® rPA EPA-823-R-18-307
\/Cr# \ Public Comment Draft
Human Health Toxicity Values for
Perfluorobutane Sulfonic Acid
(CASRN 375-73-5)
and Related Compound
Potassium Perfluorobutane Sulfonate
(CASRN 29420-49-3)
This document is a Public Comment draft. It has not been formally released by the U.S.
Environmental Protection Agency and should not at this stage be construed to represent Agency
policy. This information is distributed solely for the purpose of public review.
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Human Health Toxicity Values for
Perfluorobutane Sulfonic Acid (CASRN 375-73-5) and Related Compound
Potassium Perfluorobutane Sulfonate (CASRN 29420 49 3)
Prepared hy
I S I ji\ironmeiiuil Proleclit>n Auency
Office of Research and l)e\elopment (8101R)
National Cenier fur I -n\ironmeiital Assessment
Washington. DC 2<)460
l-P.\ Dociinienl Number S23-R-18-307
NON LMBLR 2iU8
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Disclaimer
This document is a public comment draft for review purposes only. This information is
distributed solely for the purpose of public comment. It has not been formally disseminated by
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.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Authors, Contributors, and Reviewers
CHEMICAL MANAGERS
Jason C. Lambert, PhD, DABT
Elizabeth Oesterling Owens, PhD
CONTRIBUTORS
Michelle Angrish, PhD
Xabier Arzuaga, PhD
Johanna Congleton, PhD
Ingrid Druwe, PhD
J. Allen Davis, MS
Kelly Garcia, BS
Carolyn Gigot, BA
Andrew Greenhalgh, BS
Belinda Hawkins, PhD, DABT
Ryan Jones, MSLS
Andrew Kraft, PhD
April Luke, PhD
Elizabeth Radke, PhD
Michele Taylor, PhD
Samuel Thaeker. B A
Andre WeaNcr. I'lil)
Amina Wilkins. MPI I
Michael Wright. Scl)
DRAFT DOCl MI NT PUI.PAUI.I) BY
I S I ji\iionmental Protection Agency. Office of Research and Development, National
Center for Environmental Assessment (jNCEA)
EXEC I I IM. direction
Tina liahadori. ScD NCEA Center Director
Mary Ross. Phi) NCEA Deputy Center Director
Samantha Jones. Phi) NCEA Associate Director
Emma Lavoie, PhD NCEA Assistant Center Director for Scientific Support
PRIMARY INTERNAL REVIEWERS
David Bussard, PhD
Kris Thayer, PhD
Questions regarding the contents of this document may be directed to the National Center for
Environmental Assessment in the U.S. Environmental Protection Agency's Office of Research
and Development.
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
tents
1.0 Background 1
1.1 Physical and Chemical Properties 1
1.2 Occurrence 2
1.3 Toxicokinetics 3
1.3.1 Overview 3
1.3.2 Absorption 5
1.3.3 Distribution 5
1.3.4 Metabolism 6
1.3.5 Elimination 6
1.3.6 Summary 8
2.0 Problem Formulation 9
2.1 Conceptual Model 9
2.2 Objective 11
2.3 Methods 11
2.3.1 Literature S earch 11
2.3.2 Screening Process 11
2.3.3 Study Evaluation 12
2.3.4 Data Extraction 14
2.3.5 Evidence Synthesis 15
2.3.6 Evidence Integration and I la/aid Characterization 15
2.3.7 Deri\ation of Values 17
3.0 Overview of Evidence Identification for Synthesis and Dose-Response Analysis 20
3.1 I.itcialuie Search and Screening Results 20
3.2 Siudy l-\ aluation Results 22
4.0 Evidence Synthesis: Overview of Included Studies 25
4.1 Thyroid Effects 25
4.1.1 Human Studies 25
4.1.2 Animal Studies 25
4.2 Reproducli\e l-ffecls 26
4.2.1 Human Studies 26
4.2.2 Animal Studies 27
4.3 Offspring Growth and Early Development 30
4.3.1 Human Studies 30
4.3.2 Animal Studies 30
4.4 Renal Effects 31
4.4.1 Human Studies 31
4.4.2 Animal Studies 31
4.5 Hepatic Effects 33
iii
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
4.5.1 Human Studies 33
4.5.2 Animal Studies 33
4.6 Lipids and Lipoproteins 34
4.6.1 Human Studies 34
4.6.2 Animal Studies 34
4.7 Other Effects 35
4.7.1 Human Studies 35
4.7.2 Animal Studies 35
4.8 Other Data 35
4.8.1 Tests Evaluating Genotoxicity and Mutagenicity 38
5.0 Evidence Integration and Hazard Characterization 39
5.1 Thyroid Effects 42
5.2 Developmental Effects 42
5.3 Reproductive Effects 43
5.4 Renal Effects 44
5.5 Hepatic Effects 45
5.6 Effects on Lipid or Lipoprotein I lomeostasis 46
5.7 Immune Effects 46
5.8 Evidence Integration and 1 lazard Characterization Summary 46
6.0 Derivation of Values 55
6.1 Derivation of Oral Reference Doses 55
6.1.1 Deri\alion of Candidate Subchronic RfDs 55
6.1.2 Derivation of Candidate Chronic RlDs 65
6.2 Derivation of Inhalation Reference Concentrations 69
6.3 Cancer \\ eiuht-ol-l-\ idence Descriptor and Derivation of Cancer Risk Values 69
6.4 Susceptible Populations and Life Stages 69
Appendix A: Literature Search Strategy A-l
Appendix IS: Detailed PECO Criteria B-l
Appendix C: Study Evaluation Methods C-l
Appendix D: HAWC I ser Cuide and Frequently Asked Questions D-l
Appendix E. Additional Data l-'igures E-l
Appendix F. Benchmark Dose Modeling Results F-l
Appendix G. References G-l
iv
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Figures
Figure 1. Chemical structures of PFBS and K+PFBS 1
Figure 2. Conceptual model for PFBS and/or potassium salt 10
Figure 3. Approach for evaluating epidemiological and animal toxicology studies 14
Figure 4. Literature search and screening flow diagram for PFBS (CASRN 375-73-5) 21
Figure 5. Evaluation results for epidemiological studies assessing effects of PFBS (click to
see interactive data graphic for rating rationales) 23
Figure 6. Evaluation results for animal studies assessing cfllvls of PFBS exposure (click to
see interactive data graphic for rating rationales) 24
Figure 7. Thyroid effects from K+PFBS exposure (click lo see interactive data graphic and
rationale for study evaluations for effects on the thyroid in 1I AWC) 26
Figure 8. Reproductive hormone response to k PI liS exposure (click lo sec interactive
data graphic and rationale for study e\ alualions for reproducli\ e hormone levels
in 11 AWC) " 29
Figure 9. Effects to reproductive de\ elopment and estrous cycling following PFBS
exposure (click to see ink-met i\ e data graphic) 30
Figure D-l. HAWC homepage for the public Pl-'liS assessment D-l
Figure D-2. Representati\e study list... D-2
Figure D-3. Representati\e study e\aluation pie chart with the reporting domain selected
and text populating to the l ight of pie chart D-3
Figure D-4A. Visualization example for Pl-'liS (Note that the records listed under each
column (study, experiment cndpoinl. units, study design, observation time, dose)
and data within the plot are interacti\e ) D-4
Figure I)—Hi Example pop-up w indow after clicking on interactive visualization links. (In
I 'igure D-4A the red circle for study MP (2011); male at a dose of 500 mg/kg-
day was clicked leading to the pop-up shown above. Clicking on blue text will
open a new window with descriptive data.) D-4
Figure D-5. Representati\ e data download page D-5
Figure D-6A. Example BMI) modeling navigation D-6
Figure D-6B. Example BMD session D-6
Figure E-l. Serum free and total thyroxine (T4) response in animals following K+PFBS
exposure (click to see interactive data graphic) E-l
Figure E-2. Serum total triiodothyronine (T3) response in animals following K+PFBS
exposure (click to see interactive data graphic) E-2
Figure E-3. Serum thyroid-stimulating hormone (TSH) response in animals following
K+PFBS exposure (click to see interactive data graphic) E-3
v
-------
This document is a draft for review purposes only and does not constitute Agency policy.
Figure E-4. Developmental effects (eye opening) following K+PFBS in rats (click to see
interactive data graphic) E-3
Figure E-5. Developmental effects (first estrus) following K+PFBS in rats (click to see
interactive data graphic) E-4
Figure E-6. Developmental effects (vaginal patency) following K+PFBS in rats (click to
see interactive data graphic) E-4
Figure E-7. Kidney histopathological effects following K+PFBS in rats (click to see
interactive data graphic) E-5
Figure E-8. Renal effects following K+PFBS in rats (click lo see interactive data graphic) E-6
Figure E-9. Kidney weight effects following K+PFBS in rals (click to see interactive data
graphic) E-7
Figure E-10. Liver effects following K+PFBS in rals (click lo see interactive data graphic) E-8
Figure E-l 1. Effects on lipids and lipoproteins following K+PFBS in rals and inice (click to
see interactive data graphic) E-9
Figure F-l. Candidate PODs for the derivation of 1 lie sulxiironic and chronic Rfl)s for
PFBS (click to see interacli\ e data graphic) F-3
Figure F-2. Exponential (Model 4) lor total T4 in P\TD I female offspring (litter n)
exposed GDs 1-20 (Feng et a I (2<>I7) F-6
Figure F-3. Dichotomous-TTill model for papillary iLihular ductal epithelium hyperplasia in
Po female rats I .iecicr el al. F-18
Tab
Tabic I Ph\sicochemical properties ol'Pl'BS (( \\SRN 375-73-5) and related compound
k PI liS (C.\SR\ 2^420-49-3) 2
Tabic 2 Summary of toxicokinetics of scrum PFBS (mean ± standard error) 4
Table 3. Criteria for overall e\ idence integration judgments 16
Table 4. Epidemiological studies excluded based on study evaluation 22
Table 5. Other studies 36
Table 6. Summary of noncancer data for oral exposure to PFBS (CASRN 375-73-5) and
the related compound K+PFBS (CASRN 29420-49-3) 40
Table 7. Summary of hazard characterization and evidence integration judgments 47
Table 8. PODs considered for the derivation of the subchronic RfD for K+PFBS
(CASRN 29420-49-3) 57
Table 9. UFs for the candidate subchronic RfD for thyroid effects for K+PFBS
(CASRN 29420-49-3) 62
vi
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Table 10. UFs for the candidate subchronic RfD for kidney effects for K+PFBS
(CASRN 29420-49-3) 63
Table 11. Summary of candidate noncancer subchronic reference values for K+PFBS
(CASRN 29420-49-3) 64
Table 12. Confidence descriptors for candidate subchronic RfD (Thyroid Effects) for
PFBS (CASRN 375-73-5) and the related compound K+PFBS
(CASRN 29420-49-3) 65
Table 13. Confidence descriptors for the candidate subchronic RfD (Kidney Effects) for
PFBS (CASRN 375-73-5) and the related com pound k PFBS
(CASRN 29420-49-3) 65
Table 14. UFs for the candidate chronic RfD for thyroid for k PI liS
(CASRN 29420-49-3) 66
Table 15. UFs for the candidate chronic RfD for kidney effects for k PI liS
(CASRN 29420-49-3) 67
Table 16. Summary of candidate noncancer chronic reference values for Pi"15S
(CASRN 375-73-5) and related compound k PI liS (CASRN 2942i)-4l)-3) 68
Table 17. Confidence descriptors lor candidate chronic Rll) (Thyroid Effects) for PFBS
(CASRN 375-73-5) and the related compound k PI liS (CASRN 29420-49-3) 68
Table 18. Confidence descriptors lor the candidate chronic Rll) (Kidney Effects) for PFBS
(CASRN 375-73-5) and the related compound k PFBS (CASRN 29420-49-3) 69
Table A-l. Synonyms and MI-SI I terms A-l
Table B-l. Population, exposure, comparator, and outcome criteria B-l
Table C-l Questions used to guide the de\elopment of criteria for each domain in
epidemiology studies C-2
Table C-2 Criteria lor e\aluation of exposure measurement in epidemiology studies C-5
Table C-3 Ouestions used to uuide the development of criteria for each domain in
experimental animal toxicology studies C-7
Table F-l. Candidate PODs lor the derivation of the subchronic and chronic RfDs for
PFBS (C.\SR\ 375-73-5) and the related compound K+PFBS
(CASRN 2lM2<)—N-3) F-l
Table F-2. Modeling results for total T4 in PND 1 female offspring (litter n) exposed
GDs l-20a 1-5
Table F-3. Modeling results for papillary tubular/ductal epithelium hyperplasia in Po
female ratsa F-l 7
vii
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Commonly Used Abbreviations and Acronyms
AEC
absolute eosinophil count
NHANES
National Health and Nutrition
AFFF
Aqueous Film-Forming Foam
Examination Survey
AIC
Akaike's information criterion
NOAEL
no observed adverse effect level
ALT
alanine aminotransferase
NTP
National Toxicology Program
AST
aspartate aminotransferase
NZW
New Zealand White (rabbit breed)
AUC
area under the curve
OR
odds ratio
BMD
benchmark dose
PECO
population, exposure, comparator,
BMDL
benchmark dose lower confidence
oulcome
limit
PFAA
peril uoroalkyl acid
BMDS
Benchmark Dose Software
PFAS
per- and polyfluoroalkyl substances
BMR
benchmark response
PI O A
peril uorooctanoic acid
BUN
blood urea nitrogen
PI OS
peril uorooctane sulfonic acid
BW
body weight
PI liS
perfluorobutane sulfonic acid
CA
chromosomal aberration
PI 1 [xA
perfluorolievinoic acid
CASRN
Chemical Abstracts Service Registry
P\l)
postnatal day
Number
POD
poini of departure
CHO
Chinese hamster ovary (cell line cells)
RD
re la live deviation
CPN
chronic progressive nephiopalln
Rf(
inhalation reference concentration
D3
deiodinase 3
RID
oral reference dose
DAF
dosimetric adiuslmcnl factor
ROS
reactive oxygen species
DNA
deoxyribonucleic acid
1 13
re\ erse triiodothyronine
ECP
eosinophilic caliomc pioiein
SI)
standard deviation
EPA
U.S. Em iionmcnlal Pioieclion
Aijcnc\
r:
13
3,5 -diiodo-L-thyronine
triiodothyronine
GD
ijcslalion da\
14
thyroxine
HAWC
1 leallli Asscssmcnl Workspace
TBG
thyroid binding globulin
( ollalioialive
TSH
thyroid-stimulating hormone
HED
human a|uivalenl dose
TTR
transthyretin
HPT
hypol 1 ui la 111 i c-pil u 11 a i\ -1 h yroid
UF
uncertainty factor
i.v.
intravenous
UFa
interspecies uncertainty factor
ICR
Institute ol'Cancer Research
UFc
composite uncertainty factor
KPFBS
potassium periluorolnitane sulfonate
UFd
database uncertainty factor
LD50
median lethal dose
UFh
intraspecies uncertainty factor
LOAEL
lowest observed adverse effect level
UFl
LOAEL-to-NOAEL uncertainty
MW
molecular weight
factor
NCEA
National Center for Environmental
Assessment
UFS
VLDL
subchronic-to-chronic uncertainty
factor
very low density lipoprotein
viii
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Execul
Summary of Occurrence and Health Effects
The U.S. Environmental Protection Agency (EPA) is issuing draft subchronic and chronic oral
toxicity values for perfluorobutane sulfonic acid (PFBS) (Chemical Abstracts Service Registry
Number [CASRN] 375-73-5) and its related salt, potassium perfluorobutane sulfonate (K+PFBS)
(CASRN 29420-49-3) for public comment. The toxicity assessment for PFBS is a scientific and
technical report that includes toxicity values associated with potential noncancer health effects
following oral exposure (in this case, oral reference doses (RfDs). This assessment evaluates
human health hazards. The toxicity assessment and the values contained within is not a risk
assessment as it does not include an exposure assessment nor an overall risk characterization.
Further, the toxicity assessment does not address the legal, political, social, economic, or
technical considerations involved in risk management. W'licn final, the PFBS toxicity assessment
can be used by EPA, states, tribes, and local communities, along with specific exposure and
other relevant information, to determine, under the appropriate regulations and statutes, if, and
when, it is necessary to take action to address potential risk associated with human exposures to
PFBS.
PFBS and K+PFBS are both four-carhon. Ililly fluorinaled alkane members of a large and diverse
class of linear and branched compounds known as "per- and pol\ lluoroalkyl substances," or
PFAS. In the early 2000s, concerns grew o\er the environmental persistence, long half-lives in
humans, and bioaccumulation potential of longer chain PI AS. in particular perfluorooctanoic
acid (PFOA) and pcrlluoiooclane sulfonic acid (PI OS) As a result, shorter chainPFAS such as
PFBS were developed and integrated into various consumer products and applications, as this
compound has the desired properties and characteristics associated with this class of compounds
with faster elimination from the body than PFOA and PFOS. PFBS has been found in food
contact materials, dust, and source and finished drinking water. It is also associated with
Aqueous I'ilm-I'orming Foams and used during chrome electroplating as a mist suppressant. As
such, oral intake of water and food, inhalation, and dermal contact are plausible modes of PFBS
exposure, with the oral route being the primary route of exposure. PFBS has been detected in
human urine, confirming exposure to this PFAS; however, the magnitude of human exposure
likely depends on factors such as occupation (e.g., processing and/or manufacture of PFBS or
PFBS-containing products and chrome electroplating) and living conditions (e.g., proximity to
locations that make or use PllSS-containing products and well-water use).
Human studies have examined possible associations between PFBS exposure and potential
health outcomes such as alteration of menstruation, reproductive hormones or semen parameters,
kidney function (uric acid production), lung function (induction of asthma), and lipid profile. The
ability to draw conclusions about associations was limited due to the small number of studies per
outcome. Of the examined outcomes, only asthma, serum cholesterol, and high-density
lipoprotein levels were found to exhibit a statistically significant positive association with PFBS
exposure. No studies have been identified that evaluate the association between PFBS exposure
and potential cancer outcomes.
ix
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Animal studies of repeat-dose PFBS exposure have been exclusively via the oral route, used the
potassium salt of PFBS (K+PFBS) as the source exposure material, and have examined
noncancer effects only. The available rat and mouse studies support identification of thyroid,
developmental, and kidney endpoints as potential health effects following repeated exposures in
utero and/or during adulthood. Animal studies also evaluated other health outcomes such as
liver, reproductive parameters, lipid/lipoprotein homeostasis, spleen, and hematology; however,
the available evidence does not support a clear association with PFBS exposure.
Noncancer Effects Observed Following Oral Exposure
Oral exposures to PFBS or its K+ salt in adult and developing rats and mice have been shown to
result in thyroid, developmental, and kidney effects. Thyroid effects in adult exposed rats and
mice and in developing mice were primarily expressed through significant decreases in
circulating levels of hormones such as thyroxine (T4) and triiodothyronine (T3). In early
developmental life stages in mice (e.g., newborn). decreases in thyroid hormone were
accompanied by other effects indicative of delayed maturation or reproductive development
(e.g., vaginal patency and eyes opening). Kidney weight and/or histopathological alterations
(e.g., renal tubular and ductal epithelial hyperplasia) were observed in rats following short-term
and subchronic oral exposures. Many of the kidney effects, however, occurred at higher doses
than did the thyroid and developmental effects. The limited number of human studies examining
oral PFBS exposure does not inform the potential for effects in thyroid, developing offspring, or
the renal system.
Oral Reference Doses for Noncancer fffect*
Subchronic and chronic oral RlDs were deri\ ed for IT US The hazards of potential concern
include thyroid, developmental, and kidney effects I-Tom these identified targets of PFBS
toxicity, perturbation of thyroid hormone levels (eg. thyroxine [T4]) and kidney histopathology
(e.g., papillary epithelial tubular'ductal hyperplasia) were used as endpoints for derivation of
candidate RlDs The I .IW's Recommended I se of Body Weight314 as the DefauftMethod in
Derivation of the ()ral Reference / >ose (IS I-PA. 20'llb) was used to allometrically scale a
toxicologically equivalent dose of orally administered PFBS from animals to humans. Following
the EPA's Benchmark Dose Technical (inidance Document (U.S. EPA. 2012). benchmark dose
(BMD) modeling of thyroid and kidney effects following exposure toK+PFBS resulted in a
BMDL20 human equivalent dose (FLED) of 4.2 milligrams per kilogram per day (mg/kg-day) and
a BMDL10 HED of I I 5 mg/kg-day, respectively. The HEDs associated with the thyroid and
kidney effects served as the point of departure (POD) for derivation of the candidate subchronic
and chronic RfDs for each effect
The candidate subchronic RfD for K+PFBS associated with thyroid effects was calculated by
dividing the PODhed for decreased serum total T4 observed in newborn (PND 1) mice,
conducted by Feng et al. (2017). by a composite uncertainty factor (UFc) of 100 to account for
extrapolation from mice to humans (an interspecies UF, or UFa, of 3), for interindividual
differences in human susceptibility (intraspecies UF, or UFh, of 10), and for deficiencies in the
toxicity database (database UF, or UFd, of 3) (a value of 1 was applied for subchronic-to-chronic
UF, or UFs, and LOAEL-to-NOAEL UF, or UFl) (see Table 9), yielding a candidate subchronic
RfD of 4 x 10 2 mg/kg-day. As K+PFBS is fully dissociated in water at the environmental pH
range of 4-9, data for K+PFBS were used to derive a subchronic RfD for the free acid (PFBS) by
x
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
adjusting for differences in molecular weight (MW) between K+PFBS (338.19) and PFBS
(300.10), yielding the same value of 4 x 10 2 mg/kg-day for the candidate subchronic RfD
(thyroid effects) for PFBS (free acid).
The candidate subchronic RfD for K+PFBS associated with kidney effects was calculated by
dividing the PODhed for increased papillary epithelial tubular/ductal hyperplasia in Po female
rats, conducted by Lieder et al. (200%! by a composite uncertainty factor (UFc) of 100 to
account for extrapolation from rats to humans (an interspecies UF, or UFa, of 3), for
interindividual differences in human susceptibility (intraspecies UF, or UFh, of 10), and for
deficiencies in the toxicity database (database UF, or ITd. of 3) ('a value of 1 was applied for
subchronic-to-chronic UF, orUFs, and LOAEL-to-N().\N. I I ', or UFl) (see Table 10), yielding
a candidate subchronic RfD of 1 x 10_1 mg/kg-day. As k Pl-'liS is fully dissociated in water at
the environmental pH range of 4-9, data for K+PFBS were used lo derive a subchronic RfD for
the free acid (PFBS) by adjusting for differences in molecular weight (MW) betweenK+PFBS
(338.19) and PFBS (300.10), yielding the same \alue of 1 x 10 1 mu ku-day for the candidate
subchronic RfD (kidney effect) for PFBS (free acid)
The candidate chronic RfD for K+PFBS associated w ilh thyroid effects was calculated by
dividing the PODhed for decreased serum total T4 ohsei \ed in newborn (PND 1) mice,
conducted by Feng et al. (2017). by a I \\ of 3<~>0 to account for extrapolation from mice to
humans (UFa of 3), for interindividual differences in human susceptibility (UFh of 10), and
deficiencies in the toxicity database (I I'd of I <>) (a \ alue of I was applied for UFs and UFl) (see
Table 12), yielding a chronic RfD of I I <> : mu ku-dav I .ike the candidate subchronic RfD for
thyroid, based on the data for k Pl-'liS. a candidate chronic Rll) for PFBS (free acid) of
1 x icr2 mg/kg-day was domed
The candidate chronic Rll) lor k Pl 'liS associated w ilh kidney effects was calculated by
dividing the PO Dm i n lor increased piipi llai\ epithelial tubular/ductal hyperplasia in Po female
rats, conducted In l.icdcrclal (2"n^b). In a I I \ of I .<>00 to account for extrapolation from rats
to humans (UFa of 3). for intei indi\ idual differences in human susceptibility (UFh of 10), to
account for less than chronic-duration exposure (UFs of 10) and deficiencies in the toxicity
database (I I 'd of 3) (a value of I w as applied for UFl) (see Table 15), yielding a candidate
chronic RID of I I n : mu ku-dav. Like the candidate subchronic RfD for kidney, based on the
data for K+PI liS. a candidate chronic RfD for PFBS (free acid) of 1 x 10~2 mg/kg-day was
derived.
Confidence m liic kj( c
The overall confidence in the candidate subchronic RfD for thyroid effects is medium. The
gestational exposure study conducted by Feng et al. (2017) reports administration of K+PFBS by
gavage in pregnant Institute of Cancer Research (ICR) mice (10/dose) from gestation days
(GDs) 1 to 20. This study was of good quality (i.e., high confidence) with adequate reporting and
consideration of appropriate study design, methods, and conduct (click to see risk of bias
analysis in HAWC). Confidence in the oral toxicity database for derivation of the candidate
subchronic RfD is medium because, although there are multiple short-term studies and a
subchronic-duration toxicity study in laboratory animals, a two-generation reproductive toxicity
study in rats (Lieder et al.. 2009b). and multiple developmental toxicity studies in mice and rats,
there are no studies available that have specifically evaluated neurodevelopmental effects
xi
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
following PFBS exposure. Neurodevelopmental effects are of particular concern when
perturbations in thyroid hormone occur during a sensitive early life stage, and the absence of a
study evaluating neurodevelopmental effects following PFBS exposure is a source of uncertainty
in the assessment.
The overall confidence in the candidate subchronic RfD for kidney effects is medium. The
subchronic exposure study conducted by Lieder et al. (2009b) reports administration of K+PFBS
by gavage to Sprague-Dawley rats for 90 days. This study was of good quality (i.e., high
confidence), was peer-reviewed, applied established approaches, recommendations, and best
practices, and employed an appropriate exposure design for the evaluation of systemic toxicity
endpoints (click to see risk of bias analysis in HAWC). Confidence in the oral toxicity database
for derivation of the candidate subchronic RfD is medium because although there are multiple
short-term studies and a subchronic-duration toxicity stuck in laboratory animals, and one
acceptable two-generation reproductive toxicity study in rats, the database lacks studies that have
specifically evaluated neurodevelopmental effects
The overall confidence in the candidate chronic Rll) for thyroid effects is low While the RfD
was derived using the same high-confidence principal stuck conducted bv I'enu et al. (2017) that
was used for the candidate subchronic RfD, there is increased concern pertaining to the potential
for identification of hazards following longer (i e . chronic) duration PFBS exposures. Thus, due
to the lack of a study evaluating neurocle\ elopmenlal effects and a chronic duration study,
confidence in the database specifically for a candidate chronic RID for thyroid is low.
The overall confidence in the candidate chronic Rll) lor kidney is low. While the RfD is derived
using the same high confidence principal study conducted by l.ieder et al. (2009b\ as was used
for the candidate subchronic RfD, there is increased concern pertaining to the potential for
identification of hazards following longer (i.e . chronic) duration PFBS exposures. Thus, due to
the lack of a stuck e\alualinu neurocle\ elo|">menlal effects and a chronic duration study,
confidence in the database specifically lor a candidate chronic RfD for kidney is low.
Effects oihai thar: Cancer Observed / allowing Inhalation Exposure
There are no studies a\ailable that examine toxicity in humans or experimental animals
follow inu inhalation exposure, precluding the derivation of an inhalation reference concentration
(RfC)
Evidt
Under the EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA. 2005). the Agency
concluded that there is "inadequate evidence to assess carcinogenic potential" for PFBS and
K+PFBS by either oral or inhalation routes of exposure. Therefore, the lack of data on the
carcinogenicity of PFBS and the related compound K+PFBS precludes the derivation of
quantitative estimates for either oral (oral slope factor) or inhalation (inhalation unit risk)
exposure.
xii
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
1.0 Background
1.1 Physical and Chemical Properties
Perfluorobutane sulfonic acid (PFBS) (Chemical Abstracts Service Registry Number
[CASRN] 375-73-5)1 and its related salt, potassium perfluorobutane sulfonate (K+PFBS)
(CASRN 29420-49-3), are members of the group of per- and polyfluoroalkyl substances (PFAS),
more specifically the short-chain perfluoroalkane sulfonates. For purposes of this assessment,
"PFBS" will signify the ion, acid, or any salt of PFBS. Concerns about PFBS and other PFAS
stem from the resistance of these compounds to hydrolysis, photolysis, and biodegradation,
which leads to their persistence in the environment (Sundstiom et al.. 20121 The chemical
formula of PFBS is C4HF9O3S and the chemical formula of k PFBS is C4F9KO3S. Their
respective chemical structures are presented in I'iuine 1. KPFliS differs from PFBS by being
associated with a potassium ion. The reported water solubility of each species suggests that in
aqueous environments, the sulfonate would be the predominant form. The preferential use of
K+PFBS in laboratory studies is related to the optimal dissociation of the salt to the sulfonate
(i.e., PFBS) at pH ranging from 4 to 9 (see Table I) Table I provides a table of physicochemical
properties for PFBS and K+PFBS.
c . j : « r P- J- -
PFBS K+PFBS
I'iuuie I Chemical structures of PFBS and K+PFBS.
1 The CASRN given is for linear PFBS; the source PFBS used in toxicity studies was assayed at >98% linear, suggesting some
minor proportion of other chemicals, such as branched PFBS isomers, are present. Thus, observed health effects may apply to the
total linear and branched isomers in a given exposure source.
1
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Table 1. Physicochemical properties of PFBS (CASRN 375-73-5) and related compound
K+PFBS (CASRN 29420-49-3)
Property (unit)
Value
PFBS (free acid)
K+PFBS (potassium salt)
Boiling point (°C)
80-211 (experimental)13
205-447 (predicted)0
Density (g/cm3 at 71 °C)a
ND
ND
Vapor pressure (mm Hg at 20 °C)a
ND
9.15 x 10-8
pH (unitlcss)"
ND
ND
Solubility in water (mg/L)a
56.6 at 24 °C
4i.: ai :<>¦¦(•
Molecular weight (g/mol)a
300.09
33X I'J
Dissociation constant3
NA
Fulls dissociated in water over the pH range of 4-9
Sources'.
aNICNAS C2005y
bU.S. EPA Chemistry Dashboard for CASRN 375-73-5.
CU.S. EPA Chemistry Dashboard for CASRN 2942049-3.
Notes'. °C = degrees Celsius; g/cm3 = grams per cubic centimeter: g/mol = grams per mole; mm 11(1 = millimeters of mercury;
mg/1 = milligrams per liter; NA = not applicable; ND = no data.
1.2 Occurrence
PFBS-based compounds are surfactants used primarily in the manufacture of paints, cleaning
agents, and water- and stain-repellent products and coatings They serve as replacements for
perfluorooctane sulfonic acid (PFOS) (3M. 2"i)2|i) Various sources report their occurrence in
environmental media and consumer products, including drinking water, ambient water, dust,
carpeting and carpet cleaners. Iloorwax, and food packaging
Oral exposure via drinking water might he expected in areas where contamination has been
reported. EPA Unregulated Contaminant Monitoring Reporting data for public drinking water
utilities in 2013-2015 showed levels of PFBS above the Minimum Reporting Level (> 0.09
micrograms per liter [|ig/L]) in water systems serving Alabama, Colorado, Georgia, the Northern
Mariana Islands, and PennsyKania (U.S. EPA. 2017; Hu et al.. 2016Y These utilities included
both ground and surface drinking water sources, with concentrations ranging from 0.09 to
0.37 |ig/L. The estimated combined number of people served by these water systems is more
than 340,000 (U.S. EPA. 2<>1S)
Measurements from 37 surface water bodies in the northeastern United States (metropolitan New
York area and Rhode Island) collected in 2014 showed an 85% site detection rate (Zhang et al..
2016). PFBS has also been identified in surface waters in Georgia, New Jersey, North Carolina,
and the Upper Mississippi River Basin (Post et al.. 2013; Lasier et al.. 2011; Nakavama et al..
2010; Nakavama et al.. 2007). It has been detected in wastewater treatment plant effluent,
seawater, soil, and biosolids (Houtz et al.. 2016; Zhao et al.. 2012; Sepulvado et al.. 2011Y
PFBS contamination, which has been associated with the use of Aqueous Film-Forming Foams
(AFFFs) (ESTCP. 2017; Anderson et al.. 2016). was reported at Superfund sites and areas under
assessment for Superfund designation. Contaminated sites include the former Wurtsmith Air
Force Base, Ellsworth Air Force Base, and Dover Air Force Base (Aerostar SES LLC. 2017;
2
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Anonymous. 2017; ASTSWMO. 2015). At the Wurtsmith site, PFBS was detected at a
concentration of 6.4 |ig/L in ground water contaminated by a PFAS plume originating from the fire
training area (ASTSWMO. 20151 It is also present in some drinking water samples from nearby
residential wells at low nanograms per liter concentrations, which were below the screening value
cited by the Michigan Department of Community Health (MDCH. 2015). Other sources of PFAS
and/or PFBS contamination include chrome plating operations, PFAS manufacture, and sites that
use PFAS in product formulations such as textile and electronic industries.
PFBS has also been detected in household dust and consumer products. There was a 92%
detection frequency for PFBS among 39 household dusl samples (10 from the United States)
analyzed with levels ranging from 86 nanograms per gram (ng g) for the 25th percentile to
782 ng/g for the 75th percentile (Kato et al.. 2009). In a separate study, PFBS dust levels were
measured in Boston area offices (n = 31), homes (n = 3d ). and \ chicles (n = 13) with detection
frequencies being relatively low—10%, 3%, and < >°... respectiv ely—and ranging in the low parts
per billion (Fraser et al.. 2013). Consumer products could also be an exposure source. Limited
quantitative testing showed the presence of PI liS in carpet and upholstery protectors (45.8 and
89.6 ng/g), carpet shampoo (25.7 and 911 ng/g). textiles (2 ng/g), and floor wax (143 ng/g)
purchased in the United States (Liu et al.. 2014)
PFBS was detected in fast food packaging (7/20 samples) in one U.S. study (Schaider et al..
2017). The European Food Safety Authority reported the presence of PFBS in various food and
drink items, including fruits, vegetables, cheese, and bottled water. For average adult consumers,
the estimated exposure ranges for PFBS were 0.03- I Sl) nanograms per kilogram per day
(ng/kg-day) (minimum) to n l<) 3 72 ng/kg-day (maximum) (l-FSA. 2012).
PFBS has been reported in scrum of humans in the general population. In American Red Cross
samples collected in 201 5. S 4".. had a quantifiable serum PFBS concentration; the majority of
samples were below the lower limit of quantitation (4.2 nanograms per milliliter [ng/mL]) (Olsen
et al.. 201 7) The National I Icalth and Nutrition F.xamination Survey (NHANES) 2013-2014
data reported the l>5ih percentile lor PI liS al or below the level of detection (0.1 ng/mL).
Another study with a lower limit of detection (0.013 ng/g) reported increasing levels of PFBS in
serum from primiparous nursing women in Sweden from 1996 to 2010 (Glynn et al.. 2012).
1.3 Toxii
1.3.1 Overvie
Animal evidence has show n that PITBS, like other PFAS, is well absorbed following oral
administration. PFBS distributes to all tissues of the body (Bogdanska et al.. 2014). but a study
evaluating the volume of distribution concluded that distribution is predominantly extracellular
(Olsen et al.. 2009). Because of its chemical resistance to metabolic degradation, PFBS is
primarily eliminated unchanged in urine and feces.
Two sets of investigators have conducted toxicokinetic studies in rats and monkeys (Chengelis et
al.. 2009; Olsen et al.. 2009). Olsen et al. (2009) also measured the half-life of PFBS in humans.
More recently, Bogdanska et al. (2014) and Rumpler et al. (2016) have reported limited
toxicokinetic information in mice.
Results of all studies discussed in this section are summarized in Table 2.
3
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Table 2. Summary of toxicokinetics of serum PFBS (mean ± standard error)
Species/Sex
Study design
Elimination
half-life (hr)
AUC
(jig-hr/mL)
Clearance
Volume of
distribution (L/kg)
Reference
Mice
Mice/male
Single i.v. dose (30 or 300
mg/kg)
4.6a
0.4
Ruinolcr et al. (2016)
Mice/female
Single i.v. dose (30 or
300 mg/kg)
2.5a
0.4
Ruinolcr et al. (2016)
Rats
Rats/male
Single i.v. dose (10 mg/kg)
2.1
254
0.0^94 (L/hr-kg)
0.118
Chenselis et al. (2009)
Single i.v. dose (30 mg/kg)
4.51 ±2.22
294 ± 77
ll<> '4 (L/hr)"
n '30 ±0.032
Olsen et al. (2009)
Single oral dose (30 mg/kg)
4.68 ±0.4^
163 ± 10
NA
0.676 ±0.055
Olsen et al. (20091
Rats/female
Single i.v. dose (10 mg/kg)
0.64
^2
n '1 1 (1. hr-kg)
0.288
Chenselis et al. (2009)
Single i.v. dose (30 mg/kg)
3.96 ±0.21
(>5 5
4<>lJ 4(i(l. hrY3
0.351± 0.034
Olsen et al. (2009)
Single oral dose (30 mg/ks)
" 42 0 79
X5 12
\ \
0.391± 0.105
Olsen et al. (2009)
Monkeys0
Cynomolgus
macaque/male
Single i.v. dose (10 mg/km
15 rx.5)
1.1 15 X5<>
n n l(i (L/lir-kg)
0.209 ± 0.028
Chenselis et al. (2009)
Single i.v. dose (10 mg/km
<>5.2 2" 1
24' X(.
511 141 (mL/hr)
0.254 ±0.031
Olsen et al. (2009)
Cynomolgus
macaque/female
Single i \ dose < In mu km
S 1
4X<> IXn
nn22l> n 0099 (L/hr-kg)
0.248 ± 0.045
Chenselis et al. (2009)
Single i \ dose (In ms km
X' 2 41
'5 4 1' '
'(.X 120 (mL/hr)
0.255 ±0.017
Olsen et al. (20091
Humans
Males and females
NA
2"" 4.5' days>
NA
NA
NA
Olsen et al. (2009)
Notes: AUC = area under the curve; hr = hour: i.v. = intravenous: L/lir-kg = liters per hour per kilogram; L/kg = liter per kilogram; mL/hr = milliliters per hour; (ig-hr/mL =
micrograms per hour per milliliter; NA = not available.
aRumpler et al. ('2016') is a published abstract only.
bBody weights were reported to be 0.200-0.250 kg (approximately 476 L/kg-hour).
cThe data were monitored 48 hours and 31 days postdosina for Chenaelis et al. (2009) and Olsenetal. (20091 respectively.
dOne male monkey had a serum concentration more than tenfold higher than the others at 48 hours postdosing with an estimated half-life of 26 hours.
4
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
1.3.2 Absorption
Olsen et al. (2009) conducted intravenous (i.v.) and oral uptake studies in rats (three male and
female) that were given a single oral dose (30 milligrams per kilogram [mg/kg]) of potassium
PFBS (K+PFBS). The serum area under the concentration curve (AUC) after i.v. was 294 ± 77
and 65 ± 5 (|ig-h/mL) in male and female rats, respectively, and 163 ± 10 and 85 ± 12 in males
and females, respectively, after oral dosing. The large variance in AUC for male rats after i.v.
dosing and greater AUC after oral dosing compared to i.v. dosing in females makes it difficult to
interpret these results with certainty. Peak concentrations occurred at 0.3-0.4 hours after oral
dosing, showing that absorption was fairly rapid.
1.3.3 Distribution
In a published abstract, Rumpler et al. (2016) evaluated the pharmacokinetic properties of PFBS
in CD-I mice at 8 weeks of age. Male and female mice were ui\ en a single dose of 0, 30, or
300 mg/kg PFBS via gavage. Liver and kidney were harvested 24 hours postdosing. PFBS did
not accumulate in either liver or kidney, and 1 he \ olume of distribution was 0.4 liter per kilogram
[L/kg],
Olsen et al. (2009) estimated volumes of distribution for k Pl-'liS as 0.7 and <>4 1 ,/kg in male
and female rats, respectively, and 0.25 I. kg in cynomoluus macaques, and concluded that
K+PFBS is primarily distributed in the extracellular space
Bogdanska et al. (2014) characterized the tissue distribution of ;~S-labeled PFBS in male
C57BL/6 mice. Animals (3'group) were exposed for either 1. 3. or 5 days to an average of 16 mg
of PFBS/kg/day in the diet. Following 1, 3, and 5 days of exposure, total estimated recovery of
PFBS from all tissues e\aluated was 10%, 5%, and 3 .4% of the ingested dose, respectively. The
declining recovery w ilh time reflects the lack of accumulation in tissues after the first few days,
with continued elimination in the urine. The stuck authors suggest that these low recovery rates
most likely reflect rapid excretion of PFBS and/or potentially limited uptake of the compound,
but the results of Rumpler et al. (2016) and Olsen et al. (2009) suggest that limited tissue
distribution is also a factor.
Bogdanska el al (2014) found that Mood levels of PFBS did not change when comparing values
observed after I and 5 days of exposure. As with PFOS, PFBS was found to distribute to most of
the 20 tissues examined at all exposure durations, but the levels of PFBS were significantly
lower (fivefold to forty fold lower) than those of PFOS in tissues after similar exposure to PFOS,
especially in liver and lungs (Bogdanska et al.. 2014). These differences might be attributed to
chain length-dependenl acti\ e transport of perfluorinated chemicals (Weaver et al.. 2010).
Excluding stomach and fat tissue, PFBS tissue levels increased between 1 and 3 days of
exposure, but there were no significant changes in tissue levels between 3 and 5 days of exposure
in any tissue examined. Similar to PFOS, whole bone, liver, blood, skin, and muscle accounted
for approximately 90% of the recovered PFBS at all time points. The highest tissue
concentrations outside of blood, however, were found in liver, GI tissues, kidney, and cartilage.
The significant total mass found in muscle and skin was due to the total volume of these tissues
as much as the concentration in them. The liver contained the highest tissue concentration of
PFBS at all time points, while the brain contained the lowest.
5
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Human studies were not available on lactational or transplacental transfer of PFBS, but
developmental studies in animals indicate the potential for effects in offspring following
gestational exposure (Feng et al.. 2017; York. 2003a. 20021
1.3.4 Metabolism
There is no evidence of biotransformation of PFBS. It is expected that PFBS, a short-chain (C4)
of perfluoroalkyl acids (PFAAs), is metabolically inert because of the chemical stability that also
exists in the longer chain PFAA chemicals, including perfluorohexane sulfonic acid (PFHxS)
(C6), PFOS (C8), and perfluorooctanoic acid (PFOA) (C8).
1.3.5 Elimination
To facilitate comparison of differing studies for a given species, results for elimination are
organized by species.
1.8.5.1 Mice
Rumpler et al. (2016s) dosed male and female ('l)-l mice with 0, 30, or 3<)i) nig/kg PFBS via
gavage. Trunk blood was collected at 0.5, 1, 2. 4. S. 16, 24. and 48 hours and urine at 24 hours
after dosing. The half-life of PFBS was estimated to be 4 6 hours in the male mice and 2.5 hours
in the females. Within 24 hours, more than 95% of the serum I'TBS was excreted into urine.
1.3.5.2 Rats
Chengelis et al. (2009) conducted a single-dose pharmacokinetic study in Sprague-Dawley (S-D)
rats, designed to com pare the toxicokinetic heha\ ior of PI liS to that of perfluorohexanoic acid
(PFHxA), another PI'AA In this study, 12 male and 12 female rats were each administered a
bolus dose ofPFBS (l<> niu kg) \ia i v. injection Blood samples were collected from
three animals per sex at <) 5. I. I 5. 2. 4, 8, and 24 hours after dose administration. Additionally,
to determine urinary excretion, three animals per sex were housed in metabolic cages following
dose administration and urine was collected over the following time intervals: 0-6, 6-12, and
12-24 hours postdosinu. Female rats had an approximately threefold shorter mean elimination
half-life ofPFBS in serum (0.64 hours) than male rats (2.1 hours). This could be in part due to
the difference in clearance and \ olunie of distribution; the mean apparent clearance ofPFBS
from the serum was approximately eightfold higher for female rats (0.311 L/h/kg) than for male
rats (0.0394 I. h ku) and the mean apparent volume of distribution for PFBS in the serum was
approximately 2 4-fold higher lor female rats (0.288 L/kg) than for male rats (0.118 L/kg).
Approximately 70°/.. of the administered dose ofPFBS was recovered in the urine during
24 hours postdosing regardless of sex. Using the urine data, the mean half-life values for male
rats and female rats were determined to be 3.1 and 2.4 hours, respectively; the finding of longer
urinary half-lives in males is consistent with those observed for serum half-lives.
Olsen et al. (2009) evaluated the elimination ofPFBS in S-D rats after i.v. and oral exposure to
K+PFBS. The mean terminal serum elimination half-lives following i.v. administration of
30 mg/kg K+PFBS were 4.51 ± 2.22 hours for males and 3.96 ± 0.21 hours for females. Although
there was not a statistically significant difference between the terminal serum half-lives in male
and female rats, there was a statistically significant difference in the urinary clearance rates
(p < 0.01), with female rats (469 ± 40 mL/hour) having faster clearance rates than male rats
(119 ± 34 mL/h). (Since clearance [CL] is calculated from the ratio of the volume of distribution
6
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
[Vd] to the half-life [tl/2], CL = 0.693 *Vd/tl/2, differences in Vd can lead to differences in CL,
even when tl/2 is similar between comparison groups.) For rats receiving an oral dose, terminal
serum K+PFBS elimination half-lives were significantly different (p < 0.05) for males
(tl/2 = 4.68 ± 0.43 hours) versus females (tl/2 = 7.42 ± 0.79 hours). Thus, the half-life estimates
of Olsen et al. (2009) (4-7.5 hours) are roughly twice those estimated by Chengelis et al. (2009)
based on urine data (2.4 and 3.1 hours); there is more of a difference with the values estimated
by Chengelis et al. (2009) from serum data (0.64 and 2.1 hours).
