reported that 24 hours
after an IV dose, female rats excreted 13.28% of the dose, while male rats excreted 0.70% of the
dose.
In mice, the total dose excreted in 24 hours was dose dependent, with 0.882% of a 1 mg/kg
dose and 1.654% of a 20 mg/kg dose excreted in males and 0.317% of a 1 mg/kg dose and 2.552%
of a 20 mg/kg dose excreted in females fSundstrom etal.. 20121. The lower excretion in female
versus male mice for the 1 mg/kg dose was the only situation with a greater male rodent excretion
fSundstrom etal.. 20121. Urinary excretion was slower in monkeys, with 0.102% of an IV dose
excreted in urine in 24 hours in male monkeys and 0.055% of the dose in female monkeys
(Sundstrom etal.. 2012). Unlike rodents, there was no clear difference between monkey sexes in
the amount of urinary excretion.
In addition to observations in excreta, multiple studies also estimated the rate of decrease
in serum or plasma levels of PFHxS in the form of a half-life or clearance (CL) in rats fSundstrom et
al.. 2012: Kim etal.. 2016b: Kim etal.. 2018b: Huang etal.. 2019a: Benskin et al.. 20091. While all of
these studies appear to have been conducted with appropriate quality, there is significant variation
in the results. For example, Kim etal. (2018b) estimated a CL of 228 mL/kg-day in female rats after
an intravenous (IV) dose of 4 mg/kg, while Huang etal. (2019a) estimated a CL of 46 mL/kg-day in
female rats after an oral dose of 4 mg/kg. Despite the significant variability in the results between
studies, routes of exposure, and to an extent, doses of PFHxS, a quite consistent result is that the CL
in male rats is about an order of magnitude lower than in female rats, and so the subsequent
analysis evaluates parameters for male and female rats separately.
An issue found in the PK data is that for some studies that used both IV and oral doses, the
blood AUC was higher after the oral dose than after the same dose given IV, which contradicts
classical PK analysis. For example, given doses of 4 mg/kg Kim etal. (2016b) reported an AUC
almost twice as great after oral dosing than after IV dosing in female rats, and Huang etal. (2019a)
reported an AUC 40% higher after oral dosing than after IV. By classical PK analysis, one expects
that only a fraction of an oral dose will be absorbed but that the subsequent distribution and
elimination are otherwise identical to what is observed after IV dosing. In that case, the AUC after
3-21
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
oral dosing would be less than or equal to the AUC after IV dosing, to the extent that there is limited
oral bioavailability. A key assumption in this classical analysis is that distribution and elimination
are independent of the exposure route, and EPA interpreted these discordant empirical results as
suggestive that this assumption is incorrect. EPA's analysis of PK data supported this possibility,
with a trend of greater clearance following IV exposure compared with gavage in female rats (see
3.1.6 Empirical Pharmacokinetic Analysis). The mechanistic explanation for this difference is not
obvious. Excretion could be greater after IV dosing if, immediately after dosing, a smaller
proportion of PFHxS is bound to tissue phospholipids and serum proteins compared with the oral
dosing scenario. This could occur if equilibration between bound and free PFHxS takes some time.
Absorption from the GI tract is slower and PFHxS first passes through the liver (where a significant
fraction is retained) before systemic distribution, which would allow for equilibration between free
and bound states as PFHxS enters the blood. Thus, a higher fraction of PFHxS could have been
bound when first reaching general circulation after oral dosing than after IV dosing, such that the
urinary excretion after oral dosing was slower. A similar mechanistic explanation for differences in
protein binding is that passage through the acidic environment of the stomach results in a greater
proportion of the PFHxS anion, which could facilitate binding and thus limit excretion compared
with IV exposure.
Because the toxicological bioassays that will be interpreted with the PK model used oral
administration, it was considered clearly preferable that the PK parameters used should reflect that
route of exposure. Given the oral-IV discrepancies noted above, only results from oral PK
experiments were evaluated for rats and mice. Key PK parameters from these oral PK experiments
are listed in Table 3-3.
Of note in Table 3-3, and as discussed previously, is thatthe data of Huang etal. f2019al
indicate higher CL in male and female rats given a dose of 32 mg/kg compared with 4 and
16 mg/kg. While the difference was not indicated as statistically significant, it was consistent with a
mechanism of saturable renal resorption (Yang etal.. 2009: Weaver etal.. 2010) and with the end-
of-study serum concentration data shown in Figure 3-1 (NTP. 2019). Comparing results for the
lower two doses, the CL estimated by Huang etal. (2019a) for 16 mg/kg in female rats was 25%
higher than that estimated at 4 mg/kg and the CL for 16 mg/kg in male rats was 19% higher than
that estimated at 4 mg/kg. Although not statistically significant, this was also interpreted as
consistent with some dose dependence. On the other hand, the CL reported for female rats at
32 mg/kg by Huang etal. f2019al was below that reported by Kim etal. f2016bl at 4 mg/kg and the
CL for male rats at 32 mg/kg by Huang etal. (2019a) was below that estimated from the results of
Benskin et al. (2009) presumably due to interstudy variability. Hence, subsequent PK analyses
included data for all dose levels from Huang etal. (2019a).
Overall mean CL values and confidence intervals for male and female rats, mice, and
monkeys were obtained by Bayesian PK analysis of all the oral PK data for each sex of rodents and
3-22
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
the IV PK data for each sex of monkeys (summary in Section 3.1.6, analysis details provided in
Appendix E).
Table 3-3. Summary of estimated clearance values in animals
Citation
Dose (mg/kg)
CLa(miykg-d)
Half-life (d)
n
Male rats
Benskin et al. (2009)
0.03
9.85b
15.9
7
Kim et al. (2016b)
4
7.15
6.71 (6.01-7.42)
26.9
28.9 (23.9-33.8)
5
Kim et al. (2018b)
10
6.65
4.78(3.14-6.42)
34.1
48 (29.3-64.7)
5
Huang et al. (2019a)
4
4.82
4.08(3.43-4.81)
17.6
23.4 (18.6-28.4)
3C
16
5.74
4.56(3.86-5.31)
16.5
21.9 (17.2-26.1)
3C
32
9.02
4.56(3.86-5.31)
14.8
19.6 (15.8-23.3)
3C
Population mean
5.46 (3.87-6.97)
27.3 (15.3 - 39.2)
Female rats
Kim et al. (2016b)
4
124.8
125.8 (116.4-135.6)
1.72
1.64 (1.57-1.72)
5
Kim et al. (2018b)
1
81.1
85.3 (80.2-90.6)
1.60
1.62 (1.54-1.68)
5
4
65.3
112.5 (105.6-119.6)
1.69
1.48 (1.41-1.54)
5
Huang et al. (2019a)
4
46.1
50.13 (45.4-54.9)
2.33
2.36 (2.2-2.5)
3C
16
59.0
61.8 (56.2-67.6)
2.19
2.14 (2.01-2.25)
3C
32
92.2
95.9 (87-105.3)
1.98
1.84 (1.71-1.97)
3C
Population mean
85.3 (59.7-107.9)
1.86 (1.22-2.5)
Male mice
Sundstrom et al. (2012)
1
2.94
30.5
4
20
4.83
28.0
4
Population mean
3.81 (3.54-4.08)
27.4 (25.8-29)
Female mice
Sundstrom et al. (2012)
1
2.68
24.8
4
3-23
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Citation
Dose (mg/kg)
CLa(ml7kg-d)
Half-life (d)
n
20
3.79
26.8
4
Population mean
3.14 (2.98-3.29)
26.7 (25.6-27.7)
Male monkeys
Sundstrom et al. (2012)
10
1.33
1.17 (0.72-1.64)
141
177.6 (95.5-250.6)
3
Female monkeys
Sundstrom et al. (2012)
10
1.93
2.1 (1.78-2.42)
87
74.4 (59.1-88.9)
3
Only oral exposure results are shown for rats because there were discrepancies between oral and IV data that
could not be resolved and the oral route was used in the bioassays evaluated for toxicity. Only oral dosimetry
data were available for mice and only IV dosimetry data were available for monkeys (results shown; (Sundstrom
etal., 2012)).
aValues in italics are mean (90% credible interval) from Bayesian analysis (details in Appendix E).
Calculated from reported half-life (T0.5) for n-PFHxS as CL = ln(2)*Vd/T0.5 using the geometric mean of Vd values
for male rats listed in Table 3-1. Serum time-course data were not available from Benskin et al. (2009), so results
from this study were not used in the Bayesian analysis.
cNumber of rats per time point, but each rat had blood taken at no more than two time points, so the total number
of rats used per dose level were much higher (Huang et al., 2019a).
While the results summarized in Table 3-3 were obtained by empirical analysis for total
clearance, it is worth noting the fraction of PFHxS eliminated in feces reported by Kim etal. f2018bl
was used as a means of estimating fecal clearance in humans. These data were used to estimate
total clearance for studies where renal clearance was measured and were deemed most
appropriate as primate and human-specific data were unavailable. The ratio of average PFHxS
excretion in feces versus urine was 8.2% and 7.9% in male and female rats, respectively, after IV
dosing and 15.1% and 9.0%, respectively, after oral dosing (Kim etal.. 2018b). The higher fraction
eliminated in feces after oral dosing was attributed in part to incomplete absorption by that route.
Therefore, an average value of 8% from the IV data was used for extrapolation to humans.
The excretion of PFHxS has been observed in humans both directly through measuring
PFHxS in urine and indirectly through the observation of changes in serum or plasma
concentrations over time. Changes in serum or plasma concentrations are informative of excretion
because PFHxS is not metabolized, thus any observations of decreasing concentrations in blood
after the distribution of the chemical were attributed to excretion. Most observations were within
populations with higher exposure than the general population, either workers in fluorochemical
production fOlsen etal.. 2007: Gao etal.. 2015: Fu etal.. 20161. workers at a fishery where the
waters were contaminated with PFAS fZhou etal.. 20141. or with increased exposure via
contaminated drinking water fWorlev etal.. 2017: Li etal.. 20181. For measures of clearance and
half-life, geometric means were presented unless otherwise specified because geometric means are
less influenced by extreme values that are common in these skewed distributions.
3-24
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Humans
Half-life estimates
Four studies reported half-life values for PFHxS based on observations of decreasing serum
levels in individual subjects at multiple time points after decreased exposure, either due to
retirement after occupational exposure (Olsen etal.. 2007). replacement of the foam used by
firefighters (Nilsson etal.. 2022a) or to the introduction of drinking water filtration at an
occupational site fLi etal.. 2018: Li etal.. 2022b). Li etal. (2022b) is a follow-up analysis of the
population evaluated by Li etal. f2018I Of the four studies, the Nilsson et al. study and the Olsen et
al. study fit the data for each person separately, while the two studies by Li and colleagues used a
mixed-effect statistical approach. Several plots in Olsen etal. f20071 showed declines in serum
levels over time that were very close to log-linear (i.e., showed negligible positive curvature), which
is suggestive of little effect of ongoing exposure for those subjects. However, Li etal. (2022b)
obtained a shorter half-life using data collected between 6 months and 1 year after the end of
exposure (mean ti/2 = 3.85 years) compared with using data collected 1-2.5 years after the end of
exposure (mean ti/2 = 4.33 years) or 2.5-4.5 years after the end of exposure (mean
ti/2 = 4.62 years). Positive curvature in a serum time-course plot after a decrease in exposure (for
example retirement), which is indicated by these results from Li etal. f2022bl. is evidence of
background exposure, as can be observed by examining Eq. 2 in (Bartell. 2012). The differences
among half-life values for the periods of evaluation reported by Li etal. (2022b). less than 20%, are
not statistically significant, however. Li etal. (2018) reported a mean half-life of 7.4 years in males
(n = 20) and 4.7 years in females (n = 30) aged 15-50 years old while Li etal. (2022b) reported a
median (5th, 95th percentile) half-life of 5.4 (2.34, 9.29) years (n = 114). Olsen etal. f20071
reported a half-life of 8.5 years in their cohort, which consisted of 2 females and 24 males at
retirement
The population of Li etal. (2022b) included children and the mean half-life for those
participants 1-14 years of age was 3.01 years compared with 5.26 years for participants 15-
50 years of age and 6.41 years in participants over 50 years. The much lower apparent half-life in
the 1-14-year-old group is almost certainly the result of PFHxS dilution into the growing bodies of
the youth. The intermediate half-life for participants 15-50 years of age may be partly attributed to
the difference between males (mean 5.39 years) and females (mean 4.48 years) which may reflect
higher clearance due to menstrual fluid loss for women in that age range. This difference of 17% in
half-life is in contrast to minimal differences of less than 2% between males and females aged 1-14
and less than 3.7% between males and females over age 50.
Nilsson etal. (2022a) analyzed PFHxS concentrations in firefighters after PFHxS was
removed from the formulation of the foam used for fire suppression. (97.5% of the recruited
population were male and the exact number of women in each sub-cohort was not reported, so the
results will be assumed to represent males.) The subjects had a range of serum concentrations at
the start of the study that overlapped with those found in the general population, which would
3-25
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
come from other exposure sources that are presumed to be shared by the study subjects. Since the
level of these other exposure sources is not precisely known and likely varies over time, the
contribution from them represents an uncertainty that would particularly impact half-life estimates
of subjects with initial concentrations in the general population range. Therefore, EPA chose to use
the results reported for only those subjects whose initial PFHxS concentration was greater than the
95th percentile of the general population, which ranged from just above that 95th percentile to
over 20 times higher. Nilsson etal. (2022a) reported a mean (95% CI) half-life of 7.7 (7.1, 8.3) years
for this group without background subtraction and a mean (95% CI) half-life of 6.7 (6.2, 7.2) years
after subtracting age-specific average concentrations reported for the general Australian
population. The half-life calculation assumes a simple exponential decay, which would only be
accurate with no ongoing exposure or if background exposure is constant, allowing it to be
addressed by simple subtraction, and is a reasonable estimate given results from other study
populations, albeit from the same country. The modest difference in the mean half-lives obtained
with and without background subtraction for the highly exposed group indicates that background
exposure had some impact on the observed changes in serum levels for that group, but less than
15%. Hence, the value obtained for the highly exposed group with subtraction is considered to be
appropriate for describing the elimination of the PFHxS from occupational exposure of this cohort
with a minimal level of uncertainty due to the assumptions involved.
Worlevetal. (2017) estimated a population half-life by fitting a PK model to population
mean serum concentrations at two timepoints with an estimated ingestion rate for that population.
Because Worlevetal. (2017) did not evaluate individual elimination, only measured serum levels at
two time points, and relied on an estimated exposure level, their study was considered to have
greater uncertainty than the other studies, with results that are more difficult to interpret in terms
of being a mean or geometric mean of individual values. In particular, it is possible that the drinking
water concentration was not constant as was assumed by Worlev etal. f20171 or that there were
other significant sources of ongoing exposure. Because of these methodological concerns, the
results of Worlev etal. (2017) were not used in estimating an overall average clearance for humans,
although it is noted that the corresponding clearance (0.031 mL/kg-day) is identical to the
estimated geometric mean across other studies (see Table 3-4).
As described in Volume of Distribution (in Section 3.1.2), Chiu etal. f20221 applied a one-
compartment PK model in a Bayesian analysis of human serum concentrations matched with
drinking water (DW) concentrations of several PFAS, including PFHxS, from multiple community
studies. Since the overall approach and parameter estimation method were considered sufficiently
sound, the resulting clearance was combined with other published human parameters in estimating
overall population clearance and volume of distribution (see Table 3-4).
Clearance rates estimated from half-lives
The clearance rate for a single-compartment PK model is related to the half-life and volume
of distribution by the following equation:
3-26
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
CL = ln(2)-Vd/T0.5 (3-2)
The approach for Bayesian analysis of PK data described in Appendix E was used to re-
analyze the monkey PK data from Sundstrom et al. f20121. resulting in mean volumes of
distribution of 272.0 mL/kg for males and 222.9 mL/kg for females, for which the average is
247.5 mL/kg. Using either the sex-specific Vd for corresponding segregated human studies or the
average Vd for results from mixed populations, values for total human clearance were estimated
from the half-life values:
• Li etal. f20181: 0.070 mL/kg-d in males and 0.090 mL/kg-d in females (same participants
as Li etal. (2022b)).
• Li etal. f2022bl: 0.096 mL/kg-d in male participants aged 15-50 years, 0.094 mL/kg-d in
females aged 15-50 years and 0.073 mL/kg-d in males and females aged >50 years
(participants below age 15 not included due to impact of growth)
• Nilsson etal. (2022a): 0.078 mL/kg-d in adults (age 22-82, 97%-98% males).
• Olsen etal. (2007): the clearance for each subject was calculated as described above for the
24 men and 2 women in the study.
o The geometric mean (arithmetic mean) of the resulting values is 0.071 (0.061) mL/kg-d
in males.
o Clearance in the two women ranked second and third lowest in the entire set
• Worlevetal. (2017): 0.030 mL/kg-day in men and women
These total clearance values also incorporate routes of clearance in addition to renal and
menstrual-associated clearance, which could consist of fecal clearance, shedding of skin, and
clearance due to childbirth and lactation, to the extent that these occurred in the study populations.
Urinary clearance estimates
Four studies directly evaluated urinary clearance of PFHxS in humans from matched serum
and urine concentrations (Zhang etal.. 2013b: Yao etal.. 2023a: Gao etal.. 2015: Fu etal.. 2016). Of
these studies, the ones with occupational cohorts Gao etal. f20151 and Fu etal. f20161 had much
greater exposure than the general population fZhang etal.. 2013b: Yao etal.. 2023al. Yao et al.
f2023al estimated clearance in infants, while all other studies were in adults. Their results are as
follows:
Fu etal. (2016) measured serum and urine PFHxS concentrations in matched samples from
occupationally exposed workers, and while they converted the results to half-lives for reporting
the paper states that Vd = 230 mL/kg was used for the estimate. Given a reported geometric mean
(GM) half-life of 19.9 years in men, the corresponding clearance is 0.022 mL/kg-d. The GM urinary
clearance for women in the study (reported in the text) was 0.024 mL/kg-d. That the overall
population GM was reported to be 0.023 mL/kg-d increases confidence in the CL in men back-
calculated here (0.022 mL/kg-d).
3-27
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Gao etal. (2015) did not distinguish between sexes but did distinguish between isomers of
PFHxS and found much greater clearance for the branched isomer, GM = 0.18 mL/kg-day, compared
with the linear (n-) isomer, GM = 0.04 mL/kg-d, with an overall clearance GM of 0.05 mL/kg-d for
total PFHxS, in a mixed population of men and women. The values for n- and total are between
those estimated from the half-lives of Li etal. f20181 and Olsen etal. f20071 (0.06-0.07 mL/kg-d)
and the urinary clearance values estimated by Fu etal. (2016) and Zhang etal. (2013b) (0.02-
0.03 mL/kg-d).
Zhang etal. (2013b) obtained GM values of 0.018 for men and older women and 0.028 for
younger women, which is in the range of total clearance estimated from Worlev etal. f20171. That
the GM values of Zhang etal. f2013bl are within an order of magnitude of the overall population
GM provides confidence that the true value is within an order of magnitude of those reported.
These route-specific clearance estimates do not include fecal elimination. After IV dosing
Kim etal. (2018b) measured fecal/urinary excretion rates of 8.2% and 7.9% in male and female
rats, respectively. Therefore, total excretion for Fu etal. (2016). Gao etal. (2015). and Zhang et al.
(2013b) was estimated as 1.08 times the estimated urinary excretion rates (i.e., 100% of urinary
excretion plus 8% of urinary excretion for fecal clearance) to determine an overall total clearance in
humans. The value estimated from a rat study was deemed appropriate as there is no human or
primate data on the relative amount of fecal and urinary excretion. There is uncertainty in
assuming that the relative amount of fecal and urinary excretion in humans is similar to rats, which
could be reduced by additional relevant human or primate data.
Yao etal. (2023a) estimated urinary clearance of PFHxS and other PFAS in infants, based on
the ratio of the estimated urinary excretion rate to estimated cord serum concentration. Cord blood
was collected at delivery and the concentration multiplied by two to account for the serum-to-
whole-blood ratio. Urine was collected in disposable diapers collected over the first postnatal week
and later extracted for measurements. The methods do not specify how a daily average urine
concentration was then determined from the set of samples for each infant, but it is presumed that
the extracted urine from all diapers collected during the week was mixed prior to analysis, resulting
in a "mixing cup" average concentration for the week. The resulting concentration was then
multiplied by a reported average urine elimination rate in infants of 48 mL/kg-day, rather than
using the actual urine volume collected. Since the serum concentrations and resulting urinary
elimination of breast-fed infants are expected to increase significantly after childbirth based on
reported breast milk: maternal serum distribution and breast milk ingestion rates, while the cord
blood concentration might only match the infant blood concentration at the moment of birth, the
resulting estimate of infant clearance is likely to be an overprediction of the true clearance rate.
From a population of 20 infants, the median (15th, 75th percentile) urinary clearance was 0.270
(0.108, 0.781) mL/kg-d, with a mean value 0.956 mL/kg-day, i.e., an order of magnitude higher than
the rate estimated in adults. The sample distribution is clearly skewed, with a maximum estimated
value of 11.7 mL/kg-day perhaps due to the urine sample timing issue discussed here. While
3-28
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
glomerular filtration is still developing in neonates, the expression of renal 0AT1 and 0AT3 is also
below adult levels fBueters etal.. 20201. and urinary excretion of PFNA will depend on both of
these opposing factors in a manner that cannot be quantitatively predicted. Given these
uncertainties, the results of this study will not be used quantitatively, though they indicate that
neonates will have lower serum levels of PFNA per unit exposure than adults.
Sex differences in human PFHxS PK
Zhang etal. (2013b) shows a small quantitative difference in urinary clearance between
men and older women and younger women (i.e., 0.01 mL/kg-day). It is possible that this difference
derives from differences in renal expression of renal transporters between men and women
(Murray. 20171. but it could also be due to random intersubject variability, given the overall range
of clearance observed across studies, and based on the overall range of clearance in each group, the
difference is not statistically significant. Hence, there does not appear to be a systematic difference
between men and women in the urinary clearance of PFHxS. However, Li etal. (20181 evaluated the
overall elimination of PFHxS from men and women who had previously high drinking water
exposure (i.e., after intervention to remove that exposure) and estimated a 67% higher elimination
rate in women than men between 15 and 50 years of age Li etal. (2022b).
Zhang etal. f2013bl also calculated a rate for menstrual clearance based on a study of PFOA
and PFOS that estimated menstrual blood loss using measurements of the blood quantity excreted
(Harada etal.. 2005). This estimate of menstrual blood loss was not specific to PFOA or PFOS and is
also potentially applicable to PFHxS. However, Harada et al. (2005) cite Hallberg et al. (1966) as the
source for a menstrual blood loss of 70 mL per cycle, but according to Hallberg, "The mean value of
the menstrual blood loss was 43.4 ± 2.3 mL in the entire series" [of experimental groups] and "the
upper normal limit of the menstrual blood loss is situated between 60 and 80 mL." Thus,
70 mL/cycle appears to be closer to an upper bound for healthy women and is not consistent with
the mean difference in PFHxS levels between men versus women evaluated below. More recently
Verner and Longnecker (2015) reviewed Hallberg et al. (1966). evaluated both blood loss and total
fluid loss from menstruation and concluded that the fluid lost in addition to blood was likely to be
serum, with the corresponding serum binding proteins and associated PFAS. Including this serum
loss and assuming 12.5 menstrual cycles per year, Verner and Longnecker f20151 estimated an
average yearly total serum loss of 868 mL (69.4 mL/cycle or 72.3 mL/month). Assuming an average
human female body weight of 72 kg (mean value for women 21-30 years of age from Table 8-5 of
(U.S. EPA. 2011a)). the corresponding average rate of clearance is
868 mL/(365 d)/(72 kg) = 0.033 mL/kg-d.
The U.S. EPA performed an analysis of data from NHANES for several PFAS, including
PFHxS, similar to that of Tain and Ducatman (2022). who found significantly lower levels of PFHxS
in females versus males between ages 12 and 57, a pattern also suggesting that menstruation or
other factors associated with reproductive age in women results in this difference. Specifically, EPA
analyzed the collection of NHANES waves from 2003-2004 through 2017-2018. Participants were
3-29
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
included if they were age 12 and above and if they had measured PFAS levels but were excluded if
they were pregnant or if they were currently breastfeeding. Data for women who were never
pregnant was also analyzed, since pregnancy and breastfeeding can reduce the body burden of
PFHxS. The reduction due to pregnancy and lactation will confound an attempt to evaluate
clearance outside of those lifestages, in particular for women prior to becoming pregnant for the
first time whose body burden is expected to be higher than women who have previously been
pregnant (and breast-fed their child). For all waves except 2003-2004, this information on
reproductive status was available only for women aged 20-44. This resulted in a total of 16,162
measurements. In the case for which a serum concentration was below the limit of detection (LOD),
the value was imputed with the LOD/V2. Overall, 1.5% of the PFHxS measurements were below the
LOD. This analysis was carried out in R (R Core Team, 2022), and the R package "survey" was used
to incorporate the NHANES survey strategy into the analysis and generate results applicable to the
U.S. population (Lumlev. 2004. 2023). Significant differences in serum levels in men versus women
were found for PFHxS (see Figure 3-3). Qualitatively similar results, though with a smaller
magnitude, were reported for PFOA, PFOS, and PFNA (lain and Ducatman. 2022).
Males
— • — Females
• • • • NP Females
Males + SD
Females + SD
NP Females + SD
— Males - SD
— Females - SD
NP Females - SD
c
o
00
rsl
<£>
o
<3-
00
rsl
<£)
o
rH
CNl
CNl
CNl
m
m
"3-
"3"
LD
m
i£>
UD
r--
r-
00
y
y
V
V
V
V
Y
V
V
V
V
v
V
V
V
V
V
fSl
<£>
o
4-
00
rsl
ID
o
-4
00
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
The data shown in Figure 3-3 strongly indicate a higher rate of PFAS excretion in women of
reproductive age, i.e., roughly between ages 12 and 50. The even lower levels observed in all
(nonpregnant, non-breastfeeding) women compared with never-pregnant women likely reflect the
transfer of PFHxS to the fetus and via breastfeeding. So, to evaluate the impact of reproductive
status outside of those events, EPA calculated the geometric mean ratio of mean PFHxS
concentrations in men to never-pregnant women between 20 and 52 years of age, the range over
which the concentration in never-pregnant women is fairly constant and has modest variance (see
Figure 3-3). On average, men in that age range have 76.2% higher PFHxS serum concentrations
than women, which is similar to the 67% difference in elimination rate reported by Li etal. f2022bl
for 15-50 years of age. If men are assumed to have an average clearance rate of 0.041 mL/kg-d
(weighted geometric mean clearance from Table 3-4 below), then 76.2% of this value is
0.031 mL/kg-d, which is almost identical to the average rate of menstrual blood and fluid clearance
estimated by Verner and Longnecker (2015) from the data of Hallberg etal. (1966). 0.033 mL/kg-d.
EPA recognizes that not all women menstruate regularly for various reasons. For example, Toubert
etal. (2022) found that 16% of a study population of elite female competitive climbers had
amenorrhea and the American College of Obstetricians and Gynecologists (ACOG) states, "About 1
in 25 women [4%] who are not pregnant, breastfeeding or going through menopause experience
amenorrhea at some point in their lives."5 More recently, a committee of the American Society for
Reproductive Medicine (ASRM) stated a prevalence of 3-4% (ASRM. 2024a). although it should be
noted that the supporting references are from 1973 and 1982 (Pettersson etal.. 1973: AG and E.
1982). the latter being specific to college students. Hence, this statistic may not reflect the current
prevalence in the population of reproductive-age women. In addition, 45% of women in a recent
survey under 30 years of age use hormonal birth control fProl etal.. 20241. Use of hormonal
contraceptives is known to suppress menstruation fHillard. 20141 and that suppression is
frequently a primary reason for their use fASRM. 2024bl. Ameta-analysis found that 83% of women
became pregnant within 12 months of discontinuing contraception (Girum and Wasie. 2018). which
would not be enough time for them to achieve steady state, including the menstrual-associated
clearance. Hence application of the clearance rate from Verner and Longnecker (2015) might over-
estimate the average clearance among all women of childbearing age. Further, while the association
of higher clearance with menstruation as a specific mechanism is further supported by the analysis
of Tain and Ducatman f20231. who evaluated the impact of hysterectomy, menopause, and hormone
replacement therapy, proof that PFAS are excreted in menstrual fluid at concentrations equal to
those found in blood plasma is not available.
There are also data indicating or showing sex- and age-related differences in PFHxS
pharmacokinetics by mechanisms other than menstruation. Koponenetal. (2018) estimated that
the body burden of PFHxS in girls appeared to decline between ages 6 and 10.5 while that in boys
remained about constant over this age range, resulting in significantly higher body burdens in boys
5https://www.acog.org/womens-health/faqs/amenorrhea-absence-of-periods.
3-31
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
than girls at age 10.5 (p = 0.01). While some girls reach menarche before age 11, increased
clearance by that mechanism will take years to impact serum levels, as indicated by the trend in the
NHANES data for females from ages 12-20 (Figure 3-3). Hence, the difference between boys and
girls at age 10 reported by Koponen etal. f20181 is most likely due to factors other than menstrual
clearance and probably contributes to the difference in the NHANES data for 12-16 years of age
(Figure 3-3). Zhang etal. (2013b) obtained geometric mean urinary clearance of 0.028 mL/kg-day
in women < 50 years vs. 0.018 mL/kg-day in men and older women. Renal transporters involved in
the resorption of PFAS in rodents are known to be under sex-dependent hormonal control (Weaver
etal.. 20101. so a similar dependence could contribute to the observed male-female difference in
serum levels of PFHxS from NHANES. Finally, these analyses do not account for any male-female
differences in exposure to PFHxS that may occur. Therefore, while there is strong correlative data
for menstruation as a mechanism for PFAS clearance and the hypothesis is mechanistically
plausible on a qualitative basis, there is quantitative uncertainty with regard to the extent to which
the observed difference in serum PFHxS levels depends on it.
Recognizing the caveats just noted on the extent of menstruation and its mechanistic
contribution to PFHxS clearance, the NHANES data analyzed here are taken to be empirical
evidence of higher clearance independent of the specific mechanism in nulliparous women. While
"NHANES excluded all persons in supervised care or custody in institutional settings, all active-duty
military personnel and active-duty family members living overseas, and any other U.S. citizens
residing outside of the 50 states and the District of Columbia"6 the survey does not filter for
contraceptive use or menstrual status and so provides a representative cross-section of women in
the U.S., inclusive of these factors. Hence, application of the slightly lower additional clearance
estimated from the NHANES data for women of reproductive age, 0.031 mL/kg-d, should
sufficiently account for the effect of the population variability in menstrual fluid loss among women
on the average clearance in women and provide an estimate of higher clearance that is agnostic to
the specific mechanism. The variation in menstrual fluid loss likely contributes to the variation
serum concentrations among nulliparous women shown in Figure 3-3, which indicates high
variability in this population, and the estimated mean values for each age range incorporate that
variation.
The analysis of the NHANES data assumes equal exposure of women and men but there are
recognized differences in exposure. For example, as described in Section 1.1.4, over 45% of
freshwater fish samples had detectable levels of PFHxS fStahl etal.. 20141 but it was not detected in
U.S. grocery store finfish and shellfish samples (Ruffle etal.. 2020). An analysis of NHANES fish
consumption data by the U.S. EPA (U.S. EPA. 2014a) (see Table 9a) found that the median
consumption of freshwater and estuarine finfish and shellfish in adults 21 years of age or older was
4.1 g/d in females and 6.2 g/d in males. A more recent study specific to freshwater fish
6https: //www.cms.gov/About-CMS/Agency-Information/OMH /resource-center /hcps-and-researchers /data-
tools/ sgm-clearinghouse / nhanes.
3-32
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
consumption in a Swedish population found a similarly higher consumption (normalized to body
weight) among men than women fAugustsson etal.. 20211. Intrinsic PK variability, such as that due
to variation in menstruation or glomerular function, can potentially be incorporated into a PK
analysis while variation in exposure is only indirectly related to PK. Even focusing on intrinsic,
biological factors, a comprehensive analysis of these factors on the clearance of PFHxS is not part of
the available science and EPA considers use of the standard uncertainty factor for human
variability, UFh, sufficient to address PK (and pharmacodynamic variability. This application is
discussed in more detail just below. Therefore, a comprehensive, quantitative analysis of the
specific factors that contribute to PK variability has not been conducted for this review.
While the impact of variation in menstruation is included in the estimated mean values,
since the data for nulliparous women was used for the current analysis, the additional losses
associated with pregnancy (e.g., placental loss) and breastfeeding are not incorporated into and
hence do not impact the estimated increase in average clearance among women of childbearing age.
These additional losses likely explain the observed difference between results for all women (not
pregnant or breastfeeding at the time of evaluation) and NP women shown in Figure 3-3. Further,
the subsequent application of UFh, which is presumed to include a factor of 3 for PK variability,
should sufficiently adjust for the variability in menstruation-associated clearance among women,
including those who are completely amenorrheic. (Applying the Vd values estimated from male and
female monkeys (Sundstrom etal.. 2012) to men and women respectively also led to some
difference in the corresponding half-life estimates below.) Therefore, the geometric mean ratio of
PFHxS concentration in never-pregnant women versus men, 0.5675 (1/1.762), will be used to
estimate clearance aside from that associated with menstruation in analyses of female data below,
i.e., the fraction of total clearance in women of childbearing age not associated with menstruation
will be calculated as 0.5675 times the total estimated clearance. Then, when estimating the total,
population-average clearance for women of childbearing age, it will be assumed to be 1.762-fold
higher than that estimated for males and older women.
Lorber etal. (2015) also examined the effects of ongoing blood loss through menstruation
or through frequent blood withdrawal as a medical treatment Male patients with frequent blood
withdrawal had serum concentrations 40%-50% less than males from the general population for
the chemicals observed in the study (PFOA, PFNA, PFDA, PFHxS, and PFOS). Female patients also
had a lower serum concentration than females from the general public. The trend in relation to the
number of recent blood draws or in the recency of the last blood draw was not examined for PFHxS,
but was for PFOA and PFOS, and significant associations were observed in PFOS only. This study's
analysis of the impact of menstrual blood loss was purely a modeling exercise, which was
performed for PFOA and PFOS. The authors estimated a monthly blood loss of 35 mL (which is close
to the median loss of 43.4 mL reported by Hallberg et al. f196611. 50% of which was serum,
resulting in a clearance of 17.5 mL/month, or 0.0081 mL/kg-day in a 72 kg woman. This value is
also chemical-independent and could be applied to PFHxS instead of the menstrual clearance
3-33
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
estimated by Verner and Longnecker (2015). but would not be sufficient to explain the large
difference in PFHxS serum concentrations between men and women in the NHANES data (see
Figure 3-3).
As mentioned in the distribution section (see Section 3.1.2), PFHxS has been observed in
breast milk, so lactation can act as an excretion route for a nursing mother. One study examined the
association between maternal serum concentrations and the length of breastfeeding and found a
weak, nonsignificant inverse association. There were stronger inverse associations for the other
PFAS studied, PFOA, PFOS and PFNA, suggesting that there may be less transfer of PFHxS to breast
milk than other PFAS, or that the variation in serum level between people is large compared with
the impact of breastfeeding.
Impact of kidney disease on urinary clearance in humans
lain and Ducatman (2019c) evaluated the relationship between PFHxS serum
concentrations and states of kidney disease and reported an inverted U-shape response of PFHxS
with GFR. Specifically, higher PFHxS concentrations are observed with GFR in the second and third
tertiles. (The result was also obtained in analyses stratified by sex). Since urinary excretion is
estimated to account for over 90% of clearance in males and over 50% of clearance in females, it is
mechanistically likely that reduction in GFR (second and third tertiles, with an absence of
albuminuria) will reduce clearance and so result in higher serum concentrations of PFHxS. On the
other hand, more advanced kidney failure that results in albuminuria would be expected to
increase PFHxS clearance, resulting in lower serum PFHxS concentrations, since renal resorption of
albumin-bound PFHxS should be significantly lower than resorption of unbound. A negative
correlation between PFHxS serum concentrations and albuminuria was reported flain and
Ducatman. 2019b). Together these observations indicate that variation in renal health, or the
degree of kidney disease, contributes to variation in PFHxS clearance in the population and should
be considered as a factor when evaluating epidemiological data for the relationship between PFHxS
exposure and kidney disease. Moderate kidney disease without albuminuria may result in inverse
causality (this level of disease causes an increase in PFHxS serum concentrations) and with
albuminuria inverse causality could occur (the disease results in a decrease in PFHxS serum
concentrations). However, an analysis of the specific effect of kidney disease on the population-
level PFHxS clearance (i.e., dependent on the prevalence of each level of kidney disease in the
population, stratified by age and sex) is not in the available science. EPA considers the variability in
the empirical human clearance data, already evaluated, sufficient to characterize overall population
variability in clearance, inclusive of the impact of kidney disease. Hence, further research and
analysis to specifically account for the variation in PFHxS clearance due to kidney disease was not
conducted.
3-34
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Dosimetry of linear versus branched isomers
Gao etal. f20151 is the only PK study to provide separate estimates of elimination for linear
versus branched isomers in humans. With the clearance of the branched isomer being so much
higher than the linear, the body burden is expected to be much higher for the linear than the
branched isomer, given equal exposures. Using the clearance for the sum of PFHxS accounts for the
relative prevalence of the different isomers in the serum of the participants. Therefore, the result
for mixed or total PFHxS from Gao etal. (20151 will be used in combination with the results of the
other PK studies. The result is interpreted as reasonably health-protective across all forms.
Summary of human PFHxS excretion
A summary of the clearance values reported or estimated from each of the adult human
elimination studies is provided in Table 3-4.
Table 3-4. Summary of clearance values estimated for humans
Study
(basis)
Clearance
(mL/kg-d)
(Half-life, y)
N
Notes
Chiu et al. (2022) (serum levels vs.
drinking water exposure)
0.0685
(8.30)
41
Geometric mean; 37 individuals and 4
population mean results
Fu et al. (2016) (urinary clearance with
fecal estimate3)
0.025
(18.9)
207
Geometric mean; 136 men, 71 women
Gao et al. (2015) (urinarv clearance with
fecal estimate3)
0.054
(8.70)
36
Geometric mean for total linear and branched
PFHxS; result based on 57 paired samples
from 22 men, 14 women
Li et al. (2018) (empirical half-life)
0.0698
(7.4)
24
Men aged 15-50; CL calculated from mean
half-life using Vd = 272 mL/kg
Li et al. (2018) (empirical half-life)
0.051
(8.3)
28
Women aged 15-50; CL calculated from mean
half-life using Vd = 222.9 mL/kg and
multiplying by 0.567 to remove menstrual-
associated clearance.
Olsen et al. (2007) (empirical half-life)
0.071
(7.3)
26
Geometric mean of individual clearance
values, calculated from reported half-lives as
described above; 24 men, 2 women (all
>59 yr)
Li et al. (2022b) (empirical half-life)
0.096
(5.4)
22
Males, ages 15-50; CL calculated from mean
half-life using Vd = 272.0 mL/kg
Li et al. (2022b) (empirical half-life)
0.054
(7.9)
30
Females, ages 15-50; CL calculated from
mean half-life using Vd = 222.9 mL/kg and
multiplying by 0.567 to remove menstrual-
associated clearance.
Li et al. (2022b) (empirical half-life)
0.073
(6.4)
33
Age > 50; CL calculated from mean half-life
using Vd = 247.5 mL/kg
Nilsson et al. (2022a) (empirical half-life)
0.078
(6.7)
99
Participants with initial serum PFAS
concentrations greater than the 95th
3-35
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
(basis)
Clearance
(mL/kg-d)
(Half-life, y)
N
Notes
percentile of the General Australian
Population in 2016-2017; age 22-82, 97%-
98% males; CL calculated from mean half-life
using Vd = 272.0 mL/kg
Worlev et al. (2017) (half-life fitted for PK
modelb)
0.026
(15.5)
45
Clearance calculated using Vd = 213 mL/kg
(value used in the PK model); 22 men, 23
women
Zhang et al. (2013b) (urinarv clearance
with fecal estimate3)
0.030
(14.0)
19
Geometric mean; women age <50 y; half-life
calculated from CL using Vd = 222.9 mL/kg
Zhang et al. (2013b) (urinarv clearance
with fecal estimate3)
0.019
(26.6)
64
Geometric mean; all men and women age >50
y; half-life calculated from CL using
Vd = 272.0 mL/kg
Weighted geometric mean
0.041cd
(11.5)
577
Exp Z[log(CLi)-Ni] / Z[Ni]; half-life calculated
from CL using Vd = 247.5 miykg
aReported urinary clearance was multiplied by 1.08 based on observed fecal/urinary elimination in rats after IV
dosing (Kim et al., 2018b).
bHalf-life determined from fitting PK model to geometric mean of serum concentrations measured in 2010 and
2016, accounting for estimated ongoing exposure.
Calculated for all studies except Worlev et al. (2017) due to methodological issues identified for that study and U
et al. (2018) since data for that population are included in the data of Li et al. (2022b) (see "Half-life estimates").
However, the value is identical to the two significant figures shown if results from these two studies are both
included.
dVariance around this value can be described by a weighted geometric standard deviation of 1.6, which is a
multiplicative factor, or a weighted geometric coefficient of variance of 22%.
In Table 3-4, the subset of clearance values estimated from empirical half-lives (Olsen etal..
2007: Li etal.. 2018: Li etal.. 2022bl are fairly similar to each other after adjustment for
(subtraction of] menstrual-associated clearance, and similar to the results of Chiu etal. (20221. but
are higher than most of the urinary clearance values and the results of Worlev etal. f20171. which
were based on exposure estimated from drinking water concentrations measured at one time point
and may not reflect higher exposure concentrations in preceding years. While Kim etal. (2018b)
observed fecal excretion of PFHxS in rats to be only 8% of urinary excretion after IV exposure, it is
possible that fecal excretion and other routes such as shedding of dead skin contribute enough to
the overall clearance to account for the two- to threefold difference between those estimated from
empirical half-lives (Olsen etal.. 2007: Li etal.. 2018: Li etal.. 2022b) and the estimates of urinary
clearance. In this case, the weighted geometric mean clearance shown in Table 3-4 will
underpredict overall clearance to that extent However, it also possible that the empirical half-lives
reflect urinary clearance under conditions of saturated renal resorption, which is not
representative of the general population at lower exposure levels, but Chiu etal. (20221 attempted
to exclude very highly exposed individuals (i.e., with occupational exposure) and also obtained a
relatively high clearance.
3-36
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Data on how clearance may vary as a function of age (i.e., in rat pups or children compared
with adults) and during pregnancy are mostly lacking. Li etal. f2022bl did estimate the half-life in
individuals 1-14 years of age and found it to be about one-half of that in older individuals (3 years
versus 6 years), but this is an apparent half-life that likely includes the impact of growth. As
discussed above, Yao etal. f2023al estimated urinary clearance of PFHxS in infants to be almost
seven times higher than the estimated clearance rates in adults, 0.27 versus 0.041 mL/kg-d, but the
approach used may have over-estimated the rate. Renal excretion varies in proportion to body
surface area with age over most of the lifetime but is still developing in newborns along with
expression of organic anion transporters (OATs) fBueters etal.. 20201 that are associated with
renal resorption of PFAS, and the volume of distribution may also vary with age. In the preceding
section, "Distribution in fetal tissues and children," the possible effect of changes in extracellular
water and blood volume as a fraction of BW in children was discussed. Finally, the absence of a
reliable pharmacokinetic model which can account for these factors and the likely differences in
accumulation of PFHxS in humans exposed chronically versus in experimental animals during
relatively short-term health effects studies creates uncertainty in simpler pharmacokinetic
extrapolation based on clearance. Nevertheless, the analysis of human sex differences in PFHxS
clearance from NHANES data indicates a difference in average clearance between women of
reproductive age and men of 76.5% (i.e., that clearance in women is 1.765 times greater than men),
which is quite consistent with the weighted geometric mean clearance of 0.041 mL/kg-d (computed
after reducing clearance in reproductive age by this factor, but primarily using data from men and
older women) and the average menstrual fluid clearance of 0.033 mL/kg-d from Verner and
Longnecker (2015) (That is to say that if the relative clearance in women versus men is calculated
by adding the rate of menstrual clearance from Verner and Longnecker f20151 to the average rate
estimated for men and older women, then that rate estimate is (0.041 + 0.033)/0.041 = 1.8 times
higher in women, compared with an estimate of 1.765 times higher clearance in women versus men
calculated from the NHANES data alone). To be clear, this adjustment addresses a population-
average difference between men and women, while variability in clearance among women,
including that which may be due to variation in the rate of menstrual fluid loss, is addressed by
application of the PK portion of UFh, UFh,pk = 3, which is much greater than the factor of 1.765 being
applied for menstrual-associated clearance. Hence, together with application of UFh,pk, the
estimated clearance in women of childbearing age should be sufficiently protective of women
whose clearance is no greater than that estimated for men (and older women). It should also be
noted that this factor of menstrual-associated clearance will not be applied when evaluating
dosimetry for nondevelopmental effects, such as thyroid perturbations observed in adult female
rats. For those endpoints, the analysis effectively assumes that all women are effectively
amenorrheic.
3-37
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
The limited data available for neonates and children indicate that their clearance is higher
than adults, so use of the adult value for estimating dosimetry in children should be health-
protective of that population.
While the range of values in Table 3-4 represent a range of uncertainty of fivefold, given the
number of estimates it seems unlikely that the true clearance in humans would be lower than the
minimum value of 0.019 mL/kg-d from Zhang etal. (2013b). The weighted geometric mean
clearance of 0.041 mL/kg-d is 2.2 times higher than this minimum and based on the overall
evidence was considered sound for use in estimating human equivalent doses (HEDs) for points of
departure (PODs) estimated from animal toxicity studies or blood concentrations estimated from
epidemiological evaluations, with clearance in women of reproductive age set to 1.765 times
higher, 0.072 mL/kg-d.
The clearance values shown in Tables 3-3 and 3-4 were compared with species-specific
glomerular filtration rate (GFR), with and without adjustment for serum protein binding, to
evaluate the possible role of those mechanisms. Considering the time period used by Davies and
Morris (19931. this comparison used that value for average human BW, 70 kg, which results in an
estimated GFR/BW of 2.57 L/kg-d in humans, 83,000 times greater than the empirically estimated
geometric mean clearance for humans. Kim etal. f2018bl reported an average PFHxS free fractions
(/free) of 0.00025 in humans, which led to GFR x/free = 0.64 mL/kg-d, which is still almost 16 times
greater than the geometric mean empirical clearance. Thus, it appears likely that there is significant
renal resorption of PFHxS in humans.
Comparing the human CL values to those predicted from allometric scaling of mouse and
rodent CL values shows that allometric scaling appears to overpredict human clearance rats. BW3/4
allometric scaling suggested that CL in an 80 kg human should be 4.2 times lower than in a 0.25 kg
rat and 7.2 times lower than in a 30 g mouse. Applying a factor of 4.2 to the population mean CL
values for male and female rats in Table 3-3, resulted in predictions of human male CL of
1.1 mL/kg-d and female CL of 24 mL/kg-d, one to three orders of magnitude higher than the values
estimated from human data in Table 3-4. Likewise using the CL in mice and the allometric factor of
7.2 resulted in an estimated human male CL of 0.53 mL/kg-d and female CL of 0.44 mL/kg-d,
roughly an order of magnitude higher than observed. Performing this analysis for a 6 kg male
monkey or a 4 kg female monkey produces a similar overprediction, with extrapolated clearance
values of 0.62 and 1.0 mL/kg-d after applying scaling factors of 1.9 and 2.1. In summary, this
analysis indicated that use of BW3/4 scaling would have led to an overprediction of HEDs
(effectively an underprediction of risk) by one to three orders of magnitude, depending on the
animal species and sex in which a POD was identified. Hence, the use of BW3/4 scaling was avoided
for PFHxS, but comparisons of BW3/4 scaling to the selected approach (see Section 3.1.6) were
provided for context.
3-38
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Excretion Summary
The estimated average clearance values for adult humans are listed in Table 3-5. Since the
data in Table 3-4 were adjusted to remove menstrual-associated clearance for women of
reproductive age when estimating the general, nonspecific clearance in humans, a corresponding
correction was added for women in this age range. In particular, the estimate of relative clearance
in women versus men from NHANES described above, a factor of 1.765, is consistent with the
average menstrual fluid clearance estimate ofVerner and Longnecker (2015) (0.033 mL/kg-day),
which in turn is considered reasonable given its consistency with the original menstrual fluid
volume data of Hallbergetal. T1966I While this factor is considered appropriate for deriving HEDs
for reproductive effects in women since newly available data show that maternal serum levels
remain constant or decline during pregnancy and the early postpartum period, the additional
menstrual-associated clearance factor is considered appropriate for estimating HEDs for effects
occurring in utero or otherwise correlated with maternal serum concentrations measured during
pregnancy and postpartum.
However, because the current analysis should protect younger children, men, and older
women, it was considered appropriate not to include menstrual-associated clearance when
evaluating dosimetry in humans for health effects that can occur at any point in life, such as thyroid
perturbations, even though they may have been observed in laboratory animals of reproductive
age. This choice follows the typical approach when assessing susceptible sub-populations.
3-39
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-5. Summary clearance values for humans
Population
Clearance (mL/kg-d)
(Half-life, y)
References
Human geometric mean
(general population)
0.04 rb
(11.5°)
(Zhang et al., 2013b; Olsen et al., 2007; Nilsson et al., 2022a;
Li et al., 2022b; Gao et al., 2015; Fu et al., 2016; Chiu et al..
2022)
With menstrual fluid loss
(women of reproductive
age)
0.072
(5.9d)
Increased by a factor of 1.762 over the general population
value based on analysis of NHANES data to include
menstrual-associated clearance.
aHuman clearance estimates also depend in part on volumes of distribution (Vd) estimated for monkeys by
Sundstrom et al. (2012); clearance in women of reproductive age was reduced to remove menstrual-associated
clearance before calculating the rate for the general population.
bMeasurements of urinary clearance only were corrected for estimated fecal/urinary clearance ratio of 1.08 based
on observations in rats by Kim et al. (2018b).
Calculated as ti/2 = ln(2)*Vd/CL using Vd = 247.5 mL/kg, the average of the mean values estimated by the EPA for
male (272.0 mL/kg) and female (222.9) monkeys from the data of Sundstrom et al. (2012) (see Table 3-1).
Calculated as for the general population, but using the mean estimated Vd for female monkeys (see table note c).
3.1.5. Evaluation of PBPK and PK Modeling
The PFAS protocol (Supplemental Information document, Appendix A) recommends the use
of scientifically sound and validated physiologically based pharmacokinetic (PBPK) models as the
preferred approach for dosimetry extrapolation from animals to humans, while allowing for the use
of data-informed extrapolations (such as the ratio of serum clearance values) for PFAS that lack a
scientifically sound and sufficiently validated PBPK model. If chemical-specific information is not
available or too uncertain, the protocol then recommends that doses be scaled allometrically using
body weight (BW)3/4 methods. Selection from among this hierarchy of decisions considered both
the inherent and chemical-specific uncertainty (e.g., data availability) for each approach option.
This hierarchy of recommended approaches for cross-species dosimetry extrapolation is consistent
with EPA's recommendations on using allometric scaling for the derivation of oral reference doses
(U.S. EPA. 2011b). This hierarchy preferentially prioritizes adjustments that result in reduced
uncertainty in the dosimetric extrapolation.
A PBPK model was identified for PFHxS in rats and humans fKim etal.. 2018bl The
computational code for this model was obtained from the model authors and evaluated for
consistency with the written description in the published paper, the PK data for PFHxS, known
physiology, and the accepted practices of PBPK modeling. Unfortunately, several flaws were found
in the model. One flaw, an error in the balance of blood flow through the liver, had only a moderate
impact on model predictions. A much larger issue identified is that the model had only been
calibrated to fit the oral PK data for rats and the set of model parameters selected by the model
authors to match those data included an oral bioavailability (BA) lower than is otherwise supported
by the empirical PK data. For example, the fraction absorbed by the male rat was effectively set to
39% in the model when the empirical PK analysis showed 88%-92% bioavailability. Further, when
the model was used to simulate the intravenous PK data, data to which a PK model should be
3-40
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
calibrated, the parameters were found to be completely inconsistent with these data. Figure 3-4
compares results obtained with a replication of the PBPK model, which exactly matches the
published PBPK model results for oral dosimetry, with the data and empirical PK fit for a 10 mg/kg
IV dose to male rats.
The overprediction (approximately three to four times higher than the data for male rats) of
the IV data by the Kim etal. (2018b) model indicated that distribution into the body is significantly
underpredicted by the model, which was offset in the simulations of oral dosimetry data by use of
an unrealistically low oral bioavailability. Initial efforts to refit the model to the data did not
produce acceptable fits to both the IV and oral dose PK data and involved changing model
assumptions in a way that would require separate experimental validation before use. In particular,
to match the observed rate of decline in the blood as well as the observed accumulation in urine
and feces required an assumption of another route of excretion, for which there are no data. It was
therefore determined that the published model structure and underlying assumptions did not allow
for a sufficiently sound calibration of the model to the PK data, given the currently available
understanding of PFAS pharmacokinetics.
Plasma
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Fabregaetal. (2015) developed a PBPK model describing the dosimetry of multiple PFAS in
humans, including PFHxS. A concern with this model is that the tissue:blood partition coefficients
were estimated by comparing tissue concentrations measured in cadavers with blood
concentrations from different (living) subjects, albeit from the same geographic region. Also, the
brief description provided for the estimation of the parameters for saturable renal resorption was
considered not sufficient to allow for independent reproduction of that process and it was unclear
how the two constants can be independently identified from such data. Finally, model results for
PFHxS shown by the authors underpredict an epidemiological dataset (Rvlander et al.. 2009) by
about an order of magnitude. Therefore, the model was not considered further for use in this
review.
Verneretal. f20161 developed a coupled classical PK model, wherein single-compartment
models represented the mother and fetus or child, which incorporated growth of the fetus and
child, maternal body weight changes, and a time-varying rate of milk intake to account for the
decline in g/kg-day ingested with the child's age. With parameter samples selected from
distributions by Monte Carlo sampling, maternal exposure levels for individuals from two studies
were selected to match the observed maternal serum concentration at delivery (i.e., given the
sample set of parameter values) and the PFAS concentrations in the mother and child simulated for
the first 3 years of the child's life. Measured plasma levels in children at 6 months of age were fairly
well predicted, though the model tended to underpredict the plasma levels at age three, with many
observations more than twofold higher than predicted. A version of the PK model was implemented
and its ability to predict rat PK data was evaluated as described in Appendix E.2. Unfortunately,
because of the underprediction of PFHxS concentrations in three-year-old children shown by
Verneretal. f20161 and the poor performance of the model in predicting rat PK data using
parameter values estimated for that species (see Appendix E.2), model predictions were not
considered sufficiently reliable for use in this assessment.
It is also noted that EPA's high-throughput toxicokinetics (HTTK) computational model
package (Pearce etal.. 2017) predicts dosimetry for PFHxS. However, this model currently does not
account for the activity of transporters, in particular those involved with renal resorption, so
clearance (in the absence of metabolism) is estimated as the free fraction in blood times the
glomerular filtration rate. The HTTK package estimates the half-life of PFHxS in humans to be
38 days or 0.11 years, corresponding to CL = 4.1 mL/kg-d (using Vd for female monkeys), two
orders of magnitude higher than that estimated from the empirical in vivo human data. Hence, the
HTTK model was also not considered further for use in this review.
Bil etal. (2022) used a classical two-compartment PK model structure to estimate internal
dose relative potency factors for liver toxicity observed in male rats for nine PFAS, including PFHxS.
Since the PK model parameter estimation was performed separately for each PFAS, only the results
for PFHxS need to be discussed here, but it is noted that the objective of the paper was to develop a
method for the prediction of toxicity from exposure to PFAS mixtures. For, PFHxS, Bil etal. f20221
3-42
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
used the PK data of Huang etal. (2019a). one of the studies included in EPA's analysis, and obtained
results for a single compartment (monophasic clearance) with a volume of distribution of
137 mL/kg and a half-life of 16.5 days using the data for the 16 mg/kg dose. These values are
similar to those reported by Huang etal. f2019al for that dose (144 mL/kg and 16.9 days,
respectively), but somewhat lower than the results of EPA's analysis of multiple data sets including
Huang etal. (2019a) (mean values of 217 mL/kg and 21 days). Because EPA's clearance value is
obtained from analyzing data from all three dose levels used by Huang etal. (2019a) and data from
two other studies (Kim etal.. 2016b: Kim etal.. 2018b). it is considered superior for use in
pharmacokinetic extrapolation from animal-to-human points of departure.
Sweeney f20221 developed a PBPK model for PFHxS in humans. Model simulations were
conducted for individuals from 0 to 70 years of age and results were analyzed (compared with
data) for individuals from 12 to 70 years of age. The text indicates that an adjustment factor for
ingestion in children 0-10 years of age was employed, but gestational and lactational exposure are
not mentioned and pregnancy was not simulated. The model structure and assumptions and
adjustments for physiological changes with age appear to be sound and the author has compared
model results to a comprehensive set of human PK data.
Unfortunately, the model code for Sweeney f20221 contains a mass-balance error in which
the unbound fraction in plasma (CAFREE) is calculated as the total amount in plasma (APLAS)
divided by the plasma volume, which effectively means that distribution to tissues and urinary
elimination is not restricted by the plasma protein binding. If instead one interprets APLAS as only
being the amount free in plasma, then the corresponding total amount in plasma (APLAS/FREE) is
not included in the mass balance check for the model code. EPA's review of the model code
suggested that the variable APLAS is consistent with the total amount in the plasma, not the free
amount. For example, the differential equation for APLAS sums all the PFHxS that distributes out of
the liver after absorption from the stomach (based on the amount free in the liver), rather than
being only assigned the fraction that is free in blood. However, if the total amount in blood is
AP LAS/FREE, making this correction would add an amount approximately 40 times APLAS to the
overall mass balance equation, which would then likely demonstrate an overall mass balance error.
It is possible that the mass balance error in Sweeney (2022) is related to the inability of Kim
etal. f2018bl to correctly replicate the IV dosimetry in rats, noted above, in that both point to a
central assumption that appears to be incorrect. Kim etal. f2018bl correctly calculates the mass
balance in the plasma based on the assumption that only the free fraction in the plasma can
distribute to tissues, but then fails to predict that tissue distribution after IV dosing. The central
model code used by Sweeney (2022) was originally developed by Loccisano etal. (2011). who may
have inadvertently introduced the mass balance error in an attempt to correct for an inability of the
model to predict tissue distribution and urinary elimination. The resolution of this issue may
require relaxing the assumption that the free fraction and bound fraction in the serum are strictly at
equilibrium at all times, as opposed to being treated as a dynamic equilibrium with distinct rates of
3-43
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
association and dissociation. In the latter case, the rate of distribution to tissues and urinary
elimination would be limited by the rate of dissociation, which may be more rapid than the
equilibrium fraction free multiplied by the blood flow rate to the tissues (or glomeruli). A
mathematical model that incorporates the kinetics of plasma binding and release to describe
uptake of drugs by the brain has been previously described by Robinson and Rapoport f 19861. but
adaptation of this model to the tissue distribution of PFHxS would require measurement of the
separate rates of association and dissociation, data which have not been reported. Hence,
appropriate revision of the PBPK models was not possible for use in this assessment
Irrespective of the potential impact of the mass balance error, from Table 1 of Sweeney
f20221. the model predicts urine concentrations around 2.5 times higher than Fu etal. f20161 and
3.75 times higher than measured by Zhang etal. f2013bl. indicating an overall predicted clearance
of 0.06-0.07 mL/kg-d, consistent with the results of Li etal. f20181. whose data were used for
calibration. However, the result means that application of the Sweeney (2022) would be less health-
protective than use of the weighted geometric mean clearance, 0.041 mL/kg-d (see Table 3-5) and
would not address some of the other uncertainties noted here. For both this reason and the mass
balance issue, the model was not further considered for use in the current analysis.
Most recently, Chiu etal. f20221 applied a one-compartment PK model in a Bayesian
analysis of human serum concentrations matched with drinking water (DW) concentrations of
several PFAS, including PFHxS, from multiple community studies. Since the one-compartment
model structure is essentially identical to that already evaluated by the EPA and only addresses
exposure of adults, for whom body weight is presumed fixed, it was not considered further for use
as a PK model, but the overall approach and parameter estimation method were considered
sufficiently sound that the resulting parameters were combined with other published human
parameters in estimating overall population clearance and volume of distribution (see Table 3-4).
Yao etal. f2023al used a one-compartment PK model to estimate the time-course of
multiple PFAS, including PFHxS, in human children from birth to 1 year of age. However, the model
used a constant level of intake by the child, based on the breast milk concentration measured just
after birth and the volume of breast milk ingested per day for infants <1 month of age, and did not
account for the dilution due to growth of the child over that time. Breast milk intake is expected to
peak between 3 and 6 months of age and the intake per kilogram of body weight of the infant to
decline from the first month of age through the first year
fhttps://www.epa.gov/expobox/exposure-factors-handbook-chapter-151. while concentrations of
PFHxS in maternal serum declined on average in the first month after birth (Oh etal.. 2022). Hence,
the simulations of Yao etal. (2023a) likely overpredict the actual PFHxS time-course in children
after the first month of life.
3.1.6. Empirical Pharmacokinetic Analysis
To estimate sex-specific PK parameters with measures of uncertainty for male and female
rats based on all of the published studies, including Kim etal. f2018bl. a hierarchical Bayesian
3-44
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
analysis was conducted using either a one- or a two-compartment empirical PK model. Details of
the analysis are provided in Appendix E. Results for a one-compartment model are described here
for mice and rats and results for a two-compartment model for monkeys.
Estimation of Pharmacokinetic Parameters
In classical PK theory it is expected that once a chemical is absorbed or distributed to the
blood, its excretion (clearance) is then independent of the route of administration. With IV
administration, 100% of the dose is delivered directly to the blood, while only a fraction of an oral
dose may be absorbed. Therefore, the area-under-the-curve (AUC) for blood or serum
concentration after an oral dose should be less than or at most equal to the AUC after the same dose
administered IV, and the fraction absorbed, or bioavailability, is estimated as AUCorai/AUCiv.
However, when both the IV and oral PFHxS exposure data for rats (at identical doses) were
analyzed from Kim etal. (2016b). Kim etal. (2018b) and Huang etal. (2019a) by EPA, the estimated
serum concentration AUC was consistently lower for the IV dose data than the oral dose data for
several of the datasets, with the result that the corresponding CL values were quite different, in
some cases with non-overlapping data-set-level credible intervals (see Figure 3-5). This difference
was especially evident in the female, where CL after IV dosing was higher in all cases examined.
This outcome does not match general pharmacokinetic theory, which depends on several
assumptions, including that distribution into body tissues is independent of dose route.
3-45
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Female rat
Male rat
Kim, 2016, 4 mg/kg (iv)
Kim, 2016, 4 mg/kg (iv)
Kim, 2016, 4 mg/kg (gavage)
Kim, 2018, 1 mg/kg (iv)
Kim, 2018, 1 mg/kg (gavage)
Kim, 2018, 4 mg/kg (iv)
Kim, 2018, 10 mg/kg (gavage)
Kim, 2016, 4 mg/kg (gavage)
Kim, 2018, 10 mg/kg (iv)
Kim, 2018, 4 mg/kg (gavage)
Huang, 2019, 4 mg/kg (iv)
Huang, 2019, 4 mg/kg (gavage)
Huang, 2019, 4 mg/kg (gavage)
Huang, 2019, 4 mg/kg (iv)
50 100 150 200 250
Clearance (ml/kg/d)
2
4
6
8
Clearance (ml/kg/d)
Figure 3-5. Comparison of Female (left) and Male (right) CL values for IV and
gavage exposure of equivalent dose levels from Kim etal. (2016b). Kim et al.
(2018b) and Huang etal. (2019a). The central point, a triangle for IV and a circle
for gavage, denotes the mean CL, the thicker portion of the lines are the quartiles,
and the thinner extent of the lines denote the 95th confidence interval. Note that
these clearance values are slightly different than presented in Table 3-6, because
those values were based on an analysis of only the gavage datasets, whereas the
values in the figure above are based on analysis of the gavage and IV data together
in a hierarchical Bayesian framework.
Since data of Kim etal. f2018bl show nearly identical urinary and fecal excretion after IV
versus oral dosing, it is possible that distribution into body tissues was much greater after IV
dosing perhaps because more of the IV-infused PFHxS could distribute to various tissues before it
became bound to serum proteins, while the slower absorption from oral dosing led to lower tissue
distribution. Tissue dosimetry data after both IV and oral doses, which could be used to evaluate
this hypothesis, were not available and resolution of the apparent discrepancy was considered
beyond the scope of this analysis. Because the objective was to extrapolate dosimetry from oral
exposures in animal toxicity studies to humans, given the unusual quantitative results from
classical PK analysis for IV versus oral dosimetry, only the oral dosimetry data were included in the
final analysis for rats and mice. Only IV dosimetry data were available for monkeys, so those data
were analyzed recognizing that it may not exactly represent oral kinetics. Because the empirical
data indicated that the blood AUC after IV exposure was less than after oral exposure to the same
dose for most of the experiments, it was assumed that oral bioavailability was 100% and that was
assumed in subsequent analyses.
A single study reported PK data that could be used for parameter estimation for mice and
monkeys fSundstrom etal.. 20121. While Sundstrom etal. f20121 did collect PK data after both IV
and oral administration in mice, they did not estimate a bioavailability for male mice and the
estimate of 50% availability in female mice was based on only two animals for oral dosimetry.
Therefore, the more complete datasets for 1 and 20 mg/kg oral doses provided separately were
analyzed similarly to the analysis for rats described above, assuming 100% bioavailability. The
resulting PK model fits (see Appendix E, Figure E-9) were quite good, showing that the oral PK data
for mice were consistent with this assumption. If bioavailability was significantly lower than 100%,
3-46
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
the model (assuming 100% uptake) would have overpredicted the serum concentration time
course, but this did not occur, indicating that this is a valid assumption.
Only IV data were available for monkeys fSundstrom etal.. 20121. so those data were
analyzed for that species, recognizing the resulting uncertainty in bioavailability and that there may
be differences in distribution and clearance between the two routes of administration. While the
mouse and rat PK data were adequately fit with a one-compartment model (see Appendix E, Figures
E-6 to E-9), the monkey PK clearly showed biphasic clearance from the serum, requiring a two-
compartment model, that is, one including both central and a deep tissue compartment (see
Appendix E, Figure E-10). No critical dose-response endpoints were identified in monkey, so no
determination needed to be made considering the best approach for pharmacokinetic extrapolation
from monkeys.
Values for the volume of distribution (Vd, mL/kg) and clearance (CL, mL/kg-d) were also
estimated from the Bayesian analysis for each study and dose, as well as overall population mean
values (see Appendix E). An average half-life (T1/2) was calculated from these results using the
formula, Ti/2 = ln(2) x Vd/CL. Interestingly, while the analysis showed a clear, large sex difference
in clearance and the corresponding half-life between male and female rats, almost no difference
appeared between male and female mice. The monkey results should be interpreted with some
caution, as they were based on only three animals per sex, but they suggest an intermediate case
between rats and mice, with clearance in male monkeys being 73% of female monkeys. The much
slower clearance in male rats compared with female rats is assumed to result from higher
expression of renal transporters that resorb PFHxS. The data for mice and monkeys suggest that
expression of the transporters is much less sex-dependent in those species.
Table 3-6. Pharmacokinetic parameters for rats, mice, monkeys, and humans
Study
Dose
(mg/kg)
n
Clearance (miykg-d)3
Volume of
distribution (mL7kg)a
Tl/2b(d)
Male rats
Kim et al. (2016b)
4
5
6.71 (6.01-7.42)
279.0 (234.7-323.7)
28.9
(23.9-
33.8)
Kim et al. (2018b)
4
5
4.78 (3.14-6.42)
313.7 (298.9-327.9)
48.0
(29.3-
64.7)
4
3C
4.08 (3.43-4.81)
136 (115-155.9)
23.4
(18.6-
28.4)
16
3C
4.56 (3.86-5.31)
142.7 (119.6-164.2)
21.9
(17.2-
26.1)
3-47
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Dose
(mg/kg)
n
Clearance (miykg-d)a
Volume of
distribution (mL7kg)a
Tl/2b(d)
32
3C
4.56 (3.86-5.31)
206.2 (173.6-237.4)
19.6
(15.8-
23.3)
Population mean
-
-
5.46 (3.87-6.97)
208.2 (136.3-278.1)
27.3
(15.3-
39.2)d
Female rats
Kim et al. (2016b)
4
5
105.6 (73.64-135.1)
299.2 (271.8-326.8)
1.64
(1.57-
1.72)
1
5
72.7 (56.76-88.15)
198.3 (182.9-214.5)
1.62
(1.54-
1.68)
4
5
96.41 (71.5-119.9)
240.6 (221-259.1)
1.48
(1.41-
1.54)
Huang et al. (2019a)
4
3C
42.92 (34.56-51.64)
170.7 (149.8-190.7)
2.36
(2.2-2.5)
16
3C
53.44 (42.12-64.93)
190.8 (169.9-212.7)
2.14
(2.01-
2.25)
32
3C
82.86 (60.98-103.62)
254.8(224-284.3)
1.84
(1.71-
1.97)
Population mean
-
-
98.21 (68-125.7)
222.6 (177.8-263.9)
1.86
(1.22-
2.5)d
Male mice
Sundstrom et al. (2012) (all data)
1 & 20
4C
3.81 (3.54-4.08)
150.6 (136.1-164.7)
27.4
(25.8-
29)
Female mice
Sundstrom et al. (2012) (all data)
1 & 20
4C
3.14 (2.98-3.29)
120.7 (112.8-128.9)
26.7
(25.6-
27.7)d
Male monkeys
Sundstrom et al. (2012)
10
3
1.17 (0.72-1.64)
272.0 (239.3-303.2)e
177.6
(95.5-
250.6)
Female monkeys
Sundstrom et al. (2012)
10
3
2.1 (1.78-2.42)
222.9 (197.9-249)6
74.4
(59.1-
88.9)
3-48
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Dose
(mg/kg)
n
Clearance (miykg-d)a
Volume of
distribution (mL7kg)a
Tl/2b(d)
Human
All males and females below age 12.4 y and
above age 50 y
577
0.041
222.9 (women)'
272.0 (men)'
3,768
(10.3 y)
4,598
(12.6 y)
Women 12.4-50 yr of age
0.072
222.9 (women)'
2,146
(5.9)
aValues are mean (study-level 90% credible interval) or population mean (90% credible interval).
bTi/2 = ([mean] volume of distribution [mL/kg]) x In (2) / ([mean] clearance [mL/kg-d]).
cNumber of animals per time point.
dRats displayed a large difference in half-life between sexes that mice did not. This sex-dependence was seen in
rats for many PFAS and has been linked to sex-hormone dependent changes in renal transporters (Kudo et al.,
2002). It is not fully understood why this phenomenon is different between species.
eSum of central and peripheral compartment volumes from a two-compartment PK model.
fVd in women assumed equal to the value for female monkeys, Vd in men assumed equal to male monkeys.
While the results for rats showed a fair degree of variability in CL between studies (see
Table 3-6), the range in mean values is 1.8-fold for males and 2.3-fold for females is modest and the
overall population means were obtained via a Bayesian analysis that addressed the variability both
within and among the datasets (see details in Appendix E, Section 1). Hence, these values provided
an estimate of the relationship between dose and mean serum concentration levels in rats that
appeared to be accurate to within a factor of two, which was set as an acceptable degree of
discrepancy between PK model simulations and data in EPA's Umbrella Quality Assurance Project
Plan (QAPP) for Dosimetry and Mechanism-Based Models (U.S. EPA. 2018b). and so were
considered sufficiently sound for use in cross-species extrapolation.
The assumption that the Vd derived from monkeys is a suitable surrogate for the human Vd
introduces some uncertainty to the calculated human half-life. However, Chiu etal. (2022) obtained
a mean (95% CI) Vd of 0.25 (0.15, 0.42) L/kg from their analysis of human data, which is essentially
the average of the values from male and female monkeys, 0.287 and 0.213 L/kg, respectively.
Hence, the extent of the uncertainty is judged to be minimal. Use of the value from Chiu etal. (2022)
would only change some of the estimated clearance values in Table 3-5 by less than 20%, so would
have a minimal impact on the geometric mean clearance obtained.
Clearance Versus Glomerular Filtration Rate and Free Fraction in Serum
Some mechanistic insight could be gained by comparing the clearance values shown in
Table 3-6 with species-specific glomerular filtration rate (GFR), with and without adjustment for
serum protein binding. Davies and Morris (1993) summarized GFR for multiple species. Using
0.25 kg as the species average BW for the rat, the GFR/BW for rats is 7.55 L/kg-d, which is
approximately 1,400 and 90 times higher than the population mean clearance estimated in male
and female rats, respectively.
3-49
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Binding to serum proteins plays a likely role in these very large differences. As discussed
above in the context of distribution, PFHxS binds to albumin with high affinity and it is the major
carrier of PFHxS in blood fWeiss etal.. 2009: Forsthuber etal.. 2020: Bischel etal.. 20101. This
binding may play a role in the limiting the rate of the renal excretion of PFHxS, in addition to the
role played by renal transporters. Kim etal. f2018bl measured reported PFHxS free fractions (/free)
of 0.00076 and 0.00069 in male and female rat plasma. Using these values, GFRx/free = 5.7 and
5.2 mL/kg-d in male and female rats. This alternative estimate of clearance for male rats is close to
the population mean in Table 3-6 (4.79 mL/kg-d), which could be interpreted as showing minimal
renal resorption in males, with a population mean clearance of 5.46 mL/kg-d. However, for female
rats GFRx/free is more than order of magnitude lower than the population mean clearance of
85.3 mL/kg-d. Section 3.1.5 provided further discussion ofthe fact that the PBPK model of Kim etal.
f2018bl. which assumed that tissue distribution was similarly limited by the free fraction,
underpredicted the observed short-term distribution of PFHxS in rats. Hence, while it is expected
that serum protein binding limits renal excretion (and tissue distribution) to some extent, the
reduction appears to be less than predicted by assuming that clearance is strictly limited to the
equilibrium free fraction. As noted above, Robinson and Rapoport T19861 used a mathematical
model that incorporates the kinetics of plasma binding and release to describe uptake of drugs by
the brain, supporting this conclusion. Alternatively, there could be an error in the measured free
fraction.
More qualitatively, the fact that the measured free fraction is similar in male versus female
rats indicates that it cannot explain the large sex difference in empirical clearance, and hence that
sex differences in renal resorption are likely to be a factor.
3.1.7. Model Evaluation Conclusion and Extrapolation Approach
The clearance in rats is sufficiently slow that PFHxS is expected to accumulate throughout
the course of the 28-day NTP bioassay (NTP. 2019) in male rats and for about 10 days in female
rats, as illustrated in Appendix E, Section 2. For this reason, the preferred approach would be to
perform an interspecies dose extrapolation that accounts for the time dependence of the internal
dose (i.e., bioaccumulation). Further, given the slow clearance of PFHxS in male rats, the growth of
rats during these toxicity studies could be a significant factor as increases in BW are expected dilute
the body burden from earlier exposures. Therefore, a computational model for a two-compartment
PK model was developed to describe the accumulation and elimination of PFHxS during these
experiments, with time-dependence in BW based on the empirical data for BW. Details of the model
and its evaluation against serum concentration data from NTP (2019) were provided in Appendix E,
Section 2. The Bayesian parameter estimation selected a one-compartment model for male rats and
a two-compartment model for female rats. (With distribution to the second or "deep" compartment
set to zero, the two-compartment model code simulates a one-compartment model.) While the
period of accumulation was much longer for male rats, female rats were modeled in the same way
as males for consistency. However, comparison of the PK model predictions to the plasma
3-50
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
concentrations measured at the of the NTP bioassay revealed that this simple approach was not
suitable for PFHxS due to an observed nonlinear relationship between dose and plasma
concentration, which the PK model was not able to replicate.
As noted in the Summary of Human PFHxS Excretion section, uncertainties also exist in the
potential extrapolation of such a model to developmental or other early-lifestage effects. Even
though the results for the PK model indicated that the model may be adequate for low-dose
extrapolation of dosimetry in adult animals, the failure of the model to fit the higher dose data (see
Appendix E, Section 2) and the issues identified with the published PBPK models (see Section 3.1.5)
demonstrated an incomplete understanding of PFHxS pharmacokinetics. Additional research, which
may be extensive, is needed to resolve the existing inconsistencies between the various models and
the data. Thus, a reliable PK model for PFHxS is not considered to be in the realm of available
science.
As described in Appendix E.2, for male rats in the NTP bioassay the results of the PK model
simulations provide qualitative support for an interpolation of the measured end-of-study
concentrations to estimate the end-of-study concentration at a POD dose, with the average
concentration during the bioassay then being approximated as one-half of that estimated final
concentration. (The discrepancy between the observed end-of-study concentration in male rats and
the calculated steady-state concentration shown in Appendix E, Figure E-12 show that assuming
steady state would significantly over-estimate the internal dose.) The average plasma concentration
estimated by interpolation can then be converted to an HED by assuming steady-state levels in
humans, using the estimated human clearance.
For female rats in the NTP bioassay model simulations may also be adequate at the lowest
two doses but significantly overpredictthe data at higher concentrations (see Appendix E, Figure E-
12). While an assumption of steady state may be likewise acceptable at the lower doses, it would
also overpredictthe observations at the higher doses. Hence, a similar interpolation of the observed
data could be used for female rats in the 28-day bioassay. However, measured concentrations are
not available for the multigenerational study Ramhaj etal. (2018) that could be used to evaluate a
PK model simulation of the study or estimate internal doses by interpolation. Therefore, an
alternate approach, not involving the PK model, is needed for those endpoints, described below.
Approach for Animal-Human Extrapolation of PFHxS Dosimetry
After evaluation of three published PBPK models and a two-compartment PK model for
PFHxS, it was determined that none of these options could reliably predict PFHxS dosimetry. For
observations in male rats in the NTP bioassay, the internal dose at the POD can be estimated as
described briefly just above and in Appendix E.2. An alternative to use of PK (or PBPK) models for
dosimetric extrapolation is use of data-derived extrapolation factors (DDEFs). As stated in EPA's
guidance for DDEFs fU.S. EPA. 2014b! use of these factors "maximize the use of available data and
improve the scientific support for a risk assessment" As discussed above in the Evaluation of
Pharmacokinetic Modeling and Summary of Human PFHxS Excretion sections, the estimated
3-51
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
population-average values of total CL for male and female rats and for humans were considered
sufficiently sound for use in such extrapolation, while use of BW3/4 scaling (the least preferred
option; see U.S. EPA f2011bll could lead to overprediction of HEDs by as much as three orders of
magnitude. Therefore, a DDEF calculated from the clearance values listed in Table 3-5 and Table 3-
6, was used as the next preferred option for extrapolation of the effects observed in the
multigeneration study. Specifically, the ratio of human clearance to clearance in the animal species
and sex in which a given POD was identified was used to estimate the HED for that POD. Specifically,
to extrapolate from a POD from the Ramh0i etal. (20181 ratbioassay to humans,
HED = POD x CLh/CLrat,f, (3-3)
where CLh is the clearance in humans for the appropriate population, CLrat,fm is the clearance in
female rats, whose dosing led to the in utero and lactational exposure of the Fi pups and
CL,H/CLrat,fis the DDEF. This calculation assumed the fraction absorbed or bioavailability in human
and rats, which is taken to be 100% as described in Section 3.1. In particular, the computational PK
analysis summarized in Section 3.1.6 found that the published PK data showed serum AUC after
oral exposures were higher than serum AUCs after matching IV exposures for several key studies
rather than results consistent with less than 100% oral bioavailability.
Effects observed on or before PND 7 are assumed to be the result of gestational exposure,
and the clearance in the female animal (dam) would be assumed to determine dosimetry to the
fetus and young pups. However, if effects observed in rat pups after PND 7, the clearance for the
same sex adult rat should be used since that clearance determines their internal dose. For results in
combined pups (both sexes) after PND 7, the higher clearance of female rats is used to be health
protective.
While menstruation does not occur during pregnancy and may not resume until after
weaning of the child, as described in the subsections Trend in Pregnancy and Breast Milk in 3.1.2
Distribution, studies of longitudinal changes during and after pregnancy show maternal serum
concentrations remaining fairly constant or declining through this lifestage. This likely occurs
because the long half-life of PFHxS results in slow accumulation as well as elimination, while the
increase in total body mass during pregnancy (including the fetus and placenta) results in a dilution
of the body burden as the PFHxS distributes into those growing tissues. Therefore, the serum levels
in the pregnant and postpartum woman are expected to be consistent with her serum levels at the
start of pregnancy, which are determined by her total clearance prior to pregnancy, including that
associated with menstrual fluid loss. Thus, HEDs for developmental endpoints that occur in utero
such as reduced birthweight or are based on measures of maternal serum concentration will be
calculated using the higher clearance estimated for women of childbearing age (12.4-50 years) in
Table 3-6.
However, this additional clearance clearly does not occur in young children, and as
described in Summary of Human PFHxS Elimination in Section 3.1.4, there may be differences in PK
3-52
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
among human lifestages that cannot be quantified because of a lack of empirical PK data during
childhood. While effects in adults do not involve extrapolation across lifestages, the degree of
accumulation of PFHxS in rats during a 28-day bioassay could be less than the accumulation during
a comparable portion (4%) of the human life span. Therefore, HEDs for effects observed in
experimental animals more than a few days after birth, where dosimetry in the pups or human
child may be a significant factor, and for immune effects correlated with serum concentrations
measured 5 years after birth, for which the exposure and clearance of the offspring are significant
factors, have been calculated using the population-average CLh from Table 3-6.
The key assumption made in calculating a DDEF for a given endpoint evaluated was that for
effects observed in adult male and female rats, the CL and Fabs for the corresponding rat sex from
Table 3-6 were used to calculate the DDEF. Table 3-7 shows the resulting DDEFs.
Table 3-7. Data-derived extrapolation factor (DDEF) calculations
Sex and species of observation (lifestage)
CLA (mL/kg-d)
DDEFa
Male rats (adult and male pups >PND 7)
5.46
7.51 x 10"3
Female rats (adult and female pups >PND 7),
nonreproductive/developmental effects
85.3
4.81 x 10"4
Female rats (adult), reproductive effects and effects in pups
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
extrapolated from the NTP rat bioassay by interpolation of the observed end-of-study, the
estimated internal doses are judged to have similar if not better accuracy (i.e., within 30% of the
actual values) and should notunderpredictthe internal doses by more than 1.3-fold. Given that
female rats, in particular dams in the Ramhai etal. f20181 multigenerational study, are expected to
be within a few percent of steady state by 7 days after dosing (see Appendix E, Figure E-12), the
implicit assumption of steady state for those endpoints should also provide fairly accurate
predictions and the analysis is unlikely to underpredict internal doses since male rat pups will have
lower clearance than females, hence higher internal doses. The nonmenstrual clearance value used
for humans was approximately twofold higher than the lowest value reported by or estimated from
multiple studies of PFHxS dosimetry in humans. Only a modest correction for fecal absorption
(using the ratio of fecal/urinary elimination observed in rats after IV dosing) was applied. Hence,
the average human clearance is unlikely to be more than twofold lower than the value used for HED
calculation.
The adjustment made for menstrual-associated clearance was based on PFHxS serum levels
found in the U.S. population of nulliparous women (NHANES data) and the difference estimated
between men and women matches closely the estimated average rate of menstrual fluid loss,
providing high confidence in this value. Further, because the lower clearance estimated for men and
nonreproductive-aged women is used to calculate HEDs for nondevelopmental endpoints and older
children, the assumption of higher clearance in reproductive-age women has no impact on the
estimated risk for those endpoints and lifestages. The only potential for underestimation of risk due
to the adjustment for menstrual-associated clearance is for perinatal developmental endpoints such
as birth weight. EPA recognizes that the resulting clearance for women of childbearing age may not
be protective for the children of women who do not menstruate regularly before becoming
pregnant (presuming that menstruation is in fact the mechanism causing the observed difference in
plasma concentrations). However, the PFHxS plasma concentrations in such women should be no
greater than those observed in their male counterparts (see Figure 3-3), i.e., no more than twofold
higher than the estimated mean for women of reproductive age. Therefore, while there is
recognized variability in physiological factors, specifically menstruation, that likely result in
corresponding variability in PFHxS clearance among reproductive-age women, this variability is
captured by application of the standard human interindividual uncertainty factor (UFh = 10), of
which a factor of 3 is attributed to pharmacokinetic differences across individuals (UFh.pk = 3).
Likewise, while there are uncertainties in the dosimetric extrapolation to developmental exposure
and dosimetry in children remain, there are currently no data to indicate that these are greater than
is accounted for by application of UFh,pk-
3.2. NONCANCER HEALTH EFFECTS
For each potential health effect discussed below, the synthesis describes the evidence base
of available studies. Arrays or tables summarizing endpoint results across studies within each
3-54
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
evidence stream are also provided. The effect levels presented in these arrays and tables are based
on statistical significance7 or biological significance, or both. Examples relevant to interpretations
of biological significance include consideration of the directionality of effect (e.g., statistically
significantly decreased cholesterol/triglycerides is of unclear toxicological relevance), tissue-
specific magnitude of effect (e.g., statistically nonsignificant increase of >10% in liver weight may
be considered biologically significant), and dose-dependence (e.g., a significant finding at a single,
lower dose level but not at multiple, higher dose levels may be interpreted as potentially spurious).
For this section, evidence to inform organ-/system-specific effects of PFHxS in animals following
developmental exposure is discussed in the individual organ-/system-specific sections (e.g., liver
effects after developmental exposure are discussed in the hepatic effects section and so on,
although they are generally cross-referenced to the Developmental Effects section; Section 3.2.3).
Evidence on other developmental effects (e.g., fetal growth) is only discussed in the Developmental
Effects section. Lastly, overt toxicity was not observed at any of the highest doses tested in any of
the available studies (in contrast to data available for some of the other PFAS being assessed by the
IRIS Program), and thus the potential for overt toxicity to complicate interpretation of the health
effect-specific PFHxS evidence is not a factor discussed in any of the following sections.
3.2.1. Thyroid Effects
Under normal physiologic conditions, the hypothalamic-pituitary-thyroid (HPT) axis, a
hormone regulatory system that controls the levels of thyroid hormones in the body, stimulates
neurons in the hypothalamus to release thyrotropin-releasing hormone (TRH) to stimulate
epithelial cells of the anterior pituitary gland to release thyroid stimulating hormone (TSH)
flrizarrv. 20141. The role of TSH is to stimulate the thyroid gland to release thyroxine (T4), which is
converted to triiodothyronine (T3). When increased T3 and T4 serum levels exceed a blood
concentration threshold, secretion of TRH from the hypothalamus is inhibited via a negative
feedback loop (Pilo etal.. 1990: Irizarrv. 2014). In adults, T3 and T4 play important metabolic
functions; for example, decreases in T3 and T4 serum levels in the presence of an intact HPT axis
results in a condition known as hypothyroidism, result in increased weight gain, fatigue, and dry
skin, as well as effects on the memory and a difficulty to concentrate. Conversely, increased levels of
T3 and T4 in the presence of an intact HPT axis results in a condition known as hyperthyroidism,
resulting in increased rate of metabolism, weight loss, and increased heart rate (Mullur etal.. 2014).
During fetal development and throughout early childhood, thyroid hormones play an important
role in somatic growth and development Thyroid hormones play an important role in immune
system functions (U.S. EPA. 2006) and as discussed in Section 3.2.2, low levels of thyroid hormones
during gestation are associated with altered immune system development and functions (Rivera et
al.. 2024: Funes etal.. 2022). Thyroid hormones have also been shown to play a critical role in
throughout the assessment, the phrase "statistical significance" indicates a p-value < 0.05, unless otherwise
noted.
3-55
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
neurogenesis, neuronal migration, and synaptogenesis, as well as shifting neuronal cells from a
proliferative state to a differentiation state and myelination (Gilbert etal.. 20161. In humans,
alterations of prenatal maternal T4 have been linked to declines in cognitive function in children
fKorevaar etal.. 2016: Haddow etal.. 19991. Importantly, changes in prenatal and maternal T4 have
been shown to be biologically important in the absence of changes in TSH reviewed in fZoeller and
Rovet. 2004: Vansell. 2022: Stagnaro-Green and Rovet. 2016: Rovet. 2005. 2014: Patel etal.. 2011:
Navarro et al.. 2014: Morreale de Escobar et al.. 2008: Moogetal.. 2017: Hood etal.. 1999a: Hood et
al.. 1999b: Hood and Klaassen. 2000: Dong etal.. 2015: Cuevas etal.. 2005: Berbel etal.. 20101.
Human Studies
Thirty-nine studies (reported in 44 publications) have investigated the relationship
between PFHxS exposure and thyroid hormones and/or thyroid disease in humans. All of the
available human studies examined the association between PFHxS exposure measured in blood and
thyroid hormones.
Multiple outcome-specific considerations were influential on the study evaluations. First,
for outcome ascertainment, collection of blood during a fasting state and at the same time of day for
all participants (or adjustment for time of collection) is preferred for measurement of thyroid
hormones to avoid misclassification due to diurnal variation fvan Kerkhofetal.. 20151. Studies that
did not consider these factors (e.g., by study design or adjustment) were not excluded but were
considered deficient for the outcome ascertainment domain, primarily thyroid stimulating
hormone (TSH), which is more impacted by these issues than thyroxine (T4) or triiodothyronine
(T3). However, this was not expected to result in substantial bias, and thus studies were not
downgraded in overall study confidence if lack of fasting, and consideration of diurnal variation
were the primary limitations identified. This possible outcome misclassification was expected to be
nondifferential and thus likely a bias toward the null; the domain ratings were used to assess
possible sources of inconsistency in the results. For participant selection, it was considered
important to account for current thyroid disease and/or use of thyroid medications; studies that
did not consider these factors by exclusion or another method were considered deficient for the
participant selection domain. Concurrent measurement of exposure with the outcome was
considered acceptable for this outcome since thyroid hormones can be up- or downregulated
relatively quickly in relation to the long half-life of PFHxS (half-life of T3 and T4 are in the order of
hours/days, respectively fLeboffetal.. 19821 versus years for PFHxS fLi etal.. 20181: see Section
3.1.3); thus, exposure measurement ratings were not downgraded for timing of measurement. All of
the available studies analyzed PFHxS in serum or plasma using appropriate methods as described
in the systematic review protocol (see Appendix A). Thyroid hormones were analyzed using
standard methods (e.g., immunoassays, HPLC-MS/MS) in all studies. The medium confidence
studies generally were not downgraded for participant selection, but most did not account for time
of day of blood collection and fasting which is considered likely to result in nondifferential outcome
3-56
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
misclassification (expected to be toward the null on average) for thyroid hormone measures. The
low confidence studies were generally downgraded for both the participant selection issues and
outcome ascertainment issues described above, though Liu etal. f20181 did not account for thyroid
medication use but was unique in the set of available studies in that data were collected
prospectively and the analysis was based on change in outcome, so there was less concern for the
lack of adjustment impacting the results.
In summary, 26 studies were medium confidence (Yang etal.. 2016a: Wen etal.. 2013:
Webster et al.. 2014: Wang etal.. 2013: Wang etal.. 2014a: Shah-Kulkarni et al.. 2016: Sarzo etal..
2021: Reardon et al.. 2019: Preston etal.. 2018: Liu etal.. 2018: Liang etal.. 2020: Li etal.. 2021b:
Lebeaux et al.. 2 0 2 0: Kim etal.. 2020a: Kang etal.. 2018: Inoue etal.. 2019: Gallo etal.. 2022: Dufour
etal.. 2018: Crawford etal.. 2017: Caron-Beaudoin etal.. 2019: Cakmak et al.. 2 0 2 2: Blake etal..
2018: Berg etal.. 2017: Aimuzi etal.. 2019: Aimuzi etal.. 20201 and 10 were low confidence (Zhang
etal.. 2018b: Liu etal.. 2021b: Li etal.. 2017c: Lewis etal.. 2015: Khalil etal.. 2018: Ti etal.. 2012:
Itoh etal.. 2019: Heffernan etal.. 2018: Chan etal.. 2011: Bloom etal.. 20101. Three studies were
uninformative in study evaluation (Seo etal.. 2018: Kim etal.. 2011a: Kim etal.. 2016a). Sensitivity
was a concern across studies due to narrow exposure contrasts in several studies (see sensitivity
domain in Figure 3-6), combined with the expected bias toward the null due to outcome
misclassification. Thus, null results are difficult to interpret. The medium confidence studies were
the focus of evidence synthesis; low confidence studies did not undergo data extraction but were
still considered for consistency in the direction of association. The domain ratings, populations, and
thyroid measures for each study are presented in Figure 3-6.
3-57
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Aimuzi, 2019, 5387078-
Aimuzi, 2020.6512125
Berg V, 2016, 3350759-
Blake, 2018, 5080657-
Bloom. 2010, 757875-
Cakmak. 2022, 10273369-
Caron-Beaudoin, 2019, 5097914-
Chan, 2011, 1402500-
Crawford, 2017, 3859813-
Dufour, 2018, 4354164
Gallo, 2022, 9962235-
Guo, 2021, 7410165-
Heffernan, 2018, 5079713-
Inoue, 2019, 5918599-
lloh. 2019. 5915990
Ji. 2012, 2919189-
Kang, 2018. 4937567
Khalil. 2018, 4238547
Kim, 2011, 1424975-
Kim. 2016, 3351917
Kim. 2020, 6833758
Labaaux, 2020, 6356361 -
Lewis, 2015, 3749030-
Li, 2017, 3856460
Li. 2021, 7277672
Liang, 2020,7161554-
Liu, 2018, 4238396-
Liu. 2021 10176563-
Preston, 2018. 4241056-
Reardon, 2019, 5412435-
Sarzo, 2021, 9959596
Seo, 2018, 4238334
Shah-Kulkarni, 2016, 3859821
Wang, 2013,4241230-
Wang, 2014, 2850394-
Webster, 2014, 2850208-
Wen, 2013, 2850943-
Yang, 2016, 3858535-
Zhang, 2018, 5079665-
* I Adequate {metric) or Medium confidence (overall)
- Deficient (metric) or Low confidence (overall)
^9 Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-6. Study evaluation results for epidemiology studies of PFHxS and
thyroid effects. Full details available by clicking HAWC link. Multiple
publications of the same study: Preston etal. f2018) also includes Preston et ai.
f2020I
3-58
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
The results for the association between PFHxS exposure and thyroid effects in medium
confidence studies are presented in Tables 3-8 and 3-9. Twenty-eight studies examined
associations with thyroid hormones in adults, including 13 focused on pregnant women (see Table
3-8). For T4, out of 27 studies, the results are mixed. In the 15 medium confidence studies, a few
statistically significant associations were reported (positive associations in both sexes in Cakmak et
al. (2022). positive association in women but inverse in men in Wen etal. (2013). positive
association in men >50 years of age in Li etal. (2021b). positive association in pregnant women in
Aimuzi etal. (2020). and inverse association in pregnant women in Reardonetal. (2019)). Other
nonsignificant results were also in both directions or they showed no association. The low
confidence studies were also inconsistent in direction of association for T4. Many of the inverse
associations had small magnitudes of effect and some estimates, particularly for total T4, were
imprecise (i.e., had wide confidence intervals), both of which decrease certainty in the evidence.
There is no clear pattern by exposure level or population. Nineteen studies examined associations
with T3. In the 12 medium confidence studies, most reported no association except for three studies
(Wen etal.. 2013: Crawford etal.. 2017: Aimuzi etal.. 2020) in women that reported higher levels of
T3 with higher exposure to PFHxS (statistically significant in latter two studies). Twenty-seven
studies reported on TSH, and of the 16 medium confidence studies, one reported statistically
significant higher TSH with higher exposure fReardon etal.. 20191 and one study reported a
statistically significant inverse association (Aimuzi etal.. 2020). both in pregnant women, but the
remaining studies reported no clear association.
Table 3-8. Associations between PFHxS exposure and thyroid hormone levels
in medium confidence studies of adults
Reference
Population
Median
exposure (IQR)
or as specified
(ng/mL)
Effect estimate
T4
T3
TSH
General population, adults
Cakmak et al.
(2022)
CHMS cross-
sectional study
(2007-2011),
Canada, 6,045
participants (all
ages)
1.5 (GM)
Percent
change for GM
equivalent
increase
Total T4
0.9 (0.1,1.8)*
NR
-1.1 (-4.9, 2.9)
Crawford et
al. (2017)
Time to Conceive
cross-sectional
study (2008-2009),
U.S., 99 women
1.6 (GM)
P (p-value) for
log-unit
increase
Total T4
-0.15 (0.5)
Free T4
0.01 (0.8)
Total T3
2.8(0.2)
-0.03 (0.7)
Wen et al.
(2013)
NHANES cross-
sectional study
(2007-2010), U.S.,
2.0 (GM)
P (95% CI) for
In-unit
increase
Total T4
Women
Total T3
Women
Women
3-59
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Median
exposure (IQR)
or as specified
Reference
Population
(ng/mL)
Effect estimate
T4
T3
TSH
1,181 adults (672
0.26 (0.11,
4.07 (2.23,
-0.02 (-0.13,
men, 509 women)
0.41)*
5.92)*
0.09)
Men
Men
Men
-0.03 (-0.18,
-0.08 (-1.70,
0.02 (-0.06,
0.11)
1.54)
0.52)
Free T4
Free T3
Women
Women
0.003 (-0.02,
0.003 (-0.02,
0.03)
0.03)
Men
Men
-0.02 (-0.03,
0.005 (-0.003,
-0.003)*
0.01)
Blake et al.
Fernald Community
2.7 (1.7-4.1)
Percent
Total T4
NR
1.97 (-7.73,
(2018)
Cohort (1990-
change for IQR
1.74 (-1.73,
12.7)
2008), U.S., 210
increase
5.33)
adults (81 men, 129
women)
Liu et al.
POUNDS Lost trial of
3.1(2.3-4.4)
Spearman
0-6 mo
0-6 mo
NR
(2018)
weight loss
correlation
0.04
0.01
treatment (2004-
coefficients for
6-24 mo
6-24 mo
2007) 621 adults
change in
-0.02
-0.05
(237 men, 384
hormone
women)
Gallo et al.
Veneto cross-
6.5 (3-12)
Percent
NR
NR
Women
(2022)
sectional study in
change for IQR
1.1 (-1.8,4)
high exposure area
increase
Men
(2017), Italy, 14,888
-5.5 (-11, 0.3)
adults
Li et al.
Ronneby cross-
93 in women
Percent
Free T4
Free T3
(2021b)
sectional study in
aged 20-50 yr
change
Women 20-
Women 20-
Women 20-
high exposure area
50 yr
50 yr
50 yr
(2014-2015),
0.43 (-0.08,
0.08 (-0.41,
-0.47 (-2.52,
Sweden, 2,687
0.94)
0.57)
1.62)
participants (all
Women >50 yr
Women >50 yr
Women >50 yr
ages)
0.01 (-0.57,0.6)
0.05 (-0.47,
0.63 (-1.88, 3.2)
Men 20-50 yr
0.57)
Men 20-50 yr
0.51 (-0.14,
Men 20-50 yr
-0.37 (-2.7,
1.16)
0.29 (-0.29,
2.01)
Men >50 yr
0.88)
Men >50 yr
0.73 (0.02,
Men >50 yr
-0.14 (-2.79,
1.45)*
0.26 (-0.36,
2.58)
0.89)
Pregnant women
3-60
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median
exposure (IQR)
or as specified
(ng/mL)
Effect estimate
T4
T3
TSH
Yang et al.
(2016b)
Beijing Prenatal
Exposure cross-
sectional study
(2013) 157 mother-
infant pairs
0.5
Spearman
correlation
coefficients
Total T4: 0.08
Free T4: 0.04
Total T3: 0.08
Free T3: 0.12
-0.15
Wang et al.
(2013)
Cross-sectional
analysis within
Norwegian Mother
and Child Cohort
Study (2003-2004),
Norway, 903
pregnant women
0.6 (0.4-0.8)
P (95% CI) for
In-unit
increase
NR
NR
0.01 (-0.04,
0.07)
Aimuzi et al.
(2020)
Cross-sectional
analysis within
Shanghai Birth
Cohort (2013-
2016), China, 1,885
pregnant women
0.6 (0.4-0.7)
P (95% CI) for
In-unit
increase
Free T4
0.12 (0.02,
0.22)*
Free T3
0.2 (0.05,
0.34)*
-0.12 (-0.22,
-0.01)*
Sarzo et al.
(2021)
Cross-sectional
analysis within
INMA (2003-2008),
Spain, 919 pregnant
women
0.6 (0.4-0.9)
Percent
change for
doubling (95%
CI)
Free T4
-1.6 (-7.56,
4.75)
Total T3
0.52 (-6.05,
7.54)
6.09 (-0.71,
13.4)
Wang et al.
(2014a)
Taiwan Maternal
and Infant Cohort
Study (2000-2001),
Taiwan, 285
pregnant women
and 116 neonates
0.8 (0.3-1.4)
P (95% CI) for
unit increase
Total T4
-0.13 (-0.32,
0.06)
Free T4
-0.01 (-0.02,
0.003)
Total T3
-0.002 (-0.01,
0.001)
0.11 (-0.002,
0.21)
Webster et
al. (2014)
CHirP cohort (2007-
2008), Canada, 152
women
1.0 (0.7-1.7)
P (95% CI) for
IQR increase
Free T4
-0.02 (-0.1,
0.07)
NR
0.01 (-0.05,
0.07)
Reardon et
al. (2019)
Alberta Pregnancy
Outcomes and
Nutrition cohort
(2009-2012), 494
women
1.0
P (95% CI) for
unit increase
Free T4
-0.01 (-0.01,
-0.001)*
Free T3
Not significant
0.14 (0.04,
0.25)*
Inoue et al.
(2019)
Cross-sectional
analysis within
Danish National
Birth Cohort (1996—
2002), Denmark,
1.1(0.8-1.4)
Absolute
Percent
difference
(95% CI) per
IQR increase
Free T4
-0.3 (-1.6,1)
NR
1.7 (-4.4, 8.1)
3-61
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median
exposure (IQR)
or as specified
(ng/mL)
Effect estimate
T4
T3
TSH
1,366 pregnant
women
Lebeaux et
Health Outcome
and Measures of the
Environment cohort
(2003-2006), 355
mother-infant pairs
1.6 (1.5)
P (95% CI) for
doubling
Total T4
-0.01 (-0.04,
0.02)
Free T4
0.02 (-0.01,
0.05)
Total T3
-0.01 (-0.04,
0.02)
Free T3
-0.02 (-0.04,
0)
-0.06 (-0.23,
0.11)
al. (2020)
Preston et al.
Project Viva cohort
(1999-2002), U.S.,
732 pregnant
women and 480
neonates
2.4(1.6-3.8)
P (95% CI) for
IQR increase
Total T4
-0.05 (-0.14,
0.04)
Free T4
-0.60 (-1.39,
0.19)
NR
2.89 (-2.12,
8.17)
(2018)
*p < 0.05.
GM = geometric mean.
One medium confidence study (Berg et al., 2017) is not included because quantitative results were only reported
for significant associations.
Six studies examined associations with thyroid hormones in children and/or adolescents, in
addition to studies of adults that included adolescents or all ages without stratifying results, which
were described above. All six studies (five medium confidence and one low confidence) reported
null associations between PFHxS exposure and thyroid hormones (Li etal.. 2021b: Kim etal..
2020a: Khalil etal.. 2018: Kangetal.. 2018: Gallo etal.. 2022: Caron-Beaudoin etal.. 20191
Eleven studies (9 medium confidence) examined associations with thyroid hormones in
infants. For T4,10 studies were available, including 9 of medium confidence. One study with the
highest exposure levels fPreston etal.. 20181 reported statistically significant lower levels of total
T4, driven by the association in boys, with an exposure-response gradient across quartiles. The
remaining studies reported no association. Nine studies examined associations with T3. One low
confidence study (Shah-Kulkarni etal.. 20161 reported statistically significant higher levels of T3
with higher PFHxS exposure in girls and no association in boys, while Aimuzi et al. (20191 reported
statistically significant inverse associations, strongest in boys. The remaining studies reported no
association. Ten studies examined the association between TSH and PFHxS exposure. There were
lower levels of TSH with higher PFHxS exposure in one low confidence study f Shah-Kulkarni et al..
2016). and higher levels of TSH in one study (Wang et al.. 2 014a) though neither was statistically
significant, and the confidence intervals were wide. The remaining studies reported no association.
3-62
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-9. Associations between PFHxS exposure and thyroid hormone levels
in medium confidence studies of infants
Reference
Population
Median
exposure
(IQR) or as
specified
(ng/mL)
Effect
estimate
T4
T3
TSH
Guo et al.
(2021)
Sheyang Mini Birth
Cohort Study (2009-
2010), China, 490
infants
0.1(0.1-
0.1)
P (95% CI)
for In-unit
increase
Total T4
0.04 (-0.006,
0.09)
Free T4
0.02 (-0.007,
0.05)
Total T3
0.04 (-0.003,
0.09)
Free T3
0.02 (-0.02, 0.05)
-0.10 (-0.23,
0.03)
Dufour et al.
(2018)
University Hospital of
Liege cohort (2013-
2016) 214 mother-
infant pairs
0.2
P (p-value)
for
detected
vs. not
detected
NR
NR
(0.9)
Girls
0.09 (0.5)
Boys
-0.06 (0.5)
Aimuzi et al.
(2019)
Cross-sectional
analysis from
Shanghai Obesity and
Allergy Cohort Study
(2012-2013), 568
infants
0.2 (0.1-
0.3)
P (95% CI)
for In-unit
increase
Free T4
0.06 (-0.06,
0.18)
Girls
0.03 (-0.14, 0.2)
Boys
0.1 (-0.07,0.26)
Free T3
-0.04 (-0.09,
-0.001)*
Girls
-0.08 (-0.14,
-0.02)*
Boys
-0.02 (-0.16,
-0.03)*
-0.03 (-0.06,
0.004)
Girls
-0.02 (-0.07,
0.02)
Boys
-0.04 (-0.08,
0.01)
Yang et al.
(2016b)
Beijing Prenatal
Exposure cross-
sectional study
(2013) 157 mother-
infant pairs
0.5
Spearman
correlation
coefficient
s
Total T4: -0.005
Free T4: 0.01
Total T3: -0.07
Free T3: -0.03
0.08
Wang et al.
(2014a)
Taiwan Maternal and
Infant Cohort Study
(2000-2001), Taiwan,
116 infants
0.8 (0.3-
1.4)
P (95% CI)
for unit
increase
Total T4
0.002 (-0.50,
0.50)
Free T4
-0.03 (-0.10,
0.04)
Total T3
-0.001 (-0.007,
0.004)
0.49 (-1.45,
2.43)
Lebeaux et
al. (2020)
Health Outcome and
Measures of the
Environment cohort
(2003-2006), 355
mother-infant pairs
1.6 (1.5)
P (95% CI)
for
doubling
Total T4
0.02 (-0.01,
0.06)
Free T4
-0.01 (-0.04,
0.02)
Total T3
-0.02 (-0.08,
0.03)
Free T3
-0.02 (-0.05,
0.02)
0.05 (-0.05,
0.16)
Preston et al.
(2018)
Project Viva cohort
(1999-2002), U.S.,
480 infants
2.4(1.6-
3.8)
P (95% CI)
for IQR
increase
-0.15 (-0.38,
0.08)
Girls
NR
NR
3-63
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median
exposure
(IQR) or as
specified
(ng/mL)
Effect
estimate
T4
T3
TSH
0.07 (-0.23,
0.37)
Boys
-0.46 (-0.83,
-0.1)*
Liang et al.
Cross-sectional
analysis within
Shanghai-Minhang
cohort (2012), China,
300 infants
2.7 (2.0-
3.4)
P (95% CI)
for In-unit
increase
Total T4
-0.59 (-7.94,
6.76)
Free T4
-0.32 (-0.87,
0.22)
Total T3
0 (-0.05, 0.04)
Free T3
0.02 (-0.08, 0.13)
0.43 (-1.02,
1.88)
(2020)
*p < 0.05.
One medium confidence study (Berg et al., 2017) is not included because quantitative results were only reported
for significant associations.
In addition, five studies (four medium confidence) fWenetal.. 2013: Kim etal.. 2020a: Gallo
etal.. 2022: Dufour etal.. 2018: Chan etal.. 20111 reported on the association between PFHxS and
dichotomous hyper- and hypothyroidism outcomes defined by the authors using set cutpoints. In
Wen etal. (2013). a medium confidence study there were greater odds of subclinical
hypothyroidism in men (OR 1.57, 95% CI 0.76, 3.25) and women (OR 3.10, 95% CI 1.22, 7.86), and
subclinical hyperthyroidism in women (OR 2.27, 95% CI 1.07, 4.80) and lower odds of subclinical
hyperthyroidism in men (OR 0.56, 95% CI 0.24,1.2). Subclinical hypothyroidism was defined as
TSH >5.43 mlU/L, and subclinical hyperthyroidism was defined as TSH < 0.24 mlU/L (both limited
to those without diagnosed thyroid disease). Also in adults, Dufour etal. f 20181 reported higher
odds (although not statistically significant) of hypothyroidism in pregnant women and Gallo et al.
(2022) did not report increases in thyroid disease or medication use. In the low confidence study
(Chan etal.. 2011). hypothyroxinemia in pregnant women was defined as normal TSH
concentrations with no evidence of hyperthyroidism (0.15-<4 mU/L) and free T4 in the lowest
10th percentile (<8.8 pmol/L) of the study sample). They found higher odds of hypothyroxinemia
with higher PFHxS exposure (OR 1.12, 95% CI 0.89,1.41). In children and adolescents, Kim et al.
f2020al reported lower odds of subclinical hypothyroidism with higher exposure and Gallo et al.
(2022) reported no association.
Thyroid effects summary
Overall, the evidence for the association between PFHxS exposure and thyroid effects is
inconsistent Some studies do indicate an association between thyroid hormones or subclinical
thyroid disease and PFHxS exposure, but this direction is not consistent across studies and the
associations with PFHxS exposure in most studies were null. There is also not clear coherence
3-64
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
across outcomes, with indications of associations with both hyper- and hypothyroidism and unclear
coherence of the direction of association between TSH and the other hormones. However, almost all
of the available studies were deficient in outcome ascertainment due to lack of consideration of
timing of sample collection. As discussed above, this is likely to result in nondifferential outcome
misclassification, which also is expected to bias results toward the null on average, although the
studies without this issue also reported null findings. Given these concerns, the findings across this
set of studies are difficult to interpret
Animal Studies
The toxicity evidence base for PFHxS-induced endocrine outcomes consists of three
multigenerational publications (two studies) in SD or Wistar rats fRamhai etal.. 2018: Ramhai et
al.. 2020: Bute nhoff etal.. 20091. one developmental study in Crl:CD mice fChang etal.. 20181. and
one short-term (28 day) study in SD rats (NTP. 2018a). All studies treated the animals orally to
PFHxS via gavage. Endocrine-related outcomes evaluated by these studies included: thyroid
hormones, histopathology, and endocrine organ weights including thyroid, parathyroid, and
adrenal gland weight. Potential PFHxS effects on male and female reproductive organs (e.g., testes
and ovaries) and reproductive hormones (e.g., testosterone and estradiol) that also encompass part
of the endocrine system are discussed in Male Reproductive Effects and Female reproductive
Effects sections.
Evaluation of the available animal studies showed that these were generally well conducted
for most endocrine-related endpoints. The available studies examined PFHxS endocrine toxicity
effects using doses that ranged between 0 and 10 mg/kg-day in mice (Chang etal.. 20181: 0 and
25 mg/kg-day in rats with the exception ofNTP f2018al. for which a range of 0-50 mg/kg-day in
female rats and 0-10 mg/kg-day in male rats was used. These ranges account for the
pharmacokinetic (PK) sex differences that have been observed in rats, for which PFHxS appears to
have a lower mean half-life in female rats versus their male counterparts (1.72 and 26.9 days,
respectively, after oral dosing (Kim etal.. 2016b)). No overt toxicity was observed at any of the
highest doses tested in any of the available studies. Two high confidence studies, Chang etal. (2018)
and NTP (2018a). examined PFHxS effects on histopathology endpoints; three high confidence
studies fNTP. 2018a: Chang etal.. 2018: Bute nhoff etal.. 20091 examined PFHxS effects on thyroid
gland weight Lastly, two high confidence studies (NTP. 2018a: Bute nhoff etal.. 2009) also
measured adrenal gland weights. A summary of the study evaluations for each endpoint are
presented in Figures 3-7, 3-10, and 3-11; additional details can be obtained from HAWC.
Thyroid hormones
Four studies (three high and one low confidence; see Figure 3-7, below) examined the
effects of PFHxS on levels of thyroid hormones, T3, T4, and/or TSH. One high confidence study, NTP
f2018al examined effects on serum concentrations of TSH, T3, and total and free T4 in adult
animals. The other two high confidence studies examined effects of PFHxS on serum T4 fRamhai et
3-65
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
al.. 2018). T3, and TSH (Ramhaj etal.. 20201 in exposed dams and their offspring (exposed via
lactation) through PND 22. Lastly, the fourth study was low confidence in which Chang etal. f20181
reported using a developmental study design that followed established guidelines for such studies
(OECD 422 Testing guidelines) fOECD. 20161. However, the reported study design ignored essential
components of the OECD 422 developmental toxicity screening guidelines. A necessary
requirement of the OECD guidelines is that serum T4 be measured as part of developmental toxicity
studies. The study authors did not measure T4 serum levels, under the rationale that T4 is an
"inactive hormone" and elected to measure TSH serum levels instead. It has been established that
serum TSH measures are not good indicators of potential endocrine disruption fStoker etal.. 2006:
OECD. 2016: Crofton. 20041.
Reporting quality -
Allocation -
Observational bias/blinding -
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
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 3-7. Study evaluation results for measures of thyroid hormone levels in
PFHxS animal toxicity studies. Full details available by clicking HAWC link.
NTP f2018al measured free and total T4 serum levels in Sprague Dawley and Ramhai et al.
f20181 measured total T4 serum levels in Wistar rats (see Figures 3-8 and 3-9). NTP observed a
statistically significant, dose-dependent decrease (p < 0.01) of free and total T4 levels starting at the
lowest experimental dose (0.625 mg/kg-day) in male rats (up to 60% and 78% decrease in free and
total T4 respectively); free T4 and total T4 were significantly decreased beginning at 12.5 mg/kg-
day and 6.25 mg/kg-day, respectively, in female rats (p < 0.01, up to 32% and 38 % decrease in free
and total T4 respectively). However, serum-total T4 levels are a more sensitive and reliable
measure of T4 due to sensitivity limitations in the available assays used to measure free T4. Ramhai
3-66
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
etal. (2018) reported similar findings in Wistar rat dams, with statistically significant, dose-
dependent decreases in serum-total T4 at 5 mg/kg-day and above in dams at PND 22 after exposure
from gestational day 7 (GND 7) through postnatal day 16/17 fRamhai etal.. 20181 (-26% decrease
at 5 mg/kg-day dose and up to -71% decrease at 25 mg/kg-day dose). Comparable observations
were made in the pups born to the PFHxS-exposed dams in Ramhai etal. f 20181. with statistically
significant decreases in total T4 levels in serum collected from PND 22 pups at >5 mg/kg-day
(p < 0.001, up to a 71% decrease in total T4 at 25 mg/kg-day dose and 38% decrease in total T4 at
5 mg/kg-day dose). No overt toxicity was observed at any of the highest doses tested in any of the
available studies. Effects occurred at lower concentrations of PFHxS in male rats than their female
counterparts indicating that males could be more susceptible to PFHxS effects than females (see
Figure 3-8). However, a more likely explanation is that these observations, at least in part, can be
explained by the differences in PFHxS pharmacokinetics that exist between male and female rats
(see Section 3.1). Sex differences in plasma half-life and tissue distribution have been observed for
PFHxS, wherein PFHxS-exposed male rats have a longer plasma half-life (20.7-26.9 days) versus
their female counterparts (0.9-1.7 days) (Kim etal.. 2016b!
Two studies, NTP f2018al and Ramhai etal. f20201. measured T3 in serum. NTP f2018al
observed a statistically significant and dose-dependent decrease (p < 0.05) in serum T3 levels in
male, but not female, SD rats at > 0.625 mg/kg-day (p < 0.01); Ramhai etal. f20201 in a similar
study design as Ramhaj etal. (2018). reported a significant decrease in serum T3 in Wistar rat
dams at the highest tested dose: 25 mg/kg-day at PND 22 after exposure from gestational day 7
(GND 7) through postnatal day 16/17 (p < 0.001,19% decrease). Comparable observations were
also made in the pups born from the exposed dams at PND 16/17 in which a significant decrease in
serum T3 was observed in pups of both sexes at the highest dose: 25 mg/kg-day (p < 0.001,16%
decrease).
Lastly, three studies, NTP f2018al. Chang etal. f20181 and Ramhai etal. f20201 investigated
PFHxS effects on TSH levels. None of these studies observed changes in TSH serum levels in male or
female CD1 mice, Sprague Dawley rats or Wistar rats in response to PFHxS exposure.
Taken together, and as noted in the study results reported by NTP and the combined
Ramh0j studies (Ramh0i etal.. 2018: Ramh0i etal.. 2020). these results indicate that PFHxS
exposure in rats has the ability to adversely decrease the endocrine hormones, T4, and T3, in the
absence of observed effects on TSH.
3-67
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Endpoint Name
Study Name
Experiment Name
Species Strain
Generation
Sex
Lifestage Exposed
PFHxS Effects on Animal Thyroid Hormones
Thyroid Stimulating Hormone (TSH)
NTP, 2018, 4309363
28 Day Oral
Rat Sprague-Dawley
Female
Male
7-8 week old
7-8 week old
1
1
Ramh0j, 2020, 6320959
Multigenerational Oral
Rat Wistar
P0
Female
Adult (gestation)
F1
Male
Developmental
Thyroxine (T4), Free
NTP, 2018, 4309363
28 Day Oral
Rat Sprague-Dawley
Female
7-8 week old
~ ~ i
Male
Female
7-8 week old
adult
Thyroxine (T4), Total
Ramh0j, 2018, 4442260
Multi-Generational Oral (range-finding)
Rat Wistar
P0
•-V
F1
Male
fetal and juvenile
•-T
Thyroxine (T4), Free
Ramhej, 2018, 4442260
Multi-Generational Oral (range-finding)
Rat Wistar
F1
Combined
fetal and juvenile
•-V
Thyroxine (T4), Total
Ramhcj, 2018, 4442260
Multi-Generational Oral
Rat Wistar
P0
Female
Adult (gestation)
V V
•
V V
F1
Combined
Fetal and Juvenile
TT XT
NTP, 2018, 4309363
28 Day Oral
Rat Sprague-Dawley
Female
7-8 week old
• ~ ~ ~ i
Male
7-8 week old
I—I Doses
A Significant Increase
~ Significant Decrease
•-~ ~ ~ i
Triiodothyronine (T3)
NTP, 2018, 4309363
28 Day Oral
Rat Sprague-Dawley
Female
7-8 week old
—i
Male
7-8 week old
m-W—W-W-—i
Ramh0j, 2020, 6320959
Multigenerational Oral
Rat Wistar
P0
Female
Adult (gestation)
m V
F1
Combined
Developmental
m V
i
1 T
0.01 0.1 1 10 100
Dose (mg/kg)
Figure 3-8. Summary of thyroid hormone measures in animal studies. Figure displays the three high confidence studies
included in the analysis; the sole low confidence study, Chang etal. (20181 was omitted from the analysis. Full details available by
clicking HAWC link. Details on study confidence may be found in Figure 3-7.
3-68
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
NTP. 2018. A 3093G3
s»j, 202O, 6320959 Triiodothyronine
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Histopathologv
Three high confidence studies evaluated nonneoplastic histopathologic lesions in endocrine
tissues in response to PFHxS exposure fRamhai etal.. 2020: NTP. 2018a: Butenhoff et al.. 20091
(see Figure 3-10). NTP f2018al evaluated various organs in the endocrine system including the
adrenal cortex, adrenal medulla, parathyroid gland, pituitary gland, and thyroid gland in adult male
and female Sprague-Dawley rats exposed to PFHxS for 28 days. NTP (2018a) observed no
histological lesions in any of the endocrine tissues they evaluated and made no observations of
hyperplasia or hypertrophy in the thyroids at doses up to 10 mg/kg-day in male rats or 50 mg/kg-
day in female rats. However, a 44-day study by Butenhoff et al. f20091 observed increased
incidences of hypertrophy and hyperplasia (characterized as "minimal") of thyroid follicular
epithelial cells in adult Sprague-Dawley male rats that were exposed to 3.0 mg/kg-day PFHxS (40%
incidence) and an increase in "moderate" hypertrophy and hyperplasia at 10 mg/kg-day PFHxS
(70% incidence) for up to 44 days (minimal hypertrophy/hyperplasia (20% incidence) was
observed in control animals). The study authors attributed the pathological changes in the thyroid
to changes in enzyme induction in the liver (see Serum Biomarkers of Liver Function in Section
3.2.4) that have been shown by others fSanders etal.. 19881 to result in a compensatory increase in
T4 clearance that may elicit increases in TSH hormone levels or no compensatory TSH responses.
The role of TSH in the progression of thyroid hyperplasia and hypertrophy was highlighted in
Noves etal. (2019). In the proposed Adverse Outcome Pathway (AOP) by Noves etal. (2019). the
authors illustrate that increased serum TSH may lead to thyroid hyperplasia and hypertrophy.
However, Butenhoff et al. (2009) did not measure thyroid hormone levels as part of their
experimental analysis, so this hypothesis was not tested. Lastly, Ramh0i etal. (2020) reported that
in Wistar rat dams exposed to PFHxS at doses ranging from 0.05 to 25 m/kg-day from gestational
day 7 (GND 7) through postnatal day 16/17, no PFHxS effects on thyroid histopathology were
observed. The authors reported that the thyroid glands corresponding to the high dose (25 mg/kg-
day) male pups showed "small histological changes;" however, these changes were within the
normal range and were no longer evident on PND 22. The authors did not observe hypertrophy or
hyperplasia at any time point in either the exposed dams or their offspring (Ramh0i etal.. 2020).
3-70
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reporting quality -
Allocation -
Observational bias/blinding -
Confounding/variable control -
Selective reporting and attrition -
Chemical administration and characterization -|
Exposure timing, frequency and duration -
Results presentation -
Endpoint sensitivity and specificity -
Overall confidence -
Legend
| Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
NR| Not reported
Figure 3-10. Study evaluation results for endocrine histopathology outcomes
in PFHxS animal toxicity studies. Full details available by clicking HAWC link.
Organ weights
Three studies evaluated the effect of PFHxS exposure on thyroid gland weights fRamhai et
al.. 2020: NTP. 2018a: Chang etal.. 20181 (see Figures 3-11 and 3-121. Chang etal. f20181 and NTP
f2018al observed no significant effects in adult CD1 male or female mice or in adult male or female
Sprague Dawley rats at the PFHxS doses administered in these studies (see Figure 3-12). However,
Ramh0i etal. (20201 observed a statistically significant (p < 0.05) decrease in absolute thyroid
weights (relative weights were not reported) starting at 5 mg/kg-bw-day that continued into the
highest dose tested (25 mg/kg-bw-day) in PND 22 female Wistar pups exposed to PFHxS starting at
GD7 (5 mg/kg-bw-day p < 0.05,17% decrease; 25 mg/kg-bw-day p < 0.01; 23% decrease) (see
Figure 3-12). The differences in experimental designs across these studies make it difficult to
compare the results and thus the importance of the findings reported by Ramhai etal. f20201 is
unclear.
Two studies, Butenhoff et al. (20091 and NTP (2018al evaluated the effects of PFHxS on
adrenal gland weights in SD rats. Butenhoff et al. (20091 reported no effect on absolute or relative
adrenal weight resulting from 0, 0.3,1.3, or 10 PFHxS mg/kg-day for 44 days. NTP observed a
statistically significant increase in absolute adrenal weights in female rats (at >12.5 mg/kg-day;
15% increase) and an increase in relative adrenal gland weight at 50 mg/kg-day (9% increase
p < 0.01) in female rats. NTP also reported decreases in both absolute (at >5 mg/kg-day; -13%;
3-71
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
p < 0.05) and relative adrenal weights (at >2.5 mg/kg-day; -17%; p < 0.05) in male rats. It is
unclear why there were opposing responses across sexes in the NTP study that were not observed
in Butenhoff et al. f20091 (see Figure 3-12); however, these observations could be due to the
pharmacokinetic differences between male and female animals coupled with differences in study
design between the two studies.
Overall, the organ weight changes are mixed and cannot be readily interpreted.
Reporting quality -
Allocation -
Observational bias/blinding -
Confounding/variable control -|
Selective reporting and attrition -
Chemical administration and characterization -
Exposure timing, frequency and duration -I
Results presentation -|
Endpoint sensitivity and specificity-
Overall confidence -I
Legend
I Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
NR Not reported
Figure 3-11. Study evaluation results for endocrine organ weights in PFHxS
animal toxicity studies. Full details available by clicking HAWC link.
3-72
-------
IRIS Toxicoloqical Review of Perfluorohexanesulfonic Acid and Related Salts
Endpoint Name
Thyroid Weight, Absolute
Thyroid Weight, Relative
Adrenal Glarid Weight, Absolute
Study Name
NTP, 2018, 4309363
Effect Subtype Experiment Name
Absolute 28 Day Oral
Ramhoj, 2020, 6320959 Absolute Multigenerational Oral
NTP, 2018, 4309363
NTP, 2018, 4309363
Relative
Adrenal Gland Weight, Relative
NTP, 2018, 4309363
Relative
28 Day Oral
Absolute 28 Day Oral
Adrenal Gland Weight, Left, Absolute Butenhoff, 2009, 1405789 Absolute Multi-Generational Oral
Adrenal Gland Weight, Right, Absolute Butenhoff, 2009, 1405789 Absolute Multi-Generational Oral
28 Day Oral
Adrenal Gland Weight, Left, Relative Butenhoff, 2009, 1405789 Relative Multi-Generational Oral
Adrenal Gland Weight, Right, Relative Butenhoff, 2009, 1405789 Relative Multi-Generational Oral
Animal Description
Rat, Sprague-Dawley ($)
Rat, Sprague-Dawley (o)
F1 Rat, Wistar (?)
F1 Rat, Wistar (2)
Rat, Sprague-Dawley (9)
Rat, Sprague-Dawley (J)
Rat, Sprague-Dawley (9)
Rat, Sprague-Dawley (3)
P0 Rat, Sprague-Dawley (9)
P0 Rat, Sprague-Dawley 0)
P0 Rat, Sprague-Dawley (2)
P0 Rat, Sprague-Dawley (o)
Rat, Sprague-Dawley (9)
Rat, Sprague-Dawley (S)
P0 Rat, Sprague-Dawley (9)
P0 Rat, Sprague-Dawley (c?)
P0 Rat, Sprague-Dawley ( ?)
P0 Rat, Sprague-Dawley (d>)
PFHxS Animal Endocrine Organ Weight Effects
# No significant change
A Significant increase
V Significant decrease
• •—
# •-
# •-
#-•-
• •-
• •-
-m AAA,
V?
—•
—•
—•
—•
-•-•-•-A
WW
0.001
0.01
0.1 1
Axis label
10
100
Figure 3-12. Summary of endocrine organ weight effects in animal studies. Figure displays the medium and high confidence
studies. Full details available by clicking HAWC link.
3-73
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Mechanistic Evidence and Supplemental Information
The available thyroid hormones data in rodents showed strong effects on T4 and T3 after
short-term exposure, although no effects were observed on TSH; however, a pattern of decreased
T4 without pronounced (or detectable) changes in TSH is consistent with hypothyroxinemia and
has been observed in some analyses of other PFAS, including several long-chain PFAS (e.g., PFOA
and PFOS (Kim etal.. 2018a) and short-chain (e.g., PFBS, PFBA, and PFHxA (U.S. EPA. 2021b. d,
2023b) PFAS. During pregnancy and early development, perturbations in thyroid function can have
impacts on normal growth and neurodevelopment in the offspring (Zoeller and Rovet. 2004: Y etal..
2024: Street etal.. 2024: Stagnaro-Green and Rovet. 20161. Low thyroid hormone status is also
likely associated with effects in numerous other organ systems, including the heart, bone, lung, and
intestine fWexler and Sharretts. 2007: Mochizuki et al.. 2007: Bizzarro and Gross. 2004: Bassett et
al.. 20071.
Mechanistic studies on the endocrine effects of PFHxS are scarce, with only one study
conducted in a mammalian test system. Long etal. (2013) explored the effects of PFHxS along with
other PFAS on thyroid hormone signaling and the aryl hydrocarbon receptor (AhR) using the T3-
dependent rat pituitary cell line, GH3. The authors found that PFHxS inhibited GH3 cell
proliferation in a dose-dependent manner. Additionally, the authors found that PFHxS—along with
three other PFAS (PFOS, PFNA, and PFUnA)—antagonized GH3 cell proliferation in response to
exogenous T3 treatment. The authors speculated that PFHxS may compete with T3 for binding to
thyroid hormone receptor (TR) or other cofactors to inhibit cell proliferation; however, specific
experiments testing this hypothesis were not conducted.
Other studies in nonmammalian systems (e.g., avian neuronal cells and chicken embryos)
have shown that PFHxS alters mRNA levels of thyroid hormone-responsive genes, including
transthyretin (TTR) fVongphachan etal.. 2011: Cassone etal.. 20121. TTR is a transport protein that
is secreted into the blood by the liver and by the choroid plexus into the cerebrospinal fluid. TTR
binds to thyroid hormones such as T4 and T3 in the serum and in the cerebrospinal fluid. Because
of its low affinity for thyroid hormones TTR readily disassociates from these and is therefore
responsible for the immediate delivery of T3 and T4 to various extrahepatic tissues and potentially
into the brain (Palha. 2002). Decreases in TTR may lead to decreases in T4 transport (Refetoff.
20151. Additionally, TTR plays a key role in thyroid hormone storage and transport during fetal
development PFHxS-induced decreases in TTR mRNA have been shown in nonmammalian
systems, and the above mechanism would in part assist in elucidating the mechanisms underlying
the in vivo observations pertaining to PFHxS-induced decreases T3 and T4.
Further, mechanistic studies exploring the effects of PFHxS on thyroid hormone transport
have shown that PFHxS competes with T4 for binding to TTR, but not thyroxine-binding globulin
(TBG) (Weiss etal.. 2009: Ren etal.. 2015: Renet al.. 2016: Huang etal.. 2023). However, TTR binds
only a small portion of the circulating thyroid hormones (15%-20%) fRefetoff. 20151. and
3-74
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
confirmatory studies in model systems more relevant to humans would be needed to understand
the potential role of PFHxS-induced alterations to thyroid hormone-responsive genes in humans.
Data from the ToxCast Dashboards Endocrine Disruptor Screening Program (EDSP21)
fhttps: //comptox.epa.gov/dashboard/chemical-lists/EDSPUOC] reveal that K+PFHxS was active in
a total of only 2 out of 57 endocrine-related assays (with both positive hits at PFHxS levels nearing
the cytotoxicity limit). A summary of the assay results from the EDSP21 project may be found in
Appendix C, Section 3. Briefly out of 27 estrogen receptor assays, K+PFHxS was active in one, the
ATG_ERE_CIS_up induction assay with an AC50 at 96.96 |a,M (see Figure 3-13). K+PFHxS was not
active in any of the 16 androgen receptor assays. K+PFHxS was active in one out of 13 assays
associated with perturbation of thyroid hormone signaling synthesis, or metabolism, namely the
NIS-RAIU_inhibition assay with an AC50 of 18.68 [J.M. It should be noted that the current panel of
bioactivity assays interrogating thyroid hormone dynamics is predominately targeted at receptor-
dependent agonism/antagonism, which is only one of several pathways by which the mammalian
HPT axis may be perturbed by PFAS (Noves etal.. 2019). K+PFHxS was not active in any of the three
steroidogenesis assays in the database. Overall, although not conclusive, PFHxS exhibited little in
vitro endocrine activity in these assays (>96% of assays were inactive).
Overall, the mechanistic information is scarce and inconclusive, and therefore does not
provide clear support for or against endocrine (thyroid)-modulating activity of PFHxS.
3-75
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
B
I
i
S 1
s
1 AT G_ERE_OS_up
HlTCALL ACTIVE
Potassium peruuoronexanesuffonafe (3871-99-6)
1
DTXSID 3037709
1
1
TP0000422F06
Cut Off
1
i
i
9
1 /
I /
li
o
O
o —
£ 1
/ |
r 1
O 0
s!
*01
S'
si
<1
—1
Winning Model
.4 4
I Constant Mode*
Model
AiC
| Gain-loss Model
RUSE
2 3
Log Concentration (uM)
¦ HO Model
Too
AC50
Stope
"I NI$_RAiUjnftft>ition
HlTCALL ACTIVE
Potassium peiHuoronexanesusfonate (3871-99-6)
OTXSID3O37709
TP0001679E08
Cutoff
/• i
8X
o
O • %
i
0
S\
oit
I:
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence Integration
Human studies provide conflicting evidence as to the potential effects of PFHxS on thyroid
outcomes (e.g., thyroid hormone levels). Although a few studies did suggest an association between
increasing PFHxS exposure levels and decreased circulating thyroid hormones (i.e., T4) or
subclinical thyroid disease, the associations were not consistent across studies (most studies were
null); the inconsistent findings could not be explained by differences in study design, confidence, or
other factors such as population, and there was no clear coherence across outcomes. The available
human evidence on PFHxS effects on the thyroid is indeterminate.
