EPA/635/R-23/148b
External Review Draft
www.epa.gov/iris
IRIS Toxicological Review of Perfluorohexanesulfonic Acid
(PFHxS, CASRN 355-46-4) and Related Salts
Supplemental Information—Appendices A-G
July 2023
Integrated Risk Information System
Center for Public Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
DISCLAIMER
This document is a preliminary draft for review purposes only. This information is
distributed solely for the purpose of predissemination peer review under applicable information
quality guidelines. It has not been formally disseminated by EPA. It does not represent and should
not be construed to represent any Agency determination or policy. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
This document is a draft for review purposes only and does not constitute Agency policy.
ii DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
CONTENTS
APPENDIX A. SYSTEMATIC REVIEW PROTOCOL A-l
APPENDIX B. LITERATURE SEARCH STRATEGY AND POPULATIONS, EXPOSURES, COMPARATORS,
AND OUTCOMES (PECO) CRITERIA B-l
B.l. LITERATURE SEARCH AND SCREENING STRATEGY B-l
B.2. TITLE AND ABSTRACT LEVEL SCREENING CRITERIA FOR THE INITIAL LITERATURE
SEARCHES B-8
B.3. DOCUMENTATION OF LITERATURE SEARCH UPDATES AFTER APRIL 2022 B-13
APPENDIX C. SUPPLEMENTAL APPROACHES AND DATAANALYSES C-l
C.l. PFAS CO-EXPOSURE CONSIDERATIONS AND META-ANALYSIS OF PFHXS EFFECTS ON
BIRTH WEIGHT C-l
C.l.l. Confounding Directionality and PFAS Co-exposure Statistical Approaches C-l
C.l.2. PFAS Co-exposure Correlations with PFHxS C-2
C.l.3. PFHxS and PFAS Co-exposure Study Results C-3
C.1.4. Pregnancy Hemodynamics Background C-6
C.l.5. Meta-Analysis Methods C-7
C.l.6. Meta-Analysis Results C-18
C.l.7. Sensitivity Analysis Results C-21
C.l.8. Summary of Meta-Analysis of PFHxS Effects on Birth Weight C-21
C.2. AOP-BASED APPROACH FOR EVALUATING POTENTIAL PFHXS-INDUCED MECHANISMS
OF HEPATOTOXICITY MODE OF ACTION C-23
C.2.1. Objective and Methodology C-23
C.2.2. Proposed AOP Approach for Evaluation of PFAS-lnduced Liver Toxicity C-23
C.3. SUMMARY OF RELEVANT HIGH-THROUGHPUT SCREENING ASSAYS FROM EPA'S
COMPTOX CHEMICALS DASHBOARD C-25
C.3.1. In vitro Bioreactivity Data Relevant to the Mechanisms of PFHxS-lnduced Liver
Effects C-25
C.3.2. In vitro Bioactivity Data Relevant to the Mechanisms of PFHxS-lnduced Thyroid
Effects C-29
APPENDIX D. BENCHMARK DOSE MODELING RESULTS D-l
D.l. BENCHMARK DOSE MODELING SUMMARY FOR NONCANCER ENDPOINTS D-l
D.l.l. Benchmark Dose Modeling Approaches D-l
This document is a draft for review purposes only and does not constitute Agency policy.
iii DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
D.2. BENCHMARK DOSE MODELING RESULTS FROM ANIMAL STUDIES D-27
D.2.1. Benchmark Dose Modeling Approaches D-27
APPENDIX E. DETAILED PHARMACOKINETIC ANALYSES E-l
E.l. BAYESIAN ANALYSIS OF PFHXS PHARMACOKINETICS IN RATS, MICE, AND MONKEYS E-l
E.l.l. Pharmacokinetic Model E-l
E.2. DESCRIPTION AND EVALUATION OF A SINGLE-COMPARTMENT PK APPROACH E-ll
APPENDIX F. QUALITY ASSURANCE FOR THE IRIS TOXICOLOGICAL REVIEW OF
PERFLUOROHEXANESULFONIC ACID AND RELATED SALTS F-l
APPENDIX G. SUMMARY OF PUBLIC AND EXTERNAL PEER REVIEW COMMENTS AND EPA'S
DISPOSITION G-l
REFERENCES R-l
This document is a draft for review purposes only and does not constitute Agency policy.
iv DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
TABLES
Table B-l. Summary of detailed search strategies for perfluorohexanesulfonic acid and related
salts B-l
Table B-2. Processes used to augment the search of core databases for PFHxS (355-46-4) B-6
Table B-3. Title and abstract level screening criteria for the initial literature searches B-8
Table B-4. Example DistillerSR form questions to be used for title/abstract level and full text level
screening for literature search updates from 2019 B-ll
Table B-5. Summary of decisions regarding studies identified after April 2022 B-14
Table C-l. PFAS correlation coefficients in nine mutually adjusted PFAS studies C-3
Table C-2. Impact of co-exposure exposure adjustment on birth weight results C-5
Table C-3 Details on study sample timings and strata assignments C-12
Table C-4. Meta-analysis of PFHxS on birth weight changes (in g per ln(ng/mL)) stratified by
study confidence C-20
Table C-5. Meta-analysis of PFHxS on birth weight (in g per ln(ng/mL)) stratified by sample
timing C-20
Table C-6. Sensitivity of natural log scale or natural scale re-expression for the overall and
stratified meta-analyses of birth weight C-21
Table C-7. Bioactivity summary for PFHxS from in vitro HTS assays C-28
Table C-8. Endocrine disruptor screening program 21 assay summary results C-29
Table D-l. Results specific to the low-dose slope from the piecewise- linear analyses of PFHxS
measured at age 5 years and log2 D-2
Table D-2. BMDs and BMDLs for effect of PFHxS at age 5 years on anti-tetanus antibody
concentrations at age 7 years D-7
Table D-3. Results specific to the low-dose slope from the piecewise- linear analyses of PFHxS
measured at age 5 years D-9
Table D-4. BMDs and BMDLs for effect of PFHxS at age 5 years on anti-diphtheria antibody
concentrations at age 7 years D-ll
Table D-5. Results of the linear analyses of PFHxS measured perinatally and tetanus antibodies
measured at age 5 years D-12
Table D-6. BMDs and BMDLs for effect of PFHxS measured perinatally and anti-tetanus antibody
concentrations at age 5 years D-13
Table D-7. Results of the analyses of PFHxS measured perinatally and diphtheria antibodies
measured at age 5 years D-15
Table D-8. BMDs and BMDLs for effect of PFHxS measured perinatally and anti-diphtheria
antibody concentrations at age 5 years D-16
Table D-9. BMDs and BMDLs for effect of PFHxS on decreased birth weight D-22
Table D-10. BMDs and BMDLs for effect of PFHxS on decreased birth weight by background
exposure D-23
Table D-ll. BMDs and BMDLs for effect of PFHxS on decreased birth weight by background
exposure D-24
Table D-12. BMDs and BMDLs for effect of PFHxS on decreased birth weight using meta-analysis
results conducted in log scale overall D-25
Table D-13. BMDs and BMDLs for effect of PFHxS on decreased birth weight using meta-analysis
results conducted in log scale overall D-26
Table D-14. Sources of data used in benchmark dose modeling of PFHxS endpoints from animal
studies D-28
This document is a draft for review purposes only and does not constitute Agency policy.
v DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table D-15. Dose-response data for decreased free T4 in male rats D-28
Table D-16. Benchmark dose results for decreased free T4 in male rats D-29
Table D-17. Dose-response data for decreased T4 in male rats D-30
Table D-18. Benchmark dose results for decreased Total T4 in male rats D-32
Table D-19. Dose-response data for total T4 in female rats D-34
Table D-20. Benchmark dose results for decreased total T4 in female rats D-35
Table D-21. Benchmark dose results for decreased T3 in male rats D-38
Table D-22. Dose response data for decreased free T3 in F1 combined PND16/17 rats D-41
Table D-23. Benchmark dose results for decreased T3 in F1 PND16 male rats D-42
Table E-l. Weakly informed prior distributions for pharmacokinetic parameters used in the
Bayesian analysis E-3
Table E-2. Results from prior sensitivity analysis for the three classes E-6
FIGURES
Figure C-l. Twenty-seven informative nonoverlapping perinatal studies of birth weight measures
and continuous PFHxS exposure results included in meta-analysis C-8
Figure C-2. Forest plot of 27 studies included in the meta-analysis on PFHxS exposures and
changes in birth weight C-19
Figure C-3. The proposed MOA in the figure above is based on previous analyses on PFAS-
induced C-25
Figure C-4. Bioactivity data for PFHxS from in vitro HTS ToxCast/Tox21 assays in human liver
tissues C-27
Figure C-5. Summary of positive nuclear receptor assays in human liver tissue C-28
Figure D-l. Difference in population tail probabilities D-5
Figure D-2. Difference in population tail probabilities resulting from a Vz standard deviation D-7
Figure D-3. Dose response data for male rat free T4 D-30
Figure D-4. Dose response data for male rat Total T4 D-34
Figure D-5. Dose response data for female rat Total T4 D-37
Figure D-6. Dose response data for male rat T3 D-40
Figure D-7. Dose response data for decreased T3 in F1 PND17 rats D-43
Figure E-l. Prior predictive check to ensure equal-tailed interval from prior distributions
encompass the available time-course concentration data for fitting E-6
Figure E-2. Predicted (black line with blue 90% credible interval) and observed (black circles)
serum time-courses for male (top panel) and female (bottom panel) rats after a
4 mg/kg gavage PFHxS E-7
Figure E-3. Predicted (black line with blue 90% credible interval) and observed (black circles)
serum time-courses for male (top panel) and female (bottom two panels) rats
after a 10 mg/kg gavage (both male and female) and 4 mg/kg gavage (female
only) PFHxS E-8
Figure E-4. Predicted (black line with blue 90% credible interval) and observed (black circles)
serum time-courses for male (top panel) and female (bottom panel) rats after a
4, 16, or 32 mg/kg gavage PFHxS E-9
Figure E-5. Predicted (black line with blue 90% credible interval) and observed (black circles)
serum time-courses for male (top panel) and female (bottom panel) mice after a
1 or 20 mg/kg gavage PFHxS E-10
This document is a draft for review purposes only and does not constitute Agency policy.
vi DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Figure E-6. Predicted (black line with blue 90% credible interval) and observed (black circles)
serum time-courses for male (top panel) and female (bottom panel) nonhuman
primates following a 10 mg/kg IV PFHxS dose E-ll
Figure E-7. Male and female rat body weight changes during 28-day PFHxS bioassay E-13
Figure E-8. Predicted accumulation and observed end-of-study of PFHxS in male rats as a
function of dose E-14
This document is a draft for review purposes only and does not constitute Agency policy.
vii DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
ABBREVIATIONS
AIC
Akaike's information criterion
HERO
Health and Environmental Research
ALT
alanine aminotransferase
Online
AST
aspartate aminotransferase
i.p.
intraperitoneal
atm
atmosphere
i.v.
intravenous
ATSDR
Agency for Toxic Substances and
IRIS
Integrated Risk Information System
Disease Registry
LC50
median lethal concentration
BMD
benchmark dose
LD50
median lethal dose
BMDL
benchmark dose lower confidence limit
LOAEL
lowest-observed-adverse-effect level
BMDS
Benchmark Dose Software
MN
micronuclei
BMR
benchmark response
MNPCE
micronucleated polychromatic
BUN
blood urea nitrogen
erythrocyte
BW
body weight
MOA
mode of action
CA
chromosomal aberration
MTD
maximum tolerated dose
CASRN
Chemical Abstracts Service registry
NCEA
National Center for Environmental
number
Assessment
CHO
Chinese hamster ovary (cell line cells)
NCI
National Cancer Institute
CI
confidence interval
NOAEL
no-observed-adverse-effect level
CL
confidence limit
NTP
National Toxicology Program
CNS
central nervous system
NZW
New Zealand White (rabbit breed)
CYP450
cytochrome P450
ORD
Office of Research and Development
DAF
dosimetric adjustment factor
PBPK
physiologically based pharmacokinetic
DMSO
dimethylsulfoxide
PND
postnatal day
DNA
deoxyribonucleic acid
POD
point of departure
EPA
Environmental Protection Agency
POD [AD J]
duration-adjusted POD
ER
extra risk
QSAR
quantitative structure-activity
FDA
Food and Drug Administration
relationship
FEVi
food expiratory volume of one second
RD
relative deviation
GDH
glutamate dehydrogenase
RfC
inhalation reference concentration
GGT
conflict of interest
RfD
oral reference dose
CPAD
Chemical and Pollutant Assessment
RGDR
regional gas dose ratio
Division
RNA
ribonucleic acid
CPHEA
Center for Public Health and
SAR
structure-activity relationship
Environmental Assessment
SCE
sister chromatid exchange
CYP450
cytochrome P450
SD
standard deviation
DAF
dosimetric adjustment factor
SDH
sorbitol dehydrogenase
DMSO
dimethylsulfoxide
SE
standard error
DNA
deoxyribonucleic acid
SGOT
serum glutamic oxaloacetic
EPA
Environmental Protection Agency
transaminase, also known as AST
ER
extra risk
SGPT
serum glutamic pyruvic transaminase,
FDA
Food and Drug Administration
also known as ALT
FEVi
forced expiratory volume of one second
TSCATS
Toxic Substances Control Act Test
GD
gestation day
Submissions
GDH
glutamate dehydrogenase
TWA
time-weighted average
GGT
y-glutamyl transferase
UF
uncertainty factor
GLP
Good Laboratory Practice
UFa
animal-to-human uncertainty factor
GSH
glutathione
UFd
database deficiencies uncertainty factor
GST
glutathione-^"-transferase
UFh
human variation uncertainty factor
Hb/g-A
animal blood:gas partition coefficient
UFl
LOAEL-to-NOAEL uncertainty factor
Hb/g-H
human blood:gas partition coefficient
UFs
subchronic-to-chronic uncertainty
HBCD
hexabromocyclododecane
factor
HEC
human equivalent concentration
WOS
Web of Science
HED
human equivalent dose
This document is a draft for review purposes only and does not constitute Agency policy.
viii DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
APPENDIX A. SYSTEMATIC REVIEW PROTOCOL
1 A single systematic review protocol was used to guide the development of five, separate
2 IRIS PFAS assessments (i.e., PFBA, PFHxA, PFHxS, PFNA, and PFDA). This "systematic review
3 protocol for the PFAS IRIS assessments" was released for public comment and subsequently
4 updated. The updated protocol and prior revisions can be found at the following location:
5 http://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=345065.
This document is a draft for review purposes only and does not constitute Agency policy.
A-l DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
APPENDIX B. LITERATURE SEARCH STRATEGY AND
POPULATIONS, EXPOSURES, COMPARATORS,
AND OUTCOMES (PECO) CRITERIA
B.l. LITERATURE SEARCH AND SCREENING STRATEGY
Table B-l. Summary of detailed search strategies for
perfluorohexanesulfonic acid and related salts (PubMed, Web of Science,
Toxline, TSCATS, Toxcenter)
Database
Terms
Hits
Initial strategy
PubMed
7/24/17
108427-53-8[rn] OR 355-46-4[rn] OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
Tridecafluorohexane-l-sulfonic acid"[tw] OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
tridecafluoro-l-Hexanesulfonic acid"[tw] OR "1-Hexanesulfonic acid,
l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-"[tw] OR "1-Hexanesulfonic acid,
tridecafluoro-"[tw] OR "1-Perfluorohexanesulfonic acid"[tw] OR
"Perfluoro-l-hexanesulfonate"[tw] OR "Perfluorohexane sulfonic acid"[tw]
OR "Perfluorohexane-1-sulphonic acid"[tw] OR
"Perfluorohexanesulfonate"[tw] OR "Perfluorohexanesulfonic acid"[tw]
OR "Perfluorohexylsulfonate"[tw] OR "Tridecafluorohexanesulfonic
acid"[tw] OR "tridecafluoro-l-Hexanesulfonic acid"[tw] OR "PFHxS"[tw]
396
This document is a draft for review purposes only and does not constitute Agency policy.
B-l DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Database
Terms
Hits
PubMed
04/29/2020
(("108427-53-8"[rn] OR "355-46-4"[rn] OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
Tridecafluorohexane-l-sulfonic acid"[tw] OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
tridecafluoro-l-Hexanesulfonic acid"[tw] OR "1-Hexanesulfonic acid,
l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-"[tw] OR "1-Hexanesulfonic acid,
tridecafluoro-"[tw] OR "1-Perfluorohexanesulfonic acid"[tw] OR
"Perfluoro-l-hexanesulfonate"[tw] OR "Perfluorohexane sulfonic acid"[tw]
OR "Perfluorohexane-1-sulphonic acid"[tw] OR
"Perfluorohexanesulfonate"[tw] OR "Perfluorohexanesulfonic acid"[tw]
OR "Perfluorohexylsulfonate"[tw] OR "Tridecafluorohexanesulfonic
acid"[tw] OR "tridecafluoro-l-Hexanesulfonic acid"[tw] OR "PFHxS"[tw])
AND ("2019/05/03"[Date - Publication]: "3000"[Date - Publication])
((((((((((((((((((((((((((((("108427-53-8"[rn]) OR "423-50-7"[rn]) OR "1-
Hexanesulfonic acid, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-, ion(l-)"[tw])
OR "PFHxS ion(l-)"[tw]) OR "PFHxS_ion"[tw]) OR
"Perfluorohexanesulfonate"[tw]) OR "Tridecafluorohexane-1-
sulfonate"[tw]) OR "perfluorohexyl sulfonate"[tw]) OR
"1,1,2,2,3,3,4,4,5,5,6,6,6-Tridecafluoro-l-hexanesulfonyl fluoride"[tw]) OR
"1-Hexanesulfonyl fluoride, l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-"[tw])
OR "1,1,2,2,3,3,4,4,5,5,6,6,6-Tridecafluoro-l-hexanesulfonic acid"[tw]) OR
"EC 206-587-l"[tw]) OR "EINECS 206-587-l"[tw]) OR "PFHS"[tw]) OR
"Perfluorhexan-l-sulfonsaure"[tw]) OR "Perfluorohexane sulfonic acid
(PFHxS)"[tw]) OR "Perfluorohexane-1-sulphonic acid"[tw]) OR "acide
perfluorohexane-l-sulfonique"[tw]) OR "acido perfluorohexano-1-
sulfonico"[tw]) OR "perfluorohexane-l-sulphonic acid"[tw]) OR
"perfluorohexanesulfonic acid"[tw]) OR "Ammonium
Perfluorohexanesulfonate"[tw]) OR "Ammonium
perfluorohexanesulfonate"[tw]) OR "PFHxS-H3N"[tw]) OR "PFHxS-K"[tw])
OR "Potassium Perfluorohexanesulfonate"[tw]) OR "Potassium
perfluorohexanesulfonate"[tw]) OR "Lithium
Perfluorohexanesulfonate"[tw]) OR "Lithium
perfluorohexanesulfonate"[tw]) OR "PFHxS-Li"[tw]) AND
("2019/05/03"[Date - Publication]: "3000"[Date - Publication])
116
This document is a draft for review purposes only and does not constitute Agency policy.
B-2 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Database
Terms
Hits
PubMed
04/06/2021
(("108427-53-8"[rn] OR "355-46-4"[rn] OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
Tridecafluorohexane-l-sulfonic acid"[tw] OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
tridecafluoro-l-Hexanesulfonic acid"[tw] OR "1-Hexanesulfonic acid,
l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-"[tw] OR "1-Hexanesulfonic acid,
tridecafluoro-"[tw] OR "1-Perfluorohexanesulfonic acid"[tw] OR
"Perfluoro-l-hexanesulfonate"[tw] OR "Perfluorohexane sulfonic acid"[tw]
OR "Perfluorohexane-1-sulphonic acid"[tw] OR
"Perfluorohexanesulfonate"[tw] OR "Perfluorohexanesulfonic acid"[tw]
OR "Perfluorohexylsulfonate"[tw] OR "Tridecafluorohexanesulfonic
acid"[tw] OR "tridecafluoro-l-Hexanesulfonic acid"[tw] OR "PFHxS"[tw])
AND ("2020/04/29"[Date - Publication]: "3000"[Date - Publication])
("108427-53-8"[rn] OR "423-50-7"[rn] OR "1-Hexanesulfonic acid,
1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-, ion(l-)"[tw] OR "PFHxS ion(l-)"[tw]
OR "PFHxS_ion"[tw] OR "Perfluorohexanesulfonate"[tw] OR
"Tridecafluorohexane-l-sulfonate"[tw] OR "perfluorohexyl sulfonate"[tw]
OR "1,1,2,2,3,3,4,4,5,5,6,6,6-Tridecafluoro-l-hexanesulfonyl fluoride" [tw]
OR "1-Hexanesulfonyl fluoride, l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-"[tw]
OR "1,1,2,2,3,3,4,4,5,5,6,6,6-Tridecafluoro-l-hexanesulfonic acid"[tw] OR
"EC 206-587-l"[tw] OR "EINECS 206-587-l"[tw] OR "PFHS"[tw] OR
"Perfluorhexan-l-sulfonsaure"[tw] OR "Perfluorohexane sulfonic acid
(PFHxS)"[tw] OR "Perfluorohexane-l-sulphonic acid"[tw] OR "acide
perfluorohexane-l-sulfonique"[tw] OR "acido perfluorohexano-1-
sulfonico"[tw] OR "perfluorohexane-l-sulphonic acid"[tw] OR
"perfluorohexanesulfonic acid"[tw] OR "Ammonium
Perfluorohexanesulfonate"[tw] OR "Ammonium
perfluorohexanesulfonate"[tw] OR "PFHxS-H3N"[tw] OR "PFHxS-K"[tw] OR
"Potassium Perfluorohexanesulfonate"[tw] OR "Potassium
perfluorohexanesulfonate"[tw] OR "Lithium
Perfluorohexanesulfonate"[tw] OR "Lithium
perfluorohexanesulfonate"[tw] OR "PFHxS-Li"[tw]) AND
("2020/04/29"[Date - Publication]: "3000"[Date - Publication])
28
This document is a draft for review purposes only and does not constitute Agency policy.
B-3 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Database
Terms
Hits
Web of Science
7/27/2017
TS="l,l,2,2,3,3,4,4,5,5,6,6,6-Tridecafluorohexane-l-sulfonic acid" OR
TS="l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-l-Hexanesulfonic acid" OR
TS="l-Hexanesulfonic acid, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-" OR
TS="l-Hexanesulfonic acid, tridecafluoro-" OR TS="1-
Perfluorohexanesulfonic acid" OR TS="Perfluoro-l-hexanesulfonate" OR
TS="Perfluorohexane sulfonic acid" ORTS="Perfluorohexane-l-sulphonic
acid" ORTS="Perfluorohexanesulfonate" ORTS="Perfluorohexanesulfonic
acid" ORTS="Perfluorohexylsulfonate" OR
TS="Tridecafluorohexanesulfonic acid" OR TS="tridecafluoro-l-
Hexanesulfonic acid" OR TS="PFHxS"
394
Web of Science
04/29/2020
((TS="l,l,2,2,3,3,4,4,5,5,6,6,6-Tridecafluorohexane-l-sulfonic acid" OR
TS="l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-l-Hexanesulfonic acid" OR
TS="l-Hexanesulfonic acid, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-" OR
TS="l-Hexanesulfonic acid, tridecafluoro-" OR TS="1-
Perfluorohexanesulfonic acid" OR TS="Perfluoro-l-hexanesulfonate" OR
TS="Perfluorohexane sulfonic acid" ORTS="Perfluorohexane-l-sulphonic
acid" OR TS="Perfluorohexanesulfonate" OR TS="Perfluorohexanesulfonic
acid" ORTS="Perfluorohexylsulfonate" OR
TS="Tridecafluorohexanesulfonic acid" OR TS="tridecafluoro-l-
Hexanesulfonic acid" OR TS="PFHxS") AND PY=2019-2020)
((TS="l-Hexanesulfonic acid, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-, ion(l-
)" OR TS="PFHxS ion(l-)" ORTS="PFHxS_ion" OR
TS="Perfluorohexanesulfonate" OR TS="Tridecafluorohexane-l-sulfonate"
OR TS="perfluorohexyl sulfonate" OR TS="1,1,2,2,3,3,4,4,5,5,6,6,6-
Tridecafluoro-l-hexanesulfonyl fluoride" OR TS="l-Hexanesulfonyl
fluoride, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-" OR
TS="l,l,2,2,3,3,4,4,5,5,6,6,6-Tridecafluoro-l-hexanesulfonic acid" OR
TS="EC 206-587-1" ORTS="EINECS 206-587-1" ORTS="PFHS" OR
TS="Perfluorhexan-l-sulfonsaure" OR TS="Perfluorohexane sulfonic acid
(PFHxS)" ORTS="Perfluorohexane-l-sulphonicacid" ORTS="acide
perfluorohexane-l-sulfonique" ORTS="acido perfluorohexano-1-
sulfonico" ORTS="perfluorohexane-l-sulphonic acid" OR
TS="perfluorohexanesulfonic acid" ORTS="Ammonium
Perfluorohexanesulfonate" OR TS="Ammonium
perfluorohexanesulfonate" ORTS="PFHxS-H3N" ORTS="PFHxS-K" OR
TS="Potassium Perfluorohexanesulfonate" ORTS="Potassium
perfluorohexanesulfonate" ORTS="Lithium Perfluorohexanesulfonate" OR
TS="Lithium perfluorohexanesulfonate" ORTS="PFHxS-Li") AND PY=2019-
2020)
90
This document is a draft for review purposes only and does not constitute Agency policy.
B-4 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Database
Terms
Hits
Web of Science
04/06/2021
((TS="l,l,2,2,3,3,4,4,5,5,6,6,6-Tridecafluorohexane-l-sulfonic acid" OR
TS="l,l,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-l-Hexanesulfonic acid" OR
TS="l-Hexanesulfonic acid, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-" OR
TS="l-Hexanesulfonic acid, tridecafluoro-" OR TS="1-
Perfluorohexanesulfonic acid" OR TS="Perfluoro-l-hexanesulfonate" OR
TS="Perfluorohexane sulfonic acid" ORTS="Perfluorohexane-l-sulphonic
acid" ORTS="Perfluorohexanesulfonate" ORTS="Perfluorohexanesulfonic
acid" ORTS="Perfluorohexylsulfonate" OR
TS="Tridecafluorohexanesulfonic acid" OR TS="tridecafluoro-l-
Hexanesulfonic acid" OR TS="PFHxS") AND PY=2020-2021)
((TS="l-Hexanesulfonic acid, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-, ion(l-
)" OR TS="PFHxS ion(l-)" ORTS="PFHxS_ion" OR
TS="Perfluorohexanesulfonate" OR TS="Tridecafluorohexane-l-sulfonate"
OR TS="perfluorohexyl sulfonate" OR TS="1,1,2,2,3,3,4,4,5,5,6,6,6-
Tridecafluoro-l-hexanesulfonyl fluoride" OR TS="l-Hexanesulfonyl
fluoride, 1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-" OR
TS="l,l,2,2,3,3,4,4,5,5,6,6,6-Tridecafluoro-l-hexanesulfonic acid" OR
TS="EC 206-587-1" ORTS="EINECS 206-587-1" ORTS="PFHS" OR
TS="Perfluorhexan-l-sulfonsaure" OR TS="Perfluorohexane sulfonic acid
(PFHxS)" ORTS="Perfluorohexane-l-sulphonicacid" ORTS="acide
perfluorohexane-l-sulfonique" ORTS="acido perfluorohexano-1-
sulfonico" ORTS="perfluorohexane-l-sulphonic acid" OR
TS="perfluorohexanesulfonic acid" ORTS="Ammonium
Perfluorohexanesulfonate" ORTS="Ammonium
perfluorohexanesulfonate" ORTS="PFHxS-H3N" ORTS="PFHxS-K" OR
TS="Potassium Perfluorohexanesulfonate" ORTS="Potassium
perfluorohexanesulfonate" ORTS="Lithium Perfluorohexanesulfonate" OR
TS="Lithium perfluorohexanesulfonate" ORTS="PFHxS-Li") AND PY=2020-
2021)
69
Toxline
7/21/2017
(108427-53-8[rn] OR 355-46-4[rn] OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
Tridecafluorohexane-l-sulfonic acid" OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
tridecafluoro-l-Hexanesulfonic acid" OR "1-Hexanesulfonic acid,
1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-" OR "1-Hexanesulfonic acid,
tridecafluoro-" OR "1-Perfluorohexanesulfonic acid" OR "Perfluoro-1-
hexanesulfonate" OR "Perfluorohexane sulfonic acid" OR
"Perfluorohexane-l-sulphonic acid" OR "Perfluorohexanesulfonate" OR
"Perfluorohexanesulfonic acid" OR "Perfluorohexylsulfonate" OR
"Tridecafluorohexanesulfonic acid" OR "tridecafluoro-l-Hexanesulfonic
acid" OR "PFHxS") AND ( ANEUPL [org] OR BIOSIS [org] OR CIS [org] OR
DART [org] OR EMIC [org] OR EPIDEM [org] OR HEEP [org] OR HMTC [org]
OR IPA [org] OR RISKUNE [org] OR MTGABS [org] OR NIOSH [org] OR NTIS
[org] OR PESTAB [org] OR PPBIB [org]) [not] PubMed [org] [not] pubdart
[org]
0
This document is a draft for review purposes only and does not constitute Agency policy.
B-5 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Database
Terms
Hits
SCOPUS (new
search)
4/26/2021
("108427-53-8" OR "355-46-4" OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
Tridecafluorohexane-l-sulfonic acid" OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
tridecafluoro-l-Hexanesulfonic acid" OR "1-Hexanesulfonic acid,
1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-" OR "1-Hexanesulfonic acid,
tridecafluoro-" OR "1-Perfluorohexanesulfonic acid" OR "Perfluoro-1-
hexanesulfonate" OR "Perfluorohexane sulfonic acid" OR
"Perfluorohexane-l-sulphonic acid" OR "Perfluorohexanesulfonate" OR
"Perfluorohexanesulfonic acid" OR "Perfluorohexylsulfonate" OR
"Tridecafluorohexanesulfonic acid" OR "tridecafluoro-1-Hexanesulfonic
acid" OR "PFHxS")
("108427-53-8" OR "423-50-7" OR "1-Hexanesulfonic acid,
1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-, ion(l-)" OR "PFHxS ion(l-)" OR
"PFHxSJon" OR "Perfluorohexanesulfonate" OR "Tridecafluorohexane-1-
sulfonate" OR "perfluorohexyl sulfonate" OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
Tridecafluoro-l-hexanesulfonyl fluoride" OR "1-Hexanesulfonyl fluoride,
1,1,2,2,3,3,4,4,5,5,6,6,6-tridecafluoro-" OR "1,1,2,2,3,3,4,4,5,5,6,6,6-
Tridecafluoro-l-hexanesulfonic acid" OR "EC 206-587-1" OR "EINECS 206-
587-1" OR "PFHS" OR "Perfluorhexan-l-sulfonsaure" OR "Perfluorohexane
sulfonic acid (PFHxS)" OR "Perfluorohexane-l-sulphonic acid" OR "acide
perfluorohexane-1-sulfonique" OR "acido perfluorohexano-l-sulfonico"
OR "perfluorohexane-l-sulphonic acid" OR "perfluorohexanesulfonic acid"
OR "Ammonium Perfluorohexanesulfonate" OR "Ammonium
perfluorohexanesulfonate" OR "PFHxS-H3N" OR "PFHxS-K" OR "Potassium
Perfluorohexanesulfonate" OR "Potassium perfluorohexanesulfonate" OR
"Lithium Perfluorohexanesulfonate" OR "Lithium
perfluorohexanesulfonate" OR "PFHxS-Li")
1,208
TSCATS2, TSCA
recent notices
7/21/2017
84-66-2
10
84-66-2 (8E OR FYI) TSCA
0 recent notice
Table B-2. Processes used to augment the search of core databases for PFHxS
(355-46-4)
System used
Selected key reference(s) or sources
References
identified
TSCATS3
TSCATS2
(httpsi//yosemite,epa,gov/oppts/epatscat8,nsf/ReportSearch?OpenForm)
2
Chemical Data Access Tool (CDAT)
(httpsi//iava,epa.gov/oppt chemical search/)
1
ChemView (https://java.epa.gov/chemview)
1
Resources searched for
physiochemical
property information
Agency for Toxic Substances and Disease Registry (ATSDR)
(https://www.atsdr.cdc.gov/)
Australian National Industrial Chemicals Notification and Assessment
Scheme (NICNAS) (https://www.nicnas.gov.au/chemical-information)
CAMEO Chemicals (https://cameochemicals.noaa.gov/)
5
Canada DSL List (http://webnet.oecd.org/CCRWEB/Search.aspx)
ChemlDplus (https://chem.nlm.nih.gov/chemidplus/)
This document is a draft for review purposes only and does not constitute Agency policy.
B-6 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
References
System used
Selected key reference(s) or sources
identified
ChemSpider (http://www,chemspider.com/)
Chemical Risk Information Platform (CHRIP1
(http://www.nite,KO.jp/en/chem/chrip/chrip search/systemTopl
CRC Handbook of Chemistry and Physics
(httpi//hbcponline.com/faces/contents/ContentsSearch.xhtml;isessionid=
940887S1S6F724E0E94SD3A6D04S4891)
ECHA Information on Chemicals (https://echa.europa.eu/)
eChemPortal (https://www.echemportal.org/echemportal/index.action)
SRC Fate Pointers (http://esc.svrres.com/fatepointer/search.asp)
Hazardous Substances Data Bank (HSDB) https://toxnet.nlm.nih.gov/cgi-
bin/sis/htmlgen?HSDB)
HSNO Chemical Classification and Information Database (CCID)
(http://www.epa.govt.nz/search-databases/Pages/HSNO-CCID.aspx)
Integrated Risk Information System (IRIS) (https://www.epa.gov/iris)
IARC Monographs (http://www.inchem.org/pages/iarc.html)
J-Check
(http://www.safe.nite.go.ip/icheck/search.action7request locale=en)
Kirk-Othmer Encyclopedia of Chemical Technology
(http://onlinelibrary.wiley.com/mrw/advanced/search?doi=10.1002/0471
238961)
NIEHS (https://www.niehs.nih.gov/)
OS HA Occupational Chemical Database
(https://www.osha.gov/chemicaldata/)
PubChem (https://pubchem.ncbi.nlm.nih.gov/search/index.html)
Ullmann's Encyclopedia
http://onlinelibrary.wiley.com/mrw/advanced/search?doi=10.1002/1435
6007)
US EPA ACToR (https://actor.epa.gov/actor/home.xhtml)
US EPA ChemView (https://iava.epa.gov/chemview)
US EPA Substance Registry Services (SRS)
(https://ofmpub.epa.gov/sor internet/registrv/substreg/searchandretriev
e/substancesearch/search.do)
USEPA CDAT (https://iava.epa.gov/oppt chemical search/)
US EPA Chemistry Dashboard (https://comptox.epa.gov/dashboard/)
Web based search for chemical manufacturer documents
Resources searched for
ATSDR (http://www.atsdr.cdc.gov/substances/index.asp)
2
health effects,
CalEPA OEHHA (http://www.oehha.ca.gov/risk.html)
toxicokinetic, and
OEHHAToxicity Criteria Database
mechanistic information
(http://www.oehha.ca.gov/tcdb/index.asp)
CPSC (http://www.cpsc.gov)
ECHA (http://echa.europa.eu/information-on-chemicals)
European Union Risk Assessment Reports
(https://ec.europa.eu/irc/en/publications-list)
EFSA Europe (http://www.efsa.europa.eu/)
eChemPortalb(http://www.echemportal.org/echemportal/participant/pag
e.action?pagelD=9)
eChemPortalb(http://www.echemportal.org/echemportal/participant/pag
e.action?pagelD=9)
Environment Canada
(http://www,ec,gc,ca/default,asp?lang=En&n=ECD3SC36)
This document is a draft for review purposes only and does not constitute Agency policy.
B-7 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
System used
Selected key reference(s) or sources
References
identified
Health Canada (httpsi//www,canada,ca/en/health-canada,html)
USEPA NSCEP (httpsi//www,epa.gov/nscep)
FDA (http://www.fda .gov/)
Federal Docket (httpi//www,regulations,gov)
IARC (httpi//monographs,iarc,fr/ENG/Classification/index,php)
ITER (http://www.tera,org/iter/)
Japan Existing Chemical Data Base
(http://dra4.nihs,go,ip/mhlw data/isp/SearchPageENG.jsp)
NIEHS (http://www.mehs,nih.gov/)
NICNAS (http://www.nicnas.gov.au/chemical-information)
NTP (http://ntpsearch.niehs.nih.gov/home)
WHO (http://www.who.int/ipcs/assessment/en/)
aOnly relevant TSCATS studies from these interfaces were added to the HERO project page.
beChemPortal includes the following databases: ACToR, AGRITOX, CCR, CCR DATA, CESAR, CHRIP, ECHA CHEM,
EnviChem, ESIS, GHS-J, HPVIS, HSDB, HSNO CCID, INCHEM, J-CHECK, JECDB, NICNAS PEC, OECD-HPV, OECD SIDS
IUCUD, SIDS UNEP, UK CCRMP Outputs, EPA-IRIS, and EPA-SRS.
B.2. TITLE AND ABSTRACT LEVEL SCREENING CRITERIA FOR THE
INITIAL LITERATURE SEARCHES
Table B-3. Title and abstract level screening criteria for the initial literature
searches
Inclusion criteria
Exclusion criteria
Population
• Humans
• Standard mammalian animal models,
including rat, mouse, rabbit, guinea pig,
hamster, monkey, dog
• Alternative animal models in standard
laboratory conditions (e.g., Xenopus,
zebrafish, minipig)
• Human or animal cells, tissues, or organs
(not whole animals); bacteria,
nonmammalian eukaryotes; other
nonmammalian laboratory species
• Ecological species
Exposure
• Exposure is to PFHxS compound
• Exposure via oral, inhalation, dermal,
intraperitoneal, or intravenous injection
routes
• Exposure is measured in air, dust, drinking
water, diet, gavage, or injection vehicle or
• Study population is not exposed to a PFHxS
compound
• Exposure is to a mixture only
This document is a draft for review purposes only and does not constitute Agency policy.
B-8 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Inclusion criteria
Exclusion criteria
via a biomarker of exposure (PFHxS levels in
whole blood, serum, plasma, or breastmilk)
Outcome
• Studies that include a measure of one or
more health effect endpoints including, but
not limited to, effects on reproduction,
development, developmental neurotoxicity,
liver, thyroid, immune system, nervous
system, genotoxicity, and cancer
• In vivo and/or in vitro studies related to
toxicity mechanisms, physiological
effects/adverse outcomes, and studies
useful for elucidating toxic modes of action
(MOAs)
• Qualitative or quantitative description of
absorption, distribution, metabolism,
excretion, toxicokinetic and/or
toxicodynamic models (e.g., PBPK, PBTK,
PBTK/TD)
• Studies addressing risks to infants, children,
pregnant women, occupational workers,
the elderly, and any other susceptible or
differentially exposed populations
Other
• Structure and physiochemical properties
• Reviews and regulatory documents
• Not on topic, including:
• Abstract only, inadequately reported
abstract, or no abstract and not considered
further because study was not potentially
relevant
• Bioremediation, biodegradation, or
chemical or physical treatment of PFHxS
compounds, including evaluation of
wastewater treatment technologies and
methods for remediation or contaminated
water and soil
• Ecosystem effects, studies in ecological
species that are not relevant to health
effects in humans
• Studies of environmental fate and transport
of PFHxS compounds in environmental
media
• Analytical methods for detecting/measuring
PFHxS compounds in environmental media
and use in sample preparations and assays.
This document is a draft for review purposes only and does not constitute Agency policy.
B-9 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Inclusion criteria
Exclusion criteria
• Studies describing the manufacture and use
of PFHxS compounds
• Not chemical-specific (studies that do not
involve testing of PFHxS compounds)
• Studies that describe measures of exposure
to PFHxS compounds without data on
associated health effects
This document is a draft for review purposes only and does not constitute Agency policy.
B-10 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table B-4. Example DistillerSR form questions to be used for title/abstract level and full text level screening for
literature search updates from 2019
Used in title/abstract and full text screening
Used in full text screening only
Source of
study if not
identified
If meets PECO and
from
Does the
Which PFAS
If meets PECO, which
endocrine outcome,
Question
database
article meet
If meets PECO, what
If supplemental, what type of
did the study
health outcome(s)
which endocrine
search?
PECO criteria?
type of evidence?
information?
report?
apply?
tags apply?