1.3.5.3 Monkeys
Similar to their study in rats, Chengelis et al. (2009) investigated the toxicokinetic profile of
PFBS through a series of experiments in the cynomolgus macaque (Macaca fascicularis).
Monkeys (three males and three females) were each administered a bolus i.v. dose of 10 mg/kg
PFBS. The controlled exposure to PFBS occurred 7 days after the same animals were each
administered a bolus dose of PFHxA (10 mg/kg) Blood samples were collected at 0 hours
(immediately prior to dosing) and at 1, 2, 4, 8. 24. and 48 hours after dose administration and
were analyzed to determine PFBS concentration in serum. Only a single clearance half-life was
estimated. The estimated half-life of PFBS in serum ranged from 5.8 to 20 <> hours in this
experiment, and the median half-life was 9.55 hours for the six animals.
The study of Chengelis et al. (200^) indicated that, under conditions of equivalent exposure, the
areas under the serum concentration-time cur\es (AlCs) were lower and the elimination
half-lives were shorter for PFHxA lhan those for PI liS in both S-D rats and cynomolgus
macaques. In the monkeys, for instance. PI I IxA was cleared more rapidly and resulted in a
lower AUC value (approximately an order of magnitude lower) w ith a shorter terminal half-life
(2.4-5.3 hours, data not show n in the study) than Pl-'liS at an equi\ alent dose (i.v. dose at
10 mg/kg).
Olsen eta I (2009) also evaluated the elimination of PFBS (specifically, K+PFBS) in cynomolgus
macaques after i \ dosing A significant difference in design from the study of Chengelis et al.
(200'->) is that Olsen el al. (2""^) followed Pl-'liS elimination for 31 days in monkeys (versus
48 hours), allowing Olsen and colleagues to identify both an initial clearance half-life and a
terminal phase-half-life. Olsen el al. (2009) did not observe statistically significant sex-related
differences in hall-life or clearance between male and female monkeys, unlike those observed in
rats. In monkeys, the mean terminal serum elimination half-lives, after i.v. administration of
10 mg/kg K+PFBS, were 95 27 hours in males and 83 ± 42 hours in females.
The serum half-life data in Olsen et al. (2009) clearly show a slow elimination phase in monkeys
that does not begin until 4 10 days after dosing. Chengelis et al. (2009) followed elimination for
only 48 hours, hence could not have observed this terminal clearance phase. The initial
elimination half-life (t0.5P) estimated by Olsen et al. (2009) in monkeys—13 hours for males,
11 hours for females—is essentially identical to the values estimated by Chengelis et al.
(2009)—10 or 15 hours for males (without/with outlier) and 8 hours in females. Hence the two
studies appear consistent in identifying an initial elimination half-life, but the difference in
design precluded Chengelis and colleagues from identifying the longer (terminal) half-life of
PFBS.
7
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
1.3.5.4 Humans
In addition to their experimental studies in rats and monkeys, Olsen et al. (2009) evaluated the
elimination of human serum K+PFBS in a group of workers with occupational exposure, with
serum concentrations measured up to 180 days after cessation of further K+PFBS work-related
activity. Given that the workers had been occupationally exposed, distribution into the tissues is
expected to have been complete before the observations began. Among the six subjects
(five male, one female), the geometric mean serum elimination half-life for K+PFBS was
25.8 days (95% confidence interval = 16.6-40.2 days). Urine appeared to be a major route of
elimination based on observed levels of PFBS in urine in the human study.
1.3.6 Summary
Collectively, elimination half-lives appear to be sim i lar lor mice and rats, with potential
sex-specific toxicokinetic differences being reported (i e . females appearing to have a faster
elimination rate). Humans have a longer serum elimination half-lilc ( weeks) than both rodents
(-hours) and monkeys (-days).
8
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
2.0 Problem Formulation
2.1 Conceptual Model
A conceptual model was developed to summarize the availability of data to understand potential
health hazards related to exposure to PFBS and/or K+PFBS. The potential sources of these
chemicals, the routes of exposure for biological receptors of concern (e.g., various human
activities related to ingested drinking water, and food preparation and consumption), the
potential assessment endpoints (e.g., effects such as developmental toxicity), and adverse health
effects in the populations at risk due to exposure to PFBS and 'or potassium salt are depicted in
the conceptual diagram in Figure 2.
9
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
STRESSOR
PFBS and Potassium Salt
POTENTIAL
SOURCES OF
EXPOSURE
EXPOSURE ROUTES
POTENTIAL
ORGANS/
SYSTEMS
AFFECTED
Drinking
Water
Thyroid Effects
Ambient
Water
Industrial
Uses
Consumer
Products
Food
Dust
Oral
Reproductive
Effects
Developmental
Effects
POTENTIAL
RECEPTORS IN
GENER4L
POPULATION
Adults
Children
Dermal
Renal Effects
Hepatic Effects
Lipid and
Lipoprotein
Effects
Pregnant Women
and Fetuses
Lactating Women
Aii-
Soil
Inhalation
Fire-
fightiug
Foams
LEGEND
Data Selected for
Assessment
Limited Data
Unknown
Quantitative Data
Figure 2. Conceptual model for PFBS and/or potassium salt
10
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
2.2 Objective
The overall objective of this assessment is to provide the health effects basis for the development
of oral reference doses (RfDs) for PFBS (CASRN 375-73-5) and a related compound, K+PFBS
(CASRN 29420-49-3), including the science-based decisions providing the basis for
identification of potential human health effects and estimating PODs. Based on the needs of the
EPA partner Program Offices, Regions, States, and/or Tribes as they pertain to diverse exposure
scenarios and human populations, subchronic and chronic RfDs have been derived. The
assessment includes studies and information previously |">i o\ ided in the 2014 Provisional
Peer-Reviewed Toxicity Value assessment (U.S. EPA. 2" 141') and huilds upon the amount of
literature containing studies published since that review.
2.3 Methods
2.3.1 Literature Search
Four online scientific databases (PubMed, Web of Science. Toxline, and TSCATS via Toxline)
were searched by the EPA's Health and Environmental Research Online (HERO) staff and
stored in the HERO database.2 The literature search focused on chemical name and synonyms
with no limitations on publication type. e\ idence stream (i e . human, animal, in vitro, and in
silico), or health outcomes. Full details of the search strategy for each database are presented in
appendix A. The initial database searches were conducted on July 18, 2017 and last updated on
February 28, 2018. Studies were also identified from other sources relevant to PFBS, including
studies submitted to the N\\ by the manufacturer of PI liS (i e . 3M) as part of TSCA
premanufacture nolices for other PFAS chemicals or as required under Toxic Substances Control
Act (TSCA) reporting requirements and studies referenced in prior evaluations of PFBS toxicity
(MDH. 2017; ATSDR. 201 5) In addition, on March 29, 2018, the National Toxicology Program
(NTP) published study tables and indi\ idual animal data from a 28-day toxicity study of PFBS
(http://doi.om/10.22427/NTP-DAT A-002-01 134-'»"3-0000-4). Although a peer-reviewed NTP
Technical Report for the PFBS study is not yet available, this information was included in the
assessment because these data ha\e undergone standard NTP quality assurance/control
processing and are publicly a\ ailahle. A protocol outlining the NTP study methods is available in
HERO (https://hero.epa.gov, hero index.cfm/reference/details/reference_id/4309741) (NTP.
2011). During the process of deri\ ing toxicity values, the EPA conducted further quantitative
analyses (e.g., BMI) modeling) beyond what was reported by the NTP.
2.3.2 Screening Pro~.
Two screeners independently conducted a title and abstract screening of the search results using
DistillerSR3 to identify study records that met the Population, Exposure, Comparator, Outcome
(PECO) eligibility criteria (see appendix B for a more detailed summary):
• Population: Human and nonhuman mammalian animal species (whole organism) of any
life stage and in vitro models of genotoxicity.
2The EPA's Health and Environmental Research Online (HERO) database provides access to the scientific literature
behind EPA science assessments. The database includes more than 2,500,000 scientific references and data from the
peer-reviewed literature used by the EPA to develop its regulations.
3DistillerSR is a web-based systematic review software used to screen studies available at
https://www.evidencepartners.com/products/distillersr-svstematic-review-software.
11
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
• Exposure: Any qualitative or quantitative estimates of exposure of PFBS or K+PFBS, via
oral or inhalation routes of exposure.
• Comparator: A comparison or reference population exposed to lower levels or for shorter
periods of time for humans. Exposure to vehicle-only or untreated control in animals.
• Outcome: Any examination of cancer or noncancer health outcomes.
In addition to the PECO criteria, the following additional exclusion criteria were applied,
although these study types were tracked as supplemental material as described following the
exclusion criteria:
• Records that do not contain original data such as oilier agency assessments, scientific
literature reviews, editorials, and commentaries.
• Abstract only (e.g., conference abstracts), and
• Retracted studies.
Records that were not excluded based on title and abstract screening ad\ a need to full-text review
using the same PECO eligibility criteria. Studies thai ha\ e not undergone peer re\ iew were
included if the information could be made public and sufficient details of study methods and
findings were included in the reports l-'ull-lext copies of potentially relevant records identified
from title and abstract screening were icliic\ed. stored in the III-RO database, and independently
assessed by the screeners using DistillerSR to confirm eligibility At both title/abstract and
full-text review levels, screening conflicls were rcsol\ cd In discussion between the primary
screeners in consultation with a third reviewer to rcsol\ c any remaining disagreements. During
title/abstract or full-text le\ el screening, studies that were not directly relevant to the PECO, but
could provide supplemental information, were categorized (or "tagged") by the type of
supplemental information tlicy |ii'o\ ided (e.g., re\ iew, commentary, or letter with no original
data; conference abstract, toxicokinetics; mechanistic information aside from in vitro
genotoxicity studies; other routes of exposure; exposure only). Conflict resolution was not
required during the screening process to identify supplemental information (i.e., tagging by a
single screener was sufficient to identify the study as potential supplemental information).
2.3.3 Study Evaluation
Study evaluation was conducted by one reviewer for epidemiological studies and by two
independent reviewers for animal studies using the EPA's version of Health Assessment
Workspace Collaboratee (I I.WVC), a free and open source web-based software application
designed to manage and facilitate the process of conducting literature assessments.4 For
pragmatic purposes, only one reviewer was considered necessary for epidemiological studies
because it was apparent during literature screening that the animal evidence would be most
informative for deriving toxicity values. The available outcomes in the epidemiological studies
were heterogeneous and unrelated to each other, and only a single study was available for each
outcome. This approach is consistent with recommendations from the National Academies of
Science encouraging the EPA to explore ways to make systematic review more feasible,
including a "rapid review in which components of the systematic review process are simplified
4HAWC: A Modular Web-Based Interface to Facilitate Development of Human Health Assessments of Chemicals.
https ://hawcproi ect. org/.
12
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
or omitted (e.g., the need for two independent reviewers)" (NASEM. 2017). Study evaluation
was not conducted for studies tagged as supplemental information during screening.
The general approach for evaluating epidemiology and animal toxicology was the same
(see Figure 3), but the specifics of applying the approach differed. These evaluations were
focused on the methodological approaches and completeness of reporting in the individual
studies, rather than on the direction or magnitude of the study results. Evaluation of
epidemiology studies was conducted for the following domains: exposure measures, outcome
measures, participant selection, confounding, analysis, sensitivity, and selective reporting. For
animal studies, the evaluation process focused on assessing aspects of the study design and
conduct through three broad types of evaluations: reporting quality, risk of bias, and study
sensitivity. A set of domains with accompanying core questions fall under each evaluation type
and directed individual reviewers to evaluate specific suidy characteristics. For each domain
evaluated for experimental animal studies—reporting quality, selection or performance bias,
confounding/variable control, reporting or attrition bias, exposure methods sensitivity, and
outcome measures and results display—basic considerations provided additional guidance on
how a reviewer might evaluate and judge a study for that domain. Core and prompting questions
used to guide the criteria and judgment for each domain are presented in appendix C. Key
concerns for the review of epidemiology and animal toxicology studies are potential sources of
bias (factors that could systematical I y affect the magnitude or direction of an effect in either
direction) and insensitivity (factors that limit the ability of a study to detect a true effect).
For each study in each evaluation domain. ie\ iewers reached a consensus rating regarding the
utility of the study for hazard identification, with categories oft:<><>il, adequate, deficient, not
reported, or critically deficieni These ratings were then combined across domains to reach an
overall classification of high, medium, or low confidence or uninfurinative (definitions of these
classifications are available in appendix (') The rationale for the classification, including a brief
description of any identified strengths and or limitations from the domains and their potential
impact on the o\ erall confidence determination, is documented and retrievable in HAWC.
Uninformati\e studies were not used in evidence synthesis or dose-response analysis. Studies
were e\ aluated for their suitability for each health outcome investigated and could receive
different ratings for each outcome
For epidemiological studies, exposure-specific criteria were developed prior to evaluation and
are described in detail in appendix C. In brief, standard analytical methods of measurement of
PFBS in serum or w hole-blood using quantitative techniques such as liquid chromatograph-triple
quadrupole mass spectrometry and high-pressure liquid chromatography with tandem mass
spectrometry were preferred. In addition, exposure must have been assessed in a relevant
time-window for development of the outcome.
13
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Individual study level domains
Domain judgements
Animal
Reporting Quality
Selection or Performance Bias
Confounding/Variable Control
Reporting or Attrition Bias
Exposure Methods Sensitivity
Outcome Measures and Results Display
Epidemiological
Exposure measurement
Outcome ascertainment
Population Selection
Confounding
Analysis
Sensitivity
Selective reporting
Judgement
Interpretation
0
o
Good
Adequate
Appropriate study conduct relating to the domain & minor deficiencies not
expected to influence results.
A study that may have some limitations relating to the domain, but they are not
likely to be severe or to have a notable impact on results.
•
•
Deficient
Critically
Deficient
Identified biases or deficiencies interpreted as likely to have had a notable impact
on the results or prevent reliable interpretation of study findings.
A serious flaw identified that is interpreted to be the primary driver of any
observed effect or makes the study uninterpretable. Study is not used without
exceptional justification.
Overall study rating
Rating
High
Medium
Low
Uninformative
Interpretation
No notable deficiencies or concerns identified; potential for bias unlikely or minimal;
sensitive methodology.
Possible deficiencies or concerns noted, but resulting bias or lack of sensitivity would be
unlikely to be of a notable degree.
Deficiencies or concerns were noted, and the potential for substantive bias or inadequate
sensitivity could have a significant impact on the study results or their interpretation.
Serious flaw(s) makes study results unusable for hazard identification
Figure 3. Approach for evaluating epidemiological and animal toxicology studies.
2.3.4 Data Extraction
Information on study design, methods, results, and data from animal toxicology studies were
extracted into the HAWC and are available at https://hawcprd.epa.gov/assessment/100000Q37/.
Visual graphics prepared from HAWC are embedded as hyperlinks and are fully interactive
when viewed online by way of a "click to see more" capability. Clicking on content allows
access to study evaluation ratings, methodological details, and underlying study data. The action
of clicking on content contained in those visual graphics (e.g., data points, endpoint, and study
design) will yield the underlying data supporting the visual content. NOTE. The following
browsers are fully supportedfor accessing HAWC: Google Chrome (preferred), Mozilla Fire fox,
and Apple Safari. There are errors in functionality when viewed with Internet Explorer. Study
methods and findings from epidemiological studies were described in narratives given the small
size and heterogeneity of the evidence base. Data extraction was performed by one member of
the evaluation team and checked by one to two other members. Any discrepancies in data
extraction were resolved by discussion or consultation with a third member of the evaluation
team. Digital rulers such as WebPlotDigitizer and Grab It (https://automeris.io/WebPlotDigitizer/
and https://grab-it.softl 12.com/, respectively) were used to extract numerical information from
figures. Use of digital rulers was documented during extraction. Dose levels were extracted as
reported in the study and converted to milligrams per kilogram per day(mg/kg-day) human
equivalent dose (HED) for endpoints that were considered for use in the dose-response and
derivation of toxicity values.
14
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
2.3.5 Evidence Synthesis
For the purposes of this assessment, after study evaluation, the informative evidence for each
outcome was summarized from the available human studies and, separately, the available animal
studies. This synthesis provides a short synopsis of the breadth of data available to inform each
outcome and summarizes information on the general study design, doses tested, outcomes
evaluated, and results for the endpoints of interest within each study. While the evidence
synthesis describes inferences about the methodological rigor and sensitivity of the individual
studies (i.e., study confidence) and discusses the pattern and magnitude of the experimental
findings within studies, it does not include conclusions drawn across the sets of studies (see
"Evidence Integration and Hazard Characterization," next)
2.3.6 Evidence Integration and Hazard Characte
In this assessment, the evaluation of the available evidence from informative human and animal
studies was described in an evidence integration narrative for each outcome, including overall
evidence integration judgments as to whether the data provide evidence sufficient to support a
hazard. These integrated judgments serve to characterize the extent of the a\ ailable evidence for
each outcome, including information on potential susceptible populations and life stages, as well
as important uncertainties in the interpretation of the data
The evidence integration for each health effect considered aspects of an association that might
suggest causation first introduced In Austin Bradford Hill (Mill. 1965). including the
consistency, exposure-response relationship, strength of association, biological plausibility, and
coherence of the evidence This involved weighing the PFBS-specific human and animal
evidence relating to each of these considerations within or across studies, including both
evidence that supported causation as well as e\ idence that indicated lack of support. For
example, the evaluation of consistency examined the similarity of results across studies (e.g.,
direction and magnitude) When inconsistencies across studies were identified, the evaluation
considered u hether results were "conflicting" (i e . unexplained positive and negative results in
similarly exposed human populations or in similar animal models) or "differing" (i.e., mixed
results explained by differences between human populations, animal models, exposure
conditions, or study methods), based on analyses of potentially important explanatory factors
such as confidence in studies' results (the results of higher confidence studies were emphasized),
exposure levels or duration, or differences in populations or species (including potential
susceptible groups) across studies (U.S. EPA. 2005). While consistent evidence across studies
increases support for hazard, unexplained inconsistency or conflicting evidence decreases
support for hazard. The e\ aluations of these considerations were informed by EPA guidelines,
including Guidelines for / K-w I op mental Toxicity Risk Assessment (U.S. EPA. 1991a) and
Guidelines for Reproductive Toxicity Risk Assessment (U.S. EPA. 1996b).
The overall evidence integration judgments were developed using a structured framework based
on evaluation of the considerations above (see Table 3). Using this framework, the human and
animal evidence for each health effect was judged separately as supports a hazard, equivocal, or
supports no hazard. Evidence integration judgments of supports a hazard span a range of
supportive evidence bases that can be further differentiated by the quantity and quality of
information available to rule out alternative explanations for the results. Equivocal evidence is
limited in terms of the quantity, consistency, or confidence level of the available studies and
15
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
serves to encourage additional research. Supports no hazard requires several high-confidence
studies across potentially susceptible populations with consistent null results; this judgment was
not reached in this assessment. Overall evidence integration judgments were drawn across the
human and animal conclusions, considering the available information on the human relevance of
findings in animals. Thus, for example, evidence in animals that supports a hazard alongside
equivocal human evidence in the absence of information indicating that the responses in animals
are unlikely to be relevant to humans would result in an overall judgment of supports a hazard
for that outcome.
Table 3. Criteria for overall evidence integration judgments
Animal
Human
Supports a
hazard
The evidence for effects is consistent or largely
consistent in at least one high- or medium-
confidence experiment.3 Although notable
uncertainties across studies might remain. an>
inconsistent evidence or remaining uncertainties
are insufficient to discount the cause for concern
from the positive experiments. In the stronger
scenarios, the set of experiments provide evidence
supporting a causal association across
independent laboratories or species In oilier
scenarios, including evidence lor an effect in a
single study, the experiments) demonstrate
additional support for causality such as coherent
effects across multiple related cndpninls. an
unusual niamiiiiide of effect, rarity, aue at onset,
orseverin. a stroiiu dose-response relationship,
and/or consistent ohser\ations across exposure
scenarios (c.g . route, uniiim. or duration). se\es.
or animal strains.
()ne or more high- or medium-confidence
independent studies reporting an association
betw een the exposure and the health outcome. In
general, the studs results are largely consistent or
any inconsistent results are not sufficient to
discount the cause for concern from the higher
confidence study or studies, and there is
reasonable confidence that alternative
explanations, including chance, bias, and
coiifoimdiim. have been ruled out. In situations in
w Inch onl> a single study is available, the results
of multiple studies are heterogeneous, or
allenialiN e explanations, including chance, bias
and confounding, have not been ruled out, there is
additional supporting evidence such as
associations with biologically related endpoints in
other human studies (coherence), large estimates
of risk, or strong evidence of an
exposure-response within or across studies.
Equivocal
Hie e\ idence is generally inadequate to determine
lia/ard. t his includes a lack of relevant studies
a\ailable or a scl of low-confidence experiments.
It also includes scenarios u ith a set of high- or
medium-confidence experiments that arc not
reasonably consistent or not considered
inform;)11n e to the lia/ard question under
evalualion This cateuors \\ ould also include a
single high- or medium-confidence experiment
with weak ev idence of an effect (e.g., changes in
one endpoint anionu several related endpoints,
and without additional evidence supporting
causality).
The evidence is considered inadequate to describe
an association between exposure and the health
outcome with confidence. This includes a lack of
studies available in humans, only low-confidence
studies, or considerable heterogeneity across
medium- or high-confidence studies. This also
includes scenarios in which there are serious
residual uncertainties across studies (these
uncertainties typically relate to exposure
characterization or outcome ascertainment,
including temporality) in a set of largely
consistent medium- or high-confidence studies.
16
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Animal
Human
Supports no
hazard
A set of high-confidence experiments examining
the full spectrum of related endpoints within a
type of toxicity, with multiple species, and testing
a reasonable range of exposure levels and
adequate sample size in both sexes, with none
showing any indication of effects. The data are
compelling in that the experiments have examined
the range of scenarios across which health effects
in animals could be observed, and an alternative
explanation (e.g., inadequately controlled features
of the studies' experimental designs) for the
observed lack of effects is not available. The
experiments were designed to specifically test for
effects of interest, including suitable exposure
timing and duration, post-exposure latency, and
endpoint evaluation procedures, and to address
potentially susceptible populations and life siaues
Several high-confidence studies, showing
consistently null results (e.g., an odds ratio of 1.0)
ruling out alternative explanations including
chance, bias, and confounding with reasonable
confidence. Each of the studies should have used
an optimal outcome and exposure assessment and
adequate sample size (specifically for higher
exposure groups and for sensitive populations).
The set as a whole should include the full range of
levels of exposures that human beings are known
to encouiikT. an evaluation of an exposure
response gradient, and at-risk populations and life
siaues and should be mutually consistent in not
slum iiiu au> indication of effect at any level of
exposure.
Note:
a "Experiment" refers to measurements in a single population of exposed animals (e.g., a study that included separate evaluations
of rats and of mice, or separate cohorts exposed at different life stages, would be considered as multiple experiments).
Conversely, two papers or studies that report on the same cohort of exposed animals (e.g., examining dilTerenl endpoints) would
not be considered separate experiments.
The primary evidence and rationale supporting these decisions were summarized in a single
evidence profile table to transparently convey the aspects of ihc c\ idence that were considered to
increase or decrease the hazard support for each health effect, for the purposes of this
assessment, only the integrated evidence that supports a hazard was considered for use in the
dose-response and demation of toxicity values.
2.3.7 Derivation of Values
Development of the dose-response assessment for PI BS and/or the potassium salt has followed
the general guidelines for risk assessment put forth by the National Research Council (NRC.
1983) and the EPA's l'ramework Jor //// man Health Risk Assessment to Inform Decision Making
(U.S. EPA. 2014b). Other I -PA guidelines and reviews considered in the development of this
assessment include the follow ing
• A Review <>j the Reference Dose and Reference Concentration Processes (U.S. EPA.
2002).
• A Framework for. \ssessing Health Risks of Environmental Exposures to Children (U.S.
EPA. 2006).
• Exposure Factors Handbook (U.S. EPA. 201 la).
• Recommended Use of Body Weight3/4 as the Default Method in Derivation of the Oral
Reference Dose (U.S. EPA. 201 lb).
• Benchmark Dose Technical Guidance Document (U.S. EPA. 2012).
• Child-Specific Exposure Scenarios Examples (U.S. EPA. 2014a).
17
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
The EPA's A Review of the Reference Dose and Reference Concentration Processes document
describes a multistep approach to dose-response assessment, including analysis in the range of
observation followed by extrapolation to lower levels (U.S. EPA. 20021 As described above,
prior to deriving toxicity values, the EPA conducted a comprehensive evaluation of available
human epidemiological and animal toxicity studies to identify potential health hazards and
associated dose-response information through the literature search and screening, study
evaluation, evidence synthesis, and evidence integration steps. This evaluation informed the
selection of candidate key studies and critical effects for dose-response analysis, from which the
EPA identified a critical effect and point of departure (POD) for subchronic and chronic
reference value derivation and extrapolated a selected POD lo a corresponding RfD (e.g.,
subchronic RfD). For dose-response analysis of PFBS and or the potassium salt, the EPA used
the BMD approach to identify a POD. The steps for deri\ ing ail RID using the BMD approach
are summarized below.
• Step 1: Evaluate the data to identify and characterize end points related to exposure
to PFBS chemicals. This step involved determining the rele\anl studies and adverse
effects to be considered for BMD modeling Once the appropriate data were collected,
evaluated for study quality, and characterized for acKerse outcomes, end points were
selected that were judged to be relevant (i e . Ibr the purposes of this assessment, effects
that were sufficient to support a hazard) and sensiti\ e as a function of dose (typically
defined by the no observed ad\ el se effect le\ el | M).\l II. | value). In this assessment,
these decisions were directly informed In the evidence integration judgments arrived at
for each assessed health outcome Some of the most important considerations that
influenced selection of endpoints for IJ VI I) modeling include data with dose-response,
percent change from controls, adversity of effect, and consistency across studies. For
PFBS, kidney, thyroid, and de\ elopmental endpoints were considered for toxicity value
derivations.
• Step 2: Convert the adjuslcd daily doses lo an HED. The adjusted daily doses were
coin erted to 11111) s using the I.IWs Recommended Use of Body Weight3/4 as the Default
h lethod in / K-nvation of the (hal Reference Dose (U.S. EPA. 201 lb). Study-reported
body weight (liW) was used when available; otherwise, default animal BW information
was retrieved from I S III * A (1988).
• Step 3: Select the benchmark response (BMR) level. Using the EPA's Benchmark
Dose Technical Guidance Document (U.S. EPA. 2012). the endpoints selected were
modeled. The BVIR is a predetermined change in the response rate of an adverse effect. It
serves as the hasis for obtaining the benchmark dose lower confidence limit (BMDL),
which is the ^5".. lower bound of the BMD. BMRs were identified and applied consistent
with quantal and continuous data and, when possible, informed by understanding of
biological significance.
• Step 4: BMD Model the data. This step involved fitting a statistical model to the
dose-response data that describes the data set of the identified adverse effect. Typically,
this involved selecting a family or families of models (e.g., polynomial continuous, hill
continuous, or exponential continuous) for further consideration based on the data and
experimental design. In this step, a BMDL was derived by placing confidence limits
(one- or two-sided) and a confidence level (typically 95%) on a BMD to obtain the dose
that ensures with high confidence that the BMR is not exceeded.
18
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
• Step 5: Determine a PODhed. If modeling was feasible, the estimated BMDL(HED)s
were used as PODs (i.e., PODhed). If dose-response modeling was not feasible, NOAEL
(HED)s or lowest observed adverse effect level (LOAEL) (HED)s were identified.
• Step 6: Provide rationale for selecting Uncertainty Factors (UFs). UFs were selected
in accordance with EPA guidelines considering variations in sensitivity among humans,
differences between animals and humans, the duration of exposure in the key study
compared to a lifetime of the species studied, and the potential limitations of the
toxicology database.
• Step 7: Calculate the subchronic and chronic RfDs. The RfDs were calculated by
dividing a PODhed by the selected UFs.
RfD = PODhed
UFc
where:
PODhed = The PODhed is calculated from the BMDL or NOAN. using aBW3'4
allometric scaling approach consistent with I-PA guidance (U.S. EPA. 2"| lb)
UFc = Composite UF established in accordance with l-l'A guidelines considering
variations in sensitivity among luimans. differences between animals and humans, the
duration of exposure in the key study compared to a lifetime of the species studied, and
the potential limitations of the toxicology database
19
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
3.0 Overview of Evidence Identification for Synthesis and
Dose-Response Analysis
3.1 Literature Search and Screening Results
The database searches yielded 373 unique records, with 50 records identified from additional
sources such as TSCA submissions, posted NTP study tables, peer-review recommendations, and
review of reference lists from other authoritative sources. Of the 373 studies identified, 261 were
excluded during title and abstract screening, 112 were reviewed at the full-text level, and 37
were considered relevant to the PECO eligibility criteria (see Figure 4). This included 14
epidemiologic studies (described in 17 publications), 10 /// vivo animal studies (described in 15
peer-reviewed and nonpeer-reviewed publications), and li\ c /// vin o genotoxicity studies. The
detailed search approach, including the query strings and IM-CO criteria, is provided in
appendix A and appendix B, respectively.
20
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Other
Sources
n = 50
Records after Duplicate Removal
n = 323
Excluded/Not on Topic
n = 261
Title & Abstract Screen after
Duplicates Removed
n = 373
Full-Text Screen
n = 112
Included Based on PECO
Relevance
n = 37
Human (n = 17)
Animal (n = 15)
In vitro genotoxicity (n = 5)
Excluded/Not PECO Relevant
n = 75
Not relevant to PECO (n = 18)
Not relevant to PECO, but
supplemental information (n
= 38)
Review, commentary, letter
(n = 19)
Figure 4. Literature search and screening flow diagram for PFBS (CASRN 375-73-5).
ML— 1
PubMed
n a 217
(
Toxline
n = 4
u J
wos
n s 307
V
TSCATS via
Toxline
n = 0
21
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
3.2 Study Evaluation Results
Based on the study evaluations, seven human epidemiology studies were considered
uninformative and are not discussed any further in this assessment (see Table 4). No animal
studies were considered uninformative and, thus, all animal studies identified as relevant during
literature screening were included in the evidence synthesis and dose-response analysis. Overall,
seven epidemiologic studies (described in 10 publications) and 10 in vivo animal studies
(described in 15 peer-reviewed and nonpeer-reviewed publications) were included in the
evidence synthesis and further evaluated for use in the development of toxicity values for PFBS.
As shown in Figures 5 and 6, while the database of studies on PFBS is not large, a number of
high- and medium-confidence oral exposure studies in animals were identified, as were several
medium-confidence studies in humans. No studies were identified evaluating the toxicity of
PFBS or K+PFBS following inhalation exposure or on the carcinogenicity of PFBS or K+PFBS
in humans or animals.
Table 4. Epidemiological studies excluded based on study evaluation
Reference
Outcome
Reason lor exclusion
Bao et al. (2017)
Blood pressure
1 Aii'emek poor sensitivity (96% of participants below the
1.()l) i w nh no observed association.
Berketal. (2014)
Depression
Serious concerns \\ n ti temporality between exposure and
outcome, confounding, and analysis.
Gvllenhammar et al. (2018)
Birth size, weight uaiu
1 Ati'cmely poor seiisinvity (median exposure = 0.01 ng/g)
u uli no observed association.
Kim et al. (2016)
(oimeinial
lis poili\ roidism
1 Acludcd from full statistical analysis by study authors
because of high percent below the LOD (72%).
Seo et al. (2018)
( holesierol. uric acid,
diaheles. 1 'All. ihvroid
hormones
No consideration of potential confounding.
Shiue <2<>|<.)
Sleep disiinhaiiees
Not evaluated due to nonspecific effect.
Warn: el al <2
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
,0°
oSNNeV' '\e°• '
Participant selection
Exposure measurement
Outcome ascertainment
Confounding
Analysis
Sensitivity -
Selective Reporting
Overall confidence
Good (metric)
or High
(overall)
Adequate (metric)
or Medium
(overall)
Deficient
(metric) or
Low (overall)
N/A
Not assessed due
to critical
deficiency in
other domain
I
Critically deficient
(metric) or
Uninformative
(overall)
Figure 5. Evaluation results for epidemiological studies assessing effects of PFBS
(click to see interactive data graphic for rating rationales).
23
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
0
^v>eA\^V?
Reporting
Allocation
Blinding
Variable Control
Selective Reporting & Attrition
Exposure Characterization
Utility of Study Design -
Outcome Assessment -
Results Presentation -
+ +
+ +
+ + + +
++
++
++
++
++
++
+ +
+ +
~~
++
++
++
++
++
++
NR
++
++
++
D
+ +
+ +
+ + + +
++
++
++
++
++
++
+ +
+ +
ni i
++
++
D
++
++
++
+ +
+ +
++
++
++
++
++
++
++
++ n ++ ++ ++ ++ ++ ++ ++
++ ++ ++ ++ ++ ++ ++ ++ ++ ++
Overall confidence
Good (metric)
Adequate (metric)
Deficient
Not
Critically deficient
++
or High
+
or Medium
-
(metric) or
NR
reported
(metric) or
(overall)
(overall)
Low (overall)
for metric
J
Uninfonnative (overall)
Figure 6. Evaluation results for animal studies assessing effects of PFBS exposure
(click to see interactive data graphic for rating rationales).
24
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
4.0 Evidence Synthesis: Overview of Included Studies
The database of all repeated-dose oral toxicity studies for PFBS and the related compound
K+PFBS that are potentially relevant to the derivation of RfD values includes a short-term range
finding study in rats (3M. 2000d). two 28-day studies in rats (NTP. 2018. 2011; 3M. 2001). one
subchronic-duration study in rats (Lieder et al.. 2009a; York. 2003b). one subchronic-duration
lipoprotein metabolism study in mice (Biiland et al.. 2011; 3M. 2010). three gestational exposure
studies in mice and rats (Feng et al.. 2017; York. 2003a. 2002). and one two-generation
reproductive toxicity study in rats (Lieder et al.. 2009b; York. 2003c. d, e). In addition, seven
epidemiologic studies (described in 10 publications) were identified that report on the association
between PFBS and human health effects. Specific study limitations identified during evaluation
(see HAWC) are discussed only for studies interpreted as low confidence or if a limitation
impacted a specific inference for drawing conclusions
Human and animal studies have evaluated potential effects on the thyroid, reproductive systems,
developing organism, kidneys, liver, and lipid and lipoprotein homeostasis following exposure to
PFBS. The evidence base for these outcomes is presented in this section I-'or each potential
health effect, the synthesis describes the database of human and animal studies, as well as an
array of the animal results across studies. NOAEI ,s and I ,().\l-l ,s in presented in figures and text
are based on statistical significance and 'or biological significance (e.g., directionality of effect
[statistically significantly decreased cholesterol/triglycerides is of unclear toxicological
relevance], abnormal or irregular dose-response | nonmonotonicty], tissue-specific
considerations for magnitude of effect [nonstatistically significant increase of >10% in liver
weight interpreted as biologically significant]) For this section. e\ idence to inform organ-
/system-specific effects of PI liS in animals following de\ elopniental exposure is discussed in
the individual organ-/system-specific sections (eg. reproductive cycling endpoints after
developmental exposure are discussed in "Reproductive Effects"). Other effects informing
potential developmental effects (e g . pup B\V) are discussed in the "Offspring Growth and Early
Development" section
Evidence integration analyses and overall judgments on the hazard support for each outcome
domain pro\ ided by the a\ailable human and animal studies are discussed in "Evidence
Integration and Hazard Characterization " Notably, in that section, the evidence informing
organ-/system-specific endpoints after developmental exposure was considered potentially
informative to both the de\ elopniental effects outcome domain and the organ-/system-specific
outcome domain
4.1 Thyroid Effed
4.1.1 Human Studies
No human studies were available to inform the potential for PFBS exposure to cause effects on
the thyroid.
4.1.2 Animal Studies
Two high-confidence studies evaluated the effects of PFBS exposure on thyroid, specifically
thyroid hormone levels, thyroid histopathology, and thyroid weight (NTP. 2018; Feng et al..
2017; NTP. 2011) (see Figure 7). Dams exposed to K+PFBS through gestation (gestation days
[GDs] 1-20) exhibited statistically significantly decreased total triiodothyronine (T3).
25
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
total thyroxine (T4). and free T4 (reduced 17%, 21%, and 12%, respectively, relative to control
at 200 mg/kg-day and reduced 16%, 20%, and 11%, respectively, relative to control at
500 mg/kg-day) on GD 20 at doses of 200 and 500 mg/kg-day, but not at 50 mg/kg-day (Feng et
al.. 2017). Decreased total T3 and total T4 were also reported at postnatal day (PND) 1, PND 30,
and PND 60 in offspring gestationally exposed to K+PFBS at the same doses (up to 37%
reduction in T3 and 52% reduction in T4). Increased thyroid-stimulating hormone (TSH) was
reported in dams and pubertal (PND 30) offspring (21% and 14% relative control at
200 mg/kg-day, respectively) exposed gestationally to K+PFBS. Statistically significant
dose-dependent decreases in total T3. total T4. and free T4 were also reported after exposure in
male and female rats to K+PFBS for 28 days at all doses tested (> 62.6 mg/kg-day) (NTP. 2018.
2011). The reported reductions in total T3 were up to -57% and -43% in male and female rats,
-86% and -77% in free T4, and -97% and -71% in total T4, respectively. Dose-response
graphics for T4, T3, and TSH, including effect size and variability, are included in appendix E,
Figures E-l, E-2, and E-3, respectively. Thyroid gland weight, thyroid histopathology, and
TSH levels were not changed after 28 days of PFBS exposure in male or female rats at up to
1,000 mg/kg-day (NTP. 2018. 20111
Kndpolnt Name
Study Nairn1
Study Type
Animal Description
Observation Time
PFBS Thyroid Effects
Tetraiodoihyronine
PNDI
—
~
PND30
*-•
~
PND60
M
^
~
Triiodothyronine (T3)
NTP 2018, 4309741
short-term (28 days)
Rat. Ilarian Sprague-Dawley ((f)
Day 28
~ •
Rat. Harlan Sprague-Dawley (9)
Day 28
~ ~
Feng 2017. 3856465
developmental (GDI to 20)
TO Mouse. ICR (9)
GD20
V
V
Fl Mouve. ICR (9)
PNDI
-~
~
PND30
-~
~
PND60
M
w
Thyroid Stimulating Hormone (TSH)
NTP 2018. 4309741
short-term (28 days)
Rat. Harlan Spraguc-Dawlcy ((f)
Day 28
» • •
Rat. Harlan Sprague-Dawley (9)
Day 28
~ • •
• •
Feng 2017. 3856465
developmental (GD 1 to 20)
TO Mouse. ICR (9)
GD20
M
-A
A
Fl Mouse, ICR (9)
PNDI
PND30
M
~
PND60
M
Thyroid Weight. Absolute
NTP 2018. 4309741
short-term <28 days)
Rat. Harlan Spraguc-Dawlcy «f)
Day 28
» • •
• ~
RaU Ilarian Sprague-Dawley (9)
Day 28
~ » »
• ~
Thyroid Weight. Relative
NTP 2018. 4309741
short-term <28 days)
Rat. Harlan Sprague-Dawley «f)
Day 28
~ • •
• ~
Rat. Ilarian Sprague-Dawley (9)
Day 28
~ • •
• •
Thyroid Histopathology
NTP 2018. 4309741
short-term (28 days)
Rat. Harlan Spraguc-Dawlcy ((f)
Day 28
~ • •
• *
RaU Ilarian Sprague-Dawley (9)
Day 28
~ « «
• ~
3M, 2000.4289992
short-term (10 days)
Rat. Cri: Cd (Sd) lbs Brief)
Day 11
• •—
~
Rut. Cri: Cd (Sd) lbs lir (9)
Day II
~ •
~
-1
no o ioo
200 300
400 500 600 700 800 900 1.000 1.
00
Dose imj/kp/dayl
Figure 7. Thyroid effects from K+PFBS exposure (click to see interactive data graphic and
rationale for study evaluations for effects on the thyroid in HAWC).
4.2 Reproductive Effects
4.2.1 Human Studies
Three studies that evaluate populations in China and Taiwan examined different reproductive
outcomes in women and men (Song et al.. 2018; Zhou et al.. 2017a; Zhou et al.. 2016).
26
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Although not statistically significant, one low-confidence cross-sectional study (Zhou et al..
2017a) reported adjusted odds ratios (OR) of 1.30 (95% CI: 0.54-3.12) for menorrhagia and 1.48
(95% CI: 0.54-4.03) for hypomenorrhea in preconception women in China for each one unit
change in PFBS. They also reported inverse statistically nonsignificant associations for these two
outcomes, however, based on exposure quartiles (OR range: 0.61-0.84 for the highest quartiles
relative to the referent) with no evidence of an exposure-response relationship. The potential for
reverse causation from uncertain temporality from these cross-sectional data temper any
conclusions that might be drawn from this one study.
One low-confidence study (Zhou et al.. 2016) in Taiwanese adolescents reported no clear
associations between PFBS levels and reproductive hormones among the entire population or
stratified by sex.
One low-confidence cross-sectional study (Song et al . 2"1S) examined the association between
PFBS exposure and semen parameters. There was no indication of decreased semen quality in
this study (correlation coefficients of-0.022 for semen concentration and
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
(i.e., lactation day [LD] 22), and observed no effects compared to controls at 1,000 mg/kg-day;
however, the data were not reported. These parameters were not evaluated in York (2002).
4.2.2.2 Male fertility
Two studies using S-D rats evaluated several potential responses in the male reproductive system
(NTP. 2018; Lieder et al.. 2009b 1 Male fertility parameters and reproductive effects (e.g., sperm
parameters) were generally unaffected by K+PFBS treatment in P0- and F1-generation males
observed by Lieder et al. (2009b). At the highest dose, there were statistically significant
increases in the percentage of abnormal sperm in F1 animals and decreases in testicular sperm
count in PO-generation males. In addition, the study authors report the number of spermatids per
gram testis was within the historical control of the testing facility These effects were not dose-
dependent. Alterations in parameters such as sperm count number and morphology are
considered indicative of adverse responses in the male re|">roducli\ e system (Foster and Gray.
2013; Mangelsdorf et al.. 2003; U.S. EPA. 1996a) A 28-day exposure study reported a
decreased trend in testicular spermatid count per mu testis evaluated al the time of necropsy;
however, no significant effects on other sperm measures were reported, including caudal
epididymal sperm count and sperm motility (N IP. 2<)18).
The differences in responses observed in the two available studies might have been due to
experimental design differences as Lieder et al. (200^b ) exposed P0 animals for 70 days and F1
animals during the entire period of gestation plus lactation, u hereas NTP (2018) exposed
animals for 28 days. Future studies should be developed to ascertain whether long-term and/or
gestational exposure to PFBS significantly affects sperm measures in sexually mature and
developing animals.
4.2.2.3 Reproduci ,,3 an
Reproductive hormones were ev aluated in mice (I'eng et al.. 2017) and, to a limited extent, in
rats (NTP. 2018. 2" I I) (see I 'iuure 8) I-xposure to k PFBS for 28 days resulted in a significant
trend lor increased testosterone levels in females, but not in males (NTP. 2018. 2011). The
increase in testosterone was not statistically significant when compared to control at any dose by
pairw ise analysis. Other reproductiv e hormone levels were not measured. Prenatal exposure to
PFBS at and above 200 mg. kg-dav resulted in statistically significant reduced serum estradiol
levels and increased serum luteinizing hormone levels in pubertal offspring (i.e., PND 30) (Feng
et al.. 2017). The change in serum estradiol levels, but not luteinizing hormone, continued into
adulthood in the k PLIJS-exposed offspring (i.e., PND 60). Adult PFBS-exposed offspring also
exhibited decreased serum progesterone levels at doses of 200 mg/kg-day and greater. PFBS
exposure did not alter maternal estradiol-, progesterone-, or gonadotropin-releasing hormone.
Reproductive hormone levels in males and females were not evaluated by Lieder et al. (2009b).
The changes in follicle and corpora lutea development reported in the same study, however, may
be associated with alterations in hormone production/levels, as ovarian follicles and corpora
lutea produce estrogen and progesterone, respectively (Foster and Gray. 2013; U.S. EPA. 1996a).
The hormonal effects observed in the NTP (2018) and Feng et al. (2017) studies might be
associated with adverse reproductive effects reported in these studies. Androgens, luteinizing
hormone, estradiol, and progesterone play an important role in normal development and
functions of the female reproductive system (Woldemeskel. 2017; Foster and Gray. 2013).
Alterations in the levels and production of these reproductive hormones can disrupt endocrine
28
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
signals at the hypothalamic-pituitary level and lead to delayed reproductive development and
changes in functions (Rudmann and Foley. 2018; Woldemeskel. 2017; Foster and Gray. 2013).