Evidence of thyroid toxicity resulting from PFHxS exposure in animal models exposed in
short-term and multigenerational studies showed dose-dependent effects on thyroid hormone (TH)
levels, most notably consistent decreases in serum T4 levels in rats (untested in mice) fRamhai et
al.. 2018: NTP. 2018a). Coherent and consistent decreases in T3 in rats were also observed across
studies, whereas TSH was unchanged. Thyroid organ weights and thyroid histopathology were
inconsistently or only weakly affected across studies (e.g., increased incidence of thyroid
hypertrophy and mild hyperplasia in one study and decreased thyroid weight in another, with
otherwise null results), suggesting that the TH decreases are probably not attributable to effects of
PFHxS on thyroid gland function. However, the available evidence from exposed rodents shows a
consistent, dose-dependent disruption of thyroid hormone homeostasis, characterized by
decreased T4 and T3 serum levels concurrent with unaffected, normal levels of TSH is consistent
with hypothyroxinemia and also consistent with what has been observed in other PFAS including
PFBS, PFHxA, PFBA, and PFOA (U.S. EPA. 2021b. d, 2022a: Kim etal.. 2018a). The observed effects
are also consistent with central hypothyroidism, a condition that occurs when the pituitary gland or
hypothalamus is unable to produce sufficient hormones for the thyroid gland to function properly.
It is defined as hypothyroidism due to insufficient stimulation by TSH. Patients with central
hypothyroidism have normal TSH levels but decreased levels of T4 (Gupta and Lee. 2011).
However, a thyrotropin releasing hormone (TRH) stimulation test would be needed determine if
central hypothyroidism was resulting from PFHxS exposure in the rats from the NTP (2018a) and
Ramh0i etal. (2018) studies. This test was not performed and presents another data gap in the
PFHxS evidence base. The observed TH decreases occurring in exposed adult animals and indirectly
(through the dams) exposed offspring were of a large magnitude of effect and occurred even at
PFHxS exposure levels as low as 0.625 mg/kg-day in male rats. This finding is consistent with the
published proposed thyroid disruption Adverse Outcome Pathway (AOP) byNoves etal. (2019) and
publication by Zoeller and Crofton (2005). in which the authors illustrated that endocrine
disruption in humans and rodents possess analogous key events and adverse outcomes perhaps
due to conserved biology across species (see additional discussion below). Decreased thyroid
hormone levels are judged relevant to human health, given the many similarities in the production,
regulation, and functioning of thyroid hormones between rodents and humans fZoeller and Rovet.
2004: Vansell. 2022: Stagnaro-Green and Rovet. 2016: Rovet. 2005. 2014: Navarro etal.. 2014:
3-77
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Morreale de Escobar et al.. 2008: Hoodetal.. 1999a: Hood etal.. 1999b: Hood and Klaassen. 2000:
Dong etal.. 2015: Cuevas etal.. 2005: Berbel etal.. 20101. Taken together, the available animal
evidence on endocrine effects, which is primarily based on the observed supporting decreases in
thyroid hormone levels after PFHxS exposure, is considered moderate.
Mechanistic studies examining the endocrine disrupting effects of PFHxS are scarce. In the
single mammalian study, Long etal. (2013). PFHxS, similar to other tested PFAS, inhibited cell
growth but not proliferation in the T3-dependent rat pituitary cell line, GH3. However, while this
study suggests the possibility that PFHxS might compete with THs, these data alone are insufficient
to provide support for biological plausibility.
The currently available evidence indicates that PFHxS exposure likely causes thyroid
effects in humans given sufficient exposure conditions8 (see Table 3-10). This conclusion is based
primarily on consistent and coherent decreases in thyroid hormone levels across short-term and
multigenerational studies in rats exposed to PFHxS levels >2.5 mg/kg-day (with males being more
sensitive). The pattern of available evidence in rats indicates that PFHxS, like other PFAS (U.S. EPA.
2021b: Coperchini etal.. 2017) leads to a disruption of thyroid hormone homeostasis in a pattern
similar to hypothyroxinemia. Noves etal. f20191 along with Zoeller and Crofton f20051 illustrated
that endocrine disruption in humans and rodents possess analogous key events and adverse
outcomes perhaps due to conserved biology across species, and thus these effects are considered
adverse and relevant to humans. These TH decreases could have detrimental effects on susceptible
populations as T3 and T4 are critical in brain development and bone growth during early childhood
and adolescence (Crofton. 2004). However, at present, few epidemiological studies and
toxicological studies have addressed PFHxS-induced effects in these populations, highlighting an
important data gap.
8The "sufficient exposure conditions" are more fully evaluated and defined for the identified health effects
through dose-response analysis in Section 5.
3-78
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-10. Evidence profile table for PFHxS thyroid effects
Evidence stream summary and interpretation
Evidence integration
summary judgment
Evidence from studies of exposed humans (see Human Thvroid Section)
Studies and confidence
Factors that increase
certainty
Factors that decrease
certainty
Summary and key findings
Evidence stream judgment
®©o
Evidence Indicates (likely)
Thyroid Measures &
Disease
Twenty-six medium
confidence studies
Ten low confidence
• No factors noted
• Unexplained
inconsistency
Some human studies report an
inverse association between
thyroid hormones and PFHxS
exposure, but most studies
reported null findings.
ooo
Indeterminate
Primary basis:
Moderate animal evidence for
decreased T4 and T3 in adult
and juvenile rats
Human relevance:
Effects in rats are considered
relevant to humans due to
conserved biology across
species (see Evidence
Integration section.)
Cross-stream coherence:
NA; human evidence
Evidence from in vivo animal studies (see Animal Thvroid Section)
Studies and confidence
Factors that increase
certainty
Factors that decrease
certainty
Summary and key findings
Evidence stream judgment
Thyroid Hormones
Three high confidence
studies in rats
• 28-d
• Multigenerational
• Consistent and
coherent decreases of
T4 and T3 in adult and
juvenile rats in the
absence of effects on
TSH
• Large Magnitude of
effect (up to 70%)
• Dose response in
studies
• No factors noted
Studies in rats (2 forT3 and 3
forT4) reported significant
decreases in TH levels in both
male and female rats (forT4),
or just male rats (for T3),
generally after PFHxS exposure
at >2.5 mg/kg-d.
©0©
Moderate
Based on decreased T4 and
T3
indeterminate
Susceptible Populations and
lifestages:
Young individuals exposed to
PFHxS during gestation and
early childhood may be
susceptible populations.
3-79
-------
IRIS Toxicoloqical Review of Perfluorohexanesulfonic Aria
' and Related Salts
Evidence stream summary and interpretation
Evidence integration
summary judgment
Histopathology
Three high confidence
studies in rats
• 28- and 42-d
• Multigenerational
• No factors noted
• No factors noted
Increased incidence of thyroid
hypertrophy and hyperplasia in
male rats in one study.
Organ Weights
Three high confidence
studies in rats and one
medium confidence
study in mice
• Concerning magnitude
of effect (up to 23%
decrease) in female
pups in one study
• Unexplained
inconsistency
(across studies for
thyroid weights
and across sexes
for adrenal
weights)
Decreased absolute thyroid
weight in female F1 pups at
PND 22 (one study); Increased
absolute adrenal gland weight
in female rats and decreased
absolute adrenal gland weight
in male rats (one study);
Increased relative adrenal
gland weight in female rats
(highest dose only) and
decreased a relative adrenal
gland weight in male rats (one
study).
3-80
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3.2.2. Immune Effects
Human Studies
Epidemiology studies examining immune effects of PFHxS exposure include studies on
antibody response, infectious diseases, and hypersensitivity-related outcomes, which includes
asthma, allergies, and atopic dermatitis. The health effects results were grouped across studies to
develop conclusions on the same or related outcomes for the main categories of immune response
according to immunotoxicity guidance from the World Health Organization/ International
Programme on Chemical Safety (IPCS. 20121: (1) immunosuppression, (2) sensitization or allergic
response, and (3) autoimmunity. Evidence for potential immune effects was considered within
these three categories because of common and related mechanisms within each category. Within
each category, health effects data were considered in the order of most to least informative for
immunotoxicity risk assessment (IPCS. 2012). Specifically, clinical studies on disease or immune
function assays are considered most informative, then general/observational immune assays
(lymphocyte phenotyping or cytokines), and finally endpoints such as hematology (i.e., blood
leukocyte counts) are least informative. Outcomes related to immunosuppression were considered
within two subcategories: antibody response and infectious disease. Several different outcomes,
such as asthma and food allergies, were included in the sensitization and allergic response category.
No studies were identified that evaluated outcomes related to autoimmunity.
Immunosuppression
Antibody response outcomes
The production of antigen-specific antibodies in response to an immune challenge (e.g.,
vaccination in humans or injection with sheep red blood cells in rodents) is a well-accepted
measure of immune function included in risk assessment guidelines and animal testing
requirements for immunotoxicity (U.S. EPA. 1998: IPCS. 1996. 2012: ICH Expert Working Group.
2005). The production, release, and increase in circulating levels of antigen-specific antibodies are
important for protection against infectious agents and preventing or reducing severity of influenza,
respiratory infection, colds, and other diseases as part of the humoral immune response. Reduced
antibody production is an indication of immunosuppression and may result in increased
susceptibility to infectious disease.
Evaluations for studies of antibody responses following vaccination as reported in 10
epidemiological studies (reported in 11 publications) are summarized in Figure 3-14. Among these
studies, there were analyses of several vaccinations: diphtheria (six studies), tetanus (seven
studies), measles (three studies), rubella (two studies), mumps (one study), Haemophilus
influenzae Type B (two studies), hepatitis (one study), and FluMist (one study). There were four
prospective birth cohorts, including three in the Faroe Islands and one in Norway fGranum etal..
20131. and one cohort of children beginning in their first year of life in Guinea-Bissau
3-81
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
(Timmermann etal.. 20201. The three Faroe Islands studies included non-overlapping populations
enrolled at separate times, all medium confidence, one with enrollment in 1997-2000 and
subsequent follow-up to age 7 fGrandiean etal.. 20121 and age 13 fGrandiean etal.. 2017al. one
with enrollment in 2007-2009 and follow-up to age 5 fGrandiean etal.. 2017bl. and one with
enrollment in 1986-1987 and follow-up to age 28 fShih etal.. 20211. These cohorts are thus
considered independent of each other. Some analyses in Grandjean etal. (2017b) combined new
data from the cohort born in 2007-2009 with new follow-up data from the cohort born in 1997-
2000 fGrandiean etal.. 20121: these are labeled in the results table. Given that the etiologic window
for immune effects of PFAS exposure is not known, these studies in the Faroe Islands have the
benefit of assessing multiple windows of exposure (maternal, multiple points in childhood) as well
as following outcomes over time. For example, exposures measured during infancy could have
reflected residual maternal antibodies, but the half-life of maternal antibodies is short and residual
antibodies would not be expected to exist beyond infancy and would not exist in the children at age
5 years. Similarly, vaccine boosters likely changed these children's antibody concentrations over
time, but such changes were not expected to be related to PFHxS concentration. Having multiple
windows of exposure in this study allowed for comparisons of effects. In children, there were also
two medium confidence cross-sectional studies in the U.S. and Greenland f Timmermann etal.. 2021:
Stein etal.. 2016bl and one low confidence (due to expected residual confounding) cross-sectional
study in Germany (Abraham etal.. 2020). In adults, there were two additional low confidence
studies, a short-term cohort (with exposure measured at vaccination and follow-up 30 days later)
in the United States (Stein etal.. 2016a) and a cross-sectional study in Denmark (Kielsen etal..
2016). These studies were low confidence due to concerns for potential selection bias and
confounding.
3-82
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Abraham, 2020, 6506041 -
Grandjean, 2017a, 3858518-
Grandjean, 2017b, 4239492-
Granum, 2013, 1937228-
Kielsen, 2016, 4241223-
Shih, 2021, 9959487-
Stein, 2016a, 3860111 -
Stein, 2016b, 3108691 -
Timmermann, 2020, 6833710 -
Timmermann, 2021, 9416315-
+
+
+
-
-
-
+
¦
++
+
+
-
+
+
++
+
+
-
+
+
~
+
H
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)
++
++
++
+
+
-
+
+
"
B
++
-
+
-
+
-
B
++
+
++
+
+
+
D
++
-
+
-
+
-
++ ++
B
+
++
+
+
+
+
+
+
+
++
-
+
+
+
+
+
+
++
+
+
+
Figure 3-14. Summary of evaluation of epidemiology studies of PFHxS and
antibody response immunosuppression effects. For additional details see HAWC
link.
Multiple publications of the same data are presented on the heat map as one study. Grandiean et al. (2017a) also
includes Grandiean et al. (2012).
The results for this set of studies are shown in Tables 3-11 (children) and 3-12 (adults).
Although results were mostly not statistically significant, a general inverse trend was apparent,
particularly among studies of children. Of the six medium confidence studies in children, three
(Stein etal.. 2016b: Granum etal.. 2013: Grandjean etal.. 2017al observed a statistically significant
inverse association for at least one vaccine type while the other three also reported inverse
associations in some analyses fTimmermann etal.. 2020: Timmermann etal.. 2021: Grandiean et
al.. 2017b], Antibody levels were measured in the blood of individuals of several age groups (and
therefore different lengths of time since their initial vaccination or booster vaccination) and
compared with serum PFHxS concentrations also measured at different ages. All the studies in
children reported an association between higher concentrations of PFHxS and lower anti-vaccine
antibody levels in at least some exposure-outcome analysis pairs. These associations were
statistically significant for tetanus vaccination in children at ages 5 and 7 with childhood exposure
measurement in Grandiean et al. f20121 and for rubella vaccination in Granum etal. f20131 and
Stein etal. f2016bl. There are some results in the opposite direction for sub-analyses of the Faroe
Island cohorts and in Timmermann etal. (2021). In Timmermann etal. (2020). an inverse
3-83
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
association was observed in children who had received only one measles vaccination, but a positive
association was observed in children who had received two vaccinations. Neither of these results
were statistically significant, but the exposure contrast in this study was limited, which may have
influenced their ability to detect a statistically significant effect. No biological rationale has been
identified as to whether one exposure time period is more predictive of an overall immune
response which might explain the few inconsistent results. Only one study (Timmermann etal..
20211 examined the odds ratio for not being protected against diphtheria (antibody
concentrations <0.1 IU/mL), which has clearer clinical significance than continuous changes in
antibody levels, and they reported an OR of 6.44 (95% 1.51, 27.36) among children with known
vaccination records (adjusted for area of residence, consistent with continuous antibody results).
In adults, the birth cohort with follow-up to young adulthood fShih etal.. 20211 reported
inconsistent results across exposure measurement timing windows. Results were similarly
inconsistent for antibodies to Hepatitis A and B (not shown). One low confidence study reported an
inverse association for diphtheria and tetanus vaccination (Kielsen et al.. 20161. The single study of
FluMistreported no immunosuppression (Stein etal.. 2016a).
It is plausible that the observed associations with PFHxS exposure could be explained by
confounding across PFAS. Exposure levels to other PFAS in the Faroe Islands populations were
considerably higher (blood concentrations of PFOS 17 ng/mL, PF0A4 ng/mL, PFHxS 0.6 ng/mL) at
age 5 years in Grandjean etal. (2012). and there was a moderately-high correlation between PFHxS
with PFOS and PFOA (r = 0.57 and 0.53, respectively). The authors assessed the possibility of
confounding in a follow-up paper (Budtz-l0rgensen and Grandiean. 2018) in which PFHxS effect
estimates from a piecewise-linear model were adjusted for PFOS and PFOA and there was only
limited attenuation of the observed effects of PFHxS indicating that there was still an independent
effect of PFHxS(see Appendix D, Table D-l). These two PFAS were the most important to control for
given that they were the most highly correlated with PFHxS and present at the highest
concentrations in the population. The other available studies did not perform multipollutant
modeling. In Stein etal. (2016b). correlations between PFHxS and PFOS and PFOA were moderate-
high (r = 0.6 and 0.45, respectively), while in the other studies of antibody response, specific
correlations for each pair of PFAS were not provided, so it is difficult to determine the potential for
highly correlated PFAS to confound the effect estimates. Still, seeing PFHxS associated with the
outcome in multiple studies, each of which have different exposure conditions and thus different
inter-PFAS correlations, reduces the likelihood that confounding is the explanation. Overall, while it
is not possible to rule out confounding across PFAS, the available evidence supports that it is
unlikely to completely explain the observed effects, based primarily on the multipollutant modeling
results of the Faroe Islands studies (Budtz-l0rgensen and Grandiean. 2018). Other sources of
potential confounding, including possible co-exposures such as PCBs, were controlled
appropriately.
3-84
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Despite the imprecision of many of the individual exposure-outcome analysis pairs, the
findings are generally consistent with an association between PFHxS exposure and
immunosuppression. Of the 37 antibody-to-PFHxS-exposure analyses provided in Table 3-11, 26
support a finding of decrease in antibodies with higher PFHxS concentration. While some were less
than a 1% decrease in antibody concentration per doubling of PFHxS concentration, the majority
were greater than 5% and several were greater than 10%. While clinical adversity is not clear for
these fairly small changes in antibody levels for a healthy individual, as with any continuous health
measure, by lowering the immune response of the entire population, it is likely that a subset of
people will be shifted into clinically relevant immune suppression and that people with preexisting
immunosuppression will be more severely affected. This combined with the elevated odds for lack
of protection from diphtheria in Timmermann etal. f20211 support that this is a relevant health
effect resulting from PFHxS exposure. The variability in the results, including a few 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 evidence. The inverse associations were observed despite limited sensitivity resulting
from narrow exposure contrast in some studies. While multiple of the available studies are in a
fairly specific population (i.e., Faroe Islands), this is the highest quality evidence available and the
results are directly relevant to humans in general, particularly given the similar exposure levels to
the general U.S. population. There is no evidence that differences in dietary habits (e.g., marine diet)
or social determinants of health in this population can explain the results. In summary, some
uncertainty remains resulting from variability in the response by age of exposure and outcome
measures as well as from vaccination (initial and boosters), and also due to the potential for
confounding across PFAS discussed above; but overall, the available evidence provides support for
an association between increased serum levels of PFHxS and decreased antibody production
following routine vaccinations in children and adults.
Table 3-11. Summary of PFHxS and data on antibody response to vaccines in
children
Reference, N,
confidence
PFHxS exposure
timing and
concentration in
serum
Outcome measure timing
Effect estimate as
specified
Effect estimate as
specified a
Diphtheria vaccine
(% change in antibodies
with increase in PFHxS)
Tetanus vaccine
(% change in
antibodies with
increase in PFHxS)3
Maternal; mean
(IQR): 4.4 (2.2-
8.4) ng/mL
Children (age 5), prebooster
-6.4 (-16.0 to 4.3)
-6.3 (-15.1 to 3.4)
Children (age 5),
postbooster
-3.7 (-14.1 to 7.9)
6.3 (-8.4 to 23.2)
3-85
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference, N,
confidence
PFHxS exposure
timing and
concentration in
serum
Outcome measure timing
Effect estimate as
specified
Effect estimate as
specified a
Grandiean et al.
(2012), N = 380-
537, medium
Grandiean et al.
(2017a)
1997-2000 cohort
Children (age 7)
-0.5 (-13.1 to 14.0)
4.5 (-9.6 to 20.6)
Children (age 5);
mean (IQR): 0.6
(0.5-0.9) ng/mL
Children (age 5), prebooster
5.0 (-8.9 to 21.0)
-6.3 (-17.6 to 6.5)
Children (age 5),
postbooster
-9.1 (-18.7 to 1.7)
-19.0 (-29.8 to -6.6)
Children (age 7)
-9.8 (-22.3 to 4.9)
-19.7 (-31.6 to -5.7)
Children (age 7);
mean (IQR): 0.5
(0.4-0.7) ng/mL
Children (age 13)
-10.2 (-25.7 to 8.5)
14.8 (-13.3 to 52.2)
Children (age 13);
mean (IQR): 0.4
(0.3-0.5) ng/mL
Children (age 13)
-5.5 (-22.9 to 15.8)
8.7 (-18.5 to 45.0)
Grandiean et al.
(2017b)b, N = 349,
At birth, not
reported
Children (age 5), prebooster
-3.33 (-15.28 to 10.30)
-11.31 (-21.72 to
0.49)
medium
2007-2009 cohort
(unless specified)
Infant (18 m);
median (IQR): 0.2
(0.1-0.4) ng/mL
Children (age 5), prebooster
2007-2009 cohort
7.85 (-0.38 to 16.76)
1997-2000 cohort
-12.42 (-55.25 to 71.43)
2007-2009 cohort
-2.616 (-10.08 to
5.47)
1997-2000 cohort
-5.18 (-51.71 to
86.19)
Children (age 5);
median (IQR):0.3
(0.2-0.4) ng/mL
Children (age 5), prebooster
4.26 (-15.12 to 28.08)
-4.432 (-21.26 to
15.99)
Granum et al.
(2013), N =49,
medium
Maternal 0-3 d
post-delivery;
median:
0.3 ng/mL
Children (age 3)
n/a
0.07 (-0.03 to 0.18)
Granum et al.
(2013), N = 50,
medium
Maternal 0-3 d
post-delivery;
median:
0.3 ng/mL
Children (age 3)
-0.48 (-4.64 to 3.67)
n/a
Timmermann et
al. (2021),
N = 314, medium
Children (age 7-
12)
Children (age 7-12)
Adjusted for time since
vaccine booster,
breastfeeding duration
48 (1,115)
Additionally adjusted for
area of residence
-40 (-64,1)
Adjusted for time
since vaccine
booster,
breastfeeding
duration
28 (-6, 73)
Additionally adjusted
for area of residence
-28 (-53,10)
Maternal
-53 (-87, 73)
-1 (-72, 245)
Measles vaccine
P (95%)a
Rubella vaccine
P (95%)a
Children (<1 yr)
Children (<1 yr)
-5 (-23, 18)
NR
3-86
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference, N,
confidence
PFHxS exposure
timing and
concentration in
serum
Outcome measure timing
Effect estimate as
specified
Effect estimate as
specified a
Timmermann et
al. (2020),
N = 237, medium
0.1(0.1-0.1)
Children (2 yr)
After 1 vaccine (control
group)
-11 (-34,19)
After 2 vaccines
(intervention group)
10 (-18, 48)
NR
Granum et al.
(2013), N = 50,
medium
Maternal 0-3 d
post-delivery;
median:
0.3 ng/mL
Children (age 3)
-0.04 (-0.30 to 0.22)
-0.38 (-0.66 to
-0.11)
Stein et al.
(2016b),
N = 1,101-1,190,
medium
Children (age 12-
19); mean:
2.5 ng/mL
Children (age 12-19)
-2.8 (-10.1 to 5.21)
(seropositive)
-6.0 (-9.6 to -2.2)
(seropositive)
Hib vaccine
P (95%)a
Mumps vaccine
P (95%)a
Granum et al.
(2013), N = 50,
medium
Maternal 0-3 d
post-delivery;
median:
0.7 ng/mL
Children (age 3)
-0.48 (-4.64 to 3.67
n/a
Stein et al.
(2016b),
N = 1,101-1,190,
medium
Children (age 12-
19); mean:
2.5 ng/mL
Children (age 12-19)
n/a
-2.3 (-5.5 to 0.9)
Bold font indicates p < 0.05.
One study did not report quantitative results. Abraham et al. (2020) stated in text that there were no significant
correlations of levels of PFHxS with levels of the vaccine antibodies for Hib, tetanus, or diphtheria.
aLinear regression (P or % change in antibody per twofold increase of PFHxS). Numbers in parentheses are 95%
confidence intervals.
bResults for Faroe Islands Cohort 5 (2007-2009) unless otherwise stated.
3-87
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-12. Summary of PFHxS and data on antibody response to vaccines in
adults
Reference, N,
confidence
Exposu retiming
and concentration
in serum/plasma
(ng/mL)
Outcome
measure
timing
Diphtheria vaccine
P (95%)a
Tetanus vaccine
P (95%)a
FluMist (A H1N1)
vaccine
Seroconversion RR
(95% CI)
Shih et al.
(2021), Faroe
Islands,
N = 281,
medium
Cord blood;
median (IQR) 0.2
(0.2)
Adults (age
28)
Total: 13.57 (-2.4,
32.15)
Women: 12.94
(-6.42, 36.32)
Men: 14.72 (-10.98,
47.82)
Total: 0.63 (-10.86,
13.6)
Women: 0.58 (-13.47,
16.91)
Men: 0.74 (-17.78,
23.43)
n/a
Children (age 7);
0.9 (0.4)
Total: 1.96 (-18.98,
28.31)
Women: -18.74
(-43.42, 16.68)
Men: 17.48 (-11.86,
56.59)
Total: 3.23 (-13.22,
22.79)
Women: -8.27 (-30.54,
21.15)
Men: 11.01 (-10.78,
38.13)
Children (age 14);
0.6 (0.4)
Total: -7.62 (-37.93,
37.48)
Women: -8.03
(-47.08, 59.84)
Men:-7.20 (-47.17,
62.98)
Total: -10.24 (-35.99,
25.87)
Women: -17.92
(-48.63, 31.14)
Men: -1.37 (-39.02,
59.53)
Adults (age 22);
0.5 (0.4)
Total: -8.44 (-27.27,
15.27)
Women: -15.68
(-36.26, 11.55)
Men: 8.32 (-27.37,
61.54)
Total: -3.47 (-19.88,
16.3)
Women: -10.25
(-28.45, 12.57)
Men: 11.85 (-18.98,
54.4)
Kielsen et al.
(2016), N = 12,
low
Adult (10 d post
vaccination);
median (IQR): 0.4
(0.3-0.7)
Adult-
change
from 4 d to
10 d
postvaccina
tion
-13.31 (-25.07, 0.29)
-4.35 (-13.72 to 6.04)
n/a
Stein et al.
(2016a),
N = 75,
low
Adult (18-49 yr
old), d of
vaccination; mean:
1.1
Adult (18-
49 yr old),
30 d
postvaccina
tion
n/a
n/a
by hemaglutinin
inhibition:
T2: 1.2 (0.2, 6.5)
T3: 3.1 (0.8, 12.7)
by immuno-
histochemistry:
T2: 1.1 (0.4, 2.9)
T3: 1.7 (0.6, 4.8)
Bold font indicates p < 0.05.
aLinear regression (P or % change in antibody per two-fold increase of PFHxS). Numbers in parentheses are 95%
confidence intervals.
3-88
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Infectious disease
Direct measures of infectious disease incidence or severity such as respiratory tract
infections, pneumonia or otitis media are useful for evaluating potential immunotoxicity in humans.
Increases in incidence or severity of infectious disease can be a direct consequence of impaired
immune function whether the specific functional deficit has been identified or not Given the clear
adversity of most infectious diseases, they are generally considered good measures for how
immunosuppression can affect individuals and communities. Physician diagnosis is the most
specific way to assess infectious diseases, but these are usually only available for severe diseases
and are less likely for diseases where treatment is not sought. Self-reported incidence or severity of
disease may be less reliable but may be the only way to assess diseases such as the common cold or
gastroenteritis which while less adverse, are more common and can thus provide information about
immunosuppression and susceptibility to more severe infections. In general, symptoms of infection
alone are not considered reliable measures of disease because of their lack of specificity. Antibody
levels in response to infection are also included in this section (differentiated from antibody levels
in response to vaccination, described above); the utility of these measures depends on the study
design and population due to various factors such as potential confounding and prevalence of
infection.
Ten studies examined infectious disease occurrence in children, including eight prospective
birth cohorts one cohort with exposure measurement in childhood, and one cohort examining
antibody response to Hand, Foot, and Mouth Disease (HFMD) infection in the first 3 months of life.
In addition, two studies examined infectious disease occurrence in adults, including a cross-
sectional study of COVID-19 illness severity fGrandiean etal.. 20201 and a cross-sectional study of
antibody levels in response to several persistent infections (Bulka etal.. 20211.
Study evaluations are summarized in Figure 3-15. Of the studies in children, four studies in
Japan (Goudarzi etal.. 2017). Spain (Manzano-Salgado etal.. 2019). Denmark (Dalsager etal..
2021a). and China (Wang etal.. 2022) were medium confidence, and the remaining studies were
low confidence (Zenget al.. 2019b: Kvalem etal.. 2020: Impinenetal.. 2018: Impinenet al.. 2019:
Granum et al.. 2 013: Dalsager etal.. 2016). The low confidence birth cohorts were rated as
"deficient" in outcome ascertainment due to relying on parental self-report of incidence of common
infections or symptoms, with no validation of the measures. However, because the parents are
unlikely to know their child's exposure level, this misclassification is likely to be nondifferential
with respect to exposure. In contrast, the medium confidence studies assessed physician-diagnosed
conditions and were limited to more severe illnesses (otitis media, pneumonia, varicella, and
respiratory syncytial viral infection), which likely have better parental recall. Zeng etal. (2019b)
was low confidence because the outcome is difficult to interpret in infants and there are concerns
for confounding by timing of HFMD infection as well as other limitations. The two studies in adults
were both considered medium confidence. Grandiean et al. f20201 used biobank samples and
national registry data in Denmark to examine severity of COVID-19 illness severity. There was some
3-89
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
concern for selection bias in this study due to the expectation that biobank samples were more
likely to be available for individuals with chronic health concerns. In addition, severity of COVID-19
is not a direct measure of immune suppression as other factors may contribute to illness severity.
Bulka, 2021, 7410156-
Dalsager, 2016, 3858505-
Dalsager, 2021, 7405343-
Goudarzi, 2016, 3859523-
Grandjean, 2020, 7403067 -
Granum, 2013, 1937228-
Impinen, 2018, 4238440-
Impinen, 2019, 5080609-
Kvalem, 2020, 6316210-
Manzano-Salgado, 2019, 5412076-
Wang, 2022, 10176501 -
Zeng, 2019, 5081554-
++
+
+
+
+
+
+
+
+
++
-
+
+
-
+
-
+
++
+
+
+
-
-
+
~
+
H
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 3-15. Summary of evaluation of epidemiology studies of PFHxS and
infectious disease immunosuppression effects. For additional details see HAWC
link.
Two studies (Impinen et al., 2018; Granum et al., 2013) were sub-samples of the Norwegian Mother and Child
(MoBa) cohort. The cohort sub-samples for these publications were different, so their study evaluations and
results are reported independently, but it is possible that there is some overlap in the participants. Two studies
(Dalsager et al., 2016; Dalsager et al., 2021a) were both analyses of the Odense Child Cohort. They were
evaluated separately due to their different samples and outcome measurement methods but were not
considered fully independent samples.
In children, higher odds of infectious disease with higher PFHxS levels were reported in two
of the four medium confidence studies fWang etal.. 2022: Goudarzi et al.. 2 0171 and three of the six
low confidence studies f Impinen et al.. 2019: Granum etal.. 2013: Dalsager etal.. 20161 (see T able
3-13). Wang etal. f20221 reported higher odds (though not statistically significant) of upper and
lower respiratory infection and diarrhea with higher exposure. (Goudarzi etal.. 2017) reported
higher odds of total infectious disease from birth to age 4, but only in girls, and a significant trend
was observed, but the association was nonmonotonic across quartiles. No clear explanation for why
3-90
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
these results might vary by sex is available, and none of the other studies of immunosuppression
analyzed the results stratified by sex. Impinen etal. f20191 also reported higher odds of
gastroenteritis (statistically significant from birth to age 3), but not common cold or otitis media.
Dalsager etal. f 20161 reported higher odds of diarrhea and fever (p > 0.05), but not cough or nasal
discharge. Another medium confidence study fManzano-Salgado etal.. 20191 reported an
association in the same direction, but the effect estimate was small and imprecise. Two other low
confidence studies did not observe an association between maternal PFHxS concentrations and
infections. In adults and adolescents, one study found higher persistent pathogen burden with
higher exposure fBulka etal.. 20211. In contrast, there an inverse association between PFHxS
exposure and COVID-19 illness severity. Overall, many of the studies had limited sensitivity due to
narrow exposure contrast, but there was no apparent relationship between higher study exposure
levels and observed associations. Given the inconsistency across studies, there is considerable
uncertainty in this outcome. The associations observed in some studies provide some limited
support for (and coherence with) the evidence of immunosuppression observed in the antibody
response studies.
Table 3-13. Summary of PFHxS and selected data on infectious disease in
humans
Disease
Reference,
confidence
Exposure measurement
timing and concentration
in serum/plasma (ng/mL)
Disease
assessment timing
PFHxS results
Total infectious
disease3
Dalsager et al.
(2021a), medium
Maternal; median: 0.4
From birth to age 4
HR (95% CI)
1.02 (0.90,1.16)
Goudarzi et al.
(2017)
medium
Maternal; median (IQR):
0.3 (0.2-0.4)
From birth to age 4
Adj OR (95% CI) Total:
Q2: 1.03 (0.764, 1.41)
Q3: 1.23 (0.905, 1.69)
Q4: 0.957 (0.703,1.30)
Trend p = 0.928
Male:
Q2: 0.780 (0.508,1.19)
Q3: 0.947 (0.614,1.45)
Q4: 0.708 (0.461,1.08)
Trend p = 0.223
Female:
Q2: 1.46 (0.938, 2.29)
Q3: 1.81 (1.14, 2.88)
Q4: 1.55 (0.976, 2.45)
Trend p = 0.045
Lower
respiratory tract
infection15
Impinen et al.
(2018)
low
Cord blood
From birth to age
10
Adj P (95% CI)
0.04 (-0.01, 0.09)
3-91
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Disease
Reference,
confidence
Exposure measurement
timing and concentration
in serum/plasma (ng/mL)
Disease
assessment timing
PFHxS results
Dalsager et al.
(2021a), medium
Maternal; median: 0.4
From birth to age 4
HR (95% CI)
1.01 (0.78,1.32)
Wang et al.
(2022), medium
Maternal; median (IQR):
0.6 (0.4-0.8)
Through Age 1
OR (95% CI)
10.62 (0.65, 173.7)
IRR (95% CI)
1.81 (0.27, 12.19)
Manzano-
Salgado et al.
(2019) medium
Maternal (1st trimester),
median (IQR): 0.6 (0.4—
0.8)
Age 1.5-7
1.07 (0.96,1.18)
Impinen et al.
(2019)
low
Maternal mid-pregnancy;
median (IQR): 0.7 (0.5-
0.9)
From birth to age 3
Adj RR (95% CI):
1.15 (1.06,1.24)
Age 6-7
0.92 (0.70,1.21)
Kvalem et al.
(2020) low
Child age 10; median
(IQR): 1.3 (0.9)
Age 10-16
Adj RR (95% CI)
0.98 (0.95,1.02)
Age 16 (last 12 m)
0.93 (0.74,1.18)
Gastroenteritis
(No. episodes/
frequency)
Granum et al.
(2013),
low
Maternal
0-3 d post-delivery;
median: 0.3
From birth to age 3
Adj P (95% CI)
3rd yr: 0.33 (-0.05, 0.71)
All 3 yr: 0.35 (0.10, 0.61)
Dalsager et al.
(2021a), medium
Maternal; median: 0.4
From birth to age 4
HR (95% CI)
0.85 (0.50,1.43)
Impinen et al.
(2019),
low
Maternal mid-pregnancy;
median (IQR): 0.7 (0.5-
0.9)
From birth to age 3
Adj (RR):
0.98 (0.96,1.02)
Age 6-7
1.27 (1.18,1.38)
Diarrhea
Dalsager et al.
(2016) low
Maternal; median
(range): 0.3 (0.02-1.0)
Age 1-3
OR for proportion of d with
symptoms (under/above
median)
Low exposure: Ref
Medium: 1.16 (0.66, 2.02)
High: 1.39 (0.77, 2.51)
IRR for number of d with
symptoms
Low exposure: Ref
Medium: 1.18 (0.64, 2.19)
High: 1.71 (0.92, 3.16)
Wang et al.
(2022), medium
Maternal; median (IQR):
0.6 (0.4-0.8)
Through age 1
OR (95% CI)
1.17 (0.20,6.83)
IRR (95% CI)
1.27 (0.50, 3.20)
Common cold
(No. episodes/
frequency)
Impinen et al.
(2018),
low
Cord blood; median (IQR):
0.2 (0.2-0.3)
From birth to age 2
Adj P (95% CI)
-0.01 (-0.04, 0.02)
3-92
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Disease
Reference,
confidence
Exposure measurement
timing and concentration
in serum/plasma (ng/mL)
Disease
assessment timing
PFHxS results
Granum et al.
(2013), low
Maternal
0-3 d post-delivery;
median: 0.3
From birth to age 3
Adj P (95% Cl)c
3rd yr: 0.24 (-0.03, 0.51)
All 3 yr: 0.15 (-0.02, 0.32)
Dalsager et al.
(2021a), medium
Maternal; median: 0.4
From birth to age 4
HR (95% CI) for upper
respiratory infections
1.01 (0.83,1.21)
Wang et al.
(2022), medium
Maternal; median (IQR):
0.6 (0.4-0.8)
Through Age 1
OR (95% CI)
1.49 (0.28, 7.97)
IRR (95% CI)
1.16 (0.60,2.26)
Impinen et al.
(2019), low
Maternal mid-pregnancy;
median (IQR): 0.7 (0.5-
0.9)
From birth to age 3
Adj RR (95% CI):
1.01 (1.00,1.03)
Kvalem et al.
(2020) medium
Child age 10; median
(IQR): 1.3 (0.9)
Age 10-16
Adj OR (95% CI):
Reference 1-2 colds
3-5 colds: 0.99 (0.93, 1.04)
>5: 0.97 (0.93,1.03)
Age 16 (last 12 m)
Adj OR (95% CI)
Reference 0 colds
1-2 colds: 0.98 (0.96,1.00)
>3: 0.97 (0.94,1.00)
Cough
Dalsager et al.
(2016) low
Maternal; median
(range): 0.3 (0.02-1.0)
Age 1-3
OR for proportion of d with
symptoms (under/above
median)
Low exposure: Ref
Medium: 1.04 (0.60,1.79)
High: 0.97 (0.54,1.73)
IRR for number of d with
symptoms
Low exposure: Ref
Medium: 1.14 (0.87,1.48)
High: 1.00 (0.76,1.31)
Ear infection
Granum et al.
(2013), low
Maternal
0-3 d post-delivery;
median: 0.3
From birth to age 3
No significant association
with otitis media (data not
shown)
Impinen et al.
(2019), low
Maternal mid-pregnancy;
median (IQR): 0.7 (0.5-
From birth to age 3
Adj RR (95% CI):
1.09 (1.04,1.14)
0.9)
Age 6-7
1.08 (0.93,1.25)
Throat infection
Impinen et al.
(2019), low
Maternal mid-pregnancy;
median (IQR): 0.7 (0.5-
0.9)
From birth to age 3
Adj RR (95% CI):
1.10 (1.02,1.18)
(no association with
streptococcus throat
infection)
3-93
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Disease
Reference,
confidence
Exposure measurement
timing and concentration
in serum/plasma (ng/mL)
Disease
assessment timing
PFHxS results
Pseudocroup
Impinen et al.
(2019), low
Maternal mid-pregnancy;
median (IQR): 0.7 (0.5-
0.9)
From birth to age 3
Adj RR (95% CI):
1.20 (1.11,1.30)
Fever
Dalsager et al.
(2016) low
Maternal; median
(range): 0.3 (0.02-1.0)
Age 1-3
OR for proportion of d with
symptoms (under/above
median)
Low exposure: Ref
Medium: 0.99 (0.58,1.71)
High: 1.29 (0.72, 2.28)
IRR for number of d with
symptoms
Low exposure: Ref
Medium: 1.07 (0.80,1.42)
High: 1.20 (0.89,1.62)
Hand Foot and
Mouth Disease
Virus Antibodies
Zeng et al.
(2019b), low
Cord; median (IQR): 4.0
(2.3-5.4)
Birth and age 3 mo
OR (95% CI) for HFMD
antibody concentration
below clinically protective
level
Cord blood:
1.08 (0.74,1.60)
3 mo: 1.00 (0.71,1.43)
COVID-19 illness
severity
Grandiean et al.
(2020), medium
Biobank prior to illness;
median (IQR):
0.5 (0.3-0.7)
Adulthood
OR (95% CI) for 1 unit
increase
Increased severity based on
hospitalization, admission
to intensive care and/or
death
0.52 (0.29, 0.93)*
Pathogen
burden of
persistent
infections based
on antibodies
Bulka et al.
(2021)
Mean: 1.5
Ages 12-49 yr
Relative difference (95% CI)
per doubling
12-19 yr: 1.11(1.06,1.15)*
20-49 yr: 1.02 (1.00,1.05)*
For individual pathogens,
onlyToxocara spp had
positive association
Bolded values are statistically significant,
includes Otitis media, pneumonia, RS virus, Varicella.
bLower respiratory tract infections include bronchitis, bronchiolitis, and pneumonia.
cBivariate model was statistically significant (p = 0.036) for all 3 years.
Sensitization or allergic response
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 flPCS. 20121. A chemical may be either a direct sensitizer (i.e., promote a specific
3-94
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
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.
While 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 measurements of the health
effects such as allergies or asthma and skin prick tests. Observational tests such as measures of
total IgE levels measure indicators of sensitivity and allergic responses but are not a direct
measurement of the response. The section is organized by the different types of measurements,
starting with functional measures as the most informative.
Thirteen studies (reported in 19 publications) examined hypersensitivity outcomes in
children. The study evaluations are summarized in Figure 3-16. Two of the included studies were
subsamples of the Norwegian Mother and Child (MoBa) cohort that were analyzed independently
(Impinen et al.. 2019: Granum etal.. 2013). In addition, three publications of NHANES data are
grouped together as one study because there is significant overlap in the NHANES years included in
the analysis samples (Steinetal.. 2016b: Humbletetal.. 2014: Buser and Scinicariello. 2016):
another publication examined a different year range of NHANES data and was considered
separately flackson-Browne etal.. 20201. Ten studies were prospective birth cohorts, with
exposure measured during gestation or in cord blood. These studies were performed in China
fChen etal.. 2018al. Japan fOkada etal.. 2014: Goudarzi etal.. 20161. Norway f Impinen et al.. 2 018:
Impinen et al.. 2019: Granum etal.. 2013). Greenland and Ukraine (Smitetal.. 2015). Spain
(Manzano-Salgado etal.. 2019). Denmark (Beck etal.. 2019). and the Faroe Islands (Timmermann
etal.. 2017a). In addition to the cohort studies, there was a case-control study of asthma in Taiwan
reported in multiple publications (Zhu etal.. 2016: Zhou etal.. 2017b: Dong etal.. 2013). a cohort of
children with exposure measured at age 10 fKvalem etal.. 20201. and the analyses of NHANES data,
which is cross-sectional. All the studies were considered medium confidence.,
3-95
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Beck, 2019, 5922599-
Chen, 2018, 4238372-
Dong, 2013, 1937230
Goudarzi, 2016, 3859523-
Granum, 2013, 1937228
Impinen, 2018, 4238440
Impinen, 2019, 5080609
Jackson-Browne MS et al. 2020
Kvalem, 2020, 6316210
Manzano-Salgado, 2019, 5412076-
Smit, 2015,2823268-
Stein, 2016, 3108691
Timmermann, 2017, 3858497-
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 3-16. Summary of evaluation of epidemiology studies of PFHxS and
hypersensitivity effects (e.g., asthma, allergies, and atopic dermatitis). For
additional details see HAWC link.
Multiple publications of the same data are presented on the heat map as one study. Goudarzi et al. (2016) also
includes Okada et al. (2014). Stein et al. (2016b) also includes Buser and Scinicariello (2016) and Humblet et al.
(2014).
Asthma
Twelve studies evaluated different measures related to asthma diagnosis and symptoms in
relation to PFHxS exposure (see Table 3-14). All studies were medium confidence. One study
examined asthma incidence (i.e., diagnosis within the past year, with cases identified from two
hospitals), which is the most specific measure available across studies, but which may result in
under-ascertainment because only severe cases are identified. The remaining studies examined
asthma prevalence using validated questionnaires, either "current" asthma (generally experiencing
symptoms in the past year with asthma diagnosis) or "ever" asthma (asthma diagnosis at any time
during their life). These measures are less specific than asthma incidence and the relevant etiologic
period is less clear.
Four studies examined "current" asthma and 11 studies examined "ever" asthma. Looking at
current asthma, one study (Impinen etal.. 2019) out of four reported higher odds, although this was
not statistically significant. Three studies also reported a positive association with "ever" asthma,
but with inconsistency within each study. Zeng etal. f2019al reported a strong positive, but very
3-96
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
imprecise, association in boys, and an imprecise inverse association in girls, while in Beck et al.
f20191. a strong positive association (p < 0.05) was observed in girls for doctor-diagnosed asthma,
but there was no sex-interaction with self-reported asthma. In Timmermann etal. f2017al. a
positive association was observed only in a small subgroup (4%, 22 children) of the study
population that did not receive MMR vaccination and may be due to chance. The remaining studies
showed no association with ever asthma.
The single study (reported in multiple publications) of asthma incidence (the most specific
outcome measurement available) reported higher odds of asthma in children 10-15 years of age
with higher PFHxS exposure with an exposure-response gradient observed across quartiles in the
overall population fDongetal.. 20131. The association was stronger in girls than in boys fZhu etal..
20161. although there was no significant interaction with sex hormone levels fZhou etal.. 2017bl.
The association was strong (OR >3 in highest quartile of exposure), and the outcome measurement
is likely to suffer from less outcome misclassification than would measures of asthma prevalence in
the other available studies. This medium confidence study in Taiwan also had PFHxS exposure
levels that were among the highest of the available studies, while several studies with null results
had exposure levels with narrow exposure contrast across participants, which may have reduced
sensitivity. While there is considerable uncertainty due to inconsistency in the results across
studies, the null results are not interpreted as contradictory to the positive findings given the better
sensitivity and specificity (and relatively higher exposure levels) in Dong etal. (2013).
Allergies/Allergic sensitization
Five studies, all medium confidence, evaluated allergies and allergic sensitization outcomes
(see Table 3-14). Two studies examined food allergies. Buser and Scinicariello f20161. an NHANES
analysis, reported higher odds of allergy in the second and fourth quartiles, with statistical
significance in the fourth quartile. Impinen etal. f20191 observed slightly higher, but not
statistically significant odds of current food allergies with higher exposure. Impinen etal. (2019)
also found higher, but not significant, odds of inhaled allergies. Four studies examined allergic
sensitization, and one study observed higher odds of elevated IgE with higher exposure, although
this was not monotonic as the highest odds were in the third quartile (Buser and Scinicariello.
20161. The other NHANES analysis fStein etal.. 2016bl and three other studies did not report
higher odds of sensitization with higher exposure.
Dermal allergic measures - eczema
Nine studies evaluated eczema (see Table 3-14). While the studies used different
terminology including eczema, atopic eczema, and atopic dermatitis, most assessed presence of an
itchy rash that was coming and going for at least 6 months using the International Study of Asthma
and Allergies in Childhood questionnaire. Three studies examined physician-diagnosed atopic
eczema, also collected using a questionnaire f Impinen et al.. 2 018: Impinen etal.. 2019: Granum et
al.. 20131. and Kvalemetal. f20201 used a different questionnaire for self-reported eczema. These
3-97
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
dermal response conditions can represent hypersensitivity to an antigen exposure from any route.
Two medium confidence studies reported higher odds of eczema with higher PFHxS exposure
fTimmermann etal.. 2017a: Chen etal.. 2018a! both statistically significant (in girls only for Chen
etal. f2018a]]. while two studies fOkada etal.. 2014: Kvalem etal.. 20201 reported an inverse
association. The remaining five studies reported no association. Exposure levels were highest in
Timmermann etal. (2017a). but levels in Chen etal. (2018a) were similar to the null studies, and
Okada etal. (2014). There is no apparent explanation for the inconsistency across studies on the
basis of study design, population, bias, or other factors.
Table 3-14. Summary of PFHxS and data on hypersensitivity in humans.
Reference
Exposu re
measurement timing
and concentration in
serum/plasma (ng/ml)
Hypersensitivity
measurement
timing
PFHxS OR (95% Cl)a or as
specified
Asthma incidence
GBCA
Dong et al. (2013)
Children, current;
median (IQR): 1.3 (0.6—
2.8) (without asthma)
Children (age 10-
15)
Asthma diagnosed in past yr
Q2: 1.54 (0.85, 2.77)
Q3: 2.94 (1.65, 5.25)
Q4: 3.83 (2.11, 6.93)
Trend p< 0.001
Zhou et al. (2017b)
Children (age 10-
15)
By Sex Hormone Levels
Low Testosterone
M: 2.12 (1.34,3.35)
F: 1.62 (1.08, 2.45)
High Testosterone
M: 1.43 (0.99,2.07)
F: 2.27 (1.29, 3.99)
Low Estradiol
M: 1.47 (1.00,2.15)
F: 2.39 (1.39, 4.12)
High Estradiol
M: 1.62 (1.01,2.60)
F: 1.65 (1.07, 2.55)
No significant interaction
between PFHxS and sex
hormone category
Zhu etal. (2016)
Children (age 10-
15)
By Sex
Q4 vs. Q1
M: 2.97 (1.33,6.64)
F: 5.02 (2.05,12.30)
Current asthma
Impinen et al. (2019)
Maternal mid-
pregnancy; median
(IQR): 0.7 (0.5-0.9)
From birth to age 7
1.21 (0.87,1.67)
3-98
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Exposu re
measurement timing
and concentration in
serum/plasma (ng/ml)
Hypersensitivity
measurement
timing
PFHxS OR (95% Cl)a or as
specified
Impinen et al. (2018)
Cord blood; median
(IQR): 0.2 (0.2-0.3)
From birth to age
10
0.99 (0.82,1.21)
Kvalem et al. (2020)
Child (age 10); median
(IQR): 1.3 (0.9)
Child (age 16)
Last 12 mo
RR: 1.00 (0.98, 1.02)
NHANES
Stein et al. (2016b)
Children, current;
mean: 2.5
Children (age 12-
19)
IQR increase:
0.98 (0.51,1.87)
Ever asthma
Zeng et al. (2019a)
Cord blood median
(IQR): 0.2 (0.1-0.2)
Child (age 5)
Ever asthma
2.02 (0.24, 17.24)
Girls: 0.48 (0.00, 85.33)
Boys: 3.40 (0.18, 65.11)
MoBa
Granum et al. (2013)
Maternal
0-3 d post-delivery;
median: 0.3
From birth to age 3
No significant association
(data not shown)
Impinen et al. (2019)
Maternal mid-
pregnancy; median
(IQR): 0.7 (0.5-0.9)
From birth to age 7
0.96 (0.79,1.18)
Beck et al. (2019)
Maternal, gestational
wk 8-16; median
(IQR): 0.4 (0.2-0.5)
Child (age 5)
Ever doctor-diagnosed
asthma
1.16 (0.78,1.71)
Boys: 0.89 (0.59,1.34)
Girls: 2.96 (1.26, 6.96)
Ever self-reported asthma
(>episodes of wheezing
lasting more than a d in past
12 mo)
1.18 (0.73,1.90)
Boys: 1.33 (0.66, 2.71)
Girls: 1.04 (0.55,1.98)
Manzano-Salgado et al. (2019)
medium
Maternal (1st
trimester), median
(IQR): 0.6 (0.4-0.8)
Age 1.5-7
Ever asthma
RR: 0.96 (0.74, 1.24)
Jackson-Browne et al. (2020)
Child (age 3-11); mean
(IQR): 0.8 (0.5-1.3)
Child (age 3-11)
Ever asthma
OR: 1.1 (0.9,1.3)
Kvalem et al. (2020)
Child (age 10); median
(IQR): 1.3 (0.9)
Child (age 10)
Ever asthma
RR: 0.99 (0.97, 1.01)
Child (age 10-16)
Asthma between 10 and 16 yr
RR: 1.00 (0.99, 1.02)
Smit et al. (2015)
Maternal, mean
gestational wk 24 or
Children (age 5-9)
0.91 (0.69,1.18)
3-99
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Exposu re
measurement timing
and concentration in
serum/plasma (ng/ml)
Hypersensitivity
measurement
timing
PFHxS OR (95% Cl)a or as
specified
25; mean (5th—95th):
Ukraine: 1.5 (0.5-4.1),
Greenland: 2.1 (1.0—
5.1)
Impinen et al. (2018)
Cord blood; median
(IQR): 0.2 (0.2-0.3)
From birth to age
10
0.94 (0.72,1.21)
Timmermann et al. (2017a)
Maternal, gestational
wk 34-36; median
(IQR): 4.5 (2.2-8.3)
Child (age 5)
0.99 (0.80,1.22)
Child (age 13)
0.98 (0.79,1.20)
Child (age 5); median
(IQR): 0.6 (0.4-0.9)
Child (age 5)
No MMR: 3.57 (0.95, 13.43)b
Yes MMR: 0.81 (0.58, 1.14)
Interaction p = 0.03
Child (age 13)
No MMR: 2.52 (0.77, 8.16)b
Yes MMR: 0.90 (0.63,1.27)
Interaction p = 0.10
Child (age 13); median
(IQR): 0.4 (0.3-0.5)
Child (age 13)
0.63 (0.41, 0.97)
NHANES
Humblet et al. (2014)
Children, current;
median (IQR): 2.0 (1.0,
4.1)
Children (age 12-
19)
Continuous: 0.98 (0.88-1.08)
T2: 1.07 (0.89,1.30)
T3: 0.92 (0.74,1.14)
Allergies (food)
Impinen et al. (2019)
Maternal mid-
pregnancy; median
(IQR): 0.7 (0.5-0.9)
From birth to age 7
Ever: 1.03 (0.82,1.30)
Current: 1.10 (0.86,1.41)
NHANES
Buser and Scinicariello (2016)
Children, current;
mean: 2.2
Children (age 12-
19)
Q2 1.43 (0.40,5.14)
Q3 0.99 (0.37, 2.65)
Q4 3.06 (1.35, 6.93)
Trend p = 0.11
Allergies (inhaled)
Impinen et al. (2019)
Maternal mid-
pregnancy; median
(IQR): 0.7 (0.5-0.9)
From birth to age 7
Ever: 1.18 (0.93,1.50)
Current: 1.21(0.81,1.81)
Allergies (sensitization)
Impinen et al. (2018)
Cord blood; median
(IQR): 0.2 (0.2-0.3)
From birth to age
10
Positive SPT or slgE > 0.35
kU/L
1.01 (0.84,1.21)
Kvalem et al. (2020)
Child (age 10); median
(IQR): 1.3 (0.9)
Child (age 10)
Positive skin prick test
RR: 1.01 (1.00, 1.02)
3-100
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Exposu re
measurement timing
and concentration in
serum/plasma (ng/ml)
Hypersensitivity
measurement
timing
PFHxS OR (95% Cl)a or as
specified
Child (age 16)
Positive skin prick test
RR: 1.00 (1.00, 1.01)
Timmermann et al. (2017a)
Maternal, gestational
wk 34-36; median
(IQR): 4.5 (2.2-8.3)
Children (age 13)
Positive skin prick test
0.94 (0.79,1.12)
Children (age 5)
Children (age 13)
Positive skin prick test
0.95 (0.75,1.20)
Child (age 5); median
(IQR): 0.6 (0.4-0.9)
Children (age 13)
Positive skin prick test
0.88 (0.64,1.21)
NHANES
Buser and Scinicariello
(2016)
Children, current;
mean: 2.2
Children (age 12-
19)
Sensitization (any slgE >0.35
kU/L)
Q2 1.11(0.66,1.88)
Q3 1.46 (0.79, 2.69)
Q4 1.17 (0.56,2.44)
Trend p = 0.72
Stein et al. (2016b)
Children, current;
mean: 2.5
Children (age 12-
19)
Sensitization (any slgE >0.35
kU/L)
IQR increase: 0.92 (0.66,1.28)
Eczema
MoBa
Granum et al. (2013)
Maternal
0-3 d post-delivery;
median: 0.3
From birth to age 3
Eczema and itchiness or
doctor-diagnosed atopic
eczema:
No significant association
(data not shown)
Impinen et al. (2019)
Maternal mid-
pregnancy; median
(IQR): 0.7 (0.5-0.9)
From birth to age 7
Ever: 1.09 (0.90,1.31)
Current: 1.06 (0.83,1.36)
Hokkaido
Goudarzi et al. (2016)
Maternal, gestational
wk 28-32; median
(IQR): 0.3 (0.2-0.4)
Children (age 4)
Ever:
Q2: 0.953 (0.658,1.38)
Q3: 0.910 (0.623,1.32)
Q4: 0.917 (0.626,1.34)
Trend p = 0.618
Okada et al. (2014)
Children (age 1 or
2)
Ever:
Q2 0.82 (0.60,1.13)
Q3 0.69 (0.50, 0.95)
Q4 0.79 (0.57,1.08)
Trend p = 0.08
Smit et al. (2015)
Maternal, gestational
wk 24
Children (age 5-9)
Ever: 1.03 (0.86,1.24)
Current: 0.93 (0.73,1.20)
Chen et al. (2018a)
Children (age 2)
Ever:
3-101
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Exposu re
measurement timing
and concentration in
serum/plasma (ng/ml)
Hypersensitivity
measurement
timing
PFHxS OR (95% Cl)a or as
specified
Cord blood; median
(IQR): 0.2 (0.2-0.2)
1.08 (0.62,1.85) per log-unit
increase
Q2 1.25(0.74,2.12)
Q3 1.15 (0.68,1.94)
Q4 1.14 (0.67,1.94)
Trend p = 0.73
Females only
Q2 1.43 (0.62, 3.30)
Q3 1.29 (0.55, 2.99)
Q4 2.30 (1.03, 5.15)
Trend p = 0.06
Impinen et al. (2018)
Cord blood; median
(IQR): 0.2 (0.2-0.3)
From birth to age
10
0-2 yr of age
1.06 (0.89,1.26)
Ever in 10 yr
1.00 (0.67,1.49)
Manzano-Salgado et al. (2019)
Maternal (1st
trimester), median
(IQR): 0.6 (0.4-0.8)
Age 1.5-7
Ever eczema
RR: 0.95 (0.86, 1.05)
Kvalem et al. (2020)
Child (age 10); median
(IQR): 1.3 (0.9)
Child (age 10)
Ever doctor diagnosed:
RR: 1.00 (0.98, 1.01)
Child (age 10-16)
Ever between 10 and 16 yr
RR: 0.79 (0.34, 0.99)
Child (age 16)
Current (last 12 mo)
RR: 0.78 (0.60, 1.02)
Timmermann et al. (2017a)
Maternal, gestational
wk 34-36; median
(IQR): 4.5 (2.2-8.3)
Children (age 13)
1.32 (1.08,1.62)
Children (age 5)
Children (age 13)
0.92 (0.70-1.22)
Child (age 5); median
(IQR): 0.6 (0.4-0.9)
Children (age 13)
No MMR: 1.27 (0.16, 10.15)c
Yes MMR: 0.80 (0.53, 1.20)
Interaction p = 0.66
Bold font indicates p < 0.05.
aAII estimates are presented as OR (95% CI) for the odds of the outcome per twofold increase in PFHxS
concentration unless otherwise stated.
bResults provided broken down by MMR vaccination status; yes (n = 537) or no (n = 22) when provided; some
results were not split by MMR vaccination status.
Animal Studies
Animal toxicity studies examining the effects of PFHxS on the immune system include two
(high confidence) short-term oral exposure studies performed in Sprague Dawley rats, fNTP.
3-102
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
2018a: 3M. 2000b], one (high confidence) multigenerational study in Sprague Dawley rats
fButenhoffetal.. 20091. and one (medium confidence due to lack of results presentation)
subchronic oral exposure study performed in Crl:CDl mice fChang etal.. 20181: the study details
are provided in Table 3-15. It should be noted that none of the studies in the database were
immunotoxicity-specific studies, but rather short-term or subchronic studies that focused on
reproductive endpoints but also measured general immune-related endpoints. IPCS guidance states
that a 28-day exposure period, such as those in the three studies in the evidence base, are adequate
to elicit an immune response (IPCS. 20121. The immune-relevant endpoints evaluated in these
studies include immune hematology (i.e., blood leukocyte counts), histopathology, and organ
weights (i.e., bone marrow, lymph nodes, spleen), which may inform sensitization and allergic
response and autoimmunity, categories of immunotoxicity described in guidance from the
International Programme on Chemical Safety flPCS. 20121.9 Studies were separately evaluated for
each of these endpoints; however, the overall confidence rating was the same regardless of
endpoint (see Figure 3-17; for study details please see Table 3-15 and HAWC).
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
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)
NRl Not reported
Figure 3-17. Study evaluation results of PFHxS animal toxicity studies with
immune-related endpoints. For additional details see HAWC link.
9IPCS guidance notes that "the dataset[s] for most chemicals is unlikely to contain all the data on all the
described endpoints" (IPCS. 20121.
3-103
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-15. Animal study details
Study
Experimental
model
Exposure route
Exposure doses
Duration
Immune endpoint(s)
3M (2000b)
Male and female
SD rats
Oral gavage
0, or 10 mg/kg-d
28 d
Total immune cell
counts3
histopathology, organ
weights
Butenhoff et al. (2009)
Male and female
SD rats
Oral gavage
0 or 10 mg/kg-d
F0:15 rats per sex and treatment group
(control, 0.3,1, 3, and 10 mg/kg-day) were
dosed with PFHxS or vehicle via gavage 14 d
prior to cohabitation, during cohabitation,
and until the day before euthanasia (21 d of
lactation or presumed gestation day 25 (if not
pregnant) for females and minimum of 42 d
of treatment for males). F1 offspring were not
dosed by gavage but were exposed by
placental transfer in utero and potentially
exposed via milk.
Histopathology and
organ weights
Chang et al. (2018)
Male and female
CD-I Mice
Oral gavage
0, 0.3,1, or 3 mg/kg-d
F0: Males: dosing started 14 d prior to
cohabitation for a total of 42 d until scheduled
to be euthanized. Females: dosing started 14 d
prior to cohabitation and continuing through
mating, gestation, and lactation. F0 dams were
euthanized on lactation day 22 (LD 22), which
was 1 d post-last dose. Fl: Mice were exposed
in utero and via lactation. After weaning at
postnatal d 22, pups were directly dosed with
PFHxS for an additional 14 d at the same
respective maternal doses.
Total Leukocyte counts'5
histopathology,0
organ weights
NTP(2018a)
Male and female
SD rats
Oral gavage
Males: 0, 0.625,1.25,
2.5, 5 or 10 mg/kg-d
Females: 0, 3.12, 6.25,
12.5, 25 or 50 mg/kg-d
28 d
Total immune cell counts
histopathology,
organ weights
aTotal immune cell count included detailed counts of immune cells, e.g., basophil, eosinophil counts.
bTotal leukocyte count does not include detailed counts of immune cells.
cData not shown.
3-104
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Immune hematology
A summary of the immune hematology outcomes can be found in Figure 3-18. Briefly, of the
three studies that examined immune outcomes, two[ 3M f2000bl and NTP f2018al] performed a
complete detailed analysis of blood leukocyte counts including basophils, eosinophils, leukocytes,
lymphocytes, monocytes, and neutrophils, while Chang etal. (2018) reported only total blood
leukocyte counts. 3M (2000b) and Chang etal. (2018) reported no statistically significant changes
in white blood cell counts in response to PFHxS exposure while NTP (2018a). observed a
statistically significant decrease (p < 0.05) in eosinophil counts at the 10 mg/kg-day dose in male
but not in female SD rats. However, there were no other statistically significant changes in immune
hematology parameters, and the inconsistency in findings across the two rat studies is not
explained by dose or duration of exposure, or rat strain.
Endpoint Name Study Name Study Design Animal Description Trend Test Result
Basophil Count (BASO) 3M, 2000, 3981194 28 Day Oral Rat, Crl:Cd Br ($) not reported
Rat, Crl:Cd Br (-?) not reported
9 Dose
A Significant Increase
•
•
NTP, 2018,4309363 28 Day Oral Rat, Sprague-Dawley (¦') not significant
Rat, Sprague-Dawley ( j) not significant
V Significant Decrease
Eosinophil Count (EO) 3M, 2000,3981194 28 Day Oral Rat, Crl:Cd Br (?) not reported
Rat, Crl:Cd Br (o) not reported
NTP, 2018,4309363 28 Day Oral Rat, Sprague-Dawley (->) not significant
Rat, Sprague-Dawley (o) significant
t-
•
•
w
Leukocytes, Total 3M, 2000,3981194 28 Day Oral Rat, Crl:Cd Br (i) not reported
Rat, Crl:Cd Br ( ?) not reported
•
•
NTP, 2018,4309363 28 Day Oral Rat, Sprague-Dawley (2) not significant
Rat, Sprague-Dawley (;;) not significant
Lymphocyte Count (LYMPH) 3M, 2000,3981194 28 Day Oral Rat, CrlrCd Br (y) not reported
Rat, Crl:Cd Br (if) not reported
•
•
NTP, 2018,4309363 28 Day Oral Rat, Sprague-Dawley (2) not significant
Rat, Sprague-Dawley (j) not significant
Monocyte Count (MONO) 3M, 2000,3981194 28 Day Oral Rat, CitCd Br (?) not reported
Rat, CrkCd Br (
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Histopathologv
Four studies, 3M f2000bl. Butenhoff et al. f20091. NTP f2018al. and Chang etal. f20181.
performed histological analyses of immune organs and tissues, including bone marrow, lymph
nodes, spleen, and thymus. All four studies reported that they found no PFHxS-related histological
abnormalities in the immune organs and tissues that they examined although specific results were
not reported.
Organ weights
Four studies, 3M f2000bl. Butenhoff et al. f20091. NTP f2018al. and Chang etal. f20181.
measured thymus and spleen weights of control and exposed animals, and no PFHxS-related effects
were observed.
Mechanistic Evidence and Supplemental Information
Most of the mechanistic evidence available relates most closely to potential sensitization or
allergic response outcomes. Specifically, five studies examined mechanistic endpoints related to
hypersensitization in the human studies. None of the five studies reported significant associations
between PFHxS and IgE fZhu etal.. 2016: Timmermann etal.. 2017a: Stein etal.. 2016b: Dong etal..
2013: Ashlev-Martin et al.. 20151. Among asthmatics in the Taiwan population where an association
was observed with asthma, increases in eosinophilic cationic protein concentration were
significantly associated (p = 0.004) with increasing PFHxS concentration (Dong etal.. 2013). In
addition, one study examined cord blood gene expression in relation to PFHxS levels and found that
gene changes associated with PFHxS tracked very well with a set of 27 gene changes associated
with common cold episodes (Pennings etal.. 2016): however, changes with PFHxS tracked very
poorly with a second set of 26 gene changes associated with rubella titers, and the relevance of
these gene changes to immune function in general, or antibody responses in particular, remains
unknown.
No mechanistic evidence from animal, in vitro, in silico, or other evidence streams was
identified. However, PFHxS-induced alterations in thyroid hormones may play a role in the immune
effects described above as T3 and T4 play a role in the development and normal functions of the
immune system (U.S. EPA. 2006: Montesinos and Pellizas. 2019: De Vito etal.. 2011) and conditions
such as gestational hypothyroxinemia can disrupt normal immune functions fRivera etal.. 2024:
Funes etal.. 20221. Additional research would be needed to understand potential association
between PFHxS-induced alterations in thyroid hormones and downstream alterations in immune
system development and functions.
Evidence Integration
Human studies provide moderate evidence for immune system effects following exposure to
PFHxS (see Table 3-16). Specifically, increased serum levels of PFHxS correlated with decreased
antibody responses were observed in most exposure-outcome timing combinations in multiple
3-106
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
medium confidence studies, although most results were imprecise (i.e., not statistically significant).
While variability in response by age of exposure and outcome measure (vaccine type), as well as
timing of vaccinations (initial and boosters), resulted in some uncertainty, decreases (generally
between 5% and 10%) in antibody concentration per doubling of PFHxS concentration were
observed with reasonable consistency across multiple well-conducted studies. In addition, higher
odds of infectious disease or symptoms with higher PFHxS concentrations were observed in four of
seven available studies, which is coherent with the immunosuppression observed in antibody
response studies. There are remaining sources of uncertainty in the immunosuppression evidence,
including potential confounding by other PFAS and imprecision of some effect estimates. The
evidence for sensitization or allergic response was generally inconsistent, but there was some
evidence of an association with asthma incidence. A strong positive association with doctor-
diagnosed asthma within the last year was observed in one medium confidence study, and this was
considered the most specific outcome measure available across the set of studies. However, unlike
the evidence on infectious disease, it is unclear how this finding might relate to the evidence
supporting immunosuppression, and without additional support or mechanistic understanding
(mechanistic information was predominantly null apart from a biomarker coherent with the
development of asthma observed in this same study) it does not support a stronger strength of
evidence determination. Other studies of sensitization and allergic response were inconsistent
Studies of autoimmunity were not available.
Animal studies provide indeterminate evidence for immune system effects following
exposure to PFHxS (see Table 3-16). There were no immunotoxicity-specific animal studies in the
database, but rather general toxicity or developmental toxicity studies that included immune-
related endpoints. As a result, the immune endpoints evaluated in the animal studies were less
sensitive and less informative for hazard identification than the endpoints evaluated in the human
studies available in the database. No reliable findings of PFHxS-related immune effects were
observed in high and medium confidence studies in animals exposed to PFHxS.
Taken together, the currently available evidence indicates that PFHxS likely causes
immune toxicity in humans given sufficient exposure conditions.10 This conclusion is based on
epidemiology evidence of an association between PFHxS exposure and immune effects—
specifically, immunosuppression, driven primarily by studies of antibody response following
vaccination, with median PFHxS blood concentrations in children of 0.3-2.5 ng/mL. Despite
imprecision in the results, the antibody results present a generally consistent pattern of findings
that higher prenatal and childhood concentrations of PFHxS were associated with suppression of at
least one measure of the anti-vaccine antibody response to common vaccines, and coherent findings
from more limited evidence of associations between PFHxS exposure and higher odds of infectious
disease. These associations were observed despite poor study sensitivity. While clinical adversity of
10The "sufficient exposure conditions" are more fully evaluated and defined for the identified health effects
through dose-response analysis in Section 5.
3-107
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
fairly small changes in antibody concentrations is not established, one study reported higher odds
for lack of protection from diphtheria, and there is potential for a subset of people to be more
severely affected. Some uncertainty remains resulting from variability in the response by age of
exposure and outcome measures as well as from timing of vaccination (initial and boosters) and the
potential for confounding by other PFAS.
3-108
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-16. Evidence profile table for PFHxS immune effects
Evidence stream summary and interpretation
Evidence integration
summary judgment
Evidence from studies of exposed humans (see Immune Human Studies Section)
Studies and
interpretation
Factors that increase
certainty
Factors that decrease
certainty
Summary and key
findings
Evidence stream
judgment
®©o
Evidence indicates (likely)
Based on generally
consistent evidence of
reduced antibody
response to vaccination at
median blood
concentrations of 0.2-
0.6 ng/mL
Human relevance:
Evidence comes from
epidemiological studies
(see Immune Human
Studies Section)
Cross-stream coherence:
NA: animal evidence is
indeterminate
Antibody Response to
Vaccine
• 7 medium
confidence studies
• 3 low confidence
studies
• Consistency -
Evidence is generally
consistent in the
direction of
association across
vaccine type, timing
of vaccination, and
age at antibody
response
measurement
• Low risk of bias in
studies in children
• Magnitude of effect
- Large effect size
observed in most
studies despite
limited sensitivity
• Potential for
residual
confounding
across PFAS
• Imprecision of
most findings
Studies in children
observed inverse
associations between
PFHxS exposure and
antibody levels following
vaccination in at least
some analyses. While not
all results were
statistically significant,
the direction of
association was generally
consistent across studies
and timing of exposure
and outcome measures.
®©o
Moderate
Generally consistent
evidence for
immunosuppression
with PFHxS exposure
based on lower antibody
response in multiple
medium confidence
studies, supported by
coherent but limited
results for infectious
diseases [Note: the
evidence of
hypersensitivity, based a
single well-conducted
study of asthma with
inconsistent findings
across other studies with
less robust outcome
measures, did not
contribute to this
judgment].
Infectious Disease
• 6 medium
confidence study
• 6 low confidence
studies
• Despite potential
limited sensitivity,
six studies observed
a significant positive
association for at
least one outcome
• Unexplained
inconsistency
• High risk of bias
from potential
outcome
misclassification in
low confidence
studies
2 medium and 3 low
confidence studies
reported higher odds of
infectious disease or
symptoms with higher
PFHxS exposure, including
total infectious disease,
lower respiratory
infection, throat infection,
pseudocroup, and
gastroenteritis
3-109
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration
summary judgment
Sensitization or allergic
response
• 13 medium
confidence studies
Magnitude of effect
- Large effect size in
the only study of
asthma incidence
Exposure-response
gradient observed
for asthma
incidence in one
study with the most
reliable outcome
measure
Biological
plausibility -
mechanistic change
coherent with
asthma in the only
study of asthma
incidence
Potential for
residual
confounding
across PFAS
Unexplained
inconsistency-
Inconsistent
direction of
associations
across studies for
all hypersensitivity
outcomes (with
predominantly
null findings)
One well-conducted study
reported a clear positive
association with asthma
incidence and eosinophilic
cationic protein. Of 11
other studies of asthma,
only four reported higher
odds of asthma in at least
one subpopulation but
were based on "current"
or "ever" asthma
definitions, which are less
specific. Results for
allergies/allergic
sensitization, and dermal
allergic measures had
inconsistent findings.
Evidence from In vivo Animal Studies (see Immune Animal Studies Section)
Hematology
• 2 high confidence
studies
• One medium
confidence study
Histopathology
• 3 high confidence
studies
• Low risk of bias
• Low risk of bias
• Endpoints
considered
nonspecific and
insensitive
indicators of
immune function
Decreased eosinophil
counts in one study (NTP,
2018a); however, there
were no other statistically
significant changes in
immune hematology
parameters and this
finding alone is not
considered adverse.
No PFHxS-induced effects
observed for
histopathology.
Evidence stream
judgment
ooo
Indeterminate
[noting that the immune
endpoints evaluated in
the available animal
studies are considered
insensitive or nonspecific
indicators of immune
function.]
3-110
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration
summary judgment
• 1 medium
confidence study
Organ weights
• 3 high confidence
studies
• 1 medium
confidence study
• Low risk of bias
No PFHxS-induced effects
observed for organ
weights.
C = cohort, CC = case control, CS = cross sectional.
3-111
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3.2.3. Developmental Effects
This section describes studies of PFHxS exposure and potential in utero and perinatal
effects or developmental delays, as well as effects attributable to developmental exposure. The
latter includes all studies for which exposure is limited to gestation and/or early life. Given that
some endpoints examined here, such as spontaneous abortion and preterm birth, could be driven
by either female reproductive or developmental toxicity, these endpoints are also discussed in the
context of coherence in Section 3.2.7 on Female reproductive effects. As such, this section has some
overlap with evidence synthesis and integration summaries for other health systems for which
studies evaluated the effects of developmental exposure (see Sections 3.2.5, 3.2.2, 3.2.7, 3.2.8, and
on potential Hepatic, Endocrine, and Female and Male Reproductive Effects, respectively).
Human Studies
The epidemiologic studies of possible developmental effects of PFHxS evaluate the
following endpoints: fetal and childhood growth restriction, spontaneous abortion, and gestational
duration (i.e., preterm birth and gestational age). Given that many of these endpoints could be
driven by either female reproductive or developmental toxicity, some are also discussed in the
context of coherence in the female reproductive effects section (see Section 3.2.7). The evidence
informing specific endpoints is discussed and synthesized below; however, the hazard conclusion
was determined at the level of developmental effects for the group of endpoints.
Study evaluation considerations
As detailed in the PFAS Systematic Review Protocol (see Appendix A), multiple outcome-
specific considerations 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 pregnancy 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 PFHxS levels and the developmental
effects examined here. Irrespective of study design, 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, it was considered in the study quality evaluations and as part of the overall weight
of evidence determination (see Appendix C for additional discussion of these issues).
For the Exposure domain, all the available studies analyzed PFAS in serum or plasma using
standard methods. Given the estimated long half-life of PFHxS in humans (range: 4.7 to 8.5 years;
see Section 3.1.4.), samples collected during all three trimesters (and shortly after birth) were
3-112
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
considered adequately representative of the most critical in utero exposures for fetal growth and
gestational duration measures. Many of the cross-sectional studies relied on umbilical cord
measures collected shortly after birth. Exposure measures collected close to or concurrently with
outcome ascertainment were considered etiologically relevant and acceptable for these
developmental endpoints; thus, exposure measurement ratings were not downgraded for timing of
measurement. 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 and
Harding f20061 and Perngetal. f20161 and other PFAS syntheses for potential cardiometabolic
disorders in Section 3.2.6). There is some evidence that birth weight deficits can be followed by
increased weight gain that may occur especially among those with rapid growth catch-up periods
during childhood (Perng etal.. 2016). 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 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, some spontaneous abortion studies were downgraded for
participant selection due to incomplete case ascertainment given that some pregnancy losses go
unrecognized early in pregnancy including before participants would be enrolled. This incomplete
ascertainment, referred to as left truncation, can result in bias toward the null if ascertainment of
fetal loss is not associated with PFHxS exposures (i.e., nondifferential). In some situations where
there is a true association with PFHxS, differential loss is possible, possibly causing a 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. When sufficient high and medium confidence evidence is available
for a set of related endpoints, the developmental effects synthesis is largely focused on the higher
quality endpoints (i.e., classified as good in the outcome domain).
3-113
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
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 developmental endpoints in the outcome domain judgments. Some of the adverse birth
weight 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
birth weight z-scores), as well as binary measures such as SGA (e.g., lowest decile of birth weight
stratified by gestational age and other covariates) and low birth weight (i.e., typically <2,500 g;
5 lbs., 8 oz.) 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 greater
measurement error (e.g., head circumference and body length measures such as ponderal index)
were given a rating of adequate fShinwell and Shlomo. 20031. These sources of measurement error
are expected to be nondifferential with respect to PFHxS 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 (as the standard definition of preterm
birth, and that used in these published studies, is gestational age <37 weeks). The potential for
measurement error can complicate accurate estimates of gestational age and may decrease study
sensitivity related to some of these endpoints especially when based on recall of last menstrual
period alone. However, 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 Studies were downgraded if based solely on last menstrual period
and more certainty was anticipated for studies using a combination of measures with comparisons
of any differences. Any sources of error in the classification of these endpoints should be
nondifferential with respect to PFHxS exposure and, therefore, would not be considered a major
concern for risk of bias, but could impact precision and study sensitivity.
Anogenital distance (AGD) is an externally visible marker that has been shown in animal
studies to be a sensitive indicator of prenatal androgen exposure (lower androgen levels associated
with decreased AGD, and the reverse). It is associated with other reproductive tract abnormalities,
including hypospadias and cryptorchidism in human and animal males fSathvanaravana et al..
2010: Salazar-Martinez etal.. 2004: Liu etal.. 20141: the potential adverse consequences in females
are less well defined. In boys, measures can be taken from the center of the anus to the posterior
base of the scrotum (ASD) or from the center of the anus to the cephalad insertion of the penile
(APD). In girls, there are two possible measures, the anoclitoris distance (ACD) and the
anofourchette distance (AFD). The primary outcome-specific criteria for this outcome are the use of
clearly defined protocols for measurement, ideally multiple measures of each distance (averaged),
and minimal variability in the age of participants at measurement.
3-114
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Growth restriction - fetal growth
Developmental Epidemiologic Studies
Sixty-one epidemiological publications (across 58 different studies) examining PFHxS
exposures in relation to developmental endpoints were identified in the literature search. Several
studies examined multiple endpoints, captured in separate subsections below. This included the
following: 12 studies on postnatal growth, 19 studies on gestational duration, 5 on fetal loss, 4 on
anogenital distance, 2 studies on birth defects, and 42 publications (across 39 different studies) that
examined fetal growth restriction.
Fetal growth restriction - study background
The heat map of 39 fetal growth restriction studies below does not include three
overlapping publications, such as the Woods etal. (2017) publication from the same study
population (Health Outcomes and Measures of the Environment cohort) as Shoaffetal. (2018) (see
Figures 3-19 and 3-20). For consistency, birth outcomes measures reported in Manzano-Salgado et
al. f2017al were preferred to in utero growth estimates in the Costa etal. f20191 study from the
same Environment and Childhood-Infancia y Medio Ambiente (INMA) birth cohort The smaller
population subset from the Bierregaard-Olesen et al. f20191 study is from the same Aarhus birth
cohort as Bach etal. (2016). Given disparate results shown below in this subset versus the whole
cohort for head circumference and birth length, results from the full study population in Bach et al.
(2016) are given precedent. However, the Bierregaard-Olesen etal. (2019) provide additional sex-
specific data not examined in Bach etal. (2016). Difference in results for these endpoints are
highlighted in the syntheses below but only one study is plotted for each endpoint to aid the
evaluation of consistency across studies. Five of the remaining 39 fetal growth studies fMonrov et
al.. 2008: Maekawa et al.. 2017: Lee etal.. 2013: Lee etal.. 2016: Alkhalawi etal.. 20161 are not
included in the synthesis further as they were classified as uninformative largely due to critical
study deficiencies in some risk of bias domains (e.g., confounding) or multiple domain deficiencies.
Birth weight - background of studies
As shown in Figure 3-19 and Table 3-17, there were 34 informative studies that examined
birth weight measures in relation to PFHxS exposures. This included 13 studies that examined
PFHxS in relation to continuous standardized birth weight scores. Ten of these 13 reported
standardized measures along with mean birth weight differences in relation to PFHxS. Three (Xiao
etal.. 2019: Gross etal.. 2020: Gardener etal.. 2021) of the 13 studies reported only standardized
birth weight measures, with Gardener et al. (2021) not plotted below with the others given an
atypical, dichotomized effect estimate with different scaling.
Of the 31 epidemiological studies with mean birth weight data, four fMarks etal.. 2019a:
Maisonet etal.. 2012: Lind etal.. 2017: Ashlev-Martin et al.. 20171 only reported sex-specific
findings, including a study in boys (Marks etal.. 2019a) and girls (Maisonet etal.. 2012) from the
3-115
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
ALSPAC study (see Figure 3-19). Fifteen different studies examined mean birth weight differences
across the sexes 14 each in boy and girls. Among the 27 studies with results in the overall
population, three studies fGao etal.. 2019: Eick etal.. 2020: Cao etal.. 20181 reported results based
only on categorical data.
3-116
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Alkhalawi el aL 2016
Ashley-Martin, 201
Bach, 20
Buck Louis, 20
Callan, 20
Cao, 201
Chang, 2022
Chen, 2021
Eick et al.. 2020
Gao, 2019
Gardener, 2021
Gross, 2020
Gyllenhammar, 2018
Hamm, 2010
Hjermilslev, 2020
Kashino, 2020
Kwon, 2016,
Lee, 2013, 3
Lee, 2016,
Lenters, 2016
Li, 2017
Lind, 2017
Luo, 2021
Maekawa, 2017,4
Maisonet, 2012,1
Manzano-Salgado, 2017,4
Marks, 2019, 5
Meng, 2018, 4
Monroy, 2008, 2
Sagiv, 2018, 4
Shi, 2017,
Shoaff, 2018,
Starling, 2017, 3
Valvi, 2017, 3
Wikstrom, 2020, €
Workman, 2019, S
5387135H
7021199-J
6311632-
3858531 -
3981528-
5617146-
3858512 -I
9959610 -j
3827535 -J
4619944-1
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-19. Study evaluation results for 39 epidemiological studies of birth
weight and PFHxS, For additional details see HAWC link.
3-117
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
39 Perinatal Studies of Mean Birth Weight included in study quality evaluation
3 Overlapping Study Mean
Birth Weight Studies
31 Included Mean
Birth Weight Studies
\
5 Uninformative Mean
Birth Weight Studies
27 Studies with Overall
Population Results
14 Studies with Sex-Specific
Results (4 studies examined
either only boys and/or girls)
27 Studies included in
Meta-Analysis that had
Overall Population Results
(boys and girls combined)
Confidence
Confidence
based on continuous
exposure measures
11 High Confidence Studies
10 Medium Confidence Studies
6 Low Confidence Studies
6 High Confidence Studies
6 Medium Confidence Studies
3 Low Confidence Studies
Figure 3-20. Perinatal studies of birth weight measures and subsets included
in different evaluations.
Birth weight - mean differences - background
Twenty-five of the included 31 mean birth weight studies were prospective birth cohorts,
and six were cross-sectional studies fXu etal.. 2019: Shi etal.. 2017: Li etal.. 2017b: Kwonetal..
2016: Gvllenhammar etal.. 2018: Callan etal.. 20161 (see Figures 3-21 and 3-22). Five of these six
studies relied on umbilical cord blood measures (Xu etal.. 2019: Shi etal.. 2017: Li etal.. 2017b:
Kwon etal.. 2016: Cao etal.. 2018). and one collected PFHxS blood samples in infants 3 weeks
following delivery fGvllenhammar et al.. 2018). Twenty-four studies had maternal blood measures
that were sampled during trimesters one fManzano-Salgado etal.. 2017a: Lind etal.. 2017: Buck
Louis etal.. 2018: Ashley-Martin et al.. 20171. two fHamm etal.. 20101. three fYao etal.. 2021: Valvi
etal.. 2017: Luo etal.. 2021: Kashino etal.. 2020: Gao etal.. 2019: Callan etal.. 20161. or across
multiple trimesters fWorkman etal.. 2019: Wikstrom etal.. 2020: Starling etal.. 2017: Shoaffetal..
2018: Sagivetal.. 2018: Marks etal.. 2019a: Maisonet et al.. 2 012: Lenters etal.. 2016: Hjermitslev
etal.. 2020: Eick etal.. 2020: Chen etal.. 2021: Chang etal.. 2022: Bach etal.. 2016). The study by
Meng etal. (2018) pooled exposure data from two study populations, one that measured PFHxS in
umbilical cord blood and one that measured PFHxS in maternal blood samples collected in
trimesters 1 and 2. For comparability with other studies of mean birth weight, EPA only examined
data from one measure, such as umbilical cord or maternal serum concentrations, and when
necessary, relied on other related publications (e.g., Gvllenhammar I f201711 or additional
information or data provided by study authors. When possible, EPA converted effect estimates that
were based on continuous PFHxS measures to a 1 ln-unit increase to enhance comparability across
3-118
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
studies (see Figures 3-23, 3-24, and 3-25). These results employing a common unit of measurement
were also used for the birth weight meta-analysis conducted by EPA (see Appendix C for details on
the methods employed).
Thirteen of the 31 mean birth weight studies were rated high in overall study confidence
fYao etal.. 2021: Wikstrom et al.. 2020: Valvi etal.. 2017: Starling etal.. 2017: Shoaff etal.. 2018:
Sagivetal.. 2018: Manzano-Salgado etal.. 2017a: Luo etal.. 2021: Lind etal.. 2017: Eicketal.. 2020:
Buck Louis etal.. 2018: Bach etal.. 2016: Ashley-Martin et al.. 2017). while 11 were rated medium
(Mengetal.. 2018: Maisonet et al.. 2 012: Li etal.. 2017b: Lenters etal.. 2016: Kwon etal.. 2016:
Kashino etal.. 2020: Hiermitslev etal.. 2020: Hamm etal.. 2010: Gvllenhammar et al.. 2018: Chen et
al.. 2021: Chang etal.. 20221. and 7 were classified as low fXu etal.. 2019: Workman etal.. 2019: Shi
etal.. 2017: Marks etal.. 2019a: Gao etal.. 2019: Cao etal.. 2018: Callanetal.. 20161 (see Figure 3-
19).
Of the 31 mean birth weight studies detailed in this synthesis, 13 studies (Wikstrom etal..
2020: Valvi etal.. 2017: Starling etal.. 2017: Shoaff etal.. 2018: Sagiv etal.. 2018: Meng etal.. 2018:
Marks etal.. 2019a: Maisonet etal.. 2012: Luo etal.. 2021: Li etal.. 2017b: Lenters etal.. 2016:
Gvllenhammar et al.. 2018: Ashlev-Martin et al.. 20171 were considered to have good study
sensitivity. T en studies fManzano-Salgado etal.. 2017a: Lind etal.. 2017: Kwon etal.. 2016:
Hiermitslev etal.. 2020: Hamm etal.. 2010: Eick etal.. 2020: Chen etal.. 2021: Chang etal.. 2022:
Buck Louis etal.. 2018: Bach etal.. 2016) were classified as adequate and eight were deficient (Yao
etal.. 2021: Xu etal.. 2019: Workman etal.. 2019: Shi etal.. 2017: Kashino etal.. 2020: Gao etal..
2019: Cao etal.. 2018: Callan etal.. 2016).
Birth weight - mean difference results (in grams) in overall population
Overall, 14 of the 27 different epidemiological studies that examined associations in the
overall population (i.e., both male and female neonates combined) detected some deficits in relation
to PFHxS exposures (see Figures 3-21, 3-22, and 3-23 and Table 3-17). This included 5 (Starling et
al.. 2017: Shoaff etal.. 2018: Manzano-Salgado etal.. 2017a: Buck Louis et al.. 2 018: Bach etal..
2016) out of 11 high confidence studies, 5 (Li etal.. 2017b: Kwon etal.. 2016: Hiermitslev etal..
2020: Gvllenhammar etal.. 2018: Chang etal.. 2022) out of 10 medium confidence and 4 (Xu etal..
2019: Gao etal.. 2019: Cao etal.. 2018: Callan etal.. 20161 out of 6 low confidence studies. In
contrast, four studies reported increased birth weight with PFHxS exposures while nine other
studies were null. For example, the high confidence study by Eick etal. f20201 reported
nonsignificant increased birth weight across PFHxS tertiles ((3 range: 75.7 to 82.2 g) relative to
tertile 1. The medium confidence study by Chen etal. (2021) reported a smaller and imprecise
increased mean birth weight based on continuous exposures ((3 = 27.6 g; 95% CI: -64.7,119.9 per
ln-unit increase) along with mixed results based on categorical PFHxS exposures ((3 range: -46 to
26 g).
The high confidence Manzano-Salgado etal. f2017al study showed consistent but
nonmonotonic birth weight decreases across all three upper quartiles ((3 range: -30 to -65 g), but a
3-119
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
relatively small deficit per each ln-unit increase ((3 = - 12.4 g; 95% CI: -46.2, 21.4). The latter
results were indicative of deficits seen in the five high confidence studies ((3 range: -12 to -22 g per
each ln-unit increase). Birth weight deficits detected in the five medium confidence studies were
larger ((3 range: -20 to -60 g per each ln-unit increase). This included three studies fKwon etal..
2016: Hiermitslev etal.. 2020: Gvllenhammar et al.. 20181 which reported birth weight decreases
consistent in magnitude ((3 range: -49 to -60 g per each ln-unit increase). The study by Chang et al.
(2022) reported a nonsignificant deficit per each ln-unit increase ((3 = -20 g; 95% CI: -84, 45) but
larger results for PFHxS quartiles 2 ((3 = -36 g; 95% CI: -154, 83) and 4 ((3 = -54 g; 95% CI: -173,
66). The study by Kashino etal. f20201 reported a null association with PFHxS and mean birth
weight ((3 = -1.3 g; 95% CI: -26.3, 23.6 per each ln-unit increase). They did show large differences
in multiparous participants ((3 = -81.2 g; 95% CI: -122.3, -40.1 per each ln-unit increase) but not
for primiparous participants ((3 = -2.2 g; 95% CI: -46.2, 41.7 per each ln-unit increase).
Birth weight deficits detected in two low confidence studies were consistent in magnitude
((3 range: -72 to -76 g per each ln-unit increase). The low confidence study by Gao etal. (2019)
reported larger decreased birth weight in a nonmonotonic fashion across PFHxS tertiles 2 ((3 = -
154.1 g; 95% CI: -332.2, 24.0) and 3 ((3 = -101.2 g; 95% CI: -275.5, 73.1). Across all confidence
levels, only one of 11 studies with categorical data in the overall population showed some evidence
of an exposure-response relationship ((3 range: -14 to -25 g across tertiles) and these study results
by (Cao etal.. 2018) were imprecise.
Birth weight-mean difference-overall population summary
Overall, the majority (14 of 27) of mean birth weight studies showed deficits with
increasing PFHxS exposures and interestingly some consistency in reported magnitude of deficits
by study confidence level. For example, the five high confidence studies showed consistently
smaller deficits ((3 range: -12 to -22 g per each unit increase) compared with the five medium ((3
range: -20 to -60 g) and two low ((3 range: -72 to -76 g) confidence studies. Although the majority
of low confidence studies observed larger birth weights in association with PFHxS exposure, the
estimates were consistently imprecise, and the identified methodological limitations preclude
further interpretation in that subset There was very limited evidence of exposure-response
relationships based on categorical data, but the magnitude of changes in those studies showing
deficits ranged from -25 to -101 g for the highest quantile (compared with the lowest quantile)
were comparable to those results ((3 range: -12 to -76 g per each ln-unit increase) based on the
continuous exposure expressions shown above.
Limited patterns were evident as study sensitivity, exposure levels and contrasts and other
study design elements were not explanatory for null or inverse associations detected across the
birth weight studies. The birth weight deficits in the overall population may be influenced by
hemodynamic changes during pregnancy related to exposure assessment timing as only 4 of the 14
were based on early biomarker sampling (see meta-analysis for further examination).
3-120
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Meta-analysis of mean birth weight differences
Thirty-one informative studies were identified for possible inclusion into a meta-analysis of
overall population estimates (see Figure C-l and more details on the Methods in Appendix C) if they
provided results in the overall population or in both sexes which allowed combination to estimate
an overall population result. Three of these studies with PFHxS categorical data only (Gao etal..
2019: Eick etal.. 2020: Cao etal.. 2018) were not included in the meta-analysis due to the lack of
results on a per continuous exposure increase. Results from 28 remaining publications from 27
cohorts include the other 24 studies identified in the overall population section noted above as well
as three additional studies, two of which reported sex-specific data only on boys and girls
individually fLind etal.. 2017: Ashley-Martin et al.. 20171 in the same publication. Another cohort
(ALSPAC) reported results in girls fMaisonetetal.. 20121 in one publication and boys fMarks etal..
2019a) in another and were combined for the meta-analysis.
Following scale conversions and re-expressions (to ln-unit) for some studies by U.S. EPA,
the meta-analysis of 27 studies showed negligible between-study heterogeneity (I2 = 0%), and a
small but statistically significant decrease in birthweight ((3=—7.9 g; 95% CI: -15.0, -0.7) per each
ln-unit PFHxS increase (see Figure 3-21). Statistically significant results comparable in magnitude
were also detected when restricted to 23 medium and high confidence studies ((3=-8.1 g; 95% CI:
-15.4, -0.9) and also to 22 studies thatprovided results based on some logarithmic transformation
((3 = -6.0 g; 95% CI: -15.8, 3.8).
Mean birth weight deficits were detected only among the 12 high ((3 = -6.8 g; 95% CI: -16.3,
2.8) and 11 medium ((3 = -10.0 g; 95% CI: -21.1,1.1) confidence studies. The pooled effect in the
low confidence studies was null ((3 = -1.5 g; 95% CI: -51.6, 48.7) and was based upon far fewer
studies (n = 4). Stratified mean birth weight deficits were also different based on studies with later
sample timing. The five studies that used umbilical cord samples or maternal samples collected
after pregnancy had considerably larger deficits ((3 = -28.3 g; 95% CI: -69.3,12.7) compared with
10 studies with mid- to late pregnancy sampling ((3 = -3.9 g; 95% CI: -17.7, 9.9) or to 12 studies
with sampling from early pregnancy ((3 = -7.6 g; 95% CI: -16.2,1.1). Among these 12 studies,
smaller differences were observed within the sub-set of six studies with the earliest sampling (e.g.,
based on predominately first trimester or earlier sampling and/or low mean, median or mode of
gestational age week at sampling < 10) ((3 = -3.5 g; 95 CI: -14.8, 7.9) compared with the remaining
six early sampled studies ((3 = -13.4 g; 95% CI: -26.9, 0.1).
Overall, the meta-analytical data showing a small change in mean birth weight per each ln-
unit change (i.e., a 2.7-fold increase in exposure in ng/mL within the range of observed exposures in
the study populations) support the main epidemiologic findings detailed above and provide some
limited evidence of an adverse effect on birthweight from maternal exposure to PFHxS (see
Appendix C for more detail and additional stratified analyses). The median exposure ranged from
0.16 to 10.36 ng/mL across the 27 studies with birth weight data in the meta-analysis. The pooled
birth weight estimates expressed here per each unit change are relatively small in magnitude and
3-121
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
could be larger depending on the range of exposures within a particular study population or the
range to which it is being extrapolated to. Although a gradient across sample timing was not
evident across all time periods, the pooled estimate in the five studies with post-partum sampling
was much larger. In contrast to the mid to late pregnancy sampled studies, the associations in the
early maternal biomarker sampled studies were consistent in magnitude to the pooled estimate
across all studies as well as the combined medium and high confidence studies. Thus, while some
uncertainty remains on the potential impact due to pregnancy hemodynamics especially in the later
sampled studies, the overall combined results, the early sample timing studies, as well as the higher
confidence (medium and high combined) studies do show a small but consistent association
between mean birthweight and PFHxS.
3-122
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-17. Summary of 34 epidemiologic studies of PFHxS exposure and growth restriction measures
Author
Study location, years
Sample
size
Median exposure
(range) in
serum/plasma
(ng/mL)
Birth
weight
Birth
length
HC
SGA/
LBW
High confidence studies
Ashlev-Martin et al.
(2017)
Canada,
2008-2011
1,509
1.0
(0.3, 25.0)
0 Overall
+ Boys
- Girls
Bach et al. (2016);
Bierregaard-Olesen et
al. (2019)
Denmark,
2008-2013
1,507
0.5 (
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Author
Study location, years
Sample
size
Median exposure
(range) in
serum/plasma
(ng/mL)
Birth
weight
Birth
length
HC
SGA/
LBW
Starling et al. (2017)
CO, USA,
2009-2014
598
0.8 (0.1, 10.9)
- Overall
Valvi et al. (2017)
Denmark,
1997-2000
604
4.54 (N/A)
+ Overall/
Boys/Girls
- Overall/ Boys
0 Girls
+
Overalf/Boys*
0 Girls
Wikstrom et al. (2020)
Sweden,
2007-2010
1533
1.23 (N/A)
0 Overall/Boys
/Girls
0 SGA Overall/Boys
-t SGA Girls
Xiao et al. (2019)
Faroe Islands, 1994-
1995
172
0.55 (0.1, 2.8)
- Overall/Boys/
Girls
Overall/Boys/Girls*
Overall/Boys/
Girls*
Yao et al. (2021)
China, 2010-2013
369
0.32
0 Overall
Medium confidence studies
Chang et al. (2022)
USA, 2014-2018
370
1.10 (
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Author
Study location, years
Sample
size
Median exposure
(range) in
serum/plasma
(ng/mL)
Birth
weight
Birth
length
HC
SGA/
LBW
Lenters et al. (2016)
Ukraine/Poland/
Greenland,
2002-2004
1,321
1.56, 2.28 (0.45,
5.95)e
0 Overall
Li et al. (2017b)
China,
2013
321
3.87 (ND, 20.15)
- Overall/Boys
0 Girls
Maisonet et al. (2012)
United Kingdom,
1991-1992
422
1.6 (0.2-54.8)
-Girls*3
- Girls*3
Meng et al. (2018)
Denmark,
1996-2002
2,120
~1(N/A)
0 Overall/Girls
+ Boys
TLBW
TVLBW
Low confidence studies
Callan et al. (2016)
W. Australia,
2003-2004
98
0.33 (0.06, 3.3)
- Overall
- Overall
- Overall
Cao et al. (2018)
China,
2013-2015
337
0.09 (0.03-0.31)f
- Overail3/Boysa
+ Girls
- Overall/Boys
0 Girls
Gao et al. (2019)
China, 2015-2016
132
0.24 (N/A)
- Overall
- Overall
Gross et al. (2020)
USA, 2014
98
0.108 (N/A)g
- Overall/
Boys/Girls
Marks et al. (2019a)
England,
1991-1992
447
1.9 (0.5, 74.2)
-Boys
- Boys3
0 Boys
Shi et al. (2017)
China,
2012
170
0.16 (
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
HC = head circumference; SGA = small for gestational age; LBW = low birth weight; VLBW = very low birth weight; LOQ = level of quantification; LOD = level of
detection; ND = nondetectable; N/A = not available.
Note: "Adverse effects" are indicated by both increased ORs (-) for dichotomous outcomes and negative associations (-) for the other outcomes.
/ denotes multiple groups with the same direction of associations.
aExposure-response relationship detected based on categorical data.
deduction based on categorical data, null results based on continuous data.
cHigh confidence for birth weight and Medium confidence for head circumference.
Arithmetic mean value, no median value available.
eNo range provided but 5th-95th percentiles included.
fNo range provided but 10th-90th percentiles included.
gDried blood spot PFHxS sample collected within 48 hours of birth.
3-126
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Population
Overall Study
Design
Exposure
Regression
Exposure
Confidence
Window
Coefficient
Comparison
Regression coefficient
# |3 [change in mean BWT (g)]
Buck Louis, 2018,
5016992
NICHD Fetal Growth Studies
(2009-2013), United States. 2106
mother-infant pairs
I High |
Cohort
(Prospective)
Trimester 1
-22.1
ln-unit (ng/mL)
1
1 • H
O P [change in mean BWT (g)] p<0.05
1—195% confidence interval
Manzano-Salgado
et al., 2017,
4238465
INMA cohort (2003-2008) 1202
mother-infant pairs
IHighl
Cohort
(Prospective)
Trimester 1
-30.2
Quartile 2
1
• 1 1
-64.9
Quartile 3
^ J
1
-40.3
Quartile 4
-12.4
ln-unit (ng/mL)
increase
1 #J 1
1
Bach eta!., 2016,
Aarhus Birth Cohort (2008-2013),
|High|
Cohort
Trimester 1-2
-41
Quartile 2
3981534
Denmark, 1507 mother-infant pairs
(Prospective)
1
-34
Quartile 3
-49
Quartile 4
a
-19.35
ln-unit (ng/mL)
increase
1 •—j—1
Sagiv, 2018,
Project Viva (1999-2002) 1645
IHighl
Cohort
Trimester 1-2
-37.5
Quartile 2
4238410
mother-infant pairs
(Prospective)
44.9
Quartile 3
. . ^
-10.8
Quartile 4
-3.28
ln-unit (ng/mL)
1
Wikstrom, 2020,
SELMA (2007-2010), Sweden. 1533
|High|
Cohort
Trimester 1-2
-4
Quartile 2
1
6311677
mother-infant pairs
(Prospective)
•)
-15
Quartile 3
-6
Quartile 4
-0.1
ln-unit (ng/mL)
Eicketal.. 2020,
7102797
Chemicals In Our Bodies (CIOB)
(2014-2018), US. 497 female
participants
IHighl
Cross-sectional
Trimester 1-3
82.2
Tertile 2
1
1 i
1
75.71
Tertile 3
Starling, 2017.
Healthy Start cohort (2009-2014) 628
IHighl
Cohort
Trimester 2-3
32.9
Tertile 2
1
3858473
mother-infant pairs
(Prospective)
-31.8
Tertile 3
-13.5
ln-unit (ng/mL)
1
1 •—1 1
Shoaff et al„
2018.4619944
HOME (2003-2006). United States.
345 mother-infant pairs
IHighl
Cohort
(Prospective)
Trimester 2-3, at
delivery
-20.88
ln-unit (ng/mL)
increase
1
1 • | 1
Luo et al.. 2021,
9959610
Zhujiang Hospital Cohort. China
(2017-2019) 224 mother-infant pairs
|High|
Cohort
(Prospective)
Trimester 3
-11.343
Quartile 2
1
* 1 1
-7.761
Quartile 3
1 1
-2.985
Quartile 4
. 1 !
¦12.5
ln-unit (ng/mL)
increase
'
Valvi etal., 2017,
Faroe Islands (1997-2000),
IHighl
Cohort
Trimester 3
21.6
ln-unit (ng/mL)
1
3983872
Denmark. 604 mother-infant pairs
(Prospective)
increase
Yao etal.. 2021.
9960202
Laizhou Wan Birth Cohort (LWBC)
(2010-2013) China, 369 parent-infant
pairs
IHighl
Cohort
(Prospective)
Trimester 3
-10.19
ln-unit (ng/mL)
increase
1
1
200
-150
-100
-50 0 50
100 150 200
Figure 3-21. Overall population birth weight results for 11 high confidence
PFHxS epidemiological studies.3 b For additional details see HAWC link.
BWT = birth weight.
aStudies are sorted first by overall study confidence level, then by exposure window examined.
3-127
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Population
Overall Study
Confidence
Design
Exposure
Window
Regression
Coefficient
Exposure
Comparison
Gyllenhamrnar,
2018, 4238300
POPUP (1996-2011) 381
mother-infant pairs
IMediuml
Cross-sectional
3 weeks post-birth
-53.3
In-unit (ng/mL)
increase
Kwon. 2016,
3858531
EBGRC (2006-2010) 268
mother-Infant pairs
|Medlum|
Cross-sectional
At birth
-60.1
In-unit (ng/mL)
Increase
LI. 2017, 3981356
GBCS (2013), China, 321
mother-infant pairs
IMediuml
Cross-sectional
At birth
-30.3
In-unit (ng/mL)
increase
Chang, 2022,
9959688
Emory University African American
Vaginal, Oral and Gut Microbiomo in
Pregnancy Study (2014-2018), 448
participants
|Medium|
Cohort
(Prospective)
Trimester 1-2
-36
5
Quartile 2
Quartile 3
-54
Quartile 4
-20.2
In-unit (ng/mL)
Chen. 2021,
7263985
Prospective cohort analysis from
Shanghai Birth Cohort (2015-2017).
214 mother-infant pairs
|Medium|
Cohort
(Prospective)
Trimester 1-2
5.1
Quartile 2
-45.5
Quartile 3
25.6
Quartile 4
27.6
In-unit (ng/mL)
Meng. 2018,
4829851
DNBC (1996-2002), Denmark. 3535
mother-infant pairs
IMediuml
Cohort
(prospective)
Trimester 1-2
37.3
Quartile 2
7.6
Quartile 3
8.6 Quartile 4
4.5
In-unit (ng/mL)
Hjermitslev, 2020,
5880849
ACCEPT birth cohort (2010-2011,
2013-2015), Greenland, 482
mother-infant pairs
IMediuml
Cohort
(Prospective)
Trimester 1-3
-48.62
In-unit (ng/mL)
increase
Hamm, 2010,
1290814
Alberta cohort (2005-2006) 252
mother-infant pairs
|Medium|
Cohort
(Prospective)
Trimester 2
4.3
Tertile 2
26
Tertile 3
21.9
In-unit (ng/mL)
increase
Lenters. 2016,
5617146
INUENDO (2002-2004),
Greenland'Poland/Ukraine, 1,321
mother-infant pairs
IMediuml
Cohort
(Prospective)
Trimester 2-3
-5,06
In-unit (ng/mL)
Kashino, 2020,
6311632
Hokkaido Study on Environment and
Children's Health (2003-2009),
Japan, 1985 mother-child pairs
IMediuml
Cohort
(Prospective)
Trimester 3
-2.21
-81.2
Parity (=1)
Parity (>=2)
-1.3
In-unit (ng/mL)
Cao. 2018,
5080197
Zhoukou City Longitudinal Birth
Cohort (2013-2015), China, 282
mother-infant pairs
Cohort
(Prospective)
At birth
-13.7
Tertile 2
-25.1
Tertile 3
Shi. 2017.
3827535
Haidan Hospital (2012)170
mother-infant pairs
Cross-sectional
At birth
47.3
In-unit (ng/mL)
Xu, 2019,
5381338
Cross-soctional study (2016-2017),
China. 98 mother-infant pairs
|Low|
Cross-sectional
At birth
-75.5
In-unit (ng/mL)
Workman, 2019,
5387046
Canadian Healthy Infant Longitudinal
Development (CHILD) Study
(2010-2012), Canada (414
mother-Infant pairs)
|Low|
Cohort
(Prospective)
Trimester 2-3
-6.6
In-unit(ngZmL)
increase
Callan. 2016.
3858524
AMETS (2008-2011). Australia, 98
mother-infant pairs
|Low|
Cross-sectional
Trimester 3
-72
In-unit (ng/mL)
Gao. 2019.
5387135
Affiliated Hospital of Capital Medical
University (2015-2016), China, 132
pregnant women
|Low|
Cohort
(Prospective)
Trimester 3
-154.1
Tertile 2
-101.2
Tertile 3
Regression coefficient
1
0 |3 [change in mean BWT (g)]
O P [change in mean BWT (g)) p<0.05
}—J 95% confidence interval
-150 -100 -50 0 50 100 150 200
Figure 3-22. Overall population birth weight results for 16 medium and low
confidence epidemiological studies. For additional details see HAVVC link.
BWT = birth weight.
aStudies are sorted first by overall study confidence level, then by exposure window examined.
b(Meng et al„ 2018) pooled samples from umbilical cord blood and maternal plasma during the first and second
trimesters. The remaining studies were all based on either one umbilical or maternal sample.
c(Gvllenhammar et al., 2018) results are displayed here for mean birth weight among 587 overall population
participants in the POPUP Cohort compared with a smaller sample size of 381 in their 2018 publication.
dFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g. (Gvllenhammar et al., 2018)).
eSome confidence intervals (CIs) truncated, e.g., the entire 95% CIs for these studies are: (Hjermitslev et al., 2020):
-230, 44.1; (Xu et ak20191 -272.7,121.6; (Gao etal., 2019): Tertile 2: -332.2, 24; Tertile 3: -275.5, 73.1.
3-128
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Author(s) and Year
Confidence Timing
Estimate [95% CI]
Buck Louis, 2018
High
1st
2106
Ashley-Martin, 2017
High
1st
1509
Lind, 2017
High
1st
636
Sagiv, 2018
High
1st-2nd
1645
Wikstrom, 2020
High
1 st-2nd
1533
Bach, 2016
High
1st-2nd
1507
Manzano-Salgado, 2017
High
1 st-3rd
1202
Starling, 2017
High
2nd-3rd
598
Shoaff, 2018
High
2nd-3rd
299
Valvi, 2017
High
3rd
604
Yao, 2021
High
3rd
369
Luo, 2021
High
3rd
224
Meng, 2018
Medium
1st-2nd
2120
Chang, 2022
Medium
1st-2nd
370
Chen, 2021
Medium
1st-2nd
214
Maisonet, 2012
Medium
1 st-3rd
895
Hjermitslev, 2020
Medium
1 st-3rd
266
Hamm, 2010
Medium
2nd
252
Lenters, 2016
Medium
2nd-3rd
1321
Kashino, 2020
Medium
3rd
1951
Li, 2017
Medium
Birth
321
Kwon, 2016
Medium
Birth
268
Gyllenhammar, 2018
Medium
Post-Birth
587
Workman, 2019
Low
2nd-3rd
414
Callan, 2016
Low
3rd
98
Shi, 2017
Low
Birth
170
Xu, 2019
Low
Birth
98
-22.1 [-52.5, 8.4]
7.5 [-26.6, 41.6]
3.5 [-46.7, 53.8]
-3.3 [-18.8, 12.2]
-0.1 [-38.1, 37.9]
-19.4 [-55.4, 16.7]
-12.4 [-46.2, 21.4]
-13.5 [-50.7, 23.7]
-20.9 [-55.9, 14.1]
21.6 [-25.2, 68.5]
-10.2 [-130.1, 109.7]
-12.5 [-106.8, 81.8]
4.5 [-36.0, 44.9]
-20.2 [-84.4, 44.0]
27.6 [-64.7, 119.9]
-11.2 [-28.5, 6.2]
-48.6 [-120.3, 23.0]
21.9 [-23.4, 67.2]
-5.1 [-44.5, 34.3]
-1.3 [-26.3, 23.6]
-30.0 [-83.5, 23.5]
-60.0 [-136.4, 16.3]
-53.3 [-104.5, -2.1]
-6.6 [-66.9, 53.7]
-72.0 [-194.0, 50.0]
47.3 [-23.4, 117.9]
-75.5 [-272.7, 121.7]
RE Model for All Studies (Q = 18.65, df = 26, p = 0.85; I2 = 0.0%, t2 = 0.00)
-7.9 [-15.0, -0.7]
I 1 1 1-
-280 -200 -100 0
Estimate (95% CI)
100
~~I
150
Figure 3-23. Forest plot of 27 studies included for the EPA meta-analysis on
changes in mean birth weight per each In-unit PFHxS increase.
Abbreviations: T1 = first trimester; T1-T2 = first and second trimester, T2 = second trimester; T2-T3 = second and
third trimester; T3 = third trimester; B = at birth, PB = post-birth. See Appendix C for more details.
3-129
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Birth weight - mean differences - sex-specific results
Nine of the 15 different studies with results showed some birth weight deficits in relation to
PFHxS exposures in either or both sexes (see Figures 3-24 and 3-25) although results were mixed
within and across studies. In contrast, four studies in boys ((3 range: 20 to 68 g) and two studies in
girls ((3 range: 29 to 43 g) showed nonsignificant increased birth weight per ln-unit PFHxS increase.
Seven studies in girls were null (Wikstrom etal.. 2020: Meng etal.. 2018: Manzano-Salgado et al..
2017a: Lind etal.. 2017: Li etal.. 2017b: Kashino etal.. 2020: Ashley-Martin et al.. 20171. while
three were null in boys fValvi etal.. 2017: Manzano-Salgado etal.. 2017a: Kashino etal.. 20201
Seven studies showing inverse associations were in boys and four were in girls with two of
these fGvllenhammar et al.. 2018: Bach etal.. 20161 showing decrements in both sexes. Results in
some studies were not consistent for categorical and continuous data expressions. For example,
birth weight deficits ranging from -21 to -34 g for quartiles 3 and 4 were seen in girls from the high
confidence Bach etal. (20161 study, but results were null per each ln-unit increase. Two of the four
studies noted above detected deficits in girls only (Maisonet etal.. 2012: Hiermitslev etal.. 20201.
The largest association in girls was seen in the medium confidence study by Hiermitslev et al.
(20201 ((3 = -76; 95% CI: -160, 7.7 per each ln-unit increase). The medium confidence Maisonet et
al. T20121 study showed some evidence of an exposure-response relationship ((3 range: -9 to -108 g
across PFHxS tertiles).
Five of the studies noted above showed deficits only in boys (Wikstrom etal.. 2020: Marks
etal.. 2019a: Lind etal.. 2017: Li etal.. 2017b: Cao etal.. 20181. Three studies that reported
decrements in boys (Wikstrom etal.. 2020: Marks etal.. 2019a: Lind etal.. 20171 showed
incongruent results based on continuous and categorical exposure data. For example, these studies
showed largely null results for each ln-unit increase but large deficits were seen for some upper
PFHxS exposure categories ((3 range: -51 to -104 g). The high confidence Wikstrom et al. f20201
study saw larger birthweight changes in the lowest 2 quartiles ((3 range: -39 to 51 g), but results
were largely null for quartile 4 and based on their continuous exposure data. A large deficit was
also seen in the medium confidence Li etal. f2017bl study ((3 = -53 g; 95% CI: -127, 20 per each ln-
unit increase). The low confidence Cao etal. (20181 study showed some evidence of an exposure-
response relationship in boys ((3 range: -30 to -109 g across tertiles), while results from the high
confidence Bach etal. (20161 were comparable in magnitude ((3 range: -16 to -21 g) based on the
upper three quartiles (compared with quartile 1) and for each ln-unit increase ((3 = -25 g). In the
medium confidence Gvllenhammar et al. (2018) study, results were stronger in males ((3 = -71 g;
95% CI: -150, 8 per each ln-unit increase) than females ((3 = -45 g; 95% CI: -139, -47 per each ln-
unit increase).
Birth weight - mean difference - sex-specific summary
Nine studies out of 15 (including 4 in girls and 7 in boys) showed some birth weight deficits
in relation to PFHxS exposures in either or both sexes but results were mixed within and across
3-130
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
studies. The magnitude of deficits was comparable among girls ((3 range: -45 to -76 g) and boys ((3
range: -13 to -71 g) per each ln-unitPFHxS increase; however, more studies showed deficits
among boys. No patterns were evident across confidence levels among boys, but the deficits seen in
girls were limited to medium and high confidence studies only. Two of the three low confidence
studies in boys showed inverse associations including one with evidence of an exposure-response
relationship based on categorical data. Among the five studies with categorical data, one study each
in boys and girls showed some suggestion of exposure-response relationships that were
comparable in magnitude (-108 and -109 g in tertile 3). Those results were larger in magnitude
but coherent with linear birth weight relationships detected in several studies with continuous
exposure metrics data as noted above (ranging from -25 to -76 g per each unit change in PFHxS).
Among these nine sex-specific studies, six had early biomarker samples indicative that
pregnancy hemodynamics was not likely an explanatory factor here. No other patterns by other
study characteristics were evident in the sex-specific findings including study sensitivity among the
null studies. Although the evidence may be somewhat stronger among males, the lack of consistent
patterns within and across studies and insufficiently sensitive studies to detect statistically
significant sex-specific associations preclude more definitive conclusions from being drawn.
3-131
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Regression coefficii
Good
Adequate
Aamus Birth Cohort (2008-2013).
SELMA(20D7-2010), Sweden, 1533
Trimester 1-2
er3
lit (ng'mL) inon
Quartile 2
Quaitile 3
QiiHrtile 4
Ouartllc 3
Quaitile 4
lit (ng'mL) irn
lit (ng'mL) irn
HIGH CONFIDENCE
| p [change in mean BWT (g)]
| |3 [change in mean BWT (g)] p<0.05
4 95% confidence inteival
Li, 2017. 3981358
Good
Good
Good
ACCEPT birth cohort (2010-2011.
2013-2015), Greenland. 482
mother-infant pairs
Hokkaido Study on Environment anc
Children's Health (2003-2009).
Japan, 1985 mother-child pairs
3 weeks post-birth
At birth
Trimester 1-2
lit (ng.'mL) irn
lit (ng'mL) irn
lit (ng'mL) irn
lit (ng'mL) im
lit (ng'mL) im
Zhoukou City Longitudinal Birth
Cohort (2013-2015). China. 282
mother-infant pairs
LOW CONFIDENCE
Shi, 2017,
3827535
Marks, 2019,
Deficient
Good
Tertile 3
lit (ng'mL) incr
-25C -ZOO -15'
Figure 3-24. Sex-specific male infants only mean birth weight results for 14
PFHxS epidemiological studies.a'b'c'd For additional details see HAWC link.
BWT = birth weight.
aStudies are sorted first by sex, overall study confidence level, then by exposure window(s) examined.
b(Meng et al., 2018) pooled samples from umbilical cord blood and maternal plasma during first and second
trimesters. The remaining studies were all based on either one umbilical or maternal sample.
c(Gvllenhammar et al., 2018) results are displayed here for mean birth weight among 587 overall population
participants in the POPUP Cohort compared with a smaller sample size of 381 in their 2018 publication.
dFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g. (Gvllenhammar et al., 2018)).
3-132
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Population
Study Sensitivity
Design
Exposure Window
Coefficient
Exposure Comparison
Regression coefficient
• p [change in mean BWT (g)]
Ashley-Martin,
2017.3001371
MIREC study (2008-2011) 1509
mother-infant pairs
Good
Cohort
(Prospective)
Trimester 1
-10.7
In-unit (ng.'mL) increase
HIGH CONFIDENCE ( ^
0 (3 [change in mean BWT (g)] p<0.05
H 95% confidence interval
Lind, 2017.
Odense Child Cohort (2010-2012)
At! equate
Cohort
Trimester 1
27
Quarlile 2
1
3838512
638 mother-infant pairs
(Prospective)
1 *
-37
Quart le 3
Quarlile 4
0
In-unit (ng.'mL) increase
1 1 1
Manzano-Salqado.
2017,4238465
INMA cohort (2003-2008) 1202
mother-Infant paiis
Ac equate
(Prospective)
Trimester 1
-13.D
In-unit (ng/mL) increase
Bach, 2016,
Aarhus Birth Cohort (2008-2013),
Adequate
Cohort
Trimester 1-2
6
Quart le 2
3981534
Denmark, 1507 mo1her-in*ant pairs
(Prospective)
(
-21
Quarlile 3
-34
Quarlile 4
1 •—1 1
-1.76
In-unit (ngfmL) increase
Wikstrem, 2020.
SELMA (2007-2010), Sweden, 1533
Good
Cohort
Trimester 1-2
30
Quarlile 2
1 _
6311677
mother-infant pairs
(Prospective)
1
28
Quart le 3
-16
Quarlile 4
1 • 1 1
-14
In-unit (ngfmL) increase
Valvi, 2017.
Faroe Islands (1997-2000),
Good
Cohort
Trimester 3
28.9
In-unit (ngf'mL) increase
3983872
Denmark, 604 mottier-infanl pairs
(Prospective)
Gyllenhammar,
2018.4238300
POPUP (1996-2011 )3B1
mother-infant pairs
Goad
Cross-sectional
3 weeks post-birth
-45.4
In-unil (ngfmL) increase
MEDIUM flONFinFNfiF , til
Li, 2017. 3981358
GBCS (2013), China. 321
mothor-lntant pairs
Good
Cross-sectional
At birth
-2.3
In-unit (ng.'mL) increase
1 • 1
4829851
~ NBC (1996-2002), Denmark, 3535
mother-infant pairs
Good
(Prospective)
Trimester 1-2
-6.4
In-unlt (ngfrnL) inciease
1 •> 1
Hjermitslev, 2020.
5880849
ACCEPT birth cohort (2010-2011,
2013-2015), Greenland, 482
mother-infant pairs
Adequate
Cohort
(Prospective)
Trimester 1 -3
-75.8
In-unit (ng.'mL) increase
1
Maisonel. 2012,
ALSPAC (1991-1092). U.K., 447
Good
Cohort
Trimester 1-3
-9.1
Terlile 2
1332465
mother-girl pairs
(Prospective)
-107.9
Tertile 3
• • ' 1
Koshino, 2020.
6311632
Hokkaido Study on Environment ant!
Children's Health (2003-2009),
Japan. 1985 mother-child pairs
Deficient
Cohort
(Prospective)
Trimester 3
10.51
In-unit (ng.'mL) increase
1—[•—1
Cao. 2018,
5080197
Zhoukou City Longitudinal Birth
Cohort (2013-2015). China. 282
molher-infanl pairs
Deficient
Cohort
(Prospective)
At birth
-5.1
Tertile 2
LOW CONFIDENCE
1
72.6
Shi. 2017.
Haidan Hospital (2012) 170
Deficient
Cross-sectional
At birth
43
In-unit (ng.'mL) increase
3827535
mother-infant pairs
1
-350 -300 -250 -200 -150 -100 -50 0 5D 1D0 150 200 250 300 350
Figure 3-25. Sex-specific female infants only mean birth weight results for 14
PFHxS epidemiological studies. For additional details see HAWC link.
BWT = birth weight.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bMeng et al. (2018) pooled samples from umbilical cord blood and maternal plasma during first and second
trimesters. The remaining studies were all based on either one umbilical or maternal sample.
cGvllenhammar et al. (2018) results are displayed here for mean birth weight among 587 overall population
participants in the POPUP Cohort compared with a smaller sample size of 381 in their 2018 publication.
dFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g. (Gvllenhammar et al., 2018)).
3-133
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Birth weight - standardized measures - background
Twelve of 13 studies in the overall population that reported continuous standardized birth
weight scores in relation to different PFHxS measures (see Figures 3-26 and 3-27), while the
Gardener et al. f20211 study (not included on the forest plot) examined odds of being in the lowest
standardized birthweight category (versus the top 3 birth weight z-score quartiles). Four of the 13
studies also reported sex-specific results (Xiao etal.. 2019: Wikstrom etal.. 2020: Gross etal.. 2020:
Eick etal.. 2020). while Gardener et al. (2021) only examined interactions across sex for
associations between PFHxS and standardized birth weight measures.
Among the 13 studies that examined PFHxS exposure in relation to standardized birth
weight scores in the overall population, eight were high fXiao etal.. 2019: Wikstrom etal.. 2020:
Shoaff etal.. 2018: Sagiv etal.. 2018: Gardener etal.. 2021: Eick etal.. 2020: Bach etal.. 2016:
Ashlev-Martin et al.. 2017). three were medium (Meng etal.. 2018: Hamm etal.. 2010:
Gvllenhammar et al.. 2018) and two were low (Workman et al.. 2019: Gross etal.. 2020) confidence.
Six studies had good (Wikstrom etal.. 2020: Shoaff etal.. 2018: Sagiv etal.. 2018: Meng etal.. 2018:
Gvllenhammar et al.. 2018: Ashlev-Martin et al.. 20171 study sensitivity ratings, while five were
adequate (Xiao etal.. 2019: Hamm et al.. 2010: Gardener etal.. 2021: Eick etal.. 2020: Bach etal..
20161 and two were deficient fWorkman et al.. 2019: Gross etal.. 20201.
Birth weight-standardized measures- study results
Null associations between PFHxS exposure and standardized birth weight scores were
reported in six studies (Workman etal.. 2019: Wikstrom etal.. 2020: Sagiv etal.. 2018: Hamm etal..
2010: Gvllenhammar etal.. 2018: Ashlev-Martin etal.. 2017) (see Figures 3-26 and 3-27). Similar to
results from categorical and continuous exposures in Wikstrom et al. f20201 and Sagiv etal. f 20181.
birth weight z-score results were largely null in relation to PFHxS tertiles in the high confidence
Eick etal. f20201 study in the overall population and across the sexes. They did report larger birth
weight z-scores in the overall population for tertile 3 ((3 = 0.15; 95% CI: -0.12, 0.42 compared with
tertile 1) that appeared to be driven primarily by results in females ((3 = 0.22; 95% CI: -0.18, 0.63).
The high confidence study by Gardener et al. (2021) detected nonsignificant increased odds for
their lowest standardized birthweight category (vs. the top three birth weight z-score quartiles)
across PFHxS quartiles (Q3 OR = 1.70; 95% CI: 0.81, 3.74); Q4 OR = 1.20; 95% CI: 0.55± 2.62). They
also found no statistically significant interactions for their birth weight z-score measures by sex.
Although their continuous exposure results were null per each ln-unit PFHxS increase, the
high confidence study by Bach etal. (2016) reported a small decrease in standardized birth weight
scores ((3 = -0.11; 95% CI: -0.25, 0.03) in PFHxS quartile 4 compared with quartile 1. Similar
results were seen for both tertiles 2 and 3 only ((3 range: -0.12 to -0.13) in the high confidence
Shoaff etal. (2018) study. Statistically significant results similar in magnitude were detected in the
medium confidence Meng etal. T20181 study ((3 = -0.14; 95% CI: -0.22, -0.07 per each ln-unit
PFHxS increase). Larger statistically significant lower birth weight z-scores results were reported in
3-134
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
the low confidence study by Gross et al. (2020) for the overall population ((3 = -0.65; 95% CI: -0.99,
-0.39), males (p = -0.60; 95% CI: -1.14, -0.06) and females (p = -0.77; 95% CI: -1.25, -0.29) for
PFHxS levels greater than the mean level of dried blood spot samples. Associations large in
magnitude per each ln-unit increase were also detected in the high confidence study by Xiao et al.
£2019} for the overall population ((3 = -0.74; 95% CI: -1.23, -0.26), male neonates ((3 = -0.62; 95%
CI: -1.28, 0.06), and female neonates ((3 = -0.87; 95% CI: -1.50, -0.22).
3-135
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Population
Study Sensitivity
Design
Exposure
Window
Regression
Coefficient
Exposure
Comparison
Regression coefficient
9 P [change in
O P [change in
BWT Z-Score]
BWT Z-Scorel p<0.05
Ashley-Martin,
2017. 3981371
MIREC study (2008-2011) 1509
mother-infant pairs
Good
Cohort
(Prospective)
Trimester 1
0.04
In-unit (ng'mL)
HH 35
% confidence interval
Bach et al.,2016.
3981534
Aarhus Birth Cohort (2008-2013),
Denmark. 1507 mother-infant pairs
Adequate
Cohort
(Prospective)
Trimester 1-2
-0.01
-0.03
-0.11
-0.05
Quartile 2
Quartile 3
Quartile 4
In-unit (ng'mL)
"•-It J"
• J
Sagiv. 2018.
4238410
Project Viva (1999-2002) 1645
mother-infant pairs
Good
Cohort
(Prospective)
Trimester 1-2
-0.1
0.04
0
0
Quartile 2
Quartile 3
Quartile 4
In-unit (ng'mL)
increase
i—•—pi
i '• i
i—•—i
Wikstrom, 2020.
6311677
SELMA (2007-2010), Sweden, 1533
mother-infant pairs
Good
Cohort
(Prospective)
Trimester 1-2
-0.004
-0.016
-0.008
0.007
Quartile 2
Quartile 3
Quartile 4
In-unit (ng'mL)
•i
i-i-i
Eick et al., 2020.
7102797
Chemicals In Our Bodies (CIOB)
(2014-2018). US, 497 female
participants
Adequate
Cross-sectional
Trimester 1-3
0.08
Tertila 2
1
0.15
Tertile 3
1
*
Shoaff et al..
2018, 4619944
HOME (2003-2006). United States.
345 mother-infant pairs
Good
Cohort
(Prospective)
Trimester 2-3, at
delivery
-0.12
-0.13
-0.09
Tertile 2
Tertile 3
In-unit (ngi'mL)
1 * 1
1
1 • 1 1
Xiao et al.. 2019,
Faroe Islands (1994-1995), 172
Adequate
Cohort
Trimester 3
-0.74
In-unit (ng'mL)
1
5918609
mother-infant pairs
(Prospective)
increase
~ |
Gyllenhammar,
2018.4238300
POPUP (1996-2011)381
mother-infant pairs
Good
Cross-sectional
3 weeks post-birth
-0.003
In-unit (ng'mL)
!
MengctaL 2018,
4829851
DNBC (1996-2002), Denmark, 3535
mother-infant pairs
Good
Cohort
(Prospective)
Trimester 1-2
-0.14
In-unit (ng'mL)
increase
f—•-) 1
Alberta cohort (2005-2006) 252
mother-infant pairs
Adequate
Trimester 2
1290814
(Prospective)
0.035
In-unit (ng'mL)
I—1# 1
Gross et al., 2020.
7014743
Starting Early Program (StEP)
Cohort. United Stales. 98
mother-infant pairs
Deficient
Nested
At birth
-0.65
high (>mean) vs.
|
1
Workman et al.,
2019. 5387046
Canadian Healthy Infant Longitudinal
Development (CHILD) Study
(2010-2012). Canada (414
mother-infant pairs)
Deficient
Cohort
(Prospective)
T rimester 2-3
-0.016
ln-unit(ng/mL)
increase
' 1 '
.4 -0.2 0 0.2
0
4 0.6 0
8 1
Figure 3-26. Overall population standardized birth weight results for 12
epidemiologic studies. For additional details see HAWC link.
BWT= birth weight.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
b(Xiao et al., 2019) results are truncated: the complete 95% CI ranges from -1.23 to -0.26 g.
Tor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g. (Gvllenhammar et al., 2018)).
3-136
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Population
Study Sensitivity
Design
Exposure Window
Regression
Exposure Comparison
Coefficient
Regression coefficient
# P [change in BWT Z-Score]
Wikstrom, 2020,
6311677
SELMA (2007-2010), Sweden, 1533
mother-infant pairs
Good
Cohort
(Prospective)
Trimester 1-2
-0.081
Quartile 2
BOYS
I
i—m-t-m
O P [change in BWT Z-Score] p<0.05
-0.083
Quartile 3
1 # ' 1
0.016
Quartile 4
1 • 1
-0.017
In-unit (ng/mL) increase
H-
Sagiv, 2018,
4238410
Project Viva (1999-2002) 1645
mother-infant pairs
Good
Cohort
(Prospective)
Trimester 1-2
-0.01
In-unit (ng/mL) increase
S—•-<
Eick, 2020,
7102797
Chemicals In Our Bodies (CIOB)
(2014-2018), US, 497 female
participants
Adequate
Cross-sectional
Trimester 1-3
0.14
Tertile 2
! —
1
Tertile 3
1 1
Xiao, 2019,
Faroe Islands (1994-1995), 172
Adequate
Cohort
Trimester 3
-0.62
In-unit (ng/mL) increase
5918609
mother-infant pairs
(Prospective)
Gross, 2020,
7014743
Starting Early Program (StEP)
Cohort, United States, 98
mother-infant pairs
Deficient
Nested
case-control
At birth
-0.6
high (>mean) vs. low
1
O 1
1
Wikstrom, 2020,
6311677
SELMA (2007-2010), Sweden, 1533
mother-infant pairs
Good
Cohort
(Prospective)
Trimester 1-2
0.071
Quartile 2
GIRLS
1—L«—1
0.062
Quartile 3
|—r#—1
-0.043
Quartile 4
i—#i—i
0.031
In-unit (ng/mL) increase
i—•—i
Sagiv, 2018,
4238410
Project Viva (1999-2002) 1645
mother-infant pairs
Good
Cohort
(Prospective)
Trimester 1-2
0.04
In-unit (ng/mL) increase
»-j#H
Eick, 2020,
7102797
Chemicals In Our Bodies (CIOB)
(2014-2018), US, 497 female
participants
Adequate
Cross-sectional
Trimester 1-3
0.06
Tertile 2
1 1 •
0.22
Tertile 3
Xiao, 2019,
Faroe Islands (1994-1995), 172
Adequate
Cohort
Trimester 3
-0.87
In-unit (ng/mL) increase
5918609
mother-infant pairs
(Prospective)
Gross, 2020,
7014743
Starting Early Program (StEP)
Cohort, United States, 98
mother-infant pairs
Deficient
Nested
case-control
At birth
-0.77
high (>mean) vs. low
o
1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4
0.6 0.8
1.2 1.4
Figure 3-27. Sex-stratified standardized birth weight results for five epidemiologic studies (boys above reference
line, girls below). For additional details see HAWC link.
BWT= birth weight,
aStudies are sorted first by overall study confidence level, then by Exposure Window examined.
b(Xiao et al., 2019) results are truncated: the complete 95% CI ranges from -1.5 to -0.22.
3-137
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Birth weight - summary of different measures and analyses
Twenty of 34 studies in total that examined either mean or standardized birth weight
showed some deficits in relation to PFHxS exposures. This included six of 13 studies that showed
inverse associations between PFHxS and standardized birth weight measures in the overall
population. Among the 12 studies examining continuous birth weight measures in the overall
population, 3 showed some associations of at least -0.1 in relation to either categorical or
continuous PFHxS exposures. Two other studies (one high and one low confidence) showed
stronger associations in excess of-0.74 as well as comparable results in both sexes. The high
confidence study by Gardener et al. (20211 also reported nonsignificant odds of being in the lowest
standardized birthweight category (vs. the top 3 B WT z-score quartiles) based on PFHxS quartiles 3
(OR range: 1.20 to 1.74). There was limited evidence of exposure-response relationships in support
of the continuous study results expressed per a unit change. Few patterns and minimal differences
were seen across sexes. Among the six studies in the overall population that showed some
suggestion of inverse associations, two studies (one high and one low confidence) reported large
associations consistent in magnitude for both male and female neonates. Study sensitivity did also
not seem to explain null study findings as four of these six studies had good ratings in this domain.