Answer
options
(can
select
multiple
options)
• Source
other
than
HERO
database
search
• Yes
• No
• Unclear
• Tag as
potentially
relevant
supplement
al
information
• Human
• Animal (mam-
malian models)
• In vitro or in silico
genotoxicity
• PBPK or PK model
• In vivo mechanistic or MOA
studies, including nonPECO
routes of exposure
(e.g., injection) and
populations
(e.g., nonmammalian)
• In vitro or in silico studies
(nongenotoxicity)
• ADME/toxicokinetic
(excluding models)
• Exposure assessment or
characterization (no health
outcome)
• PFAS mixture study (no
individual PFAS
comparisons)
• Human case reports or case
series
• Ecotoxicity studies
• Environmental fate or
occurrence (including food)
• PFBA
• PFHxA
• PFHxS
• PFNA
• PFDA
• General toxicity,
including body
weight, mortality,
and survival
• Cancer
• Cardiovascular,
including serum
lipids
• Endocrine
(hormone)
• Gastrointestinal
• Genotoxicity
• Growth (early life)
and development
• Hematological,
including
nonimmune/hepati/
renal clinical
chemistry measures
• Hepatic, including
liver measures and
• Adrenal
• Sex hormones
(e.g., androgen;
estrogen;
progesterone)
• Neuroendocrin
e
• Pituitary
• Steroidogenesis
• Thyroid
This document is a draft for review purposes only and does not constitute Agency policy.
B-ll DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Used in title/abstract and full text screening
Used in full text screening only
• Manufacture, engineering,
use, treatment, remediation,
or laboratory methods
• Other assessments or
records with no original data
(e.g., reviews, editorials,
commentaries)
serum markers
(e.g., ALT; AST)
• Immune/
inflammation
• Musculoskeletal
• Nervous system,
including behavior
and sensory
function
• Nutrition and
metabolic
• Ocular
• PBPK or PK model
• Renal, including
urinary measures
(e.g., protein)
• Reproductive
• Respiratory
• Skin and connective
tissue effects
This document is a draft for review purposes only and does not constitute Agency policy.
B-12 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
B.3. DOCUMENTATION OF LITERATURE SEARCH UPDATES AFTER APRIL 2022
Table B-5 documents the decisions regarding studies identified after April 2022, including a literature search update in April 2023
and studies identified in public comments received through the EPA docket on the draft IRIS PFDA assessment1. The table focuses
primarily on the new studies that met the assessment PECO criteria. Specifically, epidemiology studies that meet the PECO criteria were
identified; no experimental animal studies that meet the PECO criteria were identified. Table B-5 provides EPA's disposition on the
decision to incorporate these studies into the assessment as defined in draft Peer Review Charge question 1 (i.e., only incorporating
studies that may potentially change which hazards are identified, or notably affect the RfDs, or studies that directly inform the identified
key science issues); the charge question asks the peer reviewers to weigh in on EPA's disposition. These same criteria were applied to
certain categories of newly identified supplemental studies (i.e., ADME and mechanistic studies, including non-PECO exposure route
studies). Thus, in addition to providing this characterization for all studies that meet PECO criteria, Table B-5 also includes studies from
supplemental evidence categories that were determined to warrant inclusion in the assessment based on the criteria described above.
The decision to exclude other recently identified studies that meet these specific supplemental evidence categories is documented in
HAWC. Recently identified studies that meet supplemental evidence categories other than those above (e.g., exposure-only) were not
evaluated in this way and are tagged in HERO and HAWC along with other screening decisions (e.g., excluded studies).
1A total of 186 studies were submitted by the State of New Jersey Department of Environmental Protection and the Natural Resources Defense Council
(NRDC). There was a large amount of overlap between these studies and those already identified and screened for inclusion in the draft IRIS PFHxS
assessment by EPA before April 2022. Those not already identified were screened using the PFHxS assessment PECO criteria. Table B-5 lists as
"Commenter (on PFDA)" those studies meeting PECO criteria that were not identified before April 2022 and not identified through the routine EPA
literature update in 2023.
This document is a draft for review purposes only and does not constitute Agency policy.
B-13 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table B-5. Summary of decisions regarding studies identified after April 2022, including characterization of all
studies meeting PECO criteria and all supplemental studies added to the assessment syntheses
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
Immune Effects
Kaur et al. (2023)
Lit update
Antibody levels to
SARS-COV2 in
adults
Inverse association (beta -0.68, 95% CI -1.18, -
0.18)
No.
Findings are consistent with existing
evidence and have no impact on
immunosuppression conclusions,
particularly given that two of the new
studies are in adults and the draft
conclusions are primarily based on studies
in children.
(Porter et al., 2022)
Lit update
Antibody levels to
SARS-COV2 in
adults
Inverse association with IgG and neutralizing
antibodies in response to COVID vaccination
(statistical significance varied based on model)
(Zhang et al.,
2023b)
Lit update
Vaccine response
Inverse association with rubella antibodies (-
6.48% change, 95% CI -10.69, -2.07). Inverse
but not statistically significant association with
mumps antibodies in sub-population with
lower folate.
(Zhang et al., 2022)
Lit update
Infectious disease
Positive association with common cold at 3-11
years (OR 1.31, 95% CI 1.05,1.63) but not 12-
19 years
No. Existing evidence on infectious disease
is inconsistent and new studies do not
change current draft judgment.
(Pan et al., 2023)
Lit update
Asthma
No association with current asthma (OR 0.97,
95% CI 0.57,1.65 in Q4 vs Ql) or wheezing.
Inverse association with asthma attacks and
emergency visits.
No.
Existing evidence on asthma is inconsistent
and new studies do not change current
draft judgment.
Gavlord et al.
(2019)
Commenter (on
PFDA)
Asthma
No association with asthma diagnosis (OR
0.96, 95% CI 0.65, 1.44)
Averina et al. (2019)
Commenter (on
PFDA)
Asthma
Positive association with asthma (OR 2.18,
95% CI 1.08, 4.42 in Q4 vs Ql). No association
with allergies or eczema.
(Ammitzb0ll et al.,
2019)
Commenter (on
PFDA)
Multiple sclerosis
No association with multiple sclerosis (2%
change, 95% CI -9,15)
No.
This document is a draft for review purposes only and does not constitute Agency policy.
B-14 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
Gavlord et al.
(2020)
Commenter (on
PFDA)
Celiac disease
Positive association with celiac disease (OR
1.72, 95% CI 0.85, 3.49) with stronger effect in
women (OR 3.24, 95% CI 1.04, 10.11)
Mixed results for autoimmune conditions in
new studies would not influence draft
conclusions on immune effects.
Developmental Effects
Wang et al. (2022)
Lit update
Fetal growth
restriction (Birth
length (BL); head
circumference
(HC); birthweight
(BWT))
No sex-specific associations were observed for
birth length (BL), birth weight (BWT) and head
circumference (HC) endpoints.
BL Male P = -0.080; 95%CI: -0.062, 0.222;
BL Female P = -0.004; 95%CI: -0.310, 0.303.
HC Male P = 0.005; 95%CI: -0.180, 0.191;
HC Female P = -0.110; 95%CI: -0.345, 0.125.
BWT Male P = 0.024; 95%CI: -0.140, 0.188;
BWT Female P = -0.062; 95%CI: -0.291, 0.166.
No.
Null results observed for birth length, birth
weight and head circumference endpoints
in both female and male neonates would
not change the current draft judgment for
fetal growth restriction.
Peterson et al.
(2022)
Lit update
Fetal growth
restriction
No associations were evident across fetal
measures in relation to PFHxS exposures.
No.
Null results for fetal biometric endpoints
would not change the current draft
judgment for fetal growth restriction.
Padula et al. (2023)
Lit update
Fetal growth
restriction,
gestational
duration
No associations were evident across fetal
growth and gestational duration endpoints
[gestational age p= 0.02; 95%CI: -0.19, 0.23;
birth weight for gestational age p= -0.06;
95%CI: -0.18, 0.06; term low birth weight OR=
1.14; 95%CI: 0.46, 2.84; small for gestational
age OR= 1.25; 95%CI: 0.84,1.87; large for
gestational age OR= 0.86; 95%CI: 0.59,1.25;
preterm birth OR= 0.97; 95%CI: 0.61,1.55.
No.
Null results for all fetal growth and
gestational duration endpoints would not
change the current draft judgment for
either gestational duration or fetal growth
restriction.
Ouidir et al. (2020)
Commenter (on
PFDA)
Fetal growth
restriction
Per each PFHxS IQR increase, a statistically
significant longitudinal decrease in head
circumference (P = -0.22 mm; p-value: <0.05)
and increases in longitudinal biparietal
diameter (P = 0.07 mm; p-value: <0.05), and
femur length (P = 0.12 mm; p-value: <0.001)
No.
Study population was previously reported
in a publication already in the assessment
Buck Louis et al. (2018). New results for
longitudinal in utero measurements from
This document is a draft for review purposes only and does not constitute Agency policy.
B-15 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
were detected. Results were null for
abdominal circumference (P = 0.11 mm),
occipital-frontal diameter changes (P = -0.04
mm) and estimated fetal growth (P = 3.27 g)
(p-value/CIs not provided).
ultrasonography would not change the
current draft judgment.
Petroff etal, (2023)
Lit update
Gestational age
No association between PFHxS exposure and
gestational age (B = 0.04 ± 0.21; p=0.85).
No.
Null results for gestational age would not
change the current draft judgment for
gestational duration.
Yu et al. (2022)
Lit update
Preterm birth
Results were mixed with a non-significant
increase in risk seen for untransformed data
(OR=1.76; 95%CI: 0.91, 3.40 per each ng/mL
increase) only; transformed results were null
(OR=0.93; 95%CI: 0.80,1.08 per each In-unit
increase).
No.
Small increased risks here along with the
null results in Padula et al, (2023) and Liao
et al, (2022b) would not change the current
draft judgment for gestational duration.
Liao et al, (2022b)
Lit update
Preterm birth
Results were mixed with a statistically
significant decrease in preterm birth per each
log 10 increase (OR=0.73; 95%CI: 0.39,1.38)
driven by tertile 3 (OR=0.60; 95%CI: 0.37,
0.98); results were null for tertile 2 (OR=0.97;
95%CI: 0.63,1.50) relative to tertile 1.
No.
Inconsistent new results in three new
studies including decreased risk reported
here combined with increased risk by Yu et
al, (2022) and null results in Padula et al,
(2023) above would not change the current
draft judgment for gestational duration.
Wang et al, (2016)
Commenter (on
PFDA)
Gestational
duration
This exposure study showed a statistically
significant increase between gestational age
and concentrations of PFHxS in cord blood.
No.
Although an association was reported in an
exposure prediction model examining
gestational age and PFHxS, the study design
and analyses would likely preclude this
from inclusion into the synthesis. These
data would not contribute to the
gestational duration judgments.
Hong et al, (2022)
Lit update
Spontaneous
abortion
Inverse association (OR=0.05; 95% CI: 0.00,
7.36)
No.
Updated analysis of study that is already
included in the draft assessment.
This document is a draft for review purposes only and does not constitute Agency policy.
B-16 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
Li et al. (2022a)
Lit update
Anogenital
distance
Positive association with two AGD measures
(p<0.05)
No.
New study adds to existing inconsistency in
the AGD evidence and would not change
the draft judgment.
Hepatic
Borghese et al.
(2022)
Lit update
Liver enzymes
Positive association with AST, GGT, and ALP,
positive but not statistically significant
association with ALT and bilirubin
No.
New studies are consistent with the
existing studies and would not change the
draft judgment.
Liao et al. (2023)
Lit update
Liver enzymes
Positive association with bilirubin but not ALT,
AST, or GGT
Kim et al. (2023b)
Lit update
Liver enzymes
Positive but not statistically significant
associations with ALT, AST, and GGT
Yao et al. (2020)
Commenter (on
PFDA)
Liver enzymes
Positive association with ALT, AST, GGT
(statistically significant for GGT)
Salihovic et al.,
(2019, 6324314)
Commenter (on
PFDA)
Bile acid levels
(liver)
Inverse correlations with most bile acids
(statistically significant for GDCA)
Rantakokko et al.
(2015)
Commenter (on
PFDA)
Non-alcoholic fatty
liver disease
Inverse association with lobular inflammation
(OR 0.02, 95% CI <0.01, 0.53 for 2-4 foci per
200x field)
No.
While there are no studies of clinical liver
disease available for PFHxS in the current
draft, the new studies are inconsistent and
would not change the draft judgment of
slight for hepatic effects.
E et al. (2023)
Lit update
Liver disease
No association with liver problems (OR 0.97,
95% CI 0.72,1.30). Positive but not statistically
significant association with ALT.
Nilsson et al.
(2022b)
Lit update
Liver problems
Positive association with non-alcoholic fatty
liver disease in women but not men, with
strongest association in postmenopausal
women (OR 2.50, 95% CI 1.29, 4.85 in Q4 vs
Ql)
Cancer
Feng et al. (2022a)
Lit update
Breast cancer
No association with breast cancer (OR = 0.93,
95% CI: 0.79,1.09) per unit increase in In-
transformed plasma PFHxS levels.
No.
Inconsistent results across the new studies
showing increased risk, decreased risk and
This document is a draft for review purposes only and does not constitute Agency policy.
B-17 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
Li et al, (2022b)
Lit update
Breast cancer
Decreased risk for breast cancer (OR = 0.73,
95% CI: 0.63, 0.87) per SD increase in In-
transformed PFHxS from the adjusted model -
without LASSO (see Table S3).
no association of PFHxS and breast cancer.
In addition, one new breast cancer study
reports on the same study population as a
publication already in the assessment
Wiels0e et al, (2017). One previous study
reported significantly increased risk of
breast cancer among women <= 50 years of
age who were estrogen receptor positive;
and non-significantly decreased risk of
breast cancer among women who were
estrogen receptor negative and > 50 years
of age. Another study reported significantly
decreased risk for some genotypes.
The only study reporting on liver cancer did
not find an association with PFHxS.
The only study of renal cancer reported a
significant association observed for renal
cancer that dissipated when controlling for
other PFAS.
The available epidemiologic evidence on
PFDA and the risk of cancer remains
inadequate; the new studies are not
impactful.
Wiels0e et al,
(2018)
Commenter (on
PFDA)
Breast cancer
Increased risk for breast cancer (OR 5.45, 95%
CI 1.26, 23.8) in high vs. low PFHxS exposure
for one genotype).
Lee et al. (2020)
Commenter (on
PFDA)
Breast cancer
No association of PFHxS with mammographic
density, a strong predictor of breast cancer
(beta -0.02, p-value 0.95).
Goodrich et al,
(2022)
Lit update
Liver cancer
No association of PFHxS with liver cancer (OR
= 1.10, 95% CI: 0.56, 2.30) for PFHxS greater
than the 90th% vs less than 90th%.
Shearer et al, (2021)
Commenter (on
PFDA)
Renal cancer
Increased risk of renal cell carcinoma with
PFHxS per unit increase in log2-transformed
serum PFHxS (OR=1.27; 95% CI: 1.03, 1.56)
that attenuated when controlling for other
PFAS (OR=1.12; 95% CI: 0.88, 1.43).
Neurodevelopment
Luo et al, (2022a)
Lit update
Broad
neurodevelopmen
tal scale
Inverse but not statistically significant
association with cognitive, language, motor,
and social-emotional scores, but statistically
significant positive association with adaptive
behavior score
No.
There is inconsistency for
neurodevelopmental effects in the current
draft, and the new studies would not
This document is a draft for review purposes only and does not constitute Agency policy.
B-18 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
Oh etal. (2022b)
Lit update
Autism,
developmental
delay
Positive but not statistically significant
associations with autism spectrum disorder
and developmental delay
influence the draft judgment of slight
evidence.
Zhou et al. (2023)
Lit update
Broad
neurodevelopmen
tal scale
Inverse association with communication and
motor at 6 mos but inconsistent findings for
other measures (problem solving, personal-
social) and other visits (2,12, and 24 mos)
Li et al. (2023c)
Lit update
Broad
neurodevelopmen
tal scale
Positive association with persistently low
trajectory for communication (p<0.05), gross
motor, problem solving ability (p<0.05), and
personal-social skills, but not fine motor
Oulhote et al.
(2019)
Commenter (on
PFDA)
Broad
neurodevelopmen
tal scale
Positive association with Boston Naming Test.
No association with Strengths and Difficulties
Questionnaire.
(van Larebeke et al.,
2022)
Lit update
Broad
neurodevelopmen
tal scale
Inverse (favorable) association with incorrect
responses on the Continuous Performance
Test but not other test results
Kim et al. (2023a)
Lit update
ADHD scale
Positive though non-monotonic association
with ADHD rating scale at 8 yrs, dependent on
age at exposure measurement and sex
Male Reproductive
Luo et al. (2022b)
Lit update
Semen parameters
No association with sperm concentration of
motility
No.
Evidence is inconsistent and the new
studies would not influence the draft
conclusion.
Ma etal. (2021)
Commenter (on
PFDA)
Semen parameters
No association sperm concentration, motility,
or morphology
Rivera-Nunez et al.
(2023)
Lit update
Reproductive
hormones
Positive association with T (p<0.05), no
association with free T, El, E2, E3
No.
Evidence is inconsistent in existing studies
and the new studies would not influence
Guoetal. (2023)
Lit update
Reproductive
hormones
No association with testosterone or estradiol
(included boys and girls)
This document is a draft for review purposes only and does not constitute Agency policy.
B-19 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
Nian et al. (2020)
Commenter (on
PFDA)
Reproductive
hormones
No association with total testosterone (beta
0.079, 95% CI -0.009, 0.166 per In-unit
change), FSH, or LH
the draft synthesis conclusion of
indeterminate evidence.
Female Reproductive
Hong et al. (2022)
Lit update
In vitro
fertilization
outcomes
No association with oocyte maturation rate,
fertilization rate, high quality embryo rate.
Inverse but not statistically significant
(OR=0.60, 95% CI 0.12, 2.96) for clinical
pregnancy
No.
Evidence of an association with fecundity
and infertility is inconsistent across new
studies and was similarly inconsistent
across existing studies. The conclusion of
indeterminate evidence would likely remain
the same.
Cohen et al. (2023)
Lit update
Fecundity,
pregnancy
No association with time to pregnancy or odds
of clinical pregnancy
Luo et al. (2022c)
Lit update
Fecundity,
infertility
Lower odds of infertility (OR 0.61, 95% CI 0.45,
0.82) and higher fecundability
Tan et al. (2022)
Lit update
Infertility
Lower odds of infertility (non-monotonic
across quartiles and not statistically
significant)
(Whitworth et al.,
2016)
Commenter (on
PFDA)
Fecundity
No association (FR 0.97, 95% CI 0.90,1.1)
Ma et al. (2021)
Commenter (on
PFDA)
In vitro
fertilization
outcomes,
pregnancy
Fewer zygotes and good quality embryos with
higher exposure. No association with clinical
pregnancy.
Wang et al. (2019)
Commenter (on
PFDA)
Polycystic ovarian
syndrome
Positive but not statistically significant
association with PCOS-related infertility (OR
2.08, 95% CI 0.88, 4.93 in 3rd vs. 1st tertile)
No.
Existing evidence on gynecological
conditions is inconsistent and there is
considerable uncertainty due to potential
reverse causation. The new study does not
inform this uncertainty.
Rivera-Nunez et al.
(2023)
Lit update
Reproductive
hormones
Positive association with El, E2, E3 (p<0.05),
no association with T, FT
No.
This document is a draft for review purposes only and does not constitute Agency policy.
B-20 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
Nian et al. (2020)
Commenter (on
PFDA)
Reproductive
hormones
No association with total testosterone (beta -
0.029, 95% CI -0.090, 0.032 per In-unit
change), FSH, or LH
New studies would not change the current
draft judgment.
Liu et al. (2020a)
Commenter (on
PFDA)
Reproductive
hormones
Positive association with estradiol (6.8%
change, 95% CI 2.2,11.6)
Lin et al. (2022)
Lit update
Postpartum
hemorrhage
Higher odds of postpartum hemorrhage (OR
3.42, 95% CI 1.45, 8.07)
No.
Single study of the outcome and evidence is
not strong enough to increase certainty in
the evidence for female reproductive
effects.
Urinary
(Nilsson et al.,
2022b)
Lit update
Kidney disease,
urate
No association with kidney disease (OR 0.90,
95% CI 0.76,1.08) or urate
No.
Existing studies are inconsistent with
considerable uncertainty due to potential
reverse causation. The new studies do not
inform this uncertainty.
Liang et al. (2023)
Lit update
Glomerular
filtration rate
Higher GFR (not statistically significant)
Sood et al. (2019)
Commenter (on
PFDA)
Glomerular
filtration rate
Inverse but not statistically significant
association with eGFR (beta -10.3, 95% CI -
23.6, 3.0)
Feng et al. (2022b)
Lit update
Hyperuricemia
No association with hyperuricemia
Arrebola et al.
(2019)
Commenter (on
PFDA)
Hyperuricemia
Positive but not statistically significant
association with hyperuricemia (OR 1.33, 95%
CI 0.70, 2.54)
Yao et al. (2020)
Commenter (on
PFDA)
Uric acid
Positive association with uric acid (beta 8.44,
95% CI 2.17, 15.09)
Cardiometabolic
Haug et al. (2023)
Lit update
Serum lipids
No association with HDL or LDL cholesterol
No.
Mixed results from the new studies would
not change the current draft judgment.
Donat-Vargas et al.
(2019b)
Commenter (on
PFDA)
Serum lipids,
hypertension
No association with total cholesterol,
triglycerides, or hypertension
This document is a draft for review purposes only and does not constitute Agency policy.
B-21 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
(Batzella et al.,
2022)
Lit update
Serum lipids
Positive association with total cholesterol
(beta 1.74, 95% CI 1.36, 2.13) and LDL-
cholesterol
(Morgan et al.,
2023)
Lit update
Serum lipids
No association with total cholesterol or LDL-
cholesterol (crude analysis only)
(Nilsson et al.,
2022b)
Lit update
Serum lipids,
blood pressure,
cardiovascular
disease
Positive association with total cholesterol and
LDL-cholesterol in cross-sectional but not
prospective analysis. No association with high
blood pressure (OR 0.92, 95% CI 0.83,1.03) or
cardiovascular disease (OR 0.96, 95% CI 0.81,
1.15)
Yao et al. (2020)
Commenter (on
PFDA)
Serum lipids,
blood glucose
Positive association with total cholesterol
(beta 6.98, 95% CI 3.06,11.14), triglycerides,
and blood glucose
Mitro et al. (2020)
Lit update
Blood pressure
No association with blood pressure, BMI,
waist circumference, mid-upper arm
circumference, or skinfold thickness
Sood et al. (2019)
Commenter (on
PFDA)
Blood pressure
No association with blood pressure (beta 0.3,
95% CI-0.1, 0.7)
Lind et al. (2018)
Commenter (on
PFDA)
Carotid artery
intima-media
thickness
Positive association with IMT thickness (beta
0.015, 95% CI 0.005, 0.0025)
No.
These results support coherence with
serum lipids but would not change the
current draft judgment.
Li et al. (2023b)
Lit update
Cardiovascular
disease
No association with acute coronary syndrome
No.
New study contributes to existing
inconsistency and would not change the
current draft judgment.
Yang et al. (2022)
Lit update
Gestational
hypertension
Lower odds of gestational hypertension (OR
0.66, 95% CI 0.35,1.24) and lower continuous
blood pressure
No.
This document is a draft for review purposes only and does not constitute Agency policy.
B-22 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
Huoetal. (2020)
Lit update
Gestational
hypertension
No association with gestational hypertension
(OR 0.80, 95% CI 0.44,1.47) or preeclampsia
(OR 1.05, 95% CI 0.60, 1.83)
New studies contribute to existing
inconsistency and would not change the
current draft judgment.
Zhu and Bartell
(2022)
Lit update
Gestational
hypertension
Small positive association with hypertensive
disorders in pregnancy (OR 1.03, 95% CI 1.02,
1.04)
Xu etal. (2022)
Lit update
Gestational
diabetes
Inverse association with gestational diabetes
(OR 0.09, 95% CI 0.03, 0.22 in third tertile),
inverse association with continuous glucose
levels in oral glucose tolerance test
No.
Existing studies are inconsistent and new
studies would not change the current draft
judgment.
Zhang et al. (2023a)
Lit update
Gestational
diabetes
Positive association with gestational diabetes
(OR 3.46, 95% CI 1.64, 6.30 in 3rd tertile)
Xu et al. (2020)
Lit update
Gestational
diabetes
No association with gestational diabetes (OR
0.79, 95% CI 0.46, 1.31 in Q4 vs Ql)
Li et al. (2020)
Commenter (on
PFDA)
Gestational blood
glucose
Positive but not statistically significant
association with blood glucose in oral glucose
tolerance test (beta 0.07, 95% CI -0.06, 0.21)
Dunder et al. (2023)
Lit update
Blood glucose
No association with blood glucose
No.
Existing and new studies are primarily null,
and the current draft judgment is unlikely
to change.
Christensen et al.
(2016)
Commenter (on
PFDA)
Diabetes
No association with diabetes (OR 0.98, 95 % CI
0.69,1.16) or pre-diabetes (OR 1.00, 95% CI
0.77,1.16)
(Park et al., 2022)
Lit update
Diabetes
Positive association with incident diabetes (OR
1.58, 95% CI 1.13, 2.21 in T3 vs Tl) but not
monotonic across tertiles
Donat-Vargas et al.
(2019a)
Commenter (on
PFDA)
Diabetes risk,
insulin resistance
No increase in diabetes risk or HOMA-IR
Kim et al. (2015)
Commenter (on
PFDA)
Insulin resistance
No association with HOMA (beta -0.08, 95% CI
-0.68, 0.52)
This document is a draft for review purposes only and does not constitute Agency policy.
B-23 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
Mehta etal. (2021)
Commenter (on
PFDA)
Insulin resistance
No association with blood glucose or HOMA-IR
Brosset and Ngueta
(2022)
Lit update
Glycemic control
No association with poor glycemic control
Ye etal. (2021)
Commenter (on
PFDA)
Metabolic
syndrome
No association with metabolic syndrome (OR
1.02, 95% CI 0.93,1.13) or blood glucose,
blood pressure, serum lipids, or waist
circumference
No. Existing and new studies are primarily
null, and the current draft judgment is
unlikely to change.
Schillemans et al.
(2022)
Lit update
Adiposity
No association with BMI z-score
No. Existing and new studies are primarily
null, and the current draft judgment is
unlikely to change.
Zeng et al. (2023)
Lit update
Adiposity
No association with BMI z-score trajectory
(Harris et al., 2017)
Commenter (on
PFDA)
Adiposity
Lower PFHxS levels in obese (-8.0% difference,
95% CI -26.6,15.2 for obese vs normal)
Ji et al. (2012)
Commenter (on
PFDA)
Adiposity
Higher PFHxS concentrations in overweight
participants, but no statistical analysis
Pirard et al. (2020)
Commenter (on
PFDA)
Adiposity
No association with BMI (quantitative results
not presented)
Liu et al. (2020b)
Commenter (on
PFDA)
Adiposity
No association with BMI
Endocrine
Jensen et al. (2022)
Lit update
Thyroid hormones
No association with free T4, positive but non-
monotonic and not statistically significant
association with TSH (beta 4.05, 95% CI -1.58,
10.00)
No. Existing and new studies are primarily
null, and the current draft judgment is
unlikely to change.
Derakhshan et al.
(2022)
Lit update
Thyroid hormones
Positive association with free T4 (beta 0.13,
95% CI -0.01, 0.28) but no association with
TSH or free T3
Li et al. (2023a)
Lit update
Thyroid hormones
No association with TSH or free T4
This document is a draft for review purposes only and does not constitute Agency policy.
B-24 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
Tillaut et al. (2022)
Lit update
Thyroid hormones
No association with free T4, free T3, or TSH
Jain and Ducatman
(2019)
Commenter (on
PFDA)
Thyroid hormones
Positive association with Total T3 in
participants at higher glomerular filtration
stages.
Dufour et al. (2020)
Commenter (on
PFDA)
Thyroid disease
Inverse association with hyperthyroidism (OR
0.14, 95% CI 0.03, 0.63)
(Christensen et al.,
2016)
Commenter (on
PFDA)
Thyroid disease
Inverse association with thyroid disease (OR
0.59, 95% CI 0.20, 1.06)
(Nilsson et al.,
2022b)
Lit update
Thyroid problems,
thyroid hormones
No association with thyroid problems (OR
0.94, 95% CI 0.73,1.21). Inverse but not
statistically significant association with T4 but
not T3 or TSH.
Other
H0isager et al.
(2022)
Lit update
Bone mineral
density
Inverse association with bone mineral content
and density (p>0.05), stronger in boys
No.
Available studies are inconsistent, and
evidence would likely be indeterminate.
Zhao et al. (2022)
Lit update
Bone mineral
density
Inverse association (p>0.05) with femur bone
mineral density in women without
menopause/hysterectomy
Colicino et al.
(2020)
Lit update
Bone mineral
density
No association with lumbar spine or femur
density
Xiong et al. (2022)
Lit update
Bone mineral
density
Positive association with femur density and
inverse association with lumbar spine density
in girls only
Fan et al. (2023)
Lit update
Bone mineral
density,
osteoporosis
Positive but not statistically significant
association with osteoporosis (OR 1.23, 95% CI
0.95,1.60), inverse association with bone
mineral density (beta -0.23, 95% CI -0.33, -
0.12)
This document is a draft for review purposes only and does not constitute Agency policy.
B-25 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
Shiue (2015d)
Commenter (on
PFDA)
Oral health
No association with teeth health, ache, tooth
loss
Liao et al. (2022a)
Lit update
Hematology
Positive but not statistically significant
association with gestational anemia in the 1st
and 3rd but not 2nd trimesters. No association
with hemoglobin concentration during
pregnancy
No.
Inconsistent results in new studies.
Evidence would likely be indeterminate
overall.
Cui et al. (2022)
Lit update
Hematology
Positive association with hematocrit (3.51%
change, 95% CI 1.82, 5.24) and hemoglobin
(3.14% change, 95% CI 1.33, 4.99) during
pregnancy
Liu et al. (2022)
Lit update
Hematology
No association with white blood cells and
lymphocytes
Shiue (2015a)
Commenter (on
PFDA)
Neurologic;
Remembering
condition
No association with difficulty remembering
(RR 0.45, 95% CI 0.25-0.81 for >3 times per
week)
No.
Lack of association in available studies and
would likely be indeterminate overall.
Shiue (2015b)
Commenter (on
PFDA)
Neurologic;
Depression
No association with adult depression
Shiue (2015c)
Commenter (on
PFDA)
Neurologic;
Hearing
disturbance
No association with trouble hearing
(Gavlord et al.,
2019)
Commenter (on
PFDA)
Pulmonary
function
No association with FEV or FVC (FEV1 beta -
0.01, 95% CI -0.10, 0.08, FVC beta 0.03, 95% CI
-0.08, 0.13)
No.
Lack of association in available studies and
would likely be indeterminate overall.
Shi et al. (2023)
Lit update
Pulmonary
function
No association with forced expiratory volume
or forced volume capacity
ADME/PBPK Studies
Chiu et al. (2022)
Lit update
One-compartment
PK model fit to
data from highly
GM (95% CI) for ti/2, Vd and CL are 8.30 (5.38-
13.5) yr, 0.29 (0.17-0.45) L/kg and 0.068
(0.033-0.107) mL/kg-d. The CL is higher than
Yes.
This document is a draft for review purposes only and does not constitute Agency policy.
B-26 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Reference
Source1
Health outcome
Results summary
EPA disposition on incorporation and
characterization of impact2
exposed
communities
(after
intervention)
our previous GM and health-protective lower
bound, but in the range of other studies.
Incorporated clearance value into
calculation of overall average clearance.
See Section 3.1 in main document
Jain and Ducatman
(2022)
Lit update
PFHxS serum
levels in US
females vs. males
as a function of
age (NHANES).
In males a slow, steady increase from age 12
to > 75, but in females the levels decline from
age 12 to 30, reaching ~ Yi the levels in males,
then begin to increase around age 45.
Yes.
Semi-quantitative support for impact of
menstrual fluid loss: data are consistent
with estimated clearance values. Including
*mean* non-menstrual clearance. See
Section 3.1 in main document
Oh etal. (2022a)
Lit update
Change in
maternal PFHxS
levels from
conception to 2
yrs post-partum
Mean PFHxS serum levels decline slightly
(0.6%) during pregnancy, decline 5.6%
(statistically significant) from 0-6 mos post-
partum, then increase 0.5% from 6-24 mos
post-partum.
Yes.
Maternal concentrations at or below
concentration at conception: should be
predicted with menstrual clearance
included. See Section 3.1 in main document
Li et al. (2022c)
Lit update
PFHxS half-life in
Swedish
population after
end of high
drinking water
exposure.
Mean (95% CI) ti/2 = 4.52 (4.14, 4.99) y.
Median (5th, 95th %tile) = 5.4 (2.34, 9.29) y.
This is a bit shorter than some other studies
but overlap to a fair extent. Results may be
less impacted by ongoing background
exposure than other data.
Incorporated clearance (from reported half-
life) into calculation of overall average
clearance. See Section 3.1 in main
document
Nilsson et al.
(2022a)
Lit update
PFHxS half-lives in
Australian fire-
fighters after
change in foam
formulation.
Mean (95% CI) ti/2 estimated to be 7.8 (7.3,
8.3) yrs. Rate of decline vs. initial serum level
also evaluated and appeared to be
independent. Mean ti/2 in high exposure
group was 7.7 yrs vs. 8.2 yrs in low exposure
group.
Yes.
Incorporated clearance (from reported half-life)
into calculation of overall average clearance. See
Section 3.1 in main document
aNo animal studies were identified in the April 2023 literature search.
This document is a draft for review purposes only and does not constitute Agency policy.
B-27 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
APPENDIX C. SUPPLEMENTAL APPROACHES AND
DATA ANALYSES
C.l. PFAS CO-EXPOSURE CONSIDERATIONS AND META-ANALYSIS OF
PFHXS EFFECTS ON BIRTH WEIGHT
As noted in the polyfluoroalkyl substances (PFAS) protocol, the potential for confounding
by co-occurring PFAS to bias effect estimates is a concern in epidemiological studies despite a lack
of scientific consensus on how best to address PFAS co-exposures (and other co-occurring
contaminants) especially those exposures that may derive from different sources. The potential for
confounding across PFAS is incorporated in individual study evaluations and assessed across
studies in evidence syntheses and in the characterizations of the strength of evidence. For other
covariates like glomerular filtration rate, in general, more confidence was placed in studies that
adjusted for pregnancy hemodynamics, or if they considered this potential source of bias by
sampling PFAS levels earlier in pregnancy. More details on the considerations of the potential effect
of PFAS co-exposures and pregnancy hemodynamics follow.
C.l.l. Confounding Directionality and PFAS Co-exposure Statistical Approaches
A source of uncertainty in the epidemiological database was the potential for confounding
by other PFAS (and other co-occurring contaminants). Co-occurring PFAS that are actual
confounders (i.e., associated with both the PFAS of interest and the outcome, but not an
intermediate in the causal pathway between the two) would be considered positive confounders if
their effect estimate with the endpoint of interest is in the same direction as the main primary PFAS
of interest. In this example, such PFAS are considered positive confounders if their effect estimate
with the endpoint of interest is in the same direction as the primary PFAS of interest. If positive
confounders are not accounted for in the epidemiological study design or analysis phase, the
anticipation is that any resultant bias would be away from the null.
Statistical approaches can help address the challenges of evaluating the effects of numerous
PFAS that may be present in the environment and estimated via different biomarkers and other
exposure measures. For example, multi-pollutant models (i.e., those that adjust for at least one co-
exposure) can theoretically provide an estimate of the independent association for specific
pollutants with the endpoint of interest by controlling for the independent effect of one or more co-
exposures, thereby removing the potential confounding bias (assuming proper functional forms of
and limited measurement error related to the confounding variables). However, under certain
circumstances, when highly correlated co-exposures are included in the same model, controlling for
one co-exposure can potentially amplify the potential confounding bias of another confounder
This document is a draft for review purposes only and does not constitute Agency policy.
C-l DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
rather than removing it (Weisskopf et al, 20181. This may be especially problematic when
confounding is present and simultaneously adjusted PFAS exposures come from different exposure
sources. Dimension-reducing statistical approaches (e.g., principal component analysis and
penalized modeling based on elastic net regression) are increasingly being used for screening large
groups of chemical classes and help to prioritize specific mixtures. However, as noted by Meng et al.
f20181. these approaches might be better suited as "prediction models to screen for a wide range of
chemicals from different sources, and the interpretation of results might become less
straightforward due to the necessary standardization of exposure values." These regression model
outputs also do not provide confidence intervals (CIs), which precludes evaluations of precision.
Given these interpretation difficulties and potential for co-exposure amplification bias, it remains
unclear whether certain mutually adjusted models give a more accurate representation of the
independent effect of specific pollutants for complex PFAS mixture scenarios.
In the main developmental epidemiological syntheses, the evaluation of between-study
heterogeneity is based on single pollutant models to increase comparability of available data. An
evaluation of single-pollutant (i.e., perfluorohexanesulfonic acid [PFHxS] alone) models and other
approaches are detailed below. The objective herein is to assess whether there is any direct
evidence for confounding in the studies comparing multi-pollutant (mutually adjusted for other
PFAS) and single-pollutant (i.e., PFHxS alone with other confounders adjusted for) model results
under the assumption that multi-pollutant models may provide a better reflection of the underlying
risk in the absence of any co-amplification bias. Additional objectives of this Appendix were to
consider the potential for confounding by examining the strength of associations between co-
occurring PFAS as well as those between each PFAS and the primary endpoint of interest (e.g., birth
weight-related measures).
C.1.2. PFAS Co-exposure Correlations with PFHxS
In general, the stronger the correlation or association observed between co-exposures, and
the larger the associations between the co-exposure and endpoints such as fetal growth restriction,
the more concern there would be for potential confounding. A preliminary analysis of 22 studies in
the inventory informs Table C-l, which illustrates the direction and magnitude of the correlations
between PFAS co-exposures in the PFHxS studies that examined these measures. While it shows
that some PFAS co-occur with PFHxS (as expected given some similar anticipated sources), it also
illustrates that the magnitude of these relationships can vary across studies. For example, with the
exception of two studies for which correlations ranged from 0.30 to 0.34, most studies showed
moderate or high correlations of PFHxS with PFOAand PFOS (range: 0.47-0.75). PFHxS was
consistently weakly or moderately correlated with both PFDA and PFNA in all of these studies
(range: 0.22-0.51). These data suggest that among the other PFAS that have been evaluated, PFOS,
and to a lesser extent, PFOA, might be the co-exposures of most concern in many study settings. The
stronger correlations between PFHxS and PFOS is unsurprising because, of these chemicals, PFOS
shares the greatest structural similarity with PFHxS. In addition to the impact of the structural
This document is a draft for review purposes only and does not constitute Agency policy.
C-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
similarity on the physico-chemical properties that determine disposition of the chemicals in the
environment, these chemicals may also occur together due to their co-production in the
manufacturing process fBoucher et al. 2019: 3M. 19991. Similarly, although co-occurrence data
may be unavailable, other sulfonated PFAS that are unmeasured or less well characterized in
biomarkers may likewise represent potential co-exposures of concern.
Table C-l. PFAS correlation coefficients in nine mutually adjusted PFAS
studies
Reference
Study
confidence
Correlations with PFHxS
PFOS
PFOA
PFNA
PFDA
Ashlev-Martin et al. (2017)
High
0.55
0.47
Luo et al. (2021)
High
0.01
0.02
-0.04
-0.03
Manzano-Salgado et al. (2017)
High
0.56
0.40
0.36
Shoaff etal. (2018)"
High
~0.6
~0.4
~0.3
~0.2
Starling et al. (2017)
High
0.65
0.61
0.45
0.27
Hamm et al. (2010)
Medium
0.54
0.55
Lenters et al. (2016)
Medium
0.34
0.34
0.22
0.36
Meng et al. (2018)
Medium
0.30
0.33
0.28
Callan et al. (2016)
Low
0.75
0.71
0.51
0.44
aShoaff et al. (2018) Pearson correlation coefficients ranged from 0.32 (PFNA and PFHxS) to 0.60 (PFOA and PFOS).
The estimated correlation coefficients above are based on their related publication (Woods et al., 2017); thus, this
may slightly over-estimate the PFDA and PFNA correlation given the initial range provided by Shoaff.
C.1.3. PFHxS and PFAS Co-exposure Study Results
Nine of the PFHxS studies that were evaluated examined PFAS co-exposures including one
low confidence (Callan et al. 20161. three medium confidence (Mengetal, 2018: Lenters et al..