Kndpoint
Study
Kxposure
Animal (.roup
Observation time
PI BS Reproductive Hormone Effects
Testosterone (T)
NTP 2018,4309741
28 Day Oral
Rat. llarlan Spraguc-Dawlcy (9)
Day 28
* •
—•
Rat. Harlan Sprague-Dawley (Cf)
Day 28
» •
-• •
•—
Estrogen
Feng 2017, 3856465
20 Day Onil Gestation
P0 Mouse. ICR (9)
GD20
• Doses
Fl Mouse. ICR (9)
PNDI
A Significant Increase
PND30
~
—V
V Significant Descrease
PND60
•V
~
—v
H Dose Range
Progesterone (P4)
Feng 2017, 3856465
20 Day Oral Gestation
P0 Mouse. ICR (9)
GD20
• •
•
—•
Fl Mouse. ICR (9)
PNDI
PND30
—•
PND60
•V
~
—~
Luteinizing Hormone (LH)
Feng 2017. 3856465
20 Day Oral Gestation
Fl Mouse. ICR (9)
PNDI
• «—
•
—•
PND30
• A
~—
—~
PND60
Gonadotropin Releasing Hormone (GnRH)
Feng 2017. 3856465
20 Day Oral Gestation
PO Mouse. ICR (9>
GD20
Fl Mouse. ICR (9)
PNDI
PND30
• •
•
—~
PND60
—•
1 1 1 1 1 1 1 1 1 1 1
-100 0 100 200 300 400 500 600 700 800 900 1.000 1.100
Dose (mg/kg/day)
Figure 8. Reproductive hormone response to K+PFBS exposure (click to see interactive
data graphic and rationale for study evaluations for reproductive hormone levels in
HAWC).
4.2.2.4 Reproductive system development, including markers of sexual differentiation and
maturation (female and male)
Several measures of female reproductive development were affected by gestational K+PFBS
exposure in mice (see Figure 9). Feng et al. (2017) reported a delayed first estrous in female
PFBS-exposed offspring (> 200 mg/kg-day) compared to control. Estrous cyclicity was also
affected in K+PFBS-exposed PNDs 40-60 offspring as exhibited by a prolongation of the
diestrus stage compared to control. Estrous cycling was generally not statistically significantly
altered in P0- or F1-generation females treated with K+PFBS in the two-generation study by
Lieder et al. (2009b). An increase in the number of rats with > 6 consecutive days of diestrus was
observed in the F1 females exposed to 100 mg/kg/day; however, the increase was not present at
higher doses (Lieder et al.. 2009b). Estrous cyclicity was affected after adult exposure to
K+PFBS for 28 days exhibited by a dose-dependent prolongation of diestrus at doses of
250 mg/kg-day and greater with marginal significance at the lowest dose tested (125 mg/kg-day)
(p = 0.063) (NTP. 2018. 2011). Lieder et al. (2009b) reported a delay in the days to preputial
separation in F1 males of the 30- and 1,000-mg/kg-day groups;5 however, the measure was no
longer statistically significant when adjusted for BW. There was similarly no change in the days
to vaginal patency in F1 female rats (Lieder et al.. 2009b). Unlike Lieder et al. (2009b). Feng et
al. (2017) reported a delay in vaginal patency in F1 females after gestational exposure of
200 mg/kg-day and greater.
5 A marker of delayed reproductive development (ENR.EF3 7: _ENREF_110).
29
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
F.ndpoinl Name
Slink Name
I'Apc rimenl Name
Animal Description
Observation Time
PFBS Reproductive l>evelopmenl and Fstrous Cycling
26 Da> s of Dicstnis
Lieder. 2009. 1578545
2 Generation Oral
F1 Ral. Sprague-Dawley (9)
N-A • ~
• Doses
PI) Rat. Sprague-Dawley (9)
6 days
N—• ~ *
(—1 Dose Range
26 Days of Eilrus
Lieder. 2009, 1578545
2 Generation Oral
F1 Rat. Sprague-Dawley (9)
—• • ~
^ Significant Increase
PI) Rat. Sprague-Dawley (9)
6 days
N—l • ~
^ Significant decrease
hsirous Cycle. Diestrus
Feng 2017.11856465
20 Day Oral Gestation
Fl Mouse. ICR (9)
PND40-60
A A
VTP 2018, 4309741
28 Day Oral
Ral. Harlan Sprague-Dawley (9)
Day 28
• A A ~
Estrous Cycle, Eslrus
Feng 2017. 3856465
20 Day Oral Gestation
Fl Mouse. ICR <9)
PND40-60
~ • • ~
NTP 2018, 4309741
28 Day Oral
Ral. Harlan Sprague-Dawley (9)
Day 28
~ • • ¦ • •
Estrous Cycle. Metestrus
NTP 2018. 4309741
28 Day Oral
Rat Harlan Sprague-Dawley (9)
Day 28
~ • • 1 • ~
Estrous Cycle. Proestrus
Feng 2017. 3856465
NTP 2018, 4309741
20 Day Oral Gestation
28 Day Oral
Fl Mouse. ICR (9)
Rat Harlan Sprague-Dawley (9)
PND40-60
Day 28
M • ~
~ • • • • •
Estrous Stage. Diestrus (al sacrifice)
lieder. 2009, 1578545
2 Generation Oral
Fl Ral, Sprague-Dawley (9)
Sacrifice
—• • ~
HO Kat. Sprague-Dawley (9)
M—• • ~
F.slrotis Stage. Hstrus (al sacrifice)
Lieder. 2009.1578545
2 Generation Oral
Fl Ral. Sprague-Dawley (9)
Sacrifice
~«—• • •
PI) Rat. Sprague-Dawley (9)
~ l » » »
Estrous Sutge. Mctcslius (at sacrifice)
Lieder. 2009. 1578545
2 Generation Oral
Fl Rai. Sprague-Dawley (9)
Sacrifice
M » » »
FO Rat. Sprague-Dawley (9)
M—• • ~
Estrous Stage. Proestrus (al sacrifice)
Lieder. 2009. 1578545
2 Generation Oral
Fl Ral. Sprague-Dawley (9)
Sacrifice
¦ ¦ ~
ID Rat. Sprague-Dawley <9>
M > » »
Estrous Stages/ 21 Days
Lieder. 2009. 1578545
2 Generation Oral
Fl Rat. Sprague-Dawley (9)
M • • ~
PO Rat. Sprague-Dawley (9)
H—• ~ «
Firsl Estrous - Litter N
Feng 2017. 3856465
20 Day Oral Gesialion
Fl Mouse. ICR (9)
Beginning on PND24
~- ~ ~
Preputial Separation
lieder. 2009. 1578545
2 Generation Oral
Fl Ral. Sprague-Dawley (Cf)
Day 29 (postpartum)
«A— A
Vaginal Opening
Feng 2017. 3856465
20 Day Oral Gestation
Fl Mouse. ICR 19)
PND60
•i
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
early development following PFBS exposure, including two gestational exposure studies in rats
(York. 2003a. 2002) and one gestational exposure study in mice (Feng et al.. 2017). as well as a
two-generation study in rats (Lieder et al.. 2009b; York. 2003c) (click to see interactive graphic
for Developmental Effects in HAWC).
None of the studies identified significant effects in either rats or mice on measures of fetal
alterations (i.e., malformations and variations). BW of female offspring of PFBS-exposed mice
at doses greater than 200 mg/kg-day was statistically significantly lower than control at PND 1,
and the pups remained underweight through weaning, pubertal, and adult periods, with decreases
of approximately 25% observable in pups nearing weaning (Feng et al.. 2017). At these ages
(i.e., around PND 16), Feng et al. (2017) also reported an I 5-day developmental delay in eye
opening in pups gestationally exposed to 200 mg/kg-day PIUS and greater. Importantly, no
effects on the health of the exposed dams (e.g., weight uai n) u ere observed at any dose (Feng et
al.. 2017). Dose response graphics for eye opening, including effect size and variability, are
included in appendix E, Figure E-4. Fetal BWs (male and female) were also reduced
(approximately 10%) compared to controls follow ing gestational exposure from GDs 6 to 20 at
the highest tested dose (1,000 mg/kg-day in York (2""2)1 and 2,000 mg ku-dav in York
(2003a)]). Parental BWs and organ weights, howe\ er. were also affected to a similar degree at
those doses (Lieder et al.. 2009b; York. 2<~)<~)3c. 2002). limiting the interpretation of the results.
No statistically significant changes in I ' I - and F2-genei ation pups mean pup weight at birth and
mean pup weight at weaning were reported In l.icder el al. (Z'tii^b) or York (2003c).
Several measures of thyroid hormone de\ elopment and female reproductive development were
affected by gestational Pl-'liS exposure in mice and are described in more detail in "Thyroid
Effects" and "Reproducti\e F.ffccts." respecti\ely.
4.4
4.4.:
One mcdium-conlldcncc study (Oin el al ( 2" !(•»). u ith additional details in Bao et al. (2014).
selected 225 subjects ages 12 15 years old from a prior cohort study population in seven public
schools in northern Taiwan (Tsai et al.. 2010) and examined the association between PFBS
exposure and uric acid concentrations. There was no association between ln(PFBS) concentration
and uric acid concentrations in the total population (P = 0.0064 mg/dL increase in uric acid per
1 ln-|ig/L increase in ITBS. l>5"n (!1 = -0.22, 0.23). A nonsignificant positive association in boys
was offset by a nonsignificant negative association in girls, and there is not enough information
to determine whether there are sex-specific susceptibility differences. When PFBS exposure was
analyzed for high uric acid ( (•> mg/dL), the risk was somewhat elevated in boys (OR = 1.53;
95% CI: 0.92, 2.54), but not in girls (OR = 0.99; 95% CI: 0.58, 1.73). The potential for reverse
causation tempers any conclusions that might be able to be drawn.
4.4.2 Animal Studies
Renal effects were evaluated in high-confidence short-term and subchronic-duration exposure
studies in rats (NTP. 2018. 2011; Lieder et al.. 2009a; 3M. 2001. 2000d) and in a high-
confidence two-generation reproductive study in rats (Lieder et al.. 2009b). Endpoints evaluated
in these studies include kidney weights, histopathological changes, and serum biomarkers of
31
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
effect (see Figure E-8 and Figure E-9). Dose-response graphics for histopathological effects,
including effect size and variability, are included in appendix E, Figure E-7.
Absolute and relative kidney weights of male and female S-D rats were unchanged in rats
exposed daily for 90 days to K+PFBS at doses up to 600 mg/kg-day compared to control rats
(Lieder et al.. 2009a). This lack of effect on kidney weight was also observed in parental and F1
male and female rats of the same strain exposed to K+PFBS at doses up to 1,000 mg/kg-day
during a two-generation reproductive study (Lieder et al.. 2009b). Although none of the findings
reached statistical significance, however, an approximate 9% increase in absolute kidney weight
was observed in female S-D rats exposed to 1,000 mg/kg-day T\ PFBS for 10 days (3M. 2000d);
relative-to-body kidney weights were also increased approximately 6%-9%. This organ-weight
effect was not observed in corresponding males of the study. In a follow-on 28-day study by the
same lab, a 9%-l 1% increase in absolute and relative-to-hody kidney weight was observed in
female S-D rats exposed to 900 mg/kg-day K+P1 liS (3M. 2001). although these changes were
not statistically significant. Smaller, nonsignificant increases in kidney weight were also
observed in male rats. In another 28-day study, k Pl-'liS exposure significantly increased
absolute and relative right kidney weights in high-dose male (500 mg/ku-day) S-D rats (NTP.
2018. 2011). Only relative-to-BW kidney weights were altered in female rats, hut this effect was
significant at all tested K+PFBS doses ( 62.6 mu ku-day) Click to see interactive graphic for
Kidney Weight Effects in HAWC.
After 90 days of exposure, Lieder et al. (2""^ti) observed increased incidences of
histopathological alterations of the kidneys of male and female rats of the high-dose group
(600 mg/kg-day). Increased incidence of hyperplasia of the epithelium of renal papillary tubules
and ducts was obser\ ed in rats of both sexes (see Figures E-7 and E-8). A single incidence of
papillary necrosis in both kidneys was observed in one male in the high-dose group. Further,
focal papillary edema was observed in 3/10 rats of both sexes of the high-dose groups compared
to no evidence of this effect in control rats. Similar histopathological alterations were observed
in parental and F 1 male and female rats in the two-generation reproduction study (Lieder et al..
2009b). Compared to control rats, increased incidences of hyperplasia of the renal tubular and
ductal papillary epithelium, and focal papillary edema were observed in parental male and
female rats at PI BS doses 3<)<) mg/kg-day. Hyperplastic foci in the same locations of the
kidney were also observed in male and female F1 rats exposed to > 300 mg/kg-day PFBS across
life stages from gestation to adulthood (Lieder et al.. 2009b). Focal papillary edema was
observed in male Q I .<><><"> mg'kg-day) and female (>300 mg/kg-day) F1 rats, although this
specific alteration did not appear lobe dose-dependent in females. Although kidney alterations
such as hydronephrosis, mineralization, and tubular degeneration were observed in male or
female S-D rats after just 10 days of oral K+PFBS exposure, these effects were not significant
compared to control and/or did not appear to be dose-dependent (3M. 2000d). The same
histopathological lesions were noted in the 28-day rat study albeit with lack of significance
compared to control or dose-dependence (3M. 2001). In another 28-day oral gavage study in S-D
rats, chronic progressive nephropathy (CPN) was observed in all male and female PFBS
treatment groups and control rats, with no evidence of dose-dependence for this effect (NTP.
2018. 2011). Renal papillary necrosis was also observed in these rats but only at the highest
exposure dose (1,000 mg/kg-day).
32
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Serum levels of biomarkers indicative of kidney injury and/or function, including blood urea
nitrogen (BUN) and creatinine, have been examined across multiple studies of varying exposure
durations, and were found to be unchanged in male and female rats treated with K+PFBS at doses
up to 1,000 mg/kg-day (Lieder et al.. 2009a; 3M. 2001. 2000d). After 28 days of oral gavage
exposure in S-D rats, however, NTP (2018. 2011) observed significantly increased levels of
BUN in males (> 250 mg/kg-day). This increased circulating BUN was not observed in female
rats at doses up to 1,000 mg/kg-day. Click to see interactive graphic for other Kidney Effects in
HAWC.
4.5 Hepatic Effects
4.5.1 Human Studies
No human studies were available to inform the potential for PI-'IJS exposure to cause hepatic
effects.
4.5.2 Animal Studies
Hepatic effects were evaluated in high-confidence short-term and subchronic-duration studies in
rats (NTP. 2018. 2011; Lieder et al.. 2009a; 3±\1. 2<") 1. 2"""d) and in a high-conlidence
two-generation reproductive study in rats (Lieder el a I . 2"i)tJ|i) I '.ndpoints e\ aluated in these
studies include liver weights, histopathological changes, and serum biomarkers of effect (see
Figure E-10).
Ten days of daily oral gavage exposure to k Pl-'liS significantly increased absolute,
relative-to-body, and relali\e-lo-brain weights of li\er in adult male and female S-D rats exposed
to 1,000 mg/kg-day (3\l. 2<»)"d'). The absolute li\ er mass of male rats was increased by 36%
compared to females (22"..) A similar profile of li\ er weight alteration in S-D rats was observed
following 28 days of exposure where absolute and relative liver weights of high-dose
(900 mg ku-da\) male rats were increased 25%- 3<)% (3M. 2001). Female rats of the same
treatment dose did not experience a similar magnitude increase in absolute or relative liver
weights (4°.. (•>%) In another 2S-day study in S-l) rats, K+PFBS exposure significantly increased
absolute and relative liver weights in males (> 125 and > 62.6 mg/kg-day, respectively) and
females ( 25<) and > 125 mg kg-day, respectively) (NTP. 2018. 2011). In contrast, the livers of
male and female S-D rats exposed to K PFBS at doses up to 600 mg/kg-day for 90 days were not
significantly changed compared to respective controls (Lieder et al.. 2009a). In a two-generation
reproduction study using the same strain of rat, however, increased absolute and relative liver
weights were obser\ed in male parental rats exposed to doses of K+PFBS > 300 mg/kg-day for
approximately 70 days (J.ieder et al.. 2009b). In the F1 adult males, only relative liver weight
was significantly increased al the high dose (1,000 mg/kg-day), although terminal BW was
significantly decreased in this group compared to control.
Histopathological examination of the livers of S-D rats across three separate oral gavage studies
of increasing K+PFBS exposure duration (10-day (3M. 2000d)1; 28-day (3M. 2001)1; 90-day
(Lieder et al.. 2009a)1) did not reveal any significant dose-dependent alterations or lesions. For
example, focal/multifocal hepatic inflammation was observed in 3/10 male and 4/10 female rats
of the high-dose group (no incidence at the low- or mid-dose) compared to 6/10 male and female
rats in the control treatment groups (Lieder et al.. 2009a). The Lieder et al. (2009b)
two-generation reproduction oral gavage study did identify increased incidences of
33
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
hepatocellular hypertrophy in parental and F1 adult male rats at > 300 mg/kg-day; however, this
effect was absent in female rats at doses of K+PFBS up to 1,000 mg/kg-day. NTP (2018) and
NTP (2011) identified significantly increased incidence of hepatocellular hypertrophy in male (>
125 mg/kg-day) and female (> 500 mg/kg-day) S-D rats after 28 days of K+PFBS exposure.
Further, significantly increased cytoplasmic alteration of hepatocytes was observed in these rats
(male and female at > 500 mg/kg-day). Hepatic necrosis was also observed but was not
significant compared to control and only occurred at the high dose (1,000 mg/kg-day) in both
sexes (NTP. 2018. 2011).
In general, serum biomarkers associated with altered liver function or injury, including alanine
aminotransferase (ALT) and aspartate aminotransferase (AST), were not significantly changed in
male and female S-D rats across multiple oral gavage studies of varying exposure durations up to
90 days, at K+PFBS doses up to 1,000 mg/kg-day (Liedcr cl a 1 . 2""9a; 3M. 2001. 2000d). NTP
(2018) and NTP (2011). however, reported increased serum ,VI T and AST in male
(500 mg/kg-day only) and female (> 250 mg/kg-day for ALT; 5<)0 mg kg-day for AST) rats
exposed to K+PFBS for 28 days. Click to see interactive graphic for I .i\ cr F.ffects in HAWC.
4.6 Lipids and Lipoproteins
4.6.1 Human Studies
One low-confidence study (Zeng cl a I . 2" I 5) used l lie controls liom the case-control study of
asthma described below (Dong et a I . 2"13a) and examined llic association between PFBS
exposure and serum lipids. There was a statistically significant increase in total cholesterol
(P = 19.3 mg/dL increase per 1 uu I increase in Pl-'liS. ^5"o Cl n 6-38.0) but when PFBS
exposure was analy/cd in <.|uartilcs. no exposure-response gradient was observed.
4.6.2 Animal Stu
PFBS studies ha\e not particularly focused on perturbations in lipids or lipoproteins as a
potential health outcome, as studies ha\c typically focused only on measures of serum
cholesterol and triglyceride as part of a broader panel of clinical chemistry measures in high- or
medium-confidence rat studies of in. 2S, and 90 days (see Figure E-ll) (3M (2000d)"|; 3M
(2001) I. and l.icderclal (2<)o^i)l. respectively). Circulating levels of cholesterol and
triglycerides were unchanged in male and female S-D rats following daily oral gavage exposure
toK+PFBS for l<> days at doses up 1,000 mg/kg-day (3M. 2000d). In a similarly designed study
from the same laboratory, serum cholesterol and triglyceride levels were decreased in male rats
but at the high dose only. and. this effect was not statistically significant compared to control nor
was this effect observed in female rats of the same dose group (3M. 2001). Following exposure
for up to 90 days, cholesterol and triglycerides were unchanged in male and female rats at doses
up to 600 mg/kg-day (Lieder et al.. 2009a). PFBS was included in a multi-PFAS study
specifically designed to interrogate the mechanism of effect on lipid and lipoprotein metabolism
in a transgenic mouse line (APOE*3-Leiden CETP) that is highly responsive to fat and
cholesterol intake, consistent with human populations exposed to a western-type diet (Bijland et
al.. 2011; 3M. 2010). Adult male mice were fed a western-type, high-fat diet for 4 weeks prior to
initiation of PFBS exposure and throughout the 4-6 weeks PFBS exposure period (at
approximately 30 mg/kg-day). This study included several measures of lipid and lipoprotein
synthesis, modification, and transport or clearance such as circulating plasma levels, in vivo
clearance of very low density lipoprotein (VLDL)-like particles, fecal bile acid and sterol
34
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
excretion, hepatic lipid levels, lipase activity, VLDL-triglyceride and VLDL-apoB production,
and gene expression profiles. After 4 weeks of PFBS exposure, fasting plasma triglycerides,
cholesteryl ester transfer protein, and glycerol were significantly decreased compared to mice on
the control diet. Further, the half-life of VLDL-like particles and hepatic lipase activity, and
hepatic cholesteryl ester and free cholesterol levels were decreased (Biiland et al.. 2011; 3M
2010). Hepatic uptake of VLDL-like particles (represents fatty acid/lipid transport into hepatic
tissue) was modestly, but significantly increased compared to control mice. This increased
hepatic lipid uptake in the liver was accompanied by increased expression of genes associated
with lipid binding, activation, and metabolism (e.g., P-oxidation).
4.7 Other Effects
4.7.1 Human Studies
Two studies in China examined different immune outcomes in children (Chen et aL 2018; Dong
et aL 2013a).
One medium-confidence study reported in five pub I i cations (Pin et aL 2" I 7. Zhou et aL 2017b;
Zhou et al.. 2017a; Zhu et aL. 2016; Dong et a I . 2') I 3 b) examined the association between PFBS
exposure and asthma, asthma symptoms, pulmonary function, and related immune markers (IgE,
absolute eosinophil count [AEC], eosinophilic cation ic protein 11-CP], T-helper cell-specific
cytokines, and 16-kDa club cell secretory protein). The primary linding was a statistically
significant positive association between incident asthma (i.e . diagnosis in the previous year) and
PFBS exposure (OR [95% CT] for Q2 I 3 |<) 7. 2 3J. Q3 I 2 |n 7. 2.2], Q4: 1.9 [1.1, 3.4]). There
were also increases in AEC and ECP with increased exposure (not statistically significant with
the exception of AEC in children with asthma). There was no clear association with IgE or
T-helper cell-specific cytokines. There was also no clear association with asthma severity or
control of asthma symptoms (Dong et al.. 2013a). or pulmonary function measured with
spirometry among children u ith asthma (Qin et al . 2017). While pulmonary function could be
considered an outcome separate from asthma, the authors noted no associations in pulmonary
function (i e . in nonasthmatics across the PFAS they studied), so for these purposes, it was
considered an indicator of asthma sc\crily
One medium-confidence stuck (Chen el al.. 2018) examined the association between PFBS
exposure and atopic dermatitis and reported a nonstatistically significant increase in atopic
dermatitis with increased exposure (OR: 1.23, 95% CI: 0.74-2.04).
4.7.2 AnimalStu ..
Other effects were evaluated following exposure to PFBS, including outcomes related to the
spleen, hematological system, BW, neurotoxicity, and nonspecific clinical chemistry. These
groups of outcomes were not synthesized due to inadequate available information, uncertain
biological relevance, and/or inconsistencies across studies and sexes.
4.8 Other Data
Other studies that used PFBS or K+PFBS are described in this section. These studies are not
adequate for the determination of RfD values and were considered supportive data. These data
might include acute duration exposures, genotoxicity, mechanistic, and other studies
(see Table 5).
35
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Table 5. Other studies
Test
Materials and methods
Results
Conclusions
References
Genotoxicity
Mutagenicity
test
Salmonella typhimurium (strains TA98 and TA100) and
Escherichia coli (E. coli) (strain pKM 101) in the
presence or absence of S9. Concentrations of PFBS were
between 0-5,000 (ig/plate.
Test was iieuali\e for T \ 100 andpKMlOl
strains and et|in\ ocal for 1 \98 strain.
There is no in vitro
evidence of PFBS
mutagenicity.
NTP (2005)
Ames
S. typhimurium (strains TA98, TA100, TA1535, and
TA1537) and E. coli (strain WP2uvrA) were tested in
the presence or absence of S9 and with or without a
preincubation treatment. Concentrations ofK+PFBS
were between 0-5,000 (ig/plate.
1 lie results ofbolhmutaiioii assays
indicate lhal PFBS did not induce any
smuilicanl increase in the number of
re\ ertaiit colonics for any of the tester
strains in the presence or absence of
induced ral li\ er S')
There is no in vitro
evidence of PFBS
mutagenicity.
Pant (2001)
Genotoxicity
test
Human hepatoma (HepG2) cells were treated \\ nil
0.4 |iM to 2 mM PFBS. Intracellular ROS production
was measured by use of 2',7'-dichlorofluoresccui
diacetate and DNA damage was measured wil h ilie
comet assay.
The aniouui of k( )S and DNA strand
breaks remained unaffected by PFBS
ireaimeui
PFBS did not generate
ROS or DNA damage in
human liver cells.
Eriksen et al.
(2010)
CHO
chromosomal
aberration
Cultures of CHO cells were I rented u illi k H I !S al
concentrations ranging from <> In 5.<><)<) uu nil. wiili or
without exogenous metabolic actuation The in viim
exposure duration was 3 In-
1*1 T.Sdid uoi induce a statistically
simiilicaul increase in the percentage of
cells u nil aberrations at any of the
concent nit ions tested, either with or
u iilioul metabolic activation, in either
assa> u lien compared to the solvent
controls
Based on the negative
results in the in vitro CA
assay in CHO cells, PFBS
is not considered to be a
clastogenic agent.
Xu (2001)
Micronucleus
assay
Male and female S-D rats < 5 uroupi u ere e\posed l\\ ice
daily to K+PF'I !S In oral ga\ aue al doses of ' 1 \ <>2 5.
125, or 250 mg/ku lor 28 d
I'M >S did not induce a statistically
significant increase in the frequency of
micronucleated polychromatic
erythrocytes.
PFBS was negative for
micronuclei in the blood
of male and female rats,
indicating a lack of
genotoxic potential.
NTP (2012)
Acute duration and other routes of exposure
Acute
10 rats/group, young adult male nil (sirain not
specified), administered PFBS by ga\aue. single dose,
50, 100, 300, 600, or 800 |iL/kg and observed for 14-d
postexposure.
Mortality: 0%, 20%, 60%, 80%, and 100%
at 50, 100, 300, 600, and 800 ^iL/kg PFBS,
respectively.
Acute oral PFBS rat LD50
in male rats is 236 (iL/kg
(corresponding to
430 mg/kg).
Bomhard and
Loser ri 9961)
Low
confidence
36
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Test
Materials and methods
Results
Conclusions
References
Acute dermal
Adult (8 wk of age) male and female S-D rats (5/group)
were exposed dermally (10% of body surface area) to
500, 1,000, or 2,000 mg/kg K+PFBS for 24 hr and then
observed for 15-d postexposure for signs of clinical
toxicity, mortality, BW changes, or gross pathology
(terminus of study).
No treatment-related observations were
noted.
PFBS is not acutely toxic
via the dermal route of
exposure in rats.
3M f2000b,l
Dermal irritation
Adult (14-wk of age) female NZW rabbits (3 rabbits
total for study) were exposed dermally (6 cm2 of skin) to
500 mg K+PFBS for approximately 4 hr and then
observed for 9-d postexposure for signs of clinical
toxicity, mortality, or BW changes.
Drai/e seonuu was performed on the patch
site immediate!} follow 11m the exposure
period and 24, 48, and 72 hr postexposure.
No siuus of dermal irritation w ere
ohser\ ed. No signs of clinical lo.\ieit> or
mortality occurred \o treatment-related
alterations in P.W were noted.
PFBS did not induce
erythema, edema, or other
possible dermal findings
during the scoring periods,
indicating a lack of dermal
irritant properties in
rabbits.
3M C2000aN)
Ocular
sensitivity
Adult (16-wk of age) female NZW rabbits (3 rahhits
total for study) were exposed to approximate^ x<> mu
K+PFBS via ocular installation in the left eye for 2 sec
Eyes were flushed with 0.9% saline after 24 hr and I lien
observed and scored for up to 21 -d postexposure I he
rabbits were also followed fur clinical simis of to\icil>
or mortality/moribundity.
Hvcessive laennialioii of the left eyes
noted throughout stud> postexposure.
IJased 011 the lahoiators scoring system,
I'll >S w as "moderalcK " irritating at 24
and "2 hr postexposure
PFBS is a moderate ocular
irritant in rabbits.
3M f2000c)
Contact
hypersensitivity
Adult male (10-12 wkold) and female O w k old)
CRL:(HA)BR Hartley guinea pi us w ere injected
intradermals w ilh slerile water. I ieiiiid's ad|ii\aui. or
adjuvantcontaiuum 125 iiiu ml. FCPFBS i induction
phase). Day 7 alter induction, a petrolatum pasie
containing 0.5 g K PFBS was applied In llie pre\ ions
injection site of the uuineapigs for 4X hr (topical
induction phase). I)a> 22, a challenge dose of U.5 g
K+PFBS (petrolatum p;isiei was applied lo llie skived left
cranial flank (right flanks w ere treated w 11I1 petrolatum
paste only) (challenge phase). This challenge procedure
was repeated on Day 29. Challenge sites w ere observed
and scored following each challenge period (days 24-25
males and females and days 31 ^2 males only). Guinea
pigs were also followed for signs of clinical toxicity,
mortality/moribundity, or alterations in BW.
\o mortalities, clinical signs of toxicity, or
ehauues in BW associated with PFBS
exposure. Dermal scores were zero (no
response) in females and did not exceed 1
111 males (discreet or patchy edema), which
was not considered significant compared
to control guinea pigs exposed to Freund's
adjuvant alone.
PFBS is not considered an
allergen in the guinea pig
maximization test.
3M (2002a)
CA = chromosomal aberration; CHO = Chinese hamster ovary; cm2 = square centimeters; d = day(s); DNA = deoxyribonucleic acid; LD50 = median lethal dose; |xg/plate =
microgram per plate; |xM = micromol; mM = millimol; NZW = New Zealand White; ROS = reactive oxygen species; wk = weeks(s).
37
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
4.8.1 Tests Evaluating Genotoxicity and Mutagenicity
Genotoxic, mutagenic, and clastogenic effects of PFBS have been tested in mammalian and
prokaryotic cells in vitro (Eriksen et al.. 2010; NTP. 2005; Pant. 2001; Xu. 20011 and in rats in
vivo (NTP. 2018). PFBS was negative for mutagenicity in Escherichia coli (E. coli) strain
pKMlOl and Salmonella typhimurium strain TA100 (NTP. 2005). Mutagenicity test results were
equivocal in S. typhimurium strain TA98. Pant (2001) tested PFBS at concentrations up to
5,000 [j,g/plate in E. coli strain WP2uvrA and S. typhimurium strains TA98, TA100, TA1535,
and TA1537 in the presence or absence of exogenous metabolic activation and found no
evidence of mutagenic activity. In mammalian cells in vitro, PFBS did not generate reactive
oxygen species (ROS) or oxidative deoxyribonucleic acid damage in HepG2 cells (Eriksen et al..
2010). PFBS also failed to induce chromosomal aberrations in Chinese hamster ovary cells,
suggesting a lack of clastogenic activity (Xu. 2001). Adult male and female S-D rats exposed
twice daily to oral PFBS at doses up to 250 mg/kg for 2K days did not experience any significant
increases in micronucleated polychromatic erythrocytes, indicating a lack of genotoxic activity
(see Table 5) (NTP. 2012).
38
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
5.0 Evidence Integration and Hazard Characterization
The epidemiology database of studies of PFBS exposure and health effects consists of seven
epidemiologic studies (described in 10 publications), illustrated and summarized in the previous
section. The experimental animal database of all repeated-dose oral toxicity studies for PFBS and
the related compound K+PFBS includes a short-term range finding study in rats (3M. 2000d).
two 28-day studies in rats (NTP. 2018. 2011; 3M. 20011 one subchronic-duration study in rats
(Lieder et al.. 2009a). one subchronic-duration lipoprotein metabolism study in mice (Bijland et
al.. 2011; 3M. 2010). three gestational exposure studies in mice and rats (Feng et al.. 2017; York.
2003a. 2002). and one two-generation reproductive toxicity study in rats (Lieder et al.. 2009b).
Health outcomes evaluated across available studies included effects on the thyroid, reproductive
organs and tissues, developing offspring, kidneys, liver, and lipids/lipoproteins following oral
exposure to PFBS. Table 6 provides an overview of this database of potentially relevant studies
and effects. This table includes only the high- and medium-confidence animal studies (a single,
low-confidence animal study was not considered informative to draw ing judgments on potential
health hazard[s]); the available epidemiology studies are not included as their ability to inform
conclusions about associations was limited due to the small number of studies (typically one) per
outcome and poor sensitivity resulting from low exposure le\ els.
Following the summary of the available database in TaMe (\ narrative summaries describe the
evidence integration judgments and the primary rationales supporting these decisions for each
health effect. These narratives are supported l\\ an evidence profile table that succinctly lays out
the various factors that were judged to i ncrcase or decrease the support for hazard. While the
epidemiology studies were not influential to draw inu e\ idence integration judgments (i.e., they
were judged as equi\ ocal for all outcomes) or the deri\ ation of toxicity values (i.e., these studies
are not discussed in the next section), the general findings are summarized below to provide
context to the animal study findings and identify potential areas of future research.
39
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Table 6. Summary of noncancer data for oral exposure to PFBS (CASRN 375-73-5) and the related compound
K+PFBS (CASRN 29420-49-3)
Exposure
duration3
Reference
Study
confidence
Number of
male/female, strain,
species, study type,
study duration
Doses tested
(mg/kg-d)
HITecls observed at LOAEL
NOAEL
(mg/kg-d)
LOAEL
(mg/kg-d)
Short-term
3M (2000d)
Medium
confidence
5/5, S-D rat, K+PFBS
administered by
gavage, 10 d
0, |0(I. 1(1(1.
1,0(1(1
Increased ahsolnie and relative liver weight.
300
1,000
Short-term
3M(2001)
High
confidence
10/10, S-D rat,
K+PFBS administered
by gavage, 28 d
0, Km. iiiu,
900
Increased liver wemlil (nla 1 e i and kidney
ueiuhl (female).
300
900
Short-term
NTP (2018s);
NIP (2011)
High
confidence
10/10, S-D rat, PI liS
administered b>
gavage, twice/d. 2X d
0, 62.6, 125.
250, 500,
l,000h
1 )ecreased free T4, total T4 in males and
females Increased relative liver weight in
Icinalcs. and increased relative right kidney
weight in males.
ND
62.6
Subchronic
Lieder et al.
(2009a): York
(2003b)
High
confidence
In In. S-l) ral.
k PI US ;idnniiisieied
In ua\aue. ~ d \xk.
•JiUl
(I. (.(1. 2(1(1.
(,00
Increased incidence of renal hyperplasia in
males and females.
200
600
Subchronic
Biiland et al.
Medium
(> So. \poe:;:i-l.eiden
( 1: I P mice, k PI liS
mi diel. 4 (> w k
0. id
\lierations in lipid homeostasis (e.g., decreased
hepatic lipase, triglycerides) is of uncertain
biological significance.
ND
ND
("2011): 3M
("2010)
confidence
Developmental
Fene et al.
(2017)
High
confidence
ii In. ICR mice,
k PI I3S ;idiiiiiiisieied
In ua\aue. (il)s 1 2(>
(i. 5(1. 2(i(i.
5( i( i
1 )ecreased T3, free T4, and total T4 and
increased TSH in maternal and offspring (PND
30 only). Delayed eyes opening, vaginal
opening, and final estrous and decreased BW in
pups.
50
200
Developmental
York (2003a)
His>h
confidence
II'8, S-l) ral. k PFBS
adiiiiinsicied hy
(il)s (i—20
0, 100, 300,
1,000, 2,000
Decreased maternal feed consumption, BW
gain, and gravid uterine weight. Decreased pup
BW at doses where maternal health was
affected limiting the interpretation of the
results; thus developmental effect levels were
not determined. (Limited endpoints
evaluated—pilot study).
P0: 1,000
Fl: ND
P0: 2,000
Fl: ND
40
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Exposure
duration3
Reference
Study
confidence
Number of
male/female, strain,
species, study type,
study duration
Doses tested
(mg/kg-d)
I'.ffecls observed at LOAEL
NOAEL
(mg/kg-d)
LOAEL
(mg/kg-d)
Developmental
York ("20021
High
confidence
0/25, S-D rat, K+PFBS
administered by
gavage, GDs 6-20
0, 100, 300,
1,000
Decreased malemal feed consumption and BW
uani 1 kvi'cased pup BW at doses where
maternal heallh was affected limiting the
iiilcrpielalimi of I he results; thus developmental
cl'l'ecl le\els were not determined.
P0: 300
Fl: ND
P0: 1,000
Fl: ND
Reproductive
Lieder et al.
(2009b1; York
("2003d; York
(2003d1; York
(2003e1
High
confidence
30/30, S-D rat,
K+PFBS administered
by gavage,
two-generation
reproductive stud>
POadulis 0.
30, 100, 300,
1,000
F1 adults: 0.
30. I III). 3(1(1.
1.
P0 and F1 adults increased incidence of
hyperplasia and local papillary edema in the
kidneys of males and females
F2 pups no dose-related elfecls at the highest
dose lesied (1,000 mg/kg-d).
P0, Fl: 100
F2: 1,000
P0, Fl:
300
F2: ND
Notes'. ND = no data; ICR = Institute of Cancer Research
a Duration categories are defined as follows: Acute = exposure for < 24 hours; short term = repeated exposure lor 24 hours to < 30 days; long term (subchronic) = repeated
exposure for > 30 days < 10% lifespan for humans (> 30 days up to approximately 00 days in typically used laboratory animal species); chronic = repeated exposure for >
10% lifespan for humans (>~ 90 days to 2 years in typically used laboratory animal species) ( U.S. K1V\. 2002).
bRats were gavaged twice daily at administered doses of 0. 31.3. 62.6, 125, 250. and 500 mg/kg in N IP ("20181 andNTP ("20111.
41
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
5.1 Thyroid Effects
PFBS-induced perturbation of the thyroid was consistently observed across two species, sexes,
life stages, and exposure durations in two independent, high-confidence studies. These
perturbations involved a coherent pattern of hormonal changes. Significant changes in tissue
weight or histopathology were not observed.
Similar patterns of decreases in total T3. total T4. and free T4 were observed in PFBS-exposed
pregnant mice, nonpregnant adult female and adult male rats from a 28-day study, and
gestationally exposed female mouse offspring (NTP. 2018: Feng et al.. 2017; NTP. 2011). These
decreases were of a concerning magnitude (-20% in dams and 5<)% in offspring), they were
shown to persist at least 60 days after gestational exposure in offspring, and they exhibited a
clear dose-dependence in both studies.
Development of numerous organ systems, including neuronal, reproductive, hepatic, and
immune systems, are affected by altered thyroid homeostasis since adequate levels of thyroid
hormones are necessary for normal growth and de\ elopment in early life stages (Forhead and
Fowden. 2014; Gilbert and Zoeller. 2010; Hulherl. 2000). Thus, the observed effects of PFBS
exposure on thyroid hormone economy are biologically consistent with the reported delays and
abnormalities in organ/system development discussed below It is well-established that the
presence of sufficient thyroid hormones during the gestational and neonatal period is essential
for brain development and maturation. Studies specifically e\aluating the effect of PFBS on
neurodevelopment were not identified, leaving uncertainty as to the potential for more sensitive
developmental effects Nonetheless, the coherence of these Pl-'liS lindings, in addition to the
large number of xenohiotic exposure studies demonstrating associations between thyroid
hormone economy and decrements in early life stage growth, development, and survival,
provides support for thyroid hazard.
Taken together, the e\ idence in animals for thyroid effects supports a hazard. No studies in
humans were a\ailaMe Although there are some differences in hypothalamic-pituitary-thyroid
(HPT) regulation across species (e.g., serum hormone-binding proteins, hormone turnover rates,
and timing of/'// utero thyroid de\ elopment), rodents are generally considered to be a good
model for e\ aluating the potential for thyroid effects of chemicals in humans (Zoeller et al..
2007). Overall, based on findings in animal models considered to be informative for evaluating
the potential for thyroid effects in humans, the available evidence supports a hazard and the
thyroid is considered a potential target organ for PFBS toxicity in humans.
5.2 Developmi
Overt effects on birth parameters and early development have generally not been observed in
either rats or mice after PFBS exposure. Specifically, the available studies do not provide
evidence of effects on endpoints relating to pregnancy loss, fetal survival, or fetal alterations
(Feng et al.. 2017; Lieder et al.. 2009a; York. 2003a. c, 2002). While one mouse study indicated
pronounced decreases in female offspring BW at several ages after gestational exposure (Feng et
al.. 2017). several other studies either did not observe decreases in offspring BW or only detected
these changes when parental BWs were similarly affected (Feng et al.. 2017; Lieder et al.. 2009a;
York. 2003a. c, 2002).
42
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Delays in development have been reported following gestational PFBS exposure in mice,
including delayed development of the female reproductive organs (i.e., ovaries, uterus, and
vaginal patency), delayed and abnormal estrous cycling (i.e., first estrous and prolongation of
diestrus), and delayed eye opening (Feng et al.. 2017). Age at vaginal patency and ovarian
follicle counts (i.e., in F1 rat offspring after delivery of the F2 generation) were unaffected at
1,000 mg/kg-day in a two-generation reproductive toxicity study (Lieder et al.. 2009a). This
observed lack of effects (i.e., on vaginal patency) is inconsistent with the findings in mice and
not easily explained by the current understanding of PFBS toxicokinetics. However, Feng et al.
(2017) also noted changes in reproductive hormones that might be relevant to the delays in
female sexual development, including a decrease in serum estradiol and increased luteinizing
hormone in pubertal offspring (i.e., PND 30 [Note: progesterone was decreased at a later age,
PND 60, but not PND 30], As the changes reported in mice by Feng et al. (2017) were observed
in parallel with effects on thyroid hormone levels (discussed alxn e), it is plausible that these
developmental delays and hormonal changes could represent sequalae of reduced thyroid
function, although that was not directly tested
For the most part, developmental effects have been reported in a single study and species
(mouse); however, the findings are coherent with one another as well as ith the consequences
of decreased thyroid hormone levels. Due to the coherence across effects on the thyroid and
several interrelated developmental effects in mice (i e . delays and hormonal changes), the
evidence in animals for developmental effects supports a hazard There is no reason to expect
that the specific developmental delays obser\ ed in mice would not be directly relevant to similar
processes in humans. Thus, based on findings in animals that are presumed to be relevant to
humans, the available e\ idence supports a hazard and the de\ eloping offspring is considered a
potential target for PFBS toxicity in humans As no studies in humans were available, this
represents an area dcscr\ inu of additional research
5.3 R
Reproducti\e outcomes, including male and female fertility, pregnancy outcomes, hormone
levels, markers of reproductix e de\ elopmeni, and reproductive organ weights and
histopatholouy. have been e\aluated in a number of high-confidence studies in mice (Feng et al..
2017) and rats (\TP. 2018. 2"| |; Lieder et al.. 2009a; Lieder et al.. 2009b). In addition, three
low-confidence human studies e\ aluated potential associations between PFBS exposure and
reproductive effects (Song et al . 2018; Zhou et al.. 2017a; Zhou et al.. 2016).
PFBS exposure has resulted in no significant changes in male mating and fertility parameters,
reproductive organ weights, or reproductive hormones. While there were some slight,
statistically significant effects on male reproductive endpoints in two rat studies (specifically,
altered sperm parameters such as percentage of abnormal sperm or testicular sperm count (NTP.
2018; Lieder et al.. 2009a) and delayed preputial separation at 1,000 mg/kg-day (Lieder et al..
2009a)). these findings were not clearly dose-dependent and the levels of change were of
questionable biological significance. No significant reproductive effects in men were noted
across two human studies (Song et al.. 2018; Zhou et al.. 2016).
In general, PFBS exposure in adults has also resulted in no significant alterations in female fertility
or pregnancy outcomes in rats or mice (NTP. 2018; Feng et al.. 2017; NTP. 2011; Lieder et al..
2009a; Lieder et al.. 2009b) or in two human studies (Zhou et al.. 2017a; Zhou et al.. 2016). and
43
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
inconsistent changes in rodent reproductive organ weights were reported across studies
regardless of duration and timing of exposure. However, changes in normal estrous cyclicity,
specifically prolongation of the diestrus stage, have been reported in both nonpregnant adult rats
exposed to PFBS (NTP. 2018) and adult mouse offspring exposed gestationally from GDs 1 to
20 (Feng et al.. 2017). PFBS exposures in NTP (2018) began between 8 and 10 weeks of age;
although the exposures might overlap with some aspects of reproductive development or changes
in function during adolescence, these rats were sexually mature and thus the endpoints are
considered in the context of reproductive, rather than developmental, effects. The mouse
offspring in the study by Feng et al. (2017) also displayed delayed vaginal patency and
histopathological markers of decreased fertility (i.e., decreased follicles and corpora lutea);
however, the reproductive function of those offspring was not tested. While adult rat offspring
(Fl) in a two-generation toxicity study also exhibited variable changes in estrous cyclicity
(Lieder et al.. 2009b). including prolonged diestrus at 1<><> mu ku-day, this effect was not
observed at higher doses, limiting interpretation, and no effects on \ auinal patency were
observed. Female reproductive hormones can inform the potential for effects on reproductive
organ development, estrous cyclicity, and fertility Changes in serum hormones included
increased testosterone after exposure of female rats as adults (NTP. 2018). increased luteinizing
hormone and decreased estradiol in pubertal mice after gestational exposure (I'eng et al.. 2017).
and decreased estradiol and progesterone when these gestationally exposed mice were assessed
as adults. Overall, the pattern and timing of hormonal changes after PFBS exposure is difficult to
interpret and likely incomplete. However, the hormonal alterations after gestational PFBS
exposure in mice are most relevant to conclusions about female reproductive health.