There was a slight preponderance of inverse associations with four of the six studies using later
biomarker samples.
Overall, 17 of the 31 epidemiological studies with mean birth weight in either/both sex or
the overall population detected some deficits in relation to PFHxS exposures (see Table 3-17),
although these deficits were at times limited to sex-specific findings (Marks etal.. 2019a: Maisonet
etal.. 2012: Lind etal.. 20171 and often were not statistically significant (see Figures 3-21, 3-22, 3-
24, and 3-25). This included 14 (4 low and 5 each medium and high confidence) of the 27 studies in
the overall population. Two different studies (out of 14) with categorical data in the overall
population or either sex showed some evidence of exposure-response relationships. Overall, the
magnitude of changes in those studies showing deficits ranged from -25 to -109 g for the highest
quantile (compared with the lowest quantile). Those results were consistent in magnitude with 12
studies with continuous exposure metrics data showing birth weight-related deficits with
increasing exposures in the overall population ((3 ranging from -12 to -76 g per each unit change in
PFHxS). Seven of these ranged from -12 to -30 g and the remaining five ranged from -49 to -76 g.
These data were supported by an EPA meta-analysis that showed also showed a small birth weight
deficit (P = -7.9 g; 95% CI: -15.0, -0.7) per each ln-unit PFHxS among all 27 studies and were
consistent in magnitude ((3 range: -7 to -10 g) across 12 high confidence studies, 11 medium
confidence studies, and the combined high and medium studies. Although deficits were largest
among postpartum samples, the results among the 12 early sampled studies were comparable
(P = -7.6 g; 95% CI: -16.2,1.1) to that seen in the overall population of all 27 studies. When further
restricted to the earliest six studies, the results were closer to the null ((3 = -3.5 g; 95% CI: -14.8,
7.9).
3-138
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Limited patterns were evident in the mean birth weight findings as overall confidence,
study sensitivity, exposure levels and other study design elements were not explanatory for the null
or inverse associations. The mean birth weight differences in the overall population may be
influenced by hemodynamic changes during pregnancy, as 9 of the 14 were based on late
biomarker sampling. Similar to that seen for standardized birth measures, the sex-specific data
were more mixed in relation to sample timing as four of six studies showing birth weight deficits
were based on late biomarker collection.
Birth length - background of studies
Nineteen studies examined the relationship between PFHxS exposures and birth length in
the overall population or across sexes; one study fAlkhalawi et al.. 20161 was classified as
uninformative and is not discussed here (see Figure 3-28). Two of the 10 studies reporting sex-
specific findings did not report overall population results; both studies were from the ALSPAC
population, including a study in boys (Marks etal.. 2019a] and girls fMaisonet etal.. 20121 Two
studies (Xiao etal.. 2019: Gvllenhammar etal.. 2018) reported standardized birth length measures,
while the remaining studies examined mean birth length differences in relation to PFHxS. As noted
above, two studies (Bierregaard-Olesen etal.. 2019: Bach etal.. 2016) from the Aarhus birth cohort
are discussed when discrepancies arise or in isolation as for some sex-specific findings. They are
both listed together below in the background materials just below but only counted as one study
when evaluating consistency and between-study heterogeneity patterns.
Six of the 18 included PFHxS studies examining birth length studies were classified as high
(Xiao etal.. 2019: Valvi etal.. 2017: Manzano-Salgado etal.. 2017a: Luo etal.. 2021: Buck Louis et
al.. 2018: Bierregaard-Olesen etal.. 2019: Bach etal.. 20161. and five were medium fMaisonet etal..
2012: Kashino etal.. 2020: Hiermitslev etal.. 2020: Gvllenhammar etal.. 2018: Chen etal.. 2021)
confidence. Seven of birth length studies were classified as low confidence fXu etal.. 2019:
Workman et al.. 2019: Shi etal.. 2017: Marks etal.. 2019a: Gao etal.. 2019: Cao etal.. 2018: Callan et
al.. 2016) largely due to concerns with participant, selection, confounding, and study sensitivity. For
example, seven of those studies were considered deficient for study sensitivity (Xu etal.. 2019:
Workman et al.. 2019: Shi etal.. 2017: Kashino et al.. 2020: Gao etal.. 2019: Cao etal.. 2018: Callan
etal.. 20161. Five studies were rated good fValvi etal.. 2017: Marks etal.. 2019a: Maisonetetal..
2012: Luo etal.. 2021: Gvllenhammar etal.. 2018) and six were adequate (Xiao etal.. 2019:
Manzano-Salgado etal.. 2017a: Hiermitslev etal.. 2020: Chen etal.. 2021: Buck Louis et al.. 2018:
Bjerregaard-Olesen etal.. 2019: Bach etal.. 2016).
Birth length - overall population results
Nine of the 16 studies in the overall population reported shorter birth length in relation to
PFHxS exposure (see Figure 3-29; Table 3-17). Five of the six high confidence studies observed that
PFHxS exposure was associated with shorter birth length in at least one comparison set, including
statistically significant changes in three high confidence studies examining mean fManzano-Salgado
3-139
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
etal.. 2017a: Buck Louis et al.. 2018) or standardized birth length measures (Xiao etal.. 20191. For
example, Xiao etal. f20191 reported smaller birth length z-scores in overall population ((3 = -0.52;
95% CI: -1.04, -0.13 each ln-unitincrease). The Manzano-Salgado etal. f2017al study reported
birth length reductions consistent in magnitude across all three PFHxS quartiles ((3 range: -0.31 to
-0.33 cm), although results were largely null for each ln-unit increase ((3 = -0.09; 95% CI: -0.25,
0.09). The study by Valvi etal. (2017) reported small deficits in mean birth length in the overall
population ((3 = -0.14 cm; 95% CI: -0.35, 0.04). Given a ln-unit PFHxS increase, null results were
reported in the Bach etal. (2016) study, and their smaller subset analysis (n = 671 participants)
reported in Bierregaard-Olesen etal. f20191 (the latter data are not plotted given from same
cohort). The Bach etal. f 2 0161 study based on 1,507 participants did report decreased birth length
in the third ((3 = -0.1 cm; 95% CI: -0.5, 0.3) and fourth ((3 = -0.2 cm; 95% CI: -0.5, 0.2) quartiles
compared with the lowest quartile (not included on Figure 3-29 given overlapping population). The
study by Buck Louis et al. (2018) reported that PFHxS was associated with reductions in birth
length (and upper thigh length; the latter data not shown) in the overall population ((3 = -0.22 cm;
95% CI: -0.39, -0.05 per each ln-unit increase), as well as Black ((3 = -0.43 cm; 95% CI: -0.71,
-0.14) and Hispanic neonates ((3 = -0.34 cm; 95% CI: -0.70, 0.03).
Three out of four medium confidence studies in the overall population were null for birth
length deficits in relation to PFHxS exposures. The Chen etal. f20211 study reported a small deficit
((3 = -0.15 cm; 95% CI: -0.42, 0.11) per each ln-unit increase and nonmonotonic consistent deficits
across quartiles ((3 range: -0.33 to -0.46 cm). Three out of five low confidence studies reported
some suggestion of birth length deficits in relation to PFHxS. Although results were null for tertile 3
relative to tertile 1, the low confidence study by Cao etal. (2018) reported a statistically significant
result ((3 = -0.33 cm; 95% CI: -0.68, -0.01) for tertile 2. Compared with tertile 1, the low confidence
study by Gao etal. f 20191 reported a statistically significant result ((3 = -0.43 cm; 95% CI: -0.78,
-0.07) for tertile 2 but a smaller deficit in tertile 3 ((3 = -0.20 cm; 95% CI: -0.64, 0.25). Callan et al.
(2016) reported an imprecise deficit of -0.20 cm (95% CI: -0.78, 0.38) per each ln-unit increase. In
contrast, Xu etal. (2019) reported a large increased birth ((3 = 0.66 cm; 95% CI: -0.01,1.26 per each
ln-unit increase).
Overall, 9 (5 high, 1 medium, and 3 low confidence) out of 16 studies in the overall
population provided some evidence of birth length deficits with increasing PFHxS exposure. Some
of these results were not always internally consistent across different exposure expressions
(continuous vs. categorical). The five studies with categorical data in the overall population did not
provide any evidence of any exposure-response relationships. Although mean birth length results
for continuous PFHxS exposures were smaller, two of the three studies with PFHxS quartiles
showed deficits similar in magnitude ((3 = -0.31 to -0.46 cm). There was a consistent pattern by
sample timing among those studies demonstrating birth length deficits in the overall population, as
six of the nine studies were based on late biomarker sampling. No other patterns by study
characteristics were evident
3-140
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Participant selection
Exposure measurement
Outcome ascertainment
Confounding
Analysis
Sensitivity
Selective Reporting
Overall confidence
Legend
I Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-28. Study evaluation results for 19 epidemiological studies of birth
length and PFHxS. For additional details see HAWC link.
3-141
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Population
Study Sensitivity
Design
Exposure
Window
Regression
Coefficient
Exposure
Comparison
Regression coefficient
0 p [change in mean BL (cm)]
Mari2ano-Salgado
el al., 2017,
4238465
INMA cohort (2003-2008) 1202
molher-infanl pairs
Adequate
Cohort
(Prospective)
Trimester 1
-0.33
Quartils 2
HIGH CONFIDENCE
h-195% confidence interval
1
-0.32
Quarlile 3
-0.31
Quartila 4
-0.09
ln-unit (ng/mL)
1
Buck Louis. 2018.
5016992
NICHD Fetal Growth Studios
(2009-2013), United States. 2106
mother-infant pairs
Adequate
Cohort
(Prospective)
Trimester 1
-0.22
ln-unit (ng/mL)
1
1 • 1 1
Bach et al.. 2016.
3981534
Aarhus Birth Cohort (2008-2013),
Denmark, 1507 mother-infant pairs
Adequate
Cohort
(Prospective)
Trimester 1-2
0
ln-unit (ng/mL)
0.1
Quartile 2
-0.1
Ouartile 3
1 •—| 1
-0.2
Quarlile 4
> m i i
Lunetal., 2021
9959610
Zhujiang Hospital Cohort. China
(2017-2019) 224 mother-infant pairs
Good
Cohort
(Prospective)
Trimester 3
-0.001
Quartila 2
Quartila 3
Quartilo 4
i i i
i
i
0.04
ln-unit (ng,'mL)
Valvietal., 2017,
3983872
Faroe Islands (1997-2000),
Denmark, 604 mother-infant pairs
Good
Cohort
(Prospective)
Trimester 3
-0.14
ln-unit (ng/mL)
H #—! H
Xiaoet al,, 2019,
Faroe Islands (1994-1995), 172
Adequate
Cohort
Trimester 3
-0.59
ln-unit (ng/mL)
5918609
mother-infant pairs
(Prospective)
increase
Gyllenhammar,
2018. 4238300
POPUP (1996-2011) 381
mother-infant pairs
Good
Cross-sectional
3 weeks post-blrtt
0
ln-unit (ng/mL)
MEDIUM CONFIDENCE
Chen. 2021.
7263985
Prospective cohort analysis from
Shanghai Birth Cohort (2015-2017).
214 mother-infant pairs
Adequate
Cohort
(Prospective)
Trimester 1-2
-0.4
Quart ile 2
1
Quartile 3
-0.33
Quartile 4
-0.15
ln-unit (ng/mL)
1
1 # 1-1
Hjarmitslav, 2020,
5880849
ACCEPT birth cohort (2010-2011,
2013-2015), Greenland. 482
mother-infant pairs
Adequate
Cohort
(Prospective)
Trimester 1-3
-0.18
ln-unit (ng/mL)
1
Kashino. 2020.
6311632
Hokkaido Study on Environment and
Children's Health (2003-2009),
Japan. 1985 mother-child pairs
Deficient
Cohort
(Prospective)
Trimester 3
0.01
ln-unit (ng/mL)
-f-
Cao et al., 2018,
5080197
Zhoukou City Longitudinal Birth
Cohort (2013-2015), China. 282
mother-infant pairs
Deficient
Cohort
(Prospective)
At birth
-0.33
Tertile 2
LOW CONFIDENCE
1
-0.07
Tertile 3
•-T
Shi. 2017.
Haidan Hospital (2012) 170
Deficient
Cross-sectional
At birth
0.17
ln-unit (ng/mL)
3827535
mother-infant pairs
increase
1
Xu, 2019,
Cross-sectional study (2016-2017).
Deficient
Cross-sectional
At birth
0.66
ln-unit (ng/mL)
5381338
China, 98 mother-infant pairs
increase
Workman et al.,
2019. 5387046
Canadian Healthy Infant Longitudinal
Development (CHILD) Study
Deficient
Cohort
(Prospective)
Trimester 2-3
-0.012
In-unit(ng.'mL)
(2010-2012), Canada (414
mother-infant pairs)
1
Callan. 2016.
AMETS (2008-2011), Australia, 98
Deficient
Cross-sectional
Trimester 3
-0.2
ln-unit (ng/mL)
• ' <
3858524
mother-infant pairs
increase
1
Gao el al., 2019,
5387135
Affiliated Hospital of Capital Medical
University (2015-2016), China. 132
pregnant women
Deficient
Cohort
(Prospective)
Trimester 3
-0,43
Tertile 2
1
"" " 1
-0.2
Tertile 3
14 12 -1 0 8
6 -0.4 -0.2 0 0.2
4 0.6 0.8 1 1
2 1.4
Figure 3-29. Overall population mean birth length results for 16 PFHxS
epidemiological studies. For additional details see HAWC link.
BL = birth length.
'Studies are sorted first by overall study confidence level then by Exposure Window examined.
b(Xiao et al., 2019) and (Gyllenhammar et al., 2018) in blue text report birth length z-score data; the remaining
studies evaluate mean birth length differences.
cFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g. (Gyllenhammar et al., 2018)).
Birth length - sex-specific results
Among these 11 studies with results in either boys, girls or both, some birth length deficits
were detected in 7 different studies (see Figure 3-30). The high confidence study by Xiao et al
(20191 reported deficits in both sexes including larger and statistically significant birth length z-
scores among girls (p> = -0.72; 95% CI: -1.33, -0.12 each ln-unit increase). Sex-specific results were
null based in both sexes based on continuous (per each ln-unit increase) data in the Manzano-
Salgado et al. f2017al and Kashino et al. ( 20201 studies. Four of the remaining six studies in females
3-142
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
were null (Shi etal.. 2017: Chen etal.. 2021: Cao etal.. 2018: Bjerregaard-Olesen etal.. 20191. The
medium confidence Maisonet et al. T20121 study of girls only reported dose-dependent statistically
significant associations across exposure tertiles ((3 range: -0.52 to -0.82). The medium confidence
Hiermitslev etal. f20201 study reported deficits among female neonates only ((3 = -0.42 cm; 95%
CI: -1.07, 0.22 per each ln-unit increase).
The medium confidence Chen etal. (2021) study reported a birth length deficit
((3 = -0.15 cm; 95% CI: -0.61, 0.31 per each ln-unit increase) small in magnitude in boys only. The
high confidence study by Valvi etal. (2017) reported deficits among male neonates only
((3 = -0.22 cm; 95% CI: -0.49, 0.04 per each ln-unit increase). The low confidence study by Cao et al.
T20181 detected nonmonotonic reductions in birth length across tertiles ((3 range: -0.18 to -0.44)
in boys, while another low confidence study of boys only f Marks etal.. 2019al detected evidence of
an exposure-response relationship across PFHxS tertiles ((3 range: -0.25 to -0.39). In contrast,
increased birth length ((3 range: 0.20 to 0.40 cm per ln-unit PFHxS increase) was detected in males
in three studies (Shi etal.. 2017: Hiermitslev et al.. 2020: Bierregaard-Olesen etal.. 2019).
Figure 3-30. Thumbnail schematic of Sex-stratified birth length results for 11
epidemiologic studies (boys above reference line, girls below). For additional
details and for interactive data graphic see HAWC link.
BL = birth length.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bXiao et al. (2019) in blue text reports birth length z-score data.
Tor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g. (Cao et al., 2018)).
3-143
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Summary - Birth Length-Sex-Specific
Stronger evidence of birth length deficits was observed in males (5 of 10 studies) compared
with females (3 of 10 studies); however, these deficits were generally smaller in magnitude among
males ((3 range: -0.15 to -0.39 cm) than females ((3 range: -0.42 to -0.82 cm). In addition to the two
null studies in males, three other studies reported increased birth length in relation to PFHxS
exposures. Two of the three studies with categorical data provided evidence of an inverse
exposure-response relationships, albeit only in males (Marks etal.. 2019a) and females (Maisonet
etal.. 2012) derived from the same ALPSAC study population.
Exposure levels were higher in the studies reporting birth length deficits in males, including
the top four and five of the top six highest exposure measures of centrality reported. Besides this
and the slightly more consistent results in males in general, no other patterns across study
characteristics explained the between-study heterogeneity including the null results. For example,
there was no definitive pattern of results by study confidence across the seven different studies
(two high, three medium, and two low confidence) nor sample timing (four had early biomarker
samples compared with three with late).
Summary - Birth Length
Overall, 12 out of 18 included studies provided some evidence of birth length deficits with
increasing PFHxS exposure in either the overall population or either sex. Some of these results were
not always internally consistent across different exposure expressions (continuous vs. categorical).
Two of the seven studies with categorical data provided some evidence of any exposure-response
relationships, both of these were from sex-specific studies in the same cohort. There was no pattern
among the null studies based on study sensitivity or other study characteristics. Mean and median
exposure levels were higher among the male studies showing deficits, but this did not appear to
explain results in females or the overall population. There was not a consistent pattern by sample
timing among the studies showing inverse associations in either/both sex (four of seven had early
sampling) or the overall population (three of nine had early sampling). Among the 11 different
studies demonstrating birth length deficits, six of them relied on early sampling suggesting limited
overall potential impact of pregnancy hemodynamics.
Head circumference at birth - study background
Fourteen studies examined PFHxS in relation to head circumference measured at birth
including two studies fXiao etal.. 2019: Gvllenhammar et al.. 20181 reporting standardized head
circumference measures (see Figure 3-31). Among the other 12 studies with mean head
circumference data, 10 of these studies reported data in the overall population (Xu et al.. 2019:
Workman et al.. 2019: Valvi etal.. 2017: Manzano-Salgado etal.. 2017a: Kashino etal.. 2020:
Hiermitslev etal.. 2020: Chen etal.. 2021: Callan etal.. 2016: Buck Louis et al.. 2 018: Bierregaard-
Olesen etal.. 20191: Bach etal. f20161. Eight studies analyzed sex-specific results including two
studies fMarks etal.. 2019a: Lind etal.. 20171 that only reported sex-specific data.
3-144
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Four studies were classified as low confidence (Xu etal.. 2019: Workman et al.. 2019: Marks
etal.. 2019a: Callan etal.. 20161 and five each were medium fLind etal.. 2017: Kashino etal.. 2020:
Hiermitslev etal.. 2020: Gvllenhammar etal.. 2018: Chen etal.. 20211 and high fXiao etal.. 2019:
Valvi etal.. 2017: Manzano-Salgado etal.. 2017a: Buck Louis etal.. 2018: Bierregaard-Olesen et al..
20191: Bach etal. f20161. Seven of the 14 PFHxS studies on head circumference had adequate study
sensitivity (Xiao etal.. 2019: Manzano-Salgado et al.. 2017a: Lind etal.. 2017: Hjermitslev etal..
2020: Chen etal.. 2021: Buck Louis etal.. 2018: Bierregaard-Olesen etal.. 20191. while four were
deficient fXu etal.. 2019: Workman et al.. 2019: Kashino etal.. 2020: Callan etal.. 20161 and three
had good study sensitivity fValvi etal.. 2017: Marks etal.. 2019a: Gvllenhammar etal.. 20181.
;>,e£
Participant selection •
Exposure measurement
Outcome ascertainment-
Confounding -
Analysis
Sensitivity -
Selective Reporting -
Overall confidence -
—L
^ \^0
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
of five high confidence studies, two of four medium and one of three low confidence studies (see
Figure 3-32). Three studies detected null associations fXu etal.. 2019: Kashino etal.. 2020:
Gvllenhammar et al.. 20181. Two studies reported small increases in head circumference per each
ln-unit increase including the high confidence Valvi etal. f20171 study ((3 = 0.16 cm; 95% CI: 0.01,
0.29) and the low confidence Workman etal. f20191 study ((3 = 0.12 cm; 95% CI: -0.18, 0.42).
The high confidence Xiao etal. (2019) study reported lower head circumference z-scores in
the overall population ((3 = -0.52; 95% CI: -1.04, 0.00 per each PFHxS ln-unit increase). The high
confidence study by Bach etal. (2016) detected consistent deficits across quartiles two through
four (all (3 coefficients were -0.2 cm), but they reported null findings based on the continuous
PFHxS measure as well as in their smaller subset in a separate publication fBierregaard-Olesen et
al.. 20191 (the latter data are not plotted given from same cohort). Similarly, the high confidence
study by Manzano-Salgado etal. f2017al showed some evidence of an exposure-response
relationship across the PFHxS quartiles ((3 range: -0.08 to -0.16) but not among the continuous
exposure results ((3 = -0.01 cm; 95% CI: -0.13, 0.10). The high confidence study by Buck Louis et al.
(2018) reported a precise but small deficit in the overall population ((3 = -0.09 cm; 95% CI: -0.19,
0) and saw a statistically significant reduction in head circumference for Black neonates ((3 = -0.25
cm; 95% CI: -0.41, -0.08) per each ln-unit increase in PFHxS. Two medium confidence studies
detected an imprecise head circumference difference of -0.14 cm per each ln-unit PFHxS increase
including Hjermitslev et al. (2020) (95% CI: -0.52, 0.25) and Chen etal. (2021) (95% CI: -0.46,
0.19). A larger difference was detected in the low confidence Callan etal. (2016) study
((3 = -0.31 cm; 95% CI: -0.74, 0.12 per each ln-unit PFHxS increase).
Overall, 7 of 12 studies showed some evidence of associations between PFHxS and different
head circumference measures in the overall population. Some of these results were not always
internally consistent across different exposure expressions (continuous versus categorical). One of
two studies with categorical data showed some evidence of an exposure-response relationship
across quartiles. There was no clear pattern in study characteristics among the null studies,
although two of the four had deficient study sensitivity. Five of the seven studies were based on
early biomarker samples, so pregnancy hemodynamics did not appear to explain the study findings.
3-146
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Population
Overall Study
Study Sensitivity
Design
Exposure
Regression
Exposure
Confidence
Window
Coefficient
Comparison
Regression coefficient
9 P [change in mean HC (cm)]
Manzano-Salgado
etal.. 2017,
INMA cohort (2003-2008) 1202
mother-infant pairs
|High|
Adequate
Cohort
(Prospective)
Trimester 1
-0.08
Quarti le 2
Quartile 3
—•—i—1
O P [change in mean HC (cm)] p<0.05
I—| 95% confidence interval
-0.01
ln-unit (ng,'mL)
1 # 1
Buck Louis. 2018,
5016992
NICHD Fetal Growth Studies
<2009-2013), United States. 2106
mother-infant pairs
|High|
Adequate
Cohort
(Prospective)
Trimester 1
-0.09
ln-unit (ng/mL)
1 • 1
1
Bach el hI.. 2016,
3981534
Anrhiis Birth Cohort (2008-2013),
Denmark, 1507 mother-infant pairs
|High|
Adequate
Cohort
(Prospective)
THmasler 1-2
0
In-unil (ng/mL)
1
-0.2
Quartilc 4
1 • 1 1
Valui et al.. 2017,
Faroe Islands (1997-2000).
Good
Cohort
T rimester 3
0.16
ln-unit (ng/mL)
3983672
Denmark, 604 mother-infant pairs
(Prospective)
increase
Xiao etal.. 2019,
Faroe Islands (1994-1995). 172
I High |
Adequate
Cohort
Trimester 3
-0.52
ln-unit (ng/mL)
5918609
mother-infant pairs
(Prospective)
Gyllenhammar,
2018.4238300
POPUP (1996-2011) 381
mother-infant pairs
I Medium|
Good
Cross-sectional
3 weeks post-birth
-0.03
ln-unit (ng/mL)
Chen, 2021,
7283985
Prospective cohort analysis from
Shanghai Birth Cohort (2015-2017).
214 mother-infant pairs
| Medium |
Adequate
Cohort
(Prospective)
Trimester 1-2
-0.14
ln-unit (ng/mL)
1
' • 1 '
Hjermitslev. 2020,
5880849
ACCEPT birth cohort (2010-2011,
2013-2015), Greenland, 482
mother-infant pairs
| Medium |
Adequate
Cohort
(Prospective)
Trimester 1-3
-0.14
ln-unit (ng/mL)
^ I
I
Kashino, 2020.
6311632
Hokkaido Study on Environment and
Children's Health (2003-2009),
Japan, 1985 mother-child pairs
| Medium |
Deficient
Cohort
(Prospective)
Trimester 3
-0.07
ln-unit (ng/mL)
1 •—1—1
1
Xu. 2019.
Cross-sectional study (2016-2017).
|Lo\v|
Deficient
Cross-sectional
At birth
0.05
ln-unit (ng/mL)
5381338
China. 98 mother-infant pairs
increase
Workman etal.,
2019, 5387046
Canadian Healthy Infant Longitudinal
Development (CHILD) Sluriy
|Low|
Deficient
Cohort
(Prospective)
Trimester 2-3
0.12
ln-unit{ng/niL)
1
(2010-2012), Canada (414
mother-in fanl pairs)
1
Call an, 2016.
AMETS (2008-2011). Australia. 98
|Low|
Deficient
Cross-sectional
Trimester 3
-0.31
ln-unit (ng/mL)
3858524
mother-infant pairs
increase
-G
8 -0.6 -0.4 -0.2 0 0.2
0.4 0.6 0.8
Figure 3-32. Overall population head circumference results for 12
epidemiologic studies. For additional details see HAWC link.
HC = head circumference.
aStudies are sorted first by overall study confidence level, then by Exposure Window(s) examined.
bXiao et al. (2019) and Gvllenhammar et al. (2018) in blue text report head circumference z-score data.
cXiao et al. (2019) results are truncated: the complete 95% CI ranges from -1.04 to 0.
dFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g. (Xu et al., 2019)).
Head circumference at birth - sex and race-specific results
Eight studies examined PFHxS and head circumference differences among sexes (see
Figure 3-33). Two high confidence studies were null in both sexes fManzano-Salgado etal.. 2017a:
Bierregaard-Olesen et al.. 20191 and only one study fXiao etal.. 20191 showed inverse associations
in both sexes. Four of eight studies were null in boys, and one showed larger head circumference
differences with increasing PFHxS exposures. Five studies were null in girls and two studies
showed inverse associations between head circumference differences and PFHxS exposures.
Three of eight studies in boys and two of seven studies in girls reported inverse associations
with PFHxS. The high confidence study by Xiao etal. T20191 reported smaller head circumference z-
scores with larger results in female ((3 = -0.76; 95% CI: -0.19, 0.23 per each ln-unit increase)
compared with male ((3 = -0.26; 95% CI: -0.46, 0.07 per each ln-unit increase) neonates. All of the
other studies examined mean head circumference differences in relation to PFHxS. For example, the
medium confidence study by Hiermitslev etal. (20201 showed head circumference differences
among females only ((3 = -0.26; 95% CI: -0.73, 0.20 per each ln-unit increase). Among boys, the
medium confidence study by Kashino etal. f20201 reported head circumference differences smaller
in magnitude relation to PFHxS ((3 = -0.14 cm; 95% CI: -0.29, 0.02 per each ln-unit increase), as did
3-147
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
the medium confidence study by Lind etal. (2017) ((3 = -0.1 cm; 95% CI: -0.4, 0.2 per each ln-unit
increase). The Lind etal. f20171 study showed nonmonotonic head circumference deficits across
exposure categories ((3 range: -0.1 to -0.7 cm), including one that was statistically significant for
PFHxS quartile 3 (p = -0.7 cm; 95% CI: -1.2, -0.2).
Overall, four (1 high; 3 medium confidence) of eight studies showed some evidence of
associations between PFHxS and different head circumference measures among either or both
sexes (including three of eight studies in boys and two of seven studies in girls). No study
characteristics (i.e., study design features or study quality domains) appeared to explain between-
study heterogeneity of results including sample timing as half of the studies reporting inverse
association were based on early biomarker samples.
Regression coefficient
2019,5083648 mot
Valvi, 2017,
Trimester 1-2
(Prospective)
Cohort Trimester 3
-0.27
Trimester 1
ln-unit (ng/mL) in
0.26 ln-unit (ng/mL) in<
ln-unit (ng/mL) incre
-0.4
! in mean HC (cm)]
! in mean HC (cm)] p<0.051
H 95% confidence interval
Quartile 2
Quartile 3
¦>?&., Q,iidw.onJEraiBtiy-wrLfwriu ¦...fJ/wii'"
6311632 Children's Health (2003-2009),
Japan, 1985 mother-child pairs
Marks, 2019, ALSPAC (1991-1992), England, 457
(Prospective)
lit (ng/mL) increas
unit (ng/mL) inc
-0.05
-0.01
Tertile 3
i-unit (ng/mL) incre:
irhus Birth Cohort (2008-2013) 702
Faroe Islands (1997-2000),
Cohort
Adequate
ln-unit (ng/mL) increase GIRLS
ln-unit (ng/mL) increase
3 0.01 ln-unit (ng/mL) increase
Trimester 1 0.2 Quartile
Quartil
tit (ng/mL
ACCEPT birth cohort (2010-20
2013-2015), Greenland, 482
Children's Health (2003-2009),
Japan, 1985 mother-child pairs
Figure 3-33. Sex-stratified head circumference results for eight epidemiologic
studies (boys above reference line, girls below). For additional details see HAWC
link.
HC = head circumference.
aStudies are sorted first by overall study confidence level, then by Exposure Window(s) examined.
bXiao et al. (2019) in blue text report head circumference z-score data.
3-148
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Head circumference summary
Overall, 8 of 14 total studies showed some head circumference deficits in either sex or in the
overall population in relation to PFHxS exposures. There was fairly consistent evidence of
associations in the overall population as 6 out of 12 studies (including five of the nine high and
medium confidence studies) reported some evidence of deficits for at least one exposure
comparison. Overall, one of the three studies with categorical data showed evidence of an exposure-
response relationship in either sex or in the overall population. There was no pattern among the
null studies based on study sensitivity and exposure levels/contrasts. There was not a consistent
pattern by sample timing among those studies demonstrating head circumference deficits, as half
other studies in both the overall population and sex-specific analyses that were based on late
biomarker sampling.
Small for gestational age and low birth weight
Seven epidemiological studies included here examined associations between PFHxS
exposure and different dichotomous fetal growth restriction endpoints, such as SGA (or related
intrauterine growth retardation endpoints) fXu etal.. 2019: Wikstrom etal.. 2020: Hamm etal..
2010: Chang etal.. 20221 or low birth weight (LBW) fMeng etal.. 2018: Manzano-Salgado etal..
2017a: Hiermitslev et al.. 20201 (see Figure 3-34). Two studies were high confidence fWikstrom et
al.. 2020: Manzano-Salgado etal.. 2017a). three were medium confidence Meng etal. (2018):
(Hjermitslev etal.. 2020: Hamm etal.. 2010) and two were low confidence (Xu etal.. 2019: Chang et
al.. 2022). Two of these studies had good study sensitivity (Wikstrom etal.. 2020: Manzano- Salgado
etal.. 2017a). four had adequate study sensitivity (Wikstrom et al.. 2020: Manzano-Salgado etal..
2017a: Hiermitslev et al.. 2020: Chang etal.. 20221 while one was deficient fXu etal.. 20191. All
seven studies reported results in the overall population, while two fWikstrom etal.. 2020:
Manzano-Salgado etal.. 2017al provided results in both the overall population and across sexes.
Three (Xu etal.. 2019: Wikstrom etal.. 2020: Hamm etal.. 2010) of four SGA studies showed
some adverse associations (see Figure 3-35) in relation to PFHxS. The medium confidence study by
Hamm etal. (2010) showed increased odds (OR=2.35; 95% CI: 0.63, 8.72) in the overall population
among tertile 3 compared with tertile 1. The low confidence by Xu etal. (2019) reported showed an
even larger statistically significant odds of SGA (OR=9.14; 95% CI: 1.15, 72.8 per each ln-unit
increase). Although their overall population results were null, some of the quartile results were
elevated (OR=1.76; 95% CI: 0.79, 3.90) but in a nonmonotonic fashion. Their results based on a ln-
unit increase were largely null for both sexes. In addition to the Wikstrom et al. (2020) study, two
other studies in the overall population were null (Hiermitslev etal.. 2020: Chang etal.. 2022). The
Manzano-Salgado etal. (2017a) study was null for the overall population, girls, and boys.
Two studies reported largely null results between PFHxS and LBW in the overall population
fManzano-Salgado etal.. 2017a: Hiermitslev etal.. 20201 as did the medium confidence study by
Meng etal. f20181 based on their quartile comparisons. On the basis of the continuous exposure
3-149
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
expressions, Mengetal. (2018) reported a larger risk (0R=1.5; 95% CI: 0.7, 2.9 per each ln-unit
increase) for a very LBW (i.e., <2,260 g) measure compared with the typical LBW definition of
<2,500 g (0R=1.3; 95% CI: 0.8, 2.1). Although term LBW results were null in girls in the Manzano-
Salgado etal. f2017al study, nonsignificant increases were seen amongst boys (OR=1.33; 95% CI:
0.47, 3.82 per each ln-unit increase).
Participant selection -
Exposure measurement-
Outcome ascertainment
Selective Reporting -
Overall confidence -
_i i i i i i i
&
Confounding -
Analysis -I
Sensitivity -
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 3-34. Study evaluation results for seven epidemiological studies of
small for gestational age and low birth weight and PFHxS. For additional details
see HAWC link.
3-150
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Population
Overall Study Study Sensitivity
Confidence
Design
Exposure
Window
Regression Exposure
Coefficient Comparison
Population Description
Regression coeffic
ent
Manzano Salgado
et al.. 2017.
4238465
INMA cohort (2003 2008) 1202
mother-infant pairs
|High| Adequate
Cohon
(Prospective)
Trimester 1
0.96 In-unit (ng.'mL)
Newborns (n=i 202)
LBW
I
# 3 [SGA/LBW Relative Risk (RR)]
• 3 [SGA/LBW Relative Risk (RR)] p-=0.05
Meng el al.. 2018.
DNBC (1996-2002). Denmark. 3535
IMediuml Good
Cohort
Trimester 1-2
1.1 Quartile 2
Newooms (n=37)
1 '• 1
|—195% confidence interval
(Prospective)
0.5 Quartile 3
Newborns (n=37)
• 1 '
0.8 Quartile 4
Newborns (n=37)
1.3 In-unit (ng.'mL)
Newborns (n=37)
*
Hjermilslev. 2020,
5880849
ACCEPT birth cohort (2010-2011,
2013-2015), Greenland. 4B2
|Madium| Adequate
Cohort
(Prospective)
Trimester 1-3
1.152 In-unit (ng.'mL)
Newborns (n=4B2)
1
1
Manzano-Salgado
et al.. 2017.
4238465
INMA oohort (2003-2008) 1202
mother-Infant pairs
|High| Adequate
(prospective)
Trimester 1
1.33 In-unit (ng.'mL)
Newborn boys (n-25)
LBW Boys j
'
0.83 In-unit (ng.'mL>
Newborn girls (n=33)
LBW Gin's r
1 • 1 1
0.97 In-unlt (ng.'mL)
Newborns (n=l202)
1-1—11
Wikstrfim, 2020,
6311677
SELMA (2007-2010). Sweden, 1533
mother-infant pairs
|High| Good
Cohort
(Prospective)
Trimester 1-2
1.37 Quartile 2
Newborns (n=1533)
1
0.89 Quartile 3
Newborns
Newborn girls (n=732)
i _
i
Figure 3-35. Small for gestational age and low birth weight results for seven
epidemiologic studies. For additional details see HAWC link.
SGA = small for gestational age; LBW = low birth weight.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bLow birth weight overall population data above black reference line.
cOverall population data above black dotted line; sex-stratified data below blue dotted line.
dFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g. (Xu et al., 2019)).
Small for gestational age/low birth weight summary
Although they were not always statistically significant, five different fXu etal.. 2019:
Wikstrom etal.. 2020: Meng etal.. 2018: Manzano-Salgado etal.. 2017a: Hamm etal.. 2010) of the
seven studies examining either SGA, LBW or very LBW showed some increased risks with
increasing PFHxS exposures among the overall population or either girls or boys. The associations
were quite variable (OR range: 1.3-9.1) in magnitude including some large but imprecise increased
odds, but there was no evidence of exposure-response relationships based on categorical data in
three separate studies. There were no patterns of results based on sample timing and other
characteristics.
3-151
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Fetal growth restriction summary
Among the most accurate fetal growth restriction endpoints examined, there was
reasonably consistent evidence for birth weight deficits across different measures and types of
PFHxS exposure metrics considered. Some mean or standardized birth weight deficits were
detected in 20 of the 34 included studies, including 14 out of 26 medium and high confidence
studies. Inverse associations were also noted in 17 of 31 studies that examined mean birth weight
associations in the overall population or either sex (6 high; 6 medium and 5 low confidence). This
included smaller birth weight deficits in the overall population for five high confidence studies
((3 = -12 to -22 g), the medium and low confidence studies reported reductions ranged from -20 to
-76 g per each ln-unit PFHxS increase. Eleven out of 14 sex-specific analyses, including 9 out of
12 medium and high confidence studies, showed some deficits in either or both male and female
neonates. Results were larger based on categorical comparisons in two low confidence studies
((3 range: -108 and -109 g for highest tertiles), but also consistent in magnitude among these sex-
specific studies expressing results per each ln-unit increase in both medium ((3 range: -45 to -76 g)
and high confidence studies ((3 range: -13 to -25 g).
The findings in the overall population were supported by the meta-analysis results of 27
studies presented above (and detailed in Appendix C) that showed a small deficit ((3 = -7.9 g; 95%
CI: -15.0, -0.7 per each ln-unit increase). This overall meta-analysis birth weight result ((3 = -7.9 g)
was comparable to analyses restricted to just the high ((3 = -6.8 g) and medium ((3 = -10.0 g)
confidence studies. The analysis restricted to only studies with some early pregnancy ((3 = -7.6 g)
biomarkers was also comparable in magnitude to these results. This early pregnancy data subset
would be less prone to any potential impact of bias related to pregnancy hemodynamics. As noted
above, many of the individual study results lacked precision and were not statistically significant,
especially the sex-stratified results. Two of the 16 studies examining categorical data for the overall
population or different sexes showed evidence of exposure-response relationships.
The evidence for birth length deficits was also consistent, with all four of the high
confidence studies showing deficits with increasing PFHxS exposures. However, among the high
confidence studies based on the overall populations, the birth length results were often imprecise
and small in magnitude ((3 = -0.14 to -0.43 cm). In contrast, the results for PFHxS studies of head
circumference and ponderal index were largely null. Across these different endpoints there is some
evidence of an association between fetal growth restriction and PFHxS exposure, but important
uncertainties remain. For example, there was a pattern suggestive of potential bias in studies with
biomarker samples collected after pregnancy (i.e., postpartum), given these studies showed larger
deficits in birth weight Some additional uncertainty also remains regarding whether any other
PFAS co-exposures are likely to be confounders in these studies; as such, this could potentially
affect study findings.
3-152
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Growth restriction - postnatal growth (infancy and early childhood up to 2 years of age)
Postnatal weight, height, and head circumference - background
Thirteen studies were identified that assessed postnatal growth in relation to PFHxS (see
Figure 3-36) with each examining some measures of infant weight and/or height Two
uninformative studies (Tin etal.. 2020a: Alkhalawi etal.. 2016) are not further considered here
mainly due to deficiencies or critical deficiencies in participant selection, confounding, analysis, and
study sensitivity. As shown in Figure 3-38 and Table 3-18, 5 of the 11 included studies were
considered high confidence fZhang et al.. 2022b: Starling etal.. 2019: Shoaff etal.. 2018: Manzano-
Salgado etal.. 2017b: Gao etal.. 20221. while three each were medium f Maisonet et al.. 2 012: Tensen
etal.. 2020a: Gvllenhammar et al.. 20181 and low confidence fLee etal.. 2018: Gross etal.. 2020: Cao
etal.. 2018). Of the 11 postnatal growth studies, study sensitivity in three were considered
adequate (Starling etal.. 2019: Manzano-Salgado etal.. 2017b: Gao etal.. 2022). while four each
were good (Shoaff etal.. 2018: Maisonet etal.. 2012: Lee etal.. 2018: Gvllenhammar et al.. 2018)
and deficient (Zhang et al.. 2022b: Tensen etal.. 2020a: Gross et al.. 2020: Cao etal.. 2018) largely
owing to small exposure contrasts.
Although there was some overlap across studies, limited serial measures during infancy as
well as inconsistent age at examinations and analyses may limit some comparisons here. For
example, Zhang etal. (2022b) examined growth up to 12 months and Starling et al. (2019) took
measurements at 5 months only. Manzano-Salgado etal. (2017b) examined growth from birth until
6 months of age. Lee etal. (2018) examined postnatal growth at 2 years, while the Cao etal. (2018)
analyses were based on a mean of 19 months in participants. Gvllenhammar et al. (2018) had serial
postnatal growth measures for most endpoints at 3, 6,12 and 18 months but was limited to
36 months and beyond for BMI SDS measures. Gross etal. f20201 completed examinations at
18 months, while Maisonet et al. f20121 did so at 20 months. Tensen etal. f2020al examined
different adiposity measures at 3 and 18 months, while Gao etal. (2022) examined growth
trajectory based on serial measurements at five time periods within the first 2 years (at birth,
42 days, 6 months, 12 months, and 24 months). Shoaff etal. (2018) examined postnatal growth
with repeated measures at age 4 weeks to 2 years.
3-153
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
sf>
Participant selection
Exposure measurement
Outcome ascertainment -
Confounding -
Analysis -
Sensitivity
Selective Reporting -
Overall confidence
p3 &
_L
++ ++ ++
&
Legend
I Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
J Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-36. Study evaluation results for 13 epidemiological studies of
postnatal growth and PFHxS. For additional details see HAWC link.
Postnatal weight-standardized results
In the overall population, eight postnatal studies (four high, two medium, and two low
confidence) examined PFHxS in relation to either standardized fZhang et al.. 2022b: Starling et al..
2019: Shoaffetal.. 2018: Manzano-Salgado etal.. 2017b: Gvllenhammar etal.. 20181 or mean
weight measures (Maisonet et al.. 2 012: Lee etal.. 2018: Cao etal.. 2018) (see Figure 3-37). Three of
five studies with standardized postnatal weight measures reported some inverse associations with
PFHxS exposures, while the medium confidence Gvllenhammar etal. T20181 study of standard
deviation scores (SDS) for weight measured at 3 to 18 months was null. Results in the high
confidence study by Zhang etal. f2022bl were largely null for standardized weight measures in the
overall population and both sexes, with the only association seen for increased weight among
tertile 2 exposures among girls examined up to 12 months ((3 = 0.15; 95% CI: 0.05, 0.25).
The results in the high confidence study by Starling etal. (2019) for the overall population
and both sexes were largely null for both weight-for-age and weight-for-length z-scores, although
they reported a statistically significant lower weight-for-age z-score at 5 months of age ((3 = -0.17;
95% CI: -0.33, -0.01 per each ln-unit increase) among girls. The authors did show an exposure-
3-154
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
response relationship for weight-for-age z-scores among girls across PFHxS tertiles (T2 (3 = -0.24;
95% CI: -0.54, 0.05; T3 (3 = -0.38; 95% CI: -0.69, -0.08), but the opposite was seen for boys (T2
(3 = 0.31; 95% CI: -0.01, 0.62; T3 (3 = 0.26; 95% CI: -0.09, 0.61). Results were smaller in magnitude
but fairly comparable for weight-for-length z-scores albeit in a nonmonotonic fashion for girls ((3
range: -0.20 to -0.23).
Compared with tertile 1, the high confidence study by Shoaffetal. (2018) detected small
nonstatistically significant deficits in z-scores for several outcomes including weight-for-age and
weight-for-length for PFHxS tertile 3 ((3 range: -0.15 to -0.16). They also reported nonsignificant
results per each ln-unit increase for both weight-for-age ((3 = -0.12; 95% CI: -0.29. 0.06) and
weight-for-length ((3 = -0.12; 95% CI: -0.26. 0.01) z-scores. Although they were also not statistically
significant, small weight z-score changes from birth to 6 months of age were also reported in the
Infancia y Medio Ambiente (INMA) birth cohort ((3 = -0.09; 95% CI: -0.22, 0.03 per each ln-unit
increase) from the other high confidence Manzano-Salgado etal. (2017b) study. These results
seemed largely driven by the findings in girls ((3 = -0.13; 95% CI: -0.29, 0.03).
3-155
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Population
Overall Study
Confidence
Study Sensitivity
Design
Exposure
Window
Regression
Coefficient
Exposure
Comparison
# g [change in PNG z-soore]
O P [change in PNG z-score] p<0.05
I—195% confidence interval
Zhang et al..
2022, 9944433
Shanghai Birth Cohort (SBC)
(2013-2016), China, 2395
mother-infant pairs
|High|
Deficient
Cohort
(Prospective)
Trimester 2
-0.01
In-iinil(ngjmL)
increase
Weight-for-Age Z-Score
-0.05
Tortile 3
1
1 •—1—•
OVERALL POPULATION
Starling al al.
Healthy Start Study (2009-2014),
|Hlgh!
Adequate
Cohort
Trimester 2-3
-0.04
1
2019. 5412449
United States, 1410 mother-infant
pairs
1
-0.05
Tertile 3
1
HOME (2003-2006), United States.
345 mother-Infant pairs
IHighl
Good
Trimester 2 3 at
2018.4619944
(Prospective)
delivery
1
-0.12
_ 1
2022^9944433
Starling etal.
Shanghai Birth Cohort (SBC)
(2013-2016), China, 2395
mother-infant pairs
Healthy Start Study (2009-2014),
IHIgh!
IHighi
Daficienl
Adequate
(Prospective)
Cohort
Trimester 2-3
-0.01
0.06
0.013
0.08
ln-unil(ng/mt_)
Tertile 2
Tortile 3
In-unit (ng/mL)
1
1
1
1
1 • '
i
~ i
l
BOYS
2019.5412449
United States, 1410 mother-infant
increase
i
i
Zhang et al..
2022, 9944433
Shanghai Birth Cohort (SBC)
(2013-2018). China, 2395
mother-infant pairs
IHighl
Deficient
(Prospective)
Trimester2
0.08
0.149
0.05
In-unit(ngj'mL)
Tertile 2
Tertile 3
H • 1
1
1 1 • 1
GIRLS
Stalling al al.
Healthy Start Study (2009-2014),
|Hlgh|
Adequate
Cohort
Trlmestei 2-3
-0.17
In-unit (ng/mL)
2019. 5412449
United States. 1410 mother-infant
pairs
1 .
-0.38
Tertile 3
1 '
*
-0.01
In-unit (ng/mL)
Tertile 2
1 • 1
Weight-for-Length Z-Score
Tertile 3
Shoaff etal.,
HOME (2003-2006), United States.
IHighl
Good
Cohort
Trimester 2-3, at
-0.12
1
2016,4619944
345 mother-Infant pairs
(Prospective)
delivery
increase
*
1
Starling at al.
Healthy Start Study (2009-2014).
|Hlgh|
Adequate
Cohort
Trimester 2-3
0.06
In-unit (ng/mL)
1 m
BOYS
2019.5412449
United States. 1410 mother-infant
1
1
-0.08
In-unit (ng/mL)
1 _
- 1
GIRLS
increase
-0 23
-0.2
Tertile 3
1
Gyllsnhamrnar et
al., 2018, 4238300
POPUP (1996-2011)361
|Medium|
Good
Cross-sectional
3 weeks post-brrlh
0.07
In-unil (ng.'ml_)
1 1 ^ |
Weight Z-Score
0.07
In-unit (ng/mL)
0.07
In-unit (ng/mL)
1
0.06
In-unit (ng.'mL)
1
1 , • 1
Ma rizano-Salgado.
INMA Birth Cohort (2003-2008),
| Medium |
Deficient
Cohort
Trimester 1-3
-0.09
2017b. 4238509
Spain, 1243 mother-infant pairs
(Prospective)
-0.04
In-unit (ng/mL)
Weight Gain OVERALL POPULATION
_ 1
-0.13
In-unit (ng/mL)
1
• ' 1
BJ)Y5
GIRLS
0.7 -0.6 -0
5 -0.4 -0.3 -0
2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0
6 0.7
Figure 3-37. Standardized postnatal weight results for PFHxS epidemiological
studies. For additional details see HAVVC link.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bAge at Outcome Measurement: (Gyllenhammar et al., 2018) at 3 months, 6 months, 12 months, and 18 months
(ordered top to bottom); (Starling et al., 2019) at 5 months; (Zhang et al.. 2022b) between 42 days and
12 months; (Shoaff et al., 2018) at 4 weeks, 1 year, and 2 years; (Manzano-Salgado et al., 2017b) at 6 months.
cSolid black lines divide the figure into four categories. Listed from top to bottom they are as follows: Weight-for-
Age Z-Score, Weight-for-Length Z-Score, Weight Z-Score, and Weight Gain Z-Score
dWithin each category, overall population is located above the first blue dashed lines, boys are between the two
blue dashed lines, and girls are below the second blue dashed line.
eFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g. (Gyllenhammar et al., 2018)).
3-156
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Postnatal weight_mean - results
Three studies examined associations between PFHxS exposures and mean postnatal weight
measures fMaisonet etal.. 2012: Lee etal.. 2018: Cao etal.. 20181 (see Figures 3-38). The low
confidence study by Lee etal. f 20181 detected associations infant weight at age 2 ((3 = -200 g; 95%
CI: -420, 20) per each ln-unit increase and monotonically across PFHxS quartiles ((3 range: -160 to
-360 g). For example, a large difference was detected for quartile 4 (>1.81 ng/mL) ((3 = -360 g;
95% CI: -740, 20) compared with quartile 1 (<0.77 ng/mL). They detected weight change
associations from birth to age 2 per each ln-unit increase ((3 = -170 g; 95% CI: -330,160) but was
considerably smaller among quartile 4 exposures ((3 = -60 g; 95% CI: -400, 270). The Cao et al.
f20181 study was null for all comparisons, but they did report an imprecise postnatal
(mean = 19 months) weight difference for tertile 2 ((3 = -145 g; 95% CI: -584, 294) in the overall
population. Tertile 2 results were imprecise and in opposite directions for boys ((3 = -387 g; 95%
CI: -916,143) and girls ((3 = 155 g; 95% CI: -605, 915), while there was some suggestion of reduced
weight in tertile 3 among girls ((3 = -101 g; 95% CI: -811, 608). The medium confidence study of
girls from the ALSPAC study (Maisonet etal.. 2012) were largely null and inconsistent across tertile
((3 range: -32 to 63 g) over the first 20 months of life.
Study Population Overall Study Study Sensitivity Design Exposure Regression Exposure
Confidence Window Coefficient Comparison Regression coefficient
Maisonet etaL ALSPAC (1991-1992), U.K.. 447 |Medium| Good Cohort Trimester 1-3 -31.84 Tertile 2
2012.1332465 mother-girl pairs (Prospective)
•I
62.86 Terlile 3
1 1 • —1
Lee etal., 2018, Environment and Development of |Low| Good Cohort 0-2 years -360 Quartile 4
4238394 Children (EDC) Cohort, South Korea,
645 mother-child pairs
1 • I'
1
# p [change in mean growth weight
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Postnatal weight summary
Five of eight studies in total showed some evidence of associations in the overall population
or other sex for either mean or standardized infant weight measures. This included one high
confidence study fShoaffetal.. 20181 showing associations for both weight-for-age and weight-for-
length measures in the overall population and both low confidence studies. There was a
preponderance of inverse associations between PFHxS and infant weight among girls only (based
on three of four, including two of three weight-standardized studies and one mean weight study).
No patterns across the few studies with associations were evident.
Postnatal height standardized results
In the overall population, five postnatal studies (two high, one medium, and two low
confidence) examined PFHxS in relation to either standardized (Zhang et al.. 2022b: Shoaff etal..
2018: Gvllenhammar etal.. 2018) or mean height measures (Lee etal.. 2018: Cao etal.. 2018) (see
Figures 3-39). Five studies in total examined postnatal height measures in relation to PFHxS
including three that examined standardized postnatal height fZhang et al.. 2022b: Shoaff etal..
2018: Gvllenhammar etal.. 20181. None of these studies showed any evidence of an association
between PFHxS in relation to standardized infant height measures. The medium confidence by
Gvllenhammar et al. f20181 was null for standardized height measures in the overall population.
The high confidence study by Zhang etal. (2022b) were null for standardized height measures in
the overall population and both sexes. The high confidence study by Shoaff etal. (2018) was largely
null for length-for-age z-score for continuous ((3 = -0.07; 95% CI: -0.27, 0.14) for each ln-unit
increase and categorical PFHxS exposures (Tertile 3 (3 = -0.13; 95% CI: -0.52, 0.27).
3-158
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study Population Overall Study Study Sensitivity Design Exposure Regression Exposure
Confidence Window Coefficient Comparison
Regression coefficient
Zhang el al.. Shanghai Birth Cohort (SBC) |Hlgh| Deficient Cohort Trimester 2 -0.009 Tertlle 2
2022.9944433 (2013-2016), China. 2395 (Prospective)
mother-intent pairs -0.02 Tertlle 3
-0.06 In-unlt(ng'mL)
increase
OVERALL POPULA
0 3 [change in height Z-Soore] p<0 05
Shoaff etal.. HOME (2003-2006). United States. |High| Good Cohort Trimester 2-3. at 0.04 Tertile2
2018.4619944 345 mother-infant pairs (Prospective) delivery
-0.13 Tertile 3
-0.07 ln-unit (ngAnL)
Gyllenhammar et POPUP (1996-2011) 381 |Medium| Good Cross-sectional 3 weeks post-blrth -0.01 ln-unit (ng/mL)
0 ln-unit (ng/mL)
increase
0.01 ln-unit (ng/mL)
0.01 ln-unit (ng/mL)
increase
- •
• 1
Zhang etal.. Shanghai Birth Cohort (SBC) |High| Deficient Cohort Trimester 2 -0.006 Tertlle 2
2022,9944433 (2013-2016), China. 2395 (Prospective)
mother-Intent pairs .0.022 Tertile 3
-0.06 In-unit(ng'mL)
BOYS
1 •
0.092 Tertile 2
-0.023 Tertile 3
0 ln-unlt(ng/mL)
GIRLS
1 1
1
Figure 3-39. Standardized postnatal height results for PFHxS epidemiological
studies. For additional details see HAWC link.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bAge at Outcome Measurement: Gyllenhammar et al. (2018) at 3 months, 6 months, 12 months, and 18 months
(ordered top to bottom); Zhang et al. (2022b) between 42 days and 12 months; Shoaff et al. (2018) between
4 weeks and 2 years.
cZhang et al. (2022b) and Shoaff et al. (2018) examined length-for-age z-score.
d Data for overall population is above the solid black line, while sex-stratified data is below. Within sex-stratified
data, boys are above the black dashed line, girls are below.
eFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g. Gyllenhammar et al. (2018)).
Postnatal height mean results
Two studies fLee etal.. 2018: Cao etal.. 20181 examined associations between PFHxS
exposures and mean postnatal height measures (see Figures 3-40). The low confidence study by Lee
etal. (2018) reported statistically significant decreased mean height ((3 = -0.84 cm; 95% CI: -1.26,
-0.42 per each ln-unit increase) at age 2 as well as reductions in height ((3 = -0.89 cm; 95% CI:
-1.45, -0.33 per each ln-unit increase) from birth to age 2. They also detected exposure-response
relationships and statistically significant infant height reductions in quartiles 3 and 4 for both
weight at 2 years (Q4 (3 = -1.34 cm; 95% CI: -2.09, -0.60; Q3 (3 = -0.82 cm; 95% CI: -1.57, -0.07)
and weight change from birth to 2 year (Q4 (3 = -1.63 cm; 95% CI: -2.62, -0.64; Q3 (3 = -1.20 cm;
95% CI: -2.10, -0.30). The low confidence study by Cao et al. f20181 reported nonmonotonic
increased postnatal length in the overall population ((3 range: 0.95 to 1.42 cm across
tertiles). Similar results were seen for girls ((3 range: 1.32 to 2.01 cm across tertiles) but were null
for boys ((3 range: 0.30 to 0.32 cm across tertiles).
3-159
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Overall Study Study Sensitivity Design Exposure Regression Exposure
Confidence Window Coefficient Comparison
Regression coefficient
Cao el al.. 2018, Zhoukou City Longitudinal Birth
5080197 Cohort (2013-2015). China. 282
mother-infant pairs
OVERALL POPULATION
# 0 [change in height]
O 0 [change in height] p<0.05
I—195% confidence interval
Lee et al.. 2018, Environment and Development of |Low
4238394 Children (EDC) Cohort, South Korea,
645 mother-child pairs
Ouartile 3
Quarlile 4
In-unil (ngftnL)
Quartile 2
Ouartile 3
Quartile 4
Iri-unit (ng/mL)
> et al„ 2018, Zhoukou City Longitudinal Birth
5080197 Cohort (2013-2015), China, 282
mother-infant pairs
Tertite 2
"fertile 3
Figure 3-40. Mean postnatal height results for PFHxS epidemiological studies.
For additional details see HAWC link.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bAbove the solid black line is overall population data, while below it is sex-stratified. Within the sex-stratified data,
above the dashed blue line is boys, below is girls.
Tor Lee et al. (2018) data, above the black dashed line is data referring to at 2 years, below the line is data
referring to change from birth to 2 years.
3-160
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Rapid weight gain
Four high confidence studies f Starling etal.. 2019: Shoaff etal.. 20181: Manzano-Salgado et
al. f2017bl: fGao etal.. 20221 examined different rapid weight gain measures in relation to PFHxS
(see Figures 3-41 and 3-42). In the Health Outcomes and Measures of the Environment (HOME)
study, Shoaff etal. (2018) examined rapid growth based on weight z-scores in relation to PFHxS in
the overall population. In the Healthy Start study Starling et al. (2019) examined different rapid
weight gain measures in relation to PFHxS for the overall population and both sexes. In the
Shanghai Birth Cohort, Gao etal. (2022) examined various measures of growth trajectories in the
overall population and across sex for various postnatal growth measures. In the INMA Birth Cohort
Study, Manzano-Salgado etal. f2017bl examined rapid growth from birth to 6 months.
Two of the four studies showed some increased odds of rapid growth measures with
increasing PFHxS exposures, although results were not always internally consistent. Shoaff et al.
(2018) reported null associations for odds of weight z-score differences across tertiles (e.g., tertile
3 OR=0.95, 95% CI: 0.65,1.40). The study by Manzano-Salgado etal. (2017b) was also null for rapid
growth (OR=0.87; 95% CI: 0.72,1.04). The study by Starling et al. (2019) reported an OR of 1.49
(95% CI: 1.02, 2.18) for rapid weight gain per each ln-unit increase based on the weight-for-age z-
score data but was null for weight-for-length z-score (OR=0.95; 95% CI: 0.63,1.44).
In the Gao etal. f20221 study, most relative risks were null based on standardized weight-
for-age and weight-for-length measures in the overall population and both sexes. Compared with
the moderate-stable referent, Gao etal. (2022) reported elevated odds for the low-rising weight-for-
age z-score [WAZ] trajectory (OR=1.92; 95% CI: 1.19, 3.08 per each ln-unit PFHxS increase) in the
overall population. This seemed driven by results in males (OR=2.96; 95% CI: 1.51, 5.82 per each
ln-unit PFHxS increase) given that females showed null associations. Using a weighted quantile sum
mixture approach, they reported a statistically significant inverse association (OR=1.53; 95% CI:
1.13, 2.06 per each ln-unit PFAS Sum increase) for WAZ among low-rising participants (versus
moderate-stable) with PFHxS having the highest weight among the PFAS mixture constituents.
Among males only, Gao etal. (2022) reported increased odds for weight-for-length z-score
(WLZ) trajectory in low-rising (OR=2.43; 95% CI: 1.00, 5.87 per each ln-unit PFHxS increase) and
low-stable participants (OR=2.04; 95% CI: 0.70, 6.02 per each ln-unit PFHxS increase). Compared
with the moderate-stable referent, Gao etal. f20221 reported elevated odds in females only for the
moderate-falling (OR=1.85; 95% CI: 0.97, 3.47 per each ln-unit PFHxS increase) and high-rising
length-for-age z-score (LAZ) trajectories (OR=1.61; 95% CI: 0.41, 6.38 per each ln-unit PFHxS
increase). The odds of LAZ for high-rising participants from the overall population was null in the
single pollutant model but was elevated for the PFAS mixture metric based on a weighted quantile
sum approach (OR=1.59; 95% CI: 0.90, 2.82 per each ln-unit PFHxS increase), with PFDA having the
highest weight among the PFAS mixtures.
Although most were not statistically significant, Gao etal. f20221 reported inverse
associations in the single-PFAS models for head-circumference-for-age z-score for high-rising,
3-161
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
moderate-rising, low-rising, and low-stable versus moderate-stable participants (OR range: 0.46 to
0.71 per each ln-unit PFHxS increase). They also reported a statistically significant inverse
association (OR=0.37; 95% CI: 0.18, 0.72) for low-rising versus moderate-stable groups based on a
PFAS mixture metric (per each ln-unit increase) using a weighted quantile sum approach.
Rapid weight gain summary
Overall, two of four studies showed increased odds of rapid growth in relation to PFHxS
exposures. Although results were a bit mixed across different growth trajectory measures, there
was only evidence of inverse associations between PFHxS and rapid growth as measured by head
circumference z-scores in the Gao etal. (2022) study. In contrast, most of the associations they
detected using weight-for-age, weight-for-length and length-for-age z-scores showed increased risk
of rapid growth per each ln-unit PFHxS increase. These associations were most evident among the
weight and height measures among the participants with a low baseline growth trajectory followed
by a rapid increased trend afterward (i.e., low-rising group). These data were supported by another
study (Starling etal.. 2019) that reported a statistically significant OR (1.49; 95% CI: 1.02, 2.18 per
each ln-unit increase) for rapid weight gain based on weight-for-age z-score data only. Both of these
studies are consistent with a hypothesis that rapid weight growth in childhood may have followed
intrauterine growth retardation from PFHxS exposures. These individuals may be at most risk for
metabolic syndrome, as evidenced by changes in obesity and other health effects later in life.
3-162
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Population Overa.l S,u,y Study Sensitivity Design Ensure Regression ^ur^ Population Description R , |
S*2449 Ad0CJBtB C°h0rt 095 vs. Non-Rapid Ctb** (n=352>
5S5A* ssfsgsassr-*- Hgh GMd Trl^m u6 Tfini,e2 >
£££»., H* Adnata (Rf^ Tnm#Mer3 0.B5 .h-JjJI,
,07 'n'1naiTL'
Manzan^SB^do, m>,1A Birh^^ (2003-2008), Medium Oe'iclenl ^ Cohort ^ Trimester 1-3 0.07 In-unit (ng.'mL) Rapid Growth
H'3h Ad6CJ9te (Pr££U T™3 °'7J 'n"™mL' Moderated* (n-20..
"8 ""inc^e""' Low-ste,"e vs
0.76 L<*v-Rle.ing (n=74>vsMo<*>ta(o-StoDlo (11=483)
'¦32 Low-Stable (n=149) vs Moderate-Stable (n=403)
"S" HSM>"n"n »'i
;i ZZZZZZZ
085 HiSh-RisinQ (n=M) vs Moderate-Stable (n=2S2)
...
T
¦
Figure 3-41. Postnatal rapid growth (weight-for-age and weight-for-length z-
score) results for PFHxS epidemiological studies. For additional details see
HAWC link.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bAge at Outcome Measurement: Starling et al. (2019) at 5 months, Gao et al. (2022) modeled data (collected at
42 days, 6 months, 12 months, and 24 months).
cWeight-for-Age Z-Score data above the black reference line; weight-for-length below.
dOverall population data above the blue line; Sex-stratified data below.
eSex-Stratified data: male infants above the blue dash-dotted line; females below.
fQuantile 2 in Starling et al. (2019) represents dichotomized exposure at median (quantile 1 referent: LOD-
0.1 ng/mL; quantile 2: 0.2-3.5 ng/mL).
gThe following Gao et al. (2022) results have been truncated: 1.92 [1.19-3.08], 2.96 [1.51-5.82], 2.43 [1-5.87], and
2.04 [0.7-6.02],
3-163
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Overall Study Study Sensitivity Design Exposure Regression Exposure
Confidence Window Coefficient Comparison
(ngtaiL)
n-unit (ng'mL)
Population Description
High-Rising vs. Moderate Stable
Moderate-Falling vs Moderate Stable
Regression coefficient
n-unit (ng'mL) High-Rising vs. Moderate Stable
n-unit (ng'mLI Moderate-Falling vs. Moderate Stable
n-unit (ng'mL) High-Rising vs. Moderate Stable
n-unlt (ng'mL) Moderate-Falling vs. Moderate Stable
n-unlt (ng'mL) Newborns (Moderate-Rising vs Moderate Stable)
# [Odds Ratio ro* Rapid Growth]
0 [Odds Ratio tor Rapid Growth] p<0.05
H 95% confidence interval
T
" ?4 ™
Figure 3-42. Postnatal rapid growth (length-for-age and head circumference z-
score) results for PFHxS epidemiological studies. For additional details see
HAWC link.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bAge at Outcome Measurement: Gao et al. (2022) modeled data (collected at 42 days, 6 months, 12 months, and
24 months).
cLength-for-Age Z-Score data above the black reference line; Head Circumference Z-Score below.
dSex stratified Length-for-Age Z-Score data below blue solid line; males above blue dotted line; females below.
eOverall population data above the blue line; Sex-stratified data below.
'Female confidence intervals have been truncated; the data points are 1.61 [0.41-6.38] and 1.85 [0.97-3.47],
gFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses.
3-164
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Postnatal head circumference
Three studies examined postnatal head circumference in relation to PFHxS f Zhang etal..
2022b: Gvllenhammar et al.. 2018: Cao etal.. 20181 (see Figure 3-43). Null results were detected in
the high confidence study by Zhang etal. f2022bl for head circumference-for-age z-score per each
ln-unit PFHxS increase ((3 = -0.08; 95% CI: -0.19, 0.02). The medium confidence study by
Gvllenhammar et al. (2018) showed monotonic head circumference-for-age Z increases as children
aged from 3 to 18 months ((3 range: 0.05 to 0.12). The low confidence study by Cao etal. (2018)
reported nonmonotonic increased postnatal head circumference in the overall population ((3 range:
0.90 to 1.33 cm across tertiles). These results were comparable across boys ((3 range: 0.97 to
1.27 cm across tertiles) and girls ((3 range: 0.78 to 1.34 cm across tertiles). Overall, two of three
studies showed some evidence of increased postnatal head circumference in relation to PFHxS
exposures.
Study
Population
Overall Study
Confidence
Study Sensitivity
Design
Exposure
Window
Regression
Coefficient
Exposure
Comparison
Regression coefficient
£ 3 [change in PNG HC]
Zhang «l al..
2022. 9944433
Shanghai Birth Cohort (SBC)
(2013-2016), China, 2395
mother-infant pairs
|High|
Deficient
Cohort
(Prospective)
Trimester 2
-0.D26
-0.09
Tertile 2
ln-unit(ng.'mL)
i—••—i
H 95% confidence interval
1 • r*
Cao etal., 2018.
Zhoukou City Longitudinal Birth
Cohort (2013-2015). China. 282
mother-infant pairs
|Low|
Deficient
(Prospective)
At birth
1.33
Tertile 2
1
Tertile 3
Gyllonhammar et
al.. 2018, 4238300
POPUP (1996-2011) 381
mother-infant pairs
Medium]
Good
Cross-scctional
weeks post-birt
0.05
0.07
0.09
0.12
ln-unit (ng/mL)
ln-unit (ng/mL)
increase
ln-unit (ng/mL)
ln-unit (ng/mL)
Hj-« 1
1 • 1
OVERALL POPULATION
Zhang el al„
2022,9944433
Shanghai Birth Cohort (SBC)
(2013-2016). China, 2395
mother-infant pairs
|High|
Deficient
Cohort
(Prospective)
Trimester 2
0,045
-0.08
Tertile 3
ln-unit(ng/mL)
1 1 • 1
1
BOYS
Cao etal., 2018.
5080197
Zhoukou City Longitudinal Birth
Cohort (2013-2015). China, 282
|Low|
Deficient
(Prospective)
At birth
1.27
Tertile 2
'
1
0.97
Tertile 3
1
' w
Zhang et al..
_2Q22„2944433 _
Shanghai Birth Cohort (SBC)
(2013-20.16J_ C&ina_23a5
mother-infant pairs
|High|
Deficient
_(Prospective)
Trimester 2
0.093
0.044
0.06
Tertile 2
Tartile 3
In-unit(ngj'mL)
lf—0 1
>-!"• 1
GIRLS
Cao etal., 2018.
5080197
Zhoukou City Longitudinal Birth
Cohort (2013-2015). China, 282
|Low|
Deficient
Cohort
(Prospective)
At birth
1.34
Tertile 2
1
0.78
Tertile 3
0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4
Figure 3-43. Postnatal head circumference results for PFHxS epidemiological
studies. For additional details see HAWC link.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bAge at Outcome Measurement: Gvllenhammar et al. (2018) at 3 months, 6 months, 12 months, and 18 months
(ordered top to bottom); (Zhang et al., 2022b) between 42 days and 12 months; Cao et al. (2018) at a mean of
19 months.
cZhang et al. (2022b) reports head circumference-for-age Z-Score, Gvllenhammar et al. (2018) report head
circumference Z-Score, and Cao reported odds ratios.
dOverall population is above the solid black line, while sex-stratified data is below. Within sex-stratified data, boys
are above the dashed blue line, girls below.
eCao et al. (2018) upper and lower bounds have been truncated. For overall population, the Tertile 2 bounds are
[0.42, 2.26] and the Tertile 3 bounds are [0,1.81], For boys, the Tertile 2 bounds are [0.1, 2.43] and the Tertile 3
bounds are [-0.22, 2.16], For girls, the Tertile 2 bounds are [-0.16, 2.84] and the Tertile 3 bounds are [-0.62,
2.18],
fFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g. Gvllenhammar et al. (2018)).
3-165
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Postnatal adiposity/body mass index/ponderal index/weight status
Five studies fZhang etal.. 2022b: Starling etal.. 2019: Shoaffetal.. 2018: Tensen etal..
2020a: Gross etal.. 20201 enabled examination of different measures of infant adiposity such as
body mass index (BMI), overweight status, and ponderal index (see Figure 3-44). Three of the five
studies were null (Zhang etal.. 2022b: Starling etal.. 2019: Tensen etal.. 2020a) for associations in
the overall population, while the remaining two showed decreased measures of adiposity in
relation to PFHxS. For example, the low confidence study by Gross etal. f20201 showed an inverse
but nonsignificant association between overweight status at 18 months (OR = 0.75 g; 95% CI: 0.30
to 1.85) and dried blood spot PFHxS levels above the mean (compared with below the mean) with
similar relative risks among boys (OR=0.74; 95% CI: 0.17, 3.24) and girls (OR=0.68; 95% CI: 0.15.
3.12). The high confidence study by Shoaff et al. f20181 exposure-response relationship detected
for PFHxS and BMI z-score across tertile 2: ((3 = -0.12; 95% CI: -0.37, 0.13) andtertile 3 ((3 = -0.22;
95% CI: -0.47, 0.03) and per each ln-unit increase ((3 = -0.12; 95% CI: -0.26, 0.01).
The results were a bit more mixed when examined by sex, with two of three sex-specific
studies showing some suggestion of increased adiposity among boys only. For example, the medium
confidence by Tensen etal. f2020al reported null associations at age 3 and 18 months for
standardized (i.e., SDS) postnatal waist circumference, body mass index, and ponderal index
measures in their overall population. Although they did not detect statistically significant
interactions by sex for any endpoints evaluated, slight nonsignificant increases in boys BMI
((3 = 0.13; 95% CI: -0.34, 0.60 per each ln-unit increase) and Ponderal Index ((3 = 0.34; 95% CI:
-0.14, 0.82 per each ln-unit increase) SDS scores were noted. The high confidence study by Starling
etal. (2019) was null for infant adiposity per each ln-unit PFHxS increase among the overall
population ((3 = 0.01% change in fat mass; 95% CI: -0.67, 0.68). Results were divergent for males
((3 = 0.54% change in fat mass; 95% CI: -0.51,1.58 per each ln-unit increase) versus females
((3 = -0.42 % change in fat mass; 95% CI: -1.31, 0.47 per each ln-unit increase). Similar results were
seen in their tertile analyses with more adiposity in males ((3 range: 0.89 to 1.90% change in fat
mass) and females ((3 = -0.85 to -1.11% change in fat mass). The high confidence study by Zhang et
al. (2022b) reported null associations for PFHxS and BMI-for-age z-scores ((3 = -0.01; 95% CI -0.12,
0.09 per each ln-unit increase) in the overall population, males ((3 = -0.01; 95% CI -0.12, 0.09 per
each ln-unit increase) and females ((3 = 0.10; 95% CI: -0.01, 0.20 per each ln-unit increase).
Postnatal adiposity summary
Overall, none of the five studies in the overall population reported increased adiposity with
increasing PFHxS exposures up to age 2 years. However, two of three studies in boys did show
some suggestion of increased adiposity in relation to PFHxS exposures. None of the three studies in
girls reported increased adiposity.
3-166
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study Population Overall Study Study Sensitivity Design Exposure Regression Exposure
Confidence Window Coefficient Comparison
Regression coefficient
# p [change in adiposity measures]
© (3 [change in adiposity measures] p<0.05
I—195% confidence interval
Jensen etal.. OCC (2010-2012). Denmark, 613 |Medium| Deficient Cohort Trimester 1-2 0.04 In-unit (ng/mL)
2020,6833719 mother-infant pairs (Prospective) increase
0.13 In-unit (ng/mL)
BMI (b
1
0.03 In-unit (ng/mL)
Shoaffetal., HOME (2003-2006). United States. |High| Good Cohort Trimester 2-3. at -0.12 In-unit (ng/mL)
2018,4619944 345 mother-infant pairs (Prospective) delivery increase
-0.12 Tertile 2
-0.22 Tortile 3
1 •
1 •
1.
1
1 |
1 '
BMI Z-Score (overall pop)
Zhang etal., Shanghai Birth Cohort (SBC) |High| Deficient Cohort Trimester2 -0.01 In-unit(ng.'mL)
2022,9944433 (2013-2016), China, 2395 (Prospective) increase
mother-infant pairs
0.02 Tertile 2
-0.09 Tertile 3
°i '
^ 1 _
BMI
for-Age Z-Score (overall pop)
-0.01 ln-unit(ng/mL)
0.02 Tertile 2
-0.09 Tertile 3
1 '
1
BMI-for-Age Z-Score (boys)
0.1 In-unit(ng.'mL)
0.1 Tertile 2
0.07 Tertile 3
ti • 1
1
1 • "
hJ- •
BMI-for-Age
Z-Score (girls)
Starling et al. Healthy Start Study (2009-2014), |High| Adequate Cohort Trimester 2-3 0.01 In-unit (ng/mL)
2019,5412449 United Slates, 1410 mother-infant increase
Pa'rS 0.6 Tertile 2
-0.02 Tertile 3
ty (overall pop)
1 Adlposi
0.54 In-unit (ng/mL)
1.9 Tertile 2
0.89 Tertile 3
1 Adiposity (boys)
1
-0.42 In-unit (ng/mL)
-1.11 Tertile 2
-0.85 Terlile 3
diposity (girls)
1
1
Jensen etal,, OCC (2010-2012), Denmark, 613 |Medium| Deficient Cohort Trimester 1-2 0.08 In-unit (ng/mL)
2020,6833719 mother-infant pairs (Prospective) increase
Ponderal Index (overall pop)| 1 ^
0.34 In-unit (ng/mL)
increase
Ponderal Index (bovsl . A
1 w
0.02 In-unit (ng/mL)
Ponderal Index (girls) \—
i*
Gross etal., 2020, Starting Early Program (StEP) |Low| Deficient Nested At birth 0.75 Weight Status
7014743 Cohort, United States, 98 case-control (>mean)
Weight Status (overall pop)
1 O
1
0.74 Weight Status
(>mean)
Weight Status (boys)
1
0.68 Weight Status
(>mean)
Weight Status (girls)
1 i e
0.5 -C
.4 -0.3 -0.2 -(
1 0 o'l 02 0
3 0
4
5 0.6 0
7 0.8 0.9
Figure 3-44. Postnatal body mass index, adiposity, and ponderal index and
weight status results for PFHxS epidemiological studies. For additional details
see HAWC link.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bMeasurement types are separated by the solid black reference lines and are as follows (in descending order): BMI,
BMI Z-Score, BMI-for-Age Z-Score, Adiposity, Ponderal Index, and Weight Status.
cWithin each category, above the first dotted blue line are values for overall population, between the two dotted
lines are values for boys, and below the second dotted line are values for girls.
Postnatal growth summary
Overall, there were mixed results within and across the 13 available postnatal PFHxS
studies of postnatal growth with the most consistent evidence for postnatal weight. Five of eight
studies in total showed some evidence of associations with mean or standardized infant weight
measures including three high confidence studies in the overall population and three of four studies
in girls. No other patterns were evident Only one low confidence study out of five total studies
showed any evidence of smaller height based on either with mean or standardized height measure
in the overall population or either sex. None of three available studies showed some evidence of
decreased postnatal head circumference in relation to PFHxS exposures. In contrast, two of them
showed increased postnatal head circumference. Similarly, none of the five studies in the overall
3-167
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
population reported increased adiposity in relation to PFHxS as two studies showed decreased
measures of adiposity. The results for rapid growth measures were a bit mixed but two of four
studies showed increased odds of rapid growth in relation to PFHxS.
Although few studies examined exposure-response relationships based on categorical data
in the overall population or across sexes, three different studies did show dose-dependence for
some measures such as infant weight (one of six studies), height (one of four studies) and adiposity
(one of three studies). No study characteristics were obvious explanatory factors for between-study
heterogeneity. Few patterns by sex were evident outside a preponderance of inverse associations
between PFHxS and infant weight among girls. There was also evidence in two of three studies in
boys of increased adiposity. However, limited exposure contrasts and statistical power may have
hampered the ability to detect associations small in magnitude especially among the sexes. In
summary, the evidence was mixed across various postnatal measures and different examination
windows, with only minimal evidence of exposure-response relationships to support the
continuous exposure-scaled results. One challenge in evaluating consistency across heterogeneous
studies includes disparate periods of follow-up and assessment (e.g., childhood age at examination).
Table 3-18. Summary of 11 epidemiologic studies of PFHxS exposure and post-
natal growth measured
Author
Study
location,
years
Sample
size
Median
exposu re
(range)
in ng/mL
Weight
Height
HC
Adiposity
Rapid
growth
High confidence studies
Gao et al. (2022)
China,
2013-2016
1,350
0.54
(0.21,
3.75)
Overall
Manzano-Salgado et
al. (2017b)
Spain,
2003-2008
1,154
0.58
(0.05,
11.01)
- Overall/
Girls
0 Boys
0
Overall
Shoaff et al. (2018)
OH, USA,
2003-2006
345
1.5
(0.1,
32.5)
- Overall
- Overall
- Overall3
0
Overall
Starling et al. (2019)
CO, USA,
2009-2014
415
0.7 (0.2,
2.8)b
Overall/Girls3
+ Boysa
0 Overall
+ Boys
- Girls
Overall
Zhang et al. (2022b)
China,
2013-2016
2,395
0.53
(0, 25.4)
0
Overall/Boys
+ Girls
0 Overall/
Boys/Girls
Medium confidence studies
3-168
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Author
Study
location,
years
Sample
size
Median
exposu re
(range)
in ng/mL
Weight
Height
HC
Adiposity
Rapid
growth
Gvllenhammar et al.
(2018)
Sweden,
1996-2001
381
2.4
(0.32,
26.0)
0 Overall
Maisonet et al.
(2012)
United
Kingdom,
1991-992
422
1.6
(0.2,
54.8)
- Girls
Low confidence studies
Cao et al. (2018)
China,
2013-2015
337
0.09
0.03,
0.31°
0
Overall/Boys
+ Girls
+ Overall/
Girls
0 Boys
+ Overall/
Girls/Boys
Gross et al. (2020)
USA, 2014
98
0.108
(N/A)d
Overall/
Girls/Boys
Jensen et al. (2020a)
Denmark,
2010-2012
589
0.30
(0.08,
0.66)b
0
Overall/Girls
+ Boys
Lee et al. (2018)
S. Korea,
2012-2013
361
1.19
(0.22,
1.69)
- Overall
- Overall*3
N/A = not available.
* Denotes statistical significance at p < 0.05; 0 represents a null association; + represents a positive association;
- represents a negative association; -1" represents increased odds ratio; -l represents decreased odds ratio
Note: "Adverse effects" are indicated by both increased ORs (-) for dichotomous outcomes and negative
associations (-) for the other outcomes.
/ Denotes multiple groups with the same direction of associations.
aExposure-response relationship detected based on categorical data.
bNo range provided but 5th-95th percentiles included.
cNo range provided but 10th-90th percentiles included.
dDried blood spot PFHxS sample collected within 48 hours of birth.
3-169
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Anogenital distance
Four medium confidence studies examined the associations between PFHxS and AGD in
infants (see Figure 3-45). Reduced AGD is associated with clinically relevant outcomes in males,
including cryptorchidism, hypospadias, and lower semen quality and testosterone levels
(Thankamonv etal.. 2016). but adversity of reduced AGD is less established in females. Three
studies examined boys and girls (Lind etal.. 2017: Christensen etal.. 2021: Arbuckle etal.. 2020).
while one included boys only (Tian etal.. 2019b). All four studies were birth cohorts in Denmark
(Lind etal.. 2017). Faroe Islands (Christensen etal.. 2021) (cross-sectional analysis within cohort
sample), Canada f Arbuckle etal.. 20201. and China fTian etal.. 2019bl. In Arbuckle etal. f20201 and
Tian etal. f2019bl. AGD was measured shortly after birth (median 3.5 days). Christensen et al.
f2 0211 measured AGD at 2 weeks after the expected term date. Tian etal. f2019bl additionally
measured AGD at 6 and 12 months, and Lind etal. (2017) measured at 3 months. With greater
variability in timing of measurements, there is additional potential for misclassification with these
measures, but age at time of measurement was included in the statistical models in all studies.
II 1 1 1 1 1 1 1 1
Arbuckle, 2020, 6356900 -
Christensen, 2021, 9960218-
Lind, 2017, 3858512-
Tian, 2019, 5390052-
+
++
+
+
+
+
+
~
+
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)
+
++
+
+
-
+
+
+
+
++
++
+
+
+
n
+
+
+
+
+
Figure 3-45. Summary of study evaluation for epidemiology studies of
anogenital distance. For additional details see HAWC link.
In Lind etal. (2017). there was a statistically significant inverse association (i.e., shorter
AGD with higher exposure) with ASD among boys. The other three studies did not report decreased
AGD, despite greater exposure contrasts (see Table 3-19). In girls, there was an inverse association
with PFHxS for ACD fLind etal.. 20171. This was statistically significant with PFHxS analyzed as
continuous, although there was not a monotonic decrease across quartiles. A consistent but smaller
and nonsignificant association was also observed in the third and fourth quartiles for AFD. This
association is coherent with the decrease in testosterone observed in some studies (described
below in the Reproductive Effects section). However, in the other two studies Christensen et al.
(2021): Arbuckle etal. (2020). there was no decrease in either AGD measure with higher PFHxS
exposure.
3-170
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
AGD is a marker of androgen exposure, and thus an inverse in AGD would be expected to
correspond with a decrease in testosterone. This was not observed in the two studies of
testosterone in male neonates, but an inverse association was observed in a study of female
neonates (see Male and Female Reproductive Effects). The lack of coherence for males does not
reduce confidence in the AGD findings due to low confidence in the reproductive hormone studies.
However, the inconsistency across studies results in considerable uncertainty for an association
with AGD.
3-171
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-19. Associations between PFHxS and anogenital distance in medium
confidence epidemiology studies
Boys
Reference
Population
Median exposure
(IQR)
(ng/mL)
Effect
estimate
ASD
APD
Christensen
et al. (2021)
Cross-sectional analysis
within birth cohort in the
Faroe Islands; 232 boys at
2 wk post-term
Serum
0.2 (0.1-0.3)
P (95% CI)
for In-unit
increase
0.2 (-0.3, 0.7)
NR
Lind et al.
(2017)
Birth cohort in Denmark;
299 boys at 3 mo
Serum
0.3 (0.2-0.4)
P (95% CI)
for In-unit
increase
-1.2 (-2.3, -0.2)
-0.6 (-1.8, 0.5)
Quartiles vs.
Q1
Q2: 0.6 (-1.3, 2.4)
Q3: -0.3 (-2.1, 1.6)
Q4: -0.8 (-2.7, 1.2)
Q2: 2.6 (0.5, 4.6)
Q3: 0.9 (-1.0, 2.9)
Q4: 0.1 (-2.0, 2.3)
(Arbuckle et
al.. 2020)
Birth cohort in Canada;
198 boys at birth
Plasma
1.1(0.7-1.7)
P (95% CI)
for unit
increase
0.22 (-0.54, 0.98)
0.24 (-0.52, 1.01)
Quartiles vs.
Q1
Q2: -0.08 (-1.99, 1.83)
Q3: 0.13 (-1.80, 2.06)
Q4: 0.57 (-1.33, 2.46)
Q2: -0.91 (-2.74, 0.91)
Q3: 0.64 (-1.23, 2.51)
Q4: 0.57 (-1.30, 2.44)
Tian et al.
(2019b)
Birth cohort in China; 439
boys at birth
Plasma
2.8(2.2-3.6)
P (95% CI)
for In-unit
increase
Birth: -0.19 (-0.97, 0.58)
6 mo: 0.69 (-1.86, 3.23)
12 mo: 2.21 (-0.47, 4.89)
Birth: 0.35 (-0.55, 1.26)
6 mo: 0.04 (-2.53, 2.61)
12 mo: 0.60 (-2.62, 3.83)
Girls
Reference
Population
Median exposure
(IQR)
(ng/mL)
Effect
estimate
ACD
AFD
Christensen
et al. (2021)
Cross-sectional analysis
within birth cohort in the
Faroe Islands; 231 girls at
2 wk post term
Serum
0.2 (0.1-0.3)
P (95% CI)
for In-unit
increase
NR
-0.1 (-0.4, 0.3)
Lind et al.
(2017)
Birth cohort in Denmark;
212 girls at 3 mo
Serum
0.3 (0.2-0.4)
P (95% CI)
for In-unit
increase
-0.9 (-1.9, 0.0)
-0.3 (-1.1, 0.4)
Quartiles vs.
Q1
Q2:-1.6 (-3.4, 0.2)
Q3:-2.3 (-4.1,-0.5)
Q4: -1.6 (-3.4, 0.2)
Q2: 0.2 (-1.2, 1.6)
Q3: -0.8 (-2.2,0.6)
Q4: -0.5 (-1.6,0.9)
Arbuckle et
al. (2020)
Birth cohort in Canada;
205 girls at birth
Plasma
1.1 (0.7-1.7)
P (95% CI)
for unit
increase
0.3 (-0.47, 1.07)
0.14 (-0.79, 1.07)
Quartiles vs.
Q1
Q2: 1.01 (-0.56, 2.59)
Q3: 0.31 (-1.40, 2.02)
Q4: 0.92 (-0.94, 2.79)
Q2: 1.23 (-0.66, 3.13)
Q3: -0.51 (-2.56, 1.54)
Q4: 0.52 (-1.71, 2.75)
ASD = AGD measured from anus to the posterior base of the scrotum; APD = AGD measured from the center of the
anus to the cephalad insertion of the penile; ACD = AGD measured from the from the center of the anus to the
top of the clitoris; AFD = AGD measured from the top of the center of the anus to the posterior fourchette.
3-172
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Gestation duration
As shown in Figure 3-48,19 informative epidemiological studies assessed PFHxS in relation
to changes in gestational duration measures. All 19 studies examined gestational age, with 10 of
these providing analyses of both preterm delivery and gestational age. Fourteen of the 19
gestational duration studies were nested case-control studies or prospective cohort studies (Yang
etal.. 2022a: Workman etal.. 2019: Sagiv etal.. 2018: Meng etal.. 2018: Manzano-Salgado etal..
2017a: Maisonetetal.. 2012: Lind etal.. 2017: Huo etal.. 2020: Hiermitslev etal.. 2020: Hamm etal..
2010: Gardener etal.. 2021: Gao etal.. 2019: Buck Louis et al.. 2 018: Bach etal.. 20161. and five
were cross-sectional fXu etal.. 2019: Li etal.. 2017b: Gvllenhammar et al.. 2018: Eick etal.. 2020:
Bangma etal.. 20201. The 19 epidemiological studies examined here had maternal exposure
biomarkers collected either during trimesters one f Manzano-Salgado etal.. 2017a: Lind etal.. 2017:
Buck Louis etal.. 2018). two (Huo etal.. 2020: Hamm etal.. 2010). three (Gardener etal.. 2021: Gao
etal.. 2019) across multiple trimesters (Workman etal.. 2019: Sagiv etal.. 2018: Meng etal.. 2018:
Maisonet etal.. 2012: Hiermitslev etal.. 2020: Eick etal.. 2020: Bach etal.. 2016). or had
postpartum maternal or infant samples (Yang etal.. 2022a: Xu etal.. 2019: Li etal.. 2017b:
Gvllenhammar et al.. 2018: Bangma etal.. 20201.
Nine studies were classified as having late (defined as trimester 2 exclusive onward) and
early sampling biomarker sampling (defined as having at least some trimester 1 maternal
sampling). Four of the five-cross-sectional studies/analyses had late biomarker sampling. Among
the 14 cohort or nested case-control studies, eight studies had early biomarker sampling (Sagiv et
al.. 2018: Meng etal.. 2018: Manzano-Salgado etal.. 2017a: Maisonet etal.. 2012: Lind etal.. 2017:
Hiermitslev etal.. 2020: Buck Louis et al.. 2 018: Bach etal.. 2016). while six were classified as late
fYang etal.. 2022a: Workman etal.. 2019: Huo etal.. 2020: Hamm etal.. 2010: Gardener etal.. 2021:
Gao etal.. 20191. For examination of consistency and between-study heterogeneity, the type of
statistical analyses in addition to the type of study design was evaluated. As part of this evaluation,
cross-sectional analyses are considered for any study that used maternal serum/plasma, umbilical
cord or placental postpartum PFHxS measures in relation to gestational duration even if the data
were derived from prospective cohort or nested case-control studies (e.g., (Yang etal.. 2022al).
Preterm birth
Two fManzano-Salgado etal.. 2017a: Huo etal.. 20201 of the 10 preterm birth (<37
gestational weeks) studies reported sex-specific findings in addition to overall population results
(see Figure 3-46 and Table 3-20). Ten studies examined PFHxS and preterm birth including six high
(Sagiv etal.. 2018: Manzano-Salgado etal.. 2017a: Huo etal.. 2020: Gardener etal.. 2021: Eick etal..
2020: Bach etal.. 2016) and four medium confidence (Yang etal.. 2022a: Meng etal.. 2018:
Hiermitslev etal.. 2020: Hamm etal.. 20101 studies. Two studies had good study sensitivity (Sagiv
etal.. 2018: Meng etal.. 20181. six had adequate study sensitivity fManzano-Salgado etal.. 2017a:
3-173
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Hjermitslev etal.. 2020: Hamm etal.. 2010: Gardener etal.. 2021: Eick etal.. 2020: Bach etal.. 20161
and two were rated as deficient fYang etal.. 2022a: Huo etal.. 20201.
Six of the 10 studies showed no increased odds for preterm birth in relation to PFHxS (Yang
etal.. 2022a: Manzano-Salgado etal.. 2017a: Hiermitslev et al.. 2020: Hamm etal.. 2010: Eick etal..
2020: Bach etal.. 20161 with two reporting decreased risks (see Figure 3-47). The medium
confidence study by Hamm etal. (2010) found a statistically significant decreased exposure-
response relationship between preterm birth and the upper two PFHxS exposure tertiles (OR
range: 0.31 to 0.59). An inverse association (OR = 0.59; 95% CI: 0.33,1.06) was also detected in
girls in the largely null Manzano-Salgado etal. f2017al study.
Six studies were null for based on the overall population. The other four high and medium
confidence studies reported some increased ORs but were not always internally consistent in
direction of the effect estimates reported for different PFHxS exposure comparisons. The high
confidence Sagivetal. (2018) study reported largely null results based on continuous PFHxS
exposures but showed some associations based on their categorical analysis that were not dose-
dependent. For example, they reported an increased OR of preterm birth for PFHxS quartile 3
(OR = 1.8; 95% CI: 1.1, 3.1 for 2.5-3.7 ng/mL) and 4 (OR = 1.3; 95% CI: 0.7, 2.2 for 3.8-74.5 ng/mL)
compared with quartile one. Similarly, the medium confidence study by Mengetal. T20181 reported
no associations for the various definitions of preterm birth examined for PFHxS quartile 4 or per a
ln-unit increase. They did detect an increased OR of preterm birth for the second (OR = 2.3; 95% CI:
1.1, 4.6) and third (OR = 1.5; 95% CI: 0.7, 3.2) PFHxS quartiles compared with the first quartile.
However, small sample sizes limited the interpretation of these categorical data. The categorical
analysis in the high confidence Gardener et al. (2021) also found no dose-dependence but showed a
nonsignificant twofold increased risk of preterm birth in quartile 2 (OR = 2.11; 95% CI: 0.76, 5.81)
relative to quartile 1.
In the high confidence study by Huo etal. f20201. associations between PFHxS and different
preterm birth measures (including overall and different sub-types) were just above the null value
based on continuous or categorical exposures for the overall population. However, an association
was seen for clinically indicated preterm births for each ln-unit increase (OR = 1.58; 95% CI: 0.82,
3.05) and for tertile 3 (OR = 1.43; 95% CI: 0.66, 3.08). A small nonsignificant increased risk was also
seen for overall preterm birth (OR = 1.33; 95% CI: 0.77, 2.27 per each ln-unit) in girls only, with
larger statistically significant associations noted among girls only for the clinically indicated
preterm birth subtype (OR = 2.56; 95% CI: 1.18, 5.53).
3-174
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
a*
1 &
. ao^°' s^a v ^ :6 & \
Participant selection •
Exposure measurement-
Outcome ascertainment -
Confounding -
Analysis -
Sensitivity -
Selective Reporting -
Overall confidence -
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 10 epidemiology studies of
preterm birth. For additional details see HAWC link.
3-175
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Population
Overall Study
Confidence
Study Sensitivity
Exposure
Window
Regression
Coefficient
Exposure
Comparison
Regression coefficient
# PTB Relative Risk (RR)
9 PTB Relative Kisk (KR! p<0.0b
H 95% confidence interval
Manzano-Salgado
et al.. 2017,
INMA cohort (2003-2008) 1202
mother-infant pairs
iHighl
Adequate
Cohort
(Prospective)
Trimester 1
0.79
In-unit (ng/mL)
4238465
I
Bach et al.. 2016,
3981534
Aarhus Birth Cohort (2008-2013),
Denmark. 1507 mother-infant pairs
IHighl
Adequate
Cohort
(Prospective)
Trimester 1-2
0.72
0.97
0.68
0.78
Quartile 2
Quamle 3
I—•—4 1
1
—^ 4
In-unit (ng/mL)
1
1 >1 1
Sagiv. 2018.
4238410
Project Viva (1999-2002) 1645
mother-infant pairs
iHigh|
Good
Cohort
(Prospective)
T rimester 1 -2
0.9
Quartile 2
i—•;—i
1.3
i
In-unit (ng/mL)
i
>-•<
i
Eick etal., 2020.
7102797
Chemicals In Our Bodies (CIOB)
(2014-2018). US. 497 female
participants
|High|
Adequate
Cross-sectional
Trimester 1-3
1.32
Tertile 2
i
i
1.15
Tertile 3
Huo etal.. 2020,
6835452
Shanghai Birth Cohort (2013-2016),
China, 2849 mother-infant pairs
!High|
Deficient
(Prospective)
Tnmester 2
1.07
1.17
1.16
Tertile 2
Tertile 3
In-unit (ng/mL)
i
i
i '•
Vanguard Pilot Study of the National
Children's Study (NCS) (2009), 5420
mother-infant pairs
IHighl
Adequate
Cohort
Trimester 3
7021199
(Prospective)
Yang et al.. 2022.
10176806
Kashgar Birth Cohort (2018-2019),
China. 768 mother-infant pairs
|Medium|
Deficient
Nested
At delivery
1.01
In-unit (ng/mL)
hf*
Meng etal.. 2018.
4829851
DNBC (1996-2002). Denmark. 3535
mother-infant pairs
|Medium|
Good
Trimester 1-2
11 a
(Prospective)
i
In-unit (ng/mL)
i
»-)•—i
Hjermitslev, 2020.
5880849
ACCEPT birth cohort (2010-2011.
2013-2015). Greenland. 482
mother-infant pairs
|Medium|
Adequate
Cohort
(Prospective)
Trimester 1-3
0.813
In-unit (ng/mL)
IS1 l
i
Hamm. 2010,
1290814
Alberta cohort (2005-2006) 252
mother-infant pairs
|Medium|
Adequate
Cohort
(Prospective)
Trimester 2
0.59
Tertile 2
i
0.31
Tertile 3
Manzano-Salgado
et al., 2017,
4238465
INMA cohort (2003-2008) 1202
mother-infant pairs
|High|
Adequate
(Prospective)
Trimester 1
0.97
In-unit (ng/mL)
BOYS
.
i
Huo etal.. 2020,
6835462
Shanghai Birth Cohort (2013-2016),
China, 2849 mother-infant pairs
|High|
Deficient
Cohort
(Prospective)
Tnmester 2
1
In-unit (ng/mL)
Manza no-Sa Igario
el a!.. 2017,
4238465
INMA cohort (2003-2008) 1202
mother-infant pairs
|High|
Adequate
Cohort
(Prospective)
Trimester 1
0.59
In-unit (ng/mL)
GIRLS
i
Huo etal.. 2020.
6835452
Shanghai Birth Cohort (2013-2016),
China. 2849 mother-infant pairs
iHighl
Deficient
Cohort
(Prospective)
Tnmester 2
1.33
In-unit (ng/mL)
H-4—# 1
-1 -0.5
0.5 1 1.5 2 2.5
3.5
4.5 5 5.5 6
Figure 3-47. Preterm birth results for 10 PFHxS epidemiological studies. For
additional details see HAWC link.
PTB = preterm birth.
aStudies are sorted first by overall study confidence level then by exposure window examined.
bSex specific data below solid black line; newborn boys above dotted line, newborn girls below.
cFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g., Yang et al. (2022al).
3-176
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Gestational age z_overall population results
Seventeen of the 19 epidemiological studies examined mean gestational age data in the
overall population, with the other two only reporting sex-specific findings fMaisonet etal.. 2012:
Lind etal.. 2017) for PFHxS and gestational age relationships. Four studies reporting both sex-
specific and overall population results (Meng etal.. 2018: Manzano-Salgado etal.. 2017a: Li etal..
2017b: Hiermitslev etal.. 20201. Among the 19 studies with gestational age measures, eight were
high confidence f Sagiv etal.. 2018: Manzano-Salgado etal.. 2017a: Lindetal.. 2017: Huo etal.. 2020:
Gardener etal.. 2021: Eick etal.. 2020: Buck Louis et al.. 2 018: Bach etal.. 20161. five were medium
fYang etal.. 2022a: Meng etal.. 2018: Maisonet et al.. 2 012: Hiermitslev etal.. 2020: Gvllenhammar
etal.. 20181. and six were low confidence studies fXu etal.. 2019: Workman etal.. 2019: Li etal..
2017b: Hamm etal.. 2010: Gao etal.. 2019: Bangma et al.. 2020) (see Figure 3-48). Five (Sagiv et al..
2018: Meng etal.. 2018: Maisonet etal.. 2012: Li etal.. 2017b: Gvllenhammar et al.. 2018) of the 19
studies received a good rating in the study sensitivity domain, while eight (Manzano-Salgado etal..
2017a: Lind etal.. 2017: Hiermitslev etal.. 2020: Hamm etal.. 2010: Gardener etal.. 2021: Eick et
al.. 2020: Buck Louis et al.. 2 018: Bach etal.. 20161 were considered adequate and six were deficient
fYang etal.. 2022a: Xu etal.. 2019: Workman etal.. 2019: Huo etal.. 2020: Gao etal.. 2019: Bangma
etal.. 20201.
Six (Workman etal.. 2019: Huo etal.. 2020: Gvllenhammar et al.. 2018: Buck Louis etal..
2018: Bangma etal.. 2020: Bach etal.. 2016) of the 17 studies in the overall population reported no
associations between gestational age and PFHxS exposures, while four reported an increased
gestational age with increasing PFHxS exposures fXu etal.. 2019: Li etal.. 2017b: Hamm etal..
2010: Eick etal.. 20201 (see Table 3-20 or Figure 3-49). For example, the low confidence study by
Xu etal. T20191 reported a very large increase in gestational age ((3 = 3.38 weeks; 95% CI: -0.80,
7.55) per ln-unit increase in PFHxS. The Buck Louis et al. f20181 study was largely null in the
overall population and reported some small nonsignificant differences for black ((3 = -0.14 weeks;
95% CI: -0.34, 0.05 for each ln-unit increase) and Asian ((3 = -0.09 weeks; 95% CI: -0.40, 0.21 for
each ln-unit increase) neonates.
Seven studies reported some gestational age reductions in relation to PFHxS in the overall
population. Although their continuous PFHxS exposure results were null, the high confidence study
by Sagiv etal. f 20181 showed small nonsignificant decreases for quartiles 3 and 4 albeit not in a
nonmonotonic fashion. Although their overall population results were null, based on each ln-unit
increase,.the high confidence study by Manzano-Salgado etal. (2017a) did show a small decrease in
gestational age for quartile 4 ((3 = -0.16 weeks; 95% CI: -0.43, 0.1). The medium confidence study
by Hiermitslev etal. (2020) reported a relatively large gestational age reduction ((3 = -0.32 weeks;
95% CI: -0.72, 0.08 per each ln-unit increase). The medium confidence study by Yang etal. f2022al
showed larger gestational age reductions among term births ((3 = -0.64; 95% CI: -1.64, 0.36)
compared with preterm births ((3 = -0.20 weeks; 95% CI: -3.32, 2.93) per each ln-unit increase in
3-177
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Total PFHxS exposures. The medium confidence Mengetal. (2018) study reported a decrease based
on continuous exposure ((3 = -0.29 weeks; 95% CI: -1.15, 0.58 per each ln-unit PFHxS increase) and
small nonmonotonic decreases across quartiles ((3 range: -0.06 to -0.17 weeks). The low
confidence study by Gao etal. f20191 reported a nonmonotonic decreased gestational age in
relation to PFHxS tertiles 2 ((3 = -0.37 weeks; 95% CI: -0.82, 0.09) and 3 ((3 = -0.22 weeks; 95% CI:
-0.71, 0.27). Although there was no evidence of an exposure-response relationship, the high
confidence study by Gardener et al. (2021) reported that participants in the three upper PFHxS
quartiles had smaller gestational ages ((3 range: -0.18 to -0.75) relative to quartile 1.
Although they were not always internally consistent across exposure metrics, 7 (3 high, 3
medium, and 1 low confidence) of 17 studies in the overall population showed some gestational age
reductions in relation to PFHxS exposures. Few study characteristics appeared to be related to
patterns across the study results. For example, four of the seven studies showing inverse
associations were based on early biomarker sampling. Study sensitivity in the six (three high, one
medium, and one low confidence) may explain some of the null findings as half of the studies had
deficient ratings (one good, two adequate, and three deficient).
Participant selection A
Exposure measurement A
Outcome ascertainment A
sp6
Figure 3-48. Study evaluation results for 19 epidemiological studies of
gestational age and PFHxS. For additional details see HAWC link.
3-178
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Maruano-Saigado INMA cohort (2003-2008) 1202
etal.,2017. mother-infant pairs
4238465
Overall Study Confidence Study Sensitivity Design
|Hlgh| Adequate Cohort
Project Viva (1999-2002) 1645
Chemicals In Our Bodies (CIOB)
(2014-2018), US, 497 female
participants
Vanguard Pilot Study of the National
Children's Study (NCS) (2009), 5420
mother-infant pairs
l„ 2018, DNBC (1996-2002), Denmark. 3
Hjermitslev. 2020, ACCEPT birth cohort
5880849 2013-2015), Greenlar
mother-Infant pairs
Workman et al.. Canadian Healthy Infant Longitudinal
2019. 5387046 Development (CHILD) Study
(2010-2012). Canada (414
Gao et al., 2019. Affiliated Hospital of Capital Medical
5387135 University (2015-2016), China, 132
pregnant women
Cross-sectional study (2018-2017),
China, 98 mother-infant pairs
PTB (2015-2018), US, 122 placentas
|Medium|
IMediuml
|Medlum|
IMediuml
I Low|
|Low|
|Low|
Trimester 1-2
Deficient
Adequate
Adequate
At delivery
Trimester 1-2
Trimester 1-3
Good
Deficient
Cross-sectional
Cross-sectional
Cross-sectional
-0.025
-0.64
Trimester 3 -0.37
Quartile 2
Quartlle 3
In-unit (ng/mL) in
In-unit (ng/mL) in
Quartlle 2
Quartile 3
Tertile 3
Tertile 2
Tertile 3
In-unit (ng/mL) inert
Quartile 2
Quartlle 3
Quartile 4
n't (ng/mL) incroa:
as)
Quartile 2
Quartlle 3
unit (ng/mL) inert
unit (ng/mL) inert
-0.22 Tertile 3
0.12 In-unit (rvg/mL) inci
3.38 In-unit (ng/mL) inci
0.002 In-unit (ng/mL) Inci
# [association with GA]
o [association with GA] p<0.05
1—{95% confidence interval
Figure 3-49. Overall population gestational age results for 17 PFHxS
epidemiological studies. For additional details see HAWC link.
GA = gestational age.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bThe (Yang et al., 2022a) -0.64 per interquartile Increase value is reported in the term birth population; the -0.2
per interquartile increase value is in the preterm birth population.
cFigure 4 in Gardener et al. (2021) was used to estimate gestational age differences estimated from digitization
95% CIs were not estimable.
dFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g., Yang et al. (2022al).
eYang et al. (2022a preterm results are truncated: the complete 95% CI ranges from -3.32 to 2.93. Term results
are truncated; the complete 95% CI ranges from -1.64 to 0.36.
fXu et al. (2019; results are truncated: the complete 95% CI ranges from -0.8 to 7.55.
gUnlike other studies that relied on maternal or cord serum or plasma (in ng/mL), Bangma et al. (2020 used
placental exposure measures (in ng/g).
3-179
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Gestational age - sex-specific results
Eight (four high, three medium, and one low confidence) epidemiological studies examined
mean gestational age in relation to PFHxS in either or both sexes including one that evaluated data
in girls only fMaisonet etal.. 20121 (see Figure 3-50). None of the seven studies in boys showed
decreased gestational age with increasing PFHxS, with six studies showing null associations (Sagiv
etal.. 2018: Mengetal.. 2018: Manzano-Salgado etal.. 2017a: Lind etal.. 2017: Hjermitslevetal..
2020: Eick etal.. 2020). The low confidence study by Li etal. (2017b) reported a small increased
gestational age per each ln-unit PFHxS increase ((3 = 0.20 weeks; 95% CI: -0.02, 0.42) among boys.
Five (Sagiv etal.. 2018: Meng etal.. 2018: Manzano-Salgado etal.. 2017a: Li etal.. 2017b:
Eick etal.. 20201 of the eight studies in girls reported null associations between PFHxS and mean
gestational age, while another study fEick etal.. 20201 reported nonsignificant increased
gestational age across tertiles ((3 range: 0.18 to 0.33). Three studies in girls showed some
gestational age reductions including some that were moderately large in magnitude. The high
confidence study by Lind etal. (2017) showed some suggestion of an exposure-response
relationship for mean gestational age across the upper three PFHxS quartiles ((3 range: -0.33 to
-0.86 weeks) including a large association ((3 = -0.86 weeks; 95% CI: -1.34, -0.29) in quartile 4
(0.4-7.3 ng/mL) versus quartile 1 (0.2-0.29 ng/mL). The medium confidence study by Hiermitslev
etal. f20201 also reported a large gestational age reduction ((3 = -0.57 weeks; 95% CI: -1.04, -0.10
per each ln-unit increase). In their study population of female infants only, the medium confidence
study by Maisonet et al. (2012) reported nonstatistically significant decreases in gestational age
with some suggestion of an exposure-response relationship. They reported reduced gestational age
in the second ((3 = -0.15 weeks; 95% CI: -0.52, 0.22 for 1.3-2.0 ng/mL) and third PFHxS tertiles
((3 = -0.24 weeks; 95% CI: -0.62, 0.14 for 2.0-54.8 ng/mL) compared with the lowest tertile
(<1.3 ng/mL).
Overall, three (one high and two medium confidence) studies out of eight studies in girls
only showed reduced gestational age in relation to PFHxS exposures. Although they were not
always monotonic, both of the studies with categorical data showed some evidence of exposure-
response relationships which lends support to the findings based on continuous exposure
metrics. There was no evidence of inverse associations among boys, although half of the studies had
deficient study sensitivity. Few other study characteristics appeared to be related to patterns across
the study results; however, all three of the studies showing inverse associations in females were
based on early biomarker sampling.
3-180
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Population
Overall Study
Study Sensitivity
Design
Exposure Window Regression
Exposure Comparison
Confidence
Coefficient
Regression coefficient
# p [association with GA]
Llnd. 2017,
3858512
Odense Child Cohort (2010-2012)
638 mother-Infant paiis
lHlgh|
Adoquate
Cohoit
(Prospective)
Trimester 1 0.07
Quartlle 2
0 p. [association with GA] p«0.05
H 95% confidence Interval
0
Quartie 4
» i 1
004
In-unit (ng'mL) Incrsase
Sagiv, 2018,
4238410
Project Viva (1999-2002) 1645
mother-infant pairs
|High|
Good
(Prospective)
Trimester 1-2 -0.01
Manzano-Salgado
el al.. 2017.
4238465
INMA cohort (2003-2008) 1202
mothei-inlant pairs
|High|
Adequate
Cohort
Trimester 1 -0,09
In-unit (ng'mL) increase
i—•th
Eick et al.. 2020,
7102797
Chemicals In Our Bodies (CIOB)
(2014-2018). US. 497 female
participants
High|
Adequate
Cross-sectional
Trimester 1-3 0.18
Tertile 2
|
1
Tertile 3
Meng ct al.. 2018.
4829851
DNBC (1996-2002). Denmark. 3535
|Mcdium|
(Prospective)
Trimester 1-2 -0.06
In-unit (ng'mL) Increase
Hjermitslev, 2020.
5880849
ACCEPT birth cohort (2010-2011,
2013-2015). Greenland. 482
mother-infant pairs
|Medium|
Ad"1""*
(Prospective)
Trimester 1-3 -0.08
In-unit (ng'mL) increase
l •' l
Li, 2017, 3981358
GBCS (2013), Chins, 321
mother-infant pairs
|Lcnv|
OOM
Cross-sectional
At birth 0.2
In-unit (ng'mL) increase
i! » i
Lind. 2017.
3858512
Odense Child Cohort (2010-2012)
High |
Adequate
Cohoit
Trimester 1 -0.33
-0.33
Quartlle 2
Quartila 3
i • —i
-0.17
1 • i 1
Sagiv, 2D18,
4238410
Project Viva (1999-2002) 1645
mother-infant pairs
|High|
Good
Cohort
(Prospective)
Trimester 1-2 0
In-unit (ng'mL) increase
Maivano-Salgado
al al.. 2017,
4238465
INMA cohort (2003-2008) 1202
mothei-inlant pairs
|High|
Adequate
Cohort
1 times ter 1 0.06
In-unit (ng'mL) increase
!•—1
Eick et al.. 2020,
7102797
Chemicals In Our Bodies (CIOB)
(2014-2018). US. 497 female
parti clpants
|High|
Adequate
Cross-sectional
Trimester 1-3 0.33
Tertile 2
1
Mcngctal.. 2018.
4829851
DNBC (1996-2002), Denmark. 3535
mother-Infant pairs
|Medium|
Good
Cohort
(Prospective)
Trimester 1-2 0.01
In-unit (ng'mL) Increase
»-#—i
Hjermitslev, 2020.
5880849
ACCEPT birth cohort (2010-2011,
2013-2015). Greenland. 482
mother-infant pairs
|Medium|
Adequate
Cohort
(Prospective)
Trimester 1-3 -0,57
In-unit (ng'mL) increase
, .
Maisonet et al.,
2012, 1332465
ALSPAC (1991-1992). U.K., 447
mother-girl pairs
|Medium|
Good
Cohort
(Prospective)
Trimester 1-3 -0.15
Tertile 2
Tertile 3
¦ • 1 I
In-unit (ng'mL) increase
l
mother-infant pairs
5 -1 —0.5 0 o!5
1 1.5 2
Figure 3-50. Sex-stratified gestational age results for 8 PFHxS epidemiological
studies. For additional details see HAWC link.
GA = gestational age.
aStudies are sorted first by overall study confidence level then by Exposure Window examined.
bFor evaluation of patterns of results, EPA considered studies that collected biomarker samples concurrently or
after birth to be cross-sectional analyses (e.g., Yang et al. (2022a)).
cLind et al. (2017) results are truncated: the complete 95% CI ranges from -3.1 to 0.7.
Gestational duration summary
There was mixed evidence within and between studies examining adverse associations
between PFHxS exposure with any gestational duration measures (preterm birth or gestational
age). Out of 19 total studies, 8 different ones showed gestational duration associations with
PFHxS. Four of 10 studies showed some increased odds preterm birth and PFHxS exposures in the
overall population or either or both of the sexes, although these were not always internally
consistent Seven of 17 studies in the overall population reported mean gestational age deficits in
relation to PFHxS, while 3 of 8 studies with sex-specific data only reported inverse associations in
girls. In addition to the null studies, a few studies also reported increased gestational age related to
PFHxS exposures. Gestational age can be prone to some measurement error which may reduce the
ability of some studies to detect statistically significant results for this endpoint The preterm birth
binary endpoint may also be less impacted by this measurement error given the broad classification
of preterm versus term births.
3-181
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-20. Summary of 19 epidemiological studies of PFHxS exposure and gestational duration measures-
Author
Study location/
years
N
PFHxS median
(ng/mL) exposure
Overall confidence
descriptor
Study sensitivity
domain
PTB
GA
Bach et al. (2016)
Denmark 2008-
2013
1,507
0.5
High
Adequate
0 All
0 All
Buck Louis et al. (2018)
USA,
2009-2013
2,106
0.71
High
Adequate
0 All
Eick et al. (2020)
USA,
2014-2018
506
0.33
High
Adequate
0 All
+ All
0 Boys/Girls
Gardener et al. (2021)
USA,
2009-2013
354
0.5
High
Adequate
t All
-All
Huo et al. (2020)
China,
2013-2016
2,849
0.54
High
Deficient
0 All/Boys T Girls
0 All
Lind et al. (2017)
Denmark, 2010-
2012
636
0.3
High
Adequate
-Girls
0 Boys
Manzano-Salgado et al.
(2017a)
Spain,
2003-2008
1,202
0.58
High
Adequate
0 All/Boys -i- Girls
-All
0 Boys/Girls
Sagiv et al. (2018)
USA,
1999-2002
1,645
2.4
High
Good
t All
-All
0 Boys/Girls
Gvllenhammar et al. (2018);
2017a
Sweden,
1996-2001
381
2.4
Medium
Good
0 All
Hiermitslev et al. (2020)
Greenland,
2010-2015
266
0.51
Medium
Adequate
0 All
-All/Girls
0Boys
Maisonet et al. (2012)
United Kingdom,
1991-1992
444
1.6
Medium
Good
— Girlsb
3-182
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Author
Study location/
years
N
PFHxS median
(ng/mL) exposure
Overall confidence
descriptor
Study sensitivity
domain
PTB
GA
Meng et al. (2018)
Denmark 1996-
2002
2,132
~1
Medium
Good
T All
0 All/Boys/Girls
Hamm et al. (2010)
Canada,
2005-2006
252
2.1
Medium/ Lowc
Adequate
¦I Allb
+ All
(Yang et al., 2022a)
China, 2018-
2019
768
0.049-0.058d
Medium
Deficient
0 All
— Al ld
Bangma et al. (2020)
USA, 2015-2018
122
0.0676
Low
Deficient
0 All
Gao et al. (2019)
China, 2015-
2016
132
0.24
Low
Deficient
-All
Li et al. (2017b)
China,
2013
321
3.87
Low
Good
+ All/Boys
0 Girls
Workman et al. (2019)
Canada, 2010-
2011
414
0.44
Low
Deficient
0 All
Xu et al. (2019)
China,
2016-2017
98
0.61
(0.30-1.94)f
Low
Deficient
+ Overall
PTB = preterm birth; GA = gestational age.
* Denotes statistical significance at p < 0.05; 0 : represents a null association; + : represents a positive association; - : represents a negative association; T:
represents an increased odds ratio; -l: represents a decreased odds ratio; / implies that multiple groups shared the same classification.
Note: "Adverse effects" are indicated by both increased odds ratios () for dichotomous outcomes and negative associations (-) for the other outcomes.
aGvllenhammar I (2017) and Gvllenhammar et al. (2018) results are included here (both analyzed the POPUP cohort).
"Exposure-response relationship detected based on categorical data.
'Hamm et al. (2010) was medium confidence for PTB and low confidence for GA.
"Median range across cases and controls.
"Exposure measured in placenta (ng/g).
f5th-95th percentiles.
3-183
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Fetal loss/spontaneous abortion
Five studies reported on the relationship between PFHxS exposure and spontaneous
abortion (see Figure 3-51). A cohort of pregnant women enrolled at 8-16 weeks gestation (Jensen
etal.. 20151 was considered low confidence primarily due to loss to follow-up and the high risk of
incomplete case ascertainment (i.e., not including women with losses that occurred prior to study
enrollment, which may bias the results toward or even past the null if there is a true association
between PFHxS exposure and spontaneous abortion (Radke etal.. 201911. Liewetal. f20201 is a
case-control study that identified cases via medical registry and also has the potential to miss early
losses. However, this study was not downgraded to low confidence as loss to follow-up was not a
concern. Three additional studies were considered medium confidence, two case-control studies of
first trimester miscarriage fWikstrom etal.. 2021: Mi etal.. 20221 and a cohort of women
undergoing their first in vitro fertilization-embryo transfer treatment cycle (Wang etal.. 2021a).
Notably, Mi etal. (2022) measured sodium perfluoro-l-hexanesulfonate, a related salt, rather than
PFHxS.
Tensenetal. (2015) reported an increased OR (1.53; 95% CI: 0.99, 2.38) for spontaneous
abortion for each ln-unit increase in exposure despite study sensitivity limitations. While this study
is low confidence, the bias is unlikely to be away from the null (as described above), and thus the
limitations are unlikely to explain the observed positive association. However, the other four
studies, all medium confidence, reported no association between PFHxS exposure and early
spontaneous abortion. It is possible that there is only an association with second trimester
spontaneous abortion, but the evidence is currently not adequate to make this determination and
there is considerable uncertainty due to inconsistency across studies.
Jensen, 2015, 2850253
Liew, 2020, 6387285
Mi, 2022, 10413561
Wang, 2021, 10176703
Wikstrom, 2021, 7413606
eO
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Birth defects
Two studies examined birth defects in relation to PFHxS exposures (see Figure 3-52). The
medium confidence congenital heart defect study by Ou etal. f20211 reported null associations
risks for PFHxS >0.153 ng/mL (versus <0.153 ng/mL) for septal defects (OR = 1.07; 95% CI: 0.52,
2.22), and total heart defects (OR = 1.03; 95% CI: 0.65,1.64), although a nonsignificant inverse risk
was seen for conotruncal defects (OR = 0.64; 95% CI: 0.28,1.49). Relative to tertile 1, the low
confidence Cao etal. (2018) study showed evidence of monotonic associations between all birth
defects and PFHxS tertiles 2 (OR = 2.24; 95% CI: 1.05, 5.27) and 3 (OR = 2.54; 95% CI: 1.06, 6.13).
There is considerable uncertainty in interpreting results for broad all birth defect groupings which
decreases study sensitivity given the etiological heterogeneity across different birth defects.
Overall, there was limited evidence of associations between PFHxS and birth defect based
on the two available epidemiological studies. Despite an exposure-response relationships in one
low confidence study based on an all (i.e., total) birth defect grouping, there is currently insufficient
data for any specific birth defects to draw further conclusions given the limitations noted above.
3-185
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Participant selection -
Exposure measurement
Outcome ascertainment -
Confounding -
Analysis
Sensitivity -
Selective Reporting -
Overall confidence -
e^V' s ®V®V
Ga° O^
++
¦
++
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 3-52. Summary of study evaluation for two epidemiology studies of
birth defects. For additional details see HAWC link.
Animal Studies
Five of the available toxicology studies evaluated PFHxS-induced effects in developing
animals. Three studies exposed Wistar rats fButenhoffetal.. 2009: 3M. 20031 or CD-I mice (Chang
etal.. 20181 to PFHxS for 14 days before mating and during mating, gestation, and lactation while
Marques etal. (20211 treated CD-I mice with PFHxS from GD 1 to PND 20; one study exposed
Wistar rats from GD 7 to PND 22 (Ramhaj etal.. 2018): and a separate study using Wistar rats
treated animals from GD 7 to GD 22 and from PND 1 to PND 22 (Tetzlaffetal.. 2021). These studies
administered PFHxS (doses ranging from 0.03 to 45 mg/kg-day) via gavage and evaluated maternal
toxicity and fetal survival, growth, and morphological development The Butenhoff et al. (2009). 3M
(20031 and Chang etal. (20181 studies were evaluated as high confidence, while the Ramhai et al.
(2018). Marques etal. (2021). and Tetzlaffetal. (2021) studies were evaluated as medium
confidence (see Figure 3-53). Concerns in the Ramhai etal. (20181. Tetzlaffetal. (20211. and
3-186
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Marques etal. (2021) studies were noted for allocation, and the reporting of the number of animals
per exposure group.
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
„ - orf!-V
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Not reported
Figure 3-53. Developmental animal study evaluation heatmap. For additional
details see HAWC link.
Maternal health
The health of the dams was assessed in all available studies except Tetzlaff etal. f20211 (see
Figure 3-54). Butenhoff et al. f20091: 3M f20031 reported that Sprague Dawley rats administered
PFHxS displayed decreased maternal body weight (6% to 8% relative to controls) during the
lactation period: on PNDs 4, 6-8,11, and 13 at the lowest dose (0.3 mg/kg-day); on PNDs 7 and 8 at
3 mg/kg-day; and on PNDs 4, 6-9,11,13, and 14 atthe highestdose (10 mg/kg-day). However,
these decrements are considered minimal, the animals recovered from these effects at weaning
(PND 22), and studies in CD-I mice (Marques etal.. 2021: Chang etal.. 2018) or Wistar rats
fRamhai et al.. 20181 did not report significant PFHxS-induced effects on maternal body weight
during gestation or lactation. Maternal food consumption was also not affected in exposed rats or
mice (Chang etal.. 2018: Butenhoff et al.. 2009: 3M. 2003). Additional outcomes evaluated in F0
females included kidney and liver weights, reproductive organ weights and histopathology, and
maternal serum thyroxine levels, which are discussed in those respective sections (see Sections
3.2.3, 3.2.4, and 3.2.10). Briefly, significant treatment-related increases were observed for mean
liver weight and the incidence of histopathological findings at 3 mg/kg-day in CD-I mice (Chang et
al.. 20181. and significant treatment- and dose-related decreases were observed in serum thyroxine
levels in Wistar rats fRamhai etal.. 20181: see hepatic and thyroid effect sections (see Sections 3.2.5
and 3.2.1, respectively) for more detail.
3-187
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Fetal viability
Endpoints related to fetal and postnatal viability were measured in the Butenhoff et al.
f20091. Chang etal. f20181. Marques etal. f20211. and Ramhai etal. f20181 studies. Post-
implantation loss, perinatal loss, number of live pups, litter size, and number of stillborn pups were
not affected by PFHxS exposure in Sprague Dawley or Wistar rats (Ramhaj etal.. 2018: Butenhoff et
al.. 2009: 3M. 2003). and Marques etal. (2021) reported no PFHxS-induced effects on live births per
litter in CD-I mice. However, a similar study in CD-I mice reported that exposure to PFHxS at 1 and
3 mg/kg-day decreased the related measures of live litter size (by 14% and 12%, respectively) and
the number of pups born per litter (by 12% and 11%, respectively) f Chang etal.. 20181. An
explanation for the lack of dose-dependence of these observations is unavailable. Decreased litter
size is considered an indirect indication of preimplantation loss and resorptions flPCS. 20061. but
the Chang etal. (2018) study did not measure either of these two outcomes. This mouse study also
evaluated the number of pups born-to-implant ratio and pup survival and reported no treatment-
related effects (Chang etal.. 2018). The finding of reduced litter size and live pups per litter in mice
but not in rats exposed to higher PFHxS levels is not explainable by differences in
pharmacokinetics, study design, or study evaluation considerations. Furthermore, the toxicological
significance of these effects observed in mice is not clear as these responses did not appear to be
dose dependent; other measured developmental outcomes were not altered in the Chang et al.
(2018) study.
Fetal growth
F1 animal growth was evaluated in all available animal developmental studies. PFHxS
exposure did not affect pup body weights in male or female Sprague Dawley and Wistar rats, or in
CD-I mice fTetzlaffetal.. 2021: Ramhai etal.. 2018: Marques etal.. 2021: Chang etal.. 2018:
Butenhoff et al.. 2009: 3M. 20031. Furthermore, no significant treatment-related effects were
observed on sex ratio in Sprague Dawley and Wistar rats (Ramhaj etal.. 2018: Butenhoff etal..
2009: 3M. 2003). or in CD-I mice (Chang etal.. 2018) suggesting PFHxS exposure did not
specifically affect male or female animals.
Morphological development
Gross pathological examination of F1 pups revealed no significant exposure-related
developmental effects in exposed Sprague Dawley and Wistar rats, or CD-I mice fRamhai et al..
2018: Chang etal.. 2018: Butenhoff et al.. 2009: 3M. 2003).
Small but significant alterations in F1 AGD at birth were observed in CD-I mice and Wistar
rats (Ramh0i etal.. 2018: Chang etal.. 2018). Chang etal. (2018) reported that adjusted (i.e.,
relative to cube root body weight) PND 1 AGD was increased by 3% to 5% in male CD-I mice at
doses ranging from 0.3 to 3 mg/kg-day; and in female PND 1 mice, adjusted AGD was decreased by
5% only at the mid-dose (1 mg/kg-day). AGD is used as a phenotypical marker of androgen
3-188
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
levels/production during the masculinization programming window (Foster and Gray. 2013).11
Other phenotypical markers of androgen disruption were not altered in the available studies. On
PND 13 male nipple retention (another marker indicative of hormonal alterations fFoster and Gray.
201311 was not altered by PFHxS treatment in CD-I mice, and puberty onset was not affected in
either CD-I mice or Wistar rats fRamhai etal.. 2018: Chang et al.. 20181. Additionally, male, and
female reproductive organ weights in F1 CD-I mice (at PND 36) and Wistar rats (males at PND 16,
females at PND 17 or 22) were not affected by PFHxS treatment (Ramh0i etal.. 2018: Chang etal..
20181.
The biological significance of the small and directionally inconsistent changes in androgen-
dependent AGD measures in animal and human studies is unclear. Taken together, the available
evidence does not support an effect on reproductive organ development by PFHxS exposure in
these animal studies.
nIn rodent models and in humans AGD is longer in males when compared to females (Dean and Sharpe.
20131. Decreases in AGD are associated with androgen disruption during the masculinization programming
window fFoster and Gray. 2013: Dean and Sharpe. 20131. whereas increased AGD in females could be
indicative or increased androgen levels or activation of the androgen receptor fFoster and Gray. 20131.
Exposure to chemicals known to impair androgen synthesis or antagonize the androgen receptor have been
shown to result in decreased AGD as well as effects on other indicators of hormone disruption (e.g., increased
nipple retention) or adverse effects in the reproductive system (e.g., testicular atrophy, epididymal
malformations, testicular size, hypospadias, reduced size of the testis and accessory reproductive glands)
(Dent et al.. 20151.
3-189
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Effect
Study Name
Endpoint Name
Animal Description
PFHxS Animal Developmental Effects
Maternal Body Weight
Butenhoff, 2009, 1405789
Maternal Body Weight Gain, GD 0-20
P0 Rat, Sprague-Dawley (2)
• No significant change
+—
a
—•
~
Maternal Body Weight, PND 4
P0 Rat, Sprague-Dawley (2)
A Significant increase
V
V
Maternal Body Weight, PND 6
P0 Rat, Sprague-Dawley ( )
V Significant decrease
V
V
Maternal Body Weight, PND 8
P0 Rat, Sprague-Dawley (2)
V
V
V
Maternal Body Weight, PND 7
P0 Rat, Sprague-Dawley (2)
V
V
V
Maternal Body Weight, PND 11
P0 Rat, Sprague-Dawley (2)
V
•
V
Maternal Body Weight Change, PND 1-22
P0 Rat, Sprague-Dawley ( )
•—
—•
Chang, 2018, 4409324
Maternal Body Weight Change, GD 0-18
P0 Mouse, CD-1 (£)
~—
—•
Maternal Body Weight Change, PND 1-21
P0 Mouse, CD-1 ($)
•
•
Ramhoj, 2018, 4442260
Maternal Body Weight Change, GD 7-21
P0 Rat, Wistar (2)
Maternal Body Weight Change, PND 1-14
P0 Rat, Wistar (2)
Marques, 2021, 9960182
Maternal Body Weight
P0 Mouse, CD-1 ($)
•
Pregnancy Outcomes
Chang, 2018,4409324
Number of Pups Born to Implant Ratio
F1 Mouse, CD-1 (c?°)
~
•
Ramhoj, 2018, 4442260
Perinatal Loss
P0 Rat, Wistar (2)
*—
•
Viable Litters
P0 Rat, Wistar (2)
#—
Fetal Survival
Butenhoff, 2009, 1405789
Number of Pups Delivered
F1 Rat, Sprague-Dawley {6A 9)
~—
—•—
—~
Stillborn Pups
F1 Rat, Sprague-Dawley (o 9)
~—
—•
Viability index
F1 Rat, Sprague-Dawley (<59)
~—
—~
Lactation index
F1 Rat, Sprague-Dawley (0$)
~—
—~
Ramhoj, 2018, 4442260
Postimplantation Loss
F1 Rat, Wistar 0$)
P0 Rat, Wistar (9)
+
—•
Offspring Viability
Butenhoff, 2009, 1405789
Liveborn Pups
F1 Rat, Sprague-Dawley 09)
~—
•
—~
Chang, 2018, 4409324
Number of Pups Born per Litter
F1 Mouse, CD-1 0f)
~—
V
V
Litter Size
Chang, 2018, 4409324
Live Litter Size
F1 Mouse, CD-1 09)
V
V
Ramhoj, 2018, 4442260
Litter size (Live Pups PND1)
F1 Rat, Wistar 09)
~
—~
Marques, 2021, 9960182
Mean Litter Size, Liveborn
P0 Mouse, CD-1 ($)
•
0.01 0.1
1
10
100
mg/kg-day
Figure 3-54. PFHxS-induced developmental effects. Figure displays the high and medium confidence toxicological
studies included in the analysis. For additional details see HAWC link. Details on study confidence may be found in Figure
3-53. Note: while some of the decreases in maternal body weight were statistically significant, these small changes are of
unclear biological significance and not necessarily adverse.
3-190
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Mechanistic evidence
Thyroid hormones play a critical role in pregnancy and gestational development fZoeller
and Rovet. 2004: Y etal.. 2024: Street etal.. 2024: Stagnaro-Green and Rovet. 20161. The available
toxicological studies include evaluations of the potential impact of PFHxS exposure on thyroid
hormones during gestational development See Section 3.2.1 for a synthesis of the available
mechanistic evidence.
Evidence Integration
The currently available evidence suggests but is not sufficient to infer that PFHxS might
cause developmental effects in humans given sufficient exposure conditions.12 This judgment is
based on slight human evidence, specifically the reasonably consistent, but notably uncertain,
evidence of decreased birth weight and some coherent changes in other growth parameters from
studies of exposed humans in which PFHxS was measured preconception or either during or
shortly after pregnancy (see Table 3-21). As discussed earlier (see Appendix C for more details),
with the exception of postpartum samples which have larger deficits, consistent small (and
generally statistically significant) birth weight deficits were detected in EPA's meta-analysis of
epidemiological studies including those based on early sample timing. Overall, although there are
data that suggest changes in fetal growth are related to PFHxS exposures, additional evidence (e.g.,
more epidemiological study of PFHxS exposure on birth weight with earlier biomarker sampling
that helps to reduce uncertainties in the current evidence base) would be needed to draw a
stronger judgment.
Although not entirely consistent within and across studies, the epidemiological evidence
includes a large fetal growth restriction database with some of the most accurate endpoints
available (e.g., birth weight is generally measured with little error). The available epidemiologic
studies showing birth weight-related differences for continuous exposure data ((3 range: -12 to -
76 gper each ln-unit increase) and categorical ((3 range: -25 to -109 g for the highest quantile
compared with the lowest quantile) showed results comparable in magnitude and provided some
support of a biologic gradient, albeit the categorical data to a lesser degree given lack of
monotonicity across quantiles in most studies. For example, many studies based on continuous
exposure data (per each increasing unit change in PFHxS) showed comparable birth weight-related
deficits ranges in either boys or girls ((3 range: -13 to -76 g) or in the overall population ((3 range:
-12 to -76 g). There also was some evidence of exposure-response relationships based on
categorical data in 2 of 16 epidemiological studies, although these were predominately driven by
sex-specific findings.