2016: Hainin et al.. 20101. and five high confidence studies (Luo et al.. 2021: Shoaff et al.. 2018:
Ashley-Martin et al. 2017: Manzano-Salgado etal, 2017: Starling et al. 20171. The studies by
Hamm et al. (20101 and Luo etal. (20211 did not provide both single-pollutant and multi-pollutant
model results for the continuous exposures of interest; this lack of a direct comparison precluded
further evaluations of the potential confounding by co-occurring PFAS in that study. The results for
the seven other studies based on continuous PFHxS unit changes are compared and summarized
below to assess whether any patterns of evidence for larger associations with other PFAS occurred
and/or any direct evidence for confounding in the mutually adjusted PFAS studies examining mean
birth weight given the primary focus on this endpoint (Table C-2). Two of these studies (Starling et
This document is a draft for review purposes only and does not constitute Agency policy.
C-3 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
al, 2017: Lenters etal, 20161 examined PFAS co-exposures through elastic net regression, while
the remaining studies performed multi-pollutant modeling using ordinary least squares regression.
As shown in Table C-2, Callan etal. (2016) found large mean birth weight deficits ((3 = -72 g;
95%CI: -194, 50) in their single-pollutant model of PFHxS, which became stronger (39% increase)
following adjustment only for perfluoroundecanoic acid ((3 = -100 g; 95%CI: -221, -221 Meng et al.
(2018) reported nonsignificant deficits ((3 = -12.4 g; 95%CI: -46.2, 21.4) in a single-pollutant
PFHxS model, which became stronger (30% increase) upon adjustment of PFOS, PFOA, and PFNA in
their multi-pollutant model ((3 = -16.4 g; 95%CI: -61.0, 28.1). Starling etal. (2017) reported largely
null findings for PFHxS based on either single-pollutant or multi-pollutant models and this PFAS
was not selected based on their elastic net regression. Lenters etal. (2016) reported null results for
PFHxS in both their single-pollutant model and their elastic net regression of mutually adjusted
PFAS with only PFOA retained in the latter model. Shoaffetal. (2018) reported that their
marginally significant single-pollutant PFHxS birth weight z-score was attenuated and became
more imprecise upon multi-pollutant adjustment Meng etal. (2018) reported largely null results
for PFHxS in single-pollutant models, whereas a nonsignificant increase in mean birth weight was
seen upon multi-pollutant adjustment for PFOS, PFOA, PFNA, perfluoroheptane sulfonic acid
(PFHpS), and PFDA.
Among these limited studies, there were no consistent patterns detected in the birth weight
data for PFHxS following mutual adjustment for other correlated PFAS. For example, two of the five
studies that reported birth weight deficits in single-pollutant models were strengthened in similar
fashion (30-39%) following statistical adjustment for other PFAS. In contrast, associations in two
other studies were attenuated and another study went from an overall mean decreased birthweight
to increased birthweight following adjustment. The other two studies were null in both single and
multi-pollutant models. There were also no clear patterns of larger associations for other PFAS
examined in these studies (data not shown). Thus, strong, and consistent evidence of confounding
by other PFAS is not demonstrated in these studies, which is also supported by the lack of
differences in the PFHxS-stratified pooled estimates evaluated in the meta-analysis; if confounding
were present, there would likely be more variability detected across studies and strata given the
variable correlations noted above. While there is still uncertainty due to reported correlations
between PFHxS and some PFAS (e.g., PFOS), based on the available evidence, it seems unlikely that
the consistency of birth weight deficits demonstrated from (categorical and continuous results) in
the full set of 27 PFHxS studies examined here can be fully attributed to confounding by PFAS co-
exposures.
This document is a draft for review purposes only and does not constitute Agency policy.
C-4 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table C-2. Impact of co-exposure exposure adjustment on birth weight results (change in mean birth weight or
birth weight Z-scores) per unit change (ng/mL) in PFHxS levels
Birth
Single-PFAS model
Multi-PFAS
Elastic net
Effect of adjustment
Study
weight
results with
results with
regression
on PFHxS birth
PFAS
Reference
confidence
measure
95%Cls
95%Cls
results
weight results
adjustments
Ashley-Martin et al.
High
Birth Weight
0.08 (-0.12, 0.20)a
0.04 (-0.13, 0.21)a
N/A
Slightly Attenuated
PFOS, PFOA
(2017)
z-scores
Shoaffet al. (2018)
High
Birth Weight
-0.08 (-0.18, 0.01)a
-0.05 (-0.17, 0.06)a
N/A
Slightly Attenuated
PFOS, PFOA,
z-scoresb
PFNA
Manzano-Salgado et al.
High
Mean Birth
-12.4 (-46.2, 21.4)
-16.4 (-61.0, 28.1)
N/A
Slightly Strengthened
PFOS, PFOA,
(2017)
Weight
PFNA
Starling et al. (2017)
High
Mean Birth
-13.5 (-50.7, 23.7)
11.5 (-38.9, 61.9)
N/S
Changed direction from
PFOS, PFOA,
Weight
Negative to Positive
PFNA, PFDeA
Lenters et al. (2016)
Medium
Mean Birth
Weight
-19.1 (-40.7, 2.3)
N/A
N/S
Attenuated
PFOS, PFOA,
PFNA, PFUnDA,
PFDoDA, PFDA
Meng et al. (2018)
Medium
Mean Birth
1.7 (-40.8, 44.3)
25.0 (-10.1, 60.1)
N/A
Changed from Null to
PFOS, PFOA,
Weight
Positive
PFNA, PFDA,
PFHpS
Callan et al. (2016)
Low
Mean Birth
-72 (-194, 50)
-100 (-221, 22)
N/A
Strengthened
PFUnDA
Weight
Abbreviations: N/A: not available; N/S: PFAS not selected in final elastic net regression model. PFUnDA: perfluoroundecanoic acid; PFDeA: perfluorodecanoic
acid; PFDoDA: perfluorododecanoic acid; PFHpS: perfluoroheptanesulfonic acid.
aThe Ashley-Martin etal. (2017} study BWT z-score results are per log-10 unit increases and the Shoaff etal. (2018) study BWT z-score results are per a log-2
increase (i.e., a doubling of PFHxS exposure); all other studies presented here were for each In-unit increase based on original results from publication or EPA
re-expressions.
bThe mean birth weight result for the single-pollutant model in Shoaff et al. (2018) was -13.4 grams (95%CI: -35.9, 9.1) per each 1 ng/mL increase (JR Shoaff
Personal Communication, 10-19-18 (Shoaff, 2018)).
This document is a draft for review purposes only and does not constitute Agency policy.
C-5 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
C.1.4. Pregnancy Hemodynamics Background
Pregnancy-related hemodynamic changes that occur during pregnancy (e.g.,.increased
blood plasma volume due to decreased mean arterial pressure, increased cardiac output, and
systemic vasodilation fSagiv etal. 2018: Sanghavi and Rutherford. 2014: Chapman et al. 199811 are
complex and can lead to challenges in data interpretability when timing of PFAS sampling differs
within and across studies. These changes could lead to lower PFAS levels in plasma, due to dilution
and increased renal filtration. A decrease in PFAS levels has been noted in serial measurements for
most of some PFAS during pregnancy, namely PFOA, PFOS, and PFNA (Chen etal. 2021: Glvnn etal.
20121. These hemodynamic changes have been proposed as a potential source of bias for
associations between different PFAS and neonatal and early childhood growth measures, which is
suggested by the association between glomerular filtration rate (GFR), a marker of renal function
and, indirectly, of plasma volume expansion, and fetal growth independent of gestational age and
other maternal covariates (Morken et al. 2014: Gibson. 19731. Because PFNA concentration in
serum is expected to decrease during pregnancy due to plasma volume expansion, increased renal
excretion, and transplacental transfer, time windows earlier in pregnancy prior to this decrease
may reflect the largest insult to a developing fetus. Potential confounding is one possible
explanation for the effects of pregnancy hemodynamics, but Steenland et al. (2018) also proposed
that GFR may lead to reverse causality if increased fetal growth leads to increased maternal blood
expansion and glomerular filtration rate. These potential sources of bias are anticipated to be of
greater concern when maternal serum PFAS samples are collected later in pregnancy. Therefore, as
part of the study quality evaluations, more confidence was placed in studies that adjusted for
pregnancy hemodynamics or if they considered this potential source of bias by sampling PFAS
levels earlier in pregnancy.
Only three of the 21 PFHxS studies examined in the Developmental Effects section collected
were able to analyze maternal hemodynamic data such as GFR and albumin (i.e., a marker of plasma
volume expansion). None of these studies showed evidence of substantial confounding of the
associations between PFAS and fetal growth following statistical adjustment for GFR (Manzano-
Salgado et al. 2019: Gvllenhammar et al. 2018b). or for GFR and albumin fSagiv etal, 2018).
Although early pregnancy measures may limit this potential source of bias, the first trimester
sampling of plasma albumin and GFR in two of these studies fManzano-Salgado et al. 2019: Sagiv et
al. 20181 may not be best-suited to examine potential confounding if the sample timing did not
fully reflect pregnancy-related hemodynamic changes. However, the study by fGvllenhammar et al.
2018b) with postpartum samples along with another measurement of PFOA and PFOS based on
mid-pregnancy samples (Whitworth et al. 2012) have also shown no evidence of confounding by
albumin or GFR. These data run counter to meta-analyses for both PFOA (Steenland et al. 2018)
and PFOS (Dzierlenga et al. 2020). which have detected larger birth weight deficits for later
trimester sampling (e.g., beyond trimester one) compared with early periods. To examine whether
the overall pooled estimates for mean birth weight reported in the synthesis varied across different
This document is a draft for review purposes only and does not constitute Agency policy.
C-6 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
sampling periods and overall study confidence levels, EPA examined stratum-specific pooled
estimates for PFHxS.
C.1.5. Meta-Analysis Methods
Study Inclusion
Thirty-eight developmental epidemiological publications of PFHxS that were identified met
our aforementioned inclusion criteria. Some birth cohorts reported BWT analyses for different
subsets of their populations in different publications. To avoid double-counting of overlapping
study populations, we restricted the meta-analysis to the largest study population sample where
multiple publications existed. Specifically, the Woods et a . study overlapped with the Shoaff
etal. (2018) study from the Health Outcomes and Measures of the Environment cohort, and the
Bierregaard-Olesen et al , study overlapped with fBach et al. 2016a) from the Aarhus Birth
Cohort. Following exclusion of (Bierregaard-Olesen et al. 2019) and (Woods et al. i ), there
were 36 non-overlapping studies with developmental endpoints of mean BWT changes in relation
to PFHxS exposure which advanced to study evaluation. As shown in Figure 1, five studies
ffMaekawa etal. 2017: Alkhalawi etal. 2016: Lee and Viberg. 2013: Monrov et al. 2008)) were
classified as uninformative largely due to critical study deficiencies in at least one risk of bias
domain (e.g., confounding and participant selection) or multiple domain deficiencies. Among the
remaining 31 studies, study confidence ratings included 7 low confidence studies fMarks et al.
2019b: Workman et al. 2019: Xu etal. 2019: Cao et al. 2018: Gao et al. 2018: Shi etal. 2017:
Callan et al. 2016), 11 medium confidence studies (Chang et al. 2022: Chen etal. 2020:
Hiermitslev et al. 2020: Kashino et al. 2020: Gvllenhammar et al. 2018a: Meng et al. 2018: Li et al.
2017a: Kwon etal. 2016: Lenters et al. 2016: Maisonet etal. ; im et al. 2010). and 13 high
confidence studies fYao etal. 2021: Eick et al. 2020: Wikstrom et al. 2020: Buck et al. 2018: Sagiv
etal. 2018: Shoaff et al. 2018: Ashley-Martin etal. 2017: Lind etal. 2017: Manzano-Salgado et al.
2017: Starling et al. 2017: Valvi etal. 2017: Bach et al. 2016b).
Of the 31 informative epidemiological studies with mean BWT data, th ree studies (Eick et
al. 2020: Gao et al. 2019: Cao et al. 2018) reported categorical results only. Our primary analysis
was restricted to BWT studies that were most comparable, i.e., those based on continuous PFHxS
exposures. Among the 28 studies with results based on continuous data, 24 provided effect
estimates in the overall population (i.e., male and female combined). Data were pooled for the four
studies that only reported sex-specific findings fMarks etal. 2019b: Ashley-Martin et a ; id
etal. 2017: Maisonet etal. 2012) within each study using inverse-variance weighting to provide an
effect estimate in each study's overall population. This included male and female-specific results in
studies by (Lind et al., 2017) and (Ashley-Martin etal., 2017). The study by Maisonet et al. (2012)
and Marks et al. (2019b) reported sex-specific estimates for girls and boys in different publications
from the same Avon Longitudinal Study of Parents and Children (ALSPAC) study population. These
two studies were also pooled to obtain an effect estimate in the overall population and included in
This document is a draft for review purposes only and does not constitute Agency policy.
C-7 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Supplemental Information—Perfluorohexanesulfonic Acid and Related Salts
the meta-analysis labeled henceforth, herein, as (Maisonet et al., 20121. Including these mean
birthweight estimates pooled across sexes from these three birth cohorts fAshlev-Martin et al..
2017: Lind et al.. 2017: Maisonetet al. 20121 resulted in the inclusion of 27 nonoverlapping
informative PFHxS studies (from 28 publications) with continuous exposure expressions for the
meta-analysis.
Figure C-l. Twenty-seven informative nonoverlapping perinatal studies of
birth weight measures and continuous PFHxS exposure results included in
meta-analysis.
Data Preprocessing
EPA converted the exposure-response functions quantifying the results for the 27 available
studies (based on data from 28 publications) using different units into two common exposure
metrics of natural (i.e., per ng/mL) or natural log units (i.e., per ln(ng/mL)). For example, to
standardize the units and reduce between-study heterogeneity due to the choice of unit, different
units such as log2, logm, and per SD- or IQR-unit changes were converted into a common
logarithmic function (natural log) as described below. Four of the 27 included studies were based
on natural scale PFHxS data fSagiv etal.. 2018: Shoaff. 2018: Bach et al.. 2016a: Maisonetet al..
This document is a draft for review purposes only and does not constitute Agency policy,
C-8 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
20121. and EPA used those results to estimate what the results would have been had they been
based on a one natural log (In) unit transformation. This approach was developed by Dzierlenga et
al. (20201 and involved plotting the reported linear function for the main effect using 25th-75th
percentiles at 10 percentile intervals of the exposure distribution in each study and then fitting a
natural logarithmic function to those points. This process was repeated using the reported upper
and lower CIs to estimate the bounds of the natural log function and thus the estimated standard
error of the natural log function (i.e., standard error = (upper confidence limit - lower confidence
limit) / 3.92 (Wiggins et al. 2022)).
The meta-analysis was carried out on the natural log scale because the majority (23 out of
27) of the studies reported results on the log scale. Transformations to the log-normal scale are
typically employed in epidemiological studies to satisfy regression assumptions. However, the re-
scaling methods used by Dzierlenga et al. (20201 and Steenland et al. (2018) can also be used to
express ln-transformed data on the natural scale, which may have useful implications for dose
response. A sensitivity analysis was conducted to test the robustness of our analysis to selection of
natural or natural log scale (see Sensitivity Analysis section below).
Statistical Analysis
The meta-analysis of the 27 developmental PFHxS studies reporting mean birth weight
differences was carried out using a random effects model, which follows the assumption that each
study produced an estimate of a study-specific true effect that varies across studies (Borenstein et
al. 2009). Inverse-variance weighting was employed to minimize the influence of both sampling
variance and between-study variance on the pooled effect estimate. The amount of variation due to
study heterogeneity was captured by two metrics: the I2 statistic and Cochran's Q test The I2
statistic represents the percent of variation in the pooled estimate due to between-study
heterogeneity. The range of values shown in the Cochran's I2 guidelines (Biggins et al. 20221
informed EPA's consideration of I2 statistics <40% to represent "low" potential heterogeneity, with
values from 40-69% being "moderate," and values >70% to represent "high" heterogeneity.
Cochran's Q test evaluates whether the dispersion of study-specific estimates about the pooled
effect estimate is statistically significant via a p-value (pq), based on significance level (a) of 0.05.
Both metrics may suffer from low statistical power when few studies are available, potentially
complicating interpretation of the examinations of heterogeneity. Thus, consideration of both
measures in conjunction is recommended to identify situations for which heterogeneity may be
present (Huedo-Medina et al. 20061. While the number of studies for the overall analysis may be
large enough (n = 27) to not be subject to these concerns, some uncertainty exists for the stratified
analyses with considerably fewer studies per strata.
EPA conducted stratified analyses to evaluate whether the summary effect estimate varied
by the study confidence rating or by the timing of maternal serum sampling. As detailed in Section
3.2.3, study confidence designations included four low confidence studies (Workman et al. 2019: Xu
etal. 2019: Shi et al. 2017: Callan etal. 20161. eleven medium confidence studies (Chang et al.
This document is a draft for review purposes only and does not constitute Agency policy.
C-9 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
2022: Chen etal. ; rmitslev et al, 2020: Kashino et at, 2020: Gvllenhammar et at, 2018b:
Meng et at, 2018: Li etal., 2017c: Kwon etal., 2016: Lenters et at, 2016: MaisonetetaL 2012:
Hamm et at. 20101. and twelve high confidence studies fLuo etal.. 2021: Yao etal.. 2021: Wikstrom
etal.. 2020: Buck Louis etal.. 2018: Sagiv et at. 2018: Shoaff et at. 2018: Ashley-Martin etal.. 2017:
Lind et at. 2017: Manzano-Salgado et at. 2017: Starling etal.. 2017: Valvi et at. 2017: Bach etal..
2016a").
Sample timing strata were defined according to two strategies based on reported
gestational age (weeks) at time of biomarker collection. Strategy 1 was a three-strata approach
with subgroups, early (n = 12), late (n = 10) and post (n = 5) pregnancy. Strategy 2 was a two-strata
approach, using the same definition of early pregnancy as in Strategy 1, but combining late- and
post-pregnancy into a single stratum, late + post (n = 15). Early pregnancy included studies
reporting samples from preconception (0 days), the first trimester (0 days to 13 weeks and 6 days)
or a mixture of the first and second trimesters (0 days to 27 weeks and 6 days); late pregnancy
studies sampled in the second trimester (14 weeks and 0 days to 27 weeks and 6 days), a mixture of
the second and third trimester (14 weeks and 0 days to birth), or the third trimester only (28 weeks
and 0 days to birth); post-pregnancy studies sampled at or after birth (ACOG. 2020). Studies were
assigned to sample timing strata based on reported trimesters of sampling as well as sampling
ranges and interquartile ranges or measures of centrality when measures of spread were
unavailable (see Table C-3 below for details on sample timing distributions and strata
assignments).
The two-strata sample timing approach was also used by previous PFAS meta-analyses
(Dzierlenga et at, 2020: Steenland et at, 2018). However, as noted in Wright etal. (20231. there
may be value in differentiating late maternal samples from post-partum measures and further
refining what constitutes early sampling when a sufficient number of studies allow other
alternatives. Thus, EPA apportioned studies with late pregnancy samples from those with post-
pregnancy samples to better understand differences in sampling matrices, i.e., maternal serum
sampled during pregnancy versus umbilical cord samples or postpartum maternal serum samples
(i.e., termed post-pregnancy here). Furthermore, the use of more subgroups provides more detail
on the gradient of changes that sample timing may be associated with. A sensitivity analysis was
employed to assess the robustness of the meta-analysis results to using three strata instead of two.
Strata-specific estimates that allowed for heterogeneity were calculated using a random
effects model and a subsequent fixed effects model was used to test for statistically significant
differences across the subgroups fBorenstein et at. 20091. A p-value less than 0.05 from this
hypothesis test was indicative of no statistically significant differences between any of the strata.
Strata-specific statistical tests conducted on subgroups with lower sample sizes are subject to lower
power, susceptible to higher uncertainty and therefore should be interpreted with caution. For full
details on the computations involved in both the stratified and overall meta-analyses, please refer
to the R code developed by EPA fLarsen. 20221.
This document is a draft for review purposes only and does not constitute Agency policy.
C-10 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
1 All statistical analyses are carried out using the open-source platform, R (Version 4.0.3), and
2 all meta-analytic techniques are carried out using the meta-analysis R package, metafor
3 fViechtbauer. 20101.
This document is a draft for review purposes only and does not constitute Agency policy.
C-ll DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table C-3 Details on study sample timings and strata assignments
Study
Sampling
distribution - estimates and measures of
centrality (and spread) in gestational weeks
Time period
Sample
timing
start3
Notes
Ashley-Martin et al. (2017)
N/R (N/R)
Trimester 1
Early
Sampling time of 9.9
wks was estimated
from the trimester 1
midpoint with a
range of 6 wks to 13
wks and 6 d.
Bach et al. (2016a)
12 (9, 20)
Mode (Min, Max)
Trimesters 1 and 2
Early
Buck Louis et al. (2018)
10,13.9
(Min, Max)
Trimester 1
Early
Value of 11.9 wks
was estimated as
midpoint of the
range (10 to 13.9
wks).
Chang et al. (2022)
11.4 (9.6, 12.6)
Median (25%, 75%)
Trimesters 1, 2
Early
Chen et al. (2021)
16.3 (13.85, 20.43)
Median (Min, Max)
Trimesters 1 and 2
Early
The authors
provided additional
data, which showed
their serial measures
included overlapping
trimesters, e.g., their
trimester 1 results
encompassed
trimesters 1 and 2
samples (see Zhang
(2022)).
This document is a draft for review purposes only and does not constitute Agency policy.
C-12 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Sampling
Sample
distribution - estimates and measures of
timing
Study
centrality (and spread) in gestational weeks
Time period
start3
Notes
15.39 (7, 40)
Trimesters 1, 2,
Early
This study was
(Min, Max)
and 3
assigned to the early
strata because
sampling
predominantly
occurred earlier in
pregnancy: study
authors report that
the mean
gestational week of
Hiermitslev et al. (2020)
sampling in 2010-
2011 was week 26.2,
and in 2013-2015 all
samples were
collected before the
end of week 13. 38%
of samples were
taken in 2010-2011;
62% was collected in
2013-2015
(Bonefeld-
J0rgensen, 2022).
Unci et al. (2017)
10 (5, 12)
Median (Min, Max)
Trimester 1
Early
Maisonet et al. (2012);
10, 228 (25%, 75%);
Trimester 1, 2, and
Early
Marks et al. (2019a)a
12, 33 (25%, 75%)
3
12.3 (5.6)
Trimester 1, 2, and
Early
Sampling is reported
Mean (SD)
3
to be all in the first
Manzano-Salgado et al. (2017)
trimester (Manzano-
Salgado et al., 2017)
supporting
information clarifies
This document is a draft for review purposes only and does not constitute Agency policy.
C-13 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Study
Sampling
distribution - estimates and measures of
centrality (and spread) in gestational weeks
Time period
Sample
timing
start3
Notes
that some sampling
outside of the 1st
trimester also
occurred (Wright
and Larsen, 2022).
However, first
trimester sampling
was predominant, so
this study is
designated as
conducting "early"
sample timing.
Meng et al. (2018)
8 (4, 14)
Mean
Trimesters 1 and 2
Early
The mean is
reported in related
publication by Liew
et al. (2020).
Sagiw et al. (2018)
9 (5,19)
Median (Min, Max)
Trimesters 1 and 2
Early
Wikstrom et al. (2020)
10
Median
Trimesters 1 and 2
Early
Callan et al. (2016)
37.7 (33, 40)
(Min, Max)
Trimester 3
Late
Samples were taken
2 wks before due
date, which ranged
from 35 to 42 wks;
estimate measure of
centrality used here
of 37.7 wks.
Ha mm et al. (2010)
15.5 (15, 16)
(Min, Max)
Trimester 2
Late
Measure of
centrality of 15.5
wks estimated from
This document is a draft for review purposes only and does not constitute Agency policy.
C-14 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Study
Sampling
distribution - estimates and measures of
centrality (and spread) in gestational weeks
Time period
Sample
timing
start3
Notes
the midpoint of the
reported range.
Kashino et al. (2020)
29
Median
Trimester 3
Late
Lenters et al. (2016)
25.2
Weighted mean of medians
Trimesters 2 and 3
Late
Study authors
reported country-
specific medians: 33
wks (Poland, 18%),
25 weeks
(Greenland, 32%),
23 wks (Ukraine,
49%).
Luo et al. (2021)
39.3
Mean
Trimester 3
Late
Shoaff et al. (2018)
18.1
Weighted average
Trimesters 2 and 3,
and at delivery
Late
Study was assigned
to the late strata
instead of post
because only 5% of
samples taken at
delivery, and
sensitivity analysis
conducted by study
authors found
results robust to
Trimester 2 only.
The weighted
average was derived
the following data
provided in the
manuscript: "16-
week serum samples
This document is a draft for review purposes only and does not constitute Agency policy.
C-15 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Study
Sampling
distribution - estimates and measures of
centrality (and spread) in gestational weeks
Time period
Sample
timing
start3
Notes
if available (86%),
the 26-wk sample if
the 16-wk sample
was not available
(9%), and if neither
of those was
available, then the
samples from
delivery were used
(5%)."
Starling et al. (2017)
27 (20, 34)
Median (Min, Max)
Trimesters 2 and 3
Late
Valwi et al. {2017}
34
Exact
Trimester 3
Late
Workman et al. (2019)
28.6 (14.3, 39.6)
Median (Min, Max)
Trimesters 2, 3
Late
Data provided by
authors: Mean =
27.7 wks.
Median = 28.6 wks;
Range = 14.3-39.6
wks.
Yao et al. (2021)
39.4
Mean
Trimester 3
Late
GvIIenhammar et al. (2018b)
43 (37.9, 46.1)
Mean (Range)
Post-Birth
Post
Samples were taken
3 wks after delivery;
mean (range)
delivery date = 40
wks (34.9, 43.1).
Kwon et al. (2016)
40
Exact
At Delivery
Post
Li et al. (2017c)
39
At Delivery
Post
This document is a draft for review purposes only and does not constitute Agency policy.
C-16 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Study
Sampling
distribution - estimates and measures of
centrality (and spread) in gestational weeks
Time period
Sample
timing
start3
Notes
Mean
Shi etal. {2017}
39.8(4.2)
Mean (SD)
At Delivery
Post
Xu et al. {2019}
39.4(1.4)
Mean (SD)
At delivery
Post
Abbreviations: N/R: not reported; 25th, 75th%; 25th percentile, 75th percentile of exposure distribution; SD: standard deviation
"Maisonet et al, (2012) was rated as Medium confidence, while Marks et al, (2019b) was Low. This combined population was evaluated as Medium for stratified
analyses.
This document is a draft for review purposes only and does not constitute Agency policy.
C-17 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
C.1.6. Meta-Analysis Results
As shown in Figure C-2, the overall pooled birth weight effect estimate for the 27 studies
based on the random effects model was -7.7 g (95%CI: -14.8, -0.5) per In ng/mL increase in PFHxS
exposure. The tests for heterogeneity showed that between-study variability was negligible (I2 =
0%, pq = 0.84). The meta-analysis results stratified by study confidence are displayed in Table C-4.
The 12 high confidence studies yielded a smaller pooled effect estimate of decreased birthweight ((3
= -6.8 g; 95%CI: -16.3, 2.8) than the 11 medium ((3 = -9.6 g; 95%CI: -20.8,1.6) confidence studies;
however, the differences between strata were not statistically significant (p = 0.85). There was
negligible between-study heterogeneity for the high (I2 = 0%, pq = 0.94) and medium confidence
studies, while the low confidence subgroup exhibited "low" heterogeneity (I2 = 20.1%, pq = 0.30).
Given the small sample size of the strata (n = 4) and larger and more divergent results, the low
confidence combined estimate and heterogeneity statistics are subject to relatively more
uncertainty and should be interpreted with caution.
This document is a draft for review purposes only and does not constitute Agency policy.
C-18 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information—Perfluorohexanesulfonic Acid and Related Salts
Author(s) and Year
Timing
N
High Confidence
Buck Louis, 2018
1st
2106
Ashley-Martin, 2017
1st
1509
Lind, 2017
1st
636
Sagiv, 2018
1 st-2nd
1645
Wikstrom, 2020
1st-2nd
1533
Bach, 2016
1 st-2nd
1507
Manzano-Salgado, 2017
1 st-3rd
1202
Starling, 2017
2nd-3rd
598
Shoaff, 2018
2nd-3rd
299
Valvi, 2017
3rd
604
Yao, 2021
3rd
369
Luo, 2021
3rd
224
RE Model for High Confidence (Q
= 4.87, df =
11, p = 0.
Medium Confidence
Meng, 2018
1 st-2nd
2120
Chang, 2022
1 st-2nd
370
Chen, 2021
1 st-2nd
214
Maisonet, 2012
1st-3rd
895
Hjermitslev, 2020
1 st-3rd
266
Hamm, 2010
2nd
252
Lenters, 2016
2nd-3rd
1321
Kashino, 2020
3rd
1591
Li, 2017
Birth
321
Kwon, 2016
Birth
268
Gyllenhammar, 2018
Post-Birth
587
Estimate [95% CI]
; lz = 0.0%, Tz = 0.00)
RE Model for Medium Confidence (Q = 10.02, df = 10, p = 0.44; r = 0.0%. t
RE Model for High+Medium Confidence (Q = 15.03, df = 22, p = 0.86; I2 = 0.C
0.00)
6, t2 = 0.00)
Low Confidence
Workman, 2019
Callan, 2016
Shi, 2017
Xu, 2019
2nd-3rd
3rd
Birth
Birth
170
98
RE Model for Low Confidence (Q = 3.66, df = 3, p = 0.30; I2 = 20.1%, t" = 554.03)
T2 = e
-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]
-6.8 [-16.3, 2.8]
4.5 [-36.0, 44.9]
-20.2 [ -84.4, 44.0]
27.6 [-64.7, 119.9]
-11.2 [-28.5, 6.2]
-93.0 [-230.0, 44.0]
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]
-9.6 [-20.8, 1.6]
-8.0 [-15.2, -0.7]
-6.6 [-66.9, 53.7]
-72.0 [-194.0, 50.0]
47.3 [-23.4, 117.9]
-75.5 [-272.7, 121.7]
-1.5 [-51.6, 48.7]
RE Model for All Studies (Q = 18.88, df = 26, p = 0.84; I2 = 0.0%, t2 = 0.00)
Test for Subgroup Differences: QM = 0.34, df = 2, p = 0.85
-7.7 [-14.8, -0.5]
r
T
T
-100 0
Estimate (95% CI)
~r~
100
Figure C-2. Forest plot of 27 studies included in the meta-analysis on PFHxS
exposures and changes in birth weight.
This document is a draft for review purposes only and does not constitute Agency policy,
C-19 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table C-4. Meta-analysis of PFHxS on birth weight changes (in g per
ln(ng/mL)) stratified by study confidence
Set of studies
n
P
95% confidence
interval
I2 (%)
Pq
All Studies
27
-7.7
-14.8, -0.5
0.00
0.84
High Confidence
12
-6.8
-16.3, 2.8
0.00
0.94
Medium
Confidence
11
-9.6
-20.8, 1.6
0.00
0.44
Low Confidence
4
-1.5
-51.6, 48.7
20.1
0.30
High + Medium
Confidence
23
-8.0
-15.2,-0.7
0.00
0.86
Symbols and abbreviations: n = sample size; (5 = combined estimate of change in birth weight (g) per In (ng/mL)
PFHxS exposure; I2 = % variation in the pooled effect due to study heterogeneity; pq = p-value for Cochran's Q test
for heterogeneity.
The meta-analysis results stratified by sample timing are displayed in Table C-5. While
there are no statistically significant differences between timing subgroups in either the two-strata
or the three-strata approach (p = 0.88 and p = 0.54, respectively), there is some evidence that
estimated birth weight deficits are greater in later sampling. For example, a fourfold difference in
the three-strata approach was seen between the 12 early ((3 = -7.3 g; 95%CI: -16.0,1.4) studies and
5 studies based on post-pregnancy samples ((3= -28.3 g; 95%CI: -69.3,12.7). Differences between
early and later sampling are slightly more pronounced in the three-strata approach than in the two-
strata approach for which the pooled estimate for the late and post-pregnancy samples combined
was -8.5 g (95%CI: -21.0, 4.1). The estimate for the late pregnancy strata based on 10 studies was
null (P= -3.9 g; 95%CI: -17.7, 9.9) but was more comparable in magnitude to the early sampled
strata data ((3= -7.3 g per each PFHxS ln-unit increase) versus the post-partum studies.
Except for the post-pregnancy strata, Heterogeneity observed in each subgroup was
negligible for all strata expect the post-pregnancy stratum which had an estimated "low" percent of
variation due to heterogeneity, and non-significant Cochran's Q test (I2 = 20.1%, pq = 0.30).
However, the post-pregnancy strata have a sample size of less than ten studies, so results from
these heterogeneity tests are expected to be more uncertain.
Table C-5. Meta-analysis of PFHxS on birth weight (in g per ln(ng/mL))
stratified by sample timing
Set of studies
n
P
95% Confidence
interval
l2(%)
Pq
All Studies
27
-7.7
-14.8, -0.5
0.00
0.84
This document is a draft for review purposes only and does not constitute Agency policy.
C-20 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Set of studies
n
P
95% Confidence
interval
l2(%)
PQ
Early Pregnancy
12
-7.3
-16.0, 1.4
0.00
0.91
Late Pregnancy
10
-3.9
-17.7, 9.9
0.00
0.85
Post-Pregnancy
5
-28.3
-69.3, 12.7
39.8
0.19
Late + Post Pregnancy
15
-8.5
-21.0, 4.1
0.00
0.49
Symbols and abbreviations: n = sample size; p = combined estimate of change in birth weight in g per ln(ng/mL)
PFHxS exposure; l2= % variation in the pooled effect due to study heterogeneity; pq= p-value for the Cochran's Q
test for heterogeneity.
C.1.7. Sensitivity Analysis Results
1 The sensitivity of the meta-analysis results to re-expression was tested by comparing
2 results based on effect estimates re-expressed to the natural log scale to those converted to the
3 natural scale. Table C-6 illustrates that the pooled effect estimates of mean birth weight deficits in
4 our primary analysis are smaller and closer to the null when based on the natural scale but that the
5 patterns in magnitude seen across study confidence and sample timing strata are relatively
6 comparable.
Table C-6. Sensitivity of natural log scale or natural scale re-expression for the
overall and stratified meta-analyses of birth weight (in g per ln(ng/mL)
Set of studies
n
P (95%CI)
P (95%CI)
All Studies
27
-7.7 (-14.8, -0.5)
-2.9 (-6.4, 0.7)
Study Confidence Strata
High
12
-6.8 (-16.3, 2.8)
-1.0 (-5.3, 3.2)
Medium
11
-9.6 (-20.8, 1.6)
-6.8 (-13.2, -0.3)
Low
4
-1.5 (-51.6, 48.7)
-15.9 (-55.2, 23.4)
High + Medium
23
-8.0 (-15.2, -0.7)
-2.8 (-6.4, 0.8)
Sample Timing Strata
Early
12
-7.3 (-16.0, 1.4)
-3.1 (-8.0, 1.8)
Late
10
-3.9 (-17.7, 9.9)
0.1 (-5.7, 5.9)
Post
5
-28.3 (-69.3, 12.7)
-12.8 (-23.9,-1.6)
Late + Post
15
-8.5 (-21.0, 4.1)
-3.7 (-10.0, 2.6)
Symbols and abbreviations: n = sample size; p = pooled estimate of change in birth weight (g) per In (ng/mL) or
ng/mL PFHxS exposure; CI = confidence interval.
C.1.8. Summary of Meta-Analysis of PFHxS Effects on Birth Weight
7 Similar to the hazard synthesis of all the categorical and continuous birth weight results
8 detailed in Section 3.2.2, the meta-analysis pooled estimate of 27 studies showed a statistically
9 significant decrease in mean birth weight of 7.7 g (95%CI: -14.8, -0.5) per natural log-unit increase
This document is a draft for review purposes only and does not constitute Agency policy.
C-21 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
in maternal serum PFHxS. This overall result was similar when studies were restricted to just the
medium confidence studies ((3 = -9.6 g; 95%CI: -20.8,1.6), high confidence studies ((3 = -6.8 g;
95%CI: -16.3, 2.8) or the medium + high confidence studies ((3 = -8.0 g; 95%CI: -15.2, -0.7). The
smallest differences were null for the low confidence studies, with high confidence, medium
confidence, and a combined estimate of those two being similar to the overall pooled estimate.
These higher quality studies are anticipated to be the least susceptible to potential biases.
Similarly, although a consistent gradation in birth weight deficits was not seen across
sample timing, the early sampled stratum has a pooled estimate similar in magnitude ((3 = -7.3 g) to
the overall and the higher confidence studies. Interestingly, this value was fairly comparable to that
seen in the later pregnancy sampled studies ((3 = -3.9 g) in contrast to large differences seen in the
postpartum studies which were predominantly based on umbilical cord samples. A four-fold
difference was seen when comparing pooled estimates from the twelve early sample studies ((3 =
-7.3 g; 95%CI: -16.0,1.4) and the five studies with post-pregnancy samples ((3 = -28.3 g; 95%CI:
-69.3,12.7); however, the CI for the post-pregnancy samples is wide and completely encompasses
the CI for the early samples.
Overall, these data are suggestive of a pattern of later sampling times for PFHxS showing
larger deficits in birth weight, a pattern that may suggest greater bias in later samples. And,
although the postpartum sampled studies have considerably larger results, the small decrease of -
7.3 grams (per each ln-unit PFHxS increase) from early pregnancy sample studies was not too
dissimilar to what was seen amongst studies with maternal biomarkers sampled late in pregnancy.
In comparison to meta-analyses of PFOA fSteenland et al, 2018) and PFOS fPzierlenga et al, 2020).
a strength of our meta-analysis was the ability to separate results across three different strata.
Further investigation should be undertaken to identify if the large differences between late
pregnancy and postpartum samples is unique to PFHxS or is common among PFAS. One interesting
finding related to this topic is that a recent study reported no decrease in PFHxS serum
concentration in serial sampling during pregnancy fChen etal. 20211. This finding is in contrast to
the other long-chain PFAS which uniformly decreased during pregnancy in the same study, which
suggests that maternal physiological changes during pregnancy affect PFHxS differently than other
long-chain PFAS.
There was a suggestion of an effect in the overall estimate of all 27 studies as well as the
medium and high confidence study and the early sample subsets. The latter strata may still be
impacted by potential bias from pregnancy hemodynamics since the categorization approach was
based on samples that may have included a minority of late trimester participants. While the source
of any differences between late pregnancy and postpartum remain unclear, the data do suggest that
potential bias from pregnancy hemodynamics should continue to be examined as a source of
uncertainty in epidemiological studies. Additional research on the slowly cleared PFAS such as
PFHxS is needed to further delineate any differences and better delineate the potential impact of
pregnancy hemodynamics across the class. The meta-analytical findings along with this research in
This document is a draft for review purposes only and does not constitute Agency policy.
C-22 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
indicative of complex patterns of influence due to pregnancy hemodynamic differences that are not
completely understood.
C.2. AOP-BASED APPROACH FOR EVALUATING POTENTIAL PFHxS-
INDUCED MECHANISMS OF HEPATOTOXICITY MODE OF ACTION
C.2.1. Objective and Methodology
The goal of the qualitative analysis described here is to evaluate the available mechanistic
evidence for PFHxS-induced liver effects to assess the biological plausibility of effects observed in
animal models and identify mechanistic pathways that are conserved across species and strains of
animals and liver cell culture models and are therefore more relevant to human health. The
available mechanistic and toxicological evidence was organized and evaluated in concordance with
the frameworks used for mode of action (MOA) analysis for noncancer effects and development of
Adverse Outcome Pathways (AOP)2 fEdwards etal. 2016: Boobis etal. 2008: IPCS. 20071. PFHxS-
induced hepatic effects reported in in vivo and in cell culture studies was organized according to
the following levels of biological organization: molecular interactions, cellular effects, organ effects,
and organism effects3. As recommended in U.S. EPA (20051. the analysis described here was
focused on the concordance of key events and adverse responses across species to obtain
clarification on the relevance of animal studies to human health (U.S. EPA. 20051.
In addition to analyzing the available evidence published in the peer-reviewed literature
EPA also considered mechanistic evidence from in vitro high-throughput screening (HTS) assays on
PFHxS available from EPA's CompTox Chemicals Dashboard fhttps://comptox.epa.gov/dashboard)
fU.S. EPA. 20191. Bioactivity data from the ToxCast and Tox21 collaborative projects were also
considered at the same levels of biological organization described below. A more detailed
description of the HTS analysis and results is provided in Appendix C3.