Taken together, the evidence indicates that the developing reproductive system, particularly in
females, might be a target for PFBS toxicity. However, the potential for reproductive effects in
adults was less clear, and significant impacts on mating or fertility parameters were not observed
across the available studies. Therefore, the evidence in developing animals is considered most
informati\ e to conclusions relating to potential developmental effects (see above) and the
evidence lor reproductive effects (i.e., in adults) is equivocal. In the three studies of potential
reprodncti\ e effects in humans, no clear associations were observed, and so the evidence in
human studies is equivocal ()\ erall. leased on equivocal human and animal evidence, the
available e\ idence for reproducti\e effects is equivocal.
5.4 Renal _
Renal effects associated u ilh oral exposure to PFBS have been observed in adult or developing
rats across high- or medium-confidence gavage studies of various duration (NTP. 2018. 2011;
Lieder et al.. 2009a: Lieder el al . 2009b: 3M. 2001. 2000d).
Statistically significant increases in kidney weights have been observed in male and female rats
after short-term exposure in one study (NTP. 2018). with strong dose-dependence for changes in
relative weights in female rats at doses as low as 62.5 mg/kg-day. This study was likewise the
only study to observe changes in serum markers of renal injury, specifically increased BUN in
males at > 250 mg/kg-day. However, while several other studies noted slight increases in weights,
typically at higher PFBS doses (> 500 mg/kg-day), these nonsignificant changes were not
consistently observed across the set of available studies and no other studies reported changes in
serum markers of renal injury (Lieder et al.. 2009a: Lieder et al.. 2009b: 3M. 2001. 2000d).
44
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Kidney histopathology for some effects (i.e., CPN, hydronephrosis, tubular degeneration, and
tubular dilation) was unaffected by PFBS exposure in rats, although each of these endpoints was
not assessed across several studies (NTP. 2018. 2011; Lieder et al.. 2009a; 3M. 2000d). Mixed
results were reported for mineralization and necrosis. Both of these endpoints were noted in
females, but not males, after subchronic exposure to 600 mg/kg-day (Lieder et al.. 2009a).
whereas mineralization was unaffected in male or female rats after short-term exposure
(3M. 2000d) and necrosis was unaffected in male or female rats in short-term and 2-generation
(in both generations) studies (NTP. 2018; Lieder et al.. 2009b). Multiple markers of
inflammatory changes were consistently noted in the two longest exposure duration studies,
which were the only studies to report on these endpoints. Specifically, increases in chronic
pyelonephritis, tubular basophilia, and mononuclear cell infiltration were observed in female, but
not male, rats following subchronic exposure to 600 mg/kg-dav (l ieder et al.. 2009a). Similarly,
increases in papillary edema and hyperplasia were obser\ ed in male and female rats after
subchronic exposure to 600 mg/kg-day (Lieder el a I . 2"09a). and in both generations of rats in
the two-generation study at > 300 mg/kg-day (Lieder el al.. 2009b). u ilh female rats being more
sensitive than males.
Overall, the evidence in animals suggests an increased sensili\ ily of female nils (i.e., based on
histopathology and organ weight changes). Due primarily lo llie consistency and coherence in
renal effects observed in the subchronic-duration stuck by l.ieder et al. (2009a) and the
reproductive toxicity study by Lieder et al (2009b) in male and female rats, the evidence in
animals supports a hazard. No information liom PFBS studies informs the human relevance of
these findings. Taken together, the renal histopathology e\ idence in rodents identifies a
toxicologically significant spectrum of effects thai is presumed lo be relevant to similar changes
known to occur in humans Renal effects (i e . uric acid) were e\aluated in one low-confidence
human study and no clear association was observed. and so the evidence in human studies is
equivocal. Overall, based on findings in animals that are presumed to be relevant to humans, the
available evidence supports a hazard and indicates the kidney as a target organ of PFBS toxicity.
5 5 ?"•* £* Vf 13"""" ; - ~ ~ i"*
Hepalic effects, including organ-weight changes and histopathology associated with oral
exposures lo PLUS. have been observed in high- or medium-confidence studies in adult or
developing rats following short-term and subchronic durations (NTP. 2018. 2011; Lieder et al..
2009a; 3M. 2001. 2O00d) and in a two-generation reproductive study in rats (Lieder et al..
2009b). Increased absolute and or relative liver weights were consistently observed in male and
female rats after short-term and multigenerational exposure (NTP. 2018. 2011; Lieder et al..
2009b; 3M. 2001. 2000d) In some studies, the magnitude of the liver weight changes and the
doses at which effects occurred differed across sexes of rat, although the pattern across studies
was unclear and did not consistently indicate one sex as more sensitive. Liver histopathology,
including necrosis and inflammation, was not consistently observed across PFBS studies. One
possible exception is increases in hepatocellular hypertrophy in male rats observed across two
studies (NTP. 2018; Lieder et al.. 2009b). although female rats were unaffected in the
multigenerational study and this lesion was not observed at up to 600 mg/kg-day in the
subchronic study by Lieder et al. (2009a). The only study to observe changes in serum markers
of liver injury was NTP (2018). at > 250 mg/kg-day in females and > 500 mg/kg-day in males.
The biological relevance or significance of the observed liver effects is not clear. In particular,
45
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
the adversity of the variable changes in liver weight and observations of cellular hypertrophy is
unclear. Further, the observed lesions either occurred in only one sex of rat, were not
dose-dependent compared to control, and/or occurred only at the highest PFBS dose tested. Thus,
the evidence in animals is equivocal. Overall, based on equivocal animal evidence and a lack of
human studies, the available evidence for hepatic effects is equivocal.
5.6 Effects on Lipid or Lipoprotein Homeostasis
Few studies have examined the effects of PFBS on circulating or hepatic lipid or lipoprotein
homeostasis. It is recognized that increased circulating levels of lipids and lipoprotein products
and/or increased hepatic lipid load are clinical observations of concern in humans. However, the
lack of effect on lipid dynamics in most studies of rats exposed lo high oral K+PFBS doses for up
to 90 days and the generally modest effects in transgenic mice, designed to interrogate
mechanisms of lipid transport and metabolism, fed a high-lal. w estcrn-type diet renders this
potential health outcome of unclear toxicologica I significance at this lime. Thus, given the
inconsistent, modest effects and the unclear biological relevance of 1 hose changes in isolation
(i.e., lipids/lipoproteins were decreased, not increased) the evidence in animals is equivocal.
Effects on serum lipids were evaluated in one low -confidence human study and. although an
association was observed between increased PFBS exposure and increased lolal cholesterol, this
evidence in humans is equivocal. 0\ erall. based on equivocal e\ idence in both animal and
human studies, the available evidence for effects on lipid or lipoprotein homeostasis is equivocal.
5.7 Immune Effects
Immune effects were obsei \ed in two human studies, including associations with asthma
(Dong et al.. 2013a) and atopic dermatitis (Chen el al . 2"1S), Because of the lack of additional
evidence and some concerns about risk of bias, the evidence in human studies is equivocal.
Overall, based on equivocal evidence in human studies and a lack of animal studies, the available
evidence for immune effects is equivocal
5.8 : ¦ ' '' - rterization Summary
Based on the e\ idence integration judgments regarding the potential for PFBS exposure to cause
health effects (the narrali\c abo\e is summarized in Table 7), the animal studies informing the
potential effects of PFBS exposure on thyroid function, renal function, and development were
concluded to support hazard. Thus, for the purposes of this assessment, the animal data
supporting these outcomes were considered for use in dose-response analysis, and other data
were considered no further
46
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Table 7. Summary of hazard characterization and evidence integration judgments
Studies and confidence
Factors that increase support
for hazard
Factors that decrease
support for hazard
Summary of findings
Overall
evidence
integration
judgment and
basis
Thyroid effects
Human studies
Supports a
hazard
(animal
evidence
supports a
hazard; human
evidence is
equivocal).
The primary
basis for this
judgment is
thyroid
hormone
decreases in
mice and rats
at>
62.6 mg/kg-d.
No studies available to
evaluate
--
--
--
Animal studies (all oral savase)
Mouse Studies:
• High-confidence
gestational (GDs 1-20)
exposure studv (Fens et
al.. 2017)
Rat Studies:
• High-confidence
short-term (28-d)
toxicity studv (NTP.
2018. 2011)
• Consistent thyroid hormone
decreases (i.e., for total T3,
total T4, and free T4) across
two high-confidence studies
of varied design. The
findings were consistent
across two species. se\es. life
stages, and e\pnsui'c
durations.
• Dose-response unidicnts
were obser\ ed lor those
lh\ mid hormones
• l.arue mauiiiludes of effect
to u . up lo 5<>"reductions
hi offspring serum hormone^
\\ ere reported for those
thyroid humilities.
• \o factors noted
Similar ratterns of decreases in thvroid hormones
(i e . for total T3. total T4. and free T4) were
observed in PFBS-exposed pregnant mice and
gestational!} exposed female mouse offspring at
200 ma/ka-d (Fens et al.. 2017) and in adult
female and male rats at > 62.6 mg/ke-d (NTP.
:u|8. 201 n.
Increased TSH was reported in mouse dams and
in pubertal (PND 30) offspring following
gestational exposure (Fene et al.. 2017s). but no
changes were noted in rats exposed as adults
(NTP. 2011).
Thvroid wcieht and historatholoev were not
changed after short-term exposure in adult male or
female rats (NTP. 2018. 2011).
Developmental effects
Human studies
No studies available to
evaluate
--
--
--
47
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Studies and confidence
Factors that increase support
for hazard
Factors that decrease
support for hazard
Summary of findings
Overall
evidence
integration
judgment and
basis
Animal studies (all oral gavage)
Supports a
hazard
(animal
evidence
supports a
hazard', human
evidence is
equivocal).
The primary
basis for this
judgment is a
set of persistent
developmental
delays and
alterations in
reproductive
system
maturation in
female mice,
generally at >
200 mg/kg-d.
Mouse Studies:
• High-confidence
gestational (GDs 1-20)
exposure studv (Feng et
al.. 2017s)
Rat Studies:
• Two high-confidence
gestational exposure
(GDs 6-20) studies: a
range finding study and
a follow-up studv (York.
2003c. 2002)
• High-confidence
2-generation study
(Licdcr et al.. 2009b)
• Biologically consistent
spectrum of developmental
effects in female offspring in
a high-confidence mouse
study at doses not causing
maternal toxicity, including
pronounced and persistent
effects on BW, delays in
developmental milestones
and sexual maturation,
concordant effects on
reproductive organs, and
altered serum hormones
• Concerning magnitude of
effect (e.g., - 25"., change in
pup weight) and
dose-dependence I'm' se\ oral
parameter
• Coherence of effects \x illi
lh\ mid hormone
iiisiil'licieiics abo\ei
Note lliese effects were also
coherent w illi effects on est i ons
cyclicilN obser\ ed after
short-term o\posiiie itiaduli nils
(NTP. 2018. 201 11. hilt llns was
categorized as a reproduce e
effect (see below i
• Developmental effects
were limited to changes in
one study, sex. and
species.
• A high-confidcncc rat
study reported some
inconsistent evidence,
including lack of a dela> in
\aginal patency and lack of
clear effects on est runs
cyclicily or ovarian
morphology, although the
latter endpoint was
assessed in much older
animals. These potential
differences across species
are not explainable based
on lo\icokiiielics alone.
In the onlv mouse studv (Fens et al.. 2017).
de\ elopmental effects and altered markers of
female reproductive development or function
were observed in female offspring after
gestational PFBS exposure, including decreased
n\V. dclavcd eve opening. delaved vaginal
opening. altered cslrous cvclicitv (including
prolonged diestnis). altered reproductive
hormones (e.g.. decreased estradiol and
progesterone), and effects on reproductive organs
(e.g.. weight and ovarian morphology). Most
effects were observed al > 200 mg/kg-d, with
several changes noted al PND 60.
Endpoints relating to fcrlilitv. pregnancv.
survival, and fetal alterations were unchanged in
both rats and mice across the four available
studies, although this was not tested in mouse
olisdring (Feng et al.. 2017).
Developmental BW changes in rat offspring were
either unchanged (Lieder et al.. 2009b) or
observed only at doses causing parental toxicity
(York. 2003c. 2002).
In a rat two-generation study, while some
statistically significant findings were noted for
markers of female reproductive development or
function, thev were not dose-dependent or were of
questionable biological relevance; thus, no clear
changes in F1 offspring were noted at doses up to
1.000 mg/kg-d regarding vaginal patencv or
estrous cvcling at comparable ages to (Feng et al..
2017). or in ovarian morphologv after the F1
females gave birth to the F2 pups.
48
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Studies and confidence
Factors that increase support
for hazard
Factors that decrease
support for hazard
Summary of findings
Overall
evidence
integration
judgment and
basis
Reproductive effects
Human studies
Equivocal
Male reproductive effects
{equivocal
human and
animal
evidence).
Note: As the
strongest
evidence for
female
• Low-confidence cohort
studv (Zhou et al.. 2016)
• Low-confidence
cross-sectional study
(Sone et al.. 2018)
• No factors noted.
• Lack of clear association in
studies of low confidence
with poor sciisiiiMly (i e.
due to low exposure le\ els.
raime)
No clear associalion between PFBS exposure and
male reproducli\ e hormones (Zhou et al.. 2016) or
semen parameters (Sonu et al.. 2018s).
Female reproductive effects
• Low-confidence
cross-sectional study
(Zhou et al.. 2017a)
• Low-confidence cohort
studv (Zhou et al.. 2016)
• No factors noted.
• 1 .ack of clear association in
siudies of low confidence
with poor seusiiiN n\
(i.e., due io low exposure
levels, raimei
• Potential lor re\ erse
causalionfor meiisirual
cscle characlerisiics
No clear association between PFBS exposure and
female reproductive hormones (Zhou et al.. 2016)
or menstrual evele characteristics (Sons et al..
20 18).
reproductive
effects was in
offspring that
were
gestationally
exposed, these
findings were
considered
most relevant
Animal studies fall oral %avave)
to
developmental,
not
Male reproductive effects
Rat Studies:
• High-confidence
short-term (28-d)
toxicitv studv (NTP.
2018. 2011s)
• High-confidence
2-generation study
(Lieder et al.. 2009b)
• High-confidence
subchronic studv (Lieder
et al.. 2009a)
• No faclors noted
• \ few small, statistically
significant changes were
not dose-dependent or
were of questionable
biological relevance.
• Lack of effects on male
mating and fertility,
hormones, or reproductive
organs in rats.
Statistically significant effects on sperm health
(NTP. 2018; Lieder et al.. 2009a) and delaved
DrcDutial separation at 1.000 me/ke-d (Lieder et
al.. 2009b) were not dose-dependent, were within
the normal range of historical controls for the
laboratory, and/or were no longer significantly
changed after correcting for other variables
(e.g., BW).
Other relevant parameters (e.e.. orean weiehts.
mating success, and so forth) were unchanged in
the three studies.
reproductive,
effects.
49
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Studies and confidence
Factors that increase support
for hazard
Factors that decrease
support for hazard
Summary of findings
Overall
evidence
integration
judgment and
basis
Female reproductive effects
Mouse Studies:
• High-confidence
gestational (GDs 1-20)
exposure studv (Fens et
al.. 2017s)
Rat Studies:
• High-confidence
short-term (28-d)
toxicity studv (NTP.
2018. 2011)
• High-confidence
subchronic studv (Licdcr
et al.. 2009a)
• High-confidence
2-generation study
(Licdcr et al.. 2009b)
• Effects on markers of female
reproductive function
(i.e., estrous cyclicity) were
observed in high-confidence
studies in rats and mice.
• Changes in reproductive
serum hormones were
observed in female rats
(i.e., increased testosterone)
and mice (e.g., decreased
estradiol and progesterone).
Although the pattern of
change is difficult in interpret
and likely incomplete. 1 licrc
were no conJliclinu dala
• Lack of similar effects dm
reproductive function
(i.e., estrous c>clicil>) in a
second high-cmifidence rat
study.
• Lack of effects on female
fertility or pregnancy
measures, although this
was untested in prenatalK
exposed female mouse
offspring.
• Lack of oruau uemlil
changes in ihree nil
studies.
Note 1 lie lack of effects on
o\ariau follicles in rals did
not decrease llie support for
lia/ard pro\ ided h\ findings
mi mice, as ihe aue al endpoint
assessment was ik 25<> nm/ku-d (NTP. 2018. 2011).
female reproductive oraan weishts were reduced
mi uestationallv exposed mouse offsnrine (Fens et
al . 2017s). but were unchanged after short-term,
snhchronic. or 2-generational exposure (NTP.
2<11S. 2<11 1: Lieder et al.. 2009a: Lieder et al..
:<)<>%).
50
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Studies and confidence
Factors that increase support
for hazard
Factors that decrease
support for hazard
Summary of findings
Overall
evidence
integration
judgment and
basis
Renal effects
Human studies
Supports a
hazard.
(animal
evidence
supports a
hazard', human
evidence is
equivocal).
The primary
basis for this
judgment is
kidney
histopathology
in rats,
primarily
females, at
>300 mg/kg-d.
• Low-confidence
cross-sectional study
(Oinetal.. 2016s)
• No factors noted.
• Inconsistency across
subpopulations in single
study.
• Single stud> of low
confidence w illi concern
for potential re\ crsc
causality.
Overall, there was no clear association for PFBS
and uric acid No association observed between
PFBS and uric acid in the total population.
Increase in uric acid with increased exposure in
bov s. but decrease for mi ls (neither was
statistically significant i
Animal studies fall oral savage)
Rat Studies:
• One high-confidence
subchronic studv (Licdcr
et al.. 2009a)
• Two high-confidence
studv (NIP. 2018.2011:
3M. 2001s) and one
medium-confidence
(3M. 2000d) short-term
(10-28 d) study
• One high-confidence
2-generation study
(Licdcr et al.. 2009b)
• Two high-confidence studies
with the longest exposure
durations reported consistent
effects on kidney
histopathology in male and
female rats (females were
more sensitive).
• The histopathologic^! effects
related In iiillammaliou were
larucK dose-dependent and
of a concerning niauuitudc.
allliouah primarily al liiuli
doses i iiio or 600 mg ku-d)
• Inconsisteucv in kidnev
weiglil chanues across
studies
• Findings are from a snide
laboratory and species.
Note 1 lie general lack of
effects mi oilier pathology
cndpoinis in the slimier term
studies was not considcredto
decrease support lor hazard,
as ilns was not interpreted as
inconsistent
Increases in kidnev weisht in male and female
rats were observed in one short-term study at >
62 5 mg/kg-d, but clear changes were not
observed in the other short-term, subchronic, or
iw o-gcncration rat studies.
Kidnev historatholoev for some effects
(i.e., CPN, hydronephrosis, tubular degeneration,
and tubular dilation) was unchanged in
single-study evaluations, and mixed results across
studies were reported for mineralization and
necrosis TNTP. 2018: Lieder et al.. 2009a: Lieder
et al.. 2009b: 3M. 2000d). Multiple markers
potentiallv related to inflammation and most
notably papillary edema and hyperplasia were
increased in the two longest duration studies
(Lieder et al.. 2009a: Lieder et al.. 2009b).
without contrary evidence.
Other markers of renal iniurv. including BUN and
creatinine, were mostly unaffected across studies
(TSTTP. 2018. 2011: Lieder et al.. 2009a: Lieder et
al.. 2009b: 3M. 2001. 2000d). althoushthe NIP
study did observe effects on BUN in males at >
250 mg/kg-d.
51
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Studies and confidence
Factors that increase support
for hazard
Factors that decrease
support for hazard
Summary of findings
Overall
evidence
integration
judgment and
basis
Hepatic effects
Human studies
Equivocal
{equivocal
human and
animal
evidence).
No studies available to
evaluate
--
--
--
Animal studies fall oral gavage)
Rat Studies:
• One high-confidence
subchronic studv (Licdcr
et al.. 2009a)
• Two high-confidence
studv (NTP. 2018.2011;
3M. 2001s) and one
medium-confidence
(3M. 2000d) short-term
(10-28 d) study
• One high-confidence
2-generation study
(Licdcr et al.. 2009b)
• Consistent changes in liver
weights in rats of both sexes
across four studies. Although
the pattern (e.g., by sex and
dose) and magnitude of
changes varied across
studies, weights were
consistently increased
• Other than li\er-\\emlit
changes, llicrc were
notable unexplained
inconsistencies in the
findings ncross studies
• One high-confidence study
was ciilircK iiianisisieiil
\hsolute or rclali\ e liver weiehts were increased
mi all studies except the ''0-d exposure component
of l lie studv bv Licdcr el al. (2009a). which tested
doses up to 600 mg/kg-d.
Note. 70 d of exposure in this study did elicit
effects.
Effects generally occurred at > 300 mg/kg-d,
although one study reported effects at lower doses
i YI'P. 2011; 3M. 2001). and two others (3M.
2u(il. 2000d) observed chanees at > 900 me/ke-d.
Serum markers of liver iniurv were unchanged in
three studies (Lieder et al.. 2009a: 3M. 2001.
2000d) and increased in one short-term studv at >
250 me/ke-d (NTP. 2018. 2011).
Liver histopatholoev. specificallv hepatocellular
hypertrophy and cytoplasmic alterations in males
and females CNTP. 2018. 2011) or hvpertrophv in
females onlv (Lieder et al.. 2009a). were noted in
two studies, but not in the others.
Lipid or lipoprotein homeostasis
Human studies
Equivocal
{equivocal
human and
animal
evidence).
• Low-confidence
cross-sectional study
fZene et al.. 2015)
• No factors noted
• Association in a single
study with concern for
potential reverse causality.
Increase in total cholesterol (statistically
significant, (3 = 19.3 mg/DL increase per unit
increase in PFBS).
Animal studies
52
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Studies and confidence
Factors that increase support
for hazard
Factors that decrease
support for hazard
Summary of findings
Overall
evidence
integration
judgment and
basis
Mouse Studies (dief):
• Medium-confidence
short-term (4-6 wk)
studv (Biiland et al..
2011); transgenic mice
(human-like lipid
metabolism) were fed a
high-fat diet
Rat Studies (all oral
savage):
• Decreases in serum
cholesterol and triglycerides
were observed in male rats
and mice.
• Inconsistent evidence in
other rat studies and across
sexes.
• Small effect magnitudes
and unclear direction
(decreases) of changes are
of questionable biological
relevance and could uoi he
informed by e\ahialum
dosc-dependciicN
(i.e.. only single-dose or
high-dose effects were
observed).
Serum lipids, specificallv cholesterol and
triglyceride levels, were slightly decreased
(~2
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Studies and confidence
Factors that increase support
for hazard
Factors that decrease
support for hazard
Summary of findings
Overall
evidence
integration
judgment and
basis
Atopic dermatitis
• Medium-confidence
cohort studv (Chen et
al.. 2018s)
• No factors noted.
• Slight associations u ere
not statistically significant
in a single s|iid\ u illi
concern reua al iiiu I lie
potential for residual
confounding fe u . w illi
otherPFAS chemicals)
Nmisialislically significant increase in odds of
atopic dermatitis (OR = 1.2) with increased PFBS
exposure
Animal studies
No studies available to
evaluate
-
-
-
Notes:
a The lack of liver effects in the subchronic study w as not interpreted to significantly reduce support for hazard, as the maximum tolerated dose was 600 mg/kg-d, and other studies
reported only liver effects at > 900 mg/kg-d.
T3 = triiodothyronine; T4 = thyroxine.
54
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
6.0 Derivation of Values
The hazard and dose-response database for PFBS and the potassium salt is primarily associated
with the oral route of exposure. There are a limited number of dermal studies (see Table 5) and
no known inhalation studies. There are no known studies evaluating potential cancer effects of
PFBS. As such, only noncancer reference values are derived in this assessment for the oral route.
6.1 Derivation of Oral Reference Doses
6.1.1 Derivation of Candidate Subchronic RfDs
As described above, the hazards of potential concern for oral PI-US exposure include thyroid,
developmental, and kidney effects. Thus, available studies thai e\ aluated these effects are
considered in the derivation of oral RfDs, and selected data sets from studies with multiple
exposure levels for thyroid, developmental, and kidney effects were modeled using the EPA's
Benchmark Dose Software (BMDS) Version 2.7.
Consistent with the EPA's Benchmark Dose Icchnica/ Guidance Document (U.S. EPA. 2012).
the BMD and 95% lower confidence limit on the HMD (BMDL) were estimated using a
benchmark response (BMR) to represent a minimal. biologically significant le\el of change.
Based on BMD guidance, in the absence of information regarding the level of change that is
considered biologically significant, a BMR of 1 SD from the control mean for continuous data or
a BMR of 10% extra risk for dichotomons data is used to estimate the BMD and BMDL, and to
facilitate a consistent basis of comparison across endpoints. studies, and assessments.
For thyroid hormone ellects in pregnant females and offspring, a 2<>% RD was applied based on a
biological level of concern Multiple lines of evidence regarding the degree of thyroid hormone
disruption and developmental outcomes in pregnant dams or offspring were considered in the
identification of this HMR During de\ elopmental life stages such as gestational/fetal and
postnatal early newborn, thyroid hormones are critical in a myriad of physiological processes
associated w ith somatic growth and maturation and survival mechanisms such as thermogenesis,
pulmonary gas exchange, and cardiac development (Sferruzzi-Perri et al.. 2013; Hillman et al..
2012). Further, thyroid hormones are critically important in early neurodevelopment as they directly
influence neurogenesis, synaptogenesis, and myelination (Puig-Domingo and Vila. 2013; Stenzel
andHuttner. 2013; Patel et al.. 2"! I). Indeed, human epidemiological studies have demonstrated
key relationships between decreased levels of thyroid hormones such as T4 and in utero and
early postnatal life neurode\ elopmental status. For example, children born euthyroid but who
were exposed to thyroid hormone insufficiency in utero, present with cognitive impairments
(e.g., decreased intelligence quotient) and/or concomitant abnormalities in brain imaging
(Korevaar et al.. 2016; Lavado-Autric et al.. 2003; Mirabella et al.. 2000). With regard to what level
of decrease in thyroid hormone is sufficient for anatomical and/or functional alterations, particularly
in neurodevelopment in developing fetuses or newborns, several studies have identified a fairly
stable range across humans and experimental rodents. Neurodevelopmental and cognitive deficits
have been observed in children who experienced a 25% decrease in maternal T4 during the second
trimester in utero (Haddow et al.. 1999). In other studies, mild-to-moderate thyroid insufficiency in
pregnant women was defined as having serum T4 levels below the 10th percentile for the study
population, which was associated with a 15%—30% decrease relative to the corresponding median
(Finken et al.. 2013; Julvez et al.. 2013; Roman et al.. 2013; Henrichs et al.. 2010). Similarly,
55
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
decreases in mean maternal T4 levels of ~ 10%— 17% during pregnancy and lactation have been
found to elicit neurodevelopmental toxicity in rat offspring (Gilbert et al.. 2016; Gilbert. 2011).
As the lower end of the range of T4 changes associated with untoward developmental health
outcomes (e.g., 10%) commonly falls within normal experiment-to-experiment variation in
control values, a BMR of 20% RD from control mean was determined to be a minimally
biologically significant degree of change when performing BMD modeling on thyroid hormone
alterations in pregnant females and associated offspring. For comparison purposes only, a BMR
of 1 SD is presented as well. While significantly decreased thyroid hormone (e.g., T4 and T3)
was also observed in adult rats exposed twice daily to oral K+PFBS (NTP. 20181 a biologically
significant level of change was not determined for the BMR as il is unclear what magnitude of
hormone perturbation would be considered adverse in adult animals. As such, for thyroid
hormone effects in nonpregnant adult rats, a BMR of I SI) iVoni control mean was applied.
For effects in the developing offspring, other than the thyroid, a liMR of 1 SD change from the
control mean is used for continuous data to account lor effects occurring in a sensitive life stage.
For comparison purposes only, a 0.5 SD is presented as well.
For kidney hyperplasia data from the subchronic-duration study by Lieder el a I ( 2009a) and two-
generation reproductive toxicity study by Liederclal (2<")OJh). a BMR of I < >".. c\ tra risk was
used because it is the recommended approach for dichotomous data in the absence of information
on the minimally significant level of change
Where modeling was feasible, the estimated liMI)l.s were identified as PODs (summarized in
Table 8). Further details, including the modeling output and graphical results for the model selected
for each endpoint, can he found in IIAWC and are discussed in Appendix: F. Where dose-response
modeling was not feasible. \()AI-I.s or LOAN ,s were identified (summarized in Table 8).
In Recommended I Isc <>J Hotly Weight4 as the Default Method in Derivation of the Oral Reference
Dose (U S EPA. 2C)| lh). the I-PA endorses a hierarchy of approaches to derive human equivalent
oral exposures from data from laboratory animal species, with the preferred approach being
physiologically bused loxicokinetic modeling. Other approaches might include using some
chemical-specific information, without a complete physiologically based toxicokinetic model. In the
absence of chemical-specific models or data to inform the derivation of human equivalent oral
exposures, the I-PA endorses BWU as a default to extrapolate toxicologically equivalent doses of
orally administered agents from all laboratory animals to humans for the purpose of deriving anRfD
under certain exposure conditions. More specifically, the use of BW3 4 scaling for deriving an RfD is
recommended when the observed effects are associated with the parent compound or a stable
metabolite, but not for portal-of-entry effects. Although the pharmacokinetic study by 01 sen et al.
(2009) suggests a longer half-life for K+PFBS in humans than in rats, these results were obtained by
single-dose administration and it is uncertain whether this reflects the compound's half-life after
repeated dosing. The EPA considered the 2014 Guidance for Applying Quantitative Data to Develop
Data-Derived Extrapolation Factors for Interspecies and Intraspecies Extrapolation in determining
interspecies and intraspecies UFs (UFas and UFhs, respectively) (U.S. EPA. 2014c). Using the
decision process described in Figure 2 of that guidance (U.S. EPA 2014c). the EPA concluded that
data are inadequate to support derivation of data-derived extrapolation factors. Specifically, given the
lack of available models and data to address external dose and clearance in humans with any
certainty or the magnitude of difference in half-life across species as a function of dose or time, the
default approach of the use of BW3'4 scaling to obtain a HED is considered appropriate in this case.
56
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Table 8. PODs considered for the derivation of the subchronic RfD for K+PFBS (CASRN 29420-49-3)
Endpoint/reference
Species/life stage—
sex
PODiiid"
(mg/kg-d)
Comments
Thyroid effectsf
Total T4—Fens et al. C2017)1,
Mouse/Po—F emale
BMDL20— > ^
(BMDLisd = ' 4)
\dequale model fit
Free T4—Fens et al. (2017V3
Mouse/Po—F emale
no\i:i. "5
No models pro\ ided adequate statistical or visual fit to mean responses
TSH—Fens et al. Q017)h
Mouse/Po—F emale
NO\i:i. 7 5
No models pro\ ided adequate statistical or visual fit to mean responses
Total T4 PND 1 (fetal n f—Fens et al.
(2017)
Mouse/Fi —F emale
NO\i:i. "5
No models provided adequate fit to the data, specifically variance
Total T4 PND 1 (litter nf—Fens et al.
(2017s)
Mouse/Fi —F emal e
PATDL: 4 2
(i:\ll)Ll; X<>)
\dequaie model fit
Total T4 PND 30—Fens et al. (2017*)
Mouse/Fi —F ema 1 e
I:MI)L21)= 7.8
(UMULi; "in
\dequaie model fit
Total T4 PND 60—Fens et al. (2017)
Mouse/Fi —F emale
NOAEL = 7.5
No models provided adequate fit to the data, specifically variance
TSH PND 30—Fens et al. (2017)
Mouse 1' —Female
NO \l 1. "5
No models provided adequate statistical or visual fit to mean responses
Total T4—NTP (20IS)'1
k
il—Male
LOAM. 15 5
No models provided adequate statistical or visual fit to mean responses
k
il—1 emale
BMDI. 1 6
Adequate model fit
Free T4—NTP (2018)d
k
il—Male
1.0\ni. 15 5
No models provided adequate statistical or visual fit to mean responses
k
il—1 emale
i.o\i:i. 14 '
No models provided adequate statistical or visual fit to mean responses
Developmental effects
Eves openins (fetal nf—Fens el a 1 < 2< 11")
Mouse I ' —I'euiale
NOAEL = 7.5
No models provided adequate fit to the data, specifically variance
Eves ODcnins (litter n )c—Fens el a I 117)
Mouse/Fi—I'euiale
BMDL1sd= 14.8
(BMDL0.5SD = 7.4)
Adequate model fit
Vasinal ODcnins (fetal nf—Fens el al
(2017)
Mouse 1' —I'euiale
BMDLisd = 12.4
(BMDLo.5sd = 5.1)
Adequate model fit
Vasinal openins (litter nf—Fens et al.
(2017)
Mouse 1' —l'eui;ile
BMDLisd = 7.9
(BMDLo.5sd = 3.3)
Adequate model fit
First estrous (fetal nf—Fens et al. (2017)
Mouse/Fi —F emale
NOAEL = 7.5
No models provided adequate statistical or visual fit to mean responses
57
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Endpoint/reference
Species/life stage—
sex
PODiiid"
(mg/kg-d)
Comments
First estrous (litter n)°—Fene et al. (20171
Mouse/Fi —F emale
NOAEL = 7.5
No models provided adequate statistical or visual fit to mean responses
Kidney effects
Kidney histopathology—papillary
epithelial tubular/ductal
hyperplasia—Lieder et al. (2009a1e
Rat—Male
BMDLio = 4" ii
\da|uaic model fit
Rat—Female
BMDLio — 12 (>
\da|iiak' model fit
Kidney histopathology—papillary
epithelial tubular/ductal
hyperplasia—Lieder et al. (2009b1f
Rat/Po—Male
BMDLk :<>u
Adequate model fit
Rat/Po—Female
BMDL 1 1 5
Adequate model fit
Kidney histopathology—papillary
epithelial tubular/ductal
hyperplasia—Lieder et al. (2009b1f
Rat/Fi—Male
BMDL,. 4: 8
Adequate model lit
Rat/Fi—Female
BMDLk 2"2
\dequate model fit
Notes:
BMDLio = 10% benchmark dose lower confidence limit; BMDLisd = benchmark dose lower confidence limit for 1 SD change from the control.
a Following U.S. EPA("2011b1 guidance, animal doses from candidate principal studies were converted to HEDs through the application of a dosimetric adjustment factor (PAF),
where HED = dose x DAF. DAFs for each dose are calculated as follows: DAF = (BWa1/4 BWh1'4), where BWa = animal BW and BWh = human BW. For all DAF calculations, a
reference BWh of 80 kg ("U.S. EPA. 2011a-) was used. The default BWh assumption was updated from 70 kg to 80 kg based on NHANES data from 1999 to 2006 and represents the
mean weight for adults ages 21 and older ("U.S. EPA. 2011a-). The BWh also includes estimates of BW of pregnant women reported in the EPA's Exposure Factors Handbook
("U.S. EPA. 2011a-).
b DAFs were calculated using dam terminal BWs < I >\Y. i as reported by the study authors for ICR mice. For example, dams in the 200 mg/kg-day group had a terminal BW of
0.0399 kg. The DAF was calculated DAF = (BWa1 1 ¦+¦ BWh1, ') = (0.0399l/1 801/4) = 0.1 49. The FEED was calculated as dose x DAF = 200 x 0.149 = 29.9 mg/kg-day.
c Fetal endpoints from Feng et al. ('2017') were modeled alternatively using dose group sizes based either on total number of fetuses or dams. Given that it appears that
Feng et al. ("20171 did not use the litter as the statistical unit of analysis, it is unclear if the study-reported standard errors pertain to litters or fetuses. Alternatively, modeling fetal
endpoints using litter n or fetal n provides two modeling results that bracket the "true" variance among all fetuses in a dose group. Individual animal data were requested from
study authors but were unable to be obtained.
dDAFs were calculated using study-specific time-weighted average BW 11 >\Y. i data lorfemale and male S-D rats using study-specific weekly BW data.
e DAFs were calculated using study-specific time-w eighted average 13 Wi, data for female and male S-D rats using study-specific initial and terminal BW data. For example, female
rats in the 600 mg/kg-day group had a time-weighted average BW of 2 11.35 g. The DAF was calculated as follows: DAF = (BWa1/4 BWh1'4) = (0.211351'4 801/4) = 0.2344. The
FEED was calculated as dose x DAF = 600 x 0.2344 = 140.65.
fFor the Po-male rats (S-D), DAFs were calculated using BWs (BWa) estimated from study-specific initial and terminal BWs. For the Po-female rats, DAFs were calculated using
BWs (BWa) estimated from study-specific initial, day 70, TD 1, and LD 22 BWs. For the Fi-male rats, DAFs were calculated using BWs (BWa) estimated from study-specific
postweaning day 1 and terminal BWs. For the Fi-female rats, DAFs were calculated using BWs (BWa) estimated from study-specific postweaning day 1, day of cohabitation,
TD 1, and LD 22 BWs.
i BMD modeling methods and links to modeling inputs and results in IIA WC are found in appendix F.
FIAWC visualization: Candidate POPs for Subchronic and Chronic RIP
58
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
As presented in Table 8, effects in the thyroid were considered when determining potential PODs
for derivation of a subchronic RfD. Similar patterns of decreases in total T3. total T4. and
free T4 were observed in PFBS-exposed pregnant mice, nonpregnant adult female rats, adult
male rats, and gestationally exposed female mouse offspring (NTP. 2018; Feng et al.. 2017).
Reflex increases in TSH in response to decreased T4 or T3 were not observed in male or female
rats following 28 days of exposure. Such an increase in TSH was observed in pregnant mice and
their corresponding female offspring, at PND 30 only, with an irregular dose-response or time
course. This pattern of decreased thyroid hormone without a concomitant increase in TSH is
consistent with a human clinical status referred to as "hypothyroxinemia." Importantly, it has
been noted that milder forms of thyroid perturbation are up to 10 times more prevalent in human
populations than overt gestational hypothyroidism (Korevaar et al.. 2016; Stagnaro-Green et aL
2011). Hypothyroxinemia has been associated with impairments in neurodevelopment and/or
cognition later in life (Thompson et al.. 2018; Min et al.. 2') ) As discussed in the opening of
the "Derivation of a Subchronic RfD" section above (i.e., identification of a biologically
informed BMR for thyroid hormone effects), se\ era I studies have identified a fairly stable range
across humans and experimental rodents at which decrements in thyroid hormone levels are
associated with developmental health outcomes. This hypothyroxinemia. rather than overt
hypothyroidism, is further supported by the lack of effect on thyroid weight or tissue architecture
in rats after 28 days of PFBS exposure, which might he expected in a condition that lacks a
coordinated feedback activation of TSH (NTP. 2018). While both the Feng et al. (2017) and
NTP (2018) studies are considered high confidence, it should he noted that the NTP thyroid data
is nonpeer-reviewed and has not yet been published in a completed report. However, NTP
studies are conducted with the highest level of standards and laboratory practices and employ
study designs highly rele\anl for the systematic evaluation of chemical toxicities.
Developmental effects, as presented in Table 8, were considered in the determination of the
potential PODs for deri\ ation of a subchronic RfD. Specifically, in Feng et al. (2017).
developmental delays or abnormalities in growth (i.e., BW and eye opening), reproductive
organs (i e . ovaries, uterus, and \auinal opening), and reproductive cycling (i.e., first estrous and
prolongation of diestrns) were ohser\ed in mouse offspring. These effects were observed in mice
from litters in which thyroid hormone deficiency occurred at PND 1 and was sustained through
pubertal and adult periods (i e . PND 3d and PND 60, respectively).
As presented in Table 8, effects in the kidney were considered when determining the potential
PODs for derivation of a suhchronic RfD. Mild-to-moderate hyperplasia was reported in the
kidneys of male and female rats following subchronic-duration exposure to PFBS by Lieder et al.
(2009a) and in the P0- and I' I -generation animals of the reproductive toxicity study by Lieder et
al. (2009b). Other studies evaluating effects in the kidney were of shorter duration and thus less
suitable as a candidate principal study. Additional histopathological alterations accompanied the
hyperplasia observed in the kidney, including papillary edema and inflammatory changes,
specifically increases in chronic pyelonephritis, tubular basophilia, and mononuclear cell
infiltration (Lieder et al.. 2009a; Lieder et al.. 2009b). Across kidney histopathological effects
reported following PFBS exposure, in general, female rats were more sensitive than males.
Although histopathological effects in the kidney have not been reported in human populations
exposed to PFBS, the spectrum of renal effects observed in rodents is presumed to be relevant to
similar changes known to occur in humans.
59
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Across the body of evidence supporting hazards via the oral exposure route, candidate RfDs
were derived for the most sensitive effects in the thyroid and kidney. Across all life stages
evaluated, the thyroid (specifically, decreased thyroid hormone [T4]) was identified as the most
sensitive target of PFBS toxicity. It should be noted that developmental effects (e.g., delayed
eyes opening, vaginal opening, or first estrous) were often observed in mouse litters in which
decrements in thyroid hormone occurred (Feng et al.. 20171 However developmental effects
appeared to be less sensitive than thyroid hormone perturbations in developing mice, with PODs
ranging from 7.5-14.8 mg/kg-day compared to 4.2-8.9, respectively (see Table 8). Thus, the
thyroid and kidney effects were carried forward for derivation of candidate RfDs.
When deriving the candidate subchronic RfD based on thyroid effects, the Feng et al. (2017) and
NTP (2018) studies were both considered for potential principal study due to the observed
sensitivity of thyroid hormone decrements. However, the biological significance of
hypothyroxinemia (i.e., decreased T4) in nonpregnant adult animals is unclear; therefore, the
thyroid effects from the NTP (2018) study were not selected as a critical effect. Instead, the
gestational exposure study in mice was selected as the principal stuck for derivation of the
candidate subchronic RfD based on thyroid effects The gestational exposure study conducted by
Feng et al. (2017) reports administration of K PI liS In ga\ age in ICR mice (I <> dose) from
GDs 1 to 20. This study was of good quality (i.e . high confidence) with adequate reporting and
consideration for appropriate study design, methods, and conduct (click to see risk of bias
analysis inHAWC). Feng et al. (2017) reported statistically significantly decreased total T3, total
T4, and free T4, as well as increased TSII in dams and offspring (increased TSH PND 30 only)
gestationally exposed to PFBS.
The critical effect from the I'eng et al. (2<>17) study is decreased serum total thyroxine (T4) in
newborn (PND 1) mice (I'enu el a I . 2017). T4 and T3 are essential for normal growth of
developing offspring across animal species (for re\ iew see Forhead and Fowden (2014)). And,
previous studies have shown that exposure to other PFAS during pregnancy results in lower T4
and T3 le\ els in pregnant women and fetuses or neonates (Yang et al.. 2016; Wang et al.. 2014).
The selection of total T4 as the critical effect is Ixised on a number of key considerations (see
below ) that account for cross-species correlations in thyroid physiology and hormone dynamics
particularly within the context of a developmental life stage.
A key consideration for selection of total T4 is that this represents the aggregate of potential
thyroid endocrine signaling (i e . free T4 + protein bound T4) at any given time. Although T3 is the
more active hormone form in respondent somatic tissues, the formation of T3 is contingent upon
the deiodination of free T4 T4. not T3, is the thyroid hormone that crosses the placenta of humans
and rodents. Although free T4 might be considered a suitable measure of thyroid hormone status in
non-developmental (e.g., adult) life stages, there are some important factors associated with
maintenance of the microenvironment for a developing offspring in utero that lends credence to the
use of total T4 as the critical effect. A tightly regulated transfer of maternal thyroid hormone to a
fetus is paramount to proper development of multiple tissues and organ systems (e.g., nervous
system), especially during the early trimesters. The placenta has transporters and deiodinases that
collectively act as a gatekeeper to maintain an optimal T4 microenvironment in the fetal
compartment (Fisher. 1997; Koopdonk-Kool et al.. 1996). For example, deiodinase 3 (D3) is
highly expressed in human uterus, placenta, and amniotic membrane, where it serves a critical
role of regulating thyroid hormone transfer to the fetus through the deiodination of T4 to
60
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
transcriptionally inactive reverse triiodothyronine (rT3) or T3 to inactive 3,5-diiodo-L-thyronine
(T2). Similarly, Wasco et al. (2003) showed that D3 is highly expressed in rodent uterus and is
highly induced during pregnancy. Further, the Dio3 gene that encodes D3 has been shown to be
imprinted in the mouse (Hernandez et al.. 2002). suggesting a pivotal role for this specific
deiodinase in the mouse as well. Indeed, the human and rodent placenta have been shown to be
similarly permeable to T4 and T3 (Fisher. 1997; Calvo et al.. 1992). Due to placental barrier
functionality, free T4 levels in a pregnant dam might not be entirely representative of actual T4
status in a developing fetus. Thus, decreased total T4 in offspring is expected to be more
representative of PFBS-mediated thyroid effects and potentially associative developmental effects.
There is some difference in HPT development and functional maturation and regulation during
early life stages (e.g., timing of in utero and early postnatal thyroid development) between
humans and rodents. For example, the human thyroid gland is not formed until the second
trimester of pregnancy, whereas in rodents the fetal thyroid is not formed and functional until the
17th day of gestation, which is just days before birth (typically (il) 2") As such, rodent
neurodevelopment in the early postnatal phase is analogous to the third trimester of human
development in utero (Gilbert et al.. 2012). I lo\\e\ er. within the context of early developmental
life stages, there are several commonalities in HPT dynamics between humans and rodents such
as similar profiles of (1) thyroid hormone binding proteins. (2) hormone functional reserve, and
(3) placental deiodinase. For example. l\\ o carrier proteins thyroid binding globulin (TBG) and
transthyretin (TTR)—are primarily responsible for storage and transit of T4 in mammals. TBG is
the primary carrier of T4 in humans across all life stages (Sa\ u el al.. 1991). Importantly, in fetal
and infant rats, TBG is also the primary carrier of T4 (Saui el a I . 1989). As rats transition to
adulthood, TTR takes o\ er as the primary carrier of T4 In addition, as a relatively highly
abundant carrier protein, albumin also plays a role in thyroid hormone binding and transit in
humans and rodents: how e\ er, the relative affinity lor binding is lower than either TBG or TTR.