Taken together, some mean birth weight deficits of varying magnitude were detected in 17
of 31 studies included in the main developmental synthesis, including 14 of 27 (and 10 of 21
12The "sufficient exposure conditions" are more fully evaluated and defined for the identified health effects
through dose-response analysis in Section 5.
3-191
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
medium/high confidence) studies that examined associations in the overall population and 8 of 14
that reported mean birth weight deficits in either male or female neonates or both. According to
EPA's meta-analysis, similar birth weight deficits per ln-unit PFHxS increase were seen across all 27
studies ((3 = -7.9 g; 95% CI: -15.0, -0.7), 23 medium and high confidence studies ((3 = -8.1 g; 95%
CI: -15.4, -0.9), or for the 12 high confidence studies ((3 = -6.8 g; 95% CI: -16.3, 2.8). No gradient
was seen across confidence levels or by biomarker sample timing. Although limited by a small
sample size and considerable variation in results across studies, some deficits were detected for
five postpartum sampled studies ((3 = -28.3 g; 95% CI: -69.3,12.7) using umbilical cord samples or
maternal samples after birth; this may be reflective of bias due to pregnancy hemodynamic
changes. In contrast, 12 studies based on earlier pregnancy sampling periods (e.g., any first
trimester sampling) showed deficits ((3 = -7.3 g; 95% CI: -16.0,1.1) similar in magnitude to the
overall pooled estimate of all 27 studies and those restricted to medium and high confidence. Given
that these patterns are not consistent with what EPA has seen for other PFAS such as PFNA (Wright
etal.. 2023) and what others have reported for PFOA and PFOS Steenland etal. (2018): Dzierlenga
etal. (2020). it remains unclear whether any differences noted between late pregnancy and
postpartum samples are unique to PFHxS.
Examining birth weight differences in human populations is challenging, and it can be
difficult to differentiate pathological deficits versus natural biological variation in distributions
within study populations. The magnitude of birth weight deficits across categorical and continuous
exposures in the individual studies, for example, ranged from -12 to -109 g depending on the
exposure contrasts being compared. The meta-analysis of the 27 studies that EPA conducted
showed a small but statistically significant decrease in mean birth weight ((3 = -7.9 g; 95% CI:
-15.0, -0.7) per ln-unit increase in PFHxS. This overall result was similar when studies were
restricted to just the 12 high ((3 = -6.8 g; 95% CI: -16.3, 2.8) confidence studies or the 23 combined
medium and high confidence studies ((3 = -8.1 g; 95% CI: -15.4, -0.9). The public health significance
of small changes in birth weight noted here in this meta-analysis may not be immediately evident.
On a population level, even small changes, if causally related, can increase the number of infants at
higher risk for other co-morbidities and mortality especially during the first year of life. And,
therefore, small decrements may have a large public health impact if these shift the birth weight
distribution to include more infants in the low-birth-weight category. Additionally, decreased birth
weight has been associated with long-term adverse health outcomes such as cardiovascular disease
and diabetes fOsmond and Barker. 20001. It is recognized that variations in mean birth weight may
not be clinically relevant at the individual neonate level and that different endpoints can include a
combination of pathologically and constitutionally small infants. Small changes in mean birth
weight, a proxy for growth, may however have a public health impact if they shift the whole
distribution of birth weight to include more infants in the low-birth-weight category. This has
potential population-level ramifications due to ubiquitous PFHxS exposures and given that low
birth weight infancy is associated with higher risk for co-morbidities and mortality, especially
3-192
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
during the first year of life, and can increase risk for adverse health outcomes later in life (Osmond
and Barker. 2000: De Boo and Harding. 20061. Low birth weight endpoint is a clinically recognized
endpoint that is standardized across populations even if there is some uncertainty related to the
underlying mechanisms of different developmental endpoints examined here. The consideration of
various developmental toxicological and epidemiological measures examined here does provide
some supporting evidence of the relevance of birth weight measures. Thus, while some deficits
reported in epidemiological studies and in our meta-analyses may be relatively small in magnitude,
the associations detected in the meta-analysis would be even larger if extrapolated across the
exposure distributions reported in some of these studies. Thus, this magnitude of decrease is
considered to be adverse and of concern.
Providing some evidence for changes coherent with the observed birth weight decreases,
decreases, 5 of 7 small for gestational age and low birth weight studies showed increased risk in
relation to PFHxS exposures. Additional evidence was seen in 12 of 18 (including 9 of 16 in the
overall population) birth length studies that showed associations of smaller birth length with
increasing PFHxS exposures, including 5 of 6 available high confidence studies. These results were
small in magnitude. In addition, there was some support for these findings from coherent effects
related to postnatal weight measures (as 5 of 8 studies showed inverse associations), albeit the
other postnatal growth endpoints were null or mixed.
In addition to the uncertainty related to potential bias from pregnancy hemodynamics in
developmental epidemiological studies, a common area of concern when interpreting
epidemiological findings on individual PFAS is the potential for confounding by PFAS co-exposures.
As noted for other endpoints in general, despite extensive and advanced statistical modeling
attempts, it can be difficult at times to completely isolate an independent effect for each individual
PFAS when real-world exposures involve a myriad of sources. Although there were some moderate
to strong positive correlations between PFHxS and some other PFAS, there were no consistent
patterns in magnitude of effects detected in models that adjusted for other PFAS (see detailed
write-up in Appendix C). Thus, while confounding by other PFAS remains a general source of
uncertainty in epidemiological studies, the lack of a consistent patterns across the available studies
here does not provide strong evidence of this possibility.
The available evidence on PFHxS-induced developmental effects in animal toxicity studies is
considered slight. The available animal studies do not provide evidence coherent with the
epidemiological observations of effects on fetal growth (i.e., rodent offspring body weights were
generally unaffected). Similarly, PFHxS exposure during early developmental stages did not impact
the incidence of developmental malformations or alter reproductive organ development However,
two studies using CD-I mice reported reduced fetal viability (Yao etal.. 2023b) and reduce fetal
weight and length (Zhang etal.. 2023). and one high confidence study reported a significant
decrease in litter size and numbers of pups per litter in CD-1 mice that was not dose-dependent
f Chang etal.. 20181 (note: a single, low confidence epidemiological study evaluating an outcome
3-193
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
related to fetal survival showed a marginally statistically significant increased odds of fetal loss
with increasing PFHxS exposure). However, f Chang etal.. 20181 also reported that the number of
pups born-to-implant ratio was unaffected, and two separate high and medium confidence studies
in rats reported no significant treatment-related effects on fetal survival endpoints at the same or
higher PFHxS levels fRamhai etal.. 2018: Butenhoff etal.. 2009: 3M. 20031. Chemical-induced
reduction in litter size can provide an indirect indication of preimplantation loss (IPCS. 2006):
however, this was not evaluated in any of the available gestational PFHxS exposure studies in
animals, highlighting a significant data gap.
Several epidemiological and animal toxicity studies report alterations in AGD. However, the
biological significance of the small and directionally inconsistent changes as well as lack of
consistency with other markers of androgen-dependentphenotypical outcomes and developmental
measures adds uncertainty to the available evidence. Overall, the available studies do not support
an effect on reproductive organ development by PFHxS exposure.
Overall, the available evidence suggests but is not sufficient to infer that PFHxS exposure
may have the potential to cause developmental toxicity in humans given sufficient exposure
conditions.13 A stronger evidence integration judgment was not drawn due to some important
sources of uncertainty in the epidemiological literature (most notably, uncertainty due to potential
bias by pregnancy hemodynamics) that appear to reflect complex patterns of biological influence
that are not completely understood. Nonetheless, the consistent and coherent epidemiological
findings on fetal growth restriction warrant further examination to disentangle these uncertainties
and improve understanding of whether and to what extent PFHxS exposure during these sensitive
lifestages might contribute to growth restriction in children.
13Given the uncertainty in this judgment and the available evidence, this assessment does not derive a toxicity
value that might better define the "sufficient exposure conditions" for developing this outcome (see Section 5
discussion).
3-194
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-21. Evidence profile table for PFHxS-related developmental effects
Evidence integration summary
Evidence stream summary and interpretation
judgment
Evidence from studies of exposed humans (see Development Human Section)
©OO
Evidence from human studies-fetal growth restriction
Evidence suggests, but is not
Studies and
Factors that increase
Factors that
Summary and key
Evidence stream
sufficient to infer
confidence
certainty
decrease certainty
findings
judgment
Fetal growth (Mean
• Consistent
• Imprecision of
• 20 of 34 overall
Primary basis: Consistent human
birth weight
findings of some
some birth
birth weight
©oo
evidence of decreased birth
/z scores/small for
inverse
weight deficits
studies
Slight
weight and coherent findings
gestation age/low
associations in
• Concern for
(including 14 of
across multiple other fetal and
birth weight)
20 of 34
potential
26 medium or
Based primarily on
early-life measures of growth.
(including 14 of
confounding by
high confidence)
consistent evidence
Median PFHxS values spanned
9 high, 1 medium,
26 high or
co-exposures to
studies showed
for birth weight
from 0.09 to 10.36 ng/mL across
and 5 low confidence
medium
highly correlated
inverse
reductions and
the birth weight meta-analysis
studies
confidence)
PFAS
associations in
coherent findings for
studies.
studies
• Exposure-
the overall
other fetal and
• Inverse
dependence
population, or
postnatal weight
Human relevance: N/A (based on
associations in
limited, including
among boys or
endpoints, but
human evidence)
17 of 31 mean
monotonic
girls
strength was reduced
Cross-stream coherence: N/A
birth weight
relationships, in
• Meta-analysis
due to concern for
studies and 14 (5
only 3 of 14
conducted by US
confounding and
(animal evidence indeterminate)
high; 5 medium;
different birth
EPA showed a
limited evidence of
4 low) of 27 in
weight studies
small but
dose-dependence
Susceptible populations and
overall
with categorical
statistically
across most studies
lifestages: Pregnancy and early
population
data in overall
significant birth
with categorical
life
across all study
population or
weight deficit
data.
confidence levels
either sex; lends
(-7.9 g; 95% CI:
• Although they
limited support
-15.0, -0.7) per
varied across
to studies based
each In-unit
confidence
on continuous
PFHxS increase;
levels, some
exposure metrics
results were
reported mean
comparable in
3-195
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration summary
judgment
birth weight
• Some concern
magnitude
deficits (over
over pregnancy
across early
-100 g per In-
hemodynamic
sampled studies
unit
impacts on
and high and
increase) and
birthweight
medium
relative risks
finding as 9 of 14
confidence
were fairly large
studies in the
studies
in magnitude
overall
Statistically
population were
significant meta-
based on late
analysis results
biomarker
for mean birth
sampling
weight from
continuous
exposure metrics
(-7.9 g; 95% CI:
-15.0, -0.7 per
each In-unit
increase); this
was comparable
to high (-6.8 g)
and medium
(-10.0 g)
confidence
studies
Overall meta-
analysis birth
weight results
(-7.9 g)
comparable to
early pregnancy
(-7.6 g) studies;
suggests results
not likely due to
3-196
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration summary
judgment
pregnancy
hemodynamics
• Evidence among
6 of 13
standardized
birth weight
studies primarily
seen in high (4 of
8 high and
medium (1 of 3)
confidence
studies
• 5 of 7 studies
examining either
small for
gestational age,
low birth weight
or very low birth
weight showed
some increased
risks with
increasing PFHxS
exposures
among the
overall
population or
either girls or
boys (quite
variable in
magnitude, OR
range: 1.3-9.1)
Fetal growth
restriction (birth
length)
• Consistent
findings of some
inverse
• None of the 5
studies with
categorical data
• 9 of 16 studies
reported
adverse effects,
3-197
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration summary
judgment
6 high, 5 medium,
and 7 low confidence
studies
associations in 9
of the 16 studies
in the overall
population (5
high, 1 medium,
and 3 low
confidence)
showed dose-
dependent
associations in
the overall
population
although 2 of 3
sex-specific
analyses did
(both from same
birth cohort).
Concern for
potential
confounding by
co-exposures to
highly correlated
PFAS
Some concern
for potential bias
due to sample
timing
(pregnancy
hemodynamics)
as 6 of 9 studies
with inverse
associations
were based on
later biomarker
sampling;
although this did
not bear out in
the sex-specific
analyses.
including all 5 of
6 high and 1 of 5
medium
confidence
studies
3-198
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration summary
judgment
Fetal growth
• 8 of 14 studies in
• Concern for
• 8 of 14 studies (5
restriction (head
total showed
potential
high; 2 medium
circumference)
inverse
confounding by
and 1 low
5 high,
associations,
co-exposures to
confidence)
5 medium, and 4 low
including 7 of 12
highly correlated
reported
confidence studies
studies in the
PFAS
adverse
overall
associations,
population (4 of
including 4 of 5
5 high; 2 of 4
high confidence
medium and 1 of
studies
3 low
confidence)
• Exposure-
dependence in 1
of 2 studies with
categorical data
• Limited concern
over pregnancy
hemodynamics
as 5 of 7 studies
with inverse
associations in
the overall
population were
based on early
biomarker
sampling
Anogenital distance
(AGD)
4 medium confidence
studies
• No factors
noted
• No factors
noted
• Inverse
association
between PFHxS
exposure and
AGD in 1 of 4
medium
confidence
3-199
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence integration summary
Evidence stream summary and interpretation
judgment
studies in boys
and in 1 of 3
studies in girls
Evidence from human studies postnatal growth
Studies and
Factors that increase
Factors that
Summary and key
Evidence stream
confidence
certainty
decrease certainty
findings
judgment
Postnatal growth-
• Consistent
• Inconsistent
• 5 of 8 studies
Weight measures:
findings of
periods of
showed some
5 high, 3 medium,
inverse
follow-up and
evidence of
and 3 low confidence
associations
assessment (e.g.,
postnatal weight
studies
across 5 of the 8
childhood age at
reductions
studies of infant
examination)
which showed
weight with
precludes more
some coherence
more evidence
direct
with birth
among girls
comparison
weight deficits.
• Mixed results
across studies.
• The other
were seen
• Concern for
endpoints were
among four
potential
mixed or
studies of rapid
confounding by
provided limited
growth (2 of 4
co-exposures to
or no evidence
studies).
highly correlated
of associations.
• Limited to no
PFAS
evidence of
associations for
postnatal height
(1 of 5 studies),
head
circumference (0
of 3 studies) in
overall
population or
either sex.
3-200
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration summary
judgment
• No evidence of
associations with
adiposity (0 of 5
studies) in the
overall
population, but 2
of 3 studies did
report this for
boys.
Evidence from human studies-gestational duration
Studies and
confidence
Factors that increase
certainty
Factors that
decrease certainty
Summary and key
findings
Evidence stream
judgment
Preterm birth
6 high and 4 medium
confidence studies
• All 10 published
studies were high
or medium
confidence
• Unexplained
inconsistency
• Concern for
potential
confounding by
co-exposures to
highly
correlated
PFAS
• 4 of 10 studies
showed some
evidence of
adverse
associations
Gestational age
8 high, 5 medium,
and 6 low confidence
studies
• 4 of the 7 studies
were based on
early biomarker
sampling;
suggesting that
pregnancy
hemodynamics
may have less of
an impact in this
subset.
• Unexplained
inconsistency
• One-half of the
studies in boys
were deficient
in study
sensitivity
• Concern for
potential
confounding by
co-exposures to
• 8 of 19 studies in
total as well as 7
(3 high, 3
medium, and 1
low confidence)
of 17 studies in
the overall
population
showed some
gestational age
reductions
3-201
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration summary
judgment
• There was a
preponderance of
associations
among girls with 2
of 3 of the studies
with categorical
data showing
some exposure-
response
relationship.
highly
correlated
PFAS
• 5 of the 8 sex-
specific studies
reported
associations in
girls, while none
of the studies in
the boys did.
Spontaneous
abortion
4 medium and 1 low
confidence study
• No factors noted
• Low confidence
study reporting
an effect
• 1 low confidence
reported a
positive
association
despite bias
toward null, but
4 medium
confidence
studies reported
no associations.
Evidence from in vivo animal studies (see Developmental Animal Section)
Studies and
confidence
Factors that increase
certainty
Factors that
decrease certainty
Summary and key
findings
Evidence stream
judgment
Maternal health,
fetal viabilitv. fetal
growth,
morphological
development
2 high confidence
studies:
• GD 0-PND 22
• High confidence
studies
• Unclear
biological
significance of
small maternal
weight
changes
• Lack of
expected dose-
dependence for
• Decreased litter
size in 1 of 3
studies
• Increased fetal
death in 1 of 3
studies
• No notable
PFHxS-induced
effects on
©oo
Slight
Based evidence for
decreased litter size
and increased fetal
death, but strength
was reduced due to
concern for
3-202
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration summary
judgment
3 medium confidence
study:
• GD 7-PND 22
• GD 1-GD 19
litter size
decrease in 1
study
maternal health,
fetal viability,
fetal growth, and
gestation
duration.
• Studies did not
evaluate
preimplantation
loss
inconsistent findings
across studies using
same animal models
and similar
experimental design.
3-203
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3.2.4. Hepatic Effects
Human Studies
Nineteen epidemiology studies (reported in 21 publications) report on the relationship
between PFHxS exposure and liver effects, primarily serum liver enzymes. 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 etal..
2005). Alkaline phosphatase (ALP), bilirubin, and y-glutamyltransferase (GGT) are also routinely
used to evaluate potential hepatic toxicity (Hall etal.. 2012: EMEA. 2008: Boone etal.. 2005). but
may also indicate alterations of gall bladder, bile duct, bone disease and pancreatic health, and thus
are less specific to hepatic function than ALT and AST. Elevation of liver serum biomarkers is
frequently an indication of liver injury, although they are not as specific as functional tests, which
are currently not available for PFHxS.
Serum markers of hepatic injury
The available studies evaluated serum measures of clinical markers which inform of
potential liver damage. These include circulating aminotransferases alanine aminotransferase
(ALT) and aspartate aminotransferase (AST) which are markers of hepatocellular function/injury
(Whalan. 2015: Wang etal.. 2014b: Lala etal.. 2023). Measurements of circulating alkaline
phosphatase (ALP), bile salts/acids, and bilirubin are routinely used by clinicians to evaluate
hepatobiliary toxicity (Lala etal.. 2023). AST alterations can indicate mitochondrial and
cytoplasmic injury in liver cells (Whalan. 2015). While ALT can be altered after accumulation of
elevated fatty acid in liver cells resulting in displacement of the cytoplasm fWhalan. 2015: Amacher.
20021. ALP is produced in liver, but also in bone and intestines fWhalan. 20151 and conditions
other than liver injury (e.g., bone disease) are associated with increased ALP fYangetal.. 20141.
Increases in serum ALP are indicative of a disruption in bile flow (i.e., cholestasis) and osteoclast
activity (Yang etal.. 2014: Whalan. 2015). Changes in albumin and total protein may be indicative of
chronic liver disease, as well as damage to other organs such as kidney, pancreas, thyroid, and
gastro intestinal tract (Whalan. 2015). Lastly, increased bilirubin may be indicative of bile acid
obstruction (cholestatic injury) and hepatocellular damage fWhalan. 2015: Amacher. 20021.
Of the 17 available epidemiology studies of liver enzymes, 13 were classified as medium
confidence, three as low confidence, and one was considered uninformative (see Figure 3-55). liang
etal. (2014) was considered uninformative due to critical deficiency in the confounding domain as
well as a lack of information on participant selection (deficient) and was excluded from further
analysis. The majority of the available studies were cross-sectional studies in adults, four of which
(Omoike et al.. 2020: Lin etal.. 2010: Tain and Ducatman. 2019e: Gleason etal.. 20151 were analyses
of different NHANES study populations (1999-2004, 2007-2010, 2011-2014, 2005-2012
respectively). The author defined inclusion criteria in these NHANES studies varied across analyses
3-204
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
(e.g., Gleasonetal. (2015) included adolescents as well as adults, fasting was required in Lin et al.
f20101. individuals who were carriers of hepatitis B or C virus were not excluded in Tain and
Ducatman f2019ell. Because of the overlapping population in Omoike etal. f20201 with the
previous studies, this paper was not considered a separate study. The remaining cross-sectional
studies were in populations in Canada fCakmaketal.. 20221. Korea fKim et al.. 20231. residents
near a fluoropolymer plant (Yao etal.. 2020). pregnant women (Liao etal.. 2023). primarily
government employees in China (Nian etal.. 2019: Liu etal.. 2022a). and firefighters in Australia
(Nilsson et al.. 2022b). Cross-sectional studies were considered appropriate for liver enzymes as
there is no expectation of reverse causation; additionally, the long half-life of PFHxS increases the
likelihood of the current exposure being representative of an etiologically relevant period. These
studies were all considered medium confidence for liver enzymes, except Yao etal. f20201. which
had concerns for selection bias and confounding.
In addition, there was a cohort of elderly adults (Salihovic etal.. 2018) and a birth cohort
with follow-up into childhood (Mora etal.. 2018). In children and adolescents, in addition to the
NHANES 2007-2010 analysis in Gleasonetal. (2015) that included adolescents but did not provide
stratified estimates, Attanasio f2019bl examined NHANES data from 2013 to 2016 in adolescents. A
multicenter birth cohort examined liver enzymes in childhood and was considered medium
confidence fStratakis etal.. 20201. There was also a low confidence study of children. Khalil et al.
(2018) was a pilot cross-sectional study of 48 obese children, and there was concern for potential
for selection bias and confounding. Across the studies of liver function, liver enzymes were
analyzed appropriately in serum. Analysis of PFHxS in serum or plasma samples was also
appropriate in all studies.
3-205
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Attanasio, 2019, 5412069 -
++
+
++
+
+
+
+
+
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Cakmak, 2022, 10273369-
++
+
++
+
+
+
+
+
~
E, 2023, 10699595-
++
-
-
+
+
+
-
Gleason, 2015, 2966740-
D
Critically deficient (metric) or Uninformative (overall)
Jain and Ducatman, 2019, 5080621 -
++
+
+
+
+
+
+
+
Jiang, 2014, 2850910-
-
+
+
B
+
-
Jin, 2020, 6315720-
+
+
++
-
+
+
¦
Khalil, 2018, 4238547-
¦
+
++
¦
+
+
¦
Kim, 2023, 10754695-
+
+
+
+
+
+
+
+
Liao, 2023, 10754689-
-
+
+
+
++
+
+
Lin, 2010, 1291111 -
++
+
+
+
+
+
+
+
Liu, 2022, 10273407-
+
+
+
+
++
+
+
+
Mora, 2018, 4239224-
+
+
++
+
++
+
+
+
Nian, 2019, 5080307-
+
+
+
+
++
+
+
+
Nilsson, 2022, 10587058-
+
+
+
+
+
+
+
+
Rantakokko, 2015, 3351439-
+
++
+
+
+
-
-
Salihovic, 2018, 5083555-
+
+
++
+
+
+
+
+
Stratakis, 2020, 6833620 -
+
+
++
+
++
+
+
Yao, 2020, 7021874-
-
+
+
-
+
+
+
-
Figure 3-55. Hepatic effects human study evaluation heatmap. For additional
details see HAWC link. Multiple publications of the same study: Attanasio (2019b)
also includes Attanasio f2019al: Cakmak etal. f20221 also includes Borghese et
al. f20221
The results for the 12 medium confidence studies (10 in adults)are presented in Table 3-22.
Six studies reported small, but statistically significant, positive associations between serum ALT
and PFHxS exposure (Salihovic etal.. 2018: Liu et al.. 2022a: Kim etal.. 2023: lain and Ducatman.
2019e: Gleason etal.. 2015: Cakmak etal.. 20221. although in Tain and Ducatman f2019el. this was
observed only in obese participants. Lin etal. f 20101 and Nian etal. f20191 also reported positive
associations, but with imprecise estimates. The other two studies in adults fNilsson et al.. 2022b:
3-206
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Liao etal.. 2023) and two studies in children (Mora etal.. 2018: Attanasio. 2019bl found no
association with ALT.
For other enzymes, the direction of association varied across studies Gleason et al. f20151
and Liu etal. f2022al also reported significant positive associations with AST, ALP, and total
bilirubin and Kim etal. f20231 similarly found associations with AST and GGT but did not analyze
total bilirubin. Salihovic etal. (2018) reported a significant positive association with ALP but an
inverse association with total bilirubin. Liao etal. (2023) found a statistically significant positive
association with total bilirubin but no association with AST or GGT. Other studies reported
nonstatistically significant associations in both directions for different enzymes fNian etal.. 2019:
Cakmak etal.. 20221. The lack of consistent association with total bilirubin and ALP does not
decrease confidence in the ALT findings given that these endpoints are not specific to hepatic
toxicity fWhalan. 2015: Tamber etal.. 2023: Makris etal.. 20221. In adolescents, Attanasio f2019bl
reported positive associations with total bilirubin but no clear associations with other enzymes
analyzed continuously. There were positive associations (p > 0.05) in girls with elevated ALT, AST,
and GGT (dichotomous based on upper reference limits). The other medium confidence study in
children fStratakis etal.. 20201 did not report results for individual liver enzymes but defined liver
injury risk as having any liver enzyme concentration above the 90th percentile for the study
population. They found no association between liver injury risk and PFHxS exposure. The low
confidence study (Khalil etal.. 2018) also reported no association between PFHxS and liver
enzymes.
It is possible that the observed associations (primarily in adults) could be due to
confounding by co-occurring PFAS. In the studies that reported correlations across PFAS, the
correlations between PFHxS and PFOS, PFNA, and PFOA were moderate to high (generally around
0.6). Most of the studies did not perform multipollutant modeling but five studies did present
mixture results using various methods. In each study, the analyses were not designed to identify the
association for PFHxS with and without adjustment for other PFAS, but rather to examine the effect
of a mixture of PFAS. However, weights for each PFAS in the mixture provide an indication of which
PFAS(s) were most influential on the association with liver enzymes. PFHxS had the largest weight
only for bilirubin in one study (Liao etal.. 2023) but had the smallest positive weight for ALT and
GGT in that same study. In contrast, PFNA and PFOA had the greatest contributions in multiple
studies (Borghese etal., 2022, Stratakis etal., 2020, Kim etal., 2022). PFOS was the dominant
component to the combined effect in Liu et al. (2022), and the weight for PFHxS was considerably
lower. Overall, these results indicate a substantial concern for the PFHxS results to be confounded
by other PFAS. However, these analyses are not considered evidence that PFHxS does not have an
effect on liver enzymes, as the weights indicate only that PFHxS does not contribute much to the
models beyond what is contributed by other chemicals in the model.
3-207
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Liver disease
Four studies examine liver disease outcomes. Two cross-sectional studies of nonalcoholic
fatty liver disease (NAFLD) were both evaluated as low confidence due to concerns that exposure
measured concurrent with this chronic outcome does not represent an etiologically relevant period
(see Figure 3-55). A third study was also low confidence for self-reported "liver problems" due to
high potential for outcome misclassification and due to concurrent measurement of exposure
(Nilsson et al.. 2022b) (this study was medium confidence for liver enzymes). In children, Tin etal.
(2020b) examined participants with nonalcoholic fatty liver disease and analyzed the odds of
severe disease (nonalcoholic steatohepatitis) with PFHxS exposure. There were concerns of
confounding due to lack of adjustment for socioeconomic status and inclusion of BMI, which may lie
on the causal pathway.
Rantakokko etal. (2015) used histological findings from biopsies obtained during elective
gastric bypass operation and reported an inverse association with PFHxS exposure (OR 0.02, 95%
CI <0.01, 0.53 for 2-4 foci versus none per 200x field). Limei etal. (2023) using data from NHANES,
analyzed a surrogate for NAFLD that included several variables including liver enzymes, waist
circumference, insulin, and glucose (authors report that the area under the receiver operating
characteristic curve was 0.78 in predicting ultrasound-diagnosed NAFLD). This study reported a
positive association in women but not men, with the strongest association in postmenopausal
women (OR 2.50, 95% CI 1.29, 4.85 in quartile 4 versus quartile 1). Nilsson etal. (2022b) found no
association with self-reported liver problems (OR 0.97, 95% CI 0.72,1.30). In children with
nonalcoholic fatty liver disease, higher PFHxS exposure was associated with the presence of
nonalcoholic steatohepatitis (OR [95% CI]: 4.18 [1.64,10.7] per IQRincrease). Positive associations
were also observed with grade of steatosis (p > 0.05), lobular inflammation, portal inflammation,
ballooning (p > 0.05), and liver fibrosis flin etal.. 2020bl.
Summary of hepatic effects
Given the general consistency of direction of association for ALT across the majority of the
studies in adults, there is an indication that PFHxS exposure is associated with hepatic effects. One
low confidence study of liver histology in children flin etal.. 2020bl indicates an association
between PFHxS exposure and disease severity (i.e., nonalcoholic steatohepatitis), but these findings
should be interpreted with caution due to the potential for confounding and the nongeneralizable
study population. Studies of functional hepatic endpoints (e.g., liver disease) in adults are
inconsistent, so it is not clear whether the observed changes in liver enzymes in these studies
translate to clinical hepatic injury. However, abnormally increased serum ALT indicates impaired
liver functioning and even small increases can be predictive of liver disease (Valenti. 2021: U.S. EPA.
2022c: Park etal.. 20191. so these changes are considered adverse on their own. Changes in serum
lipids (see Section 3.2.6) and uric acid (see Section 3.2.10) were also observed. Given that
cholesterol is primarily metabolized in the liver and uric acid is associated with nonalcoholic fatty
3-208
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
liver disease, these changes may provide some coherence with the evidence of hepatotoxicity.
However, there is evidence that the associations with both ALT and serum lipids could be due to
confounding by other PFAS, which is a substantial source of uncertainty.
3-209
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-22. Associations between PFHxS and liver enzymes in medium confidence epidemiology studies
Reference
Population
Median exposure
(IQR) or as specified
Effect estimate
ALT
AST
ALP
GGT
Total bilirubin
Adults
Liao et al. (2023)
Cross-
sectional
analysis
within cohort
(2015-2019);
Canada, 420
pregnant
women
0.09 (0.05-0.14)
P (p value) for tertiles
vs. T1
T2: -2.35
(-5.34, 0.64)
T3: -0.8
(-3.82, 2.19)
T2: -0.97
(-3.16, 1.22)
T3: 1.20
(-0.99, 3.40)
NR
T2: -0.83
(-2.46, 0.89)
T3: -0.11
(-1.75, 1.53)
T2: 1.69 (0.71,
2.68)*
T3: 2.27 (1.28,
3.26)*
Nian et al.
(2019)
Cross-
sectional
(2015-2016);
China; 1,605
adults
0.7 (0.01-2.7)
% change (95% CI) for
In-unit change
0.2 (-0.8,1.2)
0.1 (-0.5,0.8)
-0.1
(-0.6,0.5)
0.4 (-0.6,1.4)
-0.3 (-1.0,0.5)
Liu et al. (2022a)
Cross-
sectional
(2018-2019);
China; 1,303
adults
0.9 (0.5-1.4)
% difference (95% CI)
vs. 25th percentile
50th: 7.69
(5.62,9.80)*
75th: 12.15
(7.66, 16.83)*
95th: 16.90
(7.86, 26.70)*
50th: 3.43
(2.11, 4.78)*
75th: 6.16
(3.32, 9.07)*
95th: 9.66
(3.95,15.68)*
50th: 0.90
(-0.22, 2.03)
75th: 0.88
(-1.46, 3.27)
95th: 0.44
(-4.10, 5.19)
50th: 5.65
(3.22, 8.14)*
75th: 9.01
(3.81,14.47)*
95th: 12.65
(2.30, 24.04)
50th: 3.05 (1.57,
4.55)*
75th: 6.44 (3.25,
9.72)*
95th: 11.40
(4.92, 18.28)*
Jain and
Ducatman
(2019e)
NHANES
cross-
sectional
(2011-2014),
U.S.; 2,883
adults
1.4
P (p-value) for log-
unit change
Nonobese
0.005 (0.8)
Obese
0.05 (<0.01)*
Nonobese
0.007 (0.6)
Obese
0.01 (0.4)
Nonobese
-0.005 (0.7)
Obese
0.006 (0.6)
Nonobese
0.008 (0.7)
Obese
0.03 (0.1)
Nonobese
0.002 (0.9)
Obese
0.04 (0.07)
3-210
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median exposure
(IQR) or as specified
Effect estimate
ALT
AST
ALP
GGT
Total bilirubin
Lin et al. (2010)
NHANES
cross-
sectional
(1999-2004),
U.S.; 2,216
adults
mean (SE)
1.7 (1.0) (women)
P (SE) for log-unit
increase
0.2 (0.5),
p = 0.7
NR
NR
0.0 (0.02),
p = 0.9
0.4 (0.2),
p = 0.06
Gleason et al.
(2015)
NHANES
cross-
sectional
(2007-2010),
U.S.; 4,333
adults
(12+ yr)
1.8(1.0-3.1)
P (95% CI) for In-unit
increase
0.02
(0.01,0.03)*
0.02
(0.01,0.03)*
0.02
(0.01,0.04)*
0.01
(-0.01,0.03)
0.03 (0.01,0.05)*
Cakmak et al.
(2022)
(Borghese et al.,
2022)
Cross-
sectional
(2007-2017);
Canada;
4,952 adults
Cycle 1: 2.2;
Cycle 2: 1.7;
Cycle: 1.0
% change (95% CI) for
GM change
1.7(0.2,3.3)*
-0.3 (-1.6,0.9)
-1.2 (-3.7,
1.3)
3.6 (-0.7, 8.0)
-0.8 (-4.8, 3.5)
1,404 adults
% change (95% CI) for
doubling
1.5 (-0.4, 3.4)
3.1(1.9,4.4)
5.9 (2.8, 9.1)
(normal
weight)
3.9(1.2,6.6)
3.2 (-2.9, 9.6)
Salihovic et al.
(2018)
Cohort
(2001-2014);
Sweden;
1,002 elderly
adults
2.1(1.6-3.4)
P (p-value) for In-unit
change
0.02
(0.0,0.03)*
NR
0.06
(0.02,0.09)*
0.03
(-0.01,0.07)
-1.0 (-1.3,-0.7)*
Kim et al. (2023)
Cross-
sectional
(2015-2017),
Korea; 1,404
adults
2.3 (1.4-3.5)
% change (95% CI) for
doubling
4.8(2.0,7.8)*
2.4(0.6, 4.3)*
NR
5.4(1.3,9.6)*
NR
3-211
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median exposure
(IQR) or as specified
Effect estimate
ALT
AST
ALP
GGT
Total bilirubin
Nilsson et al.
(2022b)
Cross-
sectional
(2013-2014),
Australia; 782
adult
firefighters
6.5 (1.8-22)
P (95% CI) for
doubling
0.01 (-0.01,
0.04)
NR
NR
NR
NR
P (95% CI) for
quartiles vs. Q1
Q2: 0.03
(-0.10, 0.15)
Q3: 0.08
(-0.07, 0.22)
Q4: 0.09
(-0.06, 0.23)
Children and adolescents
Mora et al.
Project Viva
prenatal
P (95% CI) for IQR
-0.1 (-0.4,0.2)
NR
NR
NR
NR
(2018)
birth cohort
2.4(1.6-3.8)
increase
(1999-2002),
U.S.; 682
children (7-
child
1.9 (1.2-3.4)
0.0 (-0.2,0.2)
NR
NR
NR
NR
8 yr)
Attanasio
NHANES
GM (SE)
P (95% CI) for
boys
boys
NR
boys
boys
(2019b)
cross-
male 1.3 (0.09)
quartiles vs. Q1
Q2: -0.07
Q2: -0.04
Q2: -0.09
Q2: 0.11
sectional
female 0.9 (0.06)
(-0.15,0.01)
(-0.10, 0.03)
(-0.21, 0.03)
(0.03, 0.20)
(2013-2016);
Q3: -0.09
Q3: -0.03
Q3: -0.03
Q3: 0.07
354 males
(-0.20, 0.02)
(-0.09, 0.04)
(-0.15, 0.09)
(-0.01, 0.15)
and 305
Q4: -0.02
Q4: 0.00
Q4: 0.02
Q4: 0.16
females (12—
(-0.12,0.08)
(-0.09, 0.09)
(-0.12, 0.15)
(0.07,0.26)
19 yr)
girls
Q2: -0.01
(-0.14,0.12)
Q3: 0.05
(-0.05,0.16)
Q4: 0.03
(-0.10, 0.16)
girls
Q2: 0.00
(-0.10, 0.10)
Q3: 0.07
(-0.01, 0.15)
Q4: 0.03
(-0.08, 0.14)
girls
Q2: 0.10
(-0.01, 0.20)
Q3: 0.10
(-0.01, 0.20)
Q4: 0.08
(-0.02, 0.18)
p-trend: 0.01
girls
Q2: 0.08
(-0.02, 0.18)
Q3: 0.19
(0.08, 0.30)
Q4: 0.25
(0.11, 0.40)
p-trend < 0.01*
*p < 0.05.
NR = not reported.
3-212
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Animal Studies
The toxicity database for PFHxS-induced liver effects in experimental animals consists of
two short-term exposure studies using SD rats fNTP. 2018a: 3M. 2000al: two subchronic exposure
study using AP0E*3-Leiden.CETP mice14 fBiiland etal.. 20111 or C57BL/6 mice fHe etal.. 20221:
one chronic exposure study using C57BL/6J mice (Pfohl etal.. 2020) and four multigeneration
studies using Wistar (Ramhaj etal.. 2018) or Sprague Dawley rats (Butenhoff etal.. 2009: 3M.
2003). or CD-I mice (Marques etal.. 2021: Chang etal.. 2018). All studies exposed animals orally via
either gavage fRamhai etal.. 2018: NTP. 2018a: Chang etal.. 2018: Bute nhoff etal.. 2009: 3M.
2000a. 2003) or the diet fBiiland etal.. 2011). Outcomes evaluated and reported in these studies
include histopathological effects, serum biomarkers of liver damage and lipid metabolism, and
changes in absolute and relative liver weights.
Organ weight
Four high confidence studies and five medium confidence studies evaluated PFHxS-induced
effects on liver weight (see Figure 3-56). In both rats and mice, short-term and subchronic exposure
led to increased absolute and relative liver weights15 fNTP. 2018a: Biiland etal.. 2011: 3M. 2000al
(see Figure 3-57). However, a chronic exposure study using male C57BL/6J mice reported no
significant effect on liver weight after exposure to 0.15 mg/kg-day for 29 weeks fPfohl etal.. 20201.
Two short-term (28-day) exposure studies using SD rats reported that exposure to PFHxS
increased liver weight by 8% to 54% at doses ranging from 1.25 to 10 mg/kg-day (NTP. 2018a: 3M.
2000a). Although NTP (2018a) observed increased relative and absolute liver weights in both male
and female rats, 3M (2000a) observed exposure-related changes in male rats only. A separate
subchronic exposure study using APOE*3-Leiden.CETP mice also observed increased absolute liver
weight (8%) in animals orally exposed to 6 mg/kg-day PFHxS for 42 days fBiiland etal.. 20111.
Four multigenerational toxicity studies evaluated PFHxS-induced effects on liver weights in
F0 and/or F1 animals (Ramhaj etal.. 2018: Chang etal.. 2018: Bute nhoff etal.. 2009: 3M. 2003). In
F0 generation male SD rats, exposure to 3 or 10 mg/kg-day PFHxS increased absolute and relative
liver weight by 20% to 67% when compared with controls, but no effects were observed in F0
females (Butenhoff et al.. 2009: 3M. 2003). Two similar studies using CD-I mice also measured liver
weights, but reported different effects: f Chang etal.. 20181 observed increased absolute and
relative liver weight (23% to 70%) in F0 generation (male and female) animals, whereas fMarques
etal.. 2021) reported no exposure-related changes in F0 female liver weights. Both (Chang etal..
2018) and (Marques etal.. 2021) exposed pregnant animals to similar doses of PFHxS, however
(Chang etal.. 2018) treated animals for 42 days before mating, through gestation and lactation
14APOE*3-Leiden.CETP mice is a genetically modified animal model which emulates human lipoprotein
profiles and is used to investigate cholesterol metabolism and cardiovascular disease (Veseli et al.. 20171.
15Alterations in liver weight are considered indicative of exposure-related responses such as enzyme
induction and hepatocellular hypertrophy (Thoolen et al.. 2010: Sellers et al.. 20071
3-213
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
whereas (Marques etal.. 20211 exposed F0 female animals from GD 1 to PND 20. In F1 generation
animals, significant PFHxS-induced increases in liver weight were observed in male CD-I mice
(10% increase in relative liver weight at 3 mg/kg-day) after exposure during gestation, lactation,
and post-weaning (until postnatal day 36) f Chang etal.. 20181. However, in F1 male and female SD
rats sampled on PND22 and Wistar rats sampled on PNDs 16-17, there were no significant
exposure-related changes in relative or absolute liver weights (Ramhaj etal.. 2018: Ramhaj etal..
2020: Bute nhoff etal.. 2009: 3M. 20031. In F1 male and female CD-I mice exposure to a high fat diet
plus PFHxS resulted in decreased relative, but not absolute, liver weight on PND 21. These effects
were not apparent on PND 90. Overall, the majority of the available studies report fairly consistent
increases in liver weight across lifestages following PFHxS exposure.
Reporting quality -I
Allocation A.
Observational bias/blinding A
Confounding/variable control A
Selective reporting and attrition A
Chemical administration and characterization A
Exposure timing, frequency and duration A
Results presentation A
Endpoint sensitivity and specificity -)
Overall confidence -I
+
+
+
NR
-
NR
NR
NR
NR
NR
NR
NR
NR
NR
+
-
+
NR
~~
-
Legend
I Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
|NR Not reported
Figure 3-56. PFHxS liver weight animal study evaluation heatmap. For
additional details see HAWC link.
3-214
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Eridpoint Name
Liver Weight, Absolute
Study Name
NTP, 2018, 4309363
Experiment Name
28 Day Oral
28 Day Oral
28 Day Oral
28 Day Oral
Liver Weight, Relative NTP, 2018, 4309363
Liver Weight, Absolute 3M, 2000, 3981194
Liver Weight, Relative 3M, 2000, 3981194
Liver Weight, Absolute Butenhoff, 2009, 1405789 Multi-Generational Oral
Liver Weight, Relative Butenhoff, 2009, 1405789 Multi-Generational Oral
Liver Weight, Absolute Butenhoff, 2009, 1405789 Multi-Generational Oral
Liver Weight, Relative Butenhoff, 2009, 1405789 Multi-Generational Oral
Liver Weight, Absolute Ramhej, 2018, 4442260 Multi-Generational Oral
Animal Description
Rat, Sprague-Dawley (.•?)
Rat, Sprague-Dawley (^)
Rat, Sprague-Dawley (o)
Rat, Sprague-Dawley (_)
Rat, CrlrCd Br(c?)
Rat, Crl:Cd Br<*)
Rat, CrlrCd Br(cf)
Rat, Crl:Cd Br(^)
P0 Rat, Sprague-Dawley (i )
P0 Rat, Sprague-Dawley (2)
P0 Rat, Sprague-Dawley ( ?)
P0 Rat, Sprague-Dawley (_)
F1 Rat, Sprague-Dawley (,-'')
F1 Rat, Sprague-Dawley (_)
F1 Rat, Sprague-Dawley ( -?)
F1 Rat, Sprague-Dawley (_)
P0 Rat, Wistar (2)
F1 Rat, Wistar (-5)
F1 Rat, Wistar (2)
Liver Weight, Relative Ramhoj, 2018, 4442260 Multi-Generational Oral F1 Rat, Wistar (o)
F1 Rat, Wistar (2)
Liver Weight, Absolute Chang, 2018, 4409324 Multi-Generational Oral P0 Mouse, CD-1 (.5)
P0 Mouse, CD-1 (£)
Liver Weight, Relative Chang, 2018, 4409324 Multi-Generational Oral P0 Mouse, CD-1 (•:?)
P0 Mouse, CD-1 (£)
Liver Weight, Absolute Chang, 2018, 4409324 Multi-Generational Oral F1 Mouse, CD-1 (
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Histopathologv
Histopathological lesions in the liver were reported in four high confidence studies using
Sprague Dawley rats fNTP. 2018a: Butenhoff etal.. 2009: 3M. 2000a. 20031 or mice fChang et al..
20181. two medium confidence study using Wistar rats fRamhai etal.. 20201 or CD-I mice, and one
low confidence study using C57BL/6 mice (He etal.. 2022) (see Figure 3-58).
Two short-term studies evaluated histopathological responses in male and female SD rats
after exposing animals to doses ranging from 2.5 to 10 mg/kg-day for 28 days, and one subchronic
study evaluated effects in male C57BL/6 mice treated with 60 |a,g/kg-day PFHxS for 12 weeks (NTP.
2018a: He etal.. 2022: 3M. 2000a). Statistically significant increases in the incidence of
hepatocellular hypertrophy16 (44% to 100%) were observed in male SD rats exposed to PFHxS at
doses > 2.5 mg/kg-day (NTP. 2018a). or 10 mg/k-day (3M. 2000a) (see Figure 3-59). (3M. 2000a)
also evaluated other histological responses (including hematopoietic cell foci, single cell necrosis,
coagulative necrosis, hepatocellular vacuolation, and inflammatory cell foci), but reported no
significant exposure-related effects. Both studies also report that female animals did not exhibit the
histopathological effects observed in male animals (NTP. 2018a: 3M. 2000a).
PFHxS-induced histopathological effects were also evaluated in two multigenerational
toxicity studies. In F0 generation male SD rats or male and female CD-I mice, exposure to PFHxS
caused increased incidence of histopathological effects (see Figure 3-60), primarily hepatocellular
hypertrophy. In the rat study, F0 generation animals exposed to PFHxS for 42 days to 3 or
10 mg/kg-day increased the incidence of hepatocellular hypertrophy by 90% and 100%, but other
histological responses (including focal necrosis, lipidosis, vacuolation [midzonal or multifocal], and
chronic liver inflammation) were not significantly affected in high confidence studies (Butenhoff et
al.. 2009: 3M. 2003). Similar observations were made in male F0 generation CD-I mice for which
exposure to 0.3,1, or 3 mg/kg-day PFHxS for 42 days increased hepatocellular hypertrophy and
cytoplasmic alterations by 80%, 100%, and 100%, respectively when compared with controls
fChang etal.. 20181. Furthermore, the incidence of single cell necrosis and microvesicular fatty
change were increased (40% and 60% respectively) at the highest dose, but hepatocellular cell
necrosis was not affected. Female F0 generation rats or mice used in the Butenhoff et al. (2009) and
Chang etal. (2018) studies were exposed to PFHxS for 14 days before cohabitation and continued
up to postnatal day 22. F0 generation female rats were nonresponsive to PFHxS exposure
f Butenhoff etal.. 2009: 3M. 20031. However, in F0 generation female CD-I mice cytoplasmic
vacuolation was increased by 30% at the highest dose (3 mg/kg-day) and hepatocellular
hypertrophy and cytoplasmic alterations (ground glass) were increased by 50% to 100% in all
treated animals, but these effects were not dose-dependent (Chang etal.. 2018). F1 generation CD-I
"Hepatocellular hypertrophy: a cellular response to chemical-induced stress that is considered indicative of
hepatomegaly PThoolen et al.. 2010: Cattlev and Cullen. 20181 and characterized by an increase of hepatocyte
size after exposure to xenobiotic agents (Maronpot. 20141 It may be caused by increases in mitochondria,
peroxisomes, endoplasmic reticulum, or metabolic enzyme induction (Thoolen et al.. 20101 and is often
accompanied by changes in organ weight (Cattlev and Cullen. 20181
3-216
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
mice exposed to 3 mg/kg-day PFHxS during gestation and lactation displayed statistically
significant increases in cytoplasmic alterations (63% incidence in males and 88% in females) and
hepatocellular hypertrophy (83% incidence in males and 88% in females) (see Figure 3-61), but the
incidence of hepatocellular necrosis, inflammation, and cytoplasmic vacuolation was not affected in
F1 male or female CD-I mice fChang etal.. 20181. A separate study using CD-I mice reported no
effect on male or female F1 animals exposed to 1 mg/kg-day PFHxS from GDI to PND20 (Marques
etal.. 2021). These varying responses in the two studies using CD-I mice (Marques etal.. 2021:
Chang etal.. 2018) could have been due to differences in experimental exposure durations: Chang
2018, 4409324@@author-year exposed animals before mating (14 days) and then during gestation
and lactation, whereas Marques etal. f20211 only exposed animals during gestation and lactation.
Furthermore, a separate study using Wistar rats reported no significant effects in F0 or F1 animals
exposed to PFHxS (0.05 to 25 mg/kg-day) from GD7 to PND22 fRamhai etal.. 20201.
One subchronic study evaluated effects in male C57BL/6 mice treated with 60 |a,g/kg-day
PFHxS for 12 weeks (He etal.. 2022). In male C57BL/6 mice given a high fat diet, exposure to 60
|a,g/kg-day for 12 weeks resulted in increased hepatocyte ballooning, inflammatory infiltration and
fibrosis (He etal.. 2022). These findings suggest that PFHxS may induce adverse histopathological
responses after prolonged (i.e., chronic or subchronic) exposures. However, these findings should
be interpreted with caution as several deficiencies were identified in He etal. (2022) including lack
of reporting of histopathological effect incidences, observational bias, and concerns with chemical
administration (see Figure 3-58, and follow HAWC link for additional details).
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
0
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)
NRl Not reported
Figure 3-58. Liver histopathology animal study evaluation heatmap. For
additional details see HAWC link.
3-217
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
Design
Animal Description
Dose observation time text
(mg/kg-day)
PFHxS Liver Histopathology
Hepatocellular Hypertrophy
NTP, 2018, 4309363 28 Day Oral Male Sprague Dawley (SD) Rat 0
10
Female Sprague Dawley (SD) Rat 0
3.12
6.25
12.5
3M. 2000, 3981194
Hepatocellular Vacuolatian
Hematopoietic Cell Foci
Single Cell Necrosis
Single cell necrosis
Coagulative Necrosis, Focal
Coagulative Necrosis, focal
Inflammatory Cell Foci
Liver Steatosis
Male Crl:Cd Br Rat
Female Crl:Cd Br Rat
3M, 2000, 3981194 28 Day Oral Male Crl:Cd Br Ral
Female Crl:Cd Br Rat
3M. 2000, 3981194 28 Day Oral Male Crl:Cd Br Rat
Female Crl:Cd Br Rat
3M. 2000, 3981194
3M, 2000,3981194
3M. 2000, 3981194
3M, 2000, 3981194
3M, 2000, 3981194
Marques, 2021, 9960182
28 Day Oral
28 Day Oral
28 Day Oral
28 Day Oral
28 Day Oral
Male Crl:Cd Br Rat
Female Crl:Cd Br Rat
Male Crl:Cd Br Rat
Female Cri:Cd Br Rat
Male Crl:Cd Br Rat
Female Crl:Cd Br Rat
SD P0 Female CD-1 Mice
HFD P0 Female CD-1 Mice
Day 29
Day 29
Day 29
Day 29
Day 29
Day 29
Day 29
Day 29
Day 29
Day 29
Day 29
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
Study Day 28
21 PND
21 PND
21 PND
21 PND
| Statistically significant
3 incidence
0 2 4 6 8 10 12 14 16 18 20 22 24 26
Figure 3-59. PFHxS liver histopathology observations from short-term animal
toxicology studies. Figure displays the high and medium confidence studies
included in the analysis (low confidence studies not shown). Details on study
confidence may be found in Figure 3-57. For additional details see the link to:
HAWC.
3-218
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study Name
Animal Description
Dose observation time text
(mg/kg-day)
Hepatocellular Hypertrophy Butenhoff, 2009,1405789 F0 Male Crl:CD (SD) Rat
Cytoplasmic Alteration, Ground-Glass Chang, 2018,4409324 F0 Male Crl:CDl (ICR) Mouse
F0 Female Crl:CD1 (ICR) Mouse 0
Fatty Change, Microvesicular Chang, 2018, 4409324 F0 Male Crl:CD1 (ICR) Mouse
0
0.3
F0 Female Crl:CD1 (ICR) Mouse 0
Butenhoff, 2009,1405789 F0 Male Crl:CD (SD) Rat
F0 Female Crl:CD (SD) Rat
Chang, 2018, 4409324 F0 Male Crl:CD1 (ICR) Mouse
F0 Female Crl:CD1 (ICR) Mouse 0
Marques, 2021, 9960182 SD P0 Female CD-1 Mice
HFD P0 Female CD-1 Mice
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 44
Day 44
Day 44
Day 44
Day 44
Day 44
Day 44
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
Day 42
21 PND
21 PND
21 PND
21 PND
PFHxS Liver Histopathology
0.3
Day 44
1 Day 44
3
Day 44
10
Day 44
F0 Female Crl.CD (SD) Rat
0
Day 44
10
Day 44
Chang, 2018, 4409324 F0 Male Crl:CD1 (ICR) Mouse
0
Day 42
0.3
1
Day 42
Day 42
3
Day 42
F0 Female Crl:CD1 (ICR) Mouse
0
Day 42
¦¦Statistically significant
I J incidence
10 12 14 16 18 20 22 24
incidence
Figure 3-60. PFHxS liver histopathology observations from developmental
animal toxicity studies (F0 generation animals). Figure displays the high and
medium confidence studies included in the analysis (low confidence studies not
shown). Details on study confidence may be found in Figure 3-57, For additional
details see link to: HAWC.
3-219
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Endpoint Name
Study Name
Animal Description
Dose
(mg/kg-day)
observation time text
PFHxS Liver Histopathology
Hepatocellular Hypertrophy,
Centrilobular
Chang, 2018, 4409324
F1 Male Crl:CD1 (ICR) Mouse
0
PND36
~
3
PND36
F1 Female Crl:CD1 (ICR) Mouse
0
PND36
3
PND36
Cytoplasmic Alteration
Chang,2018, 4409324
F1 Male Cr1:CD1 (ICR) Mouse
0
PND36
3
PND36
F1 Female Crl:CD1 (ICR) Mouse
0
PND36
3
PND36
Cytoplasmic Vacuolation
Chang,2018,4409324
F1 Male Crl:CD1 (ICR) Mouse
0
PND36
3
PND36
~
F1 Female Crl:CDl (ICR) Mouse
0
PND36
3
PND36
Necrosis
Chang,2018, 4409324
F1 Male Crl:CD1 (ICR) Mouse
0
PND36
3
PND36
~
F1 Female Crl:CD1 (ICR) Mouse
0
PND36
3
PND36
Inflammation, Mixed Cell
Chang,2018, 4409324
F1 Male Crl:CD1 (ICR) Mouse
0
PND36
3
PND36
~
F1 Female Crl:CD1 (ICR) Mouse
0
PND36
Statistically significant
¦ incidence
3
PND36
Liver Steatosis
Marques, 2021,9960182
SD P0 Female CD-1 Mice
0
21 PND
1 21 PND
HFD P0 Female CD-1 Mice
0
21 PND
I
1
21 PND
I
2 4 6 8 10 12 14
incidence
6 18 20 22 24
6
Figure 3-61. PFHxS liver histopathology observations from developmental
animal toxicity studies (F1 generation animals). Figure displays the high and
medium confidence studies included in the analysis (low confidence studies not
shown). Details on study confidence may be found in Figure 3-57. For additional
details see the link to: HAWC.
Serum biomarkers of liver function
Four high confidence studies and two medium confidence studies measured serum
biomarkers indicative of potential liver toxicity (see Figure 3-62). As in epidemiological studies,
serum measures of clinical markers which inform of potential liver damage in experimental studies:
circulating aminotransferases ALT and AST are markers of hepatocellular function/injury;
circulating ALP, bile salts/acids, and bilirubin are routinely used to evaluate hepatobiliary toxicity
fWhalan.2015: Hall etal.. 2012: FMFA. 2008: Boone etal.. 20051.
Two multigenerational toxicity studies report that exposure to 3 or 10 mg/kg PFHxS for 24
or 44 days statistically increased ALP in F0 generation male CD-I mice (133%) and SD rats (37%),
respectively (Chang etal.. 2018: Butenhoff et al.. 2009: 3M. 20031 (see Figure 3-63). Albumin was
also statistically increased (5%) in F0 male SD rats treated with the highest PFHxS dose (10 mg/kg-
day) fButenhoff et al.. 2009: 3M. 20031 and bilirubin was decreased by 60% in male F0 CD-I mice
treated with 3 mg/kg-day PFHxS for 42 days f Chang etal.. 20181. The study authors reported no
effects in females f Chang etal.. 2018: Butenhoff etal.. 2009: 3M. 20031. These apparent differences
in susceptibility between males and females may be attributable to the pharmacokinetics of PFHxS
in males versus females (see Section 3.1). A study using CD-I mice treated with 0, or 1 mg/kg-day
3-220
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
PFHxS during gestation (GD1-19) reported no effects on serum ALT in FO dams sampled on PND 21
or male or female F1 animals sampled on PND 5, 21, or 90 fMarques etal.. 20211.
Two short-term studies using SD rats evaluated serum levels of AST, ALT, ALP, and bilirubin
after exposure to doses ranging from 0.6 to 10 mg/kg-day PFHxS for 28 days fNTP. 2018a: 3M.
2000a). 3M f2000al reported that ALP was statistically increased by 20% in male SD rats exposed
to 10 mg/kg-day, but a similar study by NTP observed no exposure-related effects in the same rat
strain (NTP. 2018a). Serum levels of ALT or AST were not affected in male or female SD rats in
either study fNTP. 2018a: 3M. 2000a). Serum globulin levels were statistically decreased by 14% to
15% in male SD rats exposed to 10 mg/kg-day PFHxS for 28 days fNTP. 2018a: 3M. 2000al. and
bilirubin was significantly decreased by 12% to 21% in male SD rats after 28 days of exposure to
PFHxS at doses ranging from 2.5 to 10 mg/kg-day fNTP. 2018al. The 3M and NTP studies also
evaluated female animals and reported no exposure-related effects. NTP f2018al also evaluated
serum levels of albumin, bile salts/acids, and total protein in male and female SD rats and reported
no significant exposure-related effects.
A subchronic exposure high confidence study using peripubertal (5-week old) male
C57BL/6J reported a statistically significant 42% increase in ALT after exposure to 0.6 mg/kg-day
for 12 weeks fHe etal.. 20221. When compared with multigenerational and short-term studies, the
findings from He etal. f20221 suggests that PFHxS exposure may alter serum markers of liver
damage after prolonged exposure. However, the He etal. (2022) is the only available subchronic
study that reported on serum markers of liver disease and several concerns related with
experimental design and exposure methods were identified (see Figure 3-62).
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
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)
NRl Not reported
Figure 3-62. PFHxS liver serum biomarkers animal study evaluation heatmap.
For additional details see HAWC link.
3-221
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Endpoint Name
Study Name
Animal Description
Duration of exposure
PFHxS Liver/Hepatobiliary Serum Biomarkers
Alanine Aminotransferase (ALT)
NTP, 2018, 4309363
Rat, Sprague-Dawley (i)
28 Days
# No significant change
3M, 2000,3981194
Rat, Crl:Cd Br (¦-?)
28 Days
A Significant increase
•
Rat, Crl:Cd Br (?)
28 Days
V Significant decrease
•
Butenhoff, 2009, 1405789
PO Rat, Sprague-Dawley (-')
44 Days
• •—• •
P0 Rat, Sprague-Dawley ($)
PND22
• •—• •
Marques, 2021,9960182
P0 Mouse, CD-1 (J)
GD1-19
•
He,2022, 10273379
Mouse, C57BL/6 ( ;')
84 days
A
•
Aspartate Aminotransferase (AST)
NTP, 2018, 4309363
Rat, Sprague-Dawley (..-')
28 Days
Rat, Sprague-Dawley (y)
28 Days
3M, 2000, 3981194
Rat, Crl:Cd Br (;?)
28 Days
V
Rat, Crl:Cd Br (i>)
28 Days
•
Butenhoff, 2009, 1405789
P0 Rat, Sprague-Dawley (-}')
44 Days
• •—• •
P0 Rat, Sprague-Dawley ( + )
PND22
• •—• •
Alkaline Phosphatase (ALP)
NTP 2018, 4309363
Rat, Sprague-Dawley (rf1)
28 Days
Rat, Sprague-Dawley (9)
28 Days
3M, 2000, 3981194
Rat, CrtCd Br (;>)
28 Days
A
Rat, Crl:Cd Br ($)
28 Days
•
Butenhoff, 2009, 1405789
P0 Rat, Sprague-Dawley (;)
44 Days
•' •—• A
P0 Rat, Sprague-Dawley (V)
PND22
• •—• •
Chang,2018, 4409324
P0 Mouse. CD-1 (•->)
42 Days
• • A
Bile Salt/Acids
NTP, 2018, 4309363
Rat, Sprague-Dawley (. ; )
28 Days
Rat, Sprague-Dawley ($)
28 Days
Direct Bilirubin
NTP, 2018, 4309363
Rat, Sprague-Dawley (;)
28 Days
Rat, Sprague-Dawley ('*)
28 Days
Total Bilirubin
NTP, 2018, 4309363
Rat, Sprague-Dawley (.?')
28 Days
Rat, Sprague-Dawley (9)
28 Days
3M, 2000, 3981194
Rat, Crl:Cd Br (;?)
28 Days
•
Rat, Crl:Cd Br (i)
28 Days
•
Butenhoff, 2009, 1405789
P0 Rat, Sprague-Dawley (-/)
44 Days
• • • •
P0 Rat, Sprague-Dawley (v)
PND22
• • • •
Chang,2018, 4409324
PO Mouse, CD-1 (;>)
42 Days
• •—W
Triglyceride (TRIG)
Bijland, 2011, 1578502
Mouse, Apoe*3-Leiden.Cetp ( '
42 Days
A
Albumin
NTP, 2018, 4309363
Rat, Sprague-Dawley (•;')
28 Days
Rat, Sprague-Dawley {_)
28 Days
3M, 2000, 3981194
Rat, Crl:Cd Br (5)
28 Days
•
Butenhoff, 2009, 1405789
P0 Rat, Sprague-Dawley (r?)
44 Days
• •—• A
Globulin
NTP, 2018, 4309363
Rat, Sprague-Dawley ( •)
28 Days
Rat, Sprague-Dawley (£)
28 Days
3M, 2000, 3981194
Rat, Crl:Cd Br (?)
28 Days
V
Rat, Crl:Cd Br (V)
28 Days
•
Albumin/Globulin (A/G) Ratio
NTP, 2018, 4309363
Rat, Sprague-Dawley
28 Days
Rat, Sprague-Dawley (9)
28 Days
3M, 2000, 3981194
Rat, Crl:Cd Br (-"')
28 Days
A
Butenhoff, 2009, 1405789
P0 Rat, Sprague-Dawley (o)
44 Days
•—-•«—A
Total Protein
NTP, 2018, 4309363
Rat, Sprague-Dawley (¦£')
28 Days
Rat, Sprague-Dawley (V)
28 Days
Liver Cholesteryl Esters
Bijland, 2011, 1578502
Mouse, Apoe*3-Leiden.Cetp (
42 Days
•
Liver Free Cholesterol
Bijland, 2011, 1578502
Mouse, Apoe*3-Leiden.Cetp (:'
42 Days
A
Liver Triglycerides
Bijland, 2011, 1578502
Mouse, Apoe*3-Leiden.Cetp (;?
42 Days
A
i i
0.01 0.1
1 1 1
1 10 100
mg/kg-day
Figure 3-63. PFHxS liver/hepatobiliary serum biomarkers. Figure displays the
high and medium confidence studies included in the analysis (see Figure 3-62). For
additional details see HAWC link.
Hepatic lipid content
PFHxS-induced alterations in hepatic lipid levels were evaluated in one high confidence
study f Chang etal.. 2018) and four medium confidence studies (Pfohl etal.. 2020: Marques etal..
3-222
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
2021: He etal.. 2022: Bijland etal.. 20111 (see Figure 3-64). All available studies used different
strains of mice fPfohl etal.. 2020: Marques etal.. 2021: He etal.. 2022: Chang etal.. 2018: Biiland et
al.. 20111 Several studies also evaluated serum levels of lipids and cholesterol. These are discussed
in the Cardiometabolic Effects section (see Section 3.2.6) as increases in serum lipids and
cholesterol are considered risk factors for cardiovascular disease f Zhang et al.. 2022a: Y etal.. 2022:
Wang etal.. 2024: Pappan etal.. 2024: Miller etal.. 2011: Linton etal.. 2000: Gad. 2015).
In CD-I mice, exposure to PFHxS (1 mg/kg-day) during gestation and lactation increased
liver cholesterol levels in F0 females (Marques etal.. 2021) without changes in liver lipids and
triglycerides fMarques etal.. 2021: Chang etal.. 20181. In male F0 CD-I mice PFHxS exposure (0.3,
1, 3 mg/kg-day) for 42 days (before and after mating) caused an increase in hepatic microvesicular
fatty change at the high dose. These apparent differences in susceptibility between males and
females may be attributable to the pharmacokinetics of PFHxS in males versus females (see Section
3.1). Two studies using male C57BL/6J mice evaluated hepatic triglyceride and cholesterol levels in
male animals at doses ranging from 0.06 to 0.15 mg/kg-day (Pfohl etal.. 2020: He etal.. 20221.
Hepatic cholesterol levels were decreased after 29 weeks of exposure (Pfohl etal.. 2020). and
hepatic triglyceride contents and mRNA levels of genes associated with lipid synthesis, metabolism,
and transport were increased after 12 weeks fHe etal.. 20221. Pfohl etal., 2020 also measured liver
triglyceride levels after 29 weeks of exposure but observed no significant PFHxS-induced effects.
Another two animal studies used genetically modified mice: AP0E*3-Leiden.CETP mice17
and PPARa null mice fDas etal.. 2017: Biiland etal.. 20111. In AP0E*3-Leiden.CETP mice, a
genetically modified animal model used to investigate cholesterol metabolism and cardiovascular
disease (Oppi etal.. 2019). PFHxS exposure (6 mg/kg-day) increased hepatic triglyceride levels, but
free cholesterol levels were not affected fBiiland etal.. 20111. PPARa is a known master regulator
and potential pathway leading to hepatic lipid accumulation. In PPARa null and wild-type mice
PFHxS exposure (10 mg/kg-day) increased hepatic lipid content, but liver triglyceride levels were
only increased in the wild-type animals (Das etal.. 20171. These findings suggest that PFHxS
treatment-related effects include increased liver lipid content through a PPARa-independent
pathway (Das etal.. 2017). Furthermore, the same study used WY-14643, a PPARa activator, as a
positive control and observed no significant effects in hepatic lipid accumulation in WY-14643-
exposed PPARa-null animals. Das etal. f20171 also observed that PFHxS exposure did not have an
impact on fatty acid beta-oxidation in wild-type and PPARa-null animals, and a separate in vitro
experiment by the same group reported no significant exposure-related effects on rat hepatic
mitochondria fatty acid beta-oxidation. Gene expression analyses have revealed that in both wild-
type and genetically modified (PPARa-null) animals PFHxS treatment resulted in altered expression
of genes associated with peroxisomal and mitochondrial fatty acid metabolism (Das etal.. 2017)
and increased levels of genes associated with fatty acid and triglyceride transport and synthesis fHe
17APOE*3-Leiden.CETP mice is a genetically modified animal model which better emulates human lipoprotein
profiles and is used to investigate cholesterol metabolism and cardiovascular disease (Veseli et al.. 20171.
3-223
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
etal.. 2022: Das etal.. 2017). These responses were also attenuated in the PPARa-null mice (Das et
al.. 20171.
The available studies suggest that PFHxS may alter hepatic lipid metabolism in animal
models. Experiments using genetically modified animals suggest that PPARa activation plays a role
in the metabolic responses described above, but other pathways are likely involved. Overall, the
metabolic effects reported in the available studies are potential indicators of PFHxS-induced
alterations in hepatocyte function, which could eventually lead to abnormal liver metabolism and
accumulation of fatty acids resulting in fatty liver disease.18 Excessive and prolonged hepatic
accumulation of lipotoxic lipid species (e.g., free cholesterol and free fatty acids) fYounossi and
Henry. 2024: M etal.. 20241 is associated with fatty liver disease, promotion of lipotoxicity, pro-
inflammatory responses, cytotoxicity, and the progression from fatty liver disease to steatohepatitis
fS etal.. 2023: Ioannou. 2016: Idalsoaga etal.. 2020: Horn etal.. 2022: Geng etal.. 20211.
Reporting quality -
Allocation -
+
+
+
Observational bias/blinding -
NR
NR
NR
NR
NR
NR
Confounding/variable control -
+
Selective reporting and attrition -
-
+
Chemical administration and characterization -
+
-
+
Exposure timing, frequency and duration -
Results presentation -
End point sensitivity and specificity -
Overall confidence -\
I Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
I Critically deficient (metric) or Uninformative (overall)
Not reported
Not applicable
Figure 3-64. PFHxS liver hepatic lipid content study evaluation heatmap. For
additional details see HAWC link.
18Fatty liver (steatosis) is a hepatic response to moderate alcohol consumption, xenobiotic exposure, or other
factors that may alter metabolic functions (Wahlangetal.. 2013: Roth et al.. 2019: loshi-Barve et al.. 201 SI. It
is characterized by excessive lipid accumulation in hepatocytes (Angrish et al.. 2016) and is considered a
reversible response when the stimulus is temporary (Roth etal.. 2019). However, steatosis increases
susceptibility to other insults and persistent steatosis is considered a precursor to other forms of liver disease
(Roth et al.. 2019: Bessone et al.. 2019). When combined with inflammation (steatohepatitis) fatty liver can
progress to fibrosis and cirrhosis (Wahlang et al.. 2013: Roth etal.. 2019).
3-224
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Mechanistic evidence and supplemental information
Mechanistic evidence relevant to PFHxS-induced effects was collected from the peer-
reviewed literature and from in vitro high-throughput screening (HTS) assays from the ToxCast and
Tox21 databases accessed via EPAs Chemicals Dashboard. The available in vitro and in vivo studies
were evaluated based on a proposed mode of action (MOA) for liver injury for PFOS and PFOA, two
structural analogs of PFHxS and among the most well-studied PFAS (U.S. EPA. 2021c). Further, an
AOP-based approach was employed to organize and discuss the evidence according to the following
levels of biological organization: molecular events, cellular effects, organ effects, and organism
effects. Reponses informative of later two biological levels of organization are presented in the
preceding hazard sections. Refer to Appendix C for more details on the objective and methodology
of the mechanistic evaluation undertaken herein, and a description of the proposed MOA for PFAS-
induced hepatotoxicity (see Appendix C, Section 2). A detailed summary of the HTS data analysis
can be found in Appendix C, Section 3.
Molecular initiating events
The available studies have examined several nuclear receptor and cell signaling pathways
associated with chemical-induced liver toxicity. Many of the hepatic effects caused by exposure to
perfluorinated compounds such as PFHxS have been attributed to activation of the peroxisome
proliferator-activated receptor alpha (PPARa19) (U.S. EPA. 2016a. b; Rosen etal.. 2017: NTDWOI.
2017: Gleason. 2017: Das etal.. 2017). In vivo studies using SD rats or several strains of mice report
that exposure to PFHxS results in activation of PPARa and increased expression of PPARa-
responsive genes f Rosen etal.. 2017: NTP. 2018a: Das etal.. 2017: Chang etal.. 2018: Biiland etal..
20111. Two cell culture studies using rat FaO hepatoma cells or primary mouse hepatocytes also
reported altered expression of PPARa-responsive genes fRosenetal.. 2013: Biork etal.. 20211.
PFHxS also activates the human PPARa in human hepatoma cell lines Rosenmai etal. f20181 and in
primary human hepatocytes exposure was associated with increased expression of PPARa-
responsive genes (Rosen etal.. 2013). Overall, these studies suggest that PFHxS exposure can
activate PPARa in animal in vivo and in vitro studies, and in human liver cell culture models.
Animal studies also provide evidence suggesting that additional nuclear receptor pathways
may be involved in PFHxS-induced liver effects. Two studies using genetically modified animals
reported increases in absolute and relative liver weight in both wild-type and PPARa null animals
(Rosen etal.. 2017: Das etal.. 2017). However, one study (Rosen etal.. 2017) also reported that
effects such as hepatic lipid accumulation were ameliorated in PPARa-null mice. Gene expression
analyses in both wild-type and PPARa-null animals suggest hepatocellular receptors (other than
PPARa) can be affected by PFHxS exposure. These include: the constitutive androstane receptor
19PPARa is a member of the nuclear receptor superfamily that can be activated endogenously by free fatty
acid derivatives. PPARa plays a role in lipid homeostasis, but it is also associated with cell proliferation,
oxidative stress and inflammation (Mellor et al„ 2016: Li et al„ 2017a: Hall et al.. 20121.
3-225
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
(CAR), and the pregnane x receptor (PXR) (Rosen etal.. 2017: Chang etal.. 2018: Biiland etal..
2011: 3M. 20101. A 28-day study using SD rats also reported increased mRNA levels of CAR/PXR-
responsive genes in response to PFHxS exposure fNTP. 2018al. suggesting these molecular effects
are conserved across rodent models. Furthermore, PFHxS was able to activate nuclear receptors
other than PPARa, in human cells (including PPARa, RXR, LXR, FOS, and NRF2; see Appendix C).
Activation of these hepatic nuclear receptors plays an important role in regulating responses to
xenobiotics, energy and nutrient homeostasis, and development of fatty liver disease (Mellor etal..
2016: Mackowiaketal.. 2018: di Masi etal.. 2009: Angrish etal.. 20161.
Cellular effects
As discussed below, the available studies provide evidence for PFHxS-induced alterations in
reactive oxygen species production, cellular stress, inflammation, and cytotoxicity.
Excessive production of reactive oxygen species (ROS) is considered a mechanism
associated with PFAS-induced hepatocellular toxicity and progression of fatty liver to
steatohepatitis (Wahlang etal.. 2019: U.S. EPA. 2016a. b; Mendez-Sanchez etal.. 2018: Li etal..
2017a: Toshi-Barve etal.. 20151. One in vivo study using C57BL/6J mice reported increased mRNA
levels of genes associated with oxidative stress, after exposure to 0.15 mg/kg-day for 25 weeks
fPfohl etal.. 20201. Two cell culture studies using HepG2 human hepatocytes present conflicting
evidence fWielsae etal.. 2015: Oio etal.. 20211. While both studies exposed cells for the same
duration (24 hours) and similar concentrations (0, 0.02, 0.2, 2, 20, 200 |j.M in (Wielsae etal.. 2015):
and 0, 0.2, 2, 20 |a,M in fOio etal.. 202111 only Wielsae etal. f20151 observed increased intracellular
ROS production and neither study observed exposure-related changes in cellular antioxidant levels.
Pathways associated with inflammation were evaluated in C57BL/6J mice. Exposure to
PFHxS (0, and 60 to 110 mg/kg-day) for 12 weeks increased liver mRNA levels of pro-inflammatory
cytokine (IL-ip) and the pro-fibrogenic factor Colla, (He etal.. 2022). IL-ip and Colla play a role
in the loss of liver functions, and progression of hepatic steatosis to steatohepatitis and fibrotic
lesions (Veskovic etal.. 2024: Sultan etal.. 2017: He etal.. 2022).
Cytotoxicity induced by PFHxS exposure was evaluated in two cell culture studies using
HepG2 human hepatocytes fOio etal.. 2020: Oio etal.. 20211. Oio etal. f20201 reported increased
cytotoxicity at an effective dose of 183 |a,M. Oio etal. f20211. did not report PFHxS-induced changes
in cytotoxicity. However, this was a mixture study designed to evaluate the combined effects of
PFHxS with other PFAS and Ojo etal. (2021) selected concentrations below their previously
identified effective dose of 183 |a,M.
3-226
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Figure 3-65. Mode of action for PFHxS-induced liver effects. Adapted from
proposed MOA for PFAS-induced liver effects (See Systematic review protocol
for the PFAS IRIS assessments (U.S. EPA. 2019c)). The available evidence from
toxicological and mechanistic studies on PFHxS informs the key events displayed in
bold.
Conclusions from mechanistic evidence
Mechanistic evidence from in vivo and in vitro rodent cell models suggests that PFHxS
activates several hepatic xenobiotic-sensing nuclear receptors and other cell signaling pathways,
namely PPARa, PPARa, CAR, PXR, and LXR PFHxS exposure was also associated with alterations in
hepatic ROS production, cellular stress, and abnormal liver function related to lipid metabolism in
animals (including genetically modified mouse models). The molecular and cellular mechanisms
induced by PFHxS exposure in these models have been implicated in chemical-induced liver
diseases such as steatosis, steatohepatitis, and fibrosis fWahlang etal.. 2013: Mellor etal.. 2016:
Toshi-Barve etal.. 2015: Angrish etal.. 20161. and provide support for the biological plausibility of
the observed liver effects described above (i.e., histopathological responses, biomarkers of altered
liver function and lipid accumulation, and organ weight changes) in oral studies on PFHxS. Also,
these are mechanistic pathways activated by other PFAS that are known to cause hepatotoxicity
(Zhangetal.. 2024).
Available mechanistic information in human models is limited to two in vitro studies in the
peer-reviewed literature and HTS assays from the ToxCast databases accessed via EPAs Chemicals
Dashboard. As described in Appendix C-3, none of the 54 available assays in the ToxCast database
using the human hepatoma HepG2 cells were responsive to PFHxS treatment. These HTS assay
findings are inconsistent with the observations from the other two in vitro studies Wielsae etal.
(2015) and Rosenmai et al. (2018). which also used HepG2 cells and reported that PFHxS exposure
3-227
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
promotes activation of the human PPARa and increased reactive oxygen species production.
Additional studies are needed to resolve these conflicting results.
Overall, the mechanistic evidence on pathways known to be associated with liver toxicity
(i.e., increased oxidative stress, altered lipid metabolism/transport, and increased expression in
pro-inflammatory and pro-fibrogenic genes) provides biological plausibility and supports the liver
effects observed in animal bioassays (Figure 3-65). The available mechanistic evidence provides
support for a possible role for both PPARa-dependent and PPARa-independent mechanisms in the
hepatic responses to PFHxS exposure, including histopathological alterations, increased cellular
lipid content, increased liver weight, and increased levels of biomarkers of liver damage observed
in animal studies. Limited evidence from in vitro studies suggest that some responses may also be
activated in human cellular models, including nuclear receptor and transcription factor pathways
that regulate liver functions (i.e., PPARa/y, CAR, PXR, RXR, LXR, FOS, NRF2), and outcomes
indicative of oxidative stress and altered metabolism. As described above, activation of these
nuclear receptor and cell signaling pathways is associated with changes in hepatic functions, lipid
accumulation, and progression of fatty liver disease to steatohepatitis. However, inconsistencies
between the available peer-reviewed studies using human cell culture models and HTS assays from
the ToxCast database suggest that additional experiments are needed.
Considerations for potentially adaptive versus adverse responses
Increases in liver weight and hepatocyte hypertrophy were observed in rodents with PFHxS
administration in short-term oral studies and increased serum markers of liver toxicity and
histological and molecular changes associated with steatohepatitis were increased in mice after
subchronic exposures. Enlargement of the liver and/or individual hepatocytes is a common
chemical-induced response that can involve lipid accumulation (e.g., micro- or macro-vesicular
steatosis), organellar growth and proliferation (e.g., peroxisomes, endoplasmic reticulum),
increased intracellular protein levels (e.g., Phase I and II enzymes), and altered regulation of gene
expression (e.g., stress response, nuclear receptors) (reviewed by Batt and Ferrari (1995)).
Hepatocyte hypertrophy related to organelle growth and proliferation in response to activation of
xenobiotic-sensing receptors (primarily PPARa) is often considered an adaptive response (Hall et
al.. 20121. However, excessive hepatic lipid accumulation is also considered a risk factor for other
metabolic conditions such as diabetes and cardiovascular disease and as discussed above patients
with fatty liver can develop steatohepatitis fldalsoaga et al.. 20201. PFHxS-induced
histopathological effects and clinical markers considered indicative of adverse responses in the
liver (e.g., increased hepatic inflammation, elevated serum ALT damage (Hall etal.. 2012)) were
reported in the study by (He etal.. 2022). However, the histological findings from He etal. (2022)
were considered low confidence due to issues related with evidence reporting and animal allocation
to exposure groups, and the other in vivo short-term studies described above also evaluated
hepatocellular necrosis, inflammation and serum markers of liver disease and they report no
PFHxS-induced changes (see synthesis of histopathology and serum biomarkers of liver function
3-228
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
above). The available evidence from toxicological and mechanistic evidence suggests that short
term exposure to PFHxS promotes development of a phenotype consistent with fatty liver disease,
and that prolonged (i.e. subchronic or chronic) exposures have the potential to result in alterations
indicative of liver damage and progression of fatty liver disease to steatohepatitis. However, the
only animal study reporting increases in hepatic inflammation and fibrosis was considered low
confidence.
Evidence Integration
The available evidence suggests but is not sufficient to infer that exposure to PFHxS might
cause hepatobiliary system effects in humans given sufficient exposure conditions20. This is due to
limitations in the available evidence that introduce significant uncertainty (see Table 3-23).
The available evidence on PFHxS-induced hepatic effects in animal toxicity studies is
considered slight. The short-term and multigenerational animal studies provide evidence of PFHxS-
induced effects on multiple endpoints relevant to the assessment of liver responses to chemical
exposure. This includes: organ weight changes, histopathology [hepatocellular hypertrophy], lipid
accumulation, inflammation, and increased expression of serum markers of liver damage and, in
general, the responses observed in animal studies using rats or mice exhibited a dose-response
gradient. Alterations in serum biomarkers of liver/hepatobiliary function (ALT, ALP, bile
salts/acids, and globulin) were observed in SD rats fNTP. 2018a: Butenhoff etal.. 2009: 3M. 2000b.
2003). and C57BL/6J and CD-I mice (Chang etal.. 2018). One subchronic study exposed mice to
PFHxS for 12 weeks and reported a significant increase in ALT levels (He et al.. 2022). but several
concerns related with the study's experimental design and methods were identified. As described
above, responses such as alterations in ALT, ALP and albumin were not consistently observed in
similar short-term (NTP. 2018a: 3M. 2000b). subchronic and chronic (He etal.. 2022: Chang etal..
2018: Butenhoff etal.. 20091. or multigenerational studies fMarques etal.. 2021: Chang etal.. 2018:
Butenhoff et al.. 2009: 3M. 2003)
Increased liver weights were reported in SD rats after 28 to 44 days of exposure (NTP.
2018a: Butenhoff etal.. 2009: 3M. 2000b. 2003) and CD-I mice treated with PFHxS for 44 days
(Chang etal.. 2018). Alterations in histological responses were also observed in the available
studies and responses such as hepatocellular hypertrophy were consistently observed after short-
term exposure in male rats and mice (NTP. 2018a: Butenhoff et al.. 2009: 3M. 2000b) and F1
generation male and female mice f Chang etal.. 20181. He etal. f20221 reported increased hepatic
mRNA levels of pro-inflammatory and pro-fibrogenic genes and observed evidence of increased
hepatic inflammation after 12 weeks of exposure. As described above several issues were
identified with the He etal. (2022) study which lowers the confidence level of histopathological
outcomes to low. Outcomes indicative of hepatocellular degeneration (e.g., vacuolization) or injury
20The "sufficient exposure conditions" are more fully evaluated and defined for the identified health effects
through dose-response analysis in Section 5.
3-229
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
(e.g., necrosis) (Hall etal.. 20121 were unaffected in the available short-term and multigenerational
studies fChang etal.. 2018: Butenhoff etal.. 2009: 3M. 20031. Responses such as single cell necrosis
might progress to more severe effect after continued exposure fThoolen etal.. 20101. but the
available results from short-term studies are not sufficient to determine whether the observed
histological PFHxS-induced effects can progress to adverse hepatic injuries with continued
exposure. These responses are considered adaptive when applying the Hall criteria (Hall etal..
20121. As described in the synthesis section above, female animals are less responsive than males
to PFHxS-induced hepatic effects and these sex-based differences in susceptibility are likely
attributable to the pharmacokinetics of PFHxS in males versus females (see Section 3.1).
Two of the available animal studies identified in the literature search used genetically
modified murine models. Exposure to PFHxS also resulted in increased liver weight and hepatocyte
lipid accumulation in APOE*3-Leiden. CETP mice fBiiland etal.. 20111. as well as wild-type and
PPARa-null mice (Das etal.. 2017). These findings suggest that PFHxS exposure may have the
potential to promote fatty liver development, including in the absence of PPARa.
Analysis of mechanistic data from in vivo and in vitro rodent models provide biological
plausibility for the apical effects reported in the short-term and multigenerational oral studies
summarized above. Exposure to PFHxS was associated with the activation of several molecular
signaling pathways and altered cellular functions thought to be involved in the MOA for liver
toxicity of well-studied PFAS such as PFOA and PFOS (see synthesis of Mechanistic evidence and
supplemental information above for more details). Additionally, the evidence for PFHxS-mediated
liver effects point to potential PPARa-dependent and -independent pathways, which is consistent
with the mechanisms of potential hepatotoxicity for related perfluorinated compounds (U.S. EPA.
2016a. b; Li etal.. 2017a: ATSDR 20211.
Potential adverse liver effects caused by exposure to PFHxS and other PFAS have been
attributed, in part, to activation of PPARa fU.S. EPA. 2016a. b; Li etal.. 2017a: ATSDR. 20211.
However, in addition to PPARa, PFHxS exposure appears to promote activation of other nuclear
receptor pathways (PPARy, CAR, PXR, LXR, and transcriptional factors, FOS, and NRF2) and
responses indicative of oxidative stress and cellular damage were observed in human liver cell
models (see synthesis of Mechanistic studies and supplemental information above for more
details). In addition, studies of PFHxS in PPARa-null mice indicate that many of the observed
responses are unaffected by loss of PPARa-signaling. Therefore, the available evidence supports the
interpretation that PPARa-dependent and -independent mechanisms mediate PFHxS-induced
effects in animals.
The available evidence on PFHxS-induced hepatic effects in humans is considered slight.
There is some evidence of an association between PFHxS exposure and hepatic effects in human
studies that is based on largely consistent associations with liver biomarkers (primarily small
increases in ALT, a specific biomarker of potential liver injury) in the blood in multiple studies of
adults. Changes in serum lipids and uric acid provide coherence with these findings. However, the
3-230
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
available studies of functional hepatic effects are inconsistent, so it is not clear whether the
observed changes in liver enzymes observed in the biomarker studies translate into clinical hepatic
injury. Further, there is concern that the associations with ALT could be due to confounding by
other PFAS. Given the otherwise compelling nature of the evidence, an additional study
demonstrating that the association with PFHxS is independent of PFOA and PFNA would likely be
sufficient for this evidence base to reach a conclusion of moderate. As described above the available
animal evidence is considered slight based on organ weight, clinical markers of liver damage, and
related histopathology responses. However, the current evidence is insufficient to support the
adversity of the observed changes due to unclear biological significance (adversity) of the observed
responses. The evidence integration summary judgment concludes that the available evidence
suggests but is not sufficient to infer that exposure to PFHxS might cause hepatobiliary system
effects in humans21. This conclusion is based on the evidence synthesis judgments of slight for the
human and animal evidence.
21Given the uncertainty in this judgment and the available evidence, this assessment does not derive a toxicity
value that might better define the "sufficient exposure conditions" for developing this outcome (see Section 5
discussion).
3-231
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-23. Evidence profile table for oral PFHxS exposure and liver effects
Evidence stream summary and interpretation
Evidence
integration
summary
judgment
Evidence from studies of exposed humans (see Hepatic Human Studies Section)
©OO
Evidence
suggests, but is
not sufficient to
infer
Based primarily
on small increases
in ALT in men and
women, and
consistent, but
possibly not
adverse, hepatic
effects in rodents
Human relevance:
Limited studies in
human in vitro
models suggest
activation of
molecular and
cellular responses
observed in
rodent models
are relevant to
human toxicity
Cross-stream
coherence:
Studies and
confidence
Factors that increase
certainty
Factors that decrease certainty
Summary and key findings
Evidence stream
judgment
Serum Biomarkers
12 medium and 2 low
confidence studies
• Medium
confidence
studies reporting
an effect
• Consistency -
increased ALT in
adults
• Precision
• Coherence-
associations
observed for
PFHxS with serum
lipids and uric
acid may be
consistent with
hepatoxicity
• Potential for confounding by
other PFAS
• Unclear biological
significance of changes in
ALT
• Positive associations
observed between PFHxS
and ALT in 8/10 studies in
adults (6 statistically
significant).
©oo
Slight
Based on largely
consistent increases
in ALT in adults
3-232
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence
integration
summary
judgment
Alterations in
serum biomarkers
of hepatobiliary
injury were
reported in
animals and in a
few
epidemiological
studies, although
the observations
are uncertain, and
the markers
affected differed
across species.
Susceptible
populations and
life stages:
None identified,
although those
with preexisting
liver disease could
potentially be a
greater risk.
Liver disease
4 low confidence
studies
• No factors
noted
• Low confidence studies
• Inconsistency
• Positive association with a
marker of nonalcoholic
fatty liver disease in
women in one study;
inverse association in a
second study. A study of
self reported liver problems
found no association.
3-233
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence
integration
summary
judgment
• One study in children found
positive associations with
severity of nonalcoholic
steatohepatitis.
Evidence from in vivo animal studies (see Hepatic Animal Studies Section)
Studies and
confidence
Factors that increase
certainty
Factors that decrease certainty
Summary and key findings
Evidence stream
judgment
Organ Weight
5 high and 3 medium
and confidence
studies in rats and
mice
• 28-d (x2)
• 42-d
• 203-d
• Gestational (x4)
• Consistent
increases, across
studies
• Dose-response in
studies reporting
effects
• Coherence with
histopathology in
male rats and
mice
• All high or
medium
confidence
studies
• Unclear biological
significance (adversity) of
the combined hepatic
findings in animals across
endpoints
• Dose-related increases in
liver weights reported at
doses ranging from 1.25 to
50 mg/kg-d rat and mouse
studies, and a gestational
exposure study in mice
©oo
Slight
Based on consistent,
coherent, and dose-
dependent increases
in organ weight,
clinical markers of
liver damage, and
related
histopathology.
However, the current
evidence is
insufficient to
support the adversity
of the changes.
Histopathology
4 high, 1 medium, and
1 low confidence
studies in rats and
mice:
• 28-d (x2)
• 84-d
• Gestational (x3)
• Consistent
cellular
hypertrophy
across studies
and species
• Coherence with
liver weight
effects (especially
at high doses)
• Unclear biological
significance (adversity) of
histopathological changes
(e.g., no necrosis observed)
as well as the combined
hepatic findings in animals
across endpoints
• Hepatocellular lesions
observed in rats and mice
including hepatocellular
hypertrophy in mice
exposed to >0.3 mg/kg-d
and rats exposed to
2.5 mg/kg-d.
3-234
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence
integration
summary
judgment
• Dose response
• All high
confidence
studies
Serum Biomarkers
4 high confidence
studies in rats and
mice:
• 28-d (x2)
• 44-d
• 42-d
1 high and 2 medium
confidence studies in
mice
• 42-d
• 84-d
• Gestational (xl)
• Dose response
• Affected biomarker (ALP)
not specific to liver
• Inconsistent evidence on
ALT levels
• No effects on AST
• Unclear biological
significance (adversity) of
the combined hepatic
findings in animals across
endpoints
• Dose-related increases in
biomarker (ALP) in male
mice and rats exposed to 3
or 10 mg/kg-d respectively
• Increased serum ALT in 1
mouse study
• Increased marker of altered
function (tissue triglyceride
levels) in mice exposed to
6 mg/kg-d
Mechanistic evidence and supplemental information (see Mechanistic Studies and Supplemental Information Section)
Biological events or
pathways
Summary of key findings, interpretation, and limitations
Evidence stream
judgment
Molecular initiating
events — PPARa
Key findings and interpretation:
• Activation of hepatic PPARa in rat and mouse models. Some evidence of PPARa
activation in human in vitro models.
• In vivo PFHxS exposure increased expression of PPARa-responsive genes in wild-type
and hPPARa mice.
Limitations: No evidence in humanized in vivo models. Inconsistencies in peer-reviewed
and ToxCast/Tox21 studies using human hepatoma HepG2 cells.
Evidence indicates a
role for PPARa-
dependent and -
independent
pathways in the MOA
for noncancer liver
effects of PFHxS.
3-235
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence
integration
summary
judgment
Molecular initiating
events — PPARy
Key findings and interpretation:
• Activation of PPARy in mouse (in vivo) and human (in vitro) models.
• Increased expression of PPARy-responsive genes in vivo; and induction of PPARy
transactivation in human hepatoma HepG2 cells.
Limitations: Few studies and no evidence in humanized in vivo models.
Limited in vitro
studies suggest some
responses may be
activated in human
molecular/cellular
models.
Molecular initiating
events — CAR/PXR
Key findings and interpretation:
• Increased expression of CAR/PXR-responsive genes in mice.
Limitations: No evidence in humanized in vivo or in vitro models.
Molecular initiating
events — other
pathways
Key findings and interpretation:
• Limited in vivo evidence supports activation of cell signaling pathways related to
altered hepatic metabolism and oxidative/cellular stress responses (RXR, LXR, FOS,
and Nrf2).
Limitations: Few studies and no evidence in humanized in vivo or in vitro models.
Cellular effects
Key findings and interpretation:
• Increased hepatic lipid content and altered expression of genes associated with fatty
acid and triglyceride metabolism.
• Increased ROS production and markers of cellular stress/cytotoxicity in HepG2 cells.
Limitations: Few in vivo studies examining cellular toxicity, functions, other cell signaling
pathways, and no evidence in humanized in vivo models. Inconsistencies in the in vivo and
in vitro results likely due to differences in experimental model and/or design features.
3-236
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3.2.5. Neurodevelopmental Effects
The available database examining potential nervous system effects of PFHxS exposure was
composed of 17 epidemiological and 2 animal studies. All the studies in the evidence base examined
the effects of PFHxS in children or, in animal studies, exposed animals during early lifestages to
examine potential effects on neurodevelopment manifest in later lifestages (i.e., testing in newborn,
juvenile, or adult rats). Therefore, this section examines and discusses the evidence on PFHxS-
induced effects on the developing nervous system. For information on other developmental effects
please see Section 3.2.3.
Human Studies
Twenty-two studies (reported in in 31 publications) examined associations between PFHxS
exposure (measured in blood) and neurodevelopmental outcomes. Neurodevelopment is typically
assessed with a wide array of neurobehavioral or neuropsychological tests, which makes it difficult
to draw clear-cut divisions of neuropsychological categories. For example, a longer mean reaction
time (a measure of response time after a stimulus is introduced) on a continuous performance test
typically indicates inattention but may also be affected by slower information processing or motor
response. For the purposes of this review, and due partly to data availability, tests were organized
into the following categories: (1) cognition, (2) Attention Deficit Hyperactivity Disorder (ADHD) or
related behaviors, (3) social behavior or autism spectrum disorder, and (4) other outcomes. Nine
studies evaluated cognition, which comprised several endpoints including IQ, executive function,
language development, and intellectual disability. Seven studies evaluated ADHD or related
behaviors, which included ADHD diagnosis, inattention, impulsivity, hyperactivity, and
externalizing problems. Five studies evaluated social behavior and included autism spectrum
disorder (ASD) diagnosis, and two different autism screening scores, although there is overlap with
the behaviors assessed with ADHD. Given the heterogeneity in the tools and age ranges used in the
studies, it can be difficult to assess consistency within these categories. Other outcomes included
motor effects (three studies) and cerebral palsy (one study).
There were several considerations specific to the use of neuropsychological tests for
assessing children. For outcome ascertainment, tests used in a study should be appropriate for the
age range being studied and for the culture and language. Other relevant factors, such as time of day
of test administration or computer use, should have been considered, and some description of the
testing environment should have been provided. If there were multiple raters, this factor should
have been considered (e.g., statistical adjustment for rater, or analysis of interrater reliability).
While blinding to exposure is ideal, this information was not commonly reported, and it was
considered unlikely that participants or the outcome assessors would have knowledge of PFHxS
exposure levels during testing. Therefore, no blinding or lack of reporting on blinding was
determined to be unlikely to cause outcome misclassification. Evaluation of confounding was based
on the approach used by the study authors to identify potential confounders; confounders that
3-237
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
were considered potentially relevant across studies included child age and sex, maternal age,
socioeconomic status, quality of caregiving environment, prenatal tobacco exposure, and parental
mental health and IQ. It was considered preferable for analyses to use the outcome scales as
continuous variables to minimize misclassification into artificial categories and improve statistical
power fSagivetal.. 20151. although this does not apply to clinical diagnosis of conditions such as
ASD and ADHD.
The majority of available studies were birth cohorts or case-control studies nested in birth
cohorts that evaluated maternal exposure to PFHxS during pregnancy (Yao etal.. 2022: Wangetal..
2015: Vuongetal.. 2016: Spratlen etal.. 2020a: Skogheim etal.. 2021: Oulhote etal.. 2016: Oh etal..
2021: Niu etal.. 2019: Luo etal.. 2020: Liew etal.. 2018: Teddy etal.. 2017: Haver etal.. 2017: Harris
etal.. 2018: Dalsager etal.. 202 lbl. Some of these studies were considered adequate rather than
good for exposure measurement due to variations in the timing during gestation of sample
collection across participants within each study. While the half-life of PFHxS is long and exposure
levels are unlikely to have changed drastically during pregnancy, changes in hemodynamics during
pregnancy may influence levels in the blood at different points during pregnancy. In some cohort
studies, childhood exposure was measured as well fVuong et al.. 2018a: Oulhote etal.. 2016: Harris
etal.. 20181. There was one case-control study with measurements from banked maternal samples
fLvall etal.. 20181 and one case-control study with maternal samples taken concurrently with
outcome measurement (Shin etal.. 2020). In addition, there were three cross-sectional studies,
based on data from NHANES (Hoffman etal.. 2010). the C8 Health Project (Stein and Savitz. 2011).
and a survey in the United States (Gump etal.. 2011). While the exposures measured in these
studies with concurrent exposure and outcome measurement may not represent an etiologically
relevant period, particularly for capturing any influence of exposure on the genetic component of
ADHD, these studies were considered adequate for exposure measurement due to the long half-life
of PFHxS and since exposure levels are generally expected to be fairly stable over time. Reverse
causation is not a concern for these outcomes because neuropsychological performance is unlikely
to influence PFHxS levels. The study evaluations are summarized in Figure 3-66.
For data extraction and synthesis, when multiple exposure measures from different time
points (ages) were available, cross-sectional results were not extracted unless the results were
different from results from the prospective measurement
3-238
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Dalsager, 2021, 9960591 -
Gump, 2011, 3858629
Harris. 2018, 4442261 -
Hoffman, 2010, 1291112
Hoyer, 2017, 4184660
Jeddy, 2017, 3859807-
Liew. 2014, 2852208
Liew, 2015, 2851010-
Liew, 2018, 5079744
Luo, 2020, 7175034-
Lyall, 2018, 4239287
Niu. 2019, 5381527-
Oh, 2021, 7404108
Oulhote, 2016, 3789517
Shin, 2020, 6507470
Skogheim, 2019, 5918847
Skogheim, 2021, 9959649
Spratlen, 2020, 6364693
Stein, 2011, 1424971
Vuong,2016, 3352166
Wang. 2015, 3860120
Yao, 2022, 10273386
Legend
I Good (metric) or High confidence (overall)
+ I Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
I Critically deficient (metric) or Uninformative (overall)
* Multiple judgments exist
Figure 3-66. Summary of study evaluation for epidemiology studies of
neurodevelopment. Multiple publications of the same study: HOME study: Vuong
etal. (2016) also includes Vuong etal. (201 Rh). Vuong etal. (201 Ra). Vuong et
al. f20191. Braun et al. f20141. Zhang et al. f201Ral. Vuong etal. f20201. Vuong
et al. (2021a). and Vuong et al. (2021b). Project Viva: Harris etal. (201R) also
includes Harris et al. (2021). Four publications with data from the Danish National
Birth Cohort were evaluated separately due to significantly different procedures but
should not be considered independent: Liew etal. (2014). Liew etal. (2015). Liew
et al. (201R). Luo etal. (2020). Two publications with data from the Norwegian
Mother Father and Child Cohort were evaluated separately due to significantly
different selection procedures but should not be considered independent:
Skogheim etal. (2020) and Skogheim etal. (2021) For additional detail see
HAWC link.
3-239
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Cognition
Ten studies (13 publications) reported on endpoints related to cognition and PFHxS
exposure, including 9 medium confidence studies and 1 low confidence study. The medium
confidence studies are presented in Table 3-24. Among the medium confidence studies, there was a
nonstatistically significant inverse association with an exposure-response gradient across quartiles
in one study for nonverbal IQ when exposure was measured in mid-childhood (Harris etal.. 2018).
The same study also reports inverse associations between nonverbal IQ and maternal exposure
during pregnancy and between verbal IQ in mid-childhood and both exposure measures, but these
are nonmonotonic across the quartiles. Nonmonotonic associations with maternal exposure during
pregnancy were also observed for the Full-Scale Intelligence Quotient (FSIQ) at 5 years of age in
Liew etal. f 20181 and for intellectual disability in Lvall etal. f2018I Other studies reported
nonstatistically significant inverse associations with in some analyses but positive associations in
others (Yao etal.. 2022: Wang etal.. 2015: Vuong etal.. 2016): Vuong etal. (2019): (Skogheim etal..
2020: Niu etal.. 2019). with no clear pattern by endpoints, timing of exposure measurement, sex, or
any other factor. The remaining medium confidence studies did not show decreased cognition with
PFHxS exposure. Lastly, the single low confidence study fleddv etal.. 20171 reported associations in
opposite directions for multiple measures of language and communication development, and these
varied by maternal age. This could be due to social factors associated with age, but since only one
low confidence study examined this interaction, it should be interpreted with caution. Overall, while
there are some inverse associations between cognitive performance and PFHxS exposure, the
nonmonotonicity, general imprecision, and inconsistency across sub-analyses within studies make
the findings difficult to interpret It is possible that there are biological reasons for the
inconsistencies, but given the heterogeneity in study designs, the data currently do not provide
clear support for associations between PFHxS exposure and cognition in children.
Attention deficit hyperactivity disorder (ADHD) or related behaviors
Ten studies (13 publications) reported on associations between PFHxS exposure and ADHD
or behaviors potentially related to ADHD, including nine medium confidence studies and one low
confidence study. The medium confidence studies are presented in Tables 3-25 and 3-26. Six of the
studies (5 of 9 medium confidence) reported positive associations.
Two medium confidence studies examined ADHD diagnosis with PFHxS exposure measured
in children cross-sectionally and two studies were cohorts examining maternal exposure. Stein and
Savitz (2011) reported statistically significant associations between ADHD diagnosis and diagnosis
plus medication in children 5 to 18 years old and exposure-response gradients observed across
quartiles. Hoffman etal. (2010) also reported statistically significant positive associations for both
outcomes in children 12-15 years of age. Liew etal. f20151 and Skogheim etal. f20211 examined
ADHD cases identified from national registries. In Liew etal. f20151. the registry was limited to
hospital and psychiatric admissions, which likely represent only severe cases. Neither registry
3-240
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
study observed higher likelihood of ADHD with higher PFHxS exposure. All of the studies of ADHD
adjusted for sex but did not examine associations stratified by sex.
The remaining seven studies focused on behaviors. While these behaviors are not specific to
ADHD, many of them are elevated in individuals with ADHD and are used in its diagnosis.
Externalizing problems (consisting of hyperactivity and conduct subscales on the Strengths and
Difficulties Questionnaire [SDQ]) were examined in four studies (using the parent version of SDQ).
One medium confidence study (Haver etal.. 20171 reported a statistically significant positive
association for 5- to 9-year-olds with maternal exposure measured during the second trimester of
pregnancy modeled as continuous (when exposure was modeled as tertiles, there was an exposure-
response gradient across exposure groups, but it was not statistically significant). Another medium
confidence study using the SDQ reported nonstatistically significant positive associations for
externalizing internalizing, and total scores fLuo etal.. 20201. The other two study using the SDQ,
also medium confidence, did not report greater problem behaviors with higher exposure (Oulhote
etal.. 2016: Harris etal.. 20211. The SDQ is a validated instrument, but its sensitivity for ADHD has
been inconsistent in different populations (Ullebo etal.. 2011: Pritchard. 2012: Hall etal.. 20191.
Looking at other neurobehavioral tests, most had only a single study available. One study
examined impulsivity and inattention using a different tool (the Conners Continuous Performance
Test-II) and also found a nonstatistically significant positive association, for inattention but not
impulsivity in 8-year-olds with both maternal exposure and exposure measured in the children
(Vuong et al.. 2018a). In the same study population using a different tool (the Behavioral
Assessment System for Children 2 [BASC-2]), positive associations were reported with
externalizing problems, hyperactivity, internalizing problems, and attention (statistically significant
for all but the latter) when exposure was measured during gestation, but no associations were
observed when exposure was measured in children at 3 years. Another medium confidence study
found no association with behavior problems (measured using the Child Behavior Checklist) using
either maternal or childhood exposure measurement Finally, a low confidence cross-sectional
study examined inter-response time (IRT) at age 9-11 and found statistically significant decreases
in IRT, which indicates poor response inhibition (a primary deficit in children with ADHD) as the
test is designed to reward longer response times (Gump etal.. 20111.
Taken together, there is some evidence of an association between PFHxS exposure and
ADHD or potentially related behaviors. A positive association was observed in most studies (6 of
10) across a variety of populations and diagnostic tests, with an exposure-response gradient in
multiple studies. However, there is remaining uncertainty. Associations were inconsistent across
medium confidence studies. In addition, the only studies reporting an association with ADHD
diagnosis are cross-sectional, which may not represent exposure in an etiologically relevant period,
while the prospective study of ADHD diagnosis reported an inverse association, although the bias in
the cross-sectional studies would likely be toward the null due to nondifferential misclassification.
3-241
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
A few studies examined the possibility of an interaction with sex. Vuong etal. (2018a)
reported better performance (lower errors of omission) in boys with higher PFHxS ((3 = -4.5, 95%
CI: -10.0,1.0), but worse in girls ((3 = 3.2, 95% CI: -1.1, 7.4). In sex-stratified analyses in Oulhote et
al. f20161. most associations were similar in boys and girls, but some had deficits in girls but not
boys (cross-sectional analyses at 7 years for externalizing problems and related subscales). Haver
etal. (2017) reported a lack of interaction with sex (p >0.1). Evidence is not adequate to fully
assess differences in the association with ADHD or related behaviors by sex.
Social behavior or autism spectrum disorder
Nine studies (10 publications), all medium confidence, examined social behaviors or ASD
and PFHxS exposure. Five studies examined ASD diagnosis. Two studies fShin etal.. 20201: Liew et
al. f20151 reported positive associations. Liew etal. f20151 found a higher risk ratio (RR 1.10, 95%
CI: 0.92,1.33) with PFHxS exposure and Shin etal. (2020) a higher odds ratio (OR 1.36, 95% CI:
0.96,1.93). The associations in both studies became statistically significant when adjusting for
other PFAS. The other three studies ASD diagnosis reported no increase in the odds of ASD
diagnosis fSkogheim etal.. 2021: Oh etal.. 2021: Lvall etal.. 20181.
Four medium confidence studies (five publications) examined questionnaires for social
behavior. Braun etal. f20141 used the Social Responsiveness Scale at 4 and 5 years and reported a
nonsignificant positive association (more problem behaviors) ((3: 0.4, 95% CI: -1.5, 2.3); in the
same study population, Vuong etal. (2021b) used the BASC-2 questionnaire and found similar
results with poor social skills. Niu etal. (2019) examined the Ages and Stages questionnaire at
4 years of age and also reported an elevated risk ratio (p > 0.05) for personal social skills problems
with higher exposure (RR 1.60, 95% CI: 0.92, 2.80 per ln-unit increase in exposure). However,
Oulhote etal. (2016) calculated an autism screening score using the peer problems and prosocial
subscales on the SDQ at 7 years and reported an inverse association (mean difference: -0.1, 95% CI:
-0.3, 0.1). Yao etal. (2022) reported no association with the Social Development Quotient on the
Gesell Development Schedules at 1 year. Three of these studies measured PFHxS exposure in
maternal serum samples collected during pregnancy (most at 16 weeks gestation for Braun et al.
(2014). at 12-16 weeks gestation for Niu etal. (2019). and at 32 weeks gestation for Oulhote et al.
f201611: one study measured exposure in cord blood fYao etal.. 20221. and one study measured
exposure in childhood at 3 and 8 years (Vuong et al.. 2021b).
Overall, there is some evidence of an association between PFHxS exposure and autism and
social behaviors, but there is inconsistency across studies and estimates are generally imprecise,
with wide confidence intervals and lack of statistical significance. It is feasible that the
inconsistency could be explained by timing of exposure measurement, autism measurement tool, or
some other factor, but it is not possible to determine with the evidence currently available.
3-242
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Other neurodevelopmental outcomes
Four medium confidence studies reported on motor-related behaviors and PFHxS exposure.
In (Harris etal.. 20181. there was a statistically significant decrease in the visual-motor score from
the Wide Range Assessment of Visual Motor Abilities (WRAVMA) test in mid-childhood with higher
exposures, when measured cross-sectionally (mean difference (95% CI) versus Ql: Q2: -5.1 (-8.9,
-1.3); Q3: -5.0 (-9.0, -0.9), Q4: -5.0 (-9.1, -0.8)). When using a maternal exposure measure during
pregnancy, the association was nonmonotonic across the quartiles. No association was observed
between the WRAVMA total score and early childhood and maternal exposure measures. In Yao et
al. f20221. a statistically significant inverse association was reported with the Gross Motor
Development Quotient on the Gesell Development Schedules at 1 year. Conversely, in Spratlen et al.
f2020al. positive associations (better motor function on Motor Development Index on Bayleys
Scales of Infant Development) were observed with PFHxS exposure at 1, 2, and 3 years of age
(p > 0.05). An association (p > 0.05) with better fine motor skills was also observed in Niu etal.
(20191. but no association was observed with gross motor skills using the Ages and Stages
Questionnaire. Given the lack of consistency across studies, the evidence is not clear of an
association between PFHxS exposure and motor-related behaviors.
One medium confidence study examined the association of PFHxS exposure measured
during the first or second trimester of gestation with rates of cerebral palsy fLiew etal.. 20141.
Cases of congenital cerebral palsy were identified from a population-based registry. There was a
nonstatistically significant positive association with congenital cerebral palsy in boys (RR 1.2, 95%
CI: 0.9,1.7, exposure-response gradient across quartiles). No association was observed in girls (RR
1.1, 95% CI: O.6., 1.9), and when limited to girls born at term, a nonsignificant inverse association
was observed (RR 0.7, 95% CI: 0.3,1.6). Given the lack of additional studies and imprecision in the
estimate (i.e., wide confidence intervals), there is no clear evidence of an association between
PFHxS exposure and cerebral palsy.
3-243
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-24. Summary of results for medium confidence epidemiology studies of PFHxS exposure and cognitive
effects
Study name,
country,
reference(s)
Measu red
endpoint
(test used)
Exposure
measu rement
timing
Estimate
type
(adverse
direction)3
Sub-
population/
N
Group or unit
change
Exposure
median (IQR) or
range (quartiles)
Effect
estimate
CI LCL
CI UCL
Danish National
Birth Cohort,
Denmark
Liew et al. (2014)
FSIQ at 5 yr
(WPPSI)
Maternal
(median 8.7, SD
2.5 wk
gestation)
Mean
Difference
vs. Q1 (4,)
Boys
(n = 831)
Q1
1.39
-2.0
-7.0
2.9
Mean
Difference
vs. Q1 (4,)
Girls
(n = 761)
Q1
1.39
-0.7
-5.1
3.6
Health Outcomes
and Measures of
the Environment
(HOME),
U.S.
Vuong et al.
(2016)
Vuong et al.
(2019)
Vuong et al.
(2020)
FSIQ at 8 yr
(WISC-IV)
3 yr
Regression
Coefficient
w
221
Ln-unit
increase in
exposure
NR
-0.4
-2.5
1.6
Maternal
(16 ± 3 wk
gestation)
Regression
Coefficient
w
221
Ln-unit
increase in
exposure
GM 1.4
0.5
-1.8
2.9
Global
executive
function
score at
5/8 yr (BRIEF)
Maternal
(16 ± 3 wk
gestation)
Mean
Difference
m
219
Ln-unit
increase in
exposure
1.5 (0.9-2.4)
1.36
-0.41
3.12
Reading
composite
scores at 8 yr
Maternal
Regression
Coefficient
w
161
LoglO-unit
increase in
exposure
1.7
4.5
-3.1
12.0
3-244
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study name,
country,
reference(s)
Measu red
endpoint
(test used)
Exposure
measu rement
timing
Estimate
type
(adverse
direction)3
Sub-
population/
N
Group or unit
change
Exposure
median (IQR) or
range (quartiles)
Effect
estimate
CI LCL
CI UCL
Project Viva,
U.S.
Harris et al. (2018)
Word
knowledge
early
Maternal
(5-21 wk
gestation)
Mean
Difference
vs. Q1 (4,)
948
Q1
<0.1-1.6
Ref
Q2
1.7-2.4
0.7
-1.6
2.9
Harris et al. (2021)
childhoodb
(PPVT)
Q3
2.5-3.7
0.1
-2.1
2.4
Q4
3.8-43.2
0.4
-1.9
2.7
Verbal IQ
mid-
childhood15
Maternal
(5-21 wk
gestation)
Mean
Difference
vs. Q1 (4,)
851
Q1
<0.1-1.6
Ref
Q2
1.7-2.4
-2.8*
-5.1
-0.5
(KBIT)
Q3
2.5-3.7
-1.2
-3.6
1.2
Q4
3.8-43.2
0.3
-2.2
2.8
Mid-childhood
(6-10 yr)
Mean
Difference
vs. Q1 (4,)
631
Q1
<0.1-1.1
Ref
Q2
1.2-1.9
-0.8
-3.6
2.1
Q3
2.0-3.4
-0.2
-3.3
2.8
Q4
3.5-56.8
-1.7
-4.8
1.5
Nonverbal IQ
mid-
childhood15
Maternal
(5-21 wk
gestation)
Mean
Difference
vs. Q1 (4,)
862
Q1
<0.1-1.6
Ref
Q2
1.7-2.4
-3.9*
-6.9
-0.5
(KBIT)
Q3
2.5-3.7
-1.6
-4.7
1.5
Q4
3.8-43.2
-1.0
-4.2
2.2
Mid-childhood
(6-10 yr)
Mean
Difference
vs. Q1 (4,)
640
Q1
<0.1-1.1
Ref
Q2
<0.1-1.1
-0.9
-4.4
2.7
Q3
1.2-1.9
-2.3
-6.1
1.5
Q4
2.0-3.4
-2.7
-6.6
1.2
3-245
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study name,
country,
reference(s)
Measu red
endpoint
(test used)
Exposure
measu rement
timing
Estimate
type
(adverse
direction)3
Sub-
population/
N
Group or unit
change
Exposure
median (IQR) or
range (quartiles)
Effect
estimate
CI LCL
CI UCL
Global
executive
function
score at 6-
10 yr (BRIEF)
Maternal (5-
21 wk gestation)
Mean
Difference
vs. Q1 en
921
Q1
<0.1-1.6
Ref
Q2
1.7-2.4
-0.3
-1.9
1.3
Q3
2.5-3.7
0.2
-1.4
1.9
Q4
3.8-43
-1.1
-2.8
0.6
Taiwan maternal
and infant cohort
study,
Taiwan
Wang etal. (2015)
FSIQ at 5 yr
(WPPSI)
Maternal
(3rd trimester)
Regression
Coefficient
w
120
Doubling of
exposure
0.7 (0.07-1.09)
0.4
-1.1
1.9
FSIQ at 8 yr
(WISC)
Regression
Coefficient
w
120
Doubling of
exposure
0.7 (0.07-1.07)
-0.2
-1.8
1.4
WTC cohort, U.S.
Spratlen et al.
(2020a)
MDI at lyr
(BSID)
Cord blood/
maternal (1 d
post-delivery)
Regression
Coefficient
w
302
Log-unit
increase
GM 0.7 (range
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study name,
country,
reference(s)
Measu red
endpoint
(test used)
Exposure
measu rement
timing
Estimate
type
(adverse
direction)3
Sub-
population/
N
Group or unit
change
Exposure
median (IQR) or
range (quartiles)
Effect
estimate
CI LCL
CI UCL
FSIQ at 6 yr
(WPPSI)
Girls 150
0.57
-3.13
4.27
Boys 152
-1.64
-8.07
4.79
Norwegian
Mother, Father
and Child cohort,
Norway
Skogheim et al.
(2020)
Verbal
working
memory at
42 mo (CDI)
Maternal (17 wk
gestation)
Regression
coefficient
w
768
Q2
0.7 (0.5-0.9)
0.03
-0.20
0.26
Q3
0.10
-0.13
0.33
Q4
0.20
-0.03
0.44
Q5
0.21
-0.03
0.45
Nonverbal
working
memory at
42 mo (CDI)
934
Q2
-0.18
-0.38
0.03
Q3
-0.05
-0.26
0.16
Q4
-0.23
-0.44
-0.02
Q5
-0.18
-0.40
0.04
Shanghai-Minhang
cohort, China
Niu et al. (2019)
Communicati
on at 4 yr
(ASQ-3)
Maternal (12—
16 wk gestation)
Risk ratio
for
problems
m
533
Ln-unit
increase in
exposure
2.8(2.1-0.5)
1.10
0.78
1.54
Girls 236
1.46
0.79
2.70
Boys 297
0.90
0.60
1.35
Problem
solving at 4 yr
(ASQ-3)
533
0.85
0.54
1.36
Girls 236
1.06
0.40
2.78
Boys 297
0.75
0.43
1.32
Early Markers for
Autism (EMA),
U.S.
Intellectual
disability at
4-9 yr
Maternal (15—
19 wk gestation)
Odds Ratio
(or) en
622
Ln-unit
increase in
exposure
GM 1.33
1.11
0.86
1.42
160
Q1
<0.8
1.0
3-247
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study name,
country,
reference(s)
Measu red
endpoint
(test used)
Exposure
measu rement
timing
Estimate
type
(adverse
direction)3
Sub-
population/
N
Group or unit
change
Exposure
median (IQR) or
range (quartiles)
Effect
estimate
CI LCL
CI UCL
Lvall et al. (2018)
(clinical
diagnosis)
Odds Ratio
(OR) vs. Q1
m
171
Q2
0.8-<1.3
1.43
0.86
2.40
133
Q3
1.3-<2.0
1.03
0.58
1.85
157
Q4
> = 2.0
1.30
0.74
2.29
Laizhou Wan Birth
Cohort, China
Yao et al. (2022)
Adaptive
Development
Quotient at
1 yr
Cord serum
Regression
coefficient
w
274
LoglO-unit
increase in
exposure
0.3 (range 0.1-
1.1)
-1.40
-6.17
3.37
Girls 135
-2.02
-9.27
5.23
Boys 139
-1.22
-7.62
5.18
Language
Development
Quotient at
1 yr
274
3.00
-1.67
7.67
Girls 135
2.05
-4.82
8.93
Boys 139
4.02
-2.39
10.42
*p < 0.05.
FSIQ: = full-scale intelligence quotient; WPPSI = Wechsler Primary and Preschool Scales of Intelligence, WISC = Wechsler Intelligence Scale for Children,
BRIEF = Behavior Rating Inventory of Executive Function, PPVT = Peabody Picture Vocabulary Test, KBIT = Kaufman Brief Intelligence Test, BSID = Bayley Scales
of Infant Development, MDI = mental development index.
aThe arrows indicate the direction the effect estimate will be if there is an association between PFHxS and reduced cognitive performance. For some tests, a
higher score means better performance, while for other tests, a higher score means more problems.
bEarly childhood median age 3.2 years, range 2.8-6.3; Mid-childhood median age 7.7 years, range 6.6-10.9.
Table 3-25. Summary of results for medium confidence epidemiology studies of PFHxS exposure and attention
deficit hyperactivity disorder (ADHD)
Study name
Measu red
endpoint
Exposu re
measu re-
ment timing
Estimate type
(adverse
direction)3
Subpopulation/N
Group or unit
change
Exposu re
median (IQR) or range
(quartiles)
Effect
Estimate
CI LCL
CI UCL
C8 Health
Project,
U.S.
ADHD
diagnosis at 5-
18 yr (clinical)
Cross-
sectional
Odds Ratio
(OR) vs. Q1
m
10,546
Q1
0.25-<2.9 ng/mL
1.0
Q2
2.9-<5.2
1.27*
1.06
1.52
Q3
5.2-<10.1
1.43*
1.21
1.70
3-248
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study name
Measu red
endpoint
Exposu re
measu re-
ment timing
Estimate type
(adverse
direction)3
Subpopulation/N
Group or unit
change
Exposu re
median (IQR) or range
(quartiles)
Effect
Estimate
CI LCL
CI UCL
(Stein and
Savitz 2011)
04
10.1-276.4
1.53*
1.29
1.83
ADHD
diagnosis +
medication at
5-18 yr
(clinical)
Cross-
sectional
Odds Ratio
(OR) vs. Q1
m
10,546
Q1
0.25-<2.9 ng/mL
1.0
02
2.9-<5.2
1.44*
1.09
1.90
03
5.2-<10.1
1.55*
1.19
2.04
04
10.1-276.4
1.59*
1.21
2.08
NHANES (1999-
2000, 2003-
2004),
U.S.
Hoffman et al.
(2010)
ADHD at 12-
15 yr (clinical)
Cross-
sectional
Odds Ratio
(OR) m
571
One unit increase
in exposure
2.2 (2.9)
1.06*
1.02
1.11
ADHD+
medication at
12-15 yr
(clinical)
2.2 (2.9)
1.07*
1.03
1.11
Danish National
Birth Cohort,
Denmark
Liew et al.
(2015)
ADHD
diagnosis
(national
registry)
Maternal
(1st
trimester)
Risk ratio (1^)
770
In-unit increase
Controls 0.9 (0.7-1.2)
0.97
0.88
1.08
Q1
1.23
0.67*
0.54
0.83
Norwegian
Mother Father
Child Cohort,
Norway
Skogheim et al.
(2021)
ADHD
diagnosis
(national
registry)
Maternal
(2nd
trimester,
18 wk
gestation)
Odds ratio
m
1801
01
0.1-0.5
1.0
02
0.5-0.6
1.08
0.82
1.42
03
0.6-0.9
1.12
0.85
1.49
04
0.9-15
0.89
0.66
1.19
*p <0.05.
aThe arrows indicate the direction the effect estimate will be if there is an association between PFHxS and reduced behavior. For all the tests included here,
higher scores indicate more ADHD diagnosis.
3-249
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-26. Summary of results for medium confidence epidemiology studies of PFHxS exposure and behavior
Study name
Measured
endpoint
Exposu re
measure-
ment timing
Estimate type
(adverse
direction)3
Subpopulation/N
Group or unit
change
Exposure
median (IQR) or range
(quartiles)
Effect
Estimate
CI LCL
CI UCL
Faroe Island
cohort,
Denmark
Externalizing
problems at
7 yr (SDQ)
5 yr
Mean
Difference
m
508
Per doubling of
exposure
0.6 (0.5-0.9)
0
-0.36
0.37
Maternal
(32 wk
gestation)
539
4.5 (2.2-8.4)
-0.19
-0.48
0.11
(2016)
Internalizing
problems at
7 yr (SDQ)
5 yr
Mean
Difference
m
508
Per doubling of
exposure
0.6 (0.5-0.9)
-0.1
-0.43
0.22
Maternal
(32 wk
gestation)
539
4.5 (2.2-8.4)
-0.1
-0.36
0.17
Total SDQ
score at 7 yr
5 yr
Mean
Difference
m
508
Per doubling of
exposure
0.6 (0.5-0.9)
-0.1
-0.66
0.46
Maternal
(32 wk
gestation)
539
4.5 (2.2-8.4)
-0.28
-0.75
0.18
INUENDO (Bio
persistent
organochlorines
in diet and
human
fertility),
Greenland,
Ukraine, Poland
Hgver et al.
Hyperactivity
score at 5-9 yr
(SDQ)
Maternal
(median 2nd
trimester)
Regression
Coefficient
m
1,023
In-unit increase in
exposure
1.5 (10th-90th 0.7-3.4)
0.20*
0.00
0.40
Low exposure
0.2-1.2
Ref
Medium exposure
1.2-2.0
0.15
-0.30
0.60
High exposure
2.0-18.8
0.41
-0.03
0.86
Total SDQ
score at 5-9 yr
Maternal
(median 2nd
trimester)
Regression
Coefficient
m
1,023
In-unit increase in
exposure
1.5 (10th-90th 0.7-3.4)
0.45
-0.03
0.92
Low exposure
0.2-1.2
Ref
(2017)
Medium exposure
1.2-2.0
0.68
-0.04
1.38
High exposure
2.0-18.8
0.80*
0.06
1.54
Project Viva,
U.S.
Harris et al.
Externalizing
problems at 6-
10 yr (SDQ)
Maternal (5-
21 wk
gestation)
Mean
Difference vs.
Qicn
921
Q1
<0.1-1.6
Ref
Q2
1.7-2.4
0.0
-0.5
0.5
(2021)
Q3
2.5-3.7
0.6
0.0
1.1
Q4
3.8-43
0.0
-0.5
0.6
3-250
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study name
Measured
endpoint
Exposu re
measure-
ment timing
Estimate type
(adverse
direction)3
Subpopulation/N
Group or unit
change
Exposure
median (IQR) or range
(quartiles)
Effect
Estimate
CI LCL
CI UCL
Internalizing
problems at 6-
10 yr (SDQ)
Q1
<0.1-1.6
Ref
Q2
1.7-2.4
0.2
-0.3
0.6
Q3
2.5-3.7
-0.1
-0.5
0.4
Q4
3.8-43
0.2
-0.3
0.7
Total SDQ
score at 6-
10 yr
Q1
<0.1-1.6
Ref
Q2
1.7-2.4
0.2
-0.6
1.0
Q3
2.5-3.7
0.5
-0.3
1.4
Q4
3.8-43
0.2
-0.7
1.1
Danish National
Birth Cohort,
Denmark
Externalizing
problems at
7 yr
Maternal
(1st
trimester)
OR CH (odds
of elevated
score)
2421
Per doubling of
exposure
0.9 (0.7-1.3)
1.11
0.86
1.43
Internalizing
problems at
7 yr
1.18
0.88
1.58
Total SDQ
score at 7 yr
1.15
0.94
1.42
Odense Child
Cohort,
Denmark
(Dalsager et al„
2021b)
Behavior
problems (CBC)
at 2-5 yr
Maternal (8-
16 wk
gestation)
Incidence rate
ratio CH
1138
Doubling of
exposure
0.4
0.98
0.93
1.03
Odds ratio
m
0.95
0.79
1.16
18 mo
Incidence rate
ratio CH
817
0.3
0.95
0.88
1.04
Odds ratio
m
1.04
0.79
1.37
Health
Outcomes and
Measures of the
Environment
(HOME)
Impulsivity -
Commissions
at 8 yr (CPT)
3 yr
Regression
Coefficient
m
204
In-unit increase in
exposure
1.9 (1.0-3.3)
-0.6
-2.1
1.0
Maternal
(16 ± 3 wk
gestation)
1.3 (0.8-2.3)
-0.5
-1.9
0.9
3 yr
1.9 (1.0-3.3)
0.6
-2.3
3.5
3-251
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study name
Measured
endpoint
Exposu re
measure-
ment timing
Estimate type
(adverse
direction)3
Subpopulation/N
Group or unit
change
Exposure
median (IQR) or range
(quartiles)
Effect
Estimate
CI LCL
CI UCL
U.S.
Vuong et al.
(2018a)
Inattention -
Omissions at
8 yr (CPT)
Maternal
(16 ± 3 wk
gestation)
1.3 (0.8-2.3)
2.5
-0.9
6.0
Vuong et al.
(2021a)
Vuong et al.
(2021b)
Externalizing
problems
(BASC-2) at 5
and 8 yr
Maternal
(16 ± 3 wk
gestation)
Odds ratio
m
241
In-unit increase in
exposure
1.5
1.9*
1.1
3.2
Hyperactivity
(BASC-2)
2.5*
1.5
4.3
Attention
(BASC-2)
1.2
0.8
1.9
Internalizing
problems
(BASC-2)
2.0*
1.1
3.4
Externalizing
problems
(BASC-2) at 8 yr
3 yr
Regression
Coefficient
m
208
Ln-unit increase in
exposure
1.9
0.02
-1.6
1.6
Hyperactivity
(BASC-2)
-0.3
-1.9
1.2
Attention
(BASC-2)
-0.1
-1.6
1.4
Conduct
problems
(BASC-2)
0.4
-1.3
2.1
*p < 0.05.
SDQ: = Strengths and Difficulties Questionnaire, CPT = Conners continuous performance test, CBC = Child Behavior Checklist, BASC-2 = Behavioral Assessment
System for Children 2.
aThe arrows indicate the direction the effect estimate will be if there is an association between PFHxS and reduced behavior. For all the tests included here,
higher scores indicate more difficulties/behavior problems.
3-252
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Animal Studies
There were three animal studies evaluating neurodevelopmental outcomes and PFHxS
exposure: two medium confidence studies fRamhai etal.. 2020: Butenhoff et al.. 20091 and one low
confidence study fViberg etal.. 20131 (see Figure 3-67). Butenhoff et al. f20091 exposed male and
female Crl:CD Sprague Dawley rats to 0.3,1, 3, or 10 mg/kg-day daily via oral gavage starting at
14 days prior to cohabitation (Fo). Fi pups were not exposed directly but were exposed in utero and
through lactation. The study authors then assessed 5 pups per sex per litter from 10 dams using the
functional observation battery (FOB)22 and an automated motor activity assessment tool atPND22.
In the second medium confidence study Ramh0i etal. (20201 exposed Wistar dams to 0, 0.05, 5, or
25 mg/k bw-day PFHxS via gavage starting at gestational day 7 (GD 7) through postnatal day (PND)
22. After weaning one male and one female pup from each litter subsequently underwent
behavioral assessment of motor activity levels23 at each of three ages: PND 27, PND 115, and PND
340. Additionally, Viberg etal. (2013) evaluated spontaneous locomotor behavior by exposing male
and female NMRI mouse pups at postnatal day 10 (PND10) to a single oral dose of PFHxS at 0.61,
6.1, or 9.2 mg/kg-bw PFHxS. Spontaneous locomotor behavior was then evaluated at 2- and 4-
months post-exposure, and nicotine-induced behavior was evaluated at 4 months.
22FOB evaluations consisted of assessment of (1) autonomic functions: lacrimation, salivation, palpebral
closure, prominence of the eye, pupillary reaction to light, piloerection, respiration, and urination and
defecation; (2) reactivity and sensitivity: sensorimotor responses to visual, auditory, tactile and painful
stimuli; (3) excitability reactions to handling and behavior in the open field; (4) gait and sensorimotor
coordination: gait pattern in the open field, severity of gait abnormalities, air righting reaction and landing
foot splay; forelimb and hindlimb grip strength; and (5) abnormal clinical signs including convulsions,
tremors and other unusual behavior, hypotonia or hypertonia, emaciation, dehydration, unkempt appearance
and deposits around the eyes, nose or mouth. (Butenhoff et al.. 20091
23Measured in activity boxes with photocells recording horizontal activity for 30 minutes. Rearing behavior
(vertical activity) was not measured by Ramhfli et al. (20201
3-253
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
« ~~ ~~
~
NR
NR
-
NR
NR
++ ++ ++
++ ++
+
++ ++
+
++ ++
++
~
~
I 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
Figure 3-67. Confidence scores of neurodevelopmental system effects from
repeated PFHxS dose animal toxicity studies. For additional details see HAWC
link.
Functional observation battery fFOBl
One study fButenhoff etal.. 20091 reported on PFHxS effects on FOB assessment on F1
pups. The authors reported no statistically significant differences between control animals and
PFHxS treated animals on the assessments of FOB parameters.
Learning and memory
One study fRamhai etal.. 20201 reported on PFHxS effects on radial arm maze assessments
in Wistar male and female rat offspring exposed to PFHxS in utero and through lactation.
Assessments were performed at PND 115 and PND 340. The authors reported that no statistically
significant differences between control animals versus PFHxS treated animals.
Motor-related behaviors
Butenhoff et al. f20091. Ramhai etal. f20201 and Vibergetal. f20131 evaluated and
reported on locomotor activity (including anxiety-related behaviors) in response to PFHxS
exposure. The two medium confidence studies, Butenhoff et al. f20091 and Ramhai etal. f20201.
reported no statistically significant differences in motor activity in either sex with in utero and
3-254
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
lactational PFHxS dosing of dams from 0.05 to 25 mg/kg-day. One low confidence study, Viberg et
al. T20131 reported decreases in ambulatory (horizontal) activity and rearing behaviors in male and
female NMRI pups at 2 and 4 months following a single oral dose of PFHxS at 0.61, 6.1, or
9.2 mg/kg-bw PFHxS at postnatal day 10 (PND 10) during the habituation (first 20 minutes) and
end (minutes 40-60) periods of observation at 2 and 4 months after a single exposure to
9.2 mg/kg-day PFHxS on PND 9; however, the authors did not account for the potential impact of
litter effects In their experimental design, and they allocated pups to dosing groups from 3-4 litters
in an unclear fashion, reducing confidence in these findings. Taken together, the potential effects of
PFHxS exposure on motor-related behaviors in rodents remain unknown.
Mechanistic evidence and supplemental information
Seven mechanistic studies were identified relating to the potential for PFHxS to elicit
neurodevelopmental effects. Two of these studies were performed in vivo and five were performed
using in vitro models. Of the two in vivo studies, one was a follow-up to the Viberg etal. (2013)
study described above. Using the same study design as Viberg etal. (2013). and thus possessing the
same methodological limitations, Lee and Viberg f20131 examined changes in proteins24 involved
in a variety of neuronal functions in the cerebral cortex and hippocampus in NMRI male and female
mice at both 24 hours and 4 months following a single dose of PFHxS on PND 9 at either 6.2 mg/kg-
bw or 9.2 mg/kg-bw. While the authors observed significant changes in protein levels at 24 hours in
PFHxS-exposed animals the majority of these changes had resolved at the 4-month timepoint. At
4 months the only significant change was an increase in Tau protein expression (p <0.01) in the
cerebral cortex of male mice at the 6.1 mg/kg-bw dose.