C.2.2. Proposed AOP Approach for Evaluation of PFAS-Induced Liver Toxicity
The proposed MOA displayed in Figure C-3 is based on molecular initiating events, key
events, and adverse outcomes identified in previous mechanistic evaluations and reviews on PFOS
and PFOA fATSDR. 2018: Li etal. 2017b: U.S. EPA. 2016a. b), which are structurally related to
PFHxS and among the most well-studied PFAS. Additional reviews on biological pathways
2Although the World Health Organization [WHO]-International Programme on Chemical Safety (IPCS]-MOA
and the Organization for Economic Co-operation and Development (OECD]-AOP frameworks are similar in
the identification and analysis of key events following modified Bradford-Hill criteria (Meek et ai, 2014).
AOPs are chemical agnostic whereas MOA analyses are intended to inform health assessments of individual
(or groups of) chemical(s) (Edwards et ai,, 2016).
3Mechanistic evidence from individual chemicals was summarized in a supplementary (MS Excel) database to
facilitate the qualitative analysis of the outcomes reported in the available studies.
This document is a draft for review purposes only and does not constitute Agency policy.
C-23 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
associated with chemical-induced cancer and noncancer liver effects were also consulted (see
citations below). A summary of the MOA is presented below.
At the molecular level, experimental studies using in vivo and cell culture models have
shown that perfluorinated compounds such as PFOS and PFOA can activate several nuclear
receptor pathways including the constitutive androstane receptor (CAR), the pregnane X receptor
(PXR), the farnesoid X receptor (FXR), the peroxisome proliferator activated receptor alpha
(PPARa) and gamma (PPARy), estrogen receptor alpha (ERa) and other receptor-independent cell
signaling pathways (e.g., phosphatidylinositol 3-kinase-serine/threonine protein kinase (PI3KAkt)
signal transduction pathway, and the nuclear factor kappa B pathway [NFkB]) (ATSDR. 2018:
Gleason.: ; at 2017b: NIDWOI. 2017: U.S. EPA. 2016a. b). PFOS- and PFOA induced
activation of PPARa is associated with hepatocellular hypertrophy caused by peroxisome
proliferation, and increased peroxisomal fatty acid (3 oxidation and cytochrome P450 4A (CYP4A)
expression and activity fATSDR. 2018: U.S. EPA. 2016a. b), and altered cholesterol metabolism (U
etal. 2017b). Increased PPARa activity can lead to oxidative stress via induction of acyl CoA
oxidase expression and activity, and H202 production in peroxisomes (Hall et at 2012). Several
studies have used genetically modified animal and cell culture models and immortalized human cell
lines to evaluate potential PFOS or PFOA activation of the human PPARa. COS-1 cells transfected
with the murine or human PPARa were responsive to PFAS exposure (U.S. EPA. 2016a. b), and F1
generation PPARahumanized mice were responsive to PFOA-induced expression responsive genes
on GDI8, but unlike wild type animals this response was not apparent on PND 20 fU.S. EPA. 2016b:
Takacs and Abbott. 20071. Studies using human liver cell lines or humanized animal models suggest
that humans are less sensitive to PPARa activation by the perfluorinated compounds PFOS and
PFOA (reviewed in Li etal. f2017b): U.S. EPA f2016a)). PPARa has also been shown to be activated
by exposure to several PFAS, including PFOS, PFOA, PFNA, and PFHxS (ATSDR. 2018: Li et at
2017b). Although PPARa is not expressed in high levels in the liver, its activation by
pharmaceuticals and xenobiotic compounds has been proposed to be associated with hepatic
steatosis caused by lipid accumulation fAngrish et at 2016: Mellor etal. 2016).
As described above exposure to perfluorinated compounds such as PFOS and PFOA has also
been shown to activate other nuclear receptor and cell signaling pathways, including the CAR, PXR,
FXR, ERa, NFkB, and oxidative stress responsive nuclear factor erythroid 2 related factor 2 (Nrf2)
(ATSDR. 2018: Li et at 2017b: U.S. EPA. 2016a). Furthermore, experiments using null animal
models exposed to several PFAS suggest that activation of CAR/PXR occurs independently of PPARa
(ATSDR. 2018: Li et at 2017b). Previous analyses of chemical-induced hepatotoxicity suggest that
activation of these cell signaling pathways in experimental models is associated with increased
expression and activity of xenobiotic metabolizing enzymes (XMEs) floshi-Barve et at 2015: Hall et
at 2012). formation of reactive metabolites, alterations in cellular lipid metabolism fAngrish etal.
2016). and endoplasmic reticulum damage floshi-Barve et at 2015).
This document is a draft for review purposes only and does not constitute Agency policy.
C-24 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
At the cellular level, exposure to PFAS such as PFOS and PFOA has been shown to increase
reactive oxygen species production and oxidative damage to cellular macromolecules fATSDR.
2018: Li etal.. 2017b: U.S. EPA. 2016bl: promote mitochondrial damage, inhibit mitochondrial
function, activate mitochondrial-mediatedcell death fLi etal.. 2017b: U.S. EPA. 2016al: increase
endoplasmic reticulum stress fU.S. EPA. 2016bl: induce DNA damage fATSDR. 2018: U.S. EPA.
2016b); disrupt intercellular gap junction communication fATSDR. 20181: elevate
production/levels of pro-inflammatory cytokines (U.S. EPA. 2016al: alter lipid and glucose
metabolism and bile acid biosynthesis (U.S. EPA. 2016a. b); and increase hepatocellular death fLi et
al.. 2017b: U.S. EPA. 2016bl. These path ways/mechanisms are associated with toxicant-induced
liver disease and can promote steatohepatitis and fibrosis fBessone etal.. 2019: Angrish etal..
2016: Cao etal.. 2016: Toshi-Barve etal.. 2015: Wahlangetal.. 20131.
Figure C-3. The proposed MOA in the figure above is based on previous
analyses on PFAS-induced (e.g., PFOA/PFOS) liver toxicity and the role of
nuclear receptor pathways in hepatotoxicity.
C.3. SUMMARY OF RELEVANT HIGH-THROUGHPUT SCREENING ASSAYS
FROM EPA'S COMPTOX CHEMICALS DASHBOARD
C.3.1. In vitro Bioreactivity Data Relevant to the Mechanisms of PFHxS-Induced Liver Effects
In vitro high-throughput screening (HTS) assays for PFHxS were downloaded from EPA's
CompTox Chemicals Dashboard fhttps://comptox.epa.gov/dashboardl fU.S. EPA. 20191. which
provides bioactivity data from the ToxCast and Tox21 collaborative projects. Available information
most pertinent to the analysis of the potential mechanisms of PFHxS-induced liver effects was
extracted to supplement and augment mechanistic findings from studies in the peer-reviewed
This document is a draft for review purposes only and does not constitute Agency policy.
C-25 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
literature previously described. Results (active/inactive, AC50 values, and scaled activity) from in
vitro assays were obtained, filtering out background control assays and nonspecific responses from
inducible reporter gene assays analyzed in the negative fitting direction relative to the control
('_dn')4. Bioactivity data were analyzed based on the type of biological response or biological target
using the annotation structure within the ToxCast assay summary information flJ.S. EPA. 20191.
PFHxS was active in 34 of 743 assays, of which 10 were performed in human liver tissues.
PFHxS was active in 0 of 54 unique assay endpoints in human hepatoma HepG2 cells. The majority
of the active liver relevant endpoint assays were nuclear receptor assays (see Figure C-4, Table C-
9). PFHxS was able to induce reporter assays for multiple nuclear receptors including PPARa,
PPARy, RXR and LXR as well as transcriptional factors, FOS, and NRF2 (see Figure C-5, Table C-9).
Nuclear receptor activities were further investigated to provide additional information to known
interactions of PFHxS with these receptor-mediated signaling pathways in ToxCast/Tox21 assays
profiling multiple endpoints (e.g., receptor binding, co-regulator recruitment and gene
transactivation) and cell types (see Figure C-5, Table C-9). As previously mentioned, PFHxS induced
activity of specific steroid/xenobiotic sensing receptors, most notably PXR, RXR and PPAR (see
Figure C-5). PFHxS interacted with the retinoic acid (RAR, 1 out 19 assays) and the human RXR (1
out of 10) in receptor binding assays. PFHxS was active in 4 out of 27 PPAR-related assays, showing
transcriptional activation, including induction of a PPAR-response element driven reporter gene
assay, PPARa and a-dependent reporter gene expression and binding to the human PPARa.
^Inducible gene reporter assays were not optimized to detect loss of signal (U.S. EPA. 2019): therefore,
responses from assays analyzed in the negative fitting direction relative to the control ('_dn') are considered
non-specific and are not presented herein.
This document is a draft for review purposes only and does not constitute Agency policy.
C-26 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
4 5
4
3.5
& 3
t> 2.5
<
~o
a 2
U C-
QJ
1/1 15
1
0.5
0
• Cell morphology
• Nuclear receptor
• Transcriptional Factor
10
100
1000
AC50
Figure C-4. Bioactivity data for PFHxS from in vitro HTS ToxCast/Tox21 assays in
human liver tissues. Scatterplots show AC50 and scaled activity values from assays
visualized according to the type of biological response or biological target AC50 values
refer to the concentration that elicits half maximal response and the scaled activity refers to
the response value divided by the activity cutoff. Assays for which chemicals were inactive
are not displayed. Additional information on these assays can be found in Table C-8.
This document is a draft for review purposes only and does not constitute Agency policy.
C-27 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Figure C-5. Summary of positive nuclear receptor assays in human liver tissue.
Additional information on these nuclear receptor assays may be found in
Table C-9.
Table C-7. Bioactivity summary for PFHxS from in vitro HTS assays from
ToxCast/Tox21 conducted in human liver tissue and grouped by biological
response / target.
Assay namea,b
Activity call
Scaled
activity
AC50 (nM)
Assay design
Cell morphology
TOX21_MMP_ratio_up
ACTIVE
4.19
65.1
Membrane potential
reporter
Nuclear receptor
TOX21_RXR_BLA_Agonist_ratio
ACTIVE
1.31
9.28
Inducible reporter
ATG_ERE_CIS_up
ACTIVE
2.74
96.9
Inducible reporter
ATG_PPRE_CIS_up
ACTIVE
1.65
44.6
Inducible reporter
ATG_PPARa_TRANS_up
ACTIVE
1.35
11.2
Inducible reporter
ATG_PPARg_TRANS_up
ACTIVE
1.69
30.5
Inducible reporter
NVS_NR_hPPARg
ACTIVE
1.51
27.5
Binding reporter
ATG_RARg_TRANS_up
ACTIVE
1.28
6.91
Inducible reporter
ATG_LXRa_TRANS_dn
ACTIVE
1.53
14.4
Inducible reporter
This document is a draft for review purposes only and does not constitute Agency policy.
C-28 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Assay nameab
Activity call
Scaled
activity
AC50 (nM)
Assay design
ATG_PXRE_CIS_up
ACTIVE
2.77
22.0
Inducible reporter
Transcriptional factor
ATG_AP_l_CIS_up
ACTIVE
1.25
66.5
Inducible reporter
ATG_NRF2_ARE_CIS_up
ACTIVE
2.12
99.5
Inducible reporter
aData were sourced from EPA's CompTox Chemicals Dashboard (U.S. EPA, 2019).
background control assays, inactive and nonspecific responses from inducible reporter gene assays analyzed in
the negative fitting direction relative to the control ('_dn') are not presented herein.
NA = not applicable.
C.3.2. In vitro Bioactivity Data Relevant to the Mechanisms of PFHxS-Induced Thyroid
Effects
Table C-8. Endocrine disruptor screening program 21 assay summary results3
Assay name
Description
Cell model
Active/
inactive
ERE assays
ACEA_ER_80hr
Growth reporter (proliferation) assay
T47D, human breast cell line
Inactive
ATG_ERE_CIS_up
Reporter Gene Assay
HepG2 human hepatoma cell line
Active
ATG_Era_TRANS_up
Reporter Gene Assay
HepG2 human hepatoma cell line
Inactive
NVS_NR_hER
Binding reporter Assay
MCF7, human breast cell line
Inactive
OT_ER_ERaERa_0480
Binding Reporter Assay
HEK293T, Human Kidney cell line
Inactive
OT_ER_ERaERa_1440
Binding Reporter Assay
HEK293T, Human Kidney cell line
Inactive
OT_ER_ERbERb_0480
Binding Reporter Assay
HEK293T, Human Kidney cell line
Inactive
OT_ER_ERbERb_1440
Binding Reporter Assay
HEK293T, Human Kidney cell line
Inactive
OT_Era_EREGFP_0120
Reporter Gene Assay
HeLa, Human Cervix cell line
Inactive
OT_Era_EREGFP_0480
Reporter Gene Assay
HeLa, Human Cervix cell line
Inactive
TOX21_ERa_BLA_Agonist
Reporter Gene Assay
HEK293T, Human Kidney cell line
Inactive
TOX21_ERa_BLA_Antagonist
Reporter Gene Assay
HEK293T, Human Kidney cell line
Inactive
Tox_ERa_LUC_VM7_Agonist
Reporter Gene Assay
VM7, Human Ovary cell line
Inactive
Tox_ERa_LUC_VM7_Antagonist
Reporter Gene Assay
VM7, Human Ovary cell line
Inactive
Tox_ERa_LUC_VM7_Antagonist
Reporter Gene Assay
VM7, Human Ovary cell line
Inactive
Tox_ERb_BLA_Agonist
Binding Reporter Assay
HEK293T, Human Kidney cell line
Inactive
Tox_ERb_BLA_Antagonist
Binding Reporter Assay
HEK293T, Human Kidney cell line
Inactive
OT_ER_ERaERb_0480
Binding Reporter Assay
HEK293T, Human Kidney cell line
Inactive
OT_ER_ERaERb_1440
Binding Reporter Assay
HEK293T, Human Kidney cell line
Inactive
TOX21_ERa_BLA_Antagonist_viability
Viability Assay
HEK293T, Human Kidney cell line
Inactive
TOX2 l_ERa_LUC_VM7_Antagonist_0.5
nM_E2_viability
Viability Assay
VM7, Human Ovary cell line
Inactive
ATG_ERa_TRANS_dn
Reporter Gene Assay
HepG2 human hepatoma cell line
Inactive
ATG_ERA_CIS_dn
Reporter Gene Assay
HepG2 human hepatoma cell line
Inactive
ACEA_ER_AUC_viability
Viability Assay
T47D, Human Breast cell line
Inactive
TOX2 l_ERa_LUC_VM7_Antagonist_0.1
nM_E2_viability
Viability Assay
VM7, Human Ovary cell line
Inactive
TOX21_ERb_BLA_Agonist_viability
Viability Assay
HEK293T, Human Kidney cell line
Inactive
TOX21_ERb_BLA_Antagonist_viability
Viability Assay
HEK293T, Human Kidney cell line
Inactive
This document is a draft for review purposes only and does not constitute Agency policy.
C-29 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Assay name
Description
Cell model
Active/
inactive
AR assays
ATG_AR_TRANS_up
Reporter Gene Assay
HepG2 human hepatoma cell line
Inactive
OT_AR_ARELUC_AG_1440
Reporter Gene Assay
CHO-K1, Chinese Hamster Ovary
Inactive
OT_AR_ARSRC1_0480
Binding Assay
HEK293T, Human Kidney cell line
Inactive
OT_AR_ARSRC1_0960
Binding Assay
HEK293T, Human Kidney cell line
Inactive
TOX2 l_AR_BLA_Ago n ist_
Reporter Gene Assay
HEK293T, Human Kidney cell line
Inactive
TOX21_AR_LU C_Antago n i st_
Reporter Gene Assay
HEK293T, Human Kidney cell line
Inactive
TOX21_AR_LUC_M DAKB2_Agonist
Reporter Gene Assay
MDA-kb2, Human Breast cell line
Inactive
TOX21_AR_LUC_M DAKB2_Antagonist
Reporter Gene Assay
MDA-kb2, Human Breast cell line
Inactive
TOX21_AR_LUC_M DAKB2_Antagonist
Reporter Gene Assay
MDA-kb2, Human Breast cell line
Inactive
TOX21_AR_LUC_M DAKB2_Agonist
Reporter Gene Assay
MDA-kb2, Human Breast cell line
Inactive
ACEA_AR_agonist_80hr
Signaling Assay
22Rvl, Human Prostate cell line
Inactive
TOX2 l_AR_BLA_Ago n ist
Viability Assay
HEK293T, Human Kidney cell line
Inactive
TOX21_AR_LUC_M DAKB2
Viability Assay
MDA-kb2, Human Breast cell line
Inactive
ATG_AR_TRANS_up
Reporter Gene Assay
HepG2 human hepatoma cell line
Inactive
TOX21_AR_LUC_M DAKB2
Viability Assay
MDA-kb2, Human Breast cell line
Inactive
TOX21_AR_LUC_M DAKB2
Viability Assay
MDA-kb2, Human Breast cell line
Inactive
Thyroid Assays
ATG_THRal_TRANS_up
Reporter Gene Assay
Inactive
NVS_NR_hTRa_Antagonist
Binding Assay
HepG2 human hepatoma cell line
Inactive
TOX21_TSH R_Ago n ist_ratio
Signaling Assay
Human cell line (no other information
available)
Inactive
T OX21_TSH R_Antagon ist_ratio
Signaling Assay
HEK293T, Human Kidney cell line
Inactive
TOX21_TR_LUC_G H 3_Ago n ist
Reporter Gene Assay
HEK293T, Human Kidney cell line
Inactive
TOX21_TR_LUC_G H 3_An tago n ist
Reporter Gene Assay
GH3, Rat pituitary cell line
Inactive
TOX21_TR_LUC_G H 3_An tago n ist_
viability
Viability Assay
GH3, Rat pituitary cell line
Inactive
ATG_TH Ra1_TRANS_d n
Reporter Gene Assay
HepG2 human hepatoma cell line
Inactive
T OX21_TSH R_wt_ratio
Background Control Assay
HEK293T, Human Kidney cell line
Inactive
NIS_RAIU_inhibition
Binding Assay
HEK293T, Human Kidney cell line
Active
NIS_H EK293T_CTG_Cytotoxicity
Viability Assay
HEK293T, Human Kidney cell line
Inactive
Steroid Assays
NVS_ADME_hCYP19Al
Enzymatic Activity
Human, Cell Free
Inactive
TOX21_Aromatase_lnhibition
Reporter Gene Assay
MCF-7, Human Breast cell line
Inactive
TOX21_Aromatase_lnhibition_viability
Viability Assay
MCF-7, Human Breast cell line
Inactive
aCompTox Chemistry Dashboard accessed 3/5/2020
httpsi//comptox.epa.eov/dashboard/dsstoxdb/resylts?search=DTXSlD3037709#invitrodb-bioassays-toxcast-data.
This document is a draft for review purposes only and does not constitute Agency policy.
C-30 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
APPENDIX D. BENCHMARK DOSE MODELING
RESULTS
This appendix provides technical detail on dose-response evaluation and determination of
points of departure (PODs) for relevant toxicological endpoints. The endpoints are modeled using
EPA's Benchmark Dose Software (BMDS, Version 3.2). Sections E.l (noncancer) and Section E.2
(cancer) describe the common practices used in evaluating the model fit and selecting the
appropriate model for determining the POD, as outlined in the Benchmark Dose Technical Guidance
document fU.S. EPA. 20121.
D.l. BENCHMARK DOSE MODELING SUMMARY FOR NONCANCER
ENDPOINTS
D.l.l. Benchmark Dose Modeling Approaches
The endpoints selected for benchmark dose (BMD) modeling include decreased serum
antibody concentrations for tetanus and diphtheria (Budtz-l0rgensen and Grandiean. 2018a:
Grandiean etal.. 2012) and decreased birth weight (Manzano-Salgado etal.. 2019: Buck Louis etal..
2018: Shoaffetal.. 2018: Starling etal.. 2017: Bach etal.. 2016al. The internal doses reported in the
human studies were used in the BMD modeling and then converted to human equivalent doses
(HEDs) using the pharmacokinetic (PK) model described in Section 3.1 of the main document; the
modeling results are presented in this appendix.
Modeling Results for Decreased Tetanus Antibody Concentrations at 7 Years of Age and PFHxS
Measured at 5 Years of Age
Budtz-largensen and Grandiean f2018al fit multivariate models of perfluorohexanesulfonic
acid (PFHxS) measured at age 5 years, against log2-transformed anti-tetanus antibody
concentrations measured at the 7-year-old examination controlling for sex, exact age at the 7-year-
old examination, and booster type at age 5 years. Models were evaluated with additional control for
PFOS (as log2[PFOS]) and PFOA (as log2[PFOA]), and without PFOS and PFOA. Three model shapes
were evaluated by Budtz-l0rgensen and Grandiean (2018a) using likelihood ratio tests: a linear
model, a piecewise-linear model with a knot at the median PFHxS concentration, and a logarithmic
function. The logarithmic functions did not fit better than the piecewise-linear functions fBudtz-
Targensen and Grandiean. 2018al. There was evidence that the piecewise-linear model fit better
than the linear model for both the PFHxS exposure model without adjustment for PFOS and PFOA
(p = 0.002; see Budtz-l0rgensen and Grandiean (2018a) Table 3) and for the model that did adjust
for PFOS and PFOA (p = 0.05).
This document is a draft for review purposes only and does not constitute Agency policy.
D-l DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
1 Table D-1 summarizes the results from Budtz-l0rgensen and Grandiean f2018a1 for PFHxS
2 at age 5 years and tetanus antibodies at age 7 years. These regression coefficients ((3) and their
3 standard errors (SE) were computed by EPA from the published BMDs and BMDL based on a
4 benchmark response (BMR) of 5% decrease in the log2-transformed antibody concentration in
5 Table 1 of Budtz-tergensen and Grandiean f2018al.5 As Budtz-l0rgensen and Grandiean f2018al
6 log2-transformed the outcome variable, the BMR was measured in unit of log2[tetanus antibody
7 concentration] as log2(l-0.05) = 0.074 log2(IU/mL).
Table D-l. Results specific to the low-dose slope from the piecewise- linear
analyses of PFHxS measured at age 5 years and log2 (tetanus antibody
concentrations) measured at age 7 from Table 1 in Budtz-Iargensen and
Grandiean (2018a) in a single-PFAS model and in a multi-PFAS model
PFOS and
Lower bound
PFOA
Slope (P) per
SE(P)
slope
Exposure
Model shape
adjusted
ng/mL
ng/mL
Slope (P) fit
(pLB)ng/mL
PFHxS at Age 5
Piecewise
No
-2.47
0.75
p = 0.001
-3.70
PFHxS at Age 5
Piecewise
Yes
-1.85
1.12
p = 0.100
-3.70
8 Interpretation of results in Table D-l:
9 • PFHxS is a significant predictor in the single-PFAS model ((3 = -2.47; p = 0.001).
10 • Effects of PFHxS in the single-PFAS model are attenuated when log2[PFOS] and
11 log2[PFOA] are included in the model ((3 = -1.85; p = 0.100).
12 • The point estimate results for PFHxS ((3) in the single-PFAS model are potentially
13 confounded by PFOS and/or PFOA since there was a 25% reduction in the effect size for
14 PFHxS from -2.47 to -1.85 when controlling for PFOS and PFOA.
15 ° One explanation is that PFOS and/or PFOA was a confounder of the PFHxS effect and
16 controlling for those co-exposures removed confounding.
17 ° Another possibility is that controlling for co-exposures like PFOS and PFOA actually
18 induced confounding fWeisskopf et al. 2018: Weisskopf and Webster. 20171.
19 ° The reasons for the change in main effect size for PFHxS are not known. For this
20 reason, there is uncertainty in knowing which point estimate is the best
21 representation of any effect of PFHxS.
5BitdtzJgrgensen and Grandiean (2018a) computed BMDs and BMDLs using a BMR of 5% decrease in the
log2(antibody concentrations). Their formula, BMD = log2(l-BMR)/(3, can simply be reversed to solve for (3 =
log2(l-BMR)/BMD. For negative dose-response where more exposure results in lower antibody
concentration, the BMDL is based on the lower bound of |3, (|3lb). Thus, the |3lb = log2(l-BMR)/BMDL. The
SE(|3) = (P - Plb)/1.645. The p-value is the two-sided probability that Z <= SE(|3)/|3.
This document is a draft for review purposes only and does not constitute Agency policy.
D-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
• However, the lower bounds on the point estimates ((3lb) are the same for the single-
PFAS and multi-PFAS models, minimizing the effect of the potential confounding given
the lower bound is ultimately used for point-of-departure derivation.
° The definition of the RfD, which is based upon the (3lb, includes allowing for an order
of magnitude (10-fold or 1,000%) uncertainty in the estimate and the uncertainty
for potential confounding in the BMD from including, or excluding, PFOS and PFOA
here is about 25%, while there is no uncertainty for potential confounding in the
BMDL as those values are the same.
Selection of the Benchmark Response
The benchmark dose (BMD) approach involves dose-response modeling to obtain BMDs,
i.e., dose levels corresponding to specific response levels near the low end of the observable range
of the data and the lower limit of the BMD (BMDLs) to serve as potential PODs for deriving
quantitative estimates below the range of observation fU.S. EPA. 20121. Selecting a BMR to estimate
the BMDs and BMDLs involves making judgments about the statistical and biological characteristics
of the data set and about the applications for which the resulting BMDs and BMDLs will be used. An
extra risk of 10% is recommended as a standard reporting level for quantal data for toxicological
data. Biological considerations may warrant the use of a BMR of 5% or lower for some types of
effects as the basis of the POD for a reference value. However, a BMR of 1% has typically been used
for quantal human data from epidemiology studies flJ.S. EPA. 20121. although this is more typically
used for epidemiologic studies of cancer mortality within large cohorts of workers which can
support the statistical estimation of small BMRs.
A blood concentration for tetanus antibodies of 0.1 IU/mL is sometimes cited in the tetanus
literature as a "protective level" and Grandiean etal. (20171 noted that the Danish vaccine producer
Statens Serum Institut recommended the 0.1 IU/mL "cut-off" level "to determine whether antibody
concentrations could be considered protective"; and Galazka and Kardvmowicz (19891 mentions
the same concentration), but Galazka et al. (19931 argues:
"The amount of circulating antitoxin needed to ensure complete immunity against
tetanus is not known for certain. Establishment of a fixed level of tetanus antitoxin
does not take into consideration variable conditions of production and adsorption of
tetanus toxin in the anaerobic area of a wound or a necrotic umbilical stump. A
given serum level could be overwhelmed by a sufficiently large dose of toxin.
Therefore, there is no absolute protective level of antitoxin and protection results
when there is sufficient toxin-neutralizing antibody in relation to the toxin load
(Passen and Andersen. 19861."
In the absence of a clear definition of an adverse effect for a continuous endpoint like
antibody concentrations, a default BMR of 1 SD change from the control mean may be selected, as
suggested in EPA's draft Benchmark Dose Technical Guidance Document (U.S. EPA. 20121. As noted
above, a lower BMR can also be used if it can be justified on a biological and/or statistical basis.
Figure D-l replicates a figure in the technical guidance (page 23; (U.S. EPA. 201211 to show that in a
This document is a draft for review purposes oniy and does not constitute Agency poiicy.
D-3 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
1 control population in which 1.4% are considered to be at risk of having an adverse effect, a
2 downward shift in the control mean of one SD results in a ~10% extra risk of being at risk of having
3 an adverse effect
This document is a draft for review purposes only and does not constitute Agency policy.
D-4 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information—Perfluorohexanesulfonic Acid and Related Salts
Standard deviation units
Standard deviation units
Figure D-l. Difference in population tail probabilities resulting from a 1
standard deviation shift in the mean from a standard normal distribution,
illustrating the theoretical basis for a baseline BMR of 1 SD.
This document is a draft for review purposes only and does not constitute Agency policy,
D-5 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Statistically, the technical guidance additionally suggests that studies of developmental
effects can support lower BMRs. Biologically, a BMR of Vi SD is a reasonable choice as anti-tetanus
antibody concentrations prevent against tetanus, which is a rare, but severe and sometimes fatal
infection, with a case-fatality rate in the United States of 13% during 2001-2008 fiiang et al.
20181. The case-fatality rate can be more than 80% for early lifestages cases fPatel and Mehta.
19991. Selgrade (2007) suggests that specific immunotoxic effects observed in children may be
broadly indicative of developmental immunosuppression impacting these children's ability to
protect against a range of immune hazards, which has the potential to be a more adverse effect that
just a single immuno-toxic effect Thus, decrements in the ability to maintain effective levels of
tetanus antitoxins following immunization may be indicative of wider immunosuppression in these
children exposed to PFHxS. By contrast, a BMR of 1 SD may be more appropriate for an effect that
would be considered "minimally adverse." A BMR smaller than Vi SD is generally selected for severe
effects (e.g., 1% extra risk of cancer mortality); decreased antibody concentrations offer diminished
protection from severe effects but are not themselves severe effects.
Following the technical guidance (U.S. EPA. 2012). EPA derived BMDs and BMDLs
associated with a 1 SD change in the distribution of log2 (tetanus antibody concentrations), and
Vi SD change in the distribution of log2 (tetanus antibody concentrations). The SD of the
log2 (tetanus antibody concentrations) at age 7 years was estimated from the distributional data
presented in Grandiean et al. (2012) as follows: the interquartile range (IQR) of the tetanus
antibody concentrations at age 7 years in IU/mL was (0.65, 4.6). Log2-tranforming these values
provides the IQR in log2 (IU/mL) as (-0.62, 2.20). Assuming that these log2-transformed values are
reasonably represented by a normal distribution, the width of the IQR is approximately 1.35 SDs.
Thus, SD = IQR/1.35, and the SD of tetanus antibodies in log2 (IU/mL) is (2.20 - (-0.62))/1.35 = 2.09
log2 (IU/mL). To show the impact of the BMR on these results, Table D-2 presents the BMDs and
BMDLs at BMRs of % SD and 1 SD.
While there was not a clear definition of an adverse effect for a continuous endpointlike
antibody concentrations, the value of 0.1 IU/mL is sometimes cited. As a check, EPA evaluated how
much extra risk would have been associated with a BMR set at a cut-off value of 0.1 IU/mL. Using
the observed distribution of tetanus antibodies at age 7 years in log2 (IU/mL), EPA calculated that
2.8% of those values would be below the cut-off value of 0.1 IU/mL which is -3.32 log2 (IU/mL). A
BMR of Vi SD resulted in 7.9% of the values being below that cutoff, which is 5.1% extra risk and
shows that the generic guidance that a BMR of Vi SD can provide a reasonably good estimate of 5%
extra risk. Figure D-2 shows an example.
This document is a draft for review purposes only and does not constitute Agency policy.
D-6 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Tetanus antibody concentrations in Log2(IU/ml)
Figure D-2. Difference in population tail probabilities resulting from a V2
standard deviation shift in the mean from an estimation of the distribution of
log2(tetanus antibody concentrations at age 7 years).
Table D-2. BMDs and BMDLs for effect of PFHxS at age 5 years on anti-tetanus
antibody concentrations at age 7 years using a BMR of V2 SD change in
log2(tetanus antibodies concentration) and a BMR of 1 SD log2(tetanus
antibodies concentration)
Estimated without control of PFOS and PFOA
Estimated with control of PFOS and PFOA
BMR
BMD (ng/mL)
P = -2.47 per ng/mL
BMDL (ng/mL)
Plb = -3.70 per ng/mL
BMD (ng/mL)
P = -1.85 per ng/mL
BMDL (ng/mL)
Plb = -3.70 per ng/mL
KSD
0.424
0.282a
0.565
0.282
1 SD
0.847
0.565
1.130
0.565
aDenotes the selected POD.
1 The lowest serum PFHxS concentration measured at age 5 years was 0.02 ng/mL, the 5th
2 percentile was 0.2 ng/mL, and the 10th percentile was 0.3 ng/mL fGrandiean and Bateson. 20211
3 so the estimated BMDL for a BMR of Vi SD (BMDL^sd) in the single-PFAS model is at about the 10th
4 percentile of the observed distribution. No information was available to judge the fit of the model in
This document is a draft for review purposes only and does not constitute Agency policy.
D-7 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
the range of the BMDLs, but the BMD and BMDL were both within the range of observed values and
the model fit PFHxS well.
The BMDy2 sd estimate from the multi-PFAS models is 33% higher than the BMDy2 sd estimate
from the models with just PFHxS, but the BMDLy2 sd estimates are equal given the lower bounds for
both models were equal. The change in BMD estimates may, or may not, reflect control for any
potential confounding of the regression effect estimates. While it is not clear which PFAS model
provided "better" estimate of the point estimate of the effect of PFHxS, the two BMDLy2 sd estimates
that serve as the PODs are equal and EPA advanced the derivation based on common BMDLy2 sd
estimates of 0.282 ng/mL from both the single-PFAS and the multi-PFAS models. However,
confidence was somewhat diminished by the potential confounding in the main effect—even
though there was no observed confounding of the BMDL. Medium confidence in the BMDLs for
PFHxS.
For immunotoxicity related to tetanus associated with PFHxS exposure measured at
age 5 years, the POD is based on a BMR of V2 SD and a BMDL% SD of 0.282 ng/mL.
Modeling Results for Decreased Diphtheria Antibody Concentrations at 7 Years of Age and
PFHxS Measured at 5 Years of Age
Budtz-l0rgensen and Grandiean f2018al fit multivariate models of PFHxS measured at age 5
years, against log2-transformed anti-diphtheria antibody concentrations measured at the 7 year-old
examination controlling for sex, exact age at the 7 year-old examination, and booster type at age 5
years. Models were evaluated with additional control for PFOS (as log2[PF0S]) and PFOA (as
log2[PF0A]), and without PFOS and PFOA. Three model shapes were evaluated by Budtz-largerisen
and Grandiean (2018a) using likelihood ratio tests: a linear model of PFHxS, a piecewise-linear
model with a knot at the median, and a logarithmic function. The logarithmic functions did not fit
better than the piecewise-linear functions fBudtz-largensen and Grandiean. 2018a). There was
evidence that the piecewise-linear model fit better than the linear model for the PFHxS exposure
without adjustment for PFOS and PFOA (p = 0.05; see in Budtz-l0rgensen and Grandiean (2018a).
Table 3), but not for the model that adjusted for PFOS and PFOA (p = 0.44). Table D-3 summarizes
the results from Budtz-l0rgensen and Grandiean (2018a) for diphtheria in this exposure window.
These regression coefficients ((3) and their standard errors (SE) were computed by EPA from the
published BMDs and BMDL based on a BMR of Vi SD in log2(diphtheria antibody concentrations) in
Table 1 of Budtz-l0rgensen and Grandiean (2018a).
This document is a draft for review purposes only and does not constitute Agency policy.
D-8 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table D-3. Results specific to the low-dose slope from the piecewise- linear
analyses of PFHxS measured at age 5 years and log2 (diphtheria antibodies)
measured at age 7 years from Table 1 in Budtz-lorgensen and Grandiean
(2018a) in a single-PFAS model and in a multi-PFAS model
Exposure
Model shape
PFOS and
PFOA adjusted
Slope (P) per
ng/mL
SE(P)
ng/mL
Slope (P) fit
Lower bound
slope (Plb)
ng/mL
PFHxS at Age 5
Piecewise
No
-1.48
0.60
p = 0.0136
-2.47
PFHxS at Age 5
Piecewise
Yes
-0.67
1.09
p = 0.537
-2.47
1 Interpretation of results in Table D-3:
2 • PFHxS is a significant predictor in the single-PFAS model ((3 = -1.48; p = 0.01).
3 • Effects are attenuated when log2[PF0S] and log2[PF0A] are included in the model ((3 = -
4 0.67; p = 0.54).
5 • The point estimate results for PFHxS are potentially confounded by PFOS and/or PFOA
6 since there was a 55% reduction in the effect size for PFHxS from -1.48 to -0.67 when
7 controlling for PFOS and PFOA.
8 ° One explanation is that PFOS and/or PFOA was a confounder of the PFHxS effect and
9 controlling for those co-exposures removed confounding.
10 ° Another possibility is that controlling for co-exposures like PFOS and PFOA actually
11 induced confounding fWeisskopfetal.. 2018: Weisskopf and Webster. 20171.
12 ° The reasons for the change in main effect size for PFHxS are not known. For this
13 reason, there is uncertainty in knowing which point estimate is the best
14 representation of any effect of PFHxS.
15 • However, the lower bounds on the point estimates ((3lb) are the same for the single-
16 PFAS and multi-PFAS models, minimizing the effect of the potential confounding given
17 the lower bound is ultimately used for point-of-departure derivation.
18 ° The definition of the RfD, which is based upon the (3LB, includes allowing for an
19 order of magnitude (10-fold or 1,000%) uncertainty in the estimate and the
20 uncertainty for potential confounding in the BMD from including, or excluding, PFOS
21 and PFOA here is about 55%, while there is no uncertainty for potential
22 confounding in the BMDL as those values are the same.
This document is a draft for review purposes only and does not constitute Agency policy.
D-9 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Selection of the Benchmark Response
Following the technical guidance fU.S. EPA. 20121. EPA derived BMDs and BMDLs
associated with a one SD change in the distribution of log2(diphtheria antibody concentrations),
and Vi SD change in the distribution of log2(diphtheria antibody concentrations). A blood
concentration for diphtheria antibodies of 0.1 IU/ml is sometimes cited in the diphtheria literature
as a "protective level" (Grandiean et al, 2017) noted that the Danish vaccine producer Statens
Serum Institut recommended the 0.1 III /ml "cut-off" level; and Galazka et al. (1993) mentions the
same concentration), but Galazka et al. (19931 argues:
"However, it has also been shown that there is no sharply defined level of antitoxin
that gives complete protection from diphtheria (Ipsen. 1946). A certain range of
variation must be accepted; the same degree of antitoxin may give an unequal
degree of protection in different persons. Other factors may influence the
vulnerability to diphtheria including the dose and virulence of the diphtheria bacilli
and the general immune status of the person infected (Christenson and Bottiger.
19861. Thus, an antibody concentration between 0.01 and 0.09 IU/ml may be
regarded as giving basic immunity, whereas a higher titer may be needed for full
protection. In some studies that used in vitro techniques, a level of 0.1 IU/ml was
considered protective (Cellesi et al. 1989: Galazka and Kardvmowicz. 19891."
Statistically, the technical guidance suggests that studies of developmental effects can
support lower BMRs. Biologically, a BMR of Vi SD is a reasonable choice as anti-diphtheria antibody
concentrations prevent against diphtheria, which is very rare in the United States, but can cause
life-threatening airway obstruction, or cardiac failure fCollier. 19751. Among 13 cases reported in
the United States during 1996-2016, no deaths were mentioned fLiangetal. 20181. However,
diphtheria remains a potentially fatal disease in other parts of the world Galazka et 331
mentions a case fatality rate of 5-10%) and PFHxS-related changes in anti-diphtheria antibody
concentrations cannot be considered to be 'minimally adverse' given the historic lethality of
diphtheria in the absence of vaccination. Selgrade (20071 suggests that specific immuno-toxic
effects observed in children may be broadly indicative of developmental immunosuppression
impacting these children's ability to protect against a range of immune hazards—which has the
potential to be a more adverse effect that just a single immuno-toxic effect.
Following the technical guidance fU.S. EPA. 20121. EPA derived BMDs and BMDLs
associated with a one SD change in the distribution of log2(diphtheria antibody concentrations) as a
standard reporting level, and Vi SD change in the distribution of log2(diphtheria antibody
concentrations). The SD of the log2(diphtheria antibody concentrations) at age 7 years was
estimated from the distributional data presented in Grand!ean etal. (20121 as follows: the
interquartile range (IQR) of the diphtheria antibody concentrations at age 7 years in IU/mL was
(0.4,1.6). Log2-tranforming these values provides the IQR in log2(IU/mL) as (-1.32, 0.68). Assuming
that these log2-transformed values are similar to the normal distribution, the width of the IQR is
This document is a draft for review purposes only and does not constitute Agency policy.
D-10 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
approximately 1.35 SDs, thus SD = IQR/1.35, and the SD of tetanus antibodies in log2(IU/mL) is
(0.68 - (-1.32))/1.35 = 1.48 log2(IU/mL). To show the impact of the BMR on these results, Table D-4
presents the BMDs and BMDLs atBMRs of xh SD and 1 SD.