Life stage-specific differences in thyroid hormone reserve capacity between adults and neonates
have been noted. On a\ erage, iniralhyroidal thyroulobulin stores in adults are on the order of
months w hereas in neonates the functional reser\ e is approximated at less than 1 day (Gilbert
and Zocllcr. 2010; Savin et al.. 2003; Van Den Hove et al.. 1999). This suggests that the adult
thyroid has compensatory abilities not present in early life stages, making fetal/neonatal
populations particularly sensitive to perturbations in thyroid hormone economy. As human and
rodent neonates are posited to be more alike than different, the dynamic reserve capacity of T4
between species near birth and in early postpartum might not be significantly different. Human
neonates have a serum half-life of T4 of approximately 3 days (Vulsma et al.. 1989). and thyroid
tissue stores of T4 are estimated lo be less than 1 day (Van Den Hove et al.. 1999). As the
developing rodent thyroid does not begin producing its own hormone until late in gestation
(> GD 17), newborn rodent T4 levels are primarily a reflection of transplacental^ translocated
maternal hormone; and adult rats have been shown to have a serum T4 half-life of 0.5-1 day
(Choksi et al.. 2003). As such, significant differences in functional thyroid reserve capacity
between human and rodent neonates is not anticipated.
61
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Accounting for the information presented above, the candidate subchronic RfD for thyroid, based
on the BMDL20 (HED) of 4.2 mg/kg-day for serum total T4 in newborn (PND 1) mice, is derived
as follows:
Candidate Subchronic RfD for K+PFBS (Thyroid) = BMDL20 (HED) UFc
= 4.2 mg/kg-day -MOO
= 0.042 mg/kg-day
= 4 x 10"2 mg/kg-day
Table 9 summarizes the UFs for the candidate subchronic RfD for K+PFBS based on effects in
the thyroid.
Table 9. UFs for the candidate subchronic RfD for thyroid effects for K+PFBS
(CASRN 29420-49-3)
UF
Value
.liiMil'ication
UFa
3
A UFa of 3 (100 5) is applied to account fur uncertainty in characterizing llie loxicokinetic and
toxicodynamic differences between mice and humans following oral K PFBS/PFBS exposure. Some
aspects of the cross-species extrapolation of loMcokinctie and loxicodynamic processes have been
accounted for by calculating an HF,D by applvnm a l)\l' as mil lined in the EPA's Recommended Use
of Bodv Weisht3/4 as the Default Method in Derivation of the Oral Reference Dose (U.S. EPA.
201 lb): however, some residual unccrtainlv remains In I lie absence of chemical-specific data to
quantify this uncertainty, the EPA's guidance recommends use ofa UFA of 3.
UFd
3
A UFd of 3 is applied due to database deficiencies The oral exposure database contains multiple
short-term and sulvhroiiic-duralion toxicitv siudies of l;ihoi;iior\ animals (NTP. 2018; Biiland et al..
2011;NTP. 2<>| 1. "Al. 2010: Lieder et al.. 2<«>l>;i. "Al. 2t><>|. 2000d). a two-generation reproductive
toxicitv siud\ 111 nils il.icderel al . 2u(i%). ;md niulliDle dev eloDmental toxicitv studies in mice and
rats (Fenu el al . 2
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
papillary epithelial tubular/ductal hyperplasia in the female rat. The Lieder et al. (2009b) study
was peer-reviewed, applied established approaches, recommendations, and best practices, and
employed an appropriate exposure design for evaluating systemic toxicity endpoints
(Organisation for Economic Co-operation and Development [OECD] 416; the Office of
Prevention, Pesticides and Toxic Substances [OPPTS] 870.3800), and was rated high confidence.
Kidney effects observed, such as papillary epithelial tubular/ductal hyperplasia, could influence
the ability of the kidney to filter waste. Therefore, the kidney endpoint was identified as a
critical effect for derivation of a candidate subchronic RfD for K+PFBS based on the BMDLio
(HED) of 11.5 mg/kg-day for papillary epithelial tubular/ductal hyperplasia in female Po rats
(Lieder et al.. 2009b). The candidate subchronic RfD for kidney is derived as follows:
Candidate Subchronic RfD for K+PFBS (Kidney) BMDLio (HED) UFc
11.5 mg/kg-day -M00
11 12 mg/kg-day
I x 10 1 mg/kg-day
Table 10 summarizes the UFs for the candidate suMironic RfD for K Pl-'liS hased on effects in
the kidney.
Table 10. UFs for the candidate subchronic Ul'l) lor kidney effects for K+PFBS
(CASRN 29420-49-3)
UF
Value
.liisliricalioii
UFa
3
A UFa of 3 (10'°) is applied in accouui I'm- loMcokiuclic and to\icodynamic differences between rats
and humans following oral k 1*1 T.S HI iS exposure Some aspects of the cross-species extrapolation
of toxicokinclic and toxicodv iiamic processes ha\ e been accounted for by calculating an HED by
applying a DAF as outlined mi the 1 P Vs AV. miuuended Use of Body Weight3'4 as the Default Method
in Derivation of the Oral Reference Dose (U.S. EPA. 201 lb): however, some residual uncertainty
remains In the absence of chemical-specilic data to quantify this uncertainty, EPA's guidance
recommends use of a UF of '
UFd
3
\ I I ', , of 3 is applied due to database deficiencies. The oral exposure database contains multiple
short-term and subchronic-duration toxicitv studies of laboratory animals (NIP. 2018; Biiland et al..
2<>| 1: NTP. 2011: 3M. 2010: Lieder et al.. 2009a: 3M. 2001. 2000d). a two-generation reproductive
toxicity sludv in rats (Lieder et al.. 2009b). and mulliDlc developmental toxicitv studies in mice and
rats (Feuu el al.. 2017: York. 2002). However, the observation of decreased thvroid hormone is
know ii lo be a crucial clement during developmental life stages, particularly for neurodevelopment,
and t lie database is limited by the lack of developmental neurotoxicity studies. In addition, as
immunotoN icily is an effect of increasing concern across several members of the larger PFAS family,
the lack of studies evaluating this outcome following PFBS exposure is a limitation in the database.
UFh
10
A UFh of 10 is applied for interindividual variability in susceptibility in the absence of quantitative
information on the toxicokinetics and toxicodynamics of K+PFBS/PFBS in humans.
UFl
1
A UFl of 1 is applied for LOAEL-to-NOAEL extrapolation because the POD is a BMDL and the
BMR was selected based on evidence that it represented a minimal biologically significant response
level in adult rats.
UFs
1
A UFS of 1 is applied because the POD comes from a subchronic-duration study of rats.
UFC
100
Composite Uncertainty Factor = UFA x UFD x UFH x UFL x UFS
63
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Candidate subchronic RfDs derived for thyroid effects and kidney effects are presented in
Table 11.
Table 11. Summary of candidate noncancer subchronic reference values for K+PFBS
(CASRN 29420-49-3)
Critical effect
POD method
POD (HED)
(mg/kg-d)
UFc
Reference
dose
(mg/kg-d)
Decreased serum total T4 in newborn (PND 1)
mice—Fens et al. (2017).
BMDL20
4.2
100
4 x 10-2
Kidney histopathology—papillary epithelial
tubular/ductal hyperplasia in P0 female
rats—Lieder et al. (2009b)
BMDL10
1 1 5
100
1 X 10-1
The data for K+PFBS can be used to derive a subchronic RfD for the live acid (PFBS), as
K+PFBS is fully dissociated in water at the en\ iron mental pH range of 4-9 (N1CNAS. 2005Y To
calculate the subchronic RfD for the free acid, the suhchronic RfD for the potassium salt is
adjusted to compensate for differences in MW belw een k PI 'liS (338.19) and PFBS (300.10).
The subchronic RfD for PFBS (free acid) for thyroid and kidney effects are the same as the
values fortheK+PFBS salt. The calculations are as follows
Subchronic RfD Rll) for k Pl-'liS sail : (MW free acid MW salt)
for PFBS (free acid) <> <>42 mg kg-day (3'>0.10 338.19)
for Thyroid KITecls <> <>42 mg kg-day (0.89)
i> i>37 mu kg-day
4 x |0 2 mg/kg-day
Subchronic Ufl) Rll) lor k PFBS salt x (MW free acid -^MW salt)
for PI-"US (free acid) '>12 nig kg-day x (300.10 338.19)
for kidney I".fleets 0.12 mg/kg-day x (0.89)
0.11 mg/kg-day
1 x 10"1 mg/kg-day
The confidence in the candidate subchronic RfD for PFBS and K+PFBS for both thyroid and
kidney effects is medium, as explained in Tables 12 and 13.
64
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Table 12. Confidence descriptors for candidate subchronic RfD (Thyroid Effects) for PFBS
(CASRN 375-73-5) and the related compound K+PFBS (CASRN 29420-49-3)
Confidence categories
Designation
Discussion
Confidence in study
H
Confidence in the principal study is high because the overall
study design, performance, and characterization of exposure
was eood. Studv details and risk of bias analysis can be found
inHAWC.
Confidence in database
M
Confidence in the oral toxicity database for derivation of the
candidate subchronic RfD for thyroid effects is medium
because although there are multiple developmental toxicity
studies in mice and rats, no studies are available that have
specifically e\aluatcd ncurodevelopmental effects.
Confidence in candidate subchronic
RfD
M
The overall confidence in ilie candidate subchronic RfD for
thyroid effects is medium
Notes: H = high;M = medium
Table 13. Confidence descriptors for the candidate subchronic RfD (Kidney Effects) for
PFBS (CASRN 375-73-5) and the related compound K PFBS (CASRN 29420-49-3)
Confidence categories
Designation
Discussion
Confidence in study
II
( oulidcuce in ilie principal study is high because the overall
siikI\ desiuu. performance, and characterization of exposure
u as uood Studs delails and risk of bias analysis can be found
in 1 IA\Y('
Confidence in database
M
Confidence mi i lie oral loxicily database for derivation of the
candidale suhchronic RID for kidney effects is medium
because, although there arc multiple short-term studies and a
subclirouic-duralion toxicity study in laboratory animals, and
one acceplablc two-generation reproductive toxicity study in
rats, the database lacks studies that have specifically evaluated
ueurodcN elopmental effects.
Confidence in candidate suhclirouic
RfD
\1
1 lie overall confidence in the candidate subchronic RfD for
kidney effects is medium.
Notes'. H = hialiL M = medium.
The subchronic Rll) is deri\ cd lo he protective of all types of effects across studies and species
following oral suhchronic exposure and is intended to protect sensitive subpopulations and life
stages. In light of the consistent observation of the thyroid effects across life stages and the
greater dose-response sensili\ iiy, relative to the kidney, EPA is proposing to base the overall
subchronic RfD on the thyroid effects. See the Federal Register Notice announcing the
availability of the draft assessment for PFBS and requesting public review and comment on this
proposal in addition to the approaches and conclusions in the PFBS assessment.
6.1.2 Derivation of Candidate Chronic RfDs
There are no chronic-duration studies available for PFBS and K+PFBS. Therefore, based on the
same database and similar considerations for the subchronic RfD, candidate noncancer chronic
RfDs were derived for thyroid effects and kidney effects.
65
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
The candidate chronic RfD for thyroid effects following exposure to K+PFBS, based on the
BMDL20 (HED) of 4.2 mg/kg-day for serum total T4 in newborn (PND 1) mice (Feng et al..
20171 is derived as follows:
Candidate Chronic RfD for K+PFBS (Thyroid) = BMDL20 (HED) - UFc
= 4.2 mg/kg-day ^ 300
= 0.014 mg/kg-day
= 1 x 10"2 mg/kg-day
Table 14 summarizes the UFs for the candidate chronic RfD for K+PFBS based on effects in the
thyroid.
Table 14. UFs for the candidate chronic RfD for thyroid for K PFBS (CASRN 29420-49-3)
UF
Value
.liisl il'iciition
UFa
3
A UFa of 3 (100 5) is applied to account fur in\icnkinetic and toxicod> iiamic differences between
mice and humans following oral K+PFP.S I'll !S exposure. Some a spec Is of 1 lie cross-species
extrapolation of toxicokinetic and toxicnd> naiiiic processes have been accnuiiied for by calculating an
HED by applying a DAF as outlined in the EPA's Recommended Use of Body Weight3/4 as the
Default Method in Derivation of the Oral Reference Dose (U.S. EPA. 2011b): however, some residual
uncertainty remains. In the absence of chemical-speeil ie dala in quantify this uncertainty, the EPA's
guidance recommends use of a I I 'a of 3.
UFd
10
A UFd of 10 is applied to acaunii for dalahase deficiencies. 1 lie oral exposure database contains
multiple short-term and subchronic-duralioii loxicils siudies of laboratory animals (NTP. 2018:
Biiland el al.. 201.1: NTP. 2011: Licdcrctal.. 20<>lJa. "Al. 2(>u|. 2 et al.. 2017: York. 20<>21 1 low e\ or. as ih\ roid hormone is known to be critical
during developmental life stages, parlicularh fur ncurodcvclopmcnl. the database is limited by the
lack of developmental neurotoxicity studies 1 iirihcr. due to the lack of chronic duration studies, there
is additional unccrianil> reuardiim how longer-ierm exposures might impact hazard identification and
duse-ivspunsc assessmeiil fur I'l lJS \ 1a the oral route (e.g.. potentially more sensitive effects). Lastly,
as i inmunotoxicilv is an effect of increasing concern across several members of the larger PFAS
lamily. the lack of studies e\ aluating this ouicome following PFBS exposure is a limitation in the
dalabasc.
UFh
10
A U F11 of 10 is applied for inlcniidi\ idual variability in susceptibility in the absence of quantitative
information on the toxicokinetics and toxicodynamics of K+PFBS/PFBS in humans.
UFl
1
A UF|, of 1 is applied for LOAEL-to-NOAEL extrapolation because the POD is a BMDL and the
BMR w as selected based on evidence that it represented a minimal biologically significant response
level in susceptible populations such as pregnant mice and developing offspring.
UFs
1
A UFS of 1 is applied because the POD comes from a developmental study of mice. The
developmental period is recognized as a susceptible life stage in which exposure during certain time
windows (e.g., gestational) is more relevant to the induction of developmental effects than lifetime
cxDOSiirc CU.S. EPA. 1991b). The additional concern over DOtcntial hazards following loneer-term
(chronic) exposures is accounted for under the UFD above.
UFC
300
Composite UF = UFA x UFD x UFH x UFL x UFS
66
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
The candidate chronic RfD for effects of exposure to K+PFBS on the kidney, based on the
BMDLio (HED) of 11.5 mg/kg-day for papillary epithelial tubular/ductal hyperplasia in female
Po rats (Lieder et al.. 200%! is derived as follows:
Candidate Chronic RfD for K+PFBS (Kidney) = BMDLio (HED) - UFc
= 11.5 mg/kg-day ^ 1,000
= 0.12 mg/kg-day
= 1 x 10"2 mg/kg-day
Table 15 summarizes the UFs for the candidate chronic RfD for K PFBS based on effects in the
kidney.
Table 15. UFs for the candidate chronic RfD for kidney effects for K+PFBS
(CASRN 29420-49-3)
UF
Value
.luslification
UFa
3
A UFa of 3 (100 5) is applied to account I'm' in\isokinetic and toxicodynamic differences between rats
and humans following oral K+PFBS/PFI !S exposure Sonic aspects of the cross-species extrapolation
of toxicokinetic and toxicodynamic processes ha\ e been accounted for by calculating an HED by
applying a DAF as outlined in the EPA's Recommended ( se of Body Weight4 as the Default Method
inDerivation of the Oral Renin h r / '".v (U.S. 1P \. 2(>| Ihi. however, some residual uncertainty
remains. In the absence of chemical-specific data in t|iiauiil\ llns uncertainty, EPA's guidance
recommends use of a UF of '
UFd
3
A UFd of 3 is applied due to dalabase deficiencies. I he oral exposure database contains multiple
short-tenu and suhchrouic-duratioii io\icitv siudies ol l;ihoi;iior\ animals (NIP. 2018; Biiland et al..
2011: NTI\ 2u| I; 2(>|u; Lieder et al.. 2<»"Al. 2<>(>|. 2( km id), a two-generation reproductive
toxicity siud\ in mis (Lieder el al.. 200% i. and mulliDle developmental toxicity studies in mice and
rats (Fern: el al . 2
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Candidate chronic RfDs derived for thyroid effects and kidney effects are presented in Table 16.
Table 16. Summary of candidate noncancer chronic reference values for PFBS
(CASRN 375-73-5) and related compound K+PFBS (CASRN 29420-49-3)
Critical effect
POD method
POD (HED)
(mg/kg-d)
UFc
Reference
dose
(mg/kg-d)
Decreased serum total T4 in newborn (PND 1)
mice—Fens et al. (2017).
BMDL20
4.2
300
1 x 10~2
Kidney histopathology—papillary epithelial
tubular/ductal hyperplasia in P0 female
rats—Lieder et al. (2009b)
BMDL10
1 1 5
1000
1 X 10-2
Using the same calculation used for adjusting k Pl-'liS lo PI BS lor the subchronic RfDs, the
candidate chronic RfDs for PFBS (free acid) lor ihyroid and kidney effects are the same as the
values for the K+PFBS salt, as shown in Table I (¦>
The confidence in the candidate chronic RfDs for PI liS and k PFBS for both thyroid and
kidney effects is low, as explained in Table I 7 and I N below.
Table 17. Confidence descriptors for candidate chronic Ufl) ( I hyroid Effects) for PFBS
(CASRN 375-73-5) and the related compound K PI- liS (( ASUN 29420-49-3)
Confidence categories
Designation
Discussion
Confidence in study
II
( onfidcncc in the principal study is high because the overall study
design, performance, and characterization of exposure was good. Study
details and risk of bias analysis can be found in HAWC.
Confidence 111 database
L
Confidence in the oral toxicity database for derivation of the chronic
RID is low because, although there arc multiple short-term studies and
a subchronic-duralion toxicity study in laboratory animals, one
acceptable two-generation reproductive toxicity study in rats, and
multiple developmental toxicity studies in mice and rats, the database
lacks any chronic duration exposure studies or studies that have
evaluated ncurodcvclopmcntal or immunological effects.
Confidence in candidate
chronic RID
1.
The overall confidence in the candidate chronic RfD for thyroid effects
is low.
Notes'. H = high; L = low
68
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Table 18. Confidence descriptors for the candidate chronic RfD (Kidney Effects) for PFBS
(CASRN 375-73-5) and the related compound K+PFBS (CASRN 29420-49-3)
Confidence categories
Designation
Discussion
Confidence in study
H
Confidence in the principal study is high because the overall study
design, performance, and characterization of exposure was good. Study
details and risk of bias analysis can be found in HAWC.
Confidence in database
L
Confidence in the oral toxicity database for derivation of the chronic
RfD is low because, although there are multiple short-term studies and
a subchronic-duration toxicity study in laboratory animals, one
acceptable two-generation reproductive toxicity study in rats, and
multiple developmental toxicity studies in mice and rats, the database
lacks any chronic duration exposure studies or studies that have
evaluated neurodevelopmciilal or immunological effects.
Confidence in candidate
chronic RfD
L
The overall confidence in I lie candidate chronic RfD for kidney effects
is low.
Notes'. H = high; L = low
The chronic RfD is derived to be protective of all types of effects across studies and species
following oral chronic exposure and is intended to protect the population as a whole, including
potentially susceptible populations and life stages (I S LP.V 2""2Y This value should be
applied in general population risk assessments. Decisions concerning averaging exposures over
time for comparison with the RfD should consider the types of toxicological effects and specific
life stages of concern. For example, fluctuations in exposure le\ els that result in elevated
exposures during development could potentially lead to an appreciable risk, even if average
levels over the full exposure duration were less llian or equal lo the RfD. In light of the
consistent observation of the thyroid effects across life stages and the greater dose-response
sensitivity, relative to the kidney effects, EPA is proposing to base the overall chronic RfD on
the thyroid effects. See the Federal Register Notice announcing the availability of the draft
assessment for PIBS and requesting puMic review and comment on this proposal in addition to
the approaches and conclusions in the PFBS assessment.
6.2 uci ivai'nn of trice Concentrations
No published studies in\estigaling the effects of subchronic- or chronic-duration inhalation
toxicity of PIT5S and the related compound K+PFBS in humans or animals have been identified.
6.3 Cancer We ice Descriptor and Derivation of Cancer Risk Values
No studies evaluating the carcinogenicity of PFBS or K+PFBS in humans or animals were
identified. In accordance with the Guidelines for Carcinogen Risk Assessment (U.S. EPA. 20051
the EPA concluded that there is "inadequate evidence to assess carcinogenic potential" for PFBS
and K+PFBS by any route of exposure. Therefore, the lack of data on the carcinogenicity of
PFBS and the related compound K+PFBS precludes the derivation of quantitative estimates for
either oral (oral slope factor) or inhalation (inhalation unit risk) exposure.
6.4 Susceptible Populations and Life Stages
Early life stages as well as pregnant women are potentially susceptible to PFBS exposure. PFBS
has been detected in blood serum of nursing mothers, which might indicate a potential for
69
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
lactational exposure (Glynn et al.. 2012); however, information on the kinetics of lactational
transfer are lacking.
The available information suggests sex-specific variation in the toxicokinetics of PFBS in
rodents. Studies in mice and rats report clearance and elimination half-life times to be faster for
females than for males (see the "Toxicokinetics" section). For example, Chengelis et al. (2009)
reported that the mean apparent clearance of PFBS from the serum was approximately eightfold
higher for female rats (0.311 L/h/kg) than for male rats (0.0394 L/h/kg) and the mean apparent
volume of distribution for PFBS in the serum was approximately 2.4-fold higher for female rats
(0.288 L/kg) than for male rats (0.118 L/kg). Olsen et al. (2009) reported a statistically
significant difference in the urinary clearance rates (p <0.01) but not in serum half-lives, with
female rats (469 ± 40 mL/hour) having faster clearance rates than male rats (119 ± 34 mL/h).
Statistically significant sex-related differences in half-lilc or clearance were not observed
between male and female monkeys Olsen et al. Differences in the toxicokinetics in
rodents could result in sex-specific differences in toxicity studies
In vivo toxicity studies report that PFBS exposure can alter thyroid hormone levels in parental
and F1 generations animals (see "Thyroid Effects"). Thyroid hormones play a critical role in
coordinating complex developmental processes for \arions organs/systems (e g , reproductive
and nervous system), and disruption of thyroid hormone production/levels in a pregnant woman
or neonate can have persistent adverse health effects for the de\ eloping offspring (Ghassabian
and Trasande. 2018; Foster and Gray. 2013; Julvez et al.. 2" 13. Roman et al.. 2013).
Animal studies also pro\ ide e\ idence that gestationally exposed females might be a susceptible
subpopulation because of potential effects on female reproduction, including evidence of altered
ovarian follicle development and delayed vaginal opening (see "Reproductive Effects").
Furthermore, gestationally exposed females also had significantly reduced BWs and delayed eye
opening These findings suggest that de\ elopmental landmarks indicative of adverse responses
can be affected after PI"BS exposure (see "Offspring Growth and Early Development").
70
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Appendix A: Literature Search Strategy
This appendix presents the full details of the literature search strategy used to identify primary,
peer-reviewed literature pertaining to perfluorobutane sulfonic acid (PFBS) (Chemical Abstracts
Service Registry Number [CASRN] 375-73-5) and/or the potassium salt (K+PFBS)
(CASRN 45187-15-3). Initial database searches were conducted on July 18, 2017 using four
online scientific databases (PubMed, Web of Science [WOS], Toxline, and TSCATS via
Toxline) and last updated on February 28, 2018. The literature search focused on chemical name
and synonyms (see Table A-l) with no limitations on publication type, evidence stream
(i.e., human, animal, in vitro, and in silico) or health outcomes. Beyond database searches,
references were also identified from studies submitted under the Toxic Substances Control Act
(TSCA) and from review of other government documents (c u . Agency for Toxic Substances
and Disease Registry [ATSDR]) and combined u ilh the results of the database search. Search
results are retained in the EPA's Health and I-n\ iron mental Research Online (HERO) database.
Table A-l. Synonyms and MESH terms
ChemID
375-73-5
1,1,2,2. \ \4.44-Nfonafluoi'ii-l-hiiiaiicMilfonic acid
l-PerllikimhiiiMMCMilliiMic acid
Nonalluniii-I -hiiiaiicsiilfiiiiic acid
Nona IliinrnhiiiaiieMil funic acid
IVillikii'iihiiiaiicsiil Ionic acid
PI 1 iS
1.1.2. \ V \4.44-\mia lliinrohnia no-1-Mil phonic acid
PubMed (new onl\)
IVi'lliioiohiilaiic siilIonic acid
IVifliioiohiiiaiicMilfoiialc
IVi'lliioiohiilaiic siilfonalc
EPA Spreadsheet
1.1. \ \ V \4.44-\oiialliioio-l-hiiianesulfonic acid
1 -1 Jiiiaiicsiil funic acid. 1.1. W\ \4,4,4-nonafluoro-
1 -1 SiiiancMilIonic acid. nonnllnoro-
1 -1 Vrlliiorobulanosiullbiiic acid
\onallnoro-l-bulanesulfonic acid
Nonallnorobulanesulfonic acid
PI liS
Pc if 11 in in-1 -butanesulfonate
IViflnoiobutane Sulfonate
IV if 111 n robutanesulfonate
IV if 111 n robutanesulfonic acid
1 Vrfluorobutylsulfonate
45187-15-3
Note: MESH = Medical subject headings
A.l. Literature Search Strings
PubMed
375-73-5[rn] OR 45187-15-3[rn] "nonafluorobutane-1-sulfonic acid"[nm] OR
"1,1,2,2,3,3,4,4,4-Nonafluoro-l-butanesulfonic acid"[tw] OR "1-Perfluorobutanesulfonic
acid"[tw] OR "Nonafluoro-l-butanesulfonic acid"[tw] OR "Nonafluorobutanesulfonic acid"[tw]
A-l
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
OR "Perfluorobutanesulfonic acid"[tw] OR "1,1,2,2,3,3,4,4,4-Nonafluorobutane-l-sulphonic
acid"[tw] OR "Perfluorobutane sulfonic acid"[tw] OR "Perfluorobutanesulfonate"[tw] OR
"Perfluorobutane sulfonate"[tw] OR " 1-Butanesulfonic acid, l,l,2,2,3,3,4,4,4-nonafluoro-"[tw]
OR "1-Butanesulfonic acid, nonafluoro-"[tw] OR "Perfluoro-l-butanesulfonate"[tw] OR
"Perfluorobutylsulfonate"[tw] OR "Eftop FBSA"[tw] OR (PFBS[tw] AND (fluorocarbon*[tw]
ORfluorotelomer*[tw] OR polyfluoro*[tw] OR perfluoro-*[tw] OR perfluoroa*[tw] OR
perfluorob*[tw] OR perfluoroc*[tw] OR perfluorod*[tw] OR perfluoroe*[tw] OR
perfluoroh*[tw] OR perfluoron*[tw] OR perfluoroo*[tw] OR perfluorop*[tw] OR
perfluoros*[tw] OR perfluorou*[tw] OR perfluorinated[tw] OR fluorinated[tw] ORPFAS[tw]
OR PFOS[tw] OR PFOA[tw]))
wos
TS="l,l,2,2,3,3,4,4,4-Nonafluoro-l-butanesulfonic acid" OR IS "1 -Perfluorobutanesulfonic
acid" OR TS="Nonafluoro-l-butanesulfonic acid" OR TS="Nonalliioiobutanesulfonic acid" OR
TS="Perfluorobutanesulfonic acid" OR TS=" 1. 1.2.2.3,3,4,4,4-Noiialliioiol">utane-l-sulphonic
acid" OR TS="Perfluorobutane sulfonic acid" OR TS-'PerfluorobiiUincsiillonate" OR
TS="Perfluorobutane sulfonate" OR TS="l-l}iilanesulfonic acid, 1,1,2,2.3.3.4.4.4-nonafluoro-"
OR TS="1-Butanesulfonic acid, nonafluoro-" OR TS="Perlliioio-l-butancsullbnate" OR
TS="Perfluorobutylsulfonate" OR TS "l-ftop FBSA" OR ( I S PTBS AND TS=(tluorocarbon*
OR fluorotelomer* OR polyfluoro* OR perlliioro-* OR perlluoroa* OR perfluorob* OR
perfluoroc* OR perfluorod* OR perlluoroe* OR perlluoi'oh* OR perfluoron* OR perfluoroo*
OR perfluorop* OR perfluoros* OR periluoroir OR perfluorinaled OR fluorinated ORPFAS
ORPFOS OR PFO.M)
Toxline
( ( 375-73-5 [rn] OR 45 IK7-1 5-3 | i n| OR " I I 2 2 3 3 4 4 4-nonafluoro-l-butanesulfonic acid"
OR "l-perfluorobuUincsullbnic acid" OR "nonalliioro-1-butanesulfonic acid" OR
"nonafluorohuiUincsiillbiiic acid" OR "perlliiorolniuinesiilfonic acid" OR "1 1223344
4-nonafliK)i'i)l">iiUinc-l-siilphonic acid" OR "iierlliioroltiilane sulfonic acid" OR
"perfluorobutanesulfonate" OR "perfluorobutane sulfonate" OR "1-butanesulfonic acid 112 2 3
3 4 4 4-nonafluoro-" OR " I -Imuinesulfonic acid nonafluoro-" OR "perfluoro-l-butanesulfonate"
OR "perfluorobutylsulfonale" OR ' eftop fbsa" OR ( pfbs AND ( fluorocarbon* OR
fluorotelomer* OR polyfluoro" OR perfluoro* OR perfluorinated OR fluorinated OR pfas OR
pfos OR pfoa ) ) )) AND ( AM :i PL [org] OR BIOSIS [org] OR CIS [org] OR DART [org] OR
EMIC [org] OR EP11)1 AI |oiu| ORHEEP [org] ORHMTC [org] OR IP A [org] OR RISKLINE
[org] ORMTGABS [org| OR MOSH [org] ORNTIS [org] ORPESTAB [org] ORPPBIB [org]
) AND NOT PubMed [oi u] AND NOT pubdart [org]
TSCATS
375-73-5[rn] AND tscsats[org]
45187-15-3 [rn] AND tscsats[org]
A-2
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Appendix B: Detailed PECO Criteria
Table B-l. Population, exposure, comparator, and outcome criteria
PECO
element
Evidence
Population
Human: Any population (occupational; general population including children, pregnant women, and
other sensitive populations). The following study designs will be considered most informative:
controlled exposure, cohort, case-control, or cross-sectional. Note: Case reports and case series are
not the primary focus of this assessment and will be tracked as supplemental material during the study
screening process.
Animal: Nonhuman mammalian animal species (\\ hole ornauisni i of any life stage (including
preconception, in utero, lactation, peripubertal, and adiili sianesi
In vitro models of genotoxicity: The studies will he considered PI ICO-relevant. All other in vitro
studies will be tagged as "not-PECO relevant, but supplemental material."
Nonmammalian model systems/;'/? vitro/in silico NOT related lo genotoxicity: Nonmammalian
model systems (e.g., fish, amphibians, birds, and <' elegans); studies of Inn nan or animal cells,
tissues, or biochemical reactions (e.g., ligand hnidiiig assays) with in viim exposure regimens;
bioinformatics pathways of disease analysis, and or high throughput screening data. These studies
will be classified as non-PECO-relevant, bill ha\ e supplemental information
Exposure
Human: Studies providing qualitative or quantilal i\e esiiniales of exposure based on administered
dose or concentration, bioniouiioriun data te n . urine, hlood. or other specimens), environmental or
occupational-setting measures ic.n . water le\ els or air concent rations), residential location, job title
or other relevant occupational iiifornialiou 1 liiman "mixture" siudics are considered PECO-relevant
as long as the> ha\ e the per- and polylluoroalky 1 suhsiauces i HAS) of interest.
Animal: Studies pro\ idiun qualitative and quauiiiali\ c esiiniales of exposure based on administered
dose or conceuiraiiou ()ral and inhalation studies arc considered PECO-relevant. Nonoral and
noninhalation studies arc tanned as supplemental Experimental mixture studies are included as
PECO-relevant oulv if ilic> include a peri luorohutane sulfonic acid- (PFBS-) only arm. Otherwise,
mixture studies arc tanned as supplemental
\ll siudics must include exposure to PFBS. (' \SRN 375-73-5. Studies of precursor PFAS that
idcuiils au\ of the la meted PI' \S as melaholiies will also be included.
Comparator
Human \ comparison or reference population exposed to lower levels (or no exposure/exposure
he low detection lcxclsi or for shoricr periods of time. For D-R purposes, exposure-response
quantitative results niiisi he preseuied in sufficient detail such as regression coefficients presented
w illi statistical measure of \ arialiou such as RR, HR, OR, or SMR or observed cases vs. expected
cas>es i common in occupational studies); slope or linear regression coefficient (i.e., per unit increase
in a coin unions outcome i. difference in the means; or report means with results oft-test, mean
comparison b> renressiou. or other mean-comparing hypothesis test.
Animal: Qua ill ii;ii i\ c exposure versus lower or no exposure with concurrent vehicle control group.
Outcome
Cancer and noucauccr lieallli outcomes. In general, endpoints related to clinical diagnostic criteria,
disease outcomes*, liibtopalhological examination, genotoxicity, or other apical/phenotypic outcomes
will be prioritized for evidence synthesis. Based on preliminary screening work and other
assessments, the systematic review is anticipated to focus on liver (including serum lipids),
developmental, reproductive, neurological, developmental neurotoxicity, thyroid disease/disruption,
immunological, cardiovascular, and musculoskeletal outcomes.
Notes'. D-R = Dose-Response; HR = hazard ratio; OR = odds ratio; PECO = population, exposure, comparator, and outcome;
RR = risk ratio; SMR = standardized mortality ratio.
B-l
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Appendix C: Study Evaluation Methods
For each outcome in a study, in each domain, reviewers reached a consensus judgment of good,
adequate, deficient, not reported, or critically deficient. Questions used to guide the development
of criteria for each domain in epidemiology studies are presented in Table C-l and experimental
animal toxicology studies in Table C-3. These categories were applied to each evaluation domain
for each study as follows:
• Good represents a judgment that the study was conducted appropriately in relation to the
evaluation domain and any deficiencies, if present, are minor and would not be expected
to influence the study results.
• Adequate indicates a judgment that there are methodological limitations relating to the
evaluation domain, but that those limitations are not likely to be severe or to have a
notable impact on the results.
• Deficient denotes identified biases or deficiencies that are interpreted as likely to have
had a notable impact on the results or that |ii e\ ent interpretation of the study findings.
• Not reported indicates that the information necessary to evaluate the domain was not
available in the study. Generally, this term carries the same functional interpretation as
deficient for the purposes of the study confidence classification. Depending on the
number and severity of other limitations identified in the study, it may or may not be
worth reaching out to the study authors for this information.
• Critically deficient reflects a judgment that the study conduct introduced a serious flaw
that makes the ohser\ ed effect(s) uninterpiclaMc Studies with a determination of
critically deficient in an evaluation domain will almost always cause the study to be
considered inerall "uninformative".
Once the evaluation domains were rated, the identified strengths and limitations were considered
to reach a stuck confidence rating o\' high. medium, low, or uninformative for a specific health
outcome This was bused 011 the reviewer judgments across the evaluation domains and included
consideration of the likely impact the noted deficiencies in bias and sensitivity, or inadequate
reporting. ha\ e 011 the results The ratings, which reflect a consensus judgment between
reviewers, are defined as follows
• High. A w el I-conducted study with no notable deficiencies or concerns were identified;
the potential for bias is unlikely or minimal, and the study used sensitive methodology.
High confidence studies generally reflect judgments of good across all or most evaluation
domains.
• Medium '. A satisfactory (acceptable) study in which deficiencies or concerns were noted,
but the limitations are unlikely to be of a notable degree. Generally, medium confidence
studies will include adequate or good judgments across most domains, with the impact of
any identified limitation not being judged as severe.
• Low. A substandard study in which deficiencies or concerns were noted, and the potential
for bias or inadequate sensitivity could have a significant impact on the study results or
their interpretation. Typically, low confidence studies would have a deficient evaluation
for one or more domains, although some medium confidence studies could have a
deficient rating in domain(s) considered to have less influence on the magnitude or
C-l
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
direction of effect estimates. Generally, low confidence results are given less weight than
high or medium confidence results during evidence synthesis and integration and are
generally not used as the primary sources of information for hazard identification or
derivation of toxicity values unless they are the only studies available. Studies rated as
low confidence only because of sensitivity concerns about bias towards the null require
additional consideration during evidence synthesis. Observing an effect in these studies
could increase confidence, assuming the study was otherwise well-conducted.
• Uninformative'. An unacceptable study in which serious flaw(s) make the study results
unusable for informing hazard identification. Studies with critically deficient judgments
in any evaluation domain will almost always be classified as uninformative (see
explanation above). Studies with multiple deficient judgments across domains might also
be considered uninformative. Uninformative studies will not be considered further in the
synthesis and integration of evidence for hazard identification or dose response but might
be used to highlight possible research gaps
Table C-l. Questions used to guide the development of criteria for each domain in
epidemiology studies
Core question
Prompting questions
l-ollow -up questions
ExDOSure
For all:
• Does the exposure measure capture the \ariabiln> m c\posure
among the participant, considering iiiicusitv. frequency, and
duration of exposure?
• Docs ihc exposure measure reflect a rclc\ am lime w indow " If not,
can ilie relationship between measures in this lime and ihc relevant
time window he estimated reliably'
• W as ihc exposure measurement hkcl> lo he affected by a knowledge
of the outcome"
• Was the c\posurc nicasurcniciii likely Io he alfectedby the presence
of ihc outcome < i.e . rc\ erse causaliis i'1
l oi'casc-couirol studies of occupational exposures:
• Is exposure hascd on a comprehensive job history describing tasks,
setting, nine period, and use of specific materials?
l or hiomarkers of exposure, general population:
• Is a standard assa> used? What are the intra-and inter-assay
coefficients of \ anal ion? Is the assay likely to be affected by
contamination" \rc values less than the limit of detection dealt with
adequately'
What exposure time period is reflected by the biomarker? If the
half-life is short, what is the correlation between serial measurements
of exposure?
Is the degree of
exposure
misclassification likely
to vary by exposure
level?
If the correlation
between exposure
measurements is
moderate, is there an
adequate statistical
approach to ameliorate
variability in
measurements?
If there is a concern
about the potential for
bias, what is the
predicted direction or
distortion of the bias on
the effect estimate (if
there is enough
information)?
measurement
Does the
exposure
measure reliably
distinguish
between levels
of exposure in a
time window
considered most
relevant fur a
causal effect
with respeel In
the development
of the outcome'.'
C-2
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Core question
Prompting questions
Follow-up questions
Outcome
For all:
• Is outcome ascertainment likely to be affected by knowledge of, or
presence of, exposure (e.g., consider access to health care, if based
on self-reported history of diagnosis)?
For case-control studies:
• Is the comparison group without the outcome (e.g., controls in a
case-control study) based on objective criteria with little or no
likelihood of inclusion of people with the disease?
For mortality measures:
• How well does cause of death data reflect occurrence of i lie disease
in an individual? How well do mortality daia relied incidence of the
disease?
For diagnosis of disease measures:
• Is diagnosis based on standard clinical criteria? If based on
self-report of diagnosis, what is ilic \ alidm of this measure?
For laboratory-based measures (e.g., hormone le\ cKi
• Is a standard assay used0 Dues the assaj ha\ e an acceptable level of
inter-assay variabilis " Is ihe sensitivity of the assa> appropriate for
the outcome measure in ilns studs population"
Is there a concern that
any outcome
misclassification is
nondifferential,
differential, or both?
What is the predicted
direction or distortion of
the bias on the effect
estimate (if there is
enough information)?
ascertainment
Does the
outcome
measure reliably
distinguish the
presence or
absence (or
degree of
severity) of the
outcome?
ParticiDant
For longitudinal cohort:
• Did pariicipauis \ iilunleer lor I lie cohort hased on know ledge of
exposure and or preclinical disease s\ niptonis" Was entry into the
cohort or continuation in the cohort related to exposure and
outcome"
For occupational cohort
• Did cutis into the cohort heum w uh the start of the exposure?
• \\ as follow -up or outcome assessment incomplete, and if so, was
follow-up related to hoili exposure and outcome status?
• ( ould exposure produce s\ niptoms that would result in a change in
work assignment work status ("healthy worker survivor effect")?
l or case-cornrol studs.
• Were coin rols representative of population and time periods from
which eases were drawn?
• Are hospital controls selected from a group whose reason for
admission is independent of exposure?
• Could recruitment strategies, eligibility criteria, or participation rates
result in differential participation relating to both disease and
exposure?
For population-based survey:
• Was recruitment based on advertisement to people with knowledge
of exposure, outcome, and hypothesis?
Were differences in
participant enrollment
and follow-up evaluated
to assess bias?
If there is a concern
about the potential for
bias, what is the
predicted direction or
distortion of the bias on
the effect estimate (if
there is enough
information)?
Were appropriate
analyses performed to
address changing
exposures over time in
relation to symptoms?
Is there a comparison of
participants and
nonparticipants to
address whether
differential selection is
likely?
selection
Is there
evidence that
selection into or
out of the study
(or analysis
sample) was
jointly related lo
exposure and In
outcome"
C-3
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Core question
Prompting questions
Follow-up questions
Confounding
Is confounding adequately addressed by considerations in...
a. ... participant selection (matching or restriction)?
b. ... accurate information on potential confounders, and
statistical adjustment procedures?
c. ... lack of association between confounder and outcome, or
confounder and exposure in the study?
d. ... information from other sources?
Is the assessment of confounders based on a thoughtful review of
published literature, potential relationships (e.g.. as can be uained
through directed acyclic graphing), minimizing potential o\ ercontrol
(e.g., inclusion of a variable on the pathway between exposure and
outcome)?
If there is a concern
about the potential for
bias, what is the
predicted direction or
distortion of the bias on
the effect estimate (if
there is enough
information)?
Is confounding
of the effect of
the exposure
likely?
Analvsis
• Are missing outcome, exposure, and covanale data recomii/cd ;tnd,
if necessary, accounted for in the aiiak sis"
• Does the analysis appropriately consider \ ariable distributions and
modeling assumptions?
• Does the analysis appropriately consider subgroups of interest
(e.g., based on variability in exposure le\ el or duration, or
susceptibility)?
• Is an appropriate analysis used for the stud> desiuu"
• Is effect modification considered, hased on considerations developed
a priori?
• Does the study include additional analyses addressum potential
biases or limitations (i.e . seusin\ ity analyses)"
If there is a concern
about the potential for
bias, what is the
predicted direction or
distortion of the bias on
the effect estimate (if
there is enough
information)?
Do the analysis
strategy and
presentation
convey the
necessary
familiarity with
the data and
assumptions?
Sensitivitv
• Is the exposure range adequate''
• Was the appropriate population included"
• Was the length of follow-up adequate" Is the time/age of outcome
ascertainment optimal dven the uiier\al of exposure and the health
outcome?
• Are there other aspects related to risk of hias or otherwise that raise
concerns about sensitivity?
Is there a
concern that
sensitivity of the
study is not
adequate in
detect an effect''
Seleclh e
• \rc the results needed for the IRIS analysis presented (based on a
priori spccificatiou)" If not. can these results be obtained?
• \re only sialisiicalk significant results presented?
reDortinii
Is there reason
to be concerned
about selective
reporting?
Note: IRIS = Integrated Risk information System
C.l. Exposure measurement evaluation criteria
The criteria used to evaluate exposure measurement for PFBS (Table C-2) are adapted from the
criteria developed by the National Toxicology Program (NTP) Office of Health Assessment and
Translation for their assessment of the association between perfluorooctane sulfonic acid (PFOS)
and perfluorooctanoic acid (PFOA) and immune effects (NTP. 2016. 2015) and were established
prior to beginning study evaluation. Standard analytical methods for evaluating individual
per- and polyfluoroalkyl substances (PFAS) in serum or whole-blood using quantitative
techniques such as liquid chromatography-triple quadrupole mass spectrometry are preferred
C-4
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
(CDC. 2018; U.S. EPA. 2014d. e; AT SDR. 2009; CDC. 2009V The estimated serum half-life of
PFBS is approximately 1 month (Lau. 2015; Olsen et al.. 2009). so unlike for some other PFAS
with longer half-lives, current exposure might not be indicative of past exposures. Little data is
available on repeated measures of PFBS in humans over time, so the reliability of a single
measure is unclear. The timing of the exposure measurement is considered in relation to the
etiologic window for each outcome being reviewed.
Table C-2. Criteria for evaluation of exposure measurement in epidemiology studies
Exposure
measurement
rating
Crileria
Good
All of the following:
• Evidence that exposure was consistent^ assessed nsnm well-established methods that
directly measure exposure (e.g., measurement of PI ' \S in blood, serum, orplasma).