PFHxS was also shown to produce a significant repression of long-term potentiation (LTP)
(p < 0.05), which is associated with learning and memory formation processes, in adult Sprague
Dawley rats exposed via intracerebroventricular injection at the CA1 region of the hippocampus
both 10 and 100 |iM PFHxS (Zhang et al.. 2016a). However, the authors noted no remarkable
changes in field excitatory postsynaptic potential (fEPSP) amplitude (decreased LTP would be
expected to represent weaker synaptic strength and reduced fEPSP) between control and PFHxS
treated groups (Zhang etal.. 2016a). In addition, this study was performed in adult rats therefore
making it difficult to determine how relevant the effects observed by Zhang etal. f2016al are to
human neurodevelopment
24BNDF: brain derived neurotrophic factor; protein involved in canonical nerve growth (Huang and
Reichardt. 2001): CaMKII: Ca2+/calmodulin dependent protein kinase II; a serine-threonine-specific protein
kinase that is regulated by Ca2/calmodulin. Involved in a variety of neuronal processes including learning and
memory (Yamauchi. 20051 GAP43: Growth Associated Protein 43; Protein expressed at high levels in neural
growth cones during development and axonal regeneration (Rosskothen-Kuhl and Illing. 2014)
Synaptophysin: protein present in the neuroendocrine cells involved in synaptic transmission (Mcmahon et
al.. 1996): Tau: Tau proteins are a group of six highly soluble protein isoforms that are produced by
alternative splicing. Tau proteins play a role in the stability of microtubules in axons and are present in
abundance in CNS neurons (Barbier et al.. 2019).
3-255
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence from animals prenatally exposed to other per and polyfluoroalkyl substances
(PFAS) such as PFOA and PFOS, suggest that PFAS may affect neurodevelopment fZhang etal..
2016b: Shresthaetal.. 2017: Salgado etal.. 2016: Lau etal.. 2003: Kawabataetal.. 2017: Fuentes et
al.. 20071. PFAS-related effects relevant to neurodevelopment include decreased choline
acetyltransferase activity in the prefrontal cortex of exposed rats postnatally fLau etal.. 20031.
delayed neuromotor maturation (e.g., decreased resistance to backward pull-on postnatal day
[PND] 10 and 11) (Fuentes etal.. 20071.
Evidence Integration
Taken together, the available human studies were interpreted to provide slight evidence.
Specifically, five medium confidence epidemiological studies that reported some evidence of
positive associations between PFHxS exposure and ADHD or behaviors potentially related to ADHD
at median blood concentrations in the study populations of 1-5 ng/mL. In addition, several
epidemiology studies examined whether PFHxS exposure has the potential to affect the following
neurodevelopmental outcomes: cognition, social behavior and autism, and other outcomes such as
motor-related behaviors and cerebral palsy. However, associations with these neurodevelopmental
outcomes were inconsistent across studies and generally imprecise with wide confidence intervals
and lack of statistical significance, and thus did not contribute to the overall judgment for potential
neurodevelopmental effects.
Th animal evidence base consisted of three studies examining PFHxS effect on FOB and
motor function, and a single study on PFHxS effects on learning and memory. PFHxS-related effects
in these studies were null or of low confidence. Additional animal studies potentially relevant to
interpreting the outcomes examined in the epidemiology studies of PFHxS were unavailable. Thus,
the overall animal evidence was considered indeterminate (see Table 3-27).
The endocrine and nervous systems work in harmony during early development. To this
end, evidence from the endocrine evidence base was also examined to see if any of the studies in
the endocrine database could help inform PFHxS neurotoxicity. While no studies evaluated both
endocrine and neurological outcomes as part of their study designs, the prior judgment that PFHxS
exposure is likely to result in decreased levels of serum thyroxine (T4)—particularly the evidence
after developmental PFHxS exposure (for more details please see Section 3.2.1), is of potential
relevance. In rats, decreased serum T4 is correlated with adverse neurodevelopmental outcomes
fCrofton. 20041. and, in humans, a link between prenatal maternal T4 and decreased cognitive
function in children has been observed (Man etal.. 1971: Li etal.. 2010: Henrichs etal.. 2013:
Haddowetal.. 1999: Finken etal.. 2013). The lack of neurological outcome measurements in the
available endocrine studies examining PFHxS-related toxicity highlights an important data gap.
3-256
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
The available evidence suggests but is not sufficient to infer whether exposure to PFHxS
might cause neurodevelopmental effects in humans given sufficient exposure conditions25 (see
Table 3-27). This conclusion is based on slight epidemiological evidence primarily from four
medium confidence epidemiological studies that reported some evidence of positive associations
between PFHxS exposure and ADHD or behaviors potentially related to ADHD at median blood
concentrations in the study populations of 1-5 ng/mL.
25The "sufficient exposure conditions" are more fully evaluated and defined for the identified health effects
through dose-response analysis in Section 5.
3-257
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-27. Evidence profile table for PFHxS neurotoxicological effects
Evidence stream summary and interpretation
Evidence integration
summary judgment
Evidence from studies of exposed humans (see Nervous System Human Studies Section)
Studies and confidence
Factors that increase
certainty
Factors that decrease
certainty
Summary of key findings
Evidence stream judgment
©OO
Evidence suggests, but is
not sufficient to infer
Primary basis:
Based on human evidence
for increased ADHD and
related behaviors at
median blood
concentrations of 0.9-
5 ng/mL
Human relevance:
Evidence comes from
epidemiological studies
(see Nervous System
Human Studies Section)
Cross-stream coherence:
NA: animal evidence is
indeterminate
Susceptible populations:
In utero or childhood
exposure.
• ADHD or related
behaviors
• 9 medium, 1 low
confidence studies
• Exposure-
response
gradients in
multiple studies
• Mostly medium
confidence
studies, with
positive
associations in 5
of 9
• Unexplained
inconsistency
• Unclear biological
relevance of
etiologic window
in cross-sectional
studies reporting
associations
5 medium and 1 low
confidence studies
reported positive
associations between
PFHxS exposure and ADHD
or behavior consistent with
ADHD.
©oo
Slight
Based on some evidence of
an association between
PFHxS exposure or ADHD
and related behaviors,
although uncertainty
remains. Other outcomes
did not contribute to this
judgment.
Cognition
• 9 medium and 1 low
confidence studies
• No factors noted
• Unexplained
inconsistency,
including by
timing of
Inverse associations
between cognition and
PFHxS exposure were
observed in multiple
3-258
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration
summary judgment
exposure
measurement.
studies, but there were
inconsistencies across
studies and in sub-analyses
within studies.
Social behavior or ASD
• 9 medium
confidence studies
• No factors noted
• Unexplained
inconsistency
• Imprecision
Of 5 studies of ASD, 2
reported higher likelihood
of diagnosis. Other studies
of social behavior were
similarly inconsistent.
Other
neurodevelopmental
effects
• 5 medium
confidence studies
• No factors noted.
• Unexplained
inconsistency for
motor-related
behaviors
• Imprecision for
cerebral palsy
2 medium confidence
studies reported a decrease
in motor scores with higher
PFHxS exposure, while
improved motor function
was observed in two
medium confidence studies.
A medium confidence study
reported a nonstatistically
significant positive
association with cerebral
palsy in boys.
Evidence from In vivo Animal Studies (see Nervous System Animal Studies Section)
Evidence stream judgment
Studies and confidence
Factors that increase
strength
Factors that decrease
strength
Summary of key findings
Behavioral
• 2 medium 1 low
confidence studies
• No factors noted
• Low confidence
study is only one
to observe an
effect
2 medium confidence
studies reported no effects
on FOB parameters, motor
activity, or learning and
memory. The low
confidence study observed
decreases in spontaneous
behaviors.
ooo
Indeterminate
1
3-259
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3.2.6. Cardiometabolic Effects
Cardiometabolic risk refers to the likelihood of developing diabetes, heart disease, or
stroke. Contributors to this risk include a combination of cardiovascular risk factors mainly
characterized by insulin resistance, dyslipidemia, hypertension, and adiposity (obesity).
Human Studies
Serum lipids
High serum levels of lipids, specifically low-density lipoprotein (LDL) cholesterol and
triglycerides, is one of the major controllable risk factors for cardiovascular disease, including
atherosclerosis, coronary heart disease, myocardial infarction, and stroke fZhang et al.. 2022a: Y et
al.. 2022: Wang etal.. 2024: Pappan etal.. 2024: Miller etal.. 2011: Linton etal.. 2000: Gad. 2015).
Cholesterol levels are typically measured in the blood. Thirty-eight studies evaluated the
relationship between PFHxS exposure and blood lipids (i.e., cholesterol, LDL cholesterol, and
triglycerides).
Multiple outcome-specific considerations for study evaluation influenced the ratings. First,
for outcome ascertainment, collection of blood during a fasting state is preferred for all blood lipid
measures fNIH. 2020: Nigam. 20111 but lack of fasting was considered deficient for triglycerides
and LDL cholesterol (which is typically calculated using triglycerides). This is because triglyceride
levels remain elevated for several hours after a meal (Nigam. 2011). which is likely to result in
substantial outcome misclassification if there is no standardization across study participants. Self-
reported high cholesterol was also considered deficient for outcome ascertainment due to the high
likelihood of misclassifying cases as controls fNataraian et al.. 20021. Both of these issues are likely
to result in nondifferential outcome misclassification and to generally bias results toward the null.
It is also important for studies to account for factors that meaningfully influence serum lipids, most
notably use of cholesterol-lowering medications and pregnancy. Studies that did not consider these
factors by exclusion, stratification, or adjustment were considered deficient for the participant
selection domain. All of the available studies analyzed serum lipids and PFHxS in serum or plasma
using standard, appropriate methods. As described in the Endocrine Effects section, reverse
causation was considered based on binding of lipophilic chemicals (such as PFAS) to serum lipids
fChevrier. 20131. but this is unlikely to significantly bias the results because PFAS, including PFHxS,
do not preferentially bind to serum lipids fForsthuber etal.. 20201. so exposure measurements in
blood, including cross-sectional, were considered adequate for this outcome.
A summary of the study evaluations is presented in Figure 3-68, and additional details can
be obtained from HAWC. Five studies were excluded from further analysis as uninformative due to
critical deficiencies confounding in four studies (Yang etal.. 2018: Tao etal.. 2008: Seo etal.. 2018:
Rotander etal.. 2015bl and selection bias in two studies fYang etal.. 2018: Sinisalu etal.. 20211.
Twenty-four studies were classified as medium confidence for at least one serum lipid measure
3-260
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
(Zeng etal.. 2015: Yang etal.. 2020b: Tian etal.. 2021: Starling et al.. 2014b: Spratlenetal.. 2020b:
Mora etal.. 2018: Matilla-Santander et al.. 2017: Manzano-Salgado etal.. 2017b: Liu etal.. 2020a:
Lin etal.. 2019: Li etal.. 2021a: Kang etal.. 2018: Tensen etal.. 2020a: Tain and Ducatman. 2018:
Gardener etal.. 2021: Dunder etal.. 2022: Dong etal.. 2019: D alia Zuanna et al.. 2 0 21: Canova etal..
2020: Canova etal.. 2021: Cakmak etal.. 2022: Blomberg etal.. 2021: Averina etal.. 20211. although
11 of these were low confidence for triglycerides (and LDL cholesterol when calculated from
triglycerides), as described above (Zeng etal.. 2015: Starling et al.. 2014b: Matilla-Santander et al..
2017: Manzano-Salgado etal.. 2017bl. Nine studies were classified as low confidence for all serum
lipid endpoints fVarshavsky et al.. 2021: Li etal.. 2020b: Koshv etal.. 2017: Khalil etal.. 2018: Khalil
etal.. 2020: Christensen etal.. 2016: Chen etal.. 2019a: Batzella et al.. 20221.
3-261
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
***
Averina, 2021, 7410155-
+
+
+
*
*
*
Legend
Q Good (metric! or Hiah confidence (overall)
Balzella, 2022, 10273294-
+*
-
* 1 Adequate (metric) or Medium confidence (overall)
Blomberg, 2021, 8442228 -
2
+*
•
B
Critically deficient (metric) or Uninformative (overall)
Cakmak, 2022, 10273369-
H
¦
Canova, 2020, 7021512-
+~
•
B
•
Canova, 2021, 10176518-
•
+*
+~
Chen, 2019, 5387400-
-
•
-
Christensen, 2016, 3858533 -
-
Dalla Zuanna, 2021, 7277682 -
•
•
-
*
-
Dong, 2019, 5080195-
-
•
-
-
*
Dunder, 2022, 10273372 -
g
~
~
Gardener. 2021,7021199-
1
-
•
B
-
Jain, 2018, 5079656-
1
-
•
-
•
Jensen et al., 2020, 6833719 -
*
+*
•
~
**
Kang, 2018,4937567-
D
-
Khalil, 2018, 4238547-
-
•
Khalil, 2020, 7021479-
-
-
Koshy, 2017, 4238478-
•
Li, 2020, 6315681-
~
+*
B
-
Li, 2021, 7404102-
D
•
*
•
Lin, 2019, 5187597-
*
B
•
•
Lin, 2020, 6988476-
Liu, 2020, 6318644-
~
•
B
•
•
Manzano-Salgado. 2017b, 4238509 -
~
+~
*
+*
Matilla-Santander, 2017,4238432 -
+*
-
+~
Mora, 2018, 4239224-
*
*
B
•
•
Papadopoulou, 2021, 9960593-
+*
f*
Rotander, 2015, 3859842 -
Seo, 2018, 4238334-
B
1
Sinisalu, 2021, 9959547-
1
¦
Spratlen, 2020, 5915332 -
~
+~
*
B
Starling, 2014, 2850928-
*
*
Tao, 2008, 2565188-
B
B
*
Tian, 2020, 7026251 -
*
*
B
•
-
Varshavsky, 2021, 7410195-
Yang, 2018, 4238462-
B
B
*
•
Yang, 2020, 7021246-
*
+~
-
+*
Zeng, 2015, 2851005-
*
Figure 3-68. Study evaluation results for epidemiology studies of PFHxS and
blood lipids. For additional details see HAWC. Multiple publications of the same
study: Canova et al. f20201 also includes Zare leddi et al. f20211: Cakmak et al.
(2022) also includes Fisher et al. (2013). Dong et al. (2019) also includes fNelson
et al.. 20101 fHe et al.. 20181 flain and Ducatman. 2019dl flain. 20131 flain.
20141 fChristensen etal.. 20191 and fFan etal.. 20201
3-262
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
The results for the association between PFHxS exposure and blood lipids are presented in
Table 3-28. It is difficultto directly compare the magnitudes of effect across studies due to the
different analyses and data transformations (e.g., log transformations of PFHxS levels and/or lipid
levels), so the synthesis is focused primarily on direction of association.
In adults, all six medium confidence studies (reported in eight publications) examining total
cholesterol reported positive associations between total cholesterol and PFHxS exposure (Liu etal..
2020a: Lin etal.. 2019: Dunder etal.. 2022: Dong etal.. 2019: Canova etal.. 2020: Cakmak etal..
2022). with statistical significance in four (Lin etal.. 2019: Dunder etal.. 2022: Canova etal.. 2020:
Cakmak etal.. 20221. In the four studies that additionally examined exposure modeled as quartiles,
three reported a monotonic exposure-response gradient fLiu etal.. 2020a: Fisher etal.. 2013:
Canova etal.. 20201. while one reported the strongest association in the third quartile fLin etal..
20191. While the direction of association was mostly consistent across studies, in the NHANES data
reported in Dong etal. (2019). the direction of association was not consistent across NHANES study
cycles. The association was inverse (not statistically significant) in 2003-2004 and 2005-2006, but
positive (not statistically significant) in 2007-2008, 2011-2012, and 2013-2014, despite similar
exposure levels across cycles. Further, in the two studies with prospective exposure measurement,
only one found a positive association fDunder etal.. 20221. while the other found an association in
cross-sectional but not prospective analyses fLin etal.. 20191. The lack of consistent associations in
studies with prospective exposure measurement reduces certainty since these studies reduce the
likelihood of reverse causality and other sources of bias compared with cross-sectional studies.
Two low confidence studies (Li etal.. 2020b: Chen etal.. 2019a) in general population adults
also observed positive associations with total cholesterol, with the latter being statistically
significant, while a third low confidence study fLin etal.. 2020cl found no association in older
residents (55-75 years). The populations in both Lin etal. f2020cl and Li etal. f2020bl were living
in high contamination areas (in Taiwan and Sweden, respectively). In addition, two studies
examined occupational populations with PFAS exposure. These studies were low confidence due to
concerns for potential selection bias and residual confounding. Batzella et al. (2022). examining
PFAS production workers in Italy, and Khalil etal. (2020) examining firefighters in the U.S., both
reported positive, but not statistically significant associations between PFHxS and total cholesterol.
In pregnant women, two studies fYang etal.. 2020b: Starling etal.. 2014bl out of five (see
Table 3-28) reported higher total cholesterol with higher PFHxS exposure, with statistical
significance in Yang etal. f2020bl and an exposure-response gradient across quartiles in Starling et
al. (2014b). In a low confidence study of high cholesterol (Christensen etal.. 2016). no association
was observed (OR 1.01, 95% CI: 0.91,1.13), but the study is expected to be biased toward the null
due to nondifferential outcome misclassification.
Three of the medium confidence studies additionally reported analyses of dichotomous
hypercholesterolemia fLin etal.. 2019: Fisher etal.. 2013: Canova et al.. 20201. Cutoffs for high
cholesterol differed across studies: in Fisher etal. f20131 the cutoff for total cholesterol was 5.2
3-263
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
mmol/L; in Canova etal. (2020). the cutoff was 190 mg/mL, and in Lin etal. (2019). the outcome
was initiation of cholesterol-lowering medication, or total cholesterol of 240 mg/mL/LDL cutoff of
160 ng/mL). Significantly higher odds of high cholesterol (OR of 1.4-1.6 in the highest quartiles)
were reported in both Fisher etal. f20131 and Canova etal. f20201. with a monotonic exposure-
response gradient across quartiles. In Lin etal. f 20191. higher odds (not statistically significant)
were observed in an analysis of high cholesterol at baseline, but not when risk of high cholesterol
was analyzed prospectively.
Results for LDL cholesterol and triglycerides in adults were less consistent than total
cholesterol in the medium confidence studies, with most studies showing similar results across the
different outcome markers, but a few reporting inverse associations for LDL and/or triglycerides
fMatilla-Santander etal.. 2017: DaliaZuannaetal.. 2021: Cakmaketal.. 20221.
In adolescents and children, there was very limited evidence of an association, with 4 of 12
medium confidence studies reporting higher total cholesterol with higher PFHxS exposure (Zeng et
al.. 2015: Mora etal.. 2018: Kang etal.. 2018: Canova etal.. 2021). and only one reporting
statistically significance, but without an exposure-response gradient across quartiles (Canova etal..
20211. The other medium confidence studies reported no association fPapadopoulou etal.. 2021:
Manzano-Salgado etal.. 2017b: Tensen etal.. 2020a: Tain and Ducatman. 2018: Blomberg etal..
2021: Averina etal.. 20211. For triglycerides, 4 of 12 studies reported positive associations fZeng et
al.. 2015: Spratlen etal.. 2020b: Manzano-Salgado etal.. 2017b: Blomberg etal.. 2021). Of note, both
Spratlen et al. (2020b) and (Blomberg etal.. 2021) reported statistically significant positive
associations in neonates, though the third study in neonates found no association (Tian etal.. 2020).
Looking at the two studies of low confidence in adolescents (Koshv etal.. 2017) and children (Khalil
etal.. 20181. both reported higher total cholesterol with higher exposure, with the difference being
statistically significant in Koshv etal. f20171. but both had serious limitations.
Overall, there is some evidence that higher PFHxS exposure is associated with higher total
cholesterol levels in adults, with less consistent evidence for parallel changes in triglycerides. The
majority of studies in adults, including pregnant women, support this association, though there are
remaining uncertainties, including less consistent evidence for LDL cholesterol and triglycerides.
In addition, there is potential for confounding across the PFAS. In the studies with stronger
associations, there were similar associations with other PFAS, including PFOS, PFOA, and PFNA, and
PFHxS is moderately positively correlated with them. Only a minority of studies that observed
positive associations with serum lipids presented mixture modeling results that allow for
interpretation of individual PFAS contributions. In all of these studies, other PFAS had higher
weight in the mixture or results for PFHxS were attenuated after adjustment. Fan etal. (2020). an
analysis of NHANES data, found that PFNA and PFOS had the highest weights for total cholesterol,
LDL cholesterol, and triglycerides, while PFHxS had the highest weight for only HDL cholesterol.
Batzella et al. f20221 found that PFNA had the highest weight (0.48), with PFHxS second (0.38). In
Starling etal. f2014bl. the nonsignificant association with PFHxS was attenuated to null with
3-264
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
simultaneous adjustment for other PFAS. In studies where a clear association with PFHxS was not
observed, mixture modeling similarly found that other PFAS were likely drivers flain and
Ducatman. 2018: Averina etal.. 20211. which is consistent with expectations. Conversely the
association with cholesterol was still present in a study with weak correlations (~0.3) between
PFHxS and PFOS and PFOA fCakmak etal.. 20221. Still, based on the mixture analyses, confounding
by other PFAS is a substantial source of uncertainty in this evidence base. While the mixture
modeling results do not rule out a possible association between PFHxS exposure and serum lipids,
they do raise the likelihood that the observed associations may be explained by confounding.
Overall, given the general consistency across studies and the observation of exposure-
response gradients across quartiles in multiple studies, there is reasonable support for a positive
association with this outcome, but there is considerable remaining uncertainty due to substantial
concern for potential confounding by other PFAS.
Table 3-28. Associations between PFHxS exposure and blood lipids in medium
confidence epidemiology studies
Reference
Population
Median
exposure (IQR)
or as specified
(ng/mL)
Effect
estimate
Total cholesterol3
LDLa
Triglycerides3
General population, adults
Dong et al.
(2019)
NHANES
cross-
sectional
(2003-2014
pooled), U.S.;
8,950 adults
1.6
P (95% CI)
for 1 unit
increase
0.98 (-0.89, 2.85)
0.72 (-1.63,
3.06)
NR
Fisher et al.
(2013)
Cakmak et al.
(2022)
Canadian
Health
Measures
Survey
(2007-2009)
cross-
sectional,
Canada; 2,345
adults
2.2 (1.2-3.6)
P (95% CI)
for 1 log-
unit
increase
0.03 (0.01,0.05)*
0.06
(0.01,0.11)*
0.02
(-0.02,0.06)
OR (95% CI)
for high
cholesterol
vs. Q1
Q2: 1.05 (0.69,1.61)
Q3: 1.43 (0.85,1.4)
Q4: 1.57 (0.93, 2.64)
p-trend: 0.001*
NR
NR
(2007-2017);
6,045
participants
1.5 (GM)
% change
for increase
equivalent
to GM
2.8(1.1, 4.5)*
-3.8 (-9, 1.7)
-1.4 (-5.0, 2.3)
Lin et al.
(2019)
Participants
from
randomized
trial of
2.3 (1.4-3.8)
Mean diff
(95% CI) for
twofold
increase
2.24 (0.15, 4.33)*
1.32 (-0.59,
3.22)
3.91 (-1.77,
9.59)
3-265
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median
exposure (IQR)
or as specified
(ng/mL)
Effect
estimate
Total cholesterol3
LDLa
Triglycerides3
diabetes
prevention,
U.S.; 888
overweight
and
prediabetic
adults
quartiles vs.
Q1
Q2: 3.87 (-2.89,
10.63)
Q3: 9.28 (2.38,
16.19)*
Q4: 7.43 (0.53,
14.33)*
Q2: 1.22
(-4.94, 7.38)
Q3: 6.22
(-0.06,12.52)
Q4: 3.88
(-2.39,10.17)
Q2: 9.64
(-8.75, 28.03)
Q3: 16.43
(-2.34, 35.22)
Q4: 11.23
(-7.52, 29.99)
Cross-
sectional
OR (95% CI)
for high
lipids
1.08 (0.94,1.25)
NR
1.03 (0.90,
1.18)
Prospective
HR (95%)
for high
lipids
Total: 1.00 0.92
(1.09)
Placebo: 1.02 (0.89,
1.17)
Lifestyle: 1.02 (0.90,
1.15)
NR
Total: 1.14
(1.00,1.28)*
Placebo: 1.23
(1.03,1.47)*
Lifestyle: 1.19
(0.98,1.44)
Liu et al.
(2020a)
Cross-
sectional
analysis from
randomized
clinical trial of
weight loss;
326
overweight
adults
2.4 (1.6-3.6)
Means ± SE
for tertiles
Tl: 181.6 ±7.8
T2: 189.3 ±7.6
T3: 192.5 ±7.8
p-trend = 0.15
NR
Tl:
119.4 ± 11.2
T2:
133.6 ± 11.0
T3:
130.8 ± 11.2
p-trend = 0.37
Dunder et al.
(2022)
Cohort study
(2001-2004),
Sweden; 864
older adults
(70 yr at
baseline)
3.1 (2.0-5.8)
P (95% CI)
for In-unit
increase
(for lipids
over 10 yr)
0.08 (0.01, 0.15)*
0.04 (-0.01,
0.10)
0.04 (0.01,
0.07)*
Canova et al.
(2020)
Cross-
sectional
study in highly
contaminated
area (2017-
2019), Italy;
15,720 young
adults (20—
39 yr)
3.6 (1.6-7.8)
P (95% CI)
for In-unit
increase
2.02 (1.45, 2.58)*
(exposure-response
gradient across
quartiles)
1.31 (0.81,
1.8)*
0.02 (0.01,
0.02)*b
OR (95% CI)
vs. Q1 for
abnormal
lipids
Q2: 1.18 (1.06,
1.30)*
Q3: 1.19 (1.07,
1.32)*
Q4: 1.41 (1.25,
1.58)*
Q2: 1.21 (1.08,
1.35)*
Q3: 1.15 (1.02,
1.29)*
Q4: 1.37 (1.20,
1.55)*
Q2: 1.11 (0.93,
1.32)
Q3: 1.17 (0.98,
1.40)
Q4: 1.22 (1.02,
1.46)*b
Pregnant women
3-266
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median
exposure (IQR)
or as specified
(ng/mL)
Effect
estimate
Total cholesterol3
LDLa
Triglycerides3
Yang et al.
(2020b)
Pregnancy
cohort (2013-
2014), China,
436 women
0.3 (0.2-0.5)
P (95% CI)
for In-unit
increase
0.18 (0.05, 0.32)*
0.09 (0.001,
0.19)*
0.07 (-0.1,
0.24)b
Gardener et al.
(2021)
Pregnancy
cohort (2009),
U.S., 433
women
0.5 (0.3-0.9)
Means ± CI
for quartiles
No clear association
(reported only on
figure)
NR
No clear
association
(reported only
on figure)
Starling et al.
(2014b)
Norwegian
Mother and
Child cross-
sectional
analysis
(2003-2004),
Norway; 891
women
0.6 (0.4-0.9)
P (95% CI)
for In-unit
increase
3.00 (-1.75,7.76)
1.92 (-2.50,
6.33)b
-0.01 (-0.05,
0.03)b
quartiles vs.
Q1
Q2: 0.65
(-6.87,8.17)
Q3: 1.62
(-6.08,9.32)
Q4: 4.25
(-3.88,12.39)
Q2: 0.44
(-6.19, 7.08)
Q3: 0.50
(-6.15, 7.16)
Q4: 1.48
(-5.89, 8.85)b
Q2: -0.04
(-0.11, 0.02)
Q3: -0.02
(-0.10, 0.05)
Q4: -0.02
(-0.09, 0.05)b
Matilla-
Santander et
al. (2017)
INMA cross-
sectional
analysis
(2003-2008),
Spain; 1,240
women
0.6 (0.4-0.8)
% change
(95% CI) for
log-unit
increase
-0.09 (-8.25, 1.45)
NR
-4.90 (-9.16,
-0.72)*b
quartiles vs.
Q1
Q2: 1.21 (-1.05,
3.45)
Q3: 0.60 (-1.69,
2.94)
Q4: 0.70 (-1.86,
3.38)
NR
Q2: -7.69
(-14.3, -1.00)
Q3: -3.92
(-10.9, 3.05)
Q4: -7.69
(-13.9, 1.40)b
Dalla Zuanna
et al. (2021)
Cross-
sectional
study in highly
contaminated
area (2017-
2020), Italy;
319 women
2.1 (1.1-4.1)
P (95% CI)
for In-unit
increase
-4.91 (-10.06, 0.24)
-8.17 (-12.54,
-3.81)*
NR
Adolescents and children
Blomberg et
al. (2021)
(additional
results with
different
timing of
Birth cohort
(2007-2009),
Faroe Islands,
459 children
(followed to
9 yr)
0.2 (0.1-0.2)
P (95% CI)
for doubling
PFAS and
lipids at
birth
Overall
0.03 (-0.04,0.09)
Girls
0.05 (-0.03,0.14)
Boys
-.003 (-0.1, 0.09)
Overall
0.01 (-0.03,
0.05)
Girls
0.019 (-0.03,
0.07)
Boys
Overall
11(5.9,17)*b
Girls
13 (5.5,21)*
Boys
9.7 (1.9, 18)*
3-267
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Median
exposure (IQR)
or as specified
Effect
Reference
Population
(ng/mL)
estimate
Total cholesterol3
LDLa
Triglycerides3
exposure and
-0.01 (-0.06,
outcome
0.05)
measurement
are available
in the
publication)
PFAS at
Overall
Overall
Overall
birth and
lipids at 18
mo
-0.04 (-0.18, 0.1)
Girls
-0.03 (-0.22, 0.17)
Boys
-0.05 (-0.24, 0.15)
-0.05 (-0.15,
0.06)
Girls
-0.05 (-0.2,
0.1)
Boys
-0.04 (-0.19,
0.12)
3.5 (-3.9,11)
Girls
7.9 (-2.5,19)
Boys
-0.87 (-11,
9.9)
PFAS and
Overall
Overall
Overall
lipids at 9 yr
-0.02 (-0.14, 0.1)
Girls
-0.05 (-0.21,0.1)
Boys
0.02 (-0.15,0.19)
-0.06 (-0.14,
0.03)
Girls
-0.06 (-0.18,
0.06)
Boys
-0.05 (-0.18,
0.08)
-1.8 (-8.3, 5.2)
Girls
2.6 (-6.3,12)
Boys
-6.8 (-16, 3)
Jensen et al.
Birth cohort
0.3
P (95% CI)
3 mo
3 mo
3 mo
(2020a)
(2010-2012),
(5th—95th:
for 1 unit
-0.08 (-0.33, 0.17)
0.01 (-0.24,
0.18 (-0.07,
Denmark; 612
0.1-0.7)
increase
Girls
0.26)
0.44)
children
-0.11 (-0.37, 0.16)
Girls
Girls
(followed to
Boys
0.05 (-0.22,
0.21 (-0.06,
18 mo)
0.13 (-0.58,0.85)
18 mo
-0.06 (-0.32, 0.21)
Girls
-0.05 (-0.32, 0.21)
Boys
-0.10 (-1.41, 1.21)
0.32)
Boys
-0.28 (-1.01,
0.44)
18 mo
-0.06 (-0.35,
0.22)
Girls
-0.08 (-0.37,
0.21)
Boys
0.37 (-1.02,
1.76)b
0.48)
Boys
-0.02 (-0.75,
0.71)
18 mo
-0.24 (-0.51,
0.04)
Girls
-0.22 (-0.50,
0.06)
Boys
-0.62 (-1.95,
0.70)b
PaoadoDoulou
Six birth
prenatal
P (95% CI)
NR
0.03 (-0.03,
0.02 (-0.05,
et al. (2021)
cohorts,
0.5 (0.3-0.9)
for doubling
0.09)b
0.08)b
Europe, 1,301
children (6-
11 yr)
Children
0.3 (0.2-0.6)
exposure
NR
0.02 (-0.06,
0.10)b
0.00 (-0.08,
0.08)b
3-268
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median
exposure (IQR)
or as specified
(ng/mL)
Effect
estimate
Total cholesterol3
LDLa
Triglycerides3
Manzano-
Salgado et al.
(2017b)-
INMA cohort
(2003-2008),
Spain; 627
children (4 yr)
prenatal
0.6 (0.4-0.8)
(GM (IQR))
P (95% CI)
for doubling
exposure
and
cholesterol
z-score
0.02 (-0.09,0.12)
Boys: -0.02
(-0.17,0.13)
Girls: 0.04
(-0.12,0.20)
-0.01 (-0.12,
0.09)b
Boys: -0.04
(-0.18, 0.10)
Girls: 0.00
(-0.15, 0.15)
0.11 (-0.01,
0.21)b
Boys: 0.16
(0.03,0.30)*
Girls: 0.07
(-0.08, 0.22)
SDratlen et al.
(2020b)
WTC cohort
(2001-2002),
U.S.; 222
newborns
cord blood 0.7
(0.5-1.0)
%
difference
for 1%
increase
0.03 (-0.02, 0.08)
NR
0.13 (-0.04,
0.23)
Mean ratio
vs. Q1
Q2: 1.03 (0.94, 1.12)
Q3: 1.06 (0.98, 1.16)
Q4: 1.07 (0.98, 1.16)
p-trend 0.5
NR
Q2: 1.08 (-.92,
1.28)
Q3: 1.22 (1.04,
1.45)
Q4: 1.26 (1.07,
1.49)
p-trend 0.002
Kang et al.
(2018)
Korea
Environmenta
1 Health
Survey in
Children and
Adolescents
cross-
sectional
analysis
(2012-2014),
Korea, 150
children (3-
18 yr)
0.8 (0.6-1.0)
P (95% CI)
for In-unit
increase
0.99 (-9.53, 11.50)
-4.22 (-13.98,
5.53)
0.08 (-0.09,
0.25)
Averina et al.
(2021)
Cross-
sectional
study (2010-
2011),
Norway, 940
adolescents
(~16 yr)
Girls 0.8,
Boys 1.0
(GMs)
P (95% CI)
for log-unit
increase
"No association"
(data not shown)
"No
association"
(data not
shown)
"No
association"
(data not
shown)
Jain and
Ducatman
(2018)
NHANES
cross-
sectional
(2013-2014),
U.S.; 458
children (6-
11 yr)
0.9
Means (95%
CI)
Ql: 154 (149-159)
Q2: 159 (155-163)
Q3: 153 (145-161)
Q4: 158 (153-164)
p = 0.4
NR
NR
3-269
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median
exposure (IQR)
or as specified
(ng/mL)
Effect
estimate
Total cholesterol3
LDLa
Triglycerides3
Zeng et al.
Genetic and
Biomarkers
study for
Childhood
Asthma cross-
sectional
analysis
(2009-2010),
Taiwan; 225
adolescents
(12-15 yr)
1.2 (range 0.2-
10.3) (boys)
P (95% CI)
for 1 unit
increase
1.10 (-0.71,2.92)
0.99 (-0.41,
2.39)b
1.80 (-0.67,
4.27)b
(2015)
Li et al.
HOME cohort
(2003-2006);
U.S.; 186
adolescents
(12 yr)
prenatal 1.3
(0.8-2.3)
Difference
for IQR
increase
NR
NR
0.1(0.0, 0.2)
(2021a)
birth 0.6 (0.4-
1.0)
NR
NR
0.1 (-0.1, 0.3)
Mora et al.
Project Viva
cohort (1999—
2002), U.S.;
682 children
(7-8 yr)
prenatal
2.4 (1.6-3.8)
P (95% CI)
for IQR
increase
0.5 (-1.1,2.2)
similar for boys and
girls
0.5 (-0.9,1.9)
similar for boys
and girls
-0.6 (-2.0,0.8)
Boys: 0.6
(-1.9,3.1)
Girls: -1.1
(-3.1,0.1)
(2018)
child
1.9 (1.2-3.4)
-0.3 (-1.0,0.5)
Boys: -0.5 (-1.5,0.4)
Girls: 0.2 (-1.0,1.3)
-0.2 (-0.9,0.4)
Boys: -0.5
(-1.4,0.3)
Girls: 0.3
(-0.6,1.3)
-0.4 (-1.0,0.3)
similar for boys
and girls
Tian et al.
Birth cohort
(2012), China;
306 newborns
prenatal
2.7 (2.0-3.5)
P (95% CI)
for In-unit
increase
0.05 (-0.07, 0.16)
0.03 (-0.11,
0.18)
0.02 (-0.11,
0.15)
(2021)
Canova et al.
Cross-
sectional
study in highly
contaminated
area (2017-
2019), Italy;
6,669
adolescents
(14-19 yr)
and 2,693
children (8-
11 yr)
adolescents
2.8 (1.6-4.8)
P (95% CI)
for In-unit
increase
1.49 (0.60, 2.37)
1.44 (0.68,
2.19)
0.01 (-0.01,
0.02)b
(2021)
P (95% CI)
vs. Q1
Q2: 1.96 (0.20,
3.73)*
Q3: 1.72 (-0.10,
3.54)
Q4: 3.80 (1.83,
5.77)*
Q2: 2.03 (0.52,
3.55)*
Q3: 1.60 (0.05,
3.16)*
Q4: 3.65 (1.97,
5.33)
Q2: 0.01
(-0.02, 0.04)
Q3: 0.00
(-0.03, 0.03)
Q4: 0.02
(-0.02, 0.05)
children
1.9 (1.2-2.8)
P (95% CI)
for In-unit
increase
1.30 (-0.28, 2.88)
0.54 (-0.87,
1.96)
-0.01 (-0.03,
0.01)
P (95% CI)
vs. Q1
Q2: 0.46 (-0.73,
1.65)
Q2: -1.70
(-4.19, 0.8)
Q2: 0 (-0.04,
0.04)
3-270
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median
exposure (IQR)
or as specified
(ng/mL)
Effect
estimate
Total cholesterol3
LDLa
Triglycerides3
Q3: 1.68 (0.44,
2.91)*
Q4: 1.32 (0.07,
2.56)*
Q3: -1.22
(-3.81, 1.38)
Q4: 0.76
(-1.86, 3.39)
Q3: 0 (-0.04,
0.04)
Q4: -0.02
(-0.07, 0.02)
*p < 0.05.
NR = not reported.
aUnits and transformations of outcome variables varied across studies.
bLow confidence endpoint within medium confidence study.
Other risk factors for cardiovascular disease
Twenty-seven studies report on the association between PFHxS exposure and other risk
factors for cardiovascular disease, including blood pressure in the general population (18 studies),
blood pressure and hypertensive disorders during pregnancy (6 studies), atherosclerosis (2
studies), abdominal aortic calcification (1 study), and ventricular geometry (1 study). The study
evaluations for these outcomes are summarized in Figure 3-69. One study was considered high
confidence, 18 were medium confidence, and 7 were low confidence. One study (Yang etal.. 2018)
evaluating blood pressure was excluded from further analysis (uninformative) due to critical
deficiencies in participant selection and confounding.
Considering blood pressure in the general population, the majority of studies reported no
association between PFHxS exposure and higher blood pressure. A few positive associations with
hypertension or higher blood pressure were observed in studies of adolescents and young adults
(see Table 3-29). Statistically significant associations were reported in a cross-sectional study of
16-year-olds in Norway (Averina et al.. 2021) and a cohort with follow-up to 12 years of age in the
U.S. (Li etal.. 2021a). though the association was not monotonic across quartiles in Averina et al.
f20211. In a region of Italy with high PFAS contamination, a positive association was observed in
young adults aged 20-39 years (Pitter etal.. 2020) but not adolescents aged 14-19 years (Canova et
al.. 20211. Studies in non-age restricted adults fLiu etal.. 2018: Lin etal.. 2020b: Christensen etal..
2016: Christensen etal.. 2019: Chen etal.. 2019a: Bao etal.. 20171 and children fPapadopoulou et
al.. 2021: Manzano-Salgado etal.. 2017b: Khalil etal.. 2018) reported null findings with blood
pressure and/or odds of hypertension, and a biological explanation is not clear for this pattern of
results by age.
Results for hypertensive disorders of pregnancy are summarized in Table 3-30. One of four
studies of gestational hypertension Borghese etal. (2020) and two of four studies of preeclampsia
fBorghese etal.. 2020: Birukov etal.. 20211 reported positive associations, with statistical
significance in one. Conversely, two studies reported inverse associations (statistically significant in
one) with gestational hypertension (Liu etal.. 2021a: Huang et al.. 2019c). The other one study of
gestational hypertension (Birukov etal.. 2021) and two studies of preeclampsia (Starling etal..
3-271
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
2014a: Huang etal.. 2019c) reported no association. One low confidence study reported no
association between PFHxS and continuous blood pressure during pregnancy fVarshavsky et al..
20211.
No association with PFHxS exposure was observed in studies of atherosclerosis in adults
fLind etal. f 20171. medium confidence) and markers of atherosclerosis/arterial wall stiffness in
adolescents (Koshvetal. (2017). low confidence). One study examining abdominal aortic
calcification, a marker of subclinical atherosclerotic disease, reported a positive, though not
statistically significant, association in men but not women fKoskela etal.. 2022). Lastly, no
association was observed in a single medium confidence study of ventricular geometry fMobacke et
al.. 20181.
3-272
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Averina, 2021, 7410155
Bangma, 2020, 6833725 -
Bao, 2017, 3860099
Batzella, 2022, 10273294-
Birukov, 2021, 7410153
Borghese, 2020, 6833656 -
Canova, 2021, 10176518
Chen, 2019, 5387400-
Christensen, 2016, 3858533
Christensen, 2019, 5080398
Huang, 2019, 5083564-
Khalil, 2018, 4238547-
Koshy, 2017, 4238478-
Koskela A et al. 2022 -
Li, 2021, 7404102-
Lin, 2020, 6311641
Lind, 2017, 3858504-
Liu, 2018, 4238396
Liu, 2021, 9944393 -
Manzano-Salgado, 2017b, 4238509
Mobacke, 2018, 4354163-
Papadopoulou, 2021, 9960593
Pitter G, 2020, 6988479 -
Starling, 2014, 2446669-
Varshavsky, 2021, 7410195-
Yang, 2018, 4238462
Zare Jeddi, 2021, 7404065 -
Legend
I Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Figure 3-69. Study evaluation results for epidemiology studies of PFHxS and
cardiovascular disease risk factors. For additional details see HAWC link. Multiple
publications of the same study: Christensen etal. T2019) also includes Liao et al.
f2020Y
3-273
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-29. Associations between PFHxS exposure and hypertension in
medium confidence epidemiology studies in adolescents and young adults
Reference
confidence
Population
Median exposure
(IQR) or as
specified
(Hg/mL)
Effect estimate
Hypertension
Averina et al. (2021)
Cross-sectional study in
Norway; 940 adolescents
(~16 yr)
0.8 (GM in girls)
OR (95% CI) for
quartiles vs. Q1
Q2: 1.63 (0.90, 2.94)
Q3: 1.25 (0.69, 2.28)
Q4: 2.06 (1.16, 3.65)*
Li et al. (2021a)
Cohort in U.S.; 221
adolescents (follow-up
through 12 yr)
1.2 (0.9,1.8) at 8 yr
Difference for
IQR increase
(outcome
continuous blood
pressure z-score)
Systolic BP
0.2 (0.0, 0.4)*
Canova et al. (2021)
Cross-sectional study in
highly PFAS exposed
region, Italy; 6,669
adolescents (14-19 yr)
2.8(1.6-4.8)
P (95% CI) for In-
unit increase
(outcome
continuous blood
pressure)
Systolic BP
-0.22 (-0.65, 0.21)
Diastolic BP
-0.15 (-0.45, 0.16)
Pitter et al. (2020)
Cross-sectional study in
highly PFAS exposed
region, Italy; 15,786
adults (20-39 yr)
6.0 (mean)
OR (95% CI) for
quartiles vs. Q1
Q2: 1.01 (0.86, 1.19)
Q3: 1.08 (0.92, 1.27)
Q4: 1.19 (1.00, 1.41)
* p < 0.05.
Table 3-30. Associations between PFHxS exposure and gestational
hypertension and preeclampsia in medium confidence epidemiology studies
Reference
Population
Median
exposure
in ng/mL
(IQR)
Effect estimate
Gestational
hypertension
Preeclampsia
Liu et al.
(2021a)
Nested case-control
study within cohort
in China; 544 women
0.1 (0.03,
0.1)
OR (95% CI) for
tertiles vs. T1
T2: 0.41 (0.25, 0.67)*
T3: 0.29 (0.17,0.50)*
NR
Huang et al.
(2019c)
Cross-sectional study
in China; 674 women
at delivery
0.2 (0.1-0.2)
OR (95% CI) for
tertiles vs. T1
T2: 0.83 (0.31,2.22)
T3: 0.48 (0.16,1.43)
T2: 1.10 (0.36, 3.38)
T3: 0.80 (0.25,2.60)
Birukov et al.
(2021)
Cohort in Denmark;
1,436 women
0.4 (0.3-0.5)
HR (95% CI) for
doubling of
exposure
0.97 (0.66, 1.43)
1.14 (0.91, 1.42)
Starling et al.
(2014a)
Nested case-control
study within cohort
in Norway; 1,046
women
0.7 (0.5-1.0)
HR (95 CI) for
quartiles vs. Q1
NR
Q2: 0.86 (0.59, 1.26)
Q3: 1.01 (0.69, 1.49)
Q4: 0.93 (0.64, 1.36)
3-274
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median
exposure
in ng/mL
(IQR)
Effect estimate
Gestational
hypertension
Preeclampsia
Borghese et
Cohort in Canada;
1,739 women
1.0 (0.7-1.6)
OR (95% CI) for
tertiles vs. T1
T2: 1.03 (0.64,1.67)
T3: 1.39 (0.87, 2.20)
T2: 1.40 (0.54, 3.63)
T3: 3.06 (1.27, 7.39)*
al. (2020)
*p < 0.05.
Cardiovascular disease
Five studies report on the association between PFHxS and cardiovascular disease, including
coronary heart disease, myocardial infarction (heart attack), and congestive heart failure. The study
evaluations are summarized in Figure 3-70. Two studies, an analysis of NHANES data for 1999-
2014 and a prospective cohort of farmers and other rural residents, were medium confidence
(Mattsson etal.. 2015: Huang etal.. 2018). The other three were low confidence (Honda-Kohmo et
al.. 2019: Graber etal.. 2019: Christensen etal.. 20161. These cross-sectional studies were focused
on very specific populations—participants in litigation over PFAS exposure f Honda-Kohmo etal..
2019: Graber etal.. 20191 or anglers fChristensen etal.. 20161. There were concerns about
confounding in all of these studies, and for sensitivity in Graber etal. f20191 and Christensen et al.
(2016) due to small sample size. Additionally, all the studies except Mattsson etal. (2015)—which
used a national register of disease—classified cardiovascular disease based on self-report on
questionnaires, which is likely to suffer from misclassification and which could be differential in
studies wherein exposure was known due to litigation fHonda-Kohmo etal.. 2019: Graber etal..
20191 but is likely nondifferential and thus toward the null in the other studies f Huang etal.. 2018:
Christensen etal.. 20161.
In the two medium confidence studies, no association between PFHxS exposure and
coronary heart disease (Mattsson et al.. 2 015: Huang etal.. 2018) or total cardiovascular disease,
congestive heart failure, coronary heart disease, angina pectoris, myocardial infarction, or stroke
(Huang etal.. 2018) was observed. In the low confidence studies, one reported higher odds of
cardiovascular conditions with higher exposure f Graber etal.. 20191 and two reported lower odds
of coronary heart disease fHonda-Kohmo etal.. 2019: Christensen etal.. 20161. although only
results in Honda-Kohmo etal. f20191 were statistically significant. An exposure-response gradient
was observed in Honda-Kohmo etal. f20191 across quantiles.
3-275
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Christensen, 2016, 3858533-
Honda-Kohmo, 2019, 5080551
Huang, 2018, 5024212
Mattsson, 2015, 3859607 -
tP
- -
+
-
+
+
-
- +
-
-
-
+
-
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)
- +
-
-
-
+
+
-
B
~
+
-
+
+
*
-
++
flfl
+
+
Figure 3-70. Study evaluation results for epidemiology studies of PFHxS and
cardiovascular disease. For additional details see HAWC link.
Summary of cardiovascular effects
Overall, there is some evidence of an association between PFHxS exposure and serum lipids.
However, the evidence for other cardiovascular-related effects is mostly null, which indicates that
changes in serum lipids do not appear to be contributing to increased cardiovascular disease risk at
the exposure concentrations observed in these populations. However, the association with serum
lipids without an accompanying increase in disease should not be considered inconsequential given
that it will likely lead to more people taking cholesterol-lowering medications. Further, given that
serum lipids are metabolized in the liver, the changes may be supportive of hepatic (see
Section 3.2.4) rather than cardiovascular toxicity.
Metabolic effects
Diabetes
Seven studies (reported in seven publications) report on the relationship between PFHxS
exposure and diabetes (i.e., type 2 diabetes). In cross-sectional studies of PFHxS and diabetes
outcomes, there is some concern for reverse causality. Metabolic changes related to diabetes (e.g.,
impairments of renal function) may affect the amount of PFHxS measured in blood. Four out of the
seven available studies were cross-sectional and were considered low confidence studies due to
temporality and other deficiencies as noted in HAWC. Three studies (Sun etal.. 2018: Charles etal..
2020: Cardenas etal.. 2017) had prospective exposure measurement prior to development of
diabetes. Sun etal. (2018) and Charles etal. (2020) used nested case-control study designs and
Cardenas etal. f20171 used a multicenter randomized clinical trial of a diabetes prevention lifestyle
intervention. Thus, these three studies were evaluated as medium confidence. A summary of the
study evaluations for PFHxS and diabetes is presented in Figure 3-71, and additional details of the
studies can be obtained from HAWC.
3-276
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Cardenas, 2017, 4167229-
Charles, 2020, 7068869-
Conway, 2016, 3859824-
He, 2018, 4238388-
Lind, 2014, 2215376-
Sun, 2018, 4241053-
Zare Jeddi, 2021, 7404065 -
+
+
+ +
+
+
+
++
+ +
~
+
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 3-71. Summary of study evaluation for PFHxS and type 2 diabetes in
epidemiology studies. For additional details see HAWC link. Multiple publications
of the same study: He et al. (2018) also includes lain (2020) and lain (2021b).
The results for the association between PFHxS exposure and diabetes are presented in
Table 3-31. All the studies evaluated exposure and outcome associations in adults; in Conway et al.
(20161. both adults and children were included in study population. In the three studies of medium
confidence, one reported higher odds of incident diabetes with higher PFHxS exposure (Sun etal..
20181. although not statistically significant, while one reported an inverse association (also not
statistically significant) f Charles etal.. 20201 and the other reported no association fCardenas etal..
20171. In the low confidence studies, one study reported higher odds of diabetes with higher
exposure in men fHe etal.. 20181 and one in women fZare Teddi etal.. 20211. On the other hand,
there was an inverse association with PFHxS exposure in Conway etal. (2016) with higher
exposure associated with lower odds of diabetes. The third low confidence study (Lind etal.. 2014)
reported no association.
Overall, the evidence for the association between PFHxS exposure and diabetes is mixed.
There is some indication of higher odds of diabetes in three studies, one medium and two low
confidences, but other studies of similar confidence and design reported null or inverse findings,
and there was inconsistency in sex differences across the two low confidence studies reporting an
effect It is possible that the inconsistency may be explained by reverse causation as described
earlier, with inverse associations explained by the association of diabetes with albuminuria and
advanced kidney disease, which may lead to lower serum PFAS. However, insufficient data exist to
confirm this hypothesis.
3-277
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-31. Associations between PFHxS exposure and type 2 diabetes in
epidemiology studies
Reference, study
confidence
Population
Median exposure
(IQR) or as specified
Effect estimate
exposu re
change
Diabetes OR (95% CI)
Charles et al.
Prospective nested case-
control study of Norwegian
Women and Cancer Study
(2001-2006), Norway; 88
women (30-70 yr)
0.9 (5th—95th: 0.4-
4.3)
Controls
IQR change
0.80 (0.54, 1.20)
(2020), medium
Sun et al. (2018),
Prospective nested case-
control study of Nurses
Health Study II (1995-2000),
U.S.; 793 adults (32-52 yr)
2.0(1.3-3.5)
controls
tertiles vs. T1
Incident type 2
T2: 1.15 (0.79, 1.67)
T3: 1.26 (0.86, 1.86)
medium
Lind et al. (2014),
PIVUS study cross-sectional
(2001-2004), Sweden; 1,016
adults (70 yr)
2.1(1.6-3.4)
In-unit change
1.00 (0.74, 1.35)
low
Cardenas et al.
Diabetes Prevention Program
(1996-1999), U.S.; 957 adults
(25+ yr)
Geometric mean
(IQR)
2.4 (2.4)
log2-unit
change
Incident type 2
0.98 (0.86,1.12)b
(2017), medium
He et al. (2018),
NHANES cross-sectional
(2003, 2004, 2005-2006,
2007-2008, 2009-2010,
2011-2012), U.S.; 7,904
adults (20+ yr)
Mean+ SE
Male
2.9 + 0.1
Female
1.9 + 0.04
quartiles vs. Q1
Men
Q2: 1.99 (1.19, 3.33)*
Q3: 1.87 (1.15, 3.05)*
Q4: 2.31(1.37,3.91)*
Women
Q2: 0.65 (0.41, 1.03)
Q3: 0.87 (0.52, 1.43)
Q4: 1.22 (0.71, 2.11)
low
Zare Jeddi et al.
Cross-sectional study in
region with high PFAS
contamination (2017-2019),
Italy; 15,876 young adults
(20-39 yr)
3.5 (1.7-7.8)
quartiles vs. Q1
Q2: 0.97 (0.76, 1.24)
Q3: 1.23 (0.97, 1.57)
Q4: 1.06 (0.82, 1.37)
Men
Q2: 1 (0.69, 1.46)
Q3: 1.22 (0.86, 1.72)
Q4: 0.99 (0.7, 1.4)
Women
Q2: 1 (0.72, 1.39)
Q3: 1.39 (1.01, 1.91)*
Q4: 1.12 (0.8,1.58)
(2021), low
Conwav et al.
C8 Health Project cross-
sectional (2005-2006), U.S.;
66,889 children and adults
Mean + SD
5.2 + 10.4
no diabetes
Unit change
(No
transformation)
0.74 (0.71, 0.77)
(2016), low
Gestational diabetes
Six studies report on the relationship between PFHxS exposure and gestational diabetes.
The quality of gestational diabetes ascertainment was based on how screening of gestational
3-278
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
diabetes mellitus (GDM) was performed (e.g., defined by a study protocol versus doctor's diagnosis
at individual clinics). Another important consideration is that GDM associations with exposure are
not interpretable in the presence of diabetes. Thus, for participant selection, it was important for
studies to account for the diabetic status and/or the use of diabetic medications. Studies that did
not consider these factors by exclusion or stratification were considered deficient for the
participant selection domain. Overall, there were five studies that examined the association
between PFHxS exposure and gestational diabetes that were of medium confidence (Yu etal.. 2021:
Wang etal.. 2018: Valvi etal.. 2017: Shapiro etal.. 2016: Rahman et al.. 2 0191 and one study of low
confidence fMatilla-Santander etal.. 20171. A summary of the study evaluations for PFHxS and
gestational diabetes is presented in Figure 3-72, and additional details of the studies can be
obtained from HAWC.
ec^
L
Matilla-Santander, 2017, 4238432-
Rahman, 2019, 5024206-
Shapiro, 2016, 3201206-
-
+
-
+
++
-
+
-
+
++ ++
+
+
-
+
+
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)
+
+
+
+
+
-
+
+
Valvi et al., 2017, 3983872-
—
+
+
++
++
+
+
Wang, 2018, 5079666-
Yu, 2021, 7751046-
+
+
+
+
-
+
+
+
+
+
+
++
-
+
+
Figure 3-72. Heatmap of study evaluations for PFHxS and gestational diabetes.
For additional details see HAWC link.
The results for the association between PFHxS exposure and gestational diabetes for all
studies are presented in Table 3-32. Two medium confidence studies (Yu etal.. 2021: Shapiro etal..
20161 reported higher odds of GDM with PFHxS exposure, but neither was statistically significant,
and in Shapiro et al. (20161. the exposure-response gradient was nonmonotonic, with the odds
ratio highest in the second quartile. The results were generally null in the three other medium
confidence studies fWang etal.. 2018: Valvi etal.. 2017: Rahman etal.. 20191. In the low confidence
study (Matilla-Santander et al.. 2017). there were higher odds of GDM with PFHxS exposure,
although the exposure-response gradient was again nonmonotonic. Overall, there is no clear
association between PFHxS exposure and GDM.
3-279
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-32. Associations between PFHxS exposure and gestational diabetes in
epidemiology studies
Reference, study
confidence
Population
Median exposure
(IQR) in ng/mLoras
specified
Effect
estimate
exposu re
change
Gestational
diabetes
mellitus (GDM)
OR (95% CI)
Yu etal. (2021),
medium
Population-based birth cohort study in
Shanghai, China (2013-2016); 2,747
pregnant women
0.5 (0.3) in controls
Log-unit
change
1.15 (0.86, 1.54)
Wang etal. (2018),
medium
Haidian Maternal & Child Health
Hospital in Beijing, China (2013); 84
pregnant women with GDM and 168
healthy pregnant women
0.5 (0.3-0.7) in
controls
Unit
change
1.07 (0.86, 1.35)
Matilla-Santander et
al. (2017), low
Population-based birth cohort study
INMA (2003-2008); Spanish regions of
Valencia, Sabadell, and Gipuzkoa;
2,150 pregnant women (recruited
during first trimester of pregnancy)
Geometric mean
(Geometric SD)
0.6 (2.0)
Quartiles
Q2: 1.25 (0.51,
3.03)
Q3: 1.81 (0.76,
4.28)
Q4: 1.15 (0.42,
3.12)
Rahman et al.
(2019), medium
NICHD Fetal Growth Study, Singletons
(2009-2013); 2,334 pregnant women
(8-13 wk of gestation)
Geometric mean
(95% CI)
Overall cohort
0.8 (0.7-0.8)
GDM
0.7 (0.6-0.9)
SD
increment
Overall cohort3
0.95 (0.73, 1.23)
With family
history of type 2
diabetes3
1.03 (0.92,1.16)
Shapiro et al. (2016),
medium
Longitudinal birth cohort study MIREC
(2008-2011); Canada; 1,274 pregnant
women (recruited <14 wk of
gestation)
Geometric mean
(SD)
GDM 1.1 (2.0)
Non-GDM 1.0(2.3)
Quartiles
Q2: 1.6 (0.7, 3.8)
Q3: 1.4 (0.6, 3.5)
Q4: 1.2 (0.4, 3.5)
Valvi et al. (2017),
medium
National Hospital in Torshavn (1997
and 2000); Faroe Islands; 604 mother-
child pairs (recruited at 34 wk of
gestation)
Median (IQR)
4.5 (2.2, 8.5)
Doubling
1.03 (0.80, 1.33)
Blood glucose and insulin resistance
Homeostatic model assessment (HOMA) is a method for assessing insulin resistance and (3-
cell function from fasting glucose and insulin measured in the plasma (Matthews etal.. 1985). The
HOMA of insulin resistance (HOMA-IR) is often used in studies evaluating future risk of diabetes. It
is important to consider that blood glucose and insulin levels and HOMA-IR are difficult to interpret
in the presence of diabetes, especially if diabetes is treated with hypoglycemic medication since the
treatment will affect insulin production and secretion. Thus, for participant selection, the studies
should account for the diabetic status and/or the use of diabetic medications in participants.
Studies that did not consider these factors by exclusion or stratification were considered deficient
for the participant selection domain, and low confidence overall.
3-280
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Twenty-eight studies (reported in 31 publications) report on the relationship between
PFHxS exposure and blood glucose and/or insulin resistance. Of these, 15 were considered medium
confidence fYu etal.. 2021: Wang etal.. 2018: Valvi etal.. 2021: Starling etal.. 2017: Ren etal.. 2020:
Li etal.. 2021a: Kangetal.. 2018: Tensen etal.. 2018: Goodrich etal.. 2021: Gardener etal.. 2021:
Duanetal.. 2020: Christensenetal.. 2019: Cardenas etal.. 2017: Cakmaketal.. 2022: Alderete etal..
2019) and 10 were low confidence. Many of these studies did not account for diabetic status of the
participants and were thus deficient for participant selection. In addition, three studies were
uninformative due to critical deficiencies in at least one domain and are not considered further
fZhang et al.. 2019a: Yang etal.. 2018: Tiangetal.. 20141. Study evaluation results are summarized
in Figure 3-73 and additional details are available in HAWC. Fifteen studies reported on general
population adults and adolescents, one examined occupational exposure in firefighters, six studies
reported on pregnant women, and five studies reported on children.
The results for the association between PFHxS exposure and these outcomes for all studies
are presented in Table 3-33. For insulin resistance, two of the medium confidence studies in adults
(Cardenas etal.. 2017) and pregnant women (Tensen etal.. 20181 reported higher HOMA-IR with
higher PFHxS exposure (both statistically significant). The association in Tensen etal. f20181 was
observed primarily in women with high GDM risk based on predefined risk factors (BMI >27
kg/m2, family history of diabetes mellitus, present multiple pregnancy, glucosuria during
pregnancy, previous GDM, or delivery of macrosomic child). The association in women without
GDM risk was in the same direction but much smaller, which may suggest an interaction between
PFAS exposure and metabolic vulnerability, but this cannot be assessed further using the available
data. The other studies indicated no increase in insulin resistance with higher exposure. For blood
glucose, three of the medium confidence studies in pregnant women fYu etal.. 2021: Tensen etal..
20181 and 6 weeks postpartum fWang etal.. 20181 reported statistically significantly elevated
blood glucose with higher PFHxS exposure. One study in adolescents and young adults also
reported a positive association in post-puberty girls undergoing an oral glucose tolerance test, with
a significant association at the 1-hour post glucose test, but an inverse association was reported in
boys and results at other ages did not show an association (Goodrich etal.. 2021). Results in other
studies were generally null.
Overall, an association is not clear between PFHxS exposure and insulin resistance or blood
glucose. Some positive associations were observed in medium confidence studies, but this was not
consistently observed across studies, including other medium confidence studies of similar design
and power. It is possible that exposure contrast was not adequate to observe an association in these
studies, but the positive associations were observed in studies with exposure levels similar to the
null studies.
3-281
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Alderete, 2019, 5080614-
Cakmak. 2022,10273369-
Cardenas, 2017, 4167229-
Chen, 2019, 5387400-
Christensen, 2019, 5080398-
Duan, 2020, 5918597-
Fleisch, 2017, 3858513-
Gardener. 2021. 7021199-
Goodrich, 2021, 9960584-
Heffernan, 2018, 5079713
Jensen, 2018, 4354143-
Jiang, 2014, 2850910
Kang, 2018, 4937567-
Khalil, 2018, 4238547
Khalil, 2020. 7021479-
Koshy, 2017, 4238478
Li, 2021, 7404102-
Lin, 2009, 1290820-
Lin, 2020, 6988476 -
Lind, 2014, 2215376
Liu, 2018, 4238396
Ren, 2020, 6833646 -
Starling, 2017, 3858473
Valvi, 2021, 8438216-
Wang. 2018, 5079666
Yang,2018, 4238462
Yu, 2021, 7751046
Zhang, 2019, 5083675
Legend
I Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
I Critically deficient (metric) or Uninformative (overall)
Figure 3-73. Heatmap of study evaluations for insulin resistance and blood
glucose.3 For additional details see HAWC link.
'Multiple publications of the same study: Lin et a!. (2009a) also includes Nelson et al. (2010); Christensen et al,
(2019) also includes Jain (2020); Cakmak et al. (2022) also includes Fisher et al. (2013).
3-282
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-33. Associations between PFHxS exposure and insulin resistance or
blood glucose in epidemiology studies
Reference and
confidence
Population
Median
exposu re
(IQR) in ng/
mLoras
specified
Effect
estimate
Insulin resistance
(HOMA-IR)
Blood glucose
Adults and adolescents
Duan et al. (2020),
Medium
Cross-sectional
study in China in
2017; 294 adults
0.3 (20 yr)
Log
mean j^SEM
Adolescents
1.0^0.1
Adults
0.6 + 0.04
Mean j^SEMb
for log-unit
change
Adolescents
0.05 jJD.03
Adults
0.00 jJD.04
Adolescents
-0.01^0.03
Adults
-0.02 jJD.06
3-283
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Median
exposu re
(IQR) in ng/
Reference and
mLoras
Effect
Insulin resistance
confidence
Population
specified
estimate
(HOMA-IR)
Blood glucose
Goodrich et al.
SOLAR cohort
1.1 (GM) in
Differences
NR
SOLAR
(2021), Medium
(2001-2012), U.S.;
SOLAR
with high vs.
Puberty
328 children (8-
cohort; 0.8 in
low PFHxS
Girls
13 yr) with 2 years
CHS cohort
levels
Fasting: 1 (-9,
follow-up
girls
12)
Children's Health
OGTT 1 hr: 3 (-8,
Study cross-
13)
sectional analysis
Boys
within cohort
Fasting: 0 (-12,
(2002), U.S.; 137
13)
young adults (17-22
OGTT 1 hr: -7
years)
(-19, 5)
Postpuberty
Girls
Fasting: 6 (-8,
19)
OGTT 1 hr: 25
(12, 39)*
Boys
Fasting: -5 (-20,
11)
OGTT 1 hr: -25
(-40, -9)*
CHS young adult
Girls
Fasting 3 (-17,
23)
OGTT 1 hr: 26 (6,
46)
Boys
Fasting: 1 (-12,
13)
OGTT 1 hr: 3
(-10, 17)
Li etal. (2021a);
Prospective cohort
1.9 (1.0-3.3)
Adjusted
NR
Exposure in
Medium
(2003-2006); U.S.;
at age 3
difference
gestation
221 adolescents
for IQR
-0.3 (-1.4, 0.9)
(12 yr, followed
increase
3 yr
from pregnancy)
0.4 (-0.6, 1.5)
12 yr
0.5 (-0.7, 1.8)
Christensen et al.
NHANES cross-
2007-2008
Odds ratio
NR
Q2: 0.88 (0.61,
(2019); Medium
sectional (2007-
2.0 (1.1, 3.5)
(95% CI) for
1.27)
2014); U.S.; 2,975
2009-2010
quartiles vs.
Q3: 0.87 (0.59,
adults (>20 yr)
1.7 (0.9, 2.9)
Q1
1.29)
3-284
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference and
confidence
Population
Median
exposu re
(IQR) in ng/
mLoras
specified
Effect
estimate
Insulin resistance
(HOMA-IR)
Blood glucose
2011-2012
1.3 (0.8, 2.3)
2013-2014
1.4 (0.8, 2.6)
Q4: 0.85 (0.55,
1.31)
Cakmak et al.
(2022); Medium
Canadian Health
Measures Survey
cross-sectional
(2007-2017);
Canada; 6,024 all
ages
1.5 (GM)
% change for
GM increase
-0.1 (-4.1, 4.6)
0.3 (-0.6, 1.3)
Lind et al. (2014);
Low
PIVUS study cross-
sectional (2001-
2004), Sweden;
1,016 adults (70 yr)
2.1(1.6-3.4)
Beta
coefficient
(95% CI) for
In-unit
change
-0.085 (-0.14,
-0.03)*
NR
Cardenas et al.
(2017); Medium
Diabetes Prevention
Program (1996—
1999), U.S.; 957
adults (25+ yr)
GM (IQR)
2.4 (2.4)
Beta
coefficient
(95% CI) for
doubling
0.34 (0.12, 0.55)a
0.29 (-0.13, 0.70)
Lin et al. (2020c);
Low
Cross-sectional
study in high
contamination area
(2016-2017),
Taiwan; 397 older
adults (55-75 yr)
2.7
Beta
coefficient
(95% CI) for
quartiles vs.
Q1
NR
Q2: 2.42 (-4.91,
9.75)
Q3: -3.22
(-10.78,4.35)
Q4: 2.54 (-5.13,
10.21)
Khalil et al. (2020);
Low
Cross-sectional
study of firefighters
(2009), U.S. 38 men
3.1 (GM)
Beta
coefficient
(95% CI) for
log-unit
change
NR
no association
(figure only)
Liu et al. (2018);
Low
POUNDS clinical
trial (2003-2007),
U.S.; 621 adults
(30-70 yr)
Male 3.1
(2.3-4.4)
Female 1.9
(1.2-3.0)
Spearman
correlation
0.07
Change in
glucose
0-6 mo in trial:
0.02
6-24 mo: -0.02
Pregnant women
Jensen et al.
(2018); Medium
Odense Child
Cohort (OCC)
(2010-2012),
Denmark; 649
pregnant women
(15-49 yr), outcome
0.3 (0.1-0.6)
% Change
(95% CI) for
doubling
High GDM risk
9.5 (1.0,18.8)*
Low GDM risk
2.8 (-7.5, 14.3)
High GDM risk
1.7 (0.2, 3.2)*
Low GDM risk
0.2 (-1.3,1.7)
3-285
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference and
confidence
Population
Median
exposu re
(IQR) in ng/
mLoras
specified
Effect
estimate
Insulin resistance
(HOMA-IR)
Blood glucose
measured at 28 wk
gestation
Yu etal. (2021),
medium
Population-based
birth cohort study in
Shanghai, China
(2013-2016); 2,747
pregnant women
0.5 (0.3) in
controls
Beta
coefficient
(95% CI) for
log-unit
change
NR
0.003 (-0.04,
0.05)
OGTT 1 hr
0.22 (0.06, 0.37)*
OGTT 2 hr
0.08 (-0.06, 0.21)
Gardener et al.
(2021); Medium
Vanguard Pilot
Study of the
National Children's
Study cross-
sectional (2009);
U.S.; 425 pregnant
women in 3rd
trimester
0.5 (0.3-0.9)
Means (95%
CI) for
quartiles
Nonsignificant,
nonmonotonic
increase (figure
only)
NR
Wang etal. (2018);
Medium
Haidian Maternal &
Child Health
Hospital in Beijing,
China (January-
March 2013); 84
pregnant women as
GDM and 168
healthy pregnant
women, outcome
measured at 6 wk
postpartum
GDM 0.5 (0.3
-0.8)
Non-GDM 0.5
(0.3-0.7)
Odds ratio
(95% CI) for
categories of
blood
glucose (3.2—
4.74; 4.75-
5.04; 5.06-
6.84 mmol/L)
NR
GDM/non-GDM
pooled (adjusted
for status)
Medium vs.
Lowest
1.32 (0.72, 2.42)
Highest vs.
Lowest
2.29 (1.22,4.29)*
Starling et al.
(2017); Medium
Health Start cohort
at the University of
Colorado Hospital
(2009-2014); U.S.;
1,410 pregnant
women (>16 yr),
outcome measured
at mid-pregnancy
0.8 (0.5, 1.2)
% Change
(95% CI) for
categories of
exposure
NR
Group 1
-0.009 (-0.029,
0.010)
Group 2
-0.023 (-0.044,
-0.002)
Ren et al. (2020);
Medium
Shanghai-Minhang
Birth Cohort (2012);
China; 856 pregnant
women (outcome
measured at 20-
28 wk gestation)
2.8(2.1-3.6)
OR (95% CI)
for high
glucose
NR
0.89 (0.51,1.55)
Children
3-286
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference and
confidence
Population
Median
exposu re
(IQR) in ng/
mLoras
specified
Effect
estimate
Insulin resistance
(HOMA-IR)
Blood glucose
Kang et al. (2018);
Medium
Korea
Environmental
Health Survey in
Children and
Adolescents
(KorEHS-C)
subcohort (2012-
2014); South Korea;
children (3-18 yr)
Geometric
mean (SD)
0.8 (1.6)
Beta
coefficient
(95% CI) for
In-unit
change
NR
0.925 (-1.779,
2.164)
Khalil et al. (2018);
Low
Cross-sectional
study of obese
children from Lipid
Clinic at Dayton's
Children Hospital
(April-Oct. 2016);
U.S.; children (8-
12 yr)
1.1 (1.4)
Beta
coefficient
(95% CI) for
unit change
-0.11 (-0.10, 0.78)
0.00 (-2.10, 2.09)
Goodrich et al.
(2021), Medium
SOLAR cohort
(2001-2012), U.S.;
328 children (8-
13 yr) with 2 yr
follow-up
1.1 (GM) in
SOLAR
cohort; 0.8 in
CHS cohort
girls
Differences
with high vs.
low PFHxS
levels
NR
Prepuberty
Girls
Fasting -2 (-16,
12)
OGTT 1 hr -4
(-18, 10)
Boys
Fasting -7 (-15,
0)
OGTT -7 (-15, 0)
Alderete et al.
(2019); Medium
Study of Latino
Adolescents at Risk
of type 2 Diabetes
(SOLAR) cohort
(2001-2011); U.S.;
children (8-14 yr)
Geometric
mean (SD)
1.7 (2)
Beta
coefficient
(95% CI) for
In-unit
change
-0.4 (-1.7, 0.8)
0.9 (-2.5, 4.2)
Fleisch et al.
(2017); Low
Project Viva
prospective cohort
(1992-2002); U.S.;
665 mother-
children pairs
Geomean
(25%, 75%)
Prenatal
2.5(1.6,3.8)
Mid-
childhood
2.2 (1.2, 3.4)
% Change
(95% CI) for
quartiles vs.
Q1
Prenatal
Q2: -6.7 (-23.7,
14.2)
Q3: -13.5 (-29.6,
6.3)
Q4: -17.1 (-32.3,
1.6)
Mid-childhood
Q2: -5.1 (-20.9,
13.8)
NR
3-287
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference and
confidence
Population
Median
exposu re
(IQR) in ng/
mLoras
specified
Effect
estimate
Insulin resistance
(HOMA-IR)
Blood glucose
Q3: -6.7 (-22.7,
12.6)
Q4: -16.8 (-31.4,
0.8)
*P-value or p-trend < 0.05.
NR = not reported; OGTT = oral glucose tolerance test.
Metabolic syndrome
Metabolic syndrome is defined using criteria related to waist circumference, elevated
triglycerides, reduced HDL cholesterol, elevated blood pressure, and elevated fasting glucose. These
factors contribute to increase the risk of cardiovascular conditions such as atherosclerosis,
coronary heart disease and stroke fGad. 2015: Fruchart etal.. 2004: American Heart Association.
20221. Three abnormal findings out of the five factors classify a person with metabolic syndrome
(Alberti et al.. 2009). Metabolic syndrome is also associated with liver disease (see Section 3.2.4).
Six studies reported on the association between PFHxS exposure and metabolic syndrome.
One study was uninformative due to critical deficiencies in participant selection, outcome
ascertainment, and confounding (Yang etal.. 20181. The other five studies were cross-sectional
fZare Teddi etal.. 2021: Lin etal.. 2009b: Lin etal.. 2009a: Fisher etal.. 2013: Christensen etal..
20191 and considered medium confidence. A summary of the study evaluations for PFHxS and
metabolic syndrome is presented in Figure 3-74, and additional details of the studies can be
obtained from HAWC.
There was little indication of increased odds of metabolic syndrome with higher exposure
to PFHxS. One study in older adults in an area with high PFAS contamination (Lin etal.. 2020c)
reported a positive association in the fourth quartile (OR [95% CI]: 1.22 [0.66, 2.25]), but this
association was nonmonotonic across quartiles and not statistically significant. The other four
studies reported results that were null fZare Teddi etal.. 2021: Lin etal.. 2009a: Fisher etal.. 20131
or inverse fChristensen etal.. 20191.
3-288
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
* e®xe o^e
,r>
Christensen, 2019, 5080398-
++
+
+
+
+
+
+
+
H
Legend
Good (metric) or High confidence (overall)
Fisher, 2013, 2919156-
++
+
+
+
++
+
+
+
+
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Lin, 2009, 1290820-
+
+
+
+
+
+
+
+
B
Lin, 2020, 6988476 -
"
+
+
+
+
+
+
+
Yang, 2018, 4238462-
¦
+
+
"
Zare Jeddi, 2021, 7404065-
+
+
+
+
+
+
+
+
Figure 3-74. Summary of study evaluations for epidemiology studies of PFHxS
and metabolic syndrome. For additional details see HAWC link.
Adiposity
Twenty-five studies (29 publications) reported on the association between PFHxS exposure
and obesity, BMI, and/or other measures of adiposity. Two studies were excluded as uninformative
due to lack of consideration of potential confounding (Zhang et al.. 2 019a: Yang etal.. 20181. Of the
23 remaining studies, 10 were cross-sectional studies fZare Teddi etal.. 2021: Thomsen etal.. 2021:
Scinicariello etal.. 2020a: Nelson etal.. 2010: Lind etal.. 2022: Khalil etal.. 2018: Domazet etal..
2020: Christensen etal.. 2019: Chen etal.. 2019a: Canova etal.. 20211 and were classified as low
confidence because of concern that the timing of exposure measurement was not relevant to
development of this chronic outcome, similar to concerns described for diabetes. Thirteen studies
had prospective exposure measurement, including nine that examined the association between
prenatal or early-life exposure measurements and adiposity during childhood, one cohort of people
living near a uranium processing plant, one clinical trial of weight loss diets that examined weight
change, and two studies of gestational weight gain. All of the prospective studies, where exposure
was measured prior to the outcome, were classified as medium confidence. The evaluations for
these studies are summarized in Figure 3-75.
3-289
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Blake, 2018, 5080657
Bloom, 2022, 9959635
Braun, 2016, 3859836
Cariova, 2021, 10176518-
Chen, 2019, 5080578-
Chen, 2019, 5387400-
Christensen, 2019, 5080398 -I
Domazet, 2020, 6833700 -
Hartman, 2017, 3859812-
Janis, 2021, 7410181-
Karlsen, 2017, 3858520-
Khalil, 2018, 4238547-
Lind, 2022, 10176401 -
Liu, 2018, 4238396-
Manzano-Salgado, 2017b, 4238509-
Marks, 2019, 5381534-
Martinsson, 2020, 6311645 -
Mora, 2017, 3859823^
Nelson, 2010, 1291110 -|
Papadopoulou, 2021, 9960593
Romano, 2020, 7014708
Scinicariello, 2020, 6391244
Thomsen, 2021, 9959568
Yang,2018, 4238462
Zare Jeddi, 2021, 7404065
Zhang,2019, 5083675
Legend
I Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
I Critically deficient (metric) or Unirformative (overall)
Figure 3-75. Summary of study evaluations for epidemiology studies of
adiposity. For additional details see HAVVC link. Multiple publications of the same
study: Braun et al. (2016) also includes Braun etal. (2020): Liu etal. (2020c).
and Li et al. f2021a). Mora et al. f2017) also includes lanis etal. f2021).
3-290
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
The results from the studies of adiposity in children are summarized in Tables 3-34 and
3-35, which contain the continuous outcome measures and dichotomous outcome (overweight),
respectively. Most studies report null results for the associations between PFHxS and BMI, waist
circumference, or direct measures of body fat In analyses of overweight/obesity as a dichotomous
outcome, three medium confidence studies (four publications) reported positive associations
(Martinsson etal.. 2020: Liu etal.. 2020c: Braun etal.. 2016) with odds ratios or relative risks
ranging 1.16 to 1.71. However, only one study was statistically significant (Liu etal.. 2020c) and the
association in Martinsson et al. (2020) was nonmonotonic across quartiles, with an inverse
association in the third quartile and a positive association in the fourth quartile. In addition, as
described in the Developmental Effects section, one medium confidence study by Gvllenhammar et
al. f 20181 was null for weight standard deviation scores over time from 3 to 60 months of age.
In adults, one medium confidence prospective study fLiu etal.. 20181 reported no difference
in weight loss associated with PFHxS exposure but found a statistically significant increase in
weight gain associated with PFHxS exposure in women following the weight loss trial (changes in
body weight: tertile 1: 2.7 ± 0.8, tertile 2: 3.6 ± 0.9, tertile 3: 4.9 ± 0.9, p-trend: 0.009). The second
medium confidence prospective study f Blake etal.. 20181 and the low confidence cross-sectional
studies fZare Teddi etal.. 2021: Lind etal.. 2022: Christensen etal.. 2019: Chen etal.. 2019al
reported no difference in adiposity with higher PFHxS exposure. Additionally, two medium
confidence studies examined gestational weight gain. Marks etal. (2019b) and Romano etal. (2020)
reported no association with absolute gestational weight gain (stratified by baseline weight
categories under/normal weight and overweight/obese).
Overall, there is very limited evidence of an association between PFHxS exposure and
adiposity. The strongest evidence comes from a weight loss trial in adults that observed higher
weight gain following the trial, but the lack of coherence with related outcomes in the remaining
studies decreases the strength of the evidence.
3-291
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-34. Associations between maternal exposure to PFHxS and adiposity
in children
Reference,
study
confidence
Population
Median
exposure
(IQR)
(Hg/mL)
Effect
estimate
BMI
Waist
circumference
Body fat
Chen et al.
(2019b),
medium
Prospective birth
cohort in China;
404 children at
5 yr
0.2 (range
0.1-0.9)
P (95% CI)
for log-unit
change
Girls: -0.5
(-1.1,0.2)
Boys: 0.4 (-0.3,
1.1)
Girls: -1.2 (-3.1,
0.7)
Boys: 0.6 (-1.3,
2.5)
Body fat percent
Girls: -1.9 (-4.9,
1.0)
Boys: 1.8 (-0.7,
4.3)
P (95% CI)
for tertiles
(refTl)
Girls
T2: 0.2 (-0.8,
0.3)
T3: -0.2 (-0.8,
0.3)
Boys
T2: 0.1 (-0.5,
0.7)
T3: 0.2 (-0.4,
0.8)
Girls
T2: -0.4 (-2.1,
1.2)
T3: -0.4 (-2.1,
1.3)
Boys
T2: -0.2 (-1.8,
1.4)
T3: 0.5 (-1.1, 2.1)
Girls
T2: -0.8 (-3.4,
1.7)
T3: -1.9 (-4.4,
0.7)
Boys
T2: 0.2 (-2.0, 2.3)
T3: 0.7 (-1.4, 2.8)
Karlsen et al.
(2017),
medium
Birth cohort
(2007-2009),
Faroe Islands;
444 children
with follow-up at
18 mo
0.2 (0.1-
0.3)
P (95% CI)
for log-unit
increase;
T2 and T3
vs. T1
0.10
(-0.01,0.21)
T2: -0.03
(-0.23,0.17)
T3: 0.18
(-0.03,0.38)
NR
NR
371 children
with follow-up at
5 yr
0.04
(-0.07,0.15)
T2: -0.02
(-0.22,0.19)
T3: 0.07
(-0.14,0.28)
NR
NR
Papadopoulou
et al. (2021),
medium
Six birth cohorts,
Europe, 1,301
children at 6-
11 yr
prenatal
0.5 (0.3-
0.9)
P (95% CI)
for
Quartiles
vs. Q1
NR
Q2: -0.02 (-0.22,
0.17)
Q3: 0.05 (-0.18,
0.28)
Q4: 0.03 (-0.23,
0.30)
NR
Children
0.3 (0.2-
0.6)
NR
Q2: -0.12 (-0.31,
0.06)
Q3: 0.10 (-0.13,
0.32)
Q4: 0.04 (-0.22,
0.29)
NR
3-292
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference,
study
confidence
Population
Median
exposure
(IQR)
(Hg/mL)
Effect
estimate
BMI
Waist
circumference
Body fat
Thomsen et al.
(2021), low
Cross-sectional
analysis within
birth cohort
(2009),
Denmark, 109
boys at ~12 yr
0.5 (0.4-
0.7)
P (95% CI)
for log-unit
increase
NR
NR
Abdominal fat
0.03 (-0.15, 0.20)
Visceral fat
0.02 (-0.11, 0.14)
Total fat
0.01 (-0.22, 0.23)
Manzano-
Salgado et al.
(2017b),
medium
INMA birth
cohort (2003-
2008), Spain;
1,230 children
with follow-up at
4 yr
0.6 (GM)
(0.4-0.8)
P (95% CI)
for
doubling
exposure
-0.02
(-0.10,0.07)
-0.04
(-0.14,0.05)
NR
1,086 children
with follow-up at
7 yr
-0.04
(-0.14,0.06)
-0.04
(-0.12,0.04)
NR
Domazet et al.
(2020), low
Cross-sectional
analysis within
multicenter
cohort (1997),
Europe; 242
children at 9 yr
0.9 (0.7-
1.1)
% change
(95% CI)
for 10%
increase
NR
NR
Fat mass
-1.07 (-1.99,
-0.15)*
Bloom et al.
(2022),
medium
ECHO cohort
(2017-2019),
U.S. 803 children
at 4-8 yr
0.9 (0.5-
1.5)
P (95% CI)
for log-unit
increase
BMI z-score
Without
obesity
-0.06 (-0.17,
0.05)
With obesity
0.01 (-0.22,
0.24)
Without obesity
-0.06 (-0.15,
0.04)
With obesity
0.16 (-0.09, 0.40)
Fat mass
Without obesity
-0.08 (-0.42,
0.25)
With obesity
0.63 (-0.68, 1.93)
Percent body fat
Without obesity
-0.003 (-0.01,
0.01)
With obesity
0.01 (-0.02,0.04)
Scinicariello et
al. (2020a),
low
NHANES cross-
sectional study
(2013-2014),
U.S. 600 children
at 3-11 yr
0.9 (GM)
P (95% CI)
for tertiles
vs. T1
BMI z-score
T2: -0.17
(-0.47, 0.13)
T3: -0.26
(-0.57, 0.04)
Weight for age
T2: -0.30 (-0.67,
0.07)
T3: -0.42 (-0.76,
-0.08)*
NR
Khalil et al.
(2018), low
Cross-sectional
study (2016),
U.S. 48 children
with obesity at
8-12 yr
1.1(1.4)
P (95% CI)
for unit
change
0.32 (-0.76,
1.39)
NR
NR
3-293
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference,
study
confidence
Population
Median
exposure
(IQR)
(Hg/mL)
Effect
estimate
BMI
Waist
circumference
Body fat
Braun et al.
(2016); Liu et
al. (2020c);
Braun et al.
(2020); Li etal.
HOME birth
cohort (2003-
2006), U.S.; 204
children with
follow-up at 8 yr
1.4 (0.8-
2.3)
Difference
(95% CI)
Tertiles vs.
T1
T2: 0.22
(-0.10,0.54)
T3: 0.12
(-0.21,0.45)
T2: 2.7 (0.0,5.4)
T3: 1.1 (-1.7,3.9)
Body fat percent
T2: 2.3 (0.3,4.2)
T3: 1.1 (-0.9,3.1)
(2021a),
medium
212 children
with follow-up at
12 yr
P (95% CI)
for IQR
increase
BMI z-score
Prenatal
exposure
0.10 (-0.08,
0.28)
12-yr-old
exposure
0.09 (-0.14,
0.31)
Prenatal
exposure
1.73 (-0.87, 4.33)
12-yr-old
exposure
0.55 (-2.48, 3.57)
Fat mass index
Prenatal
exposure
0.10 (-0.07,0.26)
12-yr-old
exposure
0.08 (-0.11, 0.27)
Body fat percent
Prenatal
exposure
0.94 (-0.35, 2.22)
12-yr-old
exposure
0.68 (-0.79,2.15)
214 children
with follow-up at
12 yr
P (95% CI)
for IQR
increase
T2: -0.65
(-1.90, 0.65)
T3: -0.50
(-1.78, 0.76)
NR
NR
Difference
(95% CI)
Tertiles vs.
T1
Rate of BMI
change from
8-12 yr
T2: -0.06
(-0.20, 0.09)
T3: -0.01
(-0.15, 0.13)
NR
NR
186 children
with follow-up at
12 yr
Difference
(95% CI)
for IQR
change
NR
Prenatal
exposure
0.03 (-0.01, 0.08)
12-yr-old
exposure 0.02
(-0.04,0.07)
Visceral fat
Prenatal
exposure
0.09 (-0.01,0.20)
12-yr-old
exposure
0.10 (-0.05,0.26)
Hartman et al.
(2017),
medium
ALSPAC birth
cohort
(1991-1992),
United Kingdom;
359 children
with follow-up at
9 yr)
1.6 (1.3-
2.2)
P (95% CI)
for 1 unit
increase
-0.02
(-0.08,0.03)
-0.08
(-0.22,0.06)
DXA total body
fat
-0.06
(-0.21,0.09)
DXA trunk fat
-0.01
(-0.11,0.08)
3-294
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Median
Reference,
exposure
study
(IQR)
Effect
Waist
confidence
Population
(Hg/mL)
estimate
BMI
circumference
Body fat
Mora et al.
Project Viva birth
2.4(1.6-
P (95% CI)
0.01
0.03 (-0.10,0.16)
Sum of
(2017); Janis
cohort (1999—
3.8)
for IQR
(-0.03,0.05)
subscapular and
et al. (2021),
2002), U.S.;
increase
triceps skinfold
medium
1,006 children
thickness
with follow-up at
0.16 (0.01,0.31)
median 3 yr
876 children
0.01
0.11 (-0.22,0.43)
Sum of
with follow-up at
(-0.03,0.05)
subscapular and
median 7 yr
triceps skinfold
thickness
0.25 (-0.14,0.64)
DXA total fat
mass index
0.04 (-0.04,0.13)
DXA trunk fat
mass index
0.02 (-0.02,0.06)
531 children
P (95% CI)
BMI z-score
NR
Total fat mass
with follow-up at
-0.05 (-0.09,
index
13 yr
0.00)
-0.22 (-0.35,
-0.08)*
Truncal fat mass
index
-0.09 (-0.16,
-0.03)*
Canova et al.
Cross-sectional
adolescen
P (95% CI)
BMI z-score
NR
NR
(2021), low
study in highly
ts
vs. Q1
Q2: -0.08
contaminated
2.8(1.6-
(-0.15, 0)
area (2017-
4.8)
Q3: 0.01
2019), Italy;
(-0.07, 0.09)
6,669
Q4: 0.03
adolescents (14—
(-0.05, 0.12)
19 yr) and 2,693
Similar for
children (8-
boys and girls
11 yr)
children
P (95% CI)
BMI z-score
NR
NR
1.9 (1.2-
for In-unit
Q2: 0.06
2.8)
increase
(-0.08, 0.2)
Q3: -0.20
(-0.34, -0.06)*
Q4: -0.18
(-0.32, -0.03)*
*p < 0.05.
T = tertile, GM = geometric mean, DXA = dual-energy X-ray absorptiometry, NR = not reported.
3-295
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-35. Associations between maternal exposure to PFHxS and overweight
status in children in medium confidence epidemiology studies
Reference
Population
Median
exposu re
(IQR)
(Hg/mL)
Effect estimate
Overweight
Karlsen et al.
(2017)
Birth cohort (2007-2009), Faroe
Islands; 444 children with
follow-up at 18 mo
0.2 (0.1-0.3)
OR (95% CI) for
log-unit
increase;
Tertiles vs. T1
1.12 (0.97, 1.30)
T2: 1.06 (0.82,1.38)
T3: 1.24 (0.97,1.58)
371 children with follow-up at
5 yr
1.11(0.77, 1.59)
T2: 0.86 (0.47,1.55)
T3: 1.22 (0.73, 2.04)
Manzano-
Salgado et al.
INMA cohort (2003-2008),
Spain; 1,230 children with
follow-up at 4 yr
0.6 (GM)
(0.4-0.8)
RR (95% CI) for
doubling
exposure
0.96 (0.87, 1.07)
(2017b)
1,086 children with follow-up at
7 yr
0.94 (0.84, 1.05)
Martinsson et
al. (2020)
Case-control study (2003-2008),
Sweden; 1,048 children at 4 yr
0.7 (0.5-1.0)
OR (95% CI);
Quartiles vs. Q1
Q2: 0.95 (0.66,1.37)
Q3: 0.66 (0.44, 0.97)
Q4: 1.16 (0.81,1.66)
Braun et al.
(2016);
Liu et al.
(2020c)
HOME birth cohort (2003-
2006), U.S.; 204 children with
follow-up at 8 yr
1.4 (0.8-2.3)
RR (95% CI);
Tertiles vs. T1
T2: 1.33 (0.72, 2.48)
T3: 1.48 (0.75, 2.96)
212 children with follow-up at
12 yr
RR (95% CI) for
IQR increase
1.71(1.08,2.73)*
Mora et al.
(2017)
Project Viva birth cohort (1999—
2002), U.S.; 1,006 children with
follow-up at median 3 yr
2.4(1.6-3.8)
RR (95% CI) for
IQR increase
Overweight:
1.03 (0.94,1.13)
Obese:
1.02 (0.89,1.17)
876 children with follow-up at
median 7 yr
Overweight:
1.04 (0.92,1.17)
Obese:
1.07 (0.94,1.22)
Animal Studies
There are two 28-day gavage studies in SD rats fNTP. 2018b: 3M. 2000al one 4- to 6-week
oral gavage exposure study using genetically modified mice (Bijland etal.. 2011). and two
reproductive/developmental studies using CD-I mice (Changetal.. 2018) or Sprague Dawley rats
(Bute nhoff etal.. 2009: 3M. 2003) that measure effects relevant to the assessment of the
cardiovascular or metabolic systems after repeated oral dose exposure to PFHxS. The studies
report on heart weight and histopathology, and alterations of cardiometabolic endpoints such as
fasting levels of serum lipids which are considered indicative of potential cardiotoxicity fGad. 2015:
Fruchart et al.. 20041. Overall study confidence was high for cardiometabolic endpoints evaluated in
3-296
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
these studies (NTP. 2018b: Chang etal.. 2018: Butenhoff et al.. 2009: Bijland etal.. 2011: 3M. 2000a.
20031. Studies reporting on heart weight and histopathology were considered of low confidence
due to experimental design uncertainties fNTP. 2018a: Butenhoff etal.. 2009: 3M. 20031 (see Figure
3-76 and discussion below). Specifically, the exposure duration of less than a month was not
considered sufficient for evaluation of injury to the cardiovascular system fDaugherty etal.. 20171.
raising significant concerns for insensitivity.
Heart weight and histopathology
There is no clearly preferred measurement for evaluating heart weights (absolute or
relative). Some data show that heart weight is nonproportional to body weight (Bailey etal.. 20041.
other data reports that heart weight in strongly correlated with body weight, with better
correlation in males fNirogi etal.. 20141. Thus, both absolute and relative heart weights are
considered biological relevant metric for this endpoint. Absolute and relative heart weights were
not altered in SD rats exposed to PFHxS for 28 days at 0.625 to 10 mg/kg-day (NTP. 2018a: 3M.
2000a). However, one reproductive/developmental toxicity study reported decreased relative
heart/brain weight (by 8%) in F0 generation male SD rats exposed to PFHxS for 44 days (Butenhoff
etal.. 2009: 3M. 2003): the biological significance of this 8% change is unclear. Importantly, the
same study also reports that absolute and heart-to-body weight ratios were not affected in males or
females exposed to PFHxS.
Heart histopathology was evaluated in a 28-day study (NTP. 2018a) and a
reproductive/developmental toxicity study (Butenhoff etal.. 2009: 3M. 2003). Both studies used SD
rats. Exposure to PFHxS from 0.625 to 10 mg/kg-day did not cause a significant effect on the
incidence of nonneoplastic cardiovascular injury in male or female rats fNTP. 2018a: Butenhoff et
al.. 2009: 3M. 2003). As noted above, there is concern that the exposure duration of these studies
(<1 month) was too short to expect to see histological manifestations of cardiac injury.
3-297
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Chemical administration and characterization
Selective reporting and attrition
Confounding/variable control
Observational bias/blinding
Reporting quality
Allocation
B Legend
Good (metric) or High confidence (overall)
+ Adequate (metric) or Medium confidence (overall)
- Deficient (metric) or Low confidence (overall)
9 Critically deficient (metric) or Uninformative (overall)
NR| Not reported
Exposure timing, frequency and duration -
Results presentation -\
Endpoint sensitivity and specificity
Overall confidence
rm
Figure 3-76. Cardiometabolic effects, heart weight/histopathology - animal
study evaluation heatmap. For additional details see HAWC link.
Serum lipids
As described above, increased serum lipids such as cholesterol and triglycerides are
established risk factors for cardiovascular disease. Briefly, increased serum levels of cholesterol
and triglycerides due to induced hepatic production, reduced clearance, or altered enterohepatic
circulation (Roth etal.. 20241 can promote vascular endothelial cell damage and activation of pro-
inflammatory cell signals, macrophage migration to the site of injury, lipoprotein retention and
modification (e.g., oxidation), formation of foam cells, and atherogenic plaque development (Zhang
etal.. 2022a: Y etal.. 2022: Ouispe etal.. 2022: Miller etal.. 2011: Linton etal.. 2000: Gad. 20151.
Several studies also evaluated hepatic lipid accumulation after PFHxS exposure (see Section 3.2.4.).
Levels of plasma cholesterol were evaluated in two reproductive/developmental toxicity
studies (Chang etal.. 2018: Butenhoff etal.. 2009: 3M. 2003). and in four short-term exposure
studies fNTP. 2018a: Biilandetal.. 2011: 3M. 2000a). one sub-chronic study (He etal.. 2022). and
one chronic exposure study fPfohl etal.. 20201 (see Figure 3-77). In the high confidence, short-term
studies, exposure to PFHxS for 28 days resulted in a 12% to 51% reduction in serum cholesterol at
doses ranging from 1.25 to 10 mg/kg-day in male and female rats in one study f3M. 2000al and in
males only in the other fNTP. 2018al. Likewise, a separate study using male APOE*3-Leiden
CETP 26 mice reported that exposure to 6 mg/kg-day PFHxS decreased total cholesterol, HDL, and
26APOE*3-Leiden.CETP mice is a genetically modified animal model which better emulates human lipoprotein
profiles and is used to investigate cholesterol metabolism and cardiovascular disease fVeseli et al.. 20171.
3-298
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
non-HDL cholesterol (Bijland etal.. 20111. Two reproductive/developmental toxicity studies report
that PFHxS exposure for 42 to 44 days decreased serum cholesterol by 19% to 42% in male F0 SD
rats at doses ranging from 0.3 to 10 mg/kg-day fButenhoff et al.. 2009: 3M. 20031. whereas F0 CD-I
male mice treated with 10 mg/kg-day displayed a 27% reduction in cholesterol fChang etal.. 20181.
However, these effects were not observed in female Sprague Dawley rats or CD-I mice f Chang etal..
2018: Bute nhoff etal.. 2009: 3M. 2003). or male C57BL/6J mice exposed for 12 or 29 weeks to 0.06
or 0.15 mg/kg-day PFHxS in high (He etal.. 20221 or medium confidence studies fPfohl etal.. 20201
respectively (see Figure 3-78).
PFHxS exposure-induced effects on serum lipid levels and production were also measured
in rats and mice. In a high confidence study of SD rats, short-term oral exposure for 28 days
decreased serum triglyceride levels by 22% to 46% after exposures ranging from 2.5 to 10 mg/kg-
day fNTP. 2018a: 3M. 2000al. and a medium confidence study using APOE*3-Leiden.CETP mice
reported decreased serum-free fatty acids (43%) and VLDL-triglyceride production rate (74%),
very-low-density lipoprotein (VLDL) half-life, and VLDL apolipoprotein production in animals
treated with 6 mg/kg-day PFHxS (Biiland etal.. 2011). The same study reported a 75% increase in
lipoprotein lipase in exposed mice fBiiland etal.. 20111. Two high confidence
reproductive/developmental toxicity studies also evaluated PFHxS-induced alterations in other
serum lipids. In SD rats, exposure to 10 mg/kg-day, decreased serum triglycerides by 27% in F0
males (Bute nhoff etal.. 2009: 3M. 2003). but a similar study using CD-I mice did not observe
significant treatment-related changes in serum triglycerides in male or female F0 animals at PFHxS
levels up to 3 mg/kg-day (Chang etal.. 2018). Medium and high confidence studies exposing using
C57BL/6J mice to 0.15 or 0.06 mg/kg-day PFHxS for 29 or 12 weeks respectively report no
significant effect on serum triglycerides fPfohl etal.. 2020: He etal.. 20221. Overall, a consistent
pattern of dose-dependent decreases in cholesterol and other lipids in the blood of animals exposed
to PFHxS were observed across high and medium confidence studies of varied design in both rats
and mice, although effects were largely absent in female rodents and studies that exposed mice to
PFHxS at lower doses. However, as described below there are limitations in using animal models
(including the APOE-modified mice) to emulate human lipid regulation.
3-299
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
„oS>^" *.V n0A^ lft o.fc^~ fvOl
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
Legend
Good (metric) or High confidence (overall)
+ I Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
NRl Not reported
Figure 3-77. Cardiometabolic effects, serum lipids - animal study evaluation
heatmap. For additional details see HAWC link.
# No significant change
Endpoint Name
Study Name
Experiment Name
Animal Description
A. Significant increase
Total Cholesterol (CHOL)
3M, 2000, 3981194
28 Day Oral
Rat, Crl:Cd Br (;?)
V Significant decrease
V
Rat, Crl:CdBr(*)
% Dose 6
V
Pfohl, 2020, 7021592
29-Week Oral
Mouse, C57BL/6J (?)
•
V
NTP, 2018, 4309363
28 Day Oral
Rat, Sprague-Dawley (°)
Rat, Sprague-Dawley (• ')
•
•
•
<
•
Chang,2018,4409324
Multi-Generational Oral
P0 Mouse, CD-1 (c1)
• • •
Butenhoff, 2009, 1405789
Multi-Generational Oral
P0 Rat, Sprague-Dawley (o)
~ • • f
P0 Rat, Sprague-Dawley ('11')
^ ^ ^
w w ™—
•
Total Cholesterol (CHOL), Blood
He, 2022,10273379
12-Week Oral
Mouse, C57BL/6 (
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
inflammatory alterations. Unfortunately, there is no single animal model that completely
recapitulates all the features of human disease fOppi etal.. 20191. Furthermore, there are
significant differences between rodent and human cardiovascular systems that should be taken into
consideration. Murine plasma cholesterol is approximately threefold lower, the major lipoprotein in
mice is HDL, not LDL fGetz and Reardon. 20121. and differences in bile acid composition contribute
to lower intestinal absorption of cholesterol and higher cholesterol excretion (Oppi etal.. 2019).
These differences contribute to significantly lower cholesterol levels in mice when compared with
humans and having lower cholesterol levels in turn confers protection from cardiovascular injuries
such as atherosclerosis fOppi etal.. 20191.
Although the available animal evidence suggests the cardiovascular system may be
responsive to PFHxS-induced responses, additional studies using experimental models and designs
that better emulate human disease would help to fully characterize the pathology of potential
cardiometabolic responses to this chemical. Future studies should focus on the use of genetically
manipulated or experimentally induced rodent models that can emulate human metabolic and
pathological conditions (Kodavanti etal.. 20151. For example, studies aimed at evaluating vascular
injuries such as atherosclerosis should focus on the use of animal models that can generate non-
HDL-based hypercholesterolemia such as LDL Receptor or apolipoprotein E (ApoE) null mice fGetz
and Reardon. 20121 and expose animals for sufficient time to develop of arterial injuries
(Daugherty etal.. 2017). Furthermore, future studies focused on potential effects to the
cardiovascular system should include analysis of physiological and biochemical parameters (e.g.,
heart rate, blood pressure, blood gases, and oxygen consumption), which are considered indicative
of adverse responses in the cardiovascular system (Gad. 2015).
Evidence Integration
The available evidence on PFHxS-induced cardiometabolic effects in humans is considered
slight (see Table 3-36). There is some evidence of an association between PFHxS exposure and
cardiometabolic effects in humans, specifically an indication of higher serum cholesterol levels,
although evaluation of the available results supports a significant concern for potential confounding
by other PFAS that prevents drawing a stronger judgment A similar association has been noted for
some other long-chain PFAS, including PFOA and PFOS fU.S. EPA. 2024a. b). However, there is little
evidence of an association between PFHxS exposure and cardiovascular disease, functional
endpoints of cardiovascular function (e.g., blood pressure), or other related cardiovascular risk
factors. It is possible that cholesterol is a more sensitive measure to PFHxS exposure and that the
exposure levels and contrast were inadequate to detect differences in disease risk. However,
without additional evidence, the lack of coherence across outcomes reduces confidence in the
evidence of the association with cardiovascular effects but indicates that the observed changes in
serum lipids could be related to hepatic toxicity.
The evidence from animal toxicity studies on PFHxS-induced cardiometabolic effects is
considered indeterminate. Animal studies report dose-related decreases in serum cholesterol and
3-301
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
triglyceride levels in male, but not female (largely), rats and mice. The direction of the observed
responses in animals is different from the observations made in human studies (e.g., decreased
serum lipids in animals versus reported increases in humans) and these effects may be caused by
PFHxS-induced alterations in hepatic lipoprotein metabolism (see Serum Biomarkers of Liver
Function Section 3.2.5). Heart weights and histopathology were not affected in exposed animals,
although these low confidence experiments were potentially insensitive. The downstream effects of
the metabolic alterations observed in the available studies are unclear in the absence of additional
experiments and measures of adverse responses in the cardiovascular system. Further,
interpretation of such results is not possible due to major limitations in the animal toxicity
database. As described above, commonly used laboratory rodent species are relatively resistant to
cardiotoxicity effects in part due to differences in lipid profiles fVeseli etal.. 20171. Furthermore,
the available evidence on PFHxS-induced cardiometabolic effects consists of short-term and
developmental exposure studies, whereas longer study durations (between 10 to 12 weeks in mice
Daugherty etal. (2017)) are generally preferred for evaluations cardiovascular system functions
and disease (e.g., atherosclerosis). These experimental design and database deficiencies limit the
interpretation of observed cardiometabolic changes in rodents and their applicability for informing
human health hazard.
The available animal and epidemiological evidence suggests but is not sufficient to infer
whether exposure to PFHxS might cause cardiometabolic effects in humans.27 This judgment is
based primarily on consistent increases in cholesterol in humans, but with limitations in the
available epidemiological studies that introduce uncertainty (see description above; slight
evidence) and also reflects an inability to interpret the available epidemiology evidence on PFHxS-
induced cardiovascular disease as well as the general lack of animal evidence (indeterminate
evidence) available to inform this health effect
27Hazards with evidence suggests judgments are typically not advanced for dose-response modeling due to
significant uncertainties associated with the available evidence.
3-302
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-36. Evidence profile table for PFHxS exposure and cardiometabolic effects
Evidence stream summary and interpretation
Evidence integration summary
judgment
Evidence from studies of exposed humans
Studies and
confidence
Factors that increase
certainty
Factors that decrease
certainty
Summary and key
findings
Evidence stream
judgment
Serum Lipids
25 medium and 9
low confidence
studies
• Consistency in
direction of
association for
cross-sectional
analyses in adults
• Medium
confidence
studies reporting
an effect
• Exposure-
response gradient
observed in five
studies
• Potential for residual
confounding across
PFAS
• Unexplained
inconsistency among
studies with
prospective
exposure
measurement and
for all studies of LDL
cholesterol and
triglycerides
Majority of studies in
adults report higher
serum cholesterol with
higher PFHxS exposure,
including 40%-60%
increases in the odds of
high cholesterol.
©oo
Slight
Generally consistent
findings for total
cholesterol in adults.
Evidence for other
related outcomes and
age groups is
inconsistent.
©OO
Evidence suggests, but is not
sufficient to infer
Primary Basis:
based primarily on consistent
increases in cholesterol in humans,
but with limitations in the available
epidemiological studies that
introduce uncertainty.
Human relevance:
The animal models used are
considered inadequate to inform
potential human cardiometabolic
responses with confidence.
Cross-stream coherence:
Evidence in animals is indeterminate
Other
Cardiovascular Risk
Factors
1 high, 18 medium,
and 7 low confidence
studies
• No factors noted
• Unexplained
inconsistency
Positive associations
reported for
hypertension in
adolescents and young
adults, but not other
adults or children. One
of four studies of
gestational
hypertension and two
of four studies of
preeclampsia reported
a positive association.
No association between
PFHxS exposure
3-303
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration summary
judgment
atherosclerosis or
ventricular geometry
Cardiovascular
Disease
2 medium and 3 low
confidence studies
• No factors noted
• Lack of cohererl ce
across outcomes in
low confidence
studies
• Unexplained
inconsistency - No
associations in the
two medium
confidence studies
No association with
cardiovascular disease
in medium confidence
studies. Low confidence
studies report higher
odds of cardiovascular
conditions and lower
odds of coronary heart
disease
Evidence from in vivo animal studies
Studies and
confidence
Factors that increase
certainty
Factors that decrease
certainty
Summary and key
findings
Evidence stream
judgment
Heart Weight/
Histopathology
3 low confidence
studies in adult rats:
• 28-d (x2)
• 44-d
• High and
medium
confidence
studies of serum
lipid measures
• Inconsistent findings
across studies
reporting on serum
lipids.
• Unclear biological
significance of
decreases in serum
lipids.
• No observed
PFHxS-induced
effects on heart
weight or
histopathology in
short-term,
potentially
insensitive studies.
• Dose-dependent
decreases in serum
cholesterol and
triglycerides.
OOO Indeterminate
Serum Lipids
5 high confidence
studies in adult rats:
• 28-d (x2)
• 42-d
• 44-d
• 84-d
2 medium quality
study:
• 42-d
• 203-d
3-304
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3.2.7. Hematopoietic Effects
Human Studies
One epidemiology study fliang etal.. 20141 examined the association between PFHxS
exposure and hematopoietic system effects, specifically the parameters from a complete blood
count (white and red blood cells, hemoglobin, platelets). This study was considered uninformative
due to lack of consideration of confounding, and thus no human studies were synthesized for
hematopoietic effects.
Animal Studies
The toxicity database for PFHxS-induced hematopoietic system effects consists of two 28-
day studies (NTP. 2018a: 3M. 2000a) in Crl:Cd Br and Sprague-Dawley rats, respectively; and one
multigenerational study in Sprague Dawley rats (Bute nhoff etal.. 2009). All studies exposed the
animals orally via gavage. Hematopoietic system-related outcomes evaluated by these studies
included non-immune blood cells counts and clotting parameters.
Evaluation of the available animal studies showed that these were well conducted for most
hematopoietic-related endpoints. All were considered high confidence. The available studies
generally examined PFHxS hematopoietic effects using doses that ranged between 0 and 10 mg/kg-
day in rats (Butenhoff et al.. 2009: 3M. 2000a) with the exception ofNTP (2018a) in which a range
of 0-50 mg/kg-day in female rats and 0-10 mg/kg-day in male rats was used. This approach was to
account for the pharmacokinetic (PK) sex differences that have been observed in rats, in
which PFHxS appears to have a lower mean half-life in female rats versus their male counterparts
(20.7 and 26.9 days respectively fKim etal.. 2016bll. No overt toxicity was observed at any of the
highest doses tested in any of the available studies. 3M f2000al and NTP
f2018al measured PFHxS related hematopoietic effects using the following parameters: hematocrit,
hemoglobin, platelet counts, prothrombin time, and red blood cell counts. NTP (2018a) also
measured PFHxS effects on reticulocyte counts. The study by Butenhoff et al. (2009) measured
hematocrit, hemoglobin, prothrombin time, and red blood cell counts in P0 males and females after
44 days of PFHxS (Butenhoff et al.. 2009).
Figure 3-79 below summarizes the results of animal study evaluations, and Figure 3-80
summarizes the experimental studies and their findings.
3-305
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
•gP- ; io° :
Reporting quality -
Allocation -
Observational bias/blinding -
Confounding/variable control -
Selective reporting and attrition -
Chemical administration and characterization -
Exposure timing, frequency and duration -
Endpoint sensitivity and specificity-
Results presentation -
Overall confidence A
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;
NRl Not reported
Figure 3-79. Hematological animal study confidence scores from repeated
PFHxS dose animal toxicity studies. For additional details see HAWC link.
Hemostasis, the physiological process of blood coagulation after injury, is dependent on
interactions between the vasculature and circulating plasma, platelets, blood cells and their related
molecules (Harris etal.. 2012: Gale. 20111. Clinical hematology assays like those available in the
PFHxS evidence based provide insight into bone marrow28 health as well as to assess blood clotting
function. Because of the dynamic interactions between hematopoietic cells and their related
molecules, information on the hematopoietic health of an organism is gained by the interpretation
of the collective battery of assays, rather than individual assay results (Harris etal.. 20121.
Therefore, the collective information from the entirety of the data provided from these available
assays was used to determine the potential for hazard posed by PFHxS on the hematopoietic
system.
Hematocrit fHctl. hemoglobin fHbl. and red blood cell fRBCl count
The hematocrit assay measures the amount (i.e., as a percent of blood volume) of red blood
cells (RBCs) in the blood. This measurement can provide insight on oxygen delivery capacity. All
three studies measured PFHxS effects on hematocrit. Two out of the three observed effects related
to PFHxS exposure 3M f2000al observed a significant decrease (5%-6%) in hematocrit in male and
28The bone marrow is the site of blood stem cell formation. Blood stem cells transform into a variety of blood
cells with distinct functions such as white cells (immune function]; red blood cells (oxygen carrying] and
platelet cells (clotting and injury repair] (Manz et al.. 20041.
3-306
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
female Crl:Cd Br rats following 28 days of daily oral exposure to 10 mg/kg-day PFHxS (the only
tested dose). In the multigenerational study Butenhoff et al. f20091 also observed a significant
(between 6% and 8%) decrease in hematocrit in male SD rats exposed to PFHxS at >3 mg/kg-day
for 44 days in F0 rats; however, females were unaffected. Further, changes in hematocrit were not
observed by NTP f2018al in male or female SD rats exposed for 28 days to doses of PFHxS up to 10
or 50 mg/kg-day, respectively. The difference in response to PFHxS on hematocrit measures in SD
rats between the Butenhoff et al., (20091 and NTP (2018a) studies may be due to the shortened
exposure duration 28 versus 42 days respectively.
Hemoglobin is an oxygen-carrying protein found in red blood cells. Its function is to deliver
oxygen from red blood cells to organs and tissues and to transport carbon dioxide from these
tissues back to the lungs. All three studies measured hemoglobin in response to PFHxS exposure
fNTP. 2018a: Butenhoff et al.. 2009: 3M. 2000a], Similar to the results for hematocrit, Butenhoff et
al. (2009) observed a significant decrease (between 5% and 7%) in hemoglobin in male, but not
female, rats orally exposed to >1 mg/kg-day PFHxS after 44 days of exposure, while 3M (2000a)
observed a significant decrease (4%-7%) in hemoglobin in male and female rats at the only dose,
10 mg/kg-day, at day 28. Changes in hemoglobin were not observed by NTP f2018al in either male
or female SD rats exposed to a similar dose range of PFHxS for 28 days.
Red blood cells carry oxygen, and their abundance can affect how much oxygen is received
by tissues and organs. RBC count provides a screening tool to assist in diagnosing or monitoring
conditions such as anemia. All studies measured RBC counts in response to PFHxS exposure, with
similar findings as for Hctand Hb, specifically: decreased RBC counts (between 7% and 8%) at
>3 mg/kg-day in male, but not female, rats exposed to PFHxS for at least 42 days (Butenhoff et al..
20091: decreased RBC counts (between 6% and 7%) in male and female rats exposed to 10 mg/kg-
day PFHxS for 28 days f3M. 2000al: and, in the second 28-day study, no changes in RBC counts in
male or female rats at up to 10 mg/kg-day (males) or 50 mg/kg-day (females) PFHxS fNTP. 2018al.
Reticulocytes count
Reticulocytes are RBC precursors produced in the bone marrow and released into the
bloodstream where they develop into mature RBCs. Reticulocyte counts can provide information
about the health of the bone marrow and its ability to produce RBCs. Only the NTP study measured
reticulocyte counts., A significant decrease (10% -27%) in number of reticulocytes was observed in
SD male rats at >1.25 mg/kg-day and a significant increase (40%) in reticulocyte counts in female
rats at 3.12 mg/kg-day, but not higher or lower doses (NTP. 2018a). The other two studies
(Butenhoff etal.. 2009: 3M. 2000a) did not evaluate reticulocytes, preventing interpretation as to
whether a compensatory response of the bone marrow to the observed effects on red blood cell
parameters might exist
3-307
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Platelet count
Platelets are cell fragments found within the blood that are critical for clot formation when
blood vessels are damaged. Together with prothrombin time, platelet counts provide information
on coagulation potential. Two studies, 3M f2000al and NTP f2018al. measured PFHxS effects on
platelet counts. 3M (2000a) observed a significant decrease (ll%-26%) in total platelet numbers
in male and female rats exposed to 10 mg/kg-day PFHxS for 28 days. NTP (2018a) did not report
any changes in platelet counts in male or female rats exposed to PFHxS for 28 days at up to
10 mg/kg-day (males) or up to 50 mg/kg-day (females).
Prothrombin time
Prothrombin time is an assay measuring the amount of time it takes blood to clot. Two
studies, Butenhoff et al. (2009) and 3M (2000a). measured PFHxS effects on prothrombin time.
Butenhoff et al. (2009) observed a significant increase (between 3%-6%) in prothrombin time in
male, but not female, rats at 0.3, 3 and 10 mg/kg-day (doses tested: 0.3,1, 3, and 10 mg/kg-day).
Under similar study conditions, the single dose (10 mg/kg-day) 28-day study by 3M f2000al
observed that prothrombin time significantly decreased (between 5%-6%) in female rats and male
rats in response to 10 mg/kg-day PFHxS. Figure 3-80 below summarizes the study design and
results for each hematology parameter described in these three studies.
3-308
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Endpoint Name
Study Name
Experiment Name
Animal Description
Observation Time
Hematocrit (Hct)
3M, 2000, 3981194
28 Day Oral
Rat, CrlrCd Br (•:)
Study Day 28
Recovery Day 14
Recovery Day 28
Rat, Crl:Cd Br (,)
Study Day 28
Recovery Day 14
Recovery Day 28
Butenhoff, 2009, 1405789
Multi-Generational Oral
P0 Rat, Sprague-Dawley (,5l
I Day 44
NTP, 2018, 4309363
28 Day Oral
Rat, Sprague-Dawley (2)
Day 29
Rat, Sprague-Dawley (;;)
Day 29
Hemoglobin (Hb)
3M, 2000, 3981194
28 Day Oral
Rat, Crl:Cd Br (f)
Study Day 28
Recovery Day 14
Recovery Day 28
Rat, Crl:Cd Br (2)
Study Day 28
Recovery Day 14
Recovery Day 28
Butenhoff. 2009, 1405789
Multi-Generational Oral
P0 Rat, Sprague-Dawley ( ':
I Day 44
NTP. 2018, 4309363
28 Day Oral
Rat, Sprague-Dawley ( : )
Day 29
Red Blood Cells (RBC)
3M. 2000, 3981194
28 Day Oral
Rat, Sprague-Dawley (. ')
Rat, Crl:Cd Br (f?)
Day 29
Study Day 28
Recovery Day 14
Recovery Day 28
Rat, Crl:Cd Br (,)
Study Day 28
Recovery Day 14
Recovery Day 28
Butenhoff, 2009, 1405789
Multi-Generational Oral
P0 Rat, Sprague-Dawley ( J]
* Day 44
NTP, 2018, 4309363
28 Day Oral
Rat, Sprague-Dawley (¦-¦)
Day 29
Reticulocytes (RET)
NTP, 2018, 4309363
28 Day Oral
Rat, Sprague-Dawley ( J)
Rat, Sprague-Dawley (2)
Day 29
Day 29
Rat, Sprague-Dawley (;)
Day 29
Platelet (PLAT)
3M, 2000, 3981194
28 Day Oral
Rat, CrlrCd Br ( :')
Study Day 28
Recovery Day 14
Recovery Day 28
Rat, CrlrCd Br (i)
Study Day 28
Recovery Day 14
NTP. 2018, 4309363
28 Day Oral
Rat, Sprague-Dawley (2)
Recovery Day 28
Day 29
Rat, Sprague-Dawley (-; )
Day 29
Prothrombin Time (PT)
3M. 2000, 3981194
28 Day Oral
Rat, CrlrCd Br (-')
Study Day 28
Recovery Day 14
Rat, CrlrCd Br (2)
Recovery Day 28
Study Day 28
Recovery Day 14
Recovery Day 28
Butenhoff, 2009, 1405789
Multi-Generational Oral
P0 Rat, Sprague-Dawley
I Day 44
PFHxS-Related Hematopoietic Effects
v
V
# No significant change
A Significant increase
V Significant decrease
Figure 3-80. Hematopoietic effects of PFHxS exposure in animals. For
additional details see HAWC link.
Evidence Integration
The currently available evidence is inadequate to assess whether PFHxS exposure may
cause hematopoietic effects in humans. The evidence informing the potential for PFHxS exposure to
cause hematopoietic effects is limited to hematology measures in three high confidence studies in
rats, with exposure durations of 28-44 days, and which together are considered to provide slight
evidence (see Table 3-37), Two of the three studies were consistent to some degree, demonstrating
a pattern of changes in male rats. Specifically, male rats exposed to PFHxS at doses ranging from 0
to 10 mg/kg-day for 28-44 days exhibited decreases in multiple RBC parameters (i.e., Hct, Hb, and
RBCs}. However, there were inconsistencies, such as reported decreases in platelets counts in one
3-309
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
28-day study (3M. 2000a], which were not observed in a separate 28-day study with similar study
design fNTP. 2018a! Prothrombin time was reported to increase in male rats as a result of PFHxS
exposure in one study fButenhoff et al.. 20091 and decrease in male and female rats in another f3M.
2000a). Butenhoff et al. f20091 did not measure hematological parameters in female rats). There
was unexplained inconsistency across studies. The two 28-day studies fNTP. 2018a: 3M. 2000al
reported opposite findings, despite similar study designs and rat strains (the Crl:CD Br rats used by
3M (2000a) are a Sprague Dawley strain). Specifically NTP (2018a) did not observe consistent
effects on these same parameters (i.e., Hct, Hb, RBCs, and platelets were unchanged; reticulocytes
were decreased) in male animals exposed to doses of PFHxS ranging from 0.625 to 10 mg/kg-day.
Thus, there is no clear explanation (e.g., study methods; doses; exposure duration; species, strain,
or sex) for this inconsistency.
As noted above, the observations in male rats across RBC parameters and other measures
reported in 3M (2000a) and Butenhoff et al. (2009) appear somewhat coherent. RBCs play an
important role in hemostasis, as increased Hct has been shown to increase blood viscosity
(reviewed in Litvinov and Weisel (2017)). Additionally, RBCs interact with platelets and modulate
their reactivity through cell signaling molecules or through direct adhesive RBC-platelet
interactions (reviewed in Litvinov and Weisel f20171I Therefore, if RBC counts, along with Hb and
Hct measures are decreased following PFHxS exposure, then it is reasonable that an increase in
prothrombin time would be observed.
The observed effects in the study by Butenhoff et al. (2009) were dose dependent, with
effects generally observed at or greater than 3 mg/kg-day, although some changes at lower doses
were also noted. The duration dependence of these effects could not be determined; the 28-day
study by 3M f2000al that reported similar findings to those observed by Butenhoff et al. f20091
only tested 10 mg/kg-day and the PFHxS-related effects on RBC parameters were no longer
observed at or after recovery day 14. Further the magnitude of effects across the various
hematological endpoints measured (ranging from about 4% to 8%) is small and their biological
significance is questionable. The animal evidence is considered slight due to the questionable
biological significance and unexplained inconsistencies in the reported PFHxS effects on
hematology among the available studies.
The currently available evidence is inadequate to assess whether PFHxS may cause
adverse hematopoietic effects in humans given sufficient exposure conditions.29 This conclusion is
based on the three available animal studies that assessed PFHxS doses ranging from 0 to 10 mg/kg-
day in male rats.
29The "sufficient exposure conditions" are more fully evaluated and defined for the identified health effects
through dose-response analysis in Section 5.
3-310
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-37. Evidence profile table for PFHxS hematopoietic effects
Evidence stream summary and interpretation
Evidence integration summary
judgment
Evidence from studies of exposed humans (see Hematopoietic Human Studies Section)
Studies and confidence
Factors that
increase certainty
Factors that decrease
certainty
Summary and key
findings
Evidence stream
judgment
Inferences across evidence
streams
No informative studies
(1 uninformative)
No informative studies identified
ooo
Indeterminate
©OO
Evidence is inadequate
Primary basis:
Despite coherent decreases in
multiple RBC parameters in two
studies in male rats, there were
unexplained inconsistencies
across studies and an unclear
biological significance of effect
magnitude for most endpoints
Human relevance:
Without evidence to the
contrary, effects in rodent
models are considered relevant
to humans.
Cross-stream coherence:
NA; human evidence
indeterminate
Susceptible Populations
and lifestages:
NA
Evidence from in vivo animal studies (see Hematopoietic Animal Studies Section)
Studies and confidence
Factors that increase
certainty
Factors that decrease
certainty
Summary and key
findings
Evidence stream
judgment
3 high confidence
studies in rats
• All high
confidence
studies
• Unexplained
inconsistencies
across sexes and
studies.
• Unclear biological
significance of
effect magnitude
for most endpoints
(~4%-8%)
2 of the 3 studies
reported male rats
exposed for 28-44 d
exhibited small
decreases in multiple,
coherent RBC
parameters (i.e., Hct,
Hb, and RBCs), as well
as decreases in
prothrombin time.
However, these effects
were observed in both
sexes in one study,
only males in a second
study, and results
were null in the third.
©oo
Slight
3-311
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3.2.8. Female Reproductive Effects
Human Studies
Studies of possible female reproductive effects of PFHxS are available for fecundity (i.e.,
time to pregnancy), reproductive hormones, pubertal development, gynecological conditions
(endometriosis and polycystic ovary syndrome [PCOS]), ovarian reserve (including POI), menstrual
cycle characteristics, and developmental measures (anogenital distance). While the evidence for
each of these outcomes is synthesized separately, many of them are closely interconnected, with
almost all of the outcomes having the potential to influence fecundity, as well as each other. For
example, fecundity may be reduced by gynecological conditions and diminished ovarian reserve.
Both of these may influence or be influenced by reproductive hormones levels, as are menstrual
cycle characteristics, timing of pubertal development, and anogenital distance. The direction of
association across these related outcomes is not always straightforward, which complicates
considerations of coherence across outcomes. For example, low levels of anti-Mullerian hormone
(discussed with ovarian reserve) may indicate difficulty getting pregnant (i.e., decreased fecundity)
but high levels may be associated with PCOS, which may also decrease fecundity. In addition,
preterm birth and spontaneous abortion could be driven by either female reproductive or
developmental toxicity. These latter two outcomes are reviewed in the developmental section of
this assessment but are also included in the consideration of coherence across outcomes for female
reproductive effects. Reverse causation is a concern for many of the outcomes discussed in this
section, due to differences in excretion introduced by pregnancy, lactation, and menstruation.
Several studies also evaluated effects on anogenital distance (AGD), a developmental
outcome that is responsive to variations in reproductive hormones ((Foster and Gray. 2013: Dean
and Sharpe. 20131. See Section 3.2.3).
In total, 35 epidemiology studies are available for these outcomes. The study evaluations
are summarized below for each outcome or group of outcomes.
Fecundity (time to pregnancy)
Fecundity is the biological capacity to reproduce. Time to pregnancy, defined as the number
of calendar months or menstrual cycles from the time of cessation of contraception to detection of
pregnancy, is the primary outcome measure used to study fecundity. Many of the other outcomes
described in this section contribute to fecundity. There are nine epidemiology studies that report
on the association between PFHxS exposure and fecundity and related outcomes. A summary of the
study evaluations is presented in Figure 3-81, and additional details can be obtained from HAWC.
One study (Cariou etal.. 2015) was considered uninformative due to lack of consideration of any
potential confounders and excluded from further analysis. Of the remaining studies, two were
preconception cohorts and considered medium confidence fVestergaard et al.. 2012: Crawford etal..
20171. and four were pregnancy cohorts and considered low confidence fVelez etal.. 2015:
3-312
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
largensen etal.. 2014: Bach etal.. 2015: Bach etal.. 20181. The pregnancy cohorts were rated lower
due to potential selection bias from excluding women who were unable to conceive. Two studies
examined related outcomes in women undergoing treatment for infertility. Wang etal. f2021al
describes a cohort of women undergoing in vitro fertilization (IVF)-embryo transfer and reports
rates of human chorionic gonadotropin (hCG) negativity following treatment; this study was rated
medium confidence. Kim etal. (2020c) is a cross-sectional study of fertilization rate in women who
underwent fully stimulated assisted reproductive treatment at an IVF clinic; this study was rated
low confidence primarily due to concerns for residual confounding.
i i i i i i i —i—
Bach, 2015, 3981559-
Bach, 2018, 5080557-
Cariou, 2015, 3859840-
Crawford, 2017, 3859813-
Jorgensen, 2014, 2851025-
Kim, 2020, 6833596 -
Vestergaard, 2012, 1332472 -
Velez, 2015, 2851037-
Wang, 2021, 10176703-
-
+
+
+
++
-
+
-
-
+
+
+
++
¦
+
-
¦
+
-
B
-
-
+
~
+
o
*
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 3-81. Summary of study evaluation for epidemiology studies of
fecundity. For additional details see HAWC link.
The results for the association between PFHxS exposure and fecundity are presented in
Table 3-38. A fecundability ratio less than 1 indicates a decrease in fecundity/increase in time to
pregnancy. Of the seven studies, two low confidence studies (Velez etal.. 2015: Bach etal.. 20181
reported a statistically significant decrease in fecundity/increase in time to pregnancy with
increased exposure (only in parous women in Bach etal. (201811. The remaining studies reported
no decrease in fecundity. In addition to the time to pregnancy results, three studies fVestergaard et
al.. 2012: Velez etal.. 2015: Bach etal.. 20151 also analyzed infertility as an outcome. Only the low
confidence study by Velez etal. (2015) reported an increase in infertility with increased exposure
(OR:1.27 (95% CI: 1.09,1.48). Neither study of IVF outcomes (fertilization rate, hCG negativity)
reported an association between PFHxS exposure and reduced fertility.
3-313
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
There is unexplained inconsistency in the evidence for this association. A decrease in
fecundity with higher exposure was observed in two low confidence studies, but not the other four
studies, which included the two medium confidence studies. The primary limitation in both Bach et
al. f20181 and Velez etal. f20151 was the potential for selection bias resulting from enrollment of
participants during pregnancy. This approach would exclude women who were ultimately unable to
conceive. If there is a true association between PFHxS and fecundity, this would be a bias against
the most exposed women, which would likely result in an underestimate of the association.
However, if there is no association, selection would not be related to exposure, so it is unlikely to
cause bias. Thus, the observed associations should not be dismissed as due to selection bias. On the
other hand, as suggested by the authors, the lack of association in nulliparous women in Bach et al.
T20181 suggests the possibility of confounding by factors related to previous pregnancies in the
results of parous women, which could also exist in Velez etal. f20151. where the population was
only 29% nulliparous. Overall, there is considerable uncertainty in the strength of this
inconsistently observed association.
Table 3-38. Summary of results for epidemiology studies of fecundity
Reference,
confidence
Population
Exposu re
median
(IQR)
Comparison for
effect estimate
Fecundability ratio
(95% CI)
Bach et al. (2015),
low
Aarhus pregnancy cohort (2008-
2013), Denmark; 1,372 nulliparous
women
0.5 (0.4-
0.6)
0.1 ng/mL increase
1.00 (0.99,1.01)
Quartiles vs. Q1
Q2: 1.05 (0.89,1.24)
Q3: 1.06 (0.89,1.25)
Q4: 1.12 (0.94,1.32)
Bach et al. (2018),
low
Danish National Birth Cohort sub-
sample (1996-2002), Denmark
Nulliparous women (n = 638)
0.9 (0.7-
1.2)
Quartiles vs. Q1
Q2: 1.03 (0.81-1.32)
Q3: 1.05 (0.83-1.35)
Q4: 0.92 (0.72-1.18)
Parous women (n = 613)
Q2: 0.74 (0.55-1.01)
Q3: 0.79 (0.59-1.04)
Q4: 0.60 (0.45-
0.80)*
Velez et al.
(2015), low
MIREC pregnancy cohort (2008-
2011), Canada; 1,625 women (29%
nulliparous)
1
SD increase
0.91 (0.86,0.97)*
Vestergaard et al.
Preconception cohort (1992-1995),
Denmark; 222 nulliparous women
1.2 (0.9-
1.8)a
log-unit increase
1.33 (1.01,1.75)
(2012), medium
Above median vs.
below
1.29 (0.90,1.83)
Crawford et al.
(2017), medium
Time to Conceive cohort (2008-
2009), U.S.; 99 women (40%
nulliparous)
1.6 (GM)
dichotomous cutoff
75th percentile
Cycle-specific model
1.40 (0.79,2.49)
d-specific model
0.96 (0.31,1.71)
3-314
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference,
confidence
Population
Exposu re
median
(IQR)
Comparison for
effect estimate
Fecundability ratio
(95% CI)
Jgrgensen et al.
(2014), low
INUENDO pregnancy cohort (2002-
2004), Greenland, Poland, Ukraine;
938 women
1.9
In-unit increase
Pooled
0.97 (0.85,1.11)
Greenland (n = 448, 31%
nulliparous)
2.0
Tertiles vs. T1
T2: 1.05 (0.79,1.38)
T3: 0.90 (0.68,1.19)
Poland (n = 203, 92% nulliparous)
2.4
T2: 0.86 (0.57,1.30)
T3: 0.94 (0.62,1.42)
Ukraine (n = 287, 79% nulliparous)
1.6
T2: 0.85 (0.59,1.23)
T3: 1.11 (0.78,1.58)
*p < 0.05.
aln participants with pregnancy.
Reproductive hormones in females
Reproductive hormones and related proteins examined in the evaluated studies include
testosterone, estradiol, insulin like growth factor 1 (IGF-1), follicle stimulating hormone (FSH),
luteinizing hormone (LH), progesterone, as well as sex hormone-binding globulin (SHBG), all
measured in blood, or in one study, saliva. Reproductive hormone levels are associated with all of
the other female reproductive outcomes discussed in this section, but the relationships are often
complex.
Key issues for the evaluation of studies of reproductive hormones were sample collection
and processing. For testosterone, LH, FSH, and prolactin, due to diurnal variation, blood sample
collection should occur at the same time of day for all participants, and if not, time of collection
must be accounted for in the analysis. If there is no consideration of time of collection, the study is
classified as deficient for outcome ascertainment and low confidence overall for these hormones as
this is expected to result in nondifferential outcome misclassification. This applied to eight studies
(Timmermann etal.. 2022: Osterman etal.. 2008: Martin. 1978: Lopez-Espinosa et al.. 2016: Lewis
etal.. 2015: Heffernan etal.. 2018: Elavarasi etal.. 2019: Avcan. 20191. Lastly, the etiologic timing of
PFHxS exposure relevant for influencing reproductive hormones is unclear and likely dependent on
several factors, and thus all exposure windows with available data were considered, including
cross-sectional since circulating hormone levels can be rapidly upregulated or downregulated in
response to a change in exposure.