Table D-4. BMDs and BMDLs for effect of PFHxS at age 5 years on anti-
diphtheria antibody concentrations at age 7 years using a BMR of SD change
in log2 (diphtheria antibodies concentration) and a BMR of 1 SD
log2 (diphtheria antibodies concentration)
Estimated without control of PFOS and PFOA
Estimated with control of PFOS and PFOA
BMR
BMD (ng/mL)
P = -1.48 per ng/mL
BMDL (ng/mL)
Plb = -2.47 per ng/mL
BMD (ng/mL)
P = -0.67 per ng/mL
BMDL (ng/mL)
Plb = -2.47 per ng/mL
/2SD
0.500
0.300a
1.100
0.300
1 SD
1.000
0.600
2.200
0.600
aDenotes the selected POD.
The lowest serum PFHxS concentration measured at age 5 years was 0.02 ng/mL, the 5th
percentile was 0.2 ng/mL, and the 10th percentile was 0.3 ng/mL fGrandiean and Bateson. 20211
so the estimated BMDL for a BMR of xh SD (BMDLy2so) in the single-PFAS model is at 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.
The BMDy2 sd estimate from the multi-PFAS models is 2.2-fold higher than the BMDy2 sd
estimate from the model with just PFHxS, but the BMDLy2 sd is the same, which may, or may not,
reflect control for any potential confounding of the regression effect estimates. While it is not clear
which PFAS model provided the "better" estimate of the point estimate of the effect of PFHxS, the
two BMDLy2 sd estimates which serve as the PODs are equal and EPA advanced POD based on
common BMDL estimates of 0.300 ng/mL from both the single-PFAS and the multi-PFAS models
because the uncertainty regarding potential confounding of the BMDL was low. However,
confidence was diminished by the stronger potential confounding in the main effect—even though
the was no observed confounding of the BMDL, and overall confidence in the BMDLs for diphtheria
was judged to be medium/low confidence.
For immunotoxicity related to diphtheria, associated with PFHxS measured at age 5
years, the POD is based on a BMR of V2 SD and a BMDLy2 sd of 0.300 ng/mL.
Modeling Results for Decreased Tetanus Antibody Concentrations at 5 Years of Age and
Perinatal PFHxS
Budtz-l0rgensen and Grandiean f2018a1 fit multivariate models of PFHxS measured
perinatally in maternal serum, against log2-transformed anti-tetanus antibody concentrations
This document is a draft for review purposes only and does not constitute Agency policy.
D-ll DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
1 measured at the 5-year-old examination controlling for sex, and exact age at the 5-year-old
2 examination, cohort, and interaction terms between cohort and sex, and between cohort and age.
3 Models were evaluated with additional control for PFOS (as log2[PF0S]) and PFOA (as log2[PF0A]),
4 and without PFOS and PFOA. Three model shapes of PFHxS were evaluated by Budtz-largerisen and
5 Grandiean f2018al using likelihood ratio tests: a linear model, a piecewise-linear model with a knot
6 at the median, and a logarithmic function. The logarithmic functions did not fit better than the
7 piecewise-linear functions (Budtz-l0rgensen and Grandiean. 2018a). Compared to the linear model,
8 the piecewise-linear model did not fit better than the linear model for either the PFHxS exposure
9 without adjustment for PFOS and PFOA using a likelihood ratio test (p = 0.45; see Budtz-l0rgensen
10 and Grandiean (2018a) Table 3), or for the model that did adjust for PFOS and PFOA (log-j[PFOS]
11 and log2[PFOA]) (p = 0.90).
12 Table D-5 summarizes the results from Budtz-l0rgensen and Grandiean (2018a) for tetanus
13 in this exposure window. These regression coefficients ((3) and their standard errors (SE) were
14 computed by EPA from the published BMDs and BMDL based on a BMR of Vi SD change in
15 log2(tetanus antibody concentrations) in Table 2 of Budtz-largensen and Grandiean (2018a).
Table D-5. Results of the linear analyses of PFHxS measured perinatally and
tetanus antibodies measured at age 5 years from Budtz-Iargensen and
Grandiean f2018a1 in a single-PFAS model and in a multi-PFAS model.
PFOS and
Lower bound
PFOA
Slope (P) per
SE(P)
slope (Plb)
Exposure
Model shape
adjusted
ng/mL
ng/mL
Slope (P) fit
ng/mL
Perinatal PFHxS
Linear
No
-0.0238
0.0183
p = 0.19
-0.0540
Perinatal PFHxS
Linear
Yes
-0.0190
0.0184
p = 0.30
-0.0492
16 Interpretation of results in Table D-5:
17 • PFHxS is a nonsignificant predictor in the single-PFAS model ((3 = -0.0238; p = 0.190).
18
19 • Effects are attenuated when log2 [PFOS] and log2 [PFOA] are included in the model ((3 = -
20 0.019; p = 0.30).
21
22 • The point estimate results for PFHxS are potentially confounded by PFOS and/or PFOA since
23 there was a 20% reduction in the effect size for PFHxS from -0.0238 to -0.0190 when
24 controlling for PFOS and PFOA.
25
26 • One explanation is that PFOS and/or PFOA was a confounder of the PFHxS effect and
27 controlling for those co-exposures removed confounding.
28
29 • Another possibility is that controlling for co-exposures like PFOS and PFOA actually
30 induced confounding (Weisskopf et a!.. 2018; Weisskopf and Webster. 2017).
31
This document is a draft for review purposes only and does not constitute Agency policy.
D-12 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
• The reasons for the change in main effect size for PFHxS are not known. For this
reason, there is uncertainty in knowing which point estimate is the best
representation of any effect of PFHxS.
• However, the lower bound on the point estimates ((3lb) for the single-PFAS model is 35%
lower than the multi-PFAS model estimate for PFHxS.
The definition of the RfD, which is based upon the (3LB, includes allowing for an order of magnitude
(10-fold or 1,000%) uncertainty in the estimate and the uncertainty for potential confounding in
the BMD from including, or excluding, PFOS and PFOAhere is about 20%, while the uncertainty for
potential confounding in the BMDL is about 9%.
Selection of the Benchmark Response
Following the technical guidance flJ.S. EPA. 20121. EPA derived BMDs and BMDLs
associated with a one SD change in the distribution of log2(tetanus antibody concentrations), and Vi
SD change in the distribution of log2(tetanus antibody concentrations). The SD of the log2(tetanus
antibody concentrations) at age 5 years was estimated from two sets of distributional data
presented from two different cohorts of 5-year-olds that were pooled in Budtz-l0rgensen and
Grandjean f2018a1. Grandjean etal. (20121 reported on 587 5-year-olds from the cohort of children
born during 1997-2000 and in Grandjean et al. (20171 reported on 349 5-year-olds from the cohort
of children born during 2007-2009. The means and SDs were computed separately and then pooled
to describe the common SD. The IQR of the tetanus antibody concentrations in the earlier birth
cohort at age 5 years in IU/mL was (0.1, 0.51). Log2-tranforming these values provides the IQR in
log2(IU/mL) as (-3.32, -0.97). Assuming that these log2-transformed values are similar to the
normal distribution, the width of the IQR is approximately 1.35 SDs, thus SD = IQR/1.35, and the SD
of tetanus antibodies inlog2(IU/mL) is (-0.97 - (-3.32))/1.35 = 1.74 log2(IU/mL). The IQR of the
tetanus antibody concentrations in the later birth cohort at age 5 years in IU/mL was (0.1, 0.3).
Log2-tranforming these values provides the IQR in log2(IU/mL) as (-3.32, -1.74), and the SD of
tetanus antibodies in log2(IU/mL) is (-1.74 - (-3.32))/1.35 = 1.17 log2(IU/mL). The pooled variance
is a weighted sum of the independent SDs and the pooled SD was estimated as 1.55 log2(IU/mL).6
To show the impact of the BMR on these results, Table D-6 presents the BMDs and BMDLs atBMRs
ofy2 SD andl SD.
Table D-6. BMDs and BMDLs for effect of PFHxS measured perinatally and
anti-tetanus antibody concentrations at age 5 years.
Estimated without control of PFOS and PFOA
Estimated with control of PFOS and PFOA
BMR
BMD (ng/mL)
P =-0.0238 per ng/mL
BMDL (ng/mL)
(3lb = -0.0540 per ng/mL
BMD (ng/mL)
P = -0.0190 per ng/mL
BMDL (ng/mL)
Plb = -0.049 per ng/mL
6Pooled variance for tetanus in 5-year-olds = [(502-l)(1.74)A2+ (298-l)(1.17)A2]/[502+298-2] = 2.41. The
pooled SD is the square rootof2.41 which is 1.55 log2(IU/mL).
This document is a draft for review purposes only and does not constitute Agency policy.
D-13 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
% SD
32.5
14.4a
40.8
15.7
1 SD
65.1
28.7
81.6
31.5
aDenotes the POD that corresponds to the analyses of PFHxS concentrations perinatally and tetanus antibodies at
age 5 years.
The lowest perinatal maternal serum PFHxS concentration measured was 0.6 ng/mL, the
5th percentile was 1.2 ng/mL, and the 10th% was 1.5 ng/mL (Grandiean and Bateson, 20211 so the
estimated BMDLs for a BMR of xh SD (BMDLy2 sd =14.4 ng/mL) in the single-PFAS model is well
above the 10th% 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. The BMDLy2 sd estimate from the single-PFAS models was 14.4
ng/mL. The BMDL estimates from the multi-PFAS models were about 9% higher than for the single-
PFAS model.
Low confidence in the BMDLs from the PFHxS-only model (14.4 ng/mL) and in the multi-
PFAS model (15.7 ng/mL). Confidence is diminished by the low quality of the model fit for PFHxS in
either model compared with the PFHxS results from tetanus in the 5-year to 7-year exposure-
outcome window of time and there is some uncertainty regarding potential confounding.
For immunotoxicity related to tetanus, associated with PFHxS measured perinatally, the
POD is based on a BMR of Vi SD and a BMDLy2 sd of 14.4 ng/mL. Note that this result is based on a
poorly fit PFHxS regression parameter ((3) estimated as -0.024 per ng/mL (90%CI: -0.054,
0.0.0064; p = 0.19) (Budtz-l0rgensen and Grandiean, 2018b). and thus this POD is identified with
low confidence.
For immunotoxicity related to tetanus associated with PFHxS exposure measured at
age 5 years, the POD estimated for comparison purposes were based on a BMR of SD and a
BMDL.% sd of 14.4 ng/mL.
Modeling Results for Decreased Diphtheria Antibody Concentrations at 5 Years of Age and
Perinatal PFHxS
Budtz-l0rgensen and Grandiean f2018al fit multivariate models of PFHxS measured
perinatally, against log2-transformed anti-diphtheria antibody concentrations measured at the 5
year-old examination controlling for sex and age. Models were evaluated with additional control for
PFOS (as log2[PFOS]) and PFOA (as log2[PFOA]), and without PFOS and PFOA. Three model shapes
were evaluated by Budtz-l0rgensen and Grandiean f2018a1 using likelihood ratio tests: a linear
model of PFHxS, a piecewise-linear model with a knot at the median, and a logarithmic function.
The logarithmic functions did not fit better than the piecewise-linear functions fBudtz-largensen
and Grandiean. 2018al. The piecewise-linear model did not fit better than the linear model for the
PFHxS exposure without adjustment for PFOS and PFOA using a likelihood ratio test (p = 0.70; see
Budtz-]0rgensen and Grandiean f2018al Table 3), or for the model that did adjust for PFOS and
PFOA (log2[PFOS] and log^PFOA]) (p = 0.11). Table D-7 summarizes the results from Budtz-
This document is a draft for review purposes only and does not constitute Agency policy.
D-14 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
1 l0rgensen and Grandjean f2018a1 for diphtheria in this exposure window. These regression
2 coefficients ((3) and their standard errors (SE) were computed by EPA from the published BMDs
3 and BMDL based on a BMR of Vi SD change in log2(diphtheria antibody concentrations) in Table 2
4 of fBudtz-largensen and Grandiean. 2Q18aI
Table D-7. Results of the analyses of PFHxS measured perinatally and
diphtheria antibodies measured at age 5 years from Budtz-Jorgensen Bucltz-
ietrgeiiseii and Grandiean (2018b) in a single-PFAS model and in a multi-PFAS
model.
Exposure
Model shape
PFOS and
PFOA adjusted
Slope (P) per
ng/mL
SE(P)
Slope (P) fit
Lower bound
slope (Plb)
Perinatal PFHxS
Linear
No
-0.0378
0.0193
p = 0.05
-0.0696
Perinatal PFHxS
Linear
Yes
-0.0328
0.0193
p = 0.089
-0.0645
5 Interpretation of results in Table D-7:
6 • PFHxS is a significant predictor in the single-PFAS model ((3 = -0.0378; p = 0.05.)
7 • Effects of PFHxS are attenuated when PFOA and PFOA are in the model ((3 = -0.0328; p =
8 0.09).
9 • Results for PFHxS are potentially confounded by PFOS and/or PFOA since there was a
10 13% change in the effect size for PFHxS from -0.038 to -0.033 when controlling for PFOS
11 and PFOA.
12 ° One explanation is that PFOS and/or PFOA was a confounder of the PFDA effect and
13 controlling for those co-exposures removed confounding.
14 ° Another possibility is that controlling for co-exposures like PFOS and PFOA actually
15 induced confounding fWeisskopf et al, 2018: Weisskopf and Webster. 2017).
16 • The reasons for the change in main effect size for PFDA are not known. For this reason,
17 there is uncertainty in knowing which point estimate is the best representation of any
18 effect of PFDA. However, the lower bounds on the point estimates ((3lb) are similar with
19 the lower bound on the multi-PFAS model effect estimate for PFHxS only 9% lower than
20 the single-PFAS model effect estimate for PFHxS. This small difference suggests very
21 little uncertainty attributable to potential confounding of the lower bound effect
22 estimates.
23 ° The definition of the RfD, which is based upon the (3LB, includes allowing for an
24 order of magnitude (10-fold or 1,000%) uncertainty in the estimate and the
25 uncertainty for potential confounding in the BMD from including, or excluding, PFOS
26 and PFOA here is about 13%, while the uncertainty for potential confounding in the
27 BMDL is about 9%.
This document is a draft for review purposes only and does not constitute Agency policy.
D-15 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Selection of the Benchmark Response
Following the technical guidance fU.S. EPA. 20121. EPA derived BMDs and BMDLs
associated with a one SD change in the distribution of log2(tetanus antibody concentrations) as a
standard reporting level, and Vi SD change in the distribution of log2(tetanus antibody
concentrations). The SD of the log2(diphtheria antibody concentrations) at age 5 years was
estimated from two sets of distributional data presented from two different birth cohorts of 5-year-
olds that were pooled in Budtz-l0rgensen and Grandiean f2018a1. Grandiean etal. (20121 reported
on 587 5-year-olds from the cohort of children born during 1997-2000 and Grandiean et a'
reported on 349 5-year-olds from the cohort of children born during 2007-2009. The means and
SDs were computed separately and then pooled to describe the common SD. The IQR of the
diphtheria antibody concentrations in the earlier birth cohort at age 5 years in IU/mL was (0.05,
0.4). Log2-tranforming these values provides the IQR inlog2(IU/mL) as (-4.32, -1.32). Assuming that
these log2-transformed values are similar to the normal distribution, the width of the IQR is
approximately 1.35 SDs, thus SD = IQR/1.35, and the SD of diphtheria antibodies in log2(IU/mL) is
(-1.32 - (-4.32))/1.35 = 2.22 log2(IU/mL). The IQR of the diphtheria antibody concentrations in the
later birth cohort at age 5 years in IU/mL was (0.1, 0.3). Log2-tranforming these values provides the
IQR inlog2(IU/mL) as (-3.32, -1.74), and the SD of diphtheria antibodies in log2(IU/mL) is (-1.74 - (-
3.32))/1.35 = 1.17 log2(IU/mL). The pooled variance is a weighted sum of the independent SDs and
the pooled SD was estimated as 1.90 log2(IU/mL).7 To show the impact of the BMR on these results,
Table D-8 presents the BMDs and BMDLs atBMRs of xh SD and 1 SD.
Table D-8. BMDs and BMDLs for effect of PFHxS measured perinatally and
anti-diphtheria antibody concentrations at age 5 years
Estimated without control of PFOS and PFOA
Estimated with control of PFOS and PFOA
BMR
BMD (ng/mL)
P =-0.0378 per ng/mL
BMDL (ng/mL)
Plb = -0.0696 per ng/mL
BMD (ng/mL)
P =-0.0328 per ng/mL
BMDL (ng/mL)
Plb = -0.0645 per ng/mL
% SD
25.1
13.T
29.0
14.7
1 SD
50.2
27.3
57.9
29.4
aDenotes the POD that corresponds to the analyses of PFHxS concentrations perinatally and diphtheria antibodies
at age 5 years.
The lowest serum PFHxS concentration measured perinatally was 0.6 ng/mL, the 5th
percentile was 1.2 ng/mL, and the 10th% was 1.5 ng/mL fGrandiean and Bateson. 20211 so the
estimated BMD for a BMR of Vi SD (BMDLy2 sd) in the single-PFAS model is well within the observed
range. 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.
7Pooled variance for tetanus in 5-year-olds = [(502-l)(1.74)A2+ (298-l)(1.17)A2]/[502+298-2] = 2.41. The
pooled SD is the square rootof2.41 which is 1.55 log2(IU/mL).
This document is a draft for review purposes only and does not constitute Agency policy.
D-16 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
The BMD y2 sd estimate from the multi-PFAS models is 15% higher than the BMDy2 sd
estimated from the model with just PFHxS, and the BMDLy2 sd is 8% higher, which may, or may not,
reflect control for any potential confounding of the regression effect estimates. While it is not clear
which estimate is "better," the BMDLs which serve as the PODs are similar (13.7 ng/mL versus 14.7
ng/mL) and EPA advanced the derivation based on results that did not control for PFOS and PFOA
because this model appeared to fit PFHxS well (p = 0.05) and there was low uncertainty due to
potential confounding in either the BMD or the BMDL. Medium confidence in the BMDLs from
PFHxS linear model (13.7 ng/mL) since the model fit reasonably well and these BMDLs do not show
meaningful uncertainty about confounding.
For immunotoxicity related to diphtheria, associated with PFHxS measured at age 5
years, the POD is based on a BMR of V2 SD and a BMDL% SD of 13.7 ng/mL.
Modeling Results for Decreased Birth Weight Using Individual Studies
As noted in Section 5.2.1 five high confidence studies (Manzano-Salgado et al, 2019: Buck
Louis etal, 2018: Shoaff et al. 2018: Starling et al. 2017: Bach etal, 2016a) reported decreased
birth weight in infants whose mothers were exposed to PFHxS. All five studies reported their
exposure metric in units of ng/mL. Three studies reported the (3 coefficients per ln(ng/mL) or per
log2(ng/mL), one study reported a (3 coefficients per ln(l+ng/mL) and one study reported a (3
coefficients per ng/mL, along with 95% confidence intervals (CIs), estimated from linear regression
models. The logarithmic transformation of exposure yields a negative value for low numbers, which
can result in implausible results from dose-response modeling (i.e., estimated risks are negative
and unable to determine the responses at zero exposure). EPA first re-expressed the reported /?
coefficients in terms of per ng/mL, if necessary, according to Dzierlenga et al. (20201. Then EPA
used the re-expressed (3 and lower limit on the CI to estimate BMD and BMDL values using the
general equation y = mx + b, where y is birth weight and x is exposure, substituting the re-
expressed (3 values from these studies for m. The intercept b represents the baseline value of birth
weight in an unexposed population and it can be estimated through y = mx + b using an average
birth weight from an external population as y, an average exposure as x and re-expressed (3 from
the studies as m.
The CDC Wonder site (https: //wonder.cdc.gov/natalitv.html) provides vital statistics for
babies born in the United States. There were 3,791,712 all live births in the United States in 2018
according to final natality data. The mean and standard deviation of birth weight were 3261.6 ±
590.7 g (7.19 ± 1.30 lb), with 8.27% of live births falling below the public health definition of low
birth weight (i.e., <2,500 g, or 5.5 lb). The full natality data for the U.S. data on birth weight was
used as it is more relevant for deriving toxicity values for the U.S. general population than the
study-specific birthweight data. Also, the CDC Wonder database may be queried to find the exact
percentage of the population falling below the cut-off value for clinical adversity. America's
Children and the Environment (ACE) Biomonitoring on Perfluorochemicals report
This document is a draft for review purposes only and does not constitute Agency policy.
D-17 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
fhttpsi//www.epa.gov/americaschildrenenvironment/ace-biomonitoring~perfIuorochemicals-
pfcs#B61 provides the median blood serum levels of PFHxS of 0.6 ng/mL in 2015-2016 in women
aged 16 to 49, using National Health and Nutrition Examination Survey (NHANES) as data source.
These values are assumed to be representative of women of reproductive age and are subsequently
used in the estimation of BMD and BMDL values from the available five epidemiological studies.
Buck Louis etal. (20181 reported a (3 coefficient of-17.1 g (95%CI: -40.7, 6.5) per 1 SD
increase of ln(l+ng/mL), corresponding to a rescaled (3 coefficient of -53.1 g (95%CI: -126.4, 20.2)
per ln(l+ng/mL) increase, for the association between birth weight and maternal PFHxS serum
concentrations (collected during 10 weeks to 13 weeks and 6 days of pregnancy with a median of
12 weeks) in a United States cohort, based on their multiple linear regression (y = m* ln(l + x) +
b) analysis. The reported (3 coefficient can be re-expressed in terms of per ng/mL according to
Dzierlenga et at. f20201. Given the median (0.71 ng/mL) and IQR (0.44-1.23 ng/mL) of the exposure
from Buck Louis etal. (20181. EPA estimated the distribution of exposure by assuming the exposure
follows a log-normal distribution and the natural logarithm of exposure is normally distributed
with mean and standard deviation as:
Then, EPA estimated the 25th - 75th percentiles at 10 percentile intervals of the exposure
distribution and corresponding responses of reported (3 coefficient. The re-expressed (3 coefficient
is determined by minimizing the sum of squared differences between the curves generated by the
re-expressed (3 and the reported p. This resulted in a re-expressed (3 coefficient of-29.9 g (95%CI: -
71.1,11.4) per ng/mL.
Typically, for continuous data, the preferred definition of the benchmark response (BMR) is
to have a basis for what constitutes a minimal level of change in the endpointthat is biologically
significant For birth weight, there is no accepted percent change that is considered adverse.
However, there is a clinical measure for what constitutes an adverse response: babies born
weighing less than 2,500 g are considered to have low birth weight, and further, low birth weight is
associated with a wide range of health conditions throughout life fTian et al. 2019: Reves and
Manalich. 2005: Hack et al.. 19951. Given this clinical cut-off for adversity and that 8.27% of all live
births in the United States in 2018 fell below this cut-off, the hybrid approach can be used to define
the BMR. The hybrid approach is advantageous in that it harmonizes the definition of the BMR for
H = ln(q50) = ln( 0.71) = —0.35
(1)
a = ln(q75/q25)/1.349 = Zn(1.23/0.44)/1.349 = 0.75
(2)
This document is a draft for review purposes only and does not constitute Agency policy.
D-18 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
continuous data with that for dichotomous data8. Essentially, the hybrid approach involves the
estimation of the dose that increases the percentile of responses falling below (or above) some cut-
off for adversity in the tail of the response distribution. Application of the hybrid approach requires
the selection of an extra risk value for BMD estimation. In the case of birth weight, an extra risk of
5% is selected given that this level of response is typically used when modeling developmental
responses from toxicology studies, and that low birthweight confers increased risk for adverse
health effects throughout life, thus supporting a BMR lower than the standard BMR of 10% extra
risk.
Therefore, given a background response and a BMR = 5% extra risk, the BMD would be the
dose thatresults in 12.86% ofthe responses falling below the 2500 g cut-off value:
Extra Risk(ER) = (P(d) - P(0)) / (1 - P(0))
P(d) = ER(1 - P(0)) + P(0) = 0.05(1 - 0.0827) + 0.0827 = 0.1286
Using the mean birth weight for all births in the United States of 3,261.6 g with a standard
deviation of 590.7 g, EPA calculated the mean response that would be associated with the 12.86th
percentile of the distribution falling below 2,500 g. In this case, the mean birth weight would be
3169.2g. Given the median exposure of 0.60 ng/mL from ACE Biomonitoring on Perfluorochemicals
as x, the mean birth weight in the United States as y and the re-expressed (3 as m term, the
intercept b can be estimated as:
b = y - mx = 3261.6 g - (-29.9 gg)"1) 0.60^ = 3279.6 g (3)
The BMD was calculated by rearranging the equation y = mx + b and solving for x, using
3279.6 g for the b term and -29.9 for the m term. This resulted in a value of 3.70 ng/mL:
nq „
x = (y — b)/m = (3169.2 g - 3279.6 #)/(-29.9 ^(^-)_1) = 3.70 ng/mL
mL
To calculate the BMDL, the method is essentially the same except that the lower limit (LL)
on the p coefficient (-71.1) is used for the m term. However, fBuck Louis etal. 20181 reports a two-
sided 95%CI for the (3 coefficient, meaning that the lower limit of that CI corresponds to a 97.5%
one-sided lower limit. The BMDL is defined as the 95% lower limit of the BMD (i.e., corresponds to
a two-sided 90%CI), so the corresponding lower limit on the (3 coefficient needs to be calculated
before calculating the BMDL. First, the standard error of the (3 coefficient can be calculated as:
8While the explicit application of the hybrid approach is not commonly used in IRIS
dose/concentration/exposure-response analyses, the more commonly used SD-definition ofthe BMR for
continuous data is simply one specific application of the hybrid approach. The SD-definition of the BMR
assumes that the cut-off for adversity is the 1.4th percentile of a normally distributed response and that
shifting the mean of that distribution by one standard deviation approximates an extra risk of 10%.
This document is a draft for review purposes only and does not constitute Agency policy.
D-19 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
SE =
Upper Limit — Lower Limit S^mf) 1 ( 71.1 g(^jj) X)
3.92
3.92
Then the corresponding 95% one-sided lower bound on the (3 coefficient can be calculated
as:
Using this value for the m term results in a BMDL value of 1.71 ng/mL maternal serum
concentration.
Shoaff et al. (20181 reported a (3 coefficient of-13.4 g (95%CI: -35.9, 9.1) per ng/mL
increase for the association between birth weight and maternal PFHxS serum concentrations
(collected during 16 weeks of pregnancy to delivery with a median of 16 weeks) in a United States
cohort A BMD of 7.50 ng/mL was calculated from Shoaff et al. f 20181 using the same approach as
above with the same values for the mean birth weight in the United States and the reported (3
coefficient directly without re-expression.
To calculate the BMDL, the same procedure as above is used to calculate the 95% one-sided
lower limit for the reported (3 coefficient from the reported lower limit on the 95% two-sided CI of -
35.9 g per ng/mL. Using the corresponding lower limit (-32.3 gper ng/mL), a BMDL of 3.12 ng/mL
is calculated.
Starling et al. (20171 reported a (3 coefficient of -13.5 g (95%CI: -50.7, 23.7) per ln(ng/mL)
for the association between birth weight and maternal PFHxS serum concentrations (collected
during 20 to 34 weeks of pregnancy with a median of 27 weeks) in a United States cohort Given
median (0.8 ng/mL) and IQR (0.5-1.2 ng/mL) of the exposure, EPA estimated the mean (-0.22) and
standard deviation (0.65) of the natural logarithm of exposure. The re-expressed (3 coefficient is -
16.2 g (95%CI: -60.7, 28.4) per ng/mL and the intercept b is 3271.3 g. The 95% one-sided lower
limits for the re-expressed (3 coefficient is -53.6 g per ng/mL. The values of the BMD and BMDL are
6.32 ng/mL and 1.91 ng/mL, respectively.
Manzano-Salgado (2019) reported a (3 coefficient of-8.6 g (95%CI: -32.0, 14.8) per log2
(ng/mL) for the association between birth weight and maternal PFHxS serum concentrations
(collected during the first trimester of pregnancy with a mean of 12.3 weeks) in a Spanish cohort
Given the median (0.58 ng/mL) and SD (0.37 ng/mL) of the exposure, EPA estimated the mean (-
0.72) and standard deviation (0.58) of the natural logarithm of exposure. The re-expressed (3
coefficient is -24.5 g (95%CI: -91.2, 42.2) per ng/mL and the intercept is 3276.3 g. The 95% one-
sided lower limits for the re-expressed (3 coefficient is -80.5 g per ng/mL. The values of the BMD
and BMDL are 4.37 ng/mL and 1.33 ng/mL, respectively.
95% one - sided LL = p - 1.645(SE(/3)) = -29.9 g^Y1 - 1.645 (21.0 g^y1)
mL v rriL '
nq ,
= -MA9&1
This document is a draft for review purposes only and does not constitute Agency policy.
D-20 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Bach (2016b) reported a (3 coefficient of-11.0 g (95%CI: -32.0, 9.0) per IQR increase in
PFHxS (ng/mL), corresponding to a rescaled (3 coefficient of -40.7 g (95%CI: -118.5, 33.3) per
ng/mL increase, for the association between birth weight and maternal PFHxS serum
concentrations (collected 9-20 weeks of pregnancy, 96% within 13 weeks) in a Danish cohort. The
BMD of 2.87 ng/mL and BMDL 1.12 ng/mL was calculated using the rescaled (3 coefficient directly
without re-expression.
For all of the above calculations, EPA used the exact percentage (8.27%) of live births in the
United States in 2018 that fell below the cut-off of 2,500 g as the tail probability to represent the
probability of extreme ("adverse") response at zero dose (P(0)). However, this exact percentage of
8.27% was calculated without accounting for the existence of background PFHxS exposure in the
United States population (i.e., 8.27% is not the tail probability of extreme response at zero dose).
Thus, EPA considers an alternative control-group response distribution erc)), using the
study-specific intercept b obtained through equation (3) (representing the baseline value of birth
weight in an unexposed population) as and the standard deviation of U.S. population as ac, to
estimate the tail probability that fell below the cut-off of 2500 g. EPA estimated the study-specific
tail probability of live births falling below the public health definition of low birth weight (2,500 g)
as:
^ r 2500 , (x-b)2 ^ <*2500 (x-b)2
P( 0) = -= I e 2°c dx = -= I e 2*590.7 2'dx
acy/2nJ-oo 590.7V2tt J-co
_ _ ng
b y mx 3261.6 [fire—expressed * 0.60
In this alternative approach, P(0) is 9.86% if there is no background exposure (x = 0). By
using the median of serum PFHxS concentrations (0.60 ng/mL) from ACE Biomonitoring on
Perfluorochemicals as background exposure (x), the tail probabilities using this alternative
approach were study-specific and ranged from 9.35% - 9.65%. As such, the results from this
alternative approach, presented under the column of "Alternative Tail Probability" in Table D-9, are
very similar to the main results, presented under the column of "Exact Percentage" in Table D-9,
when background exposure was not accounted for while estimating the tail probability.
Table D-9 presents the BMDs and BMDLs for all individual studies considered for POD
derivation, with and without accounting for background exposure while estimating the percentage
of the population falling below the cut-off value. The BMDLs across the studies and approaches
ranged from 1.12 ng/mL to 4.20 ng/mL.
This document is a draft for review purposes only and does not constitute Agency policy.
D-21 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table D-9. BMDs and BMDLs for effect of PFHxS on decreased birth weight, by using percentage (8.27%) of live
births falling below the public health definition of low birth weight, or alternative study-specific tail probability
Study
Exposure
median (IQR
or 33-67QR or
SD)
Exposure
distribution
(A*, tf)
Reported (3
(95%CI)
Re-expressed (3
(95%CI)
g/ng/mL
Intercept
b
SE of
P
95%
one-
sided LL
of p
Exact percentage
(P(0)=8.27%)
Alternative tail probability3
BMD
(ng/
mL)
BMDL
(ng/mL)
P(0)
BMD
(ng/mL)
BMDL
(ng/mL)
(Bach et al.,
2016b)
0.47 (0.36-
0.63)
(-0.76, 0.41)
-11.0 (-32.0, 9.0)
g/IQR(ng/mL)
-40.7 (-118.5, 33.3)
3,286.1
38.74
-104.5
2.87
1.12b
9.16
%
3.44
1.34
(Buck Louis et
al., 2018)
0.71 (0.44-
1.23)
(-0.35, 0.75)
-17.1 (-40.7, 6.5)
g/SD
(ln(l+ng/ml))
-29.9 (-71.1, 11.4)
3,279.6
21.0
-64.4
3.70
1.71
9.35
%
4.63
2.14
(Manzano-
Salgado et al.,
2019)
0.58 (0.37)
(-0.72, 0.58)
-8.6 (-32.0, 14.8)
g/log2(ng/ml)
-24.5 (-91.2, 42.2)
3,276.3
34.0
-80.5
4.37
1.33
9.44
%
5.60
1.71
(Shoaff et al.,
2018)
1.5 (1.0-2.0)
(0.41, 0.79)
-13.4 (-35.9, 9.1)
g/ng/ml
-13.4 (-35.9, 9.1)
3,269.7
11.5
-32.3
7.50
3.12
9.63
%
10.11
4.20
(Starling et
al., 2017)
0.8 (0.5-1.2)
(-0.22, 0.65)
-13.5 (-50.7,
23.7)
g/ln(ng/ml)
-16.2 (-60.7, 28.4)
3,271.3
22.7
-53.6
6.32
1.91
9.58
%
8.41
2.54
aThe alternative study-specific tail probability of live births falling below the public health definition of low birth weight based on normal distribution with
intercept b as mean and standard deviation of 590.7 based on U.S. population.
Smallest BMDL using the five individual studies.
This document is a draft for review purposes only and does not constitute Agency policy.
D-22 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
ACE Biomonitoring on Perfluorochemicals also provides the median blood serum levels of
PFHxS among women ages 16 to 49 in 1999-2000 (1.3 ng/mL), in 2003-2004 (1.4 ng/mL) and in
2011-2012 (0.8 ng/mL). PA performed a sensitivity analysis by estimating BMD and BMDL using
these values as background exposures. The results for Bach etal. f2016bl presented in Table D-10,
demonstrate the robustness of EPA's approaches with alternative assumptions on background
exposures.
Table D-10. BMDs and BMDLs for effect of PFHxS on decreased birth weight by
background exposure, using the exact percentage of the population (8.27%) of
live births falling below the public health definition of low birth weight, or
alternative study-specific tail
Study
Background
exposurea
Intercept
b
Exact percentage
(P(0)= 8.27%)
Alternative tail probabilityb
BMD
(ng/mL)
BMDL
(ng/mL)
P(0)
BMD
(ng/mL)
BMDL
(ng/mL)
(Bach et
al., 2016b)
0.60
3286.1
2.87
1.12
9.16%
3.44
1.34
0.80
3294.2
3.07
1.20
8.94%
3.50
1.36
1.30
3314.6
3.57
1.39
8.39%
3.65
1.42
1.40
3318.7
3.67
1.43
8.29%
3.68
1.43
aAssumptions on background exposure for the estimation of intercept using Equation (3).
bThe tail probability of live births falling below the public health definition of low birth weight based on normal
distribution.
Uncertainty may be introduced by the re-expression of regression coefficients Linakis et al..
A sensitivity analysis was performed to compare BMD and BMDL with and without re-expression of
P coefficients using Buck Louis etal. (20181. Buck Louis et al. (20181 reported a (3 coefficient of -
17.1 g (95%CI: -40.7, 6.5) per 1 SD increase of ln(l+ng/mL), corresponding to a rescaled (3
coefficient of-53.1 g (95%CI: -126.4, 20.2) per ln(l+ng/mL) increase. The BMD of 8.12 ng/mL and
BMDL 1.79 ng/mL was calculated using the general equation y = m* ln(l + x) + b and the
rescaled (3 coefficient per ln(l+ng/mL), while assuming the median blood serum levels of PFHxS of
0.60 ng/mL. This approach removed any uncertainty associated with the re-expression of
regression coefficients in the modeling. Table D-ll shows the BMD/BMDL results at several
background exposure levels using re-expressed (3 coefficient (g/ng/mL) or reported/rescaled (3
coefficient (g/ln(l+ng/mL)) for Buck Louis et al. (20181.
This document is a draft for review purposes only and does not constitute Agency policy.
D-23 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table D-ll. BMDs and BMDLs for effect of PFHxS on decreased birth weight by
background exposure, using the exact percentage of the population (8.27%) of
live births falling below the public health definition of low birth weight, with
re-expressed p coefficient (g per ln(l+ng/mL))
Study
Background
exposurea
Intercept
b
Re-expressed (3
g/ng/mL
Reported/resca led (3
g/ In (1+ ng/mL)
BMD
(ng/mL)
BMDL
(ng/mL)
BMD
(ng/mL)
BMDL
(ng/mL)
(Buck Louis et
al., 2018)
0.60
3,279.6
3.70
1.71
8.12
1.79
0.80
3,285.5
3.90
1.80
9.26
1.94
1.30
3,300.5
4.40
2.04
12.11
2.30
1.40
3,303.4
4.50
2.08
12.68
2.36
aAssumptions on background exposure for the estimation of intercept using Equation (3).
Modeling Results for Decreased Birth Weight Using Meta-analysis Results
In addition to the above five studies, epidemiologic data were also available on another 22
studies with different reported units of (3 coefficient for the association between birth weight and
PFHxS concentrations as discussed in the Meta-Analysis Method section (see Appendix C). As noted
above, EPA was able to convert the exposure-response functions quantifying the effects for these
studies based on different units into natural log units (i.e., per ln(ng/mL)) according to Dzierlenga
etal. (20201. Two studies, Lind and Ashley-Martin etal. (20171 only reported separate
estimates for boys and girls; before performing the overall meta-analysis, these estimates were
pooled using inverse-variance weighting. The study by Maisonet et al. (2012) and Marks et al.
(2019a) only reported sex-specific estimates for girls and boys from the same population. These
two studies were also pooled to obtain an effect estimate in the overall population and included in
the meta-analysis as Maisonet et al. (20121. Meta-analyses were performed using (3 coefficient per
ln(ng/mL) of all 27 studies, since the majority of the studies reported results on log scale.
Additionally, analyses were performed using subsets of the studies to evaluate whether the
summary effect estimate varied by study confidence or by the timing of maternal serum sampling.
The results were presented in Table D-12.
The meta-analysis conducted using (3 coefficient per ln(ng/mL) for all studies (n = 27)
resulted in a (3 coefficient of-7.7 g (95%CI: -14.8, -0.5) mean birth weight per ln(ng/mL) PFHxS
increase based on a random effect model with inverse-variance weights. This (3 coefficient can be
re-expressed in terms of per ng/mL according to Dzierlenga etal. (20201. First, the distribution of
exposure for each individual study was estimated by assuming the exposure follows a log-normal
distribution. One hundred replicates of random samples (sample size was the same as the reported
sample size in each study) were then simulated from the exposure distributions for each study
included in the meta-analysis, and random samples from all studies were pooled for each replicate
to get quantiles from the pooled random samples for each replicate. Lastly, the mean quantiles
(median and IQR) from the 100 replicates were used to obtain the exposure distribution for all
This document is a draft for review purposes only and does not constitute Agency policy.
D-24 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
studies using Equation (1) and (2), since the joint distribution of the exposures are also log
normally distributed. The re-expressed (3 coefficient is -7.7 g (95%CI: -14.9, -0.5) per ng/mL.
The BMD of 12.55 ng/mL from all studies can be calculated using the same approach as
above with the same values for the mean birth weight in the United States. To calculate the BMDL,
the same procedure as above was used to calculate the 95% one-sided lower limit for the re-
expressed p coefficient Using the one-sided lower limit, a BMDL of 7.05 ng/mL is calculated.
The BMD and BMDL for the effect of PFHxS on decreased birth weight using meta-analysis
results, conducted in log scale overall, and stratified by study confidence and by sample timing, are
presented in Table D-12 below. As shown in Table D-12 (and Appendix C), the overall combined (3
coefficient of-7.7 g (95%CI: -14.8, -0.5) per ln(ng/mL) increase was robust and very comparable to
that seen for only the twelve high studies (-6.8 g; 95%CI: -16.3, 2.8) or the 23 medium and high
studies combined (-8.0 g; 95%CI: -15.2, -0.7). Similarly, the BMDLs for the earlier sampled study
subsets (6.34) were very comparable to the overall full set of studies (7.05).
EPA also conducted the analysis with the alternative approach discussed above by
considering an alternative control-group response distribution crc)). The results from this
alternative approach, presented in Table D-13 below, are very similar to the previous results.