• Exposure was assessed in a relevani lime window for developmeiu of the outcome (i.e.,
temporality is established and sufficient latency occurred prior lo disease onset).
• There is evidence that a sufficient proportion of the exposure data measurements are above
the limit of quantification forthe assa> so thai different exposure u roups canbe distinguished
based on the analyses conducted.
• The laboratory analysis included standard qnalits control measures with demonstrated
precision and accuracy.
• There is sufficient spccilicil> seiisiii\it\ and i.inue or \ arialion in exposure measurements
that would minimize potential for exposure measurement error and misclassificationby
allow ins e\posure classifications to he differentiated lie. can reliably categorize participants
into groups sneli as high vs low e\posnrei
Adequate
• Evidence that e\posnre was consistently assessed using well-established methods that
direct l\ measure e\posure (c.g . measurement of PFAS in blood, serum, orplasma), but there
were some minor concerns about qnalits control measures or other potential for
nondiffereniial niisclassilicatioii
Ok
• 1 \posnre was assessed using indirect measures (e.g., drinking water concentrations and
residential location/history, questionnaire, or occupational exposure assessment by a certified
indiist rial hvmenist) that have been validated or empirically shown to be consistent with
methods that directly measure exposure (i.e., inter-methods validation: one method vs.
another) Note This could be good if the validation was sufficient. All studies for PFBS used
direct measures.
And all of the following:
• Exposure w as assessed in a relevant time window for development of the outcome.
• Tliere is e\ idence that a sufficient proportion of the exposure data measurements are above
the limit of quantification for the assay.
• There is sufficient specificity/sensitivity and range or variation in exposure measurements
that would minimize potential for exposure measurement error and misclassification by
allowing exposure classifications to be differentiated (i.e., can reliably categorize participants
into groups such as high vs. low exposure), but there might be more uncertainty than in good.
C-5
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Exposure
measurement
rating
Criteria
Deficient
Any of the following:
• Some concern, but no direct evidence, that the exposure was assessed using poorly validated
methods.
• There is insufficient information provided about the exposure assessment, including
precision, accuracy, and level of quantification, but no evidence for concern about the
method used.
• Exposure was assessed in a relevant time window for development of the outcome. There
could be concerns about reverse causation between exposure and outcome, but there is no
direct evidence that it is present.
• There is some concern over insufficient spcu licily/scnsilivity and range or variation in
exposure measurements that may result in considerable exposure measurement error and
misclassification when exposure classifications are compared (i.e., data do not lend
themselves to reliably categorize participants into groups such as high vs. low exposure,
and/or there is considerable unceriaiiits in exposure values ilial do not allow for confidence in
the examination of small perunii chaimes in continuous exposures)
Critically
deficient
Any of the following:
• Exposure was assessed in a time window ilial is unknown or not rcle\ mil lor development of
the outcome. This could he due to clear e\ idence of re\ erse causation between exposure and
outcome, or other concerns such as the lack of lempoml ordering of exposure and disease
onset, insufficient lalenc>. or lia\ iiiu exposure measurements that are not reliable measures of
exposure during the eliolomc w indow
• Direct evidence thai hiasuas hkel>. since the exposure was assessed using methods with
poor \alidil>
• 1 !\ idence of differential exposure niisdassilicalion
C-6
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Table C-3. Questions used to guide the development of criteria for each domain in experimental animal toxicology studies
Evaluation
type
Domain-
core question
Prompting questions
Basic considerations
Reporting Quality
Reporting Quality -
Does the study report
information for evaluating the
design and conduct of the
study for the
endpoint(s)/outcome(s) of
interest?
Notes:
Reviewers should reach out to
authors to obtain missing
information when studies are
considered key for hazard
evaluation and/or dose-
response.
This domain is limited to
reporting. Other aspects of the
exposure methods,
experimental design, and
endpoint evaluation methods
are evaluated using the
domains related to risk of bias
and study sensitivity.
Does the study report the following?
• Critical information necessary to Dcrform
study evaluation:
o Species; test article name; levels and
duration of exposure; route (e.g., oral;
inhalation); qualitative or quantitative
results for at least one endpoint of interest.
• Imnortant information for evaluating the
study methods:
o Test animal: strain, sex, source, and
general husbandry procedures,
o Exposure methods: source, purity, method
of administration,
o Experimental design: frequency of
exposure, animal age and life stage during
exposure and at endpoint/outcome
evaluation,
o Endpoint evaluation methods: assays or
procedures used to measure the
endpoints/outcomes of interest.
These considerations typically do not need to be refined by
assessment teams, although in some instances the
imnortant information mav be refined depending on the
endpoints/outcomes of interest or the chemical under
investigation.
A judgment and rationale for this domain should be given
for the study. Typically, these will not change regardless of
the endpoints/outcomes investigated by the study. In the
rationale, reviewers should indicate whether the study
adhered to GLP, OECD, or other testing guidelines.
• Good: All critical and imnortant information is
reported or inferable for the endpoints/outcomes of
interest.
• Adeauate: All critical information is rcDortcd but some
imnortant information is missine. However, the
missing information is not expected to significantly
impact the study evaluation.
• Deficient. All critical information is rcDortcd but
imnortant information is missine that is cxocctcd to
significantly reduce the ability to evaluate the study.
• Critically Deficient Study report is missing any pieces of
critical information. Studies that are Critically Deficient
for reporting are Uninfonnative for the overall rating and
considered no further for evidence synthesis and
integration.
C-7
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Evaluation
type
Domain-
core question
Prompting questions
Basic considerations
Risk of Bias
Selection and performance bias
Allocation -
Were animals assigned to
experimental groups using a
method that minimizes
selection bias?
For each study:
• Did each animal or litter have an equal chance
of being assigned to any experimental group
(i.e., random allocation)?
• Is the allocation method described?
• Aside from randomization, were any steps
taken to balance variables across experimental
groups during allocation?
These considerations typically do not need to be refined by
assessment teams.
A judgment and rationale for this domain should be given
for each cohort or experiment in the study.
• Good: Experimental groups were randomized and any
specific randomization procedure was described or
inferable (e.g., computer-generated scheme). [Note that
normalization is not the same as randomization (see
response for 'Adequate') ]
• Adequate: Authors report that groups were randomized
but do not describe the specific procedure used (e.g.,
"animals were randomized"). Alternatively, authors used
a nonrandom method to control for important modifying
factors across experimental groups (e.g., body weight
normalization).
• Not Reported (interpreted as Deficient): No indication of
randomization of groups or other methods (e.g.,
normalization) to control for important modifying factors
across experimental groups.
• Critically Deficient: Bias in the animal allocations was
reported or inferable.
Observational
Bias/Blinding-
Did the study implement
measures to reduce
observational bias?
For each endpoint/outcome or grouping of
endpoints/outcomes in a study:
• Does the study report blinding or other
methods/procedures for reducing observational
bias?
• If not, did the study use a design or approach
for which such procedures can be inferred?
• What is the expected impact of failure to
implement (or report implementation) of these
methods/procedures on results?
These considerations typically do not need to be refined by
the assessment teams. [Note that it can be usefiil for teams
to identify highly subjective measures of
endpoints/outcomes where observational bias may strongly
influence results prior to performing evaluations.]
A judgment and rationale for this domain should be given
for each endpoint/outcome or group of endpoints/outcomes
investigated in the study.
• Good: Measures to reduce observational bias were
described (e.g., blinding to conceal treatment groups
during endpoint evaluation; consensus-based evaluations
of histopathology lesions).3
C-8
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Evaluation
type
Domain-
core question
Prompting questions
Basic considerations
• Adequate: Methods for reducing observational bias (e.g.,
blinding) can be inferred or were reported but described
incompletely.
• Not Reported. Measures to reduce observational bias
were not described.
o Interpreted as Adequate—The potential concern for
bias was mitigated based on use of
automated/computer driven systems, standard
laboratory kits, relatively simple, objective measures
(e.g., body or tissue weight), or screening-level
evaluations of histopathology.
o Interpreted as Deficient—The potential impact on the
results is major (e.g., outcome measures are highly
subjective).
• Critically Deficient: Strong evidence for observational
bias that could have impacted results.
Confounding/
variable control
Confounding-
Are variables with the
potential to confound or
modify results controlled for
and consistent across all
experimental groups?
For each study:
• Are there differences across the treatment
groups (e.g., co-exposures, vehicle, diet,
palatability, husbandry, health status, and so
forth) that could bias the results?
• If differences are identified, to what extent are
they expected to impact the results?
These considerations may need to be refined by assessment
teams, as the specific variables of concern can vary by
experiment or chemical.
A judgment and rationale for this domain should be given
for each cohort or experiment in the study, noting when the
potential for confounding is restricted to specific
endpoints/outcomes.
• Good: Outside of the exposure of interest, variables that
are likely to confound or modify results appear to be
controlled for and consistent across experimental groups.
• Adequate: Some concern that variables that were likely to
confound or modify results were uncontrolled or
inconsistent across groups, but are expected to have a
minimal impact on the results.
• Deficient: Notable concern that potentially confounding
variables were uncontrolled or inconsistent across groups
and are expected to substantially impact the results.
C-9
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Evaluation
type
Domain-
core question
Prompting questions
Basic considerations
• Critically Deficient Confounding variables were
presumed to be uncontrolled or inconsistent across
groups and are expected to be a primary driver of the
results.
Reporting and attrition bias
Selective Reporting and
Attrition-
Did the study report results for
all prespecified outcomes and
tested animals?
Note:
This domain does not consider
the appropriateness of the
analysis/results presentation.
This aspect of study quality is
evaluated in another domain.
For each study:
Selective reporting bias:
• Are all results presented for
endpoints/outcomes described in the methods
(see note)?
Attrition bias:
• Are all animals accounted for in the results?
• If there are discrepancies, do authors provide
an explanation (e.g., death or unscheduled
sacrifice during the study)?
• If unexplained results, omissions, and/or
attrition are identified, what is the expected
impact on the interpretation of the results?
These considerations typically do not need to be refined by
assessment teams.
A judgment and rationale for this domain should be given
for each cohort or experiment in the study.
• Good: Quantitative or qualitative results were reported
for all prespecified outcomes (explicitly stated or
inferred), exposure groups and evaluation timepoints.
Data not reported in the primary article is available from
supplemental material. If results, omissions, or animal
attrition is identified, the authors provide an explanation
and these are not expected to impact the interpretation of
the results.
• Adequate: Quantitative or qualitative results are reported
for most prespecified outcomes (explicitly stated or
inferred), exposure groups and evaluation timepoints.
Omissions and/or attrition are not explained, but are not
expected to significantly impact the interpretation of the
results.
• Deficient. Quantitative or qualitative results are missing
for many prespecified outcomes (explicitly stated or
inferred), exposure groups and evaluation timepoints
and/or high animal attrition; omissions and/or attrition
are not explained and may significantly impact the
interpretation of the results.
• Critically Deficient: Extensive results omission and/or
animal attrition is identified and prevents comparisons of
results across treatment groups.
C-10
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Evaluation
type
Domain-
core question
Prompting questions
Basic considerations
c
QJ
Jl
s
a>
o
JS
"S
S
tu
¦-
s
x
o
s.
M
W
Chemical Administration
and Characterization-
Did the study adequately
characterize exposure to the
chemical of interest and the
exposure administration
methods?
1Vote:
Consideration of the
appropriateness of the route of
exposure is not evaluated at
the individual study level.
Relevance and utility of the
routes of exposure are
considered in the PECO
criteria for study inclusion and
during evidence synthesis.
For each study:
• Does the study report the source and purity
and/or composition (e.g., identity and percent
distribution of different isomers) of the
chemical? If not, can the purity and/or
composition be obtained from the supplier
(e.g., as reported on the website)?
• Was independent analytical verification of the
test article purity and composition performed?
• Did the authors take steps to ensure the
reported exposure levels were accurate?
o For inhalation studies: Were target
concentrations confirmed using reliable
analytical measurements in chamber air?
o For oral studies: If necessary based on
consideration of chemical-specific
knowledge (e.g., instability in solution;
volatility) and/or exposure design (e.g., the
frequency and duration of exposure), were
chemical concentrations in the dosing
solutions or diet analytically confirmed?
• Are there concerns about the methods used to
administer the chemical (e.g., inhalation
chamber type, gavage volume, etc.)?
It is essential that these criteria are considered and
potentially refined by assessment teams, as the specific
variables of concern can vary by chemical.
A judgment and rationale for this domain should be given
for each cohort or experiment in the study.
• Good: Chemical administration and characterization is
complete (i.e., source, purity, and analytical verification
of the test article are provided). There are no concerns
about the composition, stability, or purity of the
administered chemical or the specific methods of
administration. For inhalation studies, chemical
concentrations in the exposure chambers are verified
using reliable analytical methods.
• Adequate: Some uncertainties in the chemical
administration and characterization are identified but
these are expected to have minimal impact on
interpretation of the results (e.g., source and vendor-
reported purity are presented, but not independently
verified; purity of the test article is suboptimal but not
concerning). For inhalation studies, actual exposure
concentrations are missing or verified with less reliable
methods.
• Deficient. Uncertainties in the exposure characterization
are identified and expected to substantially impact the
results (e.g., source of the test article is not reported;
levels of impurities are substantial or concerning;
deficient administration methods such as use of static
inhalation chambers or a gavage volume considered too
large for the species and/or life stage at exposure).
• Critically Deficient: Uncertainties in the exposure
characterization are identified, and there is reasonable
certainty that the results are largely attributable to factors
other than exposure to the chemical of interest (e.g.,
identified impurities are expected to be a primary driver
of the results).
C-ll
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Evaluation
type
Domain-
core question
Prompting questions
Basic considerations
Exposure Timing,
Frequency and Duration-
Was the timing, frequency,
and duration of exposure
sensitive for the
endpoint(s)/outcome(s) of
interest?
For each endpoint/outcome or grouping of
endpoints/outcomes in a study:
• Does the exposure period include the critical
window of sensitivity?
• Was the duration and frequency of exposure
sensitive for detecting the endpoint of interest?
Considerations for this domain are highly variable
depending on the endpoint(s)/outcome(s) of interest and
must be refined by assessment teams.
A judgment and rationale for this domain should be given
for each endpoint/outcome or group of endpoints/outcomes
investigated in the study.
• Good: The duration and frequency of the exposure was
sensitive and the exposure included the critical window
of sensitivity (if known).
• Adequate: The duration and frequency of the exposure
was sensitive and the exposure covered most of the
critical window of sensitivity (if known).
• Deficient. Hie duration and/or frequency of the exposure
is not sensitive and did not include the majority of the
critical window of sensitivity (if known). These
limitations are expected to bias the results towards the
null.
• Critically Deficient The exposure design was not
sensitive and is expected to strongly bias the results
towards the null. The rationale should indicate the
specific concern(s).
Outcome measures and results
display
Endpoint Sensitivity and
Specificity-
Are the procedures sensitive
and specific for evaluating the
endpoint(s)/outcome(s) of
interest?
Note:
Sample size alone is not a
reason to conclude an
individual study is critically
deficient.
For each endpoint/outcome or grouping of
endpoints/out comes in a study:
• Are there concerns regarding the specificity
and validity of the protocols?
• Are there serious concerns regarding the
sample size (see note)?
• Are there concerns regarding the timing of the
endpoint assessment?
Considerations for this domain are highly variable
depending on the endpoint(s)/outcome(s) of interest and
must be refined by assessment teams.
A judgment and rationale for this domain should be given
for each endpoint/outcome or group of endpoints/outcomes
investigated in the study.
Examples of potential concerns include:
• Selection of protocols that are insensitive or nonspecific
for the endpoint of interest.
• Use of unreliable methods to assess the outcome.
• Assessment of endpoints at inappropriate or insensitive
ages, or without addressing known endpoint variation
(e.g., due to circadian rhythms, estrous cyclicity, etc.).
C-12
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Evaluation
type
Domain-
core question
Prompting questions
Basic considerations
* Decreased specificity or sensitivity of the response due to
the timing of endpoint evaluation as compared to
exposure (e.g., short-acting depressant or irritant effects
of chemicals; insensitivity due to prolonged period of
nonexposure prior to testing).
Results Presentation-
Are the results presented in a
way that makes the data usable
and transparent?
For each endpoint/outcome or grouping of
endpoints/outcomes in a study:
• Does the level of detail allow for an informed
interpretation of the results?
• Are the data analyzed, compared, or presented
in a way that is inappropriate or misleading?
Considerations for this domain are highly variable
depending on the outcomes of interest and must be refined
by assessment teams.
A judgment and rationale for this domain should be given
for each endpoint/outcome or group of endpoints/outcomes
investigated in the study.
Examples of potential concerns include:
• Nonpreferred presentation such as developmental toxicity
data averaged across pups in a treatment group when
litter responses are more appropriate.
• Failing to present quantitative results.
• Pooling data when responses are known or expected to
differ substantially (e.g., across sexes or ages).
• Failing to report on or address overt toxicity when
exposure levels are known or expected to be highly toxic.
• Lack of full presentation of the data (e.g., presentation of
mean without variance data; concurrent control data are
not presented).
C-13
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
Evaluation
type
Domain-
core question
Prompting questions
Basic considerations
Overall Confidence
Overall Confidence-
Considering the identified
strengths and limitations, what
is the overall confidence rating
for the endpoint(s)/outcome(s)
of interest?
Note:
Reviewers should mark studies
that are rated lower than high
confidence only due to low
sensitivity (i.e., bias towards
the null) for additional
consideration during evidence
synthesis. If the study is
otherwise well-conducted and
an effect is obsen'ed, the
confidence may be increased.
For each endpoint/outcome or grouping of
endpoints/outcomes in a study:
• Were concerns (i.e., limitations or
uncertainties) related to the reporting quality,
risk of bias, or sensitivity identified?
• If yes, what is their expected impact on the
overall interpretation of the reliability and
validity of the study results, including (when
possible) interpretations of impacts on the
magnitude or direction of the reported effects?
Hie overall confidence rating considers the likely impact of
the noted concerns (i.e., limitations or uncertainties) in
reporting, bias, and sensitivity on the results.
A confidence rating and rationale should be given for each
endpoint/outcome or group of endpoints/outcomes
investigated in the study.
• High Confidence: No notable concerns are identified
(e.g., most or all domains rated Good).
• Medium Confidence: Some concerns are identified, but
expected to have minimal impact on the interpretation of
the results (e.g., most domains rated Adequate or Good,
may include studies with Deficient ratings if concerns are
not expected to strongly impact the magnitude or
direction of the results). Any important concerns should
be carried forward to evidence synthesis.
• Low Confidence: Identified concerns are expected to
significantly impact on the study results or their
interpretation (e.g., generally. Deficient ratings for one or
more domains). The concerns leading to this confidence
judgment must be carried forward to evidence synthesis
(see note).
• Uninformative: Serious flaw(s) that make the study
results unusable for informing hazard identification (e.g.,
generally. Critically Deficient rating in any domain;
many Deficient ratings). Uninformative studies are
considered no further in the synthesis and integration of
evidence.
Notes'. GLP =good laboratory practices; OECD = Organisation for Economic Cooperation and Development.
a For nontargeted or screening-level histopathology outcomes often used in guideline studies, blinding during the initial evaluation of tissues is generally not recommended as
masked evaluation can make "the task of separating treatment-related changes from normal variation more difficult" and "there is concern that masked review during the initial
evaluation may result in missing subtle lesions." Generally, blinded evaluations are recommended for targeted secondary review of specific tissues or in instances when there is a
predefined set of outcomes that is known or predicted to occur ("Crissmanet al.. 2004).
C-14
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Appendix D: HAWC User Guide and Frequently Asked Questions
D.l. What is HAWC and What is its Purpose?
HAWC (Health Assessment Workspace Collaborative) is an interactive expert-driven content
management system for human health assessments that is intended to promote transparency,
trackability, data usability, and understanding of the data and decisions supporting an
environmental and human health assessment. Specifically, HAWC is an interface that allows the
data and decisions supporting an assessment to be managed in modules (e.g., study evaluation,
summary study data, etc.) that can be publicly accessed on-line (see #2 below and Figure D-l).
Following literature search and screening that are conducted using HERO and DistillerSR.
HAWC manages each study included in an assessment and makes the extracted information
available via a web link that takes a user to a web page displaying study specific details and data
(e.g., study evaluation, experimental design, dosing regime, endpoints evaluated, dose response
data, etc., described in further detail below in #s 3-6). Finally, all data managed in HAWC is
fully downloadable using the blue "Download datasets" link (highlighted in the red box below)
also located in the grey navigation bar located on the assessment home page (discussed in # 1
below). Note that a user may quickly navigate HAWC by clicking on the file path (highlighted in
orange dashed box below) given in the grey row below the HAWC icon and Login bar (Figure
D-l). HAWC aims to facilitate team collaboration by scientists who develop these assessments
and enhance transparency of the process by providing online access (no user account required) to
the data and expert decisions used to evaluate potential human health hazard and risk of chemical
exposures.
~ - » o e s
Figure D-l. HAWC homepage for the public PFBS assessment.
D.2. How Do I Access HAWC?
HAWC is an open-source online application that may be accessed using the following link:
https://hawcprd.epa.gov/assessment/public/ and then selecting an available assessment. The
following browsers are fully supported for accessing HAWC: Google chrome (preferred),
Mozilla Firefox, and Apple Safari. There are errors in functionality when viewed with Internet
Explorer. No user account is required for access to public HAWC assessments. The assessments
located in HAWC are meant to accompany a textual expert synthesis of the data managed in
HAWC. Each written assessment document contains embedded URL links to the evidence in
HAWC (e.g., study evaluation, summary study data, visualizations, etc) supporting the
C O i Secure https hawcprd.epa.gov assessment/100000037/
HAwer
PuWlC ASSMCTMtflS
PFBS (2018)
Study lot
Vfeuafczstions
Assessment name PfBS
Ymm 2018
True
Public True
Midden on public page? Fas*
D-l
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
assessment text. The links embedded in an assessment document can be accessed by a mouse
click (or hover while pressing CTRL+right click).
D.3. What Can I Find in HAWC?
HAWC contains a comprehensive landscape of study details and data supporting an assessment.
Note that links are provided in the assessment text to guide the reader, but a user may also
navigate to the HAWC homepage for an assessment on their own. Once a user lands on an
assessment homepage all studies included in an assessment can be viewed by clicking the blue
"Study list" link (highlighted in the red box below) in the grey navigation pane (Figure D-2). By
clicking the study name listed in blue (under "Short citation") a user can view the full study
details, study evaluation, and experimental details and data. For example, in Figure D-2, a user
may click on 3M (2000d) (highlighted in orange dashed box below). This will take the user to
the 3M (2000d) study details page that includes a link to the study in HERO along with study
details, study evaluation, and available experimental (animal) and study population
(epidemiologic) groups.
C O i Secure https hawcprd.epa.govMdy/assessment/100000037/
~ •'©~Si
Contact About Pubic Assessments Login
Study --st
Available studies
Download dalasets
Short citation
I 3M. 2000. <289992
3M„ 2001. 4241248
3M. 2010. 3927382
Bao. 2017. 3860099
Be*. 2014.2713874
Btfand, 2011.1578502
Bombard. 1996.3859928
Bioassay Epidemiology analysis
A repeated dose range-fintfng toxicity study of T-7485 in Sprague-Dawtoy rats
STUOY NUMBER 132-006. SPONSOR: 3M Pharmaceutical. St. Paul. MN
55133-3320. TESTING FACILITY Pr.med.ca RedheSd. Redfied. AR 72132 STUOY
DATES Study Initiation; June 26.2000 Arvmal Phase Initiation: June 27.2000
Animal Phase CompSebon: Juty 7.2000 Study CompWlon: October 11.2000
3M. A 28-day oral (gavage) toxcHy study of T-7485 In Sprague-Dawiey rats. v
(Study Number 132-007). St Paul, MN: 3M Corporate Toxicology.
3M. TSCA 8(e) Substantial RaK Not*® Sulfonate-based and Cartooxytic-based ~
Fluorochemicals, Docket 8EHO-0598373 - Results from a mecharwbc
investigation of the effect of PFBS, PFKS, and PFOS on ipid and lipoprotein
metabO'sm in transgeoc m.ce. Case Number 8EHQ-10-003730H. (8EHQ-10-
00373DH). Submitted to the U.S. Environmental Protection Agency under TSCA
Section 8e. [TSCA Submission).
Bao WW et a: Gender-speotic associations between serum isomers of —
perfluoroalkyl substances and Wood pressure among Chinese: Isomers of C8
Health Projoct in China. Sconce of the Total Environment 607-608:1304-1312.
Bern M et al. Pop. heavy meta* and the blues: secondary analysis of persistant —
organic pooutants (POP), heavy metals and depressive symptoms n the NHANES
National Epidemiological Survey. British Medical Journal Open 4:e005142.
Bijland S et al. Perfluoroalkyi sulfonates cause alkyl chain length-dependent
hepatic steatosis and hypolipidemic mainly by
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
pie graph (Figure D-3). For full domain and rating details the user may click the blue
"View details" button (highlighted in the red box below). (Note that this example is given for the
3 VI (2000d)).
Risk of bias visualization
Reporting
Outcome
Assessme n t/Rosu Its
Allocation/Blinding
Exposure/Study
Design Utility
t
Reporting or AttritMniabie Control
Reporting domain
^3 Reporting of information necessary for study
evaluation
Good. Important information is provided for test
species, strain, source, sex, exposure methods,
experimental design, endpoint evaluations and the
presentation of results. The study was not conducted
under GLP guidelines; however, several reviews were
performed by the Quality Assurance Unit.
3M, 2000, 4289992 risk of bias summary
Figure D-3. Representative study evaluation pie chart with the reporting domain selected
and text populating to the right of pie chart.
D.5. How Do I Access Study Specific Information on Experimental and Study
Population Details, and Extracted Endpoint Data?
Specific information on experimental design, dosing (if animal bioassay), outcomes and
exposure (if epidemiology) and extracted endpoint data can be accessed from the study details
page by clicking on (for the 3M (2000d) study) Available animal bioassay experiments at the
bottom of the study details page. A user may click on the experiment name (highlighted in blue,
10 Day Oral) to view dosing/exposure details and available groups. Clicking on available animal
groups (e.g., Male Sprague-Dawlev or Female Sprague-Dawlev) will take the reader to a new
page with experimental group information (e.g., species/strain/sex , dosing regime information,
and available/additional end points information for animal studies; and outcome and exposure
information for epidemiologic studies. If a study reports data then the data are extracted and
managed as "available endpoints". If study authors include endpoints in the methods and results,
but do not report data the endpoint is listed under "additional endpoints" without dose-response
data. All endpoints are also clickable and contain an endpoint description, methods, and (if data
are reported) a clickable data plot (e.g.. Alanine Aminotransferase (ALT)). The description of
endpoints, methods, and data are often copied directly from the study report and, therefore, can
contain study author judgments and may not necessarily include EPA judgments on the endpoint
data that would be included in the assessment.
D.6. What Are Visualizations and How Do I Access Them?
The data managed in HAWC is displayed using visualizations that are intended to support textual
descriptions within an assessment. All visualizations can be accessed using the blue
"Visualizations" link (highlighted in the red box below) also found in the grey navigation pane
D-3
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
(Figure D-4A). Note that the available visualizations are at the discretion of the chemical
manager and are meant to accompany the assessment text. Visualizations are fully interactive.
Hovering and clicking on records in the rows and columns and data points on a plot will cause a
pop-up window to appear (Figure D-4B). This pop-up window is also interactive and clicking on
blue text within this pop-up will open a new web page with descriptive data.
Contact About Public Assessments Login
Put) Uc Assessments PFBS (2018) Visualizations
PFBS T4 (effect size, animal)
DOWNLOADS
Download dataseta
C<9*rlm*nt E ndpolrj
l»w!» Slwtfr 0**lgn
Nil" 2014 430SMS 28 CK.V 0-» '
• - •
H9WO
»r SC^^-EXrwtey 4 ' Of 28
Figure D-4A. Visualization example for PFBS. (Note that the records listed under each
column (study, experiment endpoint, units, study design, observation time, dose) and data
within the plot are interactive.)
NTP 2018, 4309741 / 28 Day Oral / Male Harlan Sprague Dawley Rat / Tetraiodothyronine (T4), Free
Study Experiment Animal Group
Endpoint
Endpoint name
Tetraiodothyronine (F4), Free
System
Endocrine
Organ Thyroid
Effect
Hormone
Effect subtype
Thyroid Hormone
Observation time
Day 28
Data reported?
~
Data extracted?
Values estimated?
"
Location in literature
R07-Hormone Summary
Expected response
adversity direction
any change from reference/control group
LOAEL
62.6 mg/kg-day
Monotonicity
not-reported
Statistical test description
Jonchkeere (trend) and Shirley or Dunn (pairwise) tests
Trend result
not reported
Power notes
Statistical analysis performed by Jonchkeere (trend) and Shirley or Dunn (pairwise) tests Statistical significance for a
treatment group indicates a significant pairwise test compared to the vehicle control group Statistical significance for the
control group indicates a significant trend test * Statistically significant at P <= 0.05 " Statistically significant at P <= 0.01
Close
Figure D-4B. Example pop-up window after clicking on interactive visualization links.
(In Figure D-4A the red circle for study NTP (2011); male at a dose of 500 mg/kg-day was
clicked leading to the pop-up shown above. Clicking on blue text will open a new window
with descriptive data.)
D-4
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
D.7. How do I download datasets?
A user may download any available dataset by first clicking on the blue "Download datasets"
link (highlighted in the red box below) in the grey navigation pane on the assessment homepage.
This takes the user to a new page where the desired data set may be selected for download as an
excel file (See representative image in Figure D-5).
e O a Secure https://hawcprd.epa.gov/assessment/100000037/downloads/
Public Assessments PFBS (2018) Downloads
Q ^ Upgrade rvow
space and aha
Contact About Public Assessments Login
AVAILABLE MOOULES
Study list
Visualizations
Download dalasets
PFBS (2018) downloads
Ail data irom HAWC aro exportable into Excel. Developer exports in JSON ft
• Literature-review
it are also available (please contact us lor more information).
Study evaluation
Microsoft Excel spreadsheet
Animal btoassay data
Complete export I End point summary
Mcrosoft Excel spreadsheet
Epidemiology data
data
Microsoft Excel spreadsheet
ln-vitro data
M cfosoft Excel spreadsheet
Visualizations
Visua&iattons can be exported inio SVG, PNG. POF. and PPTX formats; you can
Figure D-5. Representative data download page.
D.8. How Do I Access the Benchmark Dose Modeling Outputs?
Benchmark dose (BMD) modeling is performed on an endpoint by endpoint basis at the
discretion of the chemical manager. Those endpoints for which BMD modeling has been
completed are referenced in the assessment text and are available for viewing. To access BMD
modeling outputs the user can click on links included in the assessment text. Alternatively, the
user may navigate to the BMD modeling outputs by clicking on a study (e.g., Feng et al. (2017))
of interest from the Study list, an available animal bioassay experiment (in this example the 20
Day Oral Gestation), an available animal group (PO Female ICR Mice), and an endpoint of
interest (Tetraiodothyronine (T4). Free). Next navigate to the blue Actions button, click, and
scroll to "View session" (highlighted in the red box below) under BMD Modeling (Figure D-
6A). The BMD setup. Results, and Model recommendation and selection (highlighted in orange
dashed box below) are available for viewing (Figure D-6B). Selecting the BMD setup tab will
display the modeled dose-response data, the selected models and options, and all benchmark
modeling responses (BMRs). The results tab will display the BMD modeling output summary for
all models. A user may hover over a selected model row to visualize the model fit to the data. In
addition, a user may obtain the Benchmark Dose Software (BMDS) Output text by clicking the
"View" button under the "Output" column for each model that was run. The Model
recommendation and selection tab displays all models, warnings when appropriate, and the
recommendation for which models are valid, questionable, or failed to fit.
D-5
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
G i Secure https 7hawcprd.epa.gov/ani/erbdpoint/100002285/
tWx
Public Assessments PFBS (2018) Feng, 2017, 3856465 20 Day Oral Gestation PO Female ICR Mice
Contact About Public Assessments Login
atodothyronine (T4), Free
SELECTED ASSESSMENT
AVAILABLE MODULES
Study list
Visualizations
DOWNLOADS
Download dalasets
Tetraiodothyronine (T4), Free
Endpoint Details Plot
Endpoint name
Tetraiodothyronine (T4). Free
System
Endocrine
Organ
Thyroid
Effect
Hormone
Effect subty pe
Thyroid Hormone
Observation time
GD20
Additional tags
high confidence
Data reported?
~
Data extracted?
V*
Values estimated? —
Location in literature
Table 3
NOAEL
50 mg/kg-day
TetralodoWiyronli
BMP Modeling
View session
Ultei
0Doses * Study
^LOAEL
0NOAEL
{
100 200 300 400 500
pose (retykg-day)
Figure D-6A. Example BMD modeling navigation.
ft Secure httpj haweprd epa.goWbmd/ww>on/100000019/
~
Contact
About
Public Assessments
Login
Public Assessments PFBS (2018) Feng. 2017, 3856465 20 Day Oral Gestation PO Female ICR Mice Tetraiodothyronine (T4). Fre
Study list
Visualizatio
Download dalasets
BMD session
BMD setup Results Model recommendation and selection
Dose-response
Number of Animals
Dose (mg/kg-day
HED) >
0 8
7.8" 8
31*e 8
770 8
* NOAEl (No Observed Ai*eri« Ef»*c1
11 SigoftcoiMty diflerorrt from control (p < 0.05)
1 IQAEl (lOWW Owwrvod Adww E«»
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Appendix E. Additional Data Figures
Study Experiment Lndpoint
Units Study Design
Dose
lutR/kg-day)
NTP20I8
28-day oral
Free TctrBiodothyroninc
pg/ml
P0 Mouse. ICR (9. N-81
GD20
Ihigh confidence)
0
50
200
500
NTP20I8
28-day oral
Total Tetraiodothymniiie
PNDI
Ihigh confidencel
0
50
Fl Moose. JCR (9. N=IO)
P0 Moose. ICR (9. N=8)
PN030 Ihigh confidencel 0
PN1W1 Ihigh confidencel
GD20 Hugh confidence! 0
50
500
Total Tetfuiodothytonilte (T4f - ug/ml Fl Mouse. ICR (9. Nx|0|
Litter N
PNDI
Ihigh confidencel 0
50
200
statistically si»nific
9 percent control response
95% CI
-IfiO -140 -120 -100
Figure E-l. Serum free and total thyroxine (T4) response in animals following K+PFBS
exposure (click to see interactive data graphic).
E-l
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Route K\|x>Mirt' EadpuM
NTP, 20IR, 4W974I Hat, Harlan Spragtte-Oavtley (9. N=1-I0) oral gavage 2H day* Triiodothyronine (T3l Ihigh confidence! ng/dl.
Kat. Harlan Sprague-Dauley (ei\ N=9-10) oral gavage 28 days Triiodothyronine (T3)
heng. 2017. 383646S M Mouse, irR (9, N=I0>
1 Mouse. ICR 19. N=:«»
oral gavage Cat) 1 in 20 Triiodothyronine (T31
oral gavage GO! to 20 Triiodothyronine I 'D)
30
200
soo
N) Mouse. ICR (9. N=K)
oral gavage GO 1 to 20 Triiodothyronine (T3)
nj;/ml 0
50
200
500
Fl Mouse. ICR (9. N=I0)
oral gavugi- GDI u.> 20 Triiodothyronine IT?) - Litter N
ng/ml 0
statistically sigiiiflc
percent control response
9S<* CI
percent control response
Figure E-2. Serum total triiodothyronine (T3) response in animals following K+PFBS
exposure (click to see interactive data graphic).
E-2
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Study
Study Design
Exposure
observation lime
Units
Dose
(mg/kg-day)
NTP 2018.4309741
Rul. Harlan Spraguc-Dawlcy (9- N=l-10)
28 days
Day 28
ng/mL
0
62.6
125
250
500
1.000
Rat. Harlan Spraguc-Dawlcy (d\ N=9-I0)
28 days
Day 28
ng/mL
0
62.6
125
250
500
Feng 2017.3856465
Fl Mouse, ICR (9. N=10)
GDI lo20
PND1
ng/ml
0
50
200
500
PND30
ng/ml
0
50
200
500
PND60
ng/ml
0
50
200
500
PO Mouse, ICR (9, N=8)
GDI to 20
GD20
ng/ml
0
200
500
~ statistically significant
£ percent control response
M 95% CI
H
i
H
H*
-100 -80 -60 -40 -20 0 20 40 60 80 100
percent control response
Figure E-3. Serum thyroid-stimulating hormone (TSH) response in animals following
K+PFBS exposure (click to see interactive data graphic).
Experiment Endpoint
Feng. 2017. 3856465 20-day oral Eye Opening
gestation
Units Study Design
observation Confidence Dose
time (mg/kg-day)
days F1 Mouse, ECR {£, N=50} Beginning on
PND12
Eye Opening - Litter N days F1 Mouse, ECR {£, N=10) Beginning on
PND12
200
5CO
200
500
| statistically significant
0 percent control rescccse
M &5% Cf
-100 -SO
—I 1 1-
1
*
*
1 i
i
i H
>~>
i
-eo -40 -20 0 20 40
percent contra:; response
Figure E-4. Developmental effects (eye opening) following K+PFBS in rats
(click to see interactive data graphic).
E-3
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Study
Study Design
Route Exposure Kndpoint
(.nits Dose
(mg/kg-day)
O statistically signifk
6 percent control res|
HI 95% CI
Feng. 2017, 3856465 Fl Mouse. ICR <9. N=30) oral gavage GDI to 20 Estrous Cycle. Diestrus days 0
50
200
500
Fl Mouse. ICR (9. N=10) oral gavage GDI to 20 Estrous Cycle. Proestrus days 0
50
200
500
O
-100 -80 -60 -40 -20 0 20 40 60 80 100
percent control response
Figure E-5. Developmental effects (first estrus) following K+PFBS in rats
(click to see interactive data graphic).
Studs
Stud) Design
Route
Exposure
End point
Units
Dose
(mg/kg-day)
Littler. 2009. 157H545
Fl Rat. Sprague-Dawley (9. N=30)
oral gavage
Vaginal Patency
days
0
30
100
300
1.000
Feng. 2017. 3856465
Fl Mouse. ICR (9. N=30)
oral gavage
GDI to 20
Vaginal Patency
days
0
50
200
500
Fl Mouse. ICR (9. N=10)
oral gavage
GDI to 20
Vaginal Patency (litter)
days
0
50
200
500
O statistically significant
£ percent control response
M95<* CI
l I l I
-100 -80 -60-40
percent control response
Figure E-6. Developmental effects (vaginal patency) following K+PFBS in rats
(click to see interactive data graphic).
E-4
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Study
Aniniul description
Study
Munition
Ijtdpoinl
Incidence
Dmr
(mfirtis-dayi
Lieder, 2009, IS7K546
Ral. Crl:Cd#<5dHs* Br Vaf/Plmtm l9. N«!0)
"""""
wibchntaic
(90 day*)
Kidney, Edema. Focal Papillary
(VIO 40.0%)
0
60
2011
3/10(30.0%)
600
Kidney. Hyperplasia, Papillary
TpbuIaWPuctal Epithelium
0/10(0.0%)
0
60
1/10(10.0%)
200
6/10 (60.0%)
600
Rat. Crl:Cd«(Sd)Jji Br Vaf/Hiwtm (Cf, N-IO)
oral gavagc
Mihchfonic
(90 day*)
Kidney. Edema, Focal Papillary
000(01.0%)
0
60
200
3/10(30.0%)
6(10
Kwines. Hyperplasia, Papillary
Tubular,1>ucwl Epithelium
1/10(10.0%)
0
200
8/10(80.0%)
6a)
Lk»fcf. 2009. I578MS
Fl Rat. Spraguc-Dawley (9, N«30>
oral gavage
reproductive
Kidney. Edem*. Focal Papillary
0/30(0.0%)
0
7/30 (23.3%)
300
4/30(13.3%)
I.OCIO
Kidney, HypeipUvia, Papillary
TuMar/l)uctal Epithelium
2/30 (6u7%)
0
13/30(43.3%)
300
15/30(50.0%)
1.000
Fl Rat, Sprague- Dawley (Ct. N=30)
oral gavage
reproductive
Kidney. Edema, local PapilUrv
1/30(3-3%)
0
000(00%)
300
9/30(30.0%)
1.000
Kidney. HvperptavM, Papillary
Tubwlir/Ductal Epithelium
3/30(10.0%)
5/30(16.7%)
21/30 (70.0%)
0
300
1.000
PO Rat. SjWTigue-Dawtey <9, N«30)
oral gavage
reproductive
Kidney. Edema. Focal Papillary
1/30(3.3%)
8/30(26.7%)
7/30(23.3%)
0
300
1.000
Kidney1. Hyperplasia, Papillary
TuMjj,'Ductal Epithelium
3/30(10.0%)
16/30 (53.3%)
0
300
21/30(70.0%)
1.000
ft) Rat. Sprague-Diiwlcy
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Emlpuint Nunc
Study Name
F-xperimcnt Name
Animul Dncriptiun
Oburtaliiin Time
HI...] Una Nitnrycn IBIIN)
3M. 2000, 4289992
10 Day Oral
Ral. CitCd(Sd) lb. Brief)
D^ 11
Rat. Cri; Cd (Sd> lbs Hr !M. 2001. 4241246
28 D*y Oral
Rat. ClfcCD«$|» «f)
Day 29
Rat. Cit."CD(SDl |9»
Day 29
Lledcr. 2009, 1578546
90 Day Oral
Rat. CriCdWSdllgt Br V*t/Plu*rai icf)
Day 90
Ral. Crl.-Cd«NS4)lgt Br VaffPltwm <9)
Day 90
Creatinine (CURAT)
3M. 2000, 4289992
10 Day Oral
Ral. Cri: Cd (Sd) 11k Br Iff)
Day 11
Rat. Cri: Cd (Sd) lb* Brl9)
Day II
NTP 2018. 4309741
2ft Day Oral
Rat, Harlan Spraguc-Dawlcy icf i
Day 2*
Rat. Harlan Spraguc-Dawlcy i9l
Day 2ft
3M. 2001.4241246
2ft Day Oral
Rat. CltCDtSD) (if)
Day 29
Rat. CrICDtSDl i9)
Day 29
liedcr. 2009. 1578546
90 Day Oral
Rat. CrlCdfe'Sdilgs Br Vaf/I'luMm «f)
Day 90
RaL CrfcCMKSdMe* Br VaffPluMm (9)
Day 90
llydroncphrusu
3M. 2001. 4241246
28 Day Oral
Rat CtLCDiSD) I Cf)
Day 29
Day 29
Rat, CHCDtSD) <9>
Day 29
Kidney. Hydmncphrrnt»
3M. 2000, 4289992
10 Day Oral
Ral. Cri: Cd(Sd)(b« Brief)
Day II
Rat. Cri: Cd (Sd) lbs Br 19)
Day 15
Kidney. Conical Tubular Dilation. Focal
Liedcr. 21)09, 1578545
2 Generation Oral
PO Rat. Spraguc-Dawlcy (Cf)
PO Rat. Spraguc-Dawlcy <9)
LD22
FI Rat Sprague-Dawlev
Day 120
Liedcr. 2009. 1578546
90 Day Oral
RaL CrtCd^KSdllgs Bt VdCPIuMui «f)
Day 90
RaL CrLCd©
Day 90
Kidney Basophilia, Tubular. Mullifiul
Lieder. 2009, 1578546
90 Day Oral
RaL Cii:Cd@iSd|lj;» Bi VoOPIuMm ((f)
Day 90
RaL CrlCdWISdllgJ Br VaJ/Plostm i9>
Day 90
Subacute Inflammation. Cortex
3M. 2001. 4241246
2ft Day Oral
Rat. CiLCDtSD) (Cf)
Day 29
Kidney. Mononuclear Cell Infiltrate
Lledcr, 2009, 1578546
90 Day Oral
Rat. CrlCdtftSdilgs Br Vat/Plusim icf)
Day 90
RaL Crl.CdWtSdilgs Br VaBPluum i9>
Day 90
Kidney. Pyelnoephritii.. Chronic
Lietter. 2009. 1578546
90 Day Oral
Rat. C'rlCdtMSdilg* Br VatfPlutUn (cf)
Day 90
Ral. Crl.-Cd#
Day 90
Kidney, Netrmii. Papillary
Lieder, 2II0V. 1578545
2 Generation Oral
PO RaL Sprague-Dawley (tt)
PO Rat. Spraguc-Dawlcy <9)
11)22
FI RaL Sptagoc-Dawky «f I
Day 120
FI RaL Spraguc-Daw lcy (9)
Day 120
N IT 2018. 4309741
2ft Day Oral
RaL Hiirlan Sptague-Dawlfy trfi
Day 2ft
RaL Harlan Spraguc-Dawlcy (9)
Day 2ft
Ucdcr. 2009. 1578546
90 Day Oral
Rat. Cri:Cd«»tSd)l|» Br VaPPluMni icf I
Day 90
RaL Crl:Cd©cf I
Day 120
FI Rat, Spraguc-Dawlcy (9)
Day 120
Lieder, 2009, I578S46
90 Day Oral
Rai. CrlCd&tSdllgs Br Vat/PluMm (Cf )
Day 90
Ral, CrlCdWiSdllg* Br VaCPhnUn l9«
Day 90
Kidney. Minerahrotion
3M, 2000, 4289992
10 Day Oral
Rot. Cri: Cd (Sd) Ibt Br 1(f)
Day II
RaL Cri: Cd (Sd) U>» Br i9)
Day 15
Lieder. 2009, 1578546
90 Day Oral
Ral. CriCdSMSdllgs Br VaCPIustm Icf <
Day 90
RaL CrlCdiftSdlls» Br Vj0PluMin l9>
Day 90
M incr.ili/iHum
3M. 2001. 4241246
2ft Day Oral
RaL CtK'DtSDl (9)
Day 29
Mincnili/jitiun. Terminal
,3M, 2001, 4241246
2ft Day Oral
RaL Cri.CD(SD) (9)
Day 29
Kidney. Ncplmipalliy. Chnmic IViigiDMic
NTP 2018,4309741
2ft Day Oral
RaL Harlan Spiuguc-Dawley l(f 1
Day 2*
RaL Harlan Spraguc-Dawlcy l9i
Day 2ft
Kidney. Hyaline Droplet*. Cortical Tubule*
l ieder, 2009, 1578546
90 Day Oral
Rat. CriCd»»
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
K.iul|MiinI Name
Study Name
Experiment Name
Animal Description
Observation Time
PFB
Kidney Weight Effects
Kidney Weight. Absolute
Lieder. 2009. 1578545
2 Generation Oral
F2 Rat. Sprague-Dawlcy (d"9>
~ • • •—
• Doses
3M. 2000.4289992
10 Day Oral
Rat. Cri: Cd (Sdl lbs Br ((f)
Day 11
|—| Dose Range
Ral. Crl: Cd (Sd> lbs Br 4 9 >
Day 11
~ • •—
A Signiftcanl lncrcurc
3M. 2001.4241246
28 Day Oral
Rut. Crl:CD
I.D 22
M—• •—
—~
Fl Rat. Sprague-Dawley (Cf I
Day 120
N—• •—
—~
Fl Rat. Sprague-Dawley (9)
Day 120
M » •
—4
Kidney Weight, Right. Absolute
Lieder. 2(109. 1578545
2 Cicncralwn Ural
P!) Rat. Sprague-Dawley ((f)
H— •—
—•
P0 Rat. Sprague-Dawley (9)
LD 22
M—• •—
—~
Fl Rat. Sprague Dawley (rf|
Day 120
H—• •—
—~
Fl Rat, Spr.isuc-Dav.ley (9>
Day 120
M—• •—
—•
NTP 2018.4309741
28 Day Oral
Rai, Harlan Sprague-Dawley id)
Day 28
• • • ~
A
—~
Rat. Harlan Sprague-Dawley (9)
Day 28
» ~ • •
—~
Kidney Weight. 1 -eft. Relative
1 jeder. 2009.1578545
2 Generation Oral
It) Rat. Sprague-Dawley (Cfi
M—• •—
—~
P0 Rat. Sprague-Dawlcy <9)
LD22
<~«—~ •—
—~
Fl Ral. Sprague-Dawlcy (tfi
Day 120
»• • •—
—~
Fl Rat. Sprague-Dawlcy (9)
Day 120
—• •—
—~
Kidney Weight. Relative
Lieder. 2009.1578545
2 Generation Oral
F2 Ral. Sprague-Dawley (
Day 11
• • •—
—~
3M. 2001.4241246
28 Day Oral
Rat. CrliCIXSD) «fi
Day 29
t • •—
Rat.Crl:CD(SD>(9>
Day 29
( • •
A
Lieder. 2009. 1578546
90 Day Oral
Ral. l'rl:Cd<*
Day 120
M—• •—
—~
NTP 2018.4309741
28 Day Oral
Rat. Harlan Sprague-Dawley (cf)
Day 28
» • • •
A
—«
Ral, Harlan Sprague-Dawlcy 19)
Day 28
~ A A A
A
-A
-1
X)
0 100 200 300
<100 500 600 ?0o
800 900
1.000 1.