Fifteen studies (reported in 16 publications) examine potential associations between PFHxS
exposure and reproductive hormones. One study was deemed uninformative due to multiple
serious deficiencies in the participant selection, confounding and analysis domains fMcCov etal..
20171. Most studies examined only testosterone and estradiol and measured exposure and outcome
concurrently, though some studies measured additional hormones and/or measured exposure
prospectively (prenatal exposure in Maisonet et al. (2015).lensen etal. (2020b). and Timmermann
3-315
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
etal. (2022). early pregnancy for outcomes in late pregnancy (Yang etal.. 2022b). and
premenopause in Harlow etal. f202111. Eight studies f Zhang etal.. 2018b: Yang etal.. 2022b: Wang
etal.. 2021b: Timmermannetal.,2022: Lewis etal.. 2015: Heffernan etal.. 2018: Harlow etal..
2021: Barrett etal.. 20151 examined associations in adults, three studies fZhou etal.. 2016:
Maisonet etal.. 2015: Lewis etal.. 20151 in adolescents, one study fLopez-Espinosa etal.. 20161 in
children, and three studies (Yao etal.. 2019: Liu etal.. 2020b: Tensen etal.. 2020b) in infants. The
study evaluations are summarized in Figure 3-82. Six studies were considered medium confidence
and seven were low confidence. However, of the medium confidence studies, two did not consider
time of day of sample collection for hormones and were thus low confidence for testosterone (Yao
etal.. 2019: Lopez-Espinosa etal.. 20161. Notably, two studies fZhangetal.. 2018b: Heffernan etal..
20181 included participants with gynecological conditions (polycystic ovarian syndrome [PCOS]
and premature ovarian insufficiency (POI), respectively). These conditions are associated with
changes in reproductive hormone levels, and thus stratified results were used. These studies may
also be affected by reverse causality, as menstrual cyclicity is associated with both hormone levels
and these conditions, and menstrual cycle length/regularity may influence PFAS excretion
(discussed further below, see Menstrual cycle characteristics below).
3-316
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Barrett, 2015, 2850382-
++
+
+
+
++
+
+
Harlow, 2021, 8569305-
~
+
++
+
+
Heffernan, 2018, 5079713-
-
+*
+
-
>
-
Jensen, 2020, 6311643-
+
-
+
++
-
-
Lewis, 2015, 3749030-
++
+
-
+
+
-
Liu, 2020, 6569227 -
+
+
+
+
++
-
+
Lopez-Espinosa, 2016, 3859832-
++
++
+*
+
++
+
+*
Maisonet, 2015, 3859841 -
+
+
+
+
++
+
+
Mccoy, 2017, 4238432-
-
+
++
-
-
~
Timmermann, 2022, 10176553-
+
-
+
•-
-
¦
Wang, 2021, 7404063-
+*
++
++
+
+*
Yang, 2022, 10176804-
+
+
-
+
Yao, 2019, 5187556-
+
+*
++
-
-
Zhang, 2018, 5079665-
¦
¦
+*
+
-
-
¦
Zhou, 2016, 3856472-
+
+
+
+
•
-
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 3-82. Summary of study evaluations for epidemiology studies of female
reproductive hormones. For additional details see HAWC link. Multiple
publications of the same study: Yao etal. (2019) also includes Yao etal. (2021).
Estradiol
Nine studies examined estradiol levels in association with PFHxS. In six studies of adults,
one low confidence study reported lower estradiol with higher exposure in women with premature
ovarian insufficiency (POI) ((3: -0.19 (95% CI: -0.37, -0.02)) but no change in women without POI
(Zhang et al.. 2018b). Conversely, one low confidence study reported higher estradiol with higher
exposure in adult women without PCOS ((3: 223, SE 255), although this was not statistically
significant, and no change was observed in women with PCOS (Heffernan etal.. 20181. In both of
these studies, the results in controls (without POI or PCOS) are more straightforward to interpret
since the presence of these conditions may influence hormones levels and as discussed below, PFAS
levels. The remaining studies of adults, all medium confidence, including one in healthy
nonpregnant women f Barrett etal.. 20151. one in pregnant women fYang etal.. 2022bl. one in
premenopausal (or transitioning to menopause) women (Harlow etal.. 2021). and one in
3-317
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
postmenopausal women (Wang etal.. 2021bl. reported no association. In younger populations, a
single low confidence study of adolescents reported no association fZhou etal.. 20161. while a
single low confidence study of children fLopez-Espinosa etal.. 20161 reported higher ln-estradiol
levels with higher PFHxS (2.1% difference (95% CI: -2.2, 6.5)). Lastly, in one medium confidence
study of infants fYao etal.. 20191. there was higher estradiol with higher PFHxS ((3: 0.30 (95% CI:
0.27, 0.37)). Overall, there are three studies reporting higher estradiol (one statistically significant)
in at least one subpopulation, one study reporting lower estradiol, and five studies reporting no
association with PFHxS exposure. There was no apparent pattern of association by study
confidence or study sensitivity ratings/exposure levels and contrast, and thus these inconsistent
results are difficult to interpret
Testosterone
As described above, most studies were low confidence for testosterone. In adult women, there were
five studies available, all low confidence except Harlow et al. (2021). Two of these reported
nonstatistically significant inverse associations between testosterone and PFHxS exposure. Lewis et
al. f20151 reported results stratified by age group and observed stronger associations in lower ages
((3 (95% CI) for 20-<40: -3.3 (-8.7, 2.5), 40-<60: -2.4 (-8.7, 4.3), 60-80: -0.2 (-8.3, 8.7). Zhang et
al. f2018bl. also reported an inverse association in controls without POI ((3 -0.11, 95% CI: -0.27,
0.05). In contrast, Heffernan etal. f 20181 reported a statistically significant positive association in
controls without PCOS ((3 0.50, SE 0.17). Studies in pre- and post-menopausal women reported no
association (Wang etal.. 2021b: Harlow etal.. 20211. In adolescents, three studies were available.
Maisonet et al. (20151. a medium confidence study, reported higher testosterone levels in 15-year-
old girls with the increasing tertiles of PFHxS exposure, although there was no apparent exposure-
response gradient across the narrow tertiles (1.3-1.9 ng/mL ((3: 0.18 (95% CI: 0.00,0.37), and >1.9
ng/mL ((3: 0.18 (95% CI: 0.00, 0.35) compared with <1.2 ng/mL PFHxS). Lewis etal. f20151
reported an inverse association ((3 -5.3, 95% CI: -11.6,1.5) (with median exposure of 0.8 ng/mL)
while Zhou etal. (2016) reported no association (with mean PFHxS exposure of 1.2 ng/mL). One
low confidence study in children reported no association with testosterone (Lopez-Espinosa et al..
20161 with median exposure of 7 ng/mL, and one low confidence study in infants (Yao etal.. 20191
reported an inverse association ((3 = -0.16 (95% CI: -0.36, 0.04) with median exposure of
0.3 ng/mL.
Overall, there are 3 of 10 studies reporting inverse associations between testosterone and
PFHxS exposure, including two of five studies in adults, one of three studies in adolescents, zero of
one study in children, and one of one study in infants. In addition, one study in adults reported a
positive association. There was no apparent pattern of association by exposure levels. The study
with the highest exposure levels and greatest contrast (Lopez-Espinosa etal.. 20161 reported no
association, while inverse associations were observed in studies with narrow contrast f Zhang etal..
2018b: Yao etal.. 20191. although not statistically significant.
3-318
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Other hormones and related molecules
For other hormones and related molecules, Lopez-Espinosa etal. f20161 examined
associations between PFHxS and IGF-1, reporting inverse, although nonmonotonic in categorical
analyses, associations. Sex hormone-binding globulin (SHBG) was not associated with PFHxS levels
in four studies (Wang etal.. 2021b: Maisonetetal.. 2015: Heffernan etal.. 2018: Harlow etal..
2021). Barrett etal. (2015) observed no evidence of association with luteal phase progesterone in
saliva in normally cycling women, while in infants, Liu etal. (2020b) reported a small but not
statistically significant positive association (2.8% increase) with progesterone. Zhang etal. (2018b)
reported positive associations with FSH ((3 0.16, 95% CI: 0.04, 0.28) and prolactin ((3 0.11, 95% CI:
-0.01, 0.22) in women with premature ovarian sufficiency, but no association in controls, while
Harlow etal. f20211 reported an inverse association with FSH only in nulliparous women (-4.62,
95% CI; -8.60, -0.47). In Tensen etal. (2020b). there were positive associations (p > 0.05) with LH,
androstenedione, and DHEAS in infant girls. Lastly, Timmermann etal. (2022) reported a
statistically nonsignificant inverse association with prolactin in pregnant women at gestational
week 10 (3.1% decrease) but no difference at gestational week 28.
3-319
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3 c,e\6° cfJ0
sf>"
Campbell, 2016, 3860110-
++
¦
¦
+
+
+
¦
a
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)
Carwile, 2021,9959594 -
+
++
+
+
+
+
Christensen, 2011, 1290803-
+
B
++
+
+
+
+
+
y
Crawford, 2017, 3859813-
+
++
++
-
+
+
+
+
Ding, 2022, 10273297-
+
+
+
+
+
+
+
+
Ernst, 2019, 5080529-
++
+
+
+
++
-
+
+
Hammarstrand, 2021, 9959666-
+
+
-
+
+
+
-
Louis, 2012, 1597490-
+
-
+
++
-
+
-
Romano, 2016, 3981728-
+
+
++
+
+
+
Rosen, 2018, 5080190-
+
+
+
++
+
+
+
Singer, 2018, 5079732-
++
•
-
+
++
-
+
-
Timmermann, 2017, 3981439-
+
++
+
++
++
+
+
+
Timmermann, 2022, 10176553-
+
++
+
+
+
-
+
+
Vagi, 2014, 2718073-
-
-
+
-
+
-
+
-
Wang, 2017, 3856459-
-
-
-
-
+
+
+
-
Wise, 2022, 9959470-
+
-
-
-
+
+
+
-
Zhang, 2018, 5079665-
-
-
+
+
+
-
-
-
Zhou, 2017, 3859799-
++
-
-
+
+
+
+
-
Figure 3-83. Summary of study evaluation for epidemiology studies of other
female reproductive effects (menstrual cycle characteristics, gynecological
conditions, ovarian reserve, and pubertal development). For additional details
see HAWC link.
Menstrual cycle characteristics
Three epidemiology studies report on the association between PFHxS exposure and
menstrual cycle characteristics. One was a pregnancy cohort in Norway fSinger etal.. 20181. one
was a cross-sectional study of participants in a preconception cohort in China (Zhou etal.. 2017a).
and one was a cross-sectional study of reproductive aged Black women in the U.S. (Wise etal..
20221. For this outcome, there is potential for reverse causation because menstruation is one of the
mechanisms by which PFAS are removed from the body. It is expected that a longer cycle would
3-320
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
result in less clearance of PFAS, and therefore higher PFAS in the body, possibly resulting in inflated
effect estimates. Thus, all three studies were considered low confidence (see Figure 3-83). There
were also concerns for potential outcome misclassification due to self-report, since the
questionnaires used were not validated. Zhou etal. f2017al reported an increase in odds of
irregular and long cycle (OR (95% CI) for continuous exposure = 1.80 (1.17,2.77) and 1.73
(1.13,2.65), respectively), and a decrease in the odds of menorrhagia (OR = 0.14 (0.06,0.36). Singer
etal. (2018) also reported higher PFHxS levels in participants with irregular (4% change, 95% CI:
-3,11) and long cycles (5% change, 95% CI; -4,14), although neither was statistically significant
Wise etal. f20221 reported lower intensity of menstrual bleed with higher exposure, but no
difference in bleed length in days. These associations with irregular and long cycles in two studies
and lower bleeding in one study is consistent with either a true association or reverse causation
due to less PFAS excretion through menstruation compared with women with regular cycles, and it
is difficult to interpret with currently available evidence.
Gynecological conditions
Four epidemiology studies report on the association between PFHxS exposure and
endometriosis. Three of the studies were cross-sectional, which decreases confidence for this
chronic outcome due to the inability to establish temporality fWang etal.. 2017: Campbell etal..
2016: Buck Louis et al.. 20181. There is potential for reverse causality as described above since
endometriosis can influence the menstrual cycle, and this could be toward a protective direction
given that endometriosis can be associated with heavier and more frequent bleeding. Because of
this issue, these studies were classified low confidence, although the study by Buck Louis et al.
f20181 is considered stronger in other study design aspects than the remaining two studies; this
study included two groups of women, one group scheduled for surgery (laparoscopy or
laparotomy), and one group identified through a population database who underwent pelvic MRI to
identify endometriosis (Buck Louis et al.. 2018) (see Figure 3-83). The remaining two studies were
deficient for outcome ascertainment, specifically due to self-report of endometriosis diagnosis
(Campbell etal.. 2016) and case definition including only endometriosis-related infertility among
surgically confirmed cases (Wang etal.. 2017). Both of these methods are likely to include
asymptomatic cases among the controls. In addition, one study that reported results only on a
mixture of PFAS was determined to meet the PECO criteria due to very high exposure to PFHxS in
participants. Hammarstrand et al. f20211 examines a population in Ronneby, Sweden with high
PFAS contamination in drinking water. This study estimated exposure using residence location
linked to data on the municipal water supply (validated against serum measurements in a
subsample) and was thus not able to develop individual PFAS estimates. PFHxS and PFOS were
predominant in this population (subsample mean serum levels in participants living in the area at
the time of high contamination were 243 and 279, respectively, compared with 15 for PFOA), so any
effect observed can likely be largely attributed to those PFAS, but it is not possible to separate their
effects, and thus the study is considered low confidence.
3-321
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Two of the low confidence studies, including the Buck Louis et al. (2018) study, reported
slightly increased odds of endometriosis with higher exposure, although the estimates were
imprecise fBuck Louis etal. f20181: operative sample OR: 1.14 (95% CI: 0.58,2.24); population
sample OR: 1.52 (95% CI: 0.40.5.801: Campbell etal. C20161 OR (95%) versusTl: T2: 0.66
(0.37,1.19), T3: 0.47 fO. 25.0.8711. Hammarstrand etal. f20211 found no association with
endometriosis despite the very high exposure to PFHxS and PFOS.
In addition, two studies examined PCOS and PFHxS exposure, including the study in
Ronneby, Sweden (Hammarstrand etal.. 2021) described above and a case-control study in the U.S.
fVagi etal.. 20141. Vagi etal. f20141 suffers from potential for reverse causality due to association
with menstruation, similar to the studies of endometriosis. Because PCOS is associated with
irregular menstruation and thus less frequent bleeding it is possible that effect estimates will be
inflated. This study is low confidence for this reason and concerns with participant selection and
confounding. There was no association between PFHxS and PCOS, but due to the study limitations,
this is difficult to interpret. Hammarstrand et al. (2021) reported higher odds of PCOS and fibroids
in participants with the highest exposure (HR: 2.18,95% CI: 1.43, 3.34), but this is also difficult to
interpret due to the co-exposure with PFOS.
Ovarian reserve
Three studies examined the association between PFHxS exposure and ovarian reserve, an
indication of a woman's egg count or remaining reproductive potential. The available studies were
two medium confidence studies, a cohort (Crawford etal.. 2017) and a nested case-control study
(Donley etal.. 2019). examining anti-Mullerian hormone (AMH), and a low confidence case-control
study examining POI f Zhang etal.. 2018bl. AMH is commonly used as an endocrine marker for age-
related decline of ovarian reserve in healthy women, with reduced AMH an indication of small
primordial follicle pool, as well as predicting poor oocyte yield for in vitro fertilization fGrvnnerup
etal.. 2012). However, a single measurement in healthy women may not be informative in
predicting fecundity (ACOG. 2019) and, as mentioned above, elevated levels of AMH are associated
with PCOS, so these results should be interpreted with caution. In contrast to AMH, POI is a more
specific outcome (defined as an elevated FSH level greater than 25 IU/L on two occasions more
than 4 weeks apart and oligo/amenorrhea for at least 4 months in Zhang etal. f2018bl. but because
this definition is closely tied to menstruation, there are concerns for reverse causality as with the
previous outcomes, which would be expected to be biased away from the null. In Zhang et al.
(2018b). there were higher odds of POI with higher exposure, with an exposure-response gradient
across tertiles (OR (95% CI) versus tertile 1: T2: 2.04 (1.03, 4.04), T3: 6.63 (3.22, 13.65)). In
Crawford et al. (2017). there was an inverse association between AMH and PFHxS, consistent with
decreased ovarian reserve, although this was not statistically significant ((3: -0.12, p = 0.4). No
association was observed with AMH in Donley et al. f20191. despite similar exposure contrast
(median 1.6 ng/mL) in the two AMH studies and lower exposure levels in Zhang etal. f2018bl. The
results of Zhang etal. f2018bl and Crawford et al. f20171 are coherent with each other as well as
3-322
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
with the positive association with FSH observed in women with POI in Zhang etal. (2018b).
although no association was observed in control women without POI (discussed with reproductive
hormones). Overall, due to the study limitations and small number of studies, there is still
considerable uncertainty.
Pubertal development
Three medium confidence studies, including birth cohorts in Denmark (Ernst etal.. 2019)
and the U.S. (Carwile etal.. 2021) and a case-control study nested in a birth cohort in the United
Kingdom fChristensen etal.. 20111. and low confidence cross-sectional study in the U.S. fWise etal..
2022) examined timing of pubertal development with prenatal PFHxS exposure. There is potential
for reverse causation in this set of studies as early menarche may lower serum PFHxS
concentrations. Ernst etal. f20191 and Carwile etal. f20211 reported results for several pubertal
outcome measures, while Christensen et al. (2011) and Wise etal. (2022) focused on age at
menarche. In Ernst etal. (2019). with median exposure of 1.1 ng/mL (10th-90th percentile: 0.6-
1.7), the participants in the third tertile of exposure had earlier age of breast development, axillary
hair, and menarche, although none were statistically significant Looking at a combined puberty
indicator outcome, there was lower age at puberty in the third tertile (age difference -2.22 months;
95% CI: -8.37, 3.93). Carwile etal. f20211. with median exposure of 1.9 ng/mL, reported no
association with pubertal development score or peak height velocity (i.e., the age at which a child
experiences the largest increase in height, a proxy for pubertal timing). In Christensen etal. (2011).
with median exposure of 1.5 ng/mL (IQR 0.5-0.8), there was not a clear association, as there were
higher odds of earlier age at menarche when PFHxS was analyzed as dichotomous based on
above/below the median (OR 1.11; 95% CI: 0.76,1.64) but lower odds when analyzed as
continuous (OR 0.89; 95% CI: 0.65,1.22), neither statistically significant. Lastly, the low confidence
study found no association with age at menarch fWise etal.. 20221. Overall, there is considerable
uncertainty for this outcome given the inconsistency in three medium confidence studies and
imprecision of the effect estimates.
Menopause
One medium confidence study, a cohort of midlife women in the U.S., examined timing of
menopause fDingetal.. 20221. The effect estimate is in the direction of earlier onset of natural
menopause, though not statistically significant, (relative survival: 0.90, 95% CI: 0.76,1.05 for total
effect (including author-proposed mediation by FSH)). As with other outcomes related to
menstruation, there is potential for reverse causation in this association, which would be consistent
with lower PFHxS concentrations in women with earlier onset of menopause, as described in
multiple sources (Ruark etal.. 2017: Tain and Ducatman. 2022).
3-323
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Breastfeeding duration
Four medium confidence birth cohorts examined duration of breastfeeding in relation to
exposure to PFHxS measured during gestation. Six additional studies without prospective
measurement of exposure that reported analyses predicting PFNA concentrations based on past
breastfeeding duration were considered supplemental evidence because of the high probability of
reverse causation due to lactation being an elimination route (Pirard etal.. 2020: Papadopoulou et
al.. 2016: Lee etal.. 2018: Kim etal.. 2020b: Harris etal.. 2017: Ammitzb0ll etal.. 20191. The results
of the included studies are summarized in Table 3-39. One study reported that participants with
higher exposure were more likely to have cessation of any (but not exclusive) breastfeeding by 3
and 6 months, but this association was not statistically significant fRomano etal.. 20161. The other
three studies found no clear association.
Table 3-39. Associations between PFHxS and breastfeeding duration in
epidemiology studies
Reference, confidence
Population
Median
exposu re
(IQR)
Form and
units of
effect
estimate
Endpoint
Effect estimate
Risk of cessation of breastfeeding (>1 indicates earlier cessation)
Timmermann et al.
(2022)
Odense Child
Cohort (2010-
2012), Denmark,
932 women
0.4 (0.2-
0.5)
HR (95% CI)
for doubling
Cessation of any
breastfeeding
1.07 (0.98, 1.16)
Cessation of
exclusive
breastfeeding
0.93 (0.88, 1.00)
Rosen et al. (2018)
Norwegian Mother
and Child Study
(1999-2008),
Norway, 1,716
women
0.6 (0.5-
0.9)
HR (95% CI)
for IQR
change
Cessation of any
breastfeeding by
3 mo
0.88 (0.75, 1.03)
Cessation of any
breastfeeding by
6 mo
0.92 (0.82, 1.03)
Romano et al. (2016)
HOME cohort
(2003-2006), U.S.,
336 women
1.5 (0.9-
2.3)
RR (95% CI)
vs. Q1
Cessation of any
breastfeeding by
3 mo
Q2: 1.28 (0.96,
1.70)
Q3: 1.24 (0.87,
1.75)
Q4: 1.39 (0.99,
1.96)
Cessation of any
breastfeeding by
6 mo
Q2: 1.03 (0.83,
1.28)
Q3: 1.08 (0.85,
1.39)
Q4: 1.22 (0.96,
1.55)
3-324
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference, confidence
Population
Median
exposu re
(IQR)
Form and
units of
effect
estimate
Endpoint
Effect estimate
Cessation of
exclusive
breastfeeding by
3 mo
Q2: 0.95 (0.86,
1.06)
Q3: 0.89 (0.78,
1.01)
Q4: 0.94 (0.84,
1.06)
Continuous duration of breastfeeding (<0 indicates earlier cessation)
Timmermann et al.
(2017b)
Two birth cohorts in
Faroe Islands
(1997-2009),
Denmark, 1,092
women
0.5 (0.2-
5.2)
Difference
in mo (95%
CI) for
doubling
Duration of any
breastfeeding
-0.2 (-0.5, 0.2)
Duration of
exclusive
breastfeeding
-0.1 (-0.2, 0.1)
*p < 0.05.
Animal Studies
The database of animal toxicity studies for PFHxS-induced female reproductive effects
consists of five oral exposure studies that include two short-term studies in Harlan Sprague Dawley
or Crl:CD BR rats exposed for 28 days fNTP. 2018a: 3M. 2000b], two reproductive/developmental
toxicity studies in Crl:CD (SD) rats or Crl:CDl (ICR) mice with exposures starting during premating
through postnatal days (PND) 22-35 (Chang etal.. 2018: Butenhoff etal.. 2009: 3M. 2003) and a
developmental toxicity study in Wistar rats with exposure during gestion and lactation (gestational
days [GD] 7 to PND 22) fRamhai etal.. 20181. The studies evaluated several endpoints relevant to
the assessment of female reproductive toxicity, namely mating and fertility, estrous cycle, hormone
levels, histopathology, organ weight and markers of sexual differentiation and maturation fU.S.
EPA. 19961. Other developmental outcomes reported in the Ramhai etal. f20181 study are
described in the synthesis of developmental effects (see Section 3.2.3).
Mating and fertility
Mating and fertility measures (i.e., fertility index, mating index and precoital interval) were
evaluated across two high confidence studies with no outstanding issues regarding risk of bias or
sensitivity (see Figure 3-84). The studies exposed F0 female SD rats or CD-I mice to doses ranging
from 3 to 10 mg/kg-day during premating gestation, and lactation (PND 22) f Chang etal.. 2018:
Butenhoff et al.. 2009: 3M. 2003). No treatment-related effects were noted in mating and fertility
indices, including length of precoital interval in female parental animals.
3-325
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reporting quality •
Allocation •
Observational bias/blinding ¦
Confounding/variable control ¦
Selective reporting and attrition •
Chemical administration and characterization •
Exposure timing, frequency and duration
Results presentation ¦
Endpoint sensitivity and specificity-
Overall confidence
++ ++
+
+
NR
NR
++
++
++
++
++
++
++
++
++
++
++
++
++
++
s
Legend
| Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Not reported
Figure 3-84. PFHxS mating and fertility animal study evaluation heatmap. For
additional details see HAWC link.
Estrous cycle characteristics
Effects on the estrous cycle were measured in four studies: a short-term study in rats
exposed for 28 days (NTP. 2018a) and two reproductive-developmental toxicity studies in F0 rats
or mice exposed during premating gestation, and lactation (PND 22) (Chang etal.. 2018: Butenhoff
etal.. 2009: 3M. 20031. and one subchronic study that exposed ICR mice for 42 days (Yin etal..
20211 (see Figure 3-85). Two of the studies were considered high confidence (NTP. 2018a:
Butenhoff et al.. 2009: 3M. 20031 and two were considered medium confidence because of
uncertainties surrounding presentation of results and selection of animals for outcome assessment
fYin etal.. 2021: Chang etal.. 20181 (see Figure 3-85). Yin etal. f20211 reported decreased
increased estrous cycle duration in treated animals, but the remaining studies which evaluated this
outcome report that PFHxS exposure had no effects in the number of cycles, cycle length, or time in
each estrous stage (proestrus, estrus, metestrus, and diestrus) of female rats or mice exposed to
doses of 0.3-50 mg/kg-day and 0.3-3 mg/kg-day, respectively (NTP. 2018a: Chang etal.. 2018:
Butenhoff et al.. 2009: 3M. 20031.
3-326
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
End point sensitivity and specificity
Overall confidence
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)
NRl Not reported
Figure 3-85. PFHxS estrous cycle animal study evaluation heatmap. For
additional details see HAWC link.
Hormone levels
The available studies have measured reproductive hormones including testosterone, follicle
stimulating hormone (FSH), Luteinizing hormone (LH), and estrogen. Serum testosterone levels
were measured in female rats in a single short-term high confidence study with no notable
concerns in any of the study evaluation domains (NTP. 2018a) (see Figure 3-86). Female rats were
exposed to 0, 3.12, 6.25,12.5, 25, and 50 mg/kg-day PFHxS for 28 days. Serum testosterone levels
were slightly increased in PFHxS-exposed rats at all doses (9%-29% compared with controls) but
the changes were not statistically significant compared with controls and did not display a dose-
response gradient. A medium confidence study using ICR mice reported that exposure to 5 mg/kg-
day PFHxS decreased serum FSH, LH, and estrogen (Yin etal.. 2021). These observations suggest
that PFHxS exposure may alter reproductive hormones in exposed female animals, however several
issues were identified with the Yin etal. (20211 study including lack of randomization and selective
reporting. Therefore, additional studies are needed.
3-327
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Chemical administration and characterization
Exposure timing, frequency and duration
Endpoint sensitivity and specificity
Selective reporting and attrition
Confounding/variable control
Observational bias/blinding
Results presentation
Overall confidence
Reporting quality
Allocation
+ Adequate (metric) or Medium confidence (overall)
- Deficient (metric) or Low confidence (overall)
9 Critically deficient (metric) or Uninformative (overall)
JR Not reported
Legend
Good (metric) or High confidence (overall)
Figure 3-86. PFHxS hormone levels animal study evaluation heatmap. For
additional details see HAWC link.
Histopathologv
Histopathology of female reproductive organs including the ovary, uterus, vagina, and
clitoral and mammary glands were examined across four studies. Two short-term studies in rats
exposed for 28 days (NTP. 2018a: 3M. 2000b) and two reproductive-developmental toxicity studies
in F0 rats or mice exposed from 14 days of premating to PND 22 (Chang etal.. 2018: Butenhoff et
al.. 2009: 3M. 20031. Three of the studies were considered high confidence (NTP. 2018a: Butenhoff
etal.. 2009: 3M. 2000b. 20031 and one was rated as medium confidence due to deficiencies in the
presentation of histopathological findings (data were only reported qualitatively) fChang et al..
20181 (see Figure 3-87).
Bilateral dilation of the uterus (minimal to mild severity) was reported in rats in the control
(1/10 rats) and PFHxS exposure groups (1/1,1/1, 3/3, and 1/10 rats at3.12, 6.25,12.5, and
50 mg/kg-day, respectively) in the NTP (2018a) study. Although lesions were observed in 100% of
the animals evaluated in the 12.5 mg/kg-day dose group, the incidence rates were identical for the
control and high dose groups (10%) and a limited number of animals were examined in the other
exposure groups; therefore, the biological interpretation of these findings is unclear. Butenhoff et
al. f20091 and 3M f20031 also observed uterine lesions in rats (mild-moderate distention and
microphage infiltration of mostly moderate severity) but the incidence rates were not significantly
different between control and PFHxS exposure (10 mg/kg-day). Two medium confidence mouse
studies report conflicting evidence. Chang etal. (2018) reported no lesions in the uterus of CD-I
3-328
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
mice exposed to 10 mg/kg-day PFHxS for 42 days (Chang etal.. 20181. However, a similar study
also using CD-I mice exposed to 5 mg/kg-day PFHxS for 42 days reported decreased number of
secondary follicles and corpora lutea, but no effect on primordial or primary follicles fYin etal..
20211. A single case of minimal focal necrosis was reported in the mammary gland30 of rats (1/10)
at a dose of 10 mg/kg-day fButenhoff etal.. 2009: 3M. 20031 but no lesions were observed in the
mammary gland of rats exposed to doses ranging from 3.12-50 mg/kg-day in a different study
(NTP. 2018a). Histological examination of the ovaries (including primordial follicle counts), clitoral
gland and vagina showed no treatment-related effects in exposed rats or mice (NTP. 2018a: Chang
etal.. 2018: Butenhoffetal.. 2009: 3M. 2000b. 20031.
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
I 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
Figure 3-87. PFHxS female reproductive histopathology animal study
evaluation heatmap. For additional details see HAWC link.
Organ weight
There are six available animal toxicity studies that evaluated effects on reproductive organ
weights in females (i.e., ovary and uterus). One study exposed CD-I mice for 42 days fYin etal..
2021). two studies exposed SD rats for 28 days (NTP. 2018a: 3M. 2000b) and three reproductive-
developmental toxicity studies examining effects in F0 rats and mice exposed during premating
and/or gestation and lactation (PND 22) (Ramh0i etal.. 2018: Chang etal.. 2018: Butenhoff etal..
2009: 3M. 2003) and in F1 mice exposed in utero, via lactation and directly from PND 22 to PND 35
(Chang etal.. 2018). Overall study confidence was medium in the Chang etal. (2018) study due to
30Mammary gland necrosis, or breast gland necrosis is considered a benign response that is not associated
with tumor development and it is rarely observed (Genova R. 2024: B and ES. 20181.
3-329
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
incomplete reporting of organ weight data (quantitative results were not provided) (see Figure 3-
88). The study by Yinetal. f20211 was also considered medium confidence due to concerns related
to animal selection for outcome assessment There were no major concerns with respect to risk of
bias or sensitivity in the other studies deemed as high confidence fRamhai etal.. 2018: NTP. 2018a:
Butenhoff etal.. 2009: 3M. 2000b. 20031. Yin etal. f20211 reported decreased absolute (butnot
relative) ovary weight in animals exposed to 50 mg/kg-day for 42 days. However, in all other
available studies PFHxS exposure did not significantly impact ovarian and uterine weights (both
absolute and relative) in animals at doses ranging from 0.05-50 mg/kg-day in any of the studies
fRamhai etal.. 2018: NTP. 2018a: Chang etal.. 2018: Butenhoff et al.. 2009: 3M. 2000b. 20031.
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
Legend
I Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
NRl Not reported
Figure 3-88. PFHxS female reproductive organ weight animal study evaluation
heatmap. For additional details see HAWC link.
Landmarks of female reproductive system development and maturation
Markers of sexual differentiation and maturation, namely anogenital distance (AGD)31 and
onset of puberty (vaginal patency), were evaluated in F1 offspring in two reproductive-
developmental toxicity studies of medium confidence in rats exposed during gestion to PND 22
fRamhai et al.. 20181 or in mice exposed in utero, via lactation and directly from PND 22 to PND 35
f Chang etal.. 20181. Key issues related to animal allocation and presentation of results for AGD (no
adjustment for body weight32) reduced confidence in one study fRamhai etal.. 20181 (see Figure 3-
31AGD is a phenotypical marker of androgen levels during gestational development (Thankamonv et al..
20161. Increased AGD in considered indicative of an adverse response in the developing female reproductive
system (U.S. EPA. 19961
32Relative AGD adjusted to the cube root of body weight is the preferred measurement for this endpoint
(Daston and Kimmel. 1998L
3-330
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
89). Ambiguity surrounding the reporting of sample size raised potential concerns in the second
study fChangetal.. 20181.
Statistically significant reductions in relative AGD (adjusted to body weight) evaluated on
PND 1 were noted in F1 mice exposed to 1 mg/kg-day (5% compared with controls) but the effects
were not seen at other dose levels (0.3 and 3 mg/kg-day) f Chang etal.. 20181. Furthermore,
absolute AGD was unaffected by treatment in F1 mice or rats up to doses of 45 mg/kg-day (Ramhaj
etal.. 2018: Chang etal.. 20181. Similarly PFHxS had no effect on the onset of puberty (vaginal
patency) in F1 mice exposed to doses of 0.3-3 mg/kg-day.
Allocation ¦
Observational bias/blinding-
Results presentation ¦
Endpoint sensitivity and specificity
Overall confidence
++
++
+
+
NR
NR
++
++
++
++
++
++
++
++
0
++
++
+
~
+
s
Legend
I Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Not reported
Figure 3-89. PFHxS female reproductive sexual differentiation and maturation
animal study evaluation heatmap. For additional details see HAWC link.
Evidence Integration
The available studies provide inadequate evidence to determine whether PFHxS exposure
has the potential to affect female reproduction in humans. This conclusion is based on studies in
both humans and animals (see Table 3-40).
The available evidence on PFHxS-induced female reproductive effects in human studies is
considered indeterminate. Outcomes evaluated in human studies include fecundity, reproductive
hormones, pubertal development, menstrual cycle characteristics, gynecological conditions, and
ovarian reserve. Associations were observed with many of these outcomes in some studies, but
there was considerable inconsistency across studies within outcomes and uncertainty due to
considerations such as reverse causality and confounding (e.g., parity for fecundity) that reduced
3-331
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
study confidence. Looking across outcomes, there is some coherence. The observed increase in
estradiol and FSH and decrease in testosterone in some studies (one study for FSH) is coherent
with risk factors for endometriosis, which in turn is coherent with reduced ovarian reserve and
fecundity. Similarly, the decrease in anogenital distance in one study of newborn girls (see
Developmental Effects Section) is coherent with the decrease is testosterone levels in some of the
studies, including the single study in infants. These connections between the outcomes increase the
strength of the evidence, but because of the limitations described above, there is too much
uncertainty in the association to draw a stronger judgment than indeterminate.
The available animal evidence on PFHxS-induced female reproductive effects is also
considered indeterminate. One medium confidence, study using mice reported PFHxS-induced
alterations in estrus cycle, histopathology, ovary weight, and reproductive hormone levels (Yinet
al.. 20211. In all other medium and high confidence studies there were no clear exposure-related
effects were observed in reproductive organ weights, estrous cycle characteristics, histopathology,
reproductive hormones levels, and functional measures of mating and fertility. In addition to the
inconsistencies between the Yin etal. (20211 and the other available studies there are no
subchronic or chronic exposure studies available, which also limits the interpretation of the current
findings.
3-332
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-40. Evidence profile table for PFHxS exposure and female reproductive effects
Evidence stream summary and interpretation
Evidence
integration
summary
judgment
Evidence from studies of exposed humans
OOO
Evidence
inadequate
Primary Basis:
Evidence is
inconsistent across
studies or largely
null.
Human relevance:
Without evidence
to the contrary,
effects in rodent
models are
considered
relevant to
humans.
Cross-stream
coherence: N/A,
evidence
indeterminate for
Studies and confidence
Factors that increase
certainty
Factors that decrease certainty
Summary and key findings
Evidence stream
judgment
Fecundity
3 medium and 5 low
confidence studies
• No factors noted
• Unexplained inconsistency
• High risk of bias
Decreased fecundity/longer
time to pregnancy in 2 low
confidence studies, but no
effect in medium confidence
studies.
ooo
Indeterminate
Associations
between
exposure and
female
reproductive
outcomes
observed in
studies of
multiple
outcomes.
Inconsistency
across studies and
concerns for
reverse causality
and other bias
hinder
interpretation.
Reproductive hormones
7 medium and 7 low
confidence studies
• No factors noted
• Unexplained inconsistency
• High risk of bias - Most
testosterone studies were
low confidence
3 of 9 studies report higher
estradiol; 3 of 9 studies report
lower testosterone.
Pubertal development
3 medium and 2 low
confidence studies
• No factors noted
• Unexplained inconsistency
Earlier age of puberty (not
statistically significant) in one
study, but no clear association
in other studies
Menstrual cycle
3 low confidence studies
• Consistency
• Low confidence studies-
potential reverse causality
Higher odds of irregular and
long cycle in 2 studies, lower
odds of menorrhagia in 1
study, and less intense
bleeding in one study
3-333
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence
integration
summary
judgment
Gynecological conditions
5 low confidence studies
• No factors noted
• Unexplained inconsistency
• All low confidence studies-
potential reverse causality
• Imprecision of effect
estimate
Higher odds of endometriosis
in 2 of 4 studies. Lower odds
of endometriosis-related
infertility in one study. 1 of 2
studies reported higher
likelihood of PCOS, but there
is potential for confounding
by PFOS.
both human and
animal studies.
Susceptible
populations and
life stages'.
None identified.
Ovarian reserve
2 medium and 1 low
confidence studies
• Coherence in
associations
between POI and
AMH in one study
• Potential for reverse
causality
• Unexplained inconsistency
across studies of AMH
Higher odds of premature
ovarian insufficiency (POI) and
lower levels of anti-Mullerian
hormones (AMH) (in 1/2
studies)
Evidence from in vivo animal studies
Studies and confidence
Factors that increase
certainty
Factors that decrease certainty
Summary and key findings
Evidence stream
judgment
Mating and fertility
2 high confidence studies
in adult rats and mice:
• 14-d (x2)
• No factors noted
• No factors noted
No observed effects on
mating or fertility index
ooo
Indeterminate
[Note: although
no notable
findings, no long-
term studies were
available]
Estrous cycle
2 high confidence studies
in adult rats:
• 28-d
• 14-d premating to
PND22
• Unexplained inconsistency
across studies
Altered cycle duration
reported in one medium
confidence study
3-334
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence
integration
summary
judgment
2 medium confidence
study in adult mice:
• 14-d premating to
PND22
• 42-d
Hormone levels
1 high confidence study
in adult rats
• 28-d
• 1 medium confidence
study in adult mice.
• 42-d
• Lack of expected dose
response
Slight increase in
testosterone, decreased
estrogen, LH, and FSH
Histopathology
3 high confidence studies
in adult rats
• 28-d (x2)
• 14-d premating to
PND22
1 medium confidence
study
• 14-d premating to
PND22
1 low confidence study
• 42-d
• Unexplained inconsistency
across studies
Decreased number of
secondary follicles and
corpora lutea in 1 low
confidence study
3-335
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence
integration
summary
judgment
Organ weights
1 high confidence study
in adult rats
• 28-d (x2)
• 14-d premating to
PND22
3 medium confidence
studies in rats and mice
• GD 7-PND 22
• 14-d premating to
PND22
• 42-d
• Unexplained inconsistency
across studies
Decreased ovary weight
reported in 1 medium
confidence study
Developmental effects
2 medium confidence
studies in rats and mice
• GD 7-PND 22
• GD O-PND 22
No observed effects on
female reproductive organ
development
3-336
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3.2.9. Male Reproductive Effects
Human Studies
Twelve epidemiology studies (reported in 15 publications) examined the association
between PFHxS exposure and male reproductive effects. The outcomes included in these studies
were semen parameters, reproductive hormones, timing of pubertal development, and anogenital
distance. These studies are described below. Several studies also evaluated effects on anogenital
distance (AGD), a developmental outcome that is responsive to variations in reproductive
hormones ((Foster and Gray. 2013: Dean and Sharpe. 20131: see Section 3.2.3).
Semen and sperm parameters
Semen concentration and sperm motility and morphology were considered the core
endpoints for the assessment of semen parameters in the available studies. Alterations of these
endpoints is considered indicative of male reproductive toxicity (U.S. EPA. 1996: Faqi etal.. 2017).
Other outcomes, such as specific sperm morphology and motility defects, were not consistently
reported across studies and were considered secondary; these outcomes are most useful to probe
into associations observed in the core endpoints. Key considerations for the assessment of semen
parameters involve sample collection and sample analysis. Samples should be collected after an
abstinence period of 2-7 days, and analysis should take place within 2 hours of collection and
follow guidelines established by the World Health Organization (WHO. 2010). While exposure
would ideally be measured during the period of spermatogenesis rather than concurrent with the
outcome, a cross-sectional design is considered adequate because the period of spermatogenesis is
fairly short (<3 months) relative to the half-life of PFHxS (years), and there is no concern for
reverse causality with this outcome.
Five epidemiology studies (reported in seven publications) examined the association
between PFHxS exposure and semen quality. The evaluations for these studies are summarized in
Figure 3-90, and additional details can be obtained from HAWC. Three studies were medium
confidence: one was a cross-sectional analysis of male partners in a pregnancy cohort (Toft etal..
2012) and two were cross-sectional studies of healthy young men (Petersen etal.. 2022: Toensen et
al.. 20131. The remaining two studies were low confidence due to multiple identified deficiencies
and were cross-sectional studies of men seeking infertility assessment fSong etal.. 2018: Huang et
al.. 2019bl. All the studies analyzed PFHxS in serum using appropriate methods and thus exposure
misclassification is expected to be minimal.
3-337
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Huang, 2019, 5406374
Joensen, 2013, 2919160-t
Petersen, 2022, 10273364H
Song, 2018, 4220306
o^ o°^' i^s s®1
3^
Ji\y J^je
Oxl'
1
1
+
i
i
+
i
+
i
i
+
¦
~
+
a
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)
j ++ ++
++
+
+
+
+
+
B
++
+
++
-
+
+
- -
++
++
-
-
-
-
+
-
+
+
+
++
+
+
+
Figure 3-90. Semen parameters epidemiology study evaluation heatmap. For
additional details see HAWC link.
The results for the association between PFHxS exposure and semen quality in medium
confidence studies are presented in Table 3-41. The studies analyzed the outcomes differently so
the effect estimates are not directly comparable, but a negative effect estimate indicates a reduction
in sperm quality with higher exposure. There was a statistically significant and dose-dependent
decrease in normal sperm morphology in one medium confidence study fToftetal.. 20121 and an
imprecise and non-dose-dependent decrease (>10% change) in concentration in the same study
(Toftetal.. 2012). A low confidence study (Huang et al.. 2019b) reported a higher concentration
(p < 0.05) and motility (p > 0.05) with PFHxS exposure. No association was reported in the other
medium (Petersen etal.. 2022: Toensen etal.. 2013) or low (Song etal.. 2018) studies. Other
publications of the same study described in Toftetal. (2012) reported no clear association between
PFHxS exposure and sperm DNA damage fSpecht etal.. 2012: Leter etal.. 20141. indicating that
PFHxS-induced DNA damage is unlikely to explain the decreases in the percent of sperm with
normal morphology (and the slightly decreased sperm numbers) observed in Toftetal. f201211.
Exposure levels were slightly higher in Toftetal. (2012) than Toensen etal. (2013). which could
explain the differing results, but this cannot be confirmed with the currently available evidence.
3-338
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-41. Associations between PFHxS and semen sperm parameters in
medium confidence epidemiology studies
Reference
Population
Median
exposu re
(IQR)
(ng/mL)
Effect
estimate
Concentration
Motility3
(% progressively
motile)
Morphology3
(% normal)
Petersen et
al. (2022)
Cross-sectional study
of young men (2017-
2019), Denmark; 1,041
men (18-20 yr)
0.3 (P5-
P95: 0.2-
0.6)
% Change
vs. T1
T2: 0 (-12, 13)
T3: 2 (-10, 16)
T2: -7 (-12, -1)
T3: -2 (-8, 4)
T2: 1 (-10, 12)
T3: 6 (-5,18)
Joensen et
al. (2013)
Cross-sectional study
of men evaluated for
military service (2008-
2009), Denmark; 247
men (18-22 yr)
0.7 (0.5-
0.9)
P (95% CI)
for 1-unit
increase
Cubic root
transformed
0.05 (-0.12,
0.22)
% Immotile
Square
transformed
-2.82 (-232,
227)
Square root
transformed
0.12 (-0.02,
0.26)
Toft et al.
(2012)
INUENDO cohort
cross-sectional
analysis (2002-2004),
Greenland, Ukraine,
Poland; 588 men
1.1
(P33-P66:
0.7-1.5)
% Change
vs. T1
(mill/ mL)
T2: -12 (-52,
28)
T3: -11 (-57,
35)
T2: 11 (-12, 35)
T3: 10 (-18, 37)
T2: -27 (-58,
3)
T3: -35 (-70,
-1)*
*p < 0.05.
CD = critically deficient; T = tertile.
aPercent motile in population was 37% in Petersen et al. (2022), 58% in Joensen et al. (2013), and 56%-64% in Toft
et al. (2012), varying by country. Percent normal morphology in population was 6% in Petersen et al. (2022), 7% in
Joensen et al. (2013) and 6%-7% in Toft et al. (2012).
Reproductive hormones in males
Testosterone and estradiol were considered the primary endpoints for male reproductive
hormones, although findings for LH, FSH, and SHBG were also reviewed where available. Key issues
for the evaluation of these studies were sample collection and processing. For testosterone, LH, and
FSH, blood sample collection should be performed in the morning due to diurnal variation, and if
not possible, time of collection must be accounted for in the analysis. If there is no consideration of
time of collection, the study is classified as deficient for outcome ascertainment and low confidence
overall for these hormones.
Nine studies (reported in 10 publications) examined the associations between PFHxS and
male reproductive hormones. Most studies examined only testosterone and estradiol. All the
studies measured exposure and outcome concurrently which was considered appropriate since
levels of these hormones are capable of being rapidly upregulated or downregulated and they are
not expected to directly bind to or otherwise interact with circulating PFAS. Four studies (Specht et
al.. 2012: Petersen etal.. 2022: Lewis etal.. 2015: Toensen etal.. 20131 examined associations in
adults, two studies in adolescents fZhou etal.. 2016: Lewis etal.. 20151. one study in children
fLopez-Espinosa etal.. 20161. and three studies in infants fYao etal.. 2019: Liu etal.. 2020b: Jensen
etal.. 2020b). The study evaluations are summarized in Figure 3-91. Four studies were rated
3-339
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
medium in overall study confidence (Petersen etal.. 2022: Lopez-Espinosa et al.. 2016: Liu etal..
2020b: loensen etal.. 20131. and five were low confidence fZhou etal.. 2016: Yao etal.. 2019:
Spechtetal.. 2012: Lewis etal.. 2015: Tensen et al.. 2020bl. However, of the medium confidence
studies, one did not consider timing of sample collection and was thus low confidence for
testosterone fLopez-Espinosa et al.. 20161.
Jensen, 2020, 6311643-
I
+
++
I
I
+
++
I
I
+
Joensen, 2013, 2919160-
++
++
++
+
+
+
+
+
Lewis, 2015, 3749030-
++
+
-
+
+
+
+
-
Liu, 2020, 6569227-
+
+
+
+
++
-
+
+
Lopez-Espinosa, 2016, 3859832-
++ ++
+*
+
++
+
+
+#
Petersen KU et al. 2022 -
+
+
+
+
++
-
+
+
Specht, 2012, 1289939-
+
++
+*
-
+
-
-
¦
Yao, 2019, 5187556-
+
++
+*
-
++
-
+
-
Zhou, 2016, 3856472-
+
+
+
-
+
+
+
-
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 3-91. Summary of study evaluation for epidemiology studies of male
reproductive hormones. For additional details see HAWC link.
Testosterone
As described above, most studies were low confidence for testosterone. In adult men, four
studies were available and two were low confidence. In the two medium confidence studies, both
populations of young men in Sweden (loensen etal.. 2013) and Denmark (Petersen etal.. 2022). no
association was reported between PFHxS exposure and testosterone levels, at mean concentrations
of 0.7 and 0.3, respectively. Nonstatistically significant inverse associations were observed in one
low confidence study of adults (Lewis etal.. 2015). and only in age groups 20 to <40 and 40 to 60 ((3
(95% CI); for 20 to 40: -1.2 (-4.7, 2.4), for 40 to 60: -3.6 (-8.2,1.2), and 60 to 80: 3.3 (-3.8, 10.8).
The other low confidence study did not report quantitative results but stated that associations were
not consistent across countries in the study fSpecht etal.. 20121. For adolescents, one low
confidence study (Lewis etal.. 2015) reported a nonstatistically significant positive association ((3
2.4, 95% CI: -9.1,15.2), and the other reported no association (Zhou etal.. 2016). A study in
children (Lopez-Espinosa etal.. 2016) reported a nonstatistically significant inverse association ((3
-2.7, 95% CI: -6.4,1.2), while two studies in infants (Yao etal.. 2019: Tensen et al.. 2020b) reported
3-340
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
no association. Overall, there is inconsistent evidence of an association between PFHxS exposure
and testosterone. Some low confidence studies report inverse associations, but the medium
confidence studies reported no association. It is possible that this is due to differences in PFHxS
levels, as the medium confidence studies had exposure levels lower than the studies that observed
an association (median blood concentrations 0.3-0.7 ng/mL versus 1.3-1.8 ng/mL in Lewis et al.
(2015) and 8 ng/mL in Lopez-Espinosa etal. (2016)). but given the concerns for outcome
misclassification in the low confidence studies, the results are difficult to interpret
Estradiol
Six studies examined associations between PFHxS exposure and estradiol in male subjects.
Among the three medium confidence studies fPetersen etal.. 2022: Lopez-Espinosa et al.. 2016:
Toensen etal.. 20131 reported no association between increasing PFHxS exposure and estradiol.
Results across the low confidence studies are mixed, as Zhou etal. (2016) reported higher estradiol
levels with higher PFHxS exposure, while Spechtetal. (2012) reported that estradiol levels were
not consistently associated with PFHxS across countries with no data shown and Yao etal. (2019)
reported no association.
Other reproductive hormones
For other reproductive hormones, SHBG was not associated with PFHxS levels in Specht et
al. (2012). Toensen etal. (2013). or Petersen etal. (2022). FSH and LH were not associated with
PFHxS in Toensen etal. (2013) or Petersen etal. (2022) and associations were not consistent across
regions in Spechtetal. (2012). In Tensen etal. (2020b). positive butnonstatistically significant
associations were reported with LH, dehydroepiandrosterone (DHEA), dehydroepiandrosterone-
sulfate (DHEAS), androstenedione, and 17-hydroxyprogesterone (17-OHP). Liu etal. f2020bl
reported a small but not statistically significant positive association (2.7% increase) with
progesterone in infants.
Overall, there is little evidence of an association between PFHxS exposure and male
reproductive hormones, but there are limitations in the available evidence that hinder
interpretation of the null findings.
Pubertal development
Two medium confidence studies, birth cohorts in Denmark (Ernst etal.. 20191 and the U.S.
fCarwile etal.. 20211. examined timing of pubertal development with PFHxS exposure. Ernst et al.
(2019) used maternal exposure (median 1.1 ng/mL, 10th-90th percentile: 0.6-1.7) while Carwile et
al. (2021) used childhood exposure at around 8 years of age. One study reported that the
participants in the third tertile of exposure had earlier genital development, pubic hair, axillary
hair, acne, voice break, and first ejaculation, with axillary hair acne, and voice break being
statistically significant. Looking at a combined puberty indicator outcome, there was lower age of
puberty in the third tertile (age difference -6.89 (95% CI: -12.57, -1.20)) fErnstetal.. 20191. The
3-341
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
second study reported no association between PFHxS exposure and a pubertal development score
or age at peak height velocity fCarwile etal.. 20211.
Summary of human studies on male reproductive effects
Overall, there is some limited evidence of an association between PFHxS exposure and
sperm motility, timing of pubertal development, and anogenital distance, but there is considerable
uncertainty in the available data due to lack of consistency across the studies on each outcome and
lack of coherence with reproductive hormones.
Animal Studies
The database of animal toxicity studies on PFHxS-induced male reproductive effects
consists of five oral exposure studies that include two short-term studies in Harlan Sprague Dawley
rats exposed for 28 days (NTP. 2018c: 3M. 2000b). two multigeneration reproduction studies in
Crl:CD (SD) rats or Crl:CDl (ICR) mice with exposures starting during 2-week premating through
PND 22-35 (Chang etal.. 2018: Butenhoff et al.. 2009) and a single-generation reproduction study
in Wistar rats with exposure during gestation and lactation (gestational days [GD] 7 to PND 22)
fRamhai etal.. 20181. The studies evaluated several endpoints relevant to the assessment of male
reproductive toxicity, namely mating and fertility, sperm measures, hormone levels, histopathology,
organ weights, and morphological markers of sexual differentiation and maturation fU.S. EPA.
19961.
Sperm parameters
Sperm measures (count, motility, morphology, concentration, and production rate) were
evaluated in three low confidence studies that exposed animals for 28 or 44 days (see Figure 3-92).
In SD rats, exposure to PFHxS for 28 days did not impact sperm count, spermatid count, or sperm
motility. Additionally, Butenhoff et al. f20091. 3M f20031 and Chang etal. f 20181 did not observe
PFHxS-induced alterations in sperm motility, morphology, or concentration after exposing SD rats
or CD-I mice for 44 and 42 days, respectively. Overall, these results suggest that PFHxS exposure
does not affect sperm measures. However, these findings should be interpreted with caution as the
available studies were of low confidence due to experimental design features that may have
resulted in reduced sensitivity and a potential bias toward the null.33
33In rodent models such as the rat it takes approximately eight weeks for spermatogonia to develop to
spermatozoa (Foster and Gray. 20131. Damage to the spermatogonial cells would not be detected in ejaculate
or cauda epididymis samples from animals exposed for periods that are shorter than eight weeks (U.S. EPA.
19961.
3-342
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Not reported
Mil
Figure 3-92. Male reproductive animal study evaluation heatmap - sperm
measures. For additional details see HAWC link.
Histopathology
Histopathology of male reproductive organs was evaluated in two high confidence studies
and one medium confidence study (see Figure 3-93). In SD rats, exposure to PFHxS for 28 to 44 days
at doses ranging from 0.3 to 10 mg/kg-day did not affect the histopathology of the testes, preputial
glands, epididymis, or seminal vesicles (NTP. 2018c: Butenhoffetal.. 2009: 3M. 2000b. 20031.
3-343
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
^e
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization -
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Not reported
Figure 3-93. Male reproductive histopathology animal study evaluation
heatmap. For additional details see HAWC link.
Hormone levels
The effects of PFHxS exposure on reproductive hormones was evaluated in one high
confidence study using SD rats (see Figure 3-94). Exposure to PFHxS for 28 days at doses ranging
from 0.625 to 10 mg/kg-day did not affect serum testosterone levels (NTP. 2018c).
3-344
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Chemical administration and characterization A
Exposure timing, frequency and duration A
Endpoint sensitivity and specificity
Selective reporting and attrition A
Confounding/variable control A
Observational bias/blinding A
Results presentation A
Overall confidence -\
Reporting quality ^
Allocation A
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Not reported
Legend
Figure 3-94. Male reproductive animal study evaluation heatmap -
reproductive hormones. For additional details see HAWC link.
Organ weights
Potential PFHxS-induced effects on male reproductive organ weights were evaluated in
three high confidence studies using SD rats (NTP. 2018c: Butenhoff etal.. 2009: 3M. 2000b. 2003)
and one medium confidence study using Wistar rats (Ramhaj et al.. 2018) (see Figure 3-95). In SD
rats, exposure to PFHxS for 28 to 44 days at doses ranging from 0.3 to 10 mg/kg-day did not affect
the weights of the testis, epididymis, or seminal vesicle fNTP. 2018c: Butenhoff etal.. 2009: 3M.
2000b. 20031. Furthermore, gestational plus lactational exposure to PFHxS (0.05 to 25 mg/kg-day)
also did not affect organ weights for epididymis, ventral prostrates, seminal vesicles, levator ani, or
testes in Wistar rats fRamhai etal.. 20181.
3-345
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
. ,c> T-
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)
NRl Not reported
Figure 3-95. Male reproductive animal study evaluation heatmap -
reproductive organ weights. For additional details see HAWC link.
Landmarks of male reproductive system development and maturation
One medium confidence gestational exposure study evaluated PFHxS-induced effects on
androgen sensitive developmental landmarks in F1 Wistar rats fRamhai etal.. 20181. Gestational
plus lactational exposure to PFHxS at doses ranging from 0.05 to 45 mg/kg-day did not affect
anogenital distance or nipple retention in Wistar rats. The developmental effects and pregnancy
outcomes of PFHxS exposure are summarized in Section 3.2.3.
Functional measures
Functional measures were evaluated in medium and high confidence studies using mice or
rats (see Figure 3-96). PFHxS exposure for 14 days before mating at doses ranging from 0.3 to
10 mg/kg-day did not have a significant impact on mating or fertility indices in rats or mice
(Ramh0j etal.. 2018: Chang etal.. 2018: Butenhoff et al.. 2009: 3M. 2003).
3-346
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reporting quality
Allocation -
Observational bias/blinding -
Confounding/variable control -
Selective reporting and attrition -
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
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)
NRl Not reported
Figure 3-96. Male reproductive animal study evaluation heatmap - developmental
effects and functional measures. For additional details see HAWC link.
Evidence Integration
The available studies provide inadequate evidence to determine whether PFHxS exposure
has the potential to affect male reproduction in humans. This conclusion is based on studies in both
humans and animals (see Table 3-42).
The available evidence on PFHxS-induced male reproductive effects in human studies is
considered indeterminate. Outcomes evaluated in human studies include semen parameters, male
reproductive hormones, and onset of puberty. No associations were observed for reproductive
hormone measures. Exposure-related alterations in sperm morphology and age of puberty were
reported. However, considerable uncertainties were also identified that reduce the strength of
evidence (see Table 3-42).
The available evidence on PFHxS-induced male reproductive effects in animal toxicity
studies is also considered indeterminate. Experimental studies using different laboratory rodent
species measured parameters considered indicative of potential adverse responses, including
reproductive organ weights, sperm measures, histopathology, reproductive hormones, and
developmental and functional measures. No significant exposure-related effects were observed for
the measured reproductive parameters in the available studies. While a judgment of compelling
evidence of no effect was considered for characterizing the animal evidence, significant
uncertainties in the animal study database prevent judgments about PFHxS exposure and male
reproductive toxicity from being drawn. Specifically, the short exposure duration in the available
studies is considered inadequate for the evaluation of sperm measures, only a single study
evaluated androgen levels, and other reproductive hormones were not studied.
3-347
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-42. Evidence profile table for PFHxS exposure and male reproductive effects
Evidence stream summary and interpretation
Evidence integration
summary judgment
Evidence from studies of exposed humans
Studies and
confidence
Factors that increase certainty
Factors that
decrease certainty
Summary and key findings
Evidence
stream
judgment
OOO
Evidence inadequate
Sperm parameters
3 medium and 2 low
confidence studies
• No factors noted
• Unexplained
inconsistency
across studies
• Imprecision -
for sperm
concentration
Decreased normal morphology
and concentration in one
medium confidence study.
ooo
Indeterminate
Some
Primary Basis:
Evidence is inconsistent
across studies or largely
null.
Human relevance:
Without evidence to the
Reproductive
hormones
4 medium and 5 low
confidence studies
• Unexplained
inconsistency
across studies
• Low confidence
studies
Inverse association with
testosterone and estradiol in
some low confidence studies, but
medium confidence studies were
null. No association with LH or
FSH levels.
evidence of
association
with sperm
motility, and
pubertal
development.
Significant
uncertainty
due to lack of
consistency
and
coherence
contrary, effects in rodent
models are considered
relevant to humans. The
rodent and human male
reproductive systems share
many conserved features.
Pubertal development
2 medium confidence
study
• No factors
noted
Significant association between
exposure and lower puberty age
in 1 of 2 studies.
Cross-stream coherence:
N/A, human and animal
evidence indeterminate
Susceptible populations and
life stages:
N/A evidence inadequate
to draw inferences
3-348
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration
summary judgment
Evidence from in vivo animal studies
Studies and
confidence
Factors that increase certainty
Factors that
decrease certainty
Summary and key findings
Evidence
stream
judgment
Sperm parameters
3 low confidence
studies in adult rats
and mice:
• 28-d
• 44-d
• 42-d
• All low
confidence
studies - Low
sensitivity
No observed effects on sperm
measures in low confidence,
insensitive studies
ooo
Indeterminate
Certainty in
the
consistently
null findings
was reduced
due to
notable data
gaps.
Histopathology
2 high confidence
studies in adult rats:
• 28-d
• 44-d
1 medium confidence
study in adult rats
• 42-d
• High or medium
confidence in studies, with
sensitive outcome
measures and low risk of
bias.
• No factors
noted
No observed effects on
histopathological outcomes
Hormone levels
1 high confidence
study in adult rats
• 28-d
No observed effects on
testosterone levels
3-349
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Evidence stream summary and interpretation
Evidence integration
summary judgment
Organ weights
3 high confidence
studies in adult rats
• 28-d (x2)
• 44-d
1 medium confidence
study in rats
• GD7-PND22
No observed effects on
reproductive organ weights
Developmental effects
1 high confidence
study in rats
• GD7-PND22
No observed effects on male
reproductive organ development
Functional measures
2 high confidence
studies in rats and
mice
• 14-d (x2)
No observed effects on mating or
fertility index
3-350
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3.2.10. Renal Effects
Human Studies
Seventeen studies (reported in 27 publications) investigate the relationship between PFHxS
exposure and markers of renal function, primarily measures of glomerular filtration rate (GFR) and
uric acid (UA). Three studies (Zhang et al.. 2019b: Seo etal.. 2018: Rotander etal.. 2015b) were
considered uninformative due to critical deficiencies in the confounding domain (see Figure 3-97).
The remaining 14 studies were primarily cross-sectional analyses and were classified as low
confidence primarily due to concerns for reverse causality without other major methodological
limitations. In essence, as described in Watkins etal. f20131. decreased renal function (as measured
by decreased GFR or other measures) could plausibly lead to higher levels of PFAS, including
PFHxS, in the blood. This hypothesis is supported by data presented by Watkins etal. (2013).
although there is some uncertainty in the conclusions due to the use of modeled exposure data as a
negative control and the potential for the causal effect to occur in addition to reverse causality. In
contrast, others have hypothesized that renal disease may result in lower PFHxS levels, which
would result in underestimation of the effect of PFHxS exposure rather than overestimation (Tain
and Ducatman. 2019bl. A possible mechanism of this source of reverse causation in the inverse
direction is the relationship of renal function with albuminuria, which can be caused by both
hypertension and diabetes, both also common causes of kidney disease (Tain and Ducatman.
2019b). Further, there may be differences in how PFHxS excretion is affected depending on the GFR
stage or severity of renal disease, which complicates interpretation of the study findings. In any
case, the results least likely to be affected by reverse causality were analyses in four studies (four
publications) designed to assess reverse causality (e.g., stratification by glomerular filtration stage
or modeling with PFHxS as the dependent variable) fZengetal.. 2019c: Moon. 2021: Lin etal..
20211: Tain f20191: f Conway etal.. 20181 and two studies with prospective designs fLin et al..
2021): Blake etal. (2018). Of these, Lin etal. (2021) had the benefit of both prospective data
analysis and additional analyses and was thus rated as medium confidence. Across studies, because
of the potential for reverse causation, there is considerable uncertainty in interpreting the results of
the available studies. Outside of these concerns, the informative studies were well conducted and
had adequate or good ratings for all domains other than exposure measurement
3-351
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
xe<£°°
Blake, 2018, 5080657-
Cakmak, 2022, 10273369-
Chen, 2019, 5387400-
Conway, 2018, 5080465-
Jain and Ducatman, 2019, 5080477-
Kataria, 2015, 3859835-
Lin, 2020, 6988476-
Lin, 2021, 7410157-
Mao, 2020, 6988481 -
Qin, 2016, 3981721 -
Rotander, 2015, 3859842-
Sagiv, 2015, 3859838-
Seo, 2018,4238334-
Wang, 2019, 5080583-
Watkins, 2013, 2850974-
Zeng, 2019, 5918630-
Zhang, 2019, 5083675-
+
-
++
+
++
+
+
-
~
+
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)
++
-
++
+
+
+
+
-
+
-
++
+
+
¦
+
¦
B
+
-
++
+
+
+
+
-
++
-
++
+
+
+
+
-
++
-
++
+
+
+
+
-
-
-
+
+
+
+
+
-
+
+
++
+
++
+
+
+
++
-
-
+
+
+
+
-
+
-
++
+
++
+
+
-
"
-
-
--
-
-
•
++
-
++
+
+
+
+
¦
"
-
-
O
-
-
-
+
-
++
++
+
++
+
+
-
++
-
+
++
+
+
-
+
-
++
+
+
+
+
-
B
¦
-
B
¦
~
Figure 3-97. Renal effects human study evaluation heatmap. For additional
details see HAWC link.
Multiple publications of the same study: Jain and Ducatman (2019c) also includes Jain and Ducatman (2019a), Jain
(2019), Jain (2013), Jain (2021b), Jain (2020), Jain (2021a), Moon (2021), and Scinicariello et al. (2020b). These
were all analyses of NHANES data that have overlapping study populations. While only one is presented in the
heat map, each was reviewed for additional information/analyses and included in the synthesis but are not
considered independent support.
Across the 14 available studies, there is an indication of impaired renal function (i.e., lower
GFR, higher UA, creatinine, or disease) in nine fWatkins etal.. 2013: Sagiv etal.. 2015: Oin etal..
2016: Mao etal.. 2020: Lin etal.. 2020c: Lin etal.. 2021: Cakmak etal.. 2022: Blake etal.. 20181.
including multiple NHANES publications that are counted as one study f Scinicariello etal.. 2020b:
Moon. 2021: lain and Ducatman. 2019c). but there are some inconsistencies (see Table 3-43).
3-352
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Considering GFR in adults, Blake etal. (2018). Sagivetal. (2015). Moon (2021). and Lin et
al. f20211 reported lower GFR with higher exposure, all statistically significant, though the
association in Lin etal. f20211 was observed only in participants with hypertension (the direction
was in the opposite direction for participants without hypertension). A different study of NHANES
data overlapping with the population in Moon f20211 but using an alternative analytical approach,
lain and Ducatman (2019c). reported an inverted U-shape response with GFR (higher exposure
levels in the second and third tertiles than first and fourth, also observed in analyses stratified by
sex), which may reflect differences in PFHxS excretion by GFR stage. In contrast to the majority of
studies, Conway etal. f 20181 and Wang etal. f2019al reported higher GFR with higher exposure
(not statistically significant). In children and adolescents, Watkins etal. f20131 reported lower GFR
with higher exposure while Kataria et al. f20151 also reported the inverted U shape with GFR.
Looking at uric acid and hyperuricemia, positive associations were again observed in the
majority of studies, with each being statistically significant in at least one sub-group. Scinicariello et
al. (2020b) reported higher odds (unstratified by sex) with an exposure-response gradient
observed across quartiles. Zeng etal. (2019c) reported higher odds of hyperuricemia in women but
not men, while Lin etal. f2020cl reported higher uric acid in the fourth quartile in men but not
women, so there was not a consistent pattern by sex.
One study examined creatinine and found a positive association fCakmak etal.. 20221.
In the few studies of renal disease, there was a positive association with kidney stones in a
single study (Mao etal.. 2020). However, no association was observed with chronic kidney disease
in the only study that reported it (Wang etal.. 2019c).
Overall, there are generally consistent associations between impaired renal function and
PFHxS exposure, with an inverted U shape of responses across GFR stages observed in multiple
studies that may be explained by differences in the ability of the kidney to reabsorb PFAS flain and
Ducatman. 2019cl. However, the potential for reverse causation is an important source of
uncertainty. However, in the studies with less potential for reverse causation, there is an indication
that this bias is unlikely to fully explain the observed associations. Significant associations were
observed in both studies with prospective exposure measurement (Lin etal.. 2021: Blake etal..
2018). though only in participants with hypertension in Lin etal. (2021). While prospective
measurement does not eliminate the possibility of reverse causation due to ongoing exposure prior
to study enrollment, the effect is likely lower. Further, Lin etal. f20211 performed a secondary
analysis using baseline GFR as the independent variable and repeated measures of PAS as the
dependent variable and found that PFAS levels did not differ significantly by baseline GFR. A similar
analysis without repeated measures in Moon (2021) also indicated that reverse causation was not
likely to explain the results. Alternatively, if reverse causation is in the inverse direction, as
described above, the observed associations could be underestimates of the true effect, but the
available data are insufficient to determine whether this is likely to be the case.
3-353
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
While there is uncertainty in the association with renal disease, the general consistency in
the findings for uric acid and hyperuricemia may provide coherence with the increases in liver
enzymes described in Section 3.2.4. Uric acid is positively associated with both ALT fChenetal..
20161 and chronic liver disease fAfzali etal.. 20101.
Table 3-43. Associations between PFHxS exposure and renal function
Reference,
confidence
Study population
Median exposure
level (IQR)
in ng/mL
Form and units of
effect estimate
Effect estimate
Glomerular filtration rate (GFR)
Decrease indicates impaired renal function
Wang et al. (2019a),
Cross-sectional study
(2015-2016); China;
1,612 adults
0.7 (0.01,2.7)
Mean change (95% CI)
in eGFR per In-unit
change
0.24 (-0.02, 0.50)
Low
Watkins et al.
Cross-sectional study of
9,660 children in U.S.
exposed to high PFOA
IQR 1.3
Mean change (95% CI)
per IQR increase exp
-1.0 (-1.5, -0.4)*
(2013), Low
Jain and Ducatman
Cross-sectional study
(NHANES) (2007-2014);
U.S.; 6,836 adults
1.4
Adjusted geometric
means (95% CI) by
glomerular function
stage (GF-1 is normal
or high filtration; GF-
3B/4 is moderately to
severely decreased)
All participants
GF-1: 1.20(1.14-1.27)
GF-2: 1.73 (1.61-1.86)
GF-3: 1.83 (1.63-2.05)
GF-3B/4: 1.01(0.78-
1.31)
(2019c), Low
Moon (2021), Low
Cross-sectional study
(NHANES) (2003-2018);
U.S.; 14,373 adults
1.5 (0.8-2.6)
P (p-value) for In-unit
increase
-1.52 (-2.10, -0.94)*
Kataria et al. (2015),
Cross-sectional study of
1,960 adolescents in
U.S.
2
P (95 CI) for quartiles
vs. Q1
Q2: 1.4 (-3.6,6.3)
Q3: 1.9 (-3.4,7.1)
Q4: -0.3 (-4.4,3.8)
Low
Sagiv et al. (2015),
Cross-sectional study of
1,645 pregnant women
in U.S.
2.4
(1.6-3.8)
% change GFR
-4.3 (-5.3, -3.3)*
Low
Geometric means
(IQR) of exp by
quartile
Ql: 3.0 (1.9,4.3)
Q2: 2.7 (1.7,4.1)
Q3: 2.3 (1.5,3.2)
Q4: 2.2 (1.5,3.5)*
Lin etal. (2021),
Cohort study within
placebo and lifestyle
intervention arms of a
diabetes prevention
randomized controlled
trial of 875 adults in the
U.S.
2.4 (1.6-3.8)
P (95 CI) for doubling
of baseline exposure
0.21 (-0.79, 1.21)
With hypertension
-2.35 (-4.46, -0.25)*
Without hypertension
1.24 (0.09, 2.39)*
Medium
Blake etal. (2018),
Prospective cohort of
residents near a
uranium processing site
(1990-2008); U.S.; 210
adults
2.7
(1.7-4.1)
Percent change (95%
CI) in eGFR per IQR
change
-2.06 (-3.53, -0.59)*
Low
3-354
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference,
confidence
Study population
Median exposure
level (IQR)
in ng/mL
Form and units of
effect estimate
Effect estimate
Conwav et al.
(2018), Low
Cross-sectional study of
53,650 adults in U.S.
exposed to high PFOA
3.0
(1.9-4.8)
OR (95% CI) for 1-unit
increase
GF-1: 2.07 (1.69-2.55)
GF-2: 2.29 (1.86-2.81)
GF-3A: 2.37 (1.87-2.84)
GF-3B: 2.30 (1.83-2.90)
GF-4/5:1.0 (ref)
Uric acid (UA)
Increase indicates impaired renal function
Zeng et al. (2019c),
Low
Cross-sectional study of
1,612 adults in China
0.7 (0.01-2.7)
Mean difference per
log-unit increase
0.01 (-0.15, 0.03)
GF-1: -0.01 (-0.06,
0.04)
GF-2: -0.00 (-0.03,
0.03)
GF-3: 0.05 (-0.04,0.15)
GF-4: -0.04 (-0.23,
0.15)
OR (95% CI) for
hyperuricemia for log-
unit increase
1.01 (0.97, 1.06)
Women: 1.18(1.01,
1.37)*
Men: 0.99 (0.95,1.04)
Chen et al. (2019a),
Low
Cross-sectional study of
122 adults in China
GM 0.8, range 0.3-
2.4
P (95% CI) for In-unit
increase
-4.42 (-24.23, 15.38)
Qin et al. (2016),
Low
Cross-sectional study of
225 children in Taiwan
1.3
(0.6-2.8)
P (95 CI) for In-unit
increase
0.14 (0.02,0.26)*
OR (95% CI) for
quartile increase exp
and high UA
1.4(0.9,2.1)
Jain and Ducatman
(2019a), Low
Cross-sectional study
(NHANES) (2007-2014);
U.S.; 6,836 adults
1.4
P (p-value) for 1-unit
increase
In GF-1 participants
Women: 0.023 (<0.01)*
Men: 0.015 (0.06)
Scinicariello et al.
(2020b), Low
Cross-sectional study
(NHANES) (2009-2014);
U.S.; 4,917 adults
1.4 (GM)
P (95% CI) in serum
uric acid for quartiles
vs. Q1
Q2:0.14 (0.02, 0.26)*
Q3: 0.22 (0.08, 0.36)*
Q4:0.33 (0.19, 0.47)*
OR (95% CI) in
hyperuricemia for
quartiles vs. Q1
Q2: 1.15 (0.89, 1.50)*
Q3: 1.33 (0.95, 1.86)*
Q4: 1.51(1.12,2.03)*
Kataria et al. (2015),
Low
Cross-sectional study of
1,960 adolescents in the
U.S.
2
P (95% CI) for
quartiles vs. Q1
Q2: 0.04 (-0.1,0.2)
Q3: 0.05 (-0.1,0.2)
Q4: -0.05 (-0.2,0.1)
Lin et al. (2020c),
Low
Cross-sectional study
(2016-2017); Taiwan;
397 older adults (55—
75 yr)
2.7 (1.9-3.7)
P (95% CI) in serum
uric acid for quartiles
vs. Q1
Q2: 0.01 (-0.32, 0.33)
Q3:-0.1 (-0.44, 0.23)
Q4: 0.39 (0.05, 0.72)*
Women:
Q2: 0 (-0.36, 0.35)
Q3:-0.1 (-0.46, 0.26)
Q3: 0.05 (-0.31, 0.42)
Men:
Q2: -0.31 (-0.97, 0.35)
3-355
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference,
confidence
Study population
Median exposure
level (IQR)
in ng/mL
Form and units of
effect estimate
Effect estimate
Q3: 0.3 (-0.37, 0.96)
Q4: 0.89 (0.22, 1.56)*
Creatinine
Increase indicates impaired renal function
Cakmak et al.
(2022), Low
Cross-sectional study
(2007-2017); Canada;
6,045 adults
1.5 (GM)
% change per 1 mean
increase in PFDA
1.0 (0.1, 1.8)*
Chronic kidney disease
OR>l indicates more disease
Wang et al. (2019b),
Low
Cross-sectional study
(2015-2016); China;
1,612 adults
0.7 (0.01-2.7)
OR (95% CI) for
chronic kidney
disease per In-unit
change in PFDA
1.01 (0.94, 1.07)
Kidney stones
Mao etal. (2020),
Low
Cross-sectional study
(NHANES) (2007-2016);
U.S.; 8,453 adults
1.5 (0.8-2.5)
OR (95% CI) for
kidney stone history
for tertiles vs. T1
T2: 1.24 (1.03, 1.51)*
T3: 1.35 (1.10, 1.68)*
*p < 0.05.
Animal Studies
There are two 28-day oral gavage exposure studies in Sprague Dawley rats fNTP. 2018b:
3M. 2000a] and two 42-44-day exposure oral gavage studies in CD-I mice f Chang etal.. 20181 and
Sprague Dawley rats fButenhoff etal.. 2009: 3M. 20031 that measure effects relevant to the
assessment of the renal system after repeated oral dose exposure to PFHxS. The studies report on
clinical chemistry (serum) biomarkers of effect, histopathology, and organ weights. Overall study
confidence was high for most endpoints evaluated in these studies with the exception of organ
weights and serum markers in Chang etal. (2018). which had incomplete reporting of null data
(results were only discussed qualitatively) resulting in a medium confidence rating (see Figure 3-
98).
3-356
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reporting quality
Allocation
Observational bias/blinding
Confounding/variable control
Selective reporting and attrition
Chemical administration and characterization
Exposure timing, frequency and duration
Results presentation
Endpoint sensitivity and specificity
Overall confidence
Legend
Good (metric) or High confidence (overall)
Adequate (metric) or Medium confidence (overall)
Deficient (metric) or Low confidence (overall)
Critically deficient (metric) or Uninformative (overall)
Not reported
Figure 3-98. Renal effects - animal study evaluation heatmap. For additional
details see HAWC link.
Clinical chemistry
Serum biomarkers of renal injury (including blood urea nitrogen [BUN], creatinine,
creatinine kinase, and total protein) were measured in Sprague Dawley rats after short-term (28-
day) exposure fNTP. 2018b: 3M. 2000al. and two 42- or 44-day exposure studies using CD-I mice
and Sprague Dawley rats (Chang etal.. 2018: Butenhoff et al.. 2009: 3M. 2003). In the F0 generation
male Sprague Dawley rats, 44 days of exposure to PFHxS at the highest tested dose, 10 mg/kg-day,
resulted in a 31% increase in BUN when compared with controls (Butenhoff etal.. 2009: 3M. 20031.
However, no effects were observed for creatinine, creatinine kinase, or total protein in male
animals and female animals from the same study fButenhoff et al.. 2009: 3M. 20031: a similar study
using CD-I mice reported no effects on creatinine, urea nitrogen, and electrolytes in F0 generation
male and female animals exposed to same levels of PFHxS (10 mg/kg-day) for 44 days; and two 28-
day study using SD rats reported no exposure-related effects in creatinine, creatinine kinase, blood
urea nitrogen (BUN), or total protein after PFHxS exposure at doses ranging from 0.6 to 10 mg/kg-
day (NTP. 2018b: 3M. 2000al. BUN is considered a late biomarker of renal injury not normally
affected until at least half of the kidney mass is compromised (Khan etal.. 20181. The biological
significance of the PFHxS-induced BUN increase observed in the NTP study is not clear as BUN was
not affected in similar studies, and other clinical indicators of kidney damage were not altered in
the available studies.
3-357
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Histopathology
Renal histopathology was evaluated across two 28-day gavage studies fNTP. 2018a: 3M.
2000a) and one 42- to 44-day exposure toxicity study fButenhoffetal.. 2009: 3M. 20031. All studies
used Sprague Dawley rats. Exposure to PFHxS for 28 to 44 days at doses ranging from 0.3 to
10 mg/kg-day did not have any notable treatment-related impacts on kidney histopathology. One
28-day short-term study also evaluated the urinary bladder and reported no effects fNTP. 2018a).
In this study chronic progressive nephropathy34 graded as minimal occurred in the kidneys of all
exposed animals, including controls.
Organ weight
Absolute and relative (to body weight) kidney weights were measured in the two 28-day
gavage studies using Sprague Dawley rats (NTP. 2018a: 3M. 2000a) and the two 42- to 44-day
exposure studies using Sprague Dawley rats (Bute nhoff etal.. 2009: 3M. 2003) or CD-I mice (Chang
etal.. 2018). Exposure to 10 mg/kg-day PFHxS for 28 days increased relative kidney weights in
male Sprague Dawley rats fNTP. 2018al. This response was not observed in female animals fNTP.
2018a) and none of the remaining studies exposing rats or mice to similar doses and durations
(ranging from 28 to 44 days) did not observe significant PFHxS-induced changes in relative or
absolute kidney weights f Chang etal.. 2018: Butenhoff et al.. 2009: 3M. 2000a. 20031.
Evidence Integration
The available evidence suggests but is not sufficient to infer that exposure to PFHxS might
cause renal system effects in humans given sufficient exposure conditions35 (see Table 3-44).
The available evidence on PFHxS-induced renal effects in humans is considered slight. The
evidence for potential renal system effects in humans is based on reported associations between
PFHxS exposure and impaired renal function in nine out of 14 informative epidemiological studies
including several statistically significant findings. There is considerable uncertainty remaining due
to the potential for reverse causation.
The available evidence on PFHxS-induced renal effects in animal toxicity studies is also
considered indeterminate. The experimental animal evidence informing potential renal system
effects is limited to two 28-day gavage studies in Sprague Dawley rats (NTP. 2018a: 3M. 2000a).
and two 42- to 44-day exposure studies using Sprague Dawley rats fButenhoff et al.. 2009: 3M.
20031 or CD-I mice fChang etal.. 20181. The studies were generally well conducted (confidence
ratings were high/medium) and reported on relevant measurements, including serum biomarkers
of renal injury (i.e., BUN, creatinine, and creatinine kinase), kidney and urinary bladder
histopathology and kidney weights. Although a few significant findings were observed, PFHxS
34Chronic progressive nephropathy is a commonly observed spontaneous lesion frequently observed in 2 to
13-week studies using SD rats (Khan et al.. 20181.
35The "sufficient exposure conditions" are more fully evaluated and defined for the identified health effects
through dose-response analysis in Section 5.
3-358
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
exposure generally did not affect the renal system in the available studies. However, the absence of
long-term studies limits the evaluation of potential renal system toxicity in animals following
PFHxS exposure, hence a conclusion of compelling evidence of no effect was not considered
appropriate.
3-359
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 3-44. Evidence profile table for PFHxS urinary system effects
Evidence stream summary and interpretation
Evidence integration
summary judgment
Evidence from studies of exposed humans
Studies and
confidence
Factors that increase certainty
Factors that
decrease
certainty
Summary and key findings
Evidence stream
judgment
Evidence suggests but
is not sufficient to
Renal Functions
1 medium and 13 low
confidence studies
• Consistency
• Precision
• Primarily low
confidence
studies -
potential
reverse
causality
9 of 14 studies reported associations
between PFHxS exposure and
impaired renal function. Reverse
causality is an important source of
uncertainty.
®oo
Slight
infer
©OO
Primary Basis:
Generally consistent
evidence across studies
in humans.
Human relevance:
N/A
Cross-stream coherence:
N/A. Evidence in animals
is indeterminate.
Evidence from in vivo animal studies
Studies and
confidence
Factors that increase certainty
Factors that
decrease
certainty
Summary and key findings
Evidence stream
judgment
Serum Biomarkers of
Renal Injury,
Histopathology,
Organ Weights
3 high confidence
studies in adult rats:
• 28-d (x2)
• 44-d
1 medium confidence
study using mice
• 44-d
• All high or medium confidence
studies
• Unexplained
inconsistency
• Increased BUN reported in
one study, bit no effects in
remaining studies and no
response in other markers of
renal disease.
• No PFHxS-induced effects on
histopathological outcomes.
• No observed PFHxS-induced
effects on kidney weights
QQQ
Indeterminate
3-360
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
3.2.11. Other Noncancer Health Effects
Human Studies
Eleven epidemiology studies (reported in 16 publications) report on the relationship
between PFHxS exposure and musculoskeletal effects, specifically bone mineral density and
osteoporosis. Most of these studies examined continuous measures of bone mineral density, while
one study was a case-control study of osteoporosis (Banjabi etal.. 20201. Nine studies were medium
confidence and had no serious concerns for risk of bias. The case-control study was low confidence
due to concerns for differences in the selection of cases and controls. The other low confidence
study was a small pilot study with concerns for residual confounding fKhalil etal.. 20181. Study
evaluations are summarized in Figure 3-99.
" r-'S.
Banjabi, 2020, 6833613-
-
¦
++
+
+
+
-
Beglarian, 2024, 11412754-
+
++
++
++
+
+
+
Blomberg, 2022, 10618582-
+
++
++
++
-
+
+
Buckley, 2021, 9959608-
+
++
++
++
+
+
+
Carwile, 2022, 10412903-
++
B
++
+
+
+
+
Cluett, 2019, 5412438-
++
++
++
+
+
+
Fan, 2023, 10701796-
+
+
+
+
+
+
+
+
Hu, 2019, 6315798-
+
++
++
+
++
+
+
+
Hojsager, 2023, 11412360-
+
++
++
+
+
-
+
+
Jeddy, 2018, 5079850-
+
B
++
+
+
+
+
+
Khalil, 2018, 4238547-
-
++
-
+
-
+
-
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 3-99. Musculoskeletal effects human study evaluation heatmap. For
additional details see HAWC link. Multiple publications of the same study: fCarwile
etal.. 20221 also includes Colicino etal. (2020). fKhalil etal.. 20161. fKirketal..
20231. fXiong et al.. 20221. and fZhao et al.. 20221.
The results for associations with bone mineral density in medium confidence studies are
summarized in Table 3-45. The majority of studies were null, without consistency in the direction of
association across studies. Statistically significant associations were observed in two studies, with
3-361
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
total bone mineral density in Fan etal. (2023) and total femur bone mineral density in Xiong et al.
f20221. but these were in opposite directions. No apparent pattern was observed by age group or
sex (a minority of studies reported sex-stratified results which are not shown), though interactions
were observed in single studies for characteristics such as menopausal status fZhao etal.. 20221
and albuminuria fXiong etal.. 20221 that have insufficient data to explore further. One low
confidence study reported inverse but not statistically significantly associations with stiffness
index, speed of sound, and broadband ultrasound attenuation (Khalil etal.. 2018).
Two studies, including the low confidence case-control study Baniabi etal. (2020). and a
medium confidence study fFan etal.. 20231 examined osteoporosis as a dichotomous outcome. Fan
etal. f20231 reported an odds ratio of 1.23 (95% CI 0.95,1.60), while Baniabi et al. f20201 did not
report higher odds with higher exposure (OR [95% CI] vs. Ql: Q2 2.47 [0.12, 50.2], Q3 0.30 [0.01,
9.18], Q4 0.05 [0.00, 3.05]).
Overall, the evidence for an association between PFHxS exposure and musculoskeletal
effects is indeterminate. Exposure concentrations were somewhat low across studies, so it is
possible that there was insufficient sensitivity to detect an effect
Table 3-45. Associations between PFHxS exposure and bone mineral density
in medium confidence epidemiology
Reference
Population
Median
exposure (IQR)
or as specified
(ng/mL)
Effect
estimate
Total body bone
mineral density
Site-specific bone mineral
density, as specified
Blomberg et al.
(2022)
Birth cohort
with follow-up
to 9 yr, Faroe
Islands; 366
children
0.2 (0.1-0.2)
P (95% CI) for
doubling
Exposure at birth
0.1 (-0.02, 0.22)
Exposure at 9 yr
0.00 (-0.15, 0.16)
NR
Hgisager et al.
(2023)
Birth cohort
with follow-up
to 7 yr;
Denmark; 881
children
0.3 (0.4-0.8)
P (95% CI) for
doubling
Prenatal exposure
-0.02 (-0.07, 0.03)
Exposure in childhood
-0.08 (-0.16, 0.0)
NR
Fan et al.
(2023)
Cross-sectional
study, China;
1,260 adults
0.9(0.5-1.4)
P (95% CI) for
In-unit
increase
Bone mineral density T-
score
-0.23 (-0.33, -0.12)*
NR
Carwile et al.
(2022)
Xiong et al.
(2022)
Zhao et al.
(2022)
Cross-sectional
study, U.S.;
NHANES 2011-
2016; 896
adolescents
0.9(0.6-1.6)
P (95% CI) for
doubling
Bone mineral density Z-
score
Boys: -0.06 (-0.16,
0.04)
Females: 0.02 (-0.06,
0.11)
NR
NHANES 2005-
2010; 1,228
adolescents
3.9 (mean)
P (95% CI) for
1 unit
increase
NR
Total femur: 0.01 (0.00, 0.01)*
Lumbar spine: 0.00 (-0.01, 0.00)
3-362
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Reference
Population
Median
exposure (IQR)
or as specified
(ng/mL)
Effect
estimate
Total body bone
mineral density
Site-specific bone mineral
density, as specified
NHANES 2005-
2014; 6,416
adolescents and
adults
2.7 (mean)
P (95% CI) for
In-unit
increase
NR
Total femur: -0.004 (-0.01, 0.00)
Bucklev et al.
Birth cohort
with follow-up
to 12 yr, U.S.;
206 adolescents
1.3 (0.8-2.3)
P (95% CI) for
doubling
Bone mineral density Z-
score
0.00 (-0.11, 0.10)
Hip: -0.03 (-0.17, 0.11)
Femoral neck: 0.04 (-0.09, 0.18)
Spine: -0.11 (-0.25, 0.02)
(2021)
Beglarian et al.
Two cohorts
with follow-up
to adolescence,
U.S.; 441
adolescents and
young adults
1.4 (0.6-3.3)/
1.1 (0.5-3.1 by
cohort
P (95% CI) for
doubling
0.00 (-0.01, 0.01)
Trunk: 0.00 (-0.01, 0.01)
(2024)
Jeddv et al.
Birth cohort
with follow-up
to age 17, U.K.;
257 adolescent
girls
1.7(1.3-2.3)
P (95% CI) for
1 unit
increase
0.0001 (-0.002, 0.002)
NR
(2018)
Cluett et al.
Cross-sectional
analysis within
birth cohort,
U.S.; 576
children (6-
10 yr)
1.9(2.3)
P (95% CI) for
doubling
Single pollutant
-0.02 (-0.07, 0.03)
Multiple PFAS
0.04 (-0.03, 0.10)
NR
(2019)
Hu et al. (2019)
Cohort within
randomized
controlled trial
on weight loss
with 2 yr follow-
up, U.S.; 294
adults
3.2(1.7,3.9)
P (95% CI) for
SD increase
NR
Hip: 0.00 (-0.01, 0.00)
Femoral neck: 0.00 (-0.00, 0.01)
Spine: 0.00 (-0.00, 0.01)
Cross-sectional
analysis
NR
Hip: -0.01 (-0.02, 0.01)
Femoral neck: -0.01 (-0.02, 0.01)
Spine: -0.01 (-0.03, 0.01)
*p < 0.05.
NR = not reported.
Animal Studies
Several other health effects were examined in experimental animals; however, there were
very little data to inform whether PFHxS exposure might have the potential to cause these effects.
Specifically, the high confidence, 28-day rat study conducted by NTP f2018cl investigated the
potential for PFHxS exposure to cause effects on the alimentary system (including the esophagus,
large, small intestine, pancreas, salivary glands, and stomach), musculoskeletal system, and
respiratory system. For each of these systems, there were no clear PFHxS exposure-related effects
3-363
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
in male or female animals, with the exception of an observation of minimal36 olfactory epithelium
degeneration and minimal hyperplasia along with minimal suppurative inflammation in females,
but not males, in the highest exposure group (8/10 rats in 50 mg/kg-day exposure group). Overall,
the sparsity of evidence on these outcomes prevents any interpretation from being drawn.
Evidence Integration
The currently available evidence is inadequate to assess whether PFHxS may cause other
noncancer health effects in humans, including those related to the alimentary system,
musculoskeletal system, and respiratory system. In general, the data available for these health
outcomes were largely null and/or absent (i.e., indeterminate evidence from human and animal
studies) and considerable data gaps remain for these health effects.
3.3. CARCINOGENICITY
3.3.1. Cancer
The systematic review identified 12 epidemiologic studies that evaluated the risks of cancer
associated with exposures to PFHxS fYeung etal.. 2013: Wielsae etal.. 2017: Velarde etal.. 2022:
Tsai etal.. 2020: Omoike etal.. 2021: Liu etal.. 2021b: Lin etal.. 2020a: Li etal.. 2022a: Hardell et
al.. 2014: Ghisari etal.. 2017: Christensen etal.. 2016: Bonefeld-largensen etal.. 2014). Six cancer
studies (Wielsae etal.. 2017: Velarde etal.. 2022: Omoike etal.. 2021: Lin etal.. 2020a: Li etal..
2022a: Christensen etal.. 2016) were evaluated as 'Uninformative.' One study (Yeung etal.. 2013)
was screened as related to hepatocellular carcinoma cancer, but actually examined the serum and
liver concentrations of PFAS, including PFHxS, among patients who had liver transplants—some of
whom had hepatocellular carcinoma cancer; this study did not assess cancer risk and was not
evaluated for study quality.37
No animal in vivo, mutagenicity or genotoxicity studies were identified in the database.
36Minimal refers to average histological severity grade as follows: 1 = minimal; 2 = mild; 3 = moderate;
4 = marked) as determined by NTP (2018c).
37Yeungetal. (2013) was not included in the Figure 3 because this study did not assess cancer risk.
3-364
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Bonefeld-Jorgensen, 2014, 2851186-
Christensen, 2016, 3858533-
Liu, 2021, 10176563-
Omoike, 2020, 7021502-
Tsai, 2020, 6833693 -
Velarde, 2022, 9956482
Wielsoe, 2017, 3858479
Ghisari, 2017, 3860243-
Hardell, 2014, 2968084-1
Li, 2021, 9961926-1
Lin, 2020, 6835434-
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 3-100. Study evaluation results for epidemiology studies of PFHxS and
cancer. For additional details see HAWC link.
Human Studies
The study of prostate cancer fHardell etal.. 20141 was low confidence due to concern about
the exposure measurement not representing the etiologically relevant time period, potential for
confounding, insufficiencies in the analysis, and concerns about sensitivity (see Figure 3-100).
Hardell etal. (20141 reported a non-significantly increased risk of prostate cancer among men with
PFHxS concentrations in blood that were above the median value; and a higher, borderline
significant, risk of prostate cancer among men with PFHxS concentration greater than the 75th
percentile. Hardell etal. f20141 also reported that men with PFHxS concentrations above the
median and with a first-degree relative with prostate cancer were at significantly increased risk.
The study of thyroid cancer (Liu etal.. 2021b) was low confidence due to concern about the
exposure measurement not representing the etiologically relevant time period, deficiencies
regarding the outcome definition, and potential for confounding, (see Figure 3-95). Liu et al.
f2021bl reported significantly decreased risk of thyroid cancer associated with increasing quartiles
of PFHxS. The first study of breast cancer fBonefeld-T0rgensen etal.. 20141 was low confidence due
to concerns about participant selection and potential selection bias as there was: (1) no explanation
of why 29% of cases were withdrawn from the National Patient Registry, (2) no comparisons of the
subjects' details between the withdrawn cases and the originally selected cases, and (3) no
3-365
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
consideration of how the originally matched controls might no longer match the final set of cases.
Bonefeld-largensen etal. f20141 studied the effect of PFHxS on the risks of breast cancer in Danish
women using a case-control study, and initially found a significantly decreased risk of breast cancer
with increases in continuously measured PFHxS, although in subsequent analyses, excluding 72
breast cancer cases (29% of the cases) which were withdrawn from the National Patient Registry,
the effects changed slightly and lost statistical significance. The second study of breast cancer
Ghisari etal. (20171 was low confidence because it was based on the same case-control as Bonefeld-
l0rgensen etal. (20141 and had the same deficiencies. Ghisari etal. (20171 investigated genetic
polymorphisms as potential effect modifiers of the risk of PFAS on breast cancer. They reported
that none of the genetic polymorphisms evaluated was an effect modifier, but that some genotypes
(CYP1B1 Val/Val, COMT Val/Val, CYP17 A1 /A1 and CYP19 CT) were associated with significantly
decreased risks of breast cancer associated with increased PFHxS exposure. The third study of
breast cancer (Tsai etal.. 2020) was low confidence due to concern about the exposure
measurement not representing the etiologically relevant time period, potential for confounding,
and concerns about low sensitivity (see Figure 3-95). Tsai etal. (2020) reported significantly
increased risk of breast cancer per In-transformed unit increase in PFHxS concentration in blood
among women <50 years of age who were estrogen receptor positive; and non-significantly
decreased risk of breast cancer per ln-transformed unit increase in PFHxS concentration in women
<50 years of age and estrogen receptor negative and in all women >50 years of age. In summary,
the available epidemiologic evidence on PFHxS and the risk of cancer is limited and generally
uninformative.
Animal Studies
No studies were identified in the evidence base evaluating the carcinogenicity of PFHxS in
animals.
Evidence Integration Summary
The available evidence for any effect of PFHxS on the risk of developing or dying from
cancer is inconsistent Thus, the available human evidence on breast, thyroid or prostate cancer is
considered indeterminate and, overall, based on EPA guidelines (U.S. EPA. 2005). there is
inadequate information to assess carcinogenic potential.
3-366
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
4. SUMMARY OF HAZARD IDENTIFICATION
CONCLUSIONS
4.1. SUMMARY OF CONCLUSIONS FOR NONCANCER HEALTH EFFECTS
As described in detail in Section 3, the currently available evidence indicates that exposure
to perfluorohexanesulfonic acid [PFHxS] and its related salts likely results in thyroid (see
Section 3.2.1) and immune (see Section 3.2.2) effects in humans given sufficient PFHxS exposure
conditions. These judgments are based primarily on data from epidemiologic studies for immune
effects and on short-term (28-day exposure), and reproductive (gestational and postnatal
exposure) oral exposure studies in rodents for thyroid effects. Further characterizations of the
exposure conditions relating to these two identified hazards are provided in Section 5.
The hazard identification judgment that the evidence indicates PFHxS exposure is likely to
cause thyroid toxicity, specifically decreased thyroid hormones, in humans given sufficient PFHxS
exposure conditions, is based primarily on a short-term study and two multigenerational studies in
rats reporting a consistent and coherent pattern of hormonal changes at PFHxS exposure levels
>2.5 mg/kg-day. A consistent dose-dependent decrease of T4, and to a lesser extent T3, in adult and
juvenile rats, with a magnitude of effect (up to 70%) in the absence of effects in TSH was observed
(with males being more sensitive). In addition, one multigenerational study reported increased
incidence of minimal thyroid hypertrophy and moderate hyperplasia in male rats after PFHxS
exposure. Because of the similarities in thyroid hormone production between rodents and humans,
the effects in rodents were considered relevant to humans. A detailed discussion of thyroid effects
is included in Section 3.2.1.
The hazard identification judgment that the evidence indicates PFHxS exposure is likely to
cause immunotoxicity in humans given sufficient exposure conditions is based on generally
consistent evidence of reduced antibody response to vaccination at median blood concentrations of
0.2-0.6 ng/mL in children. The direction of association was generally consistent across studies and
timing of exposure and outcome measures, although not all the results were statistically significant
Further, three studies reported higher odds of infectious disease with higher PFHxS exposure,
including total infectious disease, lower respiratory infection, throat infection, pseudocroup, and
gastroenteritis. Lastly, there was some evidence of hypersensitivity, based primarily on a single
well-conducted study of asthma, although findings were inconsistent across studies. A detailed
discussion of immune effects is included in Section 3.2.2.
The evidence suggests but is not sufficient to infer that, given sufficient exposure
conditions, PFHxS exposure may result in adverse health effects on the hepatic, cardiometabolic,
renal, and neurodevelopmental systems, along with developmental effects. These judgments
4-1
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
highlight the notable data gaps and uncertainties identified in the available epidemiological and
experimental animal PFHxS studies (see Section 3.2.3, Section 3.2.4, Section 3.2.5, and Section
3.2.6). The uncertainties in the above-mentioned hazards were considered too large for developing
toxicity values (see Section 5). However, to convey some sense of the magnitude of a potential
estimate for developmental effects, calculations based on this suggestive evidence are provided for
comparison purposes. The objective was to inform the database uncertainty factor (UF) for
quantitative estimates of thyroid and immune effects.
For all other health effects described in Section 3 (i.e., male, and female reproductive,
hematopoietic, and other noncancer effects) the evidence is inadequate to assess whether PFHxS
exposure might cause effects in humans. No quantitative estimates were attempted for these health
effects.
The potential for multiorgan effects of PFHxS exposure exists. As an example, the reported
hypertrophy and hyperplasia in the follicular epithelium cells of the thyroid and in the centrilobular
hepatocytes in the FO male rats exposed to 10 mg/kg-day PFHxS (Butenhoff et al.. 2009) may be
related effects. It has been shown that exposure to compounds that cause microsomal enzyme
induction in the liver can result in a compensatory hypertrophy and hyperplasia of the thyroid due
to increased plasma turnover of T4 and TSH fSanders etal.. 1988: Butenhoff etal.. 20091. However,
as discussed in Section 3.2.1, the authors did not measure thyroid hormones as part of their study
design and therefore the reported observation that thyroid hypertrophy and hyperplasia are
compensatory mechanisms due to turnover of T4 and TSH is speculative. In addition, decreases in
T3 and T4 observed in adult and juvenile animals exposed to PFHxS could be linked to metabolic
effects as well as neurodevelopmental effects such as cognitive decline in children discussed in
detail Section 3.2.1). Lastly, the decreased immune response observed in children exposed to PFHxS
could lead to increased risk of infection as well as cancer fGermolec et al.. 20221. although neither
of these latter effects were well-studied in the available PFHxS evidence base.
4-2
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 4-1. Hazard conclusions across published EPA PFAS human health assessments
Health outcome
PFAS assessments3,15'0
PFHxS
PFDA
PFHxA
PFBA
PFBS
Gen X chemicals
PFOA
PFOS
Endocrine/ Thyroid
+
+
+
+
ND
Human: +
Animal: +/-
Human: +/-
Animal: +/-
Hepatic/Liver
+/-
¦
+
+
+
Human: +
Animal: +
Human: -
Animal: +
Developmental
+/-
+
+
+
¦
+/-
Human: +
Animal: +
Human: +
Animal: +
Reproductive
+
—
—
+/-
Human: -
Animal: +/-
ND
Immunotoxicity
¦
+
—
+/-
Human: +
Animal: +
Human: +/-
Animal: +
Renal
-
-
-
¦
+/-
Human: +/-
Animal: +/-
ND
Hematopoietic/
Hematological
-
-
¦
-
ND
+/-
ND
ND
Ocular
-
-
ND
-
ND
ND
ND
ND
Serum Lipids
-
+/-
ND
ND
-
ND
Human: +
Animal: +
Human: +
Hyperglycemia
-
-
ND
ND
ND
ND
Human: -
Animal: -
Animal: +/-
Nervous System
-
+/-
-
ND
ND
ND
Human: -
Animal: -
Animal: +/-
Cardiovascular
-
+/-
ND
ND
-
ND
ND
ND
4-3
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Health outcome
PFAS assessments3,15'0
PFHxS
PFDA
PFHxA
PFBA
PFBS
Gen X chemicals
PFOA
PFOS
Cancer
-
-
-
-
-
+/-
+/-
+/-
Assessments used multiple approaches for summarizing their noncancer hazard conclusion scales; for comparison purposes, the conclusions are presented as
follows:'+' = evidence demonstrates or evidence indicates (e.g., PFHxA), or evidence supports (e.g., PFBS);= suggestive evidence,= inadequate
evidence (e.g., PFHxA) or equivocal evidence (e.g., PFBS);= sufficient evidence to conclude no hazard (no assessment drew this conclusion); ND = no data
available for this outcome for this PFAS.
bThe assessments all followed the EPA carcinogenicity guidelines (U.S. EPA, 2005) a similar presentation to that used to summarize the noncancer judgments is
applied for the cancer hazard conclusions, as follows:'+' = carcinogenic to humans or likely to carcinogenic to humans;'+/-' = suggestive evidence of
carcinogenic potential;= inadequate information to assess carcinogenic potential;= not likely to be carcinogenic to humans(no assessment drew this
conclusion); ND = no carcinogenicity data available for this PFAS.
cThe hazard conclusions for the various EPA PFAS assessments presented in this table were not considered during evidence integration and thus did not inform
the evidence integration conclusions presented in the PFHxA assessment. Published EPA PFAS human health assessments: U.S. EPA PFDA (U.S. EPA, 2023a),
PFHxA (U.S. EPA, 2023b), PFBA (U.S. EPA, 2022b), PFBS (U.S. EPA, 2021b), Gen X Chemicals (U.S. EPA, 2021a), PFOA (U.S. EPA, 2024b), and PFOS (U.S. EPA,
2024a).
4-4
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
4.2. SUMMARY OF CONCLUSIONS FOR CARCINOGENICITY
The evidence currently available to make a judgment as to whether PFHxS exposure might
affect the development of any specific cancers is scant, inconsistent, and limited to low confidence
studies. Consistent with EPA guidance (U.S. EPA. 2005) to apply a standard descriptor as part of the
hazard narrative and to express a conclusion regarding the weight of evidence for the carcinogenic
hazard potential, a descriptor of inadequate information to assess carcinogenic potential is
applied for PFHxS.
4.3. CONCLUSIONS REGARDING SUSCEPTIBLE POPULATIONS AND
LIFESTAGES
Understanding the potential areas of susceptibility to the identified human health hazards
of PFHxS can help to inform expectations of variability in responses across individuals, as well as
uncertainties and confidence in candidate toxicity values (see Section 5.2). The available human
and animal evidence indicate that early lifestages represent a susceptible population for the
adverse effects of PFHxS exposure. High confidence experimental studies report alterations in
thyroid function, including reduced serum T4 and T3, after gestational and early postnatal PFHxS
exposures in rats (see Section 3.2.1). In addition, medium confidence epidemiological studies report
that exposure to PFHxS was associated with decreased immune response after routine vaccinations
against tetanus and diphtheria vaccines in children at ages 5 and 7 (see Section 3.2.2). Although
there are considerable uncertainties in the developmental epidemiological database (e.g., potential
impact on PFHxS biomarkers due to pregnancy hemodynamics), consistent and coherent
epidemiological findings on fetal growth restriction including several medium and high confidence
developmental epidemiological studies also provide support for examination of critical in utero
exposure windows (see Section 3.2.3).
The significant difference in clearance between male and female rats (7.2 vs. 84.1 mL/kg-
day, respectively; see Section 3.1.4 for details) implies a sex-dependent susceptibility in that
species: for given dose, blood and tissue levels are predicted and were observed to be significantly
higher in male rats than female rats. While clearance levels in male and female mice were quite
similar to each other (3.9 and 3.2 mL/kg-day), the markedly lower clearance in female mice
compared with female rats predicts a strong species difference for susceptibility to developmental
effects. Clearance values for adult humans are consistently much lower than observed in either
mice or rats (0.02-0.07 mL/kg-day), which is predicted to result in a strong species difference in
susceptibility. However, only one of the human studies that directly evaluated clearance observed a
clear sex difference, with (geometric mean) urinary clearance in younger women being about 50%
higher than men and older women (Zhang etal. (2013b): see Table 3-4). Additional clearance due
to menstrual fluid loss could significantly reduce internal doses in women of childbearing age. The
rate of menstrual fluid clearance estimated by Verner and Longnecker f20151 (0.033 mL/kg-day) is
4-5
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
only slightly lower than (80% of] the geometric mean clearance for fecal and urinary elimination
(0.041 mL/kg-day), so blood levels in a 30-year-old woman might be 55% of those in a 30-year-old
man exposed to the same dose flain and Ducatman. 20221. (EPA's analysis of PFNA concentrations
from NHANES in never-pregnant women ages versus men 20-52 years of age yielded a geometric
mean ratio of 56.75%.) In addition, serial blood measurement of PFHxS in pregnant women show
that the decrease in clearance due to the lack of menstruation during pregnancy does not result in
an increase in internal dose (Oh etal.. 20221. This implies that other pharmacokinetic changes
during pregnancy mediate the decreased clearance during that time and that the clearance for
women of reproductive age (prior to pregnancy) is also appropriate for evaluating maternal
dosimetry for developmental endpoints in humans. Animal-to-human extrapolations do account for
the species- and sex-specific clearance observed among mice and rats, so in that regard PK-related
susceptibility is addressed.
lain and Ducatman (2019c) observed 32% and 40% higher geometric mean PFHxS
concentrations in individuals with glomerular function designated GF-2 and GF-3A, reflecting mild-
moderate stages of kidney disease, respectively, vs. GF-1 (p < 0.01), though those with the most
severe kidney disease had PFHxS levels indistinguishable from GF-1. The results likely reflect that
GF-2 and GF-3A levels of disease decrease renal clearance of PFHxS, which in turn will lead to
higher risk of (other) adverse effects to individuals in this group. The impact, however, is in the
range of the pharmacokinetic portion of the human inter-individual uncertainty factor, i.e., is less
than UFh,pk = 3.
Given the effects seen in the developing individuals (i.e., altered thyroid and immune
functions), prenatal and early postnatal lifestages represent a potentially sensitive population for
the effects of PFHxS exposure. No evidence was available to inform other factors that could inform
the potential for susceptibility to PFHxS exposure including demographics, genetic variability,
health status other than renal function, behaviors or practices or social determinants. The potential
impact of these other susceptibility factors remains unknown.
4-6
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
5. DERIVATION OF TOXICITY VALUES
5.1. NONCANCER AND CANCER HEALTH EFFECT CATEGORIES
CONSIDERED
The available evidence indicates that oral exposure to perfluorohexanesulfonic acid
[PFHxS] and its related salts is likely to cause adverse immune effects in humans on the basis of the
evidence presented in human studies and adverse thyroid effects on the basis of the evidence
presented in animal toxicity studies. The dose levels associated with these two identified hazards
were considered for the derivation of reference doses (RfDs) as presented below. The available
evidence suggests but is not sufficient to infer that PFHxS exposure may result in developmental,
neurodevelopmental, cardiometabolic, and hepatic effects. Given the uncertainty in these latter
conclusions, ultimately no toxicity values were derived for these health effects. A dose-response
assessment is typically not performed for health effect judgments of "evidence suggests," although
when the database contains at least one well-conducted study, quantitative analyses may still be
useful for some purposes, such as providing a sense of the magnitude and uncertainty of estimates
for health effects of concern, ranking potential hazards, informing responses in potentially
susceptible populations and lifestages, or setting research priorities fU.S. EPA. 2005. 20201. The
available evidence on PFHxS-induced developmental effects includes high confidence
epidemiological studies in which the observed outcome (low birth weight) occurs during a
susceptible lifestage and is associated with increased lifetime risk for developing a variety of
adverse health conditions such as type 2 diabetes, cardiovascular disease, neurodevelopmental
disorders, and renal disease (Tian etal.. 2019a: Reyes and Manalich. 2005: Hack etal.. 19951. The
evidence for PFHxS-induced hepatic effects was also compelling despite being categorized as
evidence suggests, with strong suggestive evidence in both humans and animals. Well-conducted
epidemiological studies report consistent associations with serum ALT, though with potential
confounding by other PFAS as a key source of uncertainty. This was selected as the most consistent
human endpoint and over animal studies due to the preference for human evidence when available
(as described elsewhere). Thus, for comparison purposes during toxicity value derivation for the
identified (likely) PFHxS hazards of immune and thyroid effects, points of departure (PODs) were
estimated for developmental (i.e., birth weight) and hepatic (i.e., serum ALT) effects (see
Section 5.2.1). No other endpoints were considered for the derivation of toxicity values.
There are no available studies to inform the potential for PFHxS to cause adverse health
effects via inhalation exposure precluding the derivation of reference concentration (RfC) (see
Section 5.2.3). Likewise, evidence pertaining to the evaluation of carcinogenicity was considered
5-1
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
inadequate to assess carcinogenic potential of PFHxS in humans, precluding the derivation of
cancer toxicity values via any exposure route (see Section 5.3).
5.2. NONCANCER TOXICITY VALUES
Noncancer toxicity values, including reference doses (RfDs) for oral exposure and reference
concentrations (RfCs) for inhalation exposure, are estimates of an exposure for a given duration to
the human population (including susceptible subgroups and/or lifestages) that are likely to be
without an appreciable risk of adverse health effects over a lifetime. The RfD derived in
Section 5.2.1 corresponds to chronic, lifetime exposure and is the primary focus of this document.
In addition, a less-than-lifetime, subchronic toxicity value (referred to as a "subchronic RfD"), which
corresponds to exposure durations ranging from a month to 10% of the life span in humans, is
derived in Section 5.2.2. Subchronic toxicity values may be useful for certain decision-making
contexts (e.g., site-specific risk assessments with less-than-lifetime exposures). Both RfD and
subchronic RfD derivations include organ-/system-specific RfDs (osRfDs) associated with health
effect-specific PODs considered for use in deriving the RfD (or subchronic RfD). As with the
subchronic RfD, osRfDs can be useful for certain decision-making contexts (e.g., cumulative risk
assessment). Subsequent decisions related to dosimetric extrapolation, application of uncertainty
factors, and confidence in toxicity values are discussed below. No information exists to inform the
potential toxicity of inhaled PFHxS or derive an RfC; this decision is discussed in Section 5.2.3.
5.2.1. Oral Reference Dose (RfD) Derivation
Study/Endpoint Selection
Data sufficient to support dose-response analyses and POD calculations for oral exposure to
PFHxS or its salts were available for both identified human health hazards: thyroid and immune
effects. As mentioned above, although a definitive health hazard was not identified, a POD was also
calculated for developmental and hepatic effects because the evidence base for PFHxS includes
well-conducted epidemiological studies albeit with some uncertainty. In addition, derivation of a
POD for developmental and hepatic outcomes was considered informative of the potential
magnitude of effects relevant to susceptible populations and lifestages and thus might inform
toxicity value derivation for thyroid or immune effects.
Rationales for study selection, details of the POD calculations, and toxicity value estimation,
as well as determination of confidence in the derived toxicity values, are detailed in this section.
The general considerations used to prioritize studies for estimating PODs for potential use in
derivation of toxicity values are described in the IRIS PFAS Protocol (see Appendix A). Well-
conducted (i.e., high or medium confidence) human studies that were deemed influential to the
hazard conclusions were prioritized for POD derivation and compared with PODs derived from
well-conducted animal studies when possible. Such human studies were available for
developmental and immunotoxicity effects.
5-2
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
A summary of endpoints and rationales considered for toxicity value derivation is presented
below.
Thyroid effects
Human studies provide conflicting evidence as to the potential effects of PFHxS on thyroid
outcomes (e.g., thyroid hormone levels). While a few studies did suggest an association between
increasing PFHxS exposure levels and decreased circulating thyroid hormones (i.e., T4) or
subclinical thyroid disease, these associations were not consistent across studies (see Section 3.2.1
for details). Overall, the available human evidence on PFHxS effects on the thyroid was considered
indeterminate, and thus these studies were not considered for use in deriving toxicity values.
The database of animal studies examining PFHxS-induced thyroid effects includes one
short-term study in rats fNTP. 2018a: Chang etal.. 20181and four multigenerational reproductive
studies in rats and mice (three studies, four publications: (Ramhaj etal.. 2018: Ramhaj etal.. 2020:
Chang etal.. 2018: Butenhoff et al.. 2009).Of these, a study in Crd:CD mice (Chang etal.. 2018) was
judged as low confidence and thus was not considered for POD derivation, leaving three high
confidence studies in SD rats fNTP. 2018a: Butenhoff etal.. 20091 or Wistar rats fRamhai etal..
2018: Ramh0i etal.. 2020).
NTP f2018al examined effects on serum concentrations of total and free T4 in adult rats,
while Ramhai etal. f 20181 evaluated effects of PFHxS on free T4 serum levels in exposed dams and
their offspring (exposed during gestation and lactation) through PND 22. NTP (2018a) observed a
statistically significant, dose-dependent decrease (p < 0.01) of free and total T4 levels starting at the
lowest experimental dose (0.625 mg/kg-day) in male rats (up to 60% in free T4 and 78% decrease
in total T4). In female rats, T4 levels were significantly decreased beginning at higher doses
(12.5 mg/kg-day and above), with 38% decrease in free T4 and 33% decreases in total T4 atthe
highest dose (50 mg/kg-day) (p < 0.01). Ramhai etal. f20181 reported similar findings to those
reported by NTP (2018a) in Wistar rat dams, with statistically significant, dose-dependent
decreases in serum-free T4 at 5 mg/kg-day and above in dams at PND 22 after exposure from GD 7
through PND 16 or 17 (Ramh0i etal.. 2018). In addition, Ramh0i etal. (2018) also reported
statistically significant (p < 0.001) decreases in free T4 in the F1 offspring born from these PFHxS-
exposed dams, with free T4 decreases at >5.0 mg/kg-day at both the end of exposure, PND16 or 17
(26%-32% decrease), and when pups were euthanized at PND 22 (26%-71% decrease). Total T4
assay measurements are more reliable that those provided by the assays available to measure free
T4 in rodents as these are insufficiently sensitive to measure the very small quantity of unbound
(i.e., "free") T4 in circulation and therefore less reliable than total T4 measurements (personal
communication with Mary Gilbert, EPA, ORD). For this reason, total, but not free, T4 was moved
forward for POD and candidate value derivation.
Two studies measured T3 in serum fRamhai etal.. 2020: NTP. 2018al. NTP f2018al
observed a statistically significant and dose-dependent decrease (p < 0.05) in serum T3 levels in
male, but not female, SD rats at >0.625 mg/kg-day (p < 0.01). Ramhai etal. f20201 analyzed
5-3
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
samples taken in Ramhaj etal. (2018) and observed a significant decrease in serum T3 in Wistar rat
dams at the highest tested dose: 19% decrease at 25 mg/kg-day (p < 0.001) measured on PND 22
after exposure from GD 7 through postnatal day 16 or 17. Overall, for TH changes, findings for both
T4 and T3 in nonpregnant adult females were relatively insensitive as compared with adult males
and thus set aside from further consideration.
Butenhoff et al. (2009) reported increased incidences of hypertrophy/hyperplasia in the
thyroid. In this 44-day exposure study, Butenhoff et al. (2009) observed increased incidences of
hypertrophy (characterized as "minimal") of thyroid follicular epithelial cells in adult male rats that
were exposed to 0.3 mg/kg-day PFHxS and an increase in "moderate" hypertrophy at the 10 mg/kg-
day PFHxS dose for up to 44 days. Hypertrophy was not observed in control animals. Decreased
thyroid hormone levels are judged relevant to human health, given the many similarities in the
production, regulation, and functioning of thyroid hormones between rodents and humans (Zoeller
and Rovet. 2004: Vansell. 2022: Stagnaro-Green and Rovet. 2016: Rovet. 2005. 2014: Navarro etal..
2014: Morreale de Escobar et al.. 2008: Hood etal.. 1999a: Hood etal.. 1999b: Hood and Klaassen.
2000: Dong etal.. 2015: Cuevas etal.. 2005: Berbel etal.. 2010). In addition, rodents are known to
be more sensitive to increases in thyroid follicular hypertrophy and hyperplasia than humans, and
thus the observed changes in thyroid hormone levels (which are not known to suffer from this
same limitation) were preferentially advanced over these histopathological changes for deriving
points of departure and the increases in thyroid hypertrophy/hyperplasia were not considered
further (see Table 5-1).
Table 5-1. Endpoints considered for dose-response modeling and derivation
of points of departure for thyroid effects in animals
Endpoint
Study reference
and confidence
Exposure route
and duration
Test strain,
species, and
sex
POD
derivation
Notes
Decreased Total
T4
NTP (2018a), high
confidence
Gavage, 28 d
Rat/SD/Male
Yes
Dose-dependent effects
were observed across
sexes, but responses
were much more
sensitive in males, even
after considering sex-
dependent PK
differences.
Ramh0i et al.
(2018), high
confidence
Exposure in
utero and
lactation
GD 7-PND 16 or
17;
measurements
taken at PND
16/17
Rat/Wistar
/F1 Combined3
Yes
Dose-dependent effects
in combined serum from
(male plus female)
offspring were
consistent across
timepoints. Responses in
dams were much less
sensitive.
5-4
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Endpoint
Study reference
and confidence
Exposure route
and duration
Test strain,
species, and
sex
POD
derivation
Notes
Exposure in
utero and
lactation
GD7-PND 16/17
measurements
taken at PND 22
Rat/Wistar
/F1 Combined3
Yes
Gavage
GD7-PND 16;
Free T4
measured at GD
15
Rat/Wistar
/P0 Female
No
Gavage
GD7-PND 16;
Free T4
measured at
PND 22
Rat/Wistar
/P0 Female
No
Decreased T3
NTP (2018a), high
confidence
Gavage, 28 d
Rat/SD/Male
Yes
Dose-dependent effects
were only observed in
male rats.
Ramh0i et al.
Gavage
GD7-PND
16/17; T3
measured at
PND 22
Rat/Wistar
/P0 Female
No
Decrease was only
observed in exposed
dams and F1 pups at the
highest dose. Responses
in dams were much less
(2020), high
confidence
In utero and
lactation
GD7-PND 16/17
measurements
taken at PND
16/17
Rat/Wistar
/F1 Combined3
Yes
sensitive.
Thyroid
histopathology
Butenhoff et al.
(2009)
44 d
Rat/SD/PO
Male
No
Concern for potential
reduced human
relevance as compared
with TH measures.
aRamh0i et al. (2018) reported as combined male and female fetal and juvenile rats; individual female pup data not
reported.
TH = thyroid hormone.
Immune effects
Consistent findings of reduced antibody responses from human epidemiological studies
provide moderate human evidence of immunosuppression with PFHxS exposure. This conclusion is
based primarily on two medium confidence studies (reported in three publications) in children
(Grandjean etal.. 2012: Grandjean et al.. 2017b: Grandjean etal.. 2017al. supported by additional
studies in children and adults (Stein etal.. 2016b: Stein etal.. 2016a: Kielsen etal.. 2016: Granum et
5-5
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
al.. 20131. Although there may be some residual uncertainty regarding the potential for
confounding by other PFAS, including PFOA and PFOS, the evidence overall supports a concern for
immunosuppression in PFHxS-exposed humans.
The two medium confidence studies of antibody response following vaccination are birth
cohorts of similar populations in the Faroe Islands (see Table 5-2) f Grandiean etal.. 2012:
Grandjean etal.. 2017b: Grandjean etal.. 2017a). Across these studies, PFHxS exposure was
measured during gestation, and at 18 months and 5, 7, and 13 years, and measures of antibody
levels were taken at 5, 7, and 13 years for both diphtheria and tetanus. Inverse associations,
indicating immunosuppression, were generally observed between PFHxS exposure and antibody
levels across different combinations of timing of exposure and outcome measures, and similar
findings were reported for other long-chain PFAS. However, there are a minority of combinations
for which positive associations (higher antibody levels with higher PFHxS exposure) were observed
(not statistically significant). This heterogeneity in results does not have a clear biologic
explanation and the relevant etiologic window of exposure for this outcome is not known, although
Grandjean etal. (2017b) noted that associations were generally weaker for two early life windows
of PFHxS when exposures were measured at 18 months (as compared with PFHxS exposures
measured prenatally or in early infancy) antibodies were measured at age 5 years, and for PFHxS
exposures measured at 5 years of age and antibodies measured at age 5 years. Still, given the
inverse associations observed for most of the exposure-outcome combinations and the low risk of
bias in these studies (sensitivity was the primary concern), they are considered appropriate
candidates for POD derivation. In Budtz-l0rgensen and Grandjean (2018). the study authors
performed benchmark dose modeling for a subset of the data presented in these papers, specifically
antibody levels at age 7 and PFHxS concentrations at age 5, and antibody levels at age 5
(prebooster) and perinatal PFHxS concentrations. The authors selected these combinations due to
the strong inverse associations and because they are reasonably representative of the study results
across exposure/outcome combinations, so after review of the BMD methods, their exposure-
response results were used to inform the benchmark dose analyses. EPA selected a different BMR in
deriving the BMDs and BMDLs (see Appendix D, Section 1 for more details).
Table 5-2. Endpoints considered for dose-response modeling and derivation
of points of departure for immune (decreased serum antibody) effects in
humans
Study
reference and
confidence
Antibody type;
measurement timing
POD derivation
Notes
Antibody
concentrations
for diphtheria
and tetanus
Grandiean et al.
(2012); Grandiean et
al. (2017a);and
Grandiean et al.
No
Effect was generally coherent with
epidemiological evidence for other antibody
effects. However, while these results contribute
to understanding the hazard for PFHxS, the
analytic models in these specific publications
5-6
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study
reference and
confidence
Antibody type;
measurement timing
POD derivation
Notes
(2017b); medium
confidence
used log-transformed exposure and log-
transformed outcome variables and such log-log
models cannot be used for BMD calculations
and thus PODs were not derived.
Budtz-
J0rgensen and
Grandiean
(2018) using
data from
Grandiean et
al. (2012);
Grandiean et
al. (2017b);
(Grandiean et
al., 2017a)
medium
confidence
Decreased serum
anti-tetanus antibody
concentration in
children at age 7 yr
and PFHxS measured
at age 5 yr
Yes
Both vaccine antibody types and the two
exposure and outcome measurement timing
combinations were generally coherent with the
broader epidemiological evidence for antibody
effects. Results were based on analytic models
using log-transformed outcome and
untransformed exposure which were suitable
for BMD calculations and POD derivations (see
Appendix D1 for more details on BMD modeling
results).
Decreased serum
anti-diphtheria
antibody
concentration in
children at age 7 yr
and PFHxS measured
at age 5 yr
Yes
Decreased serum
anti-tetanus antibody
concentration in
children at age 5 yr
and PFHxS measured
perinatally
Yes
Decreased serum
anti-diphtheria
antibody
concentration in
children at age 5 yr
and PFHxS measured
perinatally
Yes
Developmental effects
Although the human evidence on developmental effects was highly uncertain and ultimately
judged as slight (see Section 3.2.3), the database includes several well-conducted medium and high
confidence epidemiological studies reporting birth weight deficits of varying magnitude in male or
female neonates or both. A meta-analysis of the available studies showed a small but statistically
significant decrease in birth weight per each ln-unit increase in PFHxS exposure (see Section 3.2.3;
and Appendix C). However, in contrast to previous meta-analyses for PFOS and PFOA (Dzierlenga et
al. (20201 and Steenland etal. (201811. differences in detected deficits based on sample timing were
evident for early sampled studies as well as high and medium/high confidence studies combined.
Notably large effects were seen for postpartum measures, but this stratum was based on
considerably fewer studies. This suggests that studies based on postpartum samples may be most
5-7
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
prone to potential bias from pregnancy hemodynamics, but the meta-analytical data are indicative
of complex patterns of influence due to pregnancy hemodynamic that are not completely
understood. Nevertheless, the apparent influence of pregnancy hemodynamics introduces
considerable uncertainty in the interpretation of these associations of evidence of PFHxS-induced
developmental effects and was a major contributing factor in the overall evidence integration
judgment for this health effect (see Section 3.2.3). Despite these important concerns regarding
sample timing, as noted above, derivation of a POD(s) for developmental outcomes was considered
potentially informative to toxicity value derivation for thyroid or immune effects.
For developmental effects, 22 epidemiology studies evaluated associations between PFHxS
exposure and fetal growth restriction, seven of which were considered high confidence. Three of
these high confidence studies measured maternal blood levels of PFHxS in the first trimester (Sagiv
etal.. 2018: Manzano-Salgado etal.. 2017a: Buck Louis et al.. 20181. One study each sampled in the
second (Shoaff etal.. 2018) third trimester (Valvi etal.. 2017). while two studies collected samples
across multiple trimesters (Starling etal.. 2017: Bach etal.. 2016).
Five of the seven high confidence studies reported adverse associations between birth
weight and PFHxS, with no evidence of adverse associations reported in Valvi etal. f20171 or Sagiv
etal. ("20181.
Thus, the five high confidence studies considered for illustrative use in dose-response
analysis (see Table 5-3) were: Buck Louis etal. (2018): Shoaff etal. (2018). Starling etal. (2017).
Manzano-Salgado etal. (2019). and Bach etal. (2016). These studies showed consistent results
especially when re-expressed on the ln-unit scale for consistency (range: -12 to -22 g per each ln-
unit PFHxS increase).
As previously described, while no toxicity value for developmental effects will be derived
due to the high uncertainty of any such value as compared with values based on thyroid or immune
effects, the PODs for developmental effects are still useful for the purposes delineated above in
Section 5.1.
5-8
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 5-3. Mean birth weight deficit studies considered for dose-response
modeling and derivation of points of departure for developmental effects in
humans
Study reference
and confidence
Population-overall
population, sex-
specific and all
births vs. term
births only
PFHxS
biomarker
sample timing
POD
derivation
Notes
Buck Louis et al.
(2018); high
confidence
Overall population;
term births
Trimester 1
Yes
Effect size was large in magnitude; study
showed some association for other
endpoints such as birth length deficits.
Maternal samples were collected during
trimester one (range: 10-13.9 wk) which
should minimize the pregnancy
hemodynamic impact.
Manzano-Salgado
et al. (2019); high
confidence
Overall population;
all births
Trimester 1
Yes
Results based on continuous exposure
increases were moderate in magnitude and
consistent with larger birth weight deficits
based on categorical data; study showed
some coherence across other endpoints
such as postnatal growth and other fetal
growth indices. Maternal samples were
collected during trimester one
(mean = 12.3 wk) which should minimize
the pregnancy hemodynamic impact. Multi-
PFAS models were developed.
Shoaff et al. (2018);
high confidence
Overall population;
term births
Trimester 2
Yes
Effect size was moderate in magnitude;
study showed some coherence across other
endpoints such as postnatal growth.
Although the mean reported sampling
period was 18 wk, it was variable across
study participants (range: 16-40 wk) which
may make a subset of these data (i.e., those
with later sampling) more prone to
potential bias from pregnancy
hemodynamic changes.
Starling et al.
(2017); high
confidence
Overall population;
term births
Trimesters 2-3
Yes
Effect size was moderate in magnitude.
Multi-PFAS models were developed.
Median of 27 gestational wk of sampling.
Concerns regarding the influence of
pregnancy hemodynamic changes are
generally greater for any trimester three
PFHxS measures, but authors statistically
adjusted for sampling timing.
Bach et al. (2016);
high confidence
Overall population;
sex-specific; term
births
Trimester 1-2
Yesa
Results based on continuous exposure
increases were moderate in magnitude and
consistent with larger deficits based on
categorical data and across sexes; this study
5-9
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Study reference
and confidence
Population-overall
population, sex-
specific and all
births vs. term
births only
PFHxS
biomarker
sample timing
POD
derivation
Notes
also showed some coherence across other
endpoints such as head circumference.
Maternal samples were largely collected
during trimesters one and two (range: 9-20
wk; mode: 12 wk) which may minimize the
pregnancy hemodynamic impact.
Valvi et al. (2017);
high confidence
Sex-specific; all
births
Trimester 3
No
Study reported increased birth weight (i.e.,
no adverse effects).
Sagiv et al.
(2018); high
confidence
Sex-specific; term
births
Trimester 1
No
Study showed mixed results.
aStudy reported sex-specific findings that boys have larger deficits compared with girls. The associations between
exposure and birth weight were not consistent across quantiles of exposures in girls. Results based overall
population were used for POD derivation since the general population was the target population.
Hepatic effects
Although the human evidence on hepatic effects was uncertain and judged as slight (see
Section 3.2.3), the database includes several medium and high confidence epidemiological studies
reporting increases in clinical markers of liver disease. A key source of uncertainty was the
potential for confounding by other PFAS. Despite this uncertainty, a POD was derived for serum
ALT in humans to compare with PODs derived for the other, better supported health effects. From
amongst the eight (out of 10 total) medium confidence studies that reported a positive association
with ALT in adults, the medium confidence study by Kim etal. (2023) was determined to be the
preferred choice for deriving a POD for adverse liver effects. Kim etal. f20231 used directed acyclic
graphs (DAGs) to select potential confounders and all models included age, sex, education level,
household income, smoking status, BMI, heavy drinking, regular exercise. Further, this study had
the most robust approach to analyzing mixture effects and PFHxS had a positive weight in the
mixture model (though other PFNA, PFOA, and PFOS had stronger weights). Details of the modeling
for estimating the POD for PFHxS are in Appendix D.l. Selected POD results based on the preferred
hybrid approach with two cutoff values defining adversity (i.e., >30 IU/L of ALT for women and >42
IU/L of ALT for men) are shown in Table 5-5.
5-10
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 5-4. Endpoints considered for dose-response modeling and derivation
of points of departure for liver effects in humans
Endpoint
Study reference
and confidence
Population
POD
derivation
Notes
Increased
serum ALT
Kim et al.
(2023),
medium
Adults, male and
female
Yes
Study examined a sub-population of the
Korean National Environmental Health Survey
(KoNEHS) and reported significant percentage
changes in In-ALT for log2-unit increase in PFNA
of 2.8% (95% CI: -1.2, 7.0) for men and 3.7%
(95% CI: -0.2, 7.8) for women using multiple
linear regression adjusted for age, sex,
education, income, smoking, heavy drinking,
exercise, and BMI. The regression coefficients
P were calculated as 0.0276 (95% CI: -0.0121,
0.0677) per log2 (ng/mL) PFHxS for men and
0.0363 (95% CI: -0.0020, 0.0751) ln-ALT(U/L)
per log2 (ng/mL) PFHxS for women.38
Estimation or Selection of Points of Departure (PODs)
Benchmark dose modeling
Consistent with EPA's Benchmark Dose Technical Guidance Document fU.S. EPA. 20121. the
BMD and 95% lower confidence limit on the BMD (BMDL) were estimated using a BMR to
represent a minimal, biologically significant level of change. The BMD Technical Guidance fU.S. EPA.