Table D-12. BMDs and BMDLs for effect of PFHxS on decreased birth weight
using meta-analysis results conducted in log scale overall, by study confidence
and by sample timing, using the percentage (8.27%) of live births falling
below the public health definition of low birth weight
Set of studies
Meta-analysis in log scale
P per ln(ng/mL)
(95%CI)
Re-expressed (3 per
ng/mL (95%CI)
BMD
(ng/mL)
BMDL
(ng/mL)
All studies (n = 27)
-7.7 (-14.8, -0.5)
-7.7 (-14.9, -0.5)
12.55
7.05
Study Confidence
High (n = 12)
-6.8 (-16.3, 2.8)
-7.3 (-17.5, 3.0)
13.26
6.09
Medium (n = 11)
-9.6 (-20.8, 1.6)
-8.5 (-18.4, 1.4)
11.50
5.81
Low( n = 4)
-1.5 (-51.6, 48.7)
-3.7 (-127.1, 120.0)
25.61
0.88
High + Medium (n = 23)
-8.0 (-15.2,-0.7)
-7.9 (-15.0, -0.7)
12.32
7.00
Sample Timinga
Early Pregnancy (n = 12)
-7.3 (-16.0, 1.4)
-7.7 (-16.8, 1.5)
12.68
6.34
Late Pregnancy (n = 10)
-3.9 (-17.7, 9.9)
-4.1 (-18.6, 10.4)
23.10
5.82
Post Pregnancy (n = 5)
-28.3 (-69.3, 12.7)
-11.7 (-28.7, 5.3)
8.48
3.83
Late + Post Pregnancy (n = 15)
-8.5 (-21.0, 4.1)
-7.3 (-18.1, 3.5)
13.19
5.89
aSample time periods include early pregnancy (the 1st trimester, 1st or 2nd trimester), late pregnancy
(2nd trimester, 2nd, or 3rd trimester), post pregnancy (birth and post-birth); n = number of studies;
effect estimates, p, represent change in birthweight (grams) per unit change in In (ng/mL) or ng/mL PFHxS
exposure; CI = confidence interval.
This document is a draft for review purposes only and does not constitute Agency policy.
D-25 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table D-13. BMDs and BMDLs for effect of PFHxS on decreased birth weight
using meta-analysis results conducted in log scale overall, by study confidence
and by sample timing, using the alternative study-specific tail probability of
live births falling below the public health definition of low birth weight
Set of studies
Meta-analysis in log scale
P per ln(ng/mL)
(95%CI)
Re-expressed (3 per
ng/mL (95%CI)
BMD
(ng/mL)
BMDL
(ng/mL)
All studies (n = 27)
-7.7 (-14.8, -0.5)
-7.7 (-14.9, -0.5)
17.38
9.77
Study Confidence
High (n = 12)
-6.8 (-16.3, 2.8)
-7.3 (-17.5, 3.0)
18.41
8.45
Medium (n = 11)
-9.6 (-20.8, 1.6)
-8.5 (-18.4, 1.4)
15.88
8.02
Low (n = 4)
-1.5 (-51.6, 48.7)
-3.7 (-127.1, 120.0)
36.21
1.25
High + Medium (n = 23)
-8.0 (-15.2, -0.7)
-7.9 (-15.0, -0.7)
17.06
9.69
Sample Timinga
Early Pregnancy (n = 12)
-7.3 (-16.0, 1.4)
-7.7 (-16.8, 1.5)
17.58
8.79
Late Pregnancy (n = 10)
-3.9 (-17.7, 9.9)
-4.1 (-18.6, 10.4)
32.59
8.21
Post Pregnancy (n = 5)
-28.3 (-69.3, 12.7)
-11.7 (-28.7, 5.3)
11.52
5.20
Late + Post Pregnancy (n = 15)
-8.5 (-21.0, 4.1)
-7.3 (-18.1, 3.5)
18.31
8.18
aSample time periods include early pregnancy (the 1st trimester, 1st or 2nd trimester), late pregnancy
(2nd trimester, 2nd or 3rd trimester), post pregnancy (birth and post-birth); n = number of studies;
effect estimates, p, represent change in birthweight (grams) per unit change in In (ng/mL) or ng/mL PFHxS
exposure; CI = confidence interval.
Summary of modeling results for decreased birth weight
For decreased birth weight associated with PFHxS exposure, the POD selected from the
available epidemiologic literature (considering both individual studies and the results of meta-
analyses using either high and medium confidence studies or focusing on early trimester sample
timing) is 1.12 ng/mL maternal serum concentration, based on birth weight data from Bach et al.
f2016al. The PODs from the meta-analyses of high, medium, or early sampling time studies were
higher than the PODs from individual study PODs, and thus were not considered health-protective
and were not considered further for POD selection. Of the five individual studies, fBuck Louis etal.
2018: Manzano-Salgado et al. 2017: Bach et al. 2016a) assessed maternal PFHxS serum
concentrations either exclusively or predominately in the first trimester, minimizing concerns
surrounding bias due to p re g na n cy- re 1 a te d hemodynamic effects. Additionally, use of the Bach et al.
(2016a) results expressed in natural scale eliminated uncertainty associated with the re-expression
of regression coefficients. Therefore, the POD from Bach et al. (2016a). which also was the lowest,
was ultimately selected.
This document is a draft for review purposes only and does not constitute Agency policy.
D-26 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
D.2. BENCHMARK DOSE MODELING RESULTS FROM ANIMAL STUDIES
D.2.1. Benchmark Dose Modeling Approaches
The endpoints selected for benchmark dose (BMD) modeling are listed in Table D-14. The
animal doses in the study were used in the BMD modeling and then converted to human equivalent
doses (HEDs) using the PK model described in Section 3.1 of the main document; the BMD modeling
results are presented in this appendix.
Modeling Procedure for Dichotomous Noncancer Data
BMD modeling of dichotomous noncancer data was conducted using EPA's Benchmark Dose
Software (BMDS, version 3.2). For these data, the Gamma, Logistic, Log-Logistic, Log-Probit,
Multistage, Probit, Weibull, and Dichotomous Hill models available within the software were fit
using a benchmark response (BMR) of 10% extra risk. The Multistage model is run for all
polynomial degrees up to n - 2, where n is the number of dose groups including control. Adequacy
of model fit was judged on the basis ofx2 goodness -of-fitp-value (p > 0.1), scaled residuals at the
data point (except the control) closest to the predefined benchmark response (absolute value <2.0),
and visual inspection of the model fit Among all models providing adequate fit, the benchmark dose
lower confidence limit (BMDL) from the model with the lowest Akaike's information criterion (AIC)
was selected as a potential POD when BMDL values were sufficiently close (within threefold).
Otherwise, the lowest BMDL was selected as a potential POD.
Modeling Procedure for Continuous Noncancer Data
BMD modeling of continuous noncancer data was conducted using EPA's Benchmark Dose
Software (BMDS, version 3.2). For these data, the Exponential, Hill, Polynomial, and Power models
available within the software are fit using a BMR of 1 standard deviation (SD) when no toxicological
information was available to determine an adverse level of response. When toxicological
information was available, the BMR was based on relative deviation, as outlined in the Benchmark
Dose Technical Guidance (U.S. EPA. 2012). An adequate fit is judged on the basis of x2 goodness of
fit p-value (p > 0.1), scaled residuals at the data point (except the control) closest to the predefined
benchmark response (absolute value <2.0), and visual inspection of the model fit In addition to
these three criteria for judging adequacy of model fit, a determination is made on whether the
variance across dose groups is homogeneous. If a homogeneous variance model is deemed
appropriate based on the statistical test provided by BMDS (i.e., Test 2), the final BMD results are
estimated from a homogeneous variance model. If the test for homogeneity of variance is rejected
(p < 0.05), the model is run again, while modeling the variance as a power function of the mean to
account for this nonhomogeneous variance.
Among all models providing adequate fit, the BMDL from the model with the lowest AIC was
selected as a potential POD when BMDL estimates differed by less than threefold. When BMDL
This document is a draft for review purposes only and does not constitute Agency policy.
D-27 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
1 estimates differed by greater than threefold, the model with the lowest BMDL was selected to
2 account for model uncertainty.
3 Data Used for Modeling
4 The source of the data used for modeling endpoints from animal studies is provided in
5 Table D-14. These data also are included in full in the tables below.
Table D-14. Sources of data used in benchmark dose modeling of PFHxS
endpoints from animal studies
Endpoint/Reference
Reference
HAWC link
Endocrine effects
4/ T4 Total - M
NTP (20181
https
//hawcprd,epa.gov/ani/endpoint/100S08242/
4/ T3 - M
NTP (20181
https
//hawcprd,epa.gov/ani/endpoint/100S08240/
4/ T3 - C
F1 PND 16
Ramh0i et al. (2020)
https
//hawc, epa.gov/ani/endpoint/100515830/
4, T4 Free - M
F1 (PND16)
Ramh0i et al. (2018)
RANGE FINDING
https://hawcprd.epa.gov/ani/endpoint/100508572/
4/ T4 Free - C
F1 (PND16)
Ramh0i et al. (2018)
RANGE FINDING
https://hawcprd.epa.gov/ani/endpoint/100508582/
4/ T4 Free - C
F1 (PND22)
Ramh0i et al. (2018)
MULTI-GEN
https://hawcprd.epa.gov/ani/endpoint/100508636/
4/ T4 Free - C
F1 (PND16)
Ramh0i et al. (2018)
MULTI-GEN
https://hawcprd.epa.gov/ani/endpoint/100508634/
T* Thyroid
hypertrophy/hypoplasia-
M F0 (44day)
Butenhoff et al. (2009)
https://hawcprd.epa.gov/ani/endpoint/100507599/
Decreased free T4 — male rats NTP (20181
Table D-15. Dose-response data for decreased free T4 in male rats NTP (2018)
Dose (mg/kg-d)
n
Mean
(ng/dL)
SD
0
10
1.737
0.1
0.625
10
0.817
0.067
1.25
10
0.481
0.031
2.5
10
0.357
0.022
5
10
0.386
0.032
10
10
0.385
0.027
This document is a draft for review purposes only and does not constitute Agency policy.
D-28 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table D-16. Benchmark dose results for decreased free T4 in male rats— non-
constant variance, BMR = 1 standard deviation
Models3
Test 3
(p value)
1 standard
deviation
Goodness
of fit
(p-Value)
AIC
BMDS
classification13
BMDS notes
BMD
BMDL
Exponential 2
(NCV—
normal)
0.7202
0.0623
0.0470
<0.0001
-21.2050
Questionable
Goodness of fit p-value < 0.1
BMD 3x lower than lowest non-zero dose
BMDL 3x lower than lowest non-zero dose
BMD 10x lower than lowest non-zero dose
BMDL 10x lower than lowest non-zero dose
Exponential 3
(NCV—
normal)
0.7202
0.0623
0.0470
<0.0001
-21.2050
Questionable
Goodness of fit p-value < 0.1
BMD 3x lower than lowest non-zero dose
BMDL 3x lower than lowest non-zero dose
BMD 10x lower than lowest non-zero dose
BMDL 10x lower than lowest non-zero dose
Exponential 4
(NCV—
normal)
0.7202
0.0425
0.0307
0.0002
-191.1842
Questionable
Goodness of fit p-value < 0.1
BMD 3x lower than lowest non-zero dose
BMDL 3x lower than lowest non-zero dose
BMD 10x lower than lowest non-zero dose
BMDL 10x lower than lowest non-zero dose
Exponential 5
(NCV—
normal)
0.7202
0.0707
0.0475
0.0096
-199.6740
Questionable
Goodness of fit p-value < 0.1
BMD 3x lower than lowest non-zero dose
BMDL 3x lower than lowest non-zero dose
BMDL 10x lower than lowest non-zero dose
Hill (NCV—
normal)
0.7202
0.1890
0.1498
0.0003
-192.8821
Questionable
Goodness of fit p-value < 0.1
BMD 3x lower than lowest non-zero dose
BMDL 3x lower than lowest non-zero dose
Polynomial
(5 degree)
(NCV—
normal)
0.7202
80.468
74.3240
<0.0001
-2.4126
Questionable
Goodness of fit p-value < 0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >| 1.51
actual response std. dev.
Polynomial
(4 degree)
(NCV—
normal)
0.7202
80.4680
24.7291
<0.0001
-2.4126
Questionable
Goodness of fit p-value < 0.1
BMD/BMDL ratio > 3
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >| 1.51
actual response std. dev.
Polynomial
(3 degree)
(NCV—
normal)
0.7202
80.4680
27.7219
<0.0001
-2.4126
Questionable
Goodness of fit p-value < 0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >| 1.51
actual response std. dev.
Polynomial
(2 degree)
(NCV—
normal)
0.7202
80.4680
27.8561
<0.0001
-2.4126
Questionable
Goodness of fit p-value < 0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >| 1.51
actual response std. dev.
Power (NCV—
0.7202
80.4680
54.9996
<0.0001
-2.4126
Questionable
Goodness of fit p-value < 0.1
This document is a draft for review purposes only and does not constitute Agency policy.
D-29 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Models3
Test 3
(p value)
1 standard
deviation
Goodness
of fit
(p-Value)
AIC
BMDS
classification13
BMDS notes
BMD
BMDL
normal)
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >| 1.51
actual response std. dev.
Linear
(NCV—
normal)
0.7202
80.4680
30.5047
<0.0001
-2.4126
Questionable
Goodness of fit p-value < 0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >| 1.51
actual response std. dev.
aNon-constant models failed to model the data.
^'Classification" column denotes whether a model can be considered for model selection purposes. See BMDS
User Guide: https://www.epa.gov/bmds.
NTP 2018 Male Rat Free T4 Data
2 -
0.2
0
0123456789 10
Dose
Figure D-3. Dose response data for male rat free T4 NTP (2018). X-axis is dose
(mg/kg-d) and y-axis is level of free T4 (ng/dL).
Decreased total T4 — male rats fNTP. 20181
Table D-17. Dose-response data for decreased T4 in male rats NTP (2018)
Dose (mg/kg-d)
n
Mean
(Hg/dL)
SD
0
10
4.24
0.229
0.625
10
2.39
0.078
1.25
10
1.7
0.058
This document is a draft for review purposes only and does not constitute Agency policy.
D-30 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Dose (mg/kg-d)
n
Mean
(Hg/dL)
SD
2.5
10
1.47
0.07
5
10
1.54
0.093
10
10
1.66
0.048
This document is a draft for review purposes only and does not constitute Agency policy.
D-31 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table D-18. Benchmark dose results for decreased Total T4 in male rats— non-constant variance, BMR = 1
standard deviation
Models
Test 3
(P-
Value)
1 standard deviation
Goodness of fit
(p-value)
AIC
BMDS
classification'5
BMDS notes
BMD
BMDL
Exponential 2 (NCV—normal)
0.0694
25.9593
20.9559
<0.0001
92.8124
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >11.51 actual
response std. dev.
Exponential 3 (NCV—normal)
0.0694
25.9589
20.9558
<0.0001
92.8124
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >11.51 actual
response std. dev.
Exponential 4 (NCV—normal)
0.0694
0.0410
0.0302
<0.0001
-69.8120
Questionable
Goodness of fit p-value <0.1
BMD 3x lower than lowest non-zero dose
BMDL 3x lower than lowest non-zero dose
BMD 10x lower than lowest non-zero dose
BMDL 10x lower than lowest non-zero dose
Exponential 5 (NCV—normal)
0.0694
0.0802
0.0569
<0.0001
-85.4291
Questionable
Goodness of fit p-value <0.1
BMD 3x lower than lowest non-zero dose
BMDL 3x lower than lowest non-zero dose
BMDL 10x lower than lowest non-zero dose
Hill (NCV—normal)
0.0694
0.2273
0.1763
<0.0001
-80.5334
Questionable
Goodness of fit p-value <0.1
BMDL 3x lower than lowest non-zero dose
Polynomial (5 degree) (NCV—normal)
0.0694
111.5387
69.6621
<0.0001
78.5088
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >11.51 actual
response std. dev.
Polynomial (4 degree) (NCV—normal)
0.0694
112.5746
91.2386
<0.0001
78.4897
Questionable
Goodness of fit p-value <0.1
This document is a draft for review purposes only and does not constitute Agency policy.
D-32 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Models
Test 3
(P-
Value)
1 standard deviation
Goodness of fit
(p-value)
AIC
BMDS
classification*1
BMDS notes
BMD
BMDL
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >11.51 actual
response std. dev.
Polynomial (3 degree) (NCV—normal)
0.0694
111.6385
59.1142
<0.0001
78.5079
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Polynomial (2 degree) (NCV—normal)
0.0694
112.3898
70.3271
<0.0001
78.4933
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >11.51 actual
response std. dev.
Power (NCV—
normal)
0.0694
111.6423
57.6923
<0.0001
78.5078
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >11.51 actual
response std. dev.
Linear
(NCV—normal)
0.0694
115.0030
56.6839
<0.0001
78.4446
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >11.51 actual
response std. dev.
aNonconstant models failed to model the data.
^'Classification" column denotes whether a model can be considered for model selection purposes. See BMDS User Guide: https://www.epa.gov/bmds.
This document is a draft for review purposes only and does not constitute Agency policy.
D-33 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information—Perfluorohexanesulfonic Acid and Related Salts
NTP 2018 Male Rat Total T4 Data
0.5
o
0123456789 10
Dose
Figure D-4. Dose response data for male rat Total T4 NTP T20181. X-axis is dose
(rag/kg-d) and y-axis is level of Total T4 ((.ig/dL).
Decreased total T4 — female rats fNTP. 20161
Table D-19. Dose-response data for total T4 in female rats NTP f 20181
Dose (mg/kg-d)
n
Mean
(Hg/dL)
SD
0
10
3.99
0.186
3.12
10
3.53
0.196
6.25
10
3.37
0.165
12.5
10
2.97
0.108
25
10
2.96
0.194
50
10
2.69
0.145
This document is a draft for review purposes only and does not constitute Agency policy,
D-34 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table D-20. Benchmark dose results for decreased total T4 in female rats— constant variance, BMR = 1 standard
deviation
Test 2
1 standard deviation
Goodness of fit
BMDS
Models3
(p-Value)
BMD
BMDL
(p-Value)
AIC
classification13
BMDS notes
Exponential 2
(CV—
normal)
0.4699
10.6750
8.5600
<0.0001
20.0623
Questionable
Goodness of fit p-value <0.1
| Residual for Dose Group Near BMD | > 2
| Residual at control | > 2
Modeled control response std. dev. > 11.51 actual
response std. dev.
Exponential 3
(CV—
normal)
0.4699
10.6760
8.5600
<0.0001
20.0623
Questionable
Goodness of fit p-value <0.1
| Residual for Dose Group Near BMD | > 2
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Exponential 4
(CV—
0.4699
1.2952
0.9782
0.0075
-29.7163
Questionable
Goodness of fit p-value <0.1
BMDL 3x lower than lowest non-zero dose
normal)
Exponential 5
(CV—
0.4699
1.2955
0.9784
0.0075
-29.7163
Questionable
Goodness of fit p-value <0.1
BMDL 3x lower than lowest non-zero dose
normal)
Hill (NCV—
normal)
0.4699
0.9571
0.6949
0.0358
-33.1163
Questionable
Goodness of fit p-value <0.1
BMD 3x lower than lowest non-zero dose
BMDL 3x lower than lowest non-zero dose
Polynomial
(5 degree)
(CV—
normal)
0.4699
13.4343
11.3456
<0.0001
25.0677
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Polynomial
(4 degree)
(CV—
normal)
0.4699
13.4343
10.9606
<0.0001
25.0677
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
This document is a draft for review purposes only and does not constitute Agency policy.
D-35 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Models3
Test 2
(p-Value)
1 standard deviation
Goodness of fit
(p-Value)
AIC
BMDS
classification13
BMDS notes
BMD
BMDL
Polynomial
(3 degree)
(CV—
normal)
0.4699
13.4343
10.9605
<0.0001
25.0677
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Polynomial
(2 degree)
(CV—
normal)
0.4699
13.4343
10.9605
<0.0001
25.0677
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Power (CV—
normal)
0.4699
13.4343
10.9610
<0.0001
25.0677
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Linear
(CV—
normal)
0.4699
13.4343
10.9605
<0.0001
25.0677
Questionable
Goodness of fit p-value <0.1
| Residual at control | > 2
Modeled control response std. dev. > 11.51 actual
response std. dev.
aConstant models failed to model the data.
^'Classification" column denotes whether a model can be considered for model selection purposes. See BMDS User Guide: https://www.epa.gov/bmds.
This document is a draft for review purposes only and does not constitute Agency policy.
D-36 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information—Perfluorohexanesulfonic Acid and Related Salts
Figure D-5. Dose response data for female rat Total T4 NTP f2018). X-axis is
dose (mg/kg-d] andy-axis is level of Total T4 (|j.g/dL).
This document is a draft for review purposes only and does not constitute Agency policy,
D-37 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table D-21. Benchmark dose results for decreased T3 in male rats—nonconstant variance, BMR = 1 standard
deviation
Models
Restriction3
1 standard deviation
p-Value
AIC
BMDS
classification13
BMDS notes
BMD
BMDL
Exponential 2
(NCV—
normal)
0.1150
7.7723
5.7661
<0.0001
441.4262
Questionable
Goodness of fit p-value <0.1
| Residual for Dose Group Near BMD | > 2
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Exponential 3
(NCV—
normal)
0.1150
7.7700
5.7661
<0.0001
441.4262
Questionable
Goodness of fit p-value <0.1
| Residual for Dose Group Near BMD | > 2
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Exponential 4
(NCV—
normal)
0.1150
0.1196
0.0825
0.0014
332.8680
Questionable
Goodness of fit p-value <0.1
BMD 3x lower than lowest non-zero dose
BMDL 3x lower than lowest non-zero dose
Exponential 5
(NCV—
normal)
0.1150
0.1527
0.0849
0.0005
334.4245
Questionable
Goodness of fit p-value <0.1
BMD 3x lower than lowest non-zero dose
BMDL 3x lower than lowest non-zero dose
Hill (NCV—
normal)
0.1150
0.2869
0.1542
0.0002
336.3586
Questionable
Goodness of fit p-value <0.1
BMDL 3x lower than lowest non-zero dose
Polynomial
(5 degree)
(NCV—
normal)
0.1150
19.8036
10.2767
<0.0001
439.1893
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Polynomial
(4 degree)
(NCV—
normal)
0.1150
19.8036
10.2769
<0.0001
439.1893
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
This document is a draft for review purposes only and does not constitute Agency policy.
D-38 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Models
Restriction3
1 standard deviation
p-Value
AIC
BMDS
classification13
BMDS notes
BMD
BMDL
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Polynomial
(3 degree)
(NCV—
normal)
0.1150
19.8036
10.2767
<0.0001
439.1893
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Polynomial
(2 degree)
(NCV—
normal)
0.1150
19.8036
10.2769
<0.0001
439.1893
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >| 1.51 actual
response std. dev.
Power (NCV—
normal)
0.1150
19.8036
10.2767
<0.0001
439.1893
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >11.51 actual
response std. dev.
Linear
(NCV—
normal)
0.1150
19.8036
10.2768
<0.0001
439.1893
Questionable
Goodness of fit p-value <0.1
BMD higher than maximum dose
BMDL higher than maximum dose
| Residual at control | > 2
Modeled control response std. dev. >11.51 actual
response std. dev.
aNonconstant models failed to model the data.
^'Classification" column denotes whether a model can be considered for model selection purposes. See BMDS User Guide: https://www.epa.gov/bmds.
This document is a draft for review purposes only and does not constitute Agency policy.
D-39 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information—Perfluorohexanesulfonic Acid and Related Salts
NTP 2018 Male RatT3 Data
100
M
Ł 40
30
20
10
0
0123456789 10
Dose
Figure D-6. Dose response data for male rat T3 NTP f2Q18).
X-axis is dose (mg/kg-d) and y-axis is level of T3 (ng/dL).
This document is a draft for review purposes only and does not constitute Agency policy,
D-40 DRAFT-DO NOT CITE OR QUOTE '
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Decreased T3 — F1 rats PND 17 Rainfagj et al. (20201
Table D-22. Dose response data for decreased free T3 in F1 combined
PND16/17 rats Ramhfli et al. f20201
Dose (mg/kg-d)
n
Mean
(ng/dL)
SD
0
18
99.91023737
13.57584288
0.05
14
102.8114448
5.986381078
5
14
92.91322639
8.979571512
25
14
83.80963867
11.99271652
This document is a draft for review purposes only and does not constitute Agency policy.
D-41 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Table D-23. Benchmark dose results for decreased T3 in F1 PND16 male rats — constant variance, BMR = 1
standard deviation Ramhfli etai. f20201
Model
1 standard deviation
Test 4
P-Value
AIC
BMDS
recommendation
BMDS recommendation notes
BMD
BMDL
Exponential 2 (CV -
normal)
4.670072
3.90788
<0.0001
323.9549629
Questionable
Constant variance test failed (Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near BMD| > 2
Exponential 3 (CV -
normal)
4.669809
3.907839
<0.0001
323.9549627
Questionable
Constant variance test failed (Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near BMD| > 2
Exponential 4 (CV -
normal)
1.408148
1.066629
0.0020435
303.9294537
Questionable
Constant variance test failed (Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near BMD| > 2
Exponential 5 (CV -
normal)
3.260612
1.101338
NA
305.2949793
Questionable
Constant variance test failed (Test 2 p-value < 0.05)
d.f.=0, saturated model (Goodness of fit test cannot
be calculated)
Hill (CV - normal)
3.12622
0.961024
NA
305.2949187
Questionable
Constant variance test failed (Test 2 p-value < 0.05)
BMD/BMDL ratio > 3
d.f.=0, saturated model (Goodness of fit test cannot
be calculated)
Polynomial Degree 3
(CV - normal)
5.175471
4.358225
<0.0001
326.0612444
Questionable
Constant variance test failed (Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near BMD| > 2
Polynomial Degree 2
(CV - normal)
5.175471
4.358136
<0.0001
326.0612444
Questionable
Constant variance test failed (Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near BMD| > 2
Power (CV - normal)
5.175466
4.357586
<0.0001
326.0612444
Questionable
Constant variance test failed (Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near BMD| > 2
Linear (CV - normal)
5.175471
4.358409
<0.0001
326.0612444
Questionable
Constant variance test failed (Test 2 p-value < 0.05)
Goodness of fit p-value < 0.1
| Residual for Dose Group Near BMD| > 2
This document is a draft for review purposes only and does not constitute Agency policy.
D-42 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Figure D-7. Dose response data for decreased T3 in F1 PND17 rats Ramhoj et
al. f20201. X-axis is dose (mg/kg-d), andy-axis is level of T3 (ng/dL).
This document is a draft for review purposes only and does not constitute Agency policy.
D-43 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
APPENDIX E. DETAILED PHARMACOKINETIC
ANALYSES
This appendix provides two detailed pharmacokinetic analyses. The first is a Bayesian
analysis of perfluorohexanesulfonic acid (PFHxS) pharmacokinetics in laboratory animals to
estimate key pharmacokinetic parameters. The second is the description and evaluation of a one-
compartment pharmacokinetic (PK) modeling approach for estimating internal doses, evaluated
against rat PFHxS PK data using the mean parameter values estimated for male rats in the Bayesian
estimation.
E.l. BAYESIAN ANALYSIS OF PFHxS PHARMACOKINETICS IN RATS, MICE,
AND MONKEYS
We estimated the sex-specific pharmacokinetic parameters (half-life, volume of
distribution, and clearance) of PFHxS in rats, mice, and nonhuman primates (cynomolgus monkeys)
by fitting one- and two-compartment models to the available concentration versus time data. A
Bayesian hierarchical methodology was developed to fit these models because of the need to pool
time-course concentration data across numerous studies with varying exposure scenarios within
each study. This approach allowed for each concentration-versus-time data set to be fit to each
model in which fitted parameters for each data set are sampled from a population-level distribution
that models the similarities between each data set. In addition, the Bayesian analysis allowed for
the generation of central estimates and credible intervals for the pharmacokinetic parameter of
interest (e.g., half-life, volume of distribution and clearance) using posterior distributions from the
estimated variables. Finally, the Bayesian methodology allowed for hypothesis testing of the one-
and two-compartment formulations to decide which model more appropriately fit the data.
E.l.l. Pharmacokinetic Model
To determine pharmacokinetic parameters for PFHxS, we estimated constants for both one-
and two-compartment model assumptions. For a one-compartment model assumption, the
following exponential decay functions were fit to the available data:
This document is a draft for review purposes only and does not constitute Agency policy.
E-l DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
where D represents the administered dose and V, ke, and ka represent the central compartment
volume, elimination constant, and absorption constant (for oral only) to be fit From these fitted
constants, pharmacokinetic parameters are derived:
V
Vd =
BW
In 2
CLC = Vd * ke
where Vd, ti/2, and CLC represent the volume of distribution, terminal half-life, and clearance
respectively and BW represents the animal body weight.
For the two-compartment model assumption, the following exponential decay functions
were fit to available data:
,t T CC — kdr i / kdr — &
\IV . Aoral _ j I ac
A'V _ Aorai _ k
a-p ' a\(ka-a){p-ct)
-,IV _ P ~ kdc noral _ ^dc ~ @
Di v — 'm . Dorai = k
P~a ' a\(ka-p)(a-p)
C?.cmpm =^{A,r e-" + B" e~f)
C2-clmpt(t) = ^-[Aorale-at + Borale~P* - (Aoral + Boral)e~k
where D represents the administered dose and V, a, (3, kdC, and ka represent central compartment
volume, alpha-phase elimination constant, beta-phase elimination constant, deep-to-central
compartment rate constant, and absorption constant (for oral only) to be fit From these fitted
constants, the remaining two-compartment constants (kCd: central-to-deep compartment rate
constant and ke: elimination constant) and the deep compartment volume [Vdeep) are derived by
solving:
a + p = kcd + kdc + ke
a* P = kdc*ke
kccL
Vdeep = V^
K-dc
which allows for the desired pharmacokinetic parameters to be derived using the following
equations:
„ _ V + Vdeep _ V fkcd + kdc\
d~ss ~ BW ~ BW V k^c >
In 2
^ = ~P~V
clc = bw* ke
This document is a draft for review purposes only and does not constitute Agency policy.
E-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
where Vd-ss, ti/2, and CLC represent the steady-state volume of distribution, terminal half-life, and
clearance respectively and BW represents the animal body weight.
Bayesian Inference
The fitted constants for each model structure (described above) were determined using
available time-course concentration data reported in mice and rats with the parameters for each
model estimated using a Bayesian calibration approach. As described in the main text, owing to the
discrepancy between oral and IV dosing bioavailability, only mice and rat time-course
concentration data following oral gavage dosing was used for rodent-specific fits, while only IV
dosing was available for nonhuman primate fits. In addition, for mice and nonhuman primates,
time-course data from only one study fSundstrom et al. 20121 were available and all sex-specific
data were pooled into a single data set and fit to the one- and two-compartment models described
above. However, a hierarchical Bayesian calibration approach was used to fit the observed time-
course concentration data for male and female rats using data reported from multiple studies
(Huang et al, 2019: Kim etal, 2018: Kim et al, 20161. For the two-compartment model, to ensure
parameter identifiability, a and /? were constrained to be ordered such that a > /?. This constraint
ensures the exponential terms are identifiable and do not "flip" while exploring the parameter
space during Markov-chain Monte-Carlo (MCMC) sampling. Finally, priors for each pharmacokinetic
parameter were chosen to be "weakly informative" based on prior knowledge of PFAS
pharmacokinetics (ATSDR ref) with 95% equal-tailed intervals spanning multiple order of
magnitude.
Priors for pharmacokinetic parameters are presented in Table E-l with corresponding
model-specific parameter prior distributions presented below. Finally, a sensitivity analysis on the
model priors is shown in the Prior Sensitivity Analysis section.
Table E-l. Weakly informed prior distributions for pharmacokinetic
parameters used in the Bayesian analysis
Median
Mad
Eti_3%
Eti_97%
Half-life (d)
15
12
0.88
250
Clearance (ml/kg-d)
50
49
0.32
6000
Vd-ss (ml/kg)
900
811
9.3
32822
Median, mean absolute deviation (mad), lower (eti_3%), and upper (eti_97%) of the
equal-tailed interval prior for each pharmacokinetic parameter.
For the mouse and nonhuman primate data, the following model was implemented to fit all
data reported in Sundstrom et al. (20121:
\nka ~JV(0,1)
ln7~JV(0,l)
In ke ~JV(—3,1.5) one compartment model
In kdc ~N(0,1) two compartment model
This document is a draft for review purposes only and does not constitute Agency policy.
E-3 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
o~Exp( 1)
Ct~LN(xt,a)
where Xj is the sample mean of the observed concentrations at time tt for all times reported. Model
parameter priors are derived from the pharmacokinetic parameter priors defined earlier.
For the hierarchical approach, the concentration versus time data comprised a population-
and data set-level for which model parameters were estimated. Here, each data set represented
each study/sex/dose concentration versus time data set extracted from the literature and were fit
using the model:
where xtj is the sample mean of the observed concentrations at time for data set j and ak is
study-level log-transformed standard deviation for the relative errors based on study k. Study-level
priors for ak were determined using the average log-transformed standard deviations:
where s; y is the sample standard deviation on the observed concentrations at time t[j for study k.
If Sij was available, a^j is the log-transformed standard deviation using the sample mean and
standard deviation. For studies in which sample standard deviations could not be extracted, an
average of all log-transformed standard deviations was used, which allowed for study-level prior
distributions on the error model log-transformed standard deviation:
Using this model, data set-level fitted constants were assigned priors based on a
noncentered parameterization of a population-level distribution. This reparameterization of a
typical hierarchical Bayesian model allows for increased sampling efficiency and can be more
efficient for sampling when there is limited data fBetancourt and Girolami. 20131. Finally,
nonelimination rate constants (ka and kdC) were assigned a unit normal, weakly informative prior to
aid parameter identifiability fGelman et al, 20151:
(C\°%eipt for 1-compartment model,
lJ [exempt for 2-compartment model
Cik ~LN (X;j, (T/j)
•route
This document is a draft for review purposes only and does not constitute Agency policy.
E-4 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
ln/xfca~JV(0,l)
ln^ ~JV(0,1)
Inpke ~N(—3,1.5) one compartment model
Inpkdc~N(0,1) two compartment model
In na,p ~N(—3,1.5), < na two compartment model
Ofcay,fce,a,/?,fcdc~Exp(l)
ke> a> P> kdc)_/ ~^(Mfcay,fce,a,/?,fcdc> ^ka.V.kg.a.p.kac)
For both the single-level and hierarchical approaches, one- and two-compartment model
goodness of fits were compared using the widely applicable information criteria (WAIC, fWatanabe.
201011. Pharmacokinetic parameters from the most appropriate model, as judged by the WAIC
comparison, were reported. To estimate the resulting pharmacokinetic parameters, we examined
posterior probability densities of the parameters from the WAIC-determined model and calculated
distributional estimates of the half-life, volume of distribution, and clearance using the equations
described above. The parameter space was sampled using PyMC fSalvatier et al. 20161 using four
independent Markov chains run for 10,000 iterations per chain. Posterior parameter distributions
were determined using the final 5,000 iterations of each chain ensuring an effective sample size
(ESS) greater than 10,000 fKruschke, 20211. Convergence was assessed using a potential scale
reduction factor with a maximum threshold of R = 1.05 fKruschke. 20211.
Prior Sensitivity Analysis
To investigate the impact of prior selection on posterior pharmacokinetic parameter
estimation, we conducted a sensitivity analysis on the priors used in the Bayesian analysis. Priors
were classified into three categories: weakly informed, broad, and uninformed. Weakly informed
priors are defined using the half-life, clearance, and volume of distribution described above based
on reported ranges of PFHxS pharmacokinetics with a prior predictive check demonstrating
available data for fitting fall within the prior 90% credible interval.
This document is a draft for review purposes only and does not constitute Agency policy.
E-5 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Hero: 5387170, PFHxS gavage 32.00 mg/kg Hero: 5387170, PFHxS gavage 16.00 mg/kg Hero: 5387170, PFHxS gavage 4.00 mg/kg
0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
time [days] time [days] time [days]
Figure E-l. Prior predictive check to ensure equal-tailed interval from prior
distributions encompass the available time-course concentration data for
fitting. Observed data from Kim et al. f2016Y
1 Broad priors are defined as uniform distributions spanning the 3% and 97% ETI defined
2 from the weakly informed priors and uninformed priors represent uniform priors spanning
3 multiple orders of magnitude and are essentially flat priors. The following figure compares these
4 three classes of priors and their impact on the posterior pharmacokinetic parameter distributions.
Table E-2. Results from prior sensitivity analysis for the three classes of priors
(weakly informed, broad, and uninformed). For each pharmacokinetic
parameters, mean, standard deviation (SD), lower HDI (HDI 5%), and upper
HDI (95%) are presented.
Prior
Half-life (d)
Clearance (ml/kg-d)
Vd-ss (mL/kg)
Mean
SD
HDI 5%,
95%
Mean
SD
HDI 5%,
95%
Mean
SD
HDI 5%, 95%
Weakly
informed
1.86
0.16
1.62, 2.11
84.1
12.7
64.7, 103.8
224.1
28.1
182.7, 266.4
Broad
1.85
0.16
1.61, 2.10
82.6
12.4
63.4, 101.8
218.6
26.6
177.7, 259.7
Uninformed
1.85
0.16
1.60, 2.08
82.9
12.4
63.4, 101.6
219.3
27.1
176.5, 258.6
5 Informed by these findings, EPA used the weakly informed pharmacokinetic priors for
6 fitting available time-course concentration data.
Study-specific clearance values and model fits
7 As described above, three data sets were used for the female rat-specific parameter
8 estimation for which only the oral gavage data were used for fitting (Huang etal.. 2019: Kim etal..
9 2018: Kim etal.. 20161. In addition to these three, a fourth data set fBenskin et al.. 20091 was used
10 for male rats. The sex-specific clearance value distribution obtained from fitting the three data sets
This document is a draft for review purposes only and does not constitute Agency policy.
E-6 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
1 together had means and 90% credible intervals of 61.36 (55.5 -67.17) mL/kg-day in female rats
2 and 7.15 (3.73-10.26) mL/kg-day in male rats. Comparatively sex-specific clearances in mice,
3 determined from Sundstrom etal. f20121. had means and 90% credible intervals of 3.18 (2.83-
4 3.52) mL/kg-day in female mice and 3.86 (2.83-3.52) mL/kg-day in male mice. For rat and mice
5 data, a one-compartment PK model was deemed superior for mice and rats based on the WAIC and
6 visual inspection of the plots indicating a lack of distribution and excretion phase. For nonhuman
7 primates, a clear distribution and excretion phase is observed in the data with WAIC indicating a
8 two-compartment model for fitting. Data from Sundstrom etal. (2012) had means and 90%
9 credible intervals of 2.12 (1.81-2.44) mL/kg-day in female cynomolgus monkeys and 1.39 (0.94-
10 1.83) mL/kg-day in male cynomolgus monkeys.
Population clearance (ml/{d-kg)): 7.53 (3.197 - 11.469)
Hero: 3749289 (0), PFHxS gavage 4.00 mg/kg
CLC (ml/kg/day): 5.57 (5.46 - 5.68)
time [days]
Population clearance (ml/(d-kg)): 84.10 (64.720 - 103.801)
Hero: 3749289 (0), PFHxS gavage 4.00 mg/kg
CLC (ml/kg/day): 117.84 (110.55 - 125.28)
time [days]
Figure E-2. Predicted (black line with blue 90% credible interval) and
observed (black circles) serum time-courses for male (top panel) and female
(bottom panel) rats after a 4 mg/kg gavage PFHxS. Observed data from Kim et
al. f2016V
This document is a draft for review purposes only and does not constitute Agency policy.
E-7 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Population clearance (ml/{d-kg)): 7.53 (3.197 - 11.469)
Hero: 4239569 (1), PFHxS gavage 10.00 mg/kg
CLC (ml/kg/day): 6.13 (2.84 - 9.71)
— 1 1 1 1 1 1
0.0 2.5 5.0 7.5 10.0 12.5 15.0
time [days]
Population clearance (ml/(d-kg)): 84.10 (64.720 - 103.801)
Hero: 4239569 (2), PFHxS gavage 4.00 mg/kg Hero: 4239569 (1), PFHxS gavage 1.00 mg/kg
time [days] time [days]
Figure E-3. Predicted (black line with blue 90% credible interval) and
observed (black circles) serum time-courses for male (top panel) and female
(bottom two panels) rats after a 10 mg/kg gavage (both male and female) and
4 mg/kg gavage (female only) PFHxS. Data from Kim etal. (2018).
This document is a draft for review purposes only and does not constitute Agency policy.