00
Dusc (mg/kg/day)
Figure E-9. Kidney weight effects following K+PFBS in rats (click to see interactive data
graphic).
E-7
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
had point (Same
Study Name
Experiment Name
Animal Description
Observation Time
PFBS Liver Kflecta
Alanine Aminotransferase (Al.T)
3M. 2000. 4289992
10 Day Oral
Rat. Crl: Cd (Sd) lb* Rr(9)
Day II
~ • • ~
• l>«v
Rat, Crl: Cd (Sd) lb* Br ((f)
Day 11
• ¦ »~
, I—| Dose Range
NTP 2018. 4309741
28 Day Oral
Rat. Harlan Sprague-Dawley (9)
D«y28
~ ' • A—
A
A, Significant Increase
Rot. Ilailan Sprugue -Oil»ley (Cf)
Day 28
~ « • •—
A
V Significant Desocasc
3M, 21)01,4241246
28 Day Oral
Rat. Crt:CD|SDl (9)
Day 29
*—• »-
Rat CrlrCDtSDMCfl
Day 29
•—~ •-
Lictler. 2009, 1578546
90 Day Oral
Rat. Crl:Cd#(Sd)lg* Br VaBPktiim (9»
Day 90
• i
Rat. CrltCdtiHSdllgt Br VafJPIustm ((f)
Day 90
~—« •
Aspartate Aminotransferase (AST)
3M. 2000. 4289992
10 Day Oral
Rat, Cd: Cd (Sd) lbs Br (9)
Day U
» • •
Rat. Crl: Cd (Sd) lb* Br (lgs Br Vafflfestm i9)
Day 90
• a
Rat. CrK\l(*(Sd)lgs Rr Vaf/Phwtm (rfi
Day 90
• a
Hcpaiocytc Cytoplasmic Alteration
NIP 2018. 4309741
28 Day Oral
Rat. Harlan Sprague-Dawley (9)
Day 28
» • • •
—«
Ral. Harlan Sprague-Dawley 1
Day 28
~ • • •
-«
Cellular Infiltration
NTP 2018,4309741
28 Day Oral
Rat. Harlan Sprague-Dawley (9i
Day 28
~ ~ « •
—4
RaL Harlan Sprague-Dawley (cf|
Day 28
~ • • •
—*
Mononuclear Cell Infiltrate
3M. 2000. 4289992
10 Day Oral
Rat. Crl: Cd (Sd) lbs Br ( (9)
Day 29
~ • •-
•
Rat. Crl:CD(SD) ((f)
Day 29
• • »-
~
Subacute Inflammation
3M. 2001. 4241246
28 Day Oral
Rat. CH:CIXSD) (9)
Du y 29
~—• •-
1
Rat Crt:CD«SD) ((f)
Day 29
« a »-
•
Day 43
• ¦ ¦
•
Subacute Inflammation. Terminal
3M. 2001. 4241246
28 Day Oral
Rat Crl:CD(SDi(9)
Day 29
•
Rat. CrlCDtSD) ((f)
Duy 29
~ • •
~
Inflammation. Chronic, (-ocal/Multifocal
Lieder. 2009. 1578546
90 Day Oral
Rai. C"ri:CdWtSd)lgs Br Vaf/Ptastm 191
Day 90
~—« •
Ral. Crl Br VatfPlmtm (rfi
Day 90
• •
Hepatocellular Hypotrophy
Lieder. 2009. 1578545
2 Generation Oral
PO Rat. Sprague-Dawley ((f)
•-
NT? 2018, 4309741
28 Day Oral
Ral, Harlan Sprague-Dawley (9)
Da y 28
> » » i
Rat. Harlan Sprague-Dawley ((f)
Day 28
~ • • •
Lieder. 2009. 1578545
2 (feneration Oral
Fl Rat. Sprague-Dawley (9)
Day 120
• • • »-
PI Rat. Sprague-Dawley ((f)
Day 120
~ • • •-
P0 Rut. Sprague-Dawley (9)
1 J) 22
~ • • •-
Liver. Necrosis
NTP 2018, 4309741
28 Day Oral
Rat. Harlan Sprague-Dawley (9)
Day 28
< « I •
Rat. Harlan Sprague-Dawley ((f)
Day 28
» • •
Necrosis
3M. 2001, 4241246
28 Day Oral
Rat. CrlCDtSD) ((f)
Duy 29
• • •-
Necrosis. Terminal
3M. 2001. 4241246
28 Day Oral
Rat. CriCDiSDxcf I
Day 29
~ • »-
Bile Duct Cyst
NTP 2018. 4309741
28 Day Oral
Rau Harlan Sprague-Dawley (9)
Day 28
• • • •
•
Ral. Harlan Sprague-Dawley irfl
Day 28
~ > * •
~
Hepatodiaphragmatic Nodule
NTP 2018, 4309741
28 Day Oral
Rat. Harlan Sprague-Dawley (9)
Day 28
~ • • •
•
Rau Hat Ian Sprague-Dawley tcf)
Day 28
~ • • •
•
Hemorrhage
NTP 2018, 4309741
28 Day Oral
Ral. Harlan Sprague-Dawley (9)
Day 28
• a a a
a
RaL Harlan Sprague-Dawley (C?)
Da y 28
~ • • •
•
Liver llisinpathology
3M. 2000.4289992
10 Day Oral
Ral. Crl: Cd (Sd) lbs Br (9)
Day 11
Rat. Crl: Cd (Sd) lbs Br (cf)
Day II
• • •-
3M. 2001. 42*11246
28 Day Oral
Rat. Crl:CD(SDl (9)
Duy 29
~ • •
Rat. Crl:CD(SDl (Cf)
Day 29
• • •
•
Liver Weight. Absolute
Lieder. 2009. 1578545
2 Generation Oral
F2 RaL Sprague-Dawley (tf 9)
M— »-
P0 Rat. S prague-Dawlev (9)
•• • •-
PI) Rut. Sprague-Dawley ((f)
M—• jSr
~
3M. 2000, 4289992
10 Day Oral
Rat, Crl: Cd (Sd) lbs Br(9)
Day 11
• • »-
-A
Rat. Crl: Cd iSd) lbs Br ((f)
Day II
• • •-
-A
NTP 2018, 4309741
28 Day Oral
Rat. Hartan Sprague-Dawley (Cf i
Duy 28
* * A—A—
3M. 2001, 4241246
28 Day Oral
Rol Cri:CD(SDl (9)
Day 29
• a •-
Rat. Crl:CD(SD) ((f)
Day 29
• a a-
—A
Lieder. 2009. 1578546
90 Day Oral
Rat. Crl:Cd»iSd>lg» Br VaPPIustm (9)
Day 90
~—• •
Ral, Crt:C-d#(Sd)lgs Br Vaf/Plustm (Cf)
Duy 90
Lieder. 2009. 1578545
2 Generation Oral
Fl Rai. Sprague-Dawley ((f)
Day 120
M—• •-
Liver Weight. Absolute
NTP 2018. 4309741
28 Day Oral
Rat. Harlan Sprague-Dawley (9)
Day 28
A—
A
-A
Liver Weight. Relative
Lieder. 2009. 1578545
2 Generation Oral
P0 Rat. Sprague-Dawley (9)
M—a a-
P0 Rot. Sprague-Dawley ((f )
W—• A"
-A
3M. 2000. 4289992
10 Day Oral
Rat, Crl: Cd (Sd) lbs Br (9)
Day 11
• •-
-A
Ral. Crl: Cd (Sd) lbs Br ((f)
Day II
• • •-
-A
NTP 2018, 4309741
28 Day Oral
Ral. Hartan Sprague-Dawley (9)
Day 28
~-•-A A—
A
-A
Rat. Harlan Sprague-Dawley (Cf)
Duy 28
~ A A A
A
3M. 2001. 4241246
28 Day Oral
RaL CrLCDISD) (9)
Day 29
( a a—
RatCrlCDiSDxcfl
Duy 29
~—• A-
—A
Lieder. 2009. 1S78546
90 Day Oral
Rat. Crl:Cd«KSd)lg> Rr Vaf/Plustm (9)
Duy 90
~ • •
RaL Crl:Cd®,(Sd)Igs Br Vaf/Plustm 0 100 201) 300
400 500 A00 700 81
It) 9110
uino i.
Dose img/Vg/dayl
Figure E-10. Liver effects following K+PFBS in rats (click to see interactive data graphic).
E-8
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Endpolnl Viamr
Studs Nantr
rxftrrtmntl Vattar
Aitlraal Dt«rrtp(lon 1
3hacrrMhm Tin
Al (ApuAl >
Btjland, 2011. 1)78)02
Subdnuaa. Oral
Mmnc. Apoc")4_cadeti Cetp 4 Bt VaPPkaaimirfl
Day 90
Chotnunil E«u» fCTi
Btjland. 2011. 1578502
Subthnmtc Oral
Movie. Afw'3 Lealen C«»p irfl
W«ak4
CluWaul Fir
Utffatnd. 2011. 157*502
Subdiruna.- Oral
Mnar. Apoe * JLeaieitCrtp
Day 11
IM. 2001.424124ft
28 Day Oral
Ral. CtWD(5Dl(9)
Day 29
Glyooul
Btjlaml. 2011. 1)7*502
Sttbihnxtic Oral
Motor. Apnr'5 t-nfat Cclp (rfl
Week 4
L«wt*uictti La-aae (LPL i
[inland. 2011. 1)78502
Sutiduuok.- Oral
Muw. Apue'J-LcMoiCdpKf 1
Week 4
Neutral Sterol EuRtaa
Lltjluitd. 2011.1)78503
Subclinma" Ural
Motor. Apoc1 VLcakn frtp (rfl
Week '1
' 1 rig lyanfc* (1* l( iI
«m. »**i timwi
10 Day Oral
Kat. Cltt'd (Sd)llwHr t9»
Day II
Ral.Ot rdtSdtllwBrtrfl
Day II
*M. 2001,424124ft
28 Day Oral
RiLCitCD«SD)t9)
Day 29
RaL CrtC D«' * LmJoi Crtp (rfl
Week J
VLDL Hall 4. tic
Bijltatd. 2011.1)785(12
Subdnim* Oral
Motor. Ap* • '-Irticn Crtp trf 1
Week -I
VLUL Uptake
Klffamd. 2011.1)78502
Sufadwiuc Oral
Muuir. Aptir'HralaiI'dp (cf)
Week J
Vtrj I ow Henatry 1 qwpmtriai (Villi i Apnlqioqimaeai R I ApnR) Secrtmn Km
Ktjland, 3011. 1578502
.Suhchmnc Oral
Motor. \poe' 11 ratal Crip (rfl
Week 4
Vm la* Demty Upoprotttn (VLDL» Trigl>opnde Seiivtkia Rale
Btjhrnd, 2011. 1)78)02
Suhchnxiic Oral
Motor. Apoe'J l«adeilfnp(rfl
W«k 4
nun IIDLCbvhstoul
Bijlaid, 2011. 1)78)02
Subdvunic Oral
Mo«c. Apoc'5-LcaJcai Crtp irfl
Week 4
H One Ranpc
• Dino
A SitaulVuii Incm
^ Sipnitkam I Jeter
»0 400 500 WXI 700 am 400 1.000 1.10
Unt Imp/kgAIn)
Figure E-ll. Effects on lipids and lipoproteins following K+PFBS in rats and mice
(click to see interactive data graphic).
E-9
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Appendix F. Benchmark Dose Modeling Results
F.l. Modeling of Noncancer Endpoints
As discussed in the body of the report under "Derivation of Oral Reference Doses," the
endpoints selected for benchmark dose (BMD) modeling were incidence of renal papillary
epithelial tubular/ductal hyperplasia in rats from Lieder et al. (2009a) and Lieder et al. (2009b);
thyroid hormones in pregnant mice and offspring at postnatal day (PND) 1, PND 30, and
PND 60 from Feng et al. (2017) and adult rats from NTP (2018); and developmental effects
(i.e., eye opening, first estrus, vaginal opening) from Feng et al. (2017). The animal doses in the
study, converted to human equivalent doses (HEDs), were used in the BMD modeling; the data
are available for download in Health Assessment Workspace Collaborative (HAWC). BMD
modeling was conducted by experts in quantitati\ e Benchmark Dose Software (BMDS) analysis
and interpretation. Links to the data and modeling output are included in Table F-l. The selected
point of departure (POD) (FLED) listed in Table I '-1 represents the best lining model for each
endpoint; if the data were determined to not be amenable to BMD modeling, the no observed
adverse effect level (NOAEL) or lowest observed ad\eise effect level (LOAN.) is listed.
Figure F-l illustrates the doses examined and XOAI-I.. I.OAI-I.. BMD, and benchmark dose
lower confidence limit (BMDL) \ al Lies for the polcnlial critical cllects.
Table F-l. Candidate PODs for 1 lie derivation of the suhchronic and chronic RfDs for
PFBS (CASRN 375-73-5) and the related compound K PI- liS (( ASRN 29420-49-3)
r.ndpoini/rcfcrcncc
Species/life stage—sex
Selected POD (HED)a
(mg/kg-d)
Kidney effects
Kidney hislopal lining—papillais cpilhclial luhular duclal
hvDerolasia—Liederel ;il (»>l>;i)
Ral Male
BMDLm = 47.0
kal Ivmalc
BMDLi o = 12.6
Kidnc> hisiiip;iilKilnu>—papillais cpilhclial liibular/duclal
hyperplasia—l.icdcrcl ;il (2uu I Si
Rat—Male
LOAEL = 15.5
Rat—Female
BMDL, sn= 1.6
Free T4-NTP (2018')
Rat—Male
LOAEL = 15.5
Rat—Female
LOAEL = 14.3
Total T4—Fens et al. (2017)
Mouse/Po—Female
BMDL20 — 7.8
Free T4—Fens et al. (2017s)
Mouse/Po—Female
NOAEL = 7.5
TSH—Fens et al. (2017)
Mouse/Po—Female
NOAEL = 7.5
Total T4 PND 1 (fetal n)—Fens et al. (2017)
Mouse/Fi—Female
NOAEL = 7.5
Total T4 PND 1 (litter ri)—Fens et al. (2017)
Mouse/Fi—Female
BMDLjn = 4.2
Total T4 PND 30—Fens et al. (2017)
Mouse/Fi—Female
BMDL20 — 7.8
F-l
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Endpoint/reference
Species/life stage—sex
Selected POD (HED)a
(mg/kg-d)
Total T4 PND 60—Fene et al. (20171
Mouse/Fi—Female
NOAEL = 7.5
TSH PND 30—Fene et al. (20171
Mouse/Fi—Female
NOAEL = 7.5
Developmental effects
Eves openine (fetal n)—Fens et al. (2017)
Mouse/Fi—Female
NOAEL = 7.5
Eves openine (litter n)—Fene et al. (20171
Mouse/Fi—Female
BMDLisd = 14.8
Vaeinal openine (fetal n)—Fene et al. (20171
Mouse/Fi—Female
BMDLisd ~ 12.4
Vaeinal openine (litter n)—Fene et al. (20171
Mouse/Fi—Female
BMDLisn = 7.9
First estrous (fetal n)—Fene et al. (20171
Mouse/Fi—Female
NOAEL = 7.5
First estrous (litter n)—Fene et al. (20171
Mouse I' —female
NOAEL = 7.5
Notes: BW = body weight; RID = reference dose; PFBS = pcrlluorohuiaue mi I Ionic ;icid; CASRN = Chemical
Abstracts Service Registry Number; K+PFBS = potassium peril uorobutane sulfonate. T3 = total triiodothyronine;
T4 = total thyroxine; TSH = thyroid-stimulating hormone
aFollowing U.S. EPA (201 lb") guidance, animal doses from candidate principal studies u ere converted to HEDs
through the application of a dosimetric adjustment factor < I) \l' i DAFs for each dose arc calculated as follows:
DAF = (BWa1/4 BWh14), where BW(, = animal BW and I >\V human PAY For all DAF calculations, a reference
human BW (BWh) of 80 kilogram (kg) (U.S. EPA. 19881 was used See Table 9 in assessment for full details. Links
are to the HAWC BMDS session containing lull modeling results fur iliai endpoint.
b Data from offspring,
F-2
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE
NOVEMBER 2018
study name
experiment name
animal description
endpoint name
observation time
Licder. 2009, 1578546
90 Day Oral
Rat. Crl:Cd@(Sd)lgs Br Vaf/Plustm «f)
Kidney. Hyperplasia. Papillary Tuhular/Ductal Rpithelium
90.0 days
Rat. Cri:Cd@(Sd)Igs Br Vaf/Plustm (9)
Kidney. Hyperplasia. Papillary Tuhular/Duclal Epithelium
90.0 days
Lieder, 2009, 1578545
2 Generation Oral
P0 Rat. Sprague-Dawley {(f)
Kidney, Hyperplasia, Papillary Tuhular/Ductal Epithelium
None not-reported
Pt) Rat. Sprague-Dawley (9)
Kidney. Hyperplasia. Papillary Tuhular/Ductal Epithelium
None not-reported
Fl Rat, Sprague-Dawley (cf)
Kidney, Hyperplasia, Papillary Tuhular/Ductal Epithelium
120.0 days
Fl Rat, Sprague-Dawley (9)
Kidney, Hyperplasia, Papillary Tuhular/Ductal Epithelium
120.0 days
NTP20I8.4309741
28 Day Oral
Rat. Harlan Sprague-Dawley (Cf)
Tetraiodothyronine (T4), Free
None not-reported
Rat, Harlan Sprague-Dawley (91
Tetraiodothyronine (T4), Free
None not-reported
Rat, Harlan Sprague-Dawley (Cf)
Tetraiodothyronine (T4). Total
None not-reported
Rat. Harlan Sprague-Dawley < 9)
Tetraiodothyronine (T4). Total
None not-reported
Rat, Harlan Sprague-Dawley (cf)
Thyroid Stimulating Hormone (TSH)
28.0 days
Rat. Harlan Sprague-Dawley <91
Thyroid Stimulating Hormone (TSH)
28.0 days
Rat. Harlan Sprague-Dawley (cf)
Triiodothyronine (T3)
None n<»t-rcportcd
Rat. Harlan Sprague-Dawley (9)
Triiodothyronine (T3)
None not-reported
Feng 2017.3856465
20 Day Oral Gestation
P0 Mouse. ICR (9)
Tetraiodothyronine (T4). Free
20.0 GD
Tetraiodothyronine (T4). Total
20.0 GD
Fl Mouse, ICR (9)
Tetraiodothyronine (T4), Total
Tetraiodothyronine (T4), Total - Litter N
1.0 PND
30.0 PND
60.0 PND
1.0 PND
P0 Mouse. ICR (9)
Thyroid Stimulating Hormone (TSH)
20.0 GD
Fl Mouse. ICR (9)
Thyroid Stimulating Hormone (TSH)
1.0 PND
30.0 PND
60.0 PND
P0 Mouse. ICR (9)
Triiodothyronine (T3)
20.0 GD
Fl Mouse. ICR (9)
Triiodothyronine (T3)
1.0 PND
30.0 PND
60.0 PND
Triiodothyronine (T3) - Litter N
1.0 PND
Eye Opening - Fetal N
12.0 PND
Eye Opening - Litter N
12.0 PND
First Estrous - Fetal N
24.0 PND
First Estrous - Litter N
24.0 PND
Vagina] Opening - Fetal N
24.0 PND
Vaginal Patency - Litter N
24.0 PND
PFBvS Candidate PODs for RIDs
-em
~ eo—
—•—
~
-•
o o
-e—>
-i—I I I I! II
Dose (mg/kg/day)
O BMD
) BMDL
O NOAEL
~ LOAEL
Doses
^ Dose Range
Figure F-l. Candidate PODs for the derivation of the subchronic and chronic RfDs for PFBS
(click to see interactive data graphic).
F-3
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
F.2. Modeling Procedure for Continuous Noncancer Data
BMD modeling of continuous data was conducted on the HAWC website using the U.S.
Environmental Protection Agency's (EPA's) BMDS (Version 2.7). All continuous models
available within the software were fit using a benchmark response (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 (e.g., BMR 20% relative deviation [RD] for T4 effects), as outlined in the Benchmark
Dose Technical Guidance (U.S. EPA. 2012). An adequate fit is judged based on the
%2 goodness-of-fit p- value (p > 0.1), magnitude of the scaled residuals in the vicinity of the
BMR, and visual inspection of the model fit. In addition to these three criteria forjudging
adequacy of model fit, a determination is made as to whether the \ ariance 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 eslimated from a homogeneous
variance model. If the test for homogeneity of variance is rejected (/» <> I ), the model is run
again while modeling the variance as a power function of the mean to account for this
nonhomogeneous variance. If this nonhomogcncous variance model does not adequately fit the
data (i.e., Test 3;p<0. 1), the data set is considered unsuitaMe lor BMD modeling. In cases in
which a model with # parameters = " dose-groups was lit to the data set and all parameters were
estimated and no p-value was calculated, that model was not considered for estimation of a POD
unless no other model provided adequate lit. Among all models providing adequate fit, the
BMDL from the model with the lowest Akaike's information criterion (AIC) was selected as a
potential POD when BMDT. values were sufficiently close (within threefold). Otherwise, the
lowest BMDL was selected as a potential POD from u liich to derive the oral reference
dose/inhalation reference concentration (RID RlX").
Modeling Predictions for Serum Total T4 in PND 1 Female Offspring (litter n)
The modeling results for total T4 in PND 1 female offspring (litter n) exposed gestation days
(GDs) I 2<> are shown in Table F-2. The Exponential 4 model (Figure F-2) was selected given
appropriate lit to the data and that the BMDL values differed by greater than threefold. The
output for the EPA's liMI)S model run is also provided below.
F-4
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Table F-2. Modeling results for total T4 in PND 1 female offspring (litter n) exposed
GDs l-20a
Model
Global p-
value
AIC
BMDisd
(HED)
(mg/kg-d)
BMDLisd
(HED)
(mg/kg-d)
BMD20
(HED)
(mg/kg-d)
BMDL20
(HED)
(mg/kg-d)
Residual
of interest
Linear
0.558
-4.72314
54.547
35.8919
27.9861
20.2211
0.349
Polynomial
0.558
-4.72314
54.547
35.8919
27.9861
20.2211
0.349
Power
0.558
-4.72314
54.547
35.8919
27.9861
20.2211
0.349
Hill
-999
-1.89
30.1637
8.14952
14.2873
3.26417
-0.0000023
Exponential-M2
0.7627
-5.34819
43.9284
WX
20.8317
12.5215
-0.5885
Exponential-M3
0.7627
-5.34819
43.9284
: v WX
20.8317
12.5215
-0.5885
Exponential-M4b
0.8421
-3.85031
30.3302
s x5 r i
13.2799
4.22705
-0.09388
Exponential-M5
-999
-1.89
.0 1 .(.
8.93443
14.X157
4.26686
4.453e-7
Notes'. BMD = maximum likelihood estimate of the exposure concentration associated with the selected BMR; BMDL = 95%
lower confidence limit on the BMD (subscripts denote BMR: i.e.. 20 = exposure concentration associated with 20% relative
deviation from control).
aFeng et al. ('2017').
b Selected model. Exponential 4 model was selected given appropriate fit to the data and that the BMDI. values differed by
greater than threefold. The Exponential 5 model w as not selected because it did not return a p-value.
F-5
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
cn
3
CD
IS)
c
O
Q_
in
QJ
cC
O Doses in Study
O LOAEL
O NOAEL
Exponential-M^
Dose (mg/kg-day HED)
Figure F-2. Exponential (Model 4) for total T4 in PND 1 female offspring (litter n) exposed
GDs 1-20 (Feng et al. (2017).
Exponential Model. (Version: 1.11; Date: 03/14/2017)
Input Data File: C:\Windows\TEMP\bmds-dfile-bc5js794.(d)
Gnuplot Plotting File:
Thu Aug 09 08:49:17 2018
BMDS Model Run
The form of the response function by Model:
Model 2: Y[dose] = a * exp{sign * b * dose}
Model 3: Y[dose] = a * exp{sign * (b * dose)Ad}
Model 4: Y[dose] = a * [c-(c-l) * exp{-b * dose}]
Model 5: Y[dose] = a * [c-(c-l) * exp{-(b * dose)Ad}]
F-6
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Note: Y[dose] is the median response for exposure = dose;
sign = +1 for increasing trend in data;
sign = -1 for decreasing trend.
Model 2 is nested within Models 3 and 4.
Model 3 is nested within Model 5.
Model 4 is nested within Model 5.
Dependent variable = Response
Independent variable = Dose
Data are assumed to be distributed: normally
Variance Model: exp(lnalpha +rho *ln(Y[dose]))
rho is set to 0.
A constant variance model is fit.
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 500
Relative Function Convergence has been set to:
Parameter Convergence has been set to: le-008
le-008
MLE solution provided: Exact
Initial Parameter Values
Variable Model 4
lnalpha
rho
a
b
c
d
-1.29725
0 Specified
1.512
0.0428586
0.434618
1 Specified
Parameter Estimates
Variable Model 4
lnalpha
a
b
c
Std. Err.
-1.29626
1.4541
0.0316353
0.416958
0.0611684
0.148456
0.0322218
0.222524
F-7
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
NC = No Convergence
Table of Stats From Input Data
Dose N Obs Mean Obs Std Dev
Dose
0
10
1.44
0.329
7.5
10
1.3
0.657
29.9
10
0.92
0.493
75
10
0.69
0.657
Estimated Values of Interest
0
7.5
29.9
75
Est Mean
1.454
1.275
0.9355
0.6853
Est Std
0.523
0.523
0.523
0.523
Scaled Residual
-0.08525
0.151
-0.09388
0.02816
Other models for which likelihoods are calculated:
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2:
Model A3
Model
Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)A2
Yij = Mu(i) + e(ij)
Var{e(ij)} = exp(lalpha + log(mean(i)) * rho)
Yij = Mu + e(i)
Var{e(ij)} = SigmaA2
Model
Likelihoods of
Log(likelihood)
Interest
DF
AIC
Al
5.944999
5
1.889998
A2
8.698072
8
1.396144
A3
5.944999
5
1.889998
R
0.3138778
2
3.372244
4
5.925156
4
3.850311
F-8
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Additive constant for all log-likelihoods = -36
the above values gives the log-likelihood
depend on the model parameters.
including
76. This constant added to
the term that does not
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels? (A2 vs. R)
Test 2: Are Variances Homogeneous? (A2 vs. Al)
Test 3: Are variances adequately modeled? (A2 vs. A3)
Test 6a: Does Model 4 fit the data? (A3 vs 4)
Test
Test 1
Test 2
Test 3
Test 6a
Tests of Interest
2*log(Likelihood Ratio) D. F.
16.77 6
5.506 3
5.506 3
0.03969 1
p-value
0.01017
0.1383
0.1383
0.8421
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose
levels, it seems appropriate to model the data.
The p-value for Test 2 is greater than .1. A homogeneous
variance model appears to be appropriate here.
The p-value for Test 3 is greater than .1. The modeled
variance appears to be appropriate here.
The p-value for Test 6a is greater than .1. Model 4 seems
to adequately describe the data.
Benchmark Dose Computations:
Specified Effect = 1.000000
Risk Type = Estimated standard deviations from control
Confidence Level
BMD
BMDL =
BMDU =
0.950000
30.3302
8.85171
750000
F-9
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Exponential Model. (Version: 1.11; Date: 03/14/2017)
Input Data File: C:\Windows\TEMP\bmds-dfile-nq31d_k8.(d)
Gnuplot Plotting File:
Thu Aug 09 08:49:17 2018
BMDS Model Run
The form of the response function by Model:
Model 2: Y[dose] = a * exp{sign * b * dose}
Model 3: Y[dose] = a * exp{sign * (b * dose)Ad}
Model 4: Y[dose] = a * [c-(c-l) * exp{-b * dose}]
Model 5: Y[dose] = a * [c-(c-l) * exp{-(b * dose)Ad}]
Note: Y[dose] is the median response for exposure = dose;
sign = +1 for increasing trend in data;
sign = -1 for decreasing trend.
Model 2 is nested within Models 3 and 4.
Model 3 is nested within Model 5.
Model 4 is nested within Model 5.
Dependent variable = Response
Independent variable = Dose
Data are assumed to be distributed: normally
Variance Model: exp(lnalpha +rho *ln(Y[dose]))
rho is set to 0.
A constant variance model is fit.
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 500
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
MLE solution provided: Exact
F-10
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Initial Parameter Values
Variable Model 4
lnalpha
rho
a
b
c
d
-1.29725
0 Specified
1.512
0.0428586
0.434618
1 Specified
Parameter Estimates
Variable Model 4
Std. Err.
lnalpha
a
b
c
-1.29626
1.4541
0.0316353
0.416958
0.0611684
0.148456
0.0322218
0.222524
NC = No Convergence
Table of Stats From Input Data
Dose N Obs Mean Obs Std Dev
0
10
1.44
0.329
7.5
10
1.3
0.657
29.9
10
0.92
0.493
75
10
0.69
0.657
Dose
0
7.5
29.9
75
Estimated Values of Interest
Est Mean Est Std Scaled Residual
1.454
1.275
0.9355
0.6853
0.523
0.523
0.523
0.523
-0.08525
0.151
-0.09388
0.02816
Other models for which likelihoods are calculated:
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
F-ll
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)A2
Model A3:
Yij = Mu(i) +
Var{e(ij)} = exp(lalpha + log(tnean(i)) * rho)
Model R: Yij = Mu + e(i)
Var{e(ij)} =
SigmaA2
Likelihoods of
Interest
Model
Log(likelihood)
DF
AIC
Al
5.944999
5
-1.889998
A2
8.698072
8
-1.396144
A3
5.944999
5
-1.889998
R
0.3138778
2
3.372244
4
5.925156
4
-3.850311
Additive constant for all log-likelihoods = -36.76. This constant added
the above values gives the log-likelihood including the term that does not
depend on the model parameters.
Test 1
Test 2
Test 3
Explanation of Tests
Does response and/or variances differ among Dose levels? (A2 vs. R)
Are Variances Homogeneous? (A2 vs. Al)
Are variances adequately modeled? (A2 vs. A3)
Test 6a: Does Model 4 fit the data? (A3 vs 4)
Test
Tests of Interest
2*log(Likelihood Ratio)
Test 1
Test 2
Test 3
Test 6a
16.77
5.506
5.506
0.03969
D. F.
6
3
3
1
p-value
0.01017
0.1383
0.1383
0.8421
The p-value for Test 1 is less than .05. There appears to be
difference between response and/or variances among the dose
levels, it seems appropriate to model the data.
The p-value for Test 2 is greater than .1. A homogeneous
variance model appears to be appropriate here.
F-12
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
The p-value for Test 3 is greater than .1. The modeled
variance appears to be appropriate here.
The p-value for Test 6a is greater than .1.
to adequately describe the data.
Benchmark Dose Computations:
Specified Effect = 0.100000
Risk Type = Relative deviation
Model 4 seems
Confidence Level
0.950000
BMD
BMDL
BMDU
5.94765
1.83718
18.5178
Exponential Model. (Version: 1.11; Date: 03/14/2017)
Input Data File: C:\Windows\TEMP\bmds-dfile-llwh9peo.(d)
Gnuplot Plotting File:
Thu Aug 09 08:49:17 2018
BMDS Model Run
The form of the response function by Model:
Model 2: Y[dose] = a * exp{sign * b * dose}
Model 3: Y[dose] = a * exp{sign * (b * dose)Ad}
Model 4: Y[dose] = a * [c-(c-l) * exp{-b * dose}]
Model 5:
Y[dose] = a * [c-(c-l) * exp{-(b * dose)
"d}]
Note: Y[dose] is the median response for exposure = dose
sign = +1 for increasing trend in data;
sign = -1 for decreasing trend.
Model 2 is nested within Models 3 and 4.
Model 3
is nested within Model 5.
Model 4 is nested within Model 5.
F-13
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Dependent variable = Response
Independent variable = Dose
Data are assumed to be distributed: normally
Variance Model: exp(lnalpha +rho *ln(Y[dose]))
rho is set to 0.
A constant variance model is fit.
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 500
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
MLE solution provided: Exact
Initial Parameter Values
Variable
Inalpha
rho
a
b
c
d
Model 4
-1.29725
0 Specified
1.512
0.0428586
0.434618
1 Specified
Parameter Estimates
Variable Model 4
_ —
Inalpha
a
b
c
NC = No Convergence
-1.29626
1.4541
0.0316353
0.416958
Std. Err.
0.0611684
0.148456
0.0322218
0.222524
Table of Stats From Input Data
Dose N Obs Mean Obs Std Dev
0
7.5
10
10
1.44
1.3
0.329
0.657
F-14
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
29.9
75
Dose
7.5
29.9
75
10
10
0.92
0.69
0.493
0.657
Estimated Values of Interest
Est Mean Est Std Scaled Residual
1.454
1.275
0.9355
0.6853
0.523
0.523
0.523
0.523
Other models for which likelihoods are calculated
-0.08525
0.151
-0.09388
0.02816
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Yij = Mu(i) +
e(ij)
Var{e(ij)} = Sigma(i)A2
Model A2:
Model A3: Yij = Mu(i) +
Var{e(ij)} = exp(lalpha + log(mean(i)) * rho)
Model R: Yij = Mu + e(i)
Var{e(ij)} = SigmaA2
Likelihoods of
Interest
Model
Log(likelihood)
DF
AIC
Al
5.944999
5
-1.889998
A2
8.698072
8
-1.396144
A3
5.944999
5
-1.889998
R
0.3138778
2
3.372244
4
5.925156
4
-3.850311
Additive constant for all log-likelihoods = -36.76. This constant added to the
above values gives the log-likelihood including the term that does not
depend on the model parameters.
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels? (A2 vs. R)
Test 2: Are Variances Homogeneous? (A2 vs. Al)
Test 3: Are variances adequately modeled? (A2 vs. A3)
F-15
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Test 6a: Does Model 4 fit the data? (A3 vs 4)
Test
Test 1
Test 2
Test 3
Test 6a
Tests of Interest
-2*log(Likelihood Ratio)
16.77
5.506
5.566
0.03969
D. F.
p-value
0.01017
0.1383
0.1383
0.8421
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose
levels, it seems appropriate to model the data.
The p-value for Test 2 is greater than .1. A homogeneous
variance model appears to be appropriate here.
The p-value for Test 3 is greater than .1. The modeled
variance appears to be appropriate here.
The p-value for Test 6a is greater than .1.
to adequately describe the data.
Model
4 seems
Benchmark Dose Computations:
Specified Effect = 0.200000
Risk Type = Relative deviation
Confidence Level
BMD
BMDL
BMDU
0.950000
13.2799
4.22705
39.2189
F.3. Modeling Procedure for Dichotomous Noncancer Data
BMD modeling of dicholomous noncancer data was conducted on the HAWC website using the
EPA's BMDS Version 2.7. For these data, the Gamma, Logistic, Log-Logistic, Log-Probit,
Multistage, Probit, and Weibull dichotomous models available within the software were fit using
a 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 based
on the %2 goodness-of-fit p-value (p> 0.1), scaled residuals at the data point (except the control)
closest to the predefined BMR (absolute value < 2.0), and visual inspection of the model fit. In
the cases where no best model was found to fit to the data, a reduced data set without the
high-dose group was further attempted for modeling and the result was presented along with that
of the full data set. In cases in which a model with # parameters = # dose-groups was fit to the
F-16
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
data set and all parameters were estimated and no /> value was calculated, that model was not
considered for estimation of a POD unless no other model provided adequate fit. Among all
models providing adequate fit, the BMDL from the model with the lowest 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 Predictions for Papillary Tubular/Ductal Epithelium Hyperplasia in Po Female
Rats
The modeling results papillary tubular/ductal epithelium hyperplasia in PO female rats are shown
in Table F-3. The Dichotomous Hill model (Figure F-3) was selected based on the lowest AIC.
The output for the EPA's BMDS model run is also provided below.
Table F-3. Modeling results for papillary tubular/ductal epithelium hyperplasia in Po
female ratsa
Model
Global
p-value
AIC
BMDio (HED)
(mg/kg-d)
BMDLio (HED)
(mg/kg-d)
Residual
of interest
Logistic
0.0099
154.462
44.8115
35.0795
-0.34
LogLogistic
0.2488
148.137
16.852
8.27226
-0.503
Probit
0.0121
153.975
42.7597
34.2962
-0.294
LogProbit
0.1977
147.751
29.4015
21.4928
0.193
Multistage
0.2216
147.724
18.1972
13.3757
-0.391
Gamma
0.2216
147.724
18.1971
13.3757
-0.391
Weibull
0.2216
147.724
18.1973
13.3757
-0.391
Dichotomous-Hillb
0.5993
147.623
24.915
11.4888
0.063
Notes'. BMD = maximum likelihood estimate of the exposure concentration associated with the selected BMR; BMDL = 95%
lower confidence limit on the BMD (subscripts denote BMR: i.e., 10 = exposure concentration associated with 10% extra risk).
a Lieder et al. (2009b)
b Selected model. Dichotomous Hill model was selected based on lowest AIC.
F-17
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
0.8-
"O
u
0.4-
Cll
O Doses in Study
• LOAEL
O NOAEL
• Dichotomous-H
% 0.2-
0.0-
0 50 100 150 200 250
Dose (mg/kg-day HED)
Figure F-3. Dichotomous-Hill model for papillary tubular/ductal epithelium hyperplasia in
Po female rats Lieder et al. (2009b).
Dichotomous Hill Model. (Version: 1.3; Date: 02/28/2013)
Input Data File: C: \lAlindows\TEMP\bmds-dfile-lohmpk31. (d)
Gnuplot Plotting File: C:\lAlindows\TEMP\bmds-dfile-lohmpk31.plt
Thu Aug 09 11:03:14 2018
BMDS Model Run
The form of the probability function is:
P[response] = v*g +(v-v*g)/[l+EXP(-intercept-slope*Log(dose))J
where: 0 <= g < 1, 0 < v <= 1
v is the maximum probability of response predicted by the model,
and v*g is the background estimate of that probability.
F-18
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Dependent variable = Incidence
Independent variable = Dose
Slope parameter is restricted as slope >= 1
Total number of observations = 5
Total number of records with missing values = 0
Maximum number of iterations = 500
Relative Function Convergence has been set to:
Parameter Convergence has been set to: le-008
le-008
Default Initial Parameter Values
v = 1
g = 0.1
intercept = -6.87307
slope = 1.44728
Asymptotic Correlation Matrix of Parameter Estimates
slope
v
1
-0.34
0.36
-0.45
-0.34
1
-0.47
0.45
intercept
0.36
-0.47
1
-0.99
slope
-0.45
0.45
-0.99
1
Variable
v
g
intercept
slope
Parameter Estimates
95.0% Wald Confidence Interval
Estimate Std. Err.
0.709341 0.0972023
0.116633 0.0558081
-10.1333 4.57076
2.60307
1.20367
Lower Conf. Limit
0.518828
0.00725113
-19.0918
0.243915
Model
Full model
Fitted model
Reduced model
AIC:
Analysis of Deviance Table
Log(likelihood) # Param's Deviance
-69.6709 5
-69.8116
¦93.2613
147.623
4
1
0.281529
47.1808
Test d.f.
1
4
Upper Conf. Limit
0.899854
0.226015
-1.1748
4.96223
P-value
0.5957
<.0001
F-19
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Goodness of Fit
Dose
Est.Prob. Expected
Observed
Size
0.0000
7.1000
23.4000
70.9000
236.0000
0.0827
0.0868
0.1624
0.5357
0.6990
2.482
2.604
4.872
16.072
20.971
3.000
2.000
5.000
16.000
21.000
30.000
30.000
30.000
30.000
30.000
ChiA2 = 0.28
d. f.
P-value = 0.5993
Benchmark Dose Computation
Specified effect =
Risk Type
Confidence
level =
BMD =
BMDL =
0.1
Extra risk
0.95
24.915
11.4888
Scaled
Residual
0.343
-0.392
0.063
-0.026
0.012
F-20
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Appendix G. References
3M (3M Company). (2000a). Acute dermal irritation study of T-7485 applied to New Zealand
white rabbits. (Study Number: 132-004). St. Paul, MN: 3M Corporate Toxicology.
3M (3M Company). (2000b). Acute dermal toxicity study of T-7485 applied to Sprague-Dawley
rats. (Study Number: 132-010). St. Paul, MN: 3M Corporate Toxicology.
3M (3M Company). (2000c). Acute ocular irritation study of T-7485 applied to New Zealand
white rabbits. (Study Number: 132-005). St. Paul. MN 3\l Corporate Toxicology.
3M (3M Company). (2000d). A repeated dose range-i Midinu toxicity study of T-7485 in
Sprague-Dawley rats. (Study Number: 132-00O) St Paul. \r\T 3M Pharmaceuticals.
3M (3M Company). (2001). A 28-day oral (ga\ age) toxicity study of T-7485 in Sprague-Dawley
rats. (Study Number: 132-007). St. Paul. MN 3M Corporate Toxicology.
3M (3M Company). (2002a). Delayed contact h> pci sensiti\ ity study of T-7485 in Hartley guinea
pigs (maximization test). (Study Number 132-<)| 5) St Paul, MN: 3M Corporate
Toxicology.
3M (3M Company). (2002b). Emiionmental. health, safety, and regulatory (EHSR) profile of
perfluorobutane sulfonate (I'lliS)
http://multimedia 3ni.com niws media 1723"3Q ehsi-|H'ollle-of-perfluorobutane-
sulfonate-plbs pel f
3M (3M Company). (2<)|<)) ISCA 8(c) substantial risk notice Sulfonate-based and carboxylic-
based fluorochemicals. docket 81-1 IO-()5l)8-373 - Results from a mechanistic
in\ estimation of the effect of I'lliS, I'M IS, and PTOS on lipid and lipoprotein metabolism
in transgenic mice |TSCA Submission] (8EHQ-10-00373DH). St. Paul, MN.
Aeroslar Sl-S I.I.C. (201 7) I'inal site inspection report of fire fighting foam usage at Dover Air
Force Ikise. Kent County. Delaware.