20121 sets up a hierarchy by which benchmark responses (BMRs) are selected. The first and
preferred approach uses a biological or toxicological basis to define what minimal level of response
or change is biologically significant. In the absence of information regarding the level of change that
is considered biologically significant, a BMR of 1 SD from the control mean for continuous data or a
BMR of 10% extra risk for dichotomous data is used to estimate the BMD and BMDL. The BMRs
selected for dose-response modeling of PFHxS-induced health effects are listed in Table 5-4 along
with the rationale for their selection. Further details, including the modeling output and graphical
results for the model selected for each endpoint, can be found in Appendix D. When dose-response
modeling was not feasible, or adequate modeling results were not obtained, no-observed-adverse-
effect level (NOAEL) or lowest observed adverse effect level (LOAEL) values were identified and
used as the POD.
38Percentage increase = (eP-l)*100) see Kim et al. (20231.
5-11
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 5-5. Benchmark response levels selected for BMD modeling of PFHxS
outcomes
Endpoint
BMR
Rationale
Thyroid effects
Decreased serum-total T4
1 standard deviation
No information is readily available that
allows for determining a minimally
biological significant response. The BMD
Technical Guidance (U.S. EPA, 2012)
recommends a BMR based on 1SD for
continuous endpoints when biological
information is not sufficient to identify the
BMR.
Decreased serum-total T3
Immune effects
Decreased antibody concentrations
for diphtheria and tetanus in
children
Yi standard deviation
No information is readily available that
allows for determining a minimally
biological significant response. The BMD
Technical Guidance (U.S. EPA, 2012)
recommends a BMR based on 1 SD for
continuous endpoints when biological
information is not sufficient to identify the
BMR. Diphtheria and tetanus are serious
and sometimes fatal infections. In addition,
childhood represents a sensitive lifestage
when immunosuppression during the
developmental stage may impede
children's ability to protect against a range
of immune hazards. Given the potential
severity of this outcome, a BMR ofA SD
was selected (see additional discussion in
Appendix D, Section 1.1).
Developmental effects
Decreased birth weight in humans
5% extra risk of exceeding adversity
cutoff (hybrid approach3)
A 5% extra risk is commonly used for
dichotomous developmental endpoints as
recommended by Benchmark Dose
Technical Guidance (U.S. EPA, 2012). For
birth weight, a public health definition of
low birth weight exists, and the hybrid
approach was used to estimate the dose at
which the extra risk of falling below that
cutoff equaled 5% (see Appendix D).
Hepatic effects
Increased serum ALT in humans
10% extra risk exceeding adversity
cutoff (hybrid approach3)
Both extra risks of 5% and 10% were
considered. A BMR of less than 10% can be
supported for severe or debilitating health
outcomes. Given the findings of
associations between increased ALT and
5-12
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Endpoint
BMR
Rationale
severe liver disease (Park et al., 2019), a
BMR of 5% was considered. However,
modest elevations in ALT are more likely
associated with milder forms of liver injury,
including steatosis and NAFLD (Oh et al.,
2017). Because of uncertainties in
measuring ALT, in selecting the most
appropriate upper limit of normal (and the
difficulty in interpreting specific elevations
above the upper limit of normal as
adverse), and in selecting the reference
population, a BMR of 10% extra risk was
selected as a "minimally adverse" effect
and as a standard reporting level per EPA's
Benchmark Dose Technical Guidance (U.S.
EPA, 2012). See also Appendix D Section
D.1.2.
aThe hybrid approach to defining the continuous BMR retains the full power of modeling continuous data (i.e.,
retains information on the distribution of continuous responses instead of dichotomizing the response variable)
and incorporates biological, toxicological, or clinical knowledge in setting the adversity cutoff (U.S. EPA, 2012,
2023a)
When modeling was feasible, the estimated BMDLs were used as PODs (see Table 5-6).
Further details, including the modeling output and graphical results for the model selected for each
endpoint, can be found in Appendix D. For the modeling of immune effects, potential confounding
by other PFOS and PFOA was considered in the POD derivation by comparing the effect estimates
from the analyses in and BMDLs for PFHxS from single-PFAS models against those from multi-PFAS
models controlling for PFOS and PFOA in analyses by Budtz-largensen and Grandjean (2018) (see
Appendix D, Section 1 for details). When dose-response modeling was not feasible, or adequate
modeling results were not obtained, NOAEL or LOAEL values were identified based on biological
rationales when possible and used as the POD. The PODs (based on BMD modeling or
NOAEL/LOAEL selection) for the endpoints advanced for dose-response analysis are presented in
Table 5-6 alongside the corresponding PODhedS derived based on the PK extrapolations as
described in Section 3.1.6.
5-13
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Table 5-6. Points of departure (PODs) considered for the derivation of PFHxS
candidate toxicity values
Endpoint
Study/confidence
Species/
sex
POD type
(% change if
NOAELor
LOAEL)
Free acid
POD
(mg/kg-d)f
PODjnternal
internal dose
(mg/L) or
DDEF (no
units)c
Free acid
PODHEDd
(mg/kg-d)
Thyroid
Decreased Total
T4
28-d study
NTP (2018a), hiah
confidence
SD rat, male
LOAELa (-44%)
0.684
PODinternal-
34.58
1.42 x 10"3
Multigenerational
Study
Ramh0i et al.
(2018), hiah
confidence
Wistar rat,
Combined Fi
(PND 16/17)
NOAELb (+4%)
0.051
4.81 x 10"4
2.45 x 10"5
Decreased T3
Multigenerational
Study
Ramh0i et al.
(2020), hiah
confidence
Wistar rat,
Combined Fi
(PND 16/17)
NOAELb(-7%)
5.5
4.81 x 10"4
2.65 x 10"3
28-d study
NTP (2018a), hiah
confidence
SD rat, male
LOAEL3 (-22%)
0.684
PODinternal.
34.58 (mg/L)
1.42 x 10"3
Immune (developmental)
Decreased
serum anti-
tetanus antibody
concentration in
children at age 7
and PFHxS cone
measured at
age 5
Grandiean et al.
(2012); Budtz-
J0rgensen and
Grandiean (2018),
medium
confidence
Human
(children)/both
BMDL^sd
_e
2.82 x 10"4
1.16 x 10"8
Decreased
serum anti-
diphtheria
antibody
concentration in
children at age 7
and PFHxS cone
measured at
age 5
Grandiean et al.
(2012); Budtz-
J0rgensen and
Grandiean (2018),
medium
confidence
Human
(children)/both
BMDL^sd
_e
3.00 x 10"4
1.23 x 10"8
Decreased
serum anti-
tetanus antibody
concentration in
children at age 5
and PFHxS cone
Grandiean et al.
(2012); Budtz-
J0rgensen and
Grandiean (2018),
medium
confidence
Human
(children)/both
BMDL^sd
_e
1.44 x 10"2
5.90 x 10"7
5-14
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
measured
perinatally
Decreased
serum anti-
diphtheria
antibody
concentration in
children at age 5
and PFHxS cone
measured
perinatally
Grandiean et al.
(2012); Budtz-
J0rgensen and
Grandiean (2018),
medium
confidence
Human
(children)/both
BMDL^sd
_e
1.37 X 10"2
5.62 x 10"7
Developmental6
Decreased birth
weight
Bach et al. (2016),
high confidence
Human
(newborn)/Both
BMDL5ER,
Hybrid
_e
1.12 x 10"3
8.06 x 10"8
Buck Louis et al.
(2018), hiah
confidence
Human
(newborn)/Both
BMDL5ER,
Hybrid
_e
1.71 x 10"3
1.23 x 10"7
Manzano-Salgado
et al. (2019), hiph
confidence
Human
(newborn)/Both
BMDL5Er ,
Hybrid
_e
1.33 x 10"3
9.58 x 10"8
Hepatic
Increased ALT
representing
increased risk of
liver effectsh
Kim et al. (2023),
Medium
Confidence
Human, female
BMDLEr5,
Hybrid
with cutoff of
30 lU/Lof ALT
for women
_e
2.22 x 10"3
9.10 x 10"8
Kim et al. (2023),
Medium
Confidence
Human, female
BMDLerio,
Hybrid
with cutoff of
30 lU/Lof ALT
for women
_e
4 ,18x 10"3
1.71 x 10"7
Kim et al. (2023),
Medium
Confidence
Human, male
BMDLers,
Hybrid
with cutoff of
42 lU/Lof ALT
for men
_e
3.14 x 10"3
1.29 x 10"7
Kim et al. (2023),
Medium
Confidence
Human, male
BMDLerio,
Hybrid
with cutoff of
42 lU/Lof ALT
for men
_e
6.58 x 10"3
2.70 x 10"7
aNo models provided adequate fit; therefore, a freestanding LOAEL, no NOAELwas identified as there were
statistically significant effects in the lowest dose.
bNo models provided adequate fit; therefore, NOAEL approach was used.
Tor thyroid effects in male rats from the NTP bioassay, the internal dose, PODintemai, was calculated by
interpolation of the measured end-of-study concentrations in the rats, as described in Section 3.1.7, and
PODHed= POD x CLh, where CLH is the estimated human clearance for the general population from Table 3-5. For
thyroid effects from the multigenerational study of Ramh0i et al. (2018), PODHed= POD x DDEF, where the DDEF
5-15
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
corresponding to the rat sex for the observation is taken from Table 3-7; the lower DDEF for female rats used for
observations in combined sex groups.
dFor immune and developmental effects observed at PND 16/17 in rats or associated with serum concentrations
measured in children at age 5 PODHEDwas calculated assuming steady-state serum concentrations using CLfor
human males and older women, since the endpoint is assumed to depend on serum concentrations in the
offspring, for which the lower clearance (not including the factor for menstrual-associated clearance) is relevant.
For effects observed at birth or associated with perinatal maternal serum concentrations, CL for humans included
the factor for menstrual-associated clearance, since maternal serum concentrations throughout pregnancy are
similar to or below prepregnancy concentrations, which result from the total clearance of the reproductive age
woman.
eBMD modeling was done on serum concentrations and hence there was no POD based on external dose.
fPOD for PFHxS free acid were calculated by taking the LOAEL or NOAEL and multiplying by the ratio of potassium
salt/ molecular weight of the free acid.
gAlthough PODs were derived for five birth weight studies (see above), there was less uncertainty in three
developmental epidemiological studies noted here with earlier maternal biomarker sampling (Manzano-Salgado
et al., 2019; Buck Louis et al., 2018; Bach et al., 2016).
hCutoffs for adversity based on ALT concentrations of 30 IU/L for women and 42 IU/L for men are discussed in
Appendix D.l.
Derivation of Candidate Lifetime Toxicity Values for the Reference Dose (RfD)
As discussed, below the developmental period is recognized as a susceptible lifestage when
exposure during a critical time window is more relevant to the induction of adverse effects than
lifetime exposure. Thus, the derivation of a lifetime value for developmental thyroid and immune
endpoints following PFHxS exposure is supported. Exposure during pregnancy was also considered
a potentially susceptible lifestage. Consistent with EPA guidelines fU.S. EPA. 19941. the thyroid
hormone PODs following 28-day PFHxS exposure in adult SD rats were not considered for
derivation of candidate lifetime values given the high degree of uncertainty associated with using
PODs from a 28-day rodent study to protect against effects observed in a chronic setting. However,
these endpoints were considered for the derivation of the subchronic RfD (see Section 5.2.2).
Overall, the developmental immune endpoints from epidemiological studies and thyroid endpoints,
specifically decreases in T3 and total T4, from a multigenerational rodent study of PFHxS, were
preferentially advanced for the derivation of candidate lifetime values.
For developmental immune effects, the PODintemai was 0.282 ng/mL (see Appendix
D.l). Grandjean and Bateson (2021) provides the minimum measured PFHxS concentration (0.02
ng/mL), the 5th% (0.2 ng/mL) and the 10th% (0.3 ng/mL) so the BMDL is between the 5th and 10th%
of the observed measurements. The estimated limit of detection of PFHxS was 0.007 ng/mL and
the limit of quantification was 0.02 ng/mL in serum (Haug etal.. 20091. PODhed values were derived
for decreased serum antibody levels (for both diphtheria and tetanus) in children (male and
female) at different timing of exposure and outcome measurement combinations, specifically
antibody levels at age 7 and PFHxS concentrations at age 5, and antibody levels at age 5 and
perinatal PFHxS concentrations (Budtz-largensen and Grandjean. 2018) (see Table 5-5). The
BMDLy2sD(HED) of 1.16 x 10"8 mg/kg-day for decreased serum anti-tetanus antibody concentrations at
age 7 and PFHxS measured at age 5 is selected for the derivation of osRfDs for immune effects.
5-16
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Confidence in the BMDL estimate was highest (medium confidence) for this endpoint in comparison
with other exposure-outcome combinations evaluated by Grandiean etal. f20121 and Budtz-
largensen and Grandiean f20181 based on a better fit model for PFHxS in the single-PFAS model
and less uncertainty with respect to potential confounding with other co-occurring PFAS (i.e., PFOS
and PFOA) (see Appendix D, Sectionl.l for more details). The BMDL^sdched) of 1.23 x 10~8 mg/kg-
day for decreased serum anti-diphtheria antibody concentrations at age 7 and PFHxS measured at
age 5 is also selected for the derivation of osRfDs for immune effects. Confidence in this BMDL
estimate was somewhat lower [medium/low confidence) for this endpoint than for anti-tetanus
antibody concentrations at age 7 (see Appendix D, Sectionl.l for more details). Further, although
both tetanus and diphtheria are rare in the United States, tetanus remains more of a concern
primarily among older adults, who are unvaccinated or inadequately vaccinated and therefore are
at higher risk of disease and mortality fLiang etal.. 20181. The estimated BMDL^sd
(2.82 x 10"4 mg/L) for this endpoint in the single-PFAS model is at about the 10th percentile of the
observed distribution. No information was available to judge the fit of the model in the range of the
BMDLs, but the BMD and BMDL were both within the range of observed values and the model fit
PFHxS well (see Appendix D, Section 1.1 for more details). The fact that the derived PODhed for
immune effects on both tetanus and diphtheria antibody concentrations at the same ages are
relatively close (1.16 x 10~8 mg/kg-day versus 1.23 x 10"8 mg/kg-day) lends support to the choice of
the PODhed of 1.16 x 10~8 mg/kg-day for decreased serum anti-tetanus antibody concentrations at
age 7 and PFHxS measured at age 5 for the derivation of the osRfD.
For thyroid osRfD, PODhed values were derived for decreased total thyroxine (T4) as well as
decreased triiodothyronine (T3) in a multigenerational reproductive study, with exposure
including all of gestation fRamhai etal.. 2018: Ramhai etal.. 20201 and a 28-day comprehensive
toxicity study in rats fNTP. 2018al (see Table 5-6). The PODhed of 2.14 x 10"5 for decreased total T4
in combined Fi Wistar rats is selected for the derivation of osRfD for thyroid effects as it was the
most sensitive and reliable measure of thyroid hormone function (see Table 5-6). As described
previously, although candidate toxicity values were not derived for developmental or hepatic
effects (decreased birth weight and increased ALT, respectively), PODs for these outcomes were
derived as they were considered informative of the magnitude of effects relevant to susceptible
lifestages and populations and may help inform uncertainty factor selection for developmental
immune effects and thyroid effects. The lowest PODs derived for developmental and hepatic effects
are 7- and 8-fold higher than the lowest POD derived for developmental immune effects. Thus,
derivation of an RfD based on the developmental immune effects would be protective of
developmental (decreased birth weight) and hepatic (increased ALT) effects.
Under EPA's A Review of the Reference Dose and Reference Concentration Processes (U.S. EPA.
20021 and Methods for Derivation of Inhalation Reference Concentrations and Application of
Inhalation Dosimetry fU.S. EPA. 19941. five possible areas of uncertainty and variability were
considered in deriving the candidate values for PFHxS. An explanation of these five possible areas
5-17
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
of uncertainty and variability and the values assigned to each as a designated uncertainty factor
(UF) to be applied to the candidate PODhed values are listed in Table 5-7, below.
Table 5-7. Uncertainty factors for the development of the lifetime RfD for
PFHxS
Value
Justification
UFa
1
A UFa of 1 is applied to the POD derived from developmental immune effects as these
responses were observed in epidemiological studies.
3
For thyroid effects, a UFA of 3 is applied to account for uncertainty in characterizing the
pharmacokinetic and pharmacodynamic differences between mice or rats and humans
following oral PFHxS exposure. Some aspects of the cross-species extrapolation of
pharmacokinetic processes have been accounted for using a DDEF to convert external doses
from rodents to administered doses in humans; however, residual uncertainty related to
potential pharmacodynamic differences remains.
UFh
10
A UFh of 10 is applied for developmental immune and thyroid effects. This is to account for
interindividual variability in humans in the absence of quantitative information on potential
differences in pharmacokinetics and pharmacodynamics relating to PFHxS exposure in
humans. (See discussion below for additional details).
UFS
1
A UFs of 1 is applied to reduced antibodv responses in children (Grandiean et al., 2012;
Budtz-J0rgensen and Grandiean, 2018). The developmental period is recognized as a
susceptible lifestage when exposure during a critical window of development is more
relevant than lifetime exposure in adulthood (U.S. EPA, 1991). Additional considerations for
the UFS for immune effects are discussed below.
1
A UFS of 1 is applied to thyroid effects observed in the F1 animals from reproductive study
(Ramh0i et al., 2018); the developmental period is a susceptible lifestage where exposure
during certain time windows (e.g., pregnancy and gestation) is more relevant to the induction
of developmental effects than lifetime exposure (U.S. EPA, 1991).
ufl
1
A UFl of 1 is applied for LOAEL-to-NOAEL extrapolation when the POD is a BMDL as is the
case for developmental immune endpoint or POD is a NOAEL as is the case for the thyroid
endpoint.
UFd
3
A UFD of 3 is applied to account for deficiencies and uncertainties in the database. Although
limited, the evidence base in laboratory animals consists of high/medium confidence short-
term studies in rodents and a high confidence developmental study in mice. The database for
PFHxS also includes several high/medium confidence epidemiological studies most
informative for immune and developmental effects, which are sensitive effects of PFHxS
exposure. However, uncertainties remain regarding the lack of studies examining effects with
long-term exposure in adults—including in women of reproductive age (which may have
increased susceptibility), studies of potential multigenerational effects, and studies of
postnatal development, neurotoxicity, and thyroid toxicity during developmental lifestages.
In all, the data are too sparse to conclude with certainty that the quantified developmental
effects are likely to be the most sensitive; thus, a UFD of 1 was not selected. However, a UFD
of 10 was also not selected given the availability of data from well-conducted studies on a
range of health outcomes in multiple species, including sensitive evaluations of
developmental and immune endpoints in humans. See discussion below for additional
details.
5-18
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Value
Justification
UFC
See Table
5-8
Composite Uncertainty Factor = UFA x UFH x UFS x UFL x UFD
As described in EPA's A Review of the Reference Dose and Reference Concentration Processes
(U.S. EPA. 2002). the interspecies uncertainty factor (UFa) is applied to account for extrapolation of
animal data to humans, and accounts for uncertainty regarding the pharmacokinetic and
pharmacodynamic differences across species. As is usual in the application of this uncertainty
factor, the pharmacokinetic uncertainty is mostly accounted for through the application of
dosimetric approaches for estimation of HEDs. This leaves some residual uncertainty around the
pharmacokinetics and the uncertainty surrounding pharmacodynamics. For developmental
immune effects, a UFa = 1 was applied to the POD as these responses were observed in
epidemiological studies. For thyroid effects, a UFa = 3 was applied to the POD derived from rodent
studies to account for interspecies uncertainty. While uncertainty in the pharmacokinetic processes
has largely been accounted for by using a DDEF to convert external rodent doses to human
administered doses, a UFa = 3 was applied to address the remaining pharmacokinetic uncertainty
and to address the pharmacodynamic uncertainty in extrapolating those effects to humans (see
Uncertainty in HED Calculations for more details.).
For developmental immune effects in children, a UFh of either 3 or 10 was considered.
Specifically, it can be argued that the PODs are derived from susceptible individuals because
children's immune systems are not fully formed and are presumably more sensitive to these effects
than most other populations, and thus, the UFH should be reduced (although uncertainty regarding
differences across individuals exposed during this sensitive lifestage would still remain). However,
a counter argument is that currently there are no data to compare the responses in children with
other populations or lifestages, so it is unclear whether these individuals are indeed particularly
susceptible to these specific effects. As described in U.S. EPA (2020). other factors, in addition to
lifestage, may increase susceptibility, including: demographics, genetic variability, health status,
behavior or practices, and social determinants. Ultimately, because the current evidence is
insufficient to address these uncertainties, a UFh of 10 is applied for developmental immune effects.
For thyroid effects, a UFh of 10 is applied to address differences due to intraspecies variability,
including potentially more sensitive or severe effects in susceptible populations or lifestages.
The duration extrapolation factor (UFS) accounts for the uncertainty in extrapolating from
less than chronic PFHxS exposure to lifetime exposure. A UFS = 1 was applied to the PODs for
thyroid effects as the selected POD was derived from a reproductive study with exposure
encompassing the critical window of gestation (Ramhaj etal.. 2018). This developmental window is
recognized as a susceptible lifestage when exposure is more relevant to the induction of
developmental effects than lifetime exposure (U.S. EPA. 1991). The reduced antibody responses
were measured in children 5-7 years of age, which also constitutes a sensitive lifestage. However,
given the slow clearance rates for this chemical, particularly in humans (see Table 3-5), PFHxS is
5-19
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
expected to accumulate in the body through adulthood. Therefore, it is plausible that longer
exposure durations can result in effects at lower exposure levels. Although the MOA for PFHxS-
induced immunosuppressive responses in humans is unknown, early-life exposures may alter the
immune system and lead to unpredictable outcomes later in life or during other susceptible
lifestages of reduced immunocompetence such as pregnancy, advanced lifestages, or
immunocompromised states (IPCS. 2012) that show increased sensitivity with continuous, longer-
term exposures. Still, given the expectation that the children and their mothers have been exposed
to elevated levels of PFHxS for many years, the observed effects on immune response are
considered the result of a cumulative, prolonged PFHxS exposure to the subjects from conception
until the age when the response was evaluated. Further, the consequences of perturbed immune
system function (in this case, suppressed antibody responses leading to potentially increased risk
of disease) during development are expected to be generally more severe and longer lasting that
those that manifest in healthy adults. Thus, a UFS of 1 was considered appropriate.
The database uncertainty factor (UFD) is applied to account for the potential of deriving an
under-protective reference value as a result of incomplete characterization of a chemical's toxicity
fU.S. EPA. 20021. For PFHxS, a UFd of 3 was selected to account for deficiencies and uncertainties in
the database. Although limited, the evidence base in laboratory animals consists of high/medium
confidence short-term studies in rodents and a high confidence developmental study in mice. The
database for PFHxS also includes several high/medium confidence epidemiological studies most
informative for immune and developmental effects, which are sensitive effects of PFHxS exposure.
However, uncertainties remain regarding the lack of studies examining effects with long-term
exposure in adults—including in women of reproductive age (which may have increased
susceptibility), studies of potential multigenerational effects, and studies of postnatal development,
neurotoxicity, and thyroid toxicity during developmental lifestages. 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. While the PFHxS
database does include high confidence reproductive/developmental toxicity studies in rats and
mice, these only span one-generation. Therefore, despite their quality, these studies fail to cover
potential transgenerational impacts of longer-term exposures evaluated in two-generation studies.
The availability of a two-generation multigenerational reproductive study could result in reference
values below those currently derived for PFHxS. However, the concern over a lack of two-
generation study in the available literature is diminished when the PFHxS, PFDA, PFOA, and PFOS
evidence bases are considered together. Although limited in their ability to assess reproductive
health or function, measures of possible reproductive toxicity occurred at doses equal to or higher
than those that resulted in effects in other organ systems (e.g., thyroid, liver) when measured after
exposure to PFHxS in utero through PND 22 (Ramh0i etal.. 2018). Similar results were observed for
the animal databases for PFOA and PFOS indicating reproductive effects were not uniquely
sensitive markers of toxicity for these long-chain PFAS fATSDR. 20211. Further, no notable male or
5-20
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
female reproductive effects were observed in epidemiological or toxicological studies investigating
exposure to PFHxS fMDH. 20191. Given these overall uncertainties with the database, a threefold UF
was applied.
The uncertainty factors described in Table 5-7 and the text above were applied and the
resulting candidate values are shown in Table 5-8. The candidate values are derived by dividing the
PODhed by the composite uncertainty factor:
Candidate values for PFHxS = PODhed + UFc. (3-5)
Table 5-8. Lifetime candidate values for PFHxS
Endpoint
Study/
confidence
Strain/
species/sex
Free acid
PODhed
(mg/kg-d)
UFa
UFh
UFS
ufl
UFd
UFC
Candidate
value
(mg/kg-d)
Thyroid
Decreased Total
T4
Ramh0i et al. (2018),
high confidence
Wistar rat, combined
Fi
Wistar rat,
Combined Fi
(PND 16/17)
2.45 x 10"5
3
10
1
1
3
100
2 x 10"7
Decreased T3
Multigenerational
Study
Ramh0i et al. (2020),
high confidence
Wistar rat,
Combined Fi
(PND 16/17)
2.65 x 10"3
3
10
1
1
3
100
3 x 10"5
Developmental immune effects
Decreased serum
anti-tetanus
antibody
concentration in
children at age 7
Grandiean et al.
(2012); Budtz-
J0rgensen and
Grandiean (2018);
medium confidence
Human
(children),
male and
female
1.16 x 10"8
1
10
1
1
3
30
4 x 10 10
Decreased serum
anti-diphtheria
antibody
concentration in
children at age 7
Grandiean et al.
(2012); Budtz-
J0rgensen and
Grandiean (2018);
medium confidence
Human
(children),
male and
female
1.23 x 10"8
1
10
1
1
3
30
4 x 10 10
Selection of Lifetime Toxicity Value(s)
Selection of organ- /system-specific oral reference doses fosRfPs)
Table 5-8 shows osRfDs selected for the individual organ systems identified in Section 3.2
(i.e., thyroid and developmental immune effects).
5-21
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
The value of 4 x 10"10 mg/kg-day (rounded from 3.9 x 10"10 and, separately,
4.1 x 10"10 mg/kg-day in Table 5-8 for decreased serum anti-tetanus and anti-diphtheria antibody
concentrations in children (male and female) at age 7 years and PFHxS measured at age 5 years
from the Grandiean etal. T20121 and Budtz-largensen and Grandiean f20181 was selected as the
osRfD for developmental immune effects. The respective PODhed values for these two endpoints
(decreased anti-tetanus as well as decreased anti-diphtheria antibodies) were close in value
(1.16 x 10"8 versus 1.23 x 10~8, respectively) and the candidate values round to the same toxicity
value.
For the thyroid effects, an osRfD of 2 x 10~7 mg/kg-day (rounded from 2.45 x 10~7 in
Table 5-8) was selected based on decreased total T4 in F1 pups exposed to PFHxS in the Ramhai et
al. f20181. As there was no other reason to select one POD over the other (e.g., different levels of
confidence in the POD calculations), the more sensitive POD for total T4 was selected over the POD
for T3.
The confidence decisions about the study, evidence base, quantification of the POD, and
overall RfD for these organ-/system-specific values are described in detail in Table 5-9, along with
the rationales for selection of confidence levels. In deciding overall confidence, confidence in the
evidence base is prioritized over the other confidence decisions. The overall confidence in the
osRfDs for both immune and thyroid effects is judged as medium. Selection of the overall RfD is
described in the following section.
Table 5-9. Confidence in the organ-/system-specific RfDs for PFHxS
Confidence
categories
Designation
Discussion
Thyroid 2 x 10~7 RfD = mg/kg-d
Confidence in
study3 used to
derive osRfD
High
Confidence in Ramh0i et al. (2018) was hiah and is based on a well-designed
experimental design using established approaches, recommendations, and best
practices (HAWC link).
Confidence in
evidence base
supporting this
hazard
Medium
Confidence in the evidence base for thyroid effects is medium based on
consistent findings in animals of decreases in T3 and T4 in adult and juvenile rats
in the absence of effects on TSH (Ramh0i et al., 2018; NTP, 2018a), but with
unexplained inconsistency in the available epidemiological studies and other
uncertainties (see Table 3-8).
Confidence in
quantification of
the PODhed
Medium
Confidence in the quantification of the PODhed and osRfD is medium given POD
was based on a NOAEL (data did not fit BMD models) and because a DDEF was
applied to estimate the PODhed- The uncertainty associated with the use of a
DDEF is less than the uncertainty introduced from the use of a NOAEL because
the DDEF is based on PFHxS-specific pharmacokinetic data (see Uncertainty in
HED Calculations). Considering these limitations, confidence in the POD was
medium.
Overall
confidence in
osRfD
Medium
The overall confidence in the osRfD is medium. The medium confidence in the
POD derivation is offset by the high confidence in the study and medium
confidence in the evidence base for thyroid effects.
5-22
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Confidence
categories
Designation
Discussion
Developmental Immune RfD = 4 x 1010
Confidence in
study3 used to
derive osRfD
Medium
Confidence in Grandiean et al. (2012); Budtz-J0rgensen and Grandiean (2018)
was rated as medium based on some concerns for sensitivity from narrow
exposure contrast, which decreases confidence in null associations onlv (HAWC
link).
Confidence in
evidence base
supporting this
hazard
Medium
Confidence in the evidence base for immune effects is medium based on
consistent findings of reduced antibody responses from two medium confidence
birth cohort studies (Grandiean et al., 2012; Grandiean et al., 2017b; Grandiean
et al., 2017a) and a low confidence study in adults (Grandiean et al., 2017b).
Limitations in this evidence base include the lack of epidemiological studies in
adults or long-term/chronic studies in animals, and a general lack of studies
examining effects on the immune system across different developmental
immunotoxicity categories, including sensitization and allergic response and
autoimmunity and autoimmune disease.
Confidence in
quantification of
the PODhed
Medium
The POD is based on BMD modeling within the range of the observed data and a
BMDLyiSD estimate that is associated with little uncertainty due to potential
confounding by PFOA or PFOS (see Appendix D, Section 1.1 for more details). The
PODhedS for decreased anti-tetanus and decreased anti-diphtheria antibodies
were close in value (1.16 x 10~8 vs. 1.23 x 10~8, respectively) which increases
confidence in the quantification of the PODhed- There is uncertainty as to the
most sensitive window of vulnerability with respect to the exposure/outcome
measurement timing (BMDs/BMDLs were estimated from PFHxS levels measured
at age 5 or perinatally and anti-tetanus antibody concentrations measured at
age 7 or 5) and the effect on antibodies at age 7 were more sensitive that those
measured at age 5 (see Appendix D, Section 1.1 for more details); however,
Grandiean et al. (2017b) reported that estimated PFOS and PFOA "concentrations
at 3 m and 6 m showed the strongest inverse associations with antibody
concentrations at age 5 yr, particularly for tetanus." Thus, it is possible that
adverse effects of PFHxS during infancy could be more sensitive than between
ages 5 and 7 yr.
Overall
confidence in
osRfD
Medium
The overall confidence in the osRfD is medium and is driven by medium
confidence in the evidence base for immune effects, the quantification of the
POD, and the study used for BMD modeling.
aAII study evaluation details can be found on HAWC.
5-23
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Selection of overall reference dose (RfD) and confidence statement
Table 5-10. RfD and organ-/system-specific RfDs for PFHxS
Reference dose (RfD
Basis
RfD (mg/kg-d)
Confidence
Immune (developmental)
effects
4 x 10 10
Medium
Organ-/system-specific RfDs
osRfDs)
Organ/system
Outcomes and studies
PODhed (mg/kg-d)
UFC
osRfD (mg/kg-d)a
Confidence
Thyroid
Decreased serum-total T4
in F1 Wistar rats
(Ramh0i et al., 2018)
2.45 x 10"5
100
2 x 10"7
Medium
Immune
(developmental)
Decreased serum anti-
tetanus and anti-
diphtheria antibody
concentrations measured
in children at age 7 with
PFHxS exposure
measured at age 5
Grandiean et al. (2012);
Budtz-J0rgensen and
Grandiean (2018);
Grandiean et al. (2012);
Budtz-J0rgensen and
Grandiean (2018)
1.16 xlO"9 and
1.23 x 10"9
30
4 x 10 10
Medium
aThe RfD or osRfD values for different salts of PFHxS would be calculated by multiplying the RfD or osRfD values for
the free acid of PFHxS (i.e.,, the toxicity values in the table above) by the ratio of molecular weights. For example,
_ ^ . . . ... MW apot as slum salt 438 . AAr- .
for the potassium salt the ratio would be: = — = 1.095. This same method of conversion
MW free acid 400
can be applied to other salts of PFHxS, such as the ammonium or sodium salts, using the corresponding molecular
weights.
From the identified human health effects of PFHxS and derived osRfDs for thyroid and
developmental immune effects (see Table 5-10), an RfD of 4 x 10"10 mg/kg-day was selected based
on decreased serum anti-tetanus and anti-diphtheria antibody concentrations in children. As
described in Table 5-9, confidence in the RfD is medium, based on medium confidence in the
developmental immune osRfD. This osRfD is based on the two lowest PODhedS available on PFHxS
immune effects (an evidence based interpreted with medium confidence) using a study considered
medium confidence. The selected osRfD is based on effects in children and expected to be protective
across all lifestages. The selection considered both available osRfDs as well as the overall
confidence and composite uncertainty for those osRfDs. The thyroid osRfD was based on
application of a composite uncertainty threefold greater than that applied in deriving the immune
osRfD. Further, when comparing the sensitivity of thyroid and immune osRfDs, the thyroid value is
over 3,000-fold higher. Had the osRfD for thyroid effects been chosen as the overall RfD, this would
have raised concerns over the ability of the thyroid RfD to be protective against potential immune
5-24
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
effects (and it may not be protective against other developmental effects, such as decreased birth
weight (see Table 5-7) if those other effects could be reliably quantified). Selection of the RfD on the
basis of developmental immune effects is presumed to be protective of possible thyroid and other
potential adverse health effects (including potential effects on birth weight and hepatotoxicity
[increased ALT levels]) in humans. Finally because the developmental immune osRfD is based on
effects observed in males and females, the overall RfD would be protective for both sexes.
5.2.2. Subchronic Toxicity Values for Oral Exposure (Subchronic Oral Reference Dose [RfD])
Derivation
In addition to providing an RfD for lifetime exposure in health systems, this document also
provides an RfD for less-than-lifetime ("subchronic") exposures. These candidate subchronic
toxicity values were based on the endpoints and PODs in Table 5-6 including the shorter duration
studies that were not advanced for consideration in developing the lifetime RfD. Given that the
immune and thyroid effects considered for the RfD were observed after exposure to PFHxS during
susceptible lifestages, these endpoints were also considered for the derivation of candidate
subchronic toxicity values, applying identical uncertainty factors to those used for the lifetime RfDs
(see Table 5-7).
The datasets advanced for derivation of the subchronic toxicity values were selected on the
basis of several considerations, including whether there is an endpoint with less uncertainty and/or
greater sensitivity, and whether the endpoint is protective of both sexes and all lifestages.
Ultimately, similar to the datasets advanced for the lifetime thyroid osRfD derivation, decreased
total T4 and decreased T3 endpoints from the Ramh0i etal. (2018) study was advanced over
identical endpoints from the high confidence NTP f2018al study. This is because the Ramhai et al.
f20181 study included exposure to PFHxS during gestation, this exposure is interpreted as a critical
sensitive window for effects on the developing thyroid system. Further, consistent with the decision
when estimating the lifetime osRfD, the POD for total T4 was advanced over the POD for T3 from
Ramh0i etal. (2018) given the increased sensitivity of the POD. The NOAELhed of 1.92 x 10"5 mg/kg-
day for decreased total T4 in F1 generation rats in the Ramh0i etal. (2018) study was selected for
the thyroid subchronic osRfD (see Table 5-6). The UFs applied to the derivation of a subchronic RfD
thyroid POD in rat offspring are the same as those applied in the derivation of lifetime RfD values.
See Table 5-7 for details.
Likewise, the same datasets on developmental immune effects were advanced for
derivation of the subchronic osRfD, with the same inherent confidence and uncertainties.
Selection of Subchronic Toxicity Value(s)
As described above, subchronic osRfDs associated with each health effect are presented as
they may be useful for certain decision purposes (i.e., site-specific risk assessments with less-than-
lifetime exposures). The osRfD values selected were associated with decreased serum anti-tetanus
antibody concentrations for immune effects and decreased total T4 levels for thyroid effects.
5-25
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Confidence in each osRfD is described in Table 5-10 and consider confidence in the study used to
derive the quantitative estimate, the overall health effect, specific evidence base, and quantitative
estimate for each osRfD.
Selection of Subchronic RfD and Confidence Statement
Organ-/system-specific subchronic RfD values for PFHxS selected in the previous section
are summarized in Table 5-11.
Table 5-11. Subchronic RfD organ-/system-specific RfD values for PFHxS
Subchronic reference dose (RfD)
Basis
RfD (mg/kg-d)
Confidence
Immune (developmental)
effects
4 x 10 10
Medium
Subchronic organ-/system-specific RfDs
Organ/system
Outcomes and
studies
PODHED(mg/kg-
d)
UFC
osRfD (mg/kg-d)
Confidence
Thyroid
Decreased serum T4
(free) in F1 Wistar
rats Ramh0i et al.
(2018)
2.45 x 10"5
(NOAEL)
100
2 x 10"7
Medium
Immune
(developmental)
Decreased serum
anti-tetanus and anti-
diphtheria antibody
concentrations
measured in children
at age 7 with PFHxS
exposure measured
at age 5 Grandiean et
al. (2012); Budtz-
J0rgensen and
Grandiean (2018);
Grandiean et al.
(2012); Budtz-
J0rgensen and
Grandiean (2018)
1.16-1.23 x 10"9
1.16 x 10~8 and
1.23 x 10"8
(BMDUsd)
30
4 x 10 10
Medium
From the identified targets of PFHxS toxicity and derived subchronic osRfDs (see
Table 5-10), an RfD of 4 x 10"10 mg/kg-day based on decreased serum anti-tetanus and diphtheria
antibody concentrations in children is selected for less-than-lifetime exposure. Confidence in the
RfD is medium, based on medium confidence in the immune osRfD, as described in Table 5-8. The
considerations for selecting the immune osRfD for the lifetime RfD are the same as those applied in
selecting the subchronic RfD.
5-26
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
5.2.3. Inhalation Reference Concentration (RfC) Derivation
No studies examining inhalation effects of short-term, subchronic, chronic, or gestational
exposure for PFHxS in humans or animals have been identified, precluding the derivation of an RfC.
5.3. CANCER TOXICITY VALUES
Considering the limitations in the PFHxS evidence base on cancer (see Section 3.3) and in
accordance with the Guidelines for Carcinogen Risk Assessment (U.S. EPA. 20051. EPA concluded
that based on the available evidence, a classification of "Inadequate Information to Assess
Carcinogenic Potential" of PFHxS in humans. The lack of adequate carcinogenicity data for PFHxS
precludes the derivation of quantitative estimates of cancer for either oral (e.g., an oral slope factor
[OSF]) or inhalation (e.g., an inhalation unit risk [IUR]) PFHxS exposure.
5-27
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
REFERENCES
3M (3M Company). (2000a). Sanitized report: Exploratory 28-day oral toxicity study with telomer
alcohol, telomer acrylate, PFHS and PFOS (positive control) by daily gavage in the rat
followed by a 14/28-day recovery period [TSCA Submission], (FYI-0500-01378 B; DCN
84000000018). St Paul, MN.
https://yosemite.epa.gov/oppts/epatscat8.nsf/ReportSearchView/05ADDF3067814C9785
25703B004B4119.
3M (3M Company). (2000b). Support: Protocol & report of data for exploratory 28-day oral toxicity
study in rats: Telomer alcohol, telomer acrylate, [ ], PFHS, PFOS, w/ attachments, cvr ltr
dated 050400 [TSCA Submission], (EPA/OTS Doc #8EHQ-0800-0373). St. Paul, MN.
https://ntrl.ntis.gOv/NTRL/dashboard/searchResults/titleDetail/OTS020492910.xhtml.
3M (3M Company). (2003). Oral (gavage) combined repeated dose toxicity study of T-7706 with the
reproduction/developmental toxicity screening test [TSCA Submission], (89-000811850U;
8EHQ-0903-00373). St. Paul, MN.
3M (3M Company). (2004). 28-day repeated dermal contact study of 3m test articles in Sprague-
Dawley rats final report 01G00002 with letter dated 10/04/2004 [TSCA Submission],
(8EHQ-04-00373BJ). Washington, D.C.: TSCA.
https://java.epa.gov/oppt chemical search/proxy?filename=2004-10-8EH0-04-
00373BT 8ehq 1004 00373bi.pdf.
3M (3M Company). (2010). TSCA 8(e) substantial risk notice: Sulfonate-based and carboxylic-based
fluorochemicals, docket 8EHQ-0598-373 - Results from a mechanistic investigation of the
effect of PFBS, PFHS, and PFOS on lipid and lipoprotein metabolism in transgenic mice
[TSCA Submission], (8EHQ-10-00373DH). St. Paul, MN.
Abraham. K: Mielke. H: Fromme. H: Volkel. W: Menzel. 1: 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.
http://dx.doi.org/10.1007/s00204-020-02715-4.
ACOG (American College of Obstetricians and Gynecologists). (2019). The use of antimullerian
hormone in women not seeking fertility care. Am J Obstet Gynecol 133: e274-e278.
Afzali. A: Weiss. NS: Bovko. El: Ioannou. GN. (2010). Association between serum uric acid level and
chronic liver disease in the United States. Hepatology 52: 578-589.
http://dx.doi.org/10.1002/hep.23717.
AG. B: E. K. (1982). Prevalence of oligomenorrhea and amenorrhea in a college population. Am J
Obstet Gynecol 144: 98-102. http: //dx.doi.org/10.1016/0002-9378f82190402-l.
Aimuzi. R: Luo. K: Chen. 0: Wang. H: Feng. L: Ouvang. F: Zhang. 1. (2019). Perfluoroalkyl and
polyfluoroalkyl substances and fetal thyroid hormone levels in umbilical cord blood among
newborns by prelabor caesarean delivery. Environ Int 130: 104929.
http ://dx. doi. or g/10.1016/i. envint 2019.10 49 29.
Aimuzi. R: Luo. K: Huang. R: Huo. XN: Nian. M: Ouvang. FX: Du. YT: Feng. LP: Wang. WY: Zhang. 1.
(2020). Perfluoroalkyl and polyfluroalkyl substances and maternal thyroid hormones in
early pregnancy. Environ Pollut 264: 114557.
http://dx.doi.Org/10.1016/j.envpol.2020.114557.
Alberti. KG: Eckel. RH: Grundy. SM: Zimmet. PZ: Cleeman. II: Donato. KA: Fruchart. TC: Tames. WP:
Loria. CM: Smith. SC. (2009). Harmonizing the metabolic syndrome: A joint interim
R-l
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
statement of the International Diabetes Federation Task Force on Epidemiology and
Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World
Heart Federation; International Atherosclerosis Society; and International Association for
the Study of Obesity. Circulation 120: 1640-1645.
http://dx.d0i.0rg/l 0.1161 /CTRCULATTONAHA.l09.192644.
Alderete. TL: Tin. R: Walker. PI: Valvi. D: Chen. Z: Tones. 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.
http://dx.doi.Org/10.1016/i.envint.2019.02.047.
Alesio. TL: Slitt. A: Bothun. GD. (2022). Critical new insights into the binding of poly- and
perfluoroalkyl substances (PFAS) to albumin protein. Chemosphere 287: 131979.
http ://dx. doi. or g/10.1016/i. chemosphere. 2 021.131979.
Alkhalawi. E: Kasper-Sonnenberg. M: Wilhelm. M: Volkel. W: Wittsiepe. I. (2016). Perfluoroalkyl
acids (PFAAs) and anthropometric measures in the first year of life: Results from the
Duisburg Birth Cohort J Toxicol Environ Health A 79: 1041-1049.
http://dx.doi.org/10.1080/15287394.2016.1219552.
Amacher. DE. (2002). A toxicologist's guide to biomarkers of hepatic response. Hum Exp Toxicol 21:
253-262. http://dx.d0i.0rg/l 0.1191 /0960327102ht247oa.
American Heart Association. (2022). What is Cholesterol? Dallas, TX.
https://www.heart.org/en/health-topics/cholesterol/about-
cholesterol#:~:text=High%20cholesterol%20is%20one%20of.the%20higher%20your%20
overall%20risk.
Ammitzb0ll. C: Bornsen. L: Petersen. ER: Oturai. AB: S0ndergaard. HB: Grandiean. P: Sellebierg. F.
(2019). Perfluorinated substances, risk factors for multiple sclerosis and cellular immune
activation. J Neuroimmunol 330: 90-95. http://dx.doi.Org/10.1016/j.jneuroim.2019.03.002.
Anderson. RH: Long. GC: Porter. RC: Anderson. IK. (2016). Occurrence of select perfluoroalkyl
substances at U.S. Air Force aqueous film-forming foam release sites other than fire-training
areas: Field-validation of critical fate and transport properties. Chemosphere 150: 678-685.
http://dx.doi.0rg/lO.lOl6/j.chemosphere.2Ol6.Ol.Oi4.
Angrish. MM: Kaiser. TP: Mcqueen. CA: Chorlev. BN. (2016). Tipping the balance: Hepatotoxicity and
the 4 apical key events of hepatic steatosis [Review], Toxicol Sci 150: 261-268.
http: / /dx.doi. or g/10.10 9 3 /toxsci /kfwO 18.
Arbuckle. TE: Macpherson. S: Foster. WG: Sathyanaravana. 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.
http ://dx. doi. or g/10.1016/i. reprotox. 2020.03.011.
Ashley-Martin. I: 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. http://dx.d0i.0rg/l0.1093/aje/kww213.
Ashley-Martin. 1: Dodds. L: Levy. AR: Piatt. RW: Marshall. IS: Arbuckle. TE. (2015). Prenatal
exposure to phthalates, bisphenol A and perfluoroalkyl substances and cord blood levels of
IgE, TSLP and IL-33. Environ Res 140: 360-368.
http://dx.doi.Org/10.1016/j.envres.2015.04.010.
ASRM (American Society of Reproductive Medicine). (2024a). Current evaluation of amenorrhea: A
committee opinion [Review], Fertil Steril 122: 52-61.
http ://dx. doi. or g/10.1016/i. fertnstert. 2024.02.001.
R-2
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
ASRM (American Society of Reproductive Medicine). (2024b). Reasons why young women in the
United States choose their contraceptive method. J Womens Health 30: 64-72.
http://dx.doi.org/10.1089/iwh.2019.8182.
ATSDR (Agency for Toxic Substances and Disease Registry). (2021). Toxicological profile for
perfluoroalkyls [ATSDR Tox Profile], Atlanta, GA: U.S. Department of Health and Human
Services, Public Health Service. http://dx.doi.Org/10.15620/cdc:59198.
ATSDR (Agency for Toxic Substances and Disease Registry). (2024). PFAS: Information for
clinicians [Fact Sheet], https://www.atsdr.cdc.gov/pfas/docs/PFAS-info-for-clinicians-
508.pdf.
Attanasio. R. (2019a). Association between perfluoroalkyl acids and liver function: Data on sex
differences in adolescents. Data in Brief 27: 104618.
http ://dx. doi. or g/10.1016/i. dib. 2019.104618.
Attanasio. R. (2019b). Sex differences in the association between perfluoroalkyl acids and liver
function in US adolescents: Analyses of NHANES 2013-2016. Environ Pollut 254: 113061.
http://dx.doi.Org/10.1016/j.envpol.2019.113061.
Augustsson. A: Lennqvist. T: Osbeck. CMG: Tibblin. P: Glynn. A: Nguyen. MA: Westberg. E:
Vestergren. R. (2021). Consumption of freshwater fish: A variable but significant risk factor
for PF0S exposure. Environ Res 192: 10284-10284.
Averina. M: Brox. 1: 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. http://dx.doi.Org/10.1016/i.envres.2021.110740.
Avcan. MF. (2019). Comparison of biomechanical properties of implant systems used in treatment
of proximal femur fractures. Gazi Universitesi Muhendislik Mimarlik Fakultesi Dergisi 34:
812-818. http://dx.doi.org/10.17341/gazimmfd.416539.
B. D: ES. F. (2018). Chapter 19 - The Mammary Gland. In WMHCGRBB Matthew A. Wallig (Ed.),
Fundamentals of Toxicologic Pathology (3 ed., pp. 547-563). Research Triangle Park, NC:
Innogenics, Inc., Harvard, MA, United States National Institute of Environmental Health
Sciences, http: / /dx.doi.org/10.1016/B978-0-12-809841 -7.00019-8.
Bach. CC: Bech. BH: Nohr. EA: Olsen. I: Matthiesen. NB: Bonefeld-Iargensen. EC: Bossi. R: Henriksen.
TB. (2016). Perfluoroalkyl acids in maternal serum and indices of fetal growth: The Aarhus
Birth Cohort Environ Health Perspect 124: 848-854.
http ://dx. doi. or g/10.12 89 /ehp. 1510046.
Bach. CC: Bech. BH: Nohr. EA: Olsen. I: Matthiesen. NB: Bossi. R: Uldbierg. N: Bonefeld-Iargensen.
EC: Henriksen. TB. (2015). Serum perfluoroalkyl acids and time to pregnancy in nulliparous
women. Environ Res 142: 535-541. http://dx.doi.Org/10.1016/i.envres.2015.08.007.
Bach. CC: Matthiesen. NB: Olsen. I: Henriksen. TB. (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. http://dx.doi.org/10.1289/EHP 1493.
Baduel. C: Paxman. CI: Mueller. IF. (2015). Perfluoroalkyl substances in a firefighting training
ground (FTG), distribution and potential future release. J Hazard Mater 296: 46-53.
http://dx.doi.Org/10.1016/j.jhazmat2015.03.007.
Bailey. SA: Zidell. RH: Perry. RW. (2004). Relationships between organ weight and body/brain
weight in the rat: What is the best analytical endpoint? Toxicol Pathol 32: 448-466.
http: //dx.doi.org/10.1080/01926230490465874.
Bangma. 1: Eaves. LA: Oldenburg. K: Reiner. TL: 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.
http ://dx. doi. or g/10.10 21 /acs. est 9b0 7102.
R-3
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Banjabi. AA: Li. AT: Kumosani. TA: Yousef. TM: Kannan. K. (2020). Serum concentrations of
perfluoroalkyl substances and their association with osteoporosis in a population in Jeddah,
Saudi Arabia. Environ Res 187: 109676. http: //dx.doi.Org/10.1016/i.envres.2020.109676.
Bao. WW: Oian. ZM: Geiger. SD: Liu. E: Liu. Y: Wang. SO: 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. http://dx.doi.Org/10.1016/i.scitotenv.2017.07.124.
Barber. TL: Berger. U: Chaemfa. C: Huber. S: Tahnke. A: Temme. C: Tones. KC. (2007). Analysis of per-
and polyfluorinated alkyl substances in air samples from Northwest Europe. J Environ
Monit9: 530-541. http://dx.doi.org/10.1039/b701417a.
Barbier. P: Zejneli. 0: Martinho. M: Lasorsa. A: Belle. V: Smet-Nocca. C: Tsvetkov. PO: Devred. F:
Landrieu. I. (2019). Role of Tau as a microtubule-associated protein: Structural and
functional aspects. Front Aging Neurosci 11: 204.
http: / /dx. doi. or g/10.3 3 89 /fnagi. 2019.00204.
Barrett. ES: Chen. C: Thurston. SW: Haug. LS: Sabaredzovic. A: Fjeldheim. FN: Frvdenberg. H: Lipson.
SF: Ellison. PT: Thune. I. (2015). Perfluoroalkyl substances and ovarian hormone
concentrations in naturally cycling women. Fertil Steril 103: 1261-1270.el263.
http: / /dx. doi. or g/10.1016/i. fertnstert. 2015.02.001.
Bartell. SM. (2012). Bias in half-life estimates using log concentration regression in the presence of
background exposures, and potential solutions. J Expo Sci Environ Epidemiol 22: 299-303.
http: / /dx. doi. or g/10.10 3 8 /i es. 2 012.2.
Bassett. THD: Clement-Lacroix. P: Samarut. T: Chassande. 0: Williams. GR: O'Shea. PT:
Sriskantharaiah. S: Rabier. B: Bovde. A: Howell. PGT: Weiss. R. ovE: Roux. TP: Malaval. L. uc.
(2007). Thyroid hormone excess rather than thyrotropin deficiency induces osteoporosis in
hyperthyroidism. Mol Endocrinoll095-1107. http://dx.doi.org/10.1210/me.2007-0033.
Batt. AM: Ferrari. L. (1995). Manifestations of chemically-induced liver-damage. Clin Chem 41:
1882-1887.
Batzella. E: Girardi. P: Russo. F: Pitter. G: Da Re. F: Fletcher. T: Canova. C. (2022). Perfluoroalkyl
substance mixtures and cardio-metabolic outcomes in highly exposed male workers in the
Veneto Region: A mixture-based approach. Environ Res 212: 113225.
http://dx.doi.Org/10.1016/i.envres.2022.113225.
Beck. IH: Timmermann. CAG: Nielsen. F: Schoeters. G: Tahnk. C: Kvhl. HB: Hast. A: Tensen. 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.
http://dx.doi.org/10.1186/sl2940-019-0541-z.
Beglarian. E: Costello. E: Walker. PI: Wang. H: Alderete. TL: Chen. Z: Valvi. D: Baumert. BO: Rock. S:
Rubbo. B: Aung. MT: Gilliland. FD: Goran. MI: Tones. DP: McConnell. R: Eckel. SP: Conti. DV:
Goodrich. TA: Chatzi. L. (2024). Exposure to perfluoroalkyl substances and longitudinal
changes in bone mineral density in adolescents and young adults: A multi-cohort study.
Environ Res 244: 117611. http://dx.doi.Org/10.1016/i.envres.2023.117611.
Benskin. TP: De Silva. AO: Martin. LT: Arsenault. G: Mccrindle. R: Riddell. N: Maburv. SA: Martin. TW.
(2009). Disposition of perfluorinated acid isomers in Sprague-Dawley rats; part 1: single
dose. Environ Toxicol Chem 28: 542-554. http://dx.doi.Org/10.1897/08-239.l.
Berbel. P: Navarro. D: Auso. E: Varea. E: Rodriguez. AE: Ballesta. IT: Salinas. M: Flores. E: Faura. CC:
de Escobar. GM. (2010). Role of late maternal thyroid hormones in cerebral cortex
development: An experimental model for human prematurity. Cereb Cortex 20: 1462-1475.
http: / /dx.doi. or g/10.10 9 3 /cercor /bhp 212.
Berg. V: N0st. TH: Pettersen. RD: Hansen. S: Vevhe. AS: Torde. R: Odland. 10: Sandanger. TM. (2017).
Persistent organic pollutants and the association with maternal and infant thyroid
R-4
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
homeostasis: a multipollutant assessment. Environ Health Perspect 125: 127-133.
http://dx.doi.org/10.12 89 /EHP15 2.
Bessone. F: Razori. MY: Roma. MG. (2019). Molecular pathways of nonalcoholic fatty liver disease
development and progression. Cell Mol Life Sci 76: 99-128.
http://dx.doi.Org/10.1007/s00018-018-2947-0.
Biiland. S: Rensen. PCN: Pieterman. El: Maas. ACE: van Per Hoorn. TW: van Erk. Ml: Havekes. LM:
Willems van Diik. K: Chang. SC: Ehresman. DT: Butenhoff. TL: Princen. HMG. (2011).
Perfluoroalkyl sulfonates cause alkyl chain length-dependent hepatic steatosis and
hypolipidemia mainly by impairing lipoprotein production in AP0E*3-Leiden CETP mice.
Toxicol Sci 123: 290-303. http://dx.doi.org/10.1093/toxsci/kfrl42.
Bil. W: Zeilmaker. Ml: Bokkers. BGH. (2022). Internal Relative Potency Factors for the Risk
Assessment of Mixtures of Per- and Polyfluoroalkyl Substances (PFAS) in Human
Biomonitoring. Environ Health Perspect 130: 77005. http://dx.doi.org/10.1289/EHP10009.
Birukov. A: Andersen. LB: Andersen. MS: Nielsen. TH: Nielsen. F: Kvhl. HB: Targensen. IS: Grandiean.
P: Dechend. R: Tensen. TK. (2021). Exposure to perfluoroalkyl substances and blood
pressure in pregnancy among 1436 women from the Odense Child Cohort Environ Int 151:
106442. http ://dx.doi. or g/10.1016/i.envint2021.106442.
Bischel. HN: Macmanus-Spencer. LA: Luthv. RG. (2010). Noncovalent Interactions of Long-Chain
Perfluoroalkyl Acids with Serum Albumin. Environ Sci Technol 44: 5263-5269.
http ://dx. doi. or g/10.10 21 /es 1013 3 4s.
Bischel. HN: Macmanus-Spencer. LA: Zhang. CI: Luthv. RG. (2011). Strong associations of short-
chain perfluoroalkyl acids with serum albumin and investigation of binding mechanisms.
Environ Toxicol Chem 30: 2423-2430. http: //dx.doi.org/10.1002 /etc.647.
Bizzarro. Ml: Gross. I. (2004). Effects of hormones on fetal lung development [Review], Obstet
Gynecol Clin North Am 31: 949-961, xii. http: //dx.doi.Org/10.1016/i.ogc.2004.08.001.
Bjerregaard-Olesen. C: Bach. CC: Long. M: Wielsae. M: Bech. BH: Henriksen. TB: Olsen. I: Bonefeld-
T0rgensen. EC. (2019). Associations of Fetal Growth Outcomes with Measures of the
Combined Xenoestrogenic Activity of Maternal Serum Perfluorinated Alkyl Acids in Danish
Pregnant Women. Environ Health Perspect 127: 17006.
http://dx.d0i.0rg/l 0.1289/EHP1884.
Biork. TA: Dawson. DA: Krogstad. 10: Wallace. KB. (2021). Transcriptional effects of binary
combinations of PFAS in FaO cells. Toxicology 464: 152997.
http://dx.doi.Org/10.1016/i.tox.2021.152997.
Blake. BE: Pinnev. 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. http://dx.d0i.0rg/l0.1016/i.envpol.2018.07.042.
Blomberg. A: Mortensen. I: Weihe. P: Grandiean. P. (2022). Bone mass density following
developmental exposures to perfluoroalkyl substances (PFAS): a longitudinal cohort study.
Environ Health 21: 113. http://dx.doi.org/10.1186/sl2940-022-0Q929-w.
Blomberg. AT: Haug. LS: Lindh. C: Sabaredzovic. A: Pineda. D: Takobsson. K: Nielsen. C. (2023).
Changes in perfluoroalkyl substances (PFAS) concentrations in human milk over the course
of lactation: A study in Ronneby mother-child cohort. Environ Res 219: 115096.
http ://dx. doi. or g/10.1016/i. envres. 2022.115096.
Blomberg. AT: Shih. YH: Messerlian. C: Targensen. 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. http://dx.doi.org/10.1016/i.envres.2021.111400.
Bloom. MS: Commodore. S: Ferguson. PL: Neelon. B: Pearce. TL: Baumer. A: Newman. RB: Grobman.
W: Tita. A: Roberts. 1: Skupski. D: Palomares. K: Nageotte. M: Kannan. K: Zhang. C: Wapner.
R: Vena. TE: Hunt. KT. (2022). Association between gestational PFAS exposure and Children's
R-5
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
adiposity in a diverse population. Environ Res 203: 111820.
http://dx.doi.Org/10.1016/i.envres.2021.111820.
Bloom. MS: Kannan. K: Spliethoff. HM: Tao. L: Aldous. KM: Vena. TE. (2010). Exploratory assessment
of perfluorinated compounds and human thyroid function. Physiol Behav 99: 240-245.
http://dx.doi.Org/10.1016/i.physbeh.2009.02.005.
Bonefeld-largensen. EC: Long. M: Fredslund. SO: Bossi. R: Olsen. 1. (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.
http: //dx.doi.org/10.1007/sl 0552-014-0446-7.
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.
http://dx.doi.Org/10.llll/i.1939-165X.2005.tb00041.x.
Borghese. MM: Liang. CL: Owen. 1: Fisher. M. (2022). Individual and mixture associations of
perfluoroalkyl substances on liver function biomarkers in the Canadian Health Measures
Survey. Environ Health 21: 85. http: //dx.doi.org/10.1186/sl2940-022-0Q892-6.
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. http://dx.doi.Org/10.1016/j.envint.2020.105789.
Braun. TM: 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. http://dx.doi.org/10.1002/oby.21258.
Braun. TM: Eliot. M: Papandonatos. GD: Buckley. TP: Cecil. KM: Kalkwarf. HI: Chen. A: Eaton. CB:
Kelsev. K: Lanphear. BP: Yolton. K. (2020). Gestational perfluoroalkyl substance exposure
and body mass index trajectories over the first 12 years of life. Int J Obes (Lond) 45: 25-35.
http://dx.d0i.0rg/l 0.1038/s41366-020-00717-x.
Braun. TM: Kalkbrenner. AE: lust. AC: Yolton. K: Calafat. AM: Siodin. A: Hauser. R: Webster. GM:
Chen. A: Lanphear. BP. (2014). Gestational exposure to endocrine-disrupting chemicals and
reciprocal social, repetitive, and stereotypic behaviors in 4- and 5-year-old children: the
HOME study. Environ Health Perspect 122: 513-520.
http ://dx. doi. or g/10.12 89 /ehp. 1307261.
Buck Louis. GM: Zhai. S: Smarr. MM: Grewal. I: Zhang. C: Grantz. KL: Hinkle. SN: Sundaram. R: Lee. S:
Honda. M: Oh. I: Kannan. K. (2018). Endocrine disruptors and neonatal anthropometry,
NICHD Fetal Growth Studies - Singletons. Environ Int 119: 515-526.
http: / /dx. doi. or g/10.1016/i. envint 2018.07.024.
Buck. RC: Franklin. I: Berger. U: Conder. TM: Cousins. IT: de Voogt. P: Tensen. AA: Kannan. K: Maburv.
SA: van Leeuwen. SP. (2011). Perfluoroalkyl and polyfluoroalkyl substances in the
environment: terminology, classification, and origins [Review], Integr Environ Assess
Manag 7: 513-541. http://dx.doi.org/10.1002/ieam.258.
Buckley. TP: Kuiper. TR: Lanphear. BP: Calafat. AM: Cecil. KM: Chen. A: Xu. Y: Yolton. K: Kalkwarf. HI:
Braun. TM. (2021). Associations of maternal serum perfluoroalkyl substances
concentrations with early adolescent bone mineral content and density: The Health
Outcomes and Measures of the Environment (HOME) study. Environ Health Perspect 129:
97011. http://dx.d0i.0rg/l0.1289/EHP9424.
Budtz-largensen. E: Grandjean. P. (2018). Application of benchmark analysis for mixed contaminant
exposures: Mutual adjustment of perfluoroalkylate substances associated with
immunotoxicity. PLoS ONE 13: e0205388.
http://dx.doi.org/10.1371/iournal.pone.0205388.
Bueters. R: Bael. A: Gasthuvs. E: Chen. C: Schreuder. MF: Frazier. KS. (2020). Ontogeny and Cross-
species Comparison of Pathways Involved in Drug Absorption, Distribution, Metabolism,
R-6
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
and Excretion in Neonates (Review): Kidney [Review], Drug Metab Dispos 48: 353-367.
http://dx.doi.org/10.1124/dmd.119.Q89755.
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. http://dx.doi.Org/10.1016/i.envpol.2021.116619.
Buser. MC: Scinicariello. F. (2016). Perfluoroalkyl substances and food allergies in adolescents.
Environ Int 88: 74-79. http://dx.doi.Org/10.1016/i.envint.2015.12.020.
Butenhoff. TL: Chang. SC: Ehresman. DT: York. RG. (2009). Evaluation of potential reproductive and
developmental toxicity of potassium perfluorohexanesulfonate in Sprague Dawley rats.
Reprod Toxicol 27: 331-341. http://dx.doi.Org/10.1016/i.reprotox.2009.01.004.
Byrne. S: Seguinot-Medina. S: Miller. P: Waghivi. V: von Hippel. FA: Buck. CL: Carpenter. DO. (2017).
Exposure to polybrominated diphenyl ethers and perfluoroalkyl substances in a remote
population of Alaska Natives. Environ Pollut 231: 387-395.
http://dx.doi.Org/10.1016/i.envpol.2017.08.020.
Cakmak. S: Lukina. A: Karthikevan. S: Atlas. E: Dales. R. (2022). The association between blood PFAS
concentrations and clinical biochemical measures of organ function and metabolism in
participants of the Canadian Health Measures Survey (CHMS). Sci Total Environ 827:
153900. http://dx.doi. or g/10.1016/i.scitotenv.2022.153900.
Callan. AC: Rotander. A: Thompson. K: Hevworth. 1: Mueller. IF: Odland. 10: Hinwood. AL. (2016).
Maternal exposure to perfluoroalkyl acids measured in whole blood and birth outcomes in
offspring. Sci Total Environ 569-570: 1107-1113.
http ://dx. doi. or g/10.1016/i .scitotenv. 2016.06.177.
Campbell. S: Raza. M: Pollack. AZ. (2016). Perfluoroalkyl substances and endometriosis in US
women in NHANES 2003-2006. Reprod Toxicol 65: 230-235.
http://dx.doi.Org/10.1016/i.reprotox.2016.08.009.
Canova. C: Barbieri. G: Zare Teddi. 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.
http ://dx. doi. or g/10.1016/i. envint 2020.106117.
Canova. C: Pi 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.
http://dx.doi.org/10.3390/iierphl82412881.
Cao. W: Liu. X: Liu. X: Zhou. Y: Zhang. X: Tian. H: Wang. I: Feng. S: Wu. Y: Bhatti. P: Wen. S: Sun. X.
(2018). Perfluoroalkyl substances in umbilical cord serum and gestational and postnatal
growth in a Chinese birth cohort Environ Int 116: 197-205.
http ://dx. doi. or g/10.1016/i .envint 2018.04.015.
Cardenas. A: Gold. PR: 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 Piabetes Prevention Program trial. Environ Health Perspect 125: 107001.
http://dx.doi.org/10.1289/EHP1612.
Cariou. R: Vevrand. B: Yamada. A: Berrebi. A: Zalko. P: Purand. S: Pollono. C: Marchand. P: Leblanc.
TC: Antignac. TP: 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.
http: / /dx. doi. or g/10.1016/i .envint 2015.07.014.
Carlson. LA: Angrish. M: Shirke. AY: Radke. EG: Schulz. B: Kraft. A: Tudson. R: Patlewicz. G: Blain. R:
Lin. C: Vetter. N: Lemeris. C: Hartman. P: Hubbard. H: Arzuaga. X: Pavis. A: Pishaw. LV:
Pruwe. IL: Hollinger. H:... Thayer. KA. (2022). Systematic evidence map for over one
R-7
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
hundred and fifty per- and polyfluoroalkyl substances (PFAS) [Review], Environ Health
Perspect 130: 56001. http://dx.doi.org/10.1289/EHP10343.
Caron-Beaudoin. E: Avotte. 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.
http://dx.doi.Org/10.1016/i.envint.2019.04.029.
Carwile. TL: Seshasavee. SM: Ahrens. KA: Hauser. R: Driban. IB: Rosen. CT: Gordon. CM: Fleisch. AF.
(2022). Serum PFAS and urinary phthalate biomarker concentrations and bone mineral
density in 12-19 year olds: 2011-2016 NHANES. J Clin Endocrinol Metab 107: e3343-e3352.
http: //dx.doi.org/10.1210 /clinem/dgac2 2 8.
Carwile. TL: Seshasavee. SM: Aris. IM: Rifas-Shiman. SL: Claus Henn. B: Calafat. AM: Sagiv. SK: Oken.
E: Fleisch. AF. (2021). Prospective associations of mid-childhood plasma per- and
polyfluoroalkyl substances and pubertal timing. Environ Int 156: 106729.
http://dx.doi.Org/10.1016/j.envint2021.106729.
Cassone. CG: Vongphachan. V: Chiu. S: Williams. KL: Letcher. RT: Pelletier. E: Crump. D: Kennedy.
SW. (2012). In ovo effects of perfluorohexane sulfonate and perfluorohexanoate on pipping
success, development, mRNA expression, and thyroid hormone levels in chicken embryos.
Toxicol Sci 127: 216-224. http://dx.doi.org/10.1093/toxsci/kfs072.
Cattlev. RC: Cullen. TM. (2018). Chapter 8. Liver and gall bladder. In MA Wallig; WM Haschek; CG
Rousseaux; B Bolon (Eds.), Fundamentals of toxicologic pathology (3rd ed., pp. 125-151).
Cambridge, MA: Academic Press. http://dx.doi.org/10.1016/B978-0-12-809841-7.00008-3.
CDC (Centers for Disease Control and Prevention). (2022). National report on human exposure to
environmental chemicals. Atlanta, GA: U.S. Department of Health and Human Services.
https://www.cdc.gov/exposurereport/.
Chan. E: Burstyn. I: Cherry. N: Bamforth. F: Martin. TW. (2011). Perfluorinated acids and
hypothyroxinemia in pregnant women. Environ Res 111: 559-564.
http://dx.doi.Org/10.1016/i.envres.2011.01.011.
Chang. CT: Barr. DB: Ryan. PB: Panuwet. P: Smarr. MM: Liu. K: Kannan. K: Yakimavets. V: Tan. Y: Lv.
V: Marsit. CT: Tones. DP: Corwin. El: Dunlop. AL: Liang. D. (2022). Per- and polyfluoroalkyl
substance (PFAS) exposure, maternal metabolomic perturbation, and fetal growth in
African American women: A meet-in-the-middle approach. Environ Int 158: 106964.
http: / /dx. doi. or g/10.1016/i. envint 2021.106964.
Chang. S: Butenhoff. TL: Parker. GA: Coder. PS: Zitzow. ID: Krisko. RM: Bjork. TA: Wallace. KB: Seed.
TG. (2018). Reproductive and developmental toxicity of potassium
perfluorohexanesulfonate in CD-I mice. Reprod Toxicol 78: 150-168.
http://dx.doi.Org/10.1016/i.reprotox.2018.04.007.
Charles. D: Berg. V: N0st. TH: Huber. S: Sandanger. TM: Rvlander. C. (2020). Pre- and post-
diagnostic blood profiles of perfluoroalkyl acids in type 2 diabetes mellitus cases and
controls. Environ Int 145: 106095. http://dx.doi.Org/10.1016/i.envint.2020.106095.
Chen. A: Tandarov. R: Zhou. L: Calafat. AM: Zhang. G: Urbina. EM: Sarac. 1: Augustin. DH: Caric. T:
Bockor. L: Petranovic. MZ: Novokmet. N: Missoni. S: Rudan. P: Deka. R. (2019a). 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.
http: / /dx. doi. or g/10.1016/j. scitotenv. 2019.05.250.
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. http://dx.doi.org/10.1021/acs.est7b00268.
Chen. L: Tong. C: Huo. X: Zhang. 1: Tian. Y. (2021). Prenatal exposure to perfluoroalkyl and
polyfluoroalkyl substances and birth outcomes: A longitudinal cohort with repeated
R-8
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
measurements. Chemosphere 267: 128899.
http://dx.doi.Org/10.1016/i.chemosphere.2020.128899.
Chen. 0: Huang. R: Hua. L: Guo. Y: Huang. L: Zhao. Y: Wang. X: Zhang. T. (2018a). Prenatal exposure
to perfluoroalkyl and polyfluoroalkyl substances and childhood atopic dermatitis: A
prospective birth cohort study. Environ Health 17: 1-12.
http: //dx.doi.org/10.1186/sl 2940-018-0352-7.
Chen. 0: Zhang. X: Zhao. Y: Lu. W: Wu. 1: Zhao. S: Zhang. 1: Huang. L. (2019b). Prenatal exposure to
perfluorobutanesulfonic acid and childhood adiposity: A prospective birth cohort study in
Shanghai, China. Chemosphere 226: 17-23.
http://dx.doi.Org/10.1016/i.chemosphere.2019.03.095.
Chen. S: Guo. X: Yu. S: Sun. G: Yang. H: Li. Z: Sun. Y. (2016). Association between Serum Uric Acid
and Elevated Alanine Aminotransferase in the General Population. Int J Environ Res Public
Health 13: 841. http://dx.doi.org/10.3390/iierphl3090841.
Chen. WL: Bai. FY: Chang. YC: Chen. PC: Chen. CY. (2018b). Concentrations of perfluoroalkyl
substances in foods and the dietary exposure among Taiwan general population and
pregnant women. J Food Drug Anal 26: 994-1004.
http://dx.doi.Org/10.1016/i.ifda.2017.12.011.
Chevrier. 1. (2013). Invited commentary: Maternal plasma polybrominated diphenyl ethers and
thyroid hormones-challenges and opportunities [Comment], Am J Epidemiol 178: 714-719.
http: / /dx.doi. or g/10.10 9 3/ai e /kwtl 38.
Chiu. WA: Lynch. MT: Lay. CR: Antezana. A: Malek. P: Sokolinski. S: Rogers. RD. (2022). Bayesian
estimation of human population toxicokinetics of PF0A, PF0S, PFHxS, and PFNA from
studies of contaminated drinking water. Environ Health Perspect 130: 127001.
http://dx.doi.org/10.1289/EHP10103.
Christensen. TVR: Bangash. KK: Weihe. P: Grandiean. P: Nielsen. F: Tensen. TK: Petersen. MS. (2021).
Maternal exposure to perfluoroalkyl chemicals and anogenital distance in the offspring: A
Faroese cohort study. Reprod Toxicol 104: 52-57.
http://dx.doi.Org/10.1016/i.reprotox.2021.06.016.
Christensen. KY: Maisonet. M: Rubin. C: Holmes. A: Calafat. AM: Kato. K: Flanders. WD: Heron. 1:
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. http://dx.doi.Org/10.1016/i.envint.2010.08.007.
Christensen. KY: Raymond. M: Blackowicz. M: Liu. Y: Thompson. BA: Anderson. HA: Turvk. M.
(2017). Perfluoroalkyl substances and fish consumption. Environ Res 154: 145-151.
http ://dx. doi. or g/10.1016/i. envres. 2016.12.032.
Christensen. KY: Raymond. M: Meiman. 1. (2019). Perfluoroalkyl substances and metabolic
syndrome. Int J Hyg Environ Health 222: 147-153.
http ://dx. doi. or g/10.1016/i. iiheh. 2018.08.014.
Christensen. KY: Raymond. M: Thompson. BA: Anderson. HA. (2016). Perfluoroalkyl substances in
older male anglers in Wisconsin. Environ Int91: 312-318.
http://dx.doi.Org/10.1016/j.envint2016.03.012.
Cluett. R: Seshasavee. SM: Rokoff. LB: Rifas-Shiman. SL: Ye. X: Calafat. AM: Gold. PR: Coull. B:
Gordon. CM: Rosen. CI: 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.
http://dx.doi.org/10.1289/EHP4918.
Colicino. E: Pedretti. NF: Busgang. SA: Gennings. C. (2020). Per- and poly-fluoroalkyl substances and
bone mineral density: Results from the Bayesian weighted quantile sum regression.
Environmental Epidemiology 4: 1. http://dx.doi.Org/10.1097/EE9.0000000000000092.
R-9
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Conway. B: Innes. KE: Long. D. (2016). Perfluoroalkyl substances and beta cell deficient diabetes. J
Diabetes Complications 30: 993-998. http://dx.doi.org/10.1016/i.idiacomp.2016.05.001.
Conway. BN: Badders. AN: Costacou. T: Arthur. 1M: Innes. KE. (2018). Perfluoroalkyl substances and
kidney function in chronic kidney disease, anemia, and diabetes. Diabetes Metab Syndr Obes
11: 707-716. http://dx.doi.org/10.2147/DMSO.S173809.
Coperchini. F: Awwad. 0: Rotondi. M: Santini. F: Imbriani. M: Chiovato. L. (2017). Thyroid
disruption by perfluorooctane sulfonate (PFOS) and perfluorooctanoate (PFOA) [Review], J
Endocrinol Invest 40: 105-121. http://dx.doi.org/10.1007/s40618-016-0572-z.
Costa. 0: Iniguez. C: Manzano-Salgado. CB: Amiano. P: Murcia. M: Casas. M: Irizar. A: Basterrechea.
M: Beneito. A: Schettgen. T: Sunver. 1: Vriiheid. M: Ballester. F: Lopez-Espinosa. Ml. (2019).
First-trimester maternal concentrations of polyfluoroalkyl substances and fetal growth
throughout pregnancy. Environ Int 130: 104830.
http://dx.doi.Org/10.1016/i.envint2019.05.024.
Crawford. NM: Fenton. SE: Strvnar. 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. http://dx.doi.Org/10.1016/i.reprotox.2017.01.006.
Crofton. KM. (2004). Developmental disruption of thyroid hormone: Correlations with hearing
dysfunction in rats. Risk Anal 24: 1665-1671. http: //dx.doi.org/10.1111 /i.0272-
4332.2004.00557.x.
Cuevas. E: Auso. E: Telefont. M: De Escobar. GM: Sotelo. C: Berbel. P. (2005). Transient maternal
hypothyroxinemia at onset of corticogenesis alters tangential migration of medial
ganglionic eminence-derived neurons. Eur J Neurosci 22: 541-551.
http://dx.doi.Org/10.llll/i.1460-9568.2005.04243.x.
Dalla Zuanna. T: Savitz. DA: Barbieri. G: Pitter. G: Zare Teddi. 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. http://dx.doi.org/10.1016/i.ecoenv.2020.111805.
Dalsager. L: Christensen. N: Halekoh. U: Timmermann. CAG: Nielsen. F: Kvhl. HB: Husbv. S:
Grandiean. P: Tensen. TK: Andersen. HR. (2021a). 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.
http://dx.doi.Org/10.1016/i.envint2021.106395.
Dalsager. L: Christensen. N: Husbv. S: Kvhl. H: Nielsen. F: Hast. A: Grandiean. P: Tensen. 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. http://dx.doi.Org/10.1016/i.envint.2016.08.026.
Dalsager. L: Tensen. TK: Nielsen. F: Grandiean. P: Bilenberg. N: Andersen. HR. (2021b). No
association between maternal and child PFAS concentrations and repeated measures of
ADHD symptoms at age 2% and 5 years in children from the Odense Child Cohort
Neurotoxicol Teratol 88: 107031. http://dx.doi.Org/10.1016/i.ntt2021.107031.
Darrow. DC: Soule. HC: Buckman. TE. (1928). Blood volume in normal infants and children. J Clin
Invest 5: 243-258. http://dx.doi.org/10.1172/TCI100156.
Das. KP: Wood. CR: Lin. MT: Starkov. AA: Lau. C: Wallace. KB: Corton. TC: Abbott. BP. (2017).
Perfluoroalkyl acids-induced liver steatosis: Effects on genes controlling lipid homeostasis.
Toxicology 378: 37-52. http://dx.doi.Org/10.1016/j.tox.2016.12.007.
Daston. GP: Kimmel. CA. (1998). An evaluation and interpretation of reproductive endpoints for
human health risk assessment In An evaluation and interpretation of reproductive
endpoints for human health risk assessment Washington, DC: ILSI Press.
Daugherty. A: Tall. AR: Daemen. MIA: Falk. E: Fisher. EA: Garcia-Cardena. G: Lusis. AT: Owens. AP. Ill:
Rosenfeld. ME: Virmani. R. (2017). Recommendation on design, execution, and reporting of
R-10
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
animal atherosclerosis studies: A scientific statement from the American Heart Association
[Review], Arterioscler Thromb Vase Biol 37: el31-el57.
http://dx.doi.Org/l 0.1161 /ATV.0000000000000062.
Davies. B: Morris. T. (1993). Physiological parameters in laboratory animals and humans [Review],
Pharm Res 10: 1093-1095. http://dx.doi.Org/10.1023/A:1018943613122.
De Boo. HA: Harding. IE. (2006). The developmental origins of adult disease (Barker) hypothesis
[Review], AustN Z J Obstet Gynaecol 46: 4-14. http://dx.doi.Org/10.l 111/i.1479-
82 8X. 2006.005 06.x.
De Vito. P: Incerpi. S: Pedersen. TZ: Lulv. P: Davis. FB: Davis. PI. (2011). Thyroid hormones as
modulators of immune activities at the cellular level [Review], Thyroid 21: 879-890.
http: / /dx. doi. or g/10.10 89 /thy. 2010.0429.
Dean. A: Sharpe. RM. (2013). Clinical review: Anogenital distance or digitlength ratio as measures
of fetal androgen exposure: relationship to male reproductive development and its
disorders [Review], J Clin Endocrinol Metab 98: 2230-2238.
http://dx.doi.org/10.1210/jc.2012-4057.
Dent. MP: Carmichael. PL: Tones. KC: Martin. FL. (2015). Towards a non-animal risk assessment for
anti-androgenic effects in humans [Review], Environ Int 83: 94-106.
http://dx.doi.Org/10.1016/i.envint2015.06.009.
di Masi. A: De Marinis. E: Ascenzi. P: Marino. M. (2009). Nuclear receptors CAR and PXR: Molecular,
functional, and biomedical aspects [Review], Mol Aspects Med 30: 297-343.
http://dx.doi.Org/10.1016/i.mam.2009.04.002.
Ding. N: Harlow. SD: Randolph. IF: Mukheriee. B: Batterman. S: Gold. EB: Park. SK. (2022).
Perfluoroalkyl Substances and Incident Natural Menopause in Midlife Women: The
Mediating Role of Sex Hormones. Am J Epidemiol 191: 1212-1223.
http: / /dx. doi. or g/10.109 3 /ai e /kwacO 5 2.
Domazet. SL: Tensen. 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.
http://dx.doi.Org/10.1016/i.envres.2020.110110.
Dong. GH: Tung. KY: Tsai. CH: Liu. MM: Wang. D: Liu. W: Tin. YH: Hsieh. WS: Lee. YL: Chen. PC.
(2013). Serum polyfluoroalkyl concentrations, asthma outcomes, and immunological
markers in a case-control study of Taiwanese children. Environ Health Perspect 121: 507-
513, 513e501-508. http://dx.doi.org/10.1289/elip.1205351.
Dong. H: You. SH: Williams. A: Wade. MG: Yauk. CL: Zoeller. RT. (2015). Transient maternal
hypothyroxinemia potentiates the transcriptional response to exogenous thyroid hormone
in the fetal cerebral cortex before the onset of fetal thyroid function: A messenger and
microRNA profiling study. Cereb Cortex 25: 1735-1745.
http: / /dx.doi. or g/10.10 9 3 /cercor /bht3 64.
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.
http://dx.doi.Org/10.1016/i.ecoenv.2019.02.061.
Donley. GM: Taylor. E: Teddy. Z: Namulanda. G: Hartman. TT. (2019). Association between in utero
perfluoroalkyl substance exposure and anti-Miillerian hormone levels in adolescent females
in a British cohort Environ Res 177: 108585.
http://dx.doi.Org/10.1016/i.envres.2019.108585.
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.
http://dx.doi.Org/10.1016/i.envint2019.105295.
R-ll
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Dufour. P: Pirard. C: Seghave. 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. http://dx.doi.Org/10.1016/i.envpol.2018.03.058.
Dunder. L: Lind. PM: Salihovic. S: Stubleski. 1: Karrman. A: Lind. L. (2022). Changes in plasma levels
of per- and polyfluoroalkyl substances (PFAS) are associated with changes in plasma lipids -
A longitudinal study over 10 years. Environ Res 211: 112903.
http://dx.doi.Org/10.1016/i.envres.2022.112903.
Dzierlenga. AL: Robinson. VG: Waidvanatha. S: Devito. Ml: Eifrid. MA: Gibbs. ST: Granville. CA:
Blvstone. CR. (2019). Toxicokinetics of perfluorohexanoic acid (PFHxA), perfluorooctanoic
acid (PFOA) and perfluorodecanoic acid (PFDA) in male and female Hsd:Sprague dawley SD
rats following intravenous or gavage administration. Xenobiotica 50: 1-11.
http://dx.doi.org/10.1080/00498254.2019.1683776.
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. http://dx.doi.org/10.1097/EE9.0000000000000095.
Eick. SM: Demicco. E: Valeri. L: Woodruff. TT: Morello-Frosch. R: Horn Thepaksorn. EK: Izano. MA:
Cushing. LI: Wang. Y: Smith. SC: Gao. S: Park. IS: Padula. AM. (2020). Associations between
prenatal maternal exposure to per- and polyfluoroalkyl substances (PFAS) and
polybrominated diphenyl ethers (PBDEs) and birth outcomes among pregnant women in
San Francisco. Environ Health 19: 100-100. http://dx.doi.org/10.1186/sl2940-020-00654-
2.
Elavarasi. SB: Mariam. D: Momeen. MU: Hu. 1: Guin. M. (2019). Effect of fluorination on bandgap,
first and second order hyperpolarizabilities in lithium substituted adamantane: A time
dependent density functional theory. Chem Phys Lett 715: 310-316.
http: / /dx. doi. or g/10.1016/i. cplett 2018.11.034.
EMEA (European Medicines Agency). (2008). Non-clinical guideline on drug induced hepatotoxicity.
(EMEA/CHMP/SWP/150115/2006). London, UK.
http://www.ema.europa.eu/docs/en GB/document library/Scientific guideline/2009/09/
WC500003355.pdf.
Ernst. A: Brix. N: Lauridsen. LLB: Olsen. 1: 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.
http: //dx.doi.org/10.1289 /EHP3567.
Fabrega. F: Kumar. V: Benfenati. E: Schuhmacher. M: Domingo. TL: Nadal. M. (2015). Physiologically
based pharmacokinetic modeling of perfluoroalkyl substances in the human body. Toxicol
Environ Chem 97: 814-827. http://dx.doi.org/10.1080/02772248.2015.1060976.
Fan. S: Wu. Y: Bloom. MS: Lv. 1: Chen. L: Wang. W: Li. Z: Tiang. 0: Bu. L: Shi. 1: Shi. T: Zeng. X: Zhang. L:
Zhang. Z: Yang. B: Dong. G: Feng. W. (2023). Associations of per- and polyfluoroalkyl
substances and their alternatives with bone mineral density levels and osteoporosis
prevalence: A community-based population study in Guangzhou, Southern China. Sci Total
Environ 862: 160617-160617. http://dx.doi.org/10.1016/).scitotenv.2022.160617.
Fan. Y: Lu. C: Li. X: Xu. 0: 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. http://dx.doi.Org/10.1016/i.envpol.2020.115138.
Faqi. AS: Hoberman. A: Lewis. E: Stump. D. (2017). Chapter 9 - developmental and reproductive
toxicology. In A Comprehensive Guide to Toxicology in Nonclinical Drug Development
(Second Edition). Cambridge, Maddschusetts: Academic Press.
http: //dx.doi.org/10.1016/B978-0-12-803620-4.00009-8.
Finken. Ml: van Eijsden. M: Loomans. EM: Vrijkotte. TG: Rotteveel. 1. (2013). Maternal
hypothyroxinemia in early pregnancy predicts reduced performance in reaction time tests
R-12
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
in 5- to 6-year-old offspring. J Clin Endocrinol Metab 98: 1417-1426.
http://dx.doi.org/10.1210/ic.2Q12-3389.
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. http: / /dx.doi.org/10.1016/i.envres.2012.11.006.
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.
http: //dx.doi.org/10.1289 /EHP303.
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. http: / /dx.doi. or g/10.1016/i.envint2019.105324.
Foster. P: Gray. LE. Tr. (2013). Toxic responses of the reproductive system. In CD Klaassen (Ed.),
Casarett & Doull's toxicology: The basic science of poisons (8th ed., pp. 861-906). New York,
NY: McGraw-Hill Education, https: //www.mheducation.com/highered/product/casarett-
doull-s-toxicologv-basic-science-poisons-9th-edition-klaassen/97812 59863745.html.
Friis-Hansen. B. (1961). Body water compartments in children: Changes during growth and related
changes in body composition. Pediatrics 28: 169-181.
Fromme. H: Mosch. C: Morovitz. M: Alba-Alejandre. I: Boehmer. S: Kiranoglu. M: Faber. F: Hannibal.
I: Genzel-Boroviczenv. 0: Koletzko. B: Volkel. W. (2010). Pre- and postnatal exposure to
perfluorinated compounds (PFCs). Environ Sci Technol 44: 7123-7129.
http: / /dx. doi. or g/10.10 21 /es 101184f.
Fruchart. TC: Nierman. MC: Stroes. ES: Kastelein. 11: Duriez. P. (2004). New risk factors for
atherosclerosis and patient risk assessment [Review], Circulation 109: III15-III19.
http: //dx.doi.org/10.1161 /01 .CTR0000131513.33892.5b.
Fu. 1: Gao. Y: Cui. L: Wang. T: Liang. Y: Ou. G: Yuan. B: Wang. Y: Zhang. A: Tiang. G. (2016).
Occurrence, temporal trends, and half-lives of perfluoroalkyl acids (PFAAs) in occupational
workers in China. Sci Rep 6: 38039. http: //dx.doi.org/10.1038/srep38039.
Fuentes. S: Colomina. MT: Vicens. P: Franco-Pons. N: Domingo. TL. (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.
http: / /dx.doi. or g/10.10 9 3 /toxsci /kfml 21.
Funes. SC: Rios. M: Fernandez-Fierro. A: Rivera-Perez. D: Soto. TA: Valbuena. TR: Altamirano-Lagos.
Ml: Gomez-Santander. F: Tara. EL: Zoroquiain. P: Roa. TC: Kalergis. AM: Riedel. CA. (2022).
Female offspring gestated in hypothyroxinemia and infected with human Metapneumovirus
(hMPV) suffer a more severe infection and have a higher number of activated CD8+ T
lymphocytes. Front Immunol 13: 966917. http: //dx.doi.org/10.3389 /fimmu.2022.966917.
Gad. SC. (2015). Cardiovascular toxicology and its evaluation. In MB Abou-Donia (Ed.), Mammalian
Toxicology. West Sussex, United Kingdom: John Wiley & Sons.
Gale. AT. (20111. Continuing Education Course #2: Current Understanding of Hemostasis [Review],
Toxicol Pathol 39: 273-280. http://dx.doi.org/10.1177/0192623310389474.
Gallo. E: Amidei. CB: Barbieri. G: Fabricio. ASC: Gion. M: Pitter. G: Dapra. F: Russo. F: Gregori. D:
Fletcher. T: Canova. C. (2022). Perfluoroalkyl substances and thyroid stimulating hormone
levels in a highly exposed population in the Veneto Region. Environ Res 203: 111794.
http: / /dx. doi. or g/10.1016/j .envres. 2021.111794.
Gao. K. e: Zhuang. T: Liu. X: Fu. 1: Zhang. I: Fu. 1. ie: Wang. L: Zhang. A: Liang. Y: Song. M: Tiang. 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.
http: / /dx. doi. or g/10.10 21 /acs. est 9b0 0 715.
R-13
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Gao. Y: Fu. 1: Cao. H: Wang. Y: Zhang. A: Liang. Y: Wang. T: Zhao. C: Tiang. 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.
http: / /dx. doi. or g/10.10 21 /acs. est 5b0 0 77 8.
Gao. Y: Luo. 1: Zhang. Y: Pan. C: Ren. Y: Zhang. 1: Tian. Y: Cohort. SB. (2022). Prenatal Exposure to
Per- and Polyfluoroalkyl Substances and Child Growth Trajectories in the First Two Years.
Environ Health Perspect 130: 37006. http://dx.doi.org/10.1289/EHP9875.
Gardener. H: Sun. 0: Grandiean. P. (2021). PFAS concentration during pregnancy in relation to
cardiometabolic health and birth outcomes. Environ Res 192: 110287.
http: / /dx. doi. or g/10.1016/i. envres. 2020.110287.
Geng. Y: Faber. KN: de Meijer. YE: Blokzijl. H: Moshage. H. (2021). How does hepatic lipid
accumulation lead to lipotoxicity in non-alcoholic fatty liver disease? [Review], Hepatology
International 15: 21-35. http://dx.doi.org/10.1007/sl2072-020-10121-2.
Genova R. G. arza RF. (2024). Breast Fat Necrosis. In StatPearls. Treasure Island, FL: StatPearls
Publishing LLC. https: //www.ncbi.nlm.nih.gov/books/NBK542191/.
Germolec. PR: Lebrec. H: Anderson. SE: Burleson. GR: Cardenas. A: Corsini. E: Elmore. SE: Kaplan.
BLF: Lawrence. BP: Lehmann. GM: Maier. CC: McHale. CM: Myers. LP: Pallardv. M: Roonev.
AA: Zeise. L: Zhang. L: Smith. MT. (2022). Consensus on the key characteristics of
immunotoxic agents as a basis for hazard identification. Environ Health Perspect 130:
105001. http://dx.doi.org/10.1289/EHP10800.
Getz. GS: Reardon. CA. (2012). Animal models of atherosclerosis [Review], Arterioscler Thromb
Vase Biol 32: 1104-1115. http://dx.doi.org/10.1161 /ATVBAHA.l 11.237693.
Ghisari. M: Long. M: R0ge. DM: Olsen. I: Bonefeld-I0rgensen. 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. http://dx.doi.Org/10.1016/j.envres.2017.01.020.
Gilbert. ME: Sanchez-Huerta. K: Wood. C. (2016). Mild thyroid hormone insufficiency during
development compromises activity-dependent neuroplasticity in the hippocampus of adult
male rats. Endocrinology 157: 774-787. http://dx.doi.org/10.1210/en.2015-1643.
Girum. T: Wasie. A. (2018). Return of fertility after discontinuation of contraception: a systematic
review and meta-analysis [Review], 3: 9. http://dx.doi.org/10.1186/s40834-018-0064-y.
Gleason. I. essie A. Cooper. K..eith R. Klotz. L.udith B. Post. G..loria B. Van Orden. G..eorge New Tersev
Drinking Water Quality Institute (NJDWQI). (2017). Health-based maximum contaminant
level support document: Perfluorooctanoic acid (PFOA): Appendix A. New Jersey Drinking
Water Quality Institute (NJDWQI), Health Effects Subcommittee.
https://www.state.ni.us/dep/watersupply/pdf/pfoa-appendixa.pdf.
Gleason. TA: Post. GB: Fagliano. TA. (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.
http: / /dx. doi. or g/10.1016/i .envres. 2014.10.00 4.
Glynn. A: Berger. U: Bignert. A: Ullah. S: Aune. M: Lignell. S: Darnerud. PP. (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. http://dx.doi.org/10.1021/es3Q1168c.
Goeden. HM: Greene. CW: Tacobus. TA. (2019). A transgenerational toxicokinetic model and its use in
derivation of Minnesota PFOA water guidance. J Expo Sci Environ Epidemiol 29: 183-195.
http://dx.doi.Org/10.1038/s41370-018-0110-5.
Goodrich. TA: Alderete. TL: Baumert. BO: Berhane. K: Chen. Z: Gilliland. FD: Goran. MI: Hu. X: Tones.
DP: Margetaki. K: Rock. S: Stratakis. N: Valvi. D: Walker. PI: Conti. DV: Chatzi. L. (2021).
R-14
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Exposure to Perfluoroalkyl Substances and Glucose Homeostasis in Youth. Environ Health
Perspect 129: 97002. http://dx.doi.org/10.1289/EHP9200.
Goudarzi. H: Mivashita. C: Okada. E: Kashino. I: Chen. CT: Ito. S: Araki. A: Kobavashi. 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.
http ://dx. doi.org/10.1016/i. envint 2017.01.024.
Goudarzi. H: Mivashita. C: Okada. E: Kashino. I: Kobavashi. S: Chen. CT: 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.
http ://dx. doi.org/10.1016/i .envint 2016.05.020.
Graber. TM: Alexander. C: Laumbach. RT: 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.
http://dx.doi.Org/l 0.1038/s41370-018-0096-z.
Grandiean. P: Andersen. EW: Budtz-l0rgensen. E: Nielsen. F: M0lbak. K: Weihe. P: Heilmann. C.
(2012). Serum vaccine antibody concentrations in children exposed to perfluorinated
compounds. JAMA 307: 391-397. http://dx.doi.org/10.1001/iama.2011.2034.
Grandjean. P: Bateson. T. (2021). RE: Benchmark analysis for PFAS immunotoxicity.
Grandiean. P: Heilmann. C: Weihe. P: Nielsen. F: Mogensen. UB: Budtz-l0rgensen. E. (2017a). Serum
vaccine antibody concentrations in adolescents exposed to perfluorinated compounds.
Environ Health Perspect 125: 077018. http://dx.doi.org/10.1289/EHP275.
Grandiean. P: Heilmann. C: Weihe. P: Nielsen. F: Mogensen. UB: Timmermann. A: Budtz-I0rgensen.
E. (2017b). Estimated exposures to perfluorinated compounds in infancy predict attenuated
vaccine antibody concentrations at age 5-years. J Immunotoxicol 14: 188-195.
http: //dx.doi.org/10.1080/1547691 X.2017.1360968.
Grandiean. P: Timmermann. CAG: Kruse. M: Nielsen. F: Vinholt. PI: Boding. L: Heilmann. C: M0lbak.
K. (2020). Severity of COVID-19 at elevated exposure to perfluorinated alkylates. PLoS ONE
15: e0244815. http://dx.doi.org/10.1371/iournal.pone.0244815.
Granum. B: Haug. LS: Namork. E: St0levik. SB: Thomsen. C: Aaberge. IS: van Loveren. H: L0vik. M:
Nvgaard. 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.
http://dx.doi.org/10.3109/1547691X.2012.755580.
Gross. RS: Ghassabian. A: Vandvousefi. S: Messito. Ml: Gao. C: Kannan. K: Trasande. L. (2020).
Persistent organic pollutants exposure in newborn dried blood spots and infant weight
status: A case-control study of low-income Hispanic mother-infant pairs. Environ Pollut
267: 115427. http://dx.doi.Org/10.1016/i.envpol.2020.115427.
Grvnnerup. AG: Lindhard. A: S0rensen. S. (2012). The role of anti-Miillerian hormone in female
fertility and infertility - an overview [Review], Acta Obstet Gynecol Scand 91: 1252-1260.
http: //dx.doi.org/10.1111 /i.l 600-0412.2012.01471 ,x.
Gump. BB: Wu. 0: Dumas. AK: Kannan. K. (2011). Perfluorochemical (PFC) exposure in children:
Associations with impaired response inhibition. Environ Sci Technol 45: 8151-8159.
http ://dx. doi. or g/10.10 21 /es 10 3 712 g.
Guo. I: Zhang. I: Wang. Z: Zhang. L: Oi. X: Zhang. Y: Chang. X: Wu. C: Zhou. Z. (2021). Umbilical cord
serum perfluoroalkyl substance mixtures in relation to thyroid function of newborns:
Findings from Sheyang Mini Birth Cohort Study. Chemosphere 273: 129664.
http://dx.doi.Org/10.1016/i.chemosphere.2021.129664.
Gupta. V: Lee. M. (2011). Central hypothyroidism. Indian J Endocrinol Metab 15: S99-S106.
http://dx.doi.org/10.4103/2230-8210.83337.
R-15
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Guvatt. G: Oxman. AD: Akl. EA: Kunz. R: Vist. G: Brozek. T: Norris. S: Falck-Ytter. Y: Glasziou. P:
DeBeer. H: Taeschke. R: Rind. D: Meerpohl. T: Dahm. P: Schiinemann. HT. (2011). GRADE
guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin
Epidemiol 64: 383-394. http://dx.doi.Org/10.1016/j.jclinepi.2010.04.026.
Gvllenhammar. I: Diderholm. B: Gustafsson. 1: Berger. U: Ridefelt. P: Benskin. TP: 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.
http: / /dx. doi. or g/10.1016/i. envint 2017.12.002.
Gvllenhammar I. G. lvnn A. L.ignell S. A.une M. C..antillana T. F..riden U. L.ampa E. (2017).
Utvardering av samband mellan mammors POP-belastning under graviditets- och
amningsperioden och deras barns halsa. (2215-17-008). Swedish Environmental Protection
Agency.
Hack. M: Klein. NK: Taylor. HG. (1995). Long-term developmental outcomes of low birth weight
infants [Review], Future Child 5: 176-196. http://dx.doi.org/10.2307/1602514.
Haddow. IE: Palomaki. GE: Allan. WC: Williams. TR: Knight. GT: Gagnon. 1: O'Heir. CE: Mitchell. ML:
Hermos. RT: Waisbren. SE: Faix. ID: Klein. RZ. (1999). Maternal thyroid deficiency during
pregnancy and subsequent neuropsychological development of the child. N Engl J Med 341:
549-555. http: //dx.doi.org/10.1056/NETM199908193410801.
Hall. AP: Elcombe. CR: Foster. TR: Harada. T: Kaufmann. W: Knippel. A: Kiittler. K: Malarkev. DE:
Maronpot. RR: Nishikawa. A: Nolte. T: Schulte. A: Strauss. V: York. Ml. (2012). Liver
hypertrophy: A review of adaptive (adverse and non-adverse) changes—Conclusions from
the 3rd International ESTP Expert Workshop [Review], Toxicol Pathol 40: 971-994.
http://dx.doi.org/10.1177/0192623312448935.
Hall. CL: Guo. B: Valentine. AZ: Groom. Ml: Daley. D: Saval. K: Hollis. C. (2019). The validity of the
Strengths and Difficulties Questionnaire (SDQ) for children with ADHD symptoms. PLoS
ONE 14: e0218518. http://dx.doi.org/10.1371/journal.pone.0218518.
Hallberg. L: Hogdahl. AM: Nilsson. L: Rvbo. G. (1966). Menstrual blood loss-a population study.
Variation at different ages and attempts to define normality. Acta Obstet Gynecol Scand 45:
320-351. http://dx.doi.org/10.3109/00016346609158455.
Hamm. MP: Cherry. NM: Chan. E: Martin. TW: Burstyn. I. (2010). Maternal exposure to
perfluorinated acids and fetal growth. J Expo Sci Environ Epidemiol 20: 589-597.
http: / /dx. doi. or g/10.10 3 8 /i es. 2 0 0 9.5 7.
Hammarstrand. S: Takobsson. K: Andersson. E: Xu. Y: Li. Y: Olovsson. M: Andersson. EM. (2021).
Perfluoroalkyl substances (PFAS) in drinking water and risk for polycystic ovarian
syndrome, uterine leiomyoma, and endometriosis: A Swedish cohort study. Environ Int 157:
106819. http: / /dx. doi. or g/10.1016/i .envint 2021.106819.
Hanssen. L: Dudarev. AA: Huber. S: Odland. 10: 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. http://dx.doi.Org/10.1016/i.scitotenv.2013.01.029.
Harada. K: Inoue. K: Morikawa. A: Yoshinaga. T: Saito. N: Koizumi. A. (2005). Renal clearance of
perfluorooctane sulfonate and perfluorooctanoate in humans and their species-specific
excretion. Environ Res 99: 253-261. http://dx.doi.Org/10.1016/i.envres.2004.12.003.
Harbison. RD: Bourgeois. MM: Tohnson. GT. (2015). Hamilton & Hardy's industrial toxicology (6th
ed.). Hoboken, NJ: John Wiley & Sons, Inc.
http://www.wiley.eom/WileyCDA/WileyTitle/productCd-0470929731.subiectCd-
CH50.html.
Hardell. E: Karrman. A: van Bavel. B: Bao. 1: Carlberg. M: Hardell. L. (2014). Case-control study on
perfluorinated alkyl acids (PFAAs) and the risk of prostate cancer. Environ Int 63: 35-39.
http://dx.doi.Org/10.1016/i.envint2013.10.005.
R-16
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Harlow. SD: Hood. MM: Ding. N: Mukherjee. B: Calafat. AM: Randolph. TF. Tr: Gold. EB: Park. SK.
(2021). Per- and polyfluoroalkyl substances and hormone levels during the menopausal
transition. J Clin Endocrinol Metab 106: e4427-e4437.
http: / /dx.doi. or g/10.1210/clinem/dgab476.
Harris. MH: Oken. E: Rifas-Shiman. SL: Calafat. AM: Bellinger. DC: Webster. TF: White. RF: Sagiv. SK.
(2021). Prenatal and childhood exposure to per- and polyfluoroalkyl substances (PFAS) and
child executive function and behavioral problems. Environ Res 202: 111621.
http ://dx. doi. or g/10.1016/i. envres. 2021.111621.
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.
http ://dx. doi. or g/10.1016/i. envint 2018.03.025.
Harris. MH: Rifas-Shiman. SL: Calafat. AM: Ye. X: Mora. AM: Webster. TF: Oken. E: Sagiv. SK. (2017).
Predictors of per- and polyfluoroalkyl substance (PFAS) plasma concentrations in 6-10 year
old American children. Environ Sci Technol 51: 5193-5204.
http ://dx. doi. or g/10.10 21 /acs. est 6b0 5 811.
Harris. NS: Bazvdlo. LAL: Winter. WE. (2012). Coagulation tests: a primer on hemostasis for clinical
chemists. Clinical Laboratory News 2012.
Hartman. TT: 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. http://dx.doi.org/l0.1089/chi.2016.0126.
Haug. LS: Thomsen. C: Becher. G. (2009). A sensitive method for determination of a broad range of
perfluorinated compounds in serum suitable for large-scale human biomonitoring. J
Chromatogr A 1216: 385-393. http://dx.doi.Org/10.1016/i.chroma.2008.10.113.
Haug. LS: Thomsen. C: Brantsaeter. AL: Kvalem. HE: Haugen. M: Becher. G: Alexander. 1: Meltzer. HM:
Knutsen. HK. (2010). Diet and particularly seafood are major sources of perfluorinated
compounds in humans. Environ Int 36: 772-778.
http ://dx. doi. or g/10.1016/i .envint 2010.05.016.
He. X: Tiang. 1: Zhang. XX. (2022). Environmental exposure to low-dose perfluorohexanesulfonate
promotes obesity and non-alcoholic fatty liver disease in mice fed a high-fat diet Environ
Sci Pollut Res Int 29: 49279-49290. http://dx.doi.org/10.1007/sll356-022-19369-7.
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. http://dx.doi.Org/10.1016/j.scitotenv.2017.12.186.
Heffernan. AL: Cunningham. TK: Drage. PS: Avlward. LL: Thompson. K: Viiayasarathv. S: Mueller. TF:
Atkin. SL: Sathvapalan. 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. http://dx.doi.org/10.1016/i.iiheh.2018.07.009.
Henrichs. 1: Ghassabian. A: Peeters. RP: Tiemeier. H. (2013). Maternal hypothyroxinemia and effects
on cognitive functioning in childhood: how and why? [Review], Clin Endocrinol 79: 152-
162. http://dx.doi.org/10.llll/cen.12227.
Heo. II: Lee. TW: Kim. SK: Oh. IE. (2014). Foodstuff analyses show that seafood and water are major
perfluoroalkyl acids (PFAAs) sources to humans in Korea. J Hazard Mater 279: 402-409.
http://dx.doi.Org/10.1016/j.jhazmat2014.07.004.
Hill. AB. (1965). The environment and disease: Association or causation? Proc R Soc Med 58: 295-
300. http://dx.doi.Org/10.1177/003591576505800503.
Hillard. PA. (2014). Menstrual suppression: current perspectives. International Journal of Women's
Health 6: 631-637. http://dx.doi.org/10.2147/TTWH.S46680.
R-17
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Hjermitslev. MH: Long. M: Wielsae. M: Bonefeld-Targensen. EC. (2020). Persistent organic pollutants
in Greenlandic pregnant women and indices of foetal growth: The ACCEPT study. Sci Total
Environ 698: 134118. http ://dx.doi.org/10.1016/i.scitotenv.2019.134118.
Hoffman. K: Webster. TF: Weisskopf. MG: Weinberg. 1: 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.
http ://dx. doi. or g/10.12 89 /ehp. 1001898.
H0isager. FD: Sigvaldsen. A: Andersen. MS: Tuul. A: Nielsen. F: Moller. S: Christesen. HBT: Gr0ntved.
A: Grandiean. P: lensen. TK. (2023). Prenatal and early postnatal exposure to perfluoroalkyl
substances and bone mineral content and density in the Odense Child Cohort Environ Int
181: 108264. http://dx.doi.org/1 0.1016/i.envint2023.108264.
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. http://dx.doi.Org/10.1016/i.idiacomp.2019.02.004.
Hood. A: Hashmi. R: Klaassen. CD. (1999a). Effects of microsomal enzyme inducers on thyroid-
follicular cell proliferation, hyperplasia, and hypertrophy. Toxicol Appl Pharmacol 160: 163-
170. http://dx.doi.org/10.1006/taap.1999.8752.
Hood. A: Klaassen. CD. (2000). Differential effects of microsomal enzyme inducers on in vitro
thyroxine (T4) and triiodothyronine (T3) glucuronidation. Toxicol Sci 55: 78-84.
http ://dx. doi. or g/10.109 3 /toxsci/5 5.1.78.
Hood. A: Liu. YP: Gattone. VH: Klaassen. CD. (1999b). Sensitivity of thyroid gland growth to thyroid
stimulating hormone (TSH) in rats treated with antithyroid drugs. Toxicol Sci 49: 263-271.
http ://dx. doi. or g/10.109 3 /toxsci /49.2.263.
Horn. CL: Morales. AL: Savard. C: Farrell. GC: Ioannou. GN. (2022). Role of cholesterol-associated
steatohepatitis in the development of NASH [Review], Hepatol Commun 6: 12-35.
http: / /dx. doi. or g/10.10 0 2 /hep4.1801.
H0ver. BB: Bonde. TP: T0ttenborg. SS: Ramlau-Hansen. CH: Lindh. C: Pedersen. HS: Toft. G. (2017).
Exposure to perfluoroalkyl substances during pregnancy and child behaviour at 5 to 9 years
of age. Horm Behav 101: 105-112. http://dx.doi.org/10.1016/i.vhbeh.2017.11.007.
Hu. Y: Liu. G: Rood. 1: Liang. L: Bray. GA: de Tonge. L: Coull. B: Furtado. TP: Oi. L: Grandjean. P: Sun. 0.
(2019). Perfluoroalkyl substances and changes in bone mineral density: A prospective
analysis in the POUNDS-LOST study. Environ Res 179: 108775.
http://dx.doi.Org/10.1016/i.envres.2019.108775.
Huang. El: Reichardt. LF. (2001). Neurotrophic: Roles in neuronal development and function. Annu
Rev Neurosci 24: 677-736. http://dx.doi.Org/10.1146/annurev.neuro.24.l.677.
Huang. K: Zhou. W: Fu. 1: Zhang. 0: Teng. Y: Gu. L: Fu. Y: Hu. B: Mei. Y: Zhang. H: Zhang. A: Fu. 1: Tiang.
G. (2023). Linking Transthyretin-Binding Chemicals and Free Thyroid Hormones: In Vitro to
In Vivo Extrapolation Based on a Competitive Binding Model. Environ Sci Technol 57: 9130-
9139. http://dx.doi.org/10.1021/acs.est3c01094.
Huang. M: Tiao. 1: Zhuang. P: Chen. X: Wang. I: Zhang. Y. (2018). Serum polyfluoroalkyl chemicals are
associated with risk of cardiovascular diseases in national US population. Environ Int 119:
37-46. http://dx.doi.Org/10.1016/i.envint.2018.05.051.
Huang. MC: Dzierlenga. AL: Robinson. VG: Waidvanatha. S: Devito. MI: Eifrid. MA: Granville. CA:
Gibbs. ST: Blvstone. CR. (2019a). 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: 645-655. http://dx.doi.Org/10.1016/i.toxrep.2019.06.016.
Huang. 0: Liu. L: Wu. Y: Wang. X: Luo. L: Nan. B: Zhang. 1: Tian. M: Shen. H. (2019b). Seminal plasma
metabolites mediate the associations of multiple environmental pollutants with semen
R-18
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
quality in Chinese men. Environ Int 132: 105066.
http ://dx. doi. or g/10.1016/i. envint 2019.105066.
Huang. R: Chen. 0: Zhang. L: Luo. K: Chen. L: Zhao. S: Feng. L: Zhang. T. (2019c). Prenatal exposure to
perfluoroalkyl and polyfluoroalkyl substances and the risk of hypertensive disorders of
pregnancy. Environ Health 18: 5. http://dx.doi.org/10.1186/sl2940-018-0445-3.
Humblet. 0: Diaz-Ramirez. LG: Balmes. 1R: Pinnev. SM: Hiatt. RA. (2014). Perfluoroalkyl chemicals
and asthma among children 12-19 years of age: NHANES (1999-2008). Environ Health
Perspectl22: 1129-1133. http://dx.doi.org/10.1289/ehp.1306606.
Huo. X: Zhang. L: Huang. R: Feng. L: Wang. W: Zhang. 1: Cohort. SB. (2020). Perfluoroalkyl
substances exposure in early pregnancy and preterm birth in singleton pregnancies: a
prospective cohort study. Environ Health 19: 60. http: //dx.doi.org/10.1186/sl2940-020-
00616-8.
ICH Expert Working Group. (2005). Immunotoxicity studies for human pharmaceuticals S8 11. In
ICH Harmonised Tripartite Guideline.
Idalsoaga. F: Kulkarni. AY: Mousa. OY: Arrese. M: Arab. TP. (2020). Non-alcoholic fatty liver disease
and alcohol-related liver disease: Two intertwined entities. Front Med 7.
http://dx.doi.org/10.3389/fmed.2020.0Q448.
Impinen. A: Longnecker. MP: Nvgaard. UC: London. ST: 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. http://dx.doi.Org/10.1016/i.envint.2018.12.041.
Impinen. A: Nvgaard. UC: Ladrup 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. http://dx.doi. or g/10.1016/i.envres.2017.10.012.
Inoue. K: Ritz. B: Andersen. SL: Ramlau-Hansen. CH: Haver. BB: Bech. BH: Henriksen. TB: Bonefeld-
T0rgensen. EC: Olsen. 1: Liew. Z. (2019). Perfluoroalkyl substances and maternal thyroid
hormones in early pregnancy; findings in the Danish National Birth Cohort Environ Health
Perspect 127: 117002. http://dx.doi.org/l 0.1289/EHP5482.
Ioannou. GN. (2016). The Role of Cholesterol in the Pathogenesis of NASH [Review], Trends
Endocrinol Metab 27: 84-95. http://dx.doi.Org/10.1016/i.tem.2015.ll.008.
IPCS (International Programme on Chemical Safety). (1996). Principles and methods for assessing
direct immunotoxicity associated with exposure to chemicals. Geneva, Switzerland: World
Health Organization, http://www.inchem.org/documents/ehc/ehc/ehcl80.htm.
IPCS (International Programme on Chemical Safety). (2006). Principles for evaluating health risks
in children associated with exposure to chemicals [WHO EHC], (Environmental Health
Criteria 225). Geneva, Switzerland: World Health Organization.
IPCS (International Programme on Chemical Safety). (2012). Guidance for immunotoxicity risk
assessment for chemicals. (Harmonization Project Document No. 10). Geneva, Switzerland:
World Health Organization, https: //apps.who.int/iris/handle/10665 /330098.
Irizarrv. L. (2014). Thyroid hormone toxicity [Supplemental Data], Endocrinology n/a.
Itoh. S: Araki. A: Mivashita. C: Yamazaki. K: Goudarzi. H: Minatova. M: Ait Bamai. Y: Kobavashi. 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.
http://dx.doi.Org/10.1016/i.envint2019.105139.
Tackson-Browne. MS: Eliot. M: Patti. M: Spanier. AT: Braun. TM. (2020). PFAS (per- and
polyfluoroalkyl substances) and asthma in young children: NHANES 2013-2014. Int J Hyg
Environ Health 229: 113565. http://dx.doi.Org/10.1016/j.ijheh.2020.113565.
R-19
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Tain. R. (2021a). Impact of kidney hyperfiltration on concentrations of selected perfluoroalkyl acids
among US adults for various disease groups. Environ Sci PollutRes Int 28: 21499-21515.
http://dx.doi.Org/10.1007/sll356-020-11855-0.
lain. RB. (2013). Effect of pregnancy on the levels of selected perfluoroalkyl compounds for females
aged 17-39 years: data from National Health and Nutrition Examination Survey 2003-2008.
J Toxicol Environ Health A 76: 409-421.
http: //dx.doi.org/10.1080/15287394.2013.771547.
Tain. RB. (2014). Contribution of diet and other factors to the levels of selected polyfluorinated
compounds: data from NHANES 2003-2008. Int J Hyg Environ Health 217: 52-61.
http://dx.doi.Org/10.1016/i.iiheh.2013.03.008.
Tain. RB. (2019). Synergistic impact of co-exposures to toxic metals cadmium, lead, and mercury
along with perfluoroalkyl substances on the healthy kidney function. Environ Res 169: 342-
347. http ://dx.doi.org/10.1016/i.envres.2018.11.037.
Tain. 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. http://dx.doi.org/10.1016/i.envpol.2020.115207.
Tain. RB. (2021b). Perfluoroalkyl acids and their isomers, diabetes, anemia, and albuminuria:
Variabilities with deteriorating kidney function. Ecotoxicol Environ Saf 208: 111625.
http://dx.doi.Org/10.1016/j.ecoenv.2020.111625.
Tain. RB: Ducatman. A. (2018). Associations between lipid/lipoprotein levels and perfluoroalkyl
substances among US children aged 6-11 years. Environ Pollut 243: 1-8.
http://dx.doi.Org/10.1016/i.envpol.2018.08.060.
Tain. RB: Ducatman. A. (2019a). Dynamics of associations between perfluoroalkyl substances and
uric acid across the various stages of glomerular function. Environ Sci Pollut Res Int 26:
12425-12434. http://dx.doi.org/10.1007/sll356-019-04666-5.
lain. RB: Ducatman. A. (2019b). 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. http://dx.doi.Org/10.1016/i.envres.2019.04.034.
lain. RB: Ducatman. A. (2019c). Perfluoroalkyl substances follow inverted U-shaped distributions
across various stages of glomerular function: Implications for future research. Environ Res
169: 476-482. http://dx.doi.org/1 0.1016/i.envres.2018.11.033.
Tain. RB: Ducatman. A. (2019d). Roles of gender and obesity in defining correlations between
perfluoroalkyl substances and lipid/lipoproteins. Sci Total Environ 653: 74-81.
http ://dx. doi. or g/10.1016/i. scitotenv. 2018.10.362.
Tain. RB: Ducatman. A. (2019e). 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.
http: //dx.doi.org/10.1097/TOM.00000000000Q1532.
Tain. RB: Ducatman. A. (2022). Serum concentrations of selected perfluoroalkyl substances for US
females compared to males as they age. Sci Total Environ 842: 156891.
http ://dx. doi. or g/10.1016/i .scitotenv. 2022.156891.
Tain. RB: Ducatman. A. (2023). Factors affecting serum PFAS concentrations among US females with
surgically and naturally induced menopause: data from NHANES 2003-2018. Environ Sci
Pollut Res Int 30: 84705-84724. http://dx.doi.org/10.1007/sl 1356-023-28395-v.
Tanis. TA: Rifas-Shiman. SL: Seshasavee. SM: Sagiv. S: Calafat. AM: Gold. PR: Coull. BA: Rosen. CI:
Oken. E: Fleisch. AF. (2021). Plasma concentrations of per- and polyfluoroalkyl substances
and body composition from mid-childhood to early adolescence. J Clin Endocrinol Metab
106: e3760-e3770. http://dx.d0i.0rg/l0.1210/clinem/dgabl87.
R-20
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Teddy. Z: Hartman. TT: Taylor. EV: Poteete. C: Kordas. K. (2017). Prenatal concentrations of
perfluoroalkyl substances and early communication development in British girls. Early Hum
Dev 109: 15-20. http://dx.doi.Org/10.1016/i.earlhumdev.2017.04.004.
Teddy. Z: Tobias. TH: Taylor. EV: Northstone. K: Flanders. WD: Hartman. TT. (2018). Prenatal
concentrations of perfluoroalkyl substances and bone health in British girls at age 17.
Archives of Osteoporosis 13: 84. http: //dx.doi.org/10.1007/sll657-018-0498-5.
Tensen. RC: Andersen. MS: Larsen. PV: Glintborg. D: Dalgard. C: Timmermann. CAG: Nielsen. F:
Sandberg. MB: Andersen. HR: Christesen. HT: Grandiean. P: Tensen. TK. (2020a). 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.
http://dx.doi.org/10.1289 /EHP5184.
Tensen. RC: Glintborg. D: Gade Timmermann. CA: Nielsen. F: Kvhl. HB: Frederiksen. H: Andersson.
AM: Tuul. A: Sidelmann. TT: Andersen. HR: Grandiean. P: Andersen. MS: Tensen. TK. (2020b).
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.
http://dx.doi.Org/10.1016/i.envres.2019.109101.
Tensen. RC: Glintborg. D: Timmermann. CAG: Nielsen. F: Kvhl. HB: Andersen. HR: Grandiean. P:
Tensen. TK: Andersen. M. (2018). Perfluoroalkyl substances and glycemic status in pregnant
Danish women: The Odense Child Cohort Environ Int 116: 101-107.
http: / /dx. doi. or g/10.1016/i. envint 2018.04.010.
Tensen. TK: Andersen. LB: Kvhl. HB: Nielsen. F: Christesen. HT: Grandiean. P. (2015). Association
between perfluorinated compound exposure and miscarriage in Danish pregnant women.
PLoS ONE 10: e0123496. http://dx.doi.org/10.1371/iournal.pone.0123496.
Ti. K: Kim. S: Kho. Y: Paek. D: Sakong. 1: Ha. 1: Kim. S: Choi. K. (2012). Serum concentrations of major
perfluorinated compounds among the general population in Korea: dietary sources and
potential impact on thyroid hormones. Environ Int 45: 78-85.
http: / /dx. doi. or g/10.1016/i .envint 2012.03.007.
Tiang. W: Zhang. Y: Zhu. L: Deng. 1. (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. http://dx.doi.Org/10.1016/i.envint.2013.12.001.
Tin. H: Mao. L: Xie. I: Zhao. M: Bai. X: Wen. 1: Shen. T: Wu. P. (2020a). Poly- and perfluoroalkyl
substance concentrations in human breast milk and their associations with postnatal infant
growth. Sci Total Environ 713: 136417. http://dx.doi.Org/10.1016/j.scitotenv.2019.136417.
Tin. H: Zhang. Y: Tiang. W: Zhu. L: Martin. TW. (2016). Isomer-Specific Distribution of Perfluoroalkyl
Substances in Blood. Environ SciTechnol 50: 7808-7815.
http://dx.doi.org/10.1021/acs.est6b01698.
Tin. R: Mcconnell. R: Catherine. C: Xu. S: Walker. PI: Stratakis. N: Tones. DP: Miller. GW: Peng. C:
Conti. DV: Vos. MB: Chatzi. L. (2020b). Perfluoroalkyl substances and severity of
nonalcoholic fatty liver in Children: An untargeted metabolomics approach. Environ Int
134: 105220. http://dx.doi.Org/10.1016/i.envint2019.l 05220.
Toensen. UN: Vevrand. B: Antignac. TP: Blomberg Tensen. M: Petersen. TH: Marchand. P: Skakkebaek.
NE: Andersson. AM: Le Bizec. B: Targensen. N. (2013). PFOS (perfluorooctanesulfonate) in
serum is negatively associated with testosterone levels, but not with semen quality, in
healthy men. Hum Reprod 28: 599-608. http://dx.doi.org/10.1093/humrep/des425.
Topling. 1: Henry. E: Wiedmeier. SE: Christensen. RD. (2009). Reference ranges for hematocrit and
blood hemoglobin concentration during the neonatal period: Data from a multihospital
health care system. Pediatrics 123: E333-E337. http://dx.doi.org/10.1542/peds.2008-2654.
Jordan. CD: Flood. TG: Laposata. M: Lewandrowski. KB. (1992). Normal reference laboratory values.
N Engl J Med 327: 718-724. http://dx.doi.org/10.1056/NETM199209033271009.
R-21
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
T0rgensen. KT: Specht. 10: Lenters. V: Bach. CC: Rvlander. L: Tonsson. BA: Lindh. CH: Giwercman. A:
Heederik. D: Toft. G: Bonde. TP. (2014). Perfluoroalkyl substances and time to pregnancy in
couples from Greenland, Poland and Ukraine. Environ Health 13: 116.
http://dx.d0i.0rg/l 0.1186/1476-069X-13-116.
Toshi-Barve. S: Kirpich. I: Cave. MC: Marsano. LS: McClain. CI. (2015). Alcoholic, nonalcoholic, and
toxicant-associated steatohepatitis: Mechanistic similarities and differences [Review],
CMGH 1: 356-367. http://dx.d0i.0rg/l0.1016/i.icmgh.2015.05.006.
Toubert. L: Warme. A: Larson. A: Gronhaug. G: Michael. M: Schoffl. V: Burtscher. E: Meyer. N. (2022).
Prevalence of amenorrhea in elite female competitive climbers. Front Sports Act Living 4:
895588. http://dx.doi.org/10.3389/fspor.2022.895588.
Kang. H: Kim. HS: Yoon. YS: Lee. 1: Kho. Y: Lee. 1: Chang. HI: Cho. YH: Kim. YA. (2021). Placental
transfer and composition of perfluoroalkyl substances (PFASs): A Korean birth panel of
parent-infant triads. Toxics 9: 168. http: //dx.doi.org/10.3390 /toxics9070168.
Kang. H: Lee. HK: Moon. HB: Kim. S: Lee. 1: 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.
http: / /dx. doi. or g/10.1016/i. scitotenv. 2018.07.177.
Kang. 0: Gao. F: Zhang. X: Wang. L: Liu. 1: Fu. M: Zhang. S: Wan. Y: Shen. H: Hu. 1. (2020).
Nontargeted identification of per- and polyfluoroalkyl substances in human follicular fluid
and their blood-follicle transfer. Environ Int 139: 105686.
http: / /dx. doi. or g/10.1016/i. envint 2020.105686.
Karlsen. M: Grandiean. 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. http://dx.doi.Org/10.1016/i.reprotox.2016.08.002.
Karrman. A: Domingo. TL: Llebaria. X: Nadal. M: Bigas. E: van Bavel. B: Lindstrom. G. (2010).
Biomonitoring perfluorinated compounds in Catalonia, Spain: Concentrations and trends in
human liver and milk samples. Environ Sci Pollut Res Int 17: 750-758.
http://dx.d0i.0rg/l 0.1007/si 1356-009-0178-5.
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. http://dx.doi.org/10.1289/ehp.9491.
Kashino. I: Sasaki. S: Okada. E: Matsuura. H: Goudarzi. H: Mivashita. C: Okada. E: Ito. YM: Araki. A:
Kishi. R. (2020). Prenatal exposure to 11 perfluoroalkyl substances and fetal growth: A
large-scale, prospective birth cohort study. Environ Int 136: 105355.
http://dx.doi.Org/10.1016/i.envint2019.105355.
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. http://dx.d0i.0rg/l 0.1186/sl 2940-015-0077-9.
Kato. K: Calafat. AM: Needham. LL. (2009). Polyfluoroalkyl chemicals in house dust. Environ Res
109: 518-523. http://dx.d0i.0rg/l 0.1016/i.envres.2009.01.005.
Kawabata. K: Matsuzaki. H: Nukui. S: Okazaki. M: Sakai. A: Kawashima. Y: Kudo. N. (2017).
Perfluorododecanoic acid induces cognitive deficit in adult rats. Toxicol Sci 157: 421-428.
http: / /dx.doi. or g/10.10 9 3 /toxsci /kfxO 5 8.
Khalil. N: Chen. A: Lee. M: Czerwinski. SA: Ebert. TR: Dewitt. TC: 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.
http://dx.doi.org/10.1289/ehp.1307909.
R-22
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Khalil. N: Ducatman. AM: Sinari. S: Billheimer. D: Hu. C: Littau. S: Burgess. TL. (2020). Per- and
polyfluoroalkyl substance and cardio metabolic markers in firefighters. J Occup Environ
Med 62: 1076-1081. http: //dx.doi.org/10.1097/TOM.0000000000002Q62.
Khalil. N: Ebert. TR: 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.
http ://dx. doi. or g/10.1016/i. envres. 2017.10.014.
Khan. KNM: Hard. GC: Li. X: Alden. CL. (2018). Chapter 11 - Urinary system. In Fundamentals of
Toxicologic Pathology (Third Edition). Cambridge, MA: Academic Press.
http: / /dx.doi.org/10.1016 /B978-0-12-809841 -7.00011 -3.
Kielsen. K: Shamim. Z: Ryder. LP: Nielsen. F: Grandjean. P: Budtz-largensen. E: Heilmann. C. (2016).
Antibody response to booster vaccination with tetanus and diphtheria in adults exposed to
perfluorinated alkylates. J Immunotoxicol 13: 270-273.
http: / /dx.doi.org/10.3109 /I547691 X.201 5.1 067259.
Kim. DH: Kim. UT: Kim. HY: Choi. SD: Oh. IE. (2016a). Perfluoroalkyl substances in serum from South
Korean infants with congenital hypothyroidism and healthy infants - Its relationship with
thyroid hormones. Environ Res 147: 399-404.
http ://dx. doi. or g/10.1016/i .envres. 2016.02.037.
Kim. HY: Kim. KN: Shin. CH: Lim. YH: Kim. II: Kim. BN: Hong. YC: Lee. YA. (2020a). The relationship
between perfluoroalkyl substances concentrations and thyroid function in early childhood:
A prospective cohort study. Thyroid 30: 1556-1565.
http ://dx. doi. or g/10.10 89 /thv. 2019.0436.
Kim. K: Bennett. DH: Calafat. AM: Hertz-Picciotto. I: Shin. HM. (2020b). Temporal trends and
determinants of serum concentrations of per- and polyfluoroalkyl substances among
Northern California mothers with a young child, 2009-2016. Environ Res 186: 109491.
http://dx.doi.Org/10.1016/j.envres.2020.109491.
Kim. Ml: Moon. S: Oh. BC: Tung. D: Ti. K: Choi. K: Park. YT. (2018a). Association between
perfluoroalkyl substances exposure and thyroid function in adults: A meta-analysis. PLoS
ONE 13: e0197244. http://dx.doi.org/10.1371/iournal.pone.0197244.
Kim. 01: Kim. S: Park. EY: Oh. IK: Tung. SK: Park. S: Hong. S: Teon. HL: Kim. HI: Park. B: Park. B: Kim. S:
Kim. B. (2023). Exposure to serum perfluoroalkyl substances and biomarkers of liver
function: The Korean national environmental health survey 2015-2017. Chemosphere 322:
138208. http://dx.doi.Org/10.1016/i.chemosphere.2023.138208.
Kim. S: Choi. K: Ti. K: Seo. I: Kho. Y: Park. I: Kim. S: Park. S: Hwang. I: Teon. I: Yang. H: Giesv. TP.
(2011a). Trans-placental transfer of thirteen perfluorinated compounds and relations with
fetal thyroid hormones. Environ Sci Technol 45: 7465-7472.
http ://dx. doi. or g/10.10 21 /es2 0 2 40 8a.
Kim. ST: Heo. SH: Lee. PS: Hwang. IG: Lee. YB: Cho. HY. (2016b). Gender differences in
pharmacokinetics and tissue distribution of 3 perfluoroalkyl and polyfluoroalkyl substances
in rats. Food Chem Toxicol 97: 243-255. http://dx.doi.Org/10.1016/i.fct.2016.09.017.
Kim. ST: Shin. H: Lee. YB: Cho. HY. (2018b). Sex-specific risk assessment of PFHxS using a
physiologically based pharmacokinetic model. Arch Toxicol 92: 1113-1131.
http: / /dx.doi.org/10.1007/s00204-017-2116-5.
Kim. SK: Kannan. K. (2007). Perfluorinated acids in air, rain, snow, surface runoff, and lakes:
Relative importance of pathways to contamination of urban lakes. Environ Sci Technol 41:
8328-8334. http://dx.doi.org/10.1021/es072107t.
Kim. SK: Lee. KT: Kang. CS: Tao. L: Kannan. K: Kim. KR: Kim. CK: Lee. IS: Park. PS: Yoo. YW: Ha. TY:
Shin. YS: Lee. TH. (2011b). Distribution of perfluorochemicals between sera and milk from
the same mothers and implications for prenatal and postnatal exposures. Environ Pollut
159: 169-174. http://dx.doi.Org/10.1016/i.envpol.2010.09.008.
R-23
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Kim. YR: White. N: Braunig. T: Viiayasarathv. S: Mueller. TF: Knox. CL: Harden. FA: Pacella. R: Toms.
LL. (2020c). Per- and poly-fluoroalkyl substances (PFASs) in follicular fluid from women
experiencing infertility in Australia. Environ Res 190: 109963.
http://dx.doi.Org/10.1016/j.envres.2020.109963.
Kirk. AB: Destefano. A: Martin. A: Kirk. KC: Martin. CF. (2023). A New Interpretation of Relative
Importance on an Analysis of Per and Polyfluorinated Alkyl Substances (PFAS) Exposures
on Bone Mineral Density. Int J Environ Res Public Health 20: 4539.
http ://dx. doi. or g/10.3 3 9 0 /ii erph2 0 0 5 45 3 9.
Kodavanti. UP: Russell. TC: Costa. PL. (2015). Rat models of cardiometabolic diseases: baseline
clinical chemistries, and rationale for their use in examining air pollution health effects.
Inhal Toxicol 27: 2-13. http://dx.doi.org/10.3109/08958378.2014.954166.
Koponen. 1: Winkens. K: Airaksinen. R: Berger. U: Vestergren. R: Cousins. IT: Karvonen. AM:
Pekkanen. 1: Kiviranta. H. (2018). Longitudinal trends of per- and polyfluoroalkyl
substances in children's serum. Environ Int 121: 591-599.
http://dx.doi.Org/10.1016/j.envint2018.09.006.
Korevaar. TIM: Muetzel. R: Medici. M: Chaker. L: Taddoe. VWV: de Riike. YB: Steegers. EAP: Visser.
TT: White. T: Tiemeier. H: Peeters. RP. (2016). Association of maternal thyroid function
during early pregnancy with offspring IQ and brain morphology in childhood: A population-
based prospective cohort study. Lancet Diabetes Endocrinol 4: 35-43.
http://dx.doi.org/10.1016/S2213-8587ri5100327-7.
Koshv. TT: Attina. TM: Ghassabian. A: Gilbert. 1: 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.
http: / /dx. doi. or g/10.1016/i. envint 2017.08.003.