E-8 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information—Perfluorohexanesulfonic Acid and Related Salts
Population clearance (ml/(d-kg)): 7.53 (3.197 - 11.469)
Hero: 5387170 (4), PFHxS gavage 32.00 mg/kgHero: 5387170 (3), PFHxS gavage 16.00 mg/kg Hero: 5387170 (2), PFHxS gavage 4.00 mg/kg
CLC (ml/kg/day): 9.79 (8.45 - 11.04) CLC (ml/kg/day): 5.92 (5.09 - 6.74) CLC (ml/kg/day): 5.38 (4.61 - 6.15)
time [days] time [days] time [days]
Population clearance (ml/(d-kg)): 84.10 (64.720 - 103.801)
Hero: 5387170 (5), PFHxS gavage 32.00 mg/kgHero: 5387170 (4), PFHxS gavage 16.00 mg/kg Hero: 5387170 (3), PFHxS gavage 4.00 mg/kg
CLC (ml/kg/day): 94.53 (85.43 - 103.33) CLC (ml/kg/day): 61.36 (55.58 - 67.17) CLC (ml/kg/day): 50.14 (45.02 - 55.01)
time [days] time [days] time [days]
Figure E-4. Predicted (black line with blue 90% credible interval) and
observed (black circles) serum time-courses for male (top panel) and female
(bottom panel) rats after a 4,16, or 32 mg/kg gavage PFHxS. Data from Huang
etal. (20191-
This document is a draft for review purposes only and does not constitute Agency policy,
E-9 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information—Perfluorohexanesulfonic Acid and Related Salts
Clearance (ml/(d-kg)): 3.86 (3.267 - 4.412)
time [days] time [days]
Clearance (ml/(d-kg)): 3.18 (2.829 - 3.517)
Hero: 1289834, PFHxS gavage 20.00 mg/kg Hero: 1289834, PFHxS gavage 1.00 mg/kg
time [days] time [days]
Figure E-5. Predicted (black line with blue 90% credible interval) and
observed (black circles) serum time-courses for male (top panel) and female
(bottom panel) mice after a 1 or 20 mg/kg gavage PFHxS. Data from Sundstrom
etal. f20121.
This document is a draft for review purposes only and does not constitute Agency policy,
E-10 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Clearance (ml/(d-kg)): 1.39 (0.943 - 1.825)
Hero: 1289834, PFHxS iv 10.00 mg/kg
time [days]
Clearance (ml/(d-kg)): 2.12 (1.810 - 2.444)
Hero: 1289834, PFHxS iv 10.00 mg/kg
time [days]
Figure E-6. Predicted (black line with blue 90% credible interval) and
observed (black circles) serum time-courses for male (top panel) and female
(bottom panel) nonhuman primates following a 10 mg/kg IV PFHxS dose. Data
from Sundstrom etal. (2012).
E.2. DESCRIPTION AND EVALUATION OF A SINGLE-COMPARTMENT PK
APPROACH
1 A single-compartment PK model based on that described by Verner etal. (20161 was
2 implemented. Verner etal. (20161 described linked one-compartment models for a mother and
3 fetus or child. For this analysis the sub-model for the mother was used with distribution to the
4 offspring set to zero and the model parameter inputs adjusted to use clearance (CL) rather than
5 half-life as the input parameter but given the change in parameters the model is otherwise
6 mathematically identical. The resulting differential equation for the amount of a substance in an
7 individual after oral dosing is:
dA
8 —— = FaijS x d x BW CL x x A/Vd
This document is a draft for review purposes only and does not constitute Agency policy.
E-ll DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
where A is the amount in the individual (mg), d is the dose (mg/kg-day), BW the body weight (kg),
CL is the clearance (L/kg-day), Vd is the volume of distribution (L/kg), and rm:mf is a factor to
account for nonuniform distribution between a pregnant individual and her fetus(es), in the event
that one wishes to simulate dosimetry during gestation. For the following simulations of dosimetry
in nonpregnant animals rm:mf is set to 1. While data are not available to indicate that distribution
differs in pregnancy versus nonpregnant adult animals, the term is still included to allow for this
possibility. For this analysis CL and Vd are assumed to be constant at the values determined from
PK studies in young adult animals, as described above. The concentration in an individual's blood is
then C = A-rm:mf/(Vdm-BW).
The changes in male and female rat BW observed in the NTP bioassay (28-day exposure
fNTP. 201911 are shown in Figure E-7. Internal doses of PFHxS predicted by the PK model as a
function of exposure day, using the population mean male rat parameters from Table 3-1 and Table
3-3, are shown in Figure E-8. The dose is assumed to be adjusted for changes in BW each day.
Because the animals were necropsied on day 29,1 day after the final dose, the model simulations
include a final day with zero exposure. Mean plasma PFHxS concentrations from the NTP study,
collected at time of necropsy, are shown for comparison. Very little accumulation was predicted in
female rats after exposure day 10, whereas male rats accumulated PFHxS throughout the study.
This document is a draft for review purposes only and does not constitute Agency policy.
E-12 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information—Perfluorohexanesulfonic Acid and Related Salts
350
340
330
320
_c
bJ)
| 310
>
T3
O 300
CD
290
280
270
Male Rat Body Weight
~
s
% Control
~ 0.625 mg/kg/d
~/
¦ 1.25 mg/kg/d
© 2.5 mg/kg/d
A 5 mg/kg/d
O 10 mg/kg/d
average
10
15
Study day
20
25
30
240
235
230
225
-Ł.220
"3
> 215
TJ
O
cn
210
205
200
195
• Contol
Female Rat Body Weight
~ 3.i2mg/Kg/a
¦ 6.25 mg/kg/d
~
O 12.5 mg/kg/d
¦
o
A 25 mg/kg/d
O 500 mg/kg/d
average
•
<
•
0 _
10
15
Study day
20
25
30
Figure E-7. Male and female rat body weight changes during 28-day PFHxS
bioassay. Data sets from NTP f20191 are identified by the dose (mg/kg-d).
This document is a draft for review purposes only and does not constitute Agency policy,
E-13 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information—Perfluorohexanesulfonic Acid and Related Salts
1
0
1
5
1 1
10 15
Days
I
20
I
25
I
30
o
to -
1
t
t
0
CM -
[j
cfe
E
Co .
o a)
~
J /
j
!
t
1 *
J /
, /
1 /
SimiJlations
3 12 mg.'kgj'd
6 25 mg/kg.'d
12 5mg/kg/d
- - ¦ 25 mg/kg/d
50 mg/kg/d
Data
A 3.12 mg/kg/d
+ 6 25 mg/kg/d
x 12 5 mg/kg/d
O 25 mg/kg/d
v 50 mg/kg/d
V
o
C
, 1
J f
O
lS '
E
O)
CO
Q.
1 /
¦ '
i i
i /
; /
Female Rats
\
X
V
A
o
CO
i /
\
o -
1
0
i
5
i i
10 15
Days
l
20
I
25
I
30
Figure E-8. Predicted accumulation and observed end-of-study of PFHxS in
male rats as a function of dose. The plasma concentrations, observed in the NTP
f20191 bioassay, were measured one day after the final dose, hence are plotted on
day 29. Exposure is treated as continuous for 28 days.
The y-axis scales in Figure E-8 are set to focus on the range of the experimental data and
because of nonlinearity in that data (discussed in more detail in the Pharmacokinetics section) the
upper portions of the higher concentration curves from the PK model are outside that range. For
example, when the dose to male rats was increased from 5 to 10 mg/kg-day, the resulting mean
plasma concentrations only increased from 162 to 198 mg/L, or 22%. This nonlinearity was
This document is a draft for review purposes only and does not constitute Agency policy,
E-14 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
1 suggested by the PK data, described in the introduction to this Pharmacokinetics section, for which
2 clearance appeared to be higher at higher administered doses in rats, although the difference in the
3 PK parameters may not have been statistically significant. Because the PK model evaluated is based
4 on first-order kinetics, it predicts that the plasma concentration doubles when the dose is doubled.
5 A likely mechanism for the observed nonlinearity is saturable renal resorption, allowing for faster
6 elimination of PFHxS at higher internal concentrations. It is also possible that absorption is less
7 efficient at higher dose levels. While the linear PK model is thereby shown to provide reasonable
8 predictions of internal dose at lower exposures, the results are inadequate for 2.5 mg/kg-day and
9 higher in male rats and for 12.5 mg/kg-day and higher in female rats.
This document is a draft for review purposes only and does not constitute Agency policy.
E-15 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
APPENDIX F. QUALITY ASSURANCE FOR THE IRIS
TOXICOLOGICAL REVIEW OF
PERFLUOROHEXANESULFONIC ACID AND
RELATED SALTS
This assessment is prepared under the auspices of the U.S. Environmental Protection
Agency's (EPA's) Integrated Risk Information System (IRIS) Program. The IRIS Program is housed
within the Office of Research and Development (ORD) in the Center for Public Health and
Environmental Assessment (CPHEA). EPA has an agency-wide quality assurance (QA) policy that is
outlined in the EPA Quality Manual for Environmental Programs (see CIO 2105-P-01.31 and follows
the specifications outlined in EPA Order CIO 2105.3.
As required by CIO 2105.1, ORD maintains a Quality Management Program, which is
documented in an internal Quality Management Plan (QMP). The latest version was developed in
2013 using Guidance for Developing Quality Systems for Environmental Programs fOA/G-11. An
NCEA/CPHEA-specific QMP was also developed in 2013 as an appendix to the ORD QMP. Quality
assurance for products developed within CPHEA is managed under the ORD QMP and applicable
appendices.
The IRIS Toxicological Review of perfluorohexanesulfonic acid (PFHxS) is designated as
Highly Influential Scientific Information (HISA)/Influential Scientific Information (ISI) and is
classified as QA Category A. Category A designations require reporting of all critical QA activities,
including audits. The development of IRIS assessments is done through a seven-step process.
Documentation of this process is available on the IRIS website: https: //www.epa.gov/iris/basic-
information-about-integrated-risk-info rmation-system#process.
Specific management of quality assurance within the IRIS Program is documented in a
Programmatic Quality Assurance Project Plan (PQAPP). A PQAPP is developed using the EPA
Guidance for Quality Assurance Project Plans fOA/G-51. All IRIS assessments follow the IRIS PQAPP,
and all assessment leads and team members are required to receive QA training on the IRIS PQAPP.
During assessment development, additional QAPPs may be applied for quality assurance
management They include:
This document is a draft for review purposes only and does not constitute Agency policy.
F-l DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Title
Document number
Date
Umbrella Quality Assurance Project
Plan for CPHEA PFAS Toxicity
Assessments
L-CPAD-0031652-QP-1-5
February 2023
Program Quality Assurance Project
Plan (PQAPP) for the Integrated Risk
Information System (IRIS) Program
L-CPAD-0030729-QP-1-6
June 2023
An Umbrella Quality Assurance
Project Plan (QAPP) for Dosimetry
and Mechanism-Based Models
(PBPK)
L-CPAD-0032188-QP-1-3
May 2023
Quality Assurance Project Plan
(QAPP) for Enhancements to
Benchmark Dose Software (BMDS)
L-HEEAD-0032189-QP-1-3
June 2023
1 During assessment development, this project undergoes five quality audits during
2 assessment development including:
Date
Type of audit
Major findings
Actions taken
August 2019
Technical system audit
None
None
August 2020
Technical system audit
None
None
July 2021
Technical system audit
None
None
August 2022
Technical system audit
None
None
June 2023
Technical system audit
None
None
This document is a draft for review purposes only and does not constitute Agency policy.
F-2 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
APPENDIX G. SUMMARY OF PUBLIC AND
EXTERNAL PEER REVIEW COMMENTS AND EPA'S
DISPOSITION
This document is a draft for review purposes only and does not constitute Agency policy.
G-l DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
REFERENCES
3M (3M Company). (1999). The science of organic fluorochemistry. (EPA-HQ-OPPT-2002-0043-0006).
https://www.regulations.gov/document/EPA-HQ-OPPT-2002-0043-00Q6.
ACOG (American College of Obstetricians and Gynecologists). (2020). How your fetus grows during
pregnancy. Available online at https://www.acog.org/womens-health/faqs/how-vour-fetus-
grows-during-pregnancv (accessed
Alkhalawi. E: Kasper-Sonnenberg. M: Wilhelm. M: Volkel. W; Wittsiepe. J. (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.
Ammitzboll. C: Bornsen. L: Petersen. ER: Oturai. AB: Sgndergaard. 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/i.ineuroim.2019.03.002.
Angrish, MM; Kaiser. JP; Mcqueen, CA; Chorley, BN. (2016). Tipping the balance: Hepatotoxicity and the
4 apical key events of hepatic steatosis [Review], Toxicol Sci 150: 261-268.
http://dx.doi.org/10.1093/toxsci/kfw018.
Arrebola, JP; Ramos. JJ; Bartolome, M; Esteban, M; Huetos, O; Canas, Al; Lopez-Herranz, A; Calvo, E;
Perez-Gomez, B; Castano, A; BIOAMBIENT.ES. (2019). Associations of multiple exposures to
persistent toxic substances with the risk of hyperuricemia and subclinical uric acid levels in
BIOAMBIENT.ES study. Environ Int 123: 512-521.
http://dx.doi.Org/10.1016/i.envint.2018.12.030.
Ashley-Martin. J; Dodds. L; Arbuckle. TE; Bouchard. MF; Fisher. M; Morriset. AS; Monnier. P; Shapiro.
GD; Ettinger. AS; Dallaire. R; Taback. S; Fraser. W; Piatt. RW. (2017). Maternal concentrations of
perfluoroalkyl substances and fetal markers of metabolic function and birth weight. Am J
Epidemiol 185: 185-193. http://dx.doi.org/10.1093/aie/kww213.
ATSDR (Agency for Toxic Substances and Disease Registry). (2018). Toxicological profile for
perfluoroalkyls. Draft for public comment [ATSDR Tox Profile], Atlanta, GA: U.S. Department of
Health and Human Services, Centers for Disease Control and Prevention.
https://www.atsdr.cdc.gov/toxprofiles/tp200.pdf.
Averina. M; Brox. J; Huber. S; Furberg. AS; Sgrensen. M. (2019). Serum perfluoroalkyl substances (PFAS)
and risk of asthma and various allergies in adolescents. The Troms0 study Fit Futures in
Northern Norway. Environ Res 169: 114-121. http://dx.doi.Org/10.1016/i.envres.2018.ll.005.
Bach. CC; Bech, BH; Nohr, EA; Olsen, J; Matthiesen, NB; Bonefeld-Jgrgensen, EC; Bossi, R; Henriksen, TB.
(2016a). Perfluoroalkyl acids in maternal serum and indices of fetal growth: The Aarhus Birth
Cohort. Environ Health Perspect 124: 848-854. http://dx.doi.org/10.1289/ehp.1510Q46.
Bach. CC; Vested. A; Jgrgensen, KT; Bonde, JP; Henriksen. TB; Toft. G. (2016b). Perfluoroalkyl and
polyfluoroalkyl substances and measures of human fertility: a systematic review [Review], Crit
Rev Toxicol 46: 735-755. http://dx.doi.org/10.1080/10408444.2016.1182117.
Batzella, E; Zare Jeddi, M; Pitter, G; Russo, F; Fletcher. T; Canova, C. (2022). Associations between
Mixture of Perfluoroalkyl Substances and Lipid Profile in a Highly Exposed Adult Community in
the Veneto Region. Int J Environ Res Public Health 19.
http://dx.doi.org/10.3390/iierphl91912421.
This document is a draft for review purposes only and does not constitute Agency policy.
R-l DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Benskiii, JP; De Silva, AO; Martin, LJ; Arsenault, G; Mccrindle, R; Riddell, N; Maburv, SA; Martin, JW.
(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.
Bessone, F: Razori, MV: 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/sQ0018-
018-2947-0.
Betancourt, MJ; Girolami, M. (2013). Hamiltonian monte carlo for hierarchical models. Betancourt, MJ;
Girolami, M. http://dx.doi.org/10.48550/arXiv.1312.0906.
Bjerregaard-Olesen, C; Bach, CC: Long, M: Wielsge, M: Bech, BH: Henriksen, IB: Olsen, J: Bonefeld-
J0rgensen, 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.doi.org/10.1289/EHP1884.
Bonefeld-J0rgensen, EC. (2022). RE: Hjermitslev et al. Persistent organic pollutants in Greenlandic
pregnant women and indices of foetal growth: The ACCEPT study. Available online at (accessed
Boobis, AR: Doe, JE: Heinrich-Hirsch, B: Meek, ME: Munn, S: Ruchirawat, M: Schlatter, J: Seed, J: Vickers,
C (2008). IPCS framework for analyzing the relevance of a noncancer mode of action for
humans [Review], Crit Rev Toxicol 38: 87-96. http://dx.doi.org/10.1080/104084407Q1749421.
Borenstein, M: Hedgt liggins, JPT: Rothstein, HR. (2009). Introduction to meta-analysis.
Chichester, UK: John Wiley & Sons.
Borghese, MM: Liang, CL: Owen, J: 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.
Boucher, JM: Cousins, I: Scheringer, M: Hungerbuhler, K: Wang, Z. (2019). Toward a comprehensive
global emission inventory of C_4-C_10 perfluoroalkanesulfonic acids (PFSAS) and related
precursors: Focus on the life cycle of C_6- and C_10-based products [Letter], Environ Sci Technol
Lett 6: 1-7. http://dx.doi.org/10.1021/acs.estlett.8b00531.
Brosset, E; Ngueta, G. (2022). Exposure to per- and polyfluoroalkyl substances and glycemic control in
older US adults with type 2 diabetes mellitus. Environ Res 216: 114697.
http://dx.doi.Org/10.1016/i.envres.2022.114697.
Buck, CO: Eliot, MN; Kelst, \ I i Calafat, AM: Chen, A: Ehrlich, S; Lanphear, BP: Braun, JM. (2018).
Prenatal exposure to perfluoroalkyl substances and adipocytokines: the HOME Study. Pediatr
Res 84: 854-860. http://dx.doi.org/10.1038/s41390-018-017Q-l.
Buck Louis, GM; Zhai, S; Smarr, MM: Grewal, J: Zhang, C; Grantz, KL; Hinkle, SN; Sundaram, R; Lee, S;
Honda, M; Oh, J: Kannan, K. (2018). Endocrine disruptors and neonatal anthropometry, NICHD
Fetal Growth Studies - Singletons. Environ Int 119: 515-526.
http://dx.doi.Org/10.1016/i.envint.2018.07.024.
Budtz-Jgrgensen, E: Grandiean, P. (2018a). 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.
Budtz-Jgrgensen, E: Grandiean, P. (2018b). Computational details for the paper "Application of
benchmark analysis for mixed contaminant exposures: Mutual adjustment of perfluoroalkylate
substances associated with immunotoxicity".
Butenhoff, JL: Chang, SC: Ehresman, DJ: 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.
Callan, AC: Rotander, A: Thompson, K; Heyworth, J: Mueller and, J0; Hinwood, AL. (2016).
Maternal exposure to perfluoroalkyl acids measured in whole blood and birth outcomes in
This document is a draft for review purposes only and does not constitute Agency policy.
R-2 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
offspring. Sci Total Environ 569-570: 1107-1113.
http://dx.doi.Org/10.1016/i.scitotenv.2016.06.177.
Cao, L: Quan, XB: Zeng, WJ: Yang, XO: Wang, Ml (2016). Mechanism of hepatocyte apoptosis [Review],
Journal of Cell Death 9: 19-29. http://dx.doi.org/10.4137/JCD.S39824.
Cao, W; Liu, X: Liu, X: Zhou > hang, X: Tian, H; Wang, J; 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.Org/10.1016/i.envint.2018.04.015.
Cellesi, C; Michelangeli, C; Rossolini, GM: Giovannoni, F: Rossolini, A. (1989). Immunity to diphtheria, six
to 15 years after a basic three-dose immunization schedule. Journal of Biological Standardization
17: 29-34. http://dx.doi.org/lQ.lQ16/0Q92-1157f89l9QQ25-5.
Chang, CJ; Barr, DB; Ryan, PB; Panuwet, P: Smarr, MM: Liu, K: Kannan, K; Yakimavets, V; Ta:
Marsit, G: Jones, DP: Corwin, Ei; 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.Org/10.1016/i.envint.2021.106964.
Chapman, AB: Abraham, nudio, S: Coffin, C: Merouani, A: Young, D: Johnson, A: Osorio, F:
Goldberg, C: Moore, LG: Dahms, T: Schrier, RW. (1998). Temporal relationships between
hormonal and hemodynamic changes in early human pregnancy. Kidney Int 54: 2056-2063.
http://dx.doi.Org/10.1046/i.1523-1755.1998.00217.x.
Chen, L: Tong, C: Huo, X: Zhang, J: Tian, Y. (2021). Prenatal exposure to perfluoroalkyl and
polyfluoroalkyl substances and birth outcomes: A longitudinal cohort with repeated
measurements. Chemosphere 267: 128899.
http://dx.doi.Org/10.1016/i.chemosphere.2020.128899.
Chen, M; Zhu, L; Wang, Q; Shan, G. (2020). Tissue distribution and bioaccumulation of legacy and
emerging per-and polyfluoroalkyl substances (PFASs) in edible fishes from Taihu Lake, China.
Environ Pollut 268: 115887. http://dx.doi.Org/10.1016/i.envpol.2020.115887.
Chiu, WA; Lynch, >, CR; Antezana, A: Malek, P; Sokolinski, S; Rogers, RD. (2022). Bayesian
estimation of human population toxicokinetics of PFOA, PFOS, PFHxS, and PFNA from studies of
contaminated drinking water. Environ Health Perspect 130: 127001.
http://dx.doi.org/lQ.1289/EHP101Q3.
Christensen, KY; Raymond, MR; Thompson, BA; Anderson, HA. (2016). Fish consumption, levels of
nutrients and contaminants, and endocrine-related health outcomes among older male anglers
in Wisconsin. J Occup Environ Med 58: 668-675.
http://dx.doi.org/10.1097/JOM.000000000000Q758.
Christenson, B: Bottiger, M. (1986). Serological immunity to diphtheria in Sweden in 1978 and 1984.
Scand J Infect Dis 18: 227-233. http://dx.doi.org/10.3109/00365548609Q32331.
Cohen, NJ: Yao, M: Midva, V: India-Aldana, S: Mouzica, T: Andra, SS: Narasimhan, S: Meher, AK: Arora,
M: Chan, JKY: Chan, SY: Lov, SL: Minguez-Alarcon. L: Oulhote, Y: Huang, J: Valvi, D. (2023).
Exposure to perfluoroalkyl substances and women's fertility outcomes in a Singaporean
population-based preconception cohort. Sci Total Environ 873: 162267.
http://dx.doi.Org/10.1016/i.scitotenv.2023.162267.
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/EE9.0000000000000Q92.
Collier, RJ. (1975). Diphtheria toxin: Mode of action and structure [Review], Bacteriol Rev 39: 54-85.
http://dx.doi.Org/10.1128/br.39.l.54-85.1975.
This document is a draft for review purposes only and does not constitute Agency policy.
R-3 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Cui, F; Liu H, h. >, heng, TZ; Xu, S; Xia, W; Sheng, X. (2022). Association of exposure to per- and
polyfluoroalkyl substances with hemoglobin and hematocrit during pregnancy. Ecotoxicol
Environ Saf 248: 114319. http://dx.doi.Org/10.1016/i.ecoenv.2022.114319.
Derakhshan, A; Kortenkamp, A; Shu, H; Broeren, MAC: Lindh, CH: Peeters, RP: Bornehag, CG: Demeneix,
B; Korevaar, TIM. (2022). Association of per- and polyfluoroalkyl substances with thyroid
homeostasis during pregnancy in the SELMA study. Environ Int 167: 107420.
http://dx.doi.Org/10.1016/i.envint.2022.107420.
Donat-Vargas, C; Bergdahl, I A: Tornevi, A: Wennberg, M; Somrra 'iranta, H; Koponen, J;
Rolandsson, O: Akesson, A. (2019a). Perfluoroalkyl substances and risk of type II diabetes: A
prospective nested case-control study. Environ Int 123: 390-398.
http://dx.doi.Org/10.1016/i.envint.2018.12.026.
Donat-Vargas, C; Bergdahl, I A: Tornevi, A: Wennberg, M; Sommar, J: Koponen, J: Kiviranta, H: Akesson,
A. (2019b). Associations between repeated measure of plasma perfluoroalkyl substances and
cardiometabolic risk factors. Environ Int 124: 58-65.
http://dx.doi.Org/10.1016/i.envint.2019.01.007.
Dufour, P: Pirard, C: Petrossians, P: Beckers, A: Charlier, C. (2020). Association between mixture of
persistent organic pollutants and thyroid pathologies in a Belgian population. Environ Res 181:
108922. http://dx.doi.Org/10.1016/i.envres.2019.108922.
Dunder, L: Salihovic, S: Elmstahl, S: Lind, PM: Lind, L. (2023). Associations between per- and
polyfluoroalkyl substances (PFAS) and diabetes in two population-based cohort studies from
Sweden. J Expo Sci Environ Epidemiol. http://dx.doi.org/10.1038/s41370-023-0Q529-x.
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/EE9.0000000000000Q95.
E, L; Zhang, S; Jiang, 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 Pollut Res Int 30: 44292-44303.
http://dx.doi.org/10.1007/sll356-023-25258-4.
Edwards, SW; Tan, YM; Villeneuve, PL: Meek, ME: McQueen, CA. (2016). Adverse outcome pathways-
Organizing toxicological information to improve decision making [Review], J Pharmacol Exp Ther
356: 170-181. http://dx.doi.org/lQ.1124/ipet.115.228239.
Eick, SM; Demicco, E; Valeri, L; Woodruff, Ti; Morello-Frosch, R; Horn Thepaksorn, EK; Izano, MA;
Cushing, U: Wang, Y; Smith, SC: Gao, S; Park, JS: 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/sl2940-020-0Q654-2.
Fan, S: Wn >, I Mom, MS h , i, *Ihen, L: Wan t \\, h, , Jiang, Q: Bu. L: Shi, J: Shi 1, \ ng. 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/i.scitotenv.2022.160617.
Feng, Y: Bai, Y: Lu. Y: Chen, M: Fu. M: Guan, X: Cao, Q: Yuan, F: Jie, J: Li, M: Meng, H: Wang, C: Hong, S:
Zhou, Y: Zhang, X: He, M: Guo, H. (2022a). Plasma perfluoroalkyl substance exposure and
incidence risk of breast cancer: A case-cohort study in the Dongfeng-Tongji cohort. Environ
Pollut 306: 119345. http://dx.doi.Org/10.1016/i.envpol.2022.119345.
Feng, Y; Fu, M; Guan, X: Wang, C; Meng, H; Zhou, Y; He, M; Guo, H. (2022b). Associations of exposure to
perfluoroalkyl substances with serum uric acid change and hyperuricemia among Chinese
This document is a draft for review purposes only and does not constitute Agency policy.
R-4 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
women: Results from a longitudinal study. Chemosphere 308: 136438.
http://dx.doi.Org/10.1016/i.chemosphere.2022.136438.
Galazka, A; Kardvmowicz, B. (1989). Immunity against diphtheria in adults in Poland. Epidemiol Infect
103: 587-593. http://dx.doi.org/10.1017/s095026880003Q983.
Galazka, AM: Milstien, JB; Robertson, SE; Cutts, FT. (1993). The immunological basis for immunization
module 2 : Diphtheria. (WHO/EPI/Gen/93.11-18). Galazka, AM; Milstien, JB; Robertson, SE;
Cutts, FT. http://apps.who.int/iris/bitstream/handle/10665/58891/WHQ-EPI-GEN-93.12-mod2-
eng.pdf?sequence=38&isAllowed=y.
Gao, K, e; Zhuang, T; Liu, X: Fu, J: Zhang, J: Fu, J, ie; Wang, L; Zhang, A: Liang, Y: Song, M: Jiang, G. (2019).
Prenatal Exposure to Per- and Polyfluoroalkyl Substances (PFASs) and Association between the
Placental Transfer Efficiencies and Dissociation Constant of Serum Proteins-PFAS Complexes.
Environ Sci Technol 53: 6529-6538. http://dx.doi.org/10.1021/acs.est.9b00715.
Gao, K; Fu, J: Xue, Q; Li, Y; Liang, Y: Pan, Y: Zhang, A: Jiang, G. (2018). An integrated method for
simultaneously determining 10 classes of per- and polyfluoroalkyl substances in one drop of
human serum. Anal Chim Acta 999: 76-86. http://dx.doi.Org/10.1016/i.aca.2017.10.038.
Gavlord, A: Berger, Kl: Naidu, M: Attina, TM: Gilbert, J: Koshv, TT; Han, X: Marmor, M: Shao, Y: Giusti, R:
Goldring, RM: Kannan, K: Trasande, L. (2019). Serum perfluoroalkyl substances and lung function
in adolescents exposed to the World Trade Center disaster. Environ Res 172: 266-272.
http://dx.doi.Org/10.1016/i.envres.2019.02.024.
Gavlord, A: Trasande, L: Kannan, K: Thomas, KM: Lee, S: Liu, M: Levine, J. (2020). Persistent organic
pollutant exposure and celiac disease: A pilot study. Environ Res 186: 109439.
http://dx.doi.Org/10.1016/i.envres.2020.109439.
Gelman, A: Lee, D: Guo, L. (2015). Stan: a probabilistic programming language for bayesian inference
and optimization. American Educational Research Journal 40.
http://dx.doi.org/10.3102/1076998615606113.
Gibso L (1973). Plasma volume and glomerular filtration rate in pregnancy and their relation to
differences in fetal growth. Br J Obstet Gynaecol 80: 1067-1074.
http://dx.doi.org/10.1111/i.: 28.1973.tb02981.x.
Gleason, J, essie A. Cooper, K.,eith R. Klotz, J.,udith B. Post, GJoria B. Van Orden, G.,eorge New Jersey
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.nj.us/dep/watersupply/pdf/pfoa-appendixa.pdf.
Glynn, A: Berger, U; Bignert, A: Ullah, S; Aune, M; Lignell, S; Darnerud, PQ. (2012). Perfluorinated a Iky I
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.
Goodrich, JA: Walker, D: Lin, X: Wang, H: Lim, T: McConnell, R: Conti, DV: Chatzi, L: Setiawan, VW.
(2022). Exposure to perfluoroalkyl substances and risk of hepatocellular carcinoma in a
multiethnic cohort. JHEP Rep 4: 100550. http://dx.doi.Org/10.1016/i.ihepr.2022.100550.
Grandiean, P: Andersen, EW: Budtz-Jgrgensen, E: Nielsen, F: Mglbak, 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.2011.2Q34.
Grandiean, P: Bateson, T. (2021). RE: Benchmark analysis for PFAS immunotoxicity. Available online at
(accessed
Grandiean, P; Heilmann, C; Weihe, P; Nielsen, F; Mogensen, UB; Timmermann, A: Budtz-Jgrgensen, E.
(2017). Estimated exposures to perfluorinated compounds in infancy predict attenuated vaccine
This document is a draft for review purposes only and does not constitute Agency policy.
R-5 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
antibody concentrations at age 5-years. J Immunotoxicol 14: 188-195.
http://dx.doi.org/10.1080/1547691X.2017.136Q968.
Guo, J: Huang, S: Yang. L: Zhou, J: Xu, X: Lin, S: Li, H; Xie, X: Wu. S. (2023). Association between
polyfluoroalkyl substances exposure and sex steroids in adolescents: The mediating role of
serum albumin. Ecotoxicol Environ Saf 253: 114687.
http://dx.doi.Org/10.1016/i.ecoenv.2023.114687.
Gyllenhammar, I; Benskin, JP; Sandblom, O; Bert threns, L: Lignell, S: Wiberg, K; Glynn, A. (2018a).
Perfluoroalkyl acids (PFAAs) in serum from 2-4-month-old infants: Influence of maternal serum
concentration, gestational age, breast-feeding, and contaminated drinking water. Environ Sci
Technol 52: 7101-7110. http://dx.doi.org/10.1021/acs.est.8b00770.
Gyllenhammar, I; Diderholm, B; Gustafsson, J: Berger, U; Ridefelt, P: Benskin, JP: Lignell, S; Lampa, E;
Glynn, A. (2018b). Perfluoroalkyl acid levels in first-time mothers in relation to offspring weight
gain and growth. Environ Int 111: 191-199. http://dx.doi.Org/10.1016/i.envint.2017.12.002.
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/10.2307/16Q2514.
Hall, AP: Elcombe, CR: Foster, JR: Harada, T: Kaufmann, W: Knippel, A: Klittler, K: Malarkey, 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.
Hamm, MP: Cherry, NM: Chan, E: Martin, JW: Burstyn, I. (2010), Maternal exposure to perfluorinated
acids and fetal growth. J Expo Sci Environ Epidemiol 20: 589-597.
http://dx.doi.org/10.1038/ies.2009.57.
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.org/10.1021/acs.est.6b05811.
Haug, M; Dunder, L; Lind, PM; Lind, L; Salihovic, S. (2023). Associations of perfluoroalkyl substances
(PFAS) with lipid and lipoprotein profiles. J Expo Sci Environ Epidemiol.
http://dx.doi.org/10.1038/s41370-023-0Q545-x.
Higgins, JPT; Thomas, J: Chandler, J; Curripston, M; Li, T; Page, MJ; Welch, VA. (2022). Cochrane
handbook for systematic reviews of interventions version 6.3. Higgins, JPT; Thomas, J; Chandler,
J; Cumpston, M; Li, T; Page, MJ; Welch, VA. http://www.training.cochrane.org/handbook.
Hjermitslev, MH: Long, M: Wielsge, M; Bonefeld-J0rgensen, 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.
Hgjsager, FD: Andersen, M: Juul, A: Nielsen, F: Moller, S: Christensen, HT: Grgntved, A: Grandiean, P:
Jensen, TK. (2022). Prenatal and early postnatal exposure to perfluoroalkyl substances and bone
mineral content and density in the Odense child cohort. Environ Int 167: 107417.
http://dx.doi.Org/10.1016/i.envint.2022.107417.
Hong, A: Zhuang, L: Cui, W: Lu, Q: Yang, P: Su. S: Wang, B: Zhang, G: Chen I' (2022). Per- and
polyfluoroalkyl substances (PFAS) exposure in women seeking in vitro fertilization-embryo
transfer treatment (IVF-ET) in China: Blood-follicular transfer and associations with IVF-ET
outcomes. Sci Total Environ 838: 156323. http://dx.doi.Org/10.1016/i.scitotenv.2022.156323.
Huang, MC; Dzierlenga, AL; Robinson, VG; Waidyanatha, S; Devito, MJ: Eifrid, MA: Granville, CA; Gibbs,
ST: Blystone, CR. (2019). Toxicokinetics of perfluorobutane sulfonate (PFBS), perfluorohexane-1-
sulphonic acid (PFHxS), and perfluorooctane sulfonic acid (PFOS) in male and female
This document is a draft for review purposes only and does not constitute Agency policy.
R-6 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
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.
Huedo-Medina, IB; Sanchez-Meca, J; Mann-Martinez, F; Botella, J. (2006). Assessing heterogeneity in
meta-analysis: Q statistic or 12 index? Psychol Methods 11: 193-206.
http://dx.doi.Org/10.1037/1082-989X.ll.2.193.
Huo, X; Huang, R; Ga: ), K: Aimuzi, R; Nian, M: Ao, J: Feng, L: Tian, Y: Wang, W; Ye, W; Zhang, J.
(2020). Perfluoroalkyl substances in early pregnancy and risk of hypertensive disorders of
pregnancy: A prospective cohort study. Environ Int 138: 105656.
http://dx.doi.Org/10.1016/i.envint.2020.105656.
IPCS (International Programme on Chemical Safety). (2007). Harmonization project document no. 4: Part
2: IPCS framework for analysing the relevance of a non-cancer mode of action for humans.
Geneva, Switzerland: World Health Organization.
http://www.who.int/ipcs/methods/harmonization/areas/cancer mode.pdf?ua=l.
Ipsen, J. (1946). Circulating antitoxin at the onset of diphtheria in 425 patients. J Immunol 54: 325-347.
Jain, RB: Ducatman, A. (2019). Perfluoroalkyl acids and thyroid hormones across stages of kidney
function. Sci Total Environ 696: 133994. http://dx.doi.Org/10.1016/i.scitotenv.2019.133994.
Jain, 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.Org/10.1016/i.scitotenv.2022.156891.
Jensen, RC: Glintborg, h. I iinmermann, CAG; Nielsen, F: Bove, H; Madsen, JB: Bilenberg, N: Grandiean,
P: Jensen, TK: Andersen, MS. (2022). Higher free thyroxine associated with PFAS exposure in first
trimester. The Odense Child Cohort. Environ Res 212: 113492.
http://dx.doi.Org/10.1016/i.envres.2022.113492.
Ji, K: Kim, S: Kho, Y: Sakong, J: Paek, D: Choi, K. (2012). Major perfluoroalkyl acid (PFAA) concentrations
and influence of food consumption among the general population of Daegu, Korea. Sci Total
Environ 438: 42-48. http://dx.doi.Org/10.1016/i.scitotenv.2012.08.007.
Joshi-Barve, S; Kirpich, I: Cave, MC; Marsano, LS; McClain, CJ. (2015). Alcoholic, nonalcoholic, and
toxicant-associated steatohepatitis: Mechanistic similarities and differences [Review], CMGH 1:
356-367. http://dx.doi.Org/10.1016/i.icmgh.2015.05.006.
Kashino, I: Sasaki, S; Okada, E; Matsuura, H; Goudar; yashita, C; Okada, E; Ito, YM; Araki, A: Kishi,
FL (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.envint.2019.105355.
Kaur, K; Lesseur, C; Chen, L; Andra, SS; Narasimhan, S; Pulivarthi, D; Midya, V; Ma, Y; Ibroci, E: Gigase, F;
Lieber, M: Lieb, W: Janevic, T: De Witte, LP: Bergink, V: Rommel, AS: Chen, 1,(2023). Cross-
sectional associations of maternal PFAS exposure on SARS-CoV-2 IgG antibody levels during
pregnancy. Environ Res 219: 115067. http://dx.doi.Org/10.1016/i.envres.2022.115067.
Kim, JH; Park, HY: Jeon, JD: Kho, Y: Kim, SK: Park, MS: Hong, YC. (2015). The modifying effect of vitamin C
on the association between perfluorinated compounds and insulin resistance in the Korean
elderly: a double-blind, randomized, placebo-controlled crossover trial. Eur J Nutr 55: 1011-
1020. http://dx.doi.org/10.1007/sQ0394-015-0915-0.
Kim, JI: Kim, BN: Lee, YA: Shin, CH: Hong, YC: Passing, LP: Hildebrandt, G: Lim, YH. (2023a). Association
between early-childhood exposure to perfluoroalkyl substances and ADHD symptoms: A
prospective cohort study. Sci Total Environ 879: 163081.
http://dx.doi.Org/10.1016/i.scitotenv.2023.163081.
Kim, OJ; Kim, S; P JK; Jung, SK: Park, S; Hong, S; Jeon, HL; Kim, HJ; Park, B; Park, B: Kim, S; Kim,
fL (2023b). Exposure to serum perfluoroalkyl substances and biomarkers of liver function: The
This document is a draft for review purposes only and does not constitute Agency policy.
R-7 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Korean national environmental health survey 2015-2017. Chemosphere 322: 138208.
http://dx.doi.Org/10.1016/i.chemosphere.2023.138208.
Kim, Si; Heo, SH: Lee, PS: Hwang, IG: Lee, YB: Cho, HY. (2016). Gender differences in pharmacokinetics
and tissue distribution of 3 perfluoroalkyl and polyfluoroalkyl substances in rats. Food Chem
Toxicol 97: 243-255. http://dx.doi.org/10.lQ16/iict.2Q16.Qt
Kim, Si; Shin, H; Lee, YB: Cho, HY. (2018). Sex-specific risk assessment of PFHxS using a physiologically
based pharmacokinetic model. Arch Toxicol 92: 1113-1131. http://dx.doi.org/10.1007/sQ02Q4-
017-2116-5.
Kruschke, JK. (2021). Bayesian analysis reporting guidelines. Nat Hum Behav 5: 1282-1291.
http://dx.doi.org/10.1038/s41562-021-C
Kwon, EJ; Shin, iS; Kim, BM; Shah-Kulkarni, S; Park, H; Kho, YL; Park, EA; Kim, Yi; 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/iOM.000000000000Q739.
Larsen, A. (2022). R markdown code for the meta-analysis of PFHxS on birthweight [Computer Program],
Lee, E; Kinninger, A; Ursin, G; Tseng, C; Hurley, S; Wang, M; War rk, iS; Petreas, M; Deapen, D;
Reynolds, P. (2020). Serum levels of commonly detected persistent organic pollutants and per-
and polyfluoroalkyl substances (PFASs) and mammographic density in postmenopausal women.