Anderson. RH; l.onu. (iC. Poller. R.C; Anderson. JK. (2016). Occurrence of select perfluoroalkyl
substances at I S Air force aqueous film-forming foam release sites other than fire-
training areas: J 'iekl-\alidation of critical fate and transport properties. Chemosphere 150:
678-685. http://dx.doi.orn, 10.1016/i.chemosphere.2016.01.014
Anonymous. (2017). Final technical memorandum proposed groundwater monitoring well
locations perfluorinated compound delineation, areas of concern perfluorinated
compound-1. Ellsworth Air Force Base, South Dakota. Anonymous.
ASTSWMO (Association of State and Territorial Solid Waste Management Officials). (2015).
Perfluorinated chemicals (PFCS): Perfluorooctanoic acid (PFOA) & perfluorooctane
sulfonate (PFOS) information paper.
G-l
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
AT SDR (Agency for Toxic Substances and Disease Registry). (2009). Draft toxicological profile
for perfluoroalkyls. Agency for Toxic Substances and Disease Regristry.
http://www.atsdr.cdc.gov/toxprofiles/tp200.pdf
AT SDR (Agency for Toxic Substances and Disease Registry). (2015). Draft toxicological profile
for perfluoroalkyls. https://www.atsdr.cdc.gov/toxprofiles/tp200.pdf
Bao. J: Lee. YL; Chen. PC: Jin. YH; Dong. GH. (2014). Perfluoroalkyl acids in blood serum
samples from children in Taiwan. Environ Sci Pollut Res Int 21: 7650-7655.
http://dx.doi.org/10.1007/sll356-014-2594-4
Bao. WW: Oian. ZM: Geiger. SD: Liu. E: Liu. Y: Wanu. SO. Lawrence. WR: Yang. BY: Hu.
LW: Zeng. XW: Dong. GH. (2017). Gender-specific associations between serum isomers
of perfluoroalkyl substances and blood pressure among Chinese: Isomers of C8 Health
Project in China. Sci Total Environ 607-OOX. 1304-1312.
http://dx.doi.Org/10.1016/i.scitotenv.2" I 7 07 124
Berk. M; Williams. LJ; Andreazza. A: Pasco. J A. Dodd. S: Jacka. FN. Movlan. S: Reiner. EJ;
Magalhaes. PVS. (2014). Pop, heavy metal and the hlnes- secondary analysis of persistent
organic pollutants (POP), hea\\ metals and dcpiessi\e symptoms in the \HANES
National Epidemiological Siiia ey IJM.I Open 4 e<><>5l42.
http://dx.doi.Org/10.l 136/bmiopen-2" 14-""5 142
Biiland. S: Rensen. PCN: Pieterman. I-.I. Maas. ACI-. \an dcr I loom. JW: vanErk. MJ: Havekes.
LM; Willems van Dijk. K; Chang. SC; Ehrcsman. I).l. ISulenhoff. JL; Princen. HMG.
(2011). Perfluoroalkyl sulfonates cause alkyl chain length-dependent hepatic steatosis
and hypolipidemia mainly by impairing lipoprotein production in APOE*3-Leiden CETP
mice. Toxicol Sci 123 29()-3<)3. hllp d\ doi org/10.1093/toxsci/kfrl42
Bogdanska. .1. Sundstrom. M. ISerustrom. U; Borg. D; Abedi-Valugerdi. M; Bergman. A:
Dcpi crre. J. Nobel. S. (2014) Tissue distribution of S-35-labelled
periluorobutanesulfonic acid in adult mice following dietary exposure for 1-5 days.
Chemosphere 98: 2S-30 http: d\ doi.org/10.1016/i.chemosphere.2013.09.062
Bomhard. E: Loser. I¦ (!9lK->) Acute toxicologic evaluation of perfluorobutanesulfonic acid. J
Am Coll Toxicol 1} Acute Toxic Data 15: S105.
http://dx.doi org I n I 177 10915818960150S175
Calvo. R: Obregon. MJ: Escobar del Rev. F: Morreale de Escobar. G. (1992). The rat placenta
and the transfer of thyroid hormones from the mother to fetus: Effects of maternal thyroid
status. Endocrinology 131: 357-365.
CDC (Centers for Disease Control and Prevention). (2009). Fourth national report on human
exposure to environmental chemicals. Atlanta, GA: U.S. Department of Health and
Human Services, Centers for Disease Control and Prevention.
http ://www. cdc. gov/exposurereport/
G-2
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
CDC (Centers for Disease Control and Prevention). (2018). National report on human exposure
to environmental chemicals. Updated tables, March 2018. Atlanta, GA: U.S. Department
of Health and Human Services, https://www.cdc. gov/exposurereport/
Chen. Q; Huang. R; Hua. L; Guo. Y; Huang. L; Zhao. Y; Wang. X; Zhang. J. (2018). Prenatal
exposure to perfluoroalkyl and polyfluoroalkyl substances and childhood atopic
dermatitis: A prospective birth cohort study. Environ Health 17: 1-12.
http://dx.doi.org/10.1186/sl2940-018-0352-7
Chengelis. CP: Kirkpatrick. JB; Myers. NR; Shinohara. M: Stetson. PL: Sved. DW. (2009).
Comparison of the toxicokinetic behavior of perfluoiohexanoic acid (PFHxA) and
nonafluorobutane-1 -sulfonic acid (PFBS) in cynomolgus monkeys and rats. Reprod
Toxicol 27: 400-406. http://dx.doi.org/10.10H-' i rcprolox 2"09.01.013
Choksi. NY: Jahnke. GD; StHilaire. C: Shelhv. M (2<~>03) Role of thyroid hormones in human
and laboratory animal reproductive health | Re\ iew]. Birth Delects Res B Dev Reprod
Toxicol 68: 479-491. http://dx.doi.oru I" 1""2 bdrb. 10045
Crissman. JW: Goodman. DG: Hildebrandt. PK; Maronpot. RR: Prater. DA. Riley. JH; Seaman.
WJ: Thake. DC. (2004). Best practices guideline Toxicologic histopathology. Toxicol
Pathol 32: 126-131. http://d\ doi oiu I0.108<) n I ^2o23<>4^0268756
Dong. GH: Tung. KY: Tsai. CH: Liu. MM. W aim. I). T in. W. Jin. YH: Hsieh. WS: Lee. YL:
Chen. PC. (2"13a) Serum polyfluoroalkyl concentrations, asthma outcomes, and
immunological markers in a case-control study of Taiwanese children. Environ Health
Perspect 121 507-513. 5 13e5<) I-5<)X http d\ doi.org/10.1289/ehp. 1205351
Dong. GH: Tunu. KY. Tsai. CI I. I.in. MM. Wanu. IX l.iu. W: Jin. YH: Hsieh. WS: Lee. YL:
Chen. PC (2') 13h) Serum polylluoroalkyl concentrations, asthma outcomes, and
immunological markers in a case-control study of Taiwanese children: Supplemental
materials [Supplemental Data| Environ Health Perspect 121: 1-8.
EFSA (Luropean Food Safety Authority) (2012). Perfluoroalkylated substances in food:
Occurrence and dietary exposure. EFSA J 10: 2743.
Eriksen. KT: Raaschou-Niclscn. (). S0rensen. M: Roursgaard. M: Loft. S: M0ller. P. (2010).
Genotoxic potential of the perfluorinated chemicals PFOA, PFOS, PFBS, PFNA and
PFHxA in human I lepG2 cells. MutatRes 700: 39-43.
http://dx.doi.Org/lU.1016/i.mrgentox.2010.04.024
ESTCP (Environmental Security Technology Certification Program). (2017). FAQs regarding
PFASs associated with AFFF use at U.S. military sites.
http://www.dtic.mil/dtic/tr/fulltext/u2/1044126.pdf
Feng. X: Cao. X: Zhao. S: Wang. X: Hua. X: Chen. L: Chen. L. (2017). Exposure of pregnant
mice to perfluorobutanesulfonate causes hypothyroxinemia and developmental
abnormalities in female offspring. Toxicol Sci 155: 409-419.
http://dx.doi.org/10.1093/toxsci/kfw219
G-3
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Finken. MJ; van Eiisden. M; Loomans. EM; Vrijkotte. TG; Rotteveel. J. (2013). Maternal
hypothyroxinemia in early pregnancy predicts reduced performance in reaction time tests
in 5- to 6-year-old offspring. J Clin Endocrinol Metab 98: 1417-1426.
http://dx.doi.org/10.1210/ic.2Q12-3389
Fisher. DA. (1997). Fetal thyroid function: Diagnosis and management of fetal thyroid disorders.
Clin Obstet Gynecol 40: 16-31.
Forhead. AJ; Fowden. AL. (2014). Thyroid hormones in fetal growth and prepartum maturation
[Review], J Endocrinol 221: R87-R103. http://dx.doi org,10.1530/JQE-14-0025
Foster. P: Gray. LE. Jr. (2013). Toxic responses of the reproducti\ e system. In CD Klaassen
(Ed.), Casarett & Doull's toxicology: The basic science of poisons (8th ed., pp. 861-906).
New York, NY: McGraw-Hill Education
http://www.mhprofessional.com/prodiicl php'*cat= 1 lo&ishn "071769234
Fraser. AJ: Webster. TF; Watkins. DJ: Strynar. M.I, kato. K; Calafat. AM. Vicira. VM; Mcclean.
MP. (2013). Polyfluorinated compounds in dust from homes, offices, and vehicles as
predictors of concentrations in office workers' serum l-nviron Int 60 12S-136.
http://dx.doi.org/10.1016/i.en\ inl 2o 13.08.012
Ghassabian. A: Trasande. L. (2018). Disruption in thyroid signaling pathway: A mechanism for
the effect of endocrine-disrupting chemicals on child neurodevelopment. Front
Endocrinol (T.ausanne) 9 209. http d\ doi.org doi |o 3389/fendo.2018.00204.
eCollection 2o IS
Gilbert. ME. (2011) Impact oflo\\-le\el thyroid hormone disruption induced by propylthiouracil
on brain development and function Toxicol Sci 124: 432-445.
Imp d\ doi org |o |o^3 loxsci kl'r244
Gilberl. Ml-. Ko\el. J. Chen. /. koihuchi. N. (2012). Developmental thyroid hormone
disruption pre\ alence. en\ ironmental contaminants and neurodevelopmental
consequences. Neurotoxicology 33: 842-852.
http://dx.doi.org/10 1 o 1 (-» j.neuro.2011.11.005
Gilbert. ME: Sanchc/.-l luerta. K. Wood. C. (2016). Mild thyroid hormone insufficiency during
development compromises activity-dependent neuroplasticity in the hippocampus of
adult male rats, l -ndocrinology 157: 774-787. http://dx.doi.org/10.1210/en.2Q15-1643
Gilbert. ME: Zoeller. RT. (2010). Thyroid hormones—impact on the developing brain: Possible
mechanisms of neurotoxicity. In GJ Harry; HA Tilson (Eds.), Neurotoxicology (3rd ed.,
pp. 79-111). New York, NY: Informa Healthcare.
Glynn. A: Berger. U: Bignert. A: Ullah. S: Aune. M: Lignell. S: Darnerud. PO. (2012).
Perfluorinated alkyl acids in blood serum from primiparous women in Sweden: serial
sampling during pregnancy and nursing, and temporal trends 1996-2010. Environ Sci
Technol 46: 9071-9079. http://dx.doi.org/10.1021/es3Q1168c
G-4
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Gyllenhammar. I; Diderholm. B; Gustafsson. J; Berger. U; Ridefelt P; Benskin. JP; Lignell. S;
Lam pa. E; Glynn. A. (2018). Perfluoroalkyl acid levels in first-time mothers in relation to
offspring weight gain and growth. Environ Int 111: 191-199.
http://dx.doi.Org/10.1016/i.envint.2017.12.002
Haddow. JE; Palomaki. GE; Allan. WC; Williams. JR; Knight GJ; Gagnon. J; O'Heir. CE;
Mitchell. ML; Hermos. RJ; Waisbren. SE; Faix. JD; Klein. RZ. (1999). Maternal thyroid
deficiency during pregnancy and subsequent neuropsychological development of the
child. N Engl J Med 341: 549-555. http://dx.doi.org/10.1056/NEJM19990819341Q801
Henrichs. J: Bongers-Schokking. JJ; Schenk. JJ; Ghassabian. A: Schmidt. HG: Visser. TJ:
Hooiikaas. H; de Muinck Keizer-Schrama. SM; Hofman. A: Jaddoe. VV: Visser. W:
Steegers. EA; Verhulst. FC: de Riike. YB; Tiemeier. H. (2010). Maternal thyroid
function during early pregnancy and cognitive functioning in early childhood: The
generation R study. J Clin Endocrinol Metah 95: 4227-4234
http://dx.doi.org/10.1210/ic.2010-Q4l 5
Hernandez. A. Fiering. S.; Martinez. E; Galton. \ A. St Germain. D. (2002) The gene locus
encoding iodothyronine deiodinase type 3 (Dio3) is imprinted in the Ictus and expresses
antisense transcripts. Endocrinology 134: 44S3-44SO
Hill. AB. (1965). The en\ ironmenl and disease Association or causation? Proc R Soc Med 58:
295-300.
Hillman. NH; Kallapur. SG. .lohe. All (2"I2) Physiology of transition from intrauterine to
extrauterine lilc Clin Perinatol 7(->l)-7X3
Houtz. EF: Sutton. R; Park. .IS. Sedlak. M (2<)I(•>) Poly- and perfluoroalkyl substances in
wastewater. Significance of unknown precursors. manufacturing shifts, and likely AFFF
impacts Water Res I42-I41) Imp. d\ doi.oru I".1016/i.watres.2016.02.055
Hu. \(. Andrews. DO: Lindstrom. Ali: Bruton. TA: Schaider. LA: Grandiean. P: Lohmann. R:
Cariunan. CC: Blum. A. Balan. SA; Higgins. CP: Sunderland. EM. (2016). Detection of
poly- and perfluoroalkyl substances (PFASs) in U.S. drinking water linked to industrial
sites, military lire training areas, and wastewater treatment plants. Environ Sci Technol
Lett 3: 344-35<). Imp dx.doi.org/10.1021/acs.estlett.6b00260
Hulbert. AJ. (2000). Thyroid hormones and their effects: a new perspective [Review], Biol Rev
Camb Philos Soc 75: 519-631.
Julvez. J: Alvarez-Pedrerol. M. ar; Rebagliato. M; Murcia. M; Forns. J: Garcia-Esteban. R;
Lertxundi. N: Espada. M; Tardon. A: Riano Galan. I; Sunyer. J. (2013). Thyroxine levels
during pregnancy in healthy women and early child neurodevelopment. Epidemiology
24: 150-157. htto://dx.doi.org/10.1097/EDE.ObO 13e318276ccd3
Kato. K; Calafat. AM: Needham. LL. (2009). Polyfluoroalkyl chemicals in house dust. Environ
Res 109: 518-523. http://dx.doi.Org/10.1016/i.envres.2009.01.005
G-5
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Kim. D; Kim. U. nJ; Kim. H; Choi. SD; Oh. J. (2016). Perfluoroalkyl substances in serum from
South Korean infants with congenital hypothyroidism and healthy infants - Its
relationship with thyroid hormones. Environ Res 147: 399-404.
http://dx.doi.Org/10.1016/i.envres.2016.02.037
Koopdonk-Kool. JM; de Vijlder. JJ; Veenboer. GJ: Ris-Stalpers. C: Kok. JH; Vulsma. T; Boer.
K: Visser. TJ. (1996). Type II and type III deiodinase activity in human placenta as a
function of gestational age. J Clin Endocrinol Metab 81: 2154-2158.
Korevaar. TIM: Muetzel. R; Medici. M; Chaker. L; Jaddoe. VWV: de Riike. YB; Steegers. EAP;
Visser. TJ: White. T: Tiemeier. H: Peeters. RP. (2016). Association of maternal thyroid
function during early pregnancy with offspring l() and brain morphology in childhood: A
population-based prospective cohort study. Lancet Diabetes Endocrinol 4: 35-43.
http://dx.doi.org/10.1016/S2213-8587O 5^)00327-7
Lasier. PJ; Washington. JW: Hassan. SM; Jenkins. I'M. (2011). Peilluorinated chemicals in
surface waters and sediments from northwest Georgia, USA, and their bioaccumulation
in Lumbriculus variegatus. Environ Toxicol ('hem 3<~) 2194-2201.
http://dx.doi.org/10.10Q2/etc.622
Lau. C. (2015). Perfluorinated compounds An oven iew In .IC DeWitt (Ed.), Toxicological
effects of perfluoroalkyl and polylluoroalkyl substances (pp 1-21). New York: Springer.
http://dx.doi.org/10.10Q7/97S-3-3 11 55 1 S-i) 1
Lavado-Autric. R; Auso. I-. (iaivia-Velasco. .IV. Del Carmen A rule. M; Escobar Del Rev. F;
Berbel. P; Mon eale De Escobar. G. (2<)<)3) I¦ arly maternal hypothyroxinemia alters
histogenesis and cerebral cortex cytoaivhiteclure of the progeny. J Clin Invest 111: 1073-
1082 hlln ' d\ doi oru'lO 1 172'JCHo2o2
Liedcr. I'll. Chanu. SC. York. IU.i. ISutenhol'f. .11.. (2<)09a). Toxicological evaluation of
potassium pei'lluorobuUinesullbnale in a 90-day oral gavage study with Sprague-Dawley
rats Toxicology 255 45-52 Imp dx.doi.org/10.1016/i.tox.2008.10.002
Lieder. PH; York. RG; Hakes. DC: Chang. SC: Butenhoff. JL. (2009b). A two-generation oral
gavage reproduction study with potassium perfluorobutanesulfonate (K+PFBS) in
Sprague Daw ley rats Toxicology 259: 33-45. http://dx.doi.Org/10.1016/i.tox.2009.01.027
Liu. X: Guo. Z; Krebs. k A. I'ope. RH; Roache. NF. (2014). Concentrations and trends of
perfluorinated chemicals in potential indoor sources from 2007 through 2011 in the US.
Chemosphere 98: 51-57. http://dx.doi.org/10.1016/i.chemosphere.2013.10.001
Mangelsdorf. I; Buschmann. J: Orthen. B. (2003). Some aspects relating to the evaluation of the
effects of chemicals on male fertility [Review], Regul Toxicol Pharmacol 37: 356-369.
http://dx.doi.org/10.1016/S0273-2300(03N)00026-6
MDCH (Michigan Department of Community Health). (2015). Letter health consultation:
Evaluation of drinking water near Wurtsmith Air Force Base, Oscoda, Iosco County,
Michigan. Atlanta, GA: U.S. Department of Health and Human Services, Agency for
G-6
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Toxic Substances and Disease.
https://www.atsdr.cdc.gov/HAC/phaAVurtsmithAFBAVurtsmithAFB H20 %20LHC 09
-08-2015 508.pdf
MDH (Minnesota Department of Health). (2017). Toxicological summary for: Perfluorobutane
sulfonate. Minnesota Department of Health, Health Based Guidance for Water Health
Risk Assessment Unit, Environmental Health Division.
http://www.health.state.mn.us/divs/eh/risk/guidance/gw/pfbssummary.pdf
Min. H: Dong. J: Wang. Y: Wang. Y: Teng. W: Xi. O: Chen. J (2016). Maternal
hypothyroxinemia-induced neurodevelopmental impairments in the progeny [Review],
Mol Neurobiol 53: 1613-1624. http://dx.doi.org/10.1007/s 12035-015-9101 -x
Mirabella. G: Feig. D; Astzalos. E; Perlman. K; Rovet. .11- (2< n n M The effect of abnormal
intrauterine thyroid hormone economies on infant cognitive abilities. J Pediatr Endocrinol
Metab 13: 191-194.
Nakavama. S: Strynar. MJ; Helfant. L; Egeghy. I'. Ye. X: Lindstrom. MS (2')<)7) Perfluorinated
compounds in the Cape Fear Drainage Basin in North Carolina. Em iron Sci Technol 41:
5271-5276. htto://dx.doi.org/10.1021 /es070792v
Nakavama. SF; Strynar. MJ: Reiner. JL; Delinskv. AD. Lindstrom. AB. (2010). Determination
of perfluorinated compounds in the upper Mississippi ri\ or basin. Environ Sci Technol
44:4103-4109 httn:"d\ doi.org/10. in21 csl<)<)3S2/.
NASEM (National Academies of Sciences, Engineering, and Medicine). (2017). Application of
systematic re\ iew methods in an overall strategy for evaluating low-dose toxicity from
endocrine active chemicals Washington. DC.: The National Academies Press.
http d\ doi org 1" 17220 2475S
NICN AS (National Industrial Chemicals Notification and Assessment Scheme). (2005). Existing
chemical hazard assessment report Potassium perfluorbutane sulfonate. Canberra,
Australia' Commonwealth of Australia, Australian Government Publishing.
http: www nicnas uo\ au data^assets/pdf file/0004/4927/Potassium Perfluorobutane
Sulfonate I'D I-' pdf
NRC (National Research Council) (1983). Risk assessment in the federal government:
Managing the process Washington, DC: National Academy Press.
http://dx.doi.org/lU.r7226/366
NTP (National Toxicology Program). (2005). Perfluorobutane sulfonate. Genetic toxicology -
bacterial mutagenicity. Chemical effects in biological systems (CEBS). (Study ID
A32037). Research Triangle Park, NC: National Institutes of Environmental Sciences,
National Institutes of Health.
http://tools.niehs.nih.gov/cebs3/ntpViews/?activeTab=detail&studvNumber=A32037
NTP (National Toxicology Program). (2011). NTP Protocol Outline of the Class Study of
Perfluorinated Chemicals [perfluorobutane sulfonate (PFBS), perfluorohexane sulfonate
G-7
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
potassium salt (PFHSKslt), perfluorooctane sulfonate (PFOS), perfluorohexanoic acid
(PFHxA), perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA),
perfluorodecanoic acid (PFDA) and WY 14643 (WY)] in Harlan Sprague Dawley rats
administered daily by gavage for 28-Days [NTP],
NTP (National Toxicology Program). (2012). Perfluorobutane sulfonate. Genetic toxicology -
micronucleus [NTP], Research Triangle Park, NC: U.S. Department of Health and
Human Services, National Institutes of Environmental Sciences, National Institutes of
Health. https://tools.niehs.nih.gov/cebs3/ntpViews/?studvNumber=002-01134-0002-
0000-3
NTP (National Toxicology Program). (2015). Protocol lo e\ aluate the evidence for an
association between perfluorooctanoic acid (PI OA) or periluorooctane sulfonate (PFOS)
exposure and immunotoxicity. Department of I leallh and I luman Services, Office of
Health Assessment and Translation.
https://ntp.niehs.nih.gov/ntp/ohat/pfoa pl'os protocol 2Q15'H-» 5<)S pdf
NTP (National Toxicology Program). (2016). NTP monograph on immunoloxicity associated
with exposure to perfluorooctanoic acid (PI'OA) or peril uorooctane sulfonate (PFOS).
Research Triangle Park, NC I S Department of I lea I ill and Human Sen ices, Office of
Health Assessment and Translation.
https://ntp.niehs.nih.gov/ntp olial pt'oa pl'os pl'oa pl'osmonograph 508.pdf
NTP (National Toxicology Program). (2018). 28-day e\ aluation of the toxicity (C07040) of
perfluorobutane sulfonate (PFBS) (375-73-5) on I larlan Sprague-Dawley rats exposed
via gavage | \TP| I S Department of Health and Human Services.
https://tools.niehs nih uo\ cehs3 \ ieu s/?action main.dataReview&bin id=3878
Olsen. (i\\ . Cluing. SC. \oker. PI-. (iornian. GS. lihivsman. DJ: Lieder. PH: Butenhoff JL.
(2i)i)1)) \ comparison of the pharmacokinetics of perfluorobutanesulfonate (PFBS) in
rats, monkeys, and humans Toxicology 256: 65-74.
hitp d\ doi.org/10 I"16/i.lox 2"i)8.11.008
Olsen. GW: Mair. DC: l.ange. CC. Harrington. LM: Church. TR: Goldberg. CL: Herron. RM;
Hanna. H; Nobilciii. .IB. Kios. JA: Reagen. WK; Lev. CA. (2017). Per- and
polyfluoroalkyl substances (PFAS) in American Red Cross adult blood donors, 2000-
2015. Environ Res 157 87-95. http://dx.doi.Org/10.1016/i.envres.2017.05.013
Pant. KJ. (2001). Evaluation of a test article in the Salmonella typhimurium/Escherichia coli
plate incorporation/preincubation mutation assay in the presence and absence of induced
rat liver S-9. (Sponsor's study no. 132-008). Redfield, AR: Primedica Redfield.
Patel. J: Landers. K; Li. H; Mortimer. RH; Richard. K. (2011). Thyroid hormones and fetal
neurological development. J Endocrinol 209: 1-8. http://dx.doi.org/10.1530/JOE-10-0444
Post. GB: Louis. JB; Lippincott. RL; Procopio. NA. (2013). Occurrence of perfluorinated
compounds in raw water from New Jersey public drinking water systems. Environ Sci
Technol 47: 13266-13275. http://dx.doi.org/10.1021/es402884x
G-8
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Puig-Domingo. M; Vila. L. (2013). The implications of iodine and its supplementation during
pregnancy in fetal brain development [Review], Curr Clin Pharmacol 8: 97-109.
Oin. XD; Oian. Z: Vaughn. MG; Huang. J; Ward. P; Zeng. XW; Zhou. Y; Zhu. Y; Yuan. P; Li.
M; Bai. Z; Paul. G: Hao. YT; Chen. W: Chen. PC: Dong. GH; Lee. YL. (2016). Positive
associations of serum perfluoroalkyl substances with uric acid and hyperuricemia in
children from Taiwan. Environ Pollut 212: 519-524.
http://dx.doi.Org/10.1016/i.envpol.2016.02.050
Oin. XD: Qian. ZM: Dharmage. SC: Perret. J: Geiger. SD: Rigdon. SE: Howard. S: Zeng. XW:
Hu. LW: Yang. BY: Zhou. Y: Li. M: Xu. SL: Bao. WW: Zhang. YZ: Yuan. P: Wang. J:
Zhang. C: Tian. YP: Nian. M: Xiao. X: Chen. W: Lee. YL: Dong. GH. (2017).
Association of perfluoroalkyl substances exposure with impaired lung function in
children. Environ Res 155: 15-21. http://dx doi org l".l"l(-> i envres.2017.01.025
Roman. GC: Ghassabian. A: Bongers-Schokkinu. .1.1. Jaddoe. VW. I lol'nmn. A: deRiike. YB:
Verhulst. FC: Tiemeier. H. (2013). Association of gestational maternal hypothyroxinemia
and increased autism risk. Ann Neurol 74 733-742 hltp://dx.doi.org I" 1002/ana.23976
Rudmann. DG: Foley. GL. (2018). Chapter 18. Female Reproductive System In MAWallig;
WMHaschek; CGRousseaux. li IJolon (Eds.), 1'iiiKhimeiitals of toxicologic pathology
(3rd ed., pp. 517-545). Cambridge. M A Academic Press
http://dx.doi.org/10.lQ16/B978-"-12-809841-7 dpi) I S-(-»
Rumpler. J: Das. K: Wood. ('. I .an. C: Strvnar. M. Wamlxmuh. .1 (2016). Pharmacokinetic
profiles of peilluoiol-mlane sulfonate and activation of hepatic genes in mice. Abstract
presented at Society of Toxicology Meeting. Vlarch 13 - 17, 2016, New Orleans, LA.
Savin. S. ( \eiic. I), \cdic. (). Radosa\ lic\ ic. R (2"03). Thyroid hormone synthesis and storage
in the thyroid gland of human neonates. J Pediatr Endocrinol Metab 16: 521-528.
Savu. I.. Vranckx. R: Mava. M. (ii ipois. D: Blouquit. MF: Nunez. EA. (1989). Thyroxine-
binding globulin and thyroxine-binding prealbumin in hypothyroid and hyperthyroid
developing rats. Biochim Iilophys Acta 992: 379-384.
Savu. L: Vranckx. R: Rouaxc-Romet. M: Mava. M: Nunez. EA: Treton. J: Flink. IL. (1991). A
senescence up-regulated protein: the rat thyroxine-binding globulin (TBG). Biochim
Biophys Acta 1<)1)7 N-22
Schaider. LA: Balan. SA: Blum. A: Andrews. DO: Strvnar. MJ: Dickinson. ME: Lunderberg.
DM: Lang. JR: Peaslee. GF. (2017). Fluorinated compounds in US fast food packaging.
Environ Sci Technol Lett 4: 105-111. http://dx.doi.org/10.1021/acs.estlett.6b00435
Seo. SH: Son. MH: Choi. SD: Lee. DH: Chang. YS. (2018). Influence of exposure to
perfluoroalkyl substances (PFASs) on the Korean general population: 10-year trend and
health effects. Environ Int 113: 149-161. http://dx.doi.Org/10.1016/i.envint.2018.01.025
G-9
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Sepulvado. JG; Blaine. AC; Hundal LS; Higgins. CP. (2011). Occurrence and fate of
perfluorochemicals in soil following the land application of municipal biosolids. Environ
Sci Technol 45: 8106-8112. http://dx.doi.org/10.1021/esl03903d
Sferruzzi-Perri. AN: Vaughan. OR: Forhead. AJ; Fowden. AL. (2013). Hormonal and nutritional
drivers of intrauterine growth [Review], Curr Opin Clin Nutr Metab Care 16: 298-309.
http://dx.doi.org/10.1097/MCQ.0b013e32835e3643
Shiue. I. (2016). Urinary arsenic, pesticides, heavy metals, phthalates, polyaromatic
hydrocarbons, and polyfluoroalkyl compounds are associated with sleep troubles in
adults: USANHANES, 2005-2006. Environ Sci Pol lut Res lnt24: 3108-3116.
http://dx.doi.org/10.1007/sll356-016-8Q54-6
Song. X: Tang. S: Zhu. H; Chen. Z; Zang. Z; Zhang. Y. \in. \. Wang. X: Yin. H: Zeng. F; He.
C. (2018). Biomonitoring PFAAs in blood and semen samples Investigation of a
potential link between PFAAs exposure and semen mobility in China. Environ Int 113:
50-54. http://dx.doi.Org/10.1016/i.em iiii.2" IS Dl.OlO
Stagnaro-Green. A: Abalovich. M; Alexander. E. A/.i/.i. I '. Mesiman. J: Kcuro. R. Nixon. A:
Pearce. EN: Soldin. OP: Sulli\an. S: Wiersinua. W (2') I I). Guidelines of the American
Thyroid Association for the diagnosis and management of thyroid disease during
pregnancy and postpartum. Thyroid 21 11 )S I -1 125
http://dx.doi.org/10.1089/thv.2" I I QQS7
Stenzel. D; Huttner. WIS (2"I3) Rol e of maternal thyroid hormones in the developing neocortex
and during human e\olulion Frontiers in Neuroanatomy 7: 19. http://dx.doi.org/doi:
10.3389/fnana 2') 13 <)<)<) I1) eColleclion 2d 13
Sundslrom. M. ISoudanska. .1. Pham. IIV: Allninasios. V. Nobel. S: McAlees. A: Eriksson. J:
DePierre. .IWISeruman. A (2<) 12). Radiosynthesis of perfluorooctanesulfonate (PFOS)
and |ieriluorol"mlanesullbnale (PI'BS), including solubility, partition and adhesion studies.
Chemosphere X7 X(o-X7l lilt p. dx.doi.org/10.1016/i.chemosphere.2012.01.027
Thompson. W. Russell. G: Baragwanath. G: Matthews. J: Vaidva. B; Thompson-Coon. J. (2018).
Maternal thyroid hormone insufficiency during pregnancy and risk of
neurodevelopmental disorders in offspring: A systematic review and meta-analysis. Clin
Endocrinol XX 575-5X4. http://dx.doi.org/10.Ill 1/cen. 13550
Tsai. C: Huang. J: Hwang. BF: Lee. YL. (2010). Household environmental tobacco smoke and
risks of asthma, wheeze and bronchitic symptoms among children in Taiwan. Respir Res
11:11. http://dx.doi.Org/10.l 186/1465-9921-11-11
U.S. EPA (U.S. Environmental Protection Agency). (1988). Recommendations for and
documentation of biological values for use in risk assessment (pp. 1-395). (EPA/600/6-
87/008). Cincinnati, OH: U.S. Environmental Protection Agency, Office of Research and
Development, Office of Health and Environmental Assessment.
http://cfpub. epa.gov/ncea/cfm/recordisplav. cfm?deid=34855
G-10
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
U.S. EPA (U.S. Environmental Protection Agency). (1991a). Guidelines for developmental
toxicity risk assessment. Fed Reg 56: 63798-63826.
U.S. EPA (U.S. Environmental Protection Agency). (1991b). Guidelines for developmental
toxicity risk assessment (pp. 1-71). (EPA/600/FR-91/001). Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment Forum.
http ://cfpub. epa.gov/ncea/cfm/recordisplav. cfm?deid=23162
U.S. EPA (U.S. Environmental Protection Agency). (1996a). Guidelines for reproductive toxicity
risk assessment. Fed Reg 61: 56274-56322.
U.S. EPA (U.S. Environmental Protection Agency). (I ) (Hiidclines for reproductive
toxicity risk assessment (pp. 1-143). (EPA/630 R-lW Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment I 'oi imi
https://www.epa.irov/sites/producti profiles 2< >2) A review of the reference dose and
reference concentration processes (pp. 1-1^2). (I .IW (\i<)/P-02/002F). Washington, DC:
U.S. Environmental Protection Agency, Risk Assessment Forum.
http://www.epa.gov/osa/re\ ieu -rcl'crcnce-dose-and-ieference-concentration-processes
U.S. EPA (U.S. Environmental Protection Agency) (2<)i)5). Guidelines for carcinogen risk
assessment |T.PA Report] (pp. 1-1 (•>(•>) (EPA (\i<) P-03/0<)| I')- Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment Forum.
http://www2 epa uo\ osa/guidelines-carcinouen-risk-assessment
U.S. EPA (U.S. Environmental Protection Agency). (2006). A framework for assessing health
risk of en\ironmenlal exposures to children (pp. 1-145). (EPA/600/R-05/093F).
Washington. DC: U.S. Environmental Protection Agency, Office of Research and
l)e\elopmenl, National Center for Environmental Assessment.
Imp cl'pub.epa.go\ ncca/cl'm iecordisplav.cfm?deid=158363
U.S. EPA (I S. I-n\ ironmenlal Protection Agency). (2011a). Exposure factors handbook: 2011
edition (final) | l-!PA Report], (EPA/600/R-090/052F). Washington, DC: U.S.
Environmental Protection Agency, Office of Research and Development, National Center
for Environmental Assessment.
http://cfpub.epa uo\ ncca/cfm/recordisplav.cfm?deid=236252
U.S. EPA (U.S. Environmental Protection Agency). (201 lb). Recommended use of body weight
3/4 as the default method in derivation of the oral reference dose (pp. 1-50).
(EPA/100/R11/0001). Washington, DC: U.S. Environmental Protection Agency, Risk
Assessment Forum, Office of the Science Advisor.
https://www.epa.gov/risk/recommended-use-bodv-weight-34-default-method-derivation-
oral-reference-dose
G-ll
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
U.S. EPA (U.S. Environmental Protection Agency). (2012). Benchmark dose technical guidance.
(EPA/100/R-12/001). Washington, DC: U.S. Environmental Protection Agency, Risk
Assessment Forum, https://www.epa.gov/risk/benchmark-dose-technical-guidance
U.S. EPA (U.S. Environmental Protection Agency). (2014a). Child-specific exposure scenarios
examples [EPA Report], (EPA/600/R-14/217F). Washington, DC.
http ://cfpub. epa.gov/ncea/risk/recordisplav. cfm?deid=262211
U.S. EPA (U.S. Environmental Protection Agency). (2014b). Framework for human health risk
assessment to inform decision making. Final [EP A Report] (EPA/100/R-14/001).
Washington, DC: U.S. Environmental Protection. Risk Assessment Forum.
https ://www.epa. gov/risk/framework-human-hea 11 h-iisk-assessment-inform-decisi on-
making
U.S. EPA (U.S. Environmental Protection Agency) (2014c) (Hiidancc for applying quantitative
data to develop data-derived extrapolation factors for interspecies and intraspecies
extrapolation [EPA Report], (EPA/10<) R-14 <)<)2F). Washington, DC: Risk Assessment
Forum, Office of the Science Advisor, https www epa gov/sites/piodLiclion/files/2015-
01/ documents/ddef-final .pdf
U.S. EPA (U.S. Environmental Protection Agency). (2014d) I lealth effects document for
perfluorooctane sulfonate (PI OS). (E-PA K22 R-14/<)<)2) Washington, DC: Office of
Water, Health and Ecological Criteria l)i\ ision
https://peerre\ iew \ ersar com/epa/pfoa/pdl' I lealth-l-riects-Document-for-
Perfluorooclane-Siilfoiiale-(PFOS).pdf
U.S. EPA (U.S. En\ iionmental Protection Agency) (2014e). Health effects document for
perfluorooctanoic acid (PI OA). (EPA/822 R-14/001). Washington, DC: Office of Water,
I lealtli and Ecological Criteria Di\ ision
Imps /peerreview.versar.com/epa/pfoa pelI' I Icalth-Effects-Document-for-
PeillLiorooctanoic-\cid-(PI'()\).pdf
U.S. EPA (I S linvironmenlal Protection Agency). (2014f). Provisional peer-reviewed toxicity
values for perfluorobutane sulfonate (CASRN 375-73-5) and related compound
potassium perfluorobutane sulfonate (CASRN 29420-49-3).
https://hhppit\ oml uo\ issue papers/PerfluorobutanesulfonicacidPFBS.pdf
U.S. EPA (U.S. Environmental Protection Agency). (2017). Unregulated Contaminant
Monitoring Rule 3 (2013-2015) occurrence data. Retrieved from
https://www.epa.gov/dwucmr/occurrence-data-unregulated-contaminant-monitoring-rule
U.S. EPA (U.S. Environmental Protection Agency). (2018). Safe drinking water information
system [Database], Retrieved from https://www3.epa.gov/enviro/facts/sdwis/search.html
Van Den Hove. MF; Beckers. C: Devlieger. H; De Zegher. F; De Naver. P. (1999). Hormone
synthesis and storage in the thyroid of human preterm and term newborns: effect of
thyroxine treatment [Review], Biochimie 81: 563-570.
G-12
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Vulsma. T; Gons. MH; de Viilder. JJ. (1989). Maternal-fetal transfer of thyroxine in congenital
hypothyroidism due to a total organification defect or thyroid agenesis. N Engl J Med
321: 13-16. http://dx.doi.org/10.1056/NEJM1989070632101Q3
Wang. B: Zhang. R: Jin. F: Lou. H: Mao. Y: Zhu. W: Zhou. W: Zhang. P: Zhang. J. (2017).
Perfluoroalkyl substances and endometriosis-related infertility in Chinese women.
Environ Int 102: 207-212. http://dx.doi.Org/10.1016/i.envint.2017.03.003
Wang. Y; Rogan. WJ: Chen. PC: Lien. GW: Chen. HY; Tseng. YC: Longnecker. MP: Wang. SL.
(2014). Association between maternal serum perfluoroalkyl substances during pregnancy
and maternal and cord thyroid hormones: Taiwan maternal and infant cohort study.
Environ Health Perspect 122: 529-534. http://dx.doi.org/10.1289/ehp.1306925
Wasco. EC: Martinez. E; Grant. KS: St Germain. EA; Si Germain. PL. (2003). Determinants of
iodothyronine deiodinase activities in rodent uterus. Endocrinology 144: 4253-4261.
http://dx.doi.org/10.1210/en.2003-04^"
Weaver. Y. iM; Ehresman. DJ; Butenhoff. JL. I laucnbuch. 1} (2010). Roles of ml: renal organic
anion transporters in transporting perfluorinaled carboxylates with different chain
lengths. Toxicol Sci 113: 305-3 14 http://dx.doi oru 1" 1 "93/toxsci/kfp275
Woldemeskel. M. (2017). Chapter 64 Toxicologic palhology of the reproductive system. InRC
Gupta (Ed.), Reproductive and Developmental Toxicology (Second Edition) (2nd ed., pp.
745-763). San Diego. CA: Academic Press http'^dx doi oru 10.1016/B978-0-12-804239-
7.00064-0
Xu. J. (2001). Test lor chemical induction of chromosome aberration in cultured Chinese
hamster ovary (CI l()) cells with and without metabolic activation. (Sponsor's study no.
132-<)()1)) Redliekl. AR Primedica Redfiekl
Yang. I.. I.i. .1. Lai. J: Luan. H; Cai. Z; Wang. Y; Zhao. Y; Wu. Y. (2016). Placental transfer of
periluoroalkyl substances and associations with thyroid hormones: Beijing prenatal
exposure study. Sci Rep (•> 21699. http ://dx. doi.org/10.1038/srep21699
York. RG. (2002) Oral (ga\ age) developmental toxicity study of potassium perfluorobutane
sulfonate (Pl 'liS) in mis ( Argus Research Protocol Number 418023). St. Paul, MN: 3M
Corporate Toxicology
York. RG. (2003a). Oral (salvage) dosage-range developmental toxicity study of potassium
perfluorobutane sulfonate (PFBS) in rats. (Sponsor's study no. T-7485.11). Horsham, PA:
Argus Research.
York. RG. (2003b). Oral (gavage) repeated dose 90-day toxicity study of potassium
perfluorobutane sulfonate (PFBS) in rats. (Argus Research Protocol Number 418026).
Washington, D.C.: U.S. Environmental Protection Agency.
G-13
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
York. RG. (2003c). Oral (gavage) two-generation (one litter per generation) reproduction study
of perfluorobutane sulfonate (PFBS) in rats. (Argus Research Protocol Number 418021).
Washington, D.C.: U.S. Environmental Protection Agency.
York. RG. (2003d). Oral (gavage) two-generation (one litter per generation) reproduction study
of perfluorobutane sulfonate (PFBS) in rats. Appendix D. Report tables - F1 generation
male rats. (Argus Research Protocol Number 418021). Washington, D.C.: U.S.
Environmental Protection Agency.
York. RG. (2003e). Oral (gavage) two-generation (one litter per generation) reproduction study
of perfluorobutane sulfonate (PFBS) in rats. Appendix k 11 isopathology report. (Argus
Research Protocol Number 418021). Washington. DC IS. Environmental Protection
Agency.
Zeng. XW; Qian. Z; Emo. B; Vaughn. M; Bao. .1. Pin. ,\l). /lui. Y. I.i. .T: Lee. YL; Dong. GH.
(2015). Association of polyfluoroalkyl chemical exposure with scrum lipids in children.
Sci Total Environ 512-513: 364-370. Imp dvdoi.org/10.10K-' i scilolcnv.2015.01.042
Zhang. X; Lohmann. R; Dassuncao. C; Hu. XC. Weber. \K. Vccitis. CD. Sunderland. EM.
(2016). Source attribution of poly- and perfluoroalkyl substances (PFASs) in surface
waters from Rhode Island and the New York metropolitan area. Environ Sci Technol Lett
3: 316-321. http://dx.doi.oru I" 102l/acs.estlett.6bO"255
Zhao. Z: Xie. Z: Moeller. A; Sturm. R; Tang. J: Zhang. G. an: l-binghaus. R. (2012). Distribution
and long-range transport of polyfluoroalkyl substances in the Arctic, Atlantic Ocean and
Antarctic coast. Environ Pollut 170: 71-77.
http://dx.doi.org/10.1016/i.envpol.2012.06.004
Zhou. \\ . /hanu. I.. Tonu. C. I'anu. I'. Zhao. S: Tian. Y; Tao. Y; Zhang. J: Study. SBC. (2017a).
Plasma pcrlluoroalky1 and polyfluoroalkyl substances concentration and menstrual cycle
characteristics in preconception women. Environ HealthPerspect 125: 067012.
Iittp d\ doi org I" I2S^ l-lll'IZoj
Zhou. Y; Bao. W W. Qian. X\l. Dee Geiger. S: Parrish. KL; Yang. BY: Lee. YL: Dong. GH.
(2017b) Perlluoroalkyl substance exposure and urine CC16 levels among asthmatics: A
case-control study of children. Environ Res 159: 158-163.
http://dx.doi org I" l"|(i j envres.2017.08.005
Zhou. Y: Hu. LW: Qian. ZM: Chang. JJ; King. C: Paul. G: Lin. S: Chen. PC: Lee. YL: Dong.
GH. (2016). Association of perfluoroalkyl substances exposure with reproductive
hormone levels in adolescents: By sex status. Environ Int 94: 189-195.
http://dx.doi.Org/10.1016/i.envint.2016.05.018
Zhu. Y: Oin. XD: Zeng. XW: Paul. G: Morawska. L: Su. MW: Tsai. CH: Wang. SO: Lee. YL:
Dong. GH. (2016). Associations of serum perfluoroalkyl acid levels with T-helper cell-
specific cytokines in children: By gender and asthma status. Sci Total Environ 559: 166-
173. http://dx.doi.Org/10.1016/i.scitotenv.2016.03.187
G-14
-------
This document is a draft for review purposes only and does not constitute Agency policy.
DRAFT FOR PUBLIC COMMENT - DO NOT CITE OR QUOTE NOVEMBER 2018
Zoeller. RT; Tan. SW; Tvl RW. (2007). General background on the hypothalamic-pituitary-
thyroid (HPT) axis [Review], Crit Rev Toxicol 37: 11-53.
http://dx.doi.org/10.1080/104084406Q1123446
G-15
------- |