Koskela. A: Ducatman. A: Schousboe. IT: Nahhas. RW: Khalil. N. (2022). Perfluoroalkyl Substances
and Abdominal Aortic Calcification. J Occup Environ Med 64: 287-294.
http: //dx.doi.org/10.1097/TOM.0000000000002479.
Kudo. N: Katakura. M: Sato. Y: Kawashima. Y. (2002). Sex hormone-regulated renal transport of
perfluorooctanoic acid. Chem Biol Interact 139: 301-316.
http://dx.doi.org/10.1016/S0009-2797r02100006-6.
Kvalem. HE: Nvgaard. UC: Ladrup 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.
http://dx.doi.Org/10.1016/i.envint2019.105259.
Kwon. EI: Shin. IS: Kim. BM: Shah-Kulkarni. S: Park. H: Kho. YL: Park. EA: Kim. YT: Ha. EH. (2016).
Prenatal exposure to perfluorinated compounds affects birth weight through GSTM1
polymorphism. J Occup Environ Med 58: el98-e205.
http: //dx.doi.org/10.1097/TOM.000000000000Q739.
Laitinen. TA: Koponen. 1: Koikkalainen. 1: Kiviranta. H. (2014). Firefighters' exposure to
perfluoroalkyl acids and 2-butoxyethanol present in firefighting foams. Toxicol Lett 231:
227-232. http://dx.doi.Org/10.1016/i.toxlet2014.09.007.
Lakind. IS: Naiman. I: Verner. MA: Leveque. L: Fenton. S. (2023). Per- and polyfluoroalkyl
substances (PFAS) in breast milk and infant formula: A global issue [Review], Environ Res
219: 115042. http://dx.doi.org/10.1016/i.envres.2022.115042.
Lala. V: Zubair. M: Minter. DA. (2023). Liver function tests. [Updated 2023 Jul 30], In StatPearls (pp.
1-13). Treasure Island, FL: StatPearls Publishing.
https://www.ncbi.nlm.nih.gov/books/NBK482489/.
R-24
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Lau. C: Anitole. K: Hodes. C: Lai. D: Pfahles-Hutchens. A: Seed. 1. (2007). Perfluoroalkyl acids: a
review of monitoring and toxicological findings [Review], Toxicol Sci 99: 366-394.
http://dx.doi.org/10.10 9 3 /toxsci /kfml 2 8.
Lau. C: Thibodeaux. TR: Hanson. RG: Rogers. TM: Grey. BE: Stanton. ME: Butenhoff. TL: Stevenson. LA.
(2003). Exposure to perfluorooctane sulfonate during pregnancy in rat and mouse. II:
postnatal evaluation. Toxicol Sci 74: 382-392. http://dx.doi.org/10.1093/toxsci/kfgl22.
Lebeaux. RM: Doherty. BT: Gallagher. LG: Zoeller. RT: Hoofnagle. AN: Calafat. AM: Karagas. MR:
Yolton. K: Chen. A: Lanphear. BP: Braun. TM: Romano. ME. (2020). Maternal serum
perfluoroalkyl substance mixtures and thyroid hormone concentrations in maternal and
cord sera: The HOME Study. Environ Res 185: 109395.
http://dx.doi.Org/10.1016/j.envres.2020.109395.
Leboff. MS: Kaplan. MM: Silva. IE: Larsen. PR. (1982). Bioavailability of thyroid hormones from oral
replacement preparations. Metabolism 31: 900-905. http: / /dx.doi. or g/10.1016/0 0 2 6-
0495C82190179-2.
Lee. ES: Han. S: Oh. IE. (2016). Association between perfluorinated compound concentrations in
cord serum and birth weightusing multiple regression models. Reprod Toxicol 59: 53-59.
http: / /dx. doi. or g/10.1016/i. reprotox. 2015.10.020.
Lee. I: Viberg. H. (2013). A single neonatal exposure to perfluorohexane sulfonate (PFHxS) affects
the levels of important neuroproteins in the developing mouse brain. Neurotoxicology 37:
190-196. http://dx.doi.Org/10.1016/i.neuro.2013.05.007.
Lee. YA: Kim. TH: Tung. HW: Lim. YH: Bae. S: Kho. Y: Hong. YC: Shin. CH: Yang. SW. (2018). The serum
concentrations of perfluoroalkyl compounds were inversely associated with growth
parameters in 2-year old children. Sci Total Environ 628-629: 226-232.
http: / /dx. doi. or g/10.1016/i. scitotenv. 2018.02.050.
Lee. YT: Kim. MK: Bae. I: Yang. TH. (2013). Concentrations of perfluoroalkyl compounds in maternal
and umbilical cord sera and birth outcomes in Korea. Chemosphere 90: 1603-1609.
http://dx.doi.Org/10.1016/i.chemosphere.2012.08.035.
Lenters. V: Portengen. L: Rignell-Hydbom. A: Tonsson. BA: Lindh. CH: Piersma. AH: Toft. G: Bonde.
TP: Heederik. D: Rvlander. L: Vermeulen. R. (2016). Prenatal phthalate, perfluoroalkyl acid,
and organochlorine exposures and term birth weight in three birth cohorts: multi-pollutant
models based on elastic net regression. Environ Health Perspect 124: 365-372.
http: / /dx. doi. or g/10.12 89 /ehp. 1408933.
Leter. G: Consales. C: Eleuteri. P: Uccelli. R: Specht. 10: Toft. G: Moccia. T: Budillon. A: Tonsson. BA:
Lindh. CH: Giwercman. A: Pedersen. HS: Ludwicki. IK: Zviezdai. V: Heederik. D: Bonde. TP:
Spano. M. (2014). Exposure to perfluoroalkyl substances and sperm DNA global methylation
in Arctic and European populations. Environ Mol Mutagen 55: 591-600.
http: / /dx. doi. or g/10.10 0 2 /em. 218 74.
Lewis. RC: Tohns. LE: Meeker. ID. (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. http://dx.doi.Org/10.3390/ijerphl20606098.
Li. H: Hammarstrand. S: Midberg. B: Xu. Y: Li. Y: Olsson. PS: Fletcher. T: Takobsson. K: Andersson.
EM. (2022a). Cancer incidence in a Swedish cohort with high exposure to perfluoroalkyl
substances in drinking water. Environ Res 204: 112217.
http://dx.doi.Org/10.1016/j.envres.2021.112217.
Li. 1: Cai. D: Chu. C: Li. 00: Zhou. Y: Hu. LW: Yang. BY: Dong. GH: Zeng. XW: Chen. D. (2020a).
Transplacental Transfer of Per- and Polyfluoroalkyl Substances (PFASs): Differences
between Preterm and Full-Term Deliveries and Associations with Placental Transporter
mRNA Expression. Environ SciTechnol 54: 5062-5070.
http: / /dx. doi. or g/10.10 21 /acs. est OcO 0 82 9.
R-25
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Li. K: Gao. P: Xiang. P: Zhang. X: Cui. X: Ma. LP. (2017a). Molecular mechanisms of PFOA-induced
toxicity in animals and humans: Implications for health risks [Review], Environ Int 99: 43-
54. http://dx.doi.Org/10.1016/i.envint2016.ll.014.
Li. M: Zeng. XW: Oian. ZM: Vaughn. MG: Sauve. S: Paul. G: Lin. S: Lu. L: Hu. LW: Yang. BY: Zhou. Y:
Oin. XD: Xu. SL: Bao. WW: Zhang. YZ: Yuan. P: Wang. 1: Zhang. C: Tian. YP: Nian. M: Xiao. X:
Fu. C: Dong. GH. (2017b). Isomers of perfluorooctanesulfonate (PFOS) in cord serum and
birth outcomes in China: Guangzhou Birth Cohort Study. Environ Int 102: 1-8.
http: / /dx. doi. or g/10.1016/i. envint 2017.03.006.
Li. N: Liu. Y: Papandonatos. GD: Calafat. AM: Eaton. CB: Kelsev. KT: Cecil. KM: Kalkwarf. HI: Yolton.
K: Lanphear. BP: Chen. A: Braun. TM. (2021a). Gestational and childhood exposure to per-
and polyfluoroalkyl substances and cardiometabolic risk at age 12 years. Environ Int 147:
106344. http: / /dx. doi. or g/10.1016/i .envint 2020.1063 44.
Li. Y: Andersson. A: Xu. Y: Pineda. D: Nilsson. CA: Lindh. CH: Takobsson. K: Fletcher. T. (2022b).
Determinants of serum half-lives for linear and branched perfluoroalkyl substances after
long-term high exposure—A study in Ronneby, Sweden. Environ Int 163: 107198.
http: / /dx. doi. or g/10.1016/i .envint 2022.107198.
Li. Y: Barregard. L: Xu. Y: Scott. K: Pineda. D: Lindh. CH: Takobsson. K: Fletcher. T. (2020b).
Associations between perfluoroalkyl substances and serum lipids in a Swedish adult
population with contaminated drinking water. Environ Health 19: 33.
http://dx.doi.Org/10.1186/sl2940-020-00588-9.
Li. Y: Cheng. Y: Xie. Z: Zeng. F. (2017c). Perfluorinated alkyl substances in serum of the southern
Chinese general population and potential impact on thyroid hormones. Sci Rep 7: 43380.
http: / /dx. doi. or g/10.10 3 8 /srep43 3 80.
Li. Y: Fletcher. T: Mucs. D: Scott. K: Lindh. CH: Tallying. P: Takobsson. K. (2018). Half-lives of PFOS,
PFHxS and PFOA after end of exposure to contaminated drinking water. Occup Environ Med
75: 46-51. http://dx.doi.org/10.1136/oemed-2017-104651.
Li. Y: Shan. Z: Teng. W: Yu. X: Fan. C: Teng. X: Guo. R: Wang. H: Li. T: Chen. Y: Wang. W: Chawinga. M:
Zhang. L: Yang. L: Zhao. Y: Hua. T. (2010). Abnormalities of maternal thyroid function during
pregnancy affect neuropsychological development of their children at 25-30 months. Clin
Endocrinol 72: 825-829. http://dx.doi.Org/10.llll /i.1365-2265.2009.03743.x.
Li. Y: Xu. Y: Fletcher. T: Scott. K: Nielsen. C: Pineda. D: Lindh. CH: Olsson. PS: Andersson. EM:
Takobsson. K. (2021b). Associations between perfluoroalkyl substances and thyroid
hormones after high exposure through drinking water. Environ Res 194: 110647.
http: / /dx. doi. or g/10.1016/j. envres. 2020.110647.
Liang. H: Wang. Z: Miao. M: Tian. Y: Zhou. Y: Wen. S: Chen. Y: Sun. X: Yuan. W. (2020). Prenatal
exposure to perfluoroalkyl substances and thyroid hormone concentrations in cord plasma
in a Chinese birth cohort. Environ Health 19: 127. http://dx.doi.org/10.1186/sl2940-020-
00679-7.
Liang. TL: Tiwari. T: Moro. P: Messonnier. NE: Reingold. A: Sawyer. M: Clark. TA. (2018). Prevention
of pertussis, tetanus, and diphtheria with vaccines in the United States: Recommendations
of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep 67: 1-
44. http://dx.doi.org/10.15585/mmwr.rr6702al.
Liao. 0: Tang. P: Fan. H: Song. Y: Liang. T: Huang. H: Pan. D: Mo. M: Leilei: Lin. M: Chen. T: Wei. H:
Long. I: Shao. Y: Zeng. X: Liu. S: Huang. D: Oiu. X. (2023). Association between maternal
exposure to per- and polyfluoroalkyl substances and serum markers of liver function during
pregnancy in China: A mixture-based approach. Environ Pollut 323: 121348.
http: / /dx. doi. or g/10.1016/i. envpol. 2 023.121348.
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. http://dx.doi.Org/10.1016/i.ecoenv.2020.110589.
R-26
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Liew. Z: Luo. 1: Nohr. EA: Bech. BH: Bossi. R: Arah. OA: Olsen. I. (2020). Maternal plasma
perfluoroalkyl substances and miscarriage: a nested case-control study in the Danish
National Birth Cohort. Environ Health Perspect 128: 47007.
http://dx.doi.org/10.1289/EHP6202.
Liew. Z: Ritz. B: Bach. CC: Asarnow. RF: Bech. BH: Nohr. EA: Bossi. R: Henriksen. TB: Bonefeld-
largensen. EC: Olsen. 1. (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.
http://dx.doi.org/10.1289/EHP2754.
Liew. Z: Ritz. B: Bonefeld-largensen. EC: Henriksen. TB: Nohr. EA: Bech. BH: Fei. C: Bossi. R: von
Ehrenstein. OS: Streia. E: Uldall. P: Olsen. 1. (2014). Prenatal exposure to perfluoroalkyl
substances and the risk of congenital cerebral palsy in children. Am J Epidemiol 180: 574-
581. http://dx.doi.org /10.1093/aie/kwu 179.
Liew. Z: Ritz. B: von Ehrenstein. OS: Bech. BH: Nohr. EA: Fei. C: Bossi. R: Henriksen. TB: Bonefeld-
largensen. EC: Olsen. 1. (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.
http: / /dx. doi. or g/10.12 89 /ehp. 1408412.
Limei. E: Zhang. S: Tiang. X. (2023). Association between perfluoroalkyl substances exposure and
the prevalence of nonalcoholic fatty liver disease in the different sexes: a study from the
National Health and Nutrition Examination Survey 2005-2018. Environ Sci PollutRes Int
30: 44292-44303. http://dx.doi.org/10.1007/sll356-023-25258-4.
Lin. AY: Panchangam. SC: Lo. CC. (2009a). The impact of semiconductor, electronics and
optoelectronic industries on downstream perfluorinated chemical contamination in
Taiwanese rivers. Environ Pollut 157: 1365-1372.
http://dx.doi.Org/10.1016/i.envpol.2008.ll.033.
Lin. CY: Chen. PC: Lin. YC: Lin. LY. (2009b). Association among serum perfluoroalkyl chemicals,
glucose homeostasis, and metabolic syndrome in adolescents and adults. Diabetes Care 32:
702-707. http://dx.doi.org/10.2337/dc08-l 816.
Lin. CY: Lin. LY: Chiang. CK: Wang. WT: 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. http://dx.doi.org/10.1038/aig.2009.707.
Lin. HW: Feng. HX: Chen. L: Yuan. XT: Tan. Z. (2020a). Maternal exposure to environmental
endocrine disruptors during pregnancy is associated with pediatric germ cell tumors.
Nagoya J Med Sci 82: 323-333. http://dx.doi.Org/10.18999/nagjms.82.2.323.
Lin. P: Cardenas. A: Hauser. R: Gold. PR: 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. http://dx.doi.org/10.1016/i.envint2019.05.027.
Lin. PD: Cardenas. A: Hauser. R: Gold. PR: Kleinman. KP: Hivert. MF: Calafat. AM: Webster. TF:
Horton. ES: Oken. E. (2020b). 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.
http://dx.doi.Org/10.1016/i.envint2020.105573.
Lin. PP: Cardenas. A: Hauser. R: Gold. PR: Kleinman. KP: Hivert. MF: Calafat. AM: Webster. TF:
Horton. ES: Oken. E. (2021). Per- and polyfluoroalkyl substances and kidney function:
Follow-up results from the Piabetes Prevention Program trial. Environ Int 148: 106375.
http://dx.doi.Org/10.1016/i.envint2020.106375.
Lin. TW: Chen. MK: Lin. CC: Chen. MH: Tsai. MS: Chan. PC: Hung. KY: Chen. PC. (2020c). Association
between exposure to perfluoroalkyl substances and metabolic syndrome and related
R-27
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
outcomes among older residents living near a Science Park in Taiwan. Int J Hyg Environ
Health 230: 113607. http://dx.doi.Org/10.1016/i.iiheh.2020.113607.
Lind. DV: Priskorn. L: Lassen. TH: Nielsen. F: Kvhl. HB: Kristensen. DM: Christesen. HT: Targensen.
IS: Grandjean. P: Tensen. 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. http://dx.doi.Org/10.1016/i.reprotox.2016.08.019.
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.
http: / / dx.doi.org/10.1007/s00125-013-3126-3.
Lind. PM: Lind. L: Salihovic. S: Ahlstrom. H: Michaelsson. K: Kullberg. 1: Strand. R. (2022). Serum
levels of perfluoroalkyl substances (PFAS) and body composition - A cross-sectional study
in a middle-aged population. Environ Res 209: 112677.
http: / /dx. doi. or g/10.1016/i. envres. 2022.112677.
Linton. MRF: Yancey. PG: Davies. SS: Terome. WG: Linton. EF: Song. WL: Doran. AC: Vickers. KC:
Bovce. A: Chrousos. G: Corpas. E: de Herder. WW: Dhatariva. K: Dungan. K: Holland. 1: Kalra.
S: Kaltsas. G: Kapoor. N: Koch. C.... Wilson. D. P. (2000). The role of lipids and lipoproteins in
atherosclerosis. In LJ De Groot (Ed.), Endotext South Dartmouth, MA: MDText.com, Inc.
https://www.ncbi.nlm.nih.gov/books/NBK343489/.
Litvinov. RI: Weisel. TW. (2017). Role of red blood cells in haemostasis and thrombosis. 12: 176-
183. http: / /dx. doi. o r g/10.1111 /voxs. 12331.
Liu. B: Wei. B: Mo. M: Song. Y: Tang. C: Tang. P: Guo. X: Tan. C: Liu. S: Huang. D: Oiu. X. (2021a).
Exposure to perfluoroalkyl substances in early pregnancy and the risk of hypertensive
disorders of pregnancy: A nested case-control study in Guangxi, China. Chemosphere 288:
132468. http://dx.doi.Org/10.1016/i.chemosphere.2021.132468.
Liu. C: Xu. X: Huo. X. ia. (2014). Anogenital distance and its application in environmental health
research [Review], Environ Sci Pollut Res Int 21: 5457-5464.
http://dx.doi.Org/10.1007/sll356-014-2570-z.
Liu. D: Tang. B: Nie. S: Zhao. N: He. L: Cui. I: Mao. W: Tin. H. (2023). Distribution of per- and poly-
fluoroalkyl substances and their precursors in human blood. J Hazard Mater 441: 129908.
http://dx.doi.Org/10.1016/j.jhazmat2022.129908.
Liu. G: Dhana. K: Furtado. ID: Rood. I: Zong. G: Liang. L: Oi. L: Bray. GA: Deionge. L: Coull. B:
Grandiean. P: Sun. 0. (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. http://dx.doi.org/10.1371/journal.pmed.1002502.
Liu. G: Zhang. B: Hu. Y: Rood. I: Liang. L: Oi. L: Bray. GA: Deionge. L: Coull. B: Grandiean. P: Furtado.
ID: Sun. 0. (2020a). Associations of Perfluoroalkyl substances with blood lipids and
Apolipoproteins in lipoprotein subspecies: the POUNDS-lost study. Environ Health 19: 5.
http://dx.doi.Org/10.1186/sl2940-020-0561-8.
Liu. H: Pan. Y: Tin. S: Li. Y: Zhao. L: Sun. X: Cui. 0: Zhang. B: Zheng. T: Xia. W: Zhou. A: Campana. AM:
Dai. I: Xu. S. (2020b). Associations of per-/polyfluoroalkyl substances with glucocorticoids
and progestogens in newborns. Environ Int 140: 105636.
http: / /dx. doi. or g/10.1016/i. envint 2020.105636.
Liu. I: Li. 1: 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.
http://dx.doi.Org/10.1016/j.envint2011.05.001.
Liu. II: Cui. XX: Tan. YW: Dong. PX: Ou. YO: Li. 00: Chu. C: Wu. LY: Liang. LX: Oin. ST: Zeeshan. M:
Zhou. Y: Hu. LW: Liu. R0: Zeng. XW: Dong. GH: Zhao. XM. (2022a). Per- and perfluoroalkyl
substances alternatives, mixtures and liver function in adults: A community-based
population study in China. Environ Int 163: 107179.
http://dx.doi.Org/10.1016/i.envint2022.107179.
R-28
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Liu. M: Zhang. G: Meng. L: Han. X: Li. Y: Shi. Y: Li. A: Turvk. ME: Zhang. 0: liang. G. (2021b).
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 56: 6144-6151. http://dx.doi.org/10.1021/acs.est.lc02850.
Liu. Y: Li. N: Papandonatos. GD: Calafat. AM: Eaton. CB: Kelsev. KT: Chen. A: Lanphear. BP: Cecil. KM:
Kalkwarf. HI: Yolton. K: Braun. TM. (2020c). Exposure to per- and polyfluoroalkyl substances
and adiposity at age 12years: Evaluating periods of susceptibility. Environ SciTechnol 54:
16039-16049. http://dx.doi.org/10.1021/acs.est.0c06088.
Liu. Y: Zhou. X: Wu. Y: Yang. X: Wang. Y: Li. S: Bai. X: Schlenk. D: Liu. W. (2022b). Exposure and
Blood-Cerebrospinal Fluid Barrier Permeability of PFASs in Neonates. Environ Sci Technol
Lett 9: 64-70. http://dx.doi.org/10.1021/acs.estlett.lc00862.
Loccisano. AE: Campbell. TL. Tr: Andersen. ME: Clewell. HI. 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. http: //dx. doi. or g/10.1016/i.vrtph.2010.12.004.
Long. M: Ghisari. M: Bonefeld-Targensen. EC. (2013). Effects of perfluoroalkyl acids on the function
of the thyroid hormone and the aryl hydrocarbon receptor. Environ Sci Pollut Res Int 20:
8045-8056. http://dx.doi.org/l 0.1007/sl 1356-013-1628-7.
Lopez-Espinosa. Ml: 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.
http: / /dx. doi. or g/10.12 89 /ehp. 1509869.
Lorber. M: Eaglesham. GE: Hobson. P: Toms. LM: Mueller. IF: Thompson. IS. (2015). The effect of
ongoing blood loss on human serum concentrations of perfluorinated acids. Chemosphere
118: 170-177. http://dx.doi.Org/10.1016/i.chemosphere.2014.07.093.
Lumlev. T. (2004). Analysis of complex survey samples. J Stat Softw 9.
http://dx.doi.org/10.18637/jss.v009.i08.
Lumlev. T. (2023). Survey: Analysis of complex survey samples (Version 4.2-1). Retrieved from
https://cran.r-proiectorg/package=survey
Luo. D: Wu. WX: Pan. YA: Du. BB: Shen. Ml: Zeng. LX. (2021). Associations of prenatal exposure to
per- and polyfluoroalkyl substances with the neonatal birth size and hormones in the
growth hormone/insulin-like growth factor axis. Environ Sci Technol 55: 11859-11873.
http: / /dx. doi. or g/10.10 21 /acs. est 1 cO 2 67 0.
Luo. 1: Xiao. I: Gao. Y: Ramlau-Hansen. CH: Toft. G: Li. I: Obel. C: Andersen. SL: Deziel. NC: Tseng. WL:
Inoue. K: Bonefeld-largensen. EC: Olsen. 1: Liew. Z. (2020). Prenatal exposure to
perfluoroalkyl substances and behavioral difficulties in childhood at 7 and 11 years.
Environ Res 191: 110111. http://dx.doi.Org/10.1016/i.envres.2020.110111.
Lvall. 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. http://dx.doi.org/l0.1289/EHP1830.
M. B: S. A: S. Z: A. S. (2024). Management of metabolic-associated fatty liver disease/metabolic
dysfunction-associated steatotic liver disease: from medication therapy to nutritional
interventions. Nutrients 16. http://dx.doi.org/10.3390/nul6142220.
Mackowiak. B: Hodge. 1: Stern. S: Wang. H. (2018). The roles of xenobiotic receptors: Beyond
chemical disposition [Review], Drug Metab Dispos 46: 1361-1371.
http: / /dx. doi. or g/10.112 4 /dmd. 118.081042.
Maekawa. R: Ito. R: Iwasaki. Y: Saito. K: Akutsu. K: Takatori. S: Ishii. R: Kondo. F: Arai. Y: Ohgane. I:
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.
http: / /dx. doi. or g/10.10 0 2 /rmb2.12049.
R-29
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Maisonet. M: Calafat. AM: Marcus. M: Taakkola. IT: 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.
http://dx.doi.org/10.1289/ehp.1408847.
Maisonet. M: Terrell. ML: Mcgeehin. MA: Christensen. KY: Holmes. A: Calafat. AM: Marcus. M.
(2012). Maternal concentrations of polyfluoroalkyl compounds during pregnancy and fetal
and postnatal growth in British girls. Environ Health Perspect 120: 1432-1437.
http ://dx. doi. or g/10.12 89 /ehp. 1003096.
Makris. K: Mousa. C: Cavalier. E. (2022). Alkaline phosphatases: Biochemistry functions, and
measurement [Review], CalcifTissue Int 112: 233-242. http://dx.doi.org/10.1007/s00223-
022-01048-x.
Mamsen. LS: Biorvang. RD: Mucs. D: Vinnars. MT: Papadogiannakis. N: Lindh. CH: Andersen. CY:
Damdimopoulou. P. (2019). Concentrations of perfluoroalkyl substances (PFASs) in human
embryonic and fetal organs from first, second, and third trimester pregnancies. Environ Int
124: 482-492. http://dx.doi.Org/10.1016/i.envint.2019.01.010.
Man. EB: Tones. WS: Holden. RH: Mellits. ED. (1971). Thyroid function in human pregnancy VIII
Retardation of progency aged 7 years, relationships to maternal age and maternal thyroid
function. Am J Obstet Gynecol 111: 905-916.
Manz. MG: Akashi. K: Weissman. IL. (2004). Biology of hematopoietic stem and progenitor cells. In
KG Blume; SJ Forman; FR Appelbaum (Eds.), Thomas' Hematopoietic Cell Transplantation (3
ed.). Maiden, MA: Blackwell Publishing.
Manzano-Salgado. CB: Casas. M: Lopez-Espinosa. MT: Ballester. F: Iniguez. C: Martinez. D: Costa. 0:
Santa-Marina. L: Pereda-Pereda. E: Schettgen. T: Sunver. T: Vriiheid. M. (2017a). Prenatal
exposure to perfluoroalkyl substances and birth outcomes in a Spanish birth cohort.
Environ Int 108: 278-284. http://dx.doi.Org/10.1016/i.envint2017.09.006.
Manzano-Salgado. CB: Casas. M: Lopez-Espinosa. Ml: Ballester. F: Iniguez. C: Martinez. D:
Romaguera. D: Fernandez-Barres. S: Santa-Marina. L: Basterretxea. M: Schettgen. T: Valvi. D:
Vioque. I: Sunver. 1: Vriiheid. M. (2017b). Prenatal exposure to perfluoroalkyl substances
and cardiometabolic risk in children from the Spanish INMA birth cohort study. Environ
Health Perspect 125: 097018. http://dx.d0i.0rg/l 0.1289/EHP1330.
Manzano-Salgado. CB: Granum. B: Lopez-Espinosa. MT: Ballester. F: Iniguez. C: Gascon. M: Martinez.
D: Guxens. M: Basterretxea. M: Zabaleta. C: Schettgen. T: Sunver. I: Vriiheid. 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. http://dx.doi.0rg/lO.lOl6/i.iiheh.2Oi9.O6.OO5.
Mao. W: Hu. 0: Chen. S: Chen. Y: Luo. M: Zhang. Z: Geng. 1: Wu. I: Xu. B: Chen. M. (2020).
Polyfluoroalkyl chemicals and the risk of kidney stones in US adults: A population-based
study. Ecotoxicol Environ Saf 208: 111497.
http ://dx. doi. or g/10.1016 /i. ecoenv. 2 0 2 0.11149 7.
Marks. KT: Cutler. AT: Teddy. Z: Northstone. K: Kato. K: Hartman. TT. (2019a). Maternal serum
concentrations of perfluoroalkyl substances and birth size in British boys. Int J Hyg Environ
Health 222: 889-895. http://dx.doi.org/10.1016/i.iiheh.2019.03.008.
Marks. KT: Teddy. Z: Flanders. WD: Northstone. K: Fraser. A: Calafat. AM: Kato. K: Hartman. TT.
(2019b). 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. http://dx.doi.Org/10.1016/i.reprotox.2019.08.003.
Maronpot. RR. (2014). Liver, hepatocyte - hypertrophy. In National Toxicology Program
Nonneoplastic Lesion Atlas. Research Triangle Park, NC: National Toxicology Program.
https://ntp.niehs.nih.gov/atlas/nnl/hepatobiliary-system/liver/Hepatocyte-Hypertrophy.
R-30
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Marques. ES: Agudelo. T: Kave. EM: Modaresi. SMS: Pfohl. M: Becanova. T: Wei. W: Polunas. M:
Goedken. M: Slitt. AL. (2021). The role of maternal high fat diet on mouse pup metabolic
endpoints following perinatal PFAS and PFAS mixture exposure. Toxicology 462: 152921.
http://dx.doi.Org/10.1016/j.tox.2021.152921.
Martin. CM. (1978). Verbal and spatial encoding of visual stimuli: the effects of sex, hemisphere and
yes-no judgements. Cortex 14: 227-233. http://dx.doi.org/10.1016/s0010-9452r78180048-
3.
Martinsson. M: Nielsen. C: Biork. 1: Rvlander. L: Malmqvist. E: Lindh. C: Rignell-Hvdbom. A. (2020).
Intrauterine exposure to perfluorinated compounds and overweight at age 4: A case-control
study. PLoS ONE 15: e0230137. http://dx.doi.org/10.1371/iournal.pone.0230137.
Matilla-Santander. N: Valvi. D: Lopez-Espinosa. Ml: Manzano-Salgado. CB: Ballester. F: Ibarluzea. 1:
Santa-Marina. L: Schettgen. T: Guxens. M: Sunver. 1: Vriiheid. M. (2017). Exposure to
Perfluoroalkyl Substances and Metabolic Outcomes in Pregnant Women: Evidence from the
Spanish INMA Birth Cohorts. Environ Health Perspect 125: 117004.
http://dx.doi.org/10.1289 /EHP1062.
Matthews. PR: Hosker. TP: Rudenski. AS: Navlor. BA: Treacher. DF: Turner. RC. (1985). Homeostasis
model assessment: Insulin resistance and beta-cell function from fasting plasma glucose and
insulin concentrations in man. Diabetologia 28: 412-419.
http://dx.doi.Org/10.1007/BF00280883.
Mattsson. K: Rignell-Hvdbom. A: Holmberg. S: Thelin. A: Tonsson. BA: Lindh. CH: Sehlstedt. A:
Rvlander. L. (2015). Levels of perfluoroalkyl substances and risk of coronary heart disease:
Findings from a population-based longitudinal study. Environ Res 142: 148-154.
http: / /dx. doi. or g/10.1016/i. envres. 2015.06.033.
McCoy. TA: Bangma. IT: Reiner. TL: Bowden. TA: Schnorr. 1: Slowev. M: O'Learv. T: Guillette. LI:
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. http://dx.doi.Org/10.1016/i.scitotenv.2017.06.137.
Mcmahon. HT: Bolshakov. VY: Tanz. R: Hammer. RE: Siegelbaum. SA: Sudhof. TC. (1996).
Synaptophysin, a major synaptic vesicle protein, is not essential for neurotransmitter
release. Proc Natl Acad Sci USA 93: 4760-4764. http://dx.doi.org/10.1073/pnas.93.10.4760.
MDH (Minnesota Department of Health). (2019). Health Based Guidance for Water: Toxicological
Summary for: Perfluorohexane sulfonate (PFHxS).
https://www.health.state.mn.us/communities/environment/risk/docs/guidance/gw/pfhx
s.pdf.
Mellor. CL: Steinmetz. FP: Cronin. MT. (2016). The identification of nuclear receptors associated
with hepatic steatosis to develop and extend adverse outcome pathways [Review], Crit Rev
Toxicol 46: 138-152. http: / /dx.doi.org/10.3109/10408444.2015.1089471.
Mendez-Sanchez. N: Cruz-Ramon. VC: Ramirez-Perez. PL: Hwang. TP: Barranco-Fragoso. B: Cordova-
Gallardo. I. (2018). New aspects of lipotoxicity in nonalcoholic steatohepatitis.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES 19: 2034.
http://dx.doi.org/10.3390/ijmsl9072034.
Meng. 0: Inoue. K: Ritz. B: Olsen. I: Liew. Z. (2018). Prenatal exposure to perfluoroalkyl substances
and birth outcomes; An updated analysis from the Danish national birth cohort Int J
Environ Res Public Health 15: 1832. http://dx.doi.org/10.3390/iierphl5091832.
Mi. X: Lin. SO: Zhang. XF: Li. IT: Pei. LI: Tin. F: Liao. 0: Xie. LM: Wei. LC: Hao. CI: Zhang. YW: Li. W.
(2022). Maternal perfluorinated compound exposure and risk of early pregnancy loss: A
nested case-control study [Letter], Biomed Environ Sci 35: 174-179.
http://dx.doi.org/10.3967/bes2022.Q26.
Miller. M: Stone. NT: Ballantyne. C: Bittner. V: Criqui. MH: Ginsberg. HN: Goldberg. AC: Howard. WT:
Tacobson. MS: Kris-Etherton. PM: Lennie. TA: Levi. M: Mazzone. T: Pennathur. S. (2011).
R-31
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Triglycerides and cardiovascular disease: A scientific statement from the American Heart
Association. Circulation 123: 2292-2333.
http://dx.doi.Org/l 0.1161 /CIR.ObOl 3e318216072.
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. http://dx.doi.org/10.1016/i.envint2018.03.033.
Mochizuki. K: Yagi. E: Sakaguchi. N: Mochizuki. H: Takabe. S: Kuranuki. S: Suzuki. T: Shimada. M:
Goda. T. (2007). The critical period for thyroid hormone responsiveness through thyroid
hormone receptor isoform alpha in the postnatal small intestine. Biochim Biophys Acta
1770: 609-616. http://dx.doi.org/l 0.1016/i.bbagen.2006.12.011.
Mondal. D: Weldon. RH: Armstrong. BG: Gibson. LI: Lopez-Espinosa. Ml: 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.
http://dx.doi.org/10.1289/ehp.1306613.
Monrov. 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. http://dx.doi.org/10.1016/i.envres.2008.06.001.
Montesinos. MM: Pellizas. CG. (2019). Thyroid Hormone Action on Innate Immunity. Front
Endocrinol (Lausanne) 10: 350. http://dx.doi.org/10.3389/fendo.2019.00350.
Moog. NK: Entringer. S: Heim. C: Wadhwa. PD: Kathmann. N: Buss. C. (2017). Influence of maternal
thyroid hormones during gestation on fetal brain development [Review], Neuroscience 342:
68-100. http://dx.doi.Org/10.1016/i.neuroscience.2015.09.070.
Moon. 1. (2021). Perfluoroalkyl substances (PFASs) exposure and kidney damage: Causal
interpretation using the US 2003-2018 National Health and Nutrition Examination Survey
(NHANES) datasets. Environ Pollut 288: 117707.
http://dx.doi.Org/10.1016/j.envpol.2021.117707.
Mora. AM: Fleisch. AF: Rifas-Shiman. SL: Woo Baidal. TA: Pardo. L: Webster. TF: Calafat. AM: Ye. X:
Oken. E: Sagiv. SK. (2018). Early life exposure to per- and polyfluoroalkyl substances and
mid-childhood lipid and alanine aminotransferase levels. Environ Int 111: 1-13.
http://dx.doi.Org/10.1016/j.envint2017.ll.008.
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. http://dx.doi.org/10.1289/EHP246.
Moreta. C: Tena. MT. (2014). Determination of perfluorinated alkyl acids in corn, popcorn and
popcorn bags before and after cooking by focused ultrasound solid-liquid extraction, liquid
chromatography and quadrupole-time of flight mass spectrometry. J Chromatogr A 1355:
211-218. http://dx.doi.org/10.1016/i.chroma.2014.06.018.
Morgan. RL: Thayer. KA: Bero. L: Bruce. N: Falck-Ytter. Y: Ghersi. D: Guvatt. G: Hooiimans. C:
Langendam. M: Mandrioli. D: Mustafa. RA: Rehfuess. EA: Roonev. AA: Shea. B: Silbergeld. EK:
Sutton. P: Wolfe. MS: Woodruff. TT: Verbeek. TH: Hollowav. AC: Santesso. N: Schiinemann. HI.
(2016). GRADE: Assessing the quality of evidence in environmental and occupational health.
Environ Int 92-93: 611-616. http://dx.doi.Org/10.1016/i.envint2016.01.004.
Morreale de Escobar. G: Ares. S: Berbel. P: Obregon. MT: Escobar del Rev. F. (2008). The changing
role of maternal thyroid hormone in fetal brain development [Review], Semin Perinatol 32:
380-386. http://dx.doi.Org/10.1053/j.semperi.2008.09.002.
Mullur. R: Liu. YY: Brent. GA. (2014). Thyroid hormone regulation of metabolism [Review], Physiol
Rev 94: 355-382. http://dx.doi.org/10.1152/physrev.00030.2013.
Murray. M. (2017). Trafficking and other regulatory mechanisms for organic anion transporting
polypeptides and organic anion transporters that modulate cellular drug and xenobiotic
R-32
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
influx and that are dysregulated in disease. Br J Pharmacol 174: 1908-1924.
http ://dx. doi. or g/10.1111 /bph. 13785.
Nataraian. S: Lipsitz. S. R.: Nietert. PT. (2002). Self-report of high cholesterol - Determinants of
validity in US adults. Am J Prev Med 23: 13-21. http://dx.doi.org/10.1016/s0749-
3797C02100446-4.
Navarro. D: Alvarado. M: Morte. B: Berbel. D: Sesma. 1: Pacheco. P: Morreale de Escobar. G: Bernal. 1:
Berbel. P. (2014). Late maternal hypothyroidism alters the expression of Camk4 in
neocortical subplate neurons: a comparison with Nurrl labeling. Cereb Cortex 24: 2694-
2706. http://dx.doi.org/10.1093/cercor/bhtl29.
Nelson. TW: 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. http://dx.doi.org/10.1289/ehp.0901165.
Nian. M: Li. 00: Bloom. M: Oian. ZM: Svberg. KM: Vaughn. MG: Wang. SO: Wei. 0: Zeeshan. M:
Gurram. N: Chu. C: Wang. 1: Tian. YP: Hu. LW: Liu. KK: Yang. BY: Liu. RO: 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. http://dx.doi.Org/10.1016/i.envres.2019.02.013.
Nigam. PK. (2011). Serum lipid profile: fasting or non-fasting? Indian J Clin Biochem 26: 96-97.
http://dx.d0i.0rg/l 0.1007/sl 2291 -010-0095-x.
NIH (National Institutes of Health). (2020). The PhenX Toolkit: protocol - lipid profile. Available
online at https://www.phenxtoolkitorg/protocols/view/40201
Nilsson. S: Smurthwaite. K: Avlward. LL: Kay. M: Toms. LM: King. L: Marrington. S: Barnes. C: Kirk.
MP: Mueller. TF: Braunig. 1. (2022a). Serum concentration trends and apparent half-lives of
per- and polyfluoroalkyl substances (PFAS) in Australian firefighters. Int J Hyg Environ
Health 246: 114040. http://dx.doi.Org/10.1016/i.iiIieIi.2022.114040.
Nilsson. S: Smurthwaite. K: Avlward. LL: Kay. M: Toms. LM: King. L: Marrington. S: Kirk. MP:
Mueller. TF: Braunig. I. (2022b). Associations between serum perfluoroalkyl acid (PFAA)
concentrations and health related biomarkers in firefighters. Environ Res 215: 114370.
http: / /dx. doi. or g/10.1016/i. envres. 2022.114370.
Nirogi. R: Goval. VK: Tana. S: Pandev. SK: Gothi. A. (2014). What suits best for organ weight analysis:
Review of relationship between organ weight and body / brain weight for rodent toxicity
studies. Int J Pharm Sci Res 5: 1525-1532. http: //dx.doi.org/10.13040/IIPSR.0975-
8232.5C4U 525-32.
Niu. 1: 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.
http://dx.d0i.0rg/l 0.1186/sl 2940-019-0493-3.
NTPWOI (New Jersey Prinking Water Quality Institute). (2017). Public review draft: Health-based
maximum contaminant level support document: Perfluorooctane sulfonate (PFOS) (CAS #:
1763-23-1; Chemical Formula: C8HF1703S). NJPWQI Health Effects Subcommittee.
https://cswab.org/wp-content/uploads/2018/09/New-Tersey-Scientific-Pocumentation-
Supporting-PF0S-13-ppt.pdf.
NLM (National Library of Medicine). (2013). HSPB: Perfluoro-n-nonanoic acid. Available online at
https://pubchem.ncbi.nlm.nih.gov/source/hsdb/804Q
NLM (National Library of Medicine). (2016). HSPB: Perfluorohexanoic acid. Available online at
https://pubchem.ncbi.nlm.nih.gov/source/hsdb/8299
NLM (National Library of Medicine). (2017). HSPB: Perfluorohexanesulfonic acid. Available online
at https://pubchem.ncbi.nlm.nih.gov/source/hsdb/8274
R-33
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Norwegian Environment Agency. (2018). Investigation of sources to PFHxS in the environment.
Munich, Germany: BiPRO GmbH.
https://www.miliodirektoratetno/globalassets/publikasioner/M961 ZM961.pdf.
Noves. PD: Paul-Friedman. KP: Haselman. IT: Barone Jr.. S: Crofton. KM: Gilbert. ME: Hornung. MW:
Laws. SC: Simmons. SO: Stoker. TE: Tietge. IE: Degitz. ST. (2019). Evaluating chemicals for
thyroid disruption: Opportunities and challenges with in vitro testing and adverse outcome
pathway approaches. Environ Health Perspect 127: 95001.
http://dx.doi.org/10.1289/EHP5297.
NTP (National Toxicology Program). (2018a). 28-day evaluation of the toxicity (C06100) of
perfluorohexane sulfonate potassium salt (PFHSKslt) (3871-99-6) on Harlan Sprague-
Dawley rats exposed via gavage. Available online at http: //d0i.0rg/l0.22427 /NTP-DATA-
002-03267-0002-0000-2 (accessed October 13, 2020).
NTP (National Toxicology Program). (2018b). 28-day evaluation of the toxicity (C20615) of
perfluorodecanoic acid (PFDA) (335-76-2) on Harlan Sprague-Dawley rats exposed via
gavage [NTP], U.S. Department of Health and Human Services.
http://dx.doi.Org/10.22427/NTP-DATA-002-02652-0004-0000-l.
NTP (National Toxicology Program). (2018c). TOX-96: 1-Perfluorobutanesulfonic acid (375-73-5),
potassium perfluorohexanesulfonate (3871-99-6), perfluorooctane sulfonate (1763-23-1),
WY-14643 (50892-23-4). Chemical Effects in Biological Systems (CEBS). Available online at
https://manticore.niehs.nih.gov/cebssearch/publication/T0X-96 (accessed August 6,
2018).
NTP (National Toxicology Program). (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://ntp.niehs.nih.gov/ntp/htdocs/st rpts/tox096 508.pdf.
OECD (Organisation for Economic Co-operation and Development). (2015). Working towards a
global emission inventory of PFASS: focus on PFCAS - status quo and the way forward. Paris,
France.
http://www.oecd.org/chemicalsafety/Working%20Towards%20a%20Global%20Emission
%20Inventorv%2 0of%2 OPFASS.pdf.
OECD (Organisation for Economic Co-operation and Development). (2016). Test no. 422: Combined
repeated dose toxicity study with the reproduction/developmental toxicity screening test:
2016 version. In OECD guidelines for the testing of chemicals, Section 4: Health effects.
Paris, France, http://dx.doi.org/10.1787/9789264264403-en.
Oh. I: Bennett. DH: Calafat. AM: Tancredi. D: Roa. PL: Schmidt. RT: Hertz-Picciotto. I: Shin. HM.
(2021). Prenatal exposure to per- and polyfluoroalkyl substances in association with autism
spectrum disorder in the MARBLES study. Environ Int 147: 106328.
http ://dx. doi. or g/10.1016/i. envint 2020.106328.
Oh. 1: Bennett. DH: Tancredi. DT: Calafat. AM: Schmidt. RT: Hertz-Picciotto. I: Shin. HM. (2022).
Longitudinal Changes in Maternal Serum Concentrations of Per- and Polyfluoroalkyl
Substances from Pregnancy to Two Years Postpartum. Environ Sci Technol 56: 11449-
11459. http://dx.doi.org/10.1021/acs.estlc07970.
Oh. RC: Hustead. TR: Ali. SM: Pantsari. MW. (2017). Mildly elevated liver transaminase levels:
Causes and evaluation. Am Fam Physician 96: 709-715.
Ojo. AF: Peng. C: Ng. TC. (2020). Combined effects and toxicological interactions of perfluoroalkyl
and polyfluoroalkyl substances mixtures in human liver cells (HepG2). Environ Pollut 263:
114182. http://dx.doi.Org/10.1016/i.envpol.2020.114182.
R-34
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Ojo. AF: Xia. 0: Peng. C: Ng. TC. (2021). Evaluation of the individual and combined toxicity of
perfluoroalkyl substances to human liver cells using biomarkers of oxidative stress.
Chemosphere 281: 130808. http://dx.doi.Org/10.1016/i.chemosphere.2021.130808.
Okada. E: Sasaki. S: Kashino. I: Matsuura. H: Mivashita. C: Kobavashi. S: Itoh. K: Ikeno. T: Tamakoshi.
A: Kishi. R. (2014). Prenatal exposure to perfluoroalkyl acids and allergic diseases in early
childhood. Environ Int 65: 127-134. http://dx.doi.Org/10.1016/i.envint.2014.01.007.
Olsen. GW: Burris. TM: Ehresman. DT: Froehlich. TW: Seacat. AM: Butenhoff. TL: Zobel. LR. (2007).
Half-life of serum elimination of perfluorooctanesulfonate,perfluorohexanesulfonate, and
perfluorooctanoate in retired fluorochemical production workers. Environ Health Perspect
115: 1298-1305. http: //dx.doi.org/10.1289/ehp.l 0009.
Omoike. OE: Pack. RP: Mamudu. HM: Liu. Y: Strasser. S: Zheng. S: Okoro. 1: Wang. L. (2020).
Association between per and polyfluoroalkyl substances and markers of inflammation and
oxidative stress. Environ Res 196: 110361.
http://dx.doi.Org/10.1016/i.envres.2020.110361.
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.
http://dx.doi.Org/10.1016/i.envres.2020.110329.
Oppi. S: Liischer. TF: Stein. S. (2019). Mouse models for atherosclerosis research- Which is my line?
[Review], Front Cardiovasc Med 6: 46. http://dx.doi.org/10.3389/fcvm.2019.00046.
Osmond. C: Barker. DT. (2000). Fetal, infant, and childhood growth are predictors of coronary heart
disease, diabetes, and hypertension in adult men and women [Review], Environ Health
Perspect 108: 545-553. http: / /dx.doi.org/10.2307/3454545.
Osterman. LE: Poore. RZ: Swarzenski. PW. (2008). The last 1000 years of natural and anthropogenic
low-oxygen bottom-water on the Louisiana shelf, Gulf of Mexico. Marine Micropaleontology
66: 291-303. http://dx.doi.Org/10.1016/j.marmicro.2007.10.005.
Ou. Y: Zeng. X: Lin. S: Bloom. MS: Han. F: Xiao. X: Wang. H: Matala. R: Li. X: Ou. Y: Nie. Z: Dong. G: Liu.
X. (2021). Gestational exposure to perfluoroalkyl substances and congenital heart defects: A
nested case-control pilot study. Environ Int 154: 106567.
http://dx.doi.Org/10.1016/j.envint2021.106567.
Oulhote. Y: Steuerwald. U: Debes. F: Weihe. P: Grandiean. P. (2016). Behavioral difficulties in 7-year
old children in relation to developmental exposure to perfluorinated alkyl substances
[Review], Environ Int 97: 237-245. http://dx.doi.Org/10.1016/i.envint2016.09.015.
Palha. TA. (2002). Transthyretin as a thyroid hormone carrier: Function revisited [Review], Clin
Chem Lab Med 40: 1292-1300. http://dx.doi.org/10.1515/CCLM.2002.223.
Papadopoulou. E: Sabaredzovic. A: Namork. E: Nvgaard. UC: Granum. B: Haug. LS. (2016). Exposure
of Norwegian toddlers to perfluoroalkyl substances (PFAS): The association with
breastfeeding and maternal PFAS concentrations. Environ Int 94: 687-694.
http://dx.doi.Org/10.1016/i.envint2016.07.006.
Papadopoulou. E: Stratakis. N: Basagana. X: Brantsaeter. AL: Casas. M: Fossati. S: Grazuleviciene. R:
Smastuen Haug. L: Heude. B: Maitre. L: Mceachan. RRC: Robinson. 0: Roumeliotaki. T:
Sabido. E: Borras. E: Urquiza. I: Vafeiadi. M: Zhao. Y: Slama. R: Wright. I: Conti. DV: Vriiheid.
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.
http://dx.doi.Org/10.1016/j.envint2021.106853.
Pappan. N: Awosika. AO: Rehman. A. (2024). Dyslipidemia. In B Abai (Ed.), StatPearls. Treasure
Island, FL: StatPearls Publishing LLC. https: //www.ncbi.nlm.nih.gov/books/NBK560891/.
Park. TH: Choi. I: Tun. DW: Han. SW: Yeo. YH: Nguyen. MH. (2019). Low Alanine Aminotransferase
Cut-Off for Predicting Liver Outcomes; A Nationwide Population-Based Longitudinal Cohort
Study. J Clin Med 8. http: //dx.doi.org/10.3390/icm8091445.
R-35
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Patel. 1: Landers. K: Li. H: Mortimer. RH: Richard. K. (2011). Thyroid hormones and fetal
neurological development [Review], J Endocrinol 209: 1-8. http://dx.doi.org/10.1530/TOE-
10-0444.
Pearce. RG: Setzer. RW: Strope. CL: Sipes. NS: Wambaugh. TF. (2017). Httk: R package for high-
throughput toxicokinetics. J StatSoftw 79: 1-26. http://dx.doi.org/10.18637/iss.v079.i04.
Pelch. KE: Reade. A: Kwiatkowski. CF: Merced-Nieves. FM: Cavalier. H: Schultz. K: Wolffe. T:
Varshavskv. 1. (2022). The PFAS-Tox Database: A systematic evidence map of health studies
on 29 per- and polyfluoroalkyl substances. Environ Int 167: 107408.
http: / /dx. doi. or g/10.1016/i. envint 2022.107408.
Pennings. TLA: Tennen. DGT: Nvgaard. UC: Namork. E: Haug. LS: van Loveren. H: Granum. B. (2016).
Cord blood gene expression supports that prenatal exposure to perfluoroalkyl substances
causes depressed immune functionality in early childhood. J Immunotoxicol 13: 173-180.
http://dx.doi.org/10.3109/1547691X.2015.1029147.
Perez. F: Llorca. M: Kock-Schulmever. M: Skrbic. B: Oliveira. LS: da Boit Martinello. K: Al-Dhabi. NA:
Antic. I: Farre. M: Barcelo. D. (2014). Assessment of perfluoroalkyl substances in food items
at global scale. Environ Res 135: 181-189. http://dx.doi.Org/10.1016/i.envres.2014.08.004.
Perez. F: Nadal. M: Navarro-Ortega. A: Fabrega. F: Domingo. TL: Barcelo. D: Farre. M. (2013).
Accumulation of perfluoroalkyl substances in human tissues. Environ Int 59: 354-362.
http: //dx. doi. or g/10.1016/j .envint 2013.06.004.
Perng. W. ei: Rifas-Shiman. SL: Kramer. MS: Haugaard. LK: Oken. E: Gillman. MW: Belfort. MB.
(2016). Early Weight Gain, Linear Growth, and Mid-Childhood Blood Pressure A Prospective
Study in Project Viva. Hypertension 67: 301-308.
http://dx.doi.org/10.1161/HYPERTENSIONAHA.115.06635.
Petersen. KU: Haervig. KK: Flachs. EM: Bonde. TP: Lindh. C: Hougaard. KS: Toft. G: Ramlau-Hansen.
CH: Tattenborg. SS. (2022). Per- and polyfluoroalkyl substances (PFAS) and male
reproductive function in young adulthood; a cross-sectional study. Environ Res 212:
113157. http://dx.doi.org/10.1016/i.envres.2022.113157.
Pettersson. F: Fries. H: Nillius. ST. (1973). Epidemiology of secondary amenorrhea. I. Incidence and
prevalence rates. Am J Obstet Gynecol 117: 80-86.
Pfohl. M: Ingram. L: Marques. E: Auclair. A: Barlock. B: Tamwal. R: Anderson. D: Cummings. BS: Slitt.
AL. (2020). Perfluorooctanesulfonic acid and perfluorohexanesulfonic acid alter the blood
lipidome and the hepatic proteome in a murine model of diet-induced obesity. Toxicol Sci
178: 311-324. http://dx.doi.org/10.1093/toxsci/kfaal48.
Pilo. A: Iervasi. G: Vitek. F: Ferdeghini. M: Cazzuola. F: Bianchi. R. (1990). Thyroidal and peripheral
production of 3,5,3'-triiodothyronine in humans by multicompartmental analysis. Am J
Physiol 258: E715-E726. http://dx.doi.Org/10.1152/aipendo.1990.258.4.E715.
Pirard. C: Dufour. P: Charlier. C. (2020). Background contamination of perfluoralkyl substances in a
Belgian general population. Toxicol Lett 333: 13-21.
http://dx.doi.Org/10.1016/i.toxlet.2020.07.015.
Pitter. G: Zare Teddi. 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.
http: //dx.doi.org/10.1186/sl 2940-020-00656-0.
Poothong. S: Thomsen. C: Padilla-Sanchez. TA: 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.
http: / /dx. doi. or g/10.10 21 /acs. est 7b0 3 2 9 9.
Preston. EV: Webster. TF: Claus Henn. B: Mcclean. MP: Gennings. C: Oken. E: Rifas-Shiman. SL:
Pearce. EN: Calafat. AM: Fleisch. AF: Sagiv. SK. (2020). Prenatal exposure to per- and
polyfluoroalkyl substances and maternal and neonatal thyroid function in the Project Viva
R-36
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Cohort: A mixtures approach. Environ Int 139: 105728.
http ://dx. doi. or g/10.1016/i. envint 2020.105728.
Preston. EV: Webster. TF: Oken. E: Claus Henn. B: Mcclean. MP: 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. http://dx.doi.org/10.1289/EHP2534.
Pritchard. A. (2012). The Strengths and Difficulties Questionnaire hyperactivity-inattention
subscale is more sensitive for the ADHD-combined subtype than other subtypes in 7-9-year-
old school children. Evid Based Ment Health 15:34. http://dx.doi.org/10.1136/ebmental-
2011-100482.
Prol. B: Klein. S: Rennie. C: Andeliia. S. (2024). Respondent Demographics and Contraceptive Use
Patterns in the United States: A National Survey of Family Growth Analysis. Cureus 16:
e53121. http://dx.doi.org/10.7759/cureus.53121.
Oin. XD: Oian. Z: Vaughn. MG: Huang. 1: 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. http://dx.doi.Org/10.1016/i.envpol.2016.02.050.
Ouispe. R: Sweeney. T: Varma. B: Agarwala. A: Michos. ED. (2022). Recent updates in
hypertriglyceridemia management for cardiovascular disease prevention. Curr Atheroscler
Rep 24: 767-778. http: //dx.doi.org/10.1007/sl 1883-022-01052-4.
Radke. E: Wright. Ml: Christensen. K: Lin. CI: Goldstone. AE: Glenn. B: Thayer. K. (2022).
Epidemiology evidence for health effects of 150 per- and polyfluoroalkyl substances: A
systematic evidence map [Review], Environ Health Perspect 130: 96003.
http: //dx.doi.org/10.1289 /EHP11185.
Radke. EG: Glenn. B: Galizia. A: Persad. A: Nachman. R: Bateson. T: Wright. TM: Navas-Acien. A:
Arrovave. WD: Puett. RC: Harville. EW: Pollack. AZ: Burns. IS: Lynch. CD: Sagiv. SK: Stein. C:
Cooper. GS. (2019). Development of outcome-specific criteria for study evaluation in
systematic reviews of epidemiology studies. Environ Int 130: 104884.
http://dx.doi.Org/10.1016/j.envint2019.05.078.
Rahman. ML: Zhang. C: Smarr. MM: Lee. S: Honda. M: Kannan. K: Tekola-Avele. 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.
http://dx.doi.Org/10.1016/j.envint2019.01.027.
Ramh0i. L: Hass. U: Boberg. 1: Scholze. M: Christiansen. S: Nielsen. F: Axelstad. M. (2018).
Perfluorohexane sulfonate (PFHxS) and a mixture of endocrine disrupters reduce thyroxine
levels and cause anti-androgenic effects in rats. Toxicol Sci 163: 579-591.
http: / /dx.doi. or g/10.10 93/toxsci/kfvO 5 5.
Ramhai. L: Hass. U: Gilbert. ME: Wood. C: Svingen. T: Usai. D: Vinggaard. AM: Mandrup. K: Axelstad.
M. (2020). Evaluating thyroid hormone disruption: investigations of long-term
neurodevelopmental effects in rats after perinatal exposure to perfluorohexane sulfonate
(PFHxS). Sci Rep 10: 2672. http://dx.doi.org/10.1038/s41598-020-59354-z.
Rantakokko. P: Mannisto. V: Airaksinen. R: Koponen. I: Viluksela. M: Kiviranta. H: Pihlaiamaki. 1.
(2015). Persistent organic pollutants and non-alcoholic fatty liver disease in morbidly obese
patients: A cohort study. Environ Health 14: 79. http: //dx.doi.org/10.1186/sl2940-015-
0066-z.
Reardon. ATF: Khodavari Moez. E: Dinu. I: Goruk. S: Field. CI: Kinniburgh. DW: Macdonald. AM:
Martin. TW: 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. http ://dx. doi. or g/10.1016/i .envint 2019.04.023.
R-37
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Refetoff. S. (2015). Thyroid hormone serum transport proteins. In KR Feingold; B Anawalt; A Boyce;
G Chrousos; WW de Herder; K Dhatariya; K Dungan; JM Hershman; J Hofland; S Kalra; G
Kaltsas; C Koch; P Kopp; M Korbonits; CS Kovacs; W Kuohung; B Laferrere; M Levy; EA
McGee; R McLachlan; JE Morley; M New; J Purnell; R Sahay; F Singer; MA Sperling; CA
Stratakis; DL Trence; DP Wilson (Eds.), Endotext. South Dartmouth, MA: MDtextcom Inc.
https://www.ncbi.nlm.nih.gov/books/NBK285566/.
Ren. XM: Oin. WP: Cao. LY: Zhang. I: Yang. Y: Wan. B: Guo. LH. (2016). Binding interactions of
perfluoroalkyl substances with thyroid hormone transport proteins and potential
toxicological implications. Toxicology 366-367: 32-42.
http: / /dx. doi. or g/10.1016 /i .to x. 2 016.0 8.011.
Ren. XM: Zhang. YF: Guo. LH: Oin. ZF: Lv. OY: Zhang. LY. (2015). Structure-activity relations in
binding of perfluoroalkyl compounds to human thyroid hormone T3 receptor. Arch Toxicol
89: 233-242. http://dx.doi.Org/10.1007/s00204-014-1258-v.
Ren. Y: Tin. L: Yang. F: Liang. H: Zhang. Z: Du. 1: Song. X: Miao. M: Yuan. W. (2020). Concentrations of
perfluoroalkyl and polyfluoroalkyl substances and blood glucose in pregnant women.
Environ Health 19: 88. http: //dx.doi.org/10.1186/sl2940-020-0064Q-8.
Reyes. L: Manalich. R. (2005). Long-term consequences of low birth weight [Review], Kidney Int
Suppl 68: S107-S111. http://dx.doi.org/l0.1111 /i.l523-1 755.2005.09718.x.
Rivera. TC: Opazo. MC: Hernandez-Armengol. R: Alvarez. 0: Mendoza-Leon. MI: Caamano. E: Gatica.
S: Bohmwald. K: Bueno. SM: Gonzalez. PA: Neunlist. M: Boudin. H: Kalergis. AM: Riedel. CA.
(2024). Transient gestational hypothyroxinemia accelerates and enhances ulcerative colitis-
like disorder in the male offspring. Front Endocrinol (Lausanne) 14: 1269121.
http://dx.doi.org/10.3389/fendo.2023.1269121.
Robinson. PI: Rapoport. SI. (1986). Kinetics of protein binding determine rates of uptake of drugs
by brain. Am J Physiol Regul Integr Comp Physiol 251: R1212-R1220.
http://dx.doi.Org/10.1152/ajpregu.1986.251.6.R1212.
Romano. ME: Gallagher. LG: Eliot. MN: Calafat. AM: Chen. A: Yolton. K: Lanphear. B: Braun. TM.
(2020). Per- and polyfluoroalkyl substance mixtures and gestational weight gain among
mothers in the health outcomes and measures of the environment study. Int J Hyg Environ
Health 231: 113660. http://dx.doi.Org/10.1016/i.iilieli.2020.113660.
Romano. ME: Xu. Y: Calafat. AM: Yolton. K: Chen. A: Webster. GM: Eliot. MN: Howard. CR: Lanphear.
BP: Braun. TM. (2016). Maternal serum perfluoroalkyl substances during pregnancy and
duration of breastfeeding. Environ Res 149: 239-246.
http: / /dx. doi. or g/10.1016/j. envres. 2016.04.03 4.
Rosen. EM: Brantsaeter. AL: Carroll. R: Haug. L: Singer. AB: Zhao. S: Ferguson. KK. (2018). Maternal
Plasma Concentrations of Per- and polyfluoroalkyl Substances and Breastfeeding Duration
in the Norwegian Mother and Child Cohort Environmental Epidemiology 2: e027.
http://dx.doi.Org/10.1097/EE9.0000000000000027.
Rosen. MB: Das. KP: Roonev. I: Abbott. B: Lau. C: Corton. TC. (2017). PPARa-independent
transcriptional targets of perfluoroalkyl acids revealed by transcript profiling. Toxicology
387: 95-107. http://dx.doi.org/l 0.1016/i.tox.2017.05.013.
Rosen. MB: Das. KP: Wood. CR: Wolf. CI: Abbott. BP: Lau. C. (2013). Evaluation of perfluoroalkyl acid
activity using primary mouse and human hepatocytes. Toxicology 308: 129-137.
http://dx.doi.Org/10.1016/i.tox.2013.03.011.
Rosenmai. AK: Ahrens. L: le Godec. T: Lundqvist. I: 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.
http: / /dx. doi. or g/10.10 0 2 /i at. 3 515.
R-38
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Rosskothen-Kuhl. N: Illing. RB. (2014). Gap43 transcription modulation in the adult brain depends
on sensory activity and synaptic cooperation. PLoS ONE 9: e92624.
http://dx.doi.org/10.1371/iournal.pone.0092624.
Rotander. A: Karrman. A: Toms. LM: Kay. M: Mueller. IF: Gomez Ramos. Ml. (2015a). Novel
fluorinated surfactants tentatively identified in firefighters using liquid chromatography
quadrupole time-of-flight tandem mass spectrometry and a case-control approach. Environ
Sci Technol 49: 2434-2442. http: / / dx. do i. o r g /10.10 21 /e s5 0 3 65 3 n.
Rotander. A: Toms. LM: Avlward. L: Kay. M: Mueller. IF. (2015b). Elevated levels of PFOS and PFHxS
in firefighters exposed to aqueous film forming foam (AFFF). Environ Int 82: 28-34.
http://dx.doi.Org/10.1016/i.envint2015.05.005.
Roth. K: Yang. Z: Agarwal. M: Birbeck. 1: Westrick. 1: Lvdic. T: Gurdziel. K: Petriello. MC. (2024).
Exposure of Environ Health Perspect 132: 087007. http: //dx.doi.org/10.1289 /EHP14339.
Roth. RA: Taeschke. H: Luvendvk. TP. (2019). Chapter 13: Toxic responses of the liver. In CD Klaassen
(Ed.), Casarett & Doull's toxicology: The basic science of poisons (9th ed., pp. 719-766). New
York, NY: McGraw Hill.
Rovet. IF. (2005). Children with congenital hypothyroidism and their siblings: Do they really differ?
Pediatrics 115: e52-e57. http://dx.doi.org/10.1542/peds.2004-1492.
Rovet. IF. (2014). The role of thyroid hormones for brain development and cognitive function
[Review], In G Szinnai (Ed.), Paediatric thyroidology (pp. 26-43). Basel, Switzerland: Karger.
http://dx.doi.org/10.1159/000363153.
Ruark. CD: Song. G: Yoon. M: Verner. MA: Andersen. ME: Clewell. HI: Longnecker. MP. (2017).
Quantitative bias analysis for epidemiological associations of perfluoroalkyl substance
serum concentrations and early onset of menopause. Environ Int 99: 245-254.
http: / /dx. doi. or g/10.1016/i. envint 2016.11.030.
Ruffle. B: Vedagiri. U: Bogdan. D: Maier. M: Schwach. C: Murphv-Hagan. C. (2020). Perfluoroalkyl
Substances in U.S. market basket fish and shellfish. Environ Res 190: 9932-9932.
http://dx.doi.Org/10.1016/i.envres.2020.109932.
Rvlander. C: Brustad. M: Falk. H: Sandanger. TM. (2009). Dietary predictors and plasma
concentrations of perfluorinated compounds in a coastal population from northern Norway.
J Environ Public Health 2009: 268219. http://dx.doi.org/l p'l 155/2009/268219.
S. N: MR. B: S. R: R. C: M. S: R. B: F. C: C. L: M. M: F. C: G. S: V. M: E. P. (2023). Natural Products for the
Treatment of Non-Alcoholic Fatty Liver Disease: A Comprehensive Review. Scientia
Pharmaceutica 91. http://dx.doi.org/10.3390/scipharm91040053.
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. http://dx.doi.org/10.1093/aie/kwx332.
Sagiv. SK: Rifas-Shiman. SL: Webster. TF: Mora. AM: Harris. MH: Calafat. AM: Ye. X: Gillman. MW:
Oken. E. (2015). Sociodemographic and perinatal predictors of early pregnancy per- and
polyfluoroalkyl substance (PFAS) concentrations. Environ Sci Technol 49: 11849-11858.
http://dx.doi.org/10.1021/acs.est5b02489.
Salazar-Martinez. E: Romano-Riquer. P: Yanez-Marquez. E: Longnecker. MP: Hernandez-Avila. M.
(2004). Anogenital distance in human male and female newborns: a descriptive, cross-
sectional study. Environ Health 3:8-13. http://dx.doi.org/10.1186/1476-069x-3-8.
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. http://dx.doi.Org/10.1016/i.toxlet2015.10.023.
Salihovic. S: Stubleski. 1: 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. http: / /dx.doi. or g/10.1016/i .envint 2018.04.052.
R-39
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Sanders. IE: Eigenberg. DA: Bracht. LI: Wang. WR: van Zwieten. Ml. (1988). Thyroid and liver
trophic changes in rats secondary to liver microsomal enzyme induction caused by an
experimental leukotriene antagonist (L-649,923). Toxicol Appl Pharmacol 95: 378-387.
http://dx.doi.org/10.1016/0041-008xr881903S6-0.
Sarzo. B: Ballesteros. V: Iniguez. C: Manzano-Salgado. CB: Casas. M: Llop. S: Murcia. M: Guxens. M:
Vriiheid. M: Marina. LS: Schettgen. T: Espada. M: Irizar. A: Fernandez-limenez. N: Ballester.
F: Lopez-Espinosa. MT. (2021). Maternal perfluoroalkyl substances, thyroid hormones, and
DIO genes: A Spanish cross-sectional study. Environ Sci Technol 55: 11144-11154.
http://dx.doi.org/10.1021/acs.estlcQ1452.
Sathyanaravana. S: Beard. L: Zhou. C: Grady. R. (2010). Measurement and correlates of ano-genital
distance in healthy, newborn infants. Int J Androl 33: 317-323.
http://dx.doi.Org/10.llll/i.1365-2605.2009.01044.x.
Schecter. A: Malik-Bass. N: Calafat. AM: Kato. K: Colacino. TA: Gent. TL: Hvnan. LS: Harris. TR: Malla.
S: Birnbaum. L. (2012). Polyfluoroalkyl compounds in Texas children from birth through 12
years of age. Environ Health Perspect 120: 590-594.
http: / /dx. doi. or g/10.12 89 /ehp. 1104325.
Schiinemann. H: Hill. S: Guvatt. G: Akl. EA: Ahmed. F. (2011). The GRADE approach and Bradford
Hill's criteria for causation. J Epidemiol Community Health 65: 392-395.
http://dx.doi.org/10.1136/jech.2010.119933.
Scinicariello. F: Buser. MC: Abadin. HG: Attanasio. R. (2020a). Perfluoroalkyl substances and
anthropomorphic measures in children (ages 3-11 years), NHANES 2013-2014. Environ Res
186: 109518. http: //dx.doi.org/10.1016/i.envres.2020.109518.
Scinicariello. F: Buser. MC: Balluz. L: Gehle. K: Murray. HE: Abadin. HG: Attanasio. R. (2020b).
Perfluoroalkyl acids, hyperuricemia and goutin adults: Analyses of NHANES 2009-2014.
Chemosphere 259: 127446. http://dx.doi.Org/10.1016/i.chemosphere.2020.127446.
Sellers. RS: Mortan. D: Michael. B: Roome. N: Tohnson. IK: Yano. BL: Perry. R: Schafer. K. (2007).
Society of toxicologic pathology position paper: Organ weight recommendations for
toxicology studies [Review], Toxicol Pathol 35: 751-755.
http://dx.doi.Org/10.1080/01926230701595300.
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. http://dx.doi.Org/10.1016/i.envint2018.01.025.
Shah-Kulkarni. S: Kim. BM: Hong. YC: Kim. HS: Kwon. EI: Park. H: Kim. YT: Ha. EH. (2016). Prenatal
exposure to perfluorinated compounds affects thyroid hormone levels in newborn girls.
Environ Int 94: 607-613. http://dx.doi.Org/10.1016/i.envint2016.06.024.
Shapiro. GD: Dodds. L: Arbuckle. TE: Ashley-Martin. 1: 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.
http://dx.doi.Org/10.1016/j.envres.2016.01.040.
Sheng. N: Li. 1: Liu. H: Zhang. A: Dai. I. (2016). Interaction of perfluoroalkyl acids with human liver
fatty acid-binding protein. Arch Toxicol 90: 217-227. http: / /dx. doi. o r g/10.10 0 7 /s0 0 2 0 4-
014-1391-7.
Shi. Y: Yang. L: Li. 1: Lai. I: Wang. Y: Zhao. Y: Wu. Y. (2017). Occurrence of perfluoroalkyl substances
in cord serum and association with growth indicators in newborns from Beijing.
Chemosphere 169: 396-402. http://dx.doi.Org/10.1016/i.chemosphere.2016.ll.050.
Shih. YH: Blomberg. AT: Bind. MA: Holm. D: Nielsen. F: Heilmann. C: Weihe. P: Grandiean. P. (2021).
Serum vaccine antibody concentrations in adults exposed to per- and polyfluoroalkyl
R-40
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
substances: A birth cohort in the Faroe Islands. J Immunotoxicol 18: 85-92.
http://dx.doi.org/10.1080/1547691X.2021.1922957.
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.
http ://dx. doi. or g/10.1016/i. envres. 2020.109514.
Shinwell. ES: Shlomo. M. (2003). Measured length of normal term infants changes over the first two
days of life. J Pediatr Endocrinol Metab 16: 537-540.
http://dx.doi.Org/10.1515/ipem.2003.16.4.537.
Shoaff. 1: Papandonatos. GD: Calafat. AM: Chen. A: Lanphear. BP: Ehrlich. S: Kelsev. KT: Braun. TM.
(2018). Prenatal exposure to perfluoroalkyl substances: Infant birth weight and early life
growth. Environmental Epidemiology 2: eOlO.
http://dx.doi.Org/10.1097/EE9.0000000000000010.
Shrestha. S: Bloom. MS: Yucel. R: Seegal. RF: Rei. R: Mccaffrev. RT: Wu. 0: Kannan. K: Fitzgerald. EF.
(2017). Perfluoroalkyl substances, thyroid hormones, and neuropsychological status in
older adults. Int J Hyg Environ Health 220: 679-685.
http://dx.doi.Org/10.1016/i.iiheh.2016.12.013.
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. http ://dx. doi.org/10.1016/i .envres. 2018.05.019.
Sinisalu. L: Yeung. LWY: Wang. I: Pan. Y: Dai. 1: Hvotylainen. T. (2021). Prenatal exposure to poly-
/per-fluoroalkyl substances is associated with alteration of lipid profiles in cord-blood.
Metabolomics 17: 103. http://dx.doi.org/10.1007/sll306-021-01853-9.
Skogheim. TS: Villanger. GD: Wevde. KVF: Engel. SM: Suren. P: 0ie. MG: Skogan. AH: Biele. G: Zeiner.
P: 0vergaard. KR: Haug. LS: Sabaredzovic. A: Aase. H. (2020). 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. http://dx.doi.Org/10.1016/i.iiheh.2019.10.003.
Skogheim. TS: Wevde. K. jell Vegard F: Aase. H: Engel. SM: Suren. P: Pie. MG: Biele. G: Reichborn-
Kiennerud. T: Brantsaeter. AL: Haug. LS: Sabaredzovic. A: Auveung. B: Villanger. GD. (2021).
Prenatal exposure to per- and polyfluoroalkyl substances (PFAS) and associations with
attention-deficit/hyperactivity disorder and autism spectrum disorder in children. Environ
Res 202: 111692. http://dx.doi.Org/10.1016/j.envres.2021.111692.
Smeltz. M: Wambaugh. IF: Wetmore. BA. (2023). Plasma protein binding evaluations of per- and
polyfluoroalkyl substances for category-based toxicokinetic assessment. Chem Res Toxicol
36: 870-881. http://dx.doi.org/10.1021/acs.chemrestox.3c00003.
Smit. LAM: Lenters. V: H0ver. BB: Lindh. CH: Pedersen. HS: Liermontova. I: Tonsson. BAG: Piersma.
AH: Bonde. TP: 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. http://dx.doi.org/10.llll/all.12605.
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.
http://dx.doi.Org/10.1016/j.envint2018.01.010.
Specht. 10: Hougaard. KS: Spano. M: Bizzaro. D: Manicardi. GC: Lindh. CH: Toft. G: Tonsson. BA:
Giwercman. A: Bonde. TP. (2012). Sperm DNA integrity in relation to exposure to
environmental perfluoroalkyl substances - a study of spouses of pregnant women in three
geographical regions. Reprod Toxicol 33: 577-583.
http://dx.doi.Org/10.1016/i.reprotox.2012.02.008.
R-41
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Spratlen. MT: Perera. FP: Lederman. SA: Rauh. VA: Robinson. M: Kannan. K: Trasande. L: Herbstman.
L (2020a). The association between prenatal exposure to perfluoroalkyl substances and
childhood neurodevelopment. Environ Pollut 263: 114444.
http://dx.doi.Org/10.1016/j.envpol.2020.114444.
Spratlen. MT: Perera. FP: Lederman. SA: Robinson. M: Kannan. K: Herbstman. 1: Trasande. L.
(2020b). The association between perfluoroalkyl substances and lipids in cord blood. J Clin
Endocrinol Metab 105: 43-54. http: //dx.doi.org/10.1210/clinem/dgz024.
Stagnaro-Green. A: Rovet. 1. (2016). Pregnancy: Maternal thyroid function in pregnancy - A tale of
two tails [Comment], Nat Rev Endocrinol 12: 10-11.
http ://dx. doi. or g/10.10 3 8/nrendo. 2015.212.
Stahl. LL: Snyder. BP: Olsen. AR: Kincaid. TM: Wathen. IB: Mccarty. HB. (2014). Perfluorinated
compounds in fish from U.S. urban rivers and the Great Lakes. Sci Total Environ 499: 185-
195. http://dx.doi.Org/10.1016/i.scitotenv.2014.07.126.
Starling. AP: Adgate. TL: Hamman. RF: Kechris. K: Calafat. AM: Dabelea. D. (2019). Prenatal exposure
to per- and polyfluoroalkyl substances and infant growth and adiposity: The healthy start
study. Environ Int 131: 104983. http://dx.doi.Org/10.1016/i.envint.2019.104983.
Starling. AP: Adgate. TL: Hamman. RF: Kechris. K: Calafat. AM: Ye. X: Dabelea. D. (2017).
Perfluoroalkyl substances during pregnancy and offspring weight and adiposity at birth:
Examining mediation by maternal fasting glucose in the healthy start study. Environ Health
Perspect 125: 067016. http://dx.doi.org/10.1289/EHP641.
Starling. AP: Engel. SM: Richardson. DB: Baird. DP: Haug. LS: Stuebe. AM: Klungsovr. K: Harmon. 0:
Becher. G: Thomsen. C: Sabaredzovic. A: Eggesbo. M: Hoppin. TA: Travlos. GS: Wilson. RE:
Trogstad. LI: Magnus. P. er: Longnecker. MP. (2014a). Perfluoroalkyl Substances During
Pregnancy and Validated Preeclampsia Among Nulliparous Women in the Norwegian
Mother and Child Cohort Study. Am J Epidemiol 179: 824-833.
http://dx.doi. or g/10.10 9 3/aj e/kwt43 2.
Starling. AP: Engel. SM: Whitworth. KW: Richardson. DB: Stuebe. AM: Daniels. TL: Haug. LS: Eggesb0.
M: Becher. G: Sabaredzovic. A: Thomsen. C: Wilson. RE: Travlos. GS: Hoppin. TA: Baird. DP:
Longnecker. MP. (2014b). Perfluoroalkyl substances and lipid concentrations in plasma
during pregnancy among women in the Norwegian Mother and Child Cohort Study. Environ
Int 62: 104-112. http://dx.doi.Org/10.1016/i.envint.2013.10.004.
Steenland. K: Barry. V: Savitz. D. (2018). Serum perfluorooctanoic acid and birthweight: an updated
meta-analysis with bias analysis. Epidemiology 29: 765-776.
http://dx.d0i.0rg/l 0.1097/EDE.0000000000000903.
Stein. CR: Ge. Y: Wolff. MS: Ye. X: Calafat. AM: Kraus. T: Moran. TM. (2016a). Perfluoroalkyl
substance serum concentrations and immune response to FluMist vaccination among
healthy adults. Environ Res 149: 171-178. http://dx.doi.Org/10.1016/i.envres.2016.05.020.
Stein. CR: Mcgovern. KT: Paiak. AM: Maglione. PI: Wolff. M. (2016b). Perfluoroalkyl and
polyfluoroalkyl substances and indicators of immune function in children aged 12-19 y:
National Health and Nutrition Examination Survey. Pediatr Res 79: 348-357.
http://dx.doi.org/10.1038/pr.2015.213.
Stein. CR: Savitz. DA. (2011). Serum perfluorinated compound concentration and attention
deficit/hyperactivity disorder in children 5-18 years of age. Environ Health Perspect 119:
1466-1471. http://dx.doi.org/10.1289/elip.1003538.
Stoker. TE: Ferrell. TM: Laws. SC: Cooper. RL: Buckalew. A. (2006). Evaluation of ammonium
perchlorate in the endocrine disruptor screening and testing program's male pubertal
protocol: Ability to detect effects on thyroid endpoints. Toxicology 228: 58-65.
http: / /dx. doi. or g/10.1016 /i .to x. 2 0 0 6.0 8.0 2 6.
Stratakis. N: Conti. DV: Tin. R: Margetaki. K: Valvi. D: Siskos. AP: Maitre. L: Garcia. E: Varo. N: Zhao. Y:
Roumeliotaki. T: Vafeiadi. M: Urquiza. I: Fernandez-Barres. S: Heude. B: Basagana. X: Casas.
R-42
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
M: Fossati. S: Grazuleviciene. R: Andrusaityte. S: Uppal. K: Mceachan. RR: Papadopoulou. E:
Robinson. 0: Haug. LS: Wright. T: Vos. MB: Keun. HC: Vriiheid. M: Berhane. KT: Mcconnell. R:
Chatzi. L. (2020). Prenatal exposure to perfluoroalkyl substances associated with increased
susceptibility to liver injury in children. Hepatology 72: 1758-1770.
http://dx.doi.org/10.10Q2/hep.31483.
Street. ME: Shulhai. AM: Petraroli. M: Patianna. V: Donini. V: Giudice. A: Gnocchi. M: Masetti. M:
Montani. AG: Rotondo. R: Bernasconi. S: Iughetti. L: Esposito. SM: Predieri. B. (2024). The
impact of environmental factors and contaminants on thyroid function and disease from
fetal to adult life: current evidence and future directions [Review], Front Endocrinol
(Lausanne) 15: 1429884. http://dx.doi.org/10.3389/fendo.2024.1429884.
Stubleski. 1: Salihovic. S: Lind. L: Lind. PM: van Bavel. B: Karrman. A. (2016). Changes in serum
levels of perfluoroalkyl substances during a 10-year follow-up period in a large population-
based cohort. Environ Int 95: 86-92. http://dx.doi.Org/10.1016/i.envint.2016.08.002.
Sultan. M: Ben-Ari. Z: Masoud. R: Pappo. 0: Kamari. Y: Safran. M. (2017). Interleukin-la and
Interleukin-ip play a central role in the pathogenesis of fulminant hepatic failure in mice.
PLoS ONE 12: e0184084. http://dx.doi.org/10.1371/iournal.pone.0184084.
Sun. 0: Zong. G: Valvi. D: Nielsen. F: Coull. B: Grandiean. 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. http://dx.doi.org/10.1289/EHP2619.
Sunderland. EM: Hu. XC: Dassuncao. C: Tokranov. AK: Wagner. CC: Allen. TG. (2019). A review of the
pathways of human exposure to poly- and perfluoroalkyl substances (PFASs) and present
understanding of health effects [Review], J Expo Sci Environ Epidemiol 29: 131-147.
http://dx.doi.Org/10.1038/s41370-018-0094-l.
Sundstrom. M: Chang. SC: Noker. PE: Gorman. GS: Hart. TA: Ehresman. DT: Bergman. A: Butenhoff. TL.
(2012). Comparative pharmacokinetics of perfluorohexanesulfonate (PFHxS) in rats, mice,
and monkeys. Reprod Toxicol 33: 441-451.
http://dx.doi.Org/10.1016/i.reprotox.2011.07.004.
Surma. M: Piskula. M: Wiczkowski. W: Zieliriski. H. (2017). The perfluoroalkyl carboxylic acids
(PFCAs) and perfluoroalkane sulfonates (PFSAs) contamination level in spices. European
Food Research and Technology 243: 297-307. http://dx.doi.org/10.1007/s00217-016-
2744-7.
Sweeney. M. (2022). Physiologically based pharmacokinetic (PBPK) modeling of perfluorohexane
sulfonate (PFHxS) in humans. Regul Toxicol Pharmacol 129: 1-10.
Tamber. SS: Bansal. P: Sharma. S: Singh. RB: Sharma. R. (2023). Biomarkers of liver diseases
[Review], Mol Biol Rep 50: 7815-7823. http://dx.doi.org/10.1007/sll033-023-08666-0.
Tao. L. in: Kannan. K: Aldous. KM: Mauer. MP: Eadon. GA. (2008). Biomonitoring of
perfluorochemicals in plasma of New York state personnel responding to the World Trade
Center disaster. Environ Sci Technol 42: 3472-3478. http://dx.doi.org/10.1021/es8000079.
Tetzlaff. CNR: Ramhai. L: Lardenois. A: Axelstad. M: Evrard. B: Chalmel. F: Taxvig. C: Svingen. T.
(2021). Adult female rats perinatally exposed to perfluorohexane sulfonate (PFHxS) and a
mixture of endocrine disruptors display increased body/fat weights without a
transcriptional footprint in fat cells. Toxicol Lett 339: 78-87.
http://dx.doi.Org/10.1016/i.toxlet.2020.12.018.
Thankamonv. A: Pasterski. V: Ong. KK: Acerini. CL: Hughes. IA. (2016). Anogenital distance as a
marker of androgen exposure in humans. Andrology 0: 1-10.
http: / /dx. doi. or g/10.1111 /andr. 12156.
Thomsen. ML: Henriksen. LS: Tinggaard. 1: Nielsen. F: Tensen. TK: Main. KM. (2021). Associations
between exposure to perfluoroalkyl substances and body fat evaluated by DXA and MRI in
109 adolescent boys. Environ Health 20: 73. http: //dx.doi.org/10.1186/sl2940-021-
00758-3.
R-43
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Thoolen. B: Maronpot. RR: Harada. T: Nvska. A: Rousseaux. C: Nolte. T: Malarkev. DE: Kaufmann. W:
Kiittler. K: Deschl. U: Nakae. D: Gregson. R: Vinlove. MP: Brix. AE: Singh. B: Belpoggi. F:
Ward. TM. (2010). Proliferative and nonproliferative lesions of the rat and mouse
hepatobiliary system [Review], Toxicol Pathol 38: 5S-81S.
http://dx.doi.org/10.1177/0192623310386499.
Tian. M..: Reichetzeder. C..: Li. 1..: Hocher. B.. (2019a). Low birth weight, a risk factor for diseases
in later life, is a surrogate of insulin resistance at birth. J Hypertens 37: 2123-2134.
http://dx.doi.Org/10.1097/HTH.0000000000002156.
Tian. Y: Liang. H: Miao. M: Yang. F: Ti. H: Cao. W: Liu. X: Zhang. X: Chen. A: Xiao. H: Hu. H: Yuan. W.
(2019b). Maternal plasma concentrations of perfluoroalkyl and polyfluoroalkyl substances
during pregnancy and anogenital distance in male infants. Hum Reprod 34: 1356-1368.
http://dx.doi.org/10.1093/humrep/dez058.
Tian. Y: Miao. M: Ti. H: Zhang. X: Chen. A: Wang. Z: Yuan. W: Liang. H. (2021). Prenatal exposure to
perfluoroalkyl substances and cord plasma lipid concentrations. Environ Pollut 268:
115426. http: / /dx.doi. or g/10.1016/j.envpol.2020.115426.
Tian. Z: Peter. KT: Gipe. AD: Zhao. H: Hou. F: Wark. DA: Khangaonkar. T: Kolodziei. EP: Tames. CA.
(2020). Suspect and Nontarget Screening for Contaminants of Emerging Concern in an
Urban Estuary. Environ Sci Technol 54: 889-901.
http://dx.doi.org/10.1021/acs.est9b06126.
Timmermann. CA: Budtz-l0rgensen. E: Tensen. TK: Osuna. CE: Petersen. MS: Steuerwald. U: Nielsen.
F: Poulsen. LK: Weihe. P: Grandiean. P. (2017a). Association between perfluoroalkyl
substance exposure and asthma and allergic disease in children as modified by MMR
vaccination. J Immunotoxicol 14: 39-49.
http://dx.doi.org/10.1080/1547691X.2016.1254306.
Timmermann. CA: Budtz-largensen. E: Petersen. MS: Weihe. P: Steuerwald. U: Nielsen. F: Tensen.
TK: Grandjean. P. (2017b). Shorter duration of breastfeeding at elevated exposures to
perfluoroalkyl substances. Reprod Toxicol 68: 164-170.
http: / /dx. doi. or g/10.1016/i. reprotox. 2016.07.010.
Timmermann. CAG: Andersen. MS: Budtz-largensen. E: Bove. H: Nielsen. F: Tensen. RC: Bruun. S:
Husbv. S: Grandjean. P: Tensen. TK. (2022). Pregnancy exposure to perfluoroalkyl
substances and associations with prolactin concentrations and breastfeeding in the Odense
Child Cohort. J Clin Endocrinol Metab 107: e631-e642.
http: / /dx.doi. or g/10.1210 /clinem /dgab6 3 8.
Timmermann. CAG: Tensen. KT: Nielsen. F: Budtz-largensen. E: van Per 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.
http: //dx.doi.org/10.1289 /EHP6517.
Timmermann. CAG: Pedersen. HS: Weihe. P: Bierregaard. P: Nielsen. F: Heilmann. C: Grandiean. 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. http://dx.doi.org/10.1016/i.envres.2021.111712.
Toft. G: Tonsson. BAG: Lindh. CH: Giwercman. A: Spano. M: Heederik. D: Lenters. V: Vermeulen. R:
Rvlander. L: Pedersen. HS: Ludwicki. TK: Zviezdai. V: Bonde. TP. (2012). Exposure to
perfluorinated compounds and human semen quality in arctic and European populations.
Hum Reprod 27: 2532-2540. http://dx.doi.org/10.1093/humrep/desl85.
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.
http://dx.doi.Org/10.1016/j.envint2020.105850.
R-44
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
U.S. EPA (U.S. Environmental Protection Agency). (1991). Guidelines for developmental toxicity risk
assessment. Fed Reg 56: 63798-63826.
U.S. EPA (U.S. Environmental Protection Agency). (1994). Methods for derivation of inhalation
reference concentrations and application of inhalation dosimetry [EPA Report],
(EPA600890066F). Research Triangle Park, NC.
https://cfpub.epa. gov/ncea/risk/recordisplay.cfm?deid=71993&.CFID=51174829&CFTOKE
N=25006317.
U.S. EPA (U.S. Environmental Protection Agency). (1996). Guidelines for reproductive toxicity risk
assessment. Fed Reg 61: 56274-56322.
U.S. EPA (U.S. Environmental Protection Agency). (1998). Health effects test guidelines OPPTS
870.7800 immunotoxicity. (EPA 712-C-98-351). Washington, DC: Prevention, Pesticides and
Toxic Substances, U.S. Environmental Protection Agency.
http://www.epa.gov/ocspp/pubs/frs/publications/Test Guidelines/series870.htm.
U.S. EPA (U.S. Environmental Protection Agency). (2002). A review of the reference dose and
reference concentration processes [EPA Report], (EPA630P02002F). Washington, DC.
https://www.epa.gov/sites/production/files/2014-12/documents/rfd-final.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2005). Guidelines for carcinogen risk assessment
[EPA Report], (EPA630P03001F). Washington, DC.
https://www.epa.gov/sites/production/files/2013-
09/documents/cancer guidelines final 3-25-05.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2006). A framework for assessing health risk of
environmental exposures to children. (EPA/600/R-05/093F). Washington, DC: U.S.
Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=158363.
U.S. EPA (U.S. Environmental Protection Agency). (2011a). 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://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P 100F2QS.txt
U.S. EPA (U.S. Environmental Protection Agency). (2011b). Recommended use of body weight 3/4
as the default method in derivation of the oral reference dose. (EPA100R110001).
Washington, DC. https://www.epa.gov/sites/production/files/2013-
09/documents/recommended-use-of-bw34.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2012). Benchmark dose technical guidance [EPA
Report], (EPA100R12001). Washington, DC: U.S. Environmental Protection Agency, Risk
Assessment Forum, https: //www.epa.gov/risk/benchmark-dose-technical-guidance.
U.S. EPA (U.S. Environmental Protection Agency). (2014a). Estimated fish consumption rates for the
U.S. population and selected subpopulations (NHANES 2003-2010) [EPA Report], (EPA-
820-R-14-002). Washington, DC. https://www.epa.gov/sites/production/files/2015-
01/documents/fish-consumption-rates-2014.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2014b). Guidance for applying quantitative data
to develop data-derived extrapolation factors for interspecies and intraspecies
extrapolation [EPA Report], (EPA/100/R-14/002F). Washington, DC: Risk Assessment
Forum, Office of the Science Advisor, https: //www.epa.gov/sites /production /files/2015-
01/documents/ddef-final.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2016a). Health effects support document for
perfluorooctane sulfonate (PFOS) [EPA Report], (EPA 822-R-16-002). Washington, DC: U.S.
Environmental Protection Agency, Office of Water, Health and Ecological Criteria Division.
https://www.epa.gov/sites/production/files/2016-
05/documents/pfos hesd final 508.pdf.
R-45
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
U.S. EPA (U.S. Environmental Protection Agency). (2016b). Health effects support document for
perfluorooctanoic acid (PFOA) [EPA Report], (EPA 822-R-16-003). Washington, DC: U.S.
Environmental Protection Agency, Office of Water, Health and Ecological Criteria Division.
https://www.epa.gov/sites/production/files/2016-05/documents/pfoa hesd final-
plainpdf.
U.S. EPA (U.S. Environmental Protection Agency). (2016c). The Third Unregulated Contaminant
Monitoring Rule. Available online at https: //www.epa.gov/dwucmr/third-unregulated-
contaminant-monitoring-rule
U.S. EPA (U.S. Environmental Protection Agency). (2018a). Chemistry Dashboard. Washington, DC.
Retrieved from https: //comptox.epa.gov/dashboard
U.S. EPA (U.S. Environmental Protection Agency). (2018b). An umbrella Quality Assurance Project
Plan (QAPP) for PBPK models [EPA Report], (ORD QAPP ID No: B-0030740-QP-1-1).
Research Triangle Park, NC.
U.S. EPA (U.S. Environmental Protection Agency). (2019a). ChemView [Database], Retrieved from
https://chemview.epa.gov/chemview
U.S. EPA (U.S. Environmental Protection Agency). (2019b). CompTox Chemicals Dashboard
[Database], Research Triangle Park, NC. Retrieved from
https://comptox.epa.gov/dashboard
U.S. EPA (U.S. Environmental Protection Agency). (2019c). 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://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=345065.
U.S. EPA (U.S. Environmental Protection Agency). (2020). ORD staff handbook for developing IRIS
assessments (public comment draft) [EPA Report], (EPA600R20137). Washington, DC: U.S.
Environmental Protection Agency, Office of Research and Development, Center for Public
Health and Environmental Assessment.
https://cfpub.epa.gov/ncea/iris drafts/recordisplav.cfm?deid=350086.
U.S. EPA (U.S. Environmental Protection Agency). (2021a). Human health toxicity values for
hexafluoropropylene oxide (HFPO) dimer acid and its ammonium salt (CASRN 13252-13-6
and CASRN 62037-80-3). Also known as "GenX chemicals." Final report [EPA Report], (EPA-
822R-21-010). Washington, DC: U.S. Environmental Protection Agency, Office of Water.
U.S. EPA (U.S. Environmental Protection Agency). (2021b). Human health toxicity values for
perfluorobutane sulfonic acid (CASRN 375-73-5) and related compound potassium
perfluorobutane sulfonate (CASRN 29420-49-3) [EPAReport], (EPA/600/R-20/345F).
Washington, DC: U.S. Environmental Protection Agency, Office of Research and
Development. https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=350888.
U.S. EPA (U.S. Environmental Protection Agency). (2021c). Systematic Review Protocol for the
PFBA, PFHxA, PFHxS, PFNA, and PFDA (anionic and acid forms) IRIS Assessments:
Supplemental Information—Appendix A. (EPA/635/R-19/050).
https://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=345065.
U.S. EPA (U.S. Environmental Protection Agency). (202 Id). Toxicological review of
perfluorobutanoic acid (PFBA) and related compound ammonium perfluorobutanoic acid
(public comment and external review draft, Aug 2021) [EPA Report], (EPA/635/R-
20/424a). Washington, DC: U.S. Environmental Protection Agency, Integrated Risk
Information System. https: //nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P1014IBF.txt
U.S. EPA (U.S. Environmental Protection Agency). (2022a). DRAFT: Toxicological review of
perfluorohexanoic acid [CASRN 307244] and related salts. (EPA/635/R-21/312a).
U.S. EPA (U.S. Environmental Protection Agency). (2022b). IRIS toxicological review of
perfluorobutanoic acid and related salts [EPA Report], (EPA/635/R-22/277Fa).
R-46
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Washington, D.C.: U.S. Environmental Protection Agency :: U.S. EPA.
https://cfpub.epa.gov/ncea/iris/iris documents/documents/toxreviews/0701tr.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2022c). 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.
U.S. EPA (U.S. Environmental Protection Agency). (2023a). IRIS toxicological review of
perfluorodecanoic acid (PFDA) and related salts (public comment and external review
draft). (EPA/635/R-23/056a).
https://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=3 54408.
U.S. EPA (U.S. Environmental Protection Agency). (2023b). Toxicological Review of
Perfluorohexanoic Acid (PFHxA) and Related Salts (Final Report, 2023) [EPA Report],
(EPA/635/R-23/027F).
https://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=357314.
U.S. EPA (U.S. Environmental Protection Agency). (2024a). Final Human health toxicity assessment
for perfluorooctane sulfonic acid (PF0S) and related salts [EPA Report], Washington, DC:
U.S. Environmental Protection Agency, Office of Water, Health and Ecological Criteria
Division.
U.S. EPA (U.S. Environmental Protection Agency). (2024b). Final human health toxicity assessment
for perfluorooctanoic acid (PFOA) and related salts [EPA Report], Washington, DC: U.S.
Environmental Protection Agency, Office of Water, Health and Ecological Criteria Division.
Ullebo. AK: Posserud. M: Heiervang. E: Gillberg. C: Obel. C. (2011). Screening for the attention deficit
hyperactivity disorder phenotype using the strength and difficulties questionnaire. Eur
Child Adolesc Psychiatry 20: 451-458. http://dx.doi.org/10.1007/s00787-011-0198-9.
Vagi. ST: Azziz-Baumgartner. E: Siodin. A: Calafat. AM: Dumesic. D: Gonzalez. L: Kato. K: Silva. MI: 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. http://dx.doi.org/10.1186/1472-6823-14-86.
Valenti. L.. Pelusi. S... Bianco. C... Ceriotti. F... Berzuini. A... Iogna Prat. L... Trotti. R... Malvestiti. F...
D'Ambrosio. R... Lampertico. P... Colli. A... Colombo. M... Tsochatzis. E... A. Fraquelli. M... Prati.
D... (2021). Definition of Healthy Ranges for Alanine Aminotransferase Levels: A 2021
Update. Hematology 5: 1824-1832. http://dx.doi.org/10.1002/hep4.1794.
Valvi. D: Hailund. K: Coull. BA: Nielsen. F: Weihe. P: Grandiean. P. (2021). Life-course Exposure to
Perfluoroalkyl Substances in Relation to Markers of Glucose Homeostasis in Early
Adulthood. J Clin Endocrinol Metab 106: 2495-2504.
http: / /dx.doi. or g/10.1210 /clinem /dgab2 67.
Valvi. D: Oulhote. Y: Weihe. P: Dalgard. C: Bierve. KS: Steuerwald. U: Grandiean. P. (2017).
Gestational diabetes and offspring birth size at elevated environmental pollutant exposures.
Environ Int 107: 205-215. http: / /dx.doi. or g/10.1016/i.envint2017.07.016.
van Kerkhof. LWM: Van Dvcke. KCG: Tansen. EHT. M: Beekhof. PK: van Oostrom. CTM: Ruskovska. T:
Velickova. N: Kamcev. N: Pennings. TLA: van Steeg. H: Rodenburg. W. (2015). Diurnal
variation of hormonal and lipid biomarkers in a molecular epidemiology-like setting. PLoS
ONE 10: e0135652. http://dx.doi.org/10.1371/iournal.pone.0135652.
Vanden Heuvel. TP: Kuslikis. BI: Van Rafelghem. Ml: Peterson. RE. (1991a). Disposition of
perfluorodecanoic acid in male and female rats. Toxicol Appl Pharmacol 107: 450-459.
http://dx.doi.Org/10.l 016/0041-008X(91)903 08-2.
Vanden Heuvel. TP: Kuslikis. BI: Van Rafelghem. MI: Peterson. RE. (1991b). Tissue distribution,
metabolism, and elimination of perfluorooctanoic acid in male and female rats. J Biochem
Toxicol 6: 83-92. http://dx.d0i.0rg/l 0.1002/ibt2570060202.
R-47
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Vansell. NR. (2022). Mechanisms by Which Inducers of Drug Metabolizing Enzymes Alter Thyroid
Hormones in Rats. Drug Metab Dispos 50:508-517.
http ://dx. doi. or g/10.112 4 /dmd. 121.000498.
Varshavskv. TR: Robinson. IF: Zhou. Y: Puckett. KA: Kwan. E: Buarpung. S: Aburajab. R: Gaw. SL: Sen.
S: Gao. SM: Smith. SC: Park. IS: Zakharevich. I: Gerona. RR: Fisher. ST: Woodruff. TT. (2021).
Organophosphate flame retardants, highly fluorinated chemicals, and biomarkers of
placental development and disease during mid-gestation. Toxicol Sci 181: 215-228.
http://dx.doi.org/10.1093/toxsci/kfab028.
Varsi. K: Huber. S: Averina. M: Brox. 1: Biarke-Monsen. AL. (2022). Quantitation of linear and
branched perfluoroalkane sulfonic acids (PFSAs) in women and infants during pregnancy
and lactation. Environ Int 160: 107065. http://dx.doi.Org/10.1016/j.envint2021.107065.
Velarde. MC: Chan. AFO: Saio. MET. V: Zakharevich. I: Melamed. 1: Uv. GLB: Teves. TMY: Corachea.
ATM: Valparaiso. AP: Macalindong. SS: Cabaluna. ND: Dofitas. RB: Giudice. LC: Gerona. RR.
(2022). Elevated levels of perfluoroalkyl substances in breast cancer patients within the
Greater Manila Area. Chemosphere 286 Ptl: 131545.
http://dx.doi.Org/10.1016/i.chemosphere.2021.131545.
Velez. MP: Arbuckle. TE: Fraser. WD. (2015). Maternal exposure to perfluorinated chemicals and
reduced fecundity: the MIREC study. Hum Reprod 30: 701-709.
http://dx.doi.org/10.1093/humrep/deu350.
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.
http: / /dx. doi. or g/10.10 21 /acs. est 5b0 0187.
Verner. MA: Ngueta. G: Tensen. ET: Fromme. H: Voelkel. W: Nvgaard. 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.
http://dx.doi.org/10.1021/acs.est5b04399.
Veseli. BE: Perrotta. P: De Meyer. GRA: Roth. L: Van der Donckt. C: Martinet. W: De Meyer. GRY.
(2017). Animal models of atherosclerosis. Eur J Pharmacol 816: 3-13.
http://dx.doi.Org/10.1016/j.ejphar.2017.05.010.
Veskovic. M: Peiovic. M: Sutulovic. N: Hrncic. D: Rasic-Markovic. A: Stanoilovic. 0: Mladenovic. D.
(2024). Exploring fibrosis pathophysiology in lean and obese metabolic-associated fatty
liver disease: An in-depth comparison. INTERNATIONAL JOURNAL OF MOLECULAR
SCIENCES 25. http://dx.doi.org/10.3390/iims25137405.
Vestergaard. S: Nielsen. F: Andersson. AM: Hi0llund. NH: Grandiean. P: Andersen. HR: Tensen. TK.
(2012). Association between perfluorinated compounds and time to pregnancy in a
prospective cohort of Danish couples attempting to conceive. Hum Reprod 27: 873-880.
http://dx.doi.org/10.1093/humrep/der450.
Viberg. H: Lee. I: Eriksson. P. (2013). Adult dose-dependent behavioral and cognitive disturbances
after a single neonatal PFHxS dose. Toxicology 304: 185-191.
http://dx.doi.Org/10.1016/j.tox.2012.12.013.
Vongphachan. V: Cassone. CG: Wu. DM: Chiu. SZ: Crump. D: Kennedy. SW. (2011). Effects of
perfluoroalkyl compounds on mRNA expression levels of thyroid hormone-responsive
genes in primary cultures of avian neuronal cells. Toxicol Sci 120: 392-402.
http ://dx.doi. or g/10.10 9 3/toxsci/kfq3 9 5.
Vuong. A: Yolton. K: Webster. GM: Siodin. A: Calafat. AM: Braun. TM: 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.
http ://dx. doi. or g/10.1016/j. envres. 2016.01.008.
R-48
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Vuong. AM: Braun. TM: Yolton. K: Wang. Z: Xie. C: Webster. GM: Ye. X: Calafat. AM: Dietrich. KN:
Lanphear. BP: Chen. A. (2018a). Prenatal and childhood exposure to perfluoroalkyl
substances (PFAS) and measures of attention, impulse control, and visual spatial abilities.
Environ Int 119: 413-420. http://dx.doi.Org/10.1016/j.envint2018.07.013.
Vuong. AM: Webster. GM: Yolton. K: Calafat. AM: Muckle. G: Lanphear. BP: Chen. A. (2021a).
Prenatal exposure to per- and polyfluoroalkyl substances (PFAS) and neurobehavior in US
children through 8 years of age: The HOME study. Environ Res 195: 110825.
http://dx.doi.Org/10.1016/i.envres.2021.110825.
Vuong. AM: Xie. C: Tandarov. R: Dietrich. KN: Zhang. H: Siodin. A: Calafat. AM: Lanphear. BP:
Mccandless. L: Braun. TM: 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. http://dx.doi.Org/10.1016/i.iiheh.2020.113527.
Vuong. AM: Yolton. K: Wang. Z: Xie. C: Webster. GM: Ye. X: Calafat. AM: Braun. TM: Dietrich. KN:
Lanphear. BP: Chen. A. (2018b). Childhood perfluoroalkyl substance exposure and executive
function in children at 8 years. Environ Int 119: 212-219.
http://dx.doi.Org/10.1016/i.envint2018.06.028.
Vuong. AM: Yolton. K: Xie. C: Dietrich. KN: Braun. TM: 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.
http://dx.doi.Org/10.1016/i.envres.2019.02.025.
Vuong. AM: Yolton. K: Xie. C: Dietrich. KN: Braun. TM: Webster. GM: Calafat. AM: Lanphear. BP: Chen.
A. (2021b). Childhood exposure to per- and polyfluoroalkyl substances (PFAS) and
neurobehavioral domains in children at age 8 years. Neurotoxicol Teratol 88: 107022.
http://dx.doi.Org/10.1016/i.ntt2021.107022.
Wahlang. B: Beier. II: Clair. HB: Bellis-Tones. HI: Falkner. K: Mcclain. CI: Cave. MC. (2013). Toxicant-
associated steatohepatitis [Review], Toxicol Pathol 41: 343-360.
http://dx.doi.org/10.1177/0192623312468517.
Wahlang. B: Tin. 1: Beier. II: Hardestv. IE: Daly. EF: Schnegelberger. RD: Falkner. KC: Prough. RA:
Kirpich. IA: Cave. MC. (2019). Mechanisms of environmental contributions to fatty liver
disease [Review], Curr Environ Health Rep 6: 80-94. http://dx.doi.org/10.1007/s40572-
019-00232-w.
Wang. B: Fu. 1: Gao. K: Liu. 0: Zhuang. L: Zhang. G: Long. M: Na. I: Ren. M: Wang. A: Liang. R: Shen. G:
Li. Z: Lu. 0. (2021a). Early pregnancy loss: Do Per- and polyfluoroalkyl substances matter?
Environ Int 157: 106837. http://dx.doi.Org/10.1016/j.envint2021.106837.
Wang. B: Zhang. R: Tin. F: Lou. H: Mao. Y: Zhu. W: Zhou. W: Zhang. P: Zhang. 1. (2017). Perfluoroalkyl
substances and endometriosis-related infertility in Chinese women. Environ Int 102: 207-
212. http://dx.doi.Org/10.1016/i.envint2017.03.003.
Wang. F: Zhao. C: Gao. Y: Fu. 1: Gao. K: Lv. K: Wang. K: Yue. H: Lan. X: Liang. Y: Wang. Y: Tiang. G.
(2019a). Protein-specific distribution patterns of perfluoroalkyl acids in egg yolk and
albumen samples around a fluorochemical facility. Sci Total Environ 650: 2697-2704.
http ://dx. doi. or g/10.1016/j. scitotenv. 2018.10.006.
Wang. 1: Zeng. XW: Bloom. MS: Oian. Z: Hinvard. LI: Belue. R: Lin. S: Wang. SO: Tian. YP: Yang. M:
Chu. C: Gurram. N: Hu. LW: Liu. KK: Yang. BY: Feng. D: Liu. RO: Dong. GH. (2019b). Renal
function and isomers of perfluorooctanoate (PFOA) and perfluorooctanesulfonate (PFOS):
Isomers of C8 Health Project in China. Chemosphere 218: 1042-1049.
http: / /dx. doi. or g/10.1016/i. chemosphere. 2 018.11.191.
Wang. L: Zhang. OM: Wu. ZY: Huang. XW. (2024). A significant presence in atherosclerotic
cardiovascular disease: Remnant cholesterol: A review. Medicine (Baltimore) 103: e38754.
http://dx.doi.Org/10.1097/MD.0000000000038754.
R-49
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Wang. Y: Aimuzi. R: Nian. M: Zhang. Y: Luo. K: Zhang. T. (2021b). Perfluoroalkyl substances and sex
hormones in postmenopausal women: NHANES 2013-2016. Environ Int 149: 106408.
http: / /dx. doi. or g/10.1016/i. envint 2021.106408.
Wang. Y: Rogan. WT: 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. http: //dx.doi.Org/10.1016/i.iiheh.2015.07.002.
Wang. Y: Rogan. WT: Chen. PC: Lien. GW: Chen. HY: Tseng. YC: Longnecker. MP: Wang. SL. (2014a).
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. http://dx.doi.Org/l0.1289/ehp. 1306925.
Wang. Y: Starling. AP: Haug. LS: Eggesbo. M: Becher. G: Thomsen. C: Travlos. G: King. D: Hoppin. TA:
Rogan. WT: Longnecker. MP. (2013). Association between perfluoroalkyl substances and
thyroid stimulating hormone among pregnant women: a cross-sectional study. Environ
Health 12: 76. http://dx.doi.org/10.1186/1476-069X-12-76.
Wang. Y: Wang. L: Chang. W: Zhang. Y: Zhang. Y: Liu. W. (2019c). Neurotoxic effects of
perfluoroalkyl acids: Neurobehavioral deficit and its molecular mechanism. Toxicol Lett
305: 65-72. http://dx.d0i.0rg/l 0.1016/i.toxlet2019.01.012.
Wang. Y: Zhang. L: Teng. Y: Zhang. T: Yang. L: Li. T: Lai. T: 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. http: / /dx. doi. or g/10.1016 /i .1 es. 2 018.0 3.016.
Wang. YM: Chai. SC: Brewer. CT: Chen. T. (2014b). Pregnane X receptor and drug-induced liver
injury [Review], Expert Opin Drug Metab Toxicol 10: 1521-1532.
http://dx.doi.org/10.1517/17425255.2014.963555.
Wang. Z: Shi. R: Ding. G: Yao. 0: 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.
http: / /dx. doi. or g/10.1016/i. chemosphere. 2 021.133235.
Watkins. DT: Tosson. 1: 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. http://dx.doi.org/10.1289/ehp.1205838.
Weaver. YM: Ehresman. DT: Butenhoff. TL: Hagenbuch. B. (2010). Roles of rat renal organic anion
transporters in transporting perfluorinated carboxylates with different chain lengths.
Toxicol Sci 113: 305-314. http://dx.doi.org/10.1093/toxsci/kfp275.
Webster. GM: Venners. SA: Mattman. A: Martin. TW. (2014). Associations between perfluoroalkyl
acids (PFASs) and maternal thyroid hormones in early pregnancy: a population-based
cohort study. Environ Res 133: 338-347. http://dx.doi.org/10.1016/i.envres.2014.06.012.
Weiss. TM: Andersson. PL: Lamoree. MH: Leonards. PEG: van Leeuwen. SPT: Hamers. T. (2009).
Competitive Binding of Poly- and Perfluorinated Compounds to the Thyroid Hormone
Transport Protein Transthyretin. Toxicol Sci 109: 206-216.
http: / /dx.doi. or g/10.10 9 3 /toxsci /kfp 055.
Weisskopf. MG: Seals. RM: Webster. TF. (2018). Bias amplification in epidemiologic analysis of
exposure to mixtures. Environ Health Perspect 126. http: //dx.doi.org/10.1289/EHP2450.
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.
http://dx.doi.org/10.1210/ic.2013-1282.
Wexler. TA: Sharretts. 1. (2007). Thyroid and bone [Review], Endocrinol Metab Clin North Am 36:
673-705, vi. http://dx.d0i.0rg/l 0.1016/i.ecl.2007.04.005.
R-50
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Whalan. IE. (2015). A toxicologist's guide to clinical pathology in animals: Hematology, clinical
chemistry, urinalysis. Switzerland: Springer International Publishing.
http://dx.doi.Org/l 0.1007/978-3-319-15853-2.
WHO (World Health Organization). (2010). WHO laboratory manual for the examination and
processing of human semen. In WHO laboratory manual for the examination and processing
of human semen (5th ed.). Geneva, Switzerland.
Wielsae. M: Kern. P: Bonefeld-Targensen. EC. (2017). Serum levels of environmental pollutants is a
risk factor for breast cancer in Inuit: a case control study. Environ Health 16:56.
http: // dx.doi.org/10.1186/sl 2940-017-0269-6.
Wielsae. M: Long. M: Ghisari. M: Bonefeld-largensen. EC. (2015). Perfluoroalkylated substances
(PFAS) affect oxidative stress biomarkers in vitro. Chemosphere 129: 239-245.
http://dx.doi.Org/10.1016/i.chemosphere.2014.10.014.
Wikstrom. S: Hussein. G: Lingroth Karlsson. A: Lindh. CH: Bornehag. CG. (2021). Exposure to
perfluoroalkyl substances in early pregnancy and risk of sporadic first trimester
miscarriage. Sci Rep 11: 3568. http://dx.doi.org/10.1038/s41598-021-82748-6.
Wikstrom. S: Lin. PI: Lindh. CH: Shu. H: Bornehag. CG. (2020). Maternal serum levels of
perfluoroalkyl substances in early pregnancy and offspring birth weight Pediatr Res 87:
1093-1099. http://dx.doi.Org/10.1038/s41390-019-0720-l.
Wise. LA: Wesselink. AK: Schildroth. S: Calafat. AM: Bethea. TN: Geller. RT: Coleman. CM: Fruh. V:
Claus Henn. B: Botelho. TC: Harmon. 0E: Thirkill. M: Wegienka. GR: Baird. DP. (2022).
Correlates of plasma concentrations of per- and poly-fluoroalkyl substances among
reproductive-aged Black women. Environ Res 203: 111860.
http ://dx. doi. or g/10.1016/i. envres. 2021.111860.
Woods. MM: Lanphear. BP: Braun. TM: McCandless. LC. (2017). Gestational exposure to endocrine
disrupting chemicals in relation to infant birth weight: A Bayesian analysis of the HOME
Study. Environ Health 16: 115. http://dx.doi.org/10.1186/sl2940-017-0332-3.
Workman. CE: Becker. AB: Azad. MB: Moraes. TT: Mandhane. PI: Turvev. SE: Subbarao. P: Brook. TR:
Sears. MR: Wong. CS. (2019). Associations between concentrations of perfluoroalkyl
substances in human plasma and maternal, infant, and home characteristics in Winnipeg
Canada. Environ Pollut 249: 758-766. http://dx.doi.Org/10.1016/j.envpol.2019.03.054.
Worlev. RR: Moore. SM: Tiernev. BC: Ye. X: Calafat. AM: Campbell. S: Woudneh. MB: Fisher. 1. (2017).
Per- and polyfluoroalkyl substances in human serum and urine samples from a residentially
exposed community. Environ Int 106: 135-143.
http://dx.doi.Org/10.1016/j.envint2017.06.007.
Wright. TM: Larsen. A: Rappazzo. K: Ru. H: Radke. EG: Bateson. TF. (2023). Systematic review and
meta-analysis of birth weight and PFNA exposures [Review], Environ Res 222: 115357.
http ://dx. doi. or g/10.1016/i .envres. 2023.115357.
WS (Weston Solutions Inc). (2007). Remedial investigation report. Phase 2. Fluorochemical (FC)
data assessment report for the Cottage Grove, MN site. St. Paul, MN: 3M Corporate
Toxicology, https://www.pca.state.mn.us/sites/default/files/pfc-cottagegrove-
remedialinvestigationreportpdf.
Xiao. C: Grandiean. P: Valvi. D: Nielsen. F: Tensen. 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. http://dx.doi.org/10.1210 /clinem /dgz 147.
Xiong. X: Chen. B: Wang. Z: Ma. L: Li. S: Gao. Y. (2022). Association between perfluoroalkyl
substances concentration and bone mineral density in the US adolescents aged 12-19 years
inNHANES 2005-2010. Front Endocrinol (Lausanne) 13: 980608.
http://dx.doi.org/10.3389/fendo.2022.980608.
Xu. C: Yin. S: Liu. Y: Chen. F: Zhong. Z: Li. F: Liu. K: Liu. W. (2019). Prenatal exposure to chlorinated
polyfluoroalkyl ether sulfonic acids and perfluoroalkyl acids: Potential role of maternal
R-51
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
determinants and associations with birth outcomes. J Hazard Mater 380: 120867.
http://dx.doi.Org/10.1016/i.ihazmat2019.120867.
Y. A: SC. F: E. R. (2024). Recent insights into the role of hormones during development and their
functional regulation. Front Endocrinol (Lausanne) 15.
http://dx.doi.org/10.3389/fendo.2024.1340432.
Y. L: X. C: L. Z: X. W: Y. X: Z. 0: G. L: 0. W: K. T: KS. L: CI. C: 1. Z: 1.. T. (2022). The functional role of
lipoproteins in atherosclerosis: novel directions for diagnosis and targeting therapy. 13:
491-520. http://dx.doi.org/10.14336/AD.2021.0929.
Yamauchi. T. (2005). Neuronal Ca2+/calmodulin-dependent protein kinase II - Discovery, progress
in a quarter of a century and perspective: Implication for learning and memory. Biol Pharm
Bull 28: 1342-1354. http: //dx.doi.org/10.1248/bpb.28.1342.
Yang. A: Chen. L: Xie. Z: Feng. H: Sun. F. (2016a). Constructal heat transfer rate maximization for
cylindrical pin-fin heat sinks. Applied Thermal Engineering 108: 427-435.
http: / /dx. doi. or g/10.1016/i. applthermaleng. 2016.07.150.
Yang. BY: Wu. 1: Niu. XL: He. CI: Bloom. MS: Abudoukade. M: Abulizi. M: Xu. AM: Li. BB: Li. L: Zhong.
XM: Wu. OZ: Chu. C: Luo. YN: Liu. XX: Zeng. XW: Yu. YT: Dong. GH: Zou. XG: Liu. T. (2022a).
Low-level environmental per- and polyfluoroalkyl substances and preterm birth: A nested
case-control study among a Uyghur population in northwestern China. Exposure and Health
14: 793-805. http://dx.doi.org/10.1007/sl 2403-021-00454-0.
Yang. CH: Glover. KP: Han. X. (2009). Organic anion transporting polypeptide (Oatp) lal-mediated
perfluorooctanoate transport and evidence for a renal reabsorption mechanism of Oatplal
in renal elimination of perfluorocarboxylates in rats. Toxicol Lett 190: 163-171.
http://dx.doi.Org/10.1016/i.toxlet.2009.07.011.
Yang. D: Han. 1: Hall. PR: Sun. I: Fu. 1: Kutarna. S: Houck. KA: Lalone. CA: Doering. TA: Ng. CA: Peng. H.
(2020a). Nontarget screening of per- and polyfluoroalkyl substances binding to human liver
fatty acid binding protein. Environ Sci Technol 54: 5676-5686.
http://dx.doi.Org/10.1021/acs.est0c00049.
Yang. I: Wang. H: Du. H: Fang. H: Han. M: Wang. Y: Xu. L: Liu. S: Yi. I: Chen. Y: Tiang. 0: He. G. (2022b).
Exposure to perfluoroalkyl substances was associated with estrogen homeostasis in
pregnant women. Sci Total Environ 805: 150360.
http: / /dx. doi. or g/10.1016/i. scitotenv. 2021.150360.
Yang. I: Wang. H: Du. H: Fang. H: Han. M: Xu. L: Liu. S: Yi. 1: Chen. Y: Tiang. 0: He. G. (2020b). Serum
perfluoroalkyl substances in relation to lipid metabolism in Chinese pregnant women.
Chemosphere 273: 128566. http://dx.doi.Org/10.1016/j.chemosphere.2020.128566.
Yang. L: Li. 1: Lai. I: Luan. H: Cai. Z: Wang. Y: Zhao. Y: Wu. Y. (2016b). Placental transfer of
perfluoroalkyl substances and associations with thyroid hormones: Beijing prenatal
exposure study. Sci Rep 6: 21699. http://dx.doi.org/10.1038/srep21699.
Yang. L: Wang. Z: Shi. Y: Li. I: Wang. Y: Zhao. Y: Wu. Y: Cai. Z. (2016c). Human placental transfer of
perfluoroalkyl acid precursors: Levels and profiles in paired maternal and cord serum.
Chemosphere 144: 1631-1638. http://dx.doi.Org/10.1016/i.chemosphere.2015.10.063.
Yang. 0: 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: Across-sectional study. Sci Total Environ 621: 1542-1549.
http: / /dx. doi. or g/10.1016/i .scitotenv. 2017.10.074.
Yang. X: Schnakenberg. LK: Shi. 0: Salminen. WF. (2014). Hepatic toxicity biomarkers. In RC Gupta
(Ed.), Biomarkers inToxicology (pp. 241-259). New York, NY: Academic Press.
http://dx.doi.Org/10.1016/B978-0-12-404630-6.00013-0.
Yao. I: Dong. Z: Tiang. L: Pan. Y: Zhao. M: Bai. X: Dai. I. (2023a). Emerging and Legacy Perfluoroalkyl
Substances in Breastfed Chinese Infants: Renal Clearance, Body Burden, and Implications.
Environ Health Perspect 131: 37003-37003. http: //dx.doi.org/10.1289/EHP11403.
R-52
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
Yao. 1: Pan. Y: Sheng. N: Su. Z: Guo. Y: Wang. I: Dai. 1. (2020). Novel perfluoroalkyl ether carboxylic
acids (PFECAs) and sulfonic acids (PFESAs): Occurrence and association with serum
biochemical parameters in residents living near a fluorochemical plant in China. Environ Sci
Technol54: 13389-13398. http://dx.doi.org/10.1021/acs.est0c02888.
Yao. 0: Gao. Y: Zhang. Y: Oin. 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.
http: / /dx. doi. or g/10.1016/i. chemosphere. 2 021.131521.
Yao. 0: Shi. R: Wang. C: Han. W: Gao. Y: Zhang. Y: Zhou. Y: Ding. G: Tian. Y. f20191. Cord blood per-
and polyfluoroalkyl substances, placental steroidogenic enzyme, and cord blood
reproductive hormone. Environ Int 129: 573-582.
http://dx.doi.Org/10.1016/i.envint2019.03.047.
Yao. 0: Vinturache. A: Lei. X: Wang. Z: Pan. C: Shi. R: Yuan. T: Gao. Y: Tian. Y. (2022). Prenatal
exposure to per- and polyfluoroalkyl substances, fetal thyroid hormones, and infant
neurodevelopment Environ Res 206: 112561.
http://dx.doi.Org/10.1016/i.envres.2021.112561.
Yao. W: Xu. 1: Tang. W: Gao. C: Tao. L: Yu. 1: Lv. 1: Wang. H: Fan. Y: Xu. DX: Huang. Y. (2023b).
Developmental toxicity of perfluorohexane sulfonate at human relevant dose during
pregnancy via disruption in placental lipid homeostasis. Environ Int 177: 108014.
http: / /dx. doi. or g/10.1016/i. envint 2023.108014.
Yeung. LW: Guruge. KS: Tanivasu. S: Yamashita. N: Angus. PW: Herath. CB. (2013). Profiles of
perfluoroalkyl substances in the liver and serum of patients with liver cancer and cirrhosis
in Australia. Ecotoxicol Environ Saf 96: 139-146.
http: / /dx. doi. or g/10.1016 /i. ecoenv. 2 013.0 6.0 0 6.
Yin. X: Pi. T: Cao. X: Liu. Z: Xie. 1: Zhang. S. (2021). Chronic exposure to perfluorohexane sulfonate
leads to a reproduction deficit by suppressing hypothalamic kisspeptin expression in mice.
Journal of Ovarian Research 14: 141. http: //dx.doi.org/10.1186 /sl3048-021-00903-z.
Younossi. ZM: Henry. L. (2024). Understanding the Burden of Nonalcoholic Fatty Liver Disease:
Time for Action. Diabetes Spectrum 37: 9-19. http://dx.doi.org/10.2337/dsi23-001Q.
Yu. G: Tin. M: Huang. Y: Aimuzi. R: Zheng. T: Nian. M: Tian. Y: Wang. W: Luo. Z: Shen. L: Wang. X: Du.
0: Xu. W: Zhang. 1. (2021). Environmental exposure to perfluoroalkyl substances in early
pregnancy, maternal glucose homeostasis and the risk of gestational diabetes: A prospective
cohort study. Environ Int 156: 106621. http: //dx.doi.Org/10.1016/i.envint2021.106621.
Zare Teddi. M: Dalla Zuanna. T: Barbieri. G: Fabricio. ASC: Dapra. F: Fletcher. T: Russo. F: Pitter. G:
Canova. C. (2021). Associations of Perfluoroalkyl Substances with Prevalence of Metabolic
Syndrome in Highly Exposed Young Adult Community Residents-A Cross-Sectional Study in
Veneto Region, Italy. Int J Environ Res Public Health 18: 1194.
http: / /dx. doi. or g/10.3 3 9 0 /ii erphl 8 0 3119 4.
Zeng. X: Chen. 0: Zhang. X: Li. H: Liu. 0: Li. C: Ma. M: Zhang. 1: Zhang. W: Zhang. I: Huang. L. (2019a).
Association between prenatal exposure to perfluoroalkyl substances and asthma-related
diseases in preschool children. Environ Sci Pollut Res Int 26: 29639-29648.
http://dx.doi.Org/10.1007/sll356-019-05864-x.
Zeng. XW: Bloom. MS: Dharmage. SC: Lodge. CI: Chen. D: Li. S: Guo. Y: Roponen. M: Talava. P:
Hirvonen. MR: Ma. H: Hao. YT: Chen. W: Yang. M: Chu. C: Li. 00: Hu. LW: Liu. KK: Yang. BY:
Liu. S: Fu. C: Dong. GH. (2019b). 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.
http: / /dx. doi. or g/10.1016/i. scitotenv. 2019.01.325.
Zeng. XW: Lodge. CI: Dharmage. SC: Bloom. MS: Yu. Y: Yang. M: Chu. C: Li. 00: Hu. LW: Liu. KK: Yang.
BY: Dong. GH. (2019c). Isomers of per- and polyfluoroalkyl substances and uric acid in
R-53
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
adults: Isomers of C8 Health Project in China. Environ Int 133: 105160.
http://dx.doi.Org/10.1016/i.envint2019.105160.
Zeng. XW: Oian. Z: Emo. B: Vaughn. M: Rao. 1: Oin. XD: Zhu. Y: Li. 1: Lee. YL: Dong. GH. (2015).
Association of polyfluoroalkyl chemical exposure with serum lipids in children. Sci Total
Environ 512-513: 364-370. http://dx.doi.Org/10.1016/i.scitotenv.2015.01.042.
Zhang. BH: Yin. F: Qiao. YN: Guo. SD. (2022a). Triglyceride and triglyceride-rich lipoproteins in
atherosclerosis [Review], Front Mol Biosci 9: 909151.
http: / /dx. doi. or g/10.3 3 89 /fmolb. 2 022.909151.
Zhang. H: Yolton. K: Webster. GM: Ye. X: Calafat. AM: Dietrich. KN: Xu. Y: Xie. C: Braun. TM: Lanphear.
BP: Chen. A. (2018a). Prenatal and childhood perfluoroalkyl substances exposures and
children's reading skills at ages 5 and 8 years. Environ Int 111: 224-231.
http: / /dx. doi. or g/10.1016/i. envint 2017.11.031.
Zhang. IF: Hu. LH: Xu. HY. (2024). Dietary exposure to per- and polyfluoroalkyl substances:
Potential health impacts on human liver [Review], Sci Total Environ 907: 167945.
http: / /dx. doi. or g/10.1016/j. scitotenv. 2023.167945.
Zhang. 0: Liu. W: Niu. 0: Wang. Y: Zhao. H: Zhang. H: Song. I: Tsuda. S: Saito. N. (2016a). Effects of
perfluorooctane sulfonate and its alternatives on long-term potentiation in the
hippocampus CA1 region of adult rats in vivo. Toxicology Research 5: 539-546.
http: / /dx. doi. or g/10.10 3 9 /c5 txO 0184f.
Zhang. 0: Zhao. H: Liu. W: Zhang. Z: Oin. H: Luo. F: Leng. S. (2016b). Developmental perfluorooctane
sulfonate exposure results in tau hyperphosphorylation and (3-amyloid aggregation in
adults rats: Incidence for link to Alzheimer's disease. Toxicology 347-349: 40-46.
http://dx.doi.Org/10.1016/i.tox.2016.03.003.
Zhang. S: Tan. R: Pan. R: Xiong. 1: Tian. Y: Wu. I: Chen. L. (2018b). Association of perfluoroalkyl and
polyfluoroalkyl substances with premature ovarian insufficiency in Chinese women. J Clin
Endocrinol Metab 103: 2543-2551. http://dx.d0i.0rg/l0.1210/ic.2017-02783.
Zhang. T: Sun. H: Lin. Y: Oin. X: Zhang. Y: Geng. X: Kannan. K. (2013a). 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.
http://dx.doi.org/10.1021/es400937y.
Zhang. T: Zhang. B: Bai. X: Yao. Y: Wang. L: Shu. Y: Kannan. K: Huang. X: Sun. H. (2019a). Health
status of elderly people living near e-waste recycling sites: association of e-waste
dismantling activities with legacy perfluoroalkyl substances (PFASs). Environ Sci Technol
Lett 6: 133-140. http://dx.doi.org/10.1021/acs.estlett.9b00085.
Zhang. W: Zhang. D: Zagorevski. DV: Liang. Y. (2019b). Exposure of Juncus effusus to seven
perfluoroalkyl acids: Uptake, accumulation and phytotoxicity. Chemosphere 233: 300-308.
http://dx.doi.Org/10.1016/i.chemosphere.2019.05.258.
Zhang. Y: Beesoon. S: Zhu. L: Martin. TW. (2013b). Biomonitoring of perfluoroalkyl acids in human
urine and estimates of biological half-life. Environ Sci Technol 47: 10619-10627.
http: / /dx. doi. or g/10.10 21 /es4019 0 5 e.
Zhang. Y: Pan. C: Ren. Y: Wang. Z: Luo. I: Ding. G: Vinturache. A: Wang. X: Shi. R: Ouvang. F: Zhang. 1:
Li. I: Gao. Y: Tian. Y. (2022b). Association of maternal exposure to perfluoroalkyl and
polyfluroalkyl substances with infant growth from birth to 12 months: A prospective cohort
study. Sci Total Environ 806: 151303. http: / /dx. doi. or g/10.1016/i .scitotenv. 2021.151303.
Zhang. YH: Lv. 1: Fan. YI: Tao. L: Xu. IT: Tang. WT: Sun. N: Zhao. LL: Xu. DX: Huang. YC. (2023).
Evaluating the effect of gestational exposure to perfluorohexane sulfonate on placental
development in mice combining alternative splicing and gene expression analyses. Environ
Health Perspect 131: 117011. http://dx.d0i.0rg/l 0.1289/EHP13217.
Zhao. W: Zitzow. TP: Ehresman. DT: Chang. SC: Butenhoff. TL: Forster. I: Hagenbuch. B. (2015).
Na+/taurocholate cotransporting polypeptide and apical sodium-dependent bile acid
R-54
-------
IRIS Toxicological Review of Perfluorohexanesulfonic Acid and Related Salts
transporter are involved in the disposition of perfluoroalkyl sulfonates in humans and rats.
Toxicol Sci 146: 363-373. http://dx.doi.org/10.1093/toxsci/kfvl02.
Zhao. W: Zitzow. TP: Weaver. Y: Ehresman. DT: Chang. SC: Butenhoff. TL: Hagenbuch. B. (2017).
Organic anion transporting polypeptides contribute to the disposition of perfluoroalkyl
acids in humans and rats. Toxicol Sci 156: 84-95. http: / / dx. do i. o r g /10.109 3 /toxs ci /kfw2 3 6.
Zhao. X: Lin. TY: Dong. WW: Tang. ML: Yan. SG. (2022). Per- and polyfluoroalkyl substances
exposure and bone mineral density in the U.S. population from NHANES 2005-2014. J Expo
Sci Environ Epidemiol, http: //dx.doi.org/10.1038/s41370-022-00452-7.
Zhou. W: Zhang. L: Tong. C: Fang. F: Zhao. S: Tian. Y: Tao. Y: Zhang. 1. (2017a). Plasma perfluoroalkyl
and polyfluoroalkyl substances concentration and menstrual cycle characteristics in
preconception women. Environ Health Perspect 125: 067012.
http://dx.doi.org/10.1289/EHP1203.
Zhou. Y: Hu. LW: Oian. ZM: Chang. 11: 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.
http: / /dx. doi. or g/10.1016/i. envint 2016.05.018.
Zhou. Y: Hu. LW: Oian. ZM: Geiger. SD: Parrish. KL: Dharmage. SC: Campbell. B: Roponen. M: Talava.
P: Hirvonen. MR: Heinrich. 1: Zeng. XW: Yang. BY: Oin. XD: Lee. YL: Dong. GH. (2017b).
Interaction effects of polyfluoroalkyl substances and sex steroid hormones on asthma
among children. Sci Rep 7: 899. http: //dx.doi.org/10.1038/s41598-017-01140-5.
Zhou. Z: Shi. YL: Vestergren. R: Wang. T: Liang. Y: Cai. YO. (2014). Highly elevated serum
concentrations of perfluoroalkyl substances in fishery employees from Tangxun lake, china.
Environ Sci Technol 48: 3864-3874. http://dx.doi.org/10.1021/es4057467.
Zhu. Y: Oin. XD: Zeng. XW: Paul. G: Morawska. L: Su. MW: Tsai. CH: Wang. SO: 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.
http: / /dx. doi. or g/10.1016/i. scitotenv. 2016.03.187.
Zoeller. RT: Crofton. KM. (2005). Mode of action: developmental thyroid hormone insufficiency-
neurological abnormalities resulting from exposure to propylthiouracil [Review], CritRev
Toxicol 35: 771-781. http://dx.d0i.0rg/l 0.1080/10408440591007313.
Zoeller. RT: Rovet. 1. (2004). Timing of thyroid hormone action in the developing brain: Clinical
observations and experimental findings [Review], J Neuroendocrinol 16: 809-818.
http://dx.d0i.0rg/l 0.1111 /i.l 365-2826.2004.01243.x.
R-55
------- |