Int i Environ Res Public Health 17: 606. http://dx.doi.org/10.3390/iierphl70206Q6.
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.
Lenters, V; Portengen, L; Rignell-Hvdbom, A; ionsson, BA; Lindh, CH; Piersma, AH; Toft, G; Bonde, iP;
Heederik, D; Rylander, 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.org/10.1289/ehp.1408933.
Li, A; Hou, i; Fu, J; Wang, Y; Hn > huan; i {> M; Song, M; iiang, G. (2023a). Association between
serum levels of TSH and free T4 and per- and polyfluoroalkyl compounds concentrations in
pregnant women, i Environ Sci 124: 11-18. http://dx.doi.Org/10.1016/i.ies.2021.10.026.
Li, B; Bao, Y; Xu, Y; Xie, S; Huang, i. (2017a). Vertical distribution of microbial communities in soils
contaminated by chromium and perfluoroalkyl substances. Sci Total Environ 599-600: 156-164.
http://dx.doi.Org/10.1016/i.scitotenv.2017.04.241.
Li, H; Chen, i; iingchao, L; Yang, i; Tah h I v f \n, Z; Ma, C; Liu, Y; Wang, L; Zhang, X; Guo, H.
(2023b). Association of exposure to perfluoroalkyl substances and risk of the acute coronary
syndrome: A case-control study in Shijiazhuang Hebei Province. Chemosphere 313: 137464.
http://dx.doi.Org/10.1016/i.chemosphere.2022.137464.
Li, i; Yang, L; He, G; Wang, B; Miao, M; ii, H; Wen, S; Cao, W; Yuan, W; Liang, H. (2022a). Association
between prenatal exposure to perfluoroalkyl substances and anogenital distance in female
neonates. Ecotoxicol Environ Saf 245: 114130. http://dx.doi.Org/10.1016/i.ecoenv.2022.114130.
Li, i; Yao, J; Xia, W; Dai, i; Liu, H; Pan, Y; Xu, S; Lu. S; iin, S; Li, Y; Sun, X; Zhang, B; Zheng, T; iiang, Y; iing,
L (2020). Association between exposure to per- and polyfluoroalkyl substances and blood
glucose in pregnant women. Inti Hyg Environ Health 230: 113596.
http://dx.doi.Org/10.1016/i.iiheh.2020.113596.
Li, K; Gao, P; Xiang, P; Zhang, X; Cui, X; Ma, LQ. (2017b). Molecular mechanisms of PFOA-induced toxicity
in animals and humans: Implications for health risks [Review], Environ Int 99: 43-54.
http://dx.doi.Org/10.1016/i.envint.2016.ll.014.
Li, M; Zeng, XW; Qian, ZM; Vaughn, MG; Sauve, S; Paul, G; Lin, S; Lu, L; Hu, LW; Yang, BY; Zhou, Y; Qin,
XD; Xu, SL; Bao, WW; Zhang, YZ; Yuan, P; Wang, i; Zhang, C; Tian, YP; Nian, M; Xiao, X; Fu, C;
This document is a draft for review purposes only and does not constitute Agency policy.
R-8 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Dong, GH. (2017c). Isomers of perfluorooctanesulfonate (PFOS) in cord serum and birth
outcomes in China: Guangzhou Birth Cohort Study. Environ Int 102: 1-8.
http://dx.doi.Org/10.1016/i.envint.2017.03.006.
Li. QQ: Huang. J: Cai. D: Chou. WC: Zeeshan. M: Chu. C: Zhou > L: Ma. HM: Tang. C: Kong. M: Xie. Y:
Dong, GH; Zeng, XW. (2023c). Prenatal exposure to legacy and alternative per- and
polyfluoroalkyl substances and neuropsychological development trajectories over the first 3
years of life. Environ Sci Technol 57: 3746-3757. http://dx.doi.org/10.1021/acs.est.2c07807.
Li, X: Song I * Shan, A; Huang, Y; Yang, h_ H > >»ng, Q; Yu, Y; Zheng, H: Cao, XC: Chen, D: Chen,
KX; Chen, X: Tang, NJ. (2022b). Perfluoroalkyl substances (PFASs) as risk factors for breast
cancer: a case-control study in Chinese population. Environ Health 21: 83.
http://dx.doi.org/10.1186/sl2940-022-0Q895-3.
Li, Y; Andersson, A: Xu, Y; Pineda, D: Nilsson, dh, CH; Jakobsson, K: Fletcher, T. (2022c).
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.Org/10.1016/i.envint.2022.107198.
Liang. JL: 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.
Liang. Y: Zhou. H: Zha >hen. W: Lei. L. (2023). Exposure to perfluoroalkyl and polyfluoroalkyl
substances and estimated glomerular filtration rate in adults: a cross-sectional study based on
NHANES (2017-2018). Environ Sci Pollut Res Int. http://dx.doi.org/10.1007/sll356-023-26384-9.
Liao. Q: Tang. P: Fan. H: Song. Y: Liang, J: Huang. H: Pan. D: Mo. M: Leilei: Lin. M: Chen. J: Wei. H: Long. J:
Shao, Y: Zeng, X: Liu, S: Huang, D: Qiu, 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.Org/10.1016/i.envpol.2023.121348.
Liao, Q; Tang, P; Pan, D; Song, Y; Lei, L; Liang, J: Liu, B; Lin, M; Huang, H; Mo, M; Huang, C; Wei, M; Liu, S;
Huang, D; Qiu, X. (2022a). Association of serum per- and polyfluoroalkyl substances and
gestational anemia during different trimesters in Zhuang ethnic pregnancy women of Guangxi,
China. Chemosphere 309: 136798. http://dx.doi.Org/10.1016/i.chemosphere.2022.136798.
Liao, Q; Tang, P; Song, Y; Liu, B; Huang, H; Liang, J: Lin, M; Shao, Y: Liu, S; Pan, D: Huang, D: Qiu, X.
(2022b). Association of single and multiple prefluoroalkyl substances exposure with preterm
birth: Results from a Chinese birth cohort study. Chemosphere 307: 135741.
http://dx.doi.Org/10.1016/i.chemosphere.2022.135741.
Liew, Z: Luo, J: Nohr, EA: Bech, BH: Bossi, R; Arah, OA: Olsen, J. (2020). Maternal plasma perfluoroalkyl
substances and miscarriage: a nested case-control study in the Danish National Birth Cohort.
Environ Health Perspect 128: 47007. http://dx.doi.org/10.1289/EHP6202.
Lin, M: Liao, Q: Tang, P: Song, Y: Liang, J, un: Li, J: Mu, C: Liu, S: Qiu, X: Yi. R, ui: Pang, Q: Pan, D: Zeng, X:
Huang, D. (2022). Association of maternal perfluoroalkyl substance exposure with postpartum
haemorrhage in Guangxi, China. Ecotoxicol Environ Saf 245: 114078.
http://dx.doi.Org/10.1016/i.ecoenv.2022.114078.
Linakis, MW: Landingham, CW: Longnecker, MP. Re-expressing coefficients from regression models for
inclusion in a meta-analysis. Linakis, MW: Landingham, CW: Longnecker, MP.
http://dx.doi.org/10.U01/2021.ll.02.466931.
Lind, DV; Priskorn, L; Lassen, TH; Nielsen, F; Kyhl, HB; Kristensen, DM: Christesen, HT; J0rgensen, JS;
Grandjean, P; Jensen, TK. (2017). Prenatal exposure to perfluoroalkyl substances and anogenital
This document is a draft for review purposes only and does not constitute Agency policy.
R-9 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
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, PM: Salihovic, S: Stubleski, J: Karrman, A; Lind, L (2018). Changes in plasma levels of perfluoroalkyl
substances (PFASs) are related to increase in carotid intima-media thickness over 10 years - a
longitudinal study. Environ Health 17: 59. http://dx.doi.org/10.1186/sl2940-018-0403-Q.
Liu, H; Pan, Y; Jin, S; Sun, X: Jiang, Y; Wang, Y; Ghassabian, A: Li, Y; Xia, W; Cui, Q; Zhang, B: Zhou, A: Dai,
(2020a). Associations between six common per- and polyfluoroalkyl substances and
estrogens in neonates of China. J Hazard Mater 407: 124378.
http://dx.doi.Org/10.1016/i.ihazmat.2020.124378.
Liu, J; Gao, X: Wang, Y; Leng, J: Li, J: Zhao, Y; Wu, Y. (2020b). Profiling of emerging and legacy per-
/polyfluoroalkyl substances in serum among pregnant women in China. Environ Pollut 271:
116376. http://dx.doi.Org/10.1016/i.envpol.2020.116376.
Liu, Y; Zhang, Z; Han, D: Zhao, Y: Yan, X: Cui, S. (2022). Association between environmental chemicals co-
exposure and peripheral blood immune-inflammatory indicators. Front Public Health 10:
980987. http://dx.doi.org/10.3389/fpubh.2022.980987.
Luo, D: Wu, WX: Pan, YA: Do. BB: Shen, MJ: 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.org/10.1021/acs.est.lc02670.
Luo, F: Chen, Q: Yu. G: Huo, X: Wang, H: Nian, M: Tian, Y: Xu, J: Zhang, J: Zhang, J. (2022a). Exposure to
perfluoroalkyl substances and neurodevelopment in 2-year-old children: A prospective cohort
study. Environ Int 166: 107384. http://dx.doi.Org/10.1016/i.envint.2022.107384.
Luo, K: Huang, W: Zhang, Q: Liu, X: Nian, M: Wei, M: Wang, Y: Chen, D: Chen, X: Zhang, J. (2022b).
Environmental exposure to legacy poly/perfluoroalkyl substances, emerging alternatives and
isomers and semen quality in men: A mixture analysis. Sci Total Environ 833: 155158.
http://dx.doi.Org/10.1016/i.scitotenv.2022.155158.
Luo, K; Liu, X: Zhou, W; Nian, M; Qiu, W; Yang, Y; Zhang, J. (2022c). Preconception exposure to
perfluoroalkyl and polyfluoroalkyl substances and couple fecundity: A couple-based exploration.
Environ Int 170: 107567. http://dx.doi.Org/10.1016/i.envint.2022.107567.
Ma, X; Cui, L; Cheu { hang, J: Zhang, X: Kang, Q; Jin I (2021). Parental plasma concentrations of
perfluoroalkyl substances and In Vitro fertilization outcomes. Environ Pollut 269: 116159.
http://dx.doi.Org/10.1016/i.envpol.2020.116159.
Maekawa, R; Ito, R; Iwasaki, Y; Saito, K; Akutsu, K; Takatori, S; Ishii, R; Kondo, F; Arai, Y; Ohgane, J;
Shiota, K; Makino, T: Sugino, N. (2017). Evidence of exposure to chemicals and heavy metals
during pregnancy in Japanese women. Reproductive Medicine and Biology 16: 337-348.
http://dx.doi.org/10.1002/rmb2.12049.
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.org/10.1289/ehp.1003096.
Manzano-Salgado, CB: Casas, M: Lopez-Espinosa, MJ: Ballester, F: Iniguez. C: Martinez, D: Costa, O:
Santa-Marina, L: Pereda-Pereda, E: Schettgen, T: Sunver, J: Vriiheid, M. (2017). 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.envint.2017.09.006.
Manzano-Salgado, CB: Granum, B; Lopez-Espinosa, MJ: Ballester, F; Iniguez, C; Gascon, M; Martinez, D;
Guxens, M: Basterretxea, M: Zabaleta, C: Schettgen, T; Sunver, J: Vriiheid, M: Casas, M. (2019).
Prenatal exposure to perfluoroalkyl substances, immune-related outcomes, and lung function in
This document is a draft for review purposes only and does not constitute Agency policy.
R-10 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
children from a Spanish birth cohort study. Int J Hyg Environ Health 222: 945-954.
http://dx.doi.org/10.1016/iJiheh.2019.06.005.
Marks, KJ; Cutler, AJ; Jedt arthstone, K: Kato, K: Hartman, TJ. (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, KJ: J eddy, Z; Flanders, WD: Northstone, K; Fraser, A; Calafat, AM: Kato, K; Hartman, TJ. (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.
Meek, ME: Boobis, A: Cote, I: Dellarco, V; Fotakis, G; Munn, S; Seed, J: Vickers, C. (2014). New
developments in the evolution and application of the WHO/IPCS framework on mode of
action/species concordance analysis [Review], J Appl Toxicol 34: 1-18.
http://dx.doi.org/10.1002/iat.2949.
Mehta, SS; James-Todd, T; Applebaum, KM: Bellavia, A: Coleman-Phox, K: Adler, N: Laraia, B: Epel, E:
Ei #ang, M: Park, JS: Zota, AR. (2021). Persistent organic pollutants and maternal
glycemic outcomes in a diverse pregnancy cohort of overweight women. Environ Res 193:
110551. http://dx.doi.Org/10.1016/i.envres.2020.110551.
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/10.3109/10408444.2Q15.1089471.
Meng, Q: Inoue, K: Ritz, B: Olsen, J: 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/iierphl5Q91832.
Mitro, SD: Sagiv, SK: Fleisch, AF: Jaacks, LM: Williams, PL: Rifas-Shiman, SL: Calafat, AM: Hivert, MF:
Oken, E; James-Todd, TM. (2020). Pregnancy per- and polyfluoroalkyl substance concentrations
and postpartum health in project viva: A prospective cohort. J Clin Endocrinol Metab 105:
e3415-e3426. http://dx.doi.org/10.1210/clinem/dgaa431.
Monroy, R; Morrison, K; Teo, K; Atkinson, S; Kubwabo, C; Stewart, B; Foster, WG. (2008). Serum levels of
perfluoroalkyl compounds in human maternal and umbilical cord blood samples. Environ Res
108: 56-62. http://dx.doi.Org/10.1016/i.envres.2008.06.001.
Morgan, S; Mottaleb, MA: Kraemer, MP: Moser, DK; Worley, J: Morris, AJ: Petriello, MC. (2023). Effect of
lifestyle-based lipid lowering interventions on the relationship between circulating levels of per-
and polyfluoroalkyl substances and serum cholesterol. Environ Toxicol Pharmacol 98: 104062.
http://dx.doi.Org/10.1016/i.etap.2023.104062.
Morken, NH: Travlos, GS: Wilson, RE: Eeeesbg, M: Longnecker, MP. (2014). Maternal glomerular
filtration rate in pregnancy and fetal size. PLoS ONE 9: el01897.
http://dx.doi.org/10.1371/iournal.pone.0101897.
Nian, M: Loo, K; Luo, F: Aimuzi, R; Huo, X: Chen, Q: Tian, Y: Zhang, J. (2020). Association between
prenatal exposure to PFAS and fetal sex hormones: Are the short-chain PFAS safer? Environ Sci
Technol 54: 8291-8299. http://dx.doi.org/10.1021/acs.est.0c02444.
Nilsson, S: Smurthwaite, K: Aylward, LL: Kay, M: Toms, LM: King, L: Marrington, S: Barnes, C: Kirk, MP:
Mueller, JF: Braunig, J. (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.iiheh.2022.114Q40.
Nilsson, S; Smurthwaite, K; Aylward, LL: Kay, M; Toms, LM: King, L; Marrington, S; Kirk, MP: Mueller, JF;
Braunig, J. (2022b). Associations between serum perfluoroalkyl acid (PFAA) concentrations and
health related biomarkers in firefighters. Environ Res 215: 114370.
http://dx.doi.Org/10.1016/i.envres.2022.114370.
This document is a draft for review purposes only and does not constitute Agency policy.
R-ll DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
NJDWQI (New Jersey Drinking Water Quality Institute). (2017). Public review draft: Health-based
maximum contaminant level support document: Perfluorooctane sulfonate (PFOS) (CAS #: 1763-
23-1; Chemical Formula: C8HF1703S). NJDWQI Health Effects Subcommittee.
https://cswab.org/wp-content/uploads/2018/09/New-Jersev-Scientific-Documentation-
Supporting-PFOS-13-ppt.pdf.
NIP (National Toxicology Program). (2018). 28-day evaluation of the toxicity (C06100) of
perfluorohexane sulfonate potassium salt (PFHSKsIt) (3871-99-6) on Harlan Sprague-Dawley rats
exposed via gavage. Available online at http://doi.org/10.22427/NTP-DATA-002-03267-00Q2-
0000-2 (accessed October 13, 2020).
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.
Oh, J: Bennett, DH: Tancredi, DJ: Caiafat, AM: Schmidt, RJ; Hertz-Picciotto, I: Shin, HM. (2022a).
Longitudinal Changes in Maternal Serum Concentrations of Per- and Polyfluoroalkyl Substances
from Pregnancy to Two Years Postpartum. Environ SciTechnol 56: 11449-11459.
http://dx.doi.org/10.1021/acs.est.lc07970.
Oh, J: Shin, HM: Kannan, K: Busgang, SA: Schmidt, RJ: Schweitzer, JB: Hertz-Picciotto, I: Bennett, PH.
(2022b). Childhood exposure to per- and polyfluoroalkyl substances and neurodevelopment in
the CHARGE case-control study. Environ Res 215: 114322.
http://dx.doi.Org/10.1016/i.envres.2022.114322.
Ouidir, M: Buck Louis, GM: Kanner, J: Grantz, KL: Zhang, C: Sundaram, R: Rahman, ML: Lee, S: Kannan, K:
Tekola-Avele, F: Mendola, P. (2020). Association of maternal exposure to persistent organic
pollutants in early pregnancy with fetal growth. JAMA Pediatr 174: 149-161.
http://dx.doi.org/10.1001/iamapediatrics.2019.51Q4.
Oulhote, Y; Coull, B; Bind, MA: Debes, F; Nielsen, F; Tamayo, I: Weihe, P; Grandjean, P. (2019). Joint and
independent neurotoxic effects of early life exposures to a chemical mixture: A multi-pollutant
approach combining ensemble learning and g-computation. Environmental Epidemiology 3:
e063. http://dx.doi.org/10.1 ).0000000000000063.
Padula, AM: Ning, X: Bakre, S; Barrett, ES; Bastain, T; Bennett, DH: Bloom, MS: Breton, CV; Dunlop, AL;
Eick, SM; Ferrara, A: Fleisch, A: Geiger, S; Goin, DE; Kannan, K; Karagas, MR: Korrick, S; Meeker,
JD; Morello-Frosch, R. (2023). Birth outcomes in relation to prenatal exposure to per-and
polyfluoroalkyl substances and stress in the environmental influences on child health outcomes
(echo) program [Supplemental Data], Environ Health Perspect 131: (037006) 037001-037011.
http://dx.doi.org/10.1289/EHP10723.
Pan, Z: Guo, Y: Zhou, Q: Wang, Q: Pan, S: Xu, S: Li, L. (2023). Perfluoroalkyl substance exposure is
associated with asthma and innate immune cell count in US adolescents stratified by sex.
Environ Sci Pollut Res Int 30: 52535-52548. http://dx.doi.org/10.1007/sll356-023-26Q65-7.
Park, SK: Wang, X: Ding, N: Karvonen-Gutierrez, CA; Caiafat, AM: Herman, WH: Mukheriee, B: Harlow,
SD. (2022). Per- and polyfluoroalkyl substances and incident diabetes in midlife women: the
Study of Women's Health Across the Nation (SWAN). Diabetologia 65: 1157-1168.
http://dx.doi.org/10.1007/sQ0125-022-05695-5.
Passen, EL: Andersen, BR. (1986). Clinical tetanus despite a protective level of toxin-neutralizing
antibody [Case Report], JAMA 255: 1171-1173.
http://dx.doi.org/10.1001/iama.1986.03370090093Q29.
Patel, JC; Mehta, BC. (1999). Tetanus: Study of 8,697 cases. Indian J Med Sci 53: 393-401.
This document is a draft for review purposes only and does not constitute Agency policy.
R-12 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Peterson, AK; Eckel, SP; Habre, R; Yang, T; Faham, D; Amiri, M; Grubbs, BH; Farzan, SF; Kaiiiiaii, K;
Robinson, M; Lerner, D; Al-Maravati, LA; Walker, DK; Grant, EG; Breton, CV; Bastain, TM. (2022).
Detected prenatal perfluorooctanoic acid (PFOA) exposure is associated with decreased fetal
head biometric parameters in participants experiencing higher perceived stress during
pregnancy in the MADRES cohort. 9. http://dx.doi.Org/10.1016/i.envadv.2022.100286.
Petroff, RL; Cavalcante, RG; Langen, ES; Dolinoy, DC; Padmanabhan, V; Goodrich, JM. (2023). Mediation
effects of DNA methylation and hydroxymethylation on birth outcomes after prenatal per- and
polyfluoroalkyl substances (PFAS) exposure in the Michigan mother-infant Pairs cohort. Clinical
Epigenetics 15: 49. http://dx.doi.org/10.1186/sl3148-023-Q1461-5.
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.
Porter, AK; Kleinschmidt, SE; Andres, KL; Reusch, CN; Krisko, RM; Taiwo, OA; Olsen, GW; Longnecker,
MP. (2022). Antibody response to COVID-19 vaccines among workers with a wide range of
exposure to per-and polyfluoroalkyl substances. Environ Int 169: 107537.
http://dx.doi.Org/10.1016/i.envint.2022.107537.
Ramhai, L; Hass, U; Boberg, J; Scholze, M; Christiansen, S; Nielse elstad, 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.org/10.1093/toxsci/kfv055.
Ramhai, L; Hass, U; Gilbert, ME; Wood, C; Svinge ai, 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, J; Viluksela, M; Kiviranta, H; Pihlajamaki, _L (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/sl2940-015-0Q66-z.
Reyes, L; Manalich, R. (2005). Long-term consequences of low birth weight [Review], Kidney Int Suppl
68: S107-S111. http://dx.doi.Org/10.llll/i.1523-1755.2005.09718.x.
Rivera-Nunez, Z; Kinkade, CW; Khoury, L; Brunner, J; Murphy, H; Wang, C; Kannan, K; Miller, RK;
O'Connor, TG; Barrett, ES. (2023). Prenatal perfluoroalkyl substances exposure and maternal sex
steroid hormones across pregnancy. Environ Res 220: 115233.
http://dx.doi.Org/10.1016/i.envres.2023.115233.
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.
Salvatier, J; Wiecki, TV; Fonnesbeck, C. (2016). Probabilistic programming in Python using PyMC3. PeerJ
Computer Science 2: e55. http://dx.doi 3eri-cs.55.
Sanghavi, M; Rutherford, JD. (2014). Cardiovascular physiology of pregnancy. Circulation 130: 1003-
1008. http://dx.doi.org/10.1161/CIRCULATIQNAHA.114.009029.
Schillemans, T; Iszatt, N; Remv, S; Schoeters, G; Fernandez, MF; D'Cruz, SC; Desalegn, A; Haug, LS;
Lignell, S; Lindroos, AK; Fabelova, L; Murinova, LP; Kosie alec. Z; Gabriel, C; Sarigiannis, D;
Pedraza-Diaz, S; Esteban-Lopez, M; Casta rio, A;... Akesson, A. (2022). Cross-sectional
associations between exposure to per- and polyfluoroalkyl substances and body mass index
among European teenagers in the HBM4EU aligned studies. Environ Pollut 316: 120566.
http://dx.doi.Org/10.1016/i.envpol.2022.120566.
This document is a draft for review purposes only and does not constitute Agency policy.
R-13 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Selgrade, MK. (2007). Immunotoxicity: The risk is real [Review], Toxicol Sci 100: 328-332.
http://dx.doi.org/10.1093/toxsci/kfm244.
Shearer, JJ; Callah ' ifat, AM: Huang, WY: Jones, RR: Sabbisetti, VS: Freedman, ND: Sampson, JN:
Silverman, DT: Purdue, MP: Hofmann, JN. (2021). Serum concentrations of per- and
polyfluoroalkyl substances and risk of renal cell carcinoma. J Natl Cancer Inst 113: 580-587.
http://dx.doi.org/10.1093/inci/djaal43.
Shi, S; Ding, Y; Wu, B; Hu, P; Chen, M; Dong, N: Vinturache, A; Gu, H: Dong, X: Ding, G. (2023).
Association of perfluoroalkyl substances with pulmonary function in adolescents (NHANES 2007-
2012). Environ Sci Pollut Res Int. http://dx.doi.org/10.1007/sll356-023-26119-w.
Shi, Y; Yang, L: Li, J: Lai, J: 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.
Shiue, I. (2015a). Arsenic, heavy metals, phthalates, pesticides, hydrocarbons and polyfluorinated
compounds but not parabens or phenols are associated with adult remembering condition: US
NHANES, 2011-2012. Environ Sci Pollut Res Int 22: 6381-6386.
http://dx.doi.org/10.1007/sll356-015-4261-9.
Shiue, I. (2015b). Urinary heavy metals, phthalates and polyaromatic hydrocarbons independent of
health events are associated with adult depression: USA NHANES, 2011-2012. Environ Sci Pollut
Res Int 22: 17095-17103. http://dx.doi.org/10.1007/sll356-015-4944-2.
Shiue, I. (2015c). Urinary heavy metals, phthalates, perchlorate, nitrate, thiocyanate, hydrocarbons, and
polyfluorinated compounds are associated with adult hearing disturbance: USA NHANES, 2011-
2012. Environ Sci Pollut Res Int 22: 20306-20311. http://dx.doi.org/10.1007/sll356-015-5546-8.
Shiue, I. (2015d). Urinary heavy metals, phthalates, phenols, thiocyanate, parabens, pesticides,
polyaromatic hydrocarbons but not arsenic or polyfluorinated compounds are associated with
adult oral health: USA NHANES, 2011-2012. Environ Sci Pollut Res Int 22: 15636-15645.
http://dx.doi.org/10.1007/sll356-015-4749-3.
Shoaff, J: Papandonatos, GD; Calafat, AM: Chen, A: Lanphear, BP: Ehrlich, S; Kelsey, KT; Braun, JM.
(2018). Prenatal exposure to perfluoroalkyl substances: Infant birth weight and early life growth.
Environmental Epidemiology 2: eOlO. http://dx.doi.org/10.1097/EE9.000000000000001Q.
Shoaff, JR. (2018). [Email to Michael Wright in response to a question about Shoaff et al. 2018],
Available online at (accessed
Sood, S; Ojo, AO: Adu, D; Kannan, K; Ghassabian, A: Koshy, T; Vento, SM; Pehrson, LJ; Gilbert, JF;
Arogundade, FA: Ademola, AD: Salako, BO: Raji, Y; Osafo, C; Antwi, S; Trachtman, H; Trasande, L;
Ajayi, S; Burke, D; Cooper, R; Gbadegesin, R; llori, T: Mamven, M; Olanrewaju, T: Parekh, R;
Rhule, J: Salako, T: Tayo, B: Ulasi, I: Investigators, HAKDRN. (2019). Association Between
Perfluoroalkyl Substance Exposure and Renal Function in Children With CKD Enrolled in H3Africa
Kidney Disease Research Network. 4: 1641-1645. http://dx.doi.Org/10.1016/i.ekir.2019.07.017.
Starling, AP: Adgate, JL: Hamman, RF: Kechris, K: Calafat, AM: Ye, X: Dabelea, D. (2017). Perfluoroalkyl
substances during pregnancy and offspring weight and adiposity at birth: Examining mediation
by maternal fasting glucose in the healthy start study. Environ Health Perspect 125: 067016.
http://dx.doi.org/10.1289/EHP641.
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.doi.org/10.1097/EDE.00000000000009Q3.
Sundstrom, M; Chang, SC; Noker, PE; Gorman, GS; Hart, JA: Ehresman, DJ: Bergman, A: Butenhoff, JL.
(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.
This document is a draft for review purposes only and does not constitute Agency policy.
R-14 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Takacs, ML; Abbott, BP. (2007). Activation of mouse and human peroxisome proliferator-activated
receptors (alpha, beta/delta, gamma) by perfluorooctanoic acid and perfluorooctane sulfonate.
Toxicol Sci 95: 108-117. http://dx.doi.org/10.1093/toxsci/kfll35.
Tan, Y: Zeng, Z: Liang, H: Weng, X: Ya en, J: Wei, X: Jing, C. (2022). Association
between Perfluoroalkyl and Polyfluoroalkyl Substances and Women's Infertility, NHANES 2013-
2016. Int J Environ Res Public Health 19. http://dx.doi.org/10.3390/ijerphl92215348.
Tian, M,Reichetzeder, -tocher, B, . (2019). Low birth weight, a risk factor for diseases in
later life, is a surrogate of insulin resistance at birth. J Hypertens 37: 2123-2134.
http://dx.doi.org/10.1097/HJH.00000000000Q2156.
Tillaut, H; Monfort, C; Giton, F; Warembourg, C; Rouget, F; Cordier, S; Laine, F; Gaudreau, E; Garlantezec,
R; Saint-Amour, D: Chevrier, C. (2022). Persistent organic pollutant exposure and thyroid
function among 12-year-old children. Neuroendocrinology.
http://dx.doi.org/10.1159/000528631.
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). (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). (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.
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-plain.pdf.
U.S. EPA (U.S. Environmental Protection Agency). (2019). CompTox Chemicals Dashboard [Database],
Research Triangle Park, NC. Retrieved from https://comptox.epa.gov/dashboard
Valvi, D; Oulhote, Y; Weihe, P; Dalgard, C; Bjerve, KS; Steuerwald, U; Grandjean, P. (2017). Gestational
diabetes and offspring birth size at elevated environmental pollutant exposures. Environ Int 107:
205-215. http://dx.doi.Org/10.1016/i.envint.2017.07.016.
van Larebeke, N; Koppen, G; Decraemer, S, am: Colles, A, nn; Bruckers, L; Den Hond, E; Govarts, E, va;
Morrens, B; Schettgen, T; Remy, S: Coertjens, D: Nawrot, T, im; Nelen, V; Baeyens, W; Schoeters,
CI (2022). Per- and polyfluoroalkyl substances (PFAS) and neurobehavioral function and
cognition in adolescents (2010-2011) and elderly people (2014): results from the Flanders
Environment and Health Studies (FLEHS). Environ Sci Eur 34. http://dx.doi.org/10.1186/sl2302-
022-00675-3.
Verner, MA: Ngueta, G: Jensen, ET: Fromme, H: Voelkel, W: Nygaard, UC: Granum, B: Longnecker, MP.
(2016). A simple pharmacokinetic model of prenatal and postnatal exposure to perfluoroalkyl
substances (PFASs). Environ Sci Technol 50: 978-986.
http://dx.doi.org/10.1021/acs.est.5b04399.
Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. J Stat Softw 36: 1-48.
Wahlang, B: Beier, Jl; Clair, HB: Bellis-Jones, HJ: Falkner, K: Mcclain, G: Cave, MC. (2013). Toxicant-
associated steatohepatitis [Review], Toxicol Pathol 41: 343-360.
http://dx.doi.org/10.1177/0192623312468517.
This document is a draft for review purposes only and does not constitute Agency policy.
R-15 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
Wang, B; Chen, Q; Shen, L; Zhao, S; Pang, W: Zhang, J. (2016). Perfluoroalkyl and polyfluoroalkyl
substances in cord blood of newborns in Shanghai, China: Implications for risk assessment
[Review], Environ Int 97: 7-14. http://dx.doi.Org/10.1016/i.envint.2016.10.008.
Wang. H: Li, W: Yang, J: Wan;-, >, hn, H: Han, M: Xu, t; Liu, S: Yi. J: Chen, Y: Jiang, Q: He, G. (2022).
Gestational exposure to perfluoroalkyl substances is associated with placental DNA methylation
and birth size. Sci Total Environ 858: 159747.
http://dx.doi.Org/10.1016/i.scitotenv.2022.159747.
Wang, W; Zhou, W; Wu, S; Liang, F; Li, Y; Zhang, J: Cui, L; Feng, Y; Wang, Y. (2019). Perfluoroalkyl
substances exposure and risk of polycystic ovarian syndrome related infertility in Chinese
women. Environ Pollut 247: 824-831. http://dx.doi.Org/10.1016/i.envpol.2019.01.039.
Watanabe, S. (2010). Asymptotic equivalence of Bayes cross validation and widely applicable
information criterion in singular learning theory. J Mach Learn Res 11: 3571-3594.
Weisskopf, MG; Seals, RM; Webster, IF. (2018). Bias amplification in epidemiologic analysis of exposure
to mixtures. Environ Health Perspect 126. http://dx.doi.org/10.1289/EHP2450.
Weisskopf, MG: Webster, IF. (2017). Trade-offs of personal versus more proxy exposure measures in
environmental epidemiology. Epidemiology 28: 635-643.
http://dx.doi.org/10.1097/EDE.000000000000Q686.
Whitworth, KW: Haug, LS: Baird, DP: Becher, G: Hoppin, JA: Eggesbo, M: Skiaerven, R: Thomsen, C:
Longnecker, MP. (2012). P60—Preliminary results on subfecundity and plasma concentrations of
perfluorinated compounds during pregnancy in the Norwegian Mother and Child Cohort Study,
2002-2004. Reprod Toxicol 33: 620-621. http://dx.doi.Org/10.1016/i.reprotox.2011.ll.094.
Whitworth, KW: Haug, LS: Sabaredzovic, A: Eggesbo, M: Longnecker, MP. (2016). Brief Report: Plasma
Concentrations of Perfluorooctane Sulfonamide and Time-to-pregnancy Among Primiparous
Women. Epidemiology 27: 712-715. http://dx.doi.org/10.1097/EDE.000000000000Q524.
Wielsge, M; Eiberg, H; Ghisari, M; Kern, P; Lind, O; Bonefeld-J0rgensen, EC. (2018). Genetic variations,
exposure to persistent organic pollutants and breast cancer risk - a greenlandic case-control
study. Basic & Clinical Pharmacology & Toxicology Online Pharmacology Online 123: 335-346.
http://dx.doi.org/10.llll/bcpt.130Q2.
Wielsge, M; Kern, P; Bonefeld-J0rgensen, EC. (2017). Serum levels of environmental pollutants is a risk
factor for breast cancer in Inuit: a case control study. Environ Health 16: 56.
http://dx.doi.org/10.1188/sl294Q-Qiy-Q289-8.
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/s41390-019-072Q-l.
Woods, MM: Lanphear, BP: Braun, JM: 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: Morae-, U, f\ landhane, Pi: Turvev, SE: Subbarao, P: Brook, JR:
Sears, MR: Wong, CS. (2019). Associations between concentrations of perfluoroalkyl substances
in human plasma and maternal, infant, and home characteristics in Winnipeg, Canada. Environ
Pollut 249: 758-766. http://dx.doi.Org/10.1016/i.envpol.2019.03.054.
Wright, JM: Larsen, A: Rappazzo, K: F adke, EG: Bateson, TF. (2023). Systematic review and meta-
analysis of birthweight and PFNA exposures. Environ Resll5357.
http://dx.doi.Org/10.1016/i.envres.2023.115357.
Wright, M; Larsen, A. (2022). Email correspondence with study author, Cyntia Manzano-Salgado.
Available online at (accessed
Xiong, X: Chen, B; Wang, Z; Ma, L; Li, S; Ł (2022). Association between perfluoroalkyl substances
concentration and bone mineral density in the US adolescents aged 12-19 years in NHANES
This document is a draft for review purposes only and does not constitute Agency policy.
R-16 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
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
determinants and associations with birth outcomes. J Hazard Mater 380: 120867.
http://dx.doi.Org/10.1016/i.ihazmat.2019.120867.
Xu, C; Zhang, L; Zhou, Q; Ding, J: Yin, S; Shang, X: Tia (2022). Exposure to per- and polyfluoroalkyl
substances as a risk factor for gestational diabetes mellitus through interference with glucose
homeostasis. Sci Total Environ 838: 156561. http://dx.doi.Org/10.1016/i.scitotenv.2022.156561.
H hou, Q; Zhang, J; Chen, X: Zhao, H; Lu, H; Ma, B: Wang, I; Wu, C: Ying, C: Xiong, Y; Zhou, Z: Li, X.
(2020). Exposure to elevated per- and polyfluoroalkyl substances in early pregnancy is related to
increased risk of gestational diabetes mellitus: A nested case-control study in Shanghai, China.
Environ Int 143: 105952. http://dx.doi.Org/10.1016/i.envint.2020.105952.
Yang, L; Ji, H; Lian,. H > nan, W; Song, X; Li, X; Niu, J; Shi, H; Wen, S: Miao, M. (2022). Associations of
perfluoroalkyl and polyfluoroalkyl substances with gestational hypertension and blood pressure
during pregnancy: A cohort study. Environ Res 215: 114284.
http://dx.doi.Org/10.1016/i.envres.2022.114284.
Yao, J: Pan, Y: Sheng, N: So, Z: Guo, Y: Wang, J: Dai, J. (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 Technol 54:
13389-13398. http://dx.doi.org/10.1021/acs.est.0c02888.
Yao, Q: Gao, Y: Zhang, Y: Qin, K: Liew, Z: Tian, Y. (2021). Associations of paternal and maternal per- and
polyfluoroalkyl substances exposure with cord serum reproductive hormones, placental
steroidogenic enzyme and birth weight. Chemosphere 285: 131521.
http://dx.doi.Org/10.1016/i.chemosphere.2021.131521.
Ye, WL: Chen, ZX: Xie, YQ: Kong, ML: Li, QQ: Yu, S: Chu, C: Dong, GH: Zeng, XW. (2021). Associations
between serum isomers of perfluoroalkyl acids and metabolic syndrome in adults: Isomers of C8
Health Project in China. Environ Res 196: 110430.
http://dx.doi.Org/10.1016/i.envres.2020.110430.
Yu, Y: Qin, XD: Bloom, MS: Chu, C: Dai, X: Li, QQ: Chen, ZX: Kong, ML: Xie, YQ: Meng, WJ: Yang, BY: Hu,
LW; Zeng, XW: Zhao, XM; Zhou, Y; Dong, GH. (2022). Associations of prenatal exposure to
perfluoroalkyl substances with preterm birth: A family-based birth cohort study. Environ Res
214: 113803. http://dx.doi.Org/10.1016/i.envres.2022.113803.
Zeng, X: Cher * ^ * !hao, J; Chen, Q; Yu, Z; Zhang, Y; Han, L; Chen, Y; Zhang, J. (2023). In utero
exposure to perfluoroalkyl substances and early childhood BMI trajectories: A mediation
analysis with neonatal metabolic profiles. Sci Total Environ 867: 161504.
http://dx.doi.Org/10.1016/i.scitotenv.2023.161504.
Zhang, JJ. (2022). RE: Prenatal exposure to perfluoroalkyl and polyfluoroalkyl substances and birth
outcomes: A longitudinal cohort with repeated measurements. Available online at (accessed
Zhang, Y: Chen, R: Gao, Y: Qu. J: War ao. M: Bai, X: Jin, H. (2023a). Human serum poly- and
perfluoroalkyl substance concentrations and their associations with gestational diabetes
mellitus. Environ Pollut 317: 120833. http://dx.doi.Org/10.1016/i.envpol.2022.120833.
Zhang, Y: Mustieles, V: Sun, Y: Oulhote, Y: Wang, YX: Messerlian, C. (2022). Association between serum
per- and polyfluoroalkyl substances concentrations and common cold among children and
adolescents in the United States. Environ Int 164: 107239.
http://dx.doi.Org/10.1016/i.envint.2022.107239.
Zhang, Y; Mustieles, V; Wang, YX: Sun, Q; Coull, B; Sun, Y; Slitt, A: Messerlian, C. (2023b). Red blood cell
folate modifies the association between serum per- and polyfluoroalkyl substances and
This document is a draft for review purposes only and does not constitute Agency policy.
R-17 DRAFT-DO NOT CITE OR QUOTE
-------
Supplemental Information— Perfluorohexanesulfonic Acid and Related Salts
antibody concentrations in U.S. adolescents. Environ Sci Technol 57: 2445-2456.
http://dx.doi.org/10.1021/acs.est.2cQ7152.
Zhao, X: Lin, JY: 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/s41370-022-0Q452-7.
Zhou, Y; Li, Q; Wang, P; Li, J; Zhao, W; Zhang, L; Wang, H; Cheng, Y; Shi, Zhang, Y. (2023).
Associations of prenatal PFAS exposure and early childhood neurodevelopment: Evidence from
the Shanghai Maternal-Child Pairs Cohort. Environ Int 173: 107850.
http://dx.doi.Org/10.1016/i.envint.2023.107850.
Zhu, Y; Bartell, SM. (2022). Per- and polyfluoroalkyl substances in drinking water and hypertensive
disorders of pregnancy in the United States during 2013-2015. Environmental Epidemiology 6:
e209. http://dx.doi.org/10.1097/EE9.00000000000002Q9.
This document is a draft for review purposes only and does not constitute Agency policy.
R-18 DRAFT-DO NOT CITE OR QUOTE
------